Much attention has been paid to the electrochemical techniques as a potential tool in synthetic chemistry. Electrolysis procedure can be frequently employed for oxidation and reduction of organic compounds. However, very little is known about the acid-catalyzed reaction in an anodic process although several electrochemical reactions in MeCN-LiClO₄-Pt¹ and MeOH-LiClO₄ or Et4NClO₄-Pt systems² have been explained in terms of an acid catalysis. We describe here a practical method for the transformation of epoxides 1 to ketones 2 and acetonides 3 promoted by a powerful electrogenerated acid. During our study, an electron-transfer chain reaction for an electrochemical ring opening of some epoxides has been proposed³. In contrast to this, the present electrochemical transformation of epoxides 1 to ketones 2 is an electrogenerated acid-catalyzed reaction.

A typical electrolysis condition is as follows; A mixture of 1a (0.2 mmol) and LiClO₄ (1.0 mmol) dissolved in CH₂Cl₂ (7.5 ml) - THF (0.5 ml)⁴ was electrolyzed under a constant current (10 mA for 1 min, 0.03 F/mol) at room temperature using platinum foils as electrodes in an undivided cell, affording 2a (91%) and 8-benzyloxy-2,6-dimethyloct-1-en-3-ol (4a) (8%). The electrochemical transformation of 1 to 2 is characterized by the following facts:

1. The reaction completes by passing a very small amount of electricity (0.02 - 0.1 F/mol),
2. Without electrolysis no 2a was detected and 1a was recovered,
3. By electrolysis in a divided cell 2a was obtained only from an anolyte and 1a was completely recovered from a catholyte,
4. Ketone 2a was obtained in good yield (91%) even when 1a was allowed to react in a pre-electrolyzed solution⁵.
5. Presence of pyridine (1 eq) in the pre-electrolyzed solution inhibited the conversion and no 2a was detected.
6. In the prolonged electrolysis. a dark gray lithium metal deposited on the cathode and a check of pH revealed that the electrolysis solution was acidic,
7. The electrolysis gave higher yield and selectivity than those of the usual acid-catalysis (Table 2).

Table 1

Rearrangements of Epoxides 1
to Ketones 2 and Acetonides 3

Epoxide 1	F/mol	Product 2a (3b)
a	0.09	91%
	0.20	(86%)**
b	0.06	74%
	0.17	(70%)*
c	0.03	80%
	0.10	(87%)*
d	0.04	88%**
	0.06	(83%)
e	0.01	87%f
	0.10	(88%)
f	0.03	88%**
	0.15	(85%)
g	0.06	67%
	0.09	(83%)*
h	0.06	82%c
	0.09	(85%)*
i	0.12	72%c
j	0.23	82%d
	0.72	(84%)*,g
k	1.22	30%
	2.53	97%e

a) 1 (0.2 mmol), LiClO₄ (1 mmol)
and Et₄NClO₄ (0.5 mmol) in DCM
(a-g), ClCH₂CH₂Cl (h, i, j), DCM-THF
(k) (8 ml) at room temp.
Current density (mA/cm²) 6.7
(a,b,i,k), 3.3 (d,e,g,h,j) and 1.7 (c,f).
b) Yield of 3 is indicated by (%):
^* LiClO₄, ^** Mg(ClO₄)₂.
c) At 75°C.
d) At 65°C.
e) Rxn in a pre-electrolyzed soln.
f) 1e (2.4 mmol)
g) 51°C

These experimental results strongly suggest that the electrochemical conversion is an electrogenerated acid-catalyzed reaction rather than an electron-transfer chain reaction^3,6.

The mechanism on the generation of the electrogenerated acid is not uncovered yet. However. it is plausible that oxidation of a trace amount of water existing in the electrolysis solution would generate a naked acid in an unbuffered condition which is therefore extremely strong in aprotic solvents⁷.

The effect of electrolytes is important. Perchlorates such as LiClO₄ (91%, 0.02 F/mol), NaClO₄ (90%. 0.04 F/mol), and Mg(ClO₄)₂ (79%. 0.04 F/mol in THF) significantly promoted the conversion. Use of Et4NClO₄ provided 2a in a good yield (89%) but it required more electricity (0.5 F/mol in CH₂Cl₂). Interestingly, use of Et₄NOTs in CH₂Cl₂ led to the preferential formation of 4a (57%) although the current efficiency was unsatisfactory (3.1 F/mol). In contrast, both Et₄NBr and CF₃CO₂Li did not give the desired ketones.

Use of Et₄NBr in CH₂Cl₂ resulted in a recovery of 1a even after 3.7 F/mol of electricity was passed. On electrolyzing 1a with CF₃CO₂Li, debenzylation took place in part, affording benzaldehyde (21%) and 6,7-epoxy-3,7-dimethyloctan-1-ol (19%) along with the recovered 1a (70%) after 4.5 F/mol.

Aprotic solvents such as CH₂Cl₂ (91%), ClCH₂CH₂Cl (95%), THF (85%), Me₂CO₃ (75%), and AcOEt (76%) are useful for the purpose. However. use of MeCN (simply distilled once over P₂O₅) and MeOH (treated with Mg and distilled) did not promote the electrochemical reaction.

Similarly, the electrochemical conversion of 1 into acetonides 3 was performed by electrolyzing 1 in a distilled acetone containing LiClO₄ or Mg(ClO₄)₂ at room temperature. Particularly noteworthy is the fact that the electrochemical acetonidation can be performed by use of a trace amount of LiClO₄ (1% mol eq to 1c) and a catalytic amount of electricity (0.1 F/mol).

The selectivity of the acetonidation is strictly controlled by the concentration of LiClO₄ ; [Acetonide 3c (%) : Ketone 2c (3) (LiClO₄ mol eq to 1 ). 87:5 (0.01). 76:15 (0.2). 61:28 (2.0). 51:40 (12)]. Ketone 2c increased with the concentration of LiClO₄ presumably because in the high concentration of the salt the intermediate carbenium ion would be significantly associated with LiClO₄ as proposed by Pocket⁸. thus retarding nucleophilic attack of acetone. As an electrolyte. LiClO₄ (87%. 0.1 F/mol). NaClO₄ (84%. 0.16 F/mol), Mg(ClO₄)₂ (90% 0.1 F/mol), and LiBF₄ (81%. 0.1 F/mol) were effective. However, Et₄NClO₄ gave poor yields and current efficiency (37%, 1.7 F/mol) and the use of Et₄NBr or Et₄NOTs resulted recovery of 1c even after 2.0 F/mol of electricity was passed.

The electrolysis of the corresponding diol of 1c in LiClO₄ - acetone provided 3c (85%. 0.3 F/mol) although the current efficiency was lower than that of epoxide 1c (0.1 F/mol). In contrast to this. cyclohexene oxide gave cyclohexene diol (91%) as a sole product which could not be converted to the corresponding acetonide even when 1 F/mol of electricity was passed. This result suggests that acyclic diols also can be transformed into acetonides by the present electrolytic method.

The results of the electrochemical rearrangement of epoxides 1 to ketones 2 and acetonides 3 are summarized in Table 1.

Acyclic ω-epoxyisoprenoids 1a-1f were converted to 2a-2f and 3a-3f in satisfactory yields.

Functional groups involved in 1 (OH, OAc, SO₂Ph, CHO, and C=C and C-C bonds) did not change in the electrolysis conditions.

Epoxides 1h derived from isosafrole provided 3,4-methylenedioxyphenyl-2-propanone (2h) (82%), a synthetic precursor of alpha-methyldopa⁹. Similarly, menthene oxide 1i gave a mixture of menthone and isomenthone (72%). Although epoxide 1j of a long chain fatty acid (erucic acid) is less reactive, it could be converted to the corresponding ketone 2i (82%) and acetonide 1i (84%). Electrolysis of 1k proceeded very slowly (1.22 F/mol), affording triphenylmethyl phenyl ketone 2k (10%) along with benzophenone (69%). In contrast, the reaction of 1k in a pre-electrolyzed solution provided 2k in 97% yield (After 2.53 F/mol, 1k was added to the pre-electrolyzed solution and stood for 2.5 h.).

The result of the electrochemical rearrangement (1 → 2) apparently demonstrates usefulness of the electrolysis method (Table 2) in comparison with the conventional chemical acid-catalyzed reaction. On the contrary to the electrolysis, the action of BF₃ etherate in benzene which is one of the most widely employed reagents for the epoxide ring opening induced in part Wagner-Meerwein type rearrangement, affording 7-benzyloxy-2,2,5-trimethylheptan-1-al 16%) as a by-product along with the desired ketone 2a (77%).