6983 (21) P. Desiongchamps, P. Athni, D. FreGl, A. Molaval, and C. Moreau. Can. J. Chem., 52, 3651 (1974). (22)J. M. Lehn and G. Wipff, J. Am. Chem. Soc.,B6,4048(1974). (23)The parameter p (0 6 p 6 1) defining the linear variation of the 13 parameter is rehted to the rectified distance covered by Hd along a tentative path determined by m a n s of previous calculations on more rigid geometries. A posteriori it was verified that such a path was substantialiy correct and very near to that reported in Figure 6. As a consequence it was not considered necessary to start with an iterative series of corrections of the tentative path. (24)L. Radom, J. A. Pople. and P. v. R. Schleyer, J. Am. Chem. SOC., 95,
8193 (1973). (25) Data taken from: D. D. Wagman, W. H. Evans, V. B. Parker, I. Halow, S. M. Bailey, and R. H. Schumm, hlatl. Bur. Stand. (U.S.), Tech. Note, No. 270-3 (1968). (26)J. P. Guthrie, J. Am. Chem. Soc., 96,3608 (1974). (27)M. L. Bender and R. J. Thomas, J. Am. Chem. Soc.,83,4183 (1961). (28) S.0.Eriksson, Acta Chem. S c a d . , 22, 892 (1968). (29) R. M. Pollak and M. L. Bender, J. Am. Chem. Soc., 92,7190(1970). (30)D. Drake, R. L. Schowen, and H. Jayaraman, J. Am. Chem. Soc., 95,
454 (1973). (31)J. M. Sayer and W. P. Jencks. J. Am. Chem. Soc.. 95,5637(1973).
Kinetics of the Base-Catalyzed Decomposition of a-Hydroperoxy Ketones Yasuhiko Sawaki and Yoshiro Ogata* Contribution No. 21 5 from the Department of Applied Chemistry, Faculty of Engineering, Nagoya University, Chikusa-ku. Nagoya, Japan. Received March 31, 197.5
Abstract: The alkoxide-catalyzed decomposition of 15 a-hydroperoxy ketones la-o afforded generally high yields of ketones (80-100%) and esters (70-100%). The high yields of esters show that the a-cleavage reaction proceeds predominantly via an acyclic, carbonyl addition intermediate. The formation of a small amount of carboxylic acid was caused by the reaction of hydroxide ion via the acyclic intermediate rather than via a cyclic 1,2-dioxetane. The pseudo-first-order rate constant, kobsd, with respect to 1 is proportional to [RONa] a t lower base concentrations and then approaches a constant at higher ones. The behavior suggests a reaction between RO- and the undissociated hydroperoxide, and the resulting second-order rate constant ( k l o ) for MeONa ranged from 4.2 X 10-1 A4-I sec-I for IC to 5.6 X M-l sec-l for I b in benzene-methanol at 0'. The substitutent effect on a-hydroperoxy-&.a-diphenylacetophenonesexhibited positive p values of 2.5-3.2 and 0.9- 1.7 (with u ) on benzoyl and a-phenyl rings, respectively. Rate-determining fragmentation of the carbonyl addition intermediate was suggested from the observations: (i) the facile transesterification of a-hydroperoxy esters (3), (ii) the relative reactivities of la-n, and (iii) the effect of hydroxide or hydroperoxide ion on the ester yield.
a-Hydroperoxy ketones are well known as intermediates in the autoxidation of ketones,' which were sometimes iso-
HO I
RiC-CRZR,
R,C-CR,R,
I I
I1 I
0-0
0 OOH 1
2
a, R, = R, = R, = Ph b, R , = R, = R, = Me c , R1, R, = -(CH2)4-;R3 = Ph
lated and identified to be a-ketohydroperoxide (1)* rather than 1,2-dioxetane (2).3 The autoxidation of ketones in basic media rapidly forms a-hydroperoxy ketonesIdq4 and gives a-hydroxy ketones in the presence of phosphite.s Although a-hydroperoxy ketones can yield a-diketones when R2 = H,"d,6a main alkaline reaction of 1 is the a cleavage (eq l).'d,4,7 R,C-CR,R,
II I
Results Rate. The reaction of a-hydroperoxy ketones (1 or R'OOH) with sodium alkoxide in benzene-alcohol (1: 1 in volume) at 0' was monitored by iodometry and expressed as:
--+
RICOZH
+
RZR,C=O
(1)
The values of kobsd were constant up to 7040% conversion except at low concentration of alkoxide, where the constancy was observed only at initial 10-20% conversion. The kobsd value is proportional to [RONa] at [EtONa] < 0.01 M or at [MeONa] < 0.1 and then approachs a constant (Table I). The validity of eq 2 at constant [RONa] with varying [R'OOH] is also shown in Table IIB. This result, which denies the induced radical decomposition of 1, together with the high yield of esters (>60%) suggests the unimportance of a base-catalyzed radical decomposition. Products. The reaction of l a with alkoxides produced benzophenone and benzoate (Table 11). RjC-CRZR,
+
RO-
--+
II 1
0 OOH
0 OOH
The mechanism was often written to involve 1,2-dioxetane (2) or its but an alternative acyclic mechanism was also ~uggested.~g-j Recent studies by Richardson7f and by Bordwel17' reached two different conclusions, cyclic and acyclic, respectively. We wish to report here our resultss on the alkaline decomposition of 16 a-hydroperoxy ketones, which support the acyclic mechanism for the a-carbon cleavage reaction (eq 1) as a main path. Sawaki, Ogata
R,CO,R
+
R,R,C=O
+ HO-
(3)
The yield of the ester decreased at lower [RONa] (Tables IIA and IIC) and a small amount of benzoic acid was det e ~ t e d The . ~ addition of water decreased considerably the yield of esters but only slightly the rate of decomposition (Table IIE), which suggests that HO- as well as alkoxide ion can react to form benzoic acid according to eq 3. Ester
1 Base-Catalyzed Decomposition of a-Hydroperoxy Ketones
6984 Table I. Rate Constants of Alkaline Decomposition of a-Hydroueroxy Ketones in Benzene-ROR (1: 1) at 0.0"a [RONal, M
l a + EtONa
l a + MeONa
lb + MeONab
lcc + MeONa ~
d0.064 0.131 0.49 0.92 1.36 1.87 2.26 2.15
0.001 0.002 0.005 0.010 0.020 0.050 0.10 0.20 0.30 0.50
-0.038 0.088 0.184 0.378 0.852 1.44 1.94 2.10
0.015 0.0402 0.0778 0.129 0.156 0.184
-0.71 2.09 4.34 9.2
a Observed first-order rate constants with [ 11 = 0.010 M and ROH = EtOH or MeOH. b Reaction at 25". ca-Hydroperoxy-aphenylcyclohexanone.
formation is also the predominant pathway for other a-hydroperoxy ketones as stated below. Substituent Effect. The substitutent effect for the alkoxide decomposition of a-hydroperoxy-a,a-diphenylacetophenone ( l a ) is summarized in Table 111. Products with 0.05 M RONa are the corresponding benzophenones (80- 100%) and esters (70-100%). The effect in the benzoyl ring resulted in positive p's of 2.5-3.2 (vs. u ) with 0.002-0.05 M RONa (Table IIIA). The effect on a-phenyl ring afforded lower p's of 0.9-1.7 (vs. u ) (Table IIIB). The identity of p values with various [RONa] suggests that the dissociation constants of R'OOH are of similar magnitude, if the overall rates involve the constants. The reaction of various a-hydroperoxy ketones with methoxide is shown in Table IV, where esters are formed in a yield of over 80% for most cases, The exceptional case is l o ( R I = i-Pr) (30-50% yield of ester), where the reaction is very slow and hence produced ester is partially hydrolyzed (10-20% conversion). Even if this hydrolysis is corrected, the selectivity for ester is 35-55%. However, the re-
action with HO- to produce isobutyric acid is probably operative (eq 3). The relative rate of HO- vs. MeO- is probably higher for l o ( R I = i-Pr) than for l b (R1 = Me) because of steric hindrance of i-Pr. In fact, addition of water results in a 7-8% increase in rate for the reaction of l o in benzene-methanol, l o a-Hydroperoxy cyclic ketone (IC)has the fastest rate. As for the acyclic ketones, replacing methyl with phenyl resulted in an approximately tenfold increase in the rate, Le., l b 11 (R3 = Et) > lm (R3 = PhCH2) and of l b ( R I = Me) > lo (R1 = i-Pr) reflect a steric effect ( E , ) in addition to inductive effect ( u * ) . But a quantitative treatment is impossible because of unavailable u* and E , values for C(OOH)R2R3. Temperature Effect. As shown in Table V, the changes in Ea (15-17 kcal/mol) and A S * (-5 to -11 eu) are small in spite of the ca. 200-fold difference in rate and there appears no clear-cut tendency between them. Transesterification of a-Hydroperoxy Esters. Table VI illustrates the preliminary transesterification of a-hydroperoxy ester (3) without its decomposition at 0'. This shows that eq 4 is reversible without the fragmentation of adduct OR '
+
Ph&+CO,R
R'O'
1
OOH
e
I
ph,C-C-OR
I
I
HOO 0-
3
4 PhZC-CO2R'
I
+
RO- (4)
OOH 4. The decomposition of 3 at 25' is due to the reaction with
contaminated hydroxide ion, since the addition of water, Le., HO-, dramatically increases the decomposition of 3. It was not known whether the basic decomposition of a-peroxy esters proceeds via 4 or hydrolyzed a-peroxy acid." The above results clearly suggest that the fragmentation mechanism of 4 is not substantiated.
Table 11. Rates and Products for Alkaline Decomposition of l a in Benzene-EtOH (1: 1) at O.O"a Products, %C [EtONa] ,M
Reaction conditionb
103kobsd,sec-'
0.005 0.02 0.05 0.10 0.05 0.05
[la] [la] [la] [la]
0.005 0.020 0.05
B-M (111) B-M (1:l) B-M (1:l)
(C) Effect of [MeONa] 0.088 0.374 0.852
0.05 0.05 0.05
E-M = 9:l E-M = 4:l E-M = 1 : l
(D) Effect of [ROHId 1.84 1.61 1.24
0.05 0.05 0.05
[H,O] = 0.02M [H,O] = 0.5M [H,O] = 1.OM
(E) Effect of [H,O] e 1.93 1.94 1.67
0.05 0.05
PhC0,Me
(A) Effect of [EtONa] 0.49 1.36 1.87 2.26 (B) Effect of [ l a ] = 0.00125
= 0.005 = 0.010 = 0.020
1.91 1.87 1.93
PhC0,Et 62 79 87 94
100 98 96 103
80 80 86
98 96 96
64 100 93 24.2 48.6 84.0
Ph,C=O
93 97 98 68.9 55.5
19.1 94 73
5 3f
104 101 105 105 100 97
a With [ l a ] = 0.010M and reaction time of 2-5 hr. b B = benzene, M = MeOH, and E = EtOH. CProducts were determined by GLC analysis after 2-5 hr of reaction ( 2 5 % ) . d Benzene-ROH (1: 1) and ROH = MeOH + EtOH in volume. e Reaction time of 20 min. f Benzoic acid (40%) was determined after esterification with diazomethane.
Journal of the American Chemical Society
/
97:24
/ November 26,1975
6985 Table 111. Substituent Effect for the Basic Decomposition of l a (Ph = C,H, R,C-CR,R, bOH R,
Id le la If 1g
p-MeOPh p-MePh Ph pClPh mClPh
lh li Ij
Ph Ph Ph
C,H,Ia 103k0b,d, sec-I
8
R,
01
0.002 M EtONa
R,
Ph Ph Ph Ph Ph P(Q) rb
Ph Ph Ph Ph Ph
0.01 M EtONa
0.05 M EtONa
0.01 M MeONa
0.05 M MeONa
(A) Effect on R, -0.019 0.128 -0.040 0.326 0.13 0.915 4.61 0.834 8.87 2.5 3 2.73 0.995 0.996
0.287 0.686 1.99 7.06 15.9 2.66 0.997
0.0217 0.0628 0.191 0.925 2.96 3.22 0.998
0.113 0.292 0.857 4.45 11.3 3.08 0.998
1.70 1.58 4.62 -0.91 0.896
0.0782 0.0862 0.486 1.68 0.990
0.541 0.490 3.28 1.66 0.948
(B) Effect on R, p-MeOPh p-MePh pCPh
Ph Ph Ph
0.536 0.516 1.75 1.12 0.977
P(0)
rb
a Reaction with [R'OOH] = 0.01 M in benzene-ROH (1: 1) at O.O", where ROH is EtOH or MeOH. Products with 0.05 M RONa were R,CO,R (70-100%) and R,R,C=O (80-100%). b Correlation coefficient.
Table lV. Substituent Effect for the Basic Decomposition of Various Types of 1 (Ph = C,H, or C,H,) at O.Ooa R,C-CRzR3 bOH
8
la lk 11 lm In lb lbd lod 1P
R,
R,
Ph Ph Ph Ph Ph Me Me iPr Me
Ph Ph Ph Ph Me Me Me Me Ph
Products, % R, Ph Me Et CH,Ph Me Me Me Me . Ph Ph
-(CHJ4-
IC
103k,bsd, sec-'
103k,,b M-' sec-'
Relative rate
R,CO,Me
0.852 0.721 0.0292 0.01 35 0.0261 0.00278 0.0402 0.00314 0.12 21g
17.0 14.4 0.58 0.27 0.52 0.056 0.804 0.0628 1.2 420
308 258 10.5 4.9 9.4 (1.00)
93 91 89 78 C
R,R3C=0 98 100 81 7 7c
97c 9 oe 30 -5 O f
0.078d 22 7550
96e 77h
a The reaction with [ l ] = 0.010 M and [MeONa] = 0.05 M in benzene-MeOH (1:l). bsecond-order rate constants estimated from Y = kobsd [ I ] , = k,[MeONa] [ I ] at [MeONa] = 0.05 M, where k&sd is proportional to [MeONa]. CWith 0.1 M MeONa at 25" for 1-3 hr. At 25". e With [ l b ] = 0.15 Mand [MeONa] = 0.3 M a t 25" for 4 hr. Methyl acetate was determinedbyGLC(~lO%);acetoneas2,4dinitrophenylhydrazone. f Reaction with [MeONa] = 0.2-0.5 M in methanol at 25" for 5-20 hr resulted in 20-50% conversion and the selectivity for methyl isobutyrate is shown (six runs). g Extrapolated from kobsd with 0.002-0.020 M MeONa (Table I). hYield of methyl 6-benzoylvalerate with 0.1 M MeONa. Its yield was constant (74 5 4%) with 0.002-0.2 M MeONa.
,
Table V. Temperature Effect on the Basic Decomposition of a-Hydroperoxy Ketones in Benzene-MeOHa io3kohsd. sec-'
R'OOH
[MeONa] ,M
0.0"
15.0"
25.0"
k2b(25'),M-' sec-'
EaC
Afisc
la Id lk 11 In lb
0.010 0.050 0.010 0.050 0.050 0.050
0.141 0.117 0.155 0.0292 0.0261 0.00278
0.5 82 0.539 0.660 0.160 0.130
1.44 1.35 1.61 0.424 0.367 0.0402
0.144 0.0270 0.161 0.0085 0.0073 0.00080
19.1 15.8 15.1 17.1 17.1 17.2
13.9 14.6 13.9 15.9 15.9 16.0
-A&d
10.9 8.4 1.9 7.4 5.2 9.6
aWith [R'OOH] = 0.010Min benzene-MeOH (1:l). bSeefootnote b inTableIV.C+l kcal/mol.d+3eu. Table VI. Transesterification of a-Hydroperoxy Esters (0.01 M) in the Presence of 0.1 M RONa in Benzene-ROH Ph,CCO,R I X X
R
OOH OOH OOH OOH OOH H H H
Et Et Et Me Me Me Et Et
Reaction condition@ (temp, "C; time, hr) B-M-MeONa B-M-MeONa B-M-MeONa B-E-EtONa B-E-EtONa B-E-EtONa B-M-MeONa B-M-MeONa
(0; 16) (25; 1.6) (25; 16) (0; 16) (25; 16) (0; 16) (0; 16) (25; 16)
Remaining peroxide, %b >98 >98 73 >98 74
Ph,CCO,R, %c I X R=Me
R = Et
46 43 70 42 11 0 41 91
30 21 5.5 37 52 100 59 8
Ph,C=O, % 0.4 1.4 7.2 1.8 10
a B = benzene, M = MeOH, and E = EtOH. b By iodometry. ca-Hydroperoxy esters were determined by GLC analysis after neutralization with AcOH and the reduction of the peroxides with Ph,P to the corresponding &-hydroxy esters.
Sawaki, Ogata
/ Base-Catalyzed Decomposition of a-Hydroperoxy Ketones
6986
Discussion Mechanism. Postulated mechanisms for the a fission of a-hydroperoxy ketones are as follows: Mechanism A ( 1,2-rearrangement mechanism)'*
rVH- RiCO-C-R,I
2.0
OH
RiC-C-R,
II
I
I
II
0 Rz
!LO -t
a
Rz
0
p-.
5
RiCO,H
+ R,R,C-=O
0
(5)
0
Mechanism B (dioxetane m e ~ h a n i s m ) ~ ~ J
2A 1A Mechanism C (intermolecular C=O ism)7i OOH RO OOH
I
RiC-C-R,
I 0 Rz 1
II
+ RO-
I
+
addition mechan-
+
I - 0 R, 6 RiCOzR + RzR,C=O
0.3
M
Figure 1. Effect of [ R O N a ] on k&sd for the alkaline decomposition of la (0.010 M ) in benzene-ROH ( 1 : l ) a t 0.0".
I
RIC-C-R,
0.2
0.1
[RO*.1.
I
+ HO'
(7)
The observed p (2.5-3.2) for ring substituents in the benzoyl group is similar to other cases of nucleophilic addition to carbonyl ( p = 2-3).18 The effect for a-phenyl is lower ( p = 0.9-1.7), since the a-phenyl is one atom further removed from the carbonyl. The fast reaction of l c reflects the facile carbonyl additions to cyclohexanone^.'^ The slower reaction of aliphatic a-hydroperoxy ketones than of aromatic ketones is curious, since nucleophilic additions to acetyl are usually much faster than those to b e n ~ o y l . This ~ ~ . seems ~~ to mean that the addition is not rate determining as discussed later. The kobsd value for the basic decomposition of 1 (R'OOH) increases with [RONa] but approaches a constant at high basicity (Figure 1A). This effect is explicable by mechanism C (eq 7), which involves an attack of alkoxide on 1 as follows:
Mechanism A was proposed for the spontaneous decomposition of a-hydroperoxy a-alkoxy ketone,I2 whose high reactivity arises probably from the acceleration by the aKa alkoxy group.I3 However, our peroxidic ketones are much R'OOH + RO- e R'OO' + ROH (8) more stable in organic solvents and decompose only by k9 strong acid or base and eq 5 cannot explain the high yield of R'OOH + RO' products (9) esters from 5. Assuming the products are formed by eq 10 Second, the evidences for the dioxetane mechanism are the observed carboxylic acid and chemilumines~ence.~~ But, v = k,b,d[R'OOH], = IzE~[R'OOH][RO-] (10) carboxylic acid is also formed by mechanism C ( R O = HO) and the high yield of ester cannot be explained by mechaHere, [ Is denotes stoichiometric concentration. Then, nism B, which must be an unimportant pathway if any. Probably, the high strain energy (-26 kcal/mol) of a 1,21 +% (11) dioxetane14 reduces [2A] and hence mechanism B. The ob=k,[RO;1 k , servation of the ~hemiluminescence~~ seems to be indicative The plot of l/kobsd vs. l/[RO-] is h e a r (Figure 1B) and of the operation of the cyclic mechanism even in a small exthe intercept and the slope give Kg = 3.8 and 71 M-' and tent on the basis of efficient chemiluminescence of other k9 = 0.0192 and 0.175 M - ' sec-I for the reaction of l a 1,2-dioxetanes.I But chemiluminescence is observable from with MeO- and EtO-, respectively. Peroxide l a is too unthe peroxide decomposition even by an acyclic mechanstable to determine the Kg value directly under these basic ism16aand from the reaction of fluorescein with either alkaline hydrogen peroxide'6b or tert-butyl h y d r ~ p e r o x i d e , ' ~ conditions, but this order is acceptable in comparison to the Kg value of 18 M-I for a-hydroperoxy ester 3 (R = Me) in and hence in the presence of fluorescein it is not always the methanol,21 and to the acidity difference between MeOH decisive evidence for the dioxetane mechanism. We oband EtOH.22 served no luminescence from the basic decomposition of aThe similar treatment for l b with methoxide resulted in hydroperoxy ketones (la,d,e) in the presence of fluoK8 4 M-I and k9 = 0.001 1 M-I sec-l in benzene-methrescein. I anol at 25O. On the other hand, the K8 from uv absorbThird, the high yield of esters supports mechanism C as a anceZ3was 12 h4-l. The difference (ca. threefold) is conmajor pathway. The accompanying acid may be due mostly siderable but not so large in view of the accuracy of K8 from to a reaction of hydroxide ion ( R O = H O in eq 7), since eq 12; e.g., the calculated K8 increases up to 5.6 M - ' by a there is no reason for the dramatic change of the mecha2096 increase of kobsd at lower [MeO-1. nism by the addition of only 1-2% of water (Table IIE and The k9 value of l a with EtO- (0.175 M - ' sec-') in benfootnote IO). That is, since the majority of the base in the zene-ethanol is considerably larger in comparison to k9 presence of I-2% water exists as alkoxide ion rather than with MeO- (0.019 M-l sec-I) in benzene-methanol. This HO-, it is abnormal to assume that the reaction of the alkdifference is largely due to a solvent effect. That is, the oxide (eq 7) is reduced and the reaction via mechanism B kobsd value, selectivities (Tables IID and IIE), and acidities (eq 6 ) is dramatically accelerated by addition of a little of water and alcoholsZZafford approximate k9 values of water. Mechanism C explains the results as follows.
-
kobsd
-
Journal of the American Chemical Society
/
97:24
/ November
26, 1975
6987 nism C as a predominant pathway. Usually esters were obtained in 80-1001 yield and these values are the lower limit for the acyclic mechanism, since hydroxide ion was also shown to be a potent reagent to afford carboxylic acid. Our explanation for the exceptional low yield (30-50%) of ester from l o (R1 = i-Pr) is that the reaction with HO- via mechanism C occurs to some extent in contrast to the lower estimation by Richardson et al.7f It is noteworthy that the selectivity for the ester from l o remains constant ( ~ 3 0 % ) ~ ' even if 90% of R'OOH is dissociated into R'OO-.30-31 Probably, the high strain (-26 kcal/mol) of 1,2-dioxetane reduces 2A or mechanism B to a minor reaction.
0.12, 0.035, and 0.023 M-' sec-' for HO-, EtO-, and MeO-, respectively, in benzene-EtOH-MeOH. There appears no large difference in reactivities between MeO- and EtO- in the same solvent, In contrast to the ordinary carbonyl reactions, the reactivity of HO- is usually high, resulting in a certain amount of acid formation even in the presence of a trace of water. Rate-Determining Step. The above treatments do not decide whether the rate-determining step is addition (eq 12) or fragmentation (eq 13). The following considerations lead to a conclusion that eq 13 determines the rate. RO R,C-CR,R, I/ 1
+ RO'
F+
I I
Rl-C-CR,R3
I
Experimental Section
(12)
All melting points and boiling points were uncorrected. N M R were recorded with a JNM-C60HL (Japan Electron Optics), and ir with a Perkin-Elmer 337 spectrophotometer. The GLC analysis 1 6 was performed with a Yanagimoto 550F gas chromatograph. '13 Materials. Starting ketones were obtained by various methods 6 RlC0,R -.+ R2R3C=0 + HO' (13) (see Table V I I I ) . The ketones, a, i, and j, were synthesized from (i) The facile transesterification of a-hydroperoxy esters desyl chloride and benzenes,32 the reaction temperature (time, min) being 80' ( 1 5), 80' (1 5 ) , and 110' (30). respectively (meth( R I = M e 0 or EtO) shows the mobile addition, and the od A). same is probably true for the case of ketones ( R I = alkyl or Method B is exemplified for the preparation of 2-phenylpropiophenyl). Moreover, carbonyl additions to ketones are generally much faster than those to esters, e.g., s a p ~ n i f i c a t i o n . ~ ~ phenone (k) as follows. Dropwise addition of sulfuryl chloride (27 g, 0.2 mol) to propiophenone (27 g, 0.2 mol) with stirring below (ii) The addition of HO- or HOO- reduces the yields of 40' afforded a-chloropropiophenone (80%). bp 126- 128' ( 18 ester from la, which is different from the reaction of phenyl mm). Phenylmagnesium bromide (0.12 mol) in ether was dropped benzoates and benzoic anhydride, where C=O addition is with stirring into ice water cooled a-chloro ketone (20.4 g, 0.12 rate determining (Table VII). That is, although EtO- and mol) in benzene (50 ml) and the reaction mixture was refluxed for MeO- exhibit the similar relative reactivities on both sub2 hr. The usual work-up afforded 80% yield of 2-phenylpropiophestrates, HO- is effective only for la, and HOO- is highly none (96% pure by GLC), bp 130-145' (2.5 mm), which was purified by crystallization from methanol, mp 50-51' (lit.3349'). Keeffective only for the benzoates. This suggests the rate-detones c and I were obtained similarly from the corresponding atermining C=O addition to la, since HO- is usually 10chloro ketones. 100-fold less effective than alkoxide as a nucleophileZ2and Ketone e was prepared from diphenylacetyl chloride and toluHOO- is a potent reagent with a effect, its steric requireene32and ketone d by the reaction in anisole-hexane ( 1 : l ) (method ment being less than water.13.25The high reactivity of HOC). Methods D,34 E,35 F,36and G3' are the literature methods. All is explicable by the favored attack of less hindered HOthe starting ketones were checked by GLC to be >98% pure. compared to the attack of EtO- on the hindered tertiary kePreparation of a-Hydroperoxy Ketones. a-Hydroper0xy-a.atone la, although the equilibirum constants for PhCHO diphenylacetophenones were generally obtained by method H , the RO- e PhCHO-(OR) are of the same magnitude (0.16 ethoxide catalyzed autoxidation of ketones (see Table VIII). a,aDiphenylacetophenone (2.16 g, 8 mmol) in DMF (40 ml) was M-I for HO- in water26and 0.22 M-' for MeO- in methcooled to -20' and 1 M sodium ethoxide in ethanol ( 1 1 ml, 1 1 an01~~). mmol) was added dropwise with gentle bubbling of oxygen (ca. 4 (iii) The relative reactivities, l b l n = 1:9 and 1p:la = rnin). After the gentle bubbling for 20 min, the cold reaction mix1:14 (Table IV), are abnormal (acetyl
