Impact of Carbonyl Formation on Cobalt Ripening over Titania Surface

Subscriber access provided by University of South Florida

Article

Impact of Carbonyl Formation on Cobalt Ripening over Titania Surface Kevin Kelsey Brian Duff, Leonardo Spanu, and Nicholas D. M. Hine J. Phys. Chem. C, Just Accepted Manuscript • DOI: 10.1021/acs.jpcc.7b05371 • Publication Date (Web): 03 Jul 2017 Downloaded from http://pubs.acs.org on July 10, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

The Journal of Physical Chemistry C is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 26

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 58 59 60

The Journal of Physical Chemistry

Impact of Carbonyl Formation on Cobalt Ripening over Titania Surface Kevin K. B. Duff,∗,† Leonardo Spanu,∗,‡ and Nicholas D. M. Hine∗,¶ †TCM Group, Cavendish Laboratory, University of Cambridge, UK CB3 0HE ‡Shell India Markets Private Limited, Bangalore, 560048, India ¶Department of Physics, University of Warwick, Coventry, UK CV4 7AL E-mail: [email protected]; [email protected]; [email protected]

July 3, 2017

1 ACS Paragon Plus Environment


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 58 59 60

Abstract We present density functional theory calculations of the adsorption and migration energies of different cobalt moieties on the anatase (101) surface. Surface diffusion of active metal sites is a crucial step in the ripening mechanism, one the primary causes for the loss of active surface area of cobalt Fischer-Tropsch catalyst. Our main objective is to clarify the impact of gas phase molecules on the transport properties of surfaceadsorbed cobalt. Water molecule physisorbed to the (101) surface have a negligible impact of the adsorption and diffusion of cobalt ions. On the contrary, cobalt carbonyl species have the largest effect on the energetics of adsorption and surface diffusion. The formation of Co(CO)3 is favourable in realistic reactor conditions and drastically decreases the surface binding energy, making transport via gas phase an alternative viable pathways for cobalt mobility.

Introduction Sintering is one of the possible deactivation pathways of transition metal catalysts 1,2 and is believed to affect Fischer-Tropsch (FT) catalysts particularly in the first few weeks of the catalyst lifetime. 3 Sintering is made up of two distinct processes that act to reduce active surface area: coalescence, which is the motion of large nanoparticles across the surface, and ripening, which is the breaking off, surface diffusion, and recombination of small catalyst particles. Of these, ripening is particularly difficult to probe experimentally in situ. This poses a challenge in the identification of factors that may be accelerating the process, and motivates an alternative computational approach. Due to their role in catalysis, the formation, growth, mobility, and interactions of transition metal atoms and clusters on metal oxide supports have been widely studied both experimentally and computationally using DFT. Examples include Co, 4 Ni, 5 Pd, 6–12 Ag, 10 Pt, 10,13–19 and Au 17,20–22 on both anatase and rutile TiO2 , and Ni, 23 Cu, 24 Pd, 24–26 Ag, 1,27 Pt 28,29 and others 30 on alumina. 2 ACS Paragon Plus Environment

Page 2 of 26

Page 3 of 26

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 58 59 60


To relate the results of DFT calculations to experimentally-relevant quantities such as macroscopic diffusion coefficients of adsorbate on a surface, a random walk model can be parameterised using adsorption and transition state energies. 31 Even in more complicated surfaces, where there may be more than one unique adsorption site, such as is the case for anatase (101) surface, these energies remain central to describing surface diffusion. 32 This motivates the comparison of absorption and transition state energies for a range of cobalt-containing adsorbates on the support surface as a measure of surface mobility. In the environment of a Fischer-Tropsch reactor, the catalyst surface is exposed to a range of gas-phase species, any of which may be the determining factor to cobalt transport, and thus to ripening. These species include the reactants and gas-phase products of the process and their derivatives. Several of these species can bind to surface-adsorbed cobalt and affect its adsorption and mobility properties. Since water is produced as a product of the reaction, surface-adsorbed water molecules may affect both adsorption and mobility, as well as the formation of surface species. In this article, we aim to study adsorption and diffusion of all potentially relevant species, so as to identify which gas-phase species present in a Fischer-Tropsch reactor are most active in the ripening process. The overall contribution of a species to cobalt mobility depends in fact on multiple factors: the binding of the species to the surface, the ability of the species to diffuse across the surface, and the concentration of the species on the surface. We focus in particular on three main quantities derived from DFT-based total energy differences between different systems: the absorption energy ∆Ebind , the migration barrier ∆Emig , and the formation energy ∆Eform . All DFT calculations are performed for adsorbate species containing a single cobalt attached to the anatase TiO2 (101) surface, in presence of different absorbates from gas-phase species (O, CO, and OH groups). The Wulff construction of anatase TiO2 indicates in fact that the (101) surface is dominant, with (001) expected to form as a secondary surface. 33 In the case of anatase (101), water molecules adsorb non-dissociatively: 34 we have therefore considered both dry surface and water covered (1 monolayer, (ML)) case, to clarify the role of water on



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 58 59 60

Page 4 of 26

the surface mobility. We find that the addition of water to the surface does not significantly affect the energetics of cobalt adsorption on the surface. However, we find compelling evidence that cobalt carbonyl can be expected to play a crucial role in the overall balance of cobalt transport.

Methods Adsorption, Mobility and Formation Energies The adsorption energy of a species on a surface is defined in Equation 1 as the energy required to completely remove a bonded instance of that species from the surface and return the surface to its pristine state. This is calculated as a difference in the total energies of three systems: ∆Ebind = Eads + Eslab − Eslab+ads

(1)

Here, Eslab+ads is the total energy of the surface with the adsorbate, Esurf is the total energy of the pristine surface, and Eads is the total energy of an isolated adsorbate in vacuum. We neglect thermal effects, and thus use total energies here rather than Gibbs free energies. Given two adsorption sites, a transition state between them can be identified using the nudged elastic band (NEB) method. 35,36 The energy of this transition state relative to the starting site defines the migration barrier. NEB is a chain-of-states method in which several intermediate images are connected with springs of natural length 0 and relaxed, with the spring force perpendicular to the path and the real force parallel to the path projected out. This relaxation is not global and depends on the choice of initial path, which must be considered carefully where the transition involves many coordinated processes. However for the relatively simple case of the surface motion of an adsorbate with a single surface contact, this is a minor concern, and linear interpolation can be used to construct a sensible initial path.


Page 5 of 26

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 58 59 60


The formation energy expresses the energy cost to form the adsorbate system relative to the constituent species existing in realistic ‘reservoir’ conditions within a reactor vessel. We use an approach similar to that of Finnis et al. 37,38 in our definitions of the chemical potentials of the reservoir species, and we use this particularly to estimate formation energies of cobalt carbonyl adsorbates under reactor conditions. In this approach we consider the surface-adsorbed cobalt carbonyl to be in equilibrium with a bath of the CO feedstock at a selected temperature and partial pressure representative of realistic conditions. The formation energy Eform of a species Co(CO)n on the anatase surface can be expressed as in Equation 2 in terms of the total energies of the adsorbed species and the chemical potential of the gas-phase species µCO .

Eform = Eslab+Co(CO)n − Eslab+Co − nµCO (T, pCO )

(2)

The chemical potential of the gas-phase species at a given temperature and partial pressure can be approximated by applying thermodynamics relations of an ideal gas of rigid dumbbells to the experimental Gibbs free energy of formation at standard conditions. A discussion of our calculation of these values without requiring DFT calculations on gas-phase species is provided in the supplementary information. The supplementary information also provides details of DFT calculations on bulk materials used throughout this study.

Computational Details Calculations were carried out using the plane-wave DFT package VASP. 39–42 This utilises the PAW method, with standard PAW datasets retaining 17, 12, 6, 4, and 1 electrons respectively in the valence for Co, Ti, O, C and H. The PBE exchange-correlation functional was used for all calculations. 43,44 It was determined that special treatments of metallic or highly correlated systems were not required as Bader charge analysis 45 indicates the adsorbed cobalt remains in a +2 oxidation state on the surface with only partial reduction in some cases, typically



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 58 59 60

when under-coordinated (See supplementary information for details). The model surface used throughout consisted of a 4-layer 2 × 2 slab of the (101) surface of vacuum. A 2 × 2 × 1 gamma-centred Monkhorst Pack grid was used for k-point with 14 A sampling on this surface, equivalent to a 4 × 4 × 1 grid on the primitive surface or a k-point grid spacing of 0.098 A

−1

by 0.131 A

−1

. For all calculations a plane-wave cut-off by 0.037 A −1

energy of 400 eV was applied. The lowest layer of 24 atoms was held fixed throughout, in positions corresponding to the bulk structure, to better approximate the limit of a semiinfinite slab. It is straightforward to identify the two main sites where adsorption is expected to be favourable for Co: the site that turns out to be the most favourable, which we call site A, is located between two O atoms on a ridge, with a further two O atoms relatively nearby, as shown in Figure 1(a). We have found that Co also binds strongly at a second site, which we call site B, as shown in Figure 1(b). This site enables bonding of Co to two nearby O atoms.

Figure 1: Adsorption sites of Co on the anatase (101) surface. (a) site A is the strongest adsorption site and (b) site B is a second, weaker adsorption site. We construct geometry models comprising a single Co ion and a Co ion with various functional groups adsorbed at each of these sites. The groups considered are -O, -CO and 6 ACS Paragon Plus Environment

Page 6 of 26

Page 7 of 26

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 58 59 60


-OH, which can all easily be formed from the products or reactants of the FT process. In practice the surface is rarely dry under reaction conditions, and because of the relatively strong interaction of the 5-fold coordinated surface titanium atoms with water molecules, the geometry of a single monolayer (ML) coverage is relatively stable and partially ordered. 46,47 The geometry of 1ML coverage of water on the surface is known from previous calculations, 48 allowing for the effect of surface hydration to be considered. To this end a second model surface is constructed for the 1ML hydrated surface, and calculations repeated for each of the species adsorbed on this surface. Where necessary, water molecules attached to the surface are moved or removed to make way for the Co-containing species. Using the relaxed configurations as beginning and end points, we then use the VASP Transition State Tools (VTST)

49,50

climbing-image NEB functionality to determine diffu-

sion barriers between the adsorption sites on the surface. These are defined as the difference between the minimum and maximum points on the continuous pathway between two sites which has the lowest overall maximum energy. Quantitative determination of activation energies of these migration processes would clearly require explicit modelling of thermal effects via molecular dynamics simulation. However, we expect that the static NEB calculations presented here will be predictive of the relative migration barriers between different species. In addition to comparing different species, we consider the effect on surface adsorption of adding multiple groups to the Co ion. We do this by forming Co(CO)2 and Co(CO)3 on the dry surface. A mobility study is also performed to examine the effect on surface mobility of additional groups.



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 58 59 60

Page 8 of 26

Results Dry Surface Adsorption Figure 2 depicts the adsorption energies Eads of several different species that can be formed from interactions between a surface-adsorbed cobalt ion and the reactor environment. In each case, the adsorption is made stronger by the the availability of four O atoms for the Co to coordinate with, including the four immediately available at site A and the two at site B. In some cases it is also strengthened by the sharing of the functional group with the exposed Ti atom near site A. For example, the site A configuration of CoO shown in Figure 3 has the oxygen coordinating a surface Ti, leading to an overall stronger adsorption. The site B position does not have the same features and limits Co – O coordination, resulting in a much lower adsorption energy.

330

Co

245 380

CoO

179 323

CoOH

278 239

CoCO

192 160 162

Co(CO)2 Co(CO)3

98

Site A Site B

73 Adsorption Energy (kJ/mol)

Figure 2: Adsorption energies ∆Ebind of adsorbates on the anatase (101) surface. ∆Ebind is defined in Equation 1. Site A is somewhat preferred for most adsorbates. Addition of multiple CO moieties significantly reduces the adsorption energy. 8 ACS Paragon Plus Environment

Page 9 of 26

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 58 59 60


Figure 3: CoO adsorbed at site A of the TiO2 surface, identified by a circle. While the added O atom decreases the surface coordination of the Co, it also coordinates with an exposed Ti atom on the surface. The net effect is a stronger surface adsorption than that of lone Co. The addition of CO and OH also reduces the coordination of the Co to surface O atoms. Notably, the addition of further gas phase species to the adsorbate further decreases coordination to surface O atoms, as demonstrated by CoCO, Co(CO)2 , and Co(CO)3 . Figure 4 depicts the site A adsorption configuration of each of these carbonyl-containing adsorbates. The CO groups do not interact with the surface and this decreased Co surface coordination accompanies a significant decrease in adsorption energy. It is worth noting that Co(CO)3 no longer sits between the bridging oxygens at site A as it coordinates to only 1 surface atom. No adsorbed configurations for further cobalt carbonyls were found.

Figure 4: Site A adsorption configurations of (a) CoCO (b) Co(CO)2 and (c) Co(CO)3 on the (101) anatase surface. 9 ACS Paragon Plus Environment


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 58 59 60

Page 10 of 26

Dry Surface Mobility Adsorbate mobility on the surface is central to the ripening process. We have investigated the energy barrier to motion between sites A and B using the aforementioned nudged elastic band method. The NEB chain between sites A and B was initialized by linear interpolation of the relaxed atomic positions. This results in each adsorbate traversing a path rotating around a surface oxygen atom between sites A and B. While this single path does not provide enough information to create a model of full surface diffusion, it does indicate how the inclusion of various species may affect adsorbate mobility. Figure 5 shows the mobility barrier relative to the site A energy of the adsorbate.

Co

106

CoO

201

CoOH

55

CoCO

49

Co(CO)2 Co(CO)3

38 25 Mobility Barrier (kJ/mol)

Figure 5: Site A to site B mobility barriers of several species across the anatase (101) surface, defined relative to the site A adsorption energy. The strong surface interaction observed in CoO at site A leads to a large barrier to mobility across the surface compared to other adsorbates. An intermediate state was found in the path between sites A and B with an adsorption energy of 309 kJ mol−1 and a mobility barrier from site A of 73 kJ mol−1 . The geometry of this intermediate is presented in the supplementary information. The mobility barrier of both CoCO and CoOH are about half that of Co. The addition of


Page 11 of 26

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 58 59 60


another carbonyl group further decreases the barrier, potentially due to the decreased cost of partially desorbing from the surface. Co(CO)3 has a mobility barrier that is essentially just the difference in adsorption energies at sites A and B; note that because Co(CO)3 is coordinated to only one surface O at site A this transition does not involve any breaking or forming of chemical bonds.

Effects of Surface Hydration Water is produced by the Fischer Tropsch process and is able to bind to the support surface. The structure of thin water layers on the anatase (101) surface has been previously investigated. 51 The presence of water can affect the adsorption and mobility of molecules involved in ripening. Figure 6 shows the adsorbed configuration of cobalt at sites A and B of the hydrated surface. Figure 7 shows the adsorption energies Eads of a selection of the species investigated on a surface with a single monolayer (ML) of water coverage.

Figure 6: Adsorbed configuration of a cobalt ion at (a) site A and (b) site B on the monolayerhydrated anatase (101) surface. Neither CoO nor CoOH were found to have stable adsorbed configurations. Both species induced splitting of a molecule from the water layer to form a surface-adsorbed OH group. 11 ACS Paragon Plus Environment


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 58 59 60

Co

279 219 190

CoCO 107

Co(CO)2 Co(CO)3

169 147

Page 12 of 26

322

Site A Site B Interm.

79 Adsorption Energy (kJ/mol)

Figure 7: Adsorption energies of a selection of adsorbates on the 1ML hydrated anatase (101) surface as defined in Equation 1. An intermediate state of Co(CO)2 is also identified. Figure 8 shows the configuration of CoOH at site A as a water molecule and a surfaceadsorbed cobalt ion. Figure 9 shows the configuration of a CoO at site B as a CoOH molecule. This indicates that the strong interactions between these adsorbates and the water present in the reactor may limit their lifetime and contribution to Co mobility. Because of the lack of stable adsorbed configurations, no migration barriers are reported in these cases. The addition of water to the surface does not significantly affect the adsorption energies of Co and CoCO, though it does change the surface O coordination of CoCO slightly. Reductions are seen from 330 kJ mol−1 to 322 kJ mol−1 , and from 239 kJ mol−1 to 219 kJ mol−1 respectively. Migration barriers for these two species were also investigated, and again small reductions from the dry results were found; from 106 kJ mol−1 to 103 kJ mol−1 and from 49 kJ mol−1 to 43 kJ mol−1 respectively. The presence of CO therefore still has a strong and comparable impact on Co mobility in the presence of surface hydration. In the case of Co(CO)2 on the hydrated surface, a significant reduction of the site A adsorption energy can be observed relative to that on the dry surface. Similarly, the mobility barrier, which on the dry surface is 38 kJ mol−1 , was determined to be very small on the hydrated surface at under 1 kJ mol−1 . An intermediate state was also found with an adsorption


Page 13 of 26

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 58 59 60


Figure 8: Geometry of CoOH at site A of the hydrated surface. The adsorbate interacted with the water monolayer to form a Co ion at its adsorption site and a water molecule. The dotted circle identifies the surface-adsorbed OH group created from the water monolayer.



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 58 59 60

Figure 9: Geometry of CoO at site B of the hydrated surface. The adsorbate interacted with the water monolayer to form CoOH. The solid circle identifies the adsorbed CoOH molecule. The dotted circle identifies the surface-adsorbed OH group created from the water monolayer.


Page 14 of 26

Page 15 of 26

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 58 59 60


energy of 169 kJ mol−1 . While this study is concerned with adsorption in proximity to the two main adsorption sites of a lone cobalt ion where possible, it is likely that Co(CO)2 would transition quickly into this state. This intermediate is depicted in Figure 10. The fact that such a strongly adsorbed intermediate was found only on the hydrated surface indicates the possibility of deeper effects of water coverage on adsorption and mobility that are beyond the scope of this study; here we only consider simple translational motion without chemically distinct intermediates or the effects of coordinated motion of the adsorbate and water molecules. A single adsorbed configuration of Co(CO)3 on the hydrated surface was found, with an adsorption energy of 79 kJ mol−1 . It is coordinated with the same surface O atom as it is on the dry surface.

Figure 10: Geometry of the intermediate state of Co(CO)2 found on the path between sites A and B on the hydrated surface. The existence of an intermediate state with a significantly higher adsorption energy than that of sites A or B indicates further interaction with the water monolayer beyond the scope of this study.



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 58 59 60

Page 16 of 26

Gas Phase vs. Surface Transport of Carbonyls Desorption into the gas phase and subsequent re-adsorption to the support surface is a possible cobalt transport method that could compete with surface motion, particularly for species with low adsorption energy. The larger carbonyl species investigated have notably low adsorption energies, indicating that if they form in significant quantities on the surface, then gas phase transport will likely be a competing process in sintering. We therefore investigate the formation energies of the various cobalt carbonyl species from surface-adsorbed cobalt at site A. Our aim is to determine which carbonyl species is likely to be most prevalent under typical reactor conditions. The reaction under consideration here is: Co(surf) + (CO)n(g) −−→ Co(CO)n(surf)

(3)

Standard thermochemistry data for the CO gas and the CoO solid were obtained from the NIST Chemistry Webbook. 52,53 The required elemental chemical potentials were obtained via total energy calculations for CoO and Co bulk crystals, and for graphitic carbon. These were obtained using well-converged total energy calculations performed using VASP with methodological choices designed to produce the same high level of convergence as the main calculation. Figure 11 shows formation energies parameterised to match conditions in a reactor at 500 K in equilibrium with CO gas. These results indicate that Co(CO)3 is, by a considerable margin, the most favourable species, over a wide range of CO pressure conditions. In fact, CoCO only becomes preferred at unrealistically low partial pressures, under 3 × 10−23 atm. At 10 atm Co(CO)3 remains favourable up to temperatures exceeding 1500 K. It can therefore be expected to form in conditions where surface-adsorbed cobalt ions exist on the surface in contact with CO feedstock.


Page 17 of 26

60 Formation Energy (kJ/mol)

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 58 59 60


CoCO Co(CO)2 Co(CO)3

40 20 0 -20 -40 -60 -80 -58

-57

-56

-55

-54

-53 -52 ln(pCO)

-51

-50

-49

-48

Figure 11: Formation energy of several species from surface-adsorbed cobalt at site A in equilibrium with a gas-phase CO environment at 500 K. Pressures are in atm.

Discussion The relatively high adsorption energy of atomic Co on the (101) anatase surface, compared to those of other cobalt moieties, confirms that gas-phase species have a strong impact on the ripening process. While there is a significant migration barrier between sites A and B for an isolated cobalt ion, this barrier is, with the exception of CoO, systematically lowered by the inclusion of gas-phase species from the reactor. These effects can be largely attributed to the decrease in Co coordination with the surface. At typical FT conditions, the support is more realistically described by a hydrated surface. Based on the stability of OH group formation on the surface it was determined that the water coverage is able to break down CoO and CoOH. While this does not provide information about the rate or likelihood of this event, it can be expected that these species will not be able to continue diffusing over large time scales, and thus are unlikely candidates to contribute significantly to sintering by surface transport. By contrast, the cobalt carbonyl maintained a low migration energy barrier between the two sites. Additionally, the carbonyl species had significantly lower adsorption energies on both the dry and wet surfaces com17 ACS Paragon Plus Environment


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 58 59 60

Page 18 of 26

pared to Co, providing a favourable approach to gas-phase transport. While no inhibiting effect on transport was found in the case of the lone cobalt ion and the cobalt carbonyls, there may be further effects of surface hydration on the adsorption and mobility of cobalt species beyond the scope of this study. Thermodynamic considerations indicate that the formation of cobalt carbonyl is favourable in a wide range of CO pressure conditions, in particular the weakly-adsorbed Co(CO)3 . We therefore propose that both surface and gas-phase transport of these larger cobalt carbonyl species are highly relevant to the sintering process in this system.

Conclusions We have presented a study of the energetics of Co-containing species that could be found on the surface of a typical catalyst support in a Fischer-Tropsch reactor. We have shown that both adsorption energies and surface mobility barriers of these species vary, indicating a link to the ripening process. In particular, carbon monoxide feed binding leading to the formation of mobile cobalt carbonyls has been identified as a candidate for accelerating the ripening process. Gas-phase transport is also viable given the low adsorption energy of the large carbonyls. This link to the ripening process motivates the study of the adsorption and mobility of small Co clusters in the presence of water and CO. It also motivates the study of the removal of atoms from Co nanoparticle surfaces, edges, and corners in the presence of reactor species. An ongoing study into larger cobalt clusters on anatase surfaces is being performed by our group.

Acknowledgements K.K.B. Duff acknowledges the financial support of Shell Global Solutions International BV. The authors thank Ravi Agrawal, Constant Guedon, and Heiko Oosterbeek for discussions 18 ACS Paragon Plus Environment

Page 19 of 26

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 58 59 60


and a critical reading of the manuscript. This work was performed using the Darwin Supercomputer of the University of Cambridge High Performance Computing Service (http://www.hpc.cam.ac.uk/), provided by Dell Inc. using Strategic Research Infrastructure Funding from the Higher Education Funding Council for England and funding from the Science and Technology Facilities Council.

Further Information Data used in generating results is available at https://doi.org/10.17863/CAM.10158. Supplementary information contents: • Bulk and surface structure information • Adsorption geometry information • Formation energies methodology • Bader charge analysis • Images of adsorbed geometries



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 58 59 60

References (1) Meyer, R.; Ge, Q.; Lockemeyer, J.; Yeates, R.; Lemanski, M.; Reinalda, D.; Neurock, M. An Ab Initio Analysis of Adsorption and Diffusion of Silver Atoms on Alumina Surfaces. Surf. Sci. 2007, 601, 134–145. (2) Hoflund, G. B.; Minahan, D. M. Study of Cs-Promoted , α-Alumina-Supported Silver , Ethylene-Epoxidation Catalysts. J. Catal. 1996, 162, 48–53. (3) Rytter, E.; Holmen, A. Deactivation and Regeneration of Commercial Type FischerTropsch Co-Catalysts - A Mini-Review. Catalysts 2015, 5, 478–499. (4) Galhenage, R. P.; Yan, H.; Tenney, S. A.; Park, N.; Henkelman, G.; Albrecht, P.; Mullins, D. R.; Chen, D. A. Understanding the Nucleation and Growth of Metals on TiO2: Co Compared to Au, Ni, and Pt. J. Phys. Chem. C 2013, 117, 7191–7201. (5) Wang, Y.; Su, Y.; Zhu, M.; Kang, L. Ni Cluster Nucleation and Growth on the Anatase TiO2 (101) Surface: a Density Functional Theory Study. RSC Adv. 2015, 5, 16582– 16591. (6) Zhao, W.; Lin, H.; Li, Y.; Zhang, Y.; Huang, X.; Chen, W. Growth Mechanism of Palladium Clusters on Rutile TiO2(110) Surface. J. Nat. Gas Chem. 2012, 21, 544– 555. (7) Ong, S. V.; Khanna, S. N. Theoretical Studies of the Stability and Oxidation of Pdn (n= 17) Clusters on Rutile TiO2 (110): Adsorption on the Stoichiometric Surface. J. Phys. Chem. C 2012, 116, 3105–3111. (8) Sanz, J. F.; Márquez, A. Adsorption of Pd Atoms and Dimers on the TiO2 (110) Surface: A First Principles Study. J. Phys. Chem. C 2007, 111, 3949–3955. (9) Zhang, J.; Zhang, M.; Han, Y.; Li, W.; Meng, X.; Zong, B. Nucleation and Growth of


Page 20 of 26

Page 21 of 26

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 58 59 60


Palladium Clusters on Anatase TiO2 (101) Surface: A First Principle Study. J. Phys. Chem. C 2008, 112, 19506–19515. (10) Pan, X.; Xu, Y.-J. Defect-Mediated Growth of Noble-Metal (Ag, Pt, and Pd) Nanoparticles on TiO2 with Oxygen Vacancies for Photocatalytic Redox Reactions under Visible Light. J. Phys. Chem. C 2013, 117, 17996–18005. (11) Xu, C.; Lai, X.; Zajac, G. W.; Goodman, D. W. Scanning Tunneling Microscopy Studies of the TiO2 (110) Surface: Structure and the Nucleation Growth of Pd. Phys. Rev. B 1997, 56, 13464–13482. (12) Negra, M. D.; Nicolaisen, M. N.; Li, Z.; Møller, P. J. Study of the Interactions Between the Overlayer and the Substrate in the Early Stages of Palladium Growth on TiO2(110). Surf. Sci. 2003, 540, 117–128. (13) Jiang, D.-e.; Overbury, S. H.; Dai, S. Structures and Energetics of Pt Clusters on TiO2: Interplay between Metal-Metal Bonds and Metal-Oxygen Bonds. J. Phys. Chem. C 2012, 116, 21880–21885. (14) Rieboldt, F.; Vilhelmsen, L. B.; Koust, S.; Lauritsen, J. V.; Helveg, S.; Lammich, L.; Besenbacher, F.; Hammer, B.; Wendt, S. Nucleation and Growth of Pt Nanoparticles on Reduced and Oxidized Rutile TiO2 (110). J. Chem. Phys. 2014, 141, 214702. (15) C ¸ akir, D.; G¨ ulseren, O. Adsorption of Pt and Bimetallic PtAu Clusters on the Partially Reduced Rutile (110) TiO2 Surface: A First-Principles Study. J. Phys. Chem. C 2012, 116, 5735–5746. (16) Han, Y.; Liu, C.-j.; Ge, Q. Interaction of Pt Clusters with the Anatase TiO2 (101) Surface: A First Principles Study. J. Phys. Chem. B 2006, 110, 7463–7472. (17) Gong, X.-Q.; Selloni, A.; Dulub, O.; Jacobson, P.; Diebold, U. Small Au and Pt Clusters



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 58 59 60

at the Anatase TiO2 (101) Surface: Behavior at Terraces, Steps, and Surface Oxygen Vacancies. J. Am. Chem. Soc. 2008, 130, 370–381. (18) Zhou, Y.; Muhich, C. L.; Neltner, B. T.; Weimer, A. W.; Musgrave, C. B. Growth of Pt Particles on the Anatase TiO2 (101) Surface. J. Phys. Chem. C 2012, 116, 12114–12123. (19) Han, Y.; Liu, C.-j.; Ge, Q. Effect of Surface Oxygen Vacancy on Pt Cluster Adsorption and Growth on the Defective Anatase TiO2(101) Surface. J. Phys. Chem. C 2007, 111, 16397–16404. (20) Vittadini, A.; Selloni, A. Small Gold Clusters on Stoichiometric and Defected TiO2 Anatase (101) and their Interaction with CO: A Density Functional Study. J. Chem. Phys. 2002, 117, 353–361. (21) Zhang, L.; Cosandey, F.; Persaud, R.; Madey, T. E. Initial Growth and Morphology of Thin Au Films on TiO2(110). Surf. Sci. 1999, 439, 73–85. (22) Chen, Y.; Tian, G.; Pan, K.; Tian, C.; Zhou, J.; Zhou, W.; Ren, Z.; Fu, H. In Situ Controlled Growth of Well-Dispersed Gold Nanoparticles in TiO2 Nanotube Arrays as Recyclable Substrates for Surface-Enhanced Raman Scattering. J. Chem. Soc., Dalton Trans. 2012, 41, 1020–1026. (23) Parizotto, N.; Rocha, K.; Damyanova, S.; Passos, F.; Zanchet, D.; Marques, C.; Bueno, J. Alumina-Supported Ni Catalysts Modified with Silver for the Steam Reforming of Methane: Effect of Ag on the Control of Coke Formation. Appl. Catal., A 2007, 330, 12–22. (24) Lodziana, Z.; Nørskov, J. K. Adsorption of Cu and Pd on α-Al2O3(0001) Surfaces with Different Stoichiometries. J. Chem. Phys. 2001, 115, 11261–11267.


Page 22 of 26

Page 23 of 26

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 58 59 60


(25) Corral Valero, M.; Digne, M.; Sautet, P.; Raybaud, P. DFT Study of the Interaction of a Single Palladium Atom with γ-Alumina Surfaces: the Role of Hydroxylation. Oil Gas Sci. Technol. 2006, 61, 535–545. (26) Corral Valero, M.; Raybaud, P.; Philippe, S. Influence of the Hydroxylation of γ-Al2O3 Surfaces on the Stability and Diffusion of Single Pd Atoms: A DFT Study. J. Phys. Chem. B 2006, 110, 1759–1767. (27) Meyer, R.; Lockemeyer, J.; Yeates, R.; Lemanski, M.; Reinalda, D.; Neurock, M. An Ab Initio Analysis of Adsorption and Diffusion of Silver Atoms on Partially Hydroxylated α-Al2O3(0 0 0 1) Surfaces. Chem. Phys. Lett. 2007, 449, 155–159. (28) Phatak, A.; Koryabkina, N.; Rai, S.; Ratts, J.; Ruettinger, W.; Farrauto, R.; Blau, G.; Delgass, W.; Ribeiro, F. Kinetics of the Water-Gas Shift Reaction on Pt Catalysts Supported on Alumina and Ceria. Catal. Today 2007, 123, 224–234. (29) Altman, E.; Gorte, R. The Desorption of CO from Small Pt Particles on Al2O3. Surf. Sci. 1986, 172, 71–80. (30) Rashkeev, S. N.; Sohlberg, K.; Glazoff, M. V.; Novak, J.; Pennycook, S. J.; Pantelides, S. T. Transition Metal Atoms on Different Alumina Phases: The Role of Subsurface Sites on Catalytic Activity. Phys. Rev. B 2003, 67, 115414. (31) Mak, C.; George, S. Surface Diffusion of Hydrogen on Ru(001): Transition State Theory Calculations. Chem. Phys. Lett. 1987, 135, 381–386. (32) Islam, M. M.; Calatayud, M.; Pacchioni, G. Hydrogen Adsorption and Diffusion on the Anatase TiO2 (101) Surface: A First-Principles Investigation. J. Phys. Chem. C 2011, 115, 6809–6814. (33) Vittadini, A.; Casarin, M.; Selloni, A. Chemistry of and on TiO2-Anatase Surfaces by DFT Calculations: A Partial Review. Theor. Chem. Acc. 2007, 117, 663–671. 23 ACS Paragon Plus Environment


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 58 59 60

(34) Tilocca, A.; Selloni, A. Reaction Pathway and Free Energy Barrier for Defect-Induced Water Dissociation on the (101) Aurface of TiO2-Anatase. J. Chem. Phys. 2003, 119, 7445–7450. (35) Mills, G.; Jónsson, H. Quantum and Thermal Effects in H2 Dissociative Adsorption: Evaluation of Free Energy Barriers in Multidimensional Quantum Systems. Phys. Rev. Lett. 1994, 72, 1124–1127. (36) Mills, G.; Jónsson, H.; Schenter, G. K. Reversible Work Transition State Theory: Application to Dissociative Adsorption of Hydrogen. Surf. Sci. 1995, 324, 305–337. (37) Finnis, M.; Lozovoi, A.; Alavi, A. The Oxidation of NiAl: What Can We Learn from Ab Initio Calculations? Annu. Rev. Mater. Res. 2005, 35, 167–207. (38) Hine, N. D. M.; Frensch, K.; Foulkes, W. M. C.; Finnis, M. W. Supercell Size Scaling of Density Functional Theory Formation Energies of Charged Defects. Phys. Rev. B 2009, 79, 024112. (39) Kresse, G.; Hafner, J. Ab Initio Molecular Dynamics for Liquid Metals. Phys. Rev. B 1993, 47, 558–561. (40) Kresse, G.; Hafner, J. Ab Initio Molecular-Dynamics Simulation of the Liquid-MetalAmorphous-Semiconductor Transition in Germanium. Phys. Rev. B 1994, 49, 14251– 14269. (41) Kresse, G.; Furthm¨ uller, J. Efficiency of Ab-Initio Total Energy Calculations for Metals and Semiconductors Using a Plane-Wave Basis Set. Comp. Mater. Sci. 1996, 6, 15–50. (42) Kresse, G.; Furthm¨ uller, J. Efficient Iterative Schemes for Ab Initio Total-Energy Calculations Using a Plane-Wave Basis Set. Phys. Rev. B 1996, 54, 11169–11186. (43) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865–3868. 24 ACS Paragon Plus Environment

Page 24 of 26

Page 25 of 26

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 58 59 60


(44) Perdew, J. P.; Burke, K.; Ernzerhof, M. Errata: Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1997, 78, 1396. (45) Sanville, E.; Kenny, S. D.; Smith, R.; Henkelman, G. Improved Grid-Based Algorithm for Bader Charge Allocation. J. Comput. Chem. 2007, 28, 899–908. (46) Tilocca, A.; Selloni, A. Vertical and Lateral Order in Adsorbed Water Layers on Anatase TiO2(101). Langmuir 2004, 20, 8379–8384. (47) Tilocca, A.; Selloni, A. Structure and Reactivity of Water Layers on Defect-Free and Defective Anatase TiO2(101) Surfaces. J. Phys. Chem. B 2004, 108, 47434751. (48) Vittadini, A.; Selloni, A.; Rotzinger, F. P.; Grätzel, M. Structure and Energetics of Water Adsorbed at TiO2 Anatase (101) and (001) Surfaces. Phys. Rev. Lett. 1998, 81, 2954–2957. (49) Henkelman, G.; Jónsson, H. Improved Tangent Estimate in the Nudged Elastic Band Method for Finding Minimum Energy Paths and Saddle Points. J. Chem. Phys. 2000, 113, 9978. (50) Sheppard, D.; Terrell, R.; Henkelman, G. Optimization Methods for Finding Minimum Energy Paths. J. Chem. Phys. 2008, 128 . (51) Aschauer, U. J.; Tilocca, A.; Selloni, A. Ab Initio Simulations of the Structure of Thin Water Layers on Defective Anatase TiO2 (101) Surfaces. Int. J. Quantum Chem 2015, 115, 1250–1257. (52) Burgess, D. R. Thermochemical Data, NIST Chemistry WebBook, NIST Standard Reference Database Number 69. (53) Domalski, E. S.; Hearing, E. D. Condensed Phase Heat Capacity Data, NIST Chemistry WebBook, NIST Standard Reference Database Number 69.



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 58 59 60

Table of Contents Graphic


Page 26 of 26

Impact of Carbonyl Formation on Cobalt Ripening over Titania Surface

Recommend Documents