Towards efficient Cu/ZnO catalysts for ester hydrogenolysis: The role of synthesis method

Article abstract of DOI:10.1016/j.apcata.2021.118320

Cu/ZnO catalysts represent an environmentally friendly alternative to Adkins catalysts used for ester hydrogenolysis. Cu/ZnO are mostly synthesized by co-precipitation (COP); however, other synthesis methods may help to obtain small highly dispersed Cu crystallites advantageous for catalyst activity. A comparative study on the effect of synthesis method on Cu/ZnO catalysts properties and activity is missing. Thus, we synthesized 8 wt% Cu/ZnO catalysts by five methods (COP, deposition-precipitation (DP), chemisorption-hydrolysis (CH), incipient wetness (IWI) and wet impregnation (WI)), characterized and tested them in dimethyl adipate hydrogenolysis. The CH-prepared catalyst was prone to Cu sintering, which impaired its performance. IWI led to large Cu nanoparticles, pore blocking and poor catalytic performance. COP and DP resulted in the smallest Cu nanoparticles (13?14 nm), largest Cu surface area (3.9–4.2 m² g_cat^?1) and specific surface area (40?43 m² g_cat^?1) reflected in their superior catalytic activity making DP a good alternative to COP to prepare active Cu/ZnO catalysts.

Full text of DOI:10.1016/j.apcata.2021.118320

Applied Catalysis A, General 624 (2021) 118320
Contents lists available at ScienceDirect
Applied Catalysis A, General
journal homepage: www.elsevier.com/locate/apcata
Towards efficient Cu/ZnO catalysts for ester hydrogenolysis: The role of
synthesis method
a
,
a
b
a,b
ˇ
Violetta Pospelova *, Jaroslav Aubrecht , Oleg Kikhtyanin , David Kubicka
a
´
Department of Petroleum Technology and Alternative Fuels, University of Chemistry and Technology Prague, Technicka 5, 166 28 Prague, Czech Republic
b
´
ˇ
Technopark Kralupy, University of Chemistry and Technology Prague, Namestí G. Karse 7/2, 278 01 Kralupy nad Vltavou, Czech Republic
A R T I C L E I N F O
A B S T R A C T
Keywords:
Cu/ZnO catalysts represent an environmentally friendly alternative to Adkins catalysts used for ester hydro-
genolysis. Cu/ZnO are mostly synthesized by co-precipitation (COP); however, other synthesis methods may help
to obtain small highly dispersed Cu crystallites advantageous for catalyst activity. A comparative study on the
effect of synthesis method on Cu/ZnO catalysts properties and activity is missing. Thus, we synthesized 8 wt%
Cu/ZnO catalysts by five methods (COP, deposition-precipitation (DP), chemisorption-hydrolysis (CH), incipient
wetness (IWI) and wet impregnation (WI)), characterized and tested them in dimethyl adipate hydrogenolysis.
The CH-prepared catalyst was prone to Cu sintering, which impaired its performance. IWI led to large Cu
Copper-based catalyst
Ester hydrogenolysis
Supported catalyst synthesis
nanoparticles, pore blocking and poor catalytic performance. COP and DP resulted in the smallest Cu nano-
particles (13ꢀ 14 nm), largest Cu surface area (3.9–4.2 m2 gcatꢀ 1) and specific surface area (40ꢀ 43 m2 gcat
)
ꢀ 1
reflected in their superior catalytic activity making DP a good alternative to COP to prepare active Cu/ZnO
catalysts.
1. Introduction
a temperature ranging from 170 to 205 ◦C and hydrogen pressure of
160 bar [8]. Moreover, it was confirmed that the ester conversion was
Ester hydrogenolysis is a long-established industrial process con-
verting esters to the corresponding alcohols that can be used in polymer,
cosmetics and pharmaceutical industries. Currently, the process is
operated at severe reaction conditions with temperature varying from
200 to 300 ◦C and hydrogen pressure ranging from 250 to 300 bar [1].
Moreover, the process still uses chromium-containing copper-based
catalysts developed as early as in 1931 by Adkins et al. [2]. Even though
dependent on CuO or Cu crystallite size, where the conversion was
increasing with the decreasing crystallite size [9]. However, the copper
crystallites in CuZnO co-precipitated catalysts were not stable as their
size increased and the BET surface area dropped significantly during the
experiment [8,10]. Instead of co-precipitation, high-surface area sup-
ports can be used to obtain high dispersion of Cu species. When the Cu
species are well anchored on a support, e.g. due to the use of optimized
preparation method allowing a good Cu-support interaction, highly
active catalysts with stable performance can be obtained [11].
–
the Adkins catalyst is highly efficient for the selective C O bond
hydrogenolysis, the presence of chromium limits its usage in future due
to the tightening environmental regulations. As a result, a lot of effort is
being put into finding a chromium-free alternative catalyst with the
same efficiency at milder reaction conditions.
Preparation methods such as incipient wetness impregnation, wet
impregnation, deposition-precipitation and chemisorption-hydrolysis
can be used to prepare supported Cu catalysts. These methods are not
as common as co-precipitation used typically for copper-zinc-based
catalyst synthesis, however, they might help to obtain small, highly
dispersed and stable Cu crystallites [11–14]. Cu/ZnO-based catalyst
prepared by incipient wetness impregnation was more stable in the
glycerol hydrogenolysis than a CuZnO-based catalyst synthesized by
co-precipitation as fewer polymeric products resulted in coke formation
were formed [15]. Cu/ZnO synthesized by deposition-precipitation had
While chromium plays a role of a structural promoter increasing the
stability of copper crystallites, copper is the active site for hydro-
genolysis [3]. Therefore, the copper-based catalysts such as CuZnO,
CuZnO/Al2O3, Cu/Al2O3 and Cu/SiO2 have been recently tested in ester
hydrogenolysis [4–7].
CuZnO catalysts synthesized by co-precipitation have previously
been shown to be highly efficient and selective in ester hydrogenolysis at
* Corresponding author.
E-mail address: pospelov@vscht.cz (V. Pospelova).
https://doi.org/10.1016/j.apcata.2021.118320
Received 28 May 2021; Received in revised form 16 July 2021; Accepted 11 August 2021
Available online 13 August 2021
0926-860X/© 2021 Elsevier B.V. All rights reserved.

V. Pospelova et al.
Applied Catalysis A, General 624 (2021) 118320
smaller Cu crystallites (ca. 6 nm) than CuZnO synthesized by
co-precipitation (about 9 nm) [16]. The synthesis of Cu/ZnO by
chemisorption-hydrolysis has not been previously described, however,
when using different supports such as TiO2 [17] or SiO2 [12], this
method allowed synthesizing small Cu particles. To conclude, the use of
methods different from co-precipitation for Cu/ZnO synthesis might be
beneficial from Cu crystallite size and their stability point of view.
Therefore, a comprehensive study on the effect of Cu/ZnO synthesis
method on the catalyst properties and performance in ester hydro-
genolysis might provide new insights in the activity of hydrogenolysis
catalysts.
COP: A glass beaker was first filled with 200 mL of distilled water
and heated to T = 60 ◦C under stirring at 400 RPM. An aqueous solution
of Cu(NO3)2.3H2O and Zn(NO3)2.6H2O (99.6 %, Lach:Ner, s.r.o., Ner-
atovice, Czech Republic) in the total concentration of 0.5 M of a Cu/Zn
ratio of 0.1 and an aqueous solution of a precipitant Na2CO3 (99.4 %,
Lach:Ner, s.r.o., Neratovice, Czech Republic) were pumped simulta-
neously into the beaker. The flow rate of the mixed salt solution was
fixed while that of the precipitant solution was continuously adjusted by
changing the pump performance to keep a constant pH of 7 ± 0.1. The
temperature and stirring rate were kept constant at 60 ◦C and 400 RPM,
correspondingly, during the precipitation. The resulting suspension was
aged under stirring for 90 min which caused an increase in the pH value
to 8.1–8.2. The prepared precipitate was then filtered using a vacuum
pump, washed with plenty of distilled water, and finally dried at 70 ◦C
overnight. Then it was calcined at 350 ◦C for 3 h (heating rate of
120 ◦C hꢀ 1). The catalyst was denoted as CuZnO-2-COP as copper and
zinc metals were present as mixed oxides formed from the mixed Cu,Zn-
hydroxycarbonates.
Here, we present the effect of the synthesis method (incipient
wetness impregnation, wet impregnation, deposition-precipitation,
chemisorption-hydrolysis and co-precipitation) on the final copper-
zinc catalysts properties such as Cu crystallite size, active surface area
and their activity in dimethyl adipate hydrogenolysis.
2. Experimental part
For the sake of comparison, the described COP procedure using
single solution of Zn(NO3)2.6H2O was applied for the precipitation of
hydrozincite Zn5(CO3)2(OH)6. Due to calcination at 350 ◦C for 3 h
(heating rate of 120 ◦C hꢀ 1), Zn5(CO3)2(OH)6 was decomposed to ZnO.
Both the as-prepared Zn5(CO3)2(OH)6 and calcined ZnO were used for
the synthesis of catalysts with the 8 wt% Cu loading by the WI method
described above and they are denoted as CuZnO-2-WI and Cu/ZnO-2-
WI, respectively.
2.1. Catalyst synthesis
Supported catalysts with the nominal Cu loading of 8 wt% were
synthesized by four different methods, including incipient wetness
impregnation (IWI), wet impregnation (WI), deposition-precipitation
(DP) and chemisorption-hydrolysis (CH) using a commercial ZnO sup-
port (>98 % ZnO, Albemarle Corporation, Charlotte, North Carolina,
United States, denoted as ZnO-1). For comparison, CuZnO catalyst (8 wt
% of Cu) was synthesized by a co-precipitation method (COP). Each
synthesis procedure was optimized beforehand (detailed description in
SI, Tables S1 and S2).
2.2. Catalyst characterization
The elemental composition of the calcined and spent catalysts was
determined by AAS using Agilent 280FS AA, where a mixture of acety-
lene and air was used as an atomization flame.
IWI: The pore volume of a commercial ZnO support was measured by
water titration to be equal to 0.64 mL∙gꢀ 1. 6.083 g of Cu(NO3)2.3H2O
(99.0 %, Penta, s.r.o., Prague, Czech Republic) was dissolved in the exact
volume of distilled water according to the pore volume (10.65 mL). The
nitrate solution was then added to the 18.76 g of ZnO support under
continuous mixing. The suspension was then left at room temperature
for 1 h, then dried at 90 ◦C for 16 h (heating rate of 60 ◦C hꢀ 1) and
calcined at 350 ◦C for 3 h (heating rate of 120 ◦C hꢀ 1). The catalyst was
denoted as Cu/ZnO-1-IWI.
The phase composition of the catalysts after calcination and reaction
was determined by X-Ray diffraction using a diffractometer PAN-
analytical X’Pert3 Powder and Cu Kα radiation. The XRD patterns were
recorded in the range of 2θ = 5◦–90◦. The crystallite sizes were esti-
mated using the Scherrer’s equation using the reflections at 2θ = 31.8◦,
38.6◦ and 43.3◦ for ZnO, CuO and Cu, respectively [18].
Nitrogen physisorption was measured at 77 K using a static volu-
metric adsorption system (TriFlex analyzer, Micromeritics, Norcross,
USA). The samples were degassed at 473 K (12 h) prior to N2 adsorption
analysis, in order to obtain a clean surface. The adsorption isotherms
were fitted using the Brunauer-Emmett-Teller (BET) method for the
specific surface area [19] and the BJH method for the pore size distri-
bution [20].
WI: 6.083 g of Cu(NO3)2.3H2O was dissolved in 100 mL of distilled
water. 18.76 g of ZnO support then was added to the nitrate solution
under continuous stirring at 300 RPM for 60 min at room temperature.
The suspension was then filtered, dried at 90 ◦C for 16 h (heating rate of
60 ◦C hꢀ 1) and calcined at 350 ◦C for 3 h (heating rate of 120 ◦C hꢀ 1).
The catalyst was denoted as Cu/ZnO-1-WI.
The reducibility of the calcined catalysts was determined by Quan-
tachrome ChemStar TPx instrument with a TCD detector. The samples
were exposed to a reducing gas mixture containing 10 vol% of hydrogen
in argon with a flow of 50 mL minꢀ 1. A U-shaped reactor with a sample
(typically 0.15 g) was located in a furnace and treated with the reducing
gas at a heating rate of 5 ◦C minꢀ 1 from room temperature up to 600 ◦C.
The hydrogen consumption was calculated using a calibration curve.
To evaluate the active copper specific surface area, the reactive
frontal chromatography method using N2O was used. The used instru-
ment was Autochem II 2920 (Micromeritics, United States) connected
on-line to a quadrupole mass spectrometer RGA 200 (Prevac, Poland).
The detailed procedure was described previously [8].
DP: The amount of 6.083 g of Cu(NO3)2.3H2O was dissolved in
900 mL of distilled water and placed in a 2-liter round bottom flask.
7.56 g of urea (99.5 %, Penta, s.r.o., Prague, Czech Republic) was added
to the solution and dissolved using an ultrasonic bath. ZnO support
(18.76 g) was added to the solution under constant stirring (300 RPM).
The prepared suspension was then slowly heated to 90 ◦C at the heating
rate of 60 ◦C hꢀ 1. The suspension was stirred at 90 ◦C for 24 h. The final
pH of the suspension was equal to 7. The suspension was then filtered,
dried at 90 ◦C for 16 h (heating rate of 60 ◦C hꢀ 1) and calcined at 350 ◦C
for 3 h (heating rate of 120 ◦C hꢀ 1). The catalyst was denoted as Cu/
ZnO-1-DP.
CH: The amount of 6.083 g of Cu(NO3)2.3H2O was dissolved in
100 mL of distilled water and placed in a glass beaker, where ammonia
(24 % NH3 aqueous solution, Penta, s.r.o., Prague, Czech Republic) was
added dropwise into the solution until pH = 9. Then, 18.76 g of ZnO was
added, and the suspension was stirred at 300 RPM for 10 min. After that,
the mixture was placed into an ice bath under constant stirring and
2.3. Catalyst testing
All experiments of hydrogenolysis of dimethyl adipate (DMA) (99 %,
Sigma-Aldrich, United States) were carried out in a Parr stainless steel
autoclave with a reactor volume of 300 mL. The amount of 4 g of the
catalyst was loaded into the autoclave and reduced in situ at 250 ◦C
using H2 (99.9 %, SIAD Czech, s.r.o., Czech Republic). After the reduc-
tion, 120 g of DMA was loaded into the reactor. The reaction
600 mL of water was added dropwise at the flow rate of 5 mL minꢀ 1
.
The suspension was then filtered, dried at 90 ◦C for 16 h (heating rate of
60 ◦C hꢀ 1) and calcined at 350 ◦C for 3 h (heating rate of 120 ◦C hꢀ 1).
The catalyst was denoted as Cu/ZnO-1-CH.
2

V. Pospelova et al.
Applied Catalysis A, General 624 (2021) 118320
temperature was kept at 250 ◦C, while the hydrogen pressure was kept
at 100 bar. Liquid reaction products were periodically (5, 10, 20, 40, 60,
90, 120, 180 min) withdrawn from the reactor during the experiment,
diluted with methanol (1:25 V/V) and analysed by an Agilent 7820 GC-
(CuZnO-2-COP) was compared with the supported Cu/ZnO catalysts of
8 wt% Cu loading synthesized by four different methods using a com-
mercial ZnO support (ZnO-1): incipient wetness impregnation (Cu/ZnO-
1-IWI), wet impregnation (Cu/ZnO-1-WI), deposition-precipitation (Cu/
ZnO-1-DP) and chemisorption-hydrolysis (Cu/ZnO-1-CH). Moreover,
two additional catalysts with the 8 wt% Cu loading were synthesized by
the WI method using the as-prepared Zn5(CO3)2(OH)6 (CuZnO-2-WI)
and the ZnO obtained by calcining Zn5(CO3)2(OH)6 (Cu/ZnO-2-WI) as a
support instead of the commercial ZnO support. The use of the as-
prepared Zn5(CO3)2(OH)6 for impregnation might provide a closer
contact of copper and zinc phases (as in the COP procedure) because
both CuO and ZnO would be formed simultaneously during the
calcination.
FID (HP-5 capillary column, 30 m length, 0.32 mm i.d., 0.25 μm thick).
2.4. Calculations
The catalyst performance was evaluated using three parameters:
DMA conversion, selectivity to products and product yields. The DMA
conversion was calculated using Eq. (1). Due to the absence of cracking
reactions (confirmed by the GC analysis of the products), the C6 back-
bone of DMA as well as of the reaction products was used as the basis for
the selectivity calculation (Eq. (2)). Methanol was excluded because it
was present both in the liquid and gaseous product streams, which
prevented its accurate quantification. The yields of the individual
products were calculated using Eq. (3).
The elemental composition of the calcined catalysts determined by
AAS revealed that the copper amount in the final catalysts was in the
range of 7ꢀ 10 wt% depending on the used method (shown in Table 1).
The Cu/Zn atomic ratio was in the range of 0.10ꢀ 0.12.
The phase composition of the calcined catalysts and the crystallite
sizes of CuO and ZnO phases was determined by XRD. As shown in Fig. 1,
there were sharp reflexes of ZnO in all studied catalysts, while the
sharpness of CuO reflexes differed. Taking into account the lower in-
tensity of CuO peak at 2θ = 38.6 ◦ and the absence of another CuO peak
at 2θ = 48 ◦ in CuZnO-2-COP, Cu/ZnO-1-DP and Cu/ZnO-1-CH samples,
it might be proposed that these three methods produced smaller and
highly dispersed copper oxide crystallites than the other methods.
The crystallite size of CuO was estimated using the Scherrer’s
equation using reflections at 2θ = 31.8 ◦ and 38.6 ◦ for ZnO and CuO,
respectively [18]. As it was suggested based on the XRD patterns (Fig. 1),
the smallest CuO crystallites of 7, 9 and 10 nm were obtained by the
COP, DP and CH method, respectively (Fig. 2A). The largest CuO crys-
tallites of 21 nm were in the catalyst produced by the IWI method. The
formation of large copper clusters and poor copper dispersion when
using the IWI method was also observed in case of Cu supported on TiO2
[21], SiO2 [22], SBA-15 [14] or rice husk ash [23]. It is worth
mentioning that the size of CuO was the same ranging from 15 to 17 nm
in all catalysts synthesized by WI method when using the commercial
ZnO (Cu/ZnO-1-WI), the as-prepared Zn5(CO3)2(OH)6 (CuZnO-2-WI) or
the synthesized ZnO (Cu/ZnO-2-WI) supports. As for ZnO crystallite
size, the initial crystallite size in the commercial support was 15 nm,
while after the copper introduction it changed depending on the method
used. As seen in Fig. 2B, the ZnO size remained the same in case of DP or
CH and increased slightly to 19 nm in case of WI. The more significant
increase in the ZnO crystallite size to 26 nm in IWI-synthesized catalyst
might indicate the destruction of the initial support structure that was
reported to occur during the air or solvent elimination from the pores
[24]. Previously, we have reported that the size of ZnO in CuZnO cat-
alysts (Cu/Zn = 1) was smaller in the CuZnO co-precipitated catalysts
than in a single-phase ZnO catalyst [8]. In other words, it can be sug-
gested that when copper and zinc are in intimate contact during the
synthesis, the size of ZnO is smaller. In agreement, the CuZnO-2-COP
had the smallest ZnO crystallites of 12 nm and the CuZnO-2-WI syn-
thesized using Zn5(CO3)2(OH)6 had a smaller ZnO crystallite size
compared to Cu/ZnO-2-WI where the synthesized ZnO was used.
nDMA,i ꢀ nDMA, t
DMA conversion (%) =
∙100
(1)
nDMA,i
nDMA,x
nDMA,products
Sx(%) =
Yx(%) =
∙100
(2)
DMA conversion∙Sx
(3)
100
where: nDMA is the number of DMA moles, i is the initial reaction time, t
is the sampling time, Sx is the selectivity to product x, nDMA,x is the
number of DMA moles converted to product x, nDMA,products is the number
of DMA moles converted to all products, Yx is the yield of the product x.
The intrinsic hydrogenolysis activity of a catalyst expressed as the
turnover frequency (TOFH) was calculated using Eq. (4).
dnDMA,H
dt
σCu⋅NA
mCAT ⋅SCu
TOFH
=
⋅
(4)
where: dn/dt is the change in number of moles relative to change in
time, nDMA,H is the number of DMA moles converted by hydrogenolysis
(mol), NA is the Avogadro constant equal to 6.022⋅1023 molꢀ 1, mCAT is
the mass of catalyst (g) and SCu is the specific copper surface area
(m2 gcatꢀ 1), t is the reaction time (0ꢀ 20 min period of experiment was
taken into the calculation) (s), σCu is the cross-section area of a copper
atom that is equal to 0.0154 nm2.
The Cu cluster size and Cu dispersion were calculated based on N2O
chemisorption using Eqs. (5) and (6), respectively. The hemisphere
model for the cluster shape was assumed.
6
Cu cluster size (nm ) =
(5)
ρCu∙SCu
where: ρCu is the copper density equal to 8.96 g cmꢀ 3, SCu is the specific
copper surface area (m2 gCu-1).
AN2∙MCu
Cu dispersion (%) =
(6)
wCu
where: AN2 is the number of N2 moles released by the adsorption of N2O
normalized to the standard conditions (0 ◦C; 101.325 kPa) (mol), wCu is
the copper mass in a measured sample (g), MCu is the molar mass of Cu
(g molꢀ 1).
Table 1
Catalyst elemental composition.
Calcined
Spent
Catalyst
Cu (wt%)
Cu/Zn
Cu/Zn
3. Results and discussion
Cu/ZnO-1-DP
Cu/ZnO-1-CH
Cu/ZnO-1-IWI
Cu/ZnO-1-WI
Cu/ZnO-2-WI
CuZnO-2-WI
CuZnO-2-COP
8
0.10
0.10
0.11
0.12
0.11
0.08
0.10
0.10
0.09
0.11
0.12
0.11
0.08
0.08
8
3.1. Catalyst characterization
9
10
9
3.1.1. Calcined and reduced catalysts
7
To describe the effect of the synthesis method on the catalyst prop-
erties, CuZnO catalyst (8 wt% of Cu) synthesized by co-precipitation
8
3

V. Pospelova et al.
Applied Catalysis A, General 624 (2021) 118320
that the impregnation of Zn5(CO3)2(OH)6 using WI was similar to the
COP procedure. Contrarily, when Zn5(CO3)2(OH)6 was calcined to ZnO
and used for impregnation (catalyst Cu/ZnO-2-WI), it had one of the
lowest BET surface areas (15 m2 gcatꢀ 1). The low surface area of Cu/
ZnO-2-WI as well as of Cu/ZnO-1-IWI was probably a result of both
the large size of CuO and ZnO crystallites and the decreased pore vol-
ume. In Fig. 3B, there is a clear trend showing that the pore volume
decreased with the increasing CuO crystallite size in the calcined cata-
lysts. Thus, the reason of the pore blockage might be the sintering of
copper crystallites that are unevenly distributed in the pores and/or not
thermally stable due to the poor interaction of the copper oxide clusters
with the support [11].
The measurement by N2 physisorption revealed how the method of
copper deposition on a support may influence the pore size distribution
(Fig. 4). CuZnO-2-COP had smaller pores compared to the supported
catalysts. Using DP and CH methods, the initial pore size distribution of
the commercial ZnO support was mostly preserved. However, the pore
volume decreased by 60 % when the copper was introduced by the WI
method and by 70 % when the IWI method was used. This further
supports the findings of XRD and BET results. The pore size distribution
for the additional catalysts based on ZnO-2 is presented in Fig. S1.
To describe the effect of the synthesis method on the copper oxide
reducibility, the calcined catalysts were analyzed by H2-TPR. In Fig. 5,
the H2-TPR profiles of the calcined catalysts are shown. At first glance,
the Cu/ZnO-1-IWI differed from all other catalysts, as the reduction
maximum of the Cu/ZnO-1-IWI was at 273 ◦C. It might be explained by
the existence of large CuO particles (43 nm) at the external catalyst
surface because the reduction temperature of pure CuO is shifted to
higher temperatures (ca. 300 ◦C) compared to the co-precipitated
CuZnO catalysts [8]. Contrarily, small highly dispersed CuO crystal-
lites in COP (7 nm), DP (9 nm) and CH (10 nm) samples were easily
reducible and therefore, these catalysts were completely reduced below
200 ◦C showing that the reducibility depended on the CuO crystallite
Fig. 1. XRD patterns of the calcined catalysts (▴- ZnO, ● – CuO).
The calcined catalysts were analyzed using N2 physisorption to
measure the catalyst specific surface area and pore distribution. The
initial textural properties of the supports (ZnO-1 and ZnO-2) are pre-
sented in Table S3. As seen in Table 2, despite the use of the same
commercial support, the final catalyst surface area increased in the
following
order:
Cu/ZnO-1-IWI < Cu/ZnO-1-WI < Cu/ZnO-1-
CH < ZnO-1 < Cu/ZnO-1-DP. This trend was in accordance with the
XRD-estimated CuO crystallite size for the catalysts derived from the
same ZnO support (shown in Fig. 3A). Moreover, the suggestion of the
high metal dispersion in DP, COP and CH catalysts based on the XRD
patterns was supported by the highest BET surface area values. Another
catalyst with a high BET surface area was CuZnO-2-WI showing the same
BET area of 40 m2 gcatꢀ 1 as CuZnO-2-COP, confirming, thus, the theory
Fig. 2. Crystallite size estimated from XRD (A - CuO and Cu, B – ZnO).
Table 2
Catalyst properties obtained by N2 physisorption, N2O chemisorption and H2-TPR.
calcined
reduced
Catalyst
SBET
Pore volume
Mean pore
size (nm)
Measured H2
SCu
SCu
Cu dispersion
(%)[a]
Cu cluster
(nm)[b]
ꢀ 1
ꢀ 1
ꢀ 1
(m2 gcat
)
(cm3 gꢀ 1
)
consumption (mmol gꢀ 1
)
(m2 gcat
)
(m2 gCu
)
Cu/ZnO-1-DP
Cu/ZnO-1-CH
Cu/ZnO-1-IWI
Cu/ZnO-1-WI
Cu/ZnO-2-WI
CuZnO-2-WI
CuZnO-2-COP
43
31
15
26
16
40
40
0.17
0.16
0.05
0.12
0.06
0.12
0.19
15.7
20.4
12.9
18.9
14.8
12.0
18.6
1.5
1.3
1.5
1.9
1.6
1.5
1.3
4.2
2.8
0.8
2.3
2.4
2.4
3.9
50
34
9
5.4
3.7
1.0
2.5
2.9
3.8
5.1
13.4
19.5
43.4[c]
29.3
24.7
19.0
14.0
23
27
35
48
[a]
calculated from N2O chemisorption using Eq. (6).
calculated from N2O chemisorption using Eq. (5).
[b]
[c]
calculated based on reducible copper content (70 % reducibility).
4

V. Pospelova et al.
Applied Catalysis A, General 624 (2021) 118320
Fig. 3. BET surface area (A) and pore volume (B) dependence on CuO crystallite size.
measurement is presented in Table 2. The measured hydrogen con-
sumption was in accordance with the theoretical hydrogen consumption
calculated based on the Cu content measured by AAS (Table S4). All
catalysts were fully reduced below 250 ◦C, except Cu/ZnO-1-IWI that
was fully reduced only at 350 ◦C.
To determine the active copper surface area and copper dispersion,
the N2O-RFC technique was used (the results are presented in Table 2).
The results validated that the catalyst synthesized by DP had the highest
copper dispersion. Moreover, the calculated Cu cluster size, of 13.4 nm,
was the smallest among the studied catalysts. The results are also in good
correlation with the XRD-estimated CuO size (Fig. 6A). It is clearly seen
in Fig. 6B that the specific Cu surface area was dependent on the syn-
thesis method that resulted in different BET surface area of the catalyst
prior to its reduction. The samples synthesized by different methods
using the same support had a different Cu surface area (linear trend in
Fig. 6B), while the samples synthesized by the same method using a
different support had a similar Cu surface area.
Fig. 4. Pore distribution of catalysts.
To sum up, comparing the catalysts synthesized using the same
commercial ZnO support, the most promising was Cu/ZnO-1-DP as it
had the smallest Cu crystallites (13.4 nm), the highest BET surface area
(43 m2 gcatꢀ 1) and the highest active copper surface area (4.2 m2 gcatꢀ 1).
The catalysts Cu/ZnO-1-CH and Cu/ZnO-1-WI also had comparable
physicochemical properties. Contrarily, during the Cu/ZnO-1-IWI syn-
thesis, the textural structure of the commercial ZnO apparently deteri-
orated: the pore volume decreased, the BET surface area significantly
dropped and the deposited CuO crystallites sintered. Additional cata-
lysts that were prepared using the precipitated Zn5(CO3)2(OH)6, CuZnO-
2-COP and CuZnO-2-WI, had a high BET surface area (both of
40 m2 gcatꢀ 1). Moreover, CuZnO-2-COP had the second highest active
copper surface area (3.9 m2 gcatꢀ 1). The last, Cu/ZnO-2-WI, was among
the worst catalysts based on the characterization probably due to the
copper crystallites sintering and destruction of the initial support
structure as described for the Cu/ZnO-1-IWI.
size. This was corroborated by the results of Cu/ZnO-1-WI (CuO size of
16 nm) having the reduction maximum shifted to the higher tempera-
ture. Generally, the catalysts showed at least two peaks, the first one in
the range of 163ꢀ 175 ◦C and the second one between 188ꢀ 202 ◦C.
Usually, the presence of more than one reduction peak in copper cata-
lysts was assigned to the subsequent reduction of Cu2+ to Cu+ and then
Cu+ to Cu◦ [25]. The range of reduction maxima corresponds with the
previously published results on the co-precipitated CuZnO catalysts,
where peaks typically between 175 and 195 ◦C have been observed [8].
The hydrogen consumption measured by H2-TPR before the RFC-N2O
3.1.2. Spent catalysts
Spent catalysts after reaction were analyzed by AAS to evaluate if
metal leaching occurred and by XRD to determine the phase composi-
tion and to estimate the Cu crystallite size. We assumed that there was
no Cu leaching based on the liquid products color after the experiment.
As for Zn, the change in Cu/Zn ratio compared to calcined catalysts was
monitored (reported in Table 1) and no leaching was confirmed.
The XRD patterns of the spent catalysts are presented in Fig. 7. In
most of the spent catalysts (except CuZnO-2-COP and Cu/ZnO-1-WI) the
metallic copper Cu◦ was determined as the only copper-containing
phase. In case of CuZnO-2-COP and Cu/ZnO-1-WI, the CuO and Cu2O
phase was detected, respectively. However, considering the reduction
temperature of 250 ◦C and the results of H2-TPR, it is clear that copper
was reduced during the in-situ reduction and re-oxidized only during the
catalyst withdrawal from the reactor and the subsequent manipulation
with the sample. The presence of Cu◦ crystallites of the size under 3 nm
Fig. 5. H2-TPR profiles of calcined catalysts.
5

V. Pospelova et al.
Applied Catalysis A, General 624 (2021) 118320
Fig. 6. A - Relationship between Cu cluster size from N2O-RFC and XRD, B - Dependence of SCu from N2O-RFC on BET surface area.
due to the XRD method limitation could not be excluded. The surprising
exclusive presence of Cu◦ in the Cu/ZnO-1-IWI might be explained by its
complete reduction under the experimental conditions. Another phase
that was present in spent catalysts was carbon that might be a result of
carbonaceous deposits or high molecular weight products formed during
the reaction on the catalyst surface. The zinc oxide phase remained the
only crystalline zinc-containing phase in the spent catalysts.
superiority in physicochemical properties even after the experiment
(250 ◦C, 100 bar H2, 180 min). The three catalysts synthesized using the
WI method, had relatively similar Cu crystallites (23ꢀ 26 nm) regardless
of the used ZnO support or precursor. Copper particles sintering during
the experiment occurred to the largest extent in the Cu/ZnO-1-CH
catalyst.
The estimated Cu (2θ = 43.3 ◦) and ZnO (2θ = 31.8 ◦) crystallite size
in comparison with the crystallite sizes in the calcined catalysts is shown
in Fig. 2. As it was mentioned before, in case of COP catalyst, only the
CuO phase was present, and the CuO crystallite size was calculated using
2θ = 38.6 ◦. The smallest crystallites were in COP and DP samples, the
Cu size in the rest of the catalysts was in the range of 23ꢀ 31 nm.
Surprisingly, the largest Cu crystallites were estimated for the Cu/
ZnO-1-CH sample that had very promising properties after the calcina-
tion and reduction. However, during the experiment, the Cu size was
increasing probably due to the sintering of Cu clusters. The size of ZnO
crystallites in the spent catalysts was mostly the same as in their calcined
counterparts, except two catalysts CuZnO-2-COP and CuZnO-2-WI,
where the size of ZnO increased compared to the calcined state. This
fact was in accordance with the preparation procedure as during the
CuZnO-2-COP and CuZnO-2-WI synthesis, zinc was present in form of
hydroxycarbonate, while the other catalysts were synthesized using a
commercial or synthesized ZnO support, i.e. ZnO-1 or ZnO-2 support,
respectively. It might be suggested that when both CuO and ZnO are
formed during the calcination, the contact between both metals is more
intimate and not only ZnO may affect the CuO stability but also vice
versa.
3.2. Catalytic results
The series of the co-precipitated CuZnO and supported Cu/ZnO
catalysts synthesized by seven different procedures was tested in
dimethyl adipate (DMA) hydrogenolysis using autoclave at a reaction
temperature of 250 ◦C and hydrogen pressure of 100 bar. The catalytic
performance was evaluated based on the calculated DMA conversion,
selectivity and yields of individual products.
3.2.1. Conversion
The dependence of the DMA conversion on the reaction time is
shown in Fig. 8. The catalysts might be divided into three groups: the
most active (synthesized by COP and DP), the least active (synthesized
by IWI) and the rest of the catalysts with a mediocre performance. The
conversion was very similar for the catalysts Cu/ZnO-2-WI and CuZnO-
2-WI, so, the difference between the synthesis procedures had no in-
fluence on the catalytic conversion. As discussed in part 3.1, COP- and
DP-synthesized catalysts had the smallest Cu crystallites, while IWI-
synthesized catalyst had the largest ones (Table 2). When plotted as a
function of the Cu crystallite size, the DMA conversion was observed to
decrease with the increasing Cu crystallite size (Fig. 9A). Similar
dependence was already reported for co-precipitated CuZnO catalysts
[9] but it has not been until now confirmed that this trend was valid for
both co-precipitated and supported catalysts independent of the syn-
thesis method. This is also reflected in dependence of the DMA con-
version on the specific copper surface area (Fig. 9B), where the DMA
conversion was increasing with the increasing SCu. The dependence of
the catalyst conversion on the active copper surface area was expected
as copper is the active site in the hydrogenolysis reaction. It might be
concluded that the choice of the synthesis method affected the size of the
Cu crystallites and the copper surface area, and thus, also the final
conversion over the copper-zinc catalysts.
To conclude, the catalysts synthesized by COP and DP showed their
3.2.2. Selectivity
The catalysts are compared by their selectivity to the main product,
hexane-1,6-diol (HDOL), and the reaction intermediate, methyl 6-
hydroxyhexanoate (1HMEol) in Fig. 10. With the increase in DMA
conversion, the selectivity to 1HMEol was decreasing while the selec-
tivity to HDOL was constant in the range of 18–23 % for most of the
catalysts with certain exceptions. One exception was Cu/ZnO-1-IWI,
where no HDOL was observed at all, however, 1HMEol was present
showing that hydrogenolysis of one of the methyl ester groups has
occurred. The reason was most probably the very low SCu of
Fig. 7. XRD patterns of the spent catalysts (▴- ZnO, x – CuO, * - Cu, o - C).
6

V. Pospelova et al.
Applied Catalysis A, General 624 (2021) 118320
over COP-synthesized catalyst the highest 1HMEol yield was achieved
that was probably then involved in the side reactions. Even though
different ZnO supports were used, the three catalysts prepared using the
WI method had a similar Cu crystallite size and afforded a similar yield
of both hydrogenolysis products.
Besides hydrogenolysis, other reactions such as transesterification
[4,26] or cyclization [27,28] might occur. A detailed reaction scheme
previously reported by Aubrecht et al. is shown in Fig. S3 [5]. The results
in Fig. 11 summarize the selectivity to the transesterification products
(TR) having two hexane fragments in their molecule (derived from
either DMA, 1HMEol or HDOL) denoted as “2H” products (2HdiME -
6-methoxy-6-oxohexyl methyl adipate, 2HMEol
- 6-hydroxyhexyl
methyl adipate, 2Hdiol - 6-hydroxyhexyl 6-hydroxyhexanoate). The
catalyst that deviated from the others with respect to HDOL and 1HMEol
selectivity (Cu/ZnO-1-IWI), differed from the other catalysts also in its
selectivity to TR products, represented here by the selectivity to 2H
compounds. In case of IWI-synthesized catalyst, the main product was
2HdiME (Fig. 11B) formed by TR of DMA with 1HMEol. It means that
after one methyl ester group was hydrogenolyzed, the 1HMEol product
underwent exclusively TR with DMA before being hydrogenolyzed.
Surprisingly, there was a trace amount of 2Hdiol (Fig. 11D) detected
over Cu/ZnO-1-IWI, that is formed by the reaction of 1HMEol and
HDOL, which was, however not detected in the reaction mixture at all. It
suggests that hydrogenolysis of 1HMEol occurred to a limited extent, but
HDOL immediately reacted with another 1HMEol molecule resulting in
the 2Hdiol formation.
Fig. 8. DMA conversion dependence on time. T = 250 ◦C, p = 100 bar,
600 RPM.
0.8 m2 gcatꢀ 1 that resulted in a low hydrogenolysis selectivity. Another
exception was Cu/ZnO-1-CH, where the selectivity to HDOL was
significantly decreasing with time and/or conversion. Here, it is
appropriate to remind that this catalyst had a high Cu dispersion (3.7 %)
and SCu of 2.8 m2 gcatꢀ 1, corresponding to relatively small CuO crys-
tallites (10 nm) before the reaction, and unusually large Cu crystallites
(31 nm) after the reaction. Thus, because of copper crystallites sintering,
the selectivity was decreasing as the Cu crystallites size increased.
The product yields of HDOL and 1HMEol were calculated to compare
the catalyst performance regardless of the conversion (Fig. S2). The
highest HDOL yield was achieved over DP-synthesized catalyst, while
Analogously, heavier TR products were also formed during the ex-
periments. It was possible to identify products consisting of 3 and 4
hexane backbones by GC–MS, however, the presence of even heavier
unidentified compounds with up to 7 hexane backbones was confirmed
by LC–MS. The accumulation of these high-molecular-weight com-
pounds might explain the carbon deposits present in the spent catalysts
Fig. 9. DMA conversion dependence on Cu crystallite size (A) and copper surface area (B). T = 250 ◦C, p = 100 bar, t = 180 min, 600 RPM.
Fig. 10. Selectivity to 1HMEol (A) and HDOL (B) in dependence on DMA conversion. HDOL – hexane-1,6-diol, 1HMEol - methyl 6-hydroxyhexanoate.
7

V. Pospelova et al.
Applied Catalysis A, General 624 (2021) 118320
Fig. 11. Selectivity to 2H (formed by 2 hexane backbones) transesterification products in dependence on DMA conversion. 2HdiME - 6-methoxy-6-oxohexyl methyl
adipate (B), 2HMEol - 6-hydroxyhexyl methyl adipate (C), 2Hdiol - 6-hydroxyhexyl 6-hydroxyhexanoate (D).
that were visible by XRD. The commented selectivity to such 3H and 4H
compounds is reported in Fig. S4. The selectivity trends to 3H com-
pounds were similar to those observed for 2H compounds discussed
above. Besides the above-mentioned products, the selectivity to cyclo-
pentanol and ε-caprolactone is presented and commented in detail in the
Supporting Information (Figs. S5 and S6, respectively).
3.2.3. Intrinsic activity
The DMA conversion and selectivity to different products might be
taken as the catalyst performance parameters, however, to evaluate the
copper intrinsic activity, the turnover frequency (TOF) should be
calculated. Based on the experimental catalytic data and the specific
copper active surface area data from the N2O chemisorption, the number
of the converted DMA moles and the number of the active copper atoms
was calculated and used in Eq. (4) to express the TOFH value. TOFH was
only related to DMA moles converted by hydrogenolysis. As shown in
Fig. 12, the most promising hydrogenolysis catalysts were catalysts
synthesized by COP, WI and DP. However, when WI was applied for the
copper impregnation on the Zn5(CO3)2(OH)6 and the synthesized ZnO to
prepare CuZnO-2-WI and Cu/ZnO-2-WI, respectively, the TOFH was not
as high as in case of ZnO-1. The lowest Cu intrinsic hydrogenolysis ac-
tivity was found in Cu/ZnO-1-IWI and in Cu/ZnO-1-CH. The low
intrinsic activity of Cu/ZnO-1-IWI is surely a result of suppressed
hydrogenolysis activity due to extremely low Cu surface area
(0.8 m2 gcatꢀ 1). Even though the Cu surface area of reduced Cu/ZnO-1-
CH was mediocre, it had the largest Cu crystallites after the experiment
showing that the Cu surface area could decrease due to the copper sin-
tering. It might be concluded that except these two catalysts, Cu/ZnO-1-
IWI and Cu/ZnO-1-CH, the variation of TOF for the rest of the catalysts
Fig. 12. TOFH results.
4. Conclusion
To the best of our knowledge, this is the first study describing the
effect of the synthesis method of copper-zinc catalysts on their proper-
ties and activity in dimethyl adipate hydrogenolysis. To accomplish this
goal, a CuZnO catalyst (8 wt% of Cu) synthesized by co-precipitation
(COP) was compared with the supported Cu/ZnO catalysts (8 wt% of
Cu) synthesized by four different methods using a commercial ZnO
support (ZnO-1): incipient wetness impregnation (IWI), wet impregna-
tion (WI), deposition-precipitation (DP) and chemisorption-hydrolysis
(CH).
was not significant being in the range of 2.2–2.6 ∙10-2 sꢀ 1
.
The Cu/ZnO-1-IWI had the smallest active Cu surface area, BET
8

V. Pospelova et al.
Applied Catalysis A, General 624 (2021) 118320
surface area and the largest Cu crystallites that resulted in the lowest
hydrogenolysis activity. Despite the relatively small CuO crystallites
(10 nm) before the reduction and high Cu surface area (2.8 m2 gcatꢀ 1) of
reduced catalyst, Cu/ZnO-1-CH also exhibited a low hydrogenolysis
activity explained by the sintering of Cu crystallites during the experi-
ment, plausibly due to a lower stability of Cu clusters prepared by CH in
comparison with the other methods. The rest of catalysts synthesized by
WI, DP and COP had similar TOF in the range of 2.2–2.6 ∙10-2 sꢀ 1. When
WI was applied for the copper impregnation on the Zn5(CO3)2(OH)6 and
synthesized ZnO to prepare CuZnO-2-WI and Cu/ZnO-2-WI, respec-
tively, the TOFH was not as high as when using the commercial ZnO
support. Cu/ZnO-1-DP and CuZnO-2-COP had the smallest Cu crystal-
lites (13ꢀ 14 nm), the highest BET surface area (40ꢀ 43 m2 gcatꢀ 1) and
highest active copper surface area (3.9-4.2 m2 gcatꢀ 1). The future opti-
mization of the Cu/ZnO catalysts synthesized by COP, DP and WI using
promoters would be aimed to the formation of small and stable Cu
crystallites with high hydrogenolysis activity.
Appendix A. Supplementary data
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.apcata.2021.118320.
References
[1] J. Pritchard, G.A. Filonenko, R. van Putten, E.J.M. Hensen, E.A. Pidko, Chem. Soc.
Rev. 44 (2015) 3808–3833.
[2] H. Adkins, K. Folkers, J. Am. Chem. Soc. 53 (1931) 1095–1097.
[3] T. Turek, D.L. Trimm, N.W. Cant, Catal. Rev. 36 (1994) 645–683.
ˇ
ˇ
´ˇ
´ˇ
[4] D. Kubicka, J. Aubrecht, V. Pospelova, J. Tomasek, P. Simacek, O. Kikhtyanin,
Catalan J. Commun. 111 (2018) 16–20.
´
ˇ
[5] J. Aubrecht, V. Pospelova, O. Kikhtyanin, L. Dubnova, D. Kubicka, Appl. Catal. A
608 (2020), 117889.
[6] P. Yuan, Z. Liu, W. Zhang, H. Sun, S. Liu, Chin. J. Catal. 31 (2010) 769–775.
[7] F.T. van de Scheur, L.H. Staal, Appl. Catal. A 108 (1994) 63–83.
´
ˇ
[8] V. Pospelova, J. Aubrecht, O. Kikhtyanin, K. Pacultova, D. Kubicka, ChemCatChem
11 (2019) 2169–2178.
ˇ
[9] O. Kikhtyanin, V. Pospelova, J. Aubrecht, M. Lhotka, D. Kubicka, Catalysts 8
(2018) 446.
´
ˇ
[10] V. Pospelova, J. Aubrecht, K. Pacultova, M. Lhotka, O. Kikhtyanin, D. Kubicka,
CRediT authorship contribution statement
Catal. Sci. Technol. 10 (2020) 3303–3314.
[11] P. Munnik, P.E. de Jongh, K.P. de Jong, Chem. Rev. 115 (2015) 6687–6718.
[12] A. Gervasini, M. Manzoli, G. Martra, A. Ponti, N. Ravasio, L. Sordelli, F. Zaccheria,
J. Phys. Chem. B 110 (2006) 7851–7861.
Violetta Pospelova: Investigation, Visualization, Writing - original
draft. Jaroslav Aubrecht: Writing - review & editing. Oleg Kikhtyanin:
[13] R.-P. Ye, L. Lin, Q. Li, Z. Zhou, T. Wang, C.K. Russell, H. Adidharma, Z. Xu, Y.-
G. Yao, M. Fan, Catal. Sci. Technol. 8 (2018) 3428–3449.
[14] L.-F. Chen, P.-J. Guo, L.-J. Zhu, M.-H. Qiao, W. Shen, H.-L. Xu, K.-N. Fan, Appl.
Catal. A 356 (2009) 129–136.
ˇ
Investigation, Writing - review & editing. David Kubicka: Conceptu-
alization, Writing - review & editing, Funding acquisition.
[15] S. Panyad, S. Jongpatiwut, T. Sreethawong, T. Rirksomboon, S. Osuwan, Catal.
Today 174 (2011) 59–64.
Declaration of Competing Interest
[16] R. Fan, M. Kyodo, L. Tan, X. Peng, G. Yang, Y. Yoneyama, R. Yang, Q. Zhang,
N. Tsubaki, Fuel Process. Technol. 167 (2017) 69–77.
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
[17] F. Boccuzzi, A. Chiorino, G. Martra, M. Gargano, N. Ravasio, B. Carrozzini, J. Catal.
165 (1997) 129–139.
[18] Y. Okamoto, K. Fukino, T. Imanaka, S. Teranishi, J. Phys. Chem. 87 (1983)
3747–3754.
[19] S. Brunauer, P.H. Emmett, E. Teller, J. Am. Chem. Soc. 60 (1938) 309–319.
[20] E.P. Barrett, L.G. Joyner, P.P. Halenda, J. Am. Chem. Soc. 73 (1951) 373–380.
[21] N. Ravasio, F. Zaccheria, M. Gargano, S. Recchia, A. Fusi, N. Poli, R. Psaro, Appl.
Catal. A 233 (2002) 1–6.
Acknowledgements
Financial support from the Czech Science Foundation (project No.
GA20-28093S) is greatly acknowledged. N2O-RFC characterization re-
sults were obtained by using the Large Research Infrastructure ENRE-
GAT supported by the Ministry of Education, Youth and Sports of the
Czech Republic (project No. LM2018098). The authors are grateful to
[22] A.F. Trasarti, D.J. Segobia, C.R. Apesteguía, F. Santoro, F. Zaccheria, N. Ravasio,
J. Am. Oil Chem. Soc. 89 (2012) 2245–2252.
[23] F.-W. Chang, W.-Y. Kuo, K.-C. Lee, Appl. Catal. A 246 (2003) 253–264.
[24] E. Marceau, X. Carrier, M. Che, Impregnation and Drying, Synthesis of Solid
Catalysts, 2009, pp. 59–82.
[25] M. Balaraju, V. Rekha, P.S. Sai Prasad, R.B.N. Prasad, N. Lingaiah, Catal. Lett. 126
(2008) 119–124.
´
Jana Cibulkova (UCT Prague, Czech Republic) for XRD analyses, Dana
ˇ
[26] O. Kikhtyanin, J. Aubrecht, V. Pospelova, D. Kubicka, Appl. Catal. A 606 (2020),
´
Pokorna (UCT Prague, Czech Republic) for AAS analyses, Miloslav
117825.
Lhotka (UCT Prague, Czech Republic) for N2 physisorption and H2-TPR
[27] P. Yuan, Z. Liu, T. Hu, H. Sun, S. Liu, React. Kinet. Mech. Catal. 100 (2010)
´
ˇ
´
measurements and Dagmar Fridrichova and Katerina Pacultova for N2O
427–439.
[28] S.M. dos Santos, A.M. Silva, E. Jordao, M.A. Fraga, Catal. Today 107-08 (2005)
chemisorption measurements (VSB-TUO, Czech Republic).
250–257.
9