INFORMATION TO USERS
This reproduction was made from a copy of a document sent to us for microfilming. While the most advanced technology has been used to photograph and reproduce this document, the quality of the reproduction is heavily dependent upon the quality of the material submitted.
The following explanation of techniques is provided to help clarify markings or notations which may appear on this reproduction.
1.The sign or “target” for pages apparently lacking from the document photographed is “Missing Page(s)”. If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting through an image and duplicating adjacent pages to assure complete continuity.
2. When an image on the film is obliterated with a round black mark, it is an indication of either blurred copy because of movement during exposure, duplicate copy, or copyrighted materials that should not have been filmed. For blurred pages, a good image of the page can be found in the adjacent frame. If copyrighted materials were deleted, a target note will appear listing the pages in the adjacent frame.
3. When a map, drawing or chart, etc., is part of the material being photographed, a definite method of “sectioning” the material has been followed. It is customary to begin filming at the upper left hand corner of a large sheet and to continue from left to right in equal sections with small overlaps. If necessary, sectioning is continued again—beginning below the first row and continuing on until complete.
4. For illustrations that cannot be satisfactorily reproduced by xerographic means, photographic prints can be purchased at additional cost and inserted into your xerographic copy. These prints are available upon request from the Dissertations Customer Services Department.
5. Some pages in any document may have indistinct print. In all cases the best available copy has been filmed.
University Micrdrilms International 300 N. Zeeb Road Ann Arbor, Ml 48106
8400314
Wright, Bradford Barry
I. PREPARATION AND CHARACTERIZATION OF META-XYLYLENE. SOLID STATE REACTIONS AND KINETICS OF ARYLCARBENES
The Ohio State University Ph.D. 1983
University Microfilms International300 N. Zeeb Road, Ann Arbor, Ml 48106
Copyright 1983 by Wright, Bradford Barry All Rights Reserved
PLEASE NOTE:
In all cases this material has been filmed in the best possible way from the available copy. Problems encountered with this document have been identified here with a check mark V .
1. Glossy photographs or pages______
2. Colored illustrations, paper or print______
3. Photographs with dark background______
4. Illustrations are poor copy______
5. Pages with black marks, not original copy___
6. Print shows through as there is text on both sides of ______page
7. Indistinct, broken or small print on several pages______
8. Print exceeds margin requirements______
9. Tightly bound copy with print lost in spine______
10. Computer printout pages with indistinct print______
11. P a g e (s) ______lacking when material received, and not available from school or author.
12. Page(s) ______seem to be missing in numbering only as text follows.
13. Two pages numbered______. Text follows.
14. Curling and wrinkled pages ______
15. Other______
University Microfilms International
I. PREPARATION AND CHARACTERIZATION OF M-XYLYLENE
II. SOLID STATE REACTIONS AND KINETICS OF ARYLCARBENES
DISSERTATION
Presented in Partial Fulfillment of the Requirements for
the Degree Doctor of Philosophy in the Graduate
School of The Ohio State University
By
Bradford Barry Wright, B.A.
******
The Ohio State University
1983
Reading Committee: Approved By
Matthew S. Platz
Leo A. Paquette
David J. Hart Advisor Department of Chemistry To Julie VITA
September 4, 1957. Born - Minneapolis, Minnesota
1979...... B.A., Honors, University of Oregon Honors College, Eugene, Oregon
1979...... Honorable Mention, National Science Foundation Graduate Fellowship Competition
1979-198 0...... Graduate Fellow, The Ohio State University, Columbus, Ohio
1980-198 1...... Graduate Teaching Fellow, Department of Chemistry, The Ohio State University, Columbus, Ohio
1981-1982 ...... Graduate Research Associate, Department of Chemistry, The Ohio State University, Columbus, Ohio.
1982-198 3...... Graduate Fellow, Phillips Petroleum, The Ohio State University, Columbus, Ohio
PUBLICATIONS
"Tunneling Parameters for the Hydrogen Atom Abstraction Reactions of Diphenylcarbene in a Low Temperature Matrix.", with V. P. Senthilnathan, Matthew S. Platz, and C. William McCurdy, Jr. Tetrahedron Lett. (1982), 23, 833-6.
"Meta-Xylylene, Electron Spin Resonance Spectroscopy of the Triplet State.", with Matthew S. Platz J. Am. Chem. Soc. (1983), 105, 628-30.
"Reactions of Triplet Diphenylcarbene by Hydrogen Atom Tunneling in Rigid Media.", with Matthew S. Platz, V. P. Senthilnathan, C. William McCurdy, Jr. J. Am. Chem. Soc. (1982), 104, 6494-501.
FIELD OF STUDY
Major Field: Organic Chemistry TABLE OF CONTENTS
Page DEDICATION...... i i
VITA...... iii
LIST OF TABLES...... V
LIST OF FIGURES...... vii
I. PREPARATION AND CHARACTERIZATION OF M-XYLYLENE 1
Historical Background...... 2
Theoretical Background...... 15
Results and Discussion...... 26
Additional Studies on Biradicals...... 50
II. SOLID STATE REACTIONS AND KINETICS OF TRIPLET ARYLCARBENES...... 56
Historical Background...... 57
Results and Discussion...... 70
EXPERIMENTAL...... 133
APPENDIX 1...... 158
REFERENCES AND NOTES...... 161 LIST OF TABLES
Table Page
1. Heats of Formation of the Xylylenes...... 10
2. Relative Energies of 8_ and 10...... 11
3. Typical D'/hc Values for Biradicals...... 25
k. Experimental Differences and Similarities Between Biradicals Derived from 21 and 25_...... ^0
5. Biradical Spectra from Diazocompounds...... 53
6. Biradical Spectra from Azides...... 5^
7. Product Distributions for the Reaction of Phenylcarbene with 2-Propanol as a Function of Temperature...... 63
8. Arrhenius Parameters for Some Matrix Carbene Reactions...... 68
9- Chemical and Kinetic Isotope Effects in the Reaction of Diphenylcarbene with Isotopically Labelled 2-Propanol at 77 K ...... 7*+
10. Time Dependency of on Photolysis Time Interval...... 76
11 Concentration Dependence of 76 . Olio 12. Abbreviations Used for Carbenes Presented in the Remainder of Part II...... 78
13. Carbene Rate Data in Toluene...... 79
Ik. Arrhenius Parameters for the Reaction of Arylcarbenes with Various Solvent Matrices...... 8k
15. Calculated Eckart Barrier Parameters for Various Arylcarbenes in Toluene Matrices...... 87
16. Carbene Rate Data in CCl^...... 10U
v Table Page
17. Carbene Rate Data in Isotopically Labelled Methanols...... 109
18. OH/CH Bond Insertion Product Ratios for Various Arylcarbenes in Isotopically Labelled Methanol Matrices at 77 K ...... 110
19. Signal Intensities in Arbitrary Units for Matched Tubes of Phenyldiazomethane in Isotopically Labelled Matrices...... 117
20. Signal Intensities in Arbitrary Units for Matched Tubes of Some Carbenes as a Function of Temperature in Toluene and Toluene-dg Matrices...... 118
21. Observed Maximum Splittings of Radical Generated by Reaction of Carbenes with Matrices...... 119
22. ESR Kinetic Data not Given in Text...... 1^0
23. ESR Signal Behavior of m-Xylylene as a Function of Temperature and Microwave Power...... l*tl
2k. Product Distribution Under Varied Experi mental Conditions for the Reaction of DPC with 2-Propanol...... 1^3
VI LIST OF FIGURES
Figure Page
1. Apparatus for the Matrix Isolation of Xylylenes...... 9
2. Possible Combination of Benzyl and Methyl Radicals...... 12
3. Possible Arrangements of Two Electrons In Two Orbitals...... 15
k. States Obtained on Placing Two Electrons In Two Arbitrary Orbitals...... l6
5- Typical Biradical ESR Spectrum...... 22
6. Coordinate System for the Spin Dipolar Interactions of Two Electrons...... 22
7. ESR Spectrum of 22 in Low Field. ., ...... 27
8. ESR Spectrum of £5 in Ethanol at 77 K ...... 28
9. ESR Soectrum of £3 in Ether at 77 K .•...... 30
10. ESR Spectrum of 8_ in Ethanol-dg...... 31
11. ESR Spectrum of 8^ in Methanol at 77 K ...... 32
12. ESR Spectrum of £3 in Toluene at 77 K ...... 33
13. ESR Spectrum of J3 in 2-Methyltetra- hydrofuran at 77 K ...... 3^
lU. Thermal Generation of m-Xylylene...... 36
15. Biradical Spectrum Arising from Photolysis of Diphenyldiazomethane...... 38
16 . Biradical Spectrum from Photolysis of l,3-Bis(diazomethyl)benzene in the Presence of Excess Diphenylamine...... k3
17. Biradical Spectrum Generated on Photolysis of Mesitylene at 77 K ...... UU
vii Figure Page
18. Saturation Curves for m-Xylylene at Various Temperatures...... ^7
19. Matrix Influences on Biradical Closure...... 6l
20. Arrhenius Plot Showing Tunneling Corrections...... 82
21. Eckart Potential Barrier...... 85
22. Calculated Tunneling Rates for DBT in Toluene...... 88
23. Calculated Tunneling Rates for DBT in Toluene...... 89
2k. Calculated Tunneling Rates for DBT in Toluene...... 90
25. Calculated Tunneling Rates for DPC in Toluene...... 91
26. Calculated Tunneling Rates for DPC in Toluene...... 92
27. Calculated Tunneling Rates for DPC in Toluene...... 93
28. Calculated Tunneling Rates for DPC in 2-Propanol...... 9k
29. Calculated Tunneling Rates for DPC in 2-Propanol...... 95
30. Calculated Tunneling Rates for DPC in 2-Propanol...... 96
31. Calculated Tunneling Rates for 9-AC in Toluene...... 97
32. Calculated Tunneling. Rates for FI (0.1M) in Toluene...... 98
33. Calculated Tunneling Rates for FI (O.IM) in Toluene...... 99
3^. Possible Trajectories for H Atom Abstraction...... 100
35- Proton Abstraction Mechanism of Carbene Insertion Alcohols...... 108
viii Figure
36. Ylid Mechanism for Carbene Insertion into Alcohols......
37. Surface Crossing Mechanism for Triplet Arylcarbenes with Alcohols......
38. Mechanism for Ether Formation by Triplet Carbenes in Alcohols......
39- ESR Spectrum of Radical Pair Derived from DBS in Toluene......
ItO. ESR Spectrum of Radical Pair Derived from DBS in Toluene-dg......
Ul. ESR Spectrum of Radical Pair Derived from DBS in 2-Methyltetrahydrofuran...
k2. ESR Spectrum of Radical Pair Derived from DBS in CCl^......
i*3. ESR Spectrum of Radical Pair Derived from DBS in 3-Methylpentane......
1A. ESR Spectrum of Radical Pair Derived from DBS in Methylcyclohexane......
b$. ESR Spectrum of Radical Pair Derived from FI in Toluene......
b6. ESR Spectrum of Radical Pair Derived from FI in Toluene-dg......
1+7- ESR Spectrum of Radical Pair Derived from FI in CCl^......
1*8. ESR Spectrum of Radical Pair Derived from DMA in Toluene ......
49. ESR Spectrum of Radical Pair Derived from DMA in Toluene-dg......
50. Possible Radical Pair Orientations....
51. Liquid Nitrogen-Dewar Apparatus for ESR Measurements at 77 K ......
52. Nitrogen Cooling Apparatus for ESR Measurements Above 77 K ......
53. Helium Cooling System for ESR Measurements Below 77 K ...... Figure Page
5h. Curie Plot of the Powder ESR Spectrum of DMA...... 139
55. Apparatus for Solid State Photolysis Used in Matrix Studies...... l M
56. Wave Functions for a Particle Tunneling Through a Potential Barrier...... 159
x I. PREPARATION AND CHARACTERIZATION OF M-XYLYLENE
II. SOLID STATE REACTIONS AND KINETICS OF ARYLCARBENES
By
Bradford Barry Wright, Ph.D.
The Ohio State University, 1983
Professor Matthew S. Platz, Advisor
M-Xylylene was prepared from three different precursors and was characterized by ESR spectroscopy, D'/hc = 0.011 cm ^ and E'/hc =
0.001 cm \ A triplet ground state was established for this species.
Additional biradical spectra resulting from photolysis of aryl diazo- compounds were studied and tentatively assigned spirarene structures.
The chemistry of arylcarbenes in low temperature organic matrices was investigated by ESR kinetic methods coupled with product analyses.
Hydrogen atom tunneling was suggested as the mechanism of carbene decay in hydrogen atom donating matrices. The mechanisms of decay in CCl^ and alcoholic matrices was also studied. Matrix isolated radical pairs were observed in the decay of several carbenes in low temperature matrices.
The above results were discussed in terms of a multiple matrix site model.
I PREPARATION AND CHARACTERIZATION OF M-XYLYLENE. Historical Background
The study of organic biradicals has been an area of substantial 1 2 interest and activity. Since the early work of Schlenk, Thiele, and 3 Chichibabin, who prepared compounds respectively, chemists have
sought to prepare biradical and biradicaloid molecules. Indeed ^-3^ are
crystalline solids, but most such compounds have only a fleeting exis
tence at room temperature. To observe these reactive species two stra
tegies have been used.
0
3
One approach has been to decrease the time necessary for observa tion. To this end nano- and picosecond time resolved laser flash pho tolysis has been used successfully to determine rates of reaction and spectral characteristics of a number of biradical systems.^ Alternatively, low temperature matrix isolation has been used to
slow down processes leading to destruction of the species under study.
This method was used by Buchwalter and Closs"* to obtain the Electron
Spin Resonance (ESR) spectrum of 4 at 5.5 K and has by far proved to be
the most general means of observing these reactive compounds.
5. r K O +
4
The following triplet biradicals have been observed and studied by
ESR in this manner.^-®
.a .
5 6 7
Compounds J^, 5_, 6_, and 7_ are members of the class of biradicals Q known as non-Kekule hydrocarbons. That is, there exists no classical
method for satisfying the standard rules of valence, or simpler put,
all canonical resonance structures have two or more unpaired electrons.
Note that these compounds differ fundamentally from species such as 4^
in that the Tr-electrons are delocalized over more than two atomic
orbitals. It is to this class that the title compound jm-xylylene (8>)
belongs.
The remaining isomeric xylylenes _9 and Jd) have quinoidal valence
bond structures. This difference has a profound effect on their spin multiplicity and hence their reactivity. Michl*® has labelled the latter compounds "biradicaloids". 4
tO 8 9
Both jj-and pr-xylylene have been determined to be diamagnetic
compounds and their NMR spectra have been obtained.^ One may be
tempted to assume that all compounds which can be drawn with no un
paired electrons have singlet ground states. This assumption is not
valid, for compound J_I has been prepared and found to have a triplet
ground state. 12 Steric considerations make 1 lb favored over 11a in
contrast to 3 which is planar.
,0
0
11 b
The chemistry of j>-xylylene has been reviewed by Errede and 1 Q Swarc and predominantly involves polymerization processes. This appears to be due to the distance between the reactive benzylic carbons.
In the case of jj-xylylene, where the atoms with high free valances are closer spatially, many useful trapping reagents are known, and there are examples of intramolecular cyclizations involving o-xylylene moieties leading to natural products as illustrated in Scheme 1 . ^ o 1 Scheme
0=0 o I ro
0 ^ 0 o o
M 0=0 0=0
s Significantly, all known examples of reactions involving 9 and
olefins to give cyclized products proceed with very high stereoselec
tivity consistent with singlet reactivity.^
A variety of derivatives of 9 and 10 have been prepared and shown
to be diamagnetic.^ Compounds 12 and J_3 are in fact reasonably
stable below 0 °C in contrast to 9 which dimerizes above -80 °C. This
increased stability was attributed to steric rather than electronic
factors.
1 2 13 14
1 6
Despite an abundance of information on _9 and _10 and their der
ivatives, relatively little is known about JJ and its derivatives. Aside
from J^, which has been observed by ESR, 9 1 the only spectroscopic
evidence of derivatives of J5 has been reported by Migirdycin and
Baudet. 22 These workers irradiated (2537 A),° frozen glasses containing m-xylene and mesitylene to form m-xylylene and 5-methyl-m-xylylene.
These compounds were characterized by their fluorescence and excitation spectra. Confidence in this method of generating xylylenes comes from the authors' ability to generate an identical fluorescence spectrum of
9 by irradiation of o-xylene to that obtained by Michl and Flynnfrom Based on the vibrational fine structure and calculated transition
energies for the singlet and triplet manifolds, these ni-xylylenes were
assigned triplet ground states.
Derivatives of JJ have been implicated in the formation of cyclo- 23 phanes. Indeed, in a clever approach to a derivative of j} Gajewski A / and Stang added dimethylvinylidene to dimethylfulvene and recovered an octamethyl-m-cyclophane, (Scheme 2). The cyclophane presumably arises from the dimerization of _18, however Chemically Induced Dynamic
Polarization (C1DNP), was not detected in the dimer when the reaction was run in an NMR spectrometer.
Attempts to prepare J3 by methods that successfully give 9_ and 10 have not proved fruitful. Accordingly, Tseng and Michl^ treated o^, ni-, and p^xylylene dibromides with potassium vapor in the flow system shown in Figure 1. Both o- and p^xylylene were observed spectrosco pically in the argon matrix, but no identifiable products were obtained from the jn-dibroraide.
This is in agreement with the work of Farmer et a l ^ who found by mass spectrometry that the thermal stability of the m-xylyl radical to Scheme 2. ¥
oo Argon
Alkali m etal
M icrow ave S am p le Cavity
Figure 1. Apparatus for the Matrix Isolation of Xylylenes.
v£> 10
loss of a hydrogen atom is much greater than that of the o- and p-
isomers. Further, Pollack, Raine, and Hehre^ have attempted to measure
the heats of formation of J3, 9_, and _H) by ion cyclotron double
resonance spectroscopy.Under the conditions of the experiment, 9^ and J+)
were easily produced, but J3 was not formed. Their results are given in
Table 1.
Table 1. Heats of Formation of the Xylylenes.
Compound A H ° (kcal/mol)
8 76 + 4
9 53 + 4
10 50 + 4
Borden et al^® have pointed out that AH for 8^ by this method is valid only for the singlet state of 8. Since, as will be described later, the triplet state is calculated to be 10 kcal/mol lower in energy than the singlet state, the actual value of AH for JB might be as low as 66 kcal/mol.
The instability of J5 relative to 9_ and has been shown 29 using the Perturbational Molecular Orbital (PMO) method developed by Dewar^® in which benzyl and methyl radicals are joined to give the isomeric xylylenes and styrene. The change in •ft'-energy on uniting the two fragments is given by the Huckel Non-Binding Molecular Orbital (NBMO), coefficient of each moiety at the atoms where they are joined according 11
to the equation,
A E = 2 (abenzyl)(am e t h y l ^ a = NBMO coefficient at atom of union
Figure 2 illustrates the results.
Thus styrene is by far the most stable isomer, then oi- and j>-
-xylylene at comparable energies with m-xylylene being the least sta
ble. Indeed, this simple model bears out the experimental facts reason
ably well. Fortunately, it is not necessary in this case to rely en
tirely on this quick calculation for relative energies of J8, 9^, and 10.
As mentioned earlier Borden et al^® have recently published
results of ab initio calculations on 8. After careful geometry optimization for the and 3B2 states of J3 and the *Ag state of _10
Cl calculations using a Dunning (3s,2p/2s) split-valence basis set were carried out. At the 11-fi'SDTQ Cl level of theory the relative energies were as in Table 2.
Table 2. Relative Energies of J3 and 10.
State Energy (kcal/mol)
3r a 2 0.0
XA b +10.0 A 1
la c -24.2 s
a) Lowest triplet state of JB. b) Lowest singlet state of 8. c) Lowest singlet state of 10. -2 i f f « i/rr A E „ = 0
- m (T
A E rt=- 7 6 ?
AEt = 1.52 ?
Figure 2.. Possible Combinations of Benzyl and Methyl Radicals.
i—■ to 13
This finding confirms the previous predictions of a triplet ground
state of 8 as well as the greater stability of _10 to 8^
It should be pointed out that calculations on systems of this size are very costly, and therefore the geometries to be investigated must maintain a high level of symmetry. Thus, the results quoted above for j8 strictly hold only for C2v geometries.
Finally, Berson and c o w o r k e r s *^2 have prepared two oxygenated derivatives of m-xylylene, 19b and 20b. Both compounds are prepared by photolysis of the corresponding enones, 19a and 20a and are ground state triplets. Despite their triplet nature the dominant mode of reaction appears to be through a zwitterionic singlet, as Markownikov addition to olefins and regiospecific addition of alcohols has been observed (Scheme 3).^^
O O' II
19a 19b
O
20a 20b
Whatever the reaction pathway, the existence of 19b and 20b as triplet species suggests that JJ should also have a triplet ground state and hence be observable by ESR spectroscopy. Scheme 3.
OH
ROH
CH© O CHjOR OR I i f OH OR Theoretical Background
At this point it is useful to consider the factors responsible for
determining the ground state multiplicity of a biradical species. As
shown in Figure 3, there are six ways that two electrons can be distri
buted among two orbitals x and 'Hy.
(a) i_ JL
Cb) i _ _ i
(c) 1 I
(d)
(e> 11
(f) 11
Figure 3. Possible Arrangements of Two Electrons in Two Orbitals.
If S^and ^ a r e identical, then the one electron energies of each arrangement are the same, however on addition of a second electron, coulombic repulsion between the electrons lifts the degeneracy of (a) -
(f). The six configurations above describe four low lying energy states, one triplet and three singlet states. Configurations (a) and
(b) are clearly triplets and (e) and (f) are clearly singlets. The 15 (1) Triplet + $ * ) / f ? E=hi+hj+JirKij (2) Singlet E=W Ji3+K13 (3) Singlet E= V hr (Ji5+J33)/2-(Ki/+ < W (Jii-J3j)/2)) (4) Singlet E=hi+h3+'JiitJ33)/2+(Kij2+
(hl - V (Jli-J33)/2))
J = coulomb integral >0 K = exchange integral >0 h = one electron energy E = total energy
Figure 4. States Obtained on Placing Two Electrons in Two Arbitrary Orbitals.
J—* ON 17
remaining singlet and triplet configurations are derived from the sum
and difference of (c) and (d) respectively. These states can be written
as in Figure 4 for generalized orbitals i and j which may or may not be
degenerate.
Since Jjj > (i.e.- the coulorabic repulsion between electrons
in the same orbital is higher than between electrons in different
orbitals), the quantity + ^jj)22 greaCer than JNj and singlets
(3) and (4) become higher in energy then (1) and (2). By inspection,
the energy of (1) is less than that of (2) and a triplet ground state O C results. This is the physical basis of Hund's Rule. However, if the
energy gap between the one electron energies h^ and hj becomes large
enough , then the terra in brackets in (3) dominates over the coulombic
repulsion terms and a singlet ground state biradicaloid results. Hoff-
mannJ has suggested that when the energy difference between h^ and hj
becomes greater than about 1.5 ev, a biradicaloid results.
In both biradical and biradicaloid species the high energies of the orbitals involved give rise to concomitant high chemical reac tivity.29
One method of increasing the energy gap between NBMO's is by varying the electronegativity of the biradical termini. Platz and coworkers29,92 have calculated by simple HMO methods that degenerate
NBMO's of 1,8-naphthoquinodimethane are split 0.3^ units upon inclusion of a nitrogen atom. 18
A E NBMC>— ■ °P AEN B M O “ °-3 P
To the practicing chemist the foregoing discussion is too general
to be useful in making routine predictions of the ground state multi
plicities of compounds such as 8. To avoid the need of ab initio calculations to determine ground state spin, three simple models have been developed which demonstrate considerable reliability.
The Borden and Davidson^® Model simply states that if the Huckel
NBMO's cannot be localized onto different sets of atoms a triplet ground state results out of a need to minimize electron repulsions.
Otherwise a singlet is the ground state or very close to it in energy.
This procedure may be further simplified by using the starring pro- on cedure developed by Dewar for alternant hydrocarbons. The number of starred and unstarred atoms are counted and compared. It has been shown that when the numbers of starred and unstarred atoms are equal, then OO the NBMO's are disjoint, thus the need to reduce electron repulsions will be low and a singlet state is possible. On the other hand, if there are more starred than unstarred atoms, then both NBMO's are usually confined to the set of starred atoms and a triplet results.
Ovchinnikov^3 9 has derived a similar method for determining ground state multiplicities in which the multiplicity is given by,
Multiplicity = jn-n*H*l
Interestingly, this method is claimed to work even when hetero- atoms are included in the TT -system. 19
As noted by Berson,^ there exist systems where these methods give different answers such as,
This has NBMO's which are disjoint and thus is a singlet by
Borden's method. However, Ovchinnikov predicts that a triplet will result. Obviously, this system provides a future significant test of theory.
A third method, developed by Longuet-Higgins^ is the simplest of all. The multiplicity is calculated as follows,
(5) k > N - 2T
(6) Multiplicity = k + 1 where,
k = // of NBMO'S
N = // of TT-electron centers
T = # of double bonds
For most systems line (5) can be replaced by an equality, but the inequality is required for antiaromatic annulenes. A possible success of this method over the other two has been the case of tetramethylene- ethane (6).
Both the Borden and Ovchinnikov methods predict a singlet ground state for this biradical while Longuet-Higgins predicts a triplet.
Dowd^ has observed the triplet ESR spectrum of tetramethyleneethane-dg at -196 °C. This fact and the stability of at this temperature suggest that a ground state tripletmay existin this case. 20
In the case of all three methods agree and a triplet ground
state is predicted.
■¥ * 5 starred atoms
* 3 unstarred atoms
ESR spectroscopy is the most powerful physical method of charac
terizing a molecule in the triplet state.First, an unambiguous
assignment of the type of specie giving rise to a signal such as a
doublet, triplet, or quintet is possible. This is in contrast to other
methods such as flash photolysis where assignment of transient spectra
is much more difficult and error prone.^ Second, by observing the
signal behavior as the temperature is varied, a determination can be made as to whether the observed signal results from a ground or excited state. Further, analysis of the spectrum yields information about the electron distribution of the specie under observation (e g.-D'and E' values for biradicals).
In principle, ESR is very similar to NMR except that in the former electron spin is flipped instead of nuclear spin. Due to the difference in mass between the electron and the proton the energy difference between quantized spin states is substantially larger in ESR than in
NMR. This increased energy gap between levels enhances the Boltzman population difference between the Mg levels and therefore makes ESR inherently more sensitive in its ability to detect signals than NMR. 29 ^
In this regard it is possible under normal conditions to detect para magnetic species at concentrations of 10~^ M in a continuous wave scan. 21
Traditionally, spectroscopists have used the conditions of J< (two isolated electrons), and J » A (two strongly coupled electrons), to describe a biradical and triplet respectively, where J is the electron- -electron interaction and A is the electron-nuclear hyperfine inter action.^ However, as commonly used in the literature of organic chemistry today, the terras are not mutually exclusive. Indeed many of the species presented in the previous section are described as triplet biradicals. To the organic chemist a biradical may be thought of as any specie which contains two electrons, with each electron in a degenerate NBMO. Due to the difficulty of preparing oriented crystalline samples, the majority of literature on triplet molecules concerns randomly oriented solid samples. The ability to observe discrete absorptions from a randomly oriented solid sample results from the fact that a sharp increase in the number of absorbing molecules occurs at those fields where the external magnetic field is parallel or antiparallel to one of the principal magnetic axes of the molecules. Importantly, to obtain useful triplet spectra, it is necessary for the molecules to be held rigidly in the sample and not allowed to tumble. This is because the spin-dipolar interaction of the unpaired electrons which is respon sible for the main features of the spectrum will be motionally averaged to zero if the biradical is allowed to tumble. This results in the collapse of the typical six line spectrum shown in Figure 5 into a single resonance line.^^ 22 |«— 6E- 9 2 D Figure 5. Typical Biradical ESR Spectrum. A convenient way to understand the origin of triplet spectra has been given by Wasserman.^ In this treatment two coupled electrons are placed in a magnetic field (below). The field experienced by electron A is then the vector sum of the external field and the field produced by electron B. Figure 6. Coordinate System for the Spin Dipolar Interaction of Two Electrons. 23 This field is given by the equation, liA = Ho + = Ho +/iB(3cos2 " O / r 3 where /*6is the magnetic moment of electron B, Hq is the external magnetic field, and © and r are defined as in Figure 6. This interac tion is manifested in the dipolar Hamiltonian as follows, where gg has the value 2.00, @ is the electron Bohr magneton, and is the spin of electron i in units of h. By transforming from laboratory coordinates to the principal axis system of the molecule one can write, -fo/3-E'J S x2_ [ D //3 + E'Js/ + 2/ 3 D ' S z2 which gives the complete Hamiltonian as, where g is now described by the anisotropic tensor j* and, D = 3/4 q e2 $ 2<® < 12)1 (1/r123)-(3zl22 /r,25 ) | 0(1.2 )> Here 0(1,2) is the antisymmetric spatial portion of the wave function. D' and E*are the zero field parameters and govern the line- shape of the spectrum as well as the triplet state energy level split tings when no field is present. D‘ is a measure of the average separa tion of the two electrons and decreases with increased If-conjugation 24 (see Table 3), while E* is related to the symmetry of the orbitals containing the unpaired electrons. In the case of a threefold orbital axis of symmetry, E* = 0. In practice, D‘and E1 are easily obtained from the derivative triplet spectrum as shown in Figure 5. Finally, it is possible to determine whether the biradical spec trum observed is due to a triplet ground or excited state by observing its signal intensity as a function of temperature. The relationship between the two is given by the Girie-Weiss Law as modified by Wasser- man et al,, 45 T . t = e*p(-AE/RT) r l+3exp(-AE/RT) where I is the signal intensity, T the absolute temperature, C a con stant, a n d A E the energy separation between singlet and triplet states. A E is taken as positive for a singlet ground state. For singlet ground states a signal increase will be observed on warming the sample while on the other hand, for a triplet ground state the equation reduces to, I *T = C and a decrease in signal is observed on warming. However, since at low temperatures it is difficult to populate excited states sufficiently for observation by ESR, the observation of a triplet spectrum at or below liquid nitrogen temperature indicates that the triplet is the ground state or thermally excited by less than 0.5 kcal/mol.^ 25 Table 3. Typical D'/hc Values for Biradicals. -1 Compound D'/hc (cm ) Reference 0.0248 X. 0.012 0.0128 . 0 1 0 . o 0.0266 32 0.0204 33 CHj Results and Discussion In recent years Platz and Senthilnathan^ have demonstrated that aryl carbenes readily abstract hydrogen atoms from organic matrices at cryogenic temperatures. Consequently, it was felt that it might be possible to prepare 8^ by double hydrogen atom extraction from solvent by the corresponding dicarbene 22 (Scheme 4). A literature search revealed that 22 had already been prepared from l,3-bis(diazomethyl)- benzene (21) by Wasserman et al.^As expected, 22_ was found to be a ground state quintet by ESR. Scheme 4. CH CH CH hV SH SH CH CH Subsequently, a sample of 21^ was prepared and dissolved in eth anol. Irradiation in the cavity of an ESR spectrometer at 77 K gave rise to a number of signals in the ESR spectrum. Among these were several strong absorptions at low field due to 22^ (Figure 7), and a biradical spectrum centered at 3290 G with an observed splitting of 246 26 Figure 7. ESR Spectrum of 22 at Low Field. EtOH 246 G — Figure 8. ESR Spectrum of 8 in Ethanol at 77 K. N) 0 0 29 G (Figure 8), with |D*/hc| =0.011 cm-1 and|E‘/hc| < 0.001 cm-1. Upon shuttering of the light source the quintet peaks decayed, however due to the small signal intensity of the z-transitions it was not possible to quantitatively correlate this decay with a growth in the biradical spectrum. Even under the best of conditions such a correlation would be difficult since the intermediacy of 23 complicates the kinetic scheme. Within experimental error the biradical spectrum showed no signal decay in 5 days at 77 K. Thus, if the biradical spectrum observed was in fact due to (3 then Scheme 4 promised to provide an effective means of producing it. To rule out the possibility that the spectrum in Figure 8 arose from a radical-radical matrix pair, ESR samples of 2l_ were prepared in ether, ethanol-dg, methanol, toluene, and 2-methyltetrahydrofuran (2-MTHF). If the spectrum were due to a radical pair formed by H-atom abstraction from the matrix then substantial changes in the observed maximum splitting should have been observed, however, no such change was evident (see Figures 9-13). Radical pairs generated from several aryl carbenes have been studied and the results are presented in Part II of this work. In these cases differences of over 150 G in the split tings are observed over this range of solvents. In the previously mentioned solvents, it was possible to observe the decay of J22 generated in the matrix. The low signal intensity of the signal in methanol is probably due to carbenic insertion into the OH bond of methanol.^® Since it was impossible to correlate the decay of Z2_ and the growth of the biradical spectrum at 77 K, experiments were carried out at lower temperatures through the use of a helium cooled probe. Even at Ether |------246 G — Figure 9. ESR Spectrum of 8 in Ether at 77 K. LO O EtOH-d 246G Figure 10. ESR Spectrum of 8 in Ethanol-d.. 6 Figure ll._ ESR Spectrum of 8 in Methanol U> N5 TOLUENE 246 G Figure 12. ESR Spectrum of 8 in Toluene at 77 K. CO co 2-MTHF LO Figure 13. ESR Spectrum of in 2-Methyltetrahydrofuran at 77 K. 4^ 35 20 K. (the lowest temperature achievable with this system), decay of the quintet was rapid and growth of the biradical could not be accurately measured. In an effort to further slow the rate of decay of 23 deu- teration of the matrix was tried. After some experimentation, ethanol- -dg (Etoh-dg) was chosen. In this solvent irradiation at 20 K generated only the quintet spectrum and doublet impurities. No biradical spectrum with a splitting larger than 60 G could be detected. At this tempera ture, the quintet was stable indefinitely. Upon warming the sample to 77 K the quintet spectrum decayed and the 246 G splitting grew in. To rule out anomalous temperature effects the sample was recooled to 20 K. The quintet spectrum which had been allowed to decay completely did not reappear, while the biradical spectrum remained. Therefore, the biradical spectrum is a "dark" pro duct generated by the thermal decay of 22^, presumably resulting from reaction with a CH (CD) bond of the matrix (see Figure 14). To check the validity of this last assumption, 21^ was placed in PerFluorinated Alkanes (PFA), a matrix inert to aryl carbenes at 77 49 K. ^ After extensive irradiation in this matrix strong quintet ab sorptions were observed, but no biradical spectrum was observed. The quintet absorptions were stable indefinitely at this temperature. This result rules out possible rearrangement of 22^ to give the biradical spectrum, and indeed any process not involving the solvent in some manner. Additionally, p-bis(diazomethyl)benzene (24) was prepared and photolyzed. This compound gave rise to no observable biradlcal spectra. This is consistent with the model of Scheme 4 since in this case would result and is diamagnetic. 36 (a) (b) (c) (d) (e) (a), (b), (e) at 22 K; (b), (c) at 77 K. Figure 14. Thermal Generation of m-Xylylene. 37 Despite the growing body of evidence pointing to jJ as the carrier of the biradical spectrum, the possibility that another species was involved caused much concern. Particularly disturbing was the little known fact that very similar spectra (234 G splitting) could be genera ted by photolysis of diphenyldiazomethane (25) in a variety of solvents (see Figure 15).-*® In this case too, a fixed splitting was observed for each solvent. To better understand whether the spectrum purported to be due to 8 was in fact due to this phenomenon, a study of the properties of this latter biradical specie was undertaken. Irradiation of Z5_ at 77 K or 4 K results in the immediate gener ation of a strong biradical spectrum regardless of solvent (e.g.-tol uene, toluene-dg, ether, EtOH, EtOH-dg, MeOH, PFA, CCl^, CgFg) which is stable until -70 °C at which point rapid decomposition occurs. This spectrum is observed in addition to the normal spectrum of diphenylcar- bene. However, no growth in the biradical signal could ever be observed during carbene decay. Further, the existence of the signal in PFA, CCl^, and CgFg pointed to a photochemical process as responsible for the spectral carrier. Low temperature irradiation (22 K), of 2^5 in toluene and toluene-dg gave no observable isotopic effect in the 77 K/ PFA h-p Figure 15. Biradical Spectrum Arising from Photolysis of Diphenyldiazomethane. 00 39 production of the biradical spectrum for short irradiations. To determine whether the biradical resulting from photolysis of 2_5 resulted from a photoreaction of the diazocompound, secondary photol ysis, or reaction of the generated carbene, another precursor for diphenylcarbene was sought. Although sulfite Tl_ and epoxide ^8 have been known to give diphenylcarbene on photolysis in solution,-’* it proved impossible to generate diphenylcarbene in frozen matrices by 52 this method. 0 ; > A c : A , 0 0 27 28 29 Success was achieved by photolyzing 73_ in PFA and CCl^ at 20 K. Unlike samples of 25 under these same conditions, 29 gave rise only to the spectrum of diphenylcarbene. This important result indicates that the diazo group is somehow directly involved in the production of a biradical species. There are several possibilities that come to mind which will be discussed in more detail later. The results of these experiments on both biradical spectral car riers is summarized in Table 4. Since it had been demonstrated that diazocompounds can give rise to anomalous biradical spectra, another precursor to 8 was sought. Compounds 30-34 were proposed as possible precursors. Table 4. Experimental Differences and Similarities Between Biradicals Derived from 21 and 25. 21 25 thermally generated photochemically generated solvent isotope effect no solvent isotope effect on formation on formation not observable in PFA observable in PFA spectrum solvent spectrum solvent invariant (246 G) invariant (236 G) 41 32 3 0 3 1 O II 3 3 3 4 Compounds _30 and 3J^ were both prepared and irradiated at 77 K in a variety of solvents, but failed to give biradical spectra. There are many possible reasons for this failure and they will not be mentioned here. Similarly, a sample of 34-^ was irradiated over a variety of wavelengths and solvents for extended periods of time. Within experi mental error no biradical spectrum with a 246 G splitting was observed. Attempts to prepare 32^ and ^ by intramolecular cyclizations resulted in polymer formation. Finally, a suitable alternate precursor to JJ was found. 1,3-bis- (bromomethyl)benzene (BBMB), 35, was placed in 2-MTHF with a ten molar excess of diphenylamine (DPA), (0.10M DPA and 0.010M 35). Under these conditions aryl halides have been known to accept electrons from photo ionized amine molecules resulting in cleavage of the carbon halogen bond analogous to that illustrated in Scheme 5 . ^ 42 Scheme 5. C H , B r c h 2- + D P A + 2 DPA +* + 2 Br ' C H j B r C H j - 35 Irradiation at 77 K of this sample quickly gave a complicated spectrum with a maximum splitting of 246 G (see Figure 16). Unfortun ately, the interior lines were obscured by the presence of other para magnetic species. This result was reproduced in ethanol. It was not possible to perform this experiment in PFA since the temperatures necessary to prepare a sample tube ( 50-60 °C) caused a reaction to occur between DPA and BBMB. Other inert solvents such as hexafluorobenzene and carbon tetra chloride were too opaque to generate signals by this method. Inciden tally, a variety of aromatic amines were tried as electron donors. Out of these trials only DPA emerged as having a sufficiently low chemical reactivity toward J35 while at the same time maintaining sufficient solubility in organic solvents to permit sample preparation. Irradia tion of 3_5 or DPA alone gave no biradical spectra. The ability to generate the 246 G splitting by two independent methods confirms the identity of j} as the spectral carrier. Further, confirmation was obtained by photolysis of mesitylene in ethanol with V 2537 A light. After prolonged irradiation at 77 K at this wavelength a weak biradical spectrum was observed with splitting 245 G (Figure 17). This is consistent with 5-methyl-m-xylylene and confirms Migirdycin's 77 K 246 G Figure 16. Biradical Spectrum from Photolysis of 1,3-Bis(diazomethyl)benzene in the Presence of Excess Diphenylamine. 4>- u> CH 246 G Figure 17. Biradical Spectrum Generated on Photolysis of Mesitylene at 77 K. 45 assignment. 22 However, similar treatment of m-xylene failed to yield a detectable spectrum of _8. Similarly, irradiation of the following compounds did not give rise to recognizable signals under the same conditions. OH OH CH- CH CH OH This indicates that direct photolysis is not a promising method for the generation of biradicals in general. In order to determine the ground state multiplicity of J3 samples of j8 were prepared by both methods mentioned previously and observed as the temperature was varied between 20 and 77 K. A signal increase was observed in both cases indicating a singlet ground state according to the Girie Law,^-* in stark contrast to theoretical predictions. The source of this discrepancy was revealed upon lowering the microwave power from 10 mW to .01 mW. At this setting an increase in signal intensity was observed on going to lower temperatures, indicating a triplet ground state. The reason for this change in behavior may be assigned to saturation of the signal at low temperatures. This is a fairly common problem at these temperatures in systems where spin-spin relaxation is not very efficient. A study of the signal behavior as a function of 46 microwave power and temperature was made and the results are summarized in Figure 18. The signal intensity is proportional to the square root of the microwave power. In this plot saturation effects show up as curvature in the line at each temperature. Thus only at .01 mW (the lowest setting available on the instrument) can linear behavior be observed and saturation avoided. Since the completion of this work, Professor Berson has communica ted some of his research results relating to 8_. Low temperature pho tolysis of gives an ESR biradical spectrum with the same observed splitting as found in the body of this work. 3 6 This spectrum presumably results from a Norrish Type I cleavage with subsequent ring opening to give j8. This method promises to be a syn thetically more useful precursor than the methods mentioned previously. Indeed, it appears that J3 can be trapped chemically from this precursor with either 1,3-dienes or more intriguingly monoalkenes as illustrated, INTENSITY 120 0 3 0 6 0 9 Figure18. Saturation Curves for m-Xylyleneat 0 //• VariousTemperatures. » V w m r e w o p • • 41K * 57 K 57 * 77K 47 48 this despite substantial steric strain. It appears that the chemistry of 8 will prove to be a rich area. Further experiments in this area could include the preparation of additional heteroatom substituted derivatives such as HN. NH 40 Compounds 37-39 are very interesting in that they should all three exhibit noticeably different D7 values since spin density at nitrogen is different in each structure. Thus it would be possible to quantify the electronic effects of hetetoatom substitution on this type of biradi cal. In this regard, a synthesis of 41 was attempted. H Ts H Ts 4 1 4 2 4 3 However samples of the bis(tosylhydrazone) treated with 40% aq KOH/dioxane gave a dark red solution which extracted into the aqueous phase on workup. Assuming that this color is due to the diazocompound, its high solubility in water makes it impossible to isolate by 49 convent ionn ] me.-ins. New methodology for preparing 4J will have to he developed before i] can be studied by this route. An attempt to prepare _40 from Che corresponding known diazide 43 in ethanol by irradiation at 77 K gave only very complicated spectra which defied analysis. Even if 40 were formed the existence of 3 geo metric isomers below might complicate the observed spectrum. In conclusion, j3 has been generated by independent routes and is stable indefinitely at 77 K. in hydrogen atom donating matrices. The spectral parameter lD'/hd=0.011 cm-* is, as expected, between that of Schlenk's hydrocarbon ( ID'/hd=0.008 cm-1)21 and Berson's quinomethane (ID/hcl =0.032 cm —1 ). 31 The extended conjugation of Schlenk's hydrocarbon leads to a lower value for lD‘/hcl , while on the other hand, spin-orbit and electronic perturbations due to the oxygen atom of 19b cause an increase in splitting compared to ni-xylylene itself. Further, it is in agreement with Berson's calculated value of 0.016 cm-*,^ which was obtained by semiempirical means. In this method, spin-orbit interac tions were neglected and the calculated value of |D/hcl was based entire ly upon the spin-spin dipolar interaction. Additional Studies on Biradicals Out of curiousity to better understand the nature of the biradical specie generated on photolysis of 2_5, further study was undertaken. First, to determine the scope of the phenomenon, other diazocompounds were tried. Careful attention was paid to the possiblity of saturation so that negative results (no spectrum in 30 min. of continuous irradia tion, 1000 W high pressure mercury arc, 77 K), can be considered sig nificant. The results are listed in Table 5. As can be readily seen, the phenomenon does occur in some diazo compounds and even diazirines, though selectively. From the appearance of the data, a trend between the observed splittings can be seen, that is, for those compounds which give spectra, the less delocalized sys tems give larger splittings. This is easily explained if one or both electrons are distributed through the aromatic IT'-system in the birad ical. Since radical pairs were ruled out in the case of 25, and since sample dilution of ^5 does not change the signal ratio of biradical to carbene after fixed irradiation time, bimolecular processes involving diazocompounds can be ignored. Thus, the only remaining alternatives are photorearrangements of the diazocompounds or the carbenes. As mentioned earlier though, irradiation of diphenylketene gave only a carbene spectrum. Therefore, a photorearrangement of 25_ is indicated. If one assumes that the same process is responsible for the spectra in 50 51 Che ocher cases, Chen anocher face is apparent. Compounds in which the diazo moiety is fixed relative to the aromatic ring don't give rise to spectra. Bearing this in mind, two structures come to mind as candidates for the spectrum carrier (drawn for phenyldiazomethane). 4 4 4 5 On close examination though, and after viewing models, it becomes apparent that ^4 is essentially planar and would therefore be diamag netic. On the other hand, a spiro fused structure such as 45 accounts nicely for much of the data, both in regard to the trend in splitting values and the inability of the rigid diazocompounds given to form biradicals. Examination of the structure below attempts to show the high strain energy demanded for a spiro structure in these cases. 52 Whatever the structure involved, irradiation at >350 nm was successful in reproducing Table 5. Thus extremely high energy photons are not required to effect transformation. The biradical derived from ^5 has considerable thermal stability in PFA. It showed no decay on warming from 77 K to 203 K at which point a rapid decay suddenly occurred. This behavior is consistent with a high activation energy for decomposition or reaction (whichever is occurring). It is also possible that a phase change may have caused the signal to disappear, in which case it would be even more thermally stable. As indicated in Table 5, irradiation of 1-naphthyldiazirine (46), gave rise to the same spectrum as that observed from 1-naphthyldiazome- thane (47). That this was not the result of photorearrangement of 46 to C C 47, a control experiment was carried out in which the intensity of the carbene and biradical signals were measured at precise intervals during irradiation. If the biradical arose from 47 alone, then the ratio of carbene/biradical would be larger for 47^ than for 46. In fact, the measured ratio was larger for 4^6 then for 4_7 by a factor of 1.4/1. This indicates that relative to the amount of carbene produced, diazi- rines appear to be better precursors for these biradicals. To determine whether other species might be suitable as precur sors, the photochemistry of some aryl azides which can be thought of as azodiazomethanes, has been examined by Wright and Zayas. ^ Indeed, these compounds readily give intense biradical spectra immediately on irradiation. The results are presented in Table 6. The splitting in this series is also consistent with a spiro intermediate such as 45. Of course, less is known about these systems so another type of carrier is Table 5. biradical Spectra from Diazocompounds. Compound Observed Splitting (G) 234 240 hJ\ ( S o 240 220? very complicated 300 < y none JO** none o g o none none OLJO none none none O L J O none Table 6. Biradical Spectra from Azides. Compound Observed Splitting (G) 430 £ 390 ¥NO, 240 420 + 10 NO, none possible. Further study in this matter is urgently needed. If structures of the type shown in 45 are indeed responsible, this formally represents the first experimental observations of a spirarene derivative.^ Although the following compounds have been prepared, they CO show little if any spiroaromaticity. ° F, 5 This finding is in accord with theoretical predictions which indicate that spiroconjugation will be most important for the spira- 59 renes and in particular those shown below. ' 4 8 4 9 50 In all three cases, substantial spiroconjugation is predicted; however, 48 should have the most pronounced effect. Consequently, theory predicts that 48^ will be a singlet while 49 and 50 will be triplets. It is at least an intriguing speculation that such compounds are responsible for the experimental observations presented above. II. SOLID STATE REACTIONS AND KINETICS OF TRIPLET ARYLCARBENES. 56 Historical Background The reactivity of carbenes has been an active field of study since the pioneering work of Hine^® in 1950 on dichlorocarbene. Much has been learned about the properties of carbenes in that time, especially as to how the electronic nature of the carbene influences the outcome of reactions in solution.^ However, since 1971 numerous reports have appeared in the literature indicating substantial changes in product distributions on lowering the reaction temperature. In cases where the temperature is low enough to cause the solvent to freeze, unusual chemistry has been observed for many carbenes. Often products which are formed only marginally in solution can become major components of the product mixture under matrix conditions. 51 52 In the earliest report of such behavior Moss and Dolling62 exam- ined the chemistry of phenylcarbene (52) In 2-butene matrices. These workers found that photolysis of j>I^ in neat 2-butene solutions led to >97% cyclopropanatlon. In the solid state large amounts of allylic CH carbene insertion products were formed (up to 52% at -196 °C). Triplet phenylcarbene was suggested to be responsible for the matrix chemistry 57 58 with Gl insertion occurring by an abstraction-recombination mechanism. Subsequently, Moss and Joyce^^ investigated the efiects of low temperatures on the reaction of with isobutene. Again under matrix conditions a substantial decrease in cyclopropanation was observed with a concomitant increase in the amount of Of bond insertion. These workers also examined the chemistry of fluorenylidene (54) at low temperature.^ ln neat isobutene solution photolysis of 53 gave a 94% yield of cyclopropane j>!5 with only 4% CH insertion product 5b. However, in the solid state the yield of 56 increased dramatically to 39% at the expense of J>5 which fell to 58%. The authors chose to in terpret this behavior in terms of a simple temperature effect on the Singlet-Triplet (S-T) * 5 3 5 5 5 6 equilibrium populations of _54, and, the associated rate constants for formation of 55^ and 56. Fluorenylidene is known to be a ground state triplet,^ so assuming that the singlet-triplet energy gap is greater than the activation energy for triplet reaction, and assuming that triplet 52 cannot undergo addition to isobutene, formation of 5(> from triplet 52 would increase on lowering the temperature. In related work Moss and Huselton^ extended these observations to the reaction of diphenylcarbene (57) in isobutene solvent. Again the yield of S9 relative to 5i8 increased as the reaction temperature was lowered with 59^ being formed in 98% relative yield at -196 °C under matrix conditions, while in solution at 0 °C it was formed in only 20% 59 yield. 25 57 58 59 Further evidence of matrix effects on reactivity was found by Moss and Wetter^ in the case of cyclopropylphenylcarbene. In this case intramolecular carbene rearrangement was responsible for most of the chemistry observed in isobutene solutions, while in the solid state intermolecular chemistry to form cyclopropanes predominated. This was attributed by the authors to a matrix steric constraint on unimolecular rearrangement. ■0- 0C = C -H + y* O Tomioka et al. have expanded the study of solid state carbene chemistry. The steroselectivity of cyclopropanation reaction of phen- ylcarbene derivatives with ring-substituted styrenes were investigated. In every case studied a dramatic change in the cis/trans isomer ratio was observed when a change from liquid to solid phase occurred. This was explained in terms of a stepwise addition of triplet phenylcarbene to styrene to form a 1,3-biradical which subsequently closes. Steric effects on rotation imposed by the matrix could account for changes in trans ci s 60 the cis/trans ratio on going to the solid state. Indeed, photolysis of azoalkanes 60c and 60t to give the postulated 1,3-biradical inter mediate indicated that changes in product sterochemistry do result on going to the solid state when biradical closure is involved (see Figure 19). In a study of the reactions of 52^ in 2-chloropropane, Tomioka et al. 69 found that two kinds of matrix effects were observed. In solu tion, phenylcarbene insertion into the CC1 bond to give 6J^ occurred readily and even Increased at the expense of (62 + 63) as the tem perature was lowered. Cl Ql 51 + ^ " JrL> + 0 u Cl 61 62 6 3 However, when the transition from liquid to solid phase occurred 61 became a minor product, and 62^ and 6JJ predominated. The exact origin of this matrix effect remains a mystery. A second matrix effect resulted in a dramatic increase in the yield of primary CH insertion product in the solid state. This observation was interpreted as strong evidence for steric control of carbene matrix reactions. Cohen and Schmidt^ have coined the term "Topocheraical Principle" to refer to the requirement that reactions in solids must take place with a minimum of atomic and molecular movement. Thus, relative yields in solid state reactions are often determined by transition state perturbations or by the rate of a diffusion process which establishes the right geometry for reaction.Such a diffusion process may be rotational or translational in which case it modifies the size and % Yield Figure 19. Matrix influences on Biradical Closure. O'. 62 shape of the reaction cavity. Since these factors are not important in the liquid phase it is not unexpected that results obtained in the liquid and solid states are substantially different in many instances. Further evidence of topocheraical matrix effects were found by Tomioka and Izawa^ in a study of the reactivity of arylcarbenes in alcoholic matrices. For both phenyl- and diphenylcarbene, in a variety of alcoholic solvents, a dramatic decrease in ethers resulting from carbene insertion into the Oil bond was seen. Correspondingly, the yield of CH insertion products increased under these circumstances (see Table 7). Since OH insertion is widely held to be the product of singlet arylcarbenes with alcohols while CH insertion occurs from the triplet state of arylcarbenes, 73 this change in product ratio may simply re flect thermal effects on the S-T equilibrium population of 57.^ Closer examination shows a dramatic jump in the formation of primary Qi inser tion product ji6 on going to solid state conditions. This has been interpreted^ as evidence of a topochemical effect on H atom abstrac tion processes occurring in matrices. Similar effects have been ob- served in ethers and amines as solvents. 75 The reactivity of carbenes in alcoholic matrices has been shown by Tomioka et al. 7 f\ to be highly dependent on the alcohol used. These workers found that in t-butanol matrices, the dominant product formed was a carbene dimer and not derived from reaction with the matrix (it is possible that this product was formed on thawing the matrix). In deed, when other alcohols were added to the mixture in small amounts (52) the only intermolecular products found were those of reaction with the added alcohol. In contrast, solution phase photolyses in t-butanol gave only the usual OH insertion products. This change in behavior was Table 7. Product Distributions for the Reaction of Phenylcarbene with 2- Propanol as a Function of Temperature. 6 4 6 5 66 Relative Product Yields Temp., °C 64 65 66 0 81 19 trace -72 65 35 trace -110b 39 50 11 -196b 16 68 16 a) Reference 72. b) Matrix conditions. attributed by the authors to crystallization phenomena on formation of the matrix. Even in other alcoholic and alkane matrices Tomioka, Griffin, and Nishiyama^ have demonstrated a dependence of the product distribution on the carbene precursor. Compounds 51, 67, and 68, all of which are 7 o photochemical precursors of 52, gave different product ratios when photolyzed in 2-propanol matrices, while in solution identical product distributions were obtained. Since it is reasonable that carbenes will be generated with different spatial relationships to the host molecules of the matrix when 51, 67, or 68 are used as precursors, these results are not unexpected. 67 68 Further evidence of the ability of topocheraical influences to control carbene reactions in the solid state has been found by To- mioka 79 who studied phenylcarbene reactivity in a variety of hydrogen atom donating matrices. He found that primary CH insertion generally increases under matrix conditions. For example, in isobutane a ratio of (3°/1°) CU insertion of 118 at 0 °C decreases to 4.1 at -196 °C. Similar results were obtained in isopropyl ether and 2-propanol. Since other evidence suggested that CH insertion in these matrices occurred from the triplet state by abstraction-recombination, these results were found to be in contradiction with solution and gas phase results which indicated a much higher regioselectivity should be observed at low temperatures•HO 65 The effects of carbene multiplicity on matrix reactivity have also been investigated by Tomioka et al.®* Photolysis of phenylchlorodi- azirine (69) gave clean OH insertion in methanol even at -196 °C and gave a mixture of cis and trans cyclopropanes in neat cis-2-butene at -196 °C. This contrasts with the behavior of phenylcarbene which gave 40% CH insertion in methanol and 30% allylic CH insertion in cis-2- -butene at -196 °C. This result was interpreted to mean that triplet arylcarbenes react preferentially with CH bonds by an abstraction-re- corabination mechanism while singlet arylcarbenes prefer to insert into OH or add to C=C bonds. It is possible however, that the observed products arise from reactions on warming the sample, especially since on Turro et al. did not report any significant decay occurring when the matrix UV spectrum of 70 was obtained at 77 K. Cl 6 9 70 The effects of carbene multiplicity on reactivity were further explored by Tomioka, Suzuki, and Izawa 83 who examined the effects of ring substituents on the product distributions of arylcarbenes in alcoholic matrices. These workers found that either electron donating or withdrawing substituents enhanced the amount of OH Insertion which occurred in the matrix, and have suggested that two mechanisms of OH Insertion may be operative under these conditions. An alternative explanation would suggest that substituent effects on the S-T energy gap may be responsible for this observation. In any case, it is clear 66 ChaC matrix reactions of carbenes are sensitive to substituent effects. The question of matrix effects on carbene reactivity has been o A extended to <>c-ketocarbenes as well. In contrast to the case of simple arylcarbenes, the chemistry of «<--ketocarbenes in matrices appears to be quite complex. While phenylcarbomethoxycarbene (71) shows a marked increase in Ql insertion into 2-propanol under matrix condi tions (from 4% at -77 °C to 63% at -196 °C), carboraethoxycarbene (_72) itself showed only a very slight increase on going from solution to solid phase. Also, Wolff-rearrangement of was aJ-most completely suppressed under matrix conditions while no effect was seen in the case of 72. To account for the difference in reactivity of 7^ and 72^ it was suggested that CH abstraction in these systems may also occur from the singlet excited state of the carbene. If the S-T gap in Tl_ were small, then reaction could occur in the matrix via the singlet state in the case of 72_ while a large S-T splitting for 7_1 would be due to typical triplet carbene reactivity. ° 0 - ^ C O O H H/N'COOH J2f^\p(OMe)2 71 7 2 7 3 In the case-of phenylphosphorylcarbenes (73), Tomioka et al.®^ have shown that as for arylcarbenes in alcoholic solutions OH insertion predominates. However, at low temperatures in matrices a marked in crease in CH insertion was observed at the expense of the OH insertion process. Also, as for simple arylcarbenes, dramatic increases in pri mary CH insertion were observed on going from solution to the solid phase when 2-propanol was used as solvent. 67 On the basis of the foregoing results it can be safely concluded that matrix effects on carbene reactivity in the solid state do exist, and the exact cause of these effects remains conjectural and may vary with solvent and carbene. O /' Senthilnathan and Platz have probed the nature of matrix effects on carbene reactions using an ESR kinetic approach. In this method carbenes were generated in the cavity of an ESR spectrometer under matrix conditions by photolysis of diazo precursors, and the ESR signal was monitored as a function of time. These workers found non-exponen tial decay in every matrix studied (i.e.- the apparent pseudo first order rate constant continuously decreases with time). Willard®^ has used the term composite first order kinetics to describe this type of kinetic behavior. Senthilnathan and Platz,®® and Lin and Gaspar®® have found that it is necessary to plot the ESR signal intensity vs. t ^ ^ for glasses or t ^ ® for polycrystalline matrices in order to obtain good kinetic fits over more than 50% of the signal decay. Similar behavior has been observed for the decay of alkyl radicals®® in ma trices and has been interpreted as arising from a distribution of site reactivities within the matrix.®^’®®’®^ Q(L Additionally, Senthilnathan and Platzow demonstrated that the observed rate of carbene decay in matrices depends on the concentration of diazo precursor, length of photolysis time, and the chemical nature of the matrix. These workers also showed that deuteration of the matrix could substantially slow the kinetics and have studied this effect for a number of solvent/carbene systems (see Table 8). The low Arrhenius parameters of Table 8 presented some difficulty in interpretation. On the one hand they might simply reflect Table 8. Arrhenius Parameters for Some Matrix Carbene Reactions. Carbene Matrix logA E Temp. a (kcal/mol) K Diphenylcarbene toluene 1.24 1.9 77-99 toluene-dg 3.36 2.9 77-109 ether 5.4 2.9 77-93 ether-d^Q 7.4 4.0 77-98 Fluorenylidene toluene 0.85 1.5 77-88 1-Naphthyl- toluene -0.12b 0.65b 77-91 carbene 1/3 a) Reference 85. b) Prom a plot of intensity vs t topochemical matrix effects on reaction. Alternatively, it was sugges ted that reaction of the carbene with the matrix might be occurring by II atom abstraction via a quantum mechanical tunneling mechanism. This was supported by increases in the log A values on deuteration of the matrix. In the case of 57 in toluene and toluene-dg, a value of log (A(D)/A(H)) = 2.2 was obtained (see Table 8). Caldin^ has suggested that values of log (A(H)/A(D)) > 2 are indicative of tunneling pheno mena . q i 1 Senthilnathan et al. further examined the reactivity of aryl carbenes in matrices and found that satisfactory fits of observed rate constants to calculated values could be obtained using a one-dimension al tunneling model. However, these results were not conclusive evidence for a tunneling mechanism in carbene matrix reactions. To further develop and test this model a series of experiments described in the next section were undertaken in this laboratory. Results and Discussion The use of ESR kinetic methods to probe carbene reactions in the solid state is largely limited by the ability of the experimentalist to interpret the results. The composite first order decay observed in these experiments creates difficulty in extracting useful information. Two approaches can be used. In the first, plots of ln(Intensity) vs. t 1/2' or t 1/3' are used to obtain linear fits, however, this yields rate constants in matrix units which can be difficult to interpret. Another method involves studying the signal decay curve over a short interval (20%), where the kinetics can be satisfactorily described using pseudo first order methods. Although this approach avoids the necessity of matrix units it suffers from a time dependence of the observed rate constant on the photolysis time required to generate the signal. In typical ESR kinetic measurements a sample tube containing a photochemical carbene precursor (usually a diazomethane derivative), dissolved in a frozen matrix is irradiated briefly in the cavity of the spectrometer. The light is then shuttered and the signal decay is monitored. Although the displayed spectrum is that of the signal deriv ative it can be shown for Lorentzian or Gaussian lineshapes that at the derivative signal maximum the rates of signal decay of the absorption and derivative spectra are identical. Usually, the low field carbene signal is monitored by fixing the field frequency and scanning the first 20% of the signal decay using the chart recorder scan speed to 70 71 record time. In this way a decay curve is drawn directly on the chart paper. However, although one can observe carbene disappearance, one does not necessarily know the chemistry of the decay process. ESR yields no information as to what types of products are being formed. In the case of carbene matrix reactions the observation of matrix isotope effects has been taken as evidence that the decay process observed by ESR is a reaction between the carbene and the matrix.®** Available evidence from the product studies of Moss and Tomioka 7 ^ mentioned earlier suggests that triplet carbenes largely undergo H atom abstraction reactions with the matrix. As will be detailed later, this interpretation is supported by the ESR observation of matrix radical pairs and, by the fact that no decay occurs when the carbene signal is generated in perfluorinated matrices from which atom abstraction is not favorable. The product and kinetic studies indicate that carbene decay occurs by reaction with the matrix. However, ESR is a very sensitive technique which may detect a minor process that is not indicative of the bulk chemistry of the sample. To investigate this point, a study of isotope effects on the kinetics and products in the reaction of diphenylcarbene with 2-propanol was undertaken. As can be seen from Table 9 the chemi cal and kinetic isotope effects are mutually consistent. Within experi mental error there is no OH(D) kinetic isotope effect, while substan tial CH(D) kinetic isotope effects are apparent in 2-propanol-2-dj and 2-propanol-l,1,1,3,3,3-d^ matrices. Similarly, the product distribu tions show virtually no OH(D) isotope effect, but a large CH(D) isotope effect on the yield of OH insertion products. The formation of primary CH insertion adducts is almost completely supressed in 2-propanol- 72 1,1,1,3,3,3-d^ and 2-propanol-dg matrices, and the relative yield of tertiary CH insertion adducts is reduced by 75% in 2-propanol-2-d^. This correlation between the observed chemical and kinetic isotope effects gives one confidence that the ESR kinetic method is not focus ing on minor chemical processes but rather is representative of the bulk chemistry of the sample. The results of these studies are also consistent with the findings 73 of others that insertion of arylcarbenes into the OH bond of alcohols occurs from the singlet state of the carbene. The lack of a substantial OH(D) isotope effect (such as that seen for CH(D) isotopic substitu tion) , makes abstraction of the OH hydrogen by triplet diphenylcarbene extremely unlikely. This is expected when the large energy for OH bond homolysis of 103 kcal/mol is considered. One might argue that correla tion between the chemical and kinetic isotope effects is coincidental and peculiar to diphenylcarbene in 2-propanol matrices, however the same correlation was demonstrated for a number of carbenes in isotopic- ally labelled methanols in a related study to be presented later (see Tables 18 and 19). Thus, the ESR method appears to be generally capable of observing the predominant chemical processes occurring in the solid state. The data also indicates that the reaction of diphenylcarbene with 2-propanol proceeds via the ground state triplet; H 3 • • R* 73 rather than via the low lying singlet; ' 0 ^ 0 - ^ 0X 0 In 2-propanol-dg the product studies indicate that the chemistry is predominantly that of the singlet. Yet, the rate of singlet chemistry in 2-propanol-dg is imperceptibly slow relative to the rate observed in 2-propanol, where triplet carbene chemistry prevails. The observation of deuterium isotope effects are also inconsistent with a singlet carbene interpretation. The singlet-triplet energy separation of di^ phenylcarbene is reported to be 3-5 kcal/mol.^ Thus, at 77 K the ratio of singlet/triplet carbene will be less than 10-®. Conversion of the triplet to the higher lying state at 77 K should be rate determining, followed by more rapid singlet reaction. In the latter scenario, car bene insertion into a CD bond follows the cate limiting step and no isotope effect should be observed. The observed pseudo first order rate constants for carbene matrix decay as measured by ESK are dependent on a number of experimental variables. It is a general phenomenon that as precursor photolysis time is increased, kggg decreases. This can be accounted for by assuming that a distribution of sites exists within the sample matrix. When the light source is turned on carbenes begin to form, however they can also decay. Only those matrix sites which allow reaction to occur on the time scale of the photolysis interval will lead to appreciable carbene signal decay while the light is on. The longer the light is left on, the wider is the range of matrix sites which decay. The remaining Table 9'. Chemical and Kinetic Isotope Effects in the Reaction of Diphenylcarbene with Isotopically Labelled 2-Propanols at 77 K. H OH MATRIX Ph> j _ U CH3 t J'20M,Nl 1 P h ^ CH2CH0HCH3. p i . 5 * - 0 Pr P h '^ " ^ " C H 3 OH 1 CHj -CHCH j 6 i 1 24.5 54.5 20.9 OD 1 CH3 -CHCH 3 7.5 t 1 22.8 52.4 . 24.8 OH 1 CD3 -CHCD 3 20 .5 i <1 26.4 73.6 TRACE OH 1 CH3- COCHj 15 i 3 45.1 13.6 41.3 OD CD3- CDCDj 2920 ♦ 45 67.9 32.1 TRACE a) Products normalized to 100%, corrected for GC response factors. b) Measurements made by Dr. V.P. Senthilnathan, Reference 91a. 75 carbene-matrix orientations give rise to the observed signal. At longer photolysis times these orientations are associated with slower rate constants and the value of kgy<, for 20% decay increases. Table 10 illustrates this time dependent behavior of ^q b s * To counteract the effect of this time dependency, photolysis intervals were carefully regulated so that a constant interval was used over the entire course of an experiment (e.g.- temperature dependence studies). Caution should be exercised in any attempt to compare data from different photolysis times as this can be misleading. Another effect which needs to be addressed is the dependence of koBs on sample concentration. At high concentrations the observed kinetics can become drastically slowed, presumably by the formation of microcrystals on cooling.^ Such crystals would have chemical envi ronments which are quite different and possibly less reactive than the solvent. The effects of concentration on carbene matrix lifetimes can be seen in Table 11. To avoid the complications of such effects all samples used in ESR studies reported herein were less than or equal to 0. lli. With a knowledge of the effects of experimental variables on the observed rate constants, a series of experiments were undertaken to explore the effects of the matrix environment on carbene reactivity. In particular, it was of interest to determine the validity of the multi ple site model as the source of nonexponential ESR signal decay as well as the suggestion by Senthilnathan 86 ’ 9 la that quantum mechanical tun neling was involved in carbene-matrix reactions. In order to study CH abstraction-recombination processes, the ESR kinetics of a number of arylcarbenes were investigated in toluene. It Table 10. Time Dependence of k on Photolysis ----- "T.Time • Interval.r - i 3 OBS hv time, s k0BS’ 10 S 25 12.2(15) 50 8.62(31) 75 6.60(25) 100 4.99(4) 150 3.74(3) 200 2.99(15) a) 0.1 M Diazofluorene/ toluene, 20% decay, 77K. 3 Table 11 . Concentration Dependence of (Diazofluorene] Cl/5 M s 0.034 24(3) 0.067 26(3) 0.10 26(3) 0.25 24(3) 0.30 b 0.50 b 1.00 c a) 50 s hv, 77 K, diazofluorene in toluene. b) 5% Decay observed in 16 min. c) No decay observed in 16 min. has been shown that diphenylcarbene reacts with toluene at the benzylic position by abstraction-recombination.^^ This is evident from the observation of crossover products resulting from radical-radical diffu sion (e.g.- tetraphenylethane and bibenzyl). Since there is no reason to believe ja priori that this chemistry will change in the matrix it will be assumed that the observed carbene decay in toluene matrices results from CH abstraction. This assumption is supported by the obser vation of ESR spectra attributable to triplet radical pairs on photol ysis of some arylcarbenes in toluene matrices. This is discussed fur ther at the end of this chapter. R S R SI / C1 H * R The results of ESR measurements on the reactivity of several arylcarbenes in toluene matrices are reported in Table 13. From here on all carbenes discussed will be abbreviated for simplicity according to Table 12. Based on this information it is possible to assign an approx imate reactivity sequence of arylcarbenes towards matrix CH reaction as follows: DMA » FL > 9-AC > DBT.DPC > DBS Anthronylidene (A) was not included since severely nonexponential decay was observed. Also, in this case it is not clear whether abstraction occurs exclusively at the benzylic position since this carbene is known OA to abstract hydrogen atoms from benzene. The carbenes reported in Table 13 all exhibited unusually low Arrhenius parameters. These are shown in Table 14 and may be indicative Table 12. Abbreviations Used for Carbenes Presented - " ' 1 g in the Remainder of Part II. Abbreviation Structure DBS © c © DBT © O © A © c © o FI DPC © r © DMA © c © 1-NC •V. 9-AC a) All carbenes In this table have been previously observed by other workers and shown to have triplet ground states except DMA which is a new compound. DMA was found to be a ground state triplet, but may have a low lying singlet as evidenced by curvature in the Curie plot. Table 13 . Carbene Rate Data in Toluene. Carbene hv time Temp. Cone. °C s-1 M DBS 250 -196 1.26(3) 0. -183 5.26(17) -178 9.-46(18) -170 21.1(1) -163.5 47.7(11) 5 -1S& 10.0(5) DBT 250 -196 2.10(3) -183 9.40(53) -178 15.5(1) -173 25.2(3) -168 45.1(23) 75 -196 4.38(17) -185 16.3(19) -183 19.4(13) -180 20.0(22) -177 21.9(18) 25 -196 7.60(21) -185 SI.1(57) -183 36.4(27) -180 36.9(20) -177 44.6(10) A 250 -196 5.37(8) FI -173 0.650(28) 0.3 -168 1.79(26) -163 2.35(14) -158 4.67(18) Table 13. Carbene Rate Data in Toluene, (cont.) Carbene hv time Temp. K-.no Cone. -2BS-i s °C 10 s M FI 250 -153 4.67(18) 0.3 -196 22.6(22) 0.1 -189 32.7(35) -187 45.8(10) -185 54.3(32) -183 74.8(23) 25 -196 122(15) 50 86.2(31) 75 66.0(25) 100 49.9(4) 150 37.4(3) 200 29.9(15) 9-AC 250 4.06(9) -183 10.8(4) -178 17.2(9) -173 26.8(15) -168 40.0(29) DPC 50 -196 3.65(14) -183 25.4(16) -178 41.5(22) -173 82.7(31) -168 89.6(93) 250 -196 2.10(1) -183 10.7(4) -178 18.1(11) -173 26.2(25) -168 48.6(48) Table 13. Carbene Rate Data in Toluene, (cont.) Carbene hv time Temp. Cone. s °C 10i o - V s - i M DPC 1000 -196 1.17(2) 0.1 -183 5.41(4) -178 8.60(18) -173 13.8(20) DMA -196 a a) Undetected at this temperature, but can be seen under steady state conditions at- lower temperature. LN(K) Figure 20. Arrhenius Plot Showing Tunneling Corrections. Tunneling Showing Plot Arrhenius 20. Figure 1/T Tunneling Correction Tunneling Deuteron Proton 82 83 of a tunneling mechanism. Readers who are unfamiliar with the basic principles of tunneling are referred to Appendix I. In brief, tunneling refers to the passage of a particle through (not over), a potential barrier. Bell^ has noted that increases in the reaction rate at low temperatures due to a tunneling correction can lead to curved Arrhenius plots (see Figure 20). For experiments carried out at low temperature, the apparent values of Ea and log A appear much lower than if the experiment had been carried out at high temperature. The low log A and E values of Table 14 are reminiscent of those a obtained by Senthilnathan et al.®^ in their studies on carbene matrix reactions. However, one might argue that the low Arrhenius parameters are an artifact of the matrix and/or the ESR method. Certainly, some lowering of the parameters is introduced by the method of measurement. At low temperatures where decay is slowest the fraction of matrix sites observed by ESR contains a relatively larger concentration of reactive sites than at warmer temperatures where decay is rapid and substantial decay occurs during photolysis. Thus, if one visualizes an average rate constant corresponding to a single site at the median of the tempera ture interval studied, then the rates observed at higher temperatures will be too slow while the rates at lower temperatures will be too fast. It seems unlikely though that this effect alone could result in such a dramatic lowering of the Arrhenius parameters (a decrease of 5-9 orders of magnitude in A in toluene matrices as compared to solution), when the effects of photolysis time on kQgg are less than one order of magnitude in the matrix. Alternatively, it is possible that the low parameters may result from the high matrix viscosities encountered (typically 1 0 ^ - 1 0 ^ 84 Table 14 . Arrhenius Parameters for the Reaction of Arylcarbenes with Various Solvent Matrices. Carbene hv time Solvent logA E a Rb s (kcal/raol) DBS 250 Toluene 1.2 1.8 .992 DBT 1.2 1.7 .996 75 0.4 1.3 .997 25 0.9 1.4 .975 FI 250 0.7 1.2 .956 9-AC 0.3 1.3 .992 DPC 50 2.0 1.9 .994 250 1.3 1.7 .998 1000 0.7 1.6 .999 DBS 100 c d 3o d 0.6 2.4 .953 DBT 2.7 1.8 .996 DPC 1.8 1.8 .983 DBS 50 CC1. 2.7 3.6 4 .993 DBT 0.6 1.8 .919 FI 5.0 3.8 .990 DPC 4.6 3.8 .988 a) 0.10 M except for CD^OD entries which are 0.01 M. b) Corre lation coefficient, 1 is perfect and 0 is random. 85 poise), 96 making molecular movement very difficult. Still, the prospect that tunneling might be involved in these reactions provided sufficient motivation to attempt to fit the experimental data to a one-dimensional tunneling model. 97 Leroy et al. have recently outlined a simple extension of transition-state theory which adequately describes the rate of hydrogen atom abstraction by tunneling in low temperature solids. Assuming that the abstraction reaction can be represented as the motion of the hydro gen atom along a linear path, these workers represented the potential surface for this process as shown in Figure 21. Within the reactant V(K» o i ( reaction coordinate) Figure 21. Eckart Potential Barrier potential well, only discrete energy levels are allowed corresponding to vibrational quantum states of the S-H. bond to be broken. The barrier to reaction is given by the Eckart potential function for which barrier permeability as a function of the barrier shape is analytically known. Thus each vibrational quantum state is associated with a particular tunneling rate governed by the barrier width at the classical turning point. By summing over all possible vibrational levels and allowing for 86 a Boltzmann distribution of individual systems the average microcanon- ical rate can be obtained. This is the observed macroscopic rate con stant for tunneling and is given by, ot} J P(E)F(E)exp((E -£ )/kT) dE o 6.o where P(E) is the transmission probability F(E) is the number of vibrational states with energies € n - (n+l/2)h>) < E. QO The Eckart potential function is given by, V(x) = A exp(x/b) ~ . B exp(x/b) (1 + exp(x/b) (1 + exp(x/b)) where B is the endothermicity of the process of crossing the barrier. The barrier height VQ is given by, VQ = (AfB)2/4A and b is the barrier width at half height. McCurdy 9 9 has incorporated this model into a computer program which determines Eckart barrier parameters by minimizing the difference between calculated and experimental rate constants. This program was used by Senthilnathan^3 in his study of DPC matrix reactions. Application of this computer program to the experimental results in Table 13 gave the minimized Eckart barrier parameters of Table 15. The values of FVAL (FVAL is the sum of the squares of the residuals of the data points to the calculated curve), are comparable to those obtained by Senthilnathan, 9 la and Leroy and Williams 9 7 and indicate that reasonable fits to the data are obtained in each case. For visual reference plots of calculated tunneling rates vs. experimental rates Table 15. Calculated Eckart Barrier Parameters for »'■■■ - 0 Various Arylcarbenes in Toluene Matrices. Carbene hv time V b FVAL s kcal/raol A° xlO DBT 25 • 12.83 1.169 5.28 75 13.13 1.150 5.12 250 13.24 1.158 3.45 DPC 50 12.90 1.179 14.1 250 13.18 1.164 5.16 1000 13.72 1.117 9.56 DPCb 50 13.45 1.158 0.652 100 13.77 1.131 0.561 1000 14.95 1.029 1.61 9-AC 250 13.66 1.096 0.969 FI 250 12.42 1.199 0.500 F1C 250 15.20 1.034 4.01 a) 0.1 M, Calculated from a computer program furnished by Prof. C.W. McCurdy, b) 0.3 M in 2-propanol, c) 0.3 M. Figure 22. Calculated Tunneling Rates for DBT in Toluene. in DBT for Rates Tunneling Calculated 22. Figure -4.000 -2.000 fl.OOO 2.000 1]. LI0D 0.000 EZ70YID>E 2 5 HV/TO 5 25 DE>iE/ D i 3ENZ37R0PYLI 50.00 TKELVIN 150.0 88 Figure.23. Calculated Tunneling Rates for DBT in Toluene. in DBT for Rates Tunneling Calculated Figure.23. LOG (K) o o o =r o cr c OJ O Q o i 0.000 D I ENZ 3 0 T R G P T HV/TO LI D S E :\! 5 7 E / 50.00 T KELVIN 100.0 150.0 89 Figure 24. Calculated Tunneling Rates for DBT in Toluene. DBT for Rates Tunneling Calculated 24. Figure LOG (K) o CD o I D I Be. N 2 0 I riOPTL I D L N L / 2 b O 5 5 O b 2 / L N L ID ID riOPTL I Be. 0 N 2 50.00 TKELVIN HV/lO 150.0ICO. 0 91 150.0 103.0 N t N / 0 b 5 / ri V I 0 T KELVIN T el D I P ri E N T L C R R B B R D R I L T C N E ri P 0.000 nnirh ooo '2 nno’o ooo'S- ooo'fi- ooo-9- Figure 25. Calculated Tunneling Rates for DPC In Toluene. Figure 26. Calculated Tunneling Rates for DPC in Toluene. in DPC for Rates Tunneling Calculated 26. Figure - 6.000 - 1.000 - 2.000 0.000 2.000 0.003 H P I D l B R R C L Y N 50.00 l T KELVIN N l / 2 Hb 3 V0 / I 0 100.0 150.0 92 Figure 27. Calculated Tunneling Rates for DPC In Toluene. In DPC for Rates Tunneling Calculated 27. Figure LOG IK) o o 0.000 E N E B n R C L T N E h P I D 50.00 TKELVIN C 0 HV/ i 0 H V / S 1C 0 150.0 93 94 150.0 1 CO. 0 CO. 1 / d 35 - '/ / I — p — I / / '/ 35 - d (Ml 901 Figure 28. Calculated Tunneling Rates for DPC in 2-Propanol. Figure 29. Calculated Tunneling Elates for DPC In 2-Propanol. In DPC Elates for Tunneling Calculated 29. Figure LOG (K) o o o o o o o.ooo D I P ri E N Y L C S n 3 E 1 riV/I-P !\ / 003 E 50.00 T KELVIN 100.0 150.0 95 Figure 30. Calculated Tunneling Rates for DPC in 2-Propanol. in DPC for Rates Tunneling Calculated 30. Figure -5.000 -U.000 -3.000 -2.000 -1.000 0.000 0.000 D I P H l ! \ i T L C R n 3 50.00 1 T KELVIN N E / 1N 0 E 3 0 ri 5 V / I 150.0 96 Figure 31. Calculated Tunneling Rates for 9-AC in Toluene. in 9-AC for Rates Tunneling Calculated 31. Figure -S.000 -4.000 -3.000 -2.000 -1.000 0.000 I.U00 2.000 0.000 20 nV/lO 5 /250 E N E B n R C L U r H T N R - 9 TKELVIN 1C0.0 150.0 97 iue3. acltdTneig ae frF (. ) in M) (0.1 FI for Rates Tunneling Calculated Figure.32. -G. 000 -G. 000 -U.000 -3.000 -2.00CI -1.000 0.000 20 riV/TO S /250 E N E D I L T N E R 3 U L F Toluene. 50.00 TKELVIN IC'0.0 150.0 98 Figure Figure -4.000 -2.000 Cl.UOO 2.000 4.000 0.000 20 HV/T0 5 /250 E N E D I L T N E R Q U L E 33.. acltdTneig ae frF (. ) In M) (0.1 FI for Rates Tunneling Calculated Toluene.. 50.00 T KELVIN 100.0 150.0 99 are given in Figures 22-33. Inspection of Table 15 shows that a trend exists between the photolysis time and the Eckart barrier parameters. Barrier heights increase with photolysis time while barrier widths decrease. Recalling that short photolysis times are associated with fast matrix sites while at longer times slower sites are observed, and if the tunneling model is qualitatively correct, this means that fast and slow matrix rate constants arise from changes in site geometry. A simple way to explain this trend involves the trajectory of the n H atom being abstracted relative to the carbene sp orbital. Schae fer has calculated that the preferred geometry for U atom abstrac tion by methylene has the migrating hydrogen atom approaching the o carbene sp orbital end on as shown in Figure 34. Since this will not be possible in all sites of the matrix, other geometries must be consi dered as well. H-S fast site slow site Figure 34.. Possible Trajectories for H Atom Abstraction. In these cases, the activation energy will be higher in agreement with the observed trend in VQ . To understand why b changes as it does it is necessary to consider the conditions under which these measure ments are made. From an experimental standpoint only processes which occur on the time scale of the experiment can be measured. For sites in 101 which the trajectory for abstraction is favorable, narrow barrier widths will lead to reaction before the irradiating light source is turned off. Therefore wide barriers will be observed for these sites. On the other hand at long photolysis times, even these sites are deple ted and the contribution of matrix sites with poor trajectories to the observed kinetics is enhanced. In this case a small barrier width is observed since larger widths react too slowly to be monitored on the time scale of the experiment. The results in Table 15 are intended for use in a qualitative manner due to the approximate methods used in obtaining them. Still, they can be very informative if the tunneling hypothesis is correct. At this point any objections to the validity of the tunneling approach must account for the fact that good fits are obtained for several carbenes and in two solvents. They must also account for the nice agreement found with the matrix site model. Further, tunneling has been clearly implicated in a variety of other H atom abstraction reactions. Willard and coworkershave extensively investigated the kine tics of U atom decay in organic matrices at low temperature. In alkane glasses decay is rapid even at 5 K. In fact it was necessary to use perdeuterated matrices to slow down abstraction sufficiently so that kinetic measurements were possible. In these latter matrices H atoms are visible even at 77 K. Results of these and other measurements have shown that tunneling is indeed involved. For the most part tunneling in these cases has been implicated by the observation of extremely low activation energies. Observation of E <50 cal/mol for the decay of H cl atoms in 3-methylpentane-dglasses over the temperature range of 4—50 K indicates a fundamentally different mode of reaction than in solution 102 where values of Ea >6000 cal/mol are observed for similar processes. Other evidence of tunneling in these experiments includes the observa tion of sizable matrix isotope effects. Toriyama and IwasakiiU109 have compared Willard's results to a computational tunneling model similar to the one used here with good success. Their ability to fit the data closely was taken as strong additional evidence for tunneling. A tunneling mechanism also appears to be involved in the atom transfer processes of alkyl radicals trapped in organic solvents. Methyl radicals generated in methanol glasses from 15-100 K yielded a decidedly curved Arrhenius plot with the observation of temperature independent rate constants below 40 K. Similarly, kinetic Isotope effects in excess of 1000 at 77 K have been observed, as well as inor dinately low preexponential factors ( A ^ q ^ -1.1 while A - 1 0 ^ has been established for the reaction in solution). Similar results have been observed for methyl radicals in acetonitrile and methylisocyanide matrices. In the case at hand the tunneling model makes some experimentally verifiable predictions. From Figures 22-33 it is apparent that at temperatures below 50 K little or no change in kggg with a change in temperature is expected. Although such an experiment is not impossible it is beyond the capabilities of equipment currently in this labora tory. Another prediction lies in the fact that heavy nuclei do not tunnel efficiently. This combined with the results of Scaiano et al.*®**, who have studied the solution chemistry of DPC with toluene and CC14, can be used as follows to provide strong evidence of tunneling in carbene matrix reactions. 103 At 300 K the second order rate constants for UPC decay in toluene and CCl^ are A.8 x 10^ s- ^M- * and 3.7 x 10^ s * respectively. The Efl and log A values for these reactions are, toluene E = 3.2(7) kcal/raol log A = 8.0(5) Cl CC14 Ea = 2.2(7) kcal/mol log A = 8.2(6) As can be clearly seen from the Arrhenius equation k^gg is de creased by decreasing A or increasing E , a k0BS = A exp(-Ea/RT) Since A(toluene) < A(CC14) and Ea(toluene) > Ea(CCl4) the value of kQgg(toluene) < ^Qgg^Cl^) regardless of the value of T. At 77 K. the values of kEg,g calculated from the Arrhenius parameters obtained in solution are, kgg.j,( toluene) “ 1 x IQ"1 s- *M- * kEsT (ccl4) - 1 x 102 s-1M— 1 Of course these results strictly apply only to solution phase measure ments (at 77 K. both solvents are frozen), still the same reactivity pattern can be expected. However, these calculated values are in stark contrast to the experimentally observed values in Tables 13 and 16. In carbon tetrachloride matrices very sluggish reactivity was observed. This is consistent with Tomioka's results which indicated that CL atom abstraction is not a particularly favorable reaction of triplet aryl- carbenes.^ After correcting the pseudo first order results in Tables 13 and 16 for solvent molarities the following rate constants were obtained: kQgg(toluene) “ A x 10-8 g-l^-l koBS(cclA> “ 2 x 10-5 s-1^ 1 The observed rate constant for the reaction of DPC with toluene is over Table 16. Carbene Rate Data in CCl..a ------4 Carbene hv t irae Temp. s °C 10io-?BS-i s DBS 50 -152 1.67(10) -146 3.52(27) -137 7.26(124) -129.4 19.5(6) DBT -163 9.69(24) -158 15.0(7) -150.8 16.3(13) -146.1 35.0(1) FI -169.7 10.2(5) -165.7 20.6(10) -160 51.0(55) -155.4 96.6(98) 200 -196 0.00051(37) DPC 50 -169.5 8.24(13) -163 8.82(25) -158 18.2(3) -151.2 29.4(15) 200 -196 0.0040(49) DMA -196 b -268 b a) 0.10 M. b) Not detected, but sample was discolored and bubbled vigorously on thawing. 105 1000 dimes that found with CCl^. When this result is compared to the calculated values (kgg>£.), a reversal i-n reactivity of six orders of magnitude is obtained at 77 K! This cannot be accounted for by the usual matrix site problem arguments since the availability of Cl atoms in CCl^ matrices must match or exceed the availability of benzylic U atoms in toluene. Additionally, this result is consistent with Tomioka's study of phenylcarbene in 2-chloropropane^ discussed in the previous chapter. The bond strengths of CCl^ and toluene (benzylic CH) are 73 kcal/mol and 85 kcal/mol respectively. According to the Hammond Postulate this increase in product stability should be reflected in the relative transition state energies. Thus abstraction of CL atoms from CC1A is predicted to be energetically more favorable than benzylic H atoms of toluene. Koth^^ has demonstrated by CIDNP experiments that singlet meth ylene abstracts Cl atoms preferentially to H atoms from l-chloro-2-me- thylbutane in solution. If carbene matrix processes observed by ESR proceeded through the singlet state of the carbenes then reaction with CC14 matrices should be faster than with toluene matrices. The obser vation of the opposite trend supports a triplet matrix reaction of arylcarbenes with matrices. The reader should note that this unexpected behavior is directly predicted by the tunneling model used above. Since chlorine atom tunneling need not be considered when looking at Cl atom transfer the kinetics of arylcarbenes in CCl^ matrices was investigated. The experimental results are given in Table 16. Although none of the carbenes studied had appreciable reactivity at 77 K (except 106 possibly DMA), it was possible to observe signal decay by warming the sample. As can be plainly seen much higher temperatures are required in CC14 than in toluene to see pseudo first order decay at comparable rates. The kinetics in CCl^ are still composite first order despite the apparent matrix uniformity one might expect. Although the values of log A (Table 14), are much lower than in solution, they are generally higher than in the case of toluene matrices. Certainly the activation energies seem to be in the right range when compared with the solution for DPC in CCl^ of 2.2(7) kcal/mol found by Scaiano.^^ So it appears that indeed a classical process is being observed in CCl^ matrices in which unusually low log A values are seen. Whether the matrix reaction of carbenes with CCl^ occurs from the triplet or singlet state of the carbene can not be determined at this time. The results of this experiment do show that taken alone, low Arrhenius parameters in matrix reactions are not conclusive evidence for the existence of tunneling processes. During the investigation of DPC/2-propanol products at 77 K it became clear that CH bond deuteration reduced the amount of chemistry occurring with those CH bonds and increased the yield of singlet de rived OH insertion adducts. It was felt that if an alcoholic matrix could be found where OH insertion could be made to happen exclusively, it would be possible to study pure singlet state reactions by the ESR matrix method. Methanol was a logical choice for such a solvent due to its availability in a variety of isotopically labelled forms. Further, due to its small size it contains more OH groups per CH bond than other aliphatic alcohols. Thus, if carbene reaction with CH bonds were slowed by deuteration, the OH bond might still be close enough to react in the 107 matrix. Also, the simplicity of the system meant that products could be easily analyzed and authentic samples prepared. With this in mind the chemistry of UPC, UBS, L-NC, UBT, and FL was explored in isotopically labelled methanolmatrices. The results are presented in Tables 17 and 18. In each case substantial increases in the relative yield of OH insertion product were observed on deuteration of the CH bonds of methanol. On the other hand, little product isotope effect was seen on deuteration of the OH bond. The chemical isotope effects are also reflected in the kinetic behavior of these systems. Negligible 0H(D) isotope effects are observed, but large CH(D) isotope effects are evident. Several mechanistic possibilities have been suggestedto account for the formation of OH insertion products from alcohols. The results of the above experiments allow some comments to be made on their validity. If one assumes that singlet-triplet equilibration is rapid rela tive to singlet reaction, then an abstraction-recombination mechanism such as that shown in Figure 35 can be ruled out. This follows from the fact that one would expect larger chemical and kinetic isotope effects than those in Tables 17 and 18. Of all the carbenes examined, DBT should be the most likely to react via this mechanism due to its elec tronic character. Aromatic stabilization of the singlet carbene and the product ion pair should make this mode of reaction more favorable for this than typical arylcarbenes, yet no appreciable isotope effects were observed. Also, FL which should not react easily by this mechanism due to the antiaromatic character of the product ion'pair, gives rise to even larger relative yields of OH insertion products. All these facts tend to rule out this mechanism in carbene OU insertion processes. 108 R R \ \ + c: hor ► /C_H + ~OR" R R I. rv h b' “ o r " H ROH ► coiiro Figure 35. Proton Abstraction Mechanism of Carbene Insertion into Alcohols. R u R R H V H \- H \| t* c: bR" ------► c -o r " ------► c-o r / / + / R R R Figure 36. Ylid Mechanism for Carbene Insertion into Alcohols. Table 17. Carbene Rate Data in Isotopically Labelled Methanols. Carbene Solvent Temp. Cone. °C 10 s M DBS MeOH -196 9.57(16) 0.010 MeOD 8.47(64) MeOH-d. 0.0713(20) 4 -171 0.241(31) -167 0.238(49) -157 1.46(8) -146.5 3.61(3) -135 3.52(63) DBT MeOH -1$6 64.5(18) 0.10 MeOD 66.6(15) MeOH-d. 1.14(3) 4 0.010 -187.6 106(4) -185 162(2) -183 217(15) -180 299(22) A MeOH -196 0.402(16) MeOH-d. b 4 FI MeOH 129(11) MeOD 125(9) MeOH-d4 80.5(51) DPC MeOH 22.9(16)c MeOD 22.2(l)c MeOH-d> 6.74(21) 4 -179 36.1(24) -175 81.2(44) -171 130(10) DMA -196 d a) 100 s photolysis time, b) Immeasurably slow, c) 40 s photolysis time. d) Not detected. Table 18. OH/CH Bond Insertion Product Ratios for Various Arylcarbenes in Isotopically Labelled Methanolic Matrices at 77 K. Carbene Solvent 0Hb CHb,C insertion insertion DBS MeOH 31.7(6) 68.3(6) MeOD 23.7(12) 76.3(12) MeOH-d. 60.7(10) 39.3(10) 4 DBT MeOH 75.8(25) 24.2(25) MeOD 64.2(14) 35.8(14) MeOH-d. 93.3(9) 6.7(9) 4 1-NC MeOH 82.6(2) 17,4(2) MeOD 63.0(8) 37.0(8) MeOH-d, 100 0 4 FI MeOH 95.8(5) 4.2(5) MeOD 90.8(16) 9.2(16) MeOK-d^ 100 0 DPC MeOH 52.8(9) 47.2(9) MeOD 43.2(11) 56.8(11) MeOH-d^ 88.7(15) 11.3(15) DMA MeOH 100 0 MeOD 94.5(11) 5.5(11) MeOH-d^ 100 0 a) Products reported normalized to 100%; acrylonitrile was added after samples had annealed for 40 hours to remove any remaining diazocompound. b) Products identified by GC/MS and assayed by GC and are adjusted for response factors, c) Trace amounts of double hydrogen abstraction products were also observed in each case. Ill Another possible mechanism involves yiid formation followed by proton migration as shown in Figure 36. This mechanism is consistent with the experimental results and known tendencies of carbenes to form ylid species. Direct insertion into the OH bond is also possible, but again would be expected to show much larger OH(D) isotope effects. Also, a mechanism in which a singlet-triplet state crossing occurs during reaction can be considered (see Figure 37). It is not clear what sort of isotope effect should be observed in this case. To date it has not been possible to devise a definitive experiment to determine whe ther or not this mechanism actually occurs in these systems. If one considers a model in which slow upconversion from triplet to singlet carbene is the rate determining step in OH insertion fol lowed by rapid reaction from the singlet to give products, as depicted in Figure 38, then the ESR kinetic and product studies mesh very nice ly. In this case the activation energy for the reaction of carbenes in CD^OD becomes the singlet-triplet splitting. In cases of large S-T energy gaps slower kinetics should be observed as well as higher yields of CH insertion products. In cases where small S-T gaps exist the opposite trends should be observed. Based on the ESR kinetic studies in methanol (Table 18), a S-T gap scale can be established as follows: FL > DPC > DBT > DBS > A From the product studies in methanol of Table 17, the following order is obtained: FL.l-NC > DBT,DPC > DBS Energy Figure 37. Surface Crossing Mechanism for Triplet Triplet for Mechanism Crossing Surface 37. Figure 3, Carbene + Alcohol + Carbene Carbene + Alcohol + Carbene Figure 38. Mechanism for Ether Formation Ether for Mechanism 38. Figure 3 ^Carbene ^Carbene Carbene Arylcarbenes with Alcohols. with Arylcarbenes ______by Triplet Carbenes in Alcohols. in Carbenes Triplet by A Reaction Coordinate Coordinate Reaction Ether ► ----- Avoided Crossing Avoided rpe Ether Triplet ige Ether Singlet 112 113 As can been seen the product ordering is consistent with the ordering based on ESR. To test this model a diphenylcarbene derivative was sought in which the S-T gap had been lowered. Since singlet carbenes generally prefer small dihedral bond angles and triplets prefer large ones, a smaller carbene bond angle than that of DPC or DBS was sought. 9-Diazo-10,10-dimethyl-9,10-dihydroanthracene (74) was thus prepared as a precursor to DMA. This system does not suffer from the electronic factors in fluorenylidene that destabilize the singlet. 74 DMA Although exact bond angles of DBS and DMA are not knownexam ination of molecular models suggests that the latter has a smaller bond angle. When ESR and product studies were carried out, DMA proved to be extremely reactive and give almost exclusively singlet products. Evi dence^® indicates that this carbene has a triplet ground state so a decrease in the S-T gap appears to be indicated. If one accepts the validity of this model for explaining kinetic and product results in methanol, then it is possible to gauge S-T energy gaps in carbenes by this method. For typically slow carbenes values of Ea(S-T gap) are obtained which are reasonable based on es timates for that of DPC of 3.0 kcal/mol by Closs.92 However, for more reactive carbenes decay occurs too rapidly for accurate measurements to be made with the existing instrumental set up. So even if the model is valid technical problems may limit its usefulness. 114 To this point no effort has been made to address the problem of what fraction of sites are being observed during ESR kinetic measure ments. based on the multiple matrix site model this fraction should change with photolysis time, carbene reactivity, and temperature. Experiments were therefore undertaken to determine the validity of the matrix site distribution assumption. This was done by preparing identi cal sample tubes each containing the same concentration of diazocora- pound but different isotopically labelled solvents (e.g.- toluene and toluene-dg). In the simplest case one would expect that the carbene signal intensity after a fixed amount of radiation would be larger in toluene-dg than in toluene. This is because more of the nascent car benes will have reacted with the protiated matrix during the photolysis interval than in the deuterated matrix. An unexpected result was obtained when diphenyldiazomethane was irradiated using a 1000 W high pressure Hg-Xe arc lamp with a CuSO^ filter to remove UV and IR radiation. The signal in toluene-dg was slightly smaller than in toluene. On further reflection it became apparent that light absorption due to photochemically produced species could be responsible. Since the two primary candidates (i.e.- diphenyl- carbene and diphenylmethyl radical), are known to absorb light only below 350 nm,^^ a filter of 1,4-diphenyl-1,3-butadiene was in stalled. With the removal of light with A <350 nm accomplished the experiment was repeated. This time, indeed, the signal was significantly larger in toluene- -dg than in toluene. It seems clear that in the initial experiment carbenes generated during photolysis absorbed much of the incident light in the deuterated tube, thus fewer carbenes were formed per unit 115 time, and those formed decayed very slowly. In the protiated tube the carbenes which decayed allowed more incident light to reach diazocom- pound and generate carbenes. In this way similar signals should be observed as a function of time, but more diphenyldiazomethane would be consumed. This latter prediction was not tested. With the 350 nm filter in place, it became possible to perform experiments of this kind on phenyldiazomethane, diphenyldiazomethane, and 5-diazo-I0,ll-dihydro-5H-dibenzo(a,d)cycloheptene (75). 7 5 The results of these experiments are consistent with the multiple matrix site distribution model. As expected the largest isotope effect on signal intensities occurred with phenylcarbene (PC) (see Table 19). This is by far the more reactive of the three carbenes (PC, DPC, and DBS), so decay during photolysis is more significant. These experimen tal results can also be used to determine the maximum fraction of matrix sites which are observed during ESR kinetic measurements in matrices by the method described previously. If one assumes that no carbene decay occurs in the deuterated solvent then the ratio of the signal in the protiated tube to that in the deuterated tube gives the fraction of generated carbenes still remaining in the former tube. Thus, if 10 units of signal are observed in toluene and 100 blocks of signal are observed in toluene-dg then 90 blocks of signal must have decayed from the former tube leaving only 10Z of the matrix sites to be observed. Tables 19 and 20 give the results of some of these experiments. As can be seen only a minute fraction of matrix sites are detected by ESR at 23 K in the case of PC. However, for diarylcarbenes the situation is not quite as severe. In the case of DBS at the rather warm temperature of 106.5 K about 30% of the matrix sites can still be observed. As the temperature is lowered to 97.5 K the ratio improves to about 50%. As the temperature is se verely lowered and decay processes slow down then all matrix sites are observable by the ESR method (see DPC at 50 K data). In principle it would be desirable to perform such experiments at every temperature where kinetic measurements are made during variable temperature stud ies. In this way it could be determined how severely matrix site pro blems influenced the results. Unfortunately, it is not practical to perform such measurements as standard procedure when doing kinetic experiments. Additionally, these measurements are subject to experimen tal errors of unknown magnitude and should not be quantitatively over emphasized. The study of carbene processes by ESR has been hampered by a lack of specific mechanistic information. The interpretation of a triplet carbene decay in alkane matrices as due to abstraction has to this point been based largely on assumption. Recently, though, experiments were performed which give direct evidence that radical reactions do occur under these conditions. Sampled of diazofluorene, 74, and 7j> were irradiated in various matrices at 77 K. In each case biradical spectra consistent with a radical pair was observed. The observed D/hc values are consistent with results of McBride et al.^^ and others**^ who have seen radical pairs in the solid state. The observed splitting are given in Table 21. None of the diazocorapounds used in this study exhibit Table 19. Signal Intensities In Arbitrary Units for Matched Tubes of Phenyldiazomethane in Isotopically Labelled Matrices. Solvent hv time Temp. Signal Signal . s K H Matrix D Matrix toluene 140 77 4b 64 310 4 90 370 4 100 450 4 109 520 4 112 705 4 131 1390 4 163 methylcyclo- 8 23 7 20 hexane 36 21 38 66 32 57 126 53 83 156 57 99 a) 1000 W Hg arc lamp, CuSO^ and 1,4-diphenyl-1,3-butadiene filters in place, b) All measurements are + 2 units. Table 20. Signal Intensities in Arbitrary Units for Hatched Tubes of Some Carbenes as a Function of Temperature in Toluene and Toluene-dR Matrices.3 Carbene hv time Temp. Signal s K toluene toluene-dg DPC 84 50 3b 3 264 11 11 384 15 15 504 19 19 DBS 12 97.5 7 11 24 13 22 36 19 33 48 24 43 12 106.5 8 16 24 14 31 36 19 46 48 24 61 118 a) 1000 W Hg arc lamp, CuSO^ and 1,4-diphenyl-l,3-butadiene filters in place, b) All measurements are + 2 units. Table 21. Observed Maximum Splittings of Radical Generated by Reaction of Carbenes with Matrices. Carbene Solvent Splitting ,(G) DBS Toluene 200 Toluene-dg 275 2M-THF3 270 CCl, 150 3-MP 175 • MCHc 180 FI Toluene. 180 Toluene-dg 240 cci4 none DMA Toluene 140 Toluene-dg 275 a) 2-Methyltetrahydrofuran.. b) 3-Methylpentane.. c) MethylcycloKexana. 120 biradical spectra when photolyzed in PFA, thereby eliminating the possibility that biradical species such as those of Part 1 are respon sible. The results of these experiments mesh nicely with the results of the tunneling calculations given earlier. In every case, the observed splitting in toluene is smaller than that observed in toluene-dg. In terms of a tunneling model this is to be expected. H atom tunneling can occur over a greater distance than D atom tunneling. The resulting ensemble of radical pairs have a larger physical separation in the former case and hence a smaller observed maximum splitting. That these pairs were not completely generated by a photochemical process was demonstrated by observing continued growth in the signal from DBS in all solvents studied long after the irradiating light source was shuttered (see Figures 39-49). However, this increase In the signal through outside experimental error amounted to only about 10% over 30 minutes. This indicates that most radical pairs are generated during photolysis. If they result from thermal reactions of the carbene then they must arise from the most reactive matrix sites. The fraction of sites which give rise to stable radical pairs is unknown, but must be relatively small due to the low signal Intensity of the biradical relative to the carbene observed in these experiments. These findings are consistent with the trajectory approach to matrix site reactivity. Fast matrix sites may react via H atom tunneling along the carbene 2 sp orbital leaving a matrix isolated radical pair with poor geometry for recombination. Slow matrix sites react via poorer angular trajec tories where the generated radical pair is better suited for collapse within the matrix. This is illustrated in Figure 50. DBS /TOLUENE •200 G ------( SEC 1400 200 ■ hy off 120 Figure 39. ESR Spectrum of Radical Pair Derived from DBS in Toluene. 121 d f / t o l u e n e -cj8 240 G — - SEC 6 0 0 « 315 - hv off Figure 40. ESR Spectrum of Radical Pair Derived from DBS in Toluene-dg. 122 DBS j 2-MTHF •270 G SEC 9 9 4 391 1 6 0 100 hV off Figure 41. ESR Spectrum of Radical Pair Derived from DBS in 2-Methyltetrahydrofuran. 123 DBS/cCI, ■150 G x 10 350 S x10 Figure 42, ESR Spectrum of Radical Pair Derived from DBS in Carbon Tetrachloride, 124 D B S / 3 - M F 175 G SEC 187 67 Figure 43t ESR Spectrum of Radical Pair Derived from DBS in 3-Methylpentane. 125 d b s / m c h •180 G SEC 6 7 5 — 4 0 3 136 ESH Spectru* of Rad, , DBS ln il«hyIcyclohexa„ePalr Derived *ro» DF/TOLUENE {------180 G 3 0 0 - tuJofl Figure 45. ESR Spectrum of Radical Pair Derived from FI in Toluene* 127 DBs / t o LUENE-(J8 J -2 75 G 1 SEC 1730 1091 131 20 — h v Off Figure 46. ESR Spectrunj of Radical pair Derived from FI in Toluene-dg» 128 DF /CCI4 ■100 G 1 600s. Figure 47. ESR Spectrum of FI In Carbon Tetrachloride. 129 dma / toluene Figure 48. ESR Spectrum of Radical Pair Derived from DMA in Toluene. 130 DMA/T0LUENE-d8 SEC 600 120 ' hv of I Figure 49. ESR Spectrum of Radical Pair Derived from DMA in Toluene-dg. 131 "Vc<3> HS > R0 C H C_ P 3 « 7 \ H—S ■ £ > - 0 c C D --- > c h r 0 r6 Figure 50. Possible Radical Pair Orientations. The end result is that most of the signal which will be formed is generated by the fast matrix sites during photolysis. This also ac counts for the lack of significant changes in line shape or the ob served splitting on allowing slow sites to decay, since they don't give rise to radical pairs in this model. In the future more experiments need to be devised to determine the validity of this explanation. In conclusion, there is substantial evidence for quantum mechani cal tunneling processes in triplet carbene reactions at low tempera tures. Additionally, by studying the chemical and kinetic isotope effects of matrix deuteration in alcoholic solvents, it is possible to gain qualitative information on the S-T energy gap of arylcarbenes. The results presented herein await the outcome of solution phase exper iments 104b pr0gress for more conclusive interpretation, as well as faster time resolution in the ESR experiments so that the chemistry of the fastest matrix sites can be probed. EXPERIMENTAL General Procedures Melting points were taken with an Electrothermal or Thomas-Hoover capillary melting point apparatus and are uncorrected. NMR spectra were recorded with a Varian EM-360L (60MHz), EM 390L (90 MHz), or Bruker WP-200 (200 MHz), instrument. NMR spectra were recorded on a Bruker WP-80 instrument at 20 MHz. All NMR shifts are reported in units of ppm relative to tetramethylsilane (0 ppm). Infrared spectra were recorded on a Perkin- Elmer Model 457 infrared spectrometer and are given in cm Mass spectra and exact mass measurements were made on a Consolidated Electronics MS-9 double-focusing mass spectrometer using an ionization potential of 70 eV and are reported in units of m/e. UV spectra were recorded on a Cary-14 ultraviolet-visible-infrared spec trometer (Applied Physics). Dry tetrahydrofuran (THF) was prepared by distillation from benzo- phenone ketyl. Solutions of n-butyllithium and methyllithium were titrated at room temperature with 0.1 N standard HC1. 1,2-dibromoethane was added prior to titration of one sample to determine the quantity of inorganic base present. Melting points are reported in °C. Boiling points are reported in °C/torr. 133 134 ESR Measurements ESR measurements were made with a Varian E-112 X-band ESR spectro meter equipped with a modified microwave cavity which admitted light through a series of louvres on one side. Carbene spectra were obtained at 10 mW of microwave power, while during measurements on biradicals <0.1 mW was typically used. All samples were prepared in 4 mm Suprasil quartz sample tubes. Samples were sealed under vacuum after three freeze-thaw cycles to remove traces of oxygen. Samples were stored in liquid nitrogen between experiments. Kinetics were obtained by one of two methods: a) The field was swept through the signal and the time at which the signal maximum was obtained was noted. The signal intensity was obtained by drawing in the baseline in the absence of signal and mea suring the vertical height between the baseline and the signal maximum. b) The field was positioned exactly on the derivative signal maximum in a trial run. With the field width at 0 G a short trace was made to establish the background level. The light was turned on for the allotted time interval and then shuttered. Coincident with shuttering the lamp, the pen sweep was started. Time was determined from a know ledge of the horizontal pen position and the sweep time. • 135 All kinetic measurements were obtained as a minimum of two and typically three trials. ESR measurements at 77 K. (-196 °C), were obtained with the sample immersed in a dewar of liquid nitrogen inside the microwave cavity as shown in Figure 51. ESR measurements at warmer temperatures were obtained using a nitrogen gas flow system depicted in Figure 52. Temperature was mea sured by an Omega Trendicator digital thermometer positioned just below the sample tube. Measurements at lower temperatures were obtained using the helium gas flow system drawn in Figure 53 (the transfer line and glass dewar were supplied by Air Products). Irradiation of ESR samples was accomplished using a Schoeffel 1000 W high pressure Hg-Xe lamp. A water cooled aqueous CuSO^ filter (pyrex, 0.1 M, 2 cm path length), was used to remove ultraviolet and infrared radiation. In filtered light experiments an additional water cooled filter containing 1,4-diphenyl-l,3-butadiene in p-dioxane (pyrex, 5 mg/100 ml, 2 cm path length), was used. Time was measured using timer mechanisms from Precision Scien tific. Additional ESR kinetic data not given in the body of the text is listed in Table 22. The raw data used to construct Figure 18 is given in Table 23. The data can also be used to construct a Curie plot for m-Xylylene as shown in Figure 54. liquid nitrogen dewar ^ Microwave Guide h \> Sample --- ESR Cavity- (not shown) Figure 51. Liquid Nitrogen Dewar Apparatus for ESR Measurements at 77 K. Dewar Sample Microwave (not shown> Guide ESR Cavity Nitrogen Gas Thermocouple Liquid and Heater Wire Nitrogen Figure 52. Nitrogen Cooling Apparatus for ESR Measurements Above 77K. Dewar ---- Microwave Guide Sample (not shown) ESR Cavity Helium Gas Thermocouple and Heater Wire Liquid Helium Figure 53. Helium Cooling System for ESR Measurements Below 77 K. 139 20 u •H 01 d d) 4-110 d 0.1 0.2 1 / T Kt Figure 54. Curie Plot of the Powder ESR Spectrum of DMA. Table 22. ESR Kinetic Data not Given in Text.3 Carbene Solvent hv Time Cone. Temp. k s M °C 10 s DBT ether 3 -196 3.17(13)b -185 9.86(187)b -183 12.1(21)b -181 21.9(16)b -179 50.2(48)b ether-d^Q -186 1.90(21)b -183 4.40(54)b -181 8.48(88)b -179 21.9(20)b -177 48.0(27)b 1-NC ^-PrOH-dg 36 -196 .043(50)° -188 .117(42)° -183 .284(49)° -180 .387(17)° -175 .823(5)° -168 2.14(31)° -196 .018(13)d -188 .077(20)d -183 .308(40)d -180 .331(40)d -175 .806(45)d -168 2.33(21)d a)..CuS0, filter in place, 1000 W Hg-Xe lamp, b) In units of s . c) Monitored at 1600 G. d) Monitored at 1375 G. Table 23. ESR Signal Behavior of m-Xylylene as a Function of Temperature and Microwave Power.3 Temperature Microwave Power (mW) K 0.01 0.05 0.10 0.20 0.50 1.00 25 18.1 30.5 36.4 42.0 35.3 31.2 27 18.3 —— — — — 32 15.5 — —— —— 41 15.0 17.5 37.5 43.3 53.3 54.0 48 14.6 — — — — — 57 14.4 33.0 48.6 73.8 102. 119. 77 11.7 22.5 31.9 43.6 68.8 98.8 a) Measurements made on the ESR spectrum generated from l,3-bis(diazomethyl)benzene/ ethanol after photolysis at 77 K for about 10 minutes. The spectrum was measured one hour later while cooling to eliminate signal growth from chemical causes. 141 142 Solid State Product Studies Product studies were performed as follows: Solutions of diazocompound (0.010 M, 0.300 ml) were placed in 4 mm pyrex tubing which had been thoroughly washed with ammonium hydroxide and dried prior to use. The samples were sealed under vacuum after three freeze-thaw cycles and photolyzed for 90 minutes at 77 K using two Rayonet RPR-3500 lamps (3500 A), in the home-built photochemical reactor shown in Figure 55. After photolysis was complete the samples were rapidly transferred to a storage dewar containing liquid nitrogen where they were annealed 40-50 hours. The samples were then thawed and 0.100 ml acrylonitrile (distilled over Cal^), was added to remove residual diazocompound. The samples were allowed to stand in the dark 2 hours prior to analysis to allow for complete reaction. Solvents used in samples were purchased from Aldrich Chemical (99+%), and were used without further purifi cation. Deuterated solvents were 99.5+ atom % purity. * GC analysis of product mixtures was accomplished using a Hewlett- -Packard 5830A gas chromatograph equipped with a flame ionization detector, and a 12' x 1/8" column (10% SE-30, 80/100 mesh Chromasorb W). Typical run times were 2.5-3.5 hours/injection over a temperature range of 160-230 °C. Product assignments were made on the basis of GC/MS spectra obtained on a Finnegan 4021 mass spectrometer using a 6 1 x 1/8" glass column (10% SE-30, 80/100 mesh Supel). The assignments were confirmed by comparison with the mass spectra of authentic samples described below and by coinjection. Table 24. Product Distribution Under Varied Experimental Conditions for the Reaction of DPC with 2-Propanol.3 h Time # of Lamps^ Decay Time Bond Insertion Products0 hr hr OH 3° CH 2° CH 4 5 18 19.7(4) 58.5(1) 21.8(1) 4 10 18 20.5(1) 58.2(1) 21.3(1) 4 15 18 20.6(1) 58.1(1) 21.3(1) 4 5 0 23.2(1) 59.8(1) 17.0(1) 4 5 1 22.5(1) 61.2(3) 16.3(1) 4 5 4 21.0(1) 61.2(1) 17.8(1) 4 5 38 19.7(1) 56.6(1) 23.7(1) .5 5 18 20.8(1) 57.9(8) 21.3(1) 1 5 18 20.0(1) 59.4(1) 20.6(1) 6 5 18 21.4(1) 56.6(1) 22.0(1) a) 0.1 M solutions of Diphenyldiazomethane in 2-propanol, 4 mm Pyrex tubing, b) Rayonet RPR-3500 lamps, c) Roughly 6% diphenylmethane observed in all samples; normalized to 100%. 144 V/ITH FRJWT fUCeteenOA PLATE. KtMGVCX? Fiaure 55. Apparatus tor Solid State Photolysis — ------Used in Matrix Studies. 145 In the cases of DBT and DMA methanol adducts relative response factors were not determined. The relative response factors of DBS adducts were used for both systems due to synthetic difficulties in obtaining some of the compounds in these series. The results which are given are the average of three separate sample tubes, each of which was injected twice. The influence of photolysis time, light intensity, and annealing time on product distributions were studied for DPC reactions in 2-pro panol matrices. This experiment was carried out as above using a Ray- onet RPR-100 photochemical reactor with RPR-3500 lamps. The results presented in Table 24 indicate that the only possible effect on the product distribution involves the annealing time. OH bond insertion derived ethers were prepared according to the following procedure: The appropriate benzylic alcohol (15 mM, Aldrich) was placed in 30 ml dry THF under N2. Sodium hydride (1.1 eq) was added with stirring. After addition was complete the mixture was allowed to stir overnight. The mixture was cooled to room temperature and 10 ml methyl iodide was added causing formation of a tan precipitate. After stirring for three hours the mixture was poured cautiously into 200 ml water. The aqueous layer was extracted with 3 x 100 ml portions of ether. The ether layers were combined, dried, and evaporated to give the product. Typical yields are 65-85% depending on the compound. 1-Methoxy-l,1-diphenylmethane:bp 94-5/.03 (lit.^^ 150-4/20); nmr (CC14, 90 MHz) 7.0-7.3 (m, 10H), 5.0 (s, 1H), 2.9 (s, 3H); ir (neat film) 1100 (C-0 stretch); ms 198 (M+ , base pk), 183 (loss of methyl), 146 167 (loss of methoxy); exact mass C ^ H ^ O calc. 198.104459, found 198.104918, diff. 0.000459. 9-Methoxyf luorene: mp 43-4,bp 106-8/.03 (lit.118 mp 43.5); nmr (CC14 , 90 MHz) 7.1-7.8 (m, 8H), 5.5 (s, 1H), 2.9 (s, 3H); ir (neat film) 1070 (C-0 stretch); ms 196(M+ ), 195 (loss of H), 181 (loss of methoxy, base pk); exact mass calc. 196.088809, found 196.089388, diff. 0.000579. 5-Methoxy-5H-dibenzo(a,d)cycloheptene:bp 124/.05 (lit.117 oily residue); nmr (CDGL^, 90 MHz) Two conformers A and B were observable in the ratio of 5:1; A 7.0-7.8 (m, 10H), 5.2 (s, 1H), 3.3 (s, 3H); B 7.0-7.8 (m, 10H), 4.8 (s, 1H), 3.4 (s, 3H),(lit.117 7.1 (m, 10H), 3.3 (s, 3H)); ir (neat film) 1080 (C-0 stretch); ms 222 (M+ ), 207 (loss of methyl), 191 (loss of methoxy, base pk); exact mass C^gHj^O calc. 222.104459, found 222.105070, diff. 0.000612. 10, ll-Dihydro-5-methoxy-5H-dibeicizo(a,d)cycloheptene: bp 122-4/.01 (lit.118 142-3/0.35); nmr (CC14 , 90 MHz) 7.0-7.5 (m, 8H), 5.2 (s, 1H), 2.9-3.8 (m, 4H, AA BB ), 3.4 (s, 3H); ir (neat film) 1080 (C-0 stretch); ms 224 (M+ ), 192 (loss of methanol, base pk); exact mass C16H 16° calc* 224.11.98, found 224.1198,diff. 0.0003. l-(Methoxymethyl)naphthalene: l-(Chloromethyl)naphthalene (5.0 g, 28.3 mM, Aldrich) was dissolved in 50 ml dry methanol. Sodium methoxide (45 g) was added and the mixture was stirred overnight. The reaction mixture was poured into water, extracted into CC14 , dried, and evaporated to give 4.9 g of a clear colorless liquid which was dis tilled to give 4.03 g of product as a colorless free-flowing liquid, 83% yield; bp 78-80/.05 (lit.119 133/10); nmr (CC14 , 90 MHz) 7.9-8.1 (m, 1H), 7.6-7.8 (m, 2H), 7.2-7.5 (m, 4H), 5.8 (s, 2H), 3.3 (s, 3H); ir 147 (neat film) 1110 (C-0 stretch); ms 172 (M+ , base pk), 141 (loss of methoxy); exact mass ci2H 12° calc. 172.088809, found 172.089427, diff. 0.000618. 10.11-Dihydro-5H-dibenzo(a,d)cycloheptene-5-carboxylic acid: 10,11-dibenzosuberone (5.0 g, 25.7 mM, Aldrich) was dissolved in 50 ml dry THF. Methyllithium/ether solution (22 ml, 1.5 N, 33.4 mM) was added and the mixture was stirred overnight under N2. The mixture was then poured onto crushed dry ice and extracted into 200 ml 5% aqueous KOH solution. This solution was washed with 3 x 100 ml portions of O ^ C ^ . Acidification of the aqueous layer gave a white solid which was fil tered and dried for two days in a vacuum oven to give 4.2 g of the desired acid as a white powder, 69% yield: mp 235-6 (lit.^® 237-8); nmr (acetone-dg, 90 MHz) 6.9-7.3 (m, 8H), 4.8 (s, 1H), 4.3 (s, 1H, exchanges with D2O), 3.2-3.6, 2.6-3.0 (m, 4H, AA BB ). 10.11-Dihydro-5H-dibenzo(a,d)cycloheptene-5-methanol: The acid prepared above (2.0 g, 8.4 mM) was dissolved in 50 ml dry THF. LiAlH^ (0.7 g) was added slowly with stirring under N2. The mixture was heated to reflux and stirred overnight. The reaction wafe cooled and ethanol was added to quench excess LiAlH^. The mixture was poured into 5% aqueous KOH (150 ml) and extracted with 3 x 50 ml portions of ether. The ether layers were combined, dried, and evaporated to give 2.3 g of an oily residue. Kugelrohr distillation at 125/0.3 gave 1.1 g of pro duct as a white solid, 58% yield: mp 65-6 (lit.*^* 56-9); nmr (CDCI3, 60 MHz) 7.0-7.3 (m, 8H), 4.1 (t, 2H, 7 Hz), 2.5-3.6 (m, 5H, symmetric), 1.6 (s, 1H, exchanges); ms 224 (M+ ), 206 (loss of water), 193 (loss of hydroxymethyl.base pk); exact mass CjgH^gO calc. 224.1202, found 224.1230, diff. 0.0028. 148 2-Propyl diphenylmethyl ether: Sodium (2.2 g, 97 mM) was placed in 55 ml dry 2-propanol and stirred under dry ^ until dissolved. Brorao- diphenylmethane (10 g, 40.5 mM, Aldrich) was added to the mixture at reflux. The mixture was refluxed overnight and quenched with water. The mixture was poured into 100 ml water and extracted with 3 x 100 ml portions of ether. The ether layers were combined, dried, and filtered to give a yellow oil from which the desired product was obtained by vacuum distillation, 5.96 g, 66% yield: bp 80-2/.03 (lit.^^ 155-8/14); nmr (CDCl-j, 90 MHz) 7.1-7.5 (m, 10H), 5.5 (s, 1H), 3.7 (septet, 1H, 6 Hz), 1.2 (d, 6H, 6Hz); ir (neat film) 1115 (C-0 stretch), 1385 (isopro pyl symmetric bend). Diphenylmethyllithium: D r y ether 100 ml was placed in a 500 ml flesk under N2. Diphenylmethane (27.0 g, Aldrich) was added followed by slow addition of 87 ml of 1.65 N n-butyllithium. The mixture was re fluxed with stirring for 20 hours resulting in a dark red solution which was transferred to a dry storage bottle. Titration indicated that the concentration of diphenylmethyllithium was 0.46 N. 1,l-Diphenyl-2-methyl-2-propanol: The solution prepared above (38 ml) was placed in a dry 100 ml flask under Ar at -78 °C. Acetone (3.0 ml) was added dropwise resulting in a total discharge of color. After 5 minutes the mixture had formed a white precipitate making stirring impossible. The mixture was warmed to room temperature and 50 ml water was added. The ether layer was separated and the aqueous layer was extracted with 3 x 50 ml portions of ether. The ether layers were combined, dried, and evaporated to give 5.8 g of a thick liquid which was vacuum distilled to give 1.76 g of the desired alcohol as a clear colorless viscous liquid, 44% yield: bp 110-114/.02 (lit*^ 151- 149 2/ 5.5) ;nmr (CDCI3, 90 MHz) 7.0-7.7 (in, 10H), 3.8 (s, 1H), 1.8 (br s, lli, exchanges with D2O), 1.2 (s, 6H); ir (neat film) 3580, 3460 (OH stretch). 4,4-Diphenyl-2-butanol: The diphenylmethyllithium solution pre pared above (38 ml) was cooled to -78 °C in a dry 100 ml round bottomed flask under Ar with stirring. Propylene oxide (3.0 ml, Aldrich) was added and the mixture was warmed slowly to room temperature with stir ring was continued for 20 minutes. After 20 minutes 3 ml water was added and stirring was continued for 20 minutes. The mixture was washed twice with water, d^.ed. and evaporated to give 5.8 g of an opaque white liquid rhicli was vacuum distilled to give the product as a clear colorless li1Uid, 2.6 g, 65% yield: bp 118-9/.04 (lit.125 107-9/0.1); nmr (CDCI3, 90 MHz) 7.1-7.4 (br s, 10H), 4.2 (t, 1H, 8Hz), 3.6 (sextet, 1H, 7Hz), 2.1-2.3 (m, 2H), 1.9 (s, 1H, exchanges with D2O), 1.1 (d, 3H, 7Hz); ir (neat film) 3560, 3360 (OH stretch). 5H-Dibenzo(a,d)cycloheptene: was obtained as a side product in the preparation of 5H-dibenzo(a,d)cyclohepten-5-one hydrazone: mp 128.5 (lit.126 130-1); nmr (CC14 , 90 MHz) 7.2 (s, 8H), 6.9 (s, 2H), 3.7 (s, 2H). 10,ll-Dihydro-5H-dibenzo(a,d)cycloheptene, 2,2-diphenylethanol, 9-fluorenemethanol, 2-(l-naphthyl)ethanol, 1-methylnaphthalene, di- phenylmethane, and fluorene were obtained from Aldrich Chemical and were used without further purification. 150 Additional Compounds Included in Text: 1.3-Benzenedicarboxaldehyde bis tosylhydrazone: 1,3-benzenedi- carboxaldehyde (2.6 g, 19.4 mM, Aldrich) was dissolved in a minimum amount of boiling ethanol. In a separate flask 8.0 g tosylhydrazide (42 mM) was dissolved in a minimum amount of boiling ethanol. The two solutions were combined resulting in the immediate formation of a white precipitate. After cooling, the mixture was filtered to give 9.0 g of the desired product as a white solid, 98% yield: mp 170d; nmr (DMSO-dg, 60 MHz) 7.2-8.0 (m, 12H), 3.5 (s, 2H), 2.3 (s, 6H), 11.6 (s, 2H). 1.4-Benzenedicarboxaldehyde bis tosylhydrazone: was prepared according to the above procedure as a white solid, 98% yield: mp 230d. 1.3-Bis(diazomethyl)benzene (21):The 1,3 bis tosylhydrazone (0.5 g) prepared above was dissolved in 10 ml 1,4-dioxane. Aqueous KOH (15 ml, 40%) was added and the mixture was heated to reflux with stir ring, resulting in the formation of a dark red, clear two-phase solu tion. The mixture was poured into 10% aqueous ^ 200^ solution and extracted with ether. The ether layer was washed six times with 10% aqueous Na200g and evaporated to give 145 mg of the bis diazocompound as a dark red odorous liquid which solidifies in the freezer, but melts below room temperature, 85% yield; nmr (CCl^, 90 MHz) 6.8-7.2, 6.1-6.5 (m, 4H), 4.7 (s, 2H); ir (neat film) 2060 (diazo N=N stretch). 28 1.4-Bis(diazomethyl)benzene (24): 1 was prepared by the above method in 70% yield as a dark red odorous liquid; ir (neat film) 2060 (diazo N=N stretch), (lit.^® 2100). 1,3-Bis(iodomethyl)benzene (30): was prepared by the method of Finkelstein 129 from l,3-bis(bromomethyl)benzene: pale yellow needles; 151 mp 104.5-5 (lit.129 106); nmr (CDCI3, 60 MHz) 7.2-7.5 (m, 4H), 4.4 (s, 4H). Photolysis of this material (0.1 M in benzene) in the cavity of the ESR gave only a strong free radical signal after 4 hours. 1.3-Benzenediaeetic acid di-t-butyl ester (31): According to the 131 general method of Bartlett 3.3 g ofjn-phenylenediacetic acid (17.3 mM, Aldrich) was added slowly to 4 g thionyl chloride with constant stirring at room temperature. After stirring for 18 hours most of the thionyl chloride was distilled off. Remaining SOCI2 was removed under vacuum to give 4.0 g of an orange-brown oil, 100% crude yield; ir (neat film) 1780 (00C1 carbonyl stretch). This material was dissolved in 65 ml dry benzene and 35 ml cyclo- hexane. The mixture was cooled in an ice bath to 0 °C. t-Butyl hydro peroxide (6.9 g, 35 mM, Aldrich) was added dropwise with stirring followed by careful addition of 3.0 ml pyridine. During this process the temperature was maintained below 8 °C. The mixture was stirred 30 minutes, then extracted with 2 x 100 ml portions of 10% ^SO^, 10% NaOH, and 10% ^ 2003. The organic layer was dried, evaporated, and filtered through 60 g of Florisil. The residue was placed under high vacuum for 5 hours to remove traces of solvent. A yellow oil resulted, 0.70 g, 13% overall crude yield; nmr (CDCI3, 90 MHz^ 7.1-7.3 (m, 4H), 3.7 (s, 4H), 1.3 (s, 18H); ir (neat film) 1780 (perester carbonyl stretch), 1100 (C-0 stretch). 1.3-Bis(bromomethyl)benzene (35): 1,3-bis(hydroxymethy1)benzene (20.0 g, 0.15 mol, Aldrich) was placed in 450 ml CH2CI2 in a one liter flask. PBr3 (19.0 ml, 0.20 mol) was added with mechanical stirring over 0.5 hour. The mixture was stirred for 1.5 hours at room tempera ture, then 30 ml water was added and stirring was continued for 15 152 minutes. The reaction mixture was separated in a separatory funnel and the organic layer was filtered through a 1" column containing 40 g of silica gel. The eluent was evaporated to give 31.3 g (0.12 mol) of a white lachrymatory solid, 81% yield: mp 73-4 (lit.13*3 70-2); nmr (CDCI3, 60 MHz) 7.0-7.3 (m, 4H), 4.3 (s, 4H). 1,1,2,2-Tetraphenylethane-l,2-diol sulfite (27): was prepared by 1 ^9 the method of Griffin et al. from benzpinacole (Aldrich) in 32% yield; mp 134-5 (lit.133 137-8); ir (CCl^) 1230, 1445 (sulfite stretching modes); ms parent ion at 412 not detected, but peak at 348 corresponds to loss of SO2. 1,3-Diazidobenzene: was prepared by the method of Forster et 133 al. The product was purified by chromatography on 20 g of silica gel using petroleum ether (low boiling) as eluent. The diazide is a free- -flowing yellow liquid (lit.133 mp 5); nmr (CDCI3, MHz^ 7.2-7.5 (m, 1H), 6.6-7.0 (m, 3H); ir (neat film) 2120 (azide stretch). Diphenylketene (29): was prepared according to the method of Smith and Hoehn134 in 30% yield: bp 85/.03 (lit.134 119-21/3.5); ir (neat film) 2120 (ketene carbonyl stretch); ms 194 (M+ ), 166 (loss of CD). 1,3,5-tris(1-naphthyl)-2,4,6-triazabicyclo(3.1.0)hexane: According 135 to the method of Schmitz and Ohme, 125 ml of 10 N methanolic ammonia was cooled to -40 °C in a 250 ml three necked flask equipped with a dry ice condenser, addition funnel, and magnetic stir bar. A mixture of 12.6 ml of £-butyl hypochlorite (0.11 mol) in 12.5 ml J^-butanol was added dropwise with stirring so as not to cause warming above -20 °C. After addition was complete the solution was warmed to -20 °C and 30.6 g of 1-naphthaldehyde (0.20 mol) was added rapidly. The mixture was stirred for 1 hour at -20 °C, and then was allowed to warm to room 153 temperature. After 2 hours a white precipitate had formed making stirring difficult. The mixture was allowed to stand an additional 8 hours and was suction filtered to give a sticky yellow solid. The solid was suspended in 10 ml ether and stirred vigorously for 30 minutes. Suction filtration gave 10.5 g (21.8 mM) of an off-white solid, crude yield 33%, mp 198d. The product was insoluble in water, ether, chloroform, DMSO, acetone, and benzene. 3H-3-(l-naphthyl)diazirine: The bicyclic triazine prepared above (6.0 g, 12.5 mM) was placed in 50 ml of anhydrous methanol under Ar at 0 °C. 2.5 ml t-butylhypochlorite (20.7 mM) was added dropwise and the mixture was stirred for 1.5 hour in the dark. The solution was then i poured into a solution of 50 g sodium metabisulfite in 400 ml water and stirred 30 minutes. This solution was extracted with 3 x 30 ml portions of Q ^ C ^ . The extracts were combined and dried over sodium sulfate. The solution was concentrated and chromatographed on 30 g of silica gel using petroleum ether (low boiling). The diazirine was monitored by long wave UV light and appears as a dark band on the column. 280 mg of a light yellow solid was isolated, 13% yield: mp 65d; nmr (CDClg, 200 MHz) 8.26-8.36 (m, 2H), 7.59-7.67 (m, 3H), 7.41 (d, 1H, 8.0 Hz), 6.42 (d, 1H, 8.0 Hz),2.58 (s, 1H, proton on diazirine ring); irradiation at 6.42 causes the doublet at 7.41 to collapse to a singlet with no other visible changes; ir (CCl^) 1620 medium (diazirine N=N stretch); UV X m ! (methanol) 360 nm, (hexane) 358 nm; ms base peak of 139 cor- mdX IDaA responds to C^Hy, high weight peaks present, decomposes at source temperature of 70 °C; irradiation at 77 K in ethanol gave an ESR spectrum of 1-naphthylcarbene. 10,10-Dimethylanthrone: was prepared by the method of Davis et 1 al. from 2-benzylbenzoic acid (Aldrich) in 50% overall yield: mp 102.5-104 (lit.136 103-4); nmr (CDCI3, 90 MHz) 8.4 (d, 2H, 8 Hz), 7.2-7.8 (m, 6H), 1.7 (s, 6H); ms 222 (M+ ), 207 (loss of methyl, base pk); exact mass C16H 140 calc. 222.1045, found 222.1045, diff. 0.0000. Benzophenone hydrazone and fluorenone hydrazone were purchased from Aldrich Chemical. Other hydrazones were prepared according to the following general procedure: Ketone (10 g) is placed in 20 ml anhydrous hydrazine and 10 g hydrazine sulfate. The stirred mixture is heated to reflux and moni tored by tic (20% ethyl acetate/hexane) until complete. The mixture is then poured onto 150 g of crushed ice and extracted with 3 x 100 ml portions of ether. The ether layers are combined, washed with water, dried, and evaporated to give the hydrazone. Purification was achieved by Kugelrohr sublimation. Yields range from 70-90%. 5H-Dibenzo(a,d)cyclohepten-5-one hydrazone: yellow glassy resin; nmr (CC14 , 90 MHz) 7.1-7.7 (m, 8H), 6.8 (s, 2H), 5.2-5.8 (br s, 2H, exchanges with D20 ; ir (neat film) 3380, 3280, 3200 (hydrazone NH2 stretching modes), 1620 (C=N stretching). 10,11-Dihydro-5H-dibenzo(a,d)cyclohepten-5-one hydrazone: yellow oil; nmr (CC14 , 90 MHz) 6.9-7.8 (m, 8H), 5.3-5.7 (br s, 2H, exchanges with D20), 2.7-3.3 (br s, 4H, does not exchange with D20); ir (neat film) 3400, 3280, 3200 (NH2 stretching modes), 1635 (C=N stretch). 10,10-Dimethylanthrone hydrazone: yellow oil; nmr (CC14 , 90 MHz) 7.9-8.1 (m, 1H), 7.0-7.8 (m, 7H), 5.3-5.7 (br s, 2H, exchanges with 155 D2O), 1.5 (s, 6H); ir (neat film) 3470, 3260, 3200 (hydrazone Nl^ stretching modes), 1615 (C=N stretch). Diphenyldiazomethane, diazofluorene, 1-naphthyldiazomethane, 5-diazo-5H-dibenzo(a,d)cycloheptene, and 10,1l-dihydro-5-diazo-5H- -dibenzo(a,d)cycloheptene were prepared from the corresponding hydra- zones by mercuric oxide oxidation as described below: Hydrazone (1.0 g) was dissolved in 200 ml ether and 20 g yelloW mercuric oxide (Baker) was added. Three drops 40% aqueous KOH were added and the mixture was stirred until tic showed complete consumption of the starting hydrazone. The mixture was filtered and evaporated, and the residue was dissolved in petroleum ether (low boiling). The solu tion was dried and evaporated to give usable diazocompound. Typical yields by this method range from 50-80%. Samples of higher purity were obtained by either sublimation or recrystallization from methanol. Diphenyldiazomethane (25): purple solid: mp 30 (lit.*^7 30); nmr (CDClg, 90 MHz) 7.2-7.6 (m); ir (neat film) 2030 (diazo N=N stretch). Diazof luorene (53) :red needles from methanol: mp 99d (lit.*'*'7 99d). 9-Diazo-10,10-dimethyl-9,10-dihydroanthracene (74): purple prisms from methanol: nmr (CCl^, 90 MHz) 6.8-7.5 (m, 8H), 1.7 (s, 6H); ir (CC14) 2060 (diazo N=N stretch). Carbene (DMA) D'/hc = 0.4253 cm-1, E V h c = 0.02396 cm- *. A Curie plot of signal intensity vs. 1/T showed this to be a ground state triplet, however some deviation from linear ity was observed indicating the possible presence of a very low lying singlet state. 156 1-Naphthyldiazomethane:888 recj needles from methanol: mp 4Id. 10,ll-Dihydro-5-diazo-5H-dibenzo(a,d)cycloheptene (75): purple plates from methanol: mp 78.5 d (lit.138 71d); nmr (CCl^, 90 MHz) 6.9-7.3 (m, 8H), 3.0 (s, 4H, sharp); ir (CCl^) 2020 (diazo N=N stretch). 5-Diazo-5H-dibenzo(a,d)cycloheptene: dark red plates: mp 62.5d (lit.138 62d); nmr (CC14, 90 MHz) 6.9-7.4 (m, 8H), 6.4 (s, 2H); ir (CCl^) 2025 (diazo N=N stretch). 9-Diazoanthrone: was prepared by the procedure of RegitzA1 3 from anthrone (Aldrich) in 69% yield; mp 280d (lit.13^ 280d). 9-Diazomethylanthracene: was obtained from Dr. V. P. Senthil- nathan. 4-(Diazomethyl)biphenyl: was obtained from Mr. E. C. Palik. Phenyldiazomethane (51): was prepared from benzaldehyde tosyl hydrazone (Aldrich) as a red liquid according to the method described for 1,3-bis(diazomethyl)benzene; ir (neat film) 2030 (diazo N=N stretch). 2,6-Pyridine dicarboxaldehyde: 2,6-Pyridinedimethanol (3.5 g, 25.9 mM, Aldrich) was dissolved in a minimum amount of DMSO. In a 250 ml flask were placed 100 ml CH2CI2 and 5 ml oxalyl chloride. This was cooled to -78 °C, and 8.5 ml dry DMSO was added with evolution of gas. After 5 minutes the diol solution was added rapidly with stirring. After 10 more minutes 35 ml triethylamine was added, and the mixture was stirred at -78 °C for 45 minutes, then warmed to room temperature. Addition of water followed by treatment with 10% aqueous NaOH until basic followed by extraction with ether and evaporation gave a sample of crude solid material. This material was purified by column chromato- 157 graphy on silica gel (25 g) using ether as the eluent to give the product as a white solid, 1.85 g, 51% yield: mp 122-4 sublimes (lit.^® 124); nmr (CDC13, 90 MHz) 10.2 (s, 2H), 8.1-8.4 (m, 3H). Attempted synthesis of 2,6-bis(diazomethyl)pyridine: 2,6-Pyr- idinedialdehyde (2.0 g, 15.0 mM) prepared above was dissolved in a minimum amount of boiling ethanol. In another flask 6.2 g tosylhydra- zide (33.0 mM, Aldrich) was dissolved in a minimum amount of boiling ethanol. The two solutions were poured together and allowed to cool. No precipitate was observed. On evaporation a sticky brown foam (8.1 g) was recovered. Portions of this material were heated with N,N,N/ ,N/ -tetramethylguanidine resulting in the rapid formation of a dark red solution. However, on treatment with 10% aqueous NaHCO^ all the color extracted into the aqueous layer and could not be made to extract back into any organic solvents which were immiscible with water. APPENDIX I The phenomenon of tunneling is a direct consequence of the wave particle duality of matter. The terra tunneling as used here refers to the passage of a particle with kinetic energy E through a potential barrier of energy VQ > E (see Figure 56). Classically, a particle of kinetic energy E will pass over a potential barrier VQ only for E > VQ ; at lower values of E reflection occurs. However, in the case of very small particles such as atoms, matter is not particulate in the class ical sence and is instead wavelike in nature. The wavelength X of a particle of mass m is given by the DeBroglie formula. When m is large, classical mechanics A = h/mv serves to describe the behavior of a particle, but for values of m corresponding to atoms a quantum mechanical treatment is required. It is a fundamental tenet of quantum theory that a mathematical expression exists which will accurately describe the motion of a par ticle or particles in a potential field. This expression V is an eigen function of the Schrodinger Equation It V7 = EV Further, the rules of quantum theory require that all solutions be continuous and single valued over all space. This requirement leads to the phenomenon of tunneling. 158 159 III V(x) x Figure 56. Wave Functions for a Particle Tunneling Through a Potential Barrier. 160 Solutions to the Schrodinger Equation are well known in many cases. 9"* In one dimension exact solutions are known for parabolic, rectangular, and Eckart potential fields to name a few. Figure 55 shows a generalized one-dimensional potential field. In regions I and II a sinusoidal wavefunction^ is obtained. The requirement that V be con tinuous over space (in this case all x), necessitates a wavefunction inside the barrier (region II), which continuously links the wavefunc- tions of regions I and III. This wavefunction turns out to be an expo nentially decaying curve as seen in Figure 56. The amplitude of this wavefunction is very sensitive to barrier height, shape, and especially width.98 The most widely used barrier for experimental applications is the one-dimensional Eckart barrier (described in main body of text). When this potential is inserted into the Schrodinger Equation, exact solu tions can be obtained. For an asymmetric barrier such as that shown in Figure 56 the permeability G (the chance of traversing the barrier on collision), is given by sinh2 ( irkb( 1 +A)) - sinh2 (tf kb( 1 -A)) G-(e ) = p------5------5--- p------t - sinh (irkb(l +A)) + cosh (£'ff(8rab A/h - 1)S where, k = (2mE) ^ 2/h , = ((E-B)/E) ^ 2 and A, B, and b are barrier parameters as given in the text. From a knowledge of the permeability, the tunneling correction to the classical rate constant, Jf1, can be calculated and is given by,98 161 exp(V /kT) ‘ 00 P = ------2---- I G(E)exp(-E/kT) dE kT Q The integrand is composed of two opposingfunctions; G(E) increas es with increasing E while the exponential term decreases. The product represents a distribution of transmitted particles as a function of energy. Bell^ has classified tunneling effects as follows: 1 1.1 < X1 < 4 small to moderate tunneling 4 < r significant tunneling Ordinarily, only in the latter two cases will tunneling be experi mentally observed. This will appear as curvature in the Arrhenius plot, low Arrhenius parameters, or an abnormally high reaction velocity. When these criteria are met one may conclude that tunneling is indeed in volved . REFERENCES AND NOTES 1) Schlenk, W. ; Brauns, M. Ber. (1915), 4J5, 661. 2) Thiele, J.; Balhorn, H. ibid (1904), 37, 1463. 3) Chichibabin, A. E. ibid (1907), 40, 1810. 4) For examples see: a) Turro, N. J. Tetrahedron (1982), 2^* 809. b) Kelley, D. F.; Mazur, M. R.; Rentzepis, P. M.; Berson, J. A. J. Am. Chem. Soc. (1982), 104, 3764. c) Kelley, D. F.; Rentzepis, P. M. ; Mazur, M. R.; Berson, J. A. ibid (1982), 104, 6167. 5) Buchwalter, S. L.; Closs, G. L. ibid (1975), 97_f 3857. 6) a) Dowd, P. ibid (1966), 88, 2587. b) Baseman, R. J.; Pratt, D. W.; Chow, M.; Dowd, P. ibid (19.76), 98, 5726. 7) Dowd, P. ibid (1970), 92, 1066. 8) Pagni, R.; Burnett, M. N.; Dodd, J. R. ibid (1977), 99, 1972. 9) Rule, M.; Matlin, A. R.; Seeger, D. E., Hilinski, E. F.; Dougherty, U. A.; Berson, J. A. Tetrahedron (1982), 3*5, 3280. 10) Flynn, C. R.; Michl, J. J. Am. Chem. Soc. (1974), 9jj, 3280. 11) a) Flynn, C. R.; Michl, J. ibid (1973), 95, 5802. b) Williams, D. J.; Pearson, J. M.; Levy, M. ibid (1970), 92, 1436. 12) Muller, E.; Tietz, E. Ber. (1941), 74, 807. 13) Errede, L. A.; Swarc, M. Quart. Revs. (London) (1958), lji, 301. 14) a) Errede, L. A.; Landrum, B. F. J. Am. Chem. Soc. (1957), 79, 4952. b) Errede, L. A.; Hopwood, S. L., Jr. ibid (1957), 79, 6507. c) Errede, L. A.; Hoyt, J. M. ibid (1960), 82, 436. d) Oppolzer, W. Synthesis (1978), 11, 793. e) Oppolzer, W. Angew. Chem. internat. Ed. Engl. (197777 lji, 10. 15) Oppolzer, W.; Keller, K. J. An. Chem. Soc. (1971), 93, 3836. 16) a) Bauld, N. L.; Farr, F. R.; Chang, C. Tetrahedron Lett. (1972), 24, 2443. b) Farr, F. R.; Bauld, N. L. J. An. Chem. Soc. (1970), j)2, 6695. c) Reference 15. 162 163 17) Miller, R. D.; Kok, J.; Michl, J. ibid (1976), j))3, 8510. 18) Dolbier, W. R., Jr.; Matsue, K.; Dewey, H. J.; Horak, D. V.; Michl, J. ibid (1979), 101, 2136. 19) Weiss, D. S. ibid (1975), 97, 2550. 20) Pullman, A.; Pullman, B.; Bergmann, E. D.; Berthier, G.; Fischer, E.; Ginsberg, D.; Hirschberg, Y. Bull. Chim. Soc. France (1951), 707. 21) _1 is a ground state triplet; see: a) Luckhurst, G. R.; Pedulli, G. F. J. Chem. Soc. (1971), 39^, 329. b) Kothe, G.; Summerman, W. Angew. Chem. internat. Ed. Engl. (1970), 9_, 906. 22) a) Migirdycin, E.; Baudet, J. J. Am. Chem. Soc. (1975), 97, 7400. b) Migirdycin, E. Hebd. Seances Acad. Sci. (1968), 266, 756. 23) a) Boekelheide, V. Acc. Chem. Res., (1980), _L3, 65. b) Pellegrin, M. Rec. Trav. Chim. Pays-Bas (1899), ljJ, 458. 24) Gajewski, J. J.; Chang, M. J.; Stang, P. J.; Fisk, T. E. J. An. Chem. Soc. (1980), H)2, 2096. 25) Tseng, K. L ,; Michl, J. ibid (1977), 99, 4842. 26) a) Farmer, J. B.; Marsden, D. G. H.; Lossing, F. P. J. Chem. Phys. (1955), 23, 403. b) Farmer, J. B.; Lossing, F. P.; Marsden, D. G. H.; McDowell, C. A. ibid (1956), 24, 52. 27) Pollack, S. K.; Raine, B. C.; Hehre, W. J. J. An. Chem. Soc. (1981), 103, 6308. 28) Kato, S.; Morokuma, K.; Feller, D.; Davidson, E. R.; Borden, W. T. ibid (1983), 105, 1791. 29) Platz, M. S. in "Diradicals", Borden, W. T., Ed.; Wiley-Interscience, New York, 1982. 30) Dewar, M. J. S.; Dougherty, R. C. "The PM0 Theory of Organic Chemistry", Plenum Press, New York, 1975. 31) Rule, M.; Matlin, A, R.; Dougherty, D. A.; Hilinski, E.; Berson, J. A. JT^ Aru_ Chem. Soc. (1979), 101, 5098. 32) Seeger, D. E.; Hilinski, E. F.; Berson, J. A. ibid (1982), 104, 4954. 33) a) Matlin, A. R.; Inglin, T. A.; Berson, J. A. ibid (1982), 104, 4954. b) Reference 9. 34) Borden, W. T. in "Diradicals", Borden, W. T., Ed.; 164 Wiley-Interscience, New York, 1982. 35) Hund, F. "Linienspektren und periodische Systerae der Elemente", Springer-Verlag, Berlin, 1927. 36) Hoffman, R. J. An. Chem. Soc. (1968), 90, 1475. 37) Platz, M. S.; Carrol, G.; Pierrat, F.; Zayas, J.; Aister, S. Tetrahedron (1982), _38, 777. 38) Borden, W. T.; Davidson, E. R. J. An. Chem. Soc. (1977), 99, 4587. 39) Ovchinnikov, A. A. Theor. Chim. Acta (1978), 40) Longuet-Higgins, H. C. J. Chem. Phys. (1950), JJ5, 265, 275, 283. 41) Drago, R. S. " Physical Methods in Chemistry", Saunders, Philadelphia, 1977. 42) Zupancic, J. J.; Schuster, G. B. J. Am. Chem. Soc. (1980), 102, 5958. 43) Private communication from Prof.. J. A. Berson. 44) Wasserman, E.; Snyder, L. C.; Yager, W. A. J. Chem. Phys (1964), 41, 1763. 45) Breslow, R.; Chang, H. W.; Hill, H. R.; Wasserman, E. J. Am. Chem. Soc. (1967), 89, 1112. 46) a) Senthilnathan, V. P.; Platz, M. S. ibid (1981), 102, 7637. b) Senthilnathan, V. P.; Platz, M. S. ibid (1981), 103, 5503. 47) a) Wasserman, E.; Murray, R. W.; Yager, W. A.; Trozzolo, A. M.; Smolinsky, G. ibid (1967), 8£, 507. See also b) Murahashi, S.-I.; Yoshimura, Y.; Yamamoto, Y.; Moritani I. Tetrahedron (1972), 28, 1485. c) Itoh, K. Chem. Phys. Lett. (1967), 235. d) Trozzolo, A. M.; Murray, R. W.; Smolinsky, G.; Yager, W. A., Wasserman, E. J. An. Chem. Soc. (1963), 85, 2526. 48) The reactivity of some arylcarbenes with methanol at 77 K as measured by product and kinetic analysis is presented in Part II of this work. 49) Extensive experimentation within our research group has shown that arylcarbenes do not react with this solvent at 77 K. 50) This phenomenon was first observed by Thurnauer, M. Ph.D. Thesis, University of Chicago. 51) Smith, R. L.; Manmade, A.; Griffin, G. W.; J. Heterocyc. Chem. (1969), 6, 443. 165 52) I wish to thank Prof. G. W. Griffin for supplying a sample of 28. 53) I wish to thank Prof. V. Boekelheide for supplying a sample of 34. 54) Ryl'kov, V. V.; Kholmogorov, K. E. Khim. Vys. Energ. (1970), 4_, 119. 55) The photoisomerization of diazirines to diazoalkanes is well known; for example, see Chedekel, M. R.; Skoglund, M.; Kreeger, R. L.; Shecter, H. J. An. Chem. Soc. (1976), 98, 7846. 56) Parts of the following contain unpublished research performed by Mr. J. Zayas in partial fulfillment of his Ph.D. requirement, Ohio State University. 57) Tajiri, A.; Nakajima, T. Tetrahedron (1971), 27_, 6089. 58) a) Durr, H.; Albert, K.-H.; Kausch, M. ibid (1979), 35, 1285. b) Jones, M., Jr. J. Org. Chem. (1968), J3, 2538. c) Cram, D. J.; Van Duuren, B. L. J. Am. Chem. Soc. (1955), JJ_, 3576. d) Schonleber, D. Angew. Chem. internat. Ed. Engl. (1969), 8^, 76. e) Jones, M., Jr. ibid (1964), 8^, 76. 59) Reference 57. These calculations were done at the INDO/2 level of theory and are of questionable validity. 60) Hine, J. J. An. Chem. Soc. (1950), 72, 2438. 61) For example, see Kirmse, W. "Carbene Chemistry", Academic Press, New York, 1964. 62) Moss, R. A.; Dolling, U.-H. J. Am. Chem. Soc. (1971), 93, 954. 63) Moss, R. A.; Joyce, M. A. ibid (1977), 99, 1262. 64) Moss, R. A.; Joyce, M. A. ibid (1978), 100, 4475. 65) "Carbenes", Vols. I and II, Moss, R. A. and Jones, M., Jr., Eds., John Wiley and Sons, New York, 1975. See also reference 63. 66) Moss, R. A.; Uuselton, J. K. J. An. Chem. Soc. (1978), 100, 1314. 67) Moss, R. A.; Wetter, W. P. Tetrahedron Lett. (1981), 22,997. 68) Tomioka, H.; Ozaki, Y.; Koyabu, Y.; Izawa, Y. ibid(1982), 23, 1917. 69) Tomioka, U.; Suzuki, S.; Izawa, Y. Chem. Lett. (1980), 293. 166 70) Cohen, M. D.; Schmidt, G. M. J. J. Chem. Soc. (1964), 1996. See also Cohen, M. D. Angew. Chem. internat. Ed. Engl. (1975), 14, 386. 71) Paul, 1. C.; Curtin, D. Y. Acc. Chem. Res. ( 1973), 6^, 217. 72) Tomioka, H.; Izawa, Y. J. Am. Chem. Soc. (1977), 9 9 6128. 73) See references 61-69, 72, 75-85. 74) For an elaboration of this type of argument see references 62 and 63 and references contained therein. 75) Tomioka, H.; Suzuki, S.; Izawa, Y. Bull. Chem. Soc. Jpn. (1982), _55, 492. 76) Tomioka, H.; Miwa, T.; Suzuki, S.; Izawa, Y. ibid (1980), 53, 753. 77) Tomioka, H.; Griffin, G. W.; Nishiyama, K. J. Am. Chem. Soc. (1979), 101, 6009. 78) Kristinsson, H.; Griffin, G. W. ibid (1966), 8tJ, 1579. 79) Tomioka, H. ibid (1979), 101, 256. 80) a) References 61, 65. See also b) Carr, R. W., Jr. J . Phys. Chem. (1966), 70, 1971. c) Herzog, B. M.; Carr, R. W., Jr. ibid (1967), _71, 2688. d) Halberstadt, M. L.; Crump, J. J. Photochem. (1972), _1_, 295. 81) Tomioka, H.; Ozaki, Y.; Izawa, Y. Chem. Lett. (1982), 843. 82) Turro, N. J.; Butcher, J. A., Jr.; Moss, R. A.; Gouo, W.; Munjal, R. C.; Fedorynski, M. J. Am. Chem. Soc (1980), 102, 7576. 83) Tomioka, H.; Suzuki, S.; Izawa, Y. ibid (1982), 104, 3156. 84) a) Tomioka, H.; Itoh, M.; Yamakawa, S.; Izawa, Y. J. C. S. Perkin II (1980), 603. b) Tomioka, H.; Okuno, H.; Izawa Y. ibid (1980), 1636. 85) Tomioka, H.; Inagaki, T.; Nakamura, S.; Izawa, Y. J. C. S. Perkin .1 (1979), 130. 86) a) Senthilnathan, V. P.; Platz, M. S. J. Am. Chem. Soc. (1980), 102, 7637. b) Senthilnathan, V. P.; Platz, M. S. Ibid (1981), 103, 5503. 87) Long, M. A.; Willard, J. E. J. Phys. Chem. (1970), 74, 1207. 167 88) Lin, C.-T.; Caspar, P. P. Tetrahedron Lett. (1980), 2_1, 3553. 89) a) Bol'shakov, B. V.; Tolkatchev, V. A. Chem. Phys. Lett. (1976), 41), 468. b) Bol'shakov, B. V.; Stepanov, A. A.; Tolkatchev, V. A. Internat. J. Chem. Kin. (1980), 271. 90) Caldin, E. F. Chem. Rev. (1969), 6JJ, 135. 91) a) Platz, M. S.; Senthilnathan, V. P.; Wright, B. B.; McCurdy, C. W., Jr. J. An. Chem. Soc. (1982), 104, 6494. b) Wright, B. B.; Senthilnathan, V. P.; Platz, M. S.; McCurdy, C. W., Jr. Tetrahedron Lett. (1982), 2_3, 833. 92) GLoss, G. L.; Closs, L. E. J. Am. Chem. Soc. (1969), 91_, 4549, and references contained herein. 93) Closs, G. L. in "Chemically Induced Magnetic Polarization", Lepley, A. R. and Closs, G. L., Eds., Wiley, New York, 1973. 94) a) Fleming, J. C.; Shecter, H. J. Org. Chem. (1969), _34, 3962. b) Canquis, G.; Reverdy, G. Bull. Soc. Chim. France (1975), 1841. c) Canquis, G.; Reverdy, G. Tetrahedron Lett. (1967), 1493. 95) Bell, R. P. "The Tunnel Effect in Chemistry", Chapman and Hall, New York, 1980. 96) a) Hutzler, J. S.; Colton, R. J.; Ling, A. C. J. Chem. Eng. Data (1972), 17, 324. b) Ling, A. C.; Willard, J. E. J. Phys. Chem. (1918), 72, 1918. 97) LeRoy, R. J.; Murai, H.; Williams, F. J. An. Chem. Soc. (1980), 102, 2325. 98) Shin, H. J. Chem. Phys. (1963), 39, 2934. 99) Reference 91a. Also, I wish to thank Prof. C. W. McCurdy for making the computer program available. 100) Bauschlicher, C. N.,Jr.; Bender, C. F.; Schaefer, H. F., Ill JL Am. Chem. Soc. (1976), 98, 3072. 101) a) Wang, H. Y.; Willard, J. E. J. Phys. Chem. (1979), 83, 3585. b) Wilkey, D. D.; Willard, J. E. J. Chem. Phys. (1976), 64, 3976. c) Aditya, S.; Wilkey, D. D.; Wang, H. Y., Willard, J. E. J. Phys. Chem. (1979), 83, 599. d) Long, M. A.; Willard, J. E. ibid (1970), 7_4, 1207. 102) Toriyama, K.; Iwasaki, M. ibid (1978), 182, 2056. 103) a) Hudson, R. L.; Shiotani, M.; Williams, F. Chem. Phys. Lett. (1977), toJ, 193. b) Campion, A.; Williams, F. J. Am. Chem. Soc. (1972), 94, 7633. c) LeRoy, R. J.; Sprague, E. D.; Williams, F. 168 ibid (1971), 93, 787. 104) a) LeRoy, R. J.; Sprague, E. D.; Williams, F. J. Phys. Chem. (1972), _76, 546. b) Sprague, E. D.; Williams, F. J. Am.~~Chem. Soc. (1971), 93, 787. c) Sprague, E. D.; Takeda, K.; Wang, J-T.; Williams, F. Can. J. Chem. (1974), 525, 2840. d) Wang, J-T.; Williams, F. J. An. Chem. Soc. (1972), 94, 2930. 105) a) Hadel, L. M.; Platz, M. S.; Scaiano, J. C. ibid, in press, b) Unpublished results. 106) "CRC Handbook of Chemistry and Physics", 60th Ed., Weast, R. C., Ed., CRC Press, Boca Raton, Fla., 1979. 107) a) Roth, H. D. J. An. Chem. Soc. (1971), 93, 4934. b) Roth, H. D. ibid (1972), 94, 1400. c) Roth, H. D. Acc. Chem. Res. (1977), H), 85. 108) Bethel, I). ; Newall, A. R.; Whittaker, D. J. Chem. Soc. B (1971), 23. 109) Griller, D.; Hadel, L.; Nazran, A. S.; Platz, M. S.; Wong, P. C.; Scaiano, J. C. J. Am. Chem. Soc., in press. 110) A Qirie plot obtained by irradiation of 9-diazo-10,10-dimethyl-9,10-dihydroanthracene crystals indicates that DMA is a ground state triplet. Curvature in the plot suggests that a very low lying singlet state may also exist. 111) Trozzolo, A. M. Acc. Chem. Res. (1968), _1_, 330. 112) Kasha, M. J. Opt. Soc. An. (1948), 38, 929. 113) McBride, J. M.; Vary, M. W. Tetrahedron (1982), 38, 765. 114) a) Matsuyama, T.; Yamaoka, H. J. Chem. Phys. (1978), 6*5, 331. b) Matsuyama, T.; Yamaoka, H. ibid (1970), j60, 468. 115) Kirmse, W.; Horner, L.; Hoffmann, H. Ann. (1958), 614, 19. 116) Guthrie, R. D.; Pendygraft, G. W.; Young, A. T. J. Am. Chem. Soc. (1976), 98, 587. 117) Nally, J.; Nazareno, J.; Polensk, J.; Maulding, H. V. J. Pharm. Sci. (1975), 64, 437. 118) Seidlova, V.; Protiva, M. Collect. Czech. Chem. Commun. (1967), 32, 2826; cf. C.A. (1967), 67, 90575b. 119) Kruyt, W.; Veldstra, H. Landbouwk. Tijdschr. (1951), 63, 398; cf; C.A. (1951), 45, 7287c. 120) Hoening, D. E.; Muchowski, J. M. Can. J. Chem. (1968), 46, 3665. 169 121) Cioranescu, E. ; Bucu, A.; Banciu, M. ; Nenitzescu, C. D. Rev. Roumaine Chim. ( 1965), K>, 141; cf. C.A. ( 1965), 63, 1145917 122) Schering, A. G. Ger. 1,213,856 (1966); cf. C.A. (1966), 64, P19632b. 123) Zieger, H. E.; Angres, I.; Mathisen, D. J. An. Chem. Soc. (1976), 98, 2580. 124) Bornstein, J.; Nunes, F. J. Org. Chem. (1965), 3(3, 3324. 125) Walborsky, H. M.; Pitt, C. G. J. An. Chem. Soc. (1962), 84, 4832. 126) Amiel, Y.; Ginsburg, D. Tetrahedron (1957), 1_, 19. 127) Goh, S. H.; Chan, K. C.; Chang, H. L. Aust. J. Chem. (1976), 29, 1699. 128) Murray, R. W.; Trozzolo, A. M. J. Org. Chem. (1964), 29_, 1268. 129) . Finkelstein, H. Ber. (1910), 43, 1532. 130) Wenner, W. J. Org. Chem. (1952), 523. 131) Bartlett, P. D.; Simons, D. M. J. An. Chem. Soc. (1960), 82, 1753. 132) Smith, R. L.; Manmade, A.; Griffin, G. W. J. Heterocyc. Chem. (1960), 6, 443. 133) Forster, M. O.; Fierz, H. E. J. Chem. Soc. (1907), 91, 1953. 134) Smith, L. I.; Hoehn, H. H. in "Organic Syntheses" Collective Volume 3, Horning, E. C., Ed., John Wiley and Sons, New York, 1955. 135) a) Schmitz, E.; Ohme, R. Ber. (1961), jM, 2166. See also: b) Schmitz, E. Ber. (1962), 9Jj, 688. c) Smith,R. A. G.; Knowles, J. R. C^ S^ Perkin II (1975), 686. 136) Davis, M. A.; Winthrop, S. 0.; Thomas, R. A.; Herr, F.; Charest, M-P.;Gaudry, R. J. Med. Chem. (1964), 7_, 88. 137) Baltzly, R.; Mehta, N. B.; Russell, P. B.; Brooks, R. E.; Grivsky, E. M.; Steinberg, A. M. J. Org. Chem. (1961), 26^, 3699. 138) Moritani, i.; Murahashi, S-I.; Nishino, M.; Kimura, K.; Tsubomura, H. Tetrahedron Lett. (1966), 4, 373. 139) Regitz, M. Ber. (1964), 97, 2742. 170 140) Queginer, G.; Pastour, P. C. R. Acad. Sci. C. Paris (1966), 262, 1335.