Physico-chemical and catalytic properties of Zr- and Cu–Zr ion-exchanged H-MCM-41

H-MCM-41 catalyst ion-exchanged by Zr and/or Cu-cations were synthesized and characterized by XPS, 27Al- and 29Si-MAS-NMR, FTIR of pyridine adsorption, laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) and direct current plasma atomic emission spectrometry.

The catalysts were evaluated towards 1-butene skeletal isomerization after pre-treatment with hydrogen or synthetic air.

The dispersion of metals after ion-exchange was reasonably good, as revealed by LA-ICP-MS.

Binding energies observed by XPS indicated that after reduction or oxidation pre-treatments copper is mainly present as Cu2+, except in case of reduced Cu-H-MCM-41.

Pre-treatment had no influence on the oxidation state of Zr4+ species.

Zr ion-exchange decreased the concentrations of Brønsted and Lewis acid sites.

As a result, the initial conversion of 1-butene was decreased, but the selectivity and stability of the catalyst were increased.

As reported before, the copper introduction decreased the Brønsted acid site density, but increased the conversion.

The enhancement in the catalytic activity has been connected to the presence of copper with oxidation state +1, which is formed under reaction conditions due to the hydrogen formation.

This led to a peculiar effect over oxidized bimetallic Cu–Zr-H-MCM-41 at a low temperature and high weight hourly space velocity (523 K and 15 h−1, respectively): The initial conversion of 1-butene was almost zero but increased during the time-on-stream reaching the value of about 13 mol%.

These observations are related to the reduction of Cu2+ species to Cu+ during the reaction.

The catalyst evaluation and characterization results led to a conclusion that the presence of Zr stabilizes copper’s oxidation state of Cu+2 under reductive conditions in the bimetallic Cu–Zr-H-MCM-41 catalyst.


The catalytic activity in the hydrocarbons transformations can be enhanced by a metal introduction, which usually gives beneficial bifunctional character for the catalyst.

For example, introduction of Pt increases the activity and the stability of the acidic catalyst in the isomerization of alkanes1–3 and metal alloys have been used in the transformation of hydrocarbons.4–5

In the skeletal isomerization of linear butenes (n-butene) to isobutene, the introduction of metal has not been extensively studied, most probably because the catalysts, typically microporous molecular sieves such as ferrierite, are intrinsically rather selective and stable.6

However, some beneficial enhancements in catalytic activity due to metal introduction have been observed.

The ion-exchange of Mg into ZSM-22 has been reported to increase the yield of isobutene and stability in time-on-stream in 1-butene skeletal isomerization.7

The ion-exchange increased the ratio between Lewis and Brønsted acid sites and this was inferred to be responsible for improvement in catalytic activity and suppression of coke formation.

Similar observations have been made in case of Mg, Sr and Ba ion-exchanged into ferrierite.8

On the other hand, ion-exchange with Ca, Cu and Mn into ZSM-22 considerably decreased the yield of isobutene.7

Impregnation of Zn into SAPO-11 was reported to diminish the number of the acid sites and conversion of 1-butene, while increase in the selectivity to isobutene was only minor.9

Mn, Co and Zn ion-exchanged or impregnated on SAPO-11 mainly decreased the selectivity to isobutene and conversion of 1-butene.10

A positive effect of Ba and Co introduction into clinoptilolite on the conversion and the selectivity, respectively, has been reported.11

In all of these cases, rather bulky metal cation is located in the narrow pores of the catalysts and therefore modifying the textural properties and decreasing the space around the cation.

The changes in the free space inside the pores may have influence on the prevailing reaction mechanism, since skeletal isomerization of n-butene is a shape selective reaction and even the pore size has an effect on the selectiviy.6

In fact, the presence of Zr in the pores of SAPO-11 has been reported to induce partial pore blockage and decrease the amount of external acid sites, which led to increased selectivity to isobutene.12

In addition, zirconium containing MCM-41 type silica prepared by supramolecular templating was found to be more active in 1-butene skeletal isomerization than the aluminum-doped reference,13 although in this case zirconium was incorporated in the framework.

Previously we have reported that ion-exchange of copper into H-MCM-41 increases 1-butene conversion and yield of isobutene, although the concentration of the Brønsted acid sites after ion-exchange was substantially reduced.14,15

This is peculiar, since typically higher conversion levels are obtained with higher Brønsted acid site densities.16

The improved activity was related to the presence of Cu+ species and the other copper species, i.e. Cu2+ and Cu0, were found to be inactive in the conversion of 1-butene.

The amount of the active Cu+ species, however, changed during the catalytic experiments, since considerable amount of hydrogen is formed due to coke formation.

It would be therefore beneficial for the reaction, if copper remained as Cu+ without reduction to Cu0.

In case of Pd-Zr-Y, the O2-TPD measurements demonstrated that the presence of Zr stabilizes the Pd2+,17 indicating that Zr may have a similar effect on other metal cations as well.

Furthermore, addition of Zr has been reported to stabilize the high dispersion of metals such as Ni, Pd and Pt.18

MCM-41 is an attractive material to study the effect of metal introduction on the hydrocarbons transformation reactions, as its intrinsic catalytic performance is rather moderate due to mild acidity and thus the effect of the metal introduction can be easily observed.

Moreover, the pores of MCM-41 are large enough not to be blocked if a large metal is introduced.

Hence, the impact of the metal as such can be recognized without any influence induced by the restriction changes in the channel structure.

The aim of this study was to investigate the effect of Zr introduction into Cu-H-MCM-41 and H-MCM-41.

Zr and/or Cu ion-exchanged H-MCM-41 catalysts have been prepared, characterized and evaluated as catalysts towards 1-butene skeletal isomerization.


Preparation of the catalysts

The synthesis of H-MCM-41 and Cu-H-MCM-41 has been described in detail elsewhere.14

Only the pertinent details are provided below.

Zr-H-MCM-41 was prepared by ion-exchange of H-MCM-41 using aqueous solution of zirconium tetrachloride at room temperature for 24 h.

After ion-exchange the Zr-H-MCM-41 was dried at 383 K for 12 h and calcined at 723 K for three hours in a muffle oven.

Cu–Zr-H-MCM-41 was prepared by ion-exchange of Zr-H-MCM-41 using aqueous solution of copper nitrate, followed by drying at 383 K for 12 h and calcination at 773 K for four hours.

The X-ray powder diffraction pattern proved that the introduction of Cu in MCM-41 did not change the parent structure of the mesoporous molecular sieve material.


Perkin-Elmer PHI 5400 ESCA with monochromatized Al Kα X-ray source with 90° angle with the analyzer was used for the XPS characterization of the Cu and Zr containing catalysts.

Si 2p line at 103.0 eV was used to correct the binding energy (BE) axis for sample charging upon X-ray bombardment.

Carbon 1s line was then found at (284.6 ± 0.3) eV in all measurements.

Catalyst samples were in the form of dry powder being mounted on two-sided tape for the XPS analysis.

In the photoelectron line fitting process intensity ratios were kept fixed at their theoretical values, i.e. 0.5 and 0.6667 for p and d shells, respectively.

No other impurities except of carbon were detected in the XPS analysis.

In quantitative XPS analysis, sensitivity factors were 4.798, 0.283, 0.770 and 2.216 for Cu 2p, Si 2p, Cl 2p and Zr 3d, respectively.

The catalysts were pretreated at 773 K for 2 h with synthetic air to oxidize or with hydrogen to reduce the metal particles.

Reduced catalysts were then stored in nitrogen atmosphere during sample transfers.

Catalysts without any pretreatment were also analyzed in order to detect the effect of both pretreatment methods.

The FTIR spectrometry of adsorbed pyridine was used to investigate the acidic properties of the synthesized catalysts.

The FTIR spectrometer (ATI Matson infinity spectrometer) was equipped with an in situ cell comprising ZnSe windows.

The samples were pressed into self-supporting discs (weight app.

20 mg and radius 0.65 cm), activated in vacuo at 723 K for 1 h prior to pyridine adsorption at 373 K. The physisorbed pyridine was desorbed in vacuo at 473 for 1 h before the FTIR spectra were recorded at 373 K. The effect of reduction was investigated by preparing a pellet of the Zr-H-MCM-41 catalyst after 2 h pre-treatment at 773 K under flow of hydrogen.

The Brønsted and Lewis acid sites can be distinguished by the bands of the chemisorbed pyridinium ion at 1545 cm−1 and coordinatively bonded pyridine at 1455 cm−1, respectively.19,20

The band at 1490 cm−1 is associated with pyridine adsorbed on both Brønsted and Lewis acid sites.

The peaks at 1596 and 1445 cm−1 are due to hydrogen bonded pyridine and the peaks at 1623 and 1575 cm−1 are associated with strong Lewis bond pyridine.21

The adsorption coefficients for pyridine adsorption on Brønsted and Lewis acid sites to calculate the molar amounts of corresponding concentrations, are taken from the literature.19

27Al MAS NMR spectra of the H-MCM-41 sample were measured on a Bruker AMX500 spectrometer in a 11.7 T magnetic field using a home-made probe head with 3.5 mm od rotors.

The sample spinning speed of 13 kHz was used.

KAl(SO4)2 was used as an external reference for 27Al chemical shift and intensity calibration. 29Si MAS NMR spectra were recorded on a Bruker AMX360 spectrometer in 8.5 T magnetic field using MAS probe head with 7 mm od rotors and sample spinning speed of about 5 kHz.

The chemical shift was scaled by external reference of [(CH3)3Si]2O, (δSi = 6.53 ppm).

Laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) was used to analyze the proportional amounts of copper and zirconium in the samples as well as distribution of the metal cations and aluminum.

Laser ablation system (Cetac Technologies Inc. LSX-200+) was connected to inductively coupled plasma mass spectrometer (Perkin Elmer SILEX, Elan 6100 DRC Plus).

Samples were pressed to pellets.

Continuous and representative sample transfer to the ICP-MS was done by scanning the laser beam over the sample for 2 mm line to avoid difficulties due to inhomogeneities of the sample in the final values.

The time to produce one line, approximately 2 mm, was 164 s and 80 replicates were performed for each measurement.

A 2 mm long line was scanned three to six times, after which another line was chosen on the sample surface.

The final values were calculated from the average of the net intensity means of about 10 scans.

The net intensity means of the detected elements were divided by the net intensity means of alumina in order to obtain comparable values and no standards were used.

Therefore, the obtained results are quantitatively comparable only to each other.

Energy level used was varied from 50 to 85% depending on the sample brightness.

Pulse repetition frequency was 10 Hz, spot diameter 50 μm and the measuring integration time on each elements was 400 ms/replicant.

The copper contents were determined quantitatively by direct current plasma (DCP) atomic emission spectrometry analysis (ARL SectraSpan 7).

The zirconium content of Zr-H-MCM-41 was evaluated by SEM-EDXA (SEM: LEO 1530, EDXA: Thermo Novan Vantage).

The surface areas and cumulative pore volumes were determined by nitrogen adsorption (Sorptometer 1900, Carlo Erba Intruments) using the BET method.

Catalyst evaluation

Skeletal isomerization of 1-butene to isobutene was investigated over H-, Zr-H-, Cu-H- and Cu–Zr-H-MCM-41 catalysts in a fixed-bed reactor at near atmospheric pressure.

The catalysts were pretreated with synthetic air or hydrogen at 773 K for 2 h before the experiments.

The product analyses were carried out on-line using a gas-chromatograph (Varian 3700) equipped with a flame-ionization detector (FID) and capillary column (50 m × 0.32 mm ID fused silica PLOT Al2O3 – KCl).

The reactant was diluted with nitrogen to obtain a partial pressure of 0.5 atm.

The conversion of 1-butene and the selectivity to isobutene were investigated as a function of time-on-stream (TOS) at the weight hourly space velocity (WHSV) of 5 h−1 and temperature 673 K as well as at WHSV of 15 h−1 at 623 K.

Since the double bond isomerization of 1-butene is much faster than the skeletal isomerization, the three linear butene isomers 1-butene, cis-2-butene and trans-2-butene were considered as reactants in the calculations.

Thus, the conversion of 1-butene and the selectivity to isobutene are defined as follows (Qm is quantity in moles):

Results and discussion


Metal contents

The amount of copper in both copper containing samples was 2.3 mass% measured by DCP.

Zr contents were approximated to be 1 mass% in Zr-H-MCM-41 and 0.8 mass% in Cu–Zr-H-MCM-41.

The former value is based on the calculated maximum amount of Zr present in the ion-exchange process (0.13 g of ZrCl4 and 5 g of H-MCM-41) and the latter one is derived from the LA-ICP-MS and XPS results having former one as a reference.

One should note that a full ion-exchange is expected and therefore the real quantities might be slightly lower.

The zirconium content of 1.1 mass% on the surface of Zr-H-MCM-41 was obtained by SEM-EDXA.


Due to the very high sensitivity of LA-ICP-MS, it is a very efficient method to elucidate the distribution of introduced metals on the catalyst.

With proper standards, it can be used even in quantitative manner, and indeed it has been applied in the determination of the trace elements in zeolites.22

Here the method has been used in a semi-quantitative way, because no suitable standards were available.

Therefore, the obtained values are comparable only to each other.

While scanning several times the catalyst surface, an overview of the bulk metal distribution over the catalyst can be obtained.

Examples of the LA-ICP-MS spectra are given in Fig. 1.

In general, the metals are distributed rather evenly indicating a good dispersion.

As can be seen in Fig. 1, in case of Cu-H-MCM-41, the intensity of copper is following the Al-content suggesting that copper species are balancing the framework charge and thus, ion-exchanged into H-MCM-41.

Similarly, Zr-content is constant in Zr-H-MCM-41 indicating a good dispersion.

Some minor heterogeneity of the aluminum distribution in the samples was visible.

Intensity fluctuations were also observed in copper as well as zirconium contents and few intensive peaks in the sodium intensity were observed indicating that traces of sodium are still present in the samples.

The proportional amounts of copper for Cu-H-MCM-41 and Cu–Zr-H-MCM-41 of 0.56 and 0.59, respectively, were obtained and similarly zirconium contents of 1.92 and 1.54 for Zr-H-MCM-41 and Cu–Zr-H-MCM-41, respectively.

The amount of copper in Cu-H-MCM-41 and Cu–Zr-H-MCM-41 is approximately the same, which is in line with the results obtained by XPS and DCP (see below).

Cu–Zr-H-MCM-41 contains about 20% less zirconium than Zr-H-MCM-41.

This is reasonable, since the former is a bimetallic catalyst being ion-exchanged twice and moreover, the XPS results indicated a similar trend (Table 1).


The results of the XPS analysis are given in Table 2.

Previously it has been reported that the copper species are mainly present as Cu2+ after oxidation and Cu0 after reduction of Cu-H-MCM-.4115

The Cu 2p3/2 binding energies were between 933.1–933 eV for all copper containing samples, except 932.0 eV for reduced Cu-H-MCM-41-Red.

Contarini and Kevan have reported binding energies ∼935–936 eV for octahedrally coordinated Cu2+, ∼933–934 eV for tetrahedrally coordinated Cu2+ and 932.0 eV without shake-up satellites for Cu+ or Cu0 in copper ion-exchanged X- and Y-type sodium zeolites.23

Thus, earlier characterization results together with the observed binding energies suggest that reduced Cu-H-MCM-41-Red had copper in form of Cu0 (or Cu+) but copper was present as Cu2+ in all other copper containing samples.

Therefore, reduced Cu–Zr-H-MCM-41-Red had also copper present as Cu2+.

However, extensive copper 2p satellite structures ominous to Cu2+ were not visible in the spectra of copper containing samples.

Although absence of the extensive copper 2p satellites is typically24 associated with the presence of copper as Cu+ or Cu0, in the present case this can be associated to long (3–5 h) exposure to X-irradiation, necessary to collect a sufficient amount of photoelectron pulses due to rather low copper concentration.

Long exposure to X-irradiation has been reported23,25 to significantly decrease the intensity for the satellites for Cu2+.

This interpretation is supported by the earlier characterization results of Cu-H-MCM-41-Ox,15 in which copper was mostly as Cu2+ but no satellites were observed in the XPS spectrum.

No CuO satellites were observed in the spectra as reported previously for copper-oxide-containing zirconium-doped mesoporous silica catalyst,26 which indicates that no CuO is present.

This is line with the previous observations with Cu-H-MCM-41-Ox.14

Zr 3d5/2 binding energies varied between 182.7 and 183.0 corresponding to zirconium present as Zr4+.

Thus, pre-treatments had no effect on the oxidations state of zirconium.

The binding energies demonstrate a clear difference in the oxidation states of copper in Cu-H-MCM-41-Red compared to other copper containing samples and it could be concluded that only Cu-H-MCM-41-Red contains substantial amounts of copper as Cu0 and Cu+.

Nevertheless, due to the absence of the extensive copper 2p satellites, the presence of some copper as Cu+ or Cu0 cannot be excluded in other Cu-containing samples.

The change of metal dispersion can be inferred from the change of relative intensity, because the mean-free path λ for ineleastic scattering of Cu 2p3/2 photoelectrons is ca. 1.4 nm.27

Since 95% of photoelectrons originate from within 3λ distance of the surface, the Cu 2p photoelectron flux becomes reduced by the particle growth.28

Comparison of Cu/Si atomic concentration ratios indicates significant growth of copper clusters in Cu-H-MCM-41 upon oxidation and especially reduction, whereas the Cu–Zr-MCM-41 shows much less sintering upon reduction and further improved dispersion upon oxidation treatment.

Improved dispersion of Pt on ZSM-5 by Zr introduction has been reported previously.18

After oxidation treatment of Cu–Zr-H-MCM-41, 20% of surface Zr was lost while copper dispersion was improved.

Pre-treatments had only a minor effect on the dispersion of Zr-H-MCM-41.

The amount of Cl left in the samples after ion-exchange and calcination is negligible, since the concentration of Cl was below the detection limit.

To summarize, zirconium modification on Cu-H-MCM-41 provides a better environment for Cu2+ cations to retain their oxidation state under reductive conditions, which is further confirmed by catalytic studies.

A reasonable explanation for this would be that zirconium cations balancing the framework charge (and hypothetically replacing copper cations compared to Cu-H-MCM-41) consequently increase the average distance between two adjacent copper cations.

As a consequence, the longer distance hinders the sintering of copper species.

Note that based on the Zr binding energies neither interactions nor charge transfer between Zr and Cu metals in Cu–Zr-H-MCM-41 were observed, which supports this conclusion.

This hypothesis assumes high dispersion of metals, which is expected by ion-exchange method and confirmed by LA-ICP-MS studies.

FTIR spectra of adsorbed pyridine

The FTIR spectra of adsorbed pyridine on Zr- and Cu–Zr-H-MCM-41 are given in Fig. 2 and the calculated acid site densities in Table 1.

Corresponding spectra of H-MCM-41 and Cu-H-MCM-41 from ref. 15 are presented for comparison.

After the ion-exchange about one third of the Brønsted acid sites (BAS) were left due to substitution of protons by Zr4+ and/or Cu2+ cations.

In case of Zr-H-MCM-41 the BAS density was slightly increased after treatment with hydrogen.

The increased number of BAS in accordance with the increased conversion level, as observed by catalytic evaluations studies presented below.

This is somewhat surprising, since based on the XPS results, Zr was not even partially reduced.

The copper containing samples showed a strong band at 1452 cm−1 overlapping and masking the band at 1455 cm−1, which is associated with strong Lewis bonded pyridine.

The band at 1452 cm−1 is typically assigned to pyridine adsorbed on copper.29

The integrated band areas at that region indicated that the presence of copper increased prominently the concentration of the Lewis acid sites (LAS), presuming that the molar extinction coefficients for both bands are approximately the same.

In addition to the Lewis bonded pyridine peak at 1454 cm−1 an additional adsorption band appears at 1445 cm−1 after Zr ion-exchange.

The latter one is probably associated with hydrogen bonded pyridine21 being absent in the H-MCM-41 and Cu-containing samples, possibly masked by the strong band at 1454 or 1452 cm−1.

Another possibility is that the band at 1445 cm−1 is due to pyridine bonded to zirconium cation.

The integrated area of the double peak at 1445 and 1454 cm−1 indicates that the concentration of the Lewis acid sites is decreased after Zr ion-exchange, supposing no major differences in the extinction coefficients due to ion-exchange (Table 1).

Based on the spectrum of the reduced Zr-H-MCM-41-Red, such a pretreatment slightly increases not only the BAS but also the LAS density.

The peak intensity at 1639 cm−1 attributed to a pyridinium-ion, i.e. pyridine bounded to a Brønsted acid site, is decreased in all ion-exchanged samples.

A peak at 1623 cm−1 ascribed as strong Lewis bonded pyridine is observed in all samples.

The ion-exchanged samples have a strong peak at about 1610 cm−1 being absent in the spectrum of H-MCM-41.

This band is due to pyridine adsorbed on the electron acceptor sites (Lewis sites or preferably cations in this case) being the vibration denoted as 8a CC(N) mode.

It has been reported that physisorbed pyridine on H-Y zeolite, held on the surface by weak hydrogen-bonding interactions, shows a characteristic band at 1614 cm−1.30

Furthermore, a peak at 1608 cm−1 has been ascribed to Lewis acidity over saponites.31

A peak at 1608 cm−1 has been observed in oxidized Cu-Y zeolite samples but was removed after reduction, indicating that it is due to pyridine adsorbed on Cu-cation.32

Accordingly, in case of Zr-H-MCM-41 it is reasonable to assume that the peak at 1610 cm−1 is due to same vibration mode, i.e. pyridine adsorbed on zirconium cations.

In summary, ion-exchange decreased the BAS concentration by about two thirds and even more in the case of bimetallic Cu–Zr-H-MCM-41.

Lewis acid site density was decreased after Zr ion-exchange, and increased if copper was present.

Reduction of Zr-H-MCM-41 slightly increased the amount of both acid sites

29Si and 27Al MAS NMR of H-MCM-41

The 29Si and 27Al MAS NMR spectra of H-MCM-41 are given in Fig. 3.

The spectrum of 29Si shows several broad overlapping lines.

A computer fit to the spectrum gives the lines at −108.5 ppm (65% of the total intensity), −100.6 ppm (13%), −94.8 ppm (19%) and −80 ppm (3%).

According to Kolodziejski et al.,33 the strongest peak at −108.5 ppm can be attributed to Si(OSi)4 sites, while the components at −100.6, and −94.8 ppm are due to the sites with silanol groups Si(OSi)3OH and Si(OSi)2(OH)2, respectively.

Since the sample has considerable amount of Al in the framework, the contribution to the signal from the sites Si(OSi)3(OAl), Si(OSi)2(OAl)2, Si(OSi)(OAl)3, and Si(OAl)4 in the region between −80 ppm and −100 ppm must be expected.34

The 27Al MAS spectrum of H-MCM-41 shows a signal with a maximum at 54 ppm due to tetrahedrally coordinated framework aluminum and a broad asymmetric line around 0 ppm due to octahedrally coordinated aluminum sites in accordance with the previous studies.34,35

A small component arising from the computer fit was attributed tentatively to the distorted tetrahedral or five-coordinated Al sites.

The quantitative analysis of the spectrum yield the absolute number of Al sites as follows (in mmol g−1): Framework tetrahedral Al 0.98, non-framework tetrahedral or five-coordinated Al 0.1 and octahedral Al 1.16 giving a total Al content of 2.24 mmol g−1.

The Si/Al ratio can be estimated from 27Al line intensity.

If a general chemical composition HAlSirO2(r + 1) is assumed to be valid (r = Si/Al ratio), and Al content is n = 2.24 mmol g−1 then one can calculate r = 1000/(60n) − 1 = 6.4.

This value differs slightly from the earlier reported value 2.6 obtained for Na-MCM-41 form by the X-Ray fluoresency.14

A probable reason for the difference between these values is that all Al was not detected by the 27Al line intensity, which can be lost as a result of very strong quadropolar line broadening leading to underestimation of the non-framework Al (present as e.g. polymeric aluminum oxides or oxide hydrates) content by 27Al NMR.

Surface areas and pore volumes

The specific surface areas and cumulative pore volumes of the studied moleuclar sieves are given in Table 3.

The surface areas and pore volumes are decreased after ion-exchange compared to Na-MCM-41.

The most pronounced decrease is in case of Cu–Zr-H-MCM-41, which has been ion-exchanged twice.

Catalytic evaluation

The catalysts were evaluated towards 1-butene skeletal isomerization at 623 K and WHSV of 15 h−1 as well as at 673 K and 5 h−1.

Metal containing catalysts were oxidized with synthetic air (designated as -Ox) or reduced with hydrogen (designated as -Red) at 773 K for 2 h before the measurement.

The results are given in Figs. 4 and 5.

Selectivity to isobutene as a function of conversion at 673 K is plotted in Fig. 6.

One should note that H-MCM-41 is not an optimal catalyst for 1-butene isomerization, such as ferrierite or SAPO-11.

This is mostly due to mild acidity and large pores of the mesoporous material.

Nevertheless, 1-butene isomerization can be used as a special characterization method, since the conversion depends on the number of the acid sites and in this case, a state of copper.

Effect of Cu ion-exchange

A comprehensive discussion of the effect of Cu ion-exchange into H-MCM-41 is given in .ref. 15

It was concluded that the presence of Cu+ enhances the conversion of 1-butene being reduced from Cu2+ during the reaction.

This is also clearly seen here at somewhat different conditions, at 673 K and the WHSV of 5 h−1 over Cu-H-MCM-41-Ox, which exhibits higher conversion than H-MCM-41.

Although the copper species are mainly present as Cu2+ in the beginning of the run, they are rapidly reduced to Cu+ before the first measured point at TOS = 10 min and conversion is consequently increased.

However, over Cu-H-MCM-41-Red the conversion and selectivity are comparable to H-MCM-41.

This is due to complete reduction of copper species to metallic Cu0 and consequent formation of an equivalent amount of the Brønsted acid sites (BAS) as in H-MCM-41.

At 623 K, an increase in conversion was observed over Cu-H-MCM-41-Ox during time on stream.15

This can be related to the reduction of Cu2+ species to active oxidation state of Cu+ during the hydrocarbon transformation, which mainly takes place after TOS = 10 min.15

It is notable that coke is formed during the reaction deactivating the catalyst,14,15 and the reasons for the catalyst deactivation are coke formation together with the reduction of copper to Cu0.

Effect of Zr ion-exchange

Zr ion-exchange in general increases the catalysts stability but at the same time, the conversion of 1-butene is decreased (Figs. 4 and 5).

It is reasonable to assume that the latter one is related to the decreased amount of BAS after Zr ion-exchange (Table 1).

This is supported by the fact that the reduced Zr-H-MCM-41-Red catalyst exhibits higher conversion than the oxidized one, and minor increase in the BAS density during the reduction of Zr-H-MCM-41 was observed as well (Table 1).

A positive effect on the conversion of 1-butene after Zr ion-exchange was not seen contrary to Cu ion-exchange, which indicates that similar beneficial interactions between zirconium ions and linear butenes do not exist.

However, the selectivity to isobutene over Zr-H-MCM-41-Red is about 10% higher than the selectivity over H-MCM-41 at the same conversion level (Fig. 6).

In genereal, there is almost a linear correlation between the selectivity to isobutene and the conversion at 673 K. The selectivity is slightly higher for the ion-exchanged samples.

This is reasonable, since increasing isobutene selectivity has been reported with the decreasing BAS density.16

In addition, the increased selectivity over Zr-H-MCM-41 can be also related to the decreased amount of Lewis acid sites in combination with the decreased amount of adjacent BAS due to ion-exchange, since the Lewis acid sites in combination with Brønsted acid sites have been reported to be responsible for the by-product formation.36

Decrease in the surface area due to the ion-exchange does not correlate with the catalytic activity; lowest conversion was observed over Zr-H-MCM-41, even if it has higher surface area than other evaluated materials.

Effect of ion-exchange of Zr and Cu into H-MCM-41

The conversion of 1-butene at 673 K and the WHSV of 5 h−1 over oxidized Cu–Zr-H-MCM-41-Ox is comparable to H-MCM-41 (Fig. 4).

Seemingly, over Cu–Zr-H-MCM-41-Ox the situation is equivalent to the one discussed above for Cu-H-MCM-41.

It is reasonable to assume that Cu2+ species are reduced to Cu+ already before TOS = 10 min.

The presence of Cu+ cations increases the catalytic activity but at the same time, the introduction of Zr decreases the BAS density resulting in a lower conversion level.

Thus, the conversion level after ion-exchange with Zr and Cu combined with oxidative pre-treatment is comparable to parent H-MCM-41 under studied conditions, but the selectivity to isobutene is slightly higher.

As in the case of Cu-H-MCM-41-Ox, the deactivation with TOS is due to the reduction of copper to Cu0 together with the coke formation.

At the same conditions, the conversion level is decreased over reduced Cu–Zr-H-MCM-41-Red being lower than that over H-MCM-41, probably due to the reduction of copper species to metallic copper and consequential presence of lower amounts of Cu+ species.

An additional explanation for the low activity is the reduced number of the BAS.

The behavior of Cu–Zr-H-MCM-41-Ox is peculiar at mild conditions, 623 K and the WHSV of 15 h−1.

The initial conversion of 1-butene is almost zero, but increases to 13 mol% in 260 min (Fig. 5).

This phenomenon is clearly related to in situ creation of active sites as described above for Cu-H-MCM-41-Ox: After oxidation the copper cations are mostly in the inactive oxidation state of +2, but are reduced during the run to the active species Cu+.

The reduction over Cu–Zr-H-MCM-41-Ox is obviously much slower compared to Cu-H-MCM-41-Ox.

This is an indication of the stabilizing effect of Zr on copper’s oxidation state of Cu2+ during the reaction, as observed also by XPS.

The stabilizing effect of Zr is further evidenced with reduced Cu–Zr-H-MCM-41-Red, which also exhibits an increase in the 1-butene conversion during time on stream being related to the reduction of Cu2+ to Cu+.

Similar increase in the conversion was not observed for Cu-H-MCM-41-Red.

This clearly indicates that substantial number of copper species has remained as Cu2+ in Cu–Zr-H-MCM-41-Red even after 2h of hydrogen pre-treatment at 773 K being reduced to Cu+ during the catalytic reaction.

The conversion level and selectivity are, however, far too low to be beneficial for this reaction after ion-exchange of H-MCM-41 with Cu and Zr.

Nevertheless, catalytic evaluation and XPS characterization results indicate that Zr stabilize the oxidation state +2 of copper to some extent.


H-MCM-41 catalyst were ion-exchanged by Zr and/or Cu-cations and characterized by XPS, 27Al- and 29Si-MAS-NMR, FTIR of pyridine adsorption, LA-ICP-MS and DCP as well as evaluated towards 1-butene skeletal isomerization.

The copper ion-exchange increased the conversion of 1-butene, although the Brønsted acid site density was decreased.

The reason for the activity improvement is the presence of active Cu+ species.

The ion-exchange of H-MCM-41 with Zr decreases the concentration of Brønsted and Lewis acid sites, which leads to a decrease in 1-butene conversion and simultaneous increase of the selectivity to isobutene and the catalyst stability.

The catalytic evaluation and characterization results suggest that the presence of Zr in Cu–Zr-H-MCM-41 made copper Cu2+ more resistant for reduction to Cu0.