Homogeneous oxidation reactions of propanediols at low temperatures

Article abstract of DOI:10.1002/cssc.201000142

O₂ reacts with propanediols via homogeneous pathways at 400-500 K. 1,2-Propanediol forms CH₃CHO, HCHO, and CO₂ via oxidative C-C cleavage and acetone via dehydration routes, while symmetrical 1,3-propanediol undergoes dehydration and oxidative dehydrogenation to form, almost exclusively, acrolein (ca. 90% selectivity). The products formed and their kinetic dependence on reactant concentrations are consistent with radical-mediated pathways initiated by O₂ insertion into C-H bonds in a β position relative to oxygen atoms in diol reactants. Propagation involves β-scission reactions that form hydroxyl and hydroxyalkyl radicals. Acrolein/O₂/H₂O mixtures from the homogeneous oxidation of 1,3-propanediol form acrylic acid (with 90% yield) in tandem reactors containing molybdenum-vanadium oxide catalysts. These data reveal the unique reactivity of diols, compared with triols and alkanols, in homogeneous oxidations, while also providing useful insight into the molecular basis for reactivity in biomass-derived oxygenates.

Full text of DOI:10.1002/cssc.201000142

DOI: 10.1002/cssc.201000142
Homogeneous Oxidation Reactions of Propanediols at Low
Temperatures
Eva Dꢀaz,[a, b] Marꢀa Eugenia Sad,[a, c] and Enrique Iglesia*[a]
O2 reacts with propanediols via homogeneous pathways at
400–500 K. 1,2-Propanediol forms CH3CHO, HCHO, and CO2 via
oxidative CꢀC cleavage and acetone via dehydration routes,
while symmetrical 1,3-propanediol undergoes dehydration and
oxidative dehydrogenation to form, almost exclusively, acrolein
(ca. 90% selectivity). The products formed and their kinetic de-
pendence on reactant concentrations are consistent with radi-
cal-mediated pathways initiated by O2 insertion into CꢀH
bonds in a b position relative to oxygen atoms in diol reac-
tants. Propagation involves b-scission reactions that form hy-
droxyl and hydroxyalkyl radicals. Acrolein/O2/H2O mixtures
from the homogeneous oxidation of 1,3-propanediol form
acrylic acid (with 90% yield) in tandem reactors containing
molybdenum-vanadium oxide catalysts. These data reveal the
unique reactivity of diols, compared with triols and alkanols, in
homogeneous oxidations, while also providing useful insight
into the molecular basis for reactivity in biomass-derived oxy-
genates.
Introduction
The conversion of biomass-derived feedstocks to liquid fuels
remains a formidable challenge.[1–5] Transesterification of trigly-
cerides in biodiesel production leads to glycerol as a low-cost
byproduct,[6] making routes for its subsequent chemical con-
version attractive. Glycerol can be converted into 1,3-propane-
diol by either enzymatic dehydration and subsequent reduc-
tion, or via deoxygenation.[7] 1,2-Propanediol can be produced
from glycerol via reactive distillation using copper chromite
catalysts.[4] Several studies have addressed the pyrolysis, dehy-
dration, etherification, esterification, and reforming reactions of
glycerol, as well as its oxidative conversion; the latter paths in-
clude oxidation at terminal OH groups to form glyceric and tar-
tronic acids, at secondary OH groups to give hydroxyacetone,
and at all OH groups to form ketomalonic or mesoxalic
acids.[4,8] Markets for such products remain immature, at least
in part because these oxidations exhibit low selectivities and
yields and involve inefficient liquid-phase processes using
dilute aqueous glycerol. The multifunctional nature of triols
and the concomitantly diverse conversion pathways and prod-
ucts lead, in general, to poor selectivities.[3]
lein selectivities of ca. 90% at 430 K. 1,2-Propanediol forms
acetone (ca. 25% selectivity) and also decomposes into acetal-
dehyde, formaldehyde, and CO2 via CꢀC cleavage. Rates were
strictly proportional to the reactor volume and unaffected by
contact with heated transfer lines or other surfaces, consistent
with exclusively homogeneous pathways. The effects of pro-
panediol and O2 pressures on oxidation rates and the identity
of the products formed are consistent with chain propagation
processes mediated by oxygen-containing radicals.
The high acrolein selectivity in homogeneous 1,3-propane-
diol oxidation led us to introduce the acrolein and H2O formed
together with unreacted O2 into a second reactor containing a
Mo-V oxide catalyst,[14] able to convert acrolein into acrylic acid
with 90% yields (based on propanediol reactants). This com-
bined process provides a route for concurrent oxidation of dif-
ficult-to-separate 1,2 and 1,3-propanediols, formed via glycerol
hydrogenolysis, into useful products (acrylic and acetic acids,
acetone, acetaldehyde). This study also provides a kinetic
framework relevant to the practice of these chemistries, as well
as mechanistic insights into homogeneous pathways for the
Reactions of diols derived from glycerol have also been stud-
ied, albeit less thoroughly. 1,2-Propanediol reacts with O2 to
form hydroxyacetone as the initial product and methylglyoxal
via subsequent oxidation of hydroxyacetone on Ag catalysts
(500–700 K).[9,10] Methylglyoxal selectivities decrease with in-
creasing conversion, because of sequential oxidation and
cracking reactions. In electrochemical processes catalyzed by
Au and Pt,[11–13] 1,2-propanediol oxidation led to the predomi-
nant cleavage of CꢀC bonds, while 1,3-propanediol preferen-
tially formed hydroxypropanoic acid via routes analogous to
those involved in oxidation of primary alkanols.
[a] Dr. E. Dꢀaz, Dr. M. E. Sad, Prof. E. Iglesia
Department of Chemical Engineering, University of California at Berkeley
Berkeley, CA 94720 (USA)
Fax: (+1)510-642-4778
E-mail: iglesia@berkeley.edu
[b] Dr. E. Dꢀaz
Department of Chemical and Environmental Engineering
University of Oviedo
C/Julian Claveria, s/n, 33006 Oviedo (Spain)
[c] Dr. M. E. Sad
UNL CONICET, Inst Invest Catalisis & Petroquim INCAPE, GICIC
RA-3000 Santa Fe (Argentina)
Here, we show that 1,2 and 1,3-propanediols react with O2
via previously unreported homogeneous gas-phase routes at
modest temperatures (400–600 K). 1,3-Propanediol gives acro-
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cssc.201000142.
ChemSusChem 2010, 3, 1063 – 1070
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1063

E. Iglesia et al.
oxidative activation of chemical bonds in typical biomass-de-
rived oxygenates.
vessel was bypassed. We conclude from these data that pro-
panediol oxidation products formed only within the heated
vessel at rates proportional to its volume and without contri-
butions from surface-mediated radical initiation or termination
reactions.
Results and Discussion
Alkanols (e.g., 1-propanol and 2-propanol) reacted with O2
only at significantly higher temperatures than diols (700 K, vs.
400–500 K) and predominantly formed CH4 and COx.[15] Glycerol
did not react with O2 at the temperatures of this study (430–
570 K), as also reported in previous studies.[16] Thus, the homo-
geneous pathways reported here appear to reflect a reactivity
in radical-mediated pathways that appears to be unique to
diols among C3 oxygenates.
Homogeneous 1,3-propanediol oxidation
Rates, conversions, and selectivities for the gas-phase oxidation
of 1,3-propanediol at 400–600 K and 12.4 s residence time are
shown in Figure 1. No products were detected in the absence
of O2. The absence of surface-mediated conversion pathways
was confirmed by displacing the reactor volume with glass
beads (6.5–9.0 cm3 volume) at constant residence time and
430–570 K.
Figure 1 shows that acrolein was the predominant product
formed at low temperatures (<500 K); acrolein selec-
tivities decreased with increasing temperature
(Figure 1), while acetaldehyde and formaldehyde se-
lectivities concurrently increased (Table 1). CO was
not detected in the reactor effluent (<0.2% selectivi-
ty) and CO2 selectivities were less than 2%. These
high selectivities to a single product (ca. 90% acrole-
in, 400–450 K) are unusual in homogeneous oxida-
tions, which typically form a wide range of products
because of the unselective nature of CꢀH bond acti-
vation steps involved in chain initiation and propaga-
tion steps. High acrolein selectivities were maintained
even as diol reactants reached nearly complete con-
version with increasing residence time (91% selectivi-
ty at>99% conversion; 543 K, 8.8 kPa O2, 1 kPa diol,
14.1 s residence time).
~
&
^
Figure 1. 1,3-Propanediol conversion ( ), oxidation rate ( ), and acrolein selectivity ( ) as
a function of temperature (12.4 s residence time; residence time is defined as the reactor
volume divided by the total volumetric (STP) flow rate; 0.47 kPa O2, 0.28 kPa 1,3-pro-
panediol).
1,3-Propanediol oxidation pathways were probed
by measuring the effects of residence time on prod-
uct selectivities (Table 1). Selectivities were independ-
ent of residence time, consistent with the products
Conversion rates and selectivities were independent of reac-
tor volume for a given residence time (Figure 2). Reactant con-
versions and selectivities were not influenced by changes in
the temperature of the transfer lines before or after the reactor
and conversion products were not detected when the reactor
Table 1. Conversion, reaction rate, and selectivity towards acrolein, acetal-
dehyde, and formaldehyde in the oxidation of 1,3-propanediol, as function
of residence time.[a]
t [s] Conversion [%] ꢀr13PD [mmolcmꢀ3 hꢀ1] Selectivity [%]
Acrolein CH3CHO HCHO CO2
18.6 32.9
11.9 23.1
9.6 20.0
7.4 12.0
10.2
10.7
12.4
9.3
8.1
9.6
90
91
93
92
89
93
90
6.0
6.8
5.5
6.1
8.2
4.6
6.4
1.8
2.0
1.3
1.7
1.6
0.9
1.3
2.0
0.2
0.2
0.2
1.0
1.3
2.0
6.0
8.8
8.9[b] 11.8
12.4[c] 23.1
10.9
[a] Conditions: T=430 K, 0.28 kPa 1,3-propanediol, 0.89 kPa O2. [b] Varied by
changing the reactor volume from 9 to 6.5 cm3 by displacement with quartz
beads. [c] Interpolated from the above data to allow comparisons at identi-
cal reactant conversions. t is the residence time, defined as the ratio be-
tween the reactor volume and the volumetric flow rate at inlet conditions.
formation as primary products from 1,3-propanediol via paral-
lel direct pathways (Scheme 1) and with their inertness in sec-
ondary oxidations. The CH3CHO/(HCHO+CO2) molar ratios in
Figure 2. Comparison of reaction rates obtained for oxidation at a space
3
^
*
;
time of 12.4 s and two different reactor volumes: 6.5 ( ) and 9 cm (
0.47 kPa O2, 0.28 kPa 1,3-propanediol).
1064
www.chemsuschem.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2010, 3, 1063 – 1070

Homogeneous Oxidation Reactions of Propanediols at Low Temperatures
Scheme 1. Homogeneous oxidation pathways of 1,3-propanediol.
products were near unity and unaffected by residence time,
consistent with C1 and C2 species that form concurrently via
CꢀC bond cleavage from diol-derived species. The ratio of
CO2/HCHO varied from 0.1 to 1.5, apparently because of secon-
dary decomposition of HCHO to COx. The addition of acrolein
to 1,3-propanediol/O2 reactants did not influence the effluent
composition or the diol conversion rate or selectivity.
The effects of 1,3-propanediol and O2 reactant pressures on
acrolein and acetaldehyde formation rates (430 K) are shown in
Figures 3 and 4. Acrolein formation rates increased with in-
Figure 4. Effects of O2 inlet pressure on the formation rate of a) acrolein, and
^
b) acetaldehyde [430 K, residence time 12 s, diol pressures: 0.14 kPa ( ),
~
&
0.28 kPa ( ), 0.42 kPa ( )].The dashed curves correspond to the functional
dependence in Equations (1) and (2).
Scheme 2. Activation pathways for radical-mediated reactions of 1,3-pro-
panediol.
Figure 3. Effects of 1,3-propanediol inlet pressure on the formation rate of
a) acrolein and b) acetaldehyde [430 K, residence time 12 s, O2 pressures:
routes for the activation of 1,3-propanediol molecules via H-
abstraction pathways. H-abstraction at the a-position is fa-
vored by bond energies.[18] The radicals thus formed can then
undergo three b-elimination routes to cleave CꢀC, CꢀH or Oꢀ
H bonds. CꢀC bond cleavage gives acetaldehyde and CH2OH*,
~
&
^
*
0.25 kPa ( ), 0.47 kPa ( ), 0.89 kPa ( ) ,1.62 kPa ( )]. The dashed curves cor-
respond to the functional dependence in Equations (1) and (2).
creasing 1,3-propanediol (P1,3PD) and O2 (PO ) pressures, whereas
2
*
acetaldehyde formation rates were independent of propane-
diol pressure and proportional to O2 inlet pressures:
which then forms formaldehyde and HO2 via oxidation.[19] b-
Elimination with CꢀH or OꢀH bond cleavage forms acrolein,
but oxygen addition is much faster than this last reaction. O2
readily adds to a-hydroxy-alkyl radicals to form a-hydroperoxyl
1=2
1;3PD O2
ðracroleinÞ ¼ kP3=2
P
ð1Þ
ð2Þ
*
radicals, which then eliminate HO2 to form 3-hydroxypropa-
ðracetaldehydeÞ ¼ kPO
nal.[20] 3-Hydroxypropanal then decomposes rapidly into acrole-
in via homogeneous pathways.
2
Fractional orders (1.5 in organic substrate; 0.5 in O2) for
acrolein are typical of free-radical pathways,[17] including those in-
The elementary reactions proposed in Scheme 3, together
with the assumption of pseudo-steady-state for all radical spe-
cies, leads to an acrolein formation rate equation similar to
r
itiated by OꢀO cleavage. Scheme 2 shows several plausible
ChemSusChem 2010, 3, 1063 – 1070
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
1065

E. Iglesia et al.
The pseudosteady-state hypothesis (PSSH) for all intermedi-
ates and the long-chain approximation[25] lead to an equation
for 3-hydroxypropanal formation rates (details in Supporting
Information):
rffiffiffiffiffiffiffi
2k
k8
1=2
1;3PD O2
1P3=2
P
ð3Þ
racrolein ¼ r3ꢀhydroxypropanal ¼ 2k4
3-Hydroxypropanal was not detected in the effluent stream,
because it reacts rapidly via a facile retro-Michael addition to
form acrolein.[26,27] Thus, acrolein formation rates are also given
by Equation (3), in agreement with measured rate data [Equa-
tion (1)]. The effective rate constant in Equation (1) is given by
the product of a propagation rate constant (k4) and the square
root of the ratio of initiation to termination rate constants (k1/
k7), as expected for radical-mediated pathways in the limit of
long chains.[25] The PSSH treatment for the intermediates in-
volved in acetaldehyde formation lead to the rate equation:
k4k
k5
11 PO
2
Scheme 3. 1,3-propanediol oxidation mechanism. The molecules within
squares are 3-hydroxypropanal, the fast decomposition of which leads to
acrolein
ð4Þ
ðracetaldehydeÞ ¼
This equation is consistent with the PO and P13PD dependen-
2
ces on acetaldehyde, shown in Figure 3b and 4b and with the
stronger effects of O2 pressure on acetaldehyde formation
rates compared with acrolein formation rates.
that measured experimentally [Equation (1)]. By analogy with
hydrocarbon autoxidation reactions,[21] we propose a bimolecu-
lar initiation step [Reaction (1)], in which propanediol reacts via
O2 addition into one of its CꢀH bonds to form hydroperoxide
intermediates. The weakest CꢀH bond in diols is at the a-posi-
tion;[21] and is the CꢀH bond that is most susceptible to
oxygen insertion. The resulting ROOH species then decompose
to form *OH and RO* [Reaction (2)],[22] which then react via
propagation steps that abstract H-atoms from a-carbons in
other diol molecules[Reactions (3) and (4)].[23] The a-hydroxy-
alkyl radicals formed in Reactions (3) and (4) can either react
with O2 [Reaction (5)] to form 3-hydroxypropanal and *HO2
Equation (3) for acrolein formation rates can be inserted into
the design equation for a plug-flow reactor to obtain an equa-
tion for the rate constant at each exit conversion (xA), resi-
dence time (t) and temperature (T):
xA
Z
1
t
dxA
k ¼
À
Á
ð5Þ
1=2
3=2
ð1 ꢀ xAÞ qO ꢀ 0:5xA
2
0
in which qO is the inlet O2/diol molar ratio. This equation can
2
*
*
radicals, or with HO2 [Reaction (6)] to form RO* and OH radi-
cals. Reactions (3–6) are bimolecular propagation reactions
that consume and form one radical. The unstable 1,1-dihy-
droxy-3-propanol can undergo intramolecular dehydration to
form 3-hydroxypropanal and water [Reaction (7)]. These propa-
gation reactions can terminate via RO* radical recombination
to form 3-hydroxypropanal, 1,3-propanediol, and oxygen mole-
cules [Reaction (8)]. It is remarkable that 3-hydroxypropanal is
formed—faster than initiation and termination reactions—in a
propagation step [Reaction (5)]; this is the cause of the large
product formation.
be used to estimate the rate constant at each temperature
(Figure 5) and the apparent activation energy (32+/
ꢀ2 kJmolꢀ1). This activation energy is lower than the values
previously reported of hydrocarbon fuel autoxidations (ca.
150 kJmolꢀ1[28]), indicating that OH groups in diols increase
rate constants for initiation and propagation steps. Homogene-
Alternate terminations pathways do not lead to the func-
tional form of the rate data Equation (1) reported here for ho-
mogeneous 1,3-propanediol oxidation. The RO* organic radi-
cals formed in Reaction (2) react with O*, yielding acetalde-
hyde and CH2OH*, which decomposes into formaldehyde via
subsequent oxidation [Reaction (9)]. By analogy with alkanols,
the CꢀC bond in propanediols are expected to be weaker than
the CꢀH or OꢀH bonds.[24] O* radicals form via Reaction (10),
Figure 5. Rate constant for acrolein formation as a function of reciprocal
temperature (12.4 s residence time, 0.47 kPa O2, 0.28 kPa diol)
*
while HO2 decomposes into O2 and H* via Reaction (11).
1066
www.chemsuschem.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2010, 3, 1063 – 1070

Homogeneous Oxidation Reactions of Propanediols at Low Temperatures
ous autoxidation of polypropylene glycol at 348–366 K gave
activation energies of ca. 80 kJmolꢀ1 [29]
These data, taken to-
.
gether with the unreactive nature of propanol and glycerol, in-
dicate a unique effect of the two OH groups in diols on the
rates of H-abstraction reactions.
Glycerol can be converted to acrolein and 1-hydroxyacetone
on acid catalysts,[30] but the homogeneous pathways reported
here lead to much higher acrolein yields and selectivities. 1,3-
Propanediol may be obtained from glycerol or via fermenta-
tion of carbohydrates.[31] These high diol conversions and acro-
lein selectivities lead to acrolein/O2/H2O streams that may be
conveniently introduced into a subsequent catalytic oxidation
reactor to give high yields of acrylic acid.
Figure 6. Effect of the temperature of the catalytic reactor on acrolein and
acetaldehyde conversion, and yields of oxidation products. Feed composi-
tion: 1 kPa 1,3-propanediol, 8.8 kPa O2, 10 kPa water. Homogeneous reactor:
543 K, residence time 14.1 s. Catalytic reactor: 0.2 g of Mo0.61V0.19Ox catalyst,
1.8 g catalyst·(mol 1,3-propanediol)ꢀ1. Acrolein ( ) and acetaldehyde (
~
*
)
~
*
conversions, acrylic acid ( ) and acetic acid ( ) yields. Yields were calculated
as the percentage of the entering 1,3-propanediol that appears as each
product (i.e., acrylic acid and acetic acid).
Coupling homogeneous oxidation of 1,3-propanediol with
catalytic conversion of acrolein products to acrylic acid
The acrolein/O2/H2O mixtures formed via homogeneous oxida-
tion of 1,3-propanediol resemble those used for acrylic acid
synthesis on Mo-V oxide catalysts,[32,33] and is used here in a
tandem reactor strategy to convert diols to acrylic acid without
intervening treatment of the effluent between reactors. The
homogeneous reactor was fed with a mixture of O2 (8.8 kPa)
and 1,3 propanediol (1 kPa, introduced as 10:1 H2O/diol molar
ratio) at 503–543 K (14.1 s residence time; 5.6 cm3 reactor
volume). The H2O added did not influence homogeneous oxi-
dation rates or selectivities at these conditions (503–543 K,
0.14–0.42 kPa 1,3-propanediol; 0.25–1.62 kPa O2). Complete
diol conversions were achieved at 543 K (Table 2) and the acro-
lein yield was 91%.
Homogeneous 1,2-propanediol oxidation pathways
Homogeneous oxidation of 1,2-propanediol also occurred at
conditions similar to described above for 1,3-propanediol con-
version. Acetone was the only C3 product and acetaldehyde,
formaldehyde, and CO2 were also formed via CꢀC bond cleav-
age. Acetone selectivities (23–26%) were much smaller than
for the C3 product (acrolein) of 1,3-propanediol oxidation.
Product selectivities did not depend on residence time
(Table 3), as in the case of 1,3-propanediol, consistent with
Table 3. Conversion, reaction rate, and selectivity towards acrolein, acetal-
dehyde, and formaldehyde in the oxidation of 1,2-propanediol, as a func-
tion of residence time.[a]
Table 2. Conversion and carbon selectivity in the homogeneous oxida-
tion of 1,3-propanediol.[a]
t [s] Conversion [%] ꢀr12PD [mmolcmꢀ3 hꢀ1] Selectivity [%]
Acetone CH3CHO HCHO CO2
T [K]
Conversion [%]
Carbon selectivity [%]
Acrolein
CH3CHO
HCHO
CO2
11.8 20.1
8.8 12.5
7.6 11.6
6.6 10.3
9.7
7.7
8.9
9.5
23
26
25
25
50
49
48
51
7.6
6.2
8.4
8.3
19.4
18.8
18.6
15.7
503
523
543
75
85
99.9
93
93
91
6
6.5
7.5
0.8
0.9
0.9
<0.5
<0.5
<0.5
[a] Conditions: T=430 K, 0.28 kPa 1,2-propanediol, 0.89 kPa O2.
[a] Conditions: 1kPa 1,3-propanediol, 10 kPa H2O, 8.8 kPa O2; 14.1 s resi-
dence time; 5.6 cm3 reactor volume.
their direct formation from 1,2-propanediol. Homogeneous oxi-
dation of 1,2-propanediol gave CH3CHO/(HCHO+CO2) molar
ratios near unity, as expected from their concurrent formation
via a single CꢀC bond cleavage of C3 molecules.
The effluent from this homogeneous reactor was introduced
directly into a subsequent reactor containing Mo0.61V0.19Ox cata-
lyst. Figure 6 shows the effects of temperature (513–613 K) in
the catalytic reactor on acrolein and acetaldehyde conversions
and acrylic acid and acetic acid yields. Acrolein and acetalde-
hyde conversions increased with reactor temperature and
acrylic and acetic acid were the only products detected. This
two-step process using sequential homogeneous–heterogene-
ous reactors produces acrylic acid and acetic acid yields (based
on the 1,3-propanediol fed to the homogeneous reactor) of
91% and 9%, respectively at full conversion of their respective
reactants. Acrylic acid formation rates ranged from 0.0017 to
0.0035 molgcatꢀ1 sꢀ1 at 513 to 613 K reactor temperatures.
Acetone formation rates increased linearly with increasing
diol and O2 pressures (Figures 7a and 8a), while acetaldehyde
formation rates increased with increasing O2 pressure but de-
creased as diol pressures increased (Figures 7a and 8a). These
acetone formation rate data can be accurately described by
the equation:
ðracetoneÞ ¼ kP1;2PDPO
ð6Þ
2
Scheme 4 shows a plausible sequence of elementary steps
for homogeneous oxidation of 1,2-propanediol. By analogy
ChemSusChem 2010, 3, 1063 – 1070
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
1067

E. Iglesia et al.
Scheme 4. Plausible reaction routes for 1,2-propanediol.
Figure 7. Effects of 1,2-propanediol inlet pressure on the formation rate of
a) acetone, and b) acetaldehyde [423 K, residence time 12 s, O2 pressures:
~
&
^
*
0.25 kPa ( ), 0.47 kPa ( ), 0.89 kPa ( ) ,1.62 kPa ( )]. The dashed curves cor-
gives propanal or acetone, depending on the location of the
unpaired electron along the C3 backbone, but propanal was
not detected among reaction products. b-Elimination at the
CꢀH bond forms, via addition of oxygen and elimination of
·HO2, 2-hydroxy-1-propanal or 1-hydroxy-2-propanone.[20]
The elementary reactions proposed in Scheme 5 lead to the
measured dependence rate [Equation (6)] for acetone. The
pathways for acetone formation reaction involve, as in the
case of 1,3-propanediol, bimolecular initiation via O2 addition,
in this case at CꢀH bonds in either a- or b- positions.[18]
respond to the functional dependence of acetone in Equation (6).
Figure 8. Effects of oxygen inlet pressure on the formation rate of a) ace-
tone, and b) acetaldehyde [423 K, residence time 12 s, diol pressures:
Scheme 5. Formation of acetone by homogeneous oxidation of 1,2-pro-
panediol.
~
&
^
0.14 kPa ( ), 0.28 kPa ( ), 0.42 kPa ( )].The dashed curves correspond to the
functional dependence of acetone in Equation (6).
with 1,3-propanediol oxidation, we propose that 1,2-propane-
diol reacts via O2 addition at its a- or b-CꢀH bonds to give the
corresponding radicals. b-Elimination reactions can then cleave
either CꢀO or CꢀH bonds in these radicals. CꢀO bond cleavage
According to Scheme 4, O2 addition occurs at b-CꢀH bonds
in 1,2-propanediol to form the intermediate hydroperoxide
[Reaction (1), Scheme 5]. This species can then decompose to
give OH and organic radicals [Reaction (2)]; the latter propa-
1068
www.chemsuschem.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2010, 3, 1063 – 1070

Homogeneous Oxidation Reactions of Propanediols at Low Temperatures
gate chains via H-abstraction from a- or b-carbons in diols [Re-
action (3)]. The hydroxyl-alkyl radicals formed [Reaction (4)] can
undergo b-elimination to cleave the CꢀO bond and form ace-
tone and OH radicals,[34] while the oxy-alkyl radicals form acet-
aldehyde, formaldehyde, and CO2 in termination steps [Reac-
tion (5)].
here. The symmetry of 1,3-propanediol molecules leads one
predominant product (acrolein), while 1,2-propanediol, lacking
such symmetry, involve more diverse pathways and oxidation
products.
The pathways for acetone formation in Scheme 5 and the
assumptions of steady-state concentrations for all intermedi-
ates in the limit of long radical chains[25] lead to an equation
consistent with measured rate data (details in Supporting In-
formation):
Conclusions
1,2-Propanediol and 1,3-propanediol react with O2 at 400–
500 K via homogeneous pathways unaffected by catalytic or
surface-mediated events. 1,2-Propanediol preferentially reacts
via CꢀC bond cleavage to form CH3CHO, HCHO, and CO2, with
the parallel formation of acetone with modest selectivities. In
contrast, 1,3-Propanediol forms acrolein almost exclusively (~
90% selectivities and yields) and CꢀC cleavage products as mi-
nority species. These rates and selectivities, as well as the mea-
sured dependences of rates on O2 and 1,3-propanediol pres-
sures are consistent with radical-mediated homogeneous path-
ways described in the text. High acrolein yields from homoge-
neous 1,3-propanediol oxidation allow the direct introduction
of the reactor effluent into a subsequent catalytic reaction, in
which acrolein/H2O/O2 mixtures form acrylic acid with ca. 90%
yields based on 1,3-propanediol reactants.
k1
2
ð7Þ
ðracetoneÞ ¼ P1;2PDPO
2
in which the rate constant reflects the kinetic constant of
oxygen addition to the 1,2-propanediol to form the hydroper-
oxide.
The low selectivity to acetone (ca. 25%) in homogeneous
oxidation of 1,2-propanediol is consistent with the parallel for-
mation of a- and b-hydroxy-alkyl radicals, in contrast with the
nearly exclusive formation of the a-hydroxy-alkyl radical in 1,3-
*
propanediol reactions, and with the fact that OH propagating
radicals (but not RO*) leads to acetone. In Scheme 4, the com-
pounds formed via CꢀH bond cleavage of a- and b-hydroxy-
alkyl radicals (1-hydroxy-2-propanone and 2-hydroxy-1-propa-
nal) are subsequently oxidized to 2-oxopropanal. This com-
pound reacts (Scheme 6) with *OH radicals to form a hydroxya-
cid, which then undergoes H-abstraction and CꢀC bond cleav-
age by b-elimination to form 2-propen-1-ol (which forms acet-
Experimental Section
The gas-phase oxidation of propanediols was carried out at ambi-
ent pressure in a vertical quartz tube (length 50 cm, volume
9 cm3), heated resistively and equipped with a concentric axial
thermowell and a type-K thermocouple. Quartz granules (1.0–
1.5 mm diameter) were used to vary the empty heated volume
within the reactor vessel. 1,2-Propanediol or 1,3-propanediol reac-
tants (Aldrich, 99.6%) were introduced as liquids with a syringe
pump (Cole Parmer 74900 Series) into a flowing gas stream in a
vaporization volume held at 420 K. Helium (Praxair, 99.999%) was
used as a diluent and the O2 co-reactant was introduced as a 10%
O2/He mixture, (Praxair). Molar rates were metered by electronic
controllers (Bronkhorst) and all transfer lines were kept at 420 K to
prevent condensation. Temperatures (400–600 K), flow rates (1.4ꢂ
10ꢀ4–8.3ꢂ10ꢀ4 moldiolhꢀ1), and O2/diol ratios (0–3.2) were varied
systematically throughout these experiments.
˙
aldehyde via tautomerism) and HOCO radicals, which form CO2
via H-abstraction.[35]
Reactants and products were analyzed by gas chromatography
(Hewlett–Packard 5890) using
a Carboxen-1000 column (60–
80 mesh, 5.22 mꢂ3.18 mm) with thermal conductivity detection
and a methyl silicone capillary column (HP-1; 50 mꢂ0.32 mm,
1 mm film) with flame ionization detection. The identity of reaction
products was determined from the elution time of known com-
pounds and their speciation was confirmed by mass spectrometry
(HP-6890/5973, 50 m HP-1 column). Diol conversions are reported
as the percentage of the entering reactants converted to products.
Selectivities are reported on a carbon basis as the percentage of
the converted diol reactants appearing as each product. Residence
time is defined as the ratio of the reactor volume to the inlet volu-
metric rate at standard conditions (STP). Carbon balances were
>95% in all experiments. The presence of organic peroxides in
diol reactants was ruled out for both 1,2- and 1,3-propanediol reac-
tants by using Whatman indicators (<10 ppm). Peroxides were not
detected in either reactant, indicating that the observed homoge-
neous reactions are not initiated by adventitious peroxide impuri-
ties.
Scheme 6. Pathways of 2-oxopropanal decomposition.
1,2-Propanediol reacts in presence of oxygen to give ace-
tone, acetaldehyde and formaldehyde as main reaction prod-
ucts (Table 3). Acetone is formed as a product of CꢀO bond
cleavage of b-hydroxy-alkyl radical. Acetaldehyde and formal-
dehyde are the decomposition products of 2-oxopropanal,
which forms via reactions of the CꢀH bond cleavage products
of a and b-hydroxyl-alkyl radicals. Reaction schemes proposed
for both 1,3-propanediol (Scheme 3) and 1,2-propanediol
(Scheme 5) are based on radical-like pathways and give rate
equations and selectivities consistent with the data reported
ChemSusChem 2010, 3, 1063 – 1070
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
1069

E. Iglesia et al.
A Mo-V oxide (Mo0.61V0.19Ox) catalyst[34] was used to convert acrole-
in, present in the effluent from the homogeneous 1,3-propanediol
oxidation reactor, into acrylic acid; this catalyst also converts acet-
aldehyde to acetic acid. The catalyst was prepared by mixing a so-
lution containing C2O4H2 (Fluka, 99%) and NH4VO3 (Sigma–Aldrich,
99%) in deionized water with one containing (NH4)6Mo7O24·4H2O
(Aldrich, 99.98%) at ambient temperature. The liquid was evaporat-
ed at 363 K while stirring and the residual solids were treated at
393 K in ambient air overnight and then in flowing dry air at 673 K
for 4 h (Praxair, extra dry).
[5] A. Corma, G. W. Huber, L. Sauvanaud, P. O’Connor, J. Catal. 2008, 257,
163–171.
[6] G. Knothe, J. Krahl, J. V. Gerpen, The Biodiesel Handbook, AOCS, Cham-
paign 2005.
[7] G. A. Kraus, Clean: Soil, Air, Water 2008, 36, 648–651.
[8] G. Chuchani, R. M. Dominguez, A. Rotinov. I. Martin, Int. J. Chem. Kinet.
1997, 29, 851–854.
[9] J. Shen, W. Shan, Y. Zhang, J. Du, H. Xu, K. Fan, W. Shen, Y. Tang, Chem.
Commun. 2004, 2880–2881.
[10] J. Shen, W. Shan, Y. Zhang, J. Du, H. Xu, K. Fan, W. Shen, Y. Tang, J.
Catal. 2006, 237, 94–101.
[11] E. Sokolova, Electrochim. Acta 1979, 24, 147–155.
[12] H. Huser, J. M. Leger, C. Lamy, Electrochim. Acta 1985, 30, 1409–1414.
[13] C. Alonso, J. Gonzꢃlez-Velasco, J. Appl. Electrochem. 1988, 18, 538–545.
[14] L. Giebeler, P. Kampeb, A. Wirth, A. H. Adamsa, J. Kunert, H. Fuess, H.
Vogel, J. Mol. Catal. A: Chem. 2006, 259, 309–318.
Acrolein oxidation was carried out in a tubular packed bed reactor
with plug-flow hydrodynamics placed after the homogeneous re-
actor. The reactor contained Mo0.61V0.19Ox (0.2 g) diluted with acid-
washed quartz powder (1 g) to prevent temperature gradients. The
homogeneous reactor was fed with a mixture containing 1,3-pro-
panediol (1 kPa), O2 (8.8 kPa), and H2O (10 kPa) at 543 K and 14.1 s
residence time. Complete conversion of 1,3-propanediol was
reached in the first reactor, with selectivities of 91%, 7.5%, and
0.9% for acrolein, acetaldehyde, and formaldehyde. This stream
was then introduced into the catalytic reactor described above. Re-
actant and product concentrations were measured by gas chroma-
tography using the same procedures as for the homogeneous re-
actor.
[15] Z. R. Ismagilov, S. N. Pak, V. K. Ermoleav, J. Catal. 1992, 136, 197–201.
[16] F. Wang, J.-L. Dubois, W. Ueda, J. Catal. 2009, 268, 260–267.
[17] J. I. Steinfeld, J. S. Francisco, W. L. Hase, Chemical Kinetics and Dynamics,
2
nd Ed., Prentice-Hall, New Jersey 1999.
[18] S. R. Logan, Can. J. Chem. 1991, 69, 540–544.
[19] N. I. Butkovskaya, N. Pouvesle, A. Kukui, Y. J. Mu, G. Le Bras, J. Phys.
Chem. A 2006, 110, 6833–6843.
[20] I. Hermans, T. L. Nguyen, P. A. Jacobs, J. Peeters, ChemPhysChem 2005,
6, 637–645.
[21] A. G. Davies, Organic Peroxides, Butterworth, London 1961.
[22] V. Warth, F. Battin-Leclerc, R. Fournet, P. A. Glaude, G. M. Come, G. Scac-
chi, Comput. Chem. 2000, 24, 541–560.
[23] C. F. Cullis, E. A. Warwicker, Proc. R. Soc. London Ser. A, 1961, 264, 392–
407.
Acknowledgements
[24] Y.-R. Luo, Handbook of Bond Dissociation Energies in Organic Com-
pounds, CRC Press, Boca Raton 2003.
[25] M. Boudart, Kinetics of Chemical Processes, Prentice-Hall, New Jersey
1968.
[26] L. H. Slaugh, J. P. Arhancet (Shell Oil Company), US 5304686, 1994.
[27] E. Tsukuda, S. Sato, R. Takahashi, T. Sodesawa, Catal. Commun. 2007, 8,
1349–1353.
[28] J. M. Pickard, E. G. Jones, Energy Fuels 1996, 10, 1074–1077.
[29] M. A. Semsarzadeh, H. Salehi, Eur. Polym. J. 2000, 36, 1001–1010.
[30] A. Neher, T. Haas, D. Arntz, H. Klenk, W. Girke (Degussa Aktiengesell-
schaft), US 5387720, 1995.
Prof. W. H. Green (Massachussets Institute of Technology) and
Prof. S. OrdꢁÇez (University of Oviedo) are acknowledged for
their helpful mechanistic discussions. E.D. acknowledges financial
support from the Spanish Ministry of Science and Innovation
(Josꢂ Castillejo Program). This work was supported by BP as part
of the Methane Conversion Cooperative at the University of Cali-
fornia at Berkeley and by the Director, Office of Basic Energy Sci-
ences, Chemical Sciences Division, US Department of Energy
under contract DE-AC02–05CH11231.
[31] D. M. Adkesson et al. (E. E. Dupont de Nemours and Company and A. E.
Staley Manufacturing Company), WO 101479, 2004.
[32] T. V. Andrushkevich, G. Ya Popova, Russ. Chem. Rev. 1991, 60, 1023–
1034 .
Keywords: diols · oxidation · radicals · reaction mechanisms ·
sustainable chemistry
´
[33] J. Tichy, Appl. Catal. A: Gen. 1997, 157, 363–385.
[34] R. Janik, W. Gerhardt, J. Prakt. Chemie (Leipzig) 1989, 331, 611–616.
[35] S. Yamabe, N. Tsuchida, S. Yamazaki, J. Org. Chem. 2006, 71, 1777–1783.
[36] X. Li, E. Iglesia, Appl. Catal. A: Gen. 2008, 334, 339–347.
[1] G. W. Huber, J. N. Chheda, C. J. Barrett, J. A. Dumesic, Science 2005, 308,
1446–1450.
[2] G. W. Huber, S. Iborra, A. Corma, Chem. Rev. 2006, 106, 4044–4098.
[3] A. Corma, S. Iborra, A. Velty, Chem. Rev. 2007, 107, 2411–2502.
[4] M. Pagliaro, R. Ciriminna, H. Kimura, M. Rossi, C. D. Pina, Angew. Chem.
2007, 119, 4516–4522; Angew. Chem. Int. Ed. 2007, 46, 4434–4440.
Received: May 19, 2010
Revised: June 20, 2010
Published online on September 9, 2010
1070
www.chemsuschem.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2010, 3, 1063 – 1070