STP880 Eb DL.15102 PDF

Copyright by ASTM Int'l (all rights reserved); Fri Feb 8 07:23:51 EST 2019
Downloaded/printed by
Patty Brilman (none) pursuant to License Agreement. No further reproductions authorized.
CORROSION OF METALS
UNDER THERMAL
INSULATION
A symposium sponsored by
ASTM Committees C-16 on Thermal
Insulation and G-1 on Corrosion
and the National Association of
Corrosion Engineers, the Institution
of Corrosion Science and Technology, and
the Materials Technology Institute
of the Chemical Process Industries
ASTM SPECIAL TECHNICAL PUBLICATION 880

Warren I. Pollock, E. I. du Pont de Nemours
and Company, and Jack M. Barnhart,
Thermal Insulation Manufacturers
Association, editors
ASTM Publication Code Number (PCN)

04-880000-27
1916 Race Street, Philadelphia, PA 19103
Library of Congress Cataloging in Publication Data
Main entry under title:
Corrosion of metals under thermal insulation.
(ASTM special technical publication; 880)

Papers presented at the symposium held at San
Antonio, TX, 11-13 Oct. 1983.
Includes bibliographies and index.
"ASTM publication code n u m b e r (PCN) 04-880000-27"
1. Corrosion and anti-corrosives--Congresses.
2. Insulation (Heat)--Congresses. I. Pollock,
Warren I. II. Barnhart, Jack M. III. ASTM Committee
C-16 on Thermal Insulation. IV. Series.
TA462.C6567 1 9 8 5 621.1%23 85-10616
ISBN 0-8031-0416-2
Copyright 9 by AMERICAN SOCIETY FOR TESTING AND MATERIALS 1985

Library of Congress Catalog Card Number: 85-10616
NOTE
The Society is not responsible, as a body,
for the statements and opinions
advanced in this publication.
Printed in Ann Arbor, MI

Aug. 1985
Foreword
The symposium on Corrosion of Metals Under Thermal Insulation was pre-

sented at San Antonio, TX, 11-13 Oct. 1983. The symposium was sponsored
by ASTM Committees C-16 on Thermal Insulation and G-1 on Corrosion and
by the National Association of Corrosion Engineers, the Institution of Corro-
sion Science and Technology, and The Materials Technology Institute of the
Chemical Process Industries. Warren I. Pollock, E. I. du Pont de Nemours
and Company, and Jack M. Barnhart, Thermal Insulation Manufacturers
Association, presided as chairmen of the symposium and are editors of this
publication.
Related
ASTM Publications
Atmospheric Corrosion of Metals, STP 767 (1982), 04-767000-27
Atmospheric Factors Affecting the Corrosion of Engineering Metals, STP

646 (1978), 04-646000-27
Chloride Corrosion of Steel in Concrete, STP 629 (1977), 04-629000-27
A Note of Appreciation
to Reviewers
The quality of the papers that appear in this publication reflects not only
the obvious efforts of the authors but also the unheralded, though essential,
work of the reviewers. On behalf of ASTM we acknowledge with appreciation
their dedication to high professional standards and their sacrifice of time and
effort.
A S T M C o m m i t t e e on Publications
ASTM Editorial Staff
Susan L. Gebremedhin
Janet R. Schroeder
Kathleen A. Greene
Bill Benzing
Contents
Introduction 1
TECHNICAL OVERVIEW
The Function of Thermal Insulation--JACK M. BARNHART 5
T H E PROBLEM
Factors Affecting Corrosion of Carbon Steel Under Thermal

I n s u l a t i o n - - P E T E R LAZAR, III 11
Factors Affecting the Stress Corrosion Cracking of Austenitlc

Stainless Steels Under Thermal Insulation--DALE McINTYRE 27
A Review of the European Meeting on Corrosion Under Lagging Held

in England, November 1980--JAMES A. RICHARDSON 42
THERMAL INSULATION MATERIALS
Thermal Insulation Materials: Generic Types and Their Properties--

GEORGE E. LANG 63
FIELD EXPERIENCE
Experience with Corrosion Beneath Thermal Insulation in a

Petrochemical Plant--TORE SANDBERG 71
Recent Experiences with Corrosion Beneath Thermal Insulation in a

C h e m i c a l P l a n t - - V I C T O R C. LONG AND PAULA G. CRAWLEY 86
Failure of Type 316 Stainless Steel Nozzles in Contact with Fire

Retardant M a s t i c - - 8 . J. MONIZ AND M. C. RITTER 95
External Stress Corrosion Cracking of Stainless Steel Under Thermal

Insulation--20 Years Later--WILLIAM G. ASHBAUGH 103
Shell and Jacket Corrosion of a Foamed In-Place Thermally Insulated

Liquefied Petroleum Gas Tank--DONALD O. TAYLOR AND
Copyright by ASTM Int'l (all rights reserved);
RODNEY D. Fri Feb 8 07:23:51 EST 2019
BENNETT 114
A Study of Corrosion of Steel U n d e r a Variety of T h e r m a l I n s u l a t i o n
Materials--WiLLIAM G. ASHBAUGH AND THOMAS F. LAUNDRIE 121
Behavior of a Copper W a t e r T u b e Exposed to N a t u r a l C a r b o n a c e o u s

G r a n u l a r a n d Cellulosic I n s u l a t i o n M a t e r i a l s - -
JAMES R. MYERS AND ARTHUR COHEN 132
CONTROL MEASURES
Controlling C a r b o n Steel Corrosion U n d e r I n s u l a t i o n - -

PAUL E. KRYSTOW 145
Protective C o a t i n g System Design for I n s u l a t e d or Fireproofed

Structures--PETER A. COLLINS, JOHN F. DELAHUNT, AND
DEBBIE C. MAATSCH 155
P r e v e n t i o n of Chloride Stress Corrosion C r a c k i n g U n d e r I n s u l a t i o n - -

LOUIS C. SUMBRY AND E. JEAN VEGDAHL 165
Designing to P r e v e n t Corrosion of Metals U n d e r I n s u l a t i o n - -

CHARLES T. METTAM ] 78
Use of A l u m i n u m Foil for P r e v e n t i o n of Stress Corrosion C r a c k i n g of

Austenitie Stainless Steel U n d e r T h e r m a l I n s u l a t i o n - -
JAMES A. RICHARDSON AND TREVOR FITZSIMMONS 188
U s i n g Speelfieations to Avoid Chloride Stress Corrosion C r a c k i n g - -

JOHN W. KALIS, JR. 199
Use I n s p e c t i o n as a M e a n s of R e d u c i n g Failures Caused By Corrosion

U n d e r W e t Insulation--HERBERT A. MOAK 204
TEST METHODS
A New A p p a r a t u s a n d Test Procedure for R u n n i n g A S T M C 692

Stress Corrosion C r a c k i n g Tests--FRANCIS B. HUTTO, JR.,
RALPH G. TISSOT, JR., AND THOMAS E. WHITAKER 211
C o m p a r i s o n s of Several Accelerated Corrosiveness Test M e t h o d s for

T h e r m a l I n s u l a t i n g Materials--KEITH 6. SHEPPARD,
SUNIL PATEL, MUKESH TANEJA, AND ROLF WElL 220
Index 231
STP880-EB/Aug. 1985
Introduction
Very serious corrosion problems can occur to plant equipment, tankage,

and piping components that are thermally insulated if the insulation becomes
wet. Many companies have had to repair or replace major pieces of equipment
at considerable expense. At one chemical process plant alone, the cost was re-
ported to be in the millions of dollars.
On carbon steels, the corrosion is usually of a general or pitting type. On
austenitic stainless steels, the corrosion is almost always chloride stress corro-
sion cracking. It is an insidious problem. The insulation usually hides the cor-
roding metal and the problem can go undetected for years until metal failure
occurs. This sometimes occurs five or more years after the insulation becomes
wet.
Insulation materials received from manufacturers and distributors are dry,
or nearly so. Obviously, if they remain dry there is no corrosion problem. So,
the solution to the corrosion under wet insulation problem would appear to be
fairly obvious: keep the insulation dry or protect the metal.
Unfortunately, application of these solutions is not that simple. Insulation
can get wet in storage and field erection. Weather barriers are not always in-
stalled correctly or they are not effective in fully preventing water ingress.
Weather barriers and protective coatings get damaged and are not maintained
and repaired.
To further complicate the problem, it appears that the degree of corrosion
when an insulation gets wet is dependent on the type of insulation. Some insu-
lations contain elements that promote corrosion, such as chloride stress corro-
sion cracking of austenitic stainless steels.
Inspection for the problem is often difficult. Good inspection techniques to
determine that the insulation is wet or that the metal surface is corroded or
stress cracked have not been widely available.
Many companies have developed the practice of applying a protective coat-
ing to steels to keep moisture from contacting the metal. Some do this only for
carbon steels, some only for stainless steels, some for both. What coatings to
use have varied considerably from one plant site to another.
Wet insulation is significantly less thermally efficient than dry insulation.
This alone should be a high driving force for keeping insulation dry, but, inter-
estingly, this has not been the case on many plant sites.
Little has appeared on this overall problem in the literature, and there has
not been a major conference in North America before this one. In Nov. 1980, a
Downloaded/printed by 1
Copyright9 by ASTM International www.astm.org
2 CORROSIONOF METALS
conference was held in Britain on "Corrosion Under Lagging." The success of

that meeting stimulated the organization of a similar type conference in the
U.S.
The purpose of this conference was to provide a forum for a thorough review
of the problem and the various control and inspection methods being used and
under development. Because the problem is broad based, several technical so-
cieties were cosponsors: ASTM Committee C-16 on Thermal Insulation and
Committee G-1 on Corrosion; the National Association of Corrosion Engi-
neers (NACE); Materials Technology Institute of the Chemical Process Indus-
tries (MTI); and Institution of Corrosion Science and Technology, a sponsor of
the British conference.
The conference was very successful with some 150 people attending. It pro-
vided high recognition to a costly problem where the solutions are many faceted,
as indicated by the papers in this publication.
Warren L Pollock
E. I. du Pont de Nemours & Co., Inc.. En-
gineering Department, Wilmington, DE
19898; symposiumcochairmanand editor.
Technical Overview
Jack M. Barnhart 1
The Function of Thermal Insulation
REFERENCE: Barnhart, J. M., "The Function of Thermal Insulation," Corrosion of

Metals Under Thermal Insulation, ASTM STP 880, W. I. Pollock and J. M. Barnhart,
Eds., American Society for Testing and Materials, Philadelphia, 1985, pp. 5-8.
KEY WORDS: corrosion, insulation, energy, chlorides, stress corrosion
Thermal insulations or thermal insulation systems are usually defined as

materials or combinations of materials that retard the flow of heat energy by
conductive, convective, or radiative modes of transfer or a combination of
these. In order to be effective, they must be properly applied.
Primarily, thermal insulation serves one or more of the following functions:
1. Conserve energy by reducing heat loss or gain of piping, ducts, vessels,

equipment, and structures.
2. Control surface temperatures of equipment and structures for personnel
protection and comfort.
3. Facilitate temperature control of a chemical process, a piece of equip-
ment, or a structure.
4. Prevent vapor condensation at surfaces having a temperature below the
dew point of the surrounding atmosphere.
5. Reduce temperature fluctuations within an enclosure when heating or
cooling is not needed or available.
Thermal insulations may also serve additional functions:
1. Add structural strength to a wall, ceiling, or floor section.
2. Provide support for a surface finish.
3. Impede water vapor transmission and air infiltration.
4. Prevent or reduce damage to equipment and structures from exposure to
fire and freezing conditions.
5. Reduce noise and vibration.
1Thermal Insulation Manufacturers Association, 7 Kirby Plaza, Mt. Kisco, NY 10549.
Thermal insulation is used to control heat flow in temperature ranges from

near absolute zero through 1650~ (3000~ and higher. Insulations normally
consist of the following basic materials and composites:
(1) inorganic, fibrous or cellular materials such as glass, asbestos, rock or
slag wool, calcium silicate, bonded perlite, vermiculite, and ceramic products.
(2) organic fibrous materials, such as cotton, animal hair, wood, pulp,
cane, or synthetic fibers, and organic cellular materials, such as cork, foamed
rubber, polystyrene, polyurethane, and other polymers.
(3) metallic or metalized organic reflective membranes (which must face
air, gas-filled, or evacuated spaces).
The structure of mass-type insulation may be cellular, granular, or fibrous,
providing gas-filled voids within the solid material that retard heat flow.
Reflective insulation consists of spaced, smooth-surfaced sheets made of metal
foil or foil surfaced material that derives its insulating value from a number of
reflective surfaces separated by air spaces.
The physical forms of industrial and building insulations are
(1) loose fill and cement,
(2) flexible and semirigid,
(3) rigid,
(4) reflective, and
(5) foamed in place.
Depending on design requirement, the choice of a particular thermal insu-
lation may involve a set of secondary characteristics in addition to the primary
property of low-thermal conductivity. Characteristics, such as resiliency or
rigidity, acoustical energy absorption, water vapor permeability, air flow re-
sistance, fire hazard and fire resistance, ease of application, applied cost, or
other parameters, may influence the choice among materials having almost
equal thermal performance values.
Some insulations have sufficient structural strength for use as load-bearing
materials. They may be used occasionally to support load-bearing floors, form
self-supporting partitions, or stiffen structural panels. For such applications,
one or more of the following properties of an insulation may be important:
strength in compression, tension, shear, impact, and flexure. These tempera-
ture dependent mechanical properties vary with basic composition, density,
cell size, fiber diameter and orientation, type and amount of binder (if any),
and both temperature and environmental conditioning.
The presence of water as a vapor, liquified or solid in insulation will decrease
its insulating value; it may cause deterioration of the insulation and eventual
structural damage by rot, corrosion, or the expansion action of freezing water.
Whether or not moisture accumulates within the insulation depends on the hy-
groscopic properties of the insulation, operating temperatures, ambient condi-
BARNHART ON THERMAL INSULATION FUNCTIONS 7
tions, and the effectiveness of water vapor retarders in relation to other vapor
resistances within the composite structure.
The moisture resistance depends on the basic material of the insulation and
the type of physical structure. Most insulations are hygroscopic and will gain
or lose moisture in proportion to the relative humidity of the air in contact with
the insulation. Fibrous and granular insulations permit transmission of water
vapor to the colder side of the structure. A vapor retarder should therefore be
used with these materials where moisture transmission is a factor. Other insu-
lations having a closed cellular structure are relatively impervious to water and
water vapor. Properties that express the influence of moisture include: absorp-
tion (capillarity), adsorption (hygroscopicity), and the water vapor transmis-
sion rate.
Other properties of insulating materials that may be important, depending
upon the application, include: density; resilience; resistance to settling; per-
manence; reuse or salvage value; ease of handling; dimensional uniformity
and stability; resistance to chemical action and chemical change; ease in fabri-
cating, applying, or finishing; and sizes and thickness obtainable.
In some specific applications, thermal insulation is called on to perform
another function, namely, to retard chloride induced stress corrosion.
An inherent characteristic of austenitic stainless steel is its tendency to crack
at stress points when exposed to certain corrosive environments. The mecha-
nisms of stress corrosion cracking are complex and incompletely understood,
but apparently related to certain metallurgical properties. Chloride ions con-
centrated at a stress point will catalyze crack propagation. Other halide ions
are also suspect.
Chlorides are common to most environments, so great care must be taken to
protect austenitic stainless steels. Water, dust and soil, process liquids, chemi-
cal fumes, even the air in coastal regions, contain chlorides in measurable, and
thus additively significant quantities.
Most thermal insulations will not, in themselves, cause stress corrosion
cracking as may be shown by tests. However, when exposed to environments
containing both chlorides and moisture, insulation systems may act as collect-
ing media, transmigrating and concentrating chlorides on heated stainless
steel surfaces. If, however, moisture is not present, the chloride salts cannot
migrate, and stress corrosion cracking will not take place.
Insulations may also be specially formulated to inhibit stress corrosion
cracking in the presence of chlorides through modifications in basic composi-
tion or incorporation of certain chemical additives. Stress corrosion cracking
is a metallurgical shortcoming of austenitic stainless steel. It is unrealistic to
expect an insulation to overcome this shortcoming. If the conditions are such
that stress corrosion cracking will occur, then, the very best an insulation
could hope to do is delay the inevitable. This is demonstrated by the occurrence
of stress corrosion cracking under insulations that were mostly sodium silicate,
a known " i n h i b i t o r . " Stress corrosion cracking u n d e r insulations is not a sim-

ple insulation problem; to quote from W i l l i a m G. A s h b a u g h of Union Carbide
in a p a p e r p r e s e n t e d to the National Association of Corrosion Engineers:
The inhibition of insulation by the addition of neutralizers or other agents to
the insulation is insufficient protection against externally introduced
chlorides which are the major source of stress corrosion cracking.
and from later in the p a p e r :
The author does not claim that insulation materials cannot or will never
cause stress corrosion cracking, but plant experience and laboratory screen-
ing tests indicate that most insulation materials which remain relatively dry
play only a secondary role in stress corrosion cracking. The real problem in
chemical plants exists as a result of the combination of corrosive atmosphere
and the many types of crevices, joints, and areas where atmospheric chloride
contamination and concentration can occur.
T h e real p r o b l e m a n d need of insulation m a n u f a c t u r e r s is to inform the
users of the "real p r o b l e m " a n d ways to address it. T h e insulation cannot a n d
should not be forced to overcome the shortcomings of austenitic stainless steel
when used in real world environments.
The Problem
Peter Lazar, l i p
Factors Affecting Corrosion of

Carbon Steel Under Thermal
Insulation
REFERENCE: Lazar, P., III, "Factors Affecting Corrosion of Carbon Steel Under Ther-
mal Insulation," Corrosion of Metals Under Thermal Insulation, A S T M S T P 880, W . I.
Pollock and J. M. Barnhart, Eds., American Society for Testing and Materials, Phila-
delphia, 1985, pp. 11-26.
ABSTRACT: The problem of corrosion under insulation is studied by analysis of seven

contributing factors: (1) equipment design; (2) service temperatures; (3) insulation selec-
tion; (4) paints and coatings; (5) weather barriers; (6) climate; and (7) maintenance prac-
tices. The purposes of our study are to identify optimum locations for inspection of insu-
lated equipment and to identify improvements in new equipment design and insulation, as
well as maintenance of existing systems. Emphasis is placed on equipment design features,
operating temperature, and weather barrier deficiencies to guide inspection. Weather bar-
rier maintenance is critical for prevention of corrosion on existing equipment. New insu-
lated equipment designs should provide special details so that the system can be better
sealed against water entry and allow clearance for the specified insulation thickness. On
new equipment or when reinsulating existing equipment, a high quality coating system is
recommended, since touch up opportunity does not exist.
KEY WORDS: corrosion, insulation, paints, coatings, maintenance, inspection, temper-

ature, climate, petrochemical equipment, mechanical design, weather barriers
T h e a m o u n t of c a r b o n steel lost b e c a u s e of corrosion u n d e r i n s u l a t i o n ( C U I )

is d e t e r m i n e d b y (1) wet exposure cycle characteristics ( d u r a t i o n a n d fre-
q u e n c y ) , (2) corrosivity of t h e a q u e o u s e n v i r o n m e n t , a n d (3) failure of protec-
tive barriers ( p a i n t a n d j a c k e t i n g ) . T h e r e are n u m e r o u s controllable factors in
the design, c o n s t r u c t i o n , a n d m a i n t e n a n c e of i n s u l a t e d e q u i p m e n t t h a t affect
t h e variables j u s t m e n t i o n e d , a n d therefore t h e a m o u n t of d a m a g e c a u s e d b y
C U I . I n o u r studies of C U I p r o b l e m s in o u r p l a n t s , seven controllable factors
were identified. T h e y are
IStaff engineer, Exxon Chemical Americas, P.O. Box 241, Baton Rouge, LA 70821.
(1) equipment design,

(2) service temperatures,
(3) insulation selection,
(4) protective paints and coatings,
(5) weather barriers,
(6) climate, and
(7) maintenance practices.
These factors will be examined individually to demonstrate how past common

practices supplied the requirements for corrosion. Understanding how com-
mon practices cause the conditions for CUI is leading to better inspection of
existing equipment and better design of new equipment.
Equipment Design
The design of pressure vessels, tanks, and piping generally includes numer-
ous details for support, reinforcement, and connection to other equipment.
These details include stiffening rings, gussets, brackets, reinforcing pads,
flanges, and so forth. Design of equipment, including these details, is the re-
sponsibility of engineers or designers who use construction codes to assure con-
sistently reliable designs for both insulated and noninsulated subjects. Consid-
eration of the problem of insulating those details and of leaving room for the
insulation is lacking in those codes and in the instructions to the designers;
thus, the equipment is designed like those that would not be insulated. The
weather barrier on such designs is broken frequently because of inappropriate
details for insulated equipment or the lack of space for the specified thickness
of insulation.
The consequence of broken jacketing is that more water gets into the insula-
tion at each exposure cycle, taking longer to dry, cooling the insulated equip-
ment item to temperatures where corrosion is possible, and increasing the
amount of cumulative damage. Some of the equipment details, such as
gussets, actually channel water into the insulation. There are also economic
consequences such as energy inefficiency and construction costs. The ineffi-
ciency of wet insulation is obvious. Also obvious is the cost of insulating equip-
ment not designed for insulation, as one watches insulators cut up insulation
and jacketing and sees needless hours spent installing around complicated
details.
The solution to the factor of equipment design affecting CUI is to take an in-
tegrated approach to the design. Specify the insulation thickness and type,
and the jacketing type before designing the equipment. Define acceptable
"code" details for the weatherproofing type, and specify spacing standards. In
every case, simplify the surface to be insulated.
See Figs. 1 and 2.
LAZAR ON CARBON STEEL 13
FIG. l--Opening in metaljacketing, cut by insulators to accomrnodate piping that was run too
close to a pressure vessel for the specified thickness. Piping has been moved as part of our effort to
rectify deficiencies.
Service Temperatures
Service temperature is important in CUI for two reasons:
(1) higher temperatures allow water to be present against the steel for less
time, but
(2) higher temperatures make the water more corrosive, and paints and
caulking fail sooner.
Generally, equipment that operates below freezing a large fraction of its life is
protected against corrosion; however, attachments to that equipment, which
are not as cold, are vulnerable in the transition out of the vapor barrier into
warm humid air. For the most part, corrosion associated with equipment oper-
14 CORROSION OF METALS
FIG. 2 - - J a c k e t i n g cut b.v btsulators to accommodate I-beam f o r platform support that was
closer to the vessel than the hlsulation thickness specified. Water can run on the f l a n g e a n d enter
the insulation at this point.
ating below freezing temperatures is corrosion outside of, not under insula-
tion. Corrosion of equipment operating between freezing and atmospheric
dew points suffers less localized corrosion, and corrosion rates tend to be lower
because, first, the water temperature is lower and second, because contami-
nants are continuously diluted by condensation; however, since the corrosion
occurs continuously, damage can accumulate almost as quickly as it does
under warm insulation.
Corrosion under warm insulation is more difficult to manage or understand
because of the dryout of entering water. Dryout produces surprisingly corro-
sive conditions on a cyclic basis, as well as less than adequate performance by
many protective coatings on which we often rely. The following is a summary of
some of our observations vis-a-vis warm service.
1. The temperature range of 60 to 80~ appears to account for the greatest
amount of damage, but failures have occurred even on systems operating at or
above 370~ when weatherproofing is poorly maintained.
2. On very warm equipment with relatively small weatherproofing defects,
corrosion will tend to be at points of entry of water where rapid evaporation oc-
curs. As equipment temperatures are reduced or weatherproofing defects get
larger, water is allowed to run to lower points where it is held up to dry more
slowly, or not at all.
3. Annual corrosion damage rates may exceed 1.5 m m . y -1. The cor-
rosivity is partly a function of water temperature itself, but also a function of
concentration of salts carded in with the water, drying out in the same loca-
tions repeatedly.
The temperature on some equipment varies by location, especially on towers.
For example, temperatures can range from more than 80~ on the bottom to
less than 0~ at the top. This produces extremes of corrosion condition on a
single equipment item.
See Figs. 3 through 8.
Insulation Selection
Insulation characteristics most influential on corrosion under insulation are
water absorbancy and chemical contributions to the water phase. While no in-
sulation selection will preclude the possibility of corrosion, some insulation
types leave the system less sensitive to defects in weatherproofing or paint film,
because they are nonabsorbent and chemically benign.
FIG. 3--Corrosion above an insulation support ring and around a small p i p e connection. This
is a vertical d r u m that had been heated with a steam coil at one time. Insulation rhzgs act as a hold
up f o r water entering through deficient top head weatherproofing.
FIG. 4--htsulation for personnel protection has bt the past been stopped about 2 m above
grade. Severe corrosion at the open end of the insulation system (water entry pohtt) is typical of
very hot systems such as steam lines.
Unfortunately, insulation selection has not been based on any consideration

of maintenance costs; rather, it has been based on installed cost versus energy
cost saved.
Other considerations, which are normally neglected, include:
(1) repairability of the insulation. Some has to be removed for inspection,
periodically, while some is accidentally damaged;
(2) effect of absorbency on steel corrosion costs and paint film life; and
(3) credit for cyclic service energy savings on behalf of nonabsorbent insula-
tions, because of less water needing to be boiled away.
FIG. 5--Another example of the effect of corrosion under personnel protection. Note the re-
duction h7 diameter of the pipe where indicated, corresponding to the water entry point.
These kinds of considerations are virtually impossible to model for cash flow
analysis, but they should not be ignored. It is therefore necessary to exercise
judgment when selecting insulation, beyond acceptance of calculated returns
on investments.
Cellular glass has been widely adapted by our plants for use from 150~
down, including low temperature requirements. The main advantage is zero
water absorption and reasonable installation cost. Drawbacks are that the ma-
terial is somewhat prone to breakage, and with rising temperatures, has an un-
acceptable k factor. Theoretically the hydrogen sulfide contained in the cells
would contribute to corrosion when water was present between insulation and
steel, but it is not released from those cells unless they are broken.
Calcium silicate insulation is highly water absorbent, and as such has con-
tributed to much of our corrosion problems at moderate temperatures and on
cyclic services. Some calcium silicate still in service may also have contained
corrosive salts, although this may be corrected for the most part with new
material being produced. The advantages of calcium silicate is primarily in k
factor at elevated temperature versus most block insulation types. To realize
this advantage requires that weatherproofing be in good condition and that the
system should be in steady hot service to keep the insulation dry.
FIG. 6--A steam traced line showing severe corrosion under the tracing tubing. Galvanic cor-
rosion may be a ,factor between copper tubing and steel pipe. when water frequently enters the
system.
Although extensively used in the past, polyurethane foam (PUF) is not a

popular insulation type in our plants at this time for moderate or cold services.
Reasons for this are
(1) vulnerability to damage,
(2) utter dependence on the vapor barrier because of its high level of water
absorbency.
(3) corrosivity of water because of hydrolysis of halogenated flame retard-
ants needed to make the insulation safe in the plant, and
(4) sensitivity to humidity during application.
FIG, 7--A ladder bracket on a tower operated at about 80~ Corrosion at this water entry
point is highly concentrated. Surface was abrasivly cleaned to remove scale from the corrosion
trench around the bracket.
Although a low k factor is claimed as an advantage for PUF, in practice,

humidity during application and water entry during service often degrades this
performance. The main attraction to PUF is low installation cost.
Fiberglass and mineral fiber based insulations are used judiciously in our
plant, primarily where existing equipment is spaced such that only fiberous in-
sulations will fit and do the job. Water absorption is a concern with fiberous in-
sulation, although the absorbancy of these insulations may vary greatly from
product to product.
Summarizing insulation selection and its effect on CUI, the two critical fac-
tors are absorbency, because of the effect on the amount of time required to
dry out, and on wicking tendencies, and chemical contributions to an entering
water phase, which increases its corrosivity in most cases. It is important to re-
emphasize two very important points:
1. Corrosion is possible under all types of insulation. The insulation type is

only a contributing factor.
2. Insulation selection requires consideration of a large set of advantages
and disadvantages in areas of installation and operating economics as well as
corrosion and is by no means a simple decision.
See Fig. 9.
Protective Coatings
Protective coatings, or paint, are extremely important in preventing CUI;
failure of protective coatings is essential before corrosion can occur. In the past
FIG. 8 - - G e n e r a l corrosion under a m a n w a y on a vessel with a moderate operatbtg tempera-

ture. General corrosion is characteristic o[" high water entry capacity compared to heat available
./br dryhtg.
FIG. 9--Cot7"osion u n d e r P U F on an idle unit,
the preyailing attitude has been, that a single coat of primer is adequate, on
the assumption that the weatherproofing never let water into the insulation
system. Consider the nature of the service in which a coating under insulation
serves. From that review some direction can be taken on coating selection.
First, the service is virtually an immersion service. In general, the insulation
environment is wet longer than that on the surface of most uninsulated equip-
ment, once the weather or vaporproofing is breached. Second, under warm in-
sulation the coating is obviously subject to higher temperatures than most
painted uninsulated equipment. Consideration must be given to both chemi-
cal degradation and permeability of the coating. Highly permeable coatings
allow corrosion to start behind the coating, even in the absence of breaks or
pinholes. Finally, many coatings depend on some form of sacrificial inhibitor
or are essentially only that (for example inorganic zinc rich coatings). Zinc rich
coatings have given extremely poor performance in our plants under insula-
tion. The following are possible explanations for that performance:
1. There is the possibility of reversal in the polarity of galvanic couples, with
increasing temperature.
2. Salts carried in and deposited with the water interfere with or destroy the
effectiveness of the inhibitors.
3. The subinsulation environment is not freely ventilated and may not have
adequate oxygen or carbon dioxide for film forming reactions to occur.
In general, our plants prefer a coating system directionally like a tank coat-
ing system, involving epoxy or epoxy phenolics in at least a two coat applica-
tion on an abrasive blast cleaned surface. Selection considerations include
temperature resistance, abrasion resistance, and some service rating for im-
mersion service. For warm insulation in particular, inorganic or organic zinc
rich primers are avoided. Inspection of the surface preparation is critical in
nonideal areas such as welds.
Visual inspection for the purpose of identifying the need to touch up failure
points is not possible. Unless corrosion or insulation failure causes rework of
the insulation entirely, there is no chance to do coatings work for 10 to 15
years, or more. Reluctance to spend resources on a coating system considering
these limitations would be unwise, given the corrosion problems that often
follow.
See Fig. 10.
Weather/Vaporproofing
The outer covering of the insulation system is a critical factor. First, it is the
primary barrier to water that provides the corrosive environment. Second, it is
the only part of the system that can be quickly inspected and economically re-
paired. The importance of desirable equipment design features was mentioned
earlier. The following is a review of barrier properties as factors in CUI.
The purpose of a vapor barrier is to keep both liquid and vapor out of the in-
sulation system. The purpose of a weather barrier, which should be used on
warm equipment, is to keep liquid out, but permit evaporation of any liquid
that gets in. For weatherproofing our standards require a minimum perme-
ance of 115 ng. Pa -1 9s -1 9m -2 measured according to ASTM Test Method
for Water Vapor Transmission Rate of Sheet Materials Using a Rapid Tech-
nique for Dynamic Measurement (E 398). Extensive use of metallic non-
breathing jacketing has probably contributed a great deal to our corrosion
damage. Various types of mastic are applied to the breaks in jacketing sys-
tems, trying vainly to keep water out. With time, these seals are failing because
of temperature limitations of mastics and aging characteristics. Liquid water
entering at these breaks is evaporating in the insulation system with inade-
quate opportunity for vapor to escape.
FIG. lO--Corrosion under fiberous insulation on another idle unit. In service, operating tem-
peratures were high enough that paint was not considered necessary.
At 100~ for example, each kilogram of water is going to produce almost

1.7 m 3 of vapor. Without a permeable weatherproof covering, the dew point in
the insulation will equal the temperature of the hot face, and water will con-
tinually condense on the jacketing to be reabsorbed by the insulation. The
small openings through which the water entered do not allow sufficient exit for
the vapor.
There are other factors besides permeability to consider. These include dur-
ability and maintainability, appearance, contribution to fire protection (that
is, melting point), flame spread resistance, and cost of installation. Like insu-
lation selection, jacketing selection is a hard decision.
As mentioned earlier one of the reasons that the weatherproofing is such an
important factor is that it is the only maintainable part of the system. Since
mastics deteriorate quickly, a frequent schedule of inspection and maintenance

is required. The weatherproofing or vaporproofing cannot be considered to
last as long as the design basis of the whole system, say l0 to 25 years. In prac-
tice it must be maintained every 2 to S years to remain effective.
Climate
Both regional climate and microclimate should be considered factors of cor-

rosion under insulation. Regional climate is important based on the reports of
corrosion that tend to come from the more humid locations, especially where
warm insulation is of concern.
Microclimate has to do with internal plant conditions, such as cooling tower
drift and whether or not it is a salt water system, falling condensation from
cold service equipment, subjection to steam discharges, spillage of process
condensate, and so forth. It is often because of the microclimate to which an
insulated item is subjected, that the worst corrosion and corrosion at elevated
temperatures where it is normally not expected are found. Because of micro-
climate factors, the corrosion may be taking place continuously. The corrosiv-
ity of water that enters the insulation can also be increased, by cooling tower
drift in particular.
Cooling tower drift is generally a very fine mist of water that can be carried
airborne for 100 m or more, downwind. Cooling tower water is generally recir-
culated so that the original salinity is at least two to three times higher than the
water supply. Except in locations very close to cooling towers, cooling tower
drift water does not enter the insulation directly. Instead, it dries on the jacket-
ing surface leaving a film of salts. Subsequent rain washes the collecting sur-
faces, carrying the concentrated salts to the equipment details, which are not
weatherproofed effectively. There, they enter by gravity or by wicking action
(in the case of absorbant insulation). As the cycles continue, salt concentra-
tions continually increase as water is evaporated in the system.
Controlling microclimate, like equipment design, should be considered
early, in this case, when considering equipment layout. In some cases it is pos-
sible to do something to eliminate the offending source after it is discovered.
This is the preferred alternative. In some cases, weatherproofing can be up-
graded to provide protection.
Maintenance Practices
With maintenance practices are included certain inspection practices for

this discussion. As stated earlier, routine maintenance of weatherproofing is
necessary to minimize defects in weatherproofing because of deterioration.
Another critically important aspect is the making of major defects because of
maintenance and inspection habits not oriented towards closure of the system
promptly after work is completed. The use of contract insulation services,
which are not always on hand when mechanical or inspection work is com-
pleted, probably contributes to this habit. In at least one case in our plant,
openings made in the insulation of one of our vessels for ultrasonic inspection
and inspection for external corrosion were never closed and are suspected as a
major cause of later severe corrosion damage at the low points in that insula-
tion system.
At our plant a very strong policy direction has been established that insula-
tion openings will be closed promptly. All mechanical and inspection work
cost estimates are to include insulation repair costs. This was not the case in
the past, when the insulation work was estimated separately, and approved or
not approved. If worked it would be several months and sometimes years later.
By comparison, in some of our units, loss or insulation would mean solidifi-
cation of the process stream. There, a resident insulator follows mechanical
work with immediate insulation repairs, and also pursues general insulation
maintenance at other times. In this particular unit, temperatures are too high
for corrosion as long as the weatherproofing is reasonably maintained. None
FIG. ll--Mechanical work (a nozzle addition) has been completed. The insulation was left as
shown for several weeks before insulation was closed. Note the deteriorating condition of the alkyd
primer on the exposed steel surface.
FIG. 12--Inspection openings and mechanical damage that had been left unrepaired &defi-
nitely until the present program started.
the less, the visible contrast between this and other areas of the plant shows the
benefit of a resident insulator in maintenance of insulation systems.
It should be noted that both the National Board Inspection Code and Ameri-
can Petroleum Institute (API) Pressure Vessel Inspection Code 510 require re-
moval of some insulation at least every five years on all vessels where external
corrosion is possible. In our plant alone we should be inspecting, very roughly
30 or 40 insulated vessels per year this way. The need for insulators in support
of this inspection function should be obvious.
See Figs. 11 and 12.
Conclu~on
In summary, there are numerous factors involved in causing or preventing
corrosion under insulation. They have been grouped into seven categories and
reviewed to show how they influence the three requirements for corrosion: ex-
posure cycle, corrosivity of the water, and failure of coatings. Twelve illustra-
tions were included, showing examples of equipment design details, mainte-
nance and inspection openings, corrosion at entry points in hot insulation
systems, and corrosion in moderate temperature insulation.
Dale Mclntyre 1
Factors Affecting the Stress

Corrosion Cracking of Austenitic
Stainless Steels Under Thermal
Insulation
REFERENCE: Mclntyre, D., "Factors Affecting the Stress Corrosion Cracking of Aus-
tenltic Stainless Steels Under Thermal Insulation," Corrosion of Metals Under Thermal
Insulation, ASTMSTP880, W. I. Pollock and J. M. Barnhart, Eds., American Society for
Testing and Materials, Philadelphia, 1985, pp. 27-41.
ABSTRACT- Fundamental factors affecting the stress corrosion cracking (SCC) of stain-
less steels under thermal insulation will be reviewed. Specific topics are susceptible alloys,
problem temperature ranges, sources of chloride ions, effect of halides other than chlo-
rides, effect of geographical location, effect of potential, pH and buffering agents, mech-
anisms of concentration, and mechanisms of inhibition. Field experience with closed cell
versus wicking type insulation will be discussed. The effectiveness of the weather barrier in
preventing SCC under insulation will be discussed in light of maintenance procedure and
detail design practice.
KEY WORDS: corrosion, austenitie stainless steels, insulation, stress corrosion
T h e practical u r g e n c i e s of p r o t e c t i n g p l a n t e q u i p m e n t have forced corrosion

e n g i n e e r s to base m e a s u r e s to p r e v e n t e x t e r n a l stress corrosion c r a c k i n g
(ESCC) u n d e r i n s u l a t i o n o n a series of a s s u m p t i o n s for which there is n o solid
d a t a base. These a s s u m p t i o n s will b e reviewed, a n d t h e i r effect o n the selection
of preventive m e a s u r e s will b e n o t e d . A series of conclusions, some of t h e m at
odds with past practice, will b e p r e s e n t e d b a s e d on e x p e r i m e n t a l studies con-
d u c t e d in o t h e r areas of corrosion science.
IBattelle-Houston Operations, 2223 West Loop South, Suite 320, Houston, TX 77027.
The Nature of Environment on Stainless Steels Under Insulation

Given that whatever environment exists beneath the insulation can cause
stress corrosion cracking (SCC), and assuming for the moment that this crack-
ing is fundamentally transgranular chloride SCC (an assumption that will be
reconsidered later), the obvious deduction is that the environment operates in
or cycles through the temperature ranges that allow crack propagation and
provides a source of water, a source of chlorides, and a source of oxygen or
other oxidizer. These conditions are necessary for any stress corrosion crack-
ing to occur. The other two necessary conditions, a susceptible alloy and net
tensile stress, come from fabrication and material selection processes.
An electrolyte must be present for cracking to proceed, and water is the most
likely. Two sources of water are endemic to the external surfaces of process ves-
sels: atmospheric moisture and city potable water, either as wash water or fire
water. Which is more likely to find its way under the weather barrier over insu-
lation? Rainwater is the more frequent but many engineers insist that dousing
with wash water or fire water (during deluge system testing) are by no means
unusual.
City potable water is typically near-neutral and oxygen-saturated with chlo-
ride contents varying widely at approximately 150 ppm. Mineral content will
vary widely depending on the source.
Atmospheric moisture, which comes to us as rain, fog, mist, or dew, is per-
haps less variable but more difficult to study than potable water. Considerable
data have been generated by researchers concerned with atmospheric corro-
sion of steels.
It has been estimated that 1 L of rain falling from a height of 1 km washes
326 m ~ of air. As it falls, it absorbs atmospheric gases, becoming significantly
more acidic and approaching a pH of about 5.5 in rural areas. Near industrial
areas the pH of rainwater can approach 4.5. If heavy concentrations of sulfur
dioxide are in the air, the pH may be between 3 and 4 in some local areas [1]. It
will also, of course, become saturated in oxygen as it passes through the air.
As it falls, the rain also sweeps the air of suspended salts. The amount of so-
dium chloride suspended as airborne particles varies widely and depends
strongly on the distance from the seashore. Measurements made in Africa and
Russia indicate that rainwater within 1.609 km (1 mile) of the seashore may
have 100-ppm chloride ion. The persistance of wind-borne salt particles is re-
markable; rainwater has a fairly consistent 10-ppm chloride ion even several
hundred miles inland (Fig. 1) [2]. If atmospheric moisture takes the form of
mist or fog, the concentration of chlorides has been measured at 200 to 400 ppm
within a mile of the coast.
So atmospheric moisture, which could serve as the source of electrolyte for
SCC under insulation, will be an ambient temperature and pressure thin film
of liquid with an oxygen content of 8 ppm, a pH of somewhere between 3 and
5.5, and a chloride content probably ranging from l0 to 100 ppm with excur-
sions to 400 ppm.
MclNTYRE ON STRESS CORROSION CRACKING 29
1000
100
E
o.
e~
10
X
0 t__ i i ~ , J 1 i . . . . . . I _ i , , i a~L.tal 1 I ~ * t l [
0 0.1 1 I0 I00
Distance From Sea, k m
FIG. 1--Sodiurn chloride in rainwater [2].
Sources of Chlorides
When chloride stress corrosion cracking first surfaced in the closed pore in-
sulation (CPI), it was assumed that the source of the chlorides was the insula-
tion itself.
Is this valid? Certainly some of the old magnesia insulations were particu-
larly high in chlorides, and certainly these materials would cause rapid failures
of U-bends in the ASTM Evaluating the Influence of Wicking-Type Thermal
Insulations on the Stress Corrosion Cracking Tendency of Austenitic Stainless
Steel (C 692) test. Since then, most agencies and companies specify chloride
contents in insulation that varies from 5 to 600 p p m [3]. How do these materials
compare with atmospheric moisture as a potential source of chlorides?
If one imagines 0.0929 m 2 (1 ft 2) of a stainless steel vessel surface in the hori-
zontal plane, under a nozzle, perhaps, on a column head, covered with 76.2 m m
(3 in.) thick 2.24-kg/m 3 (14-1b/ft 3) calcium silicate insulation, the amount of
chloride ion per unit area that could be leached out of that insulation is readily
calculated as a function of its original concentration, Such data are shown in
Fig. 2. If the insulation has 600-ppm bulk chlorides, the maximum allowed in
the ASTM and Military (MIL) specifications, then 10 070 mg of chloride
might deposit per square metre of stainless surface if all leachable chlorides
FIG. 2--Potential density of chloride deposits from insulation or rainwater.
ended up on the stainless. On the other end of the spectrum, insulation with
less than 5-ppm chlorides in the bulk might deposit 84-mg C I - / m 2.
These numbers take on considerable significance when considered in the
light of experiments done by Yajima and Arii of Japan's Toshiba Corporation
[4]. At 80~ these workers produced SCC in humidified air on stainless sur-
faces having deposited chloride levels of from 100 to 10 000 mg/m 2. This
would suggest that even insulation with very low levels of chlorides (between 5
and 10 ppm) could still produce SCC if conditions of concentration were right.
How do these figures compare with chlorides transported by rainwater?
Considering the hypothetical 0.0929 m 2 (1 ft 2) of stainless surface, assume
that its weather barrier has been completely ineffective, as they sometimes are
around nozzles or after some aging. Rainfall in the Gulf Coast industrial re-
gions ranges from 1016 to 1524 m m (40 to 60 in.) per year. Taking 1270 m m
(50 in.) as an average, our 0.0929 m 2 (1 ft 2) of insulation would be exposed
to 0.0928 m 2 X 1270 m m / y e a r -- 0.1179 m3/0.0010 m3/L = 118 L/year
(144 in. 2 • 50 in./year = 7200 in.3/61 in.3/L = 118 L/year) of rainfall. On
any facility more than about 1.609 km (1 mile) from the coast this rain would
be expected to contain about 10-ppm chloride ion. The insulation itself cannot
trap more than an absolute maximum of its own volume of this water, which
for a 76.2 m m (3 in.) thick block 0.0929 m 2 (1 ft 2) in area would be about 7 L.
Thus 111 L of rainwater must somehow find its way through the insulation
block per year, and these 111 L will carry 111 000 X 10 • 10 -6 = 1.1 g of
chloride. If this salt is all deposited on the underlying stainless steel this would
result in 11 721 nag of C I - per m 2 per year. On facilities closer to the coast, for
example, marine terminals, up to 117 200 mg/m2/year might be expected.
A 0.0929-m 2 (1-ft 2) rip in the weather barrier may seem unrealistic. Con-
sider, then, a 38.7-cm 2 (6-in. 2) defect. Falling rain must first saturate the en-
tire insulation block, not just the exposed area, before moisture can begin to
drip onto the stainless. Even so, the calculated chloride density is ten times
greater than that from insulation of equal chloride content.
Evaporation of rainwater results in some transfer of deposited chloride back
to the atmosphere. Actual measured rates of chloride deposition on exposed
surfaces range from 10 to 1000 mg/m2/day [2].
These values suggest that chlorides leached from insulation can indeed
cause external SCC if conditions of concentration are right. However, chlo-
rides deposited from the atmosphere are potentially many times more danger-
ous. The density of atmospheric chlorides is higher and, unlike the insulation
itself, the potential supply is infinite.
Many engineers argue that, as a source of chlorides, wash water or water
from testing deluge systems cannot be ignored. Such exposure is difficult to
quantify, since the amount of water impinging on the exposed surface of the in-
sulation will vary widely depending on the interest and enthusiasm of the
worker holding the wash hose. However, again taking 0.0929 m 2 (1 ft 2) of ex-
posed calcium silicate insulation, if 114 L (30 gal) of potable city water are al-
lowed to run into the surface per year, and assuming a chloride content of
150 ppm, the potential maximum density of deposited chlorides is 171 000 nag
C I - per m 2 per year, worse even than seacoast rainwater. However, the actual
amount of chlorides deposited would be significantly less than the calculated
value since the relatively brief, intense duration of washing, or deluge-system
testing, would cause more runoff and less evaporation and concentration than
the more gradual accumulation of moisture from rainfall. To counter this, the
more intense runoff would also result in greater dilution and depletion of any
inhibitor deposited on the surface.
Polyvinyl chloride (PVC) plants and some other chloride-based processes
are of themselves significant sources of chlorides. Resin fines can often be seen
covering vessels and piping to a depth of several millimetres in some PVC
plants; breakdown of this PVC because of water and heat can cause ESCC in
plants sited hundreds of miles from any seacoast.
Sources of Water
The work of Yajima and Arii [4] also brings focus on the second fundamental
assumption that corrosion engineers must make regarding stress corrosion
cracking under insulation. What is the mechanism by which water comes in
contact with the stainless surface? Is it rainwater soaking through the insula-
tion at defects in the weather barrier? Or does condensation or droplet forma-
tion during periods of high humidity play a part? How about runoff from other
areas of the vessel?
The assumed answer to this question (and there are only assumed answers)
is affected by assumptions to the first question, and in turn affects the choice of
preventive measure. If rainwater only is to be feared, and that rainwater must
soak through the insulation to contact the surface, then the insulation system's
efficiency as a weather barrier determines its ability to prevent SCC, and the
use of inhibited insulation is logical.
If, on the other hand, humidified air alone is sufficient to cause SCC, then to
prevent SCC the insulation system must not only function as a weather barrier
but also as a vapor barrier to prevent moisture-laden high-humidity air from
contacting the stainless surface. There was a time, when thick tarry asphaultic
insulation covers were used, when an effective vapor seal may have been
achieved. Today's metal foil weather sheeting, however, was never designed as
a vapor barrier but only as a weather barrier. Thus, if an effective vapor barrier
is necessary, some sort of barrier coating or sacrificial coating will be required.
Yajima and Arii covered thin stainless steel tubes with salt densities varying
from 100 to 10 000 m g / m 2. They then exposed these tubes in humidity cabi-
nets where temperatures could be controlled at 50 to 80~ and relative humidi-
ties from 60 to 80%. They produced SCC at 80~ at all salt densities and rela-
tive humidities down to 60%. At high salt densities they observed SCC at 60~
down to relative humidities of 70%.
These data suggest that in the presence of deposited salt humid air by itself
might be sufficient to cause SCC. How can this be?
Two mechanisms suggest themselves. Droplet formation is possible even at
relative humidities below saturation if the surface temperature is lower than
some value related to the absolute water content.
Figure 3 presents a series of curves showing the temperature difference that
would cause droplet formation at various relative humidities and ambient tem-
peratures. This effect may be significant on many vessels in cyclic service,
100
90
80
70
60
E
50
.=
_= 40
30
20
10
0 i 1 1 i _ L
5 lO 15 20 25
/~,T Degrees, C
FIG. 3--Temperature difference A T between stainless surface and ambient air temperature
.for droplet condensation [1].
whereby droplets could condense when the pipe or vessel is cool, then evapo-
rate when the vessel is heated.
Another mechanism revolves around the hygroscopic nature of salt deposits.
Figure 4 presents data showing the weight gain of sodium chloride at ambient
temperature and various relative humidities as a function of time. Note that
after two weeks at high relative humidities the salt had absorbed over three
times its original weight of water. Such a deposit should cease to be a salt parti-
cle and should become in essence a droplet of saturated brine, with an expected
conductivity on the order of 210 S 9cm -1, a better electrolyte than seawater.
Data from Fig. 4, combined with the results of Yajima and Arii, suggest that
stainless steel with salt deposits (resulting perhaps from simple atmospheric
exposure after erection.and before insulation is applied) might suffer SCC even
under an intact weather barrier if high-humidity air could reach the hot stain-
less surface.
Neither of these mechanisms has been proven to general satisfaction. Cer-
tainly the majority of SCC failures under insulation occur in areas where the
weather barrier has broken down, and where the insulation is actually wet to
the touch. However, such mechanisms are useful for understanding failures
observed on vessels inside buildings, which are never exposed to rainfall, and
the preservice cracking of vessels shipped by sea to high-temperature areas
such as the Middle East.
300
250
o
Q
_= 200
m
@
100
50
0 1" 3 7 14
Duration of Exposure, D a y s
FIG. 4-- Water absorption of salt particles as a function of time and relative humidity [2].
The Role of Potential and Inhibitors

Potential is now known to be of fundamental importance to SCC. The oxygen-
saturated rainfall or condensation to be expected on stainless surfaces under
insulation will have an open circuit potential squarely in the range of cracking.
In both dilute and concentrated chloride solutions the critical potential for
cracking appears to be less negative than about --0.200 VSHE (where SHE is
standard hydrogen electrode) thus cracking can be avoided by forcing the po-
tential more negative than --0.200 VSHE. This can be done either by lowering
the oxygen content of the solution below about 100 ppb or by driving the poten-
tial more negative with cathodic protection or inhibitor additions.
Lowering the oxygen content in the atmospheric environment would not
seem possible, but that is precisely the benefit provided by some very effective
barrier coatings. All organic coatings are permeable to some extent; however,
the nature of the polymers gives great variation in permeability to different
species. For example, the widely used epoxy coatings are fairly permeable to
water. However, they have an extremely low permeability to oxygen and thus
are an effective preventive measure in dilute atmospheres. (Note: epoxies are

sometimes less effective in, for example, PVC plants, where the breakdown
of PVC dust forms hydrochloric acid, which does not need oxygen to cause
cracking.)
Some inhibitors and all cathodic protection systems use the potential shift as
their primary preventive measure. Their effectiveness on an experimental level
is well documented; the life of the systems in the field depends on several de-
tails of application.
For example, sodium metasilicate inhibited insulation is widely used both in
the United States and in Europe. The inhibitor concentration is often specified
as some ratio of the chloride content. Figure 5 reproduces a widely used graph
from Ref 5, data that have been written into ASTM and MIL-I-24244 specifi-
cations. The ratios of inhibitor to chloride are typically at least 10 to 1.
These data were developed on sensitized test specimens, with the acceptance
criterion that not more than one out of four of the samples would be cracked
after 28 days. How many plant managers would accept a failure probability
of 0.25?
The basic assumptions when using inhibited insulation are that the water to
cause cracking will seep through the insulation, leaching out inhibitor as well
as chloride, and that any concentration mechanism that subsequently concen-
trates the chlorides will also concentrate the inhibitor. The first assumption is
1,000
MIL-I-24244
LIMITS
UNACCEPTABLE
CONTE
INHIBITOR
100
c
J ACCEPTABLE
@
J INHIBITOR
CONTENT
k
~ 10
i+
J
1 i i 9 i ,,t j i J ii J,tJ i J I i IJ_ll i I i i t it
10 100 1,000 10,000

Na+BJO3(ppm)
FIG. 5--Inhibitor to hal~des ratios required by MIL specifications and the U.S. Nuclear Reg-
ulatory Agency [5].
certainly not true for vessels in cyclic service, where condensation of droplets
directly on the stainless surface may play a role in cracking. Runoff water,
whether rainwater or wash water, can find its way to the stainless surface at
breaks in the caulking without actually soaking through the insulation. If, as
has been suggested, droplet formation at hygroscopic salts can take place,
once again the water, not having passed through the insulation, will not be
inhibited.
The assumption that inhibitor will concentrate on the hot stainless surface is
worth examining in the light of the extremely high solubility of sodium meta-
silicate in water. Consider again the hypothetical 0.0929 m 2 (1 ft 2) of exposed
insulation covering 0.0929 m 2 (1 ft 2) of stainless steel. If rainwater is allowed to
impinge on this surface, what will happen? After approximately 7 L of rain
have fallen, the insulation will be more than saturated with water. After that,
some of the rain will run off the surface, some will reevaporate back out of the
insulation, and certainly some will make its way through the insulation, carry-
ing with it chlorides and inhibitor.
Inhibitor concentrations in insulation that will pass the ASTM C 692 test
vary from 1.5 to up to 20% by weight of the insulation. In 0.0929 m 2 (1 ft 2) of
76.2 mm (3 in.) thick insulation, there would thus be from 24 to 318 g of inhibi-
tor. The initial 7 L of rainwater it would take to saturate 0.0929 m 2 (1 ft 2) of in-
sulation, would therefore have an inhibitor concentration of from 0.34 to
4.5%. The solubility of sodium metasilicate in water exceeds 5 equivalents per
litre, or 30%. Therefore, would not essentially all the inhibitor go into solution
in that first 7 L of rainwater? Why would there be any sodium metasilicate left
to inhibit the remaining 111 L of rainwater expected to fall on the hypothetical
square foot of insulation in the course of a year?
Assuming that the inhibitor leached from the insulation is reconcentrated
on the surface, very high densities of inhibitor will contact the stainless steel,
from 253 000 to 3.3 million m g / m 2. However, this is not renewable. Chloride
deposition from rainwater near the coast, or from the assumed 0.113 m 3 (30 gal)
of wash water or deluge-test water spilled on the vessel per year, accumulating
at rates as shown in Fig. 2, would result in unsafe inhibitor to silicate ratios of
less than 10 to 1 in about three months with 15 000 original p p m sodium sili-
cate in the insulation. For the highly inhibited 20% sodium silicate material,
the inhibitor to chloride ratio could drop below 10 after something less than
three years. For the more typical case of plants sited more than about 16-km
(10 miles) inland, inhibitor depletion would obviously take longer if rainwater
is the source of the chlorides, from about 2.5 years on insulation with low levels
of inhibitor up to a calculated 25 years at high inhibitor concentrations.
However, this assessment ignores the possibility of the inhibitor being dis-
solved off the stainless surface after deposition. Evaporation at 60 to 80~ al-
though rapid compared to room temperature, is still slow enough to allow
some runoff if enough water seeps through the insulation. Once again, the
MClNTYRE ON STRESS CORROSION CRACKING 37
high solubility of sodium metasilieate in water makes it possible to lose signifi-

cant quantities of inhibitor.
Therefore, the use of inhibited insulation involves not only adequate initial
concentrations of inhibitor but also careful monitoring of these concentrations
as a function of time, with replacement of insulation sections that become de-
pleted. This involves significant maintenance effort for successful long-term use.
Other preventive measures, barrier coatings and sacrificial coatings or foil
wraps, are not limited by the assumption that water must seep through the in-
sulation, or the assumption that the insulation is the predominant source of
chlorides. Barrier coatings and sacrificial methods are designed to protect, re-
gardless of the source of the water or the source of the chlorides.
The success of barrier coatings will clearly depend on defect density. Com-
panies who have tried barrier coatings have quoted widely varying failure
rates, from less than 1 to 75%. Other papers in this publication address this
issue.
The success of sacrificial methods, either metal-loaded coatings, thermal
spray, or foil wrap, will depend on the achievement of continuity between the
sacrificial anode and the stainless cathode through what is undoubtedly an in-
termittent, discontinuous electrolyte. Once again other papers in this session
will address that issue.
Temperature Ranges for SCC

The assumption of a temperature range in which cracking occurs must be
made to decide which items must be protected. There is, unfortunately, uncer-
tainty on both the low and the high end of the temperature range. Some engi-
neers consider equipment that operates above about 121~ (250~ to be safe
since water flashes off exposed surfaces very quickly at such temperatures.
However, a number of failures have been recorded on vessels operating at or
around 121~ (250~ Dillon [6] states that one must go higher, to at least
260~ (580~ where chloride salts lose their water of hydration, before one
can reasonably expect stainless steel to be free of stress corrosion under steady
state conditions. One European firm specifies preventive measures on all
equipment operating below 538~ (1000~ If equipment operates above the
assumed limit for SCC, the effect of temperature cycling during start-up and
shutdown must still be considered.
The lower temperature limit for stress corrosion cracking is more difficult to
establish than the upper limit. The crack growth rate of stainless steel in chlo-
rides increases radically above 80~ But below 80~ the observed threshold
temperature appears to depend on the patience of the observer. Some engi-
neers use 70~ Others use 65, 60, or 55~ depending on their experience and
their willingness to accept risk.
The search for a lower temperature threshold must ignore many cracks doc-
umented in vessels operating at ambient temperatures or below. Many of these

stress corrosion cracks have formed after 20 or 25 years of service, but they
have formed on vessels or piping that were never heated above Texas ambient.
Dillon [7] refers to stress corrosion cracks that occurred at cryogenic tempera-
tures in organic chloride service. It seems clear that there is nothing sacred
about 60 or 55~ that prevents stress corrosion cracking. Rather the lowering
of temperature gradually increases initiation time and reduces the crack pro-
pagation rate of SCC until other operational problems, such as pitting, control
the life.
Along these lines the data of Yajima and Arii [4] are pertinent. Their experi-
ments on thin-walled pressurized American Iron and Steel Institute (AISI) 304
(Unified Numbering System [UNS] $30400) stainless steel tubes were con-
ducted at varying temperatures in a humidity cabinet with sodium chloride on
their surfaces. Cracking occurred rapidly above 70~ with no apparent pit-
ting. At 60~ failure was still by SCC although some pitting was evident. At
50~ with some iron chloride on the surface, pitting predominated although
some SCC was observed on microscopic examination. Thus we see again a con-
tinuum rather than a threshold phenomenon.
Susceptible Materials
It has been clear for 40 years that the austenitic stainless steels have a partic-
ular susceptibility to chloride stress corrosion cracking. The vast majority of
problems have been observed on the 300 series stainless steels, simply because
these alloys make up about 50% of the world's stainless steel production. Fail-
ures have been recorded on equipment of AISI Types 304,304L, 321,347, 316,
and 316L (UNS $30400, $30403, $31600, $31603) stainless steels. Are there
any significant differences in susceptibility between these types? Intuitively
one would expect that the nonmolybdenum bearing grades, with their known
higher susceptibility to the pitting that so often provides the stress raisers from
which SCC propagates, would be more susceptible than the more highly al-
loyed grades. Would the nonmolybdenum bearing grades therefore justify
more extensive protection? Laboratory evidence on this point is contradictory.
Some data collected by operators of PVC plants would suggest that Type 316 is
indeed somewhat more resistant [7]. For the more general diversified plant no
convincing statistics have been advanced to support any significant difference
in susceptibility.
Are the L-grade materials more resistant than regular-carbon materials? No
laboratory researcher has ever been able to show that low-carbon materials re-
sist transgranular SCC better than regular-carbon materials. However, re-
ports of external intergranular SCC are becoming more and more common on
regular-carbon vessels and piping even in ambient temperature applications,
or during shipping and storage. How many SCC failures of regular-carbon
stainless steel near welds have been replaced without analysis on the assump-
tion that it was transgranular cracking and not to the sensitization known to
occur during welding?
One effect that has not been adequately researched is the effect of weld metal
composition on SCC resistance of nearby base metal. Weld rods for austenitic
stainless steels are almost always balanced to give some ferrite on cooling by sili-
con additions. Silicon is known to shift the potential of alloys to which it is added
in the noble direction. Such a shift will not hurt the resistance of the welds them-
selves, which are known to be slightly more resistant than their base metal to
SCC. But what about the galvanic effect on the adjacent base metal? Surely the
effect will be small, no more than 50 or 100 mV. But 100 mV of anodic potential
shift has been shown to profoundly affect the SCC of many alloys in many differ-
ent environments.
Structural Integrity of Cracked Vessels

One last assumption needs to be examined; the assumption that if external
stress corrosion cracking occurs it will take the form of leakage or pinhole
weeping. The great preponderance of field experience suggests that this is
true; the excellent toughness of the 300 series stainless steels is very forgiving.
However, at least three cases of gross loss of strength have been documented.
In one case from a PVC plant, a 304 stainless steel heat exchanger flange be-
came so weakened by external SCC that pieces could be broken off by hand. In
another case, a 203 m m (8-in.), schedule ten, 304L stainless steel pipe suf-
fered a catastrophic rupture approximately 0.6096 m (2 ft) in length upon hy-
drostatic test at 1034 kPa (150 psi) [8]. In a third case, a 0.4572 m (1.5 ft) long
piece of stainless steel simply fell out of a 304L stainless steel resin slurry tank
in a PVC plant [7].
None of the events described above caused serious losses. However in lethal
service any of these three incidents could have caused a major release. They
serve as a warning that the structural integrity of stainless steel vessels or pip-
ing with extensive stress corrosion cracks cannot be taken for granted.
Discussion
The considerations set forth above suggest a modification or supplementa-
tion to traditional methods for testing insulation regarding ESCC. In particu-
lar, the SCC preventive system must be addressed rather than just the insula-
tion. One suitable test, first proposed by Ashbaugh in 1967, would retain the
use of heated U-bends but would introduce test solution to seep through the in-
sulation from the top rather than wick up through the bottom. The test solu-
tion would contain 100-ppm chloride as sodium chloride and be acidified to a
pH of 4.5. The ratio of the total test solution volume to the volume of the insu-
lation block should be at least 10 to 1; the report of test results should include a
phenolthalein spot test for sodium metasilicate inhibitor on both the stainless
surface and in the insulation itself. Test solution drip rate and total test time
could be determined by round robin testing or a project undertaken by a group
industry association such as MTI or ASTM.
Some of the engineering concerns regarding stress corrosion cracking under
insulation have never been satisfactorily resolved. Major operational questions
remain, including:
1. How can a major vessel or piping run be inspected economically for ex-
ternal stress corrosion cracking without completely stripping the insulation
during a total shutdown?
2. If stress corrosion cracks are located by dye-penetrant, how can they be
characterized as to depth?
3. How can the structural integrity of a vessel with extensive stress corrosion
cracking be analyzed to either assure against catastrophic failure, or justify
replacement?
4. How can existing stress corrosion cracks be arrested until a production
window allows repair or replacement?
These are vital questions for plants trying to live with aging vessels and piping
runs. Approaches have been formulated to address each of these concerns;
perhaps they might be worthy of a symposium such as this in future years.
Conclusions
Study of available data, which are by no means complete, suggest the follow-
ing conclusions:
1. Even insulation with only 10-ppm chloride might cause chloride densi-
ties high enough to support stress corrosion cracking if the chlorides are
leached out and deposited on the stainless surface.
2. However, the density of potential deposits of environmental chlorides
(from rainwater, droplet condensation, or wash water) is 10 to 100 times greater
than leachable chlorides in the insulation itself. Environmental chlorides
would therefore appear to be more of a threat than chlorides in insulation.
3. Although the most common mode of failure is leakage or seeping, cata-
strophic failure of stainless steel vessels stress-cracked under insulation is pos-
sible and has been documented.
4. Although the most common source of the electrolyte to support cracking
appears to be water ingress (of rain, wash water, or fire water) at defects in the
weather barrier, stress corrosion cracking under intact weather barriers is pos-
sible on vessels in cyclic temperature service or on surfaces previously con-
taminated with hygroscopic salts.
5. A successful long-term preventive measure for stress corrosion cracking
must be able to handle chlorides both from the insulation and from other ex-
ternal sources, and must be able to cope with water contacting the stainless
McINTYRE ON STRESS CORROSION CRACKING 41
surface after e i t h e r s e e p i n g t h r o u g h t h e insulation, c o n d e n s i n g on t h e surface

as droplets, or r u n n i n g a l o n g t h e surface f r o m r e m o t e access points.
References
[1] Rocenfeld, I. L., Atmospheric Co~Tosion of Metals, National Association of Corrosion Engi-
neers, 1972, p. 97, 203.
[2] Rocenfeld, I. L., Atmospheric Corrosion of Metals, National Association of Corrosion Engi-
neers, 1972, p. 98.
[3] Nickolson, J. D., "Application of Thermal Insulation to Stainless Steel Surfaces," Bulletin of
the Institution of Corrosion Science and Technology, Vol. 19, No. 5, Oct. 1981.
[4] Yajima, M. and Arii, M., "Chloride Stress Corrosion Cracking of AIS1 304 Stainless Steel in
Air," Materials Performance, Dec. 1980, pp. 17-19.
[5] Gillett, J. and Johnson, K. A., "An Experimental Investigation inot Stress Corrosion Crack-
ing of Austenitic Stainless Steel Under Insulation," Corrosion Under Lagging, Institute of
Corrosion Science and Technology, Newcastle-upon-Tyne, United Kingdom, Nov. 1980.
[6] Dillon, C. P., Guidelines.for the Prevention of Stress-Corrosion Cracking of Nickel-Bearing
Stainless Steels and Nickel Alloys, MTI Manual 1, Materials Technology Institute, 1979.
[7] "Report on Vinyl Chloride Safety Association Meeting," Sept. 1983, Toronto, Canada.
[8] Dillion, C. P., Ed., Forms of Corrosion: Recognition and Prevention, National Association of
Corrosion Engineers, 1982, p. 66.
James A. Richardson 1
A Review of the European Meeting

on Corrosion Under Lagging Held in
England, November 1980
REFERENCE: Richardson, J. A., "A Review of the European Meeting on Corrosion Un-
der Lagging Held in England, November 1980," Corrosion of Metals Under Thermal
Insulation, A S T M STP 880, W. I. Pollock and J. M. Barnhart, Eds., American Society
for Testing and Materials, Philadelphia, 1985, pp. 42-59.
ABSTRACT: The Meeting on Corrosion Under Lagging was sponsored by the Institution
of Corrosion Science and Technology, the Thermal Insulation Manufacturers' and Sup-
pliers' Association, and the Thermal Insulation Users' Liaison Group. Ten papers were
read representing the interests of users, suppliers/contractors, consultants, and inspec-
tion engineers, and the abstracts of these papers are recorded in the appendix of this
paper. Additionally, the themes, lessons, or perspectives arising out of the meeting are
reviewed in the areas of corrosion phenomena/case histories; corrosivity of wet insula-
tion/fireproofing; corrosion prevention using inhibited laggings, organic coatings, metal-
lic foils, and paints and alloy fabrication/selection; and corrosion prevention through de-
sign, specification, inspection, and maintenance. Areas where a consensus emerged are
highlighted.
KEY WORDS: carbon steels, low alloy steels, austenitic stainless steels, calcium silicates,
insulation, cellular glass, ceramic fibers, mineral wool, corrosion, stress corrosion, or-
ganic coatings, aluminum foil, stainless steel foil, specifications
T h e o r i g i n a l i m p e t u s for t h e M e e t i n g o n C o r r o s i o n L a g g i n g c a m e f r o m t h e
T h e r m a l I n s u l a t i o n U s e r s L i a i s o n G r o u p in t h e U n i t e d K i n g d o m . T h e y w e r e
c o n c e r n e d t h a t a c o m m u n i c a t i o n g a p h a d d e v e l o p e d b e t w e e n t h e t h e r m a l in-
s u l a t i o n s u p p l i e r / c o n t r a c t o r a n d t h e u s e r on t h o s e a s p e c t s of s p e c i f i c a t i o n s
r e l a t i n g to c o r r o s i o n c o n t r o l . S u p p l i e r s a n d c o n t r a c t o r s w e r e not in all cases
f a m i l i a r w i t h t h e v a r i o u s c o r r o s i o n m e c h a n i s m s t h a t c a n o p e r a t e u n d e r lag-
g i n g systems, a n d w e r e t h e r e f o r e n o t n e c e s s a r i l y " s y m p a t h e t i c " t o w a r d s cor-
r o s i o n c o n t r o l s p e c i f i c a t i o n s " t h r u s t u p o n t h e m " by t h e user, with c o n s e q u e n t
J Senior corrosion engineer, Engineering Department, North East Group, Imperial Chemical
Industries PLC, P.O. Box 6, Billingham, Cleveland TS23 1LD, United Kingdom.
RICHARDSON ON CORROSION UNDER LAGGING 43
effects on achieved standards of application. On the other hand, user corro-

sion engineers were too inclined to view corrosion in terms of one aspect of the
lagging system rather than the system as a whole and had thus contrived to
produce an almost bewildering array of corrosion control specifications,
which in some cases entertained an unrealistic concept of what was achievable
under site application conditions. It followed that the meeting's objectives
were as much mutual education or information as breaking new ground in the
technical sense. It was hoped that it would provide a state of the art review of
current European practice and stimulate productive dialogue between the
various interested parties.
Given the objectives of the meeting, it seemed appropriate that it should be
organized by the two bodies representing corrosion engineering and the insu-
lation supply/contracting industry in the United Kingdom, and the task was
taken on jointly by the Institution of Corrosion Science and Technology and
the Thermal Insulation Manufacturers' and Suppliers' Association.
The meeting was a considerable success. Most European countries were
represented in the more than 100 delegates who attended. Ten papers were
read representing the interests of users (4 papers), suppliers/contractors (3
papers), consultants (2 papers), and inspection agencies (1 paper). The ab-
stracts of the papers, where available, are provided in the Appendix of this
paper.
This paper attempts to review the two-day meeting, and in particular any
themes, lessons, or perspectives arising out of discussion. It is inevitably "one
man's view" of a wide ranging meeting and may well reflect some of the bi-
ases/preferences of that man, albeit the Chairman of the Organizing Com-
mittee. Rather than review papers individually, it is considered more appro-
priate to review the meeting in terms of a number of specific technical topics
that arose when discussing corrosion under fireproofing and thermal insula-
tion systems. Individual papers are referred to by their number in the en-
closed Appendix. It is inappropriate in a review such as this to draw conclu-
sions on specific technical issues. However, those areas where a large measure
of consensus prevailed at the meeting are indicated.
Corrosion Phenomena and Case Histories

In the event, the meeting restricted itself to consideration of carbon, low
alloy, and stainless steel substrates. Accepting that water can penetrate lag-
ging or fireproofing systems, Paper 2 anticipated the following possible prob-
lems, and described their phenomenology/mechanisms in basic terms.
l. Carbon/Low Alloy Steels--The most significant problem is likely to be
pitting or general corrosion. These materials are normally "passive" in alka-
line environments, and might thus not be expected to corrode in, for example,
water extracts from cementitious fireproofing, or calcium silicate/magnesia
based thermal insulants. However, specific anions, notably the chloride ion,
which can arise either from the materials themselves, or as airborne or water-
borne contaminants, are able to break down passivity locally and initiate pit-
ting corrosion. If penetration by acidic airborne or waterborne contaminants,
for example, sulfur or nitrogen oxides, is possible, or if water extracts from
the lagging materials are acidic as in the case of certain organic cellular
foams, then general corrosion is likely to ensue. Certain specific air or water-
borne contaminants, notably the nitrate anion, can give rise to external stress
corrosion cracking of nonstress-relieved systems, particularly if a cyclic wet-
ting or drying concentration mechanism prevails. The consensus was that a
plant operating continuously or intermittently within the temperature range
approximately --5 to 200~ is at risk, although opinions varied on the limits
(particularly the upper limit) and on the definition of "intermittently."
2. Stainless Steels--By far the most significant problem is external stress
corrosion cracking of austenitic stainless steels. This is caused by chloride
ions that originate either in the insulation or fireproofing materials them-
selves, or as airborne or waterborne contaminants. The simple expedient of
thermal stress relief, normally so successful in preventing stress corrosion
cracking (SCC) of carbon or low alloy steel systems, is not normally practica-
ble in austenitic stainless steel systems. Surprisingly small levels of chloride
are sufficient to cause the problem, particularly if a cyclic wetting or drying
concentration mechanism prevails.
The consensus was that a plant operating continuously or intermittently
above approximately 60~ is at risk. There was some divergence on the signif-
icance of the risk at temperatures at approximately > 200~ on the grounds
that surfaces normally hotter than this can spend but a fraction of their life
wet. However, there was a consensus that most significant problems had oc-
curred on surfaces operating normally at temperatures < 200~
Case histories of specific carbon steel corrosion problems were presented in
Paper 8, and other cases were cited in the discussion. Corrosion beneath ce-
mentitious fireproofing had been experienced at locations in the Caribbean
Sea, United Kingdom, United States, Singapore, Japan, and Saudi Arabia.
Corrosion had usually, but not always, been exacerbated by the presence of
sea salts caused by either their inclusion in the originally cast concrete or by
their progressive accumulation from the atmosphere at coastal sites. In one
case, 50% metal loss had been experienced in 16-years service. Corrosion un-
der thermal insulation had been experienced under a range of lagging materi-
als at different temperatures. In one case, significant corrosion had been ex-
perienced after 12 years under a cold lagging system (--5~ on some tanks,
particularly on the roof and the first two shell courses. The insulant was in-
situ foamed polyurethane under a galvanized steel vapor jacket. In another
case, corrosion of a carbon steel heavy fuel oil pipeline operating at approxi-
mately 120~ beneath preformed calcium silicate insulation, metal clad, had
resulted in hydrocarbon leakage and a significant fire.
Case histories of external stress corrosion cracking of austenitic stainless

steels were presented in Paper 3, and again others were cited in the discus-
sion. The major case study concerned failure of brewery vessels and piping
systems operating at < 100~ Major vessel failures had resulted from flood-
ing (because of inadequate weatherproofing) of polyurethane foam lagging
containing chlorinated organic fire retardants, hydrolysis of which yielded
significant quantities of free chloride. Other failures were reported under cal-
cium silicate, and under virtually chloride-free insulants including fiberglass
blanket and asbestos. The majority of failures reported were from the brew-
ing or dairy industries, probably because of the large population of welded
stainless steel equipment from these industries operating in the most critical
temperature range. In this context, an interesting paper from New Zealand
[1] was submitted for discussion at the meeting (Fig. 1). It reported on 77
cases of stainless steel corrosion from all sectors of New Zealand industry.
Figure 1 is taken from the paper. Many of the "environmental cracking"
cases reported were, in fact, cases of external stress corrosion cracking under
lagging, and their impact in terms of proportion of total failures and failure
costs is self evident.
Corrosivity of Wet Insulation or Fireproofing

Two issues were discussed at the meeting:
(1) the composition of the wet lagging~fireproofing environment in rela-
tion to its corrosivity and
(2) the physical transport of aqueous fluids through lagging or fireproofing
systems.
Fireproofing was discussed in Paper 8. Cementitious coatings furnish an
alkaline environment, pH 11 to 12, which would normally be passivating to-
wards steel. Soluble chloride is the usual promoter of corrosion, and it arises
either from the use of calcium chloride accelerator or antifreeze or salt water
in the preparation of the mix, or as an airborne or waterborne contaminant.
Chloride levels as low as 700 ppm may prove sufficient to initiate corrosion.
Penetration of concrete occurs either through the matrix, where an excessive
water/cement (W/C) ratio has been used, through cracks associated with
shrinkage, thermal fluctuations, or mechanical damage, or at junctions with
steel equipment.
The wet lagging environment was addressed specifically in Papers 1, 3, 4,
7, and 8, and arose in discussion. A variety of analytical data were presented
from leaching tests on lagging materials undertaken at varying water/solid
ratios and temperatures. Unsurprisingly, there was considerable scatter in
the data, but the following consensus was identifiable:
1. Calcium silicate, fiberglass, cellular glass, and ceramic fiber furnish
neutral to alkaline environments when wetted, with pH values in the range 7
t J U A L I T Y OF SIAINLEbS STEEL
NIL FAULIY DESIGN
FAULTY
INSTALLATION
IP
FAILURE TYPES
Q U A L I T Y O F S T A I N L E S S STEEL - N I L F A U L T Y A S S E S S M E N T OF T H E
NMEN_AL CONDITIONS $19 2 O O
F A U L T Y D E S I G N - $1050
~ I~, ~F
T . _ N V I R vC~
TY I N S T A L L A T I O N - $5050
I// -"% "~ ~176 - $1 1

/ I II \ FAULTYUSAGE $150 -
/ II/ l.z AULTYw~176

FAULTY DESIGN AND
F A B R I C A T I O N - $37 6 0 0
ENVIRONMEN'[AL CRACKING DUE ]O THE

COMBINED EFFECTS OF STRESS A N D
EVAPORATION OF SALTS ON THE
FIG. 1--Categories under which stainless steel failures can be grouped and the costs of corro-
sion within these categories [ 1].
to 11. Cellular glass is free of soluble chloride. The other materials can con-
tain chloride. For example in the United Kingdom, The British Standard for
Thermal Insulating Materials (BS 3958, Part 1), for preformed calcium sili-
cate allows "up to approximately 500-ppm chloride, and a figure of 20-ppm
chloride on a water extract from calcium silicate" was quoted in discussion.
2. Mineral wool furnishes essentially neutral environments when wet, typi-
cally pH 6 to 7. Chloride levels are low, one figure of 2 ppm on a water extract
being quoted.
3. W a t e r extracts f r o m organic f o a m s can b e a p p r e c i a b l y acidic, p H values

as low as 2 to 3 b e i n g readily achievable. A d d i t i o n a l l y , where h a l o g e n a t e d fire
r e t a r d a n t species (typically h a l o g e n a t e d p h o s p h a t e esters) have b e e n i n c l u d e d
in the f o a m , water extracts can show very high levels of free halide, u p to
t h o u s a n d s of p p m , d e p e n d e n t u p o n the degree of hydrolysis achieved u n d e r
the test conditions, principally, t i m e a n d t e m p e r a t u r e .
The d a t a p r e s e n t e d in P a p e r 8 were as representative as any, a n d T a b l e s 1
t h r o u g h 3 are r e p r o d u c e d directly f r o m the p a p e r . C a r b o n steel corrosion
rates as high as 15 to 20 m p y (where m p y is mil p e r year) were r e p o r t e d for
organic f o a m a q u e o u s extracts, a n d some of the m o r e significant stainless
steel e x t e r n a l SCC p r o b l e m s r e p o r t e d at the m e e t i n g were associated with
f i r e - r e t a r d e d organic f o a m insulations.
TABLE 1--Properties of room temperature aqueous extracts from

polyurethane and phenolic foams [3].
Properties Phenolic Foam Polyurethane Foam
pH free water 3.45 6.1

pH water in foam ." "
Water pickup, g 13" 0".i
"Insufficient water absorbed by the foam after five days to allow pH
measurement.
TABLE 2--Properties of boiling water extracts from polyurethane and

phenolic foams [31.
Properties Phenolic Foam Polyurethane Foam
pH free water 2.37 8.4

pH water in foam 2.25 4.3
Water pickup, g 154 60
TABLE 3-- Physical properties of aqueous extracts from various thermal insulations [3].
Temperature,/z~" Halogens, ppm
Insulations RT 120~ 210~ CI- Br-
Polyurethane (FP) 30 45 100 20 390

Polyurethane 40 50 400
Calcium silicate 200 350 450
Mineral wool 75 700 . ,
Cellular glass 40 60 "300

Fiberglass 220 850 1200 . .
Ceramic blanket 25 40 100
"t~ = (t~ -- 32)/I.8. RT is room temperature.
It was pointed out in Paper 1 that excessive concern about the chloride
content of the lagging material per se in relation to the risk of stress corrosion
cracking of austenitic stainless steels avoids to some extent the real problem,
which is the total chloride available to the lagged surface during its lifetime.
Chloride from all sources must be considered, including that accumulated
during storage and installation, and that available from the atmosphere,
rainwater or wash water, and other liquid contaminants in service. Some in-
teresting data from the nuclear industry were quoted and are presented in
Table 4. There was little recognition of this problem in lagging system specifi-
cations, which tended to concentrate on the lagging material chloride con-
tent. Paper 1 reported the following references to chloride content in a sample
of 20 specifications from the oil, petrochemical, marine, and power genera-
tion industries in the United Kingdom:
1 specified less than 6 ppm
2 specified "low chloride content"
13 gave no specific limit
The physical properties of the various lagging systems in relation to fluid
transport were also discussed. There was a consensus that calcium silicate has
unfavorable "wicking" properties, and that closed cell foam glass is relatively
impermeable. There were differing views on the degree of permeability of or-
ganic foams. Regardless of the intrinsic permeability of the lagging material,
the role of joints in relation to the permeability of the lagging system was
emphasized.
Corrosion Prevention: Inhibited Laggings

The basic technology supporting the use of inhibited lagging to prevent
stress corrosion cracking of stainless steels was covered in Papers 1 and 7.
Figure 2 is taken from Paper 7 and shows the well known Karnes [2] graph,
together with some superimposed constraints applied by three American spe-
cifiers. Proponents of this approach to corrosion prevention argued the bene-
fits of a relatively high absolute sodium plus silicate level in preventing crack-
ing. Such materials are more tolerant to additional chloride ingress from
extraneous sources than those with lower absolute sodium plus silicate levels,
which although initially well within the acceptable Karnes' limits, are readily
affected adversely by small additional quantities of chloride.
Some contributors expressed concern about exclusive reliance on water sol-
uble inhibitors to prevent cracking. Over the lifetime of a lagging system,
progressive water ingress can obviously deplete inhibitor levels below those
necessary for cracking prevention. It was pointed out in Paper 1 that stress
TABLE 4--Measured/available sur/ace contamination levels of

chloride on stainless steel./bil [4].
Condition Chloride Level, g 9m 2
As supplied 0.0002 (measured)

6 months exposure in workshops 0.0003 (measured)
Finger marked during handling 0.0011 (measured)
Covered with finger prints 0.0054 (measured
Bead of perspiration containing approximately 1.0
1000-ppm CI (available)
100-mm insulation, density approximately 1.0
100 kg 9m 3, containing (available)
100 ppm CI
10 3
4 4- /
CORROSIVE . . . . . .
3z4. ~ MAX.o Cl o
ZONE 4-
.~, ='//; ~,oO =
~ .{/~ . .
10 2 & 4- o o
Q.
i I
m
~176176
oo ~ i ~ NON-CORROSIVE
10 o i 4 a/a ZONE
,~, ,"
i 0_ 1,~, l 0/~ 0
~n
* o ~
~,~ ~,~-
J
1
i ' 1013 i
I0 10 2 10 ~I 10 5
SODIUM + SILICATE, ppm
FIG. 2 - - K a r n e s ' d a t a [2] relating incidence of cracking to silicate/ehloride ratio of lagging

material. Numbers denote samples cracked o u t o/~[bur, except where stated. Three typieal speci-
fication constraints are indicated. Presented originally in Paper 7.
corrosion cracking has been experienced on stainless steel equipment insu-

lated with amosite asbestos bonded with as much as 20% sodium silicate, that
is, well above the sodium plus silicate to chloride ratio considered safe in
terms of the Karnes' criteria.
No experience was reported on the direct application of sodium silicate to
stainless steel surfaces before lagging application, a practice which is appar-
ently favored in the United States.
It was emphasized that inhibited laggings had been developed with the spe-
cific objective of combatting stress corrosion cracking of stainless steels. No
claims were made or any experience reported of their efficiency in relation to

general corrosion of carbon steel under laggings.
Corrosion Prevention: Organic Coatings

The scope for organic coatings in preventing corrosion was covered specifi-
cally in Papers 1, 4, 5, and 8.
Regarding carbon steel corrosion, there was a consensus that it is beneficial
to apply coatings beneath fireproofing, and beneath lagging systems on steel
surfaces operating within the susceptible temperature range cited above.
Generally practice appeared to favor the use of relatively simple finishing sys-
tems applied to wire brushed surfaces, and among the coatings cited were red
lead, zinc chromate, and epoxy or phenolic priming systems, dependent on
upper operational temperature. However, in some cases, more sophisticated
systems had obviously been felt appropriate, and in the specific case of steel in
concrete where a bond is required, the use of epoxy based systems was fa-
vored. Paper 8 also outlined the benefits of coating the concrete itself to mini-
mize water diffusion, and successful experiences with water-based vinyl co-
polymers and acrylic emulsions were cited.
A number of operators cited their use of coatings to control external stress
corrosion cracking of austenitic stainless steels. Favored coatings were those
with relatively good high temperature properties, and included silicone-al-
kyds and aluminum filled silicones. Some interesting data were presented in
Paper 4 relating to the protection efficiency of a number of commonly speci-
fied coating systems, and Table 5 is taken from the paper. The data confirm
that relatively cheap systems using single coats of paint on degreased surfaces
markedly reduced the incidence of cracking relative to bare surfaces, but are
only as efficient as they are free of holidays and impervious. In additional
tests, the silicon-alkyd paint proved no more efficient when cured before im-
mersion than when uncured. Evidently no galvanic benefit derives from the
use of aluminum-filled silicone paint. One case was reported concerning the
TABLE S--Incidence of stress corrosion cracking on coiled 304 spring specimens in boiling
saturated sodium chloride solution at 108 ~ C."
Corrosion Total N u m b e r Protection

Potential, of Cracks Efficiency,
Protection System mV/SCE~ 4 Specimens %
None (control) - 380 75
Silicone-alkyd paint, uncured -- 140 8 "89--
Aluminum-rich silicone paint --390 8 89
Zinc-rich epoxy paint --720 2 97
Aluminum foil -- 910 0 100
"Presented originally in Paper 4, see Appendix.
bPotentials recorded at the test temperature of 108~ SCE is saturated colomel electrode.
successful use of an epoxy paint system to control stress corrosion cracking

experienced during coastal site storage in the Middle East.
Stress Corrosion Cracking Prevention: Metallic Foils/Paints

The use of foils, typically 46 standard wire gage (swg), aluminum foil be-
neath lagging systems to prevent stress corrosion cracking of stainless steels
was covered in Paper 4. The foil apparently acts as a physical barrier to the
migration of small quantities of aggressive fluid towards stainless steel sur-
faces, and provides cathodic protection in "flooded" lagging systems, pre-
venting pitting/cracking initiation. Table 5 from the paper summarizes some
laboratory data that confirm the galvanic protection afforded by the foil, and
that it is more efficient than single-coat paint systems in reducing the risk of
stress corrosion crack initiation. The foil can be used on surfaces operating at
temperatures up to 500~ and is applied by simply wrapping around pipes or
vessels with overlays arranged to shed water.
The use of stainless steel foil as an alternative to aluminum foil was also
reported. It has the advantage of being usable at temperatures > 500~ but
acts strictly as a physical barrier, and can provide no galvanic protection. One
operator reported using soft iron foil at temperatures >500~ which can
provide some galvanic protection in the event of flooding of the lagging system
at lower temperatures. Both stainless steel and soft iron are more difficult to
apply than aluminum foil.
Paper 1 presented some data concerning the preferences of individual oper-
ators for foil or paint coatings to prevent stress corrosion cracking. The 20
specifications referred to above yielded the following:
7 specified paint coatings
5 specified foil
3 specified either paint or foil
5 made no reference
The same supplier also observed that in their experience, approximately 90%
of stainless steel surfaces are protected with foil, and approximately 10% with
paint coatings.
Two major concerns regarding the use of aluminum foil were voiced in dis-
cussion. The first concerned the corrosion resistances of the foil, which is < 5
rail thick. Both Papers 4 and 7 reported maintained protection efficiencies
despite a considerable degree of perforation of the foil, although it was recog-
nized that prolonged flooding could result in virtual removal of the foil. The
second concern related to the risk of liquid metal embrittlement in the event
of fire. This was also a major concern in relation to the use of zinc-rich coat-
ings, which are known to be efficient at preventing stress corrosion cracking
caused by the galvanic protection imparted to the substrate, as indicated in
Table 5.
The general p r o b l e m of liquid m e t a l e m b r i t t l e m e n t in relation to the use of

zinc a n d a l u m i n u m was a d d r e s s e d in P a p e r 6. The c r a c k i n g susceptibilities of
a range of ferritic, austenitic-ferritic, a n d fully austenitic m a t e r i a l s h a d been
d e t e r m i n e d u n d e r tensile l o a d i n g when coated with zinc or a l u m i n u m at tem-
p e r a t u r e s up to a p p r o x i m a t e l y 850~ or when welded. Some of the d a t a from
the p a p e r are p r e s e n t e d in T a b l e s 6 a n d 7. All of the test m a t e r i a l s proved
susceptible to e m b r i t t l e m e n t by zinc, in p a r t i c u l a r the austenitic m a t e r i a l s
where a nickel leaching m e c h a n i s m operates. However, none of the test mate-
rials c r a c k e d in the presence of a l u m i n u m , although in some cases, there was
evidence t h a t alloying h a d occurred. Discussion revealed a consensus t h a t it is
inadvisable to coat stainless steels with zinc-rich p a i n t s where toxic or flam-
m a b l e m a t e r i a l s are being processed. However, e m b r i t t l e m e n t or c r a c k i n g
risks are significantly lower in the case of a l u m i n u m , albeit some alloying
might occur, a n d require detection, in the event of a fire.
Stress Corrosion Cracking Prevention: Materials Fabrication/Selection
A n u m b e r of cases of stress corrosion c r a c k i n g r e p o r t e d at the m e e t i n g h a d

been caused by the ingress of fluids into laggings, not from the exterior, b u t
from the vessels or p i p i n g systems themselves. Typical f a b r i c a t i o n defects t h a t
h a d led to flooding of lagging systems included lack of fusion, porosity, a n d
piping, p a r t i c u l a r l y in t a c k a n d stitch welds used in the assembly of t a n k s .
TABLE 6--Incidence o:]'liquid metal embrittlement of

stainless steels and nickel alloys by zinc and aluminum
jor burner experiments."
Fracture~
Material Zinc Aluminum
ASTM A 285C +
SAE 4140 +
ASTM A-200T4 +
5 Cr 0.5 Mo +
7 Cr 0.5 Mo +
9 Ni +
AISI 405 + --
18 Cr 2 Mo + --
26 Cr 1 Mo + --
AF 22 + --
3 RE 60 + --
AISI 304
AISI 316
lncoloy 800 +
Hastelloy C +
"Presented originally in Paper 6.
/,Blank means not tested.
TABLE 7--Incidence of liquid metal embrittlement of

stainless steel and nickel alloys by zinc and aluminum
for welding experiments."
Cracks
Material Zinc Aluminum b
AISI 304 + m
AISI 310 +
AISI 316 +
AtSI 317 +
A1SI 321 +
AISI 347 +
Incoloy 800
Incoloy 825 +
Hastelloy B +
Hastelloy C
AF 22 m
18 Cr 2 Mo
26 Cr 1 Mo
AISI 405
"Presented originally in Paper 6.

b Blank means not tested.
The lessons in relation to appropriate construction supervision and inspection

are obvious.
In a number of instances, experience of expensive stress corrosion cracking
problems with austenitic stainless steels had resulted in the selection of more
resistant alloys for replacement vessels and piping systems. The materials
that were actively discussed at the meeting were as follows:
1. Extra Low Interstitial Ferritic Steels--The use of 18 Cr 2 Mo grades of
ferritic stainless steel for vessels and piping systems in the brewery industry
was discussed in Papers 2 and 3. These materials are immune to chloride
stress corrosion cracking and can be used at operational temperatures up to
approximately 300~ Concerns were expressed about the use of such materi-
als for welded constructions with wall thicknesses greater than " a few milli-
metres" because of the problems of achieving adequate heat-affected zone
(HAZ) toughnesses. However, Paper 3 reported on the satisfactory construc-
tion of some sizeable vessels with wall thicknesses up to 6 mm.
2. Duplex Stainless Steels--In a number of cases, duplex 18 Cr 5 Ni grades
of stainless steel had been used to construct sizeable vessels. These materials
are not immune to stress corrosion cracking but are significantly more resis-
tant in terms of tolerable temperatures and chloride levels than the conven-
tional 18 Cr 8 Ni austenitic grades of stainless steel.
3. "Super" Austenitie Stainless Steels--There was some discussion of the
scope for 20 Cr 25 Ni grades of austenitie stainless steel, which although not
immune to SCC are significantly more resistant than 18 Cr 8 Ni grades. The
consensus was that, in terms of limiting temperature, they perform similarly

to the duplex grades, although there was some divergence as to values of that
temperature within the range of approximately 140 to 180~
Corrosion Prevention: Design, Specification, Inspection, and Maintenance

A recurring theme throughout the meeting was that there are two require-
ments for controlling corrosion within lagging and fireproofing systems. On
the assumption that water can enter the system, there is a need for some anti-
corrosion measure(s) to be adopted within the system, be it inhibitors, paints,
foils, or whatever. However, there is an overriding need to keep water out of
lagging and fireproofing systems at all stages from application to retirement.
None of the available anticorrosion measures were designed for, or are able to
cope with, prolonged periods of exposure to flooded systems, regardless of the
lack of efficient insulation offered by such systems. It follows that designs and
specifications, while concerned with specific anticorrosion procedures, must
also be preoccupied with the necessity for efficient waterproofing, and that
this emphasis must be maintained through application or in-service inspec-
tion and maintenance.
There was a consensus on the desirability of consultation between designer,
operator, and application contractor at the design stage to produce a suitable
specification, which was elaborated in Paper 10. In relation to waterproofing,
a number of specific mechanical design issues were raised and are worthy of
note:
1. "Top hats" are of considerable value in shedding water away from up-
per termination joints between fireproofing and steel.
2. Joints in metal foil and cladding should always be arranged to shed
water.
3. Drainage points should be arranged at the base of insulation systems on
long vertical pipe runs, columns, and so forth to prevent water holdup.
4. Waterproof sealant should be applied around any protrusions from lag-
ging systems, such as hangers, supports, and so forth, and these should be
kept to a minimum.
5. Joints in steam tracing pipework should always be outside, preferably
beneath, the main lagging system.
While there was general recognition of the need for an appropriate non-
permeable vapor barrier on cold insulation systems, there was some diver-
gence as to the relative merits of metal cladding versus reinforced mastic coat-
ings for weatherproofing hot insulation systems.
The case for using "specialist" insulation application inspection was pre-
sented in Paper 9. The key tasks for such inspection were identified as
follows:
(1) confirmation that correct specified materials are applied in each area,
(2) approval or control of storage to avoid wetting,
(3) confirmation that insulant remains dry after application until sealing
or cladding is completed,
(4) checking correct application of vapor barrier,
(5) checking correct cladding application, including arrangement of joints
to shed water, and sealing of gaps, cut-outs, and so forth,
(6) confirming appropriate staggering of joints in multilayer insulation
systems,
(7) checking for damage at all stages of application, and
(8) witnessing laboratory tests on in-situ foamed materials.
There was also a consensus that vigilance on many of the latter points needs
to be maintained throughout the life of the lagging system. Caulking or joint-
ing materials dry out and lose flexibility, cladding or barrier systems suffer
local damage or perforation, joints leak in service and so forth. Accepting
that the corrosion control system within the lagging or fireproofing system
cannot provide unlimited containment, such problems need to be identified
and remedied, or corrosion problems are inevitable.
Finally, the in-service inspection of metal surfaces beneath lagging systems
was discussed briefly. The specific use of a magnetoscope for detecting zinc
embrittlement of austenitic stainless steel surfaces was discussed in Paper 6.
Otherwise, no new initiatives were reported at the meeting, and there ap-
peared little alternative to the costly removal of lagging/fireproofing systems
to allow access for the traditional nondestructive testing (NDT) techniques.
There was a consensus that this is an unsatisfactory situation, and that there
is a need for the development of an appropriate in-situ technique.
APPENDIX
Abstracts of Papers Presented at the Corrosion under Lagging Conference Held in
Newcastle-on-Tyne, England, Nov. 1980
Paper 1
"Thermal Insulation. Specification and Materials. Application to Stain-
less Steel Surfaces," J. D. Nicholson, Darlington Insulation Co Ltd., West
Auckland Road, Darlington, Co Durham DL30UP, United Kingdom.
The problem of external stress corrosion cracking of certain grades of stainless steel
is well known in the process industries. Soluble chlorides and fluorides in contact with
austenitic stainless steels under stressed conditions create a potential.risk, which can
be reduced if the plant designer chooses the optimum grade of steel.
Insulating materials are available which, in association with suitable accessories,
can reduce the risk of stress corrosion cracking. Materials with low content of free
chloride, the presence of barrier materials, and inhibited insulation can be of value.
Reproducibility of results is difficult under laboratory conditions. Site experience

suggests that dual temperature operation may provide greater risk than does continu-
ous high temperature operation.
A review of British Standards and of specifications that are issued by engineering
and petrochemical companies shows divergence of opinion about permissible levels of
soluble chloride.
The insulation contractor can assist in preventing dangerous conditions by the cor-
rect design and application of insulation systems.
Inspection and regular maintenance are essential to preserve the integrity of an in-
sulation system. The risk of stress corrosion attack can be reduced but not necessarily
eliminated by the use of materials of low chloride content.
There is only one way to prevent External Stress Corrosion Cracking of stainless
steel in contact with insulation: Keep the Insulation Dry at All Times.
This paper was published subsequently in the Bulletin of the Institution of Corro-
sion Science and Technology, Vol. 19, No. 5, Oct. 1981, pp. 2-5.
Paper 2
"The Basics of Corrosion Mechanisms in Lagged Steelwork," RAE
Hooper, Stainless Steels Department, BSC Sheffield Laboratories, Swindon
House, Moorgate, Rotherham $60 3AR, United Kingdom.
Process plant vessels and pipework are usually lagged to conserve heat, and occa-
sionally cold, and this has been a common practice for a very long time. The sharp rise
in fuel prices in the last decade, and the other increasing need to conserve energy, is
making the lagging of process equipment more and more necessary. The usual effect
of lagging is to raise the temperature of the outer side of the plant wall, which is most
commonly made of steel. This should be dry and free from corrosion hazards but,
unfortunately, this ideal situation frequently does not occur as moisture enters the
lagging. On some occasions this arises from undetected leaks in the process equip-
ment but more often the moisture comes in through inadequate weather protection of
the lagging. Lagging should be covered with a waterproof coating to provide protec-
tion from the rain, sea spray, or spillage or leakage from other parts of the plant.
When moisture does get into the lagging the high temperature causes evaporation and
concentration of any dissolved solids, either in the incoming water or leached from the
lagging itself. The point of maximum concentration depends upon the rate of water
ingress, the plant temperature, and the extent of temperature transients, but it is inev-
itable that the hot plant wall will see an environment concentrated in the dissolved
solids. As a result, corrosion can occur.
Carbon and low alloy steels tend to suffer general attack in hot, moist environments
but if the liquids trapped under lagging contain substances such as nitrates or caustic,
then stress corrosion cracking can also occur. This type of corrosion is typified by
intergranular cracks and can be rapid under some circumstances. Stainless steels gen-
erally suffer from pitting or crevice corrosion rather than general corrosion, and the
most harmful dissolved species are the halogens, of which chlorides are by far the most
common and most aggressive. If the stainless steel has an austenitic structure, for
example, Types 304 and 316 (Unified Numbering System [UNS]), which contain ap-
proximately 10% nickel, then stress corrosion cracking can also occur in the presence
of chlorides. This type of cracking is characteristically transgranular and highly
branched, although in some instances some intergranular cracking can also occur.
Corrosion under lagging can be prevented by a number of methods, which include
the prevention of ingress of moisture, the use of laggings which do not contain harm-
ful, leachable chemical species, the surface coating of pipes and vessels and last, but
by no means least, the correct choice of material.
Paper 3
"External Use of Austenitic Stainless Steel Vessels and Pipework--A Case
Study," D. Geary and G. Bailey, CAPCIS, Corrosion and Protection Centre,
UMIST, P.O. Box 88, Manchester M60 IQD, United Kingdom.
No abstract supplied.
Paper 4
"ICI Practice for Preventing SCC Under Lagging," M. E. D. Turner,
Consultant, 1 Norfolk Crescent, Ormesby, Cleveland T S 3 0 L Y , United
Kingdom.
Experience has shown that a significant reduction in the risk of stress corrosion of
stainless steel under lagging can be effected by the simple expedient of interposing a
layer of aluminum foil between the equipment and the lagging.
The underlying theory is explained and laboratory test and practical experience
described.
Some alternatives are discussed and the reasons for their rejection given.
Paper 5
"Use of Paint Coatings for Under-Lagging Corrosion Prevention. A User
View," F. H. Palmer, Engineering Department, BP Trading Ltd., Britannic
House, Moorgate, London EC3, United Kingdom.
No abstract supplied.
Paper 6
"Stress Induced Cracking of Steels by Molten Zinc and Aluminum," P.
Geenen, Koninklijke/Shell-Laboratorium, Shell Research B.V., Postbus
3003, 1003AA, Amsterdam, Netherlands.
After the Flixborough disaster had highlighted the risk of contact between molten
zinc and austenitic stainless steels, failures caused by such contact have been reported
regularly. Disagreement in the literature about the prevailing mechanism and igno-
rance of what materials are susceptible have stimulated our interest in this matter.
Particularly the need for a nondestructive detection technique called for an experi-
mental program.
The work was done in cooperation with the laboratory of Inorganic Chemistry and
Materials Science, Department of Chemical Engineering, Twente University of Tech-
nology, Enschede, Holland.
Specimens of the relevant materials were coated with a zinc or aluminum com-
pound, subsequently TIG welded without filler material and examined microscopi-
cally. SEM and X-ray distribution images were made. The susceptibility of a range of
alloys has been determined quantitatively by comparing the time to fracture and the
maximum loads applied during slow straining of specimens, with and without zinc
coating or aluminum metal layer.
A basis for nondestructive testing was laid by comparing observed metal attack un-
der zero strain and measured change in relative magnetic permeability by means of a
Forster Magnetoscope.
Paper 7
"An Experimental Investigation with Stress Corrosion Cracking of Aus-
tenitic Stainless Steel under Insulation," J. Gillett and K. A. Johnson, Fi-
breglass Ltd., Insulation Division, St. Helens, Merseyside, WA10 3TR,
United Kingdom.
The paper discusses the definition of insulation, its usage and the conditions that
may be encountered during application and afterwards in service.
The discussion includes the definition of when corrosion can be a problem and hav-
ing apparently restricted it to a narrow range of conditions demonstrates that, in fact,
these conditions are likely to be accounted transiently even if only infrequently and so
may be more common than is generally supposed. The possible origins of chloride ions
are listed and discussed.
The paper highlights the current approach of specifiers with the general require-
ment of low CI- being very common. However many international authorities do not
follow the U.K. approach, and require high sodium and silicate contents for inhibi-
tion purposes.
Investigation of fibreglass materials was followed by an extensive test program to
establish whether or not fibreglass materials could cause corrosion. These tests dem-
onstrate that in fact fibreglass materials could cause corrosion, in fact that fibreglass
does not cause corrosion of stainless steel. Further, the program investigates protec-
tive measures that could be taken against corrosion caused by contamination from
external sources of CI- ions.
The paper concludes with discussion of the results and recommendations for future
specifications.
Printed copies of this paper are available from Fibreglass Ltd.
Paper 8
"Corrosion Control Under Thermal Insulation and Fireproofing," J. F.
Delahunt, Exxon Research and Engineering Co., P. O. Box 101, Florham
Park, NJ 07932.
In recent years there have been an increasing number of reports concerning corro-
sion occurring on carbon steel structures and equipment that are either thermally in-
sulated for energy conservation or coated with concrete for protection from fire. In
view of this, Exxon Engineering has been involved in a number of field investigations
as well as laboratory investigative programs to evaluate the cause of corrosion and to
determine appropriate means to mitigate it. The cumulative result of these various
programs is presented within this paper, and it includes discussions concerning:
9 Examples of corrosion in refineries, petrochemical plants, and pipelines occur-
ring on insulated or fireproofed structures and equipment.
9 Descriptions of potential corrosion mechanisms.
9 Corrosion mitigation systems used to prevent attack of such equipment.
Paper 9
"Application Inspection of Insulating Materials," P. G. Blackburn and
R. G. Roberts, ITI Anti-Corrosion Ltd, 177 Hagden Lane, Watford, Hert-
fordshire WD1 8LW, United Kingdom.
In order to minimize heat losses it is important that thermal insulation is correctly
applied. Insulation inspectors, operating independently of the contractors and on be-
half of the client, can play an important part in the operation. Particularly in high-
lighting defects that would subsequently, and very expensively, waste heat or promote
corrosion.
Insulation materials should be inspected before installation for any damaged areas.
Joints should be offset, slabs or sections correctly butted, and in the case of two layer
application the inspector must ensure that horizontal and vertical joints are correctly
positioned.
Insulation on bends is a potential weakness because careless cutting on site can
leave gaps. These gaps are sometimes masked by the bands used to retain the insula-
tion in position. The bands themselves can also cause damage to fragile insulating
materials.
Metal cladding joints have to be correctly overlapped and wrongly positioned joints
can encourage water ingress, it is also important to see that joint sealants are correctly
applied or that fixing methods, such as pop rivets, do not penetrate the insulating
layer.
In general the most competent insulation inspectors are recruited from the thermal
insulating industry. Preferably they would have had at least ten years practical experi-
ence so that they are well aware of the short cuts, mistakes, and malpractices of
operatives.
Paper 10
"How to Write a Specification," I. G. Huggett, Consultant, Orchard
House, Low Worsall, Yarm, Cleveland, United Kingdom.
The specification has an important role in the lagging of process plant and is both
the interface and the means of communication between the lagging contractor and the
client. Thus, for example, whatever decisions are reached about the precautions nec-
essary to minimize corrosion, they have to be defined and put into effect by means of
the specification.
A good specification for lagging, as for other relatively complex construction activi-
ties, is hard to write and even the best cannot incorporate all that is necessary to
achieve good lagging; thus it can never be completely self-contained. Some things have
to be planned in advance by the client or others; additional communication between
client and contractor is almost always worthwhile and rarely unnecessary.
The way in which the specification is drafted, the phraseology, and the way the
ideas are expressed are always important. Frequently there is also real technical diffi-
culty in coming to decisions, and great care is necessary when trying to formulate
written rules: the provision for minimizing corrosion often falls into this category.
References
[1] Page, G. G., New Zealand Journal of Dairy Science and Technology, Vol. 15, 1980,
pp. 143-157.
[2] Karnes, H. F., American Institute of Chemical Engineers 57th National Meeting, Sept.
1965.
[3] Delahunt, J. F. Insulation Journal, Vol. 26, No. 2, Feb. 1982, p. 10.
[4] Nicholson, J. D., Bulletin of the Institution of Corrosion Science and Technology, Vol. 19,
No. 5, Oct. 1981, p. 2.
Thermal Insulation Materials
George E. Lang 1
Thermal Insulation Materials:

Generic Types and Their Properties
REFERENCE: Lang, G. E., "Thermal Insulation Materials: Generic Types and Their
Properties," Corrosion of Metals Under Thermal Insulation. " A S T M STP 880, W. I. Pol-
lock and J. M. Barnhart, Eds., American Society for Testing and Materials, Philadelphia,
1985, pp. 63-68.
ABSTRACT: Thermal insulation plays a very important function in all of our lives. With-
out it the cost of energy used in our homes would skyrocket, perishable foods would be avail-
able only within a short distance of the source, plastic items and many petroleum products
would not exist, and we would have to get used to warm beer and soupy ice cream.
Now to the generic and functional differences of the various types of thermal insulation.
One of the most defining properties of a thermal insulation is its service temperature range.
Since we are discussing generic types of insulation, and some manufacturers' recommended
service temperatures vary slightly, the du Pont Company's established recommended ser-
vice temperatures are listed in the paper. The properties, good and bad, of generic thermal
insulation materials that have a bearing on their individual influence on corrosion of the
metals they cover will be discussed.
KEY WORDS: thermal insulation, polystyrene, plastics, corrosion
T h e r m a l insulation plays a very i m p o r t a n t f u n c t i o n in all of o u r lives. W i t h -

out it t h e cost of e n e r g y u s e d in o u r h o m e s w o u l d skyrocket, p e r i s h a b l e foods
w o u l d be available only w i t h i n a short d i s t a n c e of t h e source, plastic items a n d
m a n y p e t r o l e u m p r o d u c t s w o u l d not exist, a n d we w o u l d h a v e to get u s e d to
w a r m b e e r a n d soupy ice c r e a m .
E x c e p t for o u r c o n c e r n s a b o u t h o m e h e a t i n g costs, t h e r m a l insulation is not
a p p r e c i a t e d in t h e r e m o t e c i r c u m s t a n c e s w h e r e it i n f l u e n c e s o u r lives, a n d in
very m a n y of t h e s e cases it is n e g l e c t e d .
T h e r m a l insulation, in t h e f o r m of p i p e covering, block, b a t t , a n d b l a n k e t ,
is m a n y small, to m i c r o s c o p i c a l l y small, gas-filled cells of some f o r m of solid
m a t e r i a l . T h e m a t e r i a l m a y b e f o a m e d plastic or glass, it m a y be a b l o c k or
b l a n k e t of glass or m i n e r a l fibers, or it m a y b e a c e m e n t i t i o u s m i x t u r e k n o w n
as c a l c i u m silicate or perlite silicate.
IConsultant on thermal insulation, E. I. du Pont de Nemours and Co., Inc., Engineering Service
Division, Louviers Building, Wilmington, DE 19898.
One of the most defining properties of a thermal insulation is its service tem-
perature range. The established service temperatures recommended by the du
Pont Company's engineering d e p a r t m e n t for the thermal insulation materials
it normally uses, as generic types are shown in Table 1.
The properties a n d characteristics of these generic thermal insulation mate-
rials, which have a bearing on their individual influence on corrosion of the
metals they cover, are discussed in the following sections. The information is
based on our m a n y years of experience.
I have not included c o m m e n t s on thermal efficiency. The reason is that m a n u -
facturers' literature values are for thermal conductivities determined in labora-
tories for "dry" samples. But installed thermal insulation frequently contains
more moisture t h a n these samples. This increased moisture may have come
from shipping, storage, installation, or service. The good low K values of ther-
mal insulations are quickly lost when the insulations become wet. Therefore, in
the field we frequently have poorer thermal efficiency of the installed insulation
than the K values used in the initial design calculations. In the practical sense,
comparing thermal insulations solely on the K values from "dry" samples can be
misleading. W h a t their values are in the field when they have more water in them
depends, of course, on the a m o u n t of water. W h e n thoroughly wet, m a n y have
similar high K values (Table 2).
Polystyrene Foam
9 This insulation is available in the form of polystyrene foam billets or poly-
styrene beads e x p a n d e d in a mold.
9 To be functional as a process thermal insulation, the density cannot be less
t h a n 24 k g / m 3 (1.5 lb/ft3).
9 In the prescribed density, it does not absorb or wick water as long as the cell
structure remains intact.
TABLE l--Established thermal temperatures recommended by the du Pont Conq~any's

engb~eerhtg department.
Generic Thermal
Insulation Materials Recommended ServiceTemperature, ~ (~
Polystyrenefoam --73 to 60 (-- 100 to 140)

Polyurethanefoam--rigid --73 to 82 (-- 100 to 180)
Polyisocyanurate--rigid -- 73 to 149 ( -- 100 to 300)
Flexiblefoamed elastomer +2 to 82 (+35 to 180)
Cellularglass -- 129 to 149 (--200 to 300)
Glass fiber 4 to 190 or 454 (40 to 375 or 850)
depending on type
Mineral wool 60 to 649or 982 ( 140to 1200or 1800)
depending on type
Calcium silicate 60 to 649 (140 to 1200)
Perlite silicate 60 to 593 (140 to 1100)
LANG ON THERMAL INSULATION MATERIALS 65
TABLE 2--Comparison of K values from samples

9in the field.
Samples K BTU, in./h, ~ ft 2
Dry calcium silicate about 0.42

Dry urethane " 0.20
Dry fiberglass 0.25
Thoroughly wet calcium silicate 4.0
Thoroughly wet urethane 4.0
Thoroughly wet fiberglass " 4.0
9 It has a m a x i m u m service t e m p e r a t u r e of 60~ (140~

9 Compressive resistance is 172 k P a (25 psi) average at 5 % deformation.
9 T h e insulation is destroyed by solvents other t h a n alcohol a n d is softened
u n d e r black or other final coverings that will reach 60~ (140~ or above.
9 W e have not used a significant quantity of this type insulation. No p l a n t
corrosion problems have been noted.
PolyurethaneFoam--Rigid
9 This plastic foam is primarily used for cold a n d anti-sweat service.
9 It does not absorb a n d wick water as long as the cell structure remains in-
tact. It is p e r m e a b l e to water vapor in cold service when required vapor barrier
fails. Vapor diffuses through cell walls to the t e m p e r a t u r e zone where it con-
denses a n d further to where it freezes.
9 It has a m a x i m u m service t e m p e r a t u r e of 82~ (180~
9 T h e typical average b u l k density is 132 k g / m 3 (2 lb/ft3).
9 Compressive resistance is 17 k P a (25 psi) average at 5% deformation.
9 If in continuously cold service, it does not corrode u n p r o t e c t e d metal sur-
faces. If in intermittent service to its m a x i m u m service t e m p e r a t u r e , it can
cause corrosion of u n p r o t e c t e d wet metal surfaces from released chlorides in
fire r e t a r d a n t s a n d blowing agents. The sun's ultraviolet rays decompose this
insulation.
PolyisocyanurateFoam--Rigid
9 This insulation is very fire resistant for an organic foam; it is a low flame
propagation rate plastic foam of the polyurethane family.
9 It does not absorb a n d wick water as long as the cell structure remains in-
tact. It is p e r m e a b l e to water vapor in cold service when required vapor barriers
fail.
9 M a x i m u m service t e m p e r a t u r e is 149~ (300~
9 Typical average b u l k density is 32 k g / m 3 (2 lb/ft3).
9 Compressive resistance is 17 k P a (25 psi) average at 5 % deformation.
9 When this material is exposed to heat and moisture, the cell structure in
the heated zone is damaged. The decomposition products may contain chlo-
rides from the fire retardant and blowing agent and are thus fairly aggressive as
corrodants of unprotected wet metal surfaces. The sun's ultraviolet rays de-
compose this insulation.
Flexible Foamed Elastomer

9 This insulation is a black, rubbery type flexible foamed plastic.
9 It does not readily absorb or wick water.
9 The maximum service temperature is 82~ (180~
9 Average bulk density is 96 k g / m 3 (6 lb/ft3).
9 It has poor compressive resistance.
9 Although not corrosive by itself, it supports corrosion of unprotected metal
surfaces when water is present, particularly when the water contains chlorides
from an external source.
Cellular Glass
9 This insulation is a rigid glass foam whose blowing agent contains hydro-
gen sulfide and carbon dioxide.
9 It does not absorb and wick water as long as the cell structure remains
intact.
9 The maximum service temperature is 149~ (300~
9 Average bulk density is 136 k g / m 3 (8.5 lb/ft 3)
9 Compressive resistance is 690kPa (100 psi) average.
9 W h e n water is present and the cell structure is damaged, release of the
foam blowing agent may cause corrosion on unprotected carbon steel surfaces.
Glass Fiber
9 This insulation is pure glass fiber containing various types of binders.
9 It will absorb and wick water, but is more able to drain excess moisture
than other types of insulation.
9 The maximum service temperature is 191 to 454~ (375 to 850~ depend-
ing upon type and brand.
9 The bulk densities range from 24 to 96 k g / m 3 (1.5 to 6 lb/ft 3) depending
upon type and brand.
9 It has poor compressive resistance.
9 The fact that it will wick water makes it conducive to corrosion on unpro-
tected wet metal surfaces.
Mineral Wool
9 This insulation is a mineral or metal slag fiber, basically an impure glass

fiber.
LANG ON THERMAL INSULATION MATERIALS 67
9 It readily absorbs and wicks water.

9 Maximum service temperature is 649 to 982~ (1200 to 1800~ depend-
ing upon brand.
9 Bulk densities range from 96 to 304 k g / m 3 (6 to 19 lb/ft 3) depending upon
type and brand.
9 It has poor compressive resistance; there is some improvement at higher
densities.
9 The fact that it will wick and hold water makes it conducive to corrosion on
unprotected wet metal surfaces.
Calcium Silicate
9 This insulation is a cementitious mixture, known in the trade as "white
goods."
9 It readily absorbs and wicks water. Can hold up to 400% of its own weight
of water without dripping.
9 Maximum service temperature is 649~ (1200~
9 Average bulk density is approximately 224 k g / m 3 (14 lb/ft3).
9 Compressive resistance is 621 to 1103 kPa (90 to 160 psi) at 5% deforma-
tion, depending upon brand.
9 Although its p H is initially high, 10 average, it is fairly aggressive in sup-
porting corrosion on unprotected wet metal surfaces because of its moisture re-
tention, particularly when the moisture contains chlorides from an external
source.
Perlite- Silicate
9 This insulation consists mostly of expanded perlite with sodium silicate as
a binder.
9 It may contain an ingredient to resist water absorption up to a certain tem-
perature at which point the moisture resistance burns out. When this happens,
it will absorb and wick water.
9 The maximum service temperature is 593~ (ll00~ above which shrink-
age becomes excessive.
9 Average bulk density is approximately 224 k g / m 3 (14 lb/ft3).
9 Compressive resistance is somewhat less than the calcium silicates being
in the 483 to 552 kPa (70 to 80 psi) range at 5% deformation.
9 The pH is high, being 10 plus, making it less conducive as a corrodant on
wet metal surfaces particularly when it contains a water repellant.
Conclusion
Selection of a thermal insulation depends on many factors. We believe an im-
portant characteristic to consider is what extent the insulation can contribute
to corrosion of the metal underneath if the insulation gets wet. In general, we
prefer "sterile" insulations. However, even a totally sterile insulation will

accelerate the corrosion process going on under it on an unprotected wet metal
surface.
DISCUSSION
W. D. Johnston I (written discussion)--I wish to comment on Mr. Lang's

reference to cellular glass. Mr. Lang indicates that cellular glass provides a
source of sulfuric acid that is aggressively corrosive. This is not true. FOAM-
GLAS | cellular glass insulation is not a source of sulfuric acid. In fact, under ag-
gressive hydrolysis conditions, it yields an alkaline solution of about pH = 9,
which is inconsistent with the presence of sulfuric acid.
G. E. Lang (author's closure)--My entire paper dealt primarily with corro-
sion experiences du Pont has had on its industrial plants. Those experiences
over 20 plus years have been accumulated by a number of engineering people in
addition to Roy Allen and myself. These plant experiences are confirmed by
our laboratory tests where possible.
As we heard during the symposium, severe corrosion and or failures normally
take years to occur. For this reason, a leachant prepared in a V2-htest does not
go far enough to guide us in solving our "real world" corrosion problems.
From our actual field experience on a number of vessels, I must stand behind
my statement that "cellular glass is very aggressive on unprotected carbon steel
surfaces" when in the presence of stagnant warm water.
Please note that in my paper, or discussion with you during the question
period or our one-on-one discussion after the meeting, I did not state that
FOAMGLAS causes chloride stress-corrosion cracking (SCC) of austenitic
stainless steel. In fact, in four 28-day tests over the years, we have proved
FOAMGLAS does not cause chloride SCC.
In the few field experiences where we had this SCC problem under FOAM-
GLAS, we proved the chlorides did not come from the FOAMGLAS.
To update our mutual concern about the pH of a FOAMGLAS leachant, I
have set up a test to repeat essentially the Lehigh Lab procedure and Roy Allen's
test. This test will run for one year with a monthly check of the leachant. We will
keep you advised of the results.
IPittsburgh CorningCorp., 800 Presque IsleDrive, Pittsburgh, PA 15239.
Field Experience
Tore Sandberg 1
Experience with Corrosion Beneath

Thermal Insulation in a
Petrochemical Plant
REFERENCE: Sandberg, T., "Experience with Corrosion Beneath Thermal Insulation

in a Petrochemical Plant," Corrosion of Metals Under Thermal Insulation, A S T M STP
880, W. I. Pollock and J. M. Barnhart, Eds, American Societyfor Testing and Materi-
als, Philadelphia, 1985, pp. 71-85.
ABSTRACT: A program to detect corrosion under insulation has been ongoing at the
Stenungsund ethylene plant since 1976. Experience to prioritize inspection has been de-
veloped. During the years 1978 to 1980, the number of yearly inspected lines with more
than 0.5-mm corrosion decreased from 50 to 15%. A surveyof the condition of 530 pres-
sure vesselswas carried out in 1980. Corrosionwas found under hot insulation on 75% of
not painted vessels and on 20% of painted and insulated vessels. No corrosion was found
on equipment in continuous cold service.
KEY WORDS: corrosion, atmospheric corrosion, insulation, experience
The Esso Chemical AB ethylene p l a n t is located in S t e n u n g s u n d on the

west coast of Sweden. The p l a n t was built in 1962 a n d 1963 with a m a j o r
expansion in 1969. The p l a n t capacity is 350.000 tons ethylene/year.
I n 1976 severe corrosion u n d e r i n s u l a t i o n (CUI) was f o u n d on piping, a n d a
couple of years later m a j o r corrosion was also f o u n d on other e q u i p m e n t . At
that time a p r o g r a m was started to detect corrosion a n d make required re-
pairs.
In 1980 severe corrosion on two towers caused a p l a n t shutdown. Extensive
stripping of e q u i p m e n t was carried out to complete the corrosion picture.
Several actions have been t a k e n to prevent future corrosion, such as
(1) increased resources for inspection a n d m a i n t e n a n c e ,

(2) new specifications for e q u i p m e n t details, p a i n t i n g , a n d insulation, a n d
(3) new demisters were specified for cooling tower installation.
IEngineering associate, Esso Chemical AB, Box 852, S-444 01 Stenungsund, Sweden.
Thermal Insulation Systems Used
Insulation for Hot Service

In part of the plant built in 1962 and 1963 calcium silicate insulation was
used throughout with jacketing of galvanized and painted steel plate. Process
equipment was painted with red lead before insulation.
In part of the plant built after 1963 and in a particular expansion project
1969 mineral wool insulation was used extensively with aluminum jacketing.
Most equipment was not painted before insulation.
Insulation for Cold Service

Block polyurethane insulation and foamed in-place polyurethane was used.
Jacketing was painted galvanized steel or aluminum.
As a result of corrosion findings the specification has been revised and
painting is now performed before insulating. The latest revision of the insula-
tion specification prescribes carbon steel piping and equipment with operat-
ing temperatures (continuous and intermittent services) up to 120~ -
ous, and 200~ shall be painted and insulated as follows:
1. Surface preparation--Very thorough blast cleaning in accordance with
Sa 21/2, Swedish Standard (SIS) 055900-1967 (Pictorial surface preparation
standards for painting steel surfaces).
2. Undercoat using inorganic zinc silicate to a thickness of 75 ~m.
3. Finish coat--Two-pack, high build, polyamide-cured epoxy, 50/~m to-
tal thickness.
4. Nominal dry film thickness for total system, 125/zm.
5. Use mineral wool as insulation material. Jacketing is to be carbon steel
with an overlay consisting of 55% aluminum, 43.4% zinc, and 1.6% silicon.
Aluminum is used where mechanical damage is less likely.
Environmental and Site Conditions

Stenungsund is located close to the sea. In addition, there is a strong effect
from the salt water cooling tower. Plant atmospheric corrosion is affected by
the cooling tower mainly because of fogging and salt water mist. The corro-
sion rate is estimated to increase by a factor of 2 to 5 with increasing proximity
to the cooling tower.
The local environmental conditions are as follows:
Air temperature, year average, + 7 ~
July, + 17~
January, -- 1 ~
Relative humidity, year average, 78%
SANDBERG ON A PETROCHEMICAL PLANT 73
Annual precipitation, 700 m m

Sulfur dioxide in air, 0.004 p p m
Salt in air, outside cooling tower area, 15 g / m 2, year
cooling tower area (80 000 m 2, 260 g / m 2, year
close to cooling tower (maximum), 3500 g / m 2, year
Since 1982, after installation of new demisters in the cooling tower, the
amount of salt in the air has been less than 30 g / m 2 for the cooling tower area
and less than 175 g / m 2 for regions close to the cooling tower.
Corrosion Encountered
During the years 1963 through 1969 corrosion maintenance and corrosion
inspection programs were developed. Spot examinations of equipment
showed no major concerns regarding CUI. Our inspection program was con-
sidered to be built on several years experience and was kept unchanged after
the large plant expansion in 1969.
Piping
In February 1976 the first major prohlems appeared. In an area, close to
the cooling tower, severe corrosion was found under insulation. Lines had to
be taken out of operation to be repaired.
A number of questions had to be answered:
9 Where to look (area, service conditions, type of insulation, and so forth)?
9 How to look (technique, organization)?
9 How serious were the problems?
9 What were the reasons?
9 How to prevent corrosion in the future?
9 How to prioritize the inspection?
9 Costs involved?
9 And many more questions.
During March through May 1976, a comprehensive corrosion detection
program was carried out on hydrocarbon lines, hot service insulated and with
an operating temperature less than 150~ During the three month period,
317 lines were inspected by removing sections of insulation.
The following are the results from the 1976 spring investigation: About
18% of the lines had areas with more than 1-mm corrosion. For more details
see Table 1. It was also noticed that steam-traced lines were corroded to about
the same extent as nonsteam-traced lines. Examples of typical CUI are shown
on Figs. 1 through 4.
We concluded from these results that an intensive inspection program must
be organized. A prioritization of offsite lines and facilities in the 1969 plant
TABLE 1--Inspection of 317 lines. March through May 1976. All lines hot service insulated.
Operating temperature less than 150~
Corrosion per 100 Lines
Number of Less 0.1 More

Inspected Than to 1 to 2 to Than
Area Lines 0.1 mm 1 mm 2 mm 3 mm 3 mm
1963 plant 91 56" 41 2 0 1

1969 plant 155 30 51 10 5 4
Offsite 71 20 45 17 13 5
All areas 317 35 47 9 5 4
"Example: 56% of 91 lines from 1963 had less than 0.1-mm corrosion.
was made. All high risk lines were to be inspected before the end of 1976.
From 1976 to 1981, 1430 lines were inspected. Of these lines repairs were
made as listed in Table 2.
We increased our ability to predict corrosion from 1976 to 1978 as a result
of the knowledge built up in 1976. After 1978 there was a dramatic decrease
in the number of corrosion findings. This is viewed as a result of the experi-
ence in how to prioritize an inspection program for corrosion detection. This
experience can be summarized with the following list of criteria for selection
of lines to be inspected.
Services Most Likely to Have Corrosion Present

9 Unpainted carbon steel with insulation
9 Operating temperatures from -- 10 to + 150~
9 Cyclic service
9 Poor jacketing
9 Exposed to moisture (weather or cooling tower)
9 Piping design-configuration and details
Corrosion of uninsulated lines was found to be about the same in each year
and much less severe when compared with insulated lines. During 1978
through 79, 2.5% of uninsulated lines had at least one spot corroded to half
wall thickness. The corresponding figure for insulated lines was 27%. To il-
lustrate the results of the prioritized inspection, two periods are compared in
Table 3.
Stress corrosion in American Iron and Steel Institute (AISI) Type 304 and
316 (Unified Numbering System [UNS] $30400 and $31600) stainless steel
has been experienced mainly on small diameter instrument piping. These
lines are replaced using material that is better resistant to stress corrosion.
FIG. 1--Examples of typical CUl--trapped water." (a) bottom of bend, (b) and (c) water en-
trance at vend may cause corrosion at drain. Example t~f trapped water is also shown on Fig. 8a.
FIG. 2--Example of CUl__jacketing open at support. Insulation jacket open at vertical beam
eaused severe local corrosion. The line diameter is 900 mm.
FIG. 3--Examples of typical CUl--effect of painting: (a) band of shop painting locally pre-
vented CUI and (b) inorganic zinc silicate without top coat does not protect for CU1. Insulation
removed after four years.
FIG. 4--Support of insulated line without pipe shoe.
TABLE 2--1nspection of 1.430 lines during the years 1976

through 1981, 70% of the lines are hot serv&e insulated.
Number of Repaired Lines per

Inspected 100 of Inspected
Year Lines Lines"
1976 350 4.3

1977
1978 i4; 2;18
1979 250 4.0
1980 310 2.3
1981 280 1.8
"Material is carbon steel in all repaired lines.
TABLE 3--Illustration of the result of prioritized inspection. Hot

service insulated lines (Total number 478).
Corrosion per 100 Lines
More than ~/2 More than

Period of Inspection Wall Thickness 0.5 mm
Jan. 1978 through Sept. 1979 27 50

Oct. 1979 through Dec. 1980 6 15
Pressure Vessels
Routine inspection of pressure vessels focused on internal inspection. In-

spection under insulation was carried out on a spot check basis. Significant
outside corrosion was first noted in 1978.
In 1980 major corrosion problems were identified. Severe corrosion on one
column required the plant to shutdown. An extensive survey of the condition
of insulated equipment was carried out.
In the plant there are 530 pressure vessels of which 190 are insulated.
Results of the survey are summarized in Table 4.
I m p o r t a n t e q u i p m e n t factors are as follows:
9 Operating temperature
9 Corrosion protection (painting)
9 Mechanical design details
9 Standard of insulation
9 Standard of jacketing
The environmental conditions and the equipment condition explain the
results listed in Table 4. W h a t can not be seen in the table is that cyclic tem-
perature service increases the risk of corrosion for hot and cold insulated
equipment.
The high extent of corrosion on hot insulated, unpainted equipment is a
result of the above factors. The reasons identified for corrosion on painted
equipment were ascribed mainly to poor detailed mechanical design.
Inspection Technique Developed and Employed

In a plant, with a large a m o u n t of equipment, organization of inspection is
as important as the inspection technique. For piping systems it is necessary to
have a good line register. The register should include data, such as operating
conditions, type of insulation, painting, time for inspection, inspection inter-
TABLE 4--Corrosion on pressure vessels, result of survey in 1980. a

Corrosion per 100 Vessels
Category of Vessel None Moderate Severe

Uninsulated, painted 88 12 0
Cold insulated (polyurethane) 100 . . . . . .
continuous service less than
minus 10~
Hot insulated, painted 80 20 0
Hot insulated, not painted 25 63 12
aTotal number of pressure vessels: 530, hot service insulated: 90, and cold
service insulated: 100.
vai, findings, and so forth. It is quickly evident that good updated isometric
drawings and flow diagrams greatly simplify inspection, prework, and re-
cording.
Using the data in the line register a judgment can be made about the risk of
corrosion under insulation. If there is a risk, the inspector can look for the
following details:
9 Attachments to piping and process equipment
9 Connections, especially vertical segments
9 Unsealed or damaged jacketing
9 Signs of rust
9 Low points especially if jacketing is damaged above the low point
9 Bottom of absorbent insulation
9 Dead legs
Insulation is removed for spot checking. Complete removal is only required
for exceptional cases. Inspection areas or openings in the insulation are not
recommended. Inspection openings are more likely to cause corrosion than
prevent it. Nondestructive testing methods have so far proven to be unrelia-
ble. Figure 5 gives an example where critical corrosion areas are likely to be
found.
FIG. S--Areas critical for CUI.
In a new plant, inspection of piping should be organized in such a way that

all lines in a pipe rack are inspected on the same occasion. If severe CUI is
discovered it may be more effective to select the most critical lines and inspect
line by line until the situation is under control. When a continuous inspection
level is achieved, inspection pipe rack by pipe rack is again preferable.
Our interval for external inspection is three to six years for drums, ex-
changers, and so forth. For piping our recommendation is six to twelve years.
However a first check on piping should be made three to six years from start-
up in order to obtain a general impression on piping condition and to set
future inspection timing.
SURVEY
30O
PER
YEAR
Z INSPECTED LINES
200
0
Z
iO0
i~dor S U R V E Y .
~ / ~.~ ~RGANIZING
~ ~ ~ _ _ . _ PER Y~AR
REPAIRED LINES
i . . . . . . . . . .
0 5 i0 YEARS
40
PAINTING, INSULATION
SCAFFOLDING.
30 TOTAL MAINTENANCE
< 2O
W
<
INSIJLAT]ON AND
X I0 I N S P E C T I O N M A N N ING
R E L A T E D T O C U r - [NSPEC'I'ION
' ; . . . . . 5. . . . i0 Y E A R S
PROBLEM
DISCOVERED
FIG. 6--Quantity of inspection and repairs, required resources as a result of CU1.
FIG. 7--Examples of designs to prevent CUI: (a) painted galvanized steel jacketing in service
20 years now at the end of life, and (b) valve design without gusset easier to seal.
FIG. 8--Examples of"designs to prevent CUI: (a) this design of insulation support ring should
prevent corrosion shown on the left and (b ) jaeketing details at manhole.
FIG. 9--Example of design to prevent CUI: support bracket.
Economics
During 1981, total maintenance costs for painting, insulation and scaffold-
ing were 2 million dollars averaging 39 people working per year. This amount
is four times the cost of our program for the period of 1973 through 1974.
Inspection cost for corrosion control is typically 0.2 million dollars per year.
The initial costs for developing programs and routines correspond to one to
three man-years depending on existing routines and plant size and age. Safety
and environmental aspects, together with the risk of unplanned shutdown
have justified the costs of our program.
For budgeting and resource planning it is necessary to know the incidence
frequency for the years after corrosion under insulation has been recognized.
Our experiences at Stenungsund are probably similar to those experienced at
other mature plant sites. These experiences are illustrated in Fig. 6.
This information may give guidance for those plant sites where the problem
is new and some plan to tackle it is being developed. Figures 7 through 9 show
some examples of improved detail design to prevent corrosion.
In the long run maintenance costs would be considerably reduced, provid-
ing improved techniques in painting, insulation and construction details were
utilized both in new projects and for repairs.
Victor C. L o n g 1 and Paula G. Crawley 1
Recent Experiences with Corrosion

Beneath Thermal Insulation in a
Chemical Plant
REFERENCE: Long, V. C. and Crawley, P. G., "Recent Experiences with Corrosion

Beneath Thermal Insulation in a Chemical Plant," Corrosion of Metals Under Thermal
Insulation. A S T M STP 880, W. I. Pollock and J. M. Barnhart, Eds., American Society
ABSTRACT: Corrosion of carbon and stainless steels under wet thermal insulation can
be a serious problem and can be especially prevalent in the humid Gulf Coast area. This
paper discusses an inspection program that has been in progress since late 1982 at a ten-
year old chemical plant located at La Porte, TX. The program is intended to determine
the extent of corrosion damage to major pieces of equipment that has occurred under
inhibited calcium silicate insulation finished with aluminum jacketing and to recommend
remedial action.
Key elements of a successful inspection program are discussed. A series of pictures is
presented that correlate visual external telltale signs with corrosion beneath the insula-
tion. At the time of writing, significant localized corrosion of carbon steel has been found,
as well as evidence of incipient stress corrosion cracking of austenitic stainless steel.
KEY WORDS: corrosion, thermal insulation, carbon steel, austenitic stainless steel,
inspection
C o r r o s i o n b e n e a t h t h e r m a l i n s u l a t i o n is a serious p r o b l e m . It is i n s i d i o u s in
n a t u r e a n d o f t e n g o e s u n d e t e c t e d u n t i l m a j o r d a m a g e has o c c u r r e d . I n t h e
c a s e of stainless steel, t o t a l r e p l a c e m e n t of d a m a g e d e q u i p m e n t m a y b e t h e
only r e c o u r s e . Costs c a n b e h i g h , b o t h for r e p a i r s a n d lost p r o d u c t i o n .
M o s t m a j o r c h e m i c a l p l a n t s h a v e active c o r r o s i o n m o n i t o r i n g p r o g r a m s
a i m e d at p r o c e s s side c o r r o s i o n . U n f o r t u n a t e l y , too little a t t e n t i o n is often
g i v e n to e x t e r n a l c o r r o s i o n u n d e r wet i n s u l a t i o n , or m o r e a p p r o p r i a t e l y to t h e
p r e v e n t i o n t h e r e o f . P e r h a p s o n e r e a s o n is l a c k of a w a r e n e s s , e s p e c i a l l y at
n e w e r p l a n t sites w h e r e t h e r e has n o t b e e n e n o u g h t i m e for c o r r o s i o n r e l a t e d
p r o b l e m s to d e v e l o p . C o s t is a n o t h e r r e a s o n . It is o f t e n difficult to j u s t i f y ex-
IE. I. du Pont de Nemours Company, P.O. Box 347, La Porte, TX 77571.
Copyright9 by ASTM lntcrnatkmal www.astm.org
LONG AND CRAWLEY ON A CHEMICAL PLANT 87
penditures for more effective insulation systems or an aggressive preventive

maintenance program for a problem that may not develop for 10 to 15 years.
These dollars are well spent, however, because the plant integrity will be pro-
tected and energy savings will accrue from keeping the insulation dry.
I intend to share with you our experiences with corrosion under insulation
at the La Porte Plant. The plant is located on the ship channel east of Hous-
ton, TX, a couple of miles from Galveston Bay. There are prevailing winds off
the salt water bay, and there is ample rainfall and humidity year round. The
plant is a typical chemical plant. Major equipment consists of distillation
columns (some quite large), reactors, and tanks. Typical operating tempera-
tures of the insulated equipment range from 50 to 150~ Materials of con-
struction in contact with insulation are generally carbon steel and American
Iron and Steel Institute (AISI) 304 or 316 (Unified Numbering System [UNS]
$30400 or $31600) stainless steels. The primary insulation system is calcium
silicate inhibited with sodium silicate and finished with either aluminum
jacketing or reinforced mastic. The plant age is about ten years.
General appearance of the insulation is good. However, on closer inspec-
tion numerous points for moisture intrusion can be found, primarily where
mastic is used as finish, at seals around nozzles and manways, and where
mechanical damage has occurred to the jacketing. Most of the leak points are
the result of aging and the lack of a preventive maintenance program dedi-
cated to the prevention of moisture intrusion. There are areas where the insu-
lation has become very wet, and in some cases the sodium silicate inhibitor
has been completely leached out.
Realizing that we had all of the conditions required for corrosion and be-
cause of problems at other plant sites, a program was initiated in late 1982 to
address the question of corrosion under insulation. There are four key ele-
ments to this program. They are
9 Equipment characterization
9 Inspection
9 Repair
9 Preventative Maintenance
In this paper, I shall discuss the first two elements.
Equipment characterization includes operating characteristics, location
within the plant, age, and mechanical design details. Thorough knowledge of
your equipment is important in setting equipment inspection priorities and
determining the most probable locations for corrosion for a given piece of
equipment.
Equipment that operates intermittently or below about 100~ or both is
generally more susceptible to corrosion under wet insulation than equipment
operating continuously at higher temperatures. Location is important. Is it
near a cooling tower and receives carry-over spray. Is it subject to flooding or
frequent hose-downs with chloride containing process water. Is it protected
by a deluge system that is tested frequently. A deluge system provides a very

aggressive source of water and chlorides.
Finally, design details are important. Corrosion under insulation differs
from internal corrosion in that it tends to be more localized. It is related to
points of moisture intrusion and areas of moisture concentration. It is very
important, for example, to know the location of insulation support rings.
They tend to act as dams for moisture that has entered the system from above,
and this is one area where we have seen the most indications of incipient stress
corrosion cracking of stainless steel equipment. Other important design de-
tails, of course, are materials of construction, wall thickness and, in the case
of stainless steel, location of stressed areas such as welds. Good characteriza-
tion of equipment will help in knowing where to start and where to look.
The second phase is field inspection. We have relied heavily on visual in-
spection at La Porte. While access can be a problem and cosmetic repairs to
the jacket system can obscure results, we have had fairly good success in cor-
relating tell-tale external signs with corrosion beneath.
Locating points of moisture intrusion is a good first step. Cracked caulking
and mastic around nozzles, manways, and on vessel heads are likely points of
moisture intrusion. Mechanical damage to the aluminum jacket is another.
Staining of the aluminum jacket with leached sodium silicate and corrosion
products is a good telltale sign. Also, perforation of the jacket from within
caused by reaction with moisture and the caustic sodium silicate is a sure sign
of trouble beneath.
Hot spots on the insulation surface are another sign of wet and damaged
insulation. The human hand is a simple and sensitive instrument for this test
but again access can be a problem. Temperature measurements using infra-
red (IR) techniques are good for remote areas, but when jacket temperatures
approach ambient, this technique suffers from interferences.
Once suspect areas have been identified, core sampling of the insulation
with a cork bore can be useful in confirming the presence of excessive mois-
ture or the loss of sodium silicate inhibitor. The final proof, however, is strip-
ping the insulation, clean up, and inspection of the metal for corrosion.
The following is a series of figures dealing with inspection of two distillation
columns, one stainless steel and one carbon steel. Operating temperatures are
between 80 and 90~
First the stainless steel column. From a distance the insulation system ap-
peared in good shape, but closer inspection revealed several areas with telltale
signs of wet insulation. About midway up the column, an insulation support
ring was found protruding through the aluminum jacketing. Insulation in
this area was found to be very wet, the sodium silicate inhibitor leached out,
and the aluminum jacketing perforated from sqdium silicate attack. Corro-
sion deposits were found on the stainless steel as will be seen in the next few
figures.
Moisture entered the insulation system around nozzles located a few feet
above the insulation support ring as shown in Fig. 1. Note the deteriorated
caulking. Figure 2 shows perforation of the aluminum jacketing and staining
caused by the leached sodium silicate. The insulation was removed from the
area above the insulation support ring. Crusty deposits were found just above
the ring, as well as spotty iron stains on the stainless steel. In Fig. 3, the crusty
deposits above the ring may be seen. They are very hard and difficult to
scrape off. Iron stain areas are also noted (Figs. 4 and 5) both along the crusty
deposits and more isolated specks where the stainless steel is relatively free of
deposits. These iron stains are symptomatic of corrosive attack on the stain-
less steel. One would expect to find stress corrosion cracking in these areas, if
present.
After the area was cleaned up and the metal surface polished with a sand-
ing disk, it was dye checked. Very little if any cracking was found. Figure 6
shows several small potential cracks in the column shell adjacent to a vertical
weld, a stressed area. Figure 7 shows two small pits possibly connected by a
short crack in an area where there was relatively few crusty deposits. We have
concluded that these were signs of incipient stress corrosion cracking. Fortu-
nately, there is no serious equipment damage in this area, but all the condi-
tions for stress corrosion cracking are there. We further conclude that if cor-
rective action is not taken, serious equipment damage will occur.
Next, I plan to discuss the inspection of the carbon steel column. Again,
FIG. 1--Deteriorated caulking, stainless steel vessel.
FIG. 2--Corroded aluminum jaeket.
FIG. 3--Crusty deposits on stainless steel vessel.
FIG. 4--Iron stain on stainless steel.
FIG. S--Pitting of stainless steel.
the insulation looked good from a distance, but it also suffered from moisture
intrusion. Figures 8 through 11 deal with an area near the base of the column.
Moisture entered the insulation system at a mastic joint where there was a
slight change in column diameter (Fig. 8) and around a nozzle. Telltale perfo-
ration of the aluminum jacket had occurred (Fig. 9). After the insulation was
removed, general corrosion of the carbon steel was noted. However, it was
relatively mild, just a few hundred micrometres (Fig. 10) except around a
manway and several nozzles (Fig. 11). In these areas heavy scale 7 m m thick
extended several centimetres back underneath the insulation. Metal loss
ranged from 0.5 to 2 mm. While this amount is still within the allowable
thickness loss, further corrosion must be avoided or repairs will be required.
FIG. 6--Dye check, possible cracks.
FIG. 7--Dye check, pitted area.
FIG. 8--Leak point, jacket of a carbon steel vessel.
FIG. 9--Corroded aluminum jacket.
FIG. lO--Minor pitting of carbon steel.
FIG. 11--Heavy scale, carbon steel vessel.
Summary
We have identified numerous areas of wet insulation including some where
the sodium silicate has been completely leached out. Significant corrosion of
carbon steel equipment is localized, primarily to areas around manways, noz-
zles, and other moisture collection points. Signs of incipient stress corrosion
cracking of stainless steel equipment are present, but no serious cracking has
yet been found. Without prompt remedial action and a good preventive main-
tenance program, we conclude that significant equipment damage will occur
in the future.
B. J. M o n i z 1 a n d M . C. R i t t e r 2
Failure of Type 316 Stainless Steel

Nozzles in Contact with Fire
Retardant Mastic
REFERENCE: Moniz, B. J. and Ritter, M. C., "Failure of Type 316 Stainless Steel Noz-
zles in Contact with Fire Retardant Mastic," Corrosion of Metals Under Thermal Insula-
tion, A S T M STP 880, W. I. Pollock and J. M. Barnhart, Eds., American Society for
Testing and Materials, Philadelphia, 1985, pp. 95-102.
ABSTRACT: Two vertical American Iron and Steel Institute (AISI) 316L (Unified Num-
bering System [UNS] $31603) stainless steel (SST) nozzles on the head of a process vessel
operating at 120~ leaked after six years in service. It has been plant practice to apply
mastic reinforced with glass cloth as a sealer over insulation around nozzles on vessel
heads. The cracking occurred where the fire retardant brominated mastic met the noz-
zles. Chloride stress corrosion cracking developed when the mastic seal with the nozzles
broke, allowing water to enter and chloride to concentrate. The water source is the fire
protection sprinkler system (S0-ppm chloride), which deluged the vessel eight to ten times
in six years. Up to 60 000-ppm chloride was detected on the fracture face and, signifi-
cantly, no bromide. Tests showed that the mastic contained 363-ppm leachable chloride
and 36 300-ppm leachable bromide. The remedy was to terminate insulation with a metal
cover and seal the gap between cover and nozzles with caulking compound.
KEY WORDS: stainless steels, thermal insulation, fire retardants, calcium silicates,
UNS $31600, chloride analysis, bromide analysis
D e s c r i p t i o n of Failure
A 4 . 3 6 - m - h i g h b y 3 . 6 S - m - d i a m e t e r (14-ft 4-in. h i g h by 12-ft d i a m e t e r ) 36

715-L (9700-gal) h o l d u p t a n k was m a d e of T y p e 316 stainless steel ( U n i f i e d
N u m b e r i n g S y s t e m [ U N S ] alloy $31600). It o p e r a t e d at 1 2 0 ~ a n d was de-
s i g n e d f o r 41.3 M P a (60 psig) a n d full v a c u u m . I t was s i t u a t e d in a l o c a t i o n
relatively s h e l t e r e d f r o m t h e e n v i r o n m e n t at a G u l f C o a s t p l a n t . A f t e r six
y e a r s it b e g a n l e a k i n g p r o c e s s t h r o u g h two n o z z l e s in t h e h e a d .
IE. 1. du Pont de Nemours and Company, Inc., Engineering Department, Beaumont, TX

77704.
2E. I. du Pont de Nemours and Company, Inc., Petrochemicals Department, Victoria, TX.
Figure 1 shows the location of the nozzles in the vessel head. They were 7.6
cm (3 in.) and 10.2 cm (4 in.) in diameter, both schedule 40S (wall thick-
ness approximately 5.7 mm). After the leak was detected the vessel was shut
down, the insulation stripped, and the nozzles liquid penetrant inspected to
detect cracking. The two affected nozzles were cut off and examined more
closely.
A second liquid penetrant inspection confirmed that the cracking had
started on the outside (Figs. 2 and 3). Metallographically mounted sections
taken through cracked regions indicated the highly branched morphology
characteristic of chloride stress cracking (Figs. 4 and 5).
The cracking location in both cases was just below the line of contact with
fire retardant mastic, which was applied as a sealant between the nozzles and
thermal insulation applied to the vessel head. Figure 6 shows the location of
cracking in relation to the insulation system.
Investigation of Cause
The vessel head was examined before the insulation was removed. The
mastic had separated part of the way around the cracked nozzles. Since there
was no elastomeric sealant between the nozzles and the mastic, it is quite
3" ~ 41aIlM~II~FLANGE
4" SCH~ 16,0 24"'MANWAY

FLANOE
6" 5 " ~ ] ] ~ ~ 12"
I
FIG. 1--Sketch showing location of nozzles on vessel head.
MONIZ AND RI'I-I'ER ON FIRE RETARDANT MASTIC 97
FIG. 2--Liquid penetrant inspection, 7.6-em (3-in.)nozzle.
FIG. 3--Liquid penetrant inspection, lO.2-em (4-in.)nozzle.
FIG. 4--Section through l O.2-cm r nozzle. Corrosion and stress corrosion cracking. (Ap-
proximately X100).
FIG. S--Section through 7.6-cm (3-in.) nozzle. Stress corrosion cracking. (Approximately
X200).
MONIZ AND Rll-FER ON FIRE RETARDANT MASTIC 99
A INSULATION MASTIC
B SYNTHETIC ORGANIC FIBER CLOTH,
OPEN WEAVE
C CALCIUM SILICATE INSULATION,
INHIBITED
NOZZLE WALL
OCATION OF CRACK
__ Ii : , , rFoR . , S T U . E .
\
VESSEL HEAD
FIG. 6--Insulation system on failed nozzle showing crack location.
possible that the gap had existed for a large part of the time the vessel had
been in service.
Samples of mastic and synthetic organic cloth (which make up the insula-
tion termination system) were analyzed for leachable chlorides and bromides.
The test is described in Table 1. The mastic contained 363-ppm chloride and
36 300-ppm bromide. The organic cloth contained 2010-ppm chloride; it was
not analyzed for bromide (because it is not a fire-retardant product).
In order to detect whether bromide was partly or totally responsible for
cracking, samples of the stress cracked nozzles were broken open and ana-
lyzed using energy dispersive X-ray analysis under the scanning electron mi-
croscope. There was a detection of 4.86 and 6.57 weight % chloride and, sig-
nificantly, no bromide. The results are shown in Table 2.
Note that the analyses obtained by energy dispersive X-ray analysis do not
include elements below atomic n u m b e r 20. The results are normalized, as-
suming no elements below atomic n u m b e r 20 exist.
TABLE 1 - Description of leachable halide test.
5 g of solid is mixed with 200 mL water (If very low CI is suspected, 100-mLwater is used)
The mixture is refluxed overnight at the boiling point.
The mixture is filtered.
The filtrate is analyzed by potentiometric titration for Cl and Br .
TABLE 2--Semiquantitative energy dispersive X-ray spectrometer

analysis of fracture face.
Nozzle, weight %
Element and Line 7.6 cm (3 in.) 10.2 cm (4 in.)
Silicon K alpha 3.89 2.77
Sulfur 15.3
Chloride 6.57 4186
Calcium ... 1.49
Titanium 0.22
Chromium 12194 32.11
Iron 61.77 47.06
Nickel 13.31 10.48
Copper ... 1.01
NOTE:no bromide detected.
Explanation of Failure
The investigation points to chloride stress cracking as the mode of failure.
Despite the possible presence of bromide ions, Br- does not appear to be
responsible in any way for stress cracking.
Chloride stress cracking requires a source of water and usually a chloride
ion evaporation-concentration mechanism. The source of water was most like
the fire protection deluge system (Fig. 7). The system was tested approxi-
mately eight times during the six-year life of the vessel. The deluge water con-
tains 50-ppm chloride and the vessel operates at 120~ In addition the mas-
tic and cloth contain 363- and 2010-ppm leachable chloride, respectively. The
conditions are extremely conducive for chloride stress cracking.
Failure Prevention Program

The failure led to two courses of action:
(1) redesign of insulation termination at nozzles and

(2) analyses of all insulation materials used in the plant for halides.
Redesign has involved improving the seal between the nozzle and termina-
tion using aluminum sheathing and silicone caulk. This is shown in Fig. 8.
Halide analyses have been carried out on samples of insulation materials
used at the plant site. See Table 3 for summary of the analyses. Materials with
leachable chloride in excess of 250 ppm have been rejected. The presently
accepted "rule of t h u m b " is 250 ppm in the chemical process industry. Expe-
rience suggests that material with leachable chloride below approximately
250 ppm will not cause stress cracking when in contact with 300 series stain-
less steel above the accepted threshold temperature of 60~ under evapora-
tion-concentration conditions. Note that the figures described are not hard
MONIZ AND RI-I-IER ON FIRE RETARDANT MASTIC 101
FIG. 7--Deluge system.
9~, INSUL AT K)N ll4ASTIC

B SYNTHETIC ORGANIC FIBER CLOTH, OPEN WEAVE
C CALCILJBASILICATE INSULATION, INHIBITED
D 1S MIL A L U M I N ~ SHEET
D...
I \ ,
V E ~ E L HEAD
FIG. 8--Redesign of insulation termination at nozzles.
TABLE 3-- Test results--leachable halides in insulating materials.
Leachable
Type Form Chloride Bromide
Fibrated asphalt cutback "A" cured sheet 99 < 20

Fibrated asphalt cutback "B" cured sheet 68 < 20
Open weave glass cloth as received < 34 9
Glass cloth with adhesive as received 123 ,
Fire retardant mastic "A" cured sheet 363 36 300
Fire retardant mastic "B" cured sheet 4 420 ,
Vinyl acrylic mastic "A" cured sheet 6 230 < 20
Vinyl acrylic mastic "B" cured sheet 340 100
Vinyl acrylic mastic "C" (nuclear grade) cured sheet 250 94
Vapor barrier membrane as received < 82 ,
Open weave synthetic organic fiber cloth as received 2 010 < 20
Nonsetting sealer "A" cured bead 101 ,
Nonsetting sealer "B" cured bead < 19 9
Flashing compound "A" cured bead 111 9
Flashing compound "B" cured bead 10 ,
Flashing compound "C" cured bead 68 ,
Loaded mastic polymer based cured sheet 171 < 20
Paint, silicone based cured film 120 < 25
Paint, polyamid catalyzed epoxy cured film 190 < 25
Paint, silicone rubber cured film <2 < 25
Paint, epoxy polyamid cured film 10 9
Wax paper used to cure sheet specimens as received < 226 ,
*Not determined.
a n d f a s t b u t a r e g i v e n as t h e g u i d e l i n e s of t h e i n d u s t r y . I n s o m e c a s e s (for
example, good company-wide experience) materials having leachable chlo-
r i d e i n e x c e s s of 2 5 0 p p m a r e n o t r e j e c t e d .
L a t e l y , c o m p a n y g u i d e l i n e s h a v e f u r t h e r e v o l v e d t o p a i n t s t a i n l e s s steel
with two coats of polyamide catalyzed epoxy paint, minimum dry film thick-
n e s s 0 . 0 8 t o 0 . 1 1 m m (3 t o 4 m i l ) u p t o a s e r v i c e t e m p e r a t u r e of 1 2 1 ~
(250~ Also, fiberglass reinforced polyester or thixotropic two-part epoxy
r e i n f o r c e d o p e n w e a v e c l o t h h a s s u p e r c e d e d a l u m i n u m j a c k e t i n g as t h e rec-
ommendation for heads. These materials can be buttered up to nozzles and
reportedly do not break away like glass cloth reinforced mastic.
William G. A s h b a u g h 1
External Stress Corrosion Cracking

of Stainless Steel Under Thermal
Insulation--20 Years Later
REFERENCE: Ashbaugh, W. G., "External Stress Corrosion Cracking of Stainless Steel

Under Thermal Insulation--20 Years Later," Corrosion of Metals Under Thermal Insula-
tion, ASTMSTP880, W. I. Pollock and J. M. Barnhart, Eds., American Society for Test-
ing and Materials, Philadelphia, 1985, pp. 103-113.
ABSTRACT. In the mid-1960s, a rash of costly stress corrosion cracking failures of stain-
less steel under insulation occurred. A series of tests were performed to establish the failure
mechanism and to provide information for prevention. Since that time, the Union Carbide
Corporation (UCC) Central Engineering policy has been to use protective coatings on the
stainless steel. No special insulation materials, inhibitors, or protective weather barriers
have been used to prevent external stress corrosion cracking (ESCC). During these past 20
years, there has been only a few incidents of ESCC of painted stainless steel; therefore a sat-
isfactory pay back on the cost of painting the stainless steel has been realized. There are still
questions to be answered, and optimization of the painting practice is taking place.
Although painting is not the only way to prevent ESCC, it is a proven procedure that can be
accomplished with today's technology.
KEY WORDS: austenitic stainless steel, corrosion, stress corrosion, insulation, thermal
insulation, chloride cracking, external stress corrosion cracking, external stress cracking,
protective coatings, paintings, cracks, concentration cell corrosion
D u r i n g t h e p a s t 20 years, we h a v e h a d c o n s i d e r a b l e e x p e r i e n c e in d e a l i n g
with a t t e m p t s to p r e v e n t e x t e r n a l stress corrosion c r a c k i n g ( E S C C ) of stainless
steel u n d e r t h e r m a l insulation. W i t h t h e o c c a s i o n of this m e e t i n g we t h o u g h t
t h a t it w o u l d b e a p p r o p r i a t e to s u m m a r i z e t h e s e e x p e r i e n c e s a n d i n f o r m a t i o n
in t h e h o p e t h a t it will assist i n d u s t r y in c o m b a t i n g this costly a n d w a s t e f u l f o r m
of corrosion a t t a c k . Several years ago, a survey was m a d e to e v a l u a t e t h e a n n u a l
cost of corrosion. T h e s e dollar figures are h u g e . T h e p o i n t t h a t I f o u n d interest-
ing was t h a t it was e s t i m a t e d t h a t as m u c h as t h r e e q u a r t e r s of this cost c a n b e
p r e v e n t e d u s i n g k n o w n t e c h n o l o g y . I believe t h a t in this c o n f e r e n c e t h e r e is suf-
1Manager of corrosion and materials engineering department, Union Carbide Corporation,

P.O. Box 471, Texas City, TX 77590.
ficient technology to essentially eliminate ESCC, if the knowledge is applied

and used properly. We will begin this paper with a review of our early problems
and then take you through our testing program, the conclusion we have reached
and the procedures and criteria we have employed successfully for the last 20
years to reduce and nearly eliminate this problem.
The Problem
In the early 1960s, a large chemical production unit was built at one of Union
Carbide's Gulf Coast plants. Much of the process equipment in this unit was
American Iron and Steel Institute (AISI) Type 316 L (Unified Numbering Sys-
tem [UNS] 5316) stainless steel. In less than six months, several of the stain-
less steel distillation columns developed leaks in their shells. As we investigated
this, we were shocked to find large areas of tiny cracks visible on the outside of
the column sections. The cracks appeared very much like the root system of
small plants (Figs. 1 and 2). The cracking resulted in drop-wise seeping of pro-
cess liquid through the cracks, which ultimately leaked from under the insula-
tion calling attention to the leaks. What first seemed to be very tiny leaks soon
turned out to be significant areas of badly transgranularly stress corrosion
cracked stainless steel.
As more and more insulation was removed from adjacent vessels, piping,
and columns, other areas of ESCC were found, in many cases, just the begin-
ning traces that had not yet penetrated the vessel walls. Once all the areas
where stress corrosion cracking had occurred were identified and repaired or
replaced, we turned to the question of what caused the cracking. In all cases,
the cracking occurred under cellular glass insulation with the equipment oper-
ating at moderately elevated temperatures. We knew we were looking for chlo-
rides; did they come from the Gulf Coast air, from nearby process units during
FIG. l - - E x t e r n a l stress corrosion crackhtg q/" a Type 310 vessels under cellular glass and anti-
abrasion coathtg.
ASHBAUGH ON EXTERNAL STRESS CORROSION 105
FIG. 2--Another example of ESCC. Here the sample had been bent to open up the main crack.
construction and before insulation was applied, from contact with the ground
before erection, or from the insulation? What was the source of the chlorides?
A review of all possible sources included the cellular glass insulation and the
anti-abrasion coating on the inside surface. When we tested the anti-abrasion
coating, we found that it contained polyvinyl chloride. Upon warming it up in
the presence of water, we found hydrochloric acid! Other sources of chlorides,
while feasible, were not believed to be highly involved since there was nothing
unusual about the exposure of the stainless steel equipment during its fabrica-
tion or erection compared to many other stainless steel items.
The evidence all pointed to the warm, stainless steel equipment operating in
contact with the anti-abrasion coating. Rainwater inside the insulation was the
third necessary ingredient. Any anti-abrasion coating based on polyvinyl chlo-
ride was summarily banned from use in our insulation activities from then on.
The magnitude and the cost of this particular ESCC problem focused our at-
tention on the mechanism of cracking stainless steel under insulation and
brought a number of questions to mind. What other insulation materials might
cause the stress corrosion cracking, under what conditions, and what must be
done to prevent a recurrence of this incident?
The Test Program

At this same time, other companies were suffering from similar ESCC prob-
lems, and were attributing it to one type of insulation or another. These re-
ports, together with our experience, indicated that we needed to determine
which insulation materials were safe and which would actively contribute to
stress corrosion cracking.
A simple laboratory test was developed using a horseshoe bend stainless steel
specimen that was heated and then placed in contact with the insulation mate-
rial (Fig. 3). Distilled water was fed to the interface between the insulation ma-
FIG. 3--Lab test apparatus used to determh,e (f insulation materials would cause stress corro-
sion cracking of Type 316 stainless steel
terial and the stainless steel. After an appropriate exposure time and tempera-
ture, observations were made as to whether or not cracking had developed.
This test procedure was later reported in an article in Materials Protection [1].
Evaluation of some 36 insulation materials using this procedure showed that
only four would, in themselves, cause stress corrosion cracking. If this were
true, or even close to being correct, it indicated to us that insulation as a class
was not responsible for the majority of failures.
While we were engaged in the laboratory testing phase, we had been review-
ing our company history of stainless steel corrosion problems and found scat-
tered, but widespread reports of stress corrosion cracking that were either cor-
rectly diagnosed as ESCC or were described as simply stainless steel cracking
failures without a metallurgical examination. We observed that many of these
cases of ESCC occurred in the absence of thermal insulation. For example, one
of the most prevalent trouble spots was underneath steel slip-on flanges at
stainless steel lap joints (Figs. 4 and 5).
As we completed the test program and reviewed plant operating history, it
became apparent that if we were to attack the ESCC problem by concentrating
on the type or quality of the insulation material, we would be overlooking the
bigger culprit, that is, the total surrounding presence of chlorides from a vari-
ety of sources.
Reconstructing the mechanism of attack, we suggest that the chloride and
water somehow must enter into the annular space between the insulation and
FIG. 4--Example q[ E S C C of a 316 pipe at lap joint. Chloride and water collected under slip-on
flange causing stress corrosion cracking.
FIG. 5--A stainless steel column section sample showbzg E S C C that occurred under a lap ring
flange. This area had not been pabtted as specified.
the outer surface of the stainless steel equipment. In addition, heat and stress
must also be present to allow the stress corrosion cracking to initiate and pro-
ceed. We have further concluded that heat plays a large role in the evaporation
of water and subsequent concentration of chlorides to the point where they be-
come quite aggressive. Since we have no control over the austenitic stainless
steels normal stress patterns, the only other parts of the equation we can deal
with are the presence of chlorides and water.
The use of "inhibited" insulation materials was investigated and rejected be-
cause our lab tests showed they would not prevent stress corrosion cracking
when chlorides were introducted under the insulation. Use of inhibited insula-
tion would limit the insulation designers freedom of materials selection if it
were required for all stainless steel equipment. And finally, it would not offer
protection under flanges, and so forth. This then left us with the consideration
of waterproof insulation systems to exclude water and chlorides or applying a
protective coating to the stainless steel.
To achieve a watertight protective coating for an insulation system, one
would have to produce, in effect, a pressure tight covering. This then would
have to be maintained in this pressure tight condition to prevent ingress of
water and chlorides. Our insulation advisors were quick to point out that such
protective covering would be exorbitantly expensive and that maintaining such
a system, if it were installed, would probably be impractical.
We then concluded that a protective coating on the stainless steel under the
insulation would be the most workable approach. The protective coating sys-
tem also would deal with the stress corrosion cracking problems under flanges,
pipe clamps, and other external appurtenances.
Thus, we came to the conclusion in the middle 1960s that the way to protect
stainless steel from ESCC was to paint it. The mere thought of painting stain-
less steel was at first repelling and strange to many people who had not experi-
enced the problems we have already described. The selection of which paint for
stainless steel also become quite a point of discussion, that is still going on to-
day. However, at that time we selected two paints that we felt would be suitable
based on previous experience with them in other applications.
Painting
In order to effectively proceed with painting stainless steel to protect it from
ESCC, it was necessary to select appropriate paints and generate the painting
criteria or guidelines for what to paint and where to paint. The selection of pro-
tective coatings was based on several criteria such as ease of application, good
adhesion to mill finished stainless steel, and resistance to hot water vapors.
The ability to maintain coating integrity in the presence of hot water vapors
and to withstand temperatures up to about 200~ was important. The 300~
we felt, was the safe upper limit above which stress corrosion cracking would
not occur because the equipment and insulation would be dry. If equipment
were cycling from higher to lower temperatures, it would have to be dealt with
as a special case.
Our first coating selection was a modified, carbon filled silicone that cured at
ambient temperature (Fig. 6). This product was tested on our stress corrosion
cracking apparatus and proved to be effective in preventing cracking from
chloride bearing water. One weakness of the modified silicone was its poor re-
sistance to chemical fumes and spills. We therefore added an epoxy phenolic
coating for those areas where we felt chemical exposure might occur. The selec-
FIG. 6--A portion of the first process unit to have all its stabtless steel painted--both insulated
and uninsulated areas were painted.
tion of protective coatings turned out to be much easier than the development
of a criteria for when and where to use them.
Our original project criteria, dealing with this subject, stated that the respon-
sible materials engineer assigned to the proiect or the plant corrosion engineer
would identify those critical pieces of equipment that needed to be painted.
Our intention was to spend the money for the protective coating only where it
would be most needed and where the pay off in preventing stress corrosion
cracking would be most obvious. Typical items to be painted were stainless
steel pressure vessels, equipment handling corrosive or toxic materials where
leaks would not be acceptable, or one of a kind pieces of equipment whose out-
age would seriously affect a major production unit. Any number of discussions
were held about various projects and plant maintenance jobs to sort out what
was critical and not critical, as well as what was or was not worth painting. At
times, it seemed that it would be cheaper to go ahead and paint rather than to
discuss it; but we were making a real effort to be selective in the use of the pro-
tective coatings through a conscious and logical decision process.
As a result, a regular patch work of painted stainless steel developed in the
late 1960s, and we found that inevitably we had some failures of items that were
not painted, which we wished we had painted. We also found that painting this
item and not painting that item, painting this pipe and not painting another
pipe in the same project was confusing and the cause of mistakes in the con-
struction phase. In 1970, a new unit was being designed and a decision was
made at that time to paint all the stainless steel (Fig. 7). Since most of the unit
would be stainless steel, this was a major decision and change in criteria. This
change was rather like running the first 4-min mile; once we overcame the bar-
rier and reached understanding and acceptance of the new philosophy, paint-
FIG. 7--A view of a Type 316 L column that had been in service about 12years. It was painted
with the rood(fled silicone.
ing of all stainless steel, as a general rule, became easier. Exceptions could now
be made as to where paint was not required, rather than as previously, when
painting was the exception.
Experience
Up to this point we have been describing our past experience, history, and
philosophy. Where are we today? After almost exactly 20 years of using paints
to protect stainless steel from stress corrosion cracking, our experience has
been excellent. We are convinced that a proper paint film protects stainless
steel for many years and that the small cost for this kind of insurance has been
paid for many times over.
But all is not perfect in this world and we have had some problems. We have
had piping and column sections painted where the painters did not pull back
the lap flange rings and paint underneath them: the result was ESCC. Painted
vessels had the paint damaged by steel cable slings, tearing and scarring the
paint: the result was ESCC. A series of small tanks that were insulated with cell-
ular glass, and contrary to explicit instructions, a polyvinyl chloride contain-
ing anti-abrasion coating was used as a cement to glue the cellular glass to the
painted vessel. The solvents in the adhesive/anti-abrasion coating penetrated
and destroyed the protective coating allowing the soon to be formed hydrogen
chloride to do its stress corrosion cracking work: the result was ESCC. A substi-
tution was made for paint specified, the paint failed: the result was ESCC.
Painting was specified for small process piping; the project manager needed to
save money and cut out the painting: the result was ESCC.
Another problem area is the total paint job on stainless steel equipment,
which has steel flanges and attachments. We know that the one coat of modified
silicone will not offer protection to carbon steel (Figs. 7 and 8). The epoxy phe-
nolic must be used to protect the steel. (Note: We do not allow zinc primers on
steel in contact with stainless steel.) Thus, a careful reading of our engineering
FIG. 8--A closer view of the column shows how this silicone paint has disappeared from the steel
rein/brcing rings. The paint on the stabdess steel is still in good condition.
criteria would have portions of a stainless steel vessel painted with the epoxy
phenolic and other portions painted with modified silicone (Figs. 9 and 10).
This makes for a colorful paint job but presents possibilities for confusion by
the fabricator or painter and probably extra cost compared to a single paint sys-
tem. As a result, we now use the epoxy phenolic for the entire vessel provided it
is compatible with the process temperature (Figs. 11 and 12).
Summary
In summary, our position and experience regarding stress corrosion crack-
ing of stainless steel under insulation indicates that the most practical way to
FIG. 9--A new heat exchanger that has its stainless steel surface painted with modified silicone
and its steel painted with epoxy phenolic.
FIG. IO--A combhtation of steel attd stahdess steel, each with its own paint spec(fieation.
FIG. 11--A short column section of Type 316 L with all parts pahtted whh epoxy phenolic.
FIG. 12--A large distillation column, Type 304 stahtless steel behtg insulated. Both stahtless
steel and steel portions are pab~ted with epoxy phenolic.
protect the stainless steel is with a coat of good paint. While our results over the
past 20 years have not been perfect, the few and isolated failures have essen-
tially been physical problems, which actually served to demonstrate that while
the paint film is intact, stress corrosion cracking is prevented. It is also worth
noting that the limited cracking that occurs at defects in the paint film is repair-
able. A general stress corrosion cracking failure often resists repair thus requir-
ing replacement.
The use of a protective coating on stainless steel to ward off chloride stress
corrosion cracking may be looked at like the wearing of a seat belt in an auto-
mobile. The seat belt does not guarantee that somebody will not come up and
hit you, but if it happens, your risk of serious injury and death is considerably
less. So the protective coating on stainless steel has to be evaluated against the
risk and consequences of ultimate stress corrosion cracking failaure. We can-
not predict when or how chlorides might invade the insulation system, but we
can trust the coating to protect the stainless steel when the inevitable happens.
Reference
[1] Ashbaugh, W. G., "ESCC of Stainless Steel Under Thermal Insulation," Materials Protec-
tion, May 1965, pp. 19-23.
D o n a l d O. Taylor 1 and R o d n e y D. Bennett 1
Shell and Jacket Corrosion of a

Foamed In-Place Thermally
Insulated Liquefied Petroleum
Gas Tank
REFERENCE: Taylor, D. O. and Bennett, R. D., "Shell and Jacket Corrosion of a

Foamed In-Place Thermally Insulated Liquefied Petroleum Gas Tank," Corrosion of
Metals Under Thermal Insulation, A S T M STP 880, W. I. Pollock and J. M. Barnhart,
Eds., American Society for Testing and Materials, Philadelphia, 198S, pp. 114-120.
ABSTRACT: An 87.78 m (288 ft) diameter by 24.77 m (81.25 ft) high liquid butane stor-
age tank was insulated in 1976 using vertically foamed in-place (FIP) low flame spread
polyurethane with an aluminum outer jacket. The tank had not been painted since the
vendor stated the urethane would not properly bond to painted surfaces. In 1979 corro-
sion of the aluminum jacket and aluminum rivets was noted. Chemicals, leached from the
urethane foam, resulted in streaks of discoloration down the sides of the tank. Detailed
inspection in 1980 involved cutting out sections of jacket and insulation to uncover the
tank shell. Extensive pitting was noted on the steel shell as well as the foam side of the
aluminum jacket. Corrosion is attributed to the intrusion of water from a seawater deluge
system that leached the acidic fire retardant chemicals from the urethane foam. A task
force was organized to investigate and evaluate various methods for insulating large cyro-
genic liquid storage tanks. The study involved visitations both in the United States and
Europe. As a result, detailed specifications were developed for the blasting, painting,
installation of foam glass insulation, and weather proofing of the tank. Since corrosion
had apparently reduced the tank wall thickness below that required by American Petro-
leum Institute (API) Recommended Rules for Design and Construction of Large, Welded
Low Pressure Storage Tanks (API 620), detailed analyses of mechanical properties of the
steel were investigated, a program was established to acoustically monitor the tank during
hydrotest, and procedures were established for minor localized weld repair to the deepest
pits. Details of the inspection procedure, testing, painting, and insulation specifications
are discussed.
KEY WORDS: insulation, tanks (containers), corrosion, polyurethane, butane storage
1Engineering consultant group leader of Process Department and supervisor of Technical Sup-
port Unit, Technical Engineering, respectively, Arabian American Oil Co., Process Department,
E-6100 Engineering Building, Dhahran, Saudi Arabia.
Copyright9 by ASTM lntcrnational www.astm.org
TAYLOR AND BENNETT ON AN LPG TANK 115
Experience with corrosion under insulation is widely varied. In carbon or

low alloy steel tanks the corrosion may take the form of simple rusting, or
dependent upon the material that infiltrates behind the insulation, severe pit-
ting and even stress cracking have been encountered. Stainless steel tanks
have also failed because of leaching of chlorides from insulation or ingress of
other chloride containing contaminants or both. Corrosion of the metallic
moisture barrier or deterioration of a nonmetallic moisture barrier permits
ingress of water and airborne chemicals that accelerate attack.
This paper will cover the details of the original insulation of a liquid butane
storage tank, discovery of progressive corrosion, tank testing, and the reinsu-
lation procedures and specifications.
Discussion
In 1976 an 87.78 m (288 ft) diameter by 24.77 m (81.25 ft) high liquid
butane storage tank operating at --40~ was insulated using vertically
foamed in-place polyurethane with fire retardant additives and an aluminum
sheeting weather barrier (Figs. 1 and 2). The carbon steel tank was not
painted before insulation since the vendor stated the urethane would not
properly bond to painted surfaces. No provision was made to prevent mois-
ture ingress other than a sealant between the overlap of the aluminum sheet-
ing, which was pop riveted together.
FIG. 1--Rust breakthrough on vertical f o a m e d in-place urethane insulation.
FIG. 2--Close-up of Fig. 1.
In 1979 there was external evidence in the form of large rust stains indicat-
ing failure of the insulation system. Random areas of the sheeting and poly-
urethane were removed to permit visual inspection and ultrasonic thickness
measurement of the steel shell. The polyurethane foam was generally found to
be water saturated for at least 152.4 m m (6 in.) on either side of the aluminum
overlap joint. Some areas in the lower region of the tank were wet throughout.
Earlier data gathered from inspection of insulated butane lines indicated
expected corrosion rates of 0.254 m m (0.01 in.) per year although lines had
been painted. Thus plant experience indicated that corrosion rates in excess
of 0.254 m m (0.01 in.) per year might reasonably be expected on the un-
painted steel tank shell.
Randomly sampled areas were found to have a general metal corrosion loss
of 2.54 m m (0.100 in.) plus as much as 2.54 m m of pitting in some locations!
Further investigation verified that the water leached the acidic fire retard-
ants (fluorine-bromine compounds) from the urethane. These acidic addi-
tives, plus breakdown of the urethane foam by intense sunlight and periodic
seawater deluge, resulted in acid solutions having a pH as low as 1.7. Addi-
tional high chloride water was introduced by the weekly testing of the seawa-
ter deluge system.
Coinciding with the early stages of this investigation there was an on-going
project to reinsulate 80 000 butane liquid (bbl) tanks with a horizontal
foamed in-place urethane system (Fig. 3). The procedure was modified to
paint the tank shell with red oxide primer before insulation.
An additional skirt modification to shed water was added to the bottom of
the tanks to avoid the collection of water at the interface of the aluminum
FIG. 3--Rust breakthrough on horizontal foamed in-place urethane insulation.
sheeting and the ring wall foundations. This skirt modification was used in
the construction of 7 new 750 000 bbl tanks. Five additional terminal tanks
were constructed in 1980 through 1981 with an epoxy primer, full ring wall
skirt and the horizontal foamed in-place urethane system.
In mid 1980 an Engineering Task Force Group was formed to conduct a
worldwide investigation of cold insulation systems. After consulting with
more than 20 different companies and visiting some 22 sites with cold service
storage facilities operating at temperatures from --10 ~ to --2~ (14 to 28~
the task force concluded that no two companies used the same system. Al-
though a decision had not been reached as to the optimum method of insula-
tion, it became critical to remove one of the 895 000 bbl tanks from service
because of evidence of accelerating corrosion rates on the steel shell. The tank
was stripped of insulation, sandblasted, marked off by square foot, and had
ultrasonic thickness measurements made and recorded. All pitted areas were
also minutely inspected and gaged. Results showed general corrosion loss
over the entire tank. The areas in contact with the wet polyurethane insula-
tion, more than 40% of the tank, showed heavy metal loss of 2.54 to 5.08 m m
with extensive deep pits, a corrosion rate of 0.68 to 1.35 m m per year (based
on actual service life).
In late 1980 the decision was made to use 152.4 m m (6 in.) of adhered
cellular glass insulation covered with a 2.03 m m (.080 in.) thick moisture bar-
rier of polyester mesh reinforced acrylic/aliphatic urethane. Detailed stress
calculations indicated that the steel shell could adequately be repaired by fill-
ing the critically deep pits with weld metal that was then ground flush. Ac-
cordingly, the shell was repaired, sandblasted to white metal, and coated with
an inorganic zinc silicate (Fig. 4). The tank was then hydrotested to the maxi-
mum fill level, and acoustic emission was monitored during the test to detect
any crack or flaw propagation. No indications were recorded.
The cellular glass blocks were installed in two layers. The first 101.4 mm (4
in.) thick layer was completely bonded to the tank shell and to each adjacent
block using a polyurethane adhesive. The joints in the second 50.8 mm (2 in.)
thick layer were offset 50% to prevent any joint in the second layer lining up
with a joint in the first layer. Extreme care and rigid inspection insured that
all blocks were bonded to adjacent blocks and that all joints were completely
filled with the urethane adhesive (Fig. 5). The contractor was required to have
a technical representative from the insulation manufacturer present at all
times during application of the cellular glass to provide training and inspec-
tion assistance.
After all cellular glass block had been installed, a coat of high solids acrylic
mastic was applied at 0.76-mm (0.030-in.) dry film thickness, and while still
wet a knitted open mesh polyester cloth was rolled in to achieve complete wet-
out and a smooth surface. A second coat of acrylic mastic was then applied to
achieve a total dry film thickness, including fabric, of 1.78 mm (0.070 in.).
After complete curing, inspection, and repair of any imperfections or bub-
bles, two coats of an aliphatic urethane were applied to a dry film thickness of
FIG. 4--Scaffolding--before application of foam glass.
FIG. 5--Application of f o a m glass, note "buttered joints. "
0.254 mm (0.010 in.). Thus, the total thickness of the completed vapor bar-
rier system was 2.03 mm (0.080 in.). This system is not only highly imperme-
able to moisture but is also very resistant to damage by wind blown objects. In
case of such damage it is readily repaired, access to the damaged area being
the biggest problem.
A very important aspect of the job, which is often neglected, is the proper
sealing of areas where there are penetrations through the insulation. Such
might be ladder supports, gage connections, nozzles, and so forth. The seal-
ant must be readily applied (preferably by a gun), must have excellent bond to
the surfaces to which it is applied, must be nonshrinkable (thus must have
minimal solvents), must have excellent weatherability, and should be capable
of accepting a paint cover coat if desired (Fig. 6).
Summary
The most significant factor in the massive corrosion of the shell and jacket
was determined to be the acidic fire retardant solution leached from the poly-
urethane foam by the moisture penetration of the inadequately sealed jacket
FIG. 6--Completed foam glass-barrier coated tank.
joints. The high humidity of the marine environment and the repeated exer-
cise of the deluge system accelerated the water ingress. The decision not to
paint the tank before application of the polyurethane foam permitted direct
exposure of the steel to the corrosive environment.
The new system has now been in service for over one year with no signs of
any deterioration. The fiber reinforced polyurethane exterior retains its gloss
finish and pleasing appearance. There is no evidence of moisture infiltration
at any point.
Thus we feel it has been effectively demonstrated that a proper, cost effec-
tive, insulating system can be accomplished if necessary care is given to sys-
tem selection, specification, training, and inspection plus time and money.
Acknowledgment
The authors gratefully acknowledge the cooperation of their fellow col-
leagues for their assistance in gathering data and photographs. We also wish
to thank the Arabian American Oil Company of Saudi Arabia for their per-
mission to publish and present this paper.
William G. A s h b a u g h 1 a n d T h o m a s F. Laundrie 1
A Study of Corrosion of Steel Under

a Variety of Thermal Insulation
Materials
REFERENCE: Ashbaugh, W. G. and Laundrie, T. F., "A Study of Corrosion of Steel

Under a Variety of Thermal Insulation Materials," Corrosion of Metals Under Thermal
Insulation, A S T M STP 880, W. 1. Pollock and J. M. Barnhart, Eds., American Society
ABSTRACT: Corrosion of steel equipment under thermal insulation is and has been a
costly problem in Union Carbide Corporation (UCC) chemical operations as well as other
refining and chemical plants throughout the world. The Materials Technology Institute
(MT1) has spent about $30 t h o u s a n d trying to develop a nondestructive inspection
method to detect and measure corrosion under insulation. The corrosion rate of carbon
steel under insulation depends upon a n u m b e r of variables: moisture, oxygen availability,
metal temperature, paint system, and the insulation. Little work has been published that
defines the effects different types of insulation have on the corrosion rate of carbon steel.
Some suggest that the type of insulation does make a difference. An in plant test was
designed and operated to obtain quantitative data comparing several types of insulation
and their effect on eon-osion of steel.
Twelve pieces of insulation were applied to carbon steel pipe and were exposed to the
atmosphere at Texas City, TX. Water, in addition to rain, was occasionally added to the
outside of the pipe, and steam was run through the inside of the pipe once a week in order
to accelerate the corrosion rate.
This paper will report qualitatively and quantitatively on the corrosion rates of the car-
bon steel under the various types of insulation. The data will be useful to insulation spe-
cialists in guiding their selection of the kind of insulation from the standpoint of its influ-
ence on steel corrosion.
KEY WORDS: corrosion, thermal insulation, steels, paints, coatings, pipes, thermal cy-
cling, industrial atmosphere
Corrosion of steel equipment under thermal insulation is and has been a

costly problem in the Union Carbide Corporation (UCC) chemical operations
as well as other refining and chemical plants throughout the world. The cot-
'Manager of corrosion and materials engineering, and corrosion and material engineer, re-
spectively, Union Carbide Corporation, P.O. Box 471, Texas City, TX 77590.
rosion rate of carbon steel under insulation depends upon a number of vari-
ables: moisture, oxygen availability, metal temperature, type of paint system,
and the insulation. Until now, little information has been published on the
effects different types of insulation have on the corrosion rate of carbon steel.
An accelerated test was run at a Gulf Coast plant using twelve different types
of insulation for a one year period. Qualitative and quantitative measure-
ments were made on the corrosion rates of carbon steel under the various
types of insulation.
In general, the types of insulation used had a small effect on the corrosion
rates of carbon steel compared to the other factors such as moisture and oxy-
gen availability. The corrosion rate was worse where the moisture was allowed
to collect on the pipe, such as underneath where the insulation hung away
from the pipe or at the edges of the test pieces. Some of the insulation types
drew the water away from the pipe and reduced the corrosion rate. An ex-
tended test of three years or more would be needed to bring out the subtle
differences that may exist between the types of insulation.
A protective coating of paint is effective in preventing corrosion, regardless
of the type of insulation and is still our recommended procedure.
Background
The problem of corrosion under thermal insulation is well known in the
refining and petrochemical industry. It has been a growing concern as more
and more cases come to light. Two recent examples are shown in Figs. 1 and
2. An increasing number of papers are being written on the subject. The Ma-
terials Technology Institute (MTI) has spent about $30 thousand trying to
FIG. 1--Example of corrosion under insulation. Gas heater periodically steamed out, only
had an inorganic zinc primer originally, and was insulated for personnel proteetion after it had
been in service f o r awhile. The hole resulted in afire.
ASHBAUGH AND LAUNDRIE ON CORROSION OF STEEL 123
FIG. 2--Example of corrosion under insulation. Th& area is in the kettle section of a column,
which was slightly above ambient. Upper portions of the column run well below ambient and
condense moisture, which runs down the wall. The proper coating had been used, but it had been
applied too thin. Pinholes allowed the corrosion to get started.
develop a nondestructive method to detect and measure corrosion under insu-

lation, so far without success [1,2]. A company memo based on UCC's experi-
ence and test work was written in 1980 by W. G. Ashbaugh entitled "Corro-
sion of Steel Under Thermal Insulation, The Hidden Danger."
Introduction
It has been suggested that the corrosion rate of carbon steel is influenced by
different types of insulation. Virtually no quantitative data exist on this sub-
ject. This accelerated test was designed to obtain such data under carefully
controlled conditions. This data will be useful to insulation specialists in
guiding their selection of the kind of insulation from the standpoint of its
influence on steel corrosion.
Procedure
The Test
Twelve pieces of insulation, each approximately 0.6 m long (2 it), were ap-
plied to 7.62-cm (3-in.) carbon steel pipe. The insulation was wrapped with
an aluminum weather barrier, but the ends were left open. The insulation
samples were spaced approximately 0.3 m (1 ft) apart on the pipe. The mate-
rials were then exposed to the atmosphere at the plant paint test site, located
just north of a cooling tower, for approximately one year (Figs. 3 through 5).
Before the insulation was installed, the steel pipe was sandblasted to a
near-white metal finish and painted with an epoxy-phenolic paint covering
about one-fourth of the outer circumference as a longitudinal stripe.
FIG. 3--Insulation test site north td" the cooling tower. Test stand be~ore steam tubing was
attached.
FIG. 4--Closer view of Fig. 3.
Seventy-pound steam was run through the pipe for 15 min once a week in
order to expose the steel to a cycle of cold/wet to hot/dry in the expectation
that this would disturb the semi-protective oxide films that develop as steel
rusts. When the steam was turned on, the pipe wall temperature would level
out at 93~ within 5 min.
Originally, it was hoped that rainwater would be an adequate source of
moisture for the corrosion of the pipe. However, after three months, a few
pieces of insulation were removed, and it was discovered that very little corro-
FIG. 5--Insulation being applied to the pipe by skilled insulators.
sion was occurring under the insulation. Rain fell during this period only 14
of the 90 days.
A number of holes were then punched through the weather barrier and the
insulation, and 0.24 L of water was squirted into the holes of each piece of
insulation twice a week, if it did not rain. The water used was the plant un-
treated river water.
The above steps, particularly the water addition and thermal cycling were
taken to provide an accelerated corrosion test so that useful data could be
obtained in one year's time.
Moisture sensors 2 were installed under the insulation on some of the test
pieces to detect the presence or absence of water between the pipe and the
inner surface of the insulation (Fig. 6).
The Evaluation
The wall thickness of the pipe was measured before and after the test with
an ultrasonic thickness instrument by the Plant Inspection Department. Vis-
ual observations were made, and data were recorded of the heating and water-
ing cycles of the pipe during the test. Observations were recorded for each
piece of insulation concerning the appearance of the pipe immediately after
the removal of the insulation at the end of the test (Fig. 7). Thickness mea-
surements were taken under the middle and near the edges of each of the
insulation samples. Some of the worst looking areas near the edges of the
insulation were measured in more detail (Fig. 8). It was also noted how far
from the edge of the insulation the severe corrosion occurred. The general
2Sereda Miniature Sensor, Model SMMS-01, Epitek Electronics Inc., Ogdensburg, NY.
FIG. 6 - - S e t u p near the end of the test. Plastic cups protecting moisture sensor leads can be
seen.
FIG. 7 - - T h e pipe and baked expanded Perlite insulation at the end of the test.
degree of pitting was recorded under the middle of the insulation piece on the
top of the pipe, the bottom of the pipe, and the side of the pipe. A forced
ranking of the insulation was made.
Discussion and Results
The corrosion of the bare pipe between the insulation pieces was more se-
vere than that underneath any of the insulation samples. The rate over the
FIG. 8--Part of the pipe after the test and cleanup. Ultrasonic thickness measurements were
taken at this time.
one year period averaged about 50 mil per year (mpy, where 1 mil ----0.001 in.
---- 0.0254 mm). The iron oxide formation seemed to be rather rapid in the
first half of the year and then slowed down somewhat.
A steel corrosion coupon was exposed in the test area to determine the cor-
rosion rate on bare steel by weight measurements. This coupon showed a cor-
rosion rate of 18 mpy as opposed to the 50 mpy on the pipe. This difference
was probably caused by two factors:
9 The reference sample was hanging in a vertical position, allowing it to
dry faster than the pipe.
9 The sample was not subject to the heating cycle that the pipe was put
through every week to deliberately increase the corrosion rate.
The corrosion rate of the reference coupon is much higher than the rural
atmospheric corrosion rate on steel; but is typical of a chemical industrial/
marine atmosphere, particularly downwind of a cooling tower. This spot was
chosen for the accelerated test.
The heating cycle was selected to dry out the surface of the pipe and its rust
scale, yet not dry out all of the moisture that had accumulated in the insula-
tion. Observations during the test confirmed this mode of operation.
Each week, if it did not rain, some untreated Brazos River water was
squirted among the holes of each of the insulation samples. The pipe was
watered approximately twice a week, whether it was from rain or artificial
watering. It was later discovered that the corrosion rate directly under the
water holes was virtually the same as that on the bare steel. Two points specif-
ically measured showed a corrosion rate of 47 and 53 mpy. These spots were
not included in judging the effect the insulation had on the corrosion rate of
the steel.
Several moisture sensors (Fig. 9) were used as a " G o - N o - G o " indication of
FIG. 9--Moisture sensor used in the test. Alternating strips of gold and copper set up a poten-
tial difference when an electrolyte is present gthat is, water).
moisture under the insulation. They showed that the pipe surface did dry out
rapidly while the steam heated up the pipe. The sensors failed over a period of
three months. However, enough data were obtained to show that the holes
and water addition did help bring moisture to the pipe-insulation interface,
whereas before the holes were punched, it had stayed relatively dry.
A paint stripe covering approximately one-quarter of the pipe's outside sur-
face was applied to the entire length of the pipe. It was one coat of Plasite
7122, 3 an epoxy phenolic. We normally specify that two coats of an epoxy
phenolic be put on carbon steel under insulation. The one coat system was
used to evaluate the performance of the thinner coating. This paint stripe was
6 to 7 mil thick and held up very well, both under the insulation and between
the pieces of insulation on the exposed pipe. There was some undercutting
along the edges, but this was to be expected. Overall, the paint performed
well and the corrosion rate under it was nil.
The corrosion under the edge of the insulation was almost as severe as that
outside the insulation. This was expected because there was nearly equal ac-
cess to water and oxygen at this point. Moving from the exposed steel to the
middle of a piece of insulation, the corrosion rate generally dropped off
quickly. The distance that the high corrosion rate continued to occur from the
edge of the insulation on inward, in most cases, was 2.5 to 5.0 cm (1 to 2 in.).
In all of the types of insulation tested, the corrosion was usually worse along
the bottom of the pipe than on the sides or top. Along the bottom where a
piece of insulation would hang away from the pipe, the moisture would form a
drip line and not be drawn away by the insulation. The pipe was installed on
slight angle (approximately 5 ~ see Fig. 4) to allow for draining of the conden-
sate when steam was put into the pipe. This also affected the direction the
rainwater entered, and the corrosion under the edge of the insulation was
generally worse on the "high end" than on the "low end" of each piece.
See Table 1 for a summary of the results. Table 2 lists the detailed data
"~Wisconsin Protective Coating Company, Union Carbide Corp. Paint Specification PCM-214.
acquired for the evaluations. The following observations listed in Table 2

were used in decreasing order of importance to establish the ranking in Table
1: (a) typical metal loss; (b) extent of the pitting on top, sides, and bottom; (c)
extent of the heavy corrosion in from the edge of the sample; and (d) worst
metal loss at the edge of the sample.
Two types of expanded perlite insulation were tested. For each type, we
tested one as received and one after it had been baked for about 24 h at 260~
to simulate what might happen to the insulation on a steam line or a heated
line after a period of time. The corrosion rate under the unbaked perlite insu-
lation dropped off quickly less than 2.5 cm from the edge, but the baked
perlite, which was falling apart, allowed high rates to proceed up to 7.5 cm
under the insulation. Some of the pitting that occurred in this area was as bad
as on the bare steel pipe. The baked insulation was partially broken apart and
mushy in texture. The corrosion rates in the middle of the baked and un-
baked perlite samples, however, were not significantly different from each
other.
The expanded Perlite gave the best performance of the insulation types. It
produced the lowest corrosion rates where the insulation touched the pipe
near the middle of the sample. The four different types of calcium silicate
performed about the same. The worst performers were the nonabsorbent cell-
ular glass and the mineral wool whose loose fibers stuck into the top iron
oxide layers.
TABLE 1--Results.
Rank Insulation (No.) Comments
1 expanded perlite (1) Line of pitting 2.5 to 5.0 cm wide on bottom

2 expanded perlite (2) Line of pitting 2.5 to 5.0 cm wide on bottom
3 calcium silicate (4) Rank items 3, 4, 5, and 6 are about the same in
4 calcium silicate (3) performance. Loose fitting in that insulation hung
5 calcium silicate (2) away from bottom. Drops of water hanging on
6 calcium silicate (1) pipe underneath. Insulation very heavy with water
when removed.
7 fibrous glass More pitting along bottom and edges than calcium
silicates.
8 expanded perlite baked (1) Insulation fell apart easily.
Much more pitting near the edges.
Some more pitting underneath.
9 expanded perlite baked (2) Insulation fell apart easily.
Much more pitting near the edges.
Some more pitting underneath.
10 polyurethane foam Only one where pitting was fairly uniform all
around the pipe. Very little advanced corrosion
under the edges.
11 cellular glass Corrosion in middle about as bad as that near the
edges.
12 mineral wool Some insulation stuck to pipe.
..g
TABLE 2--Corrosion under insulation data a n d c o m m e n t s . r

o
Amount of Corrosion
0
0
~D
Worst Metal Pit Depth in Middle" ~D
Typical Metal Loss at Edges, Extent Ranking by 0
Loss, i n ? in. of Edge 1-Least 5-Most Appearance J
Effect, Before 5
z
Rank" Middle Edges Up Down in. Top Side Bottom Sandblasting Comments
0
-I1
1 0.008 0.010 0.046 0.010 2 2 5 unbaked expanded perlite
line of pitting along bottom 2.5 to 5.0 cm m
--I
wide.
2 3 5 unbaked expanded perlite f--
2 0.018 0.029 0.032 0.029
line of pitting along bottom 2.5 to 5.0 cm
wide
3 0.016 0.018 0.027 0.028 1.5 2 2 4 calcium silicate types all about the same
4 0.013 0.025 0.031 0.042 2 1 2 3 calcium silicate types all about the same

7 0.014 0.034 0.054 0.030 3 2 3 4 worst case of high rates in from edge for
unbroken insulation
8 0.009 0.026 0.050 0.028 2 2 3 4 baked expanded perlite
lot of edge effect and broken insulation
9 0.016 0.037 0.062 0.028 3 1 2 4 8 same as No. 8. rank
10 0.019 0.028 0.024 0.023 1 3 3 3 10 uniform pitting on all sides, very little edge
effect
11 0.026 0.033 0.053 0.021 2 5 5 11 corrosion in middle about as bad as at
edges
12 0.035 0.040 0.053 0.055 1 to2 2 4 4 12 some insulation stuck to pipe
"See Table 1 results for insulation identification.

hl in. = 2.54 cm.
' Visual after sandblasting.
dVisual after removal of insulation.
Conclusions
Although this test produced no final answers, the results have added to our
understanding of the several mechanisms that are taking place between the
steel-water-insulation combination. We may conclude the following from the
test results:
9 Corrosion rates differed slightly under the various types of insulation. A
similar test of three years or more would be needed to bring out the subtle
differences between materials.
9 The more corrosion occurred the longer moisture was in contact with the
pipe. This was mainly on the underside of the pipe and at the elevated end of
the insulation sample.
9 Insulation that trapped the moisture against the pipe, without drawing it
away by absorbing it, produced the most corrosion. The pipe in contact with
insulation that was absorbent had a relatively lower corrosion rate.
9 The corrosion rate of the bare steel between insulation test pieces was
higher than normal because of the thermal cycling.
9 The epoxy-phenolic paint did an excellent job of protecting the steel and
preventing corrosion.
Summary
With a good weather barrier, the nonabsorbent insulation may actually
help keep the moisture away from the pipe. Where there are breaks in the
insulation covering, water can get under the insulation and be trapped
against the steel. The corrosion rate is then largely dependent on how long the
moisture is in contact with the steel and the thermal operating conditions.
Absorbent insulation types aid in removing the water from the metal surface.
They lose their advantage, however, if there is a continuous source of water
such as condensation or process leaks.
The present UCC Engineering policy of painting steel under insulation
(PCM-500, UCC Painting and Coating TechnologyManual) continues to be
the most effective way to prevent corrosion.
Acknowledgment
We wish to acknowledge the assistance of E. C. Powell, Jr. in planning the
test program and J. H. Rapp in obtaining the insulation materials and having
them installed.
References
[1] "Investigation of Nondestructive Testing Techniques for Detecting Corrosion of Steel Un-
der Insulation," MIT Report 4, July 1981.
[2] "Investigationof an Approach to Detectionof Corrosion Under Insulation," M1T Report 7,
March 1982.
James R. Myers 1 and A r t h u r Cohen 2
Behavior of a Copper Water Tube

Exposed to Natural Carbonaceous
Granular and Cellulosic Insulation
Materials
REFERENCE: Myers, J. R. and Cohen, A., "Behavior of a Copper Water Tube Exposed
to Natural Carbonaceous Granular and Cellulosic Insulation Materials," Corrosion of
Metals Under Thermal lnsulalion. A S T M STP 880, W. I. Pollock and J. M. Barnhart,
Eds., American Society for Testing and Materials, Philadelphia, 1985, pp. 132-142.
ABSTRACT: Copper is essentially i m m u n e to corrosion in most underground environ-

ments because of the protective films that naturally form on the metal's surface. In gen-
eral, there is no major concern regarding the corrosion of copper underground unless it is
exposed to certain aggressive substances. These include moist environments containing
(1) appreciable quantities of chlorides, sulfates, sulfides, or a m m o n i u m compounds or
both; (2) organic or inorganic acids or both; or (3) active anaerobic bacteria or all three. It
is also known that copper should not be directly embedded in cinders that generally con-
tain sulfides or carbon or both. Fortunately, when corrosive conditions exist, a variety of
cost effective techniques can be used to mitigate unacceptable corrosion of buried copper.
In spite of these well known general guidelines regarding the underground behavior of
copper, situations do occur where the metal is placed in contact with aggressive thermal
insulations. When these insulations are hygroscopic, have low resistivities and acidic pH
values, and contain sulfur compounds along with carbon, it is understandable that unac-
ceptable corrosion of the copper will occur.
The conditions that caused natural carbonaceous granular "thermal insulation" to de-
teriorate the underground copper tube domestic hot water system for a low-rise high den-
sity housing development are presented. Basically, the copper water tubes corroded be-
cause the "thermal insulation" absorbed large quantities of moisture, had a low resistivity
(less than 1200 fl.cm), and an acidic pH (as low as 4.8). Corrosion was facilitated by the
presence of appreciable quantities of sulfur (8000 ppm) and possibly carbon in the
"insulation."
Corrosion of the adversely affected copper water tube system was mitigated by cathodic
protection. In order to minimize shielding effects created by the complexity of the under-
ground system, deep anode beds were incorporated into the cathodic protection design.
1President, JRM Associates, 4198 Merlyn Dr., Franklin, OH 45005.

2Manager of s t a n d a r d s and safety engineering, Copper Development Association Inc.,
Greenwich Office Park 2, Greenwich, CT 06836-1840.
MYERS AND COHEN ON A COPPER WATER TUBE 133
The unaffected underground copper tube domestic cold water and steel gas distribution
systems were bonded to the cathodically protected copper hot water tubes in order to
prevent the possibility of stray current corrosion.
In addition, other insulation systems have been shown to be deleterious to copper. For
example, certain chemicallytreated cellulosic insulation materials can contain constitu-
ents that become aggressiveto copper when either moist or under high relative humidity
conditions.
KEY WORDS. corrosion, insulation, copper, tliermal insulation
The outside surface of copper water tube is essentially immune to corrosion

in most building system environments because of the protective films that
naturally form on the metal surface. In general, there is no major concern
regarding the external corrosion of copper water tube in building systems un-
less it is exposed to certain aggressive conditions or environments. These in-
clude moist environments containing (1) appreciable quantities of chlorides,
sulfates, sulfides, or ammonium compounds; (2) organic or inorganic acids;
or (3) active anaerobic bacteria [1-4]. It is also known that copper should not
be directly embedded in cinders that generally contain sulfides or carbon or
both.
In spite of these well established guidelines regarding copper, situations
occasionally occur where the metal is in contact with an aggressive thermal
insulating environment. For example, if the insulation is hygroscopic, has a
low resistivity, has an acidic pH value when wet, and contains sulfur along
with carbon, unacceptable corrosion of the copper can occur.
This paper identifies two such adverse conditions involving thermal insu-
lating substances that have been known to cause the corrosion of copper water
tube. Certain natural carbonaceous granular materials used for underground
thermal insulation contain appreciable quantities of carbon and sulfur.
These insulations have low resistivities and acidic pH values when moist.
Such an environment will be aggressive and should be avoided.
Certain chemically treated cellulosic-type insulations are aggressive to cop-
per if they contain specific constituents (for example, ammonium sulfate) and
become wet or are exposed to a high humidity environment. The corrosive
conditions created by a wet ammoniacal environment are further aggravated
if the insulation/moisture mixture has a low pH. Water extracts of certain
cellulosic type insulations have pH values as low as 3.7.
Moisture is a necessary prerequisite for even a potentially aggressive ther-
mal insulation to be corrosive. It is often difficult to prevent the ingress of
water as in the case of underground thermal insulations especially if they are
installed below the water table. Chilled water lines can condense moisture in
high humidity environments unless the thermal insulations are covered with
an effective vapor barrier.
When reasonable care is taken in the selection and installation of thermal
insulations, however, copper tube corrosion can be prevented.
Natural Carbonaceous Granular Insulation

Natural carbonaceous granular material (that is, finely pulverized subbi-
tuminous coal) has been used for insulating underground copper water tube.
Typically, properly compacted, finely pulverized, subbituminous coal should
provide copper with excellent thermal insulating characteristics if it remains
dry. Corrosion would not be expected to occur in a 10 is fl'cm resistivity
environment.
Examination of the analysis of subbituminous coal, however, suggests that
it could be aggressive to copper if the pulverized product became wet (Table
1). The wet material would be expected to corrode copper because it contains
appreciable quantities of carbon and sulfur. Further, the leaching of soluble
components from pulverized subbituminous coal would be expected to create
a low pH, low resistivity environment. Premature underground corrosion of
thermally insulated copper water tube at two locations investigated by the
authors can be associated with these aggressive conditions.
Domestic Hot Water Lines, Brooklyn, NY

Copper water tube was selected and installed in 1976 to convey domestic
hot water at a seven block, low-rise high density housing development in
Brooklyn, NY. A finely pulverized, natural carbonaceous granular material
was specified to provide thermal insulation for the underground lines. Within
three years, numerous leaks occurred in the hot water system. The leaks
occurred where the copper water tube was underground and thermally
insulated.
Examination of a tube that had been removed from service revealed signifi-
cant corrosion on the outside surface (Figs. 1 and 2). There was no evidence of
any significant corrosion of the inside surface by the water conveyed.
In general, the corroded outside tube surfaces were covered with a black
product that was overlain with a greenish material (Fig. 2). Energy dispersive
X-ray analysis (EDXA) and microchemical analysis revealed that the greenish
tinted copper corrosion product contained major quantities of copper and
sulfur (as sulfate) along with constituents normally found in soil (Fig. 3). The
green copper corrosion product consisted primarily of copper sulfate. Simi-
larly, analysis (Fig. 4) revealed that the black copper corrosion product con-
sisted primarily of copper sulfide.
Subsequent on-site examination revealed that wet thermal insulation
(nearly water saturated) existed at those sites where major corrosion had
taken place. Tube to soil potentials at these locations were generally more
positive than --0.15 V, referenced to a copper-copper sulfate electrode (that
is, the copper exhibited potentials at which active corrosion would be ex-
pected to be taking place) [7]. On-site testing also revealed that thermo-
galvanic and stray current corrosion were not involved in the corrosion pro-
TABLE 1-- Typical ultimate analysis for

subbituminous coal [5, 6].
Constituent Weight, %
carbon 64 to 76
hydrogen 4.4 to 5
nitrogen 1.2 to 1.5
sulfur 0.4 to 1
oxygen 12.8 to 17
FIG. 1--External surfaces of corroded underground copper water tube that had been ther-
mally insulated with a natural carbonaceous granular material. The performation in Specimen 5
(see encircled area) had initiated on the outside surface of the tube (-XO.4).
FIG. 2--External surfaces of additional corroded underground copper water tube that had
been thermally insulated with a natural carbonaceous granular material. Typically, the corrosion
products consisted of black copper sulfide overlain with green copper sulfate (~XO.4).
FIG. 3 - - E D X A and microchemical analyses revealed that the green copper corrosion product
on the outside surface of the tube consisted primarily of copper sulfate.
FIG. 4 - - E D X A and microchemical analyses revealed that the black copper corrosion product
on the outside surface of the tube consisted primarily of copper sulfide.
cess. Analysis of soil and insulation specimens collected during the on-site
inspection provided additional insight regarding this corrosion incident.
Typically, the resistivity of the wet thermal insulation ranged from 1 200 to
5 100 ~- cm. By contrast, wet soil resistivity adjacent to the insulation varied
from 9 500 to 12 000 fl.cm. These results established that the low resistivity
of the thermal insulation was not associated with the leaching of minerals
from the soil but could only be related to the inherent characteristics of the
wet natural carbonaceous granular material.
Belief that neither the soil nor groundwater chemistry was a factor in the
corrosion process was supported by p H measurements. The p H of the insula-
tion was as low as 4.8 whereas the p H of the adjacent soil varied from 7.7 to
9.0.
Conclusion
Based upon these results, it was concluded that the domestic hot water lines
corroded because they had been exposed to an aggressive wet thermal insula-
tion. The accelerated corrosion of the copper was caused by the presence of
carbon and sulfur in a wet insulation, which had a low resistivity and a rela-
tively low pH.
Hot Water Heating/Chilled Water Cooling System, Columbia, MO

A finely pulverized natural carbonaceous granular material was used to in-
sulate the underground copper water tubes for a hot water heating/chilled
water cooling system at a convalescent center constructed during 1974 in
Columbia, MO. Approximately four years later, leaks developed in the cop-
per tube that was exposed to a high water table. Examination of the copper
tube revealed corrosion only where the tube had been in contact with wet
insulation.
The corroded outside tube surface was covered with porous reddish-brown
cuprous oxide which was, in part, overlain with a green copper corrosion
product (Fig. 5). EDXA and microchemical analysis revealed that the green
material contained appreciable quantities of copper along with minor
amounts of sulfur and chloride, the overall corrosion products consisting pri-
marily of copper sulfate and copper chloride.
Subsequent testing revealed that the thermal insulation did not contain any
appreciable amounts of sulfur. It did however, have a relatively low p H value
of 4.5. Unfortunately, sufficient insulation material was unavailable for resis-
tivity measurements. The small amount of sulfur in this natural carbonaceous
granular material was not considered unusual since subbituminous coal
would be expected to contain varying amounts of this element.
FIG. S--External surfaces of corroded underground copper water tube that had been ther-
mally insulated with a natural carbonaceous granular material. Typically. the corrosion products
consisted of porous reddish-brown cuprous oxide overlain in part with green copper sulfate and
copper chloride ( - X 1 ) .
Conclusions
Based upon the general guidelines established for copper water tube in un-
derground building systems, it is apparent that finely pulverized natural car-
bonaceous granular insulation materials must be kept dry. If they become
wet, they become poor thermal barriers and create an environment that is
aggressive to copper and other commonly used construction materials. Cop-
per cannot exhibit its normal resistance to corrosion if subjected to wet, low
resistivity, low pH environments containing appreciable amounts of carbon,
and sulfur.
Prevention Through Cathodic Protection

Fortunately, when metallic materials are inadvertently exposed to a hostile,
underground, thermal insulation environment, they can be cathodically pro-
tected. Based in part upon the success of cathodic protection in mitigating the
corrosion of copper tube heating lines exposed to wet insulating concrete [8],
cathodic protection systems were designed and installed for the domestic hot
water lines at the Brooklyn, NY, housing development.
The cathodic protection systems consisted of a deep anode bed and rectifier
for each housing block at the development. 3 Deep anode beds were used be-
3Burke, N. D., private communicationto J. R. Myers, 1 June 1983.
cause of the unavailability of open land for conventional anode systems and
the desire to provide uniform current distribution to the underground piping.
The cathodic protection systems concurrently provide protection to the unin-
sulated copper tube cold water lines and the coated steel gas service pipes in
the area because all of the underground systems are electrically continuous.
Each of the anode beds contains eight vertically positioned, 76 mm (3 in.)
diameter, 1524 mm (60 in.) long graphite anodes evenly spaced between 17
and 61 m (55 and 205 ft) below the surface. The anodes are backfilled in each
254 mm (10 in.) diameter, vertically drilled hole with low resistivity calcined
petroleum coke breeze. Each anode bed is activated by an 80-V 34-A capacity
direct current rectifier. Individual anode leads to a junction box for each rec-
tifier; permanently installed reference electrodes and the necessary test sta-
tions were included in the cathodic protection design and installation to facili-
tate current control and system monitoring.
The results of field testing following initi~d energizing of the cathodic pro-
tection systems revealed that the thermally insulated copper water tubes
could be sufficiently polarized for effective corrosion mitigation.
Corrosive Attack by Moist Cellulosic Insulation

Largely because of its low cost and general availability, cellulosic-type insu-
lation has often been selected for reducing residential thermal losses. How-
ever, this type of insulation has a tendency to absorb moisture and, in the
presence of certain chemicals added for fire retardation, may become corro-
sive to metals.
Cellulosic insulation is manufactured by shredding waste paper, usually
newsprint, and then milling it into a finely divided fluffy fibrous material. To
reduce its rather high flammability potential, fire retardant chemicals are
added. The insulation is then bagged and shipped.
Quality control varies widely among the numerous companies that produce
this material. In some cases, it has been reported 4 that the chemical additives
have not been entirely absorbed, which has led to substantial residues from
the paper making process at the mill being found in the bottom of the ship-
ping container.
Fire retarding chemicals may or may not be corrosive. Borax, for example,
may act as a mild corrosion inhibitor for metals like copper. On the other
hand, aluminum sulfate is mildly corrosive and ammonium sulfate is ex-
tremely aggressive towards copper.
One detailed laboratory study by the authors [9], involving both failed cop-
per plumbing tube specimens and the cellulosic insulation with which it was
in contact, confirmed the presence of much ammonium sulfate with some
indication of boron. As in all such cases where the insulation has been wet or a
4Communicationfrom ASTM CommitteeC-16 on Thermal Insulation.
water extraction has been made, it is not possible to state with assurance
whether the minor constituents were in the original paper stock or were added
by the insulation manufacturer.
The copper and brass specimens in Fig. 6 showed patches of bluish mate-
rial identified as partly a m m o n i u m sulfate and partly unknown compounds.
Crystals of various intensities of blue color suggested the presence of Cu + +
ions in an otherwise colorless crystal (such as ammonium sulfate). Small re-
gions had the whitish cast sometimes seen when hydrates have lost some or all
of their water.
Some hard blue material adhering to the pipe seemed to consist mainly of
paper fibers, an unidentified binder, and a few tiny crystals. All of these ad-
herent materials contained copper and boron in excess of 10% and minor
amounts of such elements as silicon and aluminum.
Ash from the insulation remote from the pipe contained boron and alumi-
num in excess of 10% and minor amounts of calcium.
Table 2 [10] lists the compositions of fire retarding chemicals extracted
from 19 different cellulosic insulation samples. The wide variation is notewor-
thy as is the great p H range of the water extract with one sample having a p H
of only 3.7, which is distinctly acidic. As the actual formulations were not
revealed by the chemical tests, there is no apparent correlation with p H and
the chemical content.
FIG. 6--Outside surface of copper water tube displaying extensive attack caused by aggressive
cellulos& insulation ( - X1).
TABLE 2-- Composition and p H o/" ceUulosic insulation samples.
Fire Retardant Chemical, %
Sample Total Water Ammonium B o r i c Calcium Aluminum Sodium

Identification Solubles, % pH Sulfate Acid Sulfate Sulfate Carbonate
526-1 18 4.4 18
526-5 20 8.0 11 1 5
527-A 22 8.2 16 3
527-B 31 4.8 23
527-C 28 8.1 22
527-C1 24 8.2 20 5
527-D 22 8.0 13 s
527-E 26 4.5 2'6
-
527-F 21 5.9 ... 10 2
527-G 19 4.4 19
527-H 21 7.8 ... 16 3" 1
527-I 20 5.0 ... 4 1
535 24 7.4 ... 17 4
562 22 3.7 18 1 2
563-4 24 4.0 10 7 2
563 -5 19 7.7 '1"2 4 1 2
563-6 17 5.9 ... 4 6
563-7 23 6.1 ... S 8
593 17 7.7 ... 1'7
Revisions to Insulation Standards
M u c h u n f a v o r a b l e press notice [ 1 1 - 1 5 ] in addition to support f r o m the

U . S . C o n s u m e r P r o d u c t Safety C o m m i s s i o n [16] p r o m p t e d the F ed er al Gov-
e r n m e n t to issue the m u c h n e e d e d revision to F ed er al S t a n d a r d T h e r m a l In-
sulation (Loose Fill for P n e u m a t i c or P o u r e d A p p l i c a t i o n ) : Cellulosic or
W o o d F i b er (HH-I-515), which was issued in 1979.
C o n c u r r e n t A S T M C o m m i t t e e C-16 on T h e r m a l Insulation action involves
the revision of A S T M Specification for Cellulosic F i b er (Wood-Base) Loose-
Fill T h e r m a l I n s u l a t i o n (C 739) to c o n ta in a test m e t h o d using flat c o p p e r
specimens for corrosion e v a l u a t i o n u n d e r controlled conditions.
Conclusions
Cellulosic insulation u n d e r moist or h u m i d conditions m ay not be a suit-

able m a t e r i a l for use in c o n t a c t with p l u m b i n g or electrical systems because
the presence of c o n t a m i n a n t s a n d an a d d e d fire r e t a r d i n g c h e m i c a l m ay m a k e
it potentially aggressive to copper.
This is c o m p l i c a t e d by the fact t h a t moisture-free conditions c a n n o t be as-
sured in service, which, in t u r n , f u r t h e r activates t h e p o t e n t i a l corrosive
condition.
W h e n m o i s t i n s u l a t i o n s u p p o r t s f u n g a l g r o w t h , m o r e m o i s t u r e is t a k e n i n t o
t h e s y s t e m , a n d a c i d p r o d u c t s r e s u l t i n g f r o m b i o l o g i c a l p r o c e s s e s m a y also
cause corrosive attack.
References
[1] Myers, J. R. and Cohen, A., "Conditions Contributing to Underground Copper Corro-
sion," Journal td" the American Water Work Association, Aug. 1984, pp. 68-71.
[2] Denison, i. A., "Electrolytic Behavior of Ferrous and Non-Ferrous Metals in Soil Corrosion
Circuits," Transactions of the Electrochemical Society, Vol. 81, 1942, pp. 435-453.
[3[ Gilbert, P. T., "Corrosion of Copper. Lead and Lead Alloy Specimens After Burial in a
Number of Soils for Periods Up to 10 Years," Journal of the Institute of Metals, Vol. 73,
1947, pp. 139-174.
I4] Romanoff, M., "Underground Corrosion," National Bureau of Standards Circular 579,
Washington, DC, 1957.
[5] McGannon, H. E., Ed., The Makhlg. Shaphlg and Treating of Steel, Herbick and Iteld,
Pittsburgh, PA. 1971, p. 78.
[6] Haslarn, R. T. and Russell, R. P., Fuels and Their Combhlation, McGraw-Hill Book Com-
pany, New York, 1926, p. 52.
[7] Manual on Underground Corrosion ill Rural Electric Systems, Bulletin 161-23, Rural Elec-
trification Administration (U.S. Department of Agriculture), Oct. 1977.
[8l Rogers, P. C., Gross, E. E., and Husock, B., "Cathodic Protection of Underground Heat-
ing Lines," Materials Protection, Vol. 1, No. 7, July 1962, pp. 38-43.
[9] Corrosive Attack of Copper Water Tube by Cellulosic hlsulation, Copper Development As-
sociation Inc., Jan. 1978.
[10] Survey of Cellulosic hlsulation Materials, Energy Research and Development Administra-
tion. U.S. Department of Commerce, Table 1, Jan. 1977, p. 4.
[1I] "Push for Insulation Linked to Fire Peril." The New York Times, 27 Nov. 1977.
112] "Insulation Hazards Mounting," The Washhlgton Post, 1 Oct. 1977.
[13] Energy Conservation Digest, Vol. 1, No. 3, Washington, DC, 6 March 1978.
[14] Energy Conservation Digest, Vol. 1, No. 13, 24 July 1978.
[15] Chemical Week, 3 May 1978, p. 13.
[161 U.S. Consumer Product Safety Commission, Washington, DC, Public Calendar, Vol. 5,
No. 43, 28 July 1978.
Control Measures
Paul E. Krystow 1
Controlling Carbon Steel Corrosion

Under Insulation
REFERENCE: Krystow, P. W., "Controlling Carbon Steel Corrosion Under Insulation,"

Corrosion o/'Metals Under Thermal btsulation. A S T M STP 880, W. 1. Pollock and J. M.
Barnhart, Eds., American Society for Testing and Materials, Philadelphia, 1985, pp.
145-154.
ABSTRACT: Corrosion under hot and cold thermal insulation can be a serious problem
in chemical plants. Not only has corrosion under insulation caused staggering mainten-
ance costs in the millions of dollars but also has led to lost production time as well as
affected plant safety. Many of the chemical plants have experienced a variety of problems
and some of the specific experiences will be described in this symposium.
As a result, Exxon Chemical Company and many other chemical companies have em-
barked on a major effort to address the problem. In a companion paper, Exxon Chemi-
cals at Baton Rouge will describe identifiable factors that cause the problem. This paper
will review the important steps that Exxon Chemical has taken to control carbon steel
corrosion under insulation. Specifically, the review will include (1) organizing and sched-
uling more rigorous programs of inspection including typical examples, (2) preparation of
improved insulation specifications that address each of the factors affecting corrosion,
and (3) action programs required to assure that improved insulation specifications are
implemented.
An important aspect of controlling corrosion under insulation is through an appropri-
ate inspection program. Unfortunately under normal circumstances, inspection requires
the removal of the insulation during downtime, which is both costly and extremely diffi-
cult particularly on large towers and complex piping systems. There is an urgent need to
develop a nondestructive onstream examination (NDE) method to detect corrosion with-
out removal of insulation. In this connection, the Materials Technology Institute for the
Chemical Process Industries is investigating some of the NDE methods. A brief synopsis
will be made of the NDE methods employed. However, these methods have had only lim-
ited success, and a breakthrough inspection method is still needed. This need for break-
through represents an important challenge to the international scientific community.
KEY WORDS: corrosion, insulation, carbon steels, thermal insulation
Corrosion under hot and cold thermal insulation can be a serious problem
in chemical plants. Not only has corrosion under insulation caused staggering
1Engineering associate, Exxon Chemical Company, P.O. Box 271, FIorham Park, NJ
07932.
Copyright9 by ASTM International Www.astIn .Oi'g
maintenance costs in the millions of dollars, but it also has led to lost produc-
tion time as well as affected plant safety. Many of the chemical plants have
experienced a variety of problems and some of the specific experiences are
being described in the other papers in this publication.
As a result, Exxon Chemical Company and many other chemical compa-
nies have embarked on suitable programs to address the problem. In a com-
panion paper, Exxon Chemical Americas at Baton Rouge describes the iden-
tifiable factors that cause the problem. 2 This paper will review the important
steps that Exxon Chemical has taken to control carbon steel corrosion under
insulation from each of the identifiable factors.
Specifically, this presentation will review the following control measures as
shown in Fig. 1:
9 Prepare improved insulation specifications and guidelines for controlling
corrosion under insulation.
9 Organize and schedule more rigorous programs of inspection to assure
that the improved insulation specifications are implemented.
9 Visual inspection is the most reliable method, but some promising NDE
methods are emerging.
Before discussing the items shown in Fig. 1, let us quickly review the vari-
ous factors leading to corrosion under insulation as discussed in Lazar's pa-
per (Exxon Chemical Americas, Baton Route). 2 In addition, we shall give ex-
amples of how these corrosion factors are overcome through improved
insulation specifications and corrosion control guidelines. A brief synopsis of
the various figures as reproduced in the attachment is presented below.
Figure 2
Equipment operating with a process temperature ranging between --3.89
to 121.11/148.89~ (25 to 250/300~ is subject to corrosion under insula-
tion, and consideration should be given to providing suitable paint protection
underneath. If equipment operates at a temperature outside the range, then
painting is not necessary. Equipment that operates alternatively from a tem-
perature outside the range, but within the critical range for 20% of the time,
should be painted. However, suitable high-temperature paints are required
when elevated temperatures are encountered during alternate exposure
within the critical temperature range. Recommendations on coatings are cov-
ered in several of the other papers in this publication.
Equipment geometry and design is very critical and Figs. 3 and 4 illustrate
the variety of geometry configurations that can lead to corrosion problems.
Insulation type is important, not only from the cost and thermal insulation
capability, but also in respect to moisture retention and water leachable con-
2Lazar, P., in this publication, pp. 11-26.
KRYSTOW ON CARBON STEEL CORROSION 147
e PREPARE IMPROVED INSULATION SPECSAND GUIDELINES FOR

CONTROLLING CORROSION UNDERINSULATION.
e ORGANIZE AND SCHEDULE MORERIGOROUS PROGRAMSOF INSULATION

INSPECTION.
e VISUAL INSPECTION IS MOSTRELIABLE BUT SOMEPROMISING NDT

METHODS ARE EMERGING.
FIG. 1--Controlling carbon steel corrosion under insulation.
e PROCESS TEMPERATURE
- 2SOF TO 5000F
- NORMAL AND ALTERNATE OPERATIONS
e EQUIPMENT GEOPE~TRY AND ~SIGN
- ~ T EXPOSURE DURATION
- FREE DRAINAGEESSENTIAL
- AVOID ~ T SPOTS
I INSULATION TYPE
- TNERI~L PROPERTIES
- COST
- WATER LEACH~LE CONTENTS
- WATER RETENTION, DRYING
FIG. 2--Factors affecting corrosion under insulation.
tents. Insulation types, such as calcium silicate, which can absorb and retain
substantial amounts of water and leach chlorides, can lead to more problems
than other insulations such as foam glass, that is not normally subject to
moisture absorption and does not contain significant leachable material. For
critical insulation service, it is necessary to consider all the factors, including
those relating to corrosion.
Figures 3 and 4
Figures 3 and 4 illustrate the equipment geometry and the variety of attach-
ments that appear on a vessel, the points where water can enter and become
trapped and the areas where corrosion can occur underneath insulation. It is
recommended that similar guideline sketches be prepared for all types of
equipment, that is, piping, storage tanks, exchanger shells, drums, and so
forth, to facilitate proper inspection and serve for preparing appropriate me-
chanical design sketches for the attachments that will provide more adequate
weatherproofing, vapor proofing and painting.
.... TOP PLA'ITORM SUPPORT5
[..~"i'~ G LUGS
i ~ ~tAn,c~,tu~o~ ~.Ac~rrs
o"~,~^"~
~~
, ! STE'FEN~G OR SUPPORT R~G, ;'I~OF.~
r ~ A L L GUSSFrTF~ .
coNr~Ec'r;oNs
I SUPPORT RING, OFF WALL
VF-SS~. SU~OR'r ,L
~ClLF'P5
L I ON-VF.S6EL ~ III~CKETS, GUIDES
----4 F1REPROOFIIgG, VESSEL SUPPORT
FIG. 3--Inspection of vessels ambient to 121~ 1250~ for CUI.
"I-EAPPEDWATER
"~ WATERENTRY POINT

I t [ DRAINAGE
IMPEDIMENT TO
FIG. 4--Inspection of vessels ambient to 121 ~ (250~ for CU1.
Some undesirable geometry and design features that should not be used are
as follows:
(1) flat horizontal surfaces (vacuum rings),
(2) structural shapes that trap water ( " H " beams),
(3) shapes that are impossible to properly weatherproof (structural shapes,
gussets, and so forth),
(4) shapes that funnel or lead water into the insulation (angle iron brack-
ets), and
(5) inadequate spacing that causes interruption of the vapor or weather
barrier (nozzle extension, ladder brackets, and deck spacing).
A good discussion of equipment design is included in Lazar's paper. 2
Figure 5
This sketch details in Case A, a possible method for overcoming corrosion

from exposed carbon steel ledges as may exist with vacuum rings that can
30I ran
i~ " OL
O
,S
ra"~ $/5 SHE~r~='TAL
FIG. 5--1nsulation at stiffening~old insulation support rings.
allow water to wick upward into the insulation. Case B illustrates a method
that allows for drainage of water that may enter through breaks in the
weatherproofing/vapor barrier, rather than the water collecting at the vari-
ous impediments imposed by the attachments, support rings, and so forth.
Figure 6
As a secondary defense to preventing corrosion at service temperatures
from - 3 . 8 9 to 121.11/149.89~ (25 to 250/300~ it is recommended that
painting be applied to protect the steel under insulation. Not only should the
specification include type of paint, surface preparation, inspection require-
ments, and so forth but also should be satisfactory for all temperatures that
may be encountered, particularly during upset operation.
The design of weatherproofing/vapor proofing is considered to be the pri-
mary defense against corrosion and needs special attention around attach-
ments and protrusions, such as illustrated in Fig. 3. The use of proper seal-
ants around protrusions is especially important and needs to be checked and
reapplied in critical areas at least once every two years. Further specific de-
tails are included in Lazar's paper. 2
Climatic conditions are extremely important, and locations, such as the
Gulf Coast, where humidity and rainfall are high, require more attention
than in the less humid northern areas such as in Canada. Also, insulated
equipment in the vicinity of cooling towers or subject to splash from further
washing operations require special attention. Equipment operating at tem-
peratures below the dew point, which is subject to frequent sweating, is also of
special concern. Thus, the climatic condition, location, and operating condi-
I PAINT SELECTION
- SECONDARYDEFENSE
- EXTENDS SYSTEM LIFE
o WEATHER/VAPOR PROOFING
- PRIMARY DEFENSE
- PROTECTS AGAINST PHYSICAL, LIQUID AND VAPOR
- CLOSURESAROUNDPROTRUSIONS
- MASTIC LIFE LIMITING
I CLIMATE
e MAINTENANCEPRACTICES
- REPAIR INSULATIONDAMAGE
- INSPECTION PROGRAM
FIG. 6--Contributing factors.
tion will influence whether more or less stringent insulation specifications and
inspection programs are required.
Maintenance practices are extremely important. All equipment is eventu-
ally exposed to insulation damage, and a rigorous repair program is neces-
sary. The insulation repair program should not only detail the proper mainte-
nance procedure but include inspectors participation to verify that the
insulation has been properly installed. This is particularly important since
insulation work is often carried out by contract personnel who may not be
familiar with your insulation repair specifications. Often it is necessary to
complete maintenance insulation while onstream. This can be particularly
difficult with cold service insulation systems that are subject to icing. Proce-
dures have been developed using methanol spray wrapping with polyethylene
sheet and taping to prevent ice from reforming before insulation installation.
Figure 7
This figure illustrates a natural tendency that occurs with insulation main-
tenance. Not only is finishing the job important from an energy conservation
standpoint, but it would be even more important from the corrosion stand-
point if the same system as illustrated were upside down and rainwater and
atmospheric moisture could enter and collect into the piping system.
Figure 8
It is important to initiate an organized plant-wide program for controlling
corrosion under insulation. This is best accomplished through preparation of
KRYSTOW ON CARSON STEEL CORROSION 151
FIG. 7--The natural tendency that occurs with insulation maintenance.
0 DESIGN FEATURES
e COMPLEXGEOMETRYDETAILS
9 INSULATIONMATERIALS
9 PAINT SELECTION
e WEATHER/VAPORPROOFING
e PROJECTENGINEER GUIDELINES
e INSPECTIONPROCEDURESAND CHECKLISTS
e MAINTENANCEREPAIR TECHNIQUES(INCLUDING
ON-STREAMREPAIR PROCEDURES)
FIG. 8--List of what the guidelines cover.
a guideline that addresses all the concerns and pitfalls. Figure 8 lists the im-
portant items that should be included in the guidelines, and we suggest the
preparation of similar guidelines that conform with the particular needs of
your plant.
Figure 9
An important aspect of controlling corrosion under insulation is through
an appropriate inspection program. Unfortunately, under normal circum-
e VISUAL
9 ULTRASONICSANDMICROWAVE
9 RADIObRAPHY
e TANGENTIALRADIOGRAPHY
9 X-RAYWITH IMAGEINTENSIFICATIONAND
CLOSED-CIRCUIT TELEVISION
9 INFRARED
e CORDUCTIVIIYMEASUREMENTS
I POTENTIALME/LSUREMENTS
9 ELECIROMAbNEIICREISODSUSINGHALL
SENSORS
FIG. 9--Current inspection methods employed.
stances, inspection requires the removal of the insulation during downtime,

which is both costly and extremely difficult, particularly on large towers and
complex piping systems. There is an urgent need to develop a reliable nonde-
structive onstream examination (NDE) method to detect corrosion without
removal of insulation. In this connection, the Materials Technology Institute
for the Chemical Process Industries is investigating some of the N D E meth-
ods. There have been some promising methods under investigation which are
discussed in more detail in other papers in this publication. Figure 9 lists
some of the current inspection methods employed.
Figure 10
At the moment, the most reliable, positive inspection method is by direct
visual examination that requires physical removal of the insulation. It is diffi-
cult to pinpoint areas to be checked, but, most assuredly, areas where wet
e INSPECTION CONDITIONS OF WEATHER PROOFING EVERY TWO YEARS

PARTICULARLY AFTER TURNAROUNDS
e CHECK METAL JACKETNG EVERY TWO YEARSPARTICULARLY AFTER

TURNAROUNDS
e REMOVE INSULATION IN SELECT AREAS OF THE CRITICAL WET

ZONES EVERY FIVE YEARSAND INSPECT
e REPAIR AND REINSULATE AFTER INSPECTION SHALL BE WITNESSED

BY INSPECTOR
FIG. lO--Suggested guidelines for inspection of corrosion under insulation.
e ESTABLISH MORERIGOROUSPROGRAMOF INSPECTION
e PREPARE IMPROVED INSULATIONSPECIFICATIONSWHICH

MINIMIZE THE FACTORSCAUSINGCORROSION
e INITIATE PROGRAMTO ASSURE ITEMS ABOVEARE CARRIED

OUT EVEN IF COSTLY
e PROMOTENONDESTRUCTIVEMETHODSFOR DETECTING
CORROSION UNDER INSULATION - BREAKTHROUGHNEEDED
FIG. l 1--Program for overcoming corrosion under insulation.
insulation exists is where the corrosion will take place. These areas tend to be
at lower locations of equipment and piping where moisture can accumulate
and collect at obstructions. Also, equipment is subject to sweating caused by
service conditions or located in a vicinity where cooling water drift or water
washing splash can be encountered.
The cutting of small insulation windows for spot checking corrosion under
insulation is not always adequate. It is necessary to remove large areas of in-
sulation in critical wet regions and realize that there is a level of risk where
one has not looked. However, it is emphasized that dry areas of insulation do
not experience corrosion, and visual inspection need not be carried out in
these areas. Figure 10 gives a suggested guideline for inspection.
Figure 11
This is the concluding slide. It summarizes the program for overcoming
corrosion under insulation as presented in this paper.
DISCUSSION
A. S. Krisher I (written discussion)--This conference has been a very useful

and stimulating review of a widespread and serious problem. It is apparent
that a control program consisting of coating before insulation, proper design
of the system, proper installation of the system, thorough periodic inspection,
1Monsanto Co., 800 N. Lindbergh-CS6F, St. Louis, MO 63167.
and proper maintenance will substantially reduce the occurrence of corrosion

under insulation.
It is worth noting that most of the steps listed above are identical to those
required to assure optimum energy conservation. Consideration of this sub-
ject may help to justify improved practices in cases where corrosion control
alone is judged to be insufficient justification for the expense of such
practices.
A critical subject that was largely overlooked in this conference is the ques-
tion of economics. In my opinion, we do not have a reasonable basis for esti-
mating the probable costs if we do nothing to prevent the problem. There are
many cases where corrosion under insulation has necessitated very extensive
and expensive repairs. However, in a number of these cases, repairs were not
required until long after the design life of the plant had been exceeded. There
are also a large number of cases in which no preventative steps were taken and
no problems were encountered.
I believe that we need to derive a reasonable basis for estimating the proba-
bility, cost and timing of corrosion under insulation. This could then be used
to evaluate the justification of various corrective steps. I have a subjective
feeling that we may sometimes spend more time on preventative measures
than is justified.
PaulE. Kryston (author's closure)--The comments offered by Bert Krisher
are appropriate, but the response is difficult to explain adequately. It is
agreed that the economics for CUI has been somewhat overlooked in this con-
ference, but not necessarily neglected by the chemical companies concerned
with CUI. Unfortunately, the costs for unexpected downtimes and the money
spent for conducting appropriate inspection and maintenance needed to
make an economic justification are usually considered to be proprietary and
therefore cannot be discussed.
We are in total agreement with the statement that it is necessary to estab-
lish a reasonable basis for estimating the probability, cost, and timing of cor-
rosion under insulation. Most of the equipment in a chemical plant is not
subject to corrosion under insulation. This is why special emphasis was made
(see Fig. 2 and in Exxon Chemical's companion paper by Lazar) of the factors
leading to CUI.
It is important to review the factors affecting CUI to establish the probabil-
ity and the timing of CUI, and together with an evaluation of the costs of
unexpected downtimes and the costs to carry out maintenance/inspection ef-
forts, determine the economic justification of the corrective program. In
many petrochemical processes unexpected shutdowns can lead to substantial
production losses, and this is a key concern along with safety in establishing
the corrective action program. On the other hand, there are also many chemi-
cal processes where a leak caused by CUI does not lead to a serious safety
concern and a shutdown does not involve significant production losses so that
corrective action beforehand is not a vital criteria.
Peter A . Collins, 1 John F. D e l a h u n t , 1 a n d D e b b i e C. M a a t s c h 1
Protective Coating System Design

for Insulated or Fireproofed
Structures
REFERENCE: Collins, P. A., Delahunt, J. F., and Maatsch, D. C., "Protective Coating
System Design for Insulated or Fireproofed Structures," Corrosion of Metals Under
Thermallnsulation, STP880, W. I. Pollock and J. M. Barnhart, Eds., American Society
for Testing and Materials, Philadelphia, 1985, pp. 155-164,
ABSTRACT: Corrosion of carbon steel equipment beneath thermal insulation and con-
crete is a critical concern anmng refineries, chemical plants, marketing organizations,
and pipelines. Corrosion has been detected worldwide under all types of thermal insulat-
ing materials. At Exxon this concern has been addressed in a number of laboratory inves-
tigations and also during field troubleshooting assignments. This paper describes work
undertaken to develop protective coating systems suitable for equipment to be insulated
or fireproofed. These systems are described. In addition, the advantages of organic coat-
ings compared to inorganic zinc-rich coatings are reviewed. Also discussed are recently
completed novel evaluations to select coating systems for thermally insulated and ther-
mally cycled equipment.
KEY WORDS: corrosion, insulation, tests, paints, coatings, steels
Corrosion of c a r b o n steel e q u i p m e n t b e n e a t h t h e r m a l insulation a n d con-

crete is a critical concern a m o n g refineries, chemical plants, m a r k e t i n g orga-
nizations, a n d pipelines. Corrosion has been detected worldwide u n d e r all
types of t h e r m a l i n s u l a t i n g materials. At Exxon this concern has been ad-
dressed in various laboratory investigations. This paper describes work u n -
d e r t a k e n to evaluate protective coating systems suitable for e q u i p m e n t to be
insulated or fireproofed. Discussed are recently completed novel laboratory
evaluation t e c h n i q u e s to select coating systems for insulated a n d thermally
cycled e q u i p m e n t . I n addition, the advantages of organic coatings c o m p a r e d
to inorganic zinc rich coatings ( I O Z R ) are reviewed.
~Staffengineer, engineering associate, and project engineer, respectively,Exxon Research and

Engineering Co., P.O. Box 101, Florham Park, NJ 07932.
Corrosion Under Insulation Widespread

Corrosion under insulation and fireproofing has occurred in all major areas
of operation and beneath all major types of insulation in use. These problems
began to come to the fore during the late 1960s and became of major concern
in subsequent years. A major contributing factor for this corrosion problem
was changes made to materials specified for insulation systems previously:
9 I n s u l a t i o n s - - N e w cellular plastic foams (polyurethane and phenolic)
came into use. When water saturated, these are acidic and extremely corro-
sive.
9 M e t a l J a c k e t s - - R e d u c e d anywhere from 50 to 100% in thickness. Com-
position changed from copper-bearing carbon steel to carbon steel. Zinc coat-
ing thickness reduced. Such changes would lead to wind damage and early
corrosion failure.
9 Coating J a c k e t s - - M a n y systems changed, for example, hot tar three-ply
roof systems to, in many cases, 0.5 to 1.0 mm (20 to 40 mil) of synthetic rub-
ber for storage tank roofs. Long-term weather resistance would be reduced.
9 E q u i p m e n t P a i n t i n g - - A t specific locations, equipment painting reduced
the need to paint equipment operating between 0 to 93~ (30 to 200~ to
between 0 to 55~ (30 to 130~ before insulation application. This change
increased the opportunity for corrosion to occur.
9 F i r e p r o o f i n g - - E l i m i n a t e d external painting of concrete and also steel-
concrete seam protection. Therefore, water could penetrate the concrete
more readily and increased the corrosion potential.
These factors, along with external influences, such as increased use of recir-
culating salt water cooling towers, have led to increased corrosion under insu-
lation and fireproofing.
Exxon over recent years has conducted programs to mitigated such corro-
sion. These included several evaluations of paints and protective coatings that
would be applied to carbon steel equipment or structures before insulating or
fireproofing. This paper describes the following investigations:
9 Evaluation of paints and coatings under wet insulation or fireproofing.
9 Determination of the bond strength of concrete applied to coated and
uncoated steel.
9 High-temperature cyclic tests of paints and coatings under wet insula-
tion.
These programs are discussed in subsequent paragraphs.
Coating System Design for Insulated Equipment

Paints and coatings applied to insulated equipment may be designed to
withstand immersion in hot water (93~ [200~ on a continual basis. How-
COLLINS ET AL ON PROTECTIVE COATING DESIGN 157
ever, such a design would be inordinately expensive since minimum surface

preparation for such a service would be Steel Structures Painting Council-5
(SSPC-5), white abrasive blast cleaning, followed by the application of an ex-
pensive, high build sophisticated coating system. A more practical approach,
based upon our experience, is the use of SSPC-6, commercial abrasive blast
cleaning, and the application of a more easily applied, general maintenance
paint or coating. Based upon such a philosophy, several laboratory screening
evaluations were undertaken at steady state and cycling high-temperature
conditions.
Evaluations of Coatings Under Insulation/Concrete

The evaluation of paints and coatings for use under thermal insulation and
concrete consisted of two parts:
9 Comparison of bare and coated steel corrosion coupons after 30-days ex-
posure to wet insulation and concrete at 66~ (150~
9 Determination of the effect of coatings on the bond strength of steel to
concrete.
For the exposure tests, carbon steel coupons 100 by 100 by 3 mm (4 by 4 by
1/8 in.) were prepared by sandblasting and coating. The coatings evaluated
were an inorganic zinc-rich coating (IZRC), a red lead alkyd, lead chromate
phenolic, and an amine-cure epoxy. Each of the coated samples was scribed
to introduce a defect in the coating. To simulate service exposure, the pre-
pared samples were each encased in a "sandwich" of insulation (fire retard-
ant polyurethane foam and calcium silicate) or concrete (Type 1 portland ce-
ment) as shown schematically in Fig. 1. The coated test panels were exposed
~i CARLSONSTEEL
CORROSIONCOUPON
TOP VIEW
DISTILLED'
E'~D- - ' ~
WATERI h F ~ F ~'~r OLA.9~ FUNNEL
| r I~ II ~ NSIJLATON FI~IEPf'~UOF~N~
~- ~ CA~IBONSTEEL
CORROSIONCOUPON
SlD~ VIeW
FIG. 1--Specimen used to measure the corrosivity of carbon steel beneath wetted thermal
insulation and concrete fireproofing.
to an alternating wet and dry cycle using distilled water for 30 days. Tempera-
tures were maintained at 66~ (150~ Uncoated carbon steel test coupons
without the covering of insulation or fireproofing were also tested for compar-
ison.
All Coatings Except IZRC Perform Well

After the 30-day exposure, the coupons were removed and visually exam-
ined for rusting, discoloration, chalking, pitting, and other evidence of coat-
ing degradation. The steel coupons exposed without coatings exhibited vary-
ing degrees of corrosion--from light rusting under cement to heavy rust
buildup when tested under polyurethane foam. The corresponding corrosion
rates calculated from weight loss measurements ranged from 0.02S to 0.050
mm/year (1 to 2 mpy) for the sample tested under concrete to 0.175 to 0.200
ram/year (7 to 8 mpy) under the fire retarded polyurethane foam. Corrosion
pitting was also evident on these samples after the corrosion scale was re-
moved. The corrosion rate of the control sample was 0.050 to 0.075 m m / y e a r
(2 to 3 mpy).
The coated samples were generally in good condition following the test.
There was no evidence of film breakdown except for the ]ZRC as discussed
below. There was essentially no rusting, except for the samples tested under
polyurethane foam, which exhibited rapid corrosion of the steel exposed by
the scribe marks. Weight measurements of the painted samples before and
after the test indicated negligible weight change.
Because of a slight pitting on the surface of the IZRC coated coupons, cross
sections of these samples were examined metallographically. It was evident
that the coatings were undergoing attack and that corrosion products were
building up at the steel/coating interface. A comparison of an unexposed
IZRC sample to the test coupons is shown in Fig. 2.
The unexposed coating is tightly adherent to the steel substrate and con-
tains only a few small pores typical of this material. The exposed coatings
showed areas of disbonding, severe porosity, and underlying corrosion. The
corrosion products were qualitatively analyzed and found to contain silicon
and iron oxides. While attack of the IZRC was evident after testing under all
insulations and fireproofing, this degradation was most pronounced under
the polyurethane foam insulation.
Calculated corrosion rates for the inorganic zinc coating based on weight
loss, ranged from 0.050 to 0.090 mm/year (2 to 3.5 mpy) beneath polyure-
thane foam and calcium silicate, to 0.025 mm/year ( - 1 mpy) under cement.
However, because these were calculated based upon the density of zinc, they
are conservative because of the voids present in an IZRC film. Complete pene-
tration of the IZRC film would occur within approximately 1 to 3 years (based
upon the density of zinc).
FIG. 2--(a) Cross section of a panel coated with IZRC before testing. The coating is continu-
ous, tightly adherent to the steel substrate, and contained a few isolated pores. (b)Cross section
of a panel coated with IZR C after testing under polyurethane foam. Degradation of the coating
has occurred with a buildup of corrosion products at the coating/substrate interface. (X150)
Effect of Coatings on Bond Strength Examined

In the case of concrete applied to structural steel, an additional consider-
ation may be the effect of an applied coating on the adhesion or bond of the
steel to concrete. Bonding is important for the mechanical integrity of rein-
forced concrete structures. Also, well bonded concrete will prevent the in-
gress of water and other contaminants that will lead to the corrosion of the
underlying steel. Various standards now specify that concrete must be ap-
plied directly to sandblasted steel. The reason for this is that concrete will not
adhere as well to painted steel as to uncoated steel. Therefore, limited tests

were undertaken to quantify the degree of bonding of steel with and without
coatings.
To investigate bond strengths, specimens were prepared in accordance
with ASTM Test Method for Comparing Concretes on the Basis of the Bond
Developed with Reinforcing Steel (C 234). The specimens consisted of a car-
bon steel rod 1.9 by 45.7 cm (0.75 in. diameter by 18 in. long) encased in a
15.2-cm (6-in.) cube of concrete. The surface preparation and coatings ap-
plied to the steel rods included: sandblasted, lightly rusted, rusted in the at-
mosphere, coal tar epoxy, amine-cured epoxy, inorganic zinc, red lead
primer, and lead chromate primer. The coatings were applied to sandblasted
rods. The concrete was mixed to ASTM Making and Curing Concrete Test
Specimens in the Laboratory (C 192) using a cement/aggregate~water ratio of
2:3:1 by weight. The samples were cured for 30 days. The specimens were
tested by measuring the load required to break the bond of the steel rod to
concrete and pull it from the cement cube using a standard tensile machine as
shown in Fig. 3. The results of these pullout tests are shown in Fig. 4.
The best bond, as expected, resulted from the sandblasted bar, requiring
19.5 N (4400 lb) to break the concrete/steel bond. This bond strength was
F I G . 3 - - A p p a r a t u s used to measure the bond strength o f coated and uncoated steel bars to
concrete. The concrete cube was held f i x e d by the upper arm of the tensile machine, while the
lower arm pulled the bar downward. Loads were monitored to the nearest 89 N 120 Ib).
m~
~ m
Z
llio~ !
z
10,0
E
RNNNn
0
5.w0 |
Z o
o
8 z_ z
, |
FIG. 4--Pull out test results.
slightly reduced by a rust blooming of the sandblasted steel and markedly

reduced by atmospheric rusting or when concrete is applied to coated steel.
The coatings that produced the highest bond strengths were the two epoxies
and the inorganic zinc. It is somewhat surprising that the highest bond
strengths resulted from the epoxy coatings that have glossy finishes, whereas
the flat primers gave lower bond strengths. However, the results with epoxy
based coatings are supported by similar tests by other investigators [I].
Although the bond strengths of coated steel rods to concrete were less than
half that of sandblasted steel rods to concrete, the tests show that a bond does
exist between coated steel and concrete. While this work did not identify a
minimum bond strength necessary for acceptable field performance, corro-
sion of the underlying steel should not occur as long as the coating remains
intact and there is no loss of adhesion between the steel, coating, and cement.
High-Temperature Cyclic Evaluation Undertaken

High-temperature cyclic testing of paints and coatings was undertaken be-
cause of the severity of corrosion occurring on insulated equipment in this
type of intermittent service in operating plants. Testing conditions sought to
duplicate "worst case" operating conditions. Paints were exposed to tempera-
tures ranging from ambient to 315~ This severe temperature range was
used to accelerate coating failure.
The high-temperature test equipment is illustrated in Fig. 5. The unit was a

modification of the apparatus used previously to evaluate heat resistant
paints and also, temperature indicating paints. Insulated carbon steel pipes
were suspended several inches over the heat source (propane burner).
Chrome-alumel thermocouples were directly attached to the insulated pipe,
and in conjunction with a recorder, supplied a continuous temperature pro-
file of the pipe. A representative temperature profile is shown in Fig. 6.
The test procedure was designed to simulate cyclic service beneath wet in-
sulation. During each test period, the pipe was heated for 8 h, then left at
ambient temperature for the remainder of the day. Water was added daily to
the insulation-pipe interface. Approximately 1 L (1 qt) was applied at the top
of the pipe and allowed to run down the full length of the pipe. A basin welded
onto the pipe bottom prevented any excess water from running into the heat
source and created a reservoir to keep the insulation moist. The water entirely
evaporated by the end of each 8-h heating period. Each pipe was cycled over a
30-day period.
Insulation selection was also based on increasing the severity of the test.
Calcium silicate was used to insulate the lower two-thirds of the pipe, which
was the portion exposed to the higher temperatures. Calcium silicate was cho-
WMet ~
2YI" ~ 50mm
2"
r OJ F
F~m ~
C ~ l x m ,Tmml. I- 89 NPS
w i l l 9 24" k a ~
10t:m
C41r
TEST PIPE 68turn

Cattmn St~l. 2 89 NPS
S~ulndl~dWiS~ sI
Ii I,
Ii Ii i 7591~
i i
ii i
'I i "'
ii
I ].____ ,. _ !
ii ii i
215mm
B. I m u l 0 t K I Pipe
In G r o s - , ~ c i o n
t_, ~i i!il I
A. Sr,l ~ r r u m ~ O ! Tarot EQuionmm
FIG. 5--High-temperature paint testing equipment.
550 -- 288
525 274
500
I P ~ O Z60
475 -
\-
450 --
\
23Z
4Zb I 1 I 1 I i 1 t 218
3 6 9 12 15 18 21 24
I)ISIANt,L FIiOM Ii()II()M OF P I P k (IN)
FIG. 6 - - Temperature profile of test pipes.
sen for two reasons: it is frequently used by refineries, and it is easy to handle.
In addition, water leached through calcium silicate is known to be alkaline
and thereby increases the severity of the test. Fire retardant polyurethane
foam was used to insulate the upper one-third of the test pipe, even though
the pipe surface temperature exceeded the suggested operating limit for poly-
urethane foam (83~ [180~ However, polyurethane foam was included
because earlier work showed that water leached from this type of insulation,
when treated with fire retardants, was more corrosive because of an acidic pH
than other frequently used types of insulation. Therefore, it was felt that this
would provide a severe test of the protective properties of the paints and coat-
ings.
Six Coatings Evaluated

Six coatings were evaluated by the described test method. These included:
phenolic lead chromate primer, catalyzed epoxy amine primer, zinc-pig-
mented linseed oil-based paint, zinc-pigmented silicone acrylic primer, and
two epoxy phenolic coatings. These coatings were selected for several reasons.
The phenolic and amine epoxy are the most commonly used organic coatings
for protection beneath insulation. Although there is little knowledge concern-
ing their performance in high-temperature cyclic service and the test temper-
ature exceeded their suggested operating limit, they were included to provide
a basis for evaluating other coatings caused by extensive experience with their
performance. The silicone acrylic is heat-resistant paint, and it is currently
specified for use beneath insulation by some refineries. The oleoresinous coat-
ing was chosen for its temperature resistance and ability to be applied over a
lesser degree of surface preparation, unlike most of coatings tested. The two
TABLE 1- - High temperature cyclic testing.
Paint System Calcium Silicate Polyurethane Foam
Epoxy phenolic destroyed accelerated corrosion

Epoxy phenolic chalky, nonadherent severe attack
Phenolic chromate cracked, flaked, 30% destroyed destroyed
Amine epoxy extensive flaking destroyed
Oleoresinous intact thin rust scale
Acrylic silicone intact limited protection
phenolic-epoxy coatings were selected for their outstanding chemical resis-

tance and water resistance.
Results Confirm Severity of Environment

Results of the high-temperature cyclic testing confirmed the severity of the
environmental conditions represented by this test. A summary of coating per-
formance is shown in Table 1. Of the coatings tested, only the oleoresinous
and acrylic silicone can be specified for use to protect equipment that may be
thermally cycled. The other types of paints and coatings failed to some extent
beneath calcium silicate insulation and almost catastrophically failed be-
neath fire retarded polyurethane foam.
Conclusions
From these evaluations, we have reached the following conclusions:
9 Many organic maintenance primers should give satisfactory service be-
neath insulation and concrete. Lead chromate phenolics and amine epoxies
have been used satisfactorily in the field.
9 IZRC primer should not be used under insulation or concrete. Field expe-
rience also supports that conclusion.
9 For high-temperature cyclic service under insulation oleoresinous or sili-
cone acrylic coatings are recommended based upon accelerated testing.
Reference
[1] Mathey, R. G. and Cligton, J. R., "Bond of Coated Reinforcing Bars in Concrete", Journal
of the Structural Division, Proceedings of the American Society of Civil Engineers, Jan.
1976.
Louis C. Sumbry 1 and E. Jean Vegdahl 2
Prevention of Chloride Stress

Corrosion Cracking Under Insulation
REFERENCE: Sumbry, L. C. and Vegdahl, E. J., "Prevention of Chloride Stress Corro-

sion Cracking Under Insulation," Corrosion o[ Metals Under Thermal Insulation. A S T M
STP 880, W. I. Pollock and J. M. Barnhart, Eds., American Society for Testing and
Materials, Philadelphia, 1985, pp. 165-177.
ABSTRACT: Chloride stress corrosion cracking of stainless steel under insulation can
occur because of concentration of chlorides at the metal surface. This paper discusses our
company's current recommended practices for controlling this type of corrosion and
results of a research program to improve our recommended practices.
Our current recommended practice for controlling corrosion of stainless steel under
insulation is to select insulation materials low in extractable chlorides, jacket or coat insu-
lation to exclude water and chlorides from the environment, and to coat the stainless steel
equipment where applicable. Coating stainless steel to prevent contact with chlorides is
recommended with some reservation because of generally poor experience in coating
stainless steel equipment.
A research program is being conducted on coatings for stainless steel equipment. A
variety of coatings have been evaluated using adhesion and flexibility tests, outdoor expo-
sure, salt fog cabinet, environmental cabinet containing hydrogen sulfide, carbon diox-
ide, nitrogen dioxide, and wet air, a Weatherometer cabinet, elevated temperature expo-
sure (260 to 500~ thermal shock tests, and chloride permeability tests. Adhesion to
stainless steel is generally fair. For elevated temperature applications, straight silicones
are generally superior. Many coatings are adequate for ambient temperature applications
if protected from mechanical damage.
KEY WORDS: chlorides, stress corrosion, insulation, jackets, coatings, stainless steels,
mechanical bonds, chemical bonds
Chloride stress corrosion cracking of insulated stainless steel equipment is

a particular problem because chlorides tend to concentrate under the insula-
~Senior chemist, Gulf Research and Development Co., P.O. Drawer 2038, Pittsburgh, PA
15230.
2Consultant, 14120 SW Rochester Dr., Beaverton, OR 97005; formerly, engineer, Gulf Re-
search and Development Co., Pittsburgh, PA 15230.
Copyright9 by ASTM International WW'vV.astIn.OF~
tion at the surface of the metal if the insulation becomes wet. The moisture
can leach soluble chlorides out of the insulation, or may already contain chlo-
rides from the environment. At the warm metal surface the moisture is vapor-
ized, leaving behind an increasing concentration of chlorides. Then during
operation or startup/shutdown when the piping is in the susceptible tempera-
ture range, chloride stress corrosion cracking can quickly take place. Our
experience indicates that the problem is primarily within the 50 to 200~ (120
to 390~ temperature range. In this temperature range, waterborne chlo-
rides concentrate by evaporation of the water.
The exact mechanism of chloride stress corrosion cracking is still a matter
of discussion, however, the conditions required for it to take place are clear
and fairly well known. The four required factors are an austenitic stainless
steel, chloride ions, tensile stress on the metal, and certain temperature lim-
its. The cracks are generally highly branched, transgranular, and occur with
no deformation, and hence no warning.
This paper will discuss our current recommended practice for controlling
corrosion of stainless steel under insulation and results of a research program
to improve our recommended procedures for coating under insulation. As
this paper is concerned with chloride stress corrosion cracking, we will only
deal with equipment operating at 50~ (120~ and above.
Current Insulating Practice

The following is a discussion of the insulation practices that have been in
use at Gulf over the last 10 to 15 years. These include the design standards,
insulation and accessory materials, and installation practices. The final sec-
tion of this paper discusses work that is being performed to update these
practices.
Design Basis
Unless a heat loss is desired, vessels, heat exchanger units, and piping op-
erating at temperatures above 93~ (200~ are insulated. Insulation at lower
temperatures is required when it is necessary to limit heat losses or for person-
nel protection.
Codes--The basic design, assembly, and installation must conform to fed-
eral, state, and local codes or regulations. If a conflict exists among the codes
and regulations, the most stringent requirement will govern.
Insulation Thickness--The economic thickness, based on Gulf's standard
drawings for rigid type insulation, is defined as the thickness of a given insu-
lation that will save the greatest amount of energy while paying for itself
within an assigned period of time.
SUMBRY AND VEGDAHL ON CHLORIDE STRESS CORROSION 167
Materials
Insulation--The high-temperature insulation materials listed in Table 1
are used for vessels, heat exchangers, pumps, and piping, with the exceptions
that are listed below:
9 No asbestos bearing insulation materials (except mastic or asphalt base
materials) are used in new installations or as replacement material in existing
plants.
9 The insulating materials used on austenitic stainless steel surfaces are
low in extractable chlorides ( < 50 ppm). For additional protection against
chloride attack, austenitic stainless steel surfaces operating below 204~
(400~ are painted with a zinc free silicone or an epoxy-phenolic blend be-
fore insulating.
9 Below 121~ (250~ blanket insulation is avoided so as to eliminate
possible sagging and moisture pickup. Consideration should be given to the
use of cellular glass.
9 Installation of jacketing in order to provide watertight joints is empha-
sized. Stainless steel jacketing over insulation on stainless piping or equip-
ment is recommended when there is any possibility of the piping or equipment
reaching a metal temperature above 650~ (1200~ Aluminum jacketing
can be used; however, extreme care must be exercised to avoid contact be-
tween the aluminum and stainless steel. For example, a scrap from the trim-
TABLE 1--Recommended insulating material.
Material Temperature Range, ~ Application
Mineral fibers atmospheric temperature to 1038 piping

93 to 1038 vessels
Mineral fibers and atmospheric temperature to 816 piping
binder 93 to 816 vessels
Expanded perlite atmospheric temperature to 816 piping
93 to 816 vessels
Hydrous calcium to 816 piping
silicate to 816 vessels
Diatomaceous silica to 1100 piping
to 1100 vessels
Mineral wool and atmospheric temperature to 1038 pipmg
glass fiber atmospheric temperature to 1038 vessels
Nonrigid glass fiber atmospheric temperature to 177 piping
atmospheric temperature to 177 vessels
Rigid glass fiber atmospheric temperature to 232 pipmg
Cellular glass --268 to 250 piping
--268 to 250 vessels
Ceramic fiber to 1260 cements, castable
blocks and blanket
Insulation finishing atmospheric temperature to 1 0 3 8 miscellaneous finish
Insulating cement atmospheric temperature to 982 built-up insulation
ming of the jacketing coming in contact with stainless steel could melt and
cause liquid metal corrosion resulting in rapid premature failure of the stain-
less equipment.
Accessories--The accessories are as follows:
9 Materials used for welding studs and support clips are carbon steel for
carbon and alloy steel, and American Iron and Steel Institute (AISt) Type 304
(Unified Numbering System [UNS] $30400) stainless steel for austenitie
stainless steel vessels. All support rings are carbon steel if not welded directly
to alloy vessels. Welding and heat treatment of attachments meet the fabrica-
tion specifications for the vessel.
9 Tie wires, bands, band seals, wire mesh, and self-tapping screws for at-
taching insulation and aluminum sheets to equipment are 18 Cr, 8 Ni stain-
less steel. Bands and band seals may be aluminum for thin pipe jacketing.
9 Where insulation can not be used because of the irregular surface (such
as on valves, fittings, and flanges), insulating cement is used. Insulating ce-
ment exposed to the weather is waterproofed with a coat of asphalt emulsion.
Installation Practices
Surface preparation of nonaustenitic piping and equipment having operat-
ing temperatures of 10 to 121 ~ (50 to 250~ involves cleaning (Steel Struc-
ture Painting Council Grade SP-5) and coating with a rust inhibiting type
primer before applying insulation.
The following procedures are strictly observed:
1. Rigid type insulation is applied with staggered joints.
2. Double layer construction is used for equipment and piping in all ser-
vices above 260~ (500~
3. In multiple layer, rigid type insulation, both longitudinal and end joints
of the outer layer are staggered with respect to joints of inner layers. Each
layer is separately secured. Support rings for insulation are provided on
straight vertical piping runs. The maximum spacing of rings is 6 m (20 ft).
Insulated flanges are not used for vertical insulation support.
4. Insulation materials must be kept dry until protective coating or jacket-
ing is installed.
Research on Coatings for Stainless Steel

In following the current recommended practice for insulating austenitic
stainless steel, our company has experienced difficulties in successfully coat-
ing stainless steel equipment before insulating. These difficulties primarily
result from the poor adhesion of most coatings to stainless steel. Stainless
steels form a very dense smooth oxide film at their surface that greatly reduces
the amount of mechanical bonding between the steel and a coating. In addi-
tion, the oxide at the surface provides a different chemistry than a carbon
steel for chemical bonding with the coating.
In 1981, we began a research program to evaluate the performance of coat-
ings on austenitic stainless steels. It was hoped that the results of this research
would allow us to improve our recommended practices for insulating stainless
steel equipment.
A review of the literature indicated that silicones and epoxies are being
used to coat stainless steel [1-5]. Based on this information and on manufac-
turer's recommendations, seven coatings were selected for the initial test pro-
gram, including five silicones, an epoxy, and a vinyl. Coatings tested in the
second phase were selected primarily on the basis of the previous results, with
the addition of one new coating. Tables 2 and 3 give additional information
on all the coatings that have been tested.
Test Program
The tested coatings (described in Tables 4 and 5) were applied to 7.6 by
22.9-cm (3 by 9-in.) panels of Type 304 stainless steel for all tests except the
outdoor exposure. In the first phase, our in-house personnel applied the coat-
ings over solvent cleaned panels. For the second phase, the manufacturers
supplied us with coatings over solvent cleaned (only) and sandblasted (ap-
proximately l/2-mil surface profile) surface preparation.
The test program was designed to evaluate
(1) adhesion to stainless steel,
(2) flexibility, and
TABLE 2--Mantr in[ormation, coatings./orfirst phase of tests.

Coating Manufacturer's
Identification Color Used RecommendedMaximum
Number Coating Type in Tests ServiceTemperature, ~
8233 silicone black 425 continuous
540 excursions
8234 silicone black 455
8235 vinyl copolymer gray 7l continuous
82 excursions
8236 modified silicone black 400 continuous
540 excursions
8237 copolymerized
silicone white 260 continuous
290 excursions
8238 epoxy phenolic green 140
8239 silicone alkyd and clear 150
silicone
TABLE 3--Manujacturer's #ljbrmation, coatingsfor second phase of tests.
Coating Manufacturer's
Identification Color Used Recommended Maximum
Number Coating Type in Tests Service Temperature, ~
8318 silicone black 455

8321 vinyl copolymer gray 71 continuous
82 excursions
8322 epoxy polyamide black 95
8323 modified silicone black 400 continuous
540 excursions
8324 silicone black 425 continuous
540 excursions
8325 copolymerized silicone white 260 continuous
290 excursions
NOTE: Coating 8233 and 8324 are the same. Coating 8234 and 8318 are the same. Coating
8235 and 8321 are the same. Coating 8236 and 8323 are the same. Coating 8237 and 8325 are the
same.
TABLE 4--Application hzJormation r 1).
Film
Coating Type of Number Thickness,
# Coating of Coats mil dry Comments
8233 silicone 2(CS) 3 to 4 Dry spraying occurred on SS and CS surfaces

2(SS) 2.5 to 3 leaving a nonuniform film (that is, patchy
coverage).
8234 silicone 2(CS) 4 to 5 sagging problems with first coat
2(SS) 2 to 2.5 thinner not used with second eoat
8235 vinyl 2(CS) 4 to 5 second coat went on heavier than first
copolymer 2(SS) 3.5 to 4
8236 silicone 2(CS) 7 to 8
2(SS) 4 to 4.5
8237 silicone 2(CS) 5 to 7
copolymer 2(SS) 2.5
8238 epoxy- I(CS) 3 to 4 only one coat applied
phenolic I(SS) 3 to 4
8239 silicone 2(CS) 1 to 1.5 unable to achieve recommended film
alkyd 2(SS) 1 to 1.5 thickness; coating remained tacky after
five days; product slow to eure
NOTES: Material was applied over the following type surfaces: (1) carbon steel (CS)--sand-
blasted to white metal (SSPC-SP5 or NACE #1) and (2) stainless steel (SS)--solvent wiped
(SSPC-SPI). Conventional pressure pot spray equipment was used. 1 mil = 10 • 10 - s m.
TABLE S--Application information (Phase 2).

Coating
Identification Film Thickness,
Number Coating Type mil dry Comments
8318 silicone 4 to 4.5 applied over solvent cleaned

stainless steel
8318A silicone 4 to 4.5 applied over blasted
stainless steel
8321 vinyl copolymer 3 to 4.0 applied over solvent cleaned
stainless steel
8321A vinyl copolymer 3 to 4.0 applied over blasted
stainless steel
8322 epoxy polyamide 3 to 4.0 applied over solvent cleaned
stainless steel
8322A epoxy polyamide 3 to 4.0 applied over blasted
stainless steel
8323 modified silicone 2.0 applied over solvent cleaned
stainless steel
8323A modified silicone 2.0 applied over blasted
stainless steel
8324 silicone 3 to 5.0 applied over solvent cleaned
stainless steel
8324A silicone 3 to 5.0 applied over blasted
stainless steel
8325 copolymerized silicone 3 to 5.0 applied over solvent cleaned
stainless steel
8325A copolymerized silicone 3 to 5.0 applied over blasted
stainless steel
NOTE: All coating systems in Phase 2 were applied by the manufacturer. 1 mil = 10 •
10 - s m.
(3) resistance to various environments (weather, salt fog, industrial gases,

elevated temperatures, and thermal shock).
Adhesion of the coatings to stainless steel was evaluated using both the
scratch test and the Elcometer Adhesion Test. The scratch test is performed
by scribing a crosshatched ("#") pattern through the coating and firmly ap-
plying a piece of Scotch | tape. The tape is pulled off with a firm jerk, and the
adhesion evaluated by the amount of coating that comes off with the tape. In
the Elcometer Adhesion Test, metal disks, approximately 2 cm (0.75 in.) in
diameter, are attached to the coated panel using an epoxy adhesive, which is
allowed to cure for five days. An Elcometer Adhesion Tester is attached to the
disk and measures the force necessary to pull the disk and attached coating
off the stainless steel substrate. The strength of adhesion is measured in me-
gapascals (MPa).
To test coating flexibility, a coated stainless steel panel is placed over a
cylindrical 2.5-cm (1-in.) diameter mandrel and bent 180 ~ within a 15-s time
span. The coatings are inspected for cracking and disbonding.
For the outdoor exposure test, each coating is applied to 15.2 by 22.9-cm (6
by 9-in.) panels of carbon steel and 304 stainless steel, which were placed on
outdoor racks at Harmarville, PA and Port Arthur, TX. The panels are evalu-
ated annually.
Coated stainless steel panels are placed in a standard salt fog (ASTM Salt
Spray [Fog] Testing [B 117]) cabinet at 35~ (95~ and 100% relative hu-
midity. The panels are evaluated after four months and after one year.
Coated stainless steel panels are placed in an environmental cabinet con-
taining a gaseous mixture of hydrogen sulfide, carbon dioxide, nitrogen diox-
ide (0.67 volume % each), and air, saturated with water at 33~ (91~ The
panels are evaluated after 24 and 48 h.
Coated stainless steel panels are placed in a Xenon-Weatherometer (ASTM
Recommended Practice for Operating Light-Exposure Apparatus (Xenon-
Arc Type) With and Without Water for Exposure of Nonmetallic Materials
[G 26]) at 63~ (145~ where they are exposed to Xenon-arc light and cyclic
water spray (18 min water spray, and 102 min no water spray). Panels are
evaluated after 250 and 500 h.
To test for high-temperature stability, coated panels are placed in furnaces
at 260 and 400~ (SO0 and 7S0~ and evaluated after 24 h and 4 months.
Coatings that remain intact after the four-month exposure are retested for
adhesion with the Elcometer Adhesion Tester.
To test for ability to withstand thermal shock, coated panels are placed in a
furnace at 400~ (750~ for a minimum of 3 h and then immersed in room
temperature water. This cycle is repeated eight times. The panels are evalu-
ated after each cycle.
Results
Results of the adhesion, flexibility, outdoor exposure, salt fog cabinet, en-
vironmental cabinet, and Weatherometer tests are summarized in Tables 6
and 7. Tables 8 and 9 summarize the results of the elevated temperature and
thermal shock tests.
Good adhesion of a coating to stainless steel has generally been considered
a difficult goal to achieve because of the smooth oxide layer at the surface of
the metal. Our results would tend to confirm this, with most coatings having
only fair adhesion to stainless steeL. The exceptions were the epoxy polyamide
(#8322) and vinyl copolymer (#8235 and #8321), which both demonstrated
good adhesion. Adhesion of the coatings was essentially the same for solvent
cleaned (only) and sandblasted panels, indicating that the additional effort of
sandblasting is unjustified. Flexibility, which is less dependent on the sub-
strate, was good for a number of the coatings tested.
Despite only fair adhesion, most coatings performed adequately in the out-
door exposure, salt fog cabinet, environmental cabinet, and Weatherometer
tests. The copolymerized silicone (#8232 and #832S), tended to blister in tests
that constantly wetted the panel surface. The silicone alkyd, #8239, had little
o~
c
TABLE 6 - - R e s u l t s o/" tests at a m b i e n t a n d near a m b i e n t t e m p e r a t u r e s (Phase 1).
-<
Elcometer
Coating Coating Scratch Adhesion, Outdoor" Salt Fog h Environmental Z
Number Type Adhesion MPa (psi) Flexibility Exposure Cabinet Cabinet Wealtherometer
<
m
8233 silicone fair 1.25 (180) good fair (carbon poor (rust good good
steel was at edges)
rusted) -1-
t-
8234 silicone fair 0.95 (135) good good fair good (slight color change) good
8235 vinyl good 4.60 (665) good good good good (slight color change) fair (slight color change) O
z
8236 modified fair 2.25 (325) good good good good (slight color change) fair (color change)
silicone c)
-r
I-"

8237 copolymerized fair 1.40 (200) poor (cracks) fair fair (blisters) fair (blisters) fair (color change)
O
--11
silicone
8238 epoxy fair 1.65 (240) poor (cracks) good good good (color change) fair (slight color change) ~7
phenolic m
8239 silicone poor 0.50 (75) good poor (scratches poor poor (alligatoring fair (color change) co
---t
alkyd + and holes) (peeling) and cracks) m
silicone 6o
"After one year exposure. 0
t, Based on one year evaluation. -rl
"n
0
I
0
z
"4
",.4
TABLE 7 - - R e s u l t s o f tests at a m b i e n t a n d near a m b i e n t t e m p e r a t u r e s (Phase 2).
O
O
Elcometer 3o
3o
Coating Coating Scratch Adhesion, Outdoor" Salt Fog h Environmental O
Number Type Adhesion MPa (psi) Flexibility Exposure Cabinet Cabinet Weatherometer Go
5
Z
8318 silicone good 1.40 (200) fair-good N/A excellent excellent excellent
O
8318A silicone good 1.25 (183) fair-good N/A excellent excellent excellent -13
8321 vinyl excellent 2.80 (400) excellent N/A excellent fair-good good
copolymer I"11
.-q
8321A vinyl excellent 2.60 (375) excellent N/A fair fair-good good
copolymcr t"
co
8322 epoxy good 5.45 (783) excellent N/A excellent excellent excellent
polyamide
8322A epoxy good 4.75 (638) excellent N/A excellent excellent excellent
polyamide
8323 modified good 1.45 (208) excellent N/A fair fair poor

silicone
8323A modified good 1.40 (200) excellent N/A fair poor poor
silicone
8324 silicone excellent 1.40 (200) excellent N/A poor fair fair
8324A silicone excellent 1.45 (208) excellent N/A poor fair fair
8325 copoly- good 1.60 (225) fair N/A fair fair excellent
merized
silicone
8325A copoly- good 1.50 (216) fair N/A fair fair excellent
merized
silicone
N/A not available.

b Based on two months evaluation.
TABLE 8--Results of tests at elevated temperatures (Phase 1). c
W
260~ (500~ 400~ (750~ 2..<
Elcometer >
Elcometer Z
Coating Coating Adhesion," Adhesion, b Thermal
Number Type 24 h 4 Months MPa (psi) 24 h 4 Months MPa (psi) Shock <
rrl
8233 silicone good good 1.40 (200) good good 2.30 (330) good
8234 silicone good good 1.60 (230) good good (coating 1.65 (240) good "1"
lost its gloss) I'-
8235 vinyl fair poor poor (gone) poor O
copolymer (color change) (rubs off) z
8236 modified good poor poor poor C~
"1"
silicone (peeled off) (peeled off) I'-

8237 copoly- fair fair 1.10 (160) poor poor good O
-i1
merized (color change) (blisters and (blisters) (flaked off) u7
silicone color change) m
8238 epoxy poor ,,. poor (gone) poor 09
phenolic (flaked off) "4
2O
8239 silicone poor ... poor poor m
alkyd + (burned off) (burned off)
silicone O
O
a Room temperature test after 4 months exposure at 260~ (500~ "n
b Room temperature test after 4 months exposure at 400~ (750~ O
O
Z
..&
",4
Oa
TABLE9--Results of tests at elevated temperatures (Phase 2).

260~ (500~ 400~ (750~
Coating Coating Thermal
Number Type 24 h 2 Months 24 h 2 Months Shock
8318 silicone good good good good good
8318A silicone good good good good good
8321 vinyl fair fair fair poor poor
copolymer
8321A vinyl fair fair poor poor poor
copolymer
8322 epoxy fair fail poor fail poor
polyamide
8322A epoxy fair fail poor fail poor
polyamide
8323 modified good good fair fair good
silicone
8323A modified good good fair fair good
silicone
8324 silicone good good good good good
8324A silicone good good good good good
8325 copoly- good good good good good
merized
silicone
8325A copoly- good good good good good
merized
silicone
resistance to chemical or mechanical damage. One of the silicones, coating

#8233, had difficulties in several tests, primarily because of patchy coverage.
In the elevated temperature and thermal shock tests, only the straight sili-
cones (#8233, #8234, #8318, and #8324) performed adequately in both tests.
Adhesion of coatings (#8233 and #8234) to the stainless steel substrate was
actually improved after four months at 260~ (500~ probably because of
additional curing. The copolymerized silicone (#8237) performed well for
short periods of time at 260~ (500~ but blistered and lost adhesion after
four months exposure. Surprisingly, it survived the thermal shock test despite
400~ (7S0~ temperatures that exceeded the manufacturers maximum rec-
ommended temperature of 290~ (550~ Coating #8236, a modified sili-
cone, was inconsistent in its performance. As expected, coatings #8235,
#8238, #8322, and #8239 flaked or burned off in the elevated temperature
tests.
Conclusions
Results of testing to date indicate the following:
1. Little or no difference was noticed in the level of performance of coat-

ings applied over solvent cleaned stainless steel and coatings applied over
blasted stainless steel.
2. Most coatings tested perform adequately on stainless steel at ambient

and near-ambient temperatures despite only fair adhesion. However, the per-
formance of the epoxy polyamide and the vinyl copolymer were clearly supe-
rior to the others.
3. At elevated temperatures, straight silicones appear to give the best
performance.
Changes in Current Insulating Practice for Stainless Steels

In the past, some of our field personnel have been reluctant to apply coat-
ings over anything less than a blasted surface. We are now initiating field tests
of materials applied over solvent cleaned (only) stainless steel surfaces be-
cause of the results of this research program. In addition, we have identified
new materials and suppliers for our current recommended coating systems.
References
[1] Dillon, C. P. and Associates, Stress Corrosion Crackhlg o/"Stainless Steels and Nickel-Base
Alloys, Materials Technology Institute of the Chemical Process Industries, Inc., 1979, pp.
57-62, 104, 143-144.
[2] Mersberg, A. R. and Wee, F. W., Materials Performance, Vol. 19, No. 12, Columbus, OH,
Dec. 1980, p. 13.
[3] Willhelm, A. C., "Protective Coatings to Resist Salt Corrosion and Heating to 650~ '' Pro-
ceedings of the Air Force Materials Lab 50th Anniversary Technical Conference on Corro-
sion of Military and Aerospace Equipment, Denver, CO, 1967, p. 1581.
[4] Pilla, G. J. and DeLuccia, J. J., Metals Progress, Vol. 117, No. 6, June 1980, p. 57.
[5] Vegdahl, E. J., Damin, D. G., and Sumbry, L. C., "Eclectic Material Problems in the Petro-
chemical Industry," Preprint No. 17, Meeting of the National Association of Corrosion Engi-
neers, Anaheim, CA, 18-22 April 1983.
Charles T. M e t t a m 1
Designing to Prevent Corrosion of

Metals Under Insulation
REFERENCE: Mettam, C. T., "Designing to Prevent Corrosion of Metals Under Insula-

tion," Corrosion of Metals Under Thermal Insulation. A S T M STP 880, W. I. Pollock
and J. M. Barnhart, Eds., American Society for Testing and Materials, Philadelphia,
1985, pp. 178-187.
ABSTRACT: The paper concentrates on the corrosion of carbon steel under insulation.
Austenitic steel corrosion is also mentioned. Hot and cold insulation materials are dis-
cussed with the main consideration place on "cold" problems. The current solution of the
problem is to paint all carbon steel that is to be insulated and operating between -- 1 and
121~ Insulation material and thickness is then selected. A moisture barrier or a vapor
barrier is added to protect the insulation. Additional deterents added are vapor stops and
contraction/expansion joints. A European versus an American design is examined as a
conclusion. An Appendix, including protective coatings used, and drawings for addi-
tional clarification are included.
KEY WORDS: corrosion, insulation, permeability, urethane, coatings (under insulation)
F o r m a n y y e a r s t h e only " c o r r o s i o n u n d e r i n s u l a t i o n " p r o b l e m d i s c u s s e d

was stainless steel c o r r o s i o n . I n an e f f o r t to solve t h e p r o b l e m a M i l i t a r y Spec-
ification I n s u l a t i o n M a t e r i a l s , T h e r m a l , w i t h Special C o r r o s i o n a n d C h l o r i d e
R e q u i r e m e n t s ( M I L - 2 4 2 4 4 ) was f o l l o w e d t h a t a m o n g o t h e r t h i n g s l i m i t e d t h e
a m o u n t of c h l o r i d e s in i n s u l a t i o n to 600 p p m . U s i n g t h e s p e c i f i c a t i o n d i d n o t
e l i m i n a t e t h e stainless steel c o r r o s i o n p r o b l e m b e c a u s e it was t h e n f o u n d t h a t
c h l o r i d e s f r o m t h e a t m o s p h e r e also c o n t r i b u t e d to t h e c o r r o s i o n . A n i m m e d i -
ate s o l u t i o n was to a p p l y an i n e x p e n s i v e c o a t i n g to a u s t e n i t i c steel o p e r a t i n g
between -- 1 and 121~ before insulating.
Lately, severe c a r b o n steel c o r r o s i o n u n d e r i n s u l a t i o n or f i r e p r o o f i n g o r
b o t h has b e c o m e a m a i n c o n c e r n . S t u d i e s were c o n d u c t e d to i n v e s t i g a t e t h e
1Senior piping engineer, Lummus Crest Inc., 1515 Broad St., Bloomfield, NJ 07090.
MET'I'AM ON DESIGNING 179
problem. Many companies were unaware of their insulated carbon steel cor-
rosion problems. Included in our investigation was an ethylene plant con-
structed about 5 to 10 years ago. A "cold" piping system was examined. The
insulation used was urethane with an inadequate vapor barrier. The moisture
condition caused by "ice melt" was so bad that it was difficult to continue our
task and stay dry. The observed failure of the insulation system resulted in
large-scale corrosion of the structural steel supporting the vessels and pipe.
Replacement of the steel at great expense and downtime to the owner was the
inevitable outcome. Another large company recently interviewed revised their
specifications and now requires two coats of coal tar epoxy at 200/~m/eoat
under insulation. The change is indicative of a solution to a definite corrosion
problem. Corrosion under insulation is indeed a very serious condition as is
evidenced by the number of people attending this meeting. Extensive repair
must be made to rectify the problem, especially in cold service. The repairs
are very costly. They are difficult to make while systems are on stream, and in
some cases unit shutdowns are necessary before repairs can be made.
Design for Corrosion Under Insulation

Our solution to the corrosion problem is serious consideration during the
design phase of a project. The first step is the painting specification (see Ap-
pendix). Epoxy coating was selected as the primary line of defense under the
insulation. The coating works very well as the generic primer normally recom-
mended by fireproofing suppliers. As shown in a commercial blast (Appen-
dix) with 100-#m, dry film thickness is used. At higher temperatures the coat-
ing will change color and may even fail, however, any electrolytes present will
not normally cause corrosion. At these elevated temperatures the epoxy speci-
fied is hard and very resistant to acidic or basic attack. It will accept spray on
insulation as well as block and blanket type insulation.
The next step in our design is proper insulation selection. We normally
specify calcium silicate, fiberglass, mineral wool, perlite, or urethane for hot
insulation. Cellular glass and urethane are specified for cold insulation. Fi-
berglass is one type of insulation that is not recommended for cold service.
Manufacturers' literature tabulates fiber glass thickness based on tempera-
ture and relative humidity; however, our experience has been that manufac-
turers are unable to show us successful fiber glass installation for cold service.
A very recent client experienced actual failures of fiber glass cold insulation,
which reinforces our stand.
Moisture prevention is our third major consideration. We achieve success
by utilizing metal jacketing with a heat sealed moisture barrier over hot insu-
lation. Bands are placed in locations so that overlaps and wide openings do
not permit entry of water. To prevent the insulation from getting wet during
construction, it is specified that no insulation be left uncovered after working
hours. Compliance with this is very hard to obtain and the activity should be
monitored. On a job a few years ago, the heating up of an insulated line was
witnessed. Water poured from the jacketing to such an extent that it ap-
peared that a weld had failed. It turned out that the insulation was soaked
during recent rainstorms, and the water was being expelled in the warming
process.
We impose special requirements on cold insulation handling. Every effort
is made to ensure that the vapor barrier is not mechanically damaged during
installation. Injury to the barrier permits warm circulating air and subse-
quent moisture to reach the steel.
2"-
I
FIG. 1--Cold insulation for foamed in situ spacing construction.
METTAM ON DESIGNING 181
During a recent overseas project a European company's cold insulation sys-

tem was reviewed [1]. A facility insulated by them having two units similar to
the one we were constructing was visited. One unit was operating and the
other was down for a "turn around." Signs of icing on the unit in service were
not present. Plant personnel had stripped the "foamed in-place urethane"
from some of the lines and equipment on the "down" unit. We saw no signs of
corrosion, in fact, some of the painted pipe identifications were still legible
after ten years.
Figures 1 through 3 are part of the company's design that went into the
ik':: I
-o -
II_O~_
-----L L--
---P F--
FIG. 2--Insulation of fitting.
F " Z E P P E L I N ~ TYPE
FIG. 3--Form o f vesses head.
insulation project. The specifications were more than adequate. The main
design ingredients the European contractor used for prevention of corrosion
are metal jacketing, sealing tape, plastic grommets, sealer, and a government
regulation, "Insulation Works--Protection Against Corrosion for Cold and
Hot Insulation at Industrial Plants." Their methods are summarized as fol-
lows:
1. The initial step is painting the steel.

2. Complete metal jacketing is then used for the insulation of horizontal
equipment, as well as for piping, flanges, valves, and so forth. The object to
be insulated is completely covered with metal jacketing and spacers. A tele-
scopic metal jacketing can be used for the insulation of vertical equipment
and for vertical piping. Rings of flat iron with spacers are placed at a distance
of 1 m maximum. Urethane blocks are used as spacers if rings cannot be
installed. The jacketing is cut to size and installed with an overlap of 50 m m
on both horizontal and vertical seams. The sheets are fixed with self-tapping
screws at a m a x i m u m of 100 m m distance. The jacketing follows the outline
of the equipment. Curved heads are covered using the Zepplin shape as shown
in Fig. 3. In order to attain a higher stiffness the sections of sheet metal are
hemmed. Supports, skirts, and so forth in direct contact with equipment and
penetrating the insulation are insulated for a length equal to five times the
insulation thickness with a minimum of 300 mm.
ME'I-rAM ON DESIGNING 183
-
8
FOIL
FIG. 4--1nsulation of valves.
3. Foam is then pumped in through openings drilled into the metal jacket-
ing. The openings are fitted with grommets after pumping.
4. If parts of the equipment, such as manholes or nozzles, cannot be insu-
lated at the same time as the equipment, jacketing must be installed very
close to these parts and banded tight. The jacketing is cut back when the
parts are insulated. This is necessary to make a good foam connection and to
avoid thermal leaks.
5. Finally all joints, laps, and crimps of the metal jacket are covered with
A FUEXlIILE VAPI]R ilAJItRIER SEAL FOR

[~I'ANSIG~4/CONTRACTION JOINTS k~HERE
14VTR 15 k CRITICAL R(OUIREIIENT. HIlilHH
(
A F O I L / 1 F I L J ~ M I R I C COIqPT~ITE FOR THE mll'~l IZlllllllil
9[A/.S THAT | S TOUGH. T [ A R RESISTA~qT
i ~ S L J ~ PEIb~(ANC( ANO ~d[TAINS F L E X I I I I L I T Y IUlllI[III |
AT C~qYO~q|C TEIIPERATU~ES IS tJ~r~o.
NONIN~:. * - -
PIPE S I Z E Lo 9
'~~ I.t?
So 1.4
V I .$S
t0" t .e
12" 2.0~
t4" 2.21
re" 2.3r
tr 2.S4
2O" 2.79
24" 3.17
]r 3.r
M" 4.11
ffqiEFAORICAT[014 SAVES TIME I N TI4E

FI(L/) ~ li~Ika~S QUA..I?y C.J~NTIqqL
tmrT0 THE ~ACTOnV ~ S ~ V T~4s S ~ s
C~I-T0-t..I[~OSTN. ~ o F o t . ( ~ o , ~4n
~ l P q ' D g ~ WITH MASTIC A~4ESlqq[ F ~
IQ.m01NIO 1 0 THE J O I N T .
TN[ m_lnm_t. MONK~OENI~ ~O"ESlVE

| $ CI01~AIII[L~ UITH L0~ TE~'qP~JIATkJ~
INSULATIONS. r AND J ~ K E T I I 4 ~ .
I T IS ~ / ~ P L I E O IN 2 eARnLUEL STRIPS
v~,~x c ~ eE EASILY ~ E 0 TO e , n o v l ( ~
Ir~OA/dCy ANO A S T R O m C o ~ I r ( x ~ _ I ~ N O .
FIG. S--Prefabricated expansion~contraction,joint.
sealing tapes. The tape must be permanently elastic and weather resistant. As
stated previously, our practice is to specify cellular glass or urethane for cold
insulation. In our design one of the most important items other than insula-
tion thickness is the vapor barrier used. Normally used is a product with a low
water vapor permeability. Most material can be sprayed or trowelled on.
With either of these methods the thickness of the material can vary considera-
bly, unless it is closely monitored. To avoid thickness inconsistency, we spec-
ify the product to be supplied in sheets of standard thickness. Also specified is
the adhesive to be used for bonding the sheets together. Pipe fittings have not
yet been developed for sheets so a sprayed or trowelled on vapor barrier is
specified. Valves are handled slightly different (Fig. 4). The use of prefabri-
cated expansion/contraction joint seals [2] is a necessity in designing against
failures leading to corrosion (Fig. 5). Since our best design does not prevent
human error or acts of God, termination seals [3] are used (Fig. 6) to prevent
a failure from propagating into the system. Our design is monitored periodi-
cally by our engineering staff with visits to projects after they have been in
operation a while.
MET-I-AM ON DESIGNING 185
A ~LEXIBP=E VAPORBARRIER SEAL FOR INSULATION TERMINATIONS

WdERE LOW WVTR IS CRITICAL REQUIREMENT.
SEAL FABRICATEDFROM A TOUGH, TEAR RESISTANT, LOW

PERftE/gBILITY FOIL/FILH/FABRIC COMPOSITE THAT RETAINS
FLEXIBILITY AT CRYOGENIC TEMPERATURES.
DI D2
PIPE O.D.
\
THE SEALS ARE SUPPLIED PRE-SHAPED TO MATCHSTEPPED
TERMINATIONS FOR MULTILAYER INSULATION AND INCLUDE
ADHESIVES FOR BONDING THEM TO THE PIPE AND
INSULATION. PREFABRICATION ASSURES QUALITY CONTROL,
ELIMINATES HAND TAILORING IN THE FIELD AND SAVES
INSTALLATION TIME.
A PLIABLE, NONHARDENING MASTIC ADHESIVE FOR

BONOING THE SEAL TO THE INSULATION OUTER SURFACE
IS PRE-APPLIED AT THE FACTORY. IT IS COMPATIBLE WITH
TYP. INSULATIONS,MASTICS AND JACKETING USED FOR LOW
TEMPERATURE SYSTEMS. THE ADHESIVE CAN BE EASILY
WORKED IN PLACE TO OBTAIN A STRONG CONFORMAL BOND.
NORMAL D § 0 +
PIPE 1 2
SIZE 9 9 ADHESIVE FOR THE CLOSURE SEAM AND COLD
JUNCTION AT THE PIPE IS A 2-PART
URETHANE FORMULATED FOR CRYOGENIC
4" .23 .33 SERVICE. IT IS SPPLIEO AS A PRE-
E" .30 .41 MEASURED KIT SUITABLE FOR USE WITH
8" .318 .48 MANUAL OR AUTOMATICCAULKING EGIPflENT.
10" .43 .$3
12" .51 .61
14" .SG .~
16" ,61 ,71
lr ,s .TG
20" .74 .84
24" .Im .97
1.01 1.12
" I.l? 1.27
FIG. 6--Prefabricated termination seal.
A representative recently inspected a process plant that we built approxi-

mately six years ago. Some of the design features [4] of the cold insulation on
that project are shown in Fig. 7. The insulation, piping, and related areas
were in excellent condition. The only places that showed any signs of failure
were where insulation was removed in order to get at a valve or flange. The
maintenance personnel failed to reinsulate correctly, and ice is now forming
in these areas.
The selection of a good contractor, careful installation inspection, and ad-
herence to our design specifications insure an excellent insulation job.
FIG. 7--Design features of the cold insulation.
APPENDIX
Surface Preparation and Coating System for Insulated and Fireproofed Carbon and
Low Alloy Steel
Limitation
1. Where metal surface temperature is between --1 ~ and 121 ~
2. Vessels-tanks, towers, drums, fireproofed skirts, fireproofed heads, fireproofed
structural steel, fired heaters, exchangers and similar equipment, pipe hangers, and
piping.
3. Priming is to be done in the shop unless otherwise specified.
METTAM ON DESIGNING 187
TABLE l--Coating system: epoxy system~epoxy

phenolic.
Primer, Minimum Dry Film
Manufacturer Thickness, ~m
A A-1 (epoxy phenolic)

100
B B-1 (amine epoxy)
100
C C-1 (polyamide epoxy)
100
NOTES: Paint manufacturer's recommendation for

painting application, such as mixing of paint, applica-
tion equipment, and so forth, shall be ridgidly followed.
Surface Preparation
1. Before blasting, grind smooth-sharp edges, rough welds, steel silvers, and so
forth, remove all oil, and weld spatter and flux.
2. Blasting media shall be a 16 to 40 mesh nonmetallic abrasive, a Society of Auto-
motive Engineers (SAE) GL-40 steel grit, a S-230 steel shot or equivalent abrasive that
will give a m i n i m u m surface profile of 25/~m deep and maximum of 65 #m with rogue
peaks not exceeding 75/~m. Blasting media shall be free of all oil and moisture.
3. Blast clean all steel surfaces and welds to be coated to SSPC-SP6.
4. No blasted surfaces shall be allowed to remain uncoated overnight in humid
areas. All paints shall be applied to a surface free from moisture, oil, dust, grit, or any
other contaminants and discoloration. The initial blast quality shall be maintained
immediately before painting.
References
[t] Rheinhold and Mahla, "Cold Insulation for Equipment and Piping with Polyurethane
(PUR) Foamed in Situ," Project Engineering Specification, Mannheim, West Germany,
Aug. 1977.
[2] Sheldahl, "Flexible Vapor Barrier Seal for Insulated Terminations," Northfield, MN,
1982.
[3] Sheldahl, "Flexible Vapor Barrier Seal for Insulation Expansion/Contraction Joints,"
Northfield, MN, 1982.
[4] Chicago Bridge and Iron Co., "Specification for Multi-Layer Polyurethane Pipe Insula-
tion," Oak Park, IL, June 1976.
James A. Richardson 1 and Trevor Fitzsimmons 1
Use of Aluminum Foil for Prevention

of Stress Corrosion Cracking of
Austenitic Stainless Steel Under
Thermal Insulation
REFERENCE: Richardson, J. A. and Fitzsimmons, T., "Use of Aluminum Foil for Pre-
vention of Stress Corrosion Cracking of Austenltie Stainless Steel Under Thermal Insula-
tion," Corrosion of Metals Under Thermal Insulation. A S T M STP 880, W. I. Pollock
and J. M. Barnhart, Eds., American Society for Testing and Materials, Philadelphia,
1985, pp. 188-198.
ABSTRACT: For many years, it has been preferred practice within Imperial Chemical
Industries (ICI) to use aluminum foil on austenitic stainless steel surfaces operating at
temperatures within the range 60 to S00~ The foil protects in two ways. It presents a
physical barrier to chloride-containing fluids migrating through lagging materials to-
wards hot stainless steel surfaces. It also cathodically protects stainless steel in the event of
flooding of the lagging system, thereby preventing initiation of pitting and stress corrosion
cracking. ICI's experience with aluminum foil is summarized. Laboratory data are pre-
sented that confirm the galvanic protection afforded by the foil, and the efficiency of the
foil in preventing chloride stress corrosion cracking relative to a number of coating sys-
tems specified for the same purpose.
KEY WORDS: aluminum foil, stress corrosion, austenitic stainless steel, thermal insula-
tion, cathodic protection, silicone-alkyd coatings, aluminum-rich silicone coatings, zinc-
rich coatings
External stress corrosion cracking of austenitic stainless steel is a long-

standing problem in the process industries. Its phenomenology has been dis-
cussed admirably in numerous previous publications and in other papers in
this STP, and it is not necessary to describe it in detail in this paper. For the
problem to occur, three basic conditions must be met:
1Senior corrosion engineer and materials engineer, respectively, Imperial Chemical Industries
PLC, Engineering Department--North East Group, P.O. Box 6, Billingham, Cleveland TS23
1LD, United Kingdom.
RICHARDSON AND FITZSIMMONS ON ALUMINUM FOIL 189
1. The stainless steel surfaces must operate continuously, or intermit-

tently, at temperatures > approximately 60~ The few significant failures
we have experienced in Imperial Chemical Industries (ICI) over the past 10 to
15 years have been on surfaces operating at < approximately 200~
2. There must be soluble chlorides present. This topic has been discussed
in detail elsewhere} but suffice it to say that virtually all insulating materials,
perhaps with the exception of cellular glass, contain some soluble chloride,
and that chloride is a common contaminant in air, rainwater, most process
waters, and other process fluids that can enter lagging systems.
3. Water must enter the lagging system. In terms of preventing stress cor-
rosion cracking it is convenient to distinguish two "degress" of water ingress:
a. Migration--Weatherproofing/vapor sealing systems are rarely 100%
efficient. At best the lagging system will still be able to breathe allowing
access of atmospheric moisture, and in addition there may be minor leaks
in the seals allowing ingress of small quantities of rainwater, and so forth.
The extent to which such water will migrate towards, and wet stainless steel
surfaces will depend upon the "wicking" properties (if any) of the insula-
tion material, and the process temperature in relation to the dew point of
the lagging "atmosphere." Given that process temperatures may vary and
that there are diurnal and seasonal variations in temperatures and humid-
ity, there is obviously scope for alternate wetting/drying phenomena to
concentrate any chloride present at the metal surface.
b. Flooding--Where weatherproofing or vapor sealing systems have
been poorly designed or executed or both, or have deteriorated significantly
after prolonged periods of service, then ingress of substantial quantities of
water is possible. Common sources of aqueous fluids include, apart from
rainfall, wash waters, condensate from steam tracing joint leaks, quench
system waters, and process fluids leaking from joints. Permanently flooded
lagging systems offer little, if any, thermal insulation and are thus likely to
attract attention. However, intermittent flooding may well go undetected,
or at least be tolerated, together with the attendant mechanisms for carry-
ing soluble chlorides to the hot metal surface where they can concentrate in
a series of wetting/drying cycles.
Accepting these basic conditions for cracking, it follows that the first and
obvious preventative measure is to keep water out of lagging systems. Lagging
systems need their share of "good engineering practice," which means appro-
priate levels of design, inspection, and maintenance. This topic is covered in
other papers in this STP and will not be pursued further here, but its crucial
importance is acknowledged. However, most process plant operators seek to
provide some additional protection against stress corrosion cracking, and ICI
is no exception.
2Richardson, J. A., in this publication, pp. 42-59.

Development of a Cracking Prevention Policy

When the problem was first addressed within ICI approximately 15 years
ago, the defined policy objective was to identify a protection system that
would prevent stress corrosion cracking of stainless steel surfaces operating at
temperatures > 60~ resulting from "migration" of water (as defined above)
and occasional transient flooding of the lagging system. Three basic ap-
proaches to the problem were identified.
Control of Insulation Material Composition

This approach considered the possible specification of low-chloride or in-
hibited lagging materials or both in the hope of controlling the corrosivity of
aqueous extracts concentrating on hot stainless steel surfaces. It was con-
eluded that this approach was not sufficient to meet the policy objective. It
was recognized that the specification of materials with, for example, con-
trolled initial Na + SiO2/C1 + F ratios [1] could certainly reduce the risk of
cracking. However, progressive ingress of chloride via mechanisms of the type
discussed above would inevitably alter adversely the ratio and periodic flood-
ing if the lagging might well remove the soluble inhibitor. In any event, there
was a reluctance to carry the additional costs of over-specifying with respect to
national standards, and in 1970, for example, the appropriate British Stan-
dard (BS) for Thermal Insulating Materials (BS 3958, Part 2) allowed up to
"approximately 550-ppm chloride" in preformed calcium silicate. As has
been pointed out recently [2], the potential total chloride per unit area of
stainless steel surface is of more concern in practice than the chloride content
of the lagging per se.
Use of Coatings
This approach essentially involved the provision of a physical barrier be-
tween the stainless steel surface and any corrodent accumulating thereon. It
was concluded that this too was insufficient to meet the policy objective. Al-
though paints capable of operating at relatively high temperatures, for exam-
ple, silicone-based formulations were available, they could only be as efficient
in preventing cracking as they were free of holes and imperviousness. It would
be virtually impossible to achieve a defect-free paint system under a lagging
system, bearing in mind the likely damage caused during lagging application.
It would be costly to achieve a near defect-free system involving multi-coats,
spark, or sponge testing, and so forth. A relatively cheap, single-coat system
would inevitably leave a significant risk of crack initiation at defects.
Simultaneous Provision of a Physical Barrier and Galvanic Protection

It is well known that chloride stress cracking initiation in stainless steels is
potential dependent [3,4]. In particular, it is possible to prevent crack initia-
tion completely with appropriate levels of cathodic protection. It was thus

decided to explore the simultaneous provision of a physical barrier, to cope
with "migration" of water through the insulation and sacrificial cathodic pro-
tection, to cope with protection at holidays and with occasional flooding prob-
lems. Two approaches were identified:
1. Metal-Filled Paints--Aluminum-filled paints were ruled out, because
little if any galvanic benefit derives from their use because of relatively low
metal loadings and aluminum oxidation characteristics. Metallic zinc-rich
paints proved highly efficient at preventing cracking, but bearing in mind the
relatively low melting point of zinc ( - 4 2 0 ~ and the attendant concerns of
liquid metal embrittlement, 2 they were unlikely to find widespread use.
2. Aluminum Foil--This emerged rapidly as the leading contender among
the approaches evaluated. Laboratory work established its efficiency, both as
a physical barrier and in the provtsion of galvanic protection. Melting and
liquid metal embrittlement risks were considered significantly lower com-
pared with zinc. As long as thin gage foil was used, that is, light, virtually self
supporting, compliant, and easy to join, application problems and costs
could be reduced to a minimum.
The use of aluminum foil under lagging systems to prevent external stress
corrosion cracking of austenitic stainless steel surfaces operating between 60
and 500~ was first introduced into ICI lagging specifications approximately
15 years ago. It is the preferred preventative system, regardless of the lagging
material, and has been used on the large majority of austenitic stainless steel
vessels and piping systems constructed during that period. Where use of foil
has proved impracticable or undesirable for some reason (see below), paint
coatings have been used, with due recognition that, with the possible excep-
tion of metallic zinc-rich systems, they are not as efficient as foil in preventing
stress crack initiation.
Coating and Aluminum Foil Efficlencies in Preventing Cracking

The relative efficiencies of coating and aluminum foil in preventing crack-
ing have been evaluated in the laboratory on a number of occasions over ap-
proximately the past 15 years. The problem is to devise a test procedure that
allows valid statistical evaluation of the relative efficiencies of various preven-
tative systems in preventing crack initiation, while remaining relatively cheap
and convenient to perform. Coiled spring specimens are particularly conve-
nient in this respect, in providing a relatively large stressed surface area for
coating and for presentation to corrodents. In our recent tests, we standard-
ized on the use of specimens constructed from 3 mm diameter, American Iron
and Steel Institute (AISI) Type (304) (Unified Numbering System [UNS]
$30400) stainless steel wire, cold formed into a 10-turn helix with a 20-mm
internal diameter, providing a total surface area of about 65 cm 2. Typical
specimens are shown in Fig. 1.
FIG. 1--Coiled spring specimens used in stress corrosion testing.
Specimen surfaces were degreased with acetone before brush application of

single coats of paint coatings, that is, simulating site application conditions.
Alternatively, specimens were wrapped externally with 46 standard wire gage
(SWG) aluminum foil forming a cylinder around the coiled spring. No at-
tempt was made to establish electrical continuity other than crimping the foil
around individual turns in the spring. Specimens were subjected to full im-
mersion testing for periods, typically seven days, in various chloride solutions
in glass flasks under reflux at various boiling temperatures in the range 105 to
145~
At the conclusion of the tests, paint coatings were removed with appropri-
ate solvents and specimens were dye penetrant examined using fluorescent
dye under an ultraviolet (UV) lamp. The number of cracks were recorded.
The existence of cracks was confirmed by metallography, and a typical crack
is shown in section in Fig. 2. During some tests, natural corrosion potentials
were measured using a saturated calomel electrode in a glass sidearm coupled
to the test fluid by a sintered glass disk and a chloride-containing polyacry-
lamide gel. Some typical test data are presented in Tables 1 and 2.
The following conclusions emerged from the tests:
1. A single coat silicone-aklyd system, brush applied to a degreased stain-
less steel surface occasioned a 90 to 95% reduction in the incidence of crack-
ing compared with a bare surface. Curing of the coating at 120~ before test-
ing appeared to have little influence on protection efficiency at the levels of
FIG. 2--Microsection of typical stress corrosion crack in coiled spring specimen.
TABLE 1--1ncidence of stress corrosion cracking on coiled 304 spring specimens in boiling
saturated sodium chloride solution at 108~
Total Number
Corrosion Potential, of Cracks, Protection Efficiency,
Protection System mV/SCE" 4 Specimens %
None (control) -- 380 75 ...

Silicone-alkyd paint,
uncured -- 140 8 89
Aluminum-rich silicone
paint -- 390 8 89
Zinc-rich epoxy paint --720 2 97
Aluminum foil --910 0 100
"Potentials recorded at the test temperature of 108~ SCE is saturated calomel electrode.
coating integrity achieved. The apparently different protection efficiencies in

the two tests quoted were considered to arise from the different corrosivities of
the test liquors, rather than any variation in coating integrity between the
tests.
2. An aluminum filled silicone paint proved no more efficient than a
metal-free paint in preventing crack initiation, and the corrosion potentials in
Table 1 confirm little or no evidence of galvanic protection.
TABLE 2--1ncidence of stress corrosion cracking on coiled 304

spring specimens in boiling saturated calcium chloride
solution at 138~
Total Number
of Cracks, Protection Efficiency,
Protection System 4 Specimens %
None (control) 462

cured at 120~ for
1h 21 95
uncured 26 94
Aluminum foil 0 100
3. The particular zinc-rich paint tested reduced markedly the incidence of

cracking, and depressed the natural corrosion potential by approximately
350 mV cathodically. Indeed, two of the four specimens tested exhibited no
cracks at all, and it is surprising that any cracking was experienced at such
low corrosion potentials.
4. Aluminum foil proved 100% efficient at preventing cracking under the
test conditions, unsurprisingly in view of the very low natural corrosion poten-
tials achieved, involving a depression of about 530 mV relative to the control
corrosion potential. At such low potentials, the stainless steel is almost cer-
tainly fully cathodically protected [4]. The beneficial effects of aluminum foil
wrapping under Karnes test conditions have been confirmed recently [5].
The above data confirm the galvanic protection afforded by aluminum foil
under fully flooded conditions, and its advantages over single-coat paint sys-
tems. It is important to recognize that higher integrity (and costly) paint sys-
tems, that is, spark or sponge tested multi-coat systems on blasted surfaces,
would probably have performed as well as aluminum foil under test condi-
tions. Indeed certain of the single-coat specimens achieved 100% protection
in the tests quoted above, for example, one of the four springs coated with
aluminum-rich silicone quoted in Table 1 exhibited no cracking after seven
days. However, the purpose of the tests was to evaluate aluminum foil against
some relatively cheap single-coat site application procedures, and its advan-
tages in this respect have always been confirmed in laboratory testing.
Consideration of Fire Risks

In the case of plants handling flammable fluids, which can give rise to
flammable atmospheres, there are essentially three problems to address:
1. Is there an additional risk of spread of fire associated with the presence
of the foil per se?
2. Is there an additional risk of spread of fire resulting from liquid metal

embrittlement of stainless steel substrates by molten aluminum?
3. Is there an increased risk of ignition of a flammable atmosphere by in-
cendive sparks resulting from a "thermite" reaction between aluminum foil
and adjacent metal surfaces?
In relation to the first question, there is some dispute as to whether alumi-
num foil burns in air [6]. However, if exposed in the event of a fire it could
undoubtedly act as a source of hot, molten metal or oxide particles, which
propelled by expanding air, could raise surrounding materials to their igni-
tion temperatures [7]. But.if the foil is entirely contained beneath a lagging
system, this should not be possible unless the lagging material itself is flam-
mable. The presence of the foil in the context described in this paper, is thus
not considered to increase the risk of spread of fire.
Regarding the second issue, it is undoubtedly true that molten aluminum
can cause cracking of stressed austenitic stainless steel. However, laboratory
testing suggests that cracks do not initiate very readily2 and that the risks are
relatively minor compared with zinc, where the zinc/nickel reaction favors
crack initiation and propagation. There is some practical evidence to support
laboratory findings. For example, during one reported refinery fire, both
molten zinc and aluminum had access to stainless steel piping, but embrittle-
merit problems were restricted to those caused by zinc [8]. In one instance, we
have experienced melting of foil because of a temperature excursion on some
lagged stainless steel ducting, without any evidence of liquid metal attack.
Again, given the present context involving foil beneath a lagging system, the
likelihood of melting induced by external flame impingement would seem re-
mote and the risks of spread of fire caused by embrittlement acceptably low.
Regarding the "thermite" reaction risks, it is recognized that aluminum
foil if placed on rusty steel and subsequently struck by a hard object, could
ignite a flammable vapor through the creation of an incendive spark [9].
However, there is no evidence that such sparks can be produced between alu-
minum and a clean stainless steel surface, and in any event, there would inevi-
tably be some attentuation of impact energy within the lagging system where
the foil is entirely contained. It has thus been concluded that the risks associ-
ated with possible aluminum foil/stainless steel surface interactions beneath
a lagging system are acceptably low. However, foil is not used in locations
where it could contact carbon steel surfaces in service, for example, where
carbon steel backing flanges are used on stainless steel lines carrying flamma-
ble fluids. In these circumstances silicone-based paints are used.
Given the qualifications outlined above, it has been concluded that the ad-
ditional fire risks associated with the use of aluminum foil are acceptably low.
They must be balanced against the additional fire risks associated with exter-
nal stress corrosion cracking itself, which can obviously result in the leakage
of hazardous/flammable materials into the plant atmosphere.
The above comments are not to be taken as a justification for the extensive
use of aluminum alloys in high fire risk areas. ICI shares the concerns of other
major operators on the latter topic, and the comments above relate strictly to
the context of the paper.
Practical Application and Performance of Aluminum Foil

All ICI lagging specifications require that austenitic stainless steel surfaces
operating continuously or intermittently between 60 and 500~ be wrapped in
46-SWG aluminum foil. Application to pipe and vessel surfaces presents few,
if any, problems, largely because of the lightness and compliance of the foil.
Pipes are simply wrapped in the foil with 25-mm minimum overlaps, formed
so as to shed water on vertical lines. The foil can be "molded" around flanges
and fittings. On vessels, the foil is usually applied in bands in advance of the
lagging system. There are no support problems, because given a relatively few
anchor points, the foil is virtually self supporting. Obviously, there are no
support problems on the top dome. On the sides, there are usually lagging
support rings that provide anchor points. On the base, there are sprags or
other means of lagging support that can be used to support the foil. Over-
lapped, crimped joints are executed to shed water.
In the case of steam traced lines, a double wrapping of foil if specified, one
directly onto the pipe beneath the tracing and the other enclosing both tracing
and pipe in order to exclude unwanted insulation from the space between
them. Overlaps on vertical lines are again formed so as to be water shedding.
All compressive or flanged joints in steam tracing are specified to be outside,
preferably underneath, the main pipe insulation system.
Some relative typical costs of applying foil and single paint coatings are as
follows:
aluminum foil, s 2
aluminum-filled silicone, s 2
silicone-alkyd, s 2
The attractions of foil are self evident. Foil has the additional advantage that
its application by lagging trades is accepted, whereas the use of coatings re-
quires the involvement of additional trades in the lagging process, with conse-
quent additional timing/management problems.
During the approximate 15-year period since aluminum foil was first speci-
fied, there have been no failures of austenitic stainless steel surfaces caused by
external stress corrosion cracking where the foil had been applied to specifi-
cation. Admittedly, a relatively small sample of the total lagged surface area
has been available for inspection during that period, but a small number of
vessels and piping systems have had their laggings removed after periods of
service up to 10 years without any evidence of cracking. There have been a few
instances of failure of stream traced lines during this period where the protec-
tion system failed to meet specification for one or more of the following rea-
sons:
1. A single wrap of excessively thick, rigid material (thin sheet) had been
used, instead of a double wrap of foil. Rigid material tends to stand away
from pipe surface, denying intimate contact between the aluminum and the
pipe surface, and making overlaps difficult.
2. Flanged joints in the steam tracing had been located so as to make
weatherproofing of the adjacent lagging system virtually impossible, and so as
to leak into the lagging system.
3. Steam lances had clearly been pushed into the lagging to assist in clear-
ing blocked lines, with consequent excessive damage to the lagging system.
In all cases, there was clear evidence that significant quantities of conden-
sate had evaporated to dryness on pipe surfaces isolated from contact with
aluminum in the "cavities" between the excessively rigid sheet and the pipe
surfaces. Under these circumstances, galvanic protection cannot operate.
In the latter context it is appropriate to comment on corrosion of the alumi-
num foil, which is < 5-mil thick. Obviously, if such thin foil is exposed to
flooded lagging for prolonged periods then significant corrosion is bound to
occur. Experimental work reported here and elsewhere [5] indicates that gal-
vanic protection is achieved even with a significant degree of perforation of
the foil. For example, at the conclusion of the tests in saturated calcium chlo-
ride solution reported in Table 2, the foil was extensively perforated as a
result of pitting corrosion. However if the gaps in the foil achieve dimensions
such that aqueous extracts can concentrate on austenitic stainless steel sur-
faces in isolation from the foil, then galvanic protection cannot operate. Cor-
rosion of this severity had occurred in some of the steam tracing failures cited
above.
In this context it is important to recall the initial primary objective in devel-
oping the aluminum foil system, that is, that it should prevent stress corrosion
cracking arising from "migration" of water and occasional transient flooding
of the lagging system. None of the available stress corrosion cracking preven-
tion systems, aluminum foil, coatings, inhibited laggings, for example, can
provide unlimited protection during prolonged exposure to flooded insulation
systems. This automatically imposes a primary requirement for prevention
that adequate levels of weatherproofing and vapor sealing be maintained
throughout a plant's life.
Finally, the few significant failures caused by external stress corrosion
cracking we have experienced in ICI over the past 10 to 15 years have been on
surfaces either without any protection at all or protected by coatings. In the
latter category, the most significant failure involved a number of columns op-
erating in the temperature range of 110 to 180~ which had been coated with
a zinc oxide filled silicone-alkyd paint before lagging. After approximately 10
years service at a coastal site, the Type 304 columns were found to have suf-
fered extensive stress corrosion cracking. Cracks had presumably initiated at

holidays in the coating, but had propagated extensively beneath the appar-
ently intact coating. Indeed, the coating had to be removed by surface grind-
ing to facilitate crack detection.
Taken overall, our experience tends to confirm the relative efficiencies of
the various protection systems suggested by the laboratory work reported in
Tables 1 and 2, and to reinforce our view that aluminum foil is the most effi-
cient, cost effective available system for preventing initiation of stress corro-
sion cracking.
Conclusions
1. Thin, 46-SWG aluminum foil can be applied to austenitic stainless steel
surfaces beneath lagging systems to prevent chloride-induced stress corrosion
cracking.
2. The foil acts as a physical barrier to the "migration" of small quantities
of aggressive fluid towards stainless steel surfaces, and provides galvanic pro-
tection in "flooded" lagging systems, preventing initiation of and pitting and
stress corrosion cracking.
3. Laboratory tests confirm that aluminum foil is more efficient than sin-
gle-coat paint systems in reducing the risk of stress corrosion crack initiation.
4. Zinc-rich paints are also highly efficient at preventing stress corrosion
crack initiation, but there are more significant attendant concerns about
liquid metal embrittlement than in the case of aluminum foil.
Acknowledgments
A number of members of ICI's Materials Group have contributed to the
development of an aluminum foil policy over the past 10 to 15 years. In partic-
ular the authors would like to acknowledge J. G. Hines and M. E. D. Turner,
much of whose distilled wisdom is presented in this paper.
References
[1] U.S. Atomic Energy Commission Regulatory Guide 1.36, "Non-metallic Thermal Insulation
for Austenitic Stainless Steel."
[2] Nicholson, J. D., Bulletin of the Institution of Corrosion Science and Technology, Vol. 19,
No. 5, Oct. 1981, pp. 2-5.
13] Wranglen, G., An Introduction to Corrosion and Protection of Metals, Institute for Me-
tallskydd, Stockholm, Sweden, 1972, pp. 117-118.
[4] Andresen, P. A. and Duquette, D. J., Corrosion, Vol. 36, No. 8, Aug. 1980, pp. 40%415.
[5] Gillett, J. and Johnson, K. A., presented at Conference on Corrosion Under Lagging, Insti-
tution of Corrosion Science and Technology, Nov. 1980, published by Fibreglass Ltd.
[6] West, E. G., The Metallurgist and Materials Technologist, Vol. 14, No. 9, Sept. 1982,
pp. 395-398.
[7] Ruskin, A. M., The Metallurgist and Materials Technologist, Vol. 15, No. 7, July 1983,
p. 358.
[8] Cantwell, J. E. and Bryant, R. E., Hydrocarbon Processing, May 1973, pp. 114-116.
[9] Eisner, H. S., The Engineer, 17 Feb. 1967, pp. 259-260.
John W. Kalis, Jr. 1
Using Specifications to Avoid

Chloride Stress Corrosion Cracking
REFERENCE: Kalis, J. W., Jr., "Using Specifications to Avoid Chloride Stress Corro-
sion Cracking," Corrosion of Metals Under Thermal Insulation, A S T M STP 880, W. 1.
Pollock and J. M. Barnhart, Eds., American Society for Testing and Materials, Philadel-
phia, 1985, pp. 199-203.
ABSTRACT: Insulated austenitic stainless steel above the critical temperature must be
clearly identified to receive specific protection. Specify appropriate insulation and
weather barrier material to prevent moisture entry. Proper application of insulation mate-
rials is an important part of the specification. Considerations for applying additional pro-
tection to austenitic stainless steel are dependent on the operating temperature and criti-
cal service of the insulated facility.
KEY WORDS: insulation, specifications, corrosion, stainless steels, protection
It is well recognized that external stress corrosion cracking (ESCC) of aus-

tenitic stainless steel is induced by the presence of moisture, chlorides, tensile
stress, and temperature in the right combination. Most rain, domestic or
plant water, contains chlorides that can be concentrated by evaporation at the
stainless steel surface.
Dry insulation does not cause ESCC. However, insulation does provide the
vehicle for collecting or directing water to areas of the metal surface where
chlorides can be concentrated and cause ESCC.
Therefore, the insulation specification must include a moisture barrier to
prevent moisture from entering the insulation system. The specification
should specify insulation that can retard the migration of moisture and pro-
vide for additional protection, such as a coating or paint system, applied to
the surface of critical items. The specification must clearly identify all insu-
lated austenitic stainless steel pipe and equipment along with operating tem-
perature, since ESCC usually takes place above 60~ (140~
~Project specialist engineer-insulation, E. I. du Pont De Nemours Company, Inc., Wilming-

ton, DE 19898.
The primary objective is to prevent moisture from entering the insulation

system. This can be accomplished with some reliability by specifying proper
materials and insulating techniques. It should be stated that even with the
best insulation materials and installation, all systems require periodic inspec-
tion and maintenance to ensure that the weather barrier integrity is intact.
The specification must indicate that the insulation is to be stored and in-
stalled dry over a dry surface. Many insulation manufacturers state that mod-
erate amounts of water will not degrade the insulation and that once the pro-
cess is turned on, the heat will drive the moisture out of the insulation. This is
partially true, however, the chlorides in the moisture are left behind, which
can cause ESCC. The insulation must be protected from rain until the
weather barrier finish is properly applied. This is an area where most insula-
tion contractors have difficulty in complying.
Some insulation absorbs greater amounts of moisture than others. For ex-
ample, calcium silicate can allow water to migrate through the insulation of
an entire pipe system or piece of equipment. This accumulation of moisture
can reach saturation without visual detection. A small opening in the weather
barrier of a pipe elbow can allow a large amount of water to enter and gener-
ate a high concentration of chlorides on the pipe surface. This concentration
of chlorides may take place at a remote distance from the source of the mois-
ture entry.
It is preferred to specify an insulation material that retards the spread of
moisture. Polyisocyanurate insulation is relatively nonwicking and is pre-
ferred for use within its temperature limitations. For higher temperatures, a
water-resistant perlite silicate insulation is often used. Nonwicking insulation
is used on traced piping because the tracer often cannot generate sufficient
heat to drive out water from the system.
For many years the inhibiting action of sodium silicate in the insulation or
sodium silicate applied directly to the austenitic stainless steel surface was
relied upon for protection. The application of sodium silicate on the metal
surface is now specified as minimum protection, even where inhibited insula-
tion is used.
Pressure sensitive tapes are normally specified to secure lightweight insula-
tion to pipe. Many tapes have adhesive that contain chlorides. Chloride
cracking on austenitic stainless steel pipe where pressure sensitive tape was
applied directly to the pipe surface has been observed. A limited number of
chloride-free tapes are available; some of them can be certified to be chloride-
free.
No tape of any kind is permitted on the surface of any pipe or equipment,
since the monitoring of approved tapes on the job site can be difficult.
Insulation weather barriers must be carefully evaluated for suitability to
specific applications. The weather barrier must keep the insulation dry. Alu-
minum jacketing is specified on pipe and equipment for many heavy indus-
trial applications. The aluminum jacket thickness and application will vary
KALIS ON USING SPECIFICATIONS 201
for pipe and equipment. Aluminum jacketing provides an excellent low-cost

moisture barrier when installed properly. The insulator is instructed to in-
stall all aluminum jacketing in a water shed fashion. All jacket seams are to
overlap a minimum of 51 mm (2 in.) on pipe and 76 mm (3 in.) on equipment.
The weather barrier or jacket over pipe insulation is specified as 0.4-mm
(0.016 in.) thick aluminum. All fittings, such as tees and elbows, flanges, and
valves, are covered with aluminum. Aluminum end caps are used where the
pipe insulation terminates at valves and flanges to seal the insulation ends
from moisture. The aluminum end caps provide protection when the insula-
tion is removed from the flange or valve for service. In some cases the valve or
flange may not be reinsulated immediately, thus exposing the pipe insulation
end to moisture.
An aluminum weather barrier is used for most cylindrical equipment such
as vessels, columns, or tanks. For equipment up to 1.9 m (42 in.) diameter,
0.6 m (0.024 in.) thick smooth aluminum jacketing with spring tensioned
stainless steel bands is normally used. This heavier metal jacket with ten-
sioned bands will hold the jacket firmly against insulated equipment. The
bands also prevent moisture from working into the overlapped joints. For
equipment larger than 1.9 m (42 in.) in diameter, a pre-engineered panel sys-
tem is preferred. The panels are normally 2.4 m (8 ft) wide by 1.2 m (4 ft)
high. The aluminum jacket is 0.8 mm (0.032 in.) thick and with moisture-
resistant polyisocyanurate insulation permanently secured to the jacket. The
inside surface of the insulation is faced with an aluminum foil for additional
moisture protection. The vertical seam of the panel contains a neoprene gas-
ket for sealing and is interlocked for strength. The panels are secured to the
equipment with large spring-tensioned metal bands. Several plant sites have
successfully used pre-engineered panels on cylindrical equipment for over ten
years. Where large equipment is operating above the allowable temperature
for application of the pre-engineered panel, then a separate jacket material is
indicated. The deep-corrugated aluminum sheets, 0.6 mm (0.024 in.) thick,
provide fewer seams for possible water entry and can be installed with less
labor than the rolled smooth aluminum. The corrugations also provide addi-
tional strength.
Many insulation system failures on equipment begin with metal flashing
improperly designed and applied. High winds can blow away the flashing,
exposing the insulation to moisture. Heavy aluminum flashing with a mini-
mum thickness of 0.6 m m (0.024 in.) is normally used to cover projections on
equipment. For example, stiffener rings that extend beyond the equipment
wall insulation require the fabrication of metal flashing for best protection.
The flashing must be designed for strength and be properly secured to al-
low for expansion and contraction of the equipment. It must prevent water
from collecting and also must shed water. Most insulation contractors are
capable of fabricating special flashing requirements, however, it is advisable
to provide details in the specification.
It is very difficult to provide a good insulation weather barrier on irregular

shapes such as dished heads on vessels. Reinforced mastics have been speci-
fied in the past for this application but experience has indicated that frequent
maintenance is required to prevent failure. Therefore, this finish is only per-
mitted in areas where the insulation cannot get wet.
A finish of fiber-reinforced polyester (FRP) or epoxy provides excellent re-
sistance to weather, mechanical abuse, and atmospheric corrosion. This fin-
ish has no seams and can be sealed around equipment nozzles and other pro-
trusions. Some training is required to properly mix and apply FPR as a finish,
but most insulation contractors are familiar with it. FRP is specified for the
heads of equipment and irregular shapes where metal forming is not practi-
cal. Two component epoxy resins reinforced with synthetic organic fiber open
weave cloth have been used successfully. The organic fiber cloth provides
greater resilience to expansion than the glass cloth. This is important on large
equipment operating at elevated temperatures. This FRP type system is less
subject to poor weather conditions during application and has fewer compo-
nents to mix.
For insulated vertical vessels, install a 4.8 m m (3/16 in.) thick carbon steel
angle continuously seal welded around the top of the vessel. The horizontal
leg of the angle must be 12.7 mm (I/2 in.) greater than the insulation thick-
ness. This allows the side wall insulation and jacket to slip up under 50.8 mm
(2 in.) vertical leg, thus isolating the tank top insulation system from the side
wall insulation system. This prevents moisture or spilled product from getting
into the side wall insulation, should the top insulation system fail. The water
shed angle can be installed on the site should the equipment manufacturer
fail to provide it.
All protrusions or penetration through the insulation, such as nozzles, sup-
port lugs, and so forth, must be sealed with a bead of good caulking com-
pound. A joint separation greater than 3.2 mm (1/8 in.) is not acceptable for
caulking. Caulking is used only where it is impractical to provide metal flash-
ing. Properly installed metal flashing will outlast the best caulking com-
pounds. Only the silicone rubber caulking remains resilient for many years
and is resistant to higher temperatures and many chemicals. The pigmented
(color) silicone rubber caulking material is specified since it provides greater
resistance to higher temperatures and ultraviolet light than the translucent
type.
Caulking must be done immediately after the insulation jacket is installed
since moisture could enter through the open seams if left uncaulked for a
period of time. Caulking must be applied on a clean and dry surface. The
bead must never be feather-edged since the life of the seal depends on a uni-
form material thickness.
Austenitic stainless steel piping and equipment where the insulation can
get wet, considered critical to production, or contain hazardous chemicals
will be painted before insulation is applied. Areas where the insulation can
KALIS ON USING SPECIFICATIONS 203
get wet are defined as outdoors, open structures, or washdown areas or are
the presence of fire sprinklers and deluge systems that are periodically tested.
For moderately hot operating temperatures, a two-coat polyamide epoxy
system is used. For higher temperatures, a two-coat heat resistant silicone
system is specified. The equipment and piping to be painted are identified in
the insulation specification. Providing this information will alert the insula-
tion supervisor that the surface must be painted before applying the insula-
tion. Since he is usually not responsible for painting, specific information on
surface preparation and painting is normally indicated in the painting specifi-
cation. Paint provides additional protection to the austenitie stainless steel
should moisture-containing chlorides enter the insulation system.
The use of insulation systems for personnel protection on hot pipe and
equipment is discouraged. Shielding the hot surfaces with expanded metal
will eliminate potential areas for ESCC.
Conclusion
The insulation specification must identify all insulated austenitic stainless
steel pipe and equipment and provide precise requirements for protection. It
should specify low wicking insulation materials, rigid aluminum jacket
weather barrier, aluminum flashing with detailed sketches, and silicone
caulking compound. For irregular shapes, such as equipment heads, use a
reinforced polyester or epoxy finish. The specification must instruct the insu-
lator to the proper application of the materials. Complying with the above
specification should reduce the possibility of ESCC of austenitic stainless steel
under insulation.
H e r b e r t A. M o a k i
Use Inspection as a Means of

Reducing Failures Caused By
Corrosion Under Wet Insulation
REFERENCE: Moak, H. A., "Use Inspection as a Means of Reducing Failures Caused By

Corrosion Under Wet Insulation," Corrosion of Metals Under Thermal Insulation. A S T M
S T P 880, W. I. Pollock and J. M. Barnhart, Eds., American Societyfor Testing and Mate-
rials, Philadelphia, 1985, pp. 204-207.
ABSTRACT:This paper is an overviewof what to look for in inspectionto prevent the prob-
lem of corrosion under wet insulation. To assure quality, do not limit inspectionof thermal
insulation systems to just the installation. Reviewthe work of the designer as criticallyas
you do that of the installer. An excellentphysical installationcannot alwayscure a poor de-
sign. Inspect on a routine basis to look for "tell-tale" signs of pending failures and follow
up with prompt repair.
KEY WORDS: inspection, corrosion, insulation, design
To many people, inspection is regarded as something to be done "after the

fact." This is u n f o r t u n a t e because m a n y p e n d i n g failures are created during
design a n d are perpetuated during installation.
There is no question about moisture being the major cause of corrosion, b u t
when it is retained a n d concealed by thermal insulation, the conditions are
much more severe. This is why inspection should start with design. Look for
places where nozzles, hangers, thermocouples, conduit, or other penetrations
will permit water entry. O n e q u i p m e n t , start at the top a n d work down. O n pip-
ing, concentrate on penetrations through the top of insulation on horizontal
lines a n d all penetrations on vertical sections.
O n vertical e q u i p m e n t , installation of a weathershed ring at the top head is
critical (Fig. 1). P r o t r u d i n g stiffener rings are expensive to flash a n d present a
continuous problem with wicking type insulations. Where possible, modify the
1Project engineer, E. 1. DuPont de Nemours and Company, Inc., Engineering Department,

Wilmington,DE 19898.
Downloaded/printed by 2O4
Copyright9 by ASTM International WW'vV.astIn.OF~
MOAK ON INSPECTION 205
FIG. 1-- Weathershed ring.
vessel design or insulation thickness or both to permit an uninterrupted exte-

rior finish (Figs. 2 and 3).
Modifications made during design review are less expensive and more effec-
tive than modifications made during or after installation. Even the best field
application cannot always cure a poor design.
Before applying insulation, make sure that all required attachments have
been installed and that where a protective coating has been specified it has been
applied over a properly prepared surface.
Another critical inspection function that is frequently overlooked is to review
in detail with the supervision responsible for installation (contractor, mainte-
FIG. 2--Typical installation at stiffener rings.
FIG. 3--Increased insulation thickness to provide uninterrupted finish.
nance, and so forth) exactly how the work is to be done and the minimum qual-
ity level that will be accepted. Do not assume anything. An open discussion can
help to avoid later misunderstandings and possible rework.
During installation,
9 Make sure installation is dry.
9 If insulation is temporarily covered with plastic, be sure it is vented to pre-
vent condensation.
9 Use close fitting (6-mm [1/8 in.] maximum opening) metal flashing at all
openings.
9 Apply silicone caulking at all points of possible moisture entry. Apply a
minimum bead of 6-mm (n/4 in.) and do not "feather" edges. "Feathered"
edges curl and pull away from the metal.
9 Apply all jacketing so that all overlaps shed water.
Inspections following installation must be made frequently on a routine
basis by a qualified person. Damage that may seem insignificant to the un-
trained eye can have serious potential. As an example, a superficial look at
caulking or mastic may show that it is in place, however, a closer examination
will reveal cracked mastic, exposed reinforcing, and caulking that has pulled
away from metal surfaces.
Any gouge or puncture of the finish system can permit water entry. Openings
where insulation has been removed for maintenance or modification or both
presents the same problem.
Other inspection techniques, such as infrared or ultrasound, can detect "hot
MOAK ON INSPECTION 207
spots" or a change in metal thickness, but they do not detect points of water en-
try and are not a substitute for visual inspection.
Inspection without proper and prompt correction is wasted effort because
just knowing there is a problem is not enough.
Conclusion
In summary, inspection is an excellent tool but only when used properly.
This means having a qualified owner's representative to
9 Review design.
9 Inspect daily during construction (installation).
9 Periodically and routinely inspect for damage after installation.
9 Assure that necessary repairs are made promptly and properly.
This effort is the best insurance against metal corrosion and is the most cost
effective.
Test Methods
Francis B. Hutto, Jr., 1 Ralph G. Tissot, Jr., 1
Thomas E. Whitaker 1
A New Apparatus and Test

Procedure for Running ASTM C 692
Stress Corrosion Cracking Tests
REFERENCE: Hutto, F. B., Jr., Tissot, R. G., Jr., and Whitaker, T. E., "A New Appa-
ratus and Test Procedure for Running ASTM C 692 Stress Corrosion Cracking Tests,"
Corrosion of Metals Under Thermal Insulation, A S T M STP 880, W. I. Pollock and J. M.
Barnhart, Eds., American Society for Testing and Materials, Philadelphia, 1985, pp.
211-219.
ABSTRACT: A new experimental technique has been developed for running stress corro-
sion cracking tests on insulating materials. The new technique uses a steam heated pipe
for heating the coupons and uses an "upside down" dripping action rather than a "wick-
ing" action, as in ASTM Evaluating the Influence of Wicking-Type Thermal Insulations
on the Stress Corrosion Cracking Tendency of Austenitic Stainless Steel (C 692). In addi-
tion to providing a more realistic approach, the new procedure correlates well with the
old, simplifies sample coupon preparation and gives added reliability to the test results.
KEY WORDS: stress corrosion, insulation, test apparatus, external stress corrosion
cracking
ASTM Evaluating the Influence of Wicking-Type Thermal Insulations on

the Stress Corrosion Cracking Tendency of Austenitic Stainless Steel (C 692)
has long served as the basic test method for evaluating the influence of wick-
ing-type thermal insulations on the stress corrosion cracking tendency of aus-
tenitic stainless steel. The test method itself was a direct outgrowth of the
work done at du Pont by Dana and Delong, initially reported in 1956 [ 1]. This
test method and other similar procedures are generally referred to as the
"Dana Test."
Stated simply, the Dana test calls for the preparation of a precisely stressed
U-bend, which is fitted into a block of insulation, which, in turn, is partially
IDirector of research and development, senior research technician, and senior research engi-
neer, respectively, Pabco Insulation Division, Louisiana-Pacific Corporation, Fruita, CO 81521.
immersed in a dish of distilled water. During the course of the test, the U-
bend is heated to near the boiling point, such that water is wicked through the
insulation to the U-bend surface where it evaporates. If chlorides are present
in the insulation material, they tend to concentrate on the metal surface
where external stress corrosion cracking (ESCC) can occur. Based on a large
body of data developed by some investigators, it was determined that if ESCC
did not take place in 28 days, the material under testing could be considered
to be benign insofar as ESCC was concerned in the field.
It was recognized early that even though chlorides contained in insulation
materials could contribute to ESCC, chlorides introduced from outside
sources were an even more important factor to consider. Ashbaugh reported
in 1965 [2] that the "frequency of ESCC, as reported, appears to be higher in
coastal locations."
A companion test, generally referred to as the "accelerated Dana test,"
calls for the use of 1500-mg/L chloride solution in place of the distilled water
and running the test only 6 days rather than 28 days. This accelerated test was
allowed as an alternate pre-production test in Research, Development, Test,
and Evaluation (Navy) (RDT) Test Requirements for Thermal Insulating for
Use on Austenitic Stainless Steel (M12-1T), which has been abandoned in
favor of ASTM C 692.
The accelerated Dana test was testing more than the regular 28-day test. By
introducing additional chloride into the system, this test was evaluating the
ability of the insulation to inhibit the action of chlorides introduced from out-
side sources. Several organizations, including du Pont, Hercules, and On-
tario Hydro, have recognized this by requiring that an insulating material be
capable of passing the accelerated Dana test to qualify for use over austenitic
stainless steel. The addition of the accelerated Dana test as an alternate test
procedure is currently under ballot in ASTM Committee C 16 on Thermal
Insulation.
The objective of this project was to develop an improved test procedure that
reduces or eliminates uncontrollable variables in the existing ASTM C 692
test method.
Discussion
Problems with A S T M C 692

In the course of updating ASTM C 692-77, many problems with the current
procedure have come to light, particularly as related to the proposed alternate
test method utilizing chloride solution rather than distilled water.
Problems with Temperature--ASTM C 692-77 calls for the use of four test
coupons, resistance-heated in electrical series with a fifth " d u m m y " coupon
to which is attached a thermocouple for monitoring the uniformity of the cou-
pon temperature control. After establishing, by means of a surface pyrometer
that the test coupons are at the proper test temperature (test coupons run at
HUTFO ET AL ON ASTM C 692 TESTS 213
lower temperatures than the dummy, which is in air and has no liquid to wick
up), the temperature is controlled by controlling the established offset of the
dummy coupon. When each test coupon is instrumented with an attached
thermocouple, it becomes clear that the temperature uniformity require-
ments of ASTM C 692-77 are very difficult to meet (Table 1).
Problems with Evaporation R a t e - - W a t e r evaporation is impacted by many
factors including coupon temperatures, coupon fit to insulation block, tight-
ness of enclosure around test coupon, temperature of test environment, rela-
tive humidity of test environment, liquid level in test system, and local boiling
point for water. ASTM C 692-77 is in conflict with Military Specification In-
sulation Material with Special Corrosion, Chloride, and Fluoride Require-
ments (MIL-I-24244B), which allows the use of C 692-77 test procedures
"with certain restrictions." The evaporation rate specified in the MIL-I is not
allowed in ASTM C 692-77! With the accelerated Dana test, using 1S00-mg/
L chloride solution, one might expose different coupons to different amounts
of chloride, depending on the evaporation rate and whether makeup is with
chloride solution or distilled water.
Electrical Effects--There has been some conjecture about the possible ef-
fects of utilizing an electric current for resistance-heating of the coupons.
Testing Difficulty with Certain M a t e r i a l s - - A s presently constituted,
ASTM C 692-77 addresses itself only to wicking-type insulations. If the insu-
lation will not wick, it cannot be tested. There is also some question as to
whether or not wicking is a reasonable test mode, since there would be rela-
tively few "real life" situations where the insulation would be partially im-
mersed in a "wicking" mode. Insulations with a water-repellent treatment
cannot be tested "as received" but might well wick after exposure to service
temperatures. ASTM C 692-77 makes no provision for broadening the appli-
cability of the test by such means.
Searching f o r a Better Way

In reviewing other test methods and the literature, several observations
were made: (1) the hot steam pipe idea of MIL-I-24244B would appear to be a
TABLE 1--Coupon to coupon temperature variation

(typical data).
ASTM C 692, ~ (~ Pipe Tester, ~ (~
96 (205) 97 (206)
111 (231) 98 (208)
108 (226) 98 (208)
94 (201) 96 (204)
103 (217) 95 (207)
157 (315)" 98 (208)
"Dummy.
better route to temperature uniformity and (2) the dripping approach of

Ashbaugh [2] appeared to be a more realistic approach than wicking, per-
haps even applicable to nonwicking materials.
The first step in developing a new procedure was to try and duplicate the
work of Ashbaugh, using a long, steam-heated pipe that would allow the test-
ing of several coupons at once. The U-bends were tested "upside down" with
the insulation on top as shown in Fig. 1. While the coupons could be stress
corrosion cracked in this manner, it proved to be next to impossible to devise
a "dripper" that would deliver of the order of one drop per minute without
slowing down with time. After extensive experimentation, it was found that a
multichannel peristaltic p u m p was the answer. After making several modifi-
cations to the steam boiler approach, the test pipe that evolved is shown in
Fig. 2.
Since the available peristaltic p u m p would accommodate 20 channels, the
steam pipe was made long enough to accommodate 20 coupons at a time.
Developing a Test Procedure

Having established that the peristaltic p u m p could deliver 20 channels with
an accuracy of +_ 10%, we set out to determine what drip rate would match
the performance of accelerated ASTM C 692-77. After many tries, it was es-
tablished that a rate of 250 m L / d a y was the right amount. The apparatus was
set up such that each channel of the peristaltic p u m p had its own graduated
1000-mL reservoir. Thus, as a test proceeded, each channel could be individ-
ually monitored to see that the amounts being delivered were within the pre-
scribed limits.
Dripper
6
t Insulationtest
I
block
C
I ~
~'-
--
Heatedpipe
lest couporl
FIG. l - - Test sample configuration.
HUFFO ET AL ON ASTM C 692 TESTS 215
FIll valve L
45"(1050mm)
~-Rellef valve F
1.5" IPSpipe
P~essuregdge
Insulatlon
.i 5 liter bl)ller
] l ] h ~ gldhS
valve
1 JJle ['1~1~)i ~jU I a t u r

Ileal In U elumelfl
]llblJlo[ IOrL blt~( k
P
I l a b l e Lup
I
FIG. 2--Modified ASTM C 692 apparatus.
It was also judged that a more uniform start would be given the various
samples if all were presoaked for 1 h in 1500-mg/L chloride solution by par-
tial immersion m an evaporating dish as shown in Fig. 3.
The following test procedure was established:
1. Prepare test coupons and insulation samples as prescribed in C 692-77.
Reserve positions 1 and 20 for "blanks" where the corroding liquid is allowed
to drip directly on the coupons. If distilled water is being used, the blank
coupons must not crack in order to validate the test. If 1500-mg/L chloride is
being used, the blank coupons must crack in order to validate the test.
2. Place coupons side by side on the heating pipe with a dab of heat trans-
fer grease under each. The second and nineteenth coupons must be instru-
mented by placing a 28 American Wire Gage (AWG) (or finer) thermocouple
under the coupon in the heat transfer grease. The test temperature shall be
the average of the two thermocouples.
3. Stress coupons in accordance with ASTM C 692-77. Rather than drill
holes and stress with through-bolts, it was found that ordinary C-clamps
could do an equivalent job with much less effort and extraneous damage to
the coupon.
~Evaporating
dish
~ Insulation
test block
FIG. 3--Presoaking test sample.

i Waterlevel
216 CORROSIONOF ME'I"ALS
4. Carefully fit each insulation sample to its coupon by sanding the insula-
tion where necessary. Mark both coupons and samples to match, using a per-
manent coupon marking system for future reference.
5. Presoak each insulation sample in the test liquid by partial immersion
for 1 h as shown in Fig. 3.
6. Turn on pipe heater and bring to temperature (boiling point + 6 -- 3~
[ + 10 - - 5~ while the insulation samples are soaking.
7. Place the soaked insulation samples on their respective coupons and
start peristaltic pump.
8. Note starting time (start of pump).
9. With the peristaltic pump, set to deliver 250 mL/day, monitor closely to
be sure that all deliveries are within ___10% of this prescribed amount. Posi-
tive data points (cracking occurred) where the fluid delivery rate was more
than 10% high must be discarded. Negative data points where the fluid deliv-
ery rate was more than 10% low must also be discarded. Generally lower rates
of flow or no flow may be attributed to crimps or bends of the delivery tubes
within the peristaltic cassettes.
10. Monitor temperature as measured by thermocouples on Coupons
2 and 19 and keep the average of the two thermocouples within + 6 -- 3~
( + 10 --5~ of the boiling point.
11. At the completion of the test period, remove and clean all coupons in
accordance with ASTM C 692-77.
12. Examine at X30 for stress corrosion cracks in accordance with ASTM
C 692-77. If in doubt, bend the coupons on a smaller radius over the suspect
area. Real cracks will open up.
V a l i d a t i o n of T e s t P r o c e d u r e
Extensive testing was carried out with two calcium silicate materials, one
(Calsil A) which routinely passes the accelerated Dana test, and one (Calsil B)
which routinely fails the accelerated Dana test. Both materials routinely pass
the regular Dana test, as do most other insulation materials.
When 10 samples of Calsil A and 10 samples of Calsil B were tested simul-
taneously for six days using 1500-mg/L chloride solution, 10 samples of Calsil
A passed (did not crack) while 10 samples of Calsil B failed (cracked). When
Calsil A and Calsil B were tested for 28 days with distilled water, both passed.
Test Data
Several materials were tested on both the 28-day and the 6-day tests, and all
results were basically in agreement with ASTM C 692-77 (Tables 2 and 3).
HUTTO ET AL ON ASTM C 692 TESTS 217
TABLE 2-- Test data on various materials run six days at 250 mL/day of
1500-mg/L chloride solution.
Average
Temperature,
Test Material ~ (~ Result
1 CalsilA 99 (210) 10 samples, no cracks

Calsil B 10 samples, all cracked
2 CalsilA 107 (225) 2 samples, no cracks
Calsil B 2 samples, both cracked
Calsil C 2 samples, no cracks
Calsil D 2 samples, both cracked
Perlite A 2 samples, both cracked
Perlite B 2 samples, no cracks
blank 2 samples, both cracked
3 Cellular glass A 96 (205) 2 samples, both cracked
Perlite A 2 samples, both cracked
Perlite A heat treated at 260~ (500~ 2 samples, both cracked
Perlite B 2 samples, no cracks
Perlite B heat treated at 260~ (500~ 2 samples, no cracks
Fiberglass A 2 samples, both cracked
Calsil C 2 samples, no cracks
4 CalsilC 99 (210) 2 samples, no cracks
Calsil C heat treated at 260~ (500~ 2 samples, no cracks
Calsil B heat treated at 260~ (500~ 2 samples, both cracked
Noncritical Nature o f Coupon Preparation
A S T M C 697-77 would m a k e it appear that coupon p r e p a r a t i o n a n d stress-

ing are extremely critical to the reproducibility of the test. To check this
point, Calsil A a n d Calsil B were r u n side by side with various surface prepa-
rations a n d stressings as shown in T a b l e 3. Based on these data, coupon sur-
face p r e p a r a t i o n a n d the degree of post-stressing are very noncritical items.
As might be expected, the b e n d i n g process appears to have provided all of the
effective stress necessary to m a k e the test m e t h o d work. Additional post-
stressing a p p e a r s t o be unnecessary. R o u n d - r o b i n testing should quickly es-
tablish that less stringent surface p r e p a r a t i o n is necessary.
A r g o n Sensitization
A 28-day test was r u n c o m p a r i n g argon-sensitized coupons with unsen-

sitized coupons. T h e r e a p p e a r e d to be no obvious difference, in t h a t all cou-
pons passed (Table 4). T a b l e 5 shows six-day test results. W i t h a d d i t i o n a l
data, this r e q u i r e m e n t m i g h t also be modified.
TABLE 3 - - E f f e c t of surface treatment on E S C C

(six-day test at IO0oC) (212~ average temperature).
Material Surface Treatment Result
Calsil A regular C 692 preparation no cracks

Calsil B cracked
Calsil A bent first, sanded afterward no cracks
Calsil B cracked
Calsil A hand sanded, 80-grit sandpaper no cracks
Calsil B cracked
Calsil A hand cleaned with commercial stainless steel no cracks
polish
Calsil B cracked
Calsil A hand cleaned with Ajax | cleanser no cracks
Calsil B cracked
Calsil A not cleaned at all, only degreased no cracks
Calsil B cracked
Calsil A hand cleaned with number 2 coarse steel wool no cracks
Calsil B cracked
Calsil A belt sanded using a worn 80-grit belt no cracks
Calsil B cracked
Calsil A ASTM C 692 preparation, double stressed no cracks
Calsil B cracked
Calsil A ASTM C 692 preparation, no stress beyond no cracks
original bend
Calsil B cracked
TABLE 4--Effect o f argon sensitization on E S C C (28-day test

at 250-mL distilled water~day with coupons at 98~ [208~
Coupon Sample Result
Standard Calsil A no cracks

Standard Calsil A heat treated at 260~ (600~ no cracks
Argon-sensitized Calsil A no cracks
Argon-sensitized Calsil A heat treated at 260~ (600~ no cracks
Standard Calsil B no cracks
Standard Calsil B heat treated at 260~ (600~ no cracks
Argon-sensitized Calsil B no cracks
Argon-sensitized Calsil B heat treated at 260~ (600~ no cracks
TABLE 5--Six-day test using argon-sensitized coupons.
Sample Results
Catsil A no cracks (5 of 5)
Calsil B all cracked (5 of 5)
Calsil D all cracked (5 of 5)
Blank all cracked (2 of 2)
HUTTO ET AL ON ASTM C 692 TESTS 219
Conclusions
As compared with ASTM C 692-77, the proposed test method:
(1) has better temperature control,
(2) requires less operator attention,
(3) requires less critical sample preparation,
(4) is a more realistic approach to the mode of contamination, and
(5) is more broadly applicable to different types of insulating materials.
References
[1] Dana, A. W., Delong, W. B., Corrosion, Vol. 12, No. 7, July 1956, pp. 19-20.
[2] Ashbaugh, W. G., "ESCC of Stainless Steel Under Thermal Insulation," Materials Protec-
tion, May 1965, pp. 18-23.
Keith G. Sheppard, 1 S u n i l Patel, 1M u k e s h Taneja, 1 a n d
R o l f Weil 1
Comparisons of Several Accelerated

Corrosiveness Test Methods for
Thermal Insulating Materials
REFERENCE: Sheppard, K. G., Patel, S., Taneja, M., and Weil, R., "Comparisons of
Several Aeeelerated Corrosiveness Test Methods for Thermal Insulating Materials," Corro-
sion of Metals Under Thermal Insulation. A S T M STP 880, W. I. Pollock and J. M. Barn-
hart, Eds., American Society for Testing and Materials, Philadelphia, 1985, pp. 220-230.
ABSTRACT: A study was conducted to provide data on which to base a corrosiveness test
for the thermal insulating materials used in residential structures. Several possible test
methods were compared. The materials tested included celluloses containing several dif-
ferent fire-retardant additives, glass fiber, rock wool, and a urea formaldehyde foam. Be-
cause of their widely differing physical properties, testing was conducted in water leachants
made from the insulations. In addition, a test was performed that simulated the condensa-
tion conditions that might occur in a residence. It was found that two leachant-based meth-
ods could be suitable for accelerated corrosiveness testing of thermal insulation. One
method involved determining the corrosion rate of metal coupons immersed in leachant for
14 days at 45~ The other test was cyclic potentiodynamic voltammetry, which can be
completed in only a few hours.
KEY WORDS. thermal insulation, corrosion, corrosion tests, corrosiveness, electrode

potentials, anodic polarization, pitting, cellulose, glass fiber, mineral wool, urea formalde-
hyde foam
The purpose of this study was to provide data upon which to base an acceler-
ated corrosiveness test applicable to different thermal insulation materials,
particularly those used in residential structures. The Corrosion Task Group of
ASTM Subcommittee C16.31 on Chemical and Physical Properties has been
charged with the development of such a test procedure. At the present time,
ASTM Specification for Cellulosic Fiber (Wood-Base) Loose-Fill Thermal In-
IAssistant professor, graduate assistant, graduate assistant, and professor, respectively, De-
partment of Materials and Metallurgical Engineering, Stevens Institute of Technology, Hoboken,
NJ 07030.
Copyright9 by ASTM International Www.astIn .OI'~
SHEPPARD ON ACCELERATED CORROSIVENESS METHODS 221
sulation (C 739) includes a corrosion test. ASTM C 739 is for loose-fill cellulo-
sic insulation, and the corrosion test procedure is not applicable to other types
of insulation. The only other relevant ASTM standard for thermal insulation
that includes corrosion testing is ASTM Evaluating the Influence of Wicking-
Type Thermal Insulations on the Stress Cracking Tendency of Austenitic
Stainless Steel (C 692), but this covers a very specific corrosion condition. The
Corrosion Task Group is impeded in its consideration of a new uniform corro-
siveness test for thermal insulation by the severe lack of published laboratory
and field performance data on the corrosiveness of these materials. The study
reported here attempts to provide some of the data necessary to enable rational
decisions to be made on possible test procedures.
There are several desired characteristics of a new corrosiveness test. It should
be quantitative and be applicable to various insulating materials with their dif-
fering physical and chemical properties. It should provide rapid results, so that
the test can be applied to quality control during manufacture. The test should
be reproducible, and ideally from its results, the behavior of the insulating ma-
terial under typical field conditions as well as under "worse-case" conditions
should be predictable. The ability to predict performance under field conditions
is perhaps the area of most controversy in corrosion test development. It is likely
to be the least attainable of the desired characteristics because of the many fac-
tors other than the inherent corrosiveness of the insulation itself that influence
corrosion in the field.
A prerequisite for corrosion is the presence of moisture. Moisture can be pro-
vided by condensation, leakage, for example, from a faulty roof, or by absorp-
tion caused by hygroscopic components of the insulation material. As an exam-
ple of the latter, Anderson and Wilkes [1] found that in a moisture absorption
test (ASTM C 739), cellulosic insulation containing sulfate fire-retardent addi-
tives showed as much as a 40% weight gain in 24 h when exposed to an atmo-
sphere at 90% relative humidity. Weight gains of up to 76% were recorded with
longer exposures. The presence of any type of insulation may produce conden-
sation conditions in a wall cavity [2], particularly in the absence of an effective
vapor barrier. Thus, the potential for corrosion of metal components may be
created solely by the physical presence of insulation and its influence on thermal
gradients. Some characteristics of the insulation that may also influence corro-
sion are the already mentioned moisture content, the permeability of the insula-
tion to moisture, and the nature of the physical contact between insulation and
any metal components, particularly the presence of occluded areas that have be-
come anodic because of oxygen deprivation.
Another major factor that can influence corrosion is the presence in an insu-
lation material of chemicals such as fire-retardant additives, binders, and so
forth. These ingredients can affect the corrosion rate when conditions exist that
allow corrosion to occur. The physical distribution of chemicals within the in-
sulation and the stability of those chemicals over long time periods are also im-
portant considerations. In addition to the characteristics of the insulation ma-
terials themselves, the environmental variables of t e m p e r a t u r e , air-flow and

humidity, a n d their variation with time may influence corrosion.
In order to predict corrosion performance in the field and therefore take into
account the m a n y variables outlined above, it a p p e a r s necessary to develop a
model containing the various chemical, physical, and environmental p a r a m e -
ters. T h e corrosiveness of the insulation material as d e t e r m i n e d by any new lab-
oratory test would be b u t one of the factors in such a model. At present there are
insufficient d a t a to develop a suitable model. In particular, field-performance
d a t a are virtually nonexistent. T h e few published field studies [2,3] containing
d a t a on corrosion were not specifically designed to investigate corrosion. The
investigators did not look for metal c o m p o n e n t s in the buildings u n d e r study;
in fact, generally such locations were avoided. T h e availability of reliable field
d a t a is limited by the cost a n d difficulty associated with gaining access to metal
c o m p o n e n t s in the walls of houses and by problems with obtaining a d e q u a t e
environmental histories at each location. A more fruitful a p p r o a c h is likely to
be the simulation of field conditions using a test wall where full control can be
exercised over a wide range of conditions.
In view of the above considerations, it is felt t h a t at present, a new corrosive-
ness test should be limited to assessing the relative aggressive or inhibiting ef-
fects of the various ingredients of the insulating materials once conditions fav-
orable for corrosion are created. The possible test m e t h o d s suitable for a new
corrosiveness test are listed in Table 1. Some advantages and disadvantages of
TABLE 1--Possible corrosivenesstest methods.
1. Visual observation
1.1 pitting (pass-fail criterion)
1.2 general corrosion (rating)
Advantages: no sophisticated equipment needed--easily performed
Disadvantages: slow, often poorly reproducible, subject to judgement
2. Coupon weight loss
Advantages: same as 1
Disadvantages: poor indication of pitting, slow
3. Electrical resistance probes
(measures resistance change caused by cross-sectional area reduction)
Advantages: can be used with "dry" insulation
Disadvantages: poor indication of pitting
4. Linear polarization
(determines corrosion current (rate) by application of small potential change and extrapolation
of assumed linear potential-log current relationship)
Advantages: rapid quantitative measurement of general corrosion rate as function of exposure
time; fairly simple equipment requirements
Disadvantages: requires "wet" insulation or leachants, poor pitting indication
5. Voltammetry
(application of increasing and then decreasing anodic potential while current is monitored)
Advantages: rapid, some pitting indication as well as general corrosion
Disadvantages: requires "wet" insulation or leachants; relatively sophisticated equipment re-
quirements
each m e t h o d are m e n t i o n e d . T h e techniques were discussed in greater detail in

a previous publication [4].
Experimental Procedure
The t h e r m a l insulation materials studied in this p r o g r a m are listed in Table 2.
The choice of those insulation materials to be tested resulted from an initiative
by the A d Hoc Cellulose I n d u s t r y - G o v e r n m e n t Agencies G r o u p . This group
decided to supply a n u m b e r of cellulosic insulation batches to be used in various
test programs. These b a t c h e s were therefore included in the present study to-
gether with three b a t c h e s of glass-fiber insulation, one of mineral wool, a n d one
of urea f o r m a l d e h y d e (UF) foam. T h e cellulose b a t c h e s chosen by the A d Hoc
G r o u p all contained b o r a x as an additive. As borax is known to be a corrosion
inhibitor, it was d e c i d e d to include an additional cellulose b a t c h (593) contain-
ing no borax a n d known to be relatively corrosive from previous testing [1]. Un-
fortunately, because of the limited quantity of Cellulose 593 available, we were
not able to use this material in all the tests.
The test m e t h o d s (refer to Table 1) c o m p a r e d in this study included weight
loss and a p p e a r a n c e of steel coupons. The corrosion rate of steel coupons was
also d e t e r m i n e d by the polarization-resistance m e t h o d . These tests will be re-
ferred to as the " l e a c h a n t - c o u p o n " tests. In addition, values of the limiting cur-
rent density together with observations of the presence or absence of a positive
hysteresis loop were d e t e r m i n e d by p o t e n t i o d y n a m i c cyclic voltammetry (Fig. 1)
on steel. This testing will be referred to as voitammetry. It should be noted t h a t
limiting current density d e t e r m i n e d by voltammetry (that is, the value at which
the current limits as the anodic potential a p p l i e d to the specimen is increased),
is an arbitrary measure in t h a t its value is subject to mass t r a n s p o r t conditions.
TABLE 2--hlsulating materials used in test program.
Insulation Chemical Additive

Cellulose 1 1 part borax (S mole); 2 parts boric acid, 25% chemical content
Cellulose 2 2 parts borax; 1 part boric acid, 25% chemical content
Cellulose3 1 part borax; 1 part boric acid; 1 part aluminum trihydrate, 25%
chemical content
Cellulose4 1 part borax; 4 parts ammonium sulfate, 30% chemical content
Cellulose 6 2 parts borax; 2 parts boric acid; 1 part aluminum sulfate, 25%
chemical content
Glass Fiber A no binder
Glass Fiber B with binder
Glass Fiber C with binder"
Rock wool
Urea-formaldehyde foam
Cellulose 593 aluminum sulfate, 17% chemical content
"Similar density and fiber size to B, but different binder.
However, it was felt that the limiting current density provided a means for mak-
ing a relatively simple assessment of corrosiveness under standardized condi-
tions, and it had been found in earlier work [4] to correlate well, at least for cel-
lulosic materials, to the relative corrosiveness determined by embedding metal
coupons in wet insulation. Figure 1 shows examples of typical voltammetry
curves obtained in our earlier work with cellulosic insulation containing differ-
ent fire-retardant additives [4]. The limiting current density is indicated on
Curve a. Curve c shows the positive hysteresis loop on reversing the anodic po-
tential scan that is typically obtained for pitting-type corrosion.
The test methods outlined above were compared using steel specimens im-
mersed in leachants (that is, water extracts of soluble components) made from
the different insulation samples. The reason for using leachants rather than the
insulation materials themselves is that some types of insulation are relatively
hydrophobic, and it is difficult to achieve the conductive path between elec-
trodes necessary for application of the electrochemical testing methods. Pre-
vious tests [4] had shown that leachants from cellulosic insulation produced
values of relative corrosiveness similar to those obtained using the wet insula-
tion materials. It was desirable to use an existing standard leaching method if
possible. Therefore the leaching procedure used in this study was essentially
that specified in ASTM Chemical Analysis of Thermal Insulation Materials for
Leachable, Chloride, Silicate, and Sodium Ions (C 871), however the amount
of each type of insulation leached was based on the ratios of the applied densi-
ties of the materials obtained from manufacturers' data. This approach was
considered more appropriate than using equal weights, because of the widely
differing densities of the materials being compared. The weight of each mate-
(o) -- - - C e l l u l O S e A ( . S u l p h a t e - $(itel
(b) _ _ - C e l l u l o S e Borax/Boric acid - s i e e (
LIGHT CORROSION
NO PITTING
J,1LIGNTCOR"OSION
(C) - - Cellulose Borax~Boric acid-copper
1.0
A WITH PITTING J ii
UJ
U~
tl
/ / L~mlt~ng
/ / Current
Oenslt y
0 / 9
I HEAVYCORROSION~ / /
NO PITTING / /
/
- Corrosion Potenh al
- 1.010 "e I I I I
10 ~ l0 ~ 10' 10'
LOG CURRENT DENSITY ( mA-cm" )
FIG. l--Typical cyclic voltammetry curves obtained in insulation leachants.
rial leached was equivalent to that required to fill the 102 by 102 by 102 mm
(4 by 4 by 4 in.) volume used in the condensation test (to be described later).
For the leachant-coupon tests, the Society of Automotive Engineers (SAE)
1010 (Unified Numbering System [UNS] 10100) coupons 51 by 51 by 0.076 mm
(2 by 2 by 0.003 in.) were prepared and cleaned according to the ASTM C 739
procedure. Two coupons were fully immersed in leachant contained in a poly-
propylene dish with a fitted lid. Duplicate dishes were prepared for each insula-
tion batch. Duplicate blanks were prepared using distilled water in place of a
leachant. The leachant level was marked on the outside of each dish, and the
dishes maintained in an environmental chamber at 45~ and 95% relative hu-
midity for 14 days. The dishes were inspected regularly, and distilled water used
where necessary to maintain the liquid level in each dish. At the completion of
the test, the coupons were visually assessed for corrosion and then cleaned ac-
cording to the ASTM Recommended Practice for Preparing, Cleaning, and
Evaluating Corrosion Test Specimens (G 1) and their corrosion rates deter-
mined from weight loss measurements. Corrosion rate measurements obtained
by the polarization-resistance method were carried out using a similar exper-
imental arrangement except that three coupons were used in each dish. In-
sulated wires were attached to each coupon, and then the soldered joints were
lacquered. The three coupons became the anode, cathode, and reference elec-
trodes when immersed in leachant and were polarized using a Petrolite M-1010
instrument (Rohrbach Industries). This instrument operates on the polariza-
tion-resistance principle and gives a direct readout of corrosion rate. Corrosion
rate readings were monitored throughout the 14-day test period.
Specimens for voltammetry consisted of SAE 1010 steel mounted in epoxy
resin and abraded to a 600-grit finish. The area of steel exposed was 1 cm 2. Each
specimen was carefully inspected to ensure that no crevice existed between the
steel and its epoxy mount. Cyclic voltammetry curves were obtained using a
Princeton Applied Research Model 175 Programmer controlling a Model 173
Potentiostat such that the specimen potential was scanned anodically at 1 mV/s
in stagnant deaerated leachant. The cell current necessary to maintain the ap-
plied potential was plotted on a logarithmic scale versus the applied potential.
The scan was reversed at 1.0 V versus a saturated calomel reference electrode.
Specimens were allowed to establish a constant corrosion potential in the leach-
ant before commencement of the anodic scan. From the recorded curves,
values of corrosion potential and limiting current density were obtained, and the
presence or absence of a positive hysteresis loop, indicative of pitting, was noted.
A microscopic evaluation of each specimen was made after the test. As with the
leachant-coupon test, distilled water was used to run a blank experiment.
In addition to the leachant-based tests described above, the same batches of
insulation were tested by placing them dry into 102- by 102- by 102-mm (4- by 4-
by 4-in.) compartments within a Plexiglas | box, the bottom of which was a lac-
quered, water-cooled copper plate. This allowed a 10~ temperature gradient
to be established through the insulation specimens when the box was located in
an environmental chamber held at 45~ and 75% relative humidity for 65

days. SAE 1010 steel coupons prepared as previously described were embedded
horizontally at one-quarter, half, and three-quarters of the depth of each in-
sulation specimen. The dew point would have occurred at approximately half
the depth of the insulation under the conditions described with a 10~ tem-
perature gradient, creating the probability of condensation in the lower half of
each insulation specimen. On completion of the test, corrosion of the steel was
assessed by visual observation and by measurement of coupon weight loss after
cleaning by the ASTM G 1 method. This test will be referred to as the "conden-
sation" test. The same experiment was also carried out with no temperature
gradient applied through the insulation.
Results and Discussion
The test involving steel coupons embedded in initially dry insulation and ex-
posed for 65 days at 45~ and 75% relative humidity without a temperature
gradient through the insulation, resulted in negligible corrosion of the steel. No
corrosion was obtained greater than 0.03 m m per year as determined by coupon
weight loss. This result suggests that in the absence of moisture from conden-
sation or a leak, there is little likelihood of corrosion occurring in the presence
of the batches of insulation tested in the present program. In contrast, the re-
suits obtained when a 10~ temperature gradient was set up through the insu-
lation are recorded in Table 3. Corrosion is measured by coupon weight loss
and by visual assessment. Only data for the bottom coupon in each insulation
specimen are given, that is, the coupon nearest the cooled plate and therefore
within the condensation zone. These bottom coupons experienced significantly
greater corrosion, in specimens where corrosion occurred, than the two other
coupons in each insulation specimen. Corrosion rates for the upper coupons,
that is, located above the condensation zone, were all less than 0.03 m m per
year. Coupons located at the mid-depth of the insulation exhibited corrosion
rates intermediate to those obtained from the upper and lower coupons for all
insulation specimens. The data in Table 3 indicate that the coupons in glass
fiber and rock wool insulation exhibited little corrosion. Of the cellulosic insu-
lations, Celluloses 1 and 2 containing solely boron fire-retardant additives,
produced the least corrosion. Celluloses 4, 6, and 593, containing sulfate fire-
retardants, produced relatively severe corrosion.
Data from the leachant-coupon and voltammetry tests are also given in
Table 3 to allow comparison. Values of leachant pH are recorded. Only the
average corrosion rates for the 14-day test period are given, although corrosion
rates were recorded on a daily basis by the polarization resistance method. The
presence of a positive hysteresis loop in the voltammetry curves is indicated.
The visual appearance ratings of corrosion in each test are tabulated; the rating
scheme is in Footnote b of Table 3.
Table 4 shows the results of the various tests summarized in terms of corro-
TABLE 3--Corrosiveness test results./br steel.
Type of Insulation o'J

"1"
m
"u
Cellulosic Glass Fiber ..Q
Rock Distilled
1 2 3 4 6 593 A B C Wool UF Foam Water uJ
Corrosion Rates O
Z
Condensation test from 0.08 0.08 0.13 >0.2 0.1 >0.2 0.03 <0.03 <0.03 < 0.03 0.08
weight loss, m m / y O
Leachant coupon tests" m
Polarization resistance, 0.03 <0.03 <0.03 0.11 0.06 ... <0.03 <0.03 <0.03 <0.03 0.05 0.03 r-
m
mm/y
From weight loss, mm/y 0.03 <0.03 <0.03 0.2 0.10 ... <0.03 0.03 <0.03 0.03 0.08 0.03 -=4
Leachant voltammetry m
Limiting current density, 0.04 0.04 0.05 0.1 0.07 80 0.03 0.03 0.03 0.03 0.08 0.04

mA/cm 2 0
0
Corrosion potential (--ve), 254 156 248 665 266 659 177 642 356 157 483 149 ~o
mV versus SCE 0
Positive hysteresis loop yes no no yes yes no yes yes yes yes yes no
<
present m
Leachant, pH 7.6 8.2 8.2 7.7 7.4 7.7 9.0 8.3 9.2 8.0 2.9 6.4 Z
Ill
Appearance b Co
Condensation test 1 2 2 3 3 3 0 0 1 1 2
Leachant coupon test 1 2 2 3 3 ... 2 2 2 3 2 "1)"
1 2 0 m
Voltammetry 1 0 0 3 1 3 2 2 2 -4
"1"
O
~Typical amounts leached into 1 L of water. 128 g/L cellulose, 25 g/L glass fiber, and 102 g/L rock wool.
b0 = no pitting, 1 = slight pitting, 2 ---- perforations, and 3 = partial or complete coupon corrosion. f,O
I'O
DO
",4
PO
tO
Oo
0
20
0
Go
6
z
"FABLE 4--Corrosiveness rating versus distilled water. 0

"11
Type of Insulation m
[.-
Cellulose Glass Fiber
Rock
1 2 3 4 6 593 A B C Wool Foam
Corrosion rates grouping

Condensation tests" A A A A A A B B B B A

Leachant coupon tests/'
Polarization resistance B B B A A ... B B B B A
Weight loss B B B A A B B B B A
Voltammetry/, B B A A A "A' B B B B A
"A represents a corrosion rate > 0.03 mm/ y and B represents a corrosion rate _< 0.03 mm/y.
hA represents corrosion rates greater than in distilled water and B represents corrosion rates equal to or less than in distilled water.
siveness compared to distilled water. The letter A was assigned where corrosion
was greater than and B when it was equal to or less than that produced by dis-
tilled water. The condensation test results cannot be compared with distilled
water in the same way. However, in order to make a comparison to the leachant
test results, an arbitrary index was assigned to the condensation test results
such that A was given to corrosion rates greater than 0.03 m m per year and B to
corrosion rates equal to or less than 0.03 m m per year.
Reasonable agreement is seen in Table 3 between average corrosion rates ob-
tained from coupon weight loss and by the polarization-resistance method for
steel in leachants of the different insulation materials. Comparing corrosion
rates obtained in the leachant tests with those obtained in the condensation
test, one also finds reasonable agreement. Glass fiber and rock wool specimens
produced very low corrosion rates in both types of tests. Celluloses 4 and 6 con-
taining sulfates were seen to be corrosive in both types of test. Celluloses 1, 2,
and 3 had higher corrosion rates in the condensation test than in the leachant
tests.
. The limiting current densities obtained from voltammetry show good corre-
lation with the corrosion rate data of the other tests, particularly the leachant-
based tests. Corrosion potentials measured in the voltammetry, however, do
not show such good correlation. Certainly, Celluloses 4 and 593, the most cor-
rosive in the other tests, had the most active corrosion potentials. Cellulose 2,
Glass Fiber A, and Rock wool had the most noble corrosion potentials and very
low corrosion rates. However, among the other insulation batches, a correla-
tion between corrosion potential and corrosion rate is not apparent. A positive
hysteresis loop was obtained in the voltammetry curves for all specimens except
Celluloses 2, 3, and 593 and the distilled water. Microscopic observations of the
steel specimens after the tests confirmed the presence of pitting corrosion in all
cases where a positive hysteresis loop was obtained.
No direct correlation between leachant pH and corrosion rate was observed.
This is in agreement with our previous findings [4] and that of others [1,5]. The
appearance ratings in Table 3 shows only fair agreement when comparing the
different test methods for a particular insulation. The visual assessment is in-
herently subjective and likely to be unreliable as a standard test. For each
insulation batch, the A and B ratings of Table 4, comparing corrosion rates to
those in distilled water, are found to agree for all the testing methods. Excep-
tions are Celluloses 1, 2, and 3 for which greater corrosion was produced in the
condensation test than in the leachant tests. Such a rating against distilled
water might provide a pass/fail assessment of corrosiveness in any new testing
standard.
Conclusion
It appears that a leachant-based test is capable of indicating the relative cor-
rosiveness of different types of thermal insulation. Corrosiveness in this sense is
limited to a measurement of the effect of water-soluble components of the insu-

lation on the corrosion rate, provided that environmental conditions exist that
are conducive to corrosion. It has been shown that the results obtained in the
leachant-based tests compare well with those obtained in a simulation of the
type of condensation conditions that might occur in service, although the con-
ditions were somewhat severe in this case. Further simulation and field data are
needed.
Coupon weight losses determined from the leachant-coupon test or the limit-
ing current density obtained from voltammetry appear to be equally effective
tests. Voltammetry has the advantage of a short test period (a few hours) and is
therefore useful for quality control. In addition, voltammetry provides an indi-
cation of the pitting tendency of the leachable chemicals.
Based on the results reported in this paper, the ASTM C16.31 Corrosion
Task Group is conducting a round-robin program to evaluate a leachant-cou-
pon type corrosiveness test for thermal insulation. It plans to conduct a round-
robin for a voltammetry-based corrosiveness test in the near future.
References
[1] Anderson, R. W. and Wilkes, P., "Survey of Cellulosic Insulation Materials," Technical
Report ERDA 77-23, Energy Research and Development Administration, Washington, DC,
1977.
[2] "A Field Study of Moisture Damage in Walls Insulated Without a Vapor Barrier," Technical
Report ORNL/Sub 78/977261/1, Oak Ridge National Laboratory, Oak Ridge, TN, May
1980.
[3] "Minesota Retrofit Insulation In-Situ Program," Technical Report HCP/W 2843-91, U.S.
Department of Energy, Washington, DC, 1978.
[4] Sheppard, K. G. and Well, R., Thermal htsulation. Materials and Systems jbr Energy Con-
servation in the '80s. STP 789, F. A. Govan. D. M. Greason, and J. D. McAIlister, Eds.,
American Society for Testing and Materials, Philadelphia, 1982, pp. 132-144.
[5] Shen, K. K., Journal of Thermal btsulation, Vol. 3, Jan. 1980, p. 190.
STP880-EB/Aug. 1985
Author Index
A K
Ashbaugh, W. G., 103-113, 121-131 Kalis, J. W., Jr., 199-203
Krisher, A. S., 153
Krystow, P. E., 145-154
B
Bailey, G., 57
J
Barnhart, J. M., editor, 5-8
Bennett, R. D., 114-120 Johnson, K. A., 58
Blackburn, P. G., 58 Johnson, W. D., 68
C L
Cohen, A., 132-142 Lang, G. E., 63-68
Collins, P. A., 155-164 Laundrie, T. F., 121-131
Crawley, P. G., 86-94 Lazer, P., III, 11-26
Long, V. C., 86-94
D
M
Delahunt, J. F., 58, 155-164
Maatsch, D. C., 155-164
Mclntyre, D., 27-41
F
Mettam, C. T., 178-187
Fitzsimmons, T., 188-198 Moak, H. A., 204-207
Moniz, B. J., 95-102
Myers, J. R., 132-142
G
Geary, D., 57
N
Geenen, P., 57
Gillett, J., 58 Nicholson, J. D., 55
It P
Hooper, R. A. E., 56 Palmer, F. H., 57
Huggett, I. G., 59 Patel, S., 220-230
Hutto, F. B., Jr., 211-219 Pollock, W. I., editor, 1-2
Copyright9 by ASTMInternational www.astm.org
R Tissot, R.G., Jr., 211-219

Turner, M. E. D., 57
Richardson, J. A., 42-59, 188-198
Ritter, M. C., 95-102
Roberts, R. G., 58
Sandberg, T., 71-85 Vegdahl, E. J., 165-177

Sheppard, K. G., 220-230
Sumbry, L. C., 165-177
T W
Tan~a, M., 220-230 Weil, R., 220-230
Taylor, D. O., 114-120 Whitaker, T. E., 211-219
STP880-EB/Aug. 1985
Subject Index
A E 398, 22
G 1,225, 226
Alkaline environments, 45
G 26, 172
Aluminum flashing installation,
ASTM Subcommittee C16.3, Corro-
201, 202, 203
sion Task Group, 220, 221,
Aluminum foil, 51, 57
230
applications and performance of,
196-198
fire risk, 194-196
in stress corrosion cracking pre- B
vention, 188 Bond strength and coatings, 159-161
Aluminum sulfate, 139 Borax, 139
American Petroleum Institute Pres- Boron, 226
sure Vessel Inspection Code British Standard for Thermal Insula-
510, 26 tion Materials (BS 3958, Part
Ammonium sulfate, 139, 140 2), 190
Apparatus and test procedure for Bromides, 99, 100
ASTM C 692, 211
modified apparatus, illustration,
215
C
Aqueous environment (see Water)
Argon sensitization, 217 Calcium silicate insulation, 17
table, 218 corrosion under, 44-45
Asbestos in insulation, 167 in high-temperature testing, 162-
ASTM Committee C-16 on Thermal 163
Insulation, 141,212 in hot service, 72
ASTM standards moisture absorption, 147
B 117, 172 properties, 67
C 192, 160 testing, 216-218
C 234, 160 water extract, 46
C 692, 29, 36 water migration in, 200
C 692-77, 211 Capillarity, 7
problems with, 212-213 Carbon steel (see Steels, carbon)
C 739, 141,220-221,225 Cathodic protection, 138-139
C 871,224 Caulking
C 962, 221 breaks in, 36
deteriorating, 88-89 before insulation, 153

silicone, 100, 202, 203, 206 cementitious, 45
Cellular glass (see Glass insulation, in cracking prevention, 190
cellular) degradation, 21
Cellulose Industry-Government Agen- illustration, 159
cies Group, Ad Hoc, 223 effectiveness, 113, 191-194
Cellulosic insulation effects, 157
corrosiveness testing, 223 on bond strength, 159-161
leachants, 224 evaluation, 157-158
moist, 139-140 failure, 20
moisture absorption, 221 flexibility, 171, 172
Cement, insulating, 168 inorganic, zinc-rich (IOZR), 155,
Chemical plants, corrosion experi- 157, 161
ences in, 86, 145-146 performance, 158
Chemical Process Industries inspec- illustration, 159
tion methods study, 152 organic, 34, 50-51
Chiller water cooling/hot water heat- advantages, 155
ing system, 137-138 permeability, 21, 34-35
Chloride ions, 7 specification, 179
stress corrosion cracking and, 8, 44 for stainless steel, 168, 177
Chlorides test programs for, 169-170
concentration, 166, 200 waterproof, 56
density of deposits, 31 (see also Specific coatings and
illustration, 30 paints)
from fire retardants, 66 Cold insulation systems, 117
in insulation, 31, 166, 178 Cold piping system, 179
in lagging, 48 Concrete
protection against, 167 bonding, 159-161
soluble, 189 coatings under insulation of, 157-
sources of, 29-32, 48, 105 164
waterborne, 166 Condensation test, 225, 229
Climate effects, 24, 149-150 Consumer Product Safety Commis-
Coal tar epoxy coatings, 179 sion, U.S., 141
Coating system for carbon/low alloy Copper corrosion products, 134, 137
steel, 186 Copper chloride, 137
table, 187 Copper sulfate, 137, 138
Coating system design, protective, illustration, 136
155 Copper sulfide, 134
for insulating equipment, 156-157 illustration, 136
Coatings, 20-22 Copper water tube, aggressive insu-
adhesiveness, 171, 172 lation in, 132
anti-abrasion, 10S, 110 Corrosion
applying, 108-110 of copper, 133
SUBJECT INDEX 235
galvanic, illustration, 18 Corrosion under lagging, European

in liquefied petroleum gas tank, meeting on, 42
114 Corrosiveness test methods, accel-
Corrosion control erated, 220
guidelines, 150-151 Cracking (see Stress corrosion crack-
illustration, 152 ing)
under thermal insulation and fire- Cracking prevention policy, 190-191
proofing, 58
Corrosion detection program, 73-74 D
Corrosion mechanisms, 131
in lagging, 56 Dana test, 211
Corrosion phenomena accelerated, 212, 216
in carbon/low alloy steels, 43-44 Dryout, 14, 15
in stainless steels, 44-45 Dye checking, 89
Corrosion prevention illustration, 92
in copper, 138-139
costs, 85
E
design, specification, inspection,
and maintenance, 54-55 Elastomer, flexible foamed, 66
designing for, 178 Elastometric sealant, 96
features, 185, 186 Elcometer adhesion test, 171, 172
steps in, 182-185 results, tables, 173-176
inhibited laggings, 48-50 Embrittlement, liquid metal, 51,
organic coatings, 50-51 191, 195, 198
Corrosion rates of steel, 127, 131 incidence of, tables, 52, 53
Corrosion Science and Technology, Energy conservation, 5, 154
Institute of, 43 Engineering Task Force Group, 117
Corrosion under insulation (CUI) Environmental cabinet test, 172
in carbon steel results, tables, 173-176
contributing factors, 11, 156 Epoxy amine primer, 157, 163-164
illustration, 150 Epoxy coatings, 50-51, 102
controlling, 145 amine-cure, 157
in chemical plants, 86 application information, tables,
critical areas, illustration, 80 170, 171
design to prevent, illustration, 82- effectiveness, 172
84 tables, 173-176
economics of, 154 manufacturer's information, 169
examples, illustration, 75-77 permeability, 34-35
extent of, 156 phenolic, 22, 108, 110-112, 123,
factors affecting, illustration, 147 128, 163-164
in petrochemical plants, 71 effectiveness, 131
reducing failures caused by, 204 Epoxy system/epoxy phenolic, table,
of steel, 121 187
Equipment design, 146, 147, 204 phenolic, table, 47 (See also Epoxy,
effects, 12 phenolic)
features, 148 (see also Polyisocyanurate foam;
Equipment painting, 156 Polyurethane foam)
Evaporation rate problems, 213 Foils
Exposure tests, carbon steel, 157 aluminum, 51, 57
applications and performance
of, 196-198
fire risk, 194-196
F
in stress corrosion cracking pre-
Fabrication of materials, 52-54 vention, 188
Failure, stainless steel stainless steel, 51-52
explanation of, 100 Freezing effects on corrosion, 13-14
investigation of cause, 96-99
prevention program, 100-102
G
Federal Standard Thermal Insulation
(Loose Fill for Pneumatic or Galvanic protection, 190-191, 194,
Poured Application): Cellu- 197
losic or Wood Fiber (HH-I- Glass insulation, cellular, 17, 45-46
515), 141 blocks, 118
Fiberglass insulation, 19, 58 in cold insulation, 179
Fire protection deluge system, 100 performance, 129
illustration, 101 properties, 66
Fire retardants source of chlorides, 105
additives, 221,226 sources of sulfuric acid, 68
chlorides from, 66 Glass-fiber insulation
compositions of, 140 corrosiveness testing, 223, 226, 229
table, 141 properties, 66
corrosivity, 139, 221,226
halogenated, 47
H
mastic, 95
Fireproofed structures, protective Halide ions, 7
coating system design for, 155 leachable, 99, 100
Fireproofing table, 102
cementitious, 43 Halogenated flame retardants, 18
corrosion under, 44, 45-48, 156 Heat flow control, 6
Fire risks with aluminum foil, 194- Heat loss, reduction, 5
196 Hot water heating/chilled water cool-
Foam glass, moisture absorption, 147 ing system, 137-138
FOAMGLAS, 68 Hot water lines, domestic, corrosion,
Foams 134-137
fire resistance, 65 Humidity, sensitivity to, 18-19 (see
organic, 47 also Moisture; Vapor; Water)
SUBJECT INDEX 237
Hydrogen sulfide, 17 corrosiveness of, 15, 19, 133, 140,

Hygroscopicity, 7 141-142, 222
definition, 5, 58
design codes, 166
failure due to improper design and
I
application, 201
ICI lagging specifications, 191, 196 function, 5
Icing in cold insulation, 181 inhibited, 8, 107-108
Inhibitors, role of, 34-37 inspection, 25, 58-59
Inspection, 25, 58-59 illustration, 26
functions of, 205-206 programs, 146, 151-153
illustration, 26 installation, 206
programs, 146, 151-153 leaching of, 147-148, 166
in reducing corrosion-induced fail- magnesia, 29
ure, 204 maintenance, 24-26, 150
techniques, 79-81, 206 illustration, 151
costs, 85 materials, 6, 168
nondestructive, 80 aggressive, 133
Insulated structures, protective coat- cellulosic, 132
ing for, 155 control of, 190
Insulating practice corrosive attack of, 133, 140,
changes in, 177 141-142
design basis, 166 natural carbonaceous granular,
installation, 168 132
materials, 167-168 recommended, table, 167
Insulation nonwicking, 200, 203
asbestos in, 167 permeability, 23
blanket, 167 for personnel protection, illustra-
calcium silicate (see Calcium sili- tion, 16, 17
cate insulation) reflective, 6
calcium silicate/magnesia, 43 selection, 15-20, 146-147, 179
cellulosic (see Cellulosic insulation; stress corrosion cracking
Glass insulation, cellular) of austenitic stainless steel un-
characteristics, 6, 16 der, 27
closed pore (CPI), 29 chloride, !65
chlorides in, 31, 166, 178 prevention of, 55-56, 165
cold, 179 tarry asphaltic, 32
requirements, 180 thickness, 160, 205
signs of icing in, 181 illustration, 206
corrosion under water absorbancy, 15, 16, 147
controlling, 145 wet, 12, 134, 204
designing to prevent, 178 corrosivity of, 45-48
reducing failures caused by, 204 resistivity of, 137
wicking, 149, 189, 204 Liquefied petroleum gas tank, shell

evaluation of, 211,213 and jacket corrosion, 114
(see also Thermal insulation) Liquid metal embrittlement, 51, 191,
Insulation specification 195, 198
to avoid chloride stress corrosion incidence of, tables, 52, 53
cracking, 199 Liquid penetrant inspection, 96
and guidelines, 146 illustration, 97
Insulation standards, revisions, to,
141
M
"Insulation Works--Protection Against
Corrosion for Cold and Hot Maintenance practices, 24-26
Insulation at Industrial Plants," Mastic
182 deterioration of, 24
fire retardant, 95
joint, 92
J
moisture intrusion and, 87
Jacketing reinforced, 202
aluminum, 72, 167 temperature limitations, 22
corroded, illustration, 90, 93 Materials, changes in and corrosion,
leak point, illustration, 93 156
as moisture barrier, 201 Materials Technology Institute (MTI)
perforated, 88, 92 corrosion studies, 122-123
broken, 12 inspection methods study, 152
illustration, 13 Mechanical design issues, 54
corrosion, 114 Metallic foils, 51-52 (see also Alumi-
galvinized steel, 72 num foil)
heat sealed moisture barrier, 179 Metallic paints (see Paints, metal-
installation, 167-168 filled)
metallic nonbreathing, 22 Military Specification Insulation
selection, 23 Materials, Thermal, with
Joint seals, prefabricated expansion/ Special Corrosion and Chlo-
contraction, 184 ride Requirements (MIL-
24244), 35, 178, 213
Mineral wool, 19, 46
L corrosiveness testing, 223, 226, 229
Lagging in hot and cold service, 72
corrosion mechanism in, 56 performance, 129
corrosion under, 42 properties, 66-67
inhibited, 48-50 Moisture
specification, 59 atmospheric, 28, 29, 32
water in, 45, 52, 189, 197 barrier, 199 (see also Vapor bar-
Leachant-coupon tests, 223, 225, riers)
226, 229-230 as cause of corrosion, 204
Lead chromate phenolic coating, 157 intrusion points, 87, 88
SUBJECT INDEX 239
penetration, 119, 120 corrosion under, illustration, 21

prevention of, 115, 179-180, fire retardant, 115
200 application, 120
resistance, 7 in high-temperature testing, 163
leaching of, 119
water saturation, 116
N
lagging, 45
National Board Inspection Code, 26 rigid, 65
Nozzles, Type 316 stainless steel, water extract properties, table, 47
failure of, 95 Polyvinyl chloride (PVC)
breakdown, 31-32
source of chlorides, 105
O
Potential, role of, 34-37
Oleoresinous coating, 163-164 Pressure vessels, corrosion, table, 79
Outdoor exposure test, 171-172 Protective barrier, failure of, 11
results, tables, 173-176
Oxide film formation, 124, 127
in stainless steels, 168-169 R
Rainwater
P
sodium chloride in, 24, 28-31
illustration, 29
Painting pH of, 28
criteria, 108-110 source of moisture, 124
effects on corrosion, illustration, (see also Runoff)
77 Red lead alkyd, 157
Paints Refining and petrochemical industry
metal-filled, 51-52, 191 plants, 122
protection, 146 Research, Development, Test, and
(see also Coatings) Evaluation (Navy) Test Re-
Perlite insulation, types of, 129 quirements for Thermal Insu-
Perlite-silicate, 67 lating for Use on Austenitic
Petrochemical plants, corrosion Stainless Steel (M12-1T), 212
problem, 71 Runoff, 32, 36
Phenolic lead chromate primer, 163- Rusting, 115
164 bonding and, 161
Pitting, 38, 126 breakthrough, illustration, 115, 117
of carbon steel, 43 under cement, 158
of expanded perlite, 129
of stainless steel, illustration, 91
S
Plant conditions, 24, 72-73
Polyisocyanurate foam, 200, 201 Salts (see Sodium chloride)
rigid, 65-66 Scratch test, 171
Polyurethane foam (PUF), 18-19, results, tables, 173-176
21, 45, 64-65, 72 Seawater deluge, 116
Shell corrosion, 114 controlling, 145

Silicon, 39 factors affecting, 11
Silicone acrylic primer, 163-164 equipment, 156
Silicone-aklyds, 50 scaling and pitting, illustration,
testing results, 192-193 94
Silicone coatings, 108, 110, 111 carbon/low alloy
application information, tables, corrosion phenomena, 43-44
170, 171 surface preparation and coating
effectiveness, 172 system, 186-187
tables, 173-176 corrosion under thermal insulation
manufacturer's information, tables, materials, 121
169, 170 duplex stainless (18 Cr 5 Ni), 53
Silicone paint, aluminum filled, 193 extra low interstitial ferritic (18 Cr
Sodium chloride 2 Mo), 53
airborne, 28 SAE 1010, 225, 226
concentration, 15, 24 stainless
deposits, 33 AISI Type 304, 38, 39, 50, 74,
hygroscopic, 33, 36, 40 87
in rainwater, 28-31 exposure testing, 172
illustration, 29 stress corrosion cracking in,
Sodium metasilicate inhibitors, 35, 104, 106
36, 37, 39 tables, 193, 194
Sodium silicate insulation, 7 AISI Type 316, 38, 56, 74, 87
application to steel, 49 failure, 95
in chemical plants, 87, 89 coatings for, 168-177
inhibiting action,200 corrosion phenomena, 44-45
leaching, 88, 89, 94 18Cr 8 Ni, 168
Spectrometer analysis, table, 100 environment, under insulation,
Steel Structures Painting Council 28
SSPC-5, 157, 168 external stress corrosion crack-
SSPC-6, 157 ing, 103
Steels failure categories, illustration,
austenitic, 166 46
characteristics, 7 mechanical bonding, 169
insulation used with, 167 vessels, illustration, 90-91
protection, 7 Stress corrosion, chloride induced, 7
shortcomings, 8 Stress corrosion cracking
stress corrosion cracking, 27 of austenitic stainless steels
prevention, 188 factors, affecting, 27
"super" (20 Cr 25 Ni), 53-54 chloride, 29, 68, 100
susceptible materials, 38-39 mechanism of, 166
carbon, 38 prevention of, 165
corrosion under thermal insula- specifications to avoid, 199
tion transgranular, 28, 38-39
SUBJECT INDEX 241
external (ESCCC), 27, 31, 39-40 Tests

of carbon steels, 44 accelerated corrosiveness methods,
effects of surface treatment, 220
table, 218 condensation, 225, 229
fire risks in, 195 Dana, 211
induction of, 199 accelerated, 212, 216
prevention of, 103-104, 191 Elcometer Adhesion, 171, 172
in process industries, 188-189 results, tables, 173-176
stainless steel under thermal in- environment cabinet, 172
sulation, 103 results, tables, 173-176
intergranular, 38 high-temperature cyclic, 161-163
under insulation, 8, 103 results, table, 164
mechanisms of, 7 leachant-coupon, 223, 225, 226,
temperature ranges for, 37-38 229-230
transgranular, 28, 38-39, 104 outdoor exposure, 171-172
Stress corrosion cracking prevention results, tables, 173-176
aluminum foil use for, 188 salt fog, 172
under lagging, 57 results, tables, 173-176
materials fabrication/selection, scratch, 171
52-54 results, tables, 173-176
metallic foils/paints, 51-52 stress corrosion cracking, 211
program, 100-102 voltammetry, 223, 224, 226, 229-
Stress corrosion cracking tests, 211 230
Sulfur dioxide concentrations, 28 weatherometer, 172
Surface preparation for carbon/low results, tables, 173-176
alloy steel, 186-187 Thermal insulations
accelerated corrosiveness test
methods for, 220
carbon steel corrosion under, fac-
tors affecting, 11
T
characteristics of, 6
Temperature(s) corrosion in petrochemical
coating failure and, 161-163 plants, 71
control, 5 defined, 5
fluctuation reduction, 5 function, 5
problems, 212-213 inhibition of, 8
service materials and composition, 6
cold, 72 prevention of stress corrosion
corrosion-prone, 74 cracking under, 188
effects of, 13-15 Thermal Insulation Manufacturers'
hot, 72 and Suppliers Association,
recommended, table, 64 143
for stress corrosion cracking, 37- Thermal insulation materials
38 contribution to corrosion, 67
generic types and properties, 63 W

k values, tables, 64, 65
Water
selection, 67 absorbance (see Insulation)
service temperature ranges, 64
cooling tower, 24
Thermal insulation systems for hot corrosivity of, 11, 18
and cold service, 72
drainage, 149
Thermal Insulation Users Liaison
electrolytes in, 28
Group (UK), 42 ingress
Thermal shock test, 172 flooding, 189, 190, 191, 197
results, tables, 173-176 migration, 189, 190, 191, 197
protection against, 198
sources of, 32-33
U trapped, illustration, 75
Union Carbide Gulf Coast plants, as vapor, 6, 7
104, 122 (see also Moisture; Vapor)
Urea formaldehyde (UF) foam, cor- Waterproofing, 54
rosiveness testing, 223 Weather barriers, 32
Urethane in cold insulation, 179 aluminum, 115, 123
breakdown of, 33
effectiveness, 131
evaluation, 200
V
on irregular shapes, 202
Vapor, 6 purpose, 22
barriers, 18, 179 Weatherometer tests, 172
injury to, 180 results, tables, 173-176
purpose, 22 Weatherproofing, 22-24
condensation, prevention of, 5 design, 149
retarders, 7 Weathershed ring, 204
transmission rate,7 illustration, 205
(see also Moisture; Water) Wet exposure cycle characteristics,
Vaporproofing, 22-24 11
design, 149 Wicking, 149, 189, 204
(see also Waterproofing) evaluation of, 211,213
Vinyl coatings, tables, 169, 170
application information, tables,
Z
170, 171
effectiveness, 172 Zinc-rich paints, 51, 57
tables, 173-176 efficiency of, 198
Voltammetry, 223, 224, 226, 229- incidence of cracking in, 194
230 inorganic, 155, 157-159, 161

STP880 Eb DL.15102 PDF

Uploaded by

Copyright:

Available Formats

STP880 Eb DL.15102 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

STP880 Eb DL.15102 PDF

Uploaded by

Copyright:

Available Formats

Copyright by ASTM Int'l (all rights reserved); Fri Feb 8 07:23:51 EST 2019

ASTM SPECIAL TECHNICAL PUBLICATION 880

ASTM Publication Code Number (PCN)

1916 Race Street, Philadelphia, PA 19103

Main entry under title:

Corrosion of metals under thermal insulation.

(ASTM special technical publication; 880)

Copyright 9 by AMERICAN SOCIETY FOR TESTING AND MATERIALS 1985

Printed in Ann Arbor, MI

The symposium on Corrosion of Metals Under Thermal Insulation was pre-

Atmospheric Corrosion of Metals, STP 767 (1982), 04-767000-27

Atmospheric Factors Affecting the Corrosion of Engineering Metals, STP

Chloride Corrosion of Steel in Concrete, STP 629 (1977), 04-629000-27

The Function of Thermal Insulation--JACK M. BARNHART 5

Factors Affecting Corrosion of Carbon Steel Under Thermal

Factors Affecting the Stress Corrosion Cracking of Austenitlc

A Review of the European Meeting on Corrosion Under Lagging Held

THERMAL INSULATION MATERIALS

Thermal Insulation Materials: Generic Types and Their Properties--

Experience with Corrosion Beneath Thermal Insulation in a

Recent Experiences with Corrosion Beneath Thermal Insulation in a

Failure of Type 316 Stainless Steel Nozzles in Contact with Fire

External Stress Corrosion Cracking of Stainless Steel Under Thermal

Shell and Jacket Corrosion of a Foamed In-Place Thermally Insulated

Behavior of a Copper W a t e r T u b e Exposed to N a t u r a l C a r b o n a c e o u s

Controlling C a r b o n Steel Corrosion U n d e r I n s u l a t i o n - -

Protective C o a t i n g System Design for I n s u l a t e d or Fireproofed

P r e v e n t i o n of Chloride Stress Corrosion C r a c k i n g U n d e r I n s u l a t i o n - -

Designing to P r e v e n t Corrosion of Metals U n d e r I n s u l a t i o n - -

Use of A l u m i n u m Foil for P r e v e n t i o n of Stress Corrosion C r a c k i n g of

U s i n g Speelfieations to Avoid Chloride Stress Corrosion C r a c k i n g - -

Use I n s p e c t i o n as a M e a n s of R e d u c i n g Failures Caused By Corrosion

A New A p p a r a t u s a n d Test Procedure for R u n n i n g A S T M C 692

C o m p a r i s o n s of Several Accelerated Corrosiveness Test M e t h o d s for

Very serious corrosion problems can occur to plant equipment, tankage,

conference was held in Britain on "Corrosion Under Lagging." The success of

The Function of Thermal Insulation

REFERENCE: Barnhart, J. M., "The Function of Thermal Insulation," Corrosion of

KEY WORDS: corrosion, insulation, energy, chlorides, stress corrosion

Thermal insulations or thermal insulation systems are usually defined as

1. Conserve energy by reducing heat loss or gain of piping, ducts, vessels,

1Thermal Insulation Manufacturers Association, 7 Kirby Plaza, Mt. Kisco, NY 10549.

Thermal insulation is used to control heat flow in temperature ranges from

a known " i n h i b i t o r . " Stress corrosion cracking u n d e r insulations is not a sim-

and from later in the p a p e r :

Factors Affecting Corrosion of

ABSTRACT: The problem of corrosion under insulation is studied by analysis of seven

KEY WORDS: corrosion, insulation, paints, coatings, maintenance, inspection, temper-

T h e a m o u n t of c a r b o n steel lost b e c a u s e of corrosion u n d e r i n s u l a t i o n ( C U I )

(1) equipment design,

These factors will be examined individually to demonstrate how past common

Unfortunately, insulation selection has not been based on any consideration

Although extensively used in the past, polyurethane foam (PUF) is not a

Although a low k factor is claimed as an advantage for PUF, in practice,

1. Corrosion is possible under all types of insulation. The insulation type is

FIG. 8 - - G e n e r a l corrosion under a m a n w a y on a vessel with a moderate operatbtg tempera-

FIG. 9--Cot7"osion u n d e r P U F on an idle unit,

At 100~ for example, each kilogram of water is going to produce almost

mastics deteriorate quickly, a frequent schedule of inspection and maintenance