Carbochlorination of molybdenum trioxide: kinetic treatment

M. W. Ojeda¹, J. B. Rivarola¹ and O. D. Quiroga²

¹Instituto de Investigaciones en Tecnología Química (Universidad Nacional de San Luis-CONICET)
C.C. 290, 5700 San Luis, Argentina
e-mail:mojeda@unsl.edu.ar
²Instituto de Investigaciones para la Industria Química (Universidad Nacional de Salta -CONICET)
Buenos Aires 177, 4400 Salta, Argentina
email: oquiroga@unsa.edu.ar

Abstract ¾ The kinetics of MoO₃ chlorination in the presence of carbon is analyzed. The experimental data were obtained in a fixed-bed isothermal reactor, operated at atmospheric pressure between 543 and 603 K, and at different feed flows and Cl₂ concentrations. The following global reaction was obtained by identification of the reaction products: 2 MoO_3(s) + 2 Cl_2(g) + C_(s) ® 2 MoO₂Cl_2(g) + CO_2(g). On the basis of the analysis of experimental data, thermodynamic results and observations by other authors, three basic stages are distinguished: 1) Formation of the chlorinating agent; 2) Chlorination of molybdenum trioxide; 3) Carbon oxidation. This mechanism allows to explain the greater reactivity of carbochlorination reactions at low temperatures as compared to direct chlorination. A detailed scheme is proposed for the kinetic treatment, which reveals that 22 stages are involved in this complex reaction. The experimental data were processed with different kinetic models, and it was found that the"nucleation and growth"model, conveniently adapted to non catalyzed heterogeneous reactions is the one that better matches the experimental data.

Keywords ¾ Carbochlorination. Molybdenum Trioxide. Kinetic Treatment.

I. INTRODUCTION

   The chlorination of metal oxides in the presence of carbon (carbochlorination) has attracted the interest of numerous researchers, since this reaction allows to recover metals of commercial interest at temperatures lower than those required by direct chlorination. The carbochlorination of metal oxides is a complex fluid-solid reaction:

a Me_mO_n(s) + b Cl_2(g) + c C_(s) ® p P_(g) + q Q_(g) (1)

where a, b, c, p and q are stoichiometric coefficients, Me represents a metal and subscripts (s) and (g) indicate the aggregation state of the reagents and reaction products.
   It has been observed that in these reactions, the presence of carbon has a two-fold effect on the chlorination rate. On the one hand, it thermodynamically favors the formation of the metal oxichlorides, as confirmed by Nagata (1987). On the other hand, numerous authors have reported that carbon exerts a catalytic effect on chlorination reactions (Bernard et al., 1974; Barin and Schuler, 1980; Bicerolu and Gauvin, 1980; Pasquevich and Amorebieta, 1992). The mechanisms that have been proposed for the carbochlorination of metal oxides are not clear or conclusive enough, and important discrepancies exist between the numerous investigations carried out on this issue, even between those studies using the same system. So far, neither detailed nor generalized models for this type of reactions have been described in the literature. It must be noted, however, that most authors agree on the fact that the reaction should occur through the formation of gaseous intermediates, such as phosgene (Stefanyuk and Morozov, 1965; Barin and Schuler, 1980); oxichlorides (Dunn, 1979); and atomic chlorine (Barin and Schuler, 1980; Pasquevich and Amorebieta, 1992). Even though none of these species have been detected during carbochlorination reactions, it can be inferred from the studies of Amorebieta and Colussi (1985) that the most likely intermediate is atomic chlorine. It should also be pointed out that the chlorine reactions on a carbon surface depend on its surface reactivity (Chen and Back, 1979; Papirer et al., 1995; González et al., 1999). In a study of amorphous Ta₂O₅ carbochlorination at low temperatures, González et al. (1998) established the presence of O₂ in the reaction zone. In this work, we will experimentally study the effects of reaction temperature and time, flow and feed composition on the rate of MoO₃ carbochlorination, with the purpose of contributing experimental information and proposing a methodology for the kinetic treatment of complex fluid-solid reactions.

II. EXPERIMENTAL

Experimental equipment. Figure 1 shows a schematic representation of the experimental system employed for the chlorination and carbochlorination of MoO₃ where: DCM, represents the drying, control and measuring device, respectively; EO is the electric oven; S is the solid reagent sample holder; R is the reactor; T is the thermocouple; CT is the collecting tube; DT is the digital thermometer; WT stands for the washing traps; and O the outlet. The system was operated as follows: approximately 1x10^-4 kg of the sample to be chlorinated was placed on the sample holder S and N₂ was passed until the working temperature was reached; at that point the N₂ flow was stopped and the chlorinating agent was circulated at a preset flow, which was interrupted when the programmed reaction time had finished; N₂ was circulated again at room temperature for 1800 s in order to remove the remaining chlorinating agent and cool the reactor.

Figure 1. Flow sheet of the experimental system.

Gaseous reagents: The gaseous reagents used were: (a) Commercial Cl₂ (Indupa, ³ 98%); (b) N₂ (A.G.A., ³ 99.99%), used both as diluent and carrier gas.

Solid reagents: The solid reagents used were: (a) Powder MoO₃ (Analar BDH, ³ 99,99%, surface area 0.4x10³ m²/kg). (b) Amorphous powder carbon (carbon black type, Cabot Corporation, surface area 11.1x10³ m²/kg).
   The reagent solid samples used in the carbochlorination experiments were mixtures of MoO₃ and carbon particles (previously selected by particle size and dried in stove for 120 minutes at 110°C), obtained a 3/1 C/MoO₃ molar ratio and 60 minutes of mixing times. These preparation conditions were studied in a previous work (Ojeda et al., 2002).
   The characterization of starting material and residues by SEM was performed on a Philips model 515 microscopy. This device can analyze the sample fluorescent X rays, using an EDAX energy dispersive detector (EDXS)

III. EXPERIMENTAL RESULTS

   The degree of transformation was defined as: X = (m_o - m) / m_o , where m_o is the initial mass of MoO₃ and m is the mass of unreacted MoO₃.

MoO₃ direct chlorination assays. MoO₃ direct chlorination was studied between 573 and 873 K, at 0.33x10^-6 m³/s feed flow and X measurements at 1800 and 3600 s. The results obtained showed that this reaction only starts at 803 K, and is significant at 873 K.

Carbochlorination assays. Effect of reaction temperature and time: the effect of temperature on the reaction rate was studied in the 523-603 K temperature range and at reaction times varying between 300 and 7200 s, with a feed flow of 0.25x10^-6 m³/s of pure Cl₂. The results obtained indicate that the reaction starts at a temperature close to 543 K, and that conversion increases with higher temperatures and reaction times.

Effect of the feed rate: the effect of the feed rate on the reaction rate was studied with pure Cl₂ at 573 K, within the 0.00-0.67x10^-6 m³/s range. The results obtained show that the X values slightly decrease with increasing flows. This can be explained taking into account that the reaction goes through the formation of gaseous intermediates. This phenomenon has also been reported by other authors (Scheiman et al., 1962; Smith, 1977; Ruiz et al., 1994; Bidaye et al.,1999).

Effect of the feed composition: the effect of the feed composition was studied within the range between 0.5 and 1 of chlorine molar fraction at two reaction temperatures (573 and 583 K) and two flow levels (0.25x10^-6 and 0.33x10^-6 m³/s). The conversion of MoO₃ was observed to slightly increase when the chlorine concentration in the feed is higher.

IV. KINETIC TREATMENT

   The following procedure was used for the kinetic treatment of the MoO₃ carbochlorination data: 1º) formulation of a kinetic mechanism, considering the elementary reaction stages; 2º) schematic presentation and analysis of the stages involved in a complex fluid-solid reaction; 3º) selection and development of kinetic models; 4º) analysis of the experimental results obtained by the selected models; 5º) discrimination of the selected models.

Kinetic mechanism. The following was taken into account for the formulation of a kinetic mechanism: a) the results obtained by the thermodynamic equilibrium calculations performed on a set of probable chlorination and carbochlorination reactions; b) studies by other researchers suggesting that the chlorine atom is the most likely intermediate; c) chemical analyses performed at different levels of the studied factors revealed that MoO₂Cl₂ and CO₂ are the only reaction products. The following kinetic mechanism for MoO₃ carbochlorination is thus proposed:

Activation of Chlorine on the carbon surface:
2 Cl_2(g) + C_(s) ® 4 Cl*_(g) + C_(s) (2)

Chlorination by the active species
2 MoO_3(s) + 4 Cl*_(g) ® 2 MoO₂Cl_2(g) + 2 O*_(g) (3)

Carbon oxidation:
2 O*_(g) + C_(s) ® CO_2(g) (4)

Global reaction:
2 MoO_3(s) + 2 Cl_2(g) + C_(s) ® 2 MoO₂Cl_2(g) + CO_2(g) (5)

Involved stages. On the basis of the proposed reaction mechanism - Eqs.(2) to (4) - the scheme shown in Fig. 2 was designed. It can be seen that 22 stages are involved in the carbochlorination of MoO₃. Some of these are serially combined while others occur in parallel. Stages (1), (7), (15), (16) and (22) correspond to the mass transport by convection of reagents and/or products; stages (2), (6), (8), (13), (14), (17) and (21) correspond to the diffusion of the reagents and/or intermediates and/or products through the film surrounding the solids; stages (3), (5), (9), (11), (12), (18) and (20) correspond to the diffusion of the reagents and/or gaseous intermediates and/or products through the pores, and stages (4), (10) and (19) summarize the stages of formation of the reaction interface, adsorption of the reagents and/or gaseous intermediates, chemical reaction on the surface and desorption of the gaseous intermediates and/or reaction products. It must also be noted that stage (4) corresponds to the activation Cl₂ on the carbon surface to form Cl*_(g) -Eqn. (2)-, which, in order to reach the surface of MoO₃ has to overcome the resistance offered by stages (5) to (9), and, particularly in stage (7) it is likely to be transported by convention by the gas current circulating in the reactor. The chlorination of MoO₃ by the active species, Eqn. (3), is represented by stage (10), in which the reaction products MoO₂Cl_2(g) and O*_(g) are formed. MoO₂Cl_2(g) that has to overcome the resistance offered by stages (12), (14) and (16) in order to leave the reaction zone. The intermediate O*_(g), which, in order to reach the carbon surface, has to overcome the resistance offered by stages (11), (13), (15), (17) and (18). Stage (19) represents the reaction of carbon oxidation, given by Eqn. (4). The scheme presented in Fig. 2 is among the simplest that can be proposed for a carbochlorination reaction. It does not explicitly include reactions of the 2 Cl*_(g) ® Cl_2(g) type, which have a high probability of occurrence. This was due to the facts that 1º) including such reactions would make the analysis of this system more complex without contributing relevant information; and 2º) these reactions can be included when formulating each of the stages in which the intermediate Cl*_(g) is involved. The proposed scheme provides a global view of the complexity of the kinetic treatment of carbochlorination reactions and also allows to prevent the effect of certain factors upon the reaction rate.

Kinetic models. The development of theoretical kinetic models for the scheme presented in Fig. 2, considering different hypothesis, is a practically impossible task. However, if it is taken into account that the observed reaction rate is controlled by the rate of the slowest stage, it becomes possible to correlate the experimental data with the models developed for the kinetic treatment of simple gas-solid reactions. This criterion was adopted for the kinetic treatment of the MoO₃ carbochlorination experimental data, bearing in mind that the reaction mechanisms may vary under the different working conditions.

Figure 2. Steps in metallic oxide carbochlorination

Kinetic treatment results. The experimental data of MoO₃ carbochlorination were treated by 25 different models. Six models based on "nucleation and growth" theory, showed an acceptable match with the experimental data. After performing the corresponding model discrimination, the following was selected as the most probable:

X = b₁ { 1 - (1 + b₂ t) exp (-b₂ t) } (6)

where b₁=s _g W ° N_s° b m_B r_s/ k_N1 and b₂=k_N1; being t: reaction time; s _g: nucleus form coefficient; W °: initial interface surface; N_s°: number of active sites per unit of initial interface; b: stoichiometric coefficient of solid reagent B; m_B: molecular weight of solid reagent B; r_s : reaction rate per surface; k_N1: kinetic coefficient of the rate of active site appearance for the case of order 1. Eqn. (6), according to Quiroga et al. (1994), corresponds to a "nucleation and growth" type model with sequential appearance rate of spontaneous active sites, due to the presence of a chemical species with a nucleus growth factor p=1 (monodimensional growth) and a nucleation order n=1. The correlation results are shown in Fig. 3, and the estimates of the kinetic and statistical parameters are presented in Table 1.
   The activation energy values associated to coefficients b₁ and b₂ are E₁=175,890 J/mol and E₂=159,734 J/mol, respectively. The reaction rate was observed to vary exponentially with the volumetric flow in a magnitude of -0.19.
   In order to corroborate the validity of this model, scanning electron microscopy (SEM) micrographs of MoO₃ particles before and after chlorination are presented and described. Figure 4 shows micrographs of MoO₃ particles mixed with unattacked carbon.
   It must be noted that MoO₃ exhibits perfect cleavage according to í 100ý , matching the properties of molybdite (Fleischer, 1964). Figure 5 corresponds to micrographs of different samples of the reagent solid chlorinated under different conditions. It can be observed that the chemical attack has taken place on the edges of the cleavage planes, correspond ing to the crystallographic planes containing the [b] [c] axes. This effect is more marked under more aggressive chlorination conditions, in which the corrosion of the plane edges and the distance between them has significantly increased, as compared to unchlrorinated or slightly attacked samples.

Table 1. Carbochlorination of MoO₃
Parameter estimation
Model: X = b₁ [ 1 - (1 + b₂ t) exp (-b₂ t)]

   Furthermore, the attack takes place without close contact between the MoO₃ particles and carbon, which matches reports by other authors (Barin and Schuler, 1980; Pasquevich et al., 1992). From these micrographs it can be concluded that the MoO₃ particles undergo an intense chemical attack during chlorination, characterized by specifically proceeding on the edges of the cleavage planes, advancing unidirectionally. This totally matches the kinetic model proposed by this carbochlorination reaction.

Figure 3. Experimental data correlation by Model: X = b₁ [ 1 - (1 + b₂ t) exp (-b₂ t)] , under the operation conditions described in Table 1, for a chlorine molar fraction of 1.

V. CONCLUSIONS

   Direct chlorination showed that the reaction starts at 803 K and is significant at 873 K; while carbochlorination is initiated at 543 K and conversion increases with temperature and reaction time. Increasing feed flows were found to decrease the conversion of MoO₃ and increasing Cl₂ concentrations were found to increase the reaction rate slightly.
   A kinetic mechanism was proposed, given by Eqs. (2) to (4), on the basis of the thermodynamic equilibrium calculation of a set of probable chlorination and carbochlorination reactions, previous studies suggesting that the chlorine atom is the most probable intermediate, and chemical analyses which indicated that MoO₂Cl₂ and CO₂ are the only reaction products.
   A scheme was designed representing the stages of transport, adsorption, reaction and desorption involved in the reaction, matching the kinetic mechanism pro posed. This scheme allowed to see that 22 stages are involved in the reaction, from which it could be inferred that: 1º) the reaction mechanisms may change under different working conditions; 2º) the observed reaction rate is controlled by the rate of the slowest stage; 3º) the experimental data can be treated by models developed for simple gas-solid reactions.
   The experimental data of MoO₃ carbochlorination were treated by 25 different kinetic models. After model discrimination, the most probable one was of the "nucleation and growth" type with sequential appearance rate of spontaneous active sites.
   The validity of the proposed kinetic model is justified by SEM micrographs of MoO₃ particles which show that the chemical attack takes place on preferential sites located on the edges of the crystallographic planes containing the [b] [c] axes, with a unidirectional advance of the reaction front.

 

Figure 4. Micrographs of MoO₃ particles mixed with carbon, non-chlorinated.

Figure 5. Micrographs of MoO₃ particles mixed with carbon, chlorinated at 573 K and with different reaction times (a) 900 s; (b) 1800 s; (c) 2700 s; (d) 3600 s.

Acknowledgements

The authors wish to thank CONICET (Argentina) and Universidad Nacional de San Luis (Argentina), for financial support of this study.

REFERENCES

Amorebieta, V. T. and A. J. Colussi,"Direct Study of the Catalytic Decomposition of Chlorine and Chloromethanes over Carbon Films", Int. J. Chem. Kinet. 17, 849-858 (1985).         [ Links ]
Barin, I. and W. Schuler,"On the Kinetics of the Chlorination of Titanium Dioxide in the Presence of Solid Carbon", Metall. Trans. B 11, June, 199-207 (1980).         [ Links ]
Bernard, J. L., M. Camelot and J. Taverdet,"Effect of the Nature of Carbon on the Formation Reactions of Molybdenum Oxychlorides", C. R. Hebd. Seances Acad. Sci. 279, (26), 1125-1128 (1974).         [ Links ]
Bicerolu, O. and W. H. Gauvin,"The Chlorination Kinetics of Zirconium Dioxide in the Presence of Carbon", Can. J. Chem. Eng. 58, June, 357-366 (1980).         [ Links ]
Bidaye, A., S. Venkatachalam and C. Gupta,"Studies on the Chlorination of Zircon: Part I. Static Bed Investigations", Metall. Mater. Trans. B, 30, April, 205-213 (1999).         [ Links ]
Chen, C. and M. Back,"Kinetics of the Reaction of Oxygen with Thin Films of Carbon", Carbon 17, 495-503 (1979).         [ Links ]
Dunn, W. E.,"High Temperature Chlorination of Titanium Bearing Minerals: Part II", Metall. Trans. B 10, 271-277 (1979).         [ Links ]
Fleischer, M.,"New Mineral Names", The American Mineralogist 49, 1497-1498 (1964).         [ Links ]
González, J., M. Ruiz, J. Rivarola and D. Pasquevich,"Effects of Heating in Air and Chlorine Atmosphere on the Crystalline Structure of Pure Ta₂O₅ or Mixed with Carbon", J. Mat. Sc. 33, 4173-4180 (1998).         [ Links ]
González, J., M. Ruiz, A. Bohé and D. Pasquevich,"Oxidation of Carbons in the Presence of Chlorine", Carbon 37, 1979-1988 (1999).         [ Links ]
Ojeda, M. W., J. Rivarola and O. Quiroga,"Study on Chlorination of Molybdenum Trioxide mixed with Carbon Black", Mineral Engineering 15, (8), 585-591 (2002).         [ Links ]
Nagata, K.,"Selective Refining of Metals by Chlorine Gas", Tetsu to Hagane 73, (9), 1077-1081 (1987).         [ Links ]
Papirer, E., R. Lacroix, J. Donnet, G. Nansé and P. Fioux,"XPS Study of the Halogenation of Carbon Black – Part 2. Chlorination", Carbon, 33, (1), 63-72 (1995).         [ Links ]
Pasquevich, D. M. and V. T. Amorebieta,"Mass Spectrometric Study of Volatile Products during the Carbochlorination of Zirconia", Ber. Bunsengens Phys. Chem. 96, 534-537 (1992).         [ Links ]
Pasquevich, D.M., J. Andrade Gamboa and A. Caneiro,"On the role of carbon in the carbochlorination of refractory oxides", Thermochim. Acta 209, 209-222 (1992).         [ Links ]
Quiroga, O.D., J. R. Avanza and A. J. Fusco, Modelado cinético de las transformaciones fluido-sólido reactivo, Ed. Facena, Corrientes, (Argentina) (1994).         [ Links ]
Ruiz, M. del C., J. B. Rivarola and O. Quiroga,"Kinetic Study of the Reaction Between Molybdenum Trioxide and Gaseous Carbon Tetrachloride", Can. J. Chem. Eng. 72, April, 289-295 (1994).         [ Links ]
Stefanyuk, S. L. and I. S. Morozov,"Kinetics and Mechanism of chlorination of minerals (loparite, pyrochlore, zircon and euxenite)", Zh. Prikl. Khim., 38, (4),737-742 (1965).         [ Links ]
Scheimann, G., O. Immel, H. Rötger, and H. Schmidt,"Zur Thermischen Zersetzung von Tetrachlorkohlenstoffdämpfen und ihre Reaktionstechnische Auswertung", Z. Physik. Chem. Neue Folge, 32, 3-4 (1962).         [ Links ]
Smith, J. M., Chemical Engineering Kinetics. Compañía Editorial Continental, 527-8, México (1977).         [ Links ]

Received: September 16, 2001.
Accepted for publication: September 10, 2002.
Recommended by Guest Editors: J. Cerdá, S. Díaz and A. Bandoni.