Pestic. Sci. 1998, 53, 291È299

Using a Videographic System to Assess SprayDroplet Impaction and Reflection from Leaf andArtiücial Surfaces”Donald L. Reichard,1 Jane A. Cooper,2 Martin J. Bukovac3 & Robert D. Fox4*1 Deceased ; formerly, USDA-Agricultural Research Service Wooster, OH, USA2 Laboratory for Pest Control Application Technology/The Ohio State University, Wooster, OH 44691,USA3 Horticulture Department, Michigan State University, East Lansing, MI 48824, USA4 USDA-Agricultural Research Service, Wooster, OH 44691, USA

(Received 2 October 1997 ; revised version received 14 January 1998 ; accepted 24 March 1998)

Abstract : A video motion analysis system was used with two di†erent mono-disperse droplet generators to quantify droplet impaction and any consequentreÑection. By using di†erent magniÐcation/droplet generator combinations,droplet impaction was detailed at various stages. Low (7]) magniÐcation,together with a generator that produced a spray cloud, allowed determination ofthe height and numbers of droplets reÑected from plant surfaces. Higher (15])magniÐcation and a single-drop generator enabled the trajectory, and changes invelocity, of a rebounding droplet to be followed. By using high (90]) magniÐ-cation and the single-drop generator, detailed measurements of a droplet deform-ing on impact could be made. Examples of how these techniques could be usedare given. 1998 SCI(

Pestic. Sci., 53, 291È299 (1998)

Key words : reÑection ; impaction ; droplet generators ; leaf surfaces

1 INTRODUCTION

Application of foliar sprays is a complex process thatincludes such events as spray atomization, transport tothe plant, impaction on plant surfaces, retention, andfor systemic compounds, uptake by the plant.1h4 Theseprocesses all inÑuence spray e†ectiveness. Droplet reten-tion on a plant surface is only one stage in the applica-tion process, but it is critical since only retained spray isuseful for crop protection, while reÑected spray canresult in economic loss and contribute to environmentalcontamination.

Whether a droplet is retained or reÑected by a plantsurface is a function of the properties of the spray solu-tion (surface tension and viscosity), spray pattern

* To whom correspondence should be addressed.” This article is a US Government work and, as such, is in thepublic domain in the United States of America.

(droplet size and velocity),2,5h9 and the surface mor-phology (degree of pubescence, venation, Ðne-structure)and chemistry of surface functional groups of the targetsurface.3,10h13

We propose that a better understanding of the role ofspray additives, delivery systems and plant surfaces onspray efficacy can be achieved by improved measure-ment of droplet reÑection. Reichard14 used high-speedmotion photography and an analytical projector toobserve and analyze the impaction of uniform-sizeddroplets from a spray cloud. This paper describes theuse of a video motion analysis system in conjunctionwith two di†erent droplet generators to observe dropletimpaction. The new technology enables many replicatesof high resolution images to be captured, stored andanalyzed quickly and easily. The objective of this studywas to develop techniques for examining and quantify-ing droplet impaction and reÑection. Results presentedhere are given as an indication of the types of study

2911998 SCI. Pestic. Sci. 0031-613X/98/$17.50. Printed in Great Britain(

292 Donald L . Reichard et al.

possible using the techniques described. They do notattempt to explain all aspects of the impaction process.

Three techniques were developed ; all used monosizeddroplet generators coupled with a high-speed videosystem to capture droplet images. By varying the typeof generator (single-droplet or spray cloud), lighting andcamera magniÐcation, one particular aspect of theimpaction process could be examined in detail. Forexample, when a single droplet was viewed at the air/surface interface using high magniÐcation it was pos-sible to capture images of the impaction processimmediately before and during droplet impaction, anddroplet recovery or reÑection. If lower magniÐcationwas used, the trajectory of the incoming and reÑecteddroplet could be followed, making possible the calcu-lation of droplet velocity before and after impaction aswell as reÑection height. Using low magniÐcation with agenerator that produced a monosized droplet sprayallowed accurate measurement of the height and count(or under some conditions a reasonable estimate) of thenumber of droplets reÑected from a given surface.

2 EXPERIMENTAL METHODS

2.1 Spray application2.1.1 Drop-on-demand generatorA drop-on-demand (DOD) generator, similar to thatdeveloped by Young,15 was constructed to producesingle, monosized droplets.16 Electron microscope aper-tures were used to produce droplets from 100 to 500 lmin diameter, while square-cut hypodermic needlesranging from size 26 through 22 were used to producedroplets from 400 to 800 lm in diameter. Thus the dropsizes obtainable were representative of those producedunder many Ðeld conditions, but were not representa-tive of the smaller droplets (\100 lm) produced by air-assisted sprayers.17 The variability in droplet size andvelocity produced by this generator was less than 4%and 6%, respectively. All droplets had negligible veloc-ity close to the oriÐce ; thus, in order to obtain a rangeof incoming velocities, the target was placed at varyingvertical distances from the oriÐce. Maximum possiblevelocity for a given drop size was its terminal velocity,for example, 0.25 m s~1 for a 100-lm droplet and 1.20m s~1 for a 300-lm droplet. As oriÐce-to-target distanceincreased, the droplet Ñight was a†ected by air move-ment. This was undesirable since, at the magniÐcationsused, the Ðeld of view of the video system was limited,and accurate placement of the droplet was necessary toobtain a focused image. Cardboard tubes were mountedbeneath the generator when necessary to shield dropletsfrom air currents.

The advantages of this system are the ability toproduce single droplets that can be precisely placed ondi†erent parts of the target surface, in a range of sizes

commonly used in pesticide application. The system canalso be used to follow the trajectory (and thus calculatethe velocity) of incoming and reÑecting droplets, and todetermine reÑection height. A disadvantage of thesystem is the inability to produce high-velocity droplets.It can only be used to describe what happens to drop-lets traveling at terminal velocity or slower, for example,at some distance from the sprayer or after the initialdroplet impaction.

2.1.2 Berglund-L iu generatorA Model 3050 Berglund-Liu (BL) vibrating mono-disperse aerosol generator (TSI, St. Paul, MN 55164)was modiÐed to produce a stream of monosizeddroplets18 in the size range of 50È1500 lm. Variation indroplet size and velocity from this machine was lessthan 1%.

A brass ring positioned approximately 1 cm belowthe oriÐce, at the point where the ligament Ðrst brokeup into droplets, was operated at a DC voltage toinduce droplet charges sufficient only to disperse thedroplet stream, providing a spray cone coveringapproximately 1 cm2 at the impaction site. A troughwas positioned just below the oriÐce to collect the spraywhile the plant material was positioned and videosystem readied. Thus impaction always occurred onnew, dry surfaces.

An advantage of the BL over the DOD generator isthe production of a spray cloud, compared to a singledroplet, which more closely simulates hydraulic nozzleapplications. The velocity of droplets produced by theBL generator is less than those used in the Ðeld, e.g. a250-lm droplet would typically have a velocity of 7.0 ms~1 when produced using the BL generator, comparedto 20 m s~1 measured directly beneath the oriÐce of atypical hydraulic nozzle operated at 40 psi.19 However,droplet velocity just above the crop, typically 45 cmbeneath the nozzle, would have decreased to under 5 ms~1 (Downer, R.A., 1997, pers. comm.) which is compa-rable with that obtained from the BL. A further advan-tage is the ability to determine the proportion ofincoming droplets reÑected from a given surface, alongwith reÑection height.

2.2 Video motion analysis system

The video motion analysis system (VMAS) wasequipped with a monochrome video camera (ModelVC-81D, DAGE-MTI,Inc., Michigan City, IN 46360)with a newvicon tube and a high resolution imagingboard (Model 4MEG, EPIX, Inc., Northbrook, IL60062) controlled by a menu-driven motion analysissoftware package 4MIP (EPIX, Inc.), which providedbasic video operations along with a wide selection ofimage examination, processing, and measurement oper-ations.20

V ideographic assessment of droplet impaction and reÑection 293

The imaging board was operated at 24.0 MHz andprovided the necessary timing pulses to drive thecamera at 60 Ðelds per second. This resulted in a non-interlaced (Ðeld) image resolution of 640] 240 pixels ona high resolution monitor (Model MR2000, DAGE-MTI, Inc.). At this resolution, 27 successive imagescould be stored in the memory bu†ers on the imagingboard. EPIX software was used to calculate accuratehorizontal and vertical magniÐcation. Moveable rulerscould be drawn on the screen, and used to measureimage features such as droplet diameters, height ofreÑection from the surface, and distance between suc-cessive images of the same droplet.

2.3 Droplet illumination

To study droplet reÑection, a single stroboscope (Type1538-A, Genrad, Concord, MA 01742) was used toback-light droplets produced using the DOD generator.The Ñash rate of the strobe was approximately seventimes the Ðeld-sequential rate used to drive the camera,thus producing multiple images of the same droplet in asingle frame (Fig. 1).

To illuminate an incoming droplet spray from the BLdroplet generator, two stroboscopes were placed 0.05 mbehind, and to either side of, the target leaf at approx-imately 45¡ to the longitudinal axis of the camera. Thestrobe was used at a Ñash rate higher than the Ðeld rateof the camera, so that several images of the samedroplet were visible in each frame. Trajectories of reÑec-ted droplets were deÐned by the tracks of sequentialimages of each droplet (Fig. 2).

2.4 Droplet impaction

Droplet impaction and deformation were studied on thenon-reÑective surface of a glass microscope slide. AnSTI 3030-series pulsed light-beam proximity detector(Thorrat and Hanmer, Cleveland OH 44116) was placedbetween the DOD oriÐce and the glass slide. Each timea droplet passed through the beam it triggered a singleÑash of the strobe. By specifying the time intervalbetween the triggering of the beam and the Ñash of thestrobe, the position of the droplet relative to the targetwhen the droplet was illuminated could be varied. Byincreasing the time in small (0.1 ms) increments it waspossible to record the impaction process in detail frombefore impaction, through maximum deformation, todroplet recovery. Five 330-lm water droplets, each withan incoming velocity of 0.65 m s~1, were videographedat each time interval.

2.5 Rebound measurements

2.5.1 Drop-on-demandReÑection height and in-Ñight diameter of the dropletswere measured on the monitor, while velocities, bothincoming and rebounding, were calculated. For anytreatment, whether droplet size, surfactant or plantspecies, replication consisted of a single droplet impact-ing onto a new, freshly prepared leaf surface. Ten repli-cate measurements were used per variable. For theintra-species experiment, replication consisted of tensingle-droplet impactions each on a new dry area of thesame leaf for each variable.

Fig. 1. Composite of videographs illustrating an incoming and reÑecting water droplet (340 lm, 0.7 m s~1) during impaction onthe adaxial surface of a wheat leaf. (A) incoming, (B) impacting, (C-E) trajectory of rebound, (F) at maximum reÑection height.

Droplet produced using DOD generator.

294 Donald L . Reichard et al.

Fig. 2. ReÑection of water droplets (340 lm, 7.0 m s~1) from the adaxial surface of a wheat leaf, BL-generated spray.

The velocity of droplets produced using the DODgenerator was determined directly from the monitor bydividing the distance between successive images of thesame droplet by the Ñash rate of the strobe.

The trajectory of the reÑected droplet was seldomvertical, but angled. Since this angle varied within andamong species, the theoretical vertical reÑection heightwas calculated and used to compare reÑection heightsin all cases. Theoretical reÑection height, h (m), was cal-culated from:

h \ v2(2g)~1

where v is the velocity of the rebounding droplet imme-diately after impaction (m s~1), and g the accelerationdue to gravity (9.81 m s~2).

2.5.2 Berglund-L iuSuccessive images, 16.6 ms apart, of the impactionprocess were captured from before any dropletsimpacted until the leaf was thoroughly wet. The singleimage that depicted the Ðrst droplets impacting on apreviously non-sprayed surface was used for analysis.The heights of the 20 highest rebounding droplets wererecorded for each replicate, along with a count or esti-mate of the number of droplets reÑected. Ten replicateswere used per variable. To determine rebound height,an overlay of a horizontal line (an EPIX softwarefeature) was moved slowly from top to the bottom ofthe image. When the line coincided with a dropletwhich was at its maximum rebound height the dropletposition was recorded. The line was then moved down-

ward until the heights of 20 droplets had been recorded.Individual droplets could be distinguished easily whenfew droplets rebounded, but when many dropletsrebounded, the frame was studied and number esti-mated.

2.6 Plant material

Plant species used were cabbage (Brassica oleracea L.cv. Wisconsin All-Season), wheat (T riticum aestivum L.unknown (trial) variety), giant foxtail (Seteria faberiHerrm.), soybean (Glycine max L. cv. Clark) and pear(Pyrus communis L. cv. Bradford). All plants were pre-viously untreated. Leaves uniform in development andfree of damage were selected for all species. All plantswere greenhouse grown, with the exception of pear, theleaves of which were collected from a single pear treegrowing on campus. A small (D0.03] 0.01 m) portionof each leaf was excised and placed, adaxial surface up,on a glass slide covered with double-sided tape. Thisheld the leaf Ñat so that the edges did not obscure thesite of impaction from the camera. Preliminary experi-ments showed the same deposition whether the tapewas used or not. However, a single, Ñat, stationary leafwould not be a typical target in the Ðeld. This labor-atory procedure would probably result in a higherdeposit than would be obtained under Ðeld conditions.The leaf was then positioned directly underneath adroplet generator oriÐce and sprayed. For the DODgenerator, the oriÐce-to-target distance varied (0.02È1.0 m) with the study ; with the BL generator, the dis-tance was 0.12 m.

V ideographic assessment of droplet impaction and reÑection 295

2.7 MagniÐcation

For videographing droplet reÑection from a BL-generated spray, a Micro-Nikkor 55 mm f/2.8 lens wasused to produce an on-monitor magniÐcation ofapproximately 7]. For studying the reÑection of singledroplets, greater magniÐcation was required and thelens was reversed to give a magniÐcation of approx-imately 15]. When viewing images of impacting drop-lets a magniÐcation of 90] was used.

2.8 Surfactants

The surfactants used (Table 1) were representative ofthose commonly used for crop production ; all weretested at a typical Ðeld rate of 1.25 mg litre~1 exceptwhere noted. The water used in all studies was distilledand deionized and had a surface tension of 72 mN m~1at 25 ¡C.

2.9 Impaction/reÑection energy

Detailed discussions of droplet deformation andrecovery have been presented elsewhere.13,22,23 BrieÑy,droplet impaction/reÑection action is governed by theway the source kinetic energy of the incoming droplet ispartitioned during impaction. When the droplet impactson a surface, kinetic energy is transformed into otherforms of energy such as potential energy of the stretcheddroplet surface, potential energy of deformed leaf-surface waxes and the deÑection of the leaf itself. Energy

loss can occur within the droplet, at the interfacebetween the droplet and leaf surface, or to the leaf itself.ReÑection of a droplet occurs when a reforming droplethas sufficient stored energy to overcome adhesion forcesand droplet weight (potential energy) to impart a reÑec-tion velocity to the reformed droplet. Any change indroplet characteristics (incoming velocity, viscosity,surface tension), leaf surface structure or other energysources and sinks will a†ect reÑection velocity andheight. It is difficult to determine the physical propertiesof the spray mixture that relate to the dynamic condi-tions of droplet rebound. Also, the surface nature ofindividual leaves may vary greatly from area to areaand with leaf age, so a deÐnitive energy-balance modelof droplet impaction has not been developed.

3 RESULTS AND DISCUSSION

3.1 Droplet impaction

Water droplets (330 lm diameter) travelling at 0.65 ms~1 were impacted upon a glass surface. A selection ofdroplet images is shown in Fig. 3 ; each image is the onebest described by the mean height of all Ðve droplets atthe respective time interval. Droplet height decreased toa minimum (i.e. maximum deformation) approximately0.5 ms after impaction (Fig. 4). Height then increased asthe droplet started to return to a spherical shape, butenergy loss during impaction was too great to supportreÑection. If the surface had been reÑective, it can beseen from Fig. 4 that the droplet would have left thesurface in less than 1 ms.

TABLE 1Brief Description of Nonionic Surfactants Used

Surfactant Source Principal active ingredient21

Agridex Helena Blend of para†in base petroleum oil, polyol fattyChemical Co.a acid esters and polyethoxylated derivatives.

Components ine†ective as a spray adjuvant 1%.Induce Helena Blend of alkyl polyoxyalkane ether, free fatty

Chemical Co.a acids and isopropanol. Componentsine†ective as a spray adjuvant 10%.

Kinetic Helena Blend of polyalkyleneoxide modiÐedChemical Co.a polydimethylsiloxane and nonionic surfactants.

Components ine†ective as a spray adjuvant 1%.Regulaid Kalo Agricultural Polyethoxypolypropoxypropanol and alkyl

Chemicals, Inc.b 2-ethoxyethanol. Dihydroxy propane.Components ine†ective as a spray adjuvant 9.4%.

Valent Valent USA Blend of alkylarylpolyoxyethylene glycols, freeX-77 Corp.c fatty acids and isopropanol. Components ine†ective

as a spray adjuvant 10%.

a 6075 Poplar Ave., Memphis TN 38119.b 4550 W. 109 St., Overland Park KS 66211.c PO Box 8025, Walnut Creek CA 94596.

296 Donald L . Reichard et al.

Fig. 3. Composite of videographs illustrating the impaction,deformation and initial recovery of 330-lm water droplets ona glass surface. Time interval between successive images was

0.1 ms.

The changes in height, from 0.2 to 1.5 ms, were notstatistically di†erent. This was due to the variabilityamong the Ðve replicates and is indexed by the largestandard deviation. The variability was probably due totiming errors associated with the pulsed nature of theproximity detector. It is likely that use of a proximitydetector with a shorter delay between pulses wouldreduce these errors signiÐcantly.

This system is also suitable for determining changesin surface area during impaction, for accurately deter-mining the contact angle of droplets, and for evaluatingthe characteristics of a droplet as it evaporates from asurface.

3.2 Single-droplet reÑection (DOD generator)

3.2.1 Plant species and droplet sizeWater, a common carrier for pesticide sprays, with thehighest surface tension of all common liquids, is reÑec-ted from the leaves of many species while adhering toothers. All aqueous droplets impacted on cabbageleaves were reÑected, irrespective of droplet size (160È340 lm), whilst none of the droplets was reÑected fromthe smooth, easy-to-wet pear leaves (Table 2). GreaterreÑection (as indexed by height and number) wasobserved from cabbage and wheat than from soybeanand foxtail leaves at the largest droplet size. Only thelargest droplets (340 lm) were totally reÑected from thewheat surface, and droplet size did not appear to inÑu-ence the number of droplets reÑected from soybean andfoxtail leaves. For a constant droplet size (and thusincoming velocity), the reÑection heights di†ered signiÐ-cantly between plant species. An increase in droplet sizeresulted in a corresponding increase in reÑection height

Fig. 4. Changes in water droplet (330 lm, 0.65 m s~1) heightwith time during impaction on a glass microscope slide. Verti-

cal bars refer to ^ one standard deviation.

for all species. Droplet reÑection was species-dependent,varying from a four-fold increase in height for wheatwhen the incoming droplet size was increased from 160to 340 lm, compared to only a 1.3-fold increase forsoybean.

The adaxial surfaces of cabbage and wheat leaves arecovered with hydrophobic crystalline epicuticular wax(the cabbage leaves were glabrous while the wheat hada low density of trichomes), while soybean and foxtailhave pubescent leaf surfaces. If a droplet impacts a tri-chome directly it may not be reÑected because the tri-chome may act as an energy absorber.

For cabbage, a relatively small change in velocityoccurred during the impaction process (velocity of thereÑected droplet was about 30% less than beforeimpaction) compared to foxtail where the velocitydecreased by 60È70%; for both species the change invelocity was independent of drop size (Table 3).

Some reÑected droplets had sufficient energy forseveral reÑections before stopping. Under Ðeld condi-tions, the droplets would either settle on a like target orbe lost to the environment.

3.2.2 Intra-species di†erencesIntra-speciÐc variation is illustrated in Fig. 5 where 10individual water droplets (230 lm, 0.6 m s~1) wereimpacted on either two leaves of di†erent ages from thesame plant or two di†erent areas (that third of the leafcontaining the leaf tip, and the middle third of the leaf)of the adaxial surface of the same leaf. It can be seenthat the e†ect of droplet placement on reÑection wasentirely species-dependent. When the reÑection fromyoung versus old leaves was compared, less reÑectionwas observed from young soybean leaves than from oldones, while the rebound from a young cabbage leaf wassigniÐcantly greater than from a mature one. Similarly,the median part of the foxtail leaf showed twice thereÑection of that for the leaf tip, while the leaf mediangave lower reÑection than the leaf tip for wheat leaves.

V ideographic assessment of droplet impaction and reÑection 297

TABLE 2E†ect of Water Droplet Size on Theoretical ReÑection Height from the Adaxial Leaf Surface of Selected Plant Speciesa

ReÑection height (cm) and (numbers rebounding (%) )

Droplet diameter (km)

Plant 160 230 340

Cabbage 0.58 a (100) 0.96 a (100) 1.17 a (100)Wheat 0.29 b (80) 0.55 b (70) 1.19 a (100)Soybean 0.32 b (70) 0.36 b (60) 0.43 b (50)Foxtail 0.18 bc (50) 0.37 b (80) 0.54 b (70)Pear 0.00 c (0) 0.00 c (0) 0.00 c (0)

a Means separation by DMRT, means in columns with the same letter are not signiÐcantly di†erent at the 5% level.Adapted from Reference 24.

Fig. 5. ReÑection height of water droplets (230 lm, 0.6 m s~1)from di†erent areas of leaves (adaxial surface) of selected plant

species. Vertical bars refer to ^ one standard deviation.

ReÑection also varied with the species tested. Cabbage,for example, had little variation among replicates com-pared to soya, where the variability was large. Thesedata illustrate the need for randomization and sufficientreplication of plant material when investigating factorsthat a†ect droplet reÑection, since variability due to leafposition and age could obscure di†erences among thevariables being tested.

3.2.3 Droplet velocityReÑection height of water droplets impacting oncabbage leaves increased linearly as impact velocityincreased up to 0.90 m s~1 (Fig. 6). Beyond 0.90 m s~1reÑection height did not increase as incoming velocityincreased. The reÑection height of water droplets con-taining the surfactant “KineticÏ (0.2 mg litre~1) did notchange with increase in incoming velocity. Most factors

TABLE 3Comparison of Velocities of Incoming and ReÑecting Water Droplets from Adaxial Leaf Surfaces of Selected Plant Species during

Impactiona

Droplet diameter (km)

160 230 340

V elocity (m s~1) V elocity (m s~1) V elocity (m s~1)Plant bInc. cRef. dL ost bInc. cRef. dL ost bInc. cRef. dL ost

Cabbage 0.47 0.33 29 c 0.59 0.43 27 c 0.69 0.48 31 cWheat 0.49 0.21 58 b 0.62 0.27 56 b 0.68 0.48 28 cSoybean 0.48 0.21 59 b 0.62 0.20 68 b 0.68 0.19 71 bFoxtail 0.48 0.13 74 b 0.61 0.25 59 b 0.71 0.26 63 bPear 0.49 0.00 100 a 0.62 0.00 100 a 0.70 0.00 100 a

a Means separation by DMRT, means in columns with the same letter are not signiÐcantly di†erent at the 5% level.b Velocity of incoming droplet.c Velocity of reÑected droplet, immediately after reÑection.d Velocity lost during impaction (%).

298 Donald L . Reichard et al.

Fig. 6. Relationship between rebound height and incomingdroplet velocity of 350 lm water and 0.2 mg litre~1(L) (|)“KineticÏ droplets on cabbage leaves. Vertical bars refer to

^ one standard deviation.

a†ecting reÑection in this study, except surface tension,were approximately equal for water droplets with andwithout surfactant. Therefore di†erences in reÑection-height/incoming-velocity relationships were probablydue to the di†erence in the surface tension of the liquids(72 and 29 mN m~1 for water and 0.2 mg litre~1“KineticÏ, respectively). Lower surface tension of a liquidcauses the droplet to spread more on impact (or even tofragment), which increases contact area and energy loss,and reduces energy stored in the newly created surfacearea. This reduces the energy imparted to a reformingdroplet, and hence reduces reÑection height. Prelimi-nary results had indicated that concentrations of“KineticÏ greater than 0.2 mg litre~1 gave zero reÑectionfrom cabbage leaves at droplet velocities less than1.40 m s~1.

3.3 Spray-cloud reÑection (BL generator)

3.3.1 E†ect of surfactantSpray clouds of 260-lm droplets generated using theBL were impacted upon the adaxial surface of newlyfully expanded cabbage leaves. Cabbage has a hard-to-wet surface and was chosen to provide a severe test onthe e†ectiveness of the various surfactant at reducingrebound. Additional experimentation is needed to deter-mine whether the results below would be obtained fromother plant surfaces, or whether there is a surfactantspecies interaction.

All surfactants reduced the number and height ofreÑected droplets compared to water only. Thus, thepotential for improving spraying efficiency wasincreased and o†-target contamination reduced. Therewas a Ðve-fold di†erence in reduction of reÑectionheight between the most and least e†ective surfactants(Table 4). Surfactants are added to formulations for avariety of reasons, including improving the efficacy ofthe active ingredient(s), maintaining efficacy with lessactive ingredient, increasing the wetting, spreading oradhesion properties, and acting as emulsifying or dis-

TABLE 4E†ect of Surfactants (1.25 mg litre~1) on ReÑection ofAqueous Droplets (260 lm, 7 m s~1) from the Adaxial Surface

of Cabbage Leavesa

Reduction ReductionHeight of in in no.reÑection reÑection Numbers reÑected

Surfactant (cm) height (%) reÑected (%)

Water 1.96 a 88.4 aAgridex 1.67 b 15.0 65.8 b 25.6Induce 1.35 c 31.1 51.9 c 41.3Valent X-77 1.20 d 38.6 46.4 c 47.5Regulaid 0.79 e 59.9 26.8 d 69.7Kinetic 0.46 f 76.6 14.9 e 83.1

a Means separation by DMRT, means in columns with thesame letter are not signiÐcantly di†erent at the 5% level.Adapted from Reference 24.

persing agents in spray tank mixtures. Thus, reducingreÑection may not be the primary reason for using asurfactant, but an additional beneÐt of using surfactantsis to improve droplet capture.

3.3.2 E†ect of surfactant concentrationThere was good correlation between concentration ofthe surfactant “KineticÏ and rebound height (Fig. 7).Rebound was not eliminated at the concentrationstested but it was reduced signiÐcantly at the higherrates.

Impaction is a dynamic process with the dropletremaining on a leaf for less than 1 ms.3,14 The dropletundergoes extensive deformation during which newsurface is formed in a short time and unless surfactant

Fig. 7. E†ect of increasing surfactant (Kinetic) concentrationon height (y \ 2.5[ 2.8 ln x, r2\ 0.992) and number(L) (|)(y \ 47.6[ 2.3 ln x, r2\ 0.963) of water droplets (230 lm, 7 ms~1) reÑected from the adaxial surface of wheat leaves. Separa-tion of means for each concentration signiÐcant at the 5%

level, using DMRT.

V ideographic assessment of droplet impaction and reÑection 299

molecules are present at the new surface, or can di†useto it during droplet deformation, surface tension willnot be reduced and the surfactant will not reducerebound. As the surfactant concentration is increased,more surfactant molecules will reach the newly createdsurface during the impaction-process time. When thisoccurs, surface tension will be lowered and reÑectionwill decrease.

4 CONCLUSIONS

In this study, quantitative di†erences in reÑection wereobtained from a number of parameters, namely plantspecies, droplet velocity and size, surfactant type andconcentration.

Combining the video-motion analysis system witheither a single or spray monosized droplet generatorresulted in a versatile system for examining dropletimpaction. By using low magniÐcation and the BL gen-erator, the system could be used as either a fast, qualit-ative technique for assessing large di†erences inreÑection of a spray cloud, or more quantitatively bydetermining reÑection height and number of dropletsreÑected. With the DOD generator, detailed informa-tion on the incoming and outgoing velocity, along withreÑection height, of individual droplets could beobtained. By using the DOD generator and high mag-niÐcation, droplet impaction and subsequent deforma-tion could be examined.

Names of equipment and products are necessary to report fac-tually on available data ; however, the US Department ofAgriculture, Ohio State University, and Michigan State Uni-versity neither guarantee nor warrant the standard of theproduct, and the use of the name of USDA, OSU, or MSUimplies no approval of the product to the exclusion of othersthat may also be suitable.

ACKNOWLEDGEMENTS

The authors thank P. T. Keck for technical assistance.

REFERENCES

1. Young, B. W., Studies on the retention and deposit char-acteristics of pesticide sprays on foliage. Paper F-1È4 :Application Meeting, IXth CIGR Congress, East Lansing,MI, USA, 1979.

2. Stevens, P. J. G., Kimberly, M. O., Murphy, D. S. & Poli-cello, G. A., Adhesion of spray droplets to foliage : the roleof dynamic surface tension and advantages oforganosilicone surfactants, Pestic. Sci., 38 (1993) 237È45.

3. Wirth, W., Storp, S. & Jacobsen, W., Mechanisms control-ling leaf retention of agricultural spray solutions, Pestic.Sci., 33 (1991) 411È20.

4. Steurbaut, W., Adjuvants for use with foliar fungicides.Pestic. Sci., 38 (1993) 85È91.

5. Blackman, G. E., Bruce, R. S. & Holly J., Studies in theprinciples of phytotoxicity. J. Exp. Bot., 9 (1958) 175È205.

6. Kirkwood, R. C., Use and mode of action of adjuvants forherbicides : a review of some current work. Pestic. Sci., 38(1993) 93È102.

7. Furmidge, C. G. L., Physicochemical studies on agricultu-ral sprays IVÈthe retention of spray liquids on leaf sur-faces. J. Sci. Food Agric., 13 (1962) 127È40.

8. Taylor, W. A. & Shaw, G. B., The e†ect of drop speed, sizeand surfactant on the deposition of spray on barley andradish or mustard. Pestic. Sci., 14 (1983) 659È65.

9. Lake, J. R. & Marchant, J. A., The use of dimensionalanalysis in a study of drop retention on barley. Pestic. Sci.,14 (1983) 638È45.

10. Ennis, W. B., Williamson, R. E. & Dorschner, K. P.,Studies on spray retention by leaves of di†erent plants.W eeds, 1 (1952) 274È86.

11. Baker, E. A., The inÑuence of environment on leaf waxdevelopment in Brassica oleracea. New Phytol., 73 (1974)955È66.

12. Boize, L., Gudin, C. & Purdue, G., The inÑuence of leafsurface roughness on the spreading of oil spray drops.Ann. Appl. Biol., 84 (1976) 205È11

13. Spillman, J. J., Spray impaction, retention and adhesion :an introduction to basic characteristics. Pestic. Sci., 15(1984) 97È106.

14. Reichard, D. L., Drop formation and impaction on theplant. W eed T echnol., 2 (1988) 82È7.

15. Young, B. W., A device for the controlled production andplacement of individual drops. In Pesticide Formulationsand Application Systems : 5th V ol. AST M ST P 915, ed.L. D. Spicer & T. M Kaneko. Philadelphia, PA, 1986, pp.13È22.

16. Reichard, D. L., A system for producing various sizes,numbers, and frequencies of uniform-size drops. T rans.ASAE, 33 (1990) 1767È70.

17. Reichard, D. L., Retzer, H. J., Liljedahl, L. A. & Hall,F. R., Spray droplet size distributions delivered by air-blast orchard sprayers. T rans. ASAE, 20 (1977) 232È42.

18. Reichard, D. L., Brazee, R. D., Bukovac, M. J. & Fox,R. D., A system for photographically studying dropletimpaction on leaf surfaces. T rans. ASAE, 29 (1986)707È13.

19. Miller, P. C. H., Spray drift and its measurement. InApplication T echnology for Crop Protection, ed. G. A.Matthews & E. C. Hislop. CAB International, Wall-ingford, UK, 1993.

20. Fox, R. D., Reichard, D. L. & Brazee, R. D., A videoanalysis system for measuring droplet motion. AppliedEngineering in Agriculture, 8 (1992) 153È7.

21. Thomson L. H., In A Guide to Agricultural Spray Adju-vants used in the United States. Thomson Publications,PO Box 9335, Fresno, CA 93791, 1992È3.

22. Brazee, R. D., Reichard, D. R., Bukovac, M. J. & Fox,R. D., A partitioned energy transfer model for sprayimpaction on plants. J. Agri. Engr. Res., 50 (1991) 11È24.

23. Tadros, T. F., Interactions at interfaces and e†ects ontransfer and performance, Asp. Appl. Biol., 14 (1987) 1È22.

24. Bukovac, M. J., Leon, J. M., Cooper, J. A., Whitmoyer,R. E., Reichard, D. L. & Brazee, R. D., Spraydroplet :plant surface interaction and deposit formation asrelated to surfactants and spray volume. In Proc. 4thInternat. Symp. Adjuvants for Agrochemicals (1995), (FRIBulletin No. 193), ed. R. E. Gaskin, NZ, pp. 177È85.