Evaluation of a monolith-supported Pt/Al2O3 catalyst for wet oxidation of carbohydrate-containing waste streams

Article abstract of DOI:10.1021/es000887z

Catalytic oxidation of glucose and cellulose has been demonstrated in a monolith reactor, a novel contacting device for the oxidation of carbohydrate feedstocks that allows the processing of nonsoluble components without reactor plugging. Catalytic enhancement is observed for glucose oxidation, and the catalyst promotes selectivity to two-carbon carboxylic acids. It is proposed that catalytic oxidation of glucose occurs through parallel pathways: thermal oxidation to a wide range of organic acids and selective catalytic oxidation to low molecular weight acids. On the other hand, cellulose oxidation was not always enhanced by the presence of the catalyst. Here, the effect of the catalyst was to enhance the conversion of the organic acids produced during thermal oxidation. However, these organic acids also catalyzed the primary conversion of cellulose, thus the conversion of cellulose decreased as the reaction temperature was increased. A kinetic model is provided that is consistent with the inverse temperature effect observed for the cellulose oxidation experiments. Catalytic oxidation of glucose and cellulose has been demonstrated in a monolith reactor, a novel contacting device for the oxidation of carbohydrate feedstocks that allows the processing of nonsoluble components without reactor plugging. Catalytic enhancement is observed for glucose oxidation, and the catalyst promotes selectivity to two-carbon carboxylic acids. It is proposed that catalytic oxidation of glucose occurs through parallel pathways: thermal oxidation to a wide range of organic acids and selective catalytic oxidation to low molecular weight acids. On the other hand, cellulose oxidation was not always enhanced by the presence of the catalyst. Here, the effect of the catalyst was to enhance the conversion of the organic acids produced during thermal oxidation. However, these organic acids also catalyzed the primary conversion of cellulose, thus the conversion of cellulose decreased as the reaction temperature was increased. A kinetic model is provided that is consistent with the inverse temperature effect observed for the cellulose oxidation experiments.

Full text of DOI:10.1021/es000887z

Environ. Sci. Technol. 2000, 34, 3480-3488
Wet oxidation is defined as an aqueous-phase oxidation
Evaluation of a Monolith-Supported
Pt/Al2O3 Catalyst for Wet Oxidation
of Carbohydrate-Containing Waste
Streams
process brought about when an organic and/ or oxidizable
inorganic-containing liquid is mixed thoroughly with a
gaseous source of oxygen (usually air) at temperatures of
150-325 °C. Gauge pressures of 2000-20 000 kPa are
maintained to promote reaction and control evaporation
(3). Elevated temperatures and pressures enhance the
solubility of oxygen and provide a strong driving force for
oxidation.
T R E N T A . P A T R I C K A N D
M A R T I N A . A B R A H A M *
Recent reviews of wet air oxidation describe the chemistry
(4, 5), process technologies (6), and reactor design issues (7)
surrounding this waste treatment process. The liquid effluent
of a wet oxidation process is generally composed of low
molecular weight compounds consisting primarily of car-
boxylic acids and other carbonyl group compounds; CO2
and water are produced as the ultimate end products if the
reaction conditions are severe enough. Both homogeneous
catalysts (particularly soluble copper complexes) and het-
erogeneous catalysts (including copper supported on base
metal oxides) have been used to increase the rate of oxidation
(5). A recent study showed that the catalytic wet oxidation
of acetic acid could be achieved using a Pt/ Al2O3 catalyst
supported on a monolith (2). Two-phase flow within the
channels of the monolith provided a high rate of mass transfer
for oxygen through the liquid phase and to the surface of the
catalyst, thereby enhancing the conversion of the reactant
(8).
Department of Chemical and Environmental Engineering,
The University of Toledo, Toledo, Ohio 43606
Catalytic oxidation of glucose and cellulose has been
demonstrated in a monolith reactor, a novel contacting
device for the oxidation of carbohydrate feedstocks that
allows the processing of nonsoluble components without
reactor plugging. Catalytic enhancement is observed
for glucose oxidation, and the catalyst promotes selectivity
to two-carbon carboxylic acids. It is proposed that
catalytic oxidation of glucose occurs through parallel
pathways: thermal oxidation to a wide range of organic
acids and selective catalytic oxidation to low molecular weight
acids. On the other hand, cellulose oxidation was not
always enhanced by the presence of the catalyst. Here,
the effect of the catalyst was to enhance the conversion of
the organic acids produced during thermal oxidation.
However, these organic acids also catalyzed the primary
conversion of cellulose, thus the conversion of cellulose
decreased as the reaction temperature was increased. A
kinetic model is provided that is consistent with the
inverse temperature effect observed for the cellulose
oxidation experiments.
Selective oxidation (for example, the conversion of glucose
to gluconic acid) can be achieved through the use of
supported noble metal catalysts (9) and mild reaction
temperatures. Wet oxidation of glucose at 110-140 °C in a
neutral solution obtained conversions up to 40% and yielded
a diverse set of reaction products including gluconic acid,
glucaric acid, glucosone, 5-ketogluconic acid, arabonic acid,
and various products with 4 or less carbon atoms (10). In one
case (11), the conversion and the rate of catalyst deactivation
were shown to be dependent upon the pH of the solution,
but the selectivity to gluconic acid was relatively independent
of pH. Palladium supported on alumina was more active
than palladium on activated carbon for the conversion of
glucose to gluconic acid at 55 °C and yielded complete
conversion in only 2 h (12).
Introduction
As the duration of human occupied space missions becomes
longer, the current method of taking the essential elements
needed to sustain life becomes increasingly less feasible. As
a result, processes for the regeneration of air, water and food
need to be designed using a combination of physicochemical
and biological technologies. Such a system, known as a
controlled ecological life support system (CELSS), is closed
to mass transfer with its surroundings (1).
A CELSS is defined as a life support system that relies
heavily on biological subsystems for recycling (2). In a CELSS,
a system must be developed that combines the photosyn-
thetic activity of plants with physicochemical processes to
regenerate waste material into food and oxygen for a crew
(1). One such process under investigation to regenerate waste
material is catalytic wet oxidation. The main advantages of
wet air oxidation in a CELSS environment are the recovery
of useful water and the reduction of solid wastes to a very
small volume and weight of sterile, nondegradable ash. In
addition to that, other advantages include the production of
carbon dioxide and residual aqueous inorganic matter that
can be used as plant nutrients and the comparatively low
energy requirements, since water does not have to be
evaporated (2).
Although much experimental work has been correlated
using pseudo-first-order kinetics, a more detailed reaction
sequence (13) has emerged that successfully captures the
complex nature of the reaction process. The feedstock is first
converted to organic intermediates, acetic acid, or CO2
through parallel reaction pathways. The organic intermedi-
ates are unstable and are rapidly converted to acetic acid or
CO2. The acetic acid is relatively stable and is only converted
to CO2 if the reaction conditions are sufficiently severe. A
kinetic model based on these reaction pathways has now
been developed (14), and the model has also been extended
for use with heterogeneously catalyzed systems (15).
Solid waste in a CELSS includes human fecal waste,
packaging waste from food products, inedible plant biomass,
food system wastes, and other subsystems waste products,
such as salts, filters, and media. One common feature of all
the solid waste in a CELSS is the high fiber content, most of
which is cellulose. The current research effort focused on
the catalytic wet oxidation of glucose and cellulose using the
monolith catalyst previously shown to be effective. These
two model chemicals were chosen because they are common
products of the biomass conversion pretreatment step, acid
hydrolysis.
* Corresponding author tel: (419)530-8092; fax: (419)530-8086;
e-mail: mabraham@eng.utoledo.edu.
9
3 4 8 0 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 16, 2000
10.1021/es000887z CCC: $19.00
2000 Am erican Chem ical Society
Published on Web 07/14/2000

FIGURE 1. Schematic diagram of experimental system.
Acid hydrolysis reactions have the capability of converting
large particles of inedible biomass (wheat straw particles)
into soluble organic compounds (generally, 5- and 6-carbon
monosaccharides) and small particles of cellulose and lignin.
Acid hydrolysis of cellulose occurs at a faster rate than
enzymatic methods, but enzymatic methods are cleaner and
more selective than acid hydrolysis (16). A combination of
low temperatures (25-40 °C) and high acid concentration
provides glucose yields similar to those obtained using
temperatures over 200 °C and lower acid concentrations
around 1% (16). Alternatively, complete conversion of
cellulose can be obtained through reaction in supercritical
water at 400 °C (17).
The current research program was undertaken to evaluate
the potential to use a Pt/ Al2O3 catalyst supported on a
monolith for the conversion of aqueous streams containing
glucose and/ or cellulose. The monolith was deemed par-
ticularly appropriate for conversion of cellulose-containing
streams since the large channels of the monolith would not
be plugged by the insoluble cellulose particles. The reaction
products were identified, and simple rate expressions were
utilized to compare the conversion of both target compounds.
The effectiveness of the heterogeneous catalyst in treating
the waste stream containing solid particles can be evaluated
through the reaction pathways.
conductivity detector for on-line gas-phase analysis. A 1/ 4-
in. clear shield surrounds the entire high-pressure experi-
mental system. All experiments were conducted at a liquid
flow rate of 27.8 mL/ min and a gas flow rate of 1.7 standard
L/ min.
Monolith bricks, 62 channels/ cm2, coated with bare
alumina (Al2O3) washcoat were obtained from Allied-Signal
Corporation. Three monolith cores (4.9 cm in diameter and
10.9 cm in length) were stacked in the reactor. The top core
was angled 15° from the horizontal to allow liquid to drain
from the cores into the separator. The monolith cores were
wrapped with Fibrafax, a ceramic fiber insulating paper, 0.16
cm thick, manufactured by The Carborundum Company.
The wrap expands when wet, creating a tight seal between
the cores and the reactor wall. This serves to keep the cores
in place and also prevents gas and liquid from bypassing the
catalyst.
Platinum catalyst was prepared at the University of Toledo
using the method of incipient wetness. A 0.5% platinum by
weight solution was prepared by dissolving 3.189 g of
platinum(II) acetylacetonate into 400 mL of acetone. The
mixture was vigorously stirred until all of the metal salt was
dissolved. Before being impregnated, the monolith cores had
been previously dried in an oven at 120 °C for 45 min to
remove any atmospheric moisture that may have ac-
cumulated on the catalyst washcoat. After being dried, the
cores were allowed to cool for approximately 3 min and
weighed.
Each core was dipped into the platinum solution for 5
min and then inverted and again submerged into the solution
for 5 min. The solution entered the monolith channels by
capillary action. The cores were then allowed to air-dry and
were placed in nitrogen at 50 °C for 2 h. After being placed
in nitrogen, the monoliths were calcined in air at 250 °C for
6 h and reweighed. The weight difference of the cores after
calcination was subtracted from the weight of the cores before
impregnation. This weight was assumed to be representative
of the amount of platinum deposited on each core. The total
weight (combined on all three cores) of platinum impregnated
on the monolith cores was 0.6365 g, giving a platinum loading
of 0.26 wt %.
Experimental Section
The experimental system used for the catalytic wet oxidation
of glucose and cellulose consisted of three separate sub-
systems: feed, reactor, and separation. A schematic repre-
sentation of the system is included in Figure 1 and described
in greater detail elsewhere (8). The feed system was made up
of gas and liquid feeding units consisting of metering
equipment, heating equipment, high-pressure pump, and
liquid containers. The reactor system consisted of the frothing
section, the monolith catalyst, and a tubular reactor. The
separation system consisted of two separate gravity-induced
separation units, cooling tank, back pressure regulator, two
micrometering valves, and metering equipment. The gas
exiting the reactor was sent to a Hewlett-Packard 5980 gas
chromatograph (GC) containing an 80/ 100 mesh Porapak N
0.635 cm × 304.8 cm stainless steel column and a thermal
9
VOL. 34, NO. 16, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 4 8 1

TABLE 1. Summary of First-Order Rate Constants
temp
activation
energy (kJ/mol)
reactant catalyst (°C)
rate constant (min-1
)
glucose hydrolysis 118 1.80 × 10-4 ( 1.02 × 10-4 26.0 ( 22.2
136 3.86 × 10-4 ( 1.39 × 10-4
205 8.72 × 10-4 ( 1.16 × 10-4
Al2O3
110 6.13 × 10-4 ( 1.19 × 10-4 53.7 ( 21.2
149 2.98 × 10-3 ( 1.04 × 10-3
158 2.61 × 10-3 ( 9.94 × 10-5
167 7.04 × 10-3 ( 7.68 × 10-4
60 1.85 × 10-4 ( 7.61 × 10-5 38.8 ( 15.5
84 5.53 × 10-4 ( 1.08 × 10-4
Pt/Al2O3
110 2.64 × 10-3 ( 7.84 × 10-5
162 4.49 × 10-3 ( 2.61 × 10-4
126 1.51 × 10-3 ( 3.24 × 10-4 18.5 ( 15.9
156 1.88 × 10-3 ( 1.74 × 10-4
185 3.13 × 10-3 ( 4.45 × 10-4
87 3.81 × 10-3 ( 4.87 × 10-4
cellulose Al2O3
FIGURE 2. Comparison of glucose conversion obtained during
hydrolysis at 205 °C, thermal oxidation at 158 °C, and catalytic
Pt/Al2O3
oxidation over Pt/Al2O at 162 °C. Curves represent first-order model
3
111 2.70 × 10-3 ( 1.58 × 10-4
155 2.13 × 10-3 ( 4.44 × 10-4
185 2.16 × 10-3 ( 2.47 × 10-4
predictions.
Glucose concentration was measured through an enzy-
matic assay. Glucose (HK) reagent was ordered from Sigma
Diagnostics. Glucose is first phosphorylated by adenosine
triphosphate (ATP), in the reaction catalyzed by hexokinase
(HK). Glucose-6-phosphate (G-6-P) formed is then oxidized
to 6-phosphogluconate (6-PG) in the presence of nicotina-
mide adenine dinucleotide (NAD). This reaction is catalyzed
by glucose-6-phosphate dehydrogase (G-6-PDH). During this
oxidation, an equimolar amount of NAD is reduced to NADH.
The consequent increase in absorbance at 340 nm is directly
proportional to glucose concentration. The assay was cali-
brated using standard samples containing known amounts
of glucose.
The concentration of cellulose was measured through a
similar procedure starting with cellulase enzyme. Cellulose
was converted to its monomer glucose, and then the glucose
concentration was determined using the method previously
described. An incubation time of 2 h was chosen for the
cellulose decomposition reaction. Although this did not
provide sufficient time to convert all of the remaining
cellulose to glucose, a linear calibration curve relating the
absorbance at 340 nm to the cellulose concentration was
obtained.
Selected liquid-phase samples were also evaluated using
a Shimadzu HPLC, operating with a Spherisorb Octyl, 25 cm
× 4.6 cm i.d., 5 µm particles liquid chromatograph column.
The mobile phase used for the analysis was 0.2 M phosphoric
acid. The column had a flow rate of 0.8 mL/ min, and a UV
detector was used with the wavelength set at 210 nm.
FIGURE 3. Evaluation of the activation energy from the first-order
model for the decomposition of glucose.
(C6H10O6)n + 6nH2O f 6nCO2 + 11nH2
The only component detected in the gas phase was CO2;
carbon monoxide, the result of partial oxidation, was not
detected in any of the experiments performed. Hydrogen
could not be detected using the gas chromatograph.
Because the liquid phase is completely recycled, the
monolith froth reactor can be modeled as a batch reactor.
With that in mind, the longer the glucose and cellulose stayed
in the reactor, the higher the conversion of the reactants to
completely oxidized and partially oxidized products. Con-
version (XA) is defined mathematically as
moles of A reacted
moles of A fed
XA )
(1)
Results and Discussion
The complete oxidation of glucose yields carbon dioxide and
If a constant volume is considered, a reasonable assumption
for the current system, the material balance on the batch
reactor system provides (18)
water according to the following stoichiometric equation:
C6H12O6 + 6O2 f 6CO2 + 6H2O
dX
C
) -rA
(2)
When carried out in water, the oxidation reaction competes
with a hydrolysis reaction that yields hydrogen instead of
water:
A,0 dt
where -rA is expressed as mol/ (vol)(time). Often the rate of
a catalytic reaction is written in terms of the amount of catalyst
within the system or in the case of a monolith catalyst, the
wall surface area. Since the volume, the wall surface area,
and the catalyst weight are all fixed, simply multiplying by
the volume of liquid and dividing by the mass (or wall area)
of catalyst converts eq 2 into the alternative form. Regardless,
eq 2 allows the calculation of a macroscopic reaction rate by
measuring the change in glucose or cellulose concentration
as a function of experimental run time.
C6H12O6 + H3O f 6CO2 + 12H2
In a similar way, the complete oxidation of cellulose
11n
2
(C6H10O6)n +
O2 f 6nCO2 + 6H2
competes with cellulose hydrolysis:
9
3 4 8 2 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 16, 2000

FIGURE 4. Comparison of the product spectra obtained from thermal and catalytic oxidation of glucose.
We have utilized the integral data obtained from these
experiments to determine first-order reaction rate constants
for each experimental condition studied. Substituting the
first-order rate expression
negligible as compared to thermal or catalytic oxidation.
Comparison of the oxidation experiments performed at
approximately 160 °C reveals that glucose conversion by
thermal oxidation is significant even in the presence of the
platinum catalyst, although the catalyst generally increases
the glucose conversion.
-rA ) kCA,0(1 - X)
(3)
Glucose conversion by all mechanisms increased mono-
tonically with increasing temperature, as shown by the
Arrhenius plot in Figure 3. At all temperatures, the hydrolysis
rate constant was approximately an order of magnitude lower
than were the rate constants for either of the oxidation
pathways. Analysis of the first-order rate constants provided
an estimate for the activation energy as 26.0 ( 22.2 kJ/ mol
for hydrolysis, 53.7 ( 21.2 kJ/ mol for thermal oxidation, and
38.8 ( 15.5 kJ/ mol for catalytic oxidation. The effect of the
catalyst was to increase the rate of reaction at lower
temperatures and to decrease the activation energy of the
reaction; at the highest temperature, the thermal reaction
competes effectively with the catalytic case.
into eq 2, integrating, and rearranging provides
-ln(1 - X) ) kt
(4)
Conversion of the experimental data into the form indicated
by eq 4 allows estimation of the pseudo-first-order rate
constant. These values are reported in Table 1, along with
the statistical error representing the 95% confidence limits.
Because each experiment was performed at several reaction
temperatures, the first-order rate constants can be correlated
according to the Arrhenius equation:
k ) A exp(-EA/ RT)
(5)
A qualitative comparison of the reaction products is
provided in Figure 4, which indicates that a substantially
different product spectrum results from catalytic oxidation
than from thermal oxidation (comparison is for product
spectrum at 360 min). A large array of organic acids and
furans are produced, including oxalic, malic, lactic, succinic,
and acetic acids and 4-(hydroxymethyl)furfural. In the
presence of the catalyst, the yield of all of the products (with
the exception of oxalic acid) appears to be substantially
reduced. The appearance of oxalic acid as a stable product
is expected from previous experiments showing that it is
only converted to CO2 at temperatures greater than those
used within this research program (19).
The values calculated for the activation energies are also
reported in Table 1. Error analysis of the activation energy
was completed using 80% confidence limits because of the
small number of temperatures at which experiments were
completed.
Glucose Conversion. The conversion of glucose as a
function of time is compared for three different experimental
conditions in Figure 2, in which the symbols represent the
experimental values and the curves show the predictions of
the first-order rate constants provided in Table 1. The
hydrolysis experiment performed at 205 °C indicates that
glucose conversion through this reaction pathway was
9
VOL. 34, NO. 16, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 4 8 3

FIGURE 7. Yield of reaction products from catalytic oxidation (Pt/
Al2O ) of glucose at 162 °C.
3
FIGURE 8. Comparison of cellulose conversion obtained during
thermal oxidation at 126 °C and catalytic oxidation over Pt/Al2O
3
at 111 °C. Curves represent first-order model predictions.
FIGURE 5. Description of the major partial oxidation products
observed during thermal and catalytic oxidation of glucose and
cellulose.
followed by slightly lesser amounts of malic and succinic
acids with yields of approximately 0.05, and then lactic acid
with a yield of approximately 0.03. Fumaric and oxalic acid
were observed in yields of less than 0.01. Several other peaks
observed on the HPLC could not be assigned to specific
products. The carbon mass balance (mass of carbon quanti-
fied in the products/ mass of carbon in initial glucose charged)
was slightly less than 1 for all experiments, averaging 0.932
(standard deviation ) 0.080) over the course of the experi-
ment. This value is consistent with the large array of
unidentified products observed on the HPLC, each with a
small peak area.
In the case of catalytic oxidation (Figure 7), a smaller total
yield of partial oxidation products was observed, and the
distribution of products was substantially different than that
from thermal oxidation. Succinic and malic acids, produced
in significant yield during thermal oxidation, were the major
partial oxidation products of catalytic oxidation, with yields
of 0.03 g of product/ initial g of glucose. Oxalic acid, which
was previously observed as a minor product, reached a yield
of 0.025 g of product/ g of glucose, close to that observed for
malic and succinic acids. A small yield of 5-(hydroxymethyl)-
furfural reached a maximum of 0.005 at an intermediate
reaction time, after which its yield decreased. It is likely that
5-(hydroxymethyl)furfural is catalytically converted to the
four carbon organic acids through a sequential oxidation
pathway. The carbon mass balance was quite good for this
case, with an average value of 0.997 g of C product/ initial g
of glucose and a standard deviation of 0.037, suggesting that
all major products of catalytic oxidation have been identified.
Cellulose Decom position. Figure 8 provides a comparison
of the conversion of cellulose through thermal (Al2O3)
oxidation at 126 °C and catalytic (Pt/ Al2O3) oxidation at 111
°C. Cellulose conversion by the catalytic pathway was greater
than by the thermal pathway, even though the noncatalytic
FIGURE 6. Yield of reaction products from thermal oxidation (Al2O
monolith) of glucose at 158 °C.
3
To obtain a quantitative measure of the yield, the peak
area obtained from the HPLC analysis was calibrated with
liquid-phase concentration (ppmw) using known standards
to obtain an instrument response factor. This allowed each
peak to be assigned to an individual species and the area for
each peak to be related to a mass concentration (ppmw).
The species identified on the HPLC included an assortment
of low molecular weight organic acids and furans, as described
in Figure 5. On the basis of the mass yield of each species,
we obtained (and report our results as) a nondimensional
yield for each product relative to the mass of the reactant
initially in solution.
Figure 6 provides the temporal variation of product yields
from thermal oxidation of glucose at 158 °C. Acetic acid is
seen as the major product with a yield of approximately 0.06,
9
3 4 8 4 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 16, 2000

rate constant decreased as the temperature increased.
Although it is unusual for the overall reaction rate to decrease
with increasing temperature, such a result is possible when
the overall reaction is comprised of several kinetically
significant reaction steps. This evaluation serves as the basis
of the discussion that follows later.
In a manner similar to that observed with glucose, catalytic
oxidation of cellulose produced much lower yield of partial
oxidation products, as seen in Figure 10. The only product
with a large HPLC response was oxalic acid. Quantification
showed that the yield of this product was approximately 0.01,
comparable to the yields of fumaric and succinic acids.
However, in the case of thermal oxidation, higher yields of
all partial oxidation products were detected and are reported
quantitatively in Figure 11. In this case, the yield of succinic
acid reached 0.015; the yields of oxalic, fumaric, and acetic
acids were each approximately 0.005. Malic acid was also
observed and reached a maximum yield of 0.006 at an
intermediate reaction temperature. The carbon balance for
cellulose decomposition by thermal oxidation averaged 0.907
(standard deviation ) 0.091), and the carbon balance for
cellulose decomposition by catalytic oxidation averaged 1.013
(standard deviation ) 0.016). This confirms the large number
of partial oxidation products obtained in the thermal case
that were not observed during catalytic oxidation and the
likelihood that all products of catalytic oxidation were
identified using our analytical procedure.
FIGURE 9. Evaluation of the activation energy from the first-order
model for the decomposition of cellulose.
experiment was conducted at a higher temperature. The
values of the first-order rate constants, shown in Table 1,
further indicate the effectiveness of the heterogeneous
catalyst. No reactor plugging was observed at any conditions.
Note that cellulose is not soluble in water and is present as
a solid phase distinct from the heterogeneous catalyst and
therefore not likely to directly interact with the platinum
metal. Thus, it is interesting that the heterogeneous catalyst
is effective in accelerating the conversion of cellulose.
The effect of temperature on cellulose conversion is
described by the Arrhenius plot of Figure 9. For oxidation
without a catalyst, the reaction follows normal Arrhenius
behavior with an activation energy of 18.5 ( 15.9 kJ/ mol.
However, in the case of catalytic oxidation, the first-order
Finally, the conversion obtained during catalytic oxidation
of cellulose is compared to both catalytic and thermal
oxidation of glucose in Figure 12. Clearly, the cellulose is
more difficult to oxidize, with the catalytic oxidation of
cellulose being slower than the thermal oxidation of glucose.
However, as noted previously, there is no reason to believe
that the heterogeneous catalyst should interact with the solid
FIGURE 10. Comparison of the product spectra obtained from thermal and catalytic oxidation of cellulose.
9
VOL. 34, NO. 16, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 4 8 5

FIGURE 13. Proposed reaction pathways describing the conversion
of cellulose and glucose during catalytic wet oxidation over Pt/
FIGURE 11. Yield of reaction products from thermal oxidation (Al2O
3
Al2O .
monolith) of cellulose at 158 °C.
3
TABLE 2. First-Order Model Parameters for Catalytic Oxidation
of Cellulose
pre-exponential
(min-1
activation energy
(kJ/mol)
)
k0
kH
kcat
0.328
248.5
553.6
27.3
20.9
24.3
cellulose feed. Acid-catalyzed hydrolysis of cellulose is well-
known (16, 17, 20) and is one of the primary methods of
converting these insoluble feedstocks. It is the catalysis by
the acid products that leads to the inverse temperature
behavior observed in this work. As the reaction temperature
increases, the rate of catalytic oxidation is enhanced. As a
result, the acid concentration decreases, and the overall
conversion of cellulose is reduced.
FIGURE 12. Comparison of glucose conversion from thermal
oxidation at 158 °C and catalytic oxidation at 162 °C with cellulose
conversion from catalytic oxidation at 156 °C.
Further verification of this proposition is obtained through
kinetic modeling of the reaction steps. Following the
procedure outlined in Hill (21) and allowing the decomposi-
tion of cellulose to occur through acid catalysis, a mass
balance on the batch reactor system provides
cellulose particles. The slow rate of oxidation may be
indicative of the need for cellulose to depolymerize to the
extent that it can dissolve in the liquid phase prior to
undergoing oxidation. This theory is consistent with the
postulate of Adschiri et al. (17). Note that the products
observed from cellulose oxidation were the same as those
observed during glucose oxidation, justifying the choice of
glucose as an appropriate model compound to evaluate the
oxidation of carbohydrate wastes.
dC
) -(k0 + kH[H+])C
(6)
dt
where C represents the concentration of cellulose, [H+]
indicates the total acid concentration, and k0 and kH are the
rate constants for thermal and acid-catalyzed decomposition
of cellulose, respectively. If the glucose decomposition
reaction is fast relative to cellulose dissolution, then the
change in the concentration of acid with respect to reaction
time can be modeled as
Reaction Pathway Modeling. Catalytic oxidation of
glucose reveals that this reaction proceeds through parallel
reaction paths. Thermal oxidation provides a large array of
reaction products, including oxalic, succinic, and malic acids,
and 5-(hydroxymethyl)furfural. Catalytic oxidation gave
higher conversion of glucose and much greater yield of oxalic
acid. However, many of the same products are observed in
both cases, with comparable yields. Thus, we suggest that
the platinum catalyst selectively oxidizes glucose to oxalic
acid, while thermal oxidation accounts for the formation of
the other acid products.
d[H+]
) (k0 + kH[H+])C - kcat[H+]
(7)
dt
While the catalytic oxidation of glucose can be simply
described as conversion by parallel reaction pathways
(catalytic and thermal oxidation), this simple explanation
cannot account for the apparent negative activation energy
calculated during cellulose oxidation. Rather, we propose a
more detailed reaction pathway scheme, as shown in Figure
13. Here, we suggest that cellulose decomposition occurs
through acid-catalyzed depolymerization, forming a soluble
organic species depicted in Figure 13 as glucose. Then,
glucose oxidation occurs through parallel catalytic and
thermal oxidation pathways. The important component,
however, is that the organic acids that are produced through
the thermal conversion of glucose are then converted to CO2
through catalytic oxidation. Thus, the intermediate acid
products are essential for the initial depolymerization of the
where kcat represents the pseudo-first-order rate constant
for the catalytic conversion of acid products to permanent
gases. Although this pair of simultaneous differential equa-
tions cannot be solved analytically, it is possible to evaluate
the constants in eqs 6 and 7 by numerical integration of the
equations. Then, the constants are obtained by minimizing
the sum of squares error between the model prediction and
the experimental values.
Using the data from all reaction times and temperatures
provides 48 experimental measurements from which six
parameters must be evaluated. These values are shown in
Table 2, and the model predictions are compared with the
experimental values (for two temperatures) in Figure 14. The
variance of prediction, described as
9
3 4 8 6 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 16, 2000

FIGURE 14. Model prediction of acid concentration as a function of temperature and reaction time for catalytic oxidation of cellulose.
Evaluation of these results clearly suggests that conversion
of nonsoluble carbohydrate species such as cellulose should
likely be accomplished using a reactor system that allows
accumulation of the intermediate acidic products. For
example, a dual-reactor system consisting of a continuous
stirred tank reactor followed by a plug flow reactor would
allow dissolution of the cellulose through acid catalysis
followed by decomposition of the acidic products through
oxidation catalysis. The CSTR would then be optimized to
provide high conversion of cellulose to intermediate products,
without concern for the conversion of the intermediates.
The PFR is optimized for conversion of the intermediates to
achieve the required destruction efficiency of the waste
treatment process. On the other hand, we have also shown
that selective production of a specific reaction intermediate
can be achieved through catalytic conversion of cellulose
within a single vessel, raising the potential for partial catalytic
oxidation to serve as a means of converting carbohydrate
feedstocks into useful chemicals.
ssq
n - p
var )
where n is the number of experimental measurements and
p is the number of parameters, was determined to be 0.014.
Although the overall fit of the model is good, it is clear from
Figure 14 that the model predictions are better for the higher
temperature.
It is informative to compare the activation energies
obtained from the model predictions with those calculated
previously using psuedo-first-order kinetics. According to
the model, the rate constant k0 corresponds to noncatalyzed
thermal decomposition. The activation energy for the
constant k0 was found to be 20.9 kJ/ mol, which compares
favorably with the activation energy obtained from the
pseudo-first-order model of thermal decomposition of cel-
lulose calculated as 18.5 kJ/ mol. In previous experiments
with catalytic oxidation of oxalic acid using soluble copper
sulfate, an activation energy of approximately 80 kJ/ mol was
reported (19). The value of 24.3 kJ/ mol obtained in our model
suggests that platinum may be a better catalyst for conversion
of inorganic acids, as seen during the oxidation of acetic acid
(2). In addition, the model accounts for the conversion of
organic acids that are more easily oxidized than oxalic acid.
Figure 14 also reveals the model prediction of the acid
concentration, which provides further insight into the role
of the acidic products. At low temperature, the rate of
decomposition for the acid products is slow, allowing a fairly
high concentration to accumulate within the reactor. This
catalyzes the conversion of cellulose, providing the char-
acteristic “S”-shaped conversion curve. The maximum yield
of acidic products is predicted to be approximately 0.05 at
100 min, corresponding to the point at which the maximum
rate of cellulose conversion occurs. At long reaction times
(greater than 200 min), the conversion of cellulose is
approximately constant, since the thermal pathway is slow
at this low temperature. However, at the high temperature,
the acidic products are rapidly decomposed through the
catalytic reaction, and the highest yield of less than 0.01 occurs
at about 25 min. Thus, conversion of cellulose through acid
catalysis becomes a minor pathway at elevated temperatures,
and the thermal pathway becomes dominant. These com-
peting pathways for conversion of cellulose give rise to the
apparent negative activation energy within the temperature
range of these experiments.
Acknowledgments
This research was supported by NASA Ames Research Center,
under Joint Research Interchange NCC2-5215. The authors
are grateful to NASA, and in particular, Mr. John Fisher, Lead
Engineer of the Regenerative Life Support Group, for support
of this research effort. The HPLC was purchased from funds
provided by the NSF under Grant BES-9802747.
Literature Cited
(1) Eisenberg J. N.; Maszle, D. R.; Pawlowski, C. W.; Auslander, D.
M. J. Aerospace Eng. 1995, 139-147.
(2) Klinghoffer, A. A.; Cerro, R. L.; Abraham, M. A. Catal. Today
1998, 40 (1), 59-72.
(3) Copa, W. M.; Gitchel, W. B. Wet Oxidation. In Standard
Handbook of Hazardous Waste Treatment and Disposal;
Freemean, H. M., Ed.; McGraw-Hill: New York, 1989; Section
8.8.
(4) Mishra, V. S.; Mahajani, V. V.; Joshi, J. B. Ind. Eng. Chem. Res.
1995, 34, 2-48.
(5) Imamura, S. Ind. Eng. Chem. Res. 1999, 38, 1743-1753.
(6) Kolaczkowski, S. T.; Plucinski, P.; Beltran, F. J.; Rivas, F. J.;
McLurgh, D. B. Chem. Eng. J. 1999, 73, 143-160.
(7) Mills, P. L.; Chaudhari, R. V. Catal. Today 1999, 48, 17-29.
(8) Klinghoffer, A. A.; Cerro, R. L.; Abraham, M. A. Ind. Eng. Chem.
Res. 1998, 37 (4), 1203.
(9) Gallezot, P. Catal. Today 1997, 37, 405-418.
(10) Skaates, J. M.; Briggs, B. A.; Lamparter, R. A.; Baillod, C. R. Can.
J. Chem. Eng. 1981, 59, 517-521.
9
VOL. 34, NO. 16, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 4 8 7

(11) Abbadi A.; van Bekkum, H. J. Mol. Catal. A: Chem. 1995, 97,
111.
(18) Fogler, H. S. Elements of Chemical Reaction Engineering, 3rd
ed.; Prentice Hall: New York, 1999.
(12) Nikov, I.; Paev, K. Catal. Today 1995, 24, 41.
(13) Li, L. X.; Chen, P. S.; Gloyna, E. F. AIChE J. 1991, 37 (11), 1687-
1697.
(14) Zhang. Q.; Chuang,. K. T. AIChE J. 1999, 45, 145-150.
(15) Hamoudi, S.; Belkacemi, K.; Larachi, F. Chem. Eng. Sci. 1999,
54 (15-16), 3569-3576.
(19) Shende, R. V.; Mahajani, V. V. Ind. Eng. Chem. Res. 1994, 33,
3125-3130.
(20) Nguyen, Q. A.; Tucker, M. P.; Keller, F. A.; Beaty, D. A.; Connors,
K. M.; Eddy, F. P. Appl. Biochem. Biotechnol. 1999, 77 (9), 133-
142.
(21) Hill, C. G., Jr. Chemical Engineering Kinetics and Reactor Design;
John Wiley & Sons: New York, 1977.
(16) Camacho, F.; Gonzalez-Tello, P.; Jurado, E. J. Chem. Technol.
Biotechnol. 1996, 67, 350-356.
(17) Adschiri, T., M.; Sasaki, B. M.; Kabyemela, and K. Arai. Cellulose
Hydrolysis in Supercritical Water. In Reaction Engineering for
Pollution Prevention; Abraham, M. A., Hesketh, R. P., Eds.;
Elsevier Press: Amsterdam, 2000.
Received for review January 10, 2000. Revised manuscript
received April 27, 2000. Accepted June 9, 2000.
ES000887Z
9
3 4 8 8 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 16, 2000