Stability-reactivity relation on the reaction of β,β-disubstituted vinyl cations with ethanol

Article abstract of DOI:10.1021/jo9603441

A family of β,β-disubstituted α-(p-methoxyphenyl)vinyl cations has been generated by the laser flash photolysis of the corresponding vinyl bromides, and the rates of reactions of the cations with ethanol in acetonitrile have been measured at 25°C. The obs

Full text of DOI:10.1021/jo9603441

5274
J . Org. Chem. 1996, 61, 5274-5279
Sta bility-Rea ctivity Rela tion on th e Rea ction of â,â-Disu bstitu ted
Vin yl Ca tion s w ith Eth a n ol
Shinjiro Kobayashi,* Yuji Hori,† Toshinori Hasako,† Ken-ichi Koga, and Hiroshi Yamataka*,‡
The Institute for Fundamental Research of Organic Chemistry, Kyushu University, Higashi-ku,
Fukuoka 812-81, J apan, and Department of Industrial Chemistry, Faculty of Science and Engineering,
Saga University, Honjo-machi, Saga 840, J apan
Received February 20, 1996X
A family of â,â-disubstituted R-(p-methoxyphenyl)vinyl cations has been generated by the laser
flash photolysis of the corresponding vinyl bromides, and the rates of reactions of the cations with
ethanol in acetonitrile have been measured at 25 °C. The observed rate constants differ greatly
depending on the substituents, ranging from 3.05 × 105 L mol-1 s-1 to 8.18 × 107 L mol-1 s-1. The
thermodynamic stabilities of the vinyl cations have been estimated by means of ab initio MO
calculations for model compounds, which reveals that their stability is almost unaffected by â,â-
dialkyl substituents. In the present system, therefore, the stability-reactivity relation, in which
a less stable cation is expected to show higher reactivity, breaks down. Several model transition
structures have been considered in the MO calculations, and the results indicate that the â
substituents can move away from the incoming nucleophile to avoid steric congestion but that there
still exists a large steric repulsion in the transition state. The calculations indicate that the
reactivity of the vinyl cations is primarily controlled by this steric effect, which is the reason for
the breakdown of the stability-reactivity relationship.
The relationship between reactivity and selectivity is
nucleophile more slowly.4,5 Thus, the stability-reactivity
relation holds. For example, the bis(methoxyphenyl)-
methyl cation decomposes in acetonitrile/water (1:2 v/v)
with a rate constant of 105 s-1, which is considerably
faster than the corresponding rate for the tris(methoxy-
phenyl)methyl cation (101 s-1).4 Similarly, the R-phenyl-
ethyl cation, which is 4.9 kcal/mol less stable than the
cumyl cation,8 reacts with MeOH 10 times faster in HFIP
(4.3 × 107 vs 4.1 × 106 L mol-1 s-1).5
a well-known concept in organic chemistry. In relation
to this concept, Ritchie reported 20 years ago that the
relative rates of reactions of stable carbocations with a
series of nucleophiles were independent of the reactivity
of the cations.1,2 Thus, the reactivity differences of stable
cations such as crystal violet and p-nitrobenzenediazo-
nium ion in reactions with a large variety of nucleophiles
ranging from the very reactive MeO- to the less reactive
H2O were independent of the cation stabilities. These
findings have cast serious doubt on the generality of the
reactivity-selectivity relationship. Recent reports by
Mayr support this conclusion.3 It should be noted,
however, that there is an apparently similar but different
concept of organic reactivities, the stability-reactivity
relationship, in which a more stable cation is considered
to react more slowly. In the above examples, the reactiv-
ity-selectivity principle breaks down, but the stability-
reactivity relationship always holds.
Carbocations generated via solvolytic reactions are
normally highly unstable, and the reactions of the cation
intermediates with solvent or other nucleophiles are not
well elucidated since they are too fast to follow by
conventional methods. Reactivities of such unstable
cations remain to be determined. Laser flash photolysis
is one of the techniques that can be used to generate
unstable cationic species in solution and thus allows the
study of such reactions of the cations with nucleophiles.4-7
In the case of benzylic cations, laser flash photolysis
has shown that a more stable cation reacts with a
Stabilities of R-phenylvinyl and benzylic cations are
known to differ only slightly in the gas phase.8,9 On the
basis of heats of formation data compiled in the litera-
ture,10 the intrinsic stability of the vinyl cation is
calculated (eq 1, in kcal/mol) to be between those of
RH h R+ + H-
(1)
methyl and ethyl cations, i.e., 313 (R ) CH3), 287
(H2CdCH), 271 (CH3CH2), 251 ((CH3)2CH), and 234
((CH3)3C). This is true not only for these unstable
aliphatic carbocations but also for the relatively stable
benzylic cations. The relative stabilities (in kcal/mol) of
R-phenyl-substituted carbocations determined by ion
cyclotron resonance mass spectrometry have been re-
ported to decrease in the order PhC(CH3)2+ (0) > PhCH-
(CH3)+ (6) > PhC()CH2)+ (8) > PhCH2 (13).8,9 Thus,
+
the stability of R-arylvinyl cations is similar to that of
benzylic cations. In the present study, we have generated
a family of â,â-disubstituted vinyl cations from the
corresponding bromides by the laser flash technique and
have measured the rates of the reactions of the cations
with ethanol (eq 2). Ab initio molecular orbital calcula-
tions were also carried out for the cations, protonated
† Saga University.
‡ Present address: The Institute of Scientific and Industrial Re-
search, Osaka University, Ibaraki, Osaka 567, J apan.
X Abstract published in Advance ACS Abstracts, J uly 15, 1996.
(1) Ritchie, C. D. Acc. Chem. Res. 1979, 12, 42.
(2) Ritchie, C. D. Pure Appl. Chem. 1978, 50, 1281.
(3) Mayr, H.; Patz, M. Angew. Chem., Int. Ed. Engl. 1994, 33, 938
and references therein.
(4) McClelland, R. A.; Kanagasabapathy, V. M.; Banait, N. S.;
Steenken, S. J . Am. Chem. Soc. 1989, 111, 3966.
(5) McClelland, R. A.; Chan, C.; Cozens, F.; Modro, A.; Steenken, S.
Angew. Chem., Int. Ed. Engl. 1991, 30, 1337.
(7) Kobayashi, S.; Schnabel, W. Z. Naturforsch. 1992, 47b, 1319.
(8) Mishima, M.; Arima, K.; Inoue, H.; Usui, S.; Fujio, M.; Tsuno,
Y. Bull. Chem. Soc. J pn. 1995, 68, 3199.
(9) Mishima, M.; Ariga, T.; Fujio, M.; Tsuno, Y.; Kobayashi, S.;
Taniguchi, H. Chem. Lett. 1992, 1085.
(6) McClelland, R. A.; Kanagasabapathy, V. M.; Banait, N. S.;
Steenken, S. J . Am. Chem. Soc. 1991, 113, 1009.
(10) Lias, S. G.; Bartmess, J . E.; Liebman, J . F.; Holmes, J . L.;
Mallard, W. G. J . Phys. Chem. Ref. Data 1988, 17, 1.
S0022-3263(96)00344-1 CCC: $12.00 © 1996 American Chemical Society

Reaction of â,â-Disubstituted Vinyl Cations with Ethanol
J . Org. Chem., Vol. 61, No. 16, 1996 5275
alcohols, and model transition structures in order to
obtain insight into factors controlling reactivity in the
vinylic system.
Exp er im en ta l Resu lts a n d Ca lcu la tion s
Acetonitrile solutions containing â,â-disubstituted R-(p-
methoxyphenyl)vinyl bromides and EtOH at varying
concentrations were irradiated in a rectangular quartz
cell with a 5 ns laser flash of 266 nm light. Concentra-
tions of the bromides were (0.65-1.7) × 10-5 mol/L. The
photolysis of â,â-disubsituted R-(p-methoxyphenyl)vinyl
bromides resulted in the formation of transient optical
absorption spectra characteristic of substituted vinyl
cations having absorption maxima around 350 nm.11,12
The decay of this transient optical absorption was
measured at λ ) 340 nm in the presence of excess EtOH
in acetonitrile, and in all cases the decay conformed to
pseudo-first-order kinetics as shown in Figure 1. Second-
order rate constants were calculated according to eq 3,
where k0 is the decay rate constant obtained in the
absence of EtOH. The kinetic plots are illustrated in
Figure 2, and rate constants are listed in Table 1.
F igu r e 1. Decay of â,â-dimethyl-R-(p-methoxyphenyl)vinyl
cation at 340 nm at 25 °C.
kobsd ) k0 + k2[EtOH]
(3)
Ab initio MO calculations were carried out to clarify
the origin of the effect of â-substituents on the reactivity
of the vinyl cations with EtOH. In the model chosen for
the calculation, the EtOH nucleophile and the methoxy-
phenyl group of the cations were replaced by H2O and
by H or Ph, respectively, for simplicity. Vinyl cations
have the special conformational characteristic that the
vacant p orbital of the cation center lies in the same plane
as the â substituents. Therefore, in the model transition
states OH2 was located in the plane of the vinyl moiety
and the CR-O bond lengths were fixed at different values.
Thus, the calculations were carried out with three
coordinates fixed (the Câ-CR-O angle at 90.0°, the
O-CR-Câ-R1 dihedral angle at 0.0°, and the CR-O bond
length at RC-O Å) and other coordinates optimized. Three
sets of model transition states (TS1, TS2, and TS3) in
which RC-O was 2.0, 2.5, and 2.7 Å, respectively, were
considered. All structures were optimized without any
symmetry constraint at the HF/3-21G level of theory by
using the Gaussian 92 program.13 In Table 2 are shown
the structures of the model transition states, cations, and
F igu r e 2. First-order rate constants (s-1, 25 °C) for the decay
of â,â-disubstituted R-(p-methoxyphenyl)vinyl cations as a
function of ethanol concentration in acetonitrile.
Ta ble 1. Secon d -Or d er Ra te Con sta n ts of th e Rea ction
of â,â-Disu bstitu ted r-(p-Meth oxyp h en yl)vin yl Ca tion s
w ith Eth a n ol in Aceton itr ile a t 25 °C
substituent
rate constants
abbrev
R1, R2
k0 (s-1
)
k2 (L mol-1 s-1
)
C4
C5
C6
Me2
Et2
Ph2
-(CH2)4-
-(CH2)5-
-(CH2)6-
Me, Me
Et, Et
Ph, Ph
4.98 × 106
3.23 × 106
1.91 × 106
1.90 × 106
9.57 × 105
5.65 × 104
8.18 × 107
2.27 × 107
7.62 × 106
9.35 × 106
1.67 × 106
3.05 × 105
protonated vinyl alcohols. The calculated energies are
summarized in Table 3. The level of calculations is
rather low because the study required calculations for
vinyl cations that had one phenyl group at the R and two
phenyl groups at the â position, which made calculations
at a higher level impractical. However, the study com-
pares only the variation of energies of vinyl cations,
model transition structures, and the protonated vinyl
alcohol products for a series of substituted substrates.
(11) Schnabel, W.; Naito, I.; Kitamura, T.; Kobayashi, S.; Taniguchi,
H. Tetrahedron 1980, 35, 3229.
(12) Kobayashi, S.; Kitamura, T.; Taniguchi, H.; Schnabel, W. Chem.
Lett. 1983, 1117.
(13) Frisch, M. J .; Trucks, G. W.; Head-Gordon, M.; Gill, P. M.;
Wong, M. W.; Foresman, J . B.; J ohnson, B. G.; Schlegel, H. B.; Robb,
M. A.; Replogle, E. S.; Gomperts, R.; Andress, J . L.; Rachavachari, K.;
Binkley, J . S.; Gonzalez, C.; Martin, R. L.; Fox, D. J .; Defrees, D. J .;
Baker, J .; Stewart, J . J . P.; Pople, J . A. GAUSSIAN 92, Gaussian, Inc.,
Pittsburgh, PA, 1992.

5276 J . Org. Chem., Vol. 61, No. 16, 1996
Kobayashi et al.
Ta ble 2. Geom etr ica l P a r a m eter s of Ca tion s, Mod el Tr a n sition Str u ctu r es, a n d P r oton a ted Vin yl Alcoh olsa

Reaction of â,â-Disubstituted Vinyl Cations with Ethanol
J . Org. Chem., Vol. 61, No. 16, 1996 5277
Ta ble 2 (Con tin u ed )
a
ROH2 denotes protonated vinyl alcohol. All distances in Å and all angles in degrees.
Ta ble 3. Ca lcu la ted En er gies of Vin yl Ca tion s, Mod el Tr a n sition Str u ctu r es, a n d P r oton a ted Vin yl Alcoh ols for R3 ) H
a n d P h Ser ies
R1, R2
C4
vinyl cation
TS1
TS2
TS3
protonated alcohol
-420.287 104
(-191.955 390)
-459.149 583
(-230.817 957)
-497.976 657
(-269.644 658)
-382.667 417
(-154.329 850)
-460.30 8846
(-231.975 932)
-761.560 493
b
-495.890 814
-495.892 072
-495.891 475
-495.908 621
(-267.629 134)
-534.771 030
(-306.492 422)
-573.597 365
(-345.319 957)
-458.290 417
(-230.012 201)
-535.929 876
(-307.650 990)
-837.190 847
(-608.909 867)
C5
-534.748 967
-573.570 139
-458.264 353
-535.900 819
-837.155 916
-534.751 846
-573.575 606
-458.266 732
-535.905 848
-837.156 912
-534.751 980
-573.576 732
-458.267 865
-535.907 127
-837.155 944
C6
Me2
Et2
Ph2
a
b
Energies in hartrees. Numbers in parentheses are energies (in kcal/mol) for R3 ) H. â,â-Diphenylvinyl cation could not be localized
because it became a phenyl-bridged cation.
Ta ble 4. Rela tive Ca tion Sta bility on th e Ba ses of
Isod esm ic Rea ction 4a
As shown below, the calculated relative stabilities of
â-substituted vinyl cations are in good agreement with
the corresponding gas phase experimental values, and
the use of this level of theory is justified for the present
purpose.
R1, R2
R3 ) Ph
R3 ) H
C4
C5
C6
Me2
Et2
Ph2
-4.03
-4.08
-4.54
-3.10
-4.34
+1.50
-14.9
-14.5
-13.9
-9.5
Discu ssion
-14.1
Table 1 shows observed second-order rate constants for
reactions of â,â-disbustituted R-(p-methoxyphenyl)vinyl
cations with EtOH for â-substituents Me2, Et2, Ph2,
cyclobutyl, cyclopentyl, and cyclohexyl. It can be seen
that the reactivity differences arising from the effects of
subsituents at the â position are surprisingly large. In
the conventional interpretation, the reactivity of cations
is expected to correlate with their stability in such a
manner that a more stable cation shows less reactivity,
as has been observed for benzylic cations.4 However, it
appears that the stability-reactivity relation does not
hold for these vinylic cations.
The relative stabilities of â,â-disubstituted R-(p-meth-
oxyphenyl)vinyl cations were evaluated by means of ab
initio MO calculations for model compounds. They were
obtained on the basis of the isodesmic reaction of eq 4 in
which the carbocation is formed from a protonated
alcohol. The energies of these cationic species are listed
Energies (in hartrees) of H2O, H2CdC+H, H2CdC(H)OH2
,
a
+
H2CdC+Ph, and H2CdC(Ph)OH2+ are -75.585 960, -76.655 774,
-152.353 291, -305.008 928, and -380.636 878, respectively.
in Table 4. For the series with R3 ) H in eq 4, the â,â-
Me2 cation and the â,â-Et2 cation are seen to be more
stable than the parent vinyl cation by 9.5 and 14.1 kcal/
mol, respectively. A similar, though less pronounced,
trend is found for the R3 ) Ph series. These values are
comparable to the experimental gas phase stabilities of
â-substituted vinyl cations, where the stabilities of the
mono-â-substituted phenylvinyl cations have been de-
termined by ICR from the competitive protonation of
differently substituted phenylacetylenes. The â-Me- or
â-Et-substituted R-(p-methoxyphenyl)vinyl cations were
thus determined to be more stable than the parent cation
by 1.8 and 3.7 kcal/mol, respectively. However, as shown
in Figure 3, the reactivity differences for a series of â,â-
disubstituted R-(p-methoxyphenyl)vinyl cations do not
show a good correlation with cation stabilities. Thus, the
reactivity order does not arise from the relative cation
stabilities. However, it is interesting to note that ap-
parently the bulkier the R group the slower the rate.
in Table 3 and the derived relative vinyl cation stabilities

5278 J . Org. Chem., Vol. 61, No. 16, 1996
Kobayashi et al.
F igu r e 5. Plots of relative stabilities of TS1 and TS2 vs those
of TS3.
F igu r e 3. Correlation between log k2 for â,â-disubstituted
R-(p-methoxyphenyl)vinyl cations with ethanol in acetonitrile
and the cation stability.
substituent effect on the transition state is expected to
be negligible in the present system since the effect on
the stabilities of the cations (eq 4) is small. It is therefore
reasonable to assume that the difference in stability of
the model transition structures primarily arises from the
steric effect between the â substituents and the OH2
molecule (nucleophile) that is located closer to the R
group in the transition state than in the product state.
Hence, calculations were carried out for three different
transition states with increasing CR-O bond lengths
(TS1, TS2, and TS3 in Table 2).
As in the case of log k2, there is only a crude correlation
between the relative stabilities of the model transition
structures and Es. However, correlations among the
model transition structures are more informative. Figure
5 shows plots for TS1 (top) and TS2 (bottom) against TS3.
The plotted numbers are the relative energies of each
model structure given by ∆E ) E(TS) - E(R1R2CdC+Ph
+ H2O). A less negative number means that the model
structure suffers more steric repulsion and is less stable.
In both cases, the cycloalkyl derivatives for 4-, 5-, and
6-membered rings in Table 3 gave excellent linear
correlations with the slopes of 2.4 and 1.4; thus, the
cycloalkyl derivatives are well-behaved substituents. The
magnitudes of the slopes indicate that the sensitivity to
steric bulkiness is larger for a more congested structure.
In contrast, Et2- and Ph2-substituted derivatives deviate
downward from the correlation line in the bottom plots
in Figure 5. This indicates that Et2 and Ph2 in particular
exhibit smaller steric destabilization than expected from
the correlation line for TS2. A similar trend was found
in the upper plots; here, even the Me2 derivative deviates
downward, and the deviation of Ph2 is very large. These
findings indicate that a set of steric parameters that is
applicable to one reaction system may not be applied to
other systems. The results can be rationalized by as-
suming that the Me2, Et2, and Ph2 substituents are more
flexible than the cyclic substituents and can partially
accommodate steric repulsion by changing the local
conformation when the nucleophile comes very close to
them. This assumption is consistent with the optimized
geometries shown in Table 2 as described below.
F igu r e 4. Plots of log k2 vs Taft’s Es constants.
One way to estimate the degree of steric hindrance for
the approach of a nucleophile at the transition state is
to examine a correlation between log k2 and the Taft Es
values of the R1R2C groups. Such a correlation is shown
in Figure 4. However, although it has some success it is
not consistently satisfactory. This is not unexpected
because the steric requirement at the TS of the present
reaction may not be the same as that of the reference
reaction (acid hydrolysis of esters), and this is an
unavoidable limitation in the use of steric parameters.14
In order to see whether the observed reactivity differ-
ences can be rationalized in terms of the steric effect of
the â substituents, we have considered three families of
model transition structures. In the language of physical
organic chemistry there are two factors that affect the
relative stability of the model transition structures,
namely electronic and steric effects. The electronic
One way for the transition structures to accommodate
the steric congestion is to increase the CRCâC2 angle (here
(14) Taft, R. W., J r. In Steric Effects in Organic Chemistry; Newman,
M. S., Ed.; Wiley: New York, 1956; Chapter 13.

Reaction of â,â-Disubstituted Vinyl Cations with Ethanol
J . Org. Chem., Vol. 61, No. 16, 1996 5279
Figure 6 shows plots of log k2 against the relative
stabilities of the model TS3. The reasonably good
straight line obtained indicates that the reactivity of the
â,â-disubstituted R-(p-methoxyphenyl)vinyl cations is
controlled by the steric effect. It can be concluded that
the reactivities of â,â-disubstituted vinyl cations with a
nucleophile depend only slightly on their stabilities and
are primarily controlled by the steric effect. This control
by steric effects is the reason for the breakdown of the
stability-reactivity relation in the vinylic system. It
should also be emphasized that steric parameters that
are applicable to one system may not be applied success-
fully to another system because the effect of steric
congestion is different from system to system and sub-
stituents behave differently where steric requirements
differ.
Exp er im en ta l Section
F igu r e 6. Correlations between log k2 and the relative
stability of TS3.
Acetonitrile and ethanol were purified according to standard
procedures.15 â,â-Dialkyl-R-(p-methoxyphenyl)vinyl bromides
were prepared by refluxing dichloromethane solutions of the
corresponding ketone (60 mmol) and Vilsmeyer reagent (DMF
(15 mL)-PBr3(30 mL)) for 3 days. Laser flash photolysis was
carried out as reported previously.7
C2 is the carbon atom of the substituent attached to Câ
in a syn manner); this occurs for all substituents.
Additional mechanisms to evade steric congestion operate
for the transition structure of the Ph2-, Et2-, and Me2-
substituted derivatives. For Ph2, twisting of the Câ-Ph
bond provides an effective mechanism. The two phenyl
groups in the cation tend toward planarity in order to
achieve effective π-electron overlap with the double bond.
The optimal CRCâCPhC CPhC dihedral angle in the cation
1-(p-Meth oxyp h en yl)-1-br om o-2-eth yl-1-bu ten e: color-
1
less liquid; bp 140 °C/2.5 mmHg; H-NMR (CDCl3) δ 0.95 (t,
3H, J ) 7.5 Hz), 1.1 (t, 3H, J ) 7.5 Hz), 2.04 (q, 2H, J ) 7.5
Hz), 2.40 (q, 2H, J ) 7.5 Hz), 3.80 (s, 3H), 6.85 (d, 2H, J ) 8.7
Hz), 7.20 (d, 2H, J ) 8.7 Hz); mass spectrum m/ z (relative
intensity) 268 (M+, 157.2), 269 (M + 1, 30.3), 270 (M + 2, 160.9)
1
2
is 46.2°. However, this dihedral angle becomes larger
in going from TS3 (66.4°) to TS2 (66.8°) to TS1 (68.3°) as
a result of twisting of the Câ-Ph bond to avoid steric
repulsion from the incoming nucleophile. Because of this,
the change of the CRCâCR angle can be relatively small
for the Ph2 substrate; it varies from 133.0° (TS3) to 134.9°
(TS1) in contrast to the larger change for the C6 cy-
cloalkyl substrate, from 134.8° (TS3) to 138.6° (TS1). For
the Me2 and Et2 derivatives, conformational change at
the C-C bond is the mechanism adopted to avoid steric
congestion. For Me2, the two Me groups at the â position
in the cation have the conformation shown in Table 2.
The conformation in TS2 and TS3 is the same as in the
cation. However, TS1, in which OH2 comes close to the
Me group, takes up a different conformation by rotating
60° around the Câ-C2 bond. For Et2, rotation of the
CH3-CH2 bond is the preferred mechanism for absorbing
steric repulsion; the C-C bond rotates to move CH2-H
away from the nucleophile in more congested transition
structures.
[(p-Meth oxyph en yl)br om om eth ylen e]cyclobu tan e: col-
orless liquid; bp 100 ˚C/2.5 mmHg; 1H-NMR (CDCl3) δ 2.00
(q, 2H, J ) 7.8 Hz), 2.79-2.88 (m, 4H), 3.80 (s, 3H), 6.84 (d,
2H, J ) 8.7 Hz), 7.38 (d, 2H, J ) 8.7 Hz).
[(p-Meth oxyph en yl)br om om eth ylen e]cyclopen tan e: col-
orless crystal; mp 36.5-37.5 °C (EtOH); 1H-NMR(CDCl3) δ
1.70-1.78 (m, 4H), 2.27 (m, 2H), 2.47-2.2.52 (m, 2H), 3.81 (s,
3H), 6.84 (d, 2H, J ) 8.7 Hz), 7.32 (d, 2H, J ) 8.7 Hz)
[(p-Meth oxyph en yl)br om om eth ylen e]cycloh exa n e: col-
orless crystal; mp 57-59 °C (EtOH); 1H-NMR(CDCl3) δ 1.44-
1.68 (m, 6H), 2.13 (t, 2H), 2.53 (t, 2H), 3.81 (s, 3H), 6.84 (d,
2H, J ) 8.7 Hz), 7.21 (d, 2H, J ) 8.7 Hz); mass spectrum m/ z
(relative intensity) 280 (M+, 170.5), 282 (M + 2, 171.0)
Ack n ow led gm en t. We thank the Ministry of Edu-
cation, Science and Culture, J apan for financial support.
J O9603441
(15) Perrin, D. D.; Armarego, W. L. F.; Perrin, D. R. Purification of
Laboratory Chemicals; Pergamon: Oxford, 1980.