4682
Ind. Eng. Chem. Res. 1998, 37, 4682-4688
Phase Transformation and Activity of Cadmium Sulfide Photocatalysts for Hydrogen Production from Water: Role of Adsorbed Ammonia on Cadmium Sulfate Precursor Namita Sahu, Manjit K. Arora, S. N. Upadhyay, and A. S. K. Sinha* Department of Chemical Engineering and Technology, Institute of Technology, Banaras Hindu University, Varanasi 221 005, India
Adsorption-desorption of ammonia on cadmium sulfate (3CdSO4‚8H2O) has been studied. Its role in formation of cubic phase of CdS when cadmium sulfate is reacted with H2S has been investigated in detail using X-ray diffraction, electron diffraction, infrared spectroscopy, temperature-programmed desorption of NH3, and thermogravimetric analysis. Three distinct acidic sites on cadmium sulfate where ammonia can get chemisorbed have been reported. The heats of adsorption for the three sites are 24.3, 26.9, and 42.6 kcal mol-1, respectively. The origin of surface acidity is related to the formation of transition intermediates during the dehydration of 3CdSO4‚8H2O. Both Brønsted and Lewis acidities have been observed. The reaction of ammonia preadsorbed cadmium sulfate with H2S gas leads not only to the formation of larger crystallites but also to the cubic phase of CdS which has lower photocatalytic activity for hydrogen production from water. Ammonia can be completely desorbed from the surface of cadmium sulfate if it is heated above 600 K. Introduction Hydrogen has the potential to become an ideal source of energy for the future. It can be produced by photocatalysis of water using solar radiation. Cadmium sulfide is one of the suitable photocatalysts.1-4 In general, supported CdS exhibits better activity because of its high dispersion on the surface of the support. The support also provides heterojunction for photogenerated electrons and holes5,6 and restricts their recombination. Thus, poor charge separation due to the inadequate band bending in fine dispersion of photocatalysts because of smaller crystallite size7 is compensated. Earlier work carried out in our laboratory by Arora et al.8 showed that dispersion as well as distribution of CdS on the support is of critical importance, and these are strongly affected during preparation stage of the catalyst. They have further reported that alumina-supported CdS photocatalyst, prepared by impregnation of alumina hydrogel by ammoniacal solution of cadmium sulfate with subsequent drying and reaction with H2S gas at 473 K, exhbited a superior activity. This catalyst had high degree of dispersion of CdS, and also it was preferentially distributed on the surface of alumina. It was concluded that since light cannot penetrate deep inside the porous matrix of the support, a highly active CdS photocatalyst should have the active ingredient distributed only on or near the external surface of the support. It was also observed that the activity of this catalyst prepared using an ammoniacal solution of cadmium sulfate increased further when, prior to reaction with H2S, the dried mass was purged with N2 gas for 2 h during its preparation stage. Though they did not study adsorption of ammonia on cadmium sulfate, it was concluded that the observed difference in activities of catalysts prepared with and without purging with N2 gas, respectively, was due to residual ammonia remaining chemisorbed on cadmium sulfate during its reaction with H2S gas.
The present work reports the study on the adsorption-desorption of ammonia on cadmium sulfate, its role in the formation of different phases of CdS, and consequently its effect on the photocatalytic activity of alumina-supported CdS. It has been observed that CdSO4 has three different sites where ammonia can adsorb. The acidic sites are due to strain in crystallites of cadmium sulfate resulting from its partial dehydration. Ammonia, if chemisorbed, on the surface of CdSO4 during its reaction with H2S gas, results in the formation of the less active cubic phase of CdS and also the larger crystallites of CdS. Experimental Section A. Catalyst Preparation. Alumina-supported cadmium sulfide catalysts, prepared using the impregnation technique by Arora et al.,8 were used in the present study. The details of catalyst preparation are given below. Catalyst 1. A freshly prepared alumina hydrogel was impregnated with a requisite volume of 0.1 M ammoniacal solution of 3CdSO4‚8H2O to yield the final composition as CdS:Al2O3 ) 1:2 by weight. The mixture was kept stirred for 20 h. It was then dried over a water bath with continuous stirring and finally in an air oven at 383 K for 12 h. The dried granules (2 mm diameter) were packed in a tubular reactor and reacted with H2S at 473 K for 5 h. Pure H2S gas was allowed to flow over the granules at a very low flow rate. After the reaction, the granules were crushed to -200 mesh size and washed with distilled water to remove any unreacted cadmium sulfate. The catalyst was finally dried at 383 K in an air oven. High surface area alumina gel was prepared by neutralizing a concentrated solution of aluminum nitrate with ammonia at pH 9 and temperature 303 K.9 The hydrogel was aged for 48 h and then washed with distilled water until it became free of nitrate ions.
10.1021/ie980237s CCC: $15.00 © 1998 American Chemical Society Published on Web 10/30/1998
Ind. Eng. Chem. Res., Vol. 37, No. 12, 1998 4683
Catalyst 2. This catalyst was prepared by the same procedure as employed for catalyst 1. The only difference was that, prior to the reaction of cadmium sulfate with H2S, the dried granules were flushed with nitrogen for 2 h at 473 K to remove the residual ammonia chemisorbed on the surface. Catalyst 3. The preparation technique of this catalyst was the same as that of the catalyst 2 except that the reaction of CdSO4 with H2S was carried out at an elevated temperature of 673 K. B. Activity Measurement. The activities of the catalysts, as milliliter of hydrogen produced (at NTP) per gram of CdS in 2 h of irradiation, were measured in a batch photoreactor. The reactor was a 250 mL flatbottomed flask; one of its sides was also made flat to permit the entry of light though a plane wall. The reactor had a provision for measurement and control of pH and temperatue. A 150 W Philips tungsten halogen lamp was used as the light source. The choice of the light source was based on the availability as well as on the spectral characteristics of the emitted light. No UV or IR filter was used because the spectrum of the light used showed a very negligible emission in the UV range and use of pyrex glass reactor further prevented this radiation from reaching the catalyst. IR radiation can only be absorbed in the form of heat and could increase the temperature of the solution. During experimentation, the temperature could be maintained at the desired value; therefore, an IR filter was also not used. Two grams of catalyst (-200 mesh) was suspended with a magnetic stirrer in 250 mL of aqueous solution of concentration 0.01 and 0.004 M with respect to Na2S and Na2SO3, respectively. The pH was maintained at 8.6 during the experimentation by adding requisite quantities of NaOH and acetic acid. The temperature was maintained at 333 K, and the solution was deaerated by sparging with nitrogen for 2 h prior to irradiation. The gas evolved was collected by water displacement technique and analyzed by an on-line gas chromatograph using a 5 Å molecular sieve column and a thermal conductivity detector using nitrogen gas as carrier. Before entering the chromatograph, the evolved gas was passed through a cold trap to remove entrained moisture. Comparison of the retention time of the only peak that appeared on the chromatogram with the standard confirmed that the gas was only hydrogen. The reproducibility of data was tested in 3 runs for each catalyst. No significant deviation was observed and the data reported represent the average value. Intrinsic activity of a photocatalyst is largely independent of pH and temperature, since negligible changes in electronic band structure of a semiconductor are expected. Therefore, the activity of various catalysts were compared at constant pH and temperature. The optimum value of pH depends on the electyrolytes present in the solution. In the present study, with a mixture of sodium sulfide and sulfite as electrolytes, the optimum pH was observed to be 8.6 and the activities of all of the photocatalysts have been compared at this pH. Further, the intensity of light was also not varied in the present study. However, experiments were conducted by varying the amount of catalysts in the reactor and it was observed that the intensity of light was not a limiting factor. C. Catalyst Characterization. X-ray diffraction (XRD) studies were carried out at room temperature using a Philips 1710 X-ray diffractometer with Cu KR
Table 1. Hydrogen Evolved during 2 h of Irradiation catalyst
hydrogen evolved (mL at NTP g CdS-1)
1 2 3
7.0 8.1 6.1
target of radiation wavelength 1.542 Å. Electron diffraction studies were carried out on a Philips (CM-12) transmission electron diffractometer equipped with a twin lens system. The IR spectra at room temperature were obtained using JASCO double-beam infrared spectrophotometer. About 1 mg of the sample and 100 mg of KBr were finely ground together and pressed under high pressure. Adsorption-desorption of ammonia was studied using Micromeritics pulse chemisorb 2705 with temperatue-programmed desorption facilities. The samples were first heated in situ at the desired temperatures in flowing N2 for 5 h. Adsorption of ammonia on preheated samples was performed at 333 K by repeatedly injecting pulses of pure ammonia onto the sample until saturation was observed by the detector and finally desorption was carried out at the desired heating rates. The thermogravimetric analysis was carried out on a Stanton Redcroft thermal analyzer STA 780 series. Results and Discussion Activity. The activities of the catalyst for hydrogen production are given in Table 1. It is observed from the table that, in spite of having the same composition, the activities of catalysts are widely different. Catalyst 2 showed superior activity in comparison to catalyst 1. The only difference in the preparation of these catalysts was that, prior to the reaction with H2S, catalyst 2 was flushed with nitrogen gas for 2 h at 473 K to remove the chemisorbed ammonia. Catalyst 3 also showed lower activity compared to catalyst 2. Catalyst 3 was prepared by the same procedure as catalyst 2, except that the reaction of cadmium sulfate with H2S was carried out at a higher temperature of 673 K. XRD Studies. XRD analyses of the catalysts were carried out for the identification of phases and the estimation of crystallite size of CdS. The interplanar spacing (d value) for each peak was calculated, and the peaks were identified as those of CdS, γ-Al2O3 and different phases of aluminum hydroxide. The hexagonal phase of CdS was largely present in all of the catalysts. However, a small peak corresponding to d ) 2.054 Å, was also observed in all catalysts. This peak could be matched with that of either the cubic phase of CdS or aluminum hydroxide. Electron diffraction studies, as discussed below, revealed that cubic CdS was present in the catalysts. Therefore, this peak (d ) 2.054 Å) was assigned to that of cubic CdS. Other peaks of cubic CdS could not be detected because of their low intensities and overlapping by peaks of aluminum hydroxide. A comparison of the peak intensity of cubic CdS observed at d ) 2.054 Å in all three catalysts showed that this phase was substantially less in catalyst 2. In this catalyst, the solid mass was flushed with nitrogen for 2 h at 473 K during the preparation stage to remove the residual ammonia before reacting with H2S gas. The intensity of this peak further diminished in catalyst 3 which was prepared by reacting cadmium sulfate with H2S at a higher temperature of 673 K. It has been shown below that desorption of ammonia from the surface of cadmium sulfate takes place in the temper-
4684 Ind. Eng. Chem. Res., Vol. 37, No. 12, 1998 Table 2. X-ray Diffraction Results of Samples 1 and 2 experimental resultsa
JCPDS file hexagonal phase
cubic phase
sample 1
sample 2
d (Å)
I/Io
hkl
d (Å)
I/Io
hkl
d (Å)
I/Io
indexing
d (Å)
I/Io
indexing
3.58 3.36 3.16 2.45 2.068 1.898 1.791 1.761 1.731 1.679 1.584 1.520 1.398
75 60 100 25 55 40 18 45 18 4 8 2 16
100 002 101 102 110 103 200 112 201 004 202 104 203
3.36 2.90 2.058 1.753 1.680 1.453 1.337 1.298 1.186 1.120
100 40 80 60 10 20 30 10 30 30
111 200 220 311 222 400 331 420 422 333
3.584 3.359 3.161 2.447 2.067 1.898 1.789 1.761 1.731 1.680 1.579 1.518 1.399
71 63 100 18 58 36 6 37 14 2 3 1 8
H(100) H(002) H(101) H(102) H(110) H(103) H(200) H(112) H(201) H(004) H(202) H(104) H(203)
3.577 3.348 3.154 2.957 2.489 2.065 1.896 1.776 1.760 1.726 1.682 1.559 1.396
44 83 46 10 4 100 24 20 90 24 9 4 14
H(100) H(002), C(100) H(101) C(200) H(102) H(110), C(220) H(103) H(200) H(112), C(311) H(201) H(004), C(222) H(202) H(203), C(400)
a
H and C indicate hexagonal and cubic phases, respectively.
ature range of 400-500 K. Therefore, the appearance of the cubic phase of CdS in catalysts is attributed to the ammonia remaining chemisorbed on cadmium sulfate during its reaction with H2S. Purging of the solid mass with N2 gas or carrying out the sulfurization reaction with H2S at higher temperatures results in removal of ammonia from the surface as in catalysts 2 and 3, respectively. Two different forms of CdS, i.e., face-centered cubic (zinc blende structure) and hexagonal close pack (wurtzite structure), are known to exist. In the cubic structure, both the S and Cd atoms form individual fcc structure. Whereas in the other form, the positions of S atoms are hexagonal close pack. So far as next nearest neighbors are concerned, both of the structures are similar with each atom being surrounded at the corners of a regular tetrahedra by atoms of another kind. There is, however, a small difference with respect to atoms further away. In the wurtzite structure, there exists a lower possibility of interaction between atoms of the same kind. Therefore, greater ionization is more favored with the wurtzite structure. Hence, it appears that ammonia chemisorbed on cadmium sulfate reduces the ionization of atoms and, therefore, cubic (zinc blende) structure is also observed in the catalysts. When ammonia is desorbed from the surface, the fraction of cubic structure decreases. Two samples of unsupported CdS were also prepared for further study on the role of ammonia. Sample 1 was prepared by drying of an ammoniacal solution of 3CdSO4‚ 8H2O on a water bath and subsequently in an air oven at 393 K for 12 h. The dried mass was then kept in flowing N2 gas at 498 K for 5 h to remove chemisorbed ammonia from the surface of the dried cadmium sulfate. Finally, it was reacted with H2S gas at 498 K in a tubular reactor for 2 h. Sample 2 was also prepared in exactly the same manner; however, it was not purged with N2 and the reaction with H2S was carried out initially for 2 h at room temperature which was subsequently raised to 498 K to complete the reaction. The reaction was carried out at the room temperature to prevent desorption of ammonia from the surface. The samples were analyzed by XRD, and the results are shown in Table 2. It is clearly seen that the d values and relative intensities of all of the peaks of sample 1 exactly match with the standard JCPDS file for the hexagonal phase of CdS. However, the d values of sample 2 could be matched with both hexagonal and cubic phases of CdS. Due to the broadness of the peaks
and because of the fact that the d values of hexagonal and cubic phases of CdS lie very near to each other, the observed peaks could not be deconvoluted, but the peak at d ) 2.90 Å, which clearly corresponds to the cubic CdS, is observed only in sample 2. However, the intense peaks at d ) 3.16 and 2.45 Å in sample 2, which correspond to hexagonal phase of CdS, clearly indicate that as in sample 1 the hexagonal phase of CdS is the major component of sample 2. But in contrast to sample 1, sample 2 also contained the cubic phase of CdS where ammonia remained chemisorbed on the surface of cadmium sulfate during its reaction with H2S gas. Crystallite size distribution and mean crystallite size of the catalysts were calculated by analyzing the [110] peak of CdS by the Warren-Averbatch technique.10 This peak was selected for the analysis because it did not suffer from interference from adjacent peaks of aluminum hydroxide. The peak was expanded into a Fourier series. The shift and asymmetry in the peak and consequently strain in the catalysts were observed to be negligible. A highly crystalline sample of CdS, i.e., CdS heated at 873 K for 12 h in H2S gas, was taken as the reference. The mean crystallite sizes were calculated to be 13.8, 6.8, and 14.8 nm for catalysts 1, 2, and 3, respectively. It is seen that the crystallite size of catalyst is dependent on the preparation technique. Catalyst 2 has lower crystallite size than catalyst 1, where cadmium sulfate is reacted with H2S gas in presence of residual ammonia on the surface. It is further observed that catalyst 3 has the largest crystallite size. This catalyst was prepared by the same technique as catalyst 2 but at an elevated temperature of 673 K. Crystallite size distributions of various catalysts are shown in Figure 1. The distribution for each one of the catalysts is skewed toward the larger crystallite size. This is indicative of the grain growth of CdS via sintering during its preparation stage. The extent of sintering is seen to depend on the temperature (catalyst 3) and also on the presence of ammonia during the reaction of cadmium sulfate with H2S gas (catalyst 1). Though sintering is primarily governed by the temperature, the chemical environment may also affect the process. The role played by the environment on the sintering process has been elaborated by Butt and Petersen.11 Therefore, on the basis of above results, it is concluded that even though in catalyst 3 the fraction of cubic phase of CdS was the least, its activity is low because of much
Ind. Eng. Chem. Res., Vol. 37, No. 12, 1998 4685
temperatures. The peaks, which appear at ∼3240 and ∼3320 cm-1, are due to symmetric and asymmetric stretching of NH3 bonded via the electron lone pair at the nitrogen atom to a Lewis acidic site.15 The band at ∼1400 cm-1 is characteristic of NH4+. Therefore, presence of Brønsted acidity is also revealed by the formation of protonated ammonium ions. It is observed from the figure that intensities of all of the peaks due to chemisorbed ammonia diminish in the sample which was heated at 498 K. Takeshita et al.16 have postulated that the Brønsted acidity may arise from two sources by a second order interaction. One is from the water coordinated as Hδ+ M2+
O H
Figure 1. Crystallite size distribution of catalysts.
larger crystallites and consequently reduced surface area. The lower activity of catalyst 1, compared to that of catalyst 2, is due to the formation of cubic phase of CdS and also because of the appearance of larger crystallites of CdS due to the presence of residual ammonia during reaction of CdSO4 with H2S. It is relevant to add that the cubic phase of CdS has been reported12 to have a lower photocatalytic activity. Electron Diffraction. Electron diffraction studies were carried out to confirm the presence of cubic phase of CdS in catalyst 1 and 2. The d values were calculated by the standard formula13 and compared with the ASTM cards for hexagonal and cubic patterns. The results are given in Table 3. For comparison, the results of electron diffraction studies of catalyst 4 are also given in this table. Catalyst 4 was prepared by exactly the same procedure as catalyst 2 but an aqueous rather than ammoniacal solution of cadmium sulfate was used for impregnation of alumina gel. It is observed that in catalyst 4 electron diffraction patterns exactly match with that of hexagonal phase of CdS. However, the calculated d values for catalysts 1 and 2 could be matched with hexagonal as well as cubic patterns of CdS. This confirms the earlier drawn conclusion that the catalysts prepared using ammoniacal solution of cadmium sulfate contain both hexagonal as well as cubic phases. However, as discussed earlier, the intensities of the XRD peaks indicated that the concentration of the cubic phase decreased when, prior to the reaction with H2S, (catalyst 2 and 3), the solid mass was purged with N2 to remove the adsorbed ammonia (catalyst 2) or in addition when higher temperature was used during the preparation stage of catalysts (catalyst 3). IR Spectroscopy. Samples of cadmium sulfate for IR spectroscopy were prepared by drying its ammoniacal solution on a water bath at 323 K and finally in an oven at desired temperatures, i.e., 323, 393, and 498 K, respectively, for 24 h. During analysis, exposure of samples to atmosphere was carefully avoided. The IR spectra are shown in Figure 2. The peak positions were accurately located and matched with standard values of probable peaks. The spectra as a whole consist of those corresponding to H2O, SO42-, NH3, and NH4+. The peaks which are observed at wave numbers ∼1600 and ∼3520 cm-1 have been assigned to water of hydration.14 It is clearly observed that the intensity of 1600 cm-1 peak diminished when sample was subjected to higher
The other type of the Brønsted acidity may be due to the surface water acidified by a neighboring positive Lewis acid center through an inductive or field effect. The origin of the Brønsted acidity in the catalysts prepared in the present study may be ascribed to the above reasons. Origin of the Lewis acidity in metal sulfates is discussed in the next section. It is also observed from the IR spectra that the peaks due to SO42- ions has a triplet corresponding to wavenumbers 1210, 1120, and 980 cm-1. However, this peak becomes sharp and the triplets are not clearly observed when the sample is progressively heated at higher temperatures. Hydration and crystal symmetry are known to markedly influence the spectra of sulfates. Therefore, the above observation may be attributed to greater coordination of Cd2+ and SO42- owing to the dehydration of 3CdSO4‚8H2O. TPD Studies. Figure 3 shows the TPD profiles of ammonia adsorbed at 323 K on 3CdSO4‚8H2O preheated at different temperatures for 6 h. The temperature during TPD runs was set to increase from 323 to 500 K at a rate of 5 K per 60 s. The TPD profile of cadmium sulfate which was earlier subjected to heat treatment at 500 K is seen to have three distinct desorption peaks at 400, 460, and 500 K, respectively. It indicates three different type of sites (Type I, II, and III, respectively) for adsorption of NH3 on the surface of cadmium sulfate. In contrast, for cadmium sulfate subjected to high temperatures of 575 K and above, the lower temperature peak, i.e., type I, is not observed and also the highest temperature peak (type III) becomes more prominent. Further, the sample preheated to 1000 K shows a very broad desorption profile. The total acidity (mmol of NH3/g of 3CdSO4‚8H2O) was also calculated and is reported in Table 4. It is observed that the acidity of cadmium sulfate increases with heat treatment, and for cadmium sulfate heated at 650 K, the total acidity is maximum at 0.837 mmol/g of 3CdSO4‚ 8H2O. However, a further increase in the pretreatment temperature of cadmium sulfate adversely affects its acidity. Cadmium sulfate heated at 1000 K has a very low total acidity of 0.074 mmol NH3/g of 3CdSO4‚8H2O, which is even less than 10% of the maximum acidity observed for the sample heated at 650 K. It is known that, on activation, the surface of metal sulfates show acidic property. This activation arises from processes involving heat, compression, or irradiation to induce some imperfection on the otherwise regular crystal surface. Tanabe and Takeshita17 have discussed the acidic property of nickel sulfate. They
4686 Ind. Eng. Chem. Res., Vol. 37, No. 12, 1998 Table 3. Electron Diffraction Results of Various Catalysts d values (Å) for CdSa standard experimental
(fcc)
zone axisb
(hcp)
catalyst
d1
d2
d3
d1
d2
d3
d1
d2
d3
(fcc)
(hcp)
1
2.068
3.361
1.753
2.068
1.752
2.068
1.355
3.365 (002) 2.068 (110) c
1.761 (1 h 12) 1.756 (112) c
(001)
3.361
2.068 (110) 3.365 (200) c
(112)
1
(110)
c
2
2.068
3.361
1.753
1.876
1.301
(010)
2.521
2.987
1.368
c
c
c
c
(100)
4
3.200
4.161
1.731
c
c
c
1.761 (1 h 12) 1.398 (203 h) 1.398 (02 h 3) 1.731 (201)
c
4
3.365 (002) 1.898 (103 h) 3.160 (01 h 1) 3.580 (100)
(110)
3.581
2.068 (110) 3.580 (100) 2.450 (01 h 2) 3.160 (101)
(112)
2
1.753 (31 h1 h) 1.753 (3 h1 h 1) 1.337 33 h 1) 1.753 (31 h1 h) c
(110)
3.362
3.360 (111 h) 2.058 (2 h2 h 0) 2.058 (22 h 0) 3.360 (111 h) c
(112)
1
2.058 (22 h 0) 3.360 (1 h 11) 3.360 (11 h 1) 2.058 (22 h 0) c
c
(010)
a d values corresponding to r , r , and r . b fcc and hcp indicate face-centered cubic and hexagonal close pack, respectively. c No 1 2 3 corresponding standard d values.
Figure 2. IR spectrum of ammonia adsorbed on cadmium sulfate.
have reported that on heating the surface the acidity of nickel sulfate increased with loss of water and reached a maximum and then dropped down sharply until conversion to the anhydrous structure was complete. Since the amount of water was the critical factor for maximum acidity, they studied the structural nature of nickel sulfate with a special reference to the role of water and proposed that the acid centre was an empty orbital on nickel ion which appeared in an incompletely dehydrated metastable transition structure. They discussed that though the structure was strained and unstable, the crystal network and presence of water helped in its stabilization. In the present study also, the surface acidity of cadmium sulfate seems to be generated by its dehydration due to heating. Total acidity is related to transition intermediates during the process. Thermogravimetric Analysis. The structure of cadmium sulfate (3CdSO4‚8H2O) has been reported in the literature. It has two sets of Cd2+ ions with slight different environments, all are octahedrally surrounded by two H2O molecules and four sulfate oxygens. There
Figure 3. Desorption of NH3 from 3CdSO4‚8H2O preheated at temperatures as shown in the figure. Table 4. Acidity of Cadmium Sulfate pretreatment temp (K) of 3CdSO4‚8H2O
total acidity (mmol of NH3/g of 3CdSO4‚8H2O)
500 575 650 800 1000
0.252 0.483 0.832 0.263 0.074
are four kinds of crystallographically nonequivalent H2O molecules. Three quarter of H2O molecules are attached to Cd2+ ions and one quarter of them has two other water molecules and two sulfate oxygen atoms as neighbors. The dehydration of 3CdSO4‚8H2O was studied in the present study by thermogravimetric analysis. Cadmium sulfate was heated in nitrogen gas, and the
Ind. Eng. Chem. Res., Vol. 37, No. 12, 1998 4687
Figure 4. Dehydration profile of cadmium sulfate (3CdSO4‚ 8H2O).
temperature was set to increase at a rate of 10 K per 60 s. The rate of weight change is shown in Figure 4. Two peaks at 523 and 593 K are distinct. Therefore, it is concluded that the dehydration of 3CdSO4‚8H2O is a two-step process, becoming anhydrous above 600 K. A comparison of the areas of the two dehydration peaks indicates that the lower temperature peak corresponding to 523 K is due to the removal of one quarter of H2O molecules from the crystal which are not directly attached to Cd2+. Similarly the 593 K peak corresponds to the removal of three quarters of H2O molecules directly attached to Cd2+. It is further observed that though the dehydration of 3CdSO4‚8H2O becomes complete at 600 K, as discussed earlier, the total acidity continued to increase until 650 K. In this context the work reported by Coing-Boyal18 is relevant. He has reported that anhydrous cadmium sulfate has two crystal forms: an R form which appears when the salt becomes anhydrous and a β form which is stable above 1073 K. Therefore, it appears that on heating 3CdSO4‚ 8H2O above 600 K, though it becomes anhydrous, its transition from R to β phase induces strain and consequently acidity in the crystallites of anhydrous CdSO4. The strain and the strain-induced acidity appears to be highest in samples heated at 650 K. Hence, it may be concluded that dehydration of 3CdSO4‚8H2O and transformation of anhydrous CdSO4 from R to β form both affect the total acidity of cadmium sulfate. Heats of desorption of ammonia from sites I, II, and III were calculated applying the equation given by Cvetanovic´ and Amenomiya.19 For heating rates of 0.5, 1.0, 2.0, and 5.0 K per 60 s, the temperatures for maxima in peaks (Tm) for all of the sites were observed. Plots of 2 log Tm - log β vs 1/Tm for the three sites are shown in Figure 5 where β is the heating rate (K s-1). The heat of desorption for sites I, II, and III, calculated from slope of the respective line, were 24.3, 26.9, and 42.6 kcal mol-1, respectively. Conclusions Surface acidity is generated as a result of dehydration when cadmium sulfate is heated to higher temperatures. Acidity is related to strain in the cadmium sulfate due to the formation of the transition intermediates during the heating process. The stable β form of anhydrous cadmium sulfate exhibits negligible surface acidity. The heating process generates both the Brønsted and the Lewis acidic sites on the surface of cadmium sulfate. Ammonia chemisorbs on the Brønsted acidic sites as NH4+ and on the Lewis acidic sites
Figure 5. Plot of 2 log Tm - log β vs 1/Tm for acidic sites of 3CdSO4‚8H2O.
through the electron lone pair at the nitrogen atom. The temperature-programmed desorption studies reveal that cadmium sulfate has three distinct types of site for chemisorption of ammonia with heats of desorption of 24.3, 26.9, and 42.6 kcal mol-1, respectively. The reaction of cadmium sulfate with H2S gas in presence of chemisorbed ammonia results in the formation of larger crystallites of CdS and also a fraction of CdS appears in cubic structure and consequently lower photocatalytic activity of CdS. Only a partial removal of chemisorbed ammonia is possible by purging with N2 gas. On the other hand, chemisorbed ammonia can be completely removed at temperatures above 600 K but this also leads to the grain growth via sintering. Acknowledgment The authors thank Professor Dr. O. N. Srivastava, Department of Physics, Banaras Hindu University, Varanasi, India, for providing facilities to carry out X-ray diffraction and electron diffraction studies. The authors are also thankful to the Ministry of NonConventional Energy Sources, New Delhi, India, for the financial support for the project. One of the authors (M.K.A.) acknowledges the financial support in form of fellowship received from the Banaras Hindu University. Literature Cited (1) Darwent, J. R.; Porter, G. Photo-chemical Hydrogen Production Using Cadmium Sulfide Suspension. J. Chem. Soc., Chem. Commun. 1981, 4, 145. (2) Matsumura, M.; Saho, Y.; Tsubomura, H. Photocatalytic Hydrogen Production from Solutions of Sulfite Using Platinized Cadmium Sulfide Powder. J. Phys. Chem. 1983, 87, 3807. (3) Borrell, L.; Cervera-March, S.; Gimenez, J.; Simarro, R. A. Comparative Study of CdS based Semiconductor Photocatalysts for Solar Hydrogen Production from Sulphide + Sulphite Substrates. Sol. Energy Mater. Sol. Cells 1992, 25, 24. (4) De, G. C.; Roy, A. M.; Bhattacharya, S. S. Effect of n-Si on the Photocatalytic Production of Hydrogen by Pt-loaded CdS and CdS/ZnS Catalyst. Int. J. Hydrogen Energy 1996, 2 (1), 19. (5) Borgarello, E.; Serpone, N.; Gra¨tzel, M.; Pelizzetti, E. Photodecomposition of H2S in Aqueous Alkaline Media Catalyzed by RuO2 - loaded Alumina in the Presence of Cadmium Sulfide. Inorg. Chem. Acta 1986, 112, 197. (6) Kobayashi, J.; Kitaguchi, K.; Tsuiki, H.; Ueno, A. Photogeneration of Hydrogen from Water Over an Alumina Supported ZnS-CdS Catalyst. J. Chem. Soc., Faraday Trans. 1 1987, 83, 1395. (7) Wolkenstein, T. The Electron Theory of Semiconductors. Adv. Catal. 1960, 13, 189.
4688 Ind. Eng. Chem. Res., Vol. 37, No. 12, 1998 (8) Arora, M. K.; Sinha, A. S. K.; Upadhyay, S. N. Effect of Dispersion and Distribution on Activity of Alumina Supported Cadmium Sulfide Photocatalysts for Hydrogen Production from Water. Ind. Eng. Chem. Res. 1998, 37 (4), 1310. (9) Lippens, B. C.; Steggerda, J. J. Active Alumina. In Chemical Aspects of Adsorbents and Catalysts; Linsen, B. G., Ed.; Academic Press: New York, 1970. (10) Wagner, C. N. J. In Local Atomic Arrangements Studied by X-ray Diffraction; Cohen, J. B., Hilliard, J. E., Eds.; Gordon and Breach Scientific: Newark, NJ, 1966. (11) Butt, J. B.; Petersen, E. E. In Activation, Deactivation and Poisoning of Catalysts; Academic Press: New York, 1988. (12) Matsumura, M.; Furukawa, S.; Saho, Y.; Tsubomura, H. Cadmium Sulfide Photocatalyzed Hydrogen Production from Aqueous Solution of Sulfite: Effect of Crystal Structure and Preparation Method of the Catalyst. J. Phys. Chem. 1985, 89, 1327. (13) Thomas, G. Transmission Electron Microscopy of Metals; John Wiley & Sons: New York, 1962. (14) Rao, C. N. R. Chemical Applications of Infrared Spectroscopy; Academic Press: New York, 1963.
(15) Kung, H. H. Transition Metal Oxides: Surface Chemistry and Catalysis. Stud. Surf. Sci. Catal. 1989, 48, 72. (16) Takeshita, T.; Ohnishi, R.; Matsui, T.; Tanabe, K. Acid Property and Structure of a Solid Metal Sulfate Catalyst. Changes in Structure of Nickel Sulfates with Heating. J. Phys. Chem. 1965, 69 (12), 4077. (17) Tanabe, K.; Takeshita, T. Catalytic Activity of Solid Metal Sulfates. Adv. Catal. 1967, 17, 315. (18) Coing-Boyal, J. Crystallographic Data on the Two Forms of Anhydrous Cadmium Sulfate. Compt. Rend. 1961, 253, 997. (19) Cvetanovic´, R. J.; Amenomiya, Y. Application of a Temperature-Programmed Desorption Technique to Catalyst Studies. Adv. Catal. 1967, 17, 103.
Received for review April 16, 1998 Revised manuscript received August 10, 1998 Accepted August 13, 1998 IE980237S