Journal of Sol-Gel Science and Technology 36, 11–17, 2005 c 2005 Springer Science + Business Media, Inc. Manufactured in the United States.
Preparation of Ultra-Fine CuO: Comparison of Polymer Gel Methods and Conventional Precipitation Processes ∗ ´ ´ A.I. FERNANDEZ , A. CALLEJA, J.M. CHIMENOS, M.A. FERNANDEZ, X.G. CAPDEVILA, M. SEGARRA, H. XURIGUERA AND F. ESPIELL Unit of Materials Science and Metallurgical Engineering, Faculty of Chemistry, University of Barcelona, Mart´ı i Franqu`es 1, E-08028, Barcelona, Spain
ana
[email protected]
Received September 30, 2004; Accepted May 31, 2005
Abstract. Different ways of preparing ultra-fine copper oxide were examined. Polymer precursor techniques using polyvinyl alcohol and acrylamide were compared with the conventional precipitation of copper salts and further calcination. Thermal analysis, XRD and TEM were employed to monitor polymer degradation and the phase transformations leading to copper oxide and to characterise particle size. Copper oxide obtained by precipitation from copper nitrate has smaller particles size meanwhile that obtained by acrylamide combustion method agglomerates in lower extension. Differences between XRD crystallite size calculations and TEM observations of particles size were checked. Keywords: CuO, precursor chemistry, PVA, acrylamide, copper formate, copper nitrate
Introduction Copper oxide was chosen for a comparative study on the synthesis method of ultra-fine particles, because of its importance as catalyst in organic reactions and other industrial applications such as a pigment, in sweeting petroleum gases, in galvanic electrodes, metallurgical flux, and as an optical glass polishing agent. Fine copper oxide can be obtained by several methods such as precipitation from acidic copper solutions [1] with NaOH, sonochemical synthesis from copper acetate [2] or as thin film [3, 4] by means of the sol-gel dip technique for solar cell applications. Gel combustion methods provide a way towards small particle size materials, with high chemical homogeneity, with a minimum of intermediate phases or impurities. The polyvinyl alcohol (PVA) route has already being studied for preparation of specific ceramic materials such as high-temperature superconductor (HTSC) ∗ To
whom all correspondence should be addressed.
powders [5]. Evaporation of the solution containing a dissolved copper salt and PVA produces an ash that, in just one calcination, converts into the oxide. The use of acrylamide to form an auxiliary three-dimensional organic network was explored by Douy [6] using chelated cation solutions to prepare oxide precursors. As case studies, various series were performed by the polymer-gel method using PVA in which the effect of the PVA:Cu ratio was studied, or using acrylamide to evaluate whether or not to employ reticulating agent on the final particle size. The resulting gels were further heated and the ashes calcined at different temperatures to evaluate their effect. Another method of synthesis was explored by conventional precipitation of copper salts followed by thermal treatment. Three salts with relatively low temperatures of decomposition were selected: copper formate, copper nitrate and copper hydroxide, which, on being further heated, led to fine CuO. The preference for copper formate is due to its low decomposition temperature and the use of formate solutions in synthesis of HTSC
12
Fern´andez et al.
oxides [7]. Copper nitrate is also a precursor of copper oxide for catalytic purposes [8]. The aim of this study was to evaluate the copper oxide by the particles size and obtained by various methods of synthesis in order to assess the viability of an up-scaling process. Methods and Materials CuO 99.99% (DIOPMA) was used as a precursor of copper solutions for the polymer gel methods and as a source of copper cations for the precipitation techniques, mainly to avoid the presence of other ions that may interfere with the synthesis. PVA Ashes The method used here is described elsewhere [5]. It starts with an aqueous solution of copper nitrate prepared by dissolving CuO 99.99% (DIOPMA) in hot concentrated nitric acid. PVA with a polymerisation degree of 4–88 (FLUKA) is pre-dissolved in water and is added to the copper solution slowly to avoid excessive foaming. The effect of PVA concentration was evaluated by preparing three solutions whose PVA/Cu ratio ranged from 4:1 to 16:1. By means of a hot plate, evaporation took place with simultaneous thickening of the solution. In this step, thickening occurred through entanglement of the long PVA chains, which leads to a disordered network that collapses on heating into a black, crispy powder. These ashes obtained by heating the solutions with PVA/Cu ratios of 4:1, 8:1, and 16:1 were given the sample names 1, 2 and 4 PVA. Three aliquots of the sample 1PVA were calcined for 5 hrs in high-grade alumina crucibles at 400◦ , 600◦ and 800◦ C, respectively, giving the samples 1PVA400, 1PVA600 and 1PVA800. Acrylamide (AA) Gels The procedure followed for obtaining the acrylamide gels was developed by Sin and Odier [9] and up-scaled in a previous paper [10]. Copper oxide is dissolved in hot nitric acid, EDTA is added (EDTA:Cu 1:1) to complex copper cations, and pH is adjusted with ammonium hydroxide. Acrylamide 50% v/v in water is added under continuous stirring and the amount of AA is calculated as 10% of the copper solution mass (a highly diluted solution is assumed). On the mixture
being heated, the polymerisation initiator azobisisobutyrnitrile (AIBN) and the cross-linking agent bisacrylamide (BisAA) were added. After spontaneous gelling at 90◦ C, the gel was further heated until the bottom part self-ignited (Sample AA1). A second series was heated without adding the reticulating agent (Sample AA2). Three aliquots were taken from both the burned residues and heated at 400◦ , 600◦ and 800◦ C.
Precipitation/Crystallization of Copper Salts Copper nitrate was obtained by evaporating a solution of CuO 99.99% (DIOPMA) in hot nitric acid. This salt was later heated up at 350◦ C to form copper oxide (Sample CuO-NP). Thermal analysis of copper nitrate has been studied elsewhere [8, 11]. 330◦ C is reported as the temperature for the formation of stable CuO in air atmosphere. The effect of heating temperature on particle size was evaluated by heating for 5 hrs at 400◦ , 600◦ and 800◦ C. Copper formate was crystallized from a dilute solution, obtained by dissolving 6g CuO 99.99% (DIOPMA) with formic acid, and then heated at 400◦ C (Sample CuO-F). Copper oxide was obtained by slowly adding some drops of 1 M NaOH to a 1 M CuSO4 solution vigorously stirred. The precipitate thus obtained was immediately filtered under vacuum and dried at 110◦ C (Sample CuO-P).
Analysis Powder X-Ray Diffraction (Cu Kα radiation) was performed in a SIEMENS D-5000 model apparatus. Lattice parameters were refined by the least-squares method with version 3 of the CELREF program. Transmission electron microscopy (TEM) with a 200 kV Hitachi H800MT microscope evaluated particle shape and size in selected samples. Differential scanning calorimetric curves were obtained by means of a DSC30 Mettler Toledo calorimeter. DSC was used to monitor the combustion/thermal decomposition of the acrylamide gel. Thermogravimetrical heating curves (10◦ C/min in air) were recorded in a Setaram TG DTA92 thermobalance. TGA was used to monitor the decomposition of samples 1PVA, copper nitrate and copper formate. To evaluate particle size distribution, tests were run with a Beckman Coulter LS Particle Size Analyser. Though various operational conditions were
Comparison of Polymer Gel Methods and Conventional Precipitation Processes
tested, such as several solvents or dispersant agents, measures were carried out using acetone as fluid and previously submerging each sample 5 min in an ultrasound bath. Some agglomeration can not be avoided.
Results and Discussion Thermal Analysis PVA combustion/decomposition in the presence of nitrates and copper (II) ions is described elsewhere [9]. Thermogravimetric analysis was performed with pure PVA and with the sample 1PVA. As expected, decomposition temperature for sample 1PVA was less than with pure PVA (Fig. 1): a temperature above 500◦ C is needed to calcine the ash obtained completely. The increase in the PVA/Cu ratio had a catalytic effect on the rate of formation of ash while the sample was being heated. Thus sample 4 PVA was obtained in a shorter period than 2PVA and 1PVA. Nevertheless, increase in PVA leads to a reduction of CuO while heating, as the PVA decomposes and yields reducing compounds such as CO. This was corroborated in sample 4PVA, which has a brown color after calcination due to Cu2 O presence. DSC profiles for sample AA1 under air and nitrogen atmospheres are depicted in Fig. 2. In both cases an en-
Figure 1.
TG curves of 1PVA sample and pure PVA in air.
13
dothermic peak is observed around 95◦ C, corresponding to water loss. A first exothermic peak appearing at 243◦ C may be attributed to partial oxidation of the chains and exothermic reaction of NOx with organic matter. Another stronger exothermic event appears at 413◦ C under air atmosphere, corresponding with oxidation and depolymerisation that are not observed in nitrogen atmosphere. Table 1 summarises these results and includes those obtained for sample AA2, which did not have a reticulating agent added. This sample had a higher energy step for water loss, but a lower energy step for partial oxidation. This may be attributed to the absence of reticulating agent increasing the energy required to oxidise and depolymerise the sample. The TGA curve for copper formate decomposition is shown in Fig. 3. Complete decomposition takes place at 220◦ C, but a slight weight gain is recorded until 400◦ C, probably due to the presence of small amounts of Cu2 O that oxidize to CuO. These results corroborate those of Mohamed et al. [12]. X-Ray Diffraction XRD was performed on various samples (Fig. 4). Some copper nitrate still remained in the ashes obtained with PVA and acrylamide, as well as in CuO-NP sample heated at 350◦ C. Otherwise, XRD performed on
14
Fern´andez et al.
Figure 2.
DSC curves of acrylamide gel AA1 in air and N2 atmosphere.
samples heated at 400◦ C during 5 hrs confirmed the absence of nitrate. Crystallite size was calculated for all samples by the Scherrer equation: 0.94 × λ B × cos θ
Dhkl =
where λ is the wavelength (15.4056 nm for Cu Kα1 ), θ is the Bragg angle, D the crystallite mean size Table 1. Temperatures and heat release in the thermal decomposition of acrylamide gels. N2 atmosphere Sample
Figure 3.
TGA of CuO-F sample in air atmosphere.
Air atmosphere
Temperature Heat release Temperature Heat release (◦ C) J/g (◦ C) J/g
AA1
92.9
−895.6
95.0
AA1
243.2
366.9
247.5
−851.9 329.5
AA1
–
–
413.3
2603.0
AA2
97.5
−1405.2
AA2
239.9
117.4
in nm, and B the corrected full-width half-maximum (FWHM). The single peak used for calculations corresponds to a middle intensity peak for hkl (20-2) reflection. The peak widths (FWHM) were calculated by means of a least-squares fit. The results as a function of treatment temperature are summarised in Fig. 5. For a given preparation method, the calculated crystallite
Comparison of Polymer Gel Methods and Conventional Precipitation Processes
Figure 4.
XRD patterns of different samples.
15
derived from acrylamide gels AA1 and AA2, 1PVA heated at 600◦ C and CuO-P. For all cases, particles appeared as agglomerates. The 1PVA particles had an approximate particle size of about 120 nm. Both acrylamide samples AA1 and AA2 appear forming chains, being less agglomerated than 1PVA. Otherwise, sample CuO-P shows a quite different morphology: it is a cluster whose single particles have polyhedral shapes. Discrepancies between the crystallite size calculated and the observations made by TEM may be attributed to the presence of grain boundaries as the observed particles contain more than one diffracting domain.
Particle Size Distribution
Figure 5. Crystallite size, calculated with Scherrer equation, as a function of treatment temperature.
size increases with the increase in temperature, as is the case of trials AA1, AA2, 1PVA and CuO-NP. Nevertheless, for a calcination temperature higher than 100◦ C, the CuO-NP series tends to be smaller than the rest of the samples. The PVA method leads to smaller crystallite sizes than acrylamide method. Although polyacrylamide gels containing bisacrylamide have smaller pores than those without reticulating agent, thus yielding to smaller particles, the results showed unexpected behaviour, in that trial AA1 had a higher crystallite size than trial AA2. A local increase in temperature because there are more organics to remove may lead to the growth of crystallites. Transmission Electron Microscopy Morphology of CuO particles was observed by TEM. Figure 6 shows the micrographies obtained for samples
Evidences of particles agglomeration were observed measuring the particle size distribution, and analysing both, the differential number and the differential volume distribution. Results obtained from the number distribution such as the mean value and the cumulative parameters d10, d50 and d90 are listed in Table 2 for samples derived from acrylamide gels AA1 and AA2, 1PVA and CuO-NP heated at 600◦ C and CuO-P. From these data it is observed that except the CuO-P that show the higher mean particle size value, the rest show a very similar particle size distribution. Quite different are the results obtained by calculating the differential volume distribution, as can be observed in Fig. 7, where all samples showed broad distributions and much greater mean values. A first interpretation of these data is that agglomeration was not avoided with the ultrasound bath. Moreover, the fact that CuO-P has the lower difference between mean values of the number and volume particle size distributions indicates that is the one showing less particles agglomeration, what is probably related with the fact that these particles where heated at
Table 2. Results of statistics from number particle size distributions. d10
D50
d90
Mean
S.D
AA1
0.057
0.090
0.183
0.098
1.67
AA2
0.057
0.090
0.186
0.098
1.65
PVA1
0.057
0.089
0.177
0.096
1.60
CuO-NP
0.058
0.093
0.203
0.123
0.20
CuO-P
0.183
0.290
0.554
0.307
1.54
16
Fern´andez et al.
Figure 6.
TEM Micrographies of different samples obtained at 600◦ C.
explanation given about local increase of temperature may also justify more particles agglomeration.
Preliminary Considerations for an Up-Scaling Process
Figure 7. Particle size distributions measured in acetone after 5 min. in ultrasound bath.
a lower temperature. Other comparison may be done between samples AA1 and AA2, as AA1 also agglomerates and shows a volume distribution with higher values of particle size than AA2. This result agrees with the observations made by X-Ray diffraction and the
Both PVA and acrylamide processes require high energy consumption to dry and totally ignite the organic matter. Moreover, the use of acrylamide monomers at pilot plant scale is not recommended for safety reasons. However, these negative aspects may be justified when the alternatives are solid-state reactions with long and costly thermal treatments as well as unavoidable milling operations. The study of the acrylamide route may be considered when spherical and non-agglomerated CuO particles are needed, while the PVA route leads to a homogeneous particle shape and size of agglomerated CuO. Of the precipitation processes studied, the copper nitrate process leads to smaller though rather non-homogeneous and agglomerated particles.
Comparison of Polymer Gel Methods and Conventional Precipitation Processes
Nevertheless, more experimental tests are needed to evaluate parameters such as initial concentration, pH etc., in order to choose or reject the use of formic acid or the precipitation of copper hydroxide. The method using formic acid is the one that involves the least waste management. The thermal treatment of copper nitrates requires a gas effluent scrubber and the precipitation from a CuSO4 solution requires a ◦ wastewater treatment to reduce SO= 4 levels. 400 C is the minimum heating temperature that totally decomposes either copper nitrate or copper formate, with the CuO obtained from nitrate decomposition having smaller crystallites. Further study on the optimization of the precipitation conditions of copper formate is needed in order to evaluate their effects on particle size and shape, as this process for producing fine copper oxide seems the most environmentally friendly one. Conclusions CuO ultra-fine particles were prepared by the gel combustion method using acrylamide and PVA, and by the precipitation-pyrolysis method from copper salts such as nitrate, formate and hydroxide. For a given temperature of calcination, the crystallite size (calculated with the Scherrer equation) of CuO obtained by nitrate precipitation tends to be smaller than in other methods. Discrepancies were observed between the crystallite size calculated and the TEM observations. Differences in particle morphology and tendency to agglomerate were also observed by TEM. Evidences of particles
17
agglomeration were also observed in the particles size distribution. The use of bisacrylamide during the polymerisation of acrylamide gels caused an unexpected reaction, in that bigger CuO particles were obtained when a reticulating agent was used. Acknowledgments The authors wish to thank the Serveis Cient´ıficoT`ecnics of the Universitat de Barcelona and CIDEMGeneralitat de Catalunya. References 1. Y.K. Kim, D. Riu, S. Kim, and B. Kim, Mater. Lett. 54, 229 (2002). 2. R.V. Kumar, Y. Diamant, and A. Gedanken, Chem. Mater. 12, 2301 (2000). 3. S.C. Ray, Sol. Energ. Mat. Sol. C. 68, 307 (2001). 4. T. Maruyama, Sol. Energ. Mat. Sol. C. 56, 85 (1998). 5. A. Calleja, M. Segarra, I.G. Serradilla, X.G. Capdevila, A.I. Fern´andez, and F. Espiell, J. Eur. Ceram. Soc. 23, 1369 (2003). 6. A. Douy, J. Inorg. Mater. 3(7), 699 (2001). 7. J. Block and L.E. Dolhert, Mater. Lett. 11(10–12), 334 (1991). 8. Z. Ding, W. Martens, and R.L. Frost, J. Mater. Sci. Lett. 21, 1415 (2002). 9. A. Sin and P. Odier, Adv. Mater. 12, 649 (2000). 10. A. Calleja, X. Casas, I.G. Serradilla, M. Segarra, A. Sin, P. Odier, and F. Espiell, Physica. C. 372/376, 1115 (2002). 11. Z.D. Zivkovic, D.T. Zivkovic, and D.B. Grujicic, J. Therm. Anal. Calorim. 53 (1998). 12. M.A. Mohamed, A.K. Galwey, and S.A. Halawya, Thermochim. Acta. 411, 13 (2004).
