Cluster Nucleation and Growth from a Highly Supersaturated Adatom

Subscriber access provided by Washington University | Libraries Article Cluster Nucleation and Growth from a Highly Supersaturated Adatom Phase: Silver on Magnetite Roland Bliem, Rukan Kosak, Lukas Perneczky, Zbynek Novotny, Oscar Gamba, David Fobes, Zhiqiang Mao, Michael Schmid, Prof. Dr. Peter Blaha, Ulrike Diebold, and Gareth S. Parkinson ACS Nano, Just Accepted Manuscript • Publication Date (Web): 19 Jun 2014 Downloaded from http://pubs.acs.org on June 20, 2014

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.

ACS Nano 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 20 ACS Nano

1 2 3 4 5 6 7 8 Cluster Nucleation and Growth from a Highly 9 10 11 12 Supersaturated Adatom Phase: Silver on 13 14 15 Magnetite 16 17 18 † ‡ † † 19 Roland Bliem, Rukan Kosak, Lukas Perneczky, Zbynek Novotny, 20 21 † ¶ ¶ † 22 Oscar Gamba, David Fobes, Zhiqiang Mao, Michael Schmid, Peter 23 24 Blaha,‡ Ulrike Diebold,† and Gareth S. Parkinson∗,† 25 26 27 Institute of Applied Physics, Vienna University of Technology, Vienna, Austria, Institute for 28 29 Materials Chemistry, Vienna University of Technology, Vienna, Austria, and Tulane University, 30 31 New Orleans, Louisiana 70118, USA 32 33 34 E-mail: [email protected] 35 36 37 38 39 Abstract 40 √ √ 41 × ◦ 42 The atomic-scale mechanisms underlying the growth of Ag on the ( 2 2)R45 -Fe3O4(001) 43 surface were studied using scanning tunneling microscopy and density functional theory based 44 45 calculations. For coverages up to 0.5 ML, Ag adatoms populate the surface exclusively; ag- 46 47 glomeration into nanoparticles only occurs with the lifting of the reconstruction at 720 K. 48 49 Above 0.5 ML, Ag clusters nucleate spontaneously, and grow at the expense of the surround- 50 51 ing material with mild annealing. This unusual behaviour results from a kinetic barrier asso- 52 √ √ 53 ciated with the ( 2 × 2)R45◦ reconstruction, which prevents adatoms from transitioning 54 55 ∗To whom correspondence should be addressed 56 †Institute of Applied Physics, Vienna University of Technology, Vienna, Austria 57 ‡Institute for Materials Chemistry, Vienna University of Technology, Vienna, Austria 58 ¶Tulane University, New Orleans, Louisiana 70118, USA 59 60 1 ACS Paragon Plus Environment ACS Nano Page 2 of 20

1 2 3 4 to the thermodynamically favourable 3D phase. The barrier is identified as the large separa- 5 tion between stable adsorption sites, which prevents homogeneous cluster nucleation, and the 6 7 instability of the Ag dimer against decay to two adatoms. Since the system is dominated by 8 √ √ 9 kinetics so long as the ( 2 × 2)R45◦ exists, the growth is not well described by the tradi- 10 11 tional growth modes. It can be understood, however, as the result of supersaturation within an 12 13 adsorption template system. 14 15 16 17 18 Keywords 19 20 21 metal adatoms, nanoparticle growth, adsorption template, metal oxide surfaces, Fe3O4, magnetite 22 23 Adsorption templates are increasingly utilized to investigate fundamental processes at surfaces 24 25 because they offer a stable, well-defined initial state from which the evolution of a system can 26 27 be studied, and the influence of external stimuli determined. To date, moire´ lattices,1–3 surface 28 29 reconstructions,4 and ordered point defects5–7 have been found to provide the necessary combina- 30 31 tion of nucleation sites and corrugated surface potential to stabilize monodisperse metal clusters 32 33 at elevated temperatures. Such systems are promising for studies of heterogeneous catalysis, for 34 35 example, since the ability to tune the size of the clusters is ideal for studies of the size-effect. 36 √ √ 37 The adsorption template provided by the ( 2 × 2)R45◦-Fe O (001) surface is unique in 38 3 4 39 that the reconstruction stabilizes single Au adatoms to unprecedented temperatures (as high as 40 41 673 K).8 As such, this model system is potentially ideal to probe the mechanisms underlying single 42 43 9–13 44 atom catalysis under reaction conditions. However, while Au adatoms are stable thermally, the 45 46 coverage of Au adatoms is limited to 0.17 ML by the onset of cluster nucleation. In this study, we 47 48 turn our attention to the adsorption of Ag on the Fe3O4(001) surface. This choice was motivated 49 50 by our desire to further establish the universality of the adatom template for the transition metals, 51 52 to discover whether adatom coverages in excess of 0.17 ML could be achieved, and by reports of 53 14 54 highly active sub-nano Ag clusters in the literature. 55 56 In addition to studies of catalytic activity, the adatom template system also offers an ideal basis 57 58 for fundamental studies of cluster nucleation and growth; a long standing topic of fundamental 59 60 2 ACS Paragon Plus Environment Page 3 of 20 ACS Nano

1 2 3 4 research that underpins vital technologies such as microelectronics and heterogeneous catalysis. 5 6 Volmer-Weber growth (V-W, 3D crystallites) dominates in such systems since the surface energies 7 15,16 8 of metal oxides are often lower than those of metals, but Stranski-Krastanov (S-K) growth, in 9 10 which 3D crystallites coexist with a stable wetting layer, occurs in systems with strong adatom- 11 12 substrate interactions. In this paper, we use scanning tunneling microscopy and density functional 13 14 theory based calculations to study the atomistic details of Ag growth at the Fe3O4(001) surface. 15 17 16 While the thermodynamic preference is for 3D crystallites, as shown previously, our results show 17 √ √ ◦ 18 that a ( 2 × 2)R45 reconstruction strongly inﬂuences the growth kinetics, and indeed a near 19 20 complete layer of adatoms to be stabilized at room temperature. This unusual behavior stems 21 22 from the large distance between stable adsorption sites and the relative instability of the Ag dimer, 23 24 which decays into two adatoms on formation. Since kinetics dominate over thermodynamics below 25 26 720 K, the system is not well described by the traditional growth modes. We propose an alternative 27 28 description in which the surface is treated as an adsorption template, and cluster nucleation results 29 30 from supersaturation effects. 31 32 33 34 35 Results 36 37 38 The STM images presented in Figure 1 provide an overview of the Fe3O4(001) surface following 39 40 the deposition of Ag at room temperature. The Fe3O4(001) substrate is characterised by rows of 41 42 protrusions running in the ⟨110⟩ directions, which correspond to the rows of surface Fe atoms in 43 44 Figure 2(a).18–21 The Fe row direction, indicated by the cyan lines in Figure 1(a), rotates by 90◦ 45 46 from one terrace to to the next as a consequence of the inverse spinel structure. At the lowest 47 48 Ag coverage (Figure 1(a,d); 0.33 ML) Ag adatoms (yellow arrow) are observed at equivalent sites 49 √ √ 50 ◦ within the ( 2 × 2)R45 unit cell. Note that 1 ML is deﬁned as one adatom per Fe3O4(001)- 51 √ √ 52 ( 2× 2)R45◦ unit cell (1.42×1018 atoms/m2. The adatoms are statistically distributed and show 53 54 no preference for step edges or other defect sites. No clusters are observed. DFT calculations 55 56 reveal that the preferred adsorption site is identical to that observed previously for Au8 and Pd22 57 58 59 60 3 ACS Paragon Plus Environment ACS Nano Page 4 of 20

1 2 3 ≈ 4 adatoms, i.e. twofold coordinated to surface oxygen atoms, protruding 0.2 nm above the surface 5 6 plane. The Ag adatoms are strongly bound (1.44 eV) and possess a charge state of +1. Figure 2(a) 7 8 shows a perspective view of the optimized structure. Calculations of a Ag dimer find two similarly 9 10 stable configurations, flat and upright (see Figure 2(b,c) and Table 1). In both configurations the 11 12 bond length to the surface oxygen atom is larger than for a Ag adatom. In the flat case, both Ag 13 14 atoms have a charge state of +1, as the single adatom. In the upright case however, we estimate 15 16 the atom with no direct contact with the surface is neutral, while the other has a charge state of 17 18 approximately ≈+0.5. Crucially, the dimer is less stable than two adsorbed adatoms, consistent 19 20 with the lack of features attributable to Ag dimers in the experimental images at any coverage. 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 Figure 1: STM images of three different coverages of Ag/Fe3O4(001) deposited at room tempera- 48 ture. (a,d) 0.33 ML of Ag (Usample= +1.3 V, It = 0.3 nA): Stable adatoms (one highlighted by the 49 ⟨ ⟩ 50 arrow). The Fe rows of the Fe3O4(001) substrate run along the 110 directions (cyan lines).The 51 red box highlights a surface hydroxyl group, a common defect that occurs via dissociative adsorp- 23 52 tion of water in oxygen vacancies. (b,e) 0.58 ML of Ag (Usample= +1.2 V, It = 0.3 nA): High 53 adatom coverage and a few small clusters. (c,f) 1.45 ML Ag (Usample= +1.2 V, It = 0.3 nA): Full 54 55 adatom layer, clusters at step edges and on terraces. 56 57 58 When the Ag adatom coverage is increased to 0.58 ML, clusters are observed with a density 59 60 4 ACS Paragon Plus Environment Page 5 of 20 ACS Nano

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 Figure 2: Structural models of the adsorption geometry of Ag/Fe3O4(001) as determined by 26 27 DFT+U calculations: (a) Perspective view of a Ag adatom bound to surface oxygen in an ad- 28 sorption site between the Fe rows. (b) Perspective view of an upright Ag dimer bonding to surface 29 oxygen with one Ag atom. (c) Top view of a ﬂat Ag dimer aligned with the Fe rows, both Ag 30 atoms bind to the surface. 31 32 33 of 0.75×1016 m−2 (Figure 1(b,e)). The smallest clusters have an apparent height of ≈0.2 nm with 34 35 respect to the Fe rows irrespective of bias voltage, which suggests that they may be 2D. While 36 37 the excellent lattice match between Ag(001) and Fe O (001), differing by just 2.7%,17 would be 38 3 4 39 conducive to epitaxial growth, it is important to note that STM apparent heights contain contribu- 40 41 42 tions from electronic structure in addition to topography, and thus may not reﬂect the true height 43 44 of the cluster. At 1.45 ML Ag (Figure 1(c,f)), the adatom density approaches one adatom per √ √ 45 × ◦ × 16 −2 46 ( 2 2)R45 unit cell, and the cluster density increases to 2.7 10 m . The largest clusters 47 ≈ 48 observed at this coverage exhibit an apparent height of 0.8 nm, and are certainly 3D. 49 50 A STM movie comprising 67 sequential frames acquired following deposition of 0.29 ML Ag 51 52 is available as part of the Supplementary Information. On average, three adatoms change site per 53 54 frame, with the vast majority of hops occurring between neighbouring stable sites along the Fe 55 56 row direction. An analysis of the cross row adatom hopping reveals a strong preference for the 57 58 fast scanning direction of the STM ([100] compared to the crystallographically equivalent [010] 59 60 5 ACS Paragon Plus Environment ACS Nano Page 6 of 20

1 2 3 Table 1: Calculated DFT+U binding energies (in eV) per atom (compared to a free Ag atom) 4 5 and bond distances of the Ag atoms to the surface oxygen (Ag-O) and to the second Ag atom of 6 the dimer (Ag-Ag) for the stable configurations of Ag monomers and dimers at the Fe3O4(001) 7 surface. 8 9 Configuration EB/atom [eV] Ag-O bond length [A]˚ Ag-Ag bond length [A]˚ 10 free Ag dimer 0.91 - 2.55 11 12 Fe3O4/Ag single atom 1.44 2.07 - 13 Fe3O4/Ag dimer flat 1.21 2.40 2.72 14 Fe3O4/Ag dimer upright 1.25 2.30 2.60 15 16 17 18 19 20 21 22 23 24 25 26 27 Figure 3: STM image sequence (Usample= +1.2 V, It = 0.4 nA) illustrating the room temperature 28 mobility of the Ag adatoms. The cyan lines indicate the Fe row direction, the yellow arrows the 29 direction of motion of the moving adatom. In this sequence, only movement along the Fe rows is 30 31 observed. 32 33 34 direction), suggesting that this is likely tip induced. Surface hydroxyl groups do not appear to affect 35 22 36 the mobility of the Ag adatoms, in contrast to Pd adatoms, which are stabilized on Fe3O4(001) 37 24 −8 38 and TiO2(110). During the acquisition of the movie, the system was exposed to 10 mbar CO 39 −8 40 (frame 11 to 48) and 10 mbar CO2 (frame 49 to 67), which had no discernible effect on the 41 42 hopping rate. This is in stark contrast to Pd adatoms, which forms highly mobile Pd carbonyl 43 44 species on CO exposure.22 Figure 3 shows four sequential STM images taken from the movie, in 45 46 which the adatom mobility is indicated. Hopping atoms are indicated by yellow arrows, and the 47 48 local Fe row direction is again indicated by the cyan lines. 49 50 Figure 4 shows STM images acquired following deposition of 0.51 ML Ag, the lowest cover- 51 52 age for which Ag clusters were observed (cluster density ≈0.15×1016 m−2). At this coverage the 53 54 adatom distribution is extremely inhomogeneous; in some regions the local coverage is close to 55 56 1 ML coverage, e.g. the area encompassed by the yellow line, while other areas are almost com- 57 58 59 60 6 ACS Paragon Plus Environment Page 7 of 20 ACS Nano

1 2 3 4 pletely empty. The number of consecutive sites occupied along a particular row can be as high as 5 6 14, for example, in the area enclosed by the red line. The observed clusters have apparent heights 7 ≈ 8 that vary from 0.2-0.5 nm with respect to the Fe rows. 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Figure 4: STM images (Usample= +1.2 V, It = 0.3 nA) of 0.51 ML Ag/Fe3O4(001) deposited at 31 room temperature. (a) High resolution image showing inhomogeneously distributed adatoms. The 32 region encompassed by the yellow line has a local Ag coverage close to 1 ML. The purple line 33 encloses a single row in which 14 consecutive adsorption sites are occupied by Ag adatoms. (b) 34 ≈ × 16 −2 35 Large area scan showing the presence of a low density of clusters ( 0.15 10 m , examples 36 highlighted by yellow arrows). 37 38 39 In Figure 5 we plot the results of a thermal stability study conducted on samples with initial 40 41 Ag coverages of 0.2 and 1.45 ML. In the plot, the average density of adatoms counted in STM 42 43 images is shown as a function of annealing temperature. For an initial coverage of 0.2 ML, the 44 45 adatom density remains constant up to 670 K, but drops quickly to zero by 820 K. This is a clear 46 47 indication that, at least up to 670 K, cluster nucleation is not thermally induced. As the initial 48 49 coverage increases, the decrease in adatom coverage shifts to 520 K and 370 K for 0.58 ML and 50 51 1.45 ML, respectively. Thus, a somewhat paradoxical situation arises whereby an initially lower 52 53 adatom coverage leads to a higher adatom coverage between 520 and 800 K. This suggests that 54 55 56 clusters affect the stability of the adatom layer. 57 58 In Figure 6 we show two representative images acquired following the annealing experiments 59 60 7 ACS Paragon Plus Environment ACS Nano Page 8 of 20

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 Figure 5: Thermal stability of Ag adatoms for different initial coverages (and cluster densities). 28 The error bars are given by the standard deviation. The solid lines are intended as guide to the eye 29 only. In the absence of clusters (0.2 ML), Ag adatoms remain stable up to 670 K, at which point 30 clusters nucleate and the adatom coverage drops quickly to zero. At the high initial coverage, the 31 32 cluster nuclei are already present on the surface at room temperature. During annealing they grow 33 through the capture of the otherwise stable Ag adatoms beginning at 370 K. 34 35 36 described above. Figure 6(a) shows the 1.45 ML Ag/Fe3O4(001) surface annealed at 370 K. At this 37 38 temperature, clusters coexist with the adatom array, as was the case at room temperature, but the 39 40 clusters are larger and patches devoid of Ag adatoms exist in their vicinity (examples are shaded 41 42 in Figure 6). In what follows, such areas will be termed “vacancy trails”. The vacancy trails are 43 44 aligned with the local Fe row direction, giving further evidence that the preferential diffusion along 45 46 the row direction is not merely an effect caused by STM. In most cases the vacancy trails extend 47 48 from a cluster located at a step edge, along the local Fe row direction to the middle of a terrace. 49 50 Following annealing to 545 K, (Figure 6(b)) all clusters reside at either step edges or antiphase 51 √ √ 52 domain boundaries25 in the ( 2 × 2)R45◦ reconstruction. The shape of the clusters in Figure 53 54 6(b) appears elongated to the right side because the feedback loop of the STM did not adjust the 55 56 tip-sample distance quickly enough after surmounting the large Ag clusters. Clusters are again 57 58 59 60 8 ACS Paragon Plus Environment Page 9 of 20 ACS Nano

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Figure 6: Observation of the different cluster growth mechanisms. (a) STM image (Usample = 2 24 +1.2 V, It = 0.3 nA, 34 x 34 nm , 1.45 ML Ag, ﬂash-annealed at 370 K) illustrating the coexistence 25 of Ag clusters and adatoms. Vacancy trails (some are shaded in red) lead to the clusters. These 26 27 arise when the clusters capture Ag adatoms while diffusing on the surface. (b) STM image (Usample 2 28 = +1.45 V, It = 0.3 nA, 67 x 67 nm ) of 0.58 ML Ag, ﬂash-annealed at 545 K. Ag adatoms coexist 29 with large Ag clusters immobilized at step edges. Vacancy trails lead to the clusters in equal 30 distances on different terraces. This is consistent with adatom diffusion taking over as the primary 31 ⟨ ⟩ 32 cause of cluster growth. The cyan lines indicate the local Fe row directions along 110 . 33 34 35 associated with vacancy trails, which extend equal distances into the neighbouring terraces. This 36 37 is highlighted for the cluster in the center of Figure 6(b), which sits at the junction between three 38 39 terraces. From the areas of these large vacancy trails we can roughly estimate the number of atoms 40 41 contained in the clusters. For example, assuming an initial adatom coverage of 0.58 ML Ag, the 42 43 shaded region in Figure 6(b) corresponds to ≈225 Ag adatoms. 44 √ √ ◦ 45 With the lifting of the ( 2 × 2)R45 reconstruction at 720 K,26 a dramatic change occurs. 46 47 All small clusters disappear, and are replaced by sparsely distributed Ag clusters with an apparent 48 49 height ≈10 nm (see Supplementary Information). At 720 K, some Ag adatoms remain on the 50 51 surface, even in close proximity to the large clusters, but after annealing to 820 K all Ag adatoms 52 53 are incorporated into clusters. 54 55 56 57 58 59 60 9 ACS Paragon Plus Environment ACS Nano Page 10 of 20

1 2 3 4 Discussion 5 6 7 Combining the experimental data presented above with the results of the theoretical calculations, 8 9 we are able to develop a full picture of growth in the Ag/Fe3O4(001) system. At low coverage, 10 11 evaporated Ag atoms arrive with a random distribution and diffuse to occupy two-fold coordinated √ √ 12 ◦ 13 sites within the ( 2 × 2)R45 reconstruction. The adatoms are stable, well-separated, and suffi- 14 15 ciently immobile that homogeneous nucleation can not spontaneously occur at very low coverage 16 17 (as would occur in a typical Volmer-Weber growth). Monte Carlo simulations (see Supplementary 18 19 Information) show that already by 0.03 ML, incident Ag atoms begin to arrive at locations where 20 21 both neighbouring sites along the Fe row direction are already occupied by a Ag adatom. Since 22 23 no adatoms are observed in such locations at any coverage, this unfavourable Ag is efficiently 24 25 redistributed within the adatom array. Our DFT+U calculations suggest that this redistribution is 26 27 mediated by a Ag dimer, which forms readily (1.25 eV per Ag atom), but is ultimately unstable 28 29 with respect to two stable adatoms (1.44 eV/Ag atom). When the dimer decays, one of the atoms 30 31 is deposited in the next site along the row. Such a mechanism can also explain how such high 32 33 coverages can be achieved. Ag atoms arriving in rows with many occupied sites will be passed 34 35 backward and forward along the row, forming and breaking dimers, until the decay occurs adja- 36 37 cent to a vacancy. Crucially, the Ag dimer, like the Ag adatoms, remains bound at the preferred 38 39 site and cannot diffuse and meet further Ag adatoms in order to grow. This restricts the number of 40 41 simultaneously interacting Ag atoms to two, insufficient to form a stable cluster nucleus. 42 43 44 While dimer decay efficiently expands the adatom phase at low coverage, it is also responsible 45 46 for the nucleation of stable clusters above 0.5 ML. At this coverage, the Ag distribution is highly 47 48 inhomogeneous, and many consecutively occupied sites can be found in some rows (see Figure 49 50 4). If two excess Ag atoms exist in such a row, they will diffuse back and forth until they reach 51 52 an empty site, or until they occupy neighbouring sites. When dimer decay occurs adjacent to an 53 54 existing dimer, the number of Ag atoms that can simultaneously interact suddenly increases from 55 56 two to three, and a cluster nucleates. Note that this does not mean that the smallest stable clusters 57 58 contain three atoms, just that homogeneous nucleation process is initiated by this interaction. It is 59 60 10 ACS Paragon Plus Environment Page 11 of 20 ACS Nano

1 2 3 4 possible, indeed likely, that further Ag adatoms from the vicinity are incorporated to form a stable 5 6 cluster. 7 8 So how can we classify the growth of Ag at the Fe3O4(001) surface? At first glance, the 9 10 coexistence of a complete adatom monolayer and 3D clusters suggests a S-K growth mode. How- 11 12 ever, the binding energy of an adatom to the surface (1.41 eV) is significantly less than inside a 13 14 Ag cluster (≈2.9 eV), and the surface free energy of Ag(001) is significantly higher than that of 15 17 16 Fe3O4(001). Together, these two numbers show that V-W growth should be clearly favoured. The 17 18 annealing experiments shown in Figure 6 are consistent with a thermodynamic preference for 3D 19 20 crystallites, since the clusters grow at the expense of the adatom phase. The problem with this 21 22 assignment, however, is that the stable adatom coverage is far in excess of what could be expected 23 24 for a typical V-W system. Only after annealing to 800 K, when the reconstruction has been lifted, 25 √ √ 26 the system grows in V-W mode. We therefore conclude that the ( 2 × 2)R45◦ reconstruction 27 28 introduces a kinetic barrier that prevents adatoms transitioning to the thermodynamically favoured 29 30 3D phase. 31 32 Kinetically hindered 3D growth has been observed previously in several systems.27–34 For ex- 33 34 ample, a metastable Ag adatom layer (nearest-neighbour distance 0.38 nm) forms at room tem- 35 36 perature on the Si(001)-(2 × 1) surface, despite the thermodynamic preference for 3D crystallites. 37 38 Above 1 ML coverage, Ag clusters nucleate from, and grow at the expense of the adatom phase, as 39 40 observed here. It was reasoned that Ag adatoms bind more strongly to the Si substrate than they do 41 42 as monomers atop a 2D Ag island, and that additional Ag diffuses to the edge of the island where it 43 44 45 joins the adatom phase. Cluster formation was explained as the result of multiple Ag atoms meet- 46 34 47 ing in the 2nd layer, with metal-metal bonding parallel to the surface stabilizing the 3D phase. 48 49 This description cannot be applied to the Ag/Fe3O4(001) system because the distance between sta- 50 51 ble sites is too large (0.84 nm) for the adatom phase to be considered a “first layer”. Additional 52 53 Ag atoms deposited into the adatom phase do not encounter a Ag island, rather an array of isolated 54 55 adatoms. Consequently, Ag redistribution occurs within the adatom phase via formation and decay 56 57 of Ag dimers, and not by diffusion over the adatom phase. 58 59 60 11 ACS Paragon Plus Environment ACS Nano Page 12 of 20

1 2 3 4 In what follows, we develop an alternative description of the system in which cluster nucle- 5 6 ation results from local supersaturation within the adsorption template system. We first define the 7 8 saturation adatom coverage as 1 ML, i.e., the maximum amount of Ag that can be accommodated 9 10 in stable sites on the adatom template. Note that this is not related to saturation adatom coverage 11 12 in thermodynamic equilibrium, which is extremely low, but is a property of the system comparable 13 14 to the solubility limit. When additional Ag is deposited onto the saturated template, the system 15 16 contains more Ag than the solubility limit, and enters a supersaturated state. It is important to note 17 18 that supersaturation occurs locally well below the global solubility limit of 1 ML at room temper- 19 20 ature because the adatom coverage is inhomogeneous and diffusion is slow. In the supersaturated 21 22 adatom array, Ag dimers form and break, but larger nuclei cannot form. It is only when the degree 23 24 of supersaturation (the Ag coverage) is increased that the critical size is exceeded and a stable 25 26 nucleus forms. 27 28 In addition to studies of diffusion and growth, we anticipate the Fe3O4(001) adsorption tem- 29 30 plate should find significant application in studies of heterogeneous catalysis. Recently, single 31 32 atoms have been proposed to catalyse important reactions such as water-gas-shift and CO oxida- 33 34 tion,9–13 but as yet there is little direct evidence as to how single atoms perform this function. The 35 36 Fe3O4(001) adsorption template is the only model system where single atoms can be stabilized 37 38 at reaction temperatures, making it ideal to uncover the atomic-scale details. In particular, it will 39 40 be fascinating to study how the electronic structure of the adatoms, modified through their inter- 41 42 action with the metal-oxide support, affects adsorption properties. This has been proposed as a 43 44 14,35,36 45 critical factor enhancing the reactivity of sub-nano species over larger nanoparticles. Finally, 46 47 since the adatom templating property is an inherent property of the substrate, a well-defined model 48 49 system can be created to study the reactivity of any metal adatom in any reaction. 50 51 52 53 54 55 56 57 58 59 60 12 ACS Paragon Plus Environment Page 13 of 20 ACS Nano

1 2 3 4 Conclusions 5 6 7 In summary, the experiments described here highlight the power of adsorption templates for the 8 9 study of fundamental processes at the atomic scale. The atomic-scale processes responsible for 10 11 mass transport, cluster nucleation and growth are determined for a kinetically hindered V-W sys- 12 13 tem, and the relative instability of the Ag dimer is revealed as the barrier preventing Ag adatoms 14 15 transitioning to the thermodynamically preferable 3D phase. Similar behaviour can be expected in 16 17 other systems when the following conditions are met: (i) only one adatom is stable per adsorption 18 19 site (ii) the adatoms are almost immobile at the growth temperature (iii) the dimer is unstable com- 20 21 pared to two isolated adatoms, and (iv) the separation of the nucleation sites is sufficiently large 22 23 to prevent homogeneous cluster formation. In the absence of atomically resolved STM, this could 24 25 be detected via titration of the nucleation sites or by photoemission spectroscopy; a linear increase 26 27 of the adatom coverage (or decrease of nucleation sites) would be followed by sudden recovery of 28 29 much of the substrate area. 30 31 32 33 34 Methods 35 36 37 The experiments were performed on a synthetic Fe3O4 single crystal, prepared in an ultra-high 38 39 vacuum (UHV) chamber with a base pressure below 10−10 mbar by cycles of sputtering with Ar+ 40 −6 41 ions (pAr=8×10 mbar, 1 keV, 1.2 µA, 15 min) followed by annealing with electron beam heat- 42 43 −7 ing in O2 at 870 K (pO2 =10 mbar, 15 min). Higher oxygen pressures were avoided in order to 44 37 45 prevent the creation of bulk Fe vacancies or α-Fe2O3 inclusions over the long term. For the 46 47 thermal stability study, the sample was flash-annealed in UHV for 1 min at the given tempera- 48 49 ture, as measured by a K-type thermocouple attached to the sample holder. The uncertainty of the 50 51 specified temperatures is estimated at 50 K. Silver was deposited onto the as-prepared surface at 52 53 room temperature from a Mo crucible in an LN2-cooled electron beam evaporator. Ag coverages 54 √ √ 55 are specified in monolayers (ML), defined as one adatom per Fe O (001)-( 2 × 2)R45◦ unit 56 3 4 57 × 18 2 ≈ 58 cell (1.42 10 atoms/m ). Deposition rates ( 2 ML/min) were calibrated using a water-cooled 59 60 13 ACS Paragon Plus Environment ACS Nano Page 14 of 20

1 2 3 4 quartz crystal microbalance (QCM). STM measurements were performed in the adjacent analysis 5 × −11 6 chamber (base pressure 5 10 mbar) using a customized Omicron µ-STM operated in constant 7 8 current mode at room temperature with an electrochemically etched W tip. The Ag adatom cov- 9 10 erage in each case is determined directly from the STM images, except for 1.45 ML, where the 11 12 nominal QCM coverage is given. 13 14 The DFT-based theoretical calculations utilize the augmented plane wave + local orbitals (APW+lo) 15 38 16 method as embodied in the WIEN2k code. We employ the generalized gradient approximation 17 18 (GGA) of Perdew et al.39 and treat the correlated Fe-3d electrons with a Hubbard-U correction 19 20 (GGA+U) and the double counting correction in the fully localized limit. Specifically we em- 21 22 ploy an effective U value of 3.8 eV for the Fe-3d states. In the APW+lo method, space is divided 23 24 into spheres around the atoms and an interstitial region. We use atomic sphere radii of 1.86, 1.6, 25 26 and 2.1 bohr for Fe, O, and Ag respectively. The plane-wave cutoff is set to 12 Ry and the 2D 27 28 Brillouin zone is sampled with a 2×2×1 k-mesh. We use a Fermi broadening of 0.08 eV. The 29 30 Feoct-terminated Fe3O4(001) surface is modelled using a symmetric slab with 13 layers and a 2×2 31 32 supercell (192 Fe and O atoms). The free surface has been fully relaxed, while for the adatom 33 34 systems the four innermost layers have been frozen. 35 36 37 38 Acknowledgement 39 40 41 This material is based upon work supported as part of the Center for Atomic-Level Catalyst De- 42 43 sign, an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of 44 45 Science, Office of Basic Energy Sciences under Award Number DE-SC0001058. GSP acknowl- 46 47 edges support from the Austrian Science Fund project number P24925-N20. RB acknowledges a 48 49 stipend from the TU Vienna and Austrian Science Fund doctoral college SOLIDS4FUN, project 50 51 number W1243. The work at Tulane is supported by the NSF under grant DMR-1205469. 52 53 54 55 56 57 58 59 60 14 ACS Paragon Plus Environment Page 15 of 20 ACS Nano

1 2 3 4 Supporting Information Available 5 6 The supporting information contains an STM movie showing the mobility of Ag adatoms, a large- 7 8 scale STM image showing a large Ag cluster created by annealing to 720 K, additional thermal 9 10 stability data, and details about Monte-Carlo simulations. This material is available free of charge 11 12 via the Internet at http://pubs.acs.org/. 13 14 15 16 References 17 18 19 1. Berdunov, N.; Mariotto, G.; Balakrishnan, K.; Murphy, S.; Shvets, I. V. Oxide Templates for 20 21 Self-Assembling Arrays of Metal Nanoclusters. Surf. Sci. 2006, 600, L287–L290. 22 23 24 2. Corso, M.; Auwarter,¨ W.; Muntwiler, M.; Tamai, A.; Greber, T.; Osterwalder, J. Boron Nitride 25 26 Nanomesh. Science 2004, 303, 217–220. 27 28 29 3. N’Diaye, A. T.; Bleikamp, S.; Feibelman, P. J.; Michely, T. Two-Dimensional Ir Cluster Lattice 30 31 on a Graphene Moire´ on Ir(111). Phys. Rev. Lett. 2006, 97, 215501. 32 33 34 4. Chambliss, D. D.; Wilson, R. J.; Chiang, S. Nucleation of Ordered Ni Island Arrays on 35 36 Au(111) by Surface-Lattice Dislocations. Phys. Rev. Lett. 1991, 66, 1721–1724. 37 38 39 5. Kuhnke, K.; Kern, K. Vicinal Metal Surfaces as Nanotemplates for the Growth of Low- 40 41 Dimensional Structures. J. Phys.: Condens. Matter 2003, 15, S3311–S3335. 42 43 44 6. Repain, V.; Baudot, G.; Ellmer, H.; Rousset, S. Two-Dimensional Long-Range-Ordered 45 46 Growth of Uniform Cobalt Nanostructures on a Au(111) Vicinal Template. Europhys. Lett. 47 48 2002, 58, 730–736. 49 50 51 7. Schmid, M.; Kresse, G.; Buchsbaum, A.; Napetschnig, E.; Gritschneder, S.; Reichling, M.; 52 53 Varga, P. Nanotemplate with Holes: Ultrathin Alumina on Ni3Al(111). Phys. Rev. Lett. 2007, 54 55 99, 196104. 56 57 58 59 60 15 ACS Paragon Plus Environment ACS Nano Page 16 of 20

1 2 3 4 8. Novotny, Z.; Argentero, G.; Wang, Z.; Schmid, M.; Diebold, U.; Parkinson, G. S. Ordered 5 6 Array of Single Adatoms with Remarkable Thermal Stability: Au/Fe3O4(001). Phys. Rev. 7 8 Lett. 2012, 108, 216103. 9 10 9. Yang, X.-F.; Wang, A.; Qiao, B.; Li, J.; Liu, J.; Zhang, T. Single-Atom Catalysts: A New 11 12 13 Frontier in Heterogeneous Catalysis. Acc. Chem. Res. 2013, 46, 1740–1748. 14 15 10. Qiao, B.; Wang, A.; Yang, X.; Allard, L. F.; Jiang, Z.; Cui, Y.; Liu, J.; Li, J.; Zhang, T. 16 17 Single-Atom Catalysis of CO Oxidation using Pt /FeO . Nat. Chem. 2011, 3, 634–641. 18 1 x 19 20 11. Sun, S.; Zhang, G.; Gauquelin, N.; Chen, N.; Zhou, J.; Yang, S.; Chen, W.; Meng, X.; 21 22 Geng, D.; Banis, M. N. et al. Single-Atom Catalysis Using Pt/Graphene Achieved through 23 24 Atomic Layer Deposition. Sci. Rep. 2013, 3, 1775. 25 26 27 12. Fu, Q.; Saltsburg, H.; Flytzani-Stephanopoulos, M. Active Nonmetallic Au and Pt Species on 28 29 Ceria-Based Water-Gas Shift Catalysts. Science 2003, 301, 935–938. 30 31 32 13. Kyriakou, G.; Boucher, M. B.; Jewell, A. D.; Lewis, E. A.; Lawton, T. J.; Baber, A. E.; 33 34 Tierney, H. L.; Flytzani-Stephanopoulos, M.; Sykes, E. C. H. Isolated Metal Atom Geometries 35 36 as a Strategy for Selective Heterogeneous Hydrogenations. Science 2012, 335, 1209–1212. 37 38 39 14. Lei, Y.; Mehmood, F.; Lee, S.; Greeley, J.; Lee, B.; Seifert, S.; Winans, R. E.; Elam, J. W.; 40 41 Meyer, R. J.; Redfern, P. C. et al. Increased Silver Activity for Direct Propylene Epoxidation 42 43 via Subnanometer Size Effects. Science 2010, 328, 224–228. 44 45 46 15. Campbell, C. T. Ultrathin Metal Films and Particles on Oxide Surfaces: Structural, Electronic 47 48 and Chemisorptive Properties. Surf. Sci. Rep. 1997, 27, 1–111. 49 50 51 16. Baumer,¨ M.; Freund, H.-J. Metal Deposits on Well-Ordered Oxide Films. Prog. Surf. Sci. 52 53 1999, 61, 127–198. 54 55 56 17. Gatel, C.; Snoeck, E. Comparative Study of Pt, Au and Ag Growth on Fe3O4(001) Surface. 57 58 Surf. Sci. 2006, 600, 2650–2662. 59 60 16 ACS Paragon Plus Environment Page 17 of 20 ACS Nano

1 2 3 4 18. Stanka, B.; Hebenstreit, W.; Diebold, U.; Chambers, S. Surface Reconstruction of Fe3O4(001). 5 6 Surf. Sci. 2000, 448, 49–63. 7 8 19. Wiesendanger, R.; Shvets, I. V.; Burgler,¨ D.; Tarrach, G.; Guntherodt,¨ H. J.; Coey, J. M. D.; 9 10 11 Graser,¨ S. Topographic and Magnetic-Sensitive Scanning Tunneling Microscope Study of 12 13 Magnetite. Science 1992, 255, 583–586. 14 15 20. Pentcheva, R.; Wendler, F.; Meyerheim, H. L.; Moritz, W.; Jedrecy, N.; Schefﬂer, M. Jahn- 16 17 18 Teller Stabilization of a “Polar” Metal Oxide Surface: Fe3O4(001). Phys. Rev. Lett. 2005, 94, 19 20 126101. 21 22 21. Łodziana, Z. Surface Verwey Transition in Magnetite. Phys. Rev. Lett. 2007, 99, 206402. 23 24 25 22. Parkinson, G. S.; Novotny, Z.; Argentero, G.; Schmid, M.; Pavelec, J.; Kosak, R.; Blaha, P.; 26 27 Diebold, U. Carbon Monoxide-Induced Adatom Sintering in a Pd-Fe O Model Catalyst. Nat. 28 3 4 29 Mater. 2013, 12, 724–728. 30 31 32 23. Parkinson, G. S.; Novotny, Z.; Jacobson, P.; Schmid, M.; Diebold, U. Room Temperature 33 34 Water Splitting at the Surface of Magnetite. J. Am. Chem. Soc. 2011, 133, 12650–12655. 35 36 37 24. Addou, R.; Senftle, T. P.; O’Connor, N.; Janik, M. J.; van Duin, A. C. T.; Batzill, M. In- 38 39 ﬂuence of Hydroxyls on Pd Atom Mobility and Clustering on Rutile TiO (011)-2×1. ACS 40 2 41 Nano [Online early access]. DOI: 10.1021/nn501817w. Published Online: May 7, 2014. 42 43 http://pubs.acs.org/doi/abs/10.1021/nn501817w (accessed June 11, 2014). 44 45 46 25. Parkinson, G. S.; Manz, T. A.; Novotny, Z.; Sprunger, P. T.; Kurtz, R. L.; Schmid, M.; 47 48 Sholl, D. S.; Diebold, U. Antiphase Domain Boundaries at the Fe O (001) Surface. Phys. 49 3 4 50 Rev. B 2012, 85, 195450. 51 52 53 26. Bartelt, N. C.; Nie, S.; Starodub, E.; Bernal-Villamil, I.; Gallego, S.; Vergara, L.; Mc- 54 55 Carty, K. F.; De la Figuera, J. Order-Disorder Transition on the (100) Surface of Magnetite. 56 57 Phys. Rev. B 2013, 88, 235436. 58 59 60 17 ACS Paragon Plus Environment ACS Nano Page 18 of 20

1 2 3 4 27. Silly, F.; Castell, M. R. Bimodal Growth of Au on SrTiO3(001). Phys. Rev. Lett. 2006, 96, 5 6 086104. 7 8 28. Soria, F.; Poppa, H. Comparison of the Early Stages of Condensation of Cu and Ag on 9 10 11 Mo(100) with Cu and Ag on W(100). J. Vac. Sci. Technol. 1980, 17, 449–452. 12 13 29. Winau, D.; Itoh, H.; Schmid, A. K.; Ichinokawa, T. Reconstructions and Growth of Ag on 14 15 Si(001)(2×1). Surf. Sci. 1994, 303, 139–145. 16 17 18 30. Glueckstein, J. C.; Evans, M. M. R.; Nogami, J. Surface Unwetting During Growth of Ag on 19 20 Si(001). Phys. Rev. B 1996, 54, R11066–R110069. 21 22 23 31. Takeuchi, O.; Kageshima, M.; Sakama, H.; Kawazu, A. Atomically Resolved Nucleation and 24 25 Initial Growth of a Ag Three-Dimensional Island on Si(001) Substrate. Phys. Rev. B 2011, 83, 26 27 205433. 28 29 30 32. Kong, D.; Drucker, J. Tuning Ag/Si(100) Island Size, Shape, and Density. J. Appl. Phys. 2013, 31 32 114, 144310. 33 34 35 33. Ernst, K. H.; Ludviksson, A.; Zhang, R.; Yoshihara, J.; Campbell, C. T. Growth Model for 36 37 Metal Films on Oxide Surfaces: Cu on ZnO(0001)-O. Phys. Rev. B 1993, 47, 13782–13796. 38 39 40 34. Starr, D. E.; Ranney, J. T.; Larsen, J. H.; Musgrove, J. E.; Campbell, C. T. Measurement of 41 42 the Energetics of Metal Film Growth on a Semiconductor: Ag/Si(100)-(2×1). Phys. Rev. Lett. 43 44 2001, 87, 106102. 45 46 47 35. Kaden, W. E.; Wu, T.; Kunkel, W. A.; Anderson, S. L. Electronic Structure Controls Reactivity 48 49 of Size-Selected Pd Clusters Adsorbed on TiO2 Surfaces. Science 2009, 326, 826–829. 50 51 52 36. Li, H.; Li, L.; Li, Y. The Electronic Structure and Geometric Structure of Nanoclusters as 53 54 Catalytic Active Sites. Nanotechnol. Rev. 2013, 2, 515–528. 55 56 57 58 59 60 18 ACS Paragon Plus Environment Page 19 of 20 ACS Nano

1 2 3 4 37. Nie, S.; Starodub, E.; Monti, M.; Siegel, D. A.; Vergara, L.; El Gabaly, F.; Bartelt, N. C.; De la 5 6 Figuera, J.; McCarty, K. F. Insight into Magnetite’s Redox Catalysis from Observing Surface 7 8 Morphology during Oxidation. J. Am. Chem. Soc. 2013, 135, 10091–10098. 9 10 11 38. Blaha, P.; Schwarz, K.; Madsen, G. K. H.; Kvasnicka, D.; Lutz, J. WIEN2k, An Augmented 12 13 Plane Wave + Local Orbitals Program For Calculating Crystal Properties 2001, (Karlheinz 14 15 Schwarz, Techn. Univ. Wien, Austria, 2001). 16 17 18 39. Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made Simple. 19 20 Phys. Rev. Lett. 1996, 77, 3865–3868. 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 19 ACS Paragon Plus Environment ACS Nano Page 20 of 20

1 2 3 4 Graphical TOC Entry 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 20 ACS Paragon Plus Environment