compose dmylose, arid the value obtained for iodine absorption may be only a practical maximum and fortuitous. Packing considerations alone would allow about one iodine molecule for five glucose residues, since w e molecule would require about 6.S A. along the helix. Furthermore, it is hard to understand why reflections of the type (OOha) are absent if the iodine molecules are spaced 7.91 A. apart along the helix. It seems possible that their arrangement is a t random in a set of six equivalent positions per turn in the helix, giving rise to an effective sixthing of the spacing along the c axis. Unfortunately, the temperature factor for the complex i s so high that the reflection (OOCi) could not pos\ihl?- bc observed at rouni temperature This leaves the space group in doubt. Summary I
I’he amylose iodine complex has a hexag-
unit cell, uo = 12.97, GO = 7.91, dloo = 11.23 A. 2 . The unit cell confirms a helical structure for the starch-iodine complex; 12.97 A. is the diameter of the helix, 7.91 is the length of a turn in the helix. These dimensions are in good agreement with the dimensions of a space-filling model of a helix with six glucose residues per turn. :i. The starch-iodine complex can be prepared entirely without water or iodide ion if the starch is first put in the “IT” configuration by alcohol pre-ripitation. 1. Starch in the “V” configuration will absorb iodine vapor in quantity, while in the “A” and “ B ” ccmfiguration i t will not. dmylose in the “ir”configuration will absorb 26% of its own weight in iodine. This corresponds to one iodine ior six glucose residues, but it is not established that this is tile maximum ioditie absorptioii . oiid
A.
..._.
[ C O N I K I B l ~IIOX FROW I M E
NO\
KiicmveD APRIL 17, 1
A W S , 1.3 .
t b CHbMICAL I . A I ~ ( J K A l O K \ , [‘NIVEKSITY
OF ILI,INOIb]
End Group Structure of Polyvinyl Alcohol’ BY
(.
3
LIAKVEL A N D
’The recurring unit in polyvinyl alcohol has been shown t o be that indicated in Formula I and the
r-CH&C)--] L.
I
chemical behavior of the polymer is in general that which would be expected of a poly-1,3-glycoL2 However, Stauclinger and Warth3 have found that alcoholysis of a given polyvinyl acetate by different methods produces specimens of polyvinyl alcohol having different properties. Blaikie and Crozier4 have noted unusual changes in viscosity in polyvinyl acetate samples on hydrolysis and reacetylation. Also they have found that mineral acid causes polyvinyl alcohol to become insoluble. McDowell and Kenyon5 have found that saponification of polyvinyl acetate by alcoholic potassium hydroxide followed by reacetylation by means of acetic anhydride and pyridine ( 1 ) This is the sixteenth communication on the structure of vinyl polymers. For the fifteenth, see THISJOURNAL, 6S, 1647 (1943), (2) Staudinger, Frer and Stark, Ber., 60, 1782 (1927); Herrmann and Haehnel, i b i d . , 60. 1658 ( 1 9 2 7 ) ; Marvel and Denoon, 'hit JOURNAL, 60, 1046 (1938) (3) Staudinger and Warth, J W o k ! . C h i n . , 165, 261 (1940) f-11 Blaikie and Crozier, J x d J i u k Chum , 28, 1155 (1936) o n ‘lirii
J(IVRV\T
62 11.5
,lV11’
c.ESLEK INSKEEP
usually results in a reduction of chain length as measured by viscosity methods, by about onehalf. Industrial experience6 with these compounds has shown also that unusual molecular weight changes, based on viscosity determinations, on hydrolysis of polyvinyl acetates or acid treatment of polyvinyl alcohols, are not uncommon. I n the present communication are described some experiments on the hydrolysis of polyvinyl acetate samples and acid treatment of various polyvinyl alcohols in which the chain lengths measured by viscosity methods show marked changes. These changes may be in either direction depending on experimental condition^.^ Viscosity molecular weights and degrees of polymerization were calculated by nieans of the Staudinger equation using a K, value of 2.6 X lo-* for both polyvinyl acetate in benzene and polyvinyl alcohols in water containing 2% ethanol. To solutions of polyvinyl alcohol which contained (6) Private communication from Dr. B. C. Bren. (7) The samples of polyvinyl acetate and polyvinyl alcohol were kindly furnished by Dr. B. C. Bren of the Plastics Department, E . I iiii
Pont de Kernours and Company. and we are indebted for his :*id il’i
i t : r i i d i n c p r a n d Schwalhnrlr
.1
, 468, R (10:31)
Sept., 194:J
ENDGROUPSTRUCTURE OF POLYVINYL ALCOHOL
lill
acid, enough dilute alkali was added to make This treatment seemed to cause about 80% rethem slightly basic to litmus before viscosity acetylation. measurements were made. The molecular weight Three samples of polyvinyl acetate were studvalues calculated from these viscosity measure- ied. All were prepared by polymerization of ments are probably valuable only for comparaCHARTI tive purposes. CHsOH Alcohol Various treatments of a given sample of polyHzS04 D P 111 mer were carried out to determine the effect upon Polyvinyl acetate D P 66 CHaOH the degree of polymerization. The standard Alcohol treatments for polyvinyl acetate samples were, CH3ONa DP 102 (1) hydrolysis by means of methanol and sulfuric acid a t reflux temperaCHaOH ture for two hours using 0.9 mole of Polyvinyl Acetate + Alcohol acid per mole of monomer unit folDP216 CHaONa D P 222 lowed by filtering, dissolving the CHaOH polymeric alcohol in water, precipiWater Pyridine tating b y pouring the solution into Alcohol Alcohol Acetate excess ethanol or acetone, decanting, D P 349 HZSOI DP 159 (CHaC0)20 D P 145 Pyridine washing with ethanol or acetone, filtering, washing with ether, and airCHaONa /&so4 CHoOH drying; and (2) saponification by CHaOH J. CHsOH treating the acetate with methyl al- Acetate _ ~ _Alcohol f *- Pyridine Acetate Alcohol coholic sodium methoxide (using 1 DP 376 H2S04 D P 277 (C&CO)ZO D P 476 D P 163 Alcohol CHaOH mole of methoxide per mole of monoDP 77 CHaONa &So4 mer unit) for one and one-half hours, Alcohol followed b y isolation of the polymer D P 193 HzO Alcohol and purification by washing as above. HzSOd DP205 Polyvinyl alcohol samples were Alcohol given several treatments to see what D P 1440 Pyridine the effect on chain length would be. (sample aged I Acetate One treatment was to boil a 5o/c soh- two week51 (CHqC0h0 D P 421 tion in 5% sulfuric or hydroCHARTI11 chloric acid for two hours and Polyvinyl Acetate then isolate the polymer by diDP 1570 luting with four volumes of CHsOH water and pouring this solution &SO4 into five volumes of ethanol or Pyridine HzO Pyridine Acetate Alcohol v Alcohol Acetate acetone. Another treatment DP 304 (CHaC0)zO DP 481 H ~ S O ID P 506 (CHaCO)2O D P 446 was to allow the solid polymer containing a trace of mineral Solid CHaOH CHaOH aged acid to stand a t room temperatwo ture. A third treatment was Alcohol Alcohol weeks Alcohol Alcohol with 1 DP406 DP399 DP 1280 DP 1530 reacetylation by heating the polymeric alcohol on a steamSolid Pyridine stood acid I (CHaC0)zO bath with ten times its weight of a mixture of two parts of acetic Alcohol weeks Alcohol Acetate DP 522 with DP, 134 D P 264 anhydride to one part of pyriacid Solid dine for two to twenty-four Alcohol D P 567 hours. Usually the polymer dissolved and i t was isolated by acid pouring the reaction mixture Alcohol into a large volume of water. DP 384 and DP 799
E
A
lwater
-
I
c
C. S. MARW,LAND G. ESLERINSKEEP
1712
Vol. 65
a sample usually results in formation of an ester of shorter chain length. These changes in apparent chain length are very irreqlar and unpredictable in amount. A possible explanation of these changes can be given if polyvinyl alcohol has an aldehyde end group (11). Such a group would be expected to react with the alcohol groups in the same or CHART11' neighboring chains t o produce an acetal groupPolyvinyl alcohol ing (III).g The exact way in which any two D P 347 (Polyvinyl Hz0 molecules would combine would be largely a acetate prepared ---+ Alcohol with benzoyl peroxide HzSOd D P 342 matter of chance, and hence viscosity molecular catalyst and hydrolyzed weight measurements, which measure chain by CHaOh-a in CHIOH) length rather than true molecular weight,'" CHART 1' would show odd results. Thus, depending CHIOH H2O +-r Alcohol Alcohol on the size of n and y in Formula 11, the H2S04DP 484 Hzsot DP 261 viscosity molecular weight of I11 would be Polyvinyl alcohol D P 331 (H202 catalyzed HzO H.0 d dderent fraction of the true molecular and CHaOh'a and CHIOH L---weight. The fact that the unusual changes hydrolysis) FIzsO< DP 327 &So4 DP 2&8 111 dpparent chain lengths occur most freLHAKIV I quently 017 treatment with acids is in agreePolyvinyl alcohol HZO Pyrtdiiir DP 470 (H202catalyzed, Alcohol Acetate ment with the idea that acetal formation hydrolyzed with alcohol Heso4 D P 365 (CHaC0)20 DP 48'3 is involved. and sulfuric acid) Under alkaline conditions a polymer with CHARTVI1 d terminal aldehyde group (IV) might bePolyvinyl alcohol CHaOH Pyridine DP238 (HzOzcatalyzed; -----+ Alcohol ----+- Acetate come involved in a reversible aldol conCRaONaand CH,OH H-SOI DP 506 (CH~CO)ZODP + $ ? densation to give a molecule (17) of much 5aponification) higher molecular weight which under differThe transformations noted in the case of poly- ent conditions could hydrolyze back to the origivinyl alcohol samples in various experiments are nal. Also under alkaline conditions a reverse summarized briefly in Charts IV-VII. aldol condensation might occur to cause succesIt will be noted 0 that in general al-CHzCH- -CH2CH -CH2CHCH2C< + - -CH2CH- -C&C '>CJ!H-VH
0
1713
CH-CHz (AH
>,
-
~ - 1 . Calcd, for -
CHZCH=NNHC~H~(NO~)~:N,~.~~Q Found: N, 0.338.
+ CH3CN0
This agreement in molecular weight is certainly as close as OH),-, can be expected. The derivaVI tive gave the typical violet color expected to have the structure indicated in For- with alkali which is characteristic of 2,4-dinitromula VI1 or VIII. phenylhydrazones. l 2 HO CH2CH-CHzCHOH + The absorption spectra of this 2,4-dinitrophenylhydrazone derivative of polyvinyl alcohol OCOCH,), VI1 &XXHs and the corresponding derivatives of n-butyraldeo hyde and crotonaldehyde were taken in a Coleman HO CHzCHmodel 105 double monochromator spectrophotometer coupled with a Coleman model 310 elecOCOCHI)~cHzc'H 17111 trometer using a slit width of 15 mp and a cell If the &&-ending reaction is a disproportions- thickness of 13.0 mm. The range covered was tion the end group in polyvinyl acetate is prob- from 35(r900 mk These are plotted in Fig. 1 and it will be noted ably an enol acetate (IX) which on hydrolysis an aldehyde group (x)at the end of that there are certain definite similarities. would the polyvinyl alcohol chain. Infrared absorption spectral3 of polyvinyl HOH CHzCH-CH=CH alcohol and its 2,4-di4.0 nitrophenylhydrazone &LOCH,>. hCOCHa were taken with films IX 0 of the polymers de3,0 - -CHsCH- -CH,C/\H posited from water solution on fluoride slides. v' AXH ) The exact thicknesses M were not 9 2.0 An attempt to demonstrate an aldehyde end Of the group in polyvinyl alcohol by hypoiodite oxida- determined but Care tion as has been done for starch moleculesll was was taken to see that unsuccessful. Apparently, the secondary alcohol were 1.0 groups in polyvinyl alcohol are so readily oxidized Measurements were as to interfere with this method. However, it made at the range has been found possible to make a 2,4-dinitro- 5-5-6*2I.1. 0 phenylhydrazone derivative of a polyvinyl alcoIn the polyvinyl alco350 400 450 500 hol having a degree of polymerization of 331 a s hol absorption there are mfi . determined by viscosity methods which has a ni- two peaks (Fig. in Fig. 1.-Visible absorption trogen content indicative of a molecule containing the region typical of the spectra of dinitrophenylhycarbonyl group (5.8-6.2 drazones: 0,crotonaldehyde; 372 vinyl alcohol units. One of these peaks A, n-butFaldehYde; n~Poly/I) Polyvinyl Alcohol 2,4-Dinitrophenylhydrazone. vinyl alcohol. is shown by --A sample of polyvinyl alcohol (DP 331) was at 5.99 suspended in alcohol acidified with hydroc~oric both the polyvinyl alcohol and its 2,4-dinitroacid and refluxed for a few minutes with an excess phenylhydrazone and may therefore come from (12) Auwers and Kreuder, iLid., 58, 1974 (1925). Of 2,4-dinitrophenylhydrazine.The polymer was (13) We are indebted t o Mr. R . M. Whitney for determining the then filtered and washed with dcohol infrared absorption spectra of our compounds and t o Protessor A . M. to remove all free 2,4-dinitrophenylhydrazine. Buswell and Dr. R. C. Gore for the interpretation of these spectra.
-
CHzCH -CH?C/
(
\H
H '
(
(
-
-(
(
(11) Bergmann and Machemer, B r r . , 63, 316 (1980).
(14) Lecomte, C o n p f . r e n d . , 180, 1481 (1925); Thompson and Harris. Trans. Faraday Soc., 88, 87 (1942).
t
I t was thought possible that acid methyl alcnhol might react with polyvinyl alcohol to cause some ether formation. However, Zeisel determinations on polymers treated in this manner did not show the presence of any methoxyl groups.
i,
Summary
0.8 ~
.*"
0.6
4 0.4 0.2
During the hydrolysis of polyvinyl acetate and oi-I I . . also during reesterification of polyvinyl alcohol, 6.36 6.05 5.75 .5 4.5 unexpected and irregular changes in the apparent P. Fig. 2.-Infrared absorption spectra : 0, polyvinyl alcohol ; molecular weight of the polymer have been noted. X possible explanation of this behavior is that 3,its dinilrophenylhydrazone. polyvinyl alcohol map have one terminal aldehyde some residual acetyl groups which have not been group which, under acid conditions, can undergo removed by saponification. The peak a t 6.05 p acetal formation with the hydroxyl groups of which is shown by the polyvinyl alcohol is greatly neighboring molecules. Under alkaline condireduced and shifted slightly t o the higher wave tions an aldol or reverse aldol reaction may occur lengths in the derivative. This may be due to to cause changes in molecular size. terminal carbonyl in the untreated alcohol. URBAN\, f r r ~ ~ x o r s RECPIVEDMAY20, 1943 I
1
[CONTRIBUTION FROM
THE
LABORATORIES OF CHEMISTRY AND PHYSICS, ILLINOIS INSTITUTE OF TECHNOLOGY ]
The Aldol Condensation. 11. The Reaction of Isobutyraldehyde with its Aldol' BY ROBERTH. SAUNDERS, M. J. MURRAY AND FORREST F. CLEVELAND Experimental work with the aldols is complicated by the fact that even such ordinary treatment as distillation may either eliminate water to produce the unsaturated compound or dealdolize the material to the simple aldehyde. To obviate part of this difiiculty the present work has been confined t o 2,2,4,4-tetramethylaldo12(isobutyraldol) which, because of the two methyl groups in the alpha position, has practically no tendency to eliminate water during distillation. Experiments of other^^-^ which have been confirmed in the present work have shown that this aldol can be distilled without excessive dealdolization at reasonable temperatures (below l.OO). 2,2,4,4-Tetramethylaldol was first obtained by Urbain6 in 1895. Since that time a number of investigations have been undertaken to determine the nature of the condensation products formed when isobutyraldehyde is treated with (1) For the first article of this series see Saunders, Murray, Cleveland and Komarewsky, THISJOURNAL, 66, 1309 (1943) (2) The aldols mentioned in this paper are systematically named as derivatives of aldol tCHKHOHCHnCH0). (3) Usherwood. J . Chcm. SOL, 1717 (1923). (4) Grignard and Iliesco, C o n p f .rend., 190, 556 (1930). ( 5 ) Kirchbaum, Sidzb. Akod. W i s . Wien, IIb, 112, 1069 (1903) ifi) IJrbain, Bull. suc ckim., 13. 1048 (1895)
alkali.4*5,7The results of these experiments are not in accord even for the condensation under mild conditions. According to the literature, the action of potassium hydroxide solutions a t or below room temperature gives two main condensation products: a dimer that is 2,2,4,4-tetramethylaldol, and a trimer which has been identified as 2,2,4-trimethy1-3-hydroxypentyl isobutyrate. The structure of the aldol seems t o have been fairly well established by means of a number of derivatives and by the ease of decomposition into isobutyraldehyde. The ester, on the other hand, has been identified by some investigators by the fact that it is converted to the salt of isobutyric acid and to 2,2,4-trirnethyl-lj3-pentanediol when refluxed with alcoholic alkali. Such treatment, of course, is standard procedure for identifying esters, but it will be shown in the following discussion that in the present case the conclusions drawn from this test were not justified. Actually this isobutyraldehyde trimer may not have been the ester a t all, and failure to recognize this has been the cause of much confusion not only in the condensation of isobutyraldehyde but also in the condensation of other aldehydes as well. 7) Hriitchtmr, Ji'7.h 4 k o d W t r ? W I ~IIb, , 105, 829 (18R6
