Chem. Muter. 1992,4,50&502
500
to form the final TiB2 particles. An SEM micrograph (Figure 4b) of the product heated to 1500 "C clearly shows the morphology of the TiB2particles obtained. The particle size distribution measurement (with centrifugal technique) indicates that the particle sizes are between 0.05 and 1pm. The average particle size based on the above distribution was 0.30 pm. The elemental analyses of the precursor and the pyrolysis product at 1500 "C (Table 11) indicates that some boron was lost during the heat treatment, which subsequently led to a Ti:B ratio of 1:1.78 in the final product. This loss of boron is not surprising, since boron oxide is known to be volatile above 1300 "C. In addition, this B2O3 loss could be partially responsible for the weight loss between 1300 and 1500 "C. The pyrolytic yield (weight of residue) of product after heating to 1500 "C was 33.40% and is 92.3% of the amount expected based on Ti(0-nBU)4 To prepare TiB2 with a 1:2 Ti/B ratio in the final product, a precursor was synthesized which contained more boron and carbon and had a nominal composition of 2.3B/ [(Bu0)0.31(F~O)0.69Ti01.5]n. After pyrolysis at 1500 "C for 6 h, the elemental analyses (Table 11) of the resulting pyrolysis product indicated that it had a Ti:B ratio of 1:2.08 but still contained 2.57% C and 1.00% H. The oxygen content in this product was estimated to be less than 0.6%, based on the combined percentages (99.42%) of Ti, B, C, and H. The average particle size of the product was found to be 0.38pm. The low oxygen content of this product indicates that the carbothermic reduction of B203 is nearly complete at 1500 "C, a condition also observed in the borothermic/carbothermic reduction of Ti02with B4C.14 On the basis of the above observations, pure TiB2powders with submicron sizes can be prepared with this procedure at 1500 "C if the quantities of the starting materials are carefully controlled. Furthermore, we found that the pyrolysis process did not follow reaction 1but proceeded by a combination of borothermic and carbothermic reduction reactions in agreement with the results obtained by Walker.14 However, reaction 1 still can be used to describe the overall reaction, sipce it is obtained by combining reactions 3-6. Acknowledgment. Financial support for this work was provided by the Air Force Office of Scientific Research (AFOSR) under Contract No. F49620-89-C-0102. Registry No. TiB2, 12045-63-5.
Nanoparticles of Anatase by Electrostatic Spraying of an Alkoxide Solution Dong Gon Park and James M. Burlitch"
Department of Chemistry Cornell University, Ithaca, New York 14853-1301 Received December 23,1991 Revised Manuscript Received March 4, 1992 Much effort has been directed toward the synthesis of submicron particles with a narrow size distribution for the preparation of nanophase ceramic material~l-~ that have
* To whom correspondence should be addressed. (1)Bowen, H. K. Muter. Sci. Eng. 1980,44, 1-56 and references therein. (2)Matijevic, E. Muter. Res. SOC.Bull. 1989,14,18-46and references therein.
0897-4756/92/2804-0500$03.00/0
Water Nitrogen and solvent out
co?jng~
-. Metal Plate
Collecta
+
+
Heat Lamps
Heat Lamps
+
+
+
- Metal Plate I 1 II
-
II
Solution
Figure 1. Schematic diagram of the setup used for the fabrication of Ti02nanoparticles.
improved properties! This is done by consolidating and sintering nanometer-sized particles, often called nanoParticles? In most cases,the nanoparticlesare synthesized in vacuum or at low pressure by evaporation of a bulk source material. Synthesis of nanocomposite materials in solution is limited by aggregation; although additives help keep particles dispersed in solution, they have to be removed before a pure, nanophase solid can be formed.6 Because condensation reactions are enhanced during the drying stage of the sol-gel process, particles of gel-forming compounds become larger than those produced by gasphase method^.^ Titania particles have been made from metal-organic compounds by reactions in aeros01s.~~~ Particles smaller than 100 nm were made by thermal decomposition of titanium alkoxides.1° These methods, however, can be applied only to volatile compounds. When a liquid droplet, under the influence of an intense electric field, acquires charge density above the Rayleigh limit, the droplet becomes unstable and breaks up into numerous, charged small droplets; the liquid emanating from a tube forms either a single jet or a multiple jet depending on the electric field strength.11J2 Many practical applications of this phenomenon have been de(3)Matijevic, E.Muter. Res. SOC.Bull. 1990,15(1),1647 and references therein. (4)Siegel, R. W.; Ramasamy, S.;Hahn,H.; Zongquan, L.; Ting, L.; Gronsky, R. J. Muter. Res. 1988,3,1367-1372. ( 5 ) Andrm, R. P.; Auerback, R. S.; Brown, W. L.; B m , L. E.; Goddard, W. A., III; Kaldor, A.; Louie, S. G.; Moskovita, M.; Peercy, P. S.;Riley, S. J.; Siegel, R. W.; Spaepen, F.; Wang, Y. J. Muter. Res. 1989,4,704-736. (6)Kernizan, C. F.; Klabunde, K. J.; Sorensen, C. M.; Hadjipanayis, G. C. Chem. Muter. 1990,2,70-74. (7)Roy, R.; Komameni, S.; Yarbrough, W. Some New Advances with SSG-Derived Nanocomposites. In Ultrastructure Processing of Advanced Ceramics; Mackenzie, J. D., Ulrich, D. R., Eds.; John Wiley & Sons: New York, 1988;pp 571-588. (8)Matijevic, E. Annu. Rev. Muter. Sci. 1985,15,483-516and references therein. (9)Salmon, R.; Matijevic, E. Cerum. Int. 1990,16,157-163. (10)Okuyama, K.; Jeung, J. T.; Kousaka, Y.; Nguyen, H. V.; Wu, J. J.; Flagan, R. C. Chem. Eng. Sci. 1989,44,136*1375. (11)Cloupeau, M.; Prunet-Foch, B. J. Electrost. 1989,22,135-159. (12)Luttgens, U.;Rollgen, F. W.; Cook, K. D. Optical Investigation of Electrohydrodynumic Disintegration of Liquids in EH and ES MS; 38th ASMS Conference on Mass Spectrometry and Allied Topics, 1989; pp 132-133.
0 1992 American Chemical Society
Chem. Mater., Vol. 4, No.3,1992 601
Communications
.
. .
,
0.5
1
0.4
*
. g0.3
*
3 al
g0.2
t
0.1
I
0.n
0
10
30
20
40
50
Diameter (nm)
t.
a .*
.
c
+
1L
o
*
. a 6 t
. +7
0
r
. Figure 2. Bright field TEM image and size distribution of nanoparticles of the Ti02 precursor made from a 5.0 X M solution of Ti(OPf)3acac (bar = 500 nm).
Figure 3. Bright-fieldTEM image and size distributionof nanoparticles of the TiO2 precursor made from a 1.0 X 10" M solution (bar = 200 nm).
scribed, including paint and crop spraying and ink-jet printing.13 The first use of this phenomenon (called electrostatic'atomization") to make ceramics particles was described by Slamovich and Lange;14particles of zirconia, made from aqueous solutions of Z~(OAC)~, ranged from 0.1 to 5 pm.15 In the present study, a metal alkoxide in an organic solvent was transformed into an aerosol by the electrostatic "atomization" technique (electrospray). Charged droplets in the aerosol formed unaggregated nanoparticles during their flight from the spray n d e to the collector. The apparatus for electrospray is shown schematically in Figure 1. A Pyrex glass tube (12-cm inner diameter and 45 cm long) was heated by incandescent lamps. A potential of +23 kV dc was applied to both the pointed stainless steel needle and a nearby metal plate that was used to prevent the solution from creeping to the outside of the needle. A solution of the liquid precursor, Ti(OP+)3acac,15was fed through the 29-gauge needle by constant pressure from argon gas, since even small variations in pressure provoked severe instability of the spray. Under these conditions, the spray formed multijets, as
observed with a telescope. Nitrogen gas was passed through the tube at 30 cm3/s to remove evaporated solvent. If the inside dimension of the tube were large enough, the hot tube surrounding the region between the needle and the collector would restrict the particles to the central portion. Particles accumulated in a circular area, ca. 2 cm in diameter in the middle of a water-cooled, grounded metal plate to which was attached a copper grid for "EM, a piece of a conductive silicon wafer, or other substrate. Remarkably, no particles were found on "EM grids p l a d outside this circle. This "focusing" was probably caused by a combination of therm~phoresis'~ and a counter-field built up on the wall by a few charged particles. If the tube was not heated, the aerosol would condense on the wall, and the spray would gradually stop, probably as a result of the central electric field being weakened by back ionization18caused by charged particles on the wall. Particles formed from solutions of Ti(OPr')3acacthat ranged in concentration from 1.0 X to 5.0 X M. At the highest concentration the distribution was bimodal around 150 and 20 nm (Figure 2). Hydrolysis and condensation of the alkoxide were probably caused by adventitious water in the solution and in the nitrogen a t mosphere. As the concentration was decreased, the fraction of smaller particles (around 20 nm) increased, and the size of the biggest particles decreased below 100 nm. When the concentration was lower than 1.0 X M, the mean size decreased to ca. 20 nm (Figure 3). At the lower concentration the droplets presumably had more time to evaporate and disrupt before gelation occurred. Strong peaks for Ti were observed in the EDS spectra of the
(13) Bailey, A. G. Electrostatic Spmying of Liq&, Reeearch Studies Press, Ltd.: Taunton, England, 1988. (14) (a) Slamovich, E. B., Lange, F. F. Mater. Res. Soc. Symp. h o c . 1988,121,257-262. (b) Slamovich, E. B., Lange, F. F. J. Am. Ceram. SOC. 1990, 73,336a-3375.
(15) Under an atmoephere of dry argon,Ti(OP& was modified by the addition of an equimolar quantity acetyhcetone,M and the liquid reaction product, Ti(OPf)pcac, was diluted with EtOH, that had been dried over Mg(OEt)*,to the final concentration. To increase the electrical ConM solution ductivity, ammonium acetate was added into the 1.0 X of Ti(OP&acac80 that the final concentration of the salt was 1.0 X M. Modification with acac decreases the rate of hydrolysis1sand should give the dropleta more time to disrupt before gelation occurs (large increase in viscosity). (16) Leaustic, A.; Babonneau, F.; Livage, J. Chem. Mater. 1989, 1, 2W252.
(17) (a) Mercer, T. T. Aerosol Techno& in Hozard Euulwtwn; New York, 1973. (b) Friedlander,S. K.; Femandez de la Mora, J.; Gokoglu, S. A. J. Colloid Interface Sci. 1988,125,351-355. (18) White, H. J. J . Air Pollut. Control Assoc. 1974, 24, 314-338.
Chem. Mater. 1992,4,502-504
502
C'
b
,: .. .
size of the particles were unchanged. In summary, the conditions used to make nanoparticles from a gel-forming, metal+rganic compound in an organic solvent by the electrospraymethod were the following: (1) a multijet spray formed very small droplets; (2) a very low concentration of precursor facilitated the process in (1); (3) sufficient path length allowed the particles to dry before they struck the collector; (4) to keep collection efficiency high, the a e m l was "focused" into a small (5) adding NH,OAc to the solution increased the efficiency of collection. The composition of the as-formed nanoparticles, the mechanism by which the added salt increases the yield, and ways to scale up the process are under investigation. Since electrospray can be carried out under various conditions, e.g., in gases other than N2and at pressures under 1 atm, it should be feasible to apply the method to the preparation of various nanophase materials, including composites.
--
Acknowledgment. We thank Corning Inc. and the I.B.M. Corp. for financial support (Cornell University Ceramics Program). Microscope facilities, administered by the Materials Science Center, were partially funded by NSF. Registry No. noz,13463-67-7; Ti(OP&, acac, 15701-48-1; NHIOAC, 631-61-8.
:I (19) McClune, W. F., Ed.Powder Diffraction File, Set 21, JCPDS Int Center for Diff. Data, Swathmore, PA, 1980, card 21-1272.
.
.
.
. -
-
-
Figure 4. SEM images of TiOz particles made by electrospray deposition from a 1.0 X lod2M solution with ammonium acetate
(1.0 X lC3M)over a period of 21 h. (top) Magnification of 9OOX with some particles removed by a brush (bar = 10 pm), (bottom) magnification of 60000X (bar = 0.1 pm).
particles. Electron diffraction patterns (TEM) indicated that the particles were amorphous. After several hours of electmpraying, some particles had accumulated on the inside wall of the glass tube. This suggested that the central electric field may have been weakened by the back-ionization18caused by particles on the collector;the charge on the collected particles may not have been dissipated very efficiently because the nanoparticles themselveswere nonconductive. To provide some conductance, ammonium acetate was dissolved in the solution. When the salted solution was electrosprayed, no nanoparticles accumulated on the inside wall of the tube during many days of operation. Without added salt it was impossible to collect a large amount of material on any type of collector including glass, but with salt present, a much thicker layer accumulated on the collector (Figure 4, top). A portion of the deposit was easily removed from the glass substrate to show the thickness of the deposit. For these experiments, the concentration of the precursor M, from which a bimodal distrisolution was 1.0 X bution was expected, but only larger particles were seen within the limitsof resolution (Figure4b). After they were heated at 700 "C for 12 h, t h i n - f h XRD analysis of these particles showed that they contained only peaks at 25.3, 36.9, 37.8, 38.6, and 48.0' (28), consistent with the formation of anatase.19 After calcination, the appearance and
Second Harmonic Generation in Inorganic-Organic Complexes. 1. Control of Bulk Dipolar Alignment of p-Nitroaniline Crystal upon Complexing with Hydrobromic Acid Takeshi Gotoh, Noritaka Mizuno,* and Masakazu Iwamoto* Catalysis Research Center, Hokkaido University Sapporo 060, Japan Satoshi Hashimoto and Masahiro Kawasaki Research Institute of Applied Electricity Hokkaido University, Sapporo 060, Japan Masato Hashimoto and Makoto Misono Department of Synthetic Chemistry Faculty of Engineering The University of Tokyo Hongo, Bunkyo-ku, Tokyo 113, Japan Received December 27,1991 Revised Manuscript Received February 13,1992 Recently, much attention has been attracted to organic or inorganic molecular crystals with nonlinear optical properties because of their availability as practical material~.'-~For a material to exhibit second harmonic generation (SHG),it must have a noncentrosymmetriccrystal *To whom correspondence should be addressed. (1) Levine, B. F. Chem.Phys. Lett. 1976,37,516. (2) Tomaru, S.; Zembutsu, S.; Kawachi, M.; Kobaymhi, M. J. Chem. Soc., Chem. Commun. 1984,1207. (3) Wang, Y.; Eaton, D. F. Chem. phys. Lett. 1985,120,441. (4) Eaton, D. F.; Andereon, A. G.; Tam,W.; Wang, Y. J. Am. Chem. Soc. 1987,209,1886. (5) Cox, S. D.; Gier, T. E.; Stucky, G. D.Chem.Mater. 1990,2,609.
0 1992 American Chemical Society
