ISSN 0036-0236, Russian Journal of Inorganic Chemistry, 2008, Vol. 53, No. 1, pp. 55–59. © Pleiades Publishing, Ltd., 2008. Original Russian Text © G.A. Kozhuhovskaya, G.M. Rybachenko, A.K. Starkov, V.I. Kazbanov, 2008, published in Zhurnal Neorganicheskoi Khimii, 2008, Vol. 53, No. 1, pp. 60–65.

COORDINATION COMPOUNDS

Influence of Temperature Gradients on the Composition of Crystallization Products of Ammonium Chloroplatinum(IV) and Chlororhodium(III) Complex Salts G. A. Kozhuhovskaya, G. M. Rybachenko, A. K. Starkov, and V. I. Kazbanov† Institute of Chemistry and Chemical Technology, Siberian Division, Russian Academy of Sciences, Krasnoyarsk, Russia e-mail: [email protected] Received October 10, 2006

Abstract—The influence of temperature gradients on the particle size and component composition of solid phases in heterogeneous systems of ammonium chloroplatinum(IV) and chlororhodium(III) complex salts was studied. A Benard cell was implemented as applied to not only homogeneous but also heterogeneous systems. Two contributions to the cocrystallization of rhodium with platinum are distinguishable and kinetically resolved: the formation of nonequilibrium solid solutions by ammonium chlorometalates and equilibrium Ostwald ripening. The elucidation of both factors helps in finding crystallization parameters for minimizing the rhodium contamination of platinum in refining practice. DOI: 10.1134/S0036023608010099 †

More severe requirements imposed on the quality of refined platinum-metal products and refining depth necessitate further progress in the physicochemical grounds of relevant processes. Salt crystallization is the most frequent method for separating and refining platinum-group metals. Crystallization stages are nonequilibrium processes. Equilibrium inorganic systems are well documented, whereas data on the influence of disequilibration on the composition and properties of inorganic products are scarce.

preparing the most pure platinum. For this purpose, we studied crystallization in the system PtCl4–RhCl3– NH4Cl–H2O (0.1 M HCl) and, for reference, in the system PtCl4–NH4Cl–H2O (0.1 M HCl); in particular, we elucidated the effect of temperature gradients on the particle-size and chemical compositions of the crystallization products.

Crystal growth is a complex process, described by mass and heat transfer, chemical reactions, and phase transitions. It is believed that crystallization is preceded by the appearance of a nonequilibrium, heterogeneous phase between the solid and liquid regions, the properties of which differ from the properties of both the melt and the solid [1]. These properties dictate the qualitative transition of the system to a two-phase state. During crystallization in real technological processes, a system does not reach equilibrium; therefore, the influence of disequilibration on the composition and properties of crystalline products is an interesting and challenging problem for study.

The starting complex salts (NH4)2[PtCl6] and (NH4)3[RhCl6] · H2O were synthesized as described in [2]. The results of their analysis are displayed in Table 1. A chemically pure grade NH4Cl sample was used after double recrystallization. The complexes isolated were identified by elemental analysis, differential thermal analysis, IR spectra, electronic absorption spectra of aqueous solutions, and X-ray powder diffraction. IR spectra were recorded as KBr disks in the range 400–3600 cm–1 on a Specord IR-75 spectrometer. The characteristic frequencies matched the literature data. Electronic absorption spectra of freshly prepared solutions were recorded on a Specord UV-Vis spectrophotometer in the range 200–800 nm. The positions of absorption peaks matched the literature data. X-ray powder diffraction analysis was performed on a DRON-4 diffractometer in a quartz cell using monochromated radiation: for (NH4)2[PtCl6] ëÓäα with λeff = 1.79021 Å; and for (NH4)3[RhCl6] · H2O ëÓäα with λeff = 1.54180 Å. X-ray diffraction patterns were

EXPERIMENTAL

The necessity for improving the level of platinum refining from its associates, in particular rhodium and iridium, remains topical. Therefore, we chose to study ammonium salts of chloroplatinum(IV) and chlororhodium(III) complexes, whose equilibrium crystallization had been studied. We intended to study crystallization of ammonium chlorometalates under temperaturegradient conditions in order to find new methods for †

Deceased.

55

56

KOZHUHOVSKAYA et al.

Table 1. Elemental analysis of the reagents Compound (NH4)2[PtCl6] (NH4)3[RhCl6] · H2O

Metal Found, wt % Calcd., wt % Found, wt % Calcd., wt %

43.90 (0.17; 3)* 43.95 26.60 (0.50; 2) 26.54

+

Cl 47.98 (0.50; 3) 47.92 54.40 (1.30; 2) 54.86

N H4

8.05 (0.12; 3) 8.13 – –

* In parentheses, there are standard deviations and the number of degrees of freedom.

indexed to calculate unit cell parameters, which matched the literature data. Differential thermal analysis was carried out on a Q1000 Paulik-Paulik-Erdey derivatograph. Optical observation of crystals was performed in transmitting light on a Boethius RNMK 0.5 microscope. Observation showed that the product ammonium hexachloroplatinate(IV) was obtained as bright yellow crystals shaped as tetragonal prisms and blunt octahedra. No foreign impurities were observed. Ammonium hexachlororhodiate(III) monohydrate was obtained as transparent cherry-colored tetragonal- or hexagonalprismatic crystals without foreign impurities. An OR-207 potentiometer with EVL-IM3 and platinum electrodes was used in titrimetric analysis. Nonequilibrium crystallization in the system PtCl4−NH4Cl–H2O (0.1 M HCl) was studied as follows. Nonequilibrium conditions can be organized by a continuous energy flux in the reactor, e.g., by creating a temperature gradient. To set ∆T and study crystallization under these conditions, the following setup was used. A cooler was tightly mounted on top of a tube, which was cooled from the bottom. The outer diameter of the cooler was 68 mm; the outer diameter of the reactor was 74 mm. The reactor volume was 80 mL. Benard instability, a structured space with looped convective flows generated by a vertical temperature gradient in the volume, can be implemented in this reactor. An (NH4)2[PtCl6] sample was placed into the inner tube (reactor), 0.1 HCl (80 mL) was added, and the salt was dissolved with heating. Then, an NH4Cl sample was added and dissolved, and the reactor was immediately transferred onto a boiling water bath (the bottom of the reactor touched the water mirror of the bath), and the cooler was sunk into the reactor (the bottom of the cooler touched the mirror of the reaction solution). The component sample sizes in the reaction mixture are listed in Table 2. The component molar ratio (Pt : NH4Cl = 1 : 15) was maintained roughly constant, as recommended in [3], in order to provide the optimal conditions for (NH4)2[PtCl6] crystallization. The constancy of the external temperature gradient is determined by the constancy of the melting temperature of ice and the boiling temperature of water. The actual gradient inside the reaction volume is less than 100°ë because of the limited thermal conductivity of glass; this is a near-steady-state value.

The temperature gradient was maintained constant for set time intervals. Then, the cooler was rapidly removed, and the reaction solution was immediately filtered off. The precipitate was washed two times with ethanol and dried in air for 1 week. For comparison, experiments were carried out under equilibrium conditions (the solid phase was in contact with the mother solution for 3 weeks at T = 25 ± 0.1°C) and under severely nonequilibrium conditions, as follows: liquid nitrogen was poured to the reactor; the solution completely froze after 5 min; it was allowed to thaw for 30 min, then the precipitate was filtered off and treated as in preceding experiments. (NH4)2[PtCl6] crystals precipitated after drying were weighed (Table 2), and the percentage of the salt remaining in the solution was calculated. After 1 h of residence in the reactor, the concentration of the dissolved salt acquired a steady-state (within the error bar) value (about 20%). The crystalline precipitates were studied by energydispersive analysis [4] on a Boethius RNMK 0.5 microscope with direct observation of images on the projector screen. Sampling was according to [4]. Crystallization in the system PtCl4–RhCl3–NH4Cl– H2O (0.1 M HCl) was studied in the same manner. A reactor 50 mL in capacity, with the outer cooler diameter equal to 60 mm and the inner diameter of the reactor equal to 67 mm, was used in the experiments. The batch molar ratios were Pt : Rh : NH4Cl = 1 : 1 : 15. (NH4)2[PtCl6] and (NH4)3[RhCl6] · H2O samples (Table 2) were dissolved in 50 mL of 0.1 M HCl with heating and NH4Cl added, and the setup was rapidly assembled. After the set time, the solution was immediately filtered; the precipitate was washed two times with ethanol, dried in air for 1 week, and then weighed (Table 2). For comparison, experiments were carried out under equilibrium conditions (the solution was stored in a thermostat at 25 ± 0.1°ë for 3 weeks) and under severely nonequilibrium conditions (liquid nitrogen was poured into the reactor, and the frozen solution was allowed to thaw for 30 min). Then, solutions were rapidly filtered; the precipitates were washed two times with ethanol, dried in air for 1 week, and weighed (Table 2). Syntheses followed deliberately chosen schedules. A hot solution after charging was cooled either in air at room temperature

RUSSIAN JOURNAL OF INORGANIC CHEMISTRY

Vol. 53

No. 1 2008

INFLUENCE OF TEMPERATURE GRADIENTS ON THE COMPOSITION

57

Table 2. Results of experiments Precipi- Components in liquid phase, g D × 10–3 ± ∆ (Rh ± ∆), % Parametate ters weight, ∆, confidence level; (NH4)2[PtCl6] (NH4)3[RhCl6] · H2O NH4Cl (NH4)2[PtCl6] (NH4)3[RhCl6] · H2O g p = 0.95 Batch components, g

20 min 40 min 1h 2h 3h 3h 3.5 h 4.5 h 5h 504 h, 25°C 25°C 0°C –196°C

0.2454 0.2447 0.2570 0.2674 0.2449 0.2609 0.2441 0.2355 0.2506 0.2525

0.2182 0.2158 0.2136 0.2293 0.2138 0.2170 0.2115 0.2155 0.2192 0.2210

0.4598 0.4614 0.4580 0.4737 0.4801 0.4421 0.4521 0.4510 0.4555 0.4511

0.0739 0.1144 0.1206 0.1442 0.1201 0.1070 0.1275 0.1055 0.1266 0.2077

0.1715 0.1303 0.1364 0.1232 0.1248 0.1539 0.1166 0.1300 0.1240 0.0448

0.2180 0.2151 0.2311 0.2284 0.2129 0.2165 0.2112 0.2152 0.2186 0.2181

1.77 ± 0.03 3.64 ± 0.04 2.22 ± 0.03 3.24 ± 0.04 4.18 ± 0.03 3.20 ± 0.04 1.47 ± 0.04 1.47 ± 0.04 2.55 ± 0.04 2.85 ± 0.07

0.2495 0.2575 0.2508

0.2206 0.2015 0.2205

0.4430 0.2245 0.4461 0.2050 0.4467 0.2070

0.0250 0.0525 0.0438

0.2121 0.2181 0.2127

0.80 ± 0.04 0.220.02 8.07 ± 004 0.800.02 7.73 > 0.5–1

for 1 h, which approaches the parameters of (NH4)2[PtCl6] crystallization, or in a flask with ice for 30 min. The precipitates were treated as in the preceding experiments. Crystalline precipitates were studied using energydispersive analysis as in [4] (the number of measurements was 289) and emission spectral analysis (DFS-8 spectrograph). RESULTS AND DISCUSSION The preceding step was the study of crystallization in the system PtCl4–NH4Cl–H2O (0.1 M HCl) in a nearequilibrium state at 25°ë [5]. Equilibration was monitored by the absorption spectra of the liquid phase in the UV and visible. The solubility isotherm was calculated, and the crystallization products were characterized by chemical analysis and X-ray powder diffraction. Complex thermal analysis verified the absence of unbound + water. The platinum and N H 4 percentages were lower than the calculated values because of an NH4Cl impurity. Paper and thin-layer chromatography, pH titration, and coulometric titration showed the complex [Pt(H2O)Cl5]2–, which has acidic properties. Thus, equilibria in the system PtCl4–NH4Cl–H2O are complicated by hydrolysis. Crystalline phases were examined under an optical microscope at 400–600-fold magnification in the transmitting polarized light. The crystals were shaped as yellow blunt octahedra. The average crystal size decreased two- to threefold with increasing ammonium chloride percentage in the system. Anisotropic inclusions were RUSSIAN JOURNAL OF INORGANIC CHEMISTRY

0.060.01 0.160.02 0.100.01 0.160.02 0.190.01 0.120.02 0.070.02 0.070.02 0.120.02 0.370.05

observed in the polarized light as luminescent dots and strips. Electronic microscopy showed the existence of an autonomous microdisperse crystalline phase and the nonexistence of superstructures. X-ray powder diffraction also showed two phases. Reflection broadening makes it impossible to identify the autonomous crystalline phase of NH4Cl. Such objects are conventionally considered as microdisperse mixed crystals. Thus, the crystals obtained by the standard precipitation with ammonium chloride from hydrochloroplatinic acid solutions are mixtures of (NH4)2[PtCl6] and a microdisperse NH4Cl phase. The component composition of the mixed crystals was found as 98.8% ammonium hexachloroplatinate and 1.2% ammonium chloride. The results obtained for the solid phases of the nonequilibrium system PtCl4–NH4Cl–H2O (0.1 M HCl) were used to plot and analyze bar diagrams. Considering the grain-size dynamics of the crystallization products as a function of time, one can see that the particle size in the temperature-gradient mode increases with time; i.e., Ostwald ripening occurs. Along with the appearance of coarser particles, a fraction with the minimum diameter (up to 30 µm) always exists in the system, likely because of disequilibration. Upon cooling with liquid nitrogen, the particle-size distribution narrowed; the particle diameter was smaller than that in the equilibrium experiment or in the temperature-gradient experiment. This was due to the greater disequilibration of the process and the small crystallization duration.

Vol. 53

No. 1

2008

58

KOZHUHOVSKAYA et al.

Analysis of bar diagrams for the solid phases of the system PtCl4–RhCl3–NH4Cl–H2O (0.1 M HCl) showed that the crystal size increased with the time of residence in the temperature-gradient mode, because of Ostwald ripening. Because of disequilibration, however, a fineparticle fraction was always present in the system. During cooling with liquid nitrogen and ice, the crystal size decreased, and the distribution narrowed compared to that in the experiments performed under equilibrium conditions, under cooling at –25°C, or in the steady-state gradient mode. Emission spectral analysis data and the cocrystallization factors D calculated as a function of the duration of the experiment are compiled in Table 2. The cocrystallization factors and rhodium percentages are listed with the confidence level indicated in accordance with the precision verified in 3-h experiments. The cocrystallization factor D was calculated from [6] c impurity in solid ⋅ c crystallizant in liquid -. D = ------------------------------------------------------------------c impurity in liquid ⋅ c crystallizant in solid Having expressed concentrations in g/L and canceled the volumes, we arrived at RhCl 3 ⋅ PtCl 4 . Q RhCl3 ⋅ Q PtCl4 -, D = -------------------------------l s Q RhCl3 ⋅ Q PtCl4 B

l

where QB is the weight of the formal component in the solid, g; %Pt in solid ⋅ Q B s Q PtCl4 = ---------------------------------, %Pt in PtCl4 QB is the precipitate weight, g; and Q l and Q s are the component percentage in the liquid and solid phase, respectively; %Rh in solid ⋅ Q s s -, Q RhCl3 = --------------------------------%Rh in RhCl3 where %Pt in PtCl4 = 57.91; %Pt inPtCl3 = 49.17; Q l is the formal-component concentration in the liquid phase, g; Q l = Q – Qs; %Pt init 0 -, Q PtCl4 = Q ( NH4 )2 [ PtCl6 ] ⋅ ----------------------%Pt in PtCl4 where %Ptin = 43.94; Q0 is the batch weight, g; %Rh init 0 -, Q RhCl3 = Q ( NH3 )4 [ RhCl6 ] ⋅ H2 O ⋅ -------------------------%Rh in RhCl3 where % Rhin = 26.60. Optical observation of the precipitates obtained in the gradient mode in the system PtCl4–RhCl6–NH4Cl– H2O (0.1 M HCl) revealed impurity green crystals of octahedral, hexagonal-prismatic, or trigonal-prismatic habit. The amount of the impurity was approximately 1–3 particles per 800 crystals of the major phase. The average size of green crystals was 48.6 µm. With reference to [3, 7, 8], the green impurity was by analogy

IV

IV

assigned to (NH4)2[ Pt 1 – x , Rh x )Cl6 with x = 0–0.14, whose formation was studied in the equilibrated platinum–rhodium system. The study of the chemical composition of crystallization products in the system PtCl4–RhCl6–NH4Cl– H2O (0.1 M HCl) showed the following (Table 2): (i) the cocrystallization factor and the rhodium concentration in the precipitate are constant to the error bar regardless of the residence time in the temperature-gradient mode; (ii) liquid-nitrogen or ice cooling increases the rhodium concentration in the precipitate and the cocrystallization factor compared to the experiments performed under equilibrium, under cooling at 25°C, or in the steady-state temperature-gradient mode, likely, due to heavy disequilibration. The best possible result of the study of crystallization in the systems PtCl4–NH4Cl–H2O (0.1 M HCl) and PtCl4–RhCl6–NH4Cl–H2O (0.1 M HCl) in disequilibrium is our finding the crystallization mode that provides for a controlled and narrow grain-size distribution and the smallest rhodium concentration in the precipitate (the system PtCl4–RhCl6–NH4Cl–H2O (0.1 M HCl)). Preliminary experiments showed that the main difficulty was to find the parameters providing for the kinetic resolution of the nucleation and primary crystal growth stage from the Ostwald ripening stage. This work attempted to solve this problem through organizing a temperature gradient and varying cooling modes and crystallization durations. The study of the platinum and platinum–rhodium systems in the gradient mode showed that, regardless of the crystallization duration, it is impossible to avoid the contribution from Ostwald ripening. An expression of Ostwald ripening was a considerable broadening of the particle-size distribution curves. The narrowest distribution with the least grain size was achieved in experiments with short crystallization times (cooling with liquid nitrogen and ice). It was, however, impossible under the conditions studied to control the particle size, e.g., to shift the distribution toward greater sizes without considerable broadening. Likely, the Ostwald ripening contribution has not been avoided because of the unfavorable ratio of the rates during the crystallization stage and the absence of convective transport of the whole precipitate mass along the temperature gradient under the crystallization conditions employed. After reaching a certain critical size, grains descended to the lower zone of the reactor, where they experienced Ostwald ripening and coarsening. A fine-grain fraction circulated in the solution. This fraction inevitably occurred in crystalline precipitates regardless of the crystallization duration, and the kinetic resolution of stages was likely implemented for this fraction. The grain-size dynamics of precipitates was the same for both systems studied. The only distinction was that the Ostwald ripening contribution was far lower in the presence of the

RUSSIAN JOURNAL OF INORGANIC CHEMISTRY

Vol. 53

No. 1 2008

INFLUENCE OF TEMPERATURE GRADIENTS ON THE COMPOSITION

rhodium salt. Addition of ammonium hexachlororhodiate(III) monohydrate decreased the grain size and narrowed the distribution. Comparing the cocrystallization factors and rhodium concentrations in the precipitates obtained under different conditions, we can infer that the increases in the rhodium percentage in the precipitate and in the cocrystallization factor upon cooling by liquid nitrogen and ice correlate with the increase in the fine-grain fraction in the crystalline product. The fine-grain fractions contained far more rhodium than the coarser crystals formed during Ostwald ripening. In this case, we can speak of nonequilibrium solid solutions of ammonium chlorometalates. During a long exposure of a heterogeneous system to a near-equilibrium state, however, the cocrystallization factor and rhodium percentage in solid residues also increased. Here, undesirable effects were generated by the long-term Ostwald ripening and the formation of solid solutions by platinum-metal complex compounds with various oxidation numbers. There are certain optimum crystallization parameters that provide the minimum rhodium contamination of the platinum salt. In summary, two contributions are distinguishable in rhodium coprecipitation with platinum complex salts. Ostwald ripening generates coarse crystals, which are solid solutions of ammonium chloroplatinate(IV), ammonium chlororhodiate(III), and ammonium chlororhodiate(IV). A fine-grain fraction, which is noticeable in the heterogeneous system, has an elevated platinum-and-rhodium cocrystallization factor. The occurrence of rhodium in the system is likely due to the nonequilibrium occlusion and large integral surface of fine grains. In this work we managed to experimentally implement the Benard cell as applied to heterogeneous liquid–solid systems, rather than to a homogeneous system, namely for the system PtCl4–NH4Cl–H2O (0.1 M

RUSSIAN JOURNAL OF INORGANIC CHEMISTRY

59

HCl) and the system PtCl4–RhCl3–NH4Cl–H2O (0.1 M HCl). We demonstrated the feasibility of the mode in which crystallization and Ostwald ripening are kinetically resolved. The steady-state solubility of ammonium hexachloroplatinate was studied in the Benard cell. The steadystate solubility have values intrinsic to this temperature-increment mode and differs from the solubility of the salt at the marginal temperatures of this study. The results of our study of the effect of disequilibration on crystallization in the systems PtCl4–NH4Cl– H2O (0.1 M HCl) and PtCl4–RhCl3–NH4Cl–H2O (0.1 M HCl) orient us to search further for regimes that provide for convective transport of the whole mass of the crystalline product and the creation of the crystallization conditions under which a controlled, narrow, and uniform particle-size distribution is achievable and ammonium hexachloroplatinate(IV) precipitates contain a minimum rhodium impurity. REFERENCES 1. G. Nikolis and I. Prigogine, Exploring Complexity: an Introduction (Freeman, New York, 1989; Mir, Moscow, 1990). 2. Synthesis of Complexes of the Platinum-Group Metals: A Handbook (Nauka, Moscow, 1964) [in Russian]. 3. I. B. Usmanov, Candidate’s Dissertation in Chemistry (1990). 4. G. A. Kouzov, Fundamentals of Dispersion Analysis of Industrial Dusts and Ground Materials (Khimiya, Moscow, 1974) [in Russian]. 5. I. B. Usmanov, N. N. Anshits, and V. I. Kazbanov, Koord. Khim. 14 (1), 81 (1988). 6. I. V. Melikhov and M. S. Merkulov, Cocrystallization (Khimiya, Moscow, 1975) [in Russian]. 7. I. B. Usmanov, V. I. Kazbanov, and G. M. Rybachenko, Koord. Khim. 17 (11), 1547 (1991). 8. I. Usmanov, V. Kazbanov, and G. Ribachenko, Mend. Comm 3, 81 (1992).

Vol. 53

No. 1

2008