THEORETICAL ASPECTS OF THE DYEING OF CELLULOSE ACETATE RAYON BY c. L. BIRD, F. MANCHESTER AND MISS P. HARRIS Dept. of Colour Chemistry and Dyeing, The University, Leeds 2 Received 29th June, 1953 The aqueous solubility of a number of disperse dyes has been determined. Some have appreciable solubility at the usual dyeing temperature (SOo C). Even the most insoluble are solubilized to some extent by the dispersing agent normally added to the dyebath, and in these cases the rate of dyeing of cellulose acetate rayon is increased. The evidence now available is considered to support Clavel's view that dyeing takes place from a dilute, saturated aqueous solution. A study of the equilibrium distribution between fibre and dyebath of one of the more soluble disperse dyes has shown that thcre is a linear partition resembling that observed when a solute is distributed between two immiscible solvents.From the standard heat of dyeing it is inferred that two hydrogen bonds link this dye to the fibre. Cellulose acetate rayon, whilst superficially similar to viscose rayon, differs from it both physically and chemically. Viscose rayon is substantially pure cellulose, whereas in cellulose acetate rayon, consisting of " secondary " cellulose acetate, five out of every six hydroxyl groups have been repIaced by acetyl groups, which amount to about 40 % of the total weight. In consequence the fibre has a much reduced affinity for water and swells to a much smaller extent in aqueous solutions. This fact is illustrated by values for the volume term V, i.e. the volume in the water-swollen fibre available for the formation of a dye solution. Thus, Fowler and Michie 1 have assumed a volume term of 0.1 I./kg of dry cellulose acetate rayon, the corresponding figure for viscose rayon 2 being 0-45 l./kg.Similarly, Marsden and Urquhart 3 estimate the average pore size in acetate rayon to be only 5-10& whereas the corresponding figurc for viscose rayon given by Morton 4 is 20-30 A. The small diameter of the pores in acetate rayon limits the size of dye molecules able to penetrate the fibre. Meitner 5 has shown that some acid dycs of large particle radius will not penetrate viscose rayon, even when the fibre is steamed, although under these conditions they will penetrate cuprammonium rayon, which has a volume term 2 of 0.60 l./kg and is known to have pores of larger size than those of viscose rayon.When acetate rayon was first marketed it was found that, apart from the basic dyes (which have poor fastness properties) and a few monosulphonated acid dyes, the water-soluble dyes available had little affinity for the fibre. This difficulty was soon overcome, principally owing to the work of Holland Ellis of British Celanese Ltd., and of Baddiley and Shepherdson of British Dyestuffs Corporation Ltd. (now the Dyestuffs Division of I.C.1. Ltd.), who produced a new range, now known as the disperse dyes, especially for use on acetate rayon. In recent years it has been found that these dyes are also the most suitable for dyeing Nylon, Terylcne, and other synthetic fibres of a relatively hydrophobic character. The disperse dyes consist almost entirely of derivatives of azobenzene and of aminoanthraquinone, two typical examples being 4'-nitro-4-aminoazobenzene and 1 : 4-di(methylamino)anthraquinone.They are supplied as easily dispersible powders produced by milling the crude dye with a suitable anionic surface-active agent followed by drying. Dyeing is usually carried out at about $0" C in presence of 1-2 g/l. of dyebath of a dispersing agent, e.g. a synthetic detergent. 8586 DYEING OF RAYON Holland Ellis realized {hat successful dycings could only be obtained if these dyes were in a very fine state of dispersion, and manufacturers now ball-mill them to give a maximum particle size of about 2 p . The influence of particle size was shown by Vickerstaff and Waters,6 whose results arc given in columns 2 and 4 of table 1.The disperse dye used by these authors, however, has the lowest aqucous solubility of any that we have examined. Morc representative results are given in columns 3 and 5 of table 1, for a disperse dye of average aqueous solubility, but it is not suggested that there is no need for thorough ball-milling. THEORY OF DYEING AND MECHANISM OF THE DYEING mocEss.--It is necessary to account, on the one hand for the affinity of disperse dyes for secondary cellulose acetate, and on the other hand for their ability to transfer from aqueous dispersion to the interior of the fibre. To say that the dyes are " soluble " in the fibre leaves open the qucstion as to what are the forces of attraction between the solid solvent and the solute.Marsden and Urquharl,3 in their work on the swelling of cellulose acetate rayon by phenol, postulated hydrogen bonding between the hydroxyl group of the phenol and one of the acetyl groups of the secondary acetate. Since ncarly all TABLE EFFECT OF VARYING DEGREES OF DISPERSION ON DYE ADSORPTION treatmcnt crystalline dye suspended in Lissapol crystalline dye ground with Lissapol above, milled for 1 h in ball-mill ball-milled for 10 11 ball-milled for 48 h LS LS in a mortar rng dye adsorbed by Ig fibrc at 80OC in : 30 rnin 24 h ____- __ ~ _. ___._ 1 -mcthyl- p-nitro- 1-methyl- p-nitro- arnino-4- aniline a m i n o 4 aniline anilino- -5 anilino- -> anthra- N-ethyl- anthra- N-ethyl-N- quinone N-P-hyd ro x y- qui none &hydro xy- ethyl- cthyl- aniline aniline trace 3.3 0.05-0.1 7.5 1.0 6.6 2.7 7.6 1.2 7.2 3.6 7.7 1.2 7-0 4.3 7.3 2.7 7.0 5.4 7.5 the disperse dyes contain -NH2, -NHR (R = alkyl or aryl) or -OH, similar hydrogen bonding could take place between dye and fibre.There are, however, a few exception, e.g. 3-methoxybenzanthrone, and the obsolete dyes p-nitro-o- anisidine -+ dimethylaniline, p-nitro-o-anisidine --z diethylaniline, and 2 : 4-di- nitroaniline -> diethylaniline. Moreover, azobenzene itself will dye acetate rayon, to the extent of 0.40 % at 22" C. This figure may be compared with the saturation values (at 80" C) of disperse dyes on acetate rayon, many of which lie between 1 % and 2 % of pure dye. It therefore seems probable that azo and methoxy groups impart some affinity foAthe fibre, and for solvents such as ethanol, acetone and alkyl acetates, this affinity being strengthened by the presence of a -NH2, -NHR or -OH group.As far as the mechanism of dyeing is concerned it may bc assumed that the usual pattern is followed, i.e. that a surface film of dye is first produced on the fibre, followed by diffusion of dye into the interior. According to KartaschofT,7 who watched the dyeing of single acetate rayon fibres under the microscope, positively charged particles of suspended dye are attracted to the negatively charged surface of the fibre. Lauer,8 h.owever, found thc particles of dye to be negatively charged, which will ccrtainly be the case in normal dyeings when an anionic dispersing agent is present. Vickerstaff and Waters,G and later Millson and Turl,9 were unable to detect any attraction of dye particles by the fibre.C .L. BIRD, F . MANCHESTER AND MISS P . HARRIS 87 The reason why subsequent workers failed to confirm Kartaschoff's observation may be due to the presence of a dispersing agent in their experiments. Kartaschoff's used Celatine (B.D.C.) dyes consisting of aniinoanthraquinone derivatives in paste form with either sodium ricinoleate, or a hydrocarbon solvent or alcohol. In the last two cases no dispersing agent would be present. The effect of the presence of a little dispersing agent is shown in fig. 1. A fine dispersion of the dye, p-nitroaniline -+ diethylaniline, was obtained by grinding for 2 h in a mortar with water and a very little dispersing agent.The paste was then diluted and allowed to stand for 24 h. The resulting fine suspension was used for the experiments, which consisted of 10-min dyeings at 60" C , with addition of increasing amounts of purified Lissapol LS. In the virtual absence of dispering agent the suspension was unstable, i.e. somc of it was immediately adsorbed on to the fibre or on to any glass surface in contact with it. On adding dispersing agent, which is adsorbed both by the fibre and by the fine particles of dye, the rapid initial adsorption of dye by the fibre was eliminated, as illustrated by the left-hand portion of the curve in fig. 1. FIG. 1.-Effect of dispersing agent on rate of adsorption. In practicc the phenomenon observed by Kartaschoff will not normally occur, because a dispersing agent is almost always added to the dyebath.The problem to be solved consists of two parts, vk. (i) to determine how the dye, originally in fine suspension, reaches the fibre surface, and (ii) how the dye, in single molecules, diffuses into the fibre. In 1923, when the disperse dyes were just beginning to be used, Clavello suggested that dyeing takes place from a saturated solution, which is immediately replenished from the fine suspension as dye is removed from solution by the fibre. Clavel provided no evidence in support of his theory, and until recently the theory of Kartaschoff 7 was generally accepted. Kartaschoff observed that the particles of crystalline dye which were attracted to the surface of the wet fibre appeared to dissolve in the fibre on warming to 60" C.He concluded that dyeing should also take place if the dry rayon and finely powdered dye were left in contact, and he found that dyeing did in fact take place after 15 days at 60" C, or 4 days at 73" C. We have repeated some of Kartaschoff's experiments and find that dyeing will also take place if the rayon is suspended above the dye powder, thus showing, as Johnson11 has suggested, that under these conditions dyeing takes place through the vapour phase. We have not examined this phenomenon88 DYEING OF RAYON in any detail, but, as might be expected, it appears to depend on the volatility of the dye, and the property may be limited to those dyes which are known to sublime. The rate of dyeing, even when the fibre and dye are in contact, appears to be much slower at 60" or 100" C than that observed when dyeing from an aqueous dispersion.This is presumably accounted for by (i) the unswollen state of the dry fibre and (ii) the absence of any appreciable amount of water in the fibre to act as a carrier. Kartaschoff considered that the dyeing of acetate rayon was best explained by the solid solution theory of Witt. The X-ray evidence available at the time seemed to support this view; there appeared to be no fine structure in acctate rayon, so the fibre was described as a " solid colloid ", but it is now known that acetate rayon, like viscose rayon, does contain amorphous and crystalline regions, corresponding to varying degrees of disorder of the long chain molecules. Further circumstantial evidence for the solution theory was provided by the close parallel between the " solubility " of a dye in acetate rayon and its solubility in certain organic solvents; in fact, the solubility of a dye in ethyl acetatc or acetone is a measure of its suitability for colouring acetate rayon.12 The failure to repeat Kartaschoff's experiments, described above, rcvived interest in Clavel's theory, the great stumbling block to the acceptance of which is the apparent complete insolubility of many of the disperse dyes.We have now carried out solubility determinations at 80" C for a considerable number of purified disperse dyes and somc of the results are given in table 2, together with Corresponding values for 0.1 solutions of Lissapol LS (I.C.I.). TABLE 2 solubility at 80" C (mgll.) -~ dye water 01 % Lissapol LS p-nitroaniline --f aniline 9.5 11-5 p-nitroaniline 3 diphenylamine 0.5 4.0 p-nitroaniline + N-ethyl-N-/3-hydroxyethylaniline 7.0 19.0 1 : 4-diaminoanthraquinone 1 : 4-diamino-2-methoxyanthraquinone 1-methylamino-4-anilinoanthraquinone 17.0 20.0 11.0 15.0 < 0.2 3.0 The solubilities were determined by leaving a small quantity (ca.0.01 g) of crystalline dye in contact with 50 ml of either distilled water or a 0.1 % solution of Lissapol LS, using rubber-stoppered bottles partially immersed in a thermostat bath. The flow of water in the bath produced slight agitation of the carrier holding the bottles. After 3 days with water, and 24 h with the Lissapol LS solution, a sample of the liquid phase was removed by means of a pipette through a small plug of cotton-wool and the amount of dye present was estimated colorimetrically by means of a Hilger Spekkcr photoelectric absorptiometer.Since the cotton-wool adsorbed somc dye, the pipette was first filled through the plug and then emptied before taking a samplc for determination of dyc content. The more solubIe dyes can be applied from aqueous solution if a sufficiently Iargc volume of dye liquor is used-although this would not be practicabIe on the large scale -and it is almost certain that in these cases dyeing takes place on the lines suggested by Clavel. Probably more than 50 % of modern disperse dyes fall into this category. A satisfactory theory, however, must also explain the dyeing of acetate rayon by such highly insoluble dyes as l-methylamino-4-anilinoanthraquinone, which is used com- mercially.This dye imparts no visible coloration to water after 3 days at 80" C, but a slight coloration, indicating a solubility of 1-3 mg/l. was obtained at 80" C with a number of dispersing agents used at a concentration of 1 g/l. (cf. table 2). It is significant that this dye is the slowest dyeing of all the disperse dyes examined by Vickerstaff, a fact which would be anticipated if the dye has to pass through the aqueous phase, since the rate at which dye is adsorbed by the fibre is proportional to the concentration of dye in solution. It is also significant that this dye needs to be very finely ground in order to obtain maximum rate of dyeing (cf. table 1). Some rough experiments carried out mainly in order to determine the colour of various disperse dyes on acetatc rayon lend support to the view that dyeing takes place fromC .L. BIRD, F . MANCHESTER A N D MISS P . HARRIS 89 solution. A piece of acetate rayon was placed in a stoppered flask with some distilled water and a little crystalline dye, and the flask and contents kept at room temperature for one month. With the more soluble disperse dyes appreciable coloration of the rayon was observed after a few hours, and equilibrium appeared to have been reached after a few days, the water then being faintly coloured. At the other end of the scale, with l-methylamino-4-anilinoanthraquinone, only a pale blue was obtained after 1 month and the water was colourless. Although the crystalline dye was in actual contact with parts of the rayon, the colourings obtained were quite uniform from first to last and it was difficult to avoid the conclusion that dyeing had taken place via solution in the water.Further evidence is provided by the results of Corbikre,l3 who found that, with two anthraquinonoid disperse dyes, the rate of dyeing is increased by adding certain dispersing agents to the dyebath. We have obtained a similar result with 1 -methylami110-4-anilinoanthraquinone, but with more soluble disperse dyes, e.g. p-nitro-o-anisidine -> N-di(/%hydroxyethyl)aniline, the reverse effect was observed. This retarding effect is presumably similar to the well-known action of certain ethylene oxide condensates, e.g. Dispersol VL (I.C.I.) on leuco vat dyes, i.e. the dispersing agent combines with the dye and then releases it gradu- ally for combination with the fibre.The increased rate of dyeing observed by Corbikre, and illustrated by the right-hand portion of the curve in fig. 1, must be due to the fact that, with the least soluble disperse dyes, addition of dispersing agent results in considerably more dye being present in actual solution (cf. table 2). Consequently, the bottleneck caused by the low aqueous solubility of the dye is removed and the rate of dyeing is increased, in spite of the fwt that the dye must first be released from combination with the dispersing agent. Although there is now a considerable amount of evidence in favour of Clavel's theory, it does not provide a conclusive proof that dyeing takes place only via aqueous solution.In most cases the bulk of the dye is present in the dyebath in the form of a fine suspension, and it is probable that an equilibrium will be established rapidly between (i) the dissolved and suspended dye and (ii) the surface of the fibres. There will be a. little loosely adhering solid dye on the surface of thc fibres as well as in suspcnsion, as has in fact been observed by several workers in this field. If the unadsorbed dye is wholly in solution, as will often be the case at thc end of a dyeing, therc should be no superficial dye on the fibres. Thus it is possible for dye to be adsorbed on the fibre surfaces from solution or deposited from suspension. At the present time there is little indication as to how the dye passes from the surface to the interior of the fibre.The dye must pass between the long chains of the water-swollen fibre through pores, crevices or spaces in the network. These inlets appear to admit individual molecules only, and even these must not be too large. For example, we have failed to observe any diffusion of Lissapol LS (commercial sodium oleyl p-anisidide sulphonate) through a cellulose acetate film, after 24 h at 80" C . If fine particles of' dye ad- sorbed from suspension on to the fibre surfaces can break down into single mole- cules, it is possible that they may then be able to diffuse into the fibre, even if they are virtually insoluble in water. In this connection it is of interest to recall the azoic dyes, which are used on acetate rayon 14 as well as on cellulosic fibres.The aqueous solubility of these dyes appears to be even less than that of the least soluble of the disperse dyes, but when produced in situ within the water-swollen acetate fibre azoic dyes show a mobility similar to that observed in cellulosic fibres. This mobility, however, results in aggregation, not in the dispersion which must take place if individual molecules of dye are to penetrate the fibre. Dispersion via aqueous solution therefore seems a more probable mechanism, even with dyes of very low solubility, in which case the more soluble disperse dyes should show the highest rate of diffusion in the fibre. Meyer, Schuster and Bulow 15 found a constant partition ratio of 182 for the dis- tribution of o-nitroaniline between cellulose acetate and water at room temperature, over the concentration range in water of 0-180 mg/l.THE PARTITION OF DISPERSE DYES BETWEEN CELLULOSE ACETATE AND WATER.-90 DYEING OF RAYON Our work on aqueous solubilities showed that one dye, viz. p-nitro-o-anisidine -+ N-di(P-liydroxyethyl)aniline, had an aqueous solubility of 240 mg/l. at 80" C, which is suficicnt to enable partition experiments to be carried out over a wide range of concentrations. In these expcrimcnts 1 g samples of desized acetate rayon yarn were dyed at 80" C for 4 h with aqueous solutions of the crystalline dye of increasing concentration. Dyeings were carried out in stainless steel cages enclosed in Pyrex test-tubes fitted with a rubber bung. A vertical traverse was imparted to each cage via a stainless steel wire passing through a narrow glass tube in the centre of the bung.This apparatus replaced the usual open tubes in the Marney dycing machine. At the completion of dyeing a sample of the solution was removed for colorimetric analysis. The dyed yarn was rinsed, squeezed, dissolved in acetone and the dye estimated colorimetrically. For the desorption experiments dyed yarn was treated in the same apparatus at 80" C for 1 h with distilled water, followed by estimation of the dye present in solution and on the fibre. The results of thcse experiments arc illustrated in fig. 2, which indicates a linear, reversible partition over the whole range of normal dyeings. FIG. 2.-Adsorption and dcsorption of O ~ N / ~ - N = N - ( _ ~ N ( C , H ~ O M ) ~ . \- 0 adsorption. dcsorption.A similar straight line relationship has been obtained by Wahl, Arriould and SiInon,l6 using the normal dispersion technique with a series of dyes of general formula where R1 and R2 are H, OH, OC2H40H Or OC2H40C2H40H. Moreover, these authors show that a point is reached when the fibrc becomes saturated; the CUrvc then becomes a straight line almost parallel to the abscissae. The same type of curve has becn obtained by Daruwalla and Turner 17 for l-amino-4- hydroxyanthraquinone (Duranol Red 2B) ; in this case distribution was effected by steaming a cellulose acetate film in contact with a starch film containing the dye. Daruwalla and Turner consider that their results favour the solid solution theory. The solid solution concept has the mcrit of indicating the shape of the equi- librium isotherm, but on the other hand it implies that the dyeing of acetate rayonC.L . BIRD, F . MANCHESTER AND MISS P . HARRIS 91 is fundamentally different from other forms of dyeing. It secm more likely that the dyeing of acetate rayon with disperse dyes takes place in a manner similar to that observed with direct dyes and viscose rayon, i.e. dye niolccules are first adsorbed from solution on to the sites available at the surrace of the fibre and then diffuse (in solution) through the water-filled network of long-chain fibre molecules into the interior of the fibre, finally becoming fairly evenly distributed on sites located on micellar surfaces. With disperse dyes, however, the affinity for the aqueous phase, i.e. aqueous solubility, must be kept very low. Conse- quently, at the commencement of dyeing the bulk of the dye is normally present in the form of a suspension, but it cannot diffuse into the fibre until it has passed into solution.Solid solution or vapour phase dyeing probably plays a minor part in the dyeing of acetate rayon with disperse dyes, but it may be much more important with hydrophobic fibres such as Terylene and Qrlon. illustrated in fig. 2 may be uscd to determine the standard affinity (- Ap") of the dye, using the equation 18 (1) where Of is the concentration of dye in the fibre and Ds the concentration in the aqueous solution, expressed in moles/kg of completely dry fibre and moles/l., respectively, whilst Yis the volume term, the value used, viz. 0-1 I./kg, being that used by Fowler and Michie.1 The standard aanity at 80" C for the dye, p-nitro- o-anisidine --f N-di(/%hydroxyethyl)aniline is 5.87 kcal/mole, which is of similar magnitude to the corresponding figures for acid dyes on wool and direct dyes on cellulose.The value of Df/Ds, which is found to be constant, in accordance with eqn. (l), is 425. In order to determine the standard heat AH" of dyeing, affinity values were determined over a range of temperatures, using the Gilbert 19 desorption apparatus, as modified by Lemin.20 The results are given in table 3. TABLE 3 temp. ("K) 316 333 353 363 THE STANDARD AFFINITY AND HEAT OF DYEING OF A DISPERSE DYE.-The results - Ap0 = RT In (Df/VDs), - Apo (kcal) 6-35 6-1 6 5.87 5.76 Since AH" can be obtained by plotting Ay"/Tagainst 1/T, the slope giving AH", provided that a straight line relationship is obtained, as is here the case.The results in table 3 give a value for AH" of - 10.5 kcal/mole, a figure which is of similar magnitude to that obtained by Marshall and Peters 2 for two direct dycs on cellulose, viz. Durazol Red 2B and Chlorazol Brown M, which, out of a total of eight dyes, had the lowest heats of dyeing. If 5 kcal is taken as the value for a hydrogen bond, the figure of 10.5 kcal suggests that two hydrogen bonds are formed between secondary cellulose acetate and the dye, p-nitro-o-anisidine -+ N-di(P-hy droxyethy1)aniline. We desire to express our gratitude to Courtaulds' Scientific and Educational Trust Fund for a Scholarship which cnabled one of us (P. H.) to take part in this work. 1 Fowler and Michie in Vickerstaff's The Physical Chemistry of Dyeing, p. 276. 2 Marshall and Peters, J. Soc. Dyers Col., 1947, 63, 446. 3 Marsden and Urqiihart, J . Text. Imt., 1942, 33, T105. 4 Morton, Trans. Farachy SOC., 1935, 31, 281. 5 Meitner, J. Soc. Dyers Col., 1945, 61, 33. 6 Vickerstaff and Waters, J. SOC. Dyers Col., 1942, 58, 116.92 UNIMOLECULAR FILM BALANCE 7 Kartaschoff, Helv. chim. Acta, 1925, 8, 928. 8 Lauer, Kolloid-Z., 1932, 61, 91. 9 Millson and Turl, Text. Res. J., 1951, 21, 685. 10 Clavel and Stanisz, Rev. Gin. Mat. Col., 1923, 28, 145, 167 ; 1924, 28, 94, 158: 222. f * Johnson in Vickerstaff’s The Physical Chemistry of Dyeing, p. 260. 12 Mellor and Olpin, J. Soc. Dyers Col., 1951, 67, 622. 13 Corbihe, Teintex, 1948, 13, 433. 14 Mellor and Olpin, J. Soc. Dyers Col., 1947, 63, 396. 15 Meyer, Schuster and Biilow, Melliand Textilber., 1925, 6, 737. 16 Wahl, Amould and Simon, Teintex, 1952, 17, 288. 17 Daruwalla and Turner, J. Soc. Dyers Col., 1953, 69, 240. 18 Vickerstaff, The Physical Chemistry of Dyeing (Oliver and Boyd, 1950), p. 104. 19Gilbert’ Proc. Roy. Soc. A, 1944, 183, 167. 20 Lcmin in Vickcrstaff’s The Physical Chemirtry of Dyeing, p. 92.