I. Soc.Cosmet.Chem.

, 25, 621-638 (November1974)

Antiperspirants:
New Trendsin
FormulationandTestingTechnology
ERIC JUNGERMANN, Ph.D.*
PresentedDecember11, 1973,New York City

Synopsis-Aluminumchlorhydroxidehas been the most widely used active ingredient in

ANTIPERSPIRANT FORMULATIONS. Recently,new chemicals,such as basic aluminum
bromide, and combinationsof aluminum, zirconium, and other metal salts have been
introduced.In addition, new product forms are constantlybeing developed with dif-
ferent performance and cosmetic characteristics.The properties of the new active
ingredientsand the new formulationsare discussedwith respectto formulationVERSA-
TILITY, COSMETIC ELEGANCE] and EFFICACY. General methods used to evaluate
stainingpotential, and deodorantand antiperspirantefficacy of these products are re-
viewed. A normal activity method for determiningantiperspirantefficacyis compared
with a method based on a thermally controlled environment, and the results obtained
with these two techniquesare discussed.

INTRODUCTION

Antiperspirantformulationsbaseduponmetal saltssuchas aluminum,zinc,

or magnesiumchlorides,suIfates,acetates,or sulfocarbolatesas the active
ingredientshave beenknownfor a long time (1, 2). The mostimportantanti-
perspirantchemicalusedis basicaluminumchlorhydroxide(ACH) which is
safer,lesscorrosive,and readilyformulatedinto a variety of products(3).
The deodorantand antiperspirantmarket has changeddramaticallyover
the years. Until ca. 1960, the most important product forms were lotions,
creams,sticks,or powders,representing a 100 million dollar per year busi-
ness.In 1960,aerosoldeodorants,primarily alcoholicsolutionscontainingan
antimicrobialagent,cameinto the market,and by 1966,doubledthe size of
the business
(4). Early attemptsto developantiperspirantsin an aerosolform
ran into trouble becauseof the acidic nature of the active ingredients.Most
problemsinvolvedpackagingincompatibility,valve clogging,and perfume
stability.They mostlywere solvedby the mid-sixties(5), resultingin the
introductionof a numberof differenttypesof aerosolantiperspirants.

*Armour-Dial, Inc., GreyhoundTower, Phoenix,Ariz. 85077.

621
622 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

The early aerosolantiperspirants

introducedwere the so-calledpowder-in-
oil type formulationscontaining3-4% of active ingredientssuspendedin an
oil base.There were alsosolution-typeformulations,somein glassaerosols
becauseof corrosionproblems.
In 1971,a cosmeticallydifferentantiperspirantproductwas developed,an
"aerosolpowder"; another version, a so-called"hybrid" appeared in the
marketplaceone year later. The hybrid has a somewhathigher activelevel
than the regularpowder-in-oiland powderantiperspirants and liesbetween
theseproductsin cosmeticelegance.
The mostimportantperformanceattributesof an antiperspirant are its an-
hydrotic and deodoranteffects.In addition, the formulation must possess
aestheticand cosmeticqualities.One frequentlyobserveddrawbackof these
productshasbeentheir tendencyto causefabricstaining.It is the purposeof
this paper to discussrecent trendsin the formulationof aerosolantiperspi-
rantsand someof the testingmethodsusedto evaluateperformance charac-
teristics.

AEROSOL ANTIPERSPIRANT FORMULATIONS

Most aerosolantiperspirantformulationscontain the following compo-

nents: the active ingredients,usually aluminum chlorhydroxideor similar
salts,a liquid systemwhichserveseitherto solubilizeor to suspendthe active
ingredientor is part of a water-in-oilemulsion,miscellaneous additivessuch
as talc, perfume,suspendingagents,and propellants.Suspension systemsrep-
resentthe most commerciallyimportant examples.Someproductsbasedon
solutionsystemshave been marketed,but have not been too successful.
While providingsomewhatgreaterantiperspirantefficacy,solutionsare cos-
meticallylesspleasing.
Many combinations of raw materialsare availablefor the formulationof
aerosolantiperspirants and their selectionmustbe carefullyconsidered,
since
the surfacechemistryof the systemcan affect sedimentationand dispersion
characteristics of the formula.In addition, formulationsmust provide maxi-
mum antiperspirantand deodoranteffectiveness, maximumsafety,cosmetic
elegance,and minimumstaining.

Active Ingredients
Aluminumchloridehas been recognizedfor many years as an excellent
antiperspirant.However, becauseof its low pH, it will causefabric damage
and skin irritation (6). This had led to the developmentof variousbasic
aluminumcompounds which are lessacidic than the parent product.The
mostfrequentlyused of thesederivativesis basicaluminumchlorhydroxide.
Other metal saltsthat havebeenformulatedinto antiperspirants are shownin
Table I. In addition to thesecompoundswhich are consideredto interfere
 TRENDS IN ANTIPERSPIRANTS 623

Table I

Active Ingredientsfor AntiperspirantFormulations

Aluminumchlorhydrate
Aluminum chloride
Basic aluminum bromide
Basicaluminumhydroxychloride-zirconylhydroxyoxyehloride
Aluminumhydroxychloride-zirconylhydroxyoxychloride-glycinecomplex
Basic aluminum nitrate
Basicaluminum bromide-zirconyl hydroxy oxybromide
Magnesiumaluminum zirconluregluconatechloride
Basic aluminum iodide

with the sweatduct (7), othermaterialsthat have been reportedinclude

formaldehyde(8) andcompounds suchasanticholinergicscopolamine deriva-
tives (9).
Amongstthe metal saltswhich have gainedsomecommercialimportance
are basicaluminumbromide (BAB) and certainzirconiumcomplexes. Zir-
conlureoxychloride(ZrOCI2 and zirconylhydroxychlotide [ZrO(OH) C1]
have good antiperspirantactivity, but the aqueouspH of these com-
poundsis verylow. The useof zirconiumcompoundsasantiperspirant
actives
hasreqniredraisingthepH withoutcausing precipitation.
Commercially, this
has been achievedby bufferingthe zirconiumsaltswith basicaluminum
salts(10-14). Whilethiscancauseconsiderable gelling(14), it wasfound
that the additionof certainaminoacidssuchas glycine(11, 14) can control
the gelationproblem.There are alsopatentscovetingaluminumzirconium
systems usingdifferentbufferingsystems in antiperspirant
formulations (15,
16). A numberof experimental saltshave alsobeeninvestigated, including
basicaluminumnitrates,iodides,and mixed metal systems.

Suspending
Oils
Someof the cosmetic oilsusedin powder-in-oil
formulations are listedin
TableII. They includeisopropylandpropy]eneglycolestersof variouslong-
Table II

Cosmetic Oils Used in Aerosol Powder-in-oil Antiperspirants

Isopropyl myristate
Isopropylpalmirate
Mixed isopropylestersof variousfatty acids
Propy]eneglycol dicaprate
Propyleneglycol400 mono]aurate
Propyleneglycoldipelargonate
Triethyl citrate
Dibutyl phthalate
Organosilicones
624 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

chainfatty acids,homopolymers
of polypropyleneoxide,dibutylphthalate,
Tween 60, and certainorganosilieones.
Propertiesof theseoilsincludelow
odor,low viscosity,andgoodstability.

SuspendingAgents
Suspending agentsare frequentlyusedin aerosolpowder-in-oil systems
wherethe activematerialis dispersed ratherthan dissolved. The suspending
agentsreducethe rate of settlingof the dispersed materials.
An effectivesuspending agentcommonly usedis Bentone34,* an organic
derivativeof hydrousmagnesium aluminum silicate.Bentone requiresbatch
heatingand extensive homogenization.Homogenization causesthe formation
of hydrogen bondsbetweensilicasites,resultingin latticeformation.
Anotherfrequentlyusedsuspending agentis Cab-O-Si],ta fire-dryfumed
silica.Chainsare formedvia hydrogenbonds,and with shearmixing,an
effectivelattice structurecan be created.Cab-O-Silis quite sensitiveto the
presenceand the ionicnatureof othermaterials
in the formulation.
Syloid
244*is anotherhigh-porosity
micron-sized
silicathat hasalsobeenused.
The typeof suspendingagentusedin antiperspirantpowderformulations
canbe readilyidentifiedfromX-ray diffractiondeterminations.

Propellants
The propellantsusedin aerosolantiperspirants serveseveralimportant
functions: they deliverthe product,serveas diluentsand/or solvents,
and
assistin product"drying."Generalfactorssuchas bounceand coldness are
directlyrelated.Thecoldersprays generally
produce lessbounce,andthisis
a functionof theboilingpointsof thevariouspropellants
used.Several
trade-
offs are frequentlynecessary in optimizingan antiperspirantpropellant
system.

POWDER-IN-OIL FORMULATIONS

Powder-in-oil
formulations
representthe mostimportantcategory of prod-
uctspresentlyon the market.This categorycontainsthreedifferenttypes:
regularpowder-in-oil
formulations,
theso-called aerosolpowders,andhybrid
formulations.
There are very noticeabledifferencesfrom the cosmeticand
aesthetic
pointof viewbetweenthesethreeproductforms.A typicalpowder-
in-oil formulationis the following:

*NL Industries, 111 Broadway, New York, N.Y.

?Cabot, 125 High St., Boston,Mass.
$W. R. Grace& Co., Baltimore,Mcl.
 TRENDS IN ANTIPERSPIRANTS 625

Ingredient Per Cent

Aluminum chlorhydroxide 3.5
Isopropyl myristate 6.0
Cab-O-Sil O.3
Perfume 0.2
Propellant 90.0

Aerosolpowderformulations
containaboutthe samelevel of ingredientsas
powder-in-oilsversions,but in addition also contain 1% of talc. Also, the oil
levelis considerably
lower.A typicalformulais shownbelow:
Ingredient Per Cent

Aluminum chlorhydroxide 3.5

Talc 1.0
Suspendingoil 1.5
Cab-O-Sil 0.3
Perfume 0.2
Propellant 93.5

The hybrid-typeformulationsare similar to powder-in-oilformulationsex-

cept for a higher activeingredientlevel (5% vs. 3.5%).
The differences between these three formulations are summarized in Table
Ilk As can be seen,the ratio of activeingredientlevelsto the amountof oil
usedrangesfrom 0.6 for the powder-in-oilformula,to a 1.1 ratio for the hy-
brid product,and a ratio of 2.3 for the powder.
Table III

Formulation Differencesbetween AerosolPowder-in-oil,

Powder,and Hybrid Antiperspirants
Formulation Type % Active % Oil % Talc Active: Oil Ratio
Powder-in-oil 3.5 6.0 . .. 0.6
Hybrid 5.0 4.5 .. 1.1
Powder 3.5 1.5 1.0 2.3

Three hybrid-typeproductshave appearedin the marketplace.While fit-

ting into the abovegeneralformulationsystem,there have been significant
differences in the activeingredientsand the suspending oils that were used,
asshownin Table IV. All containapproximately5% of activeingredient,but
the activesdiffer considerably.The first containsthe traditionalbasicalumi-
num chlorhydrox;depowder, the second, a buffered aluminum ehlorhy-
droxide/zirconylhydroxyoxychloridecomplex,wh;le the third usesa buf-
fered aluminum chlorhydroxide/zirconylhydroxy oxychlorideglycine com-
plex.The suspending oilsalsodifferconsiderably.
The hybrid formulationsprovide somewhathigher antiperspirantefficacy
than regularpowder-in-oilformulations, probablybecauseof the higher ac-
626 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

Table IV

Formulation Differencesin Active Ingredients

and SuspendingOils of Three Hybrid Antiperspirants

Product Active Ingredients Suspending

Oil
1 5% altoninurechlorhydroxide Mixed silicones

2 5% basic aluminum hydroxychloride- Mixed silicones/

zirconylhydroxyoxychloride triethyl citrate

3 5% aluminum hydroxych]oride- Dibutyl phthalate

zirconyl hydroxychloride-
glycine complex

tive level. All three are cosmeticallyvery acceptable.Two of the products

have eliminatedthe stainingproblem often associatedwith aerosolantiper-
spirants,while onestill causesconsiderablestaining.

EVALUATION PROCEDURES

Key performancecriteria for which aerosolantiperspirantsare testedin-

clude the following:antiperspiranteffcacy, deodoranteffcacy, stainingpo-
tential, and cosmeticacceptability.
AntiperspirantE•cacy Evaluation
Antiperspirantefficacymay be measuredutilizingboth in vitro and in vivo
methods.Qualitativeprocedures fashionedafter the work of Minor (18), who
used starch-iodineindicators,involved developmentof so-calledsweat pore
patterns.Variousdyes and indicatorshave been usedin coniunctionwith pa-
pers,pastes,films,and lacquers(19). Zaheiskyand Rovensky(20) described
the useof a contactindicatorspottest which wasbasedon a colorreaction
betweenpyrogalloland ferric hydroxidein the presenceof water from sweat.
Theseprocedureshave value in locatingand studyingthe activity of sweat
glands.
Somein vitro methodsare basedupon the fact that aluminumsaltsare
astringentsand are capableof denaturingproteins(21), and uponthe obser-
vation that permeabilityto sodiumand iodide ionsincreasesafter treatment
with aluminumsalts(22).
Quantitativemeasuresof the reductionin perspirationare usuallymade
using gravimetricor humidity sensingtechniqueswith human volunteers.
Someattemptshavebeenmadeto useanimals,suchas cats,mice, or rats for
sweattestingpurposes, but the sweatglandsin thesespeciesare confinedto
the foot pads,and eventhoughthey have eccrineglands,they differ histo-
logicallyfrom thoseof man (22-24). Recently,Lansdownevaluatedthe use
of rat foodpadsas a modelfor examiningantiperspirants(25).
 TRENDS IN ANTIPERSPIRANTS 627

In humanvolunteers,gravimetricmeasurement of the total amountof per-

spirationsecretedmay be accomplished by the use of weighed pads or by
the useof dessicantcups(26). Other typesof measurement utilized resistance
hygrometry, collection coils, phosphoruspentoxide cells, and electrolytic
cells(27-30).
FactorsAffecting SweatingRate
Fredell and Read (31, 32) observedthat there are differencesin the amount
of sweat producedfrom right to left axilla, but that the ratio of these dif-
ferencesis fairly consistent.
They recommended that a ratio betweentreated
and controlaxillaebe usedin judgingeffectiveness. Segar(27) demonstrated
that sweatingis a cyclicalprocessand is not proportionalto the number of
total glands.
Uttley (33) recentlyreviewed someof the factorsinfluencingsweating
which must be consideredin a test procedure.These include: relative hu-
midity, equilibrium of sweatingrate, emotionalor mental stimulation,posi-
tion of the body during sweating,skin area being tested,skin temperature,
conditioningto environment,sexdifferences,and metabolicrate.
Basedupon our studies,we agreewith the observations of earlier workers
(19, 32, 34) that a gravimetricmethodemployingabsorbentpadsis an ade-
quate procedure.Reller (19) observedthat absorbentpadsare more accept-
ablephysiologically thancups.

Methods of Evaluation
We have testedtwo basicvariationsof the gravimetricmethod,the major
differencebeingin the procedureusedfor stimulationand collectionof per-
spiration.The first method,termed "normal activity method,"utilizes normal
environment conditions. The second method, termed "controlled environ-
ment method,"employsthermallycontrolledenvironmentalconditions.In
both methods,a ratio of sweatproducedby the left and right axilla is de-
terminedin a seriesof controlledcollections.The effectof antiperspirantma-
terialson the perspirationrate of eachindividualis determinedby comparing
the post-treatmentratio to the subjects'averagecontrolratio. For each indi-
vidual,the per centchangein sweatrate is calculatedas:

%change
insweat
rate
= 100
(1- post-treatment
average ratio
control
ratio
)
These data are statisticallytreated by applyingthe Studentt distribution
to establish85%confidence limitson the meanper centchangein sweatratio.

Normal Activity Method

In the normal activity method,a group of panelistsare recruitedwho re-
main together as a panel for at least 12 months. Prior to the start of
628 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

usingthe panel for test purposes,their left-to-rightaxillarysweatingratio is

determined12 timesovera 4-weekperiod.The averageresultfrom thesede-
terminationsis usedas the controlratio for eachpanelistfor the life of the
panel.
Beforethe start of any test, there is a 5-day periodin which the panelists
use no underarmproductsexcepta placebosoap.Following this are three
consecutivedays of antiperspiranttesting. There are 15-20 people in a
mixed-sexpanel. Test materialis applied only to the left axilla, the right
axilla servingas the control.For roll-on formulations,1 g of product is ap-
plied with a softbrush.For aerosols, a 1-seesprayfrom a distanceof 6 in. is
applieduniformlyto the axillaryvault. Sampleapplicationis madeby a tech-
nician, using a metronometo time the spray.After 5 rain, to permit evapo-
ration of any volatile materials,a preweighedmoisturecollectionpad is
applied. These pads, measuring2 in. x 2 in., are fashionedfrom sanitary
napkinsby strippingaway severallayersof cellulosefiller in order to get a
better axillary fit. Pads are held in place by stripsof hypoallergenictape.
The exteriorsurfaceof the pad itselfis coveredxvithSaranWrap in order to
retain absorbed moisture.
The panelistsare then free to pursuenormal activitiesfor 4 hours.At the
endof the 4-hourperiod,the absorbent padsare removedandplacedin tared
plasticointmentjarsandweightof perspirationis determinedby difference.

Controlled Environment Method

In the controlledenvironmentmethod,approximatelythe samenumber of

panelistsare used.They are askedto abstainfrom the useof all antiperspirant
materialsat leastoneweek prior to enrollmentthroughthe completionof the
test. Controlled sweat collectionsare made on Monday and Tuesday of the
test week. Post-treatmentsweatcollectionsare made on Wednesday,Thurs-
day, and Friday. Test materialsare assignedin sucha way that samplesare
appliedto the right axilla of half the panelistsand to the left axilla of the
remainingpanelists.Axillae are rinsedwith clear,warm water aud dried just
beforeeachapplication.For aerosolproducts,the axillaeare sprayedfrom the
distanceof 6 in. for 2 sec.Materials are applied immediatelyfollowing the
controlledsweat collectionon Tuesday, I hour prior to sweat collectionon
Wednesday,Thursday,andFriday.
Sweatingis inducedby having the subjectsit in a room maintainedat
100ø -+2øFand at a relative humidity of 35%. During the first 40 min of the
sweat stimulationperiod,the panelistshold unweighedpads of Webril,* a
nonwovencotton,in their axil]ae.Thispreliminarywarmupperiodis followed
by two successive 20-rain collectionperiods during which panelistshold
pre-weighedWebril padsin the axillae.Panelistsare allowedto drink ice wa-
*Kendall Co., Walpole, Mass.
 TRENDS IN ANTIPERSPIRANTS 629

ter asdesiredthroughout the collection

period.At the end of eachcollection
period,the padsare removedand placedin tared bottlesfor reweighing.
Antiperspirantactivityis calculatedasdescribedpreviously.

DeodorantEJficacyEvaluation
Deodoranteffectiveness is a recognizedattribute of many metal salts
(35-37). Efficacyis evaluated
by comparisonof the effectof onetreatmentin
one axilla versus a second treatment in the other axilla of the individuals in
panel groups(34, 38). Similarcomparativemethodsmay employa crossover
procedure.Productcomparisons usinga split axilla treatmentwithout cross-
overin a singlegroupsufferin reliabilitybecauseodorintensityof the axillae
of an individualvaries.However,when the test sitesin a panel are random-
izedby a crossover
procedure,
effectsdueto inherentdifferences
in the axil-
lae are minimized.
Axillary odor may be judged by the panel participantsthemselves,by
trainedjudges,or both.Lengthof deodoranteffectis usuallydeterminedat
variousintervalsduring the test, or at cessationto treatment.Evaluation may
bc madeby estimatingthe degreeof odorof both the axilhe and the under-
garmentat the sideof contact.Odor judgmentsobtainedare usuallybased
uponarbitrarynumericalscales(38-40). Becausetheseare subjectiveevalu-
ations,severalinvestigatorshave attemptedto eliminatesourcesof error
throughthe useof osmometers (41).
Utilizing a crossovermethod in which odor evaluatiouswere made by
trainedjudges,a cleardeodoranteffectwasshownfor a powdertin-oilformu-
lation and an aerosolpowder formulation,both containing3.5% alnminum
chlorhydroxide.An odorscaleof i (little or no odor) to 5 (strongor disagree-
able odor) wasused.The resultsare shownin Table V. The data from the test
andcontrolperiodsweresubiected to Student'st testandwerehighlysignifi-
cant.

A recentadvancein evaluatingdeodorantefficacyis that developedby

Dravnieksand coworkers(42, 43). Their methodis basedon the development
of chromatographic
patternsof axillaryodors.

Staining Potential
Soonafter the introductionof aerosolantiperspirants,it becameapparent
that fabricstainingin the axillaryareawasa majorproblem.Initially, formu-
lationswere testedfor stainingpropensityby applyingthe productdirectlyto
the fabric.It wasfoundthat thiswasnot a satisfactoryproceduresinceit did
notreflectactualuseconditions anda comparative
procedure wasdeveloped.
In thisprocedure,a panelis usedmadeup of 10 •nenand 10 women.Sub-
jectsare required to abstainfrownthe use of all antiperspirantsand deodor-
antsor otherproductsappliedto the axillaryarea, and are requiredto use
630 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

Table V

In Vive DeodorantEvaluationof Two Powder-In-OilAntiperspirants

(Summaryof IndividualMean Odor Scores•)

Subiect Control Period Test Period

Number Dry Powder Powder-In-Oil Dry Powder Powder-In-Oil

1 4.5 4.7 2.0 1.9

2 4.0 4.3 2.8 2.4
3 4.9 4.7 1.3 1.5
4 4.1 4.1 1.8 1.8
5 5.0 5.0 1.9 1.8
6 4.8 4.8 2.8 2.4
8 2.9 3.0 2.0 2.2
9 4.0 4.3 1.5 1.5
10 3.6 3.8 1.9 1.7
11 4.4 4.6 2.9 2.8
12 4.0 3.7 2.9 2.7
14 4.4 3.9 2.3 1.5
15 4.2 4.5 2.3 2.5
16 4.5 4.5 3.6 3.2
17 3.1 3.2 2.2 2.1
18 4.3 4.3 2.1 1.9
19 3.5 3.5 2.5 2.6
20 4.6 3.6 2.4 2.0
21 4.2 4.4 3.1 2.9
22 4.3 3.2 2.0 1.9
Mean 4.16 4.11 2.32 2.16
StandardDeviation 0.5518 0.5862 0.5677 0.5008
Standard Error 0.19.33 0.1310 0.1269 0.1119

•A five-pointscoringsystemwasused,1 beingweakand5 beingstronganddisagreeable

oflor.

onlyIvory soap*duringthe periodof the test.T-shirtsare suppliedto the

subiects at the beginningof the testperiod.A standard3.5%powder-in-oil
isopropyl myristateformulation is usedfor referencepurposes. Subiects are
evaluated for stainpropensityin eachaxillausingthefollowingschedule. On
day 1, a 3-seesprayof the productis appliedby a technicianusinga metro-
nomefor timing.ThenewT-shirtisthenworn.Fourhourslater,thespraying
procedureis repeated.Two hourslater,the T-shirtsare collectedand laun-
dered.The sameroutineis thenfollowedfor 2 moredays.At the endof the
thirdday,the axillaryareasof eachT-shirtareevaluated forintensityof stain
usinga photovolt reflectometerequipped with tristimulus
filters.Readings are
takenusingcachof the threefiltersandthe stainingvalueis calculated from
the formula:

*Procter& Gamble Co., Cincinnati,Ohio.

 TRENDS IN ANTIPERSPIRANTS 631

amber-blue •
x 10
green

A typical value for the controlformula is 0.7. A minimally visible stain

yieldsa value of 0.3. A value of 1 or higher indicatesa productwhich stains
fabric considerably.
When evaluatingtest products,the procedureis the sameas during the
standardizationexceptthat the test product is substitutedfor the conh'ol.
Applicationto the left and right axilla is randomized.Calculationsare nor-
malizedto the conh'olvalue of 0.7 so that resultsare reproducible.Someof
the resultsobtainedusingthis comparativemethodare shownin Table VI.
There were considerable variationsin stain potentialof products,depending
onproductformandthe typeof suspending oilsused.

Table VI

StainingPropensityof AerosolAntiperspirantFormulations
(Three-Day Test Period)
Formulation Formulation
(Active/Oil) Type StainingValue
5% aluminumhydroxychloride- Hybrid 1.0
zirconyl hydroxychloride-
glycinecomplex/dibutylphthalate

5% alumimnnchlorhydrate/mixed Hybrid 0.26

silicones

3.5% aluminum½hlorhydrate
/ mixed Aerosol 0.17
silicones powder

3.5% aluminumchlorhydrate/ Powder 0.70

isopropyl
myristate in-oil

CosmeticAcceptability
In consideringthe overallpropertiesof an antiperspirant, cosmeticand
perfumequalitiesare a key consideration. The dispensing oils,when not
runny,cancreatea pleasant tactilefeel.The propellant, whenoptimized, can
reducecoldness of the spraywhenappliedto the skin.Overall,a productis
cosmetically
betterwhenit goesondry,withoutcaking, andis gentleto both
skin and clothing.The besttechniquefor evaluatingthesepropertiesand
theiroveralleffecton productacceptability is by full scalemarketresearch
procedures,thoughsomelaboratory procedures for individualpropertiesare
also used.
632 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

Safety
The criteria for evaluatingsafety of antiperspirantformulationsinclude
a consideration of the effect of the product on the skin and its aerosolized
properties.Routinesafetyevaluationof a new formulationcomprisesdeter-
minationof acute oral and dermal toxicity,primary eye and skin irritation,
acute and subacuteinhalationstudiesin rabbits and monkeys,and sensitiza-
tion in guineapigsand in humans.
Aerosolantiperspirants have receivedconsiderable publicity as a resultof
potentialinhalationhazardsthat may be associated with low-level,long-term
exposureto propellantgasesand particulateaerosolsin general.Recently,
Drew (44) reportedthat persistentlung granulomas were observedin ham-
stersexposedto an aerosolized liquid systembasedon a propyleneglycol
complexof ACH at 50 mg/ma for 6 hoursa day for 10, 20, and 30 days.
Antiperspirantsare classified
asdrugsand are subjectto the Food and Drug
Administration's OTC review procedure.A requestto provide safety and
efficacydatafor reviewby the OTC panelhasbeenmade (45) and heatings
startedin the earlypart of 1974.

ANTIPERSPIRANT TEST PtESULTS

From the earlierdescription of the two methodsfor evaluatingantiperspi-

rant efficacyusedin our laboratory,namely,the normalactivity methodand
the controlledenvironmentmethod,it seemedplausibleto hypothesizethat
sincemethodologies were basicallysimilar,the resultstoo would be of the
sameorderof magnitude, or at leastprovidesimilardirectionalguidelines.

NominalActivity Method
The resultobtainedwhentestinga standardpowder-in-oilfor•nulationon
four differentoccasions
spreadovera periodof a year and a half are shown
in Table VII. The resultsare 3-day averagesand are expressed
with a con-
fidence limit of 95%.

Table VII

Antiperspirant Efficacy of a 3.5%

Aluminum Chlorhydrate Powder-in-oil
Formulation (Normal Activity Method)
Test Number Mean % Sxveat Reduction

œ0.6 ___1.8
21.6 4- œ.6
œ1.0 4- œ.4
27.8 ___œ.7
 TRENDS IN ANTIPERSPIRANTS 633

Table VIII

AntiperspirantEfficacy of Different Actives

in Powder-in-oilFormulations(Normal Activity Method)
Per Cent Mean % Sweat
Active Reduction

Aluminum
chlorhydroxide• 3.5 22.8 ñ 2.7
Basic aluminum bromide 3.5 23.6 + 4.0
Basic a]uminmn nitrate 3.5 23.2 + 3.9
Basic aluminum bromide- 3.5 25.5 ñ 4.0
zirconylhydroxyoxy-
bromide
Basicaluminum hydroxy 3.5 25.1 ___3.1
chloride-zirconylhydroxy
oxychloride
Magnesium aluminum zirconium 3.5 22.4 +__4.0
gluconatechloride
Magnesium aluminum zirconium 3.5 22.7 ñ 2.9
gluconatebromide

In a secondseries,differentactivematerialswere examinedin a powder-in-

oil formulation.The compoundstested and the efficacyresultsare listed in
Table VIII. It was found that regardlessof the activematerialused,the re-
sultsall fell within a verynarrowrange.
The resultsobtained comparingthe sweat reductionefficacyof different
types of antiperspirantformulationsincludingan aerosolpowder-in-oil,an
aerosolpowder,a hybrid formula,a roll-on,and a 20% aluminumchlorhy-
droxidesolutionare shownin Table IX. In the solutionforms (roll-on or 20%
aqueous)between35 and40%sweatreductionwasobserved.The two powder
versions
reducedperspiration
by 22%,whilethe hybridformulagavea slight-
ly higher result, namely 27%, a differencewhich is real from the statistical
point of view, and probablydue to the higherlevel of activeingredient(5%
vs.3.5%ACH ) of the hybridversion.

Table IX

AntiperspirantEfficacyof Different
FormulationTypes(Normal Activity Method)
Product Category Mean % Sweat Reduction

3.5% powder-in-oil 22.8 ñ 2.7

3.5% dry powder 22.1 q- 3.3
20% ACH solution 39.9 ___4.0
Hybrid formulation 27.0 ñ 4.1
Roll-on formulation 37.5 ñ 2.6
Placebo 2.8 ñ 2.8
634 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

Table X

Sweat Reduction as a Function of Time for

3.5% Aluminum ChlorhydroxidePowder-in-oil
Formulation (Nounal Activity Method)
Mean % Sweat Reduction

Time (hrs) Test 1 Test 2 Test 3

22.4 _ 3.8 21.1 _ 2.9 24.6 _ 2.4
20.0 __+4.3 20.3 _+_4.0 24.0 _ 2.7
18.9 __+3.6 18.0 _+ 3.7 19.2 _ B.0
14.0 __+4.1 12.4 _+ 2.6 17.1 _ 2.5

Table XI

Effect of SprayingTime (Concentration)

on Sweat Reduction(Normal Activity Method)
Duration of Spray Mean % Sweat Reduction

(see) Formula A Formula B

0.5 8.5 _ 2.5 10.7 _ 2.5

I 22.8 __+2.9 25.1 _ 4.5
2 28.0 __+3.2 30.3 _+ 3.7
4 34.9 _ 3.8 35.7 _ 3.7

Sincethe variouspowder-in-oilformulationsgave resultsin a very narrow

cfflcacyrange,regardlessof activeingredient,it was decidedto test the sen-
sitivityof the method.A seriesof sweatreductiondeterminationswere carried
out as a funct!onof time. Four collectionswere made followingone applica-
tion, at 3, 6, 9, and 12 hourson 3 successive
days.The data obtainedin three
separatetestswith a 3.5% powder-in-oilformulationare shownin Table X.
Each test, conductedindependently,producedessentiallythe sameper cent
decreasesin sweat reduction over time. The method was able to distinguish
quite successfully and distinctlybetweenrelatively small variationsin sweat
reduction.
In a secondseriesof testsin which different sprayingtimes were used,
data were similarlyobtainedafter one applicationin 3 consecutivedays.The
data are in Table XI. As the sprayingtime increased,the per cent sweatre-
duction also increased.
The effect of the aluminumchlorhydroxide particle size on efficacywas
alsoevaluated.Three differentgradeswere used:ChlorhydrolUltrafine,* one
grade with a smallerparticle size (56% retainedon 10-/xscreen),and one
with a largerparticlesize (96% retainedon 10-/xscreen).No significantdif-
ferenceswerefoundbetweenthe grades,asshownin Table XII.
øReheisChemical Co., Phoenix,Ariz.
 TRENDS IN ANTIPERSPIRANTS 635

Table XII

Effect of Particle Size

on Sweat Reduction(Normal Activity Method)
Mean % Sweat Reduction

Regular (UltraFine ACH) 22.8 ñ 2.9

Smaller particle size than 24.9 _ 3.3
UltraFine ACH (96% retained on
10-tzscreen)
Larger particle size than 23.9 ñ 3.1
UltraFine ACH (56% retained on
10-tzscreen)

Table XIII

AntiperspirantEfficacyof a 3.5% Aluminum Chlorhydrate

Powder-in-oil Formulation(ControlledEnvironment Method)
Test Mean % Sweat Reduction

1 12.6ñ 3.6
2 33.6 _ 5.2
3 16.0 _ 4.1
4 22.3 _ 9.8

Table XIV

AntiperspirantEfficacy of a 3.5% Basic Aluminum Bromide

Powder in-oil Formulation (Controlled Environment Method)
Test Mean % Sweat Reduction

1 35.0 _ 4.2
2 6.7 _ 6.0
3 21.2 ñ 4.2

Controlled Environment Method

The same3.5% powder-in-oilformulationusedin the normal methodwas

evaluated under the controlled environmental conditions. The results ob-
tained are summarizedin Table XIII. Resultsranged from 12.6%to 33.6%
sweatreduction.In anotherseries,a 3.5% basic aluminumbromide powder-
in-oil formulationwas tested.The resultsare shownin Table XIV and again
showedlittle reproducibility.

DISCUSSION OF RESULTS

In general,it was found that in our handswhen usingthe controlledenvi-

ronmentmethod,there was a wide spreadof resultswhen the samefonnu-
636 JOUBNAL OF THE SOCIETY OF COSMETIC CHEMISTS

lationsand the sameactiveingredientswere tested.Becauseof thesedifflcul-

ties, we have fewer resultsto report, and are continuingto investigatethis
technique.
On the other hand, the normal activity method gave reproducibleresults
nnder a variety of conditions.It did not show any real differencesbetween
the effectiveness of different active ingredientsunder the sameformulation
conditions.That did not mean that the normal activity method was insensi-
tive, since it very effectively distinguishedbetween powder-in-oil formula-
tions and roll-ohs,and clearly indicatedvariationsof antiperspirantefficacy
as a function of time or concentration.
The controlledenvironmentmethoddiffersfrom the normal activity meth-
od in severalways. The controlledmethodemploysa g-seespray,whereas
the normal methodutilizes a 1-seespray.Treatment and collectiontimes are
different;in particular, the time periodsfor collectionbetween the normal
and controlledmethodsdiffer in length (4 hoursvs. 40 min) and conditioning
i.e., the controlledmethodutilizesa conditioningperiod beforethe pads are
appliedfor collectionpurposes.
Another reason for the differences observed between the normal and the
controlledtestingtechniquesis possiblydueto the conditions
usedin the con-
trolledtest. Goinginto a 100øFsweatchamberat 35% humiditycan be a
challengingstimulusto the body. Consequently, not only is perspiration
elicitedasa functionof temperature,but alsodue to the challengeto the cen-
tral nervoussystem(19)
CONCLUSIONS

We have concludedthat the normalactivity methodis a more usefultool

for providingguidanceon the relative efficacyof differentformulationsand
differentactiveingredients.
The methodalsocloselyapproximates
what the
consumerexperiencesin actual use and we have been able to correlate
efficacydata with observations
reportedin large-scale,
carefullydesigned
constuner tests.

Certainadditionalconsiderations
mustbe kept in mind when considering
gravimetricproceduresfor the determinationof antiperspirantefficacy.
Bakiewicz(48) has shownthat when thermal stimulationis usedto induce
perspiration,resultscan vary with body position.Resultscan also be influ-
encedby drinkingcold liquids,variationsin the relativehumidityof the
room,or even by selectingpanelistswho are "high sweaters"or "low sweat-
ers."For example,Tronn•erandRentschler
(47) reportedthat the sameprod-
uct under controlledenvironmentalconditionsgave a 9•0%sweat reduction
whenappliedto a lowsweater,
whiletherewasa 50%reduction
with a group
of peopleclassified
ashigh sweaters.
To summarizeour viewson antiperspiranttest methodology:regardless of
the methodsusedby us or reportedin the literature,mostof our resultsfor
 TRENDS IN ANTIPERSPIRANTS 637

sweatreductionettlcacies
for aerosolpowder-in-oiland hybrid formulasfall
between20%to 30%.Thesefigurescanbe restatedby sayingthat a subject
was sweating with a 70% to 80% ettlciencyrather than his normal 100%
ettlciency.Thus, evenif laboratoryprocedures knowhow to measurethese
differences,it is debatablewhetherthe consumercan distinguishbetween
them. While not negatingthe use of thesegravimetricmethodsto provide
data to useas a guideto optimizeformulations, or to coinparenew active
ingredients,or to evaluateinteractionof materials,care must be taken.when
thesenumbersare used for promotionalpurposes.Sinall differences,even
thoughstatisticallysignificant,
shouldnot be magnifiedout of proportion.It
is importantthat management understandsand appreciatesthe differencesin
testinginethods,and recognizesthe limitationsand specializedmeaningof
the data derivedfrom thesetechniques.
(ReceivedApril 2, 1974)

REFERENCES

(1) Klarmann, E.G., The cosmetic aspects of perspiration and its control, Amer.
Per•um. Essent.Oil Rev., 52, 33-40 (1948).
(2) Salfield,H., The mechanismof perspiration,Ibid., 53, 385-7 (1949).
(3) deNavarre,M. G., The Chemistryand Manufactureof Cosmetics,2, D. Van Nostrand•
Princeton, N.J. 1962.
(4) Deodorantssprayingup a storm,Progressive Grocer,52 (8), 144-6 (1973).
(5) Reheis,Soonaerosolantiperspirants will dominatemarket,Drug Trade News (March
10, 1969).
(6) Bein, R. R., The action of antiperspirantcreamson fabrics,Proc. Sci. Sec. Toilet
GoodsAss., 4, 8 (1945).
(7) Papa, C. M., and Kligman, A.M., Mechanismsof eccrine anhydrosis.II. The
antiperspiranteffectof aluminumsalts,J. Invest. Dermatol.,49, 139 (1967).
(8) Parma,C. M., The action of antiperspirants,J. $oc. Cosmet. Chem., 17, 789-800
(1966).
(9) BritishPatent940,279(1963).
(lO) Helton, E.G., Daley, E. W., anti Erwin, J. C., Zirconlureoxychlorideantiperspirant,
Drug Cosmet.Ind., 80, 170 (1957).
(11) Grad, M., U.S. Patent2,854,382to Procterand Gamble(1958).
(12) Beckman,S. M., U.S. Patent 2,906,668 to Reheis Chemical (1959).
(13) Daley, E. W., U.S. Patent 2,814,584 to Procterand Gamble (1957).
(14) Daley, E. W., U.S. Patent 2,814,585 to Procter and Gamble (1957).
(15) Wainer, E., U.S. Patent 2,507,128 to National Lead (1950).
(16) Grote,I..W, U.S. Patent3,009,860to Chattanooga Medicine(1961).
(•7) GivaudanCorp., private communication(1973).
(18) Minor, V., Eines neues Verfahren zu der klinischen Untersuchung der Schweissa-
banderung,Deut. Z. Nervenh, 101, 302 (1927).
(19) Reller, H. H., Factors affecting auxfilary sweating,J. Soc..Cosmet. Chem., 15,
99-110 (1964).
(20) Zaheisky, •., and Rovensky,•., A comparisonof the effectivenessof several exter-
nal antiperspirants,
Ibid., 23, 775-89 (1972).
(21) Govett, T., deNavarre, M. G., Aluminum chlorhydrate: a new antiperspirant in-
gredient,Amer. Perpum.,49, 365 (1947).
(22) Ukrami, C., and Christian,J. E., An evaluationof the effectiveness
of antiperspi-
rant preparationsusing frog membraneand radioactivetracer techniques,J. Amer.
Pharm.Ass.,Sci. Ed., 42, 179 (1953).
638 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

(23) Wechs•er,H. L. and Fisher,'•. R., Eccrineglandsof the rat: Response

to induced
sweat:n•,hypertension,
u:'emia,andalternationsof sodiumstate.Arch.Dermatol.,97,
189 (1968).
(24) Alpin,R. S., Vocac,I. A., Saunders,D., andWard,I. W., Methodfor the evaluation
of antihidroticsubstances in the anaesthetised
cat, J. Pharm.Sci., 56, 449 (1967).
(25) Lansdown, A. B. G., The rat foot pad as a modelfor examining antiperspirants,
J. Soc.Cosmet.Chem.,24, 677-84 (1973).
(26) Wooding,W. M., Jass,H. E., and Ugelow,I., Statistical
evaluation
of quantitative
antiperspirant
data,Ibid., 15, 579-92 (1964).
(27) Segar,W. E., Hancock,
C. I., Sawyer,T., andBullard,R. W., Continuous
record-
ing of sweating
by resistance
hygrometry,
Pediatric&
35, 693-8 (1965).
(28)Custance,
A. C., Cyclingof sweatglandactivityrevealedby a new technique,
J. Appl.Physiol.,
17, 741 (1962).
(29)Jenkins,
J. W., Ouelette,P. A., Healy,D. J., andLana,C. D., A technique
for per-
spirationmeasurement,
Proc.Sci. Sect.ToiletGoodsAss.,42, 12-5 (1964).
(30) O'Ma]ley,W. J., and Christian,
J. E., The designof a continuousrecording
in vivo
methodof measuring sensible
perspirationovera limitedarea,J. Amer.Pharm.A•s.,
49, 398-401 (1960).
(31)Fredell,W. G., and Read,R. R., Antiperspirant
testing,Drug Cosmet.Ind., 69,
40-1, 111-5 (19•1).
(32)Fredell,W. G., and Read,R. R., Anfiperspirant-axil]ary
methodof deterlnining
effectiveness,
Proc.Sci.Sect.ToiletGoods
Ass.,15,23-7 (1951).
(•3) Uttley,M., Measurementand controloœperspiration,
]. Soc.Cosmet.
Chem.,23,
23-43 (1972).
(34)Fredell,W. G., andLongfellow,
1., Reporton evaluating
antiperspirant
anddeodor-
ant r•roducts.
Ibid., 9, 108-11 (1958).
(35)Shelley,W. B., Hurley,H. J., andNichols,
A. C., Axillaryodor-experimental
study
of the role of bacteria,apocrinesweat,and deodorants, AMA Arch. Dermatol.
Su•hilol.,
68,430-66(1953).
(36) Shehadeh,
N.H., andKligman, A.M.. The effectof topicalantibacterial
agents
on
thebacterial
floraof theaxilla,J. Invest.Dermatol.,
40, 61-71 (1963).
(37)Meyer-Rohn,J.,Uberdieursachen lastigen
korpergeruches,
Fette,SeUen,Anstrichm.,
67, 353-4 (1965).
(38)Whitehouse,
H. S.,andCarter,R. O., Evaluation
of deodorant
toiletbars,Proc.Sci.
Sect.ToiletGoodsAss.,48, 31--7 (1967).
(39)Fredell,W. G., andRead,R. R., Axillaryperspiration-odor
anddeodorization,
Ibid.,
25, 32-6 (1950).
(40)Gee,A. H., andSeldenberg,
I. L., Evaluation
of germicidal
soaps,
SoapChem.Spec.,
30, 42-5, 159-63 (1954).
(41)Barall,L. C., Measurement
of odors
in thecosmetic
indust13•,
]. Soc.Cosmet.
Chem.,
2, 4 (10•0).
(42)Dravnieks,
A., Krotoszynski,
B. K., Lieb,W. E., andIungermann,
E., Influence
of
an antibacterial
soapof variouseffluents
fromaxi]lae,Ibid.,19,611-26 (1968).
(48)Draynicks,A., Krotoszynskl,
B. K., Casely,R. E., and Taber,D., Gaschromato-
graphicevaluationof intensity
of axillaryodors,
Part2, ToiletGoods Ass.]., 1, 1,
14-9 (1969},
(44)Drew,R. T., Inhalation
studies
witha glycolcomplex
of aluminum chloride
hy-
droxide,
presentedat theAnnual
Conference
of American
Industrial
Hygiene Assn.,
Boston,Mass., 1973.
(45)Over-the-counter
antiperspirant
drugproducts
and over-the-counter
topicalanti-
bioticdrugs.Safetyand efficacy
review;requestfor dataand information,
Fed.
Regist.,38, 173,24391--3(1973).
(46) Bakiewicz,T. A., A criticalevaluationof the methodsavailablefor measurement
ofantiperspirancy,
]. Soc.Cosmet.
Chem.,
24,245-58(1973).
(47)Tronnier,
H., andRentschief,
O.,Experimentelle
untersuchungen
zurwirkungsweise
a]uminiumhaltiger
antipeispiranzien,
Ibid.,24,281-90(1973).