J. $oc. Cosmet.Chem., 21, 875-900 (Dec. 9, 1970) The Mechanismof Hair Bleaching LESZEK J. WOLFRAM, Ph.D., K. HALL, B.Sc.,and I. HUI, M.Sc.* PresentedDecember2, 1969,New York City Synopsis--Thecolor of mammalian HAIRS is due mainly to the inclusion of discrete,darkly colored MELANIN granules in the keratinized cytoplasmicprotein of the fiber-forming cells. During BLEACHING the melanin pigment undergoes irreversible physicochemical changeswhich result either in the toning down or complete elimination of the original fiber color. The modificationof the fiber protein (KERATIN) attendant upon bleaching is largely confined to the oxidation of combined CYSTINE. The cysteicacid residuesformed in this reaction causea significantchangein the distribution of electrostaticcrosslinks. INTRODUCTION Peroxide bleachingof pigmented keratin fibers has been practiced for many years,yet no published account of any comprehensive studyof the kineticsor mechanismof the processis available. Literature concerningthe physicochemical changesin the pigment is practically nonexistent; that dealing with keratin modification attendant upon bleachingis relativelyrich (1-15), but mainly devotedto generalities, and often contradictory. This is not surprisingin view of the fact that the various authors who have studied the bleaching processemployedwidely differingconditionsof treatment. As the aim of bleaching is to eliminate or tone down the natural hair color, the processis directly related to the structure and reactivity of the hair pigment. Two principal approachesto understandingthe structure of melanin have been tried. The analytical approach has * Gillette CompanyResearchInstitute, 1413 ResearchBlvd., Rockville, Md. 20850. 875 876 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS been employed with greatest successby Nicolaus and his coworkers (16, 17). It hasled to a conceptof melanin structureas a polymerof multiple subunitsjoined by multiple typesof bonding,i.e., a poikilopolymer. The syntheticapproachwasinitiated by Raper (18). It has led to the secondconceptof melanin structureas a regular polymer involvinga singletype ot5monomerjoined by a singletype of linkage, i.e., a homopolymer. Unfortunately,little effort wasmade to interpret the postulatedstructuresin terms of the melanin reactivity which remains an enigma. The pigment granulesare distributed within the cortex of the fiber and it is thusnot surprisingthat during the bleachingprocess someoxidation of the keratin matrix occurs. This is often referred to as "oxidative" or "bleaching"damage. With regard to the specificityof this oxidative attack,the lossof cystinehasbeen ascertained(5, 7) and the modification of other amino acid side chains (tyrosine, tryptophane, lysine, and arginine)hasbeen postulated(10, 13, 15). However, the majority of the publisheddata refers to the bleachingof wool, which is usually carriedout at elevatedtemperaturesin neutral or slightlyalkaline media. In the bleachingof hair, the useof ambient temperatureis compensated for by a higher pH value of the system. The publishedinformationon the physicochemical changesin hair keratin taking placeunder suchconditions is almost exclusivelyqualitative, and it is quite inadequateto serveas a firm guide for improving existing bleaching systems. This communicationis an accountof an investigationaimed at obtaining a better understanding of the complex processesassociatedwith the reaction of hydrogen peroxide with both the melanin pigment and hair proteins. MATERIALS AND METHODS The Caucasianhair,* brown and white, used in this investigation wasshampooed, rinsed,and conditionedat 65% RH and 70øF. The black poodle hair was obtained from random samplesof hair clippings. The hair was purified by Soxhlet extraction for 4 hours each with methylene chloride followed by absolute ethanol. It was then rinsed well with deionized water and conditioned as above. Commercially available, reagent grade solvents and chemical reagentsutilized in this studywere not further purified, unlessotherwise specified. * Suppliedby De Moo Brothers,New York. HAIR BLEACHING 877 The melanin wasisolatedfrom the hair by acid hydrolysisaccording to the method of Green and Happey (19). Purified black poodle hair was placed in a round-bottom flask equipped with a reflux condenser and hydrolyzedwith 6N HC1 for 4 hoursunder reflux. A liquor-to-hair ratio of 40:1 was used. The mixture was cooled and the melanin was separatedby centrifugingfor 30 tnin at approximately 1000 g's. The sedimented melanin was washed with deionized water until the solu- tion in equilibrium with melanin had a pH value of 5.2. The melanin was then rinsed several times with acetone and dried in vacuo at 60øC. Microscopicinspectionof the product indicated the absenceof any fibrouscontamination. The high purity of the isolated pigment was confirmed by examination of the melanin granules with the electron microscope. Altogether, 50 g of poodle hair was hydrolyzed,yielding 3.75 g of melanin. A nonhydrolyticmethod (20) of melanin extraction was employed mainly for a comparative examination of chemical properties. The hair samplewasmaintained at reflux for 24 hours in a phenol hydratethioglycolic acid mixture (PHT). The filtration step was omitted, due to a previousunsuccessful attempt in the isolationof the pigment. After separation of the brown pigment by centrifuging, the melanin was washed two times with fresh portions of PHT. The isolated melanin was further washed with deionized water as described above, followed by severalacetonerinsesand drying. The purified melanin contained a considerable amount of fibrous material, which was re- moved manually. Visible and UV absorptionspectraof melanin were obtained on a Perkin-Elmer Model 202 Spectrophotometer. Mechanical propertiesof hair were determined on the table model Instron. The fiberswere mounted on plastictabsat 2-in. gaugelength, equilibratedunder the desiredconditions,and stretchedto break at required ratesof extension. The broken endswere then cut off the tabs, conditioned,weighed,and the denier of the testedfiberswascalculated. Amino acid analysesof untreated and bleached hair were carried out on a Phoenix Model M-7800 Micro Analyzer.* The swellingof hair was determined by the liquid retention tech- niquedescribed by Valkoand Barnett(21). When specified, a reduction- * PhoenixPrecisionInstrument Co., 3803-05North Fifth Street,Philadelphia,Penna. 19140. 878 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS oxidationcyclewasusedjust beforethe determination. The purpose was to intensify the damageand thus allow a differentiation between sampleswhich are difficult to resolveotherwise. The hair was treated with 0.2M ammonium thioglycolate(pH 9.6, 35øC) for 10 min at a 20:1 liquor ratio, followed by a brief rinse and treatment with 0.2M H•O2 (pH 3.4, 35øC) for 5 min. The hair was then rinsed free of peroxidewith deionizedwater. Thin-layer chromatography wasemployedfor separatingthe melanin oxidation products. The adsorbentlayer wassilicagel (Merck), having a thicknessof 250 v. Plate size was 5 X 20 or 20 X 20 cm. In caseswhere the subsequent elution of components wasto be performed,the thin-layer plateswere freed of possibleimpurities by their immersionin spectral grademethanol [or at least30 min. The plateswere then ready for use without any prior activation. Sampleapplicationwasusually made by streakingthe aqueoussolutions onto the plate with the aid of a Brinkmann streaking piper. Wheneverlimited quantitiesof samplewere available,the sampleswere spottedonto the plateswith microliter pipers. Developmentof the plateswasperformedin a rectangularchamber (11 X 11.5 X 4 in.) containing300 ml of solvent,namely, ethanol:ammonia:water in the ratio of 80:4:16. Overnight equilibration of the solvent,in a closedtank, lined with solvent-soaked paper, was necessary before its usage. A solventmigration of 17 cm required development timesof 3-3.5 hoursat ambient temperature. After the plateswere developed,they were dried for about 20 min at 105øC. The resolvedcomponents were locatedby exposingthe plate to ultraviolet irradiation from a 4-watt lamp with a spectraldensity of 3600fk.* The spotsor bandswerelocatedby their fluorescence. Other methods of identification included the following spray reagents,preparedaccordingto Stahl's (22) procedure: (a) Bromocresol green(0.04%) in ethyl alcohol (b) AgNO:,-NH.•: equalpartsof 0.1NAgNO:,and5N NH• (C) 50% H2SO4 (d) Ninhydrin, 0.3% in ethanol.* * Ultra-Violet Products, Inc., San Gabriel, Calif. 91778. ?As suppliedby SigmaChemicalCo.,P.O. Box 14508,St. Louis,Mo. 63178. HAIR BLEACHING 879 RESULTS AND DISCUSSION Preliminary 0 bservations Stabilizedaqueoussolutionsof H202 undergolittle decomposition at ambient temperatureseven at high pH values. However, the introduction of a solid into a system brings about an increase in the decompositionrate which is roughly proportional to the surfacearea of the solid. An additional incrementin the rate of decompositionis observedwhenever the introduced solid undergoeschemical reaction with H202. Table I shows the rate differences obtained under such conditions. Table I Decomposition of Hydrogen Peroxide in the Presenceand Absence of Hair a H202 Decomposition (%) Time of Reaction, rain No fibers present White hair Brown hair 5 0 O8 0.8 10 0 21 2.5 20 0.3 42 6.5 30 0.5 61 60 0.9 94 14.2 90 1.4 13 5 21.1 8.0 a Bath ratio, 33:1; Initial [H20•] = 35 gl-•; pH 10.0; 35 ø C. Initially, with the reaction confinedto the surfaceand to the cuticular region of the fiber, the rate of the peroxide decmnpositionis ahnost identical for both brown and white hair. This is not surprisingin view of the similarityof the dimneterand of the che•nicalcronposition of both smnplesof hair. It is likely that the divergence in the decompositionratesobservedin the later stagesof the reactionis associated with the responseof the piginent,the granulesof which are locatedwithin the cortexof hair and thereforenot soreadilyaccessible to the peroxide. Bearing in mind the low inelanin contentof the studiedhair (,-•2%), these observationsconnote a much higher reactivity of melanin than that of keratin. Galculations based on the data of Table I show that the overall rate of peroxide decompositionin the presenceof brown hair is 9.0 X 10-2 •nM •nin-• g-•, while the correspondingvalue for white hair is 6.4 X 10-2mM min -• g-•. If the difference in H202 880 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS decompositionwas to be accountedfor by the melanin alone, then the latter would yield a value of 130 X 10-2 mM min -• g-•, a difference factor of over 20. A rate of this magnitude (95 X 10-2 mM min -• g-•) was actually observedin the study of the oxidation of melanin with H202 at pH 10 and 35øC. Assumingthat the rate of peroxidedecmnposition is related to that of oxidative reactionstaking place within the fiber, then this large difference in reactivity is obviously a desirable feature from the point of the bleaching process. A further increase in the reactivity ratio should lead to faster and less damaging bleaching. Such an approachhas been utilized in the bleaching of wool by using the iron mordanting technique (23). The preferential binding of iron by the pigment sensitizesthe latter to the peroxide attack and results in significantaccelerationof bleaching. Although qualitative observationsof colorchangeswhich hair undergoesduring bleaching,combined with quantitative evaluation of peroxide consumption,provide somemeasureof melanin reactivity, they add little to our understanding of the mechanismof the process. In addition, the intimate associationof the pigment with the hair fiber is likely to interfere with many physicochemical aspectsof the process and to obscuretheir relative importance. To obviate these difficulties eachof the components(melaninand keratin)wasexaminedseparately. It was assumedthat isolation of the pigment from its keratin environment would not significantlyaffectits chemicalbehavior. Reaction of Melanin Pigment with Hydrogen Peroxide The melanin was isolated from the hair in the form of discrete granules,approximately0.8-1.2 • long and 0.3-0.4 u thick (Fig. 1). Examination of the pigment with the electron microscopeat several magnificationlevels (5,000-50,000) did not reveal any structural organizationof the g•:anules. This wastrue for both the PHT- and HC1isolated melanin. The densityof melanin was determinedby the flotation technique (benzene/bromobenzene/3-bromochlorobenzene system)and wasfound to be 1.53 g/cmg. This is much lower than the value of 1.71 reported by Swift (24). The pigment •-anules are hygxoscopic, attaining the equilibrium regain of 16.4% at 65% relative humidity. Although no attemptwas made to identify the water binding sites,the acid and basecombining capacitiesof the pigmentwere determined(0.32 and 2.5 meq/g, re- HAIR BLEACHING 881 .' . Figure 1. Electron micrograph of melanin granules extracted by hydrolytic method spectively). It is very-likely that thesepolar residuesact as the primary centersof water sorption. Solubilizationof Melanin Pigment The cross-linked,polymeric structure of melanin manifestsitself in its high resistanceto numerous organic and inorganic solvents. Some dissolution of melanin was detected in DMSO, concentrated H.,SO4, and 1N NaOH, but only at elevatedtemperatures (100øC and above). Yet, evenprolongeddigestionwith thesesolventsleft the bulk of the pigment insoluble. Extensive treatments of melanin (up to 48 hours) with reducing agents such as thioglycohcacid, borohydride, sulfide, and sulfite produced no apparent physicalchangein the pigment. Neither did oxidation with persulfate, perchlorate, iodate, and permanganateperformed over a wide range of pH (1-10). A different behavior wasdisplayedby hydrogen peroxide. Dilute aqueoussolutions of this reagent caused disintegTationof the pigment gTanules,which slowly dissolvedin the reaction system. The solution becameintenselycolored, the dark color persistingfor a considerablelength of time, after which some fading was evident. This observation, obviously relevant to our bleaching 882 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS studies,was also very significantin the senseof a general method for melanin solubilization and its usefulness toward better characteriza- tion of this highly resistantpolymer. The reactionappearedamenableto spectroscopic techniques,and both the visible and the UV spectraof the melanin solutionswere examined. The visible region proved uninformative;a monotonicrise in optical densitywith increasingtime of reactionwas observed. The rise was very gradual and did not appear to reflect the qualitative changes takingplacein the system under investigation.Initial attempts to utilize the UV region were also of little avail becauseof the high absorptionintensity of H•O2, which overshadowed any absorption changescausedby the solubilization of the melanin. However, by resortingto the techniqueof differentialspectroscopy, the peroxide absorbancewas suppressed and the spectroscopic changesdue to the reactioncouldbe readilyfollowed. A typicalrecordingis reproduced in Fig. 2. Within a few momentsof the contact of the reagents,a well-definedabsorbance peak wasdeveloped. The peak intensityincreased with the time of the reaction and reached a maximum at the time of completedissolutionof the melanin. Then a slowdecreasein absorbance wasobservedasthe bleachingof melaninby H202 continued. mm.•..• -- õmirl•_ 15mi•----•"'• \ -0.4 •4 mirt 0.5 8 mi o7• •N 0m.i,• 0.s•z 8mln 0,9 • 1.0 1,1 1.2 1,3 1.4 1,5 3 0 350 3 WAVELENGTH (millimicrons) Figure 2. Solubilization of intact melanin in 1% H202 at pH 11.5 HAIR BLEACHING 883 The time required for dissolutionof melanin in aqueousH:O2 can be readily determined from the absorbancepeak and then used as a convenient parameter in further investigationsof the reaction. The generalexperimentalprocedureemployedin the solubilizationstudies was as follows: 1 mg of melanin was introduced into a volumetric flask containing 10 ml of aqueousH20.,. The reaction mixture was stirred magneticallyat 25øC; at a given time, aliquots were withdrawn, transferredinto a 5-ram quartz cell, and their spectrarecorded. The reference cell contained a solution of H:O2 at a concentration identical to that in the sample. Concomitant with the recording'of spectra a visual observationwasmade of the state of the melanin dispersion,and the time of its completedissolutionwas noted. Usually the reaction was followed for at least 60 min after the dissolution time. Effect o[ pH on the Rate of Solubilization--The pigment was treated with 1% aqueoussolutionsof HeO2 adjustedto different pH values by means of sodium hydroxide. Both the dissolution times (tD) and the absorbancesat tD were recorded,and are given in Table II. The reaction appearsto have a maximum rate in the region of pH just below the pK value of H20,(11.75), indicating that although the peroxide anion is definitely involved in the solubilization process,it may not be the soleattackingspecies. Table II Effect of pH on the Dissolutionof Melanin in 1% H202 at 25øC pH Time for Complete Dissolution (t•) (Min) Absorbance 10.45 30 10.80 14 1.29 1.27 11.25 9 1.15 11.55 10 1.21 12.20 14 1.21 12.70 13 1.02 Duke and Haas, who studied the homogeneousbase-catalyzeddecompositionof H== (25), noted a similar pH dependenceand accountedfor it by postulatinga reactive,cyclicintermediateformed from the neutral H=• molecule and its anion: [ .4' decomposition __ _ H•o/OxH' ßH20 +OH+0,2 H,,O,, + HO,- • 884 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS The rate of decompositionwasaccordinglyexpressedby' Rate = K[H.,O2][HO2-] the productreachinga maximumat pH = pK•2o2. Our resultsare not entirely consistentwith the aboveequation as the rate of solubilization appearsto be more affectedby the decreasein concentrationof the ionized speciesthan that of the neutral molecule. However, it shouldbe borne in mind that ionization of the acidic side chainspresent in the melanin is likely to lead to a buildup of negativechargeson the pigment. This may form an effectiveelectrostaticbarrier to the penetration of the peroxideanion and thusbe an important factorin affecting the rate of oxidation. Effect of Hydrogen Peroxide Concentrationon the Rate of Melanin Solubilization--The experimentswere run at the optimal pH of 11.5. An accelerationin the dissolutionrate wasobservedwith increasingconcentrationof the peroxide. Although a plot (Fig. 3) of the reciprocal I I I] 1 2 3 24.( 20., x 16.• i r. 12.0 .-- .-- :3 '• 8.0 © © ._• 4.0 Figure 3. Effectof H,20.o on rate of melanin dissolution HAIR BLEACHING 885 of the dissolutiontimes (tD) againstH202 concentrationyields a straight line indicative of ordinary kinetics for the bimolecular reaction, this simple dependenceis probably fortuitous in view of the heterogeneity of the solubi!i?ationprocessA scrutiny of the changesin absorbanceat to for various concentrations of H202 reveals a possibleclue to the physical mechanismof bleaching. If one assumesthat the bleaching of melanin by peroxide is an inherent part of the solubilization of the pigment granule, then the absorbancesof melanin solutionsat to should be independent of H202 concentration. This is not the case (Table III). Not only are the absorbanceintensities of dilute Heemelanin systemsvery much higher than thosewith prolongedtime, but they remain virtually unchangedfor a prolongedtime. The "lack" of bleachingis 'not caused by depletion of the reagent. Indeed, even in the most dilute solutions studied(0.01% H.oO2)the molar ratio of H202 to the melanin (indole residue)at to is at leastof the order of 5:1. Table III Maximum Absorbancesat tofor Various Concentrationsof H202 at pH 11.5 [H2021, % Absorbance [H202], % Absorbance 0.1 2.95 1.0 1.24 0.4 1.41 2.0 1.14 0.6 1.36 3.0 1.06 The resultscanbe plausiblyexplainedin termsof a two-stepprocess: solubilization of the granule followed by decolorization of the dissolved melanin. The data imply that the bleaching processis relatively slow when compared to the solubilization of the pigment and thus controls the overall rate. This hypothesiswas supportedby electromicrographicexamination of the melanin which had been subjectedto H202 treatment for various lengthsof time. At the end of the reaction time, excessperoxidewas decomposedby platinum black, the solution was filtered, and the undissolvedpigment was examined. There was little apparent change in the sizeof the granulesas a function of time. Yet, only 5 min of treatment was required to dissolveas much as half of the original weight of the melanin. Evidently the disintegrationof the pigment granuleswasvery fastonce it commenced. 886 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Physicochemical Propertiesof SolubilizedMelanin Solubilizationof melanin by dilute solutionsof H•O,., presented itself as a potentially very useful tool for better characterizationof this intractable polymer, provided that the modificationbrought about by the peroxideattackwasnot too great. It wasthoughtlikely that under mild conditions of treatment the primary reaction would be the elimination of the solubility-restrainingcross links and the overall chemicalnature of the pigment would be retained. Solubilizedmelanin wasthereforeprepared,and someof its propertieswere examinedand comparedto thoseof intact melanin,where possible. Solubilization of the granules was effected at low concentrationsof H..,O2(1%) in 0.5:1I ammonia at pH 10 and 200:1 liquor ratio. Complete dissolutionof melanin under these conditionstook place within 60 min. At this point the peroxidewasdestroyedby platinum black, the water wasremovedby evaporationon a steambath, and the product was isolatedas a water-soluble,highly lustrousmaterial. On acidificationto pH 2, the solubilizedmelanin precipitated. This product wasdenotedas melanin free acid (MFA). Its solubility behavior is typical of a polymer containingfew ionizablegroups,dissociationof which forcesthe polymeric chain into solution. Thus MFA remains insolubleat low pH and dissolves rapidly when the pH of the systemis raised above 4. The solubilizationprocessincreasedthe baseuptake capacityof the melanin from 2.7 to 3.8 meq/g. This correspondsto a neutralization equivalent of 262, and indicatesvery slight oxidative breakdown of the melanin structure. SpectroscopicSt•tdies--The potential usefulnessof infrared spectroscopyis limited when chemicallyill-defined polymersare examined; this certainly appearsto be the casefor melanin. There waspractically no changein the spectrumfollowing the solubilization. The UV region proved to be more informative. While the intact melanin showsno absorbance maxima, and a monotonic rise of absorbance with the de- creasein wavelengthbeing observed,the solubilized melanin exhibits a well-definedmaximum at 222 mv (Fig. 4). Although no positiveidentification of the absorbing sites can be made, a tentative assignmentof this maximum to a peroxide-typestructureis postulated. It is perhapsappropriateat this stageto discussin more detail the spectroscopic changesoccurringin the UV region during the solubilization of melanin. As reported earlier in this investigation, dissolution HAIR BLEACHING 887 0.4 0.5 0.6 •> I).7• 11.8 • z 0.9m ø 1.0 1.2 1.3 1.4 200 f • • • •i 2 0 • • I I I 300 • f, 1.5 WAVELENGTH (millimicrons) Figure 4. Absorbance of solubilized melanin in water of the pigment in aqueousH.,O2is characterizedby an increasein the absorbance and a formationof a well-pronounced peak. The position of this peak is congruentwith the absorbancemaximum displayedby H.,O2 itself, and varies with the changein the peroxide concentration in an identical manner. Upon destructionof the peroxidewith platinum black, the peak shiftsto a new positionillustrated in Fig. 4. The shift is accompaniedby an increasein the absorbanceintensity. The positionof the peakduring the solubilizationprocessis suggestive either of a peroxide-typecompoundexhibiting an absorptionpattern identical to that of H202 or of the generationof H202 during the reaction. The latter alternative, however, would not satisfactorilyaccount for the existenceof an absorbance maximum after the decompositionof hydrogen peroxidewith platinum black. Molecular Weight--The extent of the degradation suffered by melanin during its dissolutioncould be assessed by the determination of molecularweightchangesbroughtabout by the solubilizationprocess. Unfortunately, no data on the molecular weight of the intact melanin are available and our attemptsto determine it with a vapor pressure osmometerwere unsuccessful.In fact, it is the solubilizationprocess itself that presentedthe opportunity to assess the molecular weight of the pigment. The apparatuschosenfor this study was devisedby Bull (26) for osmoticpressuremeasurements. Measurementsfor this type usually require some form of extrapolation to infinite dilution. The advantageof Bull'sosmometeris that it eliminatesthe needfor extrapola- 888 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS tion by allowing measurementsto be made at sufficientlylow dilutions for Van't Hoff's law to be valid. Melanin solubilizedby ammoniacalhydrogenperoxide wasdialyzed for 24 hours prior to the measurements. The molecular weight was calculatedfrom Burk and Greenberg'sequation (27): M = CdRT/lOOP where C is the concentrationof polymer in gramsper 100 ml of solvent, d is the densityof the solvent,R is the gasconstant,T is the temperature, and P is the osmoticpressure. The equation reducesto C M = 2.527X 10.5dff after insertingthe numericalvaluesfor the constants. A melanin concentrationof 0.310 g per 100ml of saltsolutiongaverise to an osmotic pressureof 6.86 cm of water. The molecularweight calculatedfrom thesedata yielded the value of 11,400. A value of the same order, viz., M z 15,000, was also obtained from the molecular weight determination using the thin-layer gel-filtration technique. Free Radical Content--Samplesof melanin were examined in a Varian X-band esrspectrometer.Both the intact and solubilizedmelanin gave rise to virtually identical structureless absorption,with line widths o1:the order of 6 gaussand g valuesof 2.003. The spin density wasdeterminedfor both of the samplesby comparisonwith a known DPPH (o6cr-diphenyl-/:t-picryl-hydrazyl) standardand gave a value of 10•9 spinsper gram. The most important point emergingfrom this brief study is that the free radicalcharacterof the melanin is not affectedby the solubilization process.This meansthat theseradicalsare extremelystableand do not rely for their existenceand stability on a physicaltrapping mechanism. Decolorization--The solubilizationof melanin by H.,O• is only the first step in the reactionsequence.Prolongedtreatmentresultsin bleachingor decolorizationof the intenselydark solution. Although thehighefficacy of hydrogenperoxide(ascomparedwith otheroxidants) for the solubilizationof melanin wasclearly establishedin this investigation,it did not connoteits superiorityin the bleachingstep. Consequently, theeffectof a numberof oxidizingagentson thecolorchange of theaqueous solutions of solubilized melaninwasassessed. The reac- HAIR BLEACHING Table 889 IV Effect of Oxidizing Agent on the Bleaching of Solubilized Melanin Oxidizing Agent Bleaching Ability Conditions, pH 1-10 (NH4)2S.•O8 None KI O 3 None 1-7 K2Cr•O7 None 1-7 NaC104 12 None None 5.2 q+ qq- q- qq- q- 10 7 <3 7-8 H202 NaOCI KMnO4 CHaCOOOH 1-7 tion was carried out at room temperature. In each case,excessof the oxidant waspresentin the system. The resultsare given in Table IV. The most surprisingfinding was the high decolorizationefficacy of the permanganate,particularly in view of its inability to react with the intact melanin. Are the crosslinks which are broken during the solubilization important to the preservationof color? Or does the solubilizingaction of peroxidesensitizethe melanin polymer, e.g., by generationof labile, peroxide-typestructuralelements?Approximately 0.03 meq of KMnO4 was required to bleach 1 mg of soluble melanin to a pale yellowcolor. Assumingan averageunit weight of the melanin as 145, 1 mole equivalentof KMnO4 is utilized for 2 melanin units. The contribution of peroxy anion speciesto the bleaching process canagainbe readily seenin the caseof peraceticacid. In slightlyacidic media thisreagentis specificfor oxidativecleavageof the disulfidebonds in hair but has little effect on the melanin. The latter is, however, readily attackedunder alkaline conditionsand a maximum decolorization effectwas observedin the pH range closeto the pK value (8.2) of the peracid (28). It is worth pointingout that while the disintegration of the pigment granulesand their solubilizationare the necessaryprerequisitesfor bleaching,theseprocesses, by themselves, are not likely to affect the color of hair significantly. At best,the conversionof pigment particles into solublemelanin dye might bring about a slight changein hue. The decolorizationstep,on the other hand, althoughcontingentupon the former, is more readily perceivedand thus may be consideredot• greaterpracticalimportance. From the experimentalevidenceobtained so far, it is impossibleto elucidatethe precisechemicalnature 890 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Table V Ri Values of the Componentsof the Decolorized Melanin Rf Relative Intensity of Fluorescence 0.00 0.02 0.02 Strong Medium, very narrow band Strong 0.04 Medium 0.08 Strong 0.16 Weak 0.23 Medium 0.35 Weak-medium 0.41 Weak-medium 0.43 Medium 0.51 Weak of the reactionsoccurringduring the solubilizationof the pigment and its subsequent decolorization.In the former process, the solubilityrestraining crosslinks are eliminated and the chromophoricgroups appear to be left virtually intact. There is a goodcaseto argue that these cross links are much more labile and thus different from the residues which undergo much slower reaction in the decolorizationstep. The bleachingproper, on the other hand, reliesupon oxidative destruction of the highly conjugatedsystemof the indole residues. It is likely that the oxidation is centered initially on the benzenoid portion of the ring. The acceleration of the decolorization processobservedwith both permanganate and peracids is in accordancewith such a view (29). Some additional support for the postulatedpath of oxidative breakdowncan also be derived from the resultsof qualitative chromatographic analysisof the productsof the decolorizationreaction. All the resolvedcomponents (Table V) were identified as acids but no aromaticderivativeswere present. The test for pyrrolic acidswas also negative. The latter were detectedby Piattelli (16) in the permanganateoxidized melanin. It appears,therefore, that under ordinary bleaching conditions,which we have employedfor the preparation of decolorized melanin, even the indole nucleusundergoescompletedegradationupon fissionof the benzenoidportion of the ring to yield smaller fragments with acidic functions, such as oxalic acid which was identified as one of the decolorization products. We were unable,however,to identify positively the remaining components. The oxidation of melanin by peroxideis accompaniedby development of fluorescence which increasesin intensity with the progressof HAIR BLEACHING 891 the reaction. All the decolorizationproductsare stronglyfluorescent and indeed it was this property which greatly assistedtheir chromatographicseparation. To our knowledgethis hasbeen the first report of the phenomenon. The only other relevant report was the observation by Binnsand Swan(30) of the purple fluorescence from the synthetic melanins. Reactionof Hair Keratin with Hydrogen Peroxide The fact that the reactivityof melanin with regardto hydrogenperoxide happensto be so much higher than that of keratin almost automatically connotesthe bleachingprocessas a commercialsuccess.However, the melanin pigment representsonly a very small fraction of the fiber weight (usuallyabout 2%) and thus it is reasonableto expect that some oxidative modification of the fiber matrix will occur. Conven- tional bleachingprocesses utilize hydrogenperoxide in alkaline media at pH 10 and above;the perhydroxyanion (HO2-) is the predominant reactivespecies.The abundanceof sitesin keratin which might yield to a nucleophilicattack by this ion precludesany firm prior assignment of a specificinteraction. In addition, the presenceof some radicals derived from H202, the reactivity of which is not particularly sensitive to pH changes,addsto the uncertainty concerningthe type of reactions involved. The physicalmethodsusedto detectdamageassociated with bleachingare satisfactoryfor measuringthe extent of deterioration,but are of little value for elucidating the chemical character of the damage. The latter can best be ascertainedby chemical analysis,and such a methodwasusedasa startingpoint of this investigation. ChemicalCompositionof BleachedHair Tressesof brown hair were bleachedwith 3% Haaat pH 10 (0.5M NH3) and 35øC for 4 hours. The bleachedtresseswere sampled,the sampleswere hydrolyzed,and the hydrolyzateswere analyzedon the Phoenix M-7800 Micro Analyzer. The resultspresentedin Table V1 showconvincinglythat, under the conditionsstudied,the reactionbetweenkeratin and H202 is confinedmainly to the cystineresidues. The decrease in cystineis almostquantitativelymatchedby a corresponding increasein cysteicacid. The amino acid analysisdoesnot reveal any intermediateoxidation productsof cystinewhich might be formed during the bleachingprocess. These compoundsare, however,very unstable under alkaline condiitons, and any remaining would dis- proportionate to cystineandcysteic acidduringthe hydrolysis. 892 JOURNAL OF THE SOCIETY OF COSMETICCHEMISTS Table VI Amino Acid Compositions of Untreated and BleachedCaucasianHair Amino Acid Content in/•mol/g Amino Acid Untreated Bleached Cysteicacid Asparticacid 55 455 289 447 Threonine Serine Glutamic acid Proline 653 870 871 672 642 820 868 700 Glycine 539 525 Alanine 471 460 1380 1130 Valine Isoleucine Leucine 538 250 554 542 247 530 Tyrosine Phenylalanine Lysine 132 130 213 120 119 225 Histidine Ammonia 63 780 69 870 Arginine 512 540 • cystine During bleachingsomeof the keratin dissolves in the reaction medium. The weight lossesare, however,very small. After 4 hours' treatmentof brown hair with 3% H22 at pH 10 and 35øC, the amount of extractedprotein did not exceed1% of the fiber weight. Even smallerproteinextracts wererecordedin the caseof whitehair treated under similar conditions. Althoughwe could not yet assess the averagemolecularweight of the dissolved protein,its aminoacid contenthasbeendeterminedand the resultsare givenin Table VIII. The origin of the dissolvedfraction is uncertain,as the relatively largedifferences in aminoacidcomposition betweenthe extractand the untreatedhair argueagainsthomogeneous solubilizaton. Someof the solubleproteincouldconceivably resultfrom the destruction and elution of the melanin-keratincomplex.Sucha view is stronglysupported by the fact that the extracted proteincontains two aminoacidsnot [ound either in the bulk of hair or in the bleachingextractof the white hair. These are taurine and /•-alanine. Nevertheless,the major portion of the extractmostprobablyrepresents peptidesassociated directlywith the oxidizedcystine. The cysteicacid residueis knownto facilitate HAIR BLEACHING Table 893 VII Amino Acid Compositionof Keratin Material DissolvedDuring Bleachingof Brown Caucasian Hair Amino Acid Content, % Amino Acid Extract Cysteic acid 14.1 Untreated Whole 0.9 Taurine 0.2 Aspartic acid 9.2 6.0 Threonine 5.3 7.8 Serine ... 11.8 9.1 17.6 12.8 Proline 0.6 7.7 Glycine 7.6 4.0 Alanine 4.1 4.2 • 3.7 16.8 Valine 4.9 6.3 Isoleucine 1.7 3.3 Leucine 3.9 7.1 Tyrosine Phenylalanine •5-Alanine Lysine 2.8 2.1 0.3 3.9 2.4 2.1 Histidine 0.4 1.1 Arginine 7.8 8.9 Glutamic acid cystine Hair ... 3.1 greatly the hydrolysisof the adjacentpeptide bond and thus create favorable conditions for destructive solubilization. Swellingof BleachedHair Of the manywaysin whichtheoxidativedamageof keratinattendant upon bleachingmanifestsitself (deteriorationof tactile properties,mechanicalweakeningof the fiber, increasein alkali solubility,etc.), the increasein swelling representsa convenientmeansfor the assessment of the extent of damage. This increasein swelling is brought about by changesin the bulk of the fiber, and thus is directly related to the overall chemicalmodificationof keratin by H,•O_o. Evidencehasbeen presentedhere that of all the amino acidspresent in keratin, only the cystineundergoesa measurableextent of reaction with the peroxide;possiblythe oxidativebreakdownof disulfidebonds alone would satisfactorilyaccountfor the increasedswelling. However, the principallocusof the bleachingreaction,aswe havelearnedearlier, is the melanin granule. It is conceivablethat the oxidative destruction of the •nelanin granulesmight result in formation of discretevoidswith- 894 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS in the fiber structure. It was thought that the contribution of this factor to the total swelling characteristicsof the bleached hair could be ascertainedby specificoxidation of cystincin brown hair with peracetic acid (without attacking the melanin) or by control bleachingruns on white hair. Table VIII correlatesthe swellingdata with the extent of disulfide bondbreakdown. The swellingwasmeasuredat pH 7 usingthe liquid retentionmethodwith the reduction-oxidationcycle. Table VIII Swelling of Oxidized Hair Liquid Retention (%) Cystinc Oxidized, White Hair, Oxidized with Brown Hair, Oxidized with Brown Hair, Oxidized with % of Original H202 H20• PeraceticAcid 0 54.5 50.2 50.2 10 63.2 63.6 64.0 13 68.8 70.5 67.0 15 78.5 97.8 77.2 At the same levels of disulfide bond oxidation the bleached brown hair appearsto be more damagedthan its white counterpart. This difference could be due to the breakdownof melanin, a contentionsupportedby the fact that the damagein brown hair can be lowered when the disintegrationof pigmentis prevented(peraceticacid oxidation). It is, however, obvious that the main source of damage residesin the destructionof disulfidebondswhich not only opensthe structureof hair but providesadditional hydrophilic centersin the form of cysteic acid residues. The introductionof theseresiduessignificantlyaltersthe swellingcharacteristicof hair as illustrated in Fig. 5. A novel feature of the swellingbehavioris its unusualdependenceupon pH. A precipitousincreasein swellingof bleachedhair occursin the pH region 5-7.5, while no suchchangeis observedin caseof untreated hair. This phenomenoncan be bestexplainedby acceptingthe chargerearrangement mechanismpostulated for oxidized wool by Thompson and O'Donnell (31). The stronglyacidic c¾steicacid residuesare, in the courseof their formation, being cmnpensatedfor by the ionized basic groupsof the keratin. As a result, the carboxylicgroupsof the acidic side chains are being expelled from their salt links and remain essentially un-ionized at pH 4 and below. This decreasein their acidity HAIR BLEACHING 895 .• 46 Figure 5. Effect of pH on swellingof bleachedhair (in the intact fiber their pK = 2) is broughtabout by the growth in negativechargedensityfollowingthe formationof the cysteicacid residues. AbovepH 5 the displaced carboxylicgroupsbeginto titrate; this is reflectedin an increasein swellingwhich reachesa maximum value as complete ionization is approached. The available evidence thus stronglysuggests that the ionizationof the carboxylicgroupsleadingto increasedswelling contributes to the resultant fiber damage. The ionizationeffectcan be convincinglydemonstratedon the behaviorof a hair tresswhichwasexposedfor a shorttime (10 min) to the oxidizing actionof a dilute solutionof peraceticacid (pH 3.5). When the treated tresswasevaluatedfor feel and combing,it wasindistinguishablefrom an intact tress. However, when the oxidized tress was subsequentlyim- mersedfor a few minutesin pH 9 buffer, rinsed, and rated again, its tactile and combingpropertieswere similar to thoseof bleachedtresses. Apparently,sufficientetchingof the keratin took placein the buffer to changethe surfacecharacteristic of the oxidizedfibers. 896 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Table IX Mechanical Properties of Oxidized Fibers Yield Stress,g/den Breaking Stress,g/den Breaking Extension, Min Dry Wet Dry Wet Dry Wet ... 51.4 Time, Fiber Treatment Untrcated control 3% H20•, pH 10, 35øC 3% H•O•, pH 10, 35øC 3% H•O•, pH 10, 35øC 0.5% CHaCOOOH, 25øC (then 0.1M NH4OH, •nin) 1.10 0.41 1.92 1.68 40.1 30 1.08 0.36 1.84 1.61 41.6 56.5 90 1.07 0.28 1.79 1.05 44.2 58.1 180 1.07 0.20 1.43 0.75 42.7 57.6 30 10 1.08 0.36 1.86 1.58 41.0 53.5 20 1.02 0.30 1.78 1.30 43.5 56.5 30 1.02 0.27 1.74 1.20 41.8 58.2 60 0.99 0.20 1.72 1.10 46.1 59.1 MechanicalPropertiesof BleachedHair The disulfidebondscontribute greatly to the wet strengthof keratin fibers,which decreases almostlinearly with the cystinecontent. On the other hand, the strengthof the dry fibersis not appreciablyaffectedby the breakdown of covalentcrosslinks, being dependentlargely on the main chain length and interchain hydrogenbonding. When viewed from this standpoint,the changesin mechanicalproperties of hair keratin brought about by bleachingcan be satisfactorilyinterpreted in terms of the oxidative attack on the disulfide bondsalone. Thus, we observe(Table IX) a steadydecreasein wet moduluswith increasedtime of bleachingand virtually no changein either the modulus or ultimate strengthof dry fibers. The latter observationsupportsthe view that the extent of the main chain scissionduring the bleaching processis negligible;it wasshownby Elod (32) that the breakdownof 1% of the peptidebondsin keratin bringsabout an almost20% lossin the dry strengthof the fiber. The dry breaking extensionis slightly affectedby bleaching. Certainly, no evidenceof the brittlenessoften referred to is found. This is not changedby varyingthe rate of strainingfrom 1 to 50 in./min. The apparentretentionof dry strengthby the bleachedfibersdoesnot result from a decreasein regain. On the contrary, the regain of the fiber increaseswith th? extent of bleaching(Table X) over a wide range of tested humidities. Bearing in mind the charge rearrangement involving the cysteicand carboxylicacid residues,it would appear that this increasein regain may be directly related to the ionization of the HAIR BLEACHING Table 897 X Effect of Bleaching on Moisture Absorption by Hair Regain (%) at Relative Humidity of: Sample 17% 41% 61% 72% 84% 94% 23.9 Intact 5.4 8.7 13.5 15.9 19.2 Bleached 30 rain 5.6 9.3 13.9 16.1 19.5 25.0 Bleached 120 rain 6.0 9.8 14.4 16.6 20.1 25.9 Bleached 240 rain 6.5 10.1 14.8 17.5 20.5 28.7 carboxylicgroupsand by depressingsuch an ionization a return to regain valuescloseto thoseof intact fiber could be attained. This indeed is the case. When bleachedhair is soakedbriefly in acid, then rinsed in deionizedwater to removeany bound acid and its regain is redetermined, the value obtained is almost identical with that of untreated hair. The wet mechanicalpropertieswere measuredwith fibers immersed in pH 7 buffer (Fig. 6). However, unlike the untreatedhair, the hydration of bleachedhair is stronglypH dependentand thus should manifest itself accordinglyin the mechanicalperformanceof wet fibers. Let 1.0 '•]•C"•'• • •'• reduced 0.9 0.8 • 0.7 oxidized _: o.6 F-0.5 I 0.4 0.3 0.0 • 2I 3I 4I 5I 6I 7I 8I pH Figure 6. ElFoctof pH on yield index of oxidized and reduced hair 898 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS us briefly considerwhat are, in terms of interchain bonding, the consequencesof charge rearrangementattendant upon bleaching. The displacementof carboxylicgroupsfrom their salt linkageswith positively chargedaminogroupsby cysteicacidresiduesgreatlystabilizesthis linkage. This is simplydue to the fact that cysteicacid residuesremain ionized at low pH valuesat which the carboxylategroups,even those in the intact fiber, would protonate,leading to destructionof the electrostatic bond. Thus, breakdown of the covalent disulfide bond is com- pensatedfor somewhatby formation of a more stable electrostaticcross link. In slightly bleachedfibers only a fraction of the cystineundergoesthe oxidative breakdown,and only a few of the carboxylic•oups are displaced[TOmtheir salt links and just as many new, strongerbonds of the sametype are formed. In an extensivelybleachedfiber, the displacementis virtually complete and theoretically,the strengthof the fiber shouldbe only slightlyaffectedby increasingacidity. Someexperimental support[or this view is derived from a brief study of the effect of pH on the mechanicalperformanceof bleachedand reduced fibers. In both casesthe fraction of broken disulfide bonds was closeto 40%. We have chosen to use reduced fibers rather than intact fibers as our controls becausethe importance of charge rearrangement has to be viewed againstbackgroundsof similarly disorganizedstructures. The following test procedure was employed: Both bleached and reduced hair were soaked in 0.01N HC1 for 6 hours at 25øC and then rinsed with freshchangesof deionizedwater until no more acid wasreleasedby the keratin. The fibers were then dried, mounted on tabs, and stretched 5% in deionizedwater (calibration step). They were then released, kept in deionized H20 for 12 hours, and transferred for an additional 12 hoursto buffer solutions,in which they were restretchedagain. The force to attain the yield point wascalculatedin both casesand the ratio Yield Yield force in buffer force-calibration denotedas the yield index (Fig. 6). The resultsconformsatisfactorily to the pattern expectedon the basisof our theoretical considerations. Thus, the bleached fibers exhibit a region of maximal mechanical stability between pH 3 and 5 where the displacedcarboxylic groupsare undissociated and those ionized are bound in the salt links. An increase in pH above5 leadsto the ionization of free carboxyls,the fiber hydration increases and so does the ease of its deformation. This be- havior is sharply contrastedby that of reduced fibers. With no free HAIR BLEACHING 809 carboxyl sidechainsto ionize,the regionof theirmechanical stabili'ty staysalmostunchangedup to pH 8. This is not so under acidicconditionswherereducedfibersshowa precipitousfall in yield force.. It is obvious that the combination nation of electrostatic of the disulfide bond breakdown and elimi- cross links have a disastrous effect on mechanical performanceof the fiber. Althoughbleachedhair alsoshowssomeweakening (apparentlya sizeablefraction of acid-labile,salt links is still present), thestabilizing effect of newelectrostatic bonds, involving •he cysteicacidresiduesand the chargedbasicgroupsof arginine and lysine is prominently evident. ACKNOWLEDGEMENTS The authorsare indebtedto the followingpeoplefor their assistance in providing someof the data presentedin this paper. The electron micrographicexaminationof melanin wascarried out by Mr. A.. Dano of the Gillette Safety Razor Research Laboratories in Boston. Dr. Kokoschka of the NationalBureauof Standards examined samples of melanin in a Varian X-band esr spectrometerand Dr. R. K. Brown of WayneStateUniversitydeterminedthe molecularweightof solubilized melanin by thin-layer gel chromatography. : (ReceivedMarch 10',197,0) REFERENCES (1) Schrotter,A. V., Wasserstoffhyperoxyd als cosmeticum,Chern.Ber., 7, 980 (1874). (2) Weber, I.E., Hydrogenperoxidebleaching,J. Soc.Dyers Colour.,39, 209 (1923). (3) Trotman, S. R., Conditionsgoverningthe bleachingof wool with hydrogen peroxide, Ibid., 42, 154 (1926). (4) Smolens,H. G., Hydrogenperoxidebleachingof wool, silk and cotton under chemical control, .4ruer. Dyest. Rep., 18, 123 (1929). (5) Holmes,J. F., Modern methodsof wool bleaching, Text. Color., 55, 250 (1933). (6) Smith, A., and Harris, M., Oxidation of wool: effect of hydrogen peroxide on wool, J. Res. Nat. Bur. Stand.,16, 301 (1936). (7) Smith,A., and Harris, M., Oxidationof wool: photochemical oxidation,Ibid., 17, 97 (1936). (8) Wilson,N. C., The scientificaspectsof bleaching,Text. J. Aust., 15, 496 (1941). (9) Elod, E., The structureof the wool fiber, Melliand Textilber., 23, 313' (1942). (10) Funatsu, M., Chemical studieson hair. Mechanism of the reaction of concentrated hydrogenperoxidewith hair, Nippon Nogei Kabaku Kaishi,32, 175 (1958);C. A., 52, 12000 (1958). (11) Laxer, G., and Whewell, C. S., The measurement of damage produced by treatment of wool with solutionsof hydrogenperoxide,J. Soc.Dyers Colour.,68, 256 (1952). (12) Zahn, H., Chemical changes of wool by washing, steaming and bleaching, Text.Rundsch., 19, 573 (1964). _ 900 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS (13) Zahn, H., Chemical processesduring the bleaching of wool and human hair with hydrogenperoxide and peroxy acids,J. Soc.Cosmet.Chem., 17, 687 (1966). (14) Miro, P., Destructionof the tryptophaneof wool by oxidationwith hydrogenperoxide, Bull. Inst. Text. Ft., 17, 1165 (1963). (15) Ziegler, K., Wool bleaching, Text.-Prax., 17, 376 (1962). (16) Piattelli, M., and Nicolaus, R. A., Structure of mclanins and melanogenesis. Structure of melanin in Sepia, Tetrahedron, 15• 66 (1961). (17) Nicolaus, R. A., Piattelli, M., and Fattorusso, E., The structure of melanins and melanogenesis, Ibid., 20• 1163 (1964). (18) Raper, H. S., The tyrosinase-tyrosine reaction,Blochem.J., 24• 239 (1930). (19) Green,D. B., and Happey, F., The infra-red spectraof melanins,Cirtel, 1, 283 (1965). (20) Laxer, G., Somepropertiesof pigmentedanimal fiberswith specialreferenceto bleaching, Ph.D. Thesis, Universityof Leeds,England, 1955. (21) Valko, E. I., and Barnett G., A study of the swelling of hair in mixed aqueoussolvents, .l. Soc.Cosmet.Chem., 3, 108 (1952). (22) Stahl, E., Thin Layer Chromatography,AcademicPress,New York, 1965. (23) Harris, M., and Br•wn, A. E., Bleaching of keratinous fibrous material. U.S. Patent 2,914,374 (1959). (24) Swift, J. A., cited from discussion following the presentationof Green'spaper; Cirtel, 1• 300 (1965). Duke, F. R., and Haas, T. W., The homogeneousbase-catalyzeddecotnpositionof hydrogenperoxide,J. Phys. Chem.,65• 304 (1961). (26) Btdl, H. B., Osmoticpressureof egg albumin solutions,J. Biol. Chem., 1•7• 143 (1941). (27) Burk, N. F., and Greenberg, D. M., The physical chemistry of the proteins in nonaqueousand mixed solvents. I. The state of aggregationof certain proteins in ureawater solution,Ibid., 87, 197 (1930). (28) Everett, A. J., and Minkoff, G. J., Dissociationconstantsof some alkyl and acyl hydroperoxides,Trans Faraday Soc.,49, 410 (1953). (29) Perkin, W. H., Someexperimentson the oxidizing action of hydrogenperoxide, Proc. Chem. Soc.,London, 23• 166 (1907). (30) Binns,F., and Swan,G. A., Oxidation of somesyntheticmelanins,Chem. Ind. (London), 1957• 397. Thompson, E. O. P., and O'Donnell, I. J., Comparison of the completenessof oxidation with peraceticacid and performic acid, Aust. J. Biol. Sci., 12, 490 (1959). Elod, E.. Nowotny, H., and Zahn, H., The structureand reactivity of wool, Kolloid-Z., 9.•, 50 (1940).
© Copyright 2018