Coalescers

Filters
LOW pH UNDERDEPOSIT CORROSION ON A179 CARBON STEEL
Ms. Cheah Phaik Sim
Titan Petrochemicals Sdn. Bhd.
Pasir Gudang, Johor.
ABSTRACT
In a quench system of an ethylene plant, ater is separated from hydrocar!ons and car!on solids "ia
filters and coalescers and further stripped off of dissol"ed hydrocar!ons !efore !eing "apori#ed into steam.
It is not uncommon corrosion occurs in this system comprising of car!on steel equipment and piping. The
type of corrosion that has !een encountered is lo p$ under deposit at the re!oiler heat e%changers.
Chemicals such as amines are added to control corrosion, there!y gi"ing rise to either a change in the !ul&
corrosi"eness of the en"ironment.
In this paper, corrosion pro!lems that arise from contaminants in the quench ater are discussed and some
chemical solutions are also presented.
INTRODUCTION
Pr!ess Des!rip"i#
In an ethylene plant, direct contact of ater ith the pyrolysis furnace effluent from thermal crac&ing is
the stage of cooling. This step not only cools the gases, !ut also condenses the dilution steam and some of
the hydrocar!on portion of the gas. The resultant ater from the quench ater !ottoms is a mi%ture of
circulating ater, condensed dilution steam and condensed hydrocar!on. Part of the hot quench ater is used
for lo le"el heating ser"ices throughout the plant.
Before the ater can !e reused for dilution steam, it must !e cleaned for hydrocar!on and solids.
'quipment for the clean up of the quench ater includes oil(ater separators, filters, coalescers and stippers.
The quench ater form the oil separator is split ith some !eing sent !ac& to the quench ater recirculating
loop and some going to dilution steam ma&e up. This is sent to a second more effecti"e ater hydrocar!on
phase separation "ia filters and coalescers prior to !eing sent to a ater stripper for further remo"al of
dissol"ed hydrocar!on and gases. The )clean* quench ater is then fed directly to the dilution steam
generator and "apori#ed to steam "ia re!oilers using medium pressure steam as heating medium. Figure +
shos a typical flo diagram of a quench system.
$I%URE 1 T&pi!a' ()e#!h $'* Dia+ram
Chemi!a' Trea"me#" Pr+ram App'ie,
Blodon
,SG -e!oiler
.uench /ater
Toer
/ater Stripper
.uench /ater
'%changers
.uench /ater
Pump
,SG Feed
Pumps
Furnace
,ilution Steam
Generator 0,SG1
Blodon
Cooler
The quench ater contains high le"els of !oth solu!le and insolu!le oils. The solu!le oils are remo"ed at
the separation equipment upstream of the dilution steam generator hile the insolu!le oils hich
are typically aromatic oils cannot !e separated. 2cids from the crac&ed gas condensed in the
quench ater causing lo p$ condition. Therefore, emulsion !rea&ers are in3ected to allo the
insolu!le oils to !e remo"ed !etter in the oil(ater separator and coalescers. 2mine is in3ected in
the quench system to neutrali#e the acids and !ring up the p$ of the quench ater.
Crrsi# E#!)#"ere,
The primary corrosion encountered is the lo p$ under deposit type. 4any failures associated ith this
attac& ere e%perienced on the ,SG re!oilers. The upper half of the re!oiler tu!e !undle as found to !e
su!3ected to attac& in the form of ditches ith rough rolling contour. The thic& layer of deposits ere
essentially corrosion products.
The deposit sample as found to !e mostly of iron o%ide 056.671, ith trace amounts of silicon,
aluminum, phosphorous and sulfur species. The iron o%ide is either coming from the tu!e itself due to
corrosion or a carryo"er from the upstream equipment and process. Presence of oil as also noted.
$I%URE - Re.i'er T).e $ai')re $I%URE / Depsi" Samp'e
Ma"eria's 0 !#s"r)!"i#
The selection of tu!e !undle material for the ,SG re!oilers is 2ST4 2+85. This car!on steel has a
chemical composition of
Car!on 9.9: ; 9.+67
4anganese 9.<8 ; 9.:=7
Phosphorus, ma%. 9.9=>7
Sulfur, ma%. 9.9=>7
CORROSION MECHANISM
The definition of corrosion is the undesired attac& of a material as a result of electrochemical reactions
ith components in the en"ironment. 'lectrochemical corrosion is resulting from reactions !eteen a metal
surface and an ion?conducting en"ironment such as ater. There is at least to reactions in"ol"ed, namely
anodic reaction during hich a metal dissol"es hile releasing electrons
Fe @ Fe
<A
A <e
and a cathodic reaction, in hich the electrons formed are remo"ed
<$
A
A <e @ $<
Brganic contaminants decompose under high pressure and temperature to form organic acids, thus
reducing p$ in the process ater to >.9 and loer. 2cid promotes the iron corrosion reactions !y supplying
a reactant, $
A
. The o"erall reaction isC
<$
A
A <-CBB
?
A Fe @ Fe0-CBB1< A $<
corrosion
product
Corrosion of ferrous metal piping results in the formation of iron o%ides products many times the "olume
of the metal loss from the pipe all. 2cid attac& is usually in the form of thinning of all surfaces.
Dnder deposit corrosion occurs hen corrosion products accumulate on a metal surface. This is locali#ed
corrosion caused !y o%ygen concentration cells. 2s deposits form on metal surfaces, o%ygen le"els under the
deposit are reduced and an electrochemical corrosion cell is set up. This can cause the metal under the
deposit to corrode rapidly, leading to se"ere locali#ed corrosion and perforation of the metal surface.
DISCUSSION
The contaminants in the )clean* quench ater contain organic acids 0ma3ority acetic, some propionic and
!utyric1, ammonia and some car!onates from the crac&ed gas. Typical acetic acid le"el sampled from the
!lodon is =99 to :99 ppm. 2mmonia at stripper outlet is a!out <9 to >9 ppm. The corrosion protection
rendered !y the chemical treatment program uses a lo "olatility neutralising amine to neutrali#e all acid
species and ele"ate the ater p$ to 6.9 to 5.9. 2mmonia, an al&aline compound ill ha"e the same effect as
the amine. It is hoe"er, much more "olatile, higher neutralising capacity and !asicity than the amine.
p$ measured of the !ul& ater ere a"eraging 8>7 ithin the range of 6.9 to 5.9. Bnly <=7 ere !elo
p$ of 6.9 and <7 a!o"e p$ of 5.9. Figure E shos the operating p$ from 2pril <999 to June <99+. 2t p$
!elo 6.9, the passi"e magnetite film that protects the car!on steel ould !e destroyed and as such,
significantly increases general corrosion rate.
$I%URE 1 Opera"i#+ pH 0rm Apri' -222 " 3)#e -221
$oe"er, "isual inspection re"ealed the tu!e lea&s ere not related to tu!e all thinning as a result of
general p$ corrosion. The failures ere related to locali#ed under deposit corrosion and the lea&s ere
located at the upper section of the re!oiler here !ul& of the e"aporation is e%pected to ta&e place. The
under deposit corrosion de"eloped as a result of concentration of organic acids 0mostly acetic1 underneath
the porous deposit.
/hen the !ul& ater at al&aline p$ enters the porous deposit and ith high percentage of e"aporation
ta&ing place at the upper section of the !undle, the condition for setting up concentration cells under the
deposits li&e a miniature !oiler pre"ails. The high "olatility ammonia ill "apori#e and tra"el ith the steam
upsetting !ase(acid ratio in the ater phase. This could result in a locali#ed lo p$ en"ironment that is
highly corrosi"e toards magnetite and car!on steel and ultimately leading to tu!e lea&. Figure > illustrates
this phenomenon.
The conditions fa"ora!le for under deposit corrosion effects to ta&e place areC
!ul& ater temperature is approaching !oiling point
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
P
e
r
c
e
n
t
a
g
e

%
A
p
r
-
0
0
M
a
y
-
0
0
J
u
n
-
0
0
J
u
l
-
0
0
A
u
g
-
0
0
S
e
p
-
0
0
O
c
t
-
0
0
N
o
v
-
0
0

e
c
-
0
0
J
a
n
-
0
1
!
e
"
-
0
1
M
a
r
-
0
1
A
p
r
-
0
1
M
a
y
-
0
1
J
u
n
-
0
1
Month
DSG Blowdown pH In Control Zone
p# %n range & 8'0 to 9'0( p# ) 8'0 p# * 9'0
TURNAROUND
!!!"!!#
high heat flu% 0heat a!sor!ed per unit area1
$I%URE 4 I'')s"ra"i# 0 U#,er Depsi" L* pH Crrsi#
Both of these conditions pre"ail hen the percentage e"aporation in the !ul& ater is high. This e%plains
hy the tu!e lea&s are found only at the upper section of D?tu!e !undle and in case history presented in
Ta!le + that the medium pressure steam re!oilers ith lo e"aporation did not ha"e any tu!e lea&s ith the
amine program.
TABLE 1 Case His"r& 0 DS% Re.i'ers
P'a#"
Re.i'er Hea"i#+
Me,i)m
5 E6apra"i# Remarks
2 .uench Bil 8.: Fo corrosion
B 4P Steam >.6 Fo corrosion
C 4P Steam => Corrosion
, 4P Steam =9 Corrosion
' .uench Bil 6.< Fo corrosion
Condensate modeling system 0C4S1 simulated ater G steam chemistries in steam generators in the
presence of treatment chemicals and common contaminants. 2 simulation of the impact of concentration
effect 0>97 e"aporation or < times concentration1 of !lodon ater is shon in Ta!le <.
It can !e seen that the amine product concentrates more than the acid. That is desira!le as it helps to
maintain the ater p$. 2mmonia on the other hand due to its high "olatility shos significant drop in
concentration. The lost in ammonia from the !ul& ater cause the p$ to drop from 6.9 to >.5, there!y
creating a locali#ed lo p$ en"ironment. It is important to note that e"en ith consistently good p$ control
of 6.9 to 5.9, the possi!ility of under deposit locali#ed lo p$ corrosion still e%ists. B!"iously, hen the
!ul& ater p$ is lo, it ould !e easier to de"elop under deposit lo p$.
TABLE - C#!e#"ra"i# E00e!" Usi#+ Ami#e
B'*,*# 71225 *a"er8 B'*,*# 7425 e6apra"i#8 C#!e#"ra"i# 0a!"r
p$ @ 6.9 p$ @ >.5
Feutralising amine @ <69 ppm Feutralising amine @ >>+ ppm +.58
2cetic acid @ E99 ppm 2cetic acid @ :>: ppm +.:E
4P Steam at <:9 C
p$@6.> Bul& /ater
p$H8
Fe
=
B
E
Steam out
2mine(2mmonia
-?F$
E
(F$
E
B$
C$
=
CBB$
,eposit
$
<
B F$
E
A
B$
?

C$
=
CBB
?
$
A
)Free* Caustic -egion
F$= @ E9 ppm F$= @ <+ ppm 9.>=
In order to maintain the !lodon !ul& ater p$ at 6.9 to 5.9, non?"olatile !ase chemicals such as
products ith a caustic(phosphate 0Fa(PBE1 ratio !lend need to !e used. Caustic is a stronger al&aline than
amine !ut can itself !e corrosi"e causing gauging and em!rittlement to metals. Phosphate !uffers the ater
reducing the chance of large p$ change due to the de"elopment of high caustic concentrations. The under
deposit concentration effect ill not create an acidic locali#ed p$ as shon in Ta!le =. The same acetic and
ammonia are present in the !ul& ater. The !ul& ater p$ is ele"ated to 5.: !y +=> ppm of Fa and +9 ppm
of orthophosphate and only >9 ppm of amine instead of <69 ppm pre"iously as shon in Ta!le =. The amine
feed is much loer as it is designed to ele"ate the dilution steam p$. To control !lodon !ul& ater at 5.:
using amine alone ill require > times of current amine feed rate. Figure : shos the p$(PBE recommended
to control re!oiler corrosion.
TABLE / C#!e#"ra"i# E00e!" Usi#+ Na9PO1
B'*,*# 71225 *a"er8 B'*,*# 7425 e6apra"i#8 C#!e#"ra"i# 0a!"r
p$ @ 5.: p$ @ 5.+
Feutralising amine @ >9 ppm Feutralising amine @ 5E ppm +.66
2cetic acid @ E99 ppm 2cetic acid @ :>: ppm +.:E
F$= @ E9 ppm F$= @ <+ ppm 9.>=
Fa @ +=> ppm Fa @ <89 ppm <.9
B?PBE @ +9 ppm B?PBE @ <9 ppm <.9
$I%URE : Cr,i#a"e, pH9POE C#"r' ;#e
CONCLUSION
Io p$ under deposit corrosion is aggra"ated !y the high e"aporation rate 0=971 at the re!oiler causing
organic acids to concentrate in the quench ater particularly underneath the deposits. 2mines used may !e
sufficient to neutrali#e the !ul& ater !ut not at the e"aporation #one. 2ddition of p$(PBE chemical
Coord$nated pH"PO% Control Bo&
8'2
8'4
8'6
8'8
9
9'2
9'4
9'6
9'8
10
1'0 10'0
Orthopho'phate a' PO% (pp)*
p H a t + , C
2'0 4'0 6'0 8'0
CONTRO- ZON.
Na+,O4 - 2'8
Na+,O4 - 2'2
treatment program ill help ele"ate the p$ at the e"aporation #one !ut must !e controlled to pre"ent
formation of free caustic.
RE$ERENCES
+. 4arilyn /. Blasch&e, .uench /ater Chemistry and ,onstream Concerns, 'thylene Forum 0<99+1.
<. 4atcor Technology G Se"ices Pte. Itd., Failure 2nalysis of -e!oiler Tu!e and 2nalysis of ,eposits
0+55>1.
=. 2ST4 Standard Specification ,esignation 2+85, Seamless Cold?,ran?Io Car!on Steel $eat
'%changer and Condenser Tu!es 0+55=1.
E. Corrosion 2tlas Jol. +, 2+?2: 0+55+1.
>. Bet# $and!oo& of Industrial /ater Conditioning, 6<?6= 0+55+1.