Oxidation of Substituted Benzyl Alcohols by Quinoxalinium Dichromate. A Kinetic Study

Article abstract of DOI:10.1007/s00706-003-0143-9

Quinoxalinium dichromate (QxDC) oxidizes benzyl alcohol and substituted benzyl alcohols smoothly in dimethyl sulfoxide (DMSO) and in the presence of acid to the corresponding aldehydes. The reaction has unit dependence on each of the alcohol, QxDC, and acid concentrations. Electron-releasing substituents accelerate the reaction, whereas electron-withdrawing groups retard the reaction, and the rate data obey Hammett's relationship. The reaction constant ρ was -1.09 +/- 0.01 at 303 K. Oxidation of α,α-dideuteriobenzyl alcohol indicated the presence of a substantial primary kinetic isotope effect (k_H/k_D = 6.78 at 303 K). The reaction failed to induce the polymerization of acrylonitrile. The rates of oxidation were determined at different temperatures and the activation parameters were evaluated. The analysis of the dependence of the kinetic isotope effect on temperature indicated that the reaction involves a symmetrical cyclic transition state. A suitable mechanism is proposed.

Full text of DOI:10.1007/s00706-003-0143-9

Monatshefte f €u r Chemie 135, 483–491 (2004)
DOI 10.1007/s00706-003-0143-9
Oxidation of Substituted Benzyl
Alcohols by Quinoxalinium
Dichromate. A Kinetic Study
¨
Beytiye Ozg u¨ n and Nebahat De g˘ irmenbas
ꢀ
ꢁ
°
i
€
Gazi Universitesi, Fen-Edebiyat Fak €u ltesi, Kimya B €o l €u m €u , Teknikokullar,
6500 Ankara, Turkey
0
Received October 17, 2003; accepted (revised) October 24, 2003
Published online April 9, 2004 # Springer-Verlag 2004
Summary. Quinoxalinium dichromate (QxDC) oxidizes benzyl alcohol and substituted benzyl alcohols
smoothly in dimethyl sulfoxide (DMSO) and in the presence of acid to the corresponding aldehydes.
The reaction has unit dependence on each of the alcohol, QxDC, and acid concentrations. Electron-
releasing substituents accelerate the reaction, whereas electron-withdrawing groups retard the reaction,
and the rate data obey Hammett’s relationship. The reaction constant ꢀ was ꢂ1.09 ꢃ 0.01 at 303 K.
Oxidation of ꢁ,ꢁ-dideuteriobenzyl alcohol indicated the presence of a substantial primary kinetic
isotope effect (k =k ¼ 6.78 at 303 K). The reaction failed to induce the polymerization of acrylonitrile.
H
D
The rates of oxidation were determined at different temperatures and the activation parameters were
evaluated. The analysis of the dependence of the kinetic isotope effect on temperature indicated that
the reaction involves a symmetrical cyclic transition state. A suitable mechanism is proposed.
Keywords. Quinoxalinium dichromate; Oxidation; Kinetics; Mechanism.
Introduction
The major use of Cr(VI) reagents in synthetic chemistry is to oxidize alcohols to
carbonyl compounds. As a result, alcohol oxidation with these was extensively
investigated. A number of new Cr(VI) reagents together with special reaction
conditions have been developed. Anhydrous conditions are more conducive to
complexation of substrates with Cr(VI) species and therefore to mild oxidation.
The most significant recent Cr(VI) oxidants are quinolinium fluorochromate [1],
pyridinium chlorochromate [2], pyridinium dichromate [3], pyridinium fluorochro-
mate [4], and quinolinium chlorochromate [5]. We have been interested in the
synthesis and kinetics of reactions of complexed Cr(VI) species [6] and have
reported the kinetics and mechanism of oxidation of benzyl alcohol by potassium
ꢀ Corresponding author. E-mail: beytiye@gazi.edu.tr

€
B. Ozg €u n and N. De ꢀg irmenbas
°ꢁ
4
84
chlorochromate [7] and quinolinium chlorochromate [8]. Recently we prepared
quinoxalinium dichromate (QxDC) [9], which in dichloromethane readily oxidizes
primary and secondary alcohols to the corresponding carbonyl compounds. Here,
we examine the kinetics and mechanism of this oxidation of eleven primary
benzylic alcohols with QxDC in DMSO.
Results and Discussion
Since the results for all alcohols studied were similar with regard to the rate laws
and other experimental data, only representative data are reproduced for benzyl
alcohol (BA). The oxidation of the alcohols by QxDC led to the formation of the
corresponding aldehydes.
Rate Laws
Under pseudo-first-order conditions the pseudo-first order plots of log[QxDC] vs.
ꢂ3
time at different initial concentrations of QxDC (0.001 to 0.009 mol dm ) are all
linear with nearly the same slope indicating a first-order dependence of the rate on
the QxDC concentration (Table 1). Further, the pseudo-first-order rate constants
were independent of the initial concentration of QxDC at constant substrate con-
centration. In order to know the effect of the concentration of the alcohol on the
rate of oxidation, the kobs values were determined at different concentrations of
ꢂ3
alcohols in the range 0.04 to 0.50 mol dm (Table 1). The reaction rate increases
linearly with an increase in the concentration of alcohol (Table 1). Thus, the
Table 1. Rate constants for the oxidation of BA by QxDC in the presence of acid at 303 K
3
10 [QxDC]=M
4
10 k =s
ꢂ1
3
10 k =s
ꢂ1 ꢂ1
[
BA]=M
M
1
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.04
.06
.08
.10
.12
.15
.20
.50
.06
.06
.06
.06
.06
.06
.06
.06
.06
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
7.49
11.25
15.12
19.08
23.09
29.02
38.48
95.59
11.48
11.24
11.51
11.41
11.45
11.03
11.37
11.58
11.45
18.73
18.75
18.90
19.08
19.24
19.35
19.24
19.11
ꢀ
ꢀ Containing 10ꢂ3
M acrylonitrile

Oxidation by Quinoxalinium Dichromate
485
a
Table 2. Dependence of the reaction rate on acidity
[
TsOH]=M
0.05
0.07
0.10
0.20
0.25
4
0 k =s
ꢂ1
1
11.25
15.83
22.48
44.67
56.33
1
a
[
BA] ¼ 0.06 M, [QxDC] ¼ 0.003 M, T ¼ 303 K
reaction is first order with respect to the alcohol too. The plot of 1=k vs. 1=[BA] at
1
ꢄ
3
0 C is linear with no intercept on the 1=k axis. This indicates the absence of a
1
preformed complex or a complex whose formation constant is too low to be kineti-
cally determined.
The oxidation was found to be acid-catalyzed. The kobs values determined at
ꢂ3
increasing concentrations of TsOH in the range 0.05–0.25 mol dm increase
Table 2). The linear increase in the oxidation rate with acidity suggests the in-
(
volvement of a protonated Cr(VI) species in the rate determining step, which is a
stronger oxidant and electrophile than an unprotonated Cr(VI) species. There have
been earlier reports suggesting the involvement of such Cr(VI) species in chromic
acid oxidation [10], and protonated Cr(VI) species have been observed in the
presence of TsOH acid in nitrobenzene–dichloromethane mixtures [11].
Effect of Solvent Composition
The acid-catalyzed oxidation of BA was studied in solutions containing varying
proportions of DMSO and CH Cl (Table 3). Permittivity (D) values for DMSO-
CH Cl were estimated from the dielectric constants of the pure solvents [12] and
2
2
2
2
are recorded in Table 3. It can be seen that the solvent effect on reactivity is quite
significant. A plot of log k against the inverse of permittivity is linear (r ¼ 0.9618)
1
with a positive slope. This suggests an ion-dipole interaction in agreement [13]
with the observation that in the presence of an acid, the rate-determining step
involves a protonated Cr(VI) species. The oxidation of BA was studied in eight
organic solvents. The solubility of the reactants and the reaction of QxDC with
primary and secondary alcohols limited the choice of solvents. The values of
the rate constant, k , are recorded in Table 4. The increase in rate caused by the
obs
decrease in solvent polarity was attributed to an equilibrium shift favoring the
formation of chromate ester [14].
Effect of Temperature
The catalyzed oxidation of the substituted benzyl alcohols was studied at different
temperatures and it was observed that the rate increased with an increase in
a
Table 3. Dependence of the reaction rate on solvent composition
DMSO:CH Cl (v=v)
2
70:30
35.40
12.32
60:40
31.60
17.08
50:50
27.80
23.58
40:60
24.00
30.32
30:70
20.24
39.73
2
D
4
ꢂ1
0 k =s
1
1
a
[
QxDC] ¼ 0.003 M, [BA] ¼ 0.06 M, [TsOH] ¼ 0.05 M; T ¼ 303 K

€
4
86
B. Ozg €u n and N. De ꢀg irmenbas
°
ꢁ
a
Table 4. Effect of solvent on the oxidation of BA by QxDC at 303 K
4
10 k =s
ꢂ1
Solvent
1
DMSO
11.25
13.01
24.13
49.40
57.49
86.54
151.68
160.15
Dimethylformamide
Acetone
Dichloromethane
Tetrahydrofuran
Chloroform
Toluene
Benzene
a
[
QxDC] ¼ 0.003 M, [BA] ¼ 0.06 M, [TsOH] ¼ 0.05 M
Table 5. Temperature dependence of the oxidation of substituted benzyl alcohols by QxDC
4
ꢂ1
10 k =s
1
No.
Substituents
2
98 K
303 K
308 K
313 K
1
2
3
4
5
6
7
8
9
p-OMe
p-Me
m-Me
H
14.12
13.59
10.12
8.38
7.65
5.78
5.41
4.15
3.22
2.80
1.17
23.99
23.18
14.87
11.25
9.75
27.54
26.57
21.56
15.99
14.24
10.91
9.55
48.98
45.52
32.34
21.55
17.60
15.40
12.44
10.92
8.83
m-OMe
p-Cl
8.51
p-Br
7.86
m-Cl
m-Br
p-NO2
6.72
8.58
4.64
6.14
1
0
1
4.01
5.42
7.95
1
PhCD OH
2
1.66
2.35
3.33
[
QxDC] ¼ 0.003 M, [Alcohol] ¼ 0.06 M, [TsOH] ¼ 0.05 M
temperature (Table 5). From the linear plots of log k against the reciprocal of
temperature, the activation energy was calculated. The activation parameters were
calculated (Table 6) using the Eyring equation and the Gibbs equation [15].
Kinetic Isotope Effect
To ascertain the importance of the cleavage of the ꢁ-C–H bond in the rate-deter-
mining step, the oxidation of ꢁ,ꢁ-dideuteriobenzyl alcohol was studied at four
different temperatures (cf. Table 5). The reaction exhibited the presence of a sub-
stantial primary kinetic isotope effect (k =k ¼ 6.78 at 303 K).
H
D
Induced Polymerization of Acrylonitrile
The oxidation of BA under a nitrogen atmosphere failed to induce polymerization
of acrylonitrile. Further addition of acrylonitrile had no effect on the rate (Table 1).
Thus, a one-electron oxidation giving rise to free radicals is unlikely.

Oxidation by Quinoxalinium Dichromate
487
Table 6. Activation parameters of the oxidation of substituted benzyl alcohols by QxDC
#
#
#
Substituent
DH
kJ mol
ꢂDS
DG298
ꢂ1
ꢂ1 ꢂ1
ꢂ1
kJ mol
J mol
K
p-OMe
p-Me
m-Me
H
58.2 ꢃ 0.3
56.6 ꢃ 0.5
57.8 ꢃ 0.1
47.3 ꢃ 0.8
41.1 ꢃ 0.8
47.5 ꢃ 0.4
39.9 ꢃ 0.7
47.2 ꢃ 0.4
49.2 ꢃ 0.8
54.4 ꢃ 1.0
52.1 ꢃ 1.1
104 ꢃ 1.0
109 ꢃ 1.5
108 ꢃ 0.5
145 ꢃ 2.5
166 ꢃ 2.5
147 ꢃ 1.5
173 ꢃ 2.5
151 ꢃ 1.0
146 ꢃ 2.5
131 ꢃ 3.0
144 ꢃ 3.5
88.9 ꢃ 0.4
89.1 ꢃ 0.7
89.9 ꢃ 0.2
90.5 ꢃ 1.1
90.6 ꢃ 1.1
91.3 ꢃ 0.6
91.5 ꢃ 1.0
92.2 ꢃ 0.5
92.7 ꢃ 1.1
93.4 ꢃ 1.3
95.0 ꢃ 1.5
m-OMe
p-Cl
p-Br
m-CI
m-Br
p-NO2
PhCD OH
2
Correlation Analysis of Reactivity
A Hammett plot [16] of log k vs. ꢂ was drawn (Fig. 1) to correlate structure-
1
reactivity for these reactions. The rate increased with electron-donating groups
and decreased with respect to electron-withdrawing groups (Table 5). Electron-
donating substituents stabilize a transition state having a carbocation centre by
resonance. This facilitates the bond-breaking process in the transition state. On
the other hand, an electron-withdrawing substitutent increases the capacity of the
carbocation centre, suggesting the development of a loose or partial positive charge
at the transition state, which in turn helps the bond-making process, thus leading to
a decrease in rate. The negative ꢀ values indicate an electron-deficient transition
state and increase with increase in temperature. The ꢀ values at temperatures 298,
3
03, 308, and 313 K are ꢂ1.05, ꢂ1.09, ꢂ1.10 and ꢂ1.25. Such values are not
#
uncommon in hydride ion transfer mechanism [17]. The similarity of DG values
Fig. 1. Hammett plot for the QxDC oxidation of substituted benzyl alcohols (for numbering see
Table 5)

€
B. Ozg €u n and N. De ꢀg irmenbas
°ꢁ
4
88
Scheme 1
#
#
arises from changes in DH and DS values and stresses the probability that these
oxidation reactions involve similar rate determining steps (Table 6). The high
negative values of activation entropies support the assumption of a highly solvated
transition state due to its increased polarity. They are also indicative of the bimo-
lecular rate-determining step [18] as shown in Scheme 1.
The correlation between activation enthalpies and entropies of oxidation of the
eleven alcohols is fairly good (r ¼ 0.9891). The value of the isokinetic temperature
is 267 K. However, according to Exner [15] an isokinetic relationship between the
calculated values of activation enthalpies and entropies is often vitiated by random
experimental errors. Exner suggested an alternative method for establishing the
isokinetic relationship. Exner’s plot between log k at 298 K and at 313 K was
1
reasonably linear (r ¼ 0.9732, Fig. 2). The value of isokinetic temperature evalu-
ated from Exner’s plot is 186 K. The linear isokinetic correlation implies that all
the alcohols are oxidized by the same mechanism and the changes in the rate are
governed by changes in both the enthalpy and entropy of activation.
Mechanism
The oxidations have been studied in non-aqueous medium. Thus, hydrolysis
of dichromate may safely be ruled out. QxDC shows stronger absorbance in the
wave length region 375–320 nm than K Cr O . This indicates that in solution QxDC
2
2 7
does not dissociate to any appreciable extent. The fact that only TsOH brings about
Fig. 2. Exner plot for the QxDC oxidation of substituted benzyl alcohols (for numbering see Table 5)

Oxidation by Quinoxalinium Dichromate
489
the oxidation points out an involvement of TsOH in the oxidation process. The UV-
VIS spectrum clearly confirms the complex formation between oxidant and TsOH.
Complex formation between imidazolium dichromate-TsOH [19], pyridinium
dichromate-TsOH [20], NDC-TsOH, and INDC-TsOH [21] has already been
reported. If the formation constant of the complex is small the oxidation is dependent
on [TsOH]. The QxDC-TsOH complex is more reactive than other Cr(VI) species
because the complexation makes the oxidant a better electrophile. It behaves like a
protonated Cr(VI) species, which is a stronger oxidant and electrophile [10].
An examination of the kinetic data obtained in the present investigation showed
that the rate of oxidation of the alcohols was dependent on the first powers of the
concentrations for each reactant (substrate, oxidant, and acid) (Table 1) and QxDC-
TsOH is the likely oxidizing species.
The presence of a substantial primary kinetic isotope effect indicates that the
rupture of the C–H bond occurs after the O–H bond cleavage creating a carboca-
tion centre or a positive character in the transition state. The ꢀ values higher than
ꢂ3 generally suggest a fairly large degree of carbonium character in the transition
state [22]. The ꢀ value of ꢂ1.09 found in the present study is indicative of a partial
bonding. The negative value of the polar reaction constant together with the sub-
stantial deuterium isotope effect indicates that the transition state approaches a
carbocation in character. Hence, the transfer of a hydride from the alcohol to the
oxidant is suggested.
The hyride transfer may take place either by a cyclic process via an ester
intermediate [23] or by an acyclic one-step bimolecular process [24]. Kwart and
Nickel [23] have shown that the dependence of k =k on temperature can be
H
D
successfully employed to determine whether the loss of hydrogen proceeds through
a concerted cyclic process or by an acyclic one.
The data for protio- and deuterio-benzyl alcohols, fitted to the familiar expres-
#
ꢂDH =RT
sion, k =k ¼A =A e
, show a direct correspondence with the properties
of a symmetrical transition state in which activation-energy difference for protio-
and deuterio-compounds are equal to the difference in the zero-point energy for the
H
D
H
D
ꢂ1
respective C–H and C–D bonds (ca. 4.5 kJ mol ) and the entropies of activation of
the respective reactions are nearly equal [25, 26]. Evidence against a hydrogen
transfer occuring by a concerted one-step bimolecular process has been well docu-
mented by Bordwell [27]. It is evident that in the present study also the hydrogen
transfer does not occur by an acyclic bimolecular process. It is well established that
intrinsically concerted sigmatropic reactions characterized by a transfer of a hydro-
gen in a cyclic transition state are the only truly symmetrical processes involving a
linear hydrogen transfer [28]. Furthermore, a cyclic hydride transfer in the oxida-
tion of alcohols by Cr(VI) involves six electrons and, being a H €u ckel-type system,
is an allowed process [29]. Therefore, one can safely conclude that in the oxidation
of benzyl alcohols by QxDC, the hydride transfer occurs via a chromate ester.
Formation of ester intermediates in the oxidation of hydroxy compounds is well
documented [10]. Further, an intermolecular hydride transfer should normally lead
þ
to a correlation with ꢂ values [30]. Good correlation with Hammett’s ꢂ values is
also in accordance with the ester mechanism. A mechanism depicted in Scheme 2
accounts for the experimental results. Thus, a transition state having a planar,
cyclic symmetrical structure can be envisaged for the decomposition of the ester

€
B. Ozg €u n and N. De ꢀg irmenbas
°ꢁ
4
90
Scheme 2
intermediate (Scheme 2). A linear transition state will certainly exhibit a larger
isotope effect than a cyclic intermediate [31].
Experimental
Materials
All the alcohols were commercial products (Fluka or Aldrich) and purified by either recrystallization or
distillation. p-Toluenesulphonic acid (TsOH) was used as a source of protons. DMSO and CH Cl were
2
2
purified by the usual methods [33]. ꢁ,ꢁ-Dideuteriobenzyl alcohol was purchased from Isotec Inc.
USA), its isotopic purity was ꢅ98%.
(
Product Analysis
The product analysis was carried out under kinetic conditions. In a typical experiment BA (0.51 cm ,
3
3
.005 mol) and QxDC (0.48 g, 0.001 mol) were dissolved in 10cm of DMSO and the solution was
0
allowed to stand for ca. 24h to ensure completion of the reaction. Most of the solvent was removed by
distillation under reduced pressure. The residue was then treated with an excess of a freshly prepared
saturated solution of 2,4-dinitrophenylhydrazine in 1.5 M H SO and kept overnight in a refrigerator. The
2
4
precipitated 2,4-dinitrophenylhydrazone (DNP) was filtered off, dried, weighed, recrystallized from
ethanol, and weighed again. The product was identical (mp and mixture mp) with an authentic sample
of the DNP of benzaldehyde. The yields of DNP before and after recrystallization were 84%and 76%.
Kinetic Measurements
The pseudo-first-order conditions were attained by keeping a large excess (10-fold or greater) of the
alcohols with respect to QxDC. The reactions were carried out at constant temperature (ꢃ0.3 K) and
were followed up to 80% conversion by monitoring the decrease in [QxDC] at 355nm on a Shimadzu
UV-160A double-beam spectrophotometer. The pseudo-first-order rate constant, k , was evaluated
1
from the linear plots of ln[QxDC] versus time. The values reported were the mean of at least duplicate
runs and were reproduciple to within ꢃ4%.
Acknowledgements
The authors are grateful to the Research Foundation of Gazi University for supporting this study.
References
[
[
1] Murugesan V, Panduragan A (1992) Indian J Chem 31B: 377
2] Corey EJ, Suggs JW (1975) Tetrahedron Lett 16: 2647

Oxidation by Quinoxalinium Dichromate
491
[
[
[
[
3] Corey EJ, Schmidt G (1979) Tetrahedron Lett 20: 399
4] Bhattacharjee MN, Chaudhuri MK, Dasgupta HS, Roy N (1982) Synthesis 7: 588
5] Srinivasan R, Ramosh CV, Madhulatha W, Balasubramanian K (1996) Indian J Chem 35B: 480
€
€
6] a) Ozg €u n B, De ꢀg irmenbas
°
ꢁ N (1996) Synth Commun 26: 3601; b) Ozg €u n B, De ꢀg irmenbas
°
ꢁ N
€
(
1999) Synth Commun 29: 763; c) De ꢀg irmenbas
€
°
ꢁ N, Ozg €u n B (2002) J Inst Sci Techn Gazi Univ
1
5: 657; d) De ꢀg irmenbasꢁ N, Ozg €u n B (2003) Monatsh Chem 134: 1565; e) De ꢀg irmenbas
° °ꢁ N,
€
Ozg €u n B (2003) Monatsh Chem (accepted for publication)
€
[
[
[
7] Ozg €u n B, Pek A (1991) React Kinet Catal Lett 43: 589
€
8] Ozg €u n HB, De ꢀg irmenbas
°
ꢁ N (1997) J Chem Res (S) 32; (1997) J Chem Res (M) 220
€
9] De ꢀg irmenbas
°
ꢁ N, Ozg €u n B (2002) Monatsh Chem 133: 1417
[
[
[
10] Wiberg KB (1965) Oxidation in Organic Chemistry, part A. Academic Press, New York, p 69
11] a) Banerji KK (1978) J Chem Res (M) 2561; b) Banerji KK (1979) Ind J Chem 17A: 300
12] Furniss BS, Hannaford AJ, Smith PWG, Tatchell AR (1989) Vogel’s Textbook of Practical
Organic Chemistry. John Wiley & Sons, New York, p 1442
[13] Amis ES (1967) Solvent Effects on Reaction Rates and Mechanisms. Academic Press,
New York, p 42
[
[
[
[
[
[
[
[
[
[
[
14] Westheimer FH (1949) Chem Rev 45: 419
15] Frost AA, Pearson RG (1970) Kinetics and Mechanism. Wiley Eastern, New Delhi, p 144
16] Wiberg KB (1964) Physical Organic Chemistry. Wiley, New York, p 403
17] Harlan L, Goering & Robert RJ (1958) J Am Chem Soc 80: 3277
18] Klumpp GW (1982) Reactivity in Organic Chemistry. Wiley, New York
19] Chidambaranathan V (1997) Ph.D. Thesis, Annamalai University, Annamalainagar
20] Amutha M (1999) Ph.D. Thesis, Annamalai University, Annamalainagar
21] Karunakaran C, Chidambaranathan V (1999) Revue Roumaine de Chimie 44: 491
22] Freeman F, Yamachika NJ (1970) J Am Chem Soc 92: 3730
23] Kwart H, Nickle JH (1973) J Am Chem Soc 95: 3394
24] Rocek J, Westheimer FH, Eschenmoser A, Moldovanyi L, Schreiber J (1962) Helv Chim Acta
4
5: 2554
[
[
[
[
[
[
[
[
[
25] Kwart H, Latimer HC (1971) J Am Chem Soc 93: 3770
26] Kwart H, Slutsky J (1972) J Chem Soc Chem Commun: 1182
27] Bordwell FG (1972) Acc Chem Res 5: 374
28] Woodward RB, Hoffmann R (1969) Angew Chem 8: 781
29] Littler JS (1971) Tetrahedron 27: 81
30] Wiberg KB (1963) Physical Organic Chemistry. John Wiley, New York, p 410
31] Scott SL, Bakac A, Espenson JH (1992) J Am Chem Soc 114: 4205
32] Cainelli G, Cardillo G (1984) Chromium Oxidation in Organic Chemistry. Springer, Berlin
33] Perrin DD, Armarego WL, Perrin DR (1966) Purification of Organic Compounds. Pergamon
Press, Oxford