92 UNIMOLECULAR FILM BALANCE RESEARCHES ON MQNQLAYERS PART 4.-A STUDY OF DYEING PROCESSES BY THE USE OP THE UNIMOLECULAR FILM BALANCE BY Mrss MARGARET M. ALLINGHAM, C. H. GILES AND E. L. NEUST~DTER Dept. of Technical Chemistry, Royal Technical College, Glasgow, C. 1 Received 29th June, 1953 Possible modes of dye-fibre combination have been studied by determination of the effect of dyes, or model compounds representing dyes or fibres, as solutes in the aqueous layer below monolayers of a variety of surface-active compounds, the interpretation of the interactions being based on measurements of molecular areas and compressibilities. Cetyl acetate, representing cellulose acetate, and dimethylterephthalate, representing Terylcne and used in a mixed monolayer, both gave evidence of the formation of inter- molecular CH .. . N and CH . . . 0 bonds and it is suggested that this type of bond may be responsible for the attachment of dyes to these fibres. The known facts con- cerning the sorption of dyes by cellulose from aqueous solutions are more consistent with a mechanism of van der Waals’ attraction between fibre and dye than with one involving hydrogen bonds. It is postulated that the carbohydrate molecules are too firmly attached to water for hydrogen bonds with dyes in aqueous solution to be formed. Surface-active planar and non-planar benzidinc disazo-dyes representative of a direct cotton dye and a non-substantive isomer have been used in experiments to determine whether van der Waals’ attraction alone operates between such dyes and solutcs repre- senting cellulose.It is found that the planar dye does associate with cellobiose molecules which, as anticipated, remain surrounded by a layer of water, and so are prcvcnted from forming hydrogen bonds with the dye. Regarding the dyeing of proteins and Nylon, some evidence of intermolecular hydrogen-bonding involving aniide groups is obtained, but it is pointed out that this does not always agree with data obtained from sorption experiments on the fibres them- selves, perhaps because the fibre molecules are not always able to form intermolecular bonds with solutes, owing to competing intramolecular forces. The extensivc studies which have been devoted to the chemistry of dycing and relatcd processes in the last two decades have produced clear evidence of the oper- ation of elcctrovalent bonds as one of the important forms of linkage binding dyes to two types of fibre, viz., proteins and Nylon.16 With this exception, however, the nature of the attractions providing the driving force of the dyeing mechanism in fibres has not been clearly elucidated, though the operation ofMARGARET M.ALLINGHAM, c . H. GILES AND E. L. NEUSTADTER 93 hydrogen bonding and van der Waals’ forces have been suggested and to some extent demonstrated as actually involved in some types of dyeing. It has therefore seemed desirable to bring to bear on this problem an in- vestigation of the mutual hydrogen-bonding reactivities of organic radicals typical of the principal groups occurring in fibres and dyes. Dyeing operations are virtually always carried out in an aqueous medium and it was thought particularly important therefore to study the appropriate reactivities in presence of water.The method used in this research is that employed in the preliminary experiments already described : 4 the examination of the intcraction of pairs of hydrogen-bonding compounds as monolayer and solute, respectively, by means of thc Langmuir-Adam unimolecular film balance. So far the work has been confincd simply to measurements of film area and compressibility, but nevertheless, the results seem to havc valid use in interpretation of some dyeing phenomena, as we hope to show in this paper. A parallel investigation has bccn procccding 1 s 5 into the possibility of detecting the molar ratio of hydrogen-bondcd complexes of solutes in binary solution in a variety of solvents, including water.Some of thc prcdictions made as a result of this investigation have been tested in the present work. In thc previous experiments,4 surface-active compounds containing, e.g., amide, azo-, hydroxy- or quinonc groups, were spread on aqueous solutions of solutes containing hydrogen-bonding groups, principally phenols, and the changes in film arca and compressibility noted. Monofunctional solutcs, e.g., monohydric phenols, appear usually to cause increased solvation of groups with which they can rcact in the monolayer ; this alters the tilt of unsymmetrical molecules, and usually increases the film area and compressibility. Bifunctional solutes, e.g., dihydric phenols, react in a more distinctive manner.If the two hydroxy-groups are suficicntly separated in the solute molecule, they appear to act as a cross-link betwcen two molecules of the surface film, and this leads to a considerable increase in both area and compressibility of the film. In the special cases where the film molecule happens to contain two proton-acceptor groups separated by a distance approximatcly equal to that between the two hydroxy-groups of the solute, the two molecules appear to join in parallel as a 1 : 1-complex, e.g., quinol as solute combines thus with a benzoquinone surface-active derivative. The effect on the film in these cases is to increase the area slightly while the compressibility is hardly affectcd. Solutes without groups capable of hydrogen-bonding with the mono- layer molecules appeared to have no significant effect on the film.There were some indications also that when more than two functional groups are present in a solutc molecule, those most widely separated determine the degrce of film expans ion. Thc pronounced effect of the cross-linking agents affords a ready means of examining the hydrogcn-bonding reactivity of groups in an aqueous medium. The gcneral procedure is as follows : suppose it is desired to determine if hydrogen- bonding is possible bctween two groups, A and B. Thc group A is introduced into the molecule of the surface-active compound, and a solute molecule is chosen having a pair of groups, B, B, placed some distance apart. If hydrogen-bonding takes place between monolayer and solute molecules we may envisage a film structure built up as shown in fig.1, where A, A, are the heads of the monolayer molecules and B-B those of the bifunctional solute. If A is small and comparable in dimensions with the width of the molecules BB when tightly packed, then the whole structure should be quite compact and the increase in film arca by introduction of the solute should be predictable from a knowledge94 UNIMOLECULAR FILM BALANCE of the dimensions of the molecule BIB. Further, if more than onc solute compound containing pairs of groups BB, at different separation distances in the molecule, can be used, then the film area should increase with this separation, and the observation of such an effect would be a confirmation both of the bonding pro- perties of the groups A and B, and of the validity of this hypothesis of cross- bonding in the film.This procedure, as well as thc use of monofunctional solutes, has bcen followed here. The following representative compounds have been employed :- Model compounds representing fibres cellulose : cellulose acetate : cetyl acetate. Nylon and protcins : Terylenc : dimcthylterephthalate. cellobiose, glucosc, mannitol, sucrosc. glycine, N-methyl stearamide, octadecyl acetamidc. Dyes and rcppr-eseiitative coinpounds Crystal Violet (Colour Indcx No. 681) L-/ f Acid Magenta (Colour Indcx No. 692) CH3 SO3- Orange I (Colour Index No. 150) 3 : 3’-diethylthiacyanine iodide 3 : 3’-diethylthiacarbocyaninc iodidc NMARGARET M . ALLINGHAM, c. H . GILLS AND E . L . NEUSTADTER 95 3 : 3’-diethylthiatricarbocyanine iodide S ~s>-CH=CH-CH=CH-CH=CH-C€€==C \” ”ii N + \ / -I C2K5 Alizarinc Cyaninc WRS (Colour Index No.1063) (1 : 2 : 4 : 5 : 6 : 8-hcxahydroxy- anthraquinone sodium 3 : 7-disulphonate). 1 : 4-Diaminoanthraquinone-2-sulphonic acid (sodium salt), 1 : 5-dihydroxynaphthalene. disazo dyes from 2 : 2’- and 3 : 3’-dimethylbenzidine, coupled respectivcly with p-octa- decylphenol, p-dodecy laniline --> j?-naphthol, 2-p’-dodccylphenyl-p-bcnzoquinone, p-hexa- decyloxyanilinc - > o-cresol, y-~nethoxystearylphenonc, phenol, quinol, stearylethyl ketone. EXPERIMENTAL APPARATUS AND MATERIALS, etc.-The film balance and general technique used have already been described.3 The reagents were all purified by recrystallization, or were of analytical reagent quality.Alizarine Cyanine WRS (Bayer) was a commercial sample (purity, 62 %, the remainder probably being sodium chloride). Catalin (Stuart-type) atom models were used for determination of molecular dimensions. Microanalyses were by Dr. A. C. Symc, Royal Technical College, Glasgow. Light absorption curves wcrc determined on a Unicam SP 500 photoelectric spectrophotometer. CEI’YL AcETATE.-Dry hydrogen chloride gas was bubbled through a mixture of cetyl alcohol (0.3 mole) and glacial acetic acid (0.6 mole) for 1 h ; the mixture was then warmed several hours on the water bath. The acetate was isolated by vacuum distillation ; imp. 22” c. disAZO-COMPOUNDS.-~ecrysta~~izcd samples of 0- and rn-tolidine sulphates were tetrazotiscd and coupled with p-octadecylphenol (prcpared as previously described 3), dissolved in aqueous alcoholic sodium hydroxide.The precipitated disazo-compounds were filtered off, washed, dried and (Soxhlet) extracted with toluene. The solutions were orangc-brown and showed absorption peaks at 3100 A (3 : 3’-isomer) (sharp) and 3500 A (2 : 2’-isomer). Thc solid substances, which did not crystallize, werc orange-brown and were obtained by evaporation of the solution. M.p., 3 : 3’-isomer, 93” C ; 2 : 2‘- isomer, 95” C (found : 16, 6.2, 6.4 %, respcctively ; C22H9402N4 requires N, 6.05 %). ~-oCTADECYLACETAMrDE.-Stearoy~ chloride (0.1 6 mole) was added carefully to excess of an ice-cold solution of ammonia (sp. gr. 0-SSO), the precipitated amide filtered off, washed, dried, and treated with thionyl chloride (0.3 mole) to yieId p-stearonitrile.This was reduced by refluxing several hours with excess sodium in cthanol. On addition of hydrochloric acid and cooling, the amine hydrochloride crystallized out. This was distilled over quicklime, the distillate warmed a few minutes with acetic anhydride, and the product crystallized from glacial acetic acid with the aid of charcoal ; m.p. 79” C. The preparation of the other surface-active compounds has already been dcscribed.3~ DISCUSSION ‘Tables 1 and 2 summarize the expcrimcntal rcsults, and comments Icgarding their relationship to dyeing phenomena now follow. (A few results of thc pre- liminary work4 have been includcd where necessary for comparison, and these arc suitably distinguished in the tables.) The tangent of the angle made by the uppcr part of the force-area curve with the y-axis is a convenient measure of film compressibility, and has been quoted here in this sense. The interpretation of the results should give decisive information about solute groups which cannot form specific chemical bonds with fibre substrates. Informa- tion regarding groups which might form such bonds is less decisive, because of the uncertain factor introduced by the presence of inter- and intra-chain bonds in thc substrate itself.Unless the affinity of the solute is suficient to enable it to replace such bonds, it may not be able to co-ordinate with the substrate.TABLE 1 ,-MOLECULAR AREAS AND COMPRESSIBILITES OF MONOLAYERS Molecular area (812) at zero compression and compressibility t of surface-active substances substrate I water 11* Acid Magenta 111" D-cellobiose IV* Crystal Violet V* 1 : 4-diaminoanthraquinone sodium VI$ 3 : 3'-diethylthiacyanine iodide 2-sulphonate VIl$ 3 : 3'-diethylthiacarbocyanine iodide VIII$ 3 : 3'-diethylthiatricarbocyanine iodide IX* 1 : 5-dihydroxynaphthalene X** dimethylterephthalate (1 : 2-ratio) XI* cc-D-glucose XII* glycine XIII* 1 : 2 : 4 : 5 : 6 : 8-hexahydroxyanthra- quinone sodium 3 : 7-disulphonate XIV* mannitol XV* phenol XVI* sucrose XVlI* sulphanilic acid 3 a-naphthol XVIII* quinol Data in brackets are from previous paper.4 a b c 24, 0.04 125, 0-25 255, 0.27 84, 1-15 250, 0.65 305, 0.51 41, 0.84 35, 0.32 65, 0.42 89, 0.29 127, 0.70 48, 0-60 75, 1-66 210, 0.60 28, 0.05 135 (?) 230 (?) 59, 1.73 d e f (43, 0.16) (26, 0.02) (37, 0.07) 30, 0.17 (54, 0.31) 100, 0-84 59, 0-40 54, 0.34 45, 0.40 49, 0.32 (102, 0.42) 8 h i j 36, 0.11 (21, 0.03) 22, 0.09 (48, 0.09) 34, 0.11 85, 0.32 (44, 0.36) 90, 0.58 48, 0.16 (23, 0.05) 50, 0.53 28, 0.11 22, 0.09 50, 0.09 40, 0.10 50, 0.06 72, 0.65 (29, 0.08) 42, 0.19 (55, 0.05) j- expressed as the tangent of the angle to the vertical of the upper portion of the lI-A curve (scales as in fig.2). * 0.1 M aqueous solution. u = cetyl acetate ; $ 0.00025 M aqueous solution. ** used in binary solution with surface-active compound. b, c = 3 : 3'- and 2 : 2'-dimethylbenzidine + (p-octadecylphenol)z, respectively ; d = p-dodecylaniline -+ /3-naphthol ; e = 2-p'- dodecylphenyl-p-benzo-quinone ; f = p-hexadecyloxyaniline --f o-cresol ; g = p-methoxystearylphenone ; h = N-methylstearamide ; i = N-octadecy- acetamide ; j = 1-stearamidoanthraquinone.MARGARET M.ALLINGHAM, c . H . GILES AND E. L. NEUSTADTER 97 Robinson and Ambrose,13 e.g., have drawn attention to evidence that polypeptide molecules in water have all their hydrogen bonds satisfied by inter- or intra-chain linkages and none is left to co-ordinate with the water. CELLULOSE ACETATE.-Marsden and Urquhart 8 as the result of a thorough investigation of the sorption of phenol and related compounds by cellulose acetate, concluded that sorption occurs by reason of the formation of a hydrogen bond between the phenolic hydroxy-group and carbonyl oxygen in the acetyl group, CH3 I giving O H ------ O=C I , and this suggestion has been extended by Vicker- c I staff 16 to cover the dyeing of cellulose acetate fibre with the commonly employed " disperse " dyes. These are anthraquinone or azo-derivatives devoid of powerful water-solubilizing groups, and are necessarily applied as dispersions in water.They are probably enabled to penetrate the fibre by virtue of their extremely slight water-solubility.17 Most of these dyes contain potential proton-donating groups, usually -OH and -NH2, which might well combine in the manner suggested for phenol. TABLE 2.-AREAS (A2) AT COMMENCEMENT OF DEVELOPMENT OF MEASURABLE SURFACE PRESSURE AND AT LOWEST POINT OF UPPER n-A CURVE monoIayer a b C d e f 6 h i i substrate I 25, 20 250, 90 320, 220 (32, 25) (38, 35) 44, 34 (23,21) 40,21 (54,48) I1 90, 80 111 350,190 380,250 35, 30 39,33 85, 85 IV 72, 31 VI 70, 61 VII 95, 85 VIII 134,120 IX X XI 64,56 90, 90 V 38, 33 48,45 XI1 56,39 53,47 XI11 75,75 XIV 350, 150 XV XVI 48,26 380 330 36,29 25, 18 57,48 XVII 70,51 41,40 52,49 XVIII 95,65 45,41 56, 53 59,46 There are, however, two facts which tend to throw doubt upon the exact characterization of the hydrogen-bonding mechanism in this substrate.These are : (i) certain dyes which have affinity for cellulose acetate, e.g., nitroazobenzene derivatives (for formulae, see mighty), have no hydrogen atom in their molecule available for bonding ; (ii) the carbonyl group in ketones is protected against intermolecular hydrogen- bonding in water by the solvent and this type of protection appears also to occur with the carbonyl oxygen in the acetyl group, though this evidence is not conclusive.1 It is now found 1 that proton-acceptors, e.g.azobenzene or benzoquinone, can form complexes with aliphatic acetates, presumably through a hydrogen atom of the methyl residue in the acetyl group, which is activated by the adjacent carbonyl oxygen atom. In the present work we have sought to confirm tbis by using a film-forming acetyl compound, cetyl acetate, and studying the influence of proton-accepting D98 UNIMOLECULAR FILM BALANCE solutes, particularly those with two functional groups, beneath it. The particular solute compounds chosen included a series of three cyanine dycs with increasing lengths of nicthin chain. It seems justifiable to assume that the two nitrogen atoms, because of rcsonancc, may bc considered equally able to form a hydrogen bond with a proton-donator group in the monolayer and these substances will therefore be bifunctional. For comparison, a few other solute compounds, of symmetrical and unsymmetrical types, but containing proton-donating groups were examined also (fig. 2).TabIe 3 shows the apparent molecular areas, predicted and observed, of cetyl acetate on the various substrates ; considering the difficulty of determining the exact orientation of the molccules in the film the agreement is thought to be reasonably good. (With the particular atom models used, the thiazole ring in Molecular area (A2). FIG. 2.-Force-area curves for cetyl acetates in various substrates- V : Cyanine dye (VIII). Left to right I : water. I1 : sucrose. I11 : Cyanine dye (VI).IV : Cyanine dyc (VII). VI-VIII is difficult to construct and data for these three compounds are con- sidered less reliable. The increasing departure of the area increases of these compounds from the predicted values may be an indication of the enhanced water-attraction of the longer conjugate chains.) In calculating the dimensions of each AB-BA complex (cp. fig. 1) it has been assumed that the following criteria must be satisfied. Each of these has been formulated from a consideration of the known properties of the groups concerned, especially as revealed in their behaviour in monolayers or in hydrogen bonding experiments (cf. 1 9 3 9 49 5). (i) Each sulphonic acid group is surrounded by a layer of water approximately one molecule deep. (ii) Each carbonyl group in the acetate radical is associated with one water molecule. (iii) The unsynlmetrical sulphonates (XVII and V) when associated with the film molecules, stand with their sulphonic acid groups deep in the water and their main molecular axis oriented at ca.30" to the vertical (cf. 3), since the molecules then appear to be better balanced relative to the water phase.MARGARET M. ALLINGHAM, c. H. GILES AND E. L . NEUSTADTER 99 (iv) Where the solute molecules contain bulky substituent groups, e.g., SO3-, or are non-planar and their close mutual association is thus hindered, it is assumed that the two associated cetyl acetate molecules are accommodated in the inter- molecular space at the side of the solute molecules. With compounds XIII, 11, XVII, V and 1V this does not require the solute molecules to be disturbed from a close-packed position; with VZ, VII and VlII they must be forced apart somewhat to offer the requircd accommodation.(v) The mutual association between thc molecules of IX, which are simple and planar, is assumed to be sufficient to prevent interpenetration by molecules of cetyl acetate ; these must, therefore, be accommodated at the ends of the solute molecules. TABLE 3.-EFI-’ECT OF SOLUTES ON FILMS OF CETYL ACETATE ditnensions of AB-BA complex in surface planc (est.) (A) film area (A21 pcr mol. cctyl acetate no. VI VII VIII XI11 I1 IX XVII V IV solute (water) 3 : 3’-diethylthiacyanine iodide 3 : 3’diethylthiacarbocyanine 3 : 3’-diethylthiatricarbocyanine Alizarine Cyanine WRS Acid Magenta 1 : 5-dihydroxynaphthalene Orange I 1 : 41diaminoanthraquinone- Crystal Violet iodide iodide sodium 2-sulphonate lcngth 5 CaPPr.) 14.0 16.5 21.5 17.0 15.0 18-0 10.0 8.5 16.0 breadth 5 (appr.) 9.5 9.5 9.5 10.0 11-5 5.0 10.0 10.0 5.0 predicted found - 24 66 65 78 89 105 127 85 75 86 84 45 48 50 59 43 35 40 41 (vi) Each solute molecule is specifically attached to two cetyl acetate molecules, thus the apparent cross-sectional area of the complex AB-BA in the plane of the water surface, estimated according to the above criteria, must be halved to determine the predicted area per molecule of cetyl acetate.The results are considered good evidence that the acetate group in cetyl acetate, and therefore also in cellulose acetate, can form intermolecular hydrogen bonds with dyes and the behaviour of compounds VI-VIII suggests that the acetate group can act as the hydrogen donor.Whether it behaves thus with all types of dye is not certain, but seems likely.1 The commercial product is partially hydrolyzcd (approximately to thc diacetate) and the hydroxy-groups may also operate, but the triacetate can be coloured with the same type of dyes as the diacetate, so these groups cannot be very important in determining dyeing effects. TERYLENE.-Dimethyl terephthalate, rcpresenting this fibre, was incorporated in the solution used for spreading the film, on account of its low water-solubility. The surface-active compound used, p-dodecylaniline --t p-naphthol, has no free hydrogen, owing to the internal chelate ring it contains, and the undoubted evidence of cross-bonding obtained (tables 1 and 4) must therefore represent bonding between an azo-nitrogen atom and a hydrogen of the methyl ester group.In binary solutions also, both this ester and ethylene glycol dibenzoate, which more nearly resembles the Terylene polymer, appear to form bonds by a similar mechanism.1 At the 2: 1 azo compound/ester ratio the film niust be rather compact and as more cster molccules are added they at first cause a slight expansion and then appcar to be forced into the water below. * Note added in prooJ-The data are not inconsistent with a parallel association of azo compound and ester [see comment (by C. H. G.) on paper by Schuler and Remington], but the hydrogen bonding mcchaiiism is considered more liltely.100 UNIMOLECULAR FILM BALANCE TABLE 4.-DATA FOR MIXED FILMS OF DIMETHYL TEREPHTHALATE AND 0-HYDROXYAZO COMPOUND ratio, ester/azo-componnd ii iii 0.25: 1 58 70 0.5: 1 100 0.84 125 1 : 1 I35 0.58 145 2: 1 132 1-43 145 i apparent mol.area at zero pressure (Az), ii tangent to vertical (cf. table 1). iii commencement of curve (A2) (cf. table 2). iv commencement of upper portion (A2) (cf. table 2). iv 80 125 115 In connection with both cellulose acetate and Terylene we have therefore some evidence in favour of -C-H . . . 0- or -C-H . . . N- bonds being operative in dye sorption. This type of hydrogen bond has not been suggested before in connection with dyeing phenomena, but though not common, several examples of it are known amongst organic compounds (cf. ref. (1 l), (18)).CELLuLosE.-It is well established that for dyes to manifest substantivity for ccllulose they must have a long planar molecule, which can approach close alongside and lie parallel to the cellulose chain.69 10916 It has previously bcen assumed that the binding force holding the dye to the fibre lies in hydrogen bonds formed between hydroxy- (or ether) groups of cellulose and suitable groups in the dye. There are, however, a number of objections to this belief, especially the following : (i) it is not clear why the shape and planarity of the dye molecule so critically determine substantivity, since potential hydrogen-bonding groups are so abundant along the cellulose chain; (ii) Marsden and Urquhart 8 obtained evidence of considerable sorption of phenol from water by cellulose acetate, almost certainly due to hydrogen bonding, yet they were unable to detect any sorption wf this compound by cellulose under similar conditions; (iii) there is no correlation between the number of hydrogen-bonding groups in dye mole- cules, even if they are planar, and their substantivity for cellulose.* HYDROGEN-BONDING OF CARBOHYDRATES It has now been found 1 that the hydroxy- and ether groups in the ring forms of soluble carbohydrates, e.g.glucose or cellobiose, are unable to form inter- molecular bonds with other solutes in water, because of protection by the solvent, and a similar effect may account for the non-absorption of phenol by cellulose. The conclusion may thus be drawn that dyes do not readily form hydrogen bonds with the fibre in water, and that van der Waals’ forces alone are operative in determining their substantivity.This conclusion conforms with the known facts of substantivity on this fibre and explains why the planarity and shape of the dye molecule is such a critical factor, since the van der Waals’ forces attenuate far more rapidly with distance than electrostatic forces and, to be sufficiently effective, they will therefore require the closest possible approach of the dye molecule to the fibre. It should be explained here that the aldehyde group, unlike the ether and hydroxy-groups, is not prevented from forming intermolecular bonds by water, and the aldehyde group in the open-chain tautomer of glucose and cellobiose does in fact seem to co-ordinate with other solutes in water.1 This disturbs the equilibrium between the pyranose ring carbohydrate molecules and their open-chain forms in favour of the latter, and a noticeable bonding affinity of the aldehyde group of these compounds for, e.g., amines, results.There is, however, no evidence that under normal circumstances in water any form of * cf., e.g., (a) Benzo Violet A (C.I. no. 388) and (6) Orange I. These can form 2 and 3 intermolecular hydrogen bonds, respectively, yet (a) dyes cellulose and (b) does not. The molecule of (a) is, however, about twice as long as that of (b).MARGARET M . ALLINGHAM, C. H . GlLES AND E. L . NEUSTADTER 101 bond other than this one is operative. In cellulose, of course, the proportion of aldehyde end-groups is too small to be of significance in dyeing operations.The keto group, unlike the aldehyde group, is rendered inoperative towards other solutes by water so that the ketose sugars are unable to form any sort of intermolecular bonds in aqueous solution, and no complexes are observed between, e.g., fructose or sucrose and amines in water.1 The hydroxy-groups in open-chain alcohols, e.g., mannitol, are rather more reactive in water than those in carbohydrate rings, towards other solutes. With these facts in mind, we may consider the significance of the present monolayer experiments using carbohydrate solutes. In the first instance, if any of the soluble carbohydrates discussed did actually, through their hydroxy-groups, form bonds with film molecules, we should expect them to act as cross-linking agents, considerably expanding the film.Moreover, the longer molecules, cello- biose and sucrose, should have about twice the ef&t of the shorter glucose molecule, because our experience with these films leads us to believc that polyfunctional solutes act as cross-linking agents by means of the most widely separated active groups in their molecules. Secondly, if the hydroxy-groups are quite inoperative, then sucrose should have no effect on the films, while glucose or cellobiose should behave very similarly to each other, acting as monofunctional solutes. They should thus cause increased water attraction of the monolayer, by co-ordination therewith, and, in fact, should have a somewhat similar effect to phenol in this respect. The results shown are consistent with the second set of conditions and incon- sistent with the first.Sucrose has no effect, or at most, a slight expansive action, far less than cross-linking would produce ; in one case (c) it decreases the area slightly. Probably the expansion represents a very weak affinity of keto-groups for monolayer molecules. On the other hand, glucose and cellobiose behave very similarly to each other and have considerably more expansive action. MODEL SURFACE-ACTIVE COTTON DYES In order to obtain further evidence having a possible bearing on this problem, experiments were next made with surface-active compounds structurally similar to the benzidine series of direct cotton dyes. It has been suggested 10 that planar dyes of this series are highly substantive to cellulose because the repeat interval (10.8 A) of potential bonding groups, e.g., -N2-, in their molecules corresponds with that (10.3 A) of the cellobiose unit of cellulose, and thus the best conditions are satisfied for groups taking part in mutual hydrogen bonding, to make close contact.3 : 3’-Disubstituted benzidine derivatives give planar, substantive dyes ; 2 : 2’-derivatives give non-planar, non-substantive dyes. Disazo-compounds were thereforc prepared by coupling tetrazotized u- and rn-tolidine, respectively, with p-octadecylphenol, to give the two disazo-compounds shown in fig. 3. c18H37 s - - - brN===yfc>-<>-.-<~ - - - s CH3 ‘**.. / HO I planar I1 non-planar FIG. 3 .102 UNIMOLECULAR FILM BALANCE Both compounds were found to form stabIc films, of large apparent moIecular area, and high compressibility, as might be expected (fig.4). A consideration of the symmetry of these molecules suggests that they must rest on the water very much as pictured in fig. 3, the water surface SS being normal to the plane of the paper. If the repeat distance at right-angles to the longest molecular axis is 6 & as it is for the monoazo surface-active compounds,3 where a water molecule appears to be interposed between each pair of azo-groups, then the cross-scctional area of the planar (3 : 3’-) molecule in the film should be about 140A? The actual value found is 125 A2, so that the basis of prediction of the mode of orientation RG. +Force-area curves for surface-active disazo-compounds on various substrates, (i) compound b (planar) on water alone.(ii) compound b on cellobiose solution. (iii) compound c (non-planar) on water alonc. (iv) compound c on cellobiose solution. and packing appears fairly well founded. It is more difficult to predict the cor- responding value for the non-planar isomer, because its irregular shape must allow some interlocking to occur, but it will certainly be larger than that of the planar compound. Actually it is found to be 255 A2. EFFECT OF SOLUTES ON PLANAR ISOMERS The effcct of some carbohydrate and polyhydric alcohol molecules, as solutes in the water, upon the monolayers of these two compounds was next examined. Mannitol was chosen as a suitable polyhydric straight-chain alcohol. As already stated, the hydroxy-groups in this type of compound appcar to be able to form intcrmolecular bonds in water more readily than do those of the carbohydrates and, indeed, apparcnt cross-linking effects observed with the Cz-compound, ethylcne glycol, have already been reported.5 If mannitol acts by cross-linking, considerable expansion of the film should occur.The mannitol molecule will be expected to assume a compact, crumpled form between the azo molccules. The most compact arrangement gives a separation of ca. 6-5w between outermost hydroxy-groups. If this form wcre interposed between the azo-groups in each pair of the planar disazo-molecules the repeat distance wouid be cu. 9A, and the apparent molccular area of the disazo conipound ca. 215A2. ActualIy, a value of 210 A2 is found. This seems to substantiate the view that cross-linking occurs.Thc results to be expected to follow from the use of carbohydrates as solutes will depend upon the extent to which their hydroxy-groups (and potential aldchyde groups) are capable of intcrmolecular bonding, and several possible modes of intcraction may occur. Consider, e.g., the planar isomer spread on a solutionMARGARET M. ALLINGI-IAM, c. H . GILES AND E . L. NEUSTADTEK 103 of cellobiose. This might form a complcx in which the carbohydrate functions as a straight-chain aldehyde, as in binary aqueous solutions. An aldehyde group would in this case combine with cach azo-group and the rest of the carbohydrate molecule would probably remain dissolved in the underlying water. A study of table 1 will show that cellobiose (or glucose) acting probably in this way, in- creases the film area by from 35 to 100 %, the increase rising with differcnt types of molecule in the order of decreasing closeness of packing on water (i.c.benzo- quinone compound, anthraquinone compound, azo-compound 4s 3). In the present case, therefore, we should anticipate an increase of ca. 35-40 %, i.e. an apparent molecular area of ca. 170 &. In view, however, of the undoubted fact that in cellulose the hexose rings of the carbohydrate chain have strong enough attraction for long planar benzidine disazo-dye molecules to hold them closely in parallel, we may suppose this type of union to occur in the film. If so, the hexose ring-forms of the carbohydrate will be favoured in the equilibrium between these forms and the open-chain aldehyde, as is normally the case in watcr.We should then expect to have a 1 : 1- complex between disazo-molecule and cellobiose, which may involve either of two modes of packing, depending on the nature of the attractive force between thc molecules. On the one hand, if hydrogen bond attraction holds the two molecules together they must be in close contact. This requires a repeat distance between azo-compound molecules corresponding to an apparent molecular area of ca. 170A2. (If we suppose that the hydroxy-groups on the side not attached to the azo-compound remain bonded each to a water molecule, this value rises to ca. 185 A2.) On the other hand, if the carbohydrate molecules firmly retain a completely protecting “ atmospherc ” of water molecules, which is suggested 1 as the cause of their inability to form intermolecular bonds in water, and if this covering is one molecule thick on both sides of the hexose rings, then the repeat area should be ca.270 A2. Actually it is found to be 250 A2. This value, there- fore, seems to favour the hypothesis that no bonds are formed between azo-groups and carbohydrate hydroxy-groups. EFFECT OF SOLUTES ON NON-PLANAR ISOMERS The non-planar isomer cannot form a close-packed arrangement with cellobiose molecules. If cross-linking through hydroxy-groups were, however, to take place, a large area increase would occur. The precise magnitude of this increase is difficult to predict from models, but it appears that a change of 150-300% in the apparent area per disazo-molecule is likely.If, however, such bonds do not occur, and the solute behaves monofunctionally through the aldehyde group, we may predict an area increase (see above) of ca. lO%, or perhaps a lower value, because of the greater possibility of adjustment of the non-planar film molecules by interlocking. The actual increase observed is about 20 %. CONCLUSIONS The general conclusions, which are admittedly speculative and based on only the limited amount of indirect evidence which it has been possible to obtain in the time available, are that the results are consistent with the view that dycs are held to cellulose fibres by van der Waals’ attraction rather than by hydrogen bonds. Glucose and cellobiose in water appear to act normally as monofunctional aldehydes towards monolayer molecules.In the special case, however, when the monolayer molecules are long and planar, the increased van dcr Waals’ attraction for the solute favours a 1 : 1-longitudinally arranged complcx, in which the mode of union resembles that of dyes with cellulose. ETHYL cELLuLosE.--It was hoped to obtain more information upon the re- actions of cellulose by a study of a film-forming derivative, ethyl cellulose, a commercial material, containing some residual hydroxy-groups. A suitable104 UNIMOLECULAR FILM BALANCE material of this type was spread and found to give an apparent molecular area of about 40A2 per hexose unit. Expansion of the film does occur with some hydrogen-bonding substances, but a full interpretation of the results has not yet been made. The material appears to resemble cellulose acetate in dyeing properties.This suggests it has available hydrogen-bonding groups. These might be either the ether groups, which other work1 has shown can act as proton-donors under some circumstances, or the residual hydroxy-groups, which may have different properties in this environment than they do in cellulose itself. Further work is in progress on this subject. PROTEIN FIBRES.-The maximum sorption of a variety of inorganic and organic acid anions on wool is closely equivalent to the content of free amino-groups in the fibre and there seem good grounds for the belief that combination of anions with charged amino-groups by electrovalencies is responsible for the sorption.14 As the size and complexity of the anions increascs, the titration curve of wool with organic acids moves towards the alkaline side, i.e., the maximum sorption is reached at lower acidities.There is no close relationship between affinity and dissociation constant of the acid,14 and this increase in affinity must be attributed either to an increasing tendency for the large anions to form hydrogen bonds with -CONH-, -NH2 or -COOH, etc., groups in the fibre, or to in- creased van der Waals’ attraction, or to both causes 9 s 14s 15. This uncertainty has not yct, as far as we are aware, been resolved, though the very high sorption by wool of unionized organic acids of small molecular size 15 would seem to be attributable to hydrogen-bonding. Thc groups named above can undoubtedly form intermolecular bonds with other groups typical of dye molecules, and even with dyes themselves.This is demonstrated by the present * and previous experiments 4 and by examination of binary solutions.1. 5 Yet there appears to be no very clear correlation between the information so obtained and the evidence of sorption experiments of simple organic compounds on wool.2 We must therefore recall Robinson and Ambrose’s comment 13 quoted above, and entertain some doubt as to whether many solutes can enter into hydrogen-bonding attachment with protein fibres unless they have high affinity and are able to break down bonds already present. Phenol is able to do this, and the results of actual sorption tests with this compound are in better agreement with experiments on model compounds in binary solution or mono- layers.2 NYLoN.-The comments upon protein fibres apply also to Nylon.Anions of acids and dyes appear to combine with charged amino-groups in the fibre, which are far fewer in number than those in wool, and at low pH values the greatly increased uptake of dye anions is attributed to attraction by the amide groups, which have then become positively charged.12 This type of elcctrovalent link has not been so clearly demonstrated in wool. In attempting to study hydrogen- bonding with Nylon, there is again found to be difficulty in correlating the present and related experiments with actual sorption tests? This investigation is being continued and further results of our work on proteins and Nylon will be reported later. The authors express their thanks to Prof. P. D. Ritchie for his interest and encouragement, to Mr. A. Clunie for the construction of the film balance, to Dr. J. D. Kendell and Dr. J. Whetstone for supplying pure samples of cyanine dyes, and an anthraquinone dye, respectively, and to Imperial Chemical Industries Ltd., Dyestuffs Division, who supplied several intermediates. One of us (E. L. N.) is indebted to Tmperial Chemical Industries Ltd., Dyestuffs Division, and another (M. M. A.) to the Wool Textile Research Council, for financial assistance. * cf., c g . , the increase in area of films of the amides when spread on Orange I solutions and of azo-compounds spread on glycine.MARGARET M. ALLINGHAM, c. H . GILES AND E . L. NEUSTXDTER 105 1 Arshid, Giles, McLure, Ogilvie and Rose, to be published. 2 Chipalkatti, private communication. 3 Giles and Neustadter, J. Chem. SOC., 1952, 918. 4 Giles and Neustadter, J . Chem. Soc., 1952, 3806. 5 Giles, Rose and Vallance, J. Chem. SOC., 1952, 3799. 6 Hodgson, J. SOC. Dyers Col., 1933, 49, 213. 7 Knight, J . SOC. Dyers Col., 1950, 66, 169. 8 Marsden and Urquhart, J. Text. Inst., 1942, 33, T105. 9 Meggy, J. SOC. Dyers Col., 1950, 66, 510. 10 Paine and Rose, quoted by Vickerstaff.*6 11 Pauling, The Nature of the Chemical Bond (Cornell Univ. Press, Ithica, N.Y. 194% 12 Peters, J. SOC. Dyers Col., 1945, 61, 95. 13 Robinson and Ambrosc, Trans. Faraclay SOC., 1952, 48, 854. 14 Steinhardt, Fuggitt and Harris, J. Res. Nat. Bur. Stand., 1941, 26, 293. 15 Steinhardt, Fuggitt and Marris, J. Res. Nat. Bur. Stand., 1943, 30, 123. 16 Vickerstaff, The Physical Chemistry of Dyeing (Olivcr and Boyd, Ltd., Edinburgh, 17 Vickerstaff and Waters, J. Soc. Dyers Col., 1942, 58, 1 1 6. 18 Watson, Modern Theories of Organic Cheniistry (Oxford Univ. Press, 1941), 2nd edn. 2nd edn. 1950).