Ag Nanoclusters Revealing the Evolutionary

Oct 7, 2018 - Department of Chemistry, Carnegie Mellon University , Pittsburgh , Pennsylvania 15213 , United States. ‡ Department of Chemistry, Univ...
6 downloads 0 Views 813KB Size
Subscriber access provided by UNIV OF LOUISIANA

Article

A Correlated Series of Au/Ag Nanoclusters Revealing the Evolutionary Patterns of Asymmetric Ag Doping Yingwei Li, Tian-Yi Luo, Meng Zhou, Yongbo Song, Nathaniel L Rosi, and Rongchao Jin J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.8b08335 • Publication Date (Web): 07 Oct 2018 Downloaded from http://pubs.acs.org on October 10, 2018

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 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 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.

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 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

A Correlated Series of Au/Ag Nanoclusters Revealing the Evolutionary Patterns of Asymmetric Ag Doping Yingwei Li†§, Tian-Yi Luo‡§, Meng Zhou†§, Yongbo Song†, Nathaniel L. Rosi‡, and Rongchao Jin†* †Department

of Chemistry, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213, United States of Chemistry, University of Pittsburgh, Pittsburgh, Pennsylvania 15260, United States

‡Department

§These

authors contributed equally to this work.

Abstract: Doping of metal nanoclusters is an effective strategy for tailoring their functionalities for specific applications. To gain fundamental insight into the doping mechanism, it is of critical importance to have access to a series of correlated bimetal nanoclusters with different doping levels and further reveal the successive transformations. Herein we report asymmetric doping of Ag into an Au21 nanocluster to form a series of new Au/Ag bimetal nanoclusters and the effects of doping on the evolution of size, structure and properties based upon X-ray crystallography and optical spectroscopy analyses. The asymmetric doping discovered in the series reveals two important rules. First, the heteroatom doping-induced kernel transformation mechanism is revealed, explaining the successive conversions from Au21(S-Adm)15 with an incomplete cuboctahedral kernel to Au20Ag1(S-Adm)15 with a complete cuboctahedral Au12Ag1 kernel, then to Au19Ag4(S-Adm)15 with an icosahedral Au10Ag3 kernel. The electron density accumulated on the central Au atom(s) is rationalized to force an expansion of radial metal-metal bond angles, which triggers the cuboctahedral-to-icosahedral kernel conversion. This mechanism is generalized by elucidating several other cases. Second, through comparison of a series of seven nanoclusters (all protected by adamantanethiolate), we find that the unit cell symmetry of their crystals is correlated with the symmetry of the cluster’s kernel. Specifically, we observe a sequential change from triclinic to monoclinic to trigonal unit cell in the series with increasing kernel symmetry. The kernel structure-dependent optical properties are also discussed.

1 ACS Paragon Plus Environment

Journal of the American Chemical Society 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1. Introduction Atomically precise metal nanoclusters (NCs) possess unique structures and rich properties dictated by their structures and quantum size effects.1-6 A number of crystal structures have been reported,1 but most of them are independent structures (i.e., without mutual relationships). Therefore, mapping out the relationships is critical in order to reveal the cluster growth mechanism with size, structural evolution pattern, and the scaling laws for the size-dependences. To achieve fundamental understanding of nanocluster evolution patterns, it is of critical importance to create the correlated series of NCs so that the intrinsic connections and many fundamental issues could be revealed. In the literature, some interesting series of gold NCs have been reported. One is the series of Au28(SR)20, Au36(SR)24, Au44(SR)28, and Au52(SR)32,7 in which a Au4 helical double-stranded growth mode was observed. A second series is the alternate single-stranded evolution at both ends in the Au28(SR)20, Au34(SR)22, and Au42(SR)26 series.8 As for Au NCs protected by mixed ligands, an interesting series comprises the mono-icosahedral [Au13(PR)10X2]3+,9,10 rod-shaped bi-icosahedral [Au25(PR)10(SR)5X2]2+,11 and linear-shaped tri-icosahedral Au37(PR)10(SR)10X2)+,12 wherein PR = phosphine, X = halide. These series of gold NCs are interesting because multiple sizes are correlated, providing not only the information about the evolution of crystallographic structures, but also an exciting opportunity to find out the scaling laws on their electronic structure13 and optical properties.14,15 On the other hand, the correlated series of bimetal NCs are still relatively rare. Doping is a promising way of tailoring the functionality of NCs.16–22 Apart from replacing gold atom(s) with retained structures of the homogold counterparts where the dopants are either in definite position(s)23–28 or a distribution of sites,29–31 the doping approach also leads to many bimetal NCs with new structures and properties such as catalytic activity, photoluminescence, and singlet oxygen generation.16–28 The successful formation of the bimetal structures depends on several factors, such as the heterometal-gold bonding, metal-ligand interactions, and their geometric/electronic structure shell closing.32 The majority of the reported bimetallic NCs are independent, i.e., no correlation of their size, structure and properties. Nevertheless, some correlated series of bimetal NCs have been reported. In early work on phosphineprotected Au/Ag NCs by Teo et al., a series of “cluster of clusters” was reported,33 and the electronic structure was recently analyzed by Gao and coworkers.34 The Au/Ag series includes the icosahedral [Au9M4(PR)8X4]2+ (M = Au/Ag/Cu),35 rod-shaped [Au13Ag12(PR)10X8]+,36 triangle-shaped [Au18Ag20(PR)12X14] and [Au18Ag19(PR)12X11]2+.37,38 For thiolate-protected NCs, Zheng et al. reported [Au12+nCu32(SR)30+n]4- with n = 0, 2, 4, and 6.39 The Au12Cu32(SR)30 structure is the same as Ag44(SR)30 and Au12Ag32(SR)30,17,40 and the six Cu2(SR)5 surface unitswhich are octahedrally mounted on the Au12@Cu20 kernelcan be substituted pair-by-pair with Cu2Au(SR)6 units.39 A site-specific , pair replacement of surface motifs was also achieved recently in Au23(SR)16 via Ag doping.41 In another series, the icosahedral Au13 kernel was decorated by 2, 4, or 8 Cu atoms which were all triply coordinated by thiolate or pyridyl groups.42 The Cu doping examples demonstrate surface pattern evolution.39,42 From the perspective of controlling the electronic and optical properties,1-3 it would be more appealing if heteroatom(s) can be doped into the kernel of the NCs, as the kernel dictates the HOMO-LUMO gap and the optical properties of NCs.1-6,43-46 For kernel doping, we found an interesting case that upon heavy doping of Ag into [Au23(SR)16]- with a cuboctahedral (cubo-) kernel, a structure transformation to [Au25-xAgx(SR)18]- with an icosahedral (ico-) kernel was observed.31 The cubo- structure (in atomic layers of a/b/c) is indicative of the face-centered cubic (fcc) structure that is exclusively adopted by bulk gold. The 12-coordinate feature apparently plays an important role in stabilizing both cubo- and ico-Au13 entities.35 Beside the transformation of cubo-[Au23(SR)16]- to ico-[Au25-xAgx(SR)18]-, Zhu and coworkers recently reported a two-way isoelectronic conversion between Au25 and Au23 studied by MALDI mass spectrometry (MS).47 Xie’s group also employed MS to demonstrate that the conversion from cubo[Au23(SR)16]- to ico-[Au25(SR)18]- is caused by “surface-motif-exchange-induced symmetry-breaking core structure transformation mechanism”.48 In the case of bimetal NCs, it has been demonstrated that the 2 ACS Paragon Plus Environment

Page 2 of 16

Page 3 of 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

incoming Ag atoms first go into the kernel and trigger the conversion from cubo-Au23 to ico-Au25-xAgx at sufficient doping,41 and then the dopants further go to the surface motif(s) at heavy doping,31 but the intrinsic driving force that causes the transformation has not been elucidated. To address the dopinginduced structural conversion mechanism, a successive transformation series (i.e. multiple correlated structures in one series) would be particularly helpful, which may provide a convincing answer. Herein, we present the successful attainment of a correlated bimetal NC series starting with Au21(S-Adm)15, including Au20Ag1(S-Adm)15, Au19Ag4(S-Adm)15, Au23-xAgx(S-Adm)15 (x ~5.76), and Au23xAgx(S-Adm)15 (x ~7.44), all five being synthesized under the same conditions except adjusting the Au/Ag ratio in the precursors (see SI for details). With this successive seriesin which Ag is gradually doped into the cluster in an asymmetric manner, we are able to reveal two important rules. First, the asymmetric doping of Ag triggers the cubo-to-ico kernel transformation that is dictated by the large difference in electronegativity between Ag and Au, with this rule further generalized to other four bimetal cases that have not been elucidated mechanistically. Second, we find that the symmetries of the NC kernels greatly affect the symmetries of the unit cells of macroscopic crystals; this rule is generalized to other clusters such as the fcc-kernel series.7 2. Results and Discussion 2.1. Formation of Au19Ag4(S-Adm)15 We first discuss the Au19Ag4(S-Adm)15 NC (Figure 1A), which was obtained by co-reduction of AuI-SR and AgI-SR with NaBH4 via size focusing (Figure S1). The MALDI-MS spectrum of Au19Ag4(S-Adm)15 is shown in Figure S2. Crystallization was performed in dichloromethane/ethanol. Crystallographic details are given in SI. Au19Ag4(S-Adm)15 is crystallized in P21/n space group. X-ray crystallography found no counterion associated with the cluster (Figure 1B), thus, it should be charge neutral. Au19Ag4(S-Adm)15 has an ico-Au10Ag3 kernel (Figure 1C) with the three silver atoms forming a bottom triangle; this asymmetric doping is a new feature and different from all other reported bimetal NCs in which doping atoms were always found to distribute evenly.16–22,30,50 The top Au3 triangle in the ico-Au10Ag3 kernel is capped by a Ag(SR)3 “tri-podal paw” (Figure 1C). This quasi-planar Ag(SR)3 motif was first observed in [Ag17(SR)12]3-,49 then in the Au4Ag13 NC,50 and now in Au19Ag4, indicating that it could be a common motif of silver. The bottom of the Au10Ag3 kernel is further protected by three trimeric Au3(SR)4 staples, and the arrangement of the three trimers resembles a tri-blade fan along the quasi-C3 axis,51 which endows chirality to the Au19Ag4 cluster (Figure S3).

Figure 1. Total structure of the Au19Ag4(S-Adm)15 nanocluster: (A) An enantiomer of the chiral Au19Ag4(S-Adm)15 nanocluster; (B) Unit cell with a monoclinic arrangement; (C) Anatomy of the structure. Labels: magenta/violet = Au, light grey = Ag, yellow = S, and grey = C.

Importantly, the synthetic method of Au19Ag4 provided us an opportunity to prepare a series of correlated bimetal NCs by simply changing the AuI/AgI molar ratio in the precursors (Scheme 1); no 3 ACS Paragon Plus Environment

Journal of the American Chemical Society 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

delicate kinetic control or thermal etching processes was involved. When Au was the only metal in the system, we obtained the Au21(S-Adm)15 NC, which was the same as the cluster reported previously,52,53 evidenced by identical UV-vis and consistent MALDI-MS data (Figure S4). The kernel of Au21 can be considered as an Au12 cuboctahedron with one vertex missing (i.e. incomplete cubo-Au12). 52,53

Scheme 1. Evolution of crystal structure depending on the Au/Ag ratio in the salt precursors. Labels: magenta = Au, light grey = Ag, pink = Au or Ag, and yellow = S.

When a trace amount of silver (salts of Au:Ag = 25:1, molar) was present in the metal precursors, we obtained another cluster, Au20Ag1(S-Adm)15 (Scheme 1 and detailed structure in Figure S5, UV-vis and MALDI-MS in Figure S6). The Au20Ag1 NC was crystallized in P-1 space group. The kernel of Au20Ag1(SAdm)15 can be regarded as a complete cubo-Au12Ag1 capped by an additional Au atom (see the arrow in Figure S5), with protection by two dimers, one trimer and five bridging -S(R)- ligands. The structure of Au20Ag1(S-Adm)15 is indeed shared by Au20Ag1(S-tert-butyl)15,54 albeit the -R groups are different, which proves the similarity of adamantanethiol and tert-butylthiol in both S atoms connecting to tertiary carbons. Other Au23-xAgx(S-Adm)15 clusters with the same crystal structure as that of Au19Ag4(S-Adm)15 but with higher Ag doping were prepared simply by changing the initial molar ratio of Au:Ag in the precursors (Scheme 1); details are discussed in section 2.4. 2.2. Kernel transformation from cuboctahedron to icosahedron by asymmetric Ag doping The three NCs (Au21, Au20Ag1, and Au19Ag4, all being protected by 15 adamantanethiolates) are correlated, shedding light on the mechanism that causes the cubo-to-ico structural transformation. It is the asymmetric Ag doping that induces the transformation with increasing Ag dopants. The sequence is as follows: doping one Ag atom into the incomplete cubo-Au12 kernel of Au21(S-Adm)15 triggers the formation of a complete cubo-Au12Ag1 kernel in Au20Ag1(S-Adm)15, and further doping two more silver atoms triggers the conversion of the cubo-Au12Ag1 to the ico-Au10Ag3 kernel in the Au19Ag4(S-Adm)15. Starting with Au21(S-Adm)15, it has a Au12 kernel (cubo with a vertex missing) and this kernel is capped by two specific “hub” atoms (Au(1) and Au(2) in Figure 2a/b)on which a Au(SR)2 staple is mounted. As to the Au20Ag1(S-Adm)15, the cubo-Au12Ag1 kernel is also capped by one “hub” atom (herein Au(2), Figure 2c), on which a Au2(SR)3 staple is also attached. Furthermore, when some of the motifs on both Au21 and Au20Ag1 are omitted for clarity, the remaining structures (Figure S7) are quite similar except for the distortions in the middle planes of the two kernels (Figure 2b/d). We suppose that when Au(3) (Figure 2a/b) is replaced by Ag(3) (Figure 2c/d) in the kernel, one “hub” atom (herein Au(1)) moves into the middle plane (Figure 2b), pushes Au(4) up, and completes the cubo-Au12Ag1 kernel in Au20Ag1(SAdm)15. Then, the question is: what is the driving force? We rationalize that the large difference in electronegativity () between Ag (=1.93) and Au (=2.54) should be the reason. The asymmetric doping of Ag in the kernel leads to an increase in electron density on the central atom (Auc); the radial bonds are accordingly rationalized to repel from each other so as to expand the bond angles (Figure 2b/d).

4 ACS Paragon Plus Environment

Page 4 of 16

Page 5 of 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

Figure 2. Structural comparison between Au21(S-Adm)15 and Au20Ag1(S-Adm)15: (a/c) Front views of Au21(SAdm)15/Au20Ag1(S-Adm)15; (b/d) Top views of the middle planes in kernels of Au21(S-Adm)15/Au20Ag1(SAdm)15. Colors: magenta = Au, white = capping Au, light grey = Ag, and yellow = S.

This rationale is verified when comparing Au20Ag1(S-Adm)15 with Au19Ag4(S-Adm)15 (detailed comparison is shown in Figure S8). By Ag replacing two more kernel Au atoms adjacent to Ag(3) (e.g. Au(5) and Au(6) in Figure 3a-d), the structure of the kernel is converted from cubo-Au12Ag1 to icoAu10Ag3. The driving force is again the difference in electronegativity. As shown in Figure 3b/d, we divide the cubo- or ico-M13 kernels (M = Au/Ag) into a middle plane and a M7 bitetrahedron to illustrate the details. In the cubo-Au12Ag1 kernel (Figure 3b), the average bond angles (M-Auc-M) of its hexagonal middle plane and M7 are 59.8° and 60.5°, respectively. In Au19Ag4(S-Adm)15, the asymmetric distribution of the three Ag atoms, however, makes a difference; that is, the three Ag atoms gathering together at the bottom triangle of the kernel and one more capping Ag on the top (Figure 3c) significantly increase the electron density on the two layers of Au atoms between the top and the bottom, especially on the middle plane. A stronger repulsion between the Au-Au bonds is evidenced by the observation that the average bond angle of the middle plane (Au-Auc-Au) increases distinctly to 63.7° (Figure 3d). In order to accommodate the expansion, the initially hexagonal middle plane has to corrugate, i.e. three atoms move up and the other three move down (see white arrows in Figure 3a), forming a chair configuration (Figure 3c). Accordingly, within the M7 unit (Figure 3d), one observes the difference in the average bond angles of the upper Au4 tetrahedron (Au-Auc-Au = 69.7°) and the bottom Au1Ag3 tetrahedron (AgAuc-Ag = 60.3°), Figure 3d right.

Figure 3. Structural comparison between Au20Ag1(S-Adm)15 and Au19Ag4(S-Adm)15: (a/c) Front view of Au20Ag1(S-Adm)15/Au19Ag4(S-Adm)15; (b/d) Middle planes and M7 in the kernels of Au20Ag1(SAdm)15/Au19Ag4(S-Adm)15. Labels: magenta = Au, white = capping Au, light grey = Ag, and yellow = S.

5 ACS Paragon Plus Environment

Journal of the American Chemical Society 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The above step-by-step transformation insights lead to an important hypothesis, i.e., the large difference in electronegativity between Au and Ag in the kernel triggers the successive kernel conversion of gold and bimetal nanoclusters. Furthermore, this heteroatom-induced kernel transformation mechanism can be generalized to other transformation cases (vide infra). It should be noted that all the five nanoclusters have the same number of ligands (i.e. 15 SAdm); thus, when more atoms are in the kernel, the kernel has to become more compact so that the icosahedron is more suitable due to its smaller volume compared to the cuboctahedron.1 Moreover, the free-electron number increases from 6 to 8 when clusters with M11 kernels evolve to clusters with M13 kernels. From the electronic-structure point of view, the 8 electrons in the M13 kernel would increase the electron density (8 e/13 atoms = 0.62 e/atom) compared to the case of 6 electrons in the M11 kernel (6 e/11 atoms = 0.55 e/atom). Such a large difference makes the M-M bonds in M13 kernels more repelling to each other, and this factor may also contribute to the geometrical change from cuboctahdedron to icosahedron. 2.3. Generalization of the transformation mechanism to other doping cases Case 1. Ag doping-induced transformation of cubo-[Au23(SR)16] to ico-[Au25-xAgx(SR)18]. In previous research, it remained unclear why Ag doping could cause the transformation from the cubo[Au23(SR)16] NC55 through [Au23-xAgx(SR)16]- (avg. x = 0.9) to ico-[Au25-xAgx(SR)18]- (avg. x = 4.5), then to heavily doped ico-[Au25-xAgx(SR)18]- (avg. x = 19.4), Figure S9a/b/c.31,41 When less than one Ag atom (the statistical value of x) is doped into [Au23-xAgx(SR)16], the middle plane and M7 bitetrahedron of the cuboAu13-xAgx kernel both show an average bond angle of ~60° (Figure S9d). When more than 4 Ag atoms are doped into the Au12 shell (and heavy doping of Ag even goes into the exterior staple motifs), the electron density on the central Au atom increases, expanding all the radial bond angles (M-Auc-M) in the middle plane and M7, hence, leading to the cubo-to-ico kernel transformation (Figure S9e/f). The trend implies the critical role of Ag doping into the kernel, which triggers the structural transformation. Case 2. Cd doping-induced transformation of cubo-[Au23(SR)16] to quasi-icoAu20Cd4(SH)(SC6H11)19. Recently, Wu and coworkers obtained Au20Cd4(SH)(SC6H11)19 with a distorted icoAu11Cd2 kernel (quasi-icosahedron) by anti-galvanic reaction (i.e. Au23(SR)16 reaction with CdII(SR)2).56 We rationalize that the kernel distortion is also caused by the doping of less electronegative Cd at asymmetric positions. On a note, in Wu’s case the cubo-to-ico transformation is not yet complete, i.e., the Au11Cd2 kernel is an intermediate structure between cubo- and ico-, judging from the average M-Auc distance in the kernel and standard deviation, Table S1. Taking the cases of Ag and Cd doping together, we can show a clear map of the cubo-to-ico transformation (Figure 4a-e). In the homogold Au13 kernel (Figure 4a), no valence electron transfer occurs due to the pure Au composition. When a single atom in the cuboctahedral Au12 shell is replaced by a less electronegative atom (Figure 4b), a small electron transfer from the doping atom to Au atom would occur. However, the change is too tiny to trigger any structure transformation, and the bond angles barely change (β = α = 60o), e.g. Au12Ag1 in Au20Ag1(SR)15. At the doping number of ~2 as shown in Figure 4c, the kernel transformation starts when more valence electron density is transferred to the central Au atom, resulting in an intermediate kernel structure between the cuboctahedron and icosahedron as was observed in the Cd doping, i.e. Au11Cd2 in Au20Cd4(SH)(SR)19.56 Such an intermediate state should also exist in the Ag doping (though not yet observed). Larger bond angle(s) of Au-Auc-Au comparing to Au-Auc-M (β > α > 60o) might be a feature as well. Finally, as three or more heteroatoms are doped into the kernel (e.g. Au10Ag3 in Au19Ag4(SR)15, and Au13-xAgx in [Au23-xAgx(SR)16]- with x = 4.5), the valence electron density transfer from less electronegative metal atoms to the central Au atom is 6 ACS Paragon Plus Environment

Page 6 of 16

Page 7 of 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

enough to achieve a complete transformation of the kernel from cuboctahedron to icosahedron (Figure 4d/e). If the heteroatoms sit close to each other, the bond angles of M-Auc-M would be significantly smaller than those of Au-Auc-Au as the valence electron densities are accumulated on the Au-Au bonds (β >> α > 60o in Figure 4d); if the heteroatoms are evenly distributed in the M12 shell, the bond angles would be close to 65o (β = α = 65o in Figure 4e).

Figure 4. Hypothesis of kernel structure transformation from cuboctahedron to icosahedron by different doping number of heteroatom M with lower electronegativity compared to Au (χM 4).

Case 3. Ag doping in the rod Au25 nanoclusters. The structures of Au25 and Au13Ag12 are shown in Figure S10.11,36,57 The Au25 rod is arranged in a stagger-eclipse-stagger form, in which the two middle pentagon layers (Au1 and Au2) are connected by five thiolates (Figure S10a).11 In contrast, the Au13Ag12 rod is arranged in a stagger-stagger-stagger form, with the two middle pentagon layers (Ag1 and Ag2) connected by six halides (Figure S10b), as opposed to five.36,57 Figure S10c/d emphasizes the middle parts of the two kernels: the average bond angle (Au1-Auc-Au2) in the rod Au25 is 63.2°, but with Ag atoms in the Au13Ag12 rod, electron density accumulates on the central Au atom, driving the two eclipsed layers to twist from each other so as to exhibit larger Ag1-Auc-Ag2 bond angles (avg. 68.1°). The distortion at the waist caused by stronger electronic repulsion makes the two Ag pentagon layers staggered, breaking three Ag-Ag bonds and giving room to an additional halide (total six at the waist). An obvious shrinkage in the length of the rod is also noticed (Figure S10a/b). Case 4. Ag doped Au15Ag3(SR)14. The Ag doping of Au18(SR)14 was found to lead to a more compact Au6Ag3 kernel in the Au15Ag3(SR)14,58 of which the total structure is almost the same as that of Au18(SR)14 (Figure S11a/b).59 With the implications from the above cases, we rationalize that the electron density should be higher at the top and bottom Au3 triangles. Indeed, the six bond angles at the top and bottom for the Au9 kernel of Au18 expand from Au2-Au1-Au2 = avg. 57.0° (Figure S11c) to Ag2-Au1-Ag2 = avg. 58.8° for the Au6Ag3 kernel of Au15Ag3 (Figure S11d). Overall, the series of Au21, Au20Ag1 and Au19Ag4 NCs, together with the above four cases, points to a convincing deduction that doping of less electronegative Ag (and Cd as well) into the Au kernel of a nanocluster makes electron density accumulate on Au atom(s) at the center, and subsequent expansion of radial bond angles triggers the kernel distortion (e.g. case 4 above) or even conversion to a different structure (e.g. cases 1 to 3).

7 ACS Paragon Plus Environment

Journal of the American Chemical Society 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Pei et al. reported the 2D positive charge density map as well as Hirshfeld calculation to show the charge difference between Au18(S-C6H11)14 and Au17Ag1(S-C6H11)14 (note: the two NCs share the same structure).60 The charge at the Ag-doping position (+) increases when Au is replaced by Ag, indicating that the Ag atom loses some electron density to atoms of higher electronegativity. This corroborates our hypothesis for our system. It should be noted that in our Au20Ag1(SR)15, the single Ag atom is only bonded to Au atoms, and the Ag3 triangle in the kernel of Au19Ag4(SR)15 are only bonded to Au atoms as well. Thus, in these two clusters, the electron density would only go from Ag to Au, and accordingly make Ag partially positive (δ+) whereas Au partially negative (δ-). 2.4. Symmetry relationship between kernels and unit cells We continue to study what might happen to the crystal structure when an even higher amount of AgI is used in the precursor. Two more Au/Ag nanoclusters were obtained, with their MALDI-MS spectra and zoom-in regions shown in Figure S12. As to the crystal structures, the Au/Ag = 8:1 synthesis  which gave rise to Au19Ag4(S-Adm)15 (Figure 5a)  is set as the reference. The structure of the NC from the Au/Ag = 4:1 synthesis is determined to be Au23-xAgx(S-Adm)15 (avg. x = 5.76), with the same structure as that of Au19Ag4(S-Adm)15, and the additional Ag atoms exclusively go to the ico-kernel with average 22% partial occupancy at each position (Figure 5b, pink) except that the center and one position in the top triangle (in circle) being still 100% occupied by Au (crystallographic details and the atomic percentages of Ag at all positions are given in Table S2). The peculiar top-Au-atom (in circle) is eventually replaced by Ag when further increasing the Ag amount to Au/Ag = 2:1 in the synthesis, with average Ag occupancy at each position in the kernel (Figure 5c, pink) being ~35% (Table S3). Of note, the three staple motifs also have partial Ag occupancy (~11% at each position, Figure 5c).

Figure 5. Structures of Au23-xAgx(S-Adm)15 with x = 4 (a), avg. x = 5.76 (b), and x = 7.44 (c). Colors: magenta = Au, pink = Au or Ag, light grey = Ag, and yellow = S.

Interestingly, the symmetry of the unit cell of Au23-xAgx (avg. x = 7.44) converts to trigonal (R-3c) compared to Au19Ag4 and Au23-xAgx (avg. x = 5.76), the latter two having a monoclinic unit cell (P21/n). A careful study has indeed allowed us to map out the relationship between the cluster symmetry and the unit cell symmetry. In the literature, although a number of crystal structures of Au, Ag and bimetal NCs have been reported,1,2 there has been no attempt yet to relate the cluster structure with the corresponding unit cell. The formation of crystals from cluster solutions reflects the packing mode of clusters in certain environment, and controllable crystallization of nanoclusters and nanoparticles still remains a challenging task.61–64 After analyzing the crystal structures of our current series of S-Adm protected Au/Ag NCs, we have found some intriguing feature which can be generalized to a universal rule for the NCs. Corresponding to the higher symmetry of unit cell, the Ag(SR)3 staple motif with a C3 axis is observed in 8 ACS Paragon Plus Environment

Page 8 of 16

Page 9 of 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

Au23-xAgx (x = 7.44), like that in the newly reported [Ag46S7(SR)24]+.65 The C3 axis in Au23-xAgx (x = 7.44) leads to the higher symmetry of the unit cell in comparison to lower doped Au19Ag4 and Au23-xAgx (avg. x = 5.76) of which the Ag(SR)3 staple motifs are distorted to lose the C3 axis. Thus, even in the NCs with the same atomic structure, slight distortion might occur due to Ag doping at different levels. Several factors might cause the observed change in the unit cell. First, counterions, e.g. tetraoctylammonium (TOA+) in [Au25(SR)18]- or [Au23(SR)16]- could be a major reason.55,61 However, our Au23-xAgx(S-Adm)15 system is neutral and no counterions is present; thus, this factor can be ruled out. Second, different solvents used in crystallization may also have some effects,61-64 but in our current system, we chose the same combination of solvents to crystallize all the three Au23-xAgx(S-Adm)15 NCs with different x values (4, 5.76 and 7.44); thus, the solvent effect can also be eliminated. Third, the protecting ligand can also potentially affect the packing mode in the unit cell through ligand interactions;65–68 but adamantanethiol we used in this system does not contain any benzene ring, so no π-π or C-Hπ interaction is involved in our system. Without the complications from all these factors, our current system provides a good opportunity to reveal how the structures of the clusters can be related to the symmetry of the unit cells. Specifically, we have identified that the kernel symmetry strongly dictates the symmetry of the unit cell. To compare the unit cells of the Au and Au/Ag nanoclusters, two more NCs with the same SAdm ligand, Au24(S-Adm)16 and Au30(S-Adm)18,69,70 are included to enrich our illustration (Figure 6a-g). The incomplete cubo-kernel of Au21 is asymmetric, and its corresponding unit cell is triclinic (Figure 6a). Au20Ag1(S-Adm)15 and Au24(S-Adm)16 have the complete cubo-Au12Ag1 and cubo-Au13 kernels (Figure 6b/c), respectively, but these cubo-kernels are still asymmetric, indicated by the large standard deviation (0.263 – 0.286 Å) in their radial M-Auc distances (Tables S4/S5).

Figure 6. The kernel structures and unit cells of (a) Au21(S-Adm)15; (b) Au24(S-Adm)16; (c) Au20Ag1(S-Adm)15; (d) Au19Ag4(S-Adm)15; (e) Au23-xAgx(S-Adm)15 (avg. x = 5.45); (f) Au23-xAgx(S-Adm)15 (avg. x = 7.85); (g) Au30(SAdm)18. Colors: magenta = Au, violet = center Au, light grey = Ag, cyan = Au/Ag, yellow = S, and grey = C.

Enhanced symmetry is found for ico-kernels in the Au23-xAgx ones (Figure 6d/e/f), as the standard deviation of radial M-Auc distances is greatly reduced to 0.051 – 0.068 Å (Tables S6/S7/S8). However, the different radial bond lengths of Au-Auc and Ag-Auc somewhat reduce the symmetry of these kernels. As such, both Au19Ag4 and Au23-xAgx (avg. x = 5.76) form monoclinic unit cells (Figure 6d/e). In the case of Au23-xAgx (x = 7.44), the symmetry of ico-kernel is largely improved when the peculiar Au in the top triangle of the M13 kernel is finally replaced by Ag. With the restoration of the C3 axis as shown in Figure 6f, a corresponding trigonal unit cell is achieved. The C3 axis is also observed in the Au114 kernel of alkynyl-protected Au144,71 and the cluster crystallizes in space group R-3m, indicating an intimate connection between the C3 axis and the trigonal unit cell. Finally, the Au18 kernel of Au30(S-Adm)18 9 ACS Paragon Plus Environment

Journal of the American Chemical Society 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(Figure 6g) is highly symmetric with a C3 axis as well as an inversion center, thus Au30(S-Adm)18 crystallization leads to a highly symmetric unit cell, i.e. cubic packing. The above analysis of the Au:adamantanethiolate series leads to a conclusion, that is, the symmetry of unit cell is highly dependent on the symmetry of the kernel. The ico-kernels lead to unit cells of higher symmetry comparing to the cubo-kernels, and with more symmetry operations in the kernel, the symmetry of the unit cell improves. To further illustrate this rule, we compare the fcc kernel series of Au28, Au36, Au44, and Au52 (Figure S13a-d),7 all protected by 4-tert-butylbenzenethiolate (TBBT) so as to eliminate the ligand effects just like in the above adamantanethiolate series. All the four clusters in the TBBT series are neutral. In Au36(TBBT)24, the Au28 kernel has a C2 axis as indicated in Figure S13b, and the Au36 kernel of Au44(TBBT)28 also has an inversion center (Figure S13c). The existence of symmetric operation in these two kernels corresponds to monoclinic unit cells for Au36(TBBT)24 and Au44(TBBT)28. By contrast, the kernels of both Au28(TBBT)20 and Au52(TBBT)32 are lack of symmetry, and correspondingly, Au28(TBBT)20 and Au52(TBBT)32 form the less symmetric triclinic unit cells (Figure S13a/d). From this point of view, we found that the symmetry of unit cell is not size-dependent, unlike other properties such as the excitedstate electron dynamics14 and catalytic activity.72–74 Instead, it is largely dependent on the symmetry of the kernel. Many other cases are also consistent with this identified rule. For example, the two isomers of Au28(TBBT)20 and Au28(S-c-C6H11)20 have the same fcc Au20 kernel,75 but the Au28(S-c-C6H11)20 is more symmetric (having a C2 axis) than Au28(TBBT)20, and thus, in the crystal form, Au28(S-c-C6H11)20 shows a monoclinic unit cell, rather than triclinic for Au28(TBBT)20 (Figure S14). Similarly, as shown in Figure S15, the fcc kernel of Au36(TBBT)24 76 also differs slightly to that of Au36(SPh)24,77 with the latter losing the C2 axis and thus having a less symmetric unit cell (triclinic), whereas Au36(TBBT)24 has a monoclinic unit cell. The crystal packing symmetry is expected to largely affect the charge transport,63 photoluminescence,50 and many other properties of crystals. Besides the known factors (e.g. counterions, inter-ligand interactions, and entropy),55,61-68 the kernel symmetry that we have identified herein may provide another way for controlled packing of nanoclusters into single crystals and exploration of ensemble properties. 2.5. Optical spectroscopic studies Due to the prominent quantum confinement effects, small clusters (< 2 nm) show an energy gap between HOMO and LUMO (Eg). We performed UV-vis analyses on Au21, Au20Ag1, Au19Ag4, Au23-xAgx (avg. x = 5.76), and Au23-xAgx (avg. x = 7.44) at room temperature and low temperatures (Figure 7a-b). The spectra are plotted in the form of absorption intensity (Absorbance × Wavelength2) vs. photon energy. The optical peaks shift to higher energies and become sharper with the emergence of fine structures at 80 K, similar to the observations by Ramakrishna and coworker in the study on Au25(SR)18-.78 The five clusters differ from the cryo behavior of the plasmonic Au279 nanocluster as the latter shows a single plasmon band and the band has no shift at cryo temperatures.6 For the icosahedral Au19Ag4, Au23-xAgx (avg. x = 5.76), and Au23-xAgx (avg. x = 7.44), the lowest-energy peak blue-shifts with increasing Ag doping, indicating the participation of Ag orbital(s) in the formation of HOMO and LUMO.79,80 We observed that the two peaks in the 2.1 to 2.2 eV range gradually split apart from each other. By comparing the Eg at RT and 80 K (see labels in Figure 7a-b), Eg enlarges at low temperatures since phonons are suppressed as the temperature decreases. The large difference in Eg between Au20Ag1 and the Au23-xAgx(S-Adm)15 ones is primarily attributed to different kernel structures as revealed by a systematic comparison of the clusters with cubo- and ico-kernels. The icosahedral NCs, e.g. the Au23-xAgx(SR)15 (8e) ones, and the [Au2510 ACS Paragon Plus Environment

Page 10 of 16

Page 11 of 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

xAgx(SR)18]

ones (8e), exhibit smaller Eg values (by ~0.5 eV) than the cuboctahedral NCs, e.g. Au20Ag1(SR)15 (6e), [Au19Cd2(SR)16]- (8e), and [Au23(SR)16]- (8e), as show in Figure S16.

Figure 7. The spectra on energy scale for (A) Au21(S-Adm)15; (B) Au20Ag1(S-Adm)15; (C) Au19Ag4(S-Adm)15; (D) Au23-xAgx(S-Adm)15 (avg. x = 5.76); and (E) Au23-xAgx(S-Adm)15 (avg. x = 7.44) at room temperature (blue lines) and 80 K (magenta lines); (F) Bar diagram of optical gaps for five clusters at RT or 80K.

As for Au nanoclusters of larger size, we can take a look at Au38(SR)24 (bi-icosahedral Au23 kernel), Au36(SR)24 (fcc Au24 kernel), and Au40(SR)24 (fcc Au25 kernel), cuboctahedron is also fcc structure. Au38(SR)24 has an Eg of ~0.92 eV, while that of Au36(SR)24 is ~1.72 eV, and Au40(SR)24 shows the onset of absorbance at ~780 nm (Eg ~1.59 eV).1 From this comparison, the conclusion that icosahedral nanoclusters exhibit smaller Eg values than the cuboctahedral ones is the same as for larger Au nanoclusters. Finally, by monitoring the optical spectra of the doped nanoclusters (x=4 and 7.44), we compared their thermal stability and found that Au19Au4 is rather stable (Figure S17) owing to its structural factor (e.g., the Ag3 triangle in the Au10Ag3 kernel of Au19Ag4(S-Adm)15 being fully protected by Au atoms) and electronic factor (8e). 3. Conclusion In summary, by simply controlling the Au/Ag molar ratio in the precursors, we have obtained a series of Au/Ag nanoclusters protected by adamantanethiolate. With the unprecedented asymmetric doping mode of Ag in Au20Ag1, Au19Ag4 as well as Au23-xAgx, we are able to reveal two mechanisms.

11 ACS Paragon Plus Environment

Journal of the American Chemical Society 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

First, the electronegativity difference between Au and Ag is identified to be the driving force in triggering the cubo-to-ico kernel conversion, as the electron density accumulated on the center Au atom(s) causes stronger repulsion between radial bonds, hence, expanding the bond angles and resulting in the kernel conversion. In our system, the asymmetric doping of a single Ag in the incomplete cubo-Au12 kernel of Au21 leads to the completion of a cub-Au12Ag1 kernel in Au20Ag1. Further asymmetric doping triggers the change to an ico-Au10Ag3 kernel in Au19Ag4. This mechanism also explains how Au23 with a cubo-kernel changes to Au25-xAgx with an ico-kernel at higher Ag doping amounts, as well as the transformations of several other doped NCs. Second, the symmetry of unit cell is found to be largely dictated by the symmetry of kernel (at least for the neutral clusters). This trend is first discovered due to the distinct change to trigonal unit cell at highest Ag doping amount, then extended to other adamantanethiolate-protected clusters (seven total). The symmetry of unit cells increases from triclinic to monoclinic to trigonal, and finally to cubic system as the corresponding kernels become more and more symmetric. The result is further verified by the crystal structures of six Aun clusters of fcc kernels. Low-temperature optical absorbance is also performed. Besides the fact that Ag doping increases the energy gap of clusters with the same structure, the structure of the kernel exerts an even stronger influence on Eg. Generally, Au or Au-based bimetal clusters with cubo-kernels have larger gaps compared to those with ico-kernels. Overall, this work not only presents novel structures of Au/Ag bimetal nanoclusters, but also provides rationalization on the successive structure transformation triggered by kernel doping with atom(s) of less electronegativity. Furthermore, the dictation of unit cell symmetry by the cluster kernel symmetry is also identified, which may lead to controllable crystallization of nanoclusters and exploration of the bulk properties of the macroscopic cluster-assembled crystals in future work. Associated content Supporting Information The Supporting Information is available free of charge on the ACS Publications website. Details of the synthesis, MALDI-MS, crystallization, X-ray analysis, and supporting Figures S1-S17, Table S1-S8 (PDF) X-ray crystallographic data for Au20Ag1(SC10H15)15, Au19Ag4(SC10H15)15, Au23-xAgx(SC10H15)15 (x = 5.76), and Au23-xAgx(SC10H15)15 (x = 7.44) (CIF) Author information Corresponding author *[email protected] Acknowledgement R.J. acknowledges the financial support by the Air Force Office of Scientific Research under AFOSR award no. FA9550-15-1-9999 (FA9550-15-1-0154) and 2016 Defense University Research Instrumentation Program (DURIP). Reference (1)

Jin, R.; Zeng, C.; Zhou, M.; Chen, Y. Chem. Rev. 2016, 116, 10346–10413. 12 ACS Paragon Plus Environment

Page 12 of 16

Page 13 of 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

(2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) (23) (24) (25) (26) (27) (28) (29) (30)

Weerawardene, K. L. D. M.; Häkkinen, H.; Aikens, C. M. Ann. Rev. Phys. Chem. 2018, 69, 205–229. Xu, W. W.; Zhu, B.; Zeng, X. C.; Gao, Y. Nat. Commun. 2016, 7, 13574. Zhou, M.; Zeng, C.; Song, Y.; Padelford, J. W.; Wang, G.; Sfeir, M. Y.; Higaki, T.; Jin, R. Angew. Chem. Int. Ed. 2017, 56, 16257–16261. Weerawardene, K. L. D. M.; Aikens, C. M. J. Am. Chem. Soc. 2016, 138, 11202–11210. Higaki, T.; Zhou, M.; Lambright, K. J.; Kirschbaum, K.; Sfeir, M. Y.; Jin, R. J. Am. Chem. Soc. 2018, 140, 5691–5695. Zeng, C.; Chen, Y.; Iida, K.; Nobusada, K.; Kirschbaum, K.; Lambright, K. J.; Jin, R. J. Am. Chem. Soc. 2016, 138, 3950–3953. Dong, H.; Liao, L.; Zhuang, S.; Yao, C.; Chen, J.; Tian, S.; Zhu, M.; Liu, X.; Li, L.; Wu, Z. Nanoscale 2017, 9, 3742–3746. Briant, C. E.; Theobald, B. R.; White, J. W.; Bell, L. K.; Mingos, D. M. P.; Welch, A. J. J. Chem. Soc. Chem. Commun. 1981, 0, 201–202. Shichibu, Y.; Konishi, K. Small 2010, 11, 1216–1220. Shichibu, Y.; Negishi, Y.; Watanabe, T.; Chaki, N. K.; Kawaguchi, H.; Tsukuda, T. J. Phys. Chem. Lett. 2007, 111, 7845–7847. Jin, R.; Liu, C.; Zhao, S.; Das, A.; Xing, H.; Gayathri, C.; Xing, Y.; Rosi, N. L.; Gil, R. R.; Jin, R. ACS Nano 2015, 9, 8530–8536. Xu, W. W.; Li, Y.; Gao, Y.; Zeng, X. C. Nanoscale 2016, 8, 7396-7401. Zhou, M.; Zeng, C.; Sfeir, M. Y.; Cotlet, M.; Iida, K.; Nobusada, K.; Jin, R. J. Phys. Chem. Lett. 2017, 8, 4023–4030. Zhou, M.; Jin, R.; Sfeir, M. Y.; Chen, Y.; Song, Y.; Jin, R. Proc. Natl. Acad. Sci. 2017, 114, E4697E4705. Wang, Y.; Su, H.; Ren, L.; Malola, S.; Lin, S.; Teo, B. K.; Häkkinen, H.; Zheng, N. Angew. Chem. Int. Ed. 2016, 55, 15152–15156. Yang, H.; Wang, Y.; Huang, H.; Gell, L.; Lehtovaara, L.; Malola, S.; Häkkinen, H.; Zheng, N. Nat. Commun. 2013, 4, 2422. Hikosou, D.; Saita, S.; Miyata, S.; Miyaji, H.; Furuike, T.; Tamura, H.; Kawasaki, H. J. Phys. Chem. C 2018, 122, 12494–12501. Wan, X.-K.; Cheng, X.-L.; Tang, Q.; Han, Y.-Z.; Hu, G.; Jiang, D.; Wang, Q.-M. J. Am. Chem. Soc. 2017, 139, 9451–9454. Wang, Y.; Wan, X.-K.; Ren, L.; Su, H.; Li, G.; Malola, S.; Lin, S.; Tang, Z.; Häkkinen, H.; Teo, B. K.; et al. J. Am. Chem. Soc. 2016, 138, 3278–3281. Wang, S.; Jin, S.; Yang, S.; Chen, S.; Song, Y.; Zhang, J.; Zhu, M. Sci. Adv. 2015, 1, e1500441. Zeng, J.-L.; Guan, Z.-J.; Du, Y.; Nan, Z.-A.; Lin, Y.-M.; Wang, Q.-M. J. Am. Chem. Soc. 2016, 138, 7848–7851. Liao, L.; Zhou, S.; Dai, Y.; Liu, L.; Yao, C.; Fu, C.; Yang, J.; Wu, Z. J. Am. Chem. Soc. 2015, 137, 9511– 9514. Bootharaju, M. S.; Joshi, C. P.; Parida, M. R.; Mohammed, O. F.; Bakr, O. M. Angew. Chem.-Int. Ed. 2016, 55, 922–926. Qian, H.; Jiang, D.; Li, G.; Gayathri, C.; Das, A.; Gil, R. R.; Jin, R. J. Am. Chem. Soc. 2012, 134, 16159–16162. Negishi, Y.; Kurashige, W.; Kobayashi, T.; Yamazoe, S.; Kojima, N.; Seto, M.; Tsukuda, T. J. Phys. Chem. Lett. 2013, 4, 3579–3583. Annelies, S.; Noelia, B.; Knoppe, S.; Bürgi, T. Nanoscale 2016, 8, 11130–11135. Zhang, B.; Kaziz, S.; Li, H.; Wodka, D.; Malola, S.; Safonova, O.; Nachtegaal, M.; Mazet, C.; Dolamic, I.; Llorca, J.; Kalenius, E.; Daku, L. M. L.; Häkkinen, H.; Burgi, T.; Barrabes, N. Nanoscale 2015, 7, 17012–17019. Negishi, Y.; Iwai, T.; Ide, M. Chem. Commun. 2010, 46, 4713–4715. Kauffman, D. R.; Alfonso, D.; Matranga, C.; Qian, H.; Jin, R. J. Phys. Chem. C 2013, 117, 7914–7923. 13 ACS Paragon Plus Environment

Journal of the American Chemical Society 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(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) (61) (62)

Li, Q.; Wang, S.; Kirschbaum, K.; Lambright, K. J.; Das, A.; Jin, R. Chem. Commun. 2016, 52, 5194– 5197. Krishnadas, K. R.; Baksi, A.; Ghosh, A.; Natarajan, G.; Pradeep, T. ACS Nano 2017, 11, 6015–6023. Teo, B. K. Polyhedron 1988, 7, 2317–2320. Xu, W. W.; Zeng, X. C.; Gao, Y. Chem. Phys. Lett. 2017, 675, 35–39. Copley, R. C. B.; Mingos, D. M. P. J. Chem. Soc. Dalton Trans. 1996, 491–500. Teo, B. K.; Zhang, H.; Shi, X. Inorg. Chem. 1990, 29, 2083–2091. Teo, B. K.; Zhang, H.; Shi, X. J. Am. Chem. Soc. 1990, 112, 8552–8562. Teo, B. K.; Hong, M. C.; Zhang, H.; Huang, D. B. Angew. Chem. Int. Ed. 1987, 9, 897–900. Yang, H.; Wang, Y.; Yan, J.; Chen, X.; Zhang, X.; Häkkinen, H.; Zheng, N. J. Am. Chem. Soc. 2014, 136, 7197–7200. Desireddy, A.; Conn, B. E.; Guo, J.; Yoon, B.; Barnett, R. N.; Monahan, B. M.; Kirschbaum, K.; Griffith, W. P.; Whettern, R. L.; Landman, U.; Bigioni, T. P. Nature 2013, 501, 399–402. Li, Q.; Luo, T.-Y.; Taylor, M. G.; Wang, S.; Zhu, X.; Song, Y.; Mpourmpakis, G.; Rosi, N. L.; Jin, R. Sci. Adv. 2017, 3, e1603193. Yang, H.; Wang, Y.; Lei, J.; Shi, L.; Wu, X.; Mäkinen, V.; Lin, S.; Tang, Z.; He, J.; Häkkinen, H.; Zheng, L.; Zheng. N. J. Am. Chem. Soc. 2013, 135, 9568–9571. Ouyang, R.; Jiang, D.-e. ACS Catal. 2015, 5, 6624–6629. Tang, Q.; Jiang, D.-e. J. Phys. Chem. C 2015, 119, 2904–2909. Qian, H.; Zhu, Y.; Jin, R. Proc. Natl. Acad. Sci. 2012, 109, 696–700. Yu, Y.; Luo, Z.; Chevrier, D. M.; Leong, D. T.; Zhang, P.; Jiang, D.-e.; Xie, J. J. Am. Chem. Soc. 2014, 136, 1246–1249. Song, Y.; Abroshan, H.; Chai, J.; Kang, X.; Kim, H. J.; Zhu, M.; Jin, R. Chem. Mater. 2017, 29, 3055– 3061. Yao, Q.; Fung, V.; Sun, C.; Huang, S.; Chen, T.; Jiang, D.; Lee, J. Y.; Xie, J. Nat. Commun. 2018, 9, 1979. Conn, B. E.; Atnagulov, A.; Yoon, B.; Barnett, R. N.; Landman, U.; Bigioni, T. P. Sci. Adv. 2016, 2, e1601609. Chen, T.; Yang, S.; Chai, J.; Song, Y.; Fan, J.; Rao, B.; Sheng, H.; Yu, H.; Zhu, M. Sci. Adv. 2017, No. 3:e1700956. Qian, H.; Eckenhoff, W. T.; Zhu, Y.; Pintauer, T.; Jin, R. J. Am. Chem. Soc. 2010, 132, 8280–8281. Chen, S.; Xiong, L.; Wang, S.; Ma, Z.; Jin, S.; Sheng, H.; Pei, Y.; Zhu, M. J. Am. Chem. Soc. 2016, 138, 10754–10757. Jones, T. C.; Sementa, L.; Stener, M.; Gagnon, K. J.; Thanthirige, V. D.; Ramakrishna, G.; Fortunelli, A.; Dass, A. J. Phys. Chem. C 2017, 121, 10865–10869. Yang, S.; Chai, J.; Song, Y.; Fan, J.; Chen, T.; Wang, S.; Yu, H.; Li, X.; Zhu, M. J. Am. Chem. Soc. 2017, 139, 5668–5671. Das, A.; Li, T.; Nobusada, K.; Zeng, C.; Rosi, N. L.; Jin, R. J. Am. Chem. Soc. 2013, 135, 18264–18267. Zhu, M.; Wang, P.; Yan, N.; Chai, X.; He, L.; Zhao, Y.; Xia, N.; Yao, C.; Li, J.; Deng, H.; Zhu, Y.; Pei, Y. Wu, Z. Angew. Chem. Int. Ed. 2018, 57, 4500–4504. Liu, L.; Song, Y.; Chong, H.; Yang, S.; Xiang, J.; Jin, S.; Kang, X.; Zhang, J.; Yu, H.; Zhu, M. Nanoscale 2016, 8, 1407–1412. Xiang, J.; Li, P.; Song, Y.; Liu, X.; Chong, H.; Jin, S.; Pei, Y.; Yuan, X.; Zhu, M. Nanoscale 2015, 7, 18278–18283. Chen, S.; Wang, S.; Zhong, J.; Song, Y.; Zhang, J.; Sheng, H.; Pei, Y.; Zhu, M. Angew. Chem. Int. Ed. 2015, 127, 3188–3192. Kang, X.; Xiong, L.; Wang, S.; Pei, Y.; Zhu, M. Inorg. Chem. 2018, 57, 335-342. De Nardi, M.; Antonello, S.; Jiang, D.-e.; Pan, F. F.; Rissanen, K.; Ruzzi, M.; Venzo, A.; Zoleo, A.; Maran, F. ACS Nano 2014, 8, 8505–8512. Chen, C.-L.; Zhang, P.; Rosi, N. L. J. Am. Chem. Soc. 2008, 130, 13556–13557. 14 ACS Paragon Plus Environment

Page 14 of 16

Page 15 of 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

(63) (64) (65) (66) (67) (68) (69) (70) (71) (72) (73) (74) (75) (76) (77) (78) (79) (80)

Pinkard, A.; Champsaur, A. M.; Roy, X. Acc. Chem. Res. 2018, 51, 919–929. Zeng, C.; Chen, Y.; Kirschbaum, K.; Appavoo, K.; Sfeir, M. Y.; Jin, R. Sci. Adv. 2015, 1, e1500045. Liu, X.; Chen, J.; Yuan, J.; Li, Y.; Li, J.; Zhou, S.; Yao, C.; Liao, L.; Zhuang, S.; Zhao, Y.; Deng, H. Yang, J.; Wu, Z. Angew. Chem. 2018, 130, 11443-11447. Zeng, C.; Liu, C.; Chen, Y.; Rosi, N. L.; Jin, R. J. Am. Chem. Soc. 2016, 138, 8710–8713. Higaki, T.; Liu, C.; Zhou, M.; Luo, T.-Y.; Rosi, N. L.; Jin, R. J. Am. Chem. Soc. 2017, 139, 9994–10001. Zeng, C.; Chen, Y.; Kirschbaum, K.; Lambright, K. J.; Jin, R. Science 2016, 354, 1580–1584. Crasto, D.; Barcaro, G.; Stener, M.; Sementa, L.; Fortunelli, A.; Dass, A. J. Am. Chem. Soc. 2014, 136, 14933–14940. Higaki, T.; Liu, C.; Zeng, C.; Jin, R.; Chen, Y.; Rosi, N. L.; Jin, R. Angew. Chem. Int. Ed. 2016, 55, 6694–6697. Lei, Z.; Li, J.-J.; Wan, X.-K.; Zhang, W.-H.; Wang, Q.-M. Angew. Chem. Int. Ed. 2018, 57, 8639-8643. Chen, W.; Chen, S. Angew. Chem. Int. Ed. 2009, 48, 4386–4389. Zhang, J.; Li, Z.; Huang, J.; Liu, C.; Hong, F.; Zheng, K.; Li, G. Nanoscale 2017, 9, 16879–16886. Stamplecoskie, K. G.; Kamat, P. V. J. Am. Chem. Soc. 2014, 136, 11093–11099. Chen, Y.; Liu, C.; Tang, Q.; Zeng, C.; Higaki, T.; Das, A.; Jiang, D.; Rosi, N. L.; Jin, R. J. Am. Chem. Soc. 2016, 138, 1482–1485. Zeng, C.; Qian, H.; Li, T.; Li, G.; Rosi, N. L.; Yoon, B.; Barnett, R. N.; Whettern, R. L.; Landman, U.; Jin, R. Angew. Chem. Int. Ed. 2012, 51, 13114–13118. Nimmala, P. R.; Knoppe, S.; Jupally, V. R.; Delcamp, J. H.; Aikens, C. M.; Dass, A. J. Phys. Chem. B 2014, 118, 14157–14167. Devadas, M. S.; Bairu, S.; Qian, H.; Sinn, E.; Jin, R.; Ramakrishna, G. J. Phys. Chem. Lett. 2011, 2, 2752–2758. Kumara, C.; Aikens, C. M.; Dass, A. J. Phys. Chem. Lett. 2014, 5, 461–466. Jin, R.; Zhao, S.; Liu, C.; Zhou, M.; Panapitiya, G.; Xing, Y.; Rosi, N. L.; Lewis, J. P.; Jin, R. Nanoscale 2017, 9, 19183–19190.

15 ACS Paragon Plus Environment

Journal of the American Chemical Society 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

16 ACS Paragon Plus Environment

Page 16 of 16