COMMUNICATIONS  G . Christim H Stolzenberg. W P. Fehlhammer. J. Cheni. L S c i ~ .C%eni. C o n stitutes the major part of the entire measurement time. In this iizun. 1982. 184. communication we introduce a post-processing procedure for  W. P. Fehlhammer. M. Fritz, Chwi. R O Y .1993. 93, 1243. NOESY and ROESY spectra allowing the acquisition time to be  a) W. P. Fehlhammer, F. Degel. .4ngrii.. C % m . 1979. Y I , 80: Angcw. C h n . inr. shortened with a marginal loss in the signal-to-noise ratio and the €d. €ny/. 1979. 18. 75, b) W. P. Fehlhammer. E Degel. G. Beck. C ' i i m i . B w . 1987. 120. 461 accuracy of the integration of the cross-relaxation peaks. A  a ) W. P. Fehlhammer, G. Beck. .4nym. C'hheni. 1988. 100. 1391 ; A t ~ g c C~ .~ W . similar procedure was described for the accelerated measureinr. EL/.~ q1988, / . 27. 1344: b) J O F ~ U W ~ chr.ii1. I W ~ . 1989, 369. 105: C) c / ~ ~ , j i l . ment of heteronuclear TI values in 1975.L41 Btw 1989. 122. 1907: d ) % . : V u r u r f o ~ h B . 1989. 44, 1414: e ) W. P. Fehlhammer. S. Ahn. G Beck. J. O~-gnnorwr.C k i n . 1991. 4 / 1 . 181; 13 S. Ahn. G . Beck. W P. Fehlhammer. ihnl. 1991, 418, 365.  W. P Fehlhammer. G. Beck. J. Oi.grnionic,l Chon. 1989. 37Y.91. [XI M. Krdger. H Dreizler. D. Pi-cugschat. D. Lentz. Anyew. C h i i . 1991. /(/.?. RD 1673; Aij,q<,lr. C - / i m z . In!. Ed. Ei1,qi. 1991. .?O. 1644. Tn 21) P.J. Stang. Ajigtir. C/KW.1992. 1114. 281 :Anycw. Chwi. l i l r . .Ed. .Et?g/.1992. 31. 274: h) M Ochiai. M. Kunishima. Y. Nagao. K. Fuji. M. Shiro. E. Ftijit;~, J. '4)ii. Chciii. S(K 1986. /OH. 8281 Fig. 1. Pulse sequence of NOESY and KOESY cxperiments with relaxation delay of this reagent i n oreanometallic chemistrv have receiitlv [lO] Firht annlications .. been reported b) S u n g et al: a ) P. J. Stang, C. M. Crittcl, Or,~u~ioi~icrii//ir~.\ RD. mixing time r m . and acquisition time AQ = r,(max). 1990. Y, 3191, b ) P. J. Stang, K. Tykwinski. .I An?. Chon. SOC. 1992. 114. 4411 Alk~lethynyliodoniumsalts with longer chains. wch as [nPr-C=CIPh]' X and [nDec-C-CIPh] iX - react with cyanometalatea to yield I-cqclopentenyl NOESY and ROESY spectra contain cross peaks whose inte[socyanide complexes. that is, i n 3 carbene insertion occurs into a ;,-C-H bond grals reflect the cross-relaxation rate between magnetically active of the alkyl side cham. nuclei in the initial rate a p p r o x i r n a t i ~ n .The ~ ~ ] cross-relaxation 7 : MS (80 eV)- fri,': 461 [,!&I. 433 [ M - CO]. 405 [,M--2CO]. 377 [ M - 3COl. 349 [.bf - 4CO]. 321 [ h - - 5CO], 194 [ M - - 5CO - I]; IR(KBr). rates depend on the correlation time T , and the distance r between ?[cm '1 = 2110m (CN). 2055m. 1930vs (CO) [El. the interacting nuclei according to r - 6 . However, this relation Cry~tallographic data for 4 (6a): STOE four-circle diffractometer. only holds if the spin system relaxes back to thermal equilibrium ;.(Mu,,) = 0.71073 A. p(Mo,,) = 8.40 (8.84) c m - l . Amt,A:.4m,n =1.123:0.999 after each scan. An exact solution for incomplete relaxation be(1.230:0.858). 20 range 45211550 . temperature -60 ( + 20) C. cd SCHIIS: structure solution by Patterson method (direct methods). All H atonis were tween scans in a FAST NOESY experiment can be obtained from located. Retincnient . Method of least squares. H atoms isotropic, maximum the Liouville-von-Neumann equation.['"% 61 The integral of the residual electron density 0.425 (0.355) e . k 3 . Programs wed were XTAL 3.0. cross-peak between spin i and ,j (Ii:) is dependent on the relaxSHELX-76. ORTEP. DIFABS. Further details of the crystal structure investiation delay (RD) between scans, the mixing time T,,,, the acquigation m a y be obtained from the Fachinformationszentrum Karlsruhe. D76344 Eggensrein-Leopoldshafen ( F R G ) on quoting the depository number sition time AQ, and the relaxation matrix as given in [Eq. (a)]. CSD-57757. the names of the authors. and the journal citation. I L. Rebrovic, G. E Koser. J. Ory. C'hiwi. 1984. 4Y. 4700. M. Ochiai, T. 110. Y. Takaoka. Y. Masaki. M. Kuniihima. S. Tani. Y.Nagao. It: = [ e x ~ ( - ~ ~ , ) l ,1 , [I-exp(-f(RD + AQ)lJr = & - I J. C h i i S i r . C h r . Connnun 1990. 1 1 8 P. J. Stnnp. A M. Arif. C M . Crittell. .4nynir Chrni. 1990. 102. 307: ,411y~~ii~. Chm. / n r . GI. EnyI. 1990. 29. 287 M Ochlai, M Kunishima. Y Nagao. K. Fuji, E. Fujita, J C h i , . Sor . ( ' h r ~ t i i . Coniniuji 1987. 1708. The first factor I t i is due to the cross relaxation during the 4 + + r FAST NOESY ExperimentsAn Approach for Fast Structure Determination** Matthias Kock a n d Christian Griesinger* The determination of cross-relaxation rates by N M R spectroscopic methods['] provides interproton distances. Two-dimensional experiments like NOESY['] and ROESYC3'(Fig. 1 ) are gradually taking the place of classical one-dimensional saturation NOE experiments, since they allow the simultaneous measurement of all interproton distances in a molecule, irrespective of their number. Long longitudinal relaxation times ( T I )often determine the total measurement time. Normally, the relaxation delay between consecutive scans is on the order of two to three times the relaxation time of those protons with the longest TI values.[41For molecules in the extreme narrowing limit (molecular weight below 500 D), in particular, the relaxation delay con[*] Prof. Dr. C Griesinger. Dr. M. KOck lnstitut fur Organische Chemie der Universitit Marie-Curie-Strassc 11. D-60439 Frankfurt ( F R G ) Telefax- I n r . code + (69)5800-9128 [**I This work was supported by the Fonds der Chemischen Indnstrie. We thank Dr. M. Reggehn for discussion and V. Apelt and 0.Schedletzky for support in prepxing Figure 3. M. K. thanks the Deutsche Forschungsgemeinscha~tand the Fonds der Cheinischen lndustrie for postdoc grants. mixing time T,,,. and the second is due to the relaxation to thermal equilibrium during the relaxation delay and the acquisition time. When the T,value of a proton is approximately equal o r longer than R D + AQ, the integrals in the NOESY are too small owing to incomplete relaxation. The second term in Equation (a) is asymmetric with respect to the permutation of spins i and ,j. in contrast to the first term I,j . Therefore the cross-peak integrals in NOESY and ROESY spectra above and below the diagonal can be different for short recovery delays (Scheme 1). FAST NOESY FAST ROESY Scheme 1. Representation of thecross relaxation o between spins S, and S,. Despite the identity of the cross-relaxation rates o,,and cr,,. asymmetric NOESY and KOESY spectra result for short recovery delays when the T, values of the spins S, and S, are differcnt. This asymmetry is most notable for nuclei with strongly differing T, values. This effect has prevented the recording of NOESY spectra with short recovery delays.['] COMMUNICATIONS An exact compensation for incomplete relaxation including the effects of spin diffusion by complete relaxation matrix approaches in FAST NOESY will be presented el~ewhere.~'] Here we describe a procedure that allows the cross-peak integrals, Zij, of a NOESY spectrum with full relaxation to be obtained from the cross-peak integrals of a FAST NOESY experiment, ZiF. For small molecules the diagonal elements of the relaxation matrix f determine the TI values. Therefore the N multiexponential relaxation [d-exp( -T(RD AQ)Ij, can k = 1 be simplified to Equation (b). (Fig. 2a, b) the cross-peak integrals above and below the diagonal of the NOESY spectrum are almost identical; for RD = 0.25 s (Fig. 2c, d) they differ by up to a factor of 3. Above the diagonal (Fig. 2 a, c) the cross-peak integrals remain almost unchanged for R D = 5 s and R D = 0.25 s, because the T, values of the protons H-lOprO-Sand H-15pr"-Rare smaller or equal to RD AQ (see Table 1). + + N [g-exp(-r(RD k-1 + AQ)ljX= d - e x p ( - ( R D + AQ)/TIj) (b) 43.27 11.6 12.75 The Tl values of the protons must be determined by Inversion Recovery before the 2 D cross-relaxation experiment is recorded (Table 1). The evaluation of all Tl values requires a resolved I D NMR spectrum. i3 ~ 1.7 Table 1 . 7; relaxation times for the protons in compound 1 [a]. Proton 7; Proton 7; 1 3 4 pro-R 4p"-s 6.8 9 10Pro-R f 0P T P 4 1.684 (<0.1%) 1.724 (<0.1%) 0.782 (0.2%) 0.665 (1.6%) 1.593 (0.1 %) 1.064 (0.1 Yo) 0.598 (0.9 Yo) 0.603 (0.8 %) 1.794 (0.2%) 12 13 14 15 P - - R 15 p.0-s P2-P6, T3, T5 T2, T6 T7 2.533 ( < 0 . 1 "YO) 3.041( < O . l % ) 2.070 (<0.1%) 0.758 (0.1 %) 0.611 (1.8%) 3.200 ( t 0 . 1 Yo) 2.375 ( t 0 . 1 Yo) 1.184 (0.5%) 11 [a] The assignment of the protons is shown in Scheme 2. The T, relaxation times are given in s. the statistical errors in parentheses. The Inversion-Recovery experiment was recorded with 79 different recovery delays and evaluated with the T, routine of the Bruker UXNMR program. We demonstrate this procedure on a complex organic micromolecule (1, molecular weight 450 D, Scheme 2) with positive NOES.["] A series of NOESY spectra at 400 MHz with relaxation delays of 5.0, 3.0, 1.5, 1 .O, 0.75, 0.5, and 0.25 s were recorded. The measurements were performed at a temperature of 300 K in CDCI, with AQ = 0.53 s and z, = 0.5 s.[l This mixing time is still in the linear region of the NOE build-up (cf. ref. 11 01). CH, 1 H " Scheme 2. Structure of 1. Figure 2 shows sections of NOESY spectra with the cross peaks between H-12/H-13 and H-15p*0~R/H-10pr0~S. For RD = 5 s Angew. Chem. Int. Ed. Engl. 1994, 33. No. 3 0 VCH 6.3 6.2 6.1 1.7 1.6 1.5 - 6 -6 Fig. 2. Cross peaks in NOESY (a and b, R D = 5 s) and FAST NOESY spectra (c and d, R D = 0.25 s) between H-12/H-13 and H-15P"-R/H-10P"-Sabove (a and c) and below (b and d) the diagonal. The integrals of the cross peaks are given in italics. Figure 3 a shows a statistical evaluation of the integrals as a function of the relaxation delay RD: liF(RD)/Zij(RD = 5 s). The systematic error of the integrals obtained from the FAST NOESY experiments relative to those obtained from the experiment with R D = 5 s is dramatic, especially for the very short recovery times. lij= Ii'$l-exp(- (RD + AQ)/T, j)] (C) Compensation of the integrals in the FAST NOESY experiment according to Equation (c) provides the integrals, l i j of , an NOESY experiment with full relaxation. This is demonstrated in Figure 3 b by a statistical evaluation of the ratio between the compensated integrals from the FAST NOESY and the integrals from the NOESY experiment with R D = 5 s. The ratios are centered around 1 as expected. This is even true for the experiment with R D = 0.25 s, which had a relaxation delay 20 times shorter than the reference experiment (RD = 5 s). The interproton distances obtained from the uncorrected FAST NOESY integrals ZiF deviate from the values obtained from the NOESY spectrum by up to 15%. With compensation for incomplete relaxation, the interproton distances obtained from the FAST NOESY are similar to those obtained from the fully relaxed NOESY to within f 2 % . The acquisition time for NOESY and ROESY experiments can be reduced by a factor of 5 to 10, and the determination of interproton distances is almost as accurate as that based on an NOE spectrum with full relaxation. This allows spectrometer VerlaKsgeseiischafrmbH. 0-69451 Weinheim, 1994 0570-0833/94/0303-0333$10.00+ ,2510 333 r-''4 " a) 1.2 " " ' " ' " ' " " " " " ~ ' " t 0.2 tc R D Is1 t'",'' 1.2 t ~ ' ~ ~ ~  D. Canet. G. C. Levy. 1. R Peat. J. Mugn. Rcsun. 1975. 18. 199-204. 15) D. N e u h u s , M. P. Williamson. 7 % Niic ~ l i w 01~rr17uusr~r E f f w i in Srrurtrrrul und Conforn~urii~nol A n u h i s , VCH. New York. 1989.  .I.Bremer. G. L. Mendz. W. J. Moore. J Ani. Chon. Suc. 1984, iO6, 4691 -4696.  a ) H . L. Eaton. N . H. Andersen. .I Mu,qn Rcson. 1987. 74. 212 225: b) N. H. Andrrsen. H L. Eaton, X. Lai. Mufin. K e . w r i . U m i . 1989. 27. 515-528. [XI a ) R. Boelens, T. M. G. Koning. R. Kaptein. J M d Srruct. 1988. 173. 29931 I . b) R. Boelcns. T. M. G. Koning. G. A. van der Marel. J. H. van Boom, R. Kaptein. J Mupi. K i w w 1989, 82. 290-308. [Y] M . Kock. C. Griesinger. lecture at the 5th Chianti Workshop on Magnetic Resonance. 1993. Pisa. Italy: d manuscript is in preparation. [lo] M . Reggelin. H. Hoffmann. M. Kock. D. F. Mierke, J. Ani. C/icw. Soc. 1992, / i 4 . 3272 3277. [l I ] The spectra were irecorded with 16 scans. a spectral width of 3846 15 Hr, 384 real points in / , (TPPI) and 4096 points in i z .TPPI wa, uzcd In order to obtain pure phase spectra. Processing yielded a matrix of 1024 x 2048 pomts. Before Fourier transformation a squared cosine function was applied A baseline correction with a fifth-grade polynomial w a s applied. ~ " ~ ' 1 " " '~ ' ~' " ' ~ ~ ' ~ " " ~ ~ ~ ~ " ~ ~ ~ Small Molecules Stabilized in Inorganic Frameworks: NbI, Monomers in the Novel Layered Compound Nb,S2II9** Q Gordon J. Miller* and Jianhua Lin RD [s] - Fig 3. a ) Integral ratios for the cross peaks of1 obtained from a FAST NOESY. I,',. ( R D ) and from a NOESY experiment. /,,. ( R D = 5 s) i n percent. A contour representation with a factor of 2 for the level spacing was chosen to indicate the number of integral ratios within a given range of percentage. If a value of /,'fl(RD).I,,(5s) occurs only oncc. it is represented by one contour line. If i t occurs n times it is represented by trunc(lg, n ) contour lines. Only those 46 cross peaks were evaluated that showed no overlap in the NOESY and whose T, values could be determined in the 1D spectrum. The integrals obtained from the FAST NOESY experiments are reduced to 1;5 of those from the fully relaxed NOESY. This corresponds to an error in distance of up to & 15%. b) Same representation as in a) but ~ i t compensation h of incomplete relaxation during R D + AQ accordin? to Equation (c). The d e i tions of the integrals obtained from the FAST NOESY are reduced to *I0 which corresponds to an error i n distances of k 2 % . time to be used much more effectively and reduces the time for the structure determination. If the recovery delay (RD AQ) approaches or is smaller than the transverse relaxation times T,. B, spoil gradients should be used during the relaxation delay RD. + Rcceived: August 16, 1993 [Z 6287 I€] German version: A n p i ' . Chcni.1994. l M . 338 [I] a) R. R. Ernst. G. Bodenhausen. A. Wokaun, Prin<.ipl<~s of .Yiiclcwr M y y i i e r i c Reronuncr i n #ne m i d nt.0 Dinim?sion.s. Clarendon. Oxford. 1987: h) H. Kcssler. M . Gehrke. C . Griesinger. Angew. Chani. 1988. /OO.507-554: Aji,qnl.. Clioii. I n r . Ed. En,?/. 1988. -77, 490 536.  a ) J. Jeener. B. H. M e w . P. Bachmann. R. R. Ernst, J. Chrni.Pli~.\ 1979. 7i. 4546-4553, b) S. Macui-a, Y. Huang. D Sutei-. R. R. Ernst. .I Mqqm Ke.son. 1981, 43, 259 281.  a ) A. A. Bothiier-By. R. L. Stephens. J. Lee. C. D. Wdrren. R. W. Jeanlor. J A m Cheni. Snc 1984. /Mi. 81 1 813: b) A. Bax. D G. Davis. ,/. M q n . Keson. 1985. 63.107-213: c) C. Griesinger. R. R. Emst. ;bid 1987. 75, 261 - 271 Many types of inorganic extended solids can stabilize and, perhaps. "solubihze" molecular species that may not be thermodynamically stable in either the gaseous state or liquid solution. These extended structures can also serve as reaction templates (heterogeneous catalysts) for reactions that are otherwise kinetically unfeasible. Zeolites and clays are just two examples of a number of interesting materials with such applications, because they exhibit large channels or open layers for insertion of molecular entities."] We are presently investigating the possibilities of early transition metal cluster compounds.I2.31 especially ternary niobium and tantalum chalcogenide halides. for such purposes. When small molecules are introduced in a solid-state host, often the structure and, in particular, the topology of the host remains unchanged (except for shifts in lattice parameters) or undergoes small displacive transformations. Seldom does one observe major topological changes in the extended solid, although there may certainly be a synergistic relationship between the structures adopted by the extended framework and the enclosed molecule. In Nb7SZIt,(more precisely [Nb,S,I,],[Nbl,]), we have found this type of structural synergism. A (001) projection of the structure of Nb,S,I,, is shown at the top of Figure 1.['] This compound forms nearly level sheets composed of Nb, triangular clusters, according to the description [ N b S i , , I i ~ , T ~ ~ z 2 I z Nb,SI,, .~], which are stacked perpendicular to the monoclinic ah plane. In these layers, each Nb atom is octahedrally coordinated by one S and five I atoms, and these coordination polyhedra share faces to form a bicapped Nb, triangle with one S and one I atom as the capping atoms. The twofold capping leads to significantly shorter Nb-Nb dis- ["I [**I Prof. c;. J. Miller, Dr. J. Lin Department of Chemistry. Iowa State University Ames. IA 50011 (USA) Telefax. Int. code + (515) 294-0105 This research w a s supported both hy the Chemical Sciences Division, Oftkc of Besic Energy Sciences. U.S. Department of Energy. under Contract D-740sEng-82 and the Donors of the Petroleum Research Fund. administered by the American Chemical Society. We thank Dr. L. Miller. Ames, for the resistivity and magnetic susceptibility measurements. and Prof. Dr. U. Muller for several suggestions. ~ '