acids in general may alter the activity of ai1 ot' modified forms of the enzyme with more in-

enzyme by causing the rupture of intramolecular ternal degrees of freedom, possessing less catalytic
hydrogen bonds in the enzyme protein. The re- activity.
sults of such bond breaking may be the arising yT,lxr
yoKK,
N, 17 SEPTEMBER
RECEIVED 27, 1939

[CONTRIBUTION PROM THE BASICSCIENCE RESEARCH

LABORATORY
O F I"E C'NIVERSITY OF CINCINNATI]

Kinetics of Polyesterification: A Study of the Effects of Molecular Weight and

Viscosity on Reaction Rate'
BY PAULJ. FLORY

Iatroduction observe the effects of molecular weight and vis-

Investigators in the field of macromolecular cosity on reaction rate. Both molecular weight
chemistry have been prone to attribute the ap- and viscosity increase continuously as such reac-
parent sluggishness of some polymer reactions tions proceed. It should be possible, therefore,
either to an inherent dimin4shed reactivity of to observe the effects of molecular weight and
large molecules, or to the high viscosity of the viscosity on reaction rate, if such effects exist.
reaction medium. Various arguments have been The present investigation has been primarily
advanced in an effort to show theoretically that concerned with the kinetics of polyesterification
increase in size of the molecule and increase in reactions between glycols and dibasic acids.
viscosity should decrease reactivity. 2,3 How- The kinetics of several esterifications involving
ever, no direct experimental evidence has been monofunctional reactants have been investigated
advanced to prove even the existence of an effect also, in order that the polyesterification reactions
of either molecular weight or viscosity on reac- could be compared with similar reactions in which
tivity of polymer molecules. neither molecular weight nor viscosity underwent
The problem of the effect of viscosity on reac- appreciable change. All of these reactions have
tion rate of liquid phase reactions in general has been followed by titration of the total free car-
been the subject of frequent discussion,L but boxyl in samples removed from the reaction mix-
again conclusive experimental evidence does not ture a t suitable intervals.
seem to be available. Moelwyn-Hughes6 has In contrast to monofunctional esterification
concluded that the rate of collision between solute where there are only two species of reactants,
and solvent molecules should increase propor- many species of molecules are present during
tionally with viscosity and, hence, the rate of a polyesterification. Any one of these species bear-
reaction between solute and solvent should be pro- ing an hydroxyl group may react with any other
portional to viscosity also. More recently Rabin- bearing a carboxyl; the product of this reaction
owitch and Wood6 have shown with exceptional may react with another molecule, etc. Many re-
clarity that one should expect no effect of viscosity actions are occurring simultaneously: i. e., mono-
on reaction rate, except a t very high viscosities mers react with monomers, dimers, trimers, etc.,
(very low rates of diffusion) or when a large frac- dimers react with monomers, dimers, trimers, etc.,
tion of collisions leads to reaction. trimers react with monomers, etc., etc. A com-
The study of the kinetics of a condensation plete kinetic analysis, treating each of these reac-
polymerization offers a splendid opportunity to tions individually, is obviously hopeless. All of
(1) 4 portion of this work was included in a paper presented before these reactions, however, are chemically identical,
ction, Paint and Varnish Division of the
'ty at the Boston Meeting, September 14, and their rates may differ only in so far as re-
1939. activity is affected by molecular weight. In-
(2) C E. H. Bawn, T r a m F O Y ~ ~SOCU Y, 32, 178 (1986).
('j) H. Dostal, Monatsh , 67, 63 (1935), 70,409 (1937). H. Mark, stead of attempting to unravel the maze of in-
Xature, 140,8 (1937). dividual reactions, one may consider merely the
(4) See E A. Moclwyn-Hughes, "Kinetics of Reactions in Sofu-
tion," Oxford Univ Press., N e w York, N. Y , 1933, p 51 reaction of all functional groups, disregarding the
ii)Reference 4, pp 19, I.T9 p i wq size of the molecule to which each group is at-
(6) E Rahinowitcb and W'. C Wood Tvans Faraday SOC, 82,
1381 (1986) tached. T i , as usually has been supposed, reac-
Dec., 1939 EFFECTS
OF MOLECULAR
WEIGHT AND VISCOSITY ON REACTIONRATES 3335

tivity decreases with increase in size of the mole- to dissolve slowly, werk dissolved in hot chloroform, cooled
cule, then the ratio of the rate of polyesterifica- and then titrated. The end-point was very sharp.
Reactions catalyzed with p-toluenesulfonic acid were
tion to the rate of a simple esterification a t the carried out in the same manner except that, due to the
same degree of reaction should decrease as the lower temperature, 109". it was necessary t o reduce the
reactions progress. The results presented below pressure t o about 200 mm. in order to ensure rapid re-
show that polyesterification and esterification re- moval of water.
actions follow identical courses. Experimental Results
Experimental From the neutral equivalent determined by
Materials.-Diethylene glycol, lauryl alcohol, adipic titration one can calculate both the extent of reac-
acid, lauric acid, and caproic acid were obtained from the tion p , i. e., the ratio of the number of ester groups
Eastman Kodak Company. Decamethylene glycol, which
to the number of ester plus free carboxyl groups,
had been prepared by hydrogenation of diethyl sebacate,
was obtained from the University of Illinois. The di- and the degree of polymerization DP, which is
ethylene glycol was carefully fractionated through a car- defined as the average number of glycol and di-
borundum packed column. The boiling point was prac- basic acid residues per molecule. The number
tically constant throughout almost the entire distillation, of free carboxyl groups will equal the number of
129.5 to 130.0" a t 10.5 mm. A center cut was retained
molecules, provided that equivalent amounts of
for use. Decamethylene glycol was treated with boiling
5% aqueous caustic to remove unreduced ester, then re- glycol and dibasic acid are used and that the only
crystallized three times from ethylene dichloride. The reaction occurring is intermolecular esterification.
final product had a freezing point of 71.9 (*0.1) '. Lauryl When these conditions are fulfilled the neutral
alcohol, freezing a t 21.9 ( i 0 . l ) ', was used without further equivalent must be equal, therefore, to the aver-
purification.
Adipic acid was recrystallized from conductivity water;
age molecular weight, and
m. p. 151.5'; neutral equivalent 72.9, 73.0; calcd. 73.07. neut. equiv. - 18
Caproic acid, neutral equivalent 117.6, calcd. 116.2, and
D P = mean segment weight = 1/(1 - P) (1)
lauric acid, neutral equivalent 200.6, 200.9, calcd. 200.3, where the mean segment weight is half the mo-
were used with no further purification.
Apparatus and Procedure.-The apparatus consisted
lecular weight of the structural unit -0RO-
of a Pyrex reaction bulb of about 50 cc. capacity sealed to COR'CO- of the polymer.
the bottom of a 15-mm. tube about 25 cm. long. Through As Goldschmidt' has shown, esterification reac-
a stopper a t the top of this tube a pipet was inserted. The tions are hydrogen ion catalyzed. I n the absence
pipet was used to remove samples, and also for admitting a of a strong acid catalyst a second molecule of the
slow stream of nitrogen through the reaction mixture in
order to facilitate removal of water. The effluent stream of
carboxylic acid undergoing esterification must
nitrogen and water vapor passed through a 12-mm. tube function as catalyst. Therefore, the reaction rate
sealed t o the vertical tube just below its upper end. This is proportional to the hydroxyl concentration
side tube sloped upward at about 45' from horizontal and to the square of the carboxyl concentration.
for about 30 cm., and then led downward to a dry- When hydroxyl and carboxyl groups are present
ice trap. With the apparatus designed in this way loss of
glycol was minimized by condensation in the side arm,
in equal numbers, the process becomes a third
followed by its return to the reaction mixture. Glycol is order reaction,8 for which
lost more readily during the early stages of the reaction; 2kt = 1/C2 - l/Coa (2)
later on nearly all glycol molecules have reacted a t one
end or the other (or both) and hence are no longer volatile.
where k is the velocity constant, C is the con-
The reaction bulb and adjacent tubing were carefully centration of unreacted groups a t time t, and C,
cleaned before each experiment. This is especially im- is the initial concentration. If one neglects the
portant since esterification is an acid catalyzed reaction, effect on concentration of the decrease in volume
and hence is sensitive to traces of acids or bases. due to loss of water, this equation may be re-
Equivalent quantities of glycol (or alcohol) and acid
were placed in the reaction bulb which was then sealed onto
placed by
the vertical tube. The reaction bulb and several cm. of 2Co*kt = 1/(1 - p ) * - 1 (3)
the vertical tube above were heated with a vapor-bath, with which the data presented below will be com-
which remained within 0.2 O of constant temperature pared. The error arising from the use of (3)
throughout any one experiment. Liquids used in the
vapor-bath were toluene boiling a t logo, phenetole a t 166",
rather than (2) is appreciable only during the
and m-cresol, 202'. From time to time samples were (7) H. Goldschmidt and co-workers, 2. ghysik. Chcm., 60, 728
(1907); TO, 627 (1910); 81, 30 (1913); Bey., 99, 2208 (1896).
removed, weighed and titrated (at 20-25') with 0.1 N See also Hinshelwood and co-workers, Trons. Faraday Soc., 80,
alcoholic potassium hydroxide, using phenolphthalein 935, 1145 (1934); J. Chem. SOL, 593 (1939).
indicator. High molecular weight samples, which tended JOURNAL, 69, 466 (1937).
( 8 ) P. J . Flory, THIS
*3;33; P.%T;L j

first part of the polymerization, which is oi minor TABLE

I1
i n k e s t here. LAURYL ALCOHOL-ADIPICA C I D REACTION AT 202'
t Neutral
I
TABLE in rnin cquiv P 1/(1 - P)Z
nlRTHYLENP C L Y C O L --'DII'IC ACID AT 166 O 4 8.13 5 0 2609 1 &3
1 h'eutral 15 .id2 5307 4 53
:n niin eqlllr i' DP 1/(1 - p;z 28 742 6669 9 01
6 143 3 i) 1321 1.. l!i0 I . 34li 4.3 95 3 7425 15 08
13 161 5 2470 1.328 1.7tx G5 1327 8004 25.1
23 1S9.G 3675 1.581 3.50 (1 1 1525 8400 39 0
:37 233 Ll 4975 1.990 3 . si; 116 1772 8624 52.7
59 293.0 A080 2.55 6.47 1-28 2050 8812 70.9
38 :363 5 .D863 3.10 10. 1s 18;: 2330 8957 91.8
129
I-
131 i dl:$ 4.02 16. l i i 226 2581 ,9059 112 9
170 ,732 ,7894 4.75 2 2 . ii 267 2840 9145 136 9
203 608 ,8161 3.44 29.6 322 3105 9219 163.8
235 673 ,8349 6.06 36. '7 37% 3365 9279 192 3
270 738 850c) 6.67 44.4
321 832 .8672 7.33 56.7 diethylene glycol with adipic acid a t 166 and
397 94; ,8837 8.60 73.8 20Zo, and of the same glycol with caproic acid a t
488 1072 ,8971 9.75 95.0
596 1198 .90% 10.92 119.0
101io. It has been necessary to use two graphs
690 1309 9163 11.94 142.5 having different scales in order to show effec-
793 1404 ,9220 12.82 164.0 tively the entire course of the reaction. Dots
YO0 1307 9278 13.77 189.5 and circles are used to distinguish data obtained
1008 1569 03 14.3Y 206 in different experiments. The agreement be-
1147 1891 54 15.48 239
1370 183R $1403 16.81 2x7
tween two such sets of points indicates the re-
1606 I973 '437 18.08 :327 producibility. Data from the reactions of de-
camethylene glycol with adipic acid, lauryl aIco-
Data from two typical experiments are given in hol with adipic acid, and lauryl alcohol with
Tables I and 11. Extents of reaction and DP's lauric acid are shown in Fig. 2.
have been calculated using (1). In Figs. 1 and 2,
1:(1 - is plotted against C for the reaction of 200 I 1

14
r/

20 i '
/.
I 1
/ /'
I f
?'

15
DE-./
2027
DE-A
2027 / 1 DE-A
DE-A
166"/
166"/
/
12

1" 3
(tx2)! A'

*
400 800
Time in minutes.
Fig. 2.-Diethylene glycol-adipic acid (DE-A) and
01 I I I diethylene glycol-caproic acid (DE-C) reaction including
50 100 150 later stages of the reactions. Time values a t 202" have
Time in minutes. been multiplied by two.
Fig. 1.-Diethylene glycol-adipic acid (DE-A)
and diethylene glycol-caproic acid (DE-C) reactions
During the portion of the reaction shown in
during initial stages. Time values a t 202" have been Fig. 1 the relationship between 1/(1 - $ > 2 and t
multiplied by two i s definitely not linear, in disagreement with the
Dec., 1939 OF MOLECULAR
EFFECTS W E I G H T A N D VISCOSITY ON REACTION
RATES 3337

third order equation (3). However, a t the larger fications within each group there are no important
extents of reaction included in Figs. 2 and 3 the differences in the values of the velocity constants.
data do approach this linear relationship. Straight The lower rates of the esterifications compared
lines have been drawn through the points in with polyesterification (see Fig. 3) are due merely
these two figures in order to show clearly the to the diluting effect of the larger proportion of
range over which 1/(1 - p ) z is approximately hydrocarbon chain in the esterification reactions
linear with t. Attempts to fit the data to an equa- which have been studied.
tion for a reaction of lower order have not been These velocity constants are some ten to
successful; although such equations offer some- twenty times larger than those calculated from
what better agreement within a limited range Fairclough and Hinshelwood'slo data for esteri-
during the early portion of the reaction, they fail fication of dilute solutions of various aliphatic
completely a t higher extents of reaction.
In the case of the diethylene glycol-
caproic acid reaction, reliable data could
not be obtained beyond the range given
in Fig. 2, due to excessive loss of the com-
paratively volatile caproic acid. With
diethylene glycol and adipic acid a t ex-
tents of reaction somewhat beyond those
covered in Fig. 2 the points fell consist-
ently below the extrapolation of the
straight lines shown in the figure. This
is attributed to the effect of the inevitable
slight loss of the glycol by volatilization.
Small losses of one reactant are of no con-
sequence except a t large extents of reac- I 1 I 1 I

tion where the concentration of unreacted 200 400

functional groups is so low that a small Time in minutes.
deficiency of one functional group pro- Fig. 3.-Decamethylene glycol-adipic acid (DM-A), lauryl al-
duces an appreciable reduction in rate. cohol-adipic acid (L-A), and lauryl alcohol-lauric acid (L-L)
reactions at 202'.
This deviation sets in a t a much later
stage in the reaction of decamethylene glycol with acids in ethanol. The energy of activation, 13
adipic acid, since this glycol is less volatile. kcal., calculated from rates of the diethylene
Careful examination of the data presented in glycol-adipic acid reaction a t 166 and a t 202O,
the figures fails to reveal any significant dis- is slightly less than the value, 15.0 kcal., obtained
similarities between polyesterification and esteri- by these authors for esterifications in ethanol.
fication. The courses which they follow are so TABLEI11
similar that their points may be made to follow VELOCITYCONSTANTSFOR POLYESTERIFICATION AND
approximately the same curve by merely multi- ESTERIFICATION
plying all time values for one of them by a suit- k X lo-* a t 202'
in (equiv./1000 9.) -2

able factor. Reaction min. -1

Velocity constants for the esterifications and

H+(CHd-O-(CHh--OH +
HOOC-(CHg)cCOOH 0.41
polyesterifications included in this investigation H + ( C H ~ ) P O - ( C H ~ ) ~ O H +
H-(CHz)s-COOH 0.36
are given in Table 111. These have been evalu- HO-(CHa)I-OH + HOOC-(CHz)~-COOH -0.5
ated, using (3), from the slopes of the straight HO-(CHg)i(rOH + HOOC-(CHr)4-COOH 1.63
H-(CHa)irOH f HOOC-(CHz)cCOOH 1.67
lines in Figs. 2 and 3.9 The velocity constants H - ( C H d i r O H + H-(CHdii-COOH 2.5
fall into two groups, those for reactions of di-
ethylene glycol and ethylene glycol being lower Over the portions of the polyesterifications
than those for decarnethylene glycol and lauryl shown in Figs. 2 and 3 the molecular weight in-
creases about fourteen-fold. I n this range the
alcohol. Between polyesterifications and esteri-
(9) The velocity constant for the ethylene glycol-adipic a d d
viscosity increased from about 0.015 to 0.30 poise,
reaction could not be determined accurately due t o the volatility (10) R. A. Fairclough and C. N. Hinshelwood, J . Chcm. S o c ,
of this glycol. The experimental data have not been included here. 593 (1939).
3338 Y A C L J. FLORY VOl. G 1

in the diethylene glycol-adipic acid reaction. In controlling process. In order to test this as-
the esterification reactions the viscosity remains sumption, diethylene glycol and adipic acid were
practically constant, a t about 0.01 poise, except heated (without catalyst) a t 166' under reduced
for a probable slight decrease. The. polyesteri- pressure (about 200 mm.), instead of a t atmos-
fication reaction is unaffected, therefore, by in- pheric pressure as in other experiments. More
crease in viscosity, as well as by increase iri mo- efficient removal of moisture resulting from this
lecular weight, within the range investigated. reduction in pressure should have brought about
Several experiments on the reaction r i f diethyl- an increase in the rate of reaction, if the escape of
ene glycol with adipic acid catalyzed by a small water were the rate determining step. 'The data
amount of p-toluenesulfonic acid (0.004 mole per obtained in this experiment are given in Table IV.
mole of glycol) have been carried out at I C I D ~ .
TABLEIV
These should be second order, according to the
results of numerous researches7 on acid catalyzed RPACTIONOF DIETHYLENE GLYCOLWITH ADIPIC ACID
ITSDER REDUCEDPRESSURE AT 166'
esterifications. For a second order reaction, 1
t interpolated from
i,'(l - fi) increases linearly with time. In strib- n min P expts. at 1 atm.
ing similarity to the reaction in the absence of 6.5 0.1535 6
added catalyst, the slope of 1/(1 - p j plotted SX 5 3947 23
"_
against t (see Fig. 4) increases a t first, but becomes 0;) .5927 55
81 ,6793 84
constant later on. Even as far as DP 90 no
155 ,7633 137
tendency of the slope to decrease, as would occur 1713 7983 176
if reactivity decreased as the average molecular
weight increases, has been observed. The viscos- That the reaction velocity is not affected by the
ity increase accompanying this increase in mo. reduction in pressure is shown in the table by a
lecular weight is over two thousand-fold. comparison of time values with those for the
same extent of reaction interpolated from the
curve in Fig. l . I 3 The rate must be controlled
by factors inherent in the mechanism of the reac-
tion, and not b y escape of moisture as Schulz has
assumed.
Obviously i t is necessary to provide adequate
means for escape of water formed in the reaction;
otherwise the occurrence of the reverse reaction,
hydrolysis, would set up an equilibrium and
further polymerization would cease. According
to the usual practice, which was followed essenti-
ally here, a stream of gas bubbled through the
polyester is used to provide for escape of water.
Only a v e r y gentle gas stream is required to main-
I I I
tain the normal rate, which cannot be exceeded
0 200 400 600 by increasing the rate of bubbling.
Time in minutes. The sltiw rate of advancement of polyesteri-
Fig. 4.-Diethylene glycol-adipic acid reaction at fication when the average molecular weight be-
109' catalyzed by 0.4 mole per cent. of p-toluene- comes large, sometimes attributed in the past to a
sulfonic acid. low reactivity of large molecules or to the diffi-
It has been suggested in the past t h a t the rate
culty of removing moisture, is a consequence of
of polyester formation is governed b y the rate a t
the kinetic order (third) o f the reaction. The
which moisture formed in the reaction escapes rate, being proportional to the third power of the
Recently Schulz" concentration of unreacted functional groups, de-
sis of polyesterifica- creases rapidly as the extent of reaction increases.
tion based on the assumption that this is the rate (13) The experiment a t reduced presmre from which the data in
Table I V were taken was not extended to large extents of reaction
(11) See. for examph, W. H. Carothers, Chem Rev., 8,887 (1931). The excesdve loss of glycol under these conditions might hdvr
(12) G V. Schutz, 2 physik. Chem , AlSP, 127 (1938). vitiated rewlts at larger extents 01 reaction
Dec., 1939 EFFECTS
OF MOLECULAR
WEIGHT AND VISCOSITY ON RATES
REACTION 3339

Although initially the reaction is rapid, a very where (Fab*/FaFb)is the ratio of the partition
long time is required for close approach to com- function for the activated complex to the product
plete reaction, which is essential for attainment of of the partition functions for the reactants, and E,,
high molecular weight. is the activation energy a t the absolute zero.
Discussion Evans and Polanyi16 have shown that the parti-
The failure of the reaction to adhere to the tion function factor Fob*/FaFb must be smaller
third order equation (3) during the first stages of for the recombination of radicals (e. g., CH,)
the reaction is difficult to explain. A part of than for the recombination of atoms (e. g., H).
this deviation can be attributed to the decrease in Bawn considers that this decrease in “steric factor”
volume due to loss of water (i. e., to the error will continue indefinitely as the complexity of
arising from the use of (3) rather than (2)). This the molecule increases. Thus, in the vinyl type
will not account for the entire deviation, however. of polymerizations, in which a polymer molecule
Inasmuch as esterification is catalyzed by bearing a free radical adds on monomers con-
hydrogen ions (or hydrogen donors, i. e., acids), secutively, he concludes that the rate should de-
i t probably is sensitive to the nature of the reac- crease with each successive addition of monomer,
tion medium. For example, the change in di- the rate finally approaching zero when the poly-
electric constant which occurs as hydroxyl and mer molecule becomes very large.
carboxyl groups are converted to ester may in- An increase in complexity of one of the react-
fluence the reaction rate. I n dilute solutions, ing molecules modifies both Fab* and the prod-
which have been used almost exclusively in pre- uct FaFb. Unless the increase in complexity in-
vious work on esterification, the change in prop- volves modification a t a point in the molecule very
erties of the reaction medium as the esterifica- close to the functional group, both Fab* and
tion progresses is inappreciable. If the process is FaFb will be altered by virtually identical factors
sensitive to changes in the reaction medium, then for the added degrees of freedom. The structure
in the present experiments deviation from third in the immediate vicinity of the functional group
order may be expected, except a t large extents of of a polymer molecule is independent of its degree
reaction where one is dealing with a dilute solu- of polymerization. Hence, contrary to Bawn’s
tion of functional groups. conclusion, no decrease in rate of polymerization
The above discussion has been presented due to a decrease in the partition function factor
merely as a possible explanation for the peculiar with increasing length of the molecule should be
course of the reaction. Whatever the cause of expected.
this behavior may be, it is typical of both poly- Rabinowitch and Wood6 have shown convinc-
esterification and esterification. The fact of ingly that the actual collision rate in the liquid
primary importance to this investigation is their phase is not affected appreciably by changes in
similarity, in spite of the high molecular weight viscosity, although the rate of diffusion is mate-
and viscosity attained in the former. rially decreased. Even if the frequency of colli-
Bawn2>l4 has attempted to show, on the basis sion is decreased by low mobility of large polymer
of the transition state method developed by molecules, the duration of collisions must be pro-
Eyringls for the treatment of reaction velocity portionately increased. Hence, the concentra-
from a statistical mechanical point of view, that tion of molecules in the collided state cannot be
reactivity should decrease continually as size of affected by either molecular size or viscosity.
the molecule increases. According to the transi- Consequently, the rate of reaction should be inde-
tion state method16 pendent of these factors.
k = K* (k‘T/h) (4) The same conclusion follows directly from the
where k is the velocity constant, k’ is Boltzmann’s transition state method, since the activated com-
constant, T is the absolute temperature, h is plex is in thermodynamic equilibrium with the
Planck’s constant and K* is the constant for the reactants ; the concentration of activated com-
equilibrium between normal molecules and acti- plexes will not be affected by changes in mobility,
vated complexes. unless the mobility is too low to maintain their
K* = (Fao*/FaFb) exp (--Ea/k‘T) (5) equilibrium concentration.
(14) See also reference 3. (16) M.G. Evans and M. Polanyi, Trans. Favadoy Soc., 31, 875
(15) H. Eyring, J . Chem. Phrs., 8, 107,492 (1935). (1935). See also C. E. H. Bawn,ibid., 31, 1536 (1935).
 Kabinowitch aiid iX’ondGz!7 h a w pointed out therefore, b y simple calculation using the dis-
that at sufficiently hiqh viscosities tlic rate of tlif irihution equations referred to above.
fusion \Till beconw i i l o x ~ ti) maintain i
“equilibrium” number n i pairs of rearinnis a,i Summary
jacent tn each o t l w m the iiqirid (z. P , withirr 1. The kinetics of two polyesterifications,
colliding dislancr of raciz other) Diffusion, diethylene glycol-adipic acid, and decamethylene
rather than the passagr oi molecules through the glycol-adipic acid, have been compared with
activated state, hcconies the r a t e coiltrolling several non-polymer forming esterifications, di-
step when such \-lscosities arc reached, and the ethylene glycol-caproic acid, lauryl alcohol-
rate will no longer be independent oi molecular adipic acid, and lauryl alcohol-lauric acid. Pre-
size and viscosity. i;nr most reactions iihosc liminary experimental data on the reaction of di-
with T’er-y low variies of E? esceptcd) extremely ethylene glycol with adipic acid catalyzed by p-
high viscosities would be reqnired to effect a pcr- toluenesulfonic acid h a w been presented also.
ceptible clecrease in rat e, 2. As the reactions proceed (in the absence of
Some time ago the author published equations’s catalyst) the course followed corresponds to a
for calculating the numbers of molecules of various continuous increase in kinetic order, becoming
Gzes in a linear condensation polymer. The approximately third order a t large extents of re-
validity of these equations rested entirely on the action.
postulate, presented at that time without experi- 3. Polyesterification and esterification follow
mental prooi, that the furictional F?;roups of all similar courses, froin which i t has been concluded
molecules, regardless of size, are equally reactive. that reaction rate is not affected by either in-
Dostalls objected to the size distribution equa- crease in molecular weight or the concurrent in-
tions on the grounds that this postulate could not crease in viscosity.
possibly be valid. The present investigation fur- 4. This conclusion has been shown to be in
nishes proof that this postulate is correct. The agreement with theory, contrary to previous
quantities of molecules of various size$ in a linear theoretical deductions.
condenqation polymer can be determined reliably, 3. The slow rate of polyesterification when the
(17) See also E Rabinowitch, r r a m F a ? a d a r SOL., 33, 1235 average molecular weight is large is due to the
(1937). third order character of esterification.
(18) P J. Flory, Tars JOURNAL. SS, 1877 (1936)
l l O i H D w t a l Uonnish , TO, 3 2 i OXK! CINCINXA’I I, OHIO RECEiVED SEPTEMBER 30, 1939

/CONTRIBUTION
FROM TFIE DFPARTMFNT
OF CHEMISTRY,
THECITYCOLLEGE
OF THE COLLEGE
OF THE CITYOR NEW
\’ORK 1
Acetometallates. 11. Acetozincate
RY ALEXANDER
1,EHRMAN AXD PHILIP S K E L t

It has been shornza that the solubility of metal and a disodium zincate in both the anhydrous
s in aqueous solutions of alkali metal and hydrated condition. On the basis of the ease
hydroxides can best be explained as the formation of removal of water from these compounds i t was
of hydroxyl complex ions. M(OH)2, for example, concluded by Scholder that they were NaZn(OH)B
may dissolve to give M(OH)4-. I n the case o€ and NazZn (OH)4,respectively, and their hydrates,
zinc, Goudriaanl isolated as a solid phase the rather than the various hydrates of NaHZnOz
and NaiZnOa. Subsequently Brintzinger and
Wallach3 showed from the rate of diffusion of
zincate ions through a cellophane membrane that
OH)4,2H20. More recently, their ionic weight is consistent with the formula
rkers2 have isolated a mono- Znz(OH)* =. More light could be t h r m on the
(1) Goudriaan, Rec Irav. chim., 38, 505 (1920). question if the system NaOH-Zn(0H)z could be
(21 (a) R. Scholder, 2. angew. Chem., 46, 50’1 (1933); tb) K
Scholder and H Weber, 2. nnorg. aflpm. Chcm , 216, 355 ( 1 9 3 3 , shown to be a simple binary system with com-
fc) R Scholder and G Hendricfi rhi.3, 241, 76 (1939) ‘ Z ) H Brintzinger and J. Wallach, 2. angew. Chem., 47,61(1934).