THEMECHANISM OF
Sept., 1956
are the stoichiometric coefficients.of the ions upon ionization, then the activity of the electrolyte as a whole is a = alrz a2ra = f’ alri
+
o 1 2 r ~= f ’ a i z
+
If one now applies (3), cancels all common terms, substitutes ( l ) , and makes use of the same technique which led to ( 5 ) , one obtains the desired equation y =
log f ’ s(s
-
log
GELATION OF GELATIN y
= s(s
(13)
where r = r1 r2 and f, the mean activity coefficient, takes into account the possible deviation from free volume ideality. If the free volume fractions are now written in terms of the molality of the hydrated solute and if the average volume ratio of the electrolyte as a whole, (qsl rZsz)/r,is denoted by s, then cy12 is given by
log
THE
2.303[1
-
l)mM1 1000
- ( h -1000 s)mM1]
-
1000 (15) . .
There are some interesting special cases. If s = 1, mole fractions result, and (15) reduces to (5). For the ideal non-electrolytes, r and f’ are both unity and (15) becomes
-
(h
-
s)mM1] 1000
If there is no solvation, (16) reduces to equation 9.21 of reference 4. Equation 16 has been fitted to the data for aqueous sucrose from 0.1 to 5m with h = 8.7, ‘z = 2.7. The fit is about as good as that of 9.214with h = 0, z = 5. If z were really proportional to the molal volume, it would have been larger than h. If the Debye-Hiickel equation is applied to the solvated ions as before, one obtains for aqueous solutions logy =
z+ z- A
1
4
O.O18s(s - 1)rm
+ a ~ & + 2.303[1 - 0.018(h - rs)m] log [ I
- log[1
c1
(16)
+
IlrmM, 1000
- log
1299
- 0.018(h - rs)m]
(17)
Since a t present there is no a priori way to determine s, equation 17 remains a three parameter equation. Acknowledgment.-The author wishes to acknowledge conversations with Professor Joseph E. Mayer and Dr. Edward V. Sayre. He is indebted to Miriam Miller, Lorraine Fischer and Stuart Rideout for very generous aid with the computations.
MECHANISM OF GELATION OF GELATIN. INFLUENCE OF CERTAIN ELECTROLYTES O N THE MELTING POINTS OF GELS OF GELATIN AND CHEMICALLY MODIFIED GELATINS BYJAKE BELLO,HELENE C. A. RIESEAND JEROME R . VINOGRAD Contribution No. 9090 from the Gates and Crellin Laboratories of Chemistry of the California Institute of Technology, Pasadena California Received March 26, 1966
The effects of a series of salts and acids on the melting points of gelatin gels have been studied. The effects of ions, of the same or opposite charge, are additive. There is a correlation between the binding of ions, as indicated by pH changes, and their effects 011 the melting point. However, by the use of amino-acetylated gelatin and a guanidino-nitrated, hydroxylsulfated gelatin, it is shown that binding of anions at amino, guanidino or hydroxyl groups is not responsible for melting point changes. B y the use of carboxyl-esterified gelatin and hydroxyl-acetylated gelatin, it is shown that binding of cations at carboxyl or hydroxyl groups is not the cause d melting point reduction. Iron (111) ion, at low concentration, raises the melting point by inter- or intramolecular cross-linking through the carboxyl groups, I n agreement with previous reports polarizable anions are effective melting point reducers with diiodosalicylate being the most potent observed.
The effects of a variety of additives on the melt- setting point while others raise it. The most efing points of gelatin gels or setting points of gelatin fective melting point reducers heretofore reported solutions have been investigated by many work- ‘are sodium salicylate12 and sodium acetyltryptoSome additives lower the melting or phan.13 It has been suggested that the effect of additives is due to “attraction”12 by the gelatin or ( I ) Presented in part before the Division of Colloid Chemistry, 127th Meeting of the American Chemical Society, Cincinnati, Ohio, to some specific interaction.’4 Alternatively, it has Rlarch-April, 1955. ( 2 ) G.S. d’blcontres, Ann. c h i m . applicala. 38,272 (1948). (3) E. H . Bucliner and B. hleylink, Rec.lrau. chim., 62,337 (1943). (4) E. H. Biichner, K o l l o i d Z., 76, 1 (1936). (5) E. H . Bilchner, ibid., 78,339 (1937). (6) J. H. C. Merckel a n d P. W . Haayman, i b i d . , 87, 59 (1939). (7) C. Marie and A. Buffat, Z . p h y s i k . Chem., 130,233 (1927). ( 8 ) 9. R. Trotinan and H. S. Ball, J . S a c . C h c m . I n d . , Tranaacliona, 63, 225 (1934). (9) G . Ciranser, H. Oliuki a n d Y . Ilirnknwn, J . / f ’ n c u / l ! i .4gr., H o k kaido I m p . (:,iiu., 23, I65 (1930).
(10) P. C. Nobel, R e c . Irau. chim., 70, 601 (1951). (11) T. R. Briggs a n d E. M. C. Hieber, THISJ O U R N A L , 24, 74 (1920). (12) G.A. Feigenand I. L. Trapani, A r c h . Biochem. B i o p h y a . , 63, 184 (1954). (13) R. S. Gordon, J r . , and J. D. Ferry, Federation P r o c . , 6 , 136 (1946). (14) J. L). Ferry, “.4dvsnces in Protein Chemistry,” Vol. IV, ( M . L. Ailaun and J. T. Edsall, d i t o r x ) , Acadeiuic l’reas, IIIC., N e n ':ark, N. Y., 1948, pp. 26-27. 1
1300
VOl. 60
J. BELLO, H. C, A. RIESEAND J. R. VINOGRAD
been proposed that the effect is due to changes in the hydro1 equilibriu1n~~-~7 or to changes in the distribution of water between the medium and "gelatin micelles."'5 In the present work, we report data on new additives bearing on the role of polarizability of additives on gelation, on the effect of acids, on the question of binding of additives and on the effects of additives on chemically modified gelatins. Experimental The gelatin used in the work was Wilson Laboratories' U-COP-CO, Special Non-Pyrogenic Gelatin, of isoelectric point 9.2 (by viscosity minimum, turbidity maximum and mixed-bed, ion-exchange resin's), prepared by acid extraction of pigskins, and having an ash content of 0.5%. Deionized gelatin was prepared by passage through a column of Amberlite MB-3, mixed-bed, ion-exchange resin.'* Amino-acetylated Gelatin.-This prepartion has been described.'@ Van Slyke amino nitrogen determination (modification of Doherty and OggZ0) _ _ showed 99.5% acetylation of amino groups. Esterified Gelatin.-This was prepared by the action of methanol and thionvl chloride. Details of meoaration and characterization will be described elsewhere. Nitrated Gelatin.-Gelatin was simultaneously nitrated and sulfated by treatment with a mixture of concentrated sulfuric acid and 100% nitric acid. Details will be described elsewhere. Additives.-Unless otherwise noted, the salts were reagent grade chemicals used without further purification. Sodium henzenesulfonate was prepared by neutralization of benzenesulfonic acid and was recrystallized twice from water. The sodium salts of the following acids were prepared by neutralization without further purification: tribromoacetic acid,*' diiododacetic acid, (prepared by Angeli's methodz2for triiodoacetic acid; we were not able to prepare triiodoacetic acid following the published directions), dibromoacetic acid (prepared by the method of Petrieff*I for tribromoacetic using less bromine), bromoacetic (Eastman Kodak Co., redistilled), trichloroacetic acid (Baker and Adamson, Reagent), trifluoroacetic acid (Minnesota Mining and Manufacturing Co., 100.1% by titration), dichloroacetic acid (Matheson, Coleman and Bell, redistilled through a 12" spir:tl-packed column) and 2-hydroxy-3,5diiodosalicylic acid.23 Thiocyanic acid was prepared from barium thiocyanate and the calculated amount of sulfuric acid followed by removal of the barium sulfate by centrifugation. Melting Points.--Ten ml. of solution was placed in a 150 X 18 mm. test-tube, which was then stoppered, warmed a t 50' for 10 minutes to erase the thermal history of the sample and then placed in a 0 f 0.01" bath for 20 hours. A Neoprene ball, d:: 1.17, and 6 mm. in diameter, was inserted in the gel below its surface and the tube was warmed at a rate of 5 f 0.5'/hr. Up to 22 tubes could be measured a t one time. The melting point was taken as that temperature at which the ball reached the bottom of the tube. The reproducibility of melting points was about f 0 . 2 ' and a temperature change of 0 . 1 4 . 3 " occurred between the time the ball was first seen to fall and t.he time it reached the bottom of the tube. Tho pII of solutions of gclatin and salts of strong acids was 5-7, and for weaker acids was 8-9. I t was found that the melting point of 5% gelatin is independent of pH in this range and careful control of pH was not necessary except to ensure complete ionization of weak acids. When strongly acid or alkaline solutions were used, the 50" heating period was omitted. Since some solutions were too dense to permit the use of
. .
(15) E. H. Biicliner, Ree. trau. chim., 46, 439 (1927). J O U R N A L50, , 1194 (1926). (16) W. D. Bancroft, T H I H (17) W. D. Bsncroft and L. F. Could, {bid., 38, 197 (1934). (18) J. W.J a n u s , A. W. Kuncliington and A. G . Ward, Resezrch, 4 , 247 (1921). (19) J. Bello and J. R. Vinogrtrd, .I. ..In&.Chem. S O C . . ~1369 ~ , (1956). ( 2 0 ) A . C . Dolierty and C. L. Oag, I u d . Etig. Chem.. A n d . Ed., 15, 751 (1943). (21) W. Petrieff. Ber., 8, 730 (1875). ( 2 2 ) .\. .4ngeli. Ber., 26, 595 (1893). (23) G . H. Woollett and W. W. Jolinson, Org.Synlheaes, 14,52(1934).
Neoprene balls, glass balls, 4 mm. in diameter, were used in such cases. I n the density range in which both glass
and Neoprene balls could be used, differences of less than 0.3" in melting point were observed between the two methods. pH Changes.-The sodium salts and stock solution of deionized gelatin were brought to pH 7.30 f 0.05. The solutions were mixed and covered. After equilibration at 37 f 0.1' for ten minutes, the pH was measured with a Beckman Model G pH Meter and glasa and calomel electrodes. Measurements were taken periodically during at least ten minutes, although constancy was usually observed in less than five minutes. After each two gelatin-salt solutions were meaaured, a solution of gelatin alone and of buffer were measured as a check on the meter and electrodes. The standard deviations of 6-8 measurements with each of six salts were 0.03-0.05. Fewer measurements were made with the other salts.
Results and Discussion I n view of the fact that we have used a gelatin of high isoelectric point where other investigators used gelatins of low isoelectric point and that our method of melting point measurement diiTers from others, it was considered desirable t o determine the effects on melting point of a number of additives that have been reported previously. Our results agreed with those of other investigators. I n Table I are shown the melting points of gelatin gels containing various salts. Some of the sodium salts (salicylate, perchlorate, trichloroacetate, benzenesulfonate, thiocyanate, nitrate, chloride) were investigated a t concentrations from 0.1-1.0 M and the melting point-concentration curves were found to be linear, except for a slight curvature below 0.25 M , in agreement with previous work.6 Similar effects can be seen in Figs. 2 and 4. Based on this TABLE I MELTING POINTS OF 5% GELATINGELS CONTAINING SALTS Additive
None Sodium fluoride Sodium methaneaulfonate Sodium chloride Sodium bromide Sodium nitrate Sodium thiocyanate Sodium iodide Sodium benzenesulfonate Sodium salicylate Sodium trimethylacetate Sodium chloroacetate Sodium dichloroacetate Sodium trichloroacctate Sodium dibromoacetate Sodium tribromoacetate Sodium diiodoacetatc Sodium trifluoroacetate Sodium acetyltryptophan Sodium maleate Sodium succinate Sodium fumarate Sodium acetylenedicarboxylate Lithium chloride Lithium iodide Lithium salicylate Lithium diiodosalicylate Calcium chloride Magnesium chloride
Conco. moles liter-'
M.p., OC.
..
30.4
1.0 1 .o 1.o 1 .o 1 .o 1.0 1.0 1.0 1.0 1 .o 1 .o
34.5
1.0 1
.o
1.0 1.0 1.o 1.0 0.75 0.5 0.25 0.5 0.5 1.0 0.25 0.25 0.23 1.0 1.o
31.5 28.0 22.8 22.3 16.0 16.0 15.5 No gel 26.5 26.6 19.6 12.7 16.3 NO gel 12.3 24.1 13.7 33.7 32.3 31.5 19.4 25.4 26.7 20.5 N o gel 16.1 22.9
Sept., 1956
THEMECHANISM OF
linearity, some of the results, where noted, are extrapolated to concentration levels in excess of the solubility limits. I n Fig. 1is shown the effect of increasing aliphatic chain length on the effect of sodium salts of monoand dicarboxylic acids on the melting point of gelatin. As the aliphatic chain is increased, there is first an increase of melting point followed by a decrease as the non-polar chain increases toward the detergent range. The effectiveness of melting point reduction by additives with large non-polar groups has been pointed out by Gordon and Ferry.13 Ferry14 has also pointed out that large, polarizable groups make additives effective. Thus, the phenylsubstituted acids, the halogenated acids and the unsaturated acids are more effective melting point reducers than the simple, saturated aliphatic acids. For example disodium succinate raises the melting point 8" per mole while disodium fumarate raises it only 2.3". Both ions have trans configurations. Disodium maleate, which has a cis configuration, raises the melting point 6" per mole and disodium acetvlenedicarboxvlate. which is linear, lowers the helting pointlby 22' per mole. These results show that both polarizability and structure are important. Among the halogenated acids, the higher the atomic number of the halogen *, and the greater the number of halogen atoms, the greater is the melting point reduction. There is no correlation between acid strength and melting point reduction as the 24 acid strength decreases with increasing atomic number and in- 0' creases with increasing number of * 2 0 halogen atoms.24 The effect is not cL due to size alone, as trimethylacet- 2 ate, which is about the same size as wz 16 tribromoacetate (from Fisher5 Hirschfelder-Taylor models), lowers WI 12the melting point only one-eighth as much as the latter. The polariza8 bility of the additives is in good correlation with their effectiveness as melting point reducers. The most effective melting point reducer 4 known is lithium 2-hydroxy-3,5-diiodobenzoate (diiodosalicvlate) n " which lowers the melting by" 1206 per mole; that almost all of the effeet is probably due to the diiodoFig. 2.salicylate ion will be shown below. The effects of some acids are shown in Fig. 2. The large slope of the hydriodic acid curve may be related to the observed partial precipitation of the gelatin which occurs between 0.25 and 0.5 M concentration. That the reduction of melting point was not due to extensive degradation was shown by careful neutralization of the hydrochloric acid solutions, followed by determination of the melting point of the neutralized solutions. The data in Fig. 3 show that the melting point reduction is nearly reversible, even a t 3 M .
THE
GELATION OF GELATIN
1301
40t+c
o
2 4 6 8 10 NO. OF ALIPHATIC CARBON ATOMS.
Fig. 1.-Melting point of 5% gelatin containing sodium salts of carboxylic acids; effect of aliphatic chain length: 0 , monocarboxylic acids; 0,dicarboxylic acids; 0 , phenylsubstituted monocarboxylic acids. Abscissa is number of aliphatic carbon atoms exclusive of carboxyl groups.
A METHAWESULFONIC AGIO 0 SULFURIC ACID
0 HYDROCHLORIC ACID 0 HYOROBROHIC ACID
A BENZENESULFONIC ACID 0 HYDRIODIC ACID
1
(24) H. C. Brown, D. H. McDaniel and 0. Hatliger, in "Determinat i o n of Organic Structures by Physical Methods" (E. A. Braude and F. C. Nachod, editors) Academic Press, Inc.. New York, N. Y..1955,
p. 579.
I
3
2 MOLARITY
-
l
4
OF ACID.
-Effects of acids on the melting point of 5% gelatin.
I n Fig. 4 are shown the effects of urea and some related compounds. Guanidinium chloride is about as effective as hydrochloric acid and both are about twice as effective as urea. Another similarity between hydrogen and guanidinium ions is that guanidinium thiocyanate and thiocyanic acid (but not sodium thiocyanate) precipitate gelatin at about 0.5 M , but redissolve the precipitate a t 1 M concentration. The curve for guanidinium sulfate indicates that the melting point raising effect of sulfate ion becomes greater, per mole, as the coneent,ration increases. The additives fall into the well-known Hofmeister or lyotropic series and into a series determined
1302
J. BELLO,H. C. A. RIESEAND J. R. VINOGRAD
Vol. 60
dium chloride lowered the melting point by about 2", each of its ions was arbitrarily assigned a value of -1 and from this value and the observed melting point depressions of other salts there was calculated a molar melting point depression, a,for each ion. These have not been tabulated but can be calculated readily from Table I. The additivity of a was tested by comparing the measured melting points of mixtures of salts and of acids with results calculated from the a values. The results are shown in'Table 11. There is good agreement for all the additives tested except thiocyanic acid and, to a lesser extent, guanidinium thiocyanate. Thiocyanic acid is unstable and turned yellow. The additivity exhibited by the lithium salts makes it probable that the major effect of lithium diiodosalicylate is due to the diiodosalicylate ion. The Effect of Time on the Melting Point Reduc0 I 2 3 4 tion.-To determine whether the effect of additives is a rate or equilibrium effect, the melting [HCI], MOLES/ LITER. points of gelatin containing various salts was measured after storage a t 0" for periods up to 31 days. Fig. 3.-Reversibility of melting point reduction by hydrochloric acid: 0, hydrochloric acid; . , after neutralization of The results, Fig. 5 , show that after the first 2-3 0 ; a, original gelatin of the same concentration of gelatin days there is little additional rise in melting point. and sodium chloride as 0 . When 1 M sodium salicvlate was used (not shown on the figure) there was n o gelation in 31 days. It seems likely from the shapes of the curves that the melting point reductions represent changes in the equilibrium of the gelatingelatin or gelatin-solvent interaction. Effects of Neutral Salts on pH of Gelatin: Binding of Ions.-The possibility that binding of additives may cause the melting point changes was considered. This was first investigated by observing the effects 0 FORMAMIDE of neutral salts on the pH.27 There have been contradictory reports on the binding of small ions by gelatin. For example, Docking and Heyman28 GUANlDlNlUM ION) reported that gelatin binds thiocyanate and chloride, among others, while Williams, et found no I 2 3 4 binding of thiocyanate and Carr30 GONG. OF ADDITIVE, MOLES/LITER. and Carr and T o p ~found l ~ ~no bindFig. 4.-Effect of urea and related compounds on the melting point of 5% ing of &loride. gelatin gels. There have been reports on the by the effects on the denaturation of proteins, see, effects of salts on the pH of gelatin solutions. That i ~ ~ that some calcium, for example, Simpson and K a u ~ m a n n . ~ ~ is, Giedroyc and P r ~ y l e c k found additives that lower the melting point of gelatin magnesium and sodium salts lower the pH, but the promote the denaturation of proteins and those concentrations a t which they worked were too low that raise the melting point of gelatin protect (mostly below 0.05 M ) to be of significance i n reagainst denaturation. This is particularly notable gard to melting point effects. Scala33found that in the case of sulfate; for while guanidinium chlo- sodium and calcium chloride cause an increase in ride is a good melting point reducer and denatu- pH and that barium chloride produced varying rant, guanidinium sulfate has little effect on melt- effects at different concentrations. We have examined the effects of various salts a t 1 ing point or denaturation.26 Molar Melting Point Depression.-The fact M concentration on the pH of 5% gelatin solutions. that the effects Of various acids On melting are (27) G. Scatcliard and E. Black, ibid., 63, 88 (1949). parallel to, but greather than, the effects of the (28) A. R. Docking and E. Heyman, i b i d . , 43, 513 (1939). Williams, W. M. Saunders and J. S. Cioirelk i b i d . , 58, corresponding sodium salts, suggested that the 7,y;:9;i)V. effects of mixtures might be additive. Since SO(30) C. W. Carr, ~ ~Biochen. ~ hB ~ O .P ~ V S4. 6, , 4 1 7 (1963).
P i u
(25) R. B. Simpson and W. Kaurmann, J . Ana. Chem. Soc., 76, 5139 (1953). (26) N. F. Burke, T H IJOURNAL, ~ 47, 164 (1943).
(31) C. W. Carr and L. Topol, THISJOURNAL, 64, 176 (1950). (32) W. Giedroyc and S. J. Przylecki, Biochena. J . , 26, 465 (1931). (33) A . Scala, Ann. igiena, 44, 313 (1932).
THEMECHANISM OF THE GELATION OF GELATIN
Sept., 1956 TABLE I1 ADDITIVITY OF
LY
I
1303
I
I
I
1
5
IO
15
20
.
1
I
25
30
VALUES M.p. depression, OC. Calcd. Found from a
Additive
0 . 5 M NaCNS and 0 . 5 M NaI 1 5 . 0 15.2 0.5 M NaCNS and 0 . 5 M NaBr 10.9 11.7 0.5 M NaCNS and 0.5 M NaCl 9.1 9.2 0 . 5 M Guanidinium chloride and 0.5 M HCl 1 0 . 3 10.2 Benzenesulfonic acid, 1 M 1 9 . 0 20.5" Hydriodic acid, 1 M 20.8 20.0" Hydrobromic acid, 1 M 14.0 13.6" Methanesulfonic acid, 1 M 5.5 4.5a Thiocyanic acid, 1 M 3 3 0 . 2 20.0 Guanidinium thiocyanate, 1 M 26.4 23.6 15.0 15.2 LiI, 0.25 LiCI, 0.25 M and sodium salicylate 0.25 M 2 0 . 5 21.4 5 Calculated from the values for hydrogen ion found for hydrochloric acid.
From the data of Fig. 6 it can be seen that negative ions raise and positive ions lower the pH, the behavior expected if binding occurs. Also, there is a correlation between p I i change and melting point change for both melting point reducers and melting point raisers, although only two of the latter were investigated. The pH changes for the amino-acetylated gelatin are smaller than those for the original gelatin as is to be expected if the binding of anions occurs a t the positively charged groupsS4; the acetylated gelatin contains fewer such groups. Somewhat unexpected was the finding that sodium chloride and, to a lesser extent, bromide reduce the pH of the acetylated gelatin while raising that of the non-acetylated gelatin. This suggests that sodium binding occurs. I n the case of nonacetylated gelatin, the effect of the sodium binding is swamped by the binding of the anions. The acetylated gelatin has a net negative charge a t pH 7 , which would promote cation binding while the non-acetylated gelatin has a net positive charge which would have the opposite effect. Effects of Additives on Chemically Modified Gelatins.-It was considered possible that the correlation between pH change and melting point change was fortuitous and that the bound ions do not affect the melting point. Accordingly, we investigated the effects of a variety of salts on gelatin having one or another of its possible binding sites blocked. The effects of some additives on the melting point of the amino-acetylated gelatin are shown in Table 111.
51
-
T I M E IN DAYS. Fig. 5.-Effect of time on melting points of gels containing 1 M salts: 1, no salt; 2, sodium chloride; 3, sodium bromide; 4, guanidinium chloride; 5 , sodium thiocyanate; 6, 0.7 M sodium salicylate.
35
I
301
I
I
I
' P N 6 F
I
I
N&:3s03Na
'-0.6 -0.4 -0.2
1
CH2CICOONa
0
0.2
0.4 0.6 0.8
A pH. Fig. 6.-Correlation between pH and melting point changes for 1 M salts and 5% gelatin. CNIH&l ia guanidinium chloride; 8, data for amino-acetylated gelatin.
TABLE I11 Although sodium bromide has a slightly smaller EFFECTO F ADDITIVES O N THE MELTING POINTOF 574 GELS effect on the melting point of the acetylated gelatin than on that of the original gelatin, the effects of OF AMINO-ACETYLATED GELATIN Concn. Additive
None Sodium bromide Sodium thiocyanate Sodium iodide Sodium salicylate Lithium diiodosalicylate
moles/l.)-1
.. 1.0 10 1.0 0.5 0.06
M p.,
=c.
16 1 10.6 No gel No gel No gel 8.8
M P. reduction, OC.
.. 5.5 916.1 916.1 y16.1 7.3
(34) I. M. Klotz, Cold S p r i n g Harbor Symposia on Quanlilaliue Biology, 14, 97 (1950).
sodium thiocyanate and iodide and of lithium diiodosalicylate on the acetylated gelatin are equal to or greater than on the original gelat,in. Lithium diiodosalicylate reduced the melting points of the original and acetylated gelatins by 7.0 and 7.3", respectively. The data for this salt are especially significant as this salt is the most effective of the melting point reducers and would be expected to be the most strongly bound as, in general, the larger the ion the greater the binding.34,35 (35) J. D. Teresi and J. M. Luck, J . Biol. Chem., 199, 823 (1952).
J. BELLO, H. C. A. RIESEAND J. R. VINOGRAD
1304
The role of binding a t amino groups was also investigated by observing the effect on melting point of two additives at pH values high enough to convert most of the charged ammonium groups to uncharged (and, therefore, non-binding) amina groups. At pH 11.7, most of the amino groups are in the uncharged form since the pK values of the e-amino groups of the lysine residues are about 10.5 (see, for example, Tanford and Wagner on lysozymea6), a value in agreement with the titration curve we have obtained on gelatin and that of Ames." Also titration in the presence of l M sodium thiocyanate did not change the titration curve significantly. This suggests that little or no binding of thiocyanate occurs at high pK. The results, Table IV, show that iodide and benzenesulfonate are as effective a t pH 11.7 as a t pH 6. From the data of Tables I11 and I V it is apparent that binding at charged ammonium groups, whether or not i t occurs, is not the mechanism by which anions lower the melting point. TABLE IV EFFECTSOF SOMESALTSON THE MELTINGPOINTS OF 57$ GELATINAT VARIOUSpH VALUES Additive
Concn.,
M
BH
M.p., 'C.
M.p. reductions'
None .. 11.7 2 9 . 6 .. Potassium iodide 0.5 11.7 22.4 7.2 Sodium benzenesulfonate 0 . 5 11.7 22.5 7.1 None .. 6 30.2 .. Potassium iodide 0.5 6 23.0 7.2 Sodium benzenesulfonate 0 . 5 6 23.2 7.0 Calcium chloride 1 6 16.1 14.1 Guanidium chloride 1 6 22.0 8.2 None .. 2 27.6 .. Calcium chloride 1 2 16.1 11.5 Guanidium chloride 1 2 18.8 8.8 a Difference between the melting points with and without salt at the same pH.
The possibility of binding of anions a t guanidinium groups being responsible for melting point reduction was investigated by observing the effect of lithium diiodosalicylate and sodium thiocyanate on a gelatin having 71% of its guanidinium groups nitrated to uncharged nitroguanidino groups. Simultaneously with the nitration, the hydroxyl groups were sulfatedas by the large excess of sulfuric acid present, resulting in the introduction of 60 sulfate groups per lo5 g.39 Complete sulfation, in the absence of nitric acid, introduces 150 sulfate groups per lo5g. It was found that lithium diiodosalicylate and sodium thiocyanate were as effective in lowering the melting point of the nitrated gelatin as that of the original gelatin (see Table V). Although the effects of other melting point reducing anions on the nitrated-sulfated gelatin were not investigated, the results obtained show that binding of anions a t guarlidinium groups or a t hydroxyl groups is prob(36) C. Tanford and M . L. Wagner, J . A m . Chem. SOC.,76, 3331 (1954). (37) W.h l . Amea, J . Sci. Food Agn'c., 9, 579 (1952). (38) H. C. Reitz, R. E. Ferrel. H. S. Olcott and H. Fraenkel-Conrat, J . Am. Chsm. Soc., 68, 1024 (1946). (39) Sulfur analyses b y Elek Microanalytical Laboratories, Loa Angeles, California.
Vol. 60
TABLE V EFFECTS OF SODIUM SALTSON THE MELTINGPOINTSOF GELATINA N D MODIFIEDGELATINS' Concn. Type of gelatin
Salt
Original Thiocyanate Nitrated Thiocyanate Original Lithium diiodosalicylate Nitrated Lithium diiodosalicylate Original Acetate Acetylated Acetate Nitrated Acetate Original Succinate* Acetylated Succinateb Nitrated Succinate* Original Dibasic phosphate* Acetylated Dibasic phosphate6 Nitrated Dibasic phosphate* Original Fluoride Acetylated Fluoride Nitrated Fluoride a At pH 7 unless otherwise noted. Ir At
of salt, M
0.5 0.5 0.06 0.06 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 pH 8.9.
M.p. change, 'C.
-7.1 -7.2 -7 . 0 -7 . 0 +1.0
$0.9 +0.9 +l.9 t2.0 +1.6
$2.8 f2.6 +2.8 +1.1 f1.3 f1.4
ably not important in melting point reduction, particularly as diiodosalicylate would be expected to be the most strongly bound of the anions investigated. Similarly, in the case of anions that raise the melting point, the possibility of the effect being due to binding of the anions a t cationic sites was investigated with the aid of the acetylated and nitratedsulfated gelatins. The results are shown in Table V for anions of four different types: inorganic univalent and divalent and organic univalent and divalent. It is apparent that with both the acetylated and nitrated gelatins, the additives are as effective as with the original gelatin, showing that anions that raise t,he melting point do not do so as a result of being bound a t the amino or guanidino groups. Similarly, the effect of possible binding of cations a t the negatively charged carboxylate groups was investigated by determining the melting point reduction by guanidinium chloride and calcium chloride a t pH 6 and 2. Binding of guaiiidinium ion by a low isoelectric gelatin above pH 3 has been reported.40 Titration of gelatin with hydrochloric acid, in the presence of 1 N calcium chloride, showed a decrease of 0.6 unit in the pK of the carboxyl groups, indicating that calcium binding orcurs. At pH 2 the carboxyl groups are largely in the uncharged state. From Table IV it can be seen that guanidinium chloride is a t least as effective a t pH 2 as a t pH 6, while calcium chloride is SOYo as effective a t p H 2 as a t p H G . A similar experiment has been reported by Merckel4l for sodium salts of several acids. However, as sodium ion appears to have little effect on melting point, the data presented here for calcium and guanidinium chlorides are more significant. I n addition, the effect of calcium chloride on a gelatin having 7oy0of its carboxyl groups esterified was determined a t pH 1.6 where there are almost no free carboxylate ions. The melting point was reduced by 11.5", compared (40) M. D.Crynberg, Biochsm. Z., 262, 272 (1933). (41) J. H.C. Merckel, Kolloid Z.. 1 8 , 339 (1937).
Sept., 1956
THEMECHANISM OF THE GELATION OF GELATIN
with 11.5' for normal gelatin a t pH 2, and 14.1" for normal gelatin at p H 6. From the foregoing it appears highly unlikely that binding of cations a t negative carboxylate groups is very important in melting point reduction, although this may be of some significance in the case of calcium ion. The possibility that binding of cations to hydroxyl groups may be important in reducing the melting point was tested with calcium chloride and a gelatin having its hydroxyl groups quantitatively and selectively acetylated. The melting point of the hydroxyl-acetylated gelatin was 15.1O ; in the presence of 1 M calcium chloride there was no gelation. Since the melting point reduction produced by calcium chloride is 14" in the case of normal gelatin, it is evident that binding a t hydroxyl groups is not important in gelation. One cationic melting point raiser has been investigated, namely, iron(III), as the chloride. The effect of iron(II1) chloride on the melting point of gelatin gels is a function, a t constant gelatin concentration, of the concentration of the iron. Thus. addition of 0.25 M iron(II1) chloride to a 10% gelatin solution (to obtain final concentrations of 0.13 M and 5%, respectively) causes immediate gelation and after the usual storage a t 0" for 20 hours, the melting point of the gel is raised by 4', or 32' per mole of iron chloride. If, however, 1 M iron(II1) chloride be used, no gelation occurs till cooling, and, after 20 hours a t 0", the melting point is reduced by 8', compared with the melting point of gelatin alone a t the same pH. When a 5y0 solution of the carboxyl-esterified gelatin is made 0.13 M in iron(II1) chloride, gelation does not occur till the solution is chilled, and the melting point, after standard treatment, is not raised, but reduced by l o , or 8" per mole, corresponding to the reduction produced by 1 M iron on the original gelatin. The melting point of the esterified gelatin alone is the same as that of the original gelatin, indicating that little degradation occurred in the esterification process. The above observations can be explained by assuming coordination of iron at carboxyl groups. At low concentration of iron, cross-links are formed by coordination of iron with carboxyl groups of two gelatin molecules. Alternatively, intramolecular cross-links may stabilize a structure required for gelation, or both mechanisms may operate. A t high concentration of iron, polycoordination of iron is less extensive and the salt acts as a melting point reducer. Mechanism of Action of Electrolytes on Gelation. -The foregoing data have shown that, with the exception of iron(III), the effects of ions on the melting points of gelatin gels, are not due to binding a t the charged groups or a t hydroxyl groups. These conclusions apply only to the small ions we have studied. In the case of polyelectrolytes and micelle forming ions more complex interactions may occur.42.43 Other remaining explanations for the effects of ionic additives are: (1) an effect on the solvent and (2) interactions with parts of the gelatin molecule other than those investigated here. With regard (42) K. G. A. Pankhurst and R. C . M. Smith, Trans. Faraday Soc.. 40, 565 (1944). (43) M. Joly, Bull. 8oc. chim. biol., 80, 398 (1948).
1305
to the former, the effects of salts on the dielectric constant appear to be an inadequate explanation for the widely varying effects on melting point by the different salts. This has also been found wanting to explain the effects of some salts on the denaturation of ovalbumin.26 The available data do not permit a decision between other possible effects on the solvent, such as the structure-breaking effect44or changes in the degree of the polymerization of water.'? The last of these was proposed by Bancroft and Could who made arbitrary assumptions as to the effects of various ions. The large differences in the effects of various ions on the melting point, both as to direction and magnitude, suggest that the explanation lies in interactions between the ions and the gelatin. (It must be borne in mind, however, that the large changes in melting point may be the result of rather small changes in the configuration of the gelatin molecules.) The only parts of the gelatin molecules not considered above are the non-polar side chains and the peptide groups. Interactions between the non-polar groups and ions such as iodide or calcium are unlikely. Interactions between the non-polar groups and organic ions are possible. I n such interactions the relatively non-polar portion of the ion would be oriented toward the gelatin and the charged group toward the water. The peptide bond remains the most likely site for the action of small ions. These materials then lower the melting point by breaking peptide hydrogen bonds involved in intermolecular cross linking, and, possibly, similar bonds involved in maintaining an intramolecular configuration necessary for gelatin. The formation of segments of gelatin molecules of ordered configuration in the gelation process has been suggested by several a ~ t h o r s . ~ ~The **~ effectiveness of large ions as melting point reducers may be due to theirability tocover the peptide group. That the effectiveness of hydrogen ion may be due to interaction a t the peptide bond is suggested by the preparation of 1:1 complexes of amides and hydrohalic a ~ i d s . ~ ?The - ~ complexing ~ of anionsat peptide bonds is suggested by the finding of Steinhardt and Fugitt60 that the rate of acid hydrolysis of the amide and peptide bonds of wool depends on the nature of the anion of the acid, the large polarizable anions having the greatest catalytic effect. The anions that raise the melting point may do so by protecting the ordered segments of the gelatin from denaturation by water or by cross-linking through interaction at the peptide bonds. The former explanation is supported by the fact that melting point raisers increase the specific rotation of gelatinl5' as does cooling of a gelatin solution to gelling temperature. Also, melting point raisers protect ovalbumin against urea d e n a t u r a t i ~ n . ~ ~ If the effects of the additives are due to changing the solvent, then the observed correlation between (44) H. 5.Frank and M. W. Evans, J . Chem. Phys., 18,507 (1945). (45) C. Robinson and M. V. Bott, Nature, 168, 325 (1951). S8, 968 (1954). (46) H. Boedtker and P. Doty. TEISJOURNAL, (47) E. H. White, J . Am. Chem. Soc., 77, 6215 (1955). (48) P. Walden, Chem. Zantr., 83, [I], 122 (1912). (49) A. Werner, Ber., 86, 154 (1903). (50) J. Sbinhardt and C. H. Fugitt, J . Researrh Null. Bur. S h n d urds. 29, 315 (1942). (51) E. S t h n y , Rolloid Z.,86, 353 (1924).
1306
LEOA. WALL,MARYR. HARVEY AND MAXTRYON
melting point and pH change is fortuitous. If the effects on the melting point are due to interactions with the gelatin, such interactions would contribUte to the pH change, the major part of which is due to binding a t the charged groups. It is reasonable to assume that those ions that are most strongly bound a t charged groups would also inter-
Vol. 60
act most strongly a t the polar peptide groups. A precise description of possible complexes between ions and peptide bonds is not possible a t present. Acknowledgment.-This work was supported under Contract No. DA 49-007-MD-298 with the Office of the Surgeon General, Department of the Army, to whom we wish i o express our thanks.
OXIDATIVE DEGRADATION OF STYRENE AND a-DEUTEROSTYRENE POLYMERS BY LEOA. WALL,MARYR. HARVEYAND MAXTRYON National Bureau of Standards, Washington, D. C Received March 2S71968
The oxidation of polystyrene and a-deuterostyrene polymers in the presence of ultraviolet radiation and air a t 60' has been that of polyinvestigated. Initially, the deuterated polymer shows an increase in absorption a t 340 mp of only about styrene. A post-irradiation effect was observed which disappears upon further irradiation. The reaction occurring during post-irradiation of polystyrene consists of two first-order components with activation energies of 16 and 20-24 kcal. per mole. This may indicate the decomposition of two different hydroperoxide structures during the post-irradiation periods. However, it is suggested that one of these components may result from a cis-trans isomerization. I t is concluded tentatively that polystyrene oxidizes mainly a t the a-position of the monomer unit but that the resulting hydroperoxide, although formed, is extremely labile.
Introduction I n earlier investigations a t this Laboratory on the ultraviolet-induced oxidation of polystyrene, the carbonyl and hydroxyl group build-up was followed by infrared spectra2a and the small amountls of volatile components produced were analyzed by mass spectrometry.2b Recently the increase in ultraviolet absorption has been found to be a sensitive indication of d e g r a d a t i ~ n . ~Thus ultraviolet spectrometry, which is usually a more convenient and simpler technique t o employ than infrared or mass spectrometry, has been used in this work. Often the easily accessible regions in the infrared merely show the presence of certain groups without giving definite further information concerning the specific nature of the structure containing these groups. Mass spectrometry, on the other hand, ordinarily gives analyses of volatile fragments which are, a t least a t the low extents of reaction so far studied,2 invariably contaminated with trace impurities such as solvent and solvent oxidation products. The rate of oxygen consumption would be a more direct quantity to study from a mechanistic viewpoint. We have, however, confined ourselves in this study to the changes in ultraviolet transmission in the region of 280 to 400 mp, because previous work has shown that decreased transmission occurred in these regions during the early stages of oxidation.3 In the present work the oxidation behavior of adeuterostyrene polymer was compared with that of polystyrene, the purpose being to establish the site of radical attack, It is well-known that deuterium atoms are abstracted a t much lower rates (1) Presented a t tlie 126th Meeting of the American Chemical Society in New York, N. y . , September 12-17, 1954. (2) (a) B. G. Achhammer, M. J. Reiney and F. W. Reinhart, J . Research Nall. Bur. Slandards, 4 1 , 116 (1951); (b) B. G. Achliammer. M. J. Reiney, L. A. Wall and F. W. Reinhart, J . Polymer S c i . , 8, 555 (1952) : also Natl. Bur. Standards Circular 525, Polymer Degradation Mechanisms, p. 205 (1953). (3) M. J. Reiney, M . Tryon and B. G . Achhammer, J . Research Nall. Bur. Standards, 51, 155 (1953).
than protium atoms, and hence such a comparison should enable one to gain basic knowledge of the mechanism of oxidation. If the presence of deuterium atoms alters any of the rate-determining elementary steps of the oxidation process, then the over-all rate will be altered. An analysis of the results in conjunction with other kinetic data will aid in establishing the mechanism in more detail than has hitherto been possible. Deuterium studies have been used previously to ascertain the site of transfer processes in polymerization reactions4r5and in pyrolytic decomposition of polymers.6 Materials and Methods The polystyrene used i n this work was a sample prepared by thermal bulk polymerization a t 120" and had an approximate number average molecular weight, as determined from osmotic pressure measurements, of 237,000. This was the same highly purified polystyrene sample used in earlier work.s The poly-a-deuterostyrene sample wa8 prepared by the bulk polymerization a t 70' of a-deuterostyrene. The monomer was synthesized from acetophenone by reduction with lithium aluminum deuteride to give a-deuteromethylphenylcarbinol, which was dehydrated. Mass spectra gave the following analysis: styrene 1.90%, styrene-dl 97.4270, styrene-dz 0.54%, styrene-& 0.14%. Both polymer samples were purified of monomer, dimer and similar materials by repeated solution in benzene, followed by precipitation in methanol. The final product was dissolved in benzene, the solution frozen, and the solvent, removed by sublimation a t reduced pressure. The extent of puiification was determined by the ultraviolet ahsorption of chloroform solutions of these polymers after each cycle of solution and precipitation as recently descrihed . 3 Films of these purified polymers were cast from benzene solutions by the method described in references 2a and 3 . The film thicknesses used were approximately 0.18 mm. The films were exposed to ultraviolet radiant energy from a sunlamp in air on a rotating turntable 15, cm. from the lamp. The temperature of the table was 60 . This equipment is described in method No. 6021 of Federal Specifica(4) L. A. Wall and D. W.Brown, J . Polymer S c i . , 14, 513 (1954). (5) P. D. Bartlett and F. A. Tate, J . A n . Chem. Soc., 1 5 , 91 (1953). (6) L. A. Wall, D . W . Brown and V. E. Hart, J . Polymer Sci., 16, 157 (1955).
