Experimental investigation of various vegetable fibers as sorbent materials for oil spills

Marine Pollution Bulletin 50 (2005) 1340 1346 www.elsevier.com/locate/marpolbul Experimental investigation of various vegetable fibers as sorbent materials for oil spills T.R. Annunciado, T.H.D. Sydenstricker, S.C. Amico * Department of Mechanical Engineering, Federal University of Paraná (UFPR), P.O. Box 19.011, 81.531-990 Curitiba-PR, Brazil Abstract Oil spills are a global concern due to their environmental and economical impact. Various commercial systems have been developed to control these spills, including the use of fibers as sorbents. This research investigates the use of various vegetable fibers, namely mixed leaves residues, mixed sawdust, sisal (Agave sisalana), coir fiber (Cocos nucifera), sponge-gourd (Luffa cylindrica) and silk-floss as sorbent materials of crude oil. Sorption tests with crude oil were conducted in deionized and marine water media, with and without agitation. Water uptake by the fibers was investigated by tests in dry conditions and distillation of the impregnated sorbent. The silk-floss fiber showed a very high degree of hydrophobicity and oil sorption capacity of approximately 85 g oil/g sorbent (in 24 hours). Specific gravity measurements and buoyancy tests were also used to evaluate the suitability of these fibers for the intended application. Ó 2005 Elsevier Ltd. All rights reserved. Keywords: Oil spill; Sorbents; Vegetable fibers; Sorption experiments; Silk floss 1. Introduction * Corresponding author. Fax: +55 361 3131. E-mail address: amico@ufpr.br (S.C. Amico). Oil is one of the most important energy and raw material source for synthetic polymers and chemicals worldwide. Whenever oil is explored, transported and stored and its derivatives are used there is risk of spillage with the potential to cause significant environmental impact. Pollution by petroleum oils affects sea life, economy, tourism and leisure activities because of the coating properties of these materials. Oil spills harm the beauty of polluted sites, the strong odor can be felt miles away and the excessive growth of green algae alters sea color and the landscape. When oil is spilled into a marine environment, it is subject to several processes including spreading, drifting, evaporation, dissolution, photolysis, biodegradation and formation of water oil emulsions. Oil spreading is likely to occur, especially if the sea surface is still. In the peculiar environment of rivers, pollutants are driven along the stream. In open seas or in harbors, the consequences of pollutants are often severe because of the action of local or tidal currents (Bucas and Saliot, 2002). Viscous oils spread more slowly than less viscous ones and therefore, water temperature, along with wind speed and sea conditions have an intense effect on the extent of oil spreading. Spreading is important in determining the fate of spilled oil through evaporation, emulsification and natural dispersion. Loss of volatile fractions changes oil composition and alters its density, pour point and flash point. Emulsification and evaporation lead to a decrease in the oil water density difference, and an increase in the oil pour-point (Reed et al., 1999; Wei et al., 2003). All these processes influence the choice of oil-spill countermeasures. Nevertheless, it is essential to quickly collect the oil after a spillage and mechanical recovery 0025-326X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.marpolbul.2005.04.043

T.R. Annunciado et al. / Marine Pollution Bulletin 50 (2005) 1340 1346 1341 by sorbents is one of the most important countermeasures in marine oil-spill response (Wei et al., 2003). Two broad categories of sorption phenomena, adsorption and absorption, can be differentiated by the degree to which the sorbate molecule interacts with the sorbent phase and its freedom to migrate within the sorbent. In adsorption, solute accumulation is in general restrict to the surface or interface between the solution and adsorbent. In contrast, absorption is a process in which solute, transferred from one phase to the other, interpenetrates the sorbent phase by at least several nanometers. Sorption results from a variety of different types of attractive forces between solute, solvent and sorbent molecules. Chemical (covalent or hydrogen bonds), electrostatic (ion ion, ion dipole) and physical (Coulombic, Kiesom energy, Debye energy, London dispersion energy) forces act together, but usually one type prevails in a particular situation (Weber et al., 1991). Despite the fact that synthetic polymers as polypropylene are said to represent ideal materials for marine oil-spill recovery due to their low density, low water uptake and excellent physical and chemical resistance, these sorbents are not renewable and biodegradable. Vegetable fibers are environmentally friendly materials, with densities close to that of synthetic polymers or even lower, and may show high oil sorption capacity at a usually low cost (Wei et al., 2003). The aim of this work is therefore to investigate various vegetable fibers, namely mixed leaves residues, mixed sawdust, sisal, coir, sponge-gourd and silk-floss fibers as potential sorbent materials for the oil sector. 2. Materials and methods 2.1. Sorbents and oil Crude heavy-oil from offshore wells of the Campos/ Rio de Janeiro Brazil basin was supplied by Repar/ Petrobras and used in all sorption experiments. The oil density was 25.8 API (0.90 g/cm 3 ) and its viscosity was 34 cp at 20 C. Different vegetable fibrous materials were used as sorbents, namely, mixed leaves residues, mixed sawdust, sisal (Agave sisalana), coir (Cocos nucifera), spongegourd (Luffa cylindrica) and silk-floss (Chorisia speciosa) fibers. The leaves residues comprised a mix of leaves from different trees supplied by a local vegetable waste company and were investigated as a cheap and practical alternative when commercial sorbents are not available. Sawdust from various wood-trees was obtained from a local carpentry workshop. Sisal was supplied by Cisaf-Nutrinuts/RN, coir fiber by Embrapa/CE and sponge-gourd by Driana Cosmetic company. Silk-floss fibers were collected from local trees. The fibers were separately grounded for 30 min in a knife mill (Rome) and classified in a set of ASTM sieves (3.35 mm, 1.70 mm, 850 lm, 600 lm, 425 lm, 300 lm and 212 lm) with the aid of a mechanical sieve-shaking device (Viatest) before sampling. The silk-floss was not grounded due to its cotton-like nature, which impairs this milling procedure. Specific gravity of the fibers, except silk floss, was measured with the aid of a picnometer and hexane was chosen as the test fluid since it did not show apparent indication of reaction with the vegetable fibers used and also because it was lighter than all fibers being tested. Due to the large volume occupied by the silk-floss fibers, their specific gravity was estimated by buoyancy in pentane (0.62 g/cm 3 ). Buoyancy tests were carried out following the work of Ribeiro et al. (2000) in order to suggest fiber suitability as a sorbent for water spills. Degree of hydrophobicity of the fibers was estimated according to Ribeiro et al. (2003), who used a heterogeneous mixture of water and hexane, steering, drying and weighing, and correlated the ratio of material (i.e., fiber) transferred to the organic phase as an estimation of the degree of hydrophobicity (or oleophilicity) of the materials. 2.2. Sorption experiments For the sorption experiments, crude oil was poured into a 100-ml beaker containing 80 ml of deionized water (ph = 7.0). After that, 0.5 g of the fibrous material was gently and evenly placed onto the oil surface (Fig. 1a). For the silk-floss, only 0.1 g of fiber was used due to its bulky nature. The amount of oil in the beaker was chosen so that there was still plenty of oil remaining in the beaker after completion of the sorption test. Thus, the volume of oil used was 5 ml for all fibers except sisal and silk-floss, for which 10 and 20 ml of oil were used, respectively, due to their higher oil sorption. After a certain period of time, namely, 5, 20, 40, 60 and 1440 min (24 h), the material was removed with the aid of a nylon collector, which was then placed on the top of a filter paper, being allowed drainage under vacuum for 5 min before weighing. All tests were carried out at 20 ± 1 Cand all weighing used an analytical balance (0.001 g). The sorption was calculated as the ratio of sorbed material to dry sorbent mass (S 0 ), sorption = (S t S 0 )/ S 0, where S t is the total mass of sorbed samples. Thus, sorption is given in unities of g/g of dry sorbent, being the sorbent the various vegetable fibers used. For each type of fiber, between 3 and 5 independent sorption experiments were carried out. This technique, although of a simple nature, is expected to give reliable results, with low standard deviation on the measurements.

1342 T.R. Annunciado et al. / Marine Pollution Bulletin 50 (2005) 1340 1346 Fig. 1. Different sorption systems used in tests carried out at room temperature just after fiber placement on the oil surface: (a) static system, (b) dynamic system and (c) dry system. Other sorption conditions were experimentally simulated: (i) dynamic system (Fig. 1b): some tests were carried out under constant steering (approximately 500 rpm) in an isothermal magnetic stirrer (Sciencetool) and (ii) marine conditions: these tests followed the same methodology above mentioned except for the fact that a substitute ocean water, produced according to ASTM D1141-90, was used instead of the deionized water. Further tests were conducted to evaluate water uptake by the fibers. These were divided into: (i) dry system (Fig. 1c), following the same methodology above except for the fact that oil was poured into a beaker with no water in it and (ii) evaluation of water by distillation of the sorbent after sorption, according to ASTM D95-83. Fig. 2 shows a histogram with the results of the classification of the grounded fibers where it can be seen that different fibers showed variable length dispersion. The granulometry range 1.70 0.850 mm obtained the largest percentage of grounded fibers for leaves residues, sponge gourd and sawdust, and also obtained enough of the other two fibers (sisal and coir fiber) for the experiments. Therefore, this was the granulometry range chosen as the standard one, being used in the sorption tests. Nevertheless, preliminary sorption tests were carried out to investigate the effect of granulometry on sorption. Table 2 shows the findings of such tests conducted for 60 min, where one can notice that a reduction of fiber 3. Results and discussion The mean specific gravity of the fibers is shown in Table 1. It can be seen that sisal, the leaves residues, sawdust and coir fiber are heavier than water, whereas sponge gourd and silk-floss are lighter. Sisal and the leaves residues are particularly heavy and this may influence their performance on the sorption test, which depends on their buoyancy in water, especially for the static system. Table 1 Specific gravity of the different fibers Fiber Specific gravity (g/cm 3 ) Standard deviation Sisal a 1.26 Leaves residues 1.16 0.04 Sawdust 1.07 0.03 Coir fiber 1.01 0.02 Sponge gourd b 0.92 0.05 Silk-floss <0.62 a Extracted from Sydenstricker et al. (2003). b Extracted from Tanobe et al. (2005). Fig. 2. Histogram of the granulometry distribution of the different grounded fibers. Table 2 Oil sorption of the various fibers at different granulometry ranges Fiber >3.35 mm 0.85 1.70 mm Sisal 3.0 6.4 113 Leaves residues 1.4 2.7 93 Sawdust 4.1 6.4 56 Coir fiber 1.8 5.4 200 Sponge gourd 1.9 4.6 142 Sorption increase

T.R. Annunciado et al. / Marine Pollution Bulletin 50 (2005) 1340 1346 1343 granulometry, from greater than 3.35 mm to 0.85 1.70 mm, is responsible for an increase in sorption that may reach 205%. Thus, these tests have confirmed the importance of controlling granulometry when comparing the results for different fibers and also that sorption is indeed very dependent on the availability of surface area per gram of sorbent, as reported by Shukla et al. (2002). Fig. 3 shows the results of the sorption tests for the static system and, as expected, sorption increases with the sorption period, from 5 to 1440 min. Sorption capacity followed the general trend: silk-floss > sisal > sawdust > coir fiber > sponge gourd > leaves residues. Fig. 4 shows the results of the sorption tests for the dynamic system, where the same features mentioned above are seen, namely, (a) a continuously higher sorption for all fibers as the sorption time increases, (b) the sorption capacity of the fibers followed the general trend: silk-floss > sisal > sawdust > coir fiber > sponge Fig. 3. Sorption of the different fibers at various sorption times static system. Fig. 4. Sorption of the different fibers at various sorption times dynamic system. gourd > leaves residues and (c) a much larger sorption for the silk-floss. The very low sorption capacity of the leaves residues suggests that they may only be indicated as a cheap low-performance alternative, in case a more efficient sorbent is not available. Not all weight gain shown in Figs. 3 and 4 refers to oil sorption, since water is also incorporated into the fibers to an extent dependent on the particular fiber. Table 3 shows the evaluation of water uptake for the silk-floss fiber. Comparison of sorption results for the static (water + oil) system (column B) and the dry (oil only) system (column A) suggests that water uptake varied between 2.5% and 6.3% of the total sorbed mass (column C). However, this method will give accurate results only if the kinetics of oil sorption of these two systems is comparable and any other influencing factor is constant and due to these drawbacks, a distillation technique was also used to check these findings. In fact, a narrower range of results was obtained with distillation (column D), with water uptake in the range of 3.1 4.1%. Besides, due to the small differences found in column D of Table 3, within the experimental error, no particular trend regarding the variation of water uptake with time was identified. Table 3 also suggests that direct comparison of experimental results obtained in different systems may incur in misleading findings. Column E (sorption in the dynamic system) has given lower values than column A (oil sorption in the dry system) for up to 60 min sorption time. This may have happened because the frequent water fiber contact, a consequence of agitation, decreases sorption rate in such a way that even the combined sorption of oil and water is not sufficient to equal the absorption of oil in the dry system, being this an indication of the hydrophobicity of the silk-floss. The dynamic system, therefore, has ratified the need to use the distillation technique, which has given water uptake values (column F) in the range of 2.7 3.5%, in the same range as that for the static system. Lee et al. (1999) reported a considerable reduction in oil sorption when water and severe agitation (with an orbital shaker) is present during sorption; a reduction from 30.62 to 8.07 g diesel oil/g sorbent (74% reduction) was obtained for natural cotton. In this work, the measured oil sorption reduction was less important, around 7%, possibly due to the significantly less severe agitation, higher hyrophobicity of the silk-floss, and also because in the work of Lee, the relationship between water and oil was 250/20 whereas in this work this relationship was 80/20. Another interesting finding of Figs. 3 and 4 regards the kinetics of sorption. Fibers such as sisal and silkfloss in both systems, static and dynamic, sorb more than 80% of their 24-h (1440 min) capacity in just 5 min. Besides, except for sisal that is of more difficult packing into the beaker, the dynamic system was in general responsible for retarding sorption.

1344 T.R. Annunciado et al. / Marine Pollution Bulletin 50 (2005) 1340 1346 Table 3 Sorption of the silk-floss fiber in the different systems Sorption time Dry system Static system Dynamic system A B C D E F Oil sorption (g oil/g sorbent) Sorption Water uptake by weight difference g (% of B) Water uptake by distillation g (% of B) Sorption 5 73.9 78.8 5.0 (6.3%) 2.7 (3.4%) 68.7 2.0 (2.9%) 20 75.2 79.2 3.9 (4.9%) 72.9 40 79.5 83.9 4.4 (5.2%) 74.5 60 81.0 85.2 4.3 (5.0%) 3.5 (4.1%) 77.6 2.7 (3.5%) 1440 84.9 87.1 2.2 (2.5%) 2.7 (3.1%) 85.8 2.3 (2.7%) Water uptake by distillation g (% of E) A straightforward, quantitative comparison between oil sorption in the static and dynamic systems, however, needs to consider a combination of factors, such as water fiber and oil fiber contact, buoyancy, hydrophobicity, accessibility of dry fibers to oil once the first layers of fibers become wet, kinetics of water and oil sorption, sorption capacity and time necessary to achieve equilibrium. The net result of these factors is dependent on a particular fiber and therefore of difficult general prediction. A more detailed analysis into these aspects will be given below for the silk-floss. Considering that the water uptake for the silk-floss is minimum and independent on sorption time and agitation, all systems should approach a similar oil sorption value upon equilibrium. Fig. 5 shows the sorption evolution with time for the three systems. The static system shows higher sorption than the dry system due to the combined oil and water uptake. If the mean water uptake of 3.5% (average of column D in Table 3) is subtracted from each point, a new curve is built (Static system minus water curve in Fig. 5) which is very similar to the dry system especially for longer sorption period. For the dynamic system a different scenario is seen. The dynamic system curve is below the dry system one for most of the time, meaning that the sorption kinetics is disturbed by the agitation, that favors water fiber contact and consequently delays oil sorption, and only on the long-run oil sorption is able to approach the expected Fig. 5. Sorption evolution for the silk floss for the dry, static and dynamic systems and water contribution. value. One could expect that since there is more water fiber contact, more water would be driven to the sorbent, but this is not verified, suggesting that the fiber has a hydrophobic nature. Nonetheless, once the water-share of the sorption (3.0% average of column F in Table 3) is subtracted from the total sorption, the data (dynamic system minus water curve in Fig. 5) approach that of the dry system in the long-term. In all, irrespective of the sorption conditions, the 24-h oil sorption of the silk-floss reached approximately 85 g oil/g sorbent. This sorption capacity is much higher than those reported in the literature for other vegetable fibers. Witka-Jezewska et al. (2003) showed sorption values from different authors and a maximum of 40 g oil/g sorbent for unscoured cotton. Saito et al. (2003) reported a maximum of 16.5 g oil/g sorbent for Sugi Bark. Lee et al. (1999) reported around 7 8 g diesel oil/g of ground, refined or extracted Kenaf core and bast, whereas natural cotton showed a much higher sorption, reaching 30 g diesel oil/g sorbent. Ribeiro et al. (2003) has found 11.6 g oil/g sorbent for salvinia sp. (mostly leaves) with a 237 cp oil (Marlin). This author has also reported 2.7 g oil/g peat sorb (<4 mm in size) for a Canadian commercial product. In order to ratify the reported values, two commercial vegetable sorbents of peat sorb were analyzed. The products were tested at 20 Cin the dry system for 60 min giving 4.8 ± 0.4 and 7.1 ± 0.3 g oil/g sorbent (as-received) and 6.8 ± 0.1 and 9.8 ± 0.4 g oil/g sorbent (for controlled granulometry range: 1.70 0.850 mm). Thus, these commercial products presented higher sorption capacity than the fibers used in this study (about 4.5 g oil/g sorbent for sisal and sawdust in the same conditions and less for the others), except the silk-floss, that reaches 8.5 times that of the best commercial product. The value obtained for silk-floss is even comparable to those of synthetic materials. Wei et al. (2003) reported a maximum of 14 g crude oil/g sorbent for different commercial nonwoven polypropylene sorbents. Witka-Jezewska et al. (2003) reported sorption values of 30 g oil/g viscose rayon and 40 g oil/g polypropylene, being the highest a polyurethane foam, that reached 100 g oil/g sorbent.

T.R. Annunciado et al. / Marine Pollution Bulletin 50 (2005) 1340 1346 1345 Table 4 Water uptake of the various fibers and their degree of hydrophobicity in different water conditions Fiber Water uptake Hydrophobicity Deionized water Sisal 27 31 0.0 0.0 Leaves residues 23 33 86.9 99.3 Sawdust 21 27 56.5 87.9 Coir fiber 42 45 38.6 77.7 Sponge gourd 50 51 0.0 0.0 Silk-floss 2.5 5.0 97.6 98.2 Marine water Table 5 Buoyancy of the various fibers under different conditions Fiber Static system Dynamic system Deionized Marine Deionized Marine Sisal 0.0 0.0 0.0 0.0 Leaves residues 65.5 75.3 85.6 95.0 Sawdust 12.4 16.0 17.7 22.9 Coir fiber 20.6 90.0 49.9 98.0 Sponge gourd 3.4 3.9 3.4 8.1 Silk-floss 100.0 100.0 100.0 100.0 Preliminary evaluation of water uptake as shown in column Cof Table 3 for silk-floss, has been also used for the other fibers and the results are shown in Table 4. It can be seen that water uptake vary in a wide range for the various fibers. Sponge gourd and coir fibers have shown the highest water uptake (42 51%), whereas sisal, leaves residues and sawdust have shown intermediate values (21 33%). These values are in the same range as those reported by Pasila (2004) for the separation of oil from a mixture of deionized water and lubricating or fuel oil, around 30 56% water content and 23 53% water content for filters composed of flax and hemp fibers, respectively. However, differently from the other fibers, silk-floss showed extremely low water uptake. This low water uptake and the lower oil sorption in the dynamic system may be partially explained by a high hydrophobicity. Table 4 shows the hydrophobicity of all fibers and in fact silk-floss showed values around 98%, along with the leaves residues. An opposite behavior was shown by sisal and sponge gourd with 0.0% hydrophobicity according to the methodology used, whereas sawdust and coir fibers showed intermediate values. Fig. 6 shows the results for the static and dynamic systems when artificial marine water was used. In general, the simulation of a marine environment did not compromise sorption capacity and since the values were within the estimated experimental error, no particular trend was identified. The silk-floss showed the best sorption capacity in all situations compared to the other fibers and also a similar sorption capacity in any simulated environment, suggesting that it may be used in any case of spill control with similar oil removal efficiency. The results of the buoyancy tests are shown in Table 5. Silk-floss, with 100% buoyancy for all simulated experimental conditions, showed the opposite behavior of sisal, with 0% buoyancy. The results of the dynamic system were higher than the respective ones for the static system and, as expected, all fibers showed higher buoyancy under marine condition in comparison to the deionized water. Coir fiber, with a specific gravity of 1.01 g/cm 3, largely benefits from the small increase in density of the marine water (1.024 g/cm 3 ) in comparison to the deionized one (0.998 g/cm 3 ), showing an increase from 20.6% to 90.0% and 49.9% to 98.0% for the static and dynamic systems, respectively. Table 5 clearly indicates the inability of sisal, sponge gourd and sawdust to be used in any water oil-spill conditions due to their low buoyancy, whereas the leaves residues and the coir fiber may be used in marine environments. 4. Conclusions Fig. 6. Sorption of the different fibers for a 60-min sorption period: static and dynamic system and deionized and salty water environment. The use of different vegetable fibers as sorbents of crude oil was investigated in various simulated conditions, deionized and marine water, with and without agitation. The sorption capacity of the fibers followed the general trend: silk-floss > sisal and sawdust > coir fiber > sponge gourd > leaves residues and the sorption capacity may be further increased by reducing granulometry. The silk-floss showed a rapid oil sorption and a very high sorption capacity of approximately 85 g oil/g sorbent (in 24 h), high degree of hydrophobicity and low water uptake. The sorption capacity was around

1346 T.R. Annunciado et al. / Marine Pollution Bulletin 50 (2005) 1340 1346 8.5 12 times that of two commercial products composed of peat sorb. Several of the low-cost vegetable fibers investigated may be used in dry environments, with variable sorption capacity. On the other hand, buoyancy tests indicated the inability of sisal, sponge gourd and sawdust to be used in any water oil-spill conditions, whereas the leaves residues and the coir fiber may be adequate for marine environments. Acknowledgement The authors would like to thank PRH-24/ANP/ MCT, Repar/Petrobras, Driana buchas, Embrapa and Cisaf-Nutrinuts for their support. References Bucas, G., Saliot, A., 2002. Sea transport of animal and vegetable oils and its environmental consequences. Marine Pollution Bulletin 44 (12), 1388 1396. Lee, B., Han, J.S., Rowell, R.M., 1999. Oil sorption by lignocellulosic fibers. In: Kenaf properties, processing and products. Mississippi State University, Ag. & Bio Engineering, pp. 421 433. Pasila, A., 2004. A biological oil adsorption filter. Marine Pollution Bulletin 49 (11 12), 1006 1012. Reed, M., Johansen, O., Brandvik, P.J., Daling, P., Lewis, A., Fiocco, R., Mackay, D., Prentki, R., 1999. Oil spill modeling towards the close of the 20th century: Overview of the state of the art. Spill Science & Technology Bulletin 5 (1), 3 16. Ribeiro, T.H., Rubio, J., Smith, R.W., 2003. A dried hydrophobic aquaphyte as an oil filter for oil/water emulsions. Spill Science & Technology Bulletin 8 (5 6), 483 489. Ribeiro, T.H., Rubio, J., Smith, R.W., Rubio, J., 2000. Sorption of oils by a nonliving biomass of a Salvinia sp. Environmental Science & Technology 34, 5201 5205. Saito, M., Ishii, N., Ogura, S., Maemura, S., Suzuki, H., 2003. Development and water tank tests of Sugi bark sorbent (SBS). Spill Science & Technology Bulletin 8 (5-6), 475 482. Shukla, A., Zhang, Y.H., Dubey, P., Margrave, J.L., Shukla, S.S., 2002. The role of sawdust in the removal of unwanted materials from water. Journal of Hazardous Materials 95 (1 2), 137 152. Sydenstricker, T.H.D., Mocnnacz, S., Amico, S.C., 2003. Pull-out and other evaluations in sisal-reinforced polyester biocomposites. Polymer Testing 22 (4), 375 380. Tanobe, V.O.A., Sydenstricker, T.H.D., Munaro, M., Amico, S.C., 2005. A comprehensive characterization of chemically treated Brazilian sponge-gourds. Polymer Testing 24 (4), 474 482. Weber, W.J., Mcginley, P.M., Katz, L.E., 1991. Sorption phenomena in subsurface systems concepts, models and effects on contaminant fate and transport. Water Research 25 (5), 499 528. Wei, Q.F., Mather, R.R., Fotheringham, A.F., Yang, R.D., 2003. Evaluation of nonwoven polypropylene oil sorbents in marine oilspill recovery. Marine Pollution Bulletin 46 (6), 780 783. Witka-Jezewska, E., Hupka, J., Pieniazek, P., 2003. Investigation of oleophilic nature of straw sorbent conditioned in water. Spill Science & Technology Bulletin 8 (5 6), 561 564.