Degradation of 2,5-dihydroxy-1,4-benzoquinone by hydrogen peroxide: A combined kinetic and theoretical study

Article abstract of DOI:10.1021/jo4001178

2,5-Dihydroxy-1,4-benzoquinone (DHBQ) is one of the key chromophores formed upon aging in cellulosic materials. This study addresses the degradation mechanism of DHBQ by hydrogen peroxide to provide a solid knowledge base for optimization of bleaching sequences aiming at DHBQ removal. Kinetic analysis provided an activation energy (E_a) of 20.4 kcal/mol. Product analyses indicated the product mixture to contain malonic acid, acetic acid, and carbon dioxide. DFT(B3LYP) computation presented a plausible mechanism for the formation of these products from DHBQ. DHBQ forms intermediate I2k, having an intramolecular O-O bridge between C-2 and C-5 of the 1,4-cyclohexadione structure. This O-O bond is homolytically cleaved, and the subsequent β-fragmentation of the resulting radical forms ketene and oxaloacetic acid. While ketene yields acetic acid, oxaloacetic acid then gives malonic acid and carbon dioxide through further attack of hydrogen peroxide via an intermediate that is oxidatively decarboxylated. The calculated E_a value (23.3 kcal/mol) in the rate-determining step, i.e., the homolysis of I2k, agreed well with the experimental value. There is also a minor pathway in which the spin state changes to triplet during the homolysis of I2k; in this way two malonyl radicals are formed that are converted to two molecules of malonic acid.

Full text of DOI:10.1021/jo4001178

Article
pubs.acs.org/joc
Degradation of 2,5-Dihydroxy-1,4-benzoquinone by Hydrogen
Peroxide: A Combined Kinetic and Theoretical Study
Takashi Hosoya and Thomas Rosenau*
Department of Chemistry, University of Natural Resources and Life Sciences, Muthgasse 18, A-1190 Vienna, Austria
*
S Supporting Information
ABSTRACT: 2,5-Dihydroxy-1,4-benzoquinone (DHBQ) is one of the key chromophores
formed upon aging in cellulosic materials. This study addresses the degradation
mechanism of DHBQ by hydrogen peroxide to provide a solid knowledge base for
optimization of bleaching sequences aiming at DHBQ removal. Kinetic analysis provided
an activation energy (E ) of 20.4 kcal/mol. Product analyses indicated the product mixture
a
to contain malonic acid, acetic acid, and carbon dioxide. DFT(B3LYP) computation
presented a plausible mechanism for the formation of these products from DHBQ. DHBQ
forms intermediate I2k, having an intramolecular O−O bridge between C-2 and C-5 of
the 1,4-cyclohexadione structure. This O−O bond is homolytically cleaved, and the subsequent β-fragmentation of the resulting
radical forms ketene and oxaloacetic acid. While ketene yields acetic acid, oxaloacetic acid then gives malonic acid and carbon
dioxide through further attack of hydrogen peroxide via an intermediate that is oxidatively decarboxylated. The calculated Ea
value (23.3 kcal/mol) in the rate-determining step, i.e., the homolysis of I2k, agreed well with the experimental value. There is
also a minor pathway in which the spin state changes to triplet during the homolysis of I2k; in this way two malonyl radicals are
formed that are converted to two molecules of malonic acid.
INTRODUCTION
Scheme 1. Structures of 2,5-Dihydroxy-1,4-benzoquinone
DHBQ) and Its Resonance-Stabilized Dianion
■
(
2
,5-Dihydroxy-1,4-benzoquinone (DHBQ) has recently been
reported as one of the most important compounds that cause
discoloration in cellulosic materials. It is a nearly ubiquitous
cellulosic chromophore as it is a prime survivor of bleaching
treatments, and it is also found in aged cellulosic matrices as it
is regenerated from low-molecular weight breakdown products
1
of cellulose aging. The reason why DHBQ is rather resistant
toward bleaching chemicals and ″survives″ industrial pulp
bleaching sequences much longer than other aromatic or
quinoid chromophores is its exceptional resonance stabilization.
In alkaline media it forms a dianion with highly delocalized
double bonds. As most common bleaching agents (ozone,
hydrogen peroxide, chlorine dioxide) act primarily by attacking
double bonds, this explains the increased resistance of the
However, these works have employed DHBQ as a starting
material in organic synthesis. With the revival of the compound
that is connected to its pivotal role in cellulose yellowing, the
degradation and ″discoloration″ of DHBQ became a matter of
much interest. It is thus important to investigate the detailed
behavior of DHBQ toward bleaching agents and elucidate the
compound’s degradation mechanisms. In this report, we
communicate a clarification of the detailed mechanism of the
reaction between DHBQ and hydrogen peroxide as one of the
most common bleaching agents, combining experimental
investigation with computational studies (DFT calculations).
1
compound toward oxidative destruction and removal. The
chemistry of DHBQ has thus become a point of great interest
from several viewpoints: the pulp and paper industries are
interested in better, i.e., faster and cheaper, removal of the
compound from cellulosic pulps; conservationists working with
historic cellulosic matrices are looking for mild and compatible
ways for DHBQ destruction; and cellulose scientists search for
good methods of DHBQ detection and quantification of its
effects on cellulose properties. A better understanding of the
reactivity of DHBQ is thus quite important for a general
understanding of chemical processes in cellulose aging,
bleaching, and general cellulose science. Scheme 1 gives the
structure of DHBQ, and its resonance-stabilized dianion.
Many previous accounts have addressed the chemistry o2f
DHBQ, which mainly undergoes electrophilic substitution,
RESULTS AND DISCUSSION
■
Kinetic Analysis and Product Characterization. For
kinetic studies of the reaction of DHBQ with hydrogen
peroxide, a reaction system as simple as possible was chosen,
using aqueous, nonbuffered solutions of both reagents. In
principle, the reaction follows a second-order rate law.
However, if an excess amount of hydrogen peroxide is used
Received: January 18, 2013
Published: February 18, 2013
3
4
nucleophilic substitution, and condensation with amines.
©
2013 American Chemical Society
3176
dx.doi.org/10.1021/jo4001178 | J. Org. Chem. 2013, 78, 3176−3182

The Journal of Organic Chemistry
as under our standard conditions: 257 molar equiv of H O2
Article
(
2
relative to DHBQ; see Experimental Details), the rate law
assumes a pseudo first order. In this case, the concentration of
DHBQ [DHBQ] at reaction time t is presented by the
following equation:
−
kt
[
DHBQ] = [DHBQ] e
(1)
0
where [DHBQ] is the initial concentration of DHBQ, and k is
0
the pseudo-first-order rate constant. We can determine k from
the slope of the curve that plots ln[DHBQ] versus t. Figure 1
Figure 2. Arrhenius plot of the degradation reaction in the
temperature range from 313.15 to 353.15 K.
mol) were very similar to E , while the activation Gibbs
a
⧧
energies (Δ G° = 24.9 to 25.3 kcal/mol) were substantially
⧧
⧧
higher than Δ H° due to negative activation entropies (Δ S° =
⧧
−
14.7 to −16.3 cal/(mol·K)). These negative Δ S° values
indicate a higher order in the transition state of the rate-
determining step.
1
Through product analysis by H NMR after freeze-drying, we
found malonic acid to be the major product from DHBQ, as
indicated by the singlet at 3.2 ppm, shown in Figure 3A. The
presence of malonic acid was also confirmed by GC−MS
analysis; see Figure S1 in Supporting Information for the total-
ion chromatogram and the mass fragmentation pattern of the
trimethylsilyl (TMS) derivative. The carbon-based yield of
malonic acid was not quantitative but was almost constant
under the employed reaction conditions (44.4−55.1%); see
Table S1 in Supporting Information for the yield of malonic
acid under each reaction condition. This result indicates that
H O -degradation of DHBQ forms other major products that
Figure 1. Relationship between ln[DHBQ] and reaction time t in the
temperature range from 313.15 to 353.15 K.
shows such a plot in the range of 313.15−353.15 K. It is evident
that there is a linear relationship between ln[DHBQ] and t,
indicating correctness of the pseudo-first-order approximation.
The pseudo-first-order rate constant k follows at each
temperature from the slope of the lines; the determined k
−4
−2 −1
values ranged from 4.66 × 10 to 2.74 × 10 s , as
2
2
summarized in Table 1.
1
were undetectable by either H NMR or GC−MS after freeze-
drying.
Table 1. Experimental Pseudo-First-Order Rate Constants
1
To identify those byproducts, in situ H NMR analysis of the
(
k) and Activation Parameters in the Degradation of DHBQ
reaction in D O was performed. As shown in Figure 3B, a
2
by H O in Non-buffered Aqueous Solution
2
2
1
singlet at 1.88 ppm appeared, which did not occur in the H
a
activation parameter
NMR spectrum (DMSO-d ) after freeze-drying (Figure 3A).
6
⧧
⧧
⧧
From its chemical shift, the signal was assigned to acetic acid
(confirmed by spiking), which was volatile enough to be
evaporated from the product mixture during the freeze-drying
process. It was likely that acetic acid was not produced as the
decarboxylation product of malonic acid, but directly from
DHBQ, since malonic acid was found to be considerably stable
under the reaction conditions and formed only rather minor
amounts of acetic acid at 313.15 K over a prolonged reaction
time (1 day); see Figure 3C. The acetic acid formed was not
quantified as we were unable to find a suitable internal standard5
that additionally was stable enough toward hydrogen peroxide.
Adding the reaction mixture into a saturated aqueous
solution of calcium hydroxide resulted in formation of white
precipitate; see Figure S2 in the Supporting Information. This
indicated the presence of carbon dioxide (carbonate) in the
reaction mixture. Since carbon dioxide is also formed from
malonic acid (decarboxylation) and acetic acid (minor
temp
Δ H°
Δ S°
Δ G°
(
K)
k (s−1)
(kcal/mol)
(cal/mol·K)
(kcal/mol)
4.66 × 10−
4
3
3
3
2
313.15
323.15
333.15
343.15
353.15
19.9
19.8
19.8
19.8
19.8
−16.3
−14.7
−14.9
−16.2
−15.9
24.9
24.5
24.8
25.3
25.3
3.01 × 10−
−
7.30 × 10
9.74 × 10−
2.74 × 10−
a
The activation energy (E ) was determined to be 20.4 kcal/mol.
a
Figure 2 shows the Arrhenius plot, the logarithm of the
pseudo-first-order rate constants k (ln k) versus the reciprocal
of the reaction temperature (K), having a linear relationship
that allows determination of the various activation factors from
this plot. The activation energy (E ) of the reaction of DHBQ
with H O was determined to be 20.4 kcal/mol, as presented in
a
2
2
⧧
Table 1. The activation enthalpies (Δ H° = 19.8 to 19.9 kcal/
3
177
dx.doi.org/10.1021/jo4001178 | J. Org. Chem. 2013, 78, 3176−3182

The Journal of Organic Chemistry
Article
1e
been found in crystal structures. We concluded that we can
employ the DFT(RB3LYP) method for the calculation of the
equilibrium structures and the transition state of ionic reactions
and the DFT(UB3LYP) method for the calculation of the
radicals and the transition state of homolytic processes in the
reaction system; see the Supporting Information for a detailed
discussion of the accuracy of the DFT(B3LYP) method.
Figure 4 shows the overall reaction pathway derived from the
DFT(B3LYP) computations: DHBQ (6 carbon atoms) gives
malonic acid (3 carbon atoms), acetic acid (2 carbon atoms),
and carbon dioxide (1 carbon atom) as reaction products. In
the initial step of this degradation, hydrogen peroxide attacks
C-2 of DHBQ in a nucleophilic manner to form intermediate
I1e via transition state TS1. In this nucleophilic attack, a proton
relay from hydrogen peroxide to the C-4 carbonyl group occurs
6
via three water molecules, as shown in Scheme 2; see also
Figure S5 in Supporting Information for detailed geometry
⧧
changes. The E and the Δ H° of this step were calculated to be
a
⧧
1
4.0 and 11.5 kcal/mol, respectively. The Δ G° (28.7 kcal/
⧧
mol) of this step is much higher than the Δ H° due to the
decrease in entropy by the bond formation between DHBQ
and hydrogen peroxide.
After I1e is converted into more stable I1k, which has a keto
7
structure, O-7′ of I1k nucleophically attacks C-5 to form an
intermediate I2e with 2,5-dihydoxy-1,4-cyclohexadione struc-
ture, which has an intramolecular O−O bridge between C-2
and C-5. This intramolecular substitution involves the
formation of an anionic intermediate IA via TS2 and the
1
nucleophilic attack of anionic O-7′ of IA at C-5 via TS2 , as
2
shown in Scheme 3 and Figure 4; see also Figure S6 in
Supporting Information for detailed geometry change. E and
a
⧧
Δ H° of this process, i.e. the values for TS2 , are 12.5 and 9.8
2
1
⧧
Figure 3. H NMR spectrum in DMSO-d of (A) the product mixture
6
kcal/mol, respectively, while Δ G° (26.6 kcal/mol) is once
after freeze-drying (313.15 K/300 s), (B) the in situ reaction mixture
more much higher due to the decreased entropy compared to
from DHBQ in D O (313.15 K/300 s), and (C) the reaction mixture
2
that of the reactants. I2e then gives the more stable I2k through
in the degradation of malonic acid in D O (315.15 K/1 day). Note:
7
2
keto−enol interconversion. We also found another reaction
a)
IS = internal standard (4-bromobenzaldehyde).
pathway starting from I1k, where initial homolysis of the O−O
bond of I1k occurs, but the E of this pathway is so high (29.6
a
oxidation) under the same conditions, it was unclear at this
moment whether carbon dioxide was produced directly from
DHBQ. However, in the next section, we will clarify this issue,
showing that carbon dioxide is already formed in the
degradation of DHBQ to malonic acid.
As a summary of the experiments, we found that DHBQ
degrades mainly into malonic acid with a carbon yield of ∼50%
in the aqueous hydrogen peroxide solution. Acetic acid and
carbon dioxide were also detected but were not quantified. The
kcal/mol) that this pathway would in fact be negligible; see
Figure S7 for the transition state geometry and its energy.
The following step of the overall degradation includes the
homolysis of the O−O bond of I2k and the β-fragmentation of
the resulting radical to form ketene and oxaloacetic acid (I3)
via TS3, as shown in Figure 4. See also the schematic reaction
in Scheme 4, the detailed mechanism of which will be discussed
later. Ketene is so reactive that it will be readily attacked by
water to give acetic acid, which was also detected in the
experimental product analysis; see previous section. The
activation parameters from the Arrhenius plot were E = 20.4
a
⧧
⧧
kcal/mol, Δ H° = 19.8 to 19.9 kcal/mol, Δ S° =−14.7 to
DFT(UB3LYP)-calculated energy of TS3 (E
a
= 23.3 kcal/
mol, Δ H° = 22.4 kcal/mol, Δ G° = 33.4 kcal/mol) is
considerably higher than those of TS1, TS2 , and TS2 , and
⧧
⧧
⧧
−
16.3 cal/mol·K, and Δ G° = 24.9 to 25.3 kcal/mol.
Computational Study (DFT). Since the kinetic experi-
1
2
ments provided only limited information about the detailed
mechanism of the degradation, we carried out quantum
chemical computations at the DFT(B3LYP) level of theory
to elucidate the sequence of elemental reaction steps. First, a
test calculation for hydrogen peroxide was carried out to check
the performance of the B3LYP functional for the description of
the properties of the O−O bond: the DFT(RB3LYP) method
correctly presented the geometry of hydrogen peroxide, and the
DFT(UB3LYP) method well described the potential energy
curve of the cleavage of the O−O bond. Previous work showed
that the B3LYP/6-31+G** level of theory successfully
reproduced the varying bond lengths within DBHQ that have
this process is extremely exothermic; ΔE = −95.0 kcal/mol,
0
0
ΔH = −96.0 kcal/mol, ΔG = −91.0 kcal/mol. From these
results, we concluded that this process evidently is the rate-
determining step in the overall degradation. The calculated Ea
⧧
(23.3 kcal/mol) and Δ H° (22.4 kcal/mol) were in quite
satisfying agreement with those obtained from kinetic analysis
(see Table 1), indicating the correctness of the computed
⧧
pathway. Though the calculated Δ S° (−36.6 kcal/mol) value
is not in quantitative agreement with that of the experiment
⧧
(−16.3 kcal/mol, see Table 1), the largely negative Δ S° was
8
well reproduced by the DFT(B3LYP) computation. Note that
geometry optimization of the biradical produced from the O−
3
178
dx.doi.org/10.1021/jo4001178 | J. Org. Chem. 2013, 78, 3176−3182

The Journal of Organic Chemistry
Article
Figure 4. Reaction pathway from DHBQ to malonic acid calculated with DFT(B3LYP) method. Notes: a) Energies are presented in kcal/mol. b)
Bond lengths are presented in Å. c)The energies of these transition states and intermediates are calculated relative to the DHBQ-3H O complex
2
d)
instead of DHBQ; see Figure S4 in Supporting Information for the geometry of the DHBQ-3H O complex. DFT(UB3LYP) method was
2
employed for the energy evaluation of TS3; see also Figure S3 in Supporting Information.
Scheme 2. Nucleophilic Attack of Hydrogen Peroxide to
DHBQ
Scheme 3. Intramolecular Nucleophilic Attack in I1k To
Form I2e
kcal/mol, respectively, relative to oxaloacetic acid (I3). I4 then
O bond cleavage of I2k was not successful, resulting in the
reformation of the O−O bond again. This means that the
biradical is not a stable species and exists only transitorily in the
reaction course from I2k to ketene and I3.
The next step of the sequence is a further nucleophilic attack
of hydrogen peroxide to the carbonyl group of oxaloacetic acid
releases carbon dioxide and water to yield malonic acid with an
⧧
⧧
E of 15.2 kcal/mol, Δ H° of 12.9 kcal/mol, and Δ G° of 29.6
a
kcal/mol via TS5; see Figure S9 in Supporting Information for
detailed geometry changes. The activation barrier in TS5 is
higher than that in TS4. The step from I4 to malonic acid and
carbon dioxide via TS4 is thus the rate-determining step with
(
I3) to form intermediate I4 via TS4; see Figure S8 in
regard to the degradation of oxaloacetic acid (I3). The
⧧
Supporting Information for detailed geometry changes. The E ,
Δ H°, and Δ G° values of this process were 8.9, 6.2, and 23.3
activation barrier to TS5 (E = 15.2 kcal/mol, Δ H° = 12.9
a
a
⧧
⧧
⧧
kcal/mol, Δ G° = 29.6 kcal/mol) is much lower than that to
3
179
dx.doi.org/10.1021/jo4001178 | J. Org. Chem. 2013, 78, 3176−3182

The Journal of Organic Chemistry
Article
Scheme 4. Schematic Description of the Mechanism in the
Rate-Determining Step
⧧
⧧
TS3 (E = 23.3 kcal/mol, Δ H° = 22.4 kcal/mol, Δ G° = 33.4
a
kcal/mol, see Figure 4). Hence, it is reasonable to assume that
oxaloacetic acid (I3), once formed, is rapidly degraded and was
thus not detected in the experiment. It should be also noted
that the reaction from oxaloacetic acid (I3) to malonic acid is
known as a general method for oxidative decarboxylation of an
α-keto acid and that a previous experimental study about its
degradation mechanism well supports the presented mecha-
Figure 6. NBO charge and Mulliken spin density of important atoms
of TS3, calculated with the DFT(UB3LYP) method. Note: The
values in parentheses are the difference in NBO charge relative to I2K.
a)
9
nism.
of the C-1-centered acyl radical to form ketene and oxaloacetic
acid (I3), as shown in Scheme 4.
In the overall reaction, homolysis of the O−O bond of I2k
and the subsequent β-fragmentation of the resulting radical via
TS3 is the most important process, which is also the rate-
determining step as discussed above. Therefore, we should like
to address the detailed mechanism of this step. Figure 5 shows
the geometry changes in the reaction from I2k to ketene and
oxaloacetic acid (I3). From I2k to TS3, the O-7−O-7′ bond is
considerably elongated from 1.483 to 2.091 Å, as is the C-1−C-
According to the mechanism discussed above, the carbon-
based yield of malonic acid must be exactly 50%, since one
molecule of DHBQ gives one molecule of malonic acid.
However, as mentioned in the previous section, the maximum
yield of malonic acid was slightly above 50% in some cases; see
also Table S1 in Supporting Information. This fact led us to
investigate an alternative reaction pathway in which one
molecule of DHBQ yielded two molecules of malonic acid.
As shown in the potential energy curve for the O-7−O-7′
distance of I2k in Figure 7, the potential energy of the singlet
state is increased by the elongation of the O-7−O-7′ bond,
while that of the triplet state becomes smaller. These singlet
and triplet potential energy curves meet around a O-7−O-7′
bond length of 2.0 Å. The spin state of the reactant is thus
likely to change from singlet to triplet through a conical
intersection when the O-7−O-7′ bond is elongated to around
2.0 Å; see CI in Figure 7. When the spin state becomes a triplet,
furthermore, the reactant produces two molecules of malonyl
radical, as shown in Figure 7, and the malonyl radical would be
2
bond from 1.540 to 1.836 Å, while the O-7−C-2 distance is
shortened form 1.478 to 1.301 Å. Despite those large geometry
changes, the NBO charges of several important atoms are not
changed very much, as shown in Figure 6. These results
indicate that O-7−O-7′ cleavage is homolytic (radical) rather
than heterolytic (ionic). It is noted that the C-1−C-6, C-5−C-
6, and C-5−O-7′ bonds are not very different in length between
I2k and TS3. It is thus likely that the reaction from I2k to TS3
is the process where the homolysis of the O−O bond results in
the formation of an O-7′-centered radical and an C-1-centered
acyl radical, formed through β-fragmentation of the O-7-
centered radical, as shown in Scheme 4. This proposal is also
supported by high spin densities at O-7, C-1, and O-1 in TS3,
as shown in Figure 6. From TS3 to the product complex
10
easily converted into malonic acid, according to the literature.
The potential energy of CI was calculated to be 25.9 kcal/mol
relative to DHBQ, which value is moderately larger than the Ea
value (23.3 kcal/mol) of TS3. It is therefore likely that this
pathway involving the spin state change is one of the minor
competitive pathways, occurring as alternative to the one
shown in Figure 4. The involvement of this minor path explains
(
ketene and oxaloacetic acid (I3)), the C-1−C-6 distance is
shortened from 1.513 to 1.312 Å and the C-5−O-7′ distance
from 1.384 to 1.211 Å, while the C-5−C-6 distance is elongated
from 1.526 to 3.210 Å. These results indicate that the reaction
from TS3 to the product complex involves the β-fragmentation
Figure 5. Geometry changes in the rate-determining step calculated with the DFT(B3LYP) method. The bond lengths are presented in Å. The
product complex is the IRC product from TS3.
3
180
dx.doi.org/10.1021/jo4001178 | J. Org. Chem. 2013, 78, 3176−3182

The Journal of Organic Chemistry
Article
Figure 7. Potential energy curves in singlet and triplet spin states in 12k in relation to the O-7−O-7′ bond length, calculated with the
DFT(UB3LYP) method. The potential energy curves were obtained by free geometry optimizations with the O-7−O-7′ bond fixed at each length.
a)
b)
Notes: The potential energy of I2K is set to zero here. The triplet product was calculated with the steepest descend optimization starting from CI
c)
geometry. Potential energy relative to DHBQ is given in parentheses.
the experimental yield of malonic acid ranging slightly above
0%.
Future studies, to be reported as follow-up of this account,
will consider the influence of the reaction pH on the rate and
the mechanism of the degradation of DHBQ by H O .
5
2
2
CONCLUSIONS
■
EXPERIMENTAL AND COMPUTATIONAL DETAILS
We have investigated the degradation mechanism of DHBQ by
hydrogen peroxide in aqueous, nonbuffered medium. Kinetic
analysis and product investigation showed DHBQ to afford
malonic acid, acetic acid, and carbon dioxide, with an
experimental Ea value of 20.4 kcal/mol. DFT(B3LYP)
computation indicated that the initial step of this degradation
is the formation of intermediate I2k with a 2,5-dihydroxy-1,4-
cyclohexadione structure having an intramolecular O−O bridge
between C-2 and C-5. The O−O bond of I2k is then
homolytically cleaved, and the resulting radical undergoes β-
fragmentation to give ketene and oxaloacetic acid. This process
is the rate-determining step in the overall sequence. While
ketene is attacked by water to yield acetic acid, oxaloacetic acid
suffers rapid oxidative decarboxylation to malonic acid with
concomitant release of carbon dioxide and water. The
■
Experimental Details. All chemicals were purchased from
commercial providers. They were of highest purity available (p.a.
grade) and were used without purification. Proton nuclear magnetic
1
resonance ( H NMR) spectra were recorded on at 400.13 MHz
proton resonance frequency.
The degradation reaction was started by adding a 30% aqueous
solution of hydrogen peroxide (0.22 mL, 1.96 mM hydrogen peroxide)
to 2.0 mL of an aqueous solution of DHBQ (1.07 mg, 0.00763 mM) in
a 20 mL round-bottom flask. The solution of DHBQ was preheated at
temperatures between 313.15 and 353.15 K before addition of the
hydrogen peroxide solution. After the reaction mixture was stirred for
6
0−1800 s, the flask was put into an ice bath to stop the reaction and
then was stored at 193.15 K overnight. After freeze-drying of the
1
sample, the residue was analyzed with H NMR (DMSO-d ) and gas
6
chromatography−mass spectrometry (GC−MS, after trimethylsilyla-
tion; see below for the procedure). For quantification, 4-
bromobenzaldehyde was used as the internal standard. The
trimethylsilylation was carried out according to the following protocol:
500 μL of pyridine containing 1.5 mg of N,N-dimethy-4-amino-
pyridine was added to the freeze-dried sample, followed by the
addition of 500 μL of N,O-bis(trimethylsilyl)trifluoroacetamide
containing 10% of trimethylsilyl chloride. The reaction mixture was
left at 323.15 K for 2 h to complete the reaction and was then injected
calculated E value (23.3 kcal/mol) was in good agreement
a
with the experimental value, indicating the correctness of the
computed mechanism. We also found another minor pathway,
where the spin state changes from singlet to triplet by the
elongation of the O−O bond of I2k to give two malonyl
radicals, which would be readily converted into malonic acid.
The reaction mechanism allows some direct conclusions as
to bleaching sequences in pulp discoloration treatments. By
degradation of DHBQ, organic acids are produced that will
contribute to a decreased pH value. It is thus recommendable
to exert pH control or work in slightly alkaline medium to
neutralize the formed acids. Malonic acid is a synthon with high
methylene activity and readily undergoes aldol-type condensa-
tions. Such reactions might produce potent chromophores
directly.
For in situ 1H NMR analysis of the reaction mixture, the
degradation of DHBQ was performed in D O with the same
2
experimental procedure as described above. After the reaction was
quenched in the ice bath, the sample was immediately transferred into
1
a NMR tube under further cooling and analyzed by H NMR.
For qualitative evaluation of carbon dioxide in the reaction mixture,
0 μL of the reaction mixture was added into a saturated aqueous
solution of calcium hydroxide to observe the formation of a white
precipitate of CaCO3.
2
(
conjugated systems) and thus contribute to unstable bright-
ness in the cellulosic pulps. To avoid this fast brightness
reversion, a thorough washing of the pulp to remove all malonic
acid is paramount. An ozone bleaching step, which generates
high amounts of low-molecular weight carbonyl compounds as
potential reaction partners for residual malonic acid, should
thus be avoided as last step of the bleaching sequence.
Computational Details. The GAUSSIAN 09 program packages
11
were employed in these calculations. The geometry optimization and
energy evaluation was carried out by the DFT method with the B3LYP
1
2
functional and the PCM method in water. In the geometry
optimization, we used the key word “Opt” in GAUSSIAN 09 in case
of the calculations of the reactant, products, and intermediates. For the
3
181
dx.doi.org/10.1021/jo4001178 | J. Org. Chem. 2013, 78, 3176−3182

The Journal of Organic Chemistry
Article
transition states, we also employed the “Opt” keyword, but
additionally set the “Ts”, “Noeigentest”, and “Calcfc” options; see
the manual for GAUSSIAN 09 for the details of these descriptions.
(3) Koulouri, S.; Malamidou-Xenikaki, E.; Spyroudis, S. Tetrahedron
2005, 61, 10894.
1
1
(4) (a) Manthey, M. K.; Pyne, S. G.; Truscott, R. J. W. Aust. J. Chem.
1989, 42, 365. (b) Ikeda, M.; Kitahara, K.; Nishi, H. J. Heterocycl.
Chem. 1992, 29, 289. (c) Zhang, D.; Jin, G.-X. Organometallics 2003,
22, 2851. (d) Gellerman, G.; Rudi, A.; Kashman, Y. Tetrahedron 1994,
50, 12959. (e) Reuben, G.; Shonle, H. A. J. Am. Chem. Soc. 1946, 68,
2246. (f) Placin, F.; Clavier, G.; Najera, F.; Desvergne, J. P.; Pozzo, J.
L. Polycyclic Aromat. Compd. 2000, 19, 107. (g) Lehaire, M.-L.;
Scopelliti, R.; Herdeis, L.; Polborn, K.; Mayer, P.; Severin, K. Inorg.
Chem. 2004, 43, 1609. (h) Seillan, C.; Brisset, H.; Siri, O. Org. Lett.
2008, 10, 4013. (i) Tang, Q.; Liu, J.; Chan, H. S.; Miao, Q. Chem.
Eur. J. 2009, 15, 3965. (j) Lee, D.-C.; Cao, B.; Jang, K.; Forster, P. M. J.
Mater. Chem. 2010, 20, 867. (k) Tang, Q.; Liang, Z.; Liu, J.; Xu, J.;
Miao, Q. Chem. Commun. (Cambridge) 2010, 46, 2977. (l) Di, C.; Li,
J.; Yu, G.; Xiao, Y.; Guo, Y.; Liu, Y.; Qian, X.; Zhu, D. Org. Lett. 2008,
The 6-31G(d) basis sets were employed for H, C, and O, where a
diffuse function was added to each of O and a p-polarization function
was added to H. In all calculations, the solvation energy was evaluated
with the PCM method, where the UFF parameters were used to
determine the cavity size. It was ascertained that each equilibrium
geometry exhibited no imaginary frequency and each transition state
exhibited one imaginary frequency. In the case of homolytic cleavage
of the covalent O−O bond, the energy was evaluated using the
unrestricted-B3LYP (UB3LYP) method instead of the restricted-
B3LYP (RB3LYP) one.
Enthalpy, entropy, and Gibbs energy changes were calculated at
93.15 K. Zero-point energy, thermal energy, and entropy change were
2
evaluated with the DFT(B3LYP) method. The translational entropy in
1
3
water was computationally treated according to the literature. We
presented an example of the input file used for the geometry
optimization of TS3 in Supporting Information.
1
(
0, 3025.
5) Indirect quantification of acetic acid by using the ratio between
malonic acid and acetic acid is impossible in this case, since a
substantial part of the protons of malonic acid is deuterized by D O
during the degradation.
2
ASSOCIATED CONTENT
Supporting Information
Complete ref 11. Additional experimental details and results.
Cartesian coordinates of the optimized geometries. This
material is available free of charge via the Internet at http://
pubs.acs.org.
■
(6) We also optimized another transitions state connecting DHBQ to
*
S
I1k, in which only two water molecules are involved. However, the
calculated activation barrier (19.5 kcal/mol) was much higher than
that of TS1, indicating that at least three water molecules are necessary
to smoothly complete the nucleophilic addition; see also Figure S10 in
Supporting Information.
(
7) We did not calculated the transition states of this keto−enol
AUTHOR INFORMATION
Corresponding Author
E-mail: thomas.rosenau@boku.ac.at.
interconversion, but the barriers in this process would be quite low,
since the keto form is much more stable than the enol counterpart.
■
*
⧧
(
8) One of the reasons for the discrepancy in the Δ S° is that we
used the gas phase rotational entropy without correction, though we
corrected the translational entropy.
Notes
The authors declare no competing financial interest.
(9) Siegel, B.; Lanphear. J. Org. Chem. 1979, 44, 942.
(
10) (a) Dagaut, P.; Wallington, T. J.; Kurylo, M. J. J. Phy. Chem.
1988, 92, 3836. (b) Vaghjiani, G. L.; Ravishankara, A. R. J. Phys. Chem.
989, 93, 1948. (c) Lightfoot, P. D.; Veyret, B.; Lesclaux, R. J. Phy.
Chem. 1990, 94, 708. (d) Lightfoot, P. D.; Lesclaux, R.; Veyret, B. J.
Phy. Chem. 1990, 94, 700. (e) Wallington, T. J.; Dagaut, P.; Kurylo, M.
J. Chem. Rev. 1992, 92, 667.
ACKNOWLEDGMENTS
■
1
We performed quantum chemical calculations on the work-
station of the Sakaki group, Fukui Institute for Fundamental
Chemistry at Kyoto University, Japan, and we thank them for
access. The financial support of the Austrian Christian Doppler
Research Society (CDG) through the CD-lab (Advanced
cellulose chemistry and analytics) and of the Austrian Research
Promotion Agency (FFG, project 829443) is gratefully
acknowledged.
(11) Frisch, M. J. et al. Gaussian 09, Revision A.1; Gaussian, Inc.:
Wallingford, CT, 2009. See Supporting Information for the complete
reference.
(12) (a) Becke, A. D. Phys. Rev. A 1988, 38, 3098. (b) Becke, A. D. J.
Chem. Phys. 1983, 98, 5648.
(13) Mammen, M.; Shakhnovich, E. I.; Deutch, J. M.; Whitesides, G.
M. J. Org. Chem. 1998, 63, 3821.
REFERENCES
■
(
1) (a) Rosenau, T.; Potthast, A.; Milacher, W.; Hofinger, A.; Kosma,
P. Polymer 2004, 45, 6437. (b) Rosenau, T.; Potthast, A.; Milacher, W.;
Adorjan, I.; Hofinger, A.; Kosma, P. Cellulose 2005, 12, 197.
(
c) Rosenau, T.; Potthast, A.; Kosma, P.; Suess, H.-U.; Nimmerfroh,
N. Holzforschung 2007, 61 (6), 656. (d) Krainz, K.; Potthast, A.; Suess,
U.; Dietz, T.; Nimmerfroh, N.; Rosenau, T. Holzforschung 2009, 63
(
6), 647−656. (e) Rosenau, T.; Potthast, A.; Krainz, K.; Yoneda, Y.;
Dietz, T.; Shields, Z. P. I.; French, A. D. Cellulose 2011, 18 (6), 1623.
2) (a) Brassard, P.; L’Ecuyer, P. Can. J. Chem. 1958, 36, 1346.
b) Nicolaides, D. N.; Gautam, D. R.; Litinas, K. E.; Papamehael, T. J.
(
(
Chem. Soc., Perkin Trans. 1 2002, 1455. (c) Lang, M.; Muhlbauer, A.;
Jagers, E.; Steglich, W. Eur. J. Org. Chem. 2008, 20, 3544. (d) Shaabani,
A.; Ghadari, R.; Sarvary, A.; Rezayan, A. H. J. Org. Chem. 2009, 74,
4
372. (e) Shaabani, A.; Ghadari, R.; Ghasemi, S.; Pedarpour, M.;
Rezayan, A. H.; Sarvary, A.; Ng, S. W. J. Comb. Chem. 2009, 11, 956.
f) Misiolek, A. W.; Ichimura, A. S.; Gentner, R. A.; Huang, R. H.;
McCaffrey, V. P.; Jackson, J. E. Inorg. Chem. 2009, 48, 9005.
g) Jimenez-Alonso, S.; Estevez-Braun, A.; Ravelo, A. G.; Zarate, R.;
(
(
Lopez, M. Tetrahedron 2007, 63, 3066. (h) Jimenez-Alonso, S.; Perez-
Lomas, A. L.; Estevez-Braun, A.; Martinez, F. M.; Orellana, H. C.;
Ravelo, A. G.; Gamarro, F.; Castanys, S.; Lopez, M. J. Med. Chem.
2
008, 51, 7132−7143.
3
182
dx.doi.org/10.1021/jo4001178 | J. Org. Chem. 2013, 78, 3176−3182