• No results found

Corrosion Behavior of Various Steels for Compression, Transport and Injection for Carbon Capture and Storage

N/A
N/A
Protected

Academic year: 2021

Share "Corrosion Behavior of Various Steels for Compression, Transport and Injection for Carbon Capture and Storage"

Copied!
10
0
0

Loading.... (view fulltext now)

Full text

(1)

Energy Procedia 23 ( 2012 ) 216 – 225

1876-6102 © 2012 The Authors. Published by Elsevier Ltd. Selection and/or peer-review under responsibility of SINTEF Energi AS doi: 10.1016/j.egypro.2012.06.074

TCCSͲ6

CorrosionBehaviorofVariousSteelsforCompression,

TransportandInjectionforCarbonCaptureandStorage

AkiSebastianRuhl,ArturGoebel,AxelKranzmann

BAMFederalInstituteforMaterialsResearchandTesting,Division5.1,UnterdenEichen87,12205Berlin,Germany eͲMail:akiͲ[email protected] Abstract

Steels used within the process chain of carbon capture and storage (CCS) are exposed to conditions that are currentlynotfullytested.Inthepresentworkanumberofsteelswereselectedaspossibleconstructionmaterials: Alloyed steels for application in the compression and injection sections and low alloyed carbon steels for use as pipelines.Exposuretestswereconductedover600hoursatambientpressureinacontinuousflowofasimulated gas stream consisting of carbon dioxide with low contents of the relevant flue gas components nitrogen dioxide, sulfurdioxide,carbonmonoxide,oxygenandwater.Temperatureswereadjustedto5,60and170centigrade.



©2010PublishedbyElsevierLtd.Selectionand/orpeerͲreviewunderresponsibilityof[nameorganizer] Keywords:CCS;materialselection;gascorrosion;dewpointcorrosion;compression;pipeline;injection 1. Introduction Carboncaptureandstorage(CCS)isconsideredasnecessarytechnologytoreducefurtherincreaseof greenhousegases.Carbondioxide(CO2)iscaptureddirectlyatthesource,andmustbetransmittedto

the receiving geological formation. Steels along the whole process chain come in contact with the purified fluid consisting of CO2 and minor amounts of flue gas residuals such as oxygen (O2), sulfur

dioxide(SO2),nitrogendioxide(NO2),carbonmonoxide(CO)andwater(H2O).Corrosionofsteelsmight

beariskwithintheCCSchain,dependingontheconcentrationsofimpurities.

Compressedcarbondioxidewithverylowconcentrationsofimpuritieshasbeentransmittedthrough pipelinesforenhancedoilrecovery[1].Onlyalimitednumberofcorrosionexperimentsinsupercritical CO2 have been reported yet. Russick et al. [2] reported that corrosion occurs in water saturated

supercritical CO2 but not in pure supercritical CO2. In the presence of O2 and SO2 corrosion rates are

stronglyincreased[3Ͳ4].However,experimentsunderpressureareperformedunderstaticconditions.

© 2012 The Authors. Published by Elsevier Ltd. Selection and/or peer-review under responsibility of SINTEF Energi AS

Open access under CC BY-NC-ND license.

(2)

Once the autoclave is filled and pressurized with a certain molar amount of corrosive constituents, furtheradditionorreplacementofimpuritiesisverycomplicated.Corrosiveconstituentsareconsumed duetoreactionswiththesteel.Inthepresentworkcontinuousexperimentswithaconstantsupplyof impuritieswereconductedtoallowaccumulationofcorrosionproducts.Preliminaryobservationswere discussed elsewhere [5]. Detailed analyses of corrosion products were performed and the results are presentedanddiscussedinthispaper.

2. MethodsandMaterials

2.1. ExperimentalSetup

Customized exposure vessels were designed and constructed for parallel testing of up to 12 test specimensunderacontinuousflowfromtoptobottom.ThreereactorvesselsareshowninFig.1oneof whichisalreadyequippedwith12specimensofdifferentsteels.SamplesareconnectedtotheTeflon holders in the middle of the reactor. The Teflon holder is surrounded by a glass wall that allows observationofthespecimensduringtheexposureperiod.





Fig.1.Photographof3exposurevesselsforparalleltestingof12specimensundercontinuousflow

Temperatures of the exposure vessels were adjusted by keeping the vessels in climate chambers (typesIPPfor5°CandUFP500for60°Cand170°C,MemmertGmbH,Germany).Exposuretemperatures were adjusted to 170°C for conditions present in the compression process, 5°C as lowest occurring temperatureinasubsurfacepipeline,and60°Cfortheinjectionnozzle.

Mass flow controllers (type ELͲFLOW, Bronkhorst Mättig GmbH, Germany) were used to adjust volumeflowsofCO2,O2,CO,NO2andSO2.Thevolumeflowofcarbondioxidewas1.5L/hperreaction vessel(equivalentto20cm/min)withanadditionofvolumeflowsequivalentto1800ppmO2,750ppm CO,100ppmNO2,and70ppmSO2.Thegasesweremixedinastaticmixer.Sincethevaporpressureof NO2isverylowthegascylinder,pipes,tubesandthemassflowcontrollerwereheatedto60°Ctoavoid liquefaction. A water vapor concentration of 600 ppm (equivalent to a dew point of ca. Ͳ28°C) was adjustedbyleadingacontrollablefractionoftheCO2flowthroughawashingbottlefilledwithdeionized

(3)

water according to the simultaneous measurement with a dew point meter (type EASIDEW Online Hygrometer,MichellInstrumentsGmbH,Germany).Thedurationofexposurewas600hours.

2.2. Steels

Different steels with varying concentrations of chromium and other alloying elements were tested. ThealloyedandhighalloyedsteelsX12Cr13,X3CrNiMo13Ͳ4,X5CrNiCuNb16Ͳ4,X1NiCrMoCu32Ͳ28Ͳ7and titanium TiͲAlͲV4 (No. 3.7165) were tested at 170°C for application in compressors. Steels 42CrMo4, X20Cr13,X46Cr13,X5CrNiCuNb16Ͳ4,X2CrMnNiN22Ͳ5Ͳ2andX1NiCrMoCu32Ͳ28Ͳ7wereproposedforuse in the injection section and tested at 60°C. Soft iron as reference material and the pipeline steels L290NB,L360NBandL485MBwereexposedtothegasmixtureat5°C.Materialnumbersandchemical compositionsofthesteelsarelistedinTable1.



Table1:Materialnumbersandfractionsofselectedchemicalelementsofthesteelsdeterminedbysparkemissionspectroscopy. Steel MaterialNo. C[wt%] Co[wt%] Cr[wt%] Mn[wt%] Mo[wt%] Ni[wt%]

X12Cr13 1.4006 0,13 0,02 13,93 0,62 0,05 0,47 X3CrNiMo13Ͳ4 1.4313 0,04 0,03 13,16 0,81 0,53 3,88 X5CrNiCuNb16Ͳ4 1.4542 0,06 0,07 15,91 1,00 0,22 4,85 X1NiCrMoCu32Ͳ28Ͳ7 1.4562 0,01 0,21 26,60 1,52 6,11 31,27 42CrMo4 1.7225 0,40 0,01 1,02 0,68 0,14 0,12 X20Cr13 1.4021 0,23 0,02 13,19 0,38 0,02 0,12 X46Cr13 1.4034 0,49 0,02 13,41 0,49 0,13 0,45 X2CrMnNiN22Ͳ5Ͳ2 1.4162 0,02 0,03 21,72 4,96 0,30 1,63 Softiron 1.1018 0,00 0,00 0,01 0,04 0,00 0,02 L290NB 1.0484 0,12 0,01 0,10 1,11 0,03 0,06 L360NB 1.0582 0,13 0,00 0,03 1,30 0,00 0,03 L485MB 1.8977 0,09 0,01 0,03 1,54 0,00 0,05 

All materials were machined to a specialized specimen design (compare Fig. 2) that allowed direct connectiontothemountings.Alltestspecimensweregrindedonadiamondgrindingdisc(StruersMDͲ Piano 220, grit size 220) and cleaned in deionized water with ultrasonic to achieve comparable and reproduciblesurfaceproperties.Thesurfaceareaofthetestspecimenswasapproximately19cm2.

3. Resultsanddiscussion

3.1. Macroscopicinvestigation

Slightannealingcolorswerefoundonalloyedchromiumsteelsafter600hexposuretothesimulated CCSgasat170°C.AsshowninFig.2thehighalloyedsteelandtitaniumwereunalteredcomparedtothe initial state. Similar photographs of duplicate specimens of the steels X1NiCrMoCu32Ͳ28Ͳ7 and X3CrNiMo13Ͳ4werepreviouslyshown[5].Gravimetricanalysesconfirmedthatnoalterationintermsof

(4)

materiallossorweightincreaseoccurred.Thereforereactionratesbetweenthegasphaseandthesteels canbeconsideredasveryslow. 



Fig.2.X1NiCrMoCu32Ͳ28Ͳ7,X12Cr13,X3CrNiMo13Ͳ4,X5CrNiCuNb16Ͳ4andtitanium(fromlefttoright)after600hexposuretothe simulatedCCSgasat170°C.

Photographs of the test specimens are shown in Fig. 3. The relative humidity at 60°C was approximately0.3%.Neithercorrosionnorannealingcolorswereobserved.Allinvestigatedsteelswere unaltered.Bothlightandscanningelectronmicroscopyconfirmedthatnocorrosionoccurredunderthe appliedconditions.Neitherweightincreasenordecreasewasmeasured. 



Fig3.PhotographsofspecimensofsteelsX20Cr13,X46Cr13,42CrMo4,X2CrMnNiN22Ͳ5Ͳ2,X5CrNiCuNb16Ͳ4andX1NiCrMoCu32Ͳ 28Ͳ7(fromlefttoright)after600hexposureat60°C.

(5)

Clear corrosion products developed on unalloyed materials at 5°C. Despite differences in both elemental composition and microstructures the appearance of the four corroded steels was similar as showninFig.4.DuplicatesamplesofsoftironandlowalloyedcarbonsteelL485MBshownelsewhere weresimilar[5],thusthecorrosionproductsformationiswellreproducible. 



Fig.4.SoftironandlowalloyedsteelsL290NB,L360NBandL485MB(fromlefttoright)after600hexposuretothesimulatedCCS gasat5°C.

A mass increase was found due to the corrosion of the steels with formation of ferrous or ferric corrosion products. The gravimetric results shown in Fig. 5 indicate that the degree of corrosion was similarforthedifferentsteels.Amajordifferencebetweenthesteelswasthesizeoftheferritegrains andthepearliteinclusions.Themicrostructurethereforedidnotinfluencecorrosionrates.Theweight increaseinregardtosurfaceareawasapproximately0.8mg/cm2. 



Fig.5.Weightincreaseofduplicatesamplesofsoftironandlowalloyedpipelinesteelsafter600hexposuretothesimulatedgas mixtureat5°C. Consideringanaverageweightincreaseof13mgandafractionof20wt%ironinthecorrosionbuildͲ up(compareFig.10),anamountof2.6mgsteelhasreacted.Thevolumespecificmateriallosswas0.3 mm3(0.017mm3/cm2)within600hassumingasteeldensityof7.8mg/mm3.Furtherthethicknessloss was 17 μm within 600 hours. Linear extrapolation revealed a material loss of less than 0.002 mm per year.

3.2. Microscopicinvestigation

AmicroscopicoverviewofthecorrosionlayeronthelowalloyedcarbonsteelL485MBisshowninFig. 6. The layer reveals a uniform appearance with regular cracks. Identical corrosion products were observedonalllowalloyedsteels.Lópezetal.[6]reportedaninfluenceofmicrostructureandchemical

(6)

composition of carbon and low alloyed steels. The different grain sizes in the steels did not have a significant effect on the structure of the corrosion layer, too. Therefore a similar buildͲup can be expected on any low alloyed pipeline steel, independently from its microstructure. Punctual spots are distributed above the flawy layer. A similar appearance was presented for simulated atmospheric corrosion of an unalloyed steel after one week exposure to a gas containing 10 ppm SO2 and 90%

relativehumidityat25°C[7]. 



Fig.6.MicroscopicstructurerecordedwithscanningelectronimageofthecorrosionlayeronsteelL485MB.

Small crystalline spots became visible at higher resolutions in Fig. 7 (left). As shown below those crystals were not detected with XͲray or electron diffraction. The main fraction of the corrosion layer appeareddenseandnotcrystallineasshowninFig.7(right). 



Fig.7.Detailsofthecorrosionlayershowingcrystallinestructuresontopofthelayerandanamorphousstructurewithinacrack. ElementaldistributionsobtainedwithenergydispersiveXͲrayspectroscopyforasmallareaincross sectionofthecorrosionlayerareshowninFig.8.Carbonisthemainelementoftheresinusedtofuse thespecimen.Comparingthecorrosionlayerwiththesteelironisamainelementofthecorrosionlayer. OxygenisalsoinvolvedinthebuildͲup.Sulfurisequallydistributedinthecorrosionlayerwithoutstrata

(7)

of elevated sulfur content. Considering the elemental distributions ferrous or ferric sulfates or sulfites arepossiblecrystallinephases. 



Fig.8.CrosssectionofthecorrosionlayeronsteelL290NBwithelementaldistributionofiron(Fe),carbon(C),oxygen(O),and sulfur(S)fortheindicatedarea. Focusedionbeam(FIB)wasappliedtoachieveanothertransectimageoftheinterfacebetweensteel and corrosion buildup. The corrosion layer between the ferriticͲpearlitic steel below and a protective platinumdeposition above are shown inFig. 9. Here thecorrosion layer appeared thinner than in the metallographictransect.





Fig.9.IonimageaftertreatmentwithfocusedionbeamshowingthesteelL290NB,thecorrosionlayerandthedepositedplatinum layer.

(8)

Quantitative analyses were conducted with electron beam microanalyses (WDX) in cross section of steelL290NBfusedinepoxyresin.Oxygenwasthemaincomponentinthelayerfollowedbyiron.Asthe molarfractionofironwasmuchhigherthanthatofsulfur,ferroussulfatesorsulfitesarenotexpected to be the main phase. Johansson and Vannerberg [8] found ɲͲFeSO3ͼ3H2O, ɶͲFeSO3ͼ3H2O, FeSO3ͼ2H2O

andFeSO4ͼ7H2Oasmajorproductsofatmosphericcorrosioninthepresenceofsulfurdioxide. 



Fig.10.WavelengthdispersiveelementalquantificationofthecorrosionlayeronsteelL290NB. Themolefractionsofiron,oxygen,sulfurandnitrogenforthreemeasuringpointsareshowninFig. 10.Theelementalcompositionindicatedthepresenceofamixtureofironoxide,ironhydroxide,sulfates orsulfitesandprobablycrystallizationwater. 3.3. Phaseanalyses

XͲray diffractometric analysis on a test specimen of steel L485MB covered with a corrosion layer revealedthatthegradeofcrystallinityislow.Thegrazingincidencediffractionwithanincidentangleof 1°wasassumedtocoverawiderangeofthecorrosionlayer.Diffractionpatternsofbothconventional andgrazingincidentmeasurementsareshowninFig.11.OnlycrystallineferriteoftheferriticͲpearlitic steelwasfound. 



Fig.11.XͲraydiffractionpatternofasteelspecimenL485MBcoveredwithcorrosionproductsusingtwodifferentmodes. AnimageanddiffractionpatternsobtainedwithtransmissionelectronmicroscopyareshowninFig. 12.Crystallinecorrosionproductswerefoundafteratmosphericcorrosionwithlowerconcentrationsof SO2 and NO2 [8]. Here, low crystallinity of the corrosion layer was found with the highly resolving

(9)

electronbeamleadingtotheconclusionthatthecorrosionproductsaremainlyamorphous.Amorphous corrosionproductswithcracksareexpectedtobelessprotectivethanacontinuouscrystallinelayer. 



Fig.12.TEMimage(a)takenonaFIBlamellatakenfromcorrodedsteelL290NBandelectrondiffractionpatternsofb)corrosion layerandc)ferriteinthesteel. 4. Conclusions Slightannealingcolorswerefoundonalloyedsteelsafter600hexposuretoasimulatedgasmixture at 170°C and ambient pressure. Weight alteration was not significant. Thus steels tested for use in compressionfacilitiesareuseableunderconditionssimulatedhere.

At 60°C no indications of reactions between the gas phase and the steels were found. Under the conditionsappliedhere,noneofthetestedmaterialscanbeexcludedduetosusceptibilitytocorrosion. WhenCO2isinjectedintoanaquiferhigherwaterconcentrationsmightoccurattheinjectionnozzles.

Visiblecorrosionproductsdevelopedonsoftironandlowalloyedpipelinesteelsat5°C.Thematerial losswasextrapolatedtobelowerthan2μmperyearandthereforebelowthethresholdlimitvalueof 0.1mmperyear.Nosignificantinfluenceofmicrostructureorelementalcompositionofthesteelswas observed. Although a homogenous distribution of iron, oxygen and sulfur in the corrosion layer was detectedbyEDXandsignificantamountsofoxygen,ironandsulfurwerefoundwithWDX,nocrystalline phases were detected with both XͲray diffraction at grazing incident and electron diffraction with transmission electron microscopy. The corrosion layer can therefore be considered as amorphous and protective only to a low degree. Continuous exposure to gas mixtures under elevated pressures will providefurtherinsights.

Acknowledgements

The project is supported by the Federal Ministry of Economics and Technology on the basis of a decisionbytheGermanBundestag(contract03277902)andthirdpartyfundingbyAlstom,EnBW,EON, Vattenfall Europe and VNG. We thank Dr. Bettge for the photograph of the reaction vessel and Dr. Bettge,Mr.SimonandMr.Bohrausfortheircontributionstotheexperimentalsetup,organizingsteel

(10)

specimensanddiscussions.WearegratefultoMs.HoffmannformetallographicinvestigationsandMs. Oder for WDX elemental quantification. Dr. Dieter is acknowledged for XRD analyses and Mr. Saliwan Neumann for SEM analyses. We thank Ms. Rooch, Mr. Gesatzke and Dr. Oesterle for FIB specimen preparation and TEM investigation. Dr. Bäßler and Dr. Yevtushenko are acknowledged for valuable discussion. References [1]J.Gale,J.Davison,TransmissionofCO2—safetyandeconomicconsiderations,Energy29(2004)1319–1328. [2]E.Russick,G.A.Poulter,C.L.J.Adkins,N.R.Sorensen,Corrosiveeffectsofsupercriticalcarbondioxideandcosolventsonmetals, TheJournalofSupercriticalFluids,9(1996)43Ͳ50. [3]Y.ͲS.Choi,S.Nesic,D.Young,EffectofimpuritiesonthecorrosionbehaviourofCO2transmissionpipelinesteelinsupercritical CO2Ͳwaterenvironments,EnvironmentalScienceandTechnology,44(2010)9233Ͳ9238 [4]Y.ͲS.Choi,S.Nesic,D.Young,EffectofimpuritiesonthecorrosionbehaviourofCO2transmissionpipelinesteelinsupercritical CO2Ͳwaterenvironments,EnvironmentalScienceandTechnology,45(2011)3813Ͳ3818 [5]A.Kranzmann,T.Neddemeyer,A.S.Ruhl,D.Huenert,D.Bettge,G.Oder,R.S.Neumann,Thechallengeinunderstandingthe corrosionmechanismsunderoxyfuelcombustionconditions,InternationalJournalofGreenhouseGasControl,5(2011)S168Ͳ S178. [6]D.A.Lopez,T.Perez,S.N.Simison,Theinfluenceofmicrostructureandchemicalcompositionofcarbonandlowalloysteelsin CO2corrosion.AstateͲofͲtheͲartappraisal,Materials&Design,24(2003)561Ͳ575. [7]S.Oesch,EnvironmentaleffectsonmaterialsEffectofSO2onthecorrosionofunalloyedsteelandweatheringsteelͲAn attempttosimulateatmosphericcorrosion,WerkstoffeUndKorrosionͲMaterialsandCorrosion,47(1996)505Ͳ510. [8]L.G.Johansson,N.G.Vannerberg,TheCorrosionofUnprotectedSteelinanInertͲGasAtmosphereContainingWaterͲVapor, Oxygen,NitrogenandDifferentAmountsofSulfurͲDioxideandCarbonͲDioxide,CorrosionScience,21(1981)863Ͳ876.

References

Related documents

Papson, Brent Aaron, "Predictors of Participation and the Effects of Undergraduates' Involvement in a First-Year Emerging Leadership Program"

For each class meeting, students are required to read the assigned unit from the African Folklore Workbook. Each unit text is available on the class BlackBoard website under

In this paper we consider bounded degree bipartite graphs that have a small separator and large bandwidth, and prove that under reasonable con- ditions these are spanning subgraphs

“SMART Practice Aids – Risk Assessment offers links into professional standards and PPC guidance, reporting of audit program changes, one-click sign-offs for programs and

The TatA component of the twin- arginine translocation system locally weakens the cytoplasmic membrane of Escherichia coli upon protein substrate binding.. Sargent F, Berks

The 5704F Fire control cards provide display and alarm facilities for a wide variety of fire detection products and provides up to four fire zone inputs compatible with most flame,

Nine of the 10 patients with CXR-detected lung me- tastases were treated with curative intent by surgery only (n = 4) or surgery and chemotherapy (n = 5) and 5 remain

The changes were regular until 2005, not exceeding 1%, but between 2006 and 2030, a dramatic increase of the ratio can take place (by over 22%), which is caused by a great number