Corrosion in Amine Plant
Corrosion in Amine Plant
Corrosion in Amine Plant
Understanding corrosion in
alkanolamine gas treating plants
Part 1 &2
Proper mechanism diagnosis
optimizes amine operations
M.S. DuPart, T.R. Bacon and D.J. Edwards
GAS/SPEC Technology Group
INEOS LLC
As part of a Federal Trade Commission mandated remedy to the merger of The Dow Chemical Company and the Union Carbide Corporation,
INEOS plc was able to purchase both Dows Ethanolamines and GAS/SPEC MDEA-based specialty amine businesses. This purchase became effective on
February 12, 2001.
INEOS LLC was set up as the newly acquired company, which includes the GAS/SPEC Technology Group. All the key Ethanolamines and GAS/SPEC personnel
were retained by INEOS LLC. All GAS/SPEC products, technology and know-how became the exclusive property of INEOS on a global basis.
Reprinted from HYDROCARBONPROCESSING
, April 1993 issue, pages 75-80 and May 1993 issue, pages 89-94.
Copyright 1993 by Gulf Publishing Co., Houston, Texas. All rights reserved. Used with permission.
HYDROCARBON PROCESSING 3
M. S. DuPart, T. R. Bacon and D. J . Edwards,
Dow Chemi cal USA, Freeport, Texas
C
orrosi on i n al kanol ami ne gas treatment pl ants
r esul ts i n unschedul ed downti me, pr oducti on
l osses, reduced equi pment l i fe and even i njury or
death. Al though i t i s vi rtual l y i mpossi bl e to el i mi nate, i t
can be control l ed and mi ni mi zed. Desi gn engi neers and
pl ant operators must understand how corrosi on affects an
ami ne pl ant. Some key areas on whi ch to focus are: cor-
rosi on mechani sms, di fferent corrosi on types, corrosi on-
contri buti ng factors and preventi ve measures. Desi gn-
er s can emphasi ze pr eventi ve measur es wi th desi gn
consi derati ons, operati ng parameters and sol vent choi ce.
Di agnosi ng the correct type(s) of corrosi on that occur
in amine systems is half of the problem. This simple review
outl i nes the vari ous corrosi on types for al kanol ami ne pro-
cesses and the condi ti ons that frequentl y occur. Once you
have determi ned your corrosi on probl em, the preventi ve
desi gn suggesti ons remedy most pl ant probl ems.
Amine plants history. Regener ati ve pr ocesses have
used al kanol ami nes for CO
2
and H
2
S removal si nce the
early 1930s. However, removing H
2
S and/or CO
2
with alka-
nol ami ne-based gas condi ti oni ng sol vents posed i ts own
pr obl emscor r osi on. The ami ne i tsel f i snt the cul pr i t,
but the aci d gas that the ami ne absor bs i s. I t has been
shown that carbon steel exposed to al kanol ami nes under
an i ner t gas pad has l ower cor r osi on r ates than those
exposed to a water sol uti on under si mi l ar condi ti ons.
Extensi ve corrosi on data proves that aci d gas wi th al ka-
nol ami nes does affect cor r osi on r ates. Fi g. 1 for
monoethanol ami ne (MEA) shows that not onl y does the
aci d gas have an effect, but the aci d gas type and the rati o
of H
2
S to CO
2
al so i mpacts the corrosi on rate.
1
Data sug-
gest that MEA wi th H
2
S or CO
2
al one i s more corrosi ve
than a mi xture of the two. I n Fi g. 2, the effect of aci d gas
l oadi ng i s shown. As l oadi ng i ncr eases, cor r osi on r ates
also increase.
2
From Fig. 3, the data correlates elevated cor-
rosi on rates at hi gher temperatures.
3
Si mi l ar data (Fi g.
4) for di ethanol ami ne (DEA) rei nforce the same effects of
aci d gas and temperature as i n Fi gs. 1 to 3. However, there
i s no cor r el ati on that pr edi cts cor r osi on behavi or wi th
mi xtures of H
2
S and CO
2
i n MEA or DEA servi ce. Never-
thel ess, pl ant data i ndi cate that i n pr edomi nantl y CO
2
ser vi ce, smal l quanti ti es of H
2
S have a str ong passi ve
i nfl uence.
4
DEA corrosi on data (Fi g. 4) cl osel y paral l el the
MEA data. DEA i s gener al l y consi der ed l ess cor r osi ve
than MEA, but onl y i n rel ati ve terms.
Cor r osi on testi ng of car bon steel wi th methyl
di ethanol ami ne (MDEA)/CO
2
i ndi cates that there may be
an advantage to using MDEA from a corrosion standpoint.
5
Fig. 5 compares corrosion results of MEA, DEA and MDEA.
MDEA has significantly lower corrosion rates than MEA or
DEA. Pl ant operati ng data confi rm MDEA l aboratory cor-
rosi on test resul ts. Al so, Fi g. 5 i ndi cates that hi gher ami ne
concentr ati ons ar e mor e cor r osi ve. Whi l e MDEA i s not
PART 1
Understanding corrosion in
alkanolamine gas treating plants
Proper mechanism diagnosis optimizes amine operations
0
1 0
2 0
3 0
4 0
Corrosion rate
1 4 0 2 1 2 2 4 0
Te m p e ra tu re , F
D a ta n o t g e n e ra te d fo r H
2
S : C O
2
1 : 3 a n d H
2
S : C O
2
3 : 1 a t 2 1 2 F
H
2
S o n ly
C O
2
o n ly
H
2
S : C O
2
1 : 3
H
2
S : C O
2
3 : 1
M
i
c
r
o
n
s
p
e
r
y
e
a
r
Fig. 1. C o rro sio n ra te s fo r 1 5 % M E A with c a rb o n ste e l.
GAS PROCESSING DEVELOPMENTS
i mmune to corrosi on, i t has i nherentl y l ower corrosi vi ty
and can be used i n mi xed aci d gas appl i cati ons to reduce
corrosion tendencies. DEA has also exhibited good corrosion
servi ce i n mi xed aci d gas si tuati ons.
Corrosi on i n al kanol ami ne gas treatment uni ts gener-
al l y focuses on the cross exchanger s ri ch si de, ri ch-ami ne
pi pi ng after the cross exchanger, the sti l l and the reboi l er,
where free aci d gas and hi gher temperatures are the mai n
dri vi ng forces for corrosi on.
Types of c orrosion. Most cor r osi on fai l ur es i n al ka-
nol ami ne servi ce are attri buted to free aci d gas and hi gh
temper atur e. Other factor s account for many cor r osi on
experi ences. Studyi ng di fferent types of corrosi on reveal s
other means by whi ch corrosi on occurs.
General corrosion. Thi s i s the most wi del y r ecog-
ni zed form and i s characteri zed by uni form materi al dete-
ri orati on over the enti re exposed area.
6
Al though thi s cor-
r osi on type destr oys a gr eat deal of mater i al annual l y,
equi pment l i fe can be accuratel y moni tored and predi cted
by si mpl e corrosi on tests. Tests i ncl ude i n-l i ne corrosi on
probes, coupons or ul trasoni c thi ckness testi ng. General
corrosi on rates have been establ i shed for ami ne treatment
pl ants based on over 50 years of operati ng experi ence wi th
di fferent metal l urgy types wi th corrosi on al l owances set
accordi ngl y.
Unfortunatel y, most corrosi on fai l ures i n ami ne servi ce
are not general corrosi on, but l ocal i zed attack. Local i zed
corrosi on usual l y occurs i n speci fi c areas and goes unde-
tected unti l an unexpected or premature fai l ure occurs.
6
Galvanic corrosion. When two di ssi mi l ar metal s are
coupl ed together through a conducti ng el ectrol yte, an el ec-
tri cal potenti al i s set up between the two metal s, whi ch
causes the l ess resi stant metal to corrode.
6
For exampl e,
gal vani c corrosi on occurs when stai nl ess steel trays are
hel d together wi th car bon steel bol ts. The smal l car bon
steel bol ts, acti ng as el ectri cal anodes, preferenti al l y cor-
r oded ar e over the mor e r esi stant stai nl ess steel tr ay
because of the di fference i n el ectri cal potenti al between
the two metal s.
7
Another exampl e of gal vani c corrosi on
i s pump shafts and val ve stems made of steel , or more cor-
rosi on-resi stant materi al fai l i ng because of contact wi th
more nobl e graphi te packi ng.
6
Area effects are al so i mpor-
tant when consi deri ng gal vani c corrosi on.
6
I n the earl i er
tray exampl e, i f carbon steel trays were hel d together wi th
stai nl ess bol ts, then corrosi on woul d be mi ni mi zed due to
the very l arge anode (tray) and the smal l cathode (bol t).
Crevice corrosion resul ts when a concentrati on cel l
forms i n the crevi ce between two metal surfaces or a metal
and a non-metal l i c materi al . Crevi ce corrosi on i s caused by
oxygen depl eti on, changes i n aci di ty or by i nhi bi tor depl e-
ti on i n the cr evi ce.
6
Chl or i de i ons contr i bute gr eatl y to
thi s cor r osi on. Fr equentl y, cr evi ce cor r osi on i s found at
poi nts where tubes are attached to the heat exchanger s
tube sheet and i s more common i n the reboi l er.
8
One way
to pr event thi s type of cor r osi on i s to wel d tubes to the
tube sheet rather than rol l them.
6
Crevi ce corrosi on al so
occur s under deposi ts i n stagnant ar eas. Fi l tr ati on can
prevent sol i d deposi ti on.
Pitting corrosion i s a hi ghl y l ocal i zed and i ntense
corrosi on that resul ts i n rapi d and destructi ve penetra-
ti on at rel ati vel y few spots. Equi pment fai l ures due to pi t-
ti ng are often qui te sudden and unexpected. Hal i de i ons,
par ti cul ar l y chl or i des, and stagnant ar eas usual l y pr o-
mote pi tti ng-type cor r osi on. For pl ants usi ng cor r osi on
i nhi bi tor s, i nhi bi tor concentr ati on i s cr i ti cal when con-
tr ol l i ng thi s type of cor r osi on. I f the i nhi bi tor does not
compl etel y stop the pi tti ng, the i ntensi ty of the pi tti ng
can i ncr ease.
6
Car bon steel i s mor e r esi stant to pi tti ng
corrosi on than stai nl ess steel .
Intergranular corrosion occurs sel ecti vel y at metal -
l i c grai n boundari es and i s most frequentl y found i n stai n-
l ess steel s. Chromi um carbi de preci pi tati on promotes cor-
rosi on i n grai n boundari es by i mproper heat treatment or
wel di ng techni ques. Heat treatment outsi de the sensi ti z-
i ng r ange and l ow car bon stai nl ess steel s such as 304L
4 HYDROCARBON PROCESSING
0
2 5 0 2 7 5 3 0 0
Te m p e ra tu re , F
1 0
2 0
3 0
4 0
5 0
6 0
0 . 2 0 M /M C O
2
0 . 3 0 M /M C O
2
0 . 4 5 M /M C O
2
0 . 6 0 M /M C O
2
Corrosion rate
M
i
c
r
o
n
s
p
e
r
y
e
a
r
Fig. 2. C o rro sio n ra te s fo r 2 0 % M E A a n d C O
2
o n ly with c a rb o n ste e l.
0
Corrosion rate
2 5 0 /1 5 2 8 3 /4 5 3 2 0 /1 2 5
Te m p e ra tu re , F /p re ssu re , p sia
1 0
2 0
3 0
4 0
5 0
6 0
7 0
M
i
c
r
o
n
s
p
e
r
y
e
a
r
Fig. 3. C o rro sio n ra te s fo r 2 0 % M E A a n d C O
2
a tm o sp h e re with c a r-
b o n ste e l.
and 316L have hel ped r educe i nter gr anul ar cor r osi on.
Two other for ms of i nter gr anul ar cor r osi on, wel d decay
and kni fe l i ne attack, devel op from i mproper wel di ng or
l ack of post wel d heat treatment. Use el ectri c arc wel d-
i ng, i nstead of gas wel di ng, to pr event wel d decay and
kni fe l i ne attack on stabi l i zed stai nl ess steel s.
6
Selective leaching i s the removal of one el ement from
a metallic alloy by a corrosion-like process.
6
A classic exam-
pl e of thi s corrosi on type i s the sel ecti ve l eachi ng of copper
out of Monel or 70/30 Cu/Ni when oxygen i s present.
8
Gen-
er al l y, sel ecti ve l eachi ng i s not a concer n si nce copper-
bearing alloys are not used frequently and oxygen is absent
from most systems.
Erosion corrosion i s an i ncrease i n the rate of dete-
ri orati on or attack on a metal because of rel ati ve move-
ment between a cor r osi ve fl ui d and a metal sur face.
6
Sol i ds par ti cl es or gas bubbl es suspended i n the l i qui d
aggr avate thi s type of cor r osi on. Er osi on cor r osi on i s
pr omi nent i n pl umbi ng bends, el bows, tees and val ves
as wel l as i n pumps, bl ower s and i mpel l er s. Vel oci ty,
tur bul ence and the stabi l i ty of pr otecti ve f i l ms al so
affect the degr ee of er osi on cor r osi on. I mpi ngement
attack (cor r osi on caused by tur bul ent fl ow) and cavi ta-
ti on damage (caused by vapor bubbl e formati on and col -
l apse i n a l i qui d near a metal sur face) ar e two exam-
pl es of er osi on cor r osi on.
6
Stress corrosion cracking r esul ts fr om the si mul -
taneous presence of tensi l e stress and a corrosi ve medi a.
Stress can be ei ther resi dual i nternal stress i n the metal
or an external l y appl i ed stress.
6
Chl ori de-i nduced stress
cor r osi on cr acki ng i s a wi del y r ecogni zed for m of thi s
type of cor r osi on. Other factor s that affect the extent of
str ess cor r osi on cr acki ng ar e: oper ati ng temper atur e,
ami ne sol uti ons chemi cal composi ti on, metal composi -
ti on and str uctur e.
9
Stress corrosi on cracki ng i n ami ne uni ts has recei ved
consi derabl e attenti on si nce a recent catastrophi c rupture
of an MEA contactor at a U.S. r efi ner y that r esul ted i n
17 fatal i ti es and extensi ve property damage.
10
Al though
this failure was later attributed to hydrogen-induced crack-
i ng, a joi nt API /NACE commi ttee conducted a stress cor-
rosi on cracki ng survey of ami ne pl ants after thi s i nci dent.
Survey resul ts i ndi cated that stress corrosi on cracki ng i s
most preval ent i n MEA uni ts and to a much l esser degree
i n DEA and MDEA uni ts.
10
Hydrogen damage refers to mechani cal damage of a
metal caused by the presence of, or interaction with, hydro-
gen.
6
I n ami ne servi ce, hydrogen damage i s l i mi ted to bl i s-
ter i ng, embr i ttl ement or attack. Ami ne sol uti ons con-
tai ni ng hydr ogen sul fi de or cyani de can pr omote ei ther
hydr ogen embr i ttl ement or sul fi de str ess cr acki ng.
6,10
Hydr ogen bl i ster i ng can occur i n cor r osi ve ami ne sol u-
ti ons that contai n hydr ogen evol uti on poi sons, such as
sul fi des, arseni c compounds, cyani des and phosphorous-
contai ni ng i ons.
6
Often, hydrogen-associ ated fai l ures are
often mi staken for str ess-cor r osi on cr acki ng. Thi s may
account for the hi gh rate of stress-corrosi on cracki ng i nci -
dents reported i n MEA servi ce.
Mechanical design considerations. Several mechan-
i cal desi gn cri teri a can mi ni mi ze corrosi on. Some cri teri a
are based on common sense, but others have been devel -
oped from past operati ng experi ences.
Metallurgy. I deal l y, from an up-front capi tal cost per-
HYDROCARBON PROCESSING 5
Corrosion rate
1 4 0 2 1 2 2 4 0
Te m p e ra tu re , F
0
D a ta n o t g e n e ra te d fo r H
2
S o n ly a t 1 4 0 F
H
2
S o n ly
C O
2
o n ly
H
2
S : C O
2
1 : 3
H
2
S : C O
2
3 : 1
5
1 0
1 5
2 0
2 5
3 0
M
i
c
r
o
n
s
p
e
r
y
e
a
r
Fig. 4. C o rro sio n ra te s fo r 2 0 % D E A with c a rb o n ste e l.
Corrosion rate
3 0 %
M E A
S o lve n t
5 0 %
D E A
1 5 %
M E A
2 0 %
D E A
5 0 %
M D E A
3 0 %
M D E A
0
5
1 0
1 5
2 0
2 5
3 0
3 5
M
i
c
r
o
n
s
p
e
r
y
e
a
r
Fig. 5. H o t skin c o rro sio n te st with C O
2
a tm o sp h e re a n d c a rb o n ste e l,
7 -d a y te st a t 2 1 0 F.
Table 1. Corrosion rates of MEA and DEA
with various metals
20% MEA, 20% MEA, 20% MEA, 15% DEA,
240F 240F 230F 230F
No AG CO
2
satd H
2
S/ CO
2
H
2
S
D a ta so u rc e L a b L a b P la n t P la n t
M e ta l C o rro sio n ra te , m ic ro n s p e r ye a r
M o n e l 1 . 0 3 . 0 1 . 3 2 . 1
7 0 /3 0 C U /N I 1 0 . 0 *
3 0 4 S S < 1 . 0 < 1 . 0 < 1 . 0 < 1 . 0 * *
3 1 6 S S 1 . 0 < 1 . 0 < 1 . 0
4 1 0 S S < 1 . 0 * * *
C a rb o n ste e l 1 . 0 1 0 3 . 0 5 . 4 *
* C o u p o n c o m p le te ly d e stro ye d d u rin g te st p e rio d
* * P itte d to a m a x im u m d e p th o f 0 . 0 0 2 in . d u rin g te st p e rio d
* * * P itte d to a m a x im u m d e p th o f 0 . 0 1 3 in . d u rin g te st p e rio d
6 HYDROCARBON PROCESSING
specti ve, one woul d l i ke to fabri cate an ami ne pl ant sol el y
from carbon steel . Pl ants constructed of carbon steel have
operated successful l y wi thout corrosi on probl ems, but pro-
cess concessi ons requi red hi gh energy consumpti on. Sub-
sti tuti ng stai nl ess steel i n key pl ant ar eas (r i ch-ami ne
pi pi ng, cross exchanger tubes and i nternal s of the reboi l er
and stri pper) wi l l provi de adequate corrosi on protecti on
and al l ow the pl ant to be operated wi th opti mum energy
i nput. Tabl e 1 and Fi g. 6 show corrosi on rates for di fferent
ami nes wi th var i ous al l oys under both l abor ator y and
pl ant condi ti ons.
1113
Car bon steel exhi bi ted cor r osi on
rates, dependi ng on the test condi ti ons, from <1 mi l s per
year to > 100 mi l s per year. Stai nl ess steel , i n parti cul ar
316, performed adequately under all test conditions. 410SS
appears anodic to carbon steel in some applications. Always
use cauti on when speci fyi ng thi s materi al .
Whi l e some i nvesti gators have shown acceptabl e cor-
rosi on rates wi th Monel and other copper-beari ng al l oys,
11
i ts use i s di scour aged due to unfavor abl e pl ant exper i -
ence. Other evi dence suggests that oxi di zers and ami ne
degradati on products enhance corrosi on of these al l oys.
Tabl e 2 l i sts general metal l urgi cal gui del i nes for ami ne
pl ant constructi on. Avoi d di ssi mi l ar metal s usage to mi n-
i mi ze bi metal l i c or gal vani c corrosi on.
14
Flow in piping and vessels. Mi ni mi zi ng aci d gas
fl ashi ng, whi ch i s keepi ng the CO
2
and/or H
2
S i n sol u-
ti on, i s a major factor when control l i ng corrosi on i n ami ne
uni ts. Several i nvesti gators have recommended 3 ft/s as
the maxi mum sol uti on vel oci ty for carbon steel pi pe.
1416
Sol uti on vel oci ti es of 5 to 8 ft/s can be used wi th stai n-
l ess steel pi pi ng due to i ts hi gher i nherent corrosi on resi s-
tance. Other desi gn practi ces that wi l l reduce turbul ence
and l ocal i zed pressure drop are: i nstal l l ong radi us el bows
and seaml ess pi pe where the sol uti on changes di recti on,
el i mi nate fl ow restri cti ons (i . e. fl ange gaskets protrudi ng
i nto the l i ne), r epai r tubes that ar e not fl ush wi th tube
sheet, do not use threaded connecti ons or socket wel d fi t-
ti ngs and i nstal l l ong taper ed secti ons (opti mum cone
angl e of 7) when changi ng cross secti onal area of fl ow.
16,17
Prevent pump cavi tati on by mi ni mi zi ng sucti on pressure
drop (reduced vel oci ty, strai ght runs of pi pe, etc.) and pro-
vi di ng suffi ci ent NPSHA to i nhi bi t smal l gas bubbl e for-
mati on.
14,17
Control valves. The l ocati on of control val ves hel ps
to reduce aci d gas fl ashi ng and rel ated corrosi on. Locate
stri pper or reboi l er l evel control val ves downstream of the
l ean ami ne cool er.
14
Locate the absorber l evel control val ve
downstream of the lean/rich cross exchanger if a rich amine
fl ash tank i s not i ncl uded. I f a fl ash tank i s used, then
i nstal l the val ve between the absorber and exchanger.
14,16
When two-phase fl ow exi ti ng a control val ve i s i nevi tabl e,
cl oser attenti on to the desi gn of the contr ol val ve and
Table 2. Metallurgical guidelines
for amine plant construction
Item Material Alternate material
A b so rb e r
S h e ll C a rb o n ste e l
I n te rn a ls C a rb o n ste e l S ta in le ss ste e l
L e a n /ric h e x c h a n g e r
Tu b e s ric h ) S ta in le ss ste e l
S h e ll le a n ) C a rb o n ste e l
L e a n a m in e c o o le r
Tu b e s C a rb o n ste e l S ta in le ss ste e l
S h e ll C a rb o n ste e l
C a rb o n b e d
S h e ll C a rb o n ste e l
I n te rn a ls S ta in le ss ste e l
R e g e n e ra to r
S h e ll C a rb o n ste e l
I n te rn a ls S ta in le ss ste e l
R e b o ile r
Tu b e s S ta in le ss ste e l
S h e ll C a rb o n ste e l
Tu b e sh e e t C a rb o n ste e l
S te a m sid e c h a n n e ls C a rb o n ste e l
R e flu x c o n d e n se r
Tu b e s C a rb o n ste e l S ta in le ss ste e l
S h e ll C a rb o n ste e l
R e flu x d ru m C a rb o n ste e l
P u m p s
C a se s C a rb o n ste e l
I m p e lle rs S ta in le ss ste e l
P ip in g
R ic h a m in e S ta in le ss
L e a n a m in e C a rb o n ste e l S ta in le ss ste e l
G e n e ra l c a rb o n ste e l, 3 0 4 S S , 3 0 4 L S S , 3 1 6 S S a n d 3 1 6 L S S a re re c o m m e n d e d fo r g e n e ra l
u se . D o n o t u se c o p p e r, b ra ss o r o th e r c o p p e r b e a rin g a llo ys.
5 0
4 0
3 0
2 0
1 0
0
M
i
c
r
o
n
s
p
e
r
y
e
a
r
C a rb o n ste e l 3 0 4 S S 3 1 6 S S M o n e l
M e ta l
Corrosion rate
Fig. 6. C o rro sio n ra te s o f d iffe re n t m e ta ls with 3 0 % M E A a t 2 5 0 F
a n d C O
2
sa tu ra te d .
Amines lost during test
7 0
6 0
5 0
4 0
3 0
2 0
1 0
0
P
e
r
c
e
n
t
2 6 % D E A 1 5 % M E A 2 9 % M D E A
A m in e , wt%
Fig. 7. O x yg e n d e g ra d a tio n with M E A , D E A a n d M D E A .
downstream pi pi ng wi l l mi ni mi ze the potenti al effects of
hi gh vel oci ty two-phase fl ow.
Filtration. Adequate mechani cal and carbon fi l trati on
can si gni fi cantl y reduce corrosi on i n ami ne uni ts. Not onl y
can filtration reduce particles that promote erosion corrosion
and abr asi on of passi ve fi l ms, i t can effecti vel y r emove
hydrocarbons and, to a l esser extent, ami ne degradati on
pr oducts.
18
Fi l ter ar r ange-
ments should remove all par-
ti cl es gr eater than 5 mi -
cr ons.
14
I n di r ty ser vi ce,
i ncorporate a spare paral l el
mechani cal f i l ter i nto the
design. Exercise caution with
paper and sock-type fi l ters.
These filters can collapse and
recontaminate the solution.
14
Mechani cal fi l ters shoul d
be i nstal l ed upstr eam and
downstr eam of the car bon
bed. The upstream filter acts
as a guar d f i l ter and
extends car bon l i fe. The
downstream fi l ter prevents
car bon fi nes fr om enter i ng
the system. I n sour servi ce,
a full rich stream filter is rec-
ommended. I n CO
2
-onl y
appl i cati ons, a sl i p stream fi l ter on the l ean stream i s ade-
quate. Some pl ants have i ncorporated both ri ch and l ean
fi l teri ng wi th parti cul ar success.
Stress relieving of ami ne pl ant hardware has been
recommended for many years.
14,19
Al though cracki ng does
occur i n vessel s and pi pi ng that were stress-rel i eved, the
NACE/API survey showed that most cracki ng occurred i n
nonstress-rel i eved equi pment. Stress rel i evi ng shoul d be
appl i ed for al l types of ami ne servi ce and i s parti cul arl y
i mportant i n MEA uni ts due to thei r poor cracki ng record.
And strong evi dence suggests that they are speci al l y prone
to cracki ng.
20,21
Miscellaneous. There are a few other desi gn consi d-
erati ons that can mi ni mi ze corrosi on. I nl et knockouts are
useful i n preventi ng bri ne from enteri ng the ami ne sol u-
ti on. Remember, chl ori des can promote pi tti ng, stress cor-
rosi on cracki ng and crevi ce corrosi on. I nert or sweet gas
bl anketi ng of the surge drum and storage tank wi l l mi n-
i mi ze oxi dati on of the ami ne and pr event degr adati on
pr oduct for mati on.
14
DEA i s par ti cul ar l y suscepti bl e to
oxygen degradati on and extra attenti on shoul d be gi ven to
excl udi ng oxygen i n DEA pl ants (Fi g. 7).
22
Regenerati on pressure shoul d be kept l ow wi th the bul k
ami ne temperature not exceedi ng 255F and the reboi l er
heat source not exceedi ng 300F. Operati ng wi thi n these
guidelines will minimize thermal degradation of the amine.
Desi gni ng or modi fyi ng reboi l er tube bundl es to mi ni mi ze
bi ndi ng by omi tti ng tubes i n an X or V patter n can
reduce l ocal i zed overheati ng. Tubes arranged i n a square
pitch tube pattern simplify cleaning. I nstall multiple vapor-
exi t l i nes to prevent stagnant areas of aci d gas.
14,16
Next month: Part 2. Corrosi on case hi stori es i n ami ne
treatment l ook at actual pl ant probl ems. These are excel -
l ent exampl es on how to troubl eshoot potenti al corrosi on
probl ems i n your pl ant. Al so, each exampl e has observa-
ti ons and r ecommended cor r ecti ons. The exampl es ar e
defi ni te pl ant notebook addi ti ons for l ater reference.
LITERATURE CITED
1
McNab, A. J. and Treseder, R. S., Materi al s Requi rement For a Gas Treati ng Process,
Materials Performance, Vol . 10, No. 1, 1971.
2
Pol derman, L. D., Di l l on, C. P and Steel e, A. B., Degradati on of MEA i n Natural Gas
Treati ng Servi ce, Oil & Gas J ournal, May 16, 1955.
3
The Dow Chemi cal Co., Gas Conditioning Fact Book, 1962.
4
Kohl , A. L. and Ri esenfel d, F. C., Gas Purification, 3rd Edi ti on, Gul f Publ i shi ng Co., 1979.
5
The Dow Chemical Co., GAS/SPEC Technology Group, I nternal Corrosion Test Results, 1985.
6
Fontana, M. G. and Greene, N. D., Corrosion Engineering, 2nd Edi ti on, McGraw-Hi l l Book
Co., 1978.
7
Pearce, R. L. and DuPart, M. S., What i s an Ami ne I nhi bi tor System?, 64th Annual GPA
Conventi on, 1985.
8
Pearce, R. L. and DuPart, M. S., Corrosi on i n Gas Condi ti oni ng Pl antsAn Overvi ew,
NACE Corrosi on Paper #39, 1987.
9
Sedri ks, A. J., Corrosion of Stainless Steels, John Wi l ey & Sons, 1979.
10
Ri chert, J. P., Bagdasari an, A. J. and Shargay, C. A., Extent of Stress Corrosi on Crack-
i ng i n Ami ne Pl ants Reveal ed by Survey, Oil & Gas J ournal, June 5, 1989.
11
Lang, F. S. and Mason, J. F., Corrosi on i n Ami ne Gas Treati ng Sol uti ons, Corrosion, Vol .
14, No. 2, 1958.
12
The Dow Chemi cal Co., I nternal Corrosi on Test Resul ts, 1980.
13
The Dow Chemi cal Co., GAS/SPEC Technol ogy Group, I nternal Corrosi on Test Resul ts,
1990.
14
Campbell, J. M., Gas Conditioning and Processing, 4th Edition, Campbell Petroleum Series,
I nc., 1981.
15
Connors, J. S., Desi gni ng Ami ne Treaters To Mi ni mi ze Corrosi on, Proceedi ngs of the
1973 Gas Condi ti oni ng Conference.
16
Di ngman, J. S., Al l en, D. L. and Moore, T. L., Mi ni mi ze Corrosi on i n MEA Sweeteni ng
Uni ts, Proceedi ngs of the 1966 Gas Condi ti oni ng Conference.
17
Shei l an, M. and Smi th, R. F., Hydraul i c Fl ow Effect On Ami ne Pl ant Corrosi on, Oil &
Gas J ournal, November 19, 1984.
18
Chakma, A. and Mei sen, A., Acti vated Carbon Adsorpti on of Di ethanol ami ne, Methyl
Di ethanol ami ne and Thei r Degradati on Products, Carbon, Vol . 27, No. 4, 1989.
19
Maddox, R. N., Gas and Liquid Sweetening, 2nd Edi ti on, Campbel l Petrol eum Seri es, I nc.,
1977.
20
Schutt, H. U., New Aspects of Stress Corrosi on Cracki ng i n Monoethanol ami ne Sol u-
ti ons, NACE Corrosi on Paper #159, 1988.
21
Teevens, P. J., Toward a Better Understandi ng of the Cracki ng Behavi or of Carbon Steel
i n Al kanol ami ne Sour Gas Sweeteni ng Uni ts; I ts Detecti on, Moni tori ng and How to Avoi d
I t, NACE Corrosi on Paper #198, 1990.
22
The Gi rdl er Corp., Report to the U.S. Navy, 1950.
23
Treseder, R. S., Some Comments on Proposed Mechani sms for Carbon Di oxi de Corro-
si on, NACE Corrosi on Paper #284, 1984.
24
McFaddi n, D. E., H
2
S and CO
2
Corrosi on of Carbon Steel i n Natural Gas Processi ng
Pl ants, Oil & Gas J ournal, December 13, 1951.
25
Froni ng, H. R. and Jones, J. H., Corrosi on of Mi l d Steel i n Aqueous Monoethanol ami ne,
I ndustrial and Engineering Chemistry, Vol . 50, No. 12, December 1958.
26
The Dow Chemi cal Co., Gas Condi ti oni ng Laboratory, Ami ne Degradati on, August 27,
1965.
27
Kennard, M. L. and Mel sen, A., Mechani sms and Ki neti cs of Di ethanol ami ne Degra-
dati on, I ndustrial & Engineering Chemistry Fundamentals, Vol . 24, 1985.
28
Bl anc, C., Gral l , M. and Demarai s, G., The Part Pl ayed by Degradati on Products i n the
Corrosi on of Gas Condi ti oni ng Pl ants Usi ng DEA and MDEA, Proceedi ngs of the 1982
Gas Condi ti oni ng Conference.
HYDROCARBON PROCESSING 7
The authors
Mike DuPart is a development leader in the
GAS/SPEC Technology Group of Dow Chemical
USA. He has a wide range of experience in gas
processing from corrosion inhibitor research to
process optimization of amine plants. He holds
four U.S. patents and has coauthored several
papers on gas processing. Mr. DuPart holds a BS
degree in metallurgical engineering from the
University of Texas at El Paso.
Tom Bacon is a project manager for the
GAS/SPEC Technology Group of Dow Chemical,
USA. His responsibilities focus on development
and use of formulated amine solvents and
licensed technologies for removing acid gas impu-
rities from liquid and gas streams. He has 17
years of varied experience at Dow in research,
manufacturing, process engineering, capital pro-
ject management, and technical service and
development. He holds a BS degree in chemical
engineering from Texas A&M.
David J . Edwards is a senior development engi-
neer for the GAS/SPEC Technology Group of
Dow Chemical, USA. He is responsible for tech-
nical service and development of MDEA-based
formulated amine solvents used to remove acid
gas impurities from natural gas and synthesis
gas streams. Prior to joining GAS/SPEC in 1989,
he had 6 years of experience in technical sales
of ion exchange resins for Dow Chemical USA.
He holds a BS degree in chemical engineering
from Mississippi State University.
Most cor r osi on
f ai l ur es i n
al kanol ami ne
ser vi ce ar e
at t r i but ed t o
f r ee aci d gas
and hi gh
t emper at ur e.
8 HYDROCARBON PROCESSING
M. S. DuPart, T. R. Bacon and D. J . Edwards,
Dow Chemi cal USA, Freeport, Texas
P
roper di agnosi s i s hal f the probl em and sol uti on i n
al kanol ami ne corrosi on. I n Part 1, several com-
mon corrosi on mechani sms were i denti fi ed wi th
ami ne systems. Wi th many possi bl e mechani sms to choose
from, proper corrosi on di agnosi s i s a cumbersome task.
Unfortunatel y, the desi gner or the operati ons engi neer
must knowl edgeabl y si ft through the pl ant data and cor-
rectl y match the symptoms or observati ons wi th the prob-
abl e cause and prescri be correcti ve acti on(s).
The case hi stori es revi ew actual pl ant si tuati ons and
the observati ons from i nspecti on. They gi ve great i nsi ght
i nto a costl y pr obl em. The exampl es ar e defi ni te pl ant
notebook addi ti ons for l ater references.
As in many processes, better designs reduce costly down-
ti me and l ost producti on, i mprove equi pment l i fe and mi n-
i mi ze oper ati on acci dents or i njur i es. A better under -
standi ng of parameters affecti ng corrosi on wi th ami nes
present i mproves process operati ons and al l ows a better
ami ne sel ecti on.
Amine and process parameters. Careful sel ecti on of
ami ne and process parameters can have a posi ti ve i mpact
when reduci ng corrosi on i nci dence. Si nce free aci d gas i s
the pr i mar y contr i butor to cor r osi on, a br i ef di scussi on
about the actual mechani sm i s appr opr i ate. CO
2
cor r o-
si on i n ami ne uni ts i s caused by the reducti on of undi s-
soci ated carboni c aci d i n turbul ent areas where a passi ve
ferrous carbonate scal e i s unabl e to be l ai d down.
23
The
carboni c aci d can undergo any one of these reacti ons:
H
2
CO
3
+ e
c HCO
3