R. J. BERNI,R. R. BENERITO, AND H. M. ZIIFLE
1882
is interesting that the inverse dependence of the rate on hydrogen pressure is greater for the platinum catalyst than for the other catalysts. This suggests that the strength of adsorption of hydrogen on the platinum catalyst is greater than on the other catalysts. To summarize briefly, the results of the present work have made it possible to compare the catalytic
activities of a series of supported metals for the reaction of cyclopropane with hydrogen in a more fundamental manner than has been done previously since the metal surface areas were known from chemisorption measurements. In addition, it has been shown that the selectivity of conversion of the cyclopropane depends markedly on the particular metal chosen as a catalyst.
Kinetics of the Zinc Fluoroborate and Hydrogen Ion Catalyzed Hydrolyses of the Diglycidyl Ether of 1,4=Butanedioland of Diglycidyl Etherla
by Ralph J. Berni, Ruth R. Benerito, and Hilda M. Ziifle Southern Regional Research Laboratory,lb New Orleans, Louisiana
(Received December 7, 1964)
Kinetics of hydrolyses of the diglycidyl ether of 1,4-butanediol (DGEBD) and of diglycidyl ether (DGE) a t less than 1.0 M concentrations catalyzed by HC1 (pH 3.25) and by 0.05 M Zn(BF02 from 25 to 90" have been investigated. Semilogarithmic plots of changes in oxirane content with time indicated pseudo-first-order kinetics with both catalysts. Statistical evaluation indicated different rate constants for the opening of the first (kl) and second (k2)oxirane rings with Zn(BF& catalysis but equivalent rates with HCl catalysis. Analysis of data by the modified Swain method confirmed consecutive first-order kinetics for the two diepoxides with both catalysts. Further analysis of data with HC1 catalysis confirmed equivalent rate constants for both ring openings. With the exception of DGE a t go", relative rates (Ic2/kl)for the over-all reaction with Zn(BFS2 catalysis were less than unity at all temperatures. These over-all rate constants could not be corrected for the Hfion rate due to similarity in rates of ring opening with both catalysts. Enthalpies, entropies, and free energies of activation for ring openings were calculated. Large negative entropies ~ a t these low levels of catalysts. of activation indicated ring openings by S N mechanisms
Introduction Specific reaction rate constants for the hydrolyses of the diglycidyl ether of 1,4-butanediol (DGEBD) and of diglycidyl ether (DGE) at less than 1.0 M concentrations in the presence of 0.05 M zinc fluoroborate, Zn(BF4)2, were not available in the literature, nor were such data for the H + ion catalyzed hydrolyses of these diepoxides at pH 3.25, which was the pH value of the Zn (Bf)z-diepoxide solutions. Knowledge of these rates and their temperature coefficients was essential The Journal of Physical Chemistry
to the elucidation of certain cellulose diepoxide reactions which are catalyzed by dilute Zn(BFJ2 but not by dilute HCI. This paper presents a study of the kinetics of hydrolysis of the DGEBD and DGE when catalyzed by Zn(BF4)2 at pH 3.25 and by HC1 at pH 3.25. Rates (1) (a) Presented in part before the Division of Inorganic and Physical Chemistry at the Southeastern Regional Meeting of the American Chemical Society, Charleston, W. Va., Oct. 15-17, 1964. (b) One of the laboratories of the Southern Utilization Research and Development Division, Agricultural Research Service, U. S. Department of Agriculture.
KINETICSO F HYDROLYSES O F THE DIGLYCIDYL ETHER O F 1,4-BUTANEDIOL
of hydrolyses were measured a t temperatures ranging from 25 to 90". Enthalpies, entropies, and free energies of activation have been calculated.
Experimental Materials. The DGEBD was obtained from the Ciba Products C O . ~and was fractionally distilled. That fraction collected a t 125-127" (5.0 mm.) had an epoxide value of 0.98 equiv. of acid/100 g. (calcd., 0.99) as determined by the Durbetalti method3 and characteristic infrared absorption bands at 10.96 and 11.82 p. Infrared spectra of this diepoxide have been published previ~usly.~That portion of DGE (Shell Development Co.) boiling at 82" (5 mm.) was used. It had an epoxide value of 1.53 equiv./100 g. (calcd. 1.54) as determined by the method of Durbetaki and absorption bands characteristic of the infrared spectra reported previ~usly.~The Zn(BF4)2 was obtained from the Harshaw Chemical Co. as a 40% aqueous solution (d2541.0466 g./cc.). Procedure. For the Zn(BF4)2-catalyzed rate studies, known weights (approximately 5 g.) of the diepoxide were dissolved in 45.00 ml. of conductivity water containing 0.5626 g. of Zn(BF4)zand shaken in closed flasks immersed in an oil bath thermostatically controlled to +0.05". The pH values of the reaction mixtures, measured with a Beckman Model GS pH meter to h0.02 pH unit, were found to be constant throughout the timed reactions. The pH values, followed from 0 to 24 hr. for the reaction mixtures a t 25,40, 60, 75, and 90" were 3.25 =t 0.02. Concentrations of water and catalyst were constant throughout each reaction. For controlled rate studies in the absence of Zn(BF&, known weights (approximately 5 g.) of the diepoxide were dissolved in 45.00 ml. of conductivity water which was adjusted to a constant pH of 3.25 with HC1 a t each temperature. Aliquots were withdrawn with 1-ml. serological pipets at timed intervals and transferred to tared erlenmeyer flasks. Weighed samples were analyzed for oxirane oxygen by the method of Durbetaki3 and corrected for blank determinations, including water and catalyst in every case. The HBr reagent was standardized periodically. Treatment of Data. Cartesian graphs of the change in concentration of oxirane oxygen, T , expressed as milliequivalents per gram of solution, with reaction time for each catalyst a t each temperature showed a change in slope beyond the half-concentration point. Initially, semilogarithmic graphs of the data were plotted in attempts to detect changes in slope beyond the half-concentration point by the method of least squares. However, unlike hydrolyses of vinylcyclo-
1883
hexene dioxideJ5in which the rates of opening of the first and second oxirane rings were 50: 1 at 25" in the presence of Zn(BFb)z, the rates of opening of the two oxirane rings of DGEBD and DGE with Zn(BF& catalysis were not different enough t o allow for the determination of kl and kz by statistical evaluation of the linear semilogarithmic plots. N'evertheless, this study indicated that two linear portions were obtained with Zn(BFS2 catalysis and only one with HC1 catalysis. That is, a statistical analysis of the diepoxideHC1 data at each temperature showed that the slopes obtained when the first and second halves of the semilogarithmic plots were considered independently were not significantly different from the slope obtained by consideration of the entire line.6 Data obtained with each catalyst were also analyzed by the modified Swain method,',* which has been previously de~cribed.~ Earlier investigations of the hydrolyses of vinylcyclohexene dioxide5 and of butadiene diepoxidesg under similar conditions of Zn(BF4)2and HC1 catalysis a t like pH showed that the rates of hydrolyses of each of these diepoxides were very different in the presence of each catalyst. In fact, the over-all reaction rates for opening of both rings with Zn(BF4)2catalyst could be corrected for the rates of opening due to H f ion catalysis a t like pH for both diepoxides. In contrast, in this study, the over-all rates of ring openings of DGEBD and of DGE in the presence of Zn(BF4)2at pH 3.25 were essentially equal to the respective rates of ring openings in the presence of HC1 a t pH 3.25. Therefore, the over-all rates could not be corrected for rates of ring openings due to Hf ion catalysis. Therefore, the enthalpies, entropies, and free energies of activation for all ring openings were computed for the HC1catalyzed reactions and for the observed over-all reactions with Zn(BR)2 catalyst at pH 3.25.
Results and Discussion Figure 1 shows typical Cartesian graphs of oxirane oxygen content of DGEBD vs. reaction time a t 75" and indicates a change in rate beyond the half-concen(2) The mention of trade names and firms does not imply their endorsement by the U. s. Department of Agriculture over similar products or firms not mentioned. (3) A. J. Durbetaki, Anal. Chem., 28, 2000 (1956). (4) W. A. Patterson, ibid., 26, 823 (1954). (5) R. R. Benerito, H. M. Ziifle, R. J. Berni, and B. M. Banta, J . Phus. Chem., 67, 1750 (1963). (6) J. C. R. Li, "Introduction to Statistical Inference," Edwards Brothers, Inc., Ann Arbor, Mich., 1957, p. 344. (7) C. G. Swain, J . Am. Chem. Soc., 66, 1696 (1944). (8) A. A. Frost and R. G. Pearson, "Kinetics and Mechanism," John Wiley and Sons, Inc., New York, N. Y . , 1953, p. 158. (9) H. M. Ziifle, R. J. Berni, and R. R. Benerito, J . A p p l . PoZymeT Sci., 9 , 169 (1965).
Volume 69. *\'umber 6
June 1966
R. J. BERNI,R. R. BENERITO, AND H. M. ZIIFLE
1884
1.0,
TIME (MINS.)
Figure 1. Concentration of oxirane oxygen, T,in milliequivalents per gram of solution us. time in minutes for approximately 1 M diglycidyl ether of 1,4butanediol catalyzed by 0.05 M Z X ~ ( B F and ~ )by ~ HCl of pH 3.25 a t 75".
t
\
t
\
O.' tration point with each catalyst. Figure 2 is the .os1 I I I I I I semilogarithmic representation of these same data and 0 40 80 120 160 200 240 T I M E (MINS.1 indicates two intersecting linear segments with ZnFigure 2. Log of concentration of oxirane oxygen, T,in (BF4)z catalysis but only one straight line with HCl milliequivalents per gram of solution vs. time in minutes for catalysis. The observed change in slope beyond the approximately 1 M diglycidyl ether of 1,4butanediol half-concentration point and the linear relationships catalyzed by 0.05 M Zn(BF& and by HC1 of pH obtained upon semilogarithmic treatment of the data 3.25 at 75". indicate consecutive pseudo-first-order reaction kinetics. However, it should be noted that, with Zn(BF4)2 For diepoxide reactions, consecutive first-order recatalysis, the over-all rates of opening of the first (klJ actions may be represented as follows and second (kzz) oxirane rings were not sufficiently different to allow their determination by the usual statistical evaluation of the semilogarithmic plots.5 where A is the niolar concentration of the diepoxide, B Because of this, use of slopes of lines, as shown in is the molar concentration of monoepoxide after hyFigure 2, for evaluation of reaction rate constants drolysis of the first oxirane ring, and C is the molar would lead to erroneous results in this instance. Hence of the final product after opening of both the data were analyzed by the modified Swain m e t h ~ d , ~ , concentration ~ epoxide rings. Experbentally, C bears a direct relaand pseudo-first-order reaction kinetics was confirmed. tionship to total oxirane content, T , which is measured The fact that only one line was obtained when HC1 in this study as milliequivalents per gram of solution. catalyzed the reaction (see Figure 2) suggested equivaThe following equations, in which zero subscripts denote lent rates for the opening of the first ( k l ~ )and second initial concentrations, are for those instances, as with (kZH)oxirane rings. Subsequent analysis by the modiZn(BF4)z catalysis, where kl # kz. If kl = kz = kfI, as fied Swain meth0d~2~ showed the ratio of k2H:kLH to be unity, confirming equal rates of opening for both rings. Cartesian graphs of changes in oxirane oxygen content of DGE: lis. time at each temperature with both catalysts indicated a change in slope beyond the halfconcentration point and semilogarithmic plots of the T=2A+B (3) data were linear. Statistical analyses of the log conis indicated with HC1 catalysis, B and T can be reprecentration-time data indicated kl, and k2. were different sented by the following equations. while klH = kZH. Analysis of the data by the modified Swain method's8 confirmed the assumption of consecuB = kHtAOe-kHt (4) tive first-order kinetics and of equivalent rates of ring T = Aoe-"'(2 kRt) opening with HC1 catalysis. (5)
+
The Journal of Physical Chemistry
KINETICSO F HYDROLYSES O F THE DIGLYCIDYL ETHERO F 1,4-BUTANEDIOL
Values of klz, k~,,and relative rates k2.: kl, are given in Table I. With the exception of the ratio for DGE at go", each ratio was less than unity, indicating that the first oxirane ring opens more rapidly than the second, as was observed, but to a greater degree with vinylcyclohexene di0xide.j However, with the butadiene diepoxides, kzS:kl, was greater than unity.
Table I: Rate Constants for Opening of the First and Second Oxirane Rings in Zn( BF&Catalyzed Hydrolyses Temp., OC.
-Specific reaction rate constantski, X IO', k2n X IO', min. -1 min. -1
Relative rates k2E: ki,
25 40 60 75 90
Iliglycidyl ether of 1,4-butanediol 3.92 3.13 8.07 7.96 94.77 53.00 173.27 93.73 496.95 319.54
0.79 0.99 0.56 0.54 0.64
25 60 75 90
Diglycidyl ether 3.98 1.67 77.57 66.70 213.81 152.64 509.79 783.35
0.42 0.86 0.71 1.55
Values of kH are given in Table 11. To determine k H , use was made of the experimentally derived relationship
T = Toe-bt (6) where To is the initial concentration of oxirane oxygen, T is the concentration a t any time, t, and b is the slope of the semilogarithmic plot. The least-squares equation
Table I1 : Rate Constants for Opening of the First and Second Oxirane Rings in Dilute HC1 Catalyzed Hydrolyses
r -
Temp., 'C.
25 60 75 90
Specific reaction rate constants (klH = k,H) kH X 104,min.-' Diglycidyl ether of 1,4-butanediol Diglycidyl ether
1.95 44.10 124.88 299.83
2.64 51.32 181.30 511.89
of the semilogarithmic line a t each temperature was calculated, and the slope, b, was used to determine the specific reaction rate constant, kH, a t each temperature. The relationship between kH and b, determined by
1885
equating eq. 5 and 6 and making use of the fact that To = 2A0, is given by eq. 7. Since this relationship is
kH2t kH = b 2 (2.3026) -k 8 (2.3026)
(7)
time dependent, kH was taken as the average of values obtained a t t = 0 and t = maximum time. Comparison of kH with kl, and k2, values at the respective temperatures for each diepoxide shows them to be approximately the same order of magnitude. Figures 3 and 4 are typical of plots of the calculated concentrations of A , B , and T as functions of time with Zn(BF4)z and HC1 catalysts, respectively. For Figure 3, experimentally determined rate constants for DGEBD a t 75", kl, and k2,, were substituted in eq. 1, 2, and 3. For Figure 4, the experimentally determined
I
!I
-
0 EXPERIMENTAL
1.0
E 0.8
0.2
0
20
40
60
BO
100 120 140 160 TIME IMINS.)
180 200 220
24
Figure 3. Plot of calculated concentrations of diepoxide ( A ) , monoepoxide ( B ) ,and total oxirane oxygen ( T )as a function of time a t 75" for diglycidyl ether of 1,4butanediol catalyzed by 0.05 A ! Zn(BF& For convenience, concentrations of A are in millimoles and B and T are in milliequivalents per gram of solution. Curves are calculated on the assumption of consecutive first-order reaction; circles are experimentally determined values of T.
rate constant a t 75", kH, for DGEBD was substituted in eq. 1,4, and 5. Experimental values of T are indicated by circles. Standard errors of estimate were computed to be 0.03 and 0.02 for Zn(BF4)Z and HC1 catalysts, respectively. Figure 5 gives similar curves for A , B , and T calculated from experimentally determined rate constants for DGE at 60" with Zn(BF4)z catalysis. In this figure, curve T H is that calculated for DGE a t 60" with HC1 catalysis, and the shaded circles are the experimentally determined values with HC1 catalysis. In this figure, the latter values are included for comparative purposes and the il and B curves with HC1 catalysis are omitted for the sake of clarity. Extent Volume 69,.Vumber 6 June 1965
R. J. BERNI,R. R. BENERITO, AND H. M. ZIIFLE
1886
0.2
t 0
20
40
60
BO
100 120 140 160 TIME (MINS.1
180 200 220 240
Figure 4. Plot of calculated concentrations of diepoxide ( A ) , monoepoxide ( B ) , and total oxirane oxygen (T)as a function of time a t 75" for diglycidyl ether of l14-butanediol catalyzed by HCl a t pH 3.25. For convenience, concentrations of A are in millimoles and B and T are in milliequivalents per gram of solution. Curves are calculated on the assumption of consecutive first-order reaction in which kl = kz; circles are experimentally determined values of .'2
of agreement of experimental points (circles) with theoretical curves are given by the standard errors of estimate of 0.03 for Zn(BF& catalysis and 0.05 for HC1 catalysis. Agreement between theoretical and experimentally observed data confirmed the assumption of consecutive first-order reactions for both diepoxides at each temperature. Calculated enthalpies, entropies, and free energies of activation for hydrolyses of oxirane rings in DGEBD and in DGE with both catalysts are recorded in Table 111. Arrhenius plots of log k us. 1/T were linear. The
0'
40 ' 8 0
120
160 200
240 280 320 360 400 440 TIME WINS.)
Figure 5. Plot of calculated concentrations of diepoxide ( A ) , monoepoxide ( B ) , and total oxirane oxygen ( T ) as a function of time a t 60" for diglycidyl ether catalyzed by 0.05 111 Zn(BF&. For convenience, concentrations of A are in millimoles and B and T are in milliequivalents per gram of solution. Curves are calculated on the assumption of consecutive first-order reactions; circles are experimentally determined values of T. Curve TIIand shaded circles are for calculated and experimentally determined values, respectively, obtained a t 60" for diglycidyl ether catalyzed by HC1 a t pH 3.25.
of the enthalpy of activation, AH*. Entropies of activation, AS*, were calculated from the equation
where h: is the hydrolysis rate constant in reciprocal seconds and R, K , and h are the molar gas, Boltzniann, and Planck constants, respectively. Free energies of activation, AF*, were calculated from the equation
AF* = AH* - TAS* Table 111: Enthalpies, Entropies, and Free Energies of Activation of Oxirane Ring Openings in Zn(BF&- and HC1-Catalyzed Hydrolyses
Ring opening
Diglycidyl ether of 1 .I-butanediol Diglyoidyl ether AH*, AS*, AF*?s, AH*, AS*, AF*x, kcal./ cal./mole koal./ kcal./ cal./mole kcal./ deg. mole mole deg. mole mole
lstoxirane 2ndoxirane
Zn(BF4h 18.0 -24.8 25.4 16.1 -30.2 15.3 -33.7 25.4 20.0 -19.2
25.1 25.7
HCl 1st and 2nd oxirane
17.4 -26.6
25.4 17.4 -26.8
26.4
equation of each line was determined by the method of least squares and the slopes were used in the calculation The Journat of Physical Chemistry
480
(9)
Intercomparison of AH* values for DGEBD shows that the values for the first ring opening with Zn(BF4)z and both ring openings under HC1 catalysis are approximately the same while that for the second ring opening under Z ~ I ( B Fcatalysis ~)~ is significantly smaller. However, the more negative AS* for the second ring opening with Zn(BF&, indicative of a more ordered transitionstate complex, results in equal AF* values for all ring openings. In contrast, the AH* value for the second ring opening of DGE catalyzed by Zn(BF4)zexceeds that for the first ring opening, and there is a greater decrease in entropy of activation for the latter than for the second ring opening. These results are similar to those observed with the hydrolysis of vinylcyclohexene d i ~ x i d e . ~ With HC1 catalysis, AH* and AS* values are intermediate between those found for the first and second
KINETICS O F HYDROLYSES O F THE DIGLYCIDYL ETHERO F 1,4-BUTANEDIOL
1887
Table IV: Comparison of Specific Reaction Rate Constants" -Vinylcyclohexene
dioxide-
-Butadiene
diepoxide-
Diglycidyl ether of 1,4-butanediol
-Diglycidyl
ether--
Temp., OC.
klz/kiE
ksz/k?E
klz/hE
kdkm
klz/klE
k?a/k?E
hz/klE
25 40 60 75 90
6.6 5.0 4.2
2.0 2.4 4.5
2.5 1.6 1.5 1.6 2.3
16.5 27.7 11.4 42.4 76.9
2.0
1.6
1.5
0.6
2.1 1.4 1.7
1.2 0.8 1.1
1.5 1.2 1.0
1.3 0.8 1.5
, . .
...
...
...
...
...
...
k d k ? ~
...
a Subscripts 1 and 2 denote openings of the first and second oxirane rings; subscripts z and H denote Zn(BF& and HC1 catalyst, respectively.
ring openings with Zn(BFJ2, but more nearly equal to vinylcyclohexene dioxide; k 2 H < k l for ~ butadiene dithose for the first. The AF* values for openings of both epoxide; and ~ P H= ~ I Hfor DGEBD and DGE. rings with either catalyst are essentially equal. It is of Data in Table IV are given to show the effect of change interest to note that with HC1 catalysis, AH*,AX*, of catalyst with each diepoxide. While the pH (3.25) and AF* values for DGE are identical with those found a t which rates of DGEBD, DGE, and butadiene difor DGEBD a t the same pH, but the activation paepoxide were measured differed from that employed with rameters for these txyo diepoxides with Zn(BF4)2catalyvinylcyclohexene dioxide (2.55), certain conclusions sis differ. can nevertheless be drawn. Hydrolysis of epoxides occurs by either an S N or ~ For DGEBD and DGE, the kpz:kZH and k1,:klH an S N mechanism, ~ and by both mechanisms the reratios approximate unity and are not too different from action would follow first-order kinetics under the exk2,:k2H for vinylcyclohexene dioxide and the ratio of perimental conditions of this study. The epoxide reklz:k l H for butadiene diepoxide. Unlike these ratios, acts instantaneously with the H+ ion or Zn2+ion, but kl,:klH for vinylcyclohexene dioxide is a t least twofold in the SN1 mechanism, the rate-determining step is the greater, and kzz:kzH>> 1 for butadiene diepoxide. Of slow ring opening of the conjugate acid of the epoxide the four diepoxides under discussion, only AF* values ring to form a carbonium ion, which then combines for k l H and k2H of butadiene diepoxide and for kl, and with the anion of the water niolecule in the fast step to k2H of vinylcyclohexene dioxide differed significantly form the product. By an S N mechanism, ~ the conjufrom 25.4 kcal./nlole, and these were 26.3, 27.5, 23.0, gate acid reacts biniolecularly with the solvent moleand 30.4, respectively. These data indicate the much cule and orientation is an important factor. Experigreater effect of Zn(BF4)z as compared to HC1 a t like mentally, the hydrolyses of DGEBD and DGE folpH on the hydrolyses of vinylcyclohexene dioxide and lowed consecutive first-order reaction kinetics and obbutadiene diepoxide than on the hydrolyses of DGEBD served large negative entropies, showing the iniporand of DGE. Previously,g it was postulated that with tance of the orientation factor, were interpreted as inbutadiene diepoxide and H + ion catalysis, the hydroxyl dicative of an SN;~mechanism for the opening of both group a! to the second epoxide ring formed a fiveoxirane oxygen rings. These results are in agreement membered ring through intramolecular bonding, thus with those obtained in similar hydrolysis studies of making it more difficult for a H f ion to protonate the vinylcyclohexene dioxide6 and butadiene diepoxideg and resulting oxirane ring of l12-epoxy-3,4-butanediol when the concentration of epoxide was large with rein k2E > kl, was said to be spect to concentration of catalyst. Recently, an S N ~ due possibly to coordination of Zn2+ions with either the mechanism was also indicated by the negative activatwo oxirane rings of butadiene diepoxide or between tion volume AV*, for the hydrolysis of epichlorohydrin the hydroxyl group r~ to the ring oxygen and the oxirane of approximately 0.5 JI in presence of 0.0635 11 peroxygen of 1,2-epoxy-3,4-butanediol.With either DGchloric acid.10 EBD or DGE, there were apparently little differences A comparison of relative rates of the first and second in rates or activation parameters with either catalyst, ring openings with HCl and Zn(BF4)2 catalysts for four or, in other words, presence of Zn2f ions did not significantly increase rates of ring openings over those obdiepoxides was made. Only with butadiene diepoxide was kz, > kl,; with vinylcyclohexene dioxide, kp,