Appl. Magn. Reson. 17, 609-614 (1999)

AppliedMagnetic Resonance

Springer-Verlag 1999Printed in Austria

MRI Study of Spatial Distribution ofPhotochemical Reaction Products

A. A. Obynochny', A. G. Maryasov', K. A. I1'yasov 2 ,O. I. Gnezdilov 3, and K. M. Salikhov 3

'Institute of Chemical Kinetics and Combustion, Russian Academy of Sciences,Novosibirsk, Russian Federation

'Kazan State University, Kazan, Russian Federation' Kazan Physical-Technical Institute, Russian Academy of Sciences, Kazan, Russian Federation

Received October 22, 1999

Abstract. Spatial distribution of molecules with chemically induced dynamic nuclear polarization hasbeen studied by nuclear magnetic resonance imaging. It is shown that heating of a system during thephotolysis can cause highly nonuniform distribution of reaction products due to a convective effect.

1 Introduction

To perform exhaustive kinetic analysis of chemical reactions one should knowthe spatial distribution of concentrations C(r, t) of different molecules: initial re-agent molecules, intermediate particles, and products. The formation of interme-diate particles and products during photolysis or radiolysis is in general casenonuniform throughout a sample. For example, the light intensity decreases deepinto the sample by the Lambert-Beer law

J(x) = Joexp(ax), (1)

where x is the distance from the surface, a is determined by the extinction co-efficient and concentration of absorbing molecules. Thus, intermediates and prod-ucts of photochemical reactions are expected to be distributed nonuniformly. Theremight be other reasons for nonuniform distribution of chemical reaction products.

The spatial distribution of products, C(r, t), can be found by nuclear mag-netic resonance (NMR) tomography methods [1, 2]. In these experiments, theNMR signal is detected from a slice of a sample. The slice width 5 determinesspatial resolution. However, when S decreases, sensitivity decreases as well. TheNMR sensitivity increases if nuclear spins are polarized. Suppose that the nuclear

610 A. A. Obynochny et al.

spin polarization exceeds the equilibrium polarization by P times. In this case, thespatial resolution of magnetic resonance imaging (MRI) increases P times, i.e.,if spins are polarized then the spatial resolution equals S q/P, where .5. charac-terizes the spatial resolution of MRI for equilibrium polarization of nuclear spins.

Photochemical reactions often proceed via a formation of radical pairs and,as a result, the chemically induced dynamic nuclear polarization (CIDNP) mani-fests for reagents and products. The CIDNP effect can increase the NMR sensi-tivity by several orders of magnitude (see, e.g., [3]). Thus, one can expect thatin a case of photochemically induced radical reactions in solutions the spatialdistribution of reagents and reaction products can be observed with NMR to-mography. Note, there is another circumstance which favors an increase in thespatial resolution of the distribution of reagents and radical reaction products.In radical reactions, nonuniformly distributed radicals will contribute nonuniformlyto paramagnetic relaxation of nuclear spins, interaction with free radicals andradical pairs will nonuniformly shorten the relaxation times T l

and T2 . Thus, freeradicals can serve as contrasting agents and can increase the spatial resolutionof NMR imaging experiments.

This paper aims to study the spatial pattern of methyl-tert-butylketone pho-tolysis in tetrachloride carbon by means of the NMR tomograph. We hope togain the sensitivity due to the CIDNP effect.

2 Experimental

Our experimental setup consists of the excitation system, reaction cuvette, andrecording system. Ultraviolet (UV) radiation of a superhigh-pressure mercury lampDRSh-1000 was used as a source of excitation. The UV radiation is passingthrough a system of focusing lenses to the cuvette with the sample under study.A system of spherical mirrors was employed to collect most of the UV radia-tion. The greatest number of quanta in the cuvette in the absorption band ofketones was 6. 10" quanta per s. In experiments two types of cuvettes were used,cylindrical and spherical. The diameter of the spherical cuvette was 50 mm, thatof the cylindrical cuvettes was 45 and 15 mm. The length of the cylindrical cu-vette was 50 and 110 mm, respectively. Radiation entered the cuvettes throughflat windows. In experiments, a 3 by 10 mm' diaphragm was used. Series ofexperiments were performed without diaphragm. The experimental setup is pre-sented schematically in Fig. 1. BMT-1100S NMR tomograph (Bruker) was usedfor recording. In experiments, a transmit-receive radio-frequency coil for a hu-man shin was used. We have employed the pulse sequence to obtain tomogramsand sections in different directions and the sequences for obtaining one-dimen-sional NMR spectra. The mechanism of CIDNP formation was studied with HX-100 (Bruker) and FX-90Q (Jeol) spectrometers. CIDNP was recorded by stan-dard techniques. In methyl-tert-butylketone photolysis the concentration of ke-tone was 0.5 M/l. Before irradiation, samples were bubbled with argon for 30min. All substances and solvents were purified by standard techniques. In ex-

NMR Study of Spatial Distribution of Reaction Products 611

by

Fig. 1. The experimental setup showing a part of the cuvette studied with an NMR tomograph.

periments performed on NMR spectrometers, the deuterated reagents of the firm"Isotope" were used as inner standards. In this case, the enrichment in deute-rium was 99% and the content of basic substances was 98.5%. The deuteratedreagents were not purified.

3 Results and Discussion

The mechanism of the methyl-tert-butylketone photolysis in CC14 is schematicallyshown in Fig. 2. The characteristic feature of the photolysis of aliphatic ketonesis that it proceeds through the formation of a radical pair (R,CO R 2). Accordingto this scheme the NMR spectra obtained on the photolysis of methyl-tert-butyl-ketone solution in tetrachloride carbon manifest the CIDNP effect. The totalpolarization for all substances involved in the reaction is negative and exceedsthe total intensity of the signals of the sample NMR spectra before irradiation.In an NMR tomograph the modulus of the total intensity of all NMR signalswithin a certain slice is detected.

Figure 3 shows the tomograms of the cuvette with methyl-tert-butylketonephotolized in CC14 . In Fig. 3, the results are presented when the photolysis wasinduced with the diaphragm protecting the cuvette from heating. Figure 3bh pre-

byCH3C(0)C(CH 3)3 0 (CH3C(0)C(CH3)3)*

t 'I,CH3CO C(CH3)3 ^ CH 3CHO + (CH 3)2C=CH2r1

CH3CO + C(CH3)32CH3CO > CH 3COCOCH3 / J HC(CH3)3, (CH 3)2C=CH2

2C(CH3)3 > (CH3) 3CC(CH3) 3CC14

Fig. 2. Scheme of methyl-tert-butylketone photolysis in CCl4.

612 A. A. Obynochny et al.

Fig. 3. a Tomogram of the experimental cuvette fragment with maximal UV absorption obtained onmethyl-tert-butylketone solution in tetrachloride carbon. Number of scans is 16. 1 is the region ofmaximal absorption and maximal CIDNP signal. 2 is the intermediate region. 3 is the region of theequilibrium intensity of the NMR lines. b, g Transverse cross sections of a in the region 2 close tothe region 1. c Transverse cross section of a in the region 1. d, e, f Longitudinal cross sections ofa. h Transverse cross section of a in the region 3. AB indicates the location of the cross section. 4shows the corresponding NMR spectrum. Note the intense sharp line in the center of c and in the

left-hand side of e giving the NMR signal from the region 1.

sents the tomogram cross sections, as indicated; it shows the contrast nonuni-form spatial distribution of the products of the photochemical reaction studied.In the case exposed we observe a dark rectangular region with a width of 0.3mm and a length of 10 mm surrounded from three sides by a prominent lightregion with a width of 2 mm after which there is a smooth transition to theregion in which no reaction occurs. The cross sections of different parts of thecuvette show that the dark rectangle region contains a strong signal which ex-ceeds several times the equilibrium signal. In the light region the signal sharplydrops to zero and further the signal intensity varies from zero to the equilib-rium value (see Fig. 3).

The result presented in Fig. 3 is interpreted rather straightforwardly. Indeed,the light is absorbed in some layer. In this area the CIDNP is created. In a bulkof a sample where light does not penetrate no CIDNP is formed, nuclear spinsare in thermal equilibrium. In the system studied the total CIDNP effect is nega-tive. Thus, there should be a region where the CIDNP effect just compensatesthe equilibrium polarization of nuclear spins.

NMR Study of Spatial Distribution of Reaction Products 613

Fig. 4. Tomogram obtained on photolysis of methyl-tert-butylketone solution in tetrachloride carbonplaced in a spherical cuvette in the absence of a screen. AA indicates the cross section of tomo-

gram, 1 is the region of maximal absorption.

The MRI pattern drastically changes if the diaphragm protecting the cuvettefrom heating is removed. In this case the dark region (the region with a strongNMR signal) is shifted to the upper part of the cuvette. For example, Fig. 4 pre-sents the tomogram obtained for the spherical cuvette in the absence of the screen.It is clearly seen that on the top of the cuvette the NMR signal is maximal. TheMRI pattern depends on the shape of the cuvette. These results indicate the pres-ence of a convective flow during the photolysis experiment performed withoutdiaphragm. When the diaphragm was removed the system is heated, this heatingis nonuniform, only a thin surface layer of the liquid will be heated. The heatedregion is subjected to the Archimedian force and the heated liquid rises to thesurface which causes convective instability of the medium. The convective liquidflow will carry molecules with polarized nuclei. The nuclear spin polarizationdecreases due to the spin-lattice relaxation in the time scale around 10 s. For thistime the polarized molecules can be carried by the convective flow to a distanceof several centimeters.

4 Conclusions

Our observations demonstrate that NMR tomography allows one to study the spa-tial pattern of a reaction zone in photochemical reactions accompanied by thechemically induced nuclear spin polarization. At the same time, it was found thatthe spatial distribution of reaction products can be essentially affected by the con-vective flow. In this case the spatial distribution of reaction products is deter-mined not by the structure of a reaction zone but by hydrodynamic phenomena.

Acknowledgements

This work was supported by the Russian Foundation for Basic Research (grants96-03-32956, 96-03-32485, 96-03-40043, 96-15-97444) and INTAS (grant 96-1269).

614 A. A. Obynochny et al.: NMR Study of Spatial Distribution of Reaction Products

References

1. Lauterbur P.C.: Nature 242, 190 (1973)2. Yakimchenko O.E., Smirnov Al., Lebedev Ya.S.: Appl. Magn. Reson. 1, 1 (1990)3. Salikhov K.M., Molin Yu.N., Sagdeev R.Z., Buchachenko A.L. in: Spin Polarization and Mag-

netic Effects in Radical Reactions. Amsterdam: Elsevier 1984.4. Stark D.D., Bradley W.G. (eds.): Magnetic Resonance Imaging. St. Louis: C. V. Mosby Company

1988.5. Carrington A., McLachlan A.D. in: Introduction to Magnetic Resonance. With Application to

Chemistry and Chemical Physics. New York: Harper & Row 1967.6. Salikhov K.M., Semenov A.G., Tsvetkov Yu.D. in: Electron Spin-Echo and Its Application. No-

vosibirsk: Nauka 1976.7. Loycyansky L.G. in: Mechanics of Liquid and Gas. Moscow: Nauka 1987.

Authors' address: Key M. Salikhov, Kazan Physical-Technical Institute, Russian Academy of Sci-ences, Sibirsky trakt 10/7, 420029 Kazan, Russian Federation