INTRODUCTION

The literature contains a large number of publications summarized in several reviews [1, 2], which are devoted to synthesizing nano and microporous metals including silver by electrochemical dealloying, i.e., selective anodic dissolution of alloys in aqueous media. The interest in nanoporous silver is associated with the possibility of using it as a selective catalyst for hydrogenation of unsaturated hydrocarbons [3], electrochemical reduction of carbon dioxide [4], and electrochemical oxidation of aldehydes [5]. Such silver can serve as the anode in fuel cells [6–8] working on ethanol, ethylene glycol, and borazane and also is promising as the substrate for surface-enhanced Raman spectroscopy [9]. Exhibiting the catalytic effect only a little weaker as compared with gold and platinum-group metals, the much cheaper nanoporous silver can serve as their alternative in redox processes [10, 11]. The antibacterial properties of this material are important for challenging medicinal applications, and its sensitivity with respect to UV radiation is significant for electrical and photo engineering.

To date, the conditions and certain main mechanisms of the percolation dealloying in aqueous solutions, i.e., the selective dissolution of alloys to form bicontinuous systems of pores and ligaments, are already well known. As a rule, the active component content in the alloy should be about 40–60 mol %, and the potential applied or the current passed should provide the electrolysis conditions such that the rate of removal of electroactive metal ions to the electrolyte bulk exceeds the diffusion rate of the more noble metal on the surface of the remaining metal phase [2, 12, 13]. For silver alloys, the active components are most often zinc [14–19] and aluminum [20, 21]; the more complex compositions are also used, e.g., Ce–Cu–Ag [22] and Y–Mg–Ag [23]. As the electrolyte, the aqueous solutions of acids are used: sulfuric [14‑17], hydrochloric [18], or nitric [16]; the neutral solutions are used more rarely [18]. The authors of [23] carried out the selective dissolution of Y–Mg–M alloys (M = Ni, Cu, Pd, Ag, Au) in aqueous solutions of organic acids which served not only as oxidants but also as the surface-active substances which impeded the surface diffusion of metals.

The scarce studies of dealloying in low-temperature ionic liquids [24] have only half-revealed the possibilities of using nonaqueous media stable at high temperatures. Their advantages as compared with aqueous media may include the intensification of the process, the absence of sources of hydrogen and oxygen in status nascendi, and the possibility of exerting control over the formation of porous structures by choosing the appropriate temperature mode.

The main idea of this study is to use molten salts as the electrolyte in solving the problems associated with specific features of high-temperature dealloying. Recently, we have made the first steps toward studying this type of specific anodic dissolution for Al–Au and Cu–Au alloys and also high-grade brass in chloride and carbonate melts [25, 26]. We have found that dealloying of these alloys in molten salt mixtures proceeds faster as compared with their dealloying in aqueous solutions and can produce a microporous structure (with the pore size from 0.1 to several tens µm). It is quite evident that the nature of alloy and the temperature can affect the pore size. However, the important questions on the effect of the phase composition of original alloys and the electrolysis parameters on the porous metal morphology are still open.

As the dealloying objects, we have chosen two homogeneous Zn–Ag alloys with the Zn0.67Ag0.33 and Zn0.46Ag0.54 composition; as the electrolyte, we used the (LiCl)0.57(CsCl)0.26(KCl)0.17 eutectics with the melting point at 260°С. The difference between the standard electrode potentials of Zn and Ag in chloride melts calculated based on thermodynamic data [27] is 0.74 V at 300°С. This value is sufficiently high to allow the possibility of selective dissolution of zinc from the alloy. The concrete goal of this study was to investigate how the conditions of electrochemical dealloying in percolation mode affect the shape and size of pores in surface layers of these alloys.

EXPERIMENTAL

Ag–Zn alloys were obtained by dissolving pieces of silver wire (99.9 wt %) in molten zinc (analytical grade) in a quartz tube in argon atmosphere under (LiCl)0.57(CsCl)0.26(KCl)0.17 flux at 680°С. After homogenization, the metal was drawn into a quartz tube with the inner diameter of 3 mm. The thus prepared rods were cut to obtain cylinders with the height of 25–30 mm. An end of such cylinder was polished to mirror brightness on metallograghic complex Struers (Struers ApS, Denmark). The alloys were tested by means of XRD device Rigaku D/MAX-2200VL/PC (Rigaku, Japan) and emission spectrometer with inductively coupled plasma Optima 4300 DV (Perkin Elmer, USA).

The electrolyte was prepared from salts (reagent grade). Sodium chloride was dried by freezing out moisture at the liquid nitrogen temperature, chlorinated, and subjected to three-fold zone melting to remove oxygen-containing impurities. The further manipulations with lithium chloride were carried out in a dry box. Potassium and cesium chlorides were melted in air. Approximately 50 g of salts in the required ratios were molten in inert atmosphere and homogenized at 800°С before each experiment.

The measurements were carried out in a three-electrode cell (Fig. 1) in atmosphere of dry purified argon. The melt was placed in SU-1200 glassy-carbon crucible, which also served as the counter electrode for polarization. The Ag/AgCl reference electrode, i.e., Ag|(LiCl)0.57(CsCl)0.26(KCl)0.17 eutectics + 3 mol % AgCl, was placed into an alundum crucible with orifices in its walls to provide contact with the working electrolyte. Below, all the potential values are referred to this electrode.

Fig. 1.
figure 1

Cell for electrochemical measurements: (1) rubber stopper, (2) heat-reflecting screens (3) quartz tube, (4) Ag–Zn alloy, (5) Ag/AgCl reference electrode in alundum crucible (6) salt melt, (7) glassy carbon crucible, (8) current terminal (Mo), (9) thermocouple.

Full size image

At the temperature of 300°С, the working electrode was immersed into the melt to the depth of 2–8 mm; after this the measurements were carried out by means of Autolab 302N system (Metrohm, Switzerland). The dealloying was carried out in the potentiostatic and galvanostatic modes by varying either the applied potential or the current and also the electrolysis time. At the end of each experiment, the electrode was washed from salt by distilled water and ethanol, and then the surface of the electrode/melt boundary was assessed visually. The effective current density was calculated by the formula i = I/Sgeom, where I is the current, Sgeom is the geometrical surface of the electrode/melt interface. The X-ray microanalysis of the electrode’s end surface was carried out on electron microscopes JSM-5900LV (Jeol, Japan) and TESCAN MIRA 3 LMU (TESCAN, Czech Republic).

RESULTS AND DISCUSSION

According to the results of elemental chemical analysis, the synthesized alloys had the Zn0.67Ag0.33 and Zn0.46Ag0.54 composition. Figure 2 shows the results of XRD analysis. In the diffraction pattern of Zn0.67Ag0.33 alloy, all the peaks coincide with the lines of the spectrum of the ε-phase in the Ag–Zn system (PDF#25-1325). The peaks in diffraction pattern of the Zn0.46Ag0.54 alloy correspond to the hexagonal ζ phase and the unstable cubic β phase (PDF#29-1156 and PDF#29-1155, respectively). At the temperature above 274°С, these both modifications are transformed to the β phase [28].

Fig. 2.
figure 2

Diffraction patterns of alloys (a) Zn0.67Ag0.33 and (b) Zn0.46Ag0.54: spectrum lines of phases (1) ε, (2) ζ, and (3) β'.

Full size image

Figures 3 and 4 show the anodic voltammetric curves for samples of these alloys and also for pure metals, measured at the potential scan rate of 1 mV/s. As the anodic potential of the Zn0.67Ag0.33 increases (Fig. 3а), the current density also increases and reaches the maximum at about –0.67 V after which decreases almost 4-fold. This is apparently explained by depletion of the metal surface with zinc and also by diffusion limitations for zinc delivery from the metal bulk. The i(Е) dependence of the sample has a plateau in the potential region from –0.6 V up to the silver dissolution potential (about –0.25 V).

Fig. 3.
figure 3

Anodic voltammetric curves of alloys (a) Zn0.67Ag0.33  and  (b)  Zn0.46Ag0.54 in (LiCl)0.57(CsCl)0.26(KCl)0.17 melt at 300°С.

Full size image
Fig. 4.
figure 4

Anodic voltammetric curves of (a) Zn and (b) Ag in (LiCl)0.57(CsCl)0.26(KCl)0.17 melt at 300°С.

Full size image

In the voltammetric curve of Zn0.46Ag0.54 alloy (Fig. 3b), the zinc dissolution peak is shifted by 0.3 V in the anodic direction and the current density in the maximum is lower as compared with the Zn0.67Ag0.33 sample. The silver dissolution begins in the vicinity of 1 V.

Figure 5 shows the corrosion potentials Еcor calculated for all these samples based on cyclic voltammetric curves.

Fig. 5.
figure 5

Corrosion potentials of (1) Zn, (2) Zn0.67Ag0.33, (3) Zn0.46Ag0.54, and (4), Ag in (LiCl)0.57(CsCl)0.26(KCl)0.17 melt at 300°С.

Full size image

The potential at which the current density is equal to 1 mA/cm2 is usually considered as the critical dealloying potential Ecr [29]. Table 1 shows the values Еcor and Ecr.

Table 1. Corrosion potentials and critical potentials of dealloying
Full size table

Under potentiostatic conditions, the electrolysis was carried out at the potential by 100–400 mV more positive than the critical dealloying potential. Figure 6 shows typical chronoamperograms of these alloys: the Zn0.67Ag0.33 sample is dissolved at the potential of ‒0.60 V (0.17 V more positive than the critical dealloying potential of this alloy), the Zn0.46Ag0.54 sample is dissolved at –0.18 V (0.34 V more positive than Ecr). In both cases, the current density decreases by an order of magnitude in the first minutes, which is associated with active formation of pores.

Fig. 6.
figure 6

Chronoamperograms  of  alloys  in (LiCl)0.57(CsCl)0.26(KCl)0.17 melt at 300°С: (a) Zn0.67Ag0.33, Е = –600 mV; (b) Zn0.46Ag0.54, Е = ‒170 mV.

Full size image

Figure 7 shows microimages of samples after 30 min dealloying which allowed the initial stage of pore formation to be observed. To that moment, the surface composition of two alloys turned out to be substantially different: the zinc content in the Zn0.67Ag0.33 sample decreased by merely several percent, whereas the surface of Zn0.46Ag0.54 alloy contained no more than 1–3 mol % Zn. The reason for this, besides the more noble initial composition, may be the fact that the electrolysis in this case occurred at the potential much more positive than the critical dealloying potential.

Fig. 7.
figure 7

Microimages of the alloy surface after dealloying for 1800 s: (a, b) Zn0.67Ag0.33, Е = –600 mV; (c, d) Zn0.46Ag0.54, Е = –170 mV.

Full size image

In the galvanostatic mode, the current density was in the interval of 20–50 mA/cm2. Figure 8 shows chronopotentiograms of alloys under study. When the current of the density of 20 and 35 mA/cm2 flowed (Figs. 8a and 8b) through the Zn0.67Ag0.33 alloy, the potential of about –0.06 V was established, which corresponded to the silver surface. The increase in the current density to 50 mA/cm2 (Fig. 8c) shifted the surface potential in the negative direction in the initial moment. This effect can probably be associated with the fast formation of deep pores and (or) the dissolution of silver together with zinc. After this, the surface proceeded to acquire the more positive potential.

Fig. 8.
figure 8

Chronopotentiograms of alloys (a‑c) Zn0.67Ag0.33 and (d) Zn0.46Ag0.54 in (LiCl)0.57(CsCl)0.26(KCl)0.17 melt at the current density, mA/cm2: (a) 20, (b) 35, (c) 50, (d) 7.

Full size image

During the galvanostatic electrolysis of the Zn0.46Ag.54 alloy, even at a small current density of 7 mA/cm2, the metal surface acquired very quickly the high positive potential (Fig. 8d), i.e., became the “silver” surface.

Figure 9 shows the surface microimages of the Zn0.67Ag0.33 alloy after galvanostatic electrolysis. Dealloying with the current density of 35 mA/cm2 for 1800 s (Figs. 9e and 9f) led to the formation of porous structures with pores and ligamets of the comparable size in the interval from 0.5 to several µm. The residual content of zinc was about 10 mol %. The increase in the electrolysis time at the slightly lower current density of 20 mA/cm2 (Figs. 9a–9d) made the surface more lacy due to the decrease in the size of ligaments, while the pore diameter remained unchanged. In this case, the surface layer was virtually 100% silver. The similar structures were obtained also when the current of 50 mA/cm2 was passed for 1800 s (Figs. 9g, 9h); however, they contained a large amount of residual zinc, up to 20 mol %.

Fig. 9.
figure 9

Microimages of the surface of Zn0.67Ag0.33 alloy after dealloying in (LiCl)0.57(CsCl)0.26(KCl)0.17 melt at the current density, mA/cm2: (a–d) 20, (e, f) 35, and (g, h) 50. Electrolysis time, s: (a, b) 2300, (c, d) 12 600, and (e–h) 1800.

Full size image

Figure 10 shows the microimages of the surface of Zn0.46Ag0.54 alloy after galvanostatic electrolysis.

Fig. 10.
figure 10

Microimages of the surface of Zn0.46Ag0.54 alloy after its dealloying by the current density of 7 mA/cm2 for 3600 s.

Full size image

The dealloying at the current density of 7 mA/cm2 for 3600 s led to formation of the hierarchical structure formed by ligaments measuring from 0.5 to 2 µm and pores of up to 4 µm, which formed joints of 5–20 µm length with pores between them up to 10 µm. The zinc content in this surface layer approached zero. The morphology of the thus obtained phase of metal silver considerably differed from the more uniform structure of pores and ligaments obtained at the higher current densities. It resembled more likely a network of branched dendrites but with considerable connectivity. This is why it seems interesting to study certain functional properties of synthesized materials, namely, their mechanical and electrocatalytic properties.

Table 2 summarizes the conditions and results of experiments performed.

Table 2. Conditions and results of electrochemical dealloying
Full size table

CONCLUSIONS

(1) The possibility of percolation dealloying of single-phase alloys Zn0.67Ag0.33 and Zn0.46Ag0.54 in the (LiCl)0.57(CsCl)0.26(KCl)0.17 eutectics at 300°С to form ultraporous metals of various morphology is demonstrated.

(2) It is shown that the virtually complete removal of zinc from the surface is reached by dealloying under galvanostatic conditions at the current density of about 20 for the ε phase and 7 mA/cm2 for the β phase of the Ag–Zn alloy. The selectivity of alloy dissolution decreases as the current density increases.

(3) The galvanostatic dealloying of the Zn0.67Ag0.33 alloy leads to formation of characteristic uniform porous structures with pores and ligaments of approximately the same size lying in the interval of 0.5–5 µm.

(4) The galvanostatic dealloying of the Zn0.46Ag0.54 alloy produces the dendrite structure formed by branched corallites with the characteristic size of agglomerates of about 0.5–4 and 5–20 µm.