Effect of Structure and Substituents in the Aqueous Phase Oxidation of Alcohols and Polyols Over Au, Pd, and Au-Pd Catalysts

Article abstract of DOI:10.1007/s10562-014-1464-5

Abstract Reactivity trends for oxidation of various alcohols and polyols have been examined for carbon-supported Au, Pd, and Au-Pd catalysts. A Hammett σρ approach was used to study substituent effects, with Hammett factors (ρ) of 1.27, 1.31, and 0.40 obtained for Pd, Au, and Au-Pd catalysts, suggesting the formation of a net negative charge at the transition state of the rate limiting step. The lower ρ for the Au-Pd catalyst versus Au and Pd monometallic catalysts indicates the ability of the Au-Pd catalyst to stabilize the negative charge at the transition state, explaining the improved performance of Au-Pd bimetallic catalysts for alcohol oxidation. Hammett-Taft factors were used to explain the low selectivity of terminal diols and polyols to diacids.

Full text of DOI:10.1007/s10562-014-1464-5

Catal Lett
DOI 10.1007/s10562-014-1464-5
Effect of Structure and Substituents in the Aqueous Phase
Oxidation of Alcohols and Polyols Over Au, Pd, and Au–Pd
Catalysts
Abraham A. Rodriguez
John R. Monnier
•
Christopher T. Williams
•
Received: 15 September 2014 / Accepted: 10 December 2014
Ó Springer Science+Business Media New York 2015
Abstract Reactivity trends for oxidation of various
alcohols and polyols have been examined for carbon-sup-
ported Au, Pd, and Au–Pd catalysts. A Hammett rq
approach was used to study substituent effects, with
Hammett factors (q) of 1.27, 1.31, and 0.40 obtained for
Pd, Au, and Au–Pd catalysts, suggesting the formation of a
net negative charge at the transition state of the rate lim-
iting step. The lower q for the Au–Pd catalyst versus Au
and Pd monometallic catalysts indicates the ability of the
Au–Pd catalyst to stabilize the negative charge at the
transition state, explaining the improved performance of
Au–Pd bimetallic catalysts for alcohol oxidation. Ham-
mett–Taft factors were used to explain the low selectivity
of terminal diols and polyols to diacids.
1 Introduction
Much effort has been expended on the study of alcohol and
polyol oxidation over carbon-supported Au and Au-con-
taining catalysts. These reactions are often conducted in
strongly basic aqueous media. There remain many issues
concerning activity trends of Au versus Au–Pd bimetallic
compositions and role of O in the reaction mechanism [1–
2
3]. Other issues that remain unanswered include (1), why is
typically only the monoacid product observed for the
selective oxidation of glycerol. Diacids are rarely observed
for any catalyst system; this is especially important since
tartronic acid represents a higher value chemical interme-
diate than glyceric acid and (2), why are the oxidation rates
of alcohols so different? Oxidation rates of n-propanol are
much lower than those of either 1,2-propanediol or 1,3-
propanediol, which are in turn less reactive than glycerol
over the same catalysts at the same reaction conditions and
pH values [2–4].
Graphical Abstract
-
The role of added OH on activity has also been studied,
and, for supported Au catalysts, the rate of glycerol oxi-
-
dation is negligible when an excess of OH relative to
glycerol is not used. Likewise, for supported Pd and Pt
catalysts and Au–Pd bimetallic catalysts, the activities for
glycerol oxidation are much higher at high pH values [2, 5–
7
]. As literature reports have stated [4, 6, 8], the first step in
alcohol oxidation reactions is abstraction of a H atom from
a terminal –OH to form an alkoxy species. This can either
Keywords Alcohol oxidation ꢀ Aqueous phase ꢀ Diacids ꢀ
Hammett equation ꢀ Hammett–Taft factors
-
occur in the liquid phase by the action of free OH , as in
-
the case for supported Au catalysts, or by both free OH
and adsorbed alcohol to form an adsorbed alkoxy species,
as is presumably the case for Pd, Pt, or Au–Pd catalysts,
since Group VIII metals are active for dehydrogenation of
alcohols. However, the activities for Pt and Pd catalysts are
A. A. Rodriguez ꢀ C. T. Williams ꢀ J. R. Monnier (&)
Department of Chemical Engineering, University of South
Carolina, 301 Main Street, Columbia, SC 29208, USA
e-mail: monnier@cec.sc.edu
-
greatly enhanced when excess OH is present [2, 5–7].
123

A. A. Rodriguez et al.
Thus, the acid dissociation constants (K ) for alcohols
a
2.3 Oxidation Reactions
The aqueous phase oxidation of the different compounds
TM
should be considered when examining reactivity trends for
alcohols and polyols.
In this communication, we discuss the reactivity trends
for various alcohols and diols over carbon-supported Au,
Pd, and Au–Pd catalysts and relate activities to the dif-
ferent K values. Correlations of K values with Hammett
was performed in a 100 ml EZE-Seal
batch reactor
(Autoclave Engineers), using the same protocol as our
previously published work [4, 8]. The impeller turbine used
TM
for these experiments (Dispersimax , Autoclave Engi-
a
a
and Taft inductance factors are also made and used to offer
an explanation for the low rates of formation of diacids
from the oxidation of glycerol.
neers) has a hollow shaft with holes that are positioned
both above the liquid level and between the individual
turbine blades within the liquid. As a result, O mixing was
2
very efficient in the liquid reaction medium. The O gas
2
pressure above the liquid level is considerably higher than
in the liquid vortex at the impeller turbine which provides
2
Experimental Section
excellent radial and axial mixing of O in the liquid to
2
2
.1 Materials
minimize mass transfer effects of O in the aqueous reac-
2
tion medium, as opposed to a stirrer with solid shaft.
The reaction conditions used in this study were similar to
those that have been reported by others for alcohol oxidation
Glycerol, ethylene glycol, and 3-chloro-1-propanol were
purchased from Sigma Aldrich. 1,2-propanediol, 1,3-pro-
panediol, and 1-propanol were purchased from Alfa Aesar.
The 5 wt% Pd/CP-97 carbon catalyst (supplied by BASF
Catalysts Inc.) consisted of 4.4 nm Pd particles (24 % Pd
dispersion). The same CP-97 carbon was also used to make
the supported Au catalyst. Sodium hydroxide (NaOH)
[2–4, 8, 10], specifically T = 60 °C, P(O ) = 145 psig,
2
80 ml of 0.10 M solution of reactant alcohol, and 1.0 M
NaOH concentration.
-
pellets (BDH) were used to provide OH in the reaction
3 Results and Discussion
media. Sodium carbonate (Na CO ), gold chloride (Na
2
3
AuCl ꢀ3H O) (Alfa-Aesar), and formaldehyde (HCHO,
Oxidation reactions for eight different substrates were
carried out over three different catalysts; 5 wt% Pd/C,
1 wt% Au/C and the bimetallic Au–Pd/C catalyst (*0.5
coverage of Au on Pd). Table 1 shows reaction rate con-
stants for the different substrates over the different cata-
lysts, normalized to the amount of catalyst used. The
substrate alcohols are arranged in order of decreasing
activities for the Au–Pd/C catalysts. Rate constant calcu-
lated from a first order fit of reaction data.
4
2
3
7 %) (Sigma Aldrich) were used for preparation of the
supported Au catalyst. A Thermo Scientific Barnstead
Nanopure system supplied 18.2 MX-cm de-ionized water
that was used for all applications.
2
.2 Catalyst Preparation and Characterization
Preparation procedures for the 1 % Au/C and the Au–Pd/
C bimetallic catalyst have been reported in previously
published work [4, 8]. For the 1 % Au/C catalyst, solu-
tion of HAuCl 0.1 M (1 ml) was diluted with distilled
4
Table 1 First order reaction rate constants for oxidation of different
substrates on Pd, Au, and Au–Pd catalysts
water (10 ml) and was added to a stirred slurry of carbon
(
2 g) in distilled water (20 ml). A saturated solution of
Substrate
Rate constants 1/[(g cat)*h]
Na CO was added until a fixed pH of 10 was reached.
2
3
Pd/C
Au/C
0.46 Au–Pd/C
The slurry was allowed to stand for 1 h, then heated to
0 °C and reduced by dropwise addition of 1.5 ml of
7
Glycerol
1.59
0.00
0.24
0.13
0.00
–
1.86
0.10
0.54
0.13
0.26
–
34.31
12.71
8.28
5.17
4.11
3.91
3.75
HCHO (37 %). Preparation of the Au–Pd/C bimetallic
catalyst by electroless deposition has also been described
previously [8, 9]. The bimetallic catalyst has been char-
1,6-Hexanediol
1
,2-Propanediol
Ethylene glycol
1,3-Propanediol
acterized previously by using H titration of oxygen pre-
2
covered Pd sites and X-ray diffraction measurements [8],
showing 0.46 Au coverage on Pd at an overall compo-
sition of 1.1 wt% Au and 4.9 wt% Pd. The 5 % Pd/C
catalyst has been characterized previously by using H2
titration of oxygen pre-covered Pd sites, while the 1 %
Au/C catalyst has been characterized X-ray diffraction
measurements [8].
3
1
-Chloro,1-propanol
-Propanol
0.01
0.00
Rate constants have been normalized to the amount of catalyst used
1/h divided by mass of catalyst). Reaction conditions: T = 60 °C,
P(O ) = 145 psig, 80 ml of 0.10 M solution of reactant alcohol, and
1.0 M NaOH concentration, catalyst mass of 0.150 g for Pd/C,
.300 g for Au/C, and 0.060 g for 0.46 Au–Pd/C
(
2
0
1
23

Aqueous Phase Oxidation of Alcohols and Polyols
3
.1 Effect of K on Reaction Rate
a
For this reaction, the equilibrium constant, or K , the
a
acid dissociation constant, can be stated as:
As has been previously reported [2, 4, 11], the first step in
the oxidation of alcohols in aqueous, basic media is the
abstraction of a proton from the hydroxyl group. Most
evidence points to this step occurring in solution between
ꢁ
þ
½
RCH2O ꢂ½H ꢂ
K ¼
ð3Þ
a
½
RCH2OHꢂ
Rearranging the self-ionization constant for water
Eq. 4) and inserting it into Eq. 3 gives the dissociation
-
free OH and the alcohol, rather than adsorbed on the
(
catalyst surface [1, 5, 11]. Using the Henderson–Has-
selbalch equation (Eq. 1), the concentration of the depro-
tonated species can be calculated for each alcohol under
the reaction conditions.
constant (Eq. 5).
Kw
OHꢁꢂ
ꢁ
þ
þ
K ¼ ½OH ꢂ½H ꢂ )½H ꢂ ¼
w
ð4Þ
ð5Þ
½
ꢁ
ꢁ
½RCH O ꢂK
½
A ꢂ
2
w
Ka ¼
pH ¼ pK ꢁ
ð1Þ
ꢁ
a
½
RCH OHꢂ½OH ꢂ
½
HAꢂ
2
Equation 5 can then be solved for the concentration of
the deprotonated species,
Table 2 shows the concentration of the protonated and
deprotonated species at zero reaction time (pH 14), as well
as the percent of the substrate present as the deprotonated
species in solution. Comparison of reaction rate constants
in Table 1 with the initial concentrations of deprotonated
Ka
ꢁ
½RCH2OHꢂ½OHꢁꢂ
½
RCH2O ꢂ ¼
ð6Þ
Kw
-
In previous work [4] we showed the existence of a
alcohols (A ) in Table 2 indicates there is a strong corre-
kinetic isotope effect for the oxidation of d -ethylene gly-
4
lation between reactivity and the extent of deprotonation,
col (HOCD CD OH) over both Au/C and bimetallic Au–
2
2
or more specifically, the dissociation constant (K ). This
a
Pd/C catalysts, indicating that C–H bond scission was the
rate-limiting step. If we assume the same step is rate
determining for all terminal alcohols, the rate expression
can be written as:
correlation holds true for all three catalysts, even though
Pd, in particular, can form alkoxides by both deprotonation
in solution and adsorbed alcohols to form adsorbed alkoxy
species. In order to better understand this relationship, a
closer look at the kinetics of the rate limiting step is
necessary.
k1
ꢁ
ꢃ
ꢃ
ꢃ
RCH2O þ 2 ! RCHO þ H
ð7Þ
ð8Þ
ꢁ
rRCHO ¼ k1½RCH2O ꢂ
Previous work has shown that oxidation of alcohols does
not proceed in the absence of a base when using Au cat-
alysts, and is greatly inhibited when using Pd catalysts [1,
where * represents a catalytically active site. Equation 8 is
consistent with our results and the correlation of reaction
rates with the extent of deprotonation reported in Table 2.
Finally, Eq. 8 can be rewritten using Eq. 6 to give:
5
, 6, 11]. Most evidence points to the initial deprotonation
of a terminal –OH group of the substrate being a solution-
phase phenomenon driven by free hydroxyl ions [1, 2, 4, 6,
Ka
rRCHO ¼ k1
½RCH2OHꢂ½OHꢁꢂ;
ð9Þ
8
, 11]. The equilibrium reaction for the deprotonation of
Kw
the alcohol is written as:
where the direct dependence of the rate of the RDS on the
-
concentration of the substrate alcohol, OH concentration,
ꢁ
þ
RCH2OHRCH2O þ H
ð2Þ
and magnitude of K are readily apparent. The first order
a
dependency of alcohol conversion on RCH OH is also
2
Table 2 Acid dissociation constants for different substrates, con-
centration of protonated and deprotonated species, and extent of de-
protonation at pH 14 and t = 0
consistent with the first order decay plots of glycerol,
ethylene glycol, and 1,2-propanediol observed in our pre-
vious work [4, 8].
-
[A ]
-
Substrate
pK
a
[HA]
(mol/l)
%[HA] ? [A ]
a
a
(
mol/l)
3
.2 Hammett Correlations
Glycerol
14.15 0.041
15.10 0.007
14.90 0.011
15.10 0.007
15.10 0.007
0.059
0.093
0.089
0.093
0.093
0.099
0.099
41.5
7.40
11.2
7.40
7.40
1.20
1.00
1
1
,6-Hexanediol
,2-Propanediol
In order to determine how substrate structure is related to the
activity trends, the Hammett equation (Eq. 10) [12] was used
to compare 1-propanol to other substrates. The substrate
alcohols in this study can be viewed as substituted propanols.
Ethylene glycol
1
3
1
a
,3-Propanediol
-Chloro,1-propanol 15.93 0.001
-Propanol 16.00 0.001
ꢀ ꢁ
k
log
¼ rq
ð10Þ
k0
Initial substrate concentration of 0.1 M
123

A. A. Rodriguez et al.
a
Table 3 pK and r values for the series of substrates
Substrate
pK
a
a
r (pKa(H) - pK )
Glycerol
14.15
15.10
14.90
15.10
15.10
15.93
16.00
1.85
0.90
1.10
0.90
0.90
0.07
0.0
Ethylene glycol
1
1
1
3
1
,2-Propanediol
,3-Propanediol
,6-Hexanediol
-Chloro,1-propanol
-Propanol
Values taken from [17]
The parameters of Eq. 10 are: k , the reaction rate
0
constant for propanol conversion; k, the reaction rate
constant for each of the substituted alcohols; r, the sub-
stituent constant; and q, the reaction constant. The sub-
Fig. 1 Hammett plots for the alcohols from Table 3 over carbon-
supported Au, Pd, and Au–Pd catalysts
stituent constant r is defined as the difference of the pK of
a
the unsubstituted compound (pKa(H), propanol in the
present case) and the pK of the substituted compound
a
has, in fact, shown that Au is deposited in a uniform
manner over all Pd surface sites using the same ED bath,
consistent with the above hypothesis. The lower catalytic
activity at lower Au coverages is due to a smaller number
of these types of bifunctional sites (the activity is more
similar to Pd) and at higher coverages the extra Au atoms
produce larger domains of Au and even three-dimensional
Au structures, which lowers the numbers of Pd–Au site
pairs to give catalyst performance resembling pure Au
surfaces.
(
Table 3) [12–14]. Table 3 shows the pK and r values for
a
the series of alcohols in this study. Propanol has a pK of
a
1
6.00, while all other substrates have lower pK values.
a
This results in positive r values for all the different
substituted compounds. Aliphatic, linear compounds are
expected to exhibit only inductive effects and not the res-
onance effects observed for substituted aromatic structures
[
12]. This means the substituents will have a net electron
withdrawing effect, consistent with tabulated values for
Hammett r values for –OH and –Cl groups tabulated in the
literature [12, 15, 16].
XPS data for the series of catalysts with coverage from 0
to 1.0 [4, 8] showed that BE shifts for the Au 4f region did
not correlate with the trend in activity (shifts to lower BE
values were larger in magnitude at lower coverages),
suggesting that electronic effects are not responsible for the
increased activity of the Au–Pd catalysts. Furthermore, a
similar family of Ag–Pd/C catalysts was prepared and
studied to determine whether geometric/ensemble effects
were responsible for the enhanced activity of particular
Au–Pd compositions, since Ag should be as effective as Au
for formation of Pd ensembles. However, there was no
bimetallic effect for the Ag–Pd catalysts, suggesting that
ensemble effects were not responsible for the improved
activity of the 0.5 Au–Pd/C catalyst. Thus, the enhanced
activity of intermediate coverages of Au on the Pd surface
is due to the formation of bifunctional Au–Pd sites and the
fact that activity is a maximum suggests the bifunctional
site is a Au–Pd site pair.
Figure 1 shows the Hammett plots for the oxidation of
the series of substrates over the Au/C, Pd/C and Au–Pd/C
catalysts. Hammett reaction constant q are all positive
values, indicating the development of a negative charge at
the reaction center in the transition state of the rate limiting
step [12, 13, 18]. The difference in value of the reaction
constants q helps explain the difference in the catalytic
behavior of the three catalysts. Both Au and Pd have
similar values of q and rather similar rates of reaction
(
Table 1) for the different alcohols, suggesting similar
negative charge distributions at the transition state. The
Au–Pd catalyst, on the other hand, is approximately 20
times more active than either Au or Pd, while the reaction
constant q (0.40) is much smaller than for either Au or Pd
(
1.31 and 1.27, respectively).
As we have previously reported [8], the fact that the
maximum in activity occurs at a Au coverage of 0.5 on the
Pd surface suggests a bifunctional effect where both Pd and
Au sites are involved in the rate determining step. For non-
site specific deposition of Au on Pd, the maximum, sta-
tistical concentration of bimetallic Au–Pd site pairs occurs
at a coverage of 0.5. Previous work from our group [19]
The new insight provided by the Hammett Correlations
provides an explanation of the nature of the interaction of
reaction intermediates with the bimetallic site. The Au–Pd
site stabilizes the charge of the transition state by distrib-
uting the negative charge between Pd and Au sites to give a
nominal, five-membered ring transition state at the
1
23

Aqueous Phase Oxidation of Alcohols and Polyols
-
1,3-propanediol in Table 2. The e -withdrawing effect of
Fig. 2 Stabilization of glycerol
reaction intermediate at
transition state
the –OH group on the C of 1,2-propanediol increases the
2
acidity of the neighboring –CH OH group, which makes
2
-
?
formation of –CH O ? H easier and lowers the pKa
2
value to 14.9, compared to 16.00 for 1-propanol. For 1,3-
propanediol, the inter-positioning of the –CH – group
2
-
lowers the e -withdrawing effect of the –OH group at the
rate-limiting step (C–H bond breaking shown in Fig. 2).
We propose that the electron-rich Au site interacts more
weakly with the C–O bond (AuꢀꢀꢀO–C), permitting facile
desorption of products, while the active Pd site interacts
more strongly with the C–H bond, which increases the rate
of dehydrogenation of a Carbon. For the monometallic
catalyst, the Pd interaction would be more centered on the
C–H bond (Pd is an excellent hydrogenation–dehydroge-
nation catalyst), while Au interacts primarily with the C–O
bond, since it forms C–O bonds much more easily than C–
H bonds [8]. Gold is also a poor dehydrogenation catalyst
which would make C–H bond breaking more difficult. This
is consistent with DFT results published by Medlin et al.
C position, which then increases the pK to 15.10.
3 a
Application of the same reasoning to the effect of the
terminal carboxylate group of the glycerate anion
-
(HOCH2CHOHCOO ) from oxidation of glycerol helps to
understand the low rates of formation of tartronic acid (or
1,3-diglycerate dianion in basic media). The Taft factor for
-
the –COO– is -1.06 [15], indicating it is strongly e -
releasing. This lowers the acidity of the O–H group at the
C3 position; the –COO– group inhibits the ability to
deprotonate the remaining terminal –OH group.
A second factor inhibiting the formation of a dicar-
boxylic acid is the consumption of base during reaction.
-
Davis [22] determined the consumption of OH during
[
4], who report a decrease in the barrier for C–H scission
for the Au–Pd bimetallic catalyst compared to monome-
tallic Au, and also a decrease in coverage of strongly bound
adsorbates compared to monometallic Pd catalysts.
glycerol oxidation over an Au/TiO catalyst to be 1.92 mol
2
of hydroxyl ions consumed per mole of glycerol reacted.
The rate expression in Eq. 9 shows alcohol oxidation is
-
directly related to OH , thus the deprotonation of the
3
.3 Positive r and its Effect on Selectivity
second, terminal –OH group would be hindered to effec-
tively suppress the sequential oxidation pathway. Davis
[22] showed that for the same catalyst the oxidation rate of
Previous results for alcohol oxidation have shown that
while Au, Pd, and Au–Pd catalysts are highly selective for
the formation of carboxylic acids, selectivity to the corre-
sponding dicarboxylic acids are usually very low or non-
existent [4, 11, 20–22]. For glycerol, the most selective
secondary product is usually glycolic acid formed via C–C
bond cleavage, while terminal diols typically form only the
monocarboxylic acid [4, 10].
-
glyceric acid to tartronic acid increased for higher [OH ]/
[glyceric acid] ratios. Even so, TOF values for conversion
of glyceric acid were much lower than for glycerol con-
-
version; at an OH concentration of 0.6 M the TOF for
-
1
glyceric acid conversion was 0.1 s , compared to 4.9 s
-1
-
1
-
for glycerol, and only increased to 0.47 s when OH
concentration was increased to 2.0 M. This is consistent
with the proposed effect of –COO– to limit the ability to
deprotonate the second –OH group and the resulting
reaction inhibition.
Given the accepted mechanism for alcohol oxidation,
formation of a dicarboxylic acid likely proceeds through
sequential oxidation, first to a carboxylic acid and then to
the dicarboxylic acid [6, 22]. The pK of the second dis-
a
The results for oxidation of glycerol over the 0.460
Au–Pd/C bimetallic catalyst in this study are shown in
Fig. 3. At longer reaction times ([1.5 h) there is a slow
decrease in the selectivity to glyceric acid from 74 to
68 % at 5 h of reaction; conversely, the selectivity to
tartronic acid increased from 5 to 10 % over the same
time period. Glycolic acid (11 %) and oxalic acid (2 %),
both of which are C2 monoacid and diacid, respectively,
were also observed, but are not discussed since the C–C
cleavage necessary to form these products is an additional
complication and not the intent of this study. This slow
sociation constant for a diol must be higher than the first
one, meaning that the r value will be lower than that of the
diol. This in turn results in a low reaction rate for oxidation
of the second –OH group (or the remaining –OH group of
the monocarboxylic acid).
To better understand the effects of substituents on pKa
values, it is more useful to use Taft factors, rather than
Hammett factors, since Taft factors are more relevant to
aliphatic systems where only inductive effects are present
[
15, 16]. For Taft factors, substituent effects fall off with
increasing distance from the reaction center by a factor of
formation of tartronic acid is consistent with the proposed
-
effect of OH concentration and pK on reaction rate and
0
.36 for every –CH – group. The positive Taft factor of
2
a
1
.34 for the –OH group, indicative of an electron with-
selectivity. These results in this study are similar to those
reported by Davis [22] for oxidation of glycerol using
drawing effect can also be used to explain the difference
in pKa values for 1-propanol, 1,2-propanediol and
1.6 % Au/TiO in a batch reactor. Glycerol conversion
2
123

A. A. Rodriguez et al.
Fig. 3 a Normalized moles versus reaction time and b selectivity for monoacid (glyceric acid, circles) and diacid (tartronic acid, triangles) for
glycerol (squares) oxidation over 0.46 Au–Pd/C catalyst
was carried out for 12 h to observe the formation of
diacids. Selectivity to glyceric acid was 63 at 33 % con-
version, and increased to 69 at 85 % conversion, while
selectivity to tartronic acid was 2 and 4 %, respectively.
The conversion to diacids was very slow and the overall
selectivity to tartronic acid and oxalic acid remained
aldehyde and then the acid, eliminating the need for the
hydroxide catalyzed deprotonation in solution.
4 Conclusions
\
10 % even after 12 h of total reaction time. On the
Oxidation of a family of alcohols, diols, and polyols has
been studied to diagnose reactivity trends of different
substrates. The Hammett rq approach was used to study
the dependence of substrate structure with reactivity. The
linear dependency between r and the logarithm of the
reaction rates (relative to n-propanol) gave different
positive slopes q for the reactions of the different alcohols
for the three catalysts studied. This implies that a negative
charge is developed at the reaction center during the rate-
limiting step, and that electron withdrawing groups pro-
mote activity. The low selectivity of terminal diols and
polyols to form dicarboxylic acids (or dicarboxylate
anions in basic media) by the sequential oxidation of the
second terminal –OH group of the monocarboxylate anion
is likely due to the combination of two factors. The first is
the low reactivity of the second –OH group caused by the
other hand Prati [23] evaluated gold supported on a
commercial, basic anion resin for the oxidation of glyc-
erol in a batch reactor and observed relatively high
selectivity to tartronic acid; at 90 % glycerol conversion
the selectivities to glyceric acid and tartronic acid were
4
9 % and 16 %, respectively. However, increasing the
-
OH ]/[glycerol] ratio from 1:1 to 4:1 resulted in a
[
decrease of tartronic acid, with 60 % selectivity to gly-
ceric acid and only 8 % to tartronic acid. Turnover fre-
quency for glycerol conversion increased by a factor of
four, but the lower selectivity to tartronic acid is incon-
sistent with other published results and the consecutive
pathway for tartronic acid formation.
The present batch reactor results suggest two possibili-
ties to overcome low selectivity to dicarboxylic acid pro-
-
ducts. The first is to simply increase the OH concentration
-
inhibiting effect of the –COO group at the other termi-
of the solution [6]. This is not practical, due to the already
high concentrations that currently make these reactions
unattractive. Further, the logarithmic nature of the pH scale
nus; the second factor is the necessarily high pK of the
a
-
second –OH groups and the lower concentration of OH
remaining after formation of the monocarboxylate species.
The kinetic expression for the oxidation of the different
alcohols reveals the direct dependence of the rate-limiting
-
means that very high concentrations of OH would be
required. The second option is to tailor a catalyst that reacts
with the –OH groups of these substrates to form adsorbed
-
RO species. Such an intermediate could react to form an
step to the K of the substrate and the concentration of
a
deprotonated substrate in solution.
1
23

Aqueous Phase Oxidation of Alcohols and Polyols
Acknowledgments The authors gratefully acknowledge NSF No.
CBET-0854339 for support of Abraham A. Rodriguez during the
course of this research.
10. Biella S, Castiglioni G, Fumagalli C et al (2002) Catal Today
72:43–49
11. Ketchie WC, Murayama M, Davis RJ (2007) J Catal 250:
2
64–273
1
1
1
1
1
1
2. Johnson CD (1973) University Press, Cambridge Eng
3. Hammett LP (1935) Chem Rev 17:125–136
4. McDaniel DH, Brown HC (1958) J Org Chem 23:420–427
5. Lange NA, Dean JA (1973) McGraw-Hill, New York
6. March J (1968) McGraw-Hill, New York
7. Serjeant EP, Dempsey B (1979) International Union of Pure and
Applied Chemistry Pergamon Press, Oxford, New York
8. Hammett LP (1937) J Am Chem Soc 59:96–103
9. Rebelli J, Rodriguez AA, Ma S et al (2011) Catal Today
References
1
. Zope BN, Hibbitts DD, Neurock M, Davis RJ (2010) Science
30:74–78
3
2
3
4
. Davis SE, Ide MS, Davis RJ (2012) Green Chem 15:17–45
. Ide MS, Davis RJ (2013) J Catal 308:50–59
. Griffin MB, Rodriguez AA, Montemore MM et al (2013) J Catal
1
1
3
07:111–120
. Garcia R, Besson M, Gallezot P (1995) Appl Catal Gen
27:165–176
. Ketchie WC, Murayama M, Davis RJ (2007) Top Catal
4:307–317
. Takagaki A, Tsuji A, Nishimura S, Ebitani K (2011) Chem Lett
0:150–152
. Rodriguez AA, Williams CT, Monnier JR (2014) Appl Catal Gen
75:161–168
. Rebelli J, Detwiler M, Ma S et al (2010) J Catal 270:224–233
1
60:170–178
5
6
7
8
9
2
2
0. Dimitratos N, Lopez-Sanchez JA, Anthonykutty JM et al (2009)
Phys Chem Chem Phys 11:4952–4961
1. Dimitratos N, Porta F, Prati L (2005) Appl Catal Gen 291:
1
4
2
10–214
2
2
2. Zope BN, Davis SE, Davis RJ (2012) Top Catal 55:24–32
3. Villa A, Chan-Thaw CE, Prati L (2010) Appl Catal B Environ
4
9
6:541–547
4
123