1712
J. Phys. Chtem. 1980, 84. 1712-1717
CHI and C032-ions according to reaction 6.4 -
,Lo
CH3CI' 0
+
OH
-
A
CH4
+ C03'-
observed in the present experiments following the reaction of cis-2-butene and 03-.Apparently in this case alkoxide ions are formed as a major intermediate since these result in the production of butadiene. The oxygen-containing products are attributed to peroxy radical intermediates of the type CH3-CH=CH-CH200.. These results demonstrate that reactions between alkenes and 03-on MgO generally lead to nonselective oxidation. One would hope to gain insight into the possible role of this ion in epoxidation catalysis, but rapid surface reactions, for example, between ethylene oxide and MgO, make it difficult to obtain such information.2 Moreover, there is no evidence for the formation of thermally stable ozonide ions on more acidic metal oxide surfaces.
(6)
A major difference between the reaction of C2H4with 0- or 03-is the relatively small amount of CHI detected in the latter case. This difference may be understood by considering that reaction 3 occurs when O2 is present, whereas with 0- reaction 4 resulted in acetate ions as the dominant surface intermediate. The similar product distribution which was okiserved following the reaction of propylene with 03- suggests a related reaction mechanism. The allyl radical formed by hydrogen-atom abstraction may react by reaction 7. One H~C=CH=CH'
+ 30 2 - + A
o2
-
Acknowledgment. This work was supported by the National Science Foundation under Grant No. CHE7809811.
h
References and Notes (1) Aika, K.; Lunsford, J. H. J. Phys. Chem. 1977, 87, 1393. (2) Aika, K.; Lunsford, J. H. J. Phys. Chem. 1978, 82, 1794. (3) Ward, M. B.; Lin, M. J.; Lunsford, J. H. J. Catai. 1977, 50, 306; Yang, T. J.; Lunsford, J. H., submitted to J. Catal. (4) Takita, Y.; Lunsford, J. H. J. Phys. Chem. 1979, 83, 683. (5) Ben Taarit, Y.; Symons, M. C . R.; Tench, A. J., J . Chem. SOC., Faraday Trans. 7 1977, 73, 1149. (6) DonaMson, J. D.; Knifton, J. F.; Ross, S. D. Spectrochim. Acta 1964, 20,847. Kagel, R. 0.; Greenler, R. G. J. Chem. Phys. 1968, 49, 1638. Miyata, H.; Wakamiya, M.; Kubokawa, Y. J. Catal. 1974, 34, 117.
cannot, however, rule out the formation of peroxy radicals which would result in the formation of CO:, and HzO.I In the reactions of 1-butene with 0-,butadiene was the principal hydrocarbon product,2 and in this respect 1butene acted more like an alkane than like ethylene or propylene, whereas n-butane gave a low yi.eld of butene when reacted with 03-.4It was somewhat suprising, therefore, when significant amounts of butadiene were
A 1,2 Hydrogen Shift and Other Thermally Induced Free Radical Reactsons in X-Irradiated Methyl a-D-Glucopyranoside Single Crystals. An ESR-ENDOR Study K. P. Madden and W. A. Bernhard" Department of Radiation Biology and Biophysics, University of Rochester, Rochester, New York 14642 (Received December 17, 1979) Publication costs assisted by the National Science Foundation
Upon warming single crystals of methyl cu-D-glucopyranosideX-irradiated at 77 K, three free radical reactions have been observed. At -190 K the deprotonated primary hydroxyalkyl radical, centered at C6, converts by a 1,2 hydrogen shift to a C5-centered secondary oxyalkyl radical. Further warming to -230 K causes a conversion of the C5 secondary oxyalkyl radical to a primary hydroxyalkyl radical, centered at C2, opening the pyranose ring. It is postulated that the C2 radical, at -320 K, abstracts a hydrogen from C5, re-forming the C5 secondary oxyalkyl radical. Thus a chain reaction i s propagated through the crystal until terminated in yet another free radical product. This latter free radical has been characterized by ENDOR but its structure is uncertain.
Introduction Formation of a primary hydroxyalkyl (pHA) radical is a common event in crystalline carbohydrates containing a primary alcohol group. On warming t o higher temjperatures, the pHA radical reacts to form other more stable free radicals. The structure of these secondary radicals varies in different crystals. For example, in 5-chlorodeoxyuridine the pHA radical decay, UTider slow warming conditions, yields a radical due to hydvogen addition at the base.l In contrast, decay of the pHb, radical in 3'-cytidlylic acid (3'-CMP) leads to an allylic radicaL2 Formation of the allylic rad'ical in 3'-CMP involves a number of free radical interme&iates. The first three steps, proposed in an earlier stucly, are as follows:2 0022-3654/80/2084-17 12$0 1.OO/O
H
3'CMP
H
OH
-270K %*;t. H
*t;.
77K
H H
. oo
@ OH PHA '0 W
Y
P
OH
dpHA sOA
P
OH
dpHA
t
- Hik al
lyl
The identification of the pHA radical is certain; however, 0 1980 American Chemical Society
Single Crystal Methyl a-i>Glucopyranoside
The Journal of Physical Chemistry, Vol. 84, No. 13, 1980
1713
there is ambiguity in the deprotonated primary hydroxyalkyl (dpHA) radical assignment. Structure I is an al-
I ternative assignment also consistent with the ESR data and an intermediate believed to be important in the aqueous radiolysis of some monosaccharides. The feasibility of the above reaction sequence is made more uncertain by the second reaction involving a 1,2 hydrogen transfer since such a transfer seems to be without precedence. The mechanism by which the allylic radical is formed in 3’-CMP is important to understanding how single strand breaks may be formed in DNA. For this reason we have searched for related free radical events in carbohydrates containing a primary alcohol group. Reported here is ,a sequence of reactions in methyl aD-glucopyranoside (aMeGlu) crystals. The results show that the dpHA radical decays via a 1,2 hydrogen shift, demonstrating that a reaction sequence similar to that proposed for 3’-CM:P is feasible. After the 1,2 hydrogen shift in LvMeGlu other reactions ensue as part of a chain reaction that entailri ring opening. The mechanisms involved in these latter reactions are not conclusively determined; however, they appear to differ from the latter steps of allylic radical formation in 3’-CMP.
1
0
40
20
40
60
TIME (MINUTES) Figure 1. Plot of line heiiht of C6dpHA radical (D) and C5-sOA radical (S) as a function of time with the crystal temperature maintained at 190 K; plot of line height of C5-sOA radical (S) and CPpHA radical (I) as a function of time with the crystal temperature maintained at 230 K.
€
I
E
i
Y
Materials and Methods The same procedures were used here as in the 77 K study of ~vMeGlu.~ However, after irradiation at 77 K the crystals were slowly warmed until a free radical transformation was observed. The sample was maintained at the transformation temperature until the reaction was complete, then returned to 77 K for ESR and ENDOR measurements. The temlperature was continuously monitored by using an Ag:Fe vs. chrome1 thermocouple. Polycrystalline a1VeGlu-2,3,4,6,6,-dS, obtained frorrn Merck and Co., wa13 recrystallized several times from deuterium oxide, to convert OH to OD groups, and then ground and pressed into pellets for subsequent irradiation at 77 K. The dose for the crystals and pellets was 5 Mrd, using a Machlett OElG60 X-ray tube operating at 50 kV and 20 mA.
Results and Analysis Four free radicals have been identified in a-MeGlu crystals X-irradiated and observed at 77 K: (1) an 0 6 primary alkoxy radical, (2) an 0 2 secondary alkoxy radical, (3) a C6-centered deprotonated primary hydroxyalkyl radical (C6-dpHA), and (4) a C5-centered secondary oxyalkyl radical ( C ~ - S O A ) .Upon ~ slow warming to room temperature, four distinct changes are observed: First, by 140 K both alkoxy radicals are annealed out. Second, at -190 K the C6-dpHA radical disappears, with concomitant growth of the C5-sOA radical. The C5-sOA radicals thus formed are indistinguishable from the initial population produced at 77 K. Third, at -230 K the C5sOA radical develops an additional coupling that is exchangeable, and then rapidly converts to a new primary hydroxyalkyl radical. Finally, at 320 K the ESR spectrum changes concurrently with a change in crystal appearance, the crystal becoming milky and susceptible to fracture. After 12 h at room temperature, the ESR spectra are dominated by a inew species that is stable at room temperature. The time course of these last three con-
-
-
TIME (MINUTES) Figure 2. Plot of line height of CPpHA radical (I) and stable room temperature radical (X) as a function of time with the crystal maintained at 323 K.
versions is shown in Figures 1 and 2. A. Conversion of the C6-dpHA Radical to the C5-sOA Radical. A unirnolecular C6-dpHA C5-sOA radical conversion entails a 1,2 hydrogen shift. To determine whether or not the conversion is unimolecular an isotopic-labeling experiment was performed. LvMeGlu, deuterated at nonexchangeable 2,3,4,6, and 6’ positions and at all hydroxyl positions, was X-irradiated as a polycrystalline pellet a t 77 K and warmed to induce the C6-dpHA C5-sOA conversion. Reaction 1 shows the deuterium-+
-
C 6-d pHA CS-sOA proton positions expected for a 1,2 shift. In a fully protonated crystal the C5-sOA radical has !A: values of H6 = 88 MHz, H6’ = 12 MHz, and H4 = 93 M H z . ~If a shift occurs in the deuterated material, two sets of Ai& values could result. If D6 is replaced by H5, the values would be H6 = 88 MHz, D6’ = 2 MHz, and D4 = 13 MHz. If D6’ is replaced by H5, the values would be D6 = 12 MHz, H6’ = 12 MHz, and D4 = 13 MHz. The powder spectra of both unlabeled and labeled LvMeGlu at 200 K are shown in Figure 3. The spectrum of unlabeled aMeGlu consists of 1:2:1 triplet, due to the two large P proton hfcs. The spectrum of deuterium labeled aMeGlu spectrum consists of a pair of 1:l:l triplets; the large
12
1714
The Journal of Physical Chemktry, Vol. 84, No. 13, 1980
Madden and Bernhard
2.00
g =
I
T
h'
50 U
40 30
20
0
n
b
Z
H0
b
a
C
W
DIRECTION
Flgure 5. ENDOR transition frequencies for the CP-pHA radical as a function of static magnetk fleld dlrection: (0) experimental points for the a proton; (X) experimental points for the /3 proton. Solid lines are theoretical values generated from the hyperfine coupling tensors in Table
I. Figure 3. X-band ESR spectra of polycrystalline aMeGlu-2,3,4, 6,6'-d5 grown from D20 (top) and unlabeled aMeGlu grown from D,O (bottoni). Spectra recorded at -200 K.
TABLE I: Hyperfine Coupling and g Factor Interactions in the Primary Hydroxyalkyl Radical (C2-pHA) direction cosines
principal value,
MHz
a
Ha
Amh, 16.3 ir 0.1
C
Aia
0.801 3 0 ( ! ~ 1 ) , ~ 30(%l)e 0.896 -0.292 0.334 30(+1),' 30(ir 1 )p 0.431 0.753 -0.497 1 0 ( i 0 . 5 ) , d 10(*0.5)g
. .
Ha
A h t , 48.5 i 0.1 Ha A,,,
b
83.3 i 0 . 2
IT
-0.107
0.589
0.42
0.889
49.4
A:%, 14.6 i: 0.1 HP A .~ h t., 18.2 i: 0.1 .A,,,, HB 31.0 t 0.1 9.1
0.818 -0.46 -0.396 -0.016
Gt'z
b 0x1s
0.194
0.436 0.918 1 0 ( + l ) , h 82(%1)1
/tho, HP 21.3
I ,
42,' 32' 0.73 - 0 . 4
gimin,
0.55
2.0018 (t0.0005) 0.5
gint >
0.88
0.0
2.0027 ( i 0 . 0 0 0 5 ) 0.50 - 0 . 2 2.0045 ( i 0 . 0 0 0 5 ) OH coupling range 0-1 7 MHz
gmax 7
0.84
-
V
v t
g =
2.00
Flgure 4. X-band ESR spectra of the C2pHA radical, with H, parallel to b. Top spectrum is from H,O grown crystal and bottom is from D 2 0 grown crystal.
splitting is a proton hfc of -90 MHz and the smaller triplet coupling is a /3 deuteron hfc of -13 MHz. It can be concluded that the H5 hydrogen has shifted to the H6 position on the same molecule. If abstraction of H5 from an adjacent molecule had occurred, then hfs to three 0 deuterons and no /3 protons should have been observed. The transfer appears to be stereospecific, occurring to the H6 but not the H6' position. B. The 273 K Primary Hydroxyalkyl Radical. ESR spectra of the 273 K primary hydroxyalkyl radical are shown in Figure 4. Both ENDOR (Figure 5) and ESR data show hyperfine couplings (hfc) to an a , p, and hydroxyl proton. Analysis of the CY proton ENDOR data was
a Difference angle between the torsion angle or principal axis calculated from ESR or ENDOR data and the crystalIn-piane bisector of lographic information indicated. LC3-C2-02 ( V l ) . Perpendicular t o C3-C2-02 plane (V2). V1 X V2. e In-plane bisector of LC3-C4-04 (V3). Perpendicular t o C3-C4-04 plane (V4). g V3 x V4. 62.H3. Difference between torsion angle calculated by using eq 1 with B , = 126 MHz, B o = 0, and the torsion. angle defined by L ( 02-C2-C3)-C2-C3-H3, C4. . ,H3. Difference between torsion angle calculated by using eq 1 with B , = 126 MHz, B o = 0, and the torsion angle defined by 1(04-C4-C3)-C4-C3-H3. .
2
performed by using both a positive sense and a negative sense of rotation for the third data set? One tensor yielded principal values inconsistent with an a proton and was rejected. The a proton tensor selected and the /3 proton tensor are given in Table I. Using atomic positions from neutron diffraction data, we considered a series of possible sites for the primary hydroxyalkyl radical. The a proton hfc tensor compares favorably with either a C2 or C4 centered model. For the C2 model a 30' rotation about the C2-C3 bond gives the observed directions of the and A$ principal axes. The C4 model requires a 30'
The Journal of Physical Chemistry, Vol. 84, No. 13, 1980 1715
Single Crystal Methyl a-IEGlucopyranoside g =
^u
2.00
401
L
"
30
x
'
N
1
O
b
85 1 7
b axis
L
n
w
b
a
C
H@ D I R E C T I O N
Flgure 7. ENDOR transitlon frequencies for the stable room temperature radical as a function of magnetic field dlrectlon: (0)experimental points for AH@1;(X) experimental points for AHz; (I) experimental points for exchangeable coupling. Solld lines are theoretical values generated fr 7 the hyperfine coupling tensors in Table 11.
TABLE 11: Hyperfine Coupling Interactions in the Stable Room Temperature Radical direction cosines Figure 6. X-band ESR spectra of the stable room temperature radical, with H, parallel to b. Top spectrum from H20 grown crystal and bottom from D20 grown crystal.
rotation about the C4-C3 bond. The isotropic component of the /3 proton hfc can be estimated from eq 2 where p is the spin population on the
AEJ = p(Bo
+ Bz cos2 0)
principal value, MHz
a
b
Amax, HP 1 37.8 ( k 0 . l )
-0.526
0.381
0.850
A i ! ' , 26.1 (k0.1) Am,,-,, HQ 1 24.8 (kO.1) HP 29.6 Aho',
0.45
0.86
0.25
0.72
-0.52
0.46
H, A,,,
0.78
-0.08
-0.62
0.6
0.5
0.6
0.3
0.8
-0.5
6.1 (kO.1)
H
Ah;, 3.0 (k0.2) H
a carbon atom, Bo N 0 MHz, B2 N 126 MHz? and 0 is the
A&,
torsion angle between the LEO (lone electron orbital) symmetry axis and the C-H, bond. The torsion angles for the C2 and C4 models were calculated by using the AH" 'ftt axis as the LEO axis. Both models ive 0 = 60° and Ais! = 28 MHz. Though the observed Aln{value of 21 MHz is accommodated by either model, the direction of A%x is not. The maximum 10 proton hfc is expected along the C-.H, direction. It occurs 10' from the C2-aH3 direction and 8 2 O from the C:4-.H3 direction. The C2 model, therefore, is in better agreement with the /3 proton hfc tensor. Additional support for the C2 model is obtained by calculating the expected hydroxyl proton coupling. The isotropic component of AoH is adequately described by eq l5with Bo = 0 MHz and Bz = 67 M H Z . ~The torsion angle between the LEO and the corresponding O-H bond is found, as above, by using the direction as the LEO symmetry axis. For the C2 center, 0 equals 72'; whereas for the C4 center, 0 equals 3O. These correspond to A: values of 6 and 67 MHz, respectively. Experimental values of AoH could not be measured at sufficient orientations for an accurate tensor de termination. At orientations where measurements were possible the hydroxyl proton coupling ranged from 0 to 17 MHz, favoring the C2 model. This primary hydroxyalkyl radical (C2-pHA) is, therefore, assigned to Structure 11. -C3 (OH)H g C 2 (OH)H, I1 C. Stable Room Temperature Radical. The stable room temperature radical has an ESR spectrum characterized by a broad, isotropic doublet, with an additional 16-MHz exhangeable coupling visible when Ho is parallel to the b
A,: 3.8 AoH = 16.0 along the b axis AoH = 7.6 along the c axis
8
2 . 3 (k0.3)
C
axis (Figure 6). Three hyperfine couplings were followed by using ENDOR. The data are shown in Figure 7 and the principal values and axes for two of the hfc tensors, corresponding to the nonexchangeable protons, are given in Table 11. A complete analysis of the exchangeable hfc and g factor could not be accomplished from these data. The g factors along the a , b, and c axes are 2.002, 2.000, and 2.003. The degree of anisotropy in the two nonexchangeable proton hfc's is useful in deducing the structure of the stable room temperature radical. The lafger splitting has principal values of a /3 hydrogen in a C-CH, fragment. The anisotropy, AH@l- A2;, is 13.0 MHz. From the work of KO and Box,Pax it appears that A Z x - Ai!,, varies only slightly with changes in A%, going from -16 MHz for Ai, = 10-40 MHz to -12 MHz for Ais,, = 70-110 MHz. The anisotropy ought to provide, therefore, a gauge of the spin density on the carbon having the broken bond. In our case = 29.6 MHz and an estimate of p is 13.0/16 = 0.8. The small nonexchangeable hfc has an unusually small anisotropy, A?= - A a n = 3.8 MHz. This dipolar component is too small for a typical /3 hydrogen lying in the nodal plane of the lone electron orbital (LEO). It is also too small relative to the isotropic component to be explained by a distant proton. One explanation we suggest is that hyperfine interaction is due to an aldehydic proton. It is known from aqueous solution work that such a proton has a hfs with a small isotropic component. For example, 1.4 MHz was observed for the aldehydic proton in H,CCHO.'
1716
The Journal of Physical Chstnistty, Vol. 84, No. 13, 7980
Madden and Bernhard
Scheme I ~HOH 77K
C6-pHA
OCH,
I
uMeG I u
-H+
CHO@
-230K
-190K
OCH,
OCH,
CG-dpHA
C5-sOA
-H2
H
CHZOH
o
t
H
-320K HO
OCH,
C 1 -MeOA
C2-pHA
The exchangeable hfs, for which only a partial set of ENDOR data was obtained, is most likely due to a hydroxylic proton. The hfc’s of 16.0 and 7.6 MHz measured along the b and c axis, respectively, indicate that the OH bond must lie near the LEO nodal plane. The free radical model that is constructed from the above observations is structure 111. H
1.1
-c-c-c
I I
OH OH
Discussion Our current model for the free radical reactions stemming from electron abstraction at the primary alcohol group of aMeGlu is shown in Scheme I. The first four steps leading up to the C5-sOA radical are based on fairly definitive information. The subsequent chain reaction, Cl-MeOA C2-pHA CS-sOA* etc., C5-sOA* involves some conjecture on detail, but there is evidence that this type of sequence occurs. The last step of the scheme suggests a possible assignment for the radical that is stable a t room temperature. The assumption that the C6-dpHA radical in aMeGlu is formed via electron abstraction comes from observations made on the nucleotide 3’-CMP.2 The pHA radical in 3’-CMP must be formed by either electron abstraction or superexcitation. Superexcitation could lead to a homolytic cleavage of the C(5’)-H bond giving the pHA radical and an H atom. However, it seems unlikely that homolytic cleavage occurs because H atoms are known to react with C5 of the cytosine base and the resulting C5-H addition radical is not observed.$ Furthermore in 3’-CMP, the cytosine base serves as an excellent scavenger of electrons. This leaves e- abstraction as the mechanism of formation for the pHA radical in 3’-CMP and, as shown in reaction 1, the pHA radical probably deprotonates to give the dpHA radical. In aMeGlu the dpHA radical is the first step actually trapped. Nevertheless, we assume the cation
-
-
C6-pHA C6-dpHA sequence pertains, as in 3’-CMP. The 1,2 hydrogen shift, from C6-dpHA to C5-sOA, is demonstrated by the deuterium labeling experiments. The C6 shift is such that stereospecificity of the H(C5) H(C5) migrates to a position at C6 which, of the two C6-H bonds, is the one that ends up more highly conjugated. It is also important to note that the negative charge on the primary alcohol group may be a critical factor in promoting the 1,2 shift. In a-D-glUCOSe crystals a pHA radical is formed at C6. It does not deprotonatelO and does not partake in a 1,2 hydrogen shift.” The next step in Scheme I, C5-sOA CS-sOA*, is one explanation for the observed exchangeable hfs that occurs just before the C5-sOA spectrum converts to the C2-pHA spectrum at -230 K. The placement and role of this added proton has not been experimentally determined. The C1-methoxyalkyl (Cl-MeOA) radical is not observed. It is an expected intermediate based on the general mechanism of R,COR’ RzCO + R’, proposed by Kuhn and Wellman12 and employed by Dizdaroglu et al.13 to explain various end products formed by the dissolution of y-irradiated polycrystalline a-D-glUCOSe. Although formation of C1-MeOA intermediate seems rather certain, it is not clear how the C2-pHA radical is formed. The possibility shown in Scheme I entails a 1,2 methyl shift accompanied by cleavage of the Cl-C2 bonds. A chain reaction must occur at temperatures around 320 K in order to account for the change to a milky appearance and loss of mechanical strength in the crystal. In the aMeGlu crystal lattice, C2 is 3.1 8, from H5 of an adjacent m~lecule.’~ Abstraction of this H5 atom would result in the formation of a new C5-sOA* radical and a chain reaction would ensue, This reaction would propagate along the crystallographic c axis until an imperfection interrupts the chain. At this point a termination step leads to the stable room temperature radical. Unfortunately the structure (shown in Scheme I) assigned to the stable room temperature free radical is tenuous. This means the hydrogen elimination step, as a termination reaction, is speculative. It can be stated, however, that 0-hydroxyl elimination does not occur be-
-
-
\H
I11
-
-
0
( ?) stable RT radical
-
-
-
J. Phys. Chem. 1980, 84, 1717-1724
cause it would lead to a radical containing an (Y hydrogen hfc, which is not observed. P-Hydroxyl elimination is known to readily occur in aqueous solutions15but it is not obvious that this reaction is as facile in solids. Additional work with isotopic labels will be necessary in order to establish the structure of the stable room temperature radical and test the validity of the suggested termination reaction. On the other hand, the shift of a hydrogen from C5 to C6 of the dpHA radical opens up the possibility that such a shift occurs in 3’-CMP, as suggested earlier, and perhaps even in DNA. In the case of DNA, this would mean that a radical formed at C5’ will shift the unpaired electron population to C4’. And it is the C4’ centered radical thlat is believed to be an important intermediate in the concerted cleavage of both the 3’ and 5’ phosphodiester bonds of DNA.16p2 Acknowledgment. The authors acknowledgethe technical assistance of Kermit Mercer during these studies and thank Mason Gross for critically reading this manuscript. K.P.M. gratefully acknowledges a fellowship from the University of Rochester Department of Radiation Biology and Biophysics, and support from Mrs. Florence H. Madden. This work was partially performed under contract with USDOE at the University of Rochester De-
1717
partment of Radiation Biology and Biophysics, assigned Report No. UR-3490-1687, and partially under NSF Grant NO. PCM77-16830.
References and Notes (1) J. Huttermann, W. A. Bernhard, E. Haindl, and 0. Schmidt, J . Phy. Chem., 81, 228 (1977); H. Oloff, E. Haindl, J. Huttermann, and J. Krauss, Radiat. Res., in press. (2) W. A. Bernhard, J. Hiittermann, A. Muller, D. M. Close, and 0. W. Fouse, Radiat. Res., 68, 390 (1976). (3) K. P. Madden and W. A. Bernhard, J. Chem. Phys., 70,2431 (1979). (4) D. S. Schonland, Proc. Phys. Soc. London, 70, 788 (1959). (5) H. Muto, K. Nunome, and M. Iwasakl, J . Chem. Phys., 61, 5311 (1974). (6) W. A. Bernhard, D. M. Close, K. R. Mercer, and J. C. Corelll, Radlat. Res., 66, 19 (1976). (7) C-L. KO and H. C. Box, J . Chem. Phys., 68, 5357 (1978). (8) R. Livingston and H. Zeldes, J. Am. Chem. Soc., 88, 4333 (1966). (9) W. A. Bernhard and D. M. Close, unpublished results. (10) K. P. Madden and W. A. Bernhard, J. Phys. Chem., 83, 2643 (1979). (1 1) K. P. Madden and W. A. Bernhard, unpublished results. (12) L. P. Kuhn and C. Wellman, J . Org. C h m . , 22, 774 (1956). (13) M. Dizdaroglu, D. Henneberg, K. Neuwald, G. Schomburg, and C. von Sonntag, 2. Naturforsch. B, 32, 213 (1977). (14) G. A. Jeffery, R. K. McMullen, and S. Takagi, Acta Ctystallogr.,Sect. B, 22, 728 (1977). (15) See references In C. von Sonntag, in “Effects of Ionlzing Radiation on DNA”, J. Huttermann, W. Kohnlein, R. Teoule, and A. J. Bertinchamps, Eds., Springer-Verlag, New York, 1978, p 204. (16) M. Dizdaroglu, C. von Sonntag, and D. Schulte-Frohlinde, J. Am. Chem. Soc., 97, 2277 (1975).
ESR Study of Ithe Mechanism of an Intermolecular Cation-Exchange Reaction between 3,5-Dinitrobenzonitrile-Sodium Ion Pair and Sodium Tetraphenylborate Mario Barzaghi, Cesare Ollva, and Massimo Simonetta’ C.N.Ff. Center for the Study of Structure/ReactivityRelations and Institute of Physical Chemistry, University of Milan, 20133 Milan, Italy (Received January 18, 1980)
Analysis of the ESR spectra of 3,5-dinitrobenzonitrile-sodium in THF containing sodium tetraphenylborate indicates the formation of triple ions, which are the intermediates of an intermolecblar cation-exchangereaction in which the incoming cation goes to the uncomplexed nitro group. The corresponding kinetic and thermodynamic parameters have been derived through a nonlinear least-squares line-shape fitting of the experimental ESR spectra followed by an accurate analysis of their temperature and concentration dependence. The spin distribution in both the ion pair and the triple ion has been computed by the McLachlan method modified according to McClelland. The transport properties of the system 3,5-dinitrobenzonitrile-sodium/THF/sodium tetrapheriylborate have also been investigated. The possible use of the ionic aggregates of aromatic radical anions as spin probes in the study of concentrated solutions of electrolytes is investigated.
Introduction I t is well established that reduction of aromatic nitro derivatives by alkali metal in ethereal solvents yields strongly associated ion pair~.l-~ As the dielectric constants of ethereal solvents are sufficiently low, the ion-ion interactions are more important than ion-solvent interactions; and contact ion pairs, triple ions, quadrupoles, and even multiple clusters can be stable and numerous enough to affect the conductance6and the spectroscopic behavior of the solution^.^^'^^ Formation of triple ions is greatly favored by addition of salts with a common cation5i7t8and depends upon the dissociation constant of the added salt as well as the equilibrium constant for triple-ion formation. In this case, triple ions are the intermediates of intermolecular cation exchange between the radical pairs and the added salt. It is quite easy to envisage the formation of 0022-3654/80/2084-1717$01 .OO/O
a triple-ion intermediate in intermolecular cation exchange when the anion presents two polar sites, because the relative motions of the counterions produce line-width alternation effects in the ESR s p e ~ t r a . ~ , ~ ~ Following research on solution dynamics of free-radical anions and ion pairs,5 in this paper we investigate the rate and the mechanism of sodium exchange between the 3,5dinitrobenzonitrile (DNBN)-sodium ion pair and sodium tetraphenylborate (NaBPh,) in tetrahydrofuran (THF). Our aim is to establish whether the ionic aggregates of aromatic radical anions can be used as spin probes in the study of concentrated solutions of electrolytes.
Experimental Section DNBN (Aldrich) was used without additional purification. NaBPh, (Aldrich) was dried at 150 “C overnight 0 1980 American Chemical Society