ISSN 19907931, Russian Journal of Physical Chemistry B, 2009, Vol. 3, No. 8, pp. 1172–1186. © Pleiades Publishing, Ltd., 2009. Original Russian Text © S.P. Gubin, E.Yu. Buslaeva, 2009, published in Sverkhkriticheskie Flyuidy: Teoriya i Praktika, 2009, Vol. 4, No. 4, pp. 73–96.

Supercritical Isopropanol as a Reducing Agent for Inorganic Oxides S. P. Gubin and E. Yu. Buslaeva Kurnakov Institute of General and Inorganic Chemistry, Russian Academy of Sciences, Leninskii pr. 31, Moscow, 117907 Russia Received May 27, 2009

Abstract—Preparative methods for the reduction of simple and complex metal oxides by supercritical isopro panol (SCI) were developed. Procedures for effective work with SCI under usual laboratory conditions were suggested. Optimum reaction conditions (temperature and pressure) and reagent ratios for reactions between SCI and metal oxides were found. Disperse oxides coated by finedispersity metals that could be used as cat alysts were prepared. Simple methods for obtaining metal nanoparticles by the reduction in situ of metal oxide nanoparticles stabilized in polyethylene and synthetic silica (opal) matrices with SCI were developed. Key words: supercritical fluids, isopropanol, inorganic oxides, nanoparticles, single crystals. DOI: 10.1134/S1990793109080077

INTRODUCTION Inorganic oxides have long been extensively studied because of a broad spectrum of their uses in practice. The unique physical characteristics of inorganic oxides (magnetic, optical, electrical, etc.) determine the continuous interest of researchers in their struc ture and reactivity. In this area of materials science, there are several problems that attract attention. These are the development of methods for the preparation of suboxides, the formation of cluster structures, fine control of the oxygen stoichiometry of variablecom position phases, methods for controlling the composi tion and structure of surface layers of metal oxide sin gle crystals with a complex composition, intercalation of organic compounds, hydrogen, etc. into oxide matrices, controlled reduction of complex oxides for the preparation of substances with the required ratio between d metals in various degrees of oxidation, etc. It is known that small changes in the composition of binary oxides can result in essential changes in their electrophysical characteristics and other physical properties. Many unique physical characteristics of metal oxides (electrical, magnetic, and optical) depend on their nonstoichiometry (imperfection). Oxygen nonstoichiometry determines conductivity and conductivity type, nonmetal–metal phase transi tion temperature, the temperature of the transition to the superconducting state (Tc) in oxide highTc super conductors, and some other properties. There are cor relations between the type of magnetic ordering (mag netic susceptibility) and oxygen stoichiometry in oxides with magnetic properties. In addition to physi cal properties, the catalytic activity and reactivity of solid oxides depend substantially on nonstoichiome

try. This relates to “solid–solid,” “solid–gas,” and “solid–liquid” reactions. Changes in the oxygen sto ichiometry of oxides under mild conditions without the introduction of impurities is an important problem of modern materials science. At the same time, the number of methods for action on oxides is limited to classic reactions discov ered more than a century ago; their possibilities have been studied thoroughly, they can be used to solve cer tain problems but far from all. Currently, there are a large number of reducing agents whose properties have been studied in detail. However, most of them are neither universal nor mild. Reactions with most of the wellknown reducing agents require high pres sures and temperatures and are dangerous. For this reason, the search for new approaches to the reduc tion of oxides and changes in oxygen stoichiometry, especially that of complex oxides, is a problem of current interest. In recent years, the reduction of oxides has been attracting attention as a method for preparing metal nanoparticles, which are basic to (raw materials for) certain nanotechnology processes. It is therefore necessary to seek new nontraditional reducing agents free of the shortcomings specified. One of these is supercritical isopropanol (SCI). The unique properties of supercritical fluids (pene tration, solvation, and extraction properties etc.) are well known [1]. We suggested that these properties should cause changes in the reactivity of fluids, not only quantitative, such as, for instance, substantial reaction acceleration, but also qualitative, that is, the appearance of new reactions unknown for fluids in the subcritical state. However, only a small number of

1172

SUPERCRITICAL ISOPROPANOL AS A REDUCING AGENT

such reactions are known for traditional supercritical fluids (H2O, CO2, etc.), and they have not been stud ied systematically. The unique properties of supercritical isopropanol, its ability to hydrogenate double bonds in organic compounds and perform hydrogenolysis at single bonds such as C–N, C–O, C–P, and C–S, were dis covered by us more than 20 years ago [2]. We found that supercritical alcohols exhibited strongly different reactivities with respect to the same substrates; for instance, their hydrogenation ability decreased in the series (CH3)2CHOH > C2H5OH > CH3OH [2–4]. The purpose of this work was to study the interac tion of supercritical isopropanol with inorganic oxides, both simple and complex, including metal oxide nanoparticles embedded in polymer matrices and formed on the surface of synthetic opal micro granules. The most important problem was the devel opment of a procedure for work with supercritical iso propanol under usual laboratory conditions. EXPERIMENTAL Initial Oxide Samples We used powdered simple oxides of kh. ch. (chem ically pure), os. ch. (special purity), and ch. d. a. (pure for analysis) grades. Their phase compositions were substantiated by Xray diffraction. Complex oxides were prepared and characterized at the Laboratory of Physicochemical Analysis of Oxides, Institute of Gen eral and Inorganic Chemistry, Russian Academy of Sciences. Undoped Bi12GeO20 (BGO), Bi12SiO20 (BSO), and Bi12TiO20 (BTO) single crystals and BTO single crystals doped with manganese, chromium, nickel, vanadium, and copper with sheelite structures were grown using the Chukhral’skii method. Doping was performed by adding oxides of the corresponding metals to the initial batch. The concentrations of dopants in crystals were determined by laser mass spectrometry. They were of 0.07–0.4 wt %. The con centrations of uncontrolled impurities of other d met als did not exceed 1 × 10–4 wt %. Samples were pre pared in the form of 5 × 10 mm polished plates 0.1– 0.8 mm thick. Each plate was cut in two. One half was subjected to the action of SCI under supercritical con ditions, and the other was used as a reference. Single crystal plates with sillenite and sheelite structures were prepared at the Laboratory of Physicochemical Anal ysis of Oxides, Institute of General and Inorganic Chemistry, Russian Academy of Sciences. We used isopropanol of kh. ch. and os. ch. grades. Prior to use, isopropanol of kh. ch. grade was held over calcined molecular sieves (Tseozob 4A) for 24 h. The absence of traces of water in isopropanol of both kh. ch. and os. ch. grades was substantiated chromato graphically. RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY B

1173

The preparation of samples containing metal oxide nanoparticles in a polyethylene matrix and on the sur face of SiO2 microgranules is described below. Methods for the Preparation of Oxide Nanoparticles Stabilized in Polymer Matrices Metal oxide nanoparticles in polymer matrices (precursors for reduction with SCI) were prepared by two methods. The first method was used to synthesize metal oxide nanoparticles stabilized in a polyethylene matrix. The preparation of polyethylene + metal nanocomposites included two stages, (1) the prepara tion of oxide precursor (chloride and oxochloride) nanoparticles stabilized in polyethylene and (2) the reduction of nanoparticles in polyethylene with super critical isopropanol. At the first stage, oxide nanopar ticles stabilized in a polyethylene matrix were synthe sized following the procedure for the preparation of nanomaterials developed and extensively used at the Laboratory of the Chemistry of Nanomaterials, Insti tute of General and Inorganic Chemistry, Russian Academy of Sciences [5–7]. This was the thermal decomposition of metalcontaining compounds in a polyethylene + oil solutionmelt (280–300°C). The second method for the preparation of precur sors was used to obtain nanocomposites in an opal matrix. The synthesis of opal matrix + metal nano composites included four main stages: the synthesis of opal matrices by the Stober method [8], filling of opal matrix pores with concentrated solutions of salts of various metals, thermal treatment of opal matrices impregnated with solutions of metal salts, and the reduction of thermal decomposition products encap sulated in the opal matrix under the action of SCI. Solid samples of metal nanoparticles stabilized in polymers were dried in air and studied by the methods described below. Procedures for Work with SCI The interaction of oxides with isopropanol was studied in the regions of sub and supercritical alcohol states. The transition to the supercritical sate was per formed by increasing reactor temperature and, simul taneously, fluid pressure. Experiments at elevated pressures were performed using two autoclave proce dure variants, with sealed ampules and open quartz containers. In both variants, standard experiments were performed at temperatures substantially (by 50– 70°C or more) higher than the critical isopropanol temperature, outside the region of fluctuations usually observed close to the critical point. It is very important that the ratio between the alcohol, metal oxide, and internal ampule volume after sealing off (or free auto clave volume in the second variant with open quartz test tubes) provide the mean isopropanol density close to 0.27 g/cm3 (this is the critical density of isopro panol). Liquid alcohol and its vapor then remain in Vol. 3

No. 8

2009

1174

GUBIN, BUSLAEVA 10 9 8 7

8

6 7

5

6 4

5 4

3

3

2

2

1

1 Fig. 1. Autoclave for experiments with sealed ampules: 1, autoclave; 2, cartridge for ampules; 3, sealed Pyrex ampule; 4, metal oxide and isopropanol; 5, “supporting” isopropanol; 6, pocket for internal thermocouple; 7, flange lock; 8, valve; 9, pressure gauge; and 10, thermocouple.

equilibrium during ampule heating to the alcohol crit ical parameters (235.3°C and 5.3 MPa) [9, 10], and, as the temperature grows above the critical point, pres sure then increases approximately in correspondence with the critical isochore pressure. Substantial devia tions from this condition cause either uncontrolled pressure growth already at moderate temperatures or the formation of a lowdensity fluid with the proper ties of a gas, that is, experiment failure. The method of sealed ampules. In the variant with sealed ampules, we used 12–15 cm3 Pyrex ampules 14–15 mm in diameter with walls 0.7–1.0 mm thick. A metal oxide powder (0.5 g) or a crystalline plate (0.01–0.20 g) and isobutanol were sealed in an ampule [9, 10]. Six ampules were placed into massive metallic con tainer 2 (Fig. 1), which decreased temperature gradi ent. Isopropanol was introduced into the autoclave to make pressure in the autoclave at a working tempera ture close to pressure in ampules. At any temperature, pressures inside and outside ampules were close to each other, and ampules remained intact. The con tainer was sealed and placed in autoclave 1 with vol ume 600 ml; it had pocket 6 for ChromelAlumel ther mocouple 10. The construction ensured maximum closeness between thermocouple readings and the real reaction mixture temperature. Pressure was controlled using pressure gauge 9. The autoclave was hermetically sealedand placed into a resistance furnace. The tem perature was increased at a rate of 60–70 K/h. If pres sure in the autoclave exceeded the required value,

Fig. 2. Autoclave for experiments with open containers: 1, autoclave; 2, open ampule; 3, metal oxide; 4, isopro panol; 5, lock; 6, gasket; 7, bearing box; and 8, nut.

excess alcohol was removed through valve 8. After the required temperature was reached (under both sub and supercritical conditions), the ampules were held for 1–6 h. The autoclave was then cooled, and the container with ampules was removed from it. Ampules were opened after cooling them in liquid nitrogen. The liquid phase was separated for analyses, and the solid phase was placed into a desiccator and evacuated for 2–3 h to remove the liquid phase. If necessary, experiments in sealed ampules allowed us to retain all (including gaseous) reaction products and exclude possible effects (contaminating and cata lytic) of autoclave walls on the purity of products. The method under consideration has two short comings: (1) ampule loading and unloading are fairly labori ous tasks; (2) there is a fairly high possibility of explosion when ampules are opened because of the formation of gaseous reaction products (when the reaction deviates from stoichiometry). The method of open containers. To remove the diffi culties specified above, we suggested the second auto clave procedure variant (Fig. 2). Smallvolume auto claves made of EP943 nickel alloy [10] were used. A quartz test tube with a powder (0.5 g) or oxide plate (0.01–0.20 g) was placed into an autoclave. The amount of alcohol added was calculated from the same considerations as with the first variant. Six to eight autoclaves were simultaneously placed into an air thermostat and held at the required temperature as long as necessary (16–18 h at 285–290°C). The spe cial features of the construction of the valve of the autoclaves that we used excluded the possibility of explosion during experiments (they withstood pres

RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY B

Vol. 3

No. 8

2009

SUPERCRITICAL ISOPROPANOL AS A REDUCING AGENT

sures no less than five thousand atm) and allowed autoclaves to be safely and easily opened after experi ments and gaseous products, if formed, to be col lected. When the first variant is used, loads are in contact with quartz only. In the second variant, alcohol is in contact with autoclave walls, which raises doubts in the equality of conditions and comparability of the results obtained. To dispel these doubts, we performed special experiments, socalled idle experiments, in which isopropanol was only placed into an autoclave. Subsequent analyses showed the absence of differ ences between alcohol compositions before and after experiments. Temperature was maintained constant in both autoclave procedure variants to within 1–2°C. Errors in temperature measurements were estimated at ±10– 15°C for the first variant and ±5–7°C for the second one. This circumstance forced us to select tempera tures either noticeably lower or noticeably higher than the critical alcohol temperature to be sure of the phase state of the system. Visual procedure. Apart from the two methods described above, we sometimes used the procedure developed at Institute of General and Inorganic Chemistry, Russian Academy of Sciences, that allowed us to visually observe processes in a sealed ampule as the temperature increased. We could then fix meniscus disappearance, which was an indication of the transition of the substance into the supercritical state, and watch over the state of the solid phase [11]. Studies of the samples were performed following standard procedures. After cooling and opening of ampules (if the ampule method was used) or opening of an autoclave (if open containers were used), the solid phase was separated by filtration, dried in air, and studied by Xray diffraction, thermogravimetry, and spectral methods. Nanoparticles were additionally examined by transmission electron microscopy. Interaction of single crystals with supercritical flu ids causes the appearance of structure defects (disloca tions, block boundaries, etc.). For this reason, all the samples were polished anew before spectral measure ments. The absorption spectra were recorded on Specord M40 and Hitachi330 spectrophotometers, and the circular dichroism spectra, on a Mark3S dichrograph (Jobin Yvon). The size of nanoparticles was determined using transmission electron microscopy on a JEM100B (JEOL) instrument. For this purpose, samples were subjected to ultrasonic dispersion in an aqueousalco holic solution, and a drop of the dispersion was placed on a copper grid with carbon coating. The liquid phase was analyzed by chromatography and chromatomass spectrometry. We used a Varian 3700 chromatograph (REOPLEX column, vaporizer RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY B

1175

and detector temperatures 100°C, column tempera ture 30–40°C) and an Automass150 chromatomass spectrometer (OV1 column with a 0.25 mm inside diameter, column length 25 m, flow splitting 1 : 40, sample volume 0.2 μl, vaporizer temperature 200°C, electron impact ionization, ionizing electron energy 70 eV). RESULTS AND DISCUSSION The Interaction of Bismuth Oxide with Isopropanol in the Supercritical and NearCritical Regions First, we studied the interaction of αBi2O3 with isopropanol in the supercritical and nearcritical regions of the parameters of state. αBi2O3 was selected as a model compound for studying the main characteristics and distinguishing features of the reac tion. Bismuth oxide is a typical nontransition metal oxide; the structure of αBi2O3 and phase equilibria in the Bi–Bi2O3 system have been studied thoroughly [12, 13]. It is known that the reduction of αBi2O3 to the metal in a classic flow system in a flow of H2 occurs at 500–800°C. We found that, in a sealed ampule with isopro panol, αBi2O3 transformed into bismuth metal at 250–270°C during 20–30 min. Studies showed that the interaction of αBi2O3 at temperatures higher than critical proceeded as αBi2O3(s) + 3iC3H7OH(f) 2Bi(l) + 3(CH3)2C=O(f) + 3H2O(g). The symbol (f) denotes that the substance is in the supercritical fluid state. Bismuth metal is formed as spherical particles with various sizes (1 to 0.001 mm in diameter) or a molten regulus. The samples were finely disperse dark (almost black) powders; they were easy to grind in a mortar. According to the Xray diffraction data, the prod uct was bismuth metal. The form of bismuth metal after the reaction was likely related to the special fea tures of melt crystallization when the system was cooled [9]. An analysis of the liquid phase showed that, along with initial isopropanol, it contained acetone in an amount corresponding to the stoichiometry of the reaction and water. An experiment in a special unit that allowed the development of the reaction to be observed visually showed that, when the isopropanol critical tempera ture was reached, meniscus disappeared. There was no stratification in the system during the whole reaction, and the solid phase remained at the bottom of the ampule during the experiment [10]. It follows that the observed reaction was the solid state bismuth oxide reduction to bismuth metal under the action of SCI. We also studied the interaction of αBi2O3 with iso propanol in the subcritical region. In the first series of experiments, Bi2O3 was boiled in isopropanol at atmospheric pressure in argon for Vol. 3

No. 8

2009

1176

GUBIN, BUSLAEVA

15–18 h. We did not observe changes in the composi tion of the solid and liquid phases. The solid phase contained initial αBi2O3 only, and the liquid phase, pure isopropanol. Indeed, according to thermodynamic estimates, the reaction αBi2O3(s) + 3iC3H7OH(l) 2Bi(s) + 3(CH3)2C=O(l) + 3H2O(g) should not occur at temperatures below 100°C (the Gibbs energy of the reaction is +33.8 kJ/mol). In the second series of experiments, thermal treat ment in sealed ampules at 160–220°C (below the crit ical temperature) was used. We found that, although the interaction of iso C3H7OH with bismuth oxide at these temperatures caused changes in sample color (sample darkening), the presence of Bi metal or other phases could not be detected by Xray diffraction. Very small amounts of bismuth metal (traces) were observed in these samples by differential thermal analysis (DTA) and thermo gravimetry (TG) [9]. It can be assumed that the interaction of isopro panol with bismuth oxide begins before the critical temperature is reached but proceeds at a very low rate. An increase in reaction duration should then result in more complete reduction. Indeed, when bismuth oxide is held in isoC3H7OH at 200°C for a long time (more than 90 h), the metallic phase is gradually accu mulated in the sample. At the same time, the time during which 100% α Bi2O3 react with isopropanol at 250°C is only 30 min. It follows that this reaction, even if it occurs in the sub critical region, is sharply accelerated when the reagent is in the supercritical state. It was shown for the first time that a heterogeneous chemical reaction between solid metal oxide and a fluid occurred at a high rate at comparatively low tem peratures with metal formation. Similar results were obtained for such a strong oxi dizer as MnO2. Boiling of isopropanol with MnO2 during 15 h did not result in the formation of acetone in the liquid phase, which is obvious evidence of the nonoccurrence of reduction. The interaction of Mn oxides with isopropanol was studied at both subcritical and supercritical state parameters of the alcohol. As for bismuth oxide, three series of experiments were performed. In these exper iments, we used MnO2 of ch. d. a. grade (according to the Xray data, it was a mixture of MnO1.88/index card 5673/ and MnO2/index card 14644/) and Mn2O3 (bixbyite/index card 31825/). As with bismuth oxide, various MnO2 samples were boiled with isopropanol at atmospheric pressure in a flow of argon for 15–18 h in the first series of experi ments. No changes in the composition of the solid and liquid phases were observed. It follows that no redox processes occur in the MnO2–isoC3H7OH system at 83°C and 1 atm.

In the second series of experiments, the compo nents were heated in sealed ampules under subcritical conditions. We could observe the disappearance of the interphase boundary when the critical parameters were reached. The oxide–alcohol ratio and the vol ume of the ampule were selected such that process parameters did not reach the critical values for isopro panol at 170–220°C. In the third series of experiments, the components in sealed ampules were heated under supercritical conditions. The oxide–alcohol ratio and the volume of the ampule were then selected such that pressure and density exceeded their critical values when the temperature rose above 235°C. We observed the reduction of MnO2 to lower oxides discussed below in more detail. It was for the first time found that, after the transition to the supercritical state, isopropanol became an active reducing agent for inorganic oxides. The Interaction of SCI with Simple Oxides of Various Elements We studied the interaction of SCI with simple metal oxides. The oxides were divided in several groups with respect to SCI. The first group included oxides that largely underwent reduction to metals [14], MxOy(s) + iC3H7OH(f) M(s) + (CH3)2C=O(f) + H2O(g). Here, MxOy stands for CuO, CdO, HgO, PbxOy, Sb2O3, Bi2O3, CoxOy, PtO2, Ag2O, TeO2, and RexOy. Analyses of reaction products sometimes showed the presence of unreacted oxides and, likely, nonsto ichiometric variablecomposition oxides. The liquid phase contained acetone and small amounts (less than 1%) of the products of catalytic transformations of the alcohol and acetone under reaction conditions. This series contains Group I–VIII metal oxides, as a rule, from the third (more rarely, second) long period. It is known that these oxides are reduced in a flow of hydrogen, but under much harsher conditions (at 300–800°C and higher temperatures). Clearly, the developed method has certain advantages over classic reduction with hydrogen in a flow system. In addition, there are oxides that do not interact with SCI. These oxides, when brought in contact with SCI for a long time (16 h), remained unchanged. Such behavior was characteristic of BaO, ZnO, SnO2, GeO2, SiO2, ZrO2, HfO2, CeO2, Ga2O3, In2O3, Cr2O3, and NiO. The liquid phase after reactions with these oxides did not contain even traces of acetone. This was evidence that no interaction between these oxides and SCI occurred under experimental conditions. We studied reactions of SCI with metal oxides that underwent reduction to intermediate degrees of oxida tion. There were several metal oxides (MoO3, V2O5, MnO2, Mn2O3, and Fe2O3) that intensely interacted with SCI according to the composition of the liquid

RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY B

Vol. 3

No. 8

2009

SUPERCRITICAL ISOPROPANOL AS A REDUCING AGENT

phase, which contained substantial amounts of ace tone. According to Xray diffraction, reduction to lower oxidation states occurred in these reactions, MxOy(s) + iC3H7OH(f) MOy – z(s) + (CH3)2C=O(f) + H2O(g). Here, MoO3 MoO2, V2O5 VO2, MnO2 Mn2O3 Mn3O4, and Fe2O3 Fe3O4. Note that an increase in reaction duration even to 90 h did not cause the reduction of these oxides to lower oxidation states. Among these oxides, the inter action of manganese oxides with SCI was studied most thoroughly. As distinct from bismuth, manganese(III, IV) oxide powders are reduced to states with a lower manganese oxidation degree (Mn(II)), but not com pletely. In both cases, the reaction product is Mn3O4 mixed oxide (hausmannite/index card 24734/). We found that, when the process was performed at a tem perature higher than critical for 1 h, the reaction MnO2 + iC3H7OH(f) Mn3O4 + (CH3)2C=O(f) + H2O(g) occurred. Mn2O3 reacted similarly, Mn2O3(s) + isoC3H7OH(f) (CH3)2C=O(f) + Mn3O4(s) + H2O(g). No reduction of manganese oxides to MnO occurred. An increase in the duration of experiments to 6 h at 300°C resulted in the reduction of initial manganese oxides to Mn2+, but we were unable to fix the forma tion of the corresponding oxide, MnO. It was found in these experiments that the final reaction product was Mn(OH)2 (pyrochroite), βMnO2(s) + isoC3H7OH(f) (CH3)2C=O(f) + Mn(OH)2(s). It follows that SCI reduces manganese(3+, 4+) oxides to manganese compounds with the degree of oxidation 2+, either Mn3O4 mixed oxide or Mn(OH)2 hydroxide. The Formation of Hydroxides in Reactions of Simple Oxides with SCI The formation of M(OH)2 described above was not fortuitous. It was shown by Xray diffraction that the solid product of the interaction of Mn oxides with SCI consisted of pyrochroite Mn(OH)2 by more than 80% and βkurnakite (Mn2O3). A similar product (Nd(OH)3 hydroxide) was formed with a 100% yield in the interaction of Nd2O3 with SCI. This prompted us to perform a systematic study of the interaction of SCI with other rareearth metal oxides. The results substantiated the possibility of formation on Ln(OH)3 hydroxides (Ln = La, Pr, Nd, Sm, Eu, Gd, Tb, Dy, and Yb) in reactions between SCI and Ln2O3 [15]. Among the rareearth metal oxides studied, only cerium oxide CeO2 did not interact noticeably with SCI. RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY B

1177

Note that, in spite of almost identical experimental conditions, the intensity of interaction depended on the nature of the rareearth metal. Under the selected conditions (285°C, 16 h), complete transformation with 100% yields was only observed for lanthanum and neodymium; in other instances, part of initial oxides (20–40%) remained unchanged. The DTA and TG curves of the products contained thermal and weight loss effects that fully corresponded to those observed when La(OH)3 and Nd(OH)3 were decomposed to the corresponding oxides. The process under consideration can be of impor tance for preparative purposes, it can be used as a con venient method for synthesizing metal hydroxides from oxides. It follows that, according to reactivity toward SCI under supercritical conditions, Group I–VIII metal oxides can be divided into four groups: (1) oxides that undergo complete reduction to met als (CuO, CdO, PbxOy, TeO2, Sb2O3, Bi2O3, CoxOy, RexOy, Ag2O, HgO, and PtO2); (2) variablevalence metal oxides that undergo reduction to intermediate (lower) oxides (V2O5, Fe2O3, MnO2, Mn2O3, and MoO3); (3) Ln2O3 oxides (Ln = La, Pr, Nd, Sm, Eu, Gd, Tb, Dy, and Yb), which interact with SCI to produce metal hydroxides; (4) oxides that do not interact with SCI (Ga2O3, SnO2, ZnO, In2O3, Cr2O3, WO3, ZrO2, NiO, CeO2, SiO2, and GeO2). It was shown for the examples of SCI reactions with bismuth and manganese oxides that the process was heterogeneous in character during the whole its devel opment. The experiments performed revealed the diversity of directions of metal oxide interactions with SCI. We found original directions that required additional studies. Their deeper investigation could help us to understand the role played by the supercritical reagent state in determining the direction of interaction in heterogeneous reactions. Reactions of SCI with Complex Metal Oxides We studied the interaction of SCI with complex metal oxides Mx M 'y Oz. Two different oxides combined in one sample are materials wide spread in nature and industries. Fine changes in their oxygen stoichiometry, selective reduction of one of the components, con trolled changes in morphology, etc. are the problems to which efforts of many researchers are applied. It was important to estimate the possibilities of SCI for solv ing some of these problems. Our goal was to specify the range of reactions that occurred in the SCI–complex oxide systems. Studies of the interaction of SCI with complex oxides revealed certain special features of these reactions. Vol. 3

No. 8

2009

1178

GUBIN, BUSLAEVA

Table 1. Composition of oxide compound samples before and after interaction with SCI at 280–285°C Initial sample composition

Solid phase composition after the reaction, Xray data [ASTM]

Notes

Bi2CdO4

〈Bi, Cd〉 Bi [5–519] + Cd [5–647]

Eutectic alloy 〈Bi, Cd〉

30% Bi2O3 + 70% TeO2

Bi2Te3 [8–27] + Te [4–554]

Intermetallic compound formation

Bi16CrO27

Bi [5–519] + Cr3O4 [12–559] + Cr5O12 [18–390]

Bi2WO6

Bi [5–519] + WO3 [5–388] + WO2.90 [36–0102] + WO2 [5–431]

Bi4V2O11, BiVO4

Bi [5–519] + V3O7 [27–940] + V6O13 [27–1318]

Bi2Mo3O12

Bi [5–519] + MoO2 [32–671]

Bi2O3 · Co2O3

Bi [5–519] + Co [5–727] + CoO [9–402]

Bi12SiO20

Bi [5–519] + Bi12SiO20 – z

Bi12GeO20

Bi [5–519] + Bi12GeO20 – z

Bi12TiO20

Bi [5–519] + Bi12TiO20 – z

Bi4Ge3O12

Bi [5–519] + Bi4Ge3O12 – z

Bi2Mo3O12

Bi [5–519] + Bi2Mo3O12 – z

PbMoO4

Pb [4–686] + PbMoO4 – z

NaBi(WO4)2

Bi [5–519] + NaBiW2O8 – z

First, we studied the reactions of oxides with both metal oxide constituents belonging to the first group of simple metal oxides; that is, both ions were reduced by SCI to metals (Table 1) [16]. It was shown that, depending on the nature of the metals, products of two types could form in the reduc tion of binary oxides to metals, (1) eutectic alloys of two metallic components, e.g., Bi2CdO4 〈Bi,Cd〉, or (2) intermetallic compounds, [30%Bi2O3 + 70%TeO2] Bi2Te3. Next, we studied variants of the reduction of com plex oxides one of which belonged to oxides of the first group (that is, underwent complete reduction) and the other, to oxides of the second or third group or did not interact with SCI at all. The following variants of pro cesses were observed: (1) selective reduction with the formation of metal interlayers within a complex oxide with a layered structure, Bi2Mo3O12 Bimet (interlayers between Bi2Mo3O12 – σ layers), (2) selective reduction of one of the complex oxide components with the formation of highdispersity metal within the structure of the irreducible second oxide component (Table 1), Bi2WO6 Bi + WO3, Bi16CrO27 Bi + Cr2O3,

Metal and second component oxide formation

Change in oxygen stoichiometry (“extraction” of oxygen)

Bi4Ge3O12

Bi + GeO2,

Sb4Ge3O12

Sb + GeO2,

Bi12MO20

Bi + MO2,

where M = Si, Ge, Ti, V, etc. Note that reactions of this type provide the possi bility of obtaining metal (catalyst) particles on the sur face of a highdispersity porous irreducible oxidecar rier (SiO2, TiO2, etc.). The examples considered above demonstrate actu ally unlimited possibilities of SCI to perform the con trolled modification of oxide materials with complex compositions. Reactions proceed at a high rate, almost quantitatively, and under mild conditions. Supercritical isopropanol, which is an effective reducing agent, has certain advantages inherent in supercritical fluids. It is characterized by high penetra bility, easily overcomes hydrophobichydrophilic bar riers in sorption on surfaces, is an effective extractant, also from inside regions of solids, etc. These attractive features of SCI were used to broaden the range of reactions which take advantage of the unique possibilities of this method in the chem istry of inorganic oxides, such as (1) the reduction of oxides existing in the form of nanoparticles inside an organic polymer matrix or intercavity space of SiO2 microgranules, and (2) the reduction of powdered oxides considered above in the form of plates cut from the corresponding single crystals.

RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY B

Vol. 3

No. 8

2009

SUPERCRITICAL ISOPROPANOL AS A REDUCING AGENT

The Interaction of SCI with Metal Oxide Nanoparticles in a Polyethylene Matrix Because of the rapid development of a new direc tion, the chemistry of nanomaterials, there is a neces sity of obtaining nanoparticles with required composi tions; in practice, this is not always possible because of the high chemical activity of nanoparticles. Con trolled action of reagents on stabilized nanoparticles opens up the possibility of creating nanomaterials with required compositions and, therefore, properties. It is known that the supramolecular structure of a semicrystalline polymer (for instance, highpressure polyethylene) contains many micro and nanosized cavities and channels, which should not prevent reagents from penetrating into the regions of nanopar ticle localization and guarantees successful studies of the chemical properties of nanoparticles localized in the volume of a polymer. The permeability of polyeth ylene films to gases of various compositions has been studied fairly well. The probability of the penetration of gaseous chemical agents into cavities containing nanoparticles is fairly high. The situation with solvents or solutions of reagents is different. The hydrophobic character of polyethylene and the presence of residual oil (because of the special characteristics of the syn thesis of nanoparticles) in cavities hinders the penetra tion of reagents dissolved in water to nanoparticles. In this situation, a promising method for performing chemical transformations of certain classes of com pounds into compounds of other classes is the action of substances in the supercritical state on composition nanomaterials. It was suggested that the unique reduc ing properties of SCI could be used to reduce metal oxide nanoparticles considered above which are embedded in a polyethylene matrix. It is difficult to suggest another known reagent for performing such transformations. Neither hydrogen nor complex metal hydrides can be used for this pur pose. The possibilities of the universal method for the preparation and stabilization of metalcontaining nanoparticles developed at the Laboratory of the Chemistry of Nanomaterials, Institute of General and Inorganic Chemistry, Russian Academy of Sciences, are fairly great [17, 18]. We can change the composi tion of nanoparticles introduced (metals, oxides, chlorides) and their concentration over wide ranges by varying the initial compounds and matrix stabilizers [19]. To prevent the agglomeration of nanoparticles with the loss of properties characteristic of the nanosized state, various stabilization methods are used. Polymer matrices are most often employed for this purpose [20]. The physical properties of nanoparticles local ized in polymers, especially their magnetic character istics [21–23], have been studied fairly well. At the same time, the chemical reactivity of nanoparticles in such materials remains almost unstudied, although RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY B

1179

Table 2. Mean size of metalcontaining nanoparticles before and after interaction with SCI

Precursor

Compo sition of initial particles

Mean size of initial nanopar ticles, nm

Mean size of nanopar ticles after interaction with SCI

Bi(CH3COO)3 Bi(OH)3 BiCl3 · H2O Pb(NO3)2 Hg(CH3COO)2

Bi2O3 Bi2O3 BiOCl PbO HgO

7.0 ± 1.5 5±1 5±2 13 ± 2 22 ± 3

7.5 ± 2.0 6.5 ± 1.5 6.0 ± 1.5 14 ± 3 23 ± 3

their high chemical activity is always mentioned. Even the oxidation of metal nanoparticles was only studied occasionally [24–27] and only from the point of view of changes in the magnetic characteristics of nanoma terials. In this technology, many things depend on the accessibility of the corresponding precursors. For instance, there are numerous carbonyls and organo metallic derivatives for the preparation of Fe, Co, and Ni nanoparticles [28, 29]. At the same time, the use of organometallic compounds for the preparation of heavy nontransition metal nanoparticles is problem atic because of their high toxicity. In addition, heavy nontransition metal nanoparti cles (Bi, Pb, and Hg) have been studied to a much lesser extent than nanoparticles of other metals. Note that taskoriented studies of the reactivity of nanoparticles encapsulated in various materials and its comparison with the reactivity of the corresponding compact materials are almost absent. There are only occasional observations. At the same time, precursors necessary for the pro duction of the corresponding oxide nanoparticles have well been studied and are accessible for most of the metals. It follows that the development of a method for the reduction of oxide nanoparticles inside polymer matrices opens up the possibility of synthesizing a wide range of metal nanoparticles. First, we studied the reactions of SCI with bis muthcontaining nanoparticles (bismuth oxide, chlo ride, and oxychloride) [30] and lead and mercury oxide nanoparticles in polyethylene [31]. They were synthesized following the standard procedure of the thermal decomposition of the corresponding salts in a solutionmelt of highpressure polyethylene. The nanoparticles obtained were used as precursors for reduction under the action of SCI (Table 2). Bismuth containing nanoparticles were reduced by SCI follow ing the procedure developed earlier for powdered materials in open containers. It was shown that SCI could penetrate into a polyethylene matrix without its destruction and reduce nanoparticles localized in it Vol. 3

No. 8

2009

1180

GUBIN, BUSLAEVA (a) N 50 40 30 20 10

Bi2O3

2

50 nm

3 4 5 6 7 d, nm d = 4.8 ± 1.0 nm

(b) N 50 40 30 20 10 Bi

50 nm 2 3 4 5 6 7 d, nm d = 3.85 ± 0.9 nm

Fig. 3. Transmission electron microscopy data and particle size distributions of bismuthcontaining nanoparticles in polyethylene (a) before and (b) after treatment with SCI.

(a)

50 nm

50 nm

(b)

Fig. 4. Lead oxide nanoparticles (a) before treatment with SCI and (b) after treatment with SCI.

according to the reactions characteristic of compact powders of the same composition (Fig. 3). Similar results were obtained in the reduction of Pb and Hg oxide nanoparticles (Figs. 4, 5). We found that the chemical modification of the composition of nanoparticles inside a polymer matrix by the methods described above did not cause substan tial changes in the mean size of particles (Table 2). Among transition metal oxide nanoparticles, the most impressive results were obtained for rhenium oxides RexOy prepared by the decomposition of ammonium

perrhenate in a polyethylene solutionmelt following the standard procedure. The transmission electron microscopy and Xray diffraction data showed that, after treatment with SCI, the sample contained 3–6 nm rhenium nanoparticles [32]. It follows that metal oxides from the first group (reducible to metals under the action of SCI) present in the form of nanoparticles inside a polyethylene matrix transform into metal nanoparticles after treat ment with SCI. Importantly, treatment with SCI does not affect the polymer matrix and allows metal nano

RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY B

Vol. 3

No. 8

2009

SUPERCRITICAL ISOPROPANOL AS A REDUCING AGENT

(a)

50 nm

1181

50 nm

(b)

Fig. 5. (a) HgO and (b) Hg nanoparticles in polyethylene.

particles inaccessible to the other known synthetic methods to be prepared. We also considered the reduction in polyethylene of transition metal oxide nanoparticles that, in the form of powders, were reduced to oxides with interme diate degrees of oxidation of metal ions rather than to metals. By way of example, we performed the reduc tion of iron oxide Fe2O3 nanoparticles obtained in a polyethylene matrix following the standard procedure. The treatment of the initial composites with SCI caused the formation of Fe3O4 nanoparticles in all experiments, which was substantiated by the Xray and transmission electron microscopy data (Fig. 6) [33]. To summarize, we developed a method for the reduction of iron oxide particles embedded in a poly ethylene matrix under the action of SCI. We showed that reduction occurred similarly to the reduction of compact oxides and caused insignificant changes in the mean size of nanoparticles. The Reduction of Oxides in a Synthetic Opal Matrix We studied the reduction of metal oxide nanoparti cles on the surface of SiO2 microgranules or inserted into opal matrix cavities. In recent years, a trend has been observed toward fixing (immobilizing) small (2– 10 nm) nanoparticles on the surface of spherical microobjects, microgranules with a typical size of 0.2⎯20 μm. Such combined “micronanoobjects” have certain advantages. Nanoparticles fixed on the surface lose the ability to easily undergo compaction but remain accessible to interactions with external reagents and retain their main physical characteristics [34]. At the same time, “homogeneous” dispersions (sols, aerosols, etc.) can be created from microgran ules with nanoparticles on the surface, and various materials (films, coatings, volume samples) can be formed from them according to certain programs. The methods for constructing structures from microgran ules have been developed more thoroughly, micro granules are easier to manipulate than small nanopar ticles. After creating certain structures from such RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY B

coated microgranules by one or another method, one can be sure that the arrangement of nanoparticles in the material obtained this way is equally highly orga nized. Another reason for interest in such particles is related to the possibility of substantial changes in the physical and chemical properties of microgranules caused by their coating with nanoparticles. This can be a source of their new applications in practice. Sponta neous crystallization of SiO2 microgranules results in the formation of socalled opal matrices. Composites based on opal matrices are used for creating photon crystals for various applications. The main aspects of using various materials, including composites based on opal matrices, for creating pho ton crystals were considered in several works [35]. We studied the preparation of nanoparticles of var ious metals either on the surface of SiO2 microgran ules or inserted into opal matrix cavities [36]. Opal matrices were prepared by multistage hydro lytic polycondensation of tetraethoxysilane in an alco hol–ammonia medium followed by the sedimentation of silica globules in the absence of thermal and mechanical actions. The opal structure was hardened by temperature or hydrothermal treatment. Depend ing on synthesis conditions and hardening procedure, the diameter of SiO2 globules in our samples was of 200–300 nm according to the electron microscopy data. The size of intersphere cavities did not exceed 5– 80 nm. The samples had both hexagonal and cubic close packings of spherical silica globules [37] (Fig. 7). The preparation of opal matrix + metal nanocom posites included: (1) the impregnation of an opal matrix with con centrated solutions of saltprecursors; (2) thermal treatment; (3) the reduction of products filling intersphere opal matrix cavities with SCI. The precursors were Bi, Ag, Cu, Au, Zn, Fe, Co, Ni, Mn, Eu, Ru, Sb, and Te salts (as a rule, nitrates). First, we studied the formation of Bi nanoparticles in detail [28]. Opal matrices were impregnated with concentrated (0.5–2.0 M) solutions of Bi(NO3)3 in Vol. 3

No. 8

2009

1182

GUBIN, BUSLAEVA (a) N 50 40 30 20 10 50 nm Photomicrograph of the initial sample of Fe2O3 in polyethylene

4 5 6 7 8 9 d, nm Nanoparticle size distribution (d = 7.67 ± 1.3 nm)

(b) N 50 40 30 20 10 50 nm Photomicrograph of the sample of Fe3O4 in polyethylene after treatment with SCI

4 5 6 7 8 9 d, nm Nanoparticle size distribution (d = 7.67 ± 1.3 nm)

Fig. 6. Transmission electron microscopy data and particle size distributions of iron oxide nanoparticles in polyethylene (a) before and (b) after treatment with SCI.

the presence of polyhydric alcohols (mannitol and glycerol). This was followed by the thermal decompo sition of basic bismuth nitrate to highdispersity bis muth oxide. Thermal treatment was usually performed at 450°C during 2 h. According to the Xray and chemical analysis data, samples after thermal treatment were opal matrix + metal oxide (as a rule, amorphous) composites. The DTA curves of such samples did not contain noticeable effects up to 800°C. Nevertheless, indirect evidence of the presence of amorphous Bi2O3 in opal matrix cavities was the formation of bismuth silicates in samples annealed at 720°C (30 min). According to the Xray data, solidstate interaction between amor phous Bi2O3 and opal matrices caused the formation of a mixture of bismuth silicates Bi4Si3O12 (eulytite) + Bi2SiO5 (metastable layered bismuth silicate) in com plete agreement with the literature data. Silicate nano particles (20–30 nm) were formed on the surface of SiO2 globules [38]. The reduction of Bi2O3 in opal matrices to bismuth metal with SCI was performed at temperatures up to 300°C. The outward appearance of the samples

changed after treatment in SCI. The reduction of opal matrix + Bi(NO3)3 or opal matrix + Bi2O3 (amor phous) samples gave uniformly colored Bi + opal matrix nanocomposites (from dark violet to brown and green). Sample colors were likely determined by the size of SiO2 globules and the concentration of metal nanoparticles. The Xray data substantiated the presence of ele mental bismuth in opal matrices. According to the laser Xrayspectroscopic analysis data, the content of Bi in Bi + opal matrix nanocomposites did not exceed 23 wt %. Figure 8 shows that the diameter of bismuth nanoparticles was 80 nm in intersphere opal matrix cavities and 10–15 nm on the surface of microgran ules. The reduction of other metal salts and oxides in opal matrices to metals was performed using SCI at temperatures up to 300°C. The products were studied by Xray diffraction and transmission electron micros copy [36]. The Xray data listed in Table 3 substantiate the presence of elemental metals in opal matrices after reduction with SCI. The most clear Xray data were obtained for opal matrix + Sb (reduction of SbCl3), opal matrix + Ni

RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY B

Vol. 3

No. 8

2009

SUPERCRITICAL ISOPROPANOL AS A REDUCING AGENT

(a)

1183

(b)

Fig. 7. Close packings of spherical silica globules in opal matrices, (a) hexagonal and (b) cubic.

(reduction of Ni(NO3)2), opal matrix + Ag (reduction of AgNO3), opal matrix + Au (reduction of AuPPh3NO3), opal matrix + Cu (reduction of Cu(NO3)2), opal matrix + Co (reduction of Co(NO3)2), and opal matrix + Te (reduction of H2TeO4) composites. Zinc oxide is not reduced to zinc metal; that is, zinc oxide nanoparticles behave as powdered zinc oxide. We studied the reduction of Group IB metal salts, including Cu(NO3)2, AgNO3, and AuPPh3NO3. Metal salts were preliminarily introduced into space between spherical silica globules by opal matrix impregnation with their solutions. We found that SCI reduced copper, silver, and gold salts in opal matrices to metals. The size of nanoparti cles formed did not exceed the size of opal matrix cav ities; it was of ~40 nm, the diameter of opal matrix SiO2 globules being on the order of 260 nm. We studied the reduction of iron, nickel, and cobalt salts introduced into opal matrices by SCI. The Xray data are listed in Table 3. As with compact oxides, iron(III) compounds were not reduced to iron metal, and only Fe3O4 (magnetite) formed in opal matrix + iron(III) salt composites. At the same time, cobalt compounds were reduced to cobalt metal in opal matrices. The transmission electron microscopy data

on opal matrix + Co composites show that the size of cobalt nanoparticles does not exceed the diameter of cavities (80 nm) both on the surface of silica micro spheres and in intersphere space. Interesting results were obtained when nickel salts were used for impregnating opal matrices. After ther mal treatment, nickel oxide nanoparticles were reduced with SCI to form nickel metal nanoparticles. This was at variance with what we observed for nickel oxide powder, which did not undergo reduction. It is known that nanopowders are energy saturated systems. It can be assumed that this excess energy of nanoparticles compared with compact samples is suf ficient for changing the direction of reaction and acti vates the reduction of NiO nanoparticles to Ni metal. To summarize, SCI can be used to very effectively produce metalcontaining nanoparticles in inter sphere cavities of opal matrices and on the surface of separate SiO2 microgranules. The Special Features of SCI Interaction with Simple and Complex Oxide Single Crystals Interesting results were obtained when SCI inter acted with oxides in the form of compact samples, 1

100 nm

plates cut from single crystals. It is known that oxygen stoichiometry substantially influences the properties of crystals. With bismuth containing sillenites, the usual procedure (annealing in a vacuum) changes crystal stoichiometry not only in the oxygen but simultaneously in the cationic sublat tice. It is common knowledge that supercritical fluids can extract sometimes very complex natural com pounds (drug substances, dyes, etc.) from solid sub strates. This is usually related to the high penetrability of supercritical fluids caused by their low viscosity at a fairly high density. In the absence of direct analogy between extraction and reduction properties, it was interesting to determine the depth of fluid influence 1 We

Fig. 8. Bismuth nanoparticles in opal matrix cavities. RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY B

used sillenite and sheelite samples prepared at the Labora tory of Physicochemistry of Oxides, Institute of General and Inorganic Chemistry, Russian Academy of Sciences.

Vol. 3

No. 8

2009

1184

GUBIN, BUSLAEVA

Table 3. Phase composition of initial samples, samples after thermal treatment (TT), and samples treated with SCI (Xray data) Initial sample and TT conditions OM + Cu(NO3)2 (T = room) OM + CuO (T = 500°C) OM + AuxOy (T = 450°C) OM + AuPPhNO3 (T = room) OM + SbCl3 (T = 170°C) OM + Bi(NO3)3 (T = room) OM + Bi2O3 (T = 500°C) OM + Bi2O3 (T = 700°C) OM + H2TeO4 (T = room) OM + Fe2O3 (T = 450°C) OM + Co(NO3)2 (T = room) OM + Co(NO3)2 (T = 500°C) OM + Ni(NO3)2 (T = room) OM + Ni(NO3)2 (T = 450°C)

After TT Amorphous CuO Amorphous Halo Halo Amorphous Amorphous Bi4Si3O12, Bi2SiO5 Amorphous Amorphous Halo Co3O4 Amorphous Amorphous

on the crystal lattice of oxides. We began with the treatment of an αBi2O3 single crystal with SCI. The Interaction of Bismuth Oxide Single Crystals with SCI The experiment was performed with 5 × 5 × 1 mm plates cut from αBi2O3 single crystals. The plates were treated with SCI under standard conditions in sealed ampules during 1.5 and 2.2 h. Interaction with SCI caused the formation of a thin Bi metal layer on the surface of αBi2O3 single crystal plates. The cross sec tion of an αBi2O3 plate after treatment with SCI is shown schematically in Fig. 9. The plate surface is sur rounded by a loose layer 0.1–0.2 mm thick. Its com position (bismuth metal) was determined by the Xray and DTA methods [10]. The thickness of bismuth lay ers increases as the time of interaction with SCI grows longer. The Interaction of Complex Oxide Single Crystals with SCI We selected the well known type of complex oxides, sillenites, as objects of study. According to the experi mental data, sillenite phases with the composition Bi12MxO20 + δ can be formed with oxides of several metals. Depending on the type of atoms M, they can be stable or metastable compounds or solid solutions based on γBi2O3. The most complete information about phase equilibria in the Bi2O3–MxOy systems (M is a Group I–VIII metal) can be found in [39]. Among a large number of complex oxides, we selected sillenites as convenient model objects, which, on the one hand, contained easily reducible bismuth

After reduction with SCI Halo, Cu Halo, Cu Halo, Au Halo, Au Sb Bi Bi Bi Te Fe3O4, Fe2O3 Co Co Ni, NiO Ni

ions, and, on the other hand, included various cations of other metals, which interacted with bismuth and substantially differed in their ability to interact with SCI. It is also important that all these compounds have the same structure type. We can therefore as a first approximation ignore the possible influence of the structure of complex oxides on the character of their interaction with SCI. For instance, the interaction of a Bi12Ti1 – xMnxO20 single crystal with SCI demon strates the possibility of the selective “extraction” of oxygen from the crystal lattice of the phase with a sil lenite structure (Fig. 10). The changes observed are schematically shown in Fig. 10. The initial Bi12Ti1 – xMnxO20 single crystal has green coloration caused by the presence of Mn5+ and Mn4+ ions [10]. Changes in oxygen stoichiometry (for instance, caused by the annealing of a Bi12Ti1 – xMnxO20 crystal in a vacuum) [40] decrease the degree of man ganese oxidation to +2 (Mn5+/Mn4+ Mn2+). Mn2+ has no absorption bands in the visible range, and crystal color changes from green to yellow (the usual color of undoped Bi12TiO20 crystals). Color change from green to yellow moved deeper and deeper into the crystal as the process continued. It follows that, along with the reduction of Bi ions to Bi metal, mild extraction of oxygen from internal sample regions was possible, and nonstoichiometric oxide formed in the surface layer, 5+

Bi12Ti1 – x Mn x O20

2+

Bi12Ti1 – x Mn x O20 – δ.

Changes in oxygen stoichiometry were also studied for other samples (5 × 10 × 2 mm plates) cut from NaBiW2O8〈Cr〉, PbMoO4, Bi2(MoO4)3, and Bi12MO20 single crystals (samples with M = Si will be denoted by BSO, samples with M = Ge, by BGO, and samples

RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY B

Vol. 3

No. 8

2009

SUPERCRITICAL ISOPROPANOL AS A REDUCING AGENT Bismuth metal layer

1185

Bismuth metal surface layer Yellow layer of Bi12Ti1– xMn2+ x O20 – δ 0.5–1.0 mm thick

αBi2O3 single crystal volume

Unchanged Bi12Ti1 – xMn5(4)+ O 20 x single crystal part

Fig. 9. Cross section of αBi2O3 plate after treatment with SCI.

Fig. 10. Cross section of sillenite plate after treatment with SCI.

with M = Ti, by BTO) and BTO single crystals doped with Cr, Cu, Ni, Co, and V. We found that the action of SCI always caused the formation of thin bismuth or lead (for PbMoO4) metal layers on the surface of single crystal plates. For this reason, the crystals were anew polished before studying their absorption spectra after the action of SCI. Surface metal layers were 0.1– 0.2 mm deep. Under these layers, there were layers with a changed oxygen stoichiometry; they were 0.5– 1.0 mm deep. The depth of transformation depended on the duration of contact between single crystals and SCI and the temperature of interaction [41]. We stud ied the following processes of the reduction of the sur face layer of single crystals under the action of SCI [16]: NaBi(WO4)2 Bi (on the surface)

CONCLUSIONS

+ NaBiW2O8 –δ, NaBi(WO4)2〈Cr5+〉

As a result of our studies, a new reducing agent was introduced into synthetic inorganic chemistry. This is supercritical isopropanol, which has several advan tages compared with the known reducing agents. We determined the possibilities and range of its applica tions for the example of reactions with simple and complex oxides in various forms. ACKNOWLEDGMENTS The authors thank D.A. Astaf’ev†, D.A. Baranov, V.M. Valyashko, A.V. Egorysheva, A.A. Ivanov†, S.N. Ivicheva, Yu.F. Kargin, K.G. Kravchuk, T.B. Kuv shinova, S.N. Sigachev†, V.M. Skorikov, T.A. Strom nova†, and G.Yu. Yurkov for help, participation in joint studies, and consultations.

Bi (on the surface)

+ NaBiW2O8 –δ〈Cr4+〉, PbMoO4

Pb (on the surface) + PbMoO4,

Bi12MO20

Bi (on the surface) + Bi12MO20,

REFERENCES 1. M. D. Luique de Castro, M. Valcarcel, and M. T. Tena, Analytical Supercritical Fluid Extraction (Springer, Ber lin, 1994). 2. S. P. Gubin, V. M. Kirilets, V. N. Men’shov, et al., Izv. Akad. Nauk SSSR, Ser. Khim., No. 12, 2853 (1983). 3. S. P. Gubin, V. M. Kirilets, V. N. Men’shov, et al., Dokl. Akad. Nauk 268, 1129 (1983). 4. S. P. Gubin, Dokl. Akad. Nauk SSSR 345, 490 (1995). 5. S. P. Gubin, Colloids Surf. A 202, 155 (2002). 6. A. V. Kozinkin, V. G. Vlasenko, S. P. Gubin, et al., Neorg. Mater. 32, 422 (1996) [Inorg. Mater. 32, 376 (1996)]. 7. G. Yu. Yurkov, A. V. Kozinkin, T. I. Nedosekina, et al., Neorg. Mater. 37, 1175 (2006) [Inorg. Mater. 37, 1012 (2006)]. 8. W. Stober, A. Fink, and E. Bohn, J. Colloid Interface Sci. 26, 62 (1968). 9. E. Yu. Buslaeva, Yu. F. Kargin, K. G. Kravchuk, S. A. Sigachev, and S. P. Gubin, Zh. Neorg. Khim. 46, 380 (2001) [Russ. J. Inorg. Chem. 46, 323 (2001)].

(M = Si, Ge, Ga, V, Ti, Ti + Mn, Ti + Cr, Cu, Ni, F,e or V). We in parallel recorded the absorption spectra of the samples. It was found that, although the condi tions of treatment with SCI were identical, the degree of changes in the absorption spectra of samples with various compositions varied. The smallest change in absorption was observed for the spectra of BGO having the most perfect lattice. Conversely, the strongest changes were observed in the spectra of BTO, which had the largest number of defects. It follows that the treatment of single crystals of complex oxides with SCI can be used to change oxygen stoichiometry, that is, selectively remove oxygen from the volume of crys tals with the formation of nonstoichiometric oxides of various types. The advantages of using SCI for this purpose is its ability to selectively remove oxygen only and leave the cationic sublattice in the volume of crys tals intact.

†

RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY B

Vol. 3

Deceased. No. 8

2009

1186

GUBIN, BUSLAEVA

10. E. Yu. Buslaeva, K. G. Kravchuk, Yu. F. Kargin, and S. P. Gubin, Neorg. Mater. 38, 706 (2002) [Inorg. Mater. 38, 582 (2002)]. 11. V. M. Valyashko, Phase Equilibriums and Properties of Hydrothermal Systems (Nauka, Moscow, 1990) [in Rus sian]. 12. H. Z. Harwig, Anorg. Allg. Chem. 444, 151 (1978). 13. B. Isecke and J. Z. Osterwald, Phys. Chem. Neuer Folge 115, 17 (1979). 14. Yu. F. Kargin, E. Yu. Buslaeva, K. G. Kravchuk, and S. P. Gubin, Zh. Neorg. Khim. 48, 111 (2003) [Russ. J. Inorg. Chem. 48, 103 (2003)]. 15. S. P. Gubin, Yu. F. Kargin, E. Yu. Buslaeva, and K. G. Kravchuk, Zh. Neorg. Khim. 48, 810 (2003) [Russ. J. Inorg. Chem. 48, 720 (2003)]. 16. Yu. F. Kargin, E. Yu. Buslaeva, A. V. Egorysheva, et al., Zh. Neorg. Khim. 48, 1765 (2003) [Russ. J. Inorg. Chem. 48, 1615 (2003)]. 17. S. P. Gubin and I. D. Kosodubskii, Dokl. Akad. Nauk SSSR 272, 1155 (1983). 18. I. D. Kosobudskii, L. D. Kashkina, G. A. Petrakovskii, V. P. Piskorskii, N. M. Svirskaya, and S. P. Gubin, Vysokomol. Soedin. 27, 689 (1985). 19. S. P. Gubin and I. D. Kosobudskii, Usp. Khim. 52, 1350 (1983). 20. A. D. Pomogailo, A. S. Rozenberg, and I. E. Uflyand, Metallic Nanoparticles in Polymers (Khimiya, Moscow, 2000) [in Russian]. 21. S. Gudoshnikov, B. Liubimov, L. Matveets, et al., J. Magn. Magn. Mater. 258–259, 54 (2003). 22. S. P. Gubin, Yu. I. Spichkin, Yu. A. Koksharov, et al., J. Magn. Magn. Mater. 265, 234 (2003). 23. S. P. Gubin, Yu. A. Koksharov, G. B. Khomutov, and G. Yu. Yurkov, Usp. Khim. 74, 539 (2005). 24. W. A. De Heer, P. Milani, and A. Chatelain, Phys. Rev. Lett. 54, 488 (1990). 25. M. Respaud, J. M. Broto, H. Rakoto, and A. R. Fert, Phys. Rev. B 57, 2925 (1998). 26. F. Bodker, S. Morup, S. W. Charles, and S. Linderoth, J. Magn. Magn. Mater. 196–197, 18 (1999). 27. X. Q. Zhao, Y. Liang, and Z. Q. Hu, J. Appl. Phys. 80, 5857 (1996).

28. G. Yu. Yurkov, S. P. Gubin, and L. A. Paknratov, Neorg. Mater. 38, 186 (2002) [Inorg. Mater. 38, 137 (2002)]. 29. S. P. Gubin and Yu. A. Koksharov, Inorg. Mater. 38, 1085 (2002). 30. G. Yu. Yurkov, D. A. Astaf’ev, M. Yu. Gorkovenko, E. Yu. Buslaeva, Yu. F. Kargin, and S. P. Gubin, Zh. Neorg. Khim. 50, 1402 (2005) [Russ. J. Inorg. Chem. 50, 1300 (2005)]. 31. M. Yu. Gorkovenko, G. Yu. Yurkov, E. Yu. Buslaeva, and S. P. Gubin, Zh. Neorg. Khim. 51 (1), 1 (2006) [Russ. J. Inorg. Chem. 51, 51 (2006)]. 32. E. Yu. Buslaeva, D. A. Fstafiev, Yu. A. Koksharov, and G. Yu. Yurkov, in Proc. of the 8th Intern. Symp. on Supercritical Fluids ISSF2006, Kyoto, Japan, Nov. 5–8, 2006, PB238. 33. E. Yu. Buslaeva, D. A. Astafiev, and G. Yu. Yurkov, in Proc. of the Joint 21st AIRAPT and EHPRG Intern. Conf. on High Pressure Sci. and Technol., Catania, Italy, Sept. 17–21, 2007, p. 245. 34. S. P. Gubin, G. Yu. Yurkov, and N. A. Kataeva, Neorg. Mater. 41, 1159 (2005) [Inorg. Mater. 41, 1017 (2005)]. 35. C. Lopez, Adv. Mater. 15, 1679 (2003). 36. Yu. F. Kargin, S. N. Ivicheva, E. Yu. Buslaeva, V. D. Volodin, and G. Yu. Yurkov, Neorg. Mater. 42, 1065 (2006) [Inorg. Mater. 42, 966 (2006)]. 37. S. N. Ivicheva, Yu. F. Kargin, E. Yu. Buslaeva, and G. Yu. Yurkov, Neorg. Mater. 44, 918 (2008) [Inorg. Mater. 44, 807 (2008)]. 38. Yu. F. Kargin, S. N. Ivicheva, E. Yu. Buslaeva, T. B. Kuvshinova, V. D. Volodin, and G. Yu. Yurkov, Neorg. Mater. 42, 547 (2006) [Inorg. Mater. 42, 487 (2006)]. 39. Yu. F. Kargin, Doctoral Dissertation in Chemistry (IONKh RAN, Moscow, 1998). 40. V. I. Burkov, A. V. Egorysheva, Yu. F. Kargin, and V. V. Volkov, Neorg. Mater. 32, 860 (1996) [Inorg. Mater. 32, 762 (1996)]. 41. A. V. Egorysheva, V. I. Burkov, E. Yu. Buslaeva, Yu. F. Kargin, K. G. Kravchuk, S. P. Gubin, and V. M. Skorikov, Zh. Neorg. Khim. 47, 992 (2002) [Russ. J. Inorg. Chem. 47, 891 (2002)]

RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY B

Vol. 3

No. 8

2009