Carbon–carbon coupling is an important step in many catalytic reactions, and performing sp3–sp3 carbon–carbon coupling heterogeneously is particularly challenging. It has been reported that PdAu single-atom alloy (SAA) model catalytic surfaces are able to selectively couple methyl groups, producing ethane from methyl iodide. Herein, we extend this study to NiAu SAAs and find that Ni atoms in Au are active for C–I cleavage and selective sp3–sp3 carbon–carbon coupling to produce ethane. Furthermore, we perform ab initio kinetic Monte Carlo simulations that include the effect of the iodine atom, which was previously considered a bystander species. We find that model NiAu surfaces exhibit a similar chemistry to PdAu, but the reason for the similarity is due to the role the iodine atoms play in terms of blocking the Ni atom active sites. Specifically, on NiAu SAAs, the iodine atoms outcompete the methyl groups for occupancy of the Ni sites leaving the Me groups on Au, while on PdAu SAAs, the binding strengths of methyl groups and iodine atoms at the Pd atom active site are more similar. These simulations shed light on the mechanism of this important sp3–sp3 carbon–carbon coupling chemistry on SAAs. Furthermore, we discuss the effect of the iodine atoms on the reaction energetics and make an analogy between the effect of iodine as an active site blocker on this model heterogeneous catalyst and homogeneous catalysts in which ligands must detach in order for the active site to be accessed by the reactants.

Carbon–carbon coupling is an important step in the upgrading of light alkanes that are becoming more abundant due to higher shale gas production rates.1 Performing selective sp3–sp3 carbon coupling is still very challenging despite its importance in many processes, for example, in the production of pharmaceuticals.2,3 Typically, sp3–sp3 (Würtz) carbon coupling reactions are carried out using homogeneous catalysts; however, a number of issues plague homogeneous catalysis, such as difficulties in product–catalyst separation and catalyst deactivation over time. Therefore, there is interest in employing heterogeneous catalytic methods instead. Typically, Pd complexes are used as homogeneous catalysts, and recently, a heterogeneous Pd-atom based catalyst with superior performance to homogeneous systems has been reported for Suzuki Coupling.4 Specific to this work, PdAu single-atom alloys (SAAs) investigated using a model single crystal approach have been shown to perform selective sp3–sp3 carbon–carbon coupling when methyl iodide is used as the reactant.5 While Pd is a commonly used metal for coupling reactions, Ni is an attractive alternative of interest, in part, due to its lower cost.

In this work, model NiAu alloy surfaces were studied with a range of surface science techniques to understand their structure and reactivity toward methyl iodine while also making comparisons to previously investigated PdAu alloys for the same reaction.5 NiAu and PdAu single-atom alloys were then modeled theoretically in order to understand the mechanism of sp3–sp3 carbon–carbon coupling on the two systems and the effect of the iodine in terms of its competition with methyl groups for the active sites. This synergistic approach, combining surface science and modeling techniques, has already been proven successful in elucidating structure–function relationships and delivering an atomic-scale understanding of coupling reaction mechanisms.6–11 Specifically, we combined Temperature Programmed Desorption (TPD), Reflection Absorption Infrared Spectroscopy (RAIRS), Scanning Tunneling Microscopy (STM), Density Functional Theory (DFT), and kinetic Monte Carlo (KMC) methods to first determine the structure of the alloy surface, then quantify the performance of the model catalyst toward the coupling chemistry, and finally model these results from first principles in order to get an atomic-scale understanding of the reaction mechanism. Our results highlight the important role of iodine, which had previously been considered a bystander species, in the barrier for sp3–sp3 carbon coupling on PdAu SAA surfaces.

All experiments were performed in one of the three ultra-high vacuum chambers with base pressures below 2 × 10−10 mbar. The Au(111) single crystals were cleaned by successive Ar+ sputtering cycles (1.5 keV, 10 µA) followed by annealing to 725 K. Ni was deposited onto the Au(111) crystal held at 380 K using a Omicron Nanotechnology Focus electron beam evaporator, and the Ni coverage was determined by X-ray photoelectron spectroscopy (XPS) or CO TPD titration, which has been previously reported.12 Liquid methyl iodide (Sigma-Aldrich 99.5%) was purified by freeze–pump–thaw cycles and was then dosed through a high-precision leak valve onto the crystal. TPD experiments were then performed in the UHV chamber using a Hiden quadrupole mass spectrometer and a heating rate of 2 K s−1. After every TPD experiment, the Ni sites are lost to the Au bulk during the ramp to high temperature; therefore, the surface was cleaned and re-alloyed between experiments. STM studies were performed using a low-temperature scanning tunneling microscope (Omicron Nanotechnology) at 5 K or 80 K, as indicated. Reflection absorption infrared spectroscopy (RAIRS) experiments were performed at the Center for Functional Nanomaterials (CFN) at Brookhaven National Laboratory in a UHV system with a connected preparation chamber and an X-ray photoelectron spectroscopy (XPS) chamber for the measurement of the Ni content and cleanliness of the sample. The XPS data were obtained at room temperature using a SPECS PHOIBOS NAP 150 hemispherical analyzer and a monochromatic Al Kα x-ray source (1486.6, ∼0.25 eV linewidth) focused on the sample to a spot size <0.3 mm (0.05 eV step, 0.1 s dwell time, 5 scans, and 50 eV pass energy). The infrared data were collected using a Bruker 80 V spectrometer with a polarizer and an MCT detector (2000 scans, 4 cm−1 resolution) in a UHV chamber with KBr windows and a beam to a surface angle of ∼8°. The IR spectra were taken while the chamber was filled with 1 × 10−6 mbar CO.

The periodic Density Functional Theory (DFT) calculations were performed using the Vienna Ab initio Simulation package (VASP) 5.4.4.13–15 We used the nonlocal optB86b-vdW functional to describe the exchange–correlation potential.16–19 The core electrons were treated using the Projected Augmented Wave (PAW) method.20,21 The electronic density associated with the valence electrons was expanded on a plane wave basis set with an energy cutoff of 400 eV.

The following two sets of slabs and Monkhorst–Pack22 k-point meshes were used:

  • A four-layer p(4 × 4) Au(111) slab and a 7 × 7 × 1 k-point mesh were used to describe the reactivity of gold sites.

  • A four-layer p(3 × 1) Au(211) slab, where one of the step-edge Au atoms was substituted with Ni or Pd, and a 11 × 9 × 1 k-point mesh were used to describe the reactivity of the dopant atom at the elbow of the herringbone reconstruction.

The slabs were built after optimizing the bulk structure of Au (lattice constant of 4.127 Å). Further optimizations of the surface structures were carried out after fixing the positions of the atoms of the first two layers at their bulk positions. Transition states were optimized using a combination of algorithms (CI-NEB,23,24 DIMER,24–26 and Quasi-Newton). Vibrational frequency calculations were performed for all species.

For a given elementary step, the activation energy Ea is defined as the difference between the energy of the transition state ETS and the energy of the reactant(s) ER, while the reaction energy ΔrE is the difference between the energy of the product(s) EP and that of the reactant(s) ER,

Ea=ETSER,
ΔrE=EPER.

The Temperature Programmed Desorption (TPD) experiments were simulated within the graph-theoretical kinetic Monte Carlo (KMC) approach as implemented in Zacros 2.0.27,28 The kinetic constants were estimated within the approximations of harmonic Transition State Theory using parameters obtained from the DFT calculations (energies, frequencies, and rotational constants). The reaction and energetic patterns are given in Fig. S1. The simulations were run on periodic lattices comprising sites of four different types: top Au sites, top Pd (or Ni) sites, threefold Au sites, and threefold PdAu (or NiAu) mixed sites.

The lattices were built ensuring a 3% surface loading in Pd (or Ni) sites, randomly seeded as single atoms. Simulations on PdAu without iodine were run on a 15 300-site lattice (10 200 threefold sites and 5100 top sites) as reported in our previous work on PdAu. All other simulations were performed on a 105 336-site lattice (35 112 top sites and 70 224 threefold sites) to ensure good sampling even under the effect of dopant site poisoning by iodine, which decreases the number of available dopant sites, the only active sites considered here for the coupling reaction. Without lateral interactions, each dopant atom could stabilize up to six iodine atoms on their mixed threefold sites; however, such configurations are not realistic in view of steric hindrance effects due to the large size of the iodine atom. Therefore, to prevent such configurations from appearing in the KMC simulation, arbitrarily large lateral interactions were also added to the energetic model. In particular, we used an energy contribution of 1 eV for the interaction between two iodine atoms adsorbed on two of the six mixed threefold sites of the same dopant. Each TPD trace was obtained by averaging the TPD signals of ten simulations in order to reduce the noise due to the stochastic nature of the simulations.

Figure 1(a) shows a representative STM image of the NiAu SAA surface formed by depositing 0.02 ML Ni on a Au(111) surface at 380 K and cooling to 80 K for imaging. As has been reported before, the Ni atoms are seen alloyed into the Au surface in the regions around the elbows of the Au(111) 22 × √3 or “herringbone” reconstruction.29 These regions contain an edge dislocation that provides an entry site for the diffusing Ni atom to incorporate into the Au surface. In order to demonstrate that these Ni atom sites are isolated from one another on the Au surface, we performed RAIRS of adsorbed CO as seen in Fig. 1(b). At low loadings [0.02 monolayer (ML) Ni], only one sharp feature was observed at 2036 cm−1, consistent with the presence of isolated Ni atom sites, as no lower frequency signal due to multi-fold binding to Ni is seen. Previous RAIRS studies of CO at atop sites on Ni(111) have reported this feature at 2058 cm−1. The difference between this and our NiAu SAA result is due to the intrinsic electronic difference between isolated Ni sites vs extended Ni(111).30–32 Further evidence for the lower frequency of the IR feature on the SAA comes from our DFT calculations that put this frequency at 1991 cm−1, which is reasonably accurate when taking into consideration inherent errors in DFT calculations.

FIG. 1.

Characterization of the NiAu SAA surface and adsorption of the reactant methyl iodine: (a) 80 K STM image of the as-prepared 0.02 ML NiAu SAA, (b) CO RAIR signal from the 0.02 ML NiAu SAA, (c) 5 K STM image of methyl iodide on a 0.02 ML NiAu SAA after a 100 K anneal to equilibrate the intact molecules, and (d) 5 K STM image of methyl iodide on the 0.02 ML NiAu SAA surface after an anneal to 200 K that desorbs most of the molecules and cleaves the C–I bond of those with access to a Ni atom site.

FIG. 1.

Characterization of the NiAu SAA surface and adsorption of the reactant methyl iodine: (a) 80 K STM image of the as-prepared 0.02 ML NiAu SAA, (b) CO RAIR signal from the 0.02 ML NiAu SAA, (c) 5 K STM image of methyl iodide on a 0.02 ML NiAu SAA after a 100 K anneal to equilibrate the intact molecules, and (d) 5 K STM image of methyl iodide on the 0.02 ML NiAu SAA surface after an anneal to 200 K that desorbs most of the molecules and cleaves the C–I bond of those with access to a Ni atom site.

Close modal

After characterizing the structure of the NiAu SAA model surface, we used TPD to study the reactivity of methyl groups on the surface. Due to the lack of CH4 reactivity in ultra-high vacuum experiments, we used methyl iodide to populate the NiAu SAA surface with methyl groups and study their coupling chemistry. STM measurements of methyl iodide molecules on the surface revealed that before the reaction, methyl iodide molecules are present both at the Au(111) herringbone reconstruction elbows where Ni is present and on the bare Au surface between these regions [Fig. 1(c)]. The linear ordering of methyl iodine molecules seen in this image is caused by the soliton walls of the herringbone reconstruction [seen in Fig. 1(a)], which serve to corral the molecules between them.33 However, after an anneal to 200 K, which removes unreacted methyl iodide from the surface, it can be seen that the only molecular species present on the surface are at the Ni sites [Fig. 1(d)].

Figure 2 shows a series of TPD spectra that illustrate the reactivity of methyl iodide (MeI) on bare Au(111), a 0.02 ML NiAu SAA surface, and higher coverage (0.5 ML) Ni on Au(111). On bare Au(111), the dissociation of the C–I bond is difficult as evidenced by the fact that most of the MeI desorbs intact from the surface around 170 K before the C–I bond can be cleaved. However, small amounts of defects like step edges inherent to the Au(111) surface are able to dissociate the C–I bond, and the resulting methyl groups couple and desorb as ethane, as seen by the small ethane desorption peak at 326 K in Fig. 2(a).34 Note that the reactivity of Au(111) is so low that the desorption traces in Fig. 2(a) have been multiplied by a factor of 10 for clarity. A small methane peak around 320 K was also observed due to methyl radical rejection and subsequent hydrogenation on the chamber walls, in agreement with previous studies.34 

FIG. 2.

Reactivity of methyl iodine on Au(111) and NiAu model surfaces. (a) TPD after deposition of 1 L (Langmuir) CD3I on pure Au(111), demonstrating the relative inertness of the Au(111) surface. Note that TPD traces on bare Au(111) are multiplied by a factor of 10 to make the desorption peaks visible and that the scale bar is the same for traces (b) and (c). (b) TPD after deposition of 1 L of CD3I on the 0.02 ML NiAu SAA, showing ethane production and a small amount of methane produced from background hydrogenation of methyl groups. (c) TPD after deposition of 1 L of CD3I, showing that on a surface with a higher Ni loading (0.5 ML), the extended Ni sites lead to methyl decomposition and that the coupling pathway that leads to the formation of ethane is not in effect at this higher Ni loading.

FIG. 2.

Reactivity of methyl iodine on Au(111) and NiAu model surfaces. (a) TPD after deposition of 1 L (Langmuir) CD3I on pure Au(111), demonstrating the relative inertness of the Au(111) surface. Note that TPD traces on bare Au(111) are multiplied by a factor of 10 to make the desorption peaks visible and that the scale bar is the same for traces (b) and (c). (b) TPD after deposition of 1 L of CD3I on the 0.02 ML NiAu SAA, showing ethane production and a small amount of methane produced from background hydrogenation of methyl groups. (c) TPD after deposition of 1 L of CD3I, showing that on a surface with a higher Ni loading (0.5 ML), the extended Ni sites lead to methyl decomposition and that the coupling pathway that leads to the formation of ethane is not in effect at this higher Ni loading.

Close modal

In contrast to Au(111), when 0.03 ML of Ni is added to the Au surface, a dramatic increase in reactivity is observed as shown in Fig. 2(b), and larger amounts of ethane are seen desorbing at 253 K. At this temperature, ethane desorption is limited by its formation on the surface, and therefore, its desorption temperature is representative of the activation barrier for ethane formation, i.e., carbon–carbon coupling. Importantly, the much larger amounts of ethane evolved as compared to Au(111), and the fact that the ethane is formed around 80 K lower in temperature on the SAA compared to Au(111) reveals two important details. The first is that NiAu(111) SAAs facilitate low-temperature C–I cleavage, which leads to a larger yield of ethane from the SAA vs pure Au(111) from which most of the MeI desorbs from the surface around 170 K before it can react. Second, C–C coupling that leads to the formation of ethane occurs at much lower temperature on the NiAu SAA [250 vs 320 K from Au(111)], meaning that Ni atom sites also catalyze the coupling step of the reaction. This carbon–carbon coupling on the NiAu SAA surface is selective, and there is no evidence of decomposition products from our TPD measurements. The only other species observed, other than ethane, was methane desorbing from both the Au(111) surface as mentioned and the NiAu SAA. The use of isotopic labeling experiments enabled us to further investigate the origin of this methane. Specifically, CD3I was used to understand the role of background H2 found in all ultra-high vacuum chambers. Figure 2 shows that fully deuterated ethane (C2D6) is the primary product desorbing from the NiAu SAA surface and that the methane produced is predominantly CD3H. The production of CD3H has been observed before and is observed because of background H2 present in the chamber, which is capable of hydrogenating adsorbed methyl groups.12,35–40

Furthermore, quantitative mass spectrometry was performed on the carbon species desorbing from the NiAu SAA surface, and it was determined that there was a 1:1 relationship between Ni sites and carbon species desorbing as methane and ethane. This provides further evidence that the Ni atom sites are responsible for both C–I bond cleavage, as evidenced by the larger amounts of ethane formed on NiAu SAAs vs pure Au(111), and the sp3–sp3 carbon–carbon coupling reaction step, as demonstrated by ethane formation at a temperature of ∼75 K lower than that of pure Au(111).

In order to examine the reactivity of Ni ensembles in Au, which are larger than one atom, we performed identical experiments on a 0.5 ML NiAu model surface as seen in Fig. 2(c). Interestingly, on this surface, which contains extended Ni ensembles, the methyl groups produced from the dissociation of CD3I were not observed to undergo carbon–carbon coupling to produce ethane. Instead, they decomposed and underwent reaction with background hydrogen to produce deuterated methane CD3H, as well as further decomposition, which led to the desorption of D2 above room temperature. Together, these TPD results demonstrate that individual, isolated Ni atoms in Au(111) serve as active sites for both C–I bond cleavage that initiates the reaction of methyl iodide and selective sp3–sp3 carbon–carbon coupling of surface-bound methyl groups to form ethane.

In order to further understand the experimental results and also compare the energetics of the reaction steps to previously reported PdAu SAAs, we performed DFT and KMC simulations. The main challenge when modeling the reactivity of gold-based single-atom alloys is to describe the active site properly. As previously mentioned, the transition metals used as surface dopants (Ni and Pd) tend to alloy into the Au surface at the elbows of the herringbone reconstruction where edge dislocations are present, which have Au sites with a coordination number lower than 9. We showed in our previous work on PdAu SAAs that, while the (111) surface provides a good model for Au sites, the (211) surface is necessary to describe the reactivity of Pd sites.5 Similarly, we consider here that the Ni sites in the NiAu SAA are located at the (211) facet and that the Au sites are well described by the (111) surface. As with the PdAu SAA, the overall activation energy for the C–C coupling of methyl groups depends on the initial coverage and the associated representative initial state [see Figs. 3 and 4(a)4(c)]. For methyl group to dopant atom (Ni and Pd) ratios (denoted as σ) greater than 1, all the dopant top sites are occupied by methyl groups (referred to as Me*), and excess methyl groups are adsorbed on gold sites (referred to as Me) as represented in Fig. 4(b). For such coverages, the excess methyl groups at Au sites couple easily with dopant-bound Me* with a rather low activation energy Ea* of 0.36 eV on NiAu and 0.45 eV on PdAu (see the barrier from Me + Me* in Fig. 3). Now, for the case of σ ≤ 1, all methyl species are adsorbed on dopant atom sites that are separated spatially from one another [as represented in Fig. 4(a)], and the representative initial state for the coupling is Me* + Me* (instead of the previously described Me + Me*). For this situation, the two methyl groups (Me*) cannot react readily since they are physically distanced. In order for the reaction to occur, one of the two methyl groups must spill over to gold (Me* → Me) and then couple with dopant-bound Me*. This spillover of methyl from the dopant to Au is endothermic (see Table I) and significantly raises the activation energy (denoted by Ea**) to 0.77 eV on NiAu and 0.67 eV on PdAu (compare red and green arrows in Fig. 3).

FIG. 3.

Energetics for the coupling of methyl groups on PdAu and NiAu SAA surfaces. In this energy diagram, the Pd and Ni single-atom sites are considered at the Au(211) facet and gold sites at the Au(111) facet. Me and Me* represent methyl adsorbed on a gold top site and a single-atom top site, respectively. Energies are referenced with respect to clean surfaces and methyl groups on gold top sites (Me).

FIG. 3.

Energetics for the coupling of methyl groups on PdAu and NiAu SAA surfaces. In this energy diagram, the Pd and Ni single-atom sites are considered at the Au(211) facet and gold sites at the Au(111) facet. Me and Me* represent methyl adsorbed on a gold top site and a single-atom top site, respectively. Energies are referenced with respect to clean surfaces and methyl groups on gold top sites (Me).

Close modal
FIG. 4.

KMC-simulated TPD traces using the parameters computed at the DFT level. (a)–(c) Different initial conditions were simulated. The traces correspond to the production of ethane on the PdAu SAA (d) and NiAu SAA (e). The dashed line indicates the position of the temperature peak when iodine is explicitly taken into account in the KMC simulations. The temperature ramps used in the simulations are as follows: 2 K/s for NiAu and 1 K/s for PdAu, consistent with experimental TPDs.

FIG. 4.

KMC-simulated TPD traces using the parameters computed at the DFT level. (a)–(c) Different initial conditions were simulated. The traces correspond to the production of ethane on the PdAu SAA (d) and NiAu SAA (e). The dashed line indicates the position of the temperature peak when iodine is explicitly taken into account in the KMC simulations. The temperature ramps used in the simulations are as follows: 2 K/s for NiAu and 1 K/s for PdAu, consistent with experimental TPDs.

Close modal
TABLE I.

Reaction energies for the spillover of methyl and iodine species from the dopant atom sites to the host Au(111) metal surface and the exchange between iodine and methyl species adsorbed at the dopant sites on PdAu and NiAu SAAs. Starred (*) adsorbates are bound to the dopant [which is considered at the step edge of the (211) facet], and unstarred adsorbates are located on the Au(111) surface. The overall exchange process is the difference between the two elementary steps. Energies are given in eV.

PdAu SAANiAu SAA
Spillover elementary steps 
Me* → Me 0.22 0.41 
I* → I 0.36 0.62 
Overall exchange process 
I* + Me → I + Me* 0.14 0.21 
PdAu SAANiAu SAA
Spillover elementary steps 
Me* → Me 0.22 0.41 
I* → I 0.36 0.62 
Overall exchange process 
I* + Me → I + Me* 0.14 0.21 

We then performed KMC simulations to directly compare the experimental TPDs with simulated TPDs based on the DFT calculations. We first discuss the case in which the effect of iodine is neglected. On both PdAu and NiAu SAAs, the KMC simulated TPDs show two ethane desorption peaks [dark blue traces in Figs. 4(d) and 4(e)]. The first one at low temperature (200 K for PdAu and 160 K for NiAu) is only seen for methyl group to dopant atom ratios (σ>1, and the second ethane desorption peak at higher temperature (255 K for PdAu and 310 K for NiAu) appears for the values of σ above and below 1. This is in line with the coverage dependent activation energies (Ea* and Ea**) expected from our DFT calculations. Experimentally, σ ∼ 1, and only one peak is seen around 250 K. This is why, in our previous work on the PdAu SAA, we had attributed this peak to the coupling of two methyl groups initially distanced on two different dopant sites [initial state Me* + Me*; see Fig. 4(a)]. In the case of the NiAu SAA, this simulated peak shifts to 310 K, in poor agreement with experimental data (see Fig. 2). This discrepancy between our experimental and theoretical results suggests that iodine, which is produced via the dissociated adsorption of methyl iodide and often thought to be a bystander, may actually play a role in the carbon–carbon coupling energetics.

Our DFT calculations indicate that the dopant sites stabilize iodine atoms to a larger extent than methyl groups as can be seen in Table I. Therefore, when the systems are in a representative low-temperature initial state, all dopant sites should be occupied by iodine (I*) and all methyl groups should reside on gold sites (Me), as illustrated in Fig. 4(c). As the dopant is the active site for the coupling, the first step of the coupling reaction is the endothermic exchange of iodine with methyl (I* + Me → I + Me* in Table I). From this state, transient dopant-bound Me* can couple with a methyl group bound to gold. This brings the overall activation energy for carbon–carbon coupling to 0.59 eV for PdAu and 0.57 eV for NiAu. When taking this exchange equilibrium into account in the KMC simulations, the TPD traces show a broad peak at 235 K for PdAu and 260 K for NiAu (pink traces in Fig. 4), in close agreement with the experimental data (see Fig. 2).

In the absence of iodine, our simulations have evidenced two regimes for the C–C coupling [see Figs. 4(d) and 4(e) and Fig. S2]. When there is an excess of methyl groups, the dopant sites act as very efficient active sites and dopant-bound methyl groups couple with excess methyl groups on gold sites at very low temperature. When the methyl-to-dopant ratio becomes stoichiometric or sub-stoichiometric, every methyl group is stabilized on physically distanced Ni dopant sites. Hence, they need to overcome the thermodynamic barrier to spill over to gold and react with a methyl group at another dopant site. In this regime, where there are more active sites than reactants, the reactant itself hinders the coupling and the latter can only happen at higher temperatures. This widens the temperature range over which methyl groups are present on the surface and may open the possibility for other undesired reactions to occur. When iodine is added in stoichiometric proportion with respect to the dopant, it pushes methyl species onto gold, where they are not able to react. However, methyl groups and iodine atoms can transiently exchange, giving rise to a configuration that is particularly reactive toward C–C coupling to ethane. Because iodine poisons the active site to some extent, the low-temperature regime coupling occurs at a moderate temperature when iodine is present. Most importantly, these simulations indicate that when iodine is present, methyl groups never get trapped on the dopant site where they can only couple at a higher temperature.

While this situation of competitive binding of iodine atoms to the active sites for methyl group coupling complicates the mechanistic picture of how the coupling reaction occurs, it allows us to make an interesting comparison to common, and very well understood, homogenously catalyzed coupling reactions. Specifically, homogeneous catalysts are typically metal centers surrounded by ligands (L) as seen in Fig. 5. In order for a coupling reaction to proceed, one or two ligands must detach from the metal center so that the reactants (blue circles) bind to the metal center. The reactants subsequently couple and desorb from the metal complex, and the ligands are replaced. This is essentially the same mechanism that we have discovered for heterogeneous carbon–carbon coupling on NiAu and PdAu SAAs as shown in Fig. 5. In an analogous manner, the iodine “ligand” bound to the metal atom dopant must be exchanged with a methyl group that can then couple with a second methyl group supplied by the Au surface, and after the coupling is complete and ethane desorbs, the metal center is again occupied by iodine.

FIG. 5.

Comparing the mechanism of carbon–carbon coupling on SAA surfaces (heterogeneous) with a typical homogenously catalyzed coupling mechanism. In the heterogeneous catalytic cycle, adsorbates are exchanged between the active site M (Pd or Ni) and the non-reactive Au sites. In the homogeneous catalytic cycles, the species are exchanged between the active site M and the solvent.

FIG. 5.

Comparing the mechanism of carbon–carbon coupling on SAA surfaces (heterogeneous) with a typical homogenously catalyzed coupling mechanism. In the heterogeneous catalytic cycle, adsorbates are exchanged between the active site M (Pd or Ni) and the non-reactive Au sites. In the homogeneous catalytic cycles, the species are exchanged between the active site M and the solvent.

Close modal

Our experiments reveal that NiAu SAA surfaces, like PdAu SAAs, are able to catalyze sp3–sp3 carbon–carbon coupling of surface-bound methyl groups to form ethane. Unlike clusters of Ni or Pd (observed in alloys at higher dopant coverages), which are incapable of ethane formation and instead catalyze decomposition of the methyl groups, single Ni or Pd atoms catalyze both C–I cleavage in the reactant molecule and carbon–carbon coupling. The latter occurs at lower temperatures in the NiAu(111) SAA than on pure Au(111), providing evidence that the reaction indeed occurs at the Ni atom sites. While the iodine atom produced by C–I cleavage in methyl iodide is often thought to be a bystander in such coupling reactions, we treat it explicitly in our DFT calculations and find that surface-bound iodine atoms outcompete surface-bound methyl groups for binding to the Ni or Pd atom sites in Au. This effect is particularly pronounced for the NiAu SAA, and we find that the energy to displace the iodine atom from the Ni active site must be considered in order to accurately model the experimental results and reproduce the correct desorption temperatures with KMC. This is analogous to the mechanism of coupling reactions on homogeneous catalysts in which ligands must detach from the active metal center in order for the reactants to be activated and the coupling product formed.

See the supplementary material for graph patterns representing the elementary steps/energetic interactions used in the graph-theoretical KMC simulations and methyl coverages as a function of temperature illustrating the effect of iodine on surface reactivity.

This work was supported as part of Integrated Mesoscale Architectures for Sustainable Catalysis, an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Award No. DE-SC0012573. R.R. was funded by the Leverhulme Trust (Grant No. RPG-2018-209). Part of the research used resources at the Center for Functional Nanomaterials, Brookhaven National Laboratory, supported by the U.S. Department of Energy, Office of Basic Energy Sciences, under Contract No. DE-SC0012704. We are also grateful to the UK Materials and Molecular Modeling Hub, which is partially funded by the EPSRC (Grant No. EP/P020194/1), for computational resources.

The data that support the findings of this study are available from the corresponding author upon reasonable request.

1.
A.
Galadima
and
O.
Muraza
, “
Revisiting the oxidative coupling of methane to ethylene in the golden period of shale gas: A review
,”
J. Ind. Eng. Chem.
37
,
1
13
(
2016
).
2.
L.
Yin
and
J.
Liebscher
, “
Carbon–carbon coupling reactions catalyzed by heterogeneous palladium catalysts
,”
Chem. Rev.
107
(
1
),
133
173
(
2007
).
3.
N. T. S.
Phan
,
M.
Van Der Sluys
, and
C. W.
Jones
, “
On the nature of the active species in palladium catalyzed Mizoroki–Heck and Suzuki–Miyaura couplings—Homogeneous or heterogeneous catalysis, a critical review
,”
Adv. Synth. Catal.
348
(
6
),
609
679
(
2006
).
4.
Z.
Chen
,
E.
Vorobyeva
,
S.
Mitchell
,
E.
Fako
,
M. A.
Ortuño
,
N.
López
,
S. M.
Collins
,
P. A.
Midgley
,
S.
Richard
,
G.
Vilé
, and
J.
Pérez-Ramírez
, “
A heterogeneous single-atom palladium catalyst surpassing homogeneous systems for Suzuki coupling
,”
Nat. Nanotechnol.
13
,
702
707
(
2018
).
5.
R.
Réocreux
,
M.
Uhlman
,
T.
Thuening
,
P.
Kress
,
R.
Hannagan
,
M.
Stamatakis
, and
E. C. H.
Sykes
, “
Efficient and selective carbon–carbon coupling on coke-resistant PdAu single-atom alloys
,”
Chem. Commun.
55
(
100
),
15085
15088
(
2019
).
6.
M.
Xi
and
B. E.
Bent
, “
Mechanisms of the Ullmann coupling reaction in adsorbed monolayers
,”
J. Am. Chem. Soc.
115
(
16
),
7426
7433
(
1993
).
7.
Q.
Fan
,
J. M.
Gottfried
, and
J.
Zhu
, “
Surface-catalyzed C–C covalent coupling strategies toward the synthesis of low-dimensional carbon-based nanostructures
,”
Acc. Chem. Res.
48
(
8
),
2484
2494
(
2015
).
8.
D. H.
Fairbrother
,
X. D.
Peng
,
R.
Viswanathan
,
P. C.
Stair
,
M.
Trenary
, and
J.
Fan
, “
Carbon–carbon coupling of methyl groups on Pt(111)
,”
Surf. Sci.
285
(
1–2
),
L455
L460
(
1993
).
9.
V. K.
Kanuru
,
G.
Kyriakou
,
S. K.
Beaumont
,
A. C.
Papageorgiou
,
D. J.
Watson
, and
R. M.
Lambert
, “
Sonogashira coupling on an extended gold surface in vacuo: Reaction of phenylacetylene with iodobenzene on Au(111)
,”
J. Am. Chem. Soc.
132
(
23
),
8081
8086
(
2010
).
10.
T.
Borkowski
,
A. M.
Trzeciak
,
W.
Bukowski
,
A.
Bukowska
,
W.
Tylus
, and
L.
Kępiński
, “
Palladium(0) nanoparticles formed in situ in the Suzuki–Miyaura reaction: The effect of a palladium(II) precursor
,”
Appl. Catal., A
378
(
1
),
83
89
(
2010
).
11.
A.
Paul
,
X.
Michael
, and
B. E.
Bent
, “
Disproportionation and coupling reactions of alkyl iodides on a Au(111) surface
,”
Surf. Sci.
297
,
327
344
(
1993
).
12.
F. R.
Lucci
,
M. T.
Darby
,
M. F. G.
Mattera
,
C. J.
Ivimey
,
A. J.
Therrien
,
A.
Michaelides
,
M.
Stamatakis
, and
E. C. H.
Sykes
, “
Controlling hydrogen activation, spillover, and desorption with Pd–Au single-atom alloys
,”
J. Phys. Chem. Lett.
7
(
3
),
480
485
(
2016
).
13.
G.
Kresse
and
J.
Hafner
, “
Ab initio molecular dynamics for liquid metals
,”
Phys. Rev. B
47
,
558
561
(
1993
).
14.
G.
Kresse
and
J.
Furthmüller
, “
Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set
,”
Phys. Rev. B
54
,
11169
11186
(
1996
).
15.
G.
Kresse
and
J.
Furthmüller
, “
Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set
,”
Comput. Mater. Sci.
6
(
1
),
15
50
(
1996
).
16.
M.
Dion
,
H.
Rydberg
,
E.
Schröder
,
D. C.
Langreth
, and
B. I.
Lundqvist
, “
Van der Waals density functional for general geometries
,”
Phys. Rev. Lett.
92
,
246401
(
2004
).
17.
J.
Klimeš
,
D. R.
Bowler
, and
A.
Michaelides
, “
Chemical accuracy for the van der Waals density functional
,”
J. Phys.: Condens. Matter
22
,
022201
(
2010
).
18.
J.
Klimeš
,
D. R.
Bowler
, and
A.
Michaelides
, “
Van der Waals density functionals applied to solids
,”
Phys. Rev. B
83
,
195131
(
2011
).
19.
J.
Klimeš
and
A.
Michaelides
, “
Perspective: Advances and challenges in treating van der Waals dispersion forces in density functional theory
,”
J. Chem. Phys.
137
,
120901
120913
(
2012
).
20.
P.
Blöchl
, “
Projector augmented-wave method
,”
Phys. Rev. B
50
,
17953
(
1994
).
21.
G.
Kresse
and
D.
Joubert
, “
From ultrasoft pseudopotentials to the projector augmented-wave method
,”
Phys. Rev. B
59
,
1758
1775
(
1999
).
22.
H. J.
Monkhorst
and
J. D.
Pack
, “
Special points for Brillouin–Zone integrations
,”
Phys. Rev. B
13
,
5188
5192
(
1976
).
23.
D.
Sheppard
,
R.
Terrell
, and
G.
Henkelman
, “
Optimization methods for finding minimum energy paths
,”
J. Chem. Phys.
128
,
134106
(
2008
).
24.
G.
Henkelman
,
B. P.
Uberuaga
, and
H.
Jónsson
, “
A climbing image nudged elastic band method for finding saddle points and minimum energy paths
,”
J. Chem. Phys.
113
,
9901
(
2000
).
25.
A.
Heyden
,
A. T.
Bell
, and
F. J.
Keil
, “
Efficient methods for finding transition states in chemical reactions: Comparison of improved dimer method and partitioned rational function optimization method
,”
J. Chem. Phys.
123
,
224101
(
2005
).
26.
G.
Henkelman
and
H.
Jónsson
, “
A dimer method for finding saddle points on high dimensional potential surfaces using only first derivatives
,”
J. Chem. Phys.
111
,
7010
(
1999
).
27.
M.
Stamatakis
and
D. G.
Vlachos
, “
A graph-theoretical kinetic Monte Carlo framework for on-lattice chemical kinetics
,”
J. Chem. Phys.
134
(
21
),
214115
(
2011
).
28.
J.
Nielsen
,
M.
d’Avezac
,
J.
Hetherington
, and
M.
Stamatakis
, “
Parallel kinetic Monte Carlo simulation framework incorporating accurate models of adsorbate lateral interactions
,”
J. Chem. Phys.
139
(
22
),
224706
(
2013
).
29.
Z.-T.
Wang
,
M. T.
Darby
,
A. J.
Therrien
,
M.
El-Soda
,
A.
Michaelides
,
M.
Stamatakis
, and
E. C. H.
Sykes
, “
Preparation, structure, and surface chemistry of Ni–Au single atom alloys
,”
J. Phys. Chem. C
120
(
25
),
13574
13580
(
2016
).
30.
C. M.
Kruppe
,
J. D.
Krooswyk
, and
M.
Trenary
, “
Polarization-dependent infrared spectroscopy of adsorbed carbon monoxide to probe the surface of a Pd/Cu(111) single-atom alloy
,”
J. Phys. Chem. C
121
(
17
),
9361
9369
(
2017
).
31.
D. A.
Patel
,
R. T.
Hannagan
,
P. L.
Kress
,
A. C.
Schilling
,
V.
Çınar
, and
E. C. H.
Sykes
, “
Atomic-scale surface structure and CO tolerance of NiCu single-atom alloys
,”
J. Phys. Chem. C
123
(
46
),
28142
28147
(
2019
).
32.
R. T.
Hannagan
,
D. A.
Patel
,
L. A.
Cramer
,
A. C.
Schilling
,
P. T. P.
Ryan
,
A. M.
Larson
,
V.
Çınar
,
Y.
Wang
,
T. A.
Balema
, and
E. C. H.
Sykes
, “
Combining STM, RAIRS and TPD to decipher the dispersion and interactions between active sites in RhCu single-atom alloys
,”
ChemCatChem
12
(
2
),
488
493
(
2020
).
33.
E. C. H.
Sykes
,
B. A.
Mantooth
,
P.
Han
,
Z. J.
Donhauser
, and
P. S.
Weiss
, “
Substrate-mediated intermolecular interactions: A quantitative single molecule analysis
,”
J. Am. Chem. Soc.
127
(
19
),
7255
7260
(
2005
).
34.
A. M.
Paul
and
B. E.
Bent
, “
Alkyl coupling on copper, silver, and gold: Correlation between the coupling rate and the metal-alkyl bond strength
,”
J. Catal.
147
,
264
271
(
1994
).
35.
M. T.
Darby
,
M.
Stamatakis
,
A.
Michaelides
, and
E. C. H.
Sykes
, “
Lonely atoms with special gifts: Breaking linear scaling relationships in heterogeneous catalysis with single-atom alloys
,”
J. Phys. Chem. Lett.
9
,
5636
5646
(
2018
).
36.
F. R.
Lucci
,
M. D.
Marcinkowski
,
T. J.
Lawton
, and
E. C. H.
Sykes
, “
H2 activation and spillover on catalytically relevant Pt–Cu single atom alloys
,”
J. Phys. Chem. C
119
(
43
),
24351
24357
(
2015
).
37.
P.
Aich
,
H.
Wei
,
B.
Basan
,
A. J.
Kropf
,
N. M.
Schweitzer
,
C. L.
Marshall
,
J. T.
Miller
, and
R.
Meyer
, “
Single-atom alloy Pd–Ag catalyst for selective hydrogenation of acrolein
,”
J. Phys. Chem. C
119
(
32
),
18140
18148
(
2015
).
38.
G. X.
Pei
,
X. Y.
Liu
,
A.
Wang
,
A. F.
Lee
,
M. A.
Isaacs
,
L.
Li
,
X.
Pan
,
X.
Yang
,
X.
Wang
,
Z.
Tai
,
K.
Wilson
, and
T.
Zhang
, “
Ag alloyed Pd single-atom catalysts for efficient selective hydrogenation of acetylene to ethylene in excess ethylene
,”
ACS Catal.
5
(
6
),
3717
3725
(
2015
).
39.
M. D.
Marcinkowski
,
A. D.
Jewell
,
M.
Stamatakis
,
M. B.
Boucher
,
E. A.
Lewis
,
C. J.
Murphy
,
G.
Kyriakou
, and
E. C. H.
Sykes
, “
Controlling a spillover pathway with the molecular cork effect
,”
Nat. Mater.
12
(
6
),
523
528
(
2013
).
40.
G.
Kyriakou
,
M. B.
Boucher
,
A. D.
Jewell
,
E. A.
Lewis
,
T. J.
Lawton
,
A. E.
Baber
,
H. L.
Tierney
,
M.
Flytzani-Stephanopoulos
, and
E. C. H.
Sykes
, “
Isolated metal atom geometries as a strategy for selective heterogeneous hydrogenations
,”
Science
335
(
6073
),
1209
1212
(
2012
).

Supplementary Material