FREERADICAL REACTIONS INITIATED R Y IONIZING RADIATIONS
March ,5, 1962
[CONTRIBUTIOS FROM THE
RADIATION LABORATORY, CONTINENTAL
719
O I L COMPANY, PONCA CITY, OKLAHOMA]
Free Radical Reactions Initiated by Ionizing Radiations. I. Arrhenius Parameters for the Reactions of Hydrogen Atoms with Propane, Ethylene and Propylene BY KANGYANG RECEIVED AUGUST9, 1961 The presence of ethylene or propylene in the gamma radiolysis of propane reduces the rate of formation of hydrogen. This can be explained T h e extent of this inhibition depends both on radiolysis temperature and olefin concentration. quantitatively b y assuming the occurrence of the following reactions:
k,
+ C3H8 +H
+
k,
k,
H C2HI +C2Hs H CaHs +C3Hi p f CsH7 = 1.1 exp (-4.4 X lO3/RT); k,/k,, = 1.3 exp (-4.8 X lO3/RT); and k , / k , = 1.3 exp (-0.4 X lO3/RT) are determined based on this radical mechanism.
H The relative rates k , / k ,
ence of the rate on the scavenger concentration is Introduction In the present series of investigations we attempt in agreement with (E-2), and, second, the resulting to determine the rate constants of the reactions in- ratio of the rate constants is reasonable. Another volving hydrogen atoms produced by the gas phase test is to investigate the dependence of the ratio on radiolysis of saturated hydrocarbons. Quantitative temperature. Since rate constants for gas phase treatment of many radiolysis mechanisms demands reactions are usually well represented by the hrrone should such information, but presently available experi- henius equation, k = ;2 exp (-E* ‘RT), mental results seem to be limited and ~onflicting.’-~ obtain a straight line when In ( k m l k s ) is plotted The method to be employed here was first de- against 1/T. The present experiments confirm veloped by Back4 and Hardwicks and utilizes the this prediction. -1s the compound 11 to be radiolyzed, propane experimental observations that the presence of a scavenger, S, in the radiolysis of a compound PIT, was chosen. I t is the simplest commercially availoften diminishes the rate of formation of hydrogen able saturated hydrocarbon that does not require elaborate purification for the present purpose. M d e - H + prod. Ethylene and propylene were selected as scavengers. M -+ Hp prod. The main reason for this choice is that the relative rates of the two reactions km H + M +H2 + R (3)
+
k, H+S-+R’ k Is H S +H?
+
(4)
+ R”
(5)
The steady state approximation gives a relation
Here, ro and rs denote, respectively, the rates of formation of hydrogen molecules in the absence and presence of scavenger, while rl represents the rate of formation of hydrogen atoms; and the k’s are the respective rate constants. Hardwicks rearranged (E-1) to the form
and determined the rate constant ratio for various liquid phase reactions by plotting (rc-rs) - l against [MI/ [SI. By using essentially the same equation, Back4 and Yang and Gant7 investigated some hydrogen atom reactions in the gas phase radiolysis of saturated hydrocarbons. At present the support of the mechanism, (1) to ( 5 ) , comes from two s o ~ i r c e s . ~First, . ~ ~ ~the depend(1) E. W. R . Steacie, “Atomic and Free Radical Reactions,” Reinhold Publishing Corporation, New York, N . Y., 1954. ( 2 ) A. F. Trotman-Dickenson, “Gas Kinetics,” Butterworths Scientific Publications, London, 1955. (3) S. W. Benson, “ T h e Foundations of Chemical Kinetics,” McGraw-Hill Book Company, Xew York, N. Y.,1960. (4) R . A. Back, J . P h y s . Chenz., 6 4 , 124 (1960). ( 5 ) T. J. Hardwick, ibid., 66, 101 (1961). (6) For example see: L. M . Dorfman, ibid., 62, 29 (1968). (7) K. Y a n g a n d P. .I G a n t , ibid., (in press).
are of kinetic interest but available information is conflicting.8-10 The determination of ks’/k, using (E-2) requires the estimation of intercepts. Because of this, the resulting ratio was inaccurate; however, this ratio did not exceed 0.08 in the present experiment. If we neglect k,’;’k, compared with unity, (E-2) may be rearranged to give
where r2 ( = ro-rl) denotes the rate of reaction 2. In this form, the extrapolation of rate data to infinite concentration of scavenger (for the estimation of r l ) is not required in obtaining the desired ratio, kmlks.
In the present experiment, both n- and iso-propyl radicals are likely to form in reactions 3 and 7. K i t h the method we employed, i t was not possible to distinguish this situation. Hence, km and k, reported here are composite rate constants. Experimental Phillips’ research grade propane was degassed by repeated evacuation at liquid nitrogen temperature, then subjected (8) P. E . Allen, H. W. Melville and J. C . Robbs, Pvoc. R o y . SOC. ( L o n d o n ) , Alas, 311 (19.53). (9) B . d e B. Darwent and R . Roberts, Disc?ission Faraday SOC.,14, 5 5 (1953). (10) A. B. Callear and J. C. Robbs, Trans. Favaday SOL.,61, 638 (1955).
KANGYANG
720
to a three-stage bulb-to-bulb distillation, each time retaining the middle third. Phillips' research grade ethylene and propylene were degassed similarly but subjected to the distillation only once. Reaction vessels (Pyrex, 140 cc.) equipped with capillary constrictions a n d break-off seals were charged with 50 cm. H g of propane and varying amounts of scavengers ( a t room temperature). A thermocouple-controlled furnace completely enclosed t h e reaction vessel. The vessel used to measure reaction temperature was the same a s the one used in obtaining rate d a t a except t h a t the former was equipped with a thermocouple well. During a single experiment lasting u p to 40 minutes, temperature did not fluctuate more than f 2 ' . About 1 X 10' curies of Coco were used to initiate reaction. Energy input rate, measured by using a n ethylene dosimeter,7 was 1.4 X 10-3 e.v. per hour per molecule of propane. After irradiation, samples were connected directly to gas chromafographic equipment through break-off seals. A silica gel column at room temperature with argon carrier was used in the determination of hydrogen concentrations (10-810-2 per cent). Other details of the experimental procedures have been described p r e ~ i o u s l y . ~
Results Table I summarizes typical experimental results showing decrease in the rate of hydrogen formations with increasing scavenger concentration. In all experiments, the ratio, [H2]/[C3Hs], did not exceed 4.5 X The maximum fraction of propane decomposed is probably the same order of magnitude. TABLE I HYDROGEN FORMATION IN THE RADIOLYSIS OF AT 201" C3H6 SYSTEM IC8H6]/[CaHsl5 Irrad. time LHzI/[CaHal
x
102
x
(min.) b
THE
106
CaHe-
The data in Table I were in good agreement with (E-3). From least square treatment, we obtained ..
+ (2.3 z!= 0.1) X
(E-4)
Deviations shown in (E-4) (and in subsequent equations) are standaw deviations. The rate constant ratios obtained from similar experiments a t various temperatures are summarized in Table 11. Arrhenius plots obtained by using the data in Table I1 were good straight lines. The equations resulting from least square treatment were : km ke
log - = (0.06 zk 0.07) logh
ke
= (0.15 f 0.05)
1 - (4400 & 100) -4.58T 1
- (4800 f
k k,
=
(0.09 f 0.09)
- (400
+
Temp.
+
( k d k d X 101
1.4 1.5 3.5 4.9 5.6 10.0 9.5 12.3 14.2 15.5
Temp.
(OK.)
330 3j2 403 474 483 490
( k m / K p ) X 101
0.96 1.4 3.6 8.7 8.9 10.5
employed G values (molecules/100 e.v.) may be obtained from the relation: G = rate X 100/E. The resulting G values for 70 and r2 were, respectively, 6.4 and 1.6. These values may be compared with Back's data,4 obtained by investigating propane radiolysis (among others) a t very low conversion and a t ambient temperature. His relative rates (when recalculated to give G(H2) of 1.2 in ethylene radiolysis) give 6.3 and 1.6 for G(H2) corresponding to ro and r2 in good agreement with our results. The present result, k e / k p = 1.3 exp (-400/RT), shows that, because of the lower activation energy, hydrogen atoms attack propylene more readily than This ethylene a t temperatures lower than -520'. is in disagreement with the result of Darwent and Robertsjgwho investigated the photolysis of H2S in the presence of ethylene or propylene a t two different temperatures and reported that k e / k p . = 0.1 exp (+0.9 X 103/XT). The source of this discrepancy is not apparent. From (E-7), we estimate ke/kp = 0.7 a t 25'. This is compatible with the result of thermal methodlo (= 0.3 a t 31°), which employs platinum wire to measure heat liberated by atomic recombinations but not with the result of molybdenum oxide methods (=3.1 a t 18'). For this reason, the rate constant, k , = 4.8 x 1011 mole-' cc. set.-' a t 31' obtained by the thermal method, is used below to estimate absolute Arrhenius parameters. The present method gives only relative rate constants. To obtain individual rate constants from these results, some additional information is needed. We have taken a recent result," E* = 1.5 kcal./
(E-6)
TABLE I11 ARRHENIUSPARAMETERS, A MOLE-^ cc. SEC.-') AND E* (KCAL./MOLE) FOR SOME HYDROGEN ATOMREACTIONS YLogA-E-This
1
f 100) __ 4.58T
(OK.)
329 334 383 403 415 450 474 485 503 513
(E-5)
Hence log 2
TABLE I1 THE RELATIVERATE CONSTANTSFOR THE REACTIONS, km ke k* H 4- CJHBJ, H C2H4-+-, AND H C3K7+, AT VARIOUSTEMPERATURES
Discussion RateC X 106
0.00 30.0 4.48 8.96 1.19 40.0 3.52 5.28 2.38 40.0 2.80 4.20 4.72 40.0 2.32 3.48 5.91 40.0 2.16 3.2-I Propane pressure at room temperature = 500-501 mm. Energy input rate = 1.4 X e.v. per hour per molecule of propane. [H2]/[C3H8]/hr.
= (9.5 f 0.2) X 10-3-1~1 [MI
Vol. 54
(E-7)
Unscavenged rate 70 and completely scavenged rate r2 are of importance in discussing radiolysis mechanisms.6 Neither of these rates showed any temperature dependence (56-240'). Since the rate is expressed in [Hz]/[CZHs] per hour and energy in-
work
Reactions
H
4-C2H5
Lit."
This
work
C*H, 1 2 . 9 13.5 1.9 1 2 . 8 14.4 1.5 H C3Hs CIHT H C3H8+ Hz C3H7 13.0 (14.5)* 6 . 3 b These values 0 Darwent and Roberts, ref. 9. the reaction cf D atoms with propane.
+
+
+
+
+
Lko
4.1 5.0 (7.2)h are for
March 5 , 1962
CONDUCTANCE OF MANGANESE SULFATE IN DIOXANE-WATER MIXTURES
mole for the H-CBHGreaction, and adjusted the A factor for this reaction to give K , = 4.8 x 10" mole-' CC. sec.-l a t 31'. The result thus obtained was used to estimate the Arrhenius parameters summarized in Table 111,where the results in reference 9 are also included for comparison. The present
[CONTRIBUTION FROM THE
DEPARTMENT OF CHEMISTRY,
72 1
values for both A and E are lower in all three reactions. The source of this discrepancy is not clear. Acknowledgments.--Mr. J, D. Reedy helped with experimental work. Dr, F. H. Dickey and Dr. L. 0. Morgan offered valuable discussions and suggestions.
THE UNIVERSITY OF
MICHIGAN, ANNARBOR,MICHIGAN]
Ion Association in Polyvalent Symmetrical Electrolytes. IV. The Conductance of MnSO, and MnBDS in Dioxane-Water Mixtures B Y GORDON ATKINSON' AND CALVIN J. H.4LLADq2 RECEIVED AUGUST4, 1961 The conductance of M n m-benzene disulfonate (MnBDS) has been measured in mixtures from 0 t o 407, dioxane; and MnS04 has been studied from 0 t o 257, dioxane. T h e data are analyzed using the Fuoss-Onsager theory and K A , A0 and U J parameters obtained. As was found in the previous work on these salts in methanol-water mixtures, MnBDS is far less associated in water than MnS04. CZJ and the Walden product (A"?) are constant for both salts over the solvent range examined and the same as found in the methanol-water mixtures. However, the K A ' Sfound in the dioxane-water mixtures are uniformly lower than those found in the methanol-water mixtures of the same dielectric constant. It also is found t h a t the log K A v s , l / D plot for only MnSOa in dioxane-water mixtures is in very good agreement with the Fuoss-Bjerrum theory of ion-pairing. Various speculations on the solvation of the ions involved are advanced t o tentatively explain these results.
Recent work on the conductance of high-charge Experimental electrolyte^^-^ has demonstrated the existence of The preparation, qurification and analysis of the Mn(I1) 2-2 salts that are only very slightly associated in salts have been described previously4 as have the apparatus and modus ~ p e r a n d i . ~ A11 solutions were made up by water. These substances, the Cu(I1) and Mn(I1) weight in flask-type conductance cells. Pycnometric salts of rn-benzene disulfonic acid and the Cu(I1) densities were used for conversion to the molarity scale. salt of 4,4'-biphenyldisulfonic acid follow the Fuoss- The dioxane was purified by the standard procedurese; Onsager theoretical predictions exactly and yield and the physical constants of the dioxane-water mixtures in the calculations were those given by Fuoss.' Hyvery reasonable values of the conductance param- used drolysis corrections were not used because their calculated eters ho and Uj. One of these salts MnBDS has size fell within the experimental errors of A determination. been examined in methanol-water mixtures from Data Treatment.-Table I gives the equivalent conduct0 to 100% methanol; and as a comparison, MnS04 ance and concentration data for the two salts in the various mixtures. The phoreograms of the two salts in the was examined in the same solvent mixtures. I t solvent different solvent mixtures are so similar to those shown was found that in the range 0 to 0.25 mole fraction previously4 as to make their portrayal superfluous. methanol both salts exhibited a constant value of For the treatment of the data the Fuoss-Onsager equation a, (the ion size parameter) and the Walden product, in the form for associated electrolytes was used. We hov. I n addition, the plots of log KA as. 1 / D again have ignored the viscosity correction which has been proposed by Fuoss since its contribution to 2-2 salt conductwere linear for both salts in this range. Un- ances is unknown. fortunately, solubility problems prevented the A4 = A' - S( Ca)'Jz E( C a ) log ( Ca) 4extension of the measurements much past this J ( C a ) - K.4(Ca\j,*.\ (1) range for MnS04. For MnBDS, where the measwhere the symbols are explained in references 4 and 7, urements could be made to 100% methanol, it was (particularly 4, for 2-2 salt treatment). found that U J and hov deviated strongly from the For the actual analysis of the association we have adopted low methanol concentration values when X M ~ O H the Fuoss "y-x" method and have used the Debye-Huckel was greater than 0.23. I t also was shown that the extended equation to calculate the activity coefficients, The calculation has been programmed for the I B M association constants in this high methanol con- j+. 704 computer so t h a t maximum utilization of data can be tent range were not consistent with the low range achieved. Fuoss defines values. All of these deviations were tentatively A' = A f SCi1/2 - ECi log Ci (2) correlated with the extreme deviations of the AA\ = A' - 1 ' (3) methanol-water solvent mixtures from ideality. In an attempt to further clarify this problem y = A.\/C; (C;= a C ) (4) and to further determine the association of these and x = f,zA salts, measurements have been made in dioxanewater mixtures covering the same dielectric con- Then ? J - KAX stant range.
+
(1) D e p a r t m e n t of Chemistry, University of M a r y l a n d , College P a r k , Maryland. (2) Union Carbide h-uclear Corporation, Tuxedo, Xew York. (3) G . Atkinson, M. Yokoi a n d C . J. Hallada, J. Am. Chem. SOL, 83, 1570 (1961). ( 4 ) C. J. Hallada and G. Atkinson, ibiii., 83, 3769 (1961). ( 5 ) M. I'ukoi a n d G. Atkinson, i b i d . , 8 3 , 4 3 6 i (1961).
T h e computer first evaluates a n approximate A a1 =
AO -
CY
from
s(A/A~)v~C'/~
(6) C. A . K r a u s and R. A . Vingee, ibid., 66, 513 (1934). (7) R. M. Fuoss a n d F. Accascina, "Electrolytic Conductance Interscience Publishers, Inc., S e w York, N. Y . , 1959.
"
