MECHANISM OF SOLVENT ACTION Influence of Molecular Size and Shape on Temperature Dependence of Solvent Ability ARTHUR K. DOOLITTLE Carbide and Carbon Chemicals Corporation, South Charleston, )v. ?‘a.
T h e size and.shape of the solvent molecule influence the temperature coefficient of solvent ability, as judged by measuring the threshold concentration for solvation a t different temperatures. In the range f50” to -10’ C. small linear molecules exhibit normal behavior and become better solvents a s the temperature rises, but large linear molecules of the same homologous series behave in the opposite manner and are better solvents cold than
hot. For solvent molecules of the same molecular weight, the shape becomes the determining factor; compact molecules show the normal temperature dependence of solvent ability, and extended molecules show the inverted temperature behavior. Certain solvents exhibiting this inverted temperature dependence of solvent ability are used to prepare gels of macromolecular sobstances that liquefy on cooling.
T
sider the dependence of threshold concentration on temperature; several factors that influence this dependence will be discussed.
HE previous paper of this series (3) established a criterion of solvent ability based on the tolerance of a resin solution for a diluent. This criterion, called the “threshold concentration”, is the number of moles of solvent present per liter at the dilution ratio end point at 0.5 gram of solids per 100 cc. total volume. The dependence of threshold concentration on the molecular weight of the solvent was studied for several homologous series of solvents a t 20’ C.; it was observed that, as the molecular weight of the solvent increased, each series approached a limiting value called the “class threshold concentration”. This paper will con-
TEMPERATURE COEFFICIENT OF SOLVENT ABILITY
If solvent ability improves as the temperature is raised, the solvent has a positive temperature coefficient of solvent ability. If, however, solvent ability decreases with a rise in temperature, the solvent bas a negative coefficient of solvent ability or an inverted solubility-temperature relation. The temperature coefficient of solvent ability may vary with the level of the temperature range under consideration. Temperatures from -10’ t o 50” C. are studied here because we live in this range for the most part, and hence i t is significant with reference to the use of solvents in surface coatings and plastics. Since solvent ability ie not readily defined and we have no units for expressing its magnitude, we prefer to think of this concept in terms of the threshold concentration, ~ D R which , can be readily measured and expressed in moles of solvent per liter at the dilution ratio end point. Thus a decrease in threshold concentration represents an increase in solvent ability, and for practical pqrpoees we may therefore define the temperature coefficient of solvent ability as the negative fractional rate of change of threshold concentration per degree. Furthermore, since this coefficient varies with the level of tem-
I W
n-ACETIC ESTERS ~
-x-x
a
0
IO 0
2.0
L. \
= = 5ooc.
= 20°G.
I
6 4 0
2
40
= 50%. 8
fi
80
120 M
160
0.8 0.6
=
2OOC. -10%.
I
I
J
300
400
500
Figure 2. Nitrocellulose Threshold Concentrations vs. Molecular Weight for Di-n-Alkyl Phthalates
200
1. Dimethyl 2. Dieth 1 3. D i - d u t y 1
Figure 1. Nitrocellulose Threshold Concentrations us. Molecular Weight for n-Acetic Esters and n-2-Ketones
535
Di--hexyl 5. Di-moatyl 6. M-n-dodeoyl 4.
I N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY
536
0.6
I
I
I
200
300 M
400
Figure 3. Nitrocellulose Threshold Concentrations us. RIolecular Weight for Di-n-Alkyl Succinates with Toluene as Diluent 4. 5. 6.
1. Dimethyl
2. 3.
Diethyl Di-n-butyl
Di-n-hexyl Di-n-octyl Di-n-dodecyl
* = 50OC. = 2OOC. 0
= -IO°C,
1
c x--
~~~
A & &
0.0
2AO
3b0
M 340
J
4b0
Figure 4. Nitrocellulose Threshold Concentrations us. Afolecular Weight for Di-n-Allioxyethyl Succinates with Toluene as Diluent 1. 2.
Di(methy1 Cellosolve) 3. Di(n-butyl Cellosolve) Difethyl Cellosolve) 4. Di(n-heyy1 Cellosolve) 5. Di(n-octyl Cellosolve)
perature range, Te further define temperature coefficient of solvent ability for this paper arbitrarily as the negative of the mean fractional change in threshold concentration per a C. that is observed over the range - 10' to 50 O C. Specifically,
.
I
I
This Coefficient will be positive or negative, depending on whether solvent ability increases or decreases as the temperature rises within this range. DEPENDEhCE OF SOLVEhT ABILITY ON TEMPERATURE
INFLUENCE OF MOLECULAR WEIGHT. The dependence of threshold concentration on temperature in homologous series of solvents is influenced by the molecular weight of the solvent. Members of low molecular weight generally have positive temperature coefficients in the range - 10' to 50' C., where'as members of high molecular rreight may show an inversion of solvent ability. The point where the solvent is insensitive to temperature in this range varies with the diluent used. These observations are apparent from Figures 1 to 8, based on Tables I and 11. The relations shown by these data do not necessarily hold at temperatures above or below the range covered in this study. For example, in the case of acetone and Vinylite resin VYNW, as the temperature is raised above 50" C. (in a closed system), the solvent ability continues to improve until a level of about 7090' C. is reached. As the temperature is raised still higher, sol-
Vol. 38. No. 5
I
300 M
I
500
Figure 5, Nitrocellulose Threshold Concentrations u s . Molecular Weight for Tri-n-AllqI Phosphates with Toluene as Diluent 1. Tri-n-butyl.
2. Tri-n-ortyl.
3. Tri-n-clodec>I,
vent ability falls ofY unt'il at 134" C. pure acetone ceases to be a solvent for this resin. Some instances have been reported in the literature n here t,he solvent ability of small solvent molecules was thought to improve as t.he temperature was reduced. Most of these cases involve an alcohol or a mixture of an alcohol with some other active solvent. Byron (2) and the writer believe that, where alcohols are involved, association through hydrogen bonding affords stable enough aggregates so that the unit taking part in the solvation process is of considerably higher molecular weight than that of the simple alcohol molecule. One instance (1)TT"S noted, however, where methyl acetate was said to improve in solvent ability as the temperature was reduced to -50" C. On the assumption that this result might be accounted for by the presence of methanol as impurit,y in the methyl acetate used, since no analysis of this solvent was given, a sample of methyl acetate v m carefully purified to free it of alcohol. I t was determined that this pure ester was a poorer solvent at -50" C. than at higher temperatures. This experiment was carried out by sealing a tube containing a concentration of 0.5 gram PXP half-second nitrocellulose per 100 cc. of a mixture of methyl acetate and toluene, adjusted to 4.09 moles of solvent per liter. The tube was clear at SO", 20°, and -10" C., but the nitrocellulose came down as a gel precipitate at -50" C. The precipitate redissolved when the tube w~bsagain warmed t o room temperature. A control containing no dissolved solids was unaffected at all temperatures. IKFLCENCE OF STRCCTURE.The temperature dependence of solvent ability in the range - 10" to 50" C. is influenced far more by the shape of the solvent molecuie than by its weight. The following five isomeric esters as solvents for nitrocellulose are an illustration: Anln
- I _
At
Kame
Mol. Wt.
Lactone of 4,4-diethyl-5-hydroxypentanoic acid n-ISeptyl acetate n-Butyl pentanoate 1-Isopropyl-2-methylpropyiacetate I-Propylbutyl acetate
156.2 158.2 158.2 158.2 158.2
(from Toluene D R ) +0.0046 -0 0015
- 0.0017 -0.0024 -0.0037
The most compact shape, the lactone, has a positive temperature coefficient of solvent ability whereas the more extended structures exhibit negative coefficients or an inverted solubilitytemperature relation. Furthermore, the inversion is greater with branching and increases with the extension of the branches. Further evidence indicating that greater inversion is attained by greater extension of the branched chains is shown by the following compounds as solvents for nitrocellulose: A n /_ n -_
Name 1-Propylbutyl acetate Nonyl Celioaolve isophorate Heptadecyl Cellosolve isophorate
At (from
Mol. Wt.
Toluene D R )
158.2 364.5 466.7
-0 . 0 0 5 3
-0.0037
- 0.0087
INDUSTRIAL AND ENGINEERING CHEMISTRY
May, 1946
The structural formulas of the so-called isophorates are:
alcohol are examples of water-soluble macromolecular substances that gel on heating and liquefy on cooling. Since water solutions of gelatin gel on cooling whereas water solutions of polyvinyl alcohol gel on heating, the phenomenon is obviously not caused by the peculiar behavior of the solvent (water) but rather by some differences in the behavior of the resins. I n recent years this laboratory devoted a great deal of study to a type of sol-gel transformation that is brought about by a change in the solvent ability of the solvent rather than by a change in the di,ssolved resinous substances. TT7ith solvents of low molecular weight shawing a positive temperature coefficient of solvent ability, such gels liquefy on heating. With solvents of high molecular weight showing a negative temperature coefficient of solvent ability, gels can sometimes be made that liquefy on cooling. Since other factors such as the viscosity of the pure solvent confuse the observation of the phenomenon, it does not necessarily follow that all gels will liquefy on cooling, which are made n-ith solvents having negative temperature coefficients of solvent ability. The nature of the dissolved resinous substance also influences the sol-gel transformation phenomenon. Nitrocellulose gels liquefy readily; cellulose acetate gels are likewise satisiactory, but Vinylite resin gels are sluggish and often require several days to change phase.
0
I1
C
CHs
f\p/-
\
0
/
CHa
-
PHa PH* -
\ F /H 3
O-CHP-CH~-O-CH
HP
/CHTCHa
‘CH?
\
Konyl Cellosolve Isophorate
53P
CH-CHa
0
‘CH2
\
CH-CH2-CHS
Heptadecyl Cellosolve Isophorate
\CH2
\
0
= -IO°C.
C&
\
CHz \CHI
INFLUENCE OF POSITION OF ACTIVEGROUPS. In linear molecules containing multiple solvating groups, the position of such active groups in the molecule influences the temperature dependence of solvent ability. I n general, the inversion is greater, the farther removed the active groups are from the ends of the molecule. For example, consider the following groups of isomeric esters as solvents for nitrocellulose: Name Di-n-amyl succinate Diethyl sebacate Di-n-nonyl succinate Di-n-hexyl sebsoate Diethyl octadeoanedioate Di-n-heptyl succinate Di-n-butyl sebacate
Mol. Wt.
258 3 268.3 370.6 370.6 370 6 314 5 314 5
--An/n At
(from Toluene DR) -0 0021 +0.0012 -0 0042 -0 0038 -0 0013 -0 0041 -0 0033
In the case of each group, the isomer having the greater extension of hydrocabon chain beyond the active group exhibits the greater decrease of solvent ability with temperature. SOL-GEL TRANSFORMATION
The gelation of solutions of water-soluble macromolecular substances on cooling is a commonplace phenomenon that has been thoroughly studied. The process is reversed on heating, and the gel liquefies. The change from the sol to the gel and vice versa is called “sol-gel transformation”. Most of the systems that have been studied involve gelation on cooling and liquefaction on heating. This normal situation may be explained in the usual manner on the basis of the greater effectiveness of the solvent a t the higher temperature. A few instances are known, however, where the reverse situation occurs in water solution. Methylcellulose and polyvinyl
o*6 2b0
ab0
M 4;)o
5bo
Figure 6. Vinylite VYNS Threshold Concentrations us. Molecular Weight for Di-n-Alkyl Phthalates 1. Dimethyl Diethyl 3. Di-n-butyl 2.
4. Di-n-hexyl 5. Di-n-octyl 6. Di-n-dodecyl
Recently, certain mixed cellulose esters have been investigated.
A solvent, dibutyl Carbitol benzenephosphonate, that showed an appreciable negative temperature coefficient with cellulose acetate as the dissolved substance, gave a normal temperature dependence with cellulose acetate propionate. Furthermore, whereas a gel made with cellulose acetate liquefied readily on cooling, the reverse situation was observed when a gel was mad0 with cellulose acetate propionate; th& is, the gel liquefied on heating. These differences in behavior seem to be due to the fact that the presence of longer side chains on the mixed cellulose ester makes this resin more temperature sensitive ; therefore, more aggregation of the resin takes place than is the cme with cellulose acetate for a given reduction in temperature. Thus the tendency of the resin to aggregate outweighs the tendency toward greater solvation at the lower temperatures which would be expected by virtue of the increased solvent ability of the solvent at reduced temperatures. To liquefy a gel on cooling, the solvent must actively solvate the resinous substance, and it must have an appreciable negative temperature coefficient of solvent ability over the desired temperature range. The mechanism of the sol-gel transformation is
I N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY
538
TABLE I. m Acetate Methyl Ethyl n-Propyl n-Butyl n-Amyl n-Hexvl n-Heptyl n-Octyl n-Dodecyl
74.1 88.1 102.1 116.2 130.2 144,2 158.2 172.3 228.4
NITROCELLULOSE l/v
P
0,8942 0.8644 0.8560 0,8509 0.8480 0.8462 0.8443 0.8427 0.8403
12.1 9.80 8.36 7.32 6.51 5.86 5.33 4.89 3.68
THRESHOLD COSCENTRATIONS DR
-
,~DR
AT
THREE TEMPERATURESa DR
P
Vol. 38, No. (5
TIDR
DR
P
n ii I!
Toluene Data
2.41 2.62 2.67 2.61 2.26 1.93 1.66 1.44 0.83
12.6 10.2 8.70 7.69 6.74 6.05 5.50 5.04 3.78
2.32 2.58 2.78 2.70 2.52 2.22 1.98 1.74 1 .04
3.80 2.85 2.30 2.09 1.91 1 88 1.85 1.84 1 8.5
0 9i31 0.9368 0,9218 0.9120 0.9060 0.9012 0.8969 0,8928 0 8861
13.1 10.6 9.03 7.86 6.95 6.25 5.66 5.18 3.88
2 16 2.18 2.78 2.71 2.66 2.31 2.10 1.85
4 1 I
1.16
1 R(1
12.6 10.2 8.70 7.59 6.74
6.70 5.40 4.58 4.05 3.62 3.32 3.08 2.88 2.46
....
13.1 10.6 9.03 7.tG 6.95 6.23 5.66 5.18 3.88
0 75 0.80 0.80 0 80 0.80 0.78 0 77 0 78 0 60
,, !)O
5.50 5.04 3.78
0.88 0.89 0.90 0.87 0.86 0.82 0.79 0.75 0.53
13.0 11.2 9.36 8.10 7.15 6.39 4.85
4.29 4.23 4.11 3.93 3.64 3.30 2.36
2.57 2 15 1.83 1.64 1.54
13.6 11.2 9.36 8.10 7.15 6.39 4.85
0.68 0.79 0.88 0.95 1.00 1.00 0.92
8 10 6.20 4.98 4.1.5 3.57 3.19 2.52
.,.. ....
6.13 5.01 3.76 3.01 2.50 1.88
2.63 4.48 3.90 2.87 2 22 1.44
1 60 0 91 0 77 0 78 0.78 0.77
6.13 5.01 3.76 3.01 2.50 1.88
135 2.00 2.13 1.45
1'60 1 00 0 80 0 77
1.72 2.30 2.52 1.96 1.45 0.80
2.82 1.81
2.10
1.55 1.11 0.97 0.95 0.93
1.1506 1.096 1.041 1.004 0.983
4.91 4.17 3.27 2.68 2.28
0.111 0.095 0.090
0.9926 0,9408 0.9194
3.72 2.17 1.52
3 02 2 12
2 10 1 !ill I 87 I ,8:1 1 .RL)
n-Heptane Data Acetate Methyl Ethyl n-Propyl n-Butyl n-Amyl n H exy 1 n-Heptyl
~:E&l
74.1 88.1 102.1 116.2 130.2 144.2 158.2 172.3 228.4
.... ..., .., , . .. . .... . .. .
12.1 9.80 8.36 7.32 6.51 5.86 5.33 4.89 3.68
.. . .... ,
. . ..
0.92 0.95 0.96 0.95 0.91 0.88 0.83 0.78 0.53
6.30 5.03 4.26 3.76 3.40 3.12 2.92 2.75 2.40
.
I
.
.
....
..,. ,... ,...
.... ....
6.06
....
....
.... .... .... .... ....
....
50
r, n:!
p
3,;
si
;I .?, 5 0
3 20 2 9.i 2 42
Toluene Data $-Ketone Acetone Methyl et11TI Methyl n-propyl Methyl n-butyl Methyl n-amyl Methyl n-hexyl Methyl n-nonyl
58.1 72.1 86.1 100.2 114.2 128.2 170.3
0.7557 0.7739 0.7774 0.7840 0,7899 0.7934 0.8026
.
13.0 10.7 9.@4 7.82 6.90 6.19 4 71
4.30 4 22 4.01 3.76 3.48 3.04 2.12
2.45 2.05 1.80 1.64 1.54 1 53 1 51
0,7897 0.8053 0.8065 0.8110 0,8154 0.8182 0,8255
1.48 1 44
0.8238 0 8367 0,8355 0.8380 0,8409 0.8429 0.8484
I4 2
4 07
11 6
4 0.; 3 . 9c
9.70 8 36 7 36
0.56 4 98
3.89 3.75 3 43 2.56
1 2 1 1
?(I X(1 !I;>
71 1 53 1 4x 1 40
n-Heptane Data %Ketone
58.1 72.1 86.1 100.2 114.2 128.2 170.3
....
.... .... .... ..., .... ....
13.0 10.7 9.04 7.82 R.90 6 :l9 4.71
0.81 0.91 1.01 1.03 1.06 1.06 0.96
.... ,... .... ,... ,...
14.2
0.50
,...
11.6 9.70 8.36 7.36 6.56 4.98
0.59 0.70 0.82 0.88 0.93 0.90
1.2181 1,1437 1.0705 1 ,0287 0,9999 0.961
6.26 6.14 3.85 3.07 2.56 1.91
2.53 4.28 4.30 3.68 3.01 2.06
....
6.26 5.14 3.86 3.07 2.56 1.91
...
..,. ....
....
..(.
0 , JO 7.30 5.70 4.60 3.91 3.40 2.m
Toluene Data Phthalate Dimethyl Diethyl Di-n-hut31 Dl-n-hexyl Di-n-octyl Di-n-dodecyl
194.2 222.2 278.3 334.4 390.5 502,8
1.1634 1.0935 1,0234 0.9826 0,9555 0,925
5.99 4.91 3.68 2.94 2.45 1.84
3.00 4.44 3.21 2.33 1.80 1.09
1.78 0.97 0. 7,3 0.67 0.61
0.02
n-Heptane Data Phthalate Dimethyl Diethyl Di-n-butyl Di-n-hexyl Di-n-octyl Di-n-dodecyl
194.2 222.2 278.3 334.4 390.5 502,8
..., ....
Bucoinate Dimethyl Diethyl Di-n-hutjl Di-n-hexyl Di-n-octyl Di-n-dodecyl
Phosphate Tn-n-butyl Tri-n-oetyl Tri-n-dodecyl
....
...
...
....
...,
5.99 4.91 3.68 2.94 2.45 1.84
0.76 1.45 1.95 1.78 1.08
2.79 1.50 1.00 0.88 0.88
146.1 174.2 230.3 286.4 342.5 454.5
1,088 1.811 0.950 0 4 921 0.904 0.884
7.44 5.80 4.13 3.22 2.64 1.95
1.88 2.54 2.17 1.50 1.04 0.51
2.58 1.64 1.30 1.29 1.29 1.29
1.120 1.042 0.976 0.944 0.926 0.905
7.66 5.98 4.24 3.30 2.70 1.99
234.3 262.3 318.4 374.5 430.6
1.0918 1.041 0.988 0.966 0.934
4.66 3.99 3.10 2.56 2.17
2.10 2.70 1.86 1.40 1.04
1.50 1.08 1.08 1.07 1.06
1,1212 1.068 1.015 0.980 0.968
4.79 4.07 3.19 2.62 2.21
266,3 434.6 602,9
0.9514 0.8974 0.8782
3.57 2.07 1.46
0,126 0.121 0.1P6
0.9770 0 9191 0.8988
3.66 2.12 1.49
. I . .
....
....
.... ....
.... . . I .
I
.
,...
.
.,..
.... .... I . . .
...
,..
1.22 1.87 2.24 2..08
1:73 1.07 0.79 0.02
1.70 2.12 2.48 2.26 1.80 Freezes
2.92 1.97 1.25 1.04 0.99
Toluene Data
27.3 16.1 11 6
-
2.67 2.29 1.77 1.39 32.0 21.1 15.5
1.20
1.11 1.10 1.10
1.152 1.073 1.002 0.967 0.949
7.89
6.16
1.69 1.14 0.90 0.82 0.79
1.90 2.67 2.64 2.27 1.90 36.0 26.4 19.4
0.101
0.079 0.074
-
m = molecular nei-ht. p = densitv, t molar 0 Half-second PXP nitrocellulose l o t 100 was used. All solvents xere of very high purity (Table I\'). threshold concentration, o r moles solvent per liter at dilution ratio end-poibt = ( l / u ) / ( b R 1). volume, liters per mole; DE = di1;tion ratio: n D B
illustrated by Figure 9, based on the data of Table 111. These data were taken by measuring the viscosities of the members of a series of solutions of fixed solids content which differed only in the solvent-diluent ratios used in making them up. The mole concentration of solvent, n, in the solvent-diluent portion of each solution thus progressively changed in passing from member t o member in the series. The previous paper (3)shoLYed that, when the relative viscosities of the niembers of such a series of solutions were plotted against the reciprocal solvent concentration, l / n , the resulting
+
curve was relatively flat but suddenly rose almost vertically as reciprocal solvent concentration approached the reciprocal threshold value, l / n ~ determined ~ , independently by means of a dilution ratio titration. In the present paper a similar study is made involving two solvents and txo temperatures. In a series of nitrocellulose solutions in mixtures of l-propylbutyl acetate and toluene (Figure 9A) we are dealing with a solvent that has an appreciable negative temperature coefficient of solvent ability. T h a t is, n b o is greater than and, therefore, l/nbo is smaller than l/nlo. Consequently, if w e plot relative viscosity
-
539
INDUSTRIAL AND ENGINEERING CHEMISTRY
May, 1946 against reciprocal concentration, l/n, at these two temperatures, the resulting two curves must cross each other as they approach the vertical asymptotes (at reciprocal threshold concentration) since the 50' asymptote is reached before the 20" asymptote. If a sol is made up with asolvent concentration of n = 2.45 moles per liter (at 3.0 grams nitrocelluose per 100 cc. volatiles), it will be fluid a t 20" C. When this sol is heated to 50" C.. the mole concentration will fall to 2.39 as a result of thermal expansion. and l / n will be 0.418.
TABLE 11.
VINYLITE
RESINTHRESHOLD
CONCENTRATIONS AT THREE TEMPERATURES' C-. --ZOO C-. --IOo C.l/v D R nDR p l / v D R nDR p l / v D R nDR n-Butanol Data
-50" m
p
Phthalate Dimethyl Diethyl
194.2 222.2
1.163 5.99 0.91 3.11 1.094 4.91 1.06 2.38
Di-n-butyl Di-n-hexyl Di-n-octyl Di-n-dodecyl Succinate Dimethyl Diethyl Di-n- ropyl Dl-n-hyl Di-n-hexyl Di-n-octyl Di-n-dodecyl Phosphate Tri-n-butyl Tri-n-oct 1 Tri-n-do&cyl
278.3 334.4 390.5 502.8
1.023 0.983 0.956 0.925
146.1 174.2 202:2 230.0 286.4 342.5 454.5
1.088 7.44 0.77 4.20 1.011 5.80 0.90 3.05 0.9728 4.79 0.85 2.59 0.9500 4.13 0.80 2.30 0.9210 3.22 0.61 2.00 0.9040 2.64 Q.44 1.83 0.8840 1.95 0.15 1.70
266.3 434.6 602.9
0.9514 3.57 2.18 1..12 0.8974 2.07 1.60 0.79 0.8782 1.46 0.98 0.74
1.191 6.13 0.75 3.50 1.118 5.01 0 89 2.65
1.218 6 26 0.59 3.94 1.144 5.14 0.80 2.85
n-Heptane Data
3.68 1.24 1.64 2.94 1.16 1.36 2.45 1.12 1.19 1.84
1.047 1.006 0.977 0.943
....
fore, this sol must set to a gel at 5Q". Cooling reverses the process, and a fluid sol is restored a t 20" C., where l / n is 0.408. On the other hand, if we
'
3.76 1.15 1.75 3.01 1.20 1.37 2.50 1.19 1.14 1.88 0.90 0.99
. . . .
0.961 1.91
1.120 7.66 0.63 4.70 1.042 5.98 0.87 3.20 1.001 4.95 0.87 2.65 .0.97604.24 0.83 2.32 0.9440 3.30 0.69 1.95 0.9260 2.70 0.56 1.73
1:073 6:i6 0167 3:69 1.030 5.09 0.84 2.76 1.002 4.35 0.87 2.33 0.967 3.38 0.83 1.85 0.949 2.77 0.72 1.61
0.9770 3.66 2.22 1.14 0 9191 2 12 2.20 0 66 0:8988 1:49'1.50 0:60
0.9926 3.72 2.40 1.09 0.9408 2.17 2.88 0.56 0.9194 1.52 2.61 0.42
. . . . . . . . .
..........
-,
5.0
1.071 3.85 1.20 1.75 1.029'3.07 1.34 1.31 1.000 2.56 1.28 1.12
't,
*
5OOC.
hydroxypentanoic consider coefficientthe of solvent. case ofacid ability, a solvent (Figure such that as 9B), the has the lactone a positive 20"of asymptote 4,4-diethyl-5temperature is reached before the 50" asymptote, since nM)is less than mo. Therefore, the two curves can never cross and it is impossible to cause gelation by heating or liquefaction by cooling. Furthermore, since the two solvents have the same molecular weight, and since the same diluent (toluene) and the same solute (nitrocellulose) are used in both illustrations, it is evident that the shape of the molecule is the controlling factor that differentiates between
TABLE 111.
VISCOSITY OF
2.0 3a
-
.
m
nso
-----,-I~
Same plus toluene Lactone of 4,4-diethyl-5-hydroxypentanoicacid Same plus toluene
1-Propylbutyl acetate
SOLUTIONS AT
W
l/nso
'160
50" AND 20'
= 30 Grams/Liter
In l/v
158.2 0.8361 0.8370 ... 0.8373 0.8374
... ...
5.28 3.875 2.91 2.81 2.39
0.1893 0.2580 0.3435 0.3560 0.418
0.00722 0.00594 0.00530 0.00529
4.93 5.13 5.24 5.24 .
156.2 0.9881 0.9570 0.9232 0.8878 ... 0.8626
... ... ... .......
6.33 4.88 3.41 1.943 0.97 0.873 -0.65
0.1580 0.2100 0.2031 0.615~1 1.030 1.148 1.54
0,03813 0.01812 0.01021 0.00656 0.00507 0.00493
3.27 4.01 4.58 5.03 5.28 5.31
158.2 0.8628 0.8642 ... 0.8648 0.8649 0.8650
5.45 4.00 3.00 2.90 2.70 2.30
... ... ...
Same plus toluene
Temperature, 20° C. 0.1836 0.01180 4.44
0.2500 0.3330 0.3448 0.3705 0.434
0.00915 0,00797 0.00785 0.00766
I
Lactone of 4,4-diethyl-5-hydroxypentanoic acid Same plus toluene
t-
... ... ... ... .......
156.2 1.6107 0.9805 0.9479 0.9137 0.8897
6.47 5.00 3.50 2.00 1.00 0.90 0.72
i
1.0 :
In 1/? In??
'150
Temperature, 50' C. 1-Propylbutyl acetate
~
x
c
HALF-SECOND NITROCELLULOSE PSO
-
:
0
w - 0
Solvent
0
0.1547 0.2000 0.2859 0.5000 1,000 1.111 1.39
4.69 4.83 4.85 4.87
.
0.08813 2.43 0,03413 3.38 0.01667 4.09 0.00985 4.62 0,00741 4.90 0.007234.93
-
0.0498 0.0460 0.0529 0.0739
. . .. . . . . . . . . . . . . . ,
,
3.00 3.08 2.94 2.61
. . ., . .. , . . . . . .
. . . . . . . . . . . . . . .
0.1256 2.08 0.1231. 2.10 0.1450 0.1572 1:85 0.2503 1.39
. .. . .
. .. . .
. .. . .
1.93 2.05 2.30 2.63
. .. . .
. .. . .
. .. . .
2.36 2.59 3:OO 3.48
. .. . .
. .. . .
. . . . . . . .
-
c.' W
=
1150
120 Grams/Liter In I/?
..........
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.319 1.089 0.782 0.960 3.70 4.38
. . . .
. . . .
. . . .
-0.84 -0.09 f0.25 f0.04 -1.31 -1.48
. . . .
. . . .
. . . .
. . . .
4.11 4.10 4.33 4.99 6.59 6.79
. . . .
.
. . . .
. . . . . . . . . .
9.60 3.98 2.77 5.14 55.90 81.1
- -
-2.26 4.69 -1.38 4.76 -1.02 5.11 -1.64 6.26 -4.02 8.92 -4.409.33
Ha!f-second P X P nitrocellulose, lot 100, was used. All solvents were of very high purity (Table IV). m molecular weight: p denaity; n centration of solvent, moles/liter volatiles; W concentration of solute, grams/liter volatiles; q absolute viscosity, poises; qr relative viscosity cosity of solution/viscosity of solvent). (I
. . . .
--
con(*-
~
540
Vol. 38, No. 5
INDUSTRIAL AND ENGINEERING CHEMISTRY
An excellent example of a nitrocellulose gel that liquefies readily on cooling employs di-n-hexyl Carbitol succinate in the following proportions:
\
'i,
-.&, -*
c
-0.4 -
0.0 0.6
0.2
x
= 5OOC.
0
= -IO°C.
32 .O% di-n-hexyl Carbitol succinate 64.0% toluene 4 , 0 % dry half-second nitrocellulose 100.0% by weight
-
x - x
I
I
200
Gels are formed (3) from solutions of macromolecular substances when the solvent concentration is reduced to a value approaching the threshold concentration, provided a high enough solids content is employed to avoid precipitation. Therefore, it is not always necesmry t o have three-component syst,ems, such as have been employed in the previous examples. If a solvent is chosen in which the mole concentration of the pure solvent is close to the threshold concentration, no diluent need be added in order to prepare a gel. If, in addition, the solvent ability of the solvent improves on cooling, it may be possible to liquefy the two-component gel on cooling. Examples of two-component gels that liquefy on cooling are:
I
I
300 M 400
500
Figure 9. Viiiylite VYNS Threshold Coiicentrations RIolecular Weight for Tri-n-Alkyl Phosphates with n-Heptane as Diluent
t's.
1. Tri-n-butyl.
2. Tri-n-octyl.
3. Tri-n-dodeeyl.
9 5 . 0 7 di(1-methyl-3-propoxyhexyl) succinate
Og
sol-gel transformation is a dynamic phenomenon taking place in a relatively concentrated solution, whereas the dilution ratio titration reaches an end point at very dilute solute concentration under equilibrium conditions which may be attained gradually. T h a t is, the dilution ratio measurement does not involve a time rate of change of solvent ability which might conceivably be influenced by solute concentration. The following nitrocellulose solvents are examples of this difference in performance:
5. dry half-second nitrocellulose lOO,O% by neight
9 7 . 5 7 di(2-ethylbutyl Cellosolve) beniene phosphonate 2 . 5 8 dry cellulose acetate (Eastman .4-:2) 100.0% by weight
SUMMARY
1. Temperature dependence of solvent ability in homologous series of solvents is influenced by the molecular weight of the -~ An/n , solvent. I n general, the members of low molecular weight become better solvents with rising temperature and the members of At (from Sol-Gel Resin, high molecular weight become better solvents xith falling temName m Toluene DR) Trans. % perature. Di-n-dodecyl methane2. The shape of the solvent moleculc influences temperature phosphonate 432.6 -0,0047 Rapid 4.1 dependence of solvent ability more than its weight. In general, Tri-n-octyl phosphate 434.6 -0,0070 Slow 3.7 compact molecules have a positive temperature coefficient of solvent ability, whereas extended molecules have a negative temperature coefficient of solvent ability. 3. Branched molecules are likely to exhibit greater decrease of TABLE Iv. CRITERIA O F P U R I T Y OF C O X P O U X D S USED' solvent ability Tvith rise in temperat'ure than their linear isomers, and the amount of such inversion i s more pronounced, the greater % Ester % Bcid Succinates , 99.6 0.119b the linear extension of the branches. Di(rnethy1 Cellosolve) 99.8 0.014 4. I n linear solvent molecules containing multiple active Di(ethy1 Cellosolve) 102.2 0.106 groups, the inversion is greater, the farther removed the active Di(wbuty1 Cellosolve) 0.023 Di(n-hexyl Cellosolve) 100.6 groiips are from the ends of the molecule. Phosphates Tri-n-butyl 0.024c 5 . Gels involving organic solvents and macromolecular solutes Tri-n-octyl may be made which may be liquefied by heating or by cooling, I 0021: .B1 1 0 .0 05: 3 Tri-n-dodecyl depending on the choice of solvent,. Solvents having, a positive Lactone of 4,4-diethyl-5-hydroxypentanoicacid 98.1 0.657d temperature Coefficient of solvent ability may ordinarily be used 09.9 1-I'ropylbutyi acetate Methyl acetate 99.9 0.0005~ to produce gels that liquefy on heating. Gels that liquefy on a Othkr compounds are listed in Table 1 of the previous paper ( 3 ) . cooling may often be made froin solveilts that show a considerb As succinic acid. able negative temperature coefficient of solvent ability. C As phosphoric acid. 6 . . Linear solvent molecules are generally more effective for d As 4 4-diethyI-j-hsdrox?.pentanoic acid. use in promoting an abrupt sol-gel transformation t>han are 8 As abetio acid. branched. molecules although the latter may often exhibit a greater negative temperature coefficient of solvent ability. 7. Gels that liquefy on cooling may be made of systems of I 1 two or more components and may employ resinous substances of quite different character although 6the time rate of the sol-gel transx = 5OoC. x = 50°G. formation may vary with different resinous substances. e : 20%.
::
LITERATURE CITED
(1) Berl, E., and Koerbcr, W.,J . Am. Chem. Soc., 61, 154
.I-+-*2 I
I
0.2
0.3
9. _Figure _
4
0.4
I /n
1
I
0.5
I .o
I .5
Relative Viscosity of Solutious of Half-Second Nitrocellulose
(1939). (2) Byron, M. L., J . P h y s . Chem., 30, 116 (1926). (3) Doolittle, A. K., IND. EKQ. CHEM., 36,239 (1944). on the proglam of the Di31of Paint, Vainish, and Plast~cs Chemistry of the 1045 Ueeting-in-l'nnt, . b l E R I C A N CHEWCALS O C S E T i . PREsshim
hion
A: Mixtures of 1-propylbutyl acetate with toluene. 30 g r a m a nitrocellulooe per'liter of volatilas *
E . Mixtures of lactone of 4,4-diethyl5hydroxypentanoic acid with toluene; 120 grama nitrocellulose per liter tolatiles
