Determination of Relative Intramolecular Configuration by Nuclear Quadrupole Double Resonance BY NORBERT AND ALARICH WEIDEN WEISS Institut fur Physikalische Chemie Physikalische Chemie 111 Technische Hochschule Petersenstr. 20 D-6100 Darmstadt W. Germany Received 31st July 1978 The influence of a nuclear quadrupole spin system B on the spin echo of a nuclear quadrupole system A is governed by a l/r6 law. This relation was used to determine relative intramolecular arrangements within molecules and complex ions in solids. By heteronuclear double resonance 81Br++27Al in KA12Br7 the bridging bromine has been determined and also the Br frequencies have been assigned to certain Br-A1 bonds. Homonuclear double resonance 35Cl+-+35CIwas applied to molecular compounds of C13CC02H and the close-lying frequencies of two molecules in the asymmetric unit were split-up into the appropriate groups.Finally from isotopic 81Br t-)79Br double resonance in 3,5-dichlorophenoxy aluminiumdibromide and 4-chlorophenoxy aluminiumdi- bromide a structure is proposed for these molecules. The aluminium atoms in these dimeric compounds are connected by bridging oxygen atoms whereas the bromine atoms are in terminal positions. The use of nuclear double resonance (n.d.r.) methods applied to solids is focused on the problem of increasing the signal-to-noise ratio. By these techniques many nuclei may be studied which are inaccessible to the single resonance methods because of their low resonance frequencies and/or low abundance in the solid considered.Apart from the effort to cover as many different nuclei as possible n.d.r. offers a means of assigning the different resonance frequencies found to atoms in certain positions within a chemical unit (molecule complex ion etc.) in the solid. The prob- lem of assignment geometrical positions f-) resonance frequency is usually done by involved single crystal studies. In cases of unknown crystal structure they do not lead to a unique solution of the problem. Although an n.d.r. experiment is not a unique solution of the difficulties it is valuable in both applications mentioned above. In the following the technique proposed by Emshwiller Hahn and Kaplan' (EHK) (nuclear quadrupole-nuclear quadrupole double resonance) is applied to three different groups of chemical problems by taking advantage of the l/r6-dependence of echo amplitude damping (Y = distance of the two nuclei of interest within the solid).The first problem is the relative assignment of 81Br and 27Al quadrupole resonance frequencies within the complex ion [AI,Br,]- of the ionic solid KA12Br,. Both the 1/r6-dependence and the improvement of the signal-to-noise ratio are illustrated by application of the EHK technique to "Br +-w 27AI.2 Secondly the use of " homonuclear " double resonance ++35Cl,tested on a few molecular compounds of trichloroacetic acid (TCA) with organic molecules is described and the consequences of these experiments for relative intramolecular assignment discussed. Here n.d.r. allows the assignment of n.q.r. signals with very little frequency separation to different groups e.g.-CCI,. Finally we study how to investigate the molecular structure by using the technique as " isotopic " n.d.r. 'lBr ++ 79Br e.g. on the samples of (3,5-dichlorophenoxy aluminiumdibromide)2,(3,5-CI,C,H,0AlBr2)2 and of (4-chlorophenoxyaluminium dibr0mide)~,(4-ClC~H~OAIBr~)~. NUCLEAR QUADRUPOLE DOUBLE RESONANCE EXPERIMENTAL In applying the EHK method a pulse-pulse double resonance spectrometer was used the block diagram of which is shown in fig. I together with the applied pulse programs. In the detection channel the A-channel the power stage is a I kW wide band amplifier (1-200 MHz). The final stage in the second channel the B-channel is a 700 W amplifier with a band width of 1-30 MHz which was replaced by an amplifier covering the range of 37-42 MHz for the 35Cl f-) 35Cl n.d.r.experiments. In the isotopic n.d.r. ''Br f-) 79Br only the A-channel was used serving as the detection channel as well as the search channel. In this case the complete pulse program was fed to the sample through one probe coil whereby the 180" pulse for the B-system was following 'A'' -CHANNEL J. -BROAD RE-90" GATE ~BAND-~PROBE-)r AMPL ft t -\----+-I /+l 90" 180" FIG.1.-Block diagram and pulse programs of the nuclear quadrupole-nuclear quadrupole double resonance spectrometer. Dashed line shows B-pulse in single channel operation. immediately after the 180" pulse for the A-system (see fig. 1). The pulse lengths were individ- ually adjusted according to the different requirements of the experiments.For the "Br 27Al experiment on KAI 2Br7 the 90" A-pulse was 6-10 ps and the 180" B-pulse was about 200 ps in order to prevent line broadening due to the Fourier components. z = 300 ps was chosen for the 9Oo-.r-180" sequence. In the homonuclear 35C1-35Cl experiment the 90" pulse width was 20-40 ps (A-channel) whereas the B- pulse was extended to -250 ps; T was around I ms. In the n.d.r. experiments on (33-C12CsH30AlBr2)2 and (4-CIC6H,OAIBr,) a small permanent magnet was placed near the probe to create a weak magnetic field at the site of the nuclei. The Zeeman splitting leads to a decoupling of the nuclear spins and thereby to a lowering of the mutual spin flipping rate. Thereby the effective transversal relaxation time T2is raised and can be extended up to 400~s.Again the width of the A-pulse was lops(90"). A B-pulse of -250 ps was adequate. The samples were synthesized according to the prescriptions given in the literature. KAI2Br, prepared from the melt of 1 KBr + 2AIBr3 was zone refined and handled in closed ampoules because of its high sensitivity to moist air. The molecular compounds (C13- CC02H),X can also be prepared from a melt of the individual compound^.^ Some of the N. WEIDEN AND A. WEISS 95 TCA compounds are difficult to crystallize since they form glasses. For example the com- pound TCA-vanillin melts incongruently and has to be crystallized very slowly (12 h at 18 "0. TCANa.3H20 crystallizes from an aqueous solution of TCA neutralized with Na2C03.(3,5-ClzC6H30A1Brz)z are prepared at elevated tempera- and (4-CIC6H40AlBrz)2 tures by reaction of the respective phenoles and AlBr dissolved in CS2.4 THEORETICAL The nuclear quadrupole-nuclear quadrupole double resonance experiments reported here are based on the theory developed by Emshwiller Hahn and Kaplan.' By a conventional spin echo experiment an easily observable spin system A is moni- tored. At a time z a second spin system B is irradiated by a 180" pulse whereby the local magnetic field due to the B-spins at the sites of the A-spins hAB -pB/riB changes its sign hAB -+ -hAB. The sign reversal disturbs the phase memory of the A-spins and changes the A-echo amplitude. Assuming TIA= co the damping AE term of the A-echo due to the excitation of the B-spins by the 180" pulse is AE == 2 (CO~B)Z2E~c(2 Z).Z)TAA(~ (1) E, is the term in the A-echo due to the presence of the spin system C besides A and B. TAA describes the free induction decay signal of the A-system caused by the interactions A ++ A. Part of the second moment of the A-signal is due to the B-system and is given by (co~~) Y, yB are the gyromagnetic ratios of the A- and B-nuclei respectively; IBis the spin of the B-nuclei gjk a geometrical factor and rjk the distance between the nuclei Aj and Bk. For fixed experimental condition z = constant the terms EAc(2Z)and TAA(~~) may be considered constant and It is this l/r6-dependence which is used in our EHK experiments to determine relative assignments of n.q.r.signals with respect to the geometry of the molecules considered. RESULTS AND DISCUSSION HETERONUCLEAR DOUBLE RESONANCE The crystal structure of KA12Br7 is known.5 The geometry of the ion [AI,Br,]- is shown in fig. 2. In accordance with the crystal structure seven n.q.r. signals of 81Br in the frequency range 83.5 6 v/MHz 6 91 are detected at 77 K. No simple relation between the n.q.r. frequencies and the structure of the ion [AI,Br,]- is ap- parent. In the following Roman numeralsgive thenumbering of the atoms according to the crystal structure (fig. 2) Arabic numbers are used in listing n.q.r. frequencies coup- ling constants etc. The relation between Arabic and Roman numerals is found through the n.d.r. experiments. Using the EHK n.d.r.81Br f-+ 27AI the weak n.q.r. signals of 27A1,were found; their frequencies are given in table 1. The grouping of the 4 signals into two sets bf1*f7 (A41)) Vf ;*f&441hI and [v* 1 f-+ * d (A42)) I.'* 1f--+ * &442JI can be achieved by the following arguments NUCLEAR QUADRUPOLE DOUBLE RESONANCE from 0 < q < 1 the condition 2 3vf +-+ ;/v* + 4 31 follows. The ratios of the signal intensities 1,; ++*;/Irtt t--+*; should be -1 :2. V(AI(~Jand q(A4,J should be of comparable magnitude. Br * B ry FIG.2.-Geometry of the ion [AlzBr7]-in KAl2Br7 The assignment of the different signals is shown in table 1 together with e'qQ/h and q calculated from the tables of Livingston and Zeldes.6 The structure of the ion [AI,Br,]-is such that from the l/r6-dependence the relative positions of Br .. . vlI with respect to the atoms All and All and also with respect to the internal grouping of the Br-atoms is possible. For example Br, Brlr and Br,, (see fig. 2) should show a much stronger n.d.r. effect with All than with All,. AE of the bridging atom BrIv should be strongly influenced by Al and AI,, etc. The influence of the neighbouring ions can be estimated from the crystal structure data. The shortest interionic.distance AI-Br is >4 A. Compared with the largest bond distance Al-Br = 2.43 A the ratio (AEinter)/(A€intra) = (2.43/4)6 z 0.05. Therefore the influence of intermolecular interactions on the EHK experiment is negligible for KAI,Br,. In two series of experi- ments the relative assignment (Br(l .. . ,)) ++ A1(1,2)was determined. (a) With con- stant parameters of the A-channel the influence of the four A1 resonances on the seven "Br resonances was studied. Each of the two nuclei influences four 81Br reso- nances strongly three 81Br resonances only weakly. The influence of both2,A1 nuclei on one 81Br signal is of equal strength so this signal belongs to the bridging Br(Iv). TABLE 1 .-27AI N.Q.R. FREQUENCIES COUPLING CONSTANTS AND ASYMMETRY PARAMETERS IN KAI,Br, T = 77 K ''A](,) 1.304 2.562 8.652 0.117 1.628 3.068 10.324 0.220 A complete assignment of the next neighbours was found as given in fig. 3. This experiment also gives a unique determination of the four frequencies to the two crystallographically inequivalent A1 atoms within the ion [A12Br7]-.(b)With constant parameters of the B-channel for all "Br frequencies the four 27Al transitions were studied. The results of this experiment are in accordance with experiments (a).' The double resonance studies on KA12Br are in agreement with the single crystal n.q.r. study of Yamada on this substance. N. WEIDEN AND A. WEISS v (*IB~)/ MHZ Bq,) 83.583 f3ti2] 85.152 8.562 Bq,) 85.596 Br14) 86.142 Brl,) 66.319 10324 B561 86.928 Brol 90.906 FIG. 3-Assignment of the seven *'Br n.q.r. frequencies and the two 27Al coupling constants in KA12Br7 to terminal and bridging bromine atoms and to Al(, and AI(*,,respectively. T = 77 K. HOMONUCLEAR DOUBLE RESONANCE In molecular compounds (TCA),X and salts of TCA the mean 35Cl n.q.r.fre- When n quency of the adducts is -1-2 MHz higher than that of the ~alts.~~~*~ = 2 rn = 1 or when two or more TCA molecules are present within the asymmetric unit more than three 35Cl signals are found. 35Clc)-35Cl n.d.r. offers a chance to assign the signals to the individual CC13 groups of the TCA molecules. Two crystal struc- ture determinations are found in the literature for TCA X one is TCA itself," the other TCA pyridine-N-oxide.11y12 In each case three 35Cl signals are ob-ser~ed.~~~'~ These substances are useful to study the intermolecular and the intra- molecular influence on AE separately by an isotopic n.d.r. 35Cl ft37Cl. Numbering the C1-atoms in C13CC02H by Cl,,, CI,, and Cl(3, respectively AE in an n.d.r.experiment 35Cl(1) .c)37Cl(2) and 35Cl(1 f-) 37C1(3 contains both intermolecular and intramolecular terms. The n.d.r. 35CI(,,ft37Cl,l, 35Cl(2) ++ 37Cl(2 and 35Cl(3 ftj7C1(, etc. shows only intermolecular effects. In fig. 4 three of the n.d.r. signals of TCA found are shown. As expected from the crystal structure the . r . . s2 s3 5. QJb+w 31.717 > i 0.0 74 I I . 31.654 . 31.501 . /-i NUCLEAR QUADRUPOLE DOUBLE RESONANCE intermolecular terms are small compared with the intramolecular ones. Similar results have been gained for TCA pyridine-N-oxide. Investigating TCA X with more than one TCA within the asymmetric unit the assignment of the 35C1-n.q.r. signals to a distinct CCl is not restricted to the isotopic n.d.r.Much shorter sampling times can be achieved by 35Cl++ n.d.r. because (a) the mutual abundance of 35Cl is a factor of three larger than that of 37Cl (b) y(35Cl) z 1.2 Y(,~CI)which influences AE by a factor of (1.2)'. In total the S/N ratio improves by a factor of 4.3 compared with a 35Clf-t37Cl experiment and thereby the time for an experiment decreases by a factor of 18 to 19. FIG.5-Homonuclear n.d.r. 3'Cl ft35Cl in TCANa-3H2O. A-channel frequency is 39.383 MHz. S1 Sz S4 intermolecular n.d.r; S3. S5 intramolecular n.d.r. T= 77 K. The numbers in the figure are the respective frequencies in MHz. Four molecular compounds with TCA and the sodium salt TCANa 3H,O were studied by 35Cl ++35Cl n.d.r. In TCA * vanillin two signals overlap at 77 K (39.799 MHz).At T = 168 K this unresolved doublet splits into two lines 39.386 and 39.432 MHz. The other lines at 168 K are (in MHz) 38.919; 39.608; 39.780; 39.849 (all +0.005). In fig. 5 the n.d.r. of the upper signal (39.383 MHz) with the five other signals of TCANa 3H20 is shown. (AEintra)/(AEinter) was in all experiments 2 3-4 so that n.d.r. ,'Cl ft35Cl was done for each possible coordina- tion within one compound (15 measurements of a six line spectrum). In table 2 the results of the assignment are given. As can be seen the splitting of the n.q.r. (in which no systematic differences in the frequencies can be observed) within one compound is mainly due to the crystal field effect but not to a pronounced chemical inequivalence of the two CC1,-groups.TABLEASSIGNMENT OF 35c1FREQUENCIES TO TWO CI3C-GROUPS IN FIVE TCA COM-POUNDS. T=77 K. TCA* TCA. 2 TCA. TCA. TCANa*3H20 compound p-chlorphenol dibenzylether 1,4-dioxan vanillin -40.620I 40.370 39.855-1 1-39.715 ~' 1-40.320 -40.23 2 -40.122 39.383 39.329 V(3TI) ~ 40.1021 39.702-1 /MHz 1-40.0151 -39.709 I 39.533-j L39.462 ' 39.380-'-39.348 39.325-' 139.353 38.544 37.980 N. WEIDEN AND A. WEISS ISOTOPIC DOUBLE RESONANCE The halogen n.q.r. on (3,5-C12C6H30A1Br2)2 shows two "Br signals and two 35Cl signals at 77 K. For (4-C1C6H4OA1Br2) one 35Cl and two slBr signals are found. The halogen n.q.r. frequencies are listed in table 3. The crystal structures of these compounds are unknown. On the basis of the n.q.r. results four structure models for the molecules are possible which are shown in fig.6. The dimeric character of R R R (41 / R R FIG.&Four molecular structure models for (3,5-C12C6H30A1Br2)2 and (4-CIC6H40AIBr2)2which are in agreement with pure n.q.r. data. (4-ClC,H40AlBr2)2 was proved by determination of the molecular weight4 and is supported by the crystal structure of a similar compound dibromotrimethylsiloxy- aluminium.l5 The n.q.r. frequencies of 27Al were determined by 81Br -27Al n.d.r. at 77 K. Each substance shows only two 27Al resonances the data are listed in table 3. Since TABLE3.-N.Q.R. DATA OF (3,5-DICHLOROPHENOXY ALUMINIUMDIBROMIDE)2 AND (4-CHLORO-PHENOXY ALUMINIUMDIBROMIDE)~ AT 77 K compound (3,5-C12C6H30A1Br2)2 (4-ClCsH40AIBr2)2 [~(~lBr~~,)/(~~Br~ 86.150/103.128 86.650/ 103.732 ,,)]/MHz [v('lBr( 2,)/( 79Br( ,,)]/MHz 89.466/ 107.097 87.846/105.162 v(~~CI)/MHZ 35.382 36.428 36.35 I [v(I!Z+ c)~tt)(~'Al)]/MHz 1.620 1.806 [v(f-$ ++ +~)(27Al)]/MHz 3.152 3.211 e2qQh-'(27Al)/MHz 10.552 10.910 q(27~1) 0.148 0.317 one 27Al nuclear quadrupole coupling constant was found the two A1 atoms within the molecules are equivalent and the structures (1) and (4) of fig.6 can be neglected. From 81Br -27Al n.d.r. structures (2) and (3) are indistinguishable. However isotopic double resonance 81Br ft 79Br offers a way to solve the problem. In this experiment the models (2) and (3) are inequivalent. Assuming structure (2) to be NUCLEAR QUADRUPOLE DOUBLE RESONANCE the correct one the n.d.r."Brtl f-) 79Br(, should show a strong effect since the intramolecular distance BI-(~)-B~(~) is the shortest Br-Br distance within the molecule. The echo amplitude should be much less influenced by the EHK experiment on *lBr(, c)79Br(1) and 81Br(2) 79Br(2) respectively. f-) If structure (3) of fig. 6 were present the damping of the echo by the n.d.r. 81Br(l)c)79Br(2) and S1Br(2 +t79Br(2 would be pronounced because of the rela- tively short intramolecular distances Br(l)-Br.(2) and Brc2)-Brc2). The experiment 81Br(l,c)79Br(1) should show only a small effect. The n.d.r. experiments 81Br f-) 79Br show that structure (2) fig. 6 is the correct one for both compounds (3,5-C1,C,H3OA1Br,) and (4-C1C6H4OAlBr2),. CONCLUSION In the three different n.d.r.experiments applied here the use of the EHK method in increasing sensitivity and gaining information about the molecular geometry and the assignment of frequencies is shown. The main limitation in structural studies is that the intermolecular distances have to be considerably larger than the intramolecular ones. The ratio of those distances should be at least 1.3. For this reason the deter- mination of a relative bond length scale in molecules and complex ions may often be impossible because of the small differences in the interatomic distances. The resolu- tion of the homonuclear double resonance is high at least better than 20 kHz as shown in fig. 5 for the sodium salt of trichloroacetic acid TCANa 3H,O. Isotopic double resonance is useful for 81Br ct79Br 35Cl f-) 37Cl and should be successfully applicable in other cases e.g.123Sbf-) '?3b. We are grateful to the Deutsche Forschungsgemeinschaft for financial support of this work. M. Emshwiller E. L. Hahn and D. Kaplan Phys. Rev. 1960 118,414. N. Weiden and A]. Weiss J. Magnetic Resonance 1978 30 403. D. Biedenkapp and Al. Weiss Ber. Bunsenges. phys. Chem. 1966 70 788. T. Deeg and Al. Weiss Ber. Bunsenges. phys. Chem. 1976 80 2. E. Rytter B. E. D. Rytter and H. A. Oye Acta Crysf. 1973 B29 1541. R. Livingston and H. Zeldes Tables of Eigenvalues for Pure Quadrupole Spectra Spin 512 (Oak Ridge National Laboratory Report 1955 ORNL-1912). K. Yamada J. Sci. Hiroshima Univ. Ser. A 1977 41 77. * Yu. K. Maksyutin E. N. Gur'yanova and G.K. Semin Uspekhi Khim. 1970,39 727. Al. Weiss Adv. Nuclear Quadrupole Resonance 1974 1 1. lo P.-G. Jonsson and W. C. Hamilton J. Chem. Phys. 1972,56,4433. l1 L. GoliE and F. Lazarini Vestnik Slovensk. kem. Drustva 1974 21 17. l2 L. GoliE D. Had5 and F. Lazarini J. Chem. SOC. D 1971 860. l3 H. C. Allen Jr. J. Phys. Chem. 1953 57 501. l4 H. Chihara and N. Nakamura J. Phys. SOC. Japan 1974,37 156. l5 M. Bonamico and G. Dessy J. Chem. SOC.A 1967 1786.