Nuclear and Electron Magnetic Resonance Spectroscopies of

Nov 27, 2018 - Department of Chemistry, University of Padova , via Marzolo 1, ... Atomically precise gold nanoclusters display properties that are uns...
1 downloads 0 Views 3MB Size
Download PDF
Article Cite This: Acc. Chem. Res. 2019, 52, 44−52

pubs.acs.org/accounts

Nuclear and Electron Magnetic Resonance Spectroscopies of Atomically Precise Gold Nanoclusters Published as part of the Accounts of Chemical Research special issue “Toward Atomic Precision in Nanoscience”. Mikhail Agrachev,† Marco Ruzzi,† Alfonso Venzo,‡ and Flavio Maran*,†,§ †

Department of Chemistry, University of Padova, via Marzolo 1, 35131 Padova, Italy National Research Council, CNR-ICMATE, Department of Chemistry, University of Padova, via Marzolo 1, 35131 Padova, Italy § Department of Chemistry, University of Connecticut, 55 North Eagleville Road, Storrs, Connecticut 06269, United States Acc. Chem. Res. 2019.52:44-52. Downloaded from pubs.acs.org by IOWA STATE UNIV on 01/16/19. For personal use only.

‡

CONSPECTUS: Atomically precise gold nanoclusters display properties that are unseen in larger nanoparticles. When the number of gold atoms is sufficiently small, the clusters exhibit molecular properties. Their study requires extensive use of classic molecular physical chemistry and, thus, methods such as vibrational spectroscopies, electrochemistry, density functional theory and molecular dynamics calculations, and of course nuclear magnetic resonance (NMR) and electron paramagnetic resonance (EPR) spectroscopies. NMR and EPR studies have been mostly carried out on the benchmark, stable molecules Au25(SR)18, Au38(SR)24, Au102(SR)44, and Au144(SR)60 (where SR = thiolate). In this Account, we showcase examples primarily taken from our previous and ongoing NMR and EPR studies, which we hope will trigger further interest in the use of these sensitive, though often underutilized, techniques. Indeed, 1D and 2D NMR spectra of pure, atomically precise clusters can be very detailed and informative. Molecular clusters are molecules and, thus, have discrete energy levels and undergo stepwise oxidation or reduction. The effect of the charge state on the chemical shifts and line shapes is a function of the ligand type (ligands differ due to specific bonds with different Au atom types) and the position of the chemical group along the ligand backbone: for groups near the Au core, they can be very dramatic. Ligand-protected gold clusters are hard−soft molecules where a hard metal core is surrounded by a dynamic molecular layer. The latter provides a nanoenvironment that interfaces the cluster core with the surrounding environment and can be permeated by molecules and ions. NMR spectroscopy is especially useful to assess its structure. For example, the data show that whereas long alkanethiolates form bundles, shorter chains exhibit more conformational freedom and are quite folded. NMR spectroscopy allows studying diastereotopic effects and provides information on possible hydrogen bonds of ligands with sulfur or surface gold atoms. EPR spectroscopy is a very precise technique to check and characterize the magnetic state of gold clusters or clusters doped with foreign-metal atoms. Electron nuclear double resonance (ENDOR) provides a powerful tool to assess the interaction of an unpaired electron with nuclei, as we showed for 197Au and 1H. It can be used as a sensitive probe of the spin-density distribution in nanoclusters: for example, it showed that the singly occupied molecular orbital may span outside the Au core by nearly 6 Å. Solid-state EPR spectroscopy has provided compelling evidence that the specific ligands and the crystallinity degree are very important factors in determining the interactions between clusters in the solid state. Depending on the condition, paramagnetic, superparamagnetic, ferromagnetic, or antiferromagnetic behavior can be observed. Time-resolved EPR was successfully tested to determine the efficiency of singlet-oxygen generation via sensitization of Au25 clusters. This Account thus demonstrates some of the remarkable insights that can be gained into the properties of atomically precise clusters through detailed NMR and EPR studies.

1. INTRODUCTION

tigations. This interplay enabled understanding of how properties depend on structural as well as physical factors, such as the cluster’s charge and magnetic state. Quantum confinement is responsible for making atomically precise nanoclusters display properties unseen in larger gold nanoparticle systems. When the number of gold atoms is very small (roughly, 130 atoms or less), these clusters start exhibiting

Ligand-protected gold nanoclusters can often be prepared with atomic precision when the core diameter is smaller than 2 nm.1,2 This achievement has been made possible by refining synthetic, purification, and characterization procedures, and gathering composition and structural information by mass spectroscopy and single-crystal X-ray crystallography. This knowledge helped in interpreting the ever-growing body of physicochemical data gathered on these systems. In turn, the discovery of novel properties triggered further synthetic and structural inves© 2018 American Chemical Society

Received: September 29, 2018 Published: November 27, 2018 44

DOI: 10.1021/acs.accounts.8b00495 Acc. Chem. Res. 2019, 52, 44−52

Article

Accounts of Chemical Research molecular properties, such as a clearly detectable HOMO− LUMO (highest occupied and lowest unoccupied molecular orbitals) gap2,3 and molecular photophysical properties.1,2 In this Account, we will focus on the nuclear magnetic resonance (NMR) and electron paramagnetic resonance (EPR) behaviors of atomically precise thiolate (SR) protected gold clusters. These clusters are often referred to as monolayer-protected clusters, but these monolayers are not simple monolayers. Since the first report of the structure of a thiolated nanocluster, Au102(SR)44 (RSH = para-mercaptobenzoic acid, p-MBA),4 single crystal X-ray crystallography consistently showed that thiolate-protected Au nanoclusters can display different kinds of Au−(SR) motifs, which form the actual capping monolayer. Thiolates may bind to the Au core as bridges −(SR)−, staples −(SR)−Au−(SR)−, double staples −(SR)−Au−(SR)−Au− (SR)−, and triple staples −(SR)−Au−(SR)−Au−(SR)−Au− (SR)−, or a combination of them.1 Here we focus especially on the benchmark cases of Au 25 (SR) 18 , Au 38 (SR) 24 , and Au144(SR)60. The NMR and EPR topics have been selected to showcase the potentialities of these methods and stress the enormous advantages of using atomically precise clusters.

2. NUCLEAR MAGNETIC RESONANCE 2.1. NMR Behavior

The most studied molecular cluster is Au25(SR)18.5 In its native form, Au25(SR)18 is a diamagnetic anion ion-paired to the ntetraoctylammonium cation (Oct4N+) used, e.g., as phasetransfer agent in the seminal Brust-Schiffrin synthesis.6 Its structure was first described7,8 for Au25(SC2Ph)18− (SC2Ph = phenylethanethiolate; hereafter, we will indicate the number of carbon atoms of the alkyl chain simply as Cn). Its oxidation generates the corresponding stable neutral cluster, Au25(SR)180, whose structure was also assessed by single-crystal X-ray crystallography, for various R groups.9−13 Further oxidation yields the cation Au25(SR)18+, whose structure was also solved.14 Independently of its charge, these clusters share the same general structural features: a central Au atom is surrounded by an icosahedron of 12 Au atoms, and the resulting Au13 core is capped by six −(SR)in−Au−(SR)out−Au−(SR)in− double staples. Each staple consists of two inner thiolates (in) and one outer thiolate (out). The term inner stresses that the two terminal SR groups also bind to icosahedron Au atoms (the inner side of the cluster), whereas outer refers to the outmost thiolate of the double staple (Figure 1a). Au25(SR)18− exhibits 1H and 13C NMR spectra consisting of clearly distinguishable resonances for both ligand types and in exactly a 12 (in) to 6 (out) ratio.15,16 Here we will refer mostly to results obtained for clusters with hydrophobic ligands in C6D6 (unless otherwise stated), although similar results were also obtained for the water-soluble glutathione-capped Au25(SR)18 cluster.17 In organic solvents, controlled oxidation of hydrophobic clusters can be achieved by using the powerful bis(pentafluorobenzoyl) peroxide oxidant.18 In the presence of Au25, this peroxide undergoes an overall two-electron reduction consisting of an initial concerted dissociative electron-transfer cleavage of the O−O bond to form the carboxylate and the benzoyloxy radical, followed by diffusion controlled reduction of the latter to form a second carboxylate anion. This peroxide has distinct advantages in studies on the magnetic properties of Au clusters: no hydrogen atoms; only diamagnetic species are generated; its reduction is fully irreversible and controllable stoichiometrically. We used this strategy to make

Figure 1. (a) Structure of Au25(SC3)180, where five staples have been removed for clarity. (b) 1H NMR spectra of 1 mM Au25(SC3)180 and Au25(SC3)18− (the asterisks mark the Oct4N+ countercation) in C6D6 at 25 °C; the (α-CH)in signal (at 70 °C) is offset and enlarged. (c) Δδ values against distance (see text): C2 (green), C3 (red), C4 (blue); (αCH)in, ◆; (β-CH)in, ⬢; (α-CH)out, ▲; (γ-CH)in, ▼; (β-CH)out, ●. Vertical dashed lines group the protons at similar distances. Reproduced with permission from ref 12. Copyright 2016 The Royal Society of Chemistry.

Au25(SC2Ph)180 and the Au25(SC2Ph)18+ cation.16 Stepwise oxidation of Au25(SC2Ph)18− with 2,2,6,6-tetramethylpiperidin1-oxoammonium cation (TEMPO+) as the electron acceptor and the corresponding NMR analysis were later described.19 Oxidation of Au25(SC2Ph)18− to form Au25(SC2Ph)180 generates a species exhibiting very different 1H and 13C NMR chemical shifts.15,16 The typical effect of forming the paramagnetic cluster on the proton spectrum is illustrated in Figure 1b for Au25(SC3)18 in C6D6. Notably, a highly deshielded resonance is observed for α-(CH2)in.16 This resonance is barely observable at 25 °C, and indeed some investigations missed it.15,19 At higher temperatures, however, it is a fully detectable peak (Figure 1b): e.g., for linear alkanethiolates (SCn: n = 4− 18), δ = 22.5 ± 0.2 ppm at 65 °C.20 The solvent does not affect the spectra, but for limited chemical-shift differences.16 For 13C, the α-(C)in and β-(C)in resonances are found at −223 and +350 ppm and are very broad (Au25(SC2Ph)180, 85 °C, toluene-d8).16 This behavior is caused by the paramagnetism of Au25(SR)180 and is related to the contact interaction between the nuclear and unpaired-electron magnetic moments. That Au25(SR)180 is a paramagnet was already established via EPR measurements,21 as discussed later. 45

DOI: 10.1021/acs.accounts.8b00495 Acc. Chem. Res. 2019, 52, 44−52

Article

Accounts of Chemical Research Upon stoichiometric addition of bis(pentafluorobenzoyl) peroxide to a CD2Cl2 solution of Au25(SC2Ph)180, we obtained a pure sample of [Au25(SC2Ph)18+][C6F5CO2−].16 The 1H and 13 C NMR spectra evidence the disappearance of the peaks of Au25(SC2Ph)180 and the appearance of all (CH2)in and (CH2)out peaks with chemical shifts similar to those of Au25(SC2Ph)18−. This study established, for the first time, that the Au25(SC2Ph)18+ cation is diamagnetic, which was later confirmed by EPR.22 This is not trivial: according to the superatom model,23 Au25(SR)18− is an 8-electron system characterized by a 1S21P6 superatom electron configuration, which implies the presence of 3-fold degenerate HOMOs and, therefore, the possible formation of a triplet state upon twoelectron oxidation of Au25(SR)18− to form Au25(SR)18+. For Au25(SCH3)18z (z = −1, 0, + 1), however, we noticed (density functional theory, DFT, calculations) that the commonly assumed triple degeneracy is indeed not strictly applicable, especially upon cluster oxidation.22 For the zero and +1 charge states, a progressive increase of the gap was calculated, in good agreement with the experimental HOMO−LUMO gap measured from the UV−vis absorption data and the NMR evidence about the diamagnetism of Au25(SC2Ph)18+.16,18 The electrode kinetics associated with the Au25(SR)180/Au25(SR)18− and Au25(SR)18+/Au25(SR)180 redox couples clearly showed that the formation of the cation is affected by a larger inner reorganization energy (deformation of bond angles and lengths) than for the anion.20,22,24 This conclusion is in keeping with the structural differences between the three charge states of Au25(SC2Ph)18.7−9,13,14 Regarding quantitative analysis of the transition from the −1 to the 0 charge state, the NMR investigation was extended to other Au25(SCn)18 systems (SCn: n = 2, 3, and 4) and Au25(SMePr)18 (SMePr = 2-methyl-1-propanethiolate), for which all inner and outer resonances could be assigned by analysis of homo- and heteronuclear 2D correlation spectra.12 These specific Au25(SCn)18 clusters were selected because of the knowledge of their crystallographic structures.10−12 Oneelectron oxidation of Au25(SCn)18− shifts downfield the NMR peaks corresponding to the α, β, and γ protons of the (CH2)in groups and the α and γ protons of the (CH2)out groups; the β(CH2)out peak undergoes an upfield shift (Figure 1b). Figure 1c shows that the positive chemical-shift difference (Δδ = δradical − δanion) correlates nearly exponentially with the average distance between the central Au atom and the hydrogen atoms along the thiolate chain. These Δδ values are a measure of how far the spin density spreads outside the Au13 core,16 and we used them to estimate the corresponding isotropic hyperfine coupling constant A.12

Figure 2. (a) 1H NMR spectra of 1 mM Au25(SCn)18 in C6D6 at 25 °C (* = dichloromethane). (b) Snapshots of the MD trajectories of approximately spherical C10 and elongated C18 clusters. Adapted with permission from ref 20. Copyright 2014 American Chemical Society.

trend, which is in keeping with the outcome of associated electron transfer, IR absorption spectroscopy, and molecular dynamics (MD) analyses, is caused by intracluster self-assembly of the longer ligands to form bundles on either side of the core (Figure 2b). This makes the proton signals virtually unaffected by further elongation of the chain, as this does not affect significantly their specific chemical environment. Instead, shorter chains exhibit more conformational freedom and thus solvent-molecule penetration in between the chains can be more extensive, especially as the ligand gets shorter. For these approximately spherical clusters (Figure 2b), the upfield shifts are thus caused by benzene itself (used as the NMR solvent), which is known to affect resonances in this specific way. One interesting but also challenging problem is to relate the structure in solution, as inferred from NMR data with the support of theoretical calculations, to that observed in the crystalline state by X-ray crystallography. For example, we obtained evidence that in the crystalline state dispersion interactions between ligands of nearby Au25(SCn)180 clusters (n = 4, 5) assist these clusters to form polymers,11,13 whereas the same clusters in solution are pretty much spherical.20 A detailed analysis was carried out on Au102(p-MBA)44,25 whose structure is known and shows a ligand layer composed of 22 different ligand types.4 Using a combination of NMR techniques, DFT calculations, and MD simulations, the 1H and 13C NMR chemical shifts could be assigned to all 22 ligand types, thereby

2.2. Dynamic Monolayer

Assessing the structure of the thiolate monolayer in solution is an important topic, especially for implications in redox catalysts. We prepared an extensive series of Au25(SCn)180 clusters (SCn: n = 2−18) and studied their NMR behavior.20 It is here worth stressing that for 1H NMR spectroscopy the usefulness of focusing on the paramagnetic clusters (rather than the diamagnetic anions) is that all signals are conveniently spaced along the chemical-shift scale. For each cluster, the resonances exhibit δ values determined by the specific proton (α, β, γ, etc.) and ligand type (inner or outer). Figure 2a shows the interesting phenomenon that the δ values of corresponding resonances are pretty much constant for the longer ligands (n = 10 up to 18), whereas they consistently shift upfield for shorter ligands. This 46

DOI: 10.1021/acs.accounts.8b00495 Acc. Chem. Res. 2019, 52, 44−52

Article

Accounts of Chemical Research providing an interesting perspective on how to use NMR and computational data to assess the structure of clusters whose solid-state X-ray structure is still unknown. 2.3. Diastereotopic and Surface Effects

A seminal NMR study was reported about another molecular cluster, Au38(SC2Ph)24.26 This cluster is chiral. Its structure shows an achiral Au23 kernel capped by three monomeric staples and six dimeric staples; the latter are arranged into two groups at the two ends of the rodlike Au23 core.27 Chirality is dictated by the rotary arrangement of these groups of staples. NMR spectroscopy analysis reveals the presence of four ligand types. Most of the methylene resonances show different chemical shifts for the geminal protons. This phenomenon is known as diastereotopicity effect, and for gold clusters was previously observed with chiral ligands (L-gluthathione) protecting the achiral Au25 cluster.17 The diastereotopic chemical-shift difference (Δδd) maximizes when the methylene is near the S atom. The α-CH2 and the β-CH2 resonances were found to exhibit Δδd values as large as 0.76 and 0.17 ppm, respectively.26 Very recently, we described the spontaneous fusion of Au25(SCn)180 clusters (n = 3, 4) to form the corresponding Au38(SCn)24 cluster.28 According to the NMR data, these clusters are protected by four ligand types in a 1:1:1:1 ratio, as for Au38(SC2Ph)24.26 Besides highlighting the capabilities of NMR spectroscopy in assessing structural features in solution (i.e., that Au38(SCn)24 and Au38(SC2Ph)24 have the same structure), this study shows that the diastereotopic effects are observed up to the methylene at the γ position: α-CH2, Δδd of 0.67 (C3) and 0.69 ppm (C4); β-CH2, Δδd of 0.023 (C3) and 0.085 ppm (C4); γ-CH2 Δδd of 0.029 ppm (C4).28 No effect was detected for the resonances assigned to the outermost ligand type. Diastereotopic effects may originate from any mechanism that makes two otherwise identical hydrogen atoms different in terms of chemical environment. Very recently, we carried out an NMR study on the atomically precise Au144(SR)60 nanocluster.29 The purity and stoichiometry of the samples, Au144(SCn)60 (n = 2, 3), were assessed by a combination of optical spectroscopy, electrochemistry, and mass-spectrometry results. 1D and 2D NMR analysis allowed identification of 12 distinct ligand types in 5-fold equivalence (12 × 5 = 60), with α-CH2 resonances spanning from 3.2 to 7.9 ppm. For each ligand type, the geminal protons exhibit diastereotopic effects with values that for the αCH2 protons can be defined as gigantic, as Δδd can be as large as 2.9 ppm (Figure 3). Meanwhile, the structure of the similar Au144(SC1Ph)60 cluster could be solved and showed that this cluster is chiral as a consequence of how the 30 single staples −(SR)−Au−(SR)− distribute on the core surface.30 The Δδd values observed for both Au144(SCn)60 clusters, however, are too large to be attributable to chirality. DFT calculations indeed suggest that the presence of 12 ligand types and the corresponding diastereotopic effects could be primarily related to the presence of close C−H···S contacts causing differences in the geminal αCH2 protons. Interestingly, the NMR evidence for 12 ligand types is in contrast to the previous NMR finding in D2O of the presence of only one ligand type in a similar cluster described as Au144(p-MBA)60.31 This suggests that the actual monolayer structure in solution is a function of the specific capping ligand and different interligand interactions. Possible interactions with the surface Au atoms could contribute to make the picture even more complicated.

Figure 3. 1H,1H-homonuclear correlation spectroscopy (COSY) spectrum of 0.5 mM Au144(SC2)60 in C6D6 at 45 °C. Correlations are evidenced and the Δδd values are indicated. Reproduced with permission from ref 29. Copyright 2018 The Royal Society of Chemistry.

Evidence for other interactions of the capping ligands with the cluster surface have indeed been proposed. The complex NMR results obtained in D2O for clusters described as Au68(SR)32 and Au144(m-MBA)∼40 (m-MBA = meta-mercaptobenzoic acid), together with computational and other spectroscopic characterization, suggest the presence of different ligand types.32 The COOH groups at the meta position appear to be responsible for weak interactions between ligands and ligand−Au interactions, and possibly for the small number of capping ligands in this Au144 cluster (∼40 vs the value of 60 consistently observed for all other Au144 clusters). These interactions are seen as also involving C−OH···Au and π−Au contacts. The possibility of protons interacting with gold is an aspect that has been discussed for gold complexes.33 A very interesting case of particularly possible C−H···Au interaction has been recently reported for a diphosphine-ligated divalent Au6 cluster.34 In this small cluster, the Au core is stabilized by four m-phenylene-bridged diphosphines staples forming P−Au bonds. A comparison between its 1H and 13C NMR spectra and those of a reference complex, in which the Au6 core is missing, shows that one of the C−H groups undergoes the sizable downfield shifts of 4.37 and >10 ppm, respectively. Analysis of its structure points to a particularly short C−H···Au distance (2.62 Å) and, therefore, to possible hydrogen bonding to the Au atoms. A very similar cluster, in which the aromatic ring responsible for the above interaction was substituted with an alkyl chain, shows similar though less pronounced effects (for H, 1.13 ppm), in agreement with a now longer C−H···Au distance (2.78 Å).

3. ELECTRON PARAMAGNETIC RESONANCE 3.1. EPR Behavior

Au25(SR)180 is a paramagnet that exhibits a characteristic continuous wave EPR behavior (CW-EPR), as originally observed for R = C2Ph.21 We tested the possible formation of a triplet state for Au25(SR)18+.22 Figure 4 compares the CW-EPR 47

DOI: 10.1021/acs.accounts.8b00495 Acc. Chem. Res. 2019, 52, 44−52

Article

Accounts of Chemical Research

applied to invert a specific NMR transition, eliminating the population difference created by the first pulse. A two-pulse spin−echo sequence ensues, which allows detection of the magnetization. We performed the first ENDOR experiments with Au25(SC2)180, and focused on the interaction between the unpaired electron and 197Au.10 The hyperfine and quadrupolar 197 Au constants, which are related to spin density on the 197Au nuclei, were obtained from simulation of the ENDOR spectrum. DFT calculations, based on the actual X-ray crystallography structure of Au25(SC2)180, were in good agreement with the experimental results. Interestingly, the outcome of the DFT calculations was found to depend on the actual thiolate (SC1 vs SC2), thereby stressing the risks of oversimplifying the ligand structure in this type of calculations. As a result of this ENDORstructural-DFT analysis, we found that the 25 Au atoms belong to four equivalent subgroups: the Au central atom, a group of 8 strongly coupled core atoms, a group of 4 weakly coupled core atoms, and the staple Au atoms. A second relevant aspect is the sensitivity of ENDOR to very low electron spin densities. 1H ENDOR was used to probe the spin distribution on the protons of the thiolated ligands,12 and therefore, the spreading of the singly occupied molecular orbital (SOMO) outside the Au core. This is an interesting issue, as the spin density is generally considered localized at the Au13 core.5 We focused on four Au25(SR)180 clusters with ligands of different length and hindrance: Au 25 (SC2) 18 0 , Au 25 (SC3) 18 0 , Au25(SC4)180, and Au25(SMePr)180. The ENDOR results were also compared with the NMR data (Δδ values) and analyzed for the contribution of the two ligand types and the positions of the protons along the ligand chain. Figure 5 shows a typical

Figure 4. EPR spectra of 2 mM Au25(SC2Ph)180 before (red curve) and after oxidation (green curve), in CH2Cl2 at 6 K. The blue curve shows the simulation. The inset shows an enlarged view of the spectrum of the oxidized cluster and its simulation (for residual Au25(SC2Ph)180, see the text). Reproduced with permission from ref 22. Copyright 2013 American Chemical Society.

behaviors of Au25(SC2Ph)180 and Au25(SC2Ph)18+ in frozen CH2Cl2 solutions at 6 K. The experimental CW-EPR spectrum of Au25(SC2Ph)180 is closely reproduced by a spectral simulation, using the previously reported21 anisotropic g-tensor (g = 2.56, 2.36, 1.82) and anisotropic hyperfine coupling values (A = 71, 142, 50 MHz), with 13 197Au nuclei (spin 3/2). Au25(SC2Ph)18+ was obtained by oxidation of Au25(SC2Ph)180 with bis(pentafluorobenzoyl) peroxide. To obtain Figure 4, a slight defect of peroxide was used (94% conversion). The resulting EPR spectrum of oxidized Au25(SC2Ph)180 evidences a strong decrease in intensity (green trace). The only signal is due to remaining Au25(SC2Ph)180, whose spectrum could be successfully simulated (blue trace) using the same parameters of the original Au25(SC2Ph)180 sample. The same outcome is found at higher temperatures, although for T > 100 K the spectrum is very broad. Except for the decrease of the EPR intensity (Curie law) and line broadening (spin relaxation), increasing the temperature does not change the main spectral features, and no additional EPR signals are observable. This EPR study thus provided definite evidence that Au25(SR)18+ is diamagnetic and, together with the NMR result, that the energy difference between the first and the second HOMOs must be large enough to prevent formation of a triplet state at all temperatures investigated, in keeping with DFT calculations.22 CW-EPR was also used to test the paramagnetism of other Au25(SCn)180 clusters (n = 2, 3, 4, 6, and SR = MePr)10−12,35 in frozen solutions. We found that the anisotropic g-tensor and hyperfine coupling values are very little dependent on the ligand type and length; for example we used g = 2.56, 2.40, 1.78 for Au25(SC4)180.11 CW-EPR was also used to verify that, despite their neutral charge, foreign-metal doped Au24M(SR)180 clusters (e.g., M = Pd, Pt)35 exhibit no signal and thus behave like diamagnetic Au25(SR)18− rather than neutral Au25(SR)180.

Figure 5. 1H-ENDOR spectrum (blue) and simulation (red) for 0.5 mM Au25(SC4)180 in toluene at 5 K. The dashed lines mark the three main regions: β-(CH2)in (orange), α-(CH2)out (green), and β-(CH2)out (with some contribution from γ-(CH2)out) (black). α-(CH2)in is very broad and out of range, whereas γ-(CH2)in spreads roughly at 9−11 and 13−15 MHz. Adapted with permission from ref 12. Copyright 2016 The Royal Chemical Society.

3.2. ENDOR

spectrum and its simulation. We found that the unpaired electron provides a very precise probe of the structural features of the interface between the Au core and the ligands. Main results of this analysis are that (i) the spin-density distribution affects, in decreasing order, the α, β, and γ protons; (ii) spin density is larger for the inner than the outer ligands; (iii) orbital distribution affects atoms that can be as far as almost 6 Å from the icosahedral core; (iv) comparison of the coupling constants

The ENDOR (electron nuclear double resonance) technique is based on the electronic Overhauser effect. ENDOR and pulsed ENDOR are powerful tools to assess the interaction of an unpaired electron with nuclei, and can thus be used to probe the spin-density distribution in nanoclusters. A selective microwave pulse is applied to invert the electron-spin magnetization relative to a specific EPR transition, and then a radiofrequency pulse is 48

DOI: 10.1021/acs.accounts.8b00495 Acc. Chem. Res. 2019, 52, 44−52

Article

Accounts of Chemical Research A obtained from NMR and ENDOR measurements point to a possible increase of spin polarization at very low temperature. Another pulsed EPR technique particularly useful for the measurement of weak hyperfine couplings is electron spin echo envelope modulation (ESEEM). The pulse sequences are based on the modulation of the temporal decay of the electron spin echo due to weak hyperfine interactions between the investigated paramagnetic species and the magnetic nuclei nearby. Our preliminary results show that couplings with aromatic protons in Au25(SC2Ph)180 clusters can be indeed measured by ESEEM. This finding may pave the way to investigations on clusters with long ligands and/or large cores. 3.3. Solid-State EPR

The magnetic properties of Au nanoclusters in the solid state are much more complex than those in solution. Most studies focused on Au25(SC2Ph)180, relied on superconducting quantum interference device (SQUID) magnetometry data, and reached the conclusion that this cluster is paramagnetic,14,21,36 as it is in frozen solution.21,22 SQUID, however, detects the whole susceptibility, which may include contributions from many sources; disentangling the different components may be challenging. In EPR, on the other hand, only unpaired electrons are observed. Besides complete removal of the diamagnetic contribution, EPR allows separating the other magnetic contributions. Au25(SC4)180 crystallizes by forming bundles of linear sequences of clusters interconnected by Au−Au bonds.11 These polymers form through proper relative orientation of staples of neighboring clusters and van der Waals interactions between sufficiently long alkanethiolate ligands (twist-and-lock mechanism), such as C4 and C5.11,13 According to DFT calculations, at the intercluster experimental distance (aurophilic Au−Au bond of 3.15 Å: see inset in Figure 6b) the SOMOs of neighboring clusters allow coupling the individual spins. Indeed, the CW-EPR spectrum of the crystal reveals substantial differences compared to the behavior of the same cluster isolated in frozen solution (Figure 6a) or a cluster protected by too-short ligands, such as Au25(SC3)180, in either state. Analysis of the EPR signal shows that as T increases, the double integrated EPR intensity (I) increases (Figure 6b). As I is proportional to the magnetic susceptibility of the sample, this behavior is opposite to that observed for paramagnetic Au25 clusters in frozen solution (Curie law)22,37 and indicates that the crystal exhibits antiferromagnetic behavior. In this sense, the [Au25(SC4)180]n chain represents an ideal one-dimensional antiferromagnetic system. Analysis according to the corresponding Bonner−Fisher model allowed determining a J value of 28 meV (J is the energy difference between the nonmagnetic ground state and the higher-energy magnetic state), in excellent agreement with DFT calculations (27 meV). [Au25(SC5)180]n, which was prepared by electrocrystallization,13 shows an even shorter intercluster Au−Au distances (2.98−3.03 Å) and, therefore, is also expected to exhibit antiferromagnetic behavior. These results show that interactions between paramagnets in the solid state, which may be modulated by the actual ligands, should always be taken into consideration. We explored magnetic ordering in solid samples of Au25(SC2Ph)180,37 a cluster that does not form polymers. CW-EPR allowed us to detect very different magnetic behaviors depending on the crystalline order: paramagnetism, with weak intercluster interactions, for the amorphous film; superparamagnetism, low-temperature ferromagnetism, and paramagnetism for an

Figure 6. (a) CW-EPR of crystals and frozen solution of Au25(SC4)180 at 20 K. (b) Bonner−Fisher-Hall fit (red) to the experimental data for the crystals, expressed as rescaled relative double-integrated EPR signal values. Adapted with permission from ref 11. Copyright 2014 American Chemical Society.

ensemble of microcrystals (Figure 7); pure ferromagnetism for large single crystals. These different behaviors point to different degrees of magnetic interactions between the unpaired spins, as well as different sizes of the relevant domains. Simulations of the EPR spectra based on the superatom model pointed to the importance of both spin−orbit coupling and crystal-field distortions in determining the EPR properties. These results show that magnetic interactions between clusters in the solid state can be understood only by using atomically precise clusters and controlling both charge state and crystallinity degree. 3.4. TR-EPR and Singlet Oxygen Photosensitization

Photodynamic therapy is a therapeutic strategy in which photoproduced cytotoxic agents, such as excited singlet oxygen (1O2), are generated at the cancerous lesion.38 It has been shown that atomically precise Au nanoclusters hold promises as photosensitizers,39,40 especially because of their efficient 1O2 generation due to enhanced population in the triplet excited states, as recently noted.40 Here we draw attention on the virtually unexplored area (in general, and never for clusters) of studying this process by time-resolved EPR (TR-EPR).41,42 This very sensitive technique allows detection of photosensitized 1O2 in an air-saturated solution by using a free radical (a nitroxide) as the spin probe. Figure 8 shows the TR-EPR spectrum obtained by photoexciting (532 nm) Au25(SC3)18− in toluene in the presence of 4-oxo-2,2,6,6-tetramethyl-1-piperidinyloxy and under aerobic conditions.43 This process requires the cluster to be in the appropriate magnetic state: for Au25(SC3)180, the 49

DOI: 10.1021/acs.accounts.8b00495 Acc. Chem. Res. 2019, 52, 44−52

Article

Accounts of Chemical Research

depends on the concentration of 1O2 and thus can provide a good estimate of the photosensitization efficiency of the system. Moreover, the transients of these signals provide information about the lifetime of 1O2, i.e., the efficiency of its quenching mechanisms. This approach is indeed very promising, as we found photosensitization efficiency comparable to that of conventional porphyrin photosensitizers.

4. CONCLUDING REMARKS Using atomically precise clusters is instrumental to study really fine details of these complex systems. Besides single-crystal Xray crystallography and mass spectrometry, classic molecular physical chemistry, and thus methods such as vibrational spectroscopies, electrochemistry for electron-transfer studies, DFT and MD calculations, and of course, NMR and EPR spectroscopies, are especially valuable. Studies have been mostly carried out on the benchmark molecules Au 25 (SR) 18 , Au38(SR)24, Au102(SR)44, and Au144(SR)60. The information accumulated with these clusters also provides blueprints for understanding the structure−property relationships of new clusters and refining characterization tools. Here we have showcased examples primarily taken from our previous and ongoing NMR and EPR studies, and thus reflecting our research philosophy and interests. NMR and EPR methods have been quite underutilized. NMR spectroscopy still suffers of the still diffuse bias about the actual possibility of observing nice, molecular-like NMR signals. Most often, this is just the result of generalizing the outcome of studies on clusters defined as monodisperse, but surely not atomically precise. NMR spectroscopy of properly controlled clusters is indeed very detailed and can be very informative, as we described for selected examples. Further studies, such as the use of NMR spectroscopy to understand ligand exchange or doping with foreign-metal atoms, could not be covered. Ligandprotected gold clusters are hard−soft molecules where a hard metal core is surrounded by a dynamic molecular layer. Assessing these dynamic properties in solution is one of the most interesting areas in which NMR spectroscopy is especially useful and provides information that X-ray crystallography cannot. The monolayer has indeed more roles than just protecting the core: for example, it interfaces the cluster core with the surrounding environment, controls the diffusion of molecules and ions, generates a nanoenvironment that could be

Figure 7. Temperature effect on the CW-EPR spectra of microcrystals of Au25(SC2Ph)180. Different temperatures (K) highlight superparamagnetic, ferromagnetic, and paramagnetic behaviors. The black and the red traces correspond to low-to-high and high-to-low field scans, respectively: the difference is due to magnetic hysteresis (ferromagnetism). Adapted with permission from ref 37. Copyright 2017 American Chemical Society.

same TR-EPR experiment shows no signals. Photoexcitation of the cluster generates the singlet excited state that is then converted into the triplet excited state via an intersystem crossing. 3O2 quenching of the cluster triplet state generates 1O2 through an energy-transfer process. The magnetic interaction of 1 O2 with the nitroxide then produces spin-polarization in the radical spin sublevels. The polarized TR-EPR signals show an absorptive character, as expected when photosensitization is efficient and 1O2 is involved in the radical-triplet pair mechanism.41 The intensity of the polarized signals primarily

Figure 8. Scheme of the TR-EPR detection of singlet-oxygen generation through photoexcitation of Au25(SC3)18− in toluene (see text). 50

DOI: 10.1021/acs.accounts.8b00495 Acc. Chem. Res. 2019, 52, 44−52

Article

Accounts of Chemical Research

(7) Heaven, M. W.; Dass, A.; White, P. S.; Holt, K. M.; Murray, R. W. Crystal Structure of the Gold Nanoparticle [N(C 8 H 17 ) 4 ][Au25(SCH2CH2Ph)18]. J. Am. Chem. Soc. 2008, 130, 3754−3755. (8) 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. (9) Zhu, M.; Eckenhoff, W. T.; Pintauer, T.; Jin, R. Conversion of Anionic [Au25(SCH2CH2Ph)18]- Cluster to Charge Neutral Cluster via Air Oxidation. J. Phys. Chem. C 2008, 112, 14221−14224. (10) Dainese, T.; Antonello, S.; Gascón, J. A.; Pan, F.; Perera, N. V.; Ruzzi, M.; Venzo, A.; Zoleo, A.; Rissanen, K.; Maran, F. Au25(SEt)18, a Nearly Naked Thiolate-Protected Au25 Cluster: Structural Analysis by Single Crystal X-ray Crystallography and Electron Nuclear Double Resonance. ACS Nano 2014, 8, 3904−3912. (11) De Nardi, M.; Antonello, S.; Jiang, D.; Pan, F.; Rissanen, K.; Ruzzi, M.; Venzo, A.; Zoleo, A.; Maran, F. Gold Nanowired: A Linear (Au25)n Polymer from Au25 Molecular Clusters. ACS Nano 2014, 8, 8505−8512. (12) Agrachev, M.; Antonello, S.; Dainese, T.; Gascón, J. A.; Pan, F.; Rissanen, K.; Ruzzi, M.; Venzo, A.; Zoleo, A.; Maran, F. A Magnetic Look into the Protecting Layer of Au25 Clusters. Chem. Sci. 2016, 7, 6910−6918. (13) Antonello, S.; Dainese, T.; Pan, F.; Rissanen, K.; Maran, F. Electrocrystallization of Monolayer Protected Gold Clusters: Opening the Door to Quality, Quantity and New Structures. J. Am. Chem. Soc. 2017, 139, 4168−4174. (14) Tofanelli, M. A.; Salorinne, K.; Ni, T. W.; Malola, S.; Newell, B.; Phillips, B.; Häkkinen, H.; Ackerson, C. J. Jahn−Teller Effects in Au25(SR)18. Chem. Sci. 2016, 7, 1882−1890. (15) Parker, J. F.; Choi, J.-P.; Wang, W.; Murray, R. W. Electron SelfExchange Dynamics of the Nanoparticle Couple [Au25(SC2Ph)18]0/1− by Nuclear Magnetic Resonance Line-Broadening. J. Phys. Chem. C 2008, 112, 13976−13981. (16) Venzo, A.; Antonello, S.; Gascón, J. A.; Guryanov, I.; Leapman, R. D.; Perera, N. V.; Sousa, A.; Zamuner, M.; Zanella, A.; Maran, F. Effect of the Charge State (z = −1, 0, + 1) on the Nuclear Magnetic Resonance of Monodisperse Au25[S(CH2)2Ph]18z Clusters. Anal. Chem. 2011, 83, 6355−6362. (17) Wu, Z.; Gayathri, C.; Gil, R. R.; Jin, R. Probing the Structure and Charge State of Glutathione-Capped Au25(SG)18 Clusters by NMR and Mass Spectrometry. J. Am. Chem. Soc. 2009, 131, 6535−6542. (18) Antonello, S.; Hesari, M.; Polo, F.; Maran, F. Electron Transfer Catalysis with Monolayer Protected Au25 Clusters. Nanoscale 2012, 4, 5333−5342. (19) Liu, Z.; Zhu, M.; Meng, X.; Xu, G.; Jin, R. Electron Transfer between [Au25(SC2H4Ph)18]−TOA+ and Oxoammonium Cations. J. Phys. Chem. Lett. 2011, 2, 2104−2109. (20) Antonello, S.; Arrigoni, G.; Dainese, T.; De Nardi, M.; Parisio, G.; Perotti, L.; René, A.; Venzo, A.; Maran, F. Electron Transfer through 3D Monolayers on Au25 Clusters. ACS Nano 2014, 8, 2788− 2795. (21) Zhu, M.; Aikens, C. M.; Hendrich, M. P.; Gupta, R.; Qian, H.; Schatz, G. C.; Jin, R. Reversible Switching of Magnetism in ThiolateProtected Au25 Superatoms. J. Am. Chem. Soc. 2009, 131, 2490−2492. (22) Antonello, S.; Perera, N. V.; Ruzzi, M.; Gascón, J. A.; Maran, F. Interplay of Charge State, Lability, and Magnetism in the Molecule-like Au25(SR)18 Cluster. J. Am. Chem. Soc. 2013, 135, 15585−15594. (23) 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. (24) Antonello, S.; Holm, A. H.; Instuli, E.; Maran, F. Molecular Electron-Transfer Properties of Au38 Clusters. J. Am. Chem. Soc. 2007, 129, 9836−9837. (25) Salorinne, K.; Malola, S.; Wong, O. A.; Rithner, C. D.; Chen, X.; Ackerson, C. J.; Häkkinen, H. Conformation and Dynamics of the Ligand Shell of a Water-Soluble Au102 Nanoparticle. Nat. Commun. 2016, 7, 10401.

exploited for carrying out specific reactions, and is instrumental to facilitate reactions with other clusters or thiols. The use of EPR spectroscopy in this area is still in its infancy, despite several techniques and apparatuses available nowadays. In these studies, the clusters need to be perfectly controlled for their stoichiometry as well as their charge and thus magnetic state. EPR spectroscopy clearly showed that the actual ligands and the crystallinity degree of the samples are very important factors in determining the interactions between clusters in the solid state and, consequently, their specific magnetic behavior. ENDOR and ESEEM are very sensitive techniques and provide essential information about spin distribution in the monolayer. TR-EPR, which has been extensively used for other molecular systems, such as organic dyes and fullerenes, could lead to new discoveries especially to assess the kinetics of photoinduced processes. This Account provides a personal glimpse into the distinct capabilities of NMR and EPR spectroscopies to assess specific properties of atomically precise gold clusters. We hope that it will trigger more extensive use of these very powerful and sensitive magnetic techniques in this ever-expanding area.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: fl[email protected]. ORCID

Flavio Maran: 0000-0002-8627-6491 Notes

The authors declare no competing financial interest. Biographies Mikhail Agrachev received his Ph.D. under the supervision of Prof. Maran and is currently a Postdoctoral Fellow. His main research interests are in the magnetic properties of metal nanoclusters. Marco Ruzzi is an Associate Professor, and his research focuses on TREPR detection of excitons, radical pairs, and singlet-oxygen. Alfonso Venzo is an active retiree from the Italian National Research Council. His research interests involve NMR studies of molecular systems, such as metal nanoclusters. Flavio Maran is Professor of Physical Chemistry and Electrochemistry. His main research interests are in the physicochemical properties of monolayer-protected gold clusters, electron transfer, and molecular electrochemistry.

■

REFERENCES

(1) Protected Metal Clusters: From Fundamentals to Applications. In Frontiers of Nanoscience; Tsukuda, T., Häkkinen, H., Eds.; Elsevier: Amsterdam, 2015; Vol. 9. (2) Jin, R.; Zeng, C.; Zhou, M.; Chen, Y. Atomically Precise Colloidal Metal Nanoclusters and Nanoparticles: Fundamentals and Opportunities. Chem. Rev. 2016, 116, 10346−10413. (3) Antonello, S.; Maran, F. Molecular Electrochemistry of Monolayer-Protected Clusters. Curr. Opin. Electrochem. 2017, 2, 18− 25. (4) 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. (5) Kang, X.; Chong, H.; Zhu, M. Au25(SR)18: The Captain of the Great Nanocluster Ship. Nanoscale 2018, 10, 10758−10834. (6) Brust, M.; Walker, M.; Bethell, D.; Schiffrin, D. J.; Whyman, R. Synthesis of Thiol-derivatised Gold Nanoparticles in a Two-Phase Liquid-Liquid System. J. Chem. Soc., Chem. Commun. 1994, 0, 801−802. 51

DOI: 10.1021/acs.accounts.8b00495 Acc. Chem. Res. 2019, 52, 44−52

Article

Accounts of Chemical Research (26) Qian, H.; Zhu, M.; Gayathri, C.; Gil, R. R.; Jin, R. Chirality in Gold Nanoclusters Probed by NMR Spectroscopy. ACS Nano 2011, 5, 8935−8942. (27) Qian, H.; Eckenhoff, W. T.; Zhu, Y.; Pintauer, T.; Jin, R. Total Structure Determination of Thiolate-Protected Au38 Nanoparticles. J. Am. Chem. Soc. 2010, 132, 8280−8281. (28) Dainese, T.; Antonello, S.; Bogialli, S.; Fei, W.; Venzo, A.; Maran, F. Gold Fusion: From Au25(SR)18 to Au38(SR)24, the Most Unexpected Transformation of a Very Stable Nanocluster. ACS Nano 2018, 12, 7057−7066. (29) Dainese, T.; Agrachev, M.; Antonello, S.; Badocco, D.; Black, D. M.; Fortunelli, A.; Gascón, J. A.; Stener, M.; Venzo, A.; Whetten, R. L.; Maran, F. Atomically Precise Au144(SR)60 Nanoclusters (R = Et, Pr) are Capped by 12 Distinct Ligand Types of 5-fold Equivalence and Display Gigantic Diastereotopic Effects. Chem. Sci. 2018, DOI: 10.1039/ C8SC04092C. (30) Yan, N.; Xia, N.; Liao, L.; Zhu, M.; Jin, F.; Jin, R.; Wu, Z. Unraveling the Long-Pursued Au144 Structure by X-ray Crystallography. Sci. Adv. 2018, 4, No. eaat7259. (31) Wong, O. A.; Heinecke, C. L.; Simone, A. R.; Whetten, R. L.; Ackerson, C. J. Ligand Symmetry-Equivalence on Thiolate Protected Gold Nanoclusters Determined by NMR Spectroscopy. Nanoscale 2012, 4, 4099−4102. (32) Tero, T.-R.; Malola, S.; Koncz, B.; Pohjolainen, E.; Lautala, S.; Mustalahti, S.; Permi, P.; Groenhof, G.; Pettersson, M.; Häkkinen, H. Dynamic Stabilization of the Ligand-Metal Interface in Atomically Precise Gold Nanoclusters Au68 and Au144 Protected by metaMercaptobenzoic Acid. ACS Nano 2017, 11, 11872−11879. (33) Schmidbaur, H.; Raubenheimer, H. G.; Dobrzanska, L. The Gold-Hydrogen Bond, Au-H, and the Hydrogen Bond to Gold, Au···HX. Chem. Soc. Rev. 2014, 43, 345−380. (34) Bakar, M. A.; Sugiuchi, M.; Iwasaki, M.; Shichibu, Y.; Konishi, K. Hydrogen Bonds to Au Atoms in Coordinated Gold Clusters. Nat. Commun. 2017, 8, 576. (35) Tian, S.; Liao, L.; Yuan, J.; Yao, C.; Chen, J.; Yang, J.; Wu, Z. Structures and Magnetism of Mono-palladium and Mono-platinum Doped Au25(PET)18 Nanoclusters. Chem. Commun. 2016, 52, 9873− 9876. (36) Krishna, K. S.; Tarakeshwar, P.; Mujica, V.; Kumar, C. S. S. R. Chemically Induced Magnetism in Atomically Precise Gold Clusters. Small 2014, 10, 907−911. (37) Agrachev, M.; Antonello, S.; Dainese, T.; Ruzzi, M.; Zoleo, A.; Aprà, E.; Govind, N.; Fortunelli, A.; Sementa, L.; Maran, F. Magnetic Ordering in Gold Nanoclusters. ACS Omega 2017, 2, 2607−2617. (38) Ogilby, P. R. Singlet Oxygen: There is Indeed Something New under the Sun. Chem. Soc. Rev. 2010, 39, 3181−3209. (39) Kawasaki, H.; Kumar, S.; Li, G.; Zeng, C.; Kauffman, D. R.; Yoshimoto, J.; Iwasaki, Y.; Jin, R. Generation of Singlet Oxygen by Photoexcited Au25(SR)18 Clusters. Chem. Mater. 2014, 26, 2777−2788. (40) Ho-Wu, R.; Yau, S. H.; Goodson, T., III Efficient Singlet Oxygen Generation in Metal Nanoclusters for Two-Photon Photodynamic Therapy Applications. J. Phys. Chem. B 2017, 121, 10073−10080. (41) Kawai, A.; Mitsui, M.; Kobori, Y.; Obi, K. CIDEP in RadicalSinglet Molecular Oxygen System. Appl. Magn. Reson. 1997, 12, 405− 410. (42) Martinez, C. G.; Jockusch, S.; Ruzzi, M.; Sartori, E.; Moscatelli, A.; Turro, N. J.; Buchachenko, A. L. Chemically Induced Dynamic Electron Polarization Generated through the Interaction between Singlet Molecular Oxygen and Nitroxide Radicals. J. Phys. Chem. A 2005, 109, 10216−10221. (43) Fei, W.; Agrachev, M.; Antonello, S.; Dainese, T.; Zoleo, A.; Ruzzi, M.; Maran, F. Manuscript in preparation.

52

DOI: 10.1021/acs.accounts.8b00495 Acc. Chem. Res. 2019, 52, 44−52