Porous media combustion based hydrogen production
Zhdanok S.A.
Heat and Mass transfer Institute, National Academy of Sciences of the Republic of Belarus
15 P.Brovki str., Minsk, 220072, Belarus
Introduction
The necessity in compact and efficient hydrogen source for fuel-cell-powered vehicles stimulate
studies in reforming of gaseous and liquid hydrocarbon fuels such as methane, propane, gasoline, kerosene
or diesel fuel. One of usually used processes for hydrogen generation is partial oxidation of hydrocarbon. For
the sake of simplicity summary reaction of partial oxidation can be presented using conventional formulae of
liquid hydrocarbon [1] as follows:
2C m H n + mO2 → 2mCO + nH 2
(1)
Regardless of the fact that process (1) is exothermic, it’s implementation under usual conditions is
connected with some problems. One of them is extremely low rate of elementary chemical reactions that
additively result in reaction (1) at temperature close to adiabatic one. This could be illustrated by Fig.1 where
the methane cracking time is presented as the function of the process (1) temperature. In the case of reach
mixture of aviation kerosene with air corresponding to equivalence ratio γ≈3 adiabatic temperature is about
8000C and hydrocarbons to hydrogen conversion process time could be of order of minutes. Typical values
of adiabatic temperature for different hydrocarbons-air mixtures corresponding reaction (1) are presented at
Fig.2. Several approaches to accelerate chemical reactions leading to partial oxidation of hydrocarbon fuels
are known including plasma or catalytic processes [3-4].
1600
Methane
10
2
t, s
Tad, oC
100
t
x
1200
5-11
1100
1000
10-2
400
b
t - toluene
1400 x - ethylbenzene
1300 k - kerosene
3
1
10-4
i=1-11 - CiH2i+2
1500 b - benzene
4
102
mixture: hydrocarbon - air; P = 1 atm
k
900
1 - C2H4=0.001%
2 - 0.01%
3 - 0.1%
500
600
3
800
700
700
800
900
1000
T, oC
Fig. 1. Temperature dependence of methane cracking
induction time estimated by ethylene formation: p =
1atm.
600
0
2
4
1
100
200
300
To, oC
400
500
Fig. 2. Adiabatic temperature of hydrocarbon-air
mixture in partial oxidation process as a function of
initial temperature: P = 1 atm
The present study proposes to use "super-adiabatic effect" of filtration combustion wave propagating
in porous media to perform the reaction (1) at temperatures considerably exceeding adiabatic one. This effect
was studied in many papers and was used for the lean mixtures combustion [5-6]. The goal of present study
was to explore the extent of conversion of rich hydrocarbons-air mixtures to syngas (H2+CO) by partial
oxidation in filtration combustion process in inert porous medium. Recently some work in this direction was
done in group of Professor L. Kennedy at UIC who studied methane-air and hydrogen sulphide partial
oxidation under filtration combustion conditions [7].
1
Partial methane oxidation
The methane to hydrogen conversion process was studied with the use of the reactor that consists of
cylindrical shell filed with a randomly packed bed of inert ceramic particles (Al2O3 or other material) in
which filtration combustion wave propagates, feed and mixture ignition unit, and exhaust gas ventilation
duct to withdraw the reaction products out of the room (Fig.3,4).
Quartz
thermocouple
shell
Reactant gas
mixture
1
2
Ignition
electrodes
3
Permeable
ceramic disc
Quartz shell
4
Packed
bed
Kaowool
insulation
5
6
7
Reaction
products
Fig. 3. Schematic presentation of the single wave
filtration combustion reactor
Fig. 4 Filtration combustion reactor:
1 - air-hydrocarbons swirl jet; 2 – mixing chamber;
3 – spark ignition plug; 4 –quartz packed bed shell; 5
–electrical air heater; 6 cooler; 7 - liquid product
condenser (thermal insulation and thermocouples are
uninstalled)
In most experiments a 660mm-length quartz tube with an internal diameter of 41 mm was used as a
packed bed shell. To compensate for the different thermal expansion rates of the packed bed and the tube and
to prevent the chemical reaction between them the inside diameter of the tube was covered with a thin (1.5
mm) layer of asbestos and Kaowool insulation. Heat losses due to the conduction through the tube wall were
minimized with a 60mm layer of Kaowool insulation applied to the outside diameter of the tube. Some
unsuccessful attempts to use ceramic (Al2O3) shells were made. The upper part of the quartz tube is
connected with stainless steel ignition chamber by asbestos seal. Premixed reactant mixture is uniformly
distributed across the chamber section using permeable ceramic disc. Two nichrome wire quartz insulated
electrodes are installed to ignite the mixture by spark discharge. High-voltage pulse supply is used to power
the spark discharge.
To start the combustion wave preheating of the porous medium is required. This is done by ignition
of the mixture with an equivalence ratio close to unity at low flow rate (1.8 m3/hour) during some time (500600 sec). Once combustion becomes stable the spark is removed and equivalence ratio is gradually reduced
to desired level ( γ =4). After that self-sustained filtration combustion process begins.
2
Reactant gas mixture is produced by insertion of air and methane to a long feeding pipe. The air is taken
from the high-pressure line, while the methane is taken from the standard high-pressure tank through the
pressure regulators. Gas flow rates are controlled by needle valves and measured by standard flow meters.
The axial temperature distribution in the reactor is measured by 3 S-type (Pt-Pt-10%Rh)
thermocouples with a diameter 0.5mm. These thermocouples are hosed in a 0.5m long quartz shell 8mm in
diameter. To prevent chemical reactions between quartz shell and packed bed material a thin layer of
asbestos paper insulation was used. This layer provides a satisfactory protection of quartz shell for all bed
materials except ZrO2. Intensive chemical interaction of ZrO2 with SiO2 at high temperature usually destroys
the shell during 2-3 hours. A PC based data acquisition system was employed to read and record the
temperatures at regular intervals. The software for Windows 95 developed using Delphi 3.0 controls data
processing and storage, sampling frequency and real time graphic presentation of reactor temperature field.
Experiments on methane to hydrogen conversion in filtration combustion wave have been
performed for a constant composition reacting gas mixture corresponding to equivalence ratio F=4 and
various packed bed materials. Main properties and general appearance of these materials are given in Table
1. and in Fig.5.
Table 1. Packed bed materials properties
Material
Shape
Al2O3
Al2O3
Al2O3
ZrO2
Quartz
cylinders
spheres
spheres
grains
chips
Dimensions
Mm
D=5; l=10
6
3
2-3
∼3×6×15
Density
ρ, kg/m3
3150
2810
2720
6390
2170
Heat capacity
Cp, J/kgK
794
794
794
455
729
Porosity
0.49
0.66
0.67
0.68
0.46
Fig.5. Packed bed materials: 1 - Al2O3 cylinders; 2 - Al2O3 spheres (6 mm);
3 - Al2O3 spheres (3mm); 4 - ZrO2 grains (2-3 mm); 5 – SiO2 chips (∼3×6×15 mm
For each type of a porous material the gas flow rates were in a range of about 1.3 to 5.8 m3/hr. The
maximal value was limited by the reactor length and output heat power. Typical porous media temperature
variation in a propagating combustion wave is shown in Fig.6.
Three thermocouples recording porous body temperature variation were located at distances of 80,
200 and 310 mm from the packed bed upper surface. It is assumed that when the combustion wave front
passes through the third thermocouple position the temperature it records is close to the fully developed
(equilibrium) value for a given flow rate. This temperature was considered as a maximal one for the
combustion regime studied.
3
1500
1400
1300
1200
1100
1000
900
700
0
Temperature, C
800
600
500
400
300
200
100
0
20 0
3 00
40 0
500
6 00
70 0
8 00
90 0
T im e , s
Fig.6. Typical porous medium temperature variation in propagating combustion wave (Al2O3 cylinders; Q =
3.32 m3/hr; G = 0.78 kg/m2s)
The summary of experimental data on maximal wave temperature and reaction product composition
is given in Table 2. The most important feature of the process studied is a combustion temperature
dependence on flow rate presented in Fig.7. Experimental data show only a limited temperature growth
within certain range of gas flow rates. Constant temperature limit is attained at specific flow rates about 0.81.2 kg/m2s. This limit insignificantly differs for various porous bed materials in the range 1380-1430 0C.
Another striking feature of present results is the opposite trends of predicted and experimental data
concerning the effect of packed bed pore dimensions on the maximal combustion wave temperature.
Table 2. Product composition and combustion wave maximal
temperature for various porous media materials
Flow
rate
m3/h
Al2O3 spheres (6 mm)
G,
T max, 0C
Exit Dry Gas Composition
Converted %
2
kg/м s
H2 % N2 % CH4 % CO %
H2
CH4
1.82
1.82
2.61
2.61
0.42
0.42
0.61
0.61
1371
1394
1379
1385
24.2
24.5
24.1
23.4
54.1
55.3
54.5
54.8
13.8
11.6
11.9
11.3
7.9
8.6
9.5
10.5
42
41
41
40
53
61
59
62
4.55
1.06
1411
24.8
51.9
10.0
13.3
44
64
4.55
1.06
1406
24.7
53.9
9.7
11.7
43
67
4.55
4.55
5.71
1.06
1.06
1.33
1421
1408
1419
26.0
23.7
25.3
52.9
53.6
52.6
10.2
12.1
9.3
10.9
10.6
12.8
46
41
45
64
58
67
5.71
1.33
1428
26.0
53.0
7.2
13.8
46
75
4
Al2O3 spheres (3 mm)
Exit Dry Gas Composition
H2 % N2 % CH4
CO
%
%
Converted %
CH4
H2
Flow
rate
m3/h
G,
kg/м2s
Tmax,
0
C
1.82
2.61
2.61
2.61
0.42
0.61
0.61
0.61
1370
1373
1390
1390
17.2
18.3
20.9
22.1
56.6
57.7
53.4
53.0
17.9
14.0
13.4
13.2
8.3
10.0
12.3
11.7
28
29
36
39
41
42
53
54
3.36
4.55
5.71
0.78
1.06
1.33
1398
1407
1417
16.1
20.6
19.9
57.8
53.5
56.0
15.0
15.3
11.6
11.1
10.6
12.5
26
36
33
52
47
62
Al2O3 cylinders
Flow
rate
m3/h
G,
kg/м2s
Tmax,
0
C
Exit Dry Gas Composition
Converted %
CH4
H2 % N2 % CH4 % CO % H2
1.29
1.39
1.82
1.96
2.61
2.94
0.30
0.32
0.42
0.45
0.61
0.68
1338
1340
1370
1365
1405
1414
26.4
27.1
25.9
27.2
26.2
21.6
49.0
48.5
51.1
47.6
52.4
55.0
9.0
10.7
7.3
9.6
9.3
10.3
15.6
13.7
15.7
15.6
12.1
13.1
50
52
47
53
46
36
66
59
73
63
67
65
3.36
4.55
4.55
0.78
1.06
1.06
1415
1433
1407
22.5
23.9
23.7
53.2
54.3
53.6
10.4
8.1
8.9
13.9
13.7
13.8
39
41
41
64
72
69
SiO2 chips
Flow
rate
m3/h
G,
kg/м2s
Tmax,
0
C
0.967
1.934
2.9
2.9
0.24
0.48
0.72
0.72
1235
1320
1380
1396
19.6
25.0
29.5
28.3
61.7
58.4
57.0
56.5
11.7
8.5
4.9
5.2
7.0
8.1
8.6
10.0
29
40
48
47
65
73
84
83
3.869
0.96
1380
27.1
59.4
4.6
8.9
42
86
Exit Dry Gas Composition
Converted %
H2 % N2 % CH4 % CO % H2
CH4
5
ZrO2 grains (2-3 mm)
Flow
rate
m3/h
G,
Tmax, 0C
2
kg/м s
Exit Dry Gas Composition
H2 %
Converted %
H2
CH4
1.29
1.82
2.61
2.61
3.36
0.30
0.42
0.61
0.61
0.78
1234
1323
1376
1340
1338
18.7
19.6
20.2
21.0
19.4
N2 % CH4 % CO %
54.1
54.4
56.5
51.4
54.1
15.9
13.8
10.2
13.3
14.5
11.3
12.2
13.1
14.3
12.0
32
33
33
38
33
45
53
66
52
50
3.36
4.55
0.78
1.06
1353
1385
21.6
19.4
53.8
52.5
14.2
14.6
10.4
13.5
37
34
51
48
1450
1350
0
Maximal temperature, C
1400
1300
1250
1200
0.0
0.5
1.0
1.5
2.0
2
Flow rate, kg/m s
Fig.7. Variation in combustion wave maximal temperature with flow rate.
Ο - Al2O3 spheres (6 mm), ∆ - Al2O3 spheres (3 mm), - Al2O3 cylinders,
▲ - SiO2 chips, ● - ZrO2 grains (2-3 mm).
Along with Table 2 conversion product composition data are also presented in Fig.8-9 as functions
of gas flow rate and reactor maximal temperature. In analyzing the composition data one should take into
account both temperature and residence time as main factors effecting the process studied. While maximal
reactor temperature slightly increases or remains constant residence time decreases in inverse proportion to
the gas flow rate. For this reason output hydrogen content was almost independent on flow rate for all types
of porous bed with exception of Al2O3 cylinders, as can be seen from Fig.9. In the later case even a slight
decrease of hydrogen concentration (within 2-3%) was observed. Maximal hydrogen output was about 30%
for reactor filled with quartz chips and 26% for Al2O3 cylinders and large spheres. It should be noted that in
the case of quartz chips soot formation was clearly observed. It means that methane conversion in this case
goes both through its partially oxidation and through thermal methane cracking. The later reaction is
responsible for extra hydrogen and soot formation. The presence of methane pyrolisis mechanism manifests
itself in higher hydrogen to methane ratio in reaction products as illustrated by Fig.10. For partial oxidation
(2.1) this ratio is 2.
6
25
25
Product concentration, %
Product concentration, %
30
20
15
10
5
20
15
10
5
Al2O3 spheres (6 mm)
Al2O3 spheres (3 mm)
0
0
0.6
0.8
1.0
1.2
0.4
1.4
30
30
25
25
Product concentration, %
Product concentration, %
0.4
20
15
10
5
Al2O3 cylinders
0
0.6
0.8
1.0
1.2
1.4
20
15
10
5
SiO2 chips
0
0.2
0.4
0.6
0.8
1.0
1.2
2
0.2
0.4
0.6
0.8
1.0
2
Flow rate, kg/m s
Flow rate, kg/m s
2
2
Product concentration, %
25
20
15
10
ZrO2 grains (2-3 mm)
5
0
0.2
0.4
0.6
0.8
1.0
1.2
2
Flow rate, kg/m s
Fig.8. Variation in reaction product concentration with flow rate.
∆ - H2, ▲- CO, - CH4.
7
25
Product concentration, %
Product concentration, %
30
25
20
15
10
5
15
10
5
Al2O3 spheres (6 mm)
0
Al2O3 spheres (3 mm)
0
1360
1380
1400
1420
1440
1360
30
1380
1400
1420
30
Product concentration, %
Product concentration, %
20
25
20
15
10
5
25
20
15
10
5
Al2O3 cylinders
0
SiO2 chips
0
1300
1350
1400
1450
Maximal temperature, 0C
1200
1250
1300
1350
1400
Maximal temperature, 0C
Product concentration, %
25
20
15
10
ZrO2 grains (2-3 mm)
5
0
1200
1240
1280
1320
1360
1400
Maximal temperature, 0C
Fig.9. Variation in reaction product concentration with maximal combustion wave temperature.
∆ - H2, ▲- CO, - CH4.
8
H2 / CO
3.5
3.0
2.5
2.0
1.5
1.0
0.0
0.4
0.8
1.2
1.6
Flow rate, kg/m2s
Fig.10. Variation in H2/CO with flow rate.
Ο - Al2O3 spheres (6 mm), ∆ - Al2O3 spheres (3 mm),
▲ - SiO2 chips, ● - ZrO2 grains (2-3 mm).
- Al2O3 cylinders,
Methane conversion ratio, %
100
90
80
70
60
50
40
30
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Flow rate, kg/m2s
Fig.11. Variation in methane conversion rate with flow rate.
Ο - Al2O3 spheres (6 mm), ∆ - Al2O3 spheres (3 mm), - Al2O3 cylinders,
▲ - SiO2 chips, ● - ZrO2 grains (2-3 mm).
9
Maximal hydrogen concentrations in our experiments were about 60% of ideal value corresponding to the
reaction of methane partially oxidation with hydrogen concentration that is 41%:
2CH 4+O2 + 3.75N 2 = 2CO + 4H 2 + 3.75N 2 ,
(2)
The efficiency of conversion process (conversion rate) can be estimated by methane and hydrogen
conversion ratios. These parameters given in Table. 2 and in Fig.11 are determined as follows:
conversionH 2 =
C Hout2 C Nin2
conversionCH 4 = 1 −
in
2CCH
C Nout2
4
out
CCH
C Nin2
4
(3)
in
CCH
C Nout2
4
where C are volume component concentrations.
While methane conversion ratio for some combustion regimes reached approximately 70% (and even
85% for quartz filled reactor) these values are obviously unacceptable from the point of view of the process
practical realization. The main limitation on methane conversion ratio is imposed by reaction temperature,
which is not sufficiently high for full conversion.
Partial kerosene oxidation
The filtration combustion reactor for kerosene partial oxidation was the same as for the methane partial
oxidation (Fig.3,4). The reactor tube in this case was filled with a randomly packed bed of Al2O3 ceramic
spheres of 5-6 mm in diameter. The fuel-air mixture was produced in mixing chamber by injecting air and
kerosene through a two component swirl jet. In order to facilitate the liquid fuel evaporation the air and fuel
are preheated by electrical heaters. The fuel mixture temperature at the exit end of the mixing chamber was
in the range 220-2400C. Such temperature level is enough for complete evaporation of heaviest liquid fuel
(kerosene) fractions but prevents undesirable ignition of fuel mixture in free volume of mixing chamber.
All the experiments were performed for several values of equivalence ratio of air-fuel mixture γ in
the range 2.3-4. Theoretically optimal value γ for the reaction of partial oxidation of kerosene (1) is 2.99.
This value can be estimated using conventional formulae of aviation kerosene (m=10.33; n=20.54) and
equation (1).
Self-sustained filtration combustion was observed for all test runs.(Fig.12) Steady state maximal
combustion wave temperature (Fig.13) is strongly affected by fuel-to-air ratio. This temperature
monotonically increases from 10700C at γ=2.3 to 14050C at γ=4. This is mainly due to the growth of fuel
adiabatic temperature. The value of maximal wave temperature sets the reactor operation limit in the range of
lean fuel mixtures.
1100
5
6
3
1000
Temperature, 0C
2
7
4
900
800
1400
1350
Tmax, 0C
1
1450
1-600s
2-625s
3-700s
4-800s
5-900s
6-1000s
7-1100s
8-1200s
9-1300s
10-1400s
1200
700
1300
1250
600
1200
500
8
400
1150
9
300
10
200
1100
1050
100
0
100
200
300
400
500
600
700
800
900
Axial length, mm
2.0
2.4
2.8
3.2
3.6
4.0
Equivalence ratio
Gk = 0.769 kg/hr; Ga = 3.75 kg/hr; γ = 3.02
Fig.12. Temperature profiles evolution
for different measurements moments.
Fig.13. Steady state maximal partial kerosene
oxidation process wave temperature as the function
of equivalence ratio.
Fig.14 illustrates the effect equivalence ratio on combustion products for kerosene. The main components
of gas reaction products are N2, H2, CO. The content such of hydrocarbons as C2H2, C2H4, etc. never
10
exceeded 1%. Maximal concentrations of H2 and CO are observed in the range of equivalence ratios close to
theoretically optimal value γ=3. It should be noted that concentration of CH4 in conversion products
gradually increases with γ. An abruptly drop of H2 and CO concentrations for γ<2.6 is observed. As oxygen
concentration in fuel mixture increases the percentage of fully oxidised kerosene converted to H2O and CO2
increases. The experimental conversion product composition data qualitatively agree with the results of
thermodynamic calculations of equilibrium content. A mixture of 16.7% С6Н5С2Н5 (ethyl benzene) and
83.3% С11Н24 (undecane) was used as the model kerosene composition for such calculations. The efficiency
of kerosene conversion was estimated by conversion ratio defined as the mass ratio of H2 produced by the
process to hydrogen contained in fuel (Fig.15). This parameter was calculated using measured component
concentrations, input kerosene-to-air mass ratio and mass content of hydrogen in kerosene used (n=20.54).
Maximal values of thereby calculated conversion efficiency (up to about 93%) were observed for
equivalence ratios in the range 2.8-3.
0.25
Mole fraction
0.20
0.15
*- H2
. - CO
+ - CO2
2 - CH4
0.10
0.05
Kerosene-to-hydrogen conversion ratio
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.00
2.0
2.4
2.8
3.2
3.6
Equivalence ratio
4.0
Fig.14. The effect of input mixture composition on
H2, CO, CH4 and CO2 contents in reaction products.
2.0
2.5
3.0
3.5
4.0
Equivalence ratio
Fig.15. Kerosene to hydrogen conversion
efficiency as the function of equivalence
ratio.
It should be emphasised that process studied is energy efficient. The electric power needed for normal
operation of conversion system at nominal conditions (4.6 kg/h) is approximately 0.95 kW, while heat
release of partial oxidation reaction is 1.26 kW. Therefore, basically, one can use some heat of reaction
products to heat the components of the working mixture in recuperative heat exchangers and reduce thereby
external energy consumption greatly.
Conclusions
The experiments performed show, that there is a limit of maximal temperature growth in filtration
combustion waves for methane-air mixture with equivalence ratio γ = 4. Decreasing of particle sizes in a
filling and increasing of mixture specific mass flow rate have not allowed to reach temperatures above
1430оС.
Present study demonstrates the feasibility of partial oxidation of hydrocarbon fuels in inert medium
filtration combustion wave. The efficiency of fuel to hydrogen conversion can be rather high at some test
conditions. The most efficient conversion process for kerosene seems to be for γ ≈3. One of an attractive
features of the process is the absence of solid soot-like deposits (the intensive carbon deposition was
observed only for a quartz filling particles and never observed for Al2O3 and ZrO2 ).
The products composition basically corresponds to the temperature level and length of the hot zone.
The maximal yield hydrogen in methane partial oxidation process was observed at the level 25 % in the dried
mixture. In the best case the non reacted methane was about 6 % (in the dried mixture ) and about 10 % in
average.
11
References
1. Ragosin N.A. Reactive fuels. Moscow, 1959 (In Russian).
2. Bromberg L. Rabinovich A. Alexeev N. Cohn D.R. Plasma Reforming of Diesel Fuel. PSFC /JA –944.p.1-11.
3. Buyakov I.F., Borodin V.I., Chernuho A.P., Solntsev A.P., Kuzmitskiy A.Yu., Zh danok S.A.,
Zarutskaya N.A. A Study of Process of (CH4-H2)-mixture conversion in High-voltage discharge plasma
at atmospheric pressure // Heat/Mass transfer MIF 2000, IV Minsk International Forum, May 22-26,
2000. Vol.4, p.131-137.
4. Pereira C., Wilkenhoener R., Ahmed S., and Krumpelt M. Liquid Fuel Reformer Development,
Proceedings of the 1999 US Department of Energy Hydrogen Program Review, July 2, 1999. p.1-9.
5. K.Hanamura,R.Echigo, S.Zhdanok, Superadiabatic Combustion in Porous Media. Int.J.Heat and Mass
transfer, Vol.36,13:3201-3209, 1993.
6. S.A.Zhdanok,L.A.Kennedy, G.Koester, Superadiabatic Combustion of Methane Air Mixtures under
Filtration in a Packed Bed. Combustion and Flame 100:221-231, 1995.
7. Jacques P. Bingue, Alexei V. Saveliev, Alexander A. Fridman, Lawrence A. Kennedy , Hydrogen
production in ultra-rich filtration combustion of methane and hydrogen sulphide, Int. Journal of ,
Hydrogen Energy, 27,643-649,2002.
12