Vol Ii-4

Download as pdf or txt
Download as pdf or txt
You are on page 1of 60

Final Report

Air Products and Chemicals, Inc.

An Investigation Into The Acceptability Of Stabilized Methanol


As A Fuel For Phosphoric-Acid Fuel Cell Power Systems

February, 2001

Dr. Vernon P. Roan, Principal Investigator


Mr. J. Patrick Sterchi, Research Assistant

University of Florida Fuel Cell Research and Training Laboratory

Department of Mechanical Engineering


Gainesville, Florida
1

Table of Contents

Page
Introduction................................................................................................................

Experimental Approach............................................................................................

Background..........................................................................................................

1st Generation Rig..................................................................................


Vaporizer Design............................................................................
Reactor Design ...............................................................................
Heating Element Control................................................................
Summary ........................................................................................
2nd Generation Rig.................................................................................
1st Generation Minor Modifications...............................................
Vaporizer Design............................................................................
Superheater Design ........................................................................
Reactor Design ...............................................................................
Condensing Unit.............................................................................
Control Software ............................................................................
Summary ........................................................................................

7
7
8
10
13
13
14
16
16
16
18
21
22

........................................................................................................
Results
st
1 Generation Rig Results ....................................................................
Commercial Grade Methanol Results ............................................
Stabilized Methanol Results...........................................................
Stabilized Methanol Re-Evaluation ...............................................
Conclusions Based on 1st Generation Results ................................
2nd Generation Rig Results ...................................................................
Higher Temperature Reforming Results ........................................
Parametric Study of Contaminated Methanol Feedstock...............
Parametric Study Results ...............................................................

24
24
24
25
25
26
26
26
28
30

Conclusions ......................................................................................................

42

Recommendations ............................................................................................

45

Appendix A Stabilized Methanol Analysis Results .........................................

47

Appendix B Typical Gas Chromatograph Results ...........................................

53

Bibliography.....................................................................................................

58

Introduction
While pure gaseous hydrogen is the ideal fuel for fuel cell power systems, it is also
relatively expensive, difficult to transport, and difficult to store onboard transportation systems.
Consequently, there has been, and is, considerable interest in utilizing liquid hydrocarbon fuels
for transportation fuel cell applications. For a liquid fuel to be a feasible alternative to hydrogen,
it must be possible and practical to utilize a fuel processing system to transform the liquid fuel
into a hydrogen-rich gas. More specifically, the hydrogen-rich gas must be compatible with
long-term operation of the fuel cell system within which it would be utilized, with little or no
degradation in performance.
Clearly, the implication is that a suitable fuel-fuel processor combination must maintain
the ability to provide almost complete conversion of the liquid fuel feedstock into an acceptable
hydrogen-rich gas over a long period of time and many cycles of operation. Acceptable
hydrogen-rich gas in this case means a maximum concentration of hydrogen and virtually zero
concentration of contaminants which would degrade the fuel cell stack life and/or performance.
To meet this acceptability requirement, the fuel processor must also not undergo any significant
decrease in its ability to convert the liquid hydrocarbon in a continuing and consistent manner. A
prime mechanism for the fuel processor to experience a significant decrease in conversion ability
would be through a deactivation of the catalyst in a catalyst-driven processor such as the steam
reformer.
Catalyst deactivation can occur for many reasons including excess temperature, rapid
temperature transients, or the introduction of oxygen. However, with the introduction of a new
3

liquid fuel, it is obvious that deactivation can occur as a result of one or more contaminants
contained in the liquid fuel.
The production of stabilized methanol through a process developed by Air Products, Inc.
with the support of the Department of Energy represents the potential introduction of a new
liquid fuel produced from the immense supply of domestic coal. However, in its as produced
form, the stabilized coal-based methanol does not have the purity of commercial grade methanol.
It contains small amounts of a number of compounds including a type of mineral oil used in
the production process.
Air Products and DOE subsequently supported research to determine the feasibility of the
as produced stabilized methanol in various types of engines as well as for fuel cell
applications. The University of Florida had facilities and equipment in the Fuel Cell Research
and Training Laboratory to perform this type of research, as well as two operational methanolfueled fuel cell engines. As a result, an agreement was reached for the University of Florida to
undertake the evaluation of the stabilized methanol as a suitable fuel for steam-reformed
methanol/phosphoric acid fuel cell systems.
Since it was obviously important not to damage the operational fuel cell engines, the
proposed procedure was to fabricate small steam reformers using catalyst pellets of the type
(copper oxide-zinc oxide) typically used for methanol steam reformation. The apparatus was
designed so as to allow determination of areas of deactivation if they occurred as a result of the
stabilized methanol fuel. A parallel reformer operating simultaneously with commercial grade
methanol was used as a control.
Initial results showed very rapid degradation of the catalyst exposed to the stabilized
methanol. Results pointed to a probable coating of the catalyst with non-reacted mineral oil as

the likely cause. Subsequent tests and modifications to the rig were performed as well as the
utilization of a higher temperature catalyst to avoid rapid catalyst degradation. Higher
temperatures did, as expected, result in less effect due to the mineral oil but it also resulted in a
more unfavorable balance among hydrogen, methane, carbon dioxide, and carbon monoxide in
the reformate.
The results to date indicate that for the relatively low-temperature steam reforming of
methanol, the mineral oil would most likely have to be removed prior to being a suitable fuel.
On the other hand, the limited higher temperature results indicate that possibly for auto-thermal
reforming and probably for partial oxidation reforming, the stabilized methanol would be
acceptable as a fuel. Obviously, further experimentation is needed to verify these applications.

Experimental Approach
Background
One of the primary concerns of using stabilized methanol in a fuel cell system was the
presence of possible contaminant species in the methanol. Chemical analysis by multiple
independent testing facilities indicated the presence of small quantities of species other than
methanol present in the stabilized methanol. Partial results of these analyses are shown in Table
1. Copies of the complete analyses are located in Appendix A.

Table 1: Chemical Analysis of Stabilized Methanol (only 4 dominant specie are shown)
Air Products Inc.
Atlantic Analytical Intertek Testing
Services
LaPorte, TX Data*
Species
Laboratory
(% w/w)
(% w/w)
(% w/w)
Methanol
98.7
97.23
Ethanol
0.18
0.1433
0.678
Water
0.46
0.4286
0.529
Mineral Oil**
0.49
0.2208
0.189
* A different sample was used for LaPorte, TX data
** Actual chemical composition unknown but similar to common mineral oil

Of particular interest with respect to the different species present in the stabilized
methanol was the mineral oil. The effect that mineral oil might have on fuel cell components was
uncertain, although there was the possibility that it could have a detrimental effect on fuel cell
system hardware. For this reason, instead of using the stabilized methanol initially in the
brassboard fuel cell engine at the University of Florida (and risk permanent damage to the
system), a sub-scale experimental steam reformer rig was designed and built to determine the
reformability of the stabilized methanol.
6

1st Generation Rig


The rig initially built to conduct research using stabilized methanol (1st generation rig)
consisted of two identical reformer flow paths. This design was used so that during
experimentation, commercial grade methanol would flow through one side of the rig as a control
while stabilized methanol design would flow through the other. This allowed for a direct, realtime comparison of the behavior of the two types of methanol within each respective fuel
processor. The temperature range (250C-300C) and pressure (1atm) used during
experimentation are typical of methanol steam-reforming, and flow rates used during
experimentation were scaled to mimic the existing 25kW laboratory brassboard fuel cell engine
steam reformer. The catalyst used was a low-temperature water-gas shift catalyst (CuO/ZnO on
alumina) that is typical for the methanol steam-reforming process, and similar to that used in the
brassboard fuel cell engine reformer. Each flow path included a pumping sub-system, a vaporizer
sub-system, and a reactor sub-system. Figure 1 shows a general schematic showing how these
components were integrated to form either one of the two reformer flow paths.
Vaporizer Design
Calibrated peristaltic pumps were used to pump the water and methanol from their
respective reservoirs to the vaporizer sub-system. During experimentation, a steam to carbon
ratio of approximately 1.5:1 was used to simulate operating conditions in the brassboard system.
Each vaporizer consisted of a stainless steel tube wrapped with an electrically resistive heating
element (heat tape) as shown in Figure 2. High temperature insulation covered the entire
assembly to reduce heat loss to the environment. Each side of the entire experimental rig had one
vaporizer assembly, which consisted of separate water and methanol vaporizers. The vaporizer

Figure 1: 1st Generation Experimental Methanol Steam-Reformer Rig Schematic


(one of two parallel flow paths)

assembly was outfitted with thermocouples as shown in Figure 2, and was capable of producing
superheated vapor at 300C for the flow rates used over the course of experimentation.

Reactor Design
Superheated vapor from the vaporizers was piped to the reactor sub-system (Figure 3)
through stainless steel tubing. The reactor sub-system included ten electrically resistive
heating elements (heat bands) that added energy to the reactor bed in order to offset the
endothermic steam-reforming reaction and maintain near isothermal operation. Thermocouples
were used to monitor the temperature at the core of the reactor at the locations shown in Figure

Figure 2: Vaporizer Assembly Schematic

3, and these temperature measurements were used to thermostatically control the heat bands
individually. Removable end fittings were used at each end of the reactor to allow for removal
and replacement of catalyst. A picture of the 1st generation reactor assembly prior to installation
can be seen in Figure 4. After installation, the reactor assembly was insulated with high
temperature insulation to minimize heat loss to the environment.
Heating Element Control
Heating elements were controlled remotely by a computer in conjunction with data
acquisition hardware and a power distribution unit (PDU). Computer software utilized
temperature data from the rig to determine the desired status (on/off) of any given heating
element. A digital output module controlled the operation of the heating elements through the
power distribution unit. A schematic of the control setup is shown in Figure 5. The power
distribution unit, shown in Figure 6, housed a bank of solid state relays, with a specific relay
controlling each heating element. Each relay was energized by a 5V-power source, while the
switch side connected a 120VAC circuit to each heating element. In order to measure the power
input to each heating element, a power meter was installed in the heating element circuit. A
switchboard allowed for a single meter to monitor any one of the heating elements at a time.
Figure 7 shows the configuration of the heating element control circuit.
CO2 Purge System
One additional feature of the experimental rig was a CO2 purge system that was used to
purge hydrogen from the system after use as well as maintain the catalyst bed in a non-oxidizing
environment. Implementation of this system was straightforward and is shown in schematic form
in Figure 8.

10

Figure 3: 1st Generation Reactor Schematic

Figure 4: 1st Generation


Reactor Assembly

11

Figure 5: Control System Setup


Figure 6: Power Distribution Unit (PDU)
Figure 7: Heating Element Control Circuit
Figure 8: CO2 Purge System

(not available electronically)

12

Summary
The 1st generation rig was a useful tool in determining the feasibility of reforming
stabilized methanol using standard methanol steam-reforming operating conditions and catalysts.
Attempts to reform stabilized methanol under standard conditions in the 1st generation rig were
unsuccessful in the context of use in a fuel cell. It quickly became apparent that additional
research should be conducted in order to determine the cause(s) of, and possible solution(s) to,
the lack of ability to reform stabilized methanol at standard operating conditions. However, the
existing rig was limited in its capability to modify certain parameters, particularly increasing the
reaction temperature. Therefore, it was proposed that an improved rig be built (2nd generation
rig) that would be a modification of the 1st generation rig with some new components, and would
allow for increased flexibility with respect to operating conditions compared to the 1st rig.
2nd Generation Rig
For the 2nd generation rig, the basic premise of the 1st generation design was left
unchanged- two identical flow paths, one running chemical grade methanol, the other running
some form of contaminated methanol, were used to compare the reforming characteristics of two
types of feedstock. In this manner, a relative performance level could be characterized for a
contaminated feedstock relative to an uncontaminated feedstock.
Design of the 2nd generation reforming rig involved modifying the existing rig as well as
adding new components to improve the functionality of the unit. A schematic of the 2nd
generation experimental rig is shown in Figure 9.

13

Figure 9: 2nd Generation Experimental Methanol Steam-Reformer Rig Schematic


(one of two parallel flow paths)

1st Generation Minor Modifications


Several components of the 1st generation rig were left essentially unchanged.
Control hardware was for the most part identical to the 1st generation rig, with only slight
modifications in some hardware components. The pumping system was changed only slightly, as
the 2nd generation rig had the ability to monitor the flow rate of the liquid directly through the
use of a digital scale that measured the weight of each reservoir. In addition, in the 2nd generation
rig pumps were set up for remote operation and monitoring. A schematic of the pumping system
used is shown in Figure 10, and a picture of the pumping sub-system is shown in Figure 11.

14

Figure 10: Pumping System Schematic

Figure 11: Pumping


Sub-System Assembly

15

Vaporizer Design
Liquid methanol and water were pumped to separate vaporizers that vaporized and
superheated the fluid to a maximum temperature of approximately 300C. The 1st generation
design for the vaporizers was found to work well, so an identical design was used in the 2nd
generation rig.
Superheater Design
The superheater was a new component for the 2nd generation design. Upon leaving the
vaporizers, the mixed methanol and water vapor were piped to the superheater, which was
located immediately prior to the reactor assembly. The purpose of the superheater section was to
raise the temperature of the methanol-water vapor to the desired operating temperature of the
reactor, with a maximum temperature of 450C. The design of the superheater was similar to that
of the vaporizers, and can be seen in Figure 12. The superheater section was 15 in length, and
utilized an electrically resistive ceramic beaded heater to supply heat. The surface temperature of
the superheater was monitored at the approximate location shown on Figure 12 to ensure the
heating element was below its maximum operating temperature. The superheater heating element
was controlled based on the inlet temperature to the reactor assembly. In this way, the inlet
temperature to the reactor could be controlled by the superheater assembly upstream, ensuring a
predictable reactor inlet temperature. The entire superheater assembly was insulated using high
temperature insulation.
Reactor Design
The reactor portion of the rig was completely redesigned compared to the 1st generation
model. This was necessary due to the use of a different type of heating band

16

Figure 12: Superheater Assembly Schematic

17

that allowed for higher operating temperatures. These heaters had a maximum temperature rating
of over 800C, which was about 300C higher than in the 1st design. A schematic of the reactor
is shown in Figure 13, and a photograph of the completed reactor prior to installment on the rig
can be seen in Figure 14. Six heat bands were used on each reactor, resulting in six reaction
zones. Heat could be supplied to each of these zones individually, allowing for nearly
isothermal operation. Thermocouple probes at the reactor core, as well as surface thermocouples
at the outer surface of the reactor were used to collect temperature data that could be used to
control heat band operation. An additional feature of the 2nd generation design was the ability to
take gas samples at multiple locations throughout the reformer bed. The locations of the gas ports
are shown on Figure 13, and plug valves that could be opened or closed manually controlled
operation.
Condensing Unit
A condensing unit was required to cool reformate product gas. This process was
necessary to condense excess water, as well as non-reacted hydrocarbons, out of the product
stream. The condensing unit consisted of a condensing coil submerged in an ice water bath. This
ice water bath was contained in a refrigeration unit to delay melting as long as possible. The
temperature of the dry product gas exiting the condensing unit was monitored to ensure adequate
cooling had occurred to condense a majority of the condensable products. Condensable
products accumulated in a liquid trap and were collected for analysis. Figure 15 shows a
schematic of the condensing unit assembly.

18

Figure 13: Reactor Sub-System Schematic

19

Figure 14: 2nd Generation


Reactor Assembly

Figure 15: Condensing Unit Schematic

20

Control Software
Computer control programming was developed using Labview software. This program
allowed for data input in the form of thermocouple signals, and data output in the form of digital
signals to rig components. In addition, data logging and storage was done automatically by the
control program. For several items, including pump and exhaust fan operation, switch (on/off)
logic was all that was required. Heating elements, however, could be controlled manually or
automatically by control logic.
The automatic control of the heat bands utilized several different techniques in an attempt
to maintain nearly isothermal operation. The first control parameter was a simple thermostatic
control. Temperature settings for each of the thermocouple inputs were adjustable and controlled
the operation of respective heating elements. Each heating element was controlled by two
separate thermocouple readings. One thermocouple monitored the outer surface temperature of
the rig, another thermocouple monitored the fluid temperature at the core of the reactor. The
allowable temperature for each heating band limited the surface temperature, while the fluid
temperature was set at the desired reaction temperature. Table 2 lists the thermocouple control
signals for each respective heating element. Nomenclature is referred to Figure 2, Figure 12, and
Figure 13. Note that OR logic was used for the surface temperature of heat bands 1 through 6, as
either one of the two signals could indicate an overheating problem.

21

Table 2: Thermocouple Control Assignments


Heating Element
A. Fluid
TC Control
W
M
S
1
2
3
4
5
6

W
M
1
2
3
4
5
6
7

B. Surface
TC Control
SW
SM
SS
S1 & S2
S2 & S3
S3 & S4
S4 & S5
S5 & S6
S6 & S7

Summary
The 2nd generation rig is an experimental rig that is both versatile and practical. Figure 16
shows the complete experimental rig. The design allows for a wide range of operating
conditions, yet allows for repairs and maintenance to be timely in the event of component failure.
In many ways, the steam reforming system described is ideal for performing a parametric study
of operational variables that might impact the steam reforming process, including the presence of
hydrocarbon impurities as is the case with stabilized methanol.

22

Figure 16: 2nd Generation Experimental Rig

23

Results

Results obtained during experimentation can be divided into two categories, namely, 1st
generation rig results and 2nd generation rig results. The 1st generation rig was primarily used to
evaluate the feasibility of reforming stabilized methanol under standard methanol
steam-reforming conditions. The 2nd generation rig was used for two purposes, first, to
investigate using a higher temperature process with a different catalyst to reform stabilized
methanol, and second to perform a parametric study regarding the effect of several hydrocarbon
impurities on the methanol steam-reforming process.
1st Generation Rig Results
Commercial Grade Methanol Results
Initially, experimentation using commercial grade methanol was necessary to prove the
functionality of the experimental rig. The results of this work were favorable, as the rig appeared
to reform the methanol and water input into a hydrogen-rich gas product according to design
expectations. This was evidenced by several indicators, including the following:
i) analysis of collected condensate indicated a very low level of methanol present,
corresponding to a high level of feedstock conversion
ii) input energy required to maintain the reaction temperature was near (10%) values
predicted by thermodynamic calculations
iii) dry product analyzed in gas chromatograph consisted of predicted species in near
predicted quantites
These observations indicated that the experimental rig was indeed capable of reforming
methanol at high conversion levels under standard process operating conditions.
24

Stabilized Methanol Results


During stabilized methanol experimentation, one side of the rig was used to reform
stabilized methanol, while the other was used to reform commercial grade methanol. The results
of experimentation were immediate and obvious. The first indication of a difference in the
reforming potential of the two feedstocks was the difference in power usage by each reactor. The
stabilized methanol showed a significant decrease in power consumption, which indicated a
decrease in overall conversion of methanol. Further analysis of reformate obtained from the
stabilized methanol process showed a decrease in carbon monoxide levels, which is consistent
with more incomplete conversion. Finally, the condensable products of the reaction were
analyzed and shown to have high concentrations of methanol relative to the commercial grade
methanol feedstock. These observations led to the conclusion that with the current configuration
(operating conditions and catalyst), the reforming process using stabilized methanol was
unsatisfactory, especially for fuel cell applications. In order to ensure the validity of this
conclusion, a re-evaluation was conducted under the same operating conditions using a fresh
catalyst bed.
Stabilized Methanol Re-Evaluation
Each reactor was dismantled, cleaned and loaded with fresh catalyst prior to a reevaluation of the stabilized methanol feedstock. Using the same procedure as previous
experimentation, similar results were obtained during the second attempt at reforming stabilized
methanol. These results seemed to confirm the conclusion that the stabilized methanol
as-produced is unsuitable for reforming under standard methanol steam-reforming conditions.

25

Conclusions based on 1st Generation Results


At the time, it was theorized that the oil present in the stabilized methanol might act to
somehow deactivate the catalyst, resulting in a reduction of feedstock conversion in the reactor.
It was possible that at least some of the oil remained in liquid form at rig operating temperatures
(due to its relatively high boiling point), resulting in the physical deposition of oil on the catalyst,
effectively blocking catalytic sites and reducing active surface area. Development of a 2nd
generation experimental rig was intended to allow for experimentation that would either support
or refute this hypothesis, as well as allow for a parametric investigation that could determine
acceptable levels (if any) of oil in the feedstock fuel.
2nd Generation Rig Results
Two different lines of investigation were explored using the 2nd generation experimental
rig. First, higher reforming temperatures and a different catalyst were used to determine the
feasibility of reforming stabilized methanol at higher temperatures than those typically
associated with methanol steam-reforming. Second, a parametric study was conducted using
commercial grade methanol feedstock that had been contaminated with controlled quantities of
mineral oil and other contaminants.
Higher Temperature Reforming Results
Initial experimentation conducted on the 2nd generation rig was completed using
commercial grade methanol and a standard high-temperature water-gas shift catalyst. The
catalyst used was iron oxide-chromium oxide in composition and was intended for use in the
range of 350C to 450C. In this case, the reaction temperature selected was 400C, with all
other operating conditions (including space velocity) similar to standard methanol steamreforming conditions (as in 1st generation rig experimentation).

26

Results obtained during the higher temperature experimentation were (as expected)
markedly different than those obtained using lower temperatures and the Cu-Zn catalyst. A
summary of some results is presented in Table 3, which shows some typical results for each case
(high and low temperature) under similar operating conditions (except for temperature).

Table 3: Typical Reforming Results w/ Commercial Grade Methanol


Reforming
Methanol
C. Dry Reformate Analysis
Method

Low Temp(275C)
Cu-Zn catalyst
High Temp(400C)
Fe-Cr catalyst

Conversion
CO

(% by volume)
CO2

>99%

1-2%

23.5-24.5%

94-98%

4-5%

21-23%

CH4
<100ppm
10-12%

Experimentation using stabilized methanol was conducted under identical conditions


compared to the commercial grade feedstock. Gas chromatograph results indicated that product
species were nearly identical for the two cases. However, there was a slight difference in species
concentration. The concentration of methane was about 1% higher in the case of stabilized
methanol, while the concentration of carbon monoxide was approximately 1% lower. The
conversion level of the stabilized methanol was nearly identical to that of the commercial grade
methanol. Experimentation was continued over several days for a total of approximately 20
hours of run time. Over this period, there was no noticeable change in the product gas
composition or conversion level of the higher temperature stabilized methanol reforming
process.
During the condensation process, some type of oil film was evident on the surface of the
condensate, presumably mineral oil. This indicated that at least a small portion of the oil passed

27

unreacted through the reactor in vapor form. If used in a PEM fuel cell, this oil could
contaminate the fuel cell anode, possibly in a very short amount of time.
Results from higher temperature reforming research seem to indicate that the reaction
temperature is a significant factor when considering the effect of hydrocarbon impurities on the
steam-reforming process. It is possible that raising the reaction temperature above the boiling
point of the potential contaminant species is at least partially effective in reducing the
contaminating effect of the offending species. However, while raising the temperature of the
reaction may ensure the species will remain in vapor form, it does not necessarily result in the
decomposition of the contaminant in question.
Parametric Study of Contaminated Methanol Feedstock
Experimentation conducted using stabilized methanol indicated that some type of
contamination in the methanol results in an almost immediate degradation of the catalyst bed,
severely limiting the ability to reform the methanol into a hydrogen rich gas. It was proposed that
the mineral oil present in the stabilized methanol was the probable cause of the observed
catalyst deactivation, although a parametric study was required to confirm this hypothesis.
Experimentation proceeded along two lines of investigation. One plan was developed to
test only commercial grade methanol and establish a database that could be compared to data
obtained using contaminated feedstock. A second plan was developed to research the effect of
several specific contaminants on the reforming process.
Commercial grade experimentation. All commercial grade methanol experimentation was
conducted using a 1.5:1 molar ratio of water to methanol feedstock. ICI-Katalco 51-3 lowtemperature shift catalyst was used during all phases of investigation. Two different reaction
temperatures were examined, 255 and 275 degrees centigrade. Since methanol is a liquid at room

28

temperature and pressure, determining a standard for the gas phase of methanol is problematic.
Because of this, a space velocity based on the volumetric flow rate of liquid methanol was
utilized. The range of space velocities used during the course of experimentation was roughly
0.5hr-1 to 3.0hr-1. It was found that this range of flow rates corresponded to the region were
methanol conversion begins to drop off for the reaction temperatures indicated. That is, at a
space velocity less than 0.5hr-1, the conversion of methanol is very near 100%, above 0.5hr-1,
conversion levels begin to diminish. Since future evaluation would presumably require some
measurable change in performance, it was desirable to conduct experimentation in a region were
any degradation effect would be readily apparent.
Contaminated feedstock experimentation. The operating conditions used during
experimentation using contaminated feedstock were similar to those using commercial grade
feedstock. There were no changes in the type of catalyst used, the temperature range used, the
flow rates applied, or the water to methanol molar ratio. Three different types of contaminants
were used, including a lightweight (Drakeol-10) mineral oil supplied by Air Products, isooctane,
and kerosene. Each contaminant was evaluated over a range of concentrations. The mineral oil
was tested in the range of 0.05% to 0.5% by weight. This concentration was determined based on
the mass of methanol only, not a premix of methanol and water. Isooctane and kerosene were
tested in the range of 0.1% to 1.0% by weight. In order to reduce the necessity of catalyst
replacement and reduction, a single catalyst load was used over the course of experimentation for
any specific contaminant. For this reason, experimentation was initiated at a relatively low
concentration so that any affect on the catalyst would be gradual in occurrence. After completing
experimental runs using each of the contaminants, a series of commercial grade runs were used
to evaluate the potential removal of contamination from the catalyst bed. Upon completion of

29

experimental runs for each contaminant, the reactor was dismantled, cleaned, and charged with a
fresh load of catalyst.
Parametric Study Results
One of the important issues regarding a parametric study of a chemical reaction is the
control and monitoring of reaction temperature. The reaction temperature is exponentially related
to the reaction rate as expressed by the Arrhenius reaction model. This emphasizes the necessity
of accurately monitoring and controlling reaction temperature in order to determine reaction rates
based on conversion of the reactant species. Since a perfectly isothermal reactor is not possible to
obtain, changes in the temperature of the reactor in time and space must be considered in an
analysis of experimental results.
Temperature
Over the course of experimentation, the uncertainty in the actual temperature from point
to point and instant to instant within the reactor was probably the single largest contributing
factor to uncertainty in experimental results. Figures 17 and 18 show how the temperature at one
of the seven reactor thermocouple locations (along the centerline of the reactor) would change
with time during a typical experimental run. In the case shown, the temperature setpoint for the
reactor was 250C, and the space velocity was 1.57hr-1 (based on the liquid flow rate of
methanol). Other operating conditions produced similar results with respect to the range of
temperatures about the set-point. In general, the measured temperature remains between 245C
and 260C for a majority of the run, with an exceptional spike up to near 270C or down to
240C in some cases.

30

Commercial grade results


Evaluation of commercial grade (no contaminant) feedstock was conducted at two
different reaction temperatures, 255C and 275C. Data obtained at 255C is presented in
Figures 19 and 20.
Figure 19 shows the measured fractional methanol conversion level for separate
experimental runs at varying space velocities. A linear curve fit to the data is included to
emphasize the drop off in feedstock conversion as the space velocity (and therefore flow rate)
increases past a certain point. For fuel cell applications, a conversion level of 0.99 (99%) would
be the absolute minimum allowable conversion level without risking damage to the fuel cell
stack. The dotted line on Figure 19 is intended to indicate this limit in terms of space velocity for
the given reaction temperature.
Figure 20 presents a solution for the rate constant of the methanol decomposition
reaction,
km

CH 3OH CO + 2H 2

31

Tem perature (C)

280

T1

270
260
250
240
0

30

60
Run Tim e (m in)

90

120

90

120

90

120

90

120

Tem perature (C)

280

T2

270
260
250
240
0

30

60
Run Tim e (m in)

Tem perature (C)

280

T3

270
260
250
240
0

30

60
Run Tim e (m in)

Tem perature (C)

280

T4

270
260
250
240
0

30

60
Run Tim e (m in)

Figure 17: Typical Instantaneous Temperature Measurements over Duration of Experimental


Run for Thermocouples 1 through 4

32

Figure 18: Typical Instantaneous Temperature Measurements over Duration of Experimental


Run for Thermocouples 5 through 7
Figure 19: Fractional Methanol Conversion Using Chemical Grade Feedstock at 255C
Figure 20: Rate Constant Solution Using Chemical Grade Feedstock over Duration of
Experimental Run at 255C
(not available electronically)

33

which is considered the rate limiting reaction for the methanol steam-reforming process. The
solution for the rate constant was obtained using a plug flow model that was developed for this
research. The rate constant could be solved numerically using fractional conversion data and the
space velocity coupled with chemical equilibrium calculations. In Figure 20, the rate constant is
plotted with respect to run time, which is the total time that the catalyst bed had been operational.
Presumably, any degradation in the catalyst would be seen as a decrease in the rate constant over
the course of operation. For the commercial grade feedstock there does not appear to be any
degradation over the time period shown. This was expected, as catalyst deactivation using
commercial grade feedstock is gradual and is only observed after hundreds of hours of operation.
Figures 21 and 22 show similar data as the previous figures, with the only difference
being the reaction temperature of 275C. Notice that, in general, the space velocity at 99%
conversion level is nearly double that at 255C, which emphasizes the effect of temperature on
reaction rate. In addition, the slope of the linear curve fit is smaller in magnitude at the higher
temperature, indicating a more gradual decrease in conversion level drop off with increasing
space velocity. Figure 22 indicates that the rate constant is higher in the higher temperature case,
as expected. Again, there is no discernible degradation of the catalyst over the time period of
experimentation.
Mineral oil contaminant
Figures 23 and 24 show results obtained from experimentation using methanol
contaminated with small quantities of mineral oil at a reaction temperature of 255C. Data is
presented in similar form to commercial grade data with the following exceptions:

34

Fractional Methanol Conversion, XM

1.00

0.98

0.96

0.94

0.92

0.90
1.0

1.5

2.0

2.5

3.0

-1

Space Velocity (hr )

Figure 21: Fractional Methanol Conversion Using Chemical Grade Feedstock at 275oC

16

max 15.4

Rate Constant, km (s-1)

15
14

mean 13.7
13
12

min 11.4
11
10
0

10

15

20

25

30

35

40

Run Time (hr)

Figure 22: Rate Constant Solution Using Chemical Grade Feedstock over Duration of
Experimental Run at 275oC

35

Fractional Methanol Conversion, XM

1.00
1-4
56

0.95
9 10
8 7

12 11

0.90
0.85

13 14
20
17

0.80

18
19
21

0.75
16
15

22

0.70
0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

Space Velocity (hr-1)

Figure 23: Fractional Methanol Conversion with Mineral Oil Contaminant at 255oC

Rate Constant, km (s-1)

6
3

4 5 6
7

4
9
10

11

12

13 14
15 16

21

22

17 18 19 20

Mineral Oil Processed (mL)

0
0

10

15

20

25

30

35

40

45

50

55

Run Time (hr)

Figure 24: Rate Constant Solution and Total Volume of Mineral Oil Processed over
Duration of Experimental Run at 255oC
36

i) The number associated with each data point matches data between the two figures.
ii) Figure 24 has a secondary y-axis, which shows the total amount (volume) of
contaminant processed by the reactor. The term processed is used loosely here, as it is
probable that the oil passed through the reactor unreacted.
The effect of the mineral oil on the reaction rate is immediate and apparent. The rate
constant appears to be similar to the commercial grade case initially (about 8s-1), but it decreases
quickly over time. By approximately 20 hours of run time (runs 1-8) at a very low oil
concentration (0.05%, or 0.5 grams of oil per 1000 grams of methanol) and nominal space
velocities the rate constant had been reduced by nearly 50%. The data indicates that there is an
additional 50% drop (75% total) in the rate constant by the 40 hour mark (runs 9-16) using
similar flow rates and slightly elevated oil concentrations (0.1% and 0.5%). After completing
experimentation using contaminated feedstock, non-contaminated, commercial grade methanol
was used in an attempt to clean the catalyst (runs 17-22). Over a period of 10-15 hours of run
time, there was no noticeable increase in the reaction rate constant, indicating that any positive
effect this process might have is either not measurable using this rig or is very gradual in
occurrence.
Figures 25 and 26 show results similar to those shown in the previous two figures, only at
a higher reaction temperature. The results are very similar qualitatively, although the values are
different due to the change in experimental conditions. A similar decrease in the reaction rate
constant is seen over a relatively short amount of time. This indicates that over the range of
temperatures examined, the contaminating effect of the mineral oil is relatively independent of
temperature.

37

Fractional Methanol Conversion, XM

1.00
3 1 2

0.95

7
8

0.90

10

0.85
0.80
0.75
0.70
1.9

2.0

2.1

2.2

2.3

2.4

2.5

Space Velocity (hr-1)

Figure 25: Fractional Methanol Conversion with Mineral Oil Contaminant at 275oC

16

Rate Constant, km (s-1)

12

3
4

10
8

2
5
6

6
9

Mineral Oil Processed (mL)

14

10

0
0

10

15

20

25

Run Time (hr)

Figure 26: Rate Constant Solution and Total Volume of Mineral Oil Processed over Duration of
Experimental Run at 275oC

38

Kerosene and isooctane contaminants


The results obtained during kerosene and isooctane experimentation are shown in Figures
27 and 28, and Figures 29 and 30, respectively.
The kerosene data seems to indicate a slight reduction in the reaction rate constant over
time, although it is not as pronounced as the mineral oil case. Note also that the concentration of
kerosene in the methanol was considerably higher (0.1% to 1.0%) than for the mineral oil,
resulting in a much higher total quantity of kerosene processed for the same amount of run
time. As in the mineral oil case, the use of commercial grade methanol does not appear (at least
over the time period indicated) to replenish the activity of the catalyst to its original state.
The isooctane results were inconclusive. Any effect that the isooctane might have on the
reforming process is either too small to measure using the existing rig, or occurs over a more
gradual time scale that the experimentation conducted for this research. This is evidenced by the
data shown in Figure 30, which is very similar to data obtained during commercial grade
experimentation with respect to the values obtained and the scatter observed.

39

Fractional Methanol Conversion, X M

1.00

0.98

17

16

7 8

11
13
15

0.96

10

12
14

0.94

0.92

0.90
0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

Space Velocity (hr-1)

Figure 27: Fractional Methanol Conversion with Kerosene Contaminant at 255oC

12

30

Rate Constant, km (s-1)

24
max 9.03
1

18

4
5

mean 6.88

11

7
8

10
9

12

15

13
12

14

min 4.96

16 17

Kerosene Processed (mL)

10

0
0

10

15

20

25

30

35

40

45

Run Time (hr)

Figure 28: Rate Constant Solution and Total Volume of Kerosene Processed over Duration of
Experimental Run at 255oC

40

Fractional Methanol Conversion, XM

1.00

1 2

5
4

8
9

10

0.98

12 11 7

13 14

0.96

0.94

0.92

0.90
0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

Space Velocity (hr-1)

Figure 29: Fractional Methanol Conversion with Isooctane Contaminant at 255oC

12

Rate Constant, km (s-1)

20

max 9.15
4

13
5

mean 7.18 3
6

10

14

12

10
9

min 5.63

15

11

0
0

10

15

20

25

30

Isooctane Processed (mL)

10

25

35

Run Time (hr)

Figure 30: Rate Constant Solution and Total Volume of Isooctane Processed over Duration of
Experimental Run at 255oC
41

Conclusions

Stabilized methanol produced by Air Products, Inc. was evaluated by the


University of Florida as a candidate fuel for certain fuel cell systems. Analysis of samples
of this methanol showed that
1) the stabilized methanol is over 97% pure methanol, but contains a small
amount of water, small amounts of various hydrocarbons, and negligible quantities of
sulfur and chlorine compounds.
2) one of the hydrocarbons contained in the stabilized methanol is an oil similar to
mineral oil, apparently utilized in the methanol production process. The analyzed samples
indicated that the mineral oil represented from 0.2% to 0.5% of the methanol by
weight.
An experimental rig was successfully designed and fabricated to allow accurate
temperature control as well as determination of energy input distribution for the
endothermic steam reforming of methanol. The rig also allowed for the simultaneous
evaluation of steam reforming of commercial grade and stabilized methanol. Results from
the test rig showed that
3) at a normal steam reforming temperature (~275-285C) for methanol using a
typical low-temperature water-gas shift catalyst, the catalyst quickly became deactivated
on the side processing the stabilized methanol, while the side processing commercial
grade methanol proceeded without apparent deactivation.

42

4) the deactivation of the catalyst was caused by the physical deposition (coating)
of the contaminant mineral oil on the catalyst pellets.
5) the mineral oil coating was apparently due to the incomplete vaporization of
the oil at the reformer temperatures used, thus passing essentially as an unreacted liquid
through the reformer catalyst bed.
A subsequent modification to the test rig involved replacing the CuO/ZnO (low
temperature) catalyst with an Fe-Cr (high-temperature shift) catalyst that could be
operated at higher temperatures (up to 450C). Results from this experimentation showed
that
6) at the higher temperatures, most of the mineral oil passed through the catalyst
bed in vapor form with little evidence of catalyst degradation.
7) even though the reformer catalyst was relatively unaffected by the oil, there
was some evidence of oil in the condensate downstream of the reformer, suggesting that
at least some of the oil was passing through the reformer unreacted.
8) any remaining unreacted oil in the reformate would likely condense out at the
operating temperature of a phosphoric acid fuel cell stack (~200C), and would surely
condense in a PEM system stack (<100C). Such oil condensation in the stack would be
extremely detrimental.
9) at the higher reforming temperature (~400C) using the Fe-Cr catalyst, some of
the advantages of typical methanol steam-reforming are lost resulting in a lower
hydrogen yield as well as a greater concentration of carbon monoxide in the reformate.

43

With respect to the primary issue of the suitability of Air Products stabilized
methanol as a suitable fuel for fuel cell systems, the following conclusions can be
reached:
10) With the blend of specie as received and evaluated by the University of
Florida, stabilized methanol is not an acceptable fuel for typical low temperature steam
reforming systems.
11) Using a higher temperature catalyst and operating at approximately 400C,
the stabilized methanol might be an acceptable fuel. Additional investigation as well as
system modifications (provide a more suitable catalyst and/or additional treatment of the
reformate) to reduce the CO concentration and improve the H2 yield would be required. It
would also be necessary to ensure that no oil reached the fuel cell stack.
12) A higher temperature steam reforming configuration as described above
would also require additional overall system analyses and optimization studies to
determine if the system could be competitive with a low temperature reforming system
on an energy utilization and economic basis.
13) Although no experimentation other than for steam reforming was conducted,
based on those observations it seems likely that the stabilized methanol could be
(possibly) used satisfactorily in an autothermal reformer for phosphoric acid or a PEM
fuel cell system. At the even higher temperatures of partial oxidation (POX) reformers,
no problems due to fuel composition would be expected.
14) There is no reason to believe that there would be fuel composition issues
when using stabilized methanol in high temperature applications such as molten
carbonate or solid oxide fuel cell systems.

44

Recommendations
Methanol has been shown to be an excellent fuel for transportation applications of
phosphoric acid and PEM fuel cell systems. A primary reason for this is that methanol
can be reformed at much lower temperatures than other hydrocarbon fuels, thus providing
a lower CO concentration and a more favorable H2/CH4 balance in reformate. The result
is less aftertreatment (typically none for a phosphoric acid stack), and a more efficient
fuel processing system.
Clearly, as-produced stabilized methanol will require a higher processing
temperature to avoid catalyst contamination. It is therefore important to determine
whether the higher processing temperature will cause stabilized methanol to lose its
advantages over other hydrocarbon fuels. To accomplish this, it is recommended that the
following actions be taken:
1) Perform additional experimentation with steam reformation of the stabilized
methanol with reformer temperature, steam/methanol ratio and catalyst composition as
primary independent variables. The purpose of this research would be to determine the
combination of variables that would yield no oil downstream of the reformer while
providing the best H2 yield and lowest CO concentration in the reformate.
2) Perform systems studies to determine the effect of higher temperatures and
appropriate reformate aftertreatment on the potential advantages of methanol as
compared to other hydrocarbon fuels.

45

3) Air Products should investigate the feasibility (and additional cost) of


removing the mineral oil from the stabilized methanol, at least for the portion that would
be used for PEM and phosphoric acid fuel cell system applications.

46

Appendix A

The following appendix contains copies of the analysis of stabilized methanol


from three different sources.

Pages 48-49

Atlantic Analytical Laboratory, Inc.


Whitehouse, N.J

Pages 50-51

Intertek Testing Services


Seabrook, TX

Page 52

LaPorte, TX Process Demonstration Unit


(supplied by Air Products)

47

(not available electronically)

48

(not available electronically)

49

(not available electronically)

50

(not available electronically)

51

(not available electronically)

52

Appendix B

This appendix contains results taken from the gas chromatograph and are typical
of the following experimental conditions:

Page 54

TCD chromatogram
Low-Temperature Operation using Commercial Grade Methanol

Page 55

FID chromatogram
Low-Temperature Operation using Commercial Grade Methanol

Page 56

TCD chromatogram
High-Temperature Operation using Commercial Grade Methanol

Page 57

FID chromatogram
High-Temperature Operation using Commercial Grade Methanol

53

(not available electronically)

54

(not available electronically)

55

(not available electronically)

56

(not available electronically)

57

Bibliography

Casper, M.S. , Hydrogen Manufacture by Electrolysis, Thermal Decomposition and


Unusual Techniques, Park Ridge, NJ: Noyes Data Corporation, p.1, 1978.
Cox, K.E. and K.D. Williamson, Jr., Hydrogen: Its Technology and Implications Volume
1: Production Technology, Cleveland, OH: CRC Press Inc., 40-48, 1977.
Air Products Liquid Phase Conversion Company, Producing Methanol from Coal GasThe Air Products Clean Coal Project, Allentown, PA: 1996.
Cox, K.E. and K.D. Williamson, Jr., Hydrogen: Its Technology and Implications Volume
1V: Utilization of Hydrogen, Cleveland, OH: CRC Press Inc., p.6, 1977.
Hirschenhofer J.H., D.B. Stauffer and R.R. Engleman, Fuel Cells: A Handbook (Revision
3), Morgantown, WV: Morgantown Energy Technology Center, p. 18, 1994
Fletcher, J.H., An Investigation of Emissions and Efficiency Characteristics of a
Phosphoric Acid Fuel Cell Engine, Gainesville, FL: Ph.D. Dissertation,
University of Florida, p. 19, 1999.
Blomen, L.J. and M.N. Mugerwa, Fuel Cell Systems. First ed., New York: Plenum Press,
p. 614, 1993.
Amphlett, J.C., M.J. Evans, R.A. Jones, R.F. Mann and R.D. Weir, Hydrogen
Production by the Catalytic Steam Reforming of Methanol Part 1: The
Thermodynamics, The Canadian Journal of Chemical Engineering, 59, 720-727,
1981.
Amphlett, J.C., R.D. Klassen, R.F. Mann and B.A. Peppley, Methanol, Diesel Oil and
Ethanol as Liquid Sources of Hydrogen for PEM Fuel Cells, Piscataway, NJ:
Proceedings of the 1994 28th Intersociety Energy Conversion Engineering
Conference, 1221-1226, 1993.
Ogden, J.M., M.M. Steinbugler and T.G. Kreutz, A Comparison of Hydrogen,
Methanol, and Gasoline as Fuels for Fuel Cell Vehicles: Implications for Vehicle
Design and Infrastructrue Development, Journal of Power Sources, 79, 143-168,
1998.
Voecks, G.E., Onboard Production of Hydrogen for Vehicular Systems, Boulder, CO:
Proceedings, Renewable Fuels and Advanced Power Sources for Transportation
Workshop, 139-144, 1982.
58

Meyer, A.P., J.A.S. Bett, G. Vartanian and R.A. Sederquist, Parametric Analysis of 1.5kW Methanol Power Plant Designs, Final Technical Report for U.S. Army
Mobility Equipment Research and Development Command, Contract No.
DAAK70-77-C-0195, 1978.
Kopasz, J.P., D. Applegate, L. Ruscic, S. Ahmed and M. Krumplet, Effects of
Fuels/Contaminants on Reforming Catalyst Performance and Durability,
Richland, WA: The Annual National Laboratory R&D Meeting of the DOE Fuel
Cells for Transportation Program, June 2000.
Borup, R., M. Inbody, J. Hong, B. Morton and J. Tafoya, Fuel Processing: Fuel and Fuel
Constituents Effects on Fuel Processor and Catalyst Durability and Performance,
The Annual National Laboratory R&D Meeting of the DOE Fuel Cells for
Transportation Program, June 2000.
Farrauto, R.J. and C.H. Bartholomew, Fundamentals of Industrial Catalytic Processes,
London, UK: Blackie Academic and Professional, 1997.
Amphlett, J.C., R.A. Jones, R.F. Mann and R.D. Weir, Hydrogen Production by the
Catalytic Steam Reforming of Methanol Part 3: Kinetics of Methanol
Decomposition Using C18HC Catalyst, The Canadian Journal of Chemical
Engineering, 66, 950-956, 1988.
BASF Corporation, BASF Catalyst K 3-110, BASF Technical Leaflet, 1986.
ICI Katalco, ICI Shift Conversion Catalysts, Billingham, UK: 1998.
Myers, D., J. Krebs, T. Krause and M. Krumplet, Alternative Water-Gas Shift
Catalysts, 2000 Annual National Laboratory R&D Meeting of the DOE Fuel
Cells for Transportation Program, June 2000.
Little, A.D., Multi-Fuel Reformers for Fuel Cells Used in Transportation: Phase 1 Final
Report, Cambridge, MA: U.S. Department of Energy: Office of Transportation
Technologies, p.122, 1994.
Weise, W., B Emonts and R. Peters, Methanol Steam Reforming in a Fuel Cell Drive
System, Journal of Power Sources, 84, 187-193, 1999.
GM Global Research and Development, Research and Development of a PEM Fuel Cell
System for Transportation Applications, 1998 Fuel Cells for Transportation
Annual Progress Report, Washington, D.C.: U.S. DOE Office of Advanced
Automotive Technologies, 1998.
Dams, R.A.J., P.R. Hayter and S.C. Moore, The Processing of Alcohols, Hydrocarbons,
and Ethers to Produce Hydrogen for a PEMFC for Transportation Applications,

59

Proceedings of the 1997 32nd Intersociety Energy Conversion Engineering


Conference, 837-842, 1997.

Luengo, C.A., G. Ciampi, M.O. Cencig, C. Steckelberg and M.A. Laborde, International
Journal of Hydrogen Energy, 17, 677-681, 1992.
Espino, R.L. and J.L. Robbins, Fuel and Fuel Reforming Options for Fuel Cell Vehicles,
Annandale, NJ: Exxon Corporate Research Laboratories, 1997.
Amphlett, J.C., R.F. Mann, C. McKnight and R.D. Weir, Production of Hydrogen-Rich
Gas by Steam Reforming of Methanol Over Copper Oxide-Zinc Oxide Catalysts,
Proceedings of the 20th Intersociety Energy Conversion Engineering Conference,
772-780, 1985.
Cheng, Yu-Li, L.D. Chen, and J.P. Seaba, Thermodynamic Analysis of Fuel
Processing, SAE International Congress and Exposition, Detroit, MI: March
1999.
Amphlett, J.C., M.J. Evans, R.F. Mann and R.D. Weir, Hydrogen Production by the
Catalytic Steam Reforming of Methanol Part 2: Kinetics of Methanol
Decomposition Using Girdler C66B Catalyst, The Canadian Journal of
Chemical Engineering, 63, 605-611, 1985.
Dusterwald, H.G., B. Hohlein, H. Kraut, J. Meusinger, R. Peters and U. Stimming,
Methanol Steam-Reforming in a Catalytic Fixed Bed Reactor, Journal of
Chemical Engineering Technology, 20, 617-623, 1997.
Pour, V., J. Barton and A. Benda, Kinetics of Catalysed Reaction of Methanol with
Water Vapour, Collection of Czech Chemical Communications, 40, 2923-2934,
1975.
Nauman, E.B., Chemical Reactor Design, New York: John Wiley & Sons, 1987.
Denbigh, K., Chemical Reactor Theory, Cambridge, UK: Cambridge Univ. Press, 1965.
Rase, H.F., Chemical Reactor Design for Process Plants, New York: John Wiley & Sons,
p.226, 1977.

60

You might also like