INDUSTRIAL AND ENGINEERING CHEMISTRY
May 15, 1934
223
TABLE 111. ALKALICARBOXATES AND ACETATES IN SULFONATED and 0.10 per cent, respectively. The accuracy of these reOILS BY NEW TITRATION METHODS sults is well within the experimental error and is considered SAMPLE A SAMPLE B SAMPLEC satisfactory for the analysis of sulfonated oils. Alkalinity as free fatty acids (‘4) Found Calculated Alkalinity as soap ( A s ) Found Calculated Total alkalinity ( A t ) Found Calculated Alkalinity aa soap ( A , ) and a8 carbonate ( A d Found Calculated Sodium carbonate Found Calculated Difference Sodium acetate Found Calculated Difference
.
.
MQ. KOH/Q.
MQ KOH/Q.
MQ KOH/Q.
0.00 0.00
4.5 4.4
0.00 0.00
34.0 33.4
28.4 29.0
34.2 33.4
59.2 59.8
49.9 49.6
56.9 56.9
... ...
... ...
45.9 46.6
%
%
2.38 2.49 0.11
..
.. ..
%
.. ..
..
1.11 1.25 0.14
3.14 3.01 0.13
1.61 1.51 0.10
ACKNOWLEDGMENT The writer is greatly indebted to R. A. Pingree of the U. S. Finishing Company, Providence, R. I., for helpful suggestions and to M. B. Hart of this laboratory, who did most of the analytical work.
LITERATURE CITED Alsop, W. K., and Schultz, G. W., J. Am. Leather Chem. Assoc., 16, 525 (1921). Burton, D., and Robertshaw, G. F., J. Intern. Sac. Leather Trades Chem., 17,293 (1933). Commercial Standard CS43-32, “Grading of Sulfonated (Sulfated) Oils, Saponifiable Type,” Bur. Standards, January 26, .nnn
H a r t , R.,
are given in Table 111. Sample A was made by mixing a titrated solution of sodium carbonate with an equal amount of a neutralized sulfonated oil of known alkalinity. Sample B was a mixture of a titrated solution of acetic acid exactly neutralized with caustic soda, and an equal amount of a sulfonated oil with known alkalinity and acidity. Sample C was a mixture of 25 per cent each of the carbonate and acetate solutions, and 50 per cent of the neutralized sulfonated oil. The results in Table I11 are the averages of several determinations in each case. It will be noted that the amount of sodium carbonate found in sample A differed from the calculated value by 0.11 per cent, that the difference in sodium acetate for sample B was 0.13 per cent, and the differences in sodium carbonate and sodium acetate for sample C were 0.14
J.Am. Leather Chem. Assoc., 15, 404
(1920).
Ibid.,16, 159 (1921). H a r t , R., J. IND.EKG.CHEM.,10, !998 (1918). H a r t , R., Ibid.. 21, 85 (1929). Herbig, W., Chem. Umschau Fette, Ole? Wachse Harze, 34, 330 (1927). MoBain, J. W., and McClutchie, W. L., Jr., J . Phys. Chem., 36, 2567 (1932). Nishizawa, K., Chem. Umschau Fette, Ole, Wachse Harze, 38, 1 (1931). Trotman, S. R., and Gee, G. N., J. Soc. Dyers Colourists, 49, 132 (1933). Wizoff, Chem. Umschau Fette, O l e , Wachse Harze, 38, 34, 132 (1931). RECEIVED March 2, 1934. Based largely upon work of the Research Committee of the American Association of Textile Chemists and Colorists, the author being chairman of the Subcommittee on Methods of Analysis and Standardization of Sulfonated Oils.
Determination of Alpha-Cellulose C. KILBOURNEBUMP,870 Longmeadow St., Longmeadow, Mass. It will be seen that the effect of the concentration of t h e HE test for alpha-cellulose has long been a matter of controversy. In 1911 Jentgen (1) suggested a test alkali is much greater than the effect of time, after some for resistant cellulose and advised its use in determin- period less than 45 minutes. For this reason, runs were made ing the suitability of fibers for viscose manufacture. Since with 17.5 per cent sodium hydroxide on a sample of pulp for 2, that time so many other suggestions have been made, most of 5, 10, 30, and 45 minutes, and on a sample of absorbent, them mere variations of Jentgen’s original test, that tests have cotton for 1, 2, 5, 10, and 50 minutes. The method proposed been drawn up by a committee of the Division of Cellulose by the divisional committee of the AMERICANCHEMICAL CHEMICAL SOCIETY (3) and by SOCIETY Chemistry of the AMERICAN was followed as far as possible. The period after the German Chemical Society (4). In both cases, however, maceration was changed as required for runs down to 15, the methods are purely empirical, the determination having minutes’ total mercerization. Runs of 10 minutes’ duration been chosen which gives the most consistent results. This is were treated with the alkali in four parts, 2.5 minutes apart, admitted to be true and everyone recognizes the limitations in and were macerated continuously. Runs of 5 minutes or less were treated with the total volume of alkali all a t once and using the determination. The concentration of the alkali used in the treatment of were macerated for the full time. The results of durations cellulose by these two methods is of importance. In the less than one minute were variable and showed that the solupresent research a sample of paper was treated with sodium tion failed to wet the sample completely or to have sufficient hydroxide of strength varying from 1 to 17.5 per cent for 45 contact t o act uniformly on the sample. The data a r e minutes and for 24 hours a t room temperature following the given in Table I1 and are plotted in Figure 2. SOCIETY’S procedure prescribed by the AMERICANCHEMICAL committee. The data are offered in Table I and are plotted TABLE 11. EFFECTOF TIME in Figure 1. (17.5 per cent sodium hydroxide used)
T
TABLEI. EFFECTOF TIMEAND CONCENTRATION OF ALKALI SODIUM HYDROXIDE % 1 2 4 10 17.5
ALPHA-CELLULOSE 45 minutes 24 hours
% 96.27 93.22 91.48 88.74 87.94
% 95.78 91.94 90.73 87.41 86.40
TIM^
Min. 1 2 6 10 30 45 50
ALPHA-CELLULOS~ Cotton Pulp % % .. 98.9 98.8 88:2 98.7 87.5 98.7 86.5 86.4 86.1
..
9s:5
..
224
Vol. 6, No. 3
ANALYTICAL EDITION
In both cases, by the change in titer method: initial concentration, 5.17 there is a great loss equivalents per liter; adsorption, 0.486 to 0.493 equivalents s u f f e r e d by the per mole of cellulose. By taking the adsorption of the cotton cellulose i n t h e in the 17.5 per cent sodium hydroxide (5.25 equivalents per first few minutes liter) to be 0.490, one is able to determine approximately the of r e a c t i o n , and reduction of the concentration in the solutions during the this rate of loss treatment. With a 3-gram sample using 75 cc. of 17.5 per suddenly de- cent sodium hydroxide, the concentration will fall only 0.4 creases until the per cent, so that the end concentration will be 17.1 per cent. cellulose becomes A 4-gram sample will reduce the concentration to 17.0 per cent nearly n o n r e a c - and a 2-gram sample to 17.2 per cent. The size of the sample tive. Apparently therefore need not always be exactly 3 grams. the time before At least one author (5) offers the suggestion that the losses this break occurs in alpha-cellulose suffered during the aging of alkali pulps in and the sharpness the &cose process of the b r e a k de- are due to oxidapend u p o n t h e t i o n b y t h e air. source of the cellu- However, the action must be very lose used. f ALPHA-CELLULOSE If the a l p h a - slow. I d e n t i c a l FIGURE1. EFFECTOF T h E AND CON* CENTRATION OF ALKALI cellulose is a t - samples of absorbtacked much more ent cotton were slowly than the non-alpha-cellulosic part of the material, we treated with 17.5 can easily explain the forms of these curves. The non-alpha- per cent s o d i u m cellulosic part is peptized more rapidly until it is almost gone, hydroxide a t room the loss of alpha-cellulose during this interval being very small. temperature for 50 This is represented by the horizontal portion of the curves. hours. Both were When this part is nearly all peptized, the rate of loss immedi- instoppered flasks. ately falls off because the attack on the alpha-cellulose is Through one, air very slow. The nearly vertical rise in the curves brings this was bubbled which out clearly. The position and form of the break in the curve h a d p a s s e d will, of course, depend on the sample used. If it is almost through 17.5 per pure alpha-cellulose, as in the case of cotton, the amount of cent s o d i u m hy80 0 /5 30 75 impurities will be small and will be removed quickly. The d r o x i d e p r e v i T/M€ /N MtNffTLS break in the curve will be sharp. If there is much material ously. The resiFIGURE 3. DETERMINATION OF ALPHAin addition to the alpha-cellulose in the sample, it will take d u e s r e c o v e r e d CELLULOSEIN COMMERCIAL PULPS longer to remove it, and the change from the period when the were: a i r - f r e e , 98.7 per cent: airmain reaction is the taking out of impurities, to the period when there is only a slow attack on the alpha-cellulose, will be saturated, 98.6 per cent. The effect of oxygen in the air, then, for any alkali treatment using cold 17.5 per cent sodium more gradual. A second series of runs was made using 10 instead of 17.5 hydroxide can be neglected. per cent sodium hydroxide. The pulp was the same as that TABLE 111. EFFECT OF TIME used with the stronger alkali, in the preceding experiment. (10 per cent sodium hydroxide used) TIME ALPHA-CELLULOSE The results of this treatment are given in Table I11 and Min. % plotted in Figure 2. 2 Although the 10 5 per cent a l k a l i is 10 more effective in the shorter times, both 30 10 and 17.5 per cent 45 solutions give the same values after 45 PROPOSED METHOD minutes of reaction. From this one samIn view of these facts, the author proposes for the present ple it would appear a method for the determination of alpha-cellulose which is that the 10 per cent based on the stability of the cellulose to attack by sodium alkali changes the hydroxide. Samples are run according to the method proshape of the curve posed by the AMERNANCHEMICAL SOCIETY'S committee for a t the break some- two or three different times, say, 15, 30, and 45 minutes. what. These times should be sufficient to cause the points to fall on From the data of the vertical flat of a curve similar to those in Figure 2. By Neale ( 2 )we can cal- extrapolation back to zero time, the alpha-cellulose content culate roughly the of the original sample should be given. end concentration It is claimed as an advantage of this method over any /DO 95 90 85 of the sodium hy- other, that the time factor is eliminated. No standard time X A L P H I - CLLLOLOJL d r o xi de solutions. is necessary; in fact, as long as the points lie well beyond the FIQURE 2. EFFECT OF TIME AND He gives the follow- break, the times may be chosen to suit the sample or conCONCENTRATION ON ALPHA-CELLULOSE FROM COTTONAND PULP ing figures obtained venience. Since the attack on the alpha-cellulose is slow, 0
May 15, 1934
INDUSTRIAL AND ENGINEERING CHEMISTRY
extreme accuracy in measuring the time of reaction is not necessary. Slight error in time will result in practically no error in the ultimate alpha-cellulose result. The ratio of the sample to the reagent need not be exact, provided it is within certain limits (1 to 4 grams in 75 cc. of solution). The main precaution seems to be the use of carbonate-free alkali throughout the treatment. The suggested method has been tested on six commercial paper pulps, the alpha-cellulose content of which had been determined previously by the manufacturer, and the results are shown in Table IV and in Figure 3. TABLE Iv. DETERMINATION O F ALPHA-CELLULOSE IN COMMERCIAL PULPS MANU-
225
The author is indebted to Professor Bancroft for suggestions and criticism in this work. LITERATURE CITED (1) Jentgen, Kunststo,fe, 1, 165 (1911). (2) Neale, J. Teztile Inst., 22, T320 (1931). (3) R i t t e r e t al., IND.ENQ.CHEM.,Anal. Ed., 1, 52 (1929). (4) Sohwalbe, Papier-Fab., 23, 477, 697 (1925) [Chem. Abs., 20, 283, 502 (1926)]. (5) Waentig: Papier-Fab., 25, 112 (1921) [Chem. Abs., 22, 4974 (1928)]; 26, 64 (1928)[Brit. Chem. A h . , B47, 564 (1928)l. RECEIVED January 15, 1934. This work was done by the author in partial fulfilment of the requirements for the degree of doctor of philosophy a t Cornell University.
VALUE BY: Calculation Curve 83.60 83.75
84.02
PERCENTRESIDUE AFTER: 15min. 30min. 45min. 83.03 82.95 83.57 83.75 83.36 82.97 84.02 83.14 83.50
B
87.63
87.44 87.35
87.08 87.07
86.88
87.69
87.50
Determination of Sulfur in Benzene or Gasoline
C
91.26
91.01 91.15 91.06
91.10 90.76 90.67
90.80
90.75
91.21
91.25
Modification of A. S. T. M. Lamp
90.74
D
94.90
94.54
94.44
94.46
94.56
94.60
E F
98.54
97.87
97.70
97.63
97.89
97.97
98.42
97.33
96.97
96.94
97.28
97.10
FACTURER’S
PULP A
VALUE
H. 0. ERVIN Portland Gas & Coke Company, Portland, Ore.
I
N A RECENT article, Gillis ( 2 ) describes a modification of the standard A. S. T. M. lamp which provides for better control of the flame. A similar lamp of somewhat simpler construction has been in use in the author’s laboratory for about three years, and has been particularly successful with benzene and its blends, which ordinarily give more trouble than gasoline itself. The modified lamp is sketched in Figure 1. The glass tube A , which carries the wick as in the standard A. S. T. M. lamp, is made somewhat longer than usual to accommodate the sleeve B, which is a glass tube of slightly larger diameter. Because of the variation in tubing diameters, it has been found ALPHA-CELLULOSE advisable to make the sleeve of tubing sufficiently larger than CALCULATED DIFFERENCE IN Loss FROM: FROIM: 15 to30min. 30to45min. %/T 15 30 45 AVERAQE A to require some sort of flexible pack% % % % ing which provides the necessary fric17 7 12 97.99 97.94 97.99 97.97 tion to hold the sleeve in place. If a The average of these zero values is taken as the calculated glass-to-glass friction be used, sticking result. The method of calculating gets rid of any errors in is likely to occur a t a most inopportune plotting and drawing the curve and in estimating the extrapo- moment, with disastrous results. Even a smooth-sliding sleeve as made up may lated value. bind in use, because of expansion of SUMMARY the inner tube when heated by the comA preliminary study of the effect of concentration of alkali bustion of the material under test. A on the tentative standard method proposed by the Division of small piece of cotton wicking in the Cellulose Chemistry of the AMERICAN CHEMICALSOCIETY annular space serves as a satisfactory packing material. determination of alpha-cellulose was made. In operation the sleeve is simply In studying the effect of time on the treatment of cellulose moved up and down on the tube carrywith 17.5 per cent sodium hydroxide a characteristic curve was found which showed rapid losses in weight of cellulose a t ing the wick. If the flame is too high, first and soon approached a nearly constant value. A curve the sleeve is pushed up, and vice versa. of the same type giving the same nearly constant value was Very close adjustment may be made by FIGURE 1 rotating the sleeve slightly as it is found for 10 per cent sodium hydroxide. or down. moved up A method for the determination of alpha-cellulose has been The principle of flame control in this lamp is apparently proposed temporarily which eliminates exact specifications of time and sample-reagent ratio. The suggested method has the same as that employed in one of the lamps proposed by Edgar and Calingaert ( I ) who used a movable brass outer been tested on six commercial pulps. tube, but its construction is simpler, and the control, in the ACKNOWLEDGMENT author’s experience, more satisfactory. The six pulp samples used in testing the method suggested LITERATURE CITED above were furnished, together with the manufacturer’s data, (1) Edgar a n d Calingaert, IND. ENQ. CHEY., Anal. Ed., 2, 104 (1930). through the kindness of G. A. Richter of the Brown Company. (2) Gillis, Ibid., 5, 421 (1933). Other pulp stock was furnished by the Strathmore Paper RECEIVED DECEMBER 9, 1933. Company.
It was found more convenient and perhaps a little more accurate to calculate the percentage of alpha-cellulose back to zero time than to plot the data and estimate the value from the extrapolated curve. This was done by determining the loss for the 15-minute periods, 15 to 30 minutes’ reaction and 30 to 45 minutes’ reaction. The average of these two is designated A%/AT. Using the 15-, 30-, and 45-minute values as bases, the alpha-cellulose a t zero time can be calculated by adding to these values A%/AT, 2A%/AT, and 3A%/AT, respectively. For example, using the data from pulp F:
