CONDENSER

NITROGEN

HYDROGEN

NITROGEN

HYDROGEN.CARBONDIOXIDE

crLUmccOcom

HOTRICHMEA

"CARBON DIOXIDE

COOLRICH MEA r

MEA«HHEATEXCHANGERS

ccLUQ_CL

HOT..LEAN MEA'

REFLUX ACCUMULATOR

35-40 psigSATURATED

STEAM

/

J

150-175 psigSATURATED

STEAM

RECLAIMERl

HOT LEAN MEA

SEWERCONDENSATE

CONDENSATE

Figure I. Process flow diagram of a typical monoethanolamine carbon dioxide scrubbing unit.

Corrosion ControlIn C02 Removal Systems

Where used, "Amine Guard" has completely eliminated corrosion as an operating problem atdesign and above design conditions in carbon dioxide removal systems.

K. F. Butwell, E. N. Hawkes.and B. F. Mago, Union Carbide Corp., Tarrytown, N.Y.

In 1967, Union Carbide initiated a study on corrosionand corrosion inhibition for the monoethanolamine,water, carbon dioxide-steel system. The experimentalprogram consisted of weight-loss testing, potentiostaticpolarization experiments, electrode potentialmeasurements under self-polarization conditions, and

• characterization of surface barriers. The results of thisprogram provided the experimenters with a better un-derstanding of the corrosion mechanisms involved andallowed them to discover an effective inhibitor for thesystem, Union Carbide's "Amine Guard".

To provide a perspective for discussing our ex-perimental program and its results, we will first reviewthe monoethanolamine-water-carbon dioxide removalsystem that was used before these studies were un-dertaken, Figure 1. Very briefly, a 20% by weightmonoethanolamine solution was circulated at such a ratethat the rich solution exiting the absorber containedapproximately 0.45- to 0.50 moles of carbon dioxide permole of monoethanolamine (mol./mol.) with reboiler heatduty set such that the lean solution exiting theregenerator contained 0.10- to 0.15 (mol./mol.). Theprincipal material of construction for this system was304-316 stainless steel.

Improving the designThis design, although excellent in concept and in

operation, showed many areas for improvement,provided corrosion could be eliminated as a design andoperating constraint. One of the major i areas for im-provement lie in the elimination of alloy materials ofconstruction; hence, our choice of carbon steel for themajority of the experimental program. It should be noted

Table 1. inhibitor performancein accelerated laboratory tests.

Additive to 65% MEA SolutionPresaturated with Carbon Dioxide

Additive A Additive B Additive CRange of % Protection

Given to Mild Steel Panels

0.1 0.01 . .0 00.05 0.0050.05 0.0050 0 0.05 70-990.025 0.0025 . . . 0.025 99±1

0 0-160.1 42-990.05 99±10 0

50

that we also included stainless steel in our program sinceprocess gas(synthesis gas) is used as a heat source incertain carbon dioxide removal systems. Acid-gascondensation occurs during heat transfer; hence, theneed for 304-316 stainless steel.

The second area of improvement lies in the reductionin the size of the equipment by operating at highmonoethanolamine concentrations, higher acid-gasloadings, and lower circulation rates. Needless to say;operating at the new conditions would result in asignificant reduction in pumping and reboiler heat dutyoperating costs.

The third opportunity for improvement involved theremoval of the reclaimer to change the unit into a closedoperating system.

In brief, our goal was to develop an inhibitor systemthat would apply not only to existing carbon dioxideremoval units, but would also serve as the prime factor in

the design and operation of future units.For simplicity, our experimental program was divided

into three phases:1. A detailed literature search and several com-

prehensive discussions with experienced acid-gasremoval system personnel.

2. A fundamental study of the corrosion mechanismsinvolved in the aforementioned system.

3. An evaluation of potential corrosion inhibitors.The literature search, subsequent discussions, and

laboratory evaluations suggested the following points:1. In the absence of acidic components (e.g., carbon

dioxide and hydrogen sulfide), aqueous monethanol-amine solutions do not cause serious attack of steel in theliquid phase.

2. With mixtures of two common gases, carbondioxide and hydrogen sulfide, the corrosion behaviormay be relatively complex.

Table 2. Corrosometer probe data.

Before Inhibitor Addition

Date

12-112-212-212-712-8

Probe ElementDial

ReadingCorrosion Rate

mils/yr.

Mild Steel 59.5Mild Steel 170Mild Steel 209.5Mild Steel 905Mild Steel 1000+

1-30 304 Stainless Steel.1-312-32-42-62-72-10 304 Stainless Steel2-17 304 Stainless Steel

55304 Stainless Steel 56304 Stainless Steel 58304 Stainless Steel 59304 Stainless Steel 59.5304 Stainless Steel 60

6163

2-19 304 Stainless Steel 632-25 304 Stainless Steel 653-3 304 Stainless Steel 68

After Inhibitor Addition

3-6 304 Stainless Steel 70.53-10 304 Stainless Steel 70.53-11 304 Stainless Steel 70.53-12 304 Stainless Steel 703-13 304 Stainless Steel 70

0604.8532536.8297

03.652.343.650.911.751.231.0401.221.83

00000

3-133-133-133-14 Mild Steel 423-17 Mild Steel 42

Mild Steel " 41.5 corrosometerMild Steel 43 had notMild Steel 42.5 stabilized

00

3-183-203-213-243-263-263-284-34-10 Mild Steel

Mild Steel 41.5, Mild Steel 41.5. Mild Steel 41.5, Mild Steel 41.5, Mild Steel 41.5.'Mild Steel 41.5. Mild Steel 41.5. Mild Steel 41.5

42

4-134-214-305-35-165-306-16

Mild Steel 41.5Mild Steel 41.5Mild Steel 41.5Mild Steel 41.5Mild Steel 41.5Mild Steel 41.5Mild Steel 41.5

000000000.260000000

51

Table 3. Mild steel corrosion coupon data.

Before Inhibitor Addition

Coupon No. Days in Service Weight Loss, g.Corrosion Rate

mils/yr.

DU - 368 14 6.0789 429369 14 4.2665 301370 5 2.3303 460371 5 2.1392 423372 5 2.1911 433373 . . , 5 2.0377 403

After Inhibitor Addition

Total Days Service Days Weight Loss, g.. Corrosion RateCoupon No. in Service Between Cleanings Between Examinations mils/yr.

DU - 394 9 9 0.0027 0.3395 9 9 0.0033 0.4396 9 9 0.0027 0.3397 9 9 0.0033 0.4394 16 7 0.0019 0.3395 16 7 0.0017 0.2.396 16 7 0.0021 0.3397 16 7 0.0015 0.2394 24 8 0.0016 0.2395 24 8 0.0003 0.04396 24 8 0.0014 0.2397 24 8 ....0.0010 0.1394 34 10 „0.0015 0.1395 34 10 0.0015 0.1396 34 10 0.0008 0.1397 34 10 0.0015 0.1394 48 14 0.0011 0.10395 48 14 0.0017 0.12396 48 14 0.0009 0.06397 48 14 0.0012 0.08394 66 18 0.0000 0395 66 18 0.0000 0396 66 18 0.0001 0397 66 18 0.0002 0

*The test coupons used had a surface area of 18 sq. cm. and weighed approximately 10.5 g..

350 ton / day ammonia plant operated by C. F. Industries in Terre Haute, Ind.

52

Table 4. Mild steel corrosion coupon data.

Coupon No.

123

Rich Stream Before Inhibitor Addition

Days in Service

547

Corrosion Ratemils/yr.1,2601,1791,356

Rich Stream After Inhibitor .Addition

Coupon No.1 ....234 ...56

Days in Service.... 5 . .

4 .142828 . ....28

Corrosion Ratemils/yr.158

. . . 1883.60.20.020.02

Coupon No.

123

Reboiler Overhead Before Inhibitor Addition

Days in Service

547

Corrosion Ratemils/yr.59.067.576.0

Coupon No.123

Reboiler Overhead After Inhibitor Addition

Days in Service54

14282828

Corrosion Ratemils/yr.

0.260.200.080.020.020.02

3. Solution temperatures may approach 300°F andhigh metal surface temperatures are particularlydemanding with regard to inhibitor response.

4. Monoethanolamine solutions seem to become morecorrosive with use, perhaps for several reasons.

5. Monoethanolamine has good thermal stability inaqueous media; however, the carbonate salt can beconverted upon heating to other products, such as, N-(2-hydroxethyl) ethylene diamine (HEED). The greatercorrosivity of this conversion product has been verified.

6. Although corrosion is often of a general nature,insidious localized attack does sometimes occur. Thissuggests that a passive-active condition may occur,which could help explain some of the unpredictability ofthe corrosion problems.

7. In general, the corrosion of steel by carbon dioxide-containing monoethanolamine solutions is in agreementwith its fundamental electrochemical characteristics asindicated by potentiostatic studies.

8. Vapor phase as well as liquid phase corrosion mayoccur.

It is not within the scope of this article to discuss indetail the corrosion mechanisms involved in this system,however, it must be said that it was found that thesystem could indeed be passivated even under deaerationconditions as found in the normal acid-gas removalsystem.

Corrosion inhibitorsOur selection of inhibitors for this system involved

evaluation of three fundamental categories of inhibitors,

i.e., oxidant-type, precipitant-type, and adsorption-typeinhibitors. The results of this evaluation showed acombination of additives for corrosion control to be theoptimum, Table 1. Two points should be noted withrespect to this table. First, our test program involvedseveral hundred known additives and not one when usedalone afforded complete protection to mild steel testpanels. Second, additive C, when used alone, gave spor-adic protection (42- to 99%) to the test panels; yet,when combined with additives A and B at the properconcentration, resulted in an inhibitor "system" thatafforded complete protection to the test panels. Thisphenomenon is best described by the definition ofsynergism: the cooperative action of discrete agenciessuch that the total effect is greater than the sum of thetwo effects taken independently.

The last and most severe test of the efficacy of ourinhibitor "system" was conducted in a Parr Series 4500reactor with a monoethanolamine solution saturatedwith carbon dioxide. This experiment was conducted at300°F with pure carbon dioxide injected into the bomb tomaintain a constant pressure. The corrosion rate on mildsteel at these conditions was essentially nil.

The first commercial unit to utilize the inhibitor("Amine Guard") system was a 1000 ton/day ammoniaunit that had been in operation for approximately fouryears. At the time of injection, this unit was experiencingsome degree of corrosion as evidenced by corrosometer,coupon, and operating data. Table 2 contains thecorrosometer data before and after injection with thesedata being verified by visual inspection during ascheduled turn-around. Two comments on these data are

53

appropriate; first, one should note the efficacy of thisinhibitor with respect to carbon steel and, second, at-tention is drawn to the immediate reduction of corrosionafter injection of the inhibitor system.

Table 3 contains corrosion rate data for carbon steelcoupons located in the rich solution piping on thedischarge side of a hydraulic turbine. It is fully un-derstood by the authors that the corrosion rate on themild steel coupons is not representative of the corrosionrate with respect to the piping. However, a measurementof the coupon corrosion rates before and after injection atthis particular location does give one a good indication ofthe effectiveness of the inhibitor system.

Table 4 also contains corrosion rate data on carbonsteel coupons located in the rich solution pipingdownstream of the heat exchangers and in the reboileroverhead line to the stripper. These data were generatedin a 350/ton day ammonia plant. The data obtained atthe latter location are noteworthy in that this particularsection of piping, plus the stripper overhead piping andcondenser, were of prime concern with respect tocorrosion inhibition. The vapor pressures of the variouscomponents in our inhibitor system are relatively low,therefore, one would not expect the inhibitor system tobe in contact with the piping and vessels in these areas.However, it was found through coupon corrosion ratedata that these areas were afforded a significant degreeof corrosion inhibition. We have attributed thisphenomenon to entrainment.

As mentioned earlier, one of our goals was to operate aclosed system, i.e., one without a reclaimer. Our first testunit has been on line for approximately three and a halfyears without the reclaimer in service with no excessiveaccumulation of N-(2-hydroxyethyl) ethylene dia-mine (HEED), in organic residues, or heat-stable saltsin the circulating monoethanolamine solution. Solutionanalyses data from all other units employing "AmineGuard" are in agreement with these test data. Oneadditional comment with respect to the operation of aclosed system is warranted, though somewhat painful:the usage of monethanolamine in these systems has beenreduced by 40- to 60%.

In 1970, an optimization study was begun which in-corporated theoretical, laboratory, and operating data.This study was conducted primarily for carbon dioxideremoval systems in ammonia production. We are not atliberty at this particular time to discuss in detail theresults of our optimization program; however, we maygeneralize.

Presently, we have several units, including anessentially all carbon steel unit, on an optimizationprogram. This program consists of systematic increasesin monoethanolamine concentrations and acid-gasloadings, and decreases in solution circulation rates.These changes have resulted in significant reductions inreboiler heat duties which is our prime goal for existingunits. In addition to this, we have not evidenced anysignificant corrosion problems, foaming problems, orincreases in soltuion contaminants at the new operatingconditions.

In summaryCarbon dioxide removal systems utilizing inhibited

("Amine Guard") aqueous alkanolamine solutions havecompletely eliminated corrosion as an operating problemat design and above design conditions. A systematicprogram to determine the true value of the inhibitor is inprogress with favorable results to date. When complete,the data generated from this program, coupled withaccurate physical property data, should result in the

End view of MEA solution exchangers.

design of a monoethanolamine-carbon dioxide removalunit at substantially lower capital and operating costs.

AcknowledgmentWe would like to extend special thanks to the

Management and Operations personnel of the com-mercial test plants for their cooperation and support inevaluating this inhibiting system. Also, special thanksare due C. W. West and others for the basic research workthat resulted in the discovery of the inhibitors involved.*

B. F. Mago is a project scientist in charge ofthe Surface Phenomena and CorrosionControl Laboratory at the TarrytownTechnical Center of Union Carbide Corp. Heholds a B.S. degree in chemistry fromCanisius College.

E. N. Hawkes is a research chemist workingon the Alkanolamine Inhibitor Program atthe Tarrytown Technical Center of UnionCarbide Corp. He holds a B.A. in chemistryfrom the New York University.

K. F. Butwell is a project scientist in chargeof the gas purification and dehydrationproject at the Tarrytown Technical Center ofUnion Carbide Corp. He received hisB.S.Ch.E. degree from the University ofMichigan.

54

DISCUSSION

BUTWELL: There are three main factors that UnionCarbide takes into consideration with respect to theoptimization program: Point one, no unit utilizing theinhibitor has experienced a failure due to corrosion at thisparticular time. Secondly, when you make an operatingchange, for example, in a 1000 ton/day Ammonia CarbonDioxide Removal Unit, you must realize that this is not alaboratory test unit but an intergral part of an industrialcomplex that has a capital cost of perhaps 30 milliondollars. Therefore, a conservative approach should alwaysbe taken. The third factor, and I think we have finallyovercome this, is the precise steam and water balancearound these units. During normal operation, the steamcondensate is utilized in other areas of the complex. Thissource of water is essentially eliminated during theoptimization study as we have to vent the steam to theatmospehre; therefore, we have only six to eight hours todetermine the approximate mechanical and chemical limitsof the chemical limits of the carbon dioxide removalsystem. Once these limits have been established, condensingturbines can be put into service and our optimizationprogram can be continued to completion.JIM FINNERAN, M.W. Kellogg Co.: This developmentrepresents a very efficient way of chemical debottleneckingan ammonia plant. I believe this is the major application forthe development at this time. I have two questions I wouldlike to ask.

Would you care to comment on the MEA concentrationwith which you have experience at this time. Youmentioned 20% MEA as the starting point. Can you tell usthe MEA concentration actually used with your additive?My second question concenrs whether this inhibitor hasother applications. It occurs to me that there may be otherapplication, in other solvents, or other solutions, in whichthis inhibitor might be highly effective.BUTWELL: In answer to your second question: at thistime, we are actively engaged in five particular applicationsfor the "Amine Guard" inhibitor system:

1) We are currently utilizing the inhibitor system in afew natural gas processing plants where the hydrogensulfide feed concentration is less 5 ppm by volume.

2) It is in the developmental stage (semi-works) for theremoval of hydrogen sulfide and carbon dioxide froma sour naptha cracking operation.

3) We are finalizing our plans for the installation of theinhibitor system in a acid-gas removal unit associatedwith the Wulfe Process.

4) We are currently studying its application in theacid-gas removal system associated with a coke-oventype installation.

5) We are currently utilizing the inhibitor system in amonoethanolamine-carbon dioxide removal unit thatis very similar to the carbon dioxide removal facilityassociated with an LNG operation. The outlet carbondioxide specification concentration is maintained atless than 12ppm by volume.

The answer to your first question is somewhat complicatedby the fact that our inhibitor system is proprietary innature. In all fairness to our clients and ourselves, it wouldbe inappropriate at this particular time to discuss in detailthe operating data generated at various units during ouroptimization program. However, I will say that these

conditions are well above the design conditions.W.D. CLARK, ICI Britain. If I understood what you saidduring your paper, you had a large plant running andcorroding fairly badly, and you added the inhibitor to thesolution, and it stopped corroding.

BUTWELL: Yes sir.CLARK: Now this is rather unusual, because in most caseswhere a plant is corroding, it is heavily coated withcorrosion products and an inhibitor has difficulty in gettingdown to the metal surfaces. Did you, in fact, do anyextensive cleaning-out operations so that the inhibitor hadits best chance, or was it as simple as perhaps you implied.

BUTWELL: As mentioned earlier. Table 2 containsCorrosometer Probe data as a function of time. A review ofthis data shows the inhibitor system to be effective almostimmediately after injection of the inhibitor system.Presently, this is typical for all systems employing thisinhibitor system. However, it does not contain the inhibitordepletion rates as a function of time nor does it containsome of the anxious moments we at Carbide have hadduring the initial period (six weeks) of inhibition at a fewunits. Examination of metal surfaces at several unitsutilizing this inhibitor system shows our system to performas both a cleansing agent and a passivating agent. Inaddition to this, examinations of solid residues from theseunits show a positive test for the inhibitors. With thisbackground information, we prefer, naturally, to inject intoa unit that has been thoroughly cleaned. At theseconditions we show immediate passivation and a lowinhibitor depletion rate as expected. When we inject into asystem that is experiencing severe corrosion rates we expecta high inhibitor depletion rate for the simple reason thatour inhibitor will not only passivate the equipment metalsurfaces but also any iron particles floating around thesystem. In these particular instances we try to maintain asufficient quantity of our inhibitor in the solution such thatwe are able to continually passivate the equipment metalsurfaces. Presently, we have not experienced an equipmentfailure in these units.RALPH FREEMAN, Cherokee Nitrogen Co.: We'recurrently using a combination of sodium metavenadate andpetromeen 52 in our system. We have corrosion rates of lessthan one mil per year, and we do not use a side stream filteror a reclaimer. I'm interested in knowing whether thiswould in any way infringe upon the patents for yourcurrent procedure. We have been using this method forcorrosion protection for 5 years and as far as we're'concerned it is working very well for us. The equipmentthat we see Union Carbide portraying here seems to be verysimilar to the system we are using at Pryor, Oklahoma. Canyou tell us if there is any chance of a patent conflictbetween our corrosion inhibitors and the system you areoffering?BUTWELL: The gentleman is utilizing a combination ofinhibitors in his moneothanolamine-carbon dioxide removalsystem which he claims is effective. His question is: Are weinfringing on any of Union Carbide's patents?

As a chemical engineer from the Research andDevelopment Department, I am not qualified to answeryour particular question. A representative from our PatentDepartment could supply the answer.

55