132 ADSORPTION OF DYES BY CRYSTALS THE ADSORPTION OF DYES BY CRYSTALS BY J. WHETSTONE Imperial Chemical Industries, Ltd., Research Dept., Nobel Division, Stevenston, Ayrshire Received 18th June, 1953 Explanations of the crystal habit modifying powers of certain dyes have usually depended on the adsorption, by means specified or unspecified, of the dye molecules by the growing crystal. The present work has confirmed the importance of dye adsorption in crystal habit modification phenomena. Studies of modified crystals in which dye adsorptions, followed by overgrowth of the dye molecules, have occurred to give pleo- chroic dye inclusions have indicated, however, that adsorption is not necessarily on the habit modified plane. More frequently it is perpendicular to the modified face of the crystal, which is very reasonable if modern views on the layerwise growth processes of crystal faces are accepted.It was assumed that the polar groups of dye molecules were responsible for the ad- sorption of dyes by growing crystals, and a considerable weight of evidence has been ob- tained to support the view that adsorption is due to a close similarity in pattern between the polar groups in a dye molecule and the ions of a crystal plane. The type of crystalJ. WHETSTONE 133 plane involved depends on the nature of the dye molecule ; for instance planes containing anions and cations will be most favourabIe to the adsorption of an acid triphenyl methane dye with amino and sulphonate substituents in approximately equal numbers, whereas another plane containing anions only may be more favourable to the adsorption of an azo dye with a predominance of sulphonate substituents.The most obvious evidence of dye adsorption having occurred in a crystal grown from a saturated aqueous solution containing a suitable dye is sometimes the presence of coloured areas symmetrically disposed about the central region of the crystal. Often, however, by heating the crystal on the stage of a microscope, it can be seen that such coloured regions are not due to adsorption phenomena at all, but are due to inclusions of mother liquor, which etch the edges of their containing cavities as the temperature rises. However, in many cases it will be found that the coloured inclusions are genuinely due to the adsorption of dye in the growing crystal lattice.Usually, but not invariably, this dye adsorption is accompanied by alteration of the normal crystal habit of the substance grown from water. Many instances have been observed, however, in which crystallization from a dye-containing solution has yielded much-modified crystals without significant traces of dye adsorption having occurred, and it seems reasonable to suppose that the habit modification has been due to the adsorption of dye molecules which have been displaced by the subsequent growth processes of the crystal. When highly coloured dye inclusions are present, it might reasonably be suggested that the adsorbed dye molecules have been overgrown. Easy overgrowth of adsorbed dye molecules might be associated with a lessened obstructive effect by the dye molecules on the growth of crystal planes, so that the finding of habit modified crystals might be considered to be, on the whole, a more typical manifestation of the occurrence of dye adsorption than the production of highly coloured crystals.ADSORPTION MECHANISM.-AtfemptS to explain crystal habit modifications have always hinged on elucidating the interaction of the modifying agent and the planes of the crystallographic “form”, the growth of which is most affected. Solid solution formation of the modifier with the substrate, or compound forma- tion, or some type of physical adsorption have all been suggested. Thus Buckley suggested that adsorption was due to similarities between the oxygen triangles of the sulphonate groups of the modifying dye molecules and the oxygen triangles of the oxyacid anions such as sulphate, etc.1 This view could not be maintained when Frondel demonstrated that sulphonated dyes were capable of modifying the crystal habit of halides.2 Bunn thought that the adsorptions were due to the formation of surface solid solutions due to similarities between the primitive translations of the modifier and substrate,3 but this view could not be applied to dyes as their crystal structures were unknown, and probably in many cases non-existent owing to their colloidal properties.At the time when the author became interested in the problem of dye adsorption in growing crystals new ideas as to the nature of crystal growth phenomena were coming to the fore. Bunn and Emmett had demonstrated photographically that crystal growth in many substances was associated with the spreading from continuously renewed nuclei of “ growth layers ” over the crystal faces.4 Frank had postulated that sometimes these continuously renewed nuclei might consist of spiral dislocations of the crystal lattice,5 a view which was soon to receive practical support by the direct observation of surface irregularities on some crystal faces, sometimes taking the form of slight prominences showing evidence of spiral growth.The writer has described elsewhere the evidence which led him to suggest adsorption of foreign matter might occur on the edge-faces of the growing layers tending to build up a growing face 6 so that in fact the plane in which the adsorption has occurred may134 ADSORPTION OF DYES BY CRYSTALS be perpendicular to the plane the rate of growth of which has obviously been modified. Buckley had noted that the plane of dye adsorption sometimes appeared to be quite unrelated to the plane of habit modification.7 PLEOcmOIsM.-The simplest cases for investigation of the nature of the ad- sorption phenomena between dyes and growing crystals were those in which regular deposition of dye in broadening tracks as the crystal grew was indicated by coloured inclusions as “ hour-glass ” or ‘‘ Maltese Cross ” shapes.Photo- graphs of such crystals are included in many of Buckley’s publications.1 In certain specific examples of habit modifications these inclusions varied strongly in colour when the crystal was rotated in plane-polarized light on the microscope stage.Buckley had examined the dichroism of many such examples without coming to any definite conclusion as to the significance.7 Sometimes the colour variation is from full colour to virtually colourless, but often the effect is only the lightening and darkening of the colour, probably with some concurrent vari- ation of shade. The effect is evidently due to the change of position of the adsorbed dye molecules relative to the electric vector as the crystal is rotated, i.e. ultimately to the differences in polarizability of the dye molecules in different directions, and it was thought that observation of the colour changes should prove helpful in interpreting the adsorption behaviour of dyes on crystah. THE STUDY OF PLEOCHROIC MODIFIED CRYSTALS.- Considerable progress towards the understanding of the nature of the influences affecting the possibility of dye adsorption in crystals was achieved by the study of crystals modified with acid tri- phen ylmethane dyes, not ably Acid Magenta modified ammo- nium nitrate crystals. The tr,c normal habit of ammonium i n w m .nitrate IV consists of long prisms, elongated on (001). When the salt was very care- fully crystallized from a solu- tion containing 0.01 % Acid Magenta, large platy (010) crystals were obtained, in which the path of deposi- tion of the dye as the crystal grew was revealed by an ever-broadening track starting from the centre of the crystal, and being symmetrical about the a axis (fig. 1) forming an “hour glass ”type of inclusion. These inclusions were pleochroic, varying from colourless to pale magenta in colour as the crystals lying flat on the microscope stage were rotated in plane-polarized light.Now, the Acid Magenta dye molecule is of a very symmetrical type, and the very similar dye “ trisulphonated pararosaniline ” is completely symmetrical. It was con- firmed that exactly the same crystal habit modification of ammonium nitrate IV was given by the latter dye, and the pleochroism of the modified crystals was again similar, the maximum coloration being developed when the crystals were so positioned that the c crystallographic axis and the electric vector were parallel. The pleochroic properties of the modified crystals necessarily imply that the dye molecules in the crystal must be arranged all parallel with one another, so that the differential interaction with the polarized light ranges from a maximum to zero.In view of the symmetrical nature of the molecules, it further is necessary that the minimum of absorption must be when the electric vector is perpendicular to the plane of the molecules, and the maximum when the electric vector lies in their plane. Reference to the experimental observations of the pleochroism of the modified crystals (fig. 1) showed quite clearly that the dye molecules in the modified crystal must be standing perpendicular to the modified plane and per- pendicular to the axis of elongation of the modified crystals, i.e. they are adsorbed on or lying in (100). c PXIS FIG. 1.J . WHETSTONE 135 Trisulphonated pararosaniline Acid magenta N.D.(C.I. 676). prepared from new magenta (C.T. 678) STRUCTURAL COMPARISONS OF DYE AND CRYSTAL.-BY comparison of the ionic structure of the {loo} plane of ammonium nitrate IV and the Acid Magenta molecule drawn to scale it was apparent that a very strong similarity existed between the pattern of the polar groups of the dye molecule and the ions of the crystal plane, it being possible to superimpose the dye model on the map of the plane so that the SO3- groups coincided with NO3- ions and the NH2+ groups with NH4f ions (fig. 2). On the basis of this evidence and other similar observations it was suggested that the dye molecules were adsorbed into a growing crystal plane because of the Coulomb attractions of ions in situ for the polar groups of the dye molecules in solution, and that the strength of adsorption depended on the number of polar groups involved and their accuracy of fitting on to the crystal structure.FIG. 2.-Ammonium nitrate IV (100) and pararosaniline trisulphonate. Consecutive parallel layers of ions in a crystal structure alternate in their rela- tive dispositions of anions and cations. Where the van der Waals radii of the polar groups approximate to those of the ions in the crystal plane, which is usual with simple oxyacid salts, the dye molecules may be supposed to be held on the crystal plane by the mutual attraction of ions in the surface layer and polar groups of unlike charge (e.g. the NO3- ion for the NH2+ group and vice versa), followed by the growth of the next layer of ions, in which the polar groups will replace ions of a similar charge.It has been observed that the presence of adsorbed dye molecules does not make any perceptible difference to the crystal structure as indicated by X-ray powder photographs, which is consistent with the dye molecules being truly in136 ADSORPTION OF DYES BY CRYSTALS " solid solution ". The consistency of the deductions and predictions as to the possibility of dye adsorption according to the above scheme, derived in the course of the work, is strong support for the above statement. This explanation of dye adsorption is not unlike the theories to account for the formation of orientated overgrowths on crystals advanced by van der Merwe,g but since the repeating unit would be many times larger than the crystal unit cell, if the large dye molecules could form into a characteristic crystal structure, the cases cannot be regarded as strictly analogous.The formation of micellar ag- gregates would increase the disparity between the sizes of the structural units. In this connection it has been shown in a separate investigation that there are good grounds for supposing that in fact the single dye molecules aremore important than micelles in dye adsorption phenomena with crystals. The effect of dye adsorption on crystal growth may be related to the difficulty of completing the growth of layers of ions owing to the necessity of displacing adsorbed foreign matter, or to the lessening of the Coulomb forces between ions in situ and ions in solution owing to the interposition of the hydrocarbon matter of the dye molecules, probably with the dielectric effect also contributing, adsorption is dependent on sufficient similarity between the patterns of the polar groups of a dye molecule and the corresponding ions of a crystal plane (assuming that the dye is adequately soluble to allow of a sufficient concentration in the saline solution) and due to the Coulomb forces, normally associated with crystal growth, drawing in the polar groups of the dye molecules into ionic sites, clearly certain conditions must be satisfied for the adsorption to be really strong.The number of polar groups attached to a single dye molecule (or micelle possibly) must be a factor in determining the strength of the adsorption. The polar groups should fit snugly without overlap into the ionic sites, so that adjacent ions are not excluded and the Coulomb forces attracting the polar group are at a maximum.This latter factor is of importance. For instance, the differing effects of triphenyl- methane dyes such as Acid Magenta on the habits of sodium, potassium and ammonium nitrates seem to be connected with the differing atomic radii of the cations; it is noteworthy that alkylation of the amino groups (thus making them larger) notably decreases the habit modifying powers of the dyes. Further, the nature of the anion will exert a considerable influence on the nature of the crystal planes on which dye adsorption by sulphonated dyes is possible, For instance, an XO3- anion like nitrate or chlorate is usually of dimensions very similar to the SO3- group, and it is unlikely that the dye adsorption will be on a plane which involves that the planes of the oxygen triangles of the anion and sulphonate group will be mutually orthogonal; it is more probable that the crystal plane involved in the dye adsorption will be one perpendicular (or nearly so) to the planes of the oxygen triangles of the anions so that the oxygen triangles of the sulphonate groups correspond as nearly as possible.This viewpoint can be tested in respect of the habit modification of ammonium nitrate IV with Amaranth. The modified crystals are pleochroic, the maximum absorption occurring when the electric vector is parallel with the c axis, indicating that dye adsorption is on (100) if the molecules are adsorbed perpendicularly to the modified plane, which contains the nitrate ions.The model of the dye structure can in fact be fitted to (100) drawn to scale, with the sulphonate groups falling into anion positions. How- ever, the dye model can equally well be fitted to (OlO), the modified plane, but in this case the oxygen triangles of anion and sulphonate group are orthogonal. Even so, it may be that the observed pleochroism is due to the differential inter- action of the electric vector with the length and the breadth of the dye molecule adsorbed flat on the modified plane in this way. That this latter view cannot be maintained, however, is shown by the direction of the supposed adsorption of the dye molecules, maximum absorption would be when the electric vector is at about 45" to the c axis, not parallel with the c axis as observed.GENERAL CONSIDERATIONS AFFECTING DYE ADSORPTION IN CRYSTALS.-If dyeJ . WHETSTONE 137 When the anion is of the X04 type there are four sets of oxygen triangles and the directional effect on influencing adsorption observed with the XO3 anions is lost. Since dye adsorption is commonly perpendicular or nearly perpendicular to the plane of habit modification, this implies that many more opportunities for adsorption and a multiplicity of types of habit modification should be possible. This has in fact been observed by Buckley for potassium and ammonium per- chlorates, and to some extent for the sulphates, but here the pseudohexagonal symmetry is effective in reducing the number of crystal forms on which adsorption may take place.An interesting situation arises when dye adsorption is possible on more than one plane of a crystal lattice. In general there will be differences in the tendencies for the dye to be adsorbed on the differing planes, but owing to the reduction of the growth rate of the plane on which the dye is most strongly adsorbed at first, the opportunities for adsorption on a second more rapidly growing plane will become relatively greater, and it is often found that by increasing continuously the proportion of dye in a given solution the habit modification may be induced to change from one plane to another. Many examples of this are quoted in Buckley's work on potassium and ammonium perchlorates,lo and in an as yet unpublished study of the author's it is shown that it is possible to pre- dict with a fair degree of accuracy the habit modification changes from consider- ation of the possibilities of dye adsorption on alternative crystal planes of ammonium perchlorate. The effect of the oxygen triangles of X03 ions in conjunction with the sulphon- ate groups of the dye molecules is also lost in the alkali metal halides, which in many instances have been found to be susceptible to habit modification, although owing to their cubic symmetry the habit modifications have generally been in the direction (100) + (111).The adsorption of a dye by means of its polar groups into a single crystal plane necessitates that the dye molecule shall lie flat in the plane, i.e. the various aromatic nuclei carrying the important polar groups shall be coplanar, or very nearly so.The importance of this requirement may be very clearly demonstrated by quoting the observed results with mono- and dis-azo dyes as habit modifiers. The mono-azo dyes based on a-naphthylamine -+ /%naphthol or P-naphthylamine lose their modifying powers if a sulphonate substituent group is introduced into an 8-position, thereby destroying any possibility of coplanarity between the naph- thalene ring systems. NaOS-( ) __ S03Na \ Ponceau 6R S03Na \ Amaranth138 ADSORPTION OF DYES BY CRYSTALS Thus, while Amaranth (C.I. no. 184) is frequently a potent crystal habit modifier, the further 8-sulphonated dye Ponceau 6R (C.I. no. 186) is ineffective. bis-Azo TABLE 1 .-PREDICTIONS OF DYE ADSORPTIONS AND RESULTANT CRYSTAL HABIT MODIFICATIONS fitting of polar groups on ion sites plane anions cations of dye modification solu- bility observed dye substance Acid Magenta rnm‘m’’- (NH4)$04 (100) 2/3 good 2/3 good f.sol. *1/3 poor 1/3 sat. trisulphonated pp’p”- triamino triphenyl carbinol anhydride Violet mm’-disulphon- ated pp‘-diamino tri- phenyl carbinol an- hydride anthraquinone 2 sodium sulphonate sulphonated triphenyl pararosaniline disulplzonated Dobner’s (NH&S04 (100) 2/2 good 2/2 good sol. 1 : 4 : 5 : 8-tetramino- NaN03 (10i2) - 4/4 good sol. Ink blue pp’p”-tri- NaN03 (1120) 3/3 good - sol. or-naphthylamine 3 : 6- NH4N031V (100) 3/3 good - f. sol. disulphonate -> 8- NaN03 (1012) 3/3 good - sol. naphthol-4-sulphonate (NH&S04 (100) 3/3 good - sol. Solochrome Yellow YS (NH&S04 (100) 313 good 8-naphthylamine 6 : 8- disulphonate --f sali- cylic acid /3-naphthylamine 5 : 7- disulphonate -+ a- naphthylamine 7- NaN03 (1012) s u 1 p h o n a t e -+ p- n a p h t h o l 3 : 6-di- sulphonate /3-naphthylamine 5 : 7- d i s u l p h o n a t e - t or-naphthylamine -+ NaN03 (1012) /3:naphthol 3:6- disulphonate 3-sulphonate 4 2 mol.415 good i 11/5 sat. 314 good { 1/4 sat. Trypan Red benzidine KNO3 (010) 5/5 good j3-naphthylamine 3 : 6-disulphonate disulphonate --> a- naphthylamine 7- s u l p h o n a t e + j3- n a p h t h o l 3 : 6-di- suiphonate a-naphthylamine 3 : 6- NaN03 (1012) 515 good 1/1 sat. sol. V.S.S. (cold) V.S.S. (cold) - sol. - insol. (hot, cold) (010) tablets, plates (010) plates strong (0001) modification strong (0001 1 modification 010 (laths) f.strong (0001 modification) strong (001) modification strong (001) modification f. strong (0001) mod. (hot) no mod. (cold) f. strong (0001) mod. (hot) no mod. (cold) f. strong (001) modification no modificationJ . WHETSTONE 139 dyes containing 8-sulphonates may, however, be habit modifiers, apparently because the dye molecules can easily take up such a configuration that two aromatic ring systems are coplanar, even when the third is forced out of alignment. appears to be some kind of correlation between the type of dye molecule and the kind of crystal plane on which adsorption is likely to occur. If models of the space lattices of a few crystals are examined it can be seen that the prominent planes can be classified either as " close packed " or " stepped " types.In the former all the constituent ions (or atoms or molecules) are co-planar, but not so in the stepped planes ; the individual particles all differ somewhat in their parameters perpendicular to the plane, so that the plane might be supposed to pass through the mean of the positions of its constituent ions. These planes may consist entirely of the same types of ions or of both anions and cations. The ionic reticular density in the plane is obviously likely to be highest in the close packed planes and these are regarded as most suitable for dye adsorption; close packed planes consisting entirely of anions would appear to be most suitable for the adsorption of azo dyes, etc., with a large preponderance of sulphonate over cationic groups.Similarly close packed planes consisting entirely of cations would appear to favour the adsorption of dyes with a preponderance of amino groups over sulphonates, e.g. 1 : 4 : 5 : 8-tetraminoanthraquinone 2-sulphonate. Close packed planes with both anions and cations would appear to be very suitable for adsorption of dyes with approximately equal proportions of anionic and cationic groups. Stepped planes should be suitable for adsorption in a similar manner to the above, but owing to their much sparser packing at any one given level, they might be less frequently involved in adsorption phenomena. The available evidence appears to indicate that the above considerations are in fact important in influencing dye adsorption phenomena and it is possible successfully to predict the adsorption of dyes and the resultant crystal habit modifications on the basis of comparisons of the ionic patterns of the dye molecules and crystal planes selected according to the above guide.A number of examples of how predictions of habit modifying powers in dyes for various salts have been made by testing the possibility of the occurrence of dye adsorption on suitable planes of the crystal lattice are given in table 1. It will be seen that where habit modifications have been obtained the possibility of demonstrating a high proportion of coincidences between polar groups and appropriate ionic sites exists. The dye must also have an adequate solubility in the saturated saline solution-if the dye is not sufficiently soluble the expected habit modification cannot possibly be realized. In some cases where no habit modifications were obtainable in cold solutions, but the temperature coefficient of solubility of the dye was suitable, its solubility could be increased by crystallizing at higher temperatures, so that the expected habit modifications could be obtained. PREDICTION OF POSSIBILITY OF DYE ADSORPTION ON CRYSTAL PLANES.-There The coincidences are indicated by terms : (a) good-within 0.5 A, (c) poor-just outside 1 A, (b) satisfactory-within 1 A, and (d) no coincidence, and the number of coincidences of those possible is indicated by a fraction. In order to avoid the unpredictable effects due to shading additions and the use of unpurified dye intermediates such as might be encountered with com- mercial dye samples, the foregoing investigation was carried out using specially laboratory-prepared dyes made from purified dye intermediates. The author wishes to express his indebtedness to I.C.I. Ltd., Dyestuffs Division, Research Department, and especially to Dr. E. K. Pierpoint, for the supply of these pure dye samples.140 CALORIMETRIC STUDIES 1 Buckley, Crystal GrowtA (Wiley, New York), p. 370. 2 Frondel, Amer. Miner., 1940, 25, 91. 3 Bunn, Proc. Roy. SOC. A , 1933, 141, 567. 4 Bunn and Emmett, Faraday SOC. Discussion, 1949, 5, 119. 5 Frank, Faraday SOC. Discussion, 1949, 5, 48. 6 Whetstone, Nature, 1951, 168, 663. 7 Buckley, Faraday SOC. Discussion, 1949, 5, 243. 8 Buckley, Mem. Proc. Manchester, Lit. Phil. SOC., 1938-9, 83, 3 1. 9 van der Merwe, Faraday SOC. Discussion, 1949, 5, 201. 10 Buckley, Mem. Proc. Manchester, Lit. Phil. SOC., 1950-1, 92, 1.