Bioethanol production by a coupled fermentation/pervaporation process using silicalite membranes coated with silicone rubbers T. Ikegami a*#, D. Kitamoto b, H. Negishi c, T. Imura d, H. Yanagishita e Green Processes Group, Research Institute for Green Technology, National Institute of Advanced Industrial Science and Technology, Tsukuba, Ibaraki 305-8565, Japan a Tel. +81 (29) 861-2677; Fax +81 (29) 861-4660; email:email@example.com b Tel. +81 (29)861-4664; Fax +81 (29) 861-4660; email:firstname.lastname@example.org c Tel. +81 (29) 861-4666; Fax +81 (29) 861-4674; email:email@example.com d Tel. +81 (29) 861-4666; Fax +81 (29) 861-4674; email:firstname.lastname@example.org e Tel. +81 (29) 861-4659; Fax +81 (29) 861-4674; email: email@example.com Abstract In order to stably produce highly concentrated bioethanol, a coupled fermentation/pervaporation process was studied using ethanol-selective silicalite membranes coated with two types of silicone rubber,KE-45 and KE-108, as a hydrophobic material. Ethanol recovery was greatly improved by using a membrane coated with KE-45 silicone rubber. The recovered ethanol concentration in the permeate was 67% (w/w), and the amount of recovered ethanol from the broth was more than 10 times higher than that using a non-coated silicalite membrane. Succinic acid and glycerol created during fermentation, and glucose as the carbon source greatly interfered with the pervaporation performance of the coated membrane when used to separate an ethanol/water solution. 1. Introduction It is very important to efficiently produce bioethanol as an energy source by yeast fermentation employing renewable biomass resources, from viewpoints of not only getting ready for exhaustion of fossil energy resources in future, but also decreasing greenhouse gas emission, CO2, which causes global warming today. Bioethanol must be concentrated before use. Distillation process used generally for this purpose, however, is not cost effective. On the other hand, it is possible to concentrate bioethanol by a membrane separation technique, pervaporation which can accomplish selective removal of ethanol. By using a pervaporation separation process with ethanol- selective membranes, which are required to be hydrophobic, it is possible to concentrate low- concentration bioethanol from fermentation broths, which contain many compounds such as inorganic salts, by-products, and microorganisms, in a single step. It is reported that silicalite membrane, a polycrystalline zeolite with a negligible aluminum content, exhibits a very high pervapotarion performance in the separation of an ethanol/water mixture . By a coupled fermentation/pervaporation process employing a silicalite membrane, we have reported that an enriched ethanol solution, 85% (v/v) at maximum, could be obtained . However, the permeate flux and the ethanol concentration in the permeate decreased with time during fermentation. Considering that acetic acid is adsorbed onto silicalite membranes , it is assumed that the acidic by-products during ethanol fermentation change the properties of the silicalte membrane surface from their originally hydrophobic characteristics to hydrophilic characteristics, resulting in a decrease in ethanol permselectivity. In this study, using silicalite membranes coated with silicone rubber as a hydrophobic material to keep the membrane surface hydrophobic, the fermentation/pervaporation process was examined for concentration of bioethanol. ______ *Corresponding author # Presenting author 2. Methods and materials 2.1 Adsorption of succinic acid and glycerol onto silicalite powder One gram of silicalite powder was added to 10 ml of 0.3% (w/w) succinic acid solution or 0.8% (w/w) glycerol solution in test tubes. The samples were incubated at 30°C for 24 h while shaking at 50 rpm. After washing the silicalite powder with distilled water and drying it, the infrared spectra of the powders were measured at room temperature using a MAGNA- IR 550 spectrometer (Nicolet Japan, Japan). 2.2 Membrane preparation Silicalite membranes were hydrothermally synthesized on a stainless-steel support in the tetrapropylammonium bromide-Na2O-SiO2-H2O system, according to the previously reported method . This prepared membrane will hereafter be referred to as MC-N. 2.3 Membrane coating with silicone rubbers Two room-temperature vulcanizing-type silicone rubbers (KE45 and KE108), supplied by Shin-Etsu Kagaku Co. (Tokyo, Japan), were used to coat the silicalite membranes. KE45 which consists of one component, vulcanizes when it comes into contact with air. A hydrothermally prepared silicalite membrane was dipped into 5% (w/w) of the silicone rubber solution in hexane as a diluent for 10 s. This prepared membrane will hereafter be referred to as MC-1. The two component-type KE108 vulcanizes and develops crosslinks when it comes into contact with a catalyst. A hydrothermally prepared silicalite membrane was dipped into 3% (w/w) of this silicone rubber solution in hexane for 10 s. The prepared membrane will hereafter be referred to as MC-2. These coated membranes were dried at room temperature for 2 days. 2.4 Fermentation/pervaporation Ethanol solution Ethanol fermentation was performed with (Fermentation broth) a 20% (w/w) glucose solution (150 ml) and Vacuum gauge Vacuum line commercially available dry bakers’ yeast (1.5 g) at 30°C with homogenous stirring at 600 Enriched ethanol solution rpm, using a pervaporation apparatus shown in Fig. 1. The effective surface membrane area Liquid N2 was 19.6 cm2. The permeate was collected in Water bath a cold-trap utilizing liquid N2. The trap was periodically renewed. The permeation (total) Silicalite membrane Magnetic stirrer flux was calculated by dividing the permeate weight by the effective membrane area and the Fig. 1 Pervaporation apparatus integrated with ethanol-selective permeation time. silicalite membrane. 2.5 Analysis Glucose, ethanol, succinic acid, and glycerol concentrations in the medium were determined by high-performance liquid chromatography. Ethanol concentrations in permeate were quantitatively analyzed by gas chromatography . 3. Results and discussion Preliminary experiments showed that the main by-products were succinic acid at about 0.3% (w/w) and glycerol at about 0.8% (w/w) in the fermentation broths. As shown in Fig. 2, succinic acid was adsorbed by the powder, whereas glycerol was not. On the other hand, the adsorption of these compounds by commercially (A) silicalite available silicone rubber sheet is not observed. 1: before treatment These results indicate that non-coated silicalite 2: treated with 0.8% (w/w) glycerol membranes, which are originally hydrophobic, become hydrophilic due to adsorption of 3: treated with 0.3% (w/w) succinic acid succinic acid, and that the permeate ethanol concentration decreases when pevaporation is 4: difference between 1 and 3 attempted with non-coated membranes and that Absorbance it is useful to coat the silicalite membrane with silicone rubber to keep the membrane surface (B) silicone rubber hydrophobic. 1: before treatmen Fig. 3 shows the time courses of the 2: treated with 0.8% (w/w) glycerol recovered ethanol concentrations by the pervaporation using the three kinds of prepared 3: treated with 0.3% (w/w) succinic acid silicalite membranes. The maximal ethanol concentration recovered using the non-coated MC-N membrane was 40% (w/w) after 14 h of 3500 3000 2500 2000 1500 1000 Wavenumbers (cm-1) fermentation; thereafter the permeate ethanol Fig. 2 Diffuse reflectance infrared spectra of silicalite before concentration drastically decreased to 25% and after treatment with test solutions. (A) One gram of silicalite (w/w). 10 ml On the other hand, the silicalite powder andat 30°Cof the test solution were mixed in a test tube, and shaken for 24 h. (B) One gram of silicone rubber membranes coated with silicone rubber showed sheet and 10 ml of the test solution were mixed in a test tube, higher permeate ethanol concentrations than the and shaken at 30°C for 24 h. MC-N membrane. When using the MC-2 80 Ethanol concentration in permeate, (% w/w) membrane coated with the two-component KE- 70 MC-1 108 silicone rubber, the maximal ethanol concentration reached 59% (w/w). By applying 60 the MC-1 membrane coated with the other type 50 MC-2 of silicone rubber, KE-45, an enriched ethanol Ethanol recovery solution of up to 70% (w/w), could be obtained. 40 The total recovered amount (2.2 g ethanol) MC-N (not coated) 30 using the MC-1 membrane was 11 times higher than that (0.2 g) using the MC-N membrane. 20 0 10 20 30 40 50 The total recovered ethanol concentrations from Fermentation time, (h) the fermentation broths were 30% (w/w) with Fig. 3 Profile of batch ethanol fermentation/pervaporation membranes. the MC-N membrane, 55% (w/w) with the MC- with silicaliteof 20% (w/w) Fermentation wasand 1.5 gout with 150 ml glucose solution carried of 2 membrane, and 68% (w/w) with the MC-1 dry yeast at 30°C with stirring at 600 rpm. membrane. From the above-mentioned results, 0.14 the membrane MC-1 gave the best pervaporation 0.12 performance among the tested silicalite MC-1 membranes. Even in the case of using the Total flux, (kg/m2 h) 0.10 silicalite membranes coated with two kinds of 0.08 silicone rubber, the ethanol concentration in the permeate gradually decreased toward the end of 0.06 MC-2 the fermentations. Moreover, as shown in Fig. 4, 0.04 in all cases, the total fluxes drastically decreased with elapse of fermentation time. The 0.02 MC-N (not coated) fermentation broths contain glucose as a carbon 0 0 10 20 30 40 50 source, and by-products such as organic acids, Fermentation time, (h) besides yeast cells. In what follows, some factors Fig. 4 Changes in membrane flux through silicalite concerning the pervaporation performace was membranes with time during fermentations. investigated with the membrane MC-1 which gave the best pervaporation performance in Fig. 3. 0.18 80 Ethanol concentration in permeate, (% w/w) Fig. 5 shows the relationship between 0.16 , Water flux. succinic acid concentration in an ethanol/water 0.14 , Ethanol flux. 70 solution and pervaporation performance. It is Total flux, (kg/m2 h) 0.12 clear that not only the total flux but also the 0.10 permeate ethanol concentration were greatly 60 0.08 affected by the addition of succinic acid. The 0.06 permeate ethanol concentration was little 50 affected below 0.1% (w/w) succinic acid, while 0.04 with increasing succinic acid concentration the 0.02 total flux decreased to about 10% of that before 0 40 0 0.1 0.2 0.3 addition. The surface of silicalite membrane Succinic acid, (% w/w) which is formed from polycrystals, is rough and Fig. 5 Effect of addition of succinic acid to ethanol solution there are defects or pores between silicalite on pervaporation performance. Feed ethanol concentration: grains within the membrane . We have ca. 5% (w/w), temperature: 30°C. already reported that the rough surface of silicalite 0.18 80 Ethanol concentration in permeate, (% w/w) membrane was partially covered with the KE-45 0.16 , Water flux. , Ethanol flux. silicone rubber . Based on these facts, it is 0.14 70 assumed that the observed deterioration in 0.12 Total flux, (kg/m2 h) pervaporation performance with the MC-1 0.10 membrane resulted from the partial adsorption of 0.08 60 succinic acid onto non-coated silicalite crystals in 0.06 the membrane, causing obstruction of pathways 50 0.04 for the penetrants, water and ethanol, and 0.02 changing its surface property from hydrophobic to 0 40 hydrophilic. 0 0.2 0.4 0.8 Next, the effect of glycerol on the Glycerol concentration, (% w/w) pervaporation performance was investigated. The Fig. 6 Effect of addition of glycerol to ethanol solution on pervaporation Feed total flux decreased with increasing glycerol ca. 5% (w/w), performance. 30°C. ethanol concentration: temperature: concentration (Fig. 6). In this case, however, the extent of the decrease in total flux was clearly smaller than that caused by the addition of 0.3% (w/w) succinic acid. The degree of decrease in water flux was almost the same as that in the ethanol flux. Considering that silicalite powder and silicone rubber sheet do not adsorb glycerol as shown in Fig. 2, and that the driving force of membrane permeation is the vapor pressure gradient across the membrane , it is 0.18 80 Ethanol concentration in permeate, (% w/w) assumed that the pervaporation behavior shown 0.16 , Water flux. in Fig. 6 can be attributed to a decrease in the 0.14 , Ethanol flux. 70 vapor pressures of both ethanol and water by the Total flux, (kg/m2 h) 0.12 addition of glycerol to an ethanol/water mixture. 0.10 Also, by adding glucose to an 60 0.08 ethanol/water mixture, the significant decrease in the total flux was caused (Fig. 7). The extent 0.06 50 of the decrease in water flux was larger than that 0.04 of ethanol in this case. As a result, the permeate 0.02 ethanol concentration increased. In the batch 0 40 0 2 4 8 fermentation/pervaporation experiment carried Glucose concentration, (% w/w) out in this study, the glucose concentration in the broth decreased with elapse of fermentation Fig. 7 Effect of addition of glucose to ethanol solution on pervaporation performance. Feed ethanol concentra- time. Thus, it is clear that the deterioration in tion: ca. 5% (w/w), temperature: 30°C. the total flux shown in Fig. 4 was caused by the accumulation of glycerol in the broth, besides succinic acid, and that also glucose caused the decrease at the early stage of the fermentation 4. Conclusions In this study, highly concentrated ethanol solutions can be obtained by a fermentation/pervaporation process using silicalite membranes coated with silicone rubber acting as a hydrophobic material. Use of a silicalite membrane coated with KE-45 silicone rubber resulted in the amount of recovered ethanol from the fermentation broth being 11times higher than via the non-coated membrane. Succinic acid and glycerol as by-products had negative effects compared with the pervaporation performance shown in the ethanol/water binary systems. Succinic acid caused greater negative effects than glycerol, suggesting that the negative effect caused by succinic acid is attributable to the adsorption of succinic acid onto the membrane surface. As a method to cope with this demerit, coating with silicone rubber as presented in this study is easy and advantageous. In order to prevent the membrane surface from directly contacting acidic compounds, the surface of silicalite membrane should ideally be coated uniformly and thinly with a hydrophobic material such as silicone rubber. References  T. Sano, H. Yanagishita, Y. Kiyozumi, F. Mizukami, K. Haraya, Separation of ethanol/water mixture by silicalite membrane on pervaporation, J. Membr. Sci., 95 (1994) 221-228.  T. Ikegami, H. Yanagishita, D. Kitamoto, K. Haraya, T. Nakane, H. Matsuda, N. Koura, T, Sano, Production of highly concentrated ethanol in a coupled fermentation/pervaporation process using silicalite membranes, Biotechnol. Tech., 11 (1997) 921-924.  T. Sano, S. Ejiri, K. Yamada, Y. Kawakami, H. Yanagishita, Separation of acetic acid- water mixtures by pervaporation through silicalite membrane, J. Membr. Sci., 123 (1997) 225-233.  H. Matsuda, H. Yanagishita, H. Negishi, D. Kitamoto, T. Ikegami, K. Haraya, T. Nakane, Y. Idemoto, N. Koura, T. Sano, Improvement of ethanol selectivity of silicalite membrane in pervaporation by silicone rubber coating, J. Membr. Sci., 210 (2002) 433-437.  W. Gudernatsch, K. Kimmerle, N. Stroh, H. Chmiel, Recovery and concentration of high vapour pressure bioproducts by means of controlled membrane separation, J. Membr. Sci., 36 (1988) 331-342.