1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

Catalytic Dry Reforming of Methane: Insights from ModelSystemsKnut Wittich,[a] Michael Krämer,[b] Nils Bottke,[b] and Stephan Andreas Schunk*[a]

Catalytic dry reforming under industrially relevant conditions ofhigh pressures and high temperatures poses severe challengestowards catalyst materials and process engineering. Thedemanding conditions under which the reaction is performedlead to a coupling of reactions occurring in the gas phase andreactions which are catalyzed by the material employed ascatalyst. A profound analysis of the mechanisms occurring ingas phase and resulting products from gas phase reactions iskey to understanding part of the challenges that any catalystmaterial, irrespective of its nature, will have to cope with. Thedeposition of coke on an active catalyst is as well one of themost limiting factors for catalyst lifetime and catalyst activity indry reforming. Therefore, an understanding of the thermody-namics behind coke formation and an intricate description ofthe mechanisms driving the evolution of coke is a vital piece of

the picture. Acid-base properties of the catalyst material andthe role and nature of the active metal do also need to beconsidered. A large part of the review deals with mechanismswhich are relevant for coke gasification and insights intomaterials properties, which are relevant to allow for reactionpathways along these lines. The review article focusses onresearch results which have been achieved using modelsystems – typically the analysis of model systems is a morerewarding exercise compared to fully formulated industrialcatalyst systems, as here more elucidating structure-propertyrelationships can be drawn. Additionally the article discussesdry methane reforming in the context of alternative syngasgeneration technologies and attempts to create an applicationperspective for the reader in the context of a sustainableapproach towards carbon capture and storage.

Introduction

Carbon dioxide management in terms of carbon capture,utilization and storage (CCUS) has become an important topic inpublic discussion and is reflected in efforts in research anddevelopment both in academia and industry in the last twodecades,[1] since CO2 is one of the major greenhouse gases.[2]

From an industrial point of view CO2 is not only a waste productwith a harmful environmental footprint, but also an interestingsynthon,[3] if available at a point source. Therefore, its utilizationand implementation in industrial scale as raw material intovalue chains offers certain potentials. Dry reforming is veryoften discussed in this context as a technology of choiceregarding CCUS. However, two major crucial aspects must beconsidered for any CO2 valorization technology, of course alsofor approaches in catalytic dry reforming, as discussed by Peterset al.[4] First CO2 is a highly stable molecule and can be

considered as the thermodynamic sink for many reactionsinvolving organic substrates; this results in the fact that anyprocess that intends to valorize CO2 and to introduce reactivitywill require enough energy input or reaction partners of asufficiently high intrinsic energy level. This plays a major role inthe design of any production process targeting carbon dioxidevalorization and the choice of reaction partners. Another aspectthat should be considered is the respective duration of the CO2

fixation and the value and lifecycle of the products made in therespective reaction; all these parameters play an important rolein the evaluation of any a CO2 valorization technology relyingon CCUS.

Of course, dry reforming is not the only conversiontechnology which utilizes carbon dioxide as synthon: a range ofprocesses for the chemical valorization of CO2 are already wellestablished on an industrial level. Among several small-scaleprocesses (e.g. Kolbe-Schmitt reaction[5]), two large-scale chem-ical processes that rely on CO2 as feedstock are of importance,namely the methanol production from H2 and CO2 (Eq. 1)[6] andthe production of urea from NH3 and CO2 (Eq. 2).[7]

3 H2 þ CO2 ! H3COHþ H2O (1)

2 NH3 þ CO2 ! COðNH2Þ2 þ H2O (2)

For methanol synthesis it can be argued that most of thetechnologies commercialized are process variants based on theconversion of syngas with a certain carbon dioxide content,typically around 2 to 6 vol% and adjusted hydrogen to carbonmonoxide ratios; carbon dioxide plays an important role in thereaction mechanism over copper based catalysts and occurs in

[a] Dr. K. Wittich, Dr. S. A. Schunkhte GmbHKurpfalzring 104Heidelberg 69123 (Germany)E-mail: stephan.schunk@hte-company.de

[b] Dr. M. Krämer, Dr. N. BottkeBASF SECarl-Bosch-Strasse 38Ludwigshafen am Rhein 67056 (Germany)

This manuscript is part of a Special Collection on the “French Conference onCatalysis”. Please follow the link for more articles in the collection.

© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.This is an open access article under the terms of the Creative CommonsAttribution Non-Commercial NoDerivs License, which permits use and dis-tribution in any medium, provided the original work is properly cited, the useis non-commercial and no modifications or adaptations are made.

ReviewsDOI: 10.1002/cctc.201902142

1ChemCatChem 2020, 12, 1–19 © 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Wiley VCH Freitag, 14.02.2020

2099 / 158042 [S. 1/19] 1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

the rate determining step as key critical component. Methanolsynthesis technologies for higher carbon dioxide partialpressures are under development and heavily sought after. Thementioned technologies and large-scale productions in placewhich utilize CO2 as starting materials rely on reaction partnershaving a high reduction potential, such as hydrogen itself(Eq. 1) or ammonia (Eq. 2) and therefore are currently mainlydependent on hydrogen supply from fossil sources obtainablevia reforming technologies.

CH4 þ H2O! COþ 3 H2 DH� ¼ 206:0 kJ=mol (3)

H2Oþ CO! CO2 þ H2 DH� ¼ � 41:2 kJ=mol (4)

As catalytic dry reforming is also part of the family ofreforming technologies, we want to take a closer look at thecurrent landscape which is industrially realized. The mosteconomic routes to hydrogen rely up to now on natural gas asfossil feedstock with the consequence of an inherent carbondioxide footprint.[8] The same is true for synthesis gas or syngas

that is in general an important chemical feedstock in theproduction of high-value chemicals and fuels.[9] The term syngasin this article is used in a context that refers to H2/CO mixturesof different ratios, which are accessible by the different syngasproduction technologies. The main state of the art large-scaleproduction technologies for hydrogen and syngas today arecatalytic and non-catalytic partial oxidation reactions (autother-mal reforming ATR, combined reforming CR, partial oxidationPOX and catalytic partial oxidation CPOX; Eq. 5), the steamreforming of methane (SMR; Eq. 3) with subsequent water-gas-shift stages (Eq. 4) and combinations thereof.[10,63]

CH4 þ1=2 O2 ! COþ 2 H2 DH� ¼ � 35:6 kJ=mol (5)

Dry reforming of methane (DRM; Eq. 6), a potential syngasproduction technology based on CO2 and methane, has drawna lot of attention, especially regarding raw material change,abundant gas supply, availability of bio-based gas and thedemand of technologies with the potential of carbon dioxideimport.[11,12,13] The term “dry” in dry reforming is based on the

Dr. Nils Bottke holds position of SeniorResearch Manager at BASF SE, leading petro-chemical catalyst development group. Hestudied chemistry in Würzburg, includingresearch visits at ENSI Caen, France andColumbia University, USA, and received hisPhD from the University of Würzburg. Workingat BASF for almost 20 years, Nils has heldvarious roles mainly in R&D and technology.One of his areas of expertise is the develop-ment of new processes supporting sustainablegrowth of the chemical industry based on rawmaterial change, alternative energy input andCO2 conversion. He was leading technicalteam in the development of BASF’s CarbonManagement Program, which targets thedevelopment of breakthrough technologiesfor low emission production. The focus of hisresearch is the development of heterogenouscatalysts for petrochemical processes rangingfrom gas reforming, conversion of syngas anddehydrogenation of hydrocarbons to selectivehydrogenations as well as electrocatalysis. Herepresents BASF in this field on the board ofChemistry and Energy division of the GDCh.

Dr. Michael Krämer is a Research Scientist atBASF SE in the petrochemical catalyst devel-opment group. He studied chemistry at Saar-land University where he received his PhD inTechnical Chemistry. After spending one yearas postdoctoral research fellow at the Institutode Tecnología Química, Valencia, Spain, Mi-chael joined BASF in 2008. During his firstyears at BASF, his main research fields werepartial oxidation and epoxidation reactions.Before joining the group of petrochemicalcatalysis, Michael worked in BASF‘s Digital-ization in R&D department to develop andimplement digital tools within the heteroge-neous catalysis research. His current work

focusses on catalysts for the production andconversion of syngas.

Dr. Stephan Andreas Schunk holds the posi-tion of an Executive Expert/Vice President atBASF SE and hte GmbH and is one of thefounding members of hte GmbH. He studiedchemistry at Mainz and Frankfurt Universityand holds a PhD in chemistry from theUniversity of Frankfurt. He has over 20 years ofexperience in heterogeneous and homogene-ous catalyst R&D. Stephan also teaches on aregular basis in his role as Lecturer at theDepartment of Technical Chemistry at theUniversity of Leipzig. His fields of research arepartial oxidation in gas and liquid phase,syngas production and conversion, use ofrenewable feedstocks, and the developmentof alternative concepts for materials synthesisto foster new approaches in heterogeneouscatalysis. Stephan has a special interest inadvancing research and development throughdigitalization and has contributed to thefoundation of NFDI4Cat, the digitalizationinitiative in catalysis and chemical engineeringof the German Catalysis Society.

Knut Wittich is a Postdoctoral Researcher inthe group of Dr. Stephan Schunk at hte GmbHin Heidelberg. He did his undergraduatestudies and his PhD in chemistry at RheinischeFriedrich-Wilhelms-University Bonn. During hisPhD at the department of inorganic chemistryhis research interest focused on synthesis andcharacterization of anhydrous phosphates asmaterials for heterogeneous catalysis, espe-cially in the field of partial oxidation of shortchain hydrocarbons. After his PhD he joinedhte GmbH in 2018. His fields of research areheterogeneous catalysis, synthesis and charac-terization of inorganic materials and solid-state chemistry.

Reviews

2ChemCatChem 2020, 12, 1–19 www.chemcatchem.org © 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Wiley VCH Freitag, 14.02.2020

2099 / 158042 [S. 2/19] 1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

substitution of water in steam reforming by CO2 (Eq. 3 and 4).The idea of producing synthesis gas from CO2 and CH4 goesback to as early as 1888.[14] Compared to SMR, DRM has severalattractive aspects as the process has benefits regarding energyefficiency since no or a largely reduced steam content must beproduced for the feedgas, therefore less energy be invested forthe evaporation of water. Additionally, DRM holds the valueproposition that a syngas can be prepared of low hydrogen tocarbon monoxide ratios, an exercise which must be muchharder to achieve starting from product gases obtained by SMR.

CH4 þ CO2 ! 2COþ 2 H2 DH� ¼ 247:3 kJ=mol (6)

Therefore, a lot of research on dry reforming was and still isconducted not only because of the aspects of CO2 utilizationbut also due to fact that access to CO rich syngas, that can beobtained from dry reforming, is facilitated. Easier access of COrich syngas offers potential for certain downstream processes.However, the large-scale production of CO-rich syngas by DRMis still in its infancy and has not been realized until now on abroad front.[15,16,17] The reasons for this state is that typically astrong deactivation tendency of the applied catalysts due tosevere coke formation is observed, especially at the industriallyrelevant pressure corridor of 20 to 40 bars. This phenomenonhas already been observed in 1928 by Fischer and Tropsch overNi and Co catalysts.[18] Numerous reviews on DRM have beenwritten in recent years which mainly deal with different types ofcatalysts that are said to be less affected by cokeformation.[11,19,20,21,22,23,24]

With our review we want to give an overview on insightsand findings that we consider relevant for DRM on an industrialscale and allow insights into our collaborative researchconducted on the challenging aspects of the DRM-reaction likecoke formation mechanisms under DRM conditions. We discussin this paper learnings and alternative approaches for thedesign of catalyst materials obtained for model systems toovercome some of the inherent challenges of the reaction inorder to build a basis for the large-scale application of dryreforming in industry.

As mentioned before, dry reforming of methane has onlyrecently reached a commercial scale;[25,26] up to now steamreforming (SMR), partial oxidation (POX) and autothermalreforming (ATR) are the dominating syngas productiontechnologies.[27] POX and SMR can be combined to use theindividual heat characteristics of both reactions. The combinedprocess is called autothermal reforming (ATR) if performed in asingle reactor. In all these processes, except DRM, a hydrogenricher syngas (H2/CO�2) can be obtained and especially thecombined technologies (ATR) can be used to obtain a broadrange of syngas ratios (Figure 1). In all reforming processes the(reverse) water gas shift (RWGS) plays an important role andaffects the obtained syngas ratio[28] (Eq. 4).

Steam reforming of methane is a well-established technol-ogy and favored for syngas demands of high H2/CO ratios(Eq. 3). The produced syngas has an H2/CO ratio of approx-imately 4 which can be further increased by WGS stages toobtain a pure hydrogen feedstock (99.95%).[29,30] The endother-

mic process runs at 800–950 °C and 15–40 bar over Ni-basedcatalysts. The steam-to-carbon ratio of the feed is held ataround 2.5–3 to ensure coke free operation conditions on thecatalyst. Since the invention of steam reforming of natural gas(Eq. 3) or higher hydrocarbons (Eq. 7) in 1912 extensive studiesof catalyst materials have shown that Ni is the pareto optimumof performance and price compared to noble metal catalysts inSMR. Noble metal catalysts feature higher coking and sinteringresistance due to thermodynamic reasons in equilibrium butare no economically attractive alternative in such a large-scaleprocess like SMR.

CmHn þm H2O! m COþ ðmþ n=2Þ H2 (7)

Via partial oxidation (POX) of methane a syngas ratiotypically of 2 is obtained (Eq. 8a) which is most suitable formost downstream processes, such as the direct methanolsynthesis[31] or Fischer-Tropsch synthesis.[32,33] The slightly exo-thermic process is operated without catalysts at 1150–1500 °Cand 25–80 bar[34] and requires high purity oxygen as feed andtherefore an oxygen plant in the form of an air separationunit.[35] Currently three POX processes are run on more than 320plants worldwide, the Texaco process, the Shell process and theLurgi process.[36] By applying a catalyst (Rh, Ru, Pt or Ni) in theCPOX process it could be shown that a reduction of the processtemperature to 900–1000 °C is feasible. Further lowering of theprocess temperatures to 500–600 °C at <5 bar seems to bepossible.[37] However, CPOX approaches can still be consideredas a technology under development.[38]

CH4 þ1=2 O2 ! COþ 2 H2 DH� ¼ � 35:6 kJ � mol� 1 (8a)

CmHn þm=2 O2 ! m COþ ðn=2Þ H2 (8b)

Complete substitution of the oxidants O2 (in POX) and H2O(in SMR) by CO2 in DRM (Eq. 6) directly produces a carbon richsyngas of H2/CO�2.

Figure 1. Syngas ratios of various syngas production technologies: dryreforming (DRM), partial oxidation (POX), autothermal reforming (ATR),combined reforming and steam reforming (SMR).[39]

Reviews

3ChemCatChem 2020, 12, 1–19 www.chemcatchem.org © 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Wiley VCH Freitag, 14.02.2020

2099 / 158042 [S. 3/19] 1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

However, the challenges of this reaction, besides beingslightly more endothermic than SMR, are as already mentionedthe severe coke formation tendency due to the carbon rich feedand the industrially relevant harsh reaction conditions temper-atures over 800 °C to allow for sufficient conversion andpressures of 10–40 bars to obtain a product gas at sufficientlyhigh pressure to avoid or minimize additional compressionsteps for downstream users.

Schwab et al.[39] discussed in 2015 the aspects of dryreforming for syngas production on a large scale. Theyconcluded that state of the art dry reforming technologies forsyngas production lacks new impulses and the current state ofthe art of catalysts used for SMR shows insufficient performanceunder industrially relevant DRM conditions due to deactivationmainly by coking at pressures above 20 bars.[40,41]

Up to now two processes closely related to DRM reachedcommercialization, the CALCOR[42,43] and SPARG[44] process. TheCALCOR process was designed to achieve a syngas compositionwith maximum CO content at smaller-scales of production. Thisprocess converts methane over Ni-based catalysts and circum-vents the coke formation by applying pressures close toatmospheric pressure and therefore allows the use of high CO2

partial pressures.[45,46] Thereby a CO rich syngas with a lowerlimit of H2/CO ratios of 0.43 is obtained.[47] Further CO2 recycleand H2, H2O and CH4 separation stages allow a purification ofthe stream to obtain CO with purities of 99.2%. To integratethis process in a chemical value chain with typical downstreamprocesses further compression steps are necessary since mostdownstream technologies work with higher pressures. On theother hand, compression of syngas with H2/CO ratios below 2 istechnically not an easy task.

The sulfur passivated reforming process (SPARG) combinesthe characteristics of dry and steam reforming. Thereby a broadspectrum of H2/CO ratios is accessible. Coke formation isprevented by sulfur poisoning (0.7 coverage with sulfur) of theNi-based catalysts, which leads to a reduced reaction rate ofcarbon formation. The sulfur poisoning, however, reduces theoverall activity of the catalyst and can become a problem formany downstream processes if not properly being taken careof.[39]

The CALCOR or SPARG processes were both developed toprevent coke formation under dryer reforming conditions.Unfortunately, these processes are only feasible in certaindownstream process scenarios and therefore lack breadth inapplicability. In order to overcome these shortcomings and todevelop a more broadly applicable process a better under-standing of the coke formation mechanisms and measuresregarding catalyst developments that can cope with thechallenge are necessary.

Downstream processes for DRM requiring syngas of a lowhydrogen to carbon monoxide ratio

From an industrial point of view, it is always worthwhile to lookdown the value chain and analyze which downstream processeswill benefit from the product produced; in the case of DRM a

syngas with high carbon monoxide content. By far the mostprominent and by volume the largest capacities in reformingare devoted towards the production of hydrogen – steammethane reforming is the dominating technology used todaydue to superior economy and process efficiency. Today themain hydrogen consumers on a global scale are ammoniaproduction plants, refineries and chemical plants.

The most important downstream processes relying onsyngas of a ratio of 2 are the methanol synthesis, processesrelying on Fischer Tropsch technology for the production ofalkanes.[48] These processes require a H2/CO ratio of 2, which canbe obtained by POX or ATR technology, as well as combinedreforming as SMR with carbon dioxide import. Several down-stream technologies, however, require CO rich syngas thatcould be provided with an efficient solution for DRM in place(Figure 2). It is sometimes debated whether the reverse watergas shift reaction (Eq. 4) can be an interesting alternative inproduction of carbon monoxide rich syngas. Provided thehydrogen used in the reverse water gas shift reaction isobtained from fossil sources (as for example from SMR ofnatural gas), the efficiency of the reverse water gas shiftreaction can be debated. For certain technical reasons and froman energetic footprint any reverse water gas shift processrelying on hydrogen based on natural gas will be inferior toDRM for obvious reasons.

Downstream processes of large volume that currently relyon CO rich syngas are Oxo-processes (hydroformylation ofolefins Eq. 10) and the production of acetic acid (Eq. 9).Processes which have untapped potential in coupling to DRM,due to the upside of carbon dioxide recycling and do requirelow hydrogen to carbon monoxide ratios are the directsynthesis of dimethyl ether (DME) and the direct conversion ofsuch a syngas to olefins e.g. in the Fischer-Tropsch to Olefinsreaction (FTTO).

2 COþ 2 H2 ! H3CCOOH DH� ¼ � 435:0 kJ � mol� 1 (9)

R� CH¼CH2 þ COþ H2 ! R� CH2� CH2� CHO (10)

Currently DME can be produced on an industrial scale bymethanol dehydration at 250–400 °C and 20 bar and so derivesfrom a syngas ratio of 2. The currently commercially availabletechnology uses a two-step process,[49] combining a methanolsynthesis and a dehydration stage (Eq. 11). The co-productionof water requires an additional separation stage after thedehydration reactor.

2COþ 4 H2 ! 2 H3COH ! H3COCH3 þ H2O (11)

A more efficient pathway towards DME relies on the directsynthesis of and can be performed with a syngas of ratio 1, asyngas obtainable via dry reforming[50,51] (Eq. 12). The benefit ofthis reaction is the overall higher energy efficiency, overallhigher conversions per pass compared to methanol synthesis,the avoidance of a separate dehydration step (a separatedehydration would have to be carried out in a separate reactor).

Reviews

4ChemCatChem 2020, 12, 1–19 www.chemcatchem.org © 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Wiley VCH Freitag, 14.02.2020

2099 / 158042 [S. 4/19] 1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

A very unique feature of a direct DME process with the DRMstep, is the potential of CO2 recycling and net consumption.

1:5 CO2 þ 1:5 CH4 ! 3 COþ 3 H2 ! H3COCH3 þ CO2 (12)

Dry reforming of methane: basic considerations

Thermodynamic equilibrium of DRM

Optimizing a process always goes hand in hand with an in-depth understanding of the process characteristics. In thisregard, the thermodynamic aspects of DRM are important tounderstand the origin of coke formation and the prerequisitesfor the implementation of a commercial DRM process. Asmentioned before, CO2 and CH4 are highly stable molecules andespecially CO2 can be regarded as “thermodynamic sink”.Therefore, valorization of CO2 is always an endothermic reactionand high activation energies are necessary for the downstreamchemistry.

Thermodynamic equilibrium analysis of the DRM reactionwas performed from different groups under various reactionconditions and with different models applied. Roussière[52]

calculated thermodynamic equilibria for different pressures,temperatures and feed compositions with the help ofDETCHEMEQUIL[53] based on the Gibbs free enthalpy minimiza-tion. His investigations showed that, even if not present in thebeginning, steam/water is always observed under thermody-namic conditions due to RWGS. This poses the question of how

co-feeding of water influences the thermodynamic equilibriumin DRM. However, Roussière's thermodynamic equilibria analysisshows that small amounts of steam in the initial feed have onlya minor influence on the obtained final syngas ratio – althoughcounterintuitive: DRM at completely “dry” conditions will as aproduct gas always contain water in the product gas. This isthereby a way of removing coke deposits by gasification,without changing the syngas ratio.

Cþ H2O! COþ H2 (13)

Roussière concluded that higher temperatures are beneficialfor syngas production by DRM, as expected for an endothermicreaction and a temperature range above 800 °C is necessary toachieve degrees of conversion of methane and carbon dioxidethat render the process economic. The lowest syngas ratio canbe obtained with feed ratios CH4/CO2 close to 1, where theaddition of small amounts of water (steam) to the feed has alsoa negligible influence.

Regarding coking tendency, a thermodynamic equilibriumanalysis has been published by Aramouni et al. who additionallyconsidered carbon and the respective particle size of the activemetal of the catalyst applied in their model. In this study it wasshown that the particle size seems to have a significantinfluence on the carbon deposition and an ideal particle size foravoidance of coke was found to be 5 nm, which is in agreementwith previous results from Bengaard et al.[54] and Kim et al.,[55]

who described the thermodynamic stability limit for carbondeposits on the catalyst to be 80 atoms. This finding is furthersupported by experimental results from Goula et al.[56,57]

Figure 2. Downstream processes relying on hydrogen and syngas production. Sources for syngas are highlighted in grey, established processes in blue andprocesses under development in green. Orange boxes mark the downstream processes, which could apply a syngas with low hydrogen to carbon monoxideratio. MTBE: 2-Methoxy-2-methylpropane.

Reviews

5ChemCatChem 2020, 12, 1–19 www.chemcatchem.org © 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Wiley VCH Freitag, 14.02.2020

2099 / 158042 [S. 5/19] 1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

Chein et al.[58] focused in their thermodynamic equilibriamodelling more on high pressure conditions due to theirindustrial relevance and considered additionally the effect ofinert gas N2. They concluded that increasing pressure has anegative effect on CH4 and CO2 conversion and favors carbonand water formation, while N2 co-feeding can be used to alterthe hydrogen to carbon monoxide ratio of the product gas.Özkara-Aydınoğlu[59] carried out thermodynamic equilibria anal-ysis for different pressures at temperatures of 200–1400 °C(Figure 3) and concluded that the pressure related negativeeffect on CH4 and CO2 conversion can be compensated byincreased reaction temperature and becomes negligible attemperatures above 1200 °C.

Catalyst materials for dry reforming

From an industrial point of view catalyst performance is animportant aspect, but catalyst cost cannot be neglected. As inother reforming technologies base metals are the preferredmetals of choice in the industrial realization of DRM on a largescale. Work on catalysts containing platinum metals as activecomponent has nevertheless it’s place, as the learnings fromthis work on platinum metals in DRM delivers insights thatcontribute to the overall understanding and the advancementof science. From the previous section on thermodynamicequilibrium analyses it becomes clear that DRM under con-ditions of high to complete conversion of methane and carbondioxide requires temperatures of at least 800 °C. These con-ditions pose several challenges for a catalyst material sincemost catalysts are prone to deactivation upon sintering andcoking.

[9,60,61,62,63] Also, irreversible reactions of the active metalwith the support are plausible under these high temperatureconditions. The activity of catalysts for dry reforming dependson a range of factors as the active metal (species, reducibility,particle size), the support (kind of support, surface area, acidity,basicity, oxygen storage capacity) and interaction of both.[64,65]

Several reviews[20,21,22,23,24] summarized the developments in thefield of catalyst materials for dry reforming of methane andfocused on the properties and catalytic behavior of differentcatalyst systems. While different active metals have beenconsidered, most of the discussed systems can be eitherassigned to noble metals[66] (Rh, Ru, Pt, Ir, and Pd) or basemetals (Ni, Co). As mentioned above: while noble metals showgood performance and stability in high temperature applica-tions, they are considered uneconomical for large-scaleapplications.[67,68] Investigations from Ashcroft et al.[69] showedlittle to no coke formation can be observed over Al2O3

supported Pd, Rh, Ru and Ir catalysts at high conversioncompared to quickly deactivating Ni-based catalysts whichdisplay similar initial activity. While Ashcroft et al.[69] reportedthe highest specific rates for Rh and Ir other authors publisheddifferent orders of activity.[70,71] Mortensen et al.[72] pointed out intheir review that the respective order of activity may besignificantly influenced by details of the catalyst preparationand resulting features like the dispersion, metal-supportinteractions and the reaction conditions. They comparedexperimental results from Ferreira-Aparicio et al.[71] and RostrupNielson and Bak Hansen[70] with DFT derived orders of activityfor SMR from Jones et al.[73] The kinetics of SMR and DRM arediscussed as similar, as has been shown by Wei and Iglesia[103]

and therefore the activity should as well correlate with theadsorption energies of O and C. Based on this assumption thetwo dimensional volcano plot by Jones et al.[73] (Figure 4) showsthat catalysts based on Ru and Rh are based on theoreticalcalculations the candidates of highest intrinsic activity for DRM.In terms of base metals, Ni and Co have proven to be thecandidate metals of highest intrinsic activity. The higheradsorption energy of oxygen containing species on Co as activemetal indicates that it can most likely be poisoned by oxygen,or even go through oxidation to oxidic phases, which agreeswell with experimental studies on Co catalysts.[74,75]

Figure 3. Conversion of methane (a) and CO2 (b) and final H2/CO (c) ratiocalculated for the temperature range 200–1400 °C at 1, 10 and 20 bar. FromÖzkara-Aydınoğlu.[59] Reproduced with permission from Ref. [59] Copyright2010 Elsevier.

Reviews

6ChemCatChem 2020, 12, 1–19 www.chemcatchem.org © 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Wiley VCH Freitag, 14.02.2020

2099 / 158042 [S. 6/19] 1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

Ni is far less expensive than the noble metals Ru and Rh[76]

therefore the pareto optimum considering economic aspectsand achievable activity levels. This is reflected in literaturewhere Ni-based catalysts are the most widely studied catalystscoupled by the fact that Nickel has been the workhorse in SMRand ATR. A comprehensive overview of current findings inresearch on Ni-based catalysts is given by Jang et al.[77] in theirrecent review, where especially properties of catalysts and theirinfluence on the activity in DRM are discussed. Co as activemetal on different supports has been comprehensively re-viewed by Ruckenstein and Wang; Cobalt is discussed to displaylower intrinsic activity. Weak metal-support-interactions arediscussed as potential reason for that behavior.[78] On the otherhand, Co based catalysts are reported to display increased cokestability. The activity of Ni and Co-based catalysts can be furtherincreased by the addition of noble metal dopants.[66,79,80] Severalresearchers also investigated bimetallic Ni� Co catalysts tomerge the high activity of Ni with the coke stability of Co-basedcatalysts,[81,82,83,84] still results give no clear indications whethersynergies can be achieved.

Strong metal support interaction (SMSI) can play animportant role to achieve highly dispersed active metal sites.On the other hand, SMSI is discussed to lead to lowerreducibility of the active metal of a supported catalyst andthereby reduce either the initial activity of the material[85,86,87] orto lead to deactivation triggered by the formation of solidsolutions.[88]

In the case of DRM, Ruckenstein and Wang could prove bothof these effects have relevance for Co supported on differentoxides.[78] Co supported on SrO and BaO showed only littleactivity, which was attributed to a strong sintering of SrO andBaO particles accompanied by encapsulation of the Co sites.Catalysts supported on SiO2, Al2O3 and CaO showed good initial

activity but were prone to deactivation under reaction con-ditions. According to the investigations of the author teamMgO was the only support allowing for high activity andstability of the active metal over times of 50 h. Furthermore,Ruckenstein and Wang observed lower reducibility of the Cospecies on Al2O3 and MgO, with Co(O)/MgO mixed oxide phasesforming on MgO being very hard to reduce. The low reducibilityin the case of Co(O)/MgO arises from the identical crystalstructures of CoO and MgO (NaCl-type). From the observationsof the authors this property is also beneficial in hindering thesintering of metallic Co-species.[89] Guo et al. observed similarresults for Ni supported on Al2O3, Al2O3� MgO and MgAl2O4

under DRM conditions.[90] While Ni/Al2O3 deactivated rapidly, Ni/MgAl2O4 showed no deactivation and stable catalytic perform-ance over 200 h. It was concluded that the strong interactionsbetween Ni(O) and MgAl2O4 prevent sintering and preventedthe formation of a catalytically inactive NiAl2O4 spinel.

Kinetics of dry reforming

Several kinetic models are discussed at present in literature asappropriate to describe the mechanism of the catalyzed DRMreaction. For the reforming over Ni-based catalysts mainlymodels with different modifications are discussed and havebeen compared and summarized by Mark et al.[91] and Kathiraseret al.[92] Besides simple power law models, which cannotsufficiently describe the various elementary steps of thereaction mechanism over a wide parameter range, especiallythe Eley-Rideal (ER) type of models and Langmuir-Hinshelwood-Hougen-Watson (LHHW) models are applied. An associatedassumption made in a number of publications is that the rate-determining-step (RDS) involves the reaction of adsorbedspecies with other reactants from the gas phase.

A number of authors developed approaches for the kineticdescription of the dry reforming reaction (DRM) where CH4 andCO2 are both first adsorbed on the catalyst surface prior toreaction; two possible mechanisms are currently being dis-cussed as most relevant with regards to representativedescription of data obtained by various catalyst systems. Akpanet al.[93] assumed that the methane dissociative adsorption isthe RDS. The following reaction mechanism was postulated (*:unoccupied active site and Ox: lattice oxygen of support), whichwas validated by experimental studies over Ni/CeO2-ZrO2.

CH4 þ 2* $ CH3ð*Þ þ Hð*Þ

CH3ð*Þ þ * $ CH2

ð*Þ þ Hð*Þ

CH2ð*Þ þ * $ CHð*Þ þ Hð*Þ

CHð*Þ þ * $ C ð*Þ þ Hð*Þ

Cð*Þ þ Ox $ COþ Ox-1 þ *

CO2 þ Ox-1 $ Ox þ CO

Figure 4. Two-dimensional volcano-curve of the turnover frequencies (log10)as a function of O and C adsorption energy. T=500 °C, p=1 bar; 10%conversion. From Jones et al.[73] Reproduced with permission from Ref. [73]Copyright 2008 Elsevier.

Reviews

7ChemCatChem 2020, 12, 1–19 www.chemcatchem.org © 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Wiley VCH Freitag, 14.02.2020

2099 / 158042 [S. 7/19] 1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

4 Hð*Þ $ 2 H2 þ 4*

H2 þ Ox $ Ox-1 þ H2O

Becerra et al.[94] postulated a similar model but assumeddirect reaction of gaseous CO2 with adsorbed methyl species.

CHxð*Þ þ CO2 $ COþ x=2 H2 þ *

Erdohlyi et al.[95] reported a similar approach to the kineticdescription of the mechanism for DRM over Rh/Al2O3, whileRichard et al.[96] described the same reaction with a simplifiedLangmuir-Hinshelwood mechanism, where both gaseous spe-cies are adsorbed on the surface.

More widely applied are so called Langmuir-Hinshelwood-Hougan-Watson kinetic models derived reactionmechanisms.[97,98,99] A kinetic model built on such an approachhas first been presented by Xu and Froment[100] for the case ofsteam reforming of methane and has further been improved byZhang and Verykios.[101] Zhu et al.[102] derived a reaction mecha-nism based on experimental results and conclusions made byWei and Iglesia[103,104] that involve the step-wise activation ofadsorbed CH4 and adsorbed CO2 and oxidation of CHx (0�x<4) species by surface oxygen resulting from CO2 decomposition.The formation of surface OH species and formate-type speciesas intermediates is considered within the mechanism.

CH4 þ * $ CH4ð*Þ CH4 adsorption

CH4ð*Þ þ x* $ CHx

ð*Þ þ x Hð*Þ CH4 dissociation

CO2 þ * $ CO2ð*Þ CO2 adsorption

CO2ð*Þ þ * $ CO ð*Þ þ Oð*Þ CO2 dissociation

CO2ð*Þ þ Hð*Þ $ COOHð*Þ þ * H-assisted CO2 dissociation

COOHð*Þ þ * $ COð*Þ þ OHð*Þ

CHxð*Þ þ OHð*Þ $ CHxOHð*Þ þ *

Oxidation of adsorbed CHx via OH groups

CHxOHð*Þ þ ðxþ 1Þ* $ COð*Þ þ ðxþ 1Þ H ð*Þ

CHxð*Þ þ Oð*Þ $ CHxO

ð*Þ Oxidation of adsorbed CHx

CHxOð*Þ þ x* $ COð*Þ þ x Hð*Þ

Hð*Þ þ Hð*Þ $ H2ð*Þ þ *

There is currently no general agreement in literature on theRDS of the DRM reaction; a summary of the state of discussionis given by Kathiraser et al.[92] The generally most widelyaccepted RDS is the activation of CH4 by hydrogen abstraction,as proposed by Zhang and Verykios,[101] Wang and Au[105] andWei and Iglesia.[103] While all the mentioned models aredeveloped for Ni-based catalysts, it appears that they are also

suitable for describing the DRM reaction kinetics over sup-ported noble metal catalysts (Rh, Pt, Ir).[103]

Schulz et al. compared the respective reactions pathways(Figure 5) via isotope labeled reactants in the dry methanereforming over Ni- and Pt-based catalysts.[106] A very importantfinding of Schulz et al. refers to the fact that a pronouncedreversibility of C� H and C� O bond formation is found. Theexperiments show that the initial rates for carbon dioxideconversion are higher than for methane conversion over Nickel– a fact that is attributed to the energetically more difficultmethane activation through C� H activation over Nickel. ForPlatinum the situation is different: here the rates of conversionfor methane and carbon dioxide are almost equal.[107,108] Cokeformation on both Platinum and Nickel occur via methanedissociation and decarbonation, yet for Nickel the Boudouardreaction is a predominant coke formation pathway (Eq. 14).Also, the observed rates of coke formation on Nickel were muchhigher compared to Platinum – the lower coking rates of thelatter are attributed to higher gasification rates via methanation.Schulz also reports a differing mode of activation for carbondioxide for the two metals: while over Nickel carbon dioxide isactivated via dissociative adsorption, this activation mode doesnot occur over Platinum. For Nickel based catalysts this mode ofcarbon dioxide activation is also seen as the reason for themore pronounced coking tendency.

Understanding Coke formation and growth behavior ofcarbonaceous deposits under DRM conditions

Thermodynamics of coke formation

Understanding coke formation and growth behavior as maindeactivation pathway for catalysts applied in DRM is a major

Figure 5. Reaction scheme of dry reforming over Pt- and Ni-based catalystsas proposed by Schulz et al.[133]

Reviews

8ChemCatChem 2020, 12, 1–19 www.chemcatchem.org © 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Wiley VCH Freitag, 14.02.2020

2099 / 158042 [S. 8/19] 1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

challenge to be understood and solved. Coke can be formed bya variety of reactions including catalytic and non-catalyticpathways. Besides the generation of coke precursors and cokeby non-catalytic gas-phase reactions, which will be discussedseparately, three main pathways must be considered as relevantfor coke formation under DRM conditions. The Boudouardreaction as carbon source based on CO disproportion (Eq. 14) isgenerally more favorable at lower temperatures.

2 CO! CO2 þ C DH298 ¼ � 171 kJ � mol� 1 (14)

The reduction of CO by H2 to carbon and water is favorableat low temperatures but of low relevance due to its alwayslarger Gibbs energy, except for gas mixtures containing high H2

partial pressures.[112]

COþ H2 ! H2Oþ C DH298 ¼ 131 kJ � mol� 1 (15)

Coke formation as a result of methane decomposition(methane cracking via radical pathways and subsequently

methane decarbonation) on the other hand is thermodynami-cally favored at high temperatures.

CH4 ! 2 H2 þ C DH298 ¼ 75 kJ � mol� 1 (16)

The consequences of these different coke forming reactionscan properly be illustrated by means of CHO ternarydiagrams.[109,110,111] Kee et al.[111] for example calculated therespective equilibria for a pressure of 25 bar and varioustemperatures (Figure 6). The authors defined arbitrary the cokedeposition boundary by an equilibrium content of graphiticcarbon of 1 ppm. Coke formation is apparently thermodynami-cally favored at completely “dry” reforming conditions and cokeformation is predicted from thermodynamic considerations.Under dry reforming conditions Giehr et al.[112] predicted thatgraphite formation should be the thermodynamically preferredmodification of carbon deposits. Nevertheless, other carbonforms like fullerenes, amorphous and filamentous carbon arecommonly reported in experimental studies[46,63,85,113,114,115],although their formation should be thermodynamically lessfavored. Effective carbon removal via a range of reactions

Figure 6. C/H/O ternary diagram including limit lines for carbon deposit formation as a function of gas phase composition and temperature. Adapted fromKee et al.[111].

Reviews

9ChemCatChem 2020, 12, 1–19 www.chemcatchem.org © 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Wiley VCH Freitag, 14.02.2020

2099 / 158042 [S. 9/19] 1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

(gasification of carbon with water (reverse reaction of 15) orhydrogen (reverse reaction of 16), and the reverse Boudouardreaction (14)) are discussed below and from a thermodynamicpoint of view are effective modes of carbon removal in thetemperature range were coke formation occurs; one has to takeinto account that most reactions allowing carbon removal areexergonic in nature[112] (see Figure 7).

High pressure and temperatures above 550 °C (both ofrelevance for the industrial process) favor methane cracking,while lower temperatures may lead to pronounced cokeformation from Boudouard reaction. Both reactions have beeninvestigated for their respective influence on coke formation atstandard and reaction conditions by Giehr et al.[112] It wasconcluded that coke formation has its origin from the maincontribution from methane cracking under the experimentalconditions investigated by Giehr.

Coke formation mechanisms – molecular precursors to cokeformation

As discussed in the previous section: as diverse catalystcandidate materials are applied by various groups in dryreforming of methane it can be expected that the governingcoke formation mechanisms may be differing in nature. Some-times even conflicting findings may arise due to differingkinetics in coking and decoking pathways that are catalyzed bythe respective materials. Varying reactor geometries, testconditions and programs applied will also add to alteredfindings in the performance results, but have to be taken intoaccount when comparing results head to head. The catalystsmainly worked on in literature can roughly be divided in twogroups; the noble metals (like Pt) and refractory metals,especially Ni. As already mentioned above from an economicperspective base metals are of major interest for industrialapplications of DRM, studies comprising platinum metals are ofscientific high interest. As a second important aspect, non-catalytic reactions that proceed in the gas phase play animportant role and are an often overlooked source for coke

precursor formation, more specifically higher hydrocarbons,leading to subsequent pronounced solid coke formation on thecatalyst. At the industrially relevant high pressures and temper-atures non-catalytic gas phase reactions become of majorimportance in terms of coke precursor formation and havebeen investigated by Kahle et al.[116] for the case of DRMexperimentally and via modelling approaches. The teaminvestigated coke formation over Pt-catalysts under DRMconditions (temperatures above 850 °C and pressures above20 bar). Under these experimental conditions coke formation isobserved to start at the entrance of the catalyst bed andupstream of the catalytic zone and not in the catalyst zone itselfand downstream of the catalyst section. This distinct cokingbehavior can be related to non-catalytic homogeneous reac-tions in the gas phase which lead to the formation of highlypotent coke precursors.[117] For these gas phase reactions, areaction pathway was originally proposed by Becker andHüttinger,[118,119] who described a methane dehydrogenationpathway followed by coupling reactions of formed radicals tohigher hydrocarbons (Figure 8).

The formation of light hydrocarbons by radical recombina-tion reactions, with resulting products such as acetylene (C2H2)and small olefins may be source to the formation of evenheavier coke precursor species. Combination reactions ofacetylene, and alternatively addition and dehydrogenationsteps of small olefins like ethylene and propylene, may give riseto increasing concentration of aromatics (C6H6) and polycyclicaromatic hydrocarbons (PAHs)[120,121] which are also highlypotent coke precursors. The consequences of these findings arefundamental in nature: not only is C1 related coke build-uphighly relevant, but even if present in much lower concen-tration in the feed, coke precursors like acetylene, small olefinsand aromatics are present and any catalyst material will besubject to the consequences of their presence. The depositionof carbon through these coke precursors occurs also accordingto Cava et al., via condensation of the small hydrocarbonsunder formation of larger entities and assembles in the gasphase.[122] Kahle et al. discuss three different gas-phase reactionmechanisms as prevalent in their publication. Mechanism 1 is

Figure 7. Overview of coke forming reactions (orange: Boudouard reaction, catalytic CH4 cracking and homogeneous CH4 cracking) and removing reactions(blue: methanization, gasification and reverse Boudouard reaction) under DRM conditions.

Reviews

10ChemCatChem 2020, 12, 1–19 www.chemcatchem.org © 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Wiley VCH Freitag, 14.02.2020

2099 / 158042 [S. 10/19] 1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

based on a reaction scheme developed for isooctane combus-tion (4238 reactions), which originally stems from LawrenceLivermore National Laboratory.[123] Coupling this reactionscheme to the toluene scheme by Dagaut et al.[124] leads to amechanism consisting of 8927 elementary reactions and wasexperimentally validated by the oxidation of toluene, theignition of benzene-oxygen-argon mixtures, and the combus-tion of benzene in flames. Mechanism 2 derives from theGolovitchev group[125] and consists of 690 elementary reactionsand mechanism 3 is based on results from the Dean group,[126,127]

who included the reaction pathways for aromatics and somepolyaromatic hydrocarbons.[128] For all three mechanisms Kahleet al.[116] performed simulations for an empty reactor tube andcould show that the formation of saturated and unsaturatedhydrocarbons (C2H6, C2H4, C2H2, C6H6) can be explained via allthree mechanisms (Figure 8). Noringa et al.[121] obtained com-parable results that also confirm the findings for the evaluationof the chemical kinetics of ethylene pyrolysis at 900 °C. Kahleet al. further investigated the effect of feed variation, specificallyvariation of H2 and H2O partial pressure in the inlet gas.Hydrogen addition to the initial feed has a large effect on theinhibition of coke formation due to an inhibiting effect onmethane pyrolysis and hydrocarbon formation (Fig-ure 9).[52,129,130,131] Water addition has a smaller but still pro-nounced effect on the inhibition of the non-catalytic pyrolysispathways of methane compared to hydrogen but is still aneffective inhibitor for coke formation if used in the appropriatepartial pressure regime.

Coke formation in presence of a catalyst is reported to bestrongly dependent on the nature of the catalyst, supportmaterial, active metal and the particle size of the activemetal[132,133] since different reaction pathways are possible overdiffering catalyst materials respective intermediates formed onthe catalyst species will also differ.[134] Mainly the coke formationover base metals like Ni and noble metals like Pt have beeninvestigated.[135,136,137] Coke formation over noble metals andrefractory metals differs in terms of the primary coke sourceand absolute extent of coke formation. Schulz reports that cokeformation is catalyzed on the base metal mainly via dissociationof CO2 and the Boudouard reaction, while CH4 dissociation isthe main coke source over Pt, most probably stemming fromthe fact that Platinum has a higher potency for C� Hactivation.[106] The significantly different extent of coke forma-tion of Pt and Ni results also from higher gasification rate ofcarbon species; with Pt displaying a mechanism in which cokeis gasified via methane formation, it appears that Ni is capableof catalyzing the reverse Boudouard reaction as discussed bySchulz et al.[134]

Growth patterns of nano-sized coke deposits on catalystmaterials

The size of supported nanoparticles of a respective active metaldoes not only have a major influence on the activity of acatalyst material in reforming reactions in general and espe-

Figure 8. Methane dehydrogenation pathway according to Becker and Hüttinger[118,119] (A) and numerical product profiles as a function of axial position alongthe reactor; blank reactor tube, 10% H2 (H2O), CH4/CO2 =1; 20 (B).[115] Reproduced with permission from Ref. [115] Copyright 2009 Elsevier.

Figure 9. Coke deposition under dry reforming conditions as function of temperature and H2 concentration in the gas mixture; left-to-right: empty ceramictube, partially filled reactor (20 mL Al2O3) and totally filled reactor.[52] Reproduced with permission from Ref. [52]. Copyright 2013 T. Roussiere.

Reviews

11ChemCatChem 2020, 12, 1–19 www.chemcatchem.org © 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Wiley VCH Freitag, 14.02.2020

2099 / 158042 [S. 11/19] 1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

cially in DRM, but also plays a major role in coke formation andcoke growth mechanisms.[138,139,140] Titus et al. proposed a cokegrowth mechanism for DRM based on TEM investigations andobservations of the growth modes of coke on the catalystcandidate system NiO� MgO-ZrO2.

[141] According to their resultscoke deposition and growth on supported Nickel catalystsproceeds via a two-step process (Figure 10). In the first step agrowth of multi-walled carbon nanotubes (MWCNT) with thenanoparticles of the active metal Nickel at the tip of theMWCNT can be observed. This growth mode of carbon isdislocating the Ni-nanoparticles from the support material. Inthe following step in the so-called “filling-in phase”, largeramounts of amorphous and graphitic layered carbon aredeposited around the Ni-particles and the MWCNTs. It can bespeculated that the increased surface area provided by theMWCNT leads to the formation of this less dense coke phase inthe voids. It could also be shown that more active catalysts withincreased Ni loading tend to show an increased final amountand overall rate of coke formation, this clearly illustrates thatcoke formation is a catalyzed phenomenon for this type ofcatalyst under observation. The mechanism described by Tituset al. captures well singular spotlight observations of othergroups and can describe a range of coke patterns observed fora range of reports in literature.

Controlling coke formation: Insights from catalyst modelsystems

Addition of water to the feed

Steam reforming conditions favor much less catalyst cokingthan dry reforming conditions and therefore the addition ofcertain amounts of water can be considered as way to reducecoke deposits in DRM. As discussed above: for thermodynamicreasons small amounts of water, even for inlet gas compositionsthat are totally “dry”, will result in water being present in theoutlet gas in any case, as predicted by thermodynamics. Kahleet al. investigated in detail the addition of water to the feed inDRM and showed that non-catalytic coke precursor formationcan be reduced to a certain extent in comparison to completely“dry” conditions.[116] This is in line with the fact that theformation of methyl-radicals can occur via two pathways and

only the less energetically favored and therefore prominentpathway 2 is inflicted by water addition.

CH4 $ *CH3 þ *H Pathway 1

CH4 þ *OH$ *CH3 þ H2O Pathway 2

In presence of a catalyst, H2O addition to the feed clearlyleads to an inhibition of coke precursor formation also viagasification of once formed carbon deposits.

Cþ H2O$ H2 þ CO

Gasification of coke via cyclic operation

Coke deposits can in principle be removed via the reverseBoudouard reaction in order to restore catalytic activity. Thishas been investigated for Ni catalysts supported on SiO2, Al2O3,and TiO2 by Takenaka et al.[142] In their experiments methanedecomposition was provoked on a Nickel based catalyst at550 °C, which resulted in formation of carbon deposits. In asecond step the coke deposits were gasified in a pure CO2

atmosphere at 650 °C via the reverse Boudouard reaction. Theobtained results showed that the initial catalyst activity couldonly be recovered for Al2O3 and TiO2 supported catalysts;additionally, it was observed that no tremendous sintering ofthe Ni particles occurred. In case of Ni/SiO2 it could be seen thatthe Ni nanoparticles were significantly enlarged throughsintering through the regeneration cycle. Xu et al.[143] followedsimilar ideas of carbon dioxide-based catalyst regeneration andinvestigated the regeneration of Ni/ZrO2 after methane decom-position. Later Steib et al.[144] could show that the regenerationcan be applied after catalyst coking stemming from applicationof dry reforming conditions as well. In both cases no or onlyminor sintering of the Ni nanoparticles on the support materialwas observed which is attributed to the intrinsic metal supportinteraction between Ni and ZrO2. Steib et al. proposes acomplete cyclic catalyst regeneration scheme (Figure 11) ofactivation-reforming-coking-regeneration-reforming-steps.[145]

Figure 10. Schematic representation of the stages of carbon growth on Ni particles on MgO on ZrO2 (MWCNT: multi-walled carbon nanotubes).[140] Adaptedfrom Titus et al.[141]

Reviews

12ChemCatChem 2020, 12, 1–19 www.chemcatchem.org © 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Wiley VCH Freitag, 14.02.2020

2099 / 158042 [S. 12/19] 1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

The role of acidity and basicity of the catalyst material

The acidity/basicity of the catalyst may influence the catalystactivity and selectivity patterns. Acidity/basicity of a givencatalyst system should therefore be extremely relevant for DRM,as most coke forming reactions can be understood asunselective reaction paths. (Highly) acidic support materials areusually anticipated as non-ideal catalyst or support materials forDRM due to the intrinsic enhancement of methane cracking byacid sites.[149] Therefore, acidity usually is treated as a propertythat must be moderated to limit the latter undesired effect ofcoking. Basicity is in some papers treated as a desirablematerials property, allowing for enhanced carbon dioxideadsorption and subsequently enhanced carbon dioxideactivation.[149] Reports that treat the topic of the influence ofacidity and basicity in depth are rare, nevertheless the topic assuch is of high relevance and will be discussed in the following.Ferreira-Aparicio et al. have investigated the role of acidity andbasicity in DRM for catalysts with Ru supported on Al2O3 andSiO2.

[146] The authors conclude from their study that surfacehydroxyl groups and the capability of carbon dioxide dissocia-tion or alternatively the capability of forming formate inter-mediates play a major role in the accumulation of coke formingspecies and the resilience of the resulting catalyst material.They propose a bifunctional mechanism triggered by surfacehydroxyl groups. Das et al.[147] concluded in their investigationsof Ni supported on modified silica and alumina that a moderateacidity-basicity is best for a stable and active DRM catalyst.

The above-mentioned observations are contrasted by thefindings of Titus et al.[148] who investigated the basic and acidicsite density of Ni/MgO-ZrO2-based catalysts. Titus et al. showedthat the basic and acidic site density is balanced and negligiblein absolute concentration for materials treated prior to catalytictesting at temperatures relevant to the temperature corridorthat DRM is usually performed at. In their results they showedthat acid and base site density over MgO/ZrO2 is alsoindependent of the MgO content, beyond a certain threshold ofMgO. From that finding it was concluded that despite previousreports[141] the basic site density of different compositions in themodel system NiO� MgO-ZrO2 has only a minor or evennegligible effect on the stability under DRM conditions and thatrather a higher Ni dispersion, due to stronger interactionsbetween MgO and NiO, is responsible for an improved stabilitytowards coking.

Alternative approaches apart from oxide-based carriermaterials should be mentioned in this context.[149,150] Suchmaterials that due to their inherent carbon dioxide adsorptionproperties have proven to show enhancement in low temper-ature carbon dioxide conversion have not been applied in dryreforming – but could give a perspective beyond oxides assupport materials.

Figure 11. Scheme of cyclic reforming coking regeneration steps for Ni/ZrO2.[144] Metallic Ni0 particles are shown in red. Zirconium oxide particles in grey and

Zirconium carbonates in blue. Adapted from Steib et al.[144]

Reviews

13ChemCatChem 2020, 12, 1–19 www.chemcatchem.org © 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Wiley VCH Freitag, 14.02.2020

2099 / 158042 [S. 13/19] 1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

Controlling coke formation through stabilization of nickelnanoparticles on hexaaluminates: Understanding the reductionmechanism of the active metal Ni

Roussière et al. investigated Ni-hexaaluminates ANiyAl12-yO19-δ

(A=Ba, Sr and La and y=0.25, 0.5, 1) as catalyst candidatematerials in DRM.[151] These materials are of interest sincehexaaluminates are known for their resilient nature: theymaintain their specific surface areas even under drastic reactionconditions like high temperatures of 1300 °C[152,153,154,155] andhydrothermal load. In addition, hexaaluminates stabilize Ninanoparticles and prevent their sintering. Metallic Ni nano-particles which form under reducing conditions from Ni-containing hexaaluminates, are considered as active phase inDRM.

[156,157]

The deactivation of Ni- or Co-based catalysts is discussed tobe mainly due to coking[158] and not through sintering and lossof active surface area, but especially the latter effect isimportant to be considered, since ensembles of large Niparticles are reported to be more prone to coking.[159,160,161,162] Indepth investigations of different Ni hexaaluminates show thatthe activity and stability of the catalyst can be related to thesize and texture of the Ni0 nanoparticles.[163] While materialswith large and/or unsupported Ni0 nanoparticles are highlyactive, they tend to deactivate rapidly since the coke selectivityis enhanced. Ni hexaaluminates show decreased activity andincreased coking after calcination at higher temperaturesbeyond 1300 °C, which is in good agreement with reducedstabilizing surface defects and therefore larger particles. Themore stable catalyst materials generally are found for lower Nicontent associated with a high Ni dispersity and a strongtextural growth of Ni0 nanoparticles. Especially the texturalgrowth is related to a strong interaction of the Ni nanoparticleswith the support material. Roussière et al. developed a “coupledelectron defect migration reduction mechanism” to describethe migration of Nickel in the oxidic hexaaluminate materialand developed a rationale for the textural growth of Ni

nanoparticles during the reduction on hexaaluminate materials(Figure 12).

The mechanism can be described in three steps.* In a first step oxygen is removed from the mirror planes (MP)

of the hexaaluminate. A hypothesis which is in line with thehigher oxygen diffusion within the mirror plane and thetherefore higher reducibility of the oxygen.

* In a second step Ni2+ cations in the spinel block (SB) migrateinto the mirror plane to the surface defect sites (i. e. grainboundaries), where they are reduced to Ni0 on the outersurface of the grain.

* Ni0 nanoparticles are most probable to form at defect sitesnext to steps and edges of the material. These nucleationcenters facilitate growth and the prevailing growth pattern ofthe Nickel particles are tetrahedral particles which can beshown in STEM images.The form of the nanoparticles is reasoned to originate from

the fact that the hexagonally dense packing of oxygen on thehexaaluminate surface induces the specific growth pattern ofmetallic Ni crystallizing in an fcc lattice. This growth patternallows the formation of Ni� O bonds at the grain boundary.

Decoking mechanisms via redox-catalysis on Co/y-Al2O3

Although reported as specifically less active in DRM comparedto Ni catalysts, Co-based catalysts show superior stabilitytowards coke formation.[164] This is discussed as being related toan additional decoking mechanism catalyzed by Co-containingspecies. Based on thermodynamic calculations and steady statereactor simulations Giehr et al. could show that while Ni isreduced over the whole length of the reactor, the oxidationstate of cobalt in the model compound CoAl2O4 stronglydepends on the axial position within the reactor and therespective spatial partial pressure regime.[165] In the first part ofthe reactor cobalt is most likely to be oxidized by CO2 or H2O

Figure 12. Scheme of the coupled electron defect migration reduction mechanism proposed by Roussière (A).[133] Ni0 particles grow in tetrahedral shape (C)which can be explained in terms of preferred growth according to the fcc Ni structure (B). From Roussière et al.[133] Reproduced with permission from Ref.[133] Copyright 2014 Wiley.

Reviews

14ChemCatChem 2020, 12, 1–19 www.chemcatchem.org © 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Wiley VCH Freitag, 14.02.2020

2099 / 158042 [S. 14/19] 1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

and spinel structures are the predicted reaction products(Eq. 23).[165]

Coþ Al2O3 þ CO2 ! CoAl2O4 þ CO (23)

Besides a Langmuir Hinshelwood mechanism that can serveas a mechanistic approach for the reverse Boudouard reactionover reduced Co0

, the CoAl2O4 as a mixed metal oxide cansupply oxygen in a MvK-mechanism and oxidize carbondeposits to CO and CO2 (Eq. 24). The reaction of MAl2O4 withsolid carbon (Figure 13) is discussed in the literature as beingthermodynamically favored.[112]

CoAl2O4 þ C! Coþ Al2O3 þ CO (24)

Giehr describes that carbon deposit formation is observed inthe first reactor half through methane cracking; a fact that canstill be brought in good agreement with the enhanced cokestability. Giehr et al. suggest a catalytic cycle for carbon depositremoval,[112] which explains the superior stability of Co basedmixed metal oxides like Cobalt Spinels.[112] Similar coke removalmechanisms have been proposed for different catalyst materialsby Theofanidis et al.[166,167] (Fe� Ni/MgAl2O4) and Kim et al.[168] (Ni–Ce mixed oxide catalysts).

Conclusions

Coke formation is the main reason for deactivation in dryreforming of methane under industrially relevant conditionsdue to the necessary high temperatures (thermodynamic) andpressures (economical). Therefore, understanding coke forma-tion is an important factor for the development of novel cokeresilient catalysts systems with high activity and stability. Thesources of coke under dry reforming conditions are twofold:firstly coke can be formed directly on the catalyst surface basedon reactions like methane cracking and the Boudouard reaction

and secondly coke formation on the catalyst surface can bedriven by coke precursors derived from gas-phase cokeprecursors (acetylene, olefins and aromatics), which can also bedeposited on the catalyst. The actual coke sources, in case ofcatalytic coke formation, depend on the reaction conditionsand on the respective active metal since the reaction pathwaysvary for noble and base metals and the nature of the catalystsupport. The growth of carbon on the catalyst surface isdiscussed to proceed via the formation of carbon nanotubesunderneath the metal particles and consecutive filling of thevoids in-between.

In recent years several strategies have been developed tocome up with new concepts of multifunctional catalystmaterials that can cope with coke formation: either by theprevention or system inherent removal of coke deposits underreaction conditions. The addition of water to the feed is thewidest applicable strategy, which has, in case of small wateramounts, only limited influence of the final desired syngasratios. More catalyst specific ways to control coke formationhave been derived for Ni-based model systems where especiallystrong metal support interactions are important, either tomaintain small nanoparticles during reaction or after regener-ation procedures or to stabilize small nanoparticles right fromthe beginning. A more specific way to reduce coke formationcould be shown over Co-based catalysts, where a Mars-vanKrevelen mechanism enables a de-coking process. Overalltheoretical and experimental model systems help to understandthe challenge of coke formation under dry reforming conditionsand demonstrate ways to maintain high activity over prolongedamount of time. Several approaches are discussed in literature,which all together gives a conclusive picture of the obstaclesthat are crucial and must be overcome in order to limit thecoking on catalyst materials in the field of dry reforming.

Figure 13. Proposed reaction scheme of combined Langmuir Hinshelwood and Mars-van Krevelen pathway over Co species.[112] Graphics adapted from [112].

Reviews

15ChemCatChem 2020, 12, 1–19 www.chemcatchem.org © 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Wiley VCH Freitag, 14.02.2020

2099 / 158042 [S. 15/19] 1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

Acknowledgements

The constant support of BASF management in our efforts toachieve not only a technical product and process - but animproved understanding of the underlying phenomena leading tothe development of the SYNSPIRE™ catalyst family. Linde co-workers and management are acknowledged as reliable processpartner in the development and supply of the DRYREF™ Technol-ogy. We wish to thank hte GmbH for giving us as a team theopportunity of pushing the limits in high throughput testingtechnology in reforming catalysis and the opportunity for using itas enabler in our daily work. The great collaboration, the highlevel of science and the collaborative atmosphere of our academicpartners is acknowledged specifically of the Deutschmann groupof the Karlsruhe Institute of Technology, Gläser group of LeipzigUniversity, Lercher group of the Technical University Munich andThe DECHEMA Research Institute. Finally, the BASF and hteDRYREF™ team, all its current and former co-workers, is gratefullyacknowledged. The project work was funded by the GermanMinistry of Commerce (FKZ 0327856 und 03ET1282), the fundingand trust is gratefully acknowledged.

Keywords: dry reforming · carbon deposits · industriallycatalysis · high throughput experimentation · multi-scalemodelling

[1] International Energy Agency (IEA), Energy Technology Perspectives2017.

[2] D. Treacy, J. R. H. Ross, Preprint Papers - American Chemical Society,Division of Fuel Chemistry 2004, 4, 126.

[3] M. De Falco, G. Iaquaniello, G. Centi, CO2: A Valuable Source of Carbon;Springer-Verlag: London, 2013.

[4] M. Peters, B. Köhler, W. Kuckshinrichs, W. Leitner, P. Markewitz, T. E.Müller, ChemSusChem 2011, 4, 1216.

[5] A. S. Lindsey, H. Jeskey, Chem. Rev. 1957, 57, 583.[6] M. Behrens, Angew. Chem. Int. Ed. 2016, 55, 14906.[7] A. W. Kleij, M. North, A. Urakawa, ChemSusChem 2017, 10, 1036.[8] J. D. Holladay, J. Hu, D. L. King, Y. Wang, Catal. Today 2009, 139, 244.[9] J. Zhang, H. Wang, A. K. Dalai, Ind. Eng. Chem. Res. 2009, 48, 677.

[10] J. R. Rostrup-Nielsen, Catal. Today 1993, 18, 305.[11] J.-M. Lavoie Front. Chem. 2014, 2, 81.[12] J. Romm, The Hype about Hydrogen: Fact and Fiction in the Race to Save

the Climate, Island Press, Washington, DC 2004.[13] a) M. Meinshausen, N. Meinshausen, W. Hare, S. C. B. Raper, K. Frieler,

K. R. Knutti, D. J. Frame, M. R. Allen, Nature 2009, 458, 1158;b) P. M.Fearnside, Clim. Change 2000, 46, 115;c) T. L. Frölicher, M. Winton, J. L.Sarmiento, Nat. Clim. Change 2014, 4, 40;d) S. Wang, G. Q. Lu, G. J.Millar, Energ Fuel. 1996, 10, 896.

[14] J. Lang, Z. Phys. Chem. (Leipzig) 1888, 2, 161.[15] J. Zhang, H. Wang, A. K. Dalai, Canada. US Patent: 7,985,710, 2011.[16] R. Bhat, W. M. H. Sachtler US Patent: 5,753,143, 1998.[17] S. Wang, G. Q. Lu, Appl. Catal. A 1998, 169, 271.[18] F. Fischer, H. Tropsch, Brennstoff Chem. 1928, 3, 39.[19] M. A. A. Aziz, H. D. Setiabudhi, L. P. Teh, N. H. R. Annuar, A. A. Jalil, J.

Taiwan Inst. Chem. E. 2019, 101, 139.[20] Y. Gao, J. Jiang, Y. Meng, F. Yan, A. Aihemaiti, Energy Convers. Manage.

2018, 171, 133.[21] O. Muraza, A. Galadima, Int. J. Energy Res. 2015, 39, 1196.[22] R. Singh, A. Dhir, S. K. Mohapatra, S. K. Mahla, Biomass Conv. Bioref.

2019.[23] A. Abdulrasheed, A. A. Jalil, Y. Gambo, M. Ibrahim, H. U. Hambali,

M. Y. S. Hamid, Renewable Sustainable Energy Rev. 2019, 108, 175.[24] W.-J. Jang, J.-O. Shim, H.-M. Kim, S.-Y. Yoo, H.-S. Roh, Catal. Today

2019, 324, 15.

[25] https://www.the-linde-group.com/de/news_and_media/press_re-leases/news_20151015.html, 15.05. 2015, Linde Corporate News.

[26] https://www.linde-engineering.com/de/about-linde-engineering/suc-cess-stories/technologies-more-with-less.

[27] C. M. Kalamaras, A. M. Efstathiou, “Hydrogen Production Technologies:Current State and Future Developments,” Conference Papers in Energy2013, 2013, 9.

[28] V. Subramani, P. Sharma, L. Zhang, K. Liu, C. Song, Hydrogen andSyngas Production and Purification Technologies: Hydrocarbon Process-ing for H2 Production, Wiley, New-York 2010, 14.

[29] J. N. Armor, Appl. Catal. A 1999, 176, 159.[30] S. Sircar, T. C. Golden, Sep. Sci. Technol. 2000, 35, 667.[31] W. Boll, E. Supp, G. Hochgesand, C. Higman, P. Kalteier, W. D. Müller,

M. Kriebel, H. Schlichting, H. Tanz, Ullmann’s Encyclopedia of IndustrialChemistry, Wiley-VCH, Weinheim 2007, 75.

[32] I. Wender, Fuel Process. Technol. 1996, 48, 189.[33] A. P. E. York, T. Xiao, M. L. H. Green, J. B. Claridge, Catal. Rev. 2007, 49,

511.[34] K. Liu, G. D. Deluga, A. Bitsch-Larsen, L. D. Schmidt, L. Zhang, Hydrogen

and Syngas Production and Purification Technologies: HydrocarbonProcessing for H2, Wiley, New-York 2010, 127.

[35] J. M. Ogden, T. G. Kreutz, M. Steinbugler, A. B. Cox, J. W. White,“Hydrogen energy systems studies”, In Proceedings of the 1996 U.S.DOE Hydrogen Program Review, Miami 1996.

[36] E. Supp, M. Brejc, W. Liebner, Ullmann’s Encyclopedia of IndustrialChemistry, Wiley-VCH, Weinheim 2007, 40.

[37] R. Kikuchi, Y. Iwasa, T. Takegushi, K. Eguchi, Appl. Catal. A 2005, 281,61.

[38] A. P. E. York, T. Xiao, M. L. H. Green, J. B. Claridge, Catal. Rev. 2007, 49,511.

[39] E. Schwab, A. Milanov, S. A. Schunk, A. Behrens, N. Schödel, Chem. Ing.Tech. 2015, 87, 347.

[40] R. Burch, M. I. Petch, Appl. Catal. A 1992, 88, 39.[41] E. C. Alyea, D. He, J. Wang, Appl. Catal. A 1993, 104, 77.[42] G. Kurz, S. Teuner, Erdöl und Kohle, Erdgas, Petrochemie – Hydrocarbon

Technology, 1990, 43, 171.[43] S. C. Teuner, P. Neumann, F. von Linde, Oil gas 2001, 44, 44.[44] J. R. Rostrup-Nielsen, L. J. Christiansen, Concepts in Syngas Manufacture,

Imperial College Press, London, 2011, 233.[45] A. P. E. York, T. C. Xiao, M. L. H. Green, J. B. Claridge, Catal. Rev. Sci. Eng.

2007, 49, 511.[46] Y. T. Shah, T. H. Gardner, Catal. Rev. 2014, 56, 476.[47] S. C. Teuner, P. Neumann, F. von Linde, Oil Gas Eur. Mag. 2001, 44.[48] a) I. C. Yates, C. N. Satterfield, Energ Fuel. 1991, 5, 168;b) J. J. C.

Geerlings, M. C. Zonnevylle, C. P. M. De Groot, Surf. Sci. 1991, 241, 302.[49] F. Hayer, H. Bakhtiary-Davijany, R. Myrstad, A. Holmen, P. Pfeifer, H. J.

Venvik, Chem. Eng. J. 2011, 167, 610.[50] N. Schödel, E. Haidegger, A. Behrens, V. Göke, C. Thaller, H.

Schmaderer, EP Patent 2809640B1, 2012.[51] H. A. Wright, M. K. Robertson, W. Jiang, US Patent 0002265A1, 2018.[52] T. Roussière, Dissertation, Karlsruher Institut für Technologie, Karlsruhe

2013.[53] O. Deutschmann, S. Tischer, C. Correa, D. Chatterjee, S. Kleditzsch, V. M.

Janardhanan, N. Mladenov, H. D. Minh, H. Karadeniz, M. Hettel, H.Gossler, DETCHEM software package, 2.6 ed., www.detchem.comKarlsruhe, 2017.

[54] H. S. Bengaard, J. K. Nørskov, J. Seheste, B. S. Clausen, L. P. Nielsen,A. M. Molenbroek, J. Catal. 2002, 209, 365.

[55] J.-H. Kim, D. J. Suh, T.-J. Park, K.-L. Kim, Appl. Catal. A 2000, 197, 191.[56] O. A. Bereketidou, M. A. Goula, Catal. Today 2012, 195, 93.[57] M. A. Goula, N. D. Charisiou, K. N. Papageridis, A. Delimitis, E.

Pachatouridou, E. F. Iliopoulou, Int. J. Hydrogen Energy 2015, 40, 9183.[58] R. Y. Chein, Y. C. Chen, C. T. Yu, J. N. Chung, J. Nat. Gas Sci. Eng. 2015,

26, 617.[59] Ş. Özkara-Aydınoğlu Int. J. Hydrogen Energy 2010, 35, 12821.[60] K. Liu, C. Song, V. Subramani, Hydrogen and Syngas Production and

Purification Technologies, John Wiley & Sons, Inc., Hoboken, NJ, USA,2009.

[61] K. Liu, Hydrogen and Syngas Production and Purification Technologies,John Wiley & Sons, Hoboken 2010.

[62] U. Olsbye, T. Wurzel, L. Mleczko, Ind. Eng. Chem. Res. 1997, 36, 5180.[63] M. C. J. Bradford, M. A. Vannice, Catal. Rev. Sci. Eng. 1999, 41, 1.[64] A. D. Ballarini, S. R. de Miguel, E. L. Jablonski, O. A. Scelza, A. A. Castro,

Catal. Today 2005, 107, 481.

Reviews

16ChemCatChem 2020, 12, 1–19 www.chemcatchem.org © 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Wiley VCH Freitag, 14.02.2020

2099 / 158042 [S. 16/19] 1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

[65] A. K. Avetisov, J. R. Rostrup-Nielsen, V. L. Kuchaev, J. H. B. Hansen, A. G.Zyskin, E. N. Shapatina, J. Mol. Catal. A 2010, 315, 155.

[66] D. Pakhare, J. Spivey, Chem. Soc. Rev. 2014, 43, 7813.[67] D. L. Serrano-Lotina, Int. J. Hydrogen Energy 2013, 39, 4089.[68] A. Khajeh Talkhoncheh, M. Haghighi, J. Nat. Gas Sci. Eng. 2015, 23, 16.[69] A. T. Ashcroft, A. K. Cheetham, M. L. H. Green, P. D. F. Vernon, Nature

1991, 352, 225.[70] J. R. Rostrup-Nielsen, J. H. Bak Hansen, J. Catal. 1993, 144, 38.[71] P. Ferreira-Aparicio, A. Guerrero-Ruiz, I. Rodríguez-Ramos, I. Rodríguez-

Ramos, Appl. Catal. A 1998, 170, 177.[72] P. M. Mortensen, I. Dybkjær, Appl. Catal. A 2015, 495, 141.[73] G. Jones, J. G. Jakobsen, S. S. Shim, J. Kleis, M. P. Andersson, J.

Rossmeisl, F. Abild-Pedersen, T. Bligaard, S. Helveg, B. Hinnemann, J. R.Rostrup-Nielsen, I. Chorkendorff, J. Sehested, J. K. Nørskov, J. Catal.2008, 259, 147.

[74] K. Nagaoka, K. Takanabe, K. Aika, Appl. Catal. A 2004, 268, 151.[75] E. Ruckenstein, H. Y. Wang, J. Catal. 2002, 205, 289.[76] U. S. Geological Survey, Mineral Commodity Summaries 2013, U.S.

GeologicalSurvey, 2013.[77] W.-J. Jang, J.-O. Shim, H.-M. Kim, S.-Y. Yoo, H.-S. Roh, Catal. Today

2019, 324, 15.[78] E. Ruckenstein, H. Y. Wang, Appl. Catal. A 2000, 204, 257.[79] K. Nagaoka, K. Takanabe, K. Aika, Appl. Catal. A 2004, 268, 151.[80] N. El Hassan, M. N. Kaydouh, H. Geagea, H. El Zein, K. Jabbour, S.

Casale, H. El Zakhem, P. Massiani, Appl. Catal. A 2016, 520, 114.[81] K. Takanabe, K. Nagaoka, K. Nariai, K. Aika, J. Catal. 2005, 232, 268.[82] J. Horlyck, C. Lawrey, E. C. Lovell, R. Amal, J. Scott, Chem. Eng. J. 2018,

352, 572.[83] V. M. Gonzalez-Delacruz, R. Pereñiguez, F. Ternero, J. P. Holgado, A.

Caballero, J. Phys. Chem. C 2012, 116, 2919.[84] J. Zhang, H. Wang, A. Dalai, J. Catal. 2007, 249, 300.[85] Y. Chen, J. Ren, Catal. Lett. 1994, 29, 39.[86] A. A. Firsova, Y. P. Tyulenin, T. I. Khomenko, V. N. Korchak, O. V. Krylov,

Kinet. Catal. 2003, 44, 819.[87] J.-E. Min, Y.-J. Lee, H.-G. Park, C. Zhang, K.-W. Jun, J. Ind. Eng. Chem.

2015, 26, 375.[88] A. M. Gadalla, B. Bower, Chem. Eng. Sci. 1988, 43, 3049.[89] S. A. Stevenson, J. A. Dumesic, R. T. K. Baker, E. Ruckenstein (Eds.),

Metal-Support Interactions in Catalysis, Sintering, and Redispersion, VanNostrand Reinhold, New York 1987, 141.

[90] J. Guo, H. Lou, H. Zhao, D. Chai, X. Zheng, Appl. Catal. A 2004, 273, 75.[91] M. F. Mark, F. Mark, W. F. Maier, Chem. Eng. Technol. 1997, 20, 361.[92] Y. Kathiraser, U. Oemar, E. T. Saw, Z. Li, S. Kawi, Chem. Eng. J. 2015, 278,

62.[93] E. Akpan, Y. Sun, P. Kumar, H. Ibrahim, A. Aboudheir, R. Idem, Chem.

Eng. Sci. 2007, 62, 4012.[94] A. M. Becerra, M. E. Iriarte, A. E. Castro-Luna, React. Kinet. Catal. Lett.

2003, 79, 119.[95] A. Erdohelyi, J. Cserenyi, F. Solymosi, J. Catal. 1993, 141, 287.[96] J. T. Richardson, S. A. Paripatyadar, Appl. Catal. 1990, 61, 293.[97] J. G. Xu, G. F. Froment, AIChE J. 1989, 35, 88.[98] D. L. Trimm, Z. I. Önsan, Catal. Rev. 2001, 43, 31.[99] N. Gokon, Y. Osawa, D. Nakazawa, T. Kodama, Int. J. Hydrogen Energy

2009, 34, 1787.[100] J. Xu, G. F. Froment, AIChE J. 1989, 35, 88.[101] Z. L. Zhang, X. E. Verykios, Catal. Today 1994, 21, 589.[102] Y.-A. Zhu, D. Chen, X.-G. Zhou, W.-K. Yuan, Catal. Today 2009, 148, 260.[103] J. Wei, E. Iglesia, J. Catal. 2004, 224, 370.[104] J. Wei, E. Iglesia, J. Phys. Chem. B 2004, 108, 7253.[105] H. Y. Wang, C. T. Au, Catal. Lett. 1996, 38, 77.[106] L. A. Schulz, L. C. Kahle, K. H. Delgado, S. A. Schunk, A. Jentys, O.

Deutschmann, J. A. Lercher, Appl. Catal. A 2015, 504, 599.[107] R. Bisson, M. Sacchi, T. T. Dang, B. Yoder, P. Maroni, R. D. Beck, J. Phys.

Chem. 2007, A111, 12679.[108] S. Nave, B. Jackson, J. Chem. Phys. 2009, 130, 054701.[109] E. J. Cairns, A. D. Tevebaugh, G. J. Holm, J. Electrochem. Soc. 1963, 110,

1025.[110] Z. Jaworski, B. Zakrzewska, P. Pianko-Oprych, Rev. Chem. Eng. 2017, 33,

217.[111] H. Zhu, B. L. Kee, C. Karakaya, R. O’Hayre, R. J. Kee, Catal. Today 2019,

331, 7.[112] A. Giehr, L. Maier, S. A. Schunk, O. Deutschmann, ChemCatChem 2018,

10, 751.[113] X. Fang, C. Peng, H. Peng, W. Liu, X. Xu, X. Wang, C. Li, W. Zhou,

ChemCatChem 2015, 7, 3753.

[114] H. Dedder, K. K-hler, B. Krause, K. Mette, S. Kehl, M. Behrens, V. Scherer,M. Muhler, Catal. Sci. Technol. 2014, 4, 3317.

[115] D. San-José-Alonso, J. Juan-Juan, M. J. Illán-Gjmez, M. C. Román-Martínez, Appl. Catal. A 2009, 371, 54.

[116] L. C. S. Kahle, T. Roussière, L. Maier, K. H. Delgado, G. Wasserschaff, S. A.Schunk, O. Deutschmann, Ind. Eng. Chem. Res. 2013, 52, 11920.

[117] A. Li, K. Norinaga, W. G. Zhang, O. Deutschmann, Compos. Sci. Technol.2008, 68, 1097.

[118] A. Becker, K. J. Huttinger, Carbon 1998, 36, 213.[119] A. J. Li, O. Deutschmann, Chem. Eng. Sci. 2007, 62, 4976.[120] I. Naydenova, M. Nullmeier, J. Warnatz, P. A. Vlasov, Combust. Sci.

Technol. 2004, 176, 1667.[121] K. Norinaga, O. Deutschmann, N. Saegusa, J. Hayashi, J. Anal. Appl.

Pyrolysis 2009, 86, 148.[122] A. L. Cava, D. L. Trimm, Carbon 1978, 16, 505.[123] H. J. Curran, P. Gaffuri, W. J. Pitz, C. K. A. Westbrook, Combust. Flame

2002, 129, 253.[124] P. Dagaut, G. Pengloan, A. Ristori, Phys. Chem. Chem. Phys. 2002, 4,

1846.[125] V. Golovitchev, F. Tao, J. Chomiak, SAE [Tech. Pap.] 1999, 1, 3552.[126] C. A. Mims, R. Mauti, A. M. Dean, K. D. Rose, J. Phys. Chem. 1994, 98,

13357.[127] K. M. Walters, A. M. Dean, H. Y. Zhu, R. J. Kee, J. Power Sources 2003,

123, 182.[128] T. Kaltschmitt, L. Maier, M. Hartmann, C. Hauck, O. Deutschmann, Proc.

Combust. Inst. 2011, 33, 3177.[129] A. Li, K. Norinaga, W. G. Zhang, O. Deutschmann, Compos. Sci. Technol.

2008, 68, 1097.[130] A. Becker, Z. Hu, K. J. Huttinger, Fuel 2000, 79, 1573.[131] W. G. Zhang, Z. J. Hu, K. J. Huttinger, Carbon 2002, 40, 2529.[132] Z. Li, X. Hu, L. Zhang, S. Liu, G. Lu, Appl. Catal. A 2012, 417, 281.[133] T. Roussière, L. Schulz, K. M. Schelkle, G. Wasserschaff, A. Milanov, E.

Schwab, O. Deutschmann, A. Jentys, J. Lercher, S. A. Schunk, Chem-CatChem 2014, 6, 1447.

[134] L. A. Schulz, L. C. S. Kahle, K. H. Delgado, S. A. Schunk, A. Jentys, O.Deutschmann, J. A. Lercher, Appl. Catal. A 2015, 504, 599.

[135] J. G. Park, A. N. Cormack, J. Solid State Chem. 1996, 121, 278.[136] N. Iyi, Z. Inoue, S. Takekawa, S. Kimura, J. Solid State Chem. 1984, 54,

70.[137] P. D. Dernier, J. P. Remeika, J. Solid State Chem. 1976, 17, 245.[138] N. Abdel Karim Aramouni, Energy Convers. Manage. 2017, 150, 614.[139] J. R. Rostrup-Nielsen, J. Sehested, Adv. Catal. 2002, 47, 65.[140] F. Abild-Pedersen, J. K. Nørskov, Phys. Rev. B 2006, 73, 115419.[141] J. Titus, T. Roussière, G. Wasserschaff, S. A. Schunk, A. Milanov, E.

Schwab, G. Wagner, O. Oeckler, R. Gläser, Catal. Today 2016, 270, 68.[142] S. Takenaka, E. Kato, Y. Tomikubo, K. Otsuka, J. Catal. 2003, 219, 176.[143] B. Q. Xu, J. M. Wei, Y. T. Yu, J. L. Li, Q. M. Zhu, Top. Catal. 2003, 22, 77.[144] M. Steib, A. Jentys, J. A. Lercher, J. Phys. Conf. Ser. 2016, 712, 012049.[145] M. Steib, Y. Lou, A. Jentys, J. A. Lercher, ChemCatChem 2017, 9, 3809.[146] P. Ferreira-Aparicio, I. Rodríguez-Ramos, J. A. Anderson, A. Guerrero-

Ruiz, Appl. Catal. A. 2000, 202, 183.[147] S. Das, M. Sengupta, J. Patel, A. Bordoloi, Appl. Catal. A 2017, 545, 113.[148] J. Titus, M. Goepel, S. A. Schunk, N. Wilde, R. Gläser, Catal. Commun.

2017, 100, 76.[149] F. Liu, K. Huang, Q. Wu, S. Dai, Adv. Mater. 2017,29, 1700445.[150] Y. Zhan, Q. Han, S. Pan, X. Kan, J. Mi, F. Liu, Y. Cao, C. Au, L. Jiang, Ind.

Eng. Chem. Res. 2019, 58, 7980.[151] T. Roussière, K. M. Schelkle, S. Titlbach, G. Wasserschaff, A. Milanov, G.

Cox, E. Schwab, O. Deutschmann, L. Schulz, A. Jentys, ChemCatChem2014, 6, 1438.

[152] P. K. Sahu, B. D. Kulkarni, R. B. Khomane, S. A. Pardhy, U. D. Phalgune,P. Rajmohanan, R. Pasricha, Chem. Commun. 2003, 1876.

[153] S. Laassiri, D. Duprez, S. Royer, H. Alamdari, Catal. Sci. Technol. 2011, 1,1124.

[154] J. Gao, C. Jia, M. Zhang, F. Gu, G. Xu, Z. Zhong, F. Su, RSC Adv. 2013, 3,18156.

[155] A. J. Zarur, J. Y. Ying, Nature 2000, 403, 65.[156] L. Kunz, L. Maier, S. Tischer, O. Deutschmann, in Model. Simul. Heterog.

Catal. React., Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim,Germany, 2011, pp. 113.

[157] K. Ikkour, D. Sellam, A. Kiennemann, S. Tezkratt, O. Cherifi, Catal. Lett.2009, 132, 213.

[158] I. Luisetto, S. Tuti, E. Di Bartolomeo, Int. J. Hydrogen Energy 2012, 37,15992.

[159] V. C. H. Kroll, H. M. Swann, C. Mirodatos, J. Catal. 1996, 161, 409.

Reviews

17ChemCatChem 2020, 12, 1–19 www.chemcatchem.org © 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Wiley VCH Freitag, 14.02.2020

2099 / 158042 [S. 17/19] 1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

[160] A. Bhattacharyya, V. W. Chang, Stud. Surf. Sci. Catal. 1994, 88, 207.[161] T. H. Gardner, D. Shekhawat, D. A. Berry, M. W. Smith, M. Salazar, E. L.

Kugler, Appl. Catal. A 2007, 323, 1.[162] L. Zhang, X. Wang, B. Tan, U. S. Ozkan, J. Mol. Catal. A 2009, 297, 26.[163] L. A. Schulz, E. Schwab, O. Deutschmann, G. Wasserschaff, K. M.

Schelkle, A. Jentys, A. Milanov, J. Lercher, T. Roussière, S. A. Schunk,ChemCatChem 2014, 6, 1447.

[164] I. Luisetto, S. Tuti, E. Di Bartolomeo, Int. J. Hydrogen Energy 2012,37,15992.

[165] A. Giehr, S. Angeli, L. Maier, S. A. Schunk, O. Deutschmann, submittedto Angew. Chem..

[166] S. A. Theofanidis, V. V. Galvita, H. Poelman, G. B. Marin, ACS Catal.2015, 5, 3028.

[167] S. A. Theofanidis, R. Batchu, V. V. Galvita, H. Poelman, G. B. Marin, Appl.Catal. B 2016, 185, 42.

[168] D. K. Kim, K. Stöwe, F. Müller, W. F. Maier, J. Catal. 2007, 247, 101.

Manuscript received: November 13, 2019Revised manuscript received: January 2, 2020Accepted manuscript online: January 12, 2020Version of record online: ■■■, ■■■■

Reviews

18ChemCatChem 2020, 12, 1–19 www.chemcatchem.org © 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Wiley VCH Freitag, 14.02.2020

2099 / 158042 [S. 18/19] 1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

REVIEWS

Getting in formation: Cokeformation is the main reason for de-activation in dry reforming ofmethane under industrially relevantconditions due to the necessary hightemperatures (thermodynamic) andpressures (economical). Therefore,understanding coke formation is animportant factor for the develop-ment of novel coke resilient catalystssystems with high activity andstability.

Dr. K. Wittich, Dr. M. Krämer, Dr. N.Bottke, Dr. S. A. Schunk*

1 – 19

Catalytic Dry Reforming ofMethane: Insights from ModelSystems

Reviews

Wiley VCH Freitag, 14.02.2020

2099 / 158042 [S. 19/19] 1