Preparation of uniform zinc oxide colloids by controlled double-jet precipitationt Qiping Zhongl and Egon Matijevik" Centerfor Advanced Materials Processing, Clarkson University, Box 5814, Potsdam NY 13699-5814, USA A new and simple method for the preparation of uniform zinc oxide colloidal particles of different morphologies by controlled double-jet precipitation is described. Scanning electron microscopy and X-ray diffraction data indicate that the solids generated under different conditions have identical crystallite structures, although they are internally composite. The growths of the subunits and of the particles themselves follow different patterns. Particles formed in the presence of a surfactant (sodium dodecyl sulfate) exhibit essentially the same overall characteristics, but the particles are microporous.Colloidal zinc oxide is of interest both for its many uses and for the fundamental understanding of the formation of colloidal metal oxides. The latter is aided by the properties of the zinc ion, which appears in only one oxidation state and it does not hydrolyse as readily as some other multivalent metal ions. Traditionally, zinc oxide has been used in rubber and adhesive industries. During the past three decades, the chemical industry has opened new markets for this material, such as in the production of catalysts, semiconductors, advanced cer-amics, to name a In these highly technical applications, particle size and shape become very important parameters that determine the physical properties of the product^.^,^ A number of techniques have been developed that yield particles of different, but uniform shapes, and of narrow size distribution in pm and sub-pm range^.^,^ It has been shown that the precipitation of zinc oxide from homogeneous solu- tions can yield particles of narrow size distribution and of different morphologies, yet their X-ray pattern is always characteristic of zincite.Thus, monodispersed ZnO colloids were obtained by ageing at elevated temperatures solutions of zinc nitrate to which various bases were added.' Solutions of Zn(NO,), heated in the presence of urea resulted in uniform spherical zinc basic carbonate, which at elevated temperatures decomposed to zinc oxide.g Hydrolysis of zinc salt solutions in polyols also produced zinc oxide," but the particles so obtained contained significant amounts of carbon from organic compounds, which were difficult to remove by washing.In a much more involved two-step process, Haile and Johnson also prepared monodispersed ZnO particle^.^ As a rule, the conditions that yield uniform dispersions tend to be rather restrictive, requiring low concentrations of reac- tants. Thus, it is desirable to develop methods that could generate large quantities of these materials. One such technique is controlled double-jet precipitation (CDJP), which has been primarily employed in the photographic industry for the production of silver halide crystals of narrow size distribution, as well as of well developed habit, precise internal composition, and epita~y.ll-'~ The rate of production of these materials is about 1 mol AgX dm-3 h-l in containers as large as 2000 dm3.The CDJP technique was also used to generate relatively large crystals of various other sparingly soluble salts, including metal oxalates and sulfate~.l~*~~ Recently, uniform nanosize pure and doped barium titanates were produced by this method in a very short time and in large quantities.16,17 This study describes the use of the CDJP process for the synthesis of uniform zinc oxide particles of different morpho- logies, and offers an explanation for the mechanism of their formation. t Supported by the Procter & Gamble Company, Cincinnati, OH. 1Part of a Ph.D. thesis.Experimental Materials All chemicals of reagent grade quality were used without further purification. In order to remove any possible particulate contaminants, stock solutions were filtered through 0.2 pm pore size Nuclepore membranes. Procedures The experimental setup, schematically represented in Fig. 1, consists of a 0.50 dm3 reactor equipped with a heating jacket. Into the reactor was added 0.10dm3 of water preheated to 90 "C, with a stirring rate of 500 rpm, followed by 0.10 dm3 each of a 0.02 mol dm-3 zinc nitrate (which was later increased to 0.1 mol dm-3) and of 1.6 mol dmP3 triethanolamine (TEA) solution, introduced simultaneously through glass tubes at a constant flow rate, controlled by peristaltic pumps. The system r Th Eea t I Fig.1 Schematic presentation of the controlled double-jet precipi- tation (CDJP) reactor: the 0.50dm3 container is equipped with a heating jacket, into which Zn(N03), and triethanolamine (TEA) solutions were introduced simultaneously, under a stirring rate of 500 rpm J. Muter. Chem., 1996, 6(3),443-447 443 Fig 2 SEM images (5 mm =1 pm) of zinc oxide particles prepared at 90 "C using CDJP by adding 0 10 dm3 each of 0 02 mol dm '?(NO,), and 1 6 mol dm TEA solutions into 0 10 dm3 H,O at flow rates 0 003 (a) 0 007 (b) 0 009 (c) 0 014 (d) 0 02 (e) and 0 6 (f)dm3 min was then kept for 30 min at 90 "C, while under continuous of 0 001 mol dm sodium dodecyl sulfate (SDS) solution was agitation The addition rates varied from 0 003 to 0 60 dm3 placed into the reactor instead of water min ' in different runs, in order to evaluate the influence of After ageing the particles were separated from the super- this parameter on the particle morphology and size (Table 1) natant solution by centrifugation and washed with water It was established that accelerating the stirring had no effect several times The powders were then dried in a vacuum oven In expenments dealing with the effect of a surfactant, 0 10 dm3 at 50 "C overnight 444 J Mater Chem 1996, 6(3),443-447 Table 1 The flow rate and the addition time of Zn(NO,), and TEA solutions into the 0.50 dm3 reactor, containing 0.10 dm3 water at 90 "C, in CDJP process' flow rate/dm3 min-' 0.003 0.007 0.009 0.014 0.02 0.6 0.009 addition time/min 33 15 11 7 5 0.17 11 'In all cases, the total added solution volume was 0.10 dm3 each of 0.02 mol dmP3 Zn(NO,), and 2.6 mol dmp3 TEA.In the presence of the surfactant. Characterization Particle morphology and size were determined from scanning electron micrographs (SEM). The crystal structure was charac- terized by X-ray diffraction, and the crystallite size was calcu- lated from the analysis of the line broadening features using the Scherrer formula." The specific surface area of the powder was determined by the nitrogen adsorption method (BET) using three-point measurements. Thermogravimetry (TG) and differential thermal analyses (DTA) were carried out at scan- ning rates of 20 and 10°C min-l, respectively. Results and Discussion Particle shapes The preliminary investigations showed that the shape and size of the ZnO particles depended strongly on the flow rate of zinc nitrate and TEA solutions (Fig.2). The slow addition of reactants resulted in longer, but smaller particles, as illustrated by samples (a) and (b)in Fig. 2. With increasing flow rate, the long axis decreased [(c),(d)and (e)]and, finally, nearly spherical larger particles with a diameter of 1500+300 nm were formed (f). In the presence of ca. mol dm-3 of SDS, only much smaller spherical particles precipitated, as shown in Fig. 3 (g), which were prepared as in Fig. 2(c), but in the presence of 0.001 mol dmP3 SDS. In order to increase the solid content of the resulting particles, the concentrations of the same volume (0.10 dm3) of reagents were increased to 0.1 mol dm-3 Zn(NO,),, 3mol dm-3 TEA, and 0.005mol dm-3 SDS.Identical particles as shown as sample (g) were produced in five times higher amounts. Once SDS was added, the process seemed to be insensitive to the addition rate and reagent concentrations over a broad range; spheres of 200 & 30 nm in diameter were obtained in all cases. However, neither the addition rate nor the surfactant had any effect on the crystal structure of the resulting particles. Fig. 3 SEM image (20 mm = 1 pm) of zinc oxide particle (g)prepared under the same conditions as sample (c), except the water was replaced by 0.10 dm3 of 0.001 mol dmP3 sodium dodecyl sulfate (SDS) solution Crystallite and particle size The X-ray diffraction patterns of samples (c), (f) and (g) (Fig.4) showed that in all cases the same crystalline material was produced, regardless of the geometric shape of the particles (in the absence or in the presence of the surfactant). The corresponding d-spacings matched well with reported values (Table2).19 The crystallite sizes, based on the broadening of the [1001 and [1011 reflections, depended on the conditions of the powder preparation (Fig. 5). Thus, the crystallites of sample (a)were 107 nm long, while those of sample (f) were much smaller (20 nm), although the latter constituted the largest particles. The precipitates synthesized in the presence of SDS [sample (g)]had the smallest crystallite size (15 nm).A model, based on crystal growth by bulk diffusion and the Gibbs-Thomson effect, suggested by several pre-dicted that the number of stable nuclei should increase with I'l.1'I'I' 1200 I 1000 800 400 200 0 20 30 40 50 60 70 80 26Ndegrees Fig. 4 XRD spectra of ZnO powder samples (c),(f)and (g) Table 2 X-Ray diffraction patterns for ZnO powders prepared by CDJP process compared to literature values'' sample d-spacings, at A= 1.5418 A 2.81 2.6 2.48 1.91 1.62 1.48 1.41 1.38 1.36 2.82 2.60 2.48 1.91 1.63 1.48 1.41 1.38 1.36 2.81 2.61 2.47 1.91 1.62 1.48 1.41 1.38 1.36 2.81 2.60 2.47 1.91 1.62 1.48 1.40 1.37 1.36 2.81 2.60 2.47 1.92 1.63 1.47 1.42 1.38 1.36 2.81 2.60 2.47 1.92 1.63 1.47 1.42 1.38 1.36 2.81 2.61 2.48 1.91 1.63 1.48 1.42 1.38 -2.82 2.60 2.48 1.91 1.63 1.48 1.42 1.38 1.36 2o 1T/-l-l-i0 5 10 15 20 25 30 550 600 650 reactant flow rate/103 dm3 min-' Fig.5 Crystallite dimensions as a function of the reagents addition flow rate of Zn(NO,), and TEA solutions: 0,dimension along [1011; 0,dimension along [1001 J. Mater. Chem., 1996, 6(3), 443-447 445 Table 3 Number of stable ZnO crystallites (2)as a function of reactant addition rate (R)of Zn(NO,), and TEA solutions sample R/mol s cm Z/no cm 100x10 2 4 x lolo 2 17 x lop8 7 9 x 1o1O 300x10 ’ 12 x loll 4 67 x lo-* 2 8 x lo1’ 6 67 x 10 ’ 7 5 x lo1’ 200x10 3 7 x lo’, increasing addition rate of reactants, resulting in a smaller particle size The validity of this model was experimentally tested by the double-jet precipitation of AgBr 2o 23 The number of stable ZnO crystallites per unit suspension volume was determined from the following mass balance [eqn (l)], 2=3 VmRt/4nr3 where 2 is the number of stable nuclei, V, the molar volume and r the crystallite radius of precipitated solids, R the rate of reactant addition per unit suspension volume, and t time Here, spherical morphology is assumed for simplicity Table 3 shows the effect of reactant addition rate on the number of stable ZnO crystallites (Z), while log-log plots of 2 us R for the experimental data given in Table2 are presented in Fig 6 Clearly, the crystallites of ZnO particles produced by CDJP followed the same rule, ie, the number of stable nuclei increased, while the crystallite size decreased with increasing addition rate of reactants It is essential, however, to recognize that the crystallite size differs from the final particle size, the latter actually increases with faster addition rates (Fig 2) Obviously, the proposed mechanism applies only to the case of single-crystal growth The precipitation of zinc oxide particles must follow a process that differs from the generation of internal subunits Recently, it has been recognized that in the majority of cases the formation of ‘monodispersed’ colloids proceeds through a two-stage mechanism the primary particles (as in the case of zinc oxide) are produced first, which then aggregate to larger final products 24 However, the relationship of the resulting morphology to the precursor particles has been explained only in a few cases, such as in the formation of haematite 25 26 There are two possible mechanisms involved in the aggregation, based either on diffusion-limited or reaction-limited pro-cesses2’ The subunits have to overcome an energy barrier to effectively collide with growing particles If the driving force is 12 0-I 11 5-E2“,11 0-0, 105-loo!, ., -1 . I -I -I 80 78 76 74 72 70 log (R/mol s-’ cm-3) Fig. 6 Number of stable ZnO crystallites as a function of the reactant addition flow rate of Zn(NO,)z and TEA solutions 446 J Muter Chem, 1996, 6(3), 443-447 0-final crystallite size along final crystallite size along dimension [100]/nm dimension [10l]/nm 107 81 76 65 63 63 47 46 37 33 26 25 high enough, aggregation is rapid and the rate is limited by the diffusional motion of the colliding subunits Thus, the highly concentrated small crystallites are aggregated to large spherical particles due to their faster Brownian motion and higher collision efficiency, such as in sample (f) (Fig 2, 5, and Table 3) In the case of less concentrated larger subunits, the driving force is low, and aggregation may only happen on the certain ‘active’ sites of crystallites, especially when the latter are asymmetric or are built with different geometric faces 28 30 This kind of oriented aggregation may lead to the formation and growth of columnar or needle type particles, as in the case of sample (a) (Fig 2) The ZnO particles indeed changed systematically from elongated [sample (a)] to spherical [sample (f)]with decreasing asymmetry and smaller size of crystallite subunits (Fig 5), which is expected from the aggregation theories Surface area A closer inspection of the scanning electron micrographs (Fig 2,3) shows that the surfaces of samples (f) and (g)appear rough, indicating formation by the aggregation of small crystal- lites, as supported by the X-ray analysis This nature of the particles is also reflected in the high specific surface area as determined by BET The particle size calculated from the BET data, based on a density of 5 6 g cm for zinc oxide, yielded a diameter of ca 710 nm (specific surface area, SSA, 1 5 m2 g-l) for sample (f) and of ca 28 nm (SSA, 38 m2 8-l) for sample (g) While BET measures the specific surface area of the particles, including pores, XRD depends only on the crystalline regions of the sample It is, therefore, not surprising that the sizes calculated from BET data are larger than those determined by XRD Obviously, sample (g)(prepared with SDS) was porous, which must have been formed by aggregation of a larger number of ’*05 I 975 1 t I.I 100 200 300 400 600 TI C Fig.7 Typical TG curve for the zinc oxide powder sample (a) at a heating rate of 20°C min in air Similar curves were recorded in all cases order to develop a better understanding of the reasons for 40 A35 TI"C Fig.8 DTA curve of samples (a),(f) and (8)(heating rate 10°C min-' in air) small particles, possibly incorporating some surfactant Sample (f) consisting of larger crystallites, may owe its BET area to the surface roughness of the particles Thermal behaviour TG and DTA curves are presented in Fig 7 and 8, respectively The endothermic peak occurred around 100"C and the related mass loss arose from the loss of water, which in all cases amounted to ca 3% The exothermic peak at ca 370 "C, which was rather weak for sample (a) but much more pronounced for samples (f)and (g), corresponded to the crystallization of zinc oxide The calcined samples were re-examined by SEM and XRD, which showed that the crystallite size of (f)increased from 25 nm to 31 nm, and that of sample (g)from 15 nm to 22 nm, but no change was observed with sample (a), which agreed well with the DTA data The size and the shape of the final particles, on the other hand, remained the same in all samples Conclusions This work demonstrated the feasibility of the preparation of monodispersed zinc oxide powders by controlled double-jet precipitation The morphology and the size of the particles were dominated by a single parameter, z e the flow rate of the reactants All particles were built from crystalline subunits, but the precursors and the final products followed different growth mechanisms Obviously, the nucleation, crystallite growth, and the aggregation processes should be distinguished in the forma- tion of these solids More systematic studies are necessary in varying particle sizes and morphologies Surfactant SDS greatly affected the properties of the prod- ucts obtained by CDJP The particles so obtained were micro- porous with a high specific surface area and were made up of aggregates of very small particles whose size is ca 28 nm References 1 A S Perl, Am Cerum Soc Bull, 1993,72, 122 2 M Haase, H Weller and A Henglein, J Phys Chem ,1988,92,482 3 L Spanhel and M Anderson, J Am Chem Soc , 1991,113,2826 4 S M Haile and D W Johnson, Jr, J Am Cerum SOC,1989, 72,2004 5 L Koudelka and J Horak, J Muter Scz , 1994,29,1487 6 E Matijevic, Chem Muter, 1993,5,412 7 E Matijevic, Lungmuzr, 1994, 10, 8 8 A Chittofrati and E Matijevic, Collozds Surf, 1990,48,65 9 M Castellano and E Matijevic, Chem Muter, 1989, 1, 78 10 D Jezequel, J Guenot, N Jouini and F Fievet, J Muter Res, 1995,10,77 11 I H Leubner, in Reprographzc Technology, American Chemical Society, Washington, DC, 1982 12 C R Berry, in The Theory of the Photographic Process, 4th edn, ed T H James, Macmillan, New York, 1977 13 J Stavek, M Sipek, I Hirasawa and K Toyokura, Chem Muter, 1992,4545 14 J Stavek, T Hamslik and V Zapletal, Muter Lett, 1990,9,90 15 H Muhr, R David, J Villermaux and P H Jezequel, Chem Eng Scz , 1995,50,345 16 Y S Her, E Matijevic and M C Chon, J Muter Res, in press 17 Y S Her, E Matijevic and M C Chon, J Muter Res, 1995, 10,3106 18 E Warren, X-Ray Dzfructzon, Addison-Wesley, Reading, MA, 1969 19 Fink Index to the Powder DzfSructzon File, ASTM, 1966, NO 5-0664 20 T Sugimoto, J Colloid Interface Scz ,1992,150,208 21 V G Loginov and N B Denisova, Zh Nauchn Przkl Fotogr Kznemutogr , 1975,20,231 22 I H Leubner, R Jagannathan and R S Way, Photogr Scz Eng, 1980,24,268 23 E Klein and E Moisar, Ber Bunsenges Phys Chem ,1963,67,348 24 J Stavek and J Ulrich, Cryst Res Technol, 1994,29,465 25 M Ocaiia, R Rodriguez-Clemente and C J Serna, Adv Muter, 1995,7,212 26 J K Bailey, C J Brinker and M L Mecartney, J Colloid Interface Scz , 1993,157,l 27 J Stavek and J Ulrich, Cryst Res Technol, 1994,29,763 28 Y A Brasley, V V Peisakhov and L Y Kaplun, Usp Nuuchn Fot , 1986,24,5 29 P H Karpnski and J S Wey, J Imug Scz ,1988,32,34 30 T Sugimoto, J Imug Scz , 1989,33,203 Paper 5/05437K, Received 14th August 1995 J Muter Chem , 1996, 6(3), 443-447 447