Superatoms: Electronic and Geometric Effects on ... - ACS Publications

Feb 9, 2017 - This work will focus on the reactivity of aluminum based clusters, although these concepts may be applied to clusters of different metal...
0 downloads 20 Views 5MB Size
Article pubs.acs.org/accounts

Superatoms: Electronic and Geometric Effects on Reactivity Arthur C. Reber and Shiv N. Khanna* Department of Physics, Virginia Commonwealth University, 701 W. Grace St., Richmond, Virginia 23284, United States S Supporting Information *

CONSPECTUS: The relative role of electronic and geometric effects on the stability of clusters has been a contentious topic for quite some time, with the focus on electronic structure generally gaining the upper hand. In this Account, we hope to demonstrate that both electronic shell filling and geometric shell filling are necessary concepts for an intuitive understanding of the reactivity of metal clusters. This work will focus on the reactivity of aluminum based clusters, although these concepts may be applied to clusters of different metals and ligand protected clusters. First we highlight the importance of electronic shell closure in the stability of metallic clusters. Quantum confinement in small compact metal clusters results in the bunching of quantum states that are reminiscent of the electronic shells in atoms. Clusters with closed electronic shells and large HOMO−LUMO (highest occupied molecular orbital− lowest unoccupied molecular orbital) gaps have enhanced stability and reduced reactivity with O2 due to the need for the cluster to accommodate the spin of molecular oxygen during activation of the molecule. To intuitively understand the reactivity of clusters with protic species such as water and methanol, geometric effects are needed. Clusters with unsymmetrical structures and defects usually result in uneven charge distribution over the surface of the cluster, forming active sites. To reduce reactivity, these sites must be quenched. These concepts can also be applied to ligand protected clusters. Clusters with ligands that are balanced across the cluster are less reactive, while clusters with unbalanced ligands can result in induced active sites. Adatoms on the surface of a cluster that are bound to a ligand result in an activated adatom that reacts readily with protic species, offering a mechanism by which the defects will be etched off returning the cluster to a closed geometric shell. The goal of this Account is to argue that both geometric and electronic shell filling concepts serve as valuable organizational principles that explain a wide variety of phenomena in the reactivity of clusters. These concepts help to explain the fundamental interactions that allow for specific clusters to be described as superatoms. Superatoms are clusters that exhibit a well-defined valence. A superatom cluster’s properties may be intuitively understood and predicted based on the energy gained when the cluster obtains its optimal electronic and geometric structure. This concept has been found to be a unifying principle among a wide variety of metal clusters ranging from free aluminum clusters to ligand protected noble metal clusters and even metal−chalcogenide ligand protected clusters. Thus, the importance of electronic and geometric shell closing concepts supports the superatom concept, because the properties of certain clusters with well-defined valence are controlled by the stability that is enhanced when they retain their closed electronic and geometric shells.



INTRODUCTION

Clusters offer a nearly endless number of possible compositions and sizes, so how can we identify a central dogma, or organizing principles that could provide a consolidated framework for classifying the diverse range of observed behaviors? The first potential unifying framework is that metal clusters possess electronic shells that result from the quantum confinement of the nearly free electron gas.4 This electronic shell structure was first observed in experiments on the abundance spectra of small alkali clusters, generated in beams, which identified the existence of strikingly abundant species, termed “magic” clusters.4 Not only did magic clusters present enhanced stability in mass spectra, but their ionization potentials,5 polarizability,6 and reactivity2 exhibited periodic patterns consistent with the magic clusters having similarly

Small clusters composed of a few to a few hundred atoms exhibit unique physical, chemical, electronic, and magnetic properties that may differ significantly from the properties of bulk or those in individual atoms. Extensive research over the past 30 years has revealed many examples of the unique properties of clusters; noble gold may form clusters that are catalysts,1 the highly reactive aluminum can form unreactive clusters,2 and nonmagnetic rhodium forms magnetic clusters.3 The properties of clusters can vary dramatically with size, charged state, and composition, and the addition or removal of a single atom may transform the cluster’s properties. In the cluster size regime, every atom and even every electron counts. These size and charge dependent properties make clusters the ultimate nanoscale laboratory for identifying the role of electronic and geometric structure on reactivity. © 2017 American Chemical Society

Received: September 14, 2016 Published: February 9, 2017 255

DOI: 10.1021/acs.accounts.6b00464 Acc. Chem. Res. 2017, 50, 255−263

Article

Accounts of Chemical Research unique properties due to the periodic pattern of electronic shell closure. Geometric structure also plays a critical role in the chemical stability and properties of clusters. When reacting with polar etchants such as water or methanol, certain aluminum cluster anions that have closed electronic shells are found to be highly reactive. Clusters such as Al77(N(SiMe3)2)20 and Ni9Te6(PR3)8 form materials7,8 despite having open electronic shells. Clusters with open geometric shells have defect-like structures that result in an uneven charge distribution over the surface of the cluster that results in active sites and enhances reactivity. When the cluster forms a more compact and symmetric structure, the charge distribution becomes more even, quenching the active sites and making all sites on the cluster equally reactive. The second organizing principle for understanding the reactivity and stability of clusters is that clusters with closed geometric shells have reduced reactivity, while clusters with geometries that are akin to defects enhance the reactivity. The chemical and physical properties of specific clusters are dominated by the enhanced stability the cluster gains from closing an incomplete electronic shell, granting the cluster a well-defined valence.9 This fact along with the similarity in the orbital shapes and chemical and electronic behaviors led Khanna and Jena to propose that specific clusters could be classified as superatoms9 and could be regarded as forming a third dimension of the periodic table. Later collaborative work, largely between the Khanna and Castleman groups, popularized the concept by identifying numerous superatoms with different valence types. The concept was initially applied to ligated gold clusters by Whetten and Häkkinen.10 The concept has also been recently applied by Roy, Nuckolls, and co-workers to ligated chalcogenide clusters.8 Khanna and co-workers have also extended the concept to magnetic superatoms.11,12 The purpose of this paper is to highlight the dual organizing principles for understanding the stability of superatom clusters: the importance of both electronic shell structure and geometric shell structure for understanding the reactivity of clusters, the ultimate frontier of nanoscience.

Figure 1. Electronic structure and atomic and molecular orbitals of Cl− and Al13−.

indicate gaps in energy; double lines indicate the boundary between filled and unfilled orbitals). The number of valence electrons, Ev, is determined by eq 1, where the cluster has N atoms and Va corresponds to the number of valence electrons for each atom (1 for alkali, 2 for alkaline earth, 3 for aluminum, and 1 for noble metals). Z is the net charge of the cluster. N

Ev =

∑ Va − Z 1

(1)

For the electronic shell structure of a cluster to be well described by the spherical harmonics, the cluster must be approximately spherical and metallic. This analysis becomes more complicated when using transition metals because the ordering of the nearly free electron gas is driven by the orthogonality principle, while atomic d orbitals are quite localized and contain two nodes, which complicates their incorporation into a nearly free electron gas.13 The power of this simple model is evident when we compare the atomic orbitals that form the backbone of the periodic table to the orbitals of aluminum clusters in Figure 2. This table charts the



SUPERATOMS: ELECTRONIC SHELL CLOSURE The role of electronic shells in the stability and electronic properties of metal clusters initially arose from the observations made by Knight and co-workers4 on “magic numbers” in the mass spectra of free sodium clusters. These studies involved measuring the size-dependent abundance of clusters using mass spectrometry. The measured intensities showed enhanced abundancies for clusters containing 2, 8, 18, 20, 34, 40, etc. atoms. Clemenger and co-workers used a “jellium” model to account for the enhanced stability of these specific alkali clusters.4 In this spherical jellium model, the positive charge of the cluster ions is distributed uniformly over a sphere the size of the cluster. The valence electrons behave as a nearly free electron gas confined by the positive background. In such a potential, the electronic states are grouped into shells where the angular and radial parts of the wave functions may be described using the spherical harmonics and a quantum number n. Clusters with filled shells and a large HOMO−LUMO gap exhibit enhanced stability and reduced reactivity. Figure 1 shows the ordering of the shells in the icosahedral Al13− cluster and further reveals that the orbitals are similar both to the atomic orbitals of Cl− and the expected ordering from the jellium model. The electronic levels in Al13− correspond to |1S2| 1P6|1D102S2|2P61F14||2D01G0| shell structure (single lines

Figure 2. Periodic patterns in the orbitals of atoms and aluminum clusters. Each succeeding column corresponds to an orbital, not an additional electron.

frontier orbitals and not the number of electrons, so alkali and alkaline earth atoms are part of the same column. Periodic patterns in the valence electron count of clusters leads to periodic closed electronic shell clusters. Clusters are not always spherical, and the grouping of electronic orbitals is intimately tied to the geometry of the cluster. A vivid example of the relationship between electronic and geometric structure is that geometrical distortions can lead to the splitting of electronic shells into subshells with a large HOMO−LUMO gap granting the cluster the stability 256

DOI: 10.1021/acs.accounts.6b00464 Acc. Chem. Res. 2017, 50, 255−263

Article

Accounts of Chemical Research

character. The existence of these magnetic superatoms and the associated multiplicity has been confirmed by experiments.17

associated with a magic species. Unlike atoms, a cluster with a partially filled subshell can stabilize itself either by having a high spin multiplicity (Hund’s rule) or by undergoing a Jahn−Teller distortion to form a singlet state with a gap within a subshell. The case of Al11Mg2− illustrates this competition. Figure 3



ELECTRONIC SHELLS AND REACTIVITY WITH OXYGEN The reactivity of metallic clusters with oxygen probes the chemical stability of clusters. In oxygen etching experiments, clusters are sent through a flow-tube filled with O2 gas, and the mass spectra identifies the abundance of the reacted species to determine stability. Early experiments on the reactivity of aluminum anions with oxygen showed that aluminum clusters displayed a strong variation with size, whereas bulk aluminum is highly reactive with oxygen.2 The residual mass spectra showed large peaks at Al13−, Al23−, and Al37−. These clusters have 40, 70, and 112 electrons, respectively, which corresponds to closed electronic shells.2,18 The microscopic origin of the reduced reactivity of clusters with O2 is related to the role of spin.18 Molecular oxygen is spin triplet in its ground state, and the lowest two unfilled orbitals are a pair of minority orbitals that are antibonding in nature. The etching of aluminum clusters typically proceeds through the following reaction.19

Figure 3. Stabilization of Al11Mg2− through Hund’s coupling as compared to Jahn−Teller distortion.

shows that Al11Mg2− would require 40 valence electrons for a filled electronic shell. The cluster has 38 valence electrons so it may form an icosahedral structure with a triplet multiplicity or undergo an oblate distortion. The oblate structure is 0.03 eV more stable than the pseudoicosahedral triplet structure. This demonstrates the challenge in stabilizing magnetic moments in clusters, because unlike atoms, clusters may distort to quench their unpaired electrons. This phenomenon where a cluster that is two electrons short of a closed electronic shell undergoes an oblate deformation has been observed in many other systems.14−16 One way to stabilize magnetic species is to have a compound cluster with a combination of localized and delocalized states as we demonstrated for Na8V or FeMg8 clusters.11,12 Here the stability is afforded by the delocalized electrons while the localized d-states can breed the magnetic

Al n− + 3O2 → Al n − 4 − + 2Al 2O

(2)

The reactivity entails filling of the minority LUMO’s in O2 to activate the oxygen molecule. The filling of these orbitals may be thought of as a spin excitation from a multiplicity of 3 to 1 on the oxygen half of the complex. Aluminum and oxygen have negligible spin−orbit effects, so the reaction proceeds in a manner to conserve the overall spin.20 For clusters with an even number of electrons, conservation of spin requires that the decrease in the spin multiplicity of the oxygen be accommodated by a spin excitation of the cluster. Clusters with a large spin excitation energy (SPE) will not easily accommodate the triplet spin of oxygen forming a spin barrier to reactivity. For clusters with an odd number of electrons, spin conservation does not require a spin excitation, and the 3

Figure 4. Experimental mass spectra (top) after exposure to O2, and the HOMO−LUMO gap and SPE of the Aln−, Al4Hm−, and Agn− clusters. Blue lines and squares indicate clusters that are resistant to oxygen etching. 257

DOI: 10.1021/acs.accounts.6b00464 Acc. Chem. Res. 2017, 50, 255−263

Article

Accounts of Chemical Research reaction proceeds rapidly.18,19 The spin excitation in clusters with singlet ground states is a transfer of the minority electron into a majority unfilled state; the SPE is therefore determined by the HOMO−LUMO gap. Figure 4 shows the experimental mass spectra of Aln−, Al4Hm−, and Agn− clusters after etching with oxygen along with the calculated SPE and the HOMO− LUMO gaps. Clusters that survive the etching reaction are all marked by high SPE or HOMO−LUMO gaps as it provides a reasonable estimate of SPE. To show the microscopic phenomenon of spin transfer, Figure 5 demonstrates what

Figure 6. Value of the HOMO−LUMO gap for a variety of aluminum based clusters as a function of the number of valence electrons. Triangles indicate anions, squares indicate neutrals, and the bar is the average gap. Data for this figure is given in Table S1.

stability of the compound Au104(SR)44,26 shown in Figure 7, may be understood through the superatom concept in which

Figure 5. Spin density (blue) and corresponding O2 binding energy for several aluminum based clusters.

happens when an O2 molecule is brought from a distance to approach Al13−, as a cluster with a large SPE of 1.80 eV. The spin remains localized on the O2 complex, and the binding energy is 0.77 eV. Singlet O2 has no barrier for insertion into Al13−, which is confirmed by experiments with singlet O2.19,21 Al5− is completely etched in the experimental spectrum and has a low SPE of 0.01 eV. When O2 was brought toward Al5−, it inserted into the cluster without barrier. The spin effects also control the binding energy of the O2. Al4H3− has a relatively large SPE of 1.16 eV, and Al3− has a negligible SPE. A single O2 is brought closer to both the clusters along a direction that does not lead to breaking of the O−O bond. In both cases, the O−O bond stretched to 1.6 Å, and the spin has transferred to the cluster. Despite the similar structures, the binding energy of O2 to Al4H3− is 2.36 eV, while it is 3.46 eV for Al3−. This demonstrates that the spin excitation energy causes a shift in the potential energy surface of the reaction so clusters with high SPE are less reactive, and clusters with low SPE are more reactive. Oxygen etching experiments may then rapidly identify clusters with large HOMO−LUMO gaps (SPE) and closed electronic shells. Numerous experiments on the reactivity of Aln−, AlnHm−,22 AlmIm−,23 AlnCm−,24 MgnAln−,16 BnAlm−,25 and CuAln− clusters15 confirm that the spin excitation barrier is the dominant parameter governing the reactivity with oxygen. Figure 6 summarizes the HOMO−LUMO gaps for all the above clusters. Clusters marked by a HOMO−LUMO gap exceeding 1.0 eV are generally resistant to reaction with oxygen while clusters with smaller gaps are reactive. This 1.0 eV line is only a guideline for aluminum based clusters, and might vary for other elements with different oxygen binding energies.

Figure 7. Structure and HOMO−LUMO gap of a number of stable superatom clusters based on different classes of ligands. The asterisks indicate that the cluster has unpaired electrons.

the outer 44 gold atoms are linked to thiols, while the remaining 58 gold atoms provide valence electrons corresponding to a closed electronic shell of 58 electrons. Häkkinen and co-workers have demonstrated that nearly all of the highly stable gold−thiol clusters may be viewed as superatoms with closed shells and large HOMO−LUMO gaps.10 Other examples of highly stable superatom clusters are shown in Figure 7, with gold clusters with 8 valence electrons, Au25(SR)18−,27,28 Au13Cl2(PR3)103+,29 and Au11(PR3)103+,30 and two aluminum based clusters, the 40 valence electron Al14I3− cluster and the borane analog Al12(CR3)122− clusters,31 all exhibiting closed electronic shells. Al14I3− is also found to be resistant to O2 etching in flow-tube experiments.23 What is intriguing is that some of these crystallized, highly stable superatomic clusters have open electronic shells. For example, Ni9Te6(PR)8 and Al77(NR2)202−, which form stable assemblies, are both experimentally verified to have unpaired electrons.7,8,32 While most superatomic clusters have closed electronic shells, the fact that these clusters are highly stable despite their open electronic shell implies that the electronic shell model is not the only organizing principle that controls reactivity and that alternative concepts must play a role.



ELECTRONIC SHELL CLOSURE AND LIGAND PROTECTED SUPERATOMS Closed electronic shells with large HOMO−LUMO gaps are also observed in ligand protected clusters that form stable cluster assemblies. The addition of a covalent ligands such as a thiol, halogen, pnictide, or a group IV element changes the effective number of valence electrons to reach a closed electronic shell providing electronic stability. For example, the 258

DOI: 10.1021/acs.accounts.6b00464 Acc. Chem. Res. 2017, 50, 255−263

Article

Accounts of Chemical Research



and Al20− all survive methanol etching experiments, and Al6− also appears to survive, although the initial abundance is low enough that it is difficult to confirm this.35 How then can we intuitively understand this size selected reactivity? The interaction proceeds via the donation of lonepair electrons from water to the LUMO of the cluster (or LUMO + 1 in odd-electron species). The ability of the cluster to accept the lone pair of the water molecule is the first fundamental interaction that determines the reactivity. Because it is the ability of the cluster to accept not a single electron but a lone pair of electrons, the role of this active site on the cluster is best described as a Lewis acid. Charge density on the cluster stabilizes the Al−H bond, so the lowest barrier would occur if a Lewis acid site that accepts a lone pair was adjacent to a Lewis base site that could donate charge to stabilize the Al−H bond. This two center active site where one atom serves as a Lewis acid and a neighboring site serves as a Lewis base is called a complementary active site. In Figure 9, the LUMO + 1

SUPERATOMS: GEOMETRIC EFFECTS ON REACTIVITY Not all reactions between clusters and molecules are subjected to the same fundamental constraints as reactivity with O2. A striking example of an alternative concept for understanding reactivity is shown through the size-selective reactivity of aluminum cluster anions with water and methanol.33−35 Al23− and Al37− are etched when exposed to water or methanol, despite being resistant to etching with O2. Al11− and Al20− are resistant to reaction with water and methanol despite being reactive with O2 and having open electronic shells. A large HOMO−LUMO gap is insufficient to protect a cluster from water etching, while open shell clusters may exhibit resistance to reactivity with water. The first question is whether the reactivity corresponds to the cleavage of the O−H bond, or simple adsorption of the molecule. Al12− is the most reactive aluminum cluster anion with water and methanol, and Al13− is the least. To examine the reactivity, we examined ETS defined in eq 3. E TS = EA − E B (3) where EA is the energy barrier for breaking the OH bond starting from the absorbed state and EB is the nondissociative binding energy of the molecule. A positive value of ETS implies a slow reaction because the barrier for reaction is larger than the desorption energy, while a negative ETS implies a rapid reaction. Figure 8A shows the reaction pathways for the

Figure 9. Anatomy of a complementary active site.

protrudes from the top of the Al12− cluster signaling that this is the best binding site. On an adjacent site, there is appreciable HOMO charge density. The ETS is lowest when the O is bound to the Lewis acid site and the H to the Lewis base site. Al13− is resistant to reaction with water, and the nearly spherical structure results in the even distribution of charge over the surface of the cluster. Thus, no site on the cluster may be considered a better or worse Lewis acid or Lewis base site. These considerations showed that the reactivity is controlled by the presence of complementary active Lewis acid/Lewis base sites. A further proof of the breaking of the OH bond came from the observation that the adsorption of H2O molecules by certain clusters such as Al16−, Al17−, and Al18− led to products that were deficient from the stoichiometry of water by two or four hydrogen atoms. Al17− has two sets of aluminum atom dimers located upon opposing vertices of a 13 atom icosahedral core (Figure 10). Appreciable HOMO density is located on dimers, and the LUMO protrudes into vacuum at adjacent sites, allowing for the rapid cleavage of water. The dissociation of the first water molecule induces a neighboring Lewis active site, resulting in the dissociation of a second water molecule and two H atoms on adjacent sites that selectively combine and leave the cluster as a H2 molecule. These findings conclusively confirmed the cleavage of the O−H bond. Theoretical investigations on Al16− and Al18− showed that like Al17− these

Figure 8. (A) The reaction pathway for Al13− and Al12−. (B) The ETS, binding energy, and dissociative binding energy for the reactivity of aluminum cluster anions with methanol.

cleavage of the O−H bond of methanol with Al12− and Al13−. Al12− is found to have a methanol binding energy of 0.53 eV with a ETS of −0.25 eV, while Al13− has a methanol binding energy of 0.23 eV and an ETS of +0.26 eV. We have plotted the EB, ETS, and DBE (dissociative binding energy) of methanol to Aln− clusters in Figure 8B. Four clusters are found to have positive ETS, n = 6, 11, 13, and 20. Experimentally, Al11−, Al13−, 259

DOI: 10.1021/acs.accounts.6b00464 Acc. Chem. Res. 2017, 50, 255−263

Article

Accounts of Chemical Research

while for Al20−, the LUMO + 1 sites are delocalized over the equator of the cluster, and the HOMO sites are located at the poles. Al23− has a reactive ETS of −0.23 eV located on the edge site, while Al20− has an unreactive ETS of +0.14 eV due to the delocalized LUMO and lack of a good adjacent Lewis base site. Figure 11B summarizes the barriers for all of the Lewis acid and Lewis base sites for Al23− in which red indicates the lower barriers, and blue indicates large barriers to reactivity. The reactions are most likely to occur on the edges of the cluster indicating that the complementary pairs begin to accumulate at the edges as the cluster size increases. We suspect that this is the reason that all aluminum clusters anion larger than n = 20 are reactive. The complementary active sites not only are active in breaking OH bonds in H2O and alcohols but can also break stronger polar bonds. A reaction of particular interest is the dissociation of the carbonyl group in formaldehyde: the polar CO bond has a bond dissociation energy of 7.79 eV versus 5.17 eV for the O−H bond in water. The carbonyl bond of formaldehyde could be fragmented by complementary active sites on size-selective aluminum cluster anions.36 Here, the adjacent Lewis acid and Lewis base sites bind with O and C to stabilize a resonance structure where the carbonyl is reduced to a single bond leading to its cleavage. This demonstrates that complementary Lewis acid−Lewis base sites are useful for reactions with a variety of polar bonds.

Figure 10. Mechanism for H2 production from Al17− + 2H2O. The inset figure shows the frontier orbitals.

clusters also have an icosahedral core with two atom adjuncts atop one vertex possessing similarly comparable locations of high LUMO and HOMO probability density.34 The presence of complementary Lewis acid and Lewis base pairs is governed by the geometrical structure of the cluster. This means that clusters with symmetrical, closed geometric shells are likely to have an even distribution of charge over the surface of the cluster, while clusters with defect-like structures will generally have uneven charge distributions and become more reactive. Al20− is the large aluminum cluster anion that exhibits reduced reactivity, while Al23− has a closed electronic shell and is resistant to etching by molecular oxygen but reacts with water. The ground state structure of Al23− is a hexagonal packing with 3-fold longitudinal edges as seen in Figure 11A, and Al20− has a prolate structure that is a double icosahedron with an additional Al atom in the bottom ring. In Al23−, the frontier orbitals are located on the sharp edges of the cluster,



LIGAND INDUCED ACTIVE SITES The preceding findings demonstrate that the reactivity with polar molecules is reduced when the charge density is more uniformly distributed. One strategy for reducing the reactivity of a cluster is to use ligands. Ligands generally reduce the reactivity by changing the effective number of valence electrons through bonding. However, as these ligands are electronegative, they may also induce active sites by distorting the charge distribution over the surface of the cluster. Our recent studies on the etching of AlnIm− clusters with methanol has highlighted this issue.37,38 The cluster reaction experiments were conducted in a fast flow reactor, and theoretical studies of the reactivity and active sites of AlnIm− were examined for n = 7−14 and m = 0−2.37 Four superatom clusters with closed electronic shells, Al13−, Al13I2−, Al7I2−, and Al14I3−, were identified by peaks in the mass spectra of unreacted clusters. Al7I2− has a closed shell of 20 e−, and Al13−, Al13I2−, and Al14I3− have 40 e− closed shells, with Al13I2− undergoing hypervalent bonding analogous to I3−.39 Three of these four clusters were found to be resistant to methanol etching; however, Al14I3− was found to be reactive despite having a closed electronic shell. This dual reactive pattern is due to geometric effects even though the cluster is protected by ligands. To examine the reactivity, we investigated the transition state energies, ETS. In Figure 12, the reaction pathway for Al13I2− in its ground state is compared with the pathway for a higher energy isomer of Al13I2− that has two iodine atoms bonded to adjacent aluminum atoms.38 We see that while ETS is positive for the cluster in its ground state, it is negative for the cluster with adjacent ligands showing that the higher energy structure is reactive. The frontier orbitals reveal that the adjacent ligands have induced an active site on the opposite side of the cluster. This shows that balanced ligands will stabilize a ligand protected cluster, while unbalanced ligands may induce an active site increasing the reactivity.

Figure 11. (A) Lewis acid (blue) and Lewis base (red) sites on Al20− and Al23− with the lowest energy ETS shown. (B) Heat map of the ETS of different sites on Al23− with red being most active and blue being inactive. 260

DOI: 10.1021/acs.accounts.6b00464 Acc. Chem. Res. 2017, 50, 255−263

Article

Accounts of Chemical Research

Figure 12. Reaction pathway for Al13I2− for lowest energy structure with balanced ligands and structure with adjacent ligands.

Al14I3− has a metallic core with a 13 atom icosahedron and an adatom, with one iodine bound to the adatom and another to an Al atom on the opposite side of the cluster. The ETS at the iodine site is −0.27 eV, indicating that the cluster reacts rapidly (Figure 13). The lowest energy transition state at the metal site

Figure 14. Venn diagram indicating which clusters exhibit enhanced resistance to O2 etching and methanol or water etching.

including water, methanol, and formaldehyde. Clusters with defected structures result in an uneven distribution of charge over the surface of the cluster, forming complementary Lewis acid and Lewis base sites that make the cluster reactive. These concepts are also applicable to ligand protected clusters as the ligands can be used to control valence electron counts. Clusters with balanced ligands result in even charge distribution over the surface of the cluster, while unbalanced ligands lead to induced active sites. Ligand activated adatoms also serve as an etching mechanism that drives clusters toward closed geometric shells. As we demonstrated, the complementary active sites can even break strong carbonyl bonds, and hence these concepts could be used to generate fine chemicals. The enhanced stability that certain clusters receive when they obtain a closed electronic shell and a closed geometric shell, results in the cluster having a well-defined valence. The superatoms form a third dimension of the periodic table due to the convergence of these two concepts, and therefore these form a set of powerful organizing principles for intuitively understanding the reactivity and stability of metal clusters.

Figure 13. Reaction pathway of Al14I3− with methanol at the activated adatom site and metal site.

is +0.30 eV, suggesting that the cluster will only react at the adatom iodine site. The electronic structure reveals that a LUMO + 3 orbital located on the adatom may serve as a Lewis acid site. The ETS at the metal site is +0.30 eV, revealing that Al14I3− is unreactive at the metal site. Further analysis reveals that the adatom is only activated when bound to the iodine ligand. This analysis is confirmed by the nearly complete etching of Al14I3− in the experimental spectra.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.accounts.6b00464. Jahn−Teller distortions in clusters, details for Figure 2, and raw data for Figure 6 (PDF)



CONCLUSIONS In this Account, we have highlighted the dual role of electronic and geometric factors governing the reactivity of superatomic clusters. The quantum confinement in clusters can lead to the grouping of electronic states into shells, and the electronic orbitals are spread over multiple atoms leading to geometric effects. Consequently, both the electronic shell structure (HOMO−LUMO gaps) and geometric shell filling are necessary concepts for an intuitive understanding of the reactivity of metal clusters with a variety of etchants. Figure 14 shows a Venn diagram representing the dual reactivity patterns needed to identify clusters that are stable under multiple conditions. Aluminum and silver cluster anions that have closed electronic shells and HOMO−LUMO gap larger than 1.0 eV tend to exhibit resistance to reactivity with O2 due to their large spin excitation energy required for a conservation of spin angular momentum. Clusters with closed geometric shells tend to exhibit reduced reactivity with protic species



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Shiv N. Khanna: 0000-0002-9797-1289 Notes

The authors declare no competing financial interest. Biographies Arthur C. Reber is an Associate Professor of Physics at Virginia Commonwealth University. He received his B.A. from Earlham College and his Ph.D. from the University of Chicago. His research focuses on identifying organizing concepts in reactivity at the nanoscale. 261

DOI: 10.1021/acs.accounts.6b00464 Acc. Chem. Res. 2017, 50, 255−263

Article

Accounts of Chemical Research

Copper Cluster Anions. Phys. Rev. B: Condens. Matter Mater. Phys. 2010, 81, 195404. (16) Luo, Z.; Grover, C. J.; Reber, A. C.; Khanna, S. N.; Castleman, A. W. Probing the Magic Numbers of Aluminum−Magnesium Cluster Anions and Their Reactivity toward Oxygen. J. Am. Chem. Soc. 2013, 135, 4307−4313. (17) Zhang, X.; Wang, Y.; Wang, H.; Lim, A.; Gantefoer, G.; Bowen, K. H.; Reveles, J. U.; Khanna, S. N. On the Existence of Designer Magnetic Superatoms. J. Am. Chem. Soc. 2013, 135, 4856−4861. (18) Reber, A. C.; Khanna, S. N.; Roach, P. J.; Woodward, W. H.; Castleman, A. W. Spin Accommodation and Reactivity of Aluminum Based Clusters with O2. J. Am. Chem. Soc. 2007, 129, 16098−16101. (19) Burgert, R.; Schnöckel, H.; Grubisic, A.; Li, X.; Stokes, S. T.; Bowen, K. H.; Ganteför, G. F.; Kiran, B.; Jena, P. Spin Conservation Accounts for Aluminum Cluster Anion Reactivity Pattern with O2. Science 2008, 319, 438−442. (20) Wigner, E.; Witmer, E. E. Ü ber die Struktur der zweiatomigen Molekelspektren nach der Quantenmechanik. Eur. Phys. J. A 1928, 51, 859−886. (21) Woodward, W. H.; Eyet, N.; Shuman, N. S.; Smith, J. C.; Viggiano, A. A.; Castleman, A. W. Aluminum Cluster Anion Reactivity with Singlet Oxygen: Evidence of Al9− Stability. J. Phys. Chem. C 2011, 115, 9903−9908. (22) Roach, P. J.; Reber, A. C.; Woodward, W. H.; Khanna, S. N.; Castleman, A. W. Al4H7− Is a Resilient Building Block for Aluminum Hydrogen Cluster Materials. Proc. Natl. Acad. Sci. U. S. A. 2007, 104 (37), 14565−14569. (23) Bergeron, D. E.; Roach, P. J.; Castleman, A. W.; Jones, N. O.; Khanna, S. N. Al Cluster Superatoms as Halogens in Polyhalides and as Alkaline Earths in Iodide Salts. Science 2005, 307 (5707), 231−235. (24) Reveles, J. U.; Khanna, S. N.; Roach, P. J.; Castleman, A. W. Multiple Valence Superatoms. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 18405−18410. (25) Smith, J. C.; Reber, A. C.; Khanna, S. N.; Castleman, A. W. Boron Substitution in Aluminum Cluster Anions: Magic Clusters and Reactivity with Oxygen. J. Phys. Chem. A 2014, 118, 8485−8492. (26) Jadzinsky, P. D.; Calero, G.; Ackerson, C. J.; Bushnell, D. A.; Kornberg, R. D. Structure of a Thiol Monolayer-Protected Gold Nanoparticle at 1.1 Å Resolution. Science 2007, 318, 430−433. (27) Akola, J.; Walter, M.; Whetten, R. L.; Häkkinen, H.; Grönbeck, H. On the Structure of Thiolate-Protected Au25. J. Am. Chem. Soc. 2008, 130, 3756−3757. (28) Zhu, M.; Aikens, C. M.; Hollander, F. J.; Schatz, G. C.; Jin, R. Correlating the Crystal Structure of A Thiol-Protected Au25 Cluster and Optical Properties. J. Am. Chem. Soc. 2008, 130, 5883−5885. (29) Briant, C. E.; Theobald, B. R. C.; White, J. W.; Bell, L. K.; Mingos, D. M. P.; Welch, A. J. Synthesis and X-Ray Structural Characterization of the Centred Icosahedral Gold Cluster Compound [Aul3(PMe2Ph)10Cl2](PF6)3; the Realization of a Theoretical Prediction. J. Chem. Soc., Chem. Commun. 1981, 201−202. (30) Bertino, M. F.; Sun, Z.-M.; Zhang, R.; Wang, L.-S. Facile Syntheses of Monodisperse Ultrasmall Au Clusters. J. Phys. Chem. B 2006, 110, 21416−21418. (31) Klinkhammer, K.-W.; Uhl, W.; Wagner, J.; Hiller, W. K2[Al12iBu12], a Compound with Al12 Icosahedra. Angew. Chem., Int. Ed. Engl. 1991, 30, 179−180. (32) Chauhan, V.; Sahoo, S.; Khanna, S. N. Ni9Te6(PEt3)8C60 Is a Superatomic Superalkali Superparamagnetic Cluster Assembled Material (S3-CAM). J. Am. Chem. Soc. 2016, 138, 1916−1921. (33) Roach, P. J.; Woodward, W. H.; Castleman, A. W.; Reber, A. C.; Khanna, S. N. Complementary Active Sites Cause Size-Selective Reactivity of Aluminum Cluster Anions with Water. Science 2009, 323, 492−495. (34) Reber, A. C.; Khanna, S. N.; Roach, P. J.; Woodward, W. H.; Castleman, A. W. Reactivity of Aluminum Cluster Anions with Water: Origins of Reactivity and Mechanisms for H2 Release. J. Phys. Chem. A 2010, 114, 6071−6081. (35) Reber, A. C.; Roach, P. J.; Woodward, W. H.; Khanna, S. N.; Castleman, A. W. Edge-Induced Active Sites Enhance the Reactivity of

Shiv N. Khanna is a Commonwealth Professor of Physics at Virginia Commonwealth University. Dr. Khanna is internationally recognized for his work on “superatoms” where he and co-workers discovered that selected clusters can take on the chemical behavior of atoms in the periodic table and that materials could be developed using such clusters as building blocks, forming a third dimension of the periodic table.



ACKNOWLEDGMENTS The authors gratefully acknowledge support from the Department of Energy under award number DE-SC0006420.



ABBREVIATIONS HOMO, highest occupied molecular orbital; LUMO, lowest unoccupied molecular orbital; SPE, spin excitation energy; DBE, dissociative binding energy



REFERENCES

(1) Sanchez, A.; Abbet, S.; Heiz, U.; Schneider, W.-D.; Häkkinen, H.; Barnett, R. N.; Landman, U. When Gold Is Not Noble: Nanoscale Gold Catalysts. J. Phys. Chem. A 1999, 103, 9573−9578. (2) Leuchtner, R. E.; Harms, A. C.; Castleman, A. W. Thermal Metal Cluster Anion Reactions: Behavior of Aluminum Clusters with Oxygen. J. Chem. Phys. 1989, 91, 2753. (3) Cox, A. J.; Louderback, J. G.; Bloomfield, L. A. Experimental Observation of Magnetism in Rhodium Clusters. Phys. Rev. Lett. 1993, 71, 923−926. (4) Knight, W. D.; Clemenger, K.; de Heer, W. A.; Saunders, W. A.; Chou, M. Y.; Cohen, M. L. Electronic Shell Structure and Abundances of Sodium Clusters. Phys. Rev. Lett. 1984, 52, 2141−2143. (5) Cox, D. M.; Trevor, D. J.; Whetten, R. L.; Kaldor, A. Aluminum Clusters: Ionization Thresholds and Reactivity toward Deuterium, Water, Oxygen, Methanol, Methane, and Carbon Monoxide. J. Phys. Chem. 1988, 92, 421−429. (6) Knight, W. D.; Clemenger, K.; de Heer, W. A.; Saunders, W. A. Polarizability of Alkali Clusters. Phys. Rev. B: Condens. Matter Mater. Phys. 1985, 31, 2539−2540. (7) Ecker, A.; Weckert, E.; Schnöckel, H. Synthesis and Structural Characterization of an AI77 Cluster. Nature 1997, 387, 379−381. (8) Roy, X.; Lee, C.-H.; Crowther, A. C.; Schenck, C. L.; Besara, T.; Lalancette, R. A.; Siegrist, T.; Stephens, P. W.; Brus, L. E.; Kim, P.; Steigerwald, M. L.; Nuckolls, C. Nanoscale Atoms in Solid-State Chemistry. Science 2013, 341 (6142), 157−160. (9) Khanna, S. N.; Jena, P. Atomic Clusters: Building Blocks for a Class of Solids. Phys. Rev. B: Condens. Matter Mater. Phys. 1995, 51, 13705−13716. (10) Walter, M.; Akola, J.; Lopez-Acevedo, O.; Jadzinsky, P. D.; Calero, G.; Ackerson, C. J.; Whetten, R. L.; Grönbeck, H.; Häkkinen, H. A Unified View of Ligand-Protected Gold Clusters as Superatom Complexes. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 9157−9162. (11) Reveles, J. U.; Clayborne, P. A.; Reber, A. C.; Khanna, S. N.; Pradhan, K.; Sen, P.; Pederson, M. R. Designer Magnetic Superatoms. Nat. Chem. 2009, 1, 310−315. (12) Medel, V. M.; Reveles, J. U.; Khanna, S. N.; Chauhan, V.; Sen, P.; Castleman, A. W. Hund’s Rule in Superatoms with Transition Metal Impurities. Proc. Natl. Acad. Sci. U. S. A. 2011, 108, 10062− 10066. (13) Medel, V. M.; Reber, A. C.; Chauhan, V.; Sen, P.; Köster, A. M.; Calaminici, P.; Khanna, S. N. Nature of Valence Transition and Spin Moment in AgnV+ Clusters. J. Am. Chem. Soc. 2014, 136, 8229−8236. (14) Luo, Z.; Gamboa, G. U.; Smith, J. C.; Reber, A. C.; Reveles, J. U.; Khanna, S. N.; Castleman, A. W. Spin Accommodation and Reactivity of Silver Clusters with Oxygen: The Enhanced Stability of Ag13−. J. Am. Chem. Soc. 2012, 134, 18973−18978. (15) Roach, P. J.; Woodward, W. H.; Reber, A. C.; Khanna, S. N.; Castleman, A. W. Crystal Field Effects on the Reactivity of Aluminum262

DOI: 10.1021/acs.accounts.6b00464 Acc. Chem. Res. 2017, 50, 255−263

Article

Accounts of Chemical Research Large Aluminum Cluster Anions with Alcohols. J. Phys. Chem. A 2012, 116, 8085−8091. (36) Woodward, W. H.; Reber, A. C.; Smith, J. C.; Khanna, S. N.; Castleman, A. W. Carbonyl Bond Cleavage by Complementary Active Sites. J. Phys. Chem. C 2013, 117, 7445−7450. (37) Luo, Z.; Reber, A. C.; Jia, M.; Blades, W. H.; Khanna, S. N.; Castleman, A. W. What Determines If a Ligand Activates or Passivates a Superatom Cluster? Chem. Sci. 2016, 7, 3067−3074. (38) Abreu, M. B.; Powell, C.; Reber, A. C.; Khanna, S. N. LigandInduced Active Sites: Reactivity of Iodine-Protected Aluminum Superatoms with Methanol. J. Am. Chem. Soc. 2012, 134, 20507− 20512. (39) Bergeron, D. E.; Castleman, A. W.; Morisato, T.; Khanna, S. N. Formation of Al13I−: Evidence for the Superhalogen Character of Al13. Science 2004, 304, 84−87.

263

DOI: 10.1021/acs.accounts.6b00464 Acc. Chem. Res. 2017, 50, 255−263