- Email: [email protected]
Effect of active thermal insulation on methane and carbon dioxide concentrations in the effluent of a catalytic partial oxidation reactor for natural gas conversion to synthesis gas
Chemical Engineering Journal 281 (2015) 852–859
Contents lists available at ScienceDirect
Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej
Effect of active thermal insulation on methane and carbon dioxide concentrations in the effluent of a catalytic partial oxidation reactor for natural gas conversion to synthesis gas A. Al-Musa a,⇑, S. Shabunya b, V. Martynenko b, V. Kalinin b a National Center for Combustion and Plasma Technologies, Water and Energy Research Institute, King Abdulaziz City for Science and Technology (KACST), P. O. Box 6086, Riyadh 11442, Saudi Arabia b State Scientific Institution ‘‘A.V. Luikov Heat and Mass Transfer Institute’’, National Academy of Sciences of Belarus, 15 P. Brovki, Minsk 220072, Belarus
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Local thermodynamic equilibrium is
Ni-commercial catalyst
established in CPOX pilot scale reactors. CPOX products composition is controlled by heat transfer processes in catalyst beds. Active thermal insulation has a positive effect on conversion. Preheating up to 700 °C allows for reduction in CO2 content to 0.4 vol.%. Catalyst deterioration is observed at bottom–top gas filtration through reactor.
a r t i c l e
i n f o
Article history: Received 15 March 2015 Received in revised form 18 June 2015 Accepted 11 July 2015 Available online 17 July 2015 Keywords: Methane Natural gas Catalytic partial oxidation Syngas Isothermal equilibrium Active thermal insulation
Inlet mixture Natural gas + air H2 (T) CO(T) CH4 (T) CO2(T) H2O(T)
T
Thermocouples
Inlet mixture T
Natural gas + air
H2 (T) CO(T) C H4 (T) C O2(T) H2O(T)
Outlet mixture H2; CO; CH4; CO2; H2O
a b s t r a c t Two reactor designs with either passive or active thermal insulation were tested to produce synthesis gas from natural gas using catalytic partial oxidation on a system that exceeded laboratory scale. Preheating the working mixture significantly improves the conversion parameters. Preheating levels are limited by two factors: mixture ignition in the heater and overheating at the catalyst bed inlet area. The use of hot conversion products to provide active thermal insulation significantly improves the quality of the conversion. Because the same amount of identical catalyst was used in both reactors, the observed effect supports the hypothesis that the product composition depends primarily on the heat transfer process. Ó 2015 Elsevier B.V. All rights reserved.
1. Introduction Efforts to more efficiently use natural gas reserves have raised interest in synthesis gas (syngas) production, which is currently used for a broad number of applications [1] including modern
⇑ Corresponding author. Tel.: +966 1 481 4316; fax: +966 1 481 3880. E-mail address: [email protected] (A. Al-Musa). http://dx.doi.org/10.1016/j.cej.2015.07.041 1385-8947/Ó 2015 Elsevier B.V. All rights reserved.
Outlet mixture H2; CO; CH4; CO2;H2O
gas-to-product technologies [2] and fuel cells [3]. Steam reforming is frequently used for syngas and hydrogen production; however, this process is not ideal in terms of time or efficiency. The catalytic partial oxidation (CPOX) process is an alternative to steam reforming [4]. CPOX is more energy efficient [5]. Comparative thermodynamic analyses have confirmed that partial oxidation process has advantage over steam reforming process for syngas production [6]. Highly endothermic steam reforming processes require
A. Al-Musa et al. / Chemical Engineering Journal 281 (2015) 852–859
external heating [7] with intensive heat delivery. Steam reforming process requires 2.8 times higher energy supply than partial oxidation process to convert one mole of methane to synthesis gas at 1000 °C. In the face of rising natural gas prices, the partial oxidation process is becoming more economically attractive, especially with small and medium production capacities [6,8]. Numerous studies have described and analyzed CPOX of hydrocarbons. These studies primarily focused on methane, which is the principal constituent of natural gas. Recent reviews have covered the major findings in the field and refer to key publications on theoretical and experimental CPOX processes, catalysts, and reaction kinetics [4,9]. An overwhelming majority of these publications involved the investigation of fundamental CPOX processes and were aimed at clarifying the catalysis kinetics, developing effective catalyst compositions (active agent/promoter/carrier), studying catalyst degradation mechanisms, and increasing catalyst service life. Typically, kinetic methane CPOX studies are investigated experimentally with a laboratory-scale reactor [10–16]. It has been established that, noble and transition metals are feasible CPOX catalysts, meaning that in laboratory conditions, the thermodynamic equilibrium composition is reached for 10–50 ms. These characteristic residence times mean that the chemical reaction rate is ‘‘very high’’. At the same time, active research on laboratory scale operating in kinetic mode are conducted where the selectivities of the various components of the produced synthesis gas depend on choice of residence time [17]. Such effect is particularly observed in the membrane reactors employed to remove hydrogen efficiently from the reaction volume [18]. High flow rates are not required to exclusively study scientific aspects of catalytic processes. In laboratory studies typical flow rates are in the range of 50–300 sccm [18]. implementation of scientific concepts developed using laboratory to scale up processes to industrial plants usually causes a number of problems that do not occur and are not discussed in case of small scale installations. Despite some advantages of reactors operating in kinetic modes all industrial partial oxidation plants are exploited at equilibrium conditions. The design of an industrial-scale plant requires optimization of the heat and mass transfer processes rather than identifying effective catalysts. The most feasible catalyst is primarily chosen based on its cost and durability at approximately 1000 °C. To scale up the methane CPOX process for syngas production from laboratory scale to industrial scale, a number of engineering problems have to be solved, as the heat and mass transfer coefficients vary substantially. The temperature and concentration distributions within the reactor are formed by heat and mass transfer processes interactions. These phenomena cannot be studied properly using a laboratory-scale reactor. A few publications have been devoted to elaborating on industrial-scale reactor concepts [19,20]. The purpose of this article is to demonstrate that when developing industrial scale CPOX reactors, the greatest impact is obtained from heat transfer rather than from catalyst ‘‘efficiency’’. This study expands on the previously published work [21], and testing of the two partial oxidation reactor designs is described. The methane-air mixture CPOX experiments described previously [21] were conducted with a non-preheated inlet mixture with a wide range of CH4/air equivalence ratios. The low process temperatures [21] were due to the lack of an external energy supply. These regimes were characterized by minimal heat loss in contrast to hotter regimes, but even in these conditions, the product composition appears to be worse than with adiabatic equilibrium.
853
The four different catalysts investigated showed no significant differences among their degrees of approximation to the equilibrium composition. This fact led to the conclusion that the deviation in product composition with adiabatic equilibrium is associated with heat loss rather than the lack of residence time. In the case of an inlet mixture with a C:O ratio of 1, the obtained CPOX product composition at the reactor outlet was approximately 28% H2, 13% CO, 7% CH4, and 4% CO2 (volume percents correspond to the drained mixture), whereas the adiabatic calculation gives 32.8% H2, 15.2% CO, 5.5% CH4, and 2.6% CO2. A simple model assuming local thermodynamic equilibrium showed that the observed deviations could be explained by a non-homogeneous radial temperature profile in the catalyst bed that formed due to heat loss. The calculations predicted a temperature difference of approximately 20 °C between the reactor axis and the cylindrical catalyst surface. These calculations were based on the assumption that equilibrium was achieved, meaning that ‘‘sufficient’’ residence time was achieved for the working mixture in the catalyst bed. Although the amount of catalyst used and the flow rates of the working mixture correspond to the CPOX residence times described by Korchnak et al. [22], the issues regarding the ‘‘adequacy’’ of the catalyst sample residence time tested by Al-Musa et al. [21] and the actual value of the radial temperature difference remain disputable. The residence time required to reach equilibrium is a key characteristic of the catalyst in CPOX applications. This characteristic depends on the temperature of the fed gas [23], the catalyst grain size, the effective surface area, and the flow conditions. A theoretical study on the effect of heat transfer and particle size on CH4 partial oxidation was published previously [24]. Periodic temperature perturbations were studied, but the effects of filtration field behavior changes were not considered. In the presence of radial temperature gradients, clearly identifying the impact of residence time on the degree of the approach to equilibrium based on the composition of the outlet mixture is difficult, because the temperature profile affects the chemical process rate and the flow field in the catalyst bed. Studies of mass transfer intensity (i.e., the necessary residence time) should be carried out in a reactor operating at nearly isothermal conditions. Our aim was to improve the composition of CPOX products using preheating of the inlet mixture and by implementing a novel reactor design to decrease heat loss and therefore lower CH4 and CO2 content. 2. Materials and methods 2.1. Experimental setup of passive thermal insulation The reactor dimensions and performance have been described in detail previously [21] and are shown in Fig. 1. In the present study, the potential for increasing the conversion depth by preheating the reaction mixture in this reactor was investigated, and the temperature field in the catalytic bed was estimated. For these experiments, 12 thermocouples were installed to record the temperatures. Thermocouples 1, 3, 4, 5, 7, 9, and 11 were located along the reactor axis, whereas thermocouples 2, 6, 8, 10, and 12 were displaced from the reactor axis towards the edge of the thermal insulation (Fig. 1). Ni=c Al2 O3 catalyst produced by Liaoning Haitai Sci-Tech Development Co., Ltd., China, was used. This granular catalyst with 16 wt% Ni content has cylindrical shape with dimensions, 3.5 mm in length and 1.0 mm in diameter. BET surface area is 91.5 m2/g. The volume of the catalytic bed used in experiments is about 327 cm3.
854
A. Al-Musa et al. / Chemical Engineering Journal 281 (2015) 852–859
Fig. 1. Scheme of the passive thermal insulation experimental reactor.
Wang and Tian also reported high CH4 conversion (greater than 90%) on a Ni=a Al2 O3 catalyst [26]. When Ni was supported on a calcium aluminate carrier [27], the most efficient methane conversion was 85%. The experiments were performed using natural gas (NG) rather than pure methane to better approximate the real industrial situation. The composition of NG in volumetric percents is as follows: CH4 = 97.6%, C2H6 = 1.06%, C3H8 = 0.25%, C4H10 = 0.08%, C5H12 = 0.005%, N2 = 1.005%. Because CPOX reactors create local thermodynamic equilibrium, the composition of the final product depends only on the ratio of atomic C/O/H and the enthalpy of the inlet mixture. As NG contains almost 98% methane, the energy and equilibrium compositions between experiments using pure methane and NG differ little. The actual composition of NG was taken into account for the thermodynamic calculations. All experimental data were obtained using a prepared NG-air mixture with a C/O ratio 1. The composition of the CPOX product mixture was determined after the outlet mixture was drained. The CH4, CO, and CO2 concentrations were measured using optical sensors Gascard II produced by Edinburgh Sensors Ltd. The range of measurements of CH4 and CO is between 0 and 30 vol.%., while for CO2 this value is between 0 and 5 vol.%. H2 concentration was measured within range from 0 to 100 vol.% using katharometer
K1550 produced by Hitech Instruments Ltd. The katharometer and the optical sensors were calibrated using chromatographic data measurements obtained by 7890A GC System from Agilent Technologies. Comparative analysis of the calibration data and the previous measurements demonstrated that the H2 concentration was systematically underestimated by approximately 1.5% [21]. Once tuned, the calibration error of the optical sensors was comparable to the gas chromatograph measurement errors with the exception of CH4 concentrations lower than 0.6%, which approached the sensor’s sensitivity threshold. During the experiment, the compositions of the inlet and outlet mixtures that were measured were compared with gas chromatograph measurements. Optical sensors and katharometers allowed controlled measurements at a sufficiently high frequency for technical applications of no less than once per second. Air and NG flow controllers, mixing devices, and mixture heaters also contributed some error, even when kept at a constant setting. As a result, concentration fluctuations that were caused by system component design rather than instrumentation errors were observed at the reactor outlet even when the device was operating at ‘‘steady state’’. The average sensor reading values are presented here. Gas chromatography data can be considered instantaneous, i.e., containing a fluctuating component. In our studies, we focused on relative performance, i.e., we compared the compositions of the outlet mixture obtained from different initial conditions and tested different reactor designs. In all experiments, use of the same methods and measurement system allows us to make that assessment even using technical precision (5–10%) for the concentration measurements. An electric heater was used to preheat the working mixture. The inlet temperature of the working mixture varied between 200 and 600 °C. The upper heating limit was set to prevent ignition of the mixture in the heater or in the gas supply path prior to contact with the catalyst bed. Mixture ignition is undesirable, not only from the standpoint of an installation structural failure, but also because an increased CO2 content in the outlet mixture results. The critical heating value is dictated by the working mixture properties and heater design. To avoid combustion in the heater, the velocity of the mixture flow must be several times higher than the velocity of the mixture burning at the operating temperature. The use of an intermediate heat carrier (liquid lead or tin) to control the maximum temperature (Tmax) allows us to choose the heater channel flow cross-section and determine the acceptable flow rate based on Tmax and the burning rate of the working mixture. The channel design should exclude stagnant zones where localized burning could occur due to slow fuel and combustion product exchange within the main flow zone.
2.2. Experimental setup of the active thermal insulation reactor To reduce thermal loss, the experimental reactor design was modified to thermally protect the reactor using hot CPOX products. The modified design schematic is shown in Fig. 2. The reactor vessel was designed with an additional protective stainless steel shell engineered as ‘‘a tube within a tube’’. In this design, the hot CPOX products leaving the catalyst bed flow in the annular space between the reactor wall and the shell. Because the temperature of the CPOX product mixture was near the adiabatic temperature, the reaction zone thermal loss should be substantially lower than in the reactor with passive thermal insulation. The reactor was fabricated to test potential design advantages and disadvantages rather than to conduct systematic temperature studies. In these experiments, the reactor temperature was measured at two points: the inlet to measure the heating temperature of the working mixture and the catalytic bed outlet to measure the temperature of the CPOX product.
A. Al-Musa et al. / Chemical Engineering Journal 281 (2015) 852–859
855
Fig. 4. Temperature difference between axial and outer catalyst layer temperatures versus the inlet temperature (d – G = 4.25 Nm3/h, j – G = 2.54 Nm3/h).
Fig. 2. Scheme of the active thermal insulation experimental reactor.
3. Results and discussion 3.1. Effects of preheating the working mixture using a passive thermal insulation reactor Two series of experiments were performed using a preheated working mixture with C/O = 1 and flow rates of 2.54 and 4.25 Nm3/h. Temperature characteristics of the catalyst units are shown in Figs. 3 and 4. The concentrations of the outlet mixture components versus initial temperature are plotted in Fig. 5. The experimental measurements are shown by labels, while dashed lines present the results of adiabatic calculations. The pure, unheated methane data [21] obtained at a flow rate (G) = 3.2 Nm3/h are shown as well.
Fig. 3. On axis temperature (d – G = 4.25 Nm3/h, j – G = 2.54 Nm3/h, e – G = 3.2 Nm3/h) and outer catalyst layer temperature (s – G = 4.25 Nm3/h, h – G = 2.54 Nm3/h) versus the inlet temperature. Dashed curve is the adiabatic estimation.
The temperature difference was higher using a slower flow rate. This is associated with an effective increase in thermal conductivity when the filtration velocity increases (dispersion mechanism [25]). As expected, the temperature at the insulated catalyst bed boundary was lower than the calculated adiabatic temperature. This difference increased with increased heating. The apparent difference between the axial and calculated adiabatic temperatures was somewhat unexpected. This difference may be attributed to incomplete partial oxidation reactions even near the reactor axis, where heat loss effects were minimal. Therefore, the residence time is not sufficient at these flow rates. At a low flow rate, the temperature near the reactor axis was somewhat higher, which may be the result of lower heat intensity and mass transfer processes at lower filtration rates. In this case, disruption of the axial symmetry of the thermal pattern and filtration may influence and consequently distort the temperature profile. As expected, heating the reaction mixture at the reactor inlet led to increased H2 and CO concentrations in the CPOX products, whereas the methane and carbon dioxide concentrations decreased. This indicated improved conversion relative to the ideal partial oxidation formula: 2CH4 + O2 ? 2CO + 4H2. When comparing the outlet composition with the adiabatic equilibrium composition, the difference was clearly more pronounced with more intensive heating, which may be explained by increased thermal loss during the transition to the higher reactor temperatures, causing a decrease in residence time. The composition of the conversion product did not depend solely on the flow rate. When the catalyst amount was fixed, an increase in the flow rate should reduce the residence time, impairing conversion. On the other hand, the flow of energy through the section of the reactor is proportional to the flow rate of the working mixture, while the heat loss depends only on the temperature of the reactor shell. The increase in flow rate should have a positive effect on the quality of conversion, because it reduces the percentage of heat loss, and the output mixture should have a higher temperature. this would have been the case, if the temperature of the shell remained unchanged. However, there are the effects of increasing the temperature of the shell. The increase in the rate of filtration extends the scope of the exothermic process where the temperature is higher, and intensifies the heat transfer in a porous bed, i.e. it facilitates heat transfer to the shell. As a result, the average temperature of the reactor shell increases, and heat loss increases, too. The heat flux in the shell to the flanges of the reactor has a negative effect. When the negative and positive factors are close to each other, increasing the flow will not improve conversion. In experiments at 4.25 and 2.54 Nm3/h flow rates, these factors almost cancel each other quantitatively. As shown in Fig. 5, the measured conversion product compositions were similar (with a slight advantage) to those obtained using 4.25 Nm3/h flow rate.
856
A. Al-Musa et al. / Chemical Engineering Journal 281 (2015) 852–859
Fig. 5. Effect of inlet temperature and flow rate on CPOX product composition. d – G = 4.25 Nm3/h, j – G = 2.54 Nm3/h, e – G = 3.2 Nm3/h (pure CH4 test [21]), dashed curve – adiabatic estimation.
We increased the flow rate of the working mixture in an attempt to produce an adiabatic-like composition; however, it did not produce the expected results. Doubling the catalyst volume to increase the residence time led to increased thermal loss from the catalytic bed and a decreased outlet temperature. Thus, in experiments with a passive thermally insulated reactor, preheating the inlet mixture (C/O = 1) decreased the CH4 concentrations of the CPOX products from 7% to 3% and CO2 from 3.8% to 1%. The effects of temperature and residence time on the extent of conversion clearly could not be separated in the presence of thermal loss. 3.2. Effect of active thermal insulation on natural gas conversion The composition of the CPOX product was measured in several experiments at different flow rates and heater temperatures (Tables 1 and 2). All experiments were conducted with an NG-air working mixture of C/O 1. Experimental data were sorted by heating temperatures of the inlet mixture and flow rates, and the values corresponding to the adiabatic equilibrium are given in parentheses. Temperature gradients were not measured in the second reactor design; however, given the improvement in the composition of the outlet mixture, we concluded that they were smaller. The two different reactors were compared based on the outlet mixture quality achieved when heating the mixture to 600 °C at a flow rate of 2.52 Nm3/h (second row of Table 1). The composition was 37.3% H2, 19.7% CO, 1% CH4, and 0.5% CO2, which was substantially better than the corresponding values (34% H2, 18% CO, 3.5% CH4, and 1.2% CO2) obtained in experiment using the first reactor. Table 1 contains the results of six independent experiments. Over the course of experimentation, a slight design modification of the reactor was made (changing gaskets, thermal insulation, catalyst bed fixing), accompanied by
a partial replacement of the catalyst material. However, no special attention was paid to catalyst aging during this time period. When analyzing the data collected (Table 1), the results obtained using the same or similar parameters of the inlet mixture were scattered. We believe this was primarily the result of catalyst aging. In the passive insulation reactor experiments (Fig. 1), catalyst-aging effects were not observed. The lack of effect can be explained as follows: 1. In the passive insulation reactor design (Fig. 1), the direction of gas flow and gravity effects coincided, i.e., the catalyst particles were packed downwards and did not contribute to mechanical grain movements. In the case of the active insulation reactor design (Fig. 2), the situation was reversed, i.e., gas flowed against gravity, lifting the particles. This caused interparticle friction and produced mechanical failure. Attempts to lock the layer by mechanically pressing the upper layer with porous ceramic plates during assembly of the reactor catalyst unit did not solve this problem. Additionally, the intensive heating resulted in the formation of free space inside the reactor that increased with the operating time. Every time the installation was cooled and the reactor was disassembled, formation of free space was observed. Not surprisingly, higher flow rates intensified the mechanical destruction. 2. The experiments performed using the active insulation reactor were conducted at higher flow rates for the working mixture in an attempt to produce adiabatic-like experimental conditions. The increased flow rate accelerated all the various degradation processes including thermal, mechanical, and chemical processes. 3. Improving the adiabatic properties of the experimental unit resulted in higher temperatures in the catalyst bed, which is particularly critical for the exothermic reaction zone (hot
857
A. Al-Musa et al. / Chemical Engineering Journal 281 (2015) 852–859 Table 1 Compositions and characteristics of CPOX products produced from the active thermal insulation reactor. P, bar
C/O
Theater, C
T, C
1.10 1.18 1.25 1.50 1.50 1.72 2.00 2.11 2.14 1.30 1.31 1.41 1.55 1.73 1.82 1.94 2.18 1.95 1.34 1.67
1.0 1.0 1.02 1.02 1.0 1.02 1.0 1.02 1.02 1.0 1.02 1.02 1.02 1.02 1.0 1.02 1.02 1.0 1.0 1.0
600 600 600 600 600 600 600 600 650 700 700 700 700 700 700 700 700 705 705 720
705 850 805 835 905 850 920 870 880 930 815 835 855 865 705 875 895 845 790 870
(819) (822) (815) (821) (831) (827) (842) (835) (861) (885) (875) (877) (880) (883) (895) (886) (890) (900) (890) (906)
Gtot, Nm3/h
H2
CO
CH4
1.68 2.52 3.30 5.10 5.10 6.60 8.45 9.00 9.00 3.40 3.65 4.43 5.30 6.40 6.73 7.72 9.00 8.20 4.10 5.85
36.4 (39.2) 37.3 (39.2) 35.8 (39.1) 35.8 (38.9) 36.9 (39.0) 35.8 (38.8) 35.9 (38.7) 35.7 (38.6) 36.3 (39.1) 38.2 (39.8) 36.85 (39.8) 37.0 (39.8) 37.0 (39.7) 37.2 (39.6) 37.8 (39.6) 37.2 (39.6) 37.1 (39.5) 37.4 (39.6) 38.6 (39.8) 38.0 (39.8)
18.0 (19.8) 19.7 (19.8) 18.8 (19.7) 18.8 (19.7) 19.5 (19.7) 18.8 (19.6) 19.5 (19.6) 18.8 (19.5) 19. (19.8) 19.9 (20.1) 19.3 (20.1) 19.4 (20.1) 19.5 (20.1) 19.5 (20.0) 19.23 (20.0) 19.4 (20.0) 19.3 (20.0) 19.2 (20.0) 19.89 (20.1) 19.3 (20.1)
2.45 1.00 1.68 2.00 1.60 2.50 2.30 2.60 2.12 0.90 1.42 1.40 1.45 1.58 1.04 1.70 1.65 2.40 1.74 1.00
Gtot, Nm3/h
H2
CO
3.40 6.90 6.90 8.50 8.50
40.0 39.6 39.8 39.5 38.3
CO2 (0.95) (0.99) (1.43) (1.53) (1.11) (1.61) (1.27) (1.73) (1.43) (0.57) (0.97) (0.99) (1.03) (1.08) (0.7) (1.13) (1.18) (0.71) (0.57) (0.59)
1.45 0.46 1.07 0.97 0.50 0.93 0.60 0.93 0.83 0.42 0.78 0.75 0.73 0.70 0.69 0.70 0.70 0.54 0.42 0.65
(0.29) (0.30) (0.28) (0.31) (0.33) (0.33) (0.38) (0.36) (0.25) (0.14) (0.11) (0.11) (0.12) (0.14) (0.18) (0.15) (0.17) (0.18) (0.14) (0.14)
Table 2 Composition and characteristics of CPOX products using fresh catalyst. P, bar
C/O
Th, C
T, C
1.25 1.73 1.77 2.00 2.00
1.0 1.0 1.0 1.0 1.0
600 600 700 650 650
895 835 880 875 870
(824) (836) (894) (868) (868)
spots). The increased flow rate and decreased thermal loss may have raised the temperature in this area, thereby accelerating thermal degradation of the catalyst.
3.3. Catalyst degradation To test the hypothesis that the design of the active thermal insulation reactor increased catalyst degradation, an experiment was performed using fresh catalyst material (Table 2). Photographs of fresh catalyst and used catalyst following the experiments are shown in Fig. 6. Freshly prepared catalyst resulted in high-efficiency CPOX processes. The composition of the CPOX products was near equilibrium, and the residence time was
(39.1) (38.9) (39.6) (39.2) (39.2)
20.5 20.1 20.3 20.3 19.6
CH4 (19.8) (19.6) (20.0) (19.8) (19.8)
0.67 0.82 0.67 0.80 1.40
CO2 (1.01) (1.19) (0.69) (0.99) (0.99)
0.09 0.16 0.10 0.20 0.32
(0.31) (0.36) (0.17) (0.27) (0.27)
approximately 0.025 s at the maximum flow rate of 8.5 Nm3/h. Those data with concentrations measuring ‘‘better’’ than adiabatic were the result of composition, inlet mixture temperature, and errors in pressure measurement. The last two rows of Table 2 indicate catalyst degradation following 1 h of operation in a fixed working mode. This is clearly demonstrated by the difference in CPOX product concentrations using an identical working regime. The measured composition of the CPOX product was noticeably deteriorated. After reactor disassembly, we observed that the catalyst grains were much lighter and smaller in size (Fig. 6). Destruction of alumina-based Ni catalysts with CH4 partial oxidation was reported previously [28], whereas b SiC was proposed to be an
Fig. 6. Photographs of the catalyst samples. (a) Fresh catalyst prior to loading in the reactor and (b) used catalyst following experimentation using approximately 11 Nm3 working mixture passed through 0.25 L of catalyst.
858
A. Al-Musa et al. / Chemical Engineering Journal 281 (2015) 852–859
efficient Ni catalyst carrier that could successfully be used for syngas production from methane. 3.4. Feasibility of passive and active thermal insulation When using the passive thermal insulation CPOX reactor (Fig. 1), preheating the working mixture increased the degree of conversion, but attempts to produce syngas containing less than 0.5% CH4 and CO2 failed due to thermal loss. Preheating the inlet mixture up to 600 °C yielded CPOX products containing about 3% CH4 and 1% CO2. Sufficient superheating of the inlet mixture to compensate for thermal loss and to produce CPOX products with a desired composition is impossible due to the risk of ignition of the mixture in the heating element. A heater coil design with an intermediate heat carrier (liquid lead) allowed preheating of the NG-air mixture (C/O 1) up to 750–800 °C without ignition. However, these regimes lead to reactor overheating in the catalyst bed inlet area, which is dominated by an exothermic reaction. A stationary heating regime was not achieved, because the experiment automatically stopped when the temperature exceeded 1180 °C. Increasing the flow rate reduced thermal loss; however, no significant improvement in conversion was observed. The positive effect was diminished by reduced residence time and increased working pressure, causing an equilibrium shift to higher CO2 and CH4 concentrations. Increasing the catalyst volume only expanded the heat transfer surface, and the thermal loss increased. In the catalyst bed pipe, the effective temperature near the cylindrical surface was lower than the axial temperature due to thermal loss. As a result, most of the working gas flowed close to the cylindrical surface. Thus, a tubular reactor containing endothermic processes can lead to significant deviations in the composition of CPOX products from theoretical adiabatic equilibrium values. Although increasing the reactor diameter and reducing the catalyst bed height may reduce the cylindrical surface area (i.e., the surface of thermal loss), these options are limited due to homogenous filtration flow field requirements. In light of these considerations and the experimental data, we concluded that the passive heat insulation-type reactor design (Fig. 1) is problematic for producing syngas with CO2 and CH4 contents less than 0.5%. Optimizing the heating, insulation, and catalytic unit size will certainly reduce the experimental 3% CH4 and 1% CO2 CPOX product values. However, an optimized reactor would need to be designed for a specific, constant flow rate, i.e., the installation performance cannot be varied without affecting the composition of the CPOX products. Of note, higher hydrocarbon use would limit preheating, as these molecules are less stable and more easily oxidized than CH4. The active thermal insulation reactor design using hot CPOX products to reduce thermal loss (Fig. 2) reduced the radial temperature gradient in the catalyst bed and substantially improved hydrocarbon conversion in the CPOX process. When the inlet mixture was preheated to 650–700 °C, the CPOX products had a composition that was near equilibrium and a resident time of approximately 0.025 s. The experimental results produced 0.7% CH4 and 0.2% CO2 concentrations. Those values could be improved by lowering the working pressure in the CPOX reactor. In addition to the benefits of active thermal insulation, other notable limitations of the reactor design include: 1. The adiabatic-like reactor performance is higher at higher flow rates. Therefore, decreasing the flow rate may lead to deterioration of the composition of CPOX products. 2. Fabrication of an active thermal insulation reactor (Fig. 2) requires addressing a number of technical issues to ensure efficient operation of the CPOX reactor at 1000 °C. The inner shell
of the CPOX reactor can reach temperatures of 1200 °C or higher in exothermic reaction zones, depending on the properties of the catalyst. 3. The introduction of radial thermocouples through the two shells made the reactor more vulnerable to leakage. Therefore, this reactor design is not suitable for a systematic study of the temperature characteristics of the catalyst bed. 4. Conclusion Implementing thermal protection using reverse flow of the hot product gas does not allow for the adequate control of the reactor surface temperature because the system self-adjusts. More effective and practical thermal protection may be accomplished with controlled heating of the reactor shell, which would allow the external surface temperature of the reactor to be controlled along its entire length and adjusted for different performance regimes. This could potentially maintain an isothermal radial temperature profile to obtain a CPOX product composition close to theoretical adiabatic equilibrium estimations. Notably, heating is aimed at compensating for thermal loss, and the inlet mixture heater constitutes the primary energy input required to achieve the desired product temperature. Our results obtained using Ni-based catalysts demonstrated that this catalyst material has low stability when used in large-scale partial oxidation processes. The exact mechanism of catalyst degradation is unknown; however, if the degradation is a thermal process, another catalyst that is stable at working temperatures of 1000–1100 °C is needed. Acknowledgments Financial support from the King Abdulaziz City for Science and Technology (KACST) is gratefully acknowledged (Project No. 33-816: ‘‘Experimental and theoretical studies for chemical and thermal processes in partial oxidation reactors’’). In addition, we would like to thank our colleagues in KACST and HTMI for their valuable help. References [1] I. Wender, Fuel Process. Technol. 48 (1996) 189–297. [2] T.H. Fleisch, Syngas Chemistry Symposium, Dresden, Germany, October 4–6, 2006. [3] M. Liu, M.J.B.M. Pourquie, L. Fan, W. Halliop, et al., Fuel Cells 13 (3) (2013) 428– 440. [4] S.C. Tsang, J.B. Claridge, M.L.H. Green, Catal. Today 23 (1995) 3–15. [5] S.S. Bharadwaj, L.D. Schmidt, Fuel Process. Technol. 42 (1995) 109–127. [6] A.P.E. York, T. Xiao, M.L.H. Green, Top. Catal. 22 (2003) 345–358. [7] K.L. Hohn, L.D. Schmidt, Appl. Catal. A: General 211 (2001) 53–68. [8] V. Recupero, L. Pino, R. Di Leonardo, M. Lagana, G. Maggio, J. Power Source 71 (1998) 208–214. [9] R. Lanza, J.A. Velasco, S.G. Jaras, Catalysis 23 (2011) 50–95. [10] B.C. Enger, R. Lodeng, A. Holmen, Appl. Catal. A: General 346 (2008) 1–27. [11] M. Maestri, D.G. Vlachos, A. Beretta, G. Groppi, E. Tronconi, AIChE J. 55 (4) (2009) 993–1008. [12] A. Slagtern, H.M. Swaan, U. Olsbye, I.M. Dahl, C. Mirodatos, Catal. Today 46 (1998) 107–115. [13] D. Dalle Nogare, N.J. Degenstein, R. Horn, P. Canu, L.D. Schmidt, J. Catal. 258 (2008) 131–142. [14] R. Horn, K.A. Williams, N.J. Degenstein, A. Bitsch-Larsen, et al., J. Catal. 249 (2007) 380–393. [15] B.C. Enger, R. Lodeng, A. Holmen, Appl. Catal. A: General 364 (2009) 15–26. [16] U. Friedle, G. Veser, Chem. Eng. Sci. 54 (1999) 1325–1332. [17] J. Zaman, A. Chakma, J. Membr. Sci. 92 (1) (1994) 1–28. [18] Y.S. Cheng, M.A. Pena, K.L. Yeung, J. Taiwan Inst. Chem. Eng. 40 (3) (2009) 281– 288. [19] G. Kolios, J. Frauhammer, G. Eigenberger, Chem. Eng. Sci. 57 (2002) 1505– 1510. [20] B. Glockler, A. Gritsch, A. Morillo, G. Kolios, G. Eigenberger, Chem. Eng. Res. Des. 82 (2) (2004) 148–159. [21] A. Al-Musa, S. Shabunya, V. Martynenko, S. Al-Mayman, et al., J. Power Source 246 (2014) 473–481.
A. Al-Musa et al. / Chemical Engineering Journal 281 (2015) 852–859 [22] J.D. Korchnak, M. Dunster, A. English, Patent cooperation treaty WO90/06282, PCT/US89/05369, 1990. [23] S.-B. Tang, F.-L. Qiu, S.-J. Liu, J. Nat. Gas Chem. 5 (3) (1996) 272–277. [24] X. Ren, X. Guo, J. Nat. Gas 20 (5) (2011) 503–506. [25] M. Kaviany, Principles of Heat Transfer in Porous Media, Springer-Verlag, New York, 1991, 626p.
859
[26] H. Wang, Z. Li, S. Tian, J. Nat. Gas 12 (2003) 205–209. [27] M.A. Goula, A.A. Lemonidou, W. Griinert, M. Baerns, Catal. Today 32 (1996) 149–156. [28] P. Leroi, B. Madani, C. Pham-Huua, M.-J. Ledoux, et al., Catal. Today 91–92 (2004) 53–58.
Contents lists available at ScienceDirect
Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej
Effect of active thermal insulation on methane and carbon dioxide concentrations in the effluent of a catalytic partial oxidation reactor for natural gas conversion to synthesis gas A. Al-Musa a,⇑, S. Shabunya b, V. Martynenko b, V. Kalinin b a National Center for Combustion and Plasma Technologies, Water and Energy Research Institute, King Abdulaziz City for Science and Technology (KACST), P. O. Box 6086, Riyadh 11442, Saudi Arabia b State Scientific Institution ‘‘A.V. Luikov Heat and Mass Transfer Institute’’, National Academy of Sciences of Belarus, 15 P. Brovki, Minsk 220072, Belarus
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Local thermodynamic equilibrium is
Ni-commercial catalyst
established in CPOX pilot scale reactors. CPOX products composition is controlled by heat transfer processes in catalyst beds. Active thermal insulation has a positive effect on conversion. Preheating up to 700 °C allows for reduction in CO2 content to 0.4 vol.%. Catalyst deterioration is observed at bottom–top gas filtration through reactor.
a r t i c l e
i n f o
Article history: Received 15 March 2015 Received in revised form 18 June 2015 Accepted 11 July 2015 Available online 17 July 2015 Keywords: Methane Natural gas Catalytic partial oxidation Syngas Isothermal equilibrium Active thermal insulation
Inlet mixture Natural gas + air H2 (T) CO(T) CH4 (T) CO2(T) H2O(T)
T
Thermocouples
Inlet mixture T
Natural gas + air
H2 (T) CO(T) C H4 (T) C O2(T) H2O(T)
Outlet mixture H2; CO; CH4; CO2; H2O
a b s t r a c t Two reactor designs with either passive or active thermal insulation were tested to produce synthesis gas from natural gas using catalytic partial oxidation on a system that exceeded laboratory scale. Preheating the working mixture significantly improves the conversion parameters. Preheating levels are limited by two factors: mixture ignition in the heater and overheating at the catalyst bed inlet area. The use of hot conversion products to provide active thermal insulation significantly improves the quality of the conversion. Because the same amount of identical catalyst was used in both reactors, the observed effect supports the hypothesis that the product composition depends primarily on the heat transfer process. Ó 2015 Elsevier B.V. All rights reserved.
1. Introduction Efforts to more efficiently use natural gas reserves have raised interest in synthesis gas (syngas) production, which is currently used for a broad number of applications [1] including modern
⇑ Corresponding author. Tel.: +966 1 481 4316; fax: +966 1 481 3880. E-mail address: [email protected] (A. Al-Musa). http://dx.doi.org/10.1016/j.cej.2015.07.041 1385-8947/Ó 2015 Elsevier B.V. All rights reserved.
Outlet mixture H2; CO; CH4; CO2;H2O
gas-to-product technologies [2] and fuel cells [3]. Steam reforming is frequently used for syngas and hydrogen production; however, this process is not ideal in terms of time or efficiency. The catalytic partial oxidation (CPOX) process is an alternative to steam reforming [4]. CPOX is more energy efficient [5]. Comparative thermodynamic analyses have confirmed that partial oxidation process has advantage over steam reforming process for syngas production [6]. Highly endothermic steam reforming processes require
A. Al-Musa et al. / Chemical Engineering Journal 281 (2015) 852–859
external heating [7] with intensive heat delivery. Steam reforming process requires 2.8 times higher energy supply than partial oxidation process to convert one mole of methane to synthesis gas at 1000 °C. In the face of rising natural gas prices, the partial oxidation process is becoming more economically attractive, especially with small and medium production capacities [6,8]. Numerous studies have described and analyzed CPOX of hydrocarbons. These studies primarily focused on methane, which is the principal constituent of natural gas. Recent reviews have covered the major findings in the field and refer to key publications on theoretical and experimental CPOX processes, catalysts, and reaction kinetics [4,9]. An overwhelming majority of these publications involved the investigation of fundamental CPOX processes and were aimed at clarifying the catalysis kinetics, developing effective catalyst compositions (active agent/promoter/carrier), studying catalyst degradation mechanisms, and increasing catalyst service life. Typically, kinetic methane CPOX studies are investigated experimentally with a laboratory-scale reactor [10–16]. It has been established that, noble and transition metals are feasible CPOX catalysts, meaning that in laboratory conditions, the thermodynamic equilibrium composition is reached for 10–50 ms. These characteristic residence times mean that the chemical reaction rate is ‘‘very high’’. At the same time, active research on laboratory scale operating in kinetic mode are conducted where the selectivities of the various components of the produced synthesis gas depend on choice of residence time [17]. Such effect is particularly observed in the membrane reactors employed to remove hydrogen efficiently from the reaction volume [18]. High flow rates are not required to exclusively study scientific aspects of catalytic processes. In laboratory studies typical flow rates are in the range of 50–300 sccm [18]. implementation of scientific concepts developed using laboratory to scale up processes to industrial plants usually causes a number of problems that do not occur and are not discussed in case of small scale installations. Despite some advantages of reactors operating in kinetic modes all industrial partial oxidation plants are exploited at equilibrium conditions. The design of an industrial-scale plant requires optimization of the heat and mass transfer processes rather than identifying effective catalysts. The most feasible catalyst is primarily chosen based on its cost and durability at approximately 1000 °C. To scale up the methane CPOX process for syngas production from laboratory scale to industrial scale, a number of engineering problems have to be solved, as the heat and mass transfer coefficients vary substantially. The temperature and concentration distributions within the reactor are formed by heat and mass transfer processes interactions. These phenomena cannot be studied properly using a laboratory-scale reactor. A few publications have been devoted to elaborating on industrial-scale reactor concepts [19,20]. The purpose of this article is to demonstrate that when developing industrial scale CPOX reactors, the greatest impact is obtained from heat transfer rather than from catalyst ‘‘efficiency’’. This study expands on the previously published work [21], and testing of the two partial oxidation reactor designs is described. The methane-air mixture CPOX experiments described previously [21] were conducted with a non-preheated inlet mixture with a wide range of CH4/air equivalence ratios. The low process temperatures [21] were due to the lack of an external energy supply. These regimes were characterized by minimal heat loss in contrast to hotter regimes, but even in these conditions, the product composition appears to be worse than with adiabatic equilibrium.
853
The four different catalysts investigated showed no significant differences among their degrees of approximation to the equilibrium composition. This fact led to the conclusion that the deviation in product composition with adiabatic equilibrium is associated with heat loss rather than the lack of residence time. In the case of an inlet mixture with a C:O ratio of 1, the obtained CPOX product composition at the reactor outlet was approximately 28% H2, 13% CO, 7% CH4, and 4% CO2 (volume percents correspond to the drained mixture), whereas the adiabatic calculation gives 32.8% H2, 15.2% CO, 5.5% CH4, and 2.6% CO2. A simple model assuming local thermodynamic equilibrium showed that the observed deviations could be explained by a non-homogeneous radial temperature profile in the catalyst bed that formed due to heat loss. The calculations predicted a temperature difference of approximately 20 °C between the reactor axis and the cylindrical catalyst surface. These calculations were based on the assumption that equilibrium was achieved, meaning that ‘‘sufficient’’ residence time was achieved for the working mixture in the catalyst bed. Although the amount of catalyst used and the flow rates of the working mixture correspond to the CPOX residence times described by Korchnak et al. [22], the issues regarding the ‘‘adequacy’’ of the catalyst sample residence time tested by Al-Musa et al. [21] and the actual value of the radial temperature difference remain disputable. The residence time required to reach equilibrium is a key characteristic of the catalyst in CPOX applications. This characteristic depends on the temperature of the fed gas [23], the catalyst grain size, the effective surface area, and the flow conditions. A theoretical study on the effect of heat transfer and particle size on CH4 partial oxidation was published previously [24]. Periodic temperature perturbations were studied, but the effects of filtration field behavior changes were not considered. In the presence of radial temperature gradients, clearly identifying the impact of residence time on the degree of the approach to equilibrium based on the composition of the outlet mixture is difficult, because the temperature profile affects the chemical process rate and the flow field in the catalyst bed. Studies of mass transfer intensity (i.e., the necessary residence time) should be carried out in a reactor operating at nearly isothermal conditions. Our aim was to improve the composition of CPOX products using preheating of the inlet mixture and by implementing a novel reactor design to decrease heat loss and therefore lower CH4 and CO2 content. 2. Materials and methods 2.1. Experimental setup of passive thermal insulation The reactor dimensions and performance have been described in detail previously [21] and are shown in Fig. 1. In the present study, the potential for increasing the conversion depth by preheating the reaction mixture in this reactor was investigated, and the temperature field in the catalytic bed was estimated. For these experiments, 12 thermocouples were installed to record the temperatures. Thermocouples 1, 3, 4, 5, 7, 9, and 11 were located along the reactor axis, whereas thermocouples 2, 6, 8, 10, and 12 were displaced from the reactor axis towards the edge of the thermal insulation (Fig. 1). Ni=c Al2 O3 catalyst produced by Liaoning Haitai Sci-Tech Development Co., Ltd., China, was used. This granular catalyst with 16 wt% Ni content has cylindrical shape with dimensions, 3.5 mm in length and 1.0 mm in diameter. BET surface area is 91.5 m2/g. The volume of the catalytic bed used in experiments is about 327 cm3.
854
A. Al-Musa et al. / Chemical Engineering Journal 281 (2015) 852–859
Fig. 1. Scheme of the passive thermal insulation experimental reactor.
Wang and Tian also reported high CH4 conversion (greater than 90%) on a Ni=a Al2 O3 catalyst [26]. When Ni was supported on a calcium aluminate carrier [27], the most efficient methane conversion was 85%. The experiments were performed using natural gas (NG) rather than pure methane to better approximate the real industrial situation. The composition of NG in volumetric percents is as follows: CH4 = 97.6%, C2H6 = 1.06%, C3H8 = 0.25%, C4H10 = 0.08%, C5H12 = 0.005%, N2 = 1.005%. Because CPOX reactors create local thermodynamic equilibrium, the composition of the final product depends only on the ratio of atomic C/O/H and the enthalpy of the inlet mixture. As NG contains almost 98% methane, the energy and equilibrium compositions between experiments using pure methane and NG differ little. The actual composition of NG was taken into account for the thermodynamic calculations. All experimental data were obtained using a prepared NG-air mixture with a C/O ratio 1. The composition of the CPOX product mixture was determined after the outlet mixture was drained. The CH4, CO, and CO2 concentrations were measured using optical sensors Gascard II produced by Edinburgh Sensors Ltd. The range of measurements of CH4 and CO is between 0 and 30 vol.%., while for CO2 this value is between 0 and 5 vol.%. H2 concentration was measured within range from 0 to 100 vol.% using katharometer
K1550 produced by Hitech Instruments Ltd. The katharometer and the optical sensors were calibrated using chromatographic data measurements obtained by 7890A GC System from Agilent Technologies. Comparative analysis of the calibration data and the previous measurements demonstrated that the H2 concentration was systematically underestimated by approximately 1.5% [21]. Once tuned, the calibration error of the optical sensors was comparable to the gas chromatograph measurement errors with the exception of CH4 concentrations lower than 0.6%, which approached the sensor’s sensitivity threshold. During the experiment, the compositions of the inlet and outlet mixtures that were measured were compared with gas chromatograph measurements. Optical sensors and katharometers allowed controlled measurements at a sufficiently high frequency for technical applications of no less than once per second. Air and NG flow controllers, mixing devices, and mixture heaters also contributed some error, even when kept at a constant setting. As a result, concentration fluctuations that were caused by system component design rather than instrumentation errors were observed at the reactor outlet even when the device was operating at ‘‘steady state’’. The average sensor reading values are presented here. Gas chromatography data can be considered instantaneous, i.e., containing a fluctuating component. In our studies, we focused on relative performance, i.e., we compared the compositions of the outlet mixture obtained from different initial conditions and tested different reactor designs. In all experiments, use of the same methods and measurement system allows us to make that assessment even using technical precision (5–10%) for the concentration measurements. An electric heater was used to preheat the working mixture. The inlet temperature of the working mixture varied between 200 and 600 °C. The upper heating limit was set to prevent ignition of the mixture in the heater or in the gas supply path prior to contact with the catalyst bed. Mixture ignition is undesirable, not only from the standpoint of an installation structural failure, but also because an increased CO2 content in the outlet mixture results. The critical heating value is dictated by the working mixture properties and heater design. To avoid combustion in the heater, the velocity of the mixture flow must be several times higher than the velocity of the mixture burning at the operating temperature. The use of an intermediate heat carrier (liquid lead or tin) to control the maximum temperature (Tmax) allows us to choose the heater channel flow cross-section and determine the acceptable flow rate based on Tmax and the burning rate of the working mixture. The channel design should exclude stagnant zones where localized burning could occur due to slow fuel and combustion product exchange within the main flow zone.
2.2. Experimental setup of the active thermal insulation reactor To reduce thermal loss, the experimental reactor design was modified to thermally protect the reactor using hot CPOX products. The modified design schematic is shown in Fig. 2. The reactor vessel was designed with an additional protective stainless steel shell engineered as ‘‘a tube within a tube’’. In this design, the hot CPOX products leaving the catalyst bed flow in the annular space between the reactor wall and the shell. Because the temperature of the CPOX product mixture was near the adiabatic temperature, the reaction zone thermal loss should be substantially lower than in the reactor with passive thermal insulation. The reactor was fabricated to test potential design advantages and disadvantages rather than to conduct systematic temperature studies. In these experiments, the reactor temperature was measured at two points: the inlet to measure the heating temperature of the working mixture and the catalytic bed outlet to measure the temperature of the CPOX product.
A. Al-Musa et al. / Chemical Engineering Journal 281 (2015) 852–859
855
Fig. 4. Temperature difference between axial and outer catalyst layer temperatures versus the inlet temperature (d – G = 4.25 Nm3/h, j – G = 2.54 Nm3/h).
Fig. 2. Scheme of the active thermal insulation experimental reactor.
3. Results and discussion 3.1. Effects of preheating the working mixture using a passive thermal insulation reactor Two series of experiments were performed using a preheated working mixture with C/O = 1 and flow rates of 2.54 and 4.25 Nm3/h. Temperature characteristics of the catalyst units are shown in Figs. 3 and 4. The concentrations of the outlet mixture components versus initial temperature are plotted in Fig. 5. The experimental measurements are shown by labels, while dashed lines present the results of adiabatic calculations. The pure, unheated methane data [21] obtained at a flow rate (G) = 3.2 Nm3/h are shown as well.
Fig. 3. On axis temperature (d – G = 4.25 Nm3/h, j – G = 2.54 Nm3/h, e – G = 3.2 Nm3/h) and outer catalyst layer temperature (s – G = 4.25 Nm3/h, h – G = 2.54 Nm3/h) versus the inlet temperature. Dashed curve is the adiabatic estimation.
The temperature difference was higher using a slower flow rate. This is associated with an effective increase in thermal conductivity when the filtration velocity increases (dispersion mechanism [25]). As expected, the temperature at the insulated catalyst bed boundary was lower than the calculated adiabatic temperature. This difference increased with increased heating. The apparent difference between the axial and calculated adiabatic temperatures was somewhat unexpected. This difference may be attributed to incomplete partial oxidation reactions even near the reactor axis, where heat loss effects were minimal. Therefore, the residence time is not sufficient at these flow rates. At a low flow rate, the temperature near the reactor axis was somewhat higher, which may be the result of lower heat intensity and mass transfer processes at lower filtration rates. In this case, disruption of the axial symmetry of the thermal pattern and filtration may influence and consequently distort the temperature profile. As expected, heating the reaction mixture at the reactor inlet led to increased H2 and CO concentrations in the CPOX products, whereas the methane and carbon dioxide concentrations decreased. This indicated improved conversion relative to the ideal partial oxidation formula: 2CH4 + O2 ? 2CO + 4H2. When comparing the outlet composition with the adiabatic equilibrium composition, the difference was clearly more pronounced with more intensive heating, which may be explained by increased thermal loss during the transition to the higher reactor temperatures, causing a decrease in residence time. The composition of the conversion product did not depend solely on the flow rate. When the catalyst amount was fixed, an increase in the flow rate should reduce the residence time, impairing conversion. On the other hand, the flow of energy through the section of the reactor is proportional to the flow rate of the working mixture, while the heat loss depends only on the temperature of the reactor shell. The increase in flow rate should have a positive effect on the quality of conversion, because it reduces the percentage of heat loss, and the output mixture should have a higher temperature. this would have been the case, if the temperature of the shell remained unchanged. However, there are the effects of increasing the temperature of the shell. The increase in the rate of filtration extends the scope of the exothermic process where the temperature is higher, and intensifies the heat transfer in a porous bed, i.e. it facilitates heat transfer to the shell. As a result, the average temperature of the reactor shell increases, and heat loss increases, too. The heat flux in the shell to the flanges of the reactor has a negative effect. When the negative and positive factors are close to each other, increasing the flow will not improve conversion. In experiments at 4.25 and 2.54 Nm3/h flow rates, these factors almost cancel each other quantitatively. As shown in Fig. 5, the measured conversion product compositions were similar (with a slight advantage) to those obtained using 4.25 Nm3/h flow rate.
856
A. Al-Musa et al. / Chemical Engineering Journal 281 (2015) 852–859
Fig. 5. Effect of inlet temperature and flow rate on CPOX product composition. d – G = 4.25 Nm3/h, j – G = 2.54 Nm3/h, e – G = 3.2 Nm3/h (pure CH4 test [21]), dashed curve – adiabatic estimation.
We increased the flow rate of the working mixture in an attempt to produce an adiabatic-like composition; however, it did not produce the expected results. Doubling the catalyst volume to increase the residence time led to increased thermal loss from the catalytic bed and a decreased outlet temperature. Thus, in experiments with a passive thermally insulated reactor, preheating the inlet mixture (C/O = 1) decreased the CH4 concentrations of the CPOX products from 7% to 3% and CO2 from 3.8% to 1%. The effects of temperature and residence time on the extent of conversion clearly could not be separated in the presence of thermal loss. 3.2. Effect of active thermal insulation on natural gas conversion The composition of the CPOX product was measured in several experiments at different flow rates and heater temperatures (Tables 1 and 2). All experiments were conducted with an NG-air working mixture of C/O 1. Experimental data were sorted by heating temperatures of the inlet mixture and flow rates, and the values corresponding to the adiabatic equilibrium are given in parentheses. Temperature gradients were not measured in the second reactor design; however, given the improvement in the composition of the outlet mixture, we concluded that they were smaller. The two different reactors were compared based on the outlet mixture quality achieved when heating the mixture to 600 °C at a flow rate of 2.52 Nm3/h (second row of Table 1). The composition was 37.3% H2, 19.7% CO, 1% CH4, and 0.5% CO2, which was substantially better than the corresponding values (34% H2, 18% CO, 3.5% CH4, and 1.2% CO2) obtained in experiment using the first reactor. Table 1 contains the results of six independent experiments. Over the course of experimentation, a slight design modification of the reactor was made (changing gaskets, thermal insulation, catalyst bed fixing), accompanied by
a partial replacement of the catalyst material. However, no special attention was paid to catalyst aging during this time period. When analyzing the data collected (Table 1), the results obtained using the same or similar parameters of the inlet mixture were scattered. We believe this was primarily the result of catalyst aging. In the passive insulation reactor experiments (Fig. 1), catalyst-aging effects were not observed. The lack of effect can be explained as follows: 1. In the passive insulation reactor design (Fig. 1), the direction of gas flow and gravity effects coincided, i.e., the catalyst particles were packed downwards and did not contribute to mechanical grain movements. In the case of the active insulation reactor design (Fig. 2), the situation was reversed, i.e., gas flowed against gravity, lifting the particles. This caused interparticle friction and produced mechanical failure. Attempts to lock the layer by mechanically pressing the upper layer with porous ceramic plates during assembly of the reactor catalyst unit did not solve this problem. Additionally, the intensive heating resulted in the formation of free space inside the reactor that increased with the operating time. Every time the installation was cooled and the reactor was disassembled, formation of free space was observed. Not surprisingly, higher flow rates intensified the mechanical destruction. 2. The experiments performed using the active insulation reactor were conducted at higher flow rates for the working mixture in an attempt to produce adiabatic-like experimental conditions. The increased flow rate accelerated all the various degradation processes including thermal, mechanical, and chemical processes. 3. Improving the adiabatic properties of the experimental unit resulted in higher temperatures in the catalyst bed, which is particularly critical for the exothermic reaction zone (hot
857
A. Al-Musa et al. / Chemical Engineering Journal 281 (2015) 852–859 Table 1 Compositions and characteristics of CPOX products produced from the active thermal insulation reactor. P, bar
C/O
Theater, C
T, C
1.10 1.18 1.25 1.50 1.50 1.72 2.00 2.11 2.14 1.30 1.31 1.41 1.55 1.73 1.82 1.94 2.18 1.95 1.34 1.67
1.0 1.0 1.02 1.02 1.0 1.02 1.0 1.02 1.02 1.0 1.02 1.02 1.02 1.02 1.0 1.02 1.02 1.0 1.0 1.0
600 600 600 600 600 600 600 600 650 700 700 700 700 700 700 700 700 705 705 720
705 850 805 835 905 850 920 870 880 930 815 835 855 865 705 875 895 845 790 870
(819) (822) (815) (821) (831) (827) (842) (835) (861) (885) (875) (877) (880) (883) (895) (886) (890) (900) (890) (906)
Gtot, Nm3/h
H2
CO
CH4
1.68 2.52 3.30 5.10 5.10 6.60 8.45 9.00 9.00 3.40 3.65 4.43 5.30 6.40 6.73 7.72 9.00 8.20 4.10 5.85
36.4 (39.2) 37.3 (39.2) 35.8 (39.1) 35.8 (38.9) 36.9 (39.0) 35.8 (38.8) 35.9 (38.7) 35.7 (38.6) 36.3 (39.1) 38.2 (39.8) 36.85 (39.8) 37.0 (39.8) 37.0 (39.7) 37.2 (39.6) 37.8 (39.6) 37.2 (39.6) 37.1 (39.5) 37.4 (39.6) 38.6 (39.8) 38.0 (39.8)
18.0 (19.8) 19.7 (19.8) 18.8 (19.7) 18.8 (19.7) 19.5 (19.7) 18.8 (19.6) 19.5 (19.6) 18.8 (19.5) 19. (19.8) 19.9 (20.1) 19.3 (20.1) 19.4 (20.1) 19.5 (20.1) 19.5 (20.0) 19.23 (20.0) 19.4 (20.0) 19.3 (20.0) 19.2 (20.0) 19.89 (20.1) 19.3 (20.1)
2.45 1.00 1.68 2.00 1.60 2.50 2.30 2.60 2.12 0.90 1.42 1.40 1.45 1.58 1.04 1.70 1.65 2.40 1.74 1.00
Gtot, Nm3/h
H2
CO
3.40 6.90 6.90 8.50 8.50
40.0 39.6 39.8 39.5 38.3
CO2 (0.95) (0.99) (1.43) (1.53) (1.11) (1.61) (1.27) (1.73) (1.43) (0.57) (0.97) (0.99) (1.03) (1.08) (0.7) (1.13) (1.18) (0.71) (0.57) (0.59)
1.45 0.46 1.07 0.97 0.50 0.93 0.60 0.93 0.83 0.42 0.78 0.75 0.73 0.70 0.69 0.70 0.70 0.54 0.42 0.65
(0.29) (0.30) (0.28) (0.31) (0.33) (0.33) (0.38) (0.36) (0.25) (0.14) (0.11) (0.11) (0.12) (0.14) (0.18) (0.15) (0.17) (0.18) (0.14) (0.14)
Table 2 Composition and characteristics of CPOX products using fresh catalyst. P, bar
C/O
Th, C
T, C
1.25 1.73 1.77 2.00 2.00
1.0 1.0 1.0 1.0 1.0
600 600 700 650 650
895 835 880 875 870
(824) (836) (894) (868) (868)
spots). The increased flow rate and decreased thermal loss may have raised the temperature in this area, thereby accelerating thermal degradation of the catalyst.
3.3. Catalyst degradation To test the hypothesis that the design of the active thermal insulation reactor increased catalyst degradation, an experiment was performed using fresh catalyst material (Table 2). Photographs of fresh catalyst and used catalyst following the experiments are shown in Fig. 6. Freshly prepared catalyst resulted in high-efficiency CPOX processes. The composition of the CPOX products was near equilibrium, and the residence time was
(39.1) (38.9) (39.6) (39.2) (39.2)
20.5 20.1 20.3 20.3 19.6
CH4 (19.8) (19.6) (20.0) (19.8) (19.8)
0.67 0.82 0.67 0.80 1.40
CO2 (1.01) (1.19) (0.69) (0.99) (0.99)
0.09 0.16 0.10 0.20 0.32
(0.31) (0.36) (0.17) (0.27) (0.27)
approximately 0.025 s at the maximum flow rate of 8.5 Nm3/h. Those data with concentrations measuring ‘‘better’’ than adiabatic were the result of composition, inlet mixture temperature, and errors in pressure measurement. The last two rows of Table 2 indicate catalyst degradation following 1 h of operation in a fixed working mode. This is clearly demonstrated by the difference in CPOX product concentrations using an identical working regime. The measured composition of the CPOX product was noticeably deteriorated. After reactor disassembly, we observed that the catalyst grains were much lighter and smaller in size (Fig. 6). Destruction of alumina-based Ni catalysts with CH4 partial oxidation was reported previously [28], whereas b SiC was proposed to be an
Fig. 6. Photographs of the catalyst samples. (a) Fresh catalyst prior to loading in the reactor and (b) used catalyst following experimentation using approximately 11 Nm3 working mixture passed through 0.25 L of catalyst.
858
A. Al-Musa et al. / Chemical Engineering Journal 281 (2015) 852–859
efficient Ni catalyst carrier that could successfully be used for syngas production from methane. 3.4. Feasibility of passive and active thermal insulation When using the passive thermal insulation CPOX reactor (Fig. 1), preheating the working mixture increased the degree of conversion, but attempts to produce syngas containing less than 0.5% CH4 and CO2 failed due to thermal loss. Preheating the inlet mixture up to 600 °C yielded CPOX products containing about 3% CH4 and 1% CO2. Sufficient superheating of the inlet mixture to compensate for thermal loss and to produce CPOX products with a desired composition is impossible due to the risk of ignition of the mixture in the heating element. A heater coil design with an intermediate heat carrier (liquid lead) allowed preheating of the NG-air mixture (C/O 1) up to 750–800 °C without ignition. However, these regimes lead to reactor overheating in the catalyst bed inlet area, which is dominated by an exothermic reaction. A stationary heating regime was not achieved, because the experiment automatically stopped when the temperature exceeded 1180 °C. Increasing the flow rate reduced thermal loss; however, no significant improvement in conversion was observed. The positive effect was diminished by reduced residence time and increased working pressure, causing an equilibrium shift to higher CO2 and CH4 concentrations. Increasing the catalyst volume only expanded the heat transfer surface, and the thermal loss increased. In the catalyst bed pipe, the effective temperature near the cylindrical surface was lower than the axial temperature due to thermal loss. As a result, most of the working gas flowed close to the cylindrical surface. Thus, a tubular reactor containing endothermic processes can lead to significant deviations in the composition of CPOX products from theoretical adiabatic equilibrium values. Although increasing the reactor diameter and reducing the catalyst bed height may reduce the cylindrical surface area (i.e., the surface of thermal loss), these options are limited due to homogenous filtration flow field requirements. In light of these considerations and the experimental data, we concluded that the passive heat insulation-type reactor design (Fig. 1) is problematic for producing syngas with CO2 and CH4 contents less than 0.5%. Optimizing the heating, insulation, and catalytic unit size will certainly reduce the experimental 3% CH4 and 1% CO2 CPOX product values. However, an optimized reactor would need to be designed for a specific, constant flow rate, i.e., the installation performance cannot be varied without affecting the composition of the CPOX products. Of note, higher hydrocarbon use would limit preheating, as these molecules are less stable and more easily oxidized than CH4. The active thermal insulation reactor design using hot CPOX products to reduce thermal loss (Fig. 2) reduced the radial temperature gradient in the catalyst bed and substantially improved hydrocarbon conversion in the CPOX process. When the inlet mixture was preheated to 650–700 °C, the CPOX products had a composition that was near equilibrium and a resident time of approximately 0.025 s. The experimental results produced 0.7% CH4 and 0.2% CO2 concentrations. Those values could be improved by lowering the working pressure in the CPOX reactor. In addition to the benefits of active thermal insulation, other notable limitations of the reactor design include: 1. The adiabatic-like reactor performance is higher at higher flow rates. Therefore, decreasing the flow rate may lead to deterioration of the composition of CPOX products. 2. Fabrication of an active thermal insulation reactor (Fig. 2) requires addressing a number of technical issues to ensure efficient operation of the CPOX reactor at 1000 °C. The inner shell
of the CPOX reactor can reach temperatures of 1200 °C or higher in exothermic reaction zones, depending on the properties of the catalyst. 3. The introduction of radial thermocouples through the two shells made the reactor more vulnerable to leakage. Therefore, this reactor design is not suitable for a systematic study of the temperature characteristics of the catalyst bed. 4. Conclusion Implementing thermal protection using reverse flow of the hot product gas does not allow for the adequate control of the reactor surface temperature because the system self-adjusts. More effective and practical thermal protection may be accomplished with controlled heating of the reactor shell, which would allow the external surface temperature of the reactor to be controlled along its entire length and adjusted for different performance regimes. This could potentially maintain an isothermal radial temperature profile to obtain a CPOX product composition close to theoretical adiabatic equilibrium estimations. Notably, heating is aimed at compensating for thermal loss, and the inlet mixture heater constitutes the primary energy input required to achieve the desired product temperature. Our results obtained using Ni-based catalysts demonstrated that this catalyst material has low stability when used in large-scale partial oxidation processes. The exact mechanism of catalyst degradation is unknown; however, if the degradation is a thermal process, another catalyst that is stable at working temperatures of 1000–1100 °C is needed. Acknowledgments Financial support from the King Abdulaziz City for Science and Technology (KACST) is gratefully acknowledged (Project No. 33-816: ‘‘Experimental and theoretical studies for chemical and thermal processes in partial oxidation reactors’’). In addition, we would like to thank our colleagues in KACST and HTMI for their valuable help. References [1] I. Wender, Fuel Process. Technol. 48 (1996) 189–297. [2] T.H. Fleisch, Syngas Chemistry Symposium, Dresden, Germany, October 4–6, 2006. [3] M. Liu, M.J.B.M. Pourquie, L. Fan, W. Halliop, et al., Fuel Cells 13 (3) (2013) 428– 440. [4] S.C. Tsang, J.B. Claridge, M.L.H. Green, Catal. Today 23 (1995) 3–15. [5] S.S. Bharadwaj, L.D. Schmidt, Fuel Process. Technol. 42 (1995) 109–127. [6] A.P.E. York, T. Xiao, M.L.H. Green, Top. Catal. 22 (2003) 345–358. [7] K.L. Hohn, L.D. Schmidt, Appl. Catal. A: General 211 (2001) 53–68. [8] V. Recupero, L. Pino, R. Di Leonardo, M. Lagana, G. Maggio, J. Power Source 71 (1998) 208–214. [9] R. Lanza, J.A. Velasco, S.G. Jaras, Catalysis 23 (2011) 50–95. [10] B.C. Enger, R. Lodeng, A. Holmen, Appl. Catal. A: General 346 (2008) 1–27. [11] M. Maestri, D.G. Vlachos, A. Beretta, G. Groppi, E. Tronconi, AIChE J. 55 (4) (2009) 993–1008. [12] A. Slagtern, H.M. Swaan, U. Olsbye, I.M. Dahl, C. Mirodatos, Catal. Today 46 (1998) 107–115. [13] D. Dalle Nogare, N.J. Degenstein, R. Horn, P. Canu, L.D. Schmidt, J. Catal. 258 (2008) 131–142. [14] R. Horn, K.A. Williams, N.J. Degenstein, A. Bitsch-Larsen, et al., J. Catal. 249 (2007) 380–393. [15] B.C. Enger, R. Lodeng, A. Holmen, Appl. Catal. A: General 364 (2009) 15–26. [16] U. Friedle, G. Veser, Chem. Eng. Sci. 54 (1999) 1325–1332. [17] J. Zaman, A. Chakma, J. Membr. Sci. 92 (1) (1994) 1–28. [18] Y.S. Cheng, M.A. Pena, K.L. Yeung, J. Taiwan Inst. Chem. Eng. 40 (3) (2009) 281– 288. [19] G. Kolios, J. Frauhammer, G. Eigenberger, Chem. Eng. Sci. 57 (2002) 1505– 1510. [20] B. Glockler, A. Gritsch, A. Morillo, G. Kolios, G. Eigenberger, Chem. Eng. Res. Des. 82 (2) (2004) 148–159. [21] A. Al-Musa, S. Shabunya, V. Martynenko, S. Al-Mayman, et al., J. Power Source 246 (2014) 473–481.
A. Al-Musa et al. / Chemical Engineering Journal 281 (2015) 852–859 [22] J.D. Korchnak, M. Dunster, A. English, Patent cooperation treaty WO90/06282, PCT/US89/05369, 1990. [23] S.-B. Tang, F.-L. Qiu, S.-J. Liu, J. Nat. Gas Chem. 5 (3) (1996) 272–277. [24] X. Ren, X. Guo, J. Nat. Gas 20 (5) (2011) 503–506. [25] M. Kaviany, Principles of Heat Transfer in Porous Media, Springer-Verlag, New York, 1991, 626p.
859
[26] H. Wang, Z. Li, S. Tian, J. Nat. Gas 12 (2003) 205–209. [27] M.A. Goula, A.A. Lemonidou, W. Griinert, M. Baerns, Catal. Today 32 (1996) 149–156. [28] P. Leroi, B. Madani, C. Pham-Huua, M.-J. Ledoux, et al., Catal. Today 91–92 (2004) 53–58.