OCTOBER 1983 Vol. 108 No. 1291 Piezoelectric Crystals for Mass and Chemical Measurements A Review John F. Alder and John J. McCallum Department of Instrumentation and Analytical Science University of Manchester Institute of Science and Technology P.O. Box 88 Manchestev M60 lQD Summary of Contents Introduction Historical Theory Quartz crystal microbalance Adsorption desorption and decomposition studies Aerosols and suspended particles Electrogravimetric analysis Piezoelectric crystal detectors for chromatography Gas detection Ammonia and other nitrogen compounds Aromatics Dissolved carbon dioxide Carbon monoxide Cyanide Explosives Hydrocarbons and halogenated hydrocarbons Hydrogen Hydrogen chloride Hydrogen sulphide Mercury Methane and other hydrocarbons Organophosphorus compounds Sulphur dioxide Toluene diisoc yana t e Water Bacterial and fungal growth Solution property measurement Trace metal studies Thermal analysis Conclusions Keywords Review ; fiiezoelectric crystals Introduction In the past 100 years or so piezoelectricity has gone from a scientific curiosity to a widely exploited phenomenon with applications in both science and engineering.Since 1977 Guilbault and co-workers have published four reviews on piezoelectric crystals applied to chemical analysis two of which are essentially reviews of his group’s work,lP2 the other two being more comprehensive containing 673 and 44 references4 Bearing in mind the previous reviews we have tried to view the subject from a slightly different aspect bringing out what we consider to be the important practical aspects of the use of piezoelectric crystals in mass and chemical measurements.116 1170 ALDER AND MCCALLUM PIEZOELECTRIC CRYSTALS Analyst vd. 108 Historical Cady,5 in his definitive work on piezoelectrics relates that Coulomb was the first to conjec-ture the possible production of electricity by the application of pressure on a suitable material. Hauy and later BecquereP performed some experiments in which certain crystals showed electrical effects when compressed. Their findings reported positive results with non-piezoelectric crystals such as calcite and led to the conclusion that what they had observed was “contact electricity.” The credit of being first to observe the phenomenon of piezoelectricity falls to the Curie brothers Pierre and Jacques in 1880.’ They showed that when some crystals were compressed in particular directions an electrical potential was produced between the deformed surfaces, this potential being proportional to the applied pressure.The converse effect unforeseen by the Curies was predicted by Lippmann.8 By the end of 1881 the Curies had verified this effect and showed that the piezoelectric coefficient of quartz had the same value for the direct and converse effects. LangevinQ employed quartz plates to serve as emitters and receivers of high-frequency waves under water and this led to the development of sonar. Cady,lo*ll Pierce12*13 and others produced crystal-controlled oscillators with high stability for use as tuning devices and crystal filters for com-munication lines and radio This technology is still in wide use today and is the reason for the availability of cheap very high quality crystals for analytical work.It was found that by altering the cut angle of the quartz crystal with respect to its optical axis the temperature coefficient of the crystal could be changed. This discovery led to a wide range of crystal cuts (the “T” series of Y-cuts) that gave rise to crystals with properties suitable for many applications. Piezoelectricity remained a curiosity until the start of the First World War. Theory Crystals possessing the property of piezoelectricity can be predicted from crystallographic studies. Piezoelectricity occurs in crystals that do not possess a centre of symmetry of which there are 21 cla~ses.1~ The piezoelectric effect arises when pressure on a dielectric material deforms the crystal lattice and causes a separation of the centres of gravity of oppositely charged species which gives rise to a dipole moment in each molecule.Early mathematical treatment of this effect can be found in works by Cadys and V ~ i g t . l ~ ~ l ~ If electrodes are applied to the faces of a thin slab or rod of this material and an external current sensing circuit is connected a current will be seen to flow through the external circuit when stress is applied to the crystal. Releasing the stress causes a transient current flow in the opposite direction. If the converse effect is used and an alternating potential difference applied mechanical oscillations occur within the crystal lattice.Stable oscillations only occur at the natural resonant frequency of the crystal and at that frequency the crystal presents a low impedance to the exciting voltage. If the crystal is incorporated into the feedback loop of an oscillating circuit it becomes the frequency determining element of the circuit as its Q (quality factor) is very high typically several thousand. Quartz crystals used commercially are usually made from plates that were cut from a single crystal at specific angles to the principal optical axis. These blanks are then ground to specification by etching and lapping techniques. The blanks are chosen to contain no left-handed or electrically twinned material. The first oscillators employed crystals cut in the y - z plane with the electric field applied along the x-axis as the frequency determining element.This resulted in longitudinal vibra-tion along the y-axis and consequently length was the frequency-determining parameter and meant that high frequencies could not be obtained as the lengths required were too small to be practical. Initial work with thickness as the frequency determining dimension was plagued by inter-ference from harmonics and overtones. To overcome this Y-cut crystals vibrating in the shear mode were emp1oyed.l’ These crystals gave good results with few interfering vibration modes but were very sensitive to external temperature variations. Lack et aZ.18 showed that the temperature coefficient and frequency constant varied with the angle of rotation.Two angles were shown to have zero temperature coefficients +35”15’ (AT) and -49’00’ (BT) October 1983 FOR MASS AND CHEMICAL MEASUREMENTS. A REVIEW 1171 For the majority of piezoelectric work in the analytical chemistry field AT-cut crystals have been used. Literature covering the response shape ageing and performance of these crystals is outside the scope of this work but may be found in the electrical engineering literat~re.19-~4 I t has been known that by adding mass to the surface of the crystal one obtains a lower oscillation frequency; indeed this effect has always been used by crystal manufacturers to adjust a crystal to the desired frequency. Sauerbre~~59~~ utilised this idea as the basis of a sensitive microbalance. He derived an expression relating the change in frquency to the mass of material deposited.This expression only applies to AT-cut crystals vibrating in the thickness shear mode the electric field being applied along they-axis. The frequency ( F ) can be expressed as F = - = - v t r N 2t t where Vtr is the velocity of propagation of a transverse wave in the plane of the crystal t is the thickness and N is the frequency constant. With suitable mathematical manipulations one obtains the equation in the form where M is the mass of the plate p is its density and A is the surface area. Sauerbrey assumed that if the plate was divided into an infinite number of parallel planes along the x - z axis only those planes close to the surface would affect the frequency through their mass and not through their elastic character.The frequency change resulting from the deposition of a thin uniform film of any foreign substance is equivalent to that of a layer of quartz of the same mass. Sauerbrey developed the above equation to give a more general form : For pure shear-mode vibrations the strains are all zero at the principal faces. d F dM,F -=-F APN where dM is the mass of the deposited material. Substituting for the various constants gives for quartz where d F is the change in frequency due to the coating (Hz) F is the frequency of the plate (MHz) dM is the mass of coating deposited (g) and A is the coated area (cm2). From this, one can predict a mass sensitivity of 10-9 g Hz-l for a 10-MHz crystal. Lostis2' and later Stockbridge28 also treated the piezoelectric effect mathematically with similar results.Quartz Crystal Microbalance Oberg and L i g e n s j ~ ~ ~ used thickness shear mode AT-cut 3.5-MHz crystals as a film thick-ness monitor claiming a 1 Hz A-1 thickness increment and a total thickness increment of 2 pm. Warner and Stockbridge30 used 2.5-MHz fifth-overtone and 5-MHz fundamental crystals to measure small masses of material adsorbing on to surfaces in a vacuum system. At 2.5 MHz they claimed practical sensitivities of &lo pg cm-2 with temperature stabilisa-tion to *0.01 "C. The effect of mass on the resonance frequency was examined by Stockbridge.28 The thick-ness shear AT-mode was considered in detail and a perturbation analysis was used to derive the relationship between the frequency shift AFIF and the mass added uniformly over the electrode(s) AMIA.In the context of biomedical applications of piezoelectric devices King31 suggested inserting a vibrating crystal into the biological fluid to be examined and allowing some material t 1172 ALDER AND MCCALLUM PIEZOELECTRIC CRYSTALS Analyst VoZ. 108 adsorb on to a coating on the resonator; a low-frequency vibrating crystal was suggested for use with liquids. Richard~on,~~ reviewing King’s paper,31 noted that this was a good idea in theory but for practical purposes difficult to implement for biological systems. The effects of stressing materials on a quartz crystal microbalance were examined by Ullevig et aZ.33 They found that the radial sensitivity function of a quartz microbalance was altered by the presence of a heavily stressed film on one face of the crystal.The sensitivity of the centre of the crystal was found to increase drastically while the peripheral area became insensitive to mass changes.26 The over-all (integral) sensitivity of the crystals to a mass change was shown to be unchanged. They emphasised the need for a uniform layer over the entire active area of the crystal for meaningful data. Adsorption Desorption and Decomposition Studies Haller and White34 investigated the kinetics of formation of a butadiene polymer film exposed to 250-eV electrons at pressures between 3 x lob4 and Torr these being the conditions occurring in electron microscopes. They reported that the rate of growth of the film was pro-portional to the square root of the current density.The rate was found to increase with in-creasing pressure becoming independent at higher values. The oxidation stability of elastomers was studied by Fischer and King.35 Thin films of the elastomers 2 4 p m thick (25pg) were deposited on the surface of 9-MHz quartz crystals. Changes in the mass of the rubber due to oxidation at 150 “C were detected and recorded as frequency changes over test times of 0.5-3.0 h. The effects on the oxidation of the elastomer with respect to filler type vulcanisation inhibitors temperature and ultraviolet radiation were explored and reported. Oxygen absorption and volatility properties of submicron films of asphalt were examined by King and Corbett .36 They showed that polar aromatics absorbed oxygen very quickly, whereas asphaltenes absorb at a slow but prolonged rate.Saturates and naphthene aromatics showed virtually no absorption of oxygen but were found to be volatile. These findings were consistent with studies using other techniques. H~sseiny~~ produced adsorption isotherms for various gases (carbon monoxide carbon dioxide oxygen and ethylene) on various manganese - phosphorus compounds coated on quartz piezoelectric crystals. The ability to make these sensitive mass measurements in enclosed spaces on a microscale is one of the great attractions of the quartz crystal micro-balance. The current interest in thin-film technology for a wide variety of applications offers great potential for the exploitation of this ability in studying simple chemical systems. Aerosols and Suspended Particles The detection quantification and identification of the components of aerosols and suspended particles in air streams have always been of some importance with respect to health and safety.Indeed as the properties of aerosols have become better understood and their role in industrial toxicology better appreciated this importance is growing. Chuan3* measured the mass of airborne particles (“micro-grit” aluminium oxide) impinging on a 10-MHz crystal coated with an adhesive layer. The device was reported to yield particu-late mass concentrations down to 1 pg m-3 and could resolve individual particulates to a mass of 10-l1g. The adhesion of smaller particles was measured by integrating the frequency change. Chuan estimated a detection limit of 0.1 pg m-3 down to a minimum dimension of 1 pm although “implicit in this estimate is the assumption that the impaction efficiency is unity for all particles captured by the crystal an assumption that is not valid for particles less than about 1 pm size.” The author used temperature compensation via a second crystal not exposed to the aerosol stream and having the same temperature coefficient.We have found in this laboratory that such temperature compensation is not easy over a wide temperature range at the level of precision required for very precise work. Careful ageing and matching of the crystal tempera-ture coefficients is required if the compensation circuit is to be kept simple; otherwise tempera-ture measurement and calculated corrections will need to be employed.That is not a serious problem but will increase the complexity of data recovery systems. Olin and co-workers39@ used an electrostatic precipitator in conjunction with the quartz crystals claiming that the aerosol concentration could be measured with 5% accuracy in abou October 1983 FOR MASS AND CHEMICAL MEASUREMENTS. A REVIEW 1173 10 s. These workers examined car particulate emissions cigarette smoke and laboratory and office aerosols. In a comprehensive and detailed paper,39 they reported mass sensitivities to tobacco smoke of 15.5 and 179 Hz pg-l for 1.5- and 5.0-MHz AT-cut fundamental-mode crystals close to the theoretically predicted sensitivities ; the aerosol average concentration ranged between 7 and 14 pg m-3. Both the 1.5- and 5.0-MHz crystals stopped oscillating when the mass loading reached about 400 pg corresponding to about 6 and 60 kHz frequency change respectively.The frequency change was only slightly non-linear with added mass over that range right up to the point of cessation of oscillation. Using a 1.6-MHz AT-cut crystalq0 they claimed a sensitivity of 5.5 Hz pg-l for tobacco smoke and room aerosol. A piezoelectric device was produced to measure the size concentration of particles produced in the cloud generated by an underground nuclear explosion with the formation of a crater.41 Measurable signals were obtained from 5-pm particles having masses of the order of 3 x 10-10 g ejected a t velocities of about 210 m s-l. An instrument for the determination of the particle size distribution of aerosols was produced by Carpenter and Bren~hley.~~ Particles were deposited on to 10-MHz AT-cut quartz crystals by inertial impaction.The impactor comprised of four stages with cut-off sizes for unit density particles of 2.5 6.3 12.6 and 18.9pm. A flow-rate of 0.51 min-l was used and mass sensitivities of between 140 and 2303 599 Hz pg-l were obtained. The theoretical value from Sauerbrey’s equation was 923 Hz pg-l. The performance of quartz crystals to determine aerosol mass concentration was investigated by Daley and Lundgremq3 They examined the influence of temperature humidity particle collection characteristics response linearity and mass sensitivity. They reported that the mass-sensing ability decreased for particle sizes beginning at ca. 2 pm diameter and reached essentially zero at 20 pm.Sensitiv-ity in the 2-20 pm range could be improved by use of a viscous coating. These workers give a refreshingly detailed consideration of the effects of relative humidity. Using 10-MHz AT-cut gold electrode crystals coated with silicone grease the maximum slope of the response curve [frequency change veysws relative humidity (RH) of a clean humid atmosphere] was -0.2 Hz (yo RH)-l. Uncoated 5-MHz AT-cut platinum electrode crystals showed variable response between +5 and -5Hz over the range 12-80% RH. They also measured the effect of relative humidity on crystals that had aerosol deposited on them. Some materials gave a gradual increase over the range 5-90yo RH [room aerosol road dust calcium sulphate, uranine (a dye) methylene blue - uranine mixture] whereas sodium chloride and ammonium sulphate gave little or no response ta humidity changes up to 60% RH (sodium chloride) and 70% (ammonium sulphate) and then responded catastrophically at RH values above this, obviously by uptake of water.A coating of dimethylpolysiloxane took up less than 3% of the coating mass from atmospheres ranging from 5 to 85% RH. These humidity response characteristics little discussed in the gas-sensing literature are of considerable importance in the design of piezoelectric crystal detectors. Sem and Tsur~bayashi~~ used an electrostatic precipitator to deposit dust particles on to a piezoelectric microbalance sensor. A portable system (commercially available) developed is reported to be able to measure mass concentrations in the 100 pg m-2 range of airborne dust particles smaller than 10 pm.The major attractions of aerosol measurement using quartz crystal microbalance techniques are well illustrated in the literature and are as follows high potential sensitivity; good agree-ment with theoretical response thus largely obviating the need for continuous mass calibra-tion; wide mass and frequency response range and compatibility with the filtration and precipitation methods employed. The low-cost availability of these precision-engineered components makes them very attractive particularly linked with the practicality of disposal rather than having to clean contaminated sensors. Developments in the electronic data processing and miniaturisation of all the electrical components associated with these measure-ments wilI ensure their continued application to aerosol monitoring.Electragravimetric Analysls AT-cut quartz crystals of 1.65,3,6 and 9 MHz have been used as the cathodes in an electro-chemical ce11.45946 A current was allowed to pass for a known period of time and the crystal was removed from the cell washed and dried. The mass increase was then determined from the change in frequency. Cadmium solutions over the range 5 x t o 5 x M were examined 1174 ALDER AND MCCALLUM PIEZOELECTRIC CRYSTALS Analyst VoZ. 108 Nomura and Mimatsu4' used crystals with silver on platinum-plated gold electrodes to determine the iodide in solution. The iodide was electrodeposited at -0.05 V in a sample solution containing M potassium chloride adjusted to pH 9.8 with M sodium tetra-borate(II1) - sodium hydroxide solution.The reagent blank was passed through the detector cell until constant frequency then the sample solution containing iodide was allowed to pass for 1 min (10-6-10-5 M) or 10 min (10-7-10-6 M). The crystal surface was cleaned of iodide by electrolysis at -0.4 V after each determination. Although electrogravimetry has always played an important role in assay and certain other analytical procedures it is debatable whether the use of quartz crystals provides any major advantage other than the important one of miniaturisation and micro-determination. Coulo-metry and stripping voltammetry equal or better this performance. Piezoelectric Crystal Detectors for Chromatography The quartz piezoelectric crystal microbalance has potential as a detector for chromatography owing to its mass sensitivity simplicity and ruggedness all of which recommend it for portable and on-site remote operation.King,48 in 1964 was the first to use quartz crystals in this fashion. He used gas chromato-graphy (GC) stationary phases as the crystal coating material to improve the selectivity for certain groups of compounds. Com-parative chromatograms of a series of hydrocarbons were produced, Suggested coatings and their use are shown in Table I. TABLE I COATINGS SUGGESTED BY KING48 Coating material Detector characteristic Squalane Silicone oil Apiezon grease Polyethylene glycol Sulpholane Dinonyl phthalate Aldol 40 Tide (alkyl sulphonate) Hydrocarbon detection non-selective to compound type Selective detection of polar molecules such as aromatic, oxygenated and unsaturated compounds Silica gel Water vapour Molecular sieve Alumina Hygroscopic polymers (e.g., natural resins glues cellulose derivatives and synthetic polymers) Lead acetate Hydrogen sulphide Metallic silver Metallic copper Anthraquinonedisulphonic acid A portable room-temperature gas chromatograph was produced by Karasek and Gibbin~.*~ They reported that the instrument response was rapid and proportional to the sample size.The sensitivity was found to be in the parts per million concentration range. Compounds in the boiling range 40-200 "C were separated on columns of less than 50 cm in length with short retention times. They reported that their system could be applied for the separation of com-pounds normally unstable above room temperature.A short review of ambient-temperature gas chromatography using the piezoelectric detector appeared in 1972.50 Karasek and co-workers produced two papers dealing with the performance of the crystal detector. In the first paper,51 the crystal was coated with the same material as the column stationary phase to compensate for any loss of coating from the crystal. They reported per-formance characteristics and minimum detectable concentrations for a number of material October 1983 FOR MASS AND CHEMICAL MEASUREMENTS. A REVIEW 1175 (alkanes aromatic hydrocarbons 2-alkyl ketones and alkyl esters) over the range 2 x to 8 x 10-gg. Their second paper52 examined the effect of temperature on the detector over the range 25-100 "C.They found that the response of a piezoelectric detector for a compound eluted from a GC column was where A is the peak-area response W the total mass of the eluent y the activity coefficient of the eluent in the crystal coating PO the vapour pressure of the eluent at the operating tempera-ture F the carrier gas flow-rate and C a constant characteristic of the detector temperature of the crystal and the liquid phase used to coat the crystal. The authors pointed out the obvious consequence that as the vapour pressure of eluted compounds decreases the detector sensitivity increases and yet an adequate component vapour pressure is a requisite for GC. This trade-off requirement will therefore result in a relative narrowing of the useful range over which GC can be used effectively with adequate detector sensitivity.It is worth mentioning that this is true for a reversible system where the adsorption - desorption is physical in nature (i.e. Van der Waal's forces are predominant). If the adsorption involves chemical reaction sensitivity will be increased for components with higher vapour pressure but at a certain loss of reversibility which is of course unacceptable in a GC application. This problem pervades the whole of coated piezoelectric crystal gas detection and monitoring. The temperature of the detector is also important as one would imagine even though the authors remarked that this effect was found not to be as pronounced as was predicted from theoretical considerations. Although the reason for this was not clear it could be due to the way the effluent gas impinged upon the crystal or to the variation of the response function used to predict behaviour with temperature.Janghorbani and F r e ~ n d ~ ~ in a carefully developed paper described the response charac-teristics of a coated quartz piezoelectric crystal in terms of a partition detector for vapours dissolved in a gas stream. In developing the theory of such a partition detector they refer to chromatography theory and relate the peak-area response of the detector to an imaginary plug of the gas mixture : where A is the area under the peak due to component Y m is the constant describing the frequency change (Af) of the crystal due to an incremental mass addition (AW) to the surface (Af= mAW) and KY,X is the partition coefficient of gas Y in liquid X describing the ratio where Wy,x is the mass of gas Y in a unit volume of crystal coating material X (at equilibrium) and W is the mass of gas Y in a unit volume of gas phase.V is the volume of the liquid coating X present on the efective surface area of the crystal W is the mass of gas contained within the detector volume at equilibrium with the liquid coating and F is the flow-rate of the gas phase. The authors presented data to show good linear relationships between A and injected volume of octane hexane and pentane into a gas stream passing over a squalane-coated crystal to support their predictions. They went on to discuss the theory of response time and made some theoretical predictions, concluding that sub-second response times can be obtained by using small detector volumes.The effect of temperature on sensitivity [S Hz(conc.)-l] was also derived theoretically resulting in an equatio 1176 ALDER AND MCCALLUM PIEZOELECTRIC CRYSTALS Analyst VoZ. 108 S = HT exp A / T where A is a constant in the Clausius - Clapeyron equation : A LnP = B - T where P is the vapour pressure of solute (analyte gas in flow stream) at temperature T B is a constant and where m is defined above R is the gas constant and V is the volume of liquid coating (defined earlier) characterised by density p and relative molecular mass M I in which the analyte gas has an activity coefficient y. These equations are of considerable consequence and it is perhaps regrettable that in the literature the coefficients As the authors mentioned KY,X is directly proportional to the (chromatographic) retention volume for a given solute - substrate partition system.The authors also referred to the obvious interference problems when two solutes are present in the same element of gas stream passing through the detector; interestingly they said that, contrary to theoretical prediction “the frequency response to composite samples is invariably greater than the sum of frequency shifts due to each component when introduced separately.” Edmonds and West54 also considered the coated piezoelectric crystal as a partition detector. Starting with much the same premises as Janghorbani and Freund,S3 they considered the coated crystal exposed to an atmosphere containing the analyte vapour as a “simple static two-phase system in which a series of successive equilibrations take place.” They went on to derive theoretical response curves for the leading edge of the response veysus time curve and showed that the theory closely related to practice thus supporting the findings of the previous who used Fick’s law to derive their relationships.The limiting forms of the response equation by the two sets of workers relating frequency change54 or integrated peak area (frequency change x time)53 to concentration of analyte gas in the vapour phase are the same. Theory was shown to describe practice reasonably for n-alkanes on squalene5* and chloroform on Carbowax 20M and dinonyl phthalate. Edmonds and West went on to illustrate the importance of keeping the detector cell volume low to maximise response speed and also the need to maintain the coating material only on the most sensitive area of the crystal near the centre.26 They also illustrated a very important facet of piezoelectric devices the ease with which multi-component mixtures may in principle, be analysed.Hexane and chloroform possess different response characteristics on squalane and Carbowax coatings. Using this property an unseparated mixture of the two compounds in air was analysed yielding results that were in agreement with the known values. Adsorption measurements using a crystal are more difficult to obtain in a liquid than in the gas phase as there are greater energy losses at the crystal - liquid interface. This makes it more difficult for the crystal to maintain a stable oscillation.Schulz and King55 and Konash and B a ~ t i a a n s ~ ~ described two ways around this problem. Schulz and King55 sprayed the effluent from the liquid chromatograph on to the crystal surface the solvent evaporated and the mass of any residual solute determined from the change in oscillation frequency; a complete cycle took 60 s. Using an ethylene - propylene terpolymer in hexane a sensitivity of 0.2 Hz p.p.m.-1 of solute in hexane was obtained. They chromato-graphed butyl rubber and a mixture of polystyrene by gel permeation. The results obtained by this method compared favourably with those from a refractive index detector and better base-line return was claimed for the crystal. Konash and Bastiaans56 produced a novel cell design in which the crystal used had a resonant frequency of 7 MHz in air at 25 “C and the method of mounting appears to have compensated for the increase in energy loss by allowing the liquid to come into contact with only one face of the crystal.The data they presented are scant and inconclusive and the cause of the observed frequency changes when various solutes are added is far from clear. and m are not better defined in many instances October 1983 FOR MASS AND CHEMICAL MEASUREMENTS. A REVIEW 1177 Gas Detection Kin@’-59 proposed the use of piezoelectric crystals as sorption detectors in which the crystal is coated with a substrate that will react with or adsorb the material of interest. This sorption is a mixture of chemisorption and physisorption.Ammonia and Other Nitrogen Compounds Guilbault and co-workers were able to detect ammonia and nitrogen dioxide in the parts per lo9 (p:p.b.) range using coatings of modified Ucon 75-H-90000 or Ucon LB-300X.60 These materials were activated by passing nitrogen dioxide over the coated crystal surface for about 5 min. Infrared spectra of the exposed coatings indicated that new compounds had been formed that contained covalent nitrite groups and possibly nitroso groups. The coatings also proved to be sensitive to moisture. Ascorbic acid ascorbic acid - silver nitrate and an extract of Capsicum annuurn pods were also tried.61 Three extracts of the pods were obtained in alcohol chloroform and toluene. The chloroform extract gave a frequency change of 45 Hz for 10 p.p.m.of ammonia and 15 Hz for 0.001 p.p.m. whereas the alcohol extract gave 120 and 50 Hz respectively for a 10-ml gas sample. The toluene extract proved to be inactive. Some silver nitrate was added to the alcohol extract and this gave frequency changes of 75 and 15 Hz respectively for 1-ml samples of ammonia. The authors suggested that the active component could have been ascorbic acid and there-fore it was tried as a coating material. This gave as a response a 160-Hz change for 10 p.p.m. and 70 Hz for 0.001 p.p.m. on a 1-ml ammonia sample. The addition of silver nitrate gave 320 and 100 Hz changes respectively and a detection limit of 1 p.p.b. of ammonia was quoted for this coating. Nickel dimethylglyoxime was used as the coating in an attempt to determine the concentra-tion of ammonia in solution.s2 The effects of moisture were minimised by the use of hydro-phobic membranes between the sample solution and the crystal.Alternatively the crystal was allowed to reach equilibrium above a sample of distilled water and then transferred into the sampling chamber. A concentration of 0.15 M of aqueous ammonia was reported to give a frequency change of 135 Hz. The calibration graph was reported to be linear up to 0.45 M ammonia. L-Glutamic acid hydrochloride and pyridoxine hydrochloride were reported to have greater sensitivity and better selectivity to ammonia in ambient air than previously (or subsequently) obtained.63 Response times for both coatings were less than 1 min with complete reversibility of response after 5 min.No significant interferences were reported from other gases as long as “the effect of moisture from air was eliminated using a gas-chromatographic pre-column packed with silica gel. ” “Using pyridoxine hydrochloride as the coating material a very high sensitivity of detection of ammonia was found even in the parts per trillion range. At 1 p.p.m. of ammonia a frequency change of 1190 Hz was observed. At 0.01 p.p.m. of ammonia the frequency change obtained was about 386 Hz.” A plot of log(frequency change) veysus log (concentration) was reported to be linear over the range from 0.01 p.p.b. to 1 p.p.m. of am-monia. I t is not clear how stable 0.01 p.p.b. ammonia solutions in even silica-gel dried air would be and the lack of reproducibility data in this work engenders some scepticism.Whereas the use of dried air to determine equipment response characteristics is acceptable one should beware of trying to extrapolate results so obtained to a constant humidity atmosphere and even more to a real atmosphere where humidity changes between 10 and 95% may occur over very short periods in portable applications. For many analyte gases particularly those susceptible to hydrolysis or those which are polar drying of samples by silica gel or molecular sieves results in major losses of analyte. Edmonds et ~ 1 . ~ ~ reported using polyvinylpyrrolidone as their coating. A frequency change of the order of 700 Hz was noted for 1 p.p.m. of ammonia. Moisture present was monitored using a resistive-type humidity probe and a silver chloride-coated piezoelectric crystal.Gas-phase reactions of mono- di- and trimethylamine with various metal salts using a piezoelectric crystal were reported by Guilbault et a1.G5 The coating materials tested included iron( 111) chloride zinc chloride mercury(I1) bromide cobalt (11) chloride and zinc iodide. These workers found that for all three amines the iron salt was the most sensitive. Isotherms covering the range 20-100 mmHg pressure of the amine were drawn 1178 ALDER AND MCCALLUM PIEZOELECTRIC CRYSTALS Analyst VoZ. 108 Aromatics A Nujol mixture of trans-chlorocarbonylbis(triphenylphosphine)iridium(I) [trans-IrCl(C0) -(PPh,),] was used by Karmarker and Guilbault.66 This coating was found to be reactive to aromatic hydrocarbons but not as sensitive to aliphatics.Hence aromatics such as xylenes, benzaldehyde 1,3,5-trimethylbenzene anisole and butylbenzene could be detected whereas compounds such as hexane heptane octane and cyclohexane could be detected only at high concentrations. The effect of atmospheric moisture was reported to be nil; it is unfortunate that data to show this were not included in the paper. Toluene in ambient air was monitored using a Carbowax 550-coated crystala7 installed in a portable device. This coating gave a linear response over the range 30-300 p.p.m. No inter-ferences were observed from inorganic gases such as carbon monoxide sulphur dioxide, ammonia or nitrogen dioxide at 1000 pap.m. Organic vapours gave some interference but were insignificant at the 5% V/V level. Interference from moisture was eliminated using Nafion tubing (Du Pont Type 811).As the water is adsorbed and permeates across the walls of the tubing it is removed either by a counter-current flow of dry nitrogen sweeping the tubing or by use of a desiccant. Dissolved Carbon Dioxide Fogleman and Shumanaa examined the possibility of continuously monitoring the concentra-tion of carbon dioxide in water. AT-cut 9-or 15-MHz crystals with gold electrodes were coated with didodecylamine (DDDA) and dioctadecylamine (DODA). The coated crystal was separated from the solution by a membrane consisting of Teflon PVC or a Fluoropore filter. They found that DDDA was at least 33 times more sensitive to water than to dry carbon dioxide that 100% sulphur dioxide gave a frequency change of 84 Hz and 100yo RH (25-30 "C) gave a change of 1770 Hz in 90 min and 2736 Hz in 6 h.They reported that exposure to moisture had no effect on the sensitivity of the coating to carbon dioxide. Carbon Monoxide Ho et aLa9 used a quartz piezoelectric crystal to detect carbon monoxide concentrations in the p.p.b. and p.p.m. ranges depending on the sample size. The test sample containing the carbon monoxide was reacted with mercury(I1) oxide at 210 "C to produce mercury vapour. The mercury vapour liberated was adsorbed on the gold electrodes of the crystal (see the section on mercury). A background reference stream was generated by allowing the sample to pass through a column containing granular silver oxide to oxidise the carbon monoxide quantitatively. As there was no carbon monoxide in this stream any response was due to thermal decomposition.This also allowed the removal of any interference effect arising from the presence of hydrogen, as it was included in the decomposition background. Organics that may cause interferences were reported to be removed by either activated charcoal or molecular sieve 5A. Inorganic vapours were reported not to interfere except for sulphur dioxide which was reported to be "easily removed by a molecular sieve pre-column." The effect of moisture was studied. Water vapour was thought to affect the thermal decomposition of mercury(I1) oxide rather than by direct reaction and is known to deactivate silver oxide. These workersag removed moisture by the use of a desiccant. Calcium chloride and phosphorus(V) oxide were found to be most efficient in removing water vapour without affecting the concentration of the carbon monoxide in the air stream.This system would have practical use more as a fixed-site monitor than a portable system owing to the power requirements for the heating element around the mercury(I1) oxide cell. Cyanide Nomura and c o - ~ o r k e r s ~ ~ ~ ~ have studied the use of the piezoelectric crystal to determine cyanide in solution. AT-cut 9-MHz crystals with silver-plated gold electrodes were used. Initially,'O a constant volume of the sample or standard solution was adjusted to pH 9.6 with either a dihydrogen phosphate - borate or borate - hydroxide buffer and kept at 25 "C in a water-bath. The solution was stirred at 430 rev. min-1 and the crystal whose frequency had been previously determined was immersed and left for 15 min.The crystal was remove October 1983 FOR MASS AND CHEMICAL MEASUREMENTS. A REVIEW 1179 from the solution and washed with water then acetone. The crystal was then put into the oscillator set in an air-bath kept at 30 “C. The new frequency was determined after 1 min. A linear range of 10-7-10-5~ was quoted. EDTA was found to mask interference effects completely from cations forming cyanide complexes. Further studies were carried out and a paper appeared in the Japanese literature71 in which the frequency change obtained when the solution was allowed to pass across one side of the crystal was found to depend on density (d g cm-3) and the specific conductivity (K i2-l cm-1) and the following equation was produced : Cyanide solutions over the range 10-6-5 x One drop (5 p1) of the sample solution (adjusted to pH 10.4) was placed on the electrode of a horizontal crystal in an air-bath.72 The frequency was measured just after the drop had been applied and again 4 min later.A linear range of 10-44 x 1 0 - 3 ~ was quoted. The only interferents reported were silver(1) and mercury(I1) in the presence of EDTA. M were determined. Explosives The airborne concentration of explosive materials in production plant and airport surveil-lance is of some importance. Tomita et ~ 1 . ~ ~ used Carbowax 1000-coated 9-MHz crystals to monitor the level of mononitrotoluenes (MNT) in air to which the crystal shows greater sensi-tivity at low concentrations and low sensitivity at high concentrations.A 10-ml volume of air was injected into a nitrogen carrier gas stream flowing at 30 ml min-1. The detector cell was-kept at 50 “C and higher temperatures were reported to decrease the sensitivity markedly. The frequency change due to 7.5 p.p.m. of MNT was 74 Hz at 80 “C and 186 Hz at 50 “C. The only interferences reported were some perfumes and high concentra-tions of organic solvents for chloroform due to “dissolution of the coating.” There was some response to atmospheric humidity. It was assumed however that as the humidity of “room air is relatively constant during measurement this interference can be compensated by using air of the same humidity as the carrier gas. Actually no interference from humidity and no significant change in the sensitivity were observed when room air was used as a carrier gas and also as a diluent for MNT vapour.” It was accepted that less stability in the base line due to humidity changes might become a problem in long-term continuous measurement.Hydrocarbons and Halogenated Hydrocarbons Kindlund and Lundstrom74 used two crystals in their experiments one coated and the other an uncoated reference. The two frequencies around 12 MHz were mixed and the difference frequency was converted into a voltage and monitored by a strip-chart recorder. In their initial studies they examined fatty acids alkanes lecithins and silicone oils (DC 190). DC 190 gave the best response. These workers examined the effect of the thickness of the coating on the response using 0.5% halothane (CF,CHClBr) the effect of temperature on the solubility of some common solvents over the range -30 to +50 “C and the interference effects of moisture DC 190 did not show selectivity for halogenated hydrocarbons but appeared to discriminate against water-soluble solvents.This discrimination was probably due to a difference in the number of adsorption sites and/or (for acetone) in the heat of adsorption. Edmonds and West54 examined the behaviour of variously coated 9-MHz AT-cut crystals to chloroform and toluene with respect to detector cell volume gas flow-rate analyte concentra-tion and size of detector coating. They also reported that Pluronics 64 was the most sensitive coating for ethylbenzene o-xylene and hexane Carbowax 20RI for acetone and chloroform and squalane for cyclohexane.Hydrogen Bucur75 utilised a quartz crystal microbalance to detect hydrogen (deuterium) in an inert gas and for the determination of the deuterium content in a gaseous hydrogen - deuterium mixture. He also examined the hydrogen (deuterium) interaction with thin palladium films. The sensing elements were two AT-cut crystals coated on both sides with silver electrodes and 1180 ALDER AND MCCALLUM PIEZOELECTRIC CRYSTALS Analyst Vol. 108 thin palladium film (BT-cut crystals were also studied). One crystal was operated at a reson-ance frequency of 5.25MHz and the other at 8.35 MHz. Hydrogen was determined in nitrogen up to a 0.1 molar fraction and deuterium up to a 0.25 molar fraction the frequency change reported being 150 Hz in each instance.Using the relationship that the mass of analyte adsorbed on to the palladium was proportional to the square of the partial pressure, a plot of Af2 against molar fraction was linear up to the values noted above. He quoted a detection limit of 0.01% for hydrogen and 0.03% for deuterium. Deuterium in the mixture was determined by measuring the difference between the palladium film being saturated with hydrogen and the film being exposed to the hydrogen - deuterium mixture the extent of exchange between the two isotopes being proportional to the partial pressure of deuterium. Hydrogen Chloride Hydrogen chloride in the p.p.b. and p.p.m. concentration ranges was selectively detected in ambient air76 using either triphenylamine (TPA) or trimethylammonium chloride (TMA-HCl) coated crystals.These coatings were reported to have response times of less than 30 s. Ammonia and moisture were found to cause interference problems when the TMA-HCl coating was used. The use of a gas-chromatographic column packed with silica gel was reported to have eliminated the effect of moisture. Frequency changes quoted for TPA were 154 Hz (100 p.p.m.) 52 Hz (1 p.p.m.) 40 Hz (0.1 p.p.m.) 26 Hz (0.01 p.p.m.) and 16 Hz (0.001 p.p.m.) and for TMA-HC1 changes of ca. 400 Hz (100 p.p.m.) and ca. 180 Hz (0.001 p.p.m.). Hydrogen Sulphide Silver- copper- or lead acetate-coated crystals were suggested by King77 for the determination of hydrogen sulphide. Webber et aL7* examined the acetone extracts of a number of organic soots for sensitivity to this compound.The soots were prepared by burning the substance in air over a Bunsen burner flame and collecting the residue which was then extracted in acetone and the extract evaporated from the crystal. Table I1 shows the responses obtained; a linear relationship was reported over the range 1-50 p.p.m. Chlorobenzoic acid was used also to detect hydrogen sulphide in solution.79 In this instance a membrane protected the crystal from the effects of moisture. TABLE I1 RESPONSE OF SOOT EXTRACTS TO 10 P.P.m. OF HYDROGEN SULPHIDE7* Soot extract Frequency change/Hz Benzene . . Toluene Carbon tetrachloride Chloroform : Chloroform extract Acetone extract . . Chlorobenzoic acid Benzyl chloride . . ChIoroaniline Benzoyl chloride . . Chloroacetic acid . . Xylene .. 0 . . 0 0 I . 0 . . 10 30 35 8 . . 24 . . 0 ,. 70 Mercury He noted a decrease in the afinity for mercury at surface densities in an excess of 0.15 pg cm-2J which represents a total adsorbed mass of the order of 60 ng for the crystal used. A detection limit of 5 ng 1-1 for mercury vapour in a continuously flowing air stream and 0.7 ng for the total mass of vapour in a gaseous sampIe drawn in say soil gas measurements were quoted The gold electrodes were cleaned by heating to 400 “C. The use of gold-coated crystals as part of an air monitoring system was further developed by Scheide and co-workerssl-84 over the period 1974-81. In a11 of these papers gold-coated Bristow,so in 1972 used a gold-plated quartz crystal as a “sniffer” for mercury in soil October 1983 FOR MASS AND CHEMICAL MEASUREMENTS.A REVIEW 1181 quartz crystals were used the choice of fundamental frequency being either 5 9 or 15 MHz. Table I11 lists the responses obtained from these crystals for 50 ng 1-1 of mercury. I t was also noted that the9-MHz crystal gave a frequency change of 24 Hz for 100% relative humidity (RH) 2 Hz for 82% RH and 0 Hz for 67% RH at 25 “C. These figures appear to be low if one compares them with the values reported by Fogleman and Schuman.85 A portable system was developed and reported in 1978.84 It was designed to be used as a site monitor or as a personal dosimeter with an 8-h operation period. The accumulated dosage could be displayed by pressing the ON-display button and the average exposure obtained by a simple calculation.The gold surface could be regenerated by placing the crystal in an oven at 150 “C with hot clean air passing over the crystal. TABLE I11 RESPONSE OF GOLD-PLATED CRYSTALS TO 50 ng 1-1 OF MERCURY~~ Fundamental frequency/MHz Change in frequency/Hz 5 20 9 30-115 15 300-330 Methane and Other Hydrocarbons The level of methane and other combustible gases is of paramount importance in mine safety. Current methods of measuring methane involve the use of a flame-ionisation detector or by the use of a selective catalytic combuster (pellistor). King6 used a coated piezoelectric crystal to detect moisture arising from the selective com-bustion of methane and other hydrocarbons. The crystal was set in line after a dryer and combuster (heated filament) and the hydrogen was measured after catalytic oxidation over platinum at 150 “C.Some differentiation was obtained between ethane and methane and the alkenes ethylene and propylene by control of the filament temperature. In summing up, King pointed out the features of hydrocarbon determination as 5-min response time 10 p.p.b. detection limit reactive hydrocarbons determined directly other hydrocarbons and methane by difference (after total combustion). As the detector responds to the hydrogen combustion product (water) calibration can be easily achieved by internal electrolysis of water. Organophosphorus Compounds Guilbault and co-workers have published a number of papers dealing with the detection of these corn pound^.^^-^^ A series of inorganic salts were examined as potential coatings87 with mercury( 11) bromide examined in detail.Diisopropylmethyl phosphonate (DIMP) was used as the model compound. The crystals were coated by dipping into a 0.01 M solution of the coating material. Table IV summarises their results obtained on a vacuum line. Fair linearity is reported to exist with a vapour pressure detection limit of 10-4mmHg of DIMP. With respect to the response of the crystal to moisture, Guilbault noted “A total frequency of 400 cycles resulted from the application of 0.7 mm of the phosphonate molecules. Large amounts of air and oxygen (25 mm) caused little change in the frequency of the crystal (20 Hz). Large amounts of water vapour (25 mm) caused a 100 cycle change in the frequency of oscillation.’’ Upon application of a (high) vacuum water oxygen or air were quickly removed from the AT-cut 9- or 14-MHz crystals were used.TABLE IV RESULTS OBTAINED BY GUILBAULT FOR MAGNESIUM BROMIDE-COATED CRYSTALS87 Coating None . . . . None . . Magnesium bromide Magnesium bromide Magnesium bromide Magnesium bromide Magnesium bromide Crystal frequency/MHz * . 14 14 14 9 14 . . 14 14 Exposed to pressure of AF/Hz 3 mmHg water 25 0.7 mmHg DIMP 400 0.7 mmHg DIMP 250 2 mmHg DIMP 75 25 mmHg air 20 25 mmHg oxygen 20 25 mmHg water 10 1182 ALDER AND MCCALLUM PIEZOELECTRIC CRYSTALS Analyst VoZ. 108 crystal indicating physisorption. DIMP was removed with difficulty and the initial fre-quency was never re-attained indicating chemisorption of the DIMP.The slopes for the various coating salts examined were iron(II1) chloride 4 copper( 11) chloride 2 nickel chloride 1.1 and cadmium chloride 1 Hz p.p.m.-l; all were linear to at least 400 p.p.m. of DIMP. “For the most sensitive substrate Fe(DIMP),Cl, a sensitivity below 10 p.p.m. was attained.’’ For paraoxon (0,O-diethyl-0-9-nitrophenyl phosphate) an iron( 111) chloride - paraoxon-coated crystal reportedly gave a frequency change of 44 Hz for 100 p.p.b. (should this have been 100 p.p.m.?) (Table V) or 10 Hz for 35 p.p.m. The straight-line response was about 0.3 Hz p.p.m.-l up to 65 p.p.m. with a detection limit of 10 p.p.m. of paraoxon. These workers showed that better sensitivity was obtained if both sides of the crystal were coated.88 TABLE V RESPONSE TO PARAOXON DETECTOR FOR VARIOUS INTERFERENCES AND PARAOXONs8 Interference Dry air .. * . Laboratory air . . NO2 CO,. . Water . . so,. . Paraoxon . . A F due t o 100 p.p.b./Hz 0 1 3 2 0 8 44 Two oximes 2-pyridylaldoxime methiodide (2-PAM) and isonitrobenzoylacetone (IBA) , were found to be active towards organophosphorus compoundssg and were tried as coating materials. It was found that 2-PAM was too volatile for use in a flowing stream detector. IBA was found to be stable selective and sensitive but the reaction on the substrate surface was found to be irreversible and the sensitivity decreased with each injection of DIMP or DDVP (dimethyldichlorovinyl phosphonate) . To remedy this the cobalt complex of IBA was prepared and applied to the crystal.The resultant detector was reported to be stable, selective sensitive and completely reversible. This coating was further improved by the addition of a small amount of paraoxon to the solution containing the coating. The responses to parathion DDVP and DIMP were found to be 18 Hz (5 p.p.b.) 9 Hz (50 p.p.b.) and 14 Hz (20 p.p.b.) respectively. Interferences were only significant at levels greater than or equal to 100 p.p.m. Copper complexes bound to an experimental resin polymer (XAD-4) and a mixture of 1-dodecyl-3-hydroxyiminomethylpyridinium iodide (3-PAD) with Triton X-100 and sodium hydroxide were examined.g0 Again DIMP was used as the model compound. Table VI shows adsorption values for these coatings compared with SE-30. TABLE VI ADSORPTION OF DIMP ON TEST MATERIALS” Substance [AF(Hz)/p.p.b.DIMP]/pg coating XAD-4 - Cu2+ diamine 2.6 3-PAD 1.9 SE-30 0.4 A second paper dealing with the response of 3-PAD Triton X-100 - sodium hydroxide and other coated crystals to a number of organophosphorus compounds was published in 1981.91 AT-cut 9-MHz crystals in HC-25/U mounts with silver-plated electrodes were used. Tables VII and VIII show the responses of these coatings to the test compounds and interferences. The use of sodium hydroxide that “had a high deliquenscence which was desirable for the coating” (to permit hydrolysis of the organophosphorus compound) and which also adsorbs carbon dioxide from the atmosphere must come as a surprise to anyone working with the mass-sensitive piezoelectric crystal detectors for monitoring trace levels of organic compounds.Nonetheless the authors reported an improved sensitivity with the ternary mixture (56 October 1983 FOR MASS AND CHEMICAL MEASUREMENTS. A REVIEW 1183 TABLE VII RESPONSE OF COATINGS TO ORGANOPHOSPHORUS COMPOUNDS” AFIHz ‘ DIMP Malathion Parathiok Coating (15p.p.m.) (1 p.p.m.) (1.5p.p.m.) L-Histidine hydrochloride 30 1414 126 DL-Histidine hydrochloride . . - 474 23 Succinylcholine chloride . . 55 519 76 2-PAD 290 3-PAD . . * . 403 64 24 Succinylcholine iodide . . 41 490 44 - -Triton X-100 - 13% sodium hydroxide - 31% 3-PAD) over previously tried mixtures. They construed that the detector response must “be due to the physical and chemical adsorptions, dissolutions and associations catalysed by moisture and surfactant.” Temperature studies indicated that lower temperatures are optimal for sensitivity and higher for response time.The sensitivity was half the 22 “C value at 45 “C and the recovery time improved from 4 min at 20 “C to 1 min at 55 “C after exposure to 2 ml of 15 p.p.m. DIMP in a 30 ml min-l “room air” stream. In testing for potential interferences the ternary mixture coating was exposed to some “other common atmospheric constituents and pollutants,” voiz. ammonia carbon monoxide benzene toluene ethanol chloroform and sulphur dioxide (Table VIII) . Sulphur dioxide reacted with the sodium hydroxide and as a consequence interference for sulphur dioxide could be eliminated by using a binary coating mixture of 3-PAD and Triton X-100 when monitoring air samples containing high concentrations of SO,.Surprisingly two very common atmospheric constituents carbon dioxide and water vapour were not tested for their response which seems a regrettable omission. TABLE VIII RESPONSE OF 3-PAD MIXTURE TO INTERFERENCES” The response to 15 p.p.m. of DIMP was 610 Hz. Interference Concentration p.p.m. Ammonia Ammonia Carbon monoxide Benzene Toluene . . Ethanol . . Chloroform . . Sulphur dioxide Sulphur dioxide 1000 100 1000 . . 100 100 * . 100 100 1000 100 AFIHz 15 4 2 11 25 9 71 70 42 Sulphur Dioxide Guilbault and L o p e z - R ~ m a n ~ ~ ~ ~ reported the responses they obtained from a number of coatings. Their first paper compared sodium tetrachloromercuriate with Apiezon N SE-30, QF-1 Carbowax 20M and Versamid 900.They also examined the effect of supply voltage and ambient temperature on 9- and 14-MHz coated crystals. Table IX shows the amount of sulphur dioxide adsorbed with time. These workers chose the Carbowax 2OM-coated 9-MHz crystals as the detector for a general-purpose gas ~hromatograph.9~ The detector was found to be linear over the range 1-100 p.p.m. of sulphur dioxide had a response time of 5 s and a recovery time of about 1 min. Table X lists the responses obtained in their interference study. Frechette and C O - W O ~ ~ ~ ~ S ~ ~ ~ ~ ~ examined 14 coatings; Table XI shows the relative responses obtained. Nitrogen dioxide was found to be a common interferent. Two other coatings, tridodecylamine and tripropylamine were found to be the most sensitive to sulphur dioxide but were rejected because they had a high bleed rate and showed some irreversibility.These workers chose SDM polymer (styrene - dimethylaminopropylmaleimide 1 1 copolymer) fo 1184 ALDER AND MCCALLUM PIEZOELECTRIC CRYSTALS Analyst ‘Vd. 108 TABLE IX AMOUNT OF SULPHUR DIOXIDE ADSORBED WITH TIMEg2 SO adsorbed using the coating/pg Blank Timelmin (no coating) 0.10 -0.20 -0.30 -1 2 3 0.004 4 0.021 5 0.048 10 0.075 15 0.081 --HgCl,2-0.181 0.390 0.612 0.805 1.101 1.433 2.051 2.322 2.980 4.01 1 Apiezon N 0.126 0.148 0.172 0.182 0.188 0.210 0.235 0.281 0.298 0.340 Silicone 0.082 0.089 0.097 0.097 0.103 0.111 0.114 0.141 0.149 0.194 SE-30 Silicone 0.070 0.088 0.122 0.149 0.168 0.265 0.284 0.330 0.350 0.494 QF-1 Carbowax Versamid’ 20M 900 0.030 0.063 0.041 0.125 0.082 0.142 0.102 0.195 0.113 0.208 0.118 0.231 0.151 0.252 0.162 0.301 0.184 0.352 0.201 0.381 further study.Sample cell volumes of 2 and 0.01 cm3 were tested and it was found that the cell having the lower cell volume gave the highest sensitivity. They produced a probe with which they evaluated the detector response against substrate mass sample size and concentra-tion. It was suggested that there was the possibility of the coating becoming deactivated in the presence of large amounts of moisture. The sulphur dioxide was reported to dissolve to form sulphurous acid which would attack the coating and give a salt.An infrared study was carried out and this produced evidence supporting the hypothesis. TABLE X RESPONSES OBTAINED BY GUILBAULT AND LOPEZ-ROMANs3 FOR THEIR CARBOWAX 20M COATED CRYSTALS Gas injected Concentration of gas so Water vapour . . Air . . . . co . . k3 :: N,O . . NO2 NH 30 p.p.m. Saturated 99% 99 % 99% 99 % 99 % 99 % 99 % Response, arbitrary units 46 3 2 2 3 8 2 4 2 Guilbault and co-workersg6-S8 studied a number of amines as potential coatings including P-toluidine Amine 220 triethanolamine (TEA) Armeen 2s and quadrol [NNN’N’-tetrabis(2-hydroxypropyl)ethylenediamine] with greatest emphasis on quadrol. They reported favour-able responses with all of these compoundsg6 but found some problems P-toluidine was found to be too volatile for prolonged use and Amine 220 and Armeen 2s were found to give a curved calibration graph depending on sample volume and concentration.TEA and quadrol were reported to give linear responses over the range 10 p.p.b. to 30 p.p.m. It was reported that TABLE XI RELATIVE RESPONSE TO SULPHUR DIOXIDEs4 Compound Tridodecylamine . . Melamine . . Diallyl melamine . . Igepal CO-880 . . Cellulose nitrate . . Phenyldiethanolamine Diallylamine . . Relative response 160 4 . . 24 . . 69 . . 35 . . . . 70 150 Compound UC-W98 . . . . . . Versamid 900 PP-2040 . . . . * . . . PE-100 . . 2,2-(m-tolylin1ino)diethanol .. Tripropylamine . . . . SDM polymer . . . . Relative response 45 50 190 68 225 3 19 October 1983 FOR MASS AND CHEMICAL MEASUREMENTS. A REVIEW 1185 “moisture physically condenses on the crystal surface thereby giving a strong response, Nitrogen dioxide and moisture are the major interferences. For upper atmospheric clean air studies these interferences will not have any effect on the detection of sulphur dioxide. The use of a GC column or drier to separate sulphur dioxide nitrogen dioxide or moisture followed by separate detection will solve this problem.’’ A hydrophobic membrane filter was later added to reduce the level of moisture reaching a quadrol coated detector crystal.97 The response due to “laboratory air” was reduced from 300 to 70 Hz using a 0.45-pm pore size Acropor filter membrane.Further development of their quadrol-coated crystals led to a portable system designed to monitor car exhaust fumes and refinery stack gases.98 Four hydrophobic membrane layers were used to reduce the moisture reaching the detector. Cheney and c o - w o r k e r ~ ~ ~ J ~ examined TEA as a coating suitable for sulphur dioxide. They reported the 90,95 and 99% response times which at 25 p.p.m. were between 9.25 and 11 min and at 761 p.p.m. between 1 and 5.25 min. A wide range of sensitivities and temperature variations of sensitivity for the coating towards sulphur dioxide were reported. The authors commented that the slow response below 25 p.p.m. rendered the study of response below this level impractical. The TEA coating was modified with triisopropanolamine (TIP) on Teflon to reduce the volatility that was achieved but at the expense of sensitivity.Ethylenedinitrilotetraethanollol appeared to offer better sulphur dioxide sensitivity at higher temperatures and flow-rates. Toluene Diisocyanate Alder and Isaac102J*3 reported some studies on the use of piezoelectric crystals as part of a portable instrument for personal monitoring with emphasis on shipboard use. They outlined the requirements desired for such a use and examined a number of coatings with emphasis on polyethylene glycol 400 as a model compound for development studies. This coating was found to be unsuitable for toluene diisocyanate (TDI) because of water sensitivity and irrevers-ible TDI adsorption. Using statistical analysis of the data an extrapolated limit of detection of the order of 6 p.p.b.of toluene diisocyanate in dry air was predicted. These authors suggested that the irreversibility of the coating may be exploited to give an indication of the time-weighted average (TWA) exposure of the operator to the hazardous material. Water Gjessing et al.lo4 used two gold-plated 8-MHz quartz crystals one coated with a hygroscopic material with the other left sealed and uncoated. The materials tested included sodium fluoride barium fluoride silicon oxides magnesium fluoride and some types of glass. A hygroscopic film of SiO showed promise for water with a sensitivity of 6.5 Hz %(RH)-l a t 20 “C. A patent covering the use of coated piezoelectric crystals as a moisture analyser was filed by Kingf05 in 1969.In this patent the halides of lithium and calcium are specifically mentioned and emphasised in the claims athough a number of other salts are listed. A number of coatings were examined by Lee et a1.lo6 and gelatine was found to be the most suitable. They stated that the sensitivity of this coating was 3.8 Hz p.p.m.-l (V/V) of water with 150pg of coating. This coated crystal was used to determine the concentration of moisture in a number of different commercial gases. The results obtained were compared with two recognised techniques gravimetry and the use of the Du Pont 303 moisture analyser. The results compared favourably. Bacterial and Fungal Growth A preliminary study into the use of quartz crystals to measure the growth rate of bacteria and fungi was carried out by Downes.lO7 He was unsuccessful in measuring any mass change caused by growth of these living systems.This was ascribed to the growth rate for the cells studied which was very slow and the possibility that the cell membranes could be ruptured by the high oscillation frequency. Solution Property Measurement Nomura and co-workers examined the change in frequency obtained when a quartz crystal was immersed in organic solution^.^^*^^^^ They found that the change in oscillation frequenc 1186 ALDER AND MCCALLUM PIEZOELECTRIC CRYSTALS Analyst VoZ. 108 depended on the specific gravity and viscosity of the solvents. The frequency was shown to decrease with increasing electrolyte concentration added depending on the increase in specific conductivity.They reported that AF = ad* + b$ - c where a b and c are constants determined by the crystal d is the density and r] is the viscosity of the solvent .lo9 Trace Metal Studies In another paper Nomura et aZ. reported the extraction of the 8-quinolinate chelate of lead into chloroform.l1° This solution was then transferred into a 10-ml beaker and a platinum-plated crystal immersed in the solution. The frequency was noted after 2 min and the pro-cedure repeated with a reagent blank after thorough washing and drying. Lead concentrations over the range 3 x lo-% x 1 0 A 5 ~ in the aqueous solution were examined. Interferences from iron(III) nickel cobalt(II) zinc cadmium and silver could be masked using L-ascorbic acid and cyanide. Nomura and Maruyamalll published a curious paper reporting the examination of the stability of aqueous solutions of a number of metal ions and the selective determination of iron(II1) as its phosphate.They found that with a standard crystal the frequency change was directly proportional to the specific conductivity for up to 2 mM where deviations due to viscosity and density were reported to occur. At concentrations above 20 mM abrupt fre-quency changes were found to occur as a result of the solution being able to short-circuit the quartz. Under these conditions they reported the electrodeposition of metal ions such as Mn2+ Ni2+ Co2+ Zn2+ Cd2+ Ag+ Cu2+ and Pb2+. With aluminium and iron adsorption of some salt was thought to cause the frequency change. To investigate this in more detail and to determine iron(III) they modified a standard crystal by covering one surface with a quartz plate separated from the crystal surface by “four slim hairs laid on each edge of the quartz plate.” The lead wires of the crystal holder were coated with epoxy resin to avoid electrolysis problems.A calibration graph of AF versm iron(II1) was produced and shown to be linear over the range 1 x 10-5-1 x 1 0 - 4 ~ . Some interference studies were carried out on a 5 x 1 0 - 5 ~ iron(II1) solution. Ten-fold molar amounts were thought to interfere if changes of greater than 15% were obtained. They reported that this occurred with lead aluminium bismuth, sulphide and thiosulphite. These were fused with epoxy resin. Thermal Analysis King et aZ.112 utilised piezoelectric crystals as supports for thin film thermocouples for differ-ential thermal analysis over the temperature range - 125 to 500 “C.Their system was reported to give calorimetric data for indium that agreed well with the established literature. A quartz cryst a1 t hermogravimet ric analyser for t emperature-programmed analysis of deposited films was described by Henderson et aZ.l13 They showed that their technique was applicable at temperatures up to 570 “C but suggested that temperatures in excess of 1000 “C should be possible using lithium niobate piezoelectric crystals. Very fast heating rates are possible with this device as thermal equilibrium is reportedly easier to achieve during heating. It is suggested that this system could be applied to study the volatilisation of thin films and coatings the kinetics of surface phenomena and flammability studies of polymers as thin films rather than as the bulk properties required by conventional thermogravimetric instrumen-tation.Conclusions As part of a portable detector system quartz piezoelectric crystals have potential in the field of personal exposure monitoring. They have high mass sensitivity dependent only on the oscillation frequency of the crystal. The specificity of the crystal is determined by the in-genuity of the worker in choosing the coating. Interferences such as moisture may be removed by physical means (e.g. filters membranes October 1983 FOR MASS AND CHEMICAL MEASUREMENTS. A REVIEW 1187 pre-columns) by correction using either a second crystal or an alternative moisture sensitive device (e.g.a resistive probe). Alternatively moisture could be used as part of the chemical system and allowed to become involved in the surface reaction.g0 The frequency-change data obtained from the crystals on exposure to analyte vapour lend themselves very well to modern data processing techniques. The use of on-board microcomputers to output gas concentration information in real time will guarantee interest in these devices in the future development of portable monitoring equipment. It is a pity that all workers in this field do not pay more attention to reproducibility of results calibration and reproducibility both of coating the crystals for gas analysis and of response to the analyte and interferences. Accepting the wide variation of humidity in real situations the disregard paid to the response of detector crystals proposed for practical applications by many if not most workers is surprising.To show that a detector could still be sensitive to 10 p.p.m. of a gas over a range of humidity of 10-95% at 28 “C would not only impress workers in this but also in all the gas-sensing portable instrument fields. To show that the device still worked at the not uncommon ambient temperature of -5 “C would be even more impressive. Now that many of the limitations and advantages of the piezoelectric crystal detectors have been identified and the instrumentation and data manipulation side largely developed the time is ripe to choose those situations where the piezoelectric crystal detector is best able to succeed and to characterise its performance under truly practical conditions.Applications where only low selectivity are required for example in the monitoring of families of compounds, seem likely to succeed as the selectivity problem is undoubtedly the most limiting. The application to short-term (working day) rather than long-term monitoring with disposable crystals also seems attractive. The idea of using several sensor crystals with different responses to give data that would permit interference correction or a multi-component detection capability has crossed many people’s minds. The availability of the data handling packages to do this now renders it possible even on a portable system. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. References Webber L. M. Hlavay J. and Guilbault G. G. Mikrochim. Acta 1978 351. Guilbault G. G. Anal. Proc. 1982 19 68. Hlavay J. and Guilbault G. G. Anal. Chem. 1977 49 1890. Guilbault G. G. Ion. Sel. Electrode Rev. 1980 2 3. Cady W. G. “Piezoelectricity,” First Edition McGraw-Hill New York and London 1946. Becquerel A. C. Bull. Soc. Philomath. Paris 1820 7 149 (series 3). Curie J. and Curie P. Bull. Soc. Min. Paris 1880 3 90. Lippmann G. An. Chim. Phys. Ser. 5 1881 24 145. Langevin A. “Piezoelectricity,” First Edition McGraw-Hill New York and London 1946 p. 5. Cady W. G. Proc. Inst. Radio Eng. 1924 12 805. Cady W. G. J . Opt. SOG. Am. 1925 10 415.Pierce G. W. Proc. Am. Acad. Arts Sci. 1923 59 81. Pierce G. W. Proc. Am. Acad. Arts Sci. 1925 60 277. Phillips F. C. “An Introduction to Crystallography,” Third Edition Longmans London 1963, Voigt W. Abh. Ges. Wiss. Gottingen 1890 36 1. Voigt W. “Lehrbuch der Kristallphysik,” B. G. Teubner Leipzig First Edition 1910; Second Tillyer E. D. U.S. Pat. 1907 613 1933. Lack F. R. Willard G. W. and Fair I. E. Bell Syst. Tech. J. 1934 13 453. Bechmann R. J . Sci. Instrum. 1952 29 73. Cady W. G. Proc. Inst. Radio Eng. 1922 10 83. Warner A. W. Bell Syst. Tech. J. 1960 39 1193. Warner A. W. Fraser D. B. and Stockbridge C. D. IEEE Tram. Sonics Ultrason. 1965 SU-12 52. Spencer W. J. IEEE Trans. Sonics Ultrason. 1965 SU-12 1. Horton W. H. IEEE Trans. Sonics Ultrason.1966 SU-13 111. Sauerbrey G. Z. Phys. Verha. 1957 8 113. Sauerbrey G. Z. 2. Phys. 1959 155 206. Lostis M. Thesis University of Paris Faculty of Science 1958. Stockbridge C. D. Vac. Microbalance Tech. 1966 5 193. Oberg P. and Ligensjo J. Rev. Sci. Instrum. 1959 30 1053. Warner A. W. and Stockbridge C. D. Vac. Microbalance Tech. 1963 3 55. King W. H. Jr. Bull. N.Y. Acad. Med. 1972 48 459. p. 155; Fourth Edition Oliver and Boyd Edinburgh 1971 p. 167. Edition 1928 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82.83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 1188 ALDER AND MCCALLUM PIEZOELECTRIC CRYSTALS Analyst VoZ. 108 Richardson P. D. Bull. N . Y . Acad. Med. 1972 48 465. Ullevig D. M. Evans J. F. and Albrecht M. G. Anal. Chem. 1982 54 2341. Haller I. and White P. J. Phys. Chem. 1963 67 1784. Fischer W. F. and King W. H. Jr. Anal. Chem. 1967 39 1265. King W. H. Jr. and Corbett L. W. Anal. Chem. 1969 41 580. Hosseiny A. PhD Thesis University of Manchester 1980. Chuan R. L. J. Aerosol Sci. 1970 1 111. Olin J. G. Sem G. j. and Christenson D. L. Am. Ind. Hyg. Assoc. J. 1971 32 209. O h J . G. and Sem G. J. Atmos. Environ. 1971 5 653. Chabre A. “Study of the Impact of Fine Particles on a Piezoelectric Target in View of Determining their Mass and Concentration,” Aix-Marseilles University France 1972 (in French) ; Nucl.Sci. Abstr. 1973 37 ref. 22601. Carpenter T. E. and Brenchley D. L. Am. Ind. Hyg. Assoc. J. 1973 33 503. Daley P. S. and Lundgren D. A. Am. Ind. Hyg. Assoc. J. 1975 36 518. Sem G. J. and Tsurubayashi K. Am. Ind. Hyg. Assoc. J. 1975 36 791. Mieure J . P. and Jones J . L. Talanta 1969 16 149. Jones J. L. and Mieure J . P. Anal. Chem. 1969 41 484. Nomura T. and Mimatsu T. Anal. Chim. Acta 1982 143 237. King W. H. Jr. Anal. Chem. 1964 36 1735. Karasek F. W. and Gibbins K. R. J. Chromatogr. Sci. 1971 9 535. Smith H. R. Am. Lab. 1972 49 (Oct.) 49. Karasek F. W. and Tierney J. M. J. Chromatogr. 1974 89 31. Karasek F. W. Guy P. Hill H. H. Jr. and Tierney J. M.J. Chromatogr. 1976 124 179. Janghorbani M. and Freund H. Anal. Chem. 1973 45 325. Edmonds T. E. and West T. S. Anal. Chim. Acta 1980 117 147. Schulz W. W. and King W. H. Jr. J. Chromatogr. Sci. 1973 11 343. Konash P. L. and Bastiaans G. J. Anal. Chem. 1980 52 1929. King W. H. Jr. U.S. Pat. 3 164 004 1965. King W. H. Jr. Res. Dev. 1969 20(4) 28. King W. H. Jr. Res. Dev. 1969 20(5) 28. Karmarker K. H. and Guilbault G. G. Anal. Chim. Acta 1975 75 111. Webber L. M. and Guilbault G. G. Anal. Chem. 1976 48 2244, Webber L. M. and Guilbault G. G. Anal. Chim. Acta 1977 93 145. Hlavay J. and Guilbault G. G. Anal. Chem. 1978 50 1044. Edmonds T. E. Fraser S. M. and West T. S. Poster at Royal Society of Chemistry “International Conference on the Detection and Measurement of Hazardous Substances in the Atmosphere,” City University London 20-22 December 1982.Guilbault G. G. Lopez-Roman A. and Billedeau S. M. Anal. Chim. Acta 1972 58 421. Karmarker K. H. and Guilbault G. G. Environ. Lett. 1975 10 237. Ho M. H. Guilbault G. G. and Reitz B. Anal. Chem. 1980 52 1489. Fogleman W. W. and Shuman M. S. Anal. Lett. 1976 9 751. Ho M. H. Guilbault G. G. and Scheide E. P. Anal. Chem. 1982 54 1998. Nomura T. and Hattori O. Anal. Chim. Acta 1980 115 323. Nomura T. and Minemura A. Nippon Kagaku Kaishi 1980 1621. Nomura T. Anal. Chim. Acta 1981 124 81. Tomita Y. Ho M. H. and Guilbault G. G. Anal. Chem. 1979 51 1475. Kindlund A. and Lundstrom I. Sensors Actuators 198213 3 63. Bucur R. V. Rev. Roum. Phys. 1974 19 779. Hlavay J. and Guilbault G.G. Anal. Chem. 1978 50 965. King W. H. Jr. Anal. Chem. 1964 36 1735. Webber L. M. Karmarker K. H. and Guilbault G. G. Anal. Chim. Acta 1978 97 29. Webber L. M. and Guilbault G. G Anal. Chim. Acta 1977 93 145. Bristow Q. J. Geochim. Explor. 1972 1 55. Scheide E. P. and Taylor J. K. Am. Chem. SOC. Div. Environ. Chem. Prepr. 1974 14 329. Scheide E. P. and Taylor J. K. Am. Ind. Hyg. Assoc. J. 1975 36 897. Scheide E. P. and Taylor J. K. Environ. Scz. Technol. 1974 8 1097. Scheide E. P. and Warner R. B. j. Am. Ind. Hyg. Assoc. J - 1978 39 745. Fogleman W. W. and Shuman M. S. Anal. Lett. 1976 9 751. King W. H. Jr. Environ. Sci. Technol. 1970 4 1136. Guilbault G. G. Anal. Chim. Acta 1967 39 260. Scheide E. P. and Guilbault G. G. Anal. Chem. 1972 44 1764. Shackelford W.M. and Guilbault G. G. Anal. Chim. Ada 1974 73 383. Guilbault G. G. Affolter J. Tomita Y. and Kolesar E. S. Jr. Anal. Chem. 1981 53 2057. Guilbault G. G. Tomita Y. and Kolesar E. S. Jr. Sensors Actuators 1981/2 2 43. Guilbault G. G. and Lopez-Roman A. Environ. Lett. 1971 2 35. Lopez-Roman A. and Guilbault G. G. Anal. Lett. 1972 15 225. Frechette M. W. Fasching J. L. and Rosie D. M. Anal. Chem. 1973 45 1765. Frechette M. W. and Fasching J. L. Environ. Sci. Technol. 1973 7 1135. Karmarker K. H. and Guilbault G. G. Anal. Chim. Ada 1974 71 419. Karmarker K. H. Webber L. M. and Guilbault G. G. Environ. Lett. 1975 8 345. Karmarker K. H. Webber L. M. and Guilbault G. G. Anal. Chim. Acta 1976 81 265. Cheney J. L. and Homolya J. B. Anal. Lett. 1975 8 175 October 1983 FOR MASS AND CHEMICAL MEASUREMENTS. A REVIEW 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 1189 Cheney J. L. and Homolya J. B. Sci. Total Environ. 1976 5 69. Cheney J. L. Norwood T. and Homolya J. Anal. Lett. 1976 9 361. Alder J. F. and Isaac C. A. Anal. Chim. Acta 1981 129 163. Alder J. F. and Isaac C . A. Anal. Chim. Ada 1981 129 175. Gjessing D. T. Holm C. and Lanes T. Electron. Lett. 1967 3 156. King W. H. Jr. U.S. Pat. 3 427 864 1969. Lee C. W. Fung Y . S. and Fung K. W. Anal. Chim. Acta 1982 135 277. Downes J . Internal Report Imperial College London 1979. Nomura T. Okuhara M. Murata K. and Hattori O. Bunseki Kagaku 1981 30 417 (in Japanese Nomura T. and Okuhara M. Anal. Chim. Acta 1982 142 281. Nomura T. Yamashita T. and West T. S. Anal. Chim. Acta 1982 143 243. Nomura T. and Maruyama M. Anal. Chim. Acta 1983 147 365. King W. H. Jr. Camili C. T. and Findeis A. F. Anal. Chenz. 1968 40 1330. Henderson D. E. Di Tarantor M. B. Tinkin W. G. Ahlgren D. J. Gatenby D. A. and Shun, Received March lst 1983 Accepted May loth 1983 with English abstract). T. W. Anal. Chem. 1982 54 2067