Catalyst Distribution Strategies in Fixed-Bed Reactors for Bromine

Jan 13, 2014 - Adopting experimental data available in the literature and considering model simplifications, we rationalize effective strategies to re...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/IECR

Catalyst Distribution Strategies in Fixed-Bed Reactors for Bromine Production Maximilian Moser, Laura Rodríguez-García, and Javier Pérez-Ramírez* Institute for Chemical and Bioengineering, Department of Chemistry and Applied Biosciences, ETH Zurich, Vladimir-Prelog-Weg 1, CH-8093 Zurich, Zurich, Switzerland ABSTRACT: The gas-phase oxidation of hydrogen halides to the corresponding halogens has gained industrial relevance following the recent discovery of active and stable heterogeneous catalysts based on RuO2 and CeO2. These developments originated from the need of polyurethane and polycarbonate producers to recycle the large amounts of byproduct HCl to Cl2 for further use in the phosgenation step. More recently, the catalyzed HBr oxidation to Br2 has been regarded as an enabling technology for bromine-mediated alkane functionalization processes. The highly exothermic and corrosive character of these reactions, especially HBr oxidation, requires particular precautions regarding temperature control to guarantee stable and safe operation in connection with the catalyst, the reactor, and ultimately the plant. Adopting experimental data available in the literature and considering model simplifications, we rationalize effective strategies to recover bromine, which involve simulations of staged RuO2/TiO2 and CeO2/ZrO2 catalyst beds of variable activity in fixed-bed reactors. A cooled tubular and an adiabatic cascade configuration are considered. Options to optimize process economics by minimizing catalyst cost and number of unit operations are discussed.

1. INTRODUCTION The gas-phase oxidation of HCl to Cl2 (eq 1) has attracted considerable attention in recent years as an energy-efficient way to recycle the large amounts of byproduct HCl generated during the phosgenation step in the manufacture polyurethanes (PU) and polycarbonates (PC).1,2 Till now, this HCl was absorbed by the polyvinyl chloride (PVC) industry for the production of 1,2ethylene dichloride via the oxychlorination of ethylene. However, the faster-growing demand for PC and PU compared to PVC has gradually shifted the business model toward the in situ recovery of chlorine for phosgene synthesis. Research efforts have crystallized into the development of active and stable RuO2 and CeO2-based catalysts, exhibiting low- and high-temperature activity, respectively.3−6 The latter systems resolve the severe volatilization issues characteristic of the original Deacon’s CuCl2based catalyst and the many unsuccessful analogues developed in the course of the 20th century.1 Very recently, we have demonstrated in the laboratory scale the efficiency of the new generation of Deacon catalysts for the analogous oxidation of HBr to Br2 (eq 2).7 This process is of interest in the context of the utilization of bromine for the functionalization of abundant alkanes, such as shale gas, to the corresponding alkyl bromide followed by its reaction into a broad spectrum of fuels and chemicals.8,9 The associated bromination and elimination steps generate two moles of HBr per mole of alkane converted, and its recycling to Br2 via catalyzed oxidation stands as a robust and economic process due to the low energy requirement.7 2HCl +

1 O2 ⇌ Cl 2 + H 2O, 2

A critical success factor for the large-scale implementation of catalytic processes for HCl or HBr oxidation comprises the selection of suitable reactor systems. The highly exothermic and corrosive character of these reactions, in particular HBr oxidation, requires precautions to guarantee stable and safe operation in connection with the catalyst, the reactor, and ultimately the plant. Specifically, heat removal and temperature control are the main challenges at the reactor scale in order to avoid catalyst deactivation and thermal runaways.1,2 For example, RuO2-based catalysts are very sensitive to high local temperatures, e.g. due to hot spot formation, leading to their rapid deterioration by active phase sintering and loss.2,10 Similarly, the traditional CuCl2 and Cr2O3-based catalysts experienced severe metal volatilization due to the reaction exothermicity.11−13 For this reason, fluidized-bed reactors have been the natural choice for large-scale HCl oxidation,14−16 approaching isothermal conditions due to the highly efficient heat exchange between the well-mixed solid bed and the immersed heat exchangers. However, in a fluidized bed, catalyst attrition cannot be eliminated, and, therefore, the reactor has to be equipped with cyclones for fines separation, leading to increased complexity and cost. Recent work has reported the efficient use of fixed-bed reactors for HCl oxidation, achieving a proper heat management and temperature control while maximizing the conversion per pass due to the higher catalyst loading compared to fluidized beds. Sumitomo developed an isothermal multitubular fixed-bed reactor, in which several thousand tubes of small diameter

ΔHr0 = − 56.8 kJ mol−1 Special Issue: Massimo Morbidelli Festschrift

(1)

1 2HBr + O2 → Br2 + H 2O, 2

ΔHr0

Received: Revised: Accepted: Published:

−1

= − 138 kJ mol

(2) © 2014 American Chemical Society

9067

November 2, 2013 January 10, 2014 January 13, 2014 January 13, 2014 dx.doi.org/10.1021/ie4036675 | Ind. Eng. Chem. Res. 2014, 53, 9067−9075

Industrial & Engineering Chemistry Research

Article

providing a great heat exchange area are filled with catalyst and cooled by a circulating heat transfer salt (HTS, Figure 1a).2,17

2. METHODOLOGY The required kinetic data for the reactor modeling was taken from laboratory experiments on gas-phase HBr oxidation over supported CeO2/ZrO2 and RuO2/TiO2 catalysts in particulate form (0.4−0.6 mm). For their preparation, RuCl3 and Ce(NO3)3 precursors were dissolved in water according to the desired metal loading and subsequently applied to the carriers (rutile TiO2 and monoclinic ZrO2, respectively) by incipient wetness impregnation, followed by drying and calcination.6,7 The HBr oxidation over both CeO2/ZrO2 and RuO2/TiO2 catalysts was studied at ambient pressure in a continuous-flow fixed-bed reactor setup.7 The Arrhenius plot shows that the RuO2/TiO2 catalyst is active at ca. 100 K lower temperatures compared to CeO2/ZrO2 (Figure 2). The maximum and optimal temperature conditions

Figure 1. Fixed-bed reactor configurations analyzed in this study for catalytic HX oxidation (X = Cl, Br). (a) In cooled multitubular reactors, catalyst A (i.e., RuO2/TiO2) is placed in parallel tubes cooled by circulating heat transfer salt (HTS). A temperature maximum (hot spot) can form in the low-conversion ratio region. If the temperature rises above an allowable limit (Tmax), catalyst deactivation or damaging of the reactor are likely to occur. To overcome this problem the reactor tubes can be filled with diluted beds of catalyst A. These catalysts are staged in order of their activity (A3 < A2 < A1) starting with the least active in the low-conversion ratio region to avoid hot spot formation and increasing the activity toward the outlet to improve the HX conversion. (b) In the adiabatic reactor cascades, catalyst A is loaded into a sequence of reactors, which are separated to admit intermediate cooling of the gas stream and HX, shifting the chemical equilibrium to improve especially the HCl conversion and keeping the temperature rise below the limit (Tmax,A). The combination of catalyst A with a robust high-temperature catalyst B (i.e., CeO2/ZrO2) allows for the utilization of the adiabatic temperature rise above the temperature limit of catalyst A.

Figure 2. Arrhenius plot of HBr oxidation with experimental data (symbols) and fittings of the reaction rate (lines). The simulation shows good agreement with the experimental data.

for both systems are constrained by the limits for catalyst or reactor materials. Hence, the maximum temperature for operating a RuO2/TiO2 catalyst is set to 663 K to avoid the formation of volatile RuO4.10 CeO2-based catalysts are more stable at higher temperatures and can sustain temperatures close to the reactor material limit of 723 K (Table 1). Table 1. Kinetic and Operational Data of Catalysts in HBr Oxidation catalyst

Different variants have been practiced to minimize hot spots in the low-conversion ratio region of the tubes, such as their filling with catalysts of variable activity or inert beds as well as the application of independent cooling regions.18 Differently, Bayer developed a modular adiabatic reactor system, featuring a cascade of reactors with interstage cooling and injection of reactant (Figure 1b).19 This manuscript evaluates various strategies to optimally arrange catalytic materials of different characteristics in fixed-bed reactors for the oxidation of HBr to Br2. The high exothermicity of this reaction aggravates the heat transport and removal problem, and its process feasibility has so far not been assessed in the literature. We conceptually demonstrate that the staging of CeO2/ZrO2 and RuO2/TiO2 catalysts in cooled and adiabatic reactors embodies a viable strategy for safe and controllable HBr oxidation to effectively recover bromine under mild conditions.

RuO2/ TiO2 CeO2/ ZrO2

k0/mol Br2 h−1 kgcat−1 bar−1.5 Topt /K

metala/ wt %

Ea/ kJ mol−1

2

65

4.0·109

413

663

9

58

4.0·107

553

723

Tmax /K

a

Ruthenium or cerium content in the catalysts determined experimentally.

The apparent kinetic parameters of each catalyst, required for the reactor model, were taken from the linear fit of the data points in the Arrhenius plot (Figure 2) and are listed in Table 1. With these values the forward reaction rate constant k is solved and defined as ⎛ E ⎞ k(T ) = k 0 exp⎜ − a ⎟ ⎝ RT ⎠ 9068

(3)

dx.doi.org/10.1021/ie4036675 | Ind. Eng. Chem. Res. 2014, 53, 9067−9075

Industrial & Engineering Chemistry Research

Article

where Fi is the molar flow rate of component i, Wcat is the catalyst weight, and νi is the stoichiometric coefficient of component i. For the heat transport, a homogeneous model is used, which defines the bulk gas temperature as common temperature in the catalyst bed. Its application will result in a slightly lower catalyst surface temperature, leading to a slight underestimation of the reaction rate. The application of the Mears criterion for the radial heat transport limitation concluded that we are very close to its limit. Since the formation of radial temperature gradients is not considered, the values of the catalyst bed temperature can be slightly lower than expected. Hence, the energy balances for cooled tubular and adiabatic reactors are defined as

where k0 is the pre-exponential factor, Ea is the activation energy, T is the temperature, and R is the universal gas constant. The specific reaction rate rw was modeled by implementing a power law rate expression for both catalysts rw(T ) = ak(T )yHBr yO0.5 P1.5

(4)

2

where a is the activity factor, yi is the molar fraction of component i, and P is the total pressure. The equilibrium of the HBr oxidation is shifted far to the product side, thus, the rate equation only considers the forward reaction.7 The activity factor a is included in the rate expression in order to account for CeO2/ ZrO2 and RuO2/TiO2 catalysts with different activities. The value of a can be adjusted between 0 and 1, where the actual rate of the tested catalysts in Figure 2 is obtained, when a = 1. Lower values of a are typically obtained in practice by decreasing the content of active metal in the catalytic materials. One aspect that has not been considered in the reaction rate expression is product inhibition. The strong inhibiting effect of Cl2 and H2O in HCl oxidation has been observed over several Deacon catalysts.4,10,20 However, no kinetic investigations have been conducted for HBr oxidation to account for this effect. According to the HBr conversion profiles reported by us,7 product inhibition could be particularly relevant at high conversion levels (close to 90%), resulting in a decreased reaction rate. Thus, the simulation results under these conditions are likely to overestimate the calculated conversion and temperature. The conceptual design of the reactors was based on recent patent literature for chlorine production.19,21 Details on the cooled tubular reactor design are reported in Table 2. The bed

dT iso = dWcat

value

bed length reactor diameter catalyst bed density bed voidage overall heat transfer coefficient

0.5−1 m 0.025 m 1300 kg m−3 0.68 105 W m−2 K−1

N

∑i = 1 Fc i p, i

(6)

(7)

where Tiso is the temperature in the cooled tubular reactor, Tadi is the temperature in the adiabatic reactor, ΔHr is the reaction enthalpy for HBr oxidation, ρcat is the catalyst density, U is the overall heat transfer coefficient, D is the diameter of the reactor, cp,i is the specific heat capacity of component i, and ΔT is the temperature difference between heat transfer salt and catalyst bed. The wall temperature is assumed to be constant over the entire reactor. The heat capacities of all components were calculated using gas-phase thermochemistry data from the NIST Chemistry WebBook.22

3. RESULTS AND DISCUSSION 3.1. Cooled Multitubular Reactor. Multitubular processes may suffer from severe heat removal issues, owing to the nonuniform cooling temperature of the heat transfer salt (HTS). This phenomenon is often unavoidable along the reactor, due to the large heat uptake of the molten salt and its insufficient flow rate.23 The resulting temperature fluctuations can have a significant impact on the thermal process control of the single tubes, especially in the case of highly exothermic and sensitive processes like HBr oxidation (Figure 3a). To demonstrate this, the temperature sensitivity for a process using a single RuO2/ TiO2 catalyst bed was simulated at different HTS temperatures. Despite the high dilution of HBr in the feed stream (i.e., 10 vol % HBr), already an increase of 50 K causes severe hot spot formation of up to 200 K close to the reactor inlet reaching the limiting temperature for RuO2/TiO2. At these temperatures, hot spots promote fast catalyst deterioration, driven by the oxidation of RuO2 to volatile RuO4. In addition, the reactors experience suboptimal catalyst utilization as the HBr conversion levels off within 10% of the catalyst bed (Figure 3a). However, due to the large temperature increase, it is likely that radial temperature gradients and transport limitation in the catalyst shapes occur, which could lead to an overestimation of the temperature and the HBr conversion as mentioned in Section 2. The optimization of heat removal could assist to compensate the exothermicity and thus to reduce the hot spot temperature. However, the improvement of heat transfer in fixed-bed reactors can be very demanding and increase the process costs significantly. An alternative is the optimization of the thermal conductivity of the catalyst itself as it was elegantly investigated by Sumitomo for their HCl oxidation process.24 In the case of

voidage, taken from the patent literature, is very large (0.68), which is likely due to the large size of the cylindrical catalyst bodies (diameter = 3 mm and length = 7 mm) compared to the reactor tube (diameter = 25 mm). This reported value was used because the thermal conductivity of the catalyst filling was given, which was required for the heat transport equations. A reduction of the void fraction to a common value of 0.3 to 0.4 not only will likely increase the thermal conductivity of the packed bed but also will cause more heat formation per reactor volume leading to larger hot spots. For the reactor model, we assumed ideal gas behavior, steadystate conditions, and a plug-flow velocity profile. The pressure drop was neglected after the calculation of the Ergun equation indicated no significant impact on the reaction. No external gradients for the shaped bodies are expected following the calculation of the Carberry number. However, the Weisz-Prater criterion was close to the limit, thus internal gradients may be present to a certain extent, lowering the catalyst effectiveness. Thus, the mass balances for all components (i = O2, H2O, HBr, Br2) can be written as dFi ν = i rw(T ) dWcat νBr2

4 Dρcat

r w

−ΔHrrw(T ) dT adi = N dWcat ∑i = 1 Fc i p, i

Table 2. Data of the Cooled Tubular Reactor parameter

(−ΔH r (T) − U ΔT )

(5) 9069

dx.doi.org/10.1021/ie4036675 | Ind. Eng. Chem. Res. 2014, 53, 9067−9075

Industrial & Engineering Chemistry Research

Article

Figure 3. Axial coordinate versus temperature (top graph) in (a) to (c) and HBr conversion (bottom graph) in (a) and (c) of an cooled tubular reactor with RuO2/TiO2 (a, b) and CeO2/ZrO2 catalyst (c). The maximum operation temperatures for RuO2/TiO2 and CeO2/ZrO2 catalysts are indicated (dashed lines). The high activity and exothermicity of the reaction drives the bed temperature to extreme values, although the volume fraction of HBr and the inlet temperature are very low.

Figure 4. Axial coordinate versus temperature (top graph) and HBr conversion (bottom graph) of a cooled tubular reactor with RuO2/TiO2 (a), CeO2/ ZrO2 (b), and a staged bed (c). The staging of diluted beds and the combination of RuO2/TiO2 with CeO2/ZrO2 catalyst beds enables a controllable and safe operation to fully recover bromine under mild conditions.

HBr oxidation over a single RuO2/TiO2 catalyst bed, the increase of the overall heat transfer coefficient lowers the catalyst bed temperature gradually over the entire axial reactor coordinate, but it is not be able to counteract the rapid and vast hot spot formation, which drives the reaction to full conversion at the inlet of the reactor (Figure 3b). Still, as indicated in Section 2, it is possible that product inhibition occurs at conversion levels above 85%. However, in the simulations, full HBr conversion is achieved, since the simplified model does not consider this effect. Thus, the

temperature and conversion profiles above to 85% HBr conversion are likely higher and steeper than to be expected. Obviously, a high specific activity is not always beneficial for a catalytic process, especially, when it hampers its own stability like in the case of RuO2/TiO2. Consequently, RuO2/TiO2 needs to be replaced by a catalyst that is robust enough to cope with these harsh temperature conditions instead. CeO2/ZrO2 has shown to be stable at much higher temperatures than RuO2/TiO2 in HCl oxidation.6 In order to assess its potential, a single bed of CeO2/ ZrO2 was simulated under the same conditions as used above for 9070

dx.doi.org/10.1021/ie4036675 | Ind. Eng. Chem. Res. 2014, 53, 9067−9075

Industrial & Engineering Chemistry Research

Article

Figure 5. Parametric sensitivity analysis of HBr oxidation in an cooled tubular reactor. Temperature versus axial coordinate for a single catalyst bed (a) and a staged bed (c). Their corresponding second-order derivatives (d2T/dXHBr2) versus HBr conversion are shown in (b) and (d), respectively, with the sensitive region indicated (gray boxes). The single bed reactor (b) is temperature sensitive, while the staged-bed reactor (d) does not exhibit parametric sensitivity.

conversion, the HTS temperature was increased to 473 K. At this temperature, hot spot formation over the original RuO2/TiO2 catalyst was evident (Figure 3a). However, in the case of the staged bed, pretty homogeneous temperature and activity profiles can be obtained (Figure 4a). In addition, the utilization of the catalyst bed is largely improved as the HBr conversion progresses almost linearly along the full length of the reactor. As a result, the ruthenium metal content for the catalyst is reduced by ca. 75% compared to the original RuO2/TiO2 catalyst, assuming that the reduction of the activity scales linearly with the lower ruthenium content of the catalyst. The same strategy can be applied to CeO2/ZrO2 catalysts. Therefore, a reactor using two staged CeO2/ZrO2 beds with a = 0.5 and 1 was implemented. Under the adjusted temperature conditions, the catalyst cannot fully convert all HBr and would require a higher process temperature (Figure 4b). However, rather than increasing the temperature, the combination of

RuO2/TiO2 (Figure 3c). Indeed, no severe hot spot formation is observed when the temperature of the reactor is increased, ensuring safe operation of the CeO2/ZrO2 catalyst. The insensitivity of the CeO2/ZrO2 catalyst relates to its lower intrinsic activity compared to RuO2/TiO2, resulting in a shift of the temperature maximum toward the middle of the reactor, and a more steady increase of the HBr conversion along the full length of the catalyst bed. Thus, lowering of the specific catalyst activity appears to be a good compromise between performance and stability. In fact many processes like HCl oxidation,2,18 the oxidation of naphthalene or o-xylene to phthalic anhydride,25,26 and oxychlorination of ethylene27 use staged catalyst beds with increasing catalytic activity, which verifiably enhance their overall process performance. Following this strategy, three beds of RuO2/TiO2 with variable activity (a = 0.05, 0.1, and 0.4) were staged in a cooled reactor tube. In order to achieve full HBr 9071

dx.doi.org/10.1021/ie4036675 | Ind. Eng. Chem. Res. 2014, 53, 9067−9075

Industrial & Engineering Chemistry Research

Article

Figure 6. Axial coordinate versus temperature (top graph) and HBr conversion (bottom graph) of an adiabatic reactor with RuO2/TiO2 (a), CeO2/ ZrO2 (b), and diluted catalyst beds (c). The maximum operation temperatures for the RuO2/TiO2 (dashed line) and CeO2/ZrO2 catalysts (dotted line) are indicated in (c). The temperature rise along the adiabatic reactor depends on the inlet conditions. By optimizing the HBr feed and the catalysts activity, the maximum temperature rise can be adjusted and kept below the maximum temperature.

to the cooled reactor process, no heat removal is required. Consequently, the design is technically less complex, lowering the process cost and improving the scalability.1 The adiabatic process is very interesting, in particular for HBr oxidation, since the operation temperature is relatively low and the reaction is not limited by equilibrium. According, full HBr conversion can be attained despite the temperature rise associated with the exothermicity of the reaction. Analyses of RuO2/TiO2 and CeO2/ZrO2 beds were conducted for different inlet temperatures in the range of 383−443 K (Figure 6a, b). The temperature profiles display very steep increases of up to 300 K and full HBr conversion over both catalysts. The temperature becomes constant as soon as all HBr is oxidized and by increasing the inlet temperature less catalyst material is required to reach full HBr conversion (Figure 6a, b). Obviously, the adiabatic temperature rise should not exceed the corresponding catalyst or reactor limits, thus restricting the amount of HBr fed into a single reactor. In contrast to the cooled process, the temperature rise cannot be affected by lower catalyst activity, since this only leads to an increase of the catalyst volume and to larger reactors (Figure 6c). Therefore, a different strategy is required to optimize the adiabatic process. Slight changes of the catalyst activity and/or loading in the adiabatic cascade can lead to higher temperatures surpassing the established limit. However, the amount of heat formed during the reaction and thus the temperature rise along the reactor are limited by the amount of HBr that can be converted The interstage feeding of HBr and the heat exchange enable to optimize the temperature rise in each reactor of the cascade, thus variations in the catalyst activity and loading can be compensated to avoid a temperature overshoot. Industrial adiabatic processes use reactor cascades consisting of several consecutive catalyst beds in order to convert large amounts of reactants in combination with interstage heat removal and reactant addition. This concept is used for many key processes, e.g. HCl oxidation,1 ammonia synthesis,30 or SO2

CeO2/ZrO2 and RuO2/TiO2 in one staged bed is suggested, in order to profit from the advantages of both catalytic systems, i.e. the high activity of RuO2/TiO2 at advanced reactant depletion and the superior thermal stability of CeO2/ZrO2 (Figure 4c). In this configuration the CeO2/ZrO2 (a = 1) replaces the diluted RuO2/TiO2 bed (a = 0.05) at the inlet of the reactor tube, where fast catalyst deterioration is expected, owing to the hot spot formation. Full HBr conversion is attained, while HTS temperature is only slightly increased compared to the RuO2/ TiO2-based design (Figure 4a). In order to verify the stability of this staged-bed process under optimized conditions, a parametric sensitivity analysis was conducted (Figure 5).28 For the original reactor operation shown in Figure 3, the second order derivative of the temperature over the HBr conversion became positive before the temperature maximum (Figure 5a, b), thus indicating that slight changes in the temperature can lead to a reactor runaway. The optimized staged-bed process depicted at no point a positive second derivative and, therefore, is assumed insensitive to small temperature fluctuations (Figure 5c, d). Apart from the presented solutions, further optimization of the staged bed is possible depending on the target of the optimization, for example the catalyst cost. The amount of precious ruthenium can be decreased significantly with the novel staged bed but must be counterbalanced by cerium. Current market prices for ruthenium and cerium metal amount to 2732 USD kg−1 and 15 USD kg−1, respectively.29 Thus, the price for a catalyst filling is 20% lower for the staged bed using both RuO2 and CeO2-based catalysts (Figure 4c) than for the system with staged RuO2/TiO2 beds (Figure 4a). Still, the implementation of a multitubular reactor with these defined catalyst bed distributions is challenging, since it consists of thousands of small-diameter tubes that need to be filled precisely with the right amount of catalyst(s). 3.2. Adiabatic Reactor Cascade. In adiabatic systems, the heat of reaction drives the reactant conversion, thus, in contrast 9072

dx.doi.org/10.1021/ie4036675 | Ind. Eng. Chem. Res. 2014, 53, 9067−9075

Industrial & Engineering Chemistry Research

Article

Figure 7. Axial coordinate versus temperature (top graphs), HBr conversion (middle graphs), and the ratio of O2:HBr in the gas stream (bottom graphs) of an adiabatic reactor cascade with RuO2/TiO2 (a), CeO2/ZrO2 (b), and staged beds (c). The maximum operation temperatures for the RuO2/TiO2 and CeO2/ZrO2 catalysts are indicated (dashed lines). The catalyst amount per stage and the amount of HBr fed are provided in Table 3. The combination of RuO2/TiO2 and CeO2/ZrO2 in (c) reduces the number of unit operations and the catalyst costs while attaining full HBr conversion. Furthermore, the overoxidation of RuO2 to volatile RuO4 is prevented by avoiding strong fluctuations of the O2:HBr ratio and high temperatures in the RuO2/TiO2 beds.

oxidation.31 For HBr oxidation, RuO2/TiO2 and CeO2/ZrO2 were simulated in adiabatic cascades using six reactors each (Figure 7a, b). For this design, the conditions have been adopted from the patent literature dealing with HCl oxidation.19 The HBr feed and the catalyst weight per stage are listed in Table 3. The inlet temperature and the HBr feeds in between the stages were optimized, so that the catalysts can perform in the optimal temperature range and safely below their maximum temperatures. Consequently, the amount of catalyst was reduced for the RuO2/TiO2 and CeO2/ZrO2 cascade by 37% and 26%, respectively, compared to the HBr oxidation in a single reactor process (Figure 6). However, the O2:HBr ratio increases extremely toward the outlet of each reactor due to the full consumption of HBr (Figure 7a, b). This is especially critical in combination with the high outlet temperatures, promoting the undesired oxidation of RuO2 to RuO4 (Figure 7a). The combination of RuO2/TiO2 in the cold and CeO2/ZrO2 in the hot zone of the reactor to utilize their individual advantages could circumvent this problem. In fact, a similar approach has been proposed for HCl oxidation (Figure 1b).32 The HBr is first converted over a RuO2/TiO2 bed until reaching a temperature high enough to use the CeO2/ZrO2 catalyst. The reaction then

Table 3. Amount of HBr Feed and Catalyst Weight Per Stage in the Adiabatic Reactor Cascades RuO2/TiO2 reactor cascade

CeO2/ZrO2 reactor cascade

staged-bed reactor cascadea

position

HBr/ m3 STP h−1

Wcat/ kg

HBr/ m3 STP h−1

Wcat/ kg

HBr/ m3 STP h−1

Wcat/ kg

stage 1

6.3

2.4

6.3

5.0

10.8

stage 2

4.5

1.5

4.5

1.9

8.2

stage 3

4.0

1.5

4.0

1.9

9.3

stage 4

4.2

1.7

4.2

2.0

-

stage 5

4.9

1.5

4.9

2.0

-

stage 6

4.4

1.5

4.4

2.0

-

1.4 1.9 1.0 2.9 1.2 4.5 -

a

The top value in each stage refers to the weight of RuO2/TiO2 and the bottom value to CeO2/ZrO2.

progresses until the maximum temperature of CeO2/ZrO2 is reached (Figure 7c). Hence, the staged bed can utilize a larger 9073

dx.doi.org/10.1021/ie4036675 | Ind. Eng. Chem. Res. 2014, 53, 9067−9075

Industrial & Engineering Chemistry Research

Article

cp,i = heat capacity of component i, J mol−1 K−1 D = reactor tube diameter, m Ea = activation energy, kJ mol−1 Fi = molar flow of component i, mol STP h−1 k = forward reaction rate constant, mol Br2 h−1 kgcat−1 bar−1.5 k0 = pre-exponential factor, mol Br2 h−1 kgcat−1 bar−1.5 P = total pressure, bar Q = total volumetric flow, m3 STP h−1 U = overall heat transfer coefficient, W m−2 K−1 R = universal gas constant, 8.3145 J mol−1 K−1 rw = specific reaction rate, mol Br2 h−1 kgcat−1 T = temperature, K Wcat = weight of the catalyst bed, kg yi = molar fraction of component i, -

temperature rise compared to the single catalyst beds, thus more HBr can be converted per stage, which considerably reduces the overall number of operation units from six to three. The larger HBr inlet feed further lowers the O2:HBr ratio and causes less fluctuations along the cascade (Figure 7c). When calculating the costs of the presented results for the adiabatic cascade (Figure 7c) with the cooled reactor process (Figure 4c), it was found that the costs for the precious metal are 2.75 times lower for the former. This is because the average temperature of the cooled reactor process is kept far away from the maximum limit following the sensitivity analysis (Figure 5), while in the case of the adiabatic cascade reactor the temperature rise can be utilized much better due to the optimal control of the reactor operation as mentioned before. Despite the various simplifications considered in the simulations, novel and valuable insights into the design of halogen production processes are provided. Still, future investigation have to consider more rigorous models taking into account effects like product inhibition and radial temperature gradients, which could not be considered at this stage without further data from pilot tests. This will allow to fully assess the economic and technical aspects and to thoroughly design a feasible process for production bromine.

Greek

ΔH0r = reaction enthalpy at STP conditions, kJ mol−1 ΔHr = reaction enthalpy, kJ mol−1 ΔT = temperature difference, K ρcat = catalyst density, kg m−3 υi = stoichiometric coefficient of component i, -

Abbreviations

4. CONCLUSIONS This study evaluated catalyst distribution strategies to design feasible reactor configurations for the highly exothermic HBr oxidation to Br2. Our results show that cooled processes using diluted beds of RuO2/TiO2, CeO2/ZrO2, or combinations of these do not exhibit dangerous hot spot formation while attaining a complete recovery of bromine. The decreased sensitivity on temperature provides enhanced safety with respect to thermal runaways and optimal catalyst utilization for the production of bromine in a cooled process. The staging of RuO2/ TiO2 and CeO2/ZrO2 catalysts in an adiabatic reactor cascade is less complex and leads to a very efficient design, since CeO2/ ZrO2 can utilize the high temperatures at the end of the RuO2/ TiO2 bed, thus reducing the overall number of unit operations and catalyst cost, while yielding full HBr conversion. However, the volume fraction of HBr has to be kept low in both, cooled and adiabatic reactors, to avoid a reaction runaway. Further, a thorough optimization of the operating conditions based on rigorous simulations and pilot plant data is necessary, in order to progress and design technically feasible HBr oxidation processes. All the concepts shown and quantified in the course of the manuscript for the oxidation of HBr are applicable to the oxidation of HCl, which is certainly less demanding due to its reduced exothermicity.





REFERENCES

(1) Pérez-Ramírez, J.; Mondelli, C.; Schmidt, T.; Schlüter, O. F.-K.; Wolf, A.; Mleczko, L.; Dreier, T. Sustainable chlorine recycling via catalysed HCl oxidation: from fundamentals to implementation. Energy Environ. Sci. 2011, 4, 4786. (2) Seki, K. Development of RuO2/Rutile-TiO2 catalyst for industrial HCl oxidation process. Catal. Surv. Asia 2010, 14, 168. (3) Amrute, A. P.; Mondelli, C.; Schmidt, T.; Hauert, J.; PérezRamírez, J. Industrial RuO2-based Deacon catalysts: carrier stabilization and active phase content optimization. ChemCatChem 2013, 5, 748. (4) Teschner, D.; Farra, R.; Yao, L.; Schlögl, R.; Soerijanto, H.; Schomäcker, R.; Schmidt, T.; Szentmiklósi, L.; Amrute, A. P.; Mondelli, C.; Pérez-Ramírez, J.; Novell-Leruth, G.; López, N. An integrated approach to Deacon chemistry on RuO2-based catalysts. J. Catal. 2012, 285, 273. (5) Amrute, A. P.; Mondelli, C.; Moser, M.; Novell-Leruth, G.; López, N.; Rosenthal, D.; Farra, R.; Schuster, M. E.; Teschner, D.; Schmidt, T.; Pérez-Ramírez, J. Performance, structure, and mechanism of CeO2 in HCl oxidation to Cl2. J. Catal. 2012, 286, 287. (6) Moser, M.; Mondelli, C.; Schmidt, T.; Girgsdies, F.; Schuster, M. E.; Farra, R.; Szentmiklósi, L.; Teschner, D.; Pérez-Ramírez, J. Supported CeO2 catalysts in technical form for sustainable chlorine production. Appl. Catal., B 2013, 132−133, 123. (7) Moser, M.; Rodríguez-García, L.; Amrute, A. P.; Pérez-Ramírez, J. Catalytic bromine recovery: An enabling technology for emerging alkane functionalization processes. ChemCatChem 2013, 5, 3520. (8) Olah, G. A.; Gupta, B.; Farina, M.; Felberg, J. D.; Ip, W. M.; Husain, A.; Karpeles, R.; Lammertsma, K.; Melhotra, A. K.; Trivedi, N. J. Selective monohalogenation of methane over supported acid or platinum metal catalysts and hydrolysis of methyl halides over γalumina supported metal oxide/hydroxide catalysts. A feasible path for the oxidative conversion of methane into methyl alcohol/dimethyl ether. J. Am. Chem. Soc. 1985, 107, 7097. (9) Ding, K.; Metiu, H.; Stucky, G. D. Interplay between bromine and iodine in oxidative dehydrogenation. ChemCatChem 2013, 5, 1906. (10) Amrute, A. P.; Krumeich, F.; Mondelli, C.; Pérez-Ramírez, J. Depleted uranium catalysts for chlorine production. Chem. Sci. 2013, 4, 2209.

AUTHOR INFORMATION

Corresponding Author

*Fax: 41 44 633 1405. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■ ■

adi = adiabatic cat = catalyst HTS = heat transfer salt iso = isothermal max = maximum opt = optimal

DEDICATION Dedicated to Professor Massimo Morbidelli on the occasion of his 60th birthday. NOMENCLATURE

Symbols

a = activity factor, 9074

dx.doi.org/10.1021/ie4036675 | Ind. Eng. Chem. Res. 2014, 53, 9067−9075

Industrial & Engineering Chemistry Research

Article

(11) Johnson, A. J.; Cherniavsky, A. J. Chlorine production. US2542961, 1951. (12) Kiyoura, T.; Kogure, Y.; Nagayama, T.; Kanaya, K. Manufactoring process of chlorine. US4822589, 1989. (13) Itoh, H.; Kono, Y.; Ajioka, M.; Takenaka, S.; Kataita, M. Production process of chlorine. US4803065, 1989. (14) Pan, H. Y.; Minet, R. G.; Benson, S. W.; Tsotsis, T. T. Process of converting hydrogen chloride to chlorine. Ind. Eng. Chem. Res. 1994, 33, 2996. (15) Mortensen, M.; Minet, R. G.; Tsotsis, T. T.; Benson, S. The development of a dual fluidized-bed reactor system for the conversion of hydrogen chloride to chlorine. Chem. Eng. Sci. 1999, 54, 2131. (16) Wang, L.-W.; Han, M.-H.; Wu, Y.-L.; Wie, F.; Jin, Y. Conversion of hydrogen chloride to chlorine by catalytic oxidation. Chin. J. Process Eng. 2003, 3, 340. (17) Hibi, T.; Nishida, H.; Abekawa, H. Process for producing chlorine. US5871707, 1999. (18) Iwanaga, K.; Seki, K.; Hibi, T.; Issoh, K.; Suzuta, T.; Mori, Y.; Abe, T. The development of improved hydrogen chloride oxidation process. Sumitomo Kagaku 2004, 1. (19) Wolf, A.; Mleczko, L.; Schubert, S.; Schluter, O. F.-K. Processes and apparatus for the production of chlorine by gas-phase oxidation. US2007/0274901 A1, 2007. (20) Farra, R.; Eichelbaum, M.; Schlögl, R.; Szentmiklósi, L.; Schmidt, T.; Amrute, A. P.; Mondelli, C.; Pérez-Ramírez, J.; Teschner, D. Do observations on surface coverage-reactivity correlations always describe the true catalytic process? A case study on ceria. J. Catal. 2013, 297, 119. (21) Nishida, J.; Mori, Y. Method for producing chlorine using fixed bed reactor. US2013/0039842 A1, 2013. (22) NIST Chemistry WebBook, NIST Standard Reference Database Number 69, U.S. Department of Commerce, National Institute of Standards and Technology: USA, 2013. http://webbook.nist.gov/ chemistry (accessed 31st Oct. 4.00 p.m. GMT). (23) Calverley, E. M.; Witt, P. M.; Sweeney, J. D. Reactor runaway due to statistically driven axial activity variations in graded catalyst beds. Chem. Eng. Sci. 2012, 80, 393. (24) Hibi, T.; Abekawa, H.; Seki, K.; Suzuki, T.; Suzuta, T.; Iwanaga, S. Production of chlorine. JP2000281314, 2000. (25) Sadhukhan, P.; Petersen, E. E. Oxidation of naphthalene in packed-bed reactor with catalyst activity profile: A design scheme for improved reactor stability and higher product yield. AIChE J. 1976, 22, 808. (26) Papageorgiou, J. N.; Froment, G. F. Phthalic anhydride synthesis. Reactor optimization aspects. Chem. Eng. Sci. 1996, 51, 2091. (27) Deller, K.; Krause, H.; Schmidhammer, L.; Dafinger, W. Cylindrically formed catalyst for the oxychlorination of ethylene. US5166120, 1992. (28) Adler, J.; Enig, J. W. The critical conditions in thermal explosion theory with reactant consumption. Combust. Flame 1964, 8, 97. (29) Mineral Prices - The Global Source, Mineral Fund Advisory Pty Ltd.: Australia, 2013. http://www.mineralprices.com (accessed 30th Sept. 6:00 p.m. GMT). (30) Appl, M. Ammonia. Ullmann’s Encyclopedia of Industrial Chemistry; Wiley-VCH: New York, 2006. (31) Nodehi, A.; Mousavian, M. A. Simulation and optimization of an adiabatic multi-bed catalytic reactor for the oxidation of SO2. Chem. Eng. Technol. 2006, 30, 84. (32) Schmidt, T.; Moser, M.; Müller, W. Catalyst and method for producing chlorine by means of a gas-phase oxidation. WO2013060628, 2013.

9075

dx.doi.org/10.1021/ie4036675 | Ind. Eng. Chem. Res. 2014, 53, 9067−9075