вход по аккаунту


Free Radical Studies at Low Temperatures.

код для вставкиСкачать
Free Radical Studies at Low Temperatures
Free radicals can be produced in the gas phase and then condensed together with the
molecules of an inert gas; they can also be generated and trapped in isolated sites in
rigid solids. I R and ESR spectra of trapped radicals provide information concerning their
structure and chemical properties. The techniques employed for producing and studying
radicals are reviewed and a number of recent I R and ESR investigations of trapped
radicals are discussed. The last part of the article deals exclusively with the applications
of the rotating cryostat to the study of the ESR spectra and reactions of trapped radicals
at low temperatures.
1. Introduction
Free radicals (molecular fragments containing an unpaired electron) are active participants in a wide variety of systems undergoing chemical change 111. For
example, the pyrolysis and combustion of organic
materials involve highly reactive free radicals and the
extensive conversions that occur are due to the high
efficiency of the reactions between radicals and molecules. The high reactivity of free radicals can also
result in extreme complexity since the conversion of
reactants to products occurs through a maze of numerous, very fast radical-molecule reactions taking
place in sequence and in parallel. The very efficient
reaction between two radicals to form molecular products also causes the concentration of radicals in most
systems to be so low that, not only have the concentrations not been measured, but even the presence of
radicals in the system has usually been inferred from
the nature of the products and has not been directly
The fact that the radical concentrations have not been
measured inevitably means that there are more reaction steps than there are products and that, consequently, there is uncertainty both in the reaction
scheme adopted and in the rate constants assigned to
the assumed reaction steps. The consistency of the
data for the same radical reaction studied in entirely
different systems does, in most cases, lend support to
the reaction steps proposed and to the existence of the
radicals postulated; but, clearly, these proposed reaction mechanisms would be put on a firmer basis if the
radicals postulated could be shown to exist and if the
reactions could be observed directly and in isolation.
Of equal, if not greater, importance is the fact that the
physical and thermodynamic properties of radicals
(Le., the orbital of the unpaired electron, the vibration
frequencies, etc.) must be known before any reliable
theory of radical-molecule reaction rates can be formulated. Thus, there are strong arguments for attempt-
[*I Dr. B.
“Shell” Research Limited, Thornton Research Centre
P. 0. Box 1
Chester (England)
111 E. W. R. Steacie: Atomic and Free Radical Reactions. Reinhold, New York 1954, 2nd Edition, Vol. I, p. 14.
Angew. Chem. internat. Edit. J VoI. 7 (1968) J No. 7
ing to obtain sufficient concentrations of free radicals
in a high state of purity so that their physics and chemistry can be studied more directly.
There are two ways of producing radicals in sufficient
concentrations to be detected by physical techniques
such as ESR spectroscopy: (a) the rate of radical generation can be increased; this is done in Aash photolysis 121, pulse radiolysis 131, and recently in the radiolysis of liquid hydrocarbons [41; (b) the rates of radicalradical termination reactions can be reduced; this is
the approach adopted in the study of trapped radicals
at low temperatures, and it is with this method that
this review is concerned. The collision between radicals, which is the prerequisite of reaction, is reduced
to a very small value by forming and trapping the
radicals in isolated sites in rigid solids at low temperatures. The rates of diffusion of the radicals through
the solid matrix can be made so low that the radicals
can be preserved for years without any significant
decrease in their concentration.
Work on trapped free radicals was begun by G. N .
Lewis and his students151 in the early 1940’s. They
showed that photolysis of frozen solutions of aromatic
hydrazines with ultraviolet light resulted in the formation of free radicals which could be kept for long
periods of time. In 1951, Rice and FreamoC6Jdemonstrated another way of generating trapped radicals
when they decomposed hydrazoic acid by passing it
through a heated tube at low pressures, and trapped
out reactive intermediates from the gas phase onto a
surface cooled with liquid nitrogen. (For reviews and
details of work carried out on this topic up to 1960
see [71.)
The article will be divided into four sections dealing
with: (1) experimental methods, (2) the infrared
[2] G. Porter: Technique of Organic Chemistry. Interscience
Publ., New York 1963. 2nd Edition, Vol. 111/2, p. 1055.
131 L. M . Dorfman and M . S. Matheson in G. Porter: Progress in
Reaction Kinetics. Pergamon Press, Oxford 1965, Vol. 3, p. 237.
[4] R . W. Fessenden and R . H . Schuler, 3. chem. Physics 39,2147
151 G . N. Lewis and D. Lipkin, J. Amer. chem. SOC.64, 2801
161 F. 0. Rice and M . Freamo, J. Amer. chem. SOC. 73, 5329
[7] A . M . Bass and H . P. Broida: Formation and Trapping of
Free Radicals. Academic Press, New York 1960.
spectra of free radicals, (3) recent ESR studies, and (4)
the use of the rotating cryostat in trapped radical
2. Experimental Methods of Preparing and
Trapping Radicals
2.1. Matrices : The Matrix Isolation Technique181
It is impossible to prevent collisions between radicals
in the liquid and gas phases (except in the extremely
dilute regions of a molecular beam) and it is only in
inert solids that radicals can be kept apart. The solid
matrix used for holding the radicals in dispersed sites
several molecules apart should satisfy the following
Rigidity - Clearly the matrix must be sufficiently
rigid to prevent radicals diffusing and this is usually
ensured by working at low temperatures. However,
some work has been done with ionic solids containing
trapped radicals as, for instance, in studies of y-irradiated single crystals of carboxylic acids and their
salts 191. It is worthwhile noting that radical diffusion
occurs at a temperature well below the melting point
of the matrix; usually the radicals begin to disappear
when the temperature rises to between one-tenth and
one-third of the melting point (OK).
Inertness - The matrix must not react with the free
radicals being studied. Otherwise, not only will radicals from the matrix be generated rather than the
radicals of interest, but there is also the possibility of
“chemical” diffusion whereby the free valences migrate through the matrix by a series of atom-transfer
reactions between adjacent molecules. Additionally,
the matrix must not be decomposed by the light which
is used to photodissociate the guest reagent molecules
into radicals.
Transparency - For visible, ultraviolet, and infrared
spectroscopic studies, ideally the matrix should not
absorb radiation in the same wavelength regions as the
radicals do. The rare gases are obviously the most
suitable matrices from this viewpoint. In the case of
electron spin resonance spectroscopy, the spectra are
often considerably better resolved if the radicals are
trapped in sites in which they can undergo extensive
“tumbling” motion. This helps to average out the
effects on the spectrum of the orientation of the free
radical in the magnetic field. (In most trapped radical
studies, amorphous solids are used so that if the free
radicals are in a fixed orientation, there is an overall
random distribution of all possible orientations with
a consequent broadening of the lines.) Thus, the best
matrix would be one having large cavities in which the
radicals can be trapped and still rotate in three dimensions, as well as having sufficient rigidity to prevent
translational motion of the radicals.
[8] E. D. Becker and G. C. Pimentel, J. Chem. Physics 25, 224
[9] A . Horsfield, J. R. Morton, and D . H. Whiffen, Molecular
Physics 4, 327 (1961).
Volatility - Apart from those cases where solution$
are frozen and then irradiated, most of the techniques
require that the matrix be deposited from the vapor
phase onto a cooled surface. The matrix must therefore have sufficient volatility (-1 torr) at temperatures
below those at which the matrix molecules decompose.
The rare gases are the most commonly used matrices
but hydrocarbons, fluorocarbons, and even water have
also proved useful.
2.2. Formation of Radicals Prior to Trapping
There are two distinct ways of preparing the radicals
that are to be trapped: (a) the radicals can be formed
in the gas phase and subsequently trapped in the solid
phase by co-condensing the radicals with molecules of
inert gas; (b) the radicals can be both formed and trapped in the solid phase.
2.2.1. F o r m a t i o n o f R a d i c a l s i n t h e
Gas Phase
Electric discharges are most frequently used for dissociating gaseous molecules (especially for making
simple radicals, like H, NH, etc.). The gas mixture,
consisting mainly of a rare gas with a small amount of
the compound to be dissociated, is passed through
either an electrodeless discharge or a discharge employing electrodes (Woods tube) and the resulting
dissociated mixture is then quickly condensed onto
a cold surface located close to the end of the discharge.
This technique is simple and straight-forward to use,
but it suffers from the disadvantage that the processes
occurring in the discharge are complex and the identity
and degree of purity of the radicals produced are
sometimes uncertain.
Gaseous molecules can also be dissociated by pyrolysis [61.Klein and Scheer [lo1 have studied the reactions
of dissociated hydrogen atoms by heating a tungsten
filament to a high temperature (2000°K) in a stream
of hydrogen and then bombarding frozen solid olefins
with the partially dissociated gas. Atoms and small
molecules of carbon can also be generated by thermal
methodsI111. These methods are more useful for
producing atoms by the dissociation of diatomic molecules than for producing polyatomic free radicals.
Only atoms and simple radicals such as OH, CH, C2,
NH2, and PH2, have been trapped from the gas phase
and there is a simple reason for this. The combination
of atoms and simple radicals can only occur in the
presence of a third molecule which can remove some
of the energy released on combination. (If this energy
is not removed the nascent molecule will dissociate at
the end of the first vibration.) Because of this thirdbody requirement, the rate of combination of atoms
in the gas phase can be reduced considerably by reducing the pressure and thus sufficient concentrations
of atoms can be built up to ensure the trapping of ap-
[lo] R . Klein and M. D . Scheer, J. physic. Chem. 62, 1011 (1958).
[ll] R. L. Barger and H. P. Broida, J. chem. Physics 43, 2364,
2371 (1965).
Angew. Chem. internat. Edit.
VoI. 7 (1968) 1 No. 7
preciable numbers when the gas is condensed. More
complex radicals with a larger number of bonds to
absorb the energy released on combination do not
require a third body for combination so that it is very
difficult to achieve sufficiently large concentrations of
these radicals in the gas phase. The condensation
process itself probably results in extensive combination
since the gaseous molecules and radicals are brought
much closer together as they form the solid phase.
Nevertheless, it would be of considerable interest to be
able to trap and identify radicals from complex gaseous systems where, for example, pyrolysis or oxidation
occurs. Possibly the use of techniques similar to those
used for the mass spectrometric C121 examination of
free radicals might be helpful in this connection.
2.2.2. F o r m a t i o n o f R a d i c a l s i n t h e S o l i d
The molecules to be dissociated are dispersed widely
in the inert matrix by condensing the reagent and the
inert matrix on a cold surface from the gas phase. The
frozen solid is then irradiated with visible or ultraviolet or far-ultraviolet light, which causes photodissociation of the reagent molecules into free radicals.
Radiolysis can also be used but the reagent itself must
then serve as the matrix since the absorption of ionizing irradiation is non-selective. Photolysis of wellchosen molecules can be clean and can result in the
formation of one or only a small number of radicals.
The major problem with this method is the possible occurrence of the “cage” effect which can reduce the quantum
yield of radical formation to very low values. Photolytic dissociation of a molecule always results in the formation of two
radicals (of opposite spin) and not one, and these radicals
must separate from each other if they are not to recombine
to form the parent molecule. This separation can be difficult
in the solid phase where the two radicals are “caged” together and prevented by surrounding molecules from diffusing apart. Although the cage effect can be an important
factor, it can be overcome by a suitable choice of conditions [13,141. For instance, if a molecule dissociates to give one
small fragment that can move readily through the matrix,
then the other larger fragment will be trapped. A simple reorientation of the radicals can also result in stabilization [151.
lead to the generation of a large number of radicals in
close proximity with resultant loss of the radicals by
Once the radicals have been formed and trapped, a
variety of techniques can be used for their identification and characterization. Emission and absorption
spectrometry, X-ray diffraction, calorimetry, infrared
spectrometry, and electron spin resonance spectrometry have all been used for this purpose. Only the
latter two techniques will be discussed in this article.
3. Infrared Spectra of Trapped Radicals
The characteristics of the ultraviolet and visible absorption spectra of trapped small radicals are well
known, but it is only recently that it has been possible
to obtain the infrared spectra of radicalsCl71. The
problem has been one of obtaining sufficiently high
concentrations of radicals so that infrared absorption
by the radicals is detected above the background absorption by the matrix.
The infrared spectrum of a radical can yield information
about the frequencies of the various vibrations in the radical
and also about the structure and thermodynamic properties
of the radical. For instance, it is often possible to decide
whether a radical is planar or non-planar, or whether two
atoms of the same element are equivalent o r non-equivalent.
Thermodynamic functions such as the entropy and heat
capacity can be calculated by statistical mechanics if the
vibration frequencies and the bond lengths and angles of the
radicals are known. The frequencies can be obtained from
the infrared spectra but in most cases the bond lengths and
angles have to be guessed - this is a serious gap in our knowledge and can only be removed by observation of rotational
fine structure in the infrared spectra o r by the observation of
the microwave spectra of radicals.
3.1. Infrared CeIls Used for Trapped Radical Studies
Most of the infrared work has been carried out in a cell similar to that first described by Becker and Pimentel[81 and
shown schematically in Figure 1. A Dewar vessel containing
the refrigerant, liquid hydrogen, helium,-or-nitrogen, is sur-
A hybrid of the two methods of gas and solid phase
formation of radicals has also been employed. In this
method the sample is photolyzed while it is being
condensed and in some cases there is a considerable
enhancement in yield over the simple photolysis of the
condensed solid solution 1161. This may be due both to
a reduction in the cage effect and to the fact that every
region of the deposit has been irradiated and not
merely the surface layers. (Absorption of a large
amount of light in the first few layers of a solid can
[121 F. P. Losing and A. W.Tichner, J. chem. Physics 20, 907
[13] G. C. Pimentel in A. M . Bass and H. P. Broida: Formation
and Trapping of Free Radicals. Academic Press, New York 1960,
p. 95.
[14] I. Norman and G. Porter, Proc. Roy. SOC.(London), Ser. A
230, 399 (1955).
(151 H . Linschitz in: Informal Discussion on Free Radical
Stabilization. Faraday Society, Sheffield 1958, p. 37.
[161 E. L. Cochran, F. J. Adrian, and V. A. Bowers, J . chem.
Physics 40, 213 (1964).
Angew. Chem. internat. Edit. 1 VoI. 7 (1968)
I No. 7
Figure 1 . Schematic diagram of a low temperature cell used for infrared
examination of deposits.
1 - Liquid helium. 2 - Liquid nitrogen. 3 - Evacuated space. 4 - CsBr
window. 5 - IR and UV radiation. 6 - Inlet tube for matrix and reactants. 7 - AgI window. 8 - To IR spectrometer. 9 - Copper block and
rounded by another Dewar vessel containing liquid nitrogen,
this outer Dewar serving as a radiation shield. A copper rod
protrudes from below into the refrigerant in the inner Dewar
and to the lower end of this rod is attached a copper block
holding a silver iodide window (or other window transparent
to infrared radiation). The silver iodide window is thus
cooled via the copper block and rod. The matrix and substance under examination are co-condensed on the window,
photolyzed with ultraviolet radiation, and investigated infrared spectroscopically.
3.2. Infrared Spectra
3.2.1. T h e F o r m y l R a d i c a l , H C O
The formyl radical (see also Section 4.2.1) was the
first radical whose infrared spectrum was recorded.
Ewing, Thompson, and Pimeatelf171 ingeniously avoided the cage effect by using a reactive matrix. HI or DI
molecules were dispersed in solid carbon monoxide
and the hydrogen or deuterium atoms produced by
photodissociation reacted with the carbon monoxide
to form formyl radicals, HCO or DCO. The radicals
were detected by their absorption bands in the visible
spectral region.
Subsequently, MilIigan and Jacox
were able to
produce much larger concentrations of HCO or DCO
by the simultaneous condensation and photolysis of
mixtures of HzS or D2S and CO. These workers assigned the following frequencies: C=O stretching
frequency of 1861 cm-1 in HCO and 1800 cm-1 in
DCO; C-H and C-D stretching frequencies of 2488
and 1937cm-1, respectively; and C-H and C-D
bending frequencies of 1090 and 852 cm-1, respectively.
Perhaps the .most interesting feature of the spectra is
the very low value of the C-H stretching mode at
2488 cm-1 (C-H stretching frequencies in normal
molecules occur at ca. 3000 cm-1). This indicates
strongly that the C-H bond in HCO is considerably
weaker than that in HzCO and this is entirely in accord with kinetic evidence which indicates that
D(co-H) is 18 kcal per mole compared with a value of
88 kcal per mole for D(HCO-H)
[191. The electron spin
resonance spectrum of HCO also provides evidence of
a very weak C-H bond (see Section 4.2.1).
The thermodynamic functions of the formyl radical
can now be calculated fairly accurately since the bond
angles and bond lengths of this radical are known
from the absorption spectrum of the radical in the gas
phase [205211.
3.2.2. T h e H y d r o p e r o x y l R a d i c a l , H O z
identify. Milligan and Jacox [221 finally succeeded in
obtaining its infrared spectrum in argon at 4 ° K by
employing a technique similar to that used for the
study of the formyl radical. Argon, hydrogen iodide,
and oxygen were deposited onto a cooled cesium
bromide window with simultaneous photolysis of the
cold film. Argon was used as a diluent to prevent
possible reaction between H 0 2 and the iodine atom
also produced by photodissociation of the HI. Three
absorption bands were obtained for HI602 and these
were assigned to the H-02 stretching mode (3414
cm-I), the 0-0 stretching mode (1101 cm-I), and the
H-02 bending mode (1389 cm-1). The values of the
OH stretching and bending mode frequencies are near
to those for these modes in H202 indicating a tightly
bound hydrogen atom in the HO2 radical. This is in
accord with the recent estimation of D ( H - o ~=
47.1 & 2 kcal/mole compared to D(Ho~-H)
= 89.6 It 2
kcal/mole Q31. The 0-0 stretching frequency is intermediate between that in H202 (880 cm-1) and that in
0 2 (1580 cm-1). The spectra of H160180/H190160
mixtures are interesting in that they show that the two
oxygen atoms are not equivalent. This contradicts
theoretical calculations which indicate that HO2 has
the configuration of an isosceles triangle [241.
3.2.3. T h e M e t h y l R a d i c a l , C H 3
Although much effort has been expended in the search
for the infrared spectrum of the simplest alkyl radical,
the methyl radical, it is only very recently that an absorption due to the out-of-plane bending mode has
been observed by Andrews and Pimentel1251. The methyl
radicals were generated by condensing methyl iodide
in argon simultaneously with a beam of lithium atoms.
Although bands attributabIe to lithium iodide and
methyllithium were detected, an additional absorption
at 730cm-1, which decayed on warming, was also
observed. From a comparison of the corresponding
band obtained when deuterated methyl iodide was
used, the authors assigned the observed frequency to
an out-of-plane bending mode and calculated a force
constant of 0.2527 mdyn/A. This is close to force
constants for the same mode in BBr3 (0.29) and B13
(0.24) and thus provides some support for the planar
structure of the methyl radical. No infrared spectra of
any other alkyl radicals have yet been reported (see
also Section 4.3.1.).
3.2.4. T h e T r i f l u o r o m e t h y l R a d i c a l , C F 3
The H02 radical, believed to be one of the most important intermediates in oxidation systems, has proved
to be one of the most elusive radicals to trap and
Milligan, Jacox, and Comeford [261 have succeeded in
recording the spectrum of CF3 in a variety of matrices.
The radical was prepared by a number of different
[17] G. E. Ewing, W. E. Thompson, and G. C. Pimentel, J. chem.
Physics 32, 927 (1960).
[18] D.E.Milliganand M.E.Jacox, J.chem. Physics41,3032(1964).
[19] J. A . Kerr, Chem. Reviews 66, 465 (1966).
[20] G . Herzberg and D. A . Ramsay, Proc. Roy. SOC.(London),
Ser. A, 233, 34 (1955).
1211 J. W. C. Johns, S. H. Priddk, and D. A . Ramsay, Discuss.
Faraday SOC.35, 90 (1963).
[22] D. E. Milfigan and M. E. Jacox, J. chem. Physics 38, 2627
[23] S. N. Foner and R.L. Hudson, J. chem. Physics 36, 2681 (1962).
1241 M . E. Boyd, J. chem. Physics 37, 1317 (1962).
[25] W. L. Andrews and G. C. Pimentel, J. chem. Physics 44,2527
[26] D. E. Milligan, M. E. Jacox, and J. J. Comeford, J. chern.
Physics 44, 4058 (1966).
5 10
Angew. Chem. internal. Edit. 1 Vol. 7 (1968) / No. 7
methods but the infrared spectra from the deposits had
three features in common and all these features decayed on warming. The three bands occur at 703 cm-1,
1084 cm-1, and 1250 cm-1 and have been assigned to
an umbrella mode, to a symmetric, and to an asymmetric stretching mode respectively. The finding of
two absorptions (1084 cm-1 and 1250 cm-1) in the
region of the C-F stretching fundamentals is very
important since it shows that the CF3 radical is nonplanar. (In a planar molecule only the asymmetric
stretching mode would be seen since the symmetric
mode would involve no change in dipole moment.) In
the case of CF3 a direct comparison is possible between results in the gas phase and those in the solid
phase since Carlson ’ and Pimentel1271 have also succeeded in obtaining the infrared spectrum of the CF3
radical in the gas phase by using rapid-scan infrared
techniques in conjunction with flash photolysis. Transient absorption bands were observed whose positions
were nearly identical with those observed for the trapped CF3 radicals. Recent examination of the electron
spin resonance spectrum of CF3 has also shown that
the radical is non-planar (vide infua).
3.2.5. O t h e r R a d i c a l s
The foregoing should suffice to illustrate the application
of infrared spectroscopy to trapped radicals. The
spectra of the following radicals haveIalso been observed and interpreted but will not be described in
detail here: NH2 [28,29];NH [29,301; NF2 1311; NF, NC1
and NBr 1321; N C N [331;*CCO [341; FCO 1351; ClCO 1361;
CF2 1371; and the radical anion COi 1381.
4. The Electron Spin Resonance Spectra of
Trapped Radicais
Electron spin resonance spectroscopy has several advantages for the study of radicals: it only “sees” species
containing free electrons so that matrix transparency
[27] G. A . Carlson and G . C. Pimentel, J. chem. Physics 44, 4053
[28] D . E. Milligan and M . E. Jacox, J. chem. Physics 43, 4487
[29] D . E. Milligan and M . E. Jacox, J. chem. Physics 41, 1199
[30] K . Rosengren and G . C. Pimentel, J. chem. Physics 43, 507
[31] M . D . Harmony and R. J. Myers, J. chem. Physics 35, 1129
[32] D . E. Milligan and M. E. Jacox, J. chem. Physics 40, 2461
[33] D . E. Milligan, M. E. Jacox, and A. M . Bass, J. chem.
Physics 43, 3149 (1965).
[34] M . E. Jacox and D . E. Milligan, J. chem. Physics 43, 3734
[35] D. E. Milligan, M . E. Jacox, A . M . Bass, J. J. Comeford, and
D . E. Mann, J. chem. Physics 42, 3187 (1965).
[36] M . E. Jacox and D . E. Milligan, J. chem. Physics 43, 866
[37] D . E. Milligan, D. E. Mann, and M . E. Jacox, J. chem.
Physics 41, 7199 (1964).
[38] K. 0.Hartmann and J . C. Hisatsume, J. chem. Physics 44,
1913 (1966).
Angew. Chem. internat. Edit. / Vol. 7 (1968) 1 No. 7
is not a problem; it “sees” the very part that causes the
high reactivity of the free radical; and it is extremely
sensitive and can detect free radical concentrations as
low as 10-8 molar (cf. 139,409.
The technique depends o n the fact that the electron has a
magnetic moment (associated with its spin) which can lie
either parallel or antiparallel to a magnetic field in which the
electron is placed. In a sample containing many free electrons
those with magnetic moments lying parallel to the field have
a slightly lower energy than those lying anti-parallel to it, the
relative numbers in each state (Np and Nap respectively)
being given by the Botzmann equation:
N a p / N p = exp ( - A E / k T )
where A E is the energy of separation which is equal to g p H ,
g is a dimensionless number (g-factor) having the value of
2.0023 for a free electron, p is the Bohr magneton, and H is
the field strength in gauss (G). When electromagnetic radiation of the correct frequency falls o n the sample (hv = g P H ) it
is absorbed and some of the electrons in the low energy state
are raised to the high energy state. (v is in the microwave
region, 9000 MHz, when fields of 3000 G are used.) Of great
interest to the chemist is the fact that nuclei close to the electron can cause hyperfine structure in the absorption if these
nuclei themselves possess a magnetic moment. For example,
in a hydrogen atom (nuclear spin quantum number, I = 1/2)
the magnetic moment of the nucleus, like that of an electron,
can lie parallel o r anti-parallel to the field and these different
orientations will result in the magnetic field experienced by
the electron being either greater or less than the externally
applied field. Two absorption lines are thus observed
For hydrogen atoms, Hp,the effective field due to the proton
at the electron, has a value of about 250 G. The presence of
more than one proton in the vicinity of the electron, as in
alkyl radicals, causes further splitting of these lines and clearly
this hyperfine structure gives direct information about the
environment of the unpaired electron and hence about the
orbital which it occupies.
The interaction of the free electron with neighboring nuclei
is of two types, isotropic and anisotropic. The latter is simply
the dipole-dipole interaction between the magnetic moments
of the electron and that of the nucleus, and its magnitude is
inversely proportional to the cube of their separation and also
dependent o n the relative orientation of the two spins and
hence on the orientation of the radical in the magnetic field.
If the radicals are rigidly held and their orientations are
random, this interaction can lead to considerable broadening
of the hyperfine lines with consequent loss in resolution. A
“tumbling” motion of the radicals in their sites results in the
anisotropic interaction being averaged out to zero. The isotropic interaction, or the Fermi contact interaction 1411, only
occurs when there is a finite unpaired electron density at the
neighboring nucleus. This isotropic interaction is not averaged out to zero by molecular motion and it is this interaction
which is usually measured in the spectra of randomly
orientated trapped radicals.
ESR experiments can be carried out in a similar way to infrared experiments. The copper rod and windows (Figure 1) are
replaced by a non-conducting sapphire rod (electric insulator)
and this rod is located in a microwave cavity between the pole
[39] D . J. E. Ingram: Free Radicals as Studied by Electron Spin
Resonance. Butterworths, London 1958.
[40] J . E. Wertr, Chem. Reviews 55, 829 (1955); A. Carrington,
Quart. Rev. (Chem. SOC.,London) 17, 67 (1963); C. .Ten, in [3],
Chapter 3; M . C . R. Symons, Advances physic. org. Chem. 1,283
[41] E. Fermi, Z . Physik 30, 320 (1930).
pieces of a powerful magnet. Alternatively, the free radicals
can be formed and trapped in samples outside the spectrometer and then subsequently placed in the cavity. The samples
must, of course, be prepared, transferred, and measured at
low temperatures and silica glass Dewars are used for these
Recently Cochran, Adrian, and Bowers [441 have observed the 13C hyperfine splittings in the HCO radical,
and have determined the HCO angle as 125", thus
confirming the bent structure.
4.2.2. T h e Ethynyl a n d Vinyl R a d i c a l s ,
H C = C - a n d HzC=CH[161
4.1. a and x-Radicals
The classification into a and x-radicals is dependent on
the structure and symmetry of the orbital of the free
electron ; it also has important chemical implications.
In a-radicals the free electron is in an sp-hybridized
orbital with contributions from both the s and p
orbitals of the atom at which the free valence formally
occurs; this usually results in the free electron orbital
being non-perpendicular to the bonds attached to the
free valence atom. In x-radicals the free electron orbital has a contribution from only the p orbital of the
free valence atom and is perpendicular to the bonds to
this atom; these bonds may themselves be sp-hybrid
bonds and may be either co-linear or co-planar.
4.2. a-Radicals
Photolysis of acetylene while being co-deposited with
argon results in the formation of the ethynyl radical,
HC=C. The ESR spectrum shows a hyperfine interaction of 16.1 G with the proton. Unfortunately, the
interpretation of this interaction in terms of the structure of the radical is compIex but it does seem likely
that the radical is of a- and not x-type. Observation of
the 13C interactions (13C has a magnetic moment, 12C
does not) would provide more information about the
structure of the radical.
Vinyl radicals, CH2=CH, were formed by addition of
hydrogen atoms to acetylene diluted with argon (analogous to the system used to generate HCO). The ESR
spectrum showed that the unpaired electron interacted
with three non-equivalent protons and from the effect
of deuteration and from theoretical considerations it
4.2.1. T h e F o r m y l R a d i c a l , HCO
The formyl radical could be either linear, H-C=O
(x-radical), or bent, >C=O (a-radical). Adrian,
Cochran, and Bowers [421 produced the radical by reaction of hydrogen with solid carbon monoxide and
also by photolysis of solid formaldehyde and methyl
alcohol. The most striking feature of this work is the
large proton hyperfine interaction of 137 G observed.
(Since the resonance frequency v and the magnetic
field H are related by hv = gpH, the hyperfine interaction or splitting constant can be expressed in gauss
or in megahertz.) This large interaction together with
the pronounced anisotropy of the g-factor show that
the formyl radical is a o-type radical with the bent
structure, ( I ) .
was shown that the hyperfine splitting constants could
be assigned as follows: aHcr= 16.0 G; aHcis = 34 G ;
amrans = 68 G. Cochran, Adrian, and Bowers1161
calculate the @ angle to be between 140 " and 150 ".
In the formyl radical the cc proton interaction is considerably greater (137 G).This is explained in part by the
larger resonance of the carbon monoxide molecule and
in part by the fact that the interaction is greater the
less the angle @ (in HeO, @ = 125 "). An interesting
finding in this work is the fact that only the transisomer is found when hydrogen atoms are added to dideuterioacetylene.
The other canonical form, (2), with a loose interaction
between a free hydrogen atom and a carbon monoxide
molecule, also makes an important contribution. This
is in complete accord with the information obtained
from the infrared and kinetic studies (see Section
Brivati, Keen, and Symons[431 have also observed the
ESR spectrum of the formyl radical and come to similar conclusions.
[42] F. J. Adrian, E. L. Cochran, and V. A. Bowers, J. chem.
Physics 36, 1661 (1962).
[43] J. A . Brivati, N, Keen, and M. C. R. Symons, J. chem. SOC.
(London) 1962.237.
5 12
H / D
(51, trans
This is likely to be due to the zero point vibrational
energies of the two isomers being sufficiently different
so that only the more stable isomer is present in detectable concentrations at temperatures below 32 "K.
The ESR spectrum of the vinyl radical has also been
observed recently in liquid ethylene by Fessenden and
Schuler [41 and the only significant difference between
the [solid] liquid and solid phases is that in liquid ethylene the u proton is inverting from one side of the
molecule to the other at such a rate that the interactions with the p protons are averaged.
[44] E. L. Cochran, F. J. Adrian, and V. A. Bowers, J. chem.
Physics 44, 4626 (1966).
Angew. Chem. internat. Edit.1 Vol. 7 (1968) 1 No. 7
Fessenden 1451 has recently observed the 1x13‘2 splitting
in the vinyl radical (107.57 G) and from this has calculated @ to be 151 which is in the range suggested
by Cochran, Adrian, and Bowers. (The magnitude of
the hyperfine splitting constant of the a13C is dependdent on the amount of s character in the hybridize
orbital and this in turn will also determine the bond
angle.) All the features of the ESR spectrum of the
vinyl radical, especially the al3C interaction, show
conclusively that it is a a-type radical.
4.2.3. T r i f l u o r o m e t h y l , CF3, a n d
O t h e r F l u o r i n a t e d Methyl R a d i c a l s
Fessenden and Schuler 1461 have observed the spectra
of the interesting fluorinated methyl radical trapped
in inert gas matrices, and also of CF3 in liquid C2F6
(see also Section 3.2.4). In solid krypton and xenon
the concentrations of radicals were sufficient for the
13C splitting to be detected. From the value of the 13C
interaction, these authors have calculated that the free
electron orbital has 21 % carbon 2s character and that
the radical is non-planar (there is an angle of 17.8”
between the C-F bonds and the plane normal to the
threefold symmetryaxis). In CHF2 theorbital has10 % C
2s character and the angle is 12.7 O ; the CH2F radical
is nearly planar. There is thus a slow change from a
planar radical to a pyramidal radical as the hydrogen
atoms of the methyl group are replaced by fluorine
atoms (see Section 4.3.1). The CF3 and CFzH radical
ara again 0-type radicals.
4.2.4. T h e C y c l o p r o p y l R a d i c a l
The ESR spectrum of the cyclopropyl radical in the
liquid phase has been seen only by Fessenden and
SchuIer [41 and although it was not possible to observe
any 13C interaction, it is clear from the value of the a
hydrogen splitting that the three bonds to the free
valence carbon atom do not lie in a plane and that the
radical is of the o-type.
4.3. x-Radicals
4.3.1. T h e Methyl R a d i c a l
absorption spectrum of the radical in the gas phase 1481.
The methyl radical is thus a x-type radical with the
free electron located in a carbon 2p orbital, the hydrogen atoms being bound to the carbon through sp2hybrid orbitals. An interesting result that follows from
the different values of the 13C interaction in CH3, CD3,
CDzH, and CDH2 has been determined by Fessenden [451. These values show that there is incomplete
orbital following in these species (ie., the bonding
electrons lag behind the nuclei as the nuclear skeleton
undergoes bending vibrations). This is thought to be
due to a decrease in electron repulsions caused by the
occurrence of “bent” bonds 1491.
4.3.2. O t h e r x - R a d i c a l s
Most of the alkyl radicals appear to be of this type and
these are discussed in Section 6.1.1. The benzyl radical [501, C6H&&, also belongs to this type and in this
case the planar configuration about the -CH2 group
is favored by the occurrence of overlap of the p orbital
on this carbon atom with the x-orbitals of the benzene
4.4. Oxygen-containing Radicals
4.4.1. Alkoxy R a d i c a l s
Unfortunately, there is a conspicuous lack of information on the ESR spectra of alkoxyradicals. The samples
from the y-irradiation of frozen ethanol 1511 and ultraviolet photolysis of H~Oz/ethanolmixtures 1521, followed by annealing of the samples at 100 OK, give rise
to a slightly asymmetric single-line ESR spectrum
which has been assigned to ethyl and methyl alkoxy
radicals. No structural information is obtainable from
these spectra and the assignments cannot be regarded
as certain. The electron may not have an appreciable
interaction with protons in the alkyl group but much
more work is needed before any conclusions can be
drawn with confidence.
5. The Rotating Cryostat
5.1. Principle of the Technique
The hyperfine interaction with the protons is 23 G and
with 13C is 38.5 G, indicating that the radical is nearly
planar, in agreement with the conclusion based on the
The rotating cryostat is really a means of mixing two
solids together intimately at the molecular level and in
a controlled fashion so that single-step chemical reactions can occur and the primary products can be
preserved for further study of their chemical and phps-
1451 R. W. Fessenden, Meeting of Amer. chem. SOC.Division of
Physical Chemistry, Michigan August 1966, p. 1 1 .
1461 R . W. Fessenden and R . H . Schuler, J. chem. Physics 43,2104
[47] W.Gordy and C . G. McCormick, J. Amer. chem. SOC.78,
3243 (1956); C. K . Jen, S. N . Foner, E. L. Cochrun, and V. A.
Bowers, Physic Reviews 112, 1169 (1958); B. Smaller and M . Matheson, J. chem. Physics 28, 1169 (1958); T. Cole, H. 0.Prifchard,
N. B. Davidson, and H. M . McConneN, Molecular Physics I , 406
[48] G. Herzberg, Proc. Roy. SOC.(London) Ser. A, 262,291
[49] D. M. Schrader and M . Karplus, J. chem. Physics 40, 1593
(1 964).
[50] A . Carrington and I. C. Smith, Molecular Physics 9 (2), 137
[51] C. Chacury and E. Mayer, Nature 200, 59 (1963); J. chem.
Physics 61, 1115 (1964).
[521 B. N. Shelimov, N. V. Fok, and V. V. Voevodskii, Kinetika i
Kataliz 5, 1008 (1964).
The ESR spectrum of methyl radicals has been observed in the solid phase I471 and in the liquid phase [41.
Angew. Chem. internar. Edit. / VoI. 7 (1968) / No. 7
ical properties. For example, the reaction between a
halohydrocarbon and alkali atoms,
+ Na’
+ NaX,
+ R-
has been used to generate hydrocarbon radicals.
takes place and the product radicals, RO2, together
with any unchanged R, are covered by the next matrix
layer and trapped. If the ESR spectra of the reactant
and product radicals differ sufficiently, the extent of
the conversion can be calculated directly from the
composite spectrum. Clearly, other reactions and
other methods of generating radicals can be used in
conjunction with the technique.
The rotating cryostat has a number of advantages over the
simpler co-deposition methods where the vapors of different
compounds are condensed on a stationary cold surface. The
motion of the cold surface in the rotating cryostat allows the
two processes of condensation and chemical reaction to be
separated from each other and therefore allows more control
of the system. It is also possible to study further reactions of
the radicals initially formed, and this again can be done in a
controlled way.
5.2. Experimental Equipment
Fig. 2. Principle of the rotating cryostat.
The principle of the method is illustrated in Figure 2.
Alternate layers of the two reactants are deposited
from the gas phase onto the outer surface of a stainless
steel drum which contains liquid nitrogen and is rotated rapidly (ca. 2400 r.p.m.) in a high vacuum
( < l o - 5 torr). The gaseous reactants are directed onto
the drum from jets located close to the surface and on
opposite sides of the drum so that each is deposited on
a freshly formed surface of the other. The flow rates
are adjusted so that about nine monolayers of halohydrocarbon are deposited per revolution but only
about 1/25 of a monolayer of sodium. This “sprinkling” of the surface with sodium atoms ensures that
very few alight adjacent to each other so that the radicals are formed in isolated sites. These radicals, immobilized on the surface, are then covered and trapped
by the next layer of halohydrocarbon, and a small
number of the surface molecules of this next layer are
then converted into radicals by bombardement with
further sodium atoms. This whole process is repeated
many times until a solid deposit about 1 cm wide and
about 1 mm thick is formed which is an interleaving
spiral of a very dilute, free-radical layer and a much
more dense halohydrocarbon layer.‘At the end of the
experiment this deposit which contains about 1018-19
radicals per gram is transferred from the drum, still at
77 “K and under high vacuum, into a glass tube suitable
for electron spin resonance measurements.
The technique can be slightly modified so that radicals
can be prepared and trapped in materials other than
the parent halohydrocarbon. This is achieved simply
by deposition of the matrix as a third layer which
forms the bulk of the deposit, the flux of halohydrocarbon being reduced to a value similar to that of
sodium. In this case three jets are used instead of two.
A similar modification allows the reactions of radicals
with molecules to be studied. Before the radicals are
covered up they are bombarded with a stream of the
reactant molecules from the third jet. For example,
the radicals can be bombarded with oxygen molecules,
when the reaction
A vertical section through the apparatus is shown in Figure 3.
The drum itself is constructed from stainless steel and the
parts are argon-arc-welded or vacuum-brazed together.
Fig. 3. General arrangement of cryostat (all hatched parts rotate;
others are stationary), about 116 natural size.
1 - Filling tube for liquid nitrogen. 2 - Outlet for coolant. 3 - Terminals for heater windings. 4 - Sodium wire. 5 - To vacuum pumps.
6 - Spinning drum containing liquid nitrogen. 7 - Liquid nitrogen.
8 - Inlet for coolant.
The assembly is rotated by a belt-driven pulley attached to
the outer tube, and it is the outer tube which carries the rotating seal. The “heart” of the apparatus is the seal around
the outer shaft, which must allow the drum to be rotated at
high speeds but must also seal the shaft so that pressures of
less than 10-5 torr can be maintained in the main vessel (a
high speed pump and wide-bore tubes are also used to help
to maintain these low pressures). Two rubber O-rings, manufactured for rotating seals, are housed in two grooves with a
1 cm separation between them. The lower O-ring is the vacuAngew. Chem. internat. Edit. 1 Vol. 7 (1968) ! N o . 7
um seal, while the upper one is a retaining ring which allows
silicone oil to be pumped around the rotating shaft between
the two O-rings. This flow of oil serves both to lubricate the
seals and also to carry away the frictional heat which is
generated by rotation of the shaft. (For further details
see [ 5 3 , 5 4 1 , )
6. Applications of the Rotating Cryostat
The results obtained can be divided into two categories:
those that give information about the structure of the
radicals and those that give direct information about
the chemical reactions of radicals.
6.1. ESR Spectra and Structure of Radicals
hydrocarbon. One difference is the finding of a readily interpretable spectrum for the isopropyl iodide-sodium system
with the rotating cryostat, whereas a very complex spectrum
was observed by Ayscough and Thomson after the radiolysis
of isopropyl iodide.
The spectra of the alkyl radicals can be interpreted o n the
following basis:
1) in all the alkyl radicals only the tc and P protons (i.e. those
attached to the free valence carbon atom and the adjacent
carbon atoms) show any resolvable interaction with the unpaired electron (interaction with y protons from 0.4 to 1.1 G
can be observed in the liquid phase where the lines are narrower [ 4 9 ;
2) the tc protons, all three protons of a PCH3 group and one
of the protons of a PCHzR group have similar hyperfine
splitting constants of about 22 to 25 G;
3) the other proton in a PCH2 group and the one proton in a
PCH group have a hyperfine splitting constant of about 42 to
45 G.
These features can be explained by assuming that the
a carbon atom adopts an sp2 hybridization with a coplanar arrangement of the bonds around it and with
6.1.1. A l k y l R a d i k a l s
Several radical species may be formed in the initial
stages of photolysis or radiolysis and these primary
radicals can either react with matrix molecules to form
new radicals (the photon energy in excess of that required for bond breaking can be used for these reactions) or themselves be photolyzed into new species.
With the rotating cryostat technique, the radicals are
formed by the reaction
+ Na-
R’+ NaX
the unpaired electron occupying a p-orbital whose
axis is perpendicular to this plane. In order to explain
the inequality of the two FCHzR protons, it must be
assumed that the hyperfine interaction depends on the
orientation of the FCH bond with respect to the axis
of the p-orbital. Horsfieid, Morton, and Whiffenl91
have shown that this is so in free radicals formed by
y-irradiation of carboxylic acids and have also shown
that the angular dependence of the interaction is given
so that the position of the free valence in the radical is
determined by the position of the halogen atom in the
parent halohydrocarbon molecule. The ESR spectra
of the deposits can therefore be assigned to definite
radical species with considerable confidence.
A large number of alkyl radicals of each radical type have
been prepared in a variety of matrices and the ESR spectra in
each case can be interpreted in terms of the radical expected
from the sodium-halohydrocarbon reaction. Some typical
ESR spectra are shown in Figure 4. The results [ 5 4 , 5 5 1 are in
general agreement with those of Ayscough and Thomson 1561
who avoided some of the difficulties accompanying radiolysis
by preparing the radicals by y-irradiation of the parent halo-
Fig 4. The first derlvatwe ESR spectra of
(b) 4-heptyl radrcaI
(a) n-heptyl radical
__ . [53] A.Thomas, Trans. Faraday SOC. 57, 1979 (1961).
1541 J. E. Bennett and A. Thomas, Proc. Roy. SOC.(London) Ser.
A, 280, 123 (1964).
1551 J. E. Bennett, B. Mile, and A . Thomas, unpublished.
[56] P . B. Ayscough and C.Thomson, Trans. Faraday SOC. 58,
1471 (1962).
Angew. Chem. internat. Edit. 1 YoI. 7 (1968) j No. 7
where H B is the P proton splitting, BO and B1 are constants, and @ is the angle between the BC-H bond and
the p-orbital axis projected onto a plane perpendicular
to the C-C bond. From the values of the hyperfine
interaction the angles for the two 9 protons in alkyl
radicals were calculated to be 1 2 (44 G interaction)
and 48 (22 G interaction). The fact that the two p
C-H bonds are non-equivalent shows that there is a
barrier to rotation about the C-C bond. This does not
necessarily mean that the molecule is fixed rigidly in
any configuration but indicates that the lifetime in
any of the four possible orientations which have these
angles must be greater than 10-7 sec (the reciprocal of
the hyperfine splitting).
From the chemical viewpoint the interaction with the
cc and p protons must mean that the cc and $ C-H
bonds in the radical are weaker than in the corresponding alkane. In simple valence bond theory, this can be
described in terms of contributions from canonical
structures such as H.CHz=CHz. For an isolated
hydrogen atom the hyperfine interaction has a value
of about 500 G so that a 22 G interaction can be interpreted on this basis to mean that the hydrogen is about
4 % (22/500) dissociated in the radical. It is known
from kinetic data that the C-H bond strengths in
alkyl radicals are lower than those in the corresponding
alkane (for example: D(c?H~-H) = 98.0 kcal per mole
while D(;H~CH>-H) = 39 kcal per mole) 1191
The results from these solid phase studies are in good accord
with those obtained by Fessenden and SehuIer[41 for alkyl
radicals in the liquid phase. Any differences between corresponding spectra are due to effects of temperature and phase
o n the rates of intramolecular internal reorientation of the
fact that radicals such as methyl do not attack matrices
such as cyclohexane while the phenyl radical does SO
with great ease.
6.1.4. T h e COF a n d CSF R a d i c a l Anions[61,621
6.1.2. S u b s t i t u t e d Ally1 Radicals[571
Because the free ally1 radicals can be formed with the
free valence at a specific site, the rotating cryostat
allows a n interesting direct demonstration of the
phenomenon of resonance in organic molecules. In
two separate experiments sodium atoms were caused
to react with 3-chloro-1-butene (6) and with l-chloro2-butene (7). In the absence of resonance, radicals (8)
and ( 9 ) should have quite different ESR spectra.
+ Na.
+ NaCI
+ Na-
I 7)
+ NaCl
Deposition of sodium or potassium onto solid carbon
dioxide or solid carbon disulfide at 77 "K resulted in
the formation of ion pairs, Na@COF,etc. There was a
small hyperfine coupling to the alkali metal nucleus
which showed that the unpaired electron was not
completely transferred to the carbon dioxide or carbon
disulfide molecule. Similar ion pairs can be formed by
y-irradiation of alkali formate crystals [63,641 and the
differences between the hyperfine splitting constants
of the alkali metal atoms in the two cases (one a molecular solid, the other an ionic solid) demonstrates the
marked effect the environment can have on the unpaired electron distribution between the members of
an ion pair.
6.1.5. A l i p h a t i c K e t y l R a d i c a l Anionsr651
In fact, the radicals from both chloro compounds gave
identical spectra consisting, at 77"K, of seven lines
with an equal separation of 15 G . This shows that the
unpaired electron is interacting equally with six equivalent protons and is entirely in accord with the generally held view that structures (8) and (9) are canonical
forms of the same radical species. The one proton that
does not show any interaction at 77 "K is that on C-2.
6.1.3. T h e P h e n y l Radical[ssl
The phenyl radical is considered to be one of the most
reactive hydrocarbon radicals [59,601. This radical can
be prepared in a rotating cryostat from phenyl iodide
and can be trapped when matrices of water, benzene,
deuteriobenzene, and perfluorocyclohexane are used;
however, when matrices such as camphane, hexamethylethane, and cyclohexane are used the radical
abstracts a hydrogen atom from a matrix molecule to
form a matrix radical. The ESR spectrum indicates
that the unpaired electron interacts with the two ortho
protons (18.1 G ) and to a lesser extent with the two
meta protons (6.4 G) but not with the para proton.
This shows that the unpaired electron remains in an
spz orbital on the carbon atom at which bond scission
has occurred and the radical is thus of the a-type. The
projection of the spz orbital outwards from the aromatic ring helps to explain the very high reactivity of
the phenyl radical since appreciable bonding interactions can occur with adjacent molecules at distances
where the repulsive interactions are still low. The extreme reactivity of the phenyl radical is shown by the
[57] J . E. Bennett, B. Mile, and A . Thomas, unpublished.
[ 5 8 ] J. E. Bennett, B. Mile, and A . Thomas, Chem. Commun.
1965, 265; Proc. Roy. SOC. (London) Ser. A , 293, 246 (1966).
[591 F. J. Duncan and A . F.Trotman Dickenson, J. chem. SOC.
(London) 1962, 4672.
I601 W. Fielding and H. 0. Pritchard, J. physic. Chemistry 66,
821 (1962).
The ketyl radical anions (10) of acetone, biacetyl,
cyclopentanone, and cyclohexanone were prepared by
reaction of sodium (or potassium) atoms with the
solid ketones at 77 OK,
+ Na'
N o differences in spectra were detectable when
potassium was used instead of sodium, showing that
the electron had been transferred completely to the
ketone molecule. From the values of the proton hyperfine splitting constants, the unpaired spin density at
the carbonyl carbon was calculated to be only 70 % of
that in the corresponding alkyl radical in which hydrogen replaces the O Q group. This reduction is due to
the occurrence of an appreciable spin density at the
carbonyl oxygen atom, i.e., structures such as (11)
are important.
6.1.6. T r a p p e d E l e c t r o n s
Radiation chemistry has recently undergone a major
development since it has been shown that the primary
reducing species produced by the passage of ionizing
and A.Thomas7 Trans. Faraday
~ 1 2 & ~ ~ $B' ~ ~ t t ~
[621 J. E. Bennett, B. Mile, and A. Thomas, Trans. Faraday SOC.
63, 262 (1967).
[63] D .
Ovenall and D. H . Whiffen, Molecular Physics 4 , 135
[64] p . W . Atkins, N . Keen, and M . C. R. Synions, J. chem. SOC.
(London) 1962, 2813.
[65] J. E. Bennett, B. Mile, and A. Thomas, J. chem. SOC. (London)
A, 1968, 298.
Angew. Chem. internat. Edit.
Vol. 7 (1968) / No. 7
radiation through polar media such as water are solvated electrons and not hydrogen atoms1661 as was
formerly supposed. A solvated electron is not associated with one molecule but is stabilized by interaction
with an assembly of solvent molecules. ESR spectra of
solvated electrons can be observed in liquid ammonia [66al, in frozen alkaline water, and in alcohols [6795*1.
However, their spectrum is obscured by the spectra of
other paramagnetic species such as 0 0 , OH, and ahydroxyalkyl radicals.
The rotating cryostat has been used to prepare trapped
electrons as the sole paramagnetic species by the reaction of alkali metal atoms with ice and solid alcohols
at 77 “K 1691. All the deposits were intensely colored
and at 77 “K their ESR spectra consisted of a single
narrow line. At 170 “K the spectrum Fig. 5 of trapped
electrons in water became better resolved to reveal the
presence of seven equally spaced lines (5.6 G apart)
indicating an equal interaction with six protons of
neighboring water molecules. The most likely arrangement of the water molecules around the electron is one
with a n octahedral disposition of the protons.
50 G
one hydrogen sulfide molecule [711. This marked difference between water and hydrogen sulfide is understandable because (1) the molecules in H2S are not
highly associated and have a lower dipole moment
than molecules of water or alcohol and are thus less
likely to form suitable electron traps; (2) the sulfur
atoms has low-lying 3d orbitals which can accommodate the extra electron, whereas the oxygen does not.
6.2. Reaction Studies with the Rotating Cryostat
In these experiments the conditions under which reaction can occur are somewhat unusual in that a radical and a reactant molecule are “caged” or locked
together in adjacent sites in the solid phase at 77 OK.
The reactants are continually “colliding” with each
other as they undergo lattice vibrations and librations
of varying energy and amplitude. During the time of
twenty minutes between the formation of the “caged”
radical-molecule pair in the deposit and the recording
of the ESR spectrum, reaction can occur when the
energy in the caged region exceeds a critical activation
energy with the reactants correctly orientated for reaction. If it is assumed that the energy flows in and out
of the cage region with a frequency similar to that of
lattice vibrations (v = 1012 sec-1) and that the pseudounimolecular reaction rate is given by vPexp (-EA/RT),
it can be readily shown that virtually complete reaction
of all radical-molecule pairs occurs in the time available if E, < 5 kcal per mole ( T = 77°K) and the
steric factor P = 10-1 to 10-4 sec-1. Reactions with
activation energies in excess of 5 kcal per mole cannot
be conveniently studied because of the long times
which would be involved.
Fig. 5. The first derivative ESR spectrum of trapped electrons in ice
at 170°K.
When the samples containing trapped electrons in
alcohols were either warmed slightly or exposed to
visible light, the color of the samples faded and the
spectrum changed to that of the corresponding a-hydroxyalkyl radical, RlRzCOH 1701. The rate of this
conversion could be measured from the change in the
ESR spectrum with time, and it appeared that the ratecontrolling step was a reorientation of the molecules
around the electron trap into a configuration in which
reaction could occur.
It is interesting that solvated electrons are not formed
with hydrogen sulfide. Instead, simple negative ions
H2Se are formed in which the electron is attached to
1661 See, for instance, the papers in Rad. Res. Suppl. 4 (1964).
[66a] C. A . Hutchinson and R. C. Pastor, J. chem. Physics 21,
1959 (1953).
[67] D . Schulte-Frohlinde and K . Eiben, Z. Naturforsch. 17, 445
[68] M . J . Blandamer, L. Sheilds, and M . C. R. Symons, Nature
199, 902 (1963).
[69] J. E. Bennett, B. Mile, and A.Thomas, Nature 201, 919
[70] J . E. Bennett, B. Mile, and A. Thomas, J. chem. SOC.(London)
A, 1967, 1394, 1399.
Angew. Chem. internat. Edit. / Yol. 7 (1968) / No. 7
6.2.1. A d d i t i o n o f R a d i c a l s t o O x y g e n
The reaction
+0 2+M
+ ROz*
+ M [721
and the subsequent reactions of the peroxyl radicals,
R02, are amongst the most important steps in any
hydrocarbon oxidation at moderate temperatures
( < 500 “C). The reaction has been investigated for a
large number of radicals of widely different types (e.g.,
alkyl, benzyl, and CF3 radicals) by bombarding the
radicals with molecular oxygen.
Since the ESR spectra of the radicals R* are usually
very different from those of the corresponding peroxyl
radicals, it is a relatively simple matter to estimate the
extent of conversion of R* into R 0 2 from a composite
spectrum. The changes that occur in the case of the
tert-butyl radical are shown in Figure 6. It is worth
emphasizing that this is a direct observation on frozen
radicals and does not involve the analysis of products
1711 J. E. Bennett, B. Mile, and A . Thomas, Chem. Commun.
1966, 182.
1721 J. E. Bennett, B. Mile, and A. Thomas, Eleventh International Symposium on Combustion. The Combustion Institute,
Pittsburgh 1967, p. 853.
Fig. 6. Reaction between terf-hutyl radical and oxygen at 77 OK. First
derivative ESR spectra of (a) tert-butyl; (b) terf-butyl after bombardment with less than the stoichiometric quantity of oxygen; and (c) fenbutyl after bombardment with an excess of oxygen.
of further reaction. In all cases, even at 77 OK, compIete conversion to the peroxyl radical occurred when
sufficient oxygen was used, indicating a reaction of
high efficiency and low activation energy. This extends
to other radicals the previous conclusion about the
high efficiency of this reaction made by Dingledy and
Calvert, and Hoare and Walsh who studied the reactions of methyl and ethyl radicals with oxygen in the
gas phase [73,741.
The ESR spectra of the different peroxyl radicals are
striking in that they are almost identical, being independent of the nature of the group R, and showing
no hyperfine interaction with the protons in this group.
This indicates that the free electron occupies an orbital
that is confined almost entirely to the 0-0 region of
the radical and is unaffected by the substituent group.
Since the reactivity of radicals in terms of activation
energy is determined principally by the orbital of the
free electron it follows that all peroxyl radicals will
have similar activation energies in bond-forming
6.2.2. T h e A d d i t i o n o f R a d i c a l s t o O l e f i n s
Early studies with the rotating cryostat showed 1751 that
not all radicals react with a particular olefin. For example, phenyl radicals react with ethylene at 77 OK,
whereas tert-butyl and trichloromethyl radicals do not.
Alkyl radicals d o not react with tetradeuteroethylene
[73] D . P. DingJedy and J . G. Culvert, J. Amer. chem. SOC.81,
769 (1959); 85, 856 (1963).
[74] D. E. Hoare and A . D. Walsh, Trans. Faraday SOC.53,1102
[75] J . E. Bennett and A.Thomas, Lecture at the Sixth International Symposium on Free Radicals, Cambridge 1963, Paper N.
at 77 "K [761. The model used in the earlier work must
be modified since it assumed that reaction occurred
either on direct first collision between the radical and
the alighting molecule or very soon afterwards. Subsequent work has shown that reaction can occur at any
time from the collision of the molecule with the surface
to the recording of the ESR spectrum ofthe deposit [771.
The uncertainty in time and in the value of v (the
lattice vibration frequency) means that only upper
limit estimates of EA and P can be made. It is therefore
more profitable to study the site of radical addition in
asymmetric olefins, than to attempt such estimates.
Phenyl radical addition to I-hexene [761 leads to the
formation of a radical the ESR spectrum of which
shows that addition occurs exclusively at the terminal
C H 2 group. This is a direct observation and no inferences have to be made to arrive at this conclusion.
In the case of phenyl radical addition to C H z = C D 2 [761
a secondary deuterium isotope effect could be studied
since the phenyl radicals could add to the deuterium
end of the olefin to give the radical C ~ H S - C D ~ - C H ~ *
or to the hydrogen end to give C6H5-CHzCD2..
ESR spectra of these two radicals were different and
the relative amounts of each radical could be evaluated
from the composite spectrum. In this way it was found
that C ~ H ~ - C D ~ - C H ~ * / C ~ H S - C H ~=- C
ing that the phenyl radical had a marked preference to
add at the deuterium end of the olefin. This indicates
that the sum of the carbon-hydrogen/deuterium vibration frequencies in the activated complex is greater
than that in the ground state. This arises because,
during the reaction, a soft out-of-plane bending C-H
vibration in the olefinic trigonal carbon is converted
into a harder bending vibration in the tetrahedral
product radical.
From the observed inverse isotope ratio of 4 and a
simplified form of Bigeleisen's equation [77,781 the
difference between the sum of the frequencies in the
activated complex and in the ground state was calculated to be 570 cm-1. The total change in the sum of
frequencies for complete conversion of phenyl to
phenylethyl was estimated to be 900 crn-1, and a value
of 570 cm-1 for the activated complex is a clear proof
that the complex is product-like and that a strong
bond exists between the phenyl radical and the olefin
in the activated complex.
The advantages of the rotating cryostat for this type of study
are clearly shown here; the low temperature accentuates
small differences in reactivity (for instance, an isotope effect
of 1.4 is predicted for this reaction at 25 "C) and the direct
determination of the site of radical addition by ESR avoids
difficulties that might arise because of isotope effects that can
occur in product analysis.
1761 J. E . Bennett, B. Mile, and A . Thomas, unpublished.
1771 J. Bigeleisen, J. chem. Physics 17, 675 (1949).
[78] A . Strreitweiser j r . , R. H . Jagow, and S. Suzuki, 3. Amer.
chem. SOC.80, 2326 (1958).
Angew. Chem. internat. Edit. 1 VoI. 7 (196%) 1 No. 7
6.2.3. A b s t r a c t i o n R e a c t i o n s
R*-i-R,H + RH-f-Rl-
The activation energies for hydrogen abstraction by
most radicals are too high for detectable reaction to
occur at 77°K. Phenyl, cyclopropyl, vinyl, and CF3
radicals are the exceptions and these attack other
molecules with great ease even at 77”K[58,761. This
clearly demonstrates their higher reactivity and it is of
some interest to note that all these radicals are of the
o-type, in which the free electron is in a localized
orbital projecting away from the rest of the molecule.
Received: August 7, 1967
[A 636 IE]
German version: Angew. Chem. 80, 519 (1968)
Modification of the Orientation of Substitution Reactions on Thiophene
and Furan Derivatives
Under normal conditions, thiophene and furan derivatives are substituted in the a position,
and no convenient alternative methods for the preparation of ,&substitution products have
been available until now. The present article describes a method that permits the synthesis
of many P-substituted thiophenes and furans. In this method, the carbonyl group in aaldehydes or ketones of the thiophene and furan series is blocked by complex formation
with an excess of aluminum chloride, so that etectrophilic substitution takes place in
position 4. In another useful method, the carbonyl group is blocked by acetalization.
The acetals can be metalated in the ring by organolithium compounds.
1. Introduction
Though the term “aromaticity” [**I was first introduced more than a century ago, there is still no satisfactory
definition or experimental test that could be applied to
establish whether o r not a system is aromatic. Our
present ideas on aromatic systems are dominated by
generalizations that follow from Hiickel’s application
of the MO theory to aromatic molecules and from the
further development of this author’s views. Despite
the simplifications and approximations involved, the
resulting physical picture enables us to explain some
of the properties of such systems more or less precisely,
and to predict the existence of new, similar compounds.
However, questions concerning the properties of individual types of aromatic compounds were largely
ignored for a long time.
Peculiarities in the behavior even of compounds that
are formally very similar have inthe meantime attracted considerable attention, from both the theoretical
and the practical points of view. A comparison of the
thiophenes with the furans shows that many properties
are changed when a sulfur atom is replaced by an oxygen. Greater differences are observed on comparison
of these and other heteroaromatic compounds with
Prof. Dr. Ja. L. Goldfarb, Dr. Ju. B. VolkenStein, and
Dr. L. I. Belenkij
N. D. Zelinskij Institute of Organic Chemistry of the
Moscow V-334, Leninskij Prospekt 47 (USSR).
[**I The term “aromatic” is also used for heteroaromatic systems in this paper.
Angew. Chem. internat. Edit. / Vol. 7 (1968) No. 7
benzene derivatives. These differences are particularly
obvious in substitution reactions; the unsubstituted
hetcroaroinatic compounds can very often be regarded
purely formally as monofunctional compounds.
One peculiarity of thiophene and furan is the strict
orientation of electrophilic and protophilic substitutions, in which the substituent always becomes attached to the a position of the ring [** *]. In the few cases in
which $substitution products were found, these were
probably formed by isomerization of the a-substituted
compound formed initially [4,51. Thus P-substituted
products can generally be obtained only indirectly.
How can the orienting effect of the hetero atom be
overcome in thiophene and furan chemistry? A solution to this problem would be valuable, since e.g.
many natural products are fJ-substituted furans and
thiophenes. The $-substitution products may also
include physiologically active compounds.
[***I The deuterium exchange kinetics show that thea position of
thiophene reacts three orders of magnitude faster than the p
position in electrophilic substitutions 11, 21 and six orders of
magnitude faster in protophilic substitutions 121. The a position
in furan reacts three orders of magnitude faster than the p position in protophilic substitutions [3].
[l] K . Halvarson and L. Melander, Ark. Kemi 8, 29 (1955).
121 A . I . SatenSIein, A. G. Kamrad, I. 0 . Snpiro, Ju. I. Rnnneva,
and E. N.Zvjaginceva, Doklady Akad. Nauk SSSR 168,364 (1966).
[3] A. I. SatenStein, A . C. Kamrad, I . 0.sapiro, Ju. I. Ranneva,
and S. A . GiIIer, Chim. geterocikliieskich Soedinenij 1966, 643.
141 H. Wynberg and U.E. Wiersum, J. org. Chemistry 30, 1058
[5] N . I. h i k i n , B. L. Lebedev, V . G . Nikolskij, 0. A . Korytina,
A. V . Kessenich, and E. P . Prokofev, Izvest. Akad. Nauk SSSR,
Ser. chim. 1967, 1618.
Без категории
Размер файла
1 393 Кб
free, low, temperature, radical, studies
Пожаловаться на содержимое документа