Electrochemistry of metal adlayers on metal chalcogenides

Electrodeposition of metal adlayers on semiconductor metal chalcogenides (CdSe, CdS, PbTe, PbSe, PbS, Bi2Te3) is reviewed. Cathodic underpotential deposition of metal adlayer on metal chalcogenide is the electrochemically irreversible surface limited reaction. The irreversibility of the upd increases in the row from tellurides to selenides and further to sulfides. The underpotential shift on chalcogenide nanoparticles increases with particle size. Metal upd on chalcogenides is applied as a means of measurement of electroactive surface area of chalcogenide electrodes. The method is especially advantageous for multicomponent systems with other component not supporting upd, such as CdSe-TiO2, CdSe-ZnO. Differences of voltammetric profiles of Pb upd on Bi2Te3 and Te are applied for detection of Bi2Te3 surface contamination by elemental tellurium. The further tasks in the electrochemistry of metal adlayers are their incorporation as interlayers in layered chalcogenides and electrodeposition of superlattices. Graphical abstract Graphical abstract


Introduction
Electrodeposition of a metal atomic layer (adlayer) on a foreign substrate proceeds under thermodynamic conditions different from those of the electrodeposition on the same metal substrate. The difference which results from effect of substrate is characterized by the "underpotential shift" ΔE upd of the adlayer deposition potential vs. corresponding Nernst potential E(Me n + /Me b u l k ), where Me b u l k is bulk metal. Underpotential deposition (upd) of metal adlayers on metals and their anodic oxidation have been comprehensively reviewed [1][2][3][4][5][6][7], a less elucidated is the electrochemistry of metal adlayers on nonmetals, such as chalcogens [8,9], while the electrochemistry of metal adlayers on metal chalcogenides misses review literature, despite great significance of metal adatoms, submonolayers, and adlayers as intermediates in electrodeposition of metal chalcogenides [10] and the lately developed electrodeposition of metal-metal chalcogenide superlattice structures [11]. Metal adlayer on semiconductor metal chalcogenide is also an interesting probe object for characterization of size dependences of physical and physicochemical properties of semiconductor nanoparticles [12,13]. Due to correlations of underpotential shift with the semiconductor nanoparticle size, the routine measurement of quantum dot (QD) size by optical spectroscopy may be substituted by measurement of underpotential shift [12], when in situ optical measurement is less convenient, e.g., in multicomponent nanostructured systems. Selectivity of metal adlayer electrodeposition on metal chalcogenide components of chalcogenide-oxide heterostructures is also a useful feature which allows measurement of electroactive surface area of the chalcogenide spectral sensitizer in the sensitized wide bandgap metal oxide photoanodes [14].
The electrochemistry of metal adlayers on chalcogenides has been applied in works dedicated to layer-by-layer electrochemical assembly of metal chalcogenide materials [15][16][17][18][19][20][21][22][23]. However, most of the works present electrochemistry just of initial steps of chalcogenide assembly. This minireview is based on own research of the authors with the focus on the results published in the last 5 years. on metal chalcogenides are the very characteristic examples of solid state electrochemistry objects which depend critically on electroanalytical methods. Especially complicated for nonelectrochemical analysis are the metal adlayers on the chalcogenides of the same metal (Pb ad -PbTe, Cd ad -CdSe, etc.), as they differ just in the easily mutable chemical state from metal atoms of the substrate and this determines the use of highly sensitive electrochemical techniques for characterization of the electrochemically generated adlayers.
The upd, as well as the anodic oxidation of the electrodeposited adlayer, is the surface limited reaction which fades fast at a fixed potential, so the analytical current in stationary voltammetry is equal to zero. For this reason, potentiodynamic rather than stationary versions of voltammetry and impedance spectroscopy were used for electrochemical characterization of the adlayers. While the advantage of cyclic voltammetry (CV) is self-evident, minor explanation may be required for the use of potentiodynamic instead of stationary impedance spectroscopy for frequency response analysis of the cathodic deposition and anodic oxidation of adlayers. Figure 1 shows two limiting cases of potentiodynamic voltammetric profiles of upd with the corresponding equivalent electric circuits derived from impedance spectroscopy. A pair of perfectly reversible surface limited reactions (the cathodic deposition and the anodic oxidation of an adlayer) gives mirror symmetric cathodic and anodic peaks (Fig. 1a), while electrochemical irreversibility of upd results in the separation of the cathodic and anodic peaks (Fig. 1b). The separation of the peaks even in absence of significant mass transport discloses restrictions for reversible charge transfer, while the mirror symmetric voltammogram indicates immediate readiness of the adlayer for releasing the captured electron into the electric circuit upon the potential scan reversal, i.e., a behavior typical to electric capacitor. The corresponding capacitance C a in the Faradaic branch of the equivalent circuit of electrochemically reversible upd was found with stationary and potentiodynamic electrochemical impedance spectroscopy in various reversible upd processes on metal substrates (Pb, Cu, Bi, Ag upd on gold, [26][27][28], Cu upd on Pd [29], etc.). The capacitance C a results from oscillation of metal adatom coverage at ac probing [26,30,31]. C a is typically much higher than the double layer capacitance C d l . The name "pseudocapacitance" is often used in literature for distinguishing C a from 'true' capacitance C dl of the double layer. Despite the C a originates from the electrochemical reaction and belongs to Faradaic branch of the equivalent circuit, this pseudocapacitance is, nevertheless, a true capacitance in terms of frequency response analysis and should not be mixed with "pseudocapacitances" which have been derived lately from CV and chronopotentiometry in a great number of p u b l i c a t i o n s a b o u t s u p e r c a p a c i t o r s . T h e l a t t e r "pseudocapacitances" are usually not capacitances in terms of frequency response analysis and the use of Farad unit is inappropriate for their characterization (a detailed discussion of this issue was presented in [32]).
Metal upd on chalcogens and on metal chalcogenides, contrary to the electrochemically reversible upd, tend to the voltammetric profile of the kind shown in Fig. 1b [26]. No reverse reaction proceeds in this case at the potential of the upd peak, so keeping the system at the potential of upd in the stationary state results not only in zero direct current response of the upd but also in impossibility of obtaining any alternating current response of either forward or back reaction. That is why impedance spectrum has to be acquired in the potentiodynamic mode with inevitable truncation of low frequencies to provide sufficiently high scan rate. The other consequence of nonexistence of equilibrium between electrochemically irreversible forward and back reactions is the complexity of the underpotential shift ΔE upd measurement in such systems [9,33]. When the quasiequilibrium potential shown by vertical dashed line on the right in Fig. 1b is not well defined as a boundary between cathodic and anodic peaks in CV, it can be more accurately obtained as intercept of potential dependences of inverse charge transfer resistances of forward and back reactions derived with potentiodynamic electrochemical impedance spectroscopy (PDEIS) [9]. Peculiarities of PDEIS application for characterization of nonstationary systems were discussed in [26].

Underpotential shift and adlayer-substrate interaction
The underpotential shift ΔE upd characterizes favorability of adatoms deposition on a foreign substrate compared to the deposition on a same metal substrate and thus may serve as a merit of metal-substrate (M-S) interaction strength [6]. ΔE upd is determined by several factors [34], which are often complex to calculate: binding energies in the adlayer and between adatoms and the substrate, the influence of local surface defects of the substrate, the effects of solvent and anions. Despite complexity of ΔE upd prediction, some empirical correlations were found between ΔE upd and characteristics of the deposit and substrate. For instance, the underpotential shift in metal upd on metal typically shows correlation with differences of work functions of the substrate and the deposited metal [1]. On the contrary, no correlation with the differences of work functions was observed in metal upd on tellurium [9] ( Fig. 2a). Alternatively, the underpotential shift in metal upd on tellurium shows correlation with free energy of metal telluride formation [9] (Fig. 2b). The latter correlation is the indication of metal-chalcogen covalent interaction effect in metal adlayer growth on tellurium.
Besides the covalent adatom-substrate bonding, electron energy band structure of a nonmetallic substrate affects the upd shift, e.g., no electrodeposition of metal adatoms on a wide-bandgap semiconductor p-selenium was observed in the potential region of the expected upd under dark condition; however, the electrodeposition proceeds on illuminated p-Se and the photoelectrochemically deposited adlayer gives characteristic oxidation peak in the anodic scan [35]. Underpotential shift in upd on nanoparticles of CdS, CdSe, and some other chalcogenides is controlled by LUMO energy which is dependent on the semiconductor nanoparticle size, so the upd appears to be an unusual means for probing energy band structure of nanostructured materials [12]. Anodic oxidation of Cd adlayers on n-type CdSe and CdS is hindered by blocking of anodic current on electrode-electrolyte interface. Figure 3 illustrates role of the chalcogen atom in lead chalcogenide in upd behavior. The anodic to cathodic peak separation increases in the row PbTe-PbSe-PbS, which indicates the increasing irreversibility of the upd process and also partial suppression of the upd and adlayer anodic oxidation on PbS. The electrochemical irreversibility of the upd increases likely due to the chemical bond ionic character increase in the chalcogenides series.
Ionic compounds such as oxides are not prone to acquire metal adlayers at underpotential and this provides selective upd on chalcogenide component of chalcogenide-oxide heterostructures, which is considered further. The electrochemical irreversibility of the upd of the kind shown in Fig.  3b is the indication of restricted electron transfer between metal adatoms and the chalcogenide support, and the restriction favors the use of upd as electrochemical probe for examination of energy band structure variation of the chalcogenide with particle size.

Pb ad on Bi 2 Te 3 versus Pb ad on Te
Lead upd on bismuth telluride and on tellurium proceed in the same potential region (Fig. 4) [36][37][38], but the cathodic part of voltammetric profile of the upd on tellurium transforms from a single peak to a two-humped structure in continuous cycling. The variability of the cathodic branch of voltammetric profile in upd on tellurium is due to adlayer gradual chemical interaction with tellurium and the corresponding gradual transition from upd on tellurium to upd on telluride in continuous cycling. The underpotential deposition of metal adatoms was found to be an efficient probe for disclosure of chemical state of surface layer of tellurium atoms in the substrate, which was helpful for detection of bismuth telluride contamination by elemental tellurium in electrodeposition.
Nanoparticles of elemental tellurium which can be coelectrodeposited with bismuth telluride, especially at high Te(IV):Bi(III) atomic ratio in electrolyte solution [11], are indistinguishable from Bi 2 Te 3 in XRD analysis [37,38]. As the latter is typically used to prove phase composition of the electrodeposits, the undetectable contamination by elemental tellurium is a potential catch in the electrochemical preparation of Bi 2 Te 3 -based thermoelectric materials. Fortunately, the electrochemistry of metal adlayers provides an efficient procedure for disclosure of elemental tellurium under conditions of electrodeposition [37]. The detection of elemental Te codeposition exploits peculiarities of voltammetric profiles variation at continuous cycling of lead upd on bare Bi 2 Te 3 (Fig. 5a) and on elemental Te (Fig. 5b). The first cycles of the upd are hardly distinguishable on both substrates, but a characteristic difference appears in further cycles as the additional cathodic peak before the main peak in upd on tellurium and the characteristic peak increases with each cycle (Fig. 5b). Though Pb upd on bare Bi 2 Te 3 is also not fully reversible (Fig. 5a), its voltammetric profile just scales down to slightly lower current density showing no additional peaks. Some irreversible mass increase in first cycles stabilizes leading to reversible mass changes resulting from adlayer deposition and dissolution in further cycles on Bi 2 Te 3 . Contrary to this evolution, the mass increase in Pb upd on Te progresses with each cycle. Though main part of Pb adlayer dissolves from the electrode upon anodic oxidation, some oxidized Pb adatoms form nuclei of PbTe on Te surface, so that the upd in the following cycles proceeds both on the remaining tellurium surface and on PbTe nuclei, the potential of the emerging cathodic peak corresponds perfectly to the potential of Pb upd peak on PbTe [36]. The growing peak of Pb upd on PbTe under continuous cycling was found to be a convenient indicator of Bi 2 Te 3 contamination by elemental tellurium and this helped to optimize conditions of electrodeposition of individual Bi 2 Te 3 and multicomponent nanostructures based on Bi 2 Te 3 [11, 37].

Chalcogenide particle size effects
Besides chemical state of the chalcogen atoms on a substrate surface, the chalcogenide particle size is a significant factor for electron transfer between chalcogenide nanoparticle and metal ions in the upd. Size effects in metal upd on metal chalcogenides were studied in greater detail with the use of typical quantum dot chalcogenides CdS and CdSe [12,13] and partly PbSe.
Cd upd on CdS and CdSe were first applied in the electrochemical atomic layer epitaxy (deposition) (ECALE/E-ALD [20,[39][40][41]) of the chalcogenides, although the publications about ECALE and E-ALD presented typically just cyclic  Effect of CdS and CdSe particle size in the upd of Cd upd was examined with the chalcogenide particle size control by heating at different temperature of chemical bath deposited (CBD) films [13] and also QD films electrophoretically deposited from colloidal solution of QDs of different size [12]. QD sizes in both types of the nanostructured films were derived from optical spectra. Figure 6 shows cyclic voltammograms and inverse charge transfer resistance R ct −1 potentiodynamic profiles of cadmium upd on the CBD CdS and CdSe films. The latter dependences obtained by analysis of PDEIS spectra characterize kinetics of the upd (R ct −1 is proportional to electrochemical reaction rate constant). Figure 6c, d insets show the equivalent electric circuit which fitted well to the PDEIS data in the potential range of the upd. The absence of diffusion impedance in the circuit corresponds to negligible contribution of mass transport to the PDEIS data; however, this does not necessarily exclude mass transport in the upd, as the low frequencies which normally give information about diffusion had to be truncated during impedance spectra acquisition in the potentiodynamic mode. The most significant feature in the potentiodynamic profiles of both the current and R ct −1 is their clear dependence on the QD size-the smaller are the chalcogenide QDs the lower is the underpotential shift. The size effects in the upd on chalcogenide QDs are somewhat similar to those in upd on small metal particles [42][43][44], but metal particles show size effects at much higher size. Figure 7 presents size effects in cadmium upd (peaks C 1 ) on electrophoretically deposited (EPD) films of CdSe QDs of different size [12]. The QDs were synthesized by a method described in [45] which gives much narrower size distribution compared to chemical bath deposition and this resulted in the greater perfectness of the size effects in the upd. The figure shows also potentials of bulk Cd cathodic deposition (C 2 ) and anodic oxidation (A 2 ) for comparison with the potentials of upd, the latter proceeds both in the region of bulk metal anodic oxidation and below reversible potential E(Cd 2+ /Cd). The upd competes efficiently with bulk phase electrodeposition in a narrow region below E(Cd 2+ /Cd) due to overpotential of metal phase nucleation. Figure 8a shows in terms of absolute electrode potential the correspondence of cadmium upd onset potential on CdSe to LUMO energy levels of CdSe QDs. Based on comparison of QD size effect on LUMO level and on the upd onset potential, the upd potential dependence on QD size was explained in [12] as the effect of the charge transfer control by the size-dependent position of LUMO level: the decrease in QD size results in the upward LUMO shift, so the electron transfer from electrode to metal cation via the LUMO energy level proceeds at more negative electrode potential. Figure 8b summarizes size effect on upd onset potential on different chalcogenides. These dependences can help to better understand cathodic reduction reactions on chalcogenide nanoparticles, not just upd; e.g., electrochemical corrosion of A II B VI particles which proceeds via formation of adatoms of metal A [46]. Also, the size-dependent upd appears to be an interesting new means of evaluation of the conduction band position in chalcogenide nanomaterials [12].

Electroactive surface area measurement
Electrochemical adsorption of hydrogen on Pt has been for long time the basic method of real surface area measurement of platinum electrodes, the electrochemical adsorption of oxygen on gold is a common method of gold electrodes real surface area measurement [47]. Also, Pb [48,49], Cu [50], and Hg [51] upd and oxidation of adsorbed CO layer [52] have been used to measure real surface area of some metal electrodes. The simple principle behind the surface area measurement is that adlayer oxidation/deposition charge is proportional to the area. The similar approach has provided measurement of the electroactive surface area of metal chalcogenides by underpotential deposition of metals. In work [53], the evolution of PbSe film roughness was estimated in the course of electrodeposition. The electrochemical deposition was terminated at different time and Pb adlayer oxidation charge was measured. Figure 9 shows the resulting dependence of the adlayer oxidation charge on deposition time. The figure shows that PbSe surface area grows fast and almost linearly with time at very beginning of the deposition and the growth slows down afterwards. The real surface area became equal to geometric area by approx. third second and showed a 13-fold excess at 600 seconds of deposition.
The measurement of electroactive surface area of metal chalcogenides by metal upd is especially helpful in investigations of chalcogenide heterostructures in which the other component is incapable of acquiring metal adlayer at underpotential, so the upd proceeds selectively on one component of the heterostructure and provides information about electroactive surface area of that particular component. Figure 10 shows how this method works in the investigation of spectral sensitizing of wide-bandgap oxides by CdSe nanoparticles [14]. The photocurrent sensitized by CdSe passes maximum with the increase in number of CdSe deposition cycles of successive ionic layer adsorption and reaction (SILAR) technique when the current is referred to geometric surface area (Fig. 10a) and this may result in a wrong conclusion about variation of CdSe sensitizing activity with number of SILAR cycles. Renormalization of the same photocurrent data for the electroactive surface area of the sensitizer component provided with the application of cadmium selective upd on CdSe gave an entirely different dependence of the photocurrent shown in Fig. 10b. The figure shows that the sensitizing activity of a unit surface area of the sensitizer in fact increased with number of SILAR cycles even in the region of the photocurrent decrease of Fig. 10a. The selective upd provides in tasks of this kind a unique opportunity of deriving specific characteristics of one component of multicomponent system from measurements applied on the whole system.

From adlayers to interlayers
Bismuth telluride is an example of chalcogenide with layered structure. Bi 2 Te 3 crystal consists of Te-Bi-Te-Bi-Te quintuples connected to each other by weak van der Waals bonds. Metal atoms can be inserted between the quintuples at van der   [14] Waals planes. The stable form of the inserted bismuth is atomic bilayer-a combination of two adlayers Bi ad which interact with bismuth telluride on the one side and form Bi-Bi bonds on the other side [11,54]. The product of bismuth bilayer insertion into bismuth telluride (Bi 2 ) m (Bi 2 Te 3 ) n appears as additive to bismuth telluride at bismuth telluride electrodeposition from electrolyte with high Bi(III):Te(IV) atomic ratio [11]. Figure 11 shows typical voltammetric signatures of (Bi 2 ) m (Bi 2 Te 3 ) n at electrochemical conditions of codeposition with bismuth telluride. The potential region of the anodic peak attributed to bismuth bilayer anodic oxidation overlaps with the region of Bi ad anodic oxidation on Bi 2 Te 3 at moderate Bi(III):Te(IV) atomic ratio and shifts negatively with Bi(III) concentration increase (Fig. 11), i.e., in the opposite direction to the one expected for Nernstian shift of bulk bismuth anodic peak. The difference in Bi oxidation potentials and the opposite effect of Bi(III) concentration on the potentials of the both anodic peaks helps to distinguish them electrochemically.
The technique for individual (Bi 2 ) m (Bi 2 Te 3 ) n electrodeposition was developed [11] based on pulsed potentiostatic control of two controlled potentials-the potentials of electrodeposition and electrochemical refinement. The refinement was applied after electrodeposition in each period at the potential which corresponded to a very beginning of bismuth bilayer anodic oxidation wave. Periodic switching between the deposition and refinement potentials helped to obtain the superlattices with general formula (Bi 2 ) m (Bi 2 Te 3 ) n and variable bismuth content dependent on electrolyte concentration. The products were identified with various methods and characterized electrochemically [11]. Interestingly, the potentiostatic treatment of (Bi 2 ) m (Bi 2 Te 3 ) n at the potential of bismuth bilayer anodic oxidation resulted in the product close to bismuth telluride by atomic composition but having an expanded structure, which could be of interest for further design of complex nanostructured materials based on bismuth telluride.

Conclusions
Cathodic underpotential deposition of metal adlayer on metal chalcogenide is the electrochemically irreversible surface limited reaction; hence, the cathodic deposition of the adlayer and its anodic oxidation proceed in poorly overlapping or not overlapping potential intervals. The irreversibility of the upd increases in the row from tellurides to selenides and further to sulfides, additionally the upd is significantly hindered on sulfides, probably due to greater ionicity of Me-S bond and wider bandgap.
The underpotential shift on chalcogenide nanoparticles increases with particle size. Size effect in Cd upd on CdSe which was investigated in a greater detail results from size dependence of LUMO energy and its correspondence to the onset potential of upd. Due to the correlation of the upd onset potential and LUMO energy, the onset potential of cadmium upd appears to be an unusual electrochemical means of CdSe quantum dot size measurement, and this may be of practical use in systems with complications for in situ optical measurement of QD size.
Upd on chalcogenides has been applied as a means of measurement of electroactive surface area of chalcogenide electrodes. Especially advantageous is the application of this method for the surface area measurement of chalcogenide component of heterostructures with the other component being incapable of providing upd, such as CdSe-TiO 2 and CdSe-ZnO.
Metal adlayers on chalcogenides were found to be stable against reaction with substrate. In contrast, metal adlayer deposited on chalcogen tend to react with substrate upon prolonged cycling of the adlayer cathodic deposition and anodic oxidation. This effect was applied for detection of trace amount of tellurium on bismuth telluride surface by lead upd which showed Pb upd on PbTe voltammetric signature emergence in the continuous cycling in presence of tellurium.
An interesting further task in the electrochemistry of adlayers refers to their electrochemical incorporation inside layered chalcogenides and thus forming superlattice structure. In particular, bismuth-bismuth telluride superlattices with bismuth interlayers can be obtained by electrodeposition in pulse potentiostatic mode with control of two potentials.