Langmuir 1995,11, 4793-4796
4793
Heat of Adsorption and Thin Film Surface Area Studies of a Silica Sol-Gel Film Exposed to HCl and H20 Kent B. Pfeifer" Sandia National Laboratories, Albuquerque, New Mexico 87185 Received June 19, 1995. In Final Form: September 5, 1995@ Heat of adsorption and surface area studies have been undertaken for H20 and HCl adsorbing on a Stober sol-gel film coated on a surface acoustic wave (SAW)device. Adsorption isotherms measured with the SAW device are compared to the adsorption model of Brunauer, Emmett, and Teller (BET). The heat of adsorption for the first monolayer of H2O and HC1 was found to be 14.6 and 5.3 kcal/mol, respectively. The surface area of the film was also determined using the BET model. Results are compared to surface areas calculated from the film thickness and particle size.
Introduction
reported previ~usly.~ The system used two SAW devices: one was exposed to the atmosphere of interest, while a second Surface acoustic wave (SAW) technology has been "reference" device was sealed in an inert environment. Each applied to the problem of H2O contamination in semiSAW device was used as the feedback element of an oscillator conductor process gases.' Stober silica sol-gel-coated circuit operating at approximately 97 MHz. The frequency SAW devices have been used to detect H2O in inert outputs from the two oscillators were then mixed to form a differencefrequency that was measured using a frequency counter environments but have been shown to exhibit no response (HP-5384A). Both SAW devices were mounted in a brass test to H2O in HC1 gas; however, they do respond reversibly fixture whose temperature was actively controlled using a to the presence of HC1in an inert (N2)matrix. The purpose proportional temperature controller. The brass fixture was of this paper is to report the results of heat of adsorption designed to provide excellent radio frequency (rf)shielding for studies that were undertaken in order to characterize the the test SAW devices and to provide sufficient thermal mass to binding of H20 and HC1 by the Stober films. moderate the frequency changes due to environmental temperThe adsorption model of Brunauer, Emmett, and Teller ature fluctuations. This arrangement provided excellent com(BET) is a widely accepted technique for experimentally mon-modeambient thermal response rejection. Short-termnoise determining the surface area and heat of adsorption of fluctuations in the difference frequency were less than 20 Hz and resulted in moderate drift of less than 100 Hz over several the first monolayer of adsorbate onto a surface.2 BET hours. Data were acquired via an IBM PSI2 80386-based measurements are routinely made with N2 as the adsorcomputer running HTBasic (TransEra Corp., Provo, UT) that bate on a surface cooled to 77 K a n d are used to study the interfaced with the frequency counter via IEEE-488 and that properties of materials ranging from soils to ~ o l - g e l s . ~ - ~ also controlledthe gas flow rates from the HzOgeneration system Nitrogen is the adsorbate of choice due to its well-known (Figure 1). adsorption cross section (16.2 A2); however, typical The SAW device based oscillators respond to the adsorption instrumentation designed to make BET measurements is of a chemicalspeciesby changes in the wave propagation velocity. limited to bulk materials and does not allow for measureTo first order, the wave velocity is reduced as material is adsorbed ments of thin film adsorption. The adsorption characonto the surface of the device; this leads to a change in the frequency of oscillation according to the following relationship teristics of a bulk material may or may not be representative of the adsorption characteristics of thin films made from that material. Previously, surface acoustic wave AV = -qm V , 2 Ap (SAW) technology has been used to make BET measurements of thin silicate-based sol-gel films by adsorbing N2 where Ap is the change in surface mass density due to the adsorption of the adsorbate onto the film, Av is the change in in a He matrix onto a test film cooled to 77 K which was frequency due to the adsorbate, qm is the mass sensitivity factor deposited on a SAW device.6 In this paper, the results (1.3cm2s/g)oftheST-cut quartz substrate,899and voisthe baseline from measurements of silica sol-gel film adsorption of frequency of the SAW oscillator (-97 MHz). H20 and HC1 at 40 "C will be reported. In addition, the Common mode effects such as ambient temperature changes surface areas of the films will be calculated and compared are reduced by measuring the differencefrequency between the to estimates of the surface area based on physical exposed SAW device and the reference SAW device. Since the parameters. reference device is isolated from the environment, frequency changes due to adsorption do not occur in the reference oscillator Experimental Section frequency. Frequency changes due to adsorption occur in the measurement SAW device and appear as a similar change in the SAW Sensor System. The surface acoustic wave sensor difference frequency. Frequency changes that occur in both system used to make the measurements was similar to the system oscillators, such as those due to environmental temperature fluctuation, do not appear as a change in the differencefrequency. * e-mail,
[email protected];phone, (505)844-8105;FAX, (505) Moisture Generation System. The H2O generation system 844-1198. consisted of four mass-flow controllers (MKS model 1159B)that Abstract published in Advance A C S Abstracts, November 1, providedvarious mixtures ofH20, HC1, and N2 and were arranged 1995. such that the sensor could be quicklypurged with an atmosphere (1)Heifer, K.B.;Kelly, M. J.;Guilinger, T. R.; Peterson, D. W., Sweet of dry N2. It used tubing of electropolished 316L stainless steel J. N.; Tuck, M. R. Microcontamination Proceedings; Canon Com@
munications: San Jose, CA, 1994;p 87. (2)Sen-Levy, A,; Avnir, D. Langmuir 1993,9, 2523-2529. (3)Chiou, C. T. Enuiron. Sci. Technol. 1990,24, 1164-1166. (4)Wagner, W.; Averback, R. S.;Hahn, H.; Petry, W.; Wiedenmann, A. J. Muter. Res. 1991,6 (lo), 2193-2198. ( 5 ) Tsunoda, R., J. Colloid Interface Sci. 1989,130 (l),60-68. (6) Ricco, A. J.; Frye, G. C.; Martin, S. J. Langmurr 1989,5,273.
(7)Heifer, K.B.;Sprung, J. L.; Galloway, T. R. Sensors Actuators B 1994,22,37-45. (8) Auld, B. A. Acoustic Waves and Fields in Solids; Wiley: New York, 1973;Vol. 2. (9)ST-cut quartz is a singly rotated cut having Euler angles 1 = 0", 6 = go", and p = 132.75".
0743-746319512411-4793$09.00/0 0 1995 American Chemical Society
4794 Langmuir, Vol. 11, No. 12, 1995
Pfeifer
Vacuum
HCl Gas
4-Way
Valve
U
Figure 1. Schematic diagram of flow system used to generate various concentrations of moisture and HC1. Gas sources, purge manifold for the HCl lines, flow controllers, permeation tubes, control computer, and sensor locations are shown. and was assembled using compression-type fittings. Moisture (HzO vapor) was supplied to the system from two permeation tubes (GC Industries models GC23-7322 and GC23-7323) that leaked HzO at nominal rates of 280 and 1475ng/min into the N2 stream passing by the tube. Two permeation tubes were required to generate the required range of moisture concentrations. Analyses of the permeation tube output and the dilution in the system allow calculation of the nominal moisture concentration at the sensor as follows
101
'
'
"""'
1
KL = f i + fi + f 3
where 9 is the moisture concentration in ppm, K is the permeation constant (1.358ppm cm3/ng),L is the leak rate ofthe permeation tube in ng/min, and f i , f2, and f3 are the flow rates from each of the mass-flow controllers in standard cubic centimeters per min (sccm). The supply gases (N2 and HC1) were dried using SemiGas Systems Nanochem gas dryers models 1400 (Nz)and DP-85 (HCl). The level of HzO in the supply N2 is specified to be < l ppb and the level of HzO in the HC1 is specified to be
