I. INTRODUCTION
In spite of the small stocks in the Earth's crust, iridium is attractive for use in applications such as biology (Liu and Sadler, Reference Liu and Sadler2014; Zhong et al., Reference Zhong, Lu, Leung, Wong, Peng, Yan, Ma, Cai, Wand and Leung2015) and catalysis (Firdous et al., Reference Firdous, Janjua, Qazi and Wattoo2015; Slavcheva et al., Reference Slavcheva, Borisov, Lefterova, Petkucheva and Boshnakova2015; Tetlow et al., Reference Tetlow, de Boer, Ford, Vvedensky, Curcio, Omiciuolo, Lizzit, Baraldi and Kantorovich2016). The general approach which is used to lower the cost of the final product is a reduction of the content of the active phase in the sample because of an increase of specific surface and dilution with the cheaper metal, i.e. obtain solid solutions of metals (Potemkin et al., Reference Potemkin, Filatov, Zadesenets, Snytnikov, Shubin and Sobyanin2012; Firdous et al., Reference Firdous, Janjua, Qazi and Wattoo2015). Mentioned approach can be implemented by yielding nanoscale solid solutions (nanoalloys) by thermolysis of the single-source precursors such as a double-complex salts (DCS) (Martynova et al., Reference Martynova, Filatov, Korenev, Kuratieva, Sheludyakova, Plusnin, Shubin, Slavinskaya and Boronin2014; Asanova et al., Reference Asanova, Asanov, Zadesenets, Filatov, Plyusnin, Gerasimov and Korenev2016). The salts with oxalate anions are particularly advantageous as they have pronounced reductive properties. Synthesis of DCS simultaneously containing a first raw transition metal and a platinum family metal combined with oxalate as the ligand is appealing, for these compounds can serve as precursors of corresponding nanocrystalline alloys. Such materials can be used for development of novel catalytic systems having a reduced content of noble metals (Potemkin et al., Reference Potemkin, Semitut, Shubin, Plyusnin, Snytnikov, Makotchenko, Osadchii, Svintsitskiy, Venyaminov, Korenev and Sobyanin2014; Vedyagin et al., Reference Vedyagin, Volodin, Stoyanovskii, Kenzhin, Slavinskaya, Mishakov, Plyusnin and Shubin2014; Plyusnin et al., Reference Plyusnin, Makotchenko, Shubin, Baidina, Korolkov, Sheludyakova and Korenev2015).
In our previous work, the structure and thermal decomposition of the [Co(NH3)6][Ir(C2O4)3] compound have been studied in helium and hydrogen atmospheres (Filatov et al., Reference Filatov, Yusenko, Vikulova, Plyusnin and Shubin2009). The point of this paper is an examination and correct description of the disordered structure of final product of this DCS thermolysis—equimolar nanoalloy CoIr with domain structure.
II. EXPERIMENTAL
A. Preparation of the precursor and nanoalloys
The synthesis of the precursor was denoted in the work (Filatov et al., Reference Filatov, Yusenko, Vikulova, Plyusnin and Shubin2009). To synthesize bimetallic powders in a hydrogen atmosphere, weighted samples of the DCS (~50 mg) were placed in a quartz boat and put in a tubular quartz reactor. Heating was performed in a split furnace at the rate of 20 °C min−1. The temperature was measured with a thermocouple embedded in the oven and adjusted by Omron temperature controller. Additional external thermocouple was used to monitor the correspondence of the sample temperature in the furnace and a specified temperature.
After reaching the final temperature (500 °C in this work), the samples were tempered for 1 h, and then the hydrogen stream was switched off and the system was purged with helium for 30 min. Afterward, the heater was removed and the reactor was allowed to cool to ambient temperature in a continuous helium stream.
B. Instrumentation
Powder X-ray diffraction (XRD) studies of thermolysis products of the prepared compound were made over the 2θ range 5°–120° (step 0.03, collecting time is 3 s) on a DRON-RM4 diffractometer (CuKα-radiation, graphite monochromator on the diffracted beam, ambient temperature). The refinement of lattice parameters was performed by the full profile technique applied to full-range diffraction data using PowderCell 2.4 program (Kraus and Nolze, Reference Kraus and Nolze2000). The crystallite sizes of the metal phases were determined by the Scherrer equation [WINFIT 1.2.1 (Krumm, Reference Krumm1995)]. High-resolution transmission electron microscopy (HRTEM) measurements were performed using a JEM-2010 electron microscope (lattice plane resolution 0.14 nm at an accelerating voltage of 200 kV). Images of periodic structures were analyzed by Fourier method. Local energy-dispersive X-ray analysis (EDXA) was carried out using an EDX spectrometer (EDAX Co.) fitted with a Si (Li) detector with a resolution of 130 eV. The error of the metal particle composition determination typically was not <0.05 at.%. The samples for the HRTEM study were prepared on perforated carbon film mounted on a copper grid.
III. RESULTS AND DISCUSSION
The transition temperature Cohcp → Cofcc is equal to 420 °C. The iridium (Ir) has face-centered cubic (fcc) structure in all temperature range from room up to melting one. Thus, Köster and Horn said that cobalt (Co) and Ir form a continuous solid solution in the fcc phase and the homogeneous hexagonal close-packed (hcp) solid solution extends to ~63 at.% Ir at 500 °C (Köster and Horn, Reference Köster and Horn1952).
Relatively fast (0.5 h) heating of the sample of the complex [Co(NH3)6][Ir(C2O4)3] to 500 °C and further exposure of thermolysis product to this temperature for 1 h results in the formation of CoIr alloy with an unusual XRD pattern. The observed diffraction pattern (see Figure 1) could be interpreted as a pattern of the two-phase system because peaks of fcc and hcp modifications of Co–Ir solid solution are observed. However, it can be seen that 101 peak of hcp modification is substantially wider than 100 and 002 peaks, 102 and 103 are very broad and almost invisible (see Table I). Peak 200 of fcc structure is wider than the other peaks of this modification and slightly shifted toward lower angles. It has been suggested that the particles of the sample have the domain structure where domains of fcc structure alternate with domains of hcp structure as it was in the particles of pure Co (Tsybulya et al., Reference Tsybulya, Cherepanova, Khasin, Zaikovskii and Parmon1999; Cherepanova and Tsybulya, Reference Cherepanova and Tsybulya2004; Cherepanova, Reference Cherepanova2012).
Figure 1. (Color online) Experimental XRD pattern (solid curve) and simulated ones to fcc (squares) and hcp (triangles) structures with crystallite size about 10 nm.
Table I. The peak broadening [full width at half maximum (FWHM)] for the reflections of domain structure.
To test this assumption an appropriate simulations of diffraction patterns using respective software have been done (Cherepanova and Tsybulya, Reference Cherepanova and Tsybulya2004). The calculations are carried out on the basis of the model of one-dimensionally (1D) disordered crystal. The model represents a statistical sequence of the two-dimensionally periodic layers.
The domain structure of Co–Ir alloy was built with Co–Ir layers in hexagonal metric as in hcp structure. Unit-cell parameters determining the layer structure were calculated from the refined unit-cell parameter of fcc structure a
cub = 3.6843 Å using the formulas a = b = a
cub/
$\sqrt 2 $
= 2.605 and c = 2a
cub/
$\sqrt 3 $
= 4.254 Å. The c unit-cell parameter determines the thickness of two close-packed layers AB. This two-layered fragment is convenient to choose as one layer in the model. Unit cell contains two atoms in positions M1(0, 0, 0) and M2(2/3, 1/3, 1/2), which are statistically occupied by Co and Ir atoms (Co:Ir = 1). Structure of cubic modification AB–CA–BC… can be constructed by sequential tangential displacement of the AB layer on the vector (1/3, 2/3). Structure of hexagonal modification AB–AB… can be simulated by superposition of AB layer on itself without tangential shifts. Thereby both modifications can be simulated by setting the structure of AB layer and two different ways of superposition of the AB layer on itself. Tangential displacements of the AB layer on the vectors (1/3, 2/3) and (0, 0) were marked as first and second ways of superposition of adjacent AB layers. Average thicknesses of domains having fcc and hcp structures were varied by two statistical parameters: the probability of occurrence of the second superposition way W
2 and the conditional probability of repetition of the second superposition way after second superposition way P
22.
Let us consider sense of statistical parameters. W 1 and W 2 are the probabilities of appearance of two-layered fragments of fcc structure (AB–CA or CA–BC or BC–AB) and hcp structure (AB–AB or CA–CA or BC–BC). Parameters W 11 = W 1•P 11 and W 22 = W 2•P 22 are the probabilities of existence of three-layered fragments of fcc (for example, AB–CA–BC) and hcp (for example, AB–AB–AB) structures respectively. W 111 = W 1•P 11•P11 and W 222 = W 2•P 22•P 22 are the probabilities of appearance of four-layered fragments of fcc and hcp structures respectively and so on. Therefore, parameters P 11 and P 22 determine the average thicknesses of fcc and hcp domains, respectively. Parameters W 1 and W 2 define the ratio of the average thickness of the domains having fcc structure to the average thickness of the domains having hcp structure (W 1/W 2 = (1 − W 2)/W 2). It is enough to define only two parameters: W 1 and P 11 or W 2 and P 22. As it was mentioned earlier we varied W 2 and P 22.
It is interesting to analyze how inclusions of hcp domains into fcc matrix influence on the diffraction patterns. First, we fixed the parameter P 22 = 0.8 and let the parameter W 2 sequentially increases up to 0.2, 0.4, and 0.6. These values of parameters correspond to the average thickness of the domains having hcp structure equal to 2.1 nm and to the ratio of the average thickness of the domains of the fcc structure to the average thickness of the domains of the hcp structure: 0.8/0.2, 0.6/0.4, and 0.4/0.6. Thereby the average thickness of the domains of the fcc structure was 8.4 nm [Figure 2(b)], 3.2 nm [Figure 2(c)], and 1.4 nm [Figure 2(d)], respectively. One can see that inclusion of hcp lamellar domains into fcc structure leads to the appearance of peaks additional with respect to the peaks of the fcc structure [marked on Figure 2(a)]. These superstructural peaks arise at the positions of the peaks characteristic of the hcp structure [marked on Figure 2(b)] or near them. The higher is the value of W 2, the more intensive are the additional peaks of diffuse scattering. The width of the 111, 220, and 311 peaks hardly changes. The intensity of the 111 peak gradually decreases, the 200 peak gradually shifts toward lower angles and broadens with W 2 increase.
Figure 2. XRD patterns simulated for 10 nm size particles having (a) fcc structure; (b–h) domain structure; average thicknesses of the domains of the fcc and hcp structures: (b) 8.4 and 2.1 nm; (c) 3.2 and 2.1 nm; (d) 1.4 and 2.1 nm; (e) 0.9 and 0.9 nm; (f) 1.1 and 1.1 nm; (g) 1.4 and 1.4 nm; (h) 2.1 and 2.1 nm; (i) experimental XRD pattern.
Then, we fixed the parameter W 2 = 0.5 and varied the parameter P 22 (0.5, 0.6, 0.7, and 0.8). These values correspond to the same average thickness of the two types of domains, which are equal to 0.9 nm [Figure 2(e)], 1.1 nm [Figure 2(f)], 1.4 nm [Figure 2(g)], and 2.1 nm [Figure 2(h)], respectively. Figures show that very small average thicknesses of domains (0.9 nm) produce peaks of diffuse scattering shifted toward smaller angles relative to the positions of the 200 (fcc) and 101 (hcp) peaks. When the average thicknesses of fcc and hcp domains are about 1.1 nm, one broad peak of diffuse scattering is observed in the angle range 43°–52°. Increase in the average thickness of domains (1.4 and 2.1 nm) leads to the appearance of the diffuse peaks in the positions of the 200 (fcc) and 101 (hcp) peaks.
To fit the experimental XRD data [Figure 2(i)] the parameters W 2 and P 22 were varied with step of 0.05 and corresponding XRD patterns were simulated. Calculated R-factors are listed in Table II. The best fit was obtained for a model in which particles have about 10 nm size, average thicknesses of lamellar domains having fcc and hcp structures are 1.7 and 1.1 nm, respectively (Figure 3).
Figure 3. (Color online) Experimental (dotted curve) and simulated (solid curve) XRD patterns (average crystallite size 10 × 10 × 10 nm3, R = 6.4%).
TABLE II. R-factors (%) calculated for different values of W 2 and P 22 parameters.
For confirmation of the XRD results samples were investigated by the TEM. Figure 4 shows that particles of Co–Ir alloy consist of two type lamellar domains having fcc and hcp structures. The ratio of Co:Ir measured by the EDXA is 50:50 at.% with deviation up to 1 at.% obtained at the different locations of sample.
Figure 4. TEM photographs of nanoparticles CoIr with the domain structure.
Authors think that it is necessary to make an additional in situ XRD experiment to confirm that this domain structure is formed in the thermolysis process at and not during cooling.
IV. CONCLUSION
The XRD pattern of nanosized equimolar solid solution CoIr prepared by thermolysis of [Co(NH3)6][Ir(C2O4)3] contains peaks characteristic of both fcc and hcp structures. Moreover, 101 peak of hcp modification is substantially wider than 100 and 002 peaks, 102 and 103 are very broad and almost invisible. Peak 200 of fcc structure is wider than the other peaks of this modification and slightly shifted toward lower angles. It was shown by the simulation of XRD patterns that particles of the CoIr alloy are nanoheterogeneous and consist of lamellar domains having fcc and hcp structures. The best fit was obtained for the following model parameters: an average crystallites size is about 10 nm, average thicknesses of the fcc and hcp domains are 1.7 and 1.1, respectively. The presence of domain structure was confirmed by TEM data.
ACKNOWLEDGEMENT
This work was supported by the Russian Foundation for Basic Research (Grant no. 14-03-00129-a).
