Maeda, lngold
/
Spin Trapping of Secondary A l k y l Radicals
In contrast, the kinetic energy release distributions for the loss of CO from furan and the two pyrones show that a wide range of kinetic energies is released upon fragmentation. This implies that considerable excess energy is required for these [C4H40]+- ions to rearrange to the transition state for the dissociation step. Indeed, the loss of CO from [furan]+. has such a high activation energy2' that the energy level of the transition state (-280 kcal mol-') makes thermochemically feasible the participation of many isomeric [C4H40]+structures (cf. Table I). Therefore, for these isomeric [C4H40]+- ions no firm deductions can be made from metastable peak characteristics concerning the structures of either the reacting or the nonreacting ions. Pirkle and Dine's deuterium labeling experim e n t ~elegantly ~ demonstrate that the metastably fragmenting (but energy rich) [C4H40]+- ions from 2-pyrone d o not undergo reversible ring closure. However, this conclusion may not be extrapolated to the structure of the ions which give rise to the intense peaks at m/z 68 in the normal 70-eV mass spectra of both pyrones. These peaks largely consist of [furan]+. ions. Experimental Section The collisional activation and the metastable ion (MI) spectra were measured on a Vacuum Generators ZAB-2F double-focussing mass spectrometer of reversed Nier-Johnson geometry. A 100-pA ionizing electron beam of 70 eV and an accelerating potential of 7920 V were used; sample reservoir and ion source temperatures were - I 30 O C . The magfietic field was set to select the m/z 68 [C4H40]+* precursor ions; ionic products of their metastable decompositions (the MI spectrum) occurring in the field free drift region between the magnetic and electrostatic analyzers were measured by scanning the electrostatic analyzer (ESA) potential under conditions of good energy resolution. (Main beam width was 0. I O V at an ESA voltage of 422. I V.) Kinetic energy releases were calculated in the usual way and corrected for limiting half-height width of the main beam. C A spectra were obtained in the same way after introduction of helium as a collision gas in the collision chamber near the p focal point. An external voltage (-320 V) was applied on the chamber to separate the normal metastable peaks from the C A peaks.
4975
Acknowledgment. W e thank Dr. F. P. Lossing for measurements of ionization and appearance energies and Dr. W. H. Pirkle for a gift of 2-pyrone. J. L. Holmes thanks the National Research Council of Canada for continuing financial support. References and Notes (1) (a) University of Ottawa; (b) University of Utrecht. (2) (a) J. L. Holmes, lnt. Rev. Sci.. Phys. Chem. Ser. Two, 5, 241 (1975); (b) F. W. McLafferty and W. T. Pike, J. Am. Chem. SOC.,89, 5954 (1967). (3) W. H. Pirkle and M. Dines, J. Am. Chem. SOC.,90, 2318 (1968). (4) F. P. Lossing and J. C. Traeger, Int. J. Mass Spectrom. /on Phys., 19, 9 (1976). ( 5 ) S.W. Benson, F. R. Cruickshank. D. M. Golden, G. R. Haugen, H. E. O'Neal, A. S.Rodgers, R. Shaw, and R. Walsh, Chem. Rev., 89, 279 (1969). (6) J. L. Holmes, A. D. Osborne and G. M. Weese, lnt. J. Mass Spectrom. /on Phys., 19, 207 (1976). (7) J. L. Holmes and A. D. Osborne, lnt. J. Mass Spectrom. /on Phys., in press. (8)These E,, values are almost certainly too low; the calculation assumes that allof Ere" is measured from Tmi,, Le., that all of Ere" is partitioned into translational degrees of freedom. (9) H. M. Rosenstock, K. Draxl, B. W. Steiner. and J. T. Herron, J. Phys. Chem. Ref. Data, Suppl. 1, 6, (1977). (10) J. L. Holmes, J. K. Terlouw, and F. P. Lossing, J. Phys. Chem., 80, 2860 (19761 > - -,.
(1 1) J. L. Holmes, J. K. Terlouw, P. C. Vijfhuizen, and C. A'Campo, Org. Mass Spectrom., 14, 204 (1979). (12) J. K. Terlouw, P. C. Burgers and J. L. Holmes, J. Am. Chem. Soc., 101, 225 (1979). (13) K. Levsen and H. Schwarz, Angew. Chem., lnt. Ed. Engl., 15, 509 (1976). (14) 0. L. Chapman, C. L. McIntosh, and J. Pacansky. J. Am. Chem. SOC.,95, 244 (1973). (15) R. G.S. Pong and J. S.Shirk, J. Am. Chem. SOC.,95, 248 (1973). (16) A. Krantz, J. Am. Chem. SOC.,96, 4992 (1974). (17) E. Hedaya, R. D. Miller. D. W. McNeil, P. F. D'Angelo. and P. Schissel. J. Am. Chem. SOC.,91, 1875 (1969). (18) T. W. Bentley, Spec. Period. Rep.: Mass Spectrom., Chem. SOC.London, 4 (1977). (19) P. Brown and M. M. Green, J. Org. Chem., 32, 1681 (1967). (20) J. L. Holmes and A. D. Osborne, Int. J. Mass Spectrom. /on Phys., 23, 189 ( 1977). (21) AE ([CJH,]+.) from furan = 11.80 eV, measured with energy selected electron^.^ Note that this energy lies above the calculated thermochemical and AEthresholds for production of CO ( A h = -26.4 kcal ([CH2=C=CH2]+$ = 10.88' or [CH&=H]+. = 11.48 eVg-and even AE([cyclopropene] = 11.74 eV.9 The discontinuity in the energy release distribution (Figure 1) may reflect the co-generation of two different product ion structures. The large activation energy may represent that required for ring opening. a)
Kinetic Applications of Electron Paramagnetic Resonance Spectroscopy. 34. Rate Constants for Spin Trapping. 2. Secondary Alkyl Radicals' Y. Maeda2 and K. U. Ingold* Contributionfrom the Dii+sion of Chemistry, National Research Council of Cunada, Ottawa, Ontario, Canada K I A OR6. Received January 15, 1979
Abstract: The cyclization of the secondary 6-hepten-2-yl radical to form the primary 2-methylcyclopentylcarbinyl radical has been studied by kinetic EPR spectroscopy. The rate constant for this cyclization, k,, can be represented by log ( k , / s - ' ) = (9.8 f 0.3) - (6.4 f 0.3)/0 where 8 = 2.303RT kcal/mol. This cyclization has been used as a standard to measure the absolute rate constants for the spin trapping of secondary alkyls in benzene at 40 O C by a few of the more commonly employed spin traps. These results are compared with the previously determined4 kT values for trapping primary alkyl radicals under the same conditions. Relative k T values for trapping of the tertiary 2-phenylprop-2-yl radical have also been measured.
Spin trapping3 is an EPR spectroscopic technique that has proved immensely valuable in studies of reaction mechanisms because it allows transient radicals, R., to be "visualized" by reacting them with a spin trap, T, to form a persistent spin adduct, RT.: AT
R.+T+RT* 0002-7863/79/ 1501-4975$01.OO/O
As we have pointed out p r e v i ~ u s l y spin , ~ trapping will only achieve its full potential when quantitative data on the rate constants, kT, for trapping of all types of R- radicals become available. In our initial studies of spin-trapping rates4s5 we worked with primary alkyl radicals. I n the present paper, we report an extension of this work to secondary alkyls. Our experimental procedure is almost unchanged. In the earlier
0 1979 American Chemical Society
4976
Journal of the American CheniicalSociety
w ~ r kwe ~ relied . ~ primarily on the cyclization of the 5-hexenyl radical and distinguished between the spin adducts of the unrearranged and rearranged primary alkyls by means of I3C labeling. In the present work we have relied primarily on the cyclization of the secondary 6-hepten-2-yl radical, S., to form
S'
P'
the primary 2-methylcyclopentylcarbinyl radical, P.. I n the present case, there is no need to distinguish the unrearranged and rearranged radicals by isotopic labeling since these radicals are self-labeled by having one or two hydrogens, respectively, a t their radical centers. This rearrangement is known to be irreversibleh and to occur a t a rate that is slightly faster than the 5-hexenyl c y c l i ~ a t i o n .I~n. ~the presence of a trap two spin adducts will be formed:
+ T +ST. P. + T PT.
S-
Xr
--*
The rate constant for spin trapping is given by4
+
+
-
-
Experimental Section Materials. Diacyl Peroxides. Di(2-methyl-6-heptenoyl) peroxide ( 1 ) was prepared starting with 5-bromo-I-pentene and diethyl methylmalonate. Dry ethanolI0 (57 mL) was refluxed for 2 h over 3.9 g of clean sodium cut into small pieces. After cooling, diethyl methylmalonate (29.6 g, 0. I7 mol) was added, followed by 5-bromo- 1 -pentcne (25 g, 0. I68 mol). This mixture was refluxed for 3 h and cooled. a n d 200 niL of H2O was added. Extraction with ether followed by distillation yielded 37 g of crude diethyl 4-pentenylmethylmalonate (yield 90%). This was purified by preparative VPC ( I O ft 20% O.V. I01 colunin at 180 "C) to remove small amounts of impurities (chiefly diethyl 4-pentenylmalonate) and 22.3 g of extremely pure (>99.9%) diethyl 4-pentenylmethylmalonate was recovered. ' H N M R (CDC13, 6 values downfield from kle4Si): 1.1-2.3 ( t m. 15 H. CH3 and CCH2C), 4.2 (q, 4 H, OCHl), 4.9-5.2 (m, 2 H, HlC=C), 5.5-6.2 ( m . 1 H,C=CH). This malonic ester (21.8 g, 0.09 mol) was hydrolyzed by heating for I2 h a t 120 OC with 30 g of KOH in 100 mL of water.I1 The aqueous phase was acidified ( 2 N HCI) and extracted with ether. the extract was dried, and, on removal of the solvent, 15.8 g of solid 4pentenylmethylmalonic acid was obtained (yield 94%), mp 83.2 OC.
+
101:17
/ August
15, 1979
' H Y M R (CDCI3):d 1.1-2.3 ( s + ni.9 H,CH?andCCtl?C).4.9-5.2 H . H?C=C). 5.5-6.2 (m. I H, C=CH), 12.0 (s. 2 H. O H ) .
(111. 2
The malonic acid ( I5 g) was heated to 250 OC and the decarboxylatcd 2-methyl-6-heptenoic acid ( b p 243 "C) was distilled off continuously. ' H U M R (CDC13):6 1.1--1.2 ( d , 3 H.CH3). 1.3--2.8( m . 7 H , C H and CCH?C),4.9--5.2( m , 2 H, H?C=C). 5.5-6.2 ( m , I H. C=CH). 12.1 (s. I H. O H ) . T h e 2-methql-h-hcptenoic acid \vas converted to di(2-methyl-6'H heptenoyl) peroxide ( 1 ) bq the carbonyldiiniida7ole WMR (CDCIj): 6 1.1-1.2 (d, 3 hi. CH3). 1.3-2.8 ( m , 7 H, C H and CCH?C), 4.9-5.2 (m, 2 H, H?C=C), 5.5-6.2 (m. I H, C=CH). Titration:I3 86 f I%peroxide. Di(2-methyl-6-heptanoyl) peroxide ( 2 ) was prepared by ;I similar procedure starting with I-broniopcntanc. ' H N M R (CDC13): d 0.9-1.9 (d, m, 28 H, C H j , CCH2C). 2.1-2.6 (m, 2 H, C H ) . Titration:" 92 f 2% peroxide. 2-Bromo-6-heptene (3)was prepared starting with 5-bromo- I pcntene (25 g) which was converted (with 4.3 g of Mg in I25 inL of cthcr) to the Grignard and this was treated with 12.3 g of freshly distilled ncetaldehqde. Conventioxil workup folloued by distillation jyve 10.8 g of pure 6-hcpten-2-01 (yield 59%). Reaction of this alcohol with PBr3 according to the method of Johnson and Ow.yangt4gave, after distillation, 2.8 g of 2-bromo-6-heptene (yield 28%'. purity 96'%), bp 72 O C at 20 Torr. After prepnrative VPC (10 ft. 107~O.V. 101 column a t 130 "C) I .2 g of 99.8% purc bromide was recovered. ' H Y M R (CDC13): 6 1.5-2.2 (d, m. 9 H. CH3 and CCHlC), 4.0-4.4 (scxtet. I H , CH), 4.8-5.1 ( i n , 2 H , H?C=C). 5.4-6.2 (m, I H ,
c=c H ).
In deriving this equation it is assumed that all of the alkyl radicals formed are captured by the trap.4 This means that the rate of all other processes which lead to their consumption must be much less than the rate of trapping. In the early stages of the reaction only bimolecular reactions between alkyl radicals (rate constants ca. IO9 M-' s - ' ) ~might be able to compete with t r a ~ p i n gAn . ~ "effective" trap will therefore be one for which kT[T]([S-] [Pa]) >> IO'([S-] [P.])2andsince trap concentrations can hardly be greater than I M this means that kT >> IO9( [S.] [PSI).The limit set by this condition depends on the rate, r i , a t which alkyl radicals are produced. Quantitative spin trapping experiments are most easily done with very low rates of radical formation, i.e., with ci 1 O-x-l 0-' M s-I. This puts an effective lower limit on k T of ca. IO' M-' s - ] , though a trap with such a low kT value would be very inefficient. It is, perhaps, also worth pointing out that i f a rearrangement is to compete with spin trapping then k, kT[T]. Therefore rearrangements that have unimolecular rate constants of ca. IO' s-' can only be used to measure spin trapping rate constants with very inefficient traps. More efficient traps require faster rearrangements.
+
/
Spin Traps. The follo\\ing spin traps were either available from our carlier work4 or were prepared by the same methods! 2-methyl-2nitrosopropane (NtB), nitrosodurene (NDI, 2,4,6-tri- ferf-butylnitrosobenzene ( f - B u l N B j , 5,5-dimethylpyrroline 1-oxide ( D M P O ) , methylene-!V-tert-butyInitrone(MBIVI, and phenyl-N-terf-butylnitrone
(PBN). 2,6-Di-tert-butyl-1,4-benzoquinone (Aldrich) was purified by preparative VPC (5 ft 5% O.V. 101 column a t 160 "C) to remove a I f this was not trace (0.09%)of 2.6-di-rrrt-butyl-4-1nethoxyphenol. done, thc only EPR signal observed on reaction Rith diacyl peroxides \\as that due to 2,6-di-rert-butyl-4-1nethoxyphenox~, Xieasurement of k,. Although the tin hydride method suggests that 6-hepten-2-yl cyclizes somewhat more rapidly than 5-hexenyl,".' w e felt that it would be worthwhile to measure the rate constant, k , , more accurately and over a range of temperature. The technique employed i h that of kinetic EPR spectroscopy which w e have used previousl> to measure the Arrhenius par;inicters for the cycli7ation of 5 - h e x ~ n y I ' ~ ~ ~ ~ and 4-cyanobutl t I 7 radicals. The general experimental procedure for studying unimolecular radical reactions by this technique has been adequately described." I x General Procedure for Spin Trapping. Samples were prepared in quart7 EPR tubes b) adding aliquots of stock solutions of the spin traps to known quantities of the diacyl peroxide as described The degassed, frozen solutions were transferred to the preheated cavitj of a Varian E-4 EPR spectrometer. Spectra were recorded as a function of time and ratios of the concentrations of the different spin adducts were determined by double integration of nonoverlapping lines of known degeneracy in the first derivative spectra. The solvent was ben7ene in all cases and the radicals were produced by thermal deconipoaition of diacyl pcroxides. AI1 spin trapping rate measurements \rere made at 40 O C . Dimerization of Nitrosodurene. This spin trap is highly dimerized in bolution and since onl) the monomer will be active it is necessary to knob the equilibrium constant for dimerization. K = [dimer]/ [monomer]'. In our earlier work we obtained a value of 16 for ~ 7 , i g , the molar extinction coefficient/cm monomer a t 750 nm (which is slightly off the band maximum). and a value for K = 286 M-' a t 22 OC. Assuming that the temperature coefficient of K would be small we estiniatcd K = 250 f 25 M - l a t 40 "C. W e subsequently learned that this equilibrium constant had also been measured i n benzene at temperatures from 20 to 40 OC by Doba. Ichikawa. and Yoshida.ly They obtained (the band maximum) 91 M-I cm-' and a van't Hoff relation which can be represented as log ( K I M - ' ) = -6.80 14.5/0, where 0 = 2.303RT kcal/mol. This gives K = 2140 M-l a t 40 OC. This large discrepancy led to additional studies on this equilibrium. A vcry careful remeasurement in our laboratory a t 22 OC (the on14 temperature available for experimental reasons) gave K = 2000 f 440 M-I with t800 47 f 5 M-' cm-I. Meanwhile, measurements
+
Maeda, Ingold
/ Spin Trapping of Secondary Alkyl Radicals
4917
Scheme I
,
in Professor Yoshida's laboratory'0 gave €800 36 i 2 M-' crn-' at 29 "C and, on the assumption that is temperature independent, the van't Hoffrelation, log ( K I M - ' ) = (-8.74 f 0.10) (16.1 & 3.8)/8 at temperatures from 25 to 40 OC. This equation yields K = I565 M-' at 22 O C (which is in satisfactory agreement with our revised value considering the difficulties inherent in these experiments) and K = 322 M-' at 40 OC. I f w'e combine the two sets of data (without correcting for the different e values) we obtain log ( K I M - ' ) = - 10.72 18.9/0, which yields K = 303 M-' at 40 "C. It is this last value that we have used in our spin trapping calculations.
-+
+
+
Results Measurement of k,. The irreversible cyclization of the secondary 6-hepten-2-yl radical, S-, to the primary 2-methylcyclopentylcarbinyl radical, P-, was studied by kinetic EPR spectroscopy. The S. radicals were produced by bromine abstraction from 2-bromo-6-heptene (3) using photochemically generated tri-n-butyltin radicals in n-pentane as the solvent. Absolute concentrations (determined by calibration with DPPH)*I of S. and Pa were measured under steady photolysis a t temperatures from 183 to 232 K . The overall reaction can be represented as shown in Scheme I. Since S. and P. are sterically unhindered alkyl radicals of low molecular weight the radical-radical reactions all occur at the diffusion-controlled limit. Under such circumstances the rate constant ratio, k,/2ktP, is given k,/2ktP = [P.](l
+ [P*]/[S*])
The experimentally measured S a and P. concentrations are listed in Table 1, together with the derived values of k,/2ktP. This data can be represented by log (k,/2ktp/M) = -(2.0 f 0.3) - (4.1 f 0.3)/8 (11) The rate constant for the bimolecular self-reaction of the rearranged radical, 2ktP,was measured by the usual method" in n-pentane a t 240 K. The radical decayed with second-order kinetics and with a rate constant of 4.57 X IO9 M-' s-l, which indicates that the reaction is diffusion controlled. The variation in 2k,P with temperature can therefore be estimated from the temperature coefficient of viscosity of the solvent.'* It can be represented by
-
log ( 2 k C P / M - 's-I) = 11.79 - 2.34/8 From eq 11 and 111, the rate constant for the Sa can be represented by log ( k C / s - ' ) = (9.8 f 0.3)
(111)
P. cyclization
- (6.4 f 0.3)/0
(IV)
This equation would seem to be fairly reliable since the Arrhenius preexponential factor is in the range found for analogous cyclizations, e.g., 109.5M-' S - I for 5-hexenyllh and 109.9M-' s-I for 4-cyan0buty1.l~ Equation IV yields k , = 2.1 X I O 5 s-' a t 40 OC,slightly faster than the rate constant for the 5-hexenyl cyclization, which is 1.78 X IO5 S - I at this ternperature.l6 The ratioof these rate constants is 1.18, in good agreement with values of 1.4 a t 406 and 65 O C 7 measured by the tin hydride method.23 Measurement of kT. Trapping with 2-Methyl-2-nitroso-
1 1 ;
'
1
IOG
1 ; 1
1
~
P-NqB
S-NtE ! I Figure 1. EPR spectrum of adducts formed from 1 (6.1 X NtB ( I.3 X IO-' M ) in benzene at 40 O C .
I 1
~
~
M ) and
Table I. Radical Concentrations and k,/2ktP Ratio for the Cyclization of 6-Hepten-2-yl Radicals in n-Pentane
[S.] x
temp, K
lo8.
M 21.6 19.9 18.6 12.6 8.6 7.7 6.9 5.6 4.0
I83 I88 192 I96 205 209 214 221 232
[Pa] X IOx,
k,/2k, X IO',
M
M
9.4 7.1 10.4 12.4 19.4 19.3 19.1 22.4 22.0
1.35 0.96 1.62 2.46 6.32 6.77 7.20 11.2 14.3
propane (NtBI. Thermal decomposition of di(2-methyl-6heptenoyl) peroxide ( l ) , in benzene at 40 O C and at a con-
I s' centration of 6.1 X M in the presence of 1.3 X lo-* M NtB, gave an unexpectedly complex EPR spectrum (see Figure 1 ). Under the same conditions, di(2-methyl-6-heptanoyl) peroxide (2) gave a simple EPR spectrum consisting of a triplet
2 H ( a N = 14.8 G ) of doublets ( a H = 1.8 G). This was obviously due to the trapping of the secondary heptan-2-yl radical, He. With this guide it becomes clear that two different primary alkyls are formed from S. and are trapped by the NtB to give spin adducts having slightly different EPR parameters. The formation of two distinguishable P-NtB spin adducts is, to us, the surprising result since earlier product studies had already shown the cyclizes to give both the cis-methyl-, Pa,,,, and trans-methyl-, Pet,,,,, cyclopentylcarbinyl radicals25-27 in a S a
S.
4978
Journal of the American Chemical Society
Table II. EPK Parameters for Suin Adducts in Benzene a t 40 suin adduct" ah
O C
lO1:17
/
August 15, 1979
(Hvoerfinc Solittines in Gauss) klh
I .8 I .7 12.6 ( H a ) , 9.5 (Hb) 11.3 (Ha),' 9.8 (Hb)' 6.9 13.2 (Ha), 9.5 (Hb) 1.7 (3 H).f 22.3 19.5 (ti,), 13.5 (Hb), 0.8 ( 2 HI,,) 2.5 21.8 11.6 ( H a ) , 9.2 (Hb)
14.8 14.7 15.4 15.2 13.6 13.3 10.8 13.4 14.1
H-NtB S-NtB P,,,-NtB P,r:in,-NtB S-ND P-NDd S-1- B u ~ 8' N S-f-Bu3N Bg Nd . g P-r - B u ~ B H -- PB N H-DMPO H-MBN
/
14.6
14.2 15.2
R
2.0059 2.006 I 2.006 I 2.0061 2.0061 2.006 1 2.0041 2.0062 2.0062 h h h
[' H = heptan-2-yl, S = hepten-2-yl. P = 2-incthylcyclopentylcarbinyl:for spin adduct abbreviations see text. One hydrogen unless otherwise specified. Best values, estimated by computer simulation, Cis and trans P-adduct spectra not resolved. e Anilino spin adduct. ./ @ H + 2 H,,,,,;,. fi Nitroxide spin adduct. Not measured. Table 111. Spin Trapping 6-Hepten-2-yl. S., and the Rearranged 2-Methylcyclopentylcarbinyl, P a . Radicals with N t B in Benzene a t 40 O C
[NtB] X IO3, M 4.2 8.4 12.6 16.8 33.6
(d[S-NtB]/df), & k T X 10-S,h c l X IO9,' (d[P-NtBl/dt), +ii M-l s - ' M S-' 0.133 0.223 0.320 0.456 1.156
66.5 55.8 53.3 57.0 72.3
1.39 1.19 0.95 1.06 0.94
+
Calculated from eq I (' [P-NtB] = [P,i,-NtR] [P,,,,,-NtB]. w i t h k c = 2.1 X IO5s-'.AveragevalueofkT= (61 * 8 ) X IO5M-l s - l . " c i = d[S-NtB]/dt d[P-NtB]/dr. Initial concentration of 1 = 3.05 x 10-3 M.
+
ratio of ca. 2.7: 1 2 5 to 2.3: 1 .2h The ratio of the integrated peak areas of either of the outermost, incompletely resolved, pair of lines shown in Figure 1 is roughly 2.8: I . This allows us to assign the spectrum with the larger (by 1.4 G ) overall width to Pcl,-NtB. An additional complication in the P,,,-NtB (and also in the P,,,,,-NtB) spectrum arises because the P-CH2 hydrogens are
decomposition of 1 a t 40 O C of 3.0 X I O-'s-]. The peroxide decomposition is not exactly a "clean" first-order process. Measurements at different peroxide concentrations yield, on extrapolation to zero peroxide concentration, a rate constant of 2.8 x 10-7 s-1.30 Trapping with Nitrosodurene (ND).The EPR spectra of the two isomeric cis and trans P-ND adducts cannot be distinguished (see Table 11). As is well-known, this nitrosocompound is strongly dimerized in solution. The monomer concentration was calculated from the total N D concentration (as monomer) using an equilibrium constant, K = 303 M-' in benzene at 40 OC (see Experimental Section). The rate constant for trapping was obtained from the ratio of the initial rates of formation of S-ND and P-ND (see Table IV). The values of u , (at [ l ] = 3.05 X M) again indicate that all the alkyl radicals are trapped by the N D . Trapping with 2,4,6-Tri- tert-butylnitrosobenzene (t-Bu3NB). This trap is known to give only nitroxide spin adducts with prim,try alkyl radical^.^,'^ However, it has been reported32that with sccondary alkyls both nitroxide and anilino spin adducts are produced. We have confirmed this observation with the secondary alkyls formed by thermal decomposition of both 1 and 2. Thus, at 40 O C , 2 gives the H-nitroxide and H-anilino
I
-
pel, NtB magnetically inequivalent. This is because P,,,-NtB, like other analogous n i t r o x i d e ~ ,is~ chiral ~ . ~ ~ by virtue of the asymmetric 2-methylcyclopentyl group. The P-CH2 hydrogens are therefore inequivalent even if there is rapid internal rotation about the N-Co bond. The E P R parameters for these and the other spin adducts generated in this work are summarized in Table I I. The ratios of the initial rates of formation of the S. and P. spin adducts of N t B are given in Table 111 a t various N t B concentrations. Under the experimental conditions employed (for which k , = 2.1 X lo5 s-l) the average value of the rate constant for the trapping of secondary alkyls by N t B is calculated (by eq l) to be (6. l f 0.8) X loh s-l. In each of these experiments the initial concentration of the diacyl peroxide, 1, was 3.05 X M and the total rate of formation of spin adducts, ui, was (1.1 f 0.2) M-] s-l. Since this rate does not depend on the NtB concentration it is clear that essentially all the alkyl radicals generated are trapped in each experiment. This rate of radical formation yields a rate constant for the
H-nitroxide
H
H-onillno
spin adducts in a ratio of 0.057, independent of the trap concentration (see Table V). The absolute rate constant a t 40 OC for the spin trapping of secondary alkyls as their anilino adducts was found to be 3.1 X 1 Os M-' S - I by comparison of the rates of formation of the S-anilino and P - n i t r ~ x i d eadducts ~~ in the usual way (see Table VI). Combination of this rate constant with the data in Table V yields a value of 0.177 X lo5 M-' s - I as the rate constant for the trapping of secondary alkyls as their nitroxide adducts. Competition between Two Spin Traps. The EPR spectra for the ST. and PT. adducts with nitrones are not sufficiently different for kT values to be measured directly. However, these rate constants can be measured by competitive experiments, as described p r e v i ~ u s l y ,using ~ . ~ a trap for which kT has already been determined. The results of these experiments using H.
Spin Trapping of Secondary Alkyl Radicals
Maeda, lngold
4919
Table IV. SDin Traunine S . and P. Radicals w i t h N D in Benzene at 40 "C ~
kT x M-I
monomer [ N D ] " X IO3. M
total [ N D ] x 103. M
I .9 20.2 40.4
10.0
I .23 1.42 0.98
41 1 395 412
0.254 I .26 I.96
I .3 6.68
5-l
Calculated from eq I with k , = 2.1 X I O s s-l. Average kT = (406 f 8) X IO5 M-' s - l .
Based on K = 303 M-I (see text).
[ l ] = 3.05
[
x IO-3 M .
lable V. Spin Trapping H. with f-BulNB in Benzene at 40 "C
Scheme I1
x 103, M
(d [ H -nitrox] /df ),-.o (d[H-aniiinol/dO,-o
3.76 7.5 30. I
0.0570 0.0563 0.0573
[t-Bu3NB]
p
4.53\
.
0
2
PBN
radicals from 3.3 X I 0-3 M 2 in benzene a t 40 OC are summarized in Scheme 11. Arrows connect pairs of traps used in competition and point from the more reactive to the less reactive trap. Rate constant ratios are given on each a m ~ They are all averages of four or more individual competitive measurements at relative trap concentrations varying by a t least an order of magnitude. The validity of Scheme I 1 is attested to by its internal consistency and by the satisfactory agreement between these rate constant ratios for the nitroso traps and those obtained using the S. P. rearrangement. The absolute rate constants for secondary alkyl radical spin trapping by all the foregoing traps are summarized in Table VII. Trapping with 2,6-Di-tert-butyl-l,4-benzoquinone.Roginskii and B e l y a k ~ vrecently ~~ reported that this quinone is a useful trap for thermally generated alkyl radicals.36 The spin adducts formed with a number of tertiary37 and secondary alkyls showed hyperfine splittings by two (meta) and three (2-meta a - H of secondary alkyl) equivalent hydrogens. respectively, which indicates that they are the corresponding 4-alkoxy2,6-di-tert-butylphenoxyradicals.
-
+
*
Table VI. Spin Trapping S. and Pa Radicals with t-Bu3NB in Benzene at 40 OC [ I - B u ~ N B ] (d[S-anilino]/dt),+o ~ X . 103,~ M ~ (d[P-nitrox]/dr),-,o 30.8 120 308
k r anilino" M-l s-I
c , X IO9," M s-I
3.7 I 2.86 2.75
0.9 I 0.83 0.98
X IO5,
0.054 0. I63 0.404
Calculated from eq I with k , = 2.1 X IOs s-'. Average kT = (3.1 I f 0.43) X IO5 M-l s-I. [ l ] = 3.05 X M. f
f
Table VII. Summary of Rate Constants for Spin Trapping Alkyl Radicals in Benzene k T x 10-5, M-1 s-I prim" sech tert" 40°C 40°C 26°C spin trap 2-methyl-2-nitrosopropane (Nt B) nitrosodurene ( N D ) t r i -tert - bu t y I n it rosobenze ne (t-Bu,NB)-nitroxide ( I - Bu3N B)-anilino dimethylpyrroline I-oxide ( D M P O ) methylene-tert-butylnitrone ( M B k ) phenyl-tert-butylnitrone (PBN) 2.6-di-tert- butljl- 1,4-benzoquinonc
90 407d 4.7
f
26 31 1.3 h
61 406 0.18 3. I 4.2 13 0.68 0.93
33 900'
f 2.3 g g (0.1 g
Reference 4. This work. Reference 19. Recalculated taking A' = 303 M-I. e Revised value, ref 20. ./ Not observed. I: Not measured. Complex reaction; see text.
'!
Thermal decomposition of 2 in the presence of this quinone gave the expected adduct having three equivalent hydrogens ( a H (3 H ) = 0.9 G). Competition with N t B in benzene a t 40 "C yielded a rate constant for the trapping of H. of 0.93 X lo5 M-I s-', which is somewhat smaller than the value of 8 X 1 Os M-I s - I reported35 for the secondary alkyl radicals derived from n-decane at 50 O C . An attempt to confirm our value via peroxide 1 and the S*-,Pa rearrangement yielded a complex EPR spectrum (overall width = 3.2 G, six lines) which appeared to derive from more than one radical. A similar, but not quite identical, spectrum was obtained with n-heptanoyl peroxide (which yields n-hexyl radicals) and with acetyl peroxide (methyl radicals). Roginskii and B e l y a k ~ valso ~ ~reported that methyl (from the decomposition of tert-butoxy) gave a complex spectrum (which is probably not the same as that we obtained). They suggest that one of their observed radicals corresponds to methyl addition to the 3 position of the quinone. The present results suggest that primary alkyls are also sufficiently unhindered to add to this position.34 Certainly, in the case of benzoquinone itself, primary alkyls add to the ring very
rapidly.40 The rate constant for this reaction has been estimated40 to be 2 X I O 7 M-'s-l a t 69 O C .
Trapping Tertiary Alkyl Radicals. We could not devise an unequivocal method for measuring kT for tertiary alkyl radicals. However, relatioe kT values were obtained for the tertiary 2-phenylprop-2-yl (cumyl) radical which was generated by the thermal decomposition of a z ~ c u m e n e ~ 'in. ~benzene * a t 40 O C in the presence of pairs of spin traps. C ~ H ~ C ( C H ~ ) Z N = N C ( C H ~ ) ~2CbH5C(CH3)2 C~H~ -+
+ Nr
These data are summarized in Table VI11 together with relative kT values for the primary and secondary alkyl radicals. I n comparing the relative kT values for these three types of radical, it must be remembered that the primary and secondary
4980
Journal o j t h e American Chemical Society
Table VIII. Relative Rate Constants for Spin Trapping Alkyl Radicals in Benzene at 40 'C
spin trap
prim" 5-hexenyl
NtB ND t - Bu3NB-nitroxide r-Bu3NB-anilino DMPO M BN
(loo)< 450 5.2
PUN
29 34 I .4
sec" 6-hepten-2-
terth
Y1
cun1yl
(loo)
