933
V O L U M E 25, N O . 6, J U N E 1 9 5 3 two experimental errors. I n any event, it will be seen that good reproducibility of the pentosan analyses was obtained by correcting the ultraviolet results as indicated, and the figures so obtained are in line with those obtained by colorimetric analysis of these distillates for “true furfural.” The improved precision obtained is, of course, du‘e to the fact that errors arising from the use of the less precise colorimetric method have been eliminated by the expedient of substracting a constant value from the results of the more precise ultraviolet absorption method. This same improvement in reproducibility could be obtained with any other constant value eo used, but repeated determinations have been made which shoi5- that this factor can be checked quite (4osrly.
Table TI. Comparative Pentosan 4nal)ses by TAPPI Method and Modified TAPPI Procedure Pentosan, 7& TAPPI Method Modified Method 2.6 2.36 2.1 2.22 2.4 2.08 2.3 1.97 1.9 1.86“ 1.66a 1.6 1.62 1.3 1.1 1.50 1.3 1,25 0.9 1.07
_-
Sample S o .
Y
IO 11 12
a
0 0 0.1 Average of checking ( = k O . O 5 ) duplicate analyses.
0 . 40a 0.38
the TAPPI ca!culation for the amount of hydrosymethyl furfural is approximately compensated for by t’he 0.88 correction factor also applied by TAPPI as discussed above. The effect of this overcorrection is greater for the very low pentosan pulps so that. errors of 100% or more may result by use of the TAPPI mrthotl of calculation. MODIFIED PROCEDURE
On the basis of the foregoing results, it is the conclusion of thcx authors that the present standard TAPPI method must be modified for use with low pent.osan pulps as follows: The distillation rate should be reduced to 1.5 nil. per minute. The total quantity of distillate collected should be r e d u d to 200 ml. For pulps in the range 0.5 to 2.0% of pentosan, an aliquot of the distillat’eshould be diluted 1 : 10 with distilled water and the ultraviolet light absorption a t 277.5 mp should he read directly against a distilled water blank. Using a calibration curve establish(d with pure furfural, t h e quantity of total aldehyde (furfural plus hydrosymethyl furfural) may be calculated. From t,his figurv the true per cent of pent,osans may t)e calculated by the formuh: (grams of total aldehyde - 0,002) x 100 sample weight X 0.727 where 0.727 is the theoretical factor for converting- pentosans to _ furfural. 70
pentosans =
4CKYOW LEDGMEYT
The authors wish to thank A . F. J o h n w i for his assist:ui(v i n carrl-ing out the statiitical treatment of the data obtained LITERATURE CITED
The pentosan calculations in Table V do not include the correction factor of 0.88 which is used in the TAPPI calculations to “compensate for the incomplete conversion of pentosans to furfural.” Despite this, the results of this new method agree with those of the conventional TAPPI method within the limits of error of the latter method a s applied in these laboratories. Typical comparative results are given in Table VI. A probable explanation of these results is that the overcorrection applied by
(1) Adams, G. A., and Castagne, A. E., car^. J. Research, 26, 314 (1948). (2) Rrownlee, K. A , , ”Industrial Experimentation,” 3rd ed., p. 19, Brooklyn, S . Y., Chemical Publishing Co., 1949. (3) Ibid., p. 52. (4) Smith, E. D., and Rogers, L. N., Annual TAPPI Meeting, S c w York, N.Y., Feb. 18-21,1952. R E C E I V E for D review February 13, 1952. Accepted March 11, 1953. 1’1.1.sented before the Division of Celliilose Chemistry a t the 12lst hlevting of the AMERICAX CHEMICAL SOCIETY, l l i l n a u k e r , Wis.
Polarographic Studies of Oxygen-Containing Compounds Acid Anhydrides CONSTANTINE RICCIUTI, C. 0. WILLITS, H. B. KNIGHT, AND D4NIEL SWERN Eastern Regional Research Laboratory, Philadelphia, Pa.
C
O S T I N U E D study of the polarographic behavior of oxygencontaining functional groups in nonaqueous, lithium chloride electrolytic solution ( 4 ) has shown that a,P-unsaturated organic acid anhydrides are polarographically reducible. The polarographic reduction of acid anhydrides produces wave heights which have a linear relationship to concentration. This suggested that the polarographic method would provide a direct means for their determination. APPARATUS
A Sargent Model S X I polarograph was used to obtain the current-voltage curves. The sensitivity settings of this instrument were calibrated against known resistances. The capillary m and 1 values were 3.59 mg. per second and per 1.35 seconds, respectively, which gave a capillary constant, mz/3t’/8, of 2.46 mg.*/asec.-l/S. The 7n and t values were obtained using an open circuit with the polarographic cell maintained a t 25” =t0.1” C. and with the capillary immersed in the nonaqueous electrolytic solution. The electrolytic cell is that described by Lingane ( 3 )and modified l)v Willit- et al. (4).
.\IATERIALS
The electrolytic solution described by Lewis and Quacken~)ush ( Z ) , which consisted of 0.3 M lithium chloride in equal volume.; of ab~olutemethanol and benzene, waq used in these studies. AFHYDRIDES(Eastman Kodak, White Label grade, except as indicated). Benzoic anhydride. Melting point, 42’ to 3 ” C., 97.0% pute [determined by the “anilic acid number” method ( I ) ] . n-Butyric anhydride, 93.8% pure (1). %-Butylester of aconitic anhydride prepared in this laboratory, 97.0% pure (determined by hydrolysis and acidimetry). Caproic anhydride. Citraconic anhydride, 100.0%pure ( 1 1, Crotonic anhydride, 98.6% pure (1). Maleic anhydride, Y8.8% pure (1). Phthalic anhydride, purest commerrial material, 99.8% pure (1).
Succinic anhydride. ACJDS. .4conitic acid, American Sugar Refining Co. Benzoic acid, prepared, in the authors’ laboratory, melting point 122OC. Crotonic acid, Eimer and -4mend, reagent grade. Maleic acid, Eastman Kodak, White Label grade.
ANALYTICAL CHEMISTRY
934
Polarographic studies of acid anhydrides were undertaken to determine whether their polarographic characteristics in nonaqueous system could be used for their identification and determination. a,p-Unsaturated anhydrides were found to be .polarographicallyreducible and the corresponding acids and esters, with the exception of maleic acid, did not interfere with the polarographic waves of the anhydrides. Saturated acid anhydrides did not show polarographic reduction. a,b-Unsaturated acid anhydridea can be identified in the presence of saturated anhydrides and can be determined if the identity of the unsaturated anhydride is known. It is also possible to determine the composition of a mixture of two unsaturated anhydrides if their identity is known, and if their halfwave potentials are sufficiently widely separated.
Phthalic acid, prepared from pure phthalic anhydride by boiling with water for several minutes followed by cooling to precipitate the acid. ESTERS. Dimethyl phthalate, Eastman Kodak, White Label grade. Di-%ethyl hexyl maleate, Carbide and Carbon Chemicals Corp. Methyl benzoate, Eastman Kodak, White Label grade. Methyl crotonate, Eastman Kodak, White Label grade. Monocyclohexyl phthalate, Barrett Division of Allied Chemical & Dye Corp. Monomethyl phthalate, prepared by refluxing phthalic anhydride with methanol for 30 minutes and crystallizing from water. PROCEDURE
A weighed sample was dissolved in a measured volume of the lithium chloride, methanol-benzene, electrolytic solution and the polarographic waves were obtained as described previously (4). All half-wave potential measurements were made against the saturated calomel electrode as reference electrode and all halfwave potentials were corrected for I R drop. Concentrations of anhydrides of less than O.OO040 mole per liter showed no tendency to form maxima whereas with higher concentrations, sharp maxima were formed. The usual maximum suppressors such as methyl red and gelatin had no effect on these maxima. Materials with high viscosities such as the polyacrylates and glycerin were also tried as suppressors but without success. Although abnormally high maxima were obtained with higher concentrations, the wave heights were reproducible when measured a t the minimum point following the maximum, and these wave heights were directly proportional to the concentration of the anhydride. The linearity of the wave heights us. concentration was determined from the four to six polarograms obtained for a number of the compounds. Each polarogram was obtained from individually weighed samples. To determine if there was any interaction between the acid anhydrides and the electrolytic solution, the anhydrides were added in increments to the solution with extended intervals between additions (over one half hour). It was observed that the successive wave heights were directly proportional to the amount of anhydride added. Had there been any reaction between the anhydrides and the solvent, this proportionality would not have existed. DISCUSSION OF RESULTS
alysis of mixtures. Table I1 shows the polarographic behavior of five of the reducible anhydrides and eleven acids and esters derivable from them. The table shows that, with the exception of maleic acid, the acids or esters either are nonreducible or produce noninterfering polarographic waves. I t should thus be possible to determine benzoic, crotonic, phthalic, and aconitic anhydride in the presence of their corresponding acids or esters. In the case of maleic anhydride, however, the first wave of maleic acid = -0.84 volt), would interfere with the wave of maleic anhydride ( E l / z= -0.72 volt). It is not possible to distinguish between the acid anhydrides in a mixture of two or more when their half-wave potentials are too close to one another to permit differentiation of the separate waves. For example, the maleic anhydride wave, having a halfwave potential of -0.72 volt, cannot be distinguished from that of butyl ethoxy ester of aconitic acid anhydride ( E l p = -0.77 volt) when these anhydrides are present together. However, the E112 of the maleic anhydride wave is sufficiently distant from that of phthalic anhydride (E1/* = -1.12 volts) to permit their quantitative estimation when they are present together. Figure 1 shows a typical analysis of such a mixture. Curve A is the
Table I.
Polarographic Characteristics of Reducible Acid Anhydrides
S o . of Concn. Range, Acid Anhydride Tests Millimoles/Liter Maleic 4 22.5-51.7 Aconitic (butyl 4.7-16.7 estei, 5 6.5-63.1 Citraconic J 7.4-21.4 Phtha!ic Benzoic 6 8.8-16.2 Crotonic 5 3.5-21.6 a Corrected for I R drop.
-
Half-Wave Potentialn -0.72 +0.02 -0.77 -0.84 -1.12 -1.62 -1.62
___ id Cm.*'W/a
2.145 0 . 0 3
I 27 * n 03 5 0 . 0 'i i.76 ZO:OE 5 0 . 0 0 2.54 5 0 . 0 3 5 0 . 0 2 3.08 5 0 . 1 2 3 ~ 0 . 0 33 . 3 3 k O . 0 6
*0.02
Table 11. Comparison of Polarographic Characteristics of Acid Anhydrides, Acids, and Esters id
Of the nine anhydrides listed above only the six with a,@-unsaturated linkage8 were reducible in the nonaqueous electrolytic solution used (0 to -2 volts). A similar observation had been made in a previous study ( 4 ) when only a,@-unsaturated aldehydes and ketones were found to be reducible under these conditions. None of the saturated anhydrides was reduced, indicating that under the conditions used unsaturation ie required for reduction and that polarographic determination of a$-unsaturated anhydrides in the presence of saturated anhydrides is possible. Table I shows the polarographic characteristics of the six unsaturated acid anhydrides. It is apparent from the table that the diffusion current constants are independent of concentration in the ranges studied. A number of acids and esters corresponding to the reducible anhydrides were studied to determine whether they were also reducible and, if so, whether their half-waves were sufficiently different from that of the anhydride not to interfere with the an-
Compound Benzoic anh dride Benzoic a c i d Methyl benzoate Crotonic anhydride Crotonic acid Methyl crotonate Phthalic anhydride Phthalic acid Mono-methyl phthalate Dimethyl phthalate Mono-cyclohexyl phthalate Maleic anhydride Maleic acid
Half-Wave Potential -1.62 None' None -1.62 Xone None -1.12 Sone -1.48 Xone -1.56 -0.7: 1 . 3 8 (PO, -0.84 1 . 3 3 (pot -1.30
-
Cm.Z/3tl/I
3.08 il'one None 3.33 None None 2.54 None 2.02 None 1.86
Di-2-ethyl hexyl maleate Butyl ester of aconitic acid -0.77 1.27 anhydride -1.24 2.84 Aconitic acid 0 No reduction wave from 0.0 to -2.0 volts in the electrolytic solution used.
935
V O L U M E 25, NO. 6, J U N E 1 9 5 3
I
I
1
I
I
I
I
1
1
u)
Table 111. Analyses of Solutions of Anhydride Mixtures
W
:zoo -
n
Anhydride Mixture Solution 1 phthalic maleic Solution 2 phthalic maleic Solution 3 phthalic maleic
I
8K -01 1 5 0 F 2 W
U U
3 100
Amounts (Grams/4O Ml.) Added. Found 0.0461 0.0443 0,0451 0,0472 0.0335 0.0356 0.0953 0.0938 0,0525 0.0496 0.0677 0.0689
-
Relative Error, % -1.8 f4.7 +6.3 -1.5 -5.5 f1.8 Av. f 3 . 6
0
z
2
In
3
-
50-
n
I
0 0
I
0.5
10
1.5
2.0
Figure 1. Polarograms A.
Phthalic anhydride
B. Mixture of phthalic anhydride and maleic anhgdride
polarographic m-ave of a solution of 0.01280 mole per liter of phthalic anhydride. Curve B is the wave of the same solution of phthalic anhydride to which maleic anhydride was added. The wave height, hl, of the 0.01280 M phthalic anhydride, when present alone in the electrolytic solution, is equal to hz, the wave height of phthalic anhydride in the presence of 0.01647 ill maleic anhydride. This suggested the possibility of determining these two anhydrides when present together in mixtures. Curiously enough, mixtures of the two anhydrides do not produce the maximum obtained from solutions containing only one (Figure 1). To test the possibility of analyzing a mixture of anhydrides having distinctly different half-wave potentials, a solution (0.3 M lithium chloride, 50-50 benzene-methanol) containing weighed amounts of both phthalic and maleic anhydrides was prepared. The polarogram obtained with this solution resembled curve B ,
Figure 1, in having two distinct waves, one a t -0.75 volt due to the maleic anhydride and the other a t -1.16 volts due to the phthalic anhydride. From the proportional increases in the respective wave heights caused by adding weighed increments of phthalic and maleic anhydride, the quantities of phthalic and maleic anhydrides originally present in the mixture were calculated. Table I11 shows the results obtained using three solutions of mixtures of varying amounts of phthalic and maleic anhydride. These data show that it is possible to estimate the amounts of each of these anhydrides, when present together, with a relative percentage error of Z!Z 4%. ACKNOWLEDGMEhT
Determinations of the anilic acid numbers of the acid anhydrides were made by R. E. Koos; the butyl ester of aconitic acid anhydride was prepared by M. L. Fein and E. H. Harris, Jr. LITERATURE CITED (1) Kappelmeier, C. P. -4., and van Goor, W. R., Anal. Chim. Acta, 2. 146-9 (1948). (2) Lewis, W. and Quackenbush, F. E'., J . Am. Oil Chemists SOC., 26, 53-7 (1949). (3) Lingane, J. J., and Laitinen, H. A , , IND.ENG.CHEM., ANAL.ED., 11,504-5 (1939).
k.,
(4)Willits. C. O.,Ricciuti, C.. Knight, H. B.. and Swern, D.. AKAL. CHEM., 24,785 (1952). RECEIVED for review August 31, 1951. Bccepted March 26, 1953.
Anomalous Flow of liquids through Capillaries and Measurement of Viscosity G. F. N. CALDERWOOD, H. W. DOUGLAS, AND E. W. J. MARDLES Royal Aircraft Establishment, Farnborough, England
T
H E method of determining, by means of capillary viscometers, the viscosity of liquids, in centimeter-gram-second units, relative to that of water [LOO19 centipoises a t 20.00" C. (6)],used as the standard calibrating liquid, involves a possible error owing to the unique surface tension properties of water, the surface tension of water being about twice that of oils and other organic liquids. The design of the master or secondary viscometer with longer length capillaries than in the ordinary routine type of standard viscometer is such as to minimize, within a prescribed tolerance, any error due to surface tension differences. The standard method demands scrupulous cleaning of the viscometer with chromic acid mixture and rinsing, prior to the measurement of flow time. This cleaning is performed presumably, apart from removal of dust fibers, etc., to ensure a low contact angle of the water with the glass surface. Care is required to ensure that high contact angles are avoided, otherwise a meniscus resistance, the Jamin effect, comes into operation and may afiect precision. This meniscus resistance
becomes relatively large in microviscometers which are also adversely affected by another disturbing factor caused by the reversal of liquid flow a t the meniscus. Lillard, in a recent paper ( 4 ) , described a micromethod for determining the viscosity of oils by noting the rate of flow of an index of liquid in a tilted capillary tube. He used 0.04 ml. of liquid and claimed i~ 1% accuracy. Other workers ( 7 , 8),using this method of a column of liquid in a capillary, obtained with mercury a value 6% higher than the usually accepted coefficient of viscosity and attributed the difference to electrification a t the walls. During flow through a capillary, a drop rotates from the walls inwards so that the Poiseuille equation needs modification to allow for the reversal of flow a t the menisci. The importance of this can be shown by noting the apparent viscosity of an oil using columns of different lengths. With short lengths the apparent viscosity becomes abnormally high, reaching two or three times the normal value. The viscosity results given in Table I and Figure 1 were obtained. In an experiment a t 20" C. with
