FABRICATION AND MODIFICATION OF METALLIC NANOPOWDERS BY ELECTRICAL DISCHARGE IN LIQUIDS

N.V. Tarasenko1, A.A. Nevar1, N.A. Savastenko2, E.I. Mosunov3, N. Z. Lyakhov4, T.F.Grigoreva4

1 Institute of Physics, NAS B, Minsk, Belarus

2 Leibniz-Institute for Plasma Science and Technology, Greifswald, Germany

3The Institute of Machine Mechanics and Reliability NAS B, Minsk, Belarus

4Institute of SolidState Chemistry and Mechanochemistry, SB RAS,

18 Kutateladze, Str , Novosibirsk, 630128, Russia,

Electrical-discharge technique was developed for preparation of metallic and metal-containing nanoparticles as well as for modification of metal micropowders in liquids. The morphology and composition of the nanopowders formed under various discharge conditions were investigated by means of transmission electron microscopy and X-ray diffraction analysis. The optimal conditions for the production of titanium carbide and copper nanoparticles embedded in carbon layers were found.

Introduction

A synthesis of metallic and metal-containing nanopowders is of a great interest due to their potential applications as super hard materials [[1]], environmentally friendly fuel cellswith highly effective catalysts [[2],[3]],and so on.Transition metal carbides have been widely studied as electrocatalysts, because of their electrochemical properties and electrical conductivities. Nanosized carbon particles are suitable support materials for certain types of catalysts. Of particular interest for future catalytic applications are carbon-based materials with embeded metal nanoparticles [[4]].As long as carbon nanoparticles are relatively inert supports many studies have been conducted in order to find which pre-treatment procedures are needed to achieve optimal interaction between the support and metal species [[5]].

For any application of nanoparticles to be commercially viable low-cost production methods have to be developed. A low-temperature and non-vacuum synthesis of nanoparticles via discharge in liquid (submerged discharge) provides a versatile choice for economical preparation of various nanostructures in a controllable way. An arc discharge in liquid nitrogen has firstly been reported as a cost-effective technique for the production of carbon nanotubes in 2000 by Ishigamy et al. [[6]]. Since that time, many efforts have been devoted to develop this method. Sano et al. proposed to submerge electrodes in water instead of liquid nitrogen [[7],[8]]. They reported synthesis of carbon onions [7,8] and single-walled carbon nanohorns (SWNHs) [[9]]. In latter case, carbon nanoparticles were produced via discharge in water method with the support of gas injection. Parkansky et. al reported nanoparticles synthesis via a pulsed arc submerged in ethanol. Ni, W, steel and graphite electrodes were used [[10],[11]]. The particles composition varied from carbon to pure metal including various intermediate combinations of these materials. Bera et al. employed an arc-discharge in a palladium chloride solution to produce carbon nanotubes decorated with in situ generated Pd nanoparticles [10]. Importantly, the synthesized material contained no chlorine.

In this paper methodsbased on electrical-discharges in liquids for production oftungsten and titanium carbide as well as copper nanoparticles embedded in carbon nanostructures is reported. The capabilities of arc and spark discharges submerged in liquids for synthesis of nanoparticles as well as electrical-discharge modification of metallic powders were studied.

Experimental details

The experimental reactor (Fig. 1) consisted of four main components: a power supply system (pulse generator), the electrodes, a glass vessel and a water cooling system outside the beaker. A pulsed discharge was generated between two electrodes being immersed in 100 ml of liquid (pure (99.5%) ethanol or 0.001 M CuCl2 aqueous solution). The appropriate combinations of pairs of metallic (tungsten, titanium or copper) and graphite electrodes were used. The choice of ethanol was motivated by the fact that organic compounds play a role of a carbon source to produce nanoparticles in discharge-in-liquid system [7, [12]]. Addition of the copper chloride salt into double distilled water favored the activation of discharge process. Metal (tungsten, titanium or copper) and graphite rods with diameters of 6 mm were employed as electrodes. An optimum distance between the electrodes was kept constant at 0.3 mm to maintain a stable discharge. The discharge was initiated by applying a high-frequency voltage of 3.5 kV. The power supply provided several different types of discharges. Both direct current (dc) and alternating current (ac) arc and spark discharges were generated with repetition rates of 100 and 50 Hz, respectively. Current I(t) was recorded during the discharge as a function of time by means of an oscilloscope. The peak current of the arc discharge was 9 A with a pulse duration of 4 ms. The peak current of the pulsed spark discharge was 60 A with a pulse duration of 30 μs.

The synthesized products were obtained as colloidal solutions. After 15 min presedimentation the large particles precipitated at the vessel bottom. The top layer contained the small nanoparticles was carefully poured off into a Petry dish. These suspended nanoparticles were characterized byUV-Visible optical absorption spectroscopy, transmission electron microscopy (TEM) and X-ray diffraction analysis (XRD) for their size, morphology, crystalline structure and composition.

The optical absorption spectra of colloids were measured by UV–Visible spectrophotometer (CARY 500) using 0.5 cm quartz cuvette. Transmission electron microscopy was performed by LEO 906E (LEO, UK, Germany) microscope operated at 120 kV. A drop of solution put onto the amorphous carbon coated copper grid for TEM measurements. Thereafter the liquid was evaporated at the temperature of 80C. After the drying of colloidal solution the deposit obtained on the bottom of Petri dish was examined by XRD.Powder composition and its crystalline structure were characterized by using X-ray diffraction at CuK(D8-Advance, Bruker, Germany).

Synthesis of carbide nanopowders

Promising capabilities of the developed technique for synthesis of tungsten and titanium carbides (WC, TiC), as well as carbon-encapsulated copper nanoparticles were demonstrated using the appropriate combinations of pairs of metallic and graphite electrodes submerged into the appropriate solution. Also physical and chemical processes induced by the electrical discharges in liquids were studied to optimize the process of nanoparticles synthesis.

The results of nanoparticles preparation are summarized in the Table1. The synthesis rate varied in range of 2 – 40 mg min-1 depending on peak current and pulse duration of discharge as well as polarity of metal and graphite electrodes. The synthesis rate increased with increasing of discharge current and decreasing of pulse duration. The composition and morphology of nanoparticles were also found to depend on discharge parameters.It should be noted that there is a possibility to scale-up the process.

Table 1 summarized the variation in synthesis rate and composition of tungsten nanopowders with the discharge parameters. As a general tendency, the synthesis rate was order of magnitude higher for spark discharge than that of arc discharge. It may be due to the difference in current value [[13]]. For both arc and spark discharges, it was found that the synthesis rate is lower when tungsten was acting as a cathode. This result is consistent with literature data. For example Bera et. al reported that the consumption of anode is higher than that of cathode. [13].

Table 1. Summary of nanopowder synthesis conditions and results of nanopowder characterization byXRD

Discharge type / Electrodes / Powders yield, mg/min / XRD-analysis
W2C,
vol. % / WC1-x, vol. % / C,
vol. % / W,
vol. %
1 / ac arc / W : C / 0.2 / 7.1 / 78.1 / 14.7 / -
2 / dc arc / W(cathode):C(anode) / 0.1 / 6.2 / 90.1 / 3.7 / -
3 / dc arc / W(anode):C(cathode) / 0.2 / 6.6 / 71.5 / 21.9 / -
4 / ac spark / W : C / 2.5 / 5.8 / 32.8 / 61.4 / -
5 / dc spark / W(cathode):C(anode) / 1.2 / 57.0 / 30.7 / 8.9 / 3.3
6 / dc spark / W(anode):C(cathode) / 2.1 / 5.6 / 32.5 / 61.8 / -

As it can be seen from the Table 1, the synthesized nanopowder is a mixture of hexagonal W2C, face centered cubic WC1-xand graphite. No peaks corresponding to WO were observed. Nanopowder contained also small amount body centered cubic W when synthesis was performed by dc current spark discharge with tungsten rod acting as cathode. Here, the particular behavior of this discharge should be stressed, showing rather high ability to synthesize W2C. Moreover, in contrast to the other spark discharges, synthesized material contained relatively small amount of graphite. On the other hand, applying tungsten as a cathode material appears to reduce C content in nanopowder prepared via arc discharge, too. Generally, the content of C is higher and content of WC1-x is lower when synthesis was performed by spark discharge.

Nanoparticles prepared by arc discharge were observed in their agglomerated form. The agglomerated nanoparticles were surrounded by the grey regions, which were probably graphite layers. Thistypical view was seen everywhere in TEM images of product synthesized by arc, for both ac and dc current discharges irrespective of electrodes polarity. That fact implies that the morphology of synthesized nanopowders was governed rather by the current pulse duration and value of peak current than the polarity of the electrodes. Since nanoparticles were observed in the agglomerated form, it was difficult to measure their size correctly. We suppose that approximately 4 nm nanoparticles are formed during the arc discharge in ethanol.

Fig.1 shows the TEM image of titanium carbide nanopowder synthesized by spark discharge in ethanol. As can be see from the Fig.1 the nanoparticles were also surrounded by graphite layers. Fig. 1 demonstrates that the nanoparticles synthesized by spark were nearly spherical with a mean diameter of ~ 7 nm. The particle size distribution was rather narrow (±2 nm).The XRD pattern of synthesized sample is shown in Fig. 1 (right picture). The diffraction peaks at 6,0°; 41,8°; 60,5°; 72,4°; 76,5° and 40,7°; 50,4°; 59,0°; 66,7°; 74,1° correspond to the formation of cubic face-centered titanium carbide TiC and cubic primitive TiC2 respectively. There are some diffraction peaks with 2θvalue of 40,7°; 50,4°; 59,0°; 66,7° and 74,1°, which can be assigned to the hexagonal C. The amount of TiC reached 88.7 vol.%. The quantities of TiC2 and C in samples detected by XRD corresponded to ca. 4.7 vol.% and ca. 6.7 vol.%, respectively.

Fig. 1 TEM image (left picture) of titanium carbide nanopowder synthesized by ac spark discharge and XRD-pattern (right picture) of the sample.

Synthesis of copper-carbon composite nanostructures

Numerous studies have focused on synthesis of metal-containing carbon nanocapsules (CNCs) via submerged discharge method [8,9,[14],[15],[16]]. Because of the carbon sheets surrounding the metal core, the CNCs are protected from the environment and from degradation. The carbon coatings mean that nanoparticles are biocompatible and stable in many organic media. Thus, carbon encapsulated nanoparticles are candidate for bioengineering application, high-density data storage, magnetic toners for use in photocopiers [8,[17],[18]]. The metal containing carbon nanostructures were prepared by using the electrode from mixture of graphite and metal precursor [16, [19],[20]]. Recently Xu et al. demonstrated a possibility to synthesize Ni-, Co- and Fe-containing CNCs by an arc discharge between carbon electrodes in aqueous solution of NiSO4, CoSO4 and FeSO4 respectively [15]. In contrast to the data reported by Bera et al., the synthesized material consisted of O and S due to SO4-2 ionic precursors in the solution. Since the metal core-forming material was supplied by liquids, the production rate of CNCs was limited by the salt concentration [4]. This restriction may cause a limit to apply the submerged discharge method tothe large-scale production of CNCs.

In this paper, Cu-based nanoparticles were prepared via submerged discharge of bulk copper and graphite electrodes in a copper chloride (CuCl2) aqueous solution. Thus, material of copper electrode as well as Cu from solution was supposed to be incorporated into the resulting nanoparticles. The effect of discharge parameters and electrode composition on the morphology and composition of final products have been investigated. Additionally, synthesized material was modified by laser irradiation. The changes in nanoparticles morphology and composition were examined by transmission electron microscopy (TEM), X-ray diffraction (XRD), and UV-Vis spectroscopy.

The six types of nanoparticles suspension were prepared under different discharge parameters. The synthesis parameters are summarized in Table 2. As it can be seen, the weight changeof each electrode was generally higher, when spark discharge was generated. The anode consumption rate was higher than that of cathode irrespective to a discharge type and electrode material. However, in contrast to the literature data [4], there was no cathode gain in weight. As a general trend, the nanopowder synthesis rate was higher for spark discharge than that of arc discharge. It may be explained by the difference in current value [[21]]. For both arc and spark discharges, it was found that the synthesis rate was higher when copper was acting as an anode. There is a discrepancy between nanopowder synthesis rate and material consumption rate. The values of discrepancy, D, listed in the Table 2 were calculated as follows:

(1).

Here Rsyn is the synthesis rate of nanopowder, RCu is the consumption rate of the copper electrode and RC is the consumption rate of the graphite electrode. The discrepancy, D, depended on discharge parameters. For ac-discharges, the value of discrepancy was higher for spark discharge than that for arc discharge. For dc-discharges, this trend remained if the polarity of electrodes was taken into account. It is worth to notice here that the discrepancy between material consumption rate and nanopowder synthesis rate may be caused not only by separation of sediment fraction but by the reaction of carbon atoms with water resulting in the production of gaseous compounds [9].

Table 2.Summary of nanopowder synthesis parameters.

Type of discharge;
peak current, pulse duration /
Electrodes material
/ RCu and RC,
mg min-1 / RSyn,
mg min-1 / D,
%
1 / ac1) spark;
60 A, 30 µs / Cu / 6.7 / 5.9 / 49
C / 4.8
2 / ac arc;
10 A, 4 ms / Cu / 1.2 / 2.5 / 34
C / 2.6
3 / dc2) spark;
60 A, 30 µs / Cu (cathode electrode) / 4.7 / 2.1 / 81
C (anode electrode) / 6.1
4 / dc spark;
60 A, 30 µs / Cu (anode electrode) / 6.6 / 6.9 / 38
C (cathode electrode) / 4.6
5 / dc arc;
10 A, 4 ms / Cu (cathode electrode) / 1.1 / 1.9 / 47
C (anode electrode) / 2.5
6 / dc arc;
10 A, 4 ms / Cu (anode electrode) / 2.8 / 3.3 / 33
C (cathode electrode) / 2.1

1)Alternating current pulsed discharge

2) Direct current pulsed discharge

This coincides with the fact that the largest discrepancy (more than 80%) was observed in sample with the largest graphite electrode consumption rate (sample 3). For all samples, the synthesized powder separated into three phases, one floating in suspension, one settling at the bottom as sediment, and one as a layer of film-like material floating on the liquid surface.

The aqueous solutions of CuCl2 were discharge treated for only 20 s to acquire yellowish suspensions. The transparency of the suspensions decreased with the time during the discharge treatment. The liquids turned to dark yellow after treatment by ac-discharge for 10 min. The suspensions resulting from dc-discharge treatment were conspicuously darker when C electrode was acting as an anode. The nanoparticles suspension produced by spark and arc discharges were dark brown and dark grey respectively. It might be due to the presence of relatively large amount of carbon particles in suspension (see Table 3). The dc-discharge treated solutions were olive-green when Cu was used as the anode electrode. Yellow or green colour of suspension may indicate the oxidation of copper nanoparticles [[22]]. The presence of Cu2O nanoparticles was further confirmed by XRD analysis. No changes in colour were observed after laser irradiation of suspensions.

Figure 2 shows the absorption spectra of as prepared (a) and laser irradiated (b) suspended nanopowders synthesized by discharge treatment of aqueous solution of CuCl2 (2) for 1 min. The spectra were corrected to the contributions of solvents. The optical density increased with decrease in wavelength. Generally, the optical density of suspensions prepared by spark discharge was higher than that of suspension prepared by arc discharge. This is consistent with the fact that the nanoparticles production rate was higher when the solution was treated by spark discharge. In the spectral range of 200 – 500 nm, the optical density of the samples 1, 4, 6 was higher than that of samples 2, 3, and 5. This seems to suggest that the main parameter in determining the optical properties of suspensions was concentration of Cu-based nanoparticles. For the samples number 1 and 4, a weak absorption peak was observed at very short wavelength. According to the literature data [[23],[24]] a surface plasmon peak at wavelength of 289 nm may be attributed to the presence of very small separated Cu nanoparticles (< 4 nm in size). Though TEM examination confirmed the presence of small nanoparticles in sample 1, there were no nanoparticles with diameter less than 4 nm in sample 4. Moreover, there were no copper nanoparticles in sample 1 as revealed by the XRD (see below). More likely, the existence of weak absorption peak at 280 nm implied formation of liquid byproducts. We did not observe in the absorption spectra surface plasmon band around 570 nm. Missing of the plasmon band can be explained by copper oxidation on the particle surface [23]. This suggestion was further confirmed by XRD analysis (see below). The suspensions exhibited the same colours after laser irradiation, but absorption intensity increased for samples 3, 1 and to the lessextent for sample 5, as illustrated in Figure 2b. TEM analysis revealed the morphological similarity of irradiated samples 1, 3 and 5 (see below).

Fig. 2. Absorption spectra for the as-prepared (a) and laser modified (b) suspended nanoparticles produced by ac- (1,2) and dc- pulsed discharges (3,4,5,6). The following electrode pairs were used: Cu and C for the ac-spark (1) and ac-arc (2) discharges; Cu as a cathode electrode and C as an anode electrode for the dc-spark (3) and dc-arc (5); Cu as an anode electrode and C as a cathode electrode for the dc-spark (4) and dc-arc (6).

Figure 3 depicts the corresponding TEM images for the suspensions shown in curves 1-6 of Figure 2. Parts (a) and (b) represent the TEM views of the as-prepared and irradiated samples, respectively. Three distinct structures were observed: dark small spherical particles, dark particles surrounded by a gray shell and gray flake-like structures having diffuse contours. The small dark particles with diameter 2-5 nm were observed in samples 1, 2, 3, and 5 (marked with black ellipses in Figure 3). Some dark particles, notable when using ac spark discharge for synthesis, were bigger than 20 nm, indicating coalescence. The nanoparticles synthesized by ac arc discharge (sample 2) were surrounded by the arrowed gray regions, which were probably carbon shells, as shown in Figure 3a.

Fig.3. TEM images of nanoparticles from as-prepared (a) and irradiated (b) suspensions produced by ac- (1,2) and dc- pulsed discharges (3,4,5,6). The following electrode pairs were used: Cu and C for the ac-spark (1) and ac-arc (2) discharges; Cu as a cathode electrode and C as an anode electrode for the dc-spark (3) and dc-arc (5); Cu as an anode electrode and C as a cathode electrode for the dc-spark (4) and dc-arc (6).

As we did not have any direct evidence that the shells consisted of carbon, these nanostructures will be referred further as core-shell nanoparticles. The core-shell nanoparticles were also observed in colloid prepared by dc arc discharge between copper cathode and graphite anode (sample 5). It can be seen that core-shell nanoparticles ranged from 20 to 50 nm in diameter, while thecores within the nanoparticles varied from 8 to 25 nm. The cores were non-spherical. They seemed to compose of small particles clustered together. The flake-like structures with diffuse contours were 50 nm in size. They were observed in all samples. Samples 4 and 6 consisted mostly of structures with diffuse contours. On the basis of the above observations, the ac arc discharge and dc arc discharge with copper anode electrode seemed to be more suitable for synthesis of nanoparticles with core-shell structure.