Structure and dynamics of hydrogen bonding guests in urea inclusion compounds Marina Brustolon,' Anna L. Maniero," Alessandro Marcomini" and Ulderico Segre* "Universita di Padova, Dipartimento di Chimica Fisica, Via Loredan 2, 35131 Padova, Italy bUniversita di Modena, Dipartimento di Chimica, Via Campi 183, 41100 Modena, Italy The radical obtained by y-irradiation of the 2-nonadecanone/urea (2-NDOU) and nonadecanoic acid/urea (NDAU) inclusion compounds have been studied by EPR spectroscopy. The spectra have first-order, fast-motion line shapes with anisotropic linewidths. For both compounds the spectra show the presence of two similar species. They originate from the two possible arrangements of the molecules inside the host channels, i.e. head-to-head or head-to-tail.The relative abundance of the two conformations is obtained from the intensities of their EPR signals and is accounted for in terms of the balance between guest-to- guest and guest-to-host hydrogen bonding. The transverse relaxation rate constants for the different hyperfine components have been obtained by computer simulation of the spectra. The relaxation originates from the librational motion of the p methylene group and from the hindered rotation of the radical inside the host channel. Evidence of pretransitional effects is shown by the spin relaxation rates above the order-disorder transition in NDAU. The urea inclusion compounds (UIC) are built up by urea molecules packed in an extended hydrogen-bonded arrange- ment which contains non-intersecting channe1s.l Inside these channels it is possible to accommodate molecules whose sizes are compatible with the cavity dimensions.Common examples are given by the normal alkanes or their derivatives, as ketones, alcohols, esters and fatty acids. At room temperature the symmetry of the urea host structure is described by the space group P6,22. The channels have a hexagonal symmetry axis and their walls are arranged in spirals.2 The UICs decompose at temperatures from 378 K (n= 10) to 409 K (n=30).3It should be noted that the long- chain UICs are stable above the melting point of pure urea, T, =405.8 K. On lowering the temperature the UICs undergo a phase transition to an orthorhombic phase. The longer the chain of the included molecule, the higher the transition temperature T,.For the alkane CnH2n+2 UIC, T, ranges from 110K (n=10) to 160K (n=20) and to 220K (i1=40).~ The latter transition has been modelled as an order-disorder phase transition induced by translation-rotation coupling between the orientational order of the included molecules and the transverse acoustic phonons of the urea host latti~e.~ Fatty acids are known to be included as dimers inside the urea channels because of their ability to form hydrogen bonds between themselves.6 As a consequence, their transition to the orthorhombic phase occurs at a temperature much higher than that of the corresponding alkane. As an example, T,= 167 K [by differential thermal analysis (DTA)] for the octadecane UIC and T,=239 K for the octadecanoic acid UIC.7 Moreover, the mobility of the fatty acid guests does not change sharply at the transition, but it is found to occur in a temperature range of some tens of In the UICs the host matrix and the guest molecules form two inter-penetrating ordered structures." It has been ques- tioned whether the repetition length of the guest structure along the channels cg is related to the pitch of the host helix ch.Different behaviours have been revealed according to the nature of the guest. In the case of the alkanes cg and ch are in general incommensurate, while for undecan-5-one it has been found that 2cg =3ch.11 This difference should be related to the modes of interaction of the latter guest compounds.The ketone molecules interact via long-range dipolar forces and can undergo hydrogen bonding with the urea molecules of the host matrix. Recently, the presence of an extended hydrogen-bonded array connecting hosts and guests has been reported for undecane-2,lO-dione UIC.I2 The role of hydrogen bonding is basic in accounting for the properties of host-guest assemblies as the UICs. It is therefore worth gaining a deeper understanding of the structure and the dynamics of the hydrogen-bonding guests in the urea channels. Magnetic resonance techniques are useful tools in studying both structure and dynamics in the solid state. Studies of single crystals give more detailed information with respect to dis- ordered samples. UIC crystals are often quite tiny, which makes the use of EPR advantageous because of its higher sensitivity than NMR.13 In this paper we report on our studies of hydrogen-bonding guests in urea by means of continuous wave (CW) EPR spectroscopy.In a forthcoming paper we will present the results obtained by pulsed EPR techniques. Several species of radicals have been obtained by irradiation of the UICs with hydrocarbon derivatives such as ketones, esters, ethers and carboxylic acids.14 In all these cases it is found that the only stable species is the one obtained by removal of one of the protons in the a position to the C=O group. Therefore, the main feature of the EPR spectrum is given by a hyperfine (hf) pattern of eight lines due to the coupling of the electron spin with one a and two p protons.It is found also that the lineshape is affected strongly by the temperature T and by the angle x between the magnetic field B, and the c6 crystal axis.15 As a consequence, at some particular values of x and T some lines can overlap. In this case the values of the hf splitting must be obtained by means of numerical spectral simulation. The small hf couplings with the y and 6 protons are observable only in the case of narrow lines, and their values can be measured accurately only by the ENDOR technique.16 Generally, in the high-temperature phase, the EPR lineshape can be simulated as a first-order spectrum, i.e. as a sum of a finite number of lines with different widths. The EPR spectra offer a set of data, the hf splittings and the linewidths, which can be exploited to obtain insight into the structural and dynamical properties of the guest radicals.In our previous works on the nonadecan-10-one UIC (10- NDOU),'5*'6 we have shown that the carbon chain largely deviates from the planar all-trans configuration, and that two kinds of motions affect the spin-relaxation properties of the radicals, the internal motions of the methylene chain and the uniaxial molecular rotation inside the host channels. In this paper we extend our investigation to non-symmetrical guest molecules composed of the same number of carbon atoms, it. nonadecan-2-one and nonadecanoic acid. Unsymmetrical guest J. Mater. Chem., 1996, 6(lo), 1723-1729 1723 molecules can assume two different arrangements inside the channels, head-to-head and head-to-tail (see Fig 1) It will be shown that both structures are present and that the dynamics of the guest radicals are remarkably different in the two cases Experimental We have studied by EPR spectroscopy the long chain radicals obtained from nonadecan-2-one and nonadecanoic acid in urea (2-NDOU and NDAU, respectively) Single crystals of 2-NDOU and NDAU were grown from a 1 1 methanol-acetone (spectrophotometric grade) solution of the urea and the guest in a 60 1 molar ratio The crystals were obtained either by slow evaporation at room temperature or by slow cooling (0 01 "C min-' from 50 to 20 "C) l7 The paramagnetic probes were obtained by y-irradiation with a dose of 2 Mrad at room temperature, giving rise to the stable nradicals I and I1 0 88It I CH~-C-C-C-(CB~)~&HJII HaHP' II The EPR spectra were obtained using an X-band CW Bruker ER 200 D spectrometer interfaced with a Bruker data system ESP 1600 and equipped with a Bruker variable-temperature unit The spectra were recorded in the range 290-170K and by varying the orientation x of the magnetic field with respect to the long prism axis of the hexagonal crystals coincident with the urea channel axis and which will be henceforth indicated as the 2 axis Results EPR spectra of irradiated UIC The first-order EPR transitions of an electronic spin doublet state are uniquely specified by the set of the spin quantum numbers {M}=MI, M2 M, of the nuclei which are coupled via the hf interaction to the electronic spin The transitions are located at any of the resonant values of the magnetic field Bres {M} =Bo+ 1ajMJ (1) J=1 fl where a, are the hf coupling constants (hfcc) When the I--I PESr B 0 <OH Fig.1 Possible arrangements of unsymmetrical guests within the host channels top, head-to-head, bottom, head-to-tail 1724 J Muter Chem, 1996, 6(10), 1723-1729 dynamics of the paramagnetic molecule are rapid, the Redfield- Freed theory applies," l9 and the EPR transitions are predicted to have a Lorentzian shape The linewidth of a transition is the reciprocal of the transverse spin relaxation time T2and is written as a polynomial expansion in the quantum numbers (MI 1 -=A+ 1BkMk-t CCkM2-t 1Ekk MkMk (2)T, k k k<k where the sums are over all the coupled nuclei The coefficients are related to the correlation functions of the spin interactions which are modulated by the molecular motions For nuclei with 1=1/2, the M2 terms cannot be separated from the constant term A in eqn (2), which therefore reduces to 1 -=A'+ 1BkMk-k 1Ekk MkMk (3)T2 k k<k When the number of nuclear spins is n= 3 the EPR spectrum consists of eight lines whose widths are expressed in terms of seven coefficients by using eqn (3) Therefore, it is statistically more significant to obtain from the spectrum, by a fitting procedure, the width coefficients A', Bk and Ekk instead of the widths of all the individual lines The above theory cannot be applied when the frequencies of the molecular motions are slow with respect to the amplitude of the fluctuations of the magnetic interactions The EPR lineshape is no longer given by a simple superposition of Lorentzian lines and the proper slow-motion theory should be used2' In the low-temperature orthorhombic phase it is expected that the molecular rotation inside the urea channels is in the slow-motion regime Our study, therefore, was restric- ted to the hexagonal phase The EPR spectra of the included radicals were simulated by a least-squares fitting procedure l5 21 The adjustable param- eters are the centre of the spectrum, the hf splittings and the linewidth coefficients for the a, p and p' protons By this procedure, one obtains from the EPR spectrum a set of data regarding the structure of the radicals and their dynamical properties The results are given in the following sections Hyperfine couplings The EPR spectrum of 2-NDOU at T=290K and with B, parallel to the hexagonal C6 axis (x=O") is shown in Fig 2 together with its computer simulation The overall pattern is given by a triplet of doublets as a consequence of the accidental degeneracy of the principal hf couplings The splittings due to the methyl protons are also partially resolved However, some I II I I 3400 3440 3480 3520 3560 B1G Fig.2 Expenmental and simulated EPR spectrum of 2-NDOU at T=290K and x=O" spectral features (such as the line at very low field) cannot be accounted for by a unique set of hf interactions.We therefore hypothesised that the spectrum is due to the superposition of two different species, A and B. Their contributions may be disentangled by the following procedure. First, the lineshape of species A is computed by simulating the main features of the spectrum. Then, the computed contribution of A is sub- tracted from the spectrum to obtain the ‘experimental’ spec- trum of species B. The latter is simulated in its turn and the computed contribution of B is subtracted from the original spectrum to give a clearer ‘experimental’ lineshape for the species A. This procedure is iterated until reaching convergence. In Fig. 3 we display the computed contributions of species A and B to the total lineshape.The simulation in Fig. 2 is given by 88% A plus 12% B. The EPR spectrum of NDAU oriented with x =0 O at room temperature is displayed in Fig. 4. The overall shape of the spectrum is quite similar to that of 2-NDOU(B). In this case the presence of a second species is also apparent, which gives a minor contribution to the spectrum. However, the individual spectra cannot be resolved because of the very low intensity of the signal due to the second species. We have observed the EPR spectra of both 2-NDOU and NDAU at different orientations of the magnetic field. In the case of the ketone the contributions of the two species have been separated. The angular variations of the hf couplings constants are reported in Fig.5, together with those of 10-NDOU which were measured previously.I6 The splitting which exhibits the larger angular variation can be attributed to the coupling with the a proton, while the others originate from coupling to the p and p’protons. The rotational motion of the I 1 I O-NDOU(B) I 1 I 3-rvO 3440 3480 3520 3560 BIG Fig. 3 Computer-simulated spectra of 2-NDOU(A) and 2-NDOU(B), together with the ‘experimental’ spectra obtained by the subtraction procedure described in the text 3280 3320 3360 3400 3440 BIG Fig. 4 Experimental and simulated EPR spectra of NDAU at T= 290 K and x=O” 2-NDOU(A) 2-NDOU(B)l---l 30 10 -Q I I I 1 NDAU 10 -0306090 0 30 60 90 Xldegrees Fig.5 Angular dependence of the hf splittings for 2-NDOU, 10-NDOU and NDAU at T=290 K (0,a proton; 0, P proton; +, p’ proton) radical along its long axis is fast enough to average the anisotropic magnetic interactions which, therefore, are axially symmetrical along the Z axis.The principal values of the averaged tensors Aiand g are reported in Table 1. Fig. 5 also shows the computed angular dependence of the hf coupling Table 1 Principal components of the averaged hf tensors (in G) and g tensor at T= 290 K 10-NDOU 28.8 13.9 27.0 25.9 20.8 19.7 2.0037 2.0043 2-NDOU(A) 28.8 15.0 27.7 26.5 18.9 18.0 2.0037 2.0041 2-NDOU(B) 27.1 15.2 29.6 29.3 27.5 27.8 2.0043 2.0035 NDAU 29.3 15.4 35.5 35.4 28.9 29.0 2.0042 2.0029 J. Muter. Chem., 1996,6(lo), 1723-1729 1725 constants according to the theoretical expression 22 U’(X)=X,,~cos2 x+X12 sin2x (4) The temperature dependence of the principal value 2, in the range 200-290 K is displayed in Fig 6 for species A and B of 2-NDOU and for NDAU and 10-NDOU In the case of the ketone radicals the a proton Al,is nearly temperature independent, while both the pproton components are strongly affected by the temperature variation A similar trend was found for the parallel principal values of the hf coupling tensors On the other hand, in the case of NDAU, the components of the pproton hf tensors display a modest variation, while those of the a proton hf tensors display a modest variation, while those of the a proton increase slightly on lowering the temperature The latter behaviour is evidenced in Fig 3,where the temperature variation of the isotropic hfcc a= TrA of the a proton of NDAU is displayed with a finer T grain and an enlarged scale The hfcc value varies significantly between 210 and 240K, while it is quite constant outside this range The following features are worth noting in Fig 5 and 6 11111 40 r-l2-NDOU(A) 2-N DOU(6) -*o*** e-+++++++ -m i3 d :0 0 ~0000~ ++1 IIIIIQ10*40 11111’ 30 -*-e* **** 10-NDOU -20 +++++++++ + ~~00000000 lIllA10 I 200 240 280 200 240 280 TIK Fig.6 Temperature dependence of the 2, hf coupling principal values for 2-NDOU, 10-NDOU and NDAU measured at x=90° (0,a proton, 0, p proton, +,p proton) 2 20.5 200 250 300 TIK Fig.7 Temperature variation of the %-proton hyperfine coupling constant of NDAU 1726 J Muter Chern, 1996, 6(10), 1723-1729 (I) The hf couplings of the fl protons are inequivalent in all the four cases, but A@)and A@’)are more similar for NDAU and 2-NDOU(B) (ii) The anisotropy of the hf couplings of the pprotons is greater for 2-NDOU(A) and 10-NDOU (in) On the whole, the temperature and angular dependences of the hf splittings of 2-NDOU(A) are similar to those of 10-NDOU, while the variations of the hf splittings of 2-NDOU( B) look like those of NDAU Linewidth coefficients The spin-relaxation behaviour of radical probes in the solid state gives relevant information about the different types of motions affecting its magnetic interactions l3 In our previous study on 10-NDOU” we have shown that, in the fast-motion regime, the linewidth coefficient B, is especially affected by the molecular rotation around the long axis, which modulates the anisotropic hyperfine interaction of the a proton The value of the E,,, linewidth coefficient, instead, is related to the oscil- lation of the methylene group about its equilibrium position, because of the conformational dependence of the isotropic coupling constant of the pprotons 22 If the effects of these dynamical processes are considered to be uncoupled, it is possible to obtain simple expressions for the linewidth coefficients The value of the B, coefficient is maximum when the magnetic field is oriented at x=90° with respect to the channel axis The reorientation of the probe modulates the anisotropic interactions A, and g about their average values A, and El and the linewidth coefficient is given by 1 B, =4(P€3B0/fi2kxx-gyy)(4t xx -A, yy)G (5)0 where z, is the rotational correlation time In contrast, the linewidth coefficient E,, is expected to be angular independent The oscillation of the methylene groups modulates the isotropic coupling constants a, and a, about their averages a, and afl and the following equation is obtained E,, =2<(a,-<)(q3 --aa))tc (6) where T~ is the correlation time of the conformational motion On the grounds of the previous analysis, we have focused our attention on this pair of linewidth coefficients In the case of 2-NDOU, the A species only was considered In fact, the B species has a rather low intensity and the values of its linewidth coefficients obtained by the non-linear fitting procedure suffer from a large incertitude Therefore, it seems that the linewidth coefficients of the A species only are reliable to obtain infor- mation on the dynamical processes of the nonadecan-2-one UIC We recall that, as discussed previously, the above eqn (5) and (6) are valid as long as the motions are fast with respect to the interaction anisotropies The temperature dependences of B, and E,, for the A species of the 2-NDOU radical measured with the magnetic field perpendicular to the 2 axis are displayed in Fig 8 The data for NDAU are shown in Fig 9 The linewidth coefficient values increase on lowering the temperature, but in general they do not follow a simple trend If one assumes that the temperature dependence of a motional rate is due to an activated process, l/z =(l/zo) exp (-d/kT) (7) then a linewidth coefficient w(T)(w =B,,E,, should be fitted to the expression w(T)=woexp (d/kT) (8) In the case of the ketone radical, the E,, coefficient fits to a single activated process in the whole temperature range studied, while the B, coefficient increases up to a limiting value In the case of the carboxylic acid, instead, it is manifest that two 1 Q5 0.1 I I I I I 3 3.5 4 4.5 5 103 WT Fig.8 Temperature dependence of the linewidth coefficients B, and Epp, of2-NDOU I. I J I Qh 0.1 0.01 3.4 3.0 4.2 4.6 103 KIT Fig.9 Temperature dependence of the linewidth coefficients B, and ED,, of NDAU Table 2 Activation parameters for the linewidth coefficients of the UIC radicals 10-NDOU" 9.4 x lo4 7.5 -4.4 x 103 13 2-NDOU(A) 2.1 x lo3 18 -4.9 x lo2 19 NDAUb 1.3 x 10-5 47 -6.2~ 50 NDAU' 1.3 x 104 8.7 "Ref. 15. bT=216-231 K. 'T=238-290 K. different motional regimes are present. However, in the high- temperature region the linewidth contributions due to the motions are so small that it is not possible to obtain a reliable estimate of the E,,, coefficient. Both the coefficients can be fitted to eqn. (8) in the low-temperature region. We note that this region corresponds to the temperature interval in which the aproton hfcc shows the large variation displayed in Fig.7. The activation parameters wo and A corresponding to the straight lines in Fig. 8 and 9 are reported in Table 2. Discussion Structure By comparing the results obtained for the p protons in the different cases one notes two types of behaviour. In the cases of 10-NDOU and 2-NDOU(A),the hf coupling constants have very different values, they have a marked temperature depen- dence, and the dipolar part of the hf tensor shows an anisotropy of about 2G by rotating the crystal in the XZ plane. In the second case, 2-NDOU(B) and NDAU, the two hf coupling constants have more similar values, about 30G, they have a slight temperature dependence and they are nearly isotropic.These trends can be explained by the following consider- ations. The isotropic hf coupling of the p protons depends on the dihedral angle 8, between the z orbital of the unpaired electron and the CH, bond direction according to the relation,22 ui=acos2 el (i= i,2) (9) where a typical value for the proportionality constant is ca. 43 G. In the case of a planar conformation for the carbon chain, the values of the dihedral angles are 81.2 = 30 O and the hf coupling constants are equal. The non-equivalence of the p protons indicates that the local average conformation is distorted and the two angles are given by =y 300 (10) In the cases of NDAU and 2-NDOU(B) the distortion is small and displays a small temperature dependence. The value of the distortion angle can be obtained from the ratio ul/u2and it is given by y z3 O.On the other hand, in the cases of 10-NDOU and 2-NDOU(A) the values of a, and u2 cannot be obtained by using eqn. (lo), which takes into account a static distortion. As the large temperature variations of a, and u2 suggests, they are average values due to restricted rotation of the methylene group around the C,-C, b~nd.'~,'~It can be assumed that the dihedral angles vary between the two extrema 8, +a. The size of the average distortion angle y can be obtained by plotting a, us. u2 in the temperature range explored. In fact, it has been shown that the following relation hold^:'^,'^ a1 =ma, + 21 Gi? c 1-m1 (11) where cos(60"+27) f(y) = cos(60"-27) The values of the distortion angles have been obtained by a least-squares fit and they are y =25 O and y =42 O for 10-NDOU and 2-NDOU(A), respectively.The values of the fluctuation amplitudes at room temperature are a=4O0 for the former and a =50 O for the latter. From the above discussion we may conclude that the two species detected in the EPR spectra of 2-NDOU and NDAU are related to the two allowed arrangements of the guest molecules inside the channels (see Fig. 1). Species B of 2-NDOU and nearly all the molecules of NDAU are arranged head-to-head. In this case, the molecular pair forms a tight unit because of intra-pair hydrogen bonding. The conformation of the chain in the proximity of the molecular head, as tested by the p proton hf coupling, is quite rigid and nearly planar.In the case of the ketone this arrangement can give rise to a weak hydrogen bond between the CH3 protons of one molecule and the C=O group of the next one; on the other hand, the opposite arrangement allows the carbonyl to form stronger hydrogen bonds with the NH, groups of the urea molecules. In the NDAU case the head-to-head arrangement allows the formation of strong hydrogen bonds between the two facing COOH groups, and therefore this arrangement prevails neatly. Species A of 2-NDOU (and presumably the barely detectable second species of NDAU) are disposed in the head-to-tail arrangement. In this case, the molecular chain undergoes large oscillatory motions. If the oscillation was symmetrical about the planar zigzag conformation, the two hf coupling constants should be equal.The difference between their values indicates that the chirality of the host channel induces a chiral distortion J. Muter. Chem., 1996, 6(lo), 1723-1729 1727 in the guest molecule. It is well known that the presence of a chiral carbon atom in a radical induces the j3 protons to be inequivalent because of the different energies associated with two specular conformation^.^^ Therefore we can conclude that the formation of hydrogen bonds between the urea molecules and the guest molecule promotes a helix arrangement of the guest chain. Molecular dynamics calculations are in progress to check these conclusions.24 The temperature dependence of the proton hf couplings of the included radicals is a sensitive detector of the hexagonal to orthorhombic phase transition.In the case of the ketone radicals, the p proton hf coupling constants are discontinuous at ca. 160 K for 10-NDOU and at ca. 185 K for 2-NDOU. This variation should be related to the changes occurring in the host matrix structure, which affects the amplitude of the distortion of the CH2 group. Therefore, the above temperatures should correspond to the phase transitions for the two UICs. On the other hand, in the case of NDAU the p proton hf couplings have a small, continuous variation with temperature, while the a proton hf coupling has a significant variation in the range 240-210 K, with a maximum slope at 213+3 K, as displayed in Fig.7. We suggest that the last value corresponds to the transition temperature for NDAU. The higher value of T, for the acid UIC with respect to the ketone UICs is a confirmation of the dimerization of the guest molecules in the former case. Also the difference between the two ketone UIC transition temperatures should be related to the possibility for 2-NDOU of arranging head-to-head inside the urea channels, originating a loose dimer via hydrogen bonding. Since this disposition is adopted by a fraction of the guest molecules in the case of 2-NDOU, and by nearly all the molecules for NDAU, the transition temperature for 2-NDOU is intermedi- ate between those of 10-NDOU and NDAU. The value of the isotropic hfcc a, gives insight into the conformation of the CH, with respect to the carbonyl group.In fact, the Fermi contact coupling of the a proton is pro- portional to unpaired spin density on the carbon atom pa,22 If CH, and CO lie on the same plane, the spin density is partially delocalized over the carbonyl and pa is reduced. Therefore, the lower the value of a,, the greater the deviation of the CH,CO group from planarity. For NDAU, this deviation increases in the range 230-210 K and the carboxylic groups are expected to be more distorted in the low-temperature orthorhombic phase than in the hexagonal phase. Dynamics The activation energy values measured for the linewidth coefficients B, and E,,, have been reported in Table2. They are related to the energy barrier A, hindering the molecular rotation, and d, hindering the intramolecular conformational motion.By comparing the values for the two ketone UICs it is worth noting that the values of A, are quite comparable, while A, differs by more than a factor of two. The two radicals have equal lengths and similar interactions with the host matrix. Therefore, the difference should be related to the position of the radical centre (which is adjacent to the carbonyl group) within the molecule. The CH, group of 2-NDOU(A) can explore a larger set of orientations because of the higher probability of gauche defects at the end of the chain. In fact, there is much evidence that the motions of long alkyl chains are quite different in the middle of the chain and at its extremes.In particular, NMR,25 Raman26 and molecular dynami~s~~,~~ results have shown that in n-alkane UICs the gauche defects may be localized only at the ends of :he chains. The molecular rotation is therefore driven by the same conformational interconversion, and the energy barriers hindering the two motions are therefore very 1728 J. Muter. Chew., 1996, 6(lo), 1723-1729 similar. Further evidence of a larger freedom of motions for 2- NDOU is obtained by electron spin-echo experiments which show a dramatic reduction in the electron spin phase memory time in this case with respect to 10-NDOU.29 The values of the correlation times for the rotational and conformational motions can be obtained from the linewidth coefficients B, and E,,,, respectively, when the anisotropies of the relevant magnetic interactions are known.As long as the rotational motion is concerned, we have no direct knowledge of the x and y components of the A, and g tensors, which must be inserted into eqn. (5). However, we can obtain a reliable estimate of their values. It is known that the x, y, z components of the a proton hf tensor in a CH fragment have the relative values 1: 2 : 3.22 Therefore, from the values of A,,= A,, and Al = (Axx+ A,,)/2 in Table 1 we obtain A,, M 10 G and A,, ~20G. In an analogous manner, from the value g, = (gXx+g,,)/2in Table 1 and by assuming that component along the IT orbital is equal to the free electron value, g,,=2.0023, we have g,, M 2.006.In conclusion, the anisotropies A,, -A,, and gxx-gyy for 2-NDOU are very similar to those of 10- NDOU we reported previously,15 as is to be expected because of the similarity of the paramagnetic probes. Therefore, the ratio between the rotational correlation times for the two ketones is nearly equal to the ratio between the values of the linewid th coefficients Z, ( 1 0-NDOU) B, ( 10-NDOU)N q(2-NDOU) -Ba(2-NDOU) At T=240 K we obtained z, (10-NDOU)=0.85 ns,” and we have, therefore, q(2-NDOU) = 3.5 ns. This value corresponds to a rotational frequency which is of the order of magnitude of the interaction anisotropies that are averaged by the rotational motion, i.e. (A,, -Ayy)/h and ,uBBO(gxx-g,,)/h. Therefore, the Redfield-Freed theory cannot be applied at temperatures below 240 K, as has been discussed previously, and the anomalous behaviour of the B, coefficient of 2-NDOU(A) is explained.The dynamics of NDAU are more complex. In this case the guest molecules in the channel are oriented with their functional groups facing each other, as discussed above. The energy barrier hindering the molecular rotation, A,, is about equal to that of 10-NDOU, in the high-temperature region. At T M 230 K, i.e. about 20 K above the hexagonal to orthorhom- bic phase transition occurring at T,= 213 K, the variation of the rotational rate becomes steeper, and it can be described as an activated process in the temperature range down to T,. The energy barrier for the conformational motion also is very large and quite similar to that for the molecular rotation.In the same temperature range the hfcc a, increases by about 5%. The slowing of the motions in this range is related to the distortion of the dimeric unit formed by the carboxylic groups, which is induced by the phase transition to the low-temperature ordered phase. A similar trend has been found previously by monitoring the Raman spectra of the stearic acid UIC.8 The transition temperature for this compound is ca. 234 K and the bandwidth of the CH2 scissoring mode displays a very large variation in the range 220-250 K, while it varies slowly outside this range. Similar results have been recovered for icosan-1-01 UIC and icosanoic acid UIC.’ We conclude, therefore, that the onset of orientational order at the hexagonal to orthorhom- bic phase transition occurs together with the pretransitional effects which are monitored by the progressive distortion of the carboxylic groups.Conclusion The EPR spectra of the radicals produced by y-irradiation of unsymmetrical molecules included into the urea channels show that head-to-head and head-to-tail arrangements are possible. While the former is nearly totally dominant in the case of NDAU, both situations are present in the case of 2-NDOU. The EPR signal of head-to-tail arranged 2-NDOU is far more intense than that of the head-to-head case. This result should be compared with those obtained by solid-state NMR on 6 7 8 F. Laves, N. Nicolaides and K.C. Peng, Z. KristaIlogr., 1965, 121,258. Y. Chatani, H. Anraku and Y. Taki, Mol. Cryst. Liq. Cryst., 1978, 48,219. H. L. Casal, H. G. Camerun, E. C. Kelusky and A. P. Tulloch, J. Chem. Phys., 1984,81,4322. shorter alkanone guests (decan-2-one7 undecan-2-one and dodecan-2-one) in UIC.30This NMR study suggests that the guests are arranged randomly within the channels. Two poss- ible explanations can be put forward: (a)the degree of ordering of parallel or antiparallel dispositions is a function of chain length, so that parallel ordering is preferred for longer guest 9 10 11 12 H. L. Casal, J. Phys. Chem., 1985,89,4799. K. D. M. Harris, S. P. Smart and M. D. Hollingsworth, J. Chem. SOC., Faraday Trans., 1991,87,3423. M. D. Hollingsworth and C. R. Goss, Mol. Cryst.Liq. Cryst., 1992, 219,43. M. E. Brown and M. D. Hollingsworth, Nature (London), 1995, 376,323. ketones; (b) the radical yield for irradiated ketone/urea inclusion compounds is lower when the ketone molecules are disposed head-to-head. These two hypothetical mechanisms obviously are not alternative, but they could be both operative in determining the relative spectral intensities of parallel and 13 14 15 16 M. Brustolon and U. Segre, J. Chim. Phys., 1994,91,1820. 0.H. Griffith, J. Chem. Phys., 1964,41,1093; 1965,42,2644. U. Segre, M. Brustolon, A. L. Maniero and F. Bonon, J. Phys. Chem., 1993,97,2904. F. Bonon, M. Brustolon, A. L. Maniero and U. Segre, Appl. Magn. Reson., 1992,3, 779. antiparallel arrangements. More experiments, and comparison between the results obtained by means of different techniques on the same inclusion compound will be necessary to clarify this important point.17 18 19 20 M. S. Greenfield, R. L. Vold and R. R. Vold, J. Chem. Phys., 1985, 83, 1440, A. G. Redfield, Adv. Magn. Reson., 1966, 1, 1. J. H. Freed and G. K. Fraenkel, J. Chem. Phys., 1963,39,326. J. H. Freed, in Spin Labeling: Theory and Applications, ed. L. J. Berliner, Academic Press, 1976,vol. 1. This work has been supported by CSSMRE, C.N.R., Padova, and by MURST, Rome. Dr. P. G. Fuochi of FRAE, C.N.R., Bologna (Italy) is acknowledged for irradiating the compounds. We are grateful to Professor M. D. Hollingsworth for informing us of his results before publication. 21 22 23 24 25 M. Barzaghi and M. Simonetta, J. Magn. Reson., 1983,51, 175. N. M. Atherton, Electron Spin Resonance, Ellis Horwood, Chichester, 1973. P. Smith, W. H. Donovan, C. E. Mader, L. M. Dominguez and W. T. Koscielniak, Magn. Reson. Chem., 1995,33, 395. P. G. De Benedetti, C. Menziani and U. Segre, work in progress. F. Imashiro, D. Kuwahara, T. Nakai and T. Terao, J. Chem. Phys., 1989,90,3356. References 26 A. El Baghdadi and F. Guillaume, J. Raman Spectrosc., 1995, 26, 155. 1 K. D. M. Harris, J. Solid State Chem., 1993,106, 83. 2 A. E. Smith, Acta Crystallogr., 1952,5,224. 27 K. J. Lee, W. L. Mattice and R. G. Snyder, J. Chem. Phys., 1992, 96,9138. 3 J. Ahmad, A. J. Freeston and A. Hussein, J. Inclusion Phenom. Mol. Recognit. Chem., 1994,18,115. 4 K. Eukao, T. Horiuchi, S. Taki and K. Matsushige, Mol. Cryst. 28 29 30 A. R. George and K. D. M. Harris, J.Muter. Chem., 1994,4,1731. M. Brustolon, unpublished results. M. D. Hollingsworth, personal communication. Liq. Cryst. B, 1990,180,405. 5 R. M. Lynden-Bell, Mol. Phys., 1993,79, 313. Paper 5/05021I; Received 28th July, 1995 J. Muter. Chern., 1996, 6(lo), 1723-1729 1729