Borate-catalyzed reactions of hydrogen peroxide: Kinetics and mechanism of the oxidation of organic sulfides by peroxoborates

Article abstract of DOI:10.1002/chem.200401209

The kinetics of the oxidation of substituted phenyl methyl sulfides by hydrogen peroxide in borate/boric acid buffers were investigated as a function of pH, total peroxide concentration, and total boron concentration. Second-order rate constants at 25°C f

Full text of DOI:10.1002/chem.200401209

Borate-Catalyzed Reactions of Hydrogen Peroxide: Kinetics and Mechanism
of the Oxidation of Organic Sulfides by Peroxoborates
D. Martin Davies,*[a] Michael E. Deary,[a] Kieran Quill,[b] and Robert A. Smith[c]
Abstract: The kinetics of the oxidation
of substituted phenyl methyl sulfides
by hydrogen peroxide in borate/boric
acid buffers were investigated as a
function of pH, total peroxide concen-
tration, and total boron concentration.
Second-order rate constants at 258C
for the reaction of methyl 4-nitrophen-
mett 1 values for the reactions of a
range of substituted phenyl methyl sul-
fides and hydrogen peroxide, monoper-
oxoborate or diperoxoborate are
ꢀ1.50ꢁ0.1, ꢀ0.65ꢁ0.07 and ꢀ0.48
(two points only), respectively. The 1
values for the peroxoborates are of sig-
nificantly lower magnitude than ex-
pected from their reactivity compared
to other peroxides. Nevertheless the
negative values indicate positive
1
charge development on the sulfur atom
in the transition state consistent with
nucleophilic attack by the organic sul-
fides on the peroxoborates as with the
other peroxides. The kinetic parame-
ters, including the lack of reactivity of
peroxoboric acid, are discussed in
terms of the differences in the transi-
tion state of reactions involving peroxo-
boron species with respect to those of
other peroxides.
yl sulfide and H2O2, monoperoxobo-
ꢀ
rate, HOOB(OH)3
, or diperoxobo-
rate, (HOO)2B(OH)2ꢀ, are 8.29ꢀ10ꢀ5,
Keywords: boron · green chemistry ·
homogeneous catalysis · peroxides ·
redox chemistry · substituent effects
1.51ꢀ10ꢀ2 and 1.06ꢀ10ꢀ2 mꢀ1 sꢀ1, re-
spectively.
Peroxoboric
acid,
HOOB(OH)2, is unreactive. The Ham-
Introduction
more cost-effective and “green chemical” ways. Hydrogen
peroxide is activated through the formation of peroxycar-
boxylic acids, either from a reactive carboxylic acid deriva-
tive with a good leaving group (as in the peroxide bleach ac-
tivators)[3] or from the parent carboxylic acid and a strong
acid catalyst in preparative reactions.[1] Activation by the
formation of peroxycarboxylic acids entails a sharp decrease
in atom-efficiency that is avoided in true catalysis. The
range of catalysts for any given reaction of hydrogen perox-
ide is vast.[4] The properties of simple and effective catalysts
such as methyltrioxorhenium, [Re(CH3)O3], and tungstate,
the latter coupled with a phase-transfer catalyst, have re-
cently been reviewed.[5,2] The discovery of the hydrogen car-
bonate catalysed oxidation of organic sulfides and alkenes
through peroxymonocarbonate, HOOCO2ꢀ, is a significant
development.[6–8] Peroxymonocarbonate has been compared
with the methyltrioxorhenium system: the latter decomposes
above pH 2 in the presence of hydrogen peroxide, whereas
the hydrogen carbonate system has the advantage of being
stable and functional at neutral pH, although the equilibri-
um formation of peroxymonocarbonate is rather slow.[6]
Similarly, the functional pH range of tungstate is limited be-
cause the bisperoxotungstate anion forms the feebly active
dianion above pH 4. The present paper will report that
borate is a better catalyst than hydrogen carbonate, with the
Hydrogen peroxide is an atom-efficient and environmentally
benign oxidant. It is increasingly used in such areas as the
manufacture of pharmaceutical and fine chemicals, pulp,
paper and textile processing, bleaching and cleaning, waste
water treatment and environmental remediation.[1] More-
over, due to growing concerns about chemical pollutants,
and the resulting pressure against nitric acid, chlorine and
heavy-metal oxidants, there is a trend to use it even as a
bulk chemical oxidant.[1,2] Although it is a powerful oxidant,
the reactions of hydrogen peroxide are generally rather
slow, and the challenge is to overcome this kinetic barrier in
[a] Dr. D. M. Davies, Dr. M. E. Deary
Division of Chemical and Forensic Sciences
School of Applied Sciences, Northumbria University
Newcastle upon Tyne NE1 8ST (UK)
Fax : (+44)191-227-3519
E-mail: martin.davies@unn.ac.uk
[b] Dr. K. Quill
Borax Europe Ltd., 1 A Guildford Business Park
Guildford, Surrey, GU2 8XG (UK)
[c] Dr. R. A. Smith
U.S. Borax Inc., 26877 Tourney Road
Valencia, California 91355 (USA)
3552
ꢁ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200401209
Chem. Eur. J. 2005, 11, 3552 – 3558

FULL PAPER
rapidly formed peroxoborate (sometimes called peroxybo-
rate or perborate) species being more reactive than peroxy-
monocarbonate and extending the functional pH range into
the 8–12 region.[9]
higher boron concentrations there is NMR and Raman spec-
troscopic evidence for polymeric peroxoborate species.[17]
The work of Pizer and Tihal has opened the way for com-
prehensive studies of the reactions of peroxoboron species.
Despite this, and perhaps because of the earlier warnings in
the papers of Wilson, and Jones and co-workers, no rigorous
studies of their reactivity toward nucleophiles have been
published until now (rigorous studies involve a systematic
variation of the ratio of the total concentrations of boron
and peroxide; it is not sufficient merely to vary the concen-
tration of a perborate salt added to the reaction mix). The
present paper will report on the oxidation of substituted
phenyl methyl sulfides by hydrogen peroxide in boric acid/
borate buffer.
The oxidation of organic sulfides by hydrogen peroxide
and peroxyacids involves nucleophilic attack by the sulfur
atom on the (outer, in the case of peroxyacids) peroxidic
oxygen atom and is generally well understood.[18] Recent ab
initio calculations provide further insight to the mechanism
in aqueous solutions.[19] Edwards and co-workers found that
the rate of reaction of thioxane and hydrogen peroxide in
protic solvents, ROH, is overall second-order (first-order in
each reactant) and the rate constant shows a positive corre-
lation with the autoprotolysis (i.e., ion product, [ROH2]+
[RO]ꢀ) constant. Edwards concluded that specific solvent in-
teractions provide a mechanistic path that avoids significant
charge separation in the activated complex. He proposed
that the solvent acted as a proton-transfer agent to stabilise
the transition state by minimizing charge separation as
shown in Equation (4).
Peroxoborates are rapidly formed in solutions of hydro-
gen peroxide and borate and their salts are selectively crys-
tallised from solution to form perborates, which are avail-
able commercially and have synthetic as well as bleaching
and cleaning applications.[10,11] Over fifty years ago Edwards
reported that the pH of borate buffer solution drops as hy-
drogen peroxide is added, due to the rapid equilibrium for-
mation of peroxoborates.[12] Several years later Wilson sug-
gested the equilibria shown in Equations (1) and (2) for the
formation of monoperoxoborate and diperoxoborate, re-
spectively.[13]
BðOHÞ3 þ H2O2 Ð HOOBðOHÞ3ꢀ þ Hþ
ð1Þ
ð2Þ
HOOBðOHÞ3ꢀ þ H2O2 Ð ðHOOÞ2BðOHÞ2ꢀ þ H2O
He noted that the formation of peroxoborates has impor-
tant kinetic consequences in that some reactions in which
hydrogen peroxide functions as an electrophile may be sub-
stantially accelerated in the presence of a borate buffer,
giving as an example the oxidation of the anionic nucleo-
phile, SCNꢀ. He also stated, without an example, that where
hydrogen peroxide itself functions as a nucleophile, peroxo-
ꢀ
borate appears to be less reactive than HO2 (this is certain-
ly the case in the perhydrolysis of para-nitrophenyl ace-
tate).[14] Wilson concludes his paper with the warning “Use
of borate buffers in reactions which involve hydrogen perox-
ide should only be made with care”. This is reiterated in
later work by Jones and co-workers who give examples of
the “notorious behaviour of borate buffers in complicating
the kinetics of peroxide reactions”.[15] It is, on the one hand,
rather surprising, perhaps, that anionic peroxoborates
should show a high reactivity towards nucleophiles, particu-
larly anionic ones (and, as the present paper will report, es-
pecially since the uncharged, electron-deficient peroxoboric
acid, HOOB(OH)2 is less reactive). This may have led some
Moreover, in the aprotic solvent dioxane the reaction is
overall third-order, (second-order in hydrogen peroxide)
showing the participation of a second molecule of the perox-
ide as the proton-transfer agent, that is, R=OH in Equa-
tion (4).[20] Recent ab initio calculations on the oxidation of
dimethyl sulfide by hydrogen peroxide in aqueous solution
show that the extent of charge development in a transition
state involving a single water molecule is +0.37 on the
sulfur atom and ꢀ0.35 on the distal oxygen atom of hydro-
gen peroxide. It is concluded that the transition state in-
ꢀ
authors to suggest that HOOB(OH)3 can behave as a nu-
cleophile, with B(OH)3 acting as an uncharged leaving
group.[10,11] On the other hand the bisperoxotungstate mono-
anion is a well-known anionic electrophile, so the existence
of effective negatively charged electrophiles cannot be con-
tested.[2] In any case, over thirty years after the original ob-
servation by Edwards, Pizer and Tihal quantified the effect
of hydrogen peroxide on the pH of borate buffer. In addi-
tion to the equilibria represented by Equations (1) and (2),
they deduced the formation of peroxoboric acid, Equation (3).
ꢀ
ꢀ
volves O O bond breaking together with S O bond forma-
tion, and that hydrogen transfer occurs after the system has
passed the transition state. The calculated activation energy
in this case is much lower than that for direct 1,2-hydrogen
transfer of H2O2 and shows that solvent molecules can effi-
ciently lower the activation barrier for reaction. The activa-
tion barrier is, nevertheless, too high compared with experi-
mental values. It is, however, successfully predicted by using
mechanistic pathways involving hydrogen transfer through
BðOHÞ3 þ H2O2 Ð HOOBðOHÞ2 þ H2O
ð3Þ
The equilibrium constants for the three processes are
2.0ꢀ10ꢀ8, 2.0mꢀ1 and 0.01mꢀ1, respectively.[16] No other per-
oxoboron species are evident at the fairly low total boron
concentrations (less than or equal to 0.1m) used, although at
Chem. Eur. J. 2005, 11, 3552 – 3558
www.chemeurj.org
ꢁ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3553

D. M. Davies et al.
one or two water molecules, provided that an additional
water molecule stabilises the negative charge on the distal
oxygen atom of hydrogen peroxide.[19] This is shown in
Equation (5) for hydrogen transfer involving one water mol-
ecule.
Results
Unless stated otherwise, the absorbance dropped exponen-
tially with time showing that the rate of reaction was direct-
ly proportional to the concentration of the sulfide as shown
in Equation (6).
d½Sꢂ=dt ¼ kobs½Sꢂ
ð6Þ
Addition of the metal-ion chelating agents EDTA and
EDTMP, and the radical scavenger BPN had no effect on
the observed rate constant, kobs; though, at higher pH, low
concentrations (1ꢀ10ꢀ5 m) of the chelating agents were used
to suppress peroxide decomposition. Figure 1 shows the
Bunton has extensively discussed the effect of substituents
on the peroxomonosulfate oxidation of substituted phenyl
methyl sulfides in terms of strongly solvated transition
states.[21] The present work uses substituted phenyl methyl
sulfides to explore the nature of the reactivity of peroxobo-
rates and peroxoboric acid relative to hydrogen peroxide,
and with respect to previous studies of peroxycarboxylic
acids, peroxomonsulfate and peroxymonocarbonate.[22,21,7]
Experimental Section
Materials: The organic sulfides were obtained from Aldrich and aqueous
solutions were prepared and standardised as described previously.[23] Hy-
drogen peroxide, 30 wt.%, n-tert-butyl-a-phenylnitrone (BPN), the diso-
dium salt of ethylenediaminetetraacetate (EDTA), boric acid, sodium ni-
trate and sodium hydroxide were of the best available grade obtained
Figure 1. Effect of peroxide concentration on the observed rate constant
for reaction with (11 or 2.2)ꢀ10ꢀ5 m methyl 4-nitrophenyl sulfide in the
absence (pH 4.5–4.8) or presence (pH 9.6) of borate buffer; ionic
strength 0.1m (NaNO3). The curves represent Equations (7) or (8), using
the rate constants given in Table 1.
from
Aldrich.
Ethylenediaminetetramethylenephosphonic
acid
(EDTMP) was a gift from Warwick International Limited. All solutions
were made up in distilled water. Stock solutions of hydrogen peroxide
were standardised by using titanium sulfate, taking the molar absorptivity
of the resultant pertitanic complex as 730mꢀ1 cmꢀ1 at 410 nm.[24] Solutions
of boric acid and hydrogen peroxide were mixed and brought to the re-
quired pH with sodium hydroxide solution and the appropriate amount
of sodium nitrate solution added to bring the mix to the required ionic
strength.
effect of hydrogen peroxide concentration on kobs for the ox-
idation of methyl 4-nitrophenyl sulfide in the presence and
absence of borate. The upward curvature in the absence of
borate shows that, in addition to a first-order term, there is
a second-order component in the hydrogen peroxide con-
centration dependence, according to Equation (7), which
yields the rate constants kP1 and kP2, included in Table 1.
Kinetics: During kinetic runs well-sealed spectrophotometer cuvettes
were used in order to prevent volatilisation of the less polar sulfides,
which would give falsely high rate constants. All runs were carried out at
258C. Total peroxide concentration was at least ten times greater than
the sulfide concentration. Pseudo-first-order rate constants, kobs were ob-
tained from nonlinear regression of the monoexponential decrease in ab-
sorbance due to the disappearance of the sulfide. The reaction was fol-
lowed at 352 nm and 310 nm for methyl 4-nitrophenyl sulfide and methyl
4-methylsulfanylphenyl ketone, re-
2
ð7Þ
kobs ¼ kP1½H2O2ꢂ þ kP2½H2O2ꢂ
spectively, and 260 nm for the other
sulfides. Hydrogen peroxide absorbs in
Table 1. 4-XC6H4SCH3: Hammett sp values for the substituents and rate constantsꢁstandard deviations for
the UV range and so the variation of
peroxide concentration for runs in-
volving the other sulfides was limited
in order to keep the measured absorb-
ance within the dynamic range of the
spectrophotometer. Unless stated oth-
erwise, runs were carried out at ionic
strength 0.1m, under identical condi-
tions to those used by Pizer and Tihal
for the determination of the formation
constants of the peroxoboron spe-
cies.[16]
the reactions with one and two molecules of hydrogen peroxide and with mono- and diperoxoborate.
X
sp
kP1
kP2
kP1BOH
kP2BOH
[10ꢀ5 mꢀ1 sꢀ1
]
[10ꢀ5 mꢀ2 sꢀ1
]
[10ꢀ2 mꢀ1 sꢀ1
]
[10ꢀ2 mꢀ1 sꢀ1
]
NO2
0.78
0.50
0.23
ꢀ0.16
ꢀ0.17
ꢀ0.27
8.29ꢁ0.15
25.1ꢁ0.5
97.9ꢁ7.2
239ꢁ2
1.48ꢁ0.09
5.08ꢁ0.28
1.51ꢁ0.02
2.51ꢁ0.02
4.79ꢁ0.02
5.77ꢁ0.06
6.66ꢁ0.22
8.53ꢁ0.18
1.06ꢁ0.02
1.45ꢁ0.03
CH3C(O)
Br
[a]
[a]
–
–
ꢀ
[a]
[a]
CO2
–
–
–
–
[a]
[a]
CH3
CH3O
257ꢁ4
[a]
[a]
378ꢁ2
–
–
[a] Not determined under the experimental conditions used.
3554
ꢁ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 3552 – 3558

Borate-Catalyzed Reactions
FULL PAPER
In the presence of 0.05m borate there is evident catalysis,
which levels off at higher concentrations of hydrogen perox-
ide at which the borate catalyst becomes saturated with per-
oxide. The dependence of kobs on peroxide concentration
was fitted according to Equation (8), in which the subscripts
P1BOH, P2BOH and P1B refer to the reactions of the or-
ganic sulfide and monoperoxoborate, diperoxoborate and
peroxoboric acid, respectively.
pendence of kobs on pH was fitted by using Equation (8) in
exactly the same way as its dependence on peroxide concen-
tration. The best-fit values of kP1BOH and kP2BOH and their
standard deviations are [(1.4ꢁ0.2)ꢀ10ꢀ2 and (1.0ꢁ0.2)ꢀ
10ꢀ2 mꢀ1 sꢀ1, respectively] in agreement with, but less precise
than, the values from the peroxide concentration depend-
ence shown in Table 1. An upper limit of kP1B =4.1ꢀ
10ꢀ4 mꢀ1 sꢀ1 is estimated. This is calculated as 5kP1 by assum-
ing a possible 5% error in kobs and by using the formation
constant of the predominant low pH peroxoboron species,
HOOB(OH)2, as 0.01mꢀ1.
kobs ¼kP1½H2O2ꢂþkP2½H2O2ꢂ þ kP1BOH½ðHOÞ3BOOHꢀꢂþ
2
ð8Þ
kP2BOH½ðHOÞ2BðOOHÞ2ꢀꢂ þ kP1B½ðHOÞ2BOOHꢂ
Figure 3 shows the dependence of kobs on total boron con-
centration. The lines show the expected values of kobs calcu-
lated according to Equation (8) by using the rate constants
The values of kP1 and kP2 obtained in the absence of
borate were substituted into Equation (8), and the concen-
trations of the various species were calculated numerically
by using the measured pH of the reaction mix, the peroxo-
boron formation constants of Pizer and Tihal, and their
values for the acid ionisation constants of boric acid and hy-
drogen peroxide, 1.05ꢀ10ꢀ9 and 2.51ꢀ10ꢀ12 m, respective-
ly,[16] together with the mass balance equations for total
boron and total peroxide. Best-fit values of the rate con-
stants and their standard deviations, obtained by nonlinear
regression (inserting a routine in the equation editor of Gra-
fit 3.09b to perform the bisection method to determine the
concentrations of the various species[25]), are included in
Table 1. The curves in Figure 1 represent the best-fit rate
constants with kP1B set to zero.
Figure 2 shows the dependence of kobs on pH. It should be
noted at this point that the kinetic runs at low pH showed a
small, more rapid phase that comprised about 4% of the
total absorbance change independent of the concentration
of the peroxide between 0.19m and 1.91m (although in the
absence of peroxide there was no absorbance change, show-
ing that complex formation between borate and the sulfide
was not taking place); the subsequent absorbance change
paralleled that in the absence of borate catalyst. This small
initial phase was ignored in the calculation of kobs. The de-
Figure 3. Effect of total boric acid concentration on the observed rate
constant for reaction of hydrogen peroxide with 2.2ꢀ10ꢀ5 m methyl 4-ni-
trophenyl sulfide, pH 9.6. The lines represent Equation (8) using the rate
constants in Table 1.
in Table 1. There is excellent agreement between expected
and observed values for the runs carried out under identical
conditions to those of Pizer and Tihal. In the other runs, the
total boron concentration range is extended and the ionic
strength is allowed to increase along with the concentration
of borate. The small negative deviation from linearity in this
case is consistent with the small effect of ionic strength we
have observed (results not shown) and does not constitute
evidence for the presence of less reactive polymeric peroxo-
borate species.
The reaction of methyl 4-methylsulfanylphenyl ketone
showed an identical pattern of behaviour to that of the
methyl 4-nitrophenyl sulfide, and the more precise rate con-
stants obtained from the peroxide concentration depen-
dence runs are shown in Table 1. The estimated upper limit
for kP1B is 1.3ꢀ10ꢀ3 mꢀ1 sꢀ1. For the other sulfides the maxi-
mum peroxide concentration used was between 0.08 and
0.16m, and the contributions of second-order terms in perox-
ide concentration to the observed rate constants were negli-
gible. Best-fit rate constants are shown in Table 1. The data
are shown in Figure 4 as Hammett plots using para-substitu-
ent constants.[26] For hydrogen peroxide the slope (Hammett
Figure 2. Effect of pH on the observed rate constant for reaction of two
different concentrations of hydrogen peroxide with 1ꢀ10ꢀ4 m methyl 4-ni-
trophenyl sulfide in borate buffer; [boric acid]0 0.05m, ionic strength 0.1m
(NaNO3). The curves represent Equation (8), using the rate constants
given in the text (full lines) and Table 1 (broken lines).
Chem. Eur. J. 2005, 11, 3552 – 3558
www.chemeurj.org
ꢁ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3555

D. M. Davies et al.
strate against the pKa of the parent acid of the peroxide.
The latter quantity is used as an indicator of the leaving
group character of the parent anion formed by heterolytic
cleavage of the peroxide bond, and the greater the slope of
the plot the greater the significance of peroxide bond cleav-
age in the transition state.[18] Figure 5 shows the Brønsted
Figure 4. Hammett plots for the reactions of peroxoborates or hydrogen
peroxide and substituted phenyl methyl sulfides.
1 value) is ꢀ1.5ꢁ0.1, whereas for monoperoxoborate it is
ꢀ0.65ꢁ0.07, and for diperoxoborate it is ꢀ0.48 (two points
only).
Figure 5. Brønsted-type plot, including parent Lewis acids, for the reac-
tions of peroxides and methyl 4-nitropheny sulfide.
Discussion
Oxidation by hydrogen peroxide: For the oxidation of
methyl 4-nitrophenyl sulfide and methyl 4-methylsulfanyl-
phenyl ketone, for which the use of a high concentration of
peroxide was possible, a second-order term in its concentra-
tion appears in the rate equation, [Eq. (7)]. This has been
observed previously not only in aprotic solvents, but also in
alcohol/water mixtures.[20,6] The high concentration of hydro-
gen peroxide competes with the solvent water, either as the
proton-transfer agent or the hydrogen-bond donor (or both)
in the transition state, see Equations (4) and (5). To com-
pare the stability of the transition state containing one hy-
drogen peroxide group plus at least one water molecule
with that of the transition state containing two molecules of
hydrogen peroxide, it is necessary to divide kP1 by 55m, the
concentration of bulk water. The third-order rate constants
obtained are 1.5ꢀ10ꢀ6 and 4.6ꢀ10ꢀ6 mꢀ2 sꢀ1 for methyl 4-ni-
trophenyl sulfide and methyl 4-methylsulfanylphenyl
ketone, respectively. These values are about ten times less
than the respective values of kP2, in keeping with H2O being
a poorer proton donor/acceptor, and a weaker acid and hy-
drogen-bond donor than H2O2.
plot for methyl 4-nitrophenyl sulfide, including literature
data for its reactivity with peroxomonosulfate, peroxycar-
boxylic acids and peroxymonocarbonate,[22,21,7] as well as the
present results for hydrogen peroxide and monoperoxobo-
rate and diperoxoborate. The pKa of HOOB(OH)2, the
parent acid of diperoxoborate, is 5.7, compared with 8.98
for B(OH)3, the parent acid of monoperoxoborate (these
values are consistent with HOOꢀ being less electron donat-
ing than HOꢀ).[16] Figure 5 shows that the peroxoborates are
less reactive than peroxycarboxylic acids, though this disad-
vantage lessens with increasing pH above their pKa values,
at which the peroxycarboxylic acids become ionised. The
peroxoborates are, however, more reactive than peroxy-
monocarbonate, and this advantage increases with increas-
ing pH above the pKa of hydrogen carbonate.
It should be noted at this point that the parent acids of
the peroxoborates are Lewis acids and their pKa values
denote something different from those of the other
(Brønsted) acids in Figure 5. Thus, decreasing pKa of a
parent Brønsted acid (a measure of the equilibrium between
XOH and XOꢀ +H+, in which XOOH is the peroxide) im-
plies easier cleavage of the peroxide oxygen–oxygen bond
due to the better leaving group ability of XOꢀ, whereas de-
creasing pKa of the Lewis acids (a measure of the equilibri-
um between X+H2O and XOHꢀ +H+, in which XOOH is
the peroxide) implies more difficult cleavage of the boron
peroxide bond due to the poorer leaving group ability of X.
The line on the plot is drawn through the data points for the
peroxycarboxylic acids with Brønsted slope blg =ꢀ0.65.[22] It
is evident that peroxomonosulfate, peroxymonocarbonate
and hydrogen peroxide lie near to the line, suggesting that
their reactions are mechanistically quite similar and oxygen–
The negative Hammett 1 value for the reaction of substi-
tuted phenyl methyl sulfides and hydrogen peroxide is indi-
cative of the development of positive charge on the sulfur in
the transition state, with the sulfide behaving, as expected,
as a nucleophile.[18] The actual value of 1, ꢀ1.5, is commen-
surate with the difference in charge of the sulfur in the tran-
sition state and reactant cluster, +0.37, calculated for the
oxidation of dimethyl sulfide.[19]
Oxidation by peroxoborates: To compare the reactivity of
peroxides it is usual to construct a Brønsted-type plot of the
log of the rate constant for reaction with a particular sub-
3556
ꢁ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 3552 – 3558

Borate-Catalyzed Reactions
FULL PAPER
oxygen bond breaking is significant in the transition state.
Monoperoxoborate also lies close to the line, although this
is almost certainly fortuitous since the diperoxoborate does
not lie equally close. The rate constants for the peroxobo-
rates are very similar (dashed line on Figure 5) despite their
different pKa values. This shows that boron–oxygen bond
breaking is not significant in the transition state, consistent
with the peroxoborates not acting as nucleophiles in which
the neutral B(OH)3 or HOOB(OH)2 moieties are the re-
spective leaving groups.
than expected from their reactivity compared with the other
peroxides. This indicates that the transition states for their
reactions lie on a region of the reaction coordinate at which
at least some of the relative energy changes involved in
sulfur–oxygen bond formation, oxygen–oxygen bond break-
ing, proton transfer and solvation changes are significantly
different due to the presence of the boron atom. It is neces-
sary that any mechanism that is proposed to account for the
very different behaviour of the peroxoborates must also ac-
commodate the very low reactivity of peroxoboric acid.
A comparison of the Hammett 1 values of the peroxobo-
rates and other peroxides gives further insight into the
nature of the transition state. Figure 6 shows a reactivity se-
Peroxoboric acid: The upper limits of the rate constants for
the reaction of peroxoboric acid and the two sulfides investi-
gated are very small relative to the rate constants for the
peroxoborates. At first sight this seems contrary to what is
expected, because if the peroxoborates are acting as electro-
philes then peroxoboric acid, with its absence of a negative
charge and, moreover, the presence of the electron-deficient
boron, ought to be a better electrophile. If, however, proton
transfer provides a mechanistic path for the reaction of the
peroxoborates, as shown in the transition state structure 1
(in which the extent of charge development on the sulfur
and respective peroxide oxygen atoms in the transition
state, and any additional specific solvent interactions, are
not shown) then this path is not available to peroxoboric
acid. This is because proton transfer is less favoured due to
the positive charge located on the oxygen atom of 2, associ-
Figure 6. Reactivity–selectivity plot for the reactions of peroxides and
substituted phenyl methyl sulfides, with the 4-nitro derivative as the reac-
tivity measure.
lectivity plot including the present data and that reported
for peroxymonocarbonate, peroxycarboxylic acids and per-
oxomonosulfate.[7,22,21] This plot uses the Hammett 1 values
as a measure of the selectivity of the peroxide with respect
to the nature of the substituent on the phenyl group, and
the logarithm of the rate constant for the reaction with the
nitro derivative as a measure of the reactivity of the perox-
ide. The plot shows that the peroxoborates behave very dif-
ferently to the other peroxides. The much less negative
Hammett 1 values indicates that there is significantly less
positive charge development on the sulfur atom in the tran-
sition state. The essentially linear reactivity–selectivity rela-
tionship including hydrogen peroxide, peroxymonocarbon-
ate, the peroxycarboxylic acids and peroxomonosulfate is
consistent with all these peroxides reacting with the sulfides
by the same mechanism. This is because, according to the
Hammond postulate, the structure of the transition state be-
comes closer to that of the reactant cluster as the energy dif-
ference between the transition state and reactants becomes
smaller.[27] In the present case the more reactive peroxides
have a transition state earlier along the reaction coordinate
with less positive charge development on the sulfur atom. In
the case of the peroxoborates, they are much less selective
ated with the partial double-bond character of the boron–
oxygen bond. Hence, proton transfer as a significant part of
the reaction coordinate for the peroxoborates is strongly
suggested, because of the lack of reactivity of peroxoboric
acid. We propose that the much lower extent of positive
charge development on the sulfur atom for the reactions of
the peroxoborates, as shown by the lower magnitude of the
Hammett 1 values in Figure 6, is due to the greater impor-
tance of proton transfer in stabilising the transition state in
the case of the peroxoborates.
Finally it is worth noting that the alternative nucleophilic
attack of the sulfur atom on the inner peroxide oxygen atom
in 3 to form a positively charged intermediate could possibly
Chem. Eur. J. 2005, 11, 3552 – 3558
www.chemeurj.org
ꢁ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3557

D. M. Davies et al.
be subject to neighbouring group assistance by boron that is
well known in organoboron compounds.[28] Indeed, as de-
scribed in the results section, the kinetic runs at low pH in
the presence of boric acid showed a short rapid initial phase
consistent with the formation of an intermediate that breaks
down considerably more slowly than the rate of the uncata-
lysed reaction.
transfer in the reaction involving peroxoboric acid is less
favourable due to the positive charge on the acceptor
oxygen atom associated with the partial double-bond
character of the boron–oxygen bond.
[1] C. W. Jones, Applications of Hydrogen Peroxide and Derivatives,
The Royal Society of Chemistry, Cambridge, 1999.
[2] R. Noyori, M. Aoki, K. Sato, Chem. Commun. 2003, 1977–1986.
[3] D. M. Davies, S. J. Foggo, P. M. Paradis, J. Chem. Soc. Perkin Trans.
2 1998, 1381–1385; D. M. Davies, M. E. Deary, J. Chem. Soc. Perkin
Trans. 2 1992, 559–562; D. M. Davies, M. E. Deary, J. Chem. Soc.
Perkin Trans. 2 1991, 1549–1552.
Conclusion
The following conclusions can be drawn.
1) The reaction of hydrogen peroxide and substituted
phenyl methyl sulfides shows a negative Hammett 1
value consistent with the development of positive charge
on the nucleophilic sulfur atom, as predicted by recently
published ab initio calculations on a transition state in-
[4] B. S. Lane, K. Burgess, Chem. Rev. 2003, 103, 2457–2473.
[5] J. H. Espenson, Chem. Commun. 1999, 479–488.
[6] D. E. Richardson, H. Yao, K. M. Frank, D. A. Bennett, J. Am.
Chem. Soc. 2000, 122, 1729–1739.
[7] D. A. Bennett, H. Yao, D. E. Richardson, Inorg. Chem. 2001, 40,
2996–3001.
ꢀ
ꢀ
[8] H. Yao, D. E. Richardson, J. Am. Chem. Soc. 2000, 122, 3220–3221.
[9] A preliminary account of this work was presented at the 227th
American Chemical Society National Meeting, Anaheim, California,
USA, 28th March–1st April, 2004.
[10] D. M. Schubert, Struct. Bonding (Berlin) 2003, 105, 1–40.
[11] A. McKillop, W. R. Sanderson, Tetrahedron 1995, 51, 6145–6166.
[12] J. O. Edwards, J. Am. Chem. Soc. 1953, 75, 6154–6155.
[13] I. R. Wilson, Aust. J. Chem. 1960, 13, 582–584.
[14] D. M. Davies, M. E. Deary, J. Chem. Res. Synop. 1988, 354–355.
[15] S. B. Brown, P. Jones, A. Sugget, Prog. Inorg. Chem. 1970, 13, 159–
204.
volving S O bond formation and O O bond breaking.
The calculations also predict that transfer of hydrogen to
the distal oxygen atom of hydrogen peroxide through
one or more solvent water molecules, although an impor-
tant part of the reaction coordinate, occurs after the
system has passed the transition state, so that the devel-
oping negative charge on the distal oxygen atom is stabi-
lised by hydrogen bonding from an additional water mol-
ecule.
[16] R. Pizer, C. Tihal, Inorg. Chem. 1987, 26, 3639–3642.
[17] J. Flanagan, W. P. Griffith, R. D. Powell, A. P. West, J. Chem. Soc.
Dalton Trans. 1989, 1651–1655; B. N. Chernyshov, Russian J. Inor.
Chem. 1990, 35, 1333–1335.
[18] R. Curci, J. O. Edwards, in Organic Peroxides, Vol 1 (Ed: D. Swern),
Wiley, New York, 1970, pp. 199–264; Y. Sawaki, in Organic Perox-
ides (Ed.: W. Ando), Wiley, Chichester, 1992, pp. 426–477.
[19] J.-W. Chu, B. L. Trout, J. Am. Chem. Soc. 2004, 126, 900–908.
[20] M. A. P Dankleff, R. Curci, J. O. Edwards, H.-Y. Pyun, J. Am.
Chem. Soc. 1968, 90, 3209–3218.
[21] C. A. Bunton, H. J. Foroudian, A. Kumar, J. Chem. Soc. Perkin
Trans. 2 1995, 33–39.
[22] D. M. Davies, M. E. Deary, J. Chem. Soc. Perkin Trans. 2 1996,
2423–2430.
[23] D. M. Davies, M. E. Deary, . Chem. Soc. Perkin Trans. 2 1995, 1287–
1294.
[24] G. M. Eisenburg, Ind. Eng. Chem. 1943, 15, 327–328.
[25] A. C. Norris, Computational Chemistry, Wiley, Chichester, 1981,
p. 74.
2) At high concentrations of hydrogen peroxide a pathway
involving catalysis by a second molecule of the peroxide
occurs. This is consistent either with hydrogen peroxide
being a better proton donor/acceptor than water, or it
being a better hydrogen-bond donor, or both.
3) The reaction of hydrogen peroxide and substituted
phenyl methyl sulfides is catalysed in aqueous boric acid/
borate buffer through the formation of monoperoxobo-
rate and diperoxoborate anions.
4) The peroxoborate anions are stable and functional in the
pH range 8–12.
5) The peroxoborates are less reactive than peroxomono-
sulfate and peroxycarboxylic acids, but more reactive
than peroxymonocarbonate.
6) The peroxoborate anions, in common with other perox-
ides, are subject to nucleophilic attack of the sulfur atom
on the outer peroxide oxygen atom, although the devel-
opment of positive charge on the sulfur atom is much
less in the case of the peroxoborates.
7) Peroxoboric acid is much less reactive than the peroxo-
borates. This is strong evidence for the importance of
proton transfer through a solvent water molecule in the
reaction coordinate for the peroxoborates, since proton
[26] C. Hansch, A. Leo, R. W. Taft, Chem. Rev. 1991, 91, 165–195.
[27] A. Williams, Free Energy Relationships in Organic and Bio-organic
Chemistry, The Royal Society of Chemistry, Cambridge, 2003,
pp. 135–138.
[28] D. S. Matteson, in progress in Boron Chemistry, Vol. 3 (Eds: R. J.
Brotherton, H. Steinberg), Pergamon, New York, 1970, pp. 117–176.
Received: November 25, 2004
Published online: April 13, 2005
3558
ꢁ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 3552 – 3558