Reprinted from The Journal of Chemical Physics, Vol. 42, No. 8, 2651-2657, 15 April 1965 Printed in U. S. A. Electron Spin Resonance and Electronic Structure of the RCHOR' Ether Radicals* O. Haves Griffith Gates and Crellin Laboratories of Chemistry,f California Institute of Technology, Pasadena, California (Received 30 October1964) Single crystals of inclusion compounds formed between urea and aseries of aliphatic ethers were xirradi ated and studied by electron spin resonance. The stable, x-ray-produced free radicals were all of the general type RCHOR'. The approximate value for the spin density on the carbon atom is 0.70±0.10. The unpaired spm distnbution is discussed in terms of the Hiickel and approximate configuration interaction ^--electron molecular orbital models and the valence bond method. The theoretical spin distributions are found tobe in qualitative agreement with the experimental spin distribution. I INTRODUCTION N the preceding paper,1 the radical O . II RHC—CR' was investigated by electron spin resonance (ESR). This ketone radical is of special interest because it is one of the simplest heteroatom radicals in which each atomcontributes one electron to the* system. Radicals in which one atom contributes two electrons to the w system are also of interest, and one of the simplest examples of this type of radical is RHC-OR'. Here the spin density is primarily localized on only two atoms: the oxygen atom and an adjacent carbon atom. In this paper a positive identification of the radical RCHOR' is reported in a series of ether-urea inclusion compounds.2 Approximate values of the carbon and oxygen spin densities aredetermined from thecoupling- constant data and the spin distribution is discussed in terms of the x-electron molecular orbital and valance bond methods. EXPERIMENTAL Toprepare single crystals ofeach inclusion compound investigated, the ether was added slowly to a urea- saturated methanol solution until the inclusion com- * Work supported in part by the National Science Foundation (Grant No. GP-930), and in part by a grant from the Shell Companies Foundation. t Contribution No. 3179. 1O. H. Griffith, J. Chem. Phys. 42, 2644 (1965) (referred to as II). 2We are unaware of anyprevious investigation of this type of aliphatic ether radical in an oriented matrix. Radicals of the type ROCHC02H and ROCHC02- have been reported previously[A. Horsfield and J. R. Morton, Trans. Faraday Soc. 58, 470(1962), and references quoted therein]. However, it is difficult to estimate the effect of the carboxyl groupon both the formation and the unpaired spin distribution of the ether radical (the re moval of an a proton adjacent to a carboxyl group by x irradia tion is well known). An ESR investigation of the alcohol radical reported by Dixon and Norman3 is closely related to our work and we have occasion to compare the results obtained for the two systems. 3W. T. Dixon andR. O. C. Norman, J. Chem. Sgc. 1963, 3119. pound began to precipitate out ofsolution. Theprecipi tate was then redissolved by the addition of a slight excess of methanol, and the solution was cooled slowly from 298° to 273°K over a period of from 36 to 48 h. The resulting crystals were long hexagonal needles. The s axis of each crystal is defined as lying along the needle axis, and the plane perpendicular to the needle axis is referred to as the xy plane. Apparently no crystallographic data has been reported for these ether-urea crystals. The general hexagonal structure of urea inclusion compounds has, however, been shown to be independent of the exact nature of the linear host molecule.4 We may safely assume, therefore, that the ether-urea crystals have the tubular structure charac teristic of organic urea inclusion compounds.5 The ether-urea inclusion compounds are relatively unstable, decomposing in 1-3 h in air at room tem perature. To avoid this problem, the crystals were x irradiated at liquid-nitrogen temperatures and the majority oftheESRspectra were takenwith thesample at <~273°K, rather than at room temperature. Below 273°K the crystals were stable for at least one or two days. A few crystals, x irradiated at 273°, had 273°K ESR spectra identical to those obtained from crystals x-irradiated at 77°K. It appears, therefore, that the 273°K ESRspectra areindependent of the temperature at which the crystals were x irradiated. The other experimental details, including the x-ray tube, X-band ESR spectrometer, and cooling apparatus were the same as employed in II. RADICAL IDENTIFICATION To obtain a positive identification of the x-my produced free radicals it was necessary to investigate more than one aliphatic ether. The walls of the tubular cavities hinder intermolecular radical reactions but do not prevent intramolecular radical rearrangements and therefore there are several possible structures for the final radicals produced. Furthermore, the relative mag nitudes of the f- and7-proton coupling constants were 4W. Schlenk, Jr., Ann. Chem. 565, 204 (1949). 6A. E. Smith, Acta Cryst. 5, 224 (1952). 2651 I L2652 O. HAYES GRIFFITH I IOOMc/sec Fig. 1. The 273°K ESR spectra of an x-irradiated dibutyl ether-urea crystal with the magnetic field in the xy plane and parallel to the z axis, respectively. not known.6 However, it sufficed to investigate examples of two types of ether molecules; RCH2OCH2R and RCH2OCH3. As examples of the first type, several of the symmetricalethers were investigated briefly. Single crystals of the inclusion compounds formed between urea and dibutyl ether (di-w-butyl ether), di-w-pentyl ether, di-w-hexyl ether, di-ra-octyl ether, or di-w-decyl ether were prepared and x irradiated. The ESR spectra were qualitatively the same for all five systems. The spectra of dibutyl ether, however, was much more nearly symmetric (suggesting the presence of only one radical), therefore this compound was chosen for further study. The spectra obtained with the magnetic field along the needle axis and perpendicular to the needle axis of the dibutyl ether-urea crystal are given in Fig. 1. The methyl octyl ether (methyl M-octyl ether) urea inclusion compound was chosen as an example of a long-chain methyl ether, RCH2OCH3. The spectra ob tained from these crystals at ~273°K are shown in Fig. 2. These spectra result from one anisotropic cou pling constant, two equal and nearly isotropic coupling constants, and three small coupling constants. The small splittings are only resolvedwhen the anglebetween the magnetic-field vector and the z axis is less than ~75°. The dibutyl ether-urea spectra, on the other hand, result from one anisotropic proton coupling constant, two equal and nearly isotropic coupling con stants, and two much smaller coupling constants. Again the two small coupling constants are not resolved when the magnetic-field vector is within 15° of the xy plane. From the consideration of both sets of data it is easily seen that the radicals produced from dibutyl ether and methyl octyl ether are, respectively, CH3CH2CH2CHOCH2(CH2)2CH3 (1) 8The convention for the labeling of protons {a, /3, 7, f) and Carbon Atoms (Ci, C2, C3) used here is R- 7 -CH2- -CH2- C, -CH- C2 -0—CHZ—R'. C3 and CH3(CH2)6CH2CHOCH3 (2) The anisotropic proton coupling constant and the large isotropic coupling constants are the familiar a- and /3- proton coupling constants, a" and aP, respectively. The small coupling constants are associated with the f protons rather than the 7 protons since the spectra of Radicals (1) and (2) exhibit small triplet and quartet splittings, respectively. In addition to the above inclusion compounds, one othercompound, 1,4-diethoxybutane-urea, was investi gated in order to obtain the value of cfi for a rotating group. The radical of interest for this purpose is CH3CH2OCH2CH2CH2CH2OCHCH3. (3) The reconstructed stick spectra for this radical, along with the observed ESR spectra, are shown in Fig. 3. It is clear from Fig. 3 that Radical (3) and at least one other radical are present in the x-irradiated 1,4- diethoxybutane-urea compound. From the magnitude of the splittings of the z-orientation spectrum, the second radical is evidently CH3CH2OCH2CH2CH2CHOCH2CH3. (4) No further investigation of Radical (4) was undertaken because, for our purposes, it is essentially equivalent to Radical (2). The 273°K ESR spectra for all of the above ether radicals are isotropic with respect to rotations of the magnetic field in the xy plane and are anisotropic with respect to other rotations of the magnetic field (this is characteristic of included radicals). The g value is also very nearly isotropic. The g values measured with the magnetic field parallel and perpendicular to the 2 axis of Radicals (l)-(3) are 2.0040±0.0004 and 2.0030±0.0003, respectively. All of the radicals ob served were stable for several hours at 273CK. If the crystals were allowed to warm up to room temperature, however, the ESR signal disappeared in 15 to 30 min. XY 270°K 82°K -I IOOMc/sec Fig. 2. The ESR spectra of an x-irradiated methyl octyl ether-urea crystal. ESR OF RCHOR' ETHER RADICALS Table I. Proton hyperfine coupling constants. a~° 2653 Ether Radical a," of a/ a*„a a*/ Dibutyl ether CH3(CH,) 3OCH (CH2)2CHs 64 63 8.2 23.6 60.4 Methyl octyl ether CH3(CH,)6CHOCH3 63 63 8.5 24.3 60.4 1,4-diethoxybutane CH3CH20 (CH,)406HCH, 63 62 8.5 24.3 56.6 * a", a^, and aS are the a, (3, and f protoncoupling constants, respectively, and xy and z denote the spectra recorded with the magnetic field in the crystal line xy and i directions. The two 0 protons of the dibutyl ether radical or the methyl octyl ether radical are magnetically equivalent and the three (3protons of the 1,4-diethoxybutane radical are magnetically equivalent. b The coupling constantsreportedhereare the average values obtained from a minimum of three groups of four spectra, each group being obtained from a different crystal. All values are in units of megacycles per second. The limits of experimental error varied with the orientations of the crystal in the magnetic field and the accuracy was greatest in the xy orientation (where the differences The coupling-constant data for Radicals (l)-(3) are summarized in Table I. In addition to the ESR data obtained at 270°K, the methyl octyl ether-urea crystals were investigated over the temperature range from 290° to 40°K. There were no changes in either the line widths or the splittings over the range 290° to 240°K. Around 240°K the spectra began to show signs of broadening and the 82°K ESR lines are significantly broadened (Fig. 2). Below 80°K the spectral lines appeared to broaden slightly as the temperature was lowered, but the effect was not as pronounced. The over-all width of the 40°K ESR spectra increased ~10% over the 270°K value; this is consistent with a decrease in the amplitude of motion about the C2-Ci bonds as the temperature was lowered from 240° to 40°K. There were no rapid changes in the ESR spectra as the temperature was lowered (such as might be caused by a reorientation of the ether radicals) and all temperature effects were reversible. EXPERIMENTAL SPIN-DENSITY DISTRIBUTION a. From a-Proton Coupling-Constant Data It is immediately apparent from Table I that the values of the a-proton coupling constants for all three radicals are the same. Therefore, in addition to identi fying the radicals produced, some general conclusions may be reached regarding the unpaired spin distribution of this class of aliphatic ether radicals. Equation (4) of II will be useful in obtaining the isotropic component, Oq", from the experimental values of axya and a,". First, however, the effect of the dipolar interaction between the a proton and the spin density on the oxygen atom must be estimated. To accomplish this the unpaired spin density on Carbon Atom 2, pC, and the unpaired spin density on the oxygen atom, pOT, are assumed to between theo"and
i is similar to Eq. (5) except that in the CI case the
wavefunctions fa and fa are taken to have the full
ethylenic symmetry.19 That is
and
and
0i=(2)-Kl + ^)-1(xc+Xo) (9)
fa=(2)-l(l-S)~HXc-X0), (10)
where S is the atomic orbital overlap integral. The
functions fi and \j/2 are normalized and are rigorously
orthogonal. The CI wavefunction^Mo and Hamiltonian
3C, in this approximation are
*MO= Clipi~\~C2^2 (11)
JW Scored)T 3Ccore(2)-f"3Ccore(3)
+ (n2)-1+(r13)-1+(r23)-1. (12)
Using McConnell's definition of the spin-density
operator,20 the elements of the atomic orbital spin-
density matrix may be determined from^. In terms of
' 19 R. G. Parr, Quantum Theory ofMolecular Electronic Structure(W. A. Benjamin, Inc., New York, 1963).
20 H. M. McConnell, J. Chem. Phys. 28, 1188 (1958).
2656 HAYES GRIFFITH
the coefficients Ci and C2 these elements are
(13)
pO-=iC22(l+5)-1+iCi2(l-^)-1-CiC2(l-^)-»,
(14)
PCO'=pOC-=|C22(l+^)-1-|C12(l-5)-1, (15)
and the spin-density function p"(x, y, z) is
p*(*i y,z)=pCT|xc|2
+PCO'(xc*Xo+Xo*Xc) +pO* | xo |2. (16)
The coefficients Ci and C2 are determined by the
usual variational procedure. The matrix elements of
theresulting 2X2secular determinant may beexpanded
in terms of the AO's in much the same way as for
ethylene.19 The estimation of the core integrals of the
ether radical deserves some elaboration. The core
Hamiltonian is
3Ccore(0 " - |V(<)2-f- I/C2(.-)++ U0(i)+ ++ EW
+ cTc3(,-)0+#Ha(,-)0, (I7)
where U+ (or U++) and U° denote the potentials
due to charged and neutral atoms of the core. The
approximate eigenvalue equation21
[—2V(,)2+ cTo(.')+]xo(o = UooXou (18)
and the parameter, /3, introduced by Pariser and Parr23
is
= /3rore-(S/2)(acrore+<*o°°re) m)
fi~ (1-S2)
Fortunately the spin-density distribution depends
only on the difference between cto00™ and ac°°re and
not on the individual core integrals. This tends to
reduce the errors involved in the values of U00 and
Ucc andrenders the method ofevaluating the Coulomb
andneutral-penetration integrals less critical. Initially,
we take the orbital energies to be the negative of the
valence-state ionization potentials; Uoo = —Io=—17.3
eV and L7Cc=-/c=-H.4 eV.16'24 The Coulomb inte
grals obtained by the method of Pariser and Parr with
Zc=3.25, Z0=4.55, and Rco= 1-35 A, are (CC | CC) =
10.8 eV, (00 | 00) = 14.7 eV, and (CC | 00) = 8.2 eV.16
The neutral penetration integrals {p: qq) were calculated
by standard methods21'26 using Slater orbitals. The
contributions of these penetration integrals to ao°°re
and acWK were found to be the same within~0.3 eV,26
and therefore they were not included in the spin-
density calculations. If differential overlap isneglected,
these are all of the quantities (other than /3) entering
into the expressions for the spin-density matrix ele
ments. The appropriate value of /3 for this radical is
not available, but the values —1.5,-2.5, and —3.5 eV
span what might be considered a reasonable range.
From Eqs. (13), (14), and (15), (pC, pCO', P0*)
are (0.97, 0.16, 0.03), (0.93, 0.25, 0.07), and
(0.89, 0.32, 0.11) for 13=-1.5, -2.5, and -3.5 eV,
respectively.
The calculated values for pC* are somewhat larger
than the experimental value. This is due, in part, to
the neglect of the effect of bonding on U0o and Z7cc.
For example, in benzene and other hydrocarbons Hush
and Pople27 found that the value of Ucc is~—9.5 eV,
which is significantly less negative than the valence-
state value -11.4 eV. The effect of bonding of the hy
drocarbon cr electrons is apparently to decrease the
stability of the x electrons.16 For the ether radical the
large dipole moments present an added complication.
The ir-electron dipole moment is apparently ~1.9±
0.5 D and is in the direction C~-0+. There is also
present a large cr-electron moment. In ether molecules
this moment is 1-2 D 28'29 and is almost certainly in
can be used to eliminate the kinetic integral providing
one assumes that22
Z7oW+ += £/oc-)+- fxoo-)2—7- • (19)J TijdVj
In Eq. (18) Uo0 is the usual valence-state orbital
energy. There is, of course, a similar eigenvalue equa
tion involving xc and Ucc In other aspects, the treat
ment of the ether core parallels that of the ethylene
molecule. The expressions obtained for the oxygencore
integral ao°°re and the carbon core integral accoie are
aocore= Uoo . (CC I00) - (00 | 00)
-(C2:00)-(C,:00), (20)
accore= Ucc _,(CC | 00)- (0:C2C2)
-(d:C2C2)-(Ha:C2C2), (21)
21M. Goeppert-Mayer and A. L. Sklar, J. Chem. Phys. 6,
645 (1938). , . .
22 Alternatively, one may consider an eigenvalue equation in
volving thesecond ionization potential ofoxygen. Thespin-density
distribution obtained by the two approaches is essentially the
same if the appropriate valence-state ionization potentials are
employed.
23 R. Pariser and R. G. Parr, J. Chem. Phys. 21, 466 (1953);
24 J. Parks and R. G. Parr, J. Chem. Phys. 32, 1657 (1960).
26 K. Ruedenberg, J. Chem. Phys. 34, 1861 .(1961).
26 To obtain the penetration integrals the C-H„ was taken to
be 1.08 A. The C1-C2 and C3-0 bonds were assumed to have
normal single-bond lengths; 1.54 and 1.42 A,respectively. Equally
complete cancellation of the penetration integrals occurs if a
Slater Z value of 4.90 is used.
27 N. S. Hush and J. A. Pople, Trans. Faraday Soc. 51, 600
28 C. P. Smyth, Dielectric Behavior and Structure (McGraw-
Hill Book Company, Inc., New York, 1955).
29 L. G. Wesson, Tables of Electric Dipole Moments (The tech
nology Press, Cambridge, Massachusetts, 1948).
ESR OF RCHOR' ETHER RADICALS 2657
where ^ne and ^oh designate the functions corresponding
to the neutral structure and the structure with charge
separation, respectively. The appropriate linear com
bination of these two functions, \?vb, would normally
be obtained from the Hamiltonian 3CT by the varia
tional method. However, in this case ^Vb is entirely
equivalent to SI'mo- By expanding fa and fa in terms of
fae and ^ch, ^mo becomes
* MO:
r Ci c,
IL2(1 + S)T12(1-S)J•]to
ft
L[2(l-•S)J [2(1+5)1Itoh. (25)
the direction C+-0~. The it and a dipole moments
therefore have opposite polarity. The 7r-electron moment
of the ether radical increases the electronegativity of
the oxygen atom and this should increase the