Gold chloride clusters with Au(III) and Au(I) probed by FT-ICR mass spectrometry and MP2 theory.

PCCP View Article Online

Published on 29 January 2014. Downloaded by California Institute of Technology on 03/10/2014 11:19:30.

PAPER

Cite this: Phys. Chem. Chem. Phys., 2014, 16, 7813

View Journal | View Issue

Gold chloride clusters with Au(III) and Au(I) probed by FT-ICR mass spectrometry and MP2 theory Kono H. Lemke Microsolvated clusters of gold chloride are probed by electrospray ionization mass spectrometry (ESI-MS) and scalar relativistic electronic structure calculations. Electrospray ionization of aqueous AuCl3 leads to mononuclear clusters of types [AuCl2]+(H2O)n (n = 0–4), [AuOHCl]+(H2O)n (n = 0–1) and [AuCl2]+(HCl)2(H2O)n (n = 0–4). In addition, strong ion signals due to dinuclear [Au2Cl5 xOHx]+(H2O)n (x = 0–1) are present in ESI mass spectra of aqueous AuCl3, with the abundance of individual dinuclear species controlled by the concentration-dependent variation of the precursor complexes [AuCl2 xOHx]+(H2O)n and AuCl3. Equilibrium structures, energies and thermodynamic properties of mono- and dinuclear gold clusters have been predicted using MP2 and CCSD(T) theory, and these data have been applied to examine the influence of microsolvation on cluster stability. Specifically, results from CCSD(T) calculations indicate that non-covalently bound ion-neutral complexes Au+(Cl2)(H2O)n, with formal Au(I), are the dominant forms of mononuclear gold with n = 0–2, while higher hydrates (n 4 2) are covalently bound [AuCl2]+(H2O)n complexes in which gold exists as Au(III). MP2 calculations show that the lowest energy structure of

Received 4th December 2013, Accepted 27th January 2014 DOI: 10.1039/c3cp55109a

dinuclear gold is an ion-molecule cluster [Au2Cl(Cl2)2]+ consisting of a single-bridged digold-chloronium ion bound end-on to two dichlorine ligands, with two higher energy isomers, single-bridged [Au2Cl3(Cl2)]+ and double-bridged [Au2Cl5]+ clusters. Finally, Au Au interactions in the singly-bridged clusters [Au2Cl(Cl2)2]+(H2O)n and [Au2Cl3(Cl2)]+(H2O)n are examined employing a wide range of computational tools, including natural bond

www.rsc.org/pccp

order (NBO) analysis and localized orbital locator (LOL) profiles.

I. Introduction The structure and stability of gold halide clusters, especially those of Au(III) and Au(I), forms a topic of ongoing research interest among physical1–4 and theoretical chemists,5–10 with important applications in the field of catalysis,11 and in the Earth and environmental sciences.12–15 A common theme of these studies is that relativistic effects dominate the chemistry of gold compounds with important implications for the speciation and the oxidation state of gold. An interesting example is the unusual bonding and range of oxidation states discovered in AuCl2+/0/ . Using electrospray ionization mass spectrometry ¨der and coworkers showed that AuCl2+/0/ com(ESI-MS) Schro plexes exist as highly stable gas-phase species with fundamentally different bonding schemes and thermodynamic stabilities.3 These exciting findings were further supported by high level coupled-cluster CCSD(T) computations. Another question of interest is the extent to which gas-phase MS data correlate with the solution situation. Of particular relevance here are observed correlations between gas-phase and solution Department of Earth Sciences, University of Hong Kong, Pokfulam Road, Hong Kong, SAR, Hong Kong. E-mail: [email protected]; Fax: +852-2517-6912; Tel: +852-2241-5474

This journal is © the Owner Societies 2014

behaviour of metal halide compounds that are obtained from concentration-dependent ESI-MS measurements.16 At present, there are no ESI-MS speciation data available for the AuCl3 system; however, it is noteworthy that ESI-MS data for PdCl2,17 NiCl2,18 and CuCl219 provide strong evidence of a correlation between gas-phase cluster data and solution speciation. With regards to the solution phase, UV-studies of gold halide salts provide abundant evidence for ion pairing,14 however, the situation is not as clear for polynuclear Au. Results from UV experiments20 suggest that polynuclear Au clusters occur in solutions close to the solubility limit, however, the experimental evidence is indirect and there is no information on the stoichiometries and structures of these species. Prompted by recent ESI-MS results of AuCl2+/0/ ,3 and mass spectrometric studies on bulk solution speciation,21 this paper reports on the results of a combined ESI FT-ICR mass spectrometric and theoretical investigation of the speciation of AuCl3. The present study focuses on the ESI-MS characterization of mono- and dinuclear gold clusters, the nature of the chemical bonding in these species and their thermodynamic stability, in particular upon interaction with water. Gold clusters were prepared by electrospray ionization of aqueous AuCl3 solutions at 5–50 mM, and ESI-MS results have been complemented by NBO computations and localized-orbital locator (LOL) maps, which describe

Phys. Chem. Chem. Phys., 2014, 16, 7813--7822 | 7813

View Article Online

Paper

the organization of bonding in these clusters in terms of the local kinetic-energy density.


II. Computational methods It is well established that theoretical calculations of goldbearing clusters should include both correlation and relativistic effects,7 and as such require MP2 or higher level treatment with scalar relativistic effective core potentials (ECP). The benchmark in theoretical gold chemistry are computations at the CCSD(T) level, however, CCSD(T) scales as N7 and is computationally prohibitive and is therefore, at present, not applicable to larger species, in particular to the dinuclear Au clusters examined here. Calculations at the MP2 level, on the other hand, scale more favorably as N5, and in combination with the correlation consistent type basis set, are well suited to probe larger gold clusters.5,7 Two different gold-dichloride cations have been examined using MP2 and CCSD(T) theory in combination with double-z (cc-pVDZ-PP) and triple-z (cc-pVTZ-PP) basis sets (abbreviated as VDZ and VTZ) for Au,22 cc-pVXZ (X = D,T) for O and H23 and cc-pV(D + d)Z for Cl.24 Basis set superposition error (BSSE) corrections have not been applied in this study given that MP2/VTZ level computations for Au(I) complexes with water have shown to yield BSSE errors of around 2–3 kcal mol 1,25 which is consistent with a BSSE correction of 2.1 kcal mol 1 for the dimer complex Au+(Cl2) found here. The first set of theoretical calculations carried out include MP2 and CCSD(T) geometry optimizations and frequency computations of (i) the ion–molecule complex Au+(Cl2), the bent [AuCl2]+ ion cluster, and corresponding solvated forms. Differences in stability between the above complexes (DE) have been evaluated using MP2 and CCSD(T) level theory with VTZ basis sets and these results are summarized in Fig. 2. The spatial organization of bonding in the above complexes was examined using localized orbital locator (LOL) maps.26,27 These data are presented graphically as LOL profiles and have been chosen to highlight the bonding pattern in mononuclear and dinuclear gold clusters. In brief, the LOL function is a measure of electron localization based on the electron kinetic energy density, and is referenced against the uniform electron gas with a LOL value of 0.5. Values of LOL larger than 0.5 are characteristic of regions with localized electrons, typical of covalent bonding and lone pairs, and these are shown as green/yellow/red basins in LOL maps. LOL values smaller than 0.5 are characterized by delocalized electrons and are typical of regions with multicenter bonding and are shown in deep blue, blue and lighter blue in LOL contour plots. Structures of dinuclear Au clusters were preoptimized at the MP2/VDZ level using a Stuttgart–Dresden scalar-relativistic ECP (ECP60MDF) for Au,28 and then reoptimized at the MP2/VTZ level. Three dinuclear gold cluster configurations have been considered: (i) a structure derived from planar double-bridged Au2Cl6, clusters labeled as [Au2Cl5]+ in Fig. 5, (ii) an ion–molecule cluster consisting of a central digold-chloronium cation bound end-on to two dichlorine ligands, [Au2Cl(Cl2)2]+ and (iii) a chloronium-bridged cluster in

7814 | Phys. Chem. Chem. Phys., 2014, 16, 7813--7822

PCCP

which one Au atom interacts with Cl2 and the other binds to three Cl atoms in a T-shaped configurations, labeled as [Au2Cl3(Cl2)]+. The influence of relativistic effects on the structures and energetics of [Au2Cl5]+, [Au2Cl3(Cl2)]+ and [Au2Cl(Cl2)2]+ has been probed by comparing MP2 results using scalar relativistic ECPs with those obtained from non-relativistic ECP60MHF pseudopotentials for Au,29 and these results are discussed in terms of changes in the Au–Au interaction distance and energy. Finally, natural bond orbital (NBO) analyses have been performed to examine the charge distributions on Au and charge-transfer between Au atoms and neighboring ligands in mono- and dinuclear gold upon solvation.

III. Experimental methods Mass spectrometric experiments were performed on a Bruker Daltonics 7.0 Tesla Fourier-transform ion cyclotron resonance mass spectrometer (FT-ICR/MS). Gold chloride clusters were obtained by electrospray ionization (ESI) of aqueous AuCl3 solutions (conc. 5–50 mM) (99.99%, CAS#: 12453-07-1) prepared in deionized water (Millipore). The FT-ICR mass spectrometer is equipped with an electrospray ionization source consisting of an angled (451) spray unit operated at a flow rate of 50 mL h 1 set to a spray voltage of 3.9 kV, followed by a heated (200 1C) 180 mm long quartz capillary and two skimmers. The parameters for the ion source were as follows: capillary exit 80 V, skimmer1 20 V and skimmer2 10 V. Ion clusters produced in the ESI processes were then accumulated in the hexapole for 1 second before being transferred to the ICR cell. Typical operating pressure for the first ESI stage is 10 1 mbar, in the hexapole region: 5 10 6–8 10 7 mbar and UHV ICR cell 2 10 10 mbar. Gold is a monoisotopic system containing only 196Au, while chlorine is a double isotope system with 35Cl (75.5%) and 37Cl (24.5%). A hypothetical [AuCl2]+ ion would therefore occur in three different isotopic combinations, and would be expected to appear as a triplet in the mass spectrum with one dominant peak at m/z 266.9, which is assigned to 196Au(35Cl)2 (100%) and two smaller peaks at m/z 268.9 and 270.9, assigned to 196Au37Cl35Cl (63.9%) and 196Au(37Cl)2 (10.2%), respectively. Given the low natural abundance of 17O and 18O, ion mass spectra of the solvated complexes [AuCl2]+(H2O)n would also appear as a triplet pattern at m/z 266.9 + n18.0, where n represents the number of water molecules in the cluster. Mass spectra were recorded in positive-ion mode with a mass range of 20 to 2000 m/z, and each spectrum represents an average of 600 scans at 256k data points per scan. Mass spectrometric measurements and postprocessing were performed using Bruker XMass version 7.0.8, and theoretical mass spectra of gold clusters with different isotopic compositions (isotopologues) were calculated using the Bruker DataAnalysis program.

IV. Results and discussion A.

Experimental results

Fig. 1a–d presents positive-ion mode ESI mass spectra for electrosprayed (50 mM) aqueous AuCl3 solutions at pH 2.3. Three major mononuclear complexes have been observed in the


View Article Online


PCCP

Paper

Fig. 2 Normalized ion abundances of solvated mono- and dinuclear gold species obtained from electrosprayed aqueous solutions of AuCl3 in the concentration range 5–50 mM.

Fig. 1 Positive ion-mode ESI mass spectra of aqueous AuCl3 (50mM) showing four major families of microsolvated ions: the mononuclear complexes Au+(H2O)n, Au+(H2O)n(HCl) and [AuCl]+(H2O)n (a); the triple ions [AuCl2]+(H2O)n (n = 0–4) and [AuOHCl]+(H2O)n (n = 0–1) (b), the mononuclear ion [AuCl2]+(HCl)2(H2O)n with n r 4 (c) and the dinuclear clusters [Au2Cl4OH]+(H2O)n (n = 1,2) and [Au2Cl5]+(H2O)2 (d).

mass range m/z 260–420, being [AuCl2]+(H2O)n (m/z 266.9 + n18.0, with n = 0–4), [AuOHCl]+(H2O)n (m/z 248.9 + n18.0 with n = 1) (Fig. 1b) and [AuCl2]+(HCl)2(H2O)n (m/z 338.9 + n18.0 with n = 1–4) (Fig. 1c); ion signals have also been observed for Au+(H2O)n but these to not exceed n = 2 (see Fig. 1a). A series of ion signals at lower mass ranges were identified as protonwater clusters at m/z 37.0 and 55.0, being [H5O2+] and [H7O3+], respectively, with pH-responsive behavior forming stable clusters under lower pH conditions. In addition to mononuclear gold, Fig. 1 also presents mass spectra of three dinuclear gold clusters with m/z 570.8 + n18.0 (n = 1–2) and 606.8 + n18.0 (n = 2)


(Fig. 1d) obtained from electrosprayed AuCl3 at 50 mM, and these have been assigned the general form [Au2OHCl4]+(H2O)n and [Au2Cl5]+(H2O)n, respectively. An important quantity, linking gas-phase ESI mass spectra with solution phase data, is the measured intensity ratio between monomeric and dinuclear clusters, and the concentrationdependent change of this ratio. Here, the focus is on the ratio between monomeric [AuCl2]+(H2O)n, [AuOHCl]+(H2O)n and dinuclear gold, because the latter would be expected to form by aggregation of the mononuclear complex with AuCl3. To this end, aqueous AuCl3 solutions have been electrosprayed in the concentration range 5–50 mM and results from these ESI experiments are displayed in Fig. 2 as normalized ion abundances. The following trends can be observed: first, for monomeric complexes it is seen that an increase in concentration from 5 to 50 mM results in a decrease in [AuCl2]+(H2O)n while the mixed complexes [AuOHCl]+(H2O)n remain largely unaffected at 7% of total gold. Furthermore, Fig. 2 shows that the decrease in the abundance of [AuCl2]+(H2O)n species is accompanied by a gradual rise in the dinuclear gold fraction. For instance, at the highest electrosprayed concentrations (25–50 mM), around 20% of gold chloride species are present as dinuclear clusters, demonstrating a strong tendency toward ion aggregation with increasing AuCl3 concentration. While the increase of dinuclear clusters is consistent with previous ESI-MS metal-halide ion data,21 it is noteworthy that an increase in AuCl3 solution content leads to a decrease in the ratio between [AuCl2]+(H2O)n and [AuOHCl]+(H2O)n, which at first appears inconsistent with reported aqueous speciation data for AuCl3 nOHn.14 This decrease in abundance of [AuCl2]+(H2O)n, in particular, over the concentration range 5–10 mM is attributable to [AuCl2]+(H2O)n hydrolysis and aggregation with AuCl3 toward dinuclear gold, the former accounting for around 25% of [AuCl2]+(H2O)n loss. The majority of [AuCl2]+(H2O)n loss, however, around 75%, occurs as a consequence of gold dimerization, and is clearly reflected in the increase of the dinuclear fraction upon increase in AuCl3 solution content from 5 to 25 mM. Finally, it is noteworthy that the ratio of [Au2Cl5]+(H2O)n


View Article Online


Paper

to [Au2Cl4OH]+(H2O)n increases with increasing AuCl3 solution content, in particular in the 25–50 mM concentration range. This feature suggests that at the highest AuCl3 concentration, at pH 2.3, Cl/OH ligand switching leads to an increase in the fraction of [Au2Cl5]+(H2O)n, while the fraction of [Au2Cl4OH]+(H2O)n decreases with decreasing solution pH, which is qualitatively consistent with Cl/OH ligand switching equilibria in aqueous gold chloride solutions.14 B.

Theoretical results: [AuCl2]+(H2O)n

Fig. 3 presents MP2 equilibrium structures and relative energies (DE) of mononuclear gold complexes, both water-free and microsolvated and key geometric parameters are summarized in Table 1. At the MP2 and CCSD(T) level, the global minimum structure of gold dichloride is an ion–molecule complex Au+(Cl2) with Cs symmetry, in which molecular chlorine is bound end-on to Au+, and thus, the ion complex at m/z 267 is anticipated to be a dimer consisting of Au+ and Cl2. The MP2 predicted Au Cla distance in Au+(Cl2) is 2.283 Å and the angle j between Cl atoms (Cla–Au–Clb) is 34.21 or 106.31 for Au–Cla–Clb using MP2 and these values are close to CCSD(T) results (33.71 for ¨der and coworkers3 Cla–Au–Clb or 106.11 for Au–Cla–Clb). Schro + reported a Au Cla distance for Au (Cl2) using CCSD(T) with smaller double-zeta basis sets (being 2.45 Å), which is somewhat longer than the CCSD(T) value of 2.342 Å reported here. The bent C2v complex [AuCl2]+, shown in Fig. 3, is a second minimum, for which the Au Cl distance and the Cl–Au–Cl angle using MP2 are 2.167 Å and 95.11, respectively. For comparison, a slightly larger RAu Cl value, 2.204 Å, and smaller angle g, 93.91, is found by CCSD(T), which is in fair agreement with a reported CCSD(T) value of 2.24 Å.3 Fig. 3 also shows structures and relative energies of solvated Au clusters with up to three water molecules. For the (H2O)nAu+(Cl2) complex, one first-shell water molecule is bound to Au, while the second and third water molecules are located in the

Fig. 3 MP2 equilibrium structures of [AuCl2]+(H2O)n and Au+Cl2(H2O)n with n = 0–3 and their relative energies given in kcal mol 1.


PCCP Table 1 Equilibrium MP2 and CCSD(T) geometrical parameters of waterfree and solvated mononuclear gold chloride clustersa

Cluster

Au+Cl2 (Cs)

Au+Cl2(H2O) (Cs)

Au+Cl2(H2O)2 (C1)

Au+Cl2(H2O)3 (Cs)

Au–Cla Cla–Clb j

2.283 [2.342] 1.997 [2.016] 106.3 [106.1]

2.221 [2.314] 2.000 [2.052] 106.8 [105.2]

2.219 2.001 107.3

2.218 2.002 107.6

[AuCl2]+ Cluster (C2v) Au–Cla Au–Clb g

[AuCl2]+(H2O) [AuCl2]+(H2O)2 [AuCl2]+(H2O)3 (C2) (C1) (Cs)

2.167 [2.204] 2.203 2.167 [2.204] 2.164 95.1 [93.9] 88.4

2.194 2.194 90.4

2.197 2.200 90.5

a

MP2/VTZ bond distances in Å and angles in degrees, CCSD(T)/VTZ data in brackets; see clusters in Fig. 3 for positions of Au, Cl and O atoms. Au: scalar relativistic ECP covering 60 electrons with cc-pVTZ basis set, cc-pVTZ for O and H and cc-pV(T + d)Z for Cl.

second-shell, accepting each one hydrogen bond from first-shell water. The Au Cl distance in the ion–molecule complex Au+(Cl2) is highly sensitive to the number of water molecules coordinating the complex, with the largest change in RAu Cl occurring after attachment of the first water molecule. At the MP2 level, the value of RAu Cl in (H2O)nAu+(Cl2) decreases from 2.283 Å (n = 0), 2.221 Å (n = 1), 2.219 Å (n = 2) to 2.218 Å (n = 3). MP2 values of RAu Cl in the bent complex [AuCl2]+(H2O)n shift from 2.203 Å (n = 1), 2.194 Å (n = 2) to 2.197 Å (n = 3), with two first-shell water molecules and one hydrogen-bonded second-shell water molecule. An inspection of relative energies presented in Fig. 3 shows that ion-neutral complexes are the global minimum structures up to n = 2, and thus, the complex at m/z 285 is considered to be a non-covalently interacting trimer comprised of Au+, H2O and Cl2. Accordingly, the complex at m/z 303 can be viewed as a tetramer consisting of non-covalently interacting Au+, Cl2 and (H2O)2. From Fig. 3 it is seen that the MP2 energy differences between (H2O)nAu+(Cl2) and bent [AuCl2]+(H2O)n are 21.5 kcal mol 1 for n = 1 and 2.2 kcal mol 1 for n = 2, however, for the higher hydrate (n = 3) this trend is reversed, and the (H2O)3Au+(Cl2) complex is around 2 kcal mol 1 less stable than [AuCl2]+(H2O)3. Fig. 4 shows representative LOL profiles for four mononuclear gold species. Previously, LOL maps have demonstrated to deliver valuable information on the spatial organization of electrons in simple molecules such as water30 and more complex metal-ligand systems.31–33 As seen from the LOL profiles of Au+(Cl2) (Fig. 4a), regions of reduced kinetic energy density are found between Au and Cl2, indicating that electrons are delocalized over this complex and, in particular, in the vicinity the Au Cla contact. Regions with enhanced electron localization, on the other hand, are clearly discernible as yellow/red basins on Au and as shared and lone pairs in Cl2. The LOL map of [AuCl2]+ (Fig. 4b) shows regions with enhanced electron localization, seen as light green basins, sandwiched between Au and Cl. LOL maps shown in Fig. 4c and d further illustrate the influence that solvation has on the LOL contours separating Au and Cl. In [AuCl2]+ electrons are locally concentrated in the vicinity of the Au–Cl contact, but are partially depleted by a blue breach in [AuCl2]+(H2O). Solvation of the


View Article Online


PCCP

Paper

Fig. 4 LOL contour plots for [AuCl2]+(H2O)0,1 and Au+Cl2(H2O)0,1 in the Cla–Au–Clb plane. LOL values below 0.35 shown as light to dark blue basins denote delocalized electrons, values of LOL above 0.35 characterize localized electrons.

ion–molecule complex Au+(Cl2) (Fig. 4c) leads to moderate electron localization between Au and Cl and shortening (0.06 Å) of the Au Cla contact (Table 1). To briefly summarize, LOL analyses of the bonding pattern in mononuclear gold complexes illustrate that (i) contacts with considerable electron delocalization dominate the bonding between Au+, molecular chlorine and water in the ion–molecule complexes Au+(Cl2) and (H2O)Au+(Cl2), and these are shown as pale-blue LOL basins and (ii) attachment of H2O onto the bent complex [AuCl2]+ promotes electron delocalization between Au and Cla and an accompanying Au–Cla bond length extension (B0.04 Å). Table 2 lists MP2 level reaction enthalpies (DH), entropies (DS) and free energies (DG) for Cl2 loss from Au+(Cl2) (R1), and Au–Cl bond rupture in [AuCl2]+ and [AuCl]+ (R2) and (R3) yielding Cl and reduced gold. Previous CCSD(T) studies3 report DH values of around 23 kcal mol 1 for Cl2 loss from Au+(Cl2) (R1) and 49 kcal mol 1 for the dissociation reaction (R2) with n = 0. The MP2 and CCSD(T) predicted values of DH for (R1) reported here are 32.5 kcal mol 1 and 29.2 kcal mol 1, respectively. Cleavage of one Au–Cl bond in [AuCl2]+ (R2) yielding [AuCl]+ is endothermic by 49.7 kcal mol 1, while loss of a second Cl atom from [AuCl]+ (R3) is endothermic at 34.0 kcal mol 1 with MP2. MP2 calculations have also been conducted to examine the

effect of microsolvation on the dissociation energies for reactions (R1)–(R3), and Table 2 lists values of DH, DS and DG in which up to three water molecules are taken into account. Inclusion of one water on Au+(Cl2) and Au+ increases (by around 10 kcal) the endothermicity of reaction (R1), in other words, it is more difficult to remove Cl2 from (H2O)Au+(Cl2) than from Au+(Cl2). This trend is reversed (Cl2 loss is less endothermic) upon solvation by a second and third H2O molecule, and is predicted to be 10.9 and 7.5 kcal mol 1, respectively. Interestingly, the values of DH for (R1) with n = 2 and 3 are very close to the 10.5 kcal mol 1 water condensation enthalpy, indicating that Cl2 elimination and H2O loss (vaporization) are thermodynamically indistinguishable. MP2 results for reaction (R1) reveal a clear preference for Cl2 ligation over expulsion, however, it is unclear how the solution phase would influence the energetics of reaction (R1). This consideration is important because expulsion of Cl2 from the complex into solution would result in an entropy increase that would influence the free energy change of (R1). In order to address this problem, reaction (R1) has been modified such that Cl2 is treated as an aqueous species (Cl2(g) - Cl2(aq) with DH = 6.0 kcal mol 1 and DG = 1.7 kcal mol 1) and the solution behavior of Au+(Cl2) is approximated by the stepwise solvation with up to three water molecules. The hydration of the chlorine molecule is only mildly endergonic, and thus any entropy change due to the solvation of Cl2 would be compensated by the increase in enthalpy due to the interaction of the solvent with Cl2. Thus, the entropy change due to dissolving Cl2 into water does not appear to be the source of the enhanced stability of the (H2O)Au+(Cl2) complex. Instead, the dominant effect controlling the stability of Au+(Cl2) in solution appears to be the solvent interaction of (H2O)n 1Au+(Cl2) versus (H2O)nAu+(Cl2), and the water number-dependent changes in DG for reaction (R1). The influence of microsolvation on DG for reaction (R1) is shown in Table 2, and it is clearly seen that attachment of the first water molecule onto Au+(Cl2) shifts DG to larger values, indicating that if Au+(Cl2) is embedded in a suitable solvent environment it may serve as a good candidate for experimental detection. Another important set of reactions that may compete with Cl2 elimination are those where Au(III) is reduced in a stepwise sequence to Au(II) (reaction (R2)) and from Au(II) to Au(I) (reaction (R3)). The effects of microsolvation on (R2) are shown in Table 2. Solvation with one water molecule induces a decrease in the endothermicity, relative to the water-free system, by around 9 kcal mol 1. Introduction of a second and

Table 2 Effects of microsolvation on Cl2 expulsion and Au–Cl bond dissociation enthalpies (DH)a, entropies (DS)a and free energies (DG)a of mononuclear [AuCl2]+(H2O)n and (H2O)nAu+Cl2

n=0

n=1

n=2

n=3

Reaction

DH

DS

DG

DH

DS

DG

DH

DS

DG

DH

DS

DG

(H2O)n Au+(Cl2) - Au+(H2O)n + Cl2 (R1) (H2O)n[AuCl2]+ - (H2O)nAuCl+ + Cl (R2) (H2O)nAuCl+ - (H2O)nAu+ + Cl (R3)

32.5 49.7 34.0

19.5 24.2 17.8

26.7 42.5 28.7

41.4 40.9 48.5

30.5 27.6 26.3

32.3 32.7 40.7

10.9 58.5 21.0

25.5 37.0 20.2

3.3 47.5 15.0

7.5 60.1 20.7

23.4 40.3 20.0

0.5 48.1 14.7

a MP2/VTZ reaction enthalpies (kcal mol 1), entropies (cal mol corrected electronic energies.


1

K 1), free energies (kcal mol 1) obtained from differences between thermally-


View Article Online


Paper

PCCP

third water molecule causes reaction (R2) to become increasingly less favorable and shift from 58.5 kcal mol 1 (n = 2) to 60.1 kcal mol 1 (n = 3). The opposite occurs for reaction (R3), the reduction of [AuCl]+ to Au+: values of DH for (R3) with one, two and three water molecules are 48.5, 21.0 and 20.7 kcal mol 1, respectively. This reaction is made less endothermic by the presence of water molecules, and there is a noticeable influence of solvation on reaction energies upon inclusion of a second water molecule. The reason for this discontinuity in the enthalpy profile stems from thermochemical effects of solvation shellfilling in [AuCl]+ and [Au]+, and is attributable to the higher binding energy of H2O to [AuCl]+ (55.1 kcal mol 1) than [Au]+ (40.3 kcal mol 1). Addition of a second water molecule (n = 2) decreases the endothermicity of reaction (R3) to 20.9 kcal mol 1 and, consistent with results for n = 1, this feature in the energy profile of (R3) can be ascribed to the higher binding energy of H2O to H2O[Au]+ (50.4 kcal mol 1) than to H2O[AuCl]+ (23.3 kcal mol 1). In summary, mass spectrometric results indicate that gold dichloride complexes exist as stable microsolvated species and occur in three different oxidation states. With respect to the stability of individual Au(I), (II) and (III) complexes, the predicted MP2 values of DG for reaction (R1) indicate that Cl2 expulsion from (H2O)nAu+(Cl2) is endergonic for n = 0 and 1 but thermodynamically almost neutral for n = 2 and 3. The stepwise reduction reactions of (H2O)n[AuCl2]+ to (H2O)n[AuCl]+ and finally to (H2O)n[Au]+are highly endergonic, both for the water-free system and when microsolvation effects are included. Thus, while ion–molecule complexes in general appear to be stable against Cl2 loss, it is conceivable that ligand switching reactions involving H2O and Cl2 would occur in (H2O)nAu+(Cl2) with two or more water molecules. Table 3 provides a summary of MP2 level natural bond orbital (NBO) analyses on mononuclear gold clusters, including natural atomic charges and electron configurations for atomic Au, Cl, and O. As seen, there is a moderate charge transfer Table 3 Natural charges and electron configurationsa of Au, Cl and O in mononuclear [AuCl2]+(H2O)n and Au+Cl2(H2O)n (n = 0–1)

Atomb

Species +

(H2O)Au (Cl2)

Au+(Cl2) (H2O)[AuCl2]+

[AuCl2]+ Cl2 Cl H2O Au+3 Au+ a

Au Cla Clb O Au Cla Clb Au Cla Clb O Au Cla Clb Cla,b Cl O Au Au

MP2/VTZ theory level.

NPA charge 0.64 0.09 0.13 0.94 0.86 0.01 0.15 1.02 0.04 0.13 0.93 1.05 0.03 0.03 0.00 1.00 0.90 3.00 1.00

b

Electron configuration 6s0.575d9.736p0.05 3s1.873p4.973d0.05 3s1.933p4.913d0.02 2s1.712p5.213d0.01 6s0.205d9.926p0.02 3s1.933p4.89 3s1.923p5.043d0.03 6s0.535d9.186p0.24 3s1.923p5.103d0.01 3s1.913p5.203d0.02 2s1.712p5.203d0.01 6s0.495d9.286p0.15 3s1.943p5.073d0.02 3s1.943p5.073d0.02 3s1.923p5.043d0.03 3s23p5 2s1.742p5.153d0.01 6s05d8 6s05d10

For position of Cla and Clb, see Fig. 3.


(0.14 e) from chlorine (Clb) to gold in the ion–molecule complex Au+(Cl2). This charge transfer is accompanied by an increase in the Au 6s orbital population, while 5d orbitals are increasingly depopulated and participate less in Au–Cl bonding. Additional charge (0.22 e) is transferred to Au upon attachment of water to Au+(Cl2), and the Au 6s orbital in (H2O)Au+(Cl2) further increases in occupation relative to Au+(Cl2). On the other hand, NBO data for [AuCl2]+ reveal a small charge transfer from gold to chlorine, being 0.05 e, and a minor back-transfer of charge (0.04 e) from Au to water oxygen upon solvation by one water molecule. This charge transfer in (H2O)[AuCl2]+ coincides with an increase in the 6s and 6p orbital population; however, the 5d orbitals of Au in (H2O)[AuCl2]+ are progressively depopulated. The predicted trends in reaction energies and atomic charges discussed above have important implications for ESI-MS studies of aqueous AuCl3. First, the energetic penalty of removing Cl2 from (H2O)nAu+(Cl2) decreases with increasing solvation number, until with n = 3, water and Cl2 loss are isoenergetic; thus, one water molecule would assist in stabilizing Au+(Cl2), whereas in higher (n Z 2) hydration environments, Cl2 is more easily exchanged against H2O. This feature is particularly evident if one considers the variation in DG for (R1) as function of the number of H2O molecules. Next, consistent with theoretical results for the dissociation of Au+(Cl2) and (H2O)Au+(Cl2), the occurrence of ion signals at m/z 303 and 321 can be rationalized by the presence of the bent complexes (H2O)2[AuCl2]+ and (H2O)3[AuCl2]+. Both of these complexes are exceptionally stable against reduction toward (H2O)2[AuCl]+ by 47.5 kcal mol 1, and (H2O)3[AuCl]+ by 48.1 kcal mol 1, respectively. Given the small energy difference (2 kcal mol 1) between (H2O)nAu+(Cl2) and (H2O)n[AuCl2]+ with n = 2 and 3 one might, however, expect that a mixture of at least two isomers be present at m/z 303 and 321 in ESI-MS experiments. C.

Theoretical results: [Au2Cl5]+(H2O)n

MP2 structures of water-free and hydrated dinuclear Au clusters and corresponding relative energies (DE) are shown in Fig. 5. The global minimum structure of dinuclear Au is an ion– molecule complex, [Au2Cl(Cl2)2]+, formed between a digold chloronium ion and two Cl2 molecules, with a MP2 predicted Aua–Aub distance of 3.002 Å and a bond angle b (AuaClaAub) of 83.71 (Table 4). The MP2 Aua–Aub distance and the bond angle in [Au2Cl(Cl2)2]+ compare favorably with results from X-ray diffraction studies of [Au2Cl(Ph3P)]+,34 the reported values of RAu–Au and +Au–Cl–Au being 3.085 Å and 82.71, respectively. The singly-bridged cluster [Au2Cl3(Cl2)]+ is a second minimum structure with a Aua–Aub distance of 2.639 Å and lies around 3 kcal mol 1 above the global minimum. A third minimum energy structure, about 9 kcal mol 1 higher in energy than [Au2Cl(Cl2)2]+, was derived from the planar structure of Au2Cl6 and has a longer Aua–Aub distance (3.329 Å). The MP2 predicted Au–Au distances in [Au2Cl(Cl2)2]+ and [Au2Cl3(Cl2)]+ are significantly shorter than the sum of the van de Waals radii of two Au atoms (3.320 Å), the difference being 0.32 Å for [Au2Cl(Cl2)2]+ and 0.68 Å for [Au2Cl3(Cl2)]+. Aurophilic attraction between gold atoms in clusters, such as Au2Cl2, results in a shortening of the


View Article Online


PCCP

Fig. 5

Paper

MP2 equilibrium structures of dinuclear clusters [Au2Cl(Cl2)2]+(H2O)n, [Au2Cl3(Cl2)]+(H2O)n and [Au2Cl5]+(H2O)n with n = 0–2.

Table 4 Equilibrium MP2 structural parameters for dinuclear [Au2Cl(Cl2)2]+(H2O)n, [Au2Cl3(Cl2)]+(H2O)n and [Au2Cl5]+(H2O)n (n = 0–2)a

Cluster

[Au2Cl(Cl2)2]+ [Au2Cl(Cl2)2]+(H2O) [Au2Cl(Cl2)2]+(H2O)2

Aua–Aub Aua–Cla Aua–Clb (b) AuaClaAub (y) ClaAubClbCl

3.002, 4.343 2.249, 2.480 2.249, 2.642 83.7, 122.6 133.4, 4.9

Cluster

[Au2Cl3(Cl2)]+ [Au2Cl3(Cl2)]+(H2O) [Au2Cl3(Cl2)]+(H2O)2

Aua–Aub Aua–Clb Aub–Cla (g) AuaClClb (y) ClaAubCldCl

2.639, 3.661 2.222, 2.324 2.238, 2.611 70.2, 94.6 117.7, 109.7

2.541 2.236 2.270 66.8 179.9

2.531 2.243 2.273 66.71 179.9

Cluster

[Au2Cl5]+

[Au2Cl5]+(H2O)

[Au2Cl5]+(H2O)2

Aua–Aub Aua–Cld Aua–Cle Aub–Cla (j) AuaClbAub

3.329, 3.453 2.212, 2.311 2.206, 2.305 2.195, 2.317 91.4, 98.1

3.374 2.209 2.211 2.229 93.8

3.378 2.212 2.214 2.228 94.0

2.901 2.254 2.263 79.9 156.3

2.871 2.264 2.264 78.7 142.5

a

Scalar-relativistic MP2 bond distances in Å and angles in degrees, MP2 results for dinuclear clusters using non-relativistic Stuttgart– Dresden ECP60MHF in bold; basis sets: cc-pVTZ-PP (Au), cc-pVTZ (O, H) and cc-pV(T + d)Z (Cl); for position of atoms, see Fig. 5.

Au–Au distance;5 this feature stems from relativistic effects and is also apparent in the dinuclear Au cluster examined here. For instance, a comparison between non-relativistic MP2 RAu–Au values in [Au2Cl3(Cl2)]+ with those obtained using scalar-relativistic


MP2 calculations shows that shortening of the Au–Au distance ranges to 1.022 Å. If relativistic effects are included in [Au2Cl(Cl2)2]+ the Au–Au distance is reduced from 4.343 Å to 3.002 Å. The influence of relativistic effects on Au–Au distances may also be expressed in terms of the Au–Au interaction potential De.5,35 For instance, the Au–Au distances in Au2 (2.47 Å) and Au2Cl6 (3.401 Å) translate into values for De of 53.4 and 2.0 kcal mol 1, respectively, the former being in good agreement with an experimental value of 53.1 kcal mol 1.36 For an Au–Au distance of 3.329 Å in [Au2Cl5]+, De is 2.6 kcal mol 1; however, if the Au–Au distance is further reduced, as in [Au2Cl(Cl2)2]+ (3.002 Å) and [Au2Cl3(Cl2)]+ (2.639 Å), De shifts to 8.3 and 29.6 kcal mol 1. When relativistic effects in [Au2Cl(Cl2)2]+ and [Au2Cl3(Cl2)]+ are ignored, values of De reduce to 0.1 and 0.8 kcal mol 1, respectively. Interestingly, at 2.88 Å the Au–Au interaction energy in bulk gold (12.7 kcal mol 1) is comparable to the Au–Au interaction energy of [Au2Cl(Cl2)2]+ (8.3 kcal mol 1) but, at any event, still lower than in [Au2Cl3(Cl2)]+ (29.6 kcal mol 1). Table 4 summarizes the influence of microsolvation on the structure of all three dinuclear gold clusters. MP2 predicts that [Au2Cl5]+(H2O) is the global minimum structure, with [Au2Cl3(Cl2)]+(H2O) and [Au2Cl(Cl2)2]+(H2O) higher in energy, by 22 kcal mol 1 and 32 kcal mol 1, respectively. The most stable dihydrate maintains a structure similar to that of [Au2Cl5]+(H2O), with a second water molecule bound to firstshell water. The MP2 Au–Au distance in [Au2Cl5]+(H2O)2 is 3.378 Å, which is 0.049 Å longer than in [Au2Cl5]+; thus


View Article Online


Paper

PCCP

Fig. 6 Contour plots of the kinetic energy density for [Au2Cl(Cl2)2]+(H2O)n, [Au2Cl3(Cl2)]+(H2O)n [Au2Cl5]+(H2O)n with n = 0 and 1 in the Aua–Cl–Aub plane, showing regions of enhanced electron localization (orange and red) and basins denoting delocalization (blue); red arrows mark region of higher electron localization sandwiched between Aua and Aub in water-free and monosolvated [Au2Cl3(Cl2)]+.

microsolvation has a moderate effect on the geometry of [Au2Cl5]+ (see Table 4). Two additional dihydrate clusters were identified, [Au2Cl3(Cl2)]+(H2O)2 and [Au2Cl(Cl2)2]+(H2O)2, and these are around 28 and 47 kcal mol 1 less stable than [Au2Cl5]+(H2O)2. The structure of [Au2Cl3(Cl2)]+(H2O)2 maintains a close Au–Au distance at 2.531 Å, which is around 0.06 Å (or 3%) longer than the Au–Au distance in Au2 (2.47 Å). The Au–Au distance in [Au2Cl(Cl2)2]+(H2O)2 is 2.871 Å, but still well below the sum of the van de Waals radii of two gold atoms of 3.32 Å. The Au–Au interaction energies in [Au2Cl3(Cl2)]+(H2O)2 and [Au2Cl(Cl2)2]+(H2O)2 are relatively high, 43.2 kcal mol 1 and 13.1 kcal mol 1, respectively, and all are consistently higher than that of the corresponding water-free and monosolvated clusters. As seen from LOL plots for Au2Cl5+ (Fig. 6a) and Au2Cl5+(H2O) (Fig. 6d), solvation does not affect the bonding pattern in this cluster, instead a continuous region of electron delocalization, encompassing both Au centers and bridging Cl atoms, extends over the whole cluster. LOL contours close to the Au–Au contact in [Au2Cl3(Cl2)]+ (Fig. 6c) and [Au2Cl3(Cl2)]+(H2O) (Fig. 6f), on the other hand, show a region of enhanced electron

localization, marked as a green basin sandwiched between Au atoms, and is indicative of an increasingly stronger Au–Au interaction among these clusters. The remaining LOL domains (from light to dark blue) in water-free and solvated [Au2Cl3(Cl2)]+ are associated with delocalized electrons, and characteristic of multicenter bonding. In the case of the ion-neutral cluster [Au2Cl+(Cl2)2], LOL maps (Fig. 6b and e) reveal regions with enhanced delocalization bridging the Au2Cl+ ion to both Cl2 ligands, the latter displaying red sickle-like regions in the vicinity of Cl atoms characteristic of lone pairs. Given the high stability of hydrated dinuclear gold in ESI-MS experiments, it is instructive to compare MP2 values of DH, DS and DG for Cl2 loss from [Au2Cl(Cl2)2]+ and [Au2Cl3(Cl2)]+ and examine the effect that microsolvation has on these energy differences. Table 5 presents a summary of MP2/VTZ values of DH, DS and DG for (i) the dissociation of (H2O)n[Au2Cl5]+ (R4), that is, (H2O)n[Au2Cl5]+ - (H2O)nAuCl2+ + AuCl3, (ii) the stepwise loss of Cl2 from (H2O)n[Au2Cl(Cl2)2]+ ((R5) and (R6), see Table 5), and (iii) the loss of Cl2 from (H2O)n[Au2Cl3(Cl2)]+ (R7). The MP2 predicted DH value for (R4) is 64.5 kcal mol 1, which is comparable to the MP2 level dissociation energy for Au2Cl6 - 2AuCl3 (67.3 kcal mol 1) [Hargittai et al., 2001]. Addition of one H2O molecule (n = 1) increases the endothermicity of (R4) by 2 kcal mol 1, while inclusion of a second H2O molecule (n = 2) leads to a decrease in the endothermicity (48.8 kcal mol 1), which is due to thermochemical effects of water shell-filling about (H2O)2AuCl2+. The free energies of (R4) listed in Table 5 further show that these are too endergonic for dissociation of (H2O)n[Au2Cl5]+ to proceed at any meaningful level in the ESI-MS experiments. The next set of reactions to be considered are stepwise Cl2 expulsion processes from (H2O)n[Au2Cl(Cl2)2]+ ((R5) and (R6)) and (H2O)n[Au2Cl3(Cl2)]+ (R7). In the case of (H2O)n[Au2Cl(Cl2)2]+ with n = 0, both the first (R5) and second Cl2 expulsion reactions (R6) are endothermic by 35.1 and 36.4 kcal mol 1, respectively. Inclusion of a water molecule reduces the endothermicity of (R5) to 0.2 kcal mol 1, indicating that the first ligand switching reaction in (R5) is energetically neutral. The reaction energy for (R5) with n = 2 is predicted to have considerable exothermicity ( 9.7 kcal mol 1), and thus the presence of two water molecules favors Cl2 loss, yielding a stable (H2O)2[Au2Cl(Cl2)]+ cluster. Values of DG for reactions (R5) and (R6) become increasingly exergonic with increasing solvation number, indicating that the Cl2 loss in (H2O)n[Au2Cl(Cl2)2]+ would occur at 298 K. Reaction (R6) involves

Table 5 Effects of microsolvation on cluster dissociation and Cl2 expulsion enthalpies (DH)a, entropies (DS)a and free energies (DG)a in [Au2Cl(Cl2)2]+(H2O)n, [Au2Cl3(Cl2)]+(H2O)n and [Au2Cl5]+(H2O)n (n = 0–2)

n=0 Reaction (H2O)n[Au2Cl5] - (H2O)n[AuCl2] + AuCl3 (R4) (H2O)n[Au2Cl(Cl2)2]+ - (H2O)n[Au2Cl(Cl2)]+ + Cl2 (R5) (H2O)n[Au2Cl(Cl2)]+ - (H2O)n[Au2Cl]+ + Cl2 (R6) (H2O)n[Au2Cl3(Cl2)]+ - (H2O)n[Au2Cl3]+ + Cl2 (R7) +

+

n=2

DS

DG

DH

DS

DG

DH

DS

64.5 35.1 36.4 40.1

42.7 30.3 25.3 30.4

51.8 26.1 28.9 31.3

66.8 0.2 33.2 10.6

45.5 24.0 35.4 30.9

53.3 6.9 22.7 1.4

48.8 9.7 9.1 6.0

37.4 17.6 23.8 32.0

a MP2/VTZ dissociation enthalpies (kcal mol 1), entropies (cal mol corrected electronic energies.


n=1

DH

1

DG 37.7 14.9 2.0 3.5

K 1), free energies (kcal mol 1) obtained from differences between thermally-


View Article Online

PCCP

Paper

Table 6 Atomic charges and electron configurations of Au, Cl and O in dinuclear [Au2Cl]+, [Au2Cl(Cl2)]+, [Au2Cl(Cl2)2]+ and (H2O)2[Au2Cl(Cl2)2]+a

Species

Atom +


(H2O)2Au2Cl (Cl2)2

[Au2Cl+(Cl2)2]

[Au2Cl+(Cl2)]

Au2Cl+ H2O Au+ Au+3 Cl a

Aua Aub Cla Clb Clc Oa Ob Aua Aub Cla Clb Clc Aua Aub Cla Clb Aua Aub Cla O Au Au Cl

NPA charge 0.46 0.46 0.39 0.09 0.09 0.94 0.94 0.47 0.47 0.38 0.09 0.09 0.77 0.43 0.43 0.09 0.75 0.75 0.50 0.91 1.00 3.00 1.00

Electron configuration 6s0.565d9.786p0.10 6s0.565d9.786p0.10 3s1.863p5.503d0.03 3s1.863p4.983d0.05 3s1.863p4.983d0.05 2s1.742p5.193d0.01 2s1.742p5.193d0.01 6s0.625d9.776p0.14 6s0.625d9.776p0.14 3s1.873p5.473d0.03 3s1.863p4.983d0.05 3s1.863p4.983d0.05 6s0.285d9.896p0.05 6s0.645d9.766p0.16 3s1.893p5.513d0.02 3s1.863p4.983d0.05 6s0.305d9.896p0.06 6s0.305d9.896p0.06 3s1.933p5.543d0.02 2s1.742p5.153d0.01 6s05d10 6s05d8 3s23p5

MP2/cc-pVTZ theory level with ECP60MDF for Au.

expulsion of Cl2 from (H2O)n[Au2Cl(Cl2)]+, the dissociation enthalpies for (R6) being 33.2 kcal mol 1 (n = 1) and 9.1 kcal mol 1 (n = 2). A comparison between values of DH for (R5) and (R6) shows that inclusion of one water molecule does not significantly change the endothermicity of (R6). Solvation with a second water molecule, on the other hand, weakens the cation Cl2 interaction in (H2O)n[Au2Cl(Cl2)]+, lowering DH for (R6) to 9.1 kcal mol 1. The last reaction examined involves the removal of Cl2 from (H2O)n[Au2Cl3(Cl2)]+ (R7). When water is not present, the enthalpy of (R7) is predicted to be 40.1 kcal mol 1, and DH is made less endothermic (10.6 kcal mol 1) with n = 1. Inclusion of a second water molecule results in (R7) being less endothermic; however, the second water has a moderate effect on the dissociation enthalpy and lowers the endothermicity of (R7) by around 4 kcal mol 1 to 6.0 kcal mol 1. Table 6 summarizes results of NBO population and charge analyses of [Au2Cl]+, [Au2Cl(Cl2)]+, [Au2Cl(Cl2)2]+ and (H2O)2[Au2Cl(Cl2)2]+; the electron configuration of Au in [Au2Cl]+ is 6s0.305d9.896p0.06, and 6s0.285d9.896p0.05 for Aua and 6s0.645d9.766p0.16 for Aub in [Au2Cl(Cl2)]+. Attachment of Cl2 onto [Au2Cl(Cl2)]+ gives 6s0.625d9.776p0.14 for Aua and Aub and a corresponding charge on Au of +0.47. This electron configuration remains almost unaffected upon solvation of [Au2Cl(Cl2)2]+ with two water molecules; in other words, little charge exchange occurs between [Au2Cl(Cl2)2]+ and the surrounding water molecules. A review of NBO data in Table 6 reveals a significant transfer of charge from Cl to Au (0.25 e) in [Au2Cl]+ and this charge transfer correlates with an increase in the 6s orbital population, while the 5d orbitals of Aua and Aub are depopulated relative to Au+. The situation for [Au2Cl(Cl2)]+ is quite different: here the charge on Aua is predicted to be essentially the same (+0.78 e) as for [Au2Cl]+, while the charge


on Cl2-ligated Aub decreases, relative to [Au2Cl]+ from +0.75 to +0.43 e. Additional charge transfer occurs in [Au2Cl(Cl2)]+ upon ligation with a second Cl2 molecule, and the charge on Aua shifts from +0.78 to +0.47 e, while the charge on the bridging chlorine Cla atoms further decreases to 0.38 e. In general, the stepwise ligation of [Au2Cl]+ with Cl2 results in a progressive increase in the gold 6s orbital occupation and depopulation of the 5d orbitals, at the same time Au 6p contributions are negligible. The origin of this behavior can be traced to the high electrophilicity of the Au 6s orbital and is rooted in relativistic effects.37 In the case of [Au2Cl(Cl2)]+ and [Au2Cl(Cl2)2]+ the high electrophilicity of the 6s orbital of Au promotes partial electron transfer from Cl2 to gold in [Au2Cl(Cl2)n]+ resulting in a high per-ligand binding energy, being 35.3 kcal mol 1 (n = 1) and 36.6 kcal mol 1 (n = 2) for the Cl2 ligation reaction [Au2Cl(Cl2)n 1]+ + Cl2 - [Au2Cl(Cl2)n]+, respectively. The magnitude of this relativistic effect is large, and is easily understood by comparison with non-relativistic results. When relativistic effects are neglected in [Au2Cl]+ the Au–Au distance increases from 2.831 Å to 4.761 Å and, accordingly, the non-relativistic MP2 predictions for the Au–Au distance are 4.562 Å for [Au2Cl(Cl2)]+ and 4.344 Å for [Au2Cl(Cl2)2]+. Accordingly, the relativistic contribution to the stepwise Cl2 binding energy is significant, being 23.4 kcal mol 1 for n = 1 and 24.8 kcal mol 1 for n = 2, which in both cases constitutes around 70% of the total Cl2 binding energy.

V. Conclusion Positive-ion mode electrospray ionization of aqueous AuCl3 solutions leads to the formation of mononuclear [AuCl2 xOHx]+(H2O)n (x = 0,1), [AuCl2]+(HCl)2(H2O)n and dinuclear clusters [Au2Cl5 xOHx]+(H2O)n (x = 0,1). The analysis of concentration effects on the ESI mass spectra of mono- and dinuclear gold reveals an upward shift in the abundance of the dinuclear Au cluster fraction with increasing AuCl3 concentration, in particular [Au2Cl5]+(H2O)n. In general, results from these concentration-dependent ESI-MS experiments can be applied to characterize gold cluster species potentially present in aqueous solution, and by doing so, provide complementary speciation data to that derived from solution-phase techniques. Structures and dissociation energies in low-energy gold clusters have been determined by MP2 and CCSD(T) theory. These data show that an ion–molecule complex [Au+(Cl2)](H2O)n with n r 2 is the dominant mononuclear Au complex, in which gold exists as formal Au(I), while solvated mononuclear species with n Z 2 are predicted to be more strongly-bound [AuCl2]+(H2O)n complexes containing formal Au(III). The influence of microsolvation on the dissociation of both [Au+(Cl2)](H2O)n and [AuCl2]+(H2O)n has been probed using MP2 and CCSD(T) theory, and in the case of the former, Cl2 expulsion was found to be thermodynamically nearly neutral when microsolvation effects are considered. In the case of [AuCl2]+(H2O)n, the stepwise Au reduction reactions, of the form [AuCl2 x]+(H2O)n [AuCl1 x]+(H2O)n + Cl, are predicted to be largely endergonic


View Article Online


Paper

and vary in magnitude because of solvation shell-filling effects in [AuCl2]+, [AuCl]+ and [Au]+. MP2 calculations show that the lowest-energy dinuclear Au cluster is a non-covalently interacting trimer [Au2Cl(Cl2)2]+, consisting or a central digoldchloronium ion bound end-on to two dichlorine ligands. The charge states of individual Au atoms in the clusters [Au2Cl]+, [Au2Cl(Cl2)]+, [Au2Cl(Cl2)]+ and (H2O)2[Au2Cl(Cl2)]+ have been determined by calculating the charges transferred from the Cl2 and H2O ligands to Au together with the charges carried by the Au atoms. Results from NBO analyses of Au+, [Au2Cl]+ and (H2O)2[Au2Cl(Cl2)]+ show that the charge states of Au atoms in the former shift from 1.000, 0.7484 to 0.4632e upon ligation by two Cl2 and H2O molecules, respectively. The interactions between Au centers in [Au2Cl(Cl2)2]+(H2O)n and [Au2Cl3(Cl2)]+(H2O)n have been examined using relativistic and non-relativistic MP2 calculations and these data, together with localized orbital locator maps, show that microsolvation leads to a buildup of a region with enhanced electron localization between Au centers, but also leads to electron delocalization between the Au2Cl+ and [Au2Cl3]+ cores and Cl2 ligands. This solvation-induced localization of electrons is further accompanied by a shortening of the Au–Au contacts and an increase in the Au–Au interaction energy and, thus, explicit consideration of the solvation shell is imperative for an accurate description of the stability of dinuclear Au clusters.

Acknowledgements Computing time was provided by the HKU High Performance Computing (HPC) and Grid Computing Centers. This project was supported in part by a Hong Kong UGC Special Equipment Grant (SEG HKU09) and General Research Fund HKU 703211 from the Research Grants Council of Hong Kong.

References ´ffy, M. Kolonits, A. Schulz, T. M. Klapo ¨tke and 1 B. Re M. Hargittai, J. Am. Chem. Soc., 2000, 122, 3127. ´ffy, A. Schulz and M. Kolonits, J. Am. 2 M. Hargittai, B. Re Chem. Soc., 2001, 123, 1449. ¨der, R. Brown, P. Schwerdtfeger, X. B. Wang, 3 D. Schro X. Yang, L. S. Wang and H. Schwarz, Angew. Chem., Int. Ed., 2003, 42, 311. 4 I. J. Blackmore, A. J. Bridgeman, N. Harris, M. A. Holdaway, J. F. Rooms, E. L. Thompson and N. A. Young, Angew. Chem., Int. Ed., 2005, 44, 6746. 5 P. Schwerdtfeger, A. E. Bruce and M. R. M. Bruce, J. Am. Chem. Soc., 1998, 120, 6587. 6 P. Schwerdtfeger, Angew. Chem., Int. Ed., 2003, 42, 1892. ¨, Angew. Chem., Int. Ed., 2004, 43, 4412. 7 P. Pyykko ¨, Inorg. Chim. Acta, 2005, 358, 4113. 8 P. Pyykko


PCCP

ˆ and M. A. Omary, 9 K. A. Barakat, T. R. Cundari, H. Rabaa J. Phys. Chem. B, 2006, 110, 14645. ¨, Chem. Soc. Rev., 2008, 37, 1967. 10 P. Pyykko 11 J. Zhang, Z. He, W. Li and Y. Han, RSC Adv., 2012, 2, 4814. 12 J. A. Peck, C. D. Tait, B. I. Swanson and G. E. Brown Jr., Geochim. Cosmochim. Acta, 1991, 55, 671. 13 F. Farges, J. A. Sharps and G. E. Brown Jr., Geochim. Cosmochim. Acta, 1993, 57, 1243. 14 A. Usher, D. C. McPhail and J. Brugger, Geochim. Cosmochim. Acta, 2009, 73, 3359. 15 D. M. Sherman, Geofluids, 2010, 10, 41. ¨der, L. Ducha ćkova ´, J. Tara ´bek, M. Karwowska, 16 D. Schro ´k and P. Slavıćek, J. Am. Chem. K. J. Fijalkowski, M. Onca Soc., 2011, 133, 2444. ¨der and N. G. Tsierkezos, Int. J. Mass 17 V. ˇ S´ adek, D. Schro Spectrom., 2011, 304, 9. ¨der, Collect. Czech. Chem. Commun., 2011, 76, 351. 18 D. Schro ˇabka, N. Tsierkezos ćkova ´, J. Roithova ´, P. Milko, J. Z 19 L. Ducha ¨der, Inorg. Chem., 2010, 50, 771. and D. Schro 20 M. A. Rawashdeh-Omary, M. A. Omary and H. H. Patterson, J. Am. Chem. Soc., 2000, 122, 10371. ¨der, Phys. Chem. Chem. Phys., 2012, 14, 6382. 21 D. Schro 22 K. A. Peterson and C. Puzzarini, Theor. Chem. Acc., 2005, 114, 283. 23 T. H. Dunning Jr., J. Chem. Phys., 1989, 90, 1007. 24 T. H. Dunning Jr., K. A. Peterson and A. K. Wilson, J. Chem. Phys., 2001, 114, 9244. 25 D. Feller, E. D. Glending and W. A. de Jong, J. Chem. Phys., 1999, 110, 1475. 26 H. L. Schmider and A. D. Becke, THEOCHEM, 2000, 527, 51. 27 H. L. Schmider and A. D. Becke, J. Chem. Phys., 2002, 116, 3184. 28 D. Figgen, G. Rauhut, M. Dolg and H. Stoll, Chem. Phys., 2005, 311, 227. 29 P. Schwerdtfeger, M. Dolg, W. H. E. Schwarz, G. A. Bowmaker and P. D. W. Boyd, J. Chem. Phys., 1989, 91, 1762. 30 N. O. J. Malcolm and P. L. A. Popelier, J. Comput. Chem., 2003, 24, 437. ´s, C. Fortun õ, S. Iba ń ˜ ez, A. Martıń, P. Mastrorilli, 31 J. Fornie V. Gallo and A. C. Tsipis, Inorg. Chem., 2013, 52, 1942. 32 A. C. Tsipis, Organometallics, 2012, 31, 7206. 33 A. C. Tsipis and A. V. Stalikas, Inorg. Chem., 2012, 52, 1047. 34 H. Schmidbaur, A. Hamel, N. W. Mitzel, A. Schier and S. Nogai, Proc. Natl. Acad. Sci. U. S. A., 2002, 99, 4916. ¨ and F. Mendizabal, Inorg. Chem., 1998, 37, 3018. 35 P. Pyykko 36 K. P. Huber and G. Herzberg, Molecular Spectra and molecular structure IV: Constants of diatomic molecules, Van Nostrand, New York, 1979. 37 P. Schwerdtfeger, P. D. Boyd, S. Brienne and A. K. Burrell, Inorg. Chem., 1992, 31, 3411.


Absorption mode FTICR mass spectrometry imaging.

Peptide fragmentation products in mass spectrometry probed by infrared spectroscopy.

Surface Accessibility and Dynamics of Macromolecular Assemblies Probed by Covalent Labeling Mass Spectrometry and Integrative Modeling.

Structure of full-length p53 tumor suppressor probed by chemical cross-linking and mass spectrometry.

Interaction of angiotensin peptides and zinc metal ions probed by electrospray ionization mass spectrometry.

Comprehensive list of metabolites measured by DI-FTICR mass spectrometry in thyme plants with contrasting tolerance to drought.

D exchange mass spectrometry and isothermal titration calorimetry.

Top-down MALDI-in-source decay-FTICR mass spectrometry of isotopically resolved proteins.

Lithium formate ion clusters formation during electrospray ionization: Evidence of magic number clusters by mass spectrometry and ab initio calculations.

Different modes of lipid binding to membrane proteins probed by mass spectrometry.

Conformational changes in proteins probed by hydrogen-exchange electrospray-ionization mass spectrometry.

Solvent-induced conformational changes of polypeptides probed by electrospray-ionization mass spectrometry.

Gold finger formation studied by high-resolution mass spectrometry and in silico methods.

cAMP-dependent protein kinase (PKA) complexes probed by complementary differential scanning fluorimetry and ion mobility-mass spectrometry.

Molecular and structural characterization of dissolved organic matter during and post cyanobacterial bloom in Taihu by combination of NMR spectroscopy and FTICR mass spectrometry.

mass spectrometry history, theory, design optimization, simulations, and applications.

Spatial organization of Pseudomonas aeruginosa biofilms probed by combined matrix-assisted laser desorption ionization mass spectrometry and confocal Raman microscopy.

mass spectrometry.

Gold-bismuth clusters.

Next-generation technologies for spatial proteomics: Integrating ultra-high speed MALDI-TOF and high mass resolution MALDI FTICR imaging mass spectrometry for protein analysis.

Photodissociation dynamics of allyl chloride at 200 and 266 nm studied by time-resolved mass spectrometry and photoelectron imaging.

Prediction, Detection, and Validation of Isotope Clusters in Mass Spectrometry Data.

Structures of cycloserine and 2-oxazolidinone probed by X-ray photoelectron spectroscopy: theory and experiment.

Traveling wave ion mobility mass spectrometry and ab initio calculations of phosphoric acid clusters.