Carbon monoxide oxidation on Au(1 1 1) surface decorated by spontaneously deposited Pt

Platinum is deposited spontaneously on Au(1 1 1) surface from 1 mM H₂PtCl₆ + 1 M HClO₄ solution using multiple deposition procedure. X-ray photoelectron spectroscopy (XPS) analysis has shown that after immersion into the Pt containing solution and rinsing with water, Pt(OH)₂ resides on the Au(1 1 1) substrate. Consecutive depositions as well as in situ scanning tunneling microscopy (STM) and electrochemical measurements are performed on previously electrochemically reduced Pt/Au(1 1 1) surfaces. Only homogeneous distribution of thus deposited Pt islands is observed by in situ STM. With subsequent depositions, the width of deposited Pt islands increases, but stays lower than 10 nm, while a significant increase of Pt islands height is observed, leading to moderate increase of the coverage. Cyclic voltammetry (CV) profiles of obtained Pt/Au(1 1 1) surfaces, and CO stripping curves are recorded in 0.5 M H₂SO₄ solution. CO oxidation takes place only at higher potentials shifting negatively with increasing coverage. This is discussed with respect to Pt islands width and height distributions and to the influence of the Au(1 1 1) substrate surface.     

Electroreduction of nitrate ions on Pt(1 1 1) electrodes modified by copper adatoms

Kinetics and mechanism of nitrate ion reduction on Pt(1 1 1) and Cu-modified Pt(1 1 1) electrodes have been studied by means of cyclic voltammetry, potentiostatic current transient technique and in situ FTIRS in solutions of perchloric and sulphuric acids to elucidate the role of the background anion. Modification of platinum surface with copper adatoms or small amount of 3D-Cu crystallites was performed using potential cycling between 0.05 and 0.3 V in solutions with low concentration of copper ions, this allowed us to vary coverage θ_Cu smoothly. Following desorption of copper during the potential sweep from 0.3 to 1.0 V allowed us to estimate actual coverage of Pt surface with Cu adatoms. Another manner of the modification was also applied: copper was electrochemically deposited at several constant potentials in solutions containing 10⁻⁵ or 10⁻⁴ M Cu²⁺ and 5 mM NaNO₃ with registration of current transients of copper deposition and nitrate reduction.It has been found that nitrate reduction at the Pt(1 1 1) surface modified by copper adatoms in sulphuric acid solutions is hindered as compared to pure platinum due to induced sulphate adsorption at E < 0.3 V. Sulphate blocks the adsorption sites on the platinum surface and/or islands of epitaxial Cu(1 × 1) monolayer thus hindering the adsorption of nitrate anions and their reduction. The extent of inhibition weakly depends on the copper adatom coverage. Deposition of a small amount of bulk copper does not affect noticeably the rate of nitrate reduction.Nitrate reduction on copper-modified Pt(1 1 1) electrodes in perchloric acid solutions occurs much faster as compared to pure platinum. The steady-state currents are higher by 4 and 2 orders of magnitude at the potentials of 0.12 and 0.3 V, respectively. The catalytic effect of copper adatoms is largely caused by the facilitation of nitrate adsorption on the platinum surface near Cu_ad and/or on the islands of the Cu(1 × 1) monolayer (induced nitrate adsorption).Hydrogen adatoms block the adsorption sites on platinum for NO₃⁻ anion adsorption and inhibit reactions of nitrate reduction even at moderate surface coverage.The products of nitrate reduction in sulphuric and perchloric acids are essentially the same (NO and ammonia) irrespective of the presence or absence of Cu on the platinum surface.     

Electrosorbed carbon monoxide monolayers on Pt(1 1 1)

We review structures of high-density CO monolayers on Pt(1 1 1) surfaces in CO-saturated electrolytes or in gaseous CO at near atmospheric pressure, using surface X-ray scattering (SXS) and scanning tunneling microscopy (STM). In electrolytes, we confirmed the well-known (2 × 2)-3CO and (√19 × √19)-13CO structures and were able to study the transition between them. For gas-phase studies, we were able to stabilize extremely well-ordered CO monolayers by emersion transfer from an electrochemical cell. We found that the hexagonal close-packed (2 × 2)-3CO structure is the equilibrium phase at room temperature in ∼1 atm CO gas pressure. This commensurate (C) phase transforms continuously to an incommensurate (IC) phase at elevated temperature (a second-order phase transition). We also confirm that the (√19 × √19)-13CO structure is stable at lower CO partial pressure. This C phase transforms discontinuously to an IC phase (a first-order phase transition). A tentative phase diagram and a brief review of structure details of the (2 × 2)-3CO and (√19 × √19)-13CO phases will be presented.     

Understanding CO-stripping mechanism from NiUPD/Pt(1 1 0) in view of the measured nickel formal partial charge number upon underpotential deposition on platinum surfaces in sulphate media

We recently showed nickel-underpotential deposition (Ni-UPD) occurs on polycrystalline or single crystal platinum electrodes in acidic media. Whereas the decoupling of the nickel and hydrogen adsorption/desorption peaks is difficult for low pH, these processes can be better separated for higher pH values, typically pH > 3. However, even for platinum single crystals, high pH solutions do not enable to sufficiently separate nickel from hydrogen phenomena. As a result, electrochemistry alone cannot yield important information about Ni-UPD, such as the formal partial charge number (valency of electrosorption) and the role of the sulphate or hydrogen sulphate anions.So, we decided to couple cyclic voltammetry to electrochemical quartz crystal microbalance (EQCM). EQCM measurements enable to decorrelate the simultaneous hydrogen and nickel adsorption/desorption peaks, which we could not attempt solely with electrochemistry. The coupling between gravimetric and electrochemical measurements allows us to detect the contribution of the anions and thus to isolate that of nickel: nickel coverage can then be determined. Nearly 4/5 Ni_UPD monolayer (θ_Ni ≈ 0.8) over platinum is reached at nickel equilibrium potential for high pH solutions (5.5). The QCM and electrochemistry coupling further allows the determination of nickel formal partial charge number: ι_Ni,EQCM = 1.3 ± 0.13. Direct electrochemistry measurements (Swathirajan and Bruckenstein method) yield: ι_Ni,Pt(poly) = 1.5 ± 0.17. These two values are close, which validates the electrochemical method for the nickel/platinum system. In consequence, we used Swathirajan and Bruckenstein method for Pt(1 1 0)-(1 × 2) crystal and found: ι_{Ni,Pt(1 1 0)} ≈ 1.4 ± 0.1. Whatever the system (Ni_UPD/Pt(poly) or Ni_UPD/Pt(1 1 0)-(1 × 2)) or the experimental technique, nickel formal partial charge number is lower than nickel cation charge: ι_Ni < z_Ni = 2. In consequence, upon underpotential deposition on platinum surfaces, nickel cations discharge and then undergo additional charge exchange processes, such as anion (or water) adsorption, resulting in apparent partial nickel cation discharge. Moreover, Ni_UPD/Pt(1 1 0) surface displays high activity towards CO_ad oxidation reaction. We explain such positive effect by the possible existence of a bifunctional mechanism in which oxygenated-species-covered Ni_UPD adatoms provide the oxygen atom to CO_ad?Pt species, enabling its facile oxidation.     

Electrochemical characterization of kinked Pt(7 5 1) surface in acidic media

Kinked Pt(7 5 1) surface was prepared and its electrochemical behaviors under different pretreatment conditions in acidic media were investigated systematically by using cyclic voltammetry. The results demonstrated that the upper limit of potential scanning and cooling atmospheres after the Pt(7 5 1) having been flame-annealed significantly influence the voltammetric behavior of Pt(7 5 1) electrode. The electric charge of hydrogen adsorption-desorption slightly increases with increasing the upper limit of potential scanning. Different cooling atmospheres give rise impacts to the surface structure of Pt(7 5 1) electrode, but hardly change the amount of hydrogen adsorption-desorption sites on the electrode. In addition, the so-called third oxidation peak appears near −0.08 V in H₂SO₄ media and −0.05 V in HClO₄ solution because of the presence of (1 1 0) terrace sites on this surface, and a plausible mechanism for the formation of this current peak is discussed. The results are of importance in understanding the electroadsorption properties of the kinked Pt(7 5 1) surface, as well as in further exploration of this kinked electrode in electrocatalysis.     

Adenine adsorption on Au(1 1 1) and Au(1 0 0) electrodes: Characterisation, surface reconstruction effects and thermodynamic study

Adsorption of adenine on Au(1 1 1) and Au(1 0 0) electrodes is studied by cyclic voltammetry, impedance and chronoamperometric measurements in 0.1 M and 0.01 M KClO₄ and in 0.5 M NaF solutions. The experiments performed with flame-annealed electrodes at different contact potentials, scan potential limits and scan rates, suggest different adsorption behaviour on the unreconstructed and reconstructed surface domains. This is confirmed by comparing the results obtained with electrochemically annealed unreconstructed and with flame-annealed reconstructed surfaces. In both cases the initial electrode surface state is characterised by the E_pzc values. The adsorption on reconstructed surfaces takes place at more positive potentials than on the unreconstructed surfaces and induces the lifting of the reconstruction.The thermodynamic analysis is performed on the chronoamperometric data for adenine desorption on well characterised unreconstructed Au(1 1 1) surfaces. To this end a new methodology of the chronoamperometric experiments is introduced. Quantitative thermodynamic adsorption parameters such as surface tension, Gibbs surface excess, Gibbs energy of adsorption, potential versus Gibbs excess slope and electrosorption valency are determined. Weak chemisorption of adenine is inferred with a molecular orientation independent on the coverage and on the electrode potential. It is proposed that adsorbed adenine molecules adopt a tilted orientation at the surface to facilitate the coordination to the gold atoms.     

In situ radiotracer and voltammetric study of the formation of surface adlayers in the course of Cr(VI) reduction on polycrystalline and (1 1 1) oriented platinum

This paper is focused on the in situ radiotracer and voltammetric studies of the induced HSO₄⁻/SO₄²⁻ adsorption at Pt(poly) and Pt(1 1 1) surfaces in 0.1 mol dm⁻³ HClO₄ solution in the course of Cr(VI) electroreduction. Besides this, the sorption behavior of HSO₄⁻/SO₄²⁻ ions on bare Pt(poly) and Pt(1 1 1) electrodes is compared and discussed. From the experimental results it can be stated that: (i) although the extent of bisulfate/sulfate adsorption is strongly dependent upon the crystallographic orientation of Pt surfaces, the maximum coverage on the Pt(1 1 1) does not exceed 0.2 monolayer; (ii) the Cr(VI) electroreduction on both poly- and (1 1 1) oriented platinum proceeds via a ce (chemical-electron-transfer) mechanism to yield Pt surfaces covered with intermediate surface adlayers containing Cr(VI) particles (and reduced Cr-containing adspecies) and ‘strongly bonded’ HSO₄⁻/SO₄²⁻ ions; (iii) while the coverage of platinum surfaces by the intermediate complexes formed in the course of Cr(VI) electroreduction at E > 0.20 V is basically independent of the crystallographic orientation of the Pt electrode, the onset for rapid Cr(VI) reduction is highly affected by the nature and crystallographic orientation of the electrode.     

Kinetic study of nitrate reduction on Pt(1 1 0) electrode in perchloric acid solution

The kinetics of electrocatalytic reduction of nitrate on Pt(1 1 0) in perchloric acid was studied with cyclic voltammetry at a very low sweep rate of 1 mV s⁻¹, where pseudo-steady state condition was assumed to be achieved at each electrode potential. Stationary current-potential curves in perchloric acid in the absence of nitrate showed two peaks at 0.13 V and 0.23 V (RHE) in the so-called adsorbed hydrogen region. The nitrate reduction proceeded in the potential region of the latter peak in the pH range studied. The reaction orders with respect to NO₃⁻ and H⁺ were observed to be close to 0 and 1, respectively. The former value means that the adsorbed NO₃⁻ at a saturated coverage is one of the reactants in the rate-determining step (rds). The latter value means that hydrogen species is also a reactant above or on the rds. The Tafel slope of nitrate reduction was −66 mV per decade, which is taken to be approximately −59 mV per decade, indicating that the rds is a pure chemical reaction following electron transfer. We discuss two possible reaction schemes including bimolecular and monomolecular reactions in the rds to explain the kinetics and suggest that the reactants in the rds are adsorbed hydrogen and adsorbed NO₃⁻ with the assistance of the results in our recent report for nitrate reduction on Pt(S)[n(1 1 1) × (1 1 1)] electrodes: the nitrate reduction mechanism can be classified within the framework of the Langmuir-Hinshelwood mechanism.     

Resonance surface X-ray scattering technique to determine the structure of electrodeposited Pt ultrathin layers on Au(1 1 1) surface

It is demonstrated that resonance surface X-ray scattering (RSXS), in which incident X-ray energy close to the Pt L_III absorption edge (11.55 keV) is used, is very useful for the determination of the structure of electrodeposited Pt thin layers on a Au(1 1 1) surface. This technique was applied to characterize the structure of electrodeposited Pt layers on Au(1 1 1) substrates prepared under two extreme conditions, which are known to provide rough and atomically flat layers. Detailed structural information was obtained by RSXS measurements and it was confirmed that the structures of the Pt layers were as reported. Pt atoms of the atomically flat monolayer were found to be situated at the threefold hollow cubic closest packing (ccp) sites of the Au(1 1 1)-(1 × 1) surface.     

Electrochemical characterization of irreversibly adsorbed germanium on platinum stepped surfaces vicinal to Pt(1 0 0)

The electrochemical behavior of germanium irreversibly adsorbed at stepped surfaces vicinal to the Pt(1 0 0) pole is reported. The process taking part on the (1 0 0) terraces is evaluated from charge density measurements and calibration lines versus the terrace dimension are plotted. On the series Pt(2n − 1,1,1) having (1 1 1) monoatomic steps, the charge involved in the redox process undergone by the irreversibly adsorbed germanium is able to account for (n − 0.5) terrace atoms, thus suggesting some steric difficulties in the growth of the adlayer on the (1 0 0) terraces. Conversely, no steric problems are apparent in the series Pt(n,1,0) in which more open (1 0 0) steps are present on the (1 0 0) terraces. In this latter case the charge density under the germanium redox peaks is proportional to the number of terrace atoms. Some comparison is made with other stepped surfaces to understand the behavior and stability of germanium irreversibly adsorbed on the different platinum surface sites.     

Crystalline alloys produced by mercury electrodeposition on Pt(1 1 1) electrode at room temperature

In situ scanning tunneling microscopy (STM) and reflection high energy electron diffraction (RHEED) were used to characterize mercury film electrodeposited onto a Pt(1 1 1) electrode at room temperature. Depending on the amount of Hg deposit, two different growth modes were observed. At low Hg coverage, crystalline (0 0 0 1)Hg adlayer accompanied by 30°-rotated (1 1 1)-Pt patches was found on Pt(1 1 1). Deposition of multilayer Hg resulted in layered PtHg₂ and PtHg₄ amalgams, which grew epitaxially by aligning their (2 0 1) and planes, respectively, parallel to the Pt(1 1 1) substrate. The preference of these epitaxial relationships for the electrochemically formed Pt-Hg intermetallic compounds on Pt(1 1 1) could result from minimization of the surface energy.     

XPS study of irreversibly adsorbed arsenic on a Pt(1 1 1) electrode

The chemical composition of an irreversibly adsorbed layer of arsenic on Pt(1 1 1) in sulfuric acid solution has been studied by X-ray photoelectron spectroscopy (XPS). From the chemical shift of the As 3d level, a change in the valence state from As(0) to As(III) with positive-going electrode potential is deduced, the total amount of As on the surface remaining constant. The As coverage derived from XPS is around 0.33 ML, which is in agreement with the charge under the current peak in the cyclic voltammogram. From the coadsorption of anions, accompanying the valence transition of As, As(III) is assumed to exist as As(OH)₃ on the surface.     

Heteroepitaxial growth and characterization of 3C-SiC films on on-axis Si (1 1 0) substrates by LPCVD

Cubic SiC (3C-SiC) film has been deposited on Si (1 1 0) substrate by the low pressure chemical vapor deposition (LPCVD) with gas sources of SiH₄, C₃H₈ and carrier gas of H₂. The 3C-SiC crystalline film can be confirmed through the observations using reflection high-energy electron diffraction (RHEED) images. The X-ray diffraction (XRD) pattern and the rocking curve indicate that the (1 1 1) plane of SiC film is parallel to the surface of the Si (1 1 0) substrate and the film is of high crystallinity. The results of the field emission scanning electron microscope (FESEM) images show that the film has smooth surface morphology. Transmitted electron diffraction (TED) pattern and high resolution transmission electron microscope (HRTEM) image further confirm the high quality of the film.     

Adsorption of phosphate species on poly-oriented Pt and Pt(1 1 1) electrodes over a wide range of pH

The adsorption of phosphate anions from phosphate solutions at poly-oriented and single-crystal platinum electrodes, primarily Pt(1 1 1), was studied over a wide range of pH by cyclic voltammetry. The features observed at the poly-oriented Pt electrode in phosphate solution may be related to the different crystalline facets, the (1 1 1) orientation presenting the most significant behavior in terms of phosphate adsorption. On the reversible hydrogen electrode (RHE) scale, the phosphate adsorption strength decreases with increasing alkalinity of the solution. Qualitatively, three different pH regions can be distinguished. At pH < 6 only a broad reversible peak is observed, corresponding to the adsorption of H₂PO₄⁻ and further deprotonation to adsorbed HPO₄⁻. For 6 < pH < 11 a butterfly feature followed by one or two anodic peaks (depending on scan rate) is observed, ascribed to the adsorption of HPO₄⁻ followed by its subsequent deprotonation to adsorbed PO₄³⁻. The splitting into two or three voltammetric features, and the irreversibility of the two features at more positive potential, is ascribed to the deprotonation reaction leading to a surface species (i.e. phosphate) which needs to change its surface coordination. At pH > 11 a reversible pre-wave and a sharp spike are observed, ascribed to the co-adsorption of phosphate and hydroxide.     

Vibrational Stark tuning rates from periodic DFT calculations: CO/Pt(1 1 1)

DFT periodic calculations have been used to study the influence of an external electric field on the adsorption of CO on Pt(1 1 1). Particular attention has been focused on the determination of the CO and metal-CO vibrational Stark tuning rates. Stark tuning rates have been calculated at various CO coverages; a linear dependence between the CO Stark tuning rate and the CO surface coverage has been found. We have calculated a value of 68.94 cm⁻¹/(V/Å) for the zero-coverage limit CO Stark tuning rate, in good agreement with the experimental value of 75 ± 9 cm⁻¹/(V/Å). Like the CO Stark tuning rate, the metal-CO vibrational Stark tuning rate also increases as CO surface coverage decreases. In addition, we have found (at 0.25 ML) that the CO Stark tuning rate is similar at different adsorption sites, being only slightly larger at high-coordinated sites. CO vibrational Stark tuning rates of 45.58, 47.96, 47.61 and 48.49 cm⁻¹/(V/Å) have been calculated for ontop, bridge, hcp and fcc hollow sites, respectively. Calculations at high coverage using a (2 × 2)-3CO model yield a CO Stark tuning rate of 21.08 and 25.93 cm⁻¹/(V/Å) for ontop and three-fold hollow CO, respectively. These results show that the CO Stark tuning rate for CO adsorbed at high coordinated sites is only slightly larger than that at ontop sites. This result is in contradiction with experiments, which reported larger CO Stark tuning rates at high-coordinates sites than at ontop sites. Furthermore, the calculated metal-CO stretch is larger for ontop sites than for high-coordinated sites; this result is in disagreement with previous DFT cluster model calculations. Unfortunately, there is not experimental information available to support either result. Finally, we have also studied the CO adsorption site preference dependence on electric fields. We have found that CO adsorbs preferentially at high coordinated sites at more negative fields, and at ontop sites at more positive fields, in agreement with previous experiments and DFT cluster model calculations.     

Adsorption of camphor and 2,2′-bipyridine on Bi(1 1 1) electrode surface

Impedance spectroscopy and in situ STM methods have been used for investigation of the camphor and 2,2′-bipyridine (2,2′-BP) adsorption at the electrochemically polished Bi(1 1 1) electrode from weakly acidified Na₂SO₄ supporting electrolyte solution. The influence of electrode potential on the adsorption kinetics of camphor and 2,2′-BP on Bi(1 1 1) has been demonstrated. In the region of maximal adsorption, i.e. capacitance pit in the differential capacitance versus electrode potential curve, the heterogeneous adsorption and diffusion steps are the rate determining stages for camphor and 2,2′-BP adsorption at the Bi(1 1 1) electrode. It was found that for camphor | Bi(1 1 1) interface the stable adsorbate adlayer detectable by using the in situ STM method has been observed only at the positively charged electrode surface, where the weak co-adsorption of SO₄²⁻ anions and camphor molecules is possible. At the weakly negatively charged Bi(1 1 1) electrode surface there are only physically adsorbed camphor molecules forming the compact adsorption layer. The in situ STM data in a good agreement with impedance data indicate that a very well detectable 2,2′-BP adsorption layer is formed at Bi(1 1 1) electrode in the wide region of charge densities around the zero charge potential.     

Electrochemical growth and dissolution of Ni on bimetallic Pd/Au(1 1 1) substrates

In this work we study the electrochemical growth and dissolution of a Ni on Pd-Au(1 1 1) bimetallic surfaces using in situ scanning tunnelling microscopy. We also compare Ni deposition on monometallic electrodes, i.e. Au(1 1 1) and Pd(1ML)/Au(1 1 1), using electrochemical characterizations. Results evidence that the first Ni monolayer grows preferentially on Au(1 1 1) in a wide potential range, and that a full Ni monolayer covering the entire Pd-Au surface can be selectively dissolved from Pd islands. No such selectivity is observed upon growth of subsequent Ni atomic planes. We demonstrate that the Ni-substrate interactions play a key role in the above mentioned selectivity. The binding energy of Ni to Pd is found to be 80 meV smaller than of Ni to Au. The sign and the amplitude of this difference are discussed in light of the d band filling of the Pd-Au(1 1 1) bimetallic surface and the presence of adsorbed H on Pd before deposition.     

Oxidation of formaldehyde and ethanol on Au(1 1 1) and Au(1 1 1) modified by spontaneously deposited Ru in sulfuric acid solution

The oxidation of formaldehyde and ethanol on both pure Au(1 1 1) and Au(1 1 1) modified by approximately 0.3 monolayer (ML) of spontaneously deposited Ru was studied by cyclic voltammetry (CV) in 0.5 M H₂SO₄ solution containing either 0.25 M formaldehyde or 0.35 M ethanol. In situ scanning tunneling microscopy (STM) and CV were employed to characterize the Au(1 1 1) and Ru/Au(1 1 1) surfaces. The oxidation of HCHO on Ru/Au(1 1 1) commences at 0.1 V more negative potential than on pure Au(1 1 1). From 0.25 to 0.55 V vs. (Ag/AgCl), the reaction occurs with increasing current, showing a peak at a potential of 0.43 V. It is assumed that the increasing anodic activity of the Ru/Au(1 1 1) surface is associated with the oxidation of some reaction intermediates, facilitated by the presence of Ru in its metallic state. On the other hand, the oxidation of ethanol on Ru/Au(1 1 1) commences at 0.1 V more positive potential than on pure Au(1 1 1), and proceeds in the potential region from 0.2 to 0.5 V with significantly smaller currents, showing a peak at 0.43 V. This inhibiting effect is explained by the deactivation of the most active Au(1 1 1) step sites by high coverage with Ru islands. The appearance of a small peak at 0.43 V is most likely associated to the oxidation of some intermediates during ethanol oxidation at the Ru/Au step sites formed on the Au(1 1 1) terraces by the presence of a small coverage with Ru islands.     

Alloy formation during the electrochemical growth of a Ag-Cd ultrathin film on Au(1 1 1)

The electrodeposition of a Ag/Cd ultrathin film on a Au(1 1 1) surface and the formation of a surface alloy during this process have been studied using classical electrochemical techniques and in situ Scanning Tunneling Microscopy (STM). The films were obtained from separate electrolytes containing Ag⁺ or Cd²⁺ ions and from a multicomponent solution containing both ions. First, the polarization conditions were adjusted in order to form a Ag film by overpotential deposition. Afterwards, a Cd monolayer was formed onto this Au(1 1 1)/Ag modified surface by underpotential deposition. The voltammetric behavior of the Cd UPD and the in situ STM images indicated that the ultrathin Ag films were uniformly deposited and epitaxially oriented with respect to the Au(1 1 1) surface. Long time polarization experiments showed that a significant Ag-Cd surface alloying accompanied the formation of the Cd monolayer on the Au(1 1 1)/Ag modified surface, independent of the Ag film thickness. In the case of an extremely thin Ag layer (1 Ag ML) the STM images and long time polarization experiments revealed a solid state diffusion process of Cd, Ag, and Au atoms which can be responsible for the formation of different Ag-Cd or Au-Ag-Cd alloy phases.