Combability Measurements on Human Hair

I. Soc.Cosmet.Chem.,27, 379-398 (September1976) CombabilityMeasurements on Human Hair MARIO L. GARCIA, Ph.D. and JOSE DIAZ, B.S.* PresentedMay 29, 19...
1 downloads 2 Views 355KB Size
I. Soc.Cosmet.Chem.,27, 379-398 (September1976)

CombabilityMeasurements on Human Hair MARIO L. GARCIA, Ph.D. and JOSE DIAZ, B.S.* PresentedMay 29, 1975,SCC Seminar,St. Louis, Missouri Synopsis:An instrumentalmethodfor MEASURING the effect of cosmeticproductsor any other treatmentof the COMBABILITY of HUMAN HAIR has been developed.The required instrumentation,experimentalprocedure, and interpretation of the data are presentedin detail. The methodinvolvesthe continuousrecordingof the forces,

whichopposethe motionof a combthrougha swatchof hair. The data thusproduced consistsof graphsshowingthe forcesopposing(or generatedby) combingas a function

of the positionof the combalongthe lengthof the swatch.Examplesof applications are presented.

INTRODUCTION

Combability canbe definedasthesubjective perception of therelativecase

ordifficulty withwhichhuman haircanbecombed. It depends onthemagnitudeandonthefluctuations oftheforces thatoppose combing. Combability is an importantattribute,whichis alwaysconsidered when judgingthe"condition" of humanhair.Improvedcombability is perceived as the hair beingin bettercondition. Anotherconceptcloselyassociated with combability isthatof manageability. Still anotherfactorrelatedto combability is that of the mechanical damage,whichis doneto hair with the combing process, whichis accelerated if thehair is hardto combor to untangle.It follows that combability,due to its closeconnectionwith other desirablehair

qualities, is a veryimportantfactorin judgingthe performance of manyhair careproducts. The methoddescribedin this paperwas developedin our laboratoriesfor

the purposeof quantitativelyevaluatingcombability.It hasbeenextensively testedwith a wide varietyof hair productsand treatmentsand is now usedas a standardtest during productdevelopmentand for claim substantiation in finishedproducts.A numberof instrumentalmethodsfor evaluatingcombability have been reportedin the literature (1-3). Someof the similaritiesand differences between those methods and ours will be discussed later. It is our

opinionthat ourmethodhasadvantages in its simplicityand in the type of informationthat canbe obtainedby usingit.

*ClairolInc. Research Labs.,2 Blachley Road,Stamford, Conn.06902. 379

380

JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

In what follows, the method is first describedin full detail so that it can

easilybe usedby anyinterested laboratory.This description is followedby a selectednumberof experimentalresults,interpretationof the data, and a generaldiscussion of the method. M•.TaOD

Experimentally, the methodconsists of suspending a hair swatchfrom a force-measuring device,insertinga combcloseto the root end of the swatch, settingthe combin a straightcombingmotionthroughthe swatchat a constantspeed,andcontinuously recordingthe forcesthat resistits motionduring thistransitfromthe pointof insertiontill it clearsthe tip end of the swatch. The data resultingfrom this operationconsistof a graphshowingthe load (in grams)opposing (or generated by) combingas a functionof the position of the combalongthe lengthof the swatch.We call this grapha 'Combing Curve.'

Combingcurvescan be recordedusingdry or wet hair. Typicalexamples of thesecurvescan be seenin Fig. 1 (dry) and Fig. 2 (wet). Dry combing curves are recorded using swatches,which have been previouslyhand combed.In spiteof the precombing, they showgraduallyincreasing combing forceswhichreachmaximumvaluesat or nearthe tip endof the swatch. Wet combingcurvesare recordedusingswatches which havebeenpurposely tangledby immersing themin water.The resultingcurveshowsa highincidenceof tanglesall throughthe length of the swatch.In somecases,the combingforcesare highercloseto the tip end of the swatch. In our method,combabqityis measuredby meansof two parameters, which

can be directly obtainedfrom the combingcurves.The first parameteris 'peakcombingload' (PCL). This is the highestload (in grams)that is recordedduringthe combingof the swatch.PointsP in Figs. 1 and 2 are examplesof PCLs. If desired,PCL can be convertedto peak combingforces (PCF) (in dynes) by multiplyingthem by the accelerationof gravity (•980. cm/sec2).The secondparameteris the averagecombingload (ACL). This is the averageload duringonecombingof the swatch.It is expressed in grams ß cm units) by the distancein centimeterstraveled by the comb throughthe swatch. Both of theseparametersgive us a quantitativemeasureof how difficult

(or easy)it is to comba swatchof hair. Our methodis basedon measuring the changesthat occurin suchparameterswhen the hair is treatedwith a product.Decreasesin PCL and/or ACL, which indicate improvements in combability(and vice versa)correlatewith what is perceivedwhen the hair is combedby hand.

As couldbe expected,the absolutevaluesof the PCLs and ACLs depend on a large number of facto.rssuch as speedof combing,handlingof the hair, dimensionsof the hair swatch, curlinessof the hair, comb dimensions, combmaterial,etc., which cannotbe totally controlled.It is for thesereasons

MEASUREMENTS

ON

HUMAN

HAIR

381

IO n

• eo 0

0,e=

u i•-

,

i

,

OIIITANCE

0

FROM

TOP

Figure 1. Dry combing curves, B before treatment, A after treatment with commercial creme rinse. Combing loads ap-

pear plotted againstdistanceof combfrom top end of swatch

Im ITANCE

(CM)

i

i

i

I0 FROM

!

......

TOP

'

5 (CM)

'

'

0

Figure 2. Wet combing curves, B before treatment, A after treatment with commercial creme rinse. Combing loads appear plotted against distance of comb from top end of swatch

that changesin the values of these parametersare more important and more reproduciblethan their absolutevalues. It also follows that great caxehas to be taken so that comparisons-as, for example,before and after the use of a product-are done under experimental conditionsthat are, insofar as possible,identical. If done carefully, however, this method allowsus to measurechangesin combingforcesof the order of -+ 20 per cent. Averagechangesare calculatedby averagingthe individualvaluesmeasured on a setof replicateswatches. Equipment

The instrumentationconsistsof an Instron Tensile Tester (Metric Table Model, TM-M*) to which someattachmentshave been added (see Fig. 3). The Instron load cell B which has a range of 0 to 2,0'00g is used.Other recordingtensiletestinginstruments couldbe similarlyadapted. The attachmentsto the Instron Tester shownin Fig. 3 are as follmvs (c) CombStand:the combstandis an L-shapedaluminumpart designedto hold differenttypesof combs,it is mountedon the travelingcrossbarof the Instronby meansof two screws;(b) Comb: the comb usedin our measurementsconsistsof 8 cylindricalstainlesssteel teeth (unpolished),2.2 mm in diameter,mounted(with an interteethdistanceof 1.5 mm) on an aluminum frame. Two removable bars of the same material and dimneter as the teeth are

*Instron Corp., Canton, MA.

38•,

JOUBNAL OF THE SOCIETY OF COSMETIC CHEMISTS

cl

I

, I

Figure 3. Part of Instron Tensile Tester showingcombingattachments:(a) clamp; (b) comb; (c) L-shapedstand; (d) crossbar;(e) hair swatch;(f) load cell

mountedperpendicularto them in order to keep the hair in placeduring combing(Fig. 4). (a) Clamps:the hair as usedfor thesemeasurements is

mountedonspeciallydesigned aluminumclamps(Fig. 4). Theseclampsproduceapproximately rectangular cross-sectional swatches .Theyhavea holeon thetopfromwhichtheyarehungon a 1.5mm diametermetalrod,whichis connected to the load cell of the Instron.At the clamp,the dimensions of the cross-section of the hair swatchare2.8 cm in lengthand approximately 2 mm in wi.dth. At least two regular hand combs should be at hand for eachmeasurement. One of the combsshouldonly be usedfor cleanuntreated hair swatches. The other one should be used for the treated swatches.

Ordinaryhard rubberor nyloncombsare suitable. The measurements are doneunder standardtemperatureand humidity conditions(70 +--2øF,65-+-2 per cent RH). This requiresthe availabilityof a temperatureand humiditycontrolledroomwhereinthe tensiletestercan be locatedandoperated.

MEASUREMENTS

ON

HUMAN

HAIR

383

SamplePreparation

The preparationof swatches consists of mountingthe hair that has been selectedto be usedin the measurements, on the clampspreviouslydescribed. The uniformityof thisoperationis facilitatedby proceeding asfollows:start

by securing enoughhair for a complete set of measurements. Six swatches are recommended in routineevaluations of products. If changes in combabilityareverysmall,moreswatches mightberequiredto ascertain statistically significant changes. About10 g of hair are requi•'ed to prepareeachswatch. Althoughif couldbe desirableto useperfectlystraighthair, thisis not practical becauseit is difficultandexpensive to obtainsuchhair. Virgin European

hairwithitsnaturalsoftcurlisperfectlysuitable for thesemeasurements and should be used.

If possible, all the hairto be usedin 1 setof measurements shouldcome fromthesamebatchof commercially purchased hair.Blendingthehair is not necessary andisnotrecommended because thecoherence of thehair'snatural curl is lost,and thisresultsin excessive tangling.

The lengthof the individualhairsto be usedshouldbe uniform.A good lengthto startwith in preparingswatches is 11 in.

Measure the lengthof thehairin inches.On a toploadingbalanceweigh (to4- 0.2g) individual bundles of hair (onefor eachswatch)sothateach COMB

-:• side

top view

view

CLAMP

top view o

I

side view

1





,

INCHES

Figure 4. Top and side views of comband clamp usedin measurements

384

JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

bundle'sweightin gramsis approximately equal to its stretchedlengthin inchestimes(0.7); that is, the linear densityof eachbundleshotfidbe proximately 0.7 g/in. Handlethe bundlesgentlysothat the hair'snaturalcurl is not unnecessarily disturbed. Placcthe rootendsectionof a bundlein the throat of the clamp and dis-

tributethe hair evenlythroughout its width.Allow approximately i in of hair beyondthe throat of the clamp.Securethe male part of the clampto the otherpart by meansof the 2 screws. Proceedin the sameway with the restof the bundles.Cut the excess(root end) hair closeto the top of the clamp's jaws.Stretchthe hair swatchwith the handsand cut the tip end of the hair at a distanceof 6.5 in. fromthe closestend of the clamp.This can be done conveniently usinga laboratoryguillotine.*The amountof hair freelyhanging fromthe clampwill weigh4.5to 5.0g. The hair mountedon the clampsshouldbe cicanedin orderto remove dirt, grease, or anyforeignmaterialthatmightbe presenton the hairwhenit

is purchased. The cleaning is doneusinga 15per centaqueous solution by weightof sodiumlaurylethersulfate.The (unadjusted) pH of thissolutionis

in therangeof 7 to 8. Thesolution is liberally applied twiceto eachswatch asif it werea shampoo. Afterthe second application, the swatches arethor-

oughlyrinsedunderrunningdeionized water(at roomtemperature) for 30 min.For ourpurposes, thiscleaning treatment is morerealisticandlessarbitrarythanthecommonly usedprecleaning of hairwithvarious organic solvents.

Aftercleaning, the swatches arecombed andhungto dry andequilibrate to the standardconditions of 70øF(21.1øC)and65 per centRH for 24 hours in a controlled environment room or chamber.

ExperimentalProcedure

Theexperimental procedure consists of measuring thePCLsand/orACLs of the samehair swatches wet and/or dr,vbeforea treatmentand afterthe treatment.

Measurementson UntreatedHair-Wet Measurementsas •ollows.Step 1: turn on the Instron,allowit to warmup, and calibrate,followingits operation

manual.Step2: placeeachof theswatches to bemeasured in deionized water at roomtemperature (70øF)for at least10rainandfor no morethan30 min priorto the measurements. Thiscanbe doneby supporting the swatches by theclamps andallowing themtohangfreelyinsidea largebeaker(• 3000ml) full of deionized water.Step3: takethefirstswatchoutof thewaterandcomb it untilnoderangling isnoticedonfurthercombing. Startthisoperation using

*Harvard ApparatusCo., Inc., Dover, MA.

MEASUREMENTS

ON

1IUMAN

HAIR

385

the wide-toothsectionof a hand comband finishusingthe thin-toothsection. Step4: immersethe hair swatch3 consecutivetimesin a separatebeakercontainingaleionized water at roomtemperature.The purposeof this stepis to generatea certaindegreeof tanglingof the swatchundercontrolledconditions.This stepis very importantand care shouldbe taken to performit in the sameway eachtime.It shouldbe doneby holdingthe swatchthroughthe

damp andgentlydippingit in andout of the water3 times.After the third immersion,squeezeout the excesswater twice with the fingers.Step 5: with

the crossbarsuffieienfiyout of the way (below),hang the hair swatchfrom the load cell hook and adjustthe pen of the recorderso that the swatchplus the damp read zero.To do this,usethe balancecontrolof the recorder,which doesnot affect its calibration.Step 6: removethe swatchfrom the load cell and hang it by its damp dose to the Instron.Displacethe crossbarupward to the startingposition.At thispoint, the teeth of the combshouldbe at a dis-

taneeof 2.5 em fromthe loweredgeof the hair swatchdamp. Step7: hang the hair swatchfromthe loadcell,andusingtwofingersto fiattenthe swatch, push-guidethe hair into the comb.If done carefully,this operationinsuresa fairly even distributionof the hair betweenthe interteethspaces.Also, becausethe width of the swatchat the damp is 2.8 era,and the distancebetween the 2 outerteeth of the combis 3 era, it is simpleto have all the hair "in" the comb.Placethe thin metalrodbelowthe teeth,perpendicular to them,to pre-

ventthe hair fromcomingout of the combduringthe measurements. Step8: oncethe hair is properlyplacedin the comb,the actualmeasurement can be

started.Thisconsists of continuously recording the forcethat is requiredto movethe combdownthroughthe hair swatchat a constantspeed.This is doneby settingthe crossbar in downwardmotionwhile continuously recording the load.If a recorderintegratoris available,it shouldbe functioningso that the areaunderthe combingcurveis measured. The combingspeedwill be setat 10 em/minandthe chartspeedat 10 em/min.The sensitivity for the recorderwill be set accordingto the valuesof the forcesencountered. Step 9: repeatsteps4 through8 twicefor the sameswatchin orderto recordtriplicateruns.Step10: repeatsteps2 through9 for the remainingswatches. Step

11:afterthe wet measurements are eompleted, combthe swatches usingthe handcomb,hangthemthroughthe dampsand allowthemto dry and condition for at least24 hoursat 65 per centRH and 70øF.

Dry Measurements: Dry measurements comprise steps12 and 13. Step12: Startthe dry measurements by takingthe firstconditioned hair swatchand combingit with the hand combuntil no deranglingis noticedon further

combing. Proceed thenwithsteps 5 through 9 asbefore.Step13:repeatstep 12 for the remainingswatches.

Treatment: Givethetreatment to thehairmountedonthe clampsfol]owing the recommended instructions for the product.Use deionizedwater whenever

386

JOURNAL OF THE SOCIETY OF COSMETIC

CHEMISTS

water is needed.The amountof productto be applied to each hair swatchis

calculatedtakinginto accountthe amountrecommended for a head of hair, using100g of hairastheaverageweightof hair for adultfemales,andweight of hair in eachhair swatch.The followingformulais appliedfor thispurpose.

Weight of product (weight of hair swatch) x to apply per swatch = 100

Amount of produc

recommended for a! headof hair

If rinsingwith wateris thelaststepof the treatment,careshouldbe givento this operation.Rinsingshouldbe sufficientto eliminateexcessproduct,but not sointenseso that the effectof a productcouldbe completelyeliminated. The way in whichthe productis usedin actualpracticeshouldbe followed. For example,shampoos are rinseduntil foamis no longerevident;the same shouldthenbe accomplished with therinsinggivento thehair swatches. Once the conditionsare specified,care shouldbe exercisedin rinsingeachof the swatchesin the sameway. Rinsingconditionsshouldspecifyvolumeof water, temperatureof the water, finsingtime, and method(flowingwater or immersion). Measurementsof Treated Hair: Wet measurements on treated hair should be done right after the treatment,precededonly by a 5-rain period in which the treated swatch is allowed to relax immersed in water. The rest of

thewet treatedswatches shouldbe left hangingfromtheir clampswhile they wait for the 5-min relaxationperiod and subsequentmeasurement.The main reasonfor doingthe wet measurements fight after the treatmentis because,in practice,the hair has to be combedafter any treatment,and it is at that pointthat the userwill associate the productwith its effecton wet combability. Obviously,wet measurements can be done at a later time if this will contributeadditionalinformationon the effectof the product. Calculations

Oncethemeasurements are completedthe datarequiredto calculatechangesin combabilityare obtainedfrom the combingcurves.

The PCL for eachrun corresponds to the highestloadrecordedfor that run and is read directlyfrom the corresponding combingcurve.The load for a full-scaledeflectionusefor recordingobviouslyhasto be takeninto account. The ACL for eachrun is calculatedby first measuringthe areaunder the corresponding combingcurve (in gramscm ß units) and then dividingthe valuefor the areaby the distancein centimeters thatthe combtravelsthrough the hair in that run.This distanceis readdirectlyfromthe curve.

It hasbeenourexperience, in developing thismethod,thatper centchanges in PCL aresimilarin valueto per centchanges in ACL. For thisreason,and

MEASUREMENTS

ON Table

HUMAN

HAIR

387

I

CombabilityResultson BleachedHair Before and After the Use of a SemipermanentDye Product A

Wet Measurements70øF) Peak Combing Load (G)

Swatch Number 1 2 3 4 5 6

Before Treatment (BT) Run Number 1 2 3 700. 1500. 525. 1200. 850. 1845.

1050. 1125. 690. 1175. 1950. 1425.

1160. 925. 1000. 1210. 825. 1550.

Average Column 970. 1183. 738. 1195. 1208. 1607.

After Treatment (AT) Run Number Average Per cent 1 2 3 Column Change 980. 366. 664. 650. 436. 440.

820. 360. 422. 900. 500. 350.

320. 664. 460. 616. 480. 530.

707. 463. 515. 722. 472. 440.

--27.1 --60,9 --30.2 --39.6 --60.9 --72.7

Average BT = 1150, average AT = 553.

Per cent change = (55•50!150') X

100 = --52. per cent.

Dry Measurements(65 per cent RH, 70 øF) Peak Combing Load (G) Before Treatment

Swatch Number

RunNumber 1 2 3

After Treatment

Colnmn Average

Run Number 1 2 3

Average Percent Column Change

1 2 3 4 5

800. 416. 500. 48S. 1180.

530. 390.

784. 304.

705. 370.

236. 200.

258. 247.

220. 140.

238. 196.

--66.3 --47.0

340. 365. 800.

275. 399. 500.

372. 417. 827.

263. 275. 400.

19 •. 310. 210.

140. 255. 370.

200. 280. 327.

--46.2 --32.9 --39.5

6

620.

550.

650.

607.

]87.

155.

25,6.

200.

--32.9

Average BT = 550, averageAT = 240.

(240. -- 550.•

Per cent change: •, •.

] X100 = --56 per cent.

because theyare morereadilycalculated, the useof PCL valuesis recommended.The computations are bestillustratedby considering examples. ExampleI

The data,whichappears in TableI aretypicalandcorrespond to an experimentthat was donefor the purposeof determiningthe effectof an experimentalsemipermanent dye producton the wet and dry comb ability of bleached hair.

388

JOURNAL OF TIlE SOCIETY OF COSMETIC CHEMISTS

The columnswith headingsRun 1-3 containthe PCL valuesfor each of the 3 replicaterunsthat were recordedfor eachswatch. The averagecolumnconsists of the averagesof the runsfor eachswatch. The average before treatment (BT) and average after treatment (AT) valuesare the averagesof the averagevaluesappearingin the averagecolumns.The per centchangein PCL is calculatedusingthe expression

(Average PCL AT- Average PCL BT) x 100 =per cent change inPCL AveragePCL BT If averagecombingloadsare measured insteadof peakcombining loads, the data are treated in an identical manner. the results of the above set of measurements are summarized as follows.

Effect of experimentaldirect dye base on combability:averageper cent changein dry PCL .... 56. per cent; averageper cent changein wet PCL ....

52. per cent.

The percentchangecolumns onthe righthandsideof TableI givethe per cent changesfor the individualswatches.We have chosennot to use these numbers,i.e., their averages,to calculatethe total per cent changein PCL, dueto the treatment.Thesem•mbers, however,give,oninspection, a practical indicationof the reproducibilityof the experimentand/or treatmenteffects. Example II

In most cases,combabilitymeasurements involve the comparisonof the e.%ctson combabilityof two or more products.Even when this is not the case,a productwhoseeffect.oncombabilityhasbeenpreviouslymeasured,is normally included in the experiments.This is recommendedbecause it servesas an internalstandardwhichwill detectany bias in the resultsdue to

differences in the hair. Hair fromthe samesourceshouldalwaysbe usedin any comparativestudy.

A comparison betweenproductscanbe doneby measuring the effecton combability of eachproductindividuallyasin exampleI andthencomparing the averageper centchanges. If the effectson comb'ability of t•voproducts are sufficientlydifferent,that is, of the order of (per cent changePCL) ProductA- (per centchange PCL) ProductB • 20 per cent,andif theper cent changein PCL for most of the replicateswatchesis uniform, the validityof the observed difference frequentlycan be decidedby simple inspection of the data.If thereis any doubtand/or the data are going to be usedas part of documentation supporting claimsfor a product,the statisticalsignificance of the differences must alwaysbe established. This

is bestaccomplished by doingan analysis of varianceon the data (4). Such an analysisis illustratedas followswith data that were used to com-

paretheeffectof shampoo A (TableII) to thato.fshampoo B (TableIII) on the combability of dry humanhair. As can be seenin the Tables,8

MEASUREMENTS

ON Table

HUMAN

ttAIR

391

IV

Analysis oœVariance for Data in Table II Source of Variance

t•etween rows Between columns Row X column interaction Residual Total

Sumsof Squares 185,415.7 9,052.5 14,459.6 73,495.6 282,423.4

Degreesof Freedom m -- 1 = 7 n._,-- 1 = 1 ( n, -- 1 ) x (•n2-- 1 ) = 7 n,n,•(m-- 1) = 64 mn=na-- 1 = 73

Mean Squares 26,488.

9,052.5 2,065.7 1,148.4

equalsnumber of rows equals8 (one for each swatch). equalsnumber of columnsequals2 (one for each treatment). equalsnumberof replicationsequals5 (five for eachswatch).

swatcheswere used and 5 replicate runs were done on each swatch: It canbe seenin Table II that the useof shampoo A resultedin a decrease of the forcesrequiredto comb 7 out of 8 swatchesused.On the other hand, Table III showsthat of the 8 swatchesthat were treated with shampooB, 3 showed an increase and 5 a decrease in their PCLs.

Analysisof Variancefor Data in TableII. The statisticalparametersneeded to performthe analysisare shownin Table IV. The followingoperations are doneto determinethe significance of the differentcomponents of variance. Step1. Significance of the interaction(betweenrowsand columns)against the residual:2,065.7/1,148.4= 1.79. For degreesof freedom (dr) Nt = 7, N2 = 64 the aboveratio is not significantat the 95 per cent confidence level. (Fishervarianceratio test.) This meansthat the data do not showany detectablestatisticallysignificantinteractionbetweenthe treatmentand the PCLs of the swatches.If the interactionhad been significant,it would indicatethat the effectof the productis a functionof a characteristic of some of the swatches, in our casetheir initial beforetreatmentaveragePCL. This seldom,occurs if all the swatches are preparedfrom the samehomogeneous batchof hair. If it does,it indicatesinhomogeneity of the hair, m•d the best solutionis to preparemoreswatches and excludefrom the set of swatches thosethat haveextremelyhigh valuesfor their initial beforetreatmentPCL.

This shouldbe doneon the completeset of swatches participating in the experiment.The set would then be randomlydivided in half into subsets to be usedwith eachproduct. Step2. Poolingof the sumsof squaresof the interactionand residualand

theirdegrees of freedom:(14,459.6 q-73,495.6)/(7q-64)= 1,238.8.This numberis nowtreatedasa newmeansquarefor theresidual. Step 3. Significance of the variancedue to differencesbetweencolumns (i.e., due to shampooA treatment).The value of the ratio of the mean square of the between columnsterm and that of the new residual determines

the significanceof the "betweencolumns"varim•ce:9,052.5/1,238.8= 7.31. For degreesof freedom N• = 1 and N2 = 71 the value of the ratio indicates

392

JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Table

V

Analysisof Variance for Data in Table III

Sourceof Variance Letween rows

Sumsof Squares

133,028.0

Between columns

32.5

Row X columninteraction Residual Total

13,401.0 59,972.0 206,433.5

Degreesof Freedom n• -- 1 = 7 n,, -- i = 1

( n•- 1 ) x (,n•-- 1 ) =7 n,n,,(m-- 1) = 64 mn.,n•-- I = 73

Mean Squares

19,004.0 32.5

1,914.4 937.1

nt equalsnumberof rowsequals8 (onefor each swatch). n• eqnalsnumberof columnsequals2 (one for eachtreatment). n• equalsnumberof replicationsequals5 (five for eachswatch).

statisticalsignificance at the 99 per centconfidence level.This meansthat the decreasein PCL measuredexperimentally(-15. per cent) can be considered statisticallysignificantat the 99 per cent confidence level. Analysisof Variancel•orthe Data in Table III. The parametersneededto performthe analysisappearin Table V. Step 1. Significanceof the interaction:1,914.4/937.1= 2.04. For N• = 7, N•o= 64 the interactionis not significantat the 95 per centconfidence level. Step 2. Poolingof sumsof squaresand degreesof freedom:(13,401.+ 59,972.)/(7 + 64) = 1,033.4. Step 3. Significanceof the variancedue to between columnsdifferences: 32.5/1,033.4= 0.31. For N• = 1, N2 = 71 is not significant.

Theseresultsconfirmthat the changes noticedon combingloads(Table III) after the use of shampooB do not indicateany effectdue to this treatment.

The analysisof varianceprovidesus with a criteria to establishthe statistical significance of the observedchangesin combingforces.In order to compare the effectof shampooA to that of shampooB we canproceedasfollows.

Stepa. Calculatethe difference betweenthe averageafterandaveragebefore treatmentvaluesfor eachshampoo, i.e.,

averagedifferencefor shampooA = D^---116.8- 138.=-21.2 g average differencefor shampooB--D•, = 124.5- 123.2= 1.3 g Stepb. Calculatethe standarderror(standarddeviation)of the difference in the means,i.e., Dx and D, using the formula (5)

Standard error-

rr x N/2: O'm ¾n

MEASUREMENTS

ON

HUMAN

HAIR

393

•vhere rr is the common standard deviation of each of the before and

after treatmentmeans,and n (40) is the number of observations used to calculate the means. In our ease, rr will be given by the square root of the residualmean squarecalculatedin Step 2 of the analysisof variance. We will have the following: O'm,for shampooA --

for shampooB =

35.2 x 1.41 6.32 32.1 x 1.41 6.32

= 7.85

= 7.16

If desired, these numberscan be used to esti•nateconfidencelevel litnits

(q- L) for the differences, i.e., +L = t x O'm.The residualvarianceused to calculatethe standarderrorshas74 degreesof freedom.The corresponding value for t for the 95 per cent confidencelevel is 2.0, hence +Lx = ---2. x 7.85 = -T-15.7

-----LB = +2. X 7.16 = -----14.3

andthe changes in combingforcesfor shampoos canbe expressed asfollows: Changein PCF shampoo A = -21.2 q- 15.7g Changein PCF shampooB =

1.3 ---14.3g

Stepc. In orderto calculate thesignificance of the difference betweenthe twoaverage differences foreachshampoo we performa t test.Thevaluefor t is givenas follows: t=

Dx - DB

nt X n2

O'e '• n• + n2

in which o'eis the combinedstandarddeviationobtainedby combining

andO',•b. According to theexpression o-e• = o',• 2 x df•. + o',•b2X dfb ----1151.7 df•. + dfb --2 O'c= 33.9

where the degreesof freedom= 74. Also, n• = n• = 40. The calculatedvalue

of t equals2.97.The valueof t foundon a t-tablefor the 99 per centConfidencelevel is 2.6. The differencebetween changesin PCL producedby shampoos A and B is thus shownto be significantat the 99 per cent confidence

level.

394

JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

In the presentexample,the analysisof variancewouldhavebeensufficient to demonstratethe superiorityof shampooA over B, becauseit showedthat

shampoo A had a significant effectwhile shampoo B did not. In othercases, however,if both productsare shownto have a significanteffect,the calculationsunderstepsa to c leadingto the t-testare requiredin orderto provethe superiorityof a productoverthe otherone. APPLICATIONS

Effect of commercialhair productson combability.Table VI showsthe effectof a selection of commercial products onthe combability of originallyuntreatedhumanhair. It canbe noticedfrom the table that mosttypesof hair products, if formulatedcorrectly, canimprovethe combability of humanhair. Effectof quaternaryammoniumcompounds on wet combability.The effect of quaternaryammoniumcompounds on the combabilityof humanhair is well known(6). Figure5 showsthe effectof increasing amounts of dodecyltrimethylammonium chloridesorptedby bleachedhair on its combability. The hair was bleachedfor 60 min usinga commerciallighteningproduct.It was then treatedby immersingit in 0.05 g/100 g aqueoussolutionsof the quaternaryfor 0, 1, 5, 60, and 19.0min at roomtemperatureand then rinsing

for 20 secunderrunningdeionizedwater.The amountof quaternaryon the hair was determinedby extractionwith chloroformand further analysisusing the methodof G. V. Scott (7). Table VII showssimilardata for a set of quaternaryammoniumcompounds(8). The two levelsof uptake were pro-

T,able VI

Effect of Commercial Hair Productson Cmnbability

Product

Per Cent Changein Peak Combing Force Wet Dry

Regular shampoo ConditioningshampooA ConditioningshampooB

q-

-- 57.%

+ 31.%

Leave-in

--

4.% 23.%

-- 19.% --

27.%

creme rinse

--

40.%

--

72.%

Rinse-off creme rinse

--

69.%

--

48.%

Semipermanentdye product Oxidationdye productA Oxidationdye product13 Oxidationdye product13plus a r•nse-offcremerinse Lightener (15-min treatment) Lightener (60-rain treatment) Lightener (60-rain treatment) plus,a conditioner Conditioningsettinglotion

--

3S.%

--

34.%

--

55.%

--

15.%

q-741.%

q- 47.%

--

--

65.%

25.%

q-265.% q-760.% q- 180.%

+ 20.% +110.% -- 20.%

--

--

63.%

72.%

MEASUREMENTS

UPTAKE

ON

[ M•"a I G

HUMAN

HAIR

395

HAIR)

Figure 5. Per cent reduction in wet PCL as function of uptake of dodecy]trimethy]ammonium chloride on bleached hair

ducedby immersingthe bleachedhair in 1.0 and .05 g/100 g aqueoussolutions of the quaternariesfor 5 min at room temperatureand then rinsing 20 secaswaspreviouslydone. The data showedthat only very small amountsof thesecompoundsare neededon the hair in orderto producesignificanteffectson combability,and

that increasingthe uptakebeyondthesevaluesdoesnot resultin additional benefits.It shouldbe kept in mind that the uptakeof thesecompounds by hairis goingto be affectedby thepH of themedium,thepresence of other compounds in solution(especially anionics), andthe type of hair.

Table

VII

Effectof Uptakeof Quarternary Ammonium Compounds ontheCombability of Wet Bleached Hair

1.0 g/100 g solution5 min

Compound Tetradecyltrimethylammoniumchloride Decyltrimethy]ammoniumchloride Stearyl dimethylbenzylammoniumchloride Benzyltrimethy]ammoniumchloride Distearyl dimethylammoniumchloride

.05 g/100 g solution5 min

Uptake Percentreduc- Uptake Percentreducmg/g Hair tionin PCF mg/g Hair tionin PCF 5.9

94. per cent

.2

92. per cent

9.0

83. per cent

1.5

78. per cent

2.5

90. per cent

.8

90. per cent

4.0

54. per cent

.25

50. per cent

3.6

92. per cent

.2

9'2.per cent

396

JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS DISCUSSION

In principle,combabilitymeasurements are very simple.In practice,unless

greatcare is takenin the preparationtreatment,handlingof the swatches, andstatistical analysis of the data,the resultscanlosesignificance. This is especiallytrue when establishingsmalldifferencesbetweenproducts.The main problemarisesfrom the fact that it is difficultto producea reproducibledegreeof tanglingof the swatchesprior to the measurements. In our experience,

if the methodis followedcarefully,changesin combabilityof the orderof + 20 per centcanbe accuratelyestablishedwithouthavingto measurean impracticallylarge numberof swatches.Changesof this order appearto be closeto the lower limit of what can be subjectivelynoticed by combing swatchesby hand. In this method, the before treatment measurementsare done on hair which

hasbeencleanedusinga detergentsolution. Thisis perfectlyjustifiedin testing mostproducts because, in reality,shampooing usuallyprecedes theuseof mosthair carepreparations. In the caseswherethe effectsof shampoos are beingmeasured, it couldbe arguedthat the startingpointshouldbe unclean hair which,in manycases,hasbettercombabilitythan cleanhair. This approach,however,will introducethe unnecessary complication of havingto arbitrarilydefineand reproducibly simulatedirty hair in the laboratory.Althoughthis couldbe done,we considerthat it is justifiedto start with clean hair and defineany effectthat a shampoo canhaveon combabilityas those effectsthat canbe measuredin addition,and beyondthe effectproducedby shampoos by virtueof justcleaningthe hair. In ourmethod,we choseto quantifycombabilityin termsof PCL and ACL. In particular,PCLsare relevantin termsof whatis experienced subjectively

whilecombing hair.Thisis notonlybecause theycorrespond to thehighest forces, butalsobecause, asshown by thecombing curves, theyoccurabruptly. This characteristic of combingforceshasbeenusedby Wedderburnand Prall(3) asthebasisfor developing a methodfor measuring combability. It is likely, aspointedout by thesoauthors,that fast short-termfluctuationsin

the combing forces,i.e., "tanglenoise"or "raspiness" (2), area factorcontributingto the subjective perceptionof combingresistance. The inclusionof this effectin the evaluationof combability wouldbe mostcriticalin cases wherethe effectof two products whichproducecombing curvesof similar PCL, but dissimilar "raspiness" werebeingcompared. This situation, however,hasnotyet arisenin ourexperience, andwe findthatqualitatively PCLs increaseor decreasesimultaneously with the noiselevel in the combing curves, i.e.,smoothcurvesgivesmallPCLs,whilescratchy curvesgivelarge

ones.Short-term variations in combing forcesthusappearto givesimilarinformation to thatgivenby PCL.In the absence of a detaileddescription of the Wedderburn-Prall method,it is not presently possible to do a fair corn-

MEASUREMENTS

ON

HUMAN

HAIR

397

parisonwith ours.The methoddevelopedby Newman et al. (1) is similarto ours. It involves the insertion of a comb into a swatch of hair and the measure-

ment of the forcesopposingits motion.The authorsindicate that after less than a secondof comb motion at a rate of 1.5 mm/sec, the combingforce reachesa nearly constantvah,e, xvhichis measured.This contrastssharply

with the shapeof our combingcurves(Figs. i and 2), and indicatesthat thesemeasurements are beingdoneon hair swatchesthat, either becauseof their sizeand geometryand/or becauseof the way in whichthey are handled

priorto the measurements, do not get tangledbeforeand/or while they are beingcoinbed.This approach, althoughdesirablefromthe pointof view of improvingreproducibility,is not favoredby us,becausein reality,tanglesare almostalwaysencounteredwhile combinghair, and detanglingis an integral part of the functionof productsdevelopedto improvethe combabilityof hair. Combabilitymeasurements have been reportedto be in use at Hoffman-

LaRocheCo. Laboratory,but the methodhas not been describedin detail. The methoddevelopedby Waggonerand Scott(2), whichinvolvesthe measurementand analysisof the soundfrequenciesgeneratedwhen hair is combed,althoughit is an interestingapproach,suffersin our opinionfrom the unnecessary complications introducedby the necessityof havingto interpret

the generation of soundin termsof combing frictionalforceswhichcanmore easilyanddirectlybe measured in the firstplace. CONCLUSION

Theincreasingly sophisticated productdevelopment takingplacein ourindustryisin needof methods whichcanobjectively measure theeffectof productson humanhair. Althoughthesemethodsdo not completelyreplacethe subjective evaluation of a product's performance, theyareobviously of great helpin guidingresearch andin substantiating performance claimsmadefor hair products. Furthermore, by providingquantitative datathesemethods openthe doorto the investigation of the underlying physico-chemical phenomenainvolved in the modificationof human hair for cosmeticpurposes.

The combability methoddescribed in thispaperhasprovenveryusefulin ourlaboratory. The authors hopethatotherswill benefitfromitsuse. ACKNOWLEDGMENT

The authorswishto thankDr. MichaelWongfor assisting with the statisti-

calanalysis of thedataandforhisvaluablecomments duringthedevelopment of the method.Also,we want to thankDr. Lee Hunter for valuableinsights into the physicsof combing. (ReceivedApril 30, 1975)

398

JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS •EFEI•ENCES

(1) W. Newman, G. L. Cohen,and C. Hayes, A quantitativecharacterization of combing force, J. Soc. Cosmet.Chem., 24, 773-82 (1973). (2) W. C. Waggoner and G. V. Scott, Instrumentalmethod for the determinationof hair raspiness,J. Soc. Cosmet. Chem., 17, 171-79 (1966). (3) D. L. Wedderburnand J. K. Prall, Hair productevaluation:From Laboratorybench to consumerand back again, J. Soc. Cosmet.Chem., 24, 561-76 (1973). (4) K. A. Brownlee,IndustrialExperimentation,4th ed., ChemicalPublishingCo., Inc., 1953, Pp. 106-8.

(5) K. A. Brownlee,IndustrialExperimentation, 4th ed., ChemicalPublishingCo., Inc., 1953, Pp. 35 & 93.

(6) G. V. Scott,C. R. Robbins,and J. D. Barnhurts,Sorptionof quaternaryammonium surfactantsby human hair, J. Soc. Cosmet.Chem., 20, 135-52 (1969). (7) G. V. Scott, Spectrophotometric determinationof cationicsurfactantswith OrangeII, Anal. Chem., 40, 768-73 (1[}68). (8 M. Wong, Clairol ResearchLaboratories,unpublishedresults.

Suggest Documents