Robert G.

Bryant University of Minnesota, 207 Pleasant Street SE, Minneapolis, MN 55455 It is common jargon to say that some process is fast or slow on the "NMR time scale" implying that there is in fact some precise time scale that is particularly appropriate to NMR and which is understood by all. While there may have been some validity in this imprecise jargon when relaxation time measurements were ignored, and when NMR was confined to protons, the truly multinuclear character of NMR spectroscopy and the ready access to NMR relaxation rates provided by commercial instrumentation today makes any blanket statements about the "NMR time scale" imprecise at best and more often than not completely meaningless. What are the time scales associated with NMR? A convenient way to look a t this question is in terms of the rates of physically or chemically significant processes that can be extracted from a study of NMR spectra. The primaly features that are easily available in any spectrum are the intensity, the resonance frequencies or the chemical shifts, the fine structure or the scalar coupling patterns, and the relaxation rates associated with the lines. Each leads to a somewhat different "NMR time scale" because different rates or rate constants may he extracted from study of these different aspects of the spectrum. The intensity of a line may be used trivially to measure concentration as a function of time as in any form of spectroscopy and, since this application provides no confusion about the time scale for the reactions or the rates involved, it will not he discussed further. Fundamental Background The fundamental basis for extracting kinetic information from changes induced in an NMR spectrum by a chemical exchange event is the uncertainty principle. If a nucleus may sample more than one magnetic environment in a chemical exchange process, then the uncertainty in the energy or equivalently the resonant frequency is related to the lifetime, r , by the familiar relation
Au = hl(2ar)
(1)

Figure 1. The idealized NMR specha for a pair of exchanging resonancessep. - u, shown schematically at the extremes of arated by a frequency Au = v slow and fast exchange.

quency difference between the two lines (1,2).A pair of lines will just merge to become one at the point,
r,,,~,,,,,,,
=

tJ? TAU)-' = k-'

(2)

where here Av is the energy difference between the two lines expressed in Hz. An idealized representative situation is sketched in Fieure 1. This relation is useful for making more time scale. If it is slow, resonances are well resolved for the interchanging species, while if it is fast, resonances are averaged and only a single line is resolved. The Chemical Shift When considering the time scales appropriate to averaging two resonance lines separated by a chemical shift difference, eqn. (2) shows that the two lines will he resolved if the exchange rate is small compared with the chemical shift differ1m.r i l l Hz, Av, la.tw.e~l tl~t e w line,. ~ Only n i n g l t nwrn;me n,ill be ,,hserved i r rhr rn.hnn:v r:ttt I . hr:e wnpared t u the n l m. m r i ~ tI~ II . shirr diiterrnw. Thus. t o kmru the time >C.RC n.. the averaging process that causes coalescence of the NMR

That is, the frequency of the resonance is precisely measurable as long as the lifetime in that state is long, but as this lifetime decreases the uncertainty in the energy or resonance frequency increases and one observes a lifetime broadening. In the simplest case that the observed nucleus exchanges hetween two environments with equal populations, the spectrum collapses to a single line at the average resonance position when the exchange rate, 117, is large compared with the fre-

Representative Chemical Shift Ranges, Coupling Constants, and Time Scales

Nucleus

'H
I3C 15N "F 31P 5eCo r*9Hg

010 0200 0 900 0 300 0 700 0-15.000 0- 3,000

0 0.6 0 3 0 5.4 0 17 0 17 0-214 0 32

0.2 5-04 rns 0.2 5-75 ps 0.2 5-40 ps 0.2 5-33 ps 0.2 5-13 ps 0.2 5-Ips 0.2 S-7 ps

03 0 15 027 085 085 0-1.070 0- 160

0.2s-75 ps 0.2 s-15 ps 0.2 s- 8 ps 0.2 s- 3 ps 0.2 s- 3 ps 0.2 s- 0.2 ps 0.2 s- 1.4 ps

ZjHH-10 'J~,.-150 ljNH -50 2~nF -50 ZJp-p-20 'J~*.N -50 ' J > ~ - ? e e ~ 2,500 ~

-22 -1.5 -4.5 -4.5 -11 -4.5 -90

rns rns rns rns rns rns ps

= n k e n as the coalescence lifetimegiven by eqn. 2 for a maximum shift; eg.. 10 ppm for 'H, and a minimum resolvable shinof 1 Hz. Braken as the coalescence lifetime obtained by substitution of J for Av in eqn. 2.

Volume 60 Number 1 1

November 1983

933

lines, one must know the chemical shift difference a t least approximately. Here additional care is required about blanket statements because chemical shifts may differ greatly from one nucleus to another as shown in the tahle. t cherniwl In>pesrionof column 2 d t h e tnhlv shova t h ~ the shiit mnws diiier cuns~drrahly irtm one nu' Ieus to another. I t must be appreciated that the chemical shifts actually averaged by an exchange event generally will be some fraction of the approximate total ranges indicated. Nevertheless, the conditions for the averaging of two resonances will clearly be WIIPII ,me is h5erving protun, compared with wry diiirren~ o r ,di:dt. u,hich h ~ \ . much e larger chemiljil nuclei like tl~wrine shifts. Comparison of columns 3 and 4 representing the chemical shift ranges at two common proton resonance frenuencies. 60 and 300 MHz. voints out an additional prob--. >~~~ ~, lem. Because the chemical shift difference between two resonances increases linearly with the strength of the applied dc magnetic field, the exchange rate between two averaging resonances must be 5 times larger at 300 MHz than at 60 MHz to achieve a coalesced spectrum. Thus, to make spectra look identical on the ppm chemical shift scale, the temperature of the compound studied would have to be somewhat higher on the 300 MHz svectrometer than on the 60 MHz spectrometer
~ ~ ~

50-

40

T,, s-'

3020-

". . ..
0.I

* * .

100
L

..

. ..
io

FREWENCY, Mlz
Figure 2. Proton longitudinal relaxationrate as a function of frequency for a 0.80 mMaqueous MnCI2 solution at 286 K measured from 0.01 to 35 MHz using a field cycling technique (4).

ations in the magnetic interactions that modulate the energy levels of the nuclear spins in the sample. However, since orientation of a spin in-a magnetic field requires energy, the fluctuation spectrum must have components at the resonance or Larmor frequency of the observed nucleus for the fluctuations to be effective. The relaxation equation for the isotropic intramolecular dipole-dipole interaction is

priate for averaging NMR chemical shifts, it is clearly necessary to know the magnitude, in Hz, of the chemical shift being averaged. Scalar Coupling Essentially identical arguments apply to the averaging of scalar couplings, and it is sufficient for present purposes to replace Au by J in eqn. (2) to provide the same sort of basis for discussion of the time scales that will lead to averaging or collanse of the svectrum. If the lifetime of the nucleus in a particular environment is short compared with the reciprocal of the scalar coupling constant in this case, then the coupling will not be directly observable, and only a single, coalesced line will annear. .. Insvection of the tahle indicates that, while the scalar coupling constants do not span as wide a range of values as the chemical shifts and the scalar coupling constants are

where v is the maenetoevric ratio. h Planck's constant divided -- by 2a, r the intermoment distance, w the resonance frequency, . . the correlation time characterizing the fluctuations in and 7 r. The restmanw frequency, of cows( , dtpendz m t h r t r e n ~ t h d t h e at~i~litvl . .~ .. IIC ninenetit, field. I f . and on the marnrlwvric ratio, y, of the nucleus observed according to the relation
u .
~~ ~

I + , , , = yH12a

(4)

and distance between the coupling nuclei Relaxation The averaging of even large shifts generally still limits the accessible rates of vrocesses conveniently studied to times on the order of microieconds or longer. ~ i w e v e rNMR , relaxation spectroscopy extends this range considerably. There is a great variety of NMR relaxation rates that may be measured which all differ somewhat in the instrumental and theoretical framework required. However, the range of rates that may be studied is adequately represented by consideration of the relaxation time that is most easily understood and measured, the lonritudinal NMR relaxation time, TI, or the rate, 1lT1. l / T l i s t h e first-order rate constant that characterizes the return of nuclear magnetization to an equilibrium magnitude in the direction of the applied field strength following a perturbation. The perturbation may be as simple as dropping the sample into the field in the first place and monitoring the growth of magnetization or monitoring the return to equilibrium after the magnetization has been driven away from equilibrium by a strong r.f. pulse (3). The longitudinal NMR relaxation rate is driven bv fluctu934

Depending on the nucleus observed and the field strength, the practical resonance frequencies presently range from several MHz to a maximum of 600 MHz for protons. The usual experiment is to measure the relaxation rate at fixed frequency as a function of temperature. Equation (3) predicts that the rate will nass through a maximum when w ~ is, close to unity (0.6158 he more precise). If the structure of the molecule is known, the internuclear distances are determined so that the only unknown in eqn. (3) at this point is the correlation time characterizine reorientation of the intermoment vector, r. T h e timt, ir.ile of molwular event; that can he monit<wd bv rhe loneitudinal n~lnxiltionr3tt.i arc thus on thc urder uf the reciprocal of the resonance frequency, (2ay)-1 and with present day spectrometers may practically include times from to 10-'0s. It is to be noted that the relaxation rates about actuallv measured are usually on the order of seconds but are deterkned by much faster &lecular motions near or above the Larmor or resonance frequency. What do these times correspond to physically that may he of importance to a chemist? Any motion that modulates the magnetic interaction may contribute. These times characterize such processes as very rapid chemical interconversions, like transfers or translational or rotational motions. The longitudinal relaxation rate provides a very direct way to measure the correlation time for rotational reorientation of a protein molecule for example (4). If nuclear relaxation is driven hv interaction with an electron magnetic moment, as in a paramagnetic metal ion complex, then the time scale of events that may he monitored by relaxation is pushed to very short times. The relaxation of water nrotons hv manganese(I1) ion serves as an example. The longitLdinal refaxation equation in this case becomes

tb

Journal of Chemical Education

where the correlation times 7,and 7, may have several contributions (5-7). For the aqueous solution of the aquo comD ~ X the . correlation time for the first term in hrackets is the r t t iticmi~l (. melaritm rime t h y mrt:il c ~ ~ n l l ~ wI11Ie l ~ . x . that i t r i n t it.<ondtvrm I S the t It.,.tn,~trelmdticr~~ 1i11.e.Ti.. 11 is the nuclear electric hyperfiue coupling constant which is analogous to the scalar or J coupling between nuclei, except that in the present case an unpaired electron spin is involved. wr and w, are the Larmor precession frequencies for the proton and the electron, respectively. The frequency dependence of the relaxation rate for the water protons permits resolution of the different contributions as shown Figure 2. The inflection at low frequency is caused hy the last term in eqn. (5) and therefore reports the electron relaxation time, TI,. The higher frequency inflection results from thew, part of the first term in eqn. (5) and reports the rotational correlation time of the metal complex. If we take the inflection point as 5.5 MHz and write the electron Larmor frequency w, as ( y , / y ~ ) w= ~ 658 w ~ , then the rotational correlation time of the metal complex is the reciprocal of this inflection frequency, T , , ~ 1/(658 X 2 4 = 44 ps. Though this is a limited example, it demonstrates clearly that the study of nuclear relaxation extends the time scale accessible to investigation by nuclear magnetic resonance measurements well into the picosecond range.

confusion ahout the NMR time scale may be eliminated if the interaction that is modulated by a dynamic process is always included in a statement about the spectrum or relaxation involved. Thus, instead of suggesting that some process is fast on the "NMR time scale," which is ambiguous, indicating that the process is apparently fast or slow compared with, for example, the chemical shift difference between the two resonances which is expected to be on the order of xxx Hz (not ppm) at the fields used, will minilhize amhiguity.

Acknowledgment The data shown in Fizure 2 were obtained bv Dr. David Hickeraun ;ancl l l r . Hugh (;,,~ts:siparr or wrksupported hy i h t , Natim>alI ~ i ~ t i t u t ~ lle:iltl~ . s ~ ~ t!(;312!)4h!,MI the Ka11on.11 Science Foundation (PCM-8106054). Literature Cited

--

Summary In summary the "NMR time scale" spans the range from days to picoseconds. I suggest that much of the amhiguity and

(11 S e e f o r e x a m p i e B o u e y , F A , " N ~ ~ I e ~ r M ~ g ~ e t i i R ~ ~ ~ ~ ~ ~ ~ ~ s p ~ ~ t t t t t t ~ ~ ~ A c a P ~ e s sNew . York, 1969,Ch. VLI. (21 Ksplan, J. I., and Fraenkel. G.."NMR nfChemiraiiy Exchanging Systems? Academic P ~ e s sNew . York. 1980. 13) Farrar, T. C.. and Becker,E. D.."Pulse and Fourier Transform NMR,"AesdemicPress, New York, 1971. (0 Hsllenga, I<., and Kuenig, S. H., Bi~ihhfmisW15,4255 (1976). 151 Swift, T. J., in "NMR of Paramagnetic Moleculms,'. (Editors:I.aMar, G . N., Hunucks, Jr.. W. Dew.,Ho1m.R. H.l,AcademiePress, NeaYork, 1973, Ch. 2. (61 Dwek, R. A , "Nuclear Magnetic Resonance in Biochemistry: Applications to Enzyme Systema?Tho Ciarendon Press. Oxford,1973.

Volume 60 Number 11 November 1983

935