Environ. Sci. Technol. 2003, 37, 1013-1015
Oil Removal from Water by Selective Sorption on Hydrophobic Cotton Fibers. 1. Study of Sorption Properties and Comparison with Other Cotton Fiber-Based Sorbents GERALD DESCHAMPS,‡ HERVE CARUEL,‡ M A R I E - E L I S A B E T H B O R R E D O N , * ,‡ CHRISTOPHE BONNIN,# AND CHRISTIAN VIGNOLES† Laboratoire de Chimie Agro-Industrielle, ENSIACET, 118 Route de Narbonne, 31077 Toulouse Cedex 4, France, Anjou Recherche, Chemin de la digue, 78603 Maisons Laffitte, France, and Ge´ne´rale des Eaux, 78 Chemin des Sept Deniers, 31202 Toulouse Cedex, France
Hydrophobic cotton fibers, obtained by acylation of cellulose with fatty acid using microwaves radiations, have a high selective affinity for vegetable or mineral oil, fuel, and petroleum, in aqueous medium. Their sorption capacity (SC) (weight of liquid picked up by a given weight of sorbent) is about 20 g/g, after draining. They are reusable after simple squeezing, and their SC reaches a constant value, ca. 12 g/g. Moreover, this product is stable in water, whereas raw cotton can develop molds, after oil sorption. Besides, it is also biodegradable.
Introduction Water preservation implies recovering oils and hydrocarbon oils from water, among other pollutants. Among the main existing techniques, the use of a sorbent seems to be interesting because its function is to induce separation of oil and water so that the oil can easily be recovered (1). In this aim, the sorbent should have a high oleophilic and hydrophobic property. The sorption capacity is measured by the liquid sorption ratio. It is increased if the sorbent has the capability of drawing the oil into the material matrix, which implies a porous structure. Besides, the faster the oil is trapped, the less likely it will disperse and get away, and the easier the recover operation will be. Its buoyancy and durability in aqueous media are high; it should not retain water and react like a hydrophilic product. Its retention capability should also be high, so that the sorbed oil should not drain too quickly. It should rather be reusable, nontoxic for the environment if not reused, and preferably biodegradable (1, 2). Sorbents can be classified in three classes: polymers, natural materials, and treated cellulosic materials. The most often used polymers are polypropylene, polyethylene, or polyurethane (3). These products are quite efficient, but one of their major drawbacks is their nonbiodegradability. Natural materials include the following: (i) mineral products, such * Corresponding author phone: +33 5 62 88 57 26; fax: + 33 5 62 88 57 30; e-mail:
[email protected]. ‡ Laboratoire de Chimie Agro-Industrielle, ENSIACET. # Anjou Recherche, Chemin de la digue. † Ge ´ ne´rale des Eaux. 10.1021/es020061s CCC: $25.00 Published on Web 01/21/2003
2003 American Chemical Society
as perlite and vermiculite (4) [These materials do not show sufficient buoyancy retention and their oil sorption capacity is generally low (6 g/g) (5).]; (ii) animal materials such as wool (6); and (iii) vegetable products, such as wood (7), cotton, kapok, kenaf, or milkweed (6). Wood and kenaf have a low sorption capacity (4-8 g/g) (8, 9), whereas cotton, kapok, and milkweed show a natural hydrophobic character and a high sorption capacity, about 30-40 g/g (6). Among vegetable products, cellulosic materials can be treated in order to acquire a hydrophobic character, either by coating with a resin (10) or by reaction of hydroxyl groups with fatty reagents. The latter is the type of treatment that we have chosen to obtain hydrophobic cotton fibers. The purpose of this work is to study the oil sorption, in aqueous medium by treated cotton and to compare its performance with those of hydrophilic or raw cotton. Its reusability was also assessed.
Experimental Section Types of Sorbents. The acylation of the cellulose is conducted with a fatty acid (octanoic acid), without any solvent and by using microwaves (11). The resulting cotton fibers have a degree of substitution (DS) and a degree of polymerization (DP) values of 0.15 and 655. Their length is between 12 and 35 mm. The raw cotton fibers length is between 12 and 95 mm. They are naturally hydrophobic thanks to their surface waxes. The hydrophilic cotton fibers were preliminarily bleached, and the wax was removed. Their length was between 12 and 60 mm. Types of Pollutants. Our purpose was to choose many types of pollutants so that we could study the efficiency of our sorbent for the most common types of water pollution with oils: in restaurants with vegetable oil after washing-up; in sewers with mineral oil or fuel (in the streets) after rain; in rivers with fuel after draining from a tanker; in the sea with petroleum after an oil slick; ... Thus we have chosen these four pollutants, i.e., vegetable and mineral oil, petroleum, and fuel. The vegetable oil is named ISIO 4 from LESIEUR. It is a mix of saturated (10%) and unsaturated (42% of monounsaturated and 48% of polyunsaturated) fatty acids. The mineral oil is used for gas car motors (reference 15W40). The petroleum, nonrefined, is given by the Compagnie Ge´ne´rale des Eaux. Its composition is unknown. The fuel is a red one from a gas station, used for heating. Their characteristics are shown in Table 1. Procedure. The experience principle is based on the literature (12). In a crystallizer containing 1 L of water at 20 °C, the pollutant (280 mL) was poured to obtain a layer of about 1 cm above water. The cotton (2 g), flat cake-shaped (52 mm diameter) obtained by squeezing, was put onto the pollutant. The use of a flat cake shape avoids major problems of sorbent density variations (but it continues varying slightly). After sorption, the material was left to drip for 20 min and weighed. In the case of oil the cake was dried at 105 °C to constant weight. In the case of fuel and petroleum, water
TABLE 1. Characteristics of Studied Pollutants at 20 °C
vegetable oil mineral oil fuel petroleum
density
dynamic viscosity (Pa.s)
pollutant concn (mL/L of water)
0.91 0.86 0.82 0.87
0.065 0.265 0.002 0.031
280 280 280 280
VOL. 37, NO. 5, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1013
FIGURE 1. Vegetable oil sorption vs time for different cotton based materials (pollutant concentration ) 280 mL/L of water, 20 °C). Key: 2 raw cotton; 9 treated cotton; [ hydrophilic cotton.
FIGURE 2. Fuel sorption vs time for different cotton based materials (pollutant concentration ) 280 mL/L of water, 20 °C). Key: 2 raw cotton; 9 treated cotton; [ hydrophilic cotton. was determined by Karl Fischer measurements. The sorbed pollutant and the cotton sorption capacity (SC, weight of pollutant picked by one gram of sorbent) were thus quantified. This SC was presented in grams of oil per gram of sorbent, not in volume of oil per volume of sorbent. Indeed, considering that the sorbent is made of elastic, natural, and nonregular fibers, the flat cake can rise slightly, and it is difficult to define a constant density. Besides, after a MEB analysis, we could see that treated fibers are flat, so that we could not define fiber diameter. These parameters should be taken into account to calculate the SC in volume of oil per volume of sorbent (13). Thus, we used a traditional way to present the SC. The oil sorption kinetics for treated, hydrophilic and raw cotton were determined by immersing several cakes, during various time lengths and by quantifying the sorbed oil as a function of contact time. Sorbents Reusability. After sorption, the saturated material was put into a Bu ¨ chner funnel and squeezed for 20 min under 240 g/cm2, with a 50 mmHg vacuum. Then it was weighed in order to quantify the remaining pollutant and reused for a further sorption. The recycling efficiency was determined by following the sorption capacity evolution after several sorption/squeezing cycles.
Results and Discussion Pollutants Sorption in an Aqueous Medium. The vegetable oil and fuel sorption kinetics for the various sorbents are
FIGURE 3. Comparative pollutant sorption in aqueous medium by treated and raw cotton (pollutant concentration ) 280 mL/L of water, 20 °C). Key: 9 vegetable oil; 0 mineral oil; horizontal lines, fuel; vertical lines, petroleum. shown in Figures 1 and 2. Hydrophilic cotton behavior was different from other cotton samples. Its affinity for the considered pollutants was low. It sank into the aqueous phase while releasing the previously sorbed pollutant. On the contrary, the treated and raw cottons sorbed pollutant and reached a plateau corresponding to their SC. This behavior was observed for petroleum and mineral oil as well. Figure 3 shows the SC-values of raw and acylated cottons for the four studied pollutants. They were very close, regardless of the pollutant for the acylated cotton, and about 19-20 g/g. Moreover, the sorption was selective, because only 0.1 g of water was retained by 1 g of cotton. The high oil affinity of the sorbent is mainly due to two mechanisms: first, adsorption: after acylation, a part of the treated cellulose is on the fiber surface. Thus, the oil is adsorbed by the fibers, thanks to intermolecular interactions between oil and fatty chains grafted on cotton. Then, capillary action through diffusion of oil through the successive fiber walls, to fiber lumen. This phenomenon is important thanks to the lipophilic character of the treated cotton matrix. Raw cotton showed sorption capacities higher than those of treated cotton, with values between 23 and 30 g/g, which are comparable to the values determined by Choi (6, 14). This high affinity can be explained by the presence of waxes on raw cotton fiber surface (around 0.4-0.8%). The pollutant sorption begins through interaction with waxes (6, 15). However, despite these results, the saturation speed was lower for raw cotton (Figures 1 and 2). Indeed, contrary to treated cotton, the capillary action is not as important for raw cotton (because of the hydrophilic character of its matrix), and it may play a key role on the sorbent saturation speed, which implies an advantage in using treated cotton. Besides, this raw cotton was characterized by a lack of stability in aqueous medium over time after oil sorption, so that this sorbent could only be used to treat pollution, not to prevent it. After 10 days, a part of its matrix tended to sink into water, phenomenon already observed by Browers for materials with similar characteristics to those of raw cotton (2). Besides, molds appeared on the surface and grew over time. This led to a fall of its SC-value, of about 17%. On the contrary, treated cotton was very stable over time. It did not become moldy. However, a part of its matrix tended to sink into water after 1 month. Some tests have shown that,
TABLE 2. Comparative Reusability and Sorption Capacity of Treated and Raw Cotton after Squeezing at 20 °C initial sorption capacity (g/g)
treated cotton raw cotton 1014
9
limit sorption capacity (g/g) (10 cycles)
vegetable oil
mineral oil
fuel
petroleum
vegetable oil
mineral oil
fuel
petroleum
20 30
20 24
19 23
20 26
12.5 21
11.5 17.5
12 15
12 16
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 5, 2003
Sorbents Reusability. Cotton can be reused through numerous cycles. However, SC-value decreases regularly and reaches a level around 12 g/g for acylated cotton and 21 g/g for raw cotton, with vegetable oil (Figure 4). For the other pollutants, raw cotton was more efficient than treated cotton after several cycles (Table 2). They conserved their selective affinity for pollutants (0.1 g of water sorbed per gram of cotton).
Acknowledgments The authors thank Generale des Eaux and Anjou Recherche for financial support. FIGURE 4. Comparative reusability and sorption capacity of treated and raw cotton after vegetable oil sorption (pollutant concentration ) 280 mL/L of water, 20 °C). Key: 9 treated cotton; 0 raw cotton. after a quick drying (2 h at 105 °C), cotton regained its hydrophobic character and its initial SC-value (20 g/g). The SC-value of acylated cotton was always higher than those of natural sorbents such as kenaf (around 5 g/g (6, 9)), or peat (8 g/g (16)), or polypropylene (around 10 g/g (14)). Moreover, contrary to polymers, treated cotton is respectful of the environment, the acylation only slowing down its biodegradability. This product remains biodegradable and will not lead to supplementary pollution, even when saturated, contrarily to polymers when buried or burned. Besides, this additional pollution has be to be included in the costs. Thus, this study shows the interest of this material for a use as a sorbent of various pollutants. The acylated cotton behavior during pollutant sorption was essentially influenced by the product viscosity. Indeed, the diffusion step, through capillary action is sensitive to viscosity variations. So, the more viscous the pollutant is, the more the sorption kinetic slows down. According to Schatzberg (5), the rate of pollutant penetration into capillary is inversely proportional to the pollutant viscosity. In our case, between 10 and 30 °C, temperature did not show a big influence upon treated cotton sorption power. The kinetics were comparable, and the SC were identical, ca. 20 g/g.
Literature Cited (1) Pete, J. Nonwovens Ind. 1992, 6, 32-35. (2) Browers, S. D. Plant Eng. 1982 March 18, 219-221. (3) Zahid, M. A.; Halligan, J. E.; Johnson, R. F. Ind. Eng. Chem. Prod. Res. Dev. 1972, 11(4), 550-555. (4) Lawrence, P. A. Great Britain Patent 2,151,912, 1985. (5) Schatzberg, P. U.S. Coast Guard Report No. 724110.1/2/1; U.S. Coast Guard Headquarters, Washington, DC, 1971. (6) Choi, H.-M. J. Environ. Sci. Health 1996, Part A, 31, 6, 14411457. (7) Mazet, M.; Couillault, P.; Castillo, J. M.; Mathies, G. Fr 2708588, 1995. (8) Thengs, N.; Olsen, J. D. U.S. Patent 5,181,802, 1993. (9) Varghese, B. K.; Cleveland, T. G. Sep. Sci. Technol. 1998, 33(14), 2197-2220. (10) Orth, G. O. French Patent 2,162,689, 1973. (11) Vaca-Garcia, C.; Girardeau, S.; Deschamps, G.; Nicolas, D.; Caruel, H.; Borredon, M. E.; Gaset, A. Patent PCT WO 00/50492, 2000. (12) Croquette, J.; Bocard, C. Oil Petrochem. Pollut. 1983, 1(4), 261267. (13) Drelich, J.; Hupka, J.; Gutkowski, B. Studies Environ. Sci. 1988, 34, 207-221. (14) Choi, H.-M.; Cloud, R. M. Environ. Sci. Technol. 1992, 26, 772776. (15) Johnson, R. F.; Manjrekar, T. G., Halligan, J. E. Environ. Sci. Technol. 1973, 7, 439-443. (16) D’Hennezel, F.; Coupal, B. Can. Min. Metall. Bull. 1972, 64, 99-104.
Received for review March 13, 2002. Revised manuscript received September 25, 2002. Accepted November 18, 2002. ES020061S
VOL. 37, NO. 5, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1015
