Removal of Oil from Water Surfaces by Sorption on Unstructured Fibers Robert F. Johnson’ and Tulsidas G. Manjrekar
Textile Research Center, Texas Tech University, Lubbock, Tex. 79409 James E. Halligan2
Department of Chemical Engineering, Texas Tech University, Lubbock, Tex. 79409
w The capacity of unstructured fibers to remove crude oil from seawater was related to the chemical composition and surface properties of the fibers as well as the concentration, specific gravity, and temperature of the crude oil. The sorption capabilities of cotton exceeded those observed for all of the other synthetic and natural fibers tested. The amount of oil sorbed increased as the denier of the fiber decreased for all of the materials tested. For the data available, the critical surface tension of the solid sorbent was also related to the amount of oil sorbed. Cyclic sorption/desorption studies indicated that a simple squeezing operation was sufficient to remove most of the oil sorbed on the fibers and that recycling was feasible. Pollution of offshore and shoreline waters is not an unexpected consequence of the recent growth in offshore exploration and production, and ship transport of oil. Runoff from oil fields and refinery areas and, in some cases, process effluent from petroleum and petrochemical plants constitute a pollution hazard for inland waters. Such pollution is of the most visible and least acceptable type. It can foul beaches of recreational value as well as pose a serious hazard to the indigenous wildlife population and water supplies for human consumption. The Torrey Canyon spill in 1967 has produced an extensive evaluation (Carthy and Arthur, 1968) of deleterious effects on the ecosystem. Table I summarizes the major methods used or advocated for recent oil spill cleanup, both on- and offshore. Between the Torrey Canyon spill and the Santa Barbara Channel incident in 1969, there was no significant improvement of offshore cleanup techniques. Several more recent spill cleanup systems studies supported by the petroleum industry through the American Petroleum Institute, and various governmental agencies, indicate that few improvements have been put into practice. In the Chevron Charlie spill off the Louisiana coast (1970) and the San Francisco Bay spill (1971), procedures involving hay/ straw or skimming in conjunction with containment by booms were still used. This corresponded generally to the recommendations of the Dillingham Report (Gilmore et al., 1970a), an important cleanup system study. The failure to use other methods listed in Table I has usually been based on efficiency considerations or harm to the -ecosystem. The inadequacies in deployment and retrieval associated with hay and straw in shoreline situations have been vividly documented (White, 1971). Offshore spreading and harvesting of these sorbents is also in need of rationalization (Swift and Walkup, 1970). Other floating sorbents such as fibers and foams that have higher sorption capacity and may be structured into continuous belts are, therefore, of great potential. Sorbent deployment and retrieval ~
~
~
Present address, University of Minnesota, S t . Paul, Minn. 55101. 2 To whom correspondence should be addressed. 1
would be simplified and the need for disposal of oil-laden sorbent greatly reduced. Although several new commercial developments recognize this, few data are available on oil sorption capacities of fibers for which the conditions of experiment are described (Johnson et al., 1971; Sanuke et al., 1970; Sanuke et al., 1971; Schatzberg and Nagy, 1971; Zahid et al., 1972). Such data are indispensable for further design improvements or modifications for new applications. The purpose of this paper is to relate oil sorption to fiber chemical constitution and fiber surface properties, as well as oil concentration, specific gravity, and temperature. Some results on cyclic sorption/desorption are also given. These data on fibers in unstructured form have been used to design belts from structured fibers that have already been subjected to extensive dynamic sorption investigation a t the miniplant stage. The latter work will be valuable in assessing the validity of the rapid test procedure for oil sorption on fibrous materials described below.
Materials and M e t h o d s Fibers. The fibers were all of commercial origin, i.e., they are common items of commerce and not special experimental types. They do not incorporate any special characteristics such as nonconventional fiber cross-sectional profiles. Fibers from synthetic polymers or modified natural polymers are extruded as continuous, uncrimped filaments. For some applications, these filaments are cut into short lengths, generally 1.5-8 in., and crimped into a sawtooth configuration. All synthetic fibers and modified cellulosic fibers had the normal crimps per inch, except the 3.3 and 15.0 denier polyhexamethylene adipamide (PA) fibers which were uncrimped. The latter were cut into staple from uncrimped filament. All fibers were in the staple form, i.e., of small length to diameter ratio. Cotton had the shortest staple length, about 1 in., and wool the longest, about 5 in. The synthetic fibers had staple lengths of 3 f 1 in. Variation in staple length should not cause variation in oil sorption in the case of the staple wads used here.
Oil Spill Cleanup Methods Rapid Removal of Oil from Ecosystem Physical removal Floating sorbent mechanical harvesting Sorbent attached to revolving drums or belts Gelling agent -I-mechanical harvesting Mechanical skimming Suction pumping + separation Revolving metal drums Controlled Combustion Glass beads Particulate silicas Slow Removal of Oil from Ecosystem Dispersion + biodegradation Sinking sorbent + biodegradation Biodegradation promoters
Table I.
+
Volume 7, Number 5, May 1973
439
Table II.
Fiber Description Common name
Natural fibers cotton Wool
Modified cellulose fibers Viscose rayon Cellulose (2.5) acetate Cellulose triacetate Synthetic fibers Polypropylene Acrylic Nylon 6, 6 Polyester a
Chemical name
Symbolic term
Cellulose with 4-8% impurities, including 0.4-0.8% wax coating ...
co
2.3
wo
4.0
Cellulose Acetylated cellulosea Acetylated celluloseb
VIS CA CTA
3.0 3.0, 8.0, 17.0 2.5
Polypropylene Polyacrylonitrile Polyhexamethylene adipamide Polyethylene terephthalate
PP PAN PA PET
3.0, 18.0 3.0, 5.0, 15.0 2.3, 3.3, 15.0 2.3, 16.0
Deniers used
2.5acetyl groups per anhydroglucose unit. Three acetyl groups per anhydroglucose unit
Table I I I.
Oil Characterization
Oil source
Infrared absorbancy ratio classification Specific gravity, 60°F
APi gravity, 60’F
0.882 0.897 0.898 0.900 1.000
28.93 26.25 26.07 25.72 10.00
0.899 0.940 0.952 0.951
25.89 19.03 17.13 17.29 10.00
Viscosity, cp 6OoF
720 cm-’( 1375 cm-
810 cm-’/
1375 cm-’
810 cm-’,/ 720 cm-
1600cm-’,/ 1375 cm-
1600cm-’/ 720 cm-’
A s received state
Friendswood Field Crude Clearfork Field Crude No. 1 Clearfork Field Crude No. 2 Clearfork Field Crude No. 3 ENCO No. 6 Fuel Oil
3000
0.28 0.72 0.62 0.66 0.30
0.28 0.45 0.42 0.42 0.53
1.oo 0.63 0.68 0.63 1.79
0.27 0.41 0.40 0.38 0.41
0.98 0.58 0.65 0.57 1.38
55 203 310 300 3000
0.26 0.47 0.46 0.50 0.30
0.25 0.46 0.47 0.48 0.53
1.oo 0.97 1.01 0.97 1.79
0.24 0.36 0.39 0.43 0.41
0.92 0.77 0.83 0.86 1.38
... ... ... ...
( B u n k e r C)
As used state Friendswood Field Crude Clearfork Field Crude No. 1 Clearfork Field Crude No. 2 Clearfork Field Crude No. 3 ENCO No. 6 Fuel Oil (BunkerC)
1.000
Denier is a measure of linear density, being defined as the weight in grams of a 9000-meter length, and can also allow qualitative surface area comparisons between fibers as long as fiber density does not vary considerably. Synthetic and modified cellulosic fibers are generally delustered with titanium dioxide pigment. Three levels of delustering are normally available-bright, semidull, and dull. Bright refers to no delustrant, while semidull and dull correspond to increasing delustrant concentrations. Titanium dioxide when used as a delustrant causes asperities on the fiber surface leading to fiber roughness (Scardin0 et al., 1966). The staple fibers, which were not chemically treated or scoured before oil sorption, are summarized in Table 11. Oil Samples. Examination of crude oils in this work, almost to the exclusion of refined petroleum products, was influenced by the fact that in 35 recent major offshore spills involving 2,230,000 barrels, 80% was crude oil (Smith et al., 1970). Table I11 summarizes the oil characteristics. One lighter crude from the Friendswood Field near Houston was used, as well as three samples taken a t different times from the same Clearfork Field near Lubbock. The infrared absorbancy ratio classification was done by the method of Kawahara and Ballinger (1970). Crude oil contains low-boiling fractions which evaporate rapidly after a spill (Murphy, 1970) and often before significant cleanup operations can begin. To simulate this natural weathering and to reduce experimental variability (Sorption Procedure), the crude oils were held in thin layers in trays for 48 hr in an efficient hood a t room temperature. In the case of Clearfork Field Crude No. 1, this 48-hr treatment caused a weight loss of 16.5%. One-hundred-ninety hours would have been required to bring the 440
Environmental Science & Technology
weight loss to 22.570, a t which point further weight loss would have been negligible. Removal of volatiles in a rotary evaporator a t 50°C was about 30% (relative) less efficient than the tray method. I t was assumed that the 48-hr tray treatment was sufficient to overcome experimental variability; this was subsequently confirmed. In this “as used state,” the crudes were characterized (Table 111) as well as in the “as received state.” Sorption Procedure. Primary considerations in the design of the procedure were a simultaneous contacting of fiber with both oil and water, agitation of the heterogeneous system, avoiding devices which would give a capillary desorption effect when the fiber-oil mass was removed from the system, and a rapid, simple procedure. Several attempts led to the following procedure (Johnson et al., 1971) which was used here. One-liter beakers were inserted in jigs on the tray in a WCLID Thermostated Shaker Bath, Model 02156. No heat transfer liquid was used in the bath around the beakers, nor were the heater elements of the bath used. Most runs were made a t 25 f 1°C which was ambient temperature. Some runs were made a t 5 f 2°C by precooling the beaker contents with ice. Five-hundred milliliters of a 3-5% aqueous sodium chloride solution, to simulate seawater, were added before weighed amounts of oil were placed on the surface. One gram of fiber, in loose wad form, was introduced and the assembly shaken for 5 min a t 133 cycles/min and a mid-line displacement of 0.78 in. The fiber wad plus associated oil was removed by a wire hook inserted a t the top of the mass and allowed to drip for 1 min. The fiber plus oil was allowed to stand in an open petri dish in an efficient hood a t 25 f 1°C. Any water in the fiber-oil mass separated within 1 hr. The ap-
pearance and disappearance of water with time could be visually monitored. When all water had disappeared, generally after a maximum of 48 hr, the fiber-oil mass was weighed. The degree of shaking employed did not cause appreaciable phase mixing of oil and water but was sufficient for good contact of fiber with bath oil and water phases. The absence of a strong oil-water emulsion had the further advantage of simplifing the removal of water from the fiberoil mass after removal from the beaker. For a straightforward procedure, a simple distribution of oil between the fiber and water phases was necessary-i.e., no competitive sorption of oil by beaker wall or internal stirring devices. Glass beaker walls were found to wet less with oil than ones from Teflon (Du Pont) or various types of metal. Oil was negligibly sorbed on glass beaker walls until the fibers approached saturation. Sorption of oil on staple fiber wads under these conditions is rapid. Although a 5-min contact time was used, less would have given the same oil sorption. A 3-min contact time with cotton, for example, gave the same result within experimental precision. Cyclic Sorption/Desorption Procedure. Sorption was carried out as above, followed by passage through a pair of squeeze rolls, one stainless steel and the other hard rubber, a t maximum pressure. The fiber mass was then resubjected to the sorption procedure except that on occasion, an extraction of the fiber mass in 75 ml of trichlo-
F B COTTON IC01 W W L IWOI CELLULOSE TRIPCETATE CTAI CELLULOSE (251 bCETATE C A I VISCOSE RAYON IVISJ
DLMEE
l=25 L I'C
23 4 25 30 30
t: 5m,n
roethylene with nipping through the squeeze rolls, was inserted between nipping after sorption, and the subsequent sorption. Weighing the fiber mass a t appropriate points in the sorption/desorption sequence allowed calculation of oil on fiber. Repetition of this cycle was then carried out for the desired iteration.
Results and Discussion Figures 1 and 2 show the effect of fiber chemical constitution on oil sorbed a t increasing initial oil amounts on a constant simulated seawater surface. In Figure 1, the results for two natural fibers and three modified cellulose fibers as well as straw are given. Departures from one constant denier or oil specific gravity are consequences of materials sources but are not large enough to cause reversals of order (Figures 3 and 4 ) . One gram of cotton fiber (unscoured) is able to sorb oil almost quantitatively up to 40 grams of initial oil. The natural wax coating on cotton is important in this regard. Indeed, presoaking the cotton wad for 5 min in water before the sorption procedure, reduced oil sorption only 2-570 a t initial oil amounts below 60 grams. The highly water-absorbent forms of cotton normally encountered have the wax coating removed and have significantly reduced oil sorption after prior contact with liquid water. Above 40 grams of initial oil, sorption of oil on cotton becomes less favorable. The good performance of wool is attributed to its natural covering with wool grease (lanolin). DENlER COTTON IC01 C E L L U S E 1251ACETATE C A I POLYbCRYLONlTRlLE (PAN1
T . 25: I T I = 5 rnl"
23
30,8.0. 17.0 30,50.15.0
bC - . ,
/
3010952 /VPAN
y-)
3010940
301
201
o s -m21951
0-0-
0
%,
IO 20 35 40 50 60 INITIAL OIL ON WATER 10/500ml 3 5 % NoCl SOLUTION1
w COTTON IC01
COTTON IC01 POLYP PoLYPmpYLE~ POLYACRYLONITRILE POLYA-. IPANI POLYETHYLENE fTEREFWMbLbTE & € F W M b L b T E (PET) PCCYHEXAMETHYLEK AMPAMME (PA1
501
(rn
Ow,
0940
20 30 40 50 60 INITIAL OIL ON WATER lg/500ml 3 5 % NOCI SOLUTION1
F B POLYPROPYLENE (PPI POLYETHYLENE TEREPHTHALATE (PET1 POLYHEXAMETHYLEK IDIPAMIDE (PA)
30 .. 23 23
a
$,o
CO
0952
Figure 3. Dependence of oil sorption on fiber denier
T=25*12 t = 5 min
LE%! 91 23 SO
%LEi
CLEARFORK FIELD, LUBBOCK CLEARFORK FIELD, LUBBOCK
10
0
Figure 1. Oil sorption on natural fibers, modified Cellulose fibers, and straw
OL
T = 25tI.C t = 5 min
3 0 , 180 2 3 . 160 2 3 , 3 3 , 150
S
CLEARFORK FIELD, LUBBOCK CLEARFORK FIELD, LU0BOCK
23/0951
MNlER D
0940
0951
/ /@
PET 23m940 0
0 PA
23/0940
SUB CLEARFORK FIELD, LUBBOCK CLEARFORK FELD, LUBBOCK
0940 0951
.-.-.-.-
-o
o v 0
Ow,
10
20 30 40 50 60 INITIAL OIL ON WATER lri1500mI 3 5 % NaCl SOLUTION]
Figure 2. Oil sorption on synthetic fibers with cotton baseline
0 Ow,
0
__ v-
.-.
v
-v-
Pb 33/0940 (UNCRIMPEDI
v PA
15OM940
10 20 30 40 50 60 INITIAL OIL ON WATER lp/5OJml 3 5 % NaCl SOLUTION)
Figure 4. Dependence of oil sorption on fiber denier Volume 7 , Number 5, May 1973
441
These fibers, as the others, once coated with oil, tenaciously retain a thin layer which renders them very hydrophobic. Viscose rayon reacts abnormally in view of its similarity to cotton, both being basically cellulose, and its serrated cross-sectional profile. Figures 3 and 4 show that a high fiber surface-to-volume ratio promotes sorption. It appears t h a t the higher surface-to-volume ratio of serrated profile fibers, contrasted to round profile fibers, does not contribute significantly to oil sorption, since the cellulose acetate and cellulose triacetate curves in Figure 1 are also for serrated profile fibers. The latter two fibers, however, are much more oleophilic lhan viscose rayon. Straw was found under the conditions of test to attain its maximum oil sorption of 5 grams/gram straw a t the relatively low initial oil amount of 10 grams. This agrees well with field reports (Walkup et al., 1970; Gilmore e t al., 197Ob) and another laboratory evaluation (Schatzberg and Nagy, 1971). Sorption curves for four synthetic fibers are shown in Figure 2 . There is a clear distinction in sorption behavior among these fibers and also with reference to cotton. The polypropylene, polyester, and nylon fibers have round cross-sectional profiles while the acrylic is round with a slight indentation on one side. The polyacrylonitrile curve shown in Figure 2 is characteristically different from all others in Figures 1 and 2 ; this may be associated with generally observed oleophobic properties of this type fiber. There is no general correlation between oil sorption and equilibrium water vapor sorption (moisture regain).
w
D
CELLULOSE (251 ACETATE (CAI. BRIGHT B DULL VISCOSE RAYON (VISI, BRIGHT a DULL
M
From Figures 1 and 2 , it is obvious that the fibers have 3.5 to 6 times the capacity of straw for oil a t the 30-gram initial oil reference line. This capacity advantage of fibers is almost doubled a t the 60-gram initial oil reference line. The efficiency of straw in oil spill cleanup is often associated with its sorption of “five times its weight in oil” (Chem. Eng. News, 1970). Thus, it appears necessary to consider fibers as important candidates for oil spill cleanup operations. Sorption of model oil compounds, e.g., n-heptane, or loose fiber assemblies has been investigated (Sanuke et al., 1970). No difference in sorption capacity of different fiber types was observed. The results of our work indicate that the choice of model compounds by Sanuke was not representative. The cotton reference curve in Figure 2 exhibits data for sorption replication. The maximum deviation from the mean of replication sets run at initial oil amounts in the 40-60 range was f0.75. Experimental precision a t initial oil amounts below 40 was even better. Figures 3 and 4 give data on the relation of oil sorption to fiber surface area. In all cases, increasing denier-i.e., decreasing fiber surface area, for any one fiber type decreases oil sorbed. This indicates that surface sorption contributes to the overall sorption process. No evidence was found for diffusion of oil components into the fibers. The solubility parameters (a) for oil components and the fibers (Small, 1953) are sufficiently different that diffusion into the fibers in connection with a swelling mechanism might be expected only for polypropylene and that
T=2311’C ti5MIN
30 30
SPGR
-501
CLEARFORK Q! L
FIELD. LUBBOCK C L E A R F M FIELD. LUBBOCK
oil
S
FRIENDSWOOD FIELD, HOUSTON
0899
CLEARFORK FIELD. LUBBOCK CLEARFORK FIELD, LUBBOCK BUNKER C IENCO NO 6 FUEL)
OS40
T = 25 L 1.C I i 5 m,n
B
0951 1300
n CO 2 3 M 9 5 1
09 4 0 0 952
CO 23/0940
CA 3 O/BRiGHT/O952 CA 3 0 / W L L / 0 9 5 2 VIS 3 0 / W L L / 0 9 4 0
VIS 3O/BRIGHT/O 940
COTTON ICO) POLYPROPYLENE [PPI STRAW (S)
-0 23 30
0 0
INITIAL OIL ON WATER,
Ow,
Figure 5.
Ow,
Effect of fiber delustering on oil uptake
/
IO 20 30 40 50 60 INITIAL OIL ON WATER (4/500ml 3 5 % NoCI SOCUTION1
Figure 7.
Effect of oil specific gravity on oil uptake
A CO
23/0951 T=5*2T
/
/
-
LL
5d 01,
/v
IO
/
p’; 0
201 25 20 30 40 50 60 NTTlAL OIL ON WATER (p/500ml 3 5 % NoCI SOLUTION1
IO
0,
Figure 6. 442
COTTON (COI 2 3 DENIER OIL. CLEARFORK FIELD, LUBBOCK SP GR 0951 1.5mn
Effect of temperature
Environmental Science & Technology
30
35
xc
40
45
(dyner/cm)
Figure 8. Relationship between oil uptake and critical surface tension of fibers
COTTON ICOI, 2 3 DENIER OIL. CLEARFORK FIELD, LUBBOCK SP GR 0951 T-25.IT
i -5rn,n
OIL ON FIBER AFTER FIBER CONTACTS OIL
,,SQUEEZING OIL ON FIBER AFTER ~
OIL ON FIBER AFTER EXTRACTION
Figure 9. Discontinuous sorption/desorption of oil on cotton
has not been observed (Zahid et al., 1972). Although the individual effects of crimping and denier change could not be isolated, the difference in oil sorption between the one crimped polyamide fiber and the two uncrimped polyamide fibers in Figure 4 is striking and it is suspected that crimping was the major effect. All fibers natural, modified cellulosic, or synthetic, examined except those two uncrimped polyamides, had either natural or induced crimp of one form or another. Crimp increases the porosity of the wad-i.e., the accessible fiber surface area, leading to greatly increased oil sorption. This led to the conclusion that capillarity was an important contributor to the overall sorption process. Capillary bridging has been identified (Zahid et al., 1972) as the major mechanism in sorption of oil on structured polypropylene filament assemblies. The term sorption, as used here, describes generally the process by which the sorbate (oil) is redistributed between its first phase of residence (on a simulated seawater surface) and a second, added phase (an unstructured fiber wad). Comparisons between the various fiber types in Figures 3 and 4 can be made if fiber density is considered. At constant denier and weight, decreasing density will increase surface area and increase oil sorbed. This can be seen on comparing either polyester (PET; p = 1.38) or nylon (PA; p = 1.12) to polypropylene (PP; p = 0.91), but not on comparing P E T to PA. There does not appear to be a simple relationship between fiber density and oil sorbed. Delustering with titanium dioxide is often practiced for modified cellulosic and synthetic textile fibers. The incorporation of this white pigment, yielding semidull or dull fibers, causes fiber surface roughness on a small scale. Scardino found surface asperities of about 0.5 and 0.25 micron in diameter and height, respectively, in Ti02 delustered polyester fibers. The surface roughness difference between bright and dull cellulose acetate or viscose rayon fibers was not sufficient to cause much change in oil sorbed (Figure 5 ) . If oil is sorbed on and between fiber surfaces and not diffusing into the fiber bulk, sorption should still depend on oil viscosity. Figure 5 shows that a t initial oil amounts higher than 20 grams, oil sorbed on polypropylene increases with increasing specific gravity which is directly related to viscosity (Table 111). On cotton, the results are similar except that oil sorbed goes through a maximum between specific gravities 0.951 and 1.000, the latter corresponding to Bunker C. Straw does not show this effect. If straw is considered to be a very high denier hollow fiber, a significant contribution to oil uptake might be ex-
pected due to “internal capillary storage”-i.e., within the individual straw “fibers.” This was observed in this work, but to a greater degree for the crude oils than Bunker C. For Bunker C, capillary bridging between straw “fibers” is more important than internal capillary storage. Temperature slightly affects oil sorbed but only above initial oil amounts of 40 grams (Figure 6). Decreasing temperature increases viscosity and would be expected to increase oil sorbed (Figure 7 ) . Schatzberg and Nagy (1971) suggest that oil sorption on sorbents may be related to the critical surface tension (yc) of the solid sorbent. Figure 8 plots yc against oil sorbed where the initial oil amount was 40 grams. Unfortunately, yc values for raw (unscoured) cotton, cellulose acetate, and cellulose triacetate fibers could not be found. Cyclic sorption/desorption is indispensable to a rational oil spill cleanup system based on fibrous materials in a continuous belt. Figure 9 presents data for a discontinuous sorption/squeezing sequence with 2.5 cycles on cotton. Data are also given for a discontinuous sorption/ squeezing/extraction sequence. The extraction step is not only inefficient in reducing oil on fiber beyond that removed by squeezing, it also decreases oil sorption in the subsequent sorption step. This could be due to a reduction of oil viscosity by the trichloroethylene extractant or to a removal of the natural wax coating on cotton by this solvent. Work in progress on continuous sorption/desorption using fibrous materials in continuous belts indicates greater efficiency than might be obvious from Figure 9.
Literature Cited Carthy, J . D., Arthur, D. R., Editors, “The Biological Effects of Oil Pollution on Littoral Communities,” p p vii + 193, Field Studies Council, London, 1968. Chem. Eng Nelcs, 48 (31), 34-37 (1970). Gilmore, G. A., Smith, D. D., Rice, A. H., Shenton, E. H., Moser, W. H . , “Systems Study of Oil Spill Cleanup Procedures: Analysis of Oil Spills and Control Materials,” Vol. 1, pp v + 89 TAP1 Publication No. 4024), American Petroleum Instiiute. S e w York, N.Y., 1970a. Gilmore. G. A . . Smith. D. D.. Rice, A . H.. Shenton. E . H.. Moser. W. II., “Systems Study of Oil Spill Cleanup Procedures: Indus: try Response Plan,” Vol. 2, pp iv + 110 (API Publication No. 40251, American Petroleum Institute, New York, N.Y., 197Ob. Johnson, R. F., Manjrekar, T. G., Halligan, J . E., “Textile Materials Solution to Water Pollution Problems,” 79th Annual Conference. American Societv for Engineering Education. Annauolis, Md.. June 1971. Kawahara, F. K., Ballineer, D. G., Ind. Eng. Chem., Prod. Res. Deuelop., 9,563-8 (1976). Murphy, T . A . , R o c . Industrj-Gouernment Seminar on Oil Spill Treating Agents. Washington. D. C.. 150-64 (1970). Sanuke, H., ‘Ohtsu, T., Tomatsu, C.; J . Jap. Res.Ass. for Textiie End Uses, 11, 368-72 (1970), ( J a p ) . .Sanuke. H.. Ohtsu. T.. Tomatsu. C.. ibid.. 12. 156-61 (1971). Scardino, F. L.. Rebenfeld, L., Lyons, W: J., “Effect’of Fiber Surface Roughness on Cohesion of Textile Assemblies,” Textile Engineering Conference, Amer. SOC.Mech. Eng., Charlotte, N.C., April 1966. Schatzberr. P.. Nagv. K . V.. Proc. Joint Coni. on Preoention and Control Oil Spi‘iis, Washington, D.C., 221-33 (1971). Small, P . A . , J . Appl Chem., 3, 71-80 (1953). Smith. D.D., Gilmore. G. A , . Rice. A . H . Shenton. E . H.. Moser. W. H.. Proc. Industry-Gouernrnent Seminar on Oil Spill Treating Agents, Washington, D.C., 9-56 (1970). Swift, W. H . , Walkup, P . C., ibid., 92-108 (1970). Walkup, P. C., Blacklaw: J. R., Henager, C . H., “Oil Spill Treating Agents-A Compendium,” Battelle Memorial Institute, Richland, Wash., 1970, p 146. White, P. T., National Geographic Mag., 139,866-81 (1971). Zahid, M. A , , Halligan, J . E . , Johnson, R. F., Ind. Eng. Chem., Proc. Des. Deuelop. 11,550-55 (1972).
-
~
of
Received for recielc September 6, 1972. Accepted December 29, 1972. Volume 7, Number 5,
May 1973
443
