Nuclear magnetic resonance spectroscopy studies of 2-(2,4-dinitrobenzyl) pyridine and longlived

Transcription

1 ME SEAR PAS & MS JURALS ABUT TAT US uclear magnetic resonance spectroscopy studies of 2-(2,4-dinitrobenzyl) pyridine and longlived photoinduced products This article has been downloaded from IPscience. Please scroll down to see the full text article J. Phys. D: Appl. Phys ( The Table of ontents and more related content is available Download details: IP Address: The article was downloaded on 09/12/2009 at 10:48 Please note that terms and conditions apply.

2 J. Phys. D: Appl. Phys. 32 (1999) Printed in the UK PII: S (99) uclear magnetic resonance spectroscopy studies of 2-(2,4-dinitrobenzyl) pyridine and long-lived photoinduced products T unes, Y Eichen, M Bastos, D Burrows and P Trommsdorff ITPL/IST, Departamento de Engenharia de Materiais, Av. Rovisco Pais, Lisboa, Portugal Laboratoire de himie Supramoléculaire, Université Louis Pasteur, 4, Rue Blaise Pascal, F Strasbourg, France and the Department of hemistry, Technion-Israel Institute of Technology, aifa 32000, Israel Departamento de Química da Universidade de oimbra, 3049 oimbra, Portugal Laboratoire de Spectrométrie Physique, Université J Fourier Grenoble 1-RS (UMR 5588) BP 87, F St Martin d ères edex, France pcteresa@alfa.ist.utl.pt (T unes) Received 5 May 1999 Abstract. A study of the photo-induced and thermally activated tautomerization of 2-(2,4-dinitrobenzyl) pyridine (α-dbp) by solid state nuclear magnetic resonance spectroscopy (MR) is presented. Whereas similar high-resolution 1 and 13 spectra were obtained from polycrystalline powders before and after ultraviolet light irradiation, the corresponding spectral data obtained from PV films doped with α-dbp show significant changes. The α-dbp phototautomer lifetimes in PV films (2.4 h or 6 d) were determined from the proton MR signal recovery. Scanning electron microscopy (SEM) and differential scanning calorimetry of the doped films show that, in addition to matrix isolated molecules, α-dbp is incorporated into the polymer matrix in the form of nanocrystalline aggregates. At high guest concentrations, the crystallites can be modified by controlled solvent evaporation, as monitored by MR and SEM. The variation of the 1 spectrum with temperature shows the presence of at least two stable conformational isomers in the solid samples. 1. Introduction hemical reactions in the solid state involving reversible light-induced proton transfer are, in general, not accessible by nuclear magnetic resonance (MR) methods due to the very small steady state concentration of the metastable species characteristic of such reactions, as a consequence of the low yields and of the short lifetimes. The present study is based on the fact that, for 2-(2,4-dinitrobenzyl) pyridine (α-dbp) incorporated into polymer matrices, the phototautomer could be produced at concentrations sufficient to enable a study of the metastable state using high-resolution solid state MR methods for the observation of 1 and 13 spectra. Recent increased interest in photochromic proton transfer systems is motivated by potential applications in areas such as optical imaging, data storage and holography [1]. α-dbp is photochromic both in the solid state and in solutions. Upon irradiation with ultraviolet light (UV), a blue coloured tautomer is formed. The original discovery of the photochromism of this molecule dates back to hichababin et al [2]. Subsequently, different structural assignments of the phototautomer have been proposed and discussed, and details of the reaction mechanism continue to be the object of ongoing work [3 7]. Whereas the structure of the phototautomer can now be considered as well established, the mechanism, the intermediates, and the rates of the reaction depend on the environment (fluid solution, crystal) and temperature, are not fully elucidated. For example, the lifetime in the solid (a few hours) is much longer than in solution (16 s in ethanol solution at 280 K). Scheme 1 shows the proposed reaction mechanism and the tautomers of α- DBP: the thermodynamically stable (1) state and the (2) and tautomers, identified as the reaction intermediates both in solutions [3 and references therein] and in the crystalline state [7, 8] /99/ $ IP Publishing Ltd

3 Spectroscopy studies of 2-(2,4-dinitrobenzyl) pyridine - - hν 1 '' heat heat, hν 2 hν '' hν 1 heat '' Scheme 1. b c* (4) (5) (4) (1) (2) (2) (1) (6) (1) (2) (1) (7) (9) (8) (11) (7) (12) (5) (6) Figure 1. Molecular geometry of α-dbp [9]; bond lengths and the two shortest methylene hydrogen to nitro oxygen distances are indicated. (2) (4) (10) (9) (8) (4) Proton transfer from the benzylic group to the nitrogen atom of the pyridine ring may involve either intra- or intermolecular reactions or even a reaction involving both intraand intermolecular pathways. The argument in favour of an intramolecular reaction in the solid, is based on the observation that the shortest non-bonded interaction between the benzylic hydrogens and a basic atom is the intramolecular hydrogen bond formed between one of the oxygen atoms of the o-nitro group and one of the benzylic hydrogens. This is expected to favour the formation of the group as the reaction intermediate [9]. Figure 1 shows the molecular geometry and bond lengths for α-dbp. The two shortest benzylic hydrogen to o-nitro-oxygen distances are also indicated [9]. Important objectives of work aimed at potential applications are the gain of control of the stability of the metastable species and the increase of the overall reaction yield. hemical engineering approaches, in order to modify the proton acceptor group of DBP with the view of stabilizing the phototautomer, have been successfully used [5]. The phototautomer can also be stabilized by the environment, as is dramatically demonstrated by the differences in lifetime of the blue tautomer in different polymorphs of the same molecular entity (Khatib et al [7]). Incorporation 2109

4 T unes et al (a) Figure 3. DS curves for: (a) PV films containing 33% α-dbp; (b) α-dbp crystals. (b) Figure 2. SEM photographs obtained from (a) film 1 and (b) film 2 (see section 2 for details). of photochromic molecules into polymeric hosts offers therefore the possibility to produce tractable photoactive materials, so that the characterization of these systems is feasible. In this paper we report a study on PV films doped with high concentrations of α-dbp. The physical nature of the guest (isolated molecules or phase separated, microand nanocrystalline aggregates) was established by scanning electron microscopy (SEM) and differential scanning calorimetry (DS). For this material, the conversion to the blue phototautomer could be realized to the extent that an MR study of the photochromism of α-dbp became possible. The assignment of the MR spectral data is vindicated by a comparison with α-dbp powders, solution data, and solid state 1 spectra from selectively deuterated α-dbp and chemically related compounds. 2. Experimental 2.1. Materials α-dbp (Lancaster Synthesis) was used either as-received or after recrystallization in ethanol. Further purification of the compound was achieved by column chromatography over neutral alumina using ethanol as eluent and subsequent recrystallization in ethanol. PV ( M w , BD) was used as-received. α-dbp doped PV films ( 100 µm thick) were prepared by the controlled evaporation (40, dark) of tetrahidrofuran (TF) solutions containing 5% (w/v) Figure 4. 1 MAS MR spectra of a PV film doped with α-dbp (film 1, see section 2), obtained with a spinning rate of 4.4 kz: (a) before irradiation; (b) immediately after the end of irradiation (EI) with 366 nm UV light; (c) 18 h after EI; (d) immediately after the second irradiation; (e) several weeks after EI; and (f) 2 h after the third irradiation. 2110

5 Spectroscopy studies of 2-(2,4-dinitrobenzyl) pyridine general, 64 (m) echoes were acquired. Using this sequence, it was possible to obtain spin spin (T 2 ) relaxation weighted spectra (more correctly designated as T2 -weighted spectra because they contain the effect of static magnetic field inhomogeneity): assigning large (1 ms) or short (20 µs) values to the time delay τ (with a high value of n to allow the magnetization to return to thermal equilibrium), in order to filter out signals corresponding to rapidly dephasing magnetization (with short T2 ) or to observe all the signals, but with a smaller contribution from broader resonances, respectively. igh-resolution 13 spectra were obtained with broadband proton decoupling (solutions) or proton dipolar decoupling using the cross-polarization/magic-angle spinning (P/MAS) technique (solid state). hemical shifts are reported in parts per million from tetramethylsilane, taken as an internal (solutions) or external (solid state) reference, and were directly obtained from the positions of the maximum of the lines with an error of ±0.2 ppm, estimated from the spectra obtained with the lowest digital resolution. 3. Results and discussion 3.1. Physical characterization of DBP in PV films Figure 5. 1 MR spectra of a PV film doped with α-dbp (film 2, see section 2), obtained with a MAS rate of 10 kz: (a) before irradiation; (b) immediately after the end of irradiation (EI) with 366 nm UV light; (c) 15 m after EI; (d) 45 m after EI; (e) 2 h after EI; (f) 3 h and 15 m after EI; and (g) 10 h and 30 m after EI. PV and 33% (w/w as compared to PV) α-dbp (film 1). Film 2 was prepared by dissolving film 1 in TF and allowing the solvent to evaporate at ambient temperature, under natural light. Film 2 was thinner than film Apparatus The 366 nm line of a mercury lamp was used for the photolysis of the different samples. SEM photographs were obtained using a JSM-840 Scanning Microscope (Microspec WDX). DS measurements were made on 2 3 mg samples on a Polymer Laboratories PL-DS apparatus under a nitrogen atmosphere using a heating rate of 10 min 1. igh-resolution 1 and 13 spectra (liquid and solid samples) were recorded at and 75.5 Mz, respectively, using a Bruker MSL 300 P spectrometer. Solid state 1 spectra were run with a single radiofrequency (RF)-pulse sequence (a typical value for the RF-pulse length corresponding to a 90 magnetization flip angle was 2 µs and the time delay after data acquisition was 2 s). ccasionally, a spin-echo RF-pulse sequence [(90φ τ 180 ϕ τ acquisition nτ) m] was selected in order to obtain an enhancement of narrow signals, by reducing the contribution of other broader resonances; a quadrature RFpulse phase cycle (φ = x, x,x, x and ϕ = y, y,y, y) was used to correct for possible imperfect pulse lengths and phases. The signal acquisition was started on the top of the echo and a typical acquisition time was 14 ms; in At the high doping levels used, the distribution of α-dbp in the PV films is not homogeneous, and both matrix isolated and aggregated molecules are expected to coexist. The formation of α-dbp microcrystallites is apparent in SEM images as shown in figures 2(a) and 2(b) for films 1 and 2, respectively. rystal aggregates are identified in both samples. In film 2 the spatial distribution of the crystals in the polymer matrix is more uniform, the size is somewhat larger than in film 1 but the size distribution is narrower. DS data on the films (figure 3(a)) exhibit a weak endotherm at 88, close to the melting transition of pure α-dbp at 92, = 29.6 kjmol 1 (figure 3(b)). o transitions in this temperature region are observed in DS studies of pure PV films of the same thickness. The decrease in the melting temperature of the PV embedded α-dbp, as compared to bulk α-dbp, is expected for microcrystallites and may also reflect the existence of different crystal phases. Polymorphism is indeed common for this class of compounds and even glassy phases have been observed [6, 7, 10] Photochromism of α-dbp doped PV films The coexistence of matrix isolated α-dbp molecules and of polydispersed crystalline aggregates is apparent from the time evolution of the photoproduct ( form) monitored in the visible absorption spectrum of irradiated films as a loss of absorption around 600 nm. The decay of the absorption at 600 nm follows a complex kinetic behaviour. Both shortlived and long-lived components are observed, the latter has lifetimes of several hours. Assuming a biexponential behaviour, the analysis of the decay of the absorbance at 600 nm of PV films doped with 0.29 and 0.61% (w/w) α-dbp gave 67 and 106 s for the fast components and 5 and 2.5 h for the slow components, respectively (at ambient temperature). The shorter lifetimes are comparable to the longest lifetimes 2111

6 T unes et al Figure 6. Plot of the intensity/percentage recovery of the signal observed at 0.8 ppm in the proton spectra shown in figure 2, obtained for the time interval from the end of UV light irradiation (366 nm wavelength) to about 9.2 h later. The insert shows the linear variation of the logarithm of the recovered intensity as a function of time (see text). The reference for 100% recovery is the intensity of the signal before irradiation, 50 a.u MeDBP conjugate base Scheme 2. observed for isolated molecules in solution, while longer lifetimes of over one hour are typical for crystalline samples [3 10]. It must be pointed out here that a rate distribution, instead of a biexponential model, is expected to account more accurately for the variation in local environment of DBP in a PV matrix, which has to be taken into account specially for isolated guest molecules (probably dominant at low doping levels). In fact, using polycrystalline or amorphous matrices such as PMMA, Sixl and Warta [3] reported the observation on distribution of rate constants and of activation energies for each DBP photoreaction and thermal reaction (see scheme 1) but no data were presented, neither was the level of doping indicated MR study of the phototautomerization reaction in the solid state 1 MR spectra of PV films doped with α-dbp prior and subsequent to irradiation with 366 nm UV light are shown in figures 4 and 5 for films 1 and 2, respectively. Film 1. Spectra before and immediately after irradiation (figures 4(a) and 4(b)) show the disappearance of resonances at 1.4 ppm (narrow line) and 3.3 ppm (shoulder). After the sample had been 18 h in the dark, these signals return (figure 4(c)) and are fully recovered after several weeks (figure 4(e)). The cycle (photolysis recovery in the dark) was repeated several times (spectra obtained during three cycles are shown in figure 4). The two sharper resonances affected 2112

7 Spectroscopy studies of 2-(2,4-dinitrobenzyl) pyridine Figure 7. Proton MAS spectra of the compounds: (a) α-dbp; (b) α-dbp selectively deuterated in the benzylic groups; (c) methyl α-dbp derivative; and (d) quaternary salt of α-dbp. The following spinning rates were selected to obtain spectra (a) to (c) and spectrum (d): 4.5 and 10 kz, respectively. In this last case, the T2 -weighted mode was used with τ = 1 ms (see section 2 for details). by the photolysis are assigned to part of the protons in α- DBP crystals whereas a single broad peak at 4.7 ppm is due to the protons of the polymer matrix superimposed on the remaining α-dbp protons, with two weak bands at about ±15 ppm as assigned to be spinning side bands. From the time variation of the intensity of the signal at 1.4 ppm, the lifetime of the photoproduct in this film is estimated to be about 6 days. Film 2. Similar observations were made for film 2, which was thinner than film 1. The spectra before and immediately after photolysis are shown in figure 5. The chemical shifts of the resolved α-dbp proton resonances are slightly different from those found for film 1: 0.8 and 3.8 ppm as compared to 1.4 ppm and 3.3 for film 1. The line width, 60 z, is about half the value measured in the spectrum of film 1. In contrast to film 1, for which a full recovery of the original MR spectrum was observed, the original signal amplitude was not fully recovered in this film. Figure 6 shows the variation of the intensity of the signal at 0.8 ppm as a function of time between the end of the irradiation (EI) and about four days after EI. The intensity of the signal before irradiation (not presented) is 50 on the same scale of arbitrary units. 2113

8 T unes et al Figure 8. 1 spectra obtained with MAS at 10 kz, in T2 -weighted mode with τ = 20 µs (see section 2 for details), from: (a) glassy α-dbp, immediately after the preparation; (b) glassy α-dbp, 3 h after the preparation (without stopping MAS during this period; and (c) the difference spectrum, (b) subtracted from spectrum (a). Approximately 5 h after EI, and for the next 2 h, the signal intensity remains in the interval 8.4 ± 0.4 a.u. Assigning the value 8.4 a.u. to the equilibrium intensity (I )ofthe product and naming I 0 the signal intensity obtained 15 min after EI (1.86 a.u.), a linear plot of ln((i I )/(I 0 I )) against time, in the time interval min after EI, yields a lifetime of 2.4 h (χ 2 = 0.99) for the photoproduct (see insert in figure 6). It is worth noting that figure 6 shows continuing recovery up to much longer times and that a monoexponential model is only valid for the indicated period. For comparison, 1 MAS spectra of crystals, macroscopically observed as single crystals with dimensions in the millimetre range, were recorded (not shown here) in T2 weighted mode, with a time delay τ of 1 ms (see experimental section for details); a MAS rate of 2.06 kz was selected. Three narrow signals (line width 30 z) were identified in the low frequency range: 0.8, 1.9 and 3.8 ppm. In the experiments discussed above, the recovery of the 1 MR signal monitors the proton-transfer backreaction, provided that other factors affecting the signal can be excluded. A possible interference is the formation, upon irradiation of the films, of stable radicals, which, at high concentration, would also account for the disappearance of the α-dbp signals in the 1 spectra. owever, a broadening of the PV resonance is expected in this case, since the unpaired electrons provide an important contribution for magnetization dephasing, decreasing the spin spin relaxation time. In contrast, an estimate of T 2 from the line width of the PV signals (figure 5) indicated a decrease of the spin spin relaxation rate after UV irradiation (T 2 values obtained from figures 5(a) and 5(b) were 0.5 and 0.7 ms, respectively). These values, containing the effect of static magnetic field inhomogeneity, correspond to T2. Furthermore, while ESR signals have been observed upon photolysis of α-dbp [11], none have been for the photolysis in solution at room temperature and at 70 [12]. This indicates that the steady-state concentration of any paramagnetic species must be negligibly small but does not exclude reaction pathways with short-lived paramagnetic intermediates. owever, other results suggest that electron transfer processes may be neglected so that the observed MR changes of spectra reflect a proton transfer reaction involving the tautomeric shift of one of the benzylic protons to the nitrogen of the pyridine ring: by stabilizing the form (see scheme 1), through the addition of a neighbouring nitrogen site that may establish a hydrogen bond to the transferred proton, it was possible to increase the corresponding lifetime remarkably (in particular when the pyridine ring was replaced by a 9,10-phenanthroline group) [5]. The signal of the phototautomer of α-dbp is completely buried under the strong PV signal: indeed, in the 1 spectrum recorded for the conjugate base of the methyl derivative of α-dbp (see below) in Dl 3 solutions, the chemical shift of the benzylic group was found to be 6.1 ppm. The decrease of the spectral line width, subsequent to irradiation, of the broad PV signals is now naturally explained by the contribution of superimposed narrow signals of the phototautomer of α-dbp. In contrast to bulk crystals, for which the very large extinction coefficient of the phototautomer at the photolysis wavelength limits the penetration of light and only about 1 micron surface layer of coloured material is formed 2114

9 Spectroscopy studies of 2-(2,4-dinitrobenzyl) pyridine [6, 10], a more efficient exposure to UV light and a quantitative conversion to the phototautomer is possible in the polymer films containing a homogeneous distribution of the microcrystals. In the two films studied here, the differences observed for the MR resonance frequencies as well as the larger changes in the lifetime are attributed to polymorphism of α-dbp; whereas the lifetime and the MR frequencies measured for film 2 correlate well with data obtained in pure crystals, the much longer lifetime observed in film 1 and the shifted MR frequencies indicate the predominance of a different structure. Further interesting information contained in the MR spectra concerns the regularity of the size distribution and the shape of the crystals. arrower lines are observed for a more uniform distribution of more regular crystals in the film, so that MR offers the possibility of monitoring the uniformity of crystal distribution in the film. This conclusion is also supported by the MR observation of the grinding effect on polycrystalline DBP powder: a strong broadening of the previously well-resolved resonances. The most significant change upon UV irradiation, observed in the 1 spectra of α-dbp incorporated in PV films, is the disappearance of two signals at resonance frequencies below 4 ppm. These lines may be assigned either to the protons directly involved in the transfer, i.e. the benzylic (5) or (6) [9], or to other protons affected by the structural rearrangement ensuing the reaction. In the following sections, further MR data are presented in order to clarify these assignments MR study of α-dbp MR study. The powder spectrum of crystalline α-dbp exhibits a broad signal between 6 and 8 ppm and three other signals at 4.9, 3.3 and 0.8 ppm. The dipole dipole interaction is normally the dominant line-broadening mechanism in solids, particularly if the nuclei have similar resonance frequencies. The dipolar interaction constant varies with the distance, r,asr 3. onsequently, short internuclear distances will result in broad non-resolved spectral lines. In the present study, spectral assignments, discussed below, were facilitated by the comparison with assignments for selectively deuterated and chemically related compounds. Since the assignment of the resonances corresponding to the benzylic (5) and (6) protons involved in the tautomerization is of key interest, 1 spectra were recorded from α-dbp selectively deuterated at the two benzylic positions and from two compounds that are analogous to the form of the α-dbp. The powder spectra of these compounds are compared in figure 7 with the corresponding spectrum of α-dbp. In the selectively deuterated compound, the spin concentration, and therefore spin diffusion through dipolar interaction, is reduced so that lines from protons close to that group are narrowed. Three narrow resonances at 1.5, 1.2 and 0.8 ppm are identified instead of a single broader line, at 0.8 ppm, in the parent compound. Similarly, the 1 spectra of the conjugate bases of methyl and phenyl derivatives of α-dbp show narrow signals only between 0 and 2 ppm. The signals observed in this frequency range must therefore be assigned to protons in the two rings, while signals from (5) and (6) must lie at frequencies higher than 2 ppm. Moreover, DBP resonances from the pyridine ring are also expected to account for the broader line in the DBP spectrum, shown in figure 7(a) (the proton signals from the pyridine solvent are detected at ppm); this assignment is in agreement with the strong narrowing of this signal observed at higher temperatures and MAS rates (see section 3.5) MR study. A comparison between the 13 MR spectra of α-dbp in solutions and in the solid state (not shown here) was also performed. α-dbp forms monoclinic crystals (space group P2 l /c) with four symmetry related molecules per unit cell (Scherl et al [7], [9]). Since all molecules are equivalent in this structure, 13 P/MAS spectra obtained from crystalline powders (with proton decoupling) show, as expected, a single peak for each of the different carbon atoms of the α-dbp molecule. The resonances are in general shifted to lower ppm values relative to those obtained from TF solutions. The shift is largest for the following nuclei: (1) 3.6, (5) 2.6, (9) 2.2 and (10) 2.7 ppm. The chemical shift of (8) is unaltered in spectra obtained from the solid state and TF solution Molecular conformation: dependence on sample preparation α-dbp and its derivatives possess a certain degree of conformational flexibility facilitating the existence of a variety of solid state phases with different spectral and photochromic properties. Polymorphism has, for example, been observed for a variety of α-dbp derivatives [6, 7, 10]. The differences observed for the behaviour of aggregated microcrystalline material embedded in the two films discussed above is another example. Below, we discuss the MR spectra recorded for a variety of solid state phases of α-dbp produced by different procedures. MR assignments are facilitated in some of these phases and substantiate the assignments discussed above. We also demonstrate the existence and interconversion of different conformers of α-dbp in these materials. Figure 8 shows the 1 MAS spectra obtained from a supercooled (glassy) phase of DBP, immediately after the preparation (8(a)) and three hours later (8(b)) leaving the sample under continuous MAS conditions, at room temperature. Figure 8(c) shows the difference spectrum. The analysis of the spectra shown in figure 8 indicates the presence of two DBP isomers. Additionally, comparisons between the spectra shown in figure 8 and the spectra recorded from TF solutions (not shown here) yield a complete line assignment for the more stable species 1, which dominates the spectrum presented in figure 8(b). All the resonances of species 2 are shifted to higher ppm values. In particular, the benzylic proton resonances are found at 3.68 and 4.08 ppm for species 1 and 2, respectively. Based on this observation, we propose that the slightly different resonance frequencies (3.3 and 3.8 ppm) observed for the benzylic groups in films 1 and 2 are due to the predominance of different conformational isomers produced in the preparation of the two films. Different stable isomers were also observed for other samples 2115

10 T unes et al Figure 9. 1 spectra obtained with MAS at 10 kz, in T2 -weighted mode with τ = 1 ms (see section 2 for details), from: (a) commercial polycrystalline α-dbp at 290 K, after recrystallization from ethanol solution in the dark; (b) sample A recrystallized in TF under continuous UV (366 nm) irradiation; (c) sample B at 305 K; (d) sample B at 340 K. differently prepared, as indicated by the 1 spectra. Whereas different 1 MAS spectra were identified from commercial α-dbp (used as-received or recrystallized in TF), 13 P/MAS spectra run with 1 dipolar decoupling (not shown here), did not reveal significant changes. This observation indicates the existence of compounds with different hydrogen packing schemes in the solid state, which do not induce large changes of 13 environments. It is plausible that conformational isomers are involved, and that differences in their relative stability explain spectral changes as a function of preparation, history and purity of the samples. α-dbp crystals have also been reported to undergo a permanent colour change to yellow upon prolonged irradiation [12], and to lose their photochromic properties. This change was fully reversed by powdering the crystals, or by their recrystallization in the presence of natural light. In the following we present some experiments aimed at identifying the effect of light during sample preparation. Figure 9 shows 1 spectra of α-dbp, acquired in T2 -weighted mode (see section 2 for details), which was recrystallized under different conditions and at different temperatures as indicated. In the spectrum of α-dbp recrystallized from TF under UV irradiation (figure 9(b)) a new 1 signal at 4.9 ppm is clearly observed. This signal was also present in the spectrum of polycrystalline commercial α-dbp, recrystallized from ethanol under natural light (figure 7(a)), ruling out an assignment to a degradation product, reported to be formed after prolonged UV irradiation [12]. When the temperature is increased (figures 9(c), 9(d)) the signal at about 8 ppm, assigned to protons in the aromatic rings, sharpens, indicating an increase of librational amplitude and, consequently, the suppressing (in part) of spin diffusion, the dominant contribution for line broadening in solid state MR spectra, as already mentioned in the section 3.4. In addition, the signal at 4.9 ppm at 290 K is shifted to smaller ppm values at higher temperatures. This change is reversible. Subsequently, seven spectra (not shown here) were obtained in the temperature range 295 to 342 K; to prevent any variation of the temperature drift under MAS condition, the MAS rate (5000 ± 3 z) was kept constant during the data acquisition. The spectra were recorded in T2 -weighted mode, using a spin-echo RF pulse sequence with a time delay τ of 1 ms (see experimental section for details); the line width of the signal at 4.9 ppm, correlated with T2, did not show strong alterations with temperature. From the temperature dependence of intensity changes of the signal at 4.9 ppm, an activation energy of = 23.0 ± 4.2 kj mol 1 was determined. The spectral variation with temperature is tentatively assigned to a hydrogen atom exchange between groups for a molecular conformation in which the energy of the form is lowered. Therefore, the resonance at 4.9 ppm is assigned to the benzylic proton. This signal is absent in the spectra of selectively deuterated α-dbp or of α-dbp derivatives (figures 6 and 8). The possibility of two configurations having the dinitrophenyl group in cis or trans position with respect to the pyridyl nitrogen has been previously suggested [5]. We show that the polymeric host allows for a distribution of conformational isomers of α-dbp, a distribution that also explains the observed distribution of rate constants and thermodynamic parameters in polycrystalline or amorphous matrices such as PMMA [3]. The predominance of structural disorder of α-dbp in films 1 and 2, both before and after UV irradiation, is also indicated by the absence of wellresolved resonances in 13 P/MAS spectra of these samples. ne reason for the lack of full recovery of the α-dbp signal following photolysis may be the transformation to a different conformer, as indicated earlier. Another possible cause lies, however, in the photodegradation of these films upon prolonged photolysis when they become brittle and yellow. It appears that either α-dbp or its phototautomer is responsible for this degradation process. Studies of this reaction are in progress. 4. onclusions The study of the photochromism of α-dbp in crystals by MR would normally not be possible due to the low overall quantum yield, and to the large extinction coefficient of the photoproduct (less than 2% is converted to the blue excited form and the coloured layer is only a few microns thick). owever, information on the photochromic processes was obtained in the present work via 1 MR studies of α-dbp incorporated into PV film, at high concentration, in the form of microcrystals. The kinetics obtained for the back-reaction show the presence of long-lived species, with lifetimes 2.4 h and 6 d, whose concentration is strongly 2116

11 Spectroscopy studies of 2-(2,4-dinitrobenzyl) pyridine dependent on the experimental conditions used for the film preparation. The formation of hydrogen bonds in the solid state as well as the stabilization of different conformational isomers make spectral assignments difficult; the use of isotopic and chemical substitution as well as the use of various MR RF-pulse sequences is necessary. The coexistence of different conformational isomers in thermal equilibrium also explains the observed temperature dependence of the 1 signals. rystallization of α-dbp under permanent monochromatic irradiation (366 nm) produces differences in the MR spectrum, which are associated with the stabilization of specific conformational isomers of the molecule. Additionally, the preparation of PV/DBP films, which have a known distribution, and the regularity of α-dbp crystals, as monitored by MR as well as SEM methods, was demonstrated. Acknowledgments The authors are in debt to Professor Dietrich aarer, Professor Jean-Marie Lehn, Dr Gabriel Feio and to Dr Maria elena Gil for stimulating discussions and to Dr Zoe Pikramenou for the synthesis of the quaternary salt of α-dbp and for further support. This work was partially supported by ESPRIT project no (PRTIS) sponsored by the ommission of the European ommunities. References [1] Friedrich J and aarer D 1984 Angew. hem. Int. Ed. Engl [2] hichababin A E, Kuindshi B M and Benewolenskaja S W 1925 Ber. Deutsche hem.ges [3] Klemm E, Klemm D, Graness A and Kleinschmidt J 1980 J. Prakt. hem [4] Kimura, Sakata K and Takahashi 1993 J. Mol. Struct [5] Eichen Y, Lehn J-M, Scherl M, aarer D, asalegno R, orval A, Kuldova K and Trommsdorff P 1995 J. hem. Soc. hem. ommun. 713 [6] Eichen Y, Lehn J-M, Scherl M, aarer D, Fischer J, Deian A, orval A and Trommsdorff P 1995 Angew. hem. Int. Ed. Eng [7] orval A, Kuldova K, Eichen Y, Pikramenou Z, Lehn J-M and Trommsdorff P 1996 J. Phys. hem Scherl M, aarer D, Fischer J, Deian A, Lehn J-M and Eichen Y 1996 J. Phys. hem Khatib S, Botoshansky M, Peskin U, Scherl M, aarer D and Eichen Y 1997 J. Am. hem. Soc [8] Sixl and Warta R 1985 hem. Phys [9] Seff K and Trueblood K 1968 Acta rystallogr. B [10] Khatib S, Poplawski J M and Eichen Y unpublished results [11] A low concentration of byproducts having unpaired electrons was observed for similar irradiation conditions. The signal intensity did not change in time after irradiation. Eichen Y unpublished results. [12] ardwick R, Mosher S and Passalaigue P 1960 Trans. Faraday Soc