Sulfur ligand mediated electrochemistry of gold surfaces and nanoparticles: what, how, and why

Gold surfaces are widely used in electrochemistry whilst gold nanoparticles have very many uses, with both the surfaces and the particles often being protected by sulfur-bound organic ligands. The ligands not only provide chemical stability but also directly participate in many desired processes. This review considers the diversity of known atomic structures for gold-sulfur interfaces, how these structures facilitate a diversity of mechanisms in electrochemical applications, and why this is possible based on recent advances in the basic understanding of the electronic structure of gold-sulfur bonds. Believed once to be Au(I)-thiolate in character and hence distinctly different to physisorbed thiols and disulfides, chemisorbed bonds are shown to be Au(0)-thiyls instead. A wide range of in-situ STM electrochemical and other data is interpreted from this perspective.

To prevent clogging and the effects of harmful reagents, AuNPs have to be protected by selfassembled monolayers (SAMs), a bit like tea in a teapot brought to preserve taste and warmth by a tea cosy [15][16][17].Similarly, Au-surfaces can be protected or functionalized by SAMs, with organic thiol compounds, RSH having acquired a unique status [1,[18][19][20][21][22] as these electronically "soft" ligands bind well to electronically "soft" surface Au-atoms [23].The functionalizing residues R can be hydrophobic or hydrophilic, electrostatically charged or neutral, or spatially large or small [1,18].The multifarious RSH residues can therefore impose almost tailor-made Au-surface properties, reflected in a broad variety of molecular and biomolecular adsorption and electrochemical electron transfer (ET) patterns, in which the surface Au-S bond is the central electronic contact.This review considers the basic elements of the gold-sulfur links that make up these critical interfaces: what are they, how can they be put to work, and why they facilitate so many ancient and modern applications.

2.
What is the structure of sulfur-bonded ligand-protected gold surfaces and nanoparticles?
The atomic structure is a common element of gold-sulfur interfaces with properties independent of intended function, be this electrochemical in nature or otherwise.Indeed, a wide variety of alkanethiol-bound monolayers or just isolated single molecules have been used [24,25], including molecules with sophisticated electrochemical function [26][27][28][29] intended, e.g., for in molecular scale electronics [9].To answer the question "What?", we focus on simple systems without explicit function that disclose critical structural questions.

•
In spite of their similar structures, the butanethiols display widely different adsorption modes, disentangled in minute structural detail.Mapping of other functionalized alkanethiols have been resolved in comparable detail [18].

•
The butanethiol packing at high coverage is determined primarily by Au-S bonding and lateral intermolecular interactions.Solvation is a determinant for functionally charged alkanethiols such as cysteamine [18,36] and cysteine [18,30].

•
The Au-S bond for even the butanethiols of modest structural sophistication leads to multifarious Au-surface packing.Binding of straight chain 1-butanethiol is via molecular extraction of an Au-atom ("atomic gold mining") from the Au(111)-surface, leaving a surface vacancy, followed by sideway binding of two adsorbate molecules [24,32,33,35,40].Surface vacancies typically coalesce into surface pits.There is no space for this binding mode for the bulky tert-butanethiol 2-methyl-2-propanethiol which is the only butanethiol that binds solely on a flat Au (111)-surface.
The individual (R) and (S) enantiomers bind via "atomic mining", i.e. by Au-atom extraction from the Au (111)-surface.The racemic mixture binds in both a low-density (coverage 25 %) and a high-density (coverage 30 %) phase.While the low-density phase is bound by atomic gold mining, to squeeze in more butanethiol molecules in the high-density phase, both atomic gold mining and direct Au-S adsorption operate, Fig. 2.This is highly unusual, with both binding modes operating on the same surface.

How can sulfur-bonded ligand-protected gold surfaces and nanoparticles be put to work?
The importance of molecules bound to Au-electrodes via pendant S-linker groups is strongly 3 [26,27,29,[47][48][49].The viologens and tetrathiafulvalenes shown [26][27][28], as well as the shown transition metal complexes of ruthenium, osmium and cobalt [29], provide paradigms describing this field.These molecules display in-situ scanning tunneling current/overpotential spectroscopy (STS) consistently rationalized as two-step sequential electron transfer (ET) between tip and working electrode via the molecule.Theoretical frameworks for understanding of Au-S molecular redox systems such as these have so far rested on phenomenological, although comprehensive views and models [48].These have met with success for organic redox molecules [26,27], transition metal complexes [29], redox metalloproteins [34,50,51], and Au-S bound DNA based molecules [52].

Fig.3
As particular examples, monolayer voltammetry of redox metalloproteins and DNA-based molecules has evolved [34,48,[50][51][52][53][54][55][56][57].Much current effort is on applied aspects such as biosensing [6,12,13], electrocatalysis [2,3,5,6,58,59], biofuel cells [10,11], and bioelectrochemistry on living systems [60].However, physisorptive binding of metalloproteins via -SH or -S-S-groups rather than chemisorbed Au-S bonds involving wild type proteins with retained voltammetric activity is actually rare.The nature of the bonding is therefore just as important as the atomic configuration in understanding "how" electrochemical devices function.Still another notable case is yeast cytochrome c [67,68].The linking chemisorbed -SH group in this case is, however, located slightly below the protein surface, and both conformational reorganization and a low-spin/highspin transition in the axial heme ligand structure in the protein accompanies binding to the surface.Also, we note that human insulin adsorbs on all three low-index Au-surfaces, Au(111), Au(100) and Au(110) via intrinsic disulfide groups, mapped to single-molecule resolution by in situ STM [65].How sulfur-bounded ligand-protected surfaces can be put to work is therefore quite complex.

Why does the nature of the Au-S bond make sulfur-bonded ligand-protected gold surfaces and nanoparticles so versatile?
The questions of "what" and "how" thus both lead to significant questions concerning how it is that sulfur-bonded ligand-protected gold surfaces are so versatile.The Au-S bond can take on vastly different chemisorbed geometries and coordinations, and can still function even through much weaker physisorbed connections.There must be a unifying bonding feature that allows for this diversity, explaining "why" Au-S bonded molecular adsorption and molecular function works so well and can lead to practical devices.
These results are consistent with the in-situ STM & AFM studies [71] and provides a detailed understanding of structure and function [40].For more than 30 years consensus has been that thiol adsorption involves a redox process The reverse process is electrochemical reductive desorption, RS − -Au(I) + e − → Au(0) + RS − , with the electrochemical signals reflecting highly sensitively the general state of the SAM.
Au(I)-thiolate bonds are well known in molecular Au-S complexes [72,73], forming a prime example of coordination between electronically "soft" metal ions and "soft" ligands [23].The properties of Au atoms in planar Au surfaces are, however, quite different from those in simple compounds.A new analysis, supported prodigiously by computations and data shows instead that surface Au-S binding holds a dominant element of Au(0) and thiyl radical RS• character [40], character not too different from that found in physisorptive interactions of surfaces to thiols and disulfides.Direct contact of the thiolate anion with Au surfaces leads instead to etching or even dissolution [74].This is a real physical distinction as the chemical and spectroscopic properties of Au(I)-thiolate and Au(0)-thiyl species are quite different.
A recent study [75] of Fe-porphyrin complexes with a Fe-Cys axial ligand that mimic the heme protein cytP450 offers a striking illumination.The Fe-Cys bond involves the equilibrium Fe(III)--SR  Fe(II)-•SR that resembles the resonance scenario Au(I)- However, this equilibrium depicts real and tunable valence tautomerism between two different states distinguishable by significantly different spectroscopic properties, while Au(I)thiolate/Au(0)-thiyl is instead viewed as presenting single resonance scenarios with different electronic charge distributions in the electronic ground and excited states.
The electronic structure of Au is 5d 10 6s 1 .Common belief has been that the Au-S bond to Au surfaces bond is covalent and dominated by the open shell 6s and 3p orbitals of Au and sulfur, with the Au 5d orbitals in a less prominent role [76].However, strong Au-Au interactions within surface and in Au-Au compounds [72,73,77,78] shifts the 6s band away from the Au Fermi level and the 5d band [69], causing the filled 5d orbitals to take over bonding to the sulfur atom.
Au d

S p
The bonding in such circumstances is often dominated by van der Waals dispersion forces as the interaction of a low-lying doubly occupied orbital with a high-lying one does not lead to significant covalent or ionic bonding [40].In gold chemistry this is known as the "aurophilic effect" [72,79].
Binding schemes such as these are, however, limiting cases, mixed by s-d hybridization, polarization effects, multifarious coordination, etc.All computations and sophisticated experiments show, that Au(0)-thiyl is a tunable, highly significant structural component that typically dominates the binding of SR groups to surfaces.Related dispersion forces are also important for the binding of other electronically "soft" ligands like nitrogen bases and disulfides to Au surfaces, including as well situations where Au atoms are mined out of the surfaces [47].
This understanding rationalizes the wide range of geometrical structures at interfaces observed, as well as the robustness of their electrochemical properties.

Conclusions and perspectives
The Au-S bond on planar and AuNP Au surfaces is an important element that ties together binding and packing of a wealth of complex and less complex S-linked molecules and biomolecules on both electrochemical and non-electrochemical surfaces.Understanding the properties of this structural element is therefore paramount.On the basis of recent detailed analysis [40], it is concluded that the Au-S link holds a dominant Au(0)-thiyl character resting on strong van der Waals interactions and the aurophilic effect, in addition to secondary covalent and ionic Au-S features.There is thus a similarity and continuity between chemisorption and physisorption of sulfur compounds to gold.Aspects [40] of the Brust-Schiffrin core nanoparticle synthesis [16,17] can then be understood, as well the results of real time imaging on planar Auelectrode surfaces as reported for 1-propanethiol [31] and cysteamine [36], anticipated also for more complex molecular SAMs [80].
The nature of the Au-S link as a radical may indicate reactivity in addition to electrochemical reductive desorption.Mostly, however, radical character of the sulfurs is typically lost owing to long-range superexchange coupling that couples sulfur atoms together and diffuses the radical

Fig. 1 Fig. 2
Fig.1 illustrated in electrochemical STM and break-junction experiments.Examples of more sophisticated Au-S based redox molecules characterized at single-molecule resolution by in-situ STM technology and voltammetry on single-crystal electrodes are shown in Fig.

[ 63 ,
64].An example of a robust single-molecule-based bioelectrochemical electronic behavior based on chemisorbed Au-S head-group interactions is the four-α-helix bundle bacterial protein cyt b 562 (Escherichia coli) shown in Fig. 4 [50].Mutations in this protein could introduce cysteine (Cys) residues at chosen locations on the outside of the protein.The success of the approach is indicated by the observation of almost ideal protein film voltammetry on singlecrystal Au (111)-electrode surfaces; also in situ STS displayed the expected tunnelingcurrent/overpotential STS.The results are intriguing, however, in that single or multiple attachments at different positions on the ellipsoidal protein is naively expected to give quite different redox behavior and different ET pathways through the protein yet little variation was actually observed [50].In addition, the observed cyt b 562 conductivity was almost ten times higher than nearly all reported in-situ redox single-molecule conductivities.Coherent multi-ET has been suggested as a rationale [50], but the question as to the nature of the possibly strained Au-S bonds involved is also raised.

Fig. 1 .
Fig.1.Examples of functionalized alkanethiols studied by single-crystal voltammetry and in situ STM.Top row: Molecular structures.Bottom rows: High-resolution in situ STM images.The four images to the left are the molecules indicated, on Au(111).The four images to the right show cysteine on Au(110), with sub-molecular resolution images.Adapted from ref 48.

Fig. 2 .
Fig. 2. Shown on the left are observed STM images of the butanethiol family on Au(111), with their surface-cell parameters, coverages, and interpreted atomic structures shown on the right.Two copies of each unit cell are shown in each structure.

Fig. 4 .
Fig.4.Metalloproteins linked to Au-surfaces either via alkanethiol SAMs or directly via surface cysteine residues, studied by single-crystal voltammetry and in-situ STM.Top row: Schematic view of azurin (left) and cyt b 562 (right) on the Au(111)-surfaces.Bottom row: In situ STS of the proteins indicated.Adapted from ref 50 and 66.
or c61,62] tail group of which provides a non-covalent link to the protein[48,61,62].The reactivity of the immobilized redox metalloproteins is, however, exceedingly sensitive to the structural details of the SAM.This is strikingly illustrated by the voltammetry and in-situ STM and AFM studies of the blue Cu-enzyme nitrite reductase on single-crystal Au(111) electrode surfaces, modified by a broad variety of aliphatic and aromatic thiol SAMs The fundamental electrochemical ET processes are, however, also in focus, and structural and electrochemical mapping of single redox metalloprotein and DNA-based molecules, using techniques and new theory of in-situ STM/STS, conformational protein structural changes are insignificant.Chemisorbed Au-S bond formation is a core link in protein immobilization, but mostly via SAMs of functionalized alkanethiols, the remote, hydrophobic