Hydrogen from methane and supercritical water

Article abstract of DOI:10.1002/anie.200390240

More rapidly to the target: Screening experiments in a miniautoclave demonstrate the different effects of catalytically active additives such as Raney nickel and basic alkali-metal compounds on the reaction of methane in supercritical water. Accordingly,

Full text of DOI:10.1002/anie.200390240

Angewandte
Chemie
Reactions in Supercritical Water
Hydrogen from Methane and Supercritical
Water**
Andrea Kruse* and Eckhard Dinjus
The production of hydrogen from methane is currently
carried out industrially in a multistep process. The first stage
is reforming [Eq. (1)], during which natural gas and steam are
converted into a H2-rich synthesis gas, which still contains a
high percentage of CO, with the aid of a heterogeneous
catalyst (Ni on an oxidic support) at temperatures of 700–
9008C and pressures of up to 50 MPa. As shown in
Equation (2), the water–gas shift reaction, the CO is sub-
sequently oxidized to CO2 with the concomitant formation of
hydrogen. The summation of both reaction equations gives
Equation (3).
Figure 1. Schematic layout of the apparatus for charging the mini-
autoclave.
Nickel catalysts are used predominantly in biomass gas-
ification if methane is the target product and the conversion is
carried out at temperatures below 4008C. In the backreaction
of Equation (1) the hydrogenation of CO is catalyzed and
thus the otherwise kinetically highly inhibited formation of
methane is possible.
If the temperature is so high that the equilibrium in
Equation (1) lies on the side of hydrogen, nickel catalyzes the
reforming of methane. The addition of alkali metal salts
reduces (experimentally established) coke formation and
catalyzes the water–gas shift reaction[2b,3a,3b] [Eq. (2)].
To determine the optimal reaction temperature thermo-
dynamic calculations were carried out with the computer
program equiTherm.[4] The components CH4, H2O, H2, CO,
CO2, and soot as elemental carbon were considered. Figure 2
CH4 þ H2O Ð CO þ 3 H2
CO þ H2O Ð CO2 þ H2
CH4 þ 2 H2O Ð CO2 þ 4 H2
ð1Þ
ð2Þ
ð3Þ
There are many technical variants of the shift reaction,
which differ in the catalysts and the temperature ranges
employed (Cu/Zn/Al2O3, 160–2508C; Fe/Cr oxide, 350–
4508C). Given the diminishing oil reserves the production
of synthesis gas from methane is of growing interest, since
natural gas and also biogas can be used by way of this
intermediate stage as a carbon source for the production of,
for example, synthetic fuels.
Herein, we present the conversion of methane into
hydrogen under supercritical conditions (T> 3748C, p >
22.1 MPa) and the effect of different catalysts on the above-
mentioned partial steps. A review of the properties of
supercritical water (SCW) and reactions in this medium is
given in reference [1]. We also show that a one-step H2
synthesis with a minor CO content of less than 0.2vol% is
achievable. Hereby, experiments were carried out in a
laboratory autoclave (Figure 1) and catalysts were used which
have led to higher gas yields in the conversion of more
reactive biomass in SCW.[2]
Depending upon the desired gas composition nickel on
oxidic support material,[3b] alkali metal salts (KOH, Na2CO3,
K2CO3),[2] or coke[3b,c] have been used as “catalysts” in the
production of gas from biomass in near- and supercritical
water. Little is known about the mode of action of these
additives or the nature of the catalytically active species.
[*] Dr. A. Kruse, Prof. Dr. E. Dinjus
Institut für Technische Chemie CPV
Forschungszentrum Karlsruhe
Postfach 3640, 76021 Karlsruhe (Germany)
Fax: (+49)7247-82-2244
E-mail: andrea.kruse@itc-cpv.fzk.de
Figure 2. Thermodynamic calculations of the composition of the reac-
tion mixture under the following reaction conditions (details of molar
ratios): a) 60 MPa, H2O:CH4 =143:1, b) 30 MPa, H2O:CH4 =143:1,
c) 60 MPa, H2O:CH4 =14.3:1. (The main component, water, is not
shown.)
[**] We thank Mrs Hops and Mr. Kirschner for their experimental
support.
Angew. Chem. Int. Ed. 2003, 42, No. 8
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 1433-7851/03/4208-0909 $ 20.00+.50/0
909

Communications
shows the calculated mole fractions of CH4, H2, CO, and CO2
“seasoning”, that is extended use of the reactor.[3c] During the
pyrolysis of tert-butylbenzene in SCW Kruse found a reduc-
tion in the concentration of leached out metals and the
formation of a carbon film on the reactor surface with
increasing duration of use of the reactor.[5]
in relation to temperature under different reaction conditions.
Water is present in excess and was not illustrated for reasons
of clarity. The mole fraction of carbon is much smaller than
that of the other components (0.0005–0.008) at a molar ratio
of H2O:CH4 = 143:1; 0.03–0.065 at
a
molar ratio of
In Experiment 2(Figure 3) a KOH solution (1.1 wt%
KOH, 6508C, 64 MPa) was used instead of pure water which
led to a considerable increase in the yield of H2. The CO
content was very low at 0.02vol%. This low CO content with
high H2 yield demonstrates the catalytic activity of KOH in
the shift reaction [Eq. (2)]. A mechanism for this catalysis by
alkali metal carbonates below the critical point of water has
been postulated in the literature[6] and is formulated for KOH
in Equation (4) and (5).
H2O:CH4 = 14.3:1) and was therefore likewise omitted from
the diagram. The calculations show that at a molar ratio of
H2O:CH4 = 143:1 and 60 MPa a temperature of 6508C is
favorable for the reaction since in this case the CH4 used
should be converted mainly into CO2 and H2 (Figure 2a).
When equilibrium is reached the reaction of CH4 with water
should lead to a gas mixture with the following composition:
CO2 20, H2 79, CH4 0.7, and CO 0.3 vol%. A further increase
in temperature should lead to only a relatively small change in
the composition. A lower pressure of 30 MPa has little effect
upon the gas composition (Figure 2b). More methane re-
mains in the reaction mixture at higher methane concentra-
tions (Figure 2c).
Selected results of the investigation of the gas composition
on using different catalysts/additives are shown in Figure 3.
This selection serves to illustrate the principle effects. The
volume percentage data are in each case corrected for the
amount of methane in the dead space of the experimental
apparatus (see Experimental Section). The experimental
error in the determination of the gas composition is assumed
to be Æ 15% relative to the given volume fraction. In these
experiments CO2 was detected in only small amounts in the
gas phase since it was dissolved in the aqueous solution.
In Experiment 1 (Figure 3) methane and water were used
without additives in an “aged” miniautoclave, that is one
already used several times for the conversion of hydrocarbons
in SCW, at 56 MPa and 6508C. The measured H2 yield was
only very small since because of the “aging” of the reactor
wall possible catalytic activity was greatly reduced by, for
example, the carbon deposition at the wall, and the compo-
sition expected from the thermodynamic calculation was not
achieved for kinetic reasons. Indications of changes in
catalytic activity of a reactor wall with extended use of the
reactor are available in the literature: Antal et al. reported a
drop in methane formation in gas production from biomass
because of reduced catalytic activity of the wall through
KOH þ CO Ð HCO2K
ð4Þ
ð5Þ
HCO2K þ H2O Ð KOH þ CO2 þ H2
Formic acid has been analogously postulated as an
intermediate in the shift reaction without salt addition.[7]
Spectroscopic investigations, which confirm the high stability
of formate in SCW, contradict the proposed mechanism for
the addition of alkali metal salts.[8]
A very high yield of H2 is comprehensible only if both the
shift reaction [Eq. (2)] and the reforming step [Eq. (1)] are
catalyzed. It is conceivable that KOH activates the autoclave
wall, and hereby catalyzes the reforming step. The work of
Antal et al.[3c] demonstrates that individual components of the
autoclave material are leached out in supercritical water. This
effect is reinforced by the added base because of the increased
solubility of the oxides. Thus a detachment of the passivating
layer and an activation of the surface by the increased
leaching out of individual components by KOH addition is
likely. Indications of a catalytic effect arising from a reactor
wall consisting of a nickel alloy such as is present here on
different reactions are given in references [3c,7b]. Further-
more the leached-out components could themselves be
catalytically active.
The use of K2CO3 and NaOH (Experiment 3 (Figure 3):
1.34 wt% K2CO3, 6508C, 60 MPa; Experiment 4 (Figure 3):
0.2 m NaOH, 6508C, 33 MPa) led to a lower yield of H2. The
addition of roughened and granulated nickel sheet with a
KOH solution (Experiment 5 (Figure 3): 1.1 wt% KOH,
6608C, 60 MPa) gave a similar gas composition.
The use of a suspension of Raney nickel (Experiment 6
(Figure 3): 6508C, 53 MPa and Experiment 7 (Figure 3):
6508C and 38 MPa) led to a totally different gas composition
(CO:H2 ꢀ 1:3) and a significantly lower CH4 conversion. As
expected the reforming reaction [Eq. (1)] was catalyzed by
the addition of Raney nickel, but not the shift reaction
[Eq. (2)]. It should be noted that nickel is converted into
nickel oxide in supercritical water.[9] The catalytically active
species on addition of nickel[2a,b], both here and in biomass
gasification, is not known. The addition of KOH with
concomitant use of Raney nickel (Experiment 8 (Figure 3):
6508C, 33 MPa) led to a hydrogen yield similar to that
obtained in the absence of Raney nickel (see Experiment 2),
but with increased CO formation. A reduction in temperature
from 650 to 570¤C (Experiment 9 (Figure 3): 1.1 wt% KOH,
Figure 3. Selected experimentally determined gas compositions after reac-
tion in the miniautoclave (reaction time 15 min, 6508C).
910
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1433-7851/03/4208-0910 $ 20.00+.50/0
Angew. Chem. Int. Ed. 2003, 42, No. 8

Angewandte
Chemie
1548; c) M. J. Antal, Jr., S. G. Allen, D. Schulman, X. Xu, R. J.
570¤C, 68 MPa) caused a more drastic reduction in H2 yield
than was expected on the basis of the thermodynamic
calculations.
Divilio, Ind. Eng. Chem. Res. 2000, 39, 4040 – 4053; d) X. Xu, Y.
Matsumura, J. Stenberg, M. J. Antal, Jr., Ind. Eng. Chem. Res.
1996, 35, 2522 – 2530.
The presence of alkali metal salts leads to an increased
CH4 conversion and, in agreement with the thermodynamic
calculations, to an almost exclusive formation of H2 and CO2.
With regard to an industrial process this means that only one
process step should be needed for H2 formation from CH4.
Moreover, the solubility of CO2 in water under pressure is
significantly higher than that of H2, which means that the two
components can be readily separated. The use of the nickel
catalyst alone leads only to reforming, that is to the formation
of a CO/H2 mixture which can be used for synthesis reactions.
This means that under otherwise identical reaction conditions
the gas composition can be varied by the choice of catalyst.
Regarding, the mode of action of the catalysts, the nature
of the catalytically active species and the effect of the reactor
wall, many questions remain open which cannot be answered
with the experimental set up used here. The results presented
here can be used for the development of a new process for
hydrogen production. Detailed investigations in a continuous
reactor are currently being undertaken by us to test shorter
heating and cooling procedures and to guarantee a better
mixing of the reaction mixture.
[4] equiTherm Windows, Version 5.Y, Scienceware/VCH, 1997.
[5] A. Kruse, Kernforschungszent. Karlsruhe 1994, 5399, 32 .
[6] D. C. Elliott, R. T. Haller, L. J. Sealock, Jr ., Ind. Eng. Chem.
Prod. Res. Dev. 1983, 22, 426 – 431.
[7] a) C. F. Melius, N. E. Bergan, J. E. Shepherd, Proc. 23th Symp.
Combust. 1990, 217 – 223; b) J. Yu, P. E. Savage, Ind. Eng. Chem.
Res. 1998, 37, 2– 10.
[8] P. G. Maiella, T. B. Brill, J. Phys. Chem. A 1998, 102, 5886 – 5891.
[9] C. Kaul, H. Vogel, H. E. Exner, Materialwiss. Werkstofftech. 1999,
30, 326 – 331.
Clusters with Ge0-Atoms
[Ge8{N(SiMe3)2}6]: A Ligand-Stabilized Ge
Cluster Compound with Formally Zero-Valent Ge
Atoms**
Andreas Schnepf* and Ralf Köppe
Experimental Section
Polyhedral germanium compounds can be divided into two
broad categories: Zintl anions, which can be made soluble
through the use of chelating complexing agents such as
[2.2.2]cryptands,[1] and ligand-stabilized complexes with the
general formula GenRn (n = 4,[2] 6,[3] 8[4]), which are generated
by the use of sterically demanding ligands in the reductive
coupling of the appropriate halogen-containing precursors
with alkali or alkaline earth metals.[5] As a result, only
germanium clusters in which the average oxidation state of
the Ge atoms is < 0 (Zintl anions) or ꢁ + 1 (e.g. [Ge8tBu8X2];
X = Cl,[6] Br[7]) exist to date. Germanium cluster compounds
with an average Ge oxidation state between 0 and + 1 are
thus far unknown (in contrast to corresponding tin com-
pounds).[8]
Here we describe an approach to such compounds by
the disproportionation of subvalent germanium halides
(4/n(GeXn)!(4/nꢂ1)Ge + GeX4; n = 1, 2), during which
germanium-rich intermediates, which are passed through on
the way to elemental germanium, can be trapped by, for
example, kinetic stabilization. The same concept was success-
fully used by Schnöckel et al. for the homologous aluminum
and gallium compounds.[9] Germanium(ii) halides are not
suitable as starting materials because kinetic stabilization
Figure 1 illustrates the apparatus for charging the miniautoclave
constructed of Hastelloy, a nickel alloy (ca. 4 mL internal volume).
The autoclave was charged with the previously calculated amount of
water, an aqueous solution (KOH, K2CO3, NaOH, each pa. Merck) or
an aqueous Raney nickel suspension (Merck) under a constant flow
of CH4. The desired CH4 pressure was set with the aid of manometer
P2. The reactor was then released from the charging stand and heated
in the oven.
Owing to the design of the autoclave a dead space which is filled
with CH4 and in which no reaction occurs is located between V1, the
manometer, the bursting disk, and the heatable reaction space. The
dead volume was determined by prior volume measurement and the
amount of CH4 measured after the reaction was corrected by the
fraction contained within the dead space.
After the desired reaction time the autoclave was removed,
attached again to the apparatus and a gas sampling tube was filled for
analysis by gas chromatography. The resulting gas volume was
calculated from the pressure indicated by manometer P2.
Received: July 12, 2002
Revised: October 11, 2002 [Z19724]
[1] a) D. Bröll, C. Kaul, A. Krämer, P. Krammer, T. Richter, M. Jung,
H. Vogel, P. Zehner, Angew. Chem. 1999, 111, 3180 – 3196;
Angew. Chem. Int. Ed. 1999, 38, 2998 – 3014; b) E. Dinjus, A.
Kruse in High Pressure Chemistry; Synthetic, Mechanistic, and
Supercritical Applications (Eds.: R. van Eldik, F.-G. Klärner),
Wiley-VCH, Weinheim, 2002, S. 422 – 446.
[2] a) H. Schmieder, J. Abeln, N. Boukis, E. Dinjus, A. Kruse, M.
Kluth, G. Petrich, E. Sadri, M. Schacht, J. Supercrit. Fluids 2000,
17, 145 – 153; b) A. Kruse, D. Meier, P. Rimbrecht, M. Schacht,
Ind. Eng. Chem. Res. 2000, 39, 4842– 4848.
[*] Dr. A. Schnepf, Dr. R. Köppe
Institut für Anorganische Chemie
Universität Karlsruhe (TH)
Engesserstrasse, Geb. 30.45, 76128 Karlsruhe (Germany)
Fax: (+49)721-608-4854
E-mail: schnepf@aoc2.uni-karlsruhe.de
[**] We thank the DFG for financial support of this work through the
“Semiconductor and Metal Clusters as Building Blocks for Organ-
ized Structures” program. We also thank Prof. H. Schnöckel for
helpful discussions.
[3] a) D. C. Elliott, L. J. Sealock, Jr., E. G. Baker, R. S. Butner, Ind.
Eng. Chem. Res. 1993, 32, 1535 – 1541; b) D. C. Elliott, L. J.
Sealock, Jr., E. G. Baker, Ind. Eng. Chem. Res. 1993, 32, 1542–
Angew. Chem. Int. Ed. 2003, 42, No. 8
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 1433-7851/03/4208-0911 $ 20.00+.50/0
911