J. Soc.Cosmet.Chem.

,27, 411-431 (September1976)

Birefringence:
Polarization
Microscopy
asa QuantitativeTechniqueof

HumanHair Analysis
ROGER K. CURTIS, B.A. and DON R. TYSON, B.S.*

PresentedMay 29, 1975,SCC Seminar,St. Louis,Missouri

Synopsis:An alternative to the conventionalmethodof mechanical stress-strainanalysis
of HUMAN HAIR conditionis presentedin this paper. NUMERICAL BIREFRINGENCE
is an extremely sensitivemeasure of molecular orientation. As such, this technique has
the potential of determining hair fiber condition at a fundamental molecular level. Basic
theories of POLARIZATION MICROSCOPY are presented and utilized as the basis
of a qtmntitativetechnique developedfor the measurementof birefringencein hair. The
theories,morphologicalorigins, and contributionsof both the intrinsic and form birefringencecomponents,and the correlationof numericalbirefringencewith the mechanical
propertiesof hair are discussed.Numerical birefringence,a quantitativemeasureof the
optical anisotropicproperties of a hair fiber cortex, as a reflection of the hair strand
conditionpresentlyobservedwith mechanicalstress-strain
testing, is demonstrated.
INTRODUCTION

The traditionalmethodof deteminingthe conditionof a humanhair fiber

is by measuringits mechanicalstress-strain
characteristics.
The parameters
determined,including:Young'smodulus;forceat yield point;ultimatetensile
srength;and either break point extension,or variousparametersof elastic
recoveryhysteresis;
(1) are abstractor secondary
effectsof the basicchemical-molecular
occurrences
in cosmetictreatmentand conditioning.
Wha is
neededis an additionalsystemoi analysis,
whichby lookingat hair on a foundationmolecularlevel, is morespecific.
Crystalline,or molecularchain-sheetsubstances
of a nonisometricnature,

exhibittheopticalpropertyof birefringence
whentheyareplacedin a fieldoi
plane polarizedlight (2). This phenomenon,
due o a conditionknown as
*RedkenLaboratories,Inc., Van Nuys, California91411.
411

412

JOURNAL OF THE SOCIETY OF COSMETIC

CHEMISTS

optical anisotropism,
is a result of electronorbit polarizabilityand/or a refraetive index differentialbetweena crystaland its surroundingmedium in
which the crystalis embedded(2, 3, 4). The quantitativemeasureof birefringence,referredto asthe numericalbirefringence,
is an extremelysensitive
measureof optical anisotropism,wh'ch arisesfrom molecular orientation
5).
The cortical region of a human hair is optically anisotropic(3, 6). In its
functionas the maior fiber component(7), the cortexcontributes92 per cent
to the elasticityof the hair. This is borneout by an analysisof the correlation
matrix of a multiplelinear regression
model that explainsthe factorscontributingto elasticityin hair (8, 14).
In the course of a cosmetic treatment, be it chemical or mechanical in

nature,the conditionof a hair fiber is changed.The parameterof elasticity,

or the resistanceto and recoveryfrom deformationby force, plays a maior
role in the final fiber conditionafter treatment(1).The changeof condition
duringcosmetic
treatmentis dueto molecularbondingchanges,
whichoccur
mainly in the cortex(7, 9).
Therefore,the action of cosmetics,which affect conditionand elasticity,
shouldalsoaffectthe opticalanisotropism
of the cortex.Thus numericalbirefringence,as a very sens:tivemeasureof molecularorientation,emergesas
havingthe potentialto very accuratelydeterminehair condition.
This paper presentsa quantitativesystemof analysisof humanhair condition, baseduponthe opticalphenomenon
of numericalbirefringence.
Theory of Bitefringe
Plane polarized light is utilized to observethe phenomenonof birefringence.Ordinarylight vibratesin waves,travelingin randomplanesperpendicularto the directionof propagation.
By placinga polarizationplatein front
of a light soume,only thosewavestarvelingin planesparallelto the axisof
polarizationof the plate are transmitted;the othersbeing absorbed(Fig. 1A).
An anisotropicmaterialhas two unique opticalpropertiesas follows:(1)
any oneplaneof light wavesstrikingan anisotropicmaterialis split into two
wave-planes,
traveling90 relativeto eachother;and 45 eachrelativeto the
originalplaneof propagation(Fig. i B). Hencethe term doublerefractionor

birefringence;
(2) a planeof light passingthroughan anisotropic
material
encounters
a path of a differentrefractiveindex,and thustravelsat a different
veloc'ty,in eachdifferentdirectionof traverse(2, 5).
Therefore,a wave of planepolarizedlight strikesan anisotropic
material,is
split into two waves,oneof which,the ordinaryor simplyfastwave,is traveling througha path of lesserrefractiveindex,andthusfasterthanthe extraordinaryor slowwave,whichis travelingat a perpendicular
anglein a more
difficultpath of higher refractiveindex (2, 5).

POLARIZATION

ORDINARY

LIGHT AND PLANE POLARIZED

LIGHT

AXIS OF POLARIZATION

PLANE POLARIZED
LIGHT PASSING
THROUGH AN

ANISOTROPIC

LIGHT

'

DIRECTION Of PROPAGATION

PLANE
POLARIZED
LIGHT

MEDIUM
FAST
WAVE

POLARIZER

ORDINARY

413

MICROSCOPY

ANISOTROPIC

WAVE

MEDIUM

2 WAVE PLANES
90 TO EACH OTHER
45 EACH TO THE
PLANE POLARIZED
LIGHT

Figure 1. (A) Ordinarylight wavesbeing polarized;and (B) ordinarylight wavesfurther

split into two perpendicularwaves,each travelingat differentvelocity,while traversing anisotropicmedium

The differencein refractiveindex betweenthe two pathsof the fast and

slowwave (n,o-nl) equalsthe numericalbirefringence
(2).
As the 2 wavesadvancethroughthe material,the slowwave increasingly

lagsbehindthefastwavein directproportion
to the numerical
birefringence
(the difference
in path velocities)
and to the distanceor thickness
traveled.
This quantityof "lag,"measured
in unitsof distance,is termedretardation.
Hence the formulation,retardationequalsnumer_'cal
birefringencex thickness;or A= (n2-nl)d. The numericalbirefringenceis determinedby
measuringthe retardationand the thickness(2).
The retardationlag of the slowrelativeto the fast wave causesthe two
wavesto go out of phasewith eachother(Fig. 2). Thesetwo waves,exiting
the anisotropic
material,interferewith each other, causingan elliptical
interference
pattern.In Fig. 3(A) the two wavesare comingat youin phase
with eachother,the retardationdistancehavingbeensomefull multipleof

the wavelength
(A= nX).They are vibratingin unison,i.e., eachreaching
points1-20at the sametimefrompointsA, A' to B, B'. The ellipticalinterference
patternis centered
aroundanaxis,formedby connecting
thepointsof
intersectionof consecutive
lines,drawnfrom the wave positionat a particular

instant,
perpendicular
to thewaveaxis.In Fig.3(B), theslowwavehasbeen

414

JOURNAL OF THE SOCIETY OF COSMETIC

CHEMISTS

INCREASING RETARDATION OF THE SLOW RELATIVE TO FAST WAVE

WAVE

s,ow/w-,,

WAVE/

\_/

,,_//,

_'g,X
RETARDED

k../

1/4),RETARDED

1/,,X RETARDED

1 ),RETARDED

('INPHASE]

(IN PHASE)

DIRECTION

'

OF RETARDATION

DIRECTION
OF
TRAVEL

Figure 2. Two parallel waves of equal length X, traveling

different velocities.As slow
wave increasinglyretards from fast wave, two go in and out of phase with each other

/I"RESULTANT

/ ELLIPTICAL
,

PATTERN

/(NEGATIVE
IMAGE)
/

''

/AXIS

1 INTERFERENCE

ELLIPTICAL

/""INTERFERENCE

t J
;

.,

"
../_

AXIS
OF

POLARIZATION

{gure 3. OdettJo o e]]tJe] Jteeegee tte

ome by two Jteedg
wves; (A) pase; a () 2X o.t o pse (see tezt)

retardedsomefull multiple of the wavelengthplus one-half(A = n + 1/2X).

In this case,as the fast wave advancesfrom point A, throughpoints1-20,
finallyto pointB, the slowwave,being1/2X
out of phase,advances
frompoint

POLARIZATION

MICROSCOPY

415

Figure 4. (A) Axis of polarization;(B) slow wave axesof compensator

and hair; (C) fast
wave axesof compensator
and hair. In wave axes of hair and backgroundretardation
compensator:
parallel alignmentproducessummationof their retardations-perpendicular
alignment produces subtraction of their ret.ardations

B', throughpoints20-1,finallyto pointA'; the ellipticalinterference

pattern
hasshifted90 (2).
An importantconcept,utilized in the measurement
of numericalbirefrin-

gence,isthatof additionandsubtraction
of retardation.
If anisotropic
objects
areplacedin line (i.e.,oneontopof another)
in thepathof lightof a polarizationsystemwith their slowwave axesparallel,the total amountof retardationin the system
equalsthe stLmof the retardation
of the individualObjects.
If, however,
theslowwaveaxesarealigned
perpendicular
to eachother,the
totalretardation
is equalto the difference
betweenthe objectsasthey tend
to canceleachotherout (Fig. 4) (2).
The anisotropic
materialbeingilluminatedas hasbeendescribed
(Fig.
i(B)) is viewedthrougha secondpolarizingplate, termedthe analyzer,
whoseaxisof polarization
is 90to thatof the"bottom"
polarizer.
Onlywhen
a portionof the ellipticalinterferencepatternliesin the sameaxisasthat of
the analyzer,doeslight pass;thus,light dueto retardation,
whichin turn is
caused
by theproductof numerical
birefringence
andthickness,
canbe seen
andmeasured.
No lightis passed,
a condition
referredt.oasextinction,
when
A= n)t;maximtun
lightispassed
whenA= n + 2)t
(Figs.5 and6) (2,5).

416

JOURNAL OF THE SOCIETY OF COSMETIC

CHEMISTS

OF
ANALYZER

OF
ANALYZER

AXIS OF
POLARIZATION

A=n+l/2,

A--n,

NO LIGHT-- EXTINCTION

MAXIMUM

LIGHT

Figure 5. Fast and slow waves exist anisotropicmedium to form elliptical interference
pattern: (A) /x = n X: the interference pattern is aligned 90 to the analyzer axis; no
light can pass; (B) _X= n q-% X: the pattern is in alignment with analyzer axis; maximum light is transmitted

(.

....i
:.:.
.... ,.

.;.

.
??3 ;:

':;".... S...

.. :.

:...=:
....

.:.,. ::-.:

-.

::.

..

..

.'

. ....

.C...?d:
%

. :.:
-..-. --.:.:.;.
:....Ce

(*)
()
Figure 6. (A) () Light passingthrough monocometer of wavelength X; () polarizer; (c) 546 nm retardation compensator;() analyzer. (A) compensatorcauses1 X of
retardation;no light; (B) [2 compensators
(c) are stackedtogetherequaling 1092 nm
retardation]. Comensatorcauses1% x of retardation;m,imu]n light ansmissivn

Until now, only monochromatic

light, or light of one particularcolorand
havingonly one wavelength,hasbeen considered
in the theoryof birefringencethat hasbeenpresented.
White light is madeup of a combination
of all
wavelengths
of light in the visiblespectrum,from about400 to 700 nm. The
samephenomenaof birefringenceand retardationoccur,as has been previouslydescribed,for monochromatic
light with one exception:retardationof a

particulardistance
valuewill causesomewavelengths
or colorsof lightto be

POLABIZATION

MICBOSCOPY

417

Figure7. (A) Wedgeof aniosotropic

quartzcausing
increasing
retardation
from0 an
to'1.650tm,Noterepeating
clarkbands(representing
orderof retardation)
at intervals
of thex of whitelight:550nm.(B) [Adapted
withpermission
of thepublisher
(2)],
fromMineralOptics:Principles
andTechniques
by Win. RevellPhillips,
W. H. Freeman
andCompany,
copyright
1971.Retardation
of particular
valuetransmitting
eachcolor
Xdifferentially.
Thesummation
of theseXscauses
repeating
series
of colors

418

JOURNAL OF THE SOCIETY OF COSMETIC

CHEMISTS

in phase,someonc-halfout of phase,and othersout of phasein varyingdegreesin between.Each of the individualwavelengths

will, therefore,passthe
top analyzerin a differentialamountwhere they summate,giving a single
color or wavelength.Figure 7 is a black-and-whiterepresentation
of this
phenomenon.
For a colorinterpretationof the summationof retardation,refer
to Michael Levy'sbirefringencechart (9,).
There are two types of birefringencewhich occurin an anisotropicmaterial: intrinsicand form (9,,3, 11). Intrinsicbirefriengence
is a functionof the
anisotropicpolarizability (nonisometricmolecular orientation) of electron
orbits, and not dependentupon any particular morphology.Form birefringence,however,occurswhen crystallitesof one refractiveindex are immersed
in a medium of anotherrefractiveindex. Intrinsic birefringenceoccurson a
molecular
level,whereasformbirefringence
occursmoreon a macromolecular
or structurallevel. The measuredquantity, numericalbirefringence.is the
stunmarion
of the intrinsicbirefringenceand form birefringence(2, $, 4).
EXPERIIVIENTAI.

Apparatus

Microscope:A ReicherrZeropanresearchmicroscope
is outfittedfor polarizationasfollows(Fig. 8(A)):
(1) 100 W ouartz-iodinelamp house*; (2) strain free, 0.95 N.A., dualdialhramcondenser;*($) calibrated-rotating
polarizerand analyzer;*(4)
25 X. 0.60 N.A. NeofluarN.A.* obiective,checkedto be strainfree; (5) KPL
8 X Pol Occular*, modifiedto containa measuringreticle; (6) 360 rotating
stage*; (7) Gips Rot i Ord Compensator*(Fig. 8(B) ); and (8) Ehringhaus
Compensator
with ouartzplates*(Fig. 8(B) ).
1totaru Device: This device enables a hair fiber to be rotated 360 on its

own axisfor microscopic

viewingand to evaluatestructureand measuredimensions(Fi N. 8(C)). Two 25x75 mm glassmicro slides (i.e., Corning
2948++).one cut longitudinallywith a standardlass cutter, are neededto
make ] reuseabledevice.The 3 piecesare held togetherwith an epoxy
resinglue. It is necessary
to use only high quality lass is all parts of the
svstemas any inherentstrainpresentwill cause'%ackround"
birefringence.
A capillary tube ~ 0.5 mm i.d. ~ 1.0 mm o.d.. and 75-90 mm long (i.e.,
ScientificProductsB4195-2'*), to which a short piece of plastic tubing
(Comln Silastic$medicalgradetubing 0.035in. i.d.) is attached,holds
the hair for rotation.

The capillarytube is filled with immersionoil, n, = 1.515,of a light

to mediumviscosity(Carl ZeissImmersionsoelP).
The holdinz channelis
fillod with oil of the samerefractiveindex,but of a higher viscosityto dampen tube movement(i.e., ZiessEinschluBmittelW15*). With a coverslipin

place (Coming 22 x 40 mm Number1-%$), a nearlyhomogenous

refractive
*ReicherrDivisionof AmericanOptical ScientificInstruments,Buffalo,N.Y. 14215.
*Carl Zeiss, Inc.. New York, N.Y. 10018.
Cornng GlassWorks, ScientificGlasswareand EquipmentDept., Corning,N.Y. 14830.
**Scientific Products,McGaw Park, Illinois.

POLARIZATION

!:.

PLASTICCOVER:
SLIP

?-': CAPIL
'LAB.:':'TU
BE
;"

MICROSCOPY

HAIR

............ '::t'""::

419

25min.
...
. lmm::

:'CApIE..-LAB't

e '.'l'::::teW:'

.ROTARY:DEVICE

Figure8. Reicherr
Zeropan
microscope
aplaratus
for determination
of numerical
bire-

fringence
in hair:(A)microscope
with(a)polarizer,
(b)strain
freecondenser,
(c)rotary
stage,
(d)rotary
device,
(e)strain
freeNeofluar
obiective
, (f) analyzer,
(g)com,pensator,
(h)quartz-iodine
lightsource;
(B)detail
ofquartz
firstorder
red(i)andEhringhaus
rotary
quartz(j) compensators.
[Noteslowxvave
axisorientation
(k).];(C) detailof hair

rotary device

4g0

JOURNALOF THE SOCIETYOF COSMETICCHEMISTS

indexpathhavingplaneoutside
surfaces
is formed.Thisarrangement
vir-

tuallyeliminates
anydistortion,
dueto theconvex
surfaces
of thecapillary
tube. Anotherversionof this devicehas beenutilized in previousstudies
(12).
Stress-Strain

Tester

In thisstudy,
stress-strain
measurements
areusedasa reference
to which
thenumerical
birefringence
is compared
andcorrelated.
A single
hairfiber,
1.9cmlong,issuspended
between
a setofclamps.
Force,
applied
tooneend
by a constant
speed
motor,
elongates
thefiberat a rateof 1.5percent/sec,
;vhilebeingmonitored
ontheotherendbya straingauge
transducer.*
Stress
versusstraingraphsare plottedon XY Recorder.t
Procedures

1. Microscope
Alignment;
Withthemicroscopic
system
havingachieved
a
condition
of Kohlerillumination,
thepolarizer
andanalyzer
areinserted
and
rotatedto the correctposition.
Thisis bestdoneby firstrotatingthebottom
polarizer
to a position
of eitherqb,90, 180,or 270,andthen,whilefocusing
on anilluminatedmicroslide,
by rotatingthe top analyzeruntil maximumex-

tinction
(minimal
lighttransmission)
hasbeenreached.
If thefixedfirst-order
redcompensator
plateis nowinserted
intothecompensator
slot,a deepred
background
will appearin themicroscope
field.Thisis indicative
of a N-S,
E-W,polarizer-analyzer
alignment,
witha diagonal
slow-fast
wavealignment
of the anisotropic
compensator
(2, 5).

2. Measurement
o[ NumericalBire[ringence
in HumanHair Shafts:
To recall, retardation
equalsnumerical
birefringence
xthickness
(A=(n2-n)d).
In tryingto determine
thenumerical
birefringence,
thiscanbe rearranged
to
read:numerical
birefringence
equals
retardation/thickness
((n2- n)--A/d)
(SeeFig. 11(B)later on) (2).

A hairshaftof approximately
4 cmlengthis inserted
intothe glasscapillary
tube,intowhichthe mediumweightimmersion
oil is dra;vnfroma reservoir.
Thecapillarycanthenbe pluggedandstoredin a microhemocrit
tubesealerholder.

*Stathan Instruments,Inc., Oxnard, California 93030.

'Hevlett Packard,Inc., Palo Alto, California 94303.

POLABIZATION

MICBOSCOPY

421

Figure 9. (A) Scanningelectroninicrographshoxvingelliptical axes of human hair; (B)

hair inside rotary device for neasurementon its narrow axis; and (C) on its wide axis.
Blackline is edgeof capillarytube (arrow)

To beginthe actualmeasurement,
with the rotarydevicein placeon the
microscope
stage,a few dropsof the heavyweight immersionoil is inserted
into he slot, togetherwith the loadedcapillarytube, and a coverslipis
placedon top. This basicprocedureis repeatedeachtime a 'hairis viewed,
as miniature air bubbles,churnedby capillary rotation, becometrapped

underthe coverslip,
disrupting
thehomogeneous
refractiveindex,introducing
distortion and loss of resolution.

The rotarymicroscope
stageis positioned
sothat the hair is alignedwith its
longitudinalaxisparallelto the slovwave axisetchedon the firstorderred
compensator.

In this position,the sloxvwave axisof the hair is in the sameaxisas that of

the compensator,
givingthe hair addedretardationprovidingenhanced
viewing abfiity.
Hair is essentiallyellipticalin shape,with a wide and narrowaxis(Fig.

9(A)).The narrowaxisis equivalent

to thehickness
whenthe hair is resting
onitswideaxis,theposition
usedfor evaluating
thenumerical
birefringence
of the fiber.

The hair is placedexactlyon its naxxowaxis.At that point,bothhair edges

4-22

JOURNAL OF THE SOCIETY OF COSMETIC

:':"
":'"'

CHEMISTS

-.

i ::--!"!..:
..::".'..
'""7.'"
........
?..'..
....
Figure 10. Determinationof retdation with Einaus rotary compensator.
(Hrs
B, C, D, and E are oriented90 to the compensator.
) In seriesfrom A rough E,
colorsin hair go downin orderi.e., in leR to right directionon summation-interference
colorchart(Fig. 7). (A) hair oriented
with quar firstordercompensator;
(B) Ehringhaus
compensatorset at (note extinctionbackground);(C) retardation added to field by
compensator
with equM amount subtractedfrom hair; (D) increasingretardationfrom
compensator
bringscolorbandsto backgroundand decreasedhair centercolor; (E) 1
of retardation of hair has been subtractedout in centered black extinction"arrov."
From degree of compensatorrotation, retardationcan be cculated

will appearsharp,theretardation
colors
villbe moreor lesssymmetrically
runningdowntheshaft,andthehairwill be at its thinnest
point(Fig.9(B)).
Recordthemeasurement
of theaxiswith theviewingreticle.
Thehairis thenrotated90whereagain,.theedgeswill be sharp,the retardationcolorssymmetrical,
andwill nowbe at its widestpoint( Fig.9(C)).
A measurement
of the hair wide axis"diameter"in thisposition,togetherwith
thatof the narrowaxis,canbeput intothe formulafor the areaof anellipse,
0.5 wide x 0.5 narrowx rr, to determine the cross-sectional
area. This infor-

mationis usefulin swellingand/or stress-strain

analysis.
Next, the retardation must be measured to calculate the numerical bire-

fringence.
To dothis,thequartzrotarycompensator
isputin placeof thefirst
orderred.The hair is to remainon its wide axis,alignedas bove.Whenthe
rotarycompensator
is set at 4>,it is lying fiat and hasno effecton the system.
When it is turnedin eitherdirectionfrom 4>,it addsretardationto the microscopicfield by gettingeffectivelythickerto the traversing
light. The slow
wave axisof the rotarycompensator
is 90 to that of the first order red (Fig.

POLARIZATION

MICROSCOPY

423

THE
CALCULATION
OF
RETARDATION
WHERE.IN AN EHRINGHAUSOUARTZROTARYCOMPENSATOR:

I'X isTHE

DIFFEREE
IN! ATTHEWAVELEiTH
,

ISTHEWAVELEiTH IN 1

- REFIVE

R WHITELIGHT

INDEXDFTHEORDINY

- REFIVE
INDEX M THE ERRDINARY
IN E PTE x
i - E

LE
M ILINAT
OF E
RET
TD
ZEO TI

THE CALCULATION

OF NUMERICAL

- 0

- 1.

WAVE
- 1.Wl
OF THE

BIREFRINGENCE

Figure 11.

8(B)).By switching
compensators,
withoutchanging
the orientation
of the
hair,theslowandfastwaveaxesof the compensator
arenowopposed
90
relativeto thoseof thehair,ratherthanbeingaligned.
Therefore,
beginning
fromb,anyretardation
addedto the microscopic
fieldby rotationof the
compensator,
is subtracted
from the hair. By rotatingthe compensator,
until
nobirefringence
or lightis present
in the centerof thehairor at the edgeof
the medulla(if present),
the retardation
canbe calculated.
Figure10 portraysthisprocedure.
The compensator
readsin degrees
of rotation.Thishas
to be converted
intomicrons
of retardation
throughtablessupplied
with the
compensator,
or moreaccuratelyfrom a mathematicalformula(Fig. 11(A))
(13).

The numericalbirefrigenceis then determinedfrom the measurements

taken(Fig. 11(B)). A programmable
microcomputer
is usedfor the various
calculations.

The retardationcalculation
is baseduponthe use of white light (monochromatic
light is not necessary
in thissystem),
havinga conventional
gravity
point of k--550, with the Ehringhausquartz plate compensator
fast and

slowwave refractiveindicesbeing1.5459954
and 1.5551609,
respectively.
Precise
readings
arebaseel
upontheassumption
thattheopticalsystem
and
the measured
obiectare centeredand aligned.In taking readings,the

424

JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

compensatorshouldbe tilted in both directionsfrom qb,with their average

used in the calculationof retardation.This averagingprocesstends to
cancelout any error causedby eccentricity(13).
EXPERIMENTATION

Two experimentswere conductedto test the significanceof numericalbirefringence.asa measureof hair conditionsas follows.

(1) A total of ~ 150strandsfroma femaleCaucasian

volunteerwith virgin
hair were pickedat random.Eachof thesehairswasmeasuredfor the following parameters:wide and narrow axisdiameter,cross-sectional
area,microns
of retardation,the numericalbirefringence,
and stress-strain
curvecharacteristics.Any hairsshowingobviousmechanical/chemical
damageto the cortex
were not used.

A predictionmodelvassetup, utilizingan OlivettiP652microcomputer*

and Olivetti stepwisemultiplelinear regression
programno. 3.03* to look
at thecontribution
of numerical
birefringence
to theforceat yieldpoint(Fig.
12).Utilizingthe model,a predictionof the dependent
variable,forceat 10
per centelongation,
is madefrom the measurednumericalbirefringence,
the
cross-sectional
area,and their interaction(the productof the two). The actual predictionis accomplished
by putting the three measuredvaluesinto
the solutionvectorof the model(Table I) (14).
(2) Two hairs,approximately
8 in. long,werecut into 3 partseachand
placedin vacuoat ~ 1 x 10-5 Torr overnight.Each of the threepartswas
pickedat randomwith respectto distancefrom the scalp,and placedin
either 100 per cent glycerol,distilledwater, or a solutionformulatedof the
two, to simulatethe conditionsof 0, 100, and 50 per cent relativehumidities
(3, 15). Mechanicalstress-strain
measurements
togetherwith diameterand
numericalbirefringencemeasurements,
were taken within 30 min after immersion into the solution.
Results and Discussion

The resultsof experiments

i and 2 are summarized
in Table ! and Fig. 13,
respectively.
An attemptis being made to answertwo questionsthroughtheseexperi-

mentsas follows:(1) what is the valueof numericalbirefringence

in describingthe staticconditionof a hair fiber;and (2) how doesthe changeof
numerical
birefringence
in a hairfiberrelateto the effectof a particulartreatmentor conditioner?
Thus,twoproposed
usesof numerical
birefringence
are
tested as a measure of human hair condition.

*Olivetti
ofAmerica,
Inc.,NewYork,
N.Y.10022.

POLARIZATION

MICROSCOPY

425

MULTIPLE LINEAR REGRESSION MODEL

CORRELATING
THEOPTICAL/MECHANICAL
CHARACTERISTICS
OF HUMAN

HAIR

Y = to+ tx] +'t=X=

+ txx=+ e
WHERE:

Y = FORCEIN GRAMS@10%

ARE
THE
WEIGHTS

DEFORMATION

X;

THAT ARE ASSOCIATED

WITH EACH OF THE

= AREA IN SQUAREMICRONS

MODEL PARAMETERS

X= = NUMERICALBIREFRINGENCE

X;X2 = INTERACTION
OFXANDX=
e

= UNEXPLAINEDRANDOMERROR

Figure
I2. Model
used
totestforcontribution
andsignificance
ofindependent
param-

eters

Table I
o

'1
'1

/
/

i....3 I

CORRELATION

SOLUTION VECTOR:

FORCE (IN GRAMS)

N=

.,

MATRIX

3743.57

.OO8763

896.57

32.4061

6.1874

44.2762

9.2476

PARAMETER

DISTRIBUTIONS

143

= 4.9191 + 0.004361 (area)+ 0.7106 (areax numericalbirefrigence)

SE = 0.000952

SE = 0.1380

= 4.58

----5.084

df = 140

df = 140

P .0.01

P ,C0.01

COEFFICIENT OF DETERMINATION

0.7652

OVERALL

228.1795

df

2,140

P ' 0.001

426

JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

II

/4

=%RHVSFORCE/AREA
_""

....

........
-",._.,,
_-"..,
- ........,,_
.-._ _ __-

% RHVSNUMERICAL
BIREFRINGENCE

X10

/o

gm/
/.,,m'
X10'

7
5

6.5

o'

zs'

50'
% RELATIVE

5'

/oo'

HUMIDITY

Figure 13. Linear regression

with 95 per cent confidenceinterval of force/area and
numericalbirefringenceversusper cent relativehumidity

In the contextof this paper,the term conditionis usedsynonymously

with
the parameterdescriptions
body and manageability.
These,in turn, are more
specificallydefinedas elasticityor the resistance
to and recoveryfrom deformationinducedby externalforce.It canbe seenby the natureof thisdefinition,that hair conditionis traditionallythoughtof in a mechanical
light (1, 9).
Obviously,there is a molecularbasisto this mechanicalbehavior(7, 9, 16).
The a-helix molecularchain arrangement,basisof the keratin fibril system,
is a symmetricallyorderedconfiguration
mainly responsible
for this longitudinal stabilityin hair. In addition,both surrounding
and infiltratingthe
fiber systemis an amorphouscement-likematrix, which is high in cystine

cross-linking,
andhasbotha and/3 keratinchains(7, 16, 17, 18).
It is this relationshipbetweenthe fiber systemand matrixwhich becomes
importantasthe contributing
factorto bothopticaland mechanical
properties
of the hair (3, 7, 9, 16, 17, 18) (Fig. 14).
The mechanicalparametersof this relationshipcan be explainedin te]xns
of a two-phasemodel, which becomesapparent through experimentation
with the aqueousswellingof hair (9, 16).

A hair fiber undergoes

variouschangesdue to absorbedmoisturebetween
the relativehumiditylevelsof 0 to 100per cent.Approximately
2.7 timesless
forceis requiredto stressa hair to yield pointin a saturatedversusdry state.
At the sametime, swellingoccursat a rate of 16 per cent radiallybut only

POLARIZATION

427

MICROSCOPY
..

..

:..

:.;i;..:
% .. ..

...,
w..........4!71
?:' .-.-.
....

..;. .-- ....-:'. .... '- ..

...;:'
::.'
.:....;;

...

il '-,'-...
. ?
. .
:.? ':::'...:
,' ..
!i?'
' :;
:
.":2
:'i::.
:-...:)'
:..

....
.

:...

:". -....2'

: .'7'

....

'.-.......
?.:.."
'...%:

??":'":
:4:'

'

..

."

g" . .-::.;....

':..7:.:.":-'.'-' :

:;:..-..'' .

.'.': '-L .-.'.

.,;.:,.
{ "..,.?....'.,.
:-':,.,, '-.

..:'-:.,.:;..
';.';;..:'': .' .
.....
.,.. if!;::.

' "........ .?.:

'(.
:..
. '. ;

. :..

:.;....%

..;*.

-.. ;.-...-...----.

...-. ;.;.;'.
.-:
,..:
.:..
....
,. ........
.':;
.
........: %..
.:
--.-.
..::.;.;
.-:. ...... ....... : .....
-. ,'-.-:'.'.--,
,' "-" --

"..-.'
::..
?;:;....
' :'"d;'
.

. -.:

....., :? .?;.i:,::'"
... .....::''.... .,

97'-::
-;..':*"
'.

' -'Z-?'..**
:.--, /'

.'.i;::S::::.:'..-:-::.t
_C&.:,::'...: * ,. .?

t, .", :',:;'.
:..,:..;.-;
.........
;;4:":--::
.......

:::.:.

:...,.....
..

..;:.:...;':.'.'
-

......

..

:....,,.::.

..,.

'..::;':
.......
-

......... ..

Figure]4. (A) Scanning

d (B) transmission
electron
micrographs
of fibrilandmatrix

system
human
hakcortex.
ote fiberbundles
i.n(A) madeupof smallfibers,
andin
(B) the 80 protofibd(smallwhite oiroles),
suoundedby an amohousmatrix
(stainedrk)

428

JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

1.2 per cent longitudinally.Equally, there are changesin the torsionallongitudinal

modulesratio(9, 16).
The two-phasemodel is one in which a systemof rodletsor fibrils,hydrophobicin nature,are imbeddedin an amorphous
hydrophilicmatrix.Water
enteringthis systemaffectsthe hydrogenand salt bondsof the matrix,while
the rodlet-fibrilsremainrelativelyunaffected.The swellingexperiments
bear

thisout:the rapidrate of radialswellingis indicativeof matrixbondbreakdown,whilethe low rate of longitudinalswellingis indicativeof the fibril system molecularstability.A reductionin the longitudinalYoung'sModulusis
indicativeof a changein the summation
of the fibermatrixrelationship,
contributingto the total elasticity(3, 7, 9, 16).
Thesechangesare causedby variancein the water contentof the fiber as a
resultof variedrelativehumidity.Water, in effect,is usedas a modelconditioningagentfor the purposeof theseexperiments.

Within the realmof cosmetic

conditioning
andtreatment,therealsoexist
more severeagents(i.e., permanentwaving solutionsand bleaches),which
not only affectthe hydrogenand saltlinkages,as is the casewith water,but
also the disulphideand even, at the extreme,the fibril backbonepeptide
bonds.Theseagentsalsohavean effectequalin magnitudeon the two-phase
systemand ultimatelythe mechanical-physical
propertiesof hair (1, 7, 20).
This two-phasemechanicalmodel has a rather unique correlation:it
emergessurprisinglysimilarto the classicFrey-Wysslingopticalmodelof
compositebody-rodletbirefringence(Fig. 15) (19). In this model,rodlet or
miceIlar structuresof a crystallinenature are imbeddedin a surrounding
medium of a different refractive index, as is the casein hair. The rodlet struc-

turesthemselves
are responsible
for a contributionto the birefringenceof the
total systemin the mode of intrinsicbirefringence.In addition,however,the
relationship
betweenthe refractiveindexof the rodletsand their surrounding
mediumproducesform birefringence,
which is directly proportionalto the
differencebetweenthe 2 refractiveindicesof the components.
The summation of intrinsicand form birefringence
equalsthe total:numericalbirefringence(3, 4, 11 19).

In Table I, the significance

of a numericalbirefringence
systemto determine the static condition of a hair strand can be seen. Previous work corre-

latingthe molecularorientation.offibrousmaterial,as measured

by numerical birefringence,
with sonicpulsevelocity,which is very well correlated
with Young's
Modulusin hair,encourages
this(9.1,9.2).The correlation
matrix
explainsthe contributions
of the variousparametersto the dependentvariable,gramsforceat 10 per centdeformation,aswell as any interactions
which
occurberweenparameters.
Cross-sectional
areaoffersthe greatestsinglecontribution to the force; a large hair is expectedto require more force than a

smallone.In addition,informationaboutthe orientationor integrityof the

POLARIZATION

MICROSCOPY

419

Figure 15. C],assicalFrey-Wyssling optical model of compositebody-rodlet birefringence [with permissionof the publisher(19)]

fiber-matrixsystemcan be soughtfrom the numericalbirefringencevalue.

This correlation,
however,is low and not significant.
(It doesnot appearin the
solutionvector.
) Thisis dueto the inherentcharacteristics
of humanhair being suchthat the numericalbirefringence
hasa tendencyto be inverselyproportionalto the cross-sectional
area (the correlationof parameters2 with 1)
and, thus,is not independent.In order to get additionalinformationabout
orientationand integrity,the interactionterm of the cross-sectional
area and
numericalbirefringencemust be looked at to correctfor nonindependence.
This parameterhasa correlationof > 0.85with the gramsforce.

The solutionvectoris baseduponthe 143hairsplacedinto the model.It

canbe seenby the statistical
t-valuesthat the 2 parameters,
areaand the interactionof area and numericalbirefringence,are highly significant:P <0.01.

430

JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

If onetakesthe 2 measurements
froma hair andplacestheminto the solutionvectorformula,a predicted
valueof theforcein gramsat 10percentdeformation can be obtained.With this model, the coefficientof determination

is > 0.76.Thus,more than 76 per cent of the varianceof the dependent

variable,forcein grams,canbe explained.
This valueis significant
asis ex-

pressed
by thehighoverallF-ratio(228, 2, 140),andlowP-value(< 0.001).
With the futureinsertionof additionalpertinentvariables,aswell asa greater
samplesizebase,an additionalamountof the unexplainedrandomerror can
be explained.For moredetail aboutthis type of modelsee(14).

FromFig. 13,it canbe seenthat bothYoung's

Modulusandnmnerical
birefringence
are similarlycorrelatednegativelywith relativehumidity.A
changein relative humidity, or more precisely,a changein water content,
inverselychangesboth the mechanicalelasticityand optical anisotropic
propertiesof the hair.

Cosmeticconditioning,
or treatmentswhich affectthe conditionof hair,
changethe hydrogenand saltbondingarrangements
of the matrix,and even
the morethermodynamic
bondsof the entirecortex,in the caseof harsher

avafiablecosmetictreatments(7). As a result,there is a changein the

mechanicalstress-strain
characteristic,
which is the traditionalparameterof
measurement,
as well as a changein the refractiveindexof the matrix(1, 3,

7, 9, 16,20).This causes
a changein the formbirefringence,
in additionto a
possiblechangein the fibrilsthemselves,
causinga changein the intrinsic
birefringence(3, 4, 11, 19). Again, numericalbirefringenceis an extremely
sensitivemeasureof molecularorientationin an anisotropicmaterial (3, 5,
21).

Changesin mechanicalelasticityor conditionare a resultof changes

in
molecularbondingand orientation(1, 7, 9, 16, 20). In that light, stress-strain
analysisappearsto 'bemeasuringa secondary
parameter.
The mechanicalmethodof testingis inherentlyobtrusive.Requiredare
bothan opticaldevicefor sizedetermination
anda tensiledevicefor mechanical determination.Any inconsistency
or flaw over the spanof materialanalyzed tendsto biasthe results(1).
To build a systemof analysis,
it is necessary
to usecertainknownparametersof evaluation
as a tangiblereference
to whicha correlation
canbe established.In the case of numericalbirefringence,mechanicalstress-strain
analysishas servedthis function.
Numericalbirefringenceis presentedas an alternativequantitativesystem

of analysi.s
of humanhair conditionat the molecularlevel. With this technique, oneis able to determine,unobstrusively,
utilizing only one instrument,
at specificareason the shaft,molecularoccurrences
associated
with hair condition and conditioning.

POLARIZATION

MICROSCOPY

431

ACKNOWLEDGMENTS

We are deeply indebtedto our colleaguesof the BiologicalResearchstaff

for their assistance,
especiallyto StephenW. Platts[or graphicsupport,and to
Dr. David W. Cannell and Dr. Ronald T. Harris [or their valuable critique.
(ReceivedOctober29, 1975)
REFERENCES

(1) R. Beyak, C. F. Mlyer, and G. S. Kass, Elasticity and tensile propertiesof human
hair. I. Singlefiber test method,]. Soc. Cosm.Chem., 20, 615-26 (1969).
(2) W. R. Phillips, Mineral Optics-Principles and Techniques,W. H. Freeman and
Company, San Francisco,California, 1971, Pp. 75-101.
(3) A. R. Haly and O. A. Swanepoel,Part V: The nature of birefringence,Text. Res. ].,
31, 966-72 (1961).
(4) H. J. Woods, Physicsof Fibers, The Institute of Physics,London, 1955, Pp. 50-5.
(5) S. K]osevych,Microscopyand photomicrography,
]. Biol. Photog. Ass., 43, 123-7
(1975).
(6) R. D. B. Frasier, Birefringenceand elasticity in keratin fibers, Nature, 172, 675-6
(1953).
(7) A. H. Powitt, Some properties of human hair, presented Australian Society of
CosmeticChemistsSeminar,Terrigal, N.S.W., Australia, April 1967.
(8) Roger K. Curtis and Don R. Tyson, Redken Laboratories,Inc., Van Nuys, CA
91411, unpublishedresearch.
(9) RobertA. Wall and Le Roy D. Hunter, Normal adult hair-structure
and properties,
Cosmeticsand Perfumery,89, 31-5 (1974).
(lO) Equipmentfor polarizedlight microscopy,
Carl Zeiss,Inc., (1972).
(11) R. D. B. Frasier and T. P. MacRae, Conformationsin Fibrous Proteinsand Related

SyntheticPolypeptides,
AcademicPress,New York,1973,Pp. 159-63.
(12) Vera H. Price, Pseudoviii annulati, an unusual variant of normal hair, Arch Dermatol., 102, 356-7 (1970).
(13) Rotary Compensatorby Ehringhaus, Quartz Plates, Table of Function, Carl Zeiss
Oberkochen/Wurtt, G41-520 d/e/f.
(14) Wm. Mendenhall,Introductionto Linear Modelsand the Designand Analysisof

Experiments,
WadsworthPublishingCompany,Inc., Belmont,California,1968, Pp.
1-465.

(15) R. H. Hermans,Contributionto the Physicsof CelluloseFibres,A Study in Sorp-

tion, Density,RefractivePower,and Orientation,

ElsevierPu'blishing
Company,
Inc., Amsterdam,1946, Pp. 21-32.
(16) E. C. Bendit and M. Feughelman,Encyclopediaof Polymer Scienceand Tech-

nology,Vol. 8, JohnWiley andSons,Inc., New Yok,1968,Pp. 1-44.

(17) R. D. B. Frasier,T. P. MacRae,and G. E. Rogers,Strnctureof a -keratin,Nature,
181, 592-4 (1959).
(18) LinusPaulingand RobertB. Corey,Compound
helicalconfigurations
of polypep-

tide chains:Structure
of proteins
of the a -keratintype,Nature,171, 59-61(1953).

(19) A. Frey-Wyssling,Submicroscopic
Morphologyof Protoplasmand its Derivations,

ElsevierPublishingCompany,Inc., Amsterdam,1948, Pp. 5.7-67.

(2o) S. D. Gershon,M. A. Goldberg,and M. M. Rieger,"Permanent

Waving,"in Ces-

roetics:Scienceand Technology,
Vol. 2, 2nd. ed., M. S. Balsamand E. Sagarin,
Ed., Wiley Interscience,
New York,1972,Pp. 174-96.
(21) Henry M. Morgan, Correlation of molecular orientation measurements in fibers
(22)

by opticalbirefringence
andpulsevelocity
methods,
Text.Res.J., 32, 866-8(1962).
LowellA. Goldsmith
and HowardP. Baden,The mechanical
properties
of hair,
J. Invest. Dermatol.55, 256-9 (1970).