IOBC wprs Bulletin Vol 22(9), 1999

The pheromone sprayer: New technology in stimulus application

Josef Gödde, Heinrich Arn1 and Arshraf El-Sayed1

Steinklepper Weg 1, D-35753 Greifenstein, Germany
1Swedish University of Agricultural Sciences, Alnarp, Sweden

Abstract: In a novel device for the quantitative application of olfactory stimuli, a solution of volatile chemicals is dispersed as an aerosol by means of ultrasound. A motor-driven syringe controls the rate at which the odorant-solvent mixture is released through a glass capillary ending in a tip of 10-50 µm inner diameter. Bending vibrations of the tip disperse the released solution into fine droplets, which evaporate completely within a few centimeters of the source. Since the odorant/solvent ratio is maintained until the liquid is dispersed, the release rate of odorant can be set and calculated from the dilution factor and the speed of the syringe plunger. With the sprayer, stimulation is independent of relative vapor pressures of the components, the matrix materials and the environmental factors. The pheromone sprayer was tested successfully both in the wind tunnel and the field.

Key Words: Pheromone, ultrasound, odorant, olfactory stimulation, wind tunnel

Introduction

Quantitative application of stimuli is a long standing problem in olfactory studies. Quantification of the odor stimulation requires that a gas, mostly air, contains a known concentration of odorant. In studies with insects, quantification of the stimulus is often made with the dose of odorant deposited on a substrate such as filter paper, rubber septa or cotton wreck. The amount of substance released per unit of time and carried with an air stream to the test animal depends on parameters such as vapor pressure, the adsorption to substrate and temperature. Depending on the substrate, dose response curves for the same odorant can differ by more than two log units (Linn et al. 1984; Sanders 1990). Sensitivity characteristics determined by such a technique can at best be expressed on relative scales, unless the stimuli are quantified by means such as radioactive labeling (Kaissling 1985). Better control over the stimulus concentration is obtained with another approach, in which the odorant is allowed to equilibrate between a reservoir of a solvent-odorant mixture and a closed gas space (Färbert et al. 1997). With this method, gas concentrations are still dependent on factors such as temperature and diluting solvent, which are difficult to control for components of different vapor pressures.

The device presented here was developed for a study of optimization of a pheromone blend of the grapevine moth Lobesia botrana Denis & Schiffermüller, in a flight tunnel. Attractiveness was assessed from differences in upwind flight behavior and flight track analysis of male moths responding to synthetic blends, female extracts and female-released pheromone (El-Sayed 1996). It was known from other studies that the behavior of male moths is dependent on both the release rate of a blend and the proportions of its components (Linn et al. 1984 Linn and Roelofs 1989).

Material and methods

General Design

Our device relies on two basic ideas: first, the odorant is diluted in a biologically inert solvent. Second, this odorant-solvent mixture (OSM) is released into the air at a known rate. A critical feature is the way the OSM is released (Fig. 1): With a motor-driven syringe, the OSM is pressed through the tip of a glass capillary with an opening of 10-50 µm inner diameter (ID) at a rate of approximately 10 µl/min (OSM flow rate). The capillary tip is excited to oscillate at about 120 kHz or a harmonic of this frequency. This produces an aerosol of the solution which disperses and evaporates. In the course of the forementioned study, four devices were built: three functional prototypes to test the principle were built within less than one day each, including the syringe pump. Most data given refer to the final version which for convenience sake was composed of commercially available components and which has been used for more than three years.

At only 25 µm amplitude peak-to-peak, an oscillation of 120 kHz creates a peak tip velocity of 9.42 m s-1 and a peak tip acceleration of 7.1 * 106 ms-2. This releases one or two droplets of OSM at each oscillation. Smooth tips that are created by a pulling and cutting device tend to release one droplet at each half-oscillation. Normally we used tips that were either pulled manually (e.g. from disposable micro pipettes with ring marks for 1, 2 5 µl, ABS, Zürich Switzerland, or from Vitrex glass capillaries, Modulohm A/S, Herlev, Denmark) with the ignition flame of a bunsen-burner or with different kinds of pullers and broken by forceps. This yielded irregular tips that tended to release one droplet at only one of the extremum positions of the oscillating tip. Microscopic observations showed that if the oscillation amplitude was increased further, such tips eventually released droplets at both extremum positions, too. Thus, a flow of e.g. 1 ml/h is dispersed into droplets of ca. 2.3 picoliters (corresponding to a droplet diameter of 16.5 µm) that are sprayed into the ambient air.

Due to their small size, the droplets evaporate completely within a distance of typically < 5 cm from the capillary tip. This can be verified optically (Fig. 1). The mist of OSM droplets around the capillary tip is not symmetrical with respect to the capillary axis and depends on the geometry of the capillary tip. Further, oscillating tips produce sound radiation pressure, which is a local constant pressure component due to non-linear interaction with the air (Sutilov 1984). Such a sound radiation pressure causes a local net air flow around the tip that may produce jets and threads of mist protruding some centimeters perpendicularly or obliquely to the capillary axis (Fig. 1). Outside this radius, both odorant and solvent occur in the gas phase.

Figure 1. Diagram of the pheromone sprayer (left) and active sprayer (right).

The release rate of odorant can be controlled by two independent parameters, a) the composition of the OSM (coarse setting) and b) the flow rate generated by the syringe pump (fine setting). Since both parameters can be determined quantitatively, the amount of odorant released per unit of time can be readily calculated. If the flow of the gas (in most practical cases air) is also known, the ratio of odorant released versus gas volume directly yields the achieved mean concentration of odorant in the gas. But the gas volume, that is relevant for this consideration, is the volume of the plume in which the odorant is more or less uniformly dispersed by diffusion and eddies.

The method of exciting the tip of a glass capillary tip to vibrate at ultrasonic frequencies has been described earlier (Gödde 1986): A piezo disk of 10-25 mm OD and ca. 1-2 mm thickness with solderable silver coatings on both faces is driven at its bending mode resonance by a sine or square wave of about 20 V p/p (Fig. 1). Electrically, the piezo disks behave like leaky capacitors of about 3 nF. Driving signals were taken from standard function generators or from self wired circuits consisting of a frequency-tunable oscillator chip (e.g. XR 2206 EXAR, NE 555 Signetics) and operator amplifiers (1/2 CA083E, CA3140 RCA) as boosting stage. The latter design is advantageous in battery-based field work due to its lower power consumption (25 mA at 18 V supply voltage with XR 2206 and CA3140).

A U-shaped wire of 1 mm diameter is soldered to one face of a piezo disk in such a way that the open stems reach about 10 mm over the rim of the disk. This wire serves two purposes: It holds the piezo disk in place by spring force and transfers the oscillations to the capillary tubing. Since the tip of the capillary is an acoustical transformer, the amplitude of oscillation increases towards the end, where it may reach 100 µm. The geometry of the capillary tip may vary considerably between individual capillaries without impairing their function.

The OSM flow rate is determined by the speed of the syringe piston which built up a pressure proportional to the flow rate (Hagen-Poiseuille), the OSM head pressure. The proportional factor depends on the OSM viscosity and the geometry of the capillary tip, which may vary considerably between individual capillaries without impairing their function. With solvent of high viscosity such as water our capillaries give maximum head pressure in the range of 1 to 5 bars as determined by the shrinking of gas bubbles. Maximum head pressure must be taken into account when choosing the tubing. Apart from this the design of sprayers for constant OSM flow rate is straightforward: the size of the syringe is selected according to the desired flow rate and operation time.

The design of sprayers for time-varying flow rates requires further considerations. Ideally, the OSM flow rate should follow instantaneously the electrical signal controlling the pump speed. In reality, a change in the control signal first must overcome inertia and slip of motor and gear, as well as elastic forces in the syringe pump, until the feed of the piston is proportional to the control signal again. The new piston speed then must overcome the volume elasticity of syringe and tubing until the OSM head pressure and flow rate have stabilized. Any gas bubbles in the OSM are highly detrimental since they contribute to volume elasticity. These delays may be summarized as a time constant for OSM flow rate changes. This time constant may be less than 25 ms in case of good design (rigid frame, direct-drive syringe pump, glass syringe with high piston speed, short, non-elastic narrow tubing between syringe and capillary), but well above 5 s with an inexpensive construction (toy motor and gear, disposable syringe).

Pulsed stimuli

Many experimental paradigms require pulsed stimuli with stop and go time constants in the range of a few milliseconds. Instead of optimizing the time constant of OSM flow rate changes, such pulses are easier to achieve by sucking off the complete odor-laden air stream during stimulus-free intervals (e.g. by a valve-controlled wide tube with under-pressure downwind from the capillary tip). In these cases, the syringe pump should be deactivated during longer stimulus-free intervals.

We have used sprayers routinely for several years in a pheromone-attractiveness study with the grapevine moth, L. botrana (El-Sayed 1996). The details of the sprayer used in this study are as follows: Glass capillaries of 1.4 mm OD and 0.62 mm ID (VERTEX cat. Nr. 9505329 Modulohm A/S, Herlev, Denmark) were drawn out and broken to tip diameters of about 55 µm OD / 40 µm ID (determined by scanning microscopy). The piezo disk was driven by a function generator at 100 to 140 kHz, the frequency being adjusted visually to the resonance of the individual capillary. OSMs consisted of pheromone mixtures in various organic solvents. A given OSM was typically used for about 100 minutes and expelled at 10 µl/min by a micro injection pump (CMA/100, Carnegie Medicine AB, Stockholm). A selector valve (CMA/111, Carnegie Medicine AB, Stockholm) was used to switch between different OSMs. The time to expel the residual volume of the foregoing OSM downstream of the selector was 3 min, as determined by gas chromatography.

Testing the pheromone sprayer in the wind tunnel

Two to three-day old L. botrana males were temporally immobilized at 4°C and groups of 25 males were transferred to glass release tubes (15 cm x 2.5 cm). Bioassays commenced at the beginning of the scotophase and continued for 4 hours. Males were released individually in the wind tunnel and the following behavioral responses were recorded: activation = wing fanning, random walking in the release tube; upwind flight = beginning of upwind flight for at least 10-15 cm; landing = touchdown at the source. The significance of differences in the number of males completing successive stages in their advance towards the source were compared using Chi-square analysis 2 x 2 with Yates correction (InStat 1993).

Using the rubber caps as substrate, Arn et al. (1988) found that the addition of 2 µg of E7,Z9-12OH and 0.5 µg of Z9-12Ac to 10 µg of the main pheromone compound, E7,Z9-12Ac, elicited a rate of male response close to the rate of response elicited by calling females. According to this finding, males were tested against binary or ternary blends containing constant release of E7,Z9-12Ac (10 pg/min) plus 2 pg/min of E7,Z9-12OH and/or 0.5 pg/min of Z9-12Ac. The possible effects of ultrasound emission and of the OSM-solvent ethanol on L. botrana, were tested by initial control experiments. Solvent and ultrasound generated during evaporation was shown to affect neither the female calling nor male flight behavior (El-Sayed 1996).

The extension of the plume at the windspeed of 30 cm/s was determined empirically: the tip of a pipette filled with a solution of TiCl4 was placed at the position of the sprayer capillary tip. In air, TiCl4 is immediately converted to TiO2 that yields a white smoke that was filmed and photographed. Assuming that widening of the plume is more eddy- than diffusion-controlled, the records of the TiO2 -smoke yield the time-averaged cross sections of the plume at any distance downstream from the pheromone release site. We determined the plume diameter at 1) the point of male release, the 2) 150 cm, 3) 90 cm, and 4) 40 cm from the source.

Results

Males L. botrana, were flown to blends of E-7,Z-9-dodecadienyl acetate, Z-9-dodecenyl acetate and E-7,Z-9-dodecadien-1-ol at various proportions. When the three pheromone components were released at the same ratio as used to loaded rubber cap (Arn at al., 1988) the percentage of males landing at source ca. 45% was lower than the ca. 80% landing at the rubber cap (Fig. 2A and B). We assume that this difference was partly due to the difference in stimulus application. Maximum rate of landing at the source was obtained when E7,Z9-12OH and Z9-12Ac was released at 0.5 pg/min and 0.1 pg/min respectively. A ternary blend of 10 pg/min E7,Z9-12Ac, 0.5 pg/min E7,Z9-12OH and 0.1 pg/min Z9-12Ac significantly enhanced the response of L. botrana males (El-Sayed et al. 1999a). The attractiveness ca. 70 % landing of the newly optimized ternary blend (100: 5: 1) was close to the ternary blend (100: 20: 5) reported by Arn et al. (1988). With rubber caps, as used by Arn et al. (1988), the proportion of the mixture did not reflect the true release rates since Z9-12Ac would evaporate at a relatively faster rate than E7,Z9-12OH. In addition, the dienic acetate E7,Z9-12Ac can isomerized easily on the rubber cap. It is known that the isomerized E7,Z9-12Ac is more attractive than the pure E7,Z9-12Ac (Ideses et al. 1982).

At 30 cm/s windspeed, the time averaged plume diameters at 1) the point of male release, 2) 150 cm, 3) 90 cm, and 4) 40 cm from the source were found to be 24, 20, 16 and 13 cm, corresponding to plume cross sections of 450, 315, 200 and 135 cm2, and to plume volume flows of 13.6, 9.4, 6.0, and 4.0 liters per second, respectively. In some experiments clear reactions were seen at a release rate of 0.1 pg/min of a 200 dalton pheromone component. At the mentioned plume cross sections, this corresponds to time averaged concentrations of 6.1*10-19, 8.8*10-19, 1.4*10-18 and 2.1*10-18 moles of pheromone per liter of air, or, in other terms to mean pheromone molecule distances of roughly 1.4, 1.2, 1.1 and 0.93 mm.

In addition to L. botrana, the sprayer was tested successfully with the green budworm, Hedya nubiferana Haworth, the codling moth, Cydia pomonella L. A field set-up of the pheromone sprayer was used for trapping several species of bark beetles.

Figure 2. Response of L. botrana males to different combinations of the three synthetic pheromones compounds (E7,Z9-12Ac, E7,Z9-12OH and Z9-12Ac). (A) Loaded on rubber caps, data from Arn et al. (1988). (B) Released using the pheromone sprayer.

Discussion

Ultrasound did not affect the behavior of male L. botrana or other species tested in our experiments. It is known that many insect species are sensitive to the range of ultrasound used in our set-up (Fenton and Fullard 1981). However, most of these species are susceptible to intermittent ultrasound signal (e.g. Krasnoff and Yager 1988). In contrast, we use continuous ultrasound signal to vibrate the tip. This might explain why ultrasound emitted during pheromone evaporation did not affect the animal under test.

Little is known about the influence of the solvent on insect behavior. With ethanol or hexane at extremely low release rate of 10 µl/min, no effect on behavior or males or females was apparent in all Lepidopterous species tested. However, we observed that several Dipterous and Cleopterous species did affected when ethanol was used as a solvent. When ethanol was replaced by hexane, these species oriented normally to the odor source. One precaution should be taken during the operation of evaporation: whenever the oscillations of the capillary come to a stop, drops of OSM may ooze from the tip and create an additional, non/controllable source of odor.

It could be argued that an aerosol might persist n the form of invisibly small droplets moving downwind, especially if the vapor pressure of the odorant is significantly lower than of the solvent. However, such of droplets would be unstable for the following reason. The vapor pressure of droplets s proportional to ek/r (r = radius of the droplet, k = constant). This means that the more the OSM is dispersed and the solvent evaporates, the higher the vapor pressure and evaporation rate of the solute. Yet, one must consider the extreme case in which the solute has only a negligible vapor pressure (e.g. NaCl in water). In this case, the evaporation of the solvent increases the molar fraction of the solute. This reduces the vapor pressure of the solvent until a constant ratio solvent/solute is reached. In extreme cases this may arrest any further evaporation and really lead to an aerosol. But even then, the aggregates are extremely small (and hence invisible): At 120 kHz, a sprayer produces 1 or 2 times 432*106 droplets per hour. Hence each droplet contains only one molecule on the average for every 1.4*10-15 or 0.7*10-15 mole of solute released per hour. That means, that in our near threshold experiments, individual droplets contained in the order of 100 pheromone molecules.

We have further been confronted with the argument, that enthalpic cooling by evaporation of the solvent might alter the behavior of the animals under test. This effect can be neglected. Even with solvents of high evaporation enthalpy, such as water, the plume is cooled by <0.1°C under practical conditions.

In most of the bioassays, when natural extracts were used as lure, a relative unit was used to express the amount of natural extracts load on the dispensing substrates, e.g. Female Equivalent (FE) in the case of sex pheromone. However, there are considerable variations among individual of the same species (Bäckman et al. 1997). This makes the direct comparison between the natural extracts and synthetic compounds infeasible. With the sprayer, natural extracts can be released quantitatively e.g. pg of compound x/min, etc. This is achieved first by quantification of the chemicals in the natural extracts using a gas chromatography (El-Sayed et al. 1999b).

References

Arn, H., Rauscher, S., Guerin, P. & Buser, H.R. 1988: Sex pheromone blends of three tortricid pests in European vineyards. Agriculture, Ecosystems and Environment. 21: 111-117.

Bäckman, A.-C., Bengtsson, M. & Witzgall, P. 1997: Pheromone release by individual females of codling moth, Cydia pomonella L. (Lepidoptera: Tortricidae). J. Chem. Ecol. 23: 807-815.

El-Sayed, A. 1996: Analysis of behavior and chemical stimuli in sex attraction of Lobesia botrana (Lepidoptera: Tortricidae). PhD thesis, Zürich University.

El-Sayed, A., Gödde, J., Witzgall, P. & Arn, H. 1999a: Characterization of the pheromone blend for the grapevine moth, Lobesia botrana by using flight track recording. J. Chem. Ecol. 25: 389-400.

El-Sayed, A., Bengtsson, M., Rauscher, S., Löfqvist, J. & Witzgall, P. 1999b: Evidence for a multicomponent sex pheromone in Codling moth, Cydia pomonella (Lepidoptera: Tortricidae). Environ. Entomol. in press.

Färbert, P., Koch, U.T., Färbert, A., Staten, R.T. & Cardé, R.T. 1996: Pheromone concentration measurement with EAG in cotton fields treated for mating disruption of Pectinophora gossypiella (Lepidoptera: Gelechiidae). Environ. Entomol. 26: 1105-1116.

Fenton, M.B. & Fullard, J.H. 1981: Moth hearing and the feed strategies of bats. American Scientist 69: 266-275.

Gödde, J. 1986: A fast voltage-controlled continuous-flow miniature chemostat. J. Neurosci. Methods 16: 191-200.

Ideses, R., Klug, J.T., Shani, A., Gothilf, S. & Gurevitz, E. 1982: Sex pheromone of the European grapevine moth, Lobesia botrana Schiff. (Lepidoptera: Tortricidae): Synthesis and effect of isomeric purity on biological activity. J. Chem. Ecol. 8: 195-200.

InStat, 1993: Instat Biostatistics version 2.0. GraphPad software.

Kaissling, K.E. 1985: Odorants and their application. In: Experimental Cell Biology of Taste and Olfaction, Current Techniques and Protocols. CRC Press, Boca Raton, eds. Spielman and Brand: 366-368.

Krasnoff, S.B. & Yager, D.D. 1988: Acoustic response to a pheromonal cue in the arctiid moth Pyrrharctia isabella. Physiol. Entomol. 13: 433-440.

Linn, C.E., Bjostad, L.W., Du, J.W. & Roelofs, W.L. 1984: Redundancy in a chemical signal: Behavioral responses of male Trichoplusia ni to a six-component sex pheromone blend. J. Chem. Ecol. 10:1635-1658.

Linn, C.E. & Roelofs, W.L. 1989: Response specificity of male moths to multicomponent pheromones. Chem. Senses 14: 421-437.

Sanders, C.J. 1990: Responses of male spruce budworm moths to sex pheromone released from filter paper and rubber septa. Can. Ent. 122: 263-269.

Sutilov, V.A. 1984: Physik des Ultraschalls. In Springer Wien, New York, eds. 108-125.

table of contents