1、JOURNAL OF COLLOID AND INTERFACE SCIENCE 26, 62-69 (1968) Controlled Growth of Monodisperse Silica Spheres in the Micron Size Range 1 WERNER STOBER, ARTHUR FINK Department of Radiation Biology and Biophysics, Medical School, University of Rochester, New York 14620 AD ERNST BOHN, Max-Planck-Institut
2、fir experimentelle Medizin, GOttingen, Germany Received August 3, 1967 A system of chemical reactions has been developed which permits the controlled growth of spherical silica particles of uniform size by means of hydrolysis of alkyl silicates and subsequent condensation of silicie acid in alcoholi
3、c solutions. Ammonia is used as a morphological catalyst. Particle sizes obtained in suspension range from less than 0.05 t, to 2 in diameter. In many experimental studies which in- volve the use of colloidal suspensions of matter in form of hydrosols and aerosols, it would be desirable to have the
4、suspended phase consisting of homogeneous particles of uniform shape and size. Such monodisperse particulate suspensions offer many experi- mental and theoretical advantages. They not only facilitate easy calibration proce- dures for analytical equipment, but also simplify data reduction, evaluation
5、 and in- terpretation of experiments designed to elu- cidate physicoehemieal properties or physi- opathological effects of colloids and aerosols. The results would no longer be biased by parameters of size and shape distributions. Some monodisperse suspensions of parti- cles in the colloidal size ra
6、nge are available in form of spheres of organic high polymers (1). In the aerosol field, they are primarily used as model substances and for calibration purposes. Generators producing monodis- perse particle clouds from soluble or volatile materials are used in various aerosol studies I This paper i
7、s based in part on work performed under contract with the U.S. Atomic Energy Commission at the University of Rochester Atomic Energy Project and has been assigned Publication No. UR-49-815. and have been described in the literature (2). However, no successful attempt has been made to generate monodi
8、sperse suspen- sions of silica particles. A commercial form of highly disperse silica produced by com- bustion of silicon tetraehloride in a hydrogen torch (3) consists of primary silica spheres of sizes below 0.1 u, but they are aggregated to coarse and irregular clusters which cause a very poorly
9、defined state of suspension. The following investigation was made to explore the possibilities of producing mono- disperse suspensions of silica spheres in the colloidal size range. Such material can be used in both hydrosol and aerosol studies. It will be of particular interest to investi- gators i
10、n the medical field because of its known eytotoxicity and inhalation hazard. The experiments were based on the fact that silica particles can be produced by chemical reaction of tetraesters of silicie acid (tetraalkyl silicates) with certain solu- tions. In 1956, Kolbe (4) described the formation of
11、 silica particles by reacting tetraethyl silicate in alcoholic solution with water in the presence of certain bases. With very pure reactants he observed in several eases a slowly proceeding reaction leading to the formation of uniform spherical silica 62 CONTROLLED GROWTH OF MONODISPERSE SILICA SPH
12、ERES 63 particles. In an attempt to duplicate these findings, many of our experiments resulted in gel formation and only in a few cases did the electron mierographs show particles of ellipsoidal shape and a size range near 0.08 t. Then, a systematic study of the reaction parameters was made and afte
13、r some drastic changes of the experimental conditions, quasi-monodisperse suspensions of silica spheres of sizes up to 2 were finally obtained within less than an hour and the reaction no longer required ex- tremely pure reactants. EXPERIMENTAL Reagents Methanol, ethanol, n-propanol and n- butanol u
14、sed as solvents were of analytic reagent quality. Tetraesters of silieic acid (tetraalkyl silicates) were either supplied in technical grade (methyl, ethyl) or prepared by react- ing silicon tetrachloride and alcohol (n- propyl, n-butyl, wpentyl). All esters were redistilled before used in the exper
15、iments. Ammonia (anhydrous, 99.99%) from a tank was passed through a drying column filled with sodium hydroxide pellets and bubbled through the alcohol at OC. Con- eentrations obtained were about 8 M in methanol, 5 M in ethanol, 3 M in n-pro- panol, n-butanol and n-pentanol. Ammonium hydroxide, 26P.
16、, U.S.P., was used as supplied. Titrations indicated an ammonia concentration of 14.2 M. Instruments Electron microscopes Zeiss EM 9 and Philips EM 100 B and partiee size analyzer Zeiss TGZ 3 were used. Procedures Throughout the investigation, ammonia was used as the catalyst causing the forma- tion
17、 of spherical particles. In many cases, it was applied by adding saturated alcoholic solutions of ammonia to the reaction vessel. In other cases, particularly when high ammonia concentrations were desired, satu- rated ammonium hydroxide solution was used and the water content was taken into account.
18、 At the beginning of each series of tests, pure alcohol or alcohol mixtures, saturated alcoholic ammonia solution, ammonium hydroxide, and water were mixed in Erlen- meyer flasks with ground stoppers or in rubber sealed injection bottles in the de- sired concentrations of ammonia and water. Actual a
19、mmonia contents were measured by withdrawing small samples and titrating with 1 N hydrochloric acid. Total water contents were computed by adding up the fractional amounts introduced by the com- ponents. Subsequently the alkyl silicate was added and the flasks were mounted either on a shaker or in a
20、 water bath under ultra- sonic vibration. Some tests were made while the solution was agitated by a mag- netic stirrer. Either way of agitation was effective and kept the particles in suspen- sion after they had formed. The total amount of solution in each test varied between 50 and 110 ml. One larg
21、e scale experiment with 2 liters of solution gave the same result as a test with 80 ml of a solution containing the same concentra- tions of solutes. Except for the initial exploratory tests at low concentrations, the condensation reac- tion generally started within 10 min. This could easily be obse
22、rved, because, after an invisible hydrolytic reaction forming silieie acid, the condensation of the supersaturated silieie acid was indicated by an increasing opalescence of the mixture starting 1-5 rain after adding the tetraalkyl silicate. After this initial phase, the transition to a turbid white
23、 suspension occurred regularly within a few more minutes. As a standard procedure, samples for electron microscopic investigation were taken after 120 min, although a series of samples taken at different times from the same test solution indicated that particles sometimes reach their final size afte
24、r about 15 min. All sampling for size determination was done by dipping electron microscopic carrier grids covered with Formvar films into the suspension, subsequently putting them on filter paper to remove excessive solution from the grid, and taking electron mierographs of the particles retained o
25、n the fihn. Random mierographs of the samples tom- 64 STOBER, FINK, AND BOHN prising between 100 and 1000 particles were evaluated by means of a semiautomatic particle size analyzer. The cumulative dis- tribution curve of the particles was recorded _- _- _ -i 1.0 -0.5 10 % , o.s +% 0 01 0 5 10 15 MO
26、L/LITER H20 FIG. 1. Final particle sizes as obtained by reacting 0.28 mole/liter of tetraethyl silicate with various concentrations of water and ammonia in ethanol. in logarithmic size increments and plotted on log-probability paper. The median pro- jected diameter and, in some cases, the approximat
27、e logarithmic standard deviation were taken from the graph. RESULTS AND DISCUSSION For the different alcoholic solvents, reac- tion rates were fastest with methanol, slowest with n-butanol. Likewise, final par- titles sizes obtained under comparable con- ditions were smallest in methanol and biggest
28、 in n-butanol. However, there was a tendency toward wide size distributions with the higher alcohols. Methanol-butanol mixtures in a ratio of 1:1 provided more uniform big particles. A similar relationship with regard to reaction rates and particle sizes was found when comparing results with differe
29、nt alkyl silicates. Fastest reactions (less than 1 rain) and smallest sizes (less than 0.2 t) FIG. 2. Electron micrograph of a sample of silica spheres obtained in the ethanol-ethyl ester system. CONTROLLED GROWTH OF MONODISPERSE SILICA SPHERES 65 +.5! ! %0 ., 17 MQ D 900 7oo 3o.o i 100 20 05 0 023
30、025 027 0.22 0.24 0.26 0.28 PARTICLE SIZE IN Fro. 3. Log-probabiliW plot of the cumulative size distribution curve of the sample shown n Fig. 2. 999 -99.0 - 950 800 - 500 - 20.0 5.0 10 0.1 were observed with tetramethyl ester, while tetrapentyl ester reacted slowly (up to 24 hours for quantitative c
31、ondensation) and produced big particles which, in n-propanol and n-butanol, reached sizes somewhat scattered around 2 . More uniform par- ticles were obtained in 1:3 mixtures of methanol-n-propanol. A systematic investigation of the influence of different concentrations of water, am- monia and ester
32、 was made with the ethanol- ethyl ester system. The condensation rate depends strongly upon the water content of the system. In the absence of ammonia, the silica flocculated in irregularly shaped par- titles and no spheres could be observed under the electron microscope. Thus, am- monia apparently
33、influenced the morphol- ogy, and created spherical particles when- ever it was present during the reaction. An increase in ammonia concentration (up to 8 moles/liter) under otherwise constant ex- perimental conditions caused larger par- ticle sizes. Accordingly, the largest spheres were obtained whe
34、n the reaction mixture was saturated with ammonia. When the water concentration under these conditions was varied, maximum particle size was obtained at water concentrations around 6 moles/liter, while different ester concentra- tions between 0.02 and 0.50 mole/liter had no significant influence on
35、the particle size. The three-coordinate graph in Fig. 1 repre- sents the general correlation between par- title size, water, and ammonia concentra- tion obtained with an ester concentration of 0.28 mole/liter. The actual particle sizes observed varied between 0.05 and 0.90 and were uniform in each t
36、est. Figure 2 shows an electron mierograph of a sample from this series, and Fig. 3 indicates the cumulative size distribution curve of the same sample as plotted on log-probability paper. The geometric standard deviation derived from the graph is 1.04. This value was quite typical for most of these
37、 tests. It indicates that only 5% of the particles differ from the median size by more than 8%. The generation of particles larger than 1 could not be effeeted with the ethanol-ethyl ester system but required the use of esters of higher alcohols. Various tests with these esters indicated that, under
38、 comparable conditions, the condensation reaction slowed down with increasing molecular weight of the ester, while, at the same time, particles of larger size with a fair degree of uniformity were produced. The reaction could be further slowed down by using higher alco- hols as solvents. In these ca
39、ses, however, the median particle size and the spread of the size distribution increased simultane- ously. Sometimes the samples also con- tained two distinctly different particle sizes. Thus, special precautions had to be taken to reduce these adverse effects. To achieve this, isothermal conditions
40、 (22C) and homogeneous particle suspension maintained during the reaction by gentle agitation proved effective. An extensive investigation was made with tetrapentyl ester. Particles were grown under various conditions by using different component concentrations and several al- cohols or alcohol mixt
41、ures as solvents. In the simple ethanol-pentyl ester system, at ester concentrations sufficiently low ( 0:03 mole/liter) again had little influ- ence on the final size of the particles, but the particles obtained in these tests were definitely bigger than those grown under comparable conditions from
42、 ethyl ester. The maximum values obtained for fairly uniform batches of particles were about 1.5 in diameter. The sample shown in Fig. 4 was grown from 0.2 mole/liter of pentyl ester in a solvent saturated with ammonia and containing 5 moles/liter of water. The median particle size is 1.38 u as deri
43、ved from the cumulative size distribution curve shown as a log-probability graph in Fig. 5. This graph indicates that the sample consists of two superimposed size distributions; the more prominent one is close to a log-normal distribution having a geometric standard deviation of 1.08 and covering th
44、e upper size range. It comprises about 93% of the particles. This result is quite typical for most of the size distributions obtained in tests designed to produce particles of sizes bigger than 1 a. At high pentyl ester concentrations, the ester formed a separate phase at the bottom of the reaction
45、vessel, thus providing a substrate reservoir for the hydrolysis pro- ceeding in the upper phase. In this way, continued particle growth up to sizes 3 u in diameter was effected, but apparently, the reservoir was also continuously supplying new condensation nuclei so that the particle FIG. 4. Electro
46、n micrograph of a sample of silica spheres obtained in the ethanol-pentyl ester system. CONTROLLED GROWTH OF MONODISPERSE SILICA SPHERES 67 t,.t O1 05 07 09 2 6 20 PARTICLE SIZE IN y FIG. 5. Log-probability plot of the cumulative size distribution curve of the sample shown in Fig. 4. 999 sizes in th
47、e final dense suspension were no longer approximately uniform. Instead, the 99o sizes were scattered in high count frequen- %0 cies between values of 0.5 and 3 u. Emulsi- fying the ester during the reaction by vio- s00 lent agitation did not significantly improve these results. Tests with higher alc
48、ohols or 5O0 their mixtures with methanol proved to be more effective. While butanol gave particle 200 populations of poor uniformity, the tests with propanol were satisfactory. Maximum .50 particle sizes were obtained at an optimum water content of 5 moles/liter and at high !2 concentrations of amm
49、onia and ester. Under these conditions, median diameters were observed between 1.5 and 2 . The best growth of fairly uniform particles of 2 t* diameter was facilitated by using mix- tures of methanol with butanoI (1:1) and propanol (1:3) as solvents. Yigure 6 shows TABLE I EXPERIMENTAL ARIANGEMENT FINAL /EDIAN DIAMETER AND GEOMETRIC STANDARD DEVIATION OF SILICA SPHERES Test Components i 2 3 Final pH () eom 6a) ogeom 0z) Cgeom 10 ml pentyl ester; 10ml ammonium hydrox- 12.7 1.53 1.06 1.35 1.06 1.22 1.10 ide (sat.); 50 ml propanol satur
