polydisperse ethoxylated fatty alcohol surfactants as accelerators of cuticular penetration. 1....

Pestic. Sci. 1997, 51, 131È152

Polydisperse Ethoxylated Fatty AlcoholSurfactants as Accelerators of CuticularPenetration. 1. Effects of Ethoxy Chain Lengthand the Size of the Penetrants

Peter Baur,1* B. Terence Grayson2 & Jo� rg Scho� nherr1

1 Institut fu� r Obstbau und Baumschule, Universita� t Hannover Am Steinberg 3, 31157 Sarstedt, Germany2 Department of Biological Sciences, University of Portsmouth, King Henry I Street, Portsmouth PO12DY, UK

(Received 4 November 1996 ; accepted 28 May 1997)

Abstract : The e†ects of polydisperse ethoxylated fatty alcohol (EFA) surfactantson the penetration of six organic compounds varying in size (molar volumes,107È282 cm3 mol~1) and lipophilicity (log 0É8È6É5) were investigated usingKowastomatous isolated cuticular membranes (CM) of Citrus and pear leaves. Mobi-lities of model compounds in CM were measured by unilateral desorption fromthe outer surface (UDOS). Rate constants (k*) obtained in these experiments aredirectly proportional to di†usion coefficients and, in the absence of EFA, k*values decreased by a factor of 52 when molar volumes increased only 2É64-fold.Under UDOS conditions using micellar surfactant solutions as desorptionmedia, surfactants are sorbed in the CM and the volume fractions sorbed werefound to decrease from approximately 0É062 to 0É018 when the average numberof ethoxy groups (nE) increased from 5 to 17. In the presence of the EFA sur-factants in the CM, solute mobilities increased markedly though this e†ectdiminished with increasing nE. Surfactants with nE\ 17 a†ected solute mobi-lities only marginally. Surfactant e†ects on solute mobility increased with the sizeof the solutes leading to almost identical mobilities of the model compounds.With the current range of our model compounds, lipophilicity increased withincreasing molar volumes, though evidence is presented showing that the mobi-lities of solutes depend on their molar volumes while lipophilicity has no e†ect.E†ects of micellar aqueous solutions of polydisperse surfactants on solute mobi-lities followed the pattern observed with monodisperse ones.

Experiments simulating foliar uptake (SOFU), by applying 5-kl droplets ofsolute and increasing amounts of surfactant (1È20 g litre~1) to the outer surfaceof the CM and then monitoring the rates of appearance at the inner surfaces ofthe CM, were carried out with 1-(3-Ñuoromethylphenyl)-5-phenoxy-5H-1,2,3,4-tetrazole (WL110547 ; log and cyanazine (log and the sur-Kow \ 3É6) Kow \ 2É1)factants “GenapolÏ C-050 (GP C-050, nE 5) and “GenapolÏ C-200 (GP C-200, nE17). In the applied range, uptake increased with increase in amount of both sur-factants and both surfactants were more e†ective than polyethylene glycol (PEG400). The smaller, more lipophilic GP C-050 was able to increase the rates ofuptake of both compounds greatly, while GP C-200 had a less, though observ-able, e†ect, similar to the pattern observed in the UDOS experiments. The resultsindicated that GP C-050 penetrated the cuticle rapidly while sorption of GPC-200 was slower, though better from a concentrated residue in SOFU experi-

* To whom correspondence should be addressed.Contract grant sponsor : Shell Forschung AG.

1311997 SCI. Pestic. Sci. 0031-613X/97/$17.50. Printed in Great Britain(

132 Peter Baur, B. T erence Grayson, Jo� rg Scho� nherr

ments than from aqueous solution in UDOS experiments. This induced someacceleration e†ects by GP C-200 in SOFU experiments, but these could becountered by reductions in concentrations, and hence driving forces, for penetrat-ion at higher levels of application.

Pestic Sci., 51, 131È152, 1997No. of Figures : 12. No. of Tables : 4. No. of Refs : 49

Key words : adjuvants, di†usion, foliar uptake, mobility, permeability, surfactant

1 INTRODUCTION

Pesticides formulations for foliar application generallycontain surfactants. With regard to their intended func-tions surfactants have been classiÐed by McWhorterand Van Valkenburg as (i) spray modiÐers whichincrease spray retention and coverage of the targets and(ii) activators which improve biological e†ectiveness ofactive ingredients.1,2 In more recent work utilisingradio-labelled active ingredients, a better understandingof the mode of action of activator adjuvants has beenattempted3h10 but, despite these e†orts, the exact sitesand mechanisms of action of these activators haveremained unclear, largely due to the complexity of thefoliar uptake process.11,12 Rates of foliar uptake areproportional to permeabilities of cuticles and to drivingforces and these factors cannot be separated when onlyamounts or rates of foliar uptake from droplets aremeasured.13 Permeabilities of cuticles have been studiedusing isolated cuticles and the system aqueous donor/cuticle/aqueous receiver. In this type of study, drivingforces are known and permeabilities can be calculatedfrom the rates of penetration. Such studies have demon-strated that permeabilities of cuticles can be increasedby certain surfactants7,14h19 and, since the surfactantscannot a†ect the concentration, and hence the drivingforce, of the solute, the most likely explanation is thatthe surfactants have increased the permeability of thecuticles. This was demonstrated directly by using a newexperimental procedure called UDOS (unilateraldesorption from the outer surface).17,20,21 With thismethod, solute mobilities in cuticles and e†ects of adju-vants on solute mobility in cuticles can be measuredand di†usion coefficients estimated.22 Studies withmonodisperse ethoxylated alcohols applied as aqueousmicellar solutions revealed that (i) alcohols were moste†ective, (ii) polyethylene glycols had no inÑuence onsolute mobility in cuticles, (iii) ethoxylation decreasede†ects on solute mobility, (iv) e†ects on mobility werenot correlated with HLB and (v) e†ects on solute mobil-ity were not always obtained instantaneously.20,21 Withincreasing size (molecular mass) of surfactants, moretime was needed until maximum e†ects were observed.This indicates that surfactants must penetrate into cut-icles before they can exert an e†ect on solute mobility.From these studies it was suggested that surfactants aresorbed into cuticular waxes and increase their Ñuidity.As a consequence, the mobility of solutes sorbed in

these Ñuidised amorphous cuticular waxes areincreased. This was conÐrmed by recent work withreconstituted waxes where wax/water partition coeffi-cients for monodisperse ethoxylated alcohols and thee†ects of these surfactants on mobilities of pentachlo-rophenol and tetracosanoic acid in isolated cuticlarwaxes were measured.23,24 Surfactant e†ects on di†u-sion coefficients (D) in the wax of these two compoundsincreased with increasing amounts of surfactants sorbedin the wax. Using ESR spectroscopy it was shown thatthe Ñuidity of cuticular waxes increased in the presenceof those sorbed surfactants, which acted as plasticisers.

Sorption of ethoxylated alcohols and free alcoholsfrom aqueous solutions into isolated cuticles followssimple laws.25 Cuticle/water partition coefficients, theCMC and amount of surfactant sorbed can each be cal-culated using a set of linear equations from the numberof carbon atoms in the alcohol and the number of theethoxy groups, respectively. With this information, theamounts of surfactant in cuticles in the previous UDOSexperiments,20,21 studying e†ects of fatty alcohols andethoxylated alcohols on mobility of 2,4-D in Citrus leafcuticular membranes, were calculated. The dose-e†ectcurve obtained was linear, showing that the surfactante†ect was non-speciÐc and depended only on theamounts of surfactant sorbed in the cuticles. Thus, thenumber of the carbon atoms in the alcohol and thenumber of ethoxy groups a†ected the CMC, the parti-tion coefficient and rates of penetration, leading to dif-fering amounts sorbed, but did not a†ect the intrinsice†ectiveness of the individual homologues.

Adjuvants which increase solute mobility in cuticlesand cuticlar waxes have been termed accelerator adju-vants. As already mentioned, ethoxylation is notrequired for accelerator action, as fatty alcohols can bepowerful accelerators.21 Since fatty alcohols can besurface-active there is still a chance that surface activitymight be necessary. However, many other compoundswhich are not surface-active (Baur, P., Scho� nherr J.,unpublished results) including chlorfenvinphos26 alsoexhibit strong accelerator activity, showing that surfaceactivity is not essential. Chlorfenvinphos was found tobe a very e†ective accelerator and its own mobility inpear leaf cuticles was strongly concentration-dependent.26 A high volume fraction of chlorfenvinphoswas accompanied by a decrease of the activation energyof di†usion and the compound thus behaved in asimilar way to external plasticisers of polymers. Which

EFA Surfactants and Cuticular Penetration. 1. Ethoxy Chain L ength 133

property exactly imparts good accelerator activity to acompound is not known at this time, but the compoundmust be soluble in cuticlar waxes, since e†ects increasewith increasing concentration in cuticular waxes.24h26A liquid state and lipophilicity alone are not sufficientprerequisites.26

The above results obtained using isolated cuticles andreconstituted waxes help to understand better adjuvantaction on solute mobility in cuticles and on the per-meability of cuticles. However, rates of uptake fromdroplets and spray residues depend on permeability anddriving forces, just as electrical current depends on con-ductivity and potential gradient. If there is no potentialgradient, no current will Ñow no matter what the con-ductivity is. The same holds for foliar uptake. Slow pen-etration does not necessarily imply low permeability ofcuticles. It could result from small driving forces. Thisaspect has been discussed recently by Scho� nherr andBaur13 and Baur and Scho� nherr27 who also pointedout that the aspect of driving forces has generally beenneglected in earlier investigations, where only biologicale†ects or rates of penetration have been measured. Ifdroplets of radio-labelled active ingredients are appliedto leaves, only the rates of disappearance from thesurface can be measured, while driving forces (whichvary with time) remain unknown and their e†ects on therates of uptake are usually neither discussed noracknowledged.

Stock et al.9 studied uptake into leaves of model com-pounds di†ering in molecular weight and partition coef-Ðcients with particular reference to e†ects of ethoxylatedalcohols di†ering in numbers of ethoxy groups. Smalldroplets (0É2 kl) were applied and disappearance fromthe surface residue was monitored at various time inter-vals. Rates of uptake of the polar methylglucose werebest enhanced by surfactants having 15 or 20 ethoxygroups, while rates of uptake of the lipophilic per-methrin were highest when only six ethoxy groups wereattached to the alcohols. From the data, aC13/C14three-dimensional qualitative model was derived whichrelated rates of uptake to surfactant ethoxy content andoctanol/water partition coefficient of the model com-pounds. This model was the Ðrst systematic approach toshow a qualitative relationship between the nature ofthe penetrant and that of the accelerator adjuvantrequired for optimal penetration across a wide spectrumof lipophilicity (log [3 to ]6) of penetrants. TheKowpaper hinted that other factors such as the role of watersorption in the surface residue, molecular mass of thepenetrant, the nature of the cuticle, the pH for ionisablepenetrants, etc. would be involved in inÑuencing theactual rates of penetration of active ingredient and adju-vant. However, since the method of application was toplace deposits (which dried) on the leaves attached toplants, the e†ect of the surfactant adjuvant on the con-centrations (driving forces) of the compounds penetrat-ing was an unknown factor and this prevented true

kinetic and thermodynamic treatment of the data,which might have aided understanding of the process atthe molecular level. In order to separate these e†ects wehave used a similar set of model compounds and similarpolydisperse surfactants and have studied their e†ectson solute mobilities in, and penetration across, isolatedcuticles and related these to stationary surfactant con-centrations in cuticles and cuticular waxes.

2 MATERIALS AND METHODS

2.1 Cuticular membranes

Adaxial astomatous cuticular membranes were isolatedenzymatically from mature leaves of bitter oranges(Citrus aurantium L.) and pear (Pyrus communis L. cv.Bartlett) in 1991 and 1992 as described elsewhere.28Bitter orange plants were grown in growth chamberswhile pear leaves were taken from trees in an orchard inBavaria in July. Isolated cuticles were obtained fromthese leaves by enzymatic degradation as described pre-viously, air dried and stored for at least four weeks atabout 8¡C prior to use. During this initial storage ofisolated cuticles, their water solute permeabilitiesdecrease somewhat, which might be induced by struc-tural rearrangement of cuticular waxes and/or loss ofplasticising volatiles from cuticles.29 After about fourweeksÏ storage, permeabilities of cuticles become moreconstant and time independent.

2.2 Chemicals

All transport experiments were carried out with 14C-labelled compounds (Table 1) which, in the following,will be referred to as solutes. They were stored as solu-tions in recommended solvents in the refrigerator.Before use the solvents were removed by a gentle streamof nitrogen and the compounds redissolved in di†erentsolvents containing mainly water (see Section 2.3).Applied concentrations varied in the experimentsbetween 0É07 (cyanazine) and 1 mM (permethrin).

The surfactants studied were polydisperse fattyalcohol ethoxylates (alkyl-u-hydroxypoly (oxy-1,2-eth-anediyl)) of the “GenapolÏ (GP) C series (Hoechst AG,Frankfurt, Germany). All surfactants had the same dis-tribution of the alkyl chain length (Fig. 1A) the mainconstituent of the alkyl moiety being dodecanol (48 to58%) followed by tetradecanol (19 to 24%), hexadecanol(9 to 13%), octadecanol (5 to 9%) and decanol (0 to 3%)(data from Hoechst AG, Frankfurt). The weighted meannumber of carbons in the alcohols was 12É5. The dis-tribution of the number of ethoxy groups (E) peralcohol together with the calculated mean value aregiven in Fig. 1A. The mean value resembles theweighted mean of all E derivatives (all alkyl chainlengths). The frequency distribution of E number per


TABLE 1List of Radiolabelled Test Compounds

SpeciÐcactivity Radiochemical

Compound Chemical name (MBq mmol~1) purity (%)

Phenylureaa [Carbonyl-14C]phenylurea 148 992,4-Db 2,4-Dichlorophenoxy-[1-14C] acetic acid 329 [98Cyanazinea 2-(4-Chloro-6-[1-14C]ethylamino-1,3,5-triazin-2-ylamino) 507 99É5

-2-methylpropionitrileWL110547a 1-(3-Fluoromethylphenyl)-5-[U- 503 99

14C]phenoxy-5H-1,2,3,4-tetrazoleChlorfenvinphosa 2-Chloro-1-(2,4-dichloro-[U- 303 99

14C]phenyl)vinyl diethyl phosphatePermethrina 3-phenoxybenzyl (1RS)cis, trans-3-(-2,2-dichlorovinyl)-2,2- 68 96É7

dimethylcyclopropanecarboxylate

a Shell Research Centre, Sittingbourne, UK.b Sigma Chemie, Deisenhofen, Germany.

Fig. 1. (A) Ethylene oxide (Mol E) content per alcohol homo-logue and alkyl length distribution (see inset) of fatty alcoholethoxylates of the “GenapolÏ C series. The Ðgure in bracketsgives the weighted mean number of ethoxy groups. (B)Volume fractions of ethoxylate homologues in the cuticleC12equilibrated with a micellar aqueous solution of the respectivesurfactant. The Ðgure in brackets gives the total volume frac-

tion for each surfactant.

alcohol is very broad and does not obey the Poissondistribution which is often found for nonionic sur-factants. A Poisson distribution is obtained from acidcatalysis, while alkaline conditions result in a broaderdistribution.30 The surfactants contained 1É6 (GP C-080/100) to 4É8 (C-050)% free alcohols and 2É2 (GPC-050) to 3É5 (C-100) unidentiÐed impurities. GP C-050was liquid at room temperature while GP C-080 andGP-C100 were pastelike and GP-C200 was a waxysolid. All surfactants were free of water (\0É5%) andwere stored at 4¡C. The pH of 0É5% solutions variedbetween 5É7 (GP C-050) and 6É0 (GP C-200).

In all transport experiments a 10 g litre~1 phospho-lipid (soybean lecithin) suspension (PLS) was used ascontrol desorption medium. PLS does not change cut-iclar permeability and does not measurably penetrateinto the CM22,31 but serves as a sink for permeatedlipophilic compounds owing to its sorption proper-ties.17,32,33

2.3 Experimental

2.3.1 Unilateral Desorption from the Outer Surface(UDOS)Mobilities of compounds were determined by uni-laterally desorbing radiolabelled solutes contained inthe CM from the outer surface (UDOS). Details of themethod have been described elsewhere17,22 and onlythe principles and modiÐcations will be described here.Cuticles were inserted between the lid and the desorp-tion chamber with the morphological outer surfacefacing the chamber interior. The donor solutions con-taining the radiolabelled compounds were applied as50-kl droplets to the centre of the morphological innersurface of the CM. It is essential in these UDOS desorp-tion experiments that, during evaporation of solvents(3È6 h), the compounds are quantitatively sorbed into


the cuticle to avoid solid residues being present on itsinner surface. Owing to the di†erent speciÐc activities(Table 1) and physicochemical properties (Table 2) thedonor solutions had to be modiÐed for each compound.The amounts applied varied between 0É8 kg (cyanazine)and 20 kg (permethrin). WL110547 and chlorfenvinphoswere applied in water containing 400 and 250 g kg~1ethanol, respectively. Chlorfenvinphos theoretically hasa sufficient water solubility, but alcohol was addedbecause tiny droplets were visible on the bottom of theglass vial when dissolved in water only, indicating aslow dissolution process. 2,4-D is a weak acid with a

of 2É7334 and it was dissolved in 0É01 M lactic acidpKabu†er adjusted to pH 3 with potassium hydroxide.Phenylurea and cyanazine were Ðrst dissolved in waterand applied to Citrus cuticles. However, preliminarydesorption experiments indicated that these compoundswere not quantitatively sorbed in the cuticle, and scan-ning electron microscope (SEM) pictures showed solidresidues on the inner surface of the CM.35 In order toavoid solid residues of these compounds, polyethyl-eneglycol 400 (PEG 400) was added to the aqueoussolutions because it is an involatile liquid and hasenough solvent power for both compounds.35 Amountsof 200 kg per CM were found to be optimum in prelimi-nary experiments. PEG 400 keeps the compounds in adissolved state and does not alter cuticular per-meability. This was conÐrmed by adding PEG 400 todonor solutions of WL110547 which was easily sorbedinto the CM (see Section 3). Similarly, even the lipo-philic permethrin was not quantitatively sorbed into theCitrus CM during evaporation of the water and alcoholsolvents. In this case inert phospholipid was added tothe donor solutions. An amount of 400 kg per cuticlewas chosen since desorption rates were maximum andthe treated area was covered completely with thisamount as shown by SEM.

After evaporation of the volatile solvents of the donorsolutions, the chambers were closed by an adhesive tapeon the donor side and desorption from the outer surfacewas started the next day by pipetting into the chambera 10 g litre~1 aqueous soybean lecithin suspension

(phospholipid suspension, PLS) through a samplingport. For calculating rate constants (k*) for UDOS datait is assumed that the concentration of the desorbedcompound in the receiver is zero. This assumption isfulÐlled if PLS is used as desorption medium for lipo-philic neutral solutes which completely partition intothe liposomes. During desorption the chambers wererocked horizontally while standing with the lids facingdownward in wells of a thermostated (25¡C) aluminiumblock. At predetermined intervals the desorptionmedium (PLS) was withdrawn and replaced by a freshone. Desorption with PLS was carried out for two(Pyrus) to four (Citrus) days, and after it was continuedwith a micellar solution (5 g litre~1) of the surfactantsagain between two and four days. After the last desorp-tion step the CM were cut out and the residual radioac-tivity in the CM extracted using scintillation cocktail(Aquasafe 500, Zinsser, Frankfurt, Germany). Radioac-tivity in the desorption media and CM was assayedusing a liquid scintillation counter (Packard CA 2000counter, Downers Grove, IL, USA).

2.3.2 Simulation of foliar uptake (SOFU)The whole penetration process from a solute/surfactantresidue across the cuticle was simulated by applyingdroplets containing adjuvants and test compounds tothe outer surface of CM followed by desorption fromthe inner surfaces of CM.

Simulation of foliar uptake (SOFU) was carried outwith cyanazine and WL110547 as model compoundswith moderate and strong lipophilic character, andPyrus cuticular membranes. The amount of test com-pound was always approximately 1 kg and the donorvolume was 5 kl which corresponds to concentrationsof about 0É8 mM. The compositions of “GenapolÏ sur-factants are quite similar (see Fig. 1) and therefore onlyGP C-050 and GP C-200 were used at concentrationsof 1, 2, 10, 20 g litre~1, which is equal to amounts of 5to 100 kg. Droplets with GP C-050 spread much morethan those of GP C-200, with a slight dependence onconcentration being observed. Attempts were made tocover equal areas but, particularly at higher concentra-

TABLE 2Relative Molecular Mass (M), Molar Volumes Octanol/Water Partition Coefficients, and(Vx), (Kow)

Water Solubilities of Test Compounds(SH2O)

No. Compound M Vxa (cm3 mol~1) log K

OWSH2O

(25¡C)b (g litre~1)

1 Phenylurea 136 107 0É8 4É12 2,4-D 221 138 2É8 0É83 Cyanazine 241 177 2É1 0É174 WL110547 306 192 3É6 3 ] 10~35 Chlorfenvinphos 359 233 3É5 0É1456 Permethrin 391 282 6É1 0É2 ] 10~3

a Reference 44.b From Shell Research Ltd and for 2,4-D, Reference 34.


tions, the areas with GP-200 applications were still lessthan with GP C-050 due to the further spreading of GPC-050 some minutes after application. Control experi-ments were carried out by applying the compounds in5 kl water or in aqueous PEG 400 (40 g litre~1). Theexperimental conditions were as described above(Section 2.3.1). However, in contrast to the UDOSexperiments, the chambers with cuticles were completelyÐlled with desorption medium (PLS) and stood uprightduring the Ðrst 2 h with the sampling port being closedby adhesive tape. The bulk of the water in the 5-kldonor droplets had evaporated after 30È40 min leavinga liquid solute/surfactant residue. After the Ðrst sample(2 h), the adhesive tape closing the sampling port wasremoved and the chambers inverted so that they facedthe aluminium block for the remainder of the experi-ment (24 to 33 h).

2.3.3 Calculation of rate constants describing di†usion(k*) in and penetration (k) across cuticular membranesIn UDOS experiments, rates of di†usion were charac-terised by plotting the natural logarithm (ln) of the frac-tion of the applied test compound remaining in the CMagainst time. The ratio of the concentration of theCtradiolabelled solute in the CM at time t to the originalconcentration at t \ 0 decreases according to Ðrst-C0order kinetics.

CtC0

\ e~kRt (1)

with k* being the Ðrst-order rate constant of desorption.The asterisk denotes that the driving force is the soluteconcentration in the cuticle and this rate constant (k*) istherefore independent of solute lipophilicity asexpressed by the partition coefficient K.22 Plotting

against t yielded straight lines with slope[ ln(Ct/C0)k*. Since the volume of the CM in which solutes aresorbed in constant, the ratio is equivalent to theCt/C0expression where is the amount of1 [ Mt/M0 , Mtradiolabelled substance desorbed at time t, and theM0amount sorbed initially in the CM, which was calcu-lated by summation of the amounts desorbed plus thatremaining in the CM after termination of desorption.Because desorption was initially with PLS alone andthen subsequently with surfactant solutions, two slopeswere obtained and the ratio of desorption with sur-factant to that with PLS was used to obtain a surfactante†ect on mobility of the test compound.17,20,21

Equation (1) is only valid if the concentration of pen-etrants in the water of the receiver is zero. This isachieved through the good sorptive properties of PLS,which absorbs compounds di†ering in physicochemicalproperties into di†erent regions of the vesicles.32Aqueous surfactant solutions above the critical micelleconcentration can also solubilise lipophilic compoundsinto the interior of micelles.36 The CMC of the Genapol

surfactants used in this study di†ered only slightlyabout a value of D0É03 g litre~1 (25¡C), as determinedby the Wilhelmy plate method and was well below thelowest concentration used. Surfactant micelles wouldtherefore have also acted as sinks for the permeatedlipophilic test compounds.

In SOFU experiments, the droplet on the surface ofthe CM or the hydrated surfactant/solute residueremaining after evaporation of the bulk water served asdonor. Radio-labelled solutes and surfactant moleculesdi†use through the CM into the receiver solutionsfacing the inner surfaces of the CM. The whole processcan be approximated by an equation similar to eqn (1),with the slopes of the SOFU penetration graphs beingrate constants (k), though they are not equivalent to k*obtained in UDOS experiments, where the solute con-centration in the cuticle serves as donor :

CtC0

\ e~kt (2)

The di†erence is that, while in UDOS experiments thedonor volume remains constant during the experiment,this is generally not the case in SOFU since volatile sol-vents (water etc.) evaporate and both solutes and sur-factants penetrate into the cuticle during thisevaporation to an unknown extent leaving a surfaceresidue. This means that there are at least two steps, inseries, in the overall penetration process, i.e. Ðrstly fromthe surface residue into the cuticular wax and, secondly,from the cuticle into the adjacent cell wall. The Ðrst stepcannot be Ðrst-order since the volume, and hence theconcentration of the penetrants, changes with time thusrendering the overall process not Ðrst-order. SOFUdesorption graphs are only strictly Ðrst-order when pen-etration proceeds from a non-volatile and non-penetrating solvent on the surface of the cuticle underconstant experimental conditions in humidity and tem-perature.13,37 In this special case only, the concentra-tion of the solute in the donor decreases exponentiallywith time, which results in linear graphs according toeqn (2). Between 10 and 50 CM (replications) were usedfor each species/compound/surfactant combination inUDOS and SOFU, respectively. Variability is expressedas 95% conÐdence intervals and is caused mainly byvariability among individual CM.

3 RESULTS

3.1 Preliminary tests

3.1.1 E†ect of the amount sorbed and areas ofapplication on the rates of permeation in the UDOSmethodOwing to di†erent compositions of the donor solutionsin UDOS, the treated area of the CM varied in theseexperiments. Rate constants of desorption should not


depend on the area if a compound is sorbed into theCM when applied to the inner side. In experiments withWL110547, a compound which causes no problems inUDOS experiments, rates of permeation using di†erentdonor volumes and concentrations of WL110547 anddi†erent treated areas were measured with Citrus CM(Table 3). No signiÐcant di†erences were found, inaccord with expectations. In one experiment PEG 400was added in a 50-kl droplet using the same amount asnecessary for experiments with phenylurea and cya-nazine. Again, no di†erence in rate constants was found(Table 3).

3.1.2 E†ect of co-solvents on the rates of desorption inthe UDOS methodAs stated in Section 2.3.1, co-solvents had to be usedeither to prevent formation of solid residues of some ofthe compounds (phenylurea, cyanazine) or attain com-plete solution of permethrin during their application toCitrus CM. It was necessary to assess any e†ects theseco-solvents, particularly the involatile PEG 400 and thephospholipid (PL), may have had on the rates of per-meation. PEG 400 had no e†ects on desorption ofWL110547 from Citrus CM (Table 3) and there were noe†ects of PEG 400 or PL on either the initial rates orthe surfactant-enhanced rates of permeation ofWL110547 through Pyrus CM in comparison withapplications without the co-solvents, reported morefully in Section 3.3 (see Fig. 7).

3.1.3 Solubilisation of test compounds in surfactantmicellesSolubilisation of penetrated solutes in the surfactantmicelles is necessary, not only to maintain the concen-tration in the aqueous phase at zero, but also to avoidprecipitation of some very lipophilic and insoluble di†u-sants in the receiver solution. The lipophilic phase of10 g litre~1 PLS and 5 g litre~1 micellar surfactantsolutions (usually 0É65 ml) was generally in excess of theamount of di†used solute, with the exception of per-

TABLE 3Rate Constants of Desorption of WL110547 from Citrus CM

at Di†erent Donor Volumes and Amounts per Area

Amount per Rate constantApplied volume Areaa area k* ] 10~7 (s~1)

(kl) (cm2) (kg cm~2) (^95% CI)

5 0É13 2É65 2É4 (^1É0)10 (2 ] 5) 0É26 2É65 2É6 (^0É8)

50 0É5 6É7 1É9 (^0É8)50 0É5 1É86 2É9 (^0É5)

100 0É95 0É98 3É5 (^1É1)50 0É8 1É16 3É1 (^0É6)

(]200 kg PEG400)

a Maximum area just before complete evaporation of solvents.

methrin. In this case, owing to its low speciÐc activity,desorption from Pyrus CM with “GenapolÏ C-050 led toa di†used amount of up to 5 kg or 1É8 ] 10~8 mol per-methrin per day. This is close to the amount of sur-factant, which is estimated to be 4] 10~8 mol. Abovethe CMC the solvent power of surfactant solutionsincreases linearly with concentration, and we thereforeused a 20 g litre~1 solution of GP C-050. However, inFigs 2A and 2C (which show typical plots for popu-lations of CMs obtained in UDOS experimentsÈseeSection 3.2), rates were maximum at the shortest timeinterval (asterisks in Figs 2A and 2C), suggesting thatsolubilisation may not have been complete at the longerintervals when more permethrin had penetrated. In allother experiments with the “GenapolÏ surfactants andmodel compounds, solubilisation was not rate-limiting.35

3.2 UDOS experiments with Citrus CM

The e†ect of Genapol surfactants was investigated withall compound/surfactant combinations with Citrus CM.Typical plots of the time course of permeation of per-methrin with GP C-050 at a concentration of 20 glitre~1 in the receiving solution for populations of CMin UDOS experiments obtained with Citrus and PyrusCM are shown in Figs 2A and 2C. Initial rates ofdesorption with PLS were determined during the(kPLS* )Ðrst 65 h, in which the receiver solution contained onlyPLS. These rate constants represent solute mobilities incuticles before treatment with surfactant. On substitut-ion of the PLS solution by a receiving solution contain-ing 20 g litre~1 GP C-050, the rates increasedmarkedly, but it took between 20 and 60 h beforedesorption plots became linear again. During this timesurfactants penetrate into the CM and their concentra-tions in CM and cuticular waxes increase with timeuntil equilibrium between micellar surfactant solutionsand the CM is obtained.20 The surfactant-enhancedrates of permeation were calculated from the sub-sequent linear portion of the plots, in this case, gener-ally at the shortest interval between 112 and 119 h. Asalways observed in this type of study there was a largevariation in initial desorption rates between the CMsamples. Furthermore, as also observed previously, thegreatest rates of enhancement were with those CMsamples that gave the slowest initial rates as demon-strated in a plot of maximum e†ect (kGP Cv050* /kPLS* )against (Figs 2B and 2D), and this relationship is1/kPLS*reasonably linear. The consequence is that the sur-factant enhancement tended to reduce di†erences insolute mobilities between the CM.

CM having very low values may never reachkPLS*cuticle/surfactant sorption equilibrium, as seen in Figs.2A where the lower four desorption graphs were notlinear even after 165 h. In these cuticles permethrin


Fig. 2. (A) Time course of desorption of permethrin from Citrus CM with PLS or (indicated by the arrow) an aqueous solution of“GenapolÏ C-050. The lines represent desorption from seven individual CM. (B) Dependence of the maximum e†ect on initial rateconstants measured using PLS as desorption medium The maximum e†ect is the ratio of maximum rate constants of(kpls* ).desorption (in this case between 112 and 119 h, see asterisk) over This ratio was calculated for each CM separately. (C) andkpls* .

(D) show the analogous results for Pyrus CM (see text).

mobility and surfactant penetration were very low, suchthat sorption equilibrium and maximum surfactante†ect were not always obtained. In the response graphs(maximum e†ects versus this results in a devi-1/kPLS* )ation from linearity, the data point in brackets (Fig. 2B)lying far below the line Ðtted to the data.

Similar plots (not shown) and rates (Table 4) wereobtained for all the compound/surfactant combinations.Solute mobilities (k*) declined by a factor of 52 whenmolar volumes increased 2É64-fold (column 1, Table 4and Table 2). The mean surfactant enhancement e†ectsgenerally increased through the series of compounds (i.e.with increasing molar volumes of solutes) but decreasedwith increase in extents of ethoxylation of the sur-factants from GP C-050 to GP C-200. In fact there werenegligible enhancements by GP C-200 for all com-pounds with Citrus CM. A three-dimensional bar chartof the log[mean e†ect] for each surfactant, at a concen-

tration of 5 g litre~1 in the receiver solution, against themolar volume of the compounds clearly shows the sys-tematic trends (Fig. 3). The dependence of surfactante†ects on initial rate constants of desorption wasobtained for all compounds only with GP C-050. Coef-Ðcients of correlation were smaller with Citrus thanobserved with Pyrus CM, with the highest value ofr2\ 0É81 for permethrin (Fig. 2B).

3.3 UDOS experiments with Pyrus CM

Studies with Pyrus CM were more limited than thosewith the less permeable Citrus CM. Only WL110547was used with all “GenapolÏ surfactants and only GPC-050 with the range of compounds.

The e†ects of the “GenapolÏ surfactants withWL110547 were that solute mobilities increased after


TABLE 4Rate Constants of Test Compounds (k*) and E†ects of “GenapolÏ Surfactants on Mobilities of Test Com-

pounds in Citrus cuticular membranes

Mean e†ect (^95% CI)Genapol

k* ] 10~7 (s~1)aCompound (^95% CI) C-050 C-080 C-100 C-200

Phenylurea 24É6 (^8É5) 2É9 (^0É77) 2É13 (^0É38) 1É36 (^0É20) È2,4-D 6É76 (^2É1) 6É8 (^1É62) 2É80 (^0É40) 1É60 (^0É15) ÈCyanazine 4É37 (^1É5) 6É4 (^1É31) 1É58 (^0É61) 1É55 (^0É17) ÈWL110547 4É07 (^1É4) 15É2 (^3É8) 5É01 (^0É90) 2É91 (^0É65) ÈChlorfenvinphos 1É12 (^0É6) 16É5 (^3É2) 5É40 (^0É80) 2É10 (^0É20) ÈPermethrin 0É47 (^0É1) 82É1 (^35) 13É35 (^9É04) 2É38 (^0É29) 1É13 (^0É11)

a Arithmetic means and 95% conÐdence intervals were calculated from combined values of all experimentscarried out with the respective compound.

changing the receiver solution from PLS to solutions ofGP C-050, GP C-080 and GP C-100 (Figs 4A, B and C)but GP C-200 was barely e†ective (Fig. 4D). With GPC-050 the plots were curved over the Ðrst 48 h until therates became constant, while with GP C-080 the plotswere curved throughout the period (115 h) of the experi-ment but never reached the rates found with GP C-050.In contrast, the rates with GP C-100 were constantimmediately after changing the receiver solution butwere even slower than those with GP C-080. This wassimilar to that obtained for all compounds with CitrusCM. Enhanced rates of di†usion were reasonablyinversely related (r2\ 0É85) to initial rates of di†usion

of WL110547 with GP C-050 but not with the(1/kPLS* )other surfactants (Fig. 5). The regression line Ðtted toFig. 5A was drawn into Fig. 5B and this reveals that all

data points are below the line, that is, e†ects for a givenwere smaller with GP C-080 than for GP C-050.1/kPLS*

However, the plot of e†ects versus for the nine1/kPLS*CM in the left hand corner (corresponding to thosehaving the highest initial mobilities) is fairly linear, indi-cating that these CM reached or nearly reached sorp-tion equilibrium with GP C-080, while those CMhaving lower WL110547 mobilities did not reach sur-factant sorption equilibrium and maximum possiblee†ects. This shows that surfactant e†ects on solutemobilities in cuticles is variable and dependent on thecuticle. If surfactant permeability of CM is low, poten-tial maximum e†ects may never be realised, as with GPC-100 (Fig. 5C).

The e†ect of GP C-050 on the range of compoundswas to increase their rates of di†usion markedly, with

Fig. 3. Mean e†ects of “GenapolÏ C surfactants on mobilities of the model compounds in Citrus CM.


Fig. 4. Time course of desorption of WL110547 from Pyrus CM with PLS or (indicated by the arrow) an aqueous solution of“GenapolÏ C surfactants. The lines represent desorption from Ðve individual CM.

curvature of the plots during the Ðrst 48 h after substi-tution of the PLS receiver solution by GP C-050 solu-tion for all compounds (Fig. 6). Curvature dependedstrongly on the initial mobility in the CM (Fig. 6B),where the graph for the CM with high 2,4-D mobilityshowed an instantaneous increase in slope after chang-ing from PLS to GP C-050, while the slope for the CMwith the lowest 2,4-D mobility increased throughout theperiod. The e†ect of GP C-050 varied with of the1/kPLS*Pyrus CM for all model compounds (Fig. 7).

In order to assess the e†ect of the co-solvents neces-sary for desorption experiments with phenylurea, cya-nazine and permethrin with Citrus CM, some of thedata were obtained from applications made in theabsence and presence of either PEG 400 (phenylurea,cyanazine) or PL (permethrin). There were no discern-ible di†erences between these (Fig. 6) and thus no e†ectof these co-solvents on the rates of di†usion, either ini-tially (Fig. 6) or subsequently (Fig. 7) on the enhancedrates induced by GP C-050. These GP C-050 enhanced

rates also correlated with the initial rates of di†usion(Fig. 7), again showing that the increase was greatest forthe initially least permeable CM.

3.4 SOFU experiments with Pyrus

If WL110547 and cyanazine were applied in water or inan aqueous solution with PEG 400, the amounts whichhad penetrated the Pyrus CM 24 h after applicationwere, at best, only 15% of the amount applied. Additionof surfactants to the application solution increased therates of uptake markedly compared with the controlswithout surfactant (Fig. 8A). SOFU plots (ln [amountin the surface residue] against time) in the presence ofGP C-050 were convex (i.e. did not follow Ðrst-orderkinetics) for all amounts of GP C-050 with WL110547.Slopes increased with increasing amounts of GP C-050and, even though slopes decreased at later stages, theslopes for the two highest surfactant amounts (50 and


Fig. 5. Dependence of the e†ect of “GenapolÏ C surfactants on initial rate constants of desorption of WL110547 from Pyrus CM.

100 kg) remained higher than those with the loweramounts. As a consequence, only 45È55% of WL110547applied penetrated in the presence of 5 and 10 kg, while88È91% penetrated in the presence of 50 and 100 kgsurfactant, respectively. Similar plots with GP C-200were linear, increasing with increasing amounts of GPC-200 up to an optimum amount of 50 kg. Rates ofpenetration of WL110547 were much smaller at alltimes when applied together with GP C-200 than ifapplied together with GP C-050 (Figs 8A and B).

The more polar cyanazine penetrated somewhat moreslowly than the lipophilic WL110547 (Fig. 9), but thee†ects of the two surfactants on the rates of penetrationwere similar. The enhanced penetration of cyanazine inthe presence of GP C-050 increased with increase in theamount of GP C-050 applied up to 50 kg but, with100 kg surfactant, rates were smaller than with 50 kg(Fig. 9A). Slopes were steepest initially (up to 6 h) butbecame approximately linear thereafter. The enhance-ment of the penetration rates of cyanazine by GP C-200was small at all amounts applied, being less than with

WL110547 and GP C-200 (cf. Figs 9B, 8B). Both sets ofrates were approximately linear and varied little withthe amount of GP C-200 applied. No surface residue ofGP C-050 was visible with the lower amounts applied (5and 10 kg) after the third sample (20 h) while with GPC-200 residues could be seen throughout the entireexperiment.

4 DISCUSSION

4.1 Solute and surfactant Ñows in UDOS and SOFUexperiments

Unilateral desorption from the outer surface (UDOS)provides three sets of information : (i) e†ects of molarvolumes of solutes and temperature on mobility in CM,(ii) e†ects of surfactants on solute mobility in CM and(iii) velocity of penetration of surfactants into CM asseen from the time required to realise maximum sur-factant e†ects. In UDOS experiments the majority of


Fig. 6. Time course of desorption of the model compounds from Pyrus CM with PLS or (indicated by the arrow) an aqueoussolution of Genapol C-050. The lines represent desorption from four to eight individual CM (see text).

radio-labelled solute molecules are initially contained inthe sorption compartment of the cuticles17,22 (Fig. 10),because the sorption compartment amounts to about90% of the total volume of the CM37 and apparent par-tition coefficients of organics in cutin are at least 10

times higher than in cuticular waxes.23,24,39 When thereceiver solution is brought into contact with the outersurface of the CM, solutes start to di†use from the sorp-tion compartment across the waxy limiting skin into thereceiver solution. Rates of di†usion are determined by


Fig. 7. Dependence of the e†ect of “GenapolÏ C-050 on initial rate constants of desorption of the model compounds from PyrusCM.

solute mobility in the waxes of the limiting skin. Esti-mating solute mobility in the limiting skin requires thatdesorption media do not contain compounds whichpenetrate the CM and change the structure of cuticularwaxes. If the desorption media contain relatively small

and lipophilic surfactants which can penetrate thecuticle and are sorbed in cutin and cuticular waxes, thecuticles will be loaded with surfactants until sorptionequilibrium between micellar solutions and CM isobtained.20h25 This loading with surfactants proceeds


Fig. 8. Simulation of foliar uptake of WL110547 across PyrusCM as inÑuenced by di†erent amounts of the surfactants (A)

“GenapolÏ C-050 and (B) C-200 (means of eight to 14 CM).

from large volumes of micellar surfactant solutionswhich are repeatedly exchanged with fresh ones whensamples are taken. Surfactants cannot escape from theCM as there is no solution in contact with the innersurface of the CM (Fig. 10). The time required for equi-librium depends on the mobility of the surfactant in thelimiting layer of the CM, that is on the molar volume ofsurfactant molecules, just as with other molecules.13,22However, accelerators like or chlorfenvinphosC12E8are much more mobile than other compounds of com-parable size.26 The concentrations of ethoxylated alco-hols sorbed in cuticles when in equilibrium withmicellar solutions decreased with increasing numbers ofcarbon atoms (nC) in the alcohol and the number ofethoxy groups (nE)25

log Csurf \ 2.97[ 0.02nC[ 0.08nE (3)

where the surfactant concentration has the(Csurf)dimension mmol kg~1. This equation has been derivedusing monodisperse ethoxylated alcohols. Sorption ofpolydisperse surfactants in cuticles can be obtained bycalculating sorption for each homologue separately andsumming up. This has been done for the polydisperseGenapol surfactants using C \ 12 and the number of

Fig. 9. Simulation of foliar uptake of cyanazine across PyrusCM as inÑuenced by di†erent amounts of the surfactants (A)

“GenapolÏ C-050 and (B) C-200 (means of six to 13 CM).

ethoxy groups given in Fig. 1A. In Fig. 1B volume frac-tions of surfactant homologues have been plotted foreach polydisperse mixture. The total concentration of

in the cuticle at equilibrium corresponds to theC12Exarea under each curve and is given in parenthesis. Theactual volume fraction is even greater since the

other homologues are not included in theC14Èandvalues given in Fig. 1B, but the di†erences between theGenapols still exist. As predicted by eqn (3) the amountssorbed decrease with increasing ethoxylation both forthe individual homologues and the entire mixtures.With increasing ethoxylation the composition of thesurfactants sorbed in the cuticles changes towards morepolar homologues and the total amount sorbeddecreases, such that 3É44 times more GP C-050 wassorbed than GP C-200. Still, a signiÐcant amount ofsurfactant would be sorbed in cuticles at equilibriumeven from GP C-200 micellar solutions.

It is an important feature of UDOS experiments that,given enough time, sorption equilibrium between sur-factant micelles and cuticles is obtained, such thatsurfactant-enhanced solute mobilities can be related tostationary surfactant concentrations in the cuticle andin cuticular waxes.


Fig. 10. Schematic drawing of the cuticle and the directions of solute and surfactant Ñow in (A) UDOS and (B) SOFU(from Ref. 13).

The data of Fig. 1B refer to sorption in cuticles, andmost of the surfactants will be associated with the cutinof the sorption compartment. By comparing partitioncoefficients of monodisperse alcohol ethoxylatesobtained for the systems isolated wax/water andpolymer matrix/ water (polymer matrix refers to cuticlesfree of waxes) it was found that apparent surfactantsolubility in waxes was only one-tenth of that in thepolymer matrix.24 If sorbed compounds were homoge-neously distributed, the amount of surfactant in the lim-iting skin of the cuticles used in the present study waslower than those shown in Fig. 1B under UDOS condi-tions, i.e. under equilibrium between cuticle andaqueous solution.

In SOFU experiments, discrete amounts of radio-labelled solutes and surfactants are simultaneouslyapplied to the outer surface of the CM. The waterevaporates quickly leaving a solution of solute in neathydrated surfactant. From this residue, surfactanthomologues and radio-labelled solutes di†use into andthrough the CM, each at its individual velocity (Fig.10B). Eventually, they are collected in the receiver solu-tion and are removed from the system when PLS iswithdrawn during sampling. This process leads todecreasing amounts of both solutes and surfactants inthe surface residue and to transient concentration gra-dients in the CM.

4.2 VeriÐcation of experimental procedures

The donor solutions contained di†erent amounts ofethanol and therefore the area treated varied in UDOSexperiments. However, this has no inÑuence on the rate

constants of desorption (k*) if the di†usant is quantitat-ively sorbed in the cuticle during droplet drying. Therate constant of desorption is given by

k* \P*AVdon

(4)

where P* is a permeability (P) parameter independentof the partition coefficient K (P\ P*K) and A and Vdonare the area and the underlying volume of the donorcompartment of the cuticle, respectively.17 The ratio of

and A corresponds to the thickness of the sorptionVdoncompartment of the cuticle (Fig. 10) and in UDOSexperiments this is constant. Thus, rate constants (k*)should not be a†ected by the area covered by the donordroplet. This was experimentally conÐrmed (Table 3).

With phenylurea and cyanazine, PEG 400 was neces-sary to obtain rate constants of permeation by solu-bilising and improving sorption of these compoundsinto Citrus CM during evaporation of water of thedonor solutions. If these compounds were applied inwater only, rate constants were extremely low and couldnot be increased substantially even by very e†ective sur-factants (data not shown). With Pyrus CM, sorptionfrom aqueous donor solutions without PEG 400 wascomplete as indicated by desorption of up to 90% of theamount applied of these compounds with GP C-050(Fig. 6). When PEG 400 was used on Pyrus CM nodi†erences in the rate constants or surfactant e†ectswere found, indicating that PEG 400 acted mainly onthe sorption process, possibly by delaying crystallisationand prolonging the time for sorption into the innersurface of the CM. Furthermore, PEG 400 did not


Fig. 11. (A) Dependence of the maximum e†ect of “GenapolÏC-050 on initial rate constants of permethrin in Citrus(kpls* )and Pyrus CM (data from Fig. 2B/D). (B) Dependence of themaximum e†ect of Genapol C-050 on initial rate constants

of WL110547 in Pyrus CM of di†erent origin.(kpls* )

change the permeabilities of the CM as shown by thelack of any e†ect on the mobility of phenylurea andcyanazine with Pyrus leaf CM (Figs 7A, 7C). Probablymost of the compound was sorbed in the CM and notdissolved in PEG 400 on its inner surface, otherwise therates would have been lower owing to an increase in thedonor volume. It is theoretically possible that it acts inthat way and simultaneously increases rate constants.Thus the two e†ects may cancel each other, therebymaking it impossible to Ðnd any inÑuence in the controlexperiments with phenylurea and cyanazine with PyrusCM. However, PEG 400 contains more than 100%water under saturation humidity at the loading side andthis may support our assumption that most of the com-pound was actually sorbed into the CM.

4.3 E†ects of Genapol surfactants on mobilities ofsolutes in CM

The results from the UDOS experiments with Citrus

and Pyrus CM show that the lower ethoxylated ana-logues of these alcohol ethoxylates greatly increasedsolute mobilities in these isolated CM (Figs 2 and 3).GP C-050 substantially increased mobilities of all com-pounds and its e†ect was larger than that of GP C-080and GP C-100, while GP C-200 was nearly inert (Figs3È5). Accelerating activity of ethoxylated alcohols isproportional to the amount sorbed in cuticles andwaxes.24h26 In UDOS experiments the numbers ofcarbon atoms and ethoxy groups inÑuence the amountsorbed (eqn (3)) and hence their intrinsic acceleratoractivity. The results of this study also show that theamounts of surfactant sorbed in cuticles decreased inthe same order as their general e†ects on solute mobil-ity, i.e. GP C-050[ C-080 \ C-100 [ C-200 (Fig. 1B).

However, there is one small discrepancy in that prac-tically the same amounts of GP C-080 and C-100 arecalculated to be sorbed at equilibrium, while GP C-080was clearly more e†ective as an accelerator than GPC-100 (Fig. 3). Possibly, the actual composition of theGP C-100 lot used di†ered from the analytical data pro-vided by Hoechst, or equilibrium sorption was notobtained with all CM, such that actual surfactant con-centrations in the cuticles were smaller than those cal-culated. Data generated using Pyrus CM (Fig. 5) whichare more permeable than Citrus CM, indicate that equi-librium might not have been obtained with all CM.With the even larger GP C-200 this problem would beeven greater. From the data by Riederer et al.25 cuticle/water partition coefficients of surfactant homologuescan be calculated and a value of 3É54 for a surfactanthaving the formula is obtained. Thus, even theC12E23most polar homologue in our series was lipophilic andmore soluble in cuticles than in water. Failure of GPC-200 to increase solute mobility in cuticles could bedue to the too low volume fraction sorbed in equi-librium. Data in Fig. 1B indicate that a signiÐcantamount of surfactant (volume fraction [0É018) shouldbe sorbed at equilibrium with GP C-200, yet this sur-factant was nearly inert in this study (Figs 3È5). This isin accordance with the results obtained with mono-disperse surfactants, since extrapolation of the e†ect/volume fraction relationship for monodisperse fattyalcohol ethoxylates to low volume fractions results inan e†ect of 1 (i.e. no increase at a volume fraction ofapproximately 0É03).25

4.4 Acceleration as a†ected by initial mobilities incuticles

With all model compounds studied, the e†ect of GPC-050 varied with the initial rate of Pyrus CM(1/kPLS* )(Fig. 7) and also, though less pronounced, with CitrusCM (data not shown). Initial mobilities and 1/k*PLSvalues of the CM of CM populations used in the experi-ments di†ered considerably (cf. Fig. 7).


E†ects of GP C-050 depended on initial mobilities incuticles. The lower the initial mobility, the greater wasthe accelerator e†ect. This holds not only for a givenspecies, but is valid also for a given compound and cut-icles from di†erent species (Fig. 11A) or from the samespecies coming from di†erent years or locations (Fig.11B). In Fig. 11A the e†ect of GP C-050 on mobility ofpermethrin is plotted against for both Citrus and1/kPLS*Pyrus CM. The line drawn Ðts the combined data rea-sonably well (r2\ 0É88). A similar result was obtainedfor di†erent Pyrus CM (cv. “BartlettÏ) which were iso-lated from leaves of trees growing in a growth chamber(1990) or in an orchard (1991 and 1992), respectively.E†ects of GP C-050 on mobility of WL110547 di†eredamong samples, but a straight line could be Ðtted to thecombined data (Fig. 11B).

4.5 Acceleration as a†ected by molar volume of solutes

In Fig. 12 mean rates of desorption (log k*) with PLS(empty dots) for the model compounds (combinedvalues from all experiments obtained with the respectivecompound and species) are plotted against molarvolume With both Citrus and Pyrus CM, rate con-(Vx).stants of desorption depended exponentially on molarvolume of the model compounds, as indicated by theVxlinearity of the plot of log k* against (empty dots inVxFigs 12A and B). At high volume fractions, chlorfen-vinphos and (to a lesser degree) permethrin, both lipo-philic liquids, act as self-accelerators. While the intrinsicmobility (volume fraction] 0) for chlorfenvinphos inPyrus CM was obtained recently,26 such a value forpermethrin is not yet known. This is the reason why thevalue for permethrin, which had to be applied in thehighest amount (Section 2.3.1) with Pyrus CM (Fig.12B), is larger than predicted from its size.

Accelerator surfactants increased mobilities of largesolutes much more than those of smaller ones (Fig. 3).For both species the slope of the line of a plot log k*versus molar volume decreased substantially if GPC-050 was used as desorption medium instead of inertPLS (Ðlled dots in Figs 12A and B). In Fig. 12, log k*values for desorption with GP C-050 (Ðlled dots) withCitrus were calculated by multiplying initial rate con-stants by the mean e†ects for each compound (Table 4)while for Pyrus CM initial rate constants were multi-plied by the more correct (Section 4.3) e†ects at themean value of (Fig. 7).kPLS*

In this particular set of compounds, size (molarvolume) and the lipophilicity (log are correlatedKow)(r2\ 0É83) and, since there would also be a reasonablecorrelation for a plot of log k* against log plottingKow ,e†ects against molar volumes may appear arbitrary(Fig. 3). However, this possibility can be clearlyexcluded, since the existence of a correlation with lipo-philicity (Table 2) would mean that the rate constants

Fig. 12. Dependence of rate constants of desorption of themodel compounds on their molar volumes without(Vx) (L)and with “GenapolÏ C-050 for (A) Citrus and (B) Pyrus(…)cuticular membranes. Each point represents the mean of 17 to

89 CM. For compound identiÐcation see Table 2.

increase from phenylurea to permethrin, while theopposite was observed (Fig. 12). Rate constants as mea-sured using UDOS are proportional to di†usion coeffi-cients, which are generally independent of lipophilicitybut decrease rapidly with the size of di†usants.39 Thishas been shown for di†usion of a large number of com-pounds in CM, reconstituted cuticular waxes, humanskin and synthetic polymers.13,22,24,39h43 Due to highviscosity, the size selectivity of a solid matrix is muchmore pronounced than that in liquids.40 Alcohol eth-oxylates decrease the viscosity of cuticular waxes andincrease their Ñuidity, i.e. they become more like aliquid, which results in a reduction of their size selec-tivity. This is the reason why mobilities of large mol-ecules are more a†ected by accelerators such that e†ectsincrease with solute size and di†erences between com-pounds disappear (Fig. 12).


4.6 E†ects of surfactants on rates of penetration

In SOFU experiments, rates of penetration (J,mol m~2 s~1) are measured, which depend on the per-meance (P) of cuticles and on the concentration di†er-ence across cuticles :13

J \ P*C (5)

Permeance is a composite quantity which depends onsolute mobility (D) in cuticles, a partition coefficient (K)as measure of di†erential solubility and the thickness ofthe CM (*xCM) :

P\ DK*xCM

(6)

D is related to UDOS rate constants (k*)

D\ k* *xls*xsoco (7)

where and are the thickness of the limiting*xls *xsocoskin and the sorption compartment, respectively, asshown in Fig. 10. Combining eqns (6) and (7) and sub-stituting into eqn (5) yields

J \ k* *xls*xsoco K*xCM

*C (8)

which can be further simpliÐed since the thickness of thesorption compartment and the total thickness of theCM are similar :

J \ k* *xlsK *C (9)

Equation (9) shows that rates of penetration (J) are pro-portional to solute mobility (k*) di†erential solubility(K) and the concentration di†erence across the cuticle.The thickness of the limiting skin is constant for eachtype of CM and needs no further consideration. If thedonor compartment (surfactant residue) is constant,penetration can be treated as a desorption experimentand linear graphs obtained.

We have shown above that all surfactants, except GPC-200, signiÐcantly increase solute mobility k* and, fora given type of CM and solute, this increase dependsonly on surfactant concentration in the CM. In UDOS,surfactant concentrations increase with time and even-tually become constant, if surfactants penetrate intoCM fast enough relative to the duration of the experi-ment. In SOFU experiments, surfactant concentrationsalso increase with time, but only with slowly penetrat-ing surfactants do their concentrations become station-ary within the cuticle, since only a Ðnite amount ofsurfactant is applied. Fast-penetrating surfactants

di†use into the receiver and are removed from thesystem. Therefore, mobilities of solutes in the presenceof accelerators are often not constant in SOFU, ratherthey increase initially and eventually decrease as thesupply of surfactant in the surface residue and in theCM become depleted.

The term K *C can be considered as the driving forceof penetration13 and, with SOFU experiments, the termcan be simpliÐed because the solute concentration inthe water of the receiver is held practically zero owingto the large volume of the receiver and the sorption oflipophilic solutes in PLS vesicles. Thus, the drivingforce depends only on the concentration in the surfaceresidue, which serves as a donor once the bulk waterhas evaporated, and we can write the driving force as

The solute concentration in the surface residueKCdon .changes with time, since both solutes and surfactantspenetrate into cuticles. These changes cannot be quanti-Ðed at present, since we only measured the permeationof the solutes and not the surfactants.

The partition coefficient (K) represents the equi-librium concentration of solutes between the limitingskin and the surface residue. This quantity is not knowneither, and it may not be constant, since composition ofthe surfactant residue will change with time because thesmaller and more lipophilic homologues are likely topenetrate faster and leave the larger and more polarhomologues behind.

Owing to these complexities and unknown variables,we can discuss SOFU results only qualitatively. Ratesof penetration of WL110547 followed approximatelyÐrst-order kinetics, both in the absence and presence ofGP C-200. That is, the concentration of this lipophilicsolute in the surfactant residue on top of the CMdecreased exponentially with time. Rate constants (k)were higher in the presence of GP C-200 than in thecontrol and increased with increasing amount of sur-factant up to 50 kg (Fig. 8B). Observed rate constantsapproximately doubled on increasing GP C-200 from5 kg to 50È100 kg. GP C-200 may increase penetrationby dissolving solid WL110547 on the cuticle surface,thereby increasing the driving force, but the increase ofrates with concentration indicated that some homo-logues may also have penetrated into the limiting skinand acted as plasticisers while most remained on thesurface. The e†ect of this surfactant in this SOFUexperiment is greater than that in UDOS (Figs 5D and6D) because penetration from the neat surfactantresidue is faster than from the micellar solution anddepends only on solubility in the cuticle and not on therelative solubility, i.e. the partition coefficient. The factthat acceleration did not increase with increase in appli-cation of GP C-200 from 50 to 100 kg is most probablydue to dilution of the solute in the surfactant, or interms of eqn (9), the decrease in on doubling theCdonamount of surfactant counteracted the surfactant e†ecton mobility (k*). We have shown before that, in the


presence of surfactants which do not penetrate cuticles(such as “TweenÏ 80), rates of uptake decrease withincreasing amounts of surfactant.13

GP C-050 increased the rates of penetration ofWL110547 much more than did GP C-200 though therates of penetration were not Ðrst-order throughout theperiod of penetration with the highest rates in the 0 to6-h interval (Fig. 8A). This relatively smaller and morelipophilic surfactant penetrates cuticles rapidly andtherefore achieves maximum concentrations in the lim-iting skin and maximum e†ects on solute mobilityquickly. Surfactant e†ects increased with increasingamounts of surfactant, though it can be seen that thee†ect of adding 100 kg of GP C-050 initially produces aslower rate than 50 kg. This may be accounted for bythe competition between sufficient amounts of GPC-050 for modifying the cuticle, thus acceleratingWL110547 penetration, and excess amounts acting todilute WL110547 in the reservoir, thus reducing concen-tration and the driving force. At later intervals thelarger reservoir of 100 kg GP C-050 will eventuallyreduce, due to penetration, thereby increasing thedriving force and eventual penetration of WL110547.

Similar results were obtained with the smaller andmore polar solute, cyanazine (Fig. 9). GP C-050 greatlyincreased rate constants (maximum 1É4 ] 10~5 s~1)though the dilution e†ect with 100 kg prevailedthroughout the experiment, over 25 h. As withWL110547, GP C-200 increased the rate constants ofcyanazine penetration, but the e†ects of di†erentamounts of the surfactant were small and not signiÐ-cant.

So far the e†ect of the partition coefficient (K) has notbeen considered. Equation (9) suggests that, at constantdriving force, rates of penetration should be higher forlipophilic than for polar solutes. Rate constants of pen-etration in the absence of surfactants are not very di†er-ent with values of 2É2(^0É7) ] 10~6 s~1 for cyanazineand 1É5(^0É75) ] 10~6 s~1 for WL110547 nor aremobilities (k*) with values of 1É75 ] 10~6 s~1(cyanazine) and 1É26 ] 10~6 s~1 (WL110547). Kowvalues for cyanazine and WL110547 are 126 and 3982,respectively, which implies that WL110547 is about 30times more lipophilic than cyanazine (Table 2) and pen-etration should be better with this compound. However,in SOFU experiments, solutes are not dissolved inwater but in neat hydrated surfactant, and the cuticle/surfactant partition coefficients are not known. Owingto the good solvent power of surfactants and manyother adjuvants, especially for lipophilic solutes,35 thesepartition coefficients are much smaller, and di†erencesbetween compounds may also be reduced. For instance,chlorfenvinphos has a cuticle/water partition coefficientof more than 1000, while the partition coefficientsbetween cuticle and PEG 400 or GP C-050 are smallerthan 1, i.e. chlorfenvinphos is more soluble in theseadjuvants.27

No surface residue of GP C-050 was visible at thelower amounts applied (5 and 10 kg) after the thirdsample (20 h) and therefore both the e†ect on solutemobility and on the physical state (dissolution) of thecompounds in the surfactant change. A comparison ofrates of penetration between cyanazine and WL110547in the presence of surfactant must therefore be limitedto the higher amounts of surfactants and to the timeperiod up to 6 h. The values for cyanazine andWL110547 di†er by a factor of 2É5 (1É4 and3É5 ] 10~6 s~1). Since the applied concentrations didnot di†er and UDOS experiments showed that in thepresence of GP C-050 mobilities of di†erent compoundsare similar (Fig. 12), partition coefficients are expectedto be quite similar.

4.7 Comparing these results to previous studies on foliaruptake

We have shown that “GenapolÏ surfactants a†ect bothsolute mobilities and driving forces. “GenapolÏ C-050was sorbed in CM and penetrated cuticles most rapidly.It had the highest e†ects on solute mobilities for allsolutes in the current range, with e†ects increasing withincreasing size of the solutes. Increasing degrees of eth-oxylation of the surfactants reduced sorption fromaqueous micellar solution into cuticles and their e†ectson solute mobilities.

Stock et al.9 measured uptake into the intact leaves ofdi†erent plant species, while, in this study, isolatedastomatous cuticles were used. In fact, we compared therates of penetration of solutes and cuticular transpira-tion of isolated and non-isolated cuticles and we foundslight di†erences (Baur35 and unpublished results).However, for cuticular transpiration the coefficients ofvariation are large, above 30%, for studies both withisolated cuticles and where intact leaves of many di†er-ent plants are used, and the range of values overlap45,46and coefficients of variation are even higher for organicsolute mobility. Permeabilities of isolated CM tend todecrease somewhat during storage, possibly due to lossof volatile constituents which will modify the cuticle21and/or rearrangement of the cuticular waxes.29 Suchchanges can a†ect absolute rates, though there is noreason to expect that the di†erential e†ects of molarvolumes and lipophilicities of solutes and the e†ects ofsurfactants on penetration would be a†ected. The bene-Ðcial use of isolated cuticles for mechanistic studies hasbeen discussed and underlined repeatedly.47h49 Themain drawback against direct comparison of the resultsin this work to the results of studying the patterns ofuptake into living leaves is that enzymatically isolatedcuticles can only be obtained from a limited number ofspecies that do not have amphistomatic leaves.

This work was undertaken with the intention ofobtaining a better understanding of the mechanisms of


adjuvant action underlying the observed pattern thatpenetration of lipophilic compounds was better assistedby surfactants with low ethylene oxide contents andthat of hydrophilic compounds by surfactants with highethylene oxide content.8 The same range of compoundswas used, though, in this paper, the results are givenonly for the lipophilic compounds and a chemicallysimilar, though not identical, range of surfactants. Thesubsequent discussion will be limited to the action withthese lipophilic compounds, i.e. with log values [0.KowBoth studies with leaves and isolated cuticles haveshown that fatty alcohol surfactants can penetrate cut-icles, and those with low E content penetrate faster thanthose with high.12,20 It can be seen that, in the contextof this study, surfactants in deposits on leaf surfaces canhave two major functions. Firstly, they can dissolve thepenetrant, and that dissolution a†ects its concentrationand partition coefficient and therefore its driving forcefor permeation. Secondly, after penetration into thecuticle they can, by a plasticisation mechanism, reducethe resistance to di†usion in the regions in the cuticlethrough which the penetrant will di†use.26 A possiblereduction in the degree of crystallinity of waxes has notyet been shown. In deposits on leaf surfaces (or isolatedcuticles in the SOFU experiments of this work) bothfunctions are kinetic in the sense that, since the sur-factants will begin penetrating into the cuticle duringand after evaporation of the carrier solvents, both thedriving forces and the cuticular resistances are beingmodiÐed. These modiÐcations will vary throughout thewhole sequence of events and will be dependent on thestructure of both the surfactant (ethylene oxide contentin this work) and the penetrant.

The results of this work have reinforced thesenotions. From the UDOS experiments it is clear thatthe lowest ethoxylated surfactant, GP C-050, was themost e†ective at reducing the resistance of the CM andinducing accelerator action, and that accelerating actionacross the Citrus cuticular membrane (Fig. 3) was great-est for the largest compound, permethrin, and least forthe smallest compound, phenylurea. In a similar studywith Pyrus CM and WL110547 and cyanazine, a similarrelationship was observed. For the hydrophilic sur-factant with the most ethylene oxide units, GP C-200,there was little or no accelerating action and it can beconcluded that there was little or no partitioning of itinto the CM, at least from the aqueous solutionsemployed in the membrane cells.

In the SOFU studies with the more permeable PyrusCMs, it was again shown that GP C-050 was moree†ective than GP C-200 with both a larger, lipophiliccompound, WL110547, and a smaller, less lipophiliccompound, cyanazine. Interestingly, in these SOFUstudies, although GP C-200 was less e†ective than GPC-050, it did have signiÐcant accelerating action forboth WL110547 and cyanazine which it did not in theUDOS studies. GP C-200 improved penetration in con-

trast to PEG 400 and this suggests that it is taken up insigniÐcant amounts from neat surfactant residues indu-cing some accelerator e†ect. This is indicated also bythe fact that the compositions of GP C-050 and GPC-200 have more than 6% identical constituents (Fig.1A). These results are in good accord with thoseobserved by Stock et al.9 in their work on leavesattached to plants and reinforce the conclusion that thepenetration of lipophilic, larger compounds is bestaided by surfactants with low ethylene oxide contents.The phenomenon is probably best rationalised in termsof molar volume, since rates of di†usion are inverselyrelated to a molar volume term and any quantiÐcationin terms of partitioning or solubility will, in any case, bedifficult owing to the transient, varying nature of thecuticular wax/surfactant composition during the pen-etration process. Furthermore, lipophilic solutes willhave higher permeabilities since their solubilities in thecuticle are likely to be higher providing crystallisation/solidiÐcation (and hence reduction e†ects on the drivingforce) is avoided.

Further mechanistic aspects of surfactantsÏ mode ofaction in foliar uptake and studies on the mechanismsfor penetration of hydrophilic compounds will be thesubject of a forthcoming paper.

ACKNOWLEDGEMENTS

We are grateful to Mr B. Speight and Dr M. Aven forhelpful technical discussions and to Shell ForschungAG, Schwabenheim for Ðnancial assistance to one of us(P.B.).

REFERENCES

1. McWorther, C. G., The use of adjuvants. In Adjuvants forherbicides, ed. R. H. Hodgson. Weed Science Society ofAmerica, Champaign, Illinois, 1982.

2. Van Valkenburg, J. W., Terminology, classiÐcation andchemistry. In Adjuvants for Herbicides, ed. R. H. Hodgson.Weed Science Society of America, Champaign, Illinois,1982.

3. Stevens, P. J. G., Gaskin, R. E., Hong, S. & Zabkiewicz, J.A., Contributions of stomatal inÐltration and cuticularpenetration to enhancements of foliar uptake by sur-factants. Pestic. Sci., 33 (1991) 371È82.

4. Gaskin, R. E. & Holloway, P. J., Some physicochemicalfactors inÑuencing foliar uptake enhancement ofglyphosate-mono(isopropylammonium) by polyoxyethy-lene surfactants. Pestic. Sci., 34 (1992) 195È206.

5. Holloway, P. J. & Edgerton, B. M., E†ects of formulationwith di†erent adjuvants on foliar uptake of difenzoquatand 2,4-D: model experiments with wild oat and Ðeldbean. W eed Res., 32 (1992) 183È95.

6. Holloway, P. J., Wong, W. W.-C., Partridge, H. J.,Seaman, D. & Perry, R. B., E†ects of some nonionic poly-oxyethylene surfactants on uptake of ethirimol and diclo-butrazol from suspension formulations applied to wheatleaves. Pestic. Sci., 34 (1992) 109È18.


7. Knoche, M. & Bukovac, M. J., Studies on octylphenoxysurfactants : XI. E†ect on NAA di†usion through the iso-lated tomato fruit cuticular membrane. Pestic. Sci., 38(1993) 211È17.

8. Stock, D., Edgerton, B. M., Gaskin, R. E. & Holloway, P.J., Surfactant-enhanced foliar uptake of some organiccompounds : Interactions with two model polyoxyethylenealiphatic alcohols. Pestic. Sci., 34 (1992) 233È242.

9. Stock, D., Holloway, P. J., Grayson, B. T. & Whitehouse,P., Development of a predictive uptake model to rational-ise selection of polyoxyethylene surfactant adjuvants forfoliage-applied agrochemicals. Pestic. Sci., 37 (1993) 233È45.

10. Kirkwood, R. C., Knight, H., McKay, I. & Chandrasena,J. P. N. R., The e†ect of a range of nonylphenol sur-factants on cuticle penetration, absorption, and trans-location of water-soluble herbicides. In Adjuvants forAgrochemicals. Proc. 2nd Internat. Symp. Adjuvants forAgrochemicals (1989), ed. C. L. Foy. CRC Press, BocaRaton, FL, 1992.

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

12. Stock, D. & Holloway, P. J., Possible mechanisms forsurfactant-induced foliar uptake of agrochemicals. Pestic.Sci., 38 (1993) 165È77.

13. Scho� nherr, J. & Baur, P., Modelling penetration of plantcuticles by crop protection agents and e†ects of adjuvantson their rates of penetration. Pestic. Sci., 42 (1994) 185È208.

14. Geyer, U. & Scho� nherr, J., In-vitro test for e†ects of sur-factants and formulations on permeability of plant cut-icles.In Pesticide Formulations : Innovations andDevelopments, ed. B. Cross & H. B. Scher. ACS Sympo-sium Series 371, Washington DC, 1988, pp. 22È3.

15. Riederer, M. & Scho� nherr, J., E†ects of surfactants onwater permeability of isolated plant cuticles and on thecomposition of their cuticular waxes. Pestic. Sci., 29 (1990)85È94.

16. Scho� nherr, J., Riederer, M., Schreiber, L. & Bauer, H.,Foliar uptake of pesticides and its activation by adju-vants : Theories and methods of optimization. In PesticideChemistry : Advances in International Research Develop-ment and L egislation. ed. H. Frehse. VCH Weinheim, 1991.

17. Scho� nherr, J. & Bauer, H., Analysis of e†ects of sur-factants on permeability of plant cuticles. In Adjuvants forAgrochemicals. Proc. 2nd Internat. Symp. Adjuvants forAgrochemicals (1989), ed. C. L. Foy. CRC Press, BocaRaton, FL, 1992.

18. Tan, S. Y. & Crabtree, G. D., E†ects of nonionic sur-factants on cuticular sorption and penetration of 2,4-dich-lorphenoxy acetic acid. Pestic. Sci., 35 (1992) 299È303.

19. Chamel, A., Gambonnet, B. & Coret, J., E†ects of twoethoxylated nonylphenols on sorption and penetration ofC14-isoproturon through isolated plant cuticles. PlantPhysiol. Biochem., 30 (1992) 713È21.

20. Scho� nherr, J., E†ects of monodisperse alcohol ethoxylateson mobility of 2,4-D in isolated cuticles. Pestic. Sci., 38(1993) 155È64.

21. Scho� nherr, J., E†ects of alcohols, glycols and mono-disperse ethoxylated alcohols on mobility of 2,4-D in iso-lated plant cuticles. Pestic. Sci., 39 (1993) 213È23.

22. Bauer, H. & Scho� nherr, J., Determination of mobilities oforganic compounds in plant cuticles and correlation withmolar volumes. Pestic. Sci., 35 (1992) 1È11.

23. Schreiber, L., A mechanistic approach towards surfactant/wax interactions : E†ects of octaethyleneglycolmonodo-decylether on sorption and di†usion of organic chemicals

in reconstituted cuticular wax of barley leaves. Pestic. Sci.,45 (1995) 1È11.

24. Schreiber, L., Riederer, M., Schorn, K. & Marsh, D.,Mobilities of organic compounds in reconstituted cuticu-lar wax of barley leaves : E†ects of monodisperse alcoholethoxylates on di†usion coefficients. Pestic. Sci., in press.

25. Riederer, M., Burghardt, M., Mayer, S., Obermeier, H. &Scho� nherr, J., Sorption of monodisperse alcohol eth-oxylates and their e†ects on the mobility of 2,4-D in iso-lated plant cuticles. J. Agric. Food Chem., 43 (1995)1067È75.

26. Baur, P., Grayson, B. T. & Scho� nherr, J., Concentration-dependent mobility of chlorfenvinphos in isolated plantcuticles. Pestic. Sci., 47 (1996) 171È80.

27. Baur, P. & Scho� nherr, J., Die Aufnahme systemischerWirksto†e u� ber Bla� tter : Grundlagen und Optimierung.Gartenbauwissenschaften, 61 (1996) 105È15.

28. Scho� nherr, J. & Riederer, M., Plant cuticles sorb lipophiliccompounds during enzymatic isolation. Plant CellEnviron., 9 (1986) 459È66.

29. Geyer, U. & Scho� nherr, J., The e†ect of environment onpermeability and composition of Citrus leaf cuticles :water permeability of isolated cuticular membranes.Planta, 180 (1990) 147È54.

30. Behler, A. & Ploog, U., Fettalkoholethoxylate miteingeengter HomologenverteilungÈEntwicklung neuerAlkoxylierungskatalysatoren. Fat. Sci. T echnol., 92 (1990)49È54.

31. Zellmer, S., Die Entwicklung thermolabiler, fusionsfa� higerLiposomen fu� r die Tumortherapie und von Trans-ferosomen fu� r den Einsatz an PÑanzen. Doctoral Disser-tation, Technical University of Munich, 1992.

32. Kennedy, C. D. & Harvey, J. M., Plant growth substancesaction on lecithin and lecithin/cholesterin vesicles. Pestic.Sci., 3 (1972) 715È27.

33. Tanford, C., T he hydrophobic e†ect : Formation of micellesand biological membranes. John Wiley & Sons, New York,1973.

34. Rippen, G., Handbuch der Umweltchemikalien, 2nd edn.Ecomed, Landsberg/Lech, 1992.

35. Baur, P., Barriereeigenschaften pÑanzlicher Kutikular-membranen und ihre BeeinÑussung durch organischeXenobiotika, Doctoral Dissertation, Technical Universityof Munich, 1993.

36. Edwards, D. A., Liu, Z. & Luthy, R. G., Interactionsbetween nonionic surfactant monomers hydrophobicorganic compounds and soil. W at. Sci. T ech., 26 (1992)147È58.

37. Scho� nherr, J. & Riederer, M., Foliar penetration andaccumulation of organic chemicals in plant cuticles. Rev.Environ. Contamin. T oxicol., 108 (1989) 1È70.

38. Schreiber, L. & Scho� nherr, J., Analysis of foliar uptake ofpesticides in barley leaves : Role of epicuticular waxes andcompartmentation. Pestic. Sci., 36 (1992) 213È21.

39. Baur, P., Marzouk, H., Scho� nherr, J. & Bauer, H., Mobi-lities of organic compounds in plant cuticles as a†ected bystructure and molar volumes of chemicals and plantspecies. Planta, 199 (1996) 404È12.

40. Potts, R. O. & Guy, R. H., Predicting skin permeability.Pharmaceutical Res., 9 (1992) 663È9.

41. Crank, J. & Park, G. S. (eds), Di†usion in polymers. Aca-demic Press, London, 1968.

42. Cussler, E. L., Di†usion Mass transfer in Ñuid systems.Cambridge Univ. Press, Cambridge, 1984.

43. Schreiber, L., Kirsch, T. & Riederer, M., Transportproperties of cuticular waxes of Fagus sylvatica L andPicea abies (L.) Karst. : Estimation of size selectivity andtortuosity form di†usion coefficients of aliphatic mol-


ecules. Planta, 198 (1996) 104È9.44. Abraham, M. H. & McGowan, J. C., The use of character-

istic volumes to measure cavity terms in reversed phaseliquid chromatography. Chromatographia, 23 (1987) 243È6.

45. Kerstiens, G., Cuticular water permeance of europeantrees and shrubs grown in polluted and unpolluted atmo-spheres, and its relation to stomatal response to humidityin beech (Fagus sylvatica L.). New Phytol., 129 (1995) 495È503.

46. Baur, P., Lognormal distribution of water permeability

and organic solute mobility in plant cuticles. Plant, Celland Environment, 20 (1997) 167È77.

47. Hull, H. M., Leaf structure as related to absorption ofpesticides and other compounds. Res. Rev., 31 (1970)1È150.

48. Hartley, G. S. & Graham-Bryce, I. L., Physical principlesof pesticide behaviour. Academic Press, London, 1980.

49. Bukovac, M. J. & Petracek, P. D., Characterizing pesti-cide and surfactant penetration with isolated plant cut-icles. Pestic. Sci., 37 (1993) 179È94.

polydisperse ethoxylated fatty alcohol surfactants as accelerators of cuticular penetration. 1....

Documents