Selection of carbon catalysts for the industrial manufacture of phosgene

Abstract

An approach to the evaluation of carbon catalysts suitable for the industrial manufacture of phosgene is described. Relative reactivity and the oxidative stability of catalysts have been measured in laboratory microreactors. Comparison of the performance of full size catalyst pellets versus that of crushed catalyst allowed catalyst effectiveness factors and therefore effective gas diffusivities to be estimated. These data have been combined with a 2-dimensional model incorporating a description of heat and mass transfer to predict catalyst performance in industrial scale phosgene reactors.

---

Introduction

Phosgene is an important chemical intermediate used in the manufacture of polyurethanes, polycarbonates, pharmaceuticals and agrochemicals. It is manufactured industrially via the gas phase reaction of carbon monoxide with chlorine in the presence of an activated carbon catalyst.¹

CO + Cl₂ ⇌ COCl₂

The reaction is strongly exothermic (ΔH = −107.6 kJ mol⁻¹) and is typically carried out in a shell and tube reactor. In industrial operation, the process is typically operated using an excess of carbon monoxide and achieves essentially complete chlorine conversion, with residual chlorine levels being less than 100 ppm. Although reaction commences at 30–60 °C, peak reaction temperatures in excess of 500 °C can be obtained.

There have been relatively few laboratory studies of phosgene formation reported in the literature. Potter and Baron studied reaction kinetics between 25 and 80 °C over an ill defined granular carbon catalyst.² Csũrös and coworkers examined the effect of space velocity and carbon monoxide excess on phosgene formation between 80 and 170 °C.³ Several papers on phosgene reaction kinetics using different commercial catalysts produced by Bayer have been published by Shapatina et al.^4–6 They used Langmuir–Hinschelwood rate expressions to describe phosgene formation kinetics, with different expressions being used depending on the chlorine concentration. More recently Abrams and co-workers have described the use of a synthetic carbon material which is claimed to produce lower levels of carbon tetrachloride compared to many conventional commercially available phosgene catalysts.⁷ This catalyst is commercially available via DuPont. Manufacture of phosgene using a microfabricated packed bed reactor has also been demonstrated by Ajmera et al.⁸

The selection of appropriate catalysts for use in a commercial process is non-trivial. The literature only reports an examination of catalyst kinetics at relatively low temperatures, and these data are not easily extrapolated to industrially relevant conditions. In addition there has been no systematic comparison of the performance of different catalysts. Also, the presence of feedstock impurities such as oxygen may have an effect on the performance of catalysts in industrial operation, and has not been addressed in earlier work. In this paper we report the evaluation of a range of commercially available carbon catalysts in a series of laboratory tests and use these results in combination with a reactor model incorporating a description of heat and mass transfer to infer their behaviour in industrial scale phosgene reactors.

Experimental

A total of 7 commercially available carbon catalysts which are recommended by different suppliers for the manufacture of phosgene have been evaluated. Catalysts were supplied by Chemviron, Donau Carbon, DuPont, Norit and Pica (see Table 1 for details). A sample of a used catalyst taken from a commercially operating plant at the end of catalyst life was also examined. All catalysts were characterised via nitrogen adsorption using a Quantachrome NOVA 1000e analyser. BET surface areas were determined over the pressure range 0.01 < P/P_o < 0.1, whilst micropore volumes and external surface areas were determined using the t-plot method. The pore size distributions of the fresh and used Norit RB4C catalysts were also measured via mercury porosimetry using a Micromeritics Autopore instrument.

Table 1 Characteristics of activated carbon catalysts

Catalyst	Physical appearance	BET surface area^a (m^2 g^−1)	Total pore volume (cm^3 g^−1)	Micropore volume^b (cm^3 g^−1)	External surface area^b (m^2 g^−1)	Ash content (%)
a BET areas determined using the pressure range 0.01< P/Po < 0.1. b Determined using the t-plot method.
Chemviron Solcarb 208C DM	4 × 8 mesh granules	1240	0.53	0.49	45	<3
Chemviron Solcarb 208C DR-P	4 × 8 mesh granules	1200	0.51	0.47	45	<3
Donau Supersorbon K40	4 mm extrudates	1200	0.61	0.52	80	<10
DuPont IPC	2–3 mm spheres	325	0.39	0.05	250	<1
Norit RB4C	4 mm extrudates	1070	0.48	0.42	55	<6
Norit RX3 extra	3 mm extrudates	1250	0.56	0.47	75	<2
Picatal G201	4 × 8 mesh granules	1230	0.53	0.48	40	<4
Norit RB4C (used)	4 mm extrudates	890	0.44	0.35	80	n.a.

---

Catalyst activity measurements were carried out using a micro-reactor system. Chlorine and carbon monoxide gases were fed continuously via mass flow controllers at ∼30 cm³ min⁻¹ to a stainless steel fixed bed reactor of 5 cm length and 4 mm internal diameter. The reactor was placed within an electrically heated tube furnace and two thermocouples were placed approximately 5 mm inside the inlet and outlet of the catalyst bed. The reactor exit stream was combined with a nitrogen gas flow of 1 dm³ min⁻¹. The carbon monoxide and phosgene concentrations in the diluted reaction product were monitored continuously via IR spectroscopy – the mixture being passed through a 10 cm path length continuous flow IR gas cell equipped with NaCl windows. The gas flow from the IR cell was then passed to a scrubber system containing sodium hydroxide to absorb and neutralise unreacted chlorine and the phosgene produced. The absence of internal and external diffusion limitations in the laboratory setup was confirmed using standard methodologies.⁹ The absence of external diffusion limitations was checked by carrying out tests at increasing gas flow rates and fixed contact time; their absence was confirmed when the flow rate of each reactant was greater than 20 cm³ min⁻¹. The absence of internal (pore) diffusion limitations was evaluated by testing different catalyst particle sizes. A catalyst particle size fraction of 125–250 μm was confirmed as being suitable for comparative testing of catalysts. The use of different temperature ramp rates – 0.2, 0.5 and 1 °C min⁻¹ – was shown to have no effect on the results obtained.

In a typical experiment approximately 0.2 g of crushed, sieved catalyst (125–250 μm particle size) was charged to the reactor and after purging with nitrogen, chlorine gas was introduced at a flow of 30 cm³ min⁻¹ at room temperature. Once the temperature had stabilised, carbon monoxide gas was introduced at a flow of 33 cm³ min⁻¹ (10% molar excess) and the set point temperature of the electrical furnace was increased at a rate of 0.5 °C min⁻¹ up to a maximum of ∼50 °C. Overall CO conversions at this maximum temperature fell in the range 5–15%, so the measured phosgene concentrations were converted to phosgene formation rates expressed as mmol min⁻¹ (g catalyst)⁻¹, assuming that the reactor behaves as a differential reactor. For simplicity, the measured phosgene formation rate at the reference temperature of 40 °C was taken as a “single point” comparison of catalyst activity. Activities were also compared to the literature data of Potter and Baron, who report a reaction rate of 0.16 mmol min⁻¹ (g catalyst)⁻¹ at 40 °C.²

Measurements were also carried out on full size catalyst pellets. 10–15 individual catalyst particles were charged to the reactor with a small quantity of silicon carbide powder separating each catalyst particle. At low conversions this string of particles can be approximated to a series of CSTRs. Hence the measured reaction rate can be compared to the conversion obtained with an equivalent weight of crushed catalyst particles. The ratio of the activity of the full size catalyst pellets to that obtained on the crushed catalyst is a measure of the impact of internal pore diffusion on the catalyst performance – the catalyst effectiveness factor.

The oxidative stability of the different catalysts was evaluated using a temperature programmed reactor system. Approximately 0.1 g of crushed, sieved catalyst (250–315 μm particle size) was heated in a 4 mm internal diameter reactor under 100 ml min⁻¹ flow of 2500 ppm O₂ in He. The CO₂ concentration in the exit gas stream was measured using a Hiden HPR-20 QIC mass spectrometer system. The mass spectrometer response for CO₂ was converted to concentrations using a 5000 ppm v/v CO₂–He calibration mixture from a certified gas bottle. A temperature ramp/step programme was used as follows.

1. Ramp to 575 °C and hold for 3 hours.

2. Step and hold for 3 hours at 550 °C.

3. Step and hold for 3 hours at 500 °C.

4. Step and hold for 3 hours at 450 °C.

5. Step and hold for 3 hours at 400 °C.

6. Step and hold for 3 hours at 350 °C.

Results and discussion

The physical characteristics of the catalysts as determined by nitrogen adsorption are summarised in Table 1. All materials except the DuPont IPC catalyst exhibited typical Type I adsorption isotherms which is characteristic of microporous materials. The t-plot analyses revealed that they possessed significant micropore volume with only a limited amount of external surface area. The Dupont IPC catalyst exhibited a Type II adsorption isotherm with a BET surface area of 325 m² g⁻¹ and only a small quantity of micropores.

The rate of phosgene formation obtained with the different catalysts as a function of temperature is shown in Fig. 1. A comparison of the different catalysts is also summarised in Table 2, which reports the rate of formation of phosgene at 40 °C for each catalyst and the relative catalyst activity compared to the literature kinetics reported by Potter and Baron.²

Fig. 1 Rate of phosgene formation vs. temperature.

Table 2 Reactivity testing of carbon catalysts

Catalyst	COCl2 formation rate at 40 °C (mmol min^−1 (g cat)^−1)	Relative activity^a
a Relative to results of Potter and Baron.^2
Chemviron Solcarb 208C DM	0.62	3.9
Chemviron Solcarb 208C DR-P	0.52	3.2
Donau Supersorbon K40	0.60	3.8
DuPont IPC	0.19	1.2
Norit RB4C	0.61	3.8
Norit RX3 extra	0.76	4.4
Picatal G201	0.48	3.0

---

For all the catalysts the reaction profile follows a simple Arrhenius temperature dependence over the range of temperatures studied, where the difference in the temperatures between the inlet and outlet thermocouples was less than 1.5 °C. Activation energies were in the range 42–50 kJ mol⁻¹, which is consistent with literature data.^2–6 If the reaction temperature was increased beyond 55 °C then there was a rapid increase in the measured inlet temperature, which approached 300 °C (see later). The measured activity data clearly show that the Norit RX3 catalyst is the most active catalyst, whilst the DuPont IPC catalyst has a significantly lower activity than all the other catalysts. Apart from the DuPont material all catalysts displayed activities at 40 °C significantly higher than the values reported in the literature. There is no clear relationship between the physical properties of the catalysts reported in Table 1 and their relative activities. However, the relatively low activity of the DuPont catalyst can clearly be attributed to the mesoporous nature of this material.

The rate of phosgene formation as a function of temperature for full size catalyst extrudates/granules was also evaluated for the Norit RX3extra, Donau and Picatal catalysts. A typical result is illustrated in Fig. 2, which shows the results obtained with Donau Supersorbon K40 extrudates compared to those obtained with the crushed catalyst. The ratio of the phosgene formation rates on the full size catalyst particles versus those on the crushed catalyst is a measure of the catalyst effectiveness. Lower catalyst effectiveness means that the reaction is more limited by internal mass transfer (or diffusion within the pores of the catalyst). These data were used in conjunction with the reactor model described below to estimate the effective diffusivity (D_eff) of the reactants within the pores of the catalyst in these experiments – see Table 3. These experimentally derived values were in turn used for the prediction of the behaviour of the catalysts in a reactor model of a full scale industrial reactor.

Fig. 2 Phosgene formation vs. temperature for Donau Supersorbon K40 (crushed catalyst vs. pellets).

Table 3 Experimentally derived effective diffusivity of carbon catalysts

Catalyst	COCl2 formation rate at 40 °C (mmol min^−1 (g cat)^−1)		Effective diffusivity at 40 °C^a (m^2 s^−1)
	Crushed catalyst (125–250 μm)	Full size particles
a Determined using the reactor model.
Donau Supersorbon K40	0.60	0.48	1.0 × 10^−7
Norit RX3extra	0.76	0.68	1.15 × 10^−7
Picatal G201	0.48	0.34	7.5 × 10^−8

---

In order to gain insight into the physical changes a catalyst can undergo upon use, both nitrogen adsorption and mercury porosimetry were used to characterise the fresh and used Norit RB4C catalyst samples. Nitrogen adsorption measurements showed that both the fresh and used Norit RB4C catalysts displayed the typical Type 1 isotherms characteristic of microporous materials. Analysis of the isotherms using the t-plot method showed that there was a decrease in micropore volume from 0.42 to 0.35 cm³ g⁻¹ for the used catalyst, whilst there was an increase in the pore volume determined by mercury porosimetry from 0.52 to 0.67 cm³ g⁻¹. The pore size distributions derived from the mercury porosimetry measurements are shown in Fig. 3. There is little change in the macropore structure of the fresh and used catalysts. However, there is a substantial increase in the quantity of pores in the 10–50 nm range for the used catalyst. These data show that there is a significant loss of carbon from the catalysts upon use over time – i.e. a burn-out of the pores.

Fig. 3 Pore size distributions for fresh and used Norit RB4C catalysts.

The loss of carbon could be a result of oxidation of carbon by trace levels of oxygen present in the chlorine or the reaction of the carbon with chlorine to form carbon tetrachloride. Abrams et al. showed through the use of ¹³C labeling techniques that carbon tetrachloride formation in the phosgene process is a result of the reaction of chlorine with the carbon surface.⁷ The reactive nature of activated carbon materials with chlorine above 300 °C has also been shown by Khachatryan and Dellinger.¹⁰ However, without detailed information of the levels of trace oxygen in the chlorine used in the plant and kinetic models for the reaction of the carbon with chlorine and oxygen it is not possible to distinguish between these two potential causes of carbon loss. Nevertheless it is important to highlight the results of the oxidative stability tests below which showed that the Norit RB4C catalyst was not particularly stable to oxidation at high temperatures.

For the oxidative stability tests using the temperature programmed reactor system, the steady state mass spectrometer response was converted to a CO₂ concentration, and from these data, reaction rates as a function of temperature were calculated. The results obtained with the different catalysts are shown in the form of an Arrhenius plot (ln[rate] vs. 1/T) in Fig. 4. For clarity, the results obtained with the Chemviron catalysts are not shown, as oxidation rates were slightly higher than those obtained with the Picatal material. The results are also reported in Table 4 which summarises the CO₂ formation rate at 500 °C obtained for each catalyst as well as the calculated activation energy for CO₂ formation.

Fig. 4 Arrhenius temperature dependence of oxidation rate of carbon catalysts.

Table 4 Oxidative stability tests of carbon catalysts

Catalyst	CO2 formation rate at 500 °C (mol min^−1 (g cat)^−1)	Activation energy (kJ mol^−1)
Chemviron Solcarb 208C DM	5.8 × 10^−5	117
Chemviron Solcarb 208C DR-P	5.3 × 10^−5	127
Donau Supersorbon K40	3.0 × 10^−6	162
DuPont IPC	1.1 × 10^−5	153
Norit RB4C	2.0 × 10^−5	143
Norit RX3 extra	2.0 × 10^−6	169
Picatal G201	4.0 × 10^−5	122

---

The results obtained in the oxidative stability test for the three granular carbon catalysts (Picatal G201, Chemviron Solcarb 208C DM and Solcarb 208C DR-P) show high rates of oxidation to carbon dioxide. The Norit RB4C catalyst is slightly more stable than these three materials. The DuPont IPC catalyst is less stable than both the Norit RX3extra and Donau material. This is perhaps surprising as the work of Abrams et al. reports a high degree of oxidative stability for this catalyst.⁷ The calculated activation energies show that the oxidation of the carbons is highly sensitive to temperature, particularly for the more stable materials, such as the Norit RX3extra and Donau carbons. The measured range of activation energies (120–170 kJ mol⁻¹) are comparable to those reported in the literature.¹¹ Based on the results of oxidative stability testing the Norit RX3extra and Donau Supersorbon K40 catalysts would appear to be the most promising materials to be used for phosgene manufacture when oxygen is present in the chlorine used for phosgene manufacture.

Reactor model

A standard 2-dimensional fixed bed reactor model incorporating “best” available descriptions of reaction kinetics, fluid physical properties and heat transfer within a catalytic packed bed was developed and encoded in Aspen Custom Modeler version 12.1. The standard heat and mass balance equations for a two dimensional fixed bed reactor model are as follows:¹²

with boundary conditions as follows:

C = C₀ and T = T₀ at z = 0 and 0 ≤ r ≤ R_t

The reaction kinetics were re-parameterised from the original data obtained by Potter and Baron.¹ Their experimental data were refitted using a slightly simplified form of kinetics.

The overall conversion becomes limited by thermodynamics at temperatures above 400 °C (see Fig. 5), so an equilibrium driving force term was incorporated into the reaction rate expression. A scaling factor, R₀, was also introduced to account for the significantly higher activity found in the commercial catalysts tested in this work.

where k₁ = exp(− (1515/T − 3.70)), k₂ = exp(5330/T − 14.25), k_eq = exp(13171/T − 16.447).

Fig. 5 Equilibrium conversion of an equimolar mixture of CO + Cl₂ as a function of temperature and pressure.

Catalyst effectiveness factors (η) were calculated using a numerical approximation methodology with the measured catalyst effectiveness at low temperatures being used to determine the effective diffusivity for the different catalysts.

The effective diffusivity of the reactants within the catalyst pores, D_eff, can be derived from first principles and is defined by the following expressions: where D_AB is the bulk gas diffusivity and D_k is the Knudsen diffusivity, which is related to the average pore radius, r_pore, the average reactant molecular weight and temperature.

For typical catalysts used for phosgene synthesis average pore radii fall in the range 10–100 Å, so the diffusion of the reactants within the catalyst particles is dominated by Knudsen diffusion processes. Therefore we can write the following expression.

For activated carbons, the average pore radius, the internal catalyst voidage and tortuosity are rather poorly defined parameters, as it is known that they can possess complex pore size distributions. Therefore we have chosen to estimate values for the effective diffusivity from the effectiveness factor experiments carried out under defined conditions in the laboratory microreactor as described earlier, and then calculate the effective diffusivity within the reactor model using the following expression.

Radial heat transfer within the packed bed was assumed to comprise two heat transfer resistances in series – an effective radial conductivity within the bed itself, λ_e,r, and a fluid-wall heat transfer resistance, α_w.

Taking typical values for catalyst particle thermal conductivity λ_s and fluid thermal conductivity λ_g, the overall correlation for radial thermal conductivity was reduced to the following expression which incorporates a scaling parameter to allow a degree of flexibility in accounting for any under- or over-estimation of the radial heat transfer coefficient.

The wall heat transfer coefficient, α_w, was calculated from the correlation recommended by Dixon and Cresswell.¹³

Within the model all reaction rates, physical properties and mass and heat transfer parameters were recalculated locally at each point in the catalyst bed.

Validation of the model was carried out using both laboratory data and data on peak temperatures reported in a commercially operated phosgene plant. Fig. 6 shows the temperature profile predicted by the model compared to data obtained using the laboratory reactor with the Norit RX3extra catalyst operating at a set point temperature of 100 °C. With the overall heat transfer parameter scaled to match the latter part of the cooling curve in the experimental data excellent agreement between the measured and predicted peak temperatures is obtained.

Fig. 6 Temperature profile in laboratory reactor (set temperature = 100 °C) with crushed Norit RX3extra catalyst.

Details of a full scale phosgene reactor are provided in World Patent WO03/72237 and are summarised in Table 5.

Table 5 Details of an industrial phosgene reactor (from WO03/72237)

Reactor tube diameter	39.3 mm
Number of tubes	1256
Catalyst bed length	2.7 m
Phosgene production rate	10 000 kg h^−1
CO excess(bed inlet)	4.2%
Inlet pressure	4 bar
Coolant temperature (MCB)	60 °C
Inlet gas temperature	50 °C

---

Although the details of the catalyst are not disclosed in this patent, a peak temperature of 545 °C is reported. Fig. 7 shows the centreline axial temperature profiles predicted for the Donau, Norit RX3extra and Picatal catalysts used in this work. The predicted peak temperatures fall in the range 547–565 °C, i.e. very similar to that observed in the industrially operated phosgene reactor, confirming the suitability of the model.

Fig. 7 Centreline temperature profiles in a full scale phosgene reactor for Donau Supersorbon K40, Norit RX3extra and Picatal G201 catalysts.

The overall reactor temperature profile predicted for the Donau Supersorbon K40 catalyst is shown in Fig. 8. Here it is clear how the sharp exotherm is developed within 10–15 cm at the inlet of the reactor with radial temperature gradients of around 300 °C across the relatively small diameter of the reactor tubes. Even close to the reactor wall, temperatures in excess of 200 °C are predicted at the reactor inlet.

Fig. 8 Temperature profile in a full scale phosgene reactor with Donau Supersorbon K40 catalyst.

The model can also be used to explore the effect of changing operating conditions on the performance of a full scale phosgene reactor. The effect of varying throughput on the centreline temperature profile is shown in Fig. 9 for the reactor described above containing the Donau Supersorbon K40 catalyst. Decreasing reactor throughput results in a narrowing of the peak temperature profile accompanied by an increase in peak temperature. Increasing throughput above the reference 10 000 kg h⁻¹ phosgene production rate results in a decrease in peak temperature, but with only a small degree of broadening of the peak width. The decrease in peak temperatures with increasing throughput is readily explained by the changes in radial heat transfer with overall fluid velocities (and hence Reynolds number) through the reactor tubes. In addition to the predicted changes in temperatures, the model also predicts an exponential increase in the levels of residual chlorine exiting the reactor with increasing throughput.

Fig. 9 Effect of throughput on centreline temperature profile in a full scale phosgene reactor with Donau Supersorbon K40 catalyst.

The predicted peak temperatures for the Donau Supersorbon K40, Norit RX3extra and Picatal G201 catalysts are 547, 565 and 551 °C respectively. Oxidation rates of these catalysts derived from the results of the tests described earlier at the calculated peak temperatures are 1.2 × 10⁻⁵, 1.5 × 10⁻⁵ and 1.3 × 10⁻⁴ mol min⁻¹ g ⁻¹ respectively. This, coupled with the slightly narrower temperature peak width indicated by the model as shown in Fig. 7 would suggest that for a process operating with traces of oxygen present in the chlorine being fed to the reactor, the Donau catalyst would be the preferred catalyst.

Conclusions

The relative reactivity of different commercially available carbon catalysts suitable for phosgene manufacture has been examined in a laboratory microreactor. Although a wide range of activities has been observed, all the materials tested were significantly more active than catalysts reported in the literature. Comparison of the performance of full size catalyst pellets versus that of crushed catalyst allowed effectiveness factors and therefore effective gas diffusivities to be determined. These data were used in conjunction with a 2-dimensional reactor model incorporating heat and mass transfer to predict the performance of different catalysts within a full scale phosgene reactor.

The oxidative stability of the different catalysts in the presence of relatively low levels of oxygen has also been evaluated using a temperature programmed methodology. The different carbons exhibit very different degrees of stability with the rate of oxidation at 500 °C varying by more than an order of magnitude. This information, in conjunction with knowledge of the predicted peak reactor temperatures can be used to select preferred catalysts for use in full scale phosgene production.

Glossary

C i	Concentration of species i, kmol m^−3
c p	Specific heat of gas, kJ kg^−1 K^−1
d p	Particle diameter, m
D AB	Diffusivity of A in mixture of A and B, m^2 s^−1
D eff	Effective diffusivity, m^2 s^−1
D k	Knudsen diffusivity, m^2 s^−1
h 0	Heat transfer scaling parameter
P i	Partial pressure of species i, Pa
R	Reaction rate, kmol kg^−1 s^−1
R 0	Reaction rate scaling parameter
r	Radius, m
r pore	Average pore radius, m
Rep	Particle Reynolds number
R t	Reactor tube radius, m
T	Temperature, K
T wall	Temperature of reactor tube wall, K
u s	Superficial gas velocity, m s^−1
z	Distance along reactor, m
α w	Gas-wall heat transfer coefficient, kJ m^−2 s^−1 K^−1
ε	Bed void fraction
η	Catalyst effectiveness factor
λ e,r	Effective radial thermal conductivity, kJ m^−1 s^−1 K^−1
λ g	Gas thermal conductivity, kW m^−1 K^−1
ρ B	Catalyst bulk density, kg m^−3
ρ g	Gas density, kg m^−3
τ	Tortuosity factor

Notes and references

1. T. A. Ryan, C. Ryan, E. A. Seddon and K. R. Seddon, Phosgene and Related Carbonyl Halides, Elsevier, 1996.
2. C. Potter and S. Baron, Chem. Eng. Prog., 1951, 47, 473–479.
3. Z. Csũrös, R. Soós, I. Petneházy and G. T. Szabó, Period. Polytech., Chem. Eng., 1970, 14, 3–11.
4. E. N. Shapatina, V. L. Kuchaev, B. E. Pen’kovoi and M. I. Temkin, Kinet. Catal., 1976, 17, 559–566.
5. E. N. Shapatina, V. L. Kuchaev and M. I. Temkin, Kinet. Catal., 1977, 18, 968–972.
6. E. N. Shapatina, V. L. Kuchaev and M. I. Temkin, Kinet. Catal., 1978, 20, 972–976.
7. L. Abrams, W. V. Cicha, L. E. Manzer and S. Subramoney, Stud. Surf. Sci. Catal., 2000, 130, 455–460.
8. S. K. Ajmera, M. W. Losey, K. F. Jensen and M. A. Schmidt, AIChE J., 2001, 47, 1639–1647.
9. J. Perez-Ramirez, R. J. Berger, G. Mul, F. Kapteijn and J. A. Moulijn, Catal. Today, 2000, 60, 93–109.
10. L. Khachatryan and B. Dellinger, Chemosphere, 2003, 52, 709–716.
11. C. Li and T. C. Brown, Carbon, 2001, 39, 725–732.
12. G. F. Froment and K. B. Bischoff, Chemical Reactor Analysis and Design, Wiley, 2nd edn, 1990.
13. A. G. Dixon and D. L. Cresswell, AIChE J., 1979, 25, 663–676.

---