REACTIONS OF CARBONYLIC COMPOUNDS WITH HYDRATED ELECTRONS
(see Table IV) leaves the ratio k l / k z unchanged so that the stabilization of the hot ‘Az state is not the ratedet,ermining step. What happens to the system in this state is not clear, but crossing into the 3Azstate seems probable. In t,his state the molecule could either relax to the ground state or dissociate to triplet methylene and carbon monoxide. The net result of this latter process would be indistinguishable from reaction 5. Some Of the andconclusions Of this paper have already been published. 2o
3993
Acknowledgments. The authors wish to thank the Research Grants Branch, National Center for Air Pollution Control, Bureau of Disease Prevention and Environmental Control, for the funds which made this work possible.
(20) B. A. DeGraff and G. B. Kistiakowsky, J . Phys. Chem., 71, 1553 (1967).
Reactions of Carbonylic Compounds with Hydrated Electrons’
by Edwin J. Hart, E. M. Fielden,z and M. Anbar3 Chemistry Diaision, Argonne National Laboratory, Argonne, Illinois 60459 (Received M a y $2, 1967)
The reactivity of the hydrated electron with carbonyl compounds of the type RlCOR2 has been measured. Their rate constants, kl, vary from 3.0 X lo5 M-’ sec-l for urea to 8.0 X lo9 M-’ sec-l for cyclohexanone. For aldehydes, ketones, and carboxylic acids a linear correlation is obtained between log k and Taft’s u* function. The slope, p , equals -0.74. Amides and esters deviate from this relation because of mesomeric forms, in which the electrophilic center is localized on either the nitrogen or the alkoxy oxygen. For a group of these compounds, a p of + l . l S is found. Hydrated electron rate constants, as demonstrated by these and previous studies, provide a new parameter for exploring the electronic configuration of organic compounds.
Introduction
Experimental Section
The hydrated electron, e,,-, has been suggested as a new nucleophilic reactant for studying the structure of organic molecules and ions in aqueous solution^.^-^ I n many reactions e,,- transfers directly to an electrophilic center. Halogen atoms, for example, act as such centers on aliphatic coupound~.~Another functional group displaying appreciable reactivity toward ea,is the carbonyl group.6 Since no systematic studies on these compounds have been carried out we supplement previous r e p o r t ~ ~ J ~by ’ - ~providing rate constants for a series of organic compounds, characterized by the general formula R1COR2. R1 and Rz are H, alkyl, alkoxy, NHz, or OH (0-)groups.
The rate constants were measured by an electronpulse technique similar to that described in earlier work.l0I1l The set-up appears in Figure 1. It differs (1) Based on work performed under the auspices of the U.S.Atomic Energy Commission. (2) Physics Department, Institute of Cancer Research, Belmont, Sutton, Surrey, England. (3) Weirmann Institute of Science, Rehovoth, Israel. (4) M. Anbar and E. J. Hart, J . Am. Chem. SOC.,86,5613 (1964). (5) M.Anbar and E. J. Hart, J . Phys. Chem., 69,271 (1965). (6) M.. Anbar, Advances in Chemistry Series, No. 50, American Chemical Society, Washington, D. C., 1965,p 55. (7) E.J. Hart, J. K. Thomas, and S.Gordon, Radiation Rea. Suppl., 4, 74 (1964).
Volume 71, Number i.2 November 1967
E. J. HART,E. M. FIELDEN, AND M. ANBAR
3994
d
C
A
D
record the transient absorption due to the radiationinduced hydrated electron was identical with that described previously.1Q
Preparation of Solutions
-E
Figure 1. Electron pulse radiolysis set-up using a gas laser light source: A, helium neon laser with continuous output a t 6328 A ; l3F4 X 2 cm diameter silica irradiation cell with silv&ed Suprasil end windows; C, plane mirror; D, Amperex Type XP1002 photomultiplier; E, 2-in. thick lead pot with 0.25-in. diameter hole in lid; F, top view of optical path in cell.
from previous arrangements in the use of a gas laser (A) as a light source. Benefits gained by its use are (1) a major simplification in the optical system, and (2) complete suppression of “Cerenkov noise’’ during the pulse. These advantages are offset, to some extent, by the inherent instability of the laxer and its 60-cycle ffuctuations in intensity. For microsecond sweep rates for which most of our work was done, these instabilities posed no problems. The laser was a Perkin-Elmer Model 5200 D.C. excited helium neon laser, giving a continuous output a t 6328. A. The laser emits a parallel beam of light approximately 0.5 mwatt power. The irradiation cell (B) consisted of a quartz tube 4 cm long and 2 cm in diameter with optically polished highpurity quartz end windows. The windows of the cell were completely silvered with the exception of two small diametrically opposed areas (little larger than the laser beam diameter of 3 mm) on the front face of the cell. These unsilvered portions served as entrance and exit windows for the laser beam. By adjusting the angle a t which the laser beam entered the cell, the beam was reflected back and forth (F) between the mirrored ends of the cell before emerging. Between 2 and 10 passes of the beam along the length of the cell were readily obtained in this manner. The beam emerging from the cell was reflected by a plane mirror (C) onto a photomultiplier @) (Amperex Type XP1002) encased in a 2-in. thick lead shield (E). The monochromatic and parallel nature of the light eliminated the monochromater and quartz lens assembly normally required.’O The system of electronic amplifiers and oscilloscope used to The Journal of Physicel Chembtrv
The compounds were irradiated in unbuffered solutions, and in solutions buffered a t pH 7.2 with NaH2POsNa2HP04 and a t pH 11 with NaOH. I n the solutions where 1.0 mM ethanol was used to destroy hydroxyl radicals, the ethanol was purified by distillation through a fractionation column. The solutions were injecteds into the deaerated matrix” to produce a solution of the desired solute concentration. I n this procedure, 5 ml of the stock solution (0.01-0.05 M ) or the pure organic liquid was first deaerated by shaking with an equal volume of helium in a 10-ml syringe. After 2 to 3 min shaking, the gas was expelled and 5 cc of helium again added and the shaking repeated. By chromatographic analysis we found that each shaking reduced the oxygen concentration of the solution by a factor of 20. Consequently, the concentration of oxygen in our stock solution was reduced to about 0.5 p M after two equilibrations. Normally of the order of 0.25 ml of this solution was then injected into 50 ml of the preirradiated hydrogen saturated matrix. Previous studies11 showed that the concentration of oxygen in these solutions was less than lo-’ M .
Analysis of Decay Curves The rate constants were calculated directly from a first-order plot of the logarithm of optical density us. time. The oscillograms were analyzed by CHLOE, an automatic film scanning machine developed by the Applied Mathematics Division of Argonne.I2 CHLOE, with a maximum resolution of 1024 “bits” in both the X and Y coordinates scans an oscillogram photograph and stores the data in digital form. These data were then fed into a CDC 3600 computer which printed out the optical density (or any similar function such a log OD) os. time. A graph of log OD us. time with the “least mean squares” slope was also produced automatically. This method not only saved time but it also eliminated personal errors in the measurements of oscillograms and in calculating the errors involved in such reading. Figure 2 shows a typical decay curve (8) E. J. Hart, S. Gordon, and J. K. Thomas, J. Phys. Chem., 68, 71 (1964). (9) A. Szutka, J. K. Thomas, S. Gordon, and E. J. Hart, ibid., 69, 289 (1965). (10) J. K.Thomas, S. Gordon, and E. J. Hart, ibid., 68, 1524 (1964). (11) E.J. Hart and E. M. Fielden, ref 6,p 253. (12) D.Hodges, Applied ,Mathematics Division Technical Memorandum No. 61,1963.
REACTIONS OF CARBONYLIC COMPOUNDS WITH HYDRATED ELECTRONS
3995
Hydrated Electron Rate Constants for Some Organic Compounds
Table I:
Compound
CHrCOCN
S0"Ne
Matrix
C
HzO
H*O ILO CHiCHNOH
d
(CHWNOH
e
Alk Alk Alk
II*O
KO
f
FCH,COCH,
C00-(CH,)%C00 COO -(CH,XCOOH CF,COClL NHCOCOOCFICOOCH, OHC1ItCOOCHi C14C112COOCH~ (c11,),ccoocII, CIlrCOOC*Hs CH,COCN NII~CIItCOOCH~ CNCH,COOGH,
B B
d h d
I d e
b d C C
H*O
€LO Alk Alk Alk ILO I1,O Alk Alk Alk Alk
H,0 KO Il*O Alk Alk Alk Alk
PH
7.15 7.15 7.15 10.82 10.82 10.82 7.75 7.7s 6.7 6.7 10.86 10.86 10.0 6.0 5.19 11.2 11.04 10.62 10.65 6.R1 5.91 6.53 10.94
10.66 10.66 10.92
Concn. UIM
SlOPa~
38.3 9.56 9.56 6.54 6.54 6.54 0.25 0.25 0.500 0.150 0.500 0.250 10.0 2.0 1.00 0.40 0.133 0.406 0.519 4.16 15.33 20.45 5.60 1 .OO 1 .00 0.906
0.4839 0.1461 0.1288 0.2050 0.1981 0.20R6 0.0337 0.0310 0.2255 0.0685 0.1755 0.1025 0.133 0.100 0,0289 0.6.510 0.2442 0.3277 0.1092 0.1U43 0.1511 0.5217 0.3659 0.126R 0.1259 0,1259
-kw--
2.9 x 3.5 x 2.7 x 7.2 x 7.0 X 7.3 x 3.1 X
10''
107, 10' 10' 10'r 10' 1 0
3.0 X IO' 7.2 x 107 3.0 X 1 0 1
.o x
10'
8.R X lo"
4.0 X 10'
2.91 x 1 0
Log OD/w3ec (corrections have been made for the standard matrix and for effect of p H on the matrix). BDH Analar. Fluka pnrisq. d Fluka purum. ' BDH Lab. K & K Laboratories, Inc. Baker C P ; recrystallized three times. G. F. Smith Chem. co.
'
'
and its graphical printout of the optical density data. The rate constant wm calculated from the difference between the slopes of the solute and matrix under identical irradiation conditions. The complete equation is
k
= 2.3 X
IO' X
Dog (ODlpsec) solute
-1.05
- log (OD/psec) matrix] C
where OD = optical density, psec = unit of time of decay cuwe, C = concentration in moles/l., k = rate constant in M-I see-'. Data obtained from the compounds studied appear in Table I. The estimated error is 10%.
-1.15
I -1.25
-,.35
1
*
Results and Discussion 0.5
1.0
1.5 2 0 2.S TIME in MICROSECONDS
3.0
Figure 2. Reaction of 0.001 Af methyl glycinate with points on decay curve calculated by CHME and presented on a ealwmp plot; ----, leastsquare curve. Slope = 0.1346 lag OD/-c.
e.,- at p H 10.66:
+,
3.5
The carbonyl group is highly reactive with e,Q- but its reactivity is influenced to a major extent by the nature of adjoining groups. For example, the rate constants decrease by a factor of 20,oOOwhen the methyl groups of aeetone are replaced by the amine groups of urea. In an attempt to explain this behavior we studied
3996
E. J. HART,E. M. FIELDEN, AND M. ANBAR
Table 11: The Rates of Reaction of Carbonylic Compounds (R1COR2)with Hydrated Electrons
k,
Rz
Compound
H
Acetaldehyde Acetone Fluoroacetone Trifluoroacetone Acetic acid Pyruvate ion Acetamide Pyruvonitrile N-Acetylglycinate ion Cyclohexanone Formic acid Formamide Pivalic acid Succinic acid Oxamate ion Urea Ethyl acetate Methyl propionate Methyl glycolate Methyl glyoinate Methyl fluoroacetate Ethyl cyanoacetate Methyl trifluoroacetate Methyl pivdate Formoxime Acetoxime Glycine Alanine a
CH3 FCHz F3C OH coo “2
CN HNCHzCOO -CHz(CHzh OH “2
(CCHdaC -CHzCOO NH2 CHa CHaCHz HOCHz HzNCHz FCHz NCCHz F3C (CHs)aC Ha CHP HzNCHz HaC(NHz)CH
coo
S*Ri
M-1 sec-1
-
3.5 x 6.3 x 9.5 x 6.6 X 1.8 x 6.8 X 1.7 x 3.0 x 2.0 x 8.0 x 1.4 X 4.2 x 9.7 x 1.2 x 4.0 x 3.0 X 5.7 x 9.0 x 4.8 X 2.9 X 9.8 x 3.2 X 1.9 x 2.3 x 7.2 x 3.0 X 8.0 x 5.9 x
109 109 108 lo7
10s log 107 107 107 109
IO8 107 107 108 109 lo6 107 107 lo8 lo8 10s lo8 109 107 107 lo8 106 106
Log k
O*Ri
9.54 9.80 8.98 7.82 8.26 9.83 7.23 7.48 7.30 9.90 8.15 7.62 7.99 8.08 9.60 5.48 7.76 7.96 8.68 8.46 8.99 8.51 9.28 7.36 7.86 8.48 6.91 6.77
0 0
0 0 0 0 0 0
0 -0.1 0.49 0.49 1.55 1.55 1.2 1.2 1.35 1.45 1.45 1.45 1.45 1.35 1.45 1.45 0
0 1.55 1.55
-k
O*R2
0.49 0 1.10 2.58 1.55 0 1.2 3.6 1.1 -0.1 1.55 1.2 -0.30 0 0.0 1.2
0 -0.10 0.55 0.43 1.10 1.30 2.58 -0.30 0.49 0 0.43 0.33
0.49
0 1.10 2.58 1.55
0 1.2 3.6 1.1 -0.2 2.04 1.69 1.25 1.55 1.2 2.4 1.35 1.35 2.00 1.88 2.55 2.65 3.03 1.15 0.49 0 1.98 1.88
Oxime.
t~heeffect of the substituents on the CO group. Table I lists our data for the new compounds and Table I1 summarizes the available rate constants and Taft u* values for these and other carbonyl corn pound^.'^ The compounds considered include aldehydes, ketones, carboxylic acids, amides, esters, and oximes. Taft’s u* values14 again provide us with a suitable parameter for correlating our rate constants of carbonyl compounds. We had previously found them suitable for explaining the behavior of haloaliphatic and aromatic compounds. As the u* values are expected to be a d d i t i ~ e , ’ ~we ? ’ ~list the sum of the polar substituent, constants u* of R1and R2 in Table I1 and plot them 21s. log k(e,,R1COR2) in Figures 3 and 4. We had to estimate u* values for the COO- and NH2 groups. For COO-, u* was assumed equal to that of CH, ( u * c o ~ -= 0.0). This is justified since the values of up, urn, gp+, and urn+for COO- are all very close to zero.14 I n the case of “2, u* = 1.2was estimated on the basis of the polarity of the C-N bond manifested by its dipole ~noment.’~Thus u*NH2 is comparable with
+
The Journal of Physical Chemistry
that of
U*FC!H~
and is intermediate between
U*H
and
U*OH.
The linear dependence of log k on u* illustrated in Figure 3 describes the behavior of aldehydes, ketones, and carboxylic acids. This curve has a slope p , equal to -0.74. The rate constants of substituted carbonyl compounds decrease linearily under the influence of electron-withdrawing substituents by a factor of two hundred from acetone to pyruvonitrile. I n order to explain this result we assume that eaq- adds to an orbital of the carbonyl oxygen at,om. This assumption is plausible since an ion radical, CH&(O-)CH3, has been identified in the case of a ~ e t o n e . ’ ~Electron,~~
(13) M. Anbar and P. Neta, Intern. J. Appl. Radiation Isotopes, 18, 493 (1967).
(14) R. W. Taft in “Steric Effects in Organic Chemistry,” M. S. Newman, Ed., John Wiley and Sons, Inc., New York, N. Y., 1956, p 556. (15) J. E.
Lemer and E. Grunwald, “Rates and Equilibria of Organic Reactions,” John Wiley and Sons, Inc., New York, N. Y., 1963, p 219.
REACTIONS OF CARBONYLIC COMPOUNDS WITH HYDRATED ELECTRONS
parable p value. Their reactivity is lower than that of ketones because of the inductive effect of the OH group on the nitrogen atom. This effect lowers the dipole moment from 1.3 for propylamine to 0.9 Debye units for acetoxime. By contrast the dipole moments increase from 1.6 D. for 2-propanol to 2.9 for acetone.20 Amides and esters exhibit a completely different log k us. u* free-energy relationship from that of substituted ketones and acids. Note especially that the slopes of the lines of Figures 3 and 4 change from plus to minus. The retarding effect of the NH2 or OR groups on the reactivity of the carbonyl group with eaq- is more extensive than expected from their inductive effects. We attribute this behavior to the mesomeric effect of these substituents, an effect that can also be inferred . ~ mesomeric ~ forms from the change in Y ~ , ~ The
\
-
I
I
0
1.0
I 2.0 : u
3997
I 3.0
R 4.0
+ 0:
R
\
\
C-O-
Figure 3. Relationship between hydrated electron rate constants of carbonyl compounds and Taft’s u* function.
and
//
c-o-
//
Hz +N
RO+
depress the double-bond character of the C=O group, make the carbonyl bond nonreactive toward eaq- and create new electrophilic centers of lower reactivity. The mesomeric form of esters
Ri
\ C-O-
//
RzO +
’ c c:o
7.0
I .o
N
** 2.0
*
=I
*
3.0
+ Q2
Figure 4. Relationship between hydrated electron rate constants of esters and amides and Taft’s u* function.
withdrawing groups induce a shortening of the C=O bond, a result inferred from the increase in the stretching frequency, vcp.o.18 This bond shortening, which is directly related with 0 * , l S implies a higher electron density in the T orbitals, resulting in a decreased tendency to accommodate an additional e l e ~ t r o n . ~ Oximes behave like carbonyls and they have a com-
while nonreactive at the carbonyl bond still has a finite electron affinity a t the alkoxy oxygen. This electron affinity will obviously increase with the electron-withdrawing capability of R1. The higher reactivity of esters toward eaq- correlates with an increase in u* of R1. I n Figure 4 note that there is a hundredfold increase in reactivity between (CH&CCOOCH, (*. = 1.15) and CFaCOOCHa (u* = 3.03). Because of the mesomeric character of these molecules the reactive center moves from the carbonyl to the alkoxy group. (16) S. Gordon, E. J. Hart, and J. K. Thomas, J . Phys. Chem., 68, 1262 (1964). (17) L. Kevan, P. N. Moorthy, and J. J. Weiss, J. Am. Chem. Soc., 86, 771 (1964). (18) R. N. Jones and C. Sandorfy in ”Chemical Applications of Spectroscopy, Techniques in Organic Chemistry,” Vol. IX, Interscience Publishers, New York, N. Y., 1956,p 443. (19) L. N. Ferguson, “The Modern Structure Theory of Organic Chemistry,” Prentice-Hall, Englewood Cliffs, N. J., 1963, Chapter
V. (20) A. L. McClellan, “Tables of Experimental Dipole Moments,” Freeman and Co., San Francisco, Calif., 1963.
Volume 71, Number 1% November 1967
E. J. HART,E. 31. FIELDEN, AND M.ANBAR
3998
The reactivities of formamide and acetamide also seem to be due to the
R
\
c-o-
// H2N +
mesomeric form and they fit satisfactorily on the same log k-u* curve as the esters. The unusually low reactivity of urea and the unusually high reactivity of the oxamate ion deserve special consideration. Urea may be represented by the formula U+
HZS
\
c-o-
HzN
//
U+
without any well-defined electrophilic center. Consequently low reactivity is expected. Furthermore, in the 0.10 M urea used, this compound may be partly in a dimeric, less-reactive form. Stabilization of the mesomeric form of the oxamate ion by hydrogen bonding may account for the high reactivity of this compound. The oxamate ion may exist in the forms
0
0
\
\ / /
c-c
/
N H H
-0
0
\ / /
A
c-c
/I
0-
\
N+
/\
H
0/’ H
-
-0
0
\ / /
A
c-c
/
\
0
N+
/ ‘,, /
H
H
Hydrogen bonding stabilizes the carbonyl group on the carboxylate ion thereby providing the oxamate ion with two highly reactive electrophilic centers. This assumption explains the high rate constant of 4 X log M-I sec-1 for this compound. A much lower value
The Journal of Physical Chemistry
would have been predicted from the u* of its functional groups. The carboxylic acids studied, HCOOH, CHaCOOH, -OOC(CHz)2COOH, and (CH&CCOOH, fit fairly well on either Figure 3 or 4. Consequently, it is hard to decide whether the inductive or mesomeric effects predominate in these acids. Formic acid fits better on Figure 3, whereas (CHa)&COOH and -OOC(CH,), COOH fit better on Figure 4. Possibly both inductive and mesomeric effects contribute to the reactivity of carboxylic acids. The reactivity of eaq- with carboxylic acids has also been interpreted in terms of a Bronsted relationship (ie., taking the hydroxylic hydrogen as the reactive center) with reasonable success.21 However, one would predict succinic acid (pK,, = 5.6) to be considerably less reactive than acetic acid (pK, = 4.75) and the latter less reactive than formic acid (pKa = 3.75). On the contrary, the rate constants differ by only a factor of 2 . We conclude, therefore, that the reactivity of carboxylic acids with eaq- may be accounted for by their carbonylic functional group and that the “general acid” correlationz1 is probably not applicable in this case. The rate constants of glycine and alanine are 8 X lo6 and 6 X lo6 AI-’ sec-l, respectively.22 This low reactivity of amino acids is expected because these acids are predominantly in the form of the nonreactive zwitterion. Much higher rate constants would be predicted from the inductive effects of the amino and hydroxyl groups on the carbonyl bond. Peptides, on the other hand, like glycylglycine are more reactive than the amino acids. I n conclusion, the reactivity of hydrated electrons with carbonyl compounds behaves like other systems previously studied.6 Reactivit$y depends on the electron density at a specific reaction center. Any parameter which decreases the electron density at this center enhances the rate of attack by eaq-. Thus, hydrated electrons are again shown to be an adequate probe for estimating the electron distribution on a given substrate molecule. (21) J. Rabani, ref 6, p 242. (22) J. V. Davies, M. Ebert, and A. J. Swallow, in “Pulse Radiolysis,” J. H. Baxendale, et al., Ed., Academic Press Inc., New York, N. Y., 1965, p 165.
