15214125, 2000, 11, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/1521-4125(200011)23:11<956::AID-CEAT956>3.0.CO;2-L by Universidad Autonoma De Madrid, Wiley Online Library on [23/02/2023].
See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Communications
Chemisorption of N2O and H2 for the
Surface Determination of Copper
Catalysts
By Olaf Hinrichsen, Thomas Genger, and Martin Muhler*
Dedicated to P rofessor Dr. Manfred Baerns on the occasion
of his 65th birthday
1 Introduction
Metal surface areas are usually determined by selective
chemisorption of a suitable probe molecule (e.g. CO, H2, O2). A
lot of experimental techniques for the determination of Cu
metal areas cited in the literature are based on the dissociative
chemisorption of nitrous oxide [1±7]. In this contribution, a
comparisonofdifferentmethodsusing N2Oasaprobemolecule
is presented. The experiments comprise isothermal N2O flow
experiments, transient decomposition experiments and the
pulse flow chemisorption of N2O in an He carrier stream. The
main object of these experiments was to clarify the influence of
the reaction parameters (temperature, contact time, sample
mass, etc.) on the specific Cu surface area determined. As a
further independent method for the determination of Cu
surface areas the temperature-programmed desorption of
hydrogen (H2 TPD) has been applied [8,9].
2 Experimental
2.1 Experimental Setup
The measurements have been conducted in an all-stainless
steel microreactor setup equipped with four gas lines: He
(99.9999 %), H2 (99.9999 %), a mixture of 1 % N2O in He
(99.9995 %) and a mixture of 2.1 % H2 in He (99.9995 %) for
the reduction of the calcined catalysts. The setup allows the
completion of temperature- and concentration-programmed
experiments fully automatically and can be operated at
pressures up to 60 bar [9]. The reactor consists of a glass-
lined U-tube with an inner diameter of 4 mm. The experiments
have been conducted with 200 mg of a 250±355 lm sieve
fraction fixed in the reactor by means of quartz wool. Gas
analysis was performed by a calibrated quadrupole mass
spectrometer (Balzers GAM 400). The Cn/ZnO/Al2O3
catalysts (TG01 and TG02, approx. 50 wt.-%) used in this
study have been prepared by the conventional coprecipitation
method. Prior to each experiment the catalyst has been
flushed at 493 K for 1 h in He in order to achieve an adsorbate-
free reduced catalyst surface1).
± changed from H2 to He in order to flush out the residual H2.
[*] Dr.-Ing. O. Hinrichsen, Dr. T. Genger, Prof. Dr. M. Muhler, Ruhr-
University Bochum, Laboratory of Industrial Chemistry, D-44780
Bochum, Germany; email: olaf@techem.ruhr-uni-bochum.de
1) List of symbols at the end of the paper.
956 Ó WILEY-VCH Verlag GmbH, D-69469 Weinheim, 2000
2.2 Isothermal N2O Flow Experiment, Reactive Frontal
Chromatography (N2O RFC)
For the isothermal N2O flow experiment conducted at
room temperature under atmospheric pressure, the samples
have been cooled down to room temperature in an He
stream. Then, the gas was changed from He to a mixture of
1 % N2O in He. The N2O reacts quantitatively with the Cu
surface atoms forming gas-phase N2. The surface reaction
levels off when a breakthrough of the N2O front is observed.
The experiment is assumed to be finished after the N2O front
has reached the required value of 1 %. The formed amount of
N2 has been determined by the product of the molar flow and
the peak area under the N2 front in the time interval between
the onset of appearance of N2 and the intersection of the N2
and the N2O concentration in the exit gas. The specific Cu
metal surface area has been calculated by using a value of
1.47´1019 atoms per m2 for the mean Cu surface atom densiue,
which is the arithmetic mean value of the low index faces of
Cu, namely, the Cu(111), Cu(110), and Cu(100) crystal
surfaces.
2.3 Temperature-Programmed N2O Flow Experiments
Temperature-programmed N2O flow experiments have
been conducted according to the following sequence: After
flushing in He at 493 K the samples were cooled down to
approx. 190 K in an He stream. By changing the gas from He
to 1 % N2O in He the isothermal flow experiment was started.
The sample was subsequently heated up to a final temperature
of 693 K at 6 K/min.
2.4 N2O Pulse Flow Experiments
Pulses of N2O from a sample loop with a defined volume
(V = 1.0 ml, p = 0.1 MPa) were injected in an He stream by
switching a six-way Valco valve. The stream was passed over
the adsorbate-free reduced catalyst at room temperature and
at atmospheric pressure until the complete breakthrough of
the N2O front was observed.
2.5 Temperature-Programmed Desorption of Hydrogen
(H2 TPD)
The adsorbate-free reduced catalyst sample was cooled
down to 240 K in an He stream. Complete saturation with
atomic adsorbed hydrogen was achieved by dosing 50 Nml/
min H2 at 1.5 MPa for 1 h [9]. Then, the sample was cooled
down in H2 to 78 K by means of liquid nitrogen. The gas was
The temperature-programmed desorption experiment (Q =
100 Nml/min, p = 0.1 MPa) was started by ramping the
temperature up to 493 K at 6 K/min.
0930-7516/00/1111-0956 $ 17.50+.50/0
0930-7516/00/1111-0956 $ 17.50+.50/0 Chem. Eng. Technol. 23 (2000) 11

3 Results and Discussion
The chemisorption method of N2O is based on the
decomposition of N2O on exposed Cu atoms accompanied
by the formation of adsorbed atomic oxygen and gas-phase
nitrogen:
N2O + 2 Cus ! (Cus ± O ± Cus) + N2
Cus represents an exposed Cu surface atom and the
stoichiometry of Cus : Oads is assumed to be 2 : 1 (half-
monolayer coverage of oxygen). Several different experi-
mental techniques have been applied to determine the extent
of the reaction of N2O with the Cu surface. Dell et al. [1] and
Osinga et al. [2] measured the amount of nitrogen formed in a
static flow system by freezing out the remaining N2O and
determining the residual pressure in a conventional volu-
metric apparatus. Scholten and Konvalinka applied mass
spectrometry and thermal conductivity in a vacuum recircula-
tion system. Dvorak and Pasek [4] as well as Evans et al. [5]
used the pulse chemisorption method. Pulses of N2O were
injected in an inert gas stream and were passed over the
catalyst sample. The adsorptive decomposition of N2O in an
isothermal flow experiment for the determination of specific
metal surface areas of Cu catalysts was patented by Chinchen
et al. [6] as reactive frontal chromatography (N2O RFC).
The characteristic shape of N2 and N2O monitored in the
effluent gas during an N2O flow experiment conducted
isothermally at 300 K is displayed in Fig. 1. The gas was
changed from He to a mixture of 1 % N2O in He (Q = 10 Nml/
min) after 2.5 min. After a dead time caused by the volume of
the apparatus the eluted N2 formed by the decomposition of
N2O is recorded as a front. After a certain period the reactive
mixture of N2O in He breaks through. Finally, the experiment
was finished at the time t = 22.5 min.
Figure 1. N2O flow experiment with a Cu/ZnO/Al2O3 catalyst (TG01). Reaction
conditions: T = 300 K, p = 0.1 MPa, Q = 10 Nml/min, 1 % N2O in He, mcat = 0.2 g.
In order to determine the reproducibility relating to the
determination of the specific metal surface area, two N2O flow
experiments were conducted at the same day. The obtained
values of the specific Cu metal surface areas correspond very
Chem. Eng. Technol. 23 (2000) 11, Ó WILEY-VCH Verlag GmbH, D-69469 Weinheim, 2000
15214125, 2000, 11, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/1521-4125(200011)23:11<956::AID-CEAT956>3.0.CO;2-L by Universidad Autonoma De Madrid, Wiley Online Library on [23/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Communications
well indicating that there is no irreversible damage of the Cu
particles under the mild reaction conditions used in this
experiment. The reproducibility of the values of the specific
Cu surface area amounts to about 1 %.
During the reaction between N2O and the Cu surface a heat
of reaction of about 317 kJ/(mol N2O, Cu/ZnO) is released
[10]. The sample mass of 0.2 g results in a catalyst bed height of
(1) about 20 mm. Using a total flow rate of 10 Nml/min, the mean
contact time yields to 1.4 s. Under this reaction condition the
observed temperature rise of about 1 K can be neglected. A
reduction of the contact time to 0.35 s by increasing the total
flow rate up to 40 Nml/min results within the limits of the
measuring accuracy to corresponding values of the Cu metal
surface area. If the heat dissipation is insufficient, e.g., due to
higher N2O partial pressures, consecutive reactions could
begin resulting in irreproducible values for the Cu metal
surface areas. Therefore, it can be concluded that the N2O
RFC is a reliable and fast method for the determination of Cu
metal surface areas under the reaction conditions used in this
experiment.
Fig. 2 displays the consumption of N2O and the formation of
N2 during the application of 90 N2O pulses over the catalyst
TG02 at 300 K in an He carrier gas stream (Q = 20 Nml/min).
The enveloped line shapes are comparable with the ones of the
N2O flow experiment: Until pulse # 52 (which corresponds to a
formation of N2 of about 106 lmol/g) the N2O reacts
completely with the Cu particles generating gas-phase N2. In
the sequel of the experiment, a breakthrough of the N2O front
is observed. The N2 formed until the crossing point of the N2
and the N2O curve (pulse # 66) amounts to about 125 lmol/g.
The specific surface area was determined by N2O RFC before
and after the pulse experiment. The obtained values of 9.3 m2/
g and 9.5 m2/g, respectively, are in good agreement with the
value of 10.3 m2/g determined by the pulse flow experiment.
Figure 2. Application of N2O pulses (V = 1 ml, 1 % N2O in He, p = 0.1 MPa) at
T = 300 K over the adsorbate-free catalyst sample TG02.
Since at higher temperature the oxidation of subsurface
layers and of the bulk can occur, the most suitable reaction
temperature applied during the N2O RFC bas been discussed
very intensively in the literature during the last decades
[2,6,7,11,12]. According to Osinga et al. [2] bulk oxidation
0930-7516/00/1111-0957 $ 17.50+.50/0 957
 15214125, 2000, 11, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/1521-4125(200011)23:11<956::AID-CEAT956>3.0.CO;2-L by Universidad Autonoma De Madrid, Wiley Online Library on [23/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Communications

takes place at 423 K. However, Scholten and Konvalinka [3] conducted prior to this temperature-programmed experiment
observed an increase of the N2O consumption by increasing (see Tab. 1) both methods are leading to corresponding results.
the temperature during their studies on polycrystalline copper
and supported copper catalysts. Sengupta et al. [11] concluded
from their measurements that already at temperatures above
343 K bulk oxidation takes place. Chinchen et al. [6] made the
suggestion to conduct the N2O reactive frontal chromatogra-
phy experiment at 333 K. They reported that below 333 K the
Cu surface atoms do not react quantitatively with the N2O,
whereas at temperatures above 363 K bulk oxidation takes
place. Furthermore, Luys et al. [7] were able to show by means
of flow experiments in the temperature range from 303 K to
393 K that the surface oxidation is accompanied by a slow,
temperature-dependent oxidation of subsurface layers. Köp-
pel [12] compares the pulse method with the flow method at
333 K and at 363 K using different Cu/ZrO2 catalysts. Figure 3. Isothermal N2O flow experiment at T = 189 K followed subsequently
by a temperature-programmed N2O experiment with the catalyst sample TG01
By conducting Cu single crystal experiments under UHV
(p = 0.1 MPa, Q = 20 Nml/min, mcat = 0.2 g).
conditions Habraken [13] found out the following sequence
for the initial reactive sticking probability of N2O: s0 = 0.15 for
Cu(110), s0 = 5´10±5 for Cu(100) and s0 = 2´10±9 for Cu(111) at Table 1. Desorbed amount of N2 and H2, respectively, and calculated specific Cu
surface areas for the catalyst TG01.
room temperature. Therefore, the maximal oxygen coverage
should be achieved during the N2O RFC at 300 K displayed in Sequence of Amount of H2 or N2 Specific Cu metal
Fig. 1 after a few seconds, even in the worst case of Cu(111). experiments with TG01 [lmol/g] surface area [m2/g]
Indeed, this result is consistent with the sharp breakthrough of H2 TPD 115 18.8
the N2 and N2O fronts obtained during N2O RFC. In order to 1 % N2O RFC 217 17.8
study the influence of the temperature on the formation of H2 TPD 115 18.8
subsurface oxides during the N2O dissociation kinetics on
supported Cu catalysts in more detail, an N2O flow experi-
To clarify the competition between reversible physisorption
ment was conducted at T = 189 K with subsequently heating
and irreversible chemisorption of N2O, the dosing tempera-
the catalyst sample linearly up to 750 K in a mixture of 1 %
ture during the isothermal part of the temperature-pro-
N2O in He (Q = 20 Nml/min). Fig. 3 displays the sequence of
grammed N2O flow experiment was varied (Fig. 4). An
the experiment with the catalyst TG01. The shape of the
increase of the dosing temperature (T = 236 K) leads to a lower
breakthrough in the isothermal temperature range is identical
physisorption capability and to a larger amount of N2O per
with the one of the N2O RFC conducted at the higher
time unit that is adsorptively decomposed.
temperature of 300 K displayed in Fig. 1. The Cu metal surface
areas determined by both methods correspond within the
limits of the measuring accuracy. Since a front migrates
through the catalyst bed, there is a temperature-dependent
competition between unselective equilibrium physisorption
and Cu surface selective chemisorption of the N2O at every
point in the catalyst bed. Heating the sample leads to the
desorption of weakly adsorbed N2O. The amount of N2O
physisorbed during the isothermal part of the experiment
corresponds to the amount of N2O released at higher
temperatures. In the sequel of the experiment, the N2
formation increases in the temperature range between 250 K
and 350 K. Obviously, the formation of a half-monolayer
coverage of oxygen is superimposed by a slow subsurface
Figure 4. Variation of the reaction temperature during an isothermal N2O flow
oxidation process. Therefore, a separation between these two experiment with the catalyst sample TG01 (dosing temperature T = 236 K,
processes is hardly attainable. At temperatures higher than p = 0.1 MPa, Q = 10 Nml/min, mcat = 0.2 g).
625 K copper is oxidized by N2O to CuO [7]. The integration of
the N2 signal up to the intersection of the N2 and the N2O front Muhler et al. [8] established the temperature-programmed
results in an N2 amount of 144 lmol/g while the integration up desorption of H2 (H2 TPD) as an alternative technique for the
to the start of the heating ramp yields a value of 173 lmol/g. determination of Cu metal surface areas. Complete saturation
The amount of N2 formed up to 300 K is 216 mol/g. Compared with adsorbed atomic hydrogen was achieved by dosing pure
to the value of 217 lmol/g determined by the N2O RFC H2 for one hour at 240 K and at a pressure of 1.5 MPa [9]. In the

958 Ó WILEY-VCH Verlag GmbH, D-69469 Weinheim, 2000 0930-7516/00/1111-0958 $ 17.50+.50/0 Chem. Eng. Technol. 23 (2000) 11
 15214125, 2000, 11, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/1521-4125(200011)23:11<956::AID-CEAT956>3.0.CO;2-L by Universidad Autonoma De Madrid, Wiley Online Library on [23/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Communications

subsequent temperature-programmed desorption experi- Symbols used

ment, a narrow symmetric peak with a full width at half
maximum of only 50 K centered at about 300 K is observed mcat [g] catalyst mass
(Fig. 5, QHe = 100 Nml/min, p = 0.1 MPa, b = 6 K/min). The p [Pa] total pressure
value obtained by integration of the peak area multiplicated Q [Nml min±1] flow rate
with the molar flux results in an adsorbed amount of H2 of 115 s0 [±] initial sticking probability
lmol/g. Based on a stoichiometry of Cus : Hads = 2 at T [K] temperature
saturation, the specific Cu metal surface area can be V [ml] volume of the sample loop
calculated to 18.8 m2/g.

Greek symbol

b [K min±1] heating rate

References
[1] Dell, R. M.; Stone, F. S.; Tiley, P. F., Trans. Faraday Soc. 49 (1953) p. 195.
[2] Osinga, T. J.; Linsen, B. G.; van Beek, W. P., J. Catal. 7 (1967) p. 277.
[3] Scholten, J. J. F.; Konvalinka, J. A., Trans. Faraday Soc. 65 (1969) p. 2465.
[4] Dvorak, B.; Pasek, J., J. Catal. 18 (1970) p. 108.
[5] Evans, J. W.; Wainwright, M. S.; Bridgewater, A. J.; Young, D. J., Appl.
Catal. 7 (1983) p. 75.
[6] Chinchen, G. C.; Hay, C. M.; Vanderwell, H. D.; Waugh, K. C., J. Catal. 103
(1987) p. 79.
[7] Luys, M. J.; van Oeffelt, P. H.; Brouwer, W. G. J.; Pijers, A. P.; Scholten, J. J.
Figure 5. Temperature-programmed desorption experiment in He with the F., Appl. Catal. 46 (1989) p. 161.
catalyst sample TG01 obtained after complete saturation with adsorbed atomic [8] Muhler, M.; Nielsen, L. P.; Törnqvist, E.; Clausen, B. S.; Topsùe, H., Catal.
hydrogen. The line shapes before (solid line) and after the N2O RFC experiment Lett. 14 (1992) p. 241.
(dashed line) are not distinguishable. Reaction conditions: p = 0.1 MPa, QHe= [9] Genger, T.; Hinrichsen, O.; Muhler, M., Catal. Lett. 59 (1999) p. 137.
100 Nml/min, mcat = 0.2 g, b = 6 K/min. [10] Giamello, E.; Fubini, B.; Lauro, P.; Bossi, A., J. Catal. 87 (1984) p. 443.
[11] Sengupta, G.; Gupta, D. K.; Kundu, M. L.; Sen, S. P., J. Catal. 67 (1981) p.
223.
Furthermore, experiments were conducted in the sequence [12] Köppel, R. A., Ph. D. Thesis, ETH Zürich 1991.
H2 TPD, N2O RFC and H2 TPD with the catalyst TG01. The [13] Habraken, F. H. P. M., Ph. D. Thesis, University of Utrecht 1980.
amount of desorbed H2 and N2, respectively, as well as the
This paper was also published in German in Chem. Ing. Tech. 72 (2000) No. 1+2,
values of the specific Cu metal surface area calculated by using pp. 94±98.
the appropriate stoichiometry are displayed in Tab. 1. The _______________________
values are in good agreement showing that both methods can
be used for the quantitative determination of Cu metal surface
areas, if the N2O RFC is applied under mild conditions.
Development of Ceramic Membranes
and Adsorbents from Silicon-Organic
4 Conclusions Precursors

The extent of the subsurface oxidation is determined by
various reaction parameters, like temperature, mean contact
time of the N2O with the catalyst sample and the proportion of
Cu metal surface area and the molar flux of N2O. However, a
standardized procedure performed under mild reaction
conditions guarantees a comparison of Cu metal surface areas
measured for different supported Cu catalysts.
Under the mild reaction conditions presented here the N2O
RFC is a simple, fast and reliable method. The selection of the
experimental mode (pulse or flow experiments) is not decisive.
The temperature-programmed desorption of H2 (H2 TPD)
is an additional independent experiment for the determina-
tion of Cu metal surface areas. This method enables the
verification of the experimental conditions for the N2O RFC
which is easier to conduct.
Received: February 29, 2000 [K 2604]
By Harald Schmidt, Dietmar Koch, and Georg Grathwohl*
1 Introduction
Si-metallorganic polymers are excellent starting materials
for manufacturing ceramic powders and compacts. The topic
of this paper is the pyrolysis of such polysiloxanes, which
results in an amorphous microporous Si-O-C matrix with
favorable high temperature strength and chemical stability
compared to conventional porous organic matrices, e.g.,
polymeric membranes, filters and adsorbents.
±
[*] Dipl.-Ing. H. Schmidt, Dr.-Ing. D. Koch, Prof. Dr.-Ing. G. Grathwohl,
Keramische Werkstoffe und Bauteile, Universität Bremen, Fachbereich
Produktionstechnik, Am Biologischen Garten 3, D- 28359 Bremen; email:
grath@ceramics.uni-bremen.de

Chem. Eng. Technol. 23 (2000) 11, Ó WILEY-VCH Verlag GmbH, D-69469 Weinheim, 2000 0930-7516/00/1111-0959
0930-7516/00/1111-0959 $$ 17.50+.50/0
17.50+.50/0 959