J. Phys. Chem. 1!993,97, 1856-1861
1856
A Nitrogen-15 NMR Study of cis-Azobenzene in the Solid State
Ronald D. Curtis, James W.Hilborn, Gang Wu,Michael D. Lumsden, Roderick E. Wasylisben,' and
James A. Pincock
Department of Chemistry, Dalhousie University, Halifax, Nova Scotia, Canada B3H 453
Received: October 14, 1992
Nitrogen-15 N M R spectra have been obtained for spinning and static powder samples of solid cis-azobenzene15N2 (1). The nitrogen N M R spectra obtained with magic angle spinning exhibited spinning rate-dependent
line shapes which are characteristic of dipolar coupled homonuclear spin pairs which have the same isotropic
chemical shifts but different orientations of their shift tensors relative to the dipolar vector. Analysis of the
15N N M R spectrum of a static sample yielded the three principal components of the nitrogen chemical shift
tensors: 611 = 1006 ppm, 622 = 469 ppm, and 633 = 112 ppm. As well, comparison of observed and calculated
N M R line shapes indicated that the most shielded component, 633, is perpendicular to the C N N C plane, while
the least shielded component, 611,is in the C N N C plane and oriented approximately 15O with respect to the
N-N internuclear vector. The nitrogen chemical shift tensor in 1 is compared with those reported in transazobenzene and compounds containing the X-N==C fragment. The nitrogen chemical shift tensors in cisand trans-azobenzene are discussed in light of the results of ab initio shielding calculations on cis- and transdiazene.
Iatroductioa
-ry
The photochemical properties of the azobenzenes have been
the subject of numerous investigations in the pastla and have
provided some practical uses as dyes? as photochromic probes
for studying motional aspects of polymers,IOand as light-induced
switches which can make a polymer functional.IOMore recently,
the reversible cis-trans isomerism of azobenzenes has been
implicated in a novel "three-state" photoelectrochemical method
for ultrahigh-density data storage.11J2
The two geometric forms of azobenzene, cis and trans
(thermodynamically stable), have been investigated by X-ray
diffractionl3-17 and by several solution spectroscopy
techniq~es.l~*~J'J-22
For example, l5N NMR studies of the two
forms of azobenzene in deuterated chloroform indicate that there
is only a 17.5 ppm difference in the isotropic chemical shifts in
this solvent (cis isomer at the higher frequency).'* However,
IGLO calculations performed by Schindler on the related cisand trmdiazenes, HN=NH, suggestan isotropic shift difference
in excess of 70 ppm (trans isomer at the higher frequency) and
that there are marked changes in the three principal components
of the '5N chemical shift tensors of the two isomers.23
To help understand these differences in shielding properties,
we decided to investigate the I5N NMR spectra of solid cisazobenzene-15N2,1, which can be prepared and isolated in pure
The influence of magic angle spinning (MAS) on the NMR
spectra of homonuclear spin pairs has been the topic of several
investigationsin the pa~t.=.~~35
For example, Mariq and Waugh
suggested that the MAS NMR line shape for such a spin system
can be predicted by average Hamiltonian theory.33 Later, Kubo
and Mchwelldemonstrated that thedescriptionof thelineshape
proposed by Mariq and Waugh was not completely accurate,
and they used Floquet Hamiltonian theory to calculate the MAS
NMR line shape of a homonuclear spin system reliably at all
spinning freq~encies.~~
Regardless, both approaches predict that
if two homonuclear spins are in an environment where they are
related by a 2-fold axis but do not have an inversion center, the
principal components (6,) of the two chemical shift tensors will
have the same magnitude but different orientations in the
molecular frame of reference. Upon magic angle spinning, the
distinct orientations of the two chemical shift tensors result in
instantaneous differences in the chemical shifts; together with
the direct homonuclear dipolar interaction, the MAS NMR line
shape will be a function of the spinning specd.33-35
For a static sample, the difference in the two orientations of
the chemical shift tensors of the homonuclear spin pair also has
a significant effect on the NMR line shape. In this case, the
NMR spectrum is sensitive to the relative magnitudes of the
dipolar coupling constant, R, and the chemical shift difference.
For instance, if the chemical shifts of the two spins are always
the same for all crystallite orientations, the homonuclear twospin system is referred to as an A2 spin pair. The static NMR
spectrum for a general A2 spin pair of I = I/2 nuclei consists of
two scaled and overlapping non-axially-symmetric shielding
patterns.2s32 If, however, the chemical shift difference between
the two spins is much larger than the direct dipolar coupling
between them, this constitutes an AX spin pair; the static NMR
spectrum for either the A or X nuclei (I= l/2) is identical to the
A2 case except for different amounts of scaling for the two
subspectra. Finally, when the chemical shift difference between
the two spins is not always larger than thedipolar coupling between
them, this constitutes an AB spin pair. In this case, Zilm and
Grant have demonstrated that the static NMR line shape is
intricately related to the type of spin system (Az, AB, or AX),28
and the powder NMR line shape may be a "mixture" of all three
types of spin systems. More precisely, along some directions the
d=b
1
form from doubly enriched tr~ns-azobenzene-~~N2.2~
Dipolar
chemical shift NMR studies on a static powder sample of 1allow
us to orient the nitrogen chemical shift tensor in the molecular
frame of reference and provide the three principal components
of the chemical shift tensor, 6,, ( i = 1, 2, 3).25-32 The nitrogen
chemical shift parameters of 1 will be compared to recent solidstate 15NNMR results for truns-azobenzene-'5N232as well as
with theoretical shielding calculations on two cis- and transdiazcnes.23
Correspondence should be addressed to this author.
0022-3654/93/2097-1 856$04.00/0
(6
1993 American Chemical Society
The Journal of Physical Chemistry, Vol. 97, No. 9, 1993 1857
cis-Azobenzene in the Solid State
field Bo must be calculated and summed. The two sets of
independent Euler angles, aics, flits, and TiCS(i = A, B), are
defined in Figure 1 and represent physical parameters which are
characteristic for the particular molecular fragment containing
the homonuclear spin pair.
It should be noted that since the dipolar interaction is axially
symmetric, the overall NMR line shape for a powder sample will
be invariant to any simultaneous rotation of the two chemical
shift tensors about the dipolar vector. Thus, the orientations of
the chemical shift tensors of a spin pair cannot be determined
unambiguously from the powder NMR line shape alone. In
practice, however, this uncertainty can usually be overcome by
considering the local symmetry of the molecule and/or results
from molecular orbital shielding calculations.
r
YY
Figure 1. Angles aCs,ps,and ycsrepresent the orientationof the three
principalcomponents of one of the nitrogen chemicalshift tensors relative
to the dipolar tensor. The N,N internuclear vector is along rzz. Since
the dipolartensor is axiallysymmetric,the choiceof r,,and ryyis arbitrary.
homonuclear spin pair may constitute an A2 spin system whereas
along other directions it may be an AB or AX spin system.
In order to account for the powder NMR line shape of a
general homonuclear spin / 2 pair, one has to use exact expressions
for the four transitions, vi, and their relative intensities, Pi, as
indicated in eqs 1:
where D = [ ( V A - V B )+~B2]1/2, A = J- R(3 cos2 8 - l), and B
= J + '/2R(3 cos28- 1). In eqs 1, V A and YB describe the shielding
contributions to the resonance frequencies of spin A and spin B,
respectively; J is the isotropic indirect spin-spin coupling in hertz,
R is the direct dipolar coupling constant in hertz,
(jk)/4.rr)(h/4.rr2)rArB(
AB-^) (in the absence of anisotropy in J
and librational averaging), and 8 is the angle between the applied
magnetic field and the dipolar vector, rAB.
The orientationsof the A and B spin chemical shielding tensors
can be referenced to the dipolar interaction tensor as indicated
in Figure 1. In this case, if the angles 8 and C#I represent the
orientationof the dipolar interaction tensor relativeto the applied
magnetic field, BO,then the chemical shielding contributions to
the powder NMR line shape can be evaluated from the equation
vi
= (YiB0/2T)[ 1 - (u;1 cos2xi + u'zzcos2 y
+ uf3cos2Zi)]
(2)
where (j = 1,2,3; i = A, B) are the principal components of
the chemical shielding tensors and cos Xi, COS ul:, and cos Zi are
the directional cosines which orient the applied magnetic field in
the principal axis system of the chemical shift tensor. These
angles can be expressed by a rotational transformation as shown
in eq 3. Thus, to calculate the NMR line shape for a general
homonuclear spin l / 2 pair in a powder sample, the frequencies
and intensitiesat all orientationsof 8and C#I relative to the magnetic
Experimental Section
The general procedure used to prepare pure cis-azobenzene15N2(1) was described by H a r t l e ~ .trans-Azoben~ene-~~N2
~~
was
procured from MSD Isotopes, Montreal. This sample (0.8 g)
was dissolved in 40 mL of distilled acetone, and the solution was
put into two water jacketed photolysis tubes. After irradiation
with a Hanovia 200-Wmedium-pressure mercury lamp for 8-12
h, the acetone solution was diluted with 60 mL of distilled water,
cooled in an ice-water bath for 10 min, and then filtered. The
filtrate was washed with 6.5 mL of light petroleum ether. The
remaining aqueous acetone filtrate was extracted several times
with distilled chloroform. The combined chloroform extracts
were dried with MgSO4, filtered, and then evaporated in vacuo.
This mixture was then separated on a preparative silica gel TLC
plate (Fisher) by elution with light petroleum ether; the top band
contains residual trans-azoben~ene-~~N2
whereas the lower band
is ci~-azobenzene-~~N2.
The lower band was removed and
extracted with cold chloroform (0 "C). The extract was dried
with MgS04 and filtered, and the chloroform was removed in
vacuo to yield pure ci~-azobenzene-~~N2
(vide infra). This
procedure was repeated three times to yield approximately 280
mg of pure ci~-azobenzene-~~N2.
The purity of the final sample
was confirmed by 13CNMR in deuterated chloroform.I8 In a
freezer, the cis-azobenzene-lSN2 sample was then loaded into a
7-mm zirconium oxide rotor for solid-state NMR studies.
Between solid-stateNMR experiments, the sample was stored in
a freezer to minimize thermal isomerization to the trans form.
Nitrogen-1 5 and carbon- 13 solid-state NMR spectra were
obtained at 20.3 and 50.23 MHz, respectively, on a Bruker MSL200 NMR spectrometer. All NMR spectra were acquired with
proton cross polarization (CP) and high-power proton decoupling.
Typical 7r/2 pulse widths for IH, I5N, and 13C were between 4.5
and 6.5 ks. Contact times were typically 10-12 ms. The
FLIPBACK pulse sequence36was used to acquire all NMR spectra
with recycle times of 3-5 min between the acquisition of successive
transients. Spinning frequencies for the CP/MAS NMR experiments were 1.5-4.0 kHz. The spectral width for all 15NNMR
experiments was 62.5 kHz with data file sizes of 4096 points.
Sensitivity enhancements of 5 and 200 Hz were applied to the
free induction decays (FIDs) prior to Fourier transformation.
The ISN NMR spectra were referenced to liquid NH3 (20 "C)
at 0 ppm by setting the observed ammonium signal of solid
ISNH4I5NO3
at 23.8 ppm. The 13CNMR spectra were referenced
with respect to TMS at 0 ppm by setting the observed methylene
carbon signal of solid adamantane at 38.57 ~ p m . ~ '
The I5N NMR spectrum of a static sample of 1was simulated
using a program which carries out calculations based on eqs 1-3
( J = 0). All calculations were performed on a 80386/387
microcomputer with a Fortran algorithm which incorporates the
erpolation schemeof Alderman et a1.38 The resulting calculated
VIR spectra were convoluted with a Gaussian line broadening
runction.
Self-consistent-field(SCF) and nitrogen chemical shielding
calculations on cis- and trans-diazene were performed on a Stellar
Curtis et al.
1858 The Journal of Physicul Chemistry, Vol. 97, No. 9, 1993
1
ow
am
609
400
200
0
bDm)
Figure 2. Magic angle spinning ISN NMR spectra of 1 obtained with
a spinning frequency of (a) 2035, (b) 2825, and (c) 3374 Hz. Note the
presence of peaks due to trow-azobcnzene in (a) and (c).
GS2500 vector processing computer. The necessary SCF results
were obtained with the Gaussian 90 ab initio program.39 The
molecular geometries were optimized at the 6-3 1G* level. The
nitrogen shielding tensorswere calculated with the RPAC program
of Bouman and Hansen which uses the LORG (localized orbitallocal origin) method.40 A 6-3 11G basis set was employed,
augmented with two sets of polarization functions on nitrogen
and one set on hydrogen.
550
I40
5xI
520
510
5w
(DDm)
Figure 3. Isotropic region of the IsN CP/MAS N M R spectrum observed
for a sample of 1 spinning a t a frequency of (a) 2035, (b) 2825, and (c)
3374 Hz (again note the presenceof the trans isomer in (a) and (c)). Note
the change in both the intensity and line shape of the isotropic 'SN NMR
signal as a function of the magic angle spinning frequency.
-
experiments on the same sample of 1;only a single resonance line
(u1/2
25 Hz) was observed for the ipso carbon (Si, = 155.0
ppm). However, unlike the precursor from-ambenzene, the cisazobenzene molecule does not have an inversion center. This is
analogous to the situation reported by Kubo and McDowell for
the two 31Pnuclei of Na2P207-10H2035 and by Challoner et al.
Result6 and Discussion
in 1,2-bis(2,4,6-tri-terr-butylphenyl)diphosphene.~~
Also,Power
and Wasylishenrecentlyreported a similar effect inci~-Pt(PEt~)~Magic Angle SpinningNMR Spectn. Representative I5NCP/
C12.42 Therefore, we conclude that the spinning rate-dependent
MAS NMR spectra of 1 are shown in Figures 2 and 3. The
spectra in Figure 3 are the expanded isotropic regions of the l5N
I5NNMR line shapes observed for 1result from the fact that the
NMR spectra shown in Figure 2. The l5N NMR spectrum in
two dipolar coupled l5N nuclei have chemical shift tensors that
are identical except for their relative orientations in the molecular
Figure 2b corresponds to that of a freshly prepared sample of 1
withnoindicationof any trum-ambenzene-15N2 ( 6 ~=, 511~ p m ) ~ frameofreference.
~
It shouldbenoted that weobservednosplitting
of the 15NNMR signals for the small amount of trans-ambenzeneas is apparent in Figure 2a,c. The truns-a~obenzene-~~N2
NMR
signals appeared in the samples only after the rotors had been
15N2present in some samples (Figures 2 and 3). Extensions of
used for severalNMR experimentsbut could alwaysbe completely
the average Hamiltonian theory and Floquet theory indicate that
removed by preparative TLC. Immediately obvious from the
as the MAS spinning frequency becomes much larger than the
ISN NMR spectra in Figures 2 and 3 is a splitting of the 15N
chemical shift anisotropy, the 15N NMR line shape will collapse
resonance lines of 1. From Figure 3 it is clear that the shape and
into a singlet.
splitting of the isotropic l5N NMR signal of 1vary as a function
The isotropic chemical shift determined for the solid sample
of the spinning frequency. For instance, at a spinning frequency
of 1, bi, = 528.8 ppm, comparesvery well to the analogous values
of 2035 Hz,the splitting of the isotropic l5N peaks is in excess
reported for 1dissolved in CDCl3 and (CD3)2S0solutions.18Note
of 100 Hz. As the spinning frequency is increased, the splitting
that the difference observed between the ISN isotropic chemical
gets smaller, being approximately 65 Hz at 3374 Hz. Also,from
shifts for solid cis- and rruns-azobenzene-15N2 (-18 ppm) (Table
Figure 2a, it is clear that the line shape of the different order
I) is very similar to the analogous value determined for these two
sidebands varies.
isomers in CDClj, 17.5 ppm.'8 This observation suggests that
X-ray diffraction results for cis-azobenzene indicate that
the structural and conformational integrity of both isomers of
crystals obtained from petroleum ether are orthorhombic with
ambenzene are maintained in CDCl3solution. From the number
four equivalent molecules per unit cell,15 Each molecule is
of spinning sidebands in Figure 2, it is clear that the nitrogen
nonplanar with each phenyl ring twisted by 53O about the CN
chemical shift anisotropy of 1is quite large. This property made
bond relative to the planar N-N-Ci,
fragment. However,
the aquisitionof a 15N NMR spectrum of a static powder sample
each molecule retains a C2 symmetry axis perpendicular to the
of 1 difficult.
N=N bond. Thus, the splittings observed in the ISN CP/MAS
The Static 15N NMR Spectrum. The 15N NMR spectrum
NMR spectra of 1 are not a result of crystallographic nonequivobserved for a freshly prepared static powder sample of 1is shown
alence. This conclusion was supported by I3C CP/MAS NMR
in Figure 4a; the simulated spectrum is shown in Figure 4b. The
The Journal of Physical Chemistry, Vol. 97, No. 9, 1993 1859
cis-Azobenzene in the Solid State
I
\
I
IL
a T
N-WJ
1200
900
600
PPM
300
0
Figure 4. (a) ISNNMR spectrum obtained experimentally for a static
powder sample of I. (b) Computer simulation of the powder spectrum
based on eqs 1-3.
parameters used for calculating the NMR spectrum in Figure 4b
are 611= 1006 ppm, 622 = 469 ppm, 633 = 112 ppm, aiCS= pics
= 90' ( i = 1,2), and ylCS = -yza = 15O. The estimated errors
in the three principal components are f 2 ppm and f4O for the
three Euler angles. Theexperimentalparameters aresummarized
in Table I together with the previous solid-state NMR results for
trans-azoben~ene-~~N2.~~
Also included are chemical shielding
parameters obtained using ab initio shieldingcalculations for the
cis and trans isomers of two simple model compounds.23
The N,N bond length reported for cis-azobenzene from X-ray
diffraction experiments,r N N = 1.253 A, was used tocalculate the
direct dipolar coupling constant, R = 628 Hz.I5 Including this
value in the calculation of the static I5N NMR spectrum, the
dipolar splitting at 633 in the experimental l5N static NMR
spectrumcould only be reproduced using pia = 90°. Interestingly,
the 833 region of the experimental static I5N NMR spectrum
could be precisely simulated as an A2 spin system with R = 628
Hz. This result provides strong evidence that the orientation of
6j3 is perpendicular to the N,N internuclear vector for both
nitrogens of 1 since any deviation of pics from 90° leads to
observable effects on the 633 region of the NMR spectrum (see
q s 1-3).
For other regions of the static I5N NMR spectrum in Figure
4a, AB character is apparent. For instance, it can be seen very
clearly in Figure 4a that there are three distinct shoulders at the
6' I end of the NMR spectrum. As noted earlier, for either an
A2 or an AX spin pair there would only be two such shoulders
at the d I 1 region. In addition, there is a small peak to high
frequency of the two larger peaks in the 622 region. Again, this
extra peak results from the AB character of this homonuclear
spin system. The variations of the 61I and 622 regions of the I5N
static NMR spectrum as a function of the angle ycs (ylcs =
'y2CS) are shown for several theoretical AB NMR spectra in
Figure 5 ( a f S = pics = 90°, i = 1,2). Note that the line shape
at 633 is independent of the angle yiCS,which is consistent with
eq3 provided that pics = 90'. Furthermore, the theoreticalNMR
spectrum where yla = -yfS = 15' represents the closest fit to
the experimental 15N static NMR line shape in Figure 4a; the
simulation in Figure 4b corresponds to the theoretical 15Nstatic
NMR spectrum which has been convoluted with a Gaussian line
broadening to accurately represent the observed NMR line shape
in Figure 4a. Thus, the two orientations of the nitrogen chemical
shift tensors of 1which are compatible with ulcs= uzCs = 90°,
p l C S = p 2 C S = 9 0 0 , a n d y l C S ~ - - y 2 ~ =15' aredepictedinFigure
6.
The angle of /P= 90 f 4' determined for the nitrogens of
1 indicates that the 633 are oriented perpendicular to the N,N
dipolar vector (see Figure 1). This conclusion is consistent with
previous ~xperimenta1~~-32.~3.~~
and theor~tical23,~~
results on other
molecular fragments containing dicoordinate nitrogen involved
in double bonds. Earlier ISN NMR experiments on tramazobenzene-'5N2,32
benzylidcneaniliie,~and severaloximes43 have
indicated that the intermediate principal component, 622, in these
systems is oriented in the approximate direction bisecting the
XNY angle (i.e., where a nitrogen lone pair resides). Moreover,
previous theoretical shielding calculations on the geometric
isomers of several diazenes23and on other similar moieties" predict
an orientation placing 622 in the approximate direction of the
nitrogen lone pair (Le., ycs = 30'). Also, the LORG calculations
of the nitrogen shielding tensor of cis-diazene suggest that the
value for ycsis 24'. Thus, we conclude that the orientations of
the two 15N chemical shift tensors indicated in Figure 6a are a
reasonabledescription of theshielding environmentsin 1. Clearly,
the two nitrogen chemical shift tensors of 1are related by a 2-fold
axis; however, as predicted from the 15N CP/MAS NMR
experiments on 1, the two chemical shift tensors are obviously
not related by an inversion center.
The fact that we can calculate a static 15N NMR spectrum of
1 in close agreement with the experimental NMR spectrum
(Figure 4) by using a value for the dipolar coupling constant
based on the X-ray derived N,N bond length (R = 628 Hz)
provides two fundamental pieces of information. First, the
contributions of anisotropic indirect spin-spin interactions, W ,
to the ISNNMR spectrum of 1are negligible, as the effect of W
is to modify the effective dipolar interaction according to the
expression Rem = R - W/3. Given that the absolute value of
1J(15N,1SN)i,is typically 10-20 Hz for the N=N fragment,N
any anisotropy in the indirect spin-spin couplingtensor is expected
to be small. Second, the two I5N nuclei of 1 must be relatively
free of librational motions since any motion will average the direct
dipolar interaction and lead to smaller values of &m.47-)9 This
is in contrast to trans-azobenzene which, according to X-ray
diffraction experiments, has two different sites for the transazobenzene molecule within the lattice, one of which is significantly disordered (site II).17 This disorder was reflected in the
reported nitrogen chemical shift parameters of trans-azobenzene
as indicated in Table I.32 Thus, the nitrogen chemical shift
parameters for site I of trans-azoben~ene-~~N2
will be used for
further comparison with the corresponding values determined
for the nitrogen nuclei of 1.
Discussion. A comparison between the experimental nitrogen
shielding tensors of cis- and trans-azobenzene and the theoretical
shielding calculations on the diazenes shows several interesting
features. First, the most shielded principal components, 633, of
cis- and trans-azoben~ene-~~N2
are very similar, 112 and 109
ppm, respectively. This result was reproduced by the theoretical
calculations, although the calculated magnitudes are slightly
smaller. As well, both experiment and theory predict the
orientation of 833 to be perpendicular to the N,N double bond.
Because paramagnetic shielding results from induced electronic
circulations in a plane perpendicular to the applied magnetic
field,50 this component should bc related to u u* type mixing.
Furthermore, because the paramagnetic shielding is inversely
related to the excitation energy of the field-induced transitions,
it is clear that given the large energy separations involved in a
transition such as u -. u*, small paramagnetic shielding will
result, hence explaining the Orientation of 633. Second, the
magnitudes of 611in cis- and trans-azobenzene, 1006 and 1034
ppm, respectively, are very similar whereas they differ by over
400 ppm in the diazene systems. Also, the magnitudes of 611
calculated for the diazenes are consistently greater than in the
azobenzene system. Given that bl I is oriented perpendicular to
both the T electron manifold and the location of the nitrogen lone
pair, it is reasonable to conclude that this component will be
correlated with the n -.T* transition in these species, which is
low lying in energy due to the nonbonding nature of the nitrogen
lone pair electrons. Since A&.,
was reported to be 22 500
cm-l for trons-azobenzeneand 22 900 cm-l for cis-azobenzene,Zl
-
Curtis et al.
1860 The Journal of Physical Chemistry, Vol. 97, No. 9, 1993
TABLE I: Expcrimeotrl .ad Tbeoreticrl l5N Cbemicrrl Shift Parameters for Several Compouods Containing Either a Cis or Tram
N,N Double Bond..*(Value of acsIs 90°)
compound
61I
g22
633
6,w
P,deg
yes, deg
1006
1034
lo05
1171
1121
1555
1421
1197
1597
cis-azobenzene
rruns-azobenzene (I)
rruns-azobenzenc (11)
cis-diazenc'
cis-diazencd
rruns-diazeneC
trans-diazencd
cis-dimethyldiazene
?runs-dimethyldiazene
112
109
125
469
39 1
400
665
556
515
450
65 1
506
70
64
57
52
53
60
528.8
511.0
510.0
634.6
580.3
708.6
641.3
633.7
721.0
90
83
78
15
37
46
e
e
90
24
e
e
90
40
e
e
e
e
a Errors in the three principal components are *2 ppm and i4O for the three Euler angles. Values for truns-azobenzene-15N2 are from ref 32.
Values taken from ref 23. Converted to chemical shifts according to the expressions5ai, = 244.6 - Uii. Calculated in this lab. Not specified precisely;
however, ref 23 suggests that ,fFs
is 90° and ycs is approximately 30'.
AA
90"
?-
jJ
I
I
1200
800
I
I
400
0
PPM
F l p e 5. Calculated static NMR spectra showing the variation of the
AB line shape at the 611 and 622 regions of the NMR spectrum as a
function of the angle ycs (ylcs = -y2CS, a i C S = BiCS = 90°, i = 1.2).
a
b
F i p e 6. Two possible orientations of 61 I and 622 for 1. Note that ycs
(15O) corresponds to the angle between 611 and the N,N internuclear
vector; 1333 is perpendicular to the N,N internuclear vector and is not
shown.
the small change in SII observed experimentallyis not surprising.
The large differences predicted for 61I of cis- and rrans-diazene
are in agreement with theoretical calculationsreported by Galasso
which showed that AE,,,. for cis- and trans-diazene differs by
morethan 2000cm-l.5l The fact that theleastshieldedcomponent
in thediazenes is calculated to be greater than that in the respective
azobenzenescan, in part, be related to differencesin the electronic
structure of these two systems. Also, the calculations correspond
to a rigid, isolated molecule in the gas phase, whereas intermolecular interactions and/or motional averaging will be reflected
in the experimental results. Furthermore, it has been previously
reported that electron correlation effects need to be considered
to accurately calculate nitrogen shielding tensors when the
nitrogen atom is in an unsaturated bonding environment.23-52Since
electron correlation increases the calculated chemical shielding,
its inclusion may bring the calculated results for SIIinto closer
agreement with the experimental results for the azobenzene
system. Therefore, we conclude that the nitrogen shielding in
the diazene system is calculated to be too paramagnetic in the
Hartree-Fock limit (i.e., using a singleSlater determinantal wave
function). Finally, the orientation of 8 2 2 in the azobenzenesystem
lies approximately along the direction bisecting the NNC bond
angle (where a nitrogen lone pair would reside). This result is
also supported by the theoretical calculations. The agreement
between experiment and theory provides strong evidencethat the
orientations of the two nitrogen chemical shift tensors of cisazobenzene are as indicated in Figure 6a.
In closing, we would like to comment on the deviation between
our calculated results for the nitrogen shielding in diazene and
those of Schindler.23 In both the cis and trans isomers, our
calculated chemical shifts are lower and in closer agreement with
the azobenzene system, leading one to believe that our LORG
calculations have done a better job in determining the nitrogen
paramagnetic shielding in the diazenes. This is quite surprising
considering that the numbers quoted in Table I for the IGLO
calculations were obtained with a large basis set, [41111111,
2 11111,111 for nitrogen and [3 111,111 for hydrogen, which was
shown to be saturated by performing calculationswith a different
common originof thevector potential. In our LORG calculations,
wehaveemployeda6-31lG(2D,lP)basisset ([6311,311,11] for
nitrogen and [311,1] for hydrogen) which, although not as
extensive as Schindler's, is believed to be of near Hartree-Fock
quality. Therefore, we conclude that the deviation between our
results and the IGLO results is due predominantly to geometry
effects. Our optimized N-N bond lengths (trans, 1.210 A; cis,
1.212 A) differ from those used by Schmdler on average by 4.04
A.53 As the shielding derivative with respect to the N-N bond
length in diazene was previously calculated by Chesnutand Wright
to be quite large, -1593 ppm A-',54 differences between the two
calculations are expected.
ConclusioM
Nitrogen-15 CP/MAS and static solid-state NMR spectra of
thermally unstable ci~-azobenzene-~5N2
have been examined. The
MAS I5N NMR spectra exhibit spinning rate-dependent line
shapes which are analogous to those observed for other homonuclear dipolar coupled spin pairs. In general, such systems consist
of two spins which have the same average chemical shift; however,
at some orientations in the magnetic field their resonance
frequencies differ. In the case of 1, the resonance frequencies of
the two dipolar coupled nitrogen nuclei will differ for all
orientations of the molecule in the magnetic field except when
the CNNC fragment is perpendicular to the applied magnetic
field. As an aside, it should be emphasized that the observation
of a doublet in MAS spectra of "equiva1ent" spin pairs could be
interpreted erroneously as arising from crystallographic nonequivalence if spectra are not carefully examined as a function
cis-Azobenzene in the Solid State
of spinning rate. The low-frequency edge of the line shape for
a static powder sampk of 1 exhibits A2 character while other
regions show features characteristic of AB spectra. Analysis of
the dipolar chemical shift spectrum yields the principal components of the shift tensor and its orientation in the molecular axis
system.
Acknowledgment. We thank NSERC for financial support in
the form of a postgraduate scholarship (R.D.C., M.D.L.) and
equipment and operating grants (R.E.W., J.P.P.). We acknowledge DalhousieUniversityfor a I. W. Killam Fellowship (R.D.C.)
and W.C.SumnerFellowship(R.D.C.,
J.W.H.). Wealsothank
Dr. H. C. Jarrell at the National Research Council of Canada
for allowingus to use his NMRI software on a Sun 4 workstation.
Finally, we thank Mr. B. Millier for his assistance with the MSL200 NMR spectrometer.
References and Notes
(1) Gegiou, D.; Muszkat, K. A.; Fischer, E. 1.Am. Chem. Soc. 1968,90,
3907.
(2) Ljunggren, S.;Wettennark, G. Acru Chem. Scund. 1971,25, 1599.
(3) Brown, E. V.; Granneman, G. R. J. Am. Chem. Soc. 1975,97,621.
(4) Rau, H.; Lllddecke, E. J. Am. Chem. Soc. 1982, 104, 1616.
(5) Hofmann, H.-J.; Birner, P. J. Mol. Struct. 1977, 39, 145.
(6) Monti, S.;Orlandi, G.; Palmieri, P. Chem. Phys. 1982, 71, 87.
(7) Asano, T.; Yano, T.; Okada, T. J. Am. Chem. Soc. 1982,104,4900.
(8) Griffiths, J. Chem. Soc. Rev. 1972, 1, 481.
(9) Foris, A. In The Anulyricul Chemistry of Synthetic Dyes; Venkataraman. K., Ed.; Wiley: New York, 1977.
(10) Kumar, G. S.;Neckers, D. C. Chem. Reu. 1989, 89, 1915.
(1 1) Dagani, R. Chem. Eng. News 1990,68,6.
(12) Liu, Z. F.; Hashimoto, K.; Fujishima, A. Nurure 1990, 347. 658.
(13) Robertson, J. M. J . Chem. Soc. 1939, 232.
(14) Hampson, G. C.; Robertson, J. M. J . Chem. Soc. 1941,409.
(15) Mostad, A.; Rsmming, C. Acru Chem. Scand. 1971,25, 3561.
(16) Brown, C. J. Acru Crystullogr. 1966, 21, 146.
(17) Bouwstra. J. A.; Schouten, A.; Kroon, J. Acru Crystullogr. 1983,
c39,1121.
(18) LyEka, A. Collecr. Czech. Chem. Commun. 1982,47, 1112.
(19) Kroner, J.; Schneid, W.; Wibcrg, N.; Wrackmeyer, B.; Ziegleder, G.
J. Chem. Soc., Furuduy Trans. 2 1978, 74, 1909.
(20) Lambert, J. B.; Netzel, D. A. J. Mugn. Reson. 1977, 25, 531.
(21) Forber, C. L.; Kelusky, E. C.; Bunce, N. J.; Zerner. M. C. J . Am.
Chem. Soc. 1985, 107, 5884.
(22) Bisle, H.; Rau, H. Chem. Phys. Lett. 1975, 31, 264.
(23) Schindler, M. J . Am. Chem. Soc. 1987, 109,5950.
(24) Hartley, G. S.Nature 1937, 140, 281.
(25) VanderHart, D. L.; Gutowsky, H. S.J. Chem. Phys. 1968,49,261.
The Journal of Physical Chemistry, Vol. 97, No.9, 1993 1861
(26) van Willigen, H.; Griffin, R. G.; Haberkorn, R. A. J . Chem. Phys.
1977, 67, 5855.
(27) Linder, M.; Hbhcncr, A.; Ernst. R. R. J. Chem.Phys. 1980,73,4959.
(28) Zilm, K. W.; Grant, D. M. J . Am. Chem. Soc. 1981,103, 2913.
(29) Valentine, K. G.; Rockwell, A. L.; Gierasch, L. M.; Opella, S.J. J .
Mugn. Reson. 1987, 73, 519.
(30) Curtis, R. D.; Penner, G. H.; Power, W. P.; Wasylishen, R. E. J .
Phys. Chem. 1990,94,4000.
(31) Power, W. P.; Wasylishen, R. E. In Annuul Reports on NMR
Spectroscopy; Webb, G. A., Ed.; Academic Press: London, 1991; Vol. 23, pp
1-84.
(32) Wasylishen, R. E.; Power, W. P.; Penner, G. H.; Curtis, R. D. Cun.
J . Chem. 1989, 67, 1219.
(33) Mariq, M. M.; Waugh, J. S . J. Chem. Phys. 1979, 70, 3300.
(34) Levitt, M. H.; Raleigh, D. P.; Creuzet, F.;Griffin, R. G. J. Chem.
Phys. 1990, 92, 6347.
(35) Kubo. A.; McDowell, C. A. J. Chem. Phys. 1990, 92, 7156.
(36) Tegenfeldt, J.; Haeberlen, U.J . Mugn. Reson. 1979, 36, 453.
(37) Earl, W. L.; VanderHart, D. L. J . Mugn. Reson. 1982, 48, 35.
(38) Alderman, D. W.; Solum, M. S.;Grant, D. M. J . Chem. Phys. 1986,
84, 3717.
(39) Frisch, M.J.; Head-Gordon, M.; Trucks, G. W.; Foresman, J. B.;
Schlegel, H. B.; Raghavachari, K.; Robb, M. A.; Binkley, J. S.;Gonzalez, C.;
DeFrees, D. J.; Fox, D. J.; Whiteside, R. A.; Seeger, R.; Melius, C. F.; Baker,
J.; Martin, R. L.; Kahn, L. R.; Stewart, J. J. P.; Topiol, S.;Pople, J. A.
GAUSSIAN 90,Revision H; Gaussian, Inc.: Pittsburgh, PA, 1990.
(40) (a) Bouman, T. D.; Hansen, Aa. E. RPAC Moleculur Properties
Puckuge, Version 8.6; Southern Illinois University at Edwardsville, 1990. (b)
Bouman, T. D.; Hansen, Aa. E. In?. J . Quantum Chem., Quantum Chem.
Symp. 1989, 23, 381.
(41) Challoner, R.; N a h i , T.; McDowell, C. A. J . Chem. Phys. 1991,94,
7038.
142) Power. W. P.: Wasvlishen. R. E. Inore. Chem. 1992. 31. 2176.
(43j Wasylishen, R.E.;$enner,G. H.;PowG, W. P.;Curtis, R:D.J. Am.
Chem. Soc. 1989,111,6082.
(44) Ribas Prado, F.; Giessner-Prettre, C. J . Mugn. Reson. 1982,47, 103.
(45) Mason, J. In Multinuclear N M R Mason, J., Ed.; Plenum Press:
New York, 1987; p 335.
(46) Witanowski, M.; Stefaniak, L.; Webb, 0. A. In Annuul Reports on
NMR Spectroscopy; Webb, G. A., Ed.; Academic Press: London, 1981; Vol.
1 1B, p 456.
(47) Das. T. P.; Hahn, E. L. In Solid Srure Physics; Seitz, F., Turnbull,
D., Eds.; Academic Press: New York, 1958; Suppl. 1.
(48) Henry, E. R.; Szabo, A. J. Chem. Phys. 1985,82,4753.
(49) Millar, J. M.; Thayer, A. M.; Zax, D. B.; Pines, A. J . Am. Chem.
Soc. 1986, 108, 5113.
(50) (a) Ramsey, N. F.Phys. Reo. 1950,78,699. (b) Ibid. 1951,83,540.
(c) Ibid. 1952, 86, 243.
(51) Galasso. V. Chem. Phys. 1984,83,407.
(52) Bouman, T. D.; Hansen, Aa. E. Chem. Phys. Lett. 1992, 197, 59.
(53) Parsons, C. A.; Dykstra. C. E. J. Chem. Phys. 1979. 71, 3025.
(54) Chesnut, D. B.; Wright, D. W. J. Compur. Chem. 1991, 12, 546.
(55) Jameson, C. J.; Jameson, A. K.; Oppusunggu, D.; Wille, S.;Burrell,
P. M.; Mason, J. J . Chem. Phys. 1981, 74, 81.