SEMI-BATCH AND CONTINUOUS CARBOCHLORINATION OF ZIRCONIA IN A PILOT PLANT FLUIDIZED BED REACTOR: OPTIMIZATION OF OPERATING CONDITIONS VIA RESPONSE SURFACE METHODOLOGY
HTML Full TextSEMI-BATCH AND CONTINUOUS CARBOCHLORINATION OF ZIRCONIA IN A PILOT PLANT FLUIDIZED BED REACTOR: OPTIMIZATION OF OPERATING CONDITIONS VIA RESPONSE SURFACE METHODOLOGY
M. R. Ghasemi and S. M. Ghoreishi^{*}
Department of Chemical Engineering, Isfahan University of Technology, Isfahan 84156 - 83111, Iran.
ABSTRACT: Zirconium tetrachloride was produced in a pilot plant fluidized bed reactor via the carbochlorination reaction of zirconia. Minitab software 17 was used for the design of experiment with response surface methodology (RSM) for both semi batch and continuous processes. The operating temperature (800-1200 °C), reaction time (30-90 min) and mean particle size (MPS) (70-130µm) were chosen as the range of operating conditions at semi-batch mode. Also, the ranges of operating variables at the continuous mode were chosen to be the operating temperature(1000-1400°C), chlorine concentration (3-5mol/m^{3}) and mean particle size (MPS) (70-130µm). The coefficients of determination (R^{2}) for semi-batch and continuous modes by RSM analysis were 99.54% and 99.94%, respectively. The optimal operating conditions was predicted to be the Cl_{2} concentration of 3mol/m^{3}, temperature of 1380 °C and the mean particle size of 70µm in which the maximum chlorine conversion of 98.26% was obtained at continuous process. The optimum operating conditions for semi-batch process were found by using RSM modelling to be 1200°C, 90 min and 70 µm, in which the maximum zirconia conversion of 92.7% was obtained.
Keywords: |
Zirconia, Carbochlorination, Response surface methodology (RSM), Fluidized bed reactor (FBR)
INTRODUCTION: Zirconia (Zirconium dioxide (ZrO_{2})) is a crystalline oxide of zirconium ^{1}. Zirconium metal has unique characteristics like resistance to radiation damage ^{2}, very low absorption cross section ^{3, 4}, compatibility with uranium fuel ^{5}, resistance to corrosion ^{6}, retention of strength ^{7} and desirable response to alloying modifications ^{8}. Also, pure zirconia is used in nuclear and metallurgical industries, solid electrolytes, optical fibers and gem industries (cubic single crystals) ^{9}.
Nuclear-grade zirconium metal produces in the conventional process from zirconium tetrachloride by the magnesium reduction or fused salt electrolysis. Chlorination of pure ZrO_{2 }using chlorine in the presence of carbon at high temperature produces zirconium tetrachloride (ZrCl_{4}) ^{10}.
ZrCl_{4 }is used as the starting material in other proposed methods for the production of zirconium as an alternative to conventional route ^{11}. According to the following reaction, Zirconium tetrachloride is produced using chlorination of zirconia in the presence of carbon which is really an approximation (CO_{2} formation occurs simultaneously with CO) ^{12}:
This reaction is called carbochlorination, because, ZrO_{2 }chlorination occurs in the existence of carbon (a reducing agent) in dry conditions to make the reaction thermodynamically possible. Presence of carbon favours the formation of chlorides and decreases the oxide formation tendency by providing a low oxygen potential atmosphere. It is known that at the same high temperature, the chlorination reaction rate is slower than the carbochlorination reaction ^{9}. Fluidized bed reactor (FBR) is a contacting device that uses the principle of fluidization in its operation. FBR has many benefits, including uniform temperature distribution, favourable heat and mass transfer, low pressure drop, excellent particle mixing, operational flexibility and ease of solid handling ^{13}.
Zirconia carbochlorination is a complex non-catalytic gas-solid reaction. There are some investigations that report the chlorination of zirconia. Zirconium dioxide carbochlorination on literature contains the main kinetic parameters (such as the order of the reaction with respect to reactants and activation energy) and the reaction mechanism. These investigations often hampered by the reaction complexity because involve two solid and one gaseous reacting component.
O’ reilly et al., ^{14 }found that in the carbochlorination of ZrO_{2 }in a static bed (using excess of petroleum coke powder), the chemical reaction model controls the process in the temperature range of 1223-1373 K. The activation energy and order of reaction into Cl_{2 }were 230.3 kJ mol^{−}^{1} and 0.64, respectively. Jena et al., ^{10 }studied ZrO_{2 }chlorination kinetics in the temperature range of 973 to 1273 K. The activation energy was found to be 18.3kJ/mol. Bicerolu and Gauvin ^{11 }determined the kinetics of carbochlorination of ZrO_{2 }and the effects of temperature (1400-1950 K), chlorine concentration and carbon (graphite) content on the rate of conversion in a thermo-gravimetric analysis. They reported the activation energy and order of reaction into Cl_{2 }to be 94kJ/mol^{−}^{1} and 0.79, respectively. Moreover, they found that the carbon concentration range of 22-26% can result in maximum rate of process.
Landsberg et al., ^{15} studied the kinetics of zirconium dioxide (in the form of disk pellets surrounded by a loosely packed bed of carbon powder) chlorination in the temperature range of 1120-1320 K. They found that the rate of reaction for carbothermic chlorination was first order with respect to the chlorine concentration and the activation energy was determined to be 127.7kJ/mol. The results of different studies varied based on experimental conditions, the method of sample preparation and the type of carbonaceous agent used. A survey of the literature showed that there is no published data regarding to chlorination reaction of zirconia in the presence of carbon in a pilot scale fluidized bed reactor, in spite of the fact that it is very important in the extractive metallurgy of zirconium.
The purpose of the present work was the carbochlorination of zirconia in a pilot plant fluidized bed reactor in the semi-batch and continuous mode. No attempt was made to identify the detailed mechanism of the reaction or the reaction rate. RSM was implemented for the modelling and optimizations of operating conditions while the response (reaction conversion) was affected by chlorine concentration, temperature, reaction time and mean particle size. Furthermore, the important operating parameters affecting the reactor performance were investigated by the validated model.
Experimental Section:
MATERIALS: The gases used for this study were Nitrogen (99.9% purity) and Cl_{2 }(99.9% purity). The powdered materials were ZrO_{2} (Sooreh Co., Isfahan, Iran) (composition: Zr = 73.9%, Fe = 0.18%, Si = 0.015%, Al = 0.01% and Hf = 0.004%) and carbon (composition: fixed carbon = 95.8%, Ash = 3.2%, Volatile 0.6% and heating loss = 0.4%). The BET surface areas of ZrO_{2} and carbon were determined to be 14.4m^{2}/g and 1.5m^{2}/g, respectively (Sorptometer Kelvin 1042, ISO 9277-2010). The characterization (size and morphology) of the powdered ZrO_{2 }and carbon were determined by scanning electron microscopy (SEM) and X-ray powder diffraction (XRD). Fig. 1 and 2 shows the SEM and XRD of the powdered materials in the reactor feed, respectively. The mixture of carbon / zirconia (20/80% wt) was used to prepare briquettes using laboratory hydraulic press. Then the briquettes were powdered by means of a laboratory crusher to produce particles with mean size from 70 to 130 µm.
FIG. 1: THE SEM IMAGES OF (A) CARBON AND (B) ZIRCONIA
FIG. 2: THE XRD ANALYSIS OF (A) CARBON AND (B) ZIRCONIA
Apparatus and Procedure: Fig. 3 shows the schematic representation of the pilot-scale fluidized bed reactor (FBR) with its ancillary equipment used for the experimental investigation of carbochlorination of zirconia in semi-batch and continuous modes. The preheating part of the reactor vessel was 1.3 cm I.D., 6.4 cm O.D. and 40.6 cm long. The reaction section was 6.4 cm I.D., 11.4 cm O.D. and 55.9 cm long. The disengaging section was11.4 cm I.D., 16.5 cm O.D. and 15.2 cm long. A water-cooled brass clamp was used to supply power to the bottom of reactor. ZrCI_{4} was condensed as a fluffy material in the space condenser. Fig. 4 shows the flow sheet of the process.
FIG. 3: SCHEMATIC REPRESENTATION OF THE PILOT-SCALE FLUIDIZED BED REACTOR (FBR) WITH ITS ANCILLARY EQUIPMENT
FIG. 4: THE FLOW SHEET OF THE ELECTROTHERMAL FLUIDIZED BED CHLORINATION PROCESS
All experiments were carried out at the fixed Cl_{2} flow rate of 2 l/min. Table 1 shows the specifications of the pilot plant for the operation in semi-batch mode. In this case, the ranges of operating variable (temperature (T), mean particle size (MPS) and reaction time (t)) were presented in Table 2.
TABLE 1: OPERATING CONDITIONS OF FBR IN SEMI - BATCH MODE
Quantity | Value |
Cl_{2} flow rate | 2 L/min |
Superficial velocity of Cl_{2} in the bed | 5 cm/s |
Weight of C/ZnO_{2} mixture | 2 Kg |
Weigh percent of carbon in the mixture | 20% |
electric current | 200-250 A |
Voltage | 50-60 V |
TABLE 2: UNCODED AND CODED LEVELS OF INDEPENDENT VARIABLES USED IN THE RSM DESIGN FOR SEMI-BATCH MODE
Coded
variables |
Temperature (T) (°C) | Mean particle size (MPS) (µm) | Time (t)
(min) |
|||
-1 | 800 | 70 | 30 | |||
0 | 1000 | 100 | 60 | |||
1 | 1200 | 130 | 90 | |||
Design of Experiment: Conventionally, the classical optimization method is used by investigating the effect of one variable on the response (one-variable-at-a-time method). This optimization technique has major disadvantages like the necessity of a great deal of work (high number of required experiments) that results in increased time and expenses and the lack of investigating the interactive effects of the parameters in the process. Due to these drawbacks, the optimization is done by means of statistic methods like the response surface methodology ^{16-18}.
Response surface methodology is one of the well-established combinations of mathematical and statistical techniques for designing and building experimental model. This methodology is useful for the evaluation of relative significance of each independent variable at minimum number of experiments. The objective is to understand the interaction among various parameters and the optimization of the operating conditions to obtain the best response, simultaneously ^{19}. RSM is a useful tool which has been widely applied successfully in several fields for the optimization of operating parameters in various processes ^{20-22}.
The RSM quadratic polynomial model that includes linear, quadratic and interaction terms is as follows ^{23, 24}:
Where A_{0 }is constant; A_{i}, A_{ii} and A_{ij} are the coefficients of the linear, quadratic and interaction parameters, respectively. Z_{i} represents the variables and ε is the residual associated with the experiments; k = the number of variables.
Central composite rotatable design (CCRD) (central composite design: full factorial, cube points = 8, center points in cube = 4, axial points = 6, center points in axial = 2 and α = 1) was selected as the second order symmetrical designs in RSM for 3 independent parameters with 3 levels for each one for both semi-batch and continuous. The variables were coded according to Eq. (3):
Where X_{i }= the real value, X_{i,c.p }= the real value at the center point and = the step change in the variable X_{i}.
The independent variables for semi-batch process in this research were reaction time (t), reaction temperature (T) and mean particle size (MPS) in the range of 30-90 min (step size = 30), 800–1200 °C (step size = 200) and 70-130 µm (step size = 30), respectively. Zirconia conversion was response variable (dependent variable). Also, the independent variables for continuous mode were Chlorine concentration (C), reaction temperature (T) and mean particle size (MPS) in the range of 3-5 mol/m^{3} (step size = 1), 1000-1400 °C (step size = 200) and 70-130 µm (step size = 30), respectively. Chlorine conversion was selected as the response variable.
The levels of independent parameters (coded and uncoded) are shown in Table 2 (semi-batch mode) and Table 3 (continuous mode). Table 4 and 5 illustrate the design of experiments which were according to the coded levels in CCRD which leads to 20 experimental runs for each process mode. The implementation of RSM was performed by Minitab software (version 17).
TABLE 3: UNCODED AND CODED LEVELS OF INDEPENDENT VARIABLES USED IN THE RSM DESIGN FOR CONTINUOUS MODE
Coded variables | Temperature (T) (°C) | Mean Particle Size (MPS) (µm) | Cl_{2} concentration
(C) (mol/m^{3}) |
-1 | 1000 | 70 | 3 |
0 | 1200 | 100 | 4 |
1 | 1400 | 130 | 5 |
Analysis of variance (ANOVA), R^{2} and Adj. R^{2 }were used for the evaluation of the mathematical models. The optimum operating conditions were obtained through the first derivative of the mathematical functions for zirconia and chlorine conversion in the semi-batch and continuous processes, respectively.
TABLE 4: CENTRAL COMPOSITE ROTATABLE DESIGN FOR ZIRCONIA CONVERSION AND OBSERVED RESPONSES IN SEMI-BATCH MODE
Number of experiments | MPS | t | T | X_{ZrOz} (Experimental) (%) | X_{ZrOz} (RSM Predicted) (%) |
1 | -1 | 1 | -1 | 67.0 | 64.2 |
2 | 0 | 0 | 0 | 57.5 | 57.3 |
3 | 0 | 0 | 0 | 57.9 | 57.3 |
4 | 1 | -1 | -1 | 15.5 | 13.3 |
5 | -1 | -1 | 1 | 35.5 | 36.7 |
6 | 1 | 1 | 1 | 87.5 | 88.2 |
7 | 0 | 0 | 0 | 57.1 | 57.3 |
8 | 1 | -1 | 1 | 29.2 | 30.8 |
9 | 1 | 1 | -1 | 62.4 | 60 |
10 | 0 | 0 | 0 | 58.6 | 57.3 |
11 | -1 | 1 | 1 | 91.8 | 92.7 |
12 | -1 | -1 | -1 | 20.8 | 18.9 |
13 | 0 | 1 | 0 | 72 | 74.5 |
14 | -1 | 0 | 0 | 59 | 60.4 |
15 | 0 | 0 | 0 | 56.4 | 57.3 |
16 | 1 | 0 | 0 | 54.2 | 55.4 |
17 | 0 | 0 | -1 | 38.8 | 47 |
18 | 0 | 0 | 0 | 57.0 | 57.3 |
19 | 0 | 0 | 1 | 75.5 | 70 |
20 | 0 | -1 | 0 | 23 | 23.1 |
TABLE 5: CENTRAL COMPOSITE ROTATABLE DESIGN FOR Cl_{2} CONVERSION AND OBSERVED RESPONSES IN CONTINUOUS MODE
Number of experiments | MPS | C | T | XCl_{2} (Experimental) (%) | XCl_{2} (RSM Predicted) (%) |
1 | 0 | -1 | 0 | 83.2 | 83.1 |
2 | 0 | 0 | 0 | 79.5 | 79.7 |
3 | 0 | 0 | 1 | 94.3 | 94.9 |
4 | 0 | 1 | 0 | 73.7 | 74.5 |
5 | 0 | 0 | 0 | 78.4 | 79.7 |
6 | -1 | 0 | 0 | 81.9 | 82.0 |
7 | 0 | 0 | -1 | 32.4 | 32.6 |
8 | 1 | 0 | 0 | 77.3 | 77.9 |
9 | 1 | -1 | 1 | 94.4 | 94.2 |
10 | 0 | 0 | 0 | 80.8 | 79.7 |
11 | 1 | 1 | -1 | 23.4 | 23.1 |
12 | 0 | 0 | 0 | 79.4 | 79.7 |
13 | -1 | 1 | 1 | 95 | 97.4 |
14 | -1 | -1 | -1 | 40.1 | 40.1 |
15 | 1 | -1 | -1 | 36.8 | 36.7 |
16 | -1 | -1 | 1 | 98.2 | 98.1 |
17 | -1 | 1 | -1 | 27.7 | 27.5 |
18 | 1 | 1 | 1 | 90.1 | 89.7 |
19 | 0 | 0 | 0 | 78.6 | 79.7 |
20 | 0 | 0 | 0 | 81.3 | 79.7 |
RESULTS AND DISCUSSION:
Fitting the Model and Analysis of Experimental Data: The least squares technique was used for the experimental data in Table 4 and 5 to determine the coefficients in Eq. (2) which were summarized in Table 6 and 7 for semi-batch and continuous modes, respectively.
TABLE 6: REGRESSION COEFFICIENTS AND ANOVA OF PREDICTED SECOND-ORDER POLYNOMIAL MODEL FOR ZIRCONIA CONVERSION AT SEMI-BATCH MODE
Source | Degree of freedom | Effect | Coefficient | SE Coefficient | Adj SS | Adj MS | t-value | f-value | p-value |
Model | 11 | 8285.08 | 753.19 | 48.94 | 0.000 | ||||
constant | 57.30 | 1.52 | 37.81 | ||||||
blocks | 2 | 10.84 | 5.42 | 0.35 | 0.714 | ||||
Linear | 3 | 7976.00 | 2658.67 | 172.74 | 0.000 | ||||
MPS | 1 | -5.06 | -2.53 | 1.24 | 64.01 | 64.01 | -2.04 | 4.16 | 0.076 |
t | 1 | 51.34 | 25.67 | 1.24 | 6589.49 | 6589.49 | 20.69 | 428.14 | 0.000 |
T | 1 | 23.00 | 11.50 | 1.24 | 1322.5 | 1322.50 | 9.27 | 85.93 | 0.000 |
Square | 3 | 244.57 | 81.52 | 5.30 | 0.026 | ||||
MPS^{2} | 1 | 1.24 | 0.62 | 2.39 | 1.02 | 1.02 | 0.26 | 0.07 | 0.803 |
t^{2} | 1 | -16.96 | -8.48 | 2.39 | 193.13 | 193.13 | -3.54 | 12.55 | 0.008 |
T^{2} | 1 | 2.34 | 1.17 | 2.39 | 3.66 | 3.66 | 0.49 | 0.24 | 0.639 |
2-Way Interactions | 3 | 58.75 | 19.58 | 1.27 | 0.348 | ||||
MPS×t | 1 | 0.67 | 0.34 | 1.39 | 0.91 | 0.91 | 0.24 | 0.06 | 0.814 |
MPS×T | 1 | -0.18 | -0.09 | 1.39 | 0.06 | 0.06 | -0.06 | 0.00 | 0.951 |
t×T | 1 | 5.37 | 2.69 | 1.39 | 57.78 | 57.78 | 1.94 | 3.75 | 0.089 |
Error | 8 | 123.13 | 15.39 | ||||||
Lack of fit | 5 | 121.74 | 24.35 | 52.74 | 0.004 | ||||
Pure Error | 3 | 1.39 | 0.46 | ||||||
Total | 19 | 8408.21 |
TABLE 7: REGRESSION COEFFICIENTS AND ANOVA OF PREDICTED SECOND-ORDER POLYNOMIAL MODEL FOR Cl_{2} CONVERSION AT CONTINUOUS MODE
Source | Degree of freedom | Effect | Coefficient | SE Coefficient | Adj SS | Adj MS | t-value | f-value | p-value |
Model | 11 | 11289.9 | 1026.35 | 1302.4 | 0.000 | ||||
constant | 79.738 | 0.343 | 232.51 | ||||||
blocks | 2 | 1.8 | 0.92 | 1.16 | 0.360 | ||||
Linear | 3 | 9936.3 | 3312.11 | 4202.94 | 0.000 | ||||
MPS | 1 | -4.180 | -2.090 | 0.281 | 43.7 | 43.68 | -7.45 | 55.43 | 0.000 |
C | 1 | -8.560 | -4.280 | 0.281 | 183.2 | 183.18 | -15.25 | 232.45 | 0.000 |
T | 1 | 62.320 | 31.160 | 0.281 | 9709.5 | 9709.46 | 111.00 | 12320.93 | 0.000 |
Square | 3 | 1122.2 | 374.06 | 474.67 | 0.000 | ||||
MPS^{2} | 1 | 0.445 | 0.223 | 0.542 | 0.1 | 0.13 | 0.41 | 0.17 | 0.692 |
C^{2} | 1 | -1.855 | -0.927 | 0.542 | 2.3 | 2.31 | -1.71 | 2.93 | 0.125 |
T^{2} | 1 | -32.055 | -16.027 | 0.542 | 689.5 | 689.52 | -29.58 | 874.97 | 0.000 |
2-Way Interactions | 3 | 42.6 | 14.19 | 18.00 | 0.001 | ||||
MPS×C | 1 | -0.525 | -0.262 | 0.314 | 0.6 | 0.55 | -0.84 | 0.70 | 0.427 |
MPS×T | 1 | -0.274 | -0.138 | 0.314 | 0.2 | 0.15 | -0.44 | 0.19 | 0.673 |
C×T | 1 | 4.575 | 2.287 | 0.314 | 41.9 | 41.86 | 7.29 | 53.12 | 0.000 |
Error | 8 | 6.3 | 0.79 | ||||||
Lack of fit | 5 | 1.1 | 0.21 | 0.12 | 0.977 | ||||
Pure Error | 3 | 5.2 | 1.74 | ||||||
Total | 19 | 11296.1 |
The significance of model coefficients was determined by absolute p-value and t-value which were listed in Table 6 and 7. A high t value and a small p-value for any of the terms in the model would be an indication of more significant influence on the response variables. Model terms with p < 0.001, 0.001 ≤ p < 0.05 and p ≥ 0.05 are highly significant, significant and insignificant, respectively. In the semi-batch mode, the second-order polynomial model was determined for zirconia conversion as the function of independent parameters as follows:
The results showed that linear terms of temperature and time were highly significant (p < 0.001) while mean particle size (MPS) was insignificant (p ≥ 0.05). The quadratic term of time was significant (0.001 ≤ p < 0.05) while the quadratic terms of temperature and MPS were insignificant (p ≥ 0.05). Linear interaction between time and temperature (t×T) was significant, but the linear interactions between time and MPS (t×MPS) and between temperature and MPS (T×MPS) were insignificant. The fitted model (Eq. (4) was verified via ANOVA which has been shown in Table 6. Predicted values (RSM) versus experimental (observed) values in the semi-batch mode have been presented in Fig. 5. The R^{2} and Adj. R^{2} were 99.54% and 96.52%, respectively.
FIG. 5: PREDICTED VALUES (RSM) VERSUS EXPERIMENTAL (OBSERVED) VALUES OF ZIRCONIA CONVERSION IN SEMI-BATCH MODE
In the continuous mode, the second-order polynomial model was determined for chlorine conversion as a function of independent parameters in Eq. (5):
According to the obtained results from the ANOVA in Table 7, all linear terms of temperature, concentration and MPS were highly significant (p < 0.001). The quadratic term of temperature was highly significant while the quadratic terms of other variables were non- significant.
FIG. 6: PREDICTED VALUES (RSM) VERSUS EXPERIMENTAL (OBSERVED) VALUES OF CHLORINE CONVERSION IN CONTINUOUS MODE
Linear interaction between Cl_{2} concentration and temperature was significant, but the linear interactions between other variables were non- significant. ANOVA (Table 7) verified the fitted model and the R^{2} and Adj. R^{2} were determined to be 99.94% and 99.87%, respectively. Fig. 6 shows the predicted values (RSM) versus experimental (observed) values in the continuous mode.
Optimization: The optimum zirconia conversion of 92.7% in semi-batch mode was determined by RSM at 70µm, 1200 °C, and 90 min. RSM predictions in the continuous mode showed that the maximum Cl_{2} conversion of 98.26% was achieved at the optimal operating conditions of 70 µm, 1380 °C and 3mol/m^{3}. The precision of the modelling Cl_{2} conversion was accredited with triplicate experiments at the optimum operating conditions giving the average conversion of 98 ± 2%. Fig. 7 and 8 show the optimal operating condition for zirconia and Cl_{2} conversion based on coded variables, respectively.
FIG. 7: OPTIMUM OPERATING CONDITIONS ACCORDING TO THE CODED VARIABLES BASED ON RSM FOR SEMI-BATCH MODE
FIG. 8: OPTIMUM OPERATING CONDITIONS ACCORDING TO THE CODED VARIABLES BASED ON RSM FOR CONTINUOUS MODE
Response Surface Analysis:
Semi-Batch Mode: Fig. 9 shows the effects of mean particle size and reaction temperature on zirconia conversion at the fixed reaction time (60 min). The trend of Fig. 9 illustrates that decreasing MPS from 130 µm (code = 1) to 70µm (code = −1) increases the conversion due to the availability of larger surface area for reaction. But it is necessary to realize that there is a limit in the reduction of particle size due to the lack of bubbling fluidization and excessive entrainment in the reactor.
FIG. 9: THE EFFECTS OF MEAN PARTICLE SIZE AND TEMPERATURE (CODED VALUES) ON ZIRCONIA CONVERSION AT THE FIXED REACTION TIME OF 60 min IN SEMI-BATCH MODE, (A) RESPONSE SURFACE PLOT, (B) CONTOUR PLOT
As indicated in Table 6, MPS has a negative linear effect on zirconia conversion (p ˃ 0.05), due to the increase of reaction surface area by the decreased zirconia particle size. Temperature is an important operational parameter and it is necessary to establish the optimum temperature for the process. As summarized in Table 6, the temperature has positive linear (p < 0.001) and quadratic (p > 0.05) effects, while the interaction of temperature MPS showed insignificant influence on zirconia conversion (p = 0.951). Increasing temperature causes higher reaction rate and consequently results in higher zirconia consumption and enhance zirconia conversion.
Fig. 10 depicts the effects of reaction time and MPS on zirconia conversion at 1000 °C. Increasing the chlorination time from 30 min (code = −1) to 90 min (code = 1) was found to be directly proportional to the amount of reacted zirconia.
FIG. 10: THE EFFECTS OF MEAN PARTICLE SIZE AND REACTION TIME (CODED VALUES) ON ZIRCONIA CONVERSION AT THE TEMPERATURE OF 1000 °C IN SEMI-BATCH MODE, (A) RESPONSE SURFACE PLOT, (B) CONTOUR PLOT
Table 6 shows a positive linear effect of reaction time (p < 0.001) on the response variable at low levels of time. Also, at higher levels of chlorination reaction times, the quadratic negative effect becomes significant (0.001 ≤ p < 0.05) which results in increasing the zirconia conversion by lower slope in comparison with lower levels of time.
Moreover, interaction between MPS and time had no significant effect on conversion (p = 0.814).
The explained results for increasing reaction time and temperature are also obvious in Fig. 11 that illustrates the effect of reaction time and temperature on the zirconia conversion at the fixed MPS of 100 µm.
FIG. 11: THE EFFECTS OF REACTION TIME AND TEMPERATURE (CODED VALUES) ON ZIRCONIA CONVERSION AT THE FIXED MPS OF 100 µm IN SEMI-BATCH MODE, (A) RESPONSE SURFACE PLOT, (B) CONTOUR PLOT
ANOVA in Table 6 showed that the interaction between time and temperature is non-significant (p > 0.05). The interaction between different parameters for the carbochlorination reaction of zirconia was illustrated in Fig. 12.
FIG. 12: INTERACTION PLOTS FOR ZIRCONIA CONVERSION BASED ON RESPONSE SURFACE METHODOLOGY
Continuous Mode: The interactions among operating variables (T, C and MPS) were investigated by drawing the Cl_{2} conversion against two independent variables (x and y coordinates), and the other variable was kept constant at zero level according to Eq. (5).
The effect of inlet gas (Cl_{2}) concentration and reaction temperature on the Cl_{2}conversionwas depicted in Fig. 13 at the fixed MPS of 1000 µm. Increasing temperature and inlet Cl_{2 }concentration cause higher reaction rate and consequently results in higher Cl_{2 }consumption. At constant temperature, if the inlet chlorine concentration increases from 3 mol/m^{3 }(coded = -1) to 5 ml/m^{3} (coded = 1), higher zirconia consumption is obtained but at the same time the chlorine conversion is decreased which is the consequence of excess amount of chlorine in the reactor.
FIG. 13: THE EFFECTS OF Cl_{2 }CONCENTRATION AND REACTION TEMPERATURE (CODED VALUES) ON Cl_{2 }CONVERSION AT THE FIXED MPS OF 100 µm IN THE CONTINUOUS MODE, (A) RESPONSE SURFACE PLOT, (B) CONTOUR PLOT
Table 7 shows that Cl_{2 }concentration in the reactor feed has highly significant negative linear effect and non-significant negative quadratic effect on the Cl_{2} conversion. This effect may also be explained due to the reaction rate with respect to chlorine (about 0.7) in which the increase in reaction rate is not considerable and a high amount of inlet Cl_{2 }leaves the reactor unreacted. Also, the model analysis verified that the dependency of the Cl_{2} conversion to the interaction between Cl_{2} concentration and temperature (C×T) is highly significant. The effect of reaction temperature and time on the Cl_{2 }conversion at the fixed Cl_{2 }concentration of 4 ml/m^{3 }has been shown in Fig. 14. Increasing temperature from 1000 °C (coded = -1) to 1400 °C (coded = 1) increases the Cl_{2} conversion due to the increased reaction rate.
FIG. 14: THE EFFECT OF REACTION TEMPERATURE AND MPS ON THE Cl_{2 }CONVERSION AT THE FIXED Cl_{2 }CONCENTRATION OF 4 ml/m^{3} IN THE CONTINUOUS MODE, (A) RESPONSE SURFACE PLOT, (B) CONTOUR PLOT
Table 7 shows that temperature has a highly significant linear positive effect and a highly significant quadratic negative effect at the higher levels of temperature which results in minor changes to the conversion by increasing the reaction temperature. Increasing the Mean particle size of powdered feeds from 70 µm (coded = -1) to 130 µm (coded = 1) results in decreased conversion similar to the semi-batch mode. In this case MPS had a highly significant linear negative effect and a non-significant quadratic positive effect (Table 7). The interaction between MPS and temperature was also insignificant (p = 0.673). The Trends of MPS and Cl_{2} concentration on Cl_{2} conversion at the fixed reaction temperature of 1200 °C which is depicted in Fig. 15 are similar to the above explained influences of these two parameters. Also, the interaction between MPS and Cl_{2} concentration was insignificant for Cl_{2} conversion (p = 0.427).
FIG. 15: THE EFFECT OF MPS AND Cl_{2 }CONCENTRATION ON Cl_{2 }CONVERSION AT THE FIXED REACTION TEMPERATURE OF 1200 °C IN THE CONTINUOUS MODE, (A) RESPONSE SURFACE PLOT, (B) CONTOUR PLOT
The overall interaction between different operating parameters on the mean reaction conversion has been shown in Fig. 16. Fig. 17 shows the XRD analysis of the produced ZrCl_{4} in the carbochlorination process.
FIG. 16: INTERACTION PLOTS FOR CHLORINE CONVERSION BASED ON RESPONSE SURFACE METHODOLOGY IN CONTINUOUS MODE
FIG. 17: XRD OF THE PRODUCED ZrCl_{4} IN THE FLUIDIZED BED REACTOR
CONCLUSION: The optimal conditions for carbochlorination of zirconia in a pilot plant fluidized bed reactor were investigated in semi-batch and continuous mode. Optimization of this process was carried out via RSM and experimental investigation. RSM showed that the data were appropriately fitted to second-order polynomial model. In the continuous method, all linear parameters (MPS, C and T), quadratic term of temperature and linear interaction between Cl_{2} concentration and temperature were significant parameters on the reaction conversion. The obtained results indicated that higher temperature and reaction time as well as lower MPS, increases reaction conversion in the semi-batch mode. In the continuous method, linear terms of temperature and time, the quadratic term of time and the linear interaction between time and temperature were significant parameters.
ACKNOWLEDGEMENT: The financial support provided by Isfahan University of Technology is gratefully acknowledged.
CONFLICT OF INTEREST: There is no any conflict of interest.
REFERENCES:
- Garvie RC and Nicholson PS: Phase analysis in zirconia systems. Journal of the American Ceramic Society 1972; 55: 303-5.
- Sickafus KE, Matzke H, Hartmann T, Yasuda K, Valdez JA, Chodak III P, Nastasi M and Verrall RA: Radiation damage effects in zirconia. Journal of Nuclear Materials 1999; 274: 66-77.
- Jokar AR, Davarpanah MR and Ghamsari MK: Measurement of effective neutron cross section of the reaction 98 Mo (n, γ) 99 Mo in TRR nuclear reactor by using thermal neutrons. Annals of Nuclear Energy 2014; 63: 653-7.
- Mansy MS, Adib M, Habib N, Bashter II, Morcos HN and El-Mesiry MS: Slow neutron total cross-section, transmission and reflection calculation for poly-and mono-NaCl and PbF 2 crystals. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 2016; 384: 14-22.
- Leibowitz L, Blomquist RA and Pelton AD: Thermodynamics of the uranium-zirconium system. Journal of Nuclear Materials 1989; 167: 76-81.
- Videm K: Uniform corrosion and hydrogen pick-up of zirconium-base cladding materials in boiling water reactors. Nuclear Engineering and Design 1975; 33: 170-7.
- Lucas GE and Bement AL: The effect of a zirconium strength differential on cladding collapse predictions. Journal of Nuclear Materials 1975; 55: 246-52.
- Yao MY, Zhou BX, Li Q, Liu WQ and Chu YL: The effect of alloying modifications on hydrogen uptake of zirconium-alloy welding specimens during corrosion tests. Journal of nuclear materials 2006; 350: 195-201.
- Movahedian A, Raygan S and Pourabdoli M: The chlorination kinetics of zirconium dioxide mixed with carbon black. Thermochimica acta 2011; 512: 93-7.
- Jena PK, Brocchi EA and Reis ML: Kinetics of chlorination of zirconia in mixture with petroleum coke by chlorine gas. Metallurgical and Materials Transactions B 1999; 30: 375-81.
- Bicerolu O and Gauvin WH: The chlorination kinetics of zirconium dioxide in the presence of carbon. The Canadian Journal of Chemical Engineering 1980; 58: 357-66.
- Bohé AE and Pasquevich DM: An approach to the intrinsic kinetics of the zirconium oxide carbochlorination. Berichte der Bunsengesellschaft für physikalische Chemie 1995; 99: 1553-8.
- Pan H, Chen XZ, Liang XF, Zhu LT and Luo ZH: CFD simulations of gas–liquid–solid flow in fluidized bed reactors-A review. Powder Technology 2016; 299: 235-58.
- O'Reilly AJ, Doig ID and Ratcliffe JS: The kinetics of the chlorination of zirconium dioxide in a static bed with carbon and chlorine. Journal of Inorganic and Nuclear Chemistry 1972; 34: 2487-95.
- Landsberg A, Hoatson CL and Block FE: The chlorination kinetics of zirconium dioxide in the presence of carbon and carbon monoxide. Metallurgical and Materials Transactions B 1972; 3: 521-7.
- Davoodi P, Ghoreishi SM and Hedayati A: Optimization of supercritical extraction of galegine from Galega officinalis: Neural network modeling and experimental optimization via response surface methodology. Korean Journal of Chemical Engineering 2017; 34: 854-65.
- Ghoreishi SM, Hedayati A and Mohammadi S: Optimization of periodic static-dynamic supercritical CO_{2} extraction of taxifolin from Pinus nigra bark with ethanol as entrainer. The Journal of Supercritical Fluids 2016; 113: 53-60.
- Ghoreishi SM, Hedayati A and Mousavi SO: Quercetin extraction from Rosa damascena Mill via supercritical CO_{2}: Neural network and adaptive neuro fuzzy interface system modeling and response surface optimization. The Journal of Supercritical Fluids 2016; 112: 57-66.
- Hedayati A and Ghoreishi SM: Artificial Neural Network and Adaptive Neuro-Fuzzy Interface System Modeling of Supercritical CO2 Extraction of Glycyrrhizic Acid from Glycyrrhiza glabra Chemical Product and Process Modeling 2016; 11: 217-30.
- Hedayati A and Ghoreishi SM: Supercritical carbon dioxide extraction of glycyrrhizic acid from licorice plant root using binary entrainer: experimental optimization via response surface methodology. The Journal of Supercritical Fluids 2015; 100: 209-17.
- Ghoreishi SM, Hedayati A and Kordnejad M: Micronization of chitosan via rapid expansion of supercritical solution. The Journal of Supercritical Fluids 2016; 111: 162-70.
- Ghoreishi SM, Hedayati A and Ansari K: Experimental investigation and optimization of supercritical carbon dioxide extraction of toxic heavy metals from solid waste using different modifiers and chelating agents. The Journal of Supercritical Fluids 2016; 117: 131-7.
- Gupta VK, Agarwal S, Asif M, Fakhri A and Sadeghi N: Application of response surface methodology to optimize the adsorption performance of a magnetic graphene oxide nano-composite adsorbent for removal of methadone from the environment. Journal of Colloid and Interface Science 2017; 497: 193-200.
- Ghoreishi SM, Mortazavi SM and Hedayati A: Modeling of non-catalytic supercritical water oxidation of phenol. Chemical Product and Process Modeling 2015; 10: 243-51.
How to cite this article:
Ghasemi MR and Ghoreishi SM: Semi-batch and continuous carbochlorination of zirconia in a pilot plant fluidized bed reactor: Optimization of operating conditions via response surface methodology. Int J Pharm Sci Res 2017; 8(12): 5043-56.doi: 10.13040/IJPSR. 0975-8232.8(12).5043-56.
All © 2013 are reserved by International Journal of Pharmaceutical Sciences and Research. This Journal licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License.
Article Information
11
5043-5056
1271
232
English
IJPSR
M. R. Ghasemi and S. M. Ghoreishi*
Department of Chemical Engineering, Isfahan University of Technology, Isfahan, Iran.
ghoreshi@cc.iut.ac.ir
12 April, 2017
14 June, 2017
29 June, 2017
10.13040/IJPSR.0975-8232.8(12).5043-56
01 December, 2017