water azeotrope using compound starch-based adsorbents

Bioresource Technology 101 (2010) 6170–6176

Contents lists available at ScienceDirect

Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

Separation of ethanol/water azeotrope using compound starch-based adsorbents Yanhong Wang, Chunmei Gong, Jinsheng Sun *, Hong Gao, Shuai Zheng, Shimin Xu School of Chemical Engineering and Technology, Tianjin University, No. 92 Weijin Road, Tianjin 300072, PR China

a r t i c l e

i n f o

Article history: Received 9 October 2009 Received in revised form 23 February 2010 Accepted 24 February 2010 Available online 20 March 2010 Keywords: Ethanol dehydration Adsorption Recycled starch-based adsorbent The adsorption heat

a b s t r a c t Comparing breakthrough cures of five starch-based materials experimentally prepared for ethanol dehydration, a compound adsorptive agent ZSG-1 was formulated with high adsorption capacity, low energy and material cost. The selective water adsorption was conducted in a fixed-bed absorber packed with ZSG-1 to find the optimum conditions yielding 99.7 wt% anhydrous ethanol with high efficiency. The adsorption kinetics is well described by Bohart–Adams equation. The adsorption heat, DHabs, was calculated to be 3.16  104 J mol1 from retention data by inverse gas chromatography. Results suggested that water entrapment in ZSG-1 is a exothermic and physisorption process. Also, ZSG-1 is recyclable for on-site multiple-use and then adapt for upstream fermentation process after saturation, avoiding pollution through disposal. Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction Various techniques have been developed to break the azeotrope of ethanol and water for preparation anhydrous ethanol, such as azeotropic distillation, extractive distillation, reactive distillation and adsorptive distillation. Among the methods, the adsorptive option is particularly attractive because of its low energy consumption, which takes up 50–80% of the overall energy required by the fermentative plan (Banat et al., 2000; Hassaballah and Hills, 1990). The removal of water from ethanol using a fixed-bed adsorbent, mainly molecular sieves, is a well-known process (Sowerby and Crittenden, 1988). However molecular sieves are expensive and can only be discarded after saturated with water, which makes the process less economical. So an increasing interest has been focused on the cheaper and recyclable adsorbents. Ladisch and Dyck (1979) first found that the biomass adsorbents, composed of cellulose and starch, could be used to remove water from a wide range of organics. Since then, many approaches have been developed in ethanol dehydration using various biomass materials such as corn (Robertson et al., 1983; Crawshaw and Hills, 1990), corn meal (Hong et al., 1982; Ladisch et al., 1984; Hills and Pirzada, 1989; Hassaballah and Hills, 1990) corn grits (Bienkowski et al.,1983; Neuman et al., 1986; Crawshaw and Hills,1990), cellulosics (Hong et al., 1982; Walsh et al., 1983) and starch (Hong et al., 1982; Crawshaw and Hills, 1990). The advantages of these adsorbents include high efficiency, relatively less raw material cost and energy consumption. For instance, the energy of ethanol dehydration is 3669 kJ kg1 ethanol by calcium oxide, and 2873 by cellulose. * Corresponding author. Tel./fax: +86 22 27891755. E-mail address: [email protected] (J. Sun). 0960-8524/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2010.02.102

Whereas, the authors also found that the adsorption process, consuming only 32 kJ m3 ethanol by corn grits, was more energy-saving than the 88 kJ m3 ethanol by azeotropic distillation (Hassaballah and Hills, 1990; Lee et al., 1991). Consequently corn grits were widely used in industry to produce anhydrous ethanol of 99.8 wt% over 750 million gallons per year (Beery and Ladisch, 2001). Recently, compound starch-based adsorbents were found still more efficient in anhydrous ethanol preparation. As one of the progress, Ma (2006) developed a granular adsorbent, composed of dry pachyrhizus, dry cassava, dry potato, wheat, pachyrhizus seeding, xanthan gum and carboxymethyl amylum. The product ethanol concentration reached 99.7 wt% from initial concentration of 92.4 wt%. Starch, starch-based materials, cellulose and hemi-cellulose were reported to have an affinity for water (Ladisch and Dyck, 1979; Hong et al., 1982; Westgate and Ladisch, 1993). As such, the objective of this study is to develop a optimum compound adsorbent by evaluating adsorption capacity of corn, sticky rice, sweet potatoes, durra, and crystal sugar. A new compound cheaper compound ZSG-1 consisting of corn, sweet potatoes and foaming agent was obtained through screening in a fixed-bed absorber. The best values for the process parameters was sought to increase the capacity of ZSG-1, and kinetic and thermodynamic analyses were conducted to investigate the adsorption property of ZSG-1.

2. Methods 2.1. Raw materials All raw materials, corn, sticky rice, sweet potatoes, durra, and crystal sugar, were of food quality and purchased in the market.

6171

Y. Wang et al. / Bioresource Technology 101 (2010) 6170–6176

These materials were dried in a vacuum oven at 90 °C for 12 h, and stored in sealed package for future screening experiment.

2.2. Zsg-1 Corn flour (90–40 mesh) and the sweet potato powder (120–90 mesh) were selected as basic components for ZSG-1. To prepare the absorbent, 5 g sweet potato powder was dispersed in 120 ml boiling water under stirring for 1 min. 4 ml H2SO4 (0.1 mol l1) was dripped into 82 ml above solution in a beaker, and 75 g corn flour was added into the solution by stirring to form a paste. Then the paste was squeezed into £3  4 mm columns within 1.5 h. Finally, the columns was kept drying first at room temperature in the ventilated cabinet for 7 days, and then at 90 °C in a vacuum oven for 12 h. The finished adsorbent columns were measured through the dead weight pressure test (static pressure over 313 kPa) at room temperature. 10 cm high ZSG-1 was placed in a 100 ml measuring tube and with 5 kg stainless steel column (£2.5 cm) on it. Samples with height shrinkage of less than 2 mm (2%) after 24 h are considered to pass the test.

C2

9

S2

8

10

7

5 V2

6 C1 4

S1

V1 1

V3

Fig. 1. Diagram of G-S fixed-bed adsorption system: (1) thermostatic waterbath; V1, V2, V3 valves; (4) peristaltic pump; (5) ultrathermostat; (6) steel wire gauze; (7) quartz sand; (8) adsorbent; (9) total condenser; (10) sample bulb.

2.3. Measurement of dynamic saturated adsorbance To evaluate the sorption capacity for ZSG-1, the dynamic saturated adsorbance was measured. The experiment unit was a glass distillation column packed with steel wire gauze packing caging 1 g adsorbent in it. Vapor was supplied at the bottom from a electrically heated kettle preloaded with enough distilled water. The flow of the distillate at the top total condenser was kept at three drops per minute for 2 h to drive air out and saturate the sample adsorbent. After that, the saturated adsorbent under test was taken out and weighed within 30 s.

2.4. Influencing factors analysis The influencing factors were evaluated using a fixed-bed absorber, which was a glass pipe with an internal diameter of 25 mm and a height of 20 cm. As shown in Fig. 1, a water jacket was used to maintain the temperature. The water bath and the circulating pump were switched on 3 h before heating the absorber to the operating temperature. The water jacket temperature should be held above the dew point of the vapor product mixture to avoid condensation. As soon as vapor began to generate, open valve V1 and V3 to let ethanol–water mixture run into the flask from V1 to maintain the concentration in the flask with a peristalsis pump. The mixture was prepared approximately the same ethanol concentration as the evaporation vapor. To start the experiment, switch vapor into the condenser C1 via valve V3. Samples was taken from S1 once every 2 min, analyzed and weighed to check the concentration and the feed rate. After the ethanol concentration of vapor was stabilized, close V3 and open V2, then let vapor go into the bed and water was adsorbed selectively by the adsorbent ZSG-1. Product coming from the adsorbed was condensed by condenser C2 and collected in receiver 10. The purified product was taken from the sample point S2 once every minute at the first 15 min and once every 2 min in the following 2 h. The samples were weighed and analyzed with gas chromatography. A column (U3 mm, 200 cm long), packed with GDX-203 and 100–120 mesh adsorbent was used in a gas chromatography to analyze the composition. Analysis at 120 °C was monitored by a thermal conductivity detector, and it agreed within 0.2%. The output from the detector was monitored by a recorder.

2.5. Inverse gas chromatography experiment Inverse gas chromatography (IGC) has been widely used to investigate the interaction properties of volatile compounds (moving phase) with the stationary phase. If very small amounts of a solute are used, the solute undergoes partitioning during elution. Small amounts of solute and low flow rates of carrier gas allow the approximation of equilibrium conditions. The time that elapses from injection of the sample to the recording of the peak maximum is defined as the retention time. The net retention time is the difference between the retention time of a solute and that of an unretained indicator (Vareli et al., 2000). It is also possible to obtain thermodynamic parameters using IGC. The volume of carrier gas necessary to elute an adsorbed solute from the column under specific temperature and pressure conditions which is called the net retention volume can be calculated from the following equation (Vareli et al., 2000):

V N ¼ jF c ðtR  tA Þ

ð1Þ

where tR is the retention time of the solution, tA the retention time of the air and Fc the corrected flow rate under the conditions within the oven. In Eq. (1)

Fc ¼

T v P0  Pw  F Tf P0

ð2Þ

where F is the flow rate, Tv the temperature of the oven, Tf the temperature of the flow meter, P0 the outlet pressure of the column and Pw the vapor pressure of water inside the soap bubble flow meter. j is the compressibility factor. So

j¼

ðpt =p0 Þ2  1 ðpt =p0 Þ3  1

 3=2

ð3Þ

where Pt is inlet pressure of the column. The specific retention volume V 0g is given by the following equation:

V 0g ¼

V N 273:15  T ws

ð4Þ

where ws is the mass of stationary phase and T the temperature of the column.

6172

Y. Wang et al. / Bioresource Technology 101 (2010) 6170–6176

The heat (enthalpy) of adsorption, DHs, is given by: 0 DHs d ln V g ¼ R d1=T

ð5Þ

100 to 200 meshed ZSG-1, dried in a vacuum oven at 90 °C for 3 h and stored in desiccators, were mixed with glass beads and DMCS. The inert support was dried under the same conditions as the adsorbents and stored in desiccators. The retention time of water on the above materials was measured. Twenty parts of the adsorbent sample was mixed with 80 parts of inert support by weight to generate an analytical column and was conditioned for at least 3 h by passing carrier gas (H2) through the column in an oven at 140 °C temperature. The oven temperature fluctuation was controlled within 0.5 °C by circulating air. The thermal conductivity detector (TCD) and injection temperature were both set at 150 °C. 99.999% H2 was used as the carrier gas. Pressure was regulated by a two-stage regulator and set as 400 kPa. The pressure in the column ranged from 28 kPa to 31 kPa depending on the flow rate of the carrier gas and the oven temperature. The carrier gas velocity (flow rate) was measured by a soap bubble flow meter attached to the TCD outlet and adjusted to 50 ml min1. For every test, 2 ll of the sample was injected into the chromatograph.

Fig. 2. Breakthrough curves at variable natural starchy materials for ethanol/water.

slowest decline. Thus, sweet potatoes and corn were chosen as basic materials to make the adsorptive agent ZSG-1. 3.2. Development of high adsorption capacity adsorbents

3. Results and discussion 3.1. Selection of basic materials Starch is a mixture of amylose, a linear polymer of D-glucose units joined by R-1,4-bonds, and amylopectin, a polymer of linear, 1,4-D-glucose chains links at branches points by R-1,6-bonds (Lee et al., 1991; Ruthven, 1984). The mechanism of adsorption of water is understood to involve hydrogen bonding with the hydroxyl groups on the starch chains. Both types of starch chains, amylose and amylopectin, interact with water molecules in this manner. However, the amylopectin structures also physically trap water molecules in the matrix of chain branches. When the water molecules are trapped this way, some of the nearby hydroxyl groups become unavailable for hydrogen bonding (Lee et al., 1991; Kulik et al., 1994; Rebar et al., 1984). A number of starch-based adsorbents have previously been applied to the adsorption of water from ethanol (Bushuk and Winkler, 1957; Hong et al., 1982; Pirzada, 1984). Based on the previous research and the mechanism of adsorption of water, five typical raw materials were selected and evaluated for preparing ZSG-1. In the five representative materials, corn represents amylose, sticky rice is on behalf of amylopectin, sweet potatoes stand for the high proportion of protein starch, whereas durra is high in vegetable oil starch content and crystal sugar is a single carbohydrate adsorbent. All samples were tested in adsorption experiments under the same conditions to select the basic materials. The bottom temperature of the tower was 85 °C, fixed-bed temperature 89 °C, peristaltic pump speed 19.8 rpm, the gas velocity 0.59 g°min1 and feed concentration 92.5 wt%. The loading of corn, sticky rice and sweet potatoes was 50 g, whereas that of sugar and durra was 67.0 g to keep a same bed depth, due to different specific densities. The breakthrough curves of the five raw materials are shown in Fig. 2. All tested materials showed similar behavior in these figures. It may be observed that the main differences were the maximum product concentration and the time required for saturating the adsorbent due to differences in adsorption capacities. Compared to other adsorbents, maximum ethanol concentration was obtained by sweet potatoes, and the distillate curve of corn had the

In order to improve the adsorption capacity for industrialized application, the adsorbent must have a number of hydrophilic groups and a large network structure. Such a water-adsorbent agent accommodates hundreds or even thousands of times of water. Hassaballah and Hills (1990) reported that the adsorption ratio of ethanol and water was increased to 1:1 for the corn by the means of exploding at 22 °C. The main components of corn are protein and polysaccharide. Zhang and Liu (1998) suggested that the role of hydroxyl in the polysaccharide and amidocyanogen in the protein should not be ignored in the adsorption. Thus, the factors of explosion, size of corn, additive and foaming agent were evaluated through the breakthrough curves to find ways to improve the adsorption capacity. Shown in Fig. 3A, the exploded corn has high product concentration and long saturated time, better than the unexploded ones, because high intensity of steam explosion can make the structure of corn looser. Besides, different size holes will be generated in the adsorbent with the presence of the foaming agent. This, in turn, facilitates the diffusion of adsorbates, and increases the ability of selective adsorption. A comparison test was carried out to show the difference and validate the theory. As a result, ZSG-1 containing a foaming agent shows a high product concentration (Fig. 3B). Furthermore, Fig. 3C describes the influence of additive to the adsorption capacity. Sweet potatoes and glutinous rice flour were added into corn, namely ZSG-1 and JMT-1, respectively; while YMT-1 is the blank. According to the experiment result, the optimal additive was sweet potatoes due to higher starch and protein content. 3.3. Dynamic saturated adsorbance of ZSG-1 Table 1 gives the saturated adsorbance at different temperatures. The water adsorption capacity of ZSG-1 has been found to be similar to some molecular sieves, which is about 0.15 g/g (Wang and Yin, 2002). 3.4. Effect of some parameters on ZSG-1 absorbing water The vapor superficial velocity, the temperature in fixed bed, and the feed concentration were varied in the experiments to select

6173

Y. Wang et al. / Bioresource Technology 101 (2010) 6170–6176

Fig. 3. Development of high adsorption capacity adsorbents. (A) Breakthrough curves at indirectly heated or not. (B) Breakthrough curves at added foaming agent or not adsorbents for ethanol/water. (C) Breakthrough curves at variable additives.

Table 1 Saturated adsorption of ZSG-1 for water at variable temperature. t/°C

30

79

84

89

97

Saturated adsorbance (g/g) Adsorption capacity (kg m3)

0.1046 53.32

0.1241 63.22

0.1499 76.37

0.1511 76.95

0.1529 77.87

optimal operation conditions, which decided the length and height of the plateau of the breakthrough curves. 3.4.1. Vapor superficial velocity in fixed bed According to the Van Deemter plot of gas chromatography, the relation between height equivalent to theoretical stage (HETS) and gas superficial velocity has a valley (Cooley, 1996), which causes a suitable vapor superficial velocity necessary for an adsorption process. In the experiments on the effect of vapor superficial velocities, the flow rates were held at 0.59 and 1.31 g min1, corresponding to the superficial velocities of 0.14 m s1 and 0.30 m s1, respectively. It was conducted in a fixed ZSG-1 bed with a height of 20 cm, weight of 50.0 g and the bath temperature of 90 °C. Ethanol concentration of above 99.7 wt% was obtained at 0.14 m s1. However, lower-concentration ethanol was obtained at 0.30 m s1, as shown in Fig. 4A. These results are supported by Westgate’s

(1993) conclusion that the ethanol dehydration by adsorption was dominated by mass transfer rate of ethanol and water. Hence, superficial velocity was set as 0.14 m s1 in the subsequent adsorption operations. 3.4.2. Fixed-bed temperature Adsorptive capacity of adsorbent is considerably influenced by the fix-bed temperature. In order to find the effect of temperature on the adsorption, experiments have been carried out at 81 °C and 89 °C, respectively, keeping the superficial velocity of the overheated vapor at 0.14 m s1. As a result, 100 wt% of ethanol concentration was obtained at 81 °C, which is higher than the fix-bed temperature at 89 °C. The productivity of 99.7 wt% ethanol, calculated from the experimental result, was 0.23 g/g and 0.20 g/g for 81 °C and 89 °C, respectively. Furthermore, it was found that the plateau length of the breakthrough curve got shorter with increase in temperature, as shown in Fig. 4B. As discussed earlier, the

6174

Y. Wang et al. / Bioresource Technology 101 (2010) 6170–6176

Fig. 4. Effect of some parameters on ZSG-1 absorbing water: (A) Breakthrough curves at variable vapor superficial velocities. (B) Breakthrough curves at variable temperatures. (C) Breakthrough curves at variable feed concentrations.

sorbent exhibited higher adsorptive capacity at lower temperature, above the dew point, 78.2 °C at inlet concentration. Thus, the operation temperature was held at 81 °C in the subsequent experiments. The temperature difference of 2.8 °C kept water in the vapor mixture from condensation. 3.4.3. Effect of feed concentration Combined with the results of above experiments, the influence of feed concentration on the capacity of adsorption was also tested. Fig 4C shows the concentration profile during the adsorption process under the above discussed conditions at different feed concentrations. As the feed concentration increased from 82.1 wt% to 92.5 wt%, product ethanol concentration was enhanced significantly, with the curve plateau length getting longer. One possible reason was that less adsorbed water resulted in lower temperature rising, which improved the adsorption capacity of ZSG-1. 3.5. Enthalpy of adsorption ZSG-1 was diluted as the stationary phase because of long retention time. Table 2 shows the net retention time of water and ethanol on glass beads at the temperature ranging from 75 °C to 130 °C. The retention time for both water and ethanol were very short, between 0.0125 and 0.0234 min. Moreover, there

Table 2 The net retention time of water and ethanol on the glass beads. Column temperature/°C

tr,Eol/min g1

t r;H2 O /min g1

130 125 120 115 110 105 100 95 90 85 80 75

0.0125 0.0123 0.0140 0.0141 0.0143 0.0143 0.0152 0.0158 0.0169 0.0184 0.0203 0.0225

0.0153 0.0160 0.0157 0.0157 0.0170 0.0179 0.0174 0.0186 0.0182 0.0196 0.0205 0.0234

was little difference between them because of their different boiling points. In other words, the silanized glass beads were essentially ideal diluents in view of their poor absorptive abilities to water and ethanol. The enthalpy (heat) of adsorption (Hs) is related to the energy of interaction between the adsorbed molecules and sorption sites in the substrate, which caused the parameter yields on the exothermic or endothermic characteristics of the interaction. ln V 0g versus

6175

Y. Wang et al. / Bioresource Technology 101 (2010) 6170–6176

adsorbent performance (McKay and Bino, 1990), although it suffers from certain limitations. In order to analyze the dehydration process of ZSG-1, some basic assumptions were proposed as follows: (a) The bed temperature is constant and the heat of adsorption is ignored. (b) The adsorption of ethanol does not affect the adsorption of water, that is, ethanol and water have different adsorption sites, and the amount of ethanol is so small that it can be regarded as an inert gas. (c) The vapor superficial velocity keeps the fluctuations very low. Because the concentration changes little during the adsorption, the vapor superficial velocity is calculated at an average speed.

Fig. 6 shows the linearity between the tBc0 and ln

Fig. 5. The linear regression for ln V 0g  1=T.

1/T (Fig. 5) was plotted with the slope Hs/R, thus Hs of the ZSG-1 could be calculated. Enthalpy of adsorption for water was 3.16  104 J mol1, whereas the corresponding values for ethanol in cornmeal, wheat flour of 70% extraction rate, wheat flour of 85% extraction rate, whole meal wheat flour, soy flour and pine sawdust are significantly lower, ranging from 1.43  104 J mol1 to 2.02  104 J mol1 (Vareli et al., 1997). The negative values of change in enthalpy (DHabs) shows that the adsorption is exothermic in nature. Thus, lower temperature was favorable to adsorption. The enthalpy obtained matches physical adsorption values (1.3 to 3.8  104 J mol1). Hence, the process can be considered as physisorption. 3.6. Kinetics analysis of mass transfer Data collected during laboratory and pilot plant tests are used as the basis to design full-scale adsorption column by using a number of mathematical models. The widely used model proposed by Bohart and Adams (1920) is based on surface reaction rate theory and can be represented as in Eq. (6) (Ayoob et al., 2007). In this work, it was employed to determine the adsorption capacity N0 which is necessary for the process design.

tB ¼

  N0 Z 1 c0 ln 1  c0 u Kc0 cB



c0 cB

 1

with a correlation coefficient of 0.996 at a flow rate of 0.59 g/ min. This shows that kinetics of water adsorption by ZSG-1 is better described by BDST model. The adsorption capacity No and the adsorption rate constant K were calculated from the slope and intercept of BDST plot to be 51.8 kg m3 and 13.6 m3 kg1 h1, respectively. The adsorption process consists of three parts, namely, external diffusion, internal diffusion, surface sorption. The rate constant K as a measurement of transfer of solute from the fluid phase to the solid phase, is very small, indicating that the kinetics of the adsorption transfer process is diffusion-limited. The adsorption capacity No is an important parameter in evaluating and selecting adsorbent. No of ZSG-1 indicates it has a high adsorption capacity, implying a long duration or more limited requirements of the adsorbent once in field application. 3.7. Energy estimation of adsorption To obtain 99.7 wt% ethanol, the absorption process lasted 10 min and 10.6 g product was obtained. The results are shown in Fig. 4B. The energy consumption in the process consisted of three parts: (1) Electricity to boil the feed in the electro thermal mantel. (2) Electricity to heat the adsorbent. (3) The heat of adsorption and water bath cycle shaft power.

ð6Þ

tB = breakthrough time, h; u = vapor superficial velocity, m h1, 0.59 g min1; Z = height of the fixed bed, m; Z = 0.20 m; N0 = the adsorption capacity, kg m3 K = the adsorption rate constant, m3 kg1 h1; c0 = the feed concentration, kg m3, c0 = 0.107 kg m3 cB = penetrating concentration, kg m3, cB = 0.00959– 0.0665 kg m3. Eq. (6) can also be rearranged to yield an expression for the product of the breakthrough time (tB) times the feed concentration (c0) realizing that:

t B c0 ¼ 

  1 c0 N0 Z 1 þ ln K u cB

ð7Þ

BDST model has been regarded as the simplest approach in fixed bed analysis which enables the most rapid prediction of

Fig. 6. The linear regression for the adsorption rate and adsorption capacity of water via ZSG-1 at 89 °C.

6176

Y. Wang et al. / Bioresource Technology 101 (2010) 6170–6176

The three consumptions cannot be measured directly. Energy consumption 1, named Q1, equals to the enthalpy change between liquid (26 °C) and vapor (81 °C), which is 6153 J. Energy consumption 2, named Q2, equals to the enthalpy change of ZSG-1 from 26 °C to 81 °C, which is 7425 J; The heat of adsorption (Q3) and water bath cycle shaft power (Lee et al., 1991) are ignored, because they are less than 0.5% of (Q1 + Q2). As a result, the energy consumption Q = (Q1 + Q2)/m  1.3  106 J kg1 ethanol. 4. Conclusions A new compound adsorbent ZSG-1 for breaking the azeptrope of ethanol and water was studied using a vapor adsorption process in a fixed bed. The adsorption capacity of ZSG-1 to water is as high as that of molecular sieve. The cost of ZSG-1 was approximately two times that of the corn, and one fifth of molecular sieve. The calculated thermodynamic parameters indicated that the adsorption of water on ZSG-1 is an exothermic process. The energy cost of ZSG-1 is 1.3  106 J/kg ethanol, which is smaller than the literature value (Lee et al., 1991). Thus, ZSG-1 has the potential to be used as an effective adsorbent for ethanol dehydration in industrial applications. Furthermore, made from crop materials also for ethanol fermentation, ZSG-1 can be recycled to produce more ethanol and avoid pollution through disposal, when lost efficiency after repeated regeneration. Acknowledgements We herein express our thankfulness to the Analysis Center of Tianjin University for product purity testing. The authors are grateful to Dr. Shengqiang Wang and Dr. Yingjie Qin for their helpful comments and suggestions during preparation of this manuscript. And special appreciation to Dr. Xingang Li, Director of National Engineering Center for Distillation Technology, Tianjin University, for his help on lab condition. References Ayoob, S., Gupta, A.K., Bhakat, P.B., 2007. Analysis of breakthrough developments and modeling of fixed bed adsorption system for As(V) removal from water by modified calcined bauxite (MCB). Sep. Purif. Technol. 52, 430–438. Banat, F.A., Abu Al-Rub, F.A., Simandl, J., 2000. Analysis of vapor–liquid equilibrium of ethanol–water system via headspace gas chromatography: effect of molecular sieves. Sep. Purif. Technol. 18 (2), 111–118. Beery, K.E., Ladisch, M.R., 2001. Enzyme Microb. Technol. 28, 573–581. Bienkowski, P.R., Barthe, A., Voloch, M., Neuman, R.N., Ladisch, M.R., 1983. Breakthrough behavior of 17.5 mol% water in methanol, ethanol, isopropanol,

and t-butanol vapors passed over corn grits. Separation from water in a twostage adsorption process. Biotechnol. Bioeng. Symp. 13, 629. Bohart, G.S., Adams, E.Q., 1920. Some aspects of the behavior of charcoal with respect to chlorine. J. Am. Chem. Soc. 42, 523–544. Bushuk, W., Winkler, C.A., 1957. Sorption of organic vapors on wheat flour at 27 °C. Cereal Chem. 4, 87. Cooley, R.S., 1996. In: Schweitzer, P.A. (Ed.), Handbook of Separation Techniques for Chemical Engineers, third ed. McGraw-Hill, New York. Crawshaw, J.P., Hills, J.H., 1990. Sorption of ethanol and water by starchy materials. Ind. Eng. Chem. Res. 29, 307. Hassaballah, A.A., Hills, J.H., 1990. Drying of ethanol vapors by adsorption on corn meal. Biotechnol. Bioeng. 35, 598–608. Hills, J.H., Pirzada, I.M., 1989. Analysis and prediction of breakthrough curves for packed bed adsorption of water vapor on corn-meal. Drying of ethanol vapors by adsorption on corn meal. Chem. Eng. Res. Des. 67, 442. Hong, J., Voloch, M., Ladisch, M.R., Tsao, G.T., 1982. Adsorption of ethanol–water mixtures by biomass materials. Biotechnol. Bioeng. 24, 725. Kulik, A.S., Chris de Costa, J.R., Haverkamp, J., 1994. Water organization and molecular mobility in maize starch investigated by two-dimensional solid-state NMR. J. Agric. Food Chem. 42, 2803. Ladisch, M.R., Dyck, K.K., 1979. Dehydration of ethanol: new approach gives positive energy balance. Science 205, 898–900. Ladisch, M.R., Voloch, M., Hong, J., Bienkowski, P., Tsao, G.T., 1984. Cornmeal adsorber for dehydrating ethanol vapors. Ind. Eng. Chem. Process Des. Dev. 23, 437. Lee, J.Y., Westgate, P.J., Ladisch, M.R., 1991. Water and ethanol sorption phenomena on starch. AIChE J. 37, 1187–1195. Ma, Xiao Jian, 2006. A kind of adsorbent for ethanol dehydration, China, 200610017885.1, Biochemical Engineering Centre of Zhengzhou University. McKay, G., Bino, M.J., 1990. Fixed bed adsorption for the removal of pollutants from water. Water Environ. Pollut. 66, 33–53. Neuman, R., Voloch, M., Bienkowski, P., Ladisch, M.R., 1986. Water sorption properties of a polysaccharide adsorbent. Ind. Eng. Chem. Fundam. 25, 422. Pirzada, L.M., 1984. Ph.D. Thesis, England, University of Nottinggham. Rebar, V., Fischbach, E.R., Apostolopoulos, D., Kokini, J.L., 1984. Thermodynamics of water and ethanol adsorption on four starches as model biomass separation systems. Biotechnol. Bioeng. 26, 513. Robertson, G.H., Doyle, L.R., Pavalath, A.E., 1983. Intensive use of biomass feedstock in ethanol conversion; the alcohol–water, vapor-phase separation. Biotechnol. Bioeng. 25, 3133. Ruthven, D.M., 1984. Principles of Adsorption and Adsorption Processes. Wiley, New York. Sowerby, B., Crittenden, B.D., 1988. An experimental comparison of type A molecular sieves for drying the ethanol–water azeotrope. Gas Sep. Purif. 2, 77–83. Vareli, G.D., Demertzis, P.G., Akrida-Demertzi, K., 1997. Water and ethanol adsorption on starchy and cellulosic substrates as biomass separation systems. Z. Lebensm. Unters. Forsch. 205, 204–208. Vareli, G., Demertzis, P.G., Akrida-Demertzi, K., 2000. Effect of regeneration thermal treatment of cellulosic and starchy materials on their capacity to separate water and ethanol. J. Cereal Sci. 31, 147–154. Walsh, P.K., Liu, C.P., Findley, M.E., Liapis, A.I., Siehr, D.J., 1983. Ethanol separation from water in a two-stage adsorption process. Biotechnol. Bioeng. Symp. 13, 629. Wang, Xiudao, Yin, Zhuorong, 2002. Production anhydrous ethanol by molecular sieve. Liquor-Making Sci. Technol. 1, 24–26. Westgate, P.J., Ladisch, M.R., 1993. Sorption of organics and water on starch. Ind. Eng. Chem. Res. 32, 1676. Zhang, Jiukai, Liu, Dazhong, 1998. Progress of researches on preparation of absolute ethanol with corn starch as dehydrating agent. SDILI Journal 12 (3), 1–3.