Introduction

Methanol steam reforming (MSR) is a suitable reaction to produce hydrogen for PEM fuel cell.

$$ {\text{CH}}_{{3}} {\text{OH }} + {\text{ H}}_{{2}} {\text{O }} \to {\text{ 3H}}_{{2}} + {\text{ CO}}_{{2}} . $$
(1)

It is an endothermic reaction (∆H =  + 49.4 kJ mol−1) and thus requires high temperatures, moreover carbon monoxide, due to side reaction [1], can be formed as a by-product, both factors can affect the fuel cell performances up to ruin it [2]. Velu et al. [3] indicated that a suitable amount of oxygen could reduce CO formation and contribute to decrease the reaction temperature because the methanol oxidation (MO) reaction takes place.

$$ {\text{CH}}_{{3}} {\text{OH }} + \frac{1}{2}{\text{O}}_{{2}} \to {\text{ 2H}}_{{2}} + {\text{ CO}}_{{2}} . $$
(2)

This is an exothermic reaction (∆H = − 192.2 kJ mol−1), so it supplies heat and lowers the overall reaction temperature, moreover the presence of O2 helps to decrease the CO content in the outgoing gas mixture; MSR produces three hydrogen molecules for each methanol molecule, while MO only two. A combination of MSR and MO, the so called oxidative methanol reforming or combined methanol reforming (CMR), has been proven to be energetically favorable and helpful to minimize CO production [2].

$$ {\text{CH}}_{{3}} {\text{OH }} + \alpha {\text{H}}_{{2}} {\text{O }} + \beta \frac{1}{2} {\text{O}}_{{2}} \to \, ({2 } + \alpha ){\text{H}}_{{2}} + {\text{ CO}}_{{2}} , $$
(3)

With (α + β) = 1.

CuO–ZnO are catalysts commonly used for methanol synthesis from CO2 [4].

$$ {\text{3H}}_{{2}} + {\text{ CO}}_{{2}} \to {\text{ CH}}_{{3}} {\text{OH }} + {\text{ H}}_{{2}} {\text{O}}{.} $$
(4)

As Eq. 4 is the reverse of Eq. 1, Amphlett et al. [5, 6] studied CuO–ZnO catalysts in methanol steam reforming in the temperature range 150–370 °C, using W/F = 300–1400 kg s mol−1, finding 80% methanol conversion. Breen et al. [1] studied the system CuO–ZnO/ZrO2/Al2O3, finding a complete methanol conversion at 300 °C with W/F = 0.004 g min cm−3. Zhang et al. [7] studied the effect of ZrO2 and CeO2 on the CuO-ZnO catalysts in methanol steam reforming, and found that ZrO2 increased Cu dispersion with a consequent increase of methanol conversion and CeO2 could enhance oxygen storage with a decrease in CO production. Chang et al. [8] studied the effect of Al2O3, CeO2 and ZrO2 on catalytic performances of CuO–ZnO and found that all of them enhanced CuO and ZnO dispersion, Al2O3 decreased catalyst reducibility and weakened the reaction but gave mechanical strength to the catalyst, CeO2 increased catalyst reducibility but weakened the reaction and ZrO2 improved catalyst reducibility and promoted the reaction.

Our previous study [9, 10] focused on the system ZnO/TiO2, showed that it is active in methanol steam reforming with a complete reaction at T = 400 °C and GHSV (CH3OH) = 5 * 104 h−1. The activity of the system was attributed to the ZnO phase, TiO2 was useful to enhance surface area and to decrease CO formation, the best results were found when χZn = 0.05, segregation of ZnO was evident with higher Zn content, which caused the decrease in surface area and subsequent inhibition of the reaction. Later other authors have reported on the activity of Zn/TiO2 catalysts in the methanol steam reforming. Griboski et al. [11] ascribe the system reactivity of the system Zn/TiO2 to the zinc titanate presence and Deshmane et al. [12] relate the reactivity to the TiO2 dispersion, no consideration about zinc textural properties are reported and there is no ZnO evidence. Eaimsumang et al. [13] studied the effect of the morphology and crystalline phases of pristine TiO2 on oxidative steam reforming of methanol, showing that rutile or mixed phases are more active than anatase.

Xiao et al. [14, 15] studied the effects of TiO2 on CuO-ZnO in methanol synthesis, and found that amorphous TiO2 decreases CuO and ZnO crystallite size, changes CuO reducibility with a maximum of Cu0 surface area when TiO2 = 13% mol, which promoted highest CO2 conversion. Zhang et al. [16, 17] studied the effects of TiO2 on CuO-ZnO/Al2O3 in methanol synthesis and found that a little amount of amorphous TiO2 improves CuO dispersion, allowing a better adsorption/activation of H2 and CO2 and thus increases the reaction. Khani et al. [18] studied the hybrid thermo-photo catalytic reforming of methanol by CuZn/TiO2 (ZnO is supposed nanocrystalline or amorphous), and found a conversion of 35% and 63% at 150 and 200 °C, respectively, with a WHSV of 8.5 h−1 and H2O/CH3OH = 1. D. V. Andreev et al. [19] studied the system CuZn/TiO2 in methanol steam reforming with a WHSV = 4.43 h−1 and found a conversion of 40%, 90% and 100% at T = 375, 400 and 450 °C, in this case, Zn forms Zn2TiO4.

ZnO and TiO2 are widely studied in the literature because they are used as photocatalysts [20] and dye sensitized solar cell [21]. CuO, ZnO and TiO2 are commonly used as catalysts in different reaction [22], particularly in organic synthesis [23], moreover they have biocidal properties and this makes them useful in water decontamination [24] and stone protection in the field of cultural heritage [25, 26].

In this paper, we report preliminary results on the effect introduced by nanocrystalline TiO2 on CuO-ZnO catalyst, focusing our attention on structure, reducibility and catalytic performance in CMR, after hydrogen reducing pretreatment performed at 250 or 450 °C.

Experimental

CuO–ZnO (CZ) catalyst was prepared by co-precipitation of carbonates in a similar way to Okamoto et al. [27]. Na2CO3 (Merck), zinc nitrate hexahydrate (Aldrich) and Copper (II) nitrate hemi (pentahydrate) (Aldrich) were used as starting reagents. A water solution containing zinc nitrate (0.1 M) and copper (II) nitrate (0.04 M) was prepared under magnetic stirring and kept in N2 bubbling, then a Na2CO3 water solution (0.4 M) was added dropwise, then the temperature was raised to 80 °C for 3 h. At the end the pH of the solution was 9, the precipitate was decanted overnight, washed with distilled water to eliminate all sodium, filtered, dried at 90 °C for 12 h and calcined at 350 °C for 3 h. CuO–ZnO/TiO2 (CZT) catalyst was prepared as CZ, TiO2 (Aldrich, anatase) was added in the water solution containing copper and zinc nitrate.

CMR reactions was studied. Catalytic tests were performed in a fixed bed quartz reactor with an inner diameter of 1 cm. A suitable amount of catalyst was diluted 1/10 in sieved α-Al2O3, and then put into the reactor that had already been filled up to 2 cm with 1 mm diameter SiO2 balls. The catalysts were pre-treated in (10%) H2/He mixture at 250 °C or 450 °C for 2 h, the temperature was raised at 5 °C min−1. Catalytic tests were performed in the range of temperature 200–400 °C. A solution of methanol in water was introduced into the reactor by a syringe pump (Sage Instruments model 341 A), O2 and He flows were regulated by flowmeters. The reaction condition settings are summarized in Table 1. A blank test was performed with a reactor containing only SiO2 balls and α-Al2O3.

Table 1 Combined methanol reforming reaction conditions
Full size table

The temperature of the catalysis line from the reactor outlet to the chromatograph automatic sampling area was kept at 120 °C, so that CH3OH and H2O vapors were transported by the vector gas and analyzed together with other gas components. The reaction products were analyzed by a Carlo Erba 4300 GC, equipped with a Hayesept D column, products were detected by a thermal conductivity detector.

In the following CZ(T) refers to the as calcined catalysts, CZ(T)250(or 450) refers to the catalyst exhaust, 250 or 450 indicate the hydrogen pretreatment temperature.

A Philips PW 100 diffractometer with a Ni-filtered Cu Kα (λCu = 1.54 Å) radiation was used to record powder X-ray diffraction patterns. Spectra were collected on fresh and exhaust catalysts. Crystallite sizes (d) were calculated by the Debye Scherrers equation:

$$ {\text{d }} = K\lambda /\beta \cos \theta , $$

Here k is a constant (0.9), λ is X-ray wavelength, β is the FWHM diffraction peak and θ is the diffraction angle.

A Varian Spectra A-30 Instrument atomic absorption was used to determine elemental composition. Surface area of catalysts were calculated by the BET isotherm of N2 adsorption by a Micromeritics ASAP 2010 analyzer after having degassed the catalyst in vacuum at 170 °C.

Results and discussion

Catalysts have been prepared in such a way to get the same ZnO/CuO ratio, with the aim to study the effect introduced by the nanocrystalline TiO2 presence. We report chemical composition and surface area of fresh CZ and CZT catalysts in Table 2. The nanocrystalline TiO2 presence slightly increases surface area.

Table 2 Chemical composition and surface area of ​​the catalysts
Full size table

CH3OH, H2O, CO, CO2 and H2 were found during catalytic tests, no other species were detected. CH3OH conversion, CO selectivity and H2 yield were calculated as follows:

$${\text{CH}}_{3} {\text{OH conversion}} = \frac{{\left[ {{\text{CO}}} \right] + \left[ {{\text{CO}}_{2} } \right]}}{{\left[ {{\text{CO}}} \right] + \left[ {{\text{CO}}_{2} } \right] + \left[ {{\text{CH}}_{3} {\text{OH}}} \right]}}, $$
$$\mathrm{CO\, selectivity}=\frac{\left[\mathrm{CO}\right]}{\left[\mathrm{CO}\right]+\left[{\mathrm{CO}}_{2}\right]},$$
$$ {\text{H}}_{2} \;{\text{yield}} = \frac{{\left[ {{\text{H}}_{2} } \right]}}{{\left[ {{\text{CO}}} \right] + \left[ {{\text{CO}}_{2} } \right] + \left[ {{\text{CH}}_{3} {\text{OH}}} \right]}}. $$

Catalyst performances obtained in CMR reaction after hydrogen pretreatment at 450 °C are compared for ZC and ZCT in Fig. 1.

Fig. 1
figure 1

Catalyst performances obtained for ZCT (black circles) and ZC (green triangles) after hydrogen pretreatment at 450 °C in H2/He (10%) for CMR. Reaction conditions: GCH3OHHSV = 55.000 h−1, O2 = 2.0%, H2O = 18.5%, CH3OH = 16.9%, He = 62.6%. (Color figure online)

Full size image

Methanol conversions (Fig. 1a) and hydrogen yields (Fig. 1b) as a function of the temperature have the classical sigmoidal shape. Nanocrystalline TiO2 presence in the catalyst decreases both of them in the fast increasing part of the sigmoid (T = 250 and 300 °C), however it allows to get a slightly higher value of hydrogen yield when the reaction reaches completeness. CO selectivity increases with the temperature (Fig. 1c) and with the methanol conversion (Fig. 1d) for both catalysts, it is evident that the nanocrystalline TiO2 presence decreases the CO production.

Catalyst performances of ZC and ZCT after hydrogen pretreatment at 250 °C are compared in Fig. 2. Also in this case methanol conversion (Fig. 2a) and hydrogen yield (Fig. 2b) values have a sigmoidal shape. TiO2 presence decreases the reaction at lower temperature and increases the reaction at higher temperature, so the completeness of the reaction is reached at a lower temperature. Hydrogen yield trends (Fig. 2b) seem to reflect methanol conversions trends. CO selectivity values have again an increasing trend as a function of the temperature (Fig. 2c) and the methanol conversion (Fig. 2d). The nanocrystalline TiO2 is useful to decreases CO selectivity as a function methanol conversion when methanol conversion is higher than 50%.

Fig. 2
figure 2

Catalyst performances obtained for ZCT (black circles) and ZC (green triangles) pretreated in hydrogen at 250 °C for CMR. Reaction conditions: GCH3OHHSV = 55.000 h−1, O2 = 2.0%, H2O = 18.5%, CH3OH = 16.9%, He = 62.6%. (Color figure online)

Full size image

Resuming nanocrystalline TiO2 presence affects CuO–ZnO catalytic activity, it decreases the reactivity when the hydrogen reducing pretreatment is performed at 450 °C and increases the reactivity when the hydrogen reducing pretreatment is performed at 250 °C. Lower CO production is observed with the catalyst containing nanocrystalline TiO2 independently of the temperature of the pretreatment.

The methanol conversion and the hydrogen yield obtained after the catalytic tests after the two different hydrogen pretreatment temperatures are compared in Fig. 3 for ZCT (Fig. 3a) and ZC (Fig. 3b). It can be easily seen that the hydrogen pretreatment temperature affects the methanol conversion and hydrogen yield of ZCT, but does not change those of ZC. It is now clear that TiO2 makes the catalyst performance more sensible to the temperature of the hydrogen pretreatment.

Fig. 3
figure 3

influence of the hydrogen temperature pretreatment on methanol conversion and hydrogen yield of ZCT (a) and ZC (b). Black circles = CH3OH conversion after H2 pretreatment at 250 °C, pink triangles CH3OH conversion after H2 pretreatment at 450 °C, green triangle upside = hydrogen yield after H2 pretreatment at 250 °C, blue rhombus = hydrogen yield after H2 pretreatment at 450 °C. Reaction conditions: GCH3OHHSV = 55.000 h−1, O2 = 2.0%, H2O = 18.5%, CH3OH = 16.9%, He = 62.6%. (Color figure online)

Full size image

Investigation about differences caused by the hydrogen pretreatment temperature have been studied by X-ray diffraction, which has been performed on the fresh and exhaust catalysts. The diffractograms are reported in Fig. 4, TiO2 (Aldrich) anatase diffraction pattern (T) is also reported for a comparison. Fig. 4a displays diffraction patterns of fresh ZC, ZCT and T. ZC and ZCT clearly present ZnO in wurtzite phase (JCPDS Card No. 36-1451) [28] with (100), (002) and (101) reflections at 2θ = 31.7, 34.5 and 36.2. Anatase phase (JCPDS: No 21-1272) [29] is the only component of T and is present in ZCT, showing its highest reflection (101) at 2θ = 25.3, secondary reflections are also present, as labelled in Fig. 4a. Fig. 4b shows ZC and ZCT diffractograms in the region 28° ≤ 2θ ≤ 42°, recorded as a result of a greater number of acquisitions, to get a higher signal/noise ratio. Moreover, since (111) CuO diffraction peak almost overlaps the (004) anatase diffraction peak, a suitable subtraction of the TiO2 spectra has been performed. CuO in tenorite phase (111) reflection at 2θ = 38.8 (JCPDS card no. 45-0937) [30] is visible with a very low intensity for both catalysts.

Fig. 4
figure 4

X-ray diffraction pattern of ZC and ZCT catalysts (a, b inset enlarged) CZ 450 (c) and CZT 250 and CZT 450 (d). A anatase TiO2, W wurtzite ZnO, T tenorite CuO, C cuprite Cu2O

Full size image

X-ray diffraction patterns of exhaust ZC and ZCT are reported in Figs. 4c, d, respectively, compared with those of the fresh ones. Wurtzite is the dominant phase in ZC450 catalyst (Fig. 4c), tenorite phase is present, both phases show narrower peaks than CZ diffractogram, meaning that the hydrogen pretreatment and catalytic tests promote an increase in crystallite size. Anatase and wurtzite are dominant phases in all ZCT (250 or 450) diffractograms (Fig. 4d), tenorite phase is still present with low intensity. Cu2O in cuprite phase (JCPDF no. 78-2076) [31] is visible in ZCT250 by (200) diffraction peak at 2θ = 42.3, the main Cu2O diffraction peak (111) at 2θ = 36.5 is hidden by the more intense wurtzite (101) peak in the same region. Cu0 (JCPDF no. 85-1326) [31] formation in CZT450 is pointed out by (111) and (200) diffraction peaks at 2θ = 43.4° and 50.4°, respectively.

In Table 3, we report the crystallite size of all phases as calculated by Debye Scherrer’s equation. X-ray analysis reveals the nanocrystalline structure of ZC and ZCT fresh catalysts; wurtzite and tenorite phase crystallite sizes are only slightly increased by the nanocrystalline TiO2 presence. Catalytic tests and hydrogen pretreatments cause complex changes in the structure. Comparing ZC with ZC450 we can observe that the H2 reducing pretreatment and catalytic test induce a little increase in crystallite size on both wurtzite and tenorite, but a good dispersion is still present. Viceversa comparing ZCT with ZCT250 and ZCT 450 a higher increase in ZnO crystallite size is observed, together with a gradual chemical reduction of CuO to Cu2O and Cu0. All ZCT250 phases are still well dispersed, it’s not so in ZCT450, where the formation of submicron phases of ZnO and Cu0 is found. Our experiments show that the nanocrystalline TiO2 presence slightly increases ZnO and CuO crystallite size and makes CuO more reducible. This result is apparently in contrast with the literature [14, 32], where amorphous TiO2, increased dispersion of ZnO and CuO and decreased CuO reducibility. Nanocrystalline TiO2 has oxygen vacancies, which are as electron donors and form n-type levels, in amorphous TiO2 n-type level formation cannot be ensured. On the other hand, electron donors centers can interact with CuO and ZnO ionic structure, causing a lower dispersion and a higher reducibility of CuO.

Table 3 Crystallite size (nm)
Full size table

We had shown that nanocrystalline TiO2 presence enhanced the CRM reaction if the hydrogen pretreatment was carried out at T = 250 °C (Fig. 2), and decreased the reaction if the hydrogen pretreatment was carried out at T = 450 °C (Fig. 1). We had also shown that this variation was due to a change in the catalytic performance of ZCT induced by the different hydrogen pretreatment temperature, while ZC performance remained unchanged (Fig. 3). We can now deduce that a nanosized structure of CuO and ZnO is essential to get good catalytic performance, highest methanol conversion and hydrogen yield of ZCT250 are ascribed to the formation of nanosized and Cu2O, while lowest methanol conversion and hydrogen yield of ZCT450 are ascribed to formation of submicron ZnO and Cu0.

Moreover, since CZT250 and CZT450 produce less CO than CZ250 and CZ450 respectively, we attribute the ability to decrease CO formation to the nanocrystalline TiO2 presence. It has been previously established [1, 33] that CO formation during methanol reforming reactions over CuO–ZnO catalysts is due the retro water gas shift reaction (r-WGS):

$$ {\text{CO}}_{{2}} + {\text{ H}}_{{2}} \to {\text{ CO }} + {\text{ H}}_{{2}} {\text{O}}{.} $$
(5)

Thus, a catalyst can facilitate this reaction if it is able to attract and adsorb the molecule of CO2 or H2. Both molecules have a linear shape; H2 molecule has a net electric dipole moment of zero, thus its adsorption could be facilitate by a catalyst with no electric charge on the surface. The CO2 molecule, with no net electric dipole, has a δ− charge at the ends and a δ+ charge in the center, because of steric/electric and probabilistic factors, its adsorption could be facilitated by the presence of positive charges, and be made more difficult by the presence of negative charges on the surface of the catalyst. It is still the presence of n-type levels on the nanocrystalline TiO2 that will tend to impede the CO2 adsorption and subsequently the r-WGS reaction, resulting in a final decrease of CO formation. These considerations are not in contrast with Zhang et al. [17], who attributed to TiO2 the ability to facilitate the adsorption of H2 and CO2 in CZT catalysts, because TiO2 was amorphous or highly dispersed, such as not to be detectable by X-ray diffraction. In that case, the presence of n-type levels could not be ensured as already explained.

Conclusions

In this work, we have shown that nanocrystalline TiO2 affects CuO-ZnO structure and catalytic activity in CMR. Nanocrystalline TiO2 causes a slight increase in ZnO (wurtzite) crystallite size and does not significantly change the size of CuO crystallites (tenorite) after calcination at 350 °C. Both wurtzite and tenorite phases are strongly influenced by nanocrystalline TiO2 presence after catalytic tests: wurtzite phase undergoes an increase in crystallite size, this effect is moderate if the hydrogen pretreatment is performed at 250 °C and is stronger the hydrogen pretreatment is performed at 450 °C, causing the formation of submicrometer crystallites. Tenorite phase undergoes a slight increase in crystallite size, resulting always nano sized and well dispersed. However, nanocrystalline TiO2 presence makes tenorite more chemically reducible, which causes the formation of nanosized Cu2O (cuprite) and submicron Cu0 after catalytic tests performed after hydrogen pretreatment at 250 °C and 450 °C.

It has been shown that TiO2 presence increases CuO–ZnO catalytic activity upon hydrogen pretreatment at 250 °C and decreases it upon pretreatment at 450 °C.

The best catalytic performances of the CZT250 are thus associated to the nano sized ZnO and well dispersed CuO and Cu2O, the worst catalytic performances of CZT 450 are in turn associated to the loss of the nanocrystalline nature of the catalyst with the formation of submicron ZnO and Cu0.

In the end, it has been proven that the nanocrystalline TiO2 plays a clear role in limiting the production of CO, independently of the pretreatment temperature. This ability of nanocrystalline TiO2 is ascribed to the presence of the oxygen vacancies, which act as electron donors, which contribute to hinder the CO2 and H2 surface adsorption, which is a necessary step to produce CO by r-WGS reaction.