Anal. Chem. 1986, 58, 7873-7877 (5) ACRubae, A. Y. Ph.D. Thesis, Pennsylvania State Universlty. 1978. (6) Rinder, D. F.; Flecker, J. R. Bull. Envkon. Contam. Toxlcol. 1981, 28, 375-380. (7) . . Chae, K.; Cho, L. K.; McKinney, J. D. J. Agric. food Chem. 1977, 25, 1207-1209. (8) Wie, S. I.; Hammock, B. D. J. Agric. food Chem. 1982, 30, 949-957. (9) Newsome, W. H.; Shields. J. B. J. Agric. food Chem. 1981, 29, 220-222. (10) Monroe, D. Anal. Chem. 1984, 56,920-931. (11) Newsome, W. H. J. Agric. food Chem. 1985, 33,528. (12) Bullhrant, C. M. Br. Med. J . 1988, 1 . 1272-1273. (13) Smith, P.; Heath, D. CRC CrR Rev. Toxicol. 1978, 4 , 411-445. (14) Staiff, D. C.; Comer, S. W.; Armstrong, J. F.; Woife, H. R. Bull. Environ. Contam. Toxicol. 1975, 14. 334-340. (15) Maddy, K. T. “Occupational Illnesses and Injuries due to Exposure to Paraquat as Reported by Physicians in California”; 1980 and other yearly reports. (16) Lewin, R. Science 1985. 229,257-258. (17) Glri, S. N.; Curry, D. L.; Stabenfeldt, 0.; Spangier, W. L.; Chandler, D. B.; Schiedt, M. J. Environ. Res. 1983, 30, 80-88. (18) Zavala, D. C.; Rhodes, M. L. Chest 1978, 74, 418-420. (19) Gage, J. C. Br. J. Ind. Med. 1988, 25,304. (20) Byass. J. B.; Lake, J. R. Pestic. Sci. 1977, 8 , 117. (21) Jellinek, S. fed. Reg. 1979, 44, 59956. (22) Seiber, J. N.; Woodrow, J. E. Arch. Environ. Contarn. Toxicoi. 1981, IO, 133-149. (23) Khan, S. U. J. Agric. FoodChem. 1974, 22,860-867. (24) Khan, S. U. Bull. Envkon. Contam. Toxicol. 1975, 14, 745-749. (25) Calderbank, A. Envlron. Qual. Saf. 1975, 4 , 136. (26) Pack, D. E. I n Analvtical Methods for PestlcMes, Plant Qowih Regulators, and food AWItives; Zweig, G., Ed.; Academic Press: New York, 1987; Vol. 5, pp 473-481. (27) Soderquist, C. J.; Crosby, D. G. Bull. Envlron. Contarn. Toxicol. 1972, 8, 363-368. (28) Van Dlk. A.; Ebberink, R.; de Groot, G.; Maes, R. A. A. J. Anal. Toxicoi. lb77, I , 151-154. (29) Needham, L.; Paschal, D.; Roiien, J.; Liddle, J.; Bayse, D. J. Chromatogr. Scl. 1979, 17, 87. (30) Pryde. A.; Darby, F. J. J. Chromatogr. 1975, 115, 107-116.
1873
(31) Niewola, 2.; Waish, S. T.; Davies, G. E. Int. J. Immunopharmacoi. 1983, 5: 21 1. (32) Factori, D.; Hunter, W. M. Clin. Chim. Acta 1980, 100, 81-90, (33) Levitt, T. Lancet 1977, 2 (8033), 358. (34) Niewola, 2.; Hayward, C.; Symington, B. A.; Robson, R. T. Clin. Chim. Acta 1985. 148. 149-156. (35) ErbigerlB. F.;-Borek, F i Beiser, S. M.; Lieberman, S. J . Biol. Chem. 1957. 228. 1090-1094. (36) Parker,& ‘W. Radioimmunoassay of BiolOgicai/y Active Compounds; Prentlce-bll: Engiewood Cliffs: NJ, 1976; 239 pp. (37) Voller, A.; Bidwell, D. E.;Bartlett, A. Bull. W. H. 0. 1978, 53,55-65. (38) Durham, W. F.; Wolfe, H. R. Bull. W . H . 0.1982, 25, 75. (39) Analysis of Paraauat ResMues Method RM-8- 10; Chevron Chemical Co.,-July 8, 1975: (40) Chester, G.; Ward, R. J. Arch. Environ. Contam. Toxicoi. 1984, 13, 551-563. (41) Wojeck, G. A.; Price, J. F.; Nigg, H. N.; Stamper, J. H. Arch. Environ. Contam. Toxicoi. 1983, 12, 65-70. (42) Documentation of the ThreshoM Lhnit Values for Substances in Workroom Air; Amer. Conf. Ind. Hygienist: Cincinnati, OH, 1971. (43) DuBois, D.; DuBois, E. F. Arch. Int. Med. 1918, 17, 863. (44) Berkow, S. G. Am. J. Surgery. 1931, X I (2), 315. (45) Draffan, G. H.; Clare, R. A.; Davies, D. L.; Hawksworth, G.; Murray, S.; Davies, D. S. J. Chromatogr. 1977, 139,311-320. (46) Stewart, M. J.; Levitt, T.; Jarvie, D. R. Ciin. Chim. Acta 1979, 994, 253-257. (47) Proudfoot, A. T.; Stewart. M. S.; Levitt, T.; Widdop, B. Lancet 1979, 2 (8138), 330-332. (48) Thornpsett, S. L. Acta Pharmacoi. Toxicol. 1970, 28,356. (49) Glri, S. N.; Lunsman, P. Toxicoi. Leff. 19818 9,93-100. ’
RECEIVED for review November 4,1985. Accepted February 24, 1986. This work was supported in part by the Western Region Pesticide Impact Assessment Program, and NIEHS Grant R01-ES02710-04. B. Hammock was supported by NIEHS Research Career Development Award 5K04ES00107-05.
Quantitative Study of Acoustic Emission from a Model Chemical Process R. M. Belchamber, D. Betteridge, M. P. CoIlihs,* Trevor Lilley, C. Z. Marczewski, and A. P. Wade
BP Research Centre, Sunbury on Thames, Middlesex TW16 7LN, UK
The potentlal of acoustic emission (AE) as a method for monitorlng chemkal processes has been demonstrated. For thls study, a model process, the hydration of slllca gel, has been lnvestlgated In some detaU. As sllca gel hydrates, gas Is evolved whlch causes the granules to fracture and produce sound. integrated AE signals have been used to glve quantitative Information as to the amount of material present, the average particle 8/28, and the percentage wafer content of the partially hydrated siilca gel. Pattern recognition techniques have classlfled four types of AE signals from the reaction. Kinetic lnformatlon has also been obtained about the reactlon mechanism involved by analysls of the t h e intervals between events. This analysis suggests that the hydration of slllca gel is a two-step process wRh a parallel, non-AE-produclng pathway. These results indicate that AE monitoring will prove to be an effectlve method of gaining often urllque Information about many dlfflcult-to-study chemlcal systems.
The purpose of this paper is to investigate the usefulness of acoustic emission (AE)monitoring as a method for studying 0003-2700/86/0358-1873$01.50/0
certain physical and chemical processes. Betteridge et al. (1) have demonstrated that many chemical reactions are accompanied by the emission of acoustic energy. These range from the obvious such as the “fizz” that accompanies effervescent reactions to the low-level emissions that are produced by certain crystalline-phase changes. Other workers have observed AE from the precipitation of dichloro(pyrazine)zinc(II) ( Z ) , the gelation reaction of sodium carbonate and calcium chloride (3),phase transitions of p-cresol and MBBA liquid crystals (4), and the polymorphic transformations of a number of DTA temperature standards (5). AE has found more widespread use in areas other than the strictly chemical; it is used for corrosion monitoring (6) and is routinely used in the plastics industry for studying fracture mechanisms and in nondestructive testing. Applications in these areas have been comprehensively reviewed by Drouillard and Hamstad (7). However, in our opinion, it has not been established whether AE is a useful quantitative tool for studying chemical reactions, or what the appropriate analytical techniques are for extracting useful chemical information from AE signals. For this study, the hydration of silica gel was selected as a model system. This reaction is both acoustically active and not readily amenable to other monitoring techniques. It was 0 1986 American Chemical Society
1874
ANALYTICAL CHEMISTRY, VOL. 58, NO. 8, JULY 1986 PTFE CAP
-@
B W
O0
rI
0
SAMPLE COMPARTMENT
ALUMMUM BLOCK GLASS CELL 4 C d CAPACITY ACOUSTIC EMLSSI)(I
TRANSDUCE
FELT PAD
Flgwe 1. Acoustic emission cell. 00
hoped that experience gained with this system could be extrapolated into other areas. When dry silica gel granules are mixed with water, large volumes of gas (air) are released. The volume of gas produced has been measured and found to be 0.85 mL/g of dry silica gel (25 "C, 1 bar), whereas the pore volume, measured by a volumetric nitrogen adsorption technique (81,was 0.36 mL/g. If this gas cannot escape quickly enough the internal pressure builds up and the granule fractures with a clearly audible report. Lower level emissions caused by the formation and bursting of the gas bubbles are also detectable. AE signals are very rich in information, making their interpretation difficult. Integration of the measured AE signal is sufficient to obtain simple quantitative information. However, if we limit ourselves solely to making this type of measurement considerable useful data are being disregarded. In most systems AE is not a continuous process but a series of discrete events, in the case of silica gel the fracture of a granule or the bursting of a bubble. Analysis of these individual signals, particularly in a pattern recognition sense, may then be used to deduce the dominant AE-producing mechanism. This type of analysis has been used in our laboratory to characterize the AE signals produced as various composite materials start to fail under load (9-11). As the reaction progresses, the rate a t which AE occurs changes; analysis of the time intervals between adjacent events can be used to study the kinetics of the processes involved. With the silica gel system, we have chosen to follow these three types of measurement and analysis, viz., integration of the measured AE signal, pattern recognition on the signal, and analysis of the time intervals between adjacent AE events, to evaluate the potential of AE monitoring for an otherwise difficult system to study. EXPERIMENTAL SECTION 1. Reagents. Self-indicating, desiccant grade silica gel (Fisons,
UK) was used. This was ground and sieved according to the requirements of the experiment. Drying was accomplished by baking the prepared silica gel for 16 h at 150 OC and then cooling in a vacuum desiccator. 2. Acoustic Emission Cell. A purpose built "AE cell" was constructed out of a solid aluminum block (Figure 1)such that acoustic and electromagnetic interferencewere minimized. It was equipped with a reference compartment containing a similar transducer so that any remaining interferencecould be identified. In practice, the hydration of silica gel produced AE of sufficiently high energy to allow low amplifier gains to be used, and interference did not pose a problem. 3. Measurement and Integration of AE Signals. The acoustic emissions were detected by using a broad-band piezoelectric transducer (Bruel and Kjaer, Denmark, type 8312) quipped with an integral 40-dB preamplifier. The signal was further amplified (10 dB) and band-pass filtered (100 kHz-2 MHz) by using an acoustic emission amplifier (Bruel and Kjaer, type 2638).
I 02
I Ob
I 06
WEIGHT OF SILICA GEL
I
I
OB
10
[QI
F@we 2. Integrated acoustic emission as a function of sample weight.
This amplifier, in addition to having a normal ac output, has a "peak detector" output, which gives a heavily damped dc signal proportional to the peak amplitude of the AE signal. For integration purposes this output was sampled at 100 Hz by the analogue-&digital converter and integrated over an appropriate interval, usually 60 s, by which time the AE had died down to a negligible level. 4. Acquisition of Discrete AE Signals. Individual signals were acquired from the ac output of the AE amplifier using a transient recorder (Data Laboratories, type DL 920), a sampling rate of 10 MHz being used. The transient recorder was interfaced to a computer via an IEEE-488 bus. Digital records of the signals were transferred to the computer and stored on disk for subsequent analysis. 5. Timing AE Events. The apparatus for timing AE events has been previously described in detail (9). The amplified and fiitered ac signal was fed to a computer that allowed signah above a selected amplitude to be detected. Multiple firing of this comparator by a single AE was prevented by using a monostable to provide a 10-ms dead time between detecting one event and the next. A computer-controlledclock module (Digital Equipment Corp. (DEC),type MNCKW) was then used to time the intervals between AE events. RESULTS AND ANALYSIS 1. Quantitative. Three experiments were carried out using
the integrated AE signal in an effort to establish whether it could be related in a quantitative manner to the silica gel system. The first of these was to seek whether a quantitative relationship could be established between the integrated AE and the mass of reactants present. Aliquots of dried silica gel of similar particle size were accurately weighed and then added to 20 mL of water in the cell, the resulting signal being integrated over 100 s. The results are plotted in Figure 2. The measured signal increases as the weight of silica gel is increased, although the relationship is not linear. Presumably, this is due to the average center of acoustic activity moving further away from the transducer and also large attenuation of the AE signal in the increased thickness of silica gel. The effect of water content of the silica gel was investigated. Samples of silica gel (0.1 g dry weight, 1.0-1.2-mm size range) were hydrated by exposure in a moist laboratory atmosphere for various lengths of time. The samples were reweighed and then added to water and the AE integrated for 60 s. The results are plotted in Figure 3. Results from this experiment and the previous one indicate that provided the mass (and volume) of the reactants is kept constant, a linear relationship can be drawn between the integrated AE and the original water content. No AE activity was detected when the water content exceeded 23% w/w. This corresponds to the amount
ANALYTICAL CHEMISTRY, VOL. 58, NO. 8, JULY 1988
1875
PC 1
'b
0
I I 00
I
I 8
I
I
12
I 20
16
WATER CONTENT OF SILICA GEL ( X w/wl
Flgure 3. Integrated acoustic emission as a function of initial water content of the slllca gel.
-t
04-
Lo
3 Y
03-
