- Email: [email protected]
A novel separation process for olefin gas purification: Effect of operating parameters on separation performance and process optimization
Journal of the Taiwan Institute of Chemical Engineers 40 (2009) 511–517
Contents lists available at ScienceDirect
Journal of the Taiwan Institute of Chemical Engineers journal homepage: www.elsevier.com/locate/jtice
A novel separation process for olefin gas purification: Effect of operating parameters on separation performance and process optimization Maryam Takht Ravanchi a,b, Tahereh Kaghazchi a,*, Ali Kargari a, Mansoureh Soleimani a a
Department of Chemical Engineering, Center of Excellence for Petrochemical Engineering, Amirkabir University of Technology (Tehran Polytechnic), No. 424, Hafez Ave., PO Box 15875-4413, Tehran, Iran b National Petrochemical Company, Research and Technology Co., No. 12, Sarv Alley, Shirazi South Street, Molla Sadra Ave., PO Box 14358-8471, Tehran, Iran
A R T I C L E I N F O
A B S T R A C T
Article history: Received 20 June 2008 Received in revised form 17 February 2009 Accepted 18 February 2009
Separation of propylene–propane mixtures using facilitated transport membrane is potentially a novel separation process for olefin gas purification. The main purpose of this study was to find optimum values of the process parameters using the Taguchi approach. The Taguchi method was selected as the statistical technique since it allows the main effects to be estimated with a minimum number of experimental runs. Moreover, it makes use of fractional factorial and orthogonal arrays to identify the factors and the optimum factor setting for each experimental run. Trans-membrane pressure and carrier concentration were two influential parameters that affect the separation performance of the present membrane system. These control factors in three levels were considered in the Taguchi analysis. L9 orthogonal array has been used to determine the signal-to-noise (S/N) ratio. Analysis of variance (ANOVA) was used to determine the optimum conditions. It indicated that carrier concentration has the most contribution (72%) in the membrane separation of propylene–propane mixture. Moreover, to achieve an optimum operating condition, trans-membrane pressure and carrier concentration should be set at 120 kPa and 20 wt.%, respectively. According to the Taguchi approach, by setting control factors at optimum values a product with 99.801 (vol.%) propylene was obtained. A verification test was also performed to check the optimum condition. Experimental results confirmed optimum values obtained by the Taguchi analysis and showed that at optimum operating conditions, a product with 99.63 (vol.%) propylene was obtained. ß 2009 Taiwan Institute of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
Keywords: Facilitated transport membrane Propylene Propane Silver nitrate Taguchi analysis ANOVA
1. Introduction In the petrochemical industry, olefins such as ethylene and propylene are the most important chemicals used for the production of polyethylene, polypropylene, styrene, ethyl benzene, ethylene dichloride, acrylonitrile, and isopropanol. Various petrochemical streams contain olefin and other saturated hydrocarbons. These streams typically originate from steam cracking units (ethylene production), catalytic cracking units (motor gasoline production), or the dehydrogenation of paraffins (Agam et al., 2001; Meyers, 1986). An important step in the manufacture of olefins is large-scale separation of the olefin from the corresponding paraffin. During the years, different processes have been used for the separation of olefin–paraffin mixtures, such as low-temperature distillation, extractive distillation, physical or chemical adsorption
* Corresponding author. Tel.: +98 21 64543152; fax: +98 21 66405847. E-mail address: [email protected] (T. Kaghazchi).
and physical or chemical absorption (Bryan, 2004; Eldridge, 1993). Currently, this separation is carried out by cryogenic distillation, which is highly energy-intensive due to the cryogenic temperatures required for the process and low relative volatilities of components. Distillation columns are often up to 300 ft tall and typically contain over 200 trays. With reflux ratios greater than 10, a very high energy input is required for the distillation process. This large capital expense and energy cost have created incentive for extensive research in this area of separations. Nowadays, membrane technologies are becoming more frequently used for the separation of wide varying mixtures in the petrochemical-related industries. The range of applications covers the supply of pure or enriched gases such as He, N2 and O2 from air, the separation of acid gases such as CO2 and H2S, the separation of H2 in the petrochemical and chemical industries, the separation of gold and mercury from industrial wastes and the separation of hydrocarbons. An overview of various types of membrane processes can be found in the literature (Baker, 2002, 2004; Kaghazchi et al., 2006; Kargari et al., 2004a, 2006a,b; Mohammadi et al., 2008; Mulder, 1996; Nabieyan et al., 2007; Takht Ravanchi et al., 2009).
1876-1070/$ – see front matter ß 2009 Taiwan Institute of Chemical Engineers. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.jtice.2009.02.007
M. Takht Ravanchi et al. / Journal of the Taiwan Institute of Chemical Engineers 40 (2009) 511–517
512
Nomenclature C n SP S/N y
concentration number of experiments separation percent signal-to-noise ratio response at each experiment
For the separation of olefin–paraffin mixtures, membrane separation technology has been proposed as an alternative to distillation because of its low cost and simple operation. However, the separation of olefin/paraffin mixtures using conventional polymeric membranes has not been effective because the physicochemical properties of olefins and paraffins such as their molecular size and solubility are largely indistinguishable (Kim et al., 2004; Semenova, 2004). Results obtained by researchers (Bai et al., 1998; Burns and Koros, 2003; Krol et al., 2001; Sridhar and Khan, 1999; Staudt-Bickel and Koros, 2000; Tanaka et al., 1996) show that the application of such membranes is not attractive for industrial purposes because of the relatively low separation factors obtained. Currently, the best polymeric membranes exhibit propylene/ propane separation factors of less than 10 which is still insufficient for practical use. A very effective way to increase the membrane selectivity is to incorporate some carrier in the membrane. This type of membrane is called facilitated transport membrane. The most commonly used facilitated transport membrane types are immobilized liquid membranes (ILMs) and solvent-swollen fixed-site carrier membranes. ILMs, which are also named supported liquid membranes (SLMs), are made by impregnating a microporous membrane with a solution containing the carrier. The carrier solution is held within the pores of the membrane by capillary forces (Cussler, 1994; Teramoto et al., 1986). Nonporous solvent-swollen fixed-site carrier membranes have been used to improve the mechanical stability of facilitated transport membranes. Water-swollen fixedsite carrier membranes exhibit excellent separation properties. However, unless water and/or plasticizers, such as glycerin, are present in the polymeric membrane matrix, tight ion-paring of the silver salt occurs, resulting in very low mobility of the carrier species and, consequently, very low gas fluxes. As a result, fixedsite carrier membranes must be operated continuously in a watervapor-saturated environment. Solid polymer electrolytes are an alternative type of facilitated transport membrane (Pinnau et al., 1997). In these systems, the salt dissolves in a polymer matrix in the solid state and dissociates into anions and cations. The essential feature that distinguishes a solid polymer electrolyte from polymer/salt systems based on fixed-site carrier membranes is that ionic motions in a solid polymer electrolyte take place without a solvent or plasticizer being present (Kang et al., 2006). Some of the results of propylene–propane separation using solid polymer electrolyte membranes are summarized in Table 1. Separation of propylene–propane mixture using a supported liquid membrane, which is the subject of the present study, is a noble research area. In facilitated transport membranes, trans-membrane pressure and carrier concentration are two significant parameters that control the membrane separation performance. In this paper, the effect of these influential parameters on the membrane performance for the separation of propylene–propane mixtures was studied experimentally and via Taguchi approach. As most of the petrochemical propylene–propane mixtures are 50:50 (vol.%), in this study, separation of a 50:50 (vol.%) propylene–propane at room temperature (298 5 K) was investigated.
Table 1 C3H6/C3H8 permeation data in solid polymer electrolyte membranes. Membranea
Separation factor
Reference
PMMA/AgBF4 PBMA/AgBF4 PVP/AgBF4 POZ/AgBF4 PVMK/AgBF4 PEOx/AgBF4 PVA/AgSbF6 PAAm/AgBF4
40b 35b 50b 48b 54b 58b 130 170b
Kim et al. (2003) Kim et al. (2003) Kim et al. (2003) Kim et al. (2003) Kim et al. (2000) Yoon et al. (2000) Kim et al. (2002) Park et al. (2001)
a A detailed chemical description of the abbreviations used for polymer materials is given in Takht Ravanchi et al. (2009). b Feed mixture: 50 vol.% propylene, 50 vol.% propane.
2. Facilitated transport mechanism It is reported that some transition metals react reversibly and selectively with olefins in the solution. The ability of the transition metal ion as a carrier is largely dependant on the intensity of the p complexation with olefin. Metal electronegativity and lattice energy of the transition metal salt are two parameters that affect the intensity of this reversible reaction. If the electronegativity of the metal is so high, the metal is not suitable for the facilitated carrier because it is susceptible to irreversible reaction with the p electrons of the alkene. If the electronegativity is too low, the metal is impractical as a carrier due to its weak interaction with the alkene. It was investigated that the preferred electronegativity for metal is in the range of 1.6–2.3 (Kang et al., 2001; Takht Ravanchi et al., 2009). To increase the reversible reactivity of the transition metal ion with the alkene, the anion of the transition metal plays an important role in determining the intensity and the rate of the interaction between a carrier and alkene. With the lower lattice energy of the transition metal salt, the anion forms a weak ionic bond or ion pair with the cation. It is preferable to select the anion of the transition metal salt that has low lattice energy with respect to the cation. It is reported that the preferable metal salt for facilitated transport must have lattice energy of less than 2500 kJ/mol, by which the tendency of the interaction between anion and cation in the metal salt can be reduced. Facilitated transport of olefin is shown in Fig. 1. Olefin is complexed with a complexing agent (Ag+) incorporated in the membrane at high-pressure side. The complex diffuses owing to its concentration difference across the membrane from the highpressure side to the low-pressure side, where decomplexation
Fig. 1. Facilitated transport mechanism in membrane separation.
M. Takht Ravanchi et al. / Journal of the Taiwan Institute of Chemical Engineers 40 (2009) 511–517
513
Table 2 The detailed analysis of feed gases. Component
Propylene
Propane
N2 CH4 C2H6 C2H4 C3H8 C3H6
– <1 ppm <1 ppm <1 ppm 2622 ppm 99.74 mole%
– 5 ppm 545 ppm 1 ppm 99.79 mole% 731 ppm
takes place to release the olefin. The complexing agent regenerated from the decomplexation, diffuses back to the high-pressure side due to its concentration difference between the low- and highpressure sides. This completes a facilitated transport cycle, and the complexing agent repeats the cycles. Because of the complexation, the concentration of the olefin in the membrane is increased, and the transport of the olefin is thus facilitated. On the other hand, paraffin cannot complex with the complexing agent, and the majority of the paraffin is thus rejected by the membrane. The concentration of the paraffin in the membrane is small by physical solubility, and its transport rate through the membrane is thus low. Therefore, the facilitated transport membrane can give a high olefin/paraffin selectivity to yield permeate with high olefin purity. Among various transition metal ions, silver ion was selected as propylene carrier according to two criteria, i.e. its electronegativity and its salt’s lattice energy. As the main goal of this paper is to propose a substitute process for cryogenic distillation of propylene–propane mixture and this process must be economically feasible, silver nitrate (AgNO3) was selected as carrier salt for propylene.
Schematic diagram of the membrane module is shown in Fig. 2. The lower and upper compartments, with inner diameter 110 mm, were made from stainless steel (AISI 316). Polyvinilydene diflouride (PVDF) flat sheet membranes (Durapore from Millipore, thickness 125 mm, pore size 0.22 mm, filter diameter 142 mm) were used as the support of the liquid membrane. After being immersed in the carrier solution, the membrane filter was sandwiched between two compartments of the module. Once prepared, the membrane filter could be used for 3–4 weeks with no change in separation and permeation properties.
3. Experiments
3.3. Experimental setup
3.1. Chemicals
The schematic diagram of the experimental setup is shown in Fig. 3. All tubing used to connect all parts of the setup was stainless steel (AISI 316). The experimental procedure is as follows. Propylene and propane, after passing through mass flow controllers (Brooks Instruments, model 5850S), were mixed and entered the humidifier. The humidified feed passes through a temperature control system and enters the membrane cell. A combination of a heater and a cooler were used as the temperature control system. The feed gas was introduced to the upper compartment of the cell and the sweep gas, nitrogen, was supplied to the lower compartment. The main product, permeate, was
Industrial grade propylene from Tabriz Petrochemical Company and industrial grade propane from Tehran Refinery Complex were used as feed gases and pure nitrogen was used as sweep gas. The detailed analysis of the feed gases is presented in Table 2. Silver nitrate (AgNO3, GR Pro Analysis) which was purchased from Merck Co. was used as the carrier of propylene. An aqueous solution of silver nitrate was prepared by dissolving silver nitrate in deionized water. All chemicals were used without further purification.
Fig. 2. Schematic diagram of membrane module.
3.2. Membrane module
Fig. 3. Schematic diagram of the supported liquid membrane system.
514
M. Takht Ravanchi et al. / Journal of the Taiwan Institute of Chemical Engineers 40 (2009) 511–517
collected from the lower compartment and the secondary product, retentate, was collected from the upper compartment. A back pressure regulator (BPR, Tescom, Germany) was used on the retentate line to control the pressure of the system. During all experiments, sweep gas was at atmospheric pressure. The experiments were conducted at room temperature (298 5 K). All the experimental data were obtained after an initial permeation period of 4–6 h. Each experiment was repeated two times. 3.4. Analysis The gas composition was determined by a gas chromatograph (Agilent 6890N) equipped with a flame ionization detector (FID, Agilent Technologies Inc., column, HP Al/S, 0.53 mm in diameter, and 50 m in length).
4. Design of experiments
Fig. 4. Flowchart of the Taguchi method (Nikbakht et al., 2007).
- conducting the array experiment and - analyzing and verifying the results.
Design of experiments is an invaluable tool for identifying critical parameters, optimizing chemical processes and identifying operating regions for the process (Godbert, 2000). The Taguchi method is a powerful problem solving technique for improving process performance. It reduces scrap rates, rework costs and manufacturing costs due to excessive variability in processes (Antony and Antony, 2001). The techniques for laying out experiments when multiple factors are involved, has been known for a long time and is popularly known as the factorial design of experiments. This method helps researchers to determine the possible combinations of factors and to identify the best combination. However, in industrial settings, it is extremely costly to run a number of experiments to test all combinations. The Taguchi approach developed rules to carry out the experiments, which further simplified and standardized the design of experiments, along with minimizing the number of factor combinations that would be required to test the factor effects. So the Taguchi method has been chosen for this study. The Taguchi method was developed by Genichi Taguchi between 1950 and 1960. This method (Peace, 1993; Taguchi, 1990) is a systematic application of design and analysis of experiments for the purpose of designing and improving product quality. In recent years, the Taguchi method has become a powerful tool for improving productivity during research and development so that, high quality products can be produced quickly and at low cost. Optimization of process parameters is the key step in the Taguchi method in achieving a high quality without increasing the cost. This is because optimization of process parameters can improve performance characteristics and the optimal process parameters obtained from the Taguchi method are insensitive to the variation of environmental conditions and other noise factors. Basically, classical process parameter design is complex and not easy to use. Especially, a large number of experiments have to be carried out when the number of the process parameters increases. To solve this task, the Taguchi method uses a special design of orthogonal arrays to study the entire process parameter space with a small number of experiments only (Kargari et al., 2004b). In the Taguchi approach, orthogonal arrays and analysis of variance (ANOVA) are used as the tools of analysis. A brief overview of the process followed by the Taguchi’s approach to parameter design is provided in Fig. 4. These steps can be grouped as
In separation processes, separation percent is the most important criteria by which process performance can be evaluated. Thus, in the present study it is chosen as quality characteristic. Noise factors are those parameters which are either uncontrollable or are too expensive to control such as variation of environmental operating conditions. Control parameters are those design factors that can be set and maintained. The levels for each control parameter must be chosen at this point. The number of levels for each control parameter defines the experimental region. The array experiment is designed by selecting the appropriate orthogonal array for the control parameters. Orthogonal array for a particular project depends on the number of factors and their levels. In order to analyze the results, the Taguchi method uses a statistical measure of performance called signal-to-noise (S/N) ratio. The S/N ratio can be used to determine the product quality and to compare the product performance. It can be combined with the orthogonal array for the design of experiments to improve product and process. In its simplest form, the S/N ratio is the ratio of the mean (signal) to the standard deviation (noise). The Taguchi method uses the S/N ratio to measure the separation percent deviation from the desired value. There are three standard S/N ratios, i.e., bigger-is-better, smaller-is-better and nominal-is-best quality characteristics. As in the separation of propylene–propane mixture propylene is the desirable product, and S/N ratio was calculated on the basis of propylene concentration in the product, the more the S/N ratio the better the separation performance. For the larger the better responses, as in this study, the following relation is used to calculate S/N ratio: ! n 1X 1 S=N ¼ 10 log (2) n i¼1 y2i
- planning an array experiment to determine the effects of the control factors,
In the Taguchi analysis for experimental design, important factors that have influence on process and their levels must be
Quality characteristic is the output or the response variable to be observed. The quality characteristic of this study is separation percent (SP) which is defined as below: SP ¼
C permeate C feed C feed
(1)
where n is the number of experiments and yi is the response at each experiment (Montgomery, 1991). 5. Results and discussion 5.1. Taguchi approach
M. Takht Ravanchi et al. / Journal of the Taiwan Institute of Chemical Engineers 40 (2009) 511–517 Table 3 Separation factors and their levels. Factor
Unit
Level 1
Level 2
Level 3
Trans-membrane pressure Carrier concentration
kPa wt.%
50 5
100 10
120 20
determined at first. Trans-membrane pressure and carrier concentration were two influential parameters that affect the separation performance of a facilitated transport membrane. Thus, they were chosen as separation factors. Three levels set for each of these factors, which are demonstrated in Table 3. According to the Taguchi parameter design methodology and with the aid of Qualitek-4 software, one experimental design should be selected for the controllable factors. In the present work, an L9 orthogonal array was used. The experimental layout for the separation process parameters using this orthogonal array is shown in Table 4. Each row of this table represents a run, which is a specific set of factor levels to be tested. Nine experiments with the arrangement of Table 4 were performed. Separation percent was calculated by Eq. (1) and S/N ratio was calculated by Eq. (2). The results are shown in Table 5. As it can be seen in this table, when trans-membrane pressure was kept constant and carrier concentration was increased, S/N ratio was increased (experiments 1, 2, 3 or 4, 5, 6 or 7, 8, 9). The same trend can be observed when carrier concentration was kept constant and trans-membrane pressure was increased (experiments 1, 4, 7 or 2, 5, 8 or 3, 6, 9). Basically, the larger the response, the better the performance characteristic. As can be seen in Table 5, increasing the trans-membrane pressure and carrier concentration increases S/N ratio and separation percent. The larger the S/N ratio, the larger the contribution of one separation factor at that level for the separation. Moreover, entry No. 9 tops all the other eight entries with regard to separation percent and S/N ratio. Thus entry 9 levels were selected as optimum conditions. In other words, by selection of this entry, the best response with minimum required conditions (cost) will be attainable. The Taguchi analysis of separation data are presented in Table 6 and Figs. 5 and 6. As it can be seen, with an increase in factor levels the Taguchi responses were increased as well. The highest response was obtained when factors are at their highest levels. The Taguchi analysis predicts that optimum conditions at which highest product concentration obtained is when transmembrane pressure and carrier concentration are in their highest values (i.e. 120 kPa and 20 wt.% respectively). According to this analysis, at optimum condition propylene concentration in the product will be 99.801 vol.%. 5.2. Experimental results According to the Taguchi’s experimental layout (Table 4), some experiments were conducted the results of which are demon-
1 2 3 4 5 6 7 8 9
Table 5 The separation percent and the signal to noise ratio based upon experimental results. Experiment number
Separation percent
S/N ratio (db)
1 2 3 4 5 6 7 8 9
95.60 96.72 98.94 96.96 98.22 99.20 97.20 98.48 99.26
39.81 39.86 39.95 39.87 39.92 39.96 39.88 39.93 39.97
strated in Fig. 7. As can be seen in this figure, when transmembrane pressure and carrier concentration are in their highest levels, the highest concentration of propylene can be obtained in the permeate product, i.e. the highest separation percent was obtained. Permeability and selectivity are two important parameters by which the separation performance of a membrane can be evaluated. In Table 7 these values were reported. Facilitated transport is a combination of two processes: absorption (on the feed side) and stripping (on the permeate side). Increasing the pressure is in favor of absorption and decreasing the pressure is in favor of stripping. Thus, increasing the feed pressure increases the absorbed propylene on the feed side. Due to the pressure difference between feed side and permeate side, the complexed propylene is decomplexed on the permeate side. Therefore, the more the trans-membrane pressure, the more the driving force for separation and the more the propylene concentration in the product. This concept is in agreement with what is observed in Fig. 7. In facilitated transport membranes, propylene permeation occurs via two mechanisms: Fickian diffusion and facilitation transport. In the presence of carrier in the membrane system, propylene was permeated via Fickian diffusion and facilitated transport. Based upon facilitated transport mechanism, when more carriers were available in the membrane, more propylene can be transported along the membrane thickness, propylene concentration in the product was increased and separation performance was improved. The same results were observed in Fig. 7, i.e. when a 20 wt% carrier solution was used in the membrane, highest propylene concentration was obtained. According to Fig. 7, when trans-membrane pressure and carrier concentration are at their highest levels (i.e. 120 kPa and 20 wt.%, respectively) highest propylene concentration in the product (99.63 vol.%) is obtained which results an error equal to 0.17% from the Taguchi analysis prediction. 5.3. Analysis of variance Taguchi-oriented practitioners often use the analysis of variance to verify the factors that influence the mean response.
Table 4 Experimental layout using L9 orthogonal array. Experiment number
515
Factors Trans-membrane pressure
Carrier concentration
50 50 50 100 100 100 120 120 120
5 10 20 5 10 20 5 10 20
Table 6 The responses for the Taguchi analysis of the separation data. Response
L1 L2 L3 L2 L1 L3 L1 L3 L2
Factors Trans-membrane pressure
Carrier concentration
98.543 99.063 99.156 0.52 0.613 0.093
98.293 98.903 99.566 0.61 1.273 0.663
516
M. Takht Ravanchi et al. / Journal of the Taiwan Institute of Chemical Engineers 40 (2009) 511–517 Table 7 Propylene permeability and propylene–propane selectivity. Trans-membrane pressure (kPa)
Propylene permeability (Barrer)
Carrier concentration: 5 wt.% 50 100 120
Fig. 5. Main effect of trans-membrane pressure on product concentration (Taguchi analysis).
Fig. 6. Main effect of carrier concentration on product concentration (Taguchi analysis).
Selectivity
27.26 63.41 68.20
44.45 64.79 70.43
Carrier concentration: 10 wt.% 50 100 120
89.13 139.56 149.56
59.97 111.36 130.58
Carrier concentration: 20 wt.% 50 100 120
260.53 425.05 488.06
187.68 249.00 269.27
Sum of squares (SS), mean square (variance), the ratio of factor variance on error variance (F) and percent of contribution of each factor on the response (P) are important parameters used in analysis of variance. The purpose of the ANOVA is to investigate which process parameters significantly affect the performance characteristic. This is accomplished by separating the total variability of each level, which is measured by the sum of the squared deviations from the total mean of the responses, into contribution by each process parameter and the error. The percentage contribution by each of the process parameters in the total sum of the squared deviations can be used to evaluate the importance of the process parameter change on the performance characteristic. In addition, the F-test named after Fisher (1925) can also be used to determine which parameters have a significant effect on the performance characteristic. Usually, the change of the process parameter has significant effect on the performance characteristic when the F-value is large. Results of ANOVA which are shown in Table 8 indicate that carrier concentration with the contribution of 71.686% on the final response is the most significant process parameter for affecting the performance characteristic. From Fisher tables (Fisher, 1925) with 90% confidence, F10,2,4 = 4.32. According to this value, the F-value for transmembrane pressure and carrier concentration are greater than the corresponding values of Fisher tables, so the tests are reliable with 90% confidence. The contribution of error is 11%, which is in reasonable range of errors. 6. Conclusion
Fig. 7. Effect of trans-membrane pressure on product concentration with carrier concentration as parameter (experimental results).
In the separation of propylene–propane mixture via facilitated transport mechanism, trans-membrane pressure and carrier concentration were two important parameters. The influence of these parameters on the separation performance of a membrane system was investigated experimentally and via the Taguchi approach. Both analyses confirmed that the more the trans-membrane pressure and the more the carrier concentration, the better the separation performance.
Table 8 Statistical results of ANOVA. Factor
DOF
Sum of squares
Variance
F-ratio
Trans-membrane pressure Carrier concentration Other/error
2 2 4
0.656 2.435 0.179
0.328 1.217 0.044
7.311 27.107 –
Total
8
3.271
–
–
Contribution percentage 17.330 71.686 10.984 100.000
M. Takht Ravanchi et al. / Journal of the Taiwan Institute of Chemical Engineers 40 (2009) 511–517
Acknowledgement The authors gratefully acknowledge the financial support received in the form of a research grant (research project No.: 83111149) from the National Petrochemical Company- Research and Technology Center (NPC-RT), Iran. References Agam, G., G. Dagan, J. Gilron, V. Krakov, and N. Tsesin, ‘‘Recovery of Olefins from Gasous Mixtures,’’ WO Patent, 0117664 (2001). Antony, J. and F. J. Antony, ‘‘Teaching the Taguchi Method to Industrial Engineers,’’ Work Study, 50 (4), 141 (2001). Bai, S., S. Sridhar, and A. A. Khan, ‘‘Metal-Ion Mediated Separation of Propylene from Propane Using PPO Membranes,’’ J. Membr. Sci., 147, 131 (1998). Baker, R. W., ‘‘Future Directions of Membrane Gas Separation Technology,’’ Ind. Eng. Chem. Res., 41, 1393 (2002). Baker, R. W., Membrane Technology and Applications, John Wiley, England (2004). Bryan, P. F., ‘‘Removal of Propylene from Fuel-Grade Propane,’’ Separat. Purif. Rev., 33, 157 (2004). Burns, R. L. and W. J. Koros, ‘‘Defining the Challenges for C3H6/C3H8 Separation Using Polymeric Membranes,’’ J. Membr. Sci., 211, 299 (2003). Cussler, E. L., Facilitated and Active Transport, Paul, D.R. & Y.P. Yampolskii (Eds.), CRC Press, Florida, USA (1994). Eldridge, R. B., ‘‘Olefin/Paraffin Separation Technology: A Review,’’ Ind. Eng. Chem. Res., 32, 2208 (1993). Fisher, R. A., Statistical Methods for Research Workers, Oliver and Boyd, London (1925). Godbert, S., ‘‘Experimental Design—Identifying Factors that Affect Responses,’’ In Design and Analysis in Chemical Research, Tranter, R.L. (Ed.), Sheffield Academic Press, England (2000). Kaghazchi, T., A. Kargari, R. Yegani, and A. Zare, ‘‘Emulsion Liquid Membrane Pertraction of L-lysine from Dilute Aqueous Solutions by D2EHPA Mobile Carrier,’’ Desalination, 190, 161 (2006). Kang, Y. S., H. Ch. Park, J. O. Won, S. U. Hong, and T. M. Kwon, ‘‘Facilitated Transport Separation Membranes Using Solid State Polymer Electrolytes’’, U.S. Patent, 2001/ 0015334 (2001). Kang, Y. S., J. H. Kim, J. Won, and H. S. Kim, Solid-State Facilitated Membranes for Separation of Olefins/Paraffins and Oxygen/Nitrogen, Yampolskii, Y., I. Pinnau & B. Freeman (Eds.), John Wiley, England (2006). Kargari, A., T. Kaghazchi, and M. Soleimani, ‘‘Extraction of Gold (III) Ions from Aqueous Solutions Using Polyamine Type Surfactant Liquid Membrane,’’ Can. J. Chem. Eng., 82, 1301 (2004a). Kargari, A., T. Kaghazchi, M. Sohrabi, and M. Soleimani, ‘‘Batch Extraction of Gold (III) Ions from Aqueous Solutions using Emulsion Liquid Membrane via Facilitated Carrier Transport,’’ J. Membr. Sci., 233, 1 (2004b). Kargari, A., T. Kaghazchi, M. Sohrabi, and M. Soleimani, ‘‘Application of Experimental Design to Emulsion Liquid Membrane Pertraction of Gold (III) Ions from Aqueous Solutions,’’ Iran. J. Chem. Eng., 3, 76 (2006a). Kargari, A., T. Kaghazchi, B. Mardangahi, and M. Soleimani, ‘‘Experimental and Modeling of Selective Separation of Gold (III) Ions from Aqueous Solutions by Emulsion Liquid Membrane System,’’ J. Memb. Sci., 279, 389 (2006b).
517
Kim, H. S., J. H. Ryu, H. Kim, B. S. Ahn, and Y. S. Kang, ‘‘Reversible Olefin Complexation by Silver Ions in Dry Poly(vinyl methyl ketone) Membrane and Its Application to Olefin/Paraffin Separations,’’ Chem. Commun., 14, 1261 (2000). Kim, J. H., B. R. Min, K. B. Lee, J. Won, and Y. S. Kang, ‘‘Coordination Structure of Various Ligands in Crosslinked PVA to Silver Ions for Facilitated Olefin Transport,’’ Chem. Commun., 22, 2732 (2002). Kim, J. H., B. R. Min, J. Won, S. H. Joo, H. S. Kim, and Y. S. Kang, ‘‘Role of Polymer Matrix in Polymer–Silver Complexes for Structure, Interaction, and Facilitated Olefin Transport,’’ Macromolecules, 36, 6183 (2003). Kim, J. H., S. M. Park, J. Won, and Y. S. Kang, ‘‘Dependence of Facilitated Olefin Transport on the Thickness of Silver Polymer Electrolyte Membranes,’’ J. Membr. Sci., 236, 209 (2004). Krol, J. J., M. Boerrigter, and G. H. Koops, ‘‘Polymeric Hollow Fiber Gas Separation Membranes: Preparation and the Suppression of Plasticization in Propane–Propylene Environments,’’ J. Membr. Sci., 184, 275 (2001). Meyers, R. A., Handbook of Petroleum Refining Processes, McGraw-Hill, USA (1986). Mohammadi, S., T. Kaghazchi, and A. Kargari, ‘‘A Model for Metal Ion Pertraction through Supported Liquid Membrane,’’ Desalination, 219, 324 (2008). Montgomery, D. C., Design and Analysis of Experiments, John Wiley, New York, USA (1991). Mulder, M., Basic Principles of Membrane Technology, 2nd Ed., Kluwer Academic Publishers, The Netherlands (1996). Nabieyan, B., T. Kaghazchi, A. Kargari, A. Mahmoudian, and M. Soleimani, ‘‘Bench-Scale Simultaneous Extraction and Stripping of Iodine Using Bulk Liquid Membrane System,’’ Desalination, 214, 167 (2007). Nikbakht, R., M. Sadrzadeh, and T. Mohammadi, ‘‘Effect of Operating Parameters on Concentration of Citric Acid Using Electrodialysis,’’ J. Food Eng., 83, 596 (2007). Park, Y. S., J. Won, and Y. S. Kang, ‘‘Facilitated Transport of Olefin through Solid PAAm and PAAm-Graft Composite Membranes with Silver Ions,’’ J. Membr. Sci., 183, 163 (2001). Peace, G. S., Taguchi Methods: A Hand-on Approach, Addison-Wesley, Reading, MA (1993). Pinnau, I., L. G. Toy, and C. S. Casillas, ‘‘Olefin Separation Membrane and Process,’’ U.S. Patent, 5, 670, 051 (1997). Semenova, S. I., ‘‘Polymer Membranes for Hydrocarbon Separation and Removal,’’ J. Membr. Sci., 231, 189 (2004). Sridhar, S. and A. A. Khan, ‘‘Simulation Studies for the Separation of Propylene and Propane by Ethylcellulose Membrane,’’ J. Membr. Sci., 159, 209 (1999). Staudt-Bickel, C. and W. J. Koros, ‘‘Olefin/Paraffin Gas Separations with 6FDA Based Polyimide Membranes,’’ J. Membr. Sci., 170, 205 (2000). Taguchi, G., Introduction to Quality Engineering, Asian Productivity Organization, Tokyo, Japan (1990). Takht Ravanchi, M., T. Kaghazchi, and A. Kargari, ‘‘Application of Membrane Separation Processes in Petrochemical Industry: A Review,’’ Desalination, 235, 199 (2009). Tanaka, K., A. Taguchi, J. Q. Hao, H. Kita, and K. Okamoto, ‘‘Permeation and Separation Properties of Polyimide Membranes to Olefin and Paraffins,’’ J. Membr. Sci., 121, 197 (1996). Teramoto, M., H. Matsuyama, T. Yamashiro, and Y. Katayama, ‘‘Separation of Ethylene from Ethane by Supported Liquid Membranes Containing Silver Nitrate as Carrier,’’ J. Chem. Eng. Jpn., 19, 419 (1986). Yoon, Y., J. Won, and Y. S. Kang, ‘‘Polymer Electrolyte Membranes Containing Silver Ion for Facilitated Olefin Transport,’’ Macromolecules, 33, 3185 (2000).
Contents lists available at ScienceDirect
Journal of the Taiwan Institute of Chemical Engineers journal homepage: www.elsevier.com/locate/jtice
A novel separation process for olefin gas purification: Effect of operating parameters on separation performance and process optimization Maryam Takht Ravanchi a,b, Tahereh Kaghazchi a,*, Ali Kargari a, Mansoureh Soleimani a a
Department of Chemical Engineering, Center of Excellence for Petrochemical Engineering, Amirkabir University of Technology (Tehran Polytechnic), No. 424, Hafez Ave., PO Box 15875-4413, Tehran, Iran b National Petrochemical Company, Research and Technology Co., No. 12, Sarv Alley, Shirazi South Street, Molla Sadra Ave., PO Box 14358-8471, Tehran, Iran
A R T I C L E I N F O
A B S T R A C T
Article history: Received 20 June 2008 Received in revised form 17 February 2009 Accepted 18 February 2009
Separation of propylene–propane mixtures using facilitated transport membrane is potentially a novel separation process for olefin gas purification. The main purpose of this study was to find optimum values of the process parameters using the Taguchi approach. The Taguchi method was selected as the statistical technique since it allows the main effects to be estimated with a minimum number of experimental runs. Moreover, it makes use of fractional factorial and orthogonal arrays to identify the factors and the optimum factor setting for each experimental run. Trans-membrane pressure and carrier concentration were two influential parameters that affect the separation performance of the present membrane system. These control factors in three levels were considered in the Taguchi analysis. L9 orthogonal array has been used to determine the signal-to-noise (S/N) ratio. Analysis of variance (ANOVA) was used to determine the optimum conditions. It indicated that carrier concentration has the most contribution (72%) in the membrane separation of propylene–propane mixture. Moreover, to achieve an optimum operating condition, trans-membrane pressure and carrier concentration should be set at 120 kPa and 20 wt.%, respectively. According to the Taguchi approach, by setting control factors at optimum values a product with 99.801 (vol.%) propylene was obtained. A verification test was also performed to check the optimum condition. Experimental results confirmed optimum values obtained by the Taguchi analysis and showed that at optimum operating conditions, a product with 99.63 (vol.%) propylene was obtained. ß 2009 Taiwan Institute of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
Keywords: Facilitated transport membrane Propylene Propane Silver nitrate Taguchi analysis ANOVA
1. Introduction In the petrochemical industry, olefins such as ethylene and propylene are the most important chemicals used for the production of polyethylene, polypropylene, styrene, ethyl benzene, ethylene dichloride, acrylonitrile, and isopropanol. Various petrochemical streams contain olefin and other saturated hydrocarbons. These streams typically originate from steam cracking units (ethylene production), catalytic cracking units (motor gasoline production), or the dehydrogenation of paraffins (Agam et al., 2001; Meyers, 1986). An important step in the manufacture of olefins is large-scale separation of the olefin from the corresponding paraffin. During the years, different processes have been used for the separation of olefin–paraffin mixtures, such as low-temperature distillation, extractive distillation, physical or chemical adsorption
* Corresponding author. Tel.: +98 21 64543152; fax: +98 21 66405847. E-mail address: [email protected] (T. Kaghazchi).
and physical or chemical absorption (Bryan, 2004; Eldridge, 1993). Currently, this separation is carried out by cryogenic distillation, which is highly energy-intensive due to the cryogenic temperatures required for the process and low relative volatilities of components. Distillation columns are often up to 300 ft tall and typically contain over 200 trays. With reflux ratios greater than 10, a very high energy input is required for the distillation process. This large capital expense and energy cost have created incentive for extensive research in this area of separations. Nowadays, membrane technologies are becoming more frequently used for the separation of wide varying mixtures in the petrochemical-related industries. The range of applications covers the supply of pure or enriched gases such as He, N2 and O2 from air, the separation of acid gases such as CO2 and H2S, the separation of H2 in the petrochemical and chemical industries, the separation of gold and mercury from industrial wastes and the separation of hydrocarbons. An overview of various types of membrane processes can be found in the literature (Baker, 2002, 2004; Kaghazchi et al., 2006; Kargari et al., 2004a, 2006a,b; Mohammadi et al., 2008; Mulder, 1996; Nabieyan et al., 2007; Takht Ravanchi et al., 2009).
1876-1070/$ – see front matter ß 2009 Taiwan Institute of Chemical Engineers. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.jtice.2009.02.007
M. Takht Ravanchi et al. / Journal of the Taiwan Institute of Chemical Engineers 40 (2009) 511–517
512
Nomenclature C n SP S/N y
concentration number of experiments separation percent signal-to-noise ratio response at each experiment
For the separation of olefin–paraffin mixtures, membrane separation technology has been proposed as an alternative to distillation because of its low cost and simple operation. However, the separation of olefin/paraffin mixtures using conventional polymeric membranes has not been effective because the physicochemical properties of olefins and paraffins such as their molecular size and solubility are largely indistinguishable (Kim et al., 2004; Semenova, 2004). Results obtained by researchers (Bai et al., 1998; Burns and Koros, 2003; Krol et al., 2001; Sridhar and Khan, 1999; Staudt-Bickel and Koros, 2000; Tanaka et al., 1996) show that the application of such membranes is not attractive for industrial purposes because of the relatively low separation factors obtained. Currently, the best polymeric membranes exhibit propylene/ propane separation factors of less than 10 which is still insufficient for practical use. A very effective way to increase the membrane selectivity is to incorporate some carrier in the membrane. This type of membrane is called facilitated transport membrane. The most commonly used facilitated transport membrane types are immobilized liquid membranes (ILMs) and solvent-swollen fixed-site carrier membranes. ILMs, which are also named supported liquid membranes (SLMs), are made by impregnating a microporous membrane with a solution containing the carrier. The carrier solution is held within the pores of the membrane by capillary forces (Cussler, 1994; Teramoto et al., 1986). Nonporous solvent-swollen fixed-site carrier membranes have been used to improve the mechanical stability of facilitated transport membranes. Water-swollen fixedsite carrier membranes exhibit excellent separation properties. However, unless water and/or plasticizers, such as glycerin, are present in the polymeric membrane matrix, tight ion-paring of the silver salt occurs, resulting in very low mobility of the carrier species and, consequently, very low gas fluxes. As a result, fixedsite carrier membranes must be operated continuously in a watervapor-saturated environment. Solid polymer electrolytes are an alternative type of facilitated transport membrane (Pinnau et al., 1997). In these systems, the salt dissolves in a polymer matrix in the solid state and dissociates into anions and cations. The essential feature that distinguishes a solid polymer electrolyte from polymer/salt systems based on fixed-site carrier membranes is that ionic motions in a solid polymer electrolyte take place without a solvent or plasticizer being present (Kang et al., 2006). Some of the results of propylene–propane separation using solid polymer electrolyte membranes are summarized in Table 1. Separation of propylene–propane mixture using a supported liquid membrane, which is the subject of the present study, is a noble research area. In facilitated transport membranes, trans-membrane pressure and carrier concentration are two significant parameters that control the membrane separation performance. In this paper, the effect of these influential parameters on the membrane performance for the separation of propylene–propane mixtures was studied experimentally and via Taguchi approach. As most of the petrochemical propylene–propane mixtures are 50:50 (vol.%), in this study, separation of a 50:50 (vol.%) propylene–propane at room temperature (298 5 K) was investigated.
Table 1 C3H6/C3H8 permeation data in solid polymer electrolyte membranes. Membranea
Separation factor
Reference
PMMA/AgBF4 PBMA/AgBF4 PVP/AgBF4 POZ/AgBF4 PVMK/AgBF4 PEOx/AgBF4 PVA/AgSbF6 PAAm/AgBF4
40b 35b 50b 48b 54b 58b 130 170b
Kim et al. (2003) Kim et al. (2003) Kim et al. (2003) Kim et al. (2003) Kim et al. (2000) Yoon et al. (2000) Kim et al. (2002) Park et al. (2001)
a A detailed chemical description of the abbreviations used for polymer materials is given in Takht Ravanchi et al. (2009). b Feed mixture: 50 vol.% propylene, 50 vol.% propane.
2. Facilitated transport mechanism It is reported that some transition metals react reversibly and selectively with olefins in the solution. The ability of the transition metal ion as a carrier is largely dependant on the intensity of the p complexation with olefin. Metal electronegativity and lattice energy of the transition metal salt are two parameters that affect the intensity of this reversible reaction. If the electronegativity of the metal is so high, the metal is not suitable for the facilitated carrier because it is susceptible to irreversible reaction with the p electrons of the alkene. If the electronegativity is too low, the metal is impractical as a carrier due to its weak interaction with the alkene. It was investigated that the preferred electronegativity for metal is in the range of 1.6–2.3 (Kang et al., 2001; Takht Ravanchi et al., 2009). To increase the reversible reactivity of the transition metal ion with the alkene, the anion of the transition metal plays an important role in determining the intensity and the rate of the interaction between a carrier and alkene. With the lower lattice energy of the transition metal salt, the anion forms a weak ionic bond or ion pair with the cation. It is preferable to select the anion of the transition metal salt that has low lattice energy with respect to the cation. It is reported that the preferable metal salt for facilitated transport must have lattice energy of less than 2500 kJ/mol, by which the tendency of the interaction between anion and cation in the metal salt can be reduced. Facilitated transport of olefin is shown in Fig. 1. Olefin is complexed with a complexing agent (Ag+) incorporated in the membrane at high-pressure side. The complex diffuses owing to its concentration difference across the membrane from the highpressure side to the low-pressure side, where decomplexation
Fig. 1. Facilitated transport mechanism in membrane separation.
M. Takht Ravanchi et al. / Journal of the Taiwan Institute of Chemical Engineers 40 (2009) 511–517
513
Table 2 The detailed analysis of feed gases. Component
Propylene
Propane
N2 CH4 C2H6 C2H4 C3H8 C3H6
– <1 ppm <1 ppm <1 ppm 2622 ppm 99.74 mole%
– 5 ppm 545 ppm 1 ppm 99.79 mole% 731 ppm
takes place to release the olefin. The complexing agent regenerated from the decomplexation, diffuses back to the high-pressure side due to its concentration difference between the low- and highpressure sides. This completes a facilitated transport cycle, and the complexing agent repeats the cycles. Because of the complexation, the concentration of the olefin in the membrane is increased, and the transport of the olefin is thus facilitated. On the other hand, paraffin cannot complex with the complexing agent, and the majority of the paraffin is thus rejected by the membrane. The concentration of the paraffin in the membrane is small by physical solubility, and its transport rate through the membrane is thus low. Therefore, the facilitated transport membrane can give a high olefin/paraffin selectivity to yield permeate with high olefin purity. Among various transition metal ions, silver ion was selected as propylene carrier according to two criteria, i.e. its electronegativity and its salt’s lattice energy. As the main goal of this paper is to propose a substitute process for cryogenic distillation of propylene–propane mixture and this process must be economically feasible, silver nitrate (AgNO3) was selected as carrier salt for propylene.
Schematic diagram of the membrane module is shown in Fig. 2. The lower and upper compartments, with inner diameter 110 mm, were made from stainless steel (AISI 316). Polyvinilydene diflouride (PVDF) flat sheet membranes (Durapore from Millipore, thickness 125 mm, pore size 0.22 mm, filter diameter 142 mm) were used as the support of the liquid membrane. After being immersed in the carrier solution, the membrane filter was sandwiched between two compartments of the module. Once prepared, the membrane filter could be used for 3–4 weeks with no change in separation and permeation properties.
3. Experiments
3.3. Experimental setup
3.1. Chemicals
The schematic diagram of the experimental setup is shown in Fig. 3. All tubing used to connect all parts of the setup was stainless steel (AISI 316). The experimental procedure is as follows. Propylene and propane, after passing through mass flow controllers (Brooks Instruments, model 5850S), were mixed and entered the humidifier. The humidified feed passes through a temperature control system and enters the membrane cell. A combination of a heater and a cooler were used as the temperature control system. The feed gas was introduced to the upper compartment of the cell and the sweep gas, nitrogen, was supplied to the lower compartment. The main product, permeate, was
Industrial grade propylene from Tabriz Petrochemical Company and industrial grade propane from Tehran Refinery Complex were used as feed gases and pure nitrogen was used as sweep gas. The detailed analysis of the feed gases is presented in Table 2. Silver nitrate (AgNO3, GR Pro Analysis) which was purchased from Merck Co. was used as the carrier of propylene. An aqueous solution of silver nitrate was prepared by dissolving silver nitrate in deionized water. All chemicals were used without further purification.
Fig. 2. Schematic diagram of membrane module.
3.2. Membrane module
Fig. 3. Schematic diagram of the supported liquid membrane system.
514
M. Takht Ravanchi et al. / Journal of the Taiwan Institute of Chemical Engineers 40 (2009) 511–517
collected from the lower compartment and the secondary product, retentate, was collected from the upper compartment. A back pressure regulator (BPR, Tescom, Germany) was used on the retentate line to control the pressure of the system. During all experiments, sweep gas was at atmospheric pressure. The experiments were conducted at room temperature (298 5 K). All the experimental data were obtained after an initial permeation period of 4–6 h. Each experiment was repeated two times. 3.4. Analysis The gas composition was determined by a gas chromatograph (Agilent 6890N) equipped with a flame ionization detector (FID, Agilent Technologies Inc., column, HP Al/S, 0.53 mm in diameter, and 50 m in length).
4. Design of experiments
Fig. 4. Flowchart of the Taguchi method (Nikbakht et al., 2007).
- conducting the array experiment and - analyzing and verifying the results.
Design of experiments is an invaluable tool for identifying critical parameters, optimizing chemical processes and identifying operating regions for the process (Godbert, 2000). The Taguchi method is a powerful problem solving technique for improving process performance. It reduces scrap rates, rework costs and manufacturing costs due to excessive variability in processes (Antony and Antony, 2001). The techniques for laying out experiments when multiple factors are involved, has been known for a long time and is popularly known as the factorial design of experiments. This method helps researchers to determine the possible combinations of factors and to identify the best combination. However, in industrial settings, it is extremely costly to run a number of experiments to test all combinations. The Taguchi approach developed rules to carry out the experiments, which further simplified and standardized the design of experiments, along with minimizing the number of factor combinations that would be required to test the factor effects. So the Taguchi method has been chosen for this study. The Taguchi method was developed by Genichi Taguchi between 1950 and 1960. This method (Peace, 1993; Taguchi, 1990) is a systematic application of design and analysis of experiments for the purpose of designing and improving product quality. In recent years, the Taguchi method has become a powerful tool for improving productivity during research and development so that, high quality products can be produced quickly and at low cost. Optimization of process parameters is the key step in the Taguchi method in achieving a high quality without increasing the cost. This is because optimization of process parameters can improve performance characteristics and the optimal process parameters obtained from the Taguchi method are insensitive to the variation of environmental conditions and other noise factors. Basically, classical process parameter design is complex and not easy to use. Especially, a large number of experiments have to be carried out when the number of the process parameters increases. To solve this task, the Taguchi method uses a special design of orthogonal arrays to study the entire process parameter space with a small number of experiments only (Kargari et al., 2004b). In the Taguchi approach, orthogonal arrays and analysis of variance (ANOVA) are used as the tools of analysis. A brief overview of the process followed by the Taguchi’s approach to parameter design is provided in Fig. 4. These steps can be grouped as
In separation processes, separation percent is the most important criteria by which process performance can be evaluated. Thus, in the present study it is chosen as quality characteristic. Noise factors are those parameters which are either uncontrollable or are too expensive to control such as variation of environmental operating conditions. Control parameters are those design factors that can be set and maintained. The levels for each control parameter must be chosen at this point. The number of levels for each control parameter defines the experimental region. The array experiment is designed by selecting the appropriate orthogonal array for the control parameters. Orthogonal array for a particular project depends on the number of factors and their levels. In order to analyze the results, the Taguchi method uses a statistical measure of performance called signal-to-noise (S/N) ratio. The S/N ratio can be used to determine the product quality and to compare the product performance. It can be combined with the orthogonal array for the design of experiments to improve product and process. In its simplest form, the S/N ratio is the ratio of the mean (signal) to the standard deviation (noise). The Taguchi method uses the S/N ratio to measure the separation percent deviation from the desired value. There are three standard S/N ratios, i.e., bigger-is-better, smaller-is-better and nominal-is-best quality characteristics. As in the separation of propylene–propane mixture propylene is the desirable product, and S/N ratio was calculated on the basis of propylene concentration in the product, the more the S/N ratio the better the separation performance. For the larger the better responses, as in this study, the following relation is used to calculate S/N ratio: ! n 1X 1 S=N ¼ 10 log (2) n i¼1 y2i
- planning an array experiment to determine the effects of the control factors,
In the Taguchi analysis for experimental design, important factors that have influence on process and their levels must be
Quality characteristic is the output or the response variable to be observed. The quality characteristic of this study is separation percent (SP) which is defined as below: SP ¼
C permeate C feed C feed
(1)
where n is the number of experiments and yi is the response at each experiment (Montgomery, 1991). 5. Results and discussion 5.1. Taguchi approach
M. Takht Ravanchi et al. / Journal of the Taiwan Institute of Chemical Engineers 40 (2009) 511–517 Table 3 Separation factors and their levels. Factor
Unit
Level 1
Level 2
Level 3
Trans-membrane pressure Carrier concentration
kPa wt.%
50 5
100 10
120 20
determined at first. Trans-membrane pressure and carrier concentration were two influential parameters that affect the separation performance of a facilitated transport membrane. Thus, they were chosen as separation factors. Three levels set for each of these factors, which are demonstrated in Table 3. According to the Taguchi parameter design methodology and with the aid of Qualitek-4 software, one experimental design should be selected for the controllable factors. In the present work, an L9 orthogonal array was used. The experimental layout for the separation process parameters using this orthogonal array is shown in Table 4. Each row of this table represents a run, which is a specific set of factor levels to be tested. Nine experiments with the arrangement of Table 4 were performed. Separation percent was calculated by Eq. (1) and S/N ratio was calculated by Eq. (2). The results are shown in Table 5. As it can be seen in this table, when trans-membrane pressure was kept constant and carrier concentration was increased, S/N ratio was increased (experiments 1, 2, 3 or 4, 5, 6 or 7, 8, 9). The same trend can be observed when carrier concentration was kept constant and trans-membrane pressure was increased (experiments 1, 4, 7 or 2, 5, 8 or 3, 6, 9). Basically, the larger the response, the better the performance characteristic. As can be seen in Table 5, increasing the trans-membrane pressure and carrier concentration increases S/N ratio and separation percent. The larger the S/N ratio, the larger the contribution of one separation factor at that level for the separation. Moreover, entry No. 9 tops all the other eight entries with regard to separation percent and S/N ratio. Thus entry 9 levels were selected as optimum conditions. In other words, by selection of this entry, the best response with minimum required conditions (cost) will be attainable. The Taguchi analysis of separation data are presented in Table 6 and Figs. 5 and 6. As it can be seen, with an increase in factor levels the Taguchi responses were increased as well. The highest response was obtained when factors are at their highest levels. The Taguchi analysis predicts that optimum conditions at which highest product concentration obtained is when transmembrane pressure and carrier concentration are in their highest values (i.e. 120 kPa and 20 wt.% respectively). According to this analysis, at optimum condition propylene concentration in the product will be 99.801 vol.%. 5.2. Experimental results According to the Taguchi’s experimental layout (Table 4), some experiments were conducted the results of which are demon-
1 2 3 4 5 6 7 8 9
Table 5 The separation percent and the signal to noise ratio based upon experimental results. Experiment number
Separation percent
S/N ratio (db)
1 2 3 4 5 6 7 8 9
95.60 96.72 98.94 96.96 98.22 99.20 97.20 98.48 99.26
39.81 39.86 39.95 39.87 39.92 39.96 39.88 39.93 39.97
strated in Fig. 7. As can be seen in this figure, when transmembrane pressure and carrier concentration are in their highest levels, the highest concentration of propylene can be obtained in the permeate product, i.e. the highest separation percent was obtained. Permeability and selectivity are two important parameters by which the separation performance of a membrane can be evaluated. In Table 7 these values were reported. Facilitated transport is a combination of two processes: absorption (on the feed side) and stripping (on the permeate side). Increasing the pressure is in favor of absorption and decreasing the pressure is in favor of stripping. Thus, increasing the feed pressure increases the absorbed propylene on the feed side. Due to the pressure difference between feed side and permeate side, the complexed propylene is decomplexed on the permeate side. Therefore, the more the trans-membrane pressure, the more the driving force for separation and the more the propylene concentration in the product. This concept is in agreement with what is observed in Fig. 7. In facilitated transport membranes, propylene permeation occurs via two mechanisms: Fickian diffusion and facilitation transport. In the presence of carrier in the membrane system, propylene was permeated via Fickian diffusion and facilitated transport. Based upon facilitated transport mechanism, when more carriers were available in the membrane, more propylene can be transported along the membrane thickness, propylene concentration in the product was increased and separation performance was improved. The same results were observed in Fig. 7, i.e. when a 20 wt% carrier solution was used in the membrane, highest propylene concentration was obtained. According to Fig. 7, when trans-membrane pressure and carrier concentration are at their highest levels (i.e. 120 kPa and 20 wt.%, respectively) highest propylene concentration in the product (99.63 vol.%) is obtained which results an error equal to 0.17% from the Taguchi analysis prediction. 5.3. Analysis of variance Taguchi-oriented practitioners often use the analysis of variance to verify the factors that influence the mean response.
Table 4 Experimental layout using L9 orthogonal array. Experiment number
515
Factors Trans-membrane pressure
Carrier concentration
50 50 50 100 100 100 120 120 120
5 10 20 5 10 20 5 10 20
Table 6 The responses for the Taguchi analysis of the separation data. Response
L1 L2 L3 L2 L1 L3 L1 L3 L2
Factors Trans-membrane pressure
Carrier concentration
98.543 99.063 99.156 0.52 0.613 0.093
98.293 98.903 99.566 0.61 1.273 0.663
516
M. Takht Ravanchi et al. / Journal of the Taiwan Institute of Chemical Engineers 40 (2009) 511–517 Table 7 Propylene permeability and propylene–propane selectivity. Trans-membrane pressure (kPa)
Propylene permeability (Barrer)
Carrier concentration: 5 wt.% 50 100 120
Fig. 5. Main effect of trans-membrane pressure on product concentration (Taguchi analysis).
Fig. 6. Main effect of carrier concentration on product concentration (Taguchi analysis).
Selectivity
27.26 63.41 68.20
44.45 64.79 70.43
Carrier concentration: 10 wt.% 50 100 120
89.13 139.56 149.56
59.97 111.36 130.58
Carrier concentration: 20 wt.% 50 100 120
260.53 425.05 488.06
187.68 249.00 269.27
Sum of squares (SS), mean square (variance), the ratio of factor variance on error variance (F) and percent of contribution of each factor on the response (P) are important parameters used in analysis of variance. The purpose of the ANOVA is to investigate which process parameters significantly affect the performance characteristic. This is accomplished by separating the total variability of each level, which is measured by the sum of the squared deviations from the total mean of the responses, into contribution by each process parameter and the error. The percentage contribution by each of the process parameters in the total sum of the squared deviations can be used to evaluate the importance of the process parameter change on the performance characteristic. In addition, the F-test named after Fisher (1925) can also be used to determine which parameters have a significant effect on the performance characteristic. Usually, the change of the process parameter has significant effect on the performance characteristic when the F-value is large. Results of ANOVA which are shown in Table 8 indicate that carrier concentration with the contribution of 71.686% on the final response is the most significant process parameter for affecting the performance characteristic. From Fisher tables (Fisher, 1925) with 90% confidence, F10,2,4 = 4.32. According to this value, the F-value for transmembrane pressure and carrier concentration are greater than the corresponding values of Fisher tables, so the tests are reliable with 90% confidence. The contribution of error is 11%, which is in reasonable range of errors. 6. Conclusion
Fig. 7. Effect of trans-membrane pressure on product concentration with carrier concentration as parameter (experimental results).
In the separation of propylene–propane mixture via facilitated transport mechanism, trans-membrane pressure and carrier concentration were two important parameters. The influence of these parameters on the separation performance of a membrane system was investigated experimentally and via the Taguchi approach. Both analyses confirmed that the more the trans-membrane pressure and the more the carrier concentration, the better the separation performance.
Table 8 Statistical results of ANOVA. Factor
DOF
Sum of squares
Variance
F-ratio
Trans-membrane pressure Carrier concentration Other/error
2 2 4
0.656 2.435 0.179
0.328 1.217 0.044
7.311 27.107 –
Total
8
3.271
–
–
Contribution percentage 17.330 71.686 10.984 100.000
M. Takht Ravanchi et al. / Journal of the Taiwan Institute of Chemical Engineers 40 (2009) 511–517
Acknowledgement The authors gratefully acknowledge the financial support received in the form of a research grant (research project No.: 83111149) from the National Petrochemical Company- Research and Technology Center (NPC-RT), Iran. References Agam, G., G. Dagan, J. Gilron, V. Krakov, and N. Tsesin, ‘‘Recovery of Olefins from Gasous Mixtures,’’ WO Patent, 0117664 (2001). Antony, J. and F. J. Antony, ‘‘Teaching the Taguchi Method to Industrial Engineers,’’ Work Study, 50 (4), 141 (2001). Bai, S., S. Sridhar, and A. A. Khan, ‘‘Metal-Ion Mediated Separation of Propylene from Propane Using PPO Membranes,’’ J. Membr. Sci., 147, 131 (1998). Baker, R. W., ‘‘Future Directions of Membrane Gas Separation Technology,’’ Ind. Eng. Chem. Res., 41, 1393 (2002). Baker, R. W., Membrane Technology and Applications, John Wiley, England (2004). Bryan, P. F., ‘‘Removal of Propylene from Fuel-Grade Propane,’’ Separat. Purif. Rev., 33, 157 (2004). Burns, R. L. and W. J. Koros, ‘‘Defining the Challenges for C3H6/C3H8 Separation Using Polymeric Membranes,’’ J. Membr. Sci., 211, 299 (2003). Cussler, E. L., Facilitated and Active Transport, Paul, D.R. & Y.P. Yampolskii (Eds.), CRC Press, Florida, USA (1994). Eldridge, R. B., ‘‘Olefin/Paraffin Separation Technology: A Review,’’ Ind. Eng. Chem. Res., 32, 2208 (1993). Fisher, R. A., Statistical Methods for Research Workers, Oliver and Boyd, London (1925). Godbert, S., ‘‘Experimental Design—Identifying Factors that Affect Responses,’’ In Design and Analysis in Chemical Research, Tranter, R.L. (Ed.), Sheffield Academic Press, England (2000). Kaghazchi, T., A. Kargari, R. Yegani, and A. Zare, ‘‘Emulsion Liquid Membrane Pertraction of L-lysine from Dilute Aqueous Solutions by D2EHPA Mobile Carrier,’’ Desalination, 190, 161 (2006). Kang, Y. S., H. Ch. Park, J. O. Won, S. U. Hong, and T. M. Kwon, ‘‘Facilitated Transport Separation Membranes Using Solid State Polymer Electrolytes’’, U.S. Patent, 2001/ 0015334 (2001). Kang, Y. S., J. H. Kim, J. Won, and H. S. Kim, Solid-State Facilitated Membranes for Separation of Olefins/Paraffins and Oxygen/Nitrogen, Yampolskii, Y., I. Pinnau & B. Freeman (Eds.), John Wiley, England (2006). Kargari, A., T. Kaghazchi, and M. Soleimani, ‘‘Extraction of Gold (III) Ions from Aqueous Solutions Using Polyamine Type Surfactant Liquid Membrane,’’ Can. J. Chem. Eng., 82, 1301 (2004a). Kargari, A., T. Kaghazchi, M. Sohrabi, and M. Soleimani, ‘‘Batch Extraction of Gold (III) Ions from Aqueous Solutions using Emulsion Liquid Membrane via Facilitated Carrier Transport,’’ J. Membr. Sci., 233, 1 (2004b). Kargari, A., T. Kaghazchi, M. Sohrabi, and M. Soleimani, ‘‘Application of Experimental Design to Emulsion Liquid Membrane Pertraction of Gold (III) Ions from Aqueous Solutions,’’ Iran. J. Chem. Eng., 3, 76 (2006a). Kargari, A., T. Kaghazchi, B. Mardangahi, and M. Soleimani, ‘‘Experimental and Modeling of Selective Separation of Gold (III) Ions from Aqueous Solutions by Emulsion Liquid Membrane System,’’ J. Memb. Sci., 279, 389 (2006b).
517
Kim, H. S., J. H. Ryu, H. Kim, B. S. Ahn, and Y. S. Kang, ‘‘Reversible Olefin Complexation by Silver Ions in Dry Poly(vinyl methyl ketone) Membrane and Its Application to Olefin/Paraffin Separations,’’ Chem. Commun., 14, 1261 (2000). Kim, J. H., B. R. Min, K. B. Lee, J. Won, and Y. S. Kang, ‘‘Coordination Structure of Various Ligands in Crosslinked PVA to Silver Ions for Facilitated Olefin Transport,’’ Chem. Commun., 22, 2732 (2002). Kim, J. H., B. R. Min, J. Won, S. H. Joo, H. S. Kim, and Y. S. Kang, ‘‘Role of Polymer Matrix in Polymer–Silver Complexes for Structure, Interaction, and Facilitated Olefin Transport,’’ Macromolecules, 36, 6183 (2003). Kim, J. H., S. M. Park, J. Won, and Y. S. Kang, ‘‘Dependence of Facilitated Olefin Transport on the Thickness of Silver Polymer Electrolyte Membranes,’’ J. Membr. Sci., 236, 209 (2004). Krol, J. J., M. Boerrigter, and G. H. Koops, ‘‘Polymeric Hollow Fiber Gas Separation Membranes: Preparation and the Suppression of Plasticization in Propane–Propylene Environments,’’ J. Membr. Sci., 184, 275 (2001). Meyers, R. A., Handbook of Petroleum Refining Processes, McGraw-Hill, USA (1986). Mohammadi, S., T. Kaghazchi, and A. Kargari, ‘‘A Model for Metal Ion Pertraction through Supported Liquid Membrane,’’ Desalination, 219, 324 (2008). Montgomery, D. C., Design and Analysis of Experiments, John Wiley, New York, USA (1991). Mulder, M., Basic Principles of Membrane Technology, 2nd Ed., Kluwer Academic Publishers, The Netherlands (1996). Nabieyan, B., T. Kaghazchi, A. Kargari, A. Mahmoudian, and M. Soleimani, ‘‘Bench-Scale Simultaneous Extraction and Stripping of Iodine Using Bulk Liquid Membrane System,’’ Desalination, 214, 167 (2007). Nikbakht, R., M. Sadrzadeh, and T. Mohammadi, ‘‘Effect of Operating Parameters on Concentration of Citric Acid Using Electrodialysis,’’ J. Food Eng., 83, 596 (2007). Park, Y. S., J. Won, and Y. S. Kang, ‘‘Facilitated Transport of Olefin through Solid PAAm and PAAm-Graft Composite Membranes with Silver Ions,’’ J. Membr. Sci., 183, 163 (2001). Peace, G. S., Taguchi Methods: A Hand-on Approach, Addison-Wesley, Reading, MA (1993). Pinnau, I., L. G. Toy, and C. S. Casillas, ‘‘Olefin Separation Membrane and Process,’’ U.S. Patent, 5, 670, 051 (1997). Semenova, S. I., ‘‘Polymer Membranes for Hydrocarbon Separation and Removal,’’ J. Membr. Sci., 231, 189 (2004). Sridhar, S. and A. A. Khan, ‘‘Simulation Studies for the Separation of Propylene and Propane by Ethylcellulose Membrane,’’ J. Membr. Sci., 159, 209 (1999). Staudt-Bickel, C. and W. J. Koros, ‘‘Olefin/Paraffin Gas Separations with 6FDA Based Polyimide Membranes,’’ J. Membr. Sci., 170, 205 (2000). Taguchi, G., Introduction to Quality Engineering, Asian Productivity Organization, Tokyo, Japan (1990). Takht Ravanchi, M., T. Kaghazchi, and A. Kargari, ‘‘Application of Membrane Separation Processes in Petrochemical Industry: A Review,’’ Desalination, 235, 199 (2009). Tanaka, K., A. Taguchi, J. Q. Hao, H. Kita, and K. Okamoto, ‘‘Permeation and Separation Properties of Polyimide Membranes to Olefin and Paraffins,’’ J. Membr. Sci., 121, 197 (1996). Teramoto, M., H. Matsuyama, T. Yamashiro, and Y. Katayama, ‘‘Separation of Ethylene from Ethane by Supported Liquid Membranes Containing Silver Nitrate as Carrier,’’ J. Chem. Eng. Jpn., 19, 419 (1986). Yoon, Y., J. Won, and Y. S. Kang, ‘‘Polymer Electrolyte Membranes Containing Silver Ion for Facilitated Olefin Transport,’’ Macromolecules, 33, 3185 (2000).