December 1947
INDUSTRIAL AND ENGINEERING CHEMISTRY
verted t o new type of operation; loiv temperature regeneration allows use of ordinary steel a n d keeps corrosion at a minimum; and odor nuisance during regeneration is eliminated, since disulfides are formed as a by-product. COYC LU SIOK s
1659
the rombination being knon-n as tannin solutizcr process. The authors also wish t o acknowledge the assistance of members of t h a t organization as regards information on the more recent commercial applications. LITERATURE CITED
T h e use of mercaptan oxidation catalysts, soluble in alkaline treating reagents, offers a n attractive, econoniical means of regenerating solutions used in the removal of mercaptans from gasoline. T h e ease of regeneration possible through the use of oxidation catalysts materially iniproves operations of nicrcaptan extraction systems. ACKNOWLEDGMENT
The authors wish t o express their appreciation t o J. B. Rather of Socony-Vacuum Oil Company, Inc., for permission t o publish this paper. Since t h e paper was originallv offered for publication, have bcerl n i t h t h e Shell ~ ~ ~corn: ~ pany for the joint licensing of Shell’s solutizer process incorporating the described tannin-catalyzed air regeneration technique, ~~
(1) Happel, Cauley. and Kelly, Proc. Am. Petroleum Inst., 23, 3, 67-77 (1912). (2) Kalichevskr-, “Chemical Refining of Petroleum,” p. 149. A.C.8. .\lonograph Seriea. To. 68, New York, Chemical Catalog Co., 1933. E s c . CHEM.,23,354-7 (1931). (3) Lachman, IXD. (4) Perkins and Everst in Allen’s “Commercial Organic Aiialy&,” Vol. 5 , p. 6 (1927). ( 3 Russell, Chem. R e t s . , 17, 1 5 5 4 6 (1933). (‘) Yabroff and Border, Rqfiner S a t i u a l Gasoline Mfr., 18, 171-0, PO3 (1939). RECEIVED J u n e 6, 1946. Piesented before t h e Dir-ision of Petroleum Chemistry a t t h e 110th .\leering of t h e A M E R I C A N C H E M I C A L S O C I E T Y . l ~ p ~ ~ ~ t Chicago, Ill. This paper was originally submitted J u n e 10, 1943, b u t could not be p,lbljshed because of >?artime restrictions. Recent der-elopments ]lave been included in this presrntation.
Sorption of Nitrogen and Water Vapor on Textile Fibers JOHN W.ROWEN ;I&D R. L. BL.4INE S a t i o n a l Bitrearc of Sturtdnrds, W h s h i n g t o n , D . C . J\Ieasurenients were made of the adsorption of nitrogen and water vapor on six purified textile fibers and titanium dioxide. All the fibers had a relatively lo\+ capacity for adsorption of nitrogen as conipared with capacity for adsorption of water vapor. The surface area values ranged from 0.31 square meter per gram for njlon to 0.98 square meter per gram for viscose rayon. The values of the free surface energies of adsorption as calculated by the Gibbs adsorption equation were the same for wool, cotton, silk, and rajon fibers but differed for the two sjnthetir pollmers, nylon and acetate rajon.
T
H E R E is evidence t h a t surface characteristics of textile fibers have some effect on the properties of the finished textile fabrics. Examples of such evidence of surface properties are apparent in the phenomenon of lvater-repellency ( 7 , 28), in the loss of light-reflectance (Si?), and in t’he difference in moisturesorption capacities of fabrics (22, 23, 26) after various treatments. While the iniportance of surface characteristics and properties has long been recognized in t h e technology of colloids and other sciences, the relation of these surface properties t o the properties of fabrics is not, well understood. Several yorkers ( 2 , 3, 13, 26, 27, 29, 50) have interpreted the adsorption of tvater by some textile materials (such as cellulose, u-001, nylon, etc.) as a phenomenon in multimolecular ( 1 2 ) adsorption. If one assumes t h a t the adsorption is a multilayer phenomenon, one m a y calculate t h a t t h e surface area accessible to the first layer is in the range from 100 t o 200 square meters per gram. This range of values is several orders of magnitude greater than the value (0.60 square meter per gram) calculated by Emmett and DeWitt (18)from nitrogen adsorptionmeasurements on paper cellulose. On the other hand, it is not appreciably greater than t h e value (53.0 square meters per gram) reported by Purves et al. ( 2 ) working n-ith nitrogen on a sn-ollen cotton cellulose.
In order to obtain reliable surface area measurements of textile materials it is necessary t o measure the surface area under conditions ensuring true physical adsorption. Brunaucr, Emmett, and Teller ( 1 2 ) and Harkins a n d J u r a (41) have shown t h a t the adsorption isotherms of nitrogen at - 195 O C. (the boiling point of nitrogen, a t which temperature nitrogen is physically adsorbed on solids) may be used t o calculate the true surface areas of a great many crystalline and amorphous materials. Knowledge of the molecular area or surface area ( 8 )is believed necessary in order t o evaluate such properties a s the free surface energy ( 5 ) of adsorption and the Fvork of adhesion ( 2 0 ) betrveen the liquid and the solid phase. It therefore seemed highly desirable t o measure the surface areas of the more common textile fibers-cotton, wool, viscose, silk, nylon, and cellulose acetate-anti t o obtain the Tvater adsorption isotherms of the same samples used in the niti,ogen adsorption experiments. The work lirescnted here represents a portion of the preliminary work of a larger program sponsored a n d supported by the Office of t,he Quartermaster General, War Department. It is intended t o provide comparative d a t a on the adsorption of t h e two vapors on six different purified textile mat,erials. 3lATERIALS AND THEIR PREPAR4TION
Six purified textile fibers and a sample of titanium dioxide (anatase) n-ere used in this study. The cotton, wool, silk, and rayon were from the samples .?tudied b y Wiegerink (31). T h e titanium dioxide was employed as a reference standard. T h e cotton was purified by extracting 8 hours n-ith alcohol a n d 8 hours T i t h ether, followed b y four T h e wool was washings in distilled water at 50 t o 60” C:. extracted with ether a n d alcohol until t h e extracting liquid was free of residue traces, then m s h e d at 50” t o 60” C. in distilled water. T h e ravi silk was degummed in soap solution, extracted with alcohol a n d ether, then thoroughly washed in distilled water. T h e viscose rayon \vas washed four t,imes in a 1% water solution
1660
INDUSTRIAL AND ENGINEERING CHEMISTRY
Vol. 39, No. 12
yarn were obtained by placing them over phosphorus pentoxide until constant weight was obtained-about 5 days. No loss of weight occurred from the fifth to the thirtieth day. After this drying period, the fibers were in equilibrium with a n atmosphere Jvhose relative water vapor partial pressure was less than 0.01. After the dry equilibrium point had been attained, the moisture content of the sample was determined at a series of increasing relative water vapor pressures. The rate of change of weight was plotted against time a t each partial pressure and the equilibrium value, usually attained nithin 2-2 to 48 hours, was indicated by the rate curve. The adsorption of nitrogen was determined by a method similar to the one used by Brunauer and Emmctt (16). The saturation vapor pressure of the nitrogen was determined by use of a vapor pressure Y 0.1I I I I 1 I I I 3.2 03 04 05 0.6 G.7 C.8 79 thermometer, using the purified gas employed for the adsorption experiment. Figure 1 . Adsorption Isotherms of Water and Nitrogen on The low temperature adsorption tests were made on Titanium Dioxide 20- t o 50-gram samples of the fibers. The sample t d b was packed as tightly as possible for each deof ammonia at 50" to BO" C. and t h m washed four times ~ v i t h termination t o reduce the dead s p a c ~correction. For the distilled water a t the same temperature. experiments reported in this article, the tevtile fibers n-ere evacThe sample of nylon used in this study was supplied by E. I. uated for 16 to 20 hours at room temperature (about 25' C.) du Pont de Semours &- Co. and v a s 40-deniei 13-filament yarn t o a pressure of about 10-6 mm. of mercury. Other prelimhaving one half turn per inch of Z tmirt. Six hours' extraction with ethyl ether proved sufficient for complete removal of the inary tests viere made in which the temperature of evacua0.4Gc of oil-base lubricant applied during spinning. The acetate tion n-as higher and one experiment was made of the cottonrayon was 150-denier 46-filament yarn having 2.5 turns per fiber surface available to oxvgen. For the ]OK temperature inch and was supplied by the American Viscose Co. The adsorption of titanium dioxide the temperature of evacuation sample n a s washed four times in distilled water at 50" t o 60" C. was 200" C. Helium was used in evaluating the dead space in the sample bulb. TEST METHODS The sorption of water by the textiles was determined gravimetrically. An analytical balance was used in conjunction with the conditioning apparatus described by Carson and IT'orthington (14). I n this apparatus the partial pressure of the water is maintained by means of saturated salt solutions and the samples may be weighed vithout removal from the conditioned atmosphere. The moisture vapor pressures were obtained from wetand dry-bulb temperatures and also by means of a Dunmore (1) electric hygrometer. T h e dry weights of the 5-gram samples of
AD SORPTION \I EA S U R EM E 3 TS
-4s shown in Figures 1, 2, and 3, the adsorption isotherms of both rvater vapor and nitrogen were sigmoid-shaped, corresponding to type I1 of the five types classified by Brunauer (11). This type of adsorption curve is characteristic of physical adsorption on rigid solids with pore diameters larger than the molecular diameters of the adsorbed molecules (17). However, it
O E 240 n
07
N
5.0 0
x P
J
g
3 cc m 5
a
Of
LL
40
l9
0 :
W a
n
8
Q
210-
l& 2 0 0
X
-
D
190180a: 170160150
0
2W
V A
WOOL VISCOSE SILK COTTON ACETATE NYLON
I IO0
9 z (3 W
0 m
(L
t O? z
2 0
B
5 B
VI
VI
02
d
2
='
1.0
0.
oc 00
01
0 3 0 4 0.5 0 6 RELATIVE NITROGEN VAPOR PRESSURE P/P, 0 2
07
Figure 2. Adsorption Isotherms of Nitrogen on Six Textile Fibers at 19.5" C.
-
z 2a
70 3 110 8 100
60
90 80
50
Ln
5
2 111
80
(L
04
2 111
90
3.0
z W
2 a
5
a
6
230 220-
2
> 4 111
w
5 LL
w
40
70 60 I 50 40 30 20 IO 0
d
I 3.0
2
I
0
0.1
02 0 3 0 4 0 5I 0 6 0 7 0 8 RELATIVE WATER VAPOR PRESSURE PIP.
09
I O
Figure 3. -4dsorption Isotherms of Water on Six Textile Fibers at 25' C.
INDUSTRIAL AND ENGINEERING CHEMISTRY
December 1947
1661
also occurs with nonrigid so1id.q in which the ad~ o r p t i o nsites are not restricted to the surface. It may he noted from Figure 1 that the millimoles of water vapor and nitrogen adsorbed b y the titanium dioside differ only by a small percentage (16). This is true despite the fact that the nitrogen was adsorbed after thorough evacuation, whereas water vapor was adsorbed from mixtures in air after the sample had been dried over phosphorus pentoxide. On the other hand, comparing the adsorption isotherms of the textile fibers in Figures 2 and 3 and the data in Table I reveals that, the orders of magnitude of adsorption capacity of nitrogen and water vapor are iiot the same. All the fibers had relatively (on. capacity for the adsorption of nitrogen as compared n-ith the relatively high capacit,y for adsorption of water vapor. The wool and viscose rayon fibers had slightly Figure 1. 13.E.T. Plot. for Obtaining Surface Areas ircessible to greater nitrogen-adsorption capacities than silk Titrogen and cotton, and the adsorption capacities of the acetate and nylon fibers were the lov-est. In the the acetate and nylon had the loxest capacities. Tl~c~re '\vas, adsorption of rvater vapor, the wool and viscose rayon fibers however, a certain lack of parallelism and shifting of rc~lative also had greater adsorpt,ion capacity than silk and cotton, and position. Sufficient study has not been made to indi,,:itr. the significance of these differences. CA LCL'LATED QUANTITIES .\Iillinioles oi Water Vapor per G r a m of Solid Viscose \?*
