J O U R N A L O F C H E M I S T R Y Materials Reversibly thermochromic systems based on pH-sensitive spirolactone-derived functional dyes Stephen M. Burkinshaw, John GriYths and Andrew D. Towns* Department of Colour Chemistry, University of Leeds, Leeds, UK LS2 9JT. E-mail: ccdadt@leeds.ac.uk Received 30th July 1998, Accepted 9th October 1998 Composites formulated from pH-sensitive colour formers mixed with fatty acid co-solvents and acidic developers have been prepared and their thermochromic properties investigated.Possible explanations for the thermochromic eVect have been considered and evidence is presented in support of a mechanism based on phase changes occurring within the compositions during heating and cooling. Introduction Colour formers that are pH-sensitive have been utilised commercially in the production of thermographic recording materials for around thirty years.The main outlet for such technology is facsimile paper in which the colour formers are designed to change irreversibly from colourless to coloured states on the application of heat.1,2 However, thermochromic systems are known which utilise colour formers in order to produce thermally-triggered reversible switching between coloured and colourless states.3 In addition to a pH-sensitive colorant, these systems also contain a readily fusible solid cosolvent and a colour developer.The co-solvent is a relatively low-melting hydrophobic compound that acts as a medium in which the colour former and developer can interact; the material typically has a long chain aliphatic character and may be a fatty acid, amide or alcohol.Provided that the mixture is formulated correctly, a striking colour change from and phenolic developers. From the observed influence of coloured to colourless occurs upon heating the composition molecular structure of the developer on thermochromism, and above its melting point, the original colour returning when also the significance of fractional composition on the eYcacy the material solidifies through cooling.Reversibly thermo- of thermochromism, a mechanism for the process has been chromic systems based on acidic developers and spirolactone proposed. colour formers are typical of the formulations encountered in the literature: an example of the latter component is Crystal Results and discussion Violet lactone (1), which in its lactone form, 1a, is colourless, but on ring-opening (Scheme 1) converts to the intense blue Synthesis and properties of the colour formers species 1b.The ring-opening may be induced by the addition The colour formers employed in this investigation are typical of a proton or through an increase in the polarity or hydrogenof the classes of colorant used industrially: diarylphthalide bonding ability of the host environment.The reaction is fully (1), fluoran (2), vinylphthalide (3) and fluorene (4). Crystal reversible. Violet lactone, 1, is readily available commercially. The fluoran 2 was prepared by condensation of the benzoylbenzoic acid 5 with 4-acetylaminophenol (6a) followed by hydrolysis to give fluoran 7, the primary amino group of which was then dibenzylated (Scheme 2).6 Scheme 1 Despite the many patent applications concerning thermochromic materials based on co-solvent/developer/colour former compositions, little has been published that addresses the mechanism of the thermochromic eVect.3–5 This paper describes the behaviour of thermochromic compositions formulated with fatty acids, spirolactone colour formers (1–4) © British Crown Copyright 1998/MOD.Published with the permission Scheme 2 of the Controller of Her Britannic Majesty’s Stationery OYce. J. Mater. Chem., 1998, 8, 2677–2683 2677The phthalide 3 was synthesised by condensing Michler’s compositions, some coloured solid remained, which dissolved at higher temperatures to complete the colour loss.Reducing ethylene (8) with the tetrachloro-substituted benzoylbenzoic acid 9 (Scheme 3).7 the proportions of co-solvent to developer and colour former from 505251 to 505151 and 255151 made little diVerence to the initial colour intensity, but raised the temperature at which complete colour loss occurred (Table 2). However, in all cases, the onset temperature coincided with the start of melting of the composition, which was little diVerent from that of stearic acid itself (mp 67–69 °C).Compositions were prepared from the colour formers and Bisphenol A using fatty acids other than stearic acid. Onset of colour loss was again observed to be related to the melting Scheme 3 point of the co-solvent. For example, 505251 (fatty acid5Bisphenol A51) compositions prepared using lauric acid, While 4 has been obtained from an aminospirolactone by myristic acid and palmitic acid as the co-solvent component diazotisation and intramolecular coupling8 or treatment with exhibited colour loss onset temperatures of 45 °C, 52 °C and sodium sulfite in concentrated sulfuric acid followed by 62 °C, respectively, which correspond well with the melting addition of copper powder,9 a method involving an intramol- points of the co-solvents (42–46 °C, 51–53 °C and 61–63 °C ecular Friedel–Crafts reaction10 was chosen owing to the respectively).12 For a given co-solvent, varying the mixture availability of the starting material, 1 (Scheme 4).ratio of the components had little eVect on the point at which the composition started to melt or change colour and the switching temperature was determined solely by the temperature at which each formulation began to melt.The colour change was found in all cases to be fully reversible: no colour loss was noted on repeated heating and cooling of the samples. For example, when the composition (505251) based on 2 was subjected to thirty cycles of heating to 90–100 °C and allowing to cool below 40 °C, no significant Scheme 4 diVerences in reflectance minima were noted.Each colour former showed strong infrared absorption in the 1750–1760 cm-1 range characteristic of phthalides.11 In its Mechanism of the thermochromic eVect solid form, 3 displayed photochromism in that the initially green crystals turned darker and more yellow in sunlight, While the mechanism of irreversible thermochromism of spiroreturning to their original colour in the dark.lactone-based thermographic materials that change from a The lactones, when examined by thin layer chromatography, colourless to a permanently coloured state on heating is well rapidly became coloured on contact with alumina or silica, as understood,13 little is known about the mode of operation of was expected with such polar media: 2, 3 and 4 gave dark the reversible coloured-to-colourless compositions described green, turquoise and pale green spots, respectively.The colour above, despite the volume of patent material which has of the spots paralleled the absorption maxima of the derivatives appeared and the widespread use of such compositions (in in acetic acid solution (Table 1). microencapsulated form) in textile and novelty goods.Theories involving steric considerations3 or phase separa- Preparation and properties of the compositions tion3–5 have been put forward in an attempt to explain why these compositions have only slight colour, if any, when Thermochromic compositions were prepared from each colour molten, and yet are intensely coloured in the solid state. One former 1–4 in the following manner: stearic acid, Bisphenol A theory proposes that steric factors determine the generation and colour former in the ratio 505251 were mixed and heated and loss of colour, whereas another explains the phenomenon above the melting point of the fatty acid to produce weakly in terms of temperature-driven phase changes within the coloured or colourless solutions, which were then rapidly composition.However, direct experimental evidence support- cooled by the addition of cold water causing simultaneous ing either mechanism is surprisingly lacking. It is useful first solidification and coloration of the wax. The reflectance to summarise the two mechanisms and then describe our own minima of the dried, powdered compositions accorded well observations which strongly support the second mechanism.with the absorption maxima of the corresponding colour former in acetic acid. Sterically-induced mechanism of thermochromism. In this Each composition exhibited reversible thermochromism. At premise, it is argued that the existence of the coloured species the onset of melting of the composition, colour loss started to (e.g. 1b, Scheme 1) is more favourable sterically in the solid occur and within a few degrees above the melting point, the matrix of the composition compared to that for the colourless majority of the colour had been lost. Depending on the form (e.g. 1a). Whereas the more restricted environment of fractional composition and colour former, the diVerence in the solid composition favours the coloured, relatively planar temperature between the onset of colour loss and complete ring-opened structure over the tetrahedral colourless spirolac- loss of colour varied from only a few degrees to over 40 °C tone structure, in the molten composition, the smaller steric (Table 2).At 75 °C, the melts of the 505251 (stearic requirement of the coloured species is less important and acid5Bisphenol A5dye) samples based on 2 and 3 were almost conversion to the lactone (and thus colour loss) can take place colourless, although it was observed that after melting of the more readily.The argument is based on the supposition that steric factors regulate the position of the ring-opening equilib- Table 1 Absorption maxima of the colour formers in acetic acid (99%) rium: cooling and solidification of the composition, which Colour former lmax/nm when molten contains the colour former predominantly in the colourless form, moves the equilibrium towards the coloured 2 609, 464, 437 form since the planar geometry of the latter structure is more 3 706 favourable in the solid. 4 915, 836, 621 However, X-ray diVraction data and molecular modelling 2678 J.Mater. Chem., 1998, 8, 2677–2683Table 2 Temperatures of melting, colour loss onset and completion for compositions of stearic acid/Bisphenol A/colour former Colour Mixture Onset of colour End point of former ratio Melting point/°C loss/°C colour loss/°C 2 505151 67–68.5 67 72 2 505251 67–69 67 ca. 83 2 255151 67–68.5 67 90 3 505251 67–69 67 ca. 110 cast doubt on this theory. While X-ray crystallography has confirmed the non-planar nature of the spirolactone colour formers, like fluorans,14 the ring-opened coloured species have also been found to be non-planar. Thus: (1) the aromatic rings of Crystal Violet Lactone cation (1b) complexed with metal iodides have a propeller-like orientation;15 (2) the cation from benzofluoran 10 complexed with metal iodides has a near-planar xanthene component, although the carboxy-substituted ring is almost perpendicular to it;16 (3) Scheme 5 suYciently high degree for satisfactory colour development.Conversely, too little co-solvent will prevent complete dissolution on melting and inhibit colour loss. The consequence of a the ring-opened hydrochloride salt of the anilinofluoran 11 system lacking any co-solvent, or the use of one that does not was found to have a structure in which the pendant phenyl melt in the temperature range of interest, would be that the ring is at right angles to, and the anilino substituent twisted colour change would be more gradual and less complete.There out of plane of, the xanthene structure.17 will be no sudden removal of material into solution and Molecular modelling reproduced the non-planar features of consequent rapid colour loss.the structures in the first two examples listed above. Other The eVectiveness of a developer can therefore be anticipated ring-opened species were predicted to be significantly non- to depend not only on its acidity, but also on its solubility in planar, for example, the structure of 3, in which none of the the co-solvent.In order to verify this suggestion and to test aromatic rings were co-planar because of the crowded nature the validity of the phase separation theory, a series of Nof the molecule. From the X-ray studies and molecular modelling, it is clear that the coloured species are not planar and do not possess significantly smaller steric requirements over the corresponding colourless lactone forms.Thus a mechanism based on sterically- driven thermochromism seems unlikely. acylaminophenols 6 (R=CH3 to C7H15, C17H35) of varying hydrophobicity, and thus solubility, has been synthesised. Phase separation mechanism of thermochromism. A more attractive theory for the mechanism of the thermochromic Each member has been used as a developer in stearic acidbased compositions in conjunction with a variety of colour eVect assumes that phase separation plays a major role.Thus little or no colour generation occurs in the molten composition formers, and the eVect of developer structure on thermochromism measured. because the colour former and acidic developer are dissolved in the co-solvent; the average environment experienced by the The consequences of using diVerent ratios of components have also been examined.In addition to giving rise to a chromogenic compound is therefore relatively non-polar, which encourages lactonisation, so that the equilibrium is well dilution eVect, a high ratio of co-solvent to developer and colour former should, in terms of the phase separation theory, over to the ring-closed, colourless spirolactone side.On cooling, the solubilities of the colour former and developer fall, so inhibit precipitation of the solutes, leading to a low degree of colour development, whereas too little co-solvent should pre- that eventually a proportion of these two components separates from the bulk monodisperse solution phase on solidification.vent complete dissolution on melting and restrict colour loss. To examine these eVects, a series of compositions of diVering The developer precipitates, bringing colour former with it (or vice versa). In this phase, the colour former experiences a more mixture ratios based on stearic acid and incorporating the developers 6 were prepared. polar environment and intimate contact with the developer, resulting in ring-opening and generation of colour.On heating, All the developers 6a–h caused colour development, but to varying degrees depending not only on the colour former used, the two phases merge and the colour former returns to its colourless spirolactone form. A simplified diagrammatic rep- but also on the structure of the developer, the ratio of the components and the age of the composition (see following resentation of the phase changes is shown in Scheme 5.This explanation of the mechanism has many implications. Sections). In addition to having the correct balance of acidity of the developer and basicity of the colour former, a balance must EVect of developer structure. The colour intensity of each composition in the solid state was determined by measuring also be struck between the solubilities and solvating powers of the components in the composition. The theory suggests the reflectance of the powdered material at the lmax of the colour former.The results are summarised in Fig. 1–3. All that the developer must be soluble enough to dissolve in the molten fatty acid, but not so soluble that it does not precipitate showed a similar pattern of colour development.Thus developers with short alkyl chains (6a, 6b) conferred relatively pale when the composition solidifies. Also, if too much co-solvent is present in the system, the solutes will not precipitate to a colours; inserting one or two methylene fragments into the J. Mater. Chem., 1998, 8, 2677–2683 2679in conjunction with 3 (Fig. 3); reflectance values of compositions containing 6c or 6d approached those achieved with Bisphenol A. While colour development decreased with the employment of developers of longer chain length, the colour yields were significantly greater than without any developer. Both the visible and near infrared absorption of compositions formulated with 4 exhibited a strong dependence on the structure of the developer (Fig. 4). All these findings conform to the predictions of the phase separation theory. The developers with short alkyl chains (6a, 6b) are polar and not very soluble in the co-solvent (stearic acid), so that they conferred only pale colours when the compositions solidified on cooling, because much of the developer did not dissolve in the melt during composition prep- Fig. 1 Reflectance of powdered 505151 stearic acid–6–1 compositions aration. On the other hand, the developers with long alkyl at 610 nm. chains (6f–h) were found to be very soluble in the molten compositions, but gave only weakly coloured waxes, presumably because their high solubility inhibited phase separation on cooling. The developers possessing alkyl chains of intermediate length (6c, 6d) seemed to have the required balance of satisfactory solubility in the molten stearic acid and poor solubility in the cold wax, so that in these cases, the highest amount of phase separated material was produced and colour development was greatest.EVect of fractional composition on thermochromism. The influence of the mixture ratio on the colour strengths of solid stearic acid56a–g51 compositions of ratio 2551:1, 5051:1 and 10051:1 is depicted in Fig. 5. Increasing the ratio of co-solvent to developer and colour former generally lowered colour strength of the solids as anticipated. In addition to diluting Fig. 2 Reflectance of powdered 255151 stearic acid–6–2 compositions the colour, the increased proportion of co-solvent reduces at 605 nm.colour development by inhibiting precipitation of the components on solidification of the molten composition. Consequently, while the pattern of colour strength in relation to developer chain length as described above for the 505151 series (see Section above) was generally reproduced in the 255151 and 1005151 formulations, the average colour strength Fig. 3 Reflectance of powdered 505251 stearic acid–6–3 compositions at 720 nm. chain of the latter to give 6c–d caused deepening of the colour of the compositions, but further lengthening of the alkyl residue decreased the colour intensity. Fig. 4 Reflectance of powdered 505251 stearic acid–6–4 compositions In the case of the 505151 stearic acid5developer51 composi- at 655 nm and 910 nm.tions, addition of 6a or 6b brought a reduction in minimum reflectance of around 20% compared to the formulation lacking developer (Fig. 1); use of the propyl and butyl analogues led to a further lowering of the minimum, while the colour yield with 6e was close to the level obtained with 6a and 6b. The developers with the longest chains produced the highest reflectances (weakest colours), the formulation containing 6g having a reflectance approaching that corresponding to the residual colour of 1 in stearic acid alone, i.e.a composition with no developer at all. A similar pattern was shown by the series based on 2 (Fig. 2). Colour development was generally greater than in the previous series; even the long chain developers produced significantly lower reflectances than the mixture containing the Fig. 5 Reflectance of powdered 255151, 505151 and 1005151 stearic acid–6–1 compositions at 610 nm. colour former alone. The developers were particularly eVective 2680 J. Mater. Chem., 1998, 8, 2677–2683of the former series was stronger than that of the 505151 displayed by the analogous range of 505151 formulations shown in Fig. 6. compositions, whereas the average minimum reflectance of the latter series was higher. Wax constituting a 1005151 mixture Week-old compositions were generally more weakly coloured than freshly prepared material. In certain cases, the of stearic acid56g51 was almost as pale as the corresponding material lacking developer, suggesting that the acylamino- colour change occurred suYciently rapidly to be perceptible to the naked eye; for example, molten 255151 and 505151 phenol in the former composition had little tendency to separate from the stearic acid phase and/or form a separate stearic acid56g51 compositions gave dark blue waxes when first quenched, which, over a few seconds, faded to pale blue.phase with 1. The compositions containing 6b did not conform to the The corresponding 505251 composition based on 3 lost colour over the course of a few minutes, while changes in the colours typical pattern: colour yield slightly increased on raising the proportion of co-solvent from 505151 to 1005151, presumably of 255151 stearic acid56f–g51 and 505251 stearic acid56f–g51 formulations were noticeable after a few hours.because more developer can go into solution and consequently separate with colour former on cooling.The phenomenon can be explained in terms of phase separation: rapid cooling of the molten mixture brings about brief colour development through the initial co-separation of devel- EVect of ageing on the solid composites. Whereas all the coloured solid compositions containing Bisphenol A remained oper and colour former; secondary separation then takes place, whereupon colour former crystallises out of the developer (or intensely coloured over long periods of time, the compositions formulated with the 4-acylaminophenols tended to lose their vice versa) so that colour is lost as the former is no longer intimately associated with the latter.When the incompatibility initial colour on standing with varying degrees of rapidity (Fig. 6–8). The series of 255151 and 1005151 stearic acid5651 between colour former and developer is particularly great, separation occurs rapidly enough for the colour of the wax to compositions exhibited a similar pattern of colour loss to that fade perceptibly over the course of several seconds or minutes. Thus not only is the solubility of the developer in the cosolvent an important factor, but the solubility of the colour former in the developer (or association between them) is of major relevance.Experimental Thin layer chromatography was performed using alumina plates (DC Alufolien Aluminiumoxid 150 F254 neutral type T, Merck). Melting points were determined on an Electrothermal melting point apparatus and are uncorrected. Thermal, FTIR and elemental analyses were performed on a DuPont 2000 diVerential scanning calorimeter, using a Perkin-Elmer 1740 spectrophotometer and in the Department of Chemistry of the Fig. 6 Time dependence of reflectance of powdered 505151 stearic acid–6–1 compositions at 610 nm. University of Leeds, respectively. Reflectance measurements were obtained by means of a Perkin-Elmer Lambda 9 UV/visible/NIR spectrophotometer.Preparation of the colour formers and developers Fluoran colour former 2. 4-Acetylaminophenol 6a (4.98 g, 33 mmol) and 2-(4-N,N-diethylamino-2-hydroxybenzoyl )benzoic acid 5 (8.91 g, 28 mmol) were added portionwise to stirred sulfuric acid (98%, 15 ml ) over 20 min so that the temperature remained under 45 °C. The mixture was heated and stirred at 55 °C for 22 h, water (15 ml ) added, maintaining the temperature below 85 °C, and then heated with stirring for 3 h at 90–95 °C.The intense red solution was cooled to room temperature and poured into a solution of sodium hydroxide (18 g) in water (130 ml ) at such a rate that the temperature Fig. 7 Time dependence of reflectance of powdered 255151 stearic did not exceed 80 °C; aqueous ammonia (32%, ca. 8 ml) was acid–6–2 compositions at 605 nm. added to make the pH of the purple suspension weakly alkaline. The solid was collected, washed with very dilute aqueous ammonia and dried to give crude 2¾-amino-6¾-N,Ndiethylaminofluoran 7. Crude 7 (8.00 g, 21 mmol) was added to sodium carbonate (5.71 g, 54 mmol), benzyl chloride (98.5%, 6.75 ml, 58 mmol) and water (20 ml ), before heating to 90 °C overnight to give a dark green mixture.This was cooled to room temperature, the aqueous phase decanted oV and the residual dark green tarry product washed several times with water. Stirring with ethanol (20 ml ) at 65 °C for 30 min furnished a green suspension which was cooled to room temperature and filtered. The collected solid was washed with a little ethanol (5 ml, twice) and then butanone (5 ml ) to give crude 2 as a pale green powder (10.04 g, 86% crude yield).A portion (1.00 g) of this material was recrystallised twice (ethanol–DMF, 1551) aVord- Fig. 8 Time dependence of reflectance of powdered 505251 stearic acid–6–3 compositions at 720 nm. ing analytically pure, pale beige crystals (0.58 g, mp J. Mater.Chem., 1998, 8, 2677–2683 2681172.5–174 °C). Microanalysis found C, 80.5; H, 6.1; N, 5.0% 10.15% (C26H27N3O2 requires C, 75.52; H, 6.58; N, 10.16%). DSC indicated reasonable purity, analysis revealing a small (C38H34O3N2 requires C, 80.5; H, 6.0; N, 4.9%). DSC and TLC (alumina, 951 toluene–ethyl acetate) indicated purity, endotherm at 248.4 °C and a sharp, major endotherm at 261.7 °C, with decomposition occurring at around 316 °C.TLC showing a sharp endotherm at 172.3 °C and a single dark green spot respectively. FTIR (KBr)/cm-1: 1752 (lactone (alumina, 951 toluene–ethyl acetate) revealed a single green CLO). spot of Rf 0.57. FTIR (KBr)/cm-1: 1752 (lactone CLO). Vinylphthalide colour former 3. Magnesium (98%, 4.2 g, Developers 6a–h. Apart from the readily available 6a, the 0.17 mol ) and diethyl ether (20 ml ) were stirred under nitrogen developers 6 were synthesised from 4-aminophenol by the as iodomethane (99%, 25.0 g, 0.17 mol) was added dropwise methods of Fierz-David and Kuster.20 The 4-propionyl ana- over 20 min at such a rate as to maintain a continuous logue 6b was obtained using propionic anhydride, whereas the exotherm and gentle refluxing.The mixture was gently refluxed longer alkyl chain derivatives (R=C3H7 to C7H15) were for 25 min by which time almost all of the magnesium had synthesised from the corresponding acid chlorides, while the dissolved. The reaction mixture was protected from light and stearoyl derivative 6h was prepared by condensation with a solution of Michler’s ketone (98%, 9.0 g, 33 mmol) in toluene stearic acid.(280 ml ) at room temperature was run in. The mixture was stirred under nitrogen in the absence of light at ambient temperature for 22 h, before cautiously quenching with water Preparation of the compositions (200 ml ) to give a light blue emulsion, which was stirred for 15 min. A solution of ammonium chloride (30 g) and acetic Formulations of 5051:1 stearic acid (2.00 g), developer 6 acid (99%, 15 ml ) in water (150 ml ) was added and the whole (Table 3) and 1 (0.040 g) were prepared by heating the mixture stirred for 3.5 h.The aqueous phase was washed twice with to 100–105 °C and stirring until dissolution was complete, or toluene (50 ml ) and the extracts combined with the organic until no more dissolution occurred, and quenching with the phase, from which the solvent was removed by rotary evapor- addition of ca. 20 ml of cold water. The solidified wax was ation after drying overnight with magnesium sulfate. The pale filtered oV, allowed to air-dry overnight and powdered. blue-green residue (8.59 g, 98% crude yield, mp 116–120 °C) Series of compositions of ratio 255151 and 1005151 were was recrystallised in ethanol, furnishing Michler’s ethylene (8) also prepared in a similar manner.Formulations containing 2 as a pale blue lustrous solid (7.15 g, mp 121–123 °C, lit.,18 (255151), 3 (505251) and 4 (505251) were made by the same 121–122 °C). procedure, except that the mixtures were heated to 95–100 °C, A mixture of 8 (1.33 g, 5.0 mmol), 2-(4-N,N-dimethylamino- 105–110 °C and 100–105 °C, respectively.The ratios strictly benzoyl )-3,4,5,6-tetrachlorobenzoic acid 919 (2.04 g, only apply to compositions containing 6a; the amounts of 5.0 mmol) and acetic anhydride (7.5 ml ) was stirred and developer listed in Table 3 are equimolar so that, within a heated to reflux for 20 min. The dark green mixture was particular series of compositions, each member of the range allowed to cool and poured into a mixture of toluene (50 ml ), consists of a fixed molar ratio of developer to colour former ice (50 ml ) and aqueous ammonia (32%, 10 ml ).The emulsion and co-solvent. was destroyed by addition of a little dichloromethane; the The procedure for measuring the reflectance of the powdered organic phase was collected and the aqueous phase extracted formulations involved packing the material into a glass-fronted with more dichloromethane.The combined extracts were cylindrical cell comprising two close-fitting components. The washed several times with water, dried (magnesium sulfate) composition was compressed in the cell by insertion of the and rotary evaporated to dryness, giving a dark yellow- cylinder and the reflectance measured (scan speed brown solid (2.94 g, 90% crude yield, mp 220–222 °C). 240 nm min-1); the powder was displaced and re-compressed Recrystallisation (methoxyethanol, methoxyethanol–DMF before taking a second measurement. The technique was found 1251 twice) furnished 3 as green needle-like crystals (1.03 g, to yield consistent results. mp 227.5–229 °C), which darken in sunlight, but return to their original colour in the dark.Microanalysis found C, 61.9; Molecular modelling H, 4.8; N, 6.1; Cl, 20.8% (C34H31N3O2Cl4 requires C, 62.3; H, 4.8; N, 6.4; Cl, 21.6%). DSC indicated reasonable purity, The ‘Hyperchem’ software package (Release 3 For Windows, showing a single endotherm at 226.6 °C. TLC (alumina, 951 Autodesk Inc.) was used in an attempt to visualise the geometry toluene–ethyl acetate) revealed one blue-green spot.FTIR of the ring-opened colour former molecules. The procedure (KBr)/cm-1: 1758 (lactone CLO). involved optimising the geometry by using molecular mechanical methods, conducting the iterative energy-minimising rou- Fluorene colour former 4. Aluminium chloride (96%, 50.0 g, tines to the desired energy gradient (0.01 kcal A° -1 mol-1) 0.36 mol ), urea (7.5 g, 0.12 mol) and aluminium chloride with the Polak–Ribiere algorithm. The MM+ force field was hexahydrate (99%, 0.75 g, 3.1 mmol) were stirred and heated used as it was deemed the most appropriate for relatively to 125 °C.Crystal Violet Lactone (1) (97%, 5.2 g, 12 mmol) small molecules like the colour formers. For the sake of was added and the mixture stirred at 135–140 °C for 4 h.simplicity, the colour formers were assumed to ring open to Heating was continued for another 18 h by which time the give a free carboxylic acid group, i.e. the nature of the reaction mixture had become a solid mass. The reaction was developer and the type of association was not considered in quenched by gradual addition of water (400 ml ) to give a the modelling.blue-green suspension, which was treated with hydrogen peroxide (30%, 1.2 ml ) and stirred for 1.5 h. The suspension was extracted several times with dichloromethane; the combined extracts were washed and dried over magnesium sulfate. Table 3 Masses of developers 6 used in 505151 compositions Removal of the drying agent and solvent furnished a dull green solid (4.43 g, 89% crude yield).Three recrystallisations Developer Mass/g Developer Mass/g (toluene/charcoal ) gave 4 as buV crystals (1.20 g, mp 6a 0.0400±0.0004 6e 0.0549±0.0005 225–227.5 °C). Repeated recrystallisation of 0.40 g of this 6b 0.0438±0.0004 6f 0.0586±0.0005 material from toluene and ethanol aVorded analytically-pure 6c 0.0475±0.0004 6g 0.0623±0.0005 pale cream crystals (0.07 g, mp 237–242 °C, lit.,8 244–246 °C, 6d 0.0512±0.0005 6h 0.0995±0.0005 lit.,10 240–245 °C).Microanalysis found C, 75.3; H, 6.7; N, 2682 J. Mater. Chem., 1998, 8, 2677–2683the reverse of that observed for the three-component com- Conclusions posites discussed above. Presumably the b-estradiol assumes Thermochromic systems have been prepared which deliver the roles of both developer and co-solvent as not only does it sharp reversible changes from coloured to colourless states have a phenolic residue, typical of the first component, but it when heated above their melting points. A mechanism for the also has a hydrophobic alkyl skeleton, characteristic of the phenomenon based on steric considerations was rejected on second.In the molten composite, the colorant is ring-opened the basis of X-ray crystallographic data and molecular model- and thus coloured, because it is dissolved in, and can interact ling predictions.An explanation in terms of phase changes with, the b-estradiol, whereas on solidification, the colorant was investigated by preparing compositions from developers separates from the bulk phase, ring-closes and the formulation of diVering hydrophobicity.The findings lent weight to such loses its colour. a theory, although more direct evidence of the existence of the Within the framework of the proposed mechanism, it would proposed phases and phase changes is necessary for confir- appear that another approach to produce coloured-to-colourmation of its validity. less formulations comprising only two components is possible: Several influences on colour yield were identified; for in these, a colorant of a design which combines the functions example, developer structure had a significant eVect.The of both colour former and developer, for example a spirolacdevelopers with short alkyl chains were polar and not very tone bearing phenolic residues, is formulated with a co-solvent.soluble in the co-solvent, so that they conferred only pale Provided the colorant is readily soluble in the melt, separates colours when the compositions solidified on cooling, whereas from the bulk phase on solidification and has the correct the developers with long alkyl chains were found to be very balance of basicity and acidity, the composition will be reversisoluble in the molten compositions, but gave only weakly bly thermochromic: in the molten composition, lactonisation coloured waxes, presumably because their high solubility will be favoured, while cooling will lead to separation of a inhibited phase separation on cooling.The developers pos- colorant phase in which intermolecular interaction between sessing alkyl chains of intermediate length seemed to have the the lactone and phenolic elements can occur causing ringrequired balance of satisfactory solubility in the molten stearic opening and colour generation.acid and poor solubility in the cold wax, so that in these cases, the highest amount of phase separated material was produced Acknowledgement on solidification and colour development was greatest.As anticipated, colour strengths were observed to be dependent The authors wish to thank the Defence Clothing and Textiles on mixture ratio: increasing the ratio of co-solvent to developer Agency (Colchester, UK) for the funding and support of and colour former generally lowered colour yields. However, this work. in certain cases, an increase in the proportion of co-solvent relative to developer improved colour development as more References of the latter could be dissolved and consequently separate with colour former on cooling.While formulations containing 1 F. Jones, Rev. Prog. Coloration, 1989, 19, 20. Bisphenol A have not been observed to change colour over 2 A. R. Katritzky, Z.-X. Zhang, H.-Y. Lang, N. Jubran, L. M. Leichter and N. Sweeny, J.Mater. Chem., 1997, 7, 1399. time, compositions prepared from the N-acylaminophenol 3 D. Aitken, S. M. Burkinshaw, J. GriYths and A. D. Towns, Rev. developers were found to fade to diVerent extents with varying Prog. Coloration, 1996, 26, 1. degrees of rapidity from over the course of seconds to weeks. 4 K. Naito, Appl. Phys. Lett., 1995, 67, 211. The phenomenon may arise through phase separation of the 5 J.GriYths, in ChemiChromics 95 Conference Papers, Manchester, colour former and developer from each other after the initial 1995. 6 J. Schofield, personal communication. separation of the developer together with the colour former 7 NCR Corp., US Pat. 4 119 776 (1978). on cooling. 8 Yamamoto Kagaku Gosei, Eur. Pat. 124 377 (1984). If, as suggested by the above findings, the explanation of 9 Kanzaki Paper Manufacturing Co.Ltd., Eur. Pat. 209 259 (1986). the thermochromism lies in phase separation, there are a 10 Yamamoto Kagaku Gosei, Eur. Pat. 278 614 (1988). number of conclusions that can be drawn concerning the 11 D. H. Williams and I. Fleming, Spectroscopic methods in organic design and optimisation of compositions based on co-solvent/ chemistry, 3rd edn., McGraw-Hill, 1980. 12 D. Aitken, unpublished work. developer/colour former combinations. The developer must 13 J. GriYths, J. Soc. Dyers Colour., 1988, 104, 416. possess just enough solubility in the molten co-solvent to cause 14 M. Kubata, H. Yoshioka, K. Nakatsu, M. Matsumoto and complete dissolution and yet have suYciently poor solubility Y. Sato, in Chemistry of Functional Dyes, ed. Z. Yoshida and in the cooled composition to maximise phase separation. In Y. Shirota, Mita Press, Tokyo, 1989, p. 223. addition to the colour former having satisfactory solubility in 15 G. Rihs and C. D. Weis, Dyes Pigm., 1991, 15, 107. the co-solvent and the correct basicity for colour changes of 16 G. Rihs and C. D. Weis, Dyes Pigm., 1991, 15, 165. 17 J. Sueyoshi, M. Kubata, H. Yoshioka, K. Nakatsu and Y. Hatano, high contrast, the colorant must be compatible with the in Chemistry of Functional Dyes, ed. Z. Yoshida and Y. Shirota, developer, otherwise phase separation between these two com- Mita Press, Tokyo, 1992, vol. 2, p. 34. ponents will occur after the initial separation from the co- 18 P. PfeiVer and R. Wizinger, Ann., 1928, 461, 132. solvent, resulting in loss of colour after solidification of the 19 A. Haller and H. Umbgrove, C.R. Hebd. Seances Acad. Sci., 1899, composition. 129, 90. Phase separation can also explain the thermochromism of a 20 H.E. Fierz-David and W. Kuster, Helv. Chim. Acta, 1939, 22, 82. two-component system based on 1 and b-estradiol that gives a change from colourless to coloured on melting,4 which is Paper 8/05994B J. Mater. Chem., 1998, 8, 2677–2683 2683