Ind. Eng. Chem. Res. 1994,33, 2047-2055 2047

Production of Hydrogen from Methanol. 1. Catalyst Characterization Studies

Introduction

Raphael 0. Idem and Narendra N. Bakhshi' Catalysis and Chemical Reaction Engineering Laboratory, Department of Chemical Engineering, University of Saskatchewan, Saskatoon, Saskatchewan, Canada S7N OW0

Various Cu-A1 catalysts with copper concentration ranging from 0 to 57.3 w t % were prepared by coprecipitation techniques for the production of H2 from methanol. These catalysts were calcined a t various temperatures and then reduced in H2 a t 300 "C. The catalysts were characterized thoroughly at appropriate stages of preparation. Results show that the dried catalysts contained crystalline species such as Cuz(OH)zC03 and Cu(OH)2 and noncrystalline species such as Al(OH)C03, Al(OH)3, and Cuz(OH)2C03. Their proportions and thermal stability depended on copper concentration. Calcined catalysts contained Cu2(0H)2C03, CuO, Cu20, and CuAlzO4 reducible species, Their proportions, reducibilities, and extents of reduction with hydrogen depended strongly on copper concentration and calcination temperature. The acid site distribution on calcined catalysts varied drastically with copper concentration and calcination temperature. On the other hand, BET surface area and pore structure depended not only on copper concentration and calcination temperature but also whether the catalyst was reduced or unreduced.

Hydrogen is consumed in large quantities in many industrial applications. Currently, it is being seriously considered for use commercially as an energy vector such as in fuel cells and mobile vehicles due to its high efficiency, and also for environmental reasons (Uehara, 1992). It is produced largely from hydrocarbon reforming and as off- gases from many refinery operations.

However, in Western Canada, with the large need for hydrogen in such heavy industries as fertilizer, chemicals, and petrochemicals manufacturing, petroleum refining, and oil sands and heavy oil upgrading, its demand will certainly have to be supplemented from outside the region (Yildirim et al., 1991). Consequently, in addition to the traditional hydrogen sources, other sources will have to be tapped.

Although other methods are available, the use of methanol as a chemical carrier to circumvent the problems of hydrogen storage and transportation appears to be the most promising option, especially, when viewed against the favorable economics of long-distance transportation (de Bokxet al., 1988; Kliman, 1983; Othmer, 1982). Thus, methanol can be manufactured cheaply where natural gas is abundant and transported conveniently to where the demand for hydrogen outstrips supply. Recently, low- temperaturelhigh-yield processes for methanol manufac- ture from methane have been proposed (Periana et al., 1993; Hickman and Schmidt, 1993). Hydrogen can be made by the steam reforming of methanol. Theoretically, more hydrogen can be produced than is available in methanol alone when steam reforming is employed.

In recent literature, studies on hydrogen production from methanol have been carried out using copper-containing catalysts such as Cu-ZnO/A1203, Cu/ZnO, Cu-Ni/AlzOs, Cu/A1203, Cu/Cr203, Cu/SiOz, Cu-Mn-Cr-Si, and Cu- Ni/AlP04 (Cheng, 1989; Amphlett et al., 1991; Isogai et al., 1989). Cu/ZnO catalysts have been studied in great detail, and it is interesting to note that they are also used commercially for low-temperature methanol synthesis. According to Amphlett et al. (19911, since the reverse reactions for methanol decomposition and reforming are those generally used for the overall methanol synthesis, the forward and reverse reaction paths are likely to be similar. Consequently, Cu/ZnO catalysts would have

0SSS-5SS5/94/2633-204~~04.50/0

similar effects on both methanol decomposition and reforming as on methanol synthesis. However, there is evidence in the literature (Kobayashi et al., 1976; Mat- sukata et al., 1988) that catalysts based on Cu-A1 are more active for the steam reforming of methanol than the Cu/ ZnO catalysts.

Also, it is known from catalysis literature that physical and chemical properties exert a strong influence on the catalyst performance. These properties are controlled by independent variables such as catalyst constituents, composition, and preparation methods.

A number of studies have been carried out using copper- based catalysts to analyze the effects of some of these variables on catalyst performance (Amphlett et al., 1990; Cheng, 1989; Isogai et al., 1989; Matsukata et al., 1988; Agaras et al., 1988; Takezawa et al., 1982; Minochi et al., 1979; Kobayashi et al., 1976). However, apart from the Cu/ZnO catalysts, information regarding the character- istics of other catalysts is scant. Consequently, the relationship between catalyst characteristics and catalyst performance in the production of hydrogen from methanol is not clearly understood.

It was therefore decided to carry out an in-depth characterization study on these catalysts. Eighteen co- precipitated copper-aluminum catalysts with copper concentration ranging from 0 to 57.3 wt 5% were used in this investigation.

Experimental Section

Six catalyst precursors containing different proportions of copper and aluminum were prepared by coprecipitation techniques. They were calcined at 3 temperatures and then reduced in H2 a t 300 "C. Various techniques were employed in their characterization at each stage of preparation.

Catalyst Preparation. A. Materials. The source of copper was copper nitrate trihydrate [Cu(NO3)2*3H20; ACS grade, obtained from Johnson Matthey Company, Ward Hill, MA] while that for aluminum was aluminum nitrate nonahydrate [Al(N03)~9HzO; >99.5 5% pure; ob- tained from BDH chemicals Ltd., Toronto, Canada].

B. Coprecipitation Procedure. Six solutions con- taining copper nitrate trihydratelaluminum nitrate non-

1994 American Chemical Society \

2048 Ind. Eng. Chem. Res., Vol. 33, No. 9, 1994

ahydrate (C/A) weight ratios of 0.05,0.17,0.32, and 0.52, copper nitrate trihydrate only (C), and aluminum nitrate nonahydrate only (A), each with a concentration of 105 g of (copper nitrate trihydrate + aluminum nitrate non- ahydrate)/L of distilled water, were prepared. Precipita- tion of each precursor was effected by adding the aqueous solution containing the appropriate C/A weight ratio to 1.1 times its stoichiometric requirement of a well-stirred 0.5 M aq.ueous solution of sodium carbonate (pH 11.5) maintained at 40 "C in a 3-L glass flask. Vigorous stirring of the resulting slurry (pH of about 7.5) continued for 10 min. The precipitate was filtered and washed several times with warm water and finally with cold water, and then dried overnight in air a t 110 "C.

C. Catalyst Calcination. Each dried precipitate (catalyst precursor) was pelletized in a hydraulic press under a compacting pressure of 10 000 psig, crushed, and sieved into particle size ranging from -8 + 10 mesh prior to calcination. Three calcination temperatures were used, namely, 300,500, and 700 "C. This procedure resulted in 18 catalysts. Calcination was carried out by placing a 20-g sample of the catalyst (contained in a nickel crucible) in a muffle furnace maintained at the desired temperature. After 3 h, the sample was removed and allowed to cool at room temperature.

D. Catalyst Reduction. The calcined catalysts were reduced in hydrogen. The particular procedure depended on the property under consideration. However, reduction was performed at 300 "C in all cases. Details are given under Catalyst Characterization.

Catalyst Characterization. Characterization of cata- lysts was performed at each preparation stage. This included catalyst precursors, unreduced calcined catalysts, and reduced calcined catalysts.

Catalyst Precursors. Uncalcined catalyst precursors were characterized in order to determine their elemental composition, species present and their distribution, their thermal stability, and, consequently, the minimum cal- cination temperature.

A. Chemical Analysis. Bulk copper, aluminum, and sodium elemental compositions were determined by inductively coupled plasma-atomic emission spectroscopy (ICP-AES) and the amount of carbon was determined by combustion. Also, nitrogen, hydrogen, and carbon analyses were performed with a C H N Perkin Elmer elemental analyzer. The amounts of carbon obtained from the two types of analyses for each sample were in good agreement. Oxygen was obtained by difference.

B. Differential Scanning Calorimetry (DSC). Information on the thermal characteristics, minimum calcination temperature, and species present was obtained using the differential scanning calorimetry (DSC) tech- nique. The DSC analysis was performed using a Mettler TA 4000 thermal analyzer with a plug-in Mettler DSC-20 cell. A t least two runs were conducted for each sample to ensure reproducibility. All the runs began at 30 "C and ended at 600 "C at a heating rate of 10 "C/min. The heat flow due to decomposition or phase change in the sample was recorded as a function of temperature.

C. Powder X-ray Diffraction (XRD) Measure- ments. Powder XRD analyses were performed in order to further verify the species present in the catalyst precursors as well as determine their degree of crystallinity. The XRD measurements were made with a Phillips diffractometer using Fe Kru radiation in the scanning angle (28) range of 10-90" at a scanning speed of 2 deg/min.

D. Scanning Electron Microscopy and Electron Probe Microanalysis. Both scanning electron micros-

copy (SEM) and electron probe microanalysis (EPMA) were performed on catalyst precursors using a JEOL super probe (Model JXA-8600) with an accelerating voltage of 15 kV and a probe current of 10 nA. An energy-dispersive spectrometer (EDS) was used to initially verify the presence of copper, aluminum, and sodium at various areas of each precursor sample. Local concentrations of copper and aluminum were measured at various points using LiF and TAP wavelength-dispersive spectrometers (WDS), respectively.

Calcined Catalysts. After calcination, the following techniques were used for catalyst characterization.

A. Temperature-Programmed Reduction (TPR). The reduction of a reducible material with hydrogen is a topochemical reaction capable of supplying a great deal of information about the material. Consequently, the temperature-programmed-reduction technique was em- ployed for the verification of the nature of the species present in the catalysts.

The analyses were performed a t atmospheric pressure in a conventional flow system. The equipment and procedure used were similar to the ones described by Varmaet al. (1984). TPR analysis was conducted for both reduced and unreduced catalysts. That of unreduced but calcined catalysts consisted of placing about 0.1 g of the sample in the reactor and heating it in the reducing gas stream at a linearly programmed rate of 4 "C/min from 25 to 600 "C.

TPR analysis of catalysts reduced isothermally at 300 "C consisted of placing about 0.2 g of the calcined sample in the reactor, reducing it in situ a t 300 "C in pure hydrogen at a flow rate of 60 mL/min for 2 h, allowing it to cool to room temperature in flowing nitrogen, and then starting the TPR program.

B. Temperature-Programmed Desorption (TPD) of "3. This analysis was performed in order to obtain information about the acidity of the calcined catalysts.

The equipment and procedure used were similar to those described by Hidalgo et al. (1983). The carrier gas used was N2 (99.995% purity; obtained from Linde) a t a flow rate of 60 mL/min, while a mixture of 1 vol % NH3 in N2 (high purity; also obtained from Linde) was used as the adsorbing gas. About 0.5 g of the calcined sample was used for each experiment. Adsorption was performed at 120 "C to eliminate the effects of physically adsorbed "3. All TPD runs began at 120 "C and ended at 700 "C at a program rate of 8 "C/min.

C. BET Analyses. BET measurements were per- formed to determine the BET surface area, pore volume, pore size, and pore size distribution of catalysts. The analyses were conducted using Micromeritics gas adsorp- tion equipment (Model ASAP 2000) with nitrogen (99.995% purity; obtained from Linde) as the analysis gas. BET surface area measurements and pore structure analyses were made for both reduced and unreduced catalysts.

In the case of reduced catalysts, reduction was effected in the chemisorption arm of the Micromeritics equipment a t 300 "C using pure hydrogen (99.995% purity; obtained from Linde) a t a flow rate of 100 mL/min for 2 h.

Prior to BET analyses, each catalyst was evacuated at 290 "C to a vacuum of 3 pmHg. Evacuation continued for a minimum of 10 h for unreduced catalysts and for 2 h for reduced catalysts.

D. Low-Temperature 0 2 Chemisorption Analyses. Copper metal area was measured using the low-tempera- ture 0 2 chemisorption technique proposed by Vasilevich et al. (1976). Chemisorption experiments were conducted

.

Ind. Eng. Chem. Res., Vol. 33, No. 9, 1994 2049

mixture of nitrates of copper, manganese, magnesium, and chromium, observed very broad peaks superimposed on a very high background and only copper hydroxycarbonate was identified in some of the samples. This occurrence was attributed by these authors to the formation of precipitates with crystallite size less than 50 nm.

Since all catalysts (except p-1) contained copper species with concentrations capable of being detected by XRD analysis, it was necessary to determine their compositional variations using SEM and EPMA in order to account for this apparent amorphous behavior. This is discussed below.

B. SEM and EPMA. A typical electron micrograph obtained from the backscattered electron (BSE) image is given in Figure 2 for catalyst p-5 containing 27.8 wt % copper. In the figure, copper species are seen as the bright patches of size 3-9 pm distributed throughout a dark background which, from BSE images, varied in darkness from light gray to dark indicating an increasing concen- tration of aluminum species. The grain sizes in the dark region were not detectable up to the nanometer level. Electron probe microanalysis using WDS showed that the bright patches contained only copper species while the copper/aluminum atomic ratio decreased from 1.07 in the light gray areas to 0.31 in the dark region. These results show the existence of two main areas in the catalyst precursor. These are a large dark region with a high degree of interdiffusion between copper and aluminum species which is X-ray amorphous (noncrystalline) and a small overall bright area containing copper species (Cu2(0H)2- COa and Cu(OH)2) capable of diffracting X-rays. A similar distribution of bright patches &e., crystalline copper species), though smaller in amount and size, was observed for catalyst p-4 containing 24.1 wt 7% copper. In the case of catalysts p-2 and p-3 containing 5.79 and 15.3 wt % copper, respectively, there were no observable bright areas. A complete interdiffusion of copper and aluminum species was obtained. Although the Cu/Zn atomic ratios were different, Himelfarb et al. (1985), in a characterization study of Cu/ZnO catalyst precursors, also observed drastic compositional variations. These authors used a variety of techniques to show that the distribution and structures of species depended on copper concentration.

Additional characterization of the coprecipitated Cu- A1 catalyst precursors was achieved using elemental composition and DSC techniques.

C. Elemental Composition. Virtually no nitrogen was observed in any of the precipitates and as such was not included in Table 1. The absence of nitrogen indicates the absence of nitrates in the precipitates. This can be attributed to the high solubility of nitrate salts in water and the large probability of being washed away during filtration as well as to high pH of the medium during precipitation. Minochi et al. (1979) have shown that a high pH during precipitation resulted in the exclusion of the anions of the source materials from the precipitates. Similarly, since both ICP-AES and EPMA did not detect any sodium, its concentration was not included in Table

Table 1. Elemental Commsition of Catalyst Precursors ~ ~~

elemental cornpition (wt % CIA catalyst

ratio4 identity copper aluminum carbon hydrogen oxygen

0.05 p 2 5.79 16.40 2.50 2.08 73.22

0.52 p-5 27.80 12.80 3.18 2.71 53.51

A p-1 0.00 27.04 2.11 2.68 68.17

0.17 p-3 15.30 14.00 2.74 2.14 65.82 0.32 p-4 24.10 13.70 2.93 2.60 54.67

C p-6 57.33 0.00 5.36 1.10 36.21

4 C/A, copper nitrate trihydrate/aluminum nitrate nonahydrate weight ratio; C, copper nitrate trihydrate only; A, aluminum nitrate nonahydrate only.

40 I I

0 ; I I I I I I I I , 10 20 30 40 50 00

Scanning angle (Ze), deg Figure 1. XRD pattern of catalyst p 6 .

in the chemisorption arm of the adsorption equipment (Model ASAP 2000) using oxygen (99.995% purity; obtained from Linde) as the analysis gas a t a temperature of 78 K (Klier et al., 1982). Before chemisorption measurements, each catalyst was reduced at 300 "C for 2 h in hydrogen (flow rate 100 mL/min) and evacuated at this temperature for 1 h. This was followed by evacuation at room temperature for 30 min, and at the analysis temperature (78 K) for 30 min. Corrections were made for the effects of physical adsorption on the catalysts using the method reported by Fierro (1990). Copper dispersion and average crystallite size were calculated according to the method reported by Bhatia et al. (1978).

Results and Discussion

Catalyst Precursors. A combination of various char- acterization techniques, namely, XRD, SEM, EPMA, DSC, and elemental composition were used to elucidate the nature of the dried catalyst precipitates (Le., catalyst precursors). They were designated as p-1, p-2, p-3, p-4, p-5, and p-6. Their elemental compositions are given in Table 1.

A. X-ray Diffraction (XRD) Measurements. XRD measurements were performed to identify the component phases as well as to determine the degree of crystallinity of catalyst precursors as a function of copper concentration.

The XRD spectra of the catalyst precursor containing copper only (catalyst p-6) is given in Figure 1. From the figure, major and minor peaks are observed a t scanning angles characteristic of copper hydroxycarbonate and copper hydroxide, respectively. For precursors of other compositions, the XRD spectra showed only trace peaks superimposed on a high background. Only patterns characteristic of copper hydroxycarbonate were identified with intensities diminishing as the copper concentrations decreased, showing that the species were X-ray amorphous and that their degree of crystallinity decreased with copper concentration. Amphlett et al. (1990), working on a

1. Table 1 shows that the precursors contain significant

amounts of carbon and hydrogen. These results are consistent with those from XRD measurements in which copper hydroxycarbonate was identified. The presence of hydroxycarbonates could be attributed to the use of Na2C03 for precipitation (Ghiotti and Boccuzzi, 1987). Table 1 also shows a monotonic increase in the proportion of carbon with copper concentration. Comparison of the carbon compositions of the precursors with zero copper

2050 Ind. Eng. Chem. Res., Val. 33, No. 9,1994

I

I

Figure 2. Electron micrograph of catalyst precursor containing 27.8 wt ' i , copper.

30 concentration and zero aluminum concentration (catalysts p-1 and p-6) shows that the hydroxycarbonate was formed mostly with copper. Consequently, the carbon require- ment increased with copper concentration, However, not all the copper present resulted in the formation of the hydroxycarbonate especially for samples with higher copper concentration. The remaining copper was present in the form of Cu(0H)las wasverified from XRDanalysis.

Calculations, based on the stoichiometric requirements of AI, 0, and for the C available, showed that only a small fraction of AI was present in the form of aluminum hydroxycarbonate for catalyst p-1.

D. Differential Scanning Calorimetry (DSC) Analysis. DSC was performed in order to provide further evidence for the presence of the various species, evaluate

minimum temperature suitable for calcination. The DSC

s 20 E

0 m a

~ .I

2 10

5 er a

U

- .I

their thermal behavior, and, consequently, determine the

spectra for all dried catalyst precursors are presented in Figure 3. Peaks, peak temperatures, and the corresponding characteristic species are derived from DSC spectra, and the results are summarized in Table 2.

For the sample with zero copper concentration (curve l), the endothermic peak around 190 "C represents the abstraction of water from aluminum hydroxide and the decomposition of the small amount of aluminum hy-

0

2 0 2

-10

droxycarbonate present in the precursor. For other catalysts of varying copper concentrations,

theendothermicpeakat ca. 160 'Crepresents the removal of the physically adsorbed water from the material, while the endothermic peak a t ca. 200 'C represents the abstraction of chemically bound water from aluminum hydroxide, copper hydroxide, and copper hydroxycar- bonate and the decomposition of the small amount of aluminum hydroxycarbonate that might be present.

The endothermic peak a t ca. 300 "C observed in the DSC spectra of precursors with higher copper concentra- tion represents solely the decomposition of the crystalline

100 200 300 400

Temperature ( C ) Figure 3, DSC spectra of catalysts precursors. 1.0 w t % Cu; 2,5.79 wt%Cu;3,15.3wt%Cu;4,24.1wt%Cu;5,27.0wt%Cu;6,57.3 w t % cu.

copper hydroxycarbonate. This is because the DSC spectra of the bright green colored catalyst precursor p-6 (containing copper only), which was identified by powder XRD to contain copper hydroxycarbonate as the major component and copper hydroxideas the minor component,

Ind. Eng. Chem. Res., Vol. 33, No. 9, 1994 2051 Table 2. Summary of DSC Analysis of Catalyst Precursors

no. of peaks copper concn (wt % ) peak temp ("C) 0.00 1 190 5.79 2 160,200 15.3 2 150,200 24.0 3 160,200,300 27.8 3 150,200,300 57.336 2 180,300

0 c, crystalline; n, noncrystalline; b, both crystalline and noncrystallir

Table 3. Summary of TPR Analyses of Reduced and Unreduced Catalysts

reducible copper reducible copper catalyst copper calcn no. peak speciesin species in concn temp of temp unreduced EORO reduced (wt %) ("0 peaks ("C) samples (%) samples

0.00 300 0 none 0 none 500 0 none 0 none 700 0 none 0 none

5.79 300 1 411 Cuz(OH)zCO3 0 Cuz(0H)zCOs 500 1 389 Cuz(0H)zCOs 0 CUZ(OH)ZCO~

700 2 343 Cuz(0H)zCOs Cui(0H)zCOs

15.30 300 1 363 Cuz(OH)zCO3 80 Cuz(0H)zCOs 500 1 345 Cuz(0H)zCOs 90 Cuz(0H)zCOs

cuAlo4 CAzOi

280 cU0,CuAloi 40 CuAlzOi

cuAlzo4 C d z O i 700 1 275 CuO,CuAl204 100 none

24.00 300 2 243 CuO 100 none

500 2 264 Cu0,CuA120r 100 none

700 2 211 Cuz0,CuO 100 none

27.8 300 2 263 CuO 100 none

500 2 263 Cu0,CuA120i 100 none

700 2 211 Cuz0,CuO 100 none

57.33b 300 1 258 CuO 100 none 500 1 262 CuO 100 none 700 1 264 CuO 100 none

359 Cuz(0H)zCOs

355 Cuz(OH)zC03

264 CUAlzOi

350 Cuz(0H)zCOs

341 Cuz(OH)zC03

277 CUAlzOi

a EOR, percent extent of reduction. Prepared without aluminum.

gave major and minor DSC peaks at 300 and 200 "C, respectively (curve 6).

The absence of the 300 "C peak in the DSC spectra of catalysts with low copper concentration confirms the absence of crystalline Cuz(OH)&03 species in precursors p-2 and p-3 containing 5.73 and 15.3 wt % copper, respectively. This seems to suggest that noncrystalline copper species are more stable than the crystalline copper species present only in catalysts with high copper con- centration. Thermal stability under this condition is characteristic of the dissolution of the species in the support (Ghiotti and Boccuzzi, 1987). This result is consistent with those from XRD, SEM, and EPMA measurements.

It is often assumed that all the copper hydroxy salts present in dried catalyst precursors would decompose on calcination at about 500 K (Ghiotti and Boccuzzi, 1987). However, the DSC results from our investigation show that higher temperatures are necessary for the decom- position of the above species especially in the X-ray amorphous area for coprecipitated Cu-A1 catalysts. It also shows clearly that thermal stability of these species depends strongly on copper concentration.

Thus, the DSC analysis shows that the minimum temperature for calcination is 300 OC. Details of the species still present in the catalysts after calcination a t various temperatures are given in Table 3.

specieso

M(OH)a (n); M(OH)COs (n) Al(OH)s (n); AI(OH)COs (n); Cuz(0H)zCOs (n) M(OH)s (n); M(OH)COs (n); CuA0H)zCOs (n) M(OH)s (n); M(OH)COs (n); Cu(0H)z (c); Cuz(0H)zCOs (b) M(OH)s (n); M(OH)COs (n); Cu(0H)z (c); Cud0H)zCOs (b) Cu(OH)z (c); Cuz(0H)zCOs (c)

le phases present. * Prepared without aluminum.

Calcined Catalysts, A. Temperat ureprogrammed Reduction of Unreduced Catalysts. The TPR spectra of unreduced but calcined catalysts given in Figure 4 were used in identifying the reducible species present. The peak temperatures and the corresponding species are summarized in Table 3.

The sharp peaks a t ca. 210 "C in Figure 4 represent CUZO. In our study, this species was observed for catalysts with copper concentration greater than or equal to 24 wt 5% and calcined at ca. 700 "C. Although Sivaraj and Kanta Rao (1988) have reported the presence of the CuzO species in catalysts prepared by deposition-precipitation and containing a high concentration of copper, the effects of calcination temperature were not reflected since their investigation was conducted for catalysts calcined only at 400 "C. Our results show that the presence of this species is a function of both copper concentration and calcination temperature.

The sharp peaks at ca. 262 "C represent CuO. This was observed in catalysts with copper concentrations greater than 24 wt % and calcined at or above the determined minimum temperature, and for lower copper concentra- tions of 15.3 and 5.79 wt 5% but calcined at or above 700 OC. However, the peak width in the latter catalysts is an indication of an interaction of the species with the support.

The TPR peaks at ca. 360 "C represent noncrystalline copper hydroxycarbonate. The sharp peaks in this category represent the noncrystalline hydroxycarbonate exclusively. This was observed in all catalyst compositions calcined at 300 OC. However, no CuO peak was observed for catalyst precursors p-2 and p-3 calcined at 300 "C. The absence of the CuO peak in the TPR spectra of these calcined catalysts confirms the absence of both crystalline Cuz(OH)zC03 and Cu(0H)z in their precursors.

The wide peaks at ca. 360 OC observed for all catalyst compositions calcined a t 500 "C indicate that, in addition to the noncrystalline copper hydroxycarbonate, there was also some degree of either interdiffusion of this species and the support or interaction between A1203 and the CuO obtained from the decomposition of the noncrystalline copper hydroxycarbonate. This interaction would lead to the formation of the spinel CuA1204. In general, it was observed that the greater the copper concentration, the more this peak approached the CuO peak. This shift could be taken as a measure of the progress in the decomposition of the noncrystalline copper hydroxycarbonate to CuO. As mentioned previously, this shows the effects of copper concentration on the thermal stability of noncrystalline copper hydroxycarbonate.

Also, the broadness of the peaks with decreasing copper concentration shows that species-support interaction or interdiffusion was greater a t lower copper loading. Higher copper concentration appeared to favor bulk CuO forma- tion.

The broad peaks observed after ca. 350 "C represent the spinel CuA1204. This species was found to be present in catalysts of all compositions calcined at ca. 700 OC. Also,

2052 Ind. Eng. Chem. Res., Vol. 33, No. 9, 1994

5.79 wt % copper

n 30

a Q

h

c) .cI

I

I Y d

.E 20 4 v

51 0 .r. Y

s" : 10 s :

0

100 200 300 400 500 600

Reduction Temperature (C)

15.3 wt R copper

A

100 200 300 400 500 600

Reduction Temperature (C)

Q 30

a a

c z 2 20 d

a" 5 10

:

Y d

Y

a 0 d c)

0 u

0

24.1 wt % copper

100 200 300 400 500 600

Reduction Temperature (C)

I 27.8 wt $6 copper

100 200 300 400 500 600

Reduction Temperature (C)

Figure 4. TPR spectra of calcined Catalysts. Calcination temperature ("(3): A, 300; B, 500, C, 700.

it was noticed from Figure 4 that the fraction of CuA1204 present in catalysts increased with decreasing copper concentration.

The TPR spectra of catalyst p-6 (copper only) calcined at 300 OC gave only one peak around 262 OC which is characteristic of CuO showing that all the crystalline copper hydroxycarbonate had decomposed into CuO even at a calcination temperature as low as 300 "C. This is consistent with the results from DSC experiments.

The TPR results show that, irrespective of copper concentration, all the Cuz(OH)zC03 decomposed into CuO upon calcination at 700 OC. However, a t this temperature, part of the CuO was found to combine with A1203 to form CuA1284.

The use of the TPR technique for species identification in impregnated CuO/A1203 Catalysts has been reported by Agaras et al. (1988) and Robinson and Mol (1988). These authors observed only CuO and CuAl204 species for various copper loadings and catalyst calcination temperatures. Species in coprecipitated Cu-A1 catalysts have not been previously fully identified. We have thus attempted to provide an exhaustive identification of the species present in these catalysts.

In this study, the evaluation of reducibility was based on the lowest TPR peak temperature. This is consistent with the initial reduction temperature suggested by Klisurski (1970). All the TPR peak temperatures are given in Table 3. However, considering the lowest peak tem- perature alone for each catalyst, it was observed that

Ind. Eng. Chem. Res., Vol. 33, No. 9, 1994 2053 Table 4. Summary of BET Analysis of Reduced and Unreduced Catalysts

catalyst

cu calcn BETsurface % contrib of mesopore concn temp area (m2/g) to pore vol (wt %) ("C) unreduced reduced unreduced reduced 0.00 300 144 a 89 b

500 133 a 90 b 700 111 a 88 b

5.79 300 133 141 90 97 500 129 129 90 98 700 106 102 90 99

15.3 300 121 115 90 99 500 116 109 90 97 700 93 106 90 98

24.0 300 98 93 84 97 500 93 91 85 98 700 84 107 87 98

27.8 300 93 89 89 99 500 92 88 84 99 700 76 96 88 95

57.33c 300 46 0.8 98 b 500 10 b b b 700 2 b b b

a Not reducible. * Not performed. Prepared without aluminum.

reducibility increased with copper concentration and calcination temperature. This relationship is due to the presence of different types and proportions of reducible copper species as well as their varying degrees of interaction or interdiffusion with aluminum species. These charac- teristics are controlled by catalyst composition and calcination temperature.

For similar reasons, catalysts with low copper concen- trations exhibited a greater increase in reducibility with calcination temperature than those with high copper concentration. Table 3 shows that catalysts with low concentrations of copper and calcined a t 300 "C contain noncrystalline Cuz(OH)2C03 as the reducible species with the lowest reduction temperature. On the other hand, in similar catalysts calcined at 700 "C, the copper species with the lowest reduction temperature is CuO. Thus, the difference in catalyst reducibility between calcinations at 300 "C and those at 700 "C for catalysts with low copper concentration is essentially the difference between the reduction temperatures of these two copper species. Similarly, the difference for the catalysts with high copper concentration reflects the difference between the reduction temperature of CuO obtained from crystalline Cu2(0H)2- COS and Cu(OH)2 decomposition by calcination at 300 "C and that of Cu20 obtained by calcination at 700 "C. Clearly, the difference is greater with catalysts of low copper concentration.

B. Temperature-Programmed Reduction of Re- duced Catalysts. TPR spectra were also obtained for reduced catalysts in order to verify the species still present as well as to determine the extent of reduction. The extent of reduction (EOR) was calculated as follows:

where EOR = extent of reduction, Cuu = reducible copper species in unreduced catalyst, CUR = reducible copper species in reduced catalyst, i = type of reducible copper species, and xi = weight fraction of i copper species. The reducible copper species was calculated from hydrogen consumption obtained in the TPR experiments. The reduction stoichiometry of the species was taken into consideration in the calculation as described by Varma et al. (1984). These results are summarized in Table 3. It was observed that the extent of reduction increased with copper concentration and calcination temperature.

The EOR results provide further evidence that there was a greater degree of interdiffusion of reducible species and A1203 with decreasing copper concentration.

C. BET Analysis. BET surface area, pore volume, and average diameter were measured for reduced and unreduced catalysts. The results for BET surface area are given in Table 4.

It was observed that while the BET surface area of unreduced catalysts decreased with increasing copper concentration and calcination temperature, that of reduced catalysts with high copper concentration exhibited a minimum with calcination temperature. The average pore diameters measured for both reduced and unreduced catalysts were between 13.3 and 18.8 nm. This shows that the pores were mostly in the mesopore size range. The mesopore size range used in all analyses was from 2 to 50 nm.

Pore size distribution was evaluated in terms of the percent contribution of mesopores to pore volume. The results are also presented in Table 4 for reduced and unreduced catalysts. It was observed from the table that mesopore contribution to pore volume was greater in reduced than unreduced catalysts,

Reported BET data for Cu-A1 catalysts have been few and quite limited in scope. They have also been restricted to unreduced but calcined catalysts (Sivaraj and Kanta Rao, 1988; Stephanopoulus et al., 1987; Friedman et al., 1978). In this study, extensive BET data have been collected and used to show more clearly the effects of copper concentration, calcination temperature, and iso- thermal reduction on the BET surface area and pore structure of coprecipitated Cu-AI catalysts.

D. Temperature-Programmed Desorption (TPD) of NH3 Analysis. Coprecipitated Cu-A1 catalysts have not been. previously characterized for acidity. Conse- quently, we decided to study the acid site distribution for these catalysts using the TPD technique. Effects of catalyst composition and calcination temperature on acid site distribution were investigated. Typical TPD spectra are presented in Figure 5a and Figure 5b for catalysts p-2 (low copper concentration) and p-5 (high copper concen- tration), respectively. Two types of desorption peaks were identified; the low-temperature peaks indicating the weak acid sites and the high-temperature peaks indicating the strong acid sites.

Also, it was observed that the peak temperatures of both the weak and strong acid sites decreased with increase in copper concentration. It was interesting to observe that the high temperature desorption peak (strong acid site) was practically absent for catalysts containing 24.1 wt 7% copper or more and calcined at or above 500 "C (see Figure 5b). Apparently, the high temperature desorption peaks disappeared for those calcined a t 500 "C but reappeared for those calcined at 700 "C in the case of catalysts containing less than 24.1 wt % copper (see Figure 5a). At this time, this behavior is not fully understood.

Similar type of work with alumina and silica-alumina has been performed by Amenomiya and Cvetanovic (19701, who observed drastic variations in acid site distribution with heating temperature, evacuation temperature, and catalyst composition. Although the mechanism is still unclear, these workers attributed this behavior to the variation in growth rates of the two types of acid sites with differing catalyst treatments.

2054 Ind. Eng. Chem. Res., Vol. 33, No. 9, 1994 20

h

Y .cI

m

U I

t'

-E d

e s a"

s" a

h Y - a 10

a U .LI

4

0

20

h

Y .I m

51 I

t' CI Y

2 4 v

10

e a"

;;

O m

0

51 .r(

8 4

0

5.79 wt I copper

100 200 300 400 S O 0 600 700 800

Desorption Temperature (C)

27.8 w t % copper

100 200 300 400 500 600 700 800

Desorption Temperature ( C ) Figure 5. TPD spectra of calcined catalysts. Calcination tempera- ture ( OC): A, 300; B, 500; C, 700.

E. Low-Temperature 0 2 Chemisorption Analysis. Results of copper surface area, dispersion, and average crystallite size of the reduced catalysts measured by the 0 2 chemisorption technique are presented in Table 5. It was observed that copper surface area exhibited a maxi- mum with copper concentration at 24.1 wt % copper for all three calcination temperatures. For each calcination temperature, however, the difference between the maxi- mum and minimum copper surface area increased with calcination temperature. Kaushik et al. (1991) measured the copper surface areas of Cu/A1203 catalysts prepared by deposition-precipitation and also reported a maximum with copper concentration.

Table 5. Summary of 0 2 Chemisorption Analyses of Reduced Catalysts.

catalyst copper av calcn copper surface % copper

crystallite temp concn area copper ("C) (wt %) (m2/g) dispersion size (nm) 300 5.79 0 NA NA

15.3 30.8 11.6 6.4 24.1 36.5 9.9 8.8 27.8 33.0 9.5 10.9

15.3 33.4 20.4 6.6 24.1 40.2 10.7 8.0 27.8 34.5 10.0 10.4

700 5.79 16.0 39.5 2.6 15.3 69.7 29.5 3.5 24.1 79.0 25.2 4.1 27.8 41.2 13.7 7.6

500 5.79 0 NA NA

NA, not applicable.

However, their results were limited to catalysts calcined at 400 "C. Table 5 shows that copper surface area increased with calcination temperature. These results for copper surface area are consistent with those of Matsukata et al. (1988) also for coprecipitated Cu-Al catalysts.

It was interesting to observe that, for the range of copper concentrations of 5.79-27.8 wt % , the relationship between BET surface area of calcined catalysts and metallic copper area reflected opposing responses of these characteristics to calcination temperature. As mentioned previously, the large extent of decomposition of various species when the catalyst is calcined at a high temperature results in a net increase in average pore size and a consequent decrease in BET surface area. On the other hand, high calcination temperature results in a large extent of catalyst reduction and greater interaction between copper and aluminum species to form a larger fraction of CuAl204. Both effects will lead to an increase in metallic copper area.

In general, copper dispersion increased with decreasing copper concentration and increasing calcination tempera- ture. Conversely, the average crystallite size was observed to decrease with decreasing copper concentration and increasing calcination temperature. Apparently, this average crystallite size will depend on what fraction of the total metallic copper obtained after reduction at 300 "C comes from the noncrystalline Cuz(OH)zCO3. As was shown previously, this fraction increases with calcination temperature and decreasing copper concentration. Mat- sukata et al. (1988), working with coprecipitated Cu-A1 catalysts, have reported an increasing degree of dispersion of copper with calcination temperature for catalysts containing up to 30 wt % copper. According to these authors, the variation of copper dispersion with calcination temperature exhibited a maximum at 500 OC for catalysts of higher copper concentrations. These results, together with those from our SEM, EPMA, TPR, and 0 2 chemi- sorption studies show that increase in calcination tem- perature leads to increased dispersion of the noncrystalline copper species on account of the formation of larger amounts of CuA1204. Also, the results show that the crystalline copper species undergo agglomeration because of the absence of A1203 support where the crystalline copper species are located. Since the opposing effect of the latter phenomenon increases with copper concentration, the difference in average copper crystallite sizes a t high and low calcination temperatures will be larger in catalysts with low copper concentrations.

Conclusions 1. The species contained in coprecipitated Cu-A1

catalysts using NazC03 for precipitation are crystalline Cu(0H)z and Cuz(OH)zCO3 and noncrystalline Cuz(OH)2- C03, Al(OH)3, and Al(OH)C03. Their proportions and thermal stabilities were found to depend on copper concentration.

2. Calcining these catalysts results in the formation of CuzO, CuO, Cuz(OH)zCOs, and CuAlz04 as the reducible species. The reducibilities and extents of reduction with hydrogen of the calcined catalysts were found to be strong functions of copper concentration and calcination tem- perature.

3. The acid site distribution on calcined catalystsvaried drastically with copper concentration and calcination temperature. 4. BET surface area and pore structure depended on

copper concentration and calcination temperature and, also, on whether the catalyst was reduced or unreduced.

5. Copper dispersion increased with calcination tem- perature and decreasing copper concentration.

Ind. Eng. Chem. Res., Vol. 33, No. 9, 1994 2055 Himelfarb, P. B.; Simmons, G. W.; Klier, K.; Herman, R. G. Precursors

of the copper-zinc oxide methanol synthesis catalysts. J. Catal.

Hurst, N. W.; Gentry, S. J.; Jones, A. Temperature programmed reduction. Catal. Rev.-Sci. Eng. 1982,24 (2), 233-309.

Isogai, N.; Takagawa, M.; Watabe, K.; Yoneoka, M.; Yamagishi, K. Catalyst composition for decomposition of methanol. European Patent 0 342 614, 1989.

Kaushik, V. K.; Sivaraj, C.; Kanta Rao, P. ESCA characterization of copper/alumina catalysts prepared by deposition-precipitation using urea hydrolysis. Appl. Surf. Sci. 1991, 51, 27-33.

Klier, K. Methanol synthesis. Adu. Catal. 1982,31, 243-313. Kliman, M. L. Methanol, natural gas, and the development of

alternative transportation fuels. Energy 1983,8, 859. Kliiurski, D. Relationship between the selectivity and activity of

oxide catalysts in the oxidation of methanol and the initial temperature of their reduction. Kinet. Katal. 1970,ll (l), 263- 264.

Kobayashi, H.; Takezawa, N.; Minochi, C. Methanol reforming reaction over copper containing mixed oxides. Chem. Lett. 1976,

Matsukata, M.; Uemiya, S.; Kikuchi, E. Copper-alumina spinel catalysts for steam reforming of methanol. Chem. Lett. 1988,761- 764.

Minochi, C.; Kobayashi, H.; Takezawa, N. The catalytic activities of methanol reforming catalysts and their preparations. Chem. Lett. 1979,507-510.

Othmer,D.F. Methanolis thebest wayto bringAlaskagastomarket. Oil Gas J. 1982, 80 (44), 84.

Periana, R. A.; Taube, 0. J.; Evitt, E. R.; Loffler, D. G.; Wentrcek, P. R.; Voss, G.; Masuda, T. A mercury-catalyzed, high-yield system for the oxidation of methane to methanol. Science 1993,259,340- 343.

Robinson, W. A. M.; Mol, J. C. Characterization and catalytic activity of copper/aluminamethanol synthesis catalysts. Appl. Catal. 1988, 44,165.

Sivaraj, C.; Kanta Rao, P. Characterization of copper/alumina catalysts prepared by deposition-precipitation using urea hy- drolysis, 1. Nitrous oxide decomposition and reaction of ethanol. Appl. Catal. 1988,45, 103.

Stephanopoulus, M.; Gavalas, G. R.; Jothimurugesan, K.; Lew, S.; Sharma, P. K.; Bagajewicz, M. J.; Patrick, V. ‘Detailed studies of novel regenerable sorbents for high temperature coal gasification”; Final report prepared for the US department of Energy, Agreement

Takezawa, N.; Kobayashi, H.; Hirose, A.; Shimokowabe, M.; Taka- hashi, K. Steam reforming of methanol on copper-silica catalysts; effects of copper loading and calcination temperature on reaction. Appl. Catal. 1982,4,127.

Uehara, I. R & D Program on Hydrogen Energy in Japan. Tenth Anniversary Conference of Hydrogen Industry Council, Kanan- askis, Alberta, Canada; Hydrogen Industry Council, 700, 4th Avenue S.W.: Calgary, Canada, 1992.

Varma, R. L.; Mathews, J. F.; Bakhshi, N. N. Characterization of manganese oxide-supported metal catalysts by temperature- programmed reduction. In Frontiers in Chemical Engineering; Doraiswamy, L. K., Mashelka, R. A., Eds.; Wiley Eastern Limited New Delhi, 1984, Vol. 11, pp 118-128.

Vasilevich, A. A.; Shpiro, G. P.; Alekseev, A. M.; Semonova, T. A.; Markina, M. I.; Vasil’eva, T. A,; Budkina, 0. G. A method of determining the surface of copper and some examples of its use for low-temperature catalysts for carbon monoxide conversion. Kinet. Katal. 1975, 16 (6), 1571-1577.

Yildirim, E. Optimum hydrogen utilization in upgrading and refining. Alberta Hydrogen Research Program Workshop, Calgary, Canada; Alberta Energy, Research and Technology; 1991; p 10.

1985,93,442-450.

1347-1350.

NO. DE-FC21-85MC22193; 1987.

Received for review December 27, 1993 Revised manuscript received April 29, 1994

Accepted May 31, 1994O

Acknowledgment The financial support provided by the Canadian Com-

monwealth Scholarship and Fellowship Plan, and the Natural Sciences and Engineering Research Council of Canada is gratefully acknowledged.

Literature Cited Agaras, H.; Carrella, G.; Larborde, M. A. Copper catalysts for the

steam reforming of methanol. Appl. Catal. 1988,45, 53-60. Amenomiya, Y.; Cvetanovic, R. J. Active sites for alumina and silica-

alumina as observed by temperature programmed desorption. J. Catal. 1970,18, 329.

Amphlett, J. C.; Mann, R. F.; Peppley, B. A.; Stokes, D. M. Preparation and characterization of three methanol steam reforming catalysts based on copper-magnesia with additions of manganese, chromium and vanadium oxides. Proc. Can. Symp. Catal. 1990, l l th , 392- 401.

Amphlett, J. C.; Mann, R. F.; Peppley, B. A.; Stokes, D. M. Some design considerations for a catalytic methanol steam reformer for a PEM fuel cell power generating system. Proc. Intersoc. Energy Convers. Eng. Conf 1991,26th (3) 642-649.

Bhatia, S.; Bakhshi, N. N.; Mathews, J. F. Characterization and methanation activity of supported nickel catalysts. Can. J. Chem. Eng. 1978,56, 575-581.

Cheng, W. H. Carbon dioxide calcination of methanol dissociation catalysts. US Patent 4826798,1989.

de Bokx, P. K.; Balkernende, A. R.; Geus, J. W. The mechanism and kinetics of methane formation by decomposition of methanol on a Ni/SiOz catalyst. J. Catal. 1988,117,467-484.

Fierro, J. L. G. Chemisorption of probe molecules. In Spectroscopic characterization of heterogenous catalysts; Studies in Surface Science and Catalysis 57B; Fierro, J. L. G., Ed.; Elsevier: New York, 1990, 1-66.

Friedman, R. M.; Freeman, J. J.; Lytle, F. W. Characterization of Cu/A1203 catalysts. J. Catal. 1978,55, 10.

Ghiotti, G.; Boccuzzi, F. Chemical and physical properties of copper based catalysts for CO shiftreactionandmethanolsynthmis. Catal. Rev.-Sci. Eng. 1987, 29 (2 & 3), 151-182.

Hickman, D. A.; Schmidt, L. D. Production of syngas by direct catalytic oxidation of methane. Science 1993,259, 343-346.

Hidalgo, C. V.; Itoh, H.; Hattori, T.; Niwa, M.; Murakami, Y. Measurement of acidity of various zeolites by temperature- programmed desorption of ammonia. J. Catal. 1984,85,362-369. @ Abstract published in Advance ACSAbstracts, July 15,1994.