Journal of Molecular Spectroscopy 238 (2006) 158–167
www.elsevier.com/locate/jms
Infrared spectroscopic investigation of higher diamondoids
Jos Oomens a,*, Nick Polfer a, Olivier Pirali a, Yuko Ueno b,c, Roha Maboudian b,
Paul W. May d, Jacob Filik d, Jeremy E. Dahl e, Shenggao Liu e, Robert M.K. Carlson
a
e
FOM Institute for Plasma Physics ‘‘Rijnhuizen’’, Edisonbaan 14, 3439 MN Nieuwegein, The Netherlands
b
Department of Chemical Engineering, University of California Berkeley, Berkeley, CA 94720, USA
c
NTT Microsystem Integration Laboratories, Atsugi, Kanagawa 243-0198, Japan
d
School of Chemistry, University of Bristol, Bristol BS8 1TS, UK
e
Molecular Diamond Technology, Chevron Technology Ventures, Richmond, CA 94802, USA
Received 11 April 2006
Available online 5 May 2006
Abstract
Attenuated total reflection Fourier transform infrared spectra are recorded for a number of diamond molecules (or higher diamondoids)
in the spectral range between 650 and 3000 cm 1. Molecules studied are diamantane, triamantane, [121]tetramantane, [123]tetramantane,
[1(2,3)4]pentamantane, [12(3)4]pentamantane, 3-methyl-[1(2,3)4]pentamantane, and [12312]hexamantane (cyclohexamantane). Spectral
trends are clearly recognized throughout the spectra revealing the general fingerprint of this class of molecules. In general, the spectra show
good agreement with density functional theory calculations at the B3LYP/D95(d,p) level of theory.
Ó 2006 Elsevier Inc. All rights reserved.
Keywords: Diamondoids; Infrared; Spectroscopy
1. Introduction
Diamondoid molecules consist of a diamond-like carbon cage, where all carbon atoms are sp3 hybridised, and
dangling bonds at the edges of the systems are terminated
with hydrogen atoms. The smallest member of the family,
adamantane (C10H16) is made up of the central cage of a
single diamond unit cell and was first synthesized by carbocation equilibration in 1957 by von Rague Schleyer [1]. The
next members of the family are diamantane and triamantane possessing two and three diamond cages, respectively.
For higher diamondoids, starting with tetramantane (four
cages), structural isomers can be formed.
Diamondoids show remarkable rigidity, strength, and
thermodynamic stability, as well as interesting electronic
properties [2], which may be of use in chemical, polymer,
and pharmaceutical applications, as well as in
nanotechnology.
*
Corresponding author. Fax: +31 30 603 12 04.
E-mail address: joso@huygens.rijnh.nl (J. Oomens).
0022-2852/$ - see front matter Ó 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.jms.2006.05.001
Moreover, there has been substantial interest in the
spectroscopic properties of this class of molecules from
an astrophysical view-point. Not only have nanometersized ‘diamond-like’ molecules been found in meteorites
[3], the occurrence of these compounds as isolated gasphase molecules in the interstellar medium, has also been
suggested, based on the observation of infrared absorption
[4] and emission [5,6] bands around 3.5 lm (2880 cm 1) in
the spectra of protostars as well as the post-AGB object
HR 4049 [6]. This band has been assigned to the tertiary
sp3 carbon (i.e., carbon bound to three carbon atoms)
C–H stretching mode [7], but other explanations have also
been put forward [8,9]. Further evidence for the existence
of interstellar diamondoid molecules could be obtained
by comparing the complete vibrational spectra of such
diamondoids with astronomical data [10].
Although the very smallest members of the diamondoid
family can now be readily synthesized, this is not the case
for species larger than triamantane. Hence, in spite of the
wide interest in diamondoids, relatively few studies have
been reported on the characterization of these molecules.
J. Oomens et al. / Journal of Molecular Spectroscopy 238 (2006) 158–167
A recent breakthrough in diamondoid research was the discovery that higher diamondoids can be isolated from petroleum, which was shown for species up to the size
undecamantane (11 cages) [11]. As a result, various analytical techniques including GC/MS and NMR have been
applied to identify and characterize these novel species [12].
From a spectroscopic point of view, only Raman spectra
of a selection of diamondoids with a size ranging from adamantane to [121321]heptamantane have been published
very recently [13]. In addition, theoretical methods have
been applied to predict the vibrational spectra of some of
these molecules. An extensive computational study of
adamantane and some of its ionic and dehydrogenated
Adamantane
C10H16
Diamantane
C14H20
Triamantane
C18H24
159
radical species was reported by Yan et al. [14]. For the adamantyl closed-shell cation, an experimental infrared spectrum was also obtained [15]. The Raman spectrum of
crystalline cyclohexamantane was interpreted with density
functional theory (DFT) calculations by Richardson
et al. [16]. An infrared spectroscopic study on deposited
diamondoid nanoparticles (5–350 nm in size) focusing on
the 3 lm spectral range was successfully interpreted with
DFT computations on individual molecular diamondoid
structures, giving insight into the size dependence of the
spectra [17].
To our knowledge, no infrared spectra have to date been
reported for individual higher diamondoid molecules.
Moreover, no direct comparison between theoretical and
experimental spectra for these molecules has been reported.
In this contribution, we present the first infrared spectra for
a number of higher diamondoids, and present an analysis
based on DFT computations of the vibrational spectra.
The systems studied are diamantane, triamantane,
[121]tetramantane, [123]tetramantane, [1(2,3)4]pentamantane, [12(3)4]pentamantane, and [12312]hexamantane
(cyclohexamantane), which are shown in Fig. 1. (Note that
we follow the nomenclature introduced in ref. [18].) In
addition, we present the infrared spectrum for the
[1(2,3)4]pentamantane molecule methylated at the position
of the arrow in Fig. 1 (3-methyl-[1(2,3)4]pentamantane).
2. Methods
[121]
P[123]
Tetramantanes
C22H28
[1(23)4]
[12(3)4]
Pentamantanes
C26H32
[12312]Hexamantane
C26H30
Fig. 1. Higher diamondoid structures considered in this work. Hydrogen
atoms have been omitted for clarity; all structures are fully saturated. The
smallest member of the family, adamantane, is shown for comparison. The
arrow indicates the carbon atom where by the substitution of a hydrogen
atom for a methyl group, the 3-methyl-[1(2,3)4]pentamantane is formed.
2.1. Experimental
Higher diamondoid species (size going up to undecamantane) have been isolated from petroleum oil by vacuum distillation, high-temperature pyrolytic destruction of
non-diamondoid species combined with chromatographic
techniques [11]. The size and shape separations are performed using high performance liquid chromatography in
ultrahigh purity. The molecules studied here are neutrals,
fully sp3 hybridized molecules with completely hydrogenated surfaces plus one mono-methylated species.
The attenuated total reflection Fourier transform infrared (ATR-FTIR) spectra of the diamondoid molecules
were recorded with the use of a Thermo Nicolet Avatar
370 FT-IR spectrometer and an ATR accessory with a
ZnSe ATR crystal. The spectra of [12(3)4]pentamantane
and 3-methyl-[1(2,3)4]pentamantane were recorded in Bristol (UK) at slightly higher resolution, using a Perkin-Elmer
‘Spectrum One’ FTIR spectrometer. The higher resolution
was required to better resolve the bands in the congested
spectra of these molecules. All spectra are obtained from
pure, powdered diamondoid samples, and measurements
were performed at room temperature. The intensity of
the ATR absorbance spectra are corrected for the penetration depth of the evanescent wave, which is proportional to
the wavelength.
One can notice that two enantiomers (P and M) exist for
the molecule [123]tetramantane. Since the light source used
160
J. Oomens et al. / Journal of Molecular Spectroscopy 238 (2006) 158–167
to obtain the ATR-FTIR spectra is not polarized, the spectrum recorded is not dependent on the enantiomer studied.
The structure of the P[123]tetramantane is given in the
Fig. 1, but the recorded spectrum corresponds to a racemic
mixture of the P and M enantiomers.
2.2. Computational
Geometry optimization and computation of the harmonic vibrational frequencies were performed using the
Becke3LYP hybrid density functional [19] and Dunning’s
double zeta basis set with polarization functions,
D95(d,p). This method has been shown to give reliable
results for numerous organic compounds, such as a large
variety of polycyclic aromatic hydrocarbons [20]. The computations are carried out at an IBM cluster1600 system
located at the SARA Supercomputer center in Amsterdam
using the Gaussian98 program package [21].
The high symmetry of many of the systems was conveniently used to reduce the computational cost of the calcuTable 1
Empirical scaling factors for DFT computed frequencies
Formula
Diamantane
Triamantane
[121]Tetramantane
[123]Tetramantane
[1(2,3)4]Pentamantane
[12(3)4]Pentamantane
[12312]Hexamantane
3-Methyl-[1(2,3)4]
pentamantane
C14H20
C18H24
C22H28
C22H28
C26H32
C26H32
C26H30
C27H34
Sym
D3d
C2v
C2h
C2
Td
Cs
D3d
C3v
Frequency scaling
Mid-IR
3 lm
0.9916
0.986
0.986
0.988
0.975
0.985
0.990
0.976
0.949
0.947
0.947
0.948
0.947
0.941
0.948
0.945
lations. All calculations converged and yielded no negative
frequencies. For comparison with the experimental spectra,
the calculated spectra have been convoluted with a 5 cm 1
FWHM Gaussian line profile (except for the spectra
recorded at higher resolution in Bristol, where a 2.5 cm 1
FWHM Gaussian convolution was used in the mid-infrared range).
Due to the generally substantial differences in anharmonicity for the CH stretching modes in the 3 lm spectral
range as compared to the modes in the mid-infrared range,
empirical scaling factors are commonly determined for the
two spectral ranges separately. In the 3 lm region, a scaling
of the computed frequencies was applied so that the most
intense peak matches exactly. In the mid-infrared range,
the scaling was (arbitrarily) chosen such that the strong
peak occurring around 1050–1150 cm 1 shows exact agreement. Thus, average scaling factors of 0.986 and 0.947 are
found for the mid-infrared and CH stretching regions (see
Table 1), respectively, which appears to be nicely in accord
with the overall value of 0.961 reported for the B3LYP
functional [22]. Moreover, the 3 lm scaling factor of
0.947 is very close to that of 0.946 reported by Chen
et al. [17] for computed harmonic frequencies of large diamondoids using the same functional but with the 6-31G(d)
basis set.
3. Results and discussion
The spectra are displayed in Figs. 2–9, compared
‘‘back-to-back’’ with the scaled computed spectra. As
expected, the spectra of the more symmetric species are
much simplified relative to those of the less symmetric
ones, even if they are bigger in size; a striking example
DIAMANTANE
60
3
30
0
0
-30
-3
-60
600
800
1000
1200
1400
1600 2700
wavenumber (cm-1)
Fig. 2. Infrared spectrum of diamantane.
2800
2900
3000
161
J. Oomens et al. / Journal of Molecular Spectroscopy 238 (2006) 158–167
TRIAMANTANE
2
30
0
0
-30
-2
600
800
1000
1200
1400
1600 2700
2800
2900
3000
2800
2900
3000
wavenumber (cm-1)
Fig. 3. Infrared spectrum of triamantane.
[121]TETRAMANTANE
4
50
2
0
0
-2
-50
-4
600
800
1000
1200
1400
1600 2700
wavenumber (cm-1)
Fig. 4. Infrared spectrum of [121]tetramantane.
is observed by comparing the spectra of [123]tetramantane (C2) with those of [1(2,3)4]pentamantane (Td) or
[12312]hexamantane (D3d). On the other hand, one can
also notice that the spectra of two different molecules
belonging to the same symmetry group can be quite
different. Compare for example the spectra of adamantane (see ref. [23]) and [1(2,3)4]pentamantane (both
Td symmetry) or diamantane and [12312]hexamantane
(both D3d symmetry). However, despite the dense spectrum for the lower symmetry species, the experimental-
versus-theoretical match in the mid-infrared frequency
range is usually acceptable. For instance, the spectra of
[123]tetramantane (Fig. 5) and [12(3)4]pentamantane
(Fig. 7), recorded at higher resolution because of the
highly congested mid-infrared spectrum, show good
agreement with theory.
At this point we could discuss the spectrum of each molecule individually, however, we have chosen to discuss the
cyclohexamantane spectrum in somewhat more detail and
then give a more general discussion of the spectra of the
162
J. Oomens et al. / Journal of Molecular Spectroscopy 238 (2006) 158–167
[123]TETRAMANTANE
40
2
20
1
0
0
-1
-20
-2
-40
600
800
1000
1200
1400
1600 2700
2800
2900
3000
2800
2900
3000
wavenumber (cm-1)
Fig. 5. Infrared spectrum of [123]tetramantane.
[1(2,3)4]PENTAMANTANE
1
10
0
0
-1
600
-10
800
1000
1200
1400
1600 2700
wavenumber (cm-1)
Fig. 6. Infrared spectrum of [1(2,3)4]pentamantane.
remaining molecules. Experimental peak listings and computed optimized structures for all molecules can be found
in the Supplementary material. The choice for hexamantane is not completely arbitrary since due to the high symmetry of the molecule, it represents a relatively simple and
clean spectrum. Moreover, this particular higher diamondoid seems to have received more attention than others
recently [12,16]. It is the largest species in the series investigated here. Finally, we briefly investigate the effect of
methylation on the spectrum by discussing the spectrum
of 3-methyl-[1(2,3)4]pentamantane in somewhat more
detail.
3.1. IR spectrum of cyclohexamantane
[12312]Hexamantane or cyclohexamantane belongs to
the D3d symmetry group and this high symmetry simplifies
the infrared spectrum considerably. Of the 162 vibrational
modes, only those of A2u and Eu symmetry are infrared
allowed. The Eu modes are doubly degenerate. From the
163
J. Oomens et al. / Journal of Molecular Spectroscopy 238 (2006) 158–167
[12(3)4]PENTAMANTANE
40
2
20
0
0
-20
-2
-40
600
800
1000
1200
1400
1600 2700
2800
2900
3000
wavenumber (cm-1)
Fig. 7. Infrared spectrum of [12(3)4]pentamantane. The resolution of the calculated spectrum is 2.5 cm 1.
[12312]HEXAMANTANE
3
50
0
0
-50
-3
600
800
1000
1200
1400
1600 2700
2800
2900
3000
wavenumber (cm-1)
Fig. 8. Infrared spectrum of [12312]hexamantane.
comparison of the experimental and theoretical spectra
shown in Fig. 8, it is quite straightforward to obtain an
assignment for the majority of the bands. Some difficulties
are encountered in the 1250–1400 cm 1 range, where more
bands, and with more intensity, are observed than calculated. In the CH stretching range around 2800–2900 cm 1, the
proximity of several very intense modes causes severe congestion in the experimental spectrum, which makes a secure
assignment very difficult. The proposed assignments are
listed in Table 2, where those in the above mentioned wavenumber ranges should be regarded as tentative at best.
Table 2 shows that using the scaling factors of Table 1
gives a deviation of less than 5 cm 1 for all bands except
for the bands in the 1250–1400 cm 1 range. In addition,
for the band observed at 1445 cm 1, assigned securely as a
CH2 scissoring mode, the deviation of more than 20 cm 1
appears to be anomalously large. In fact, two closely spaced
scissoring modes are predicted, one of Eu and one of A2u
164
J. Oomens et al. / Journal of Molecular Spectroscopy 238 (2006) 158–167
4 3-METHYL-[1(2,3)4]PENTAMANTANE
40
2
20
2700
2800
2900
3000
0
0
-20
-2
-40
-4
600
800
1000
1200
1400
1600 2700
2800
2900
3000
wavenumber (cm-1)
Fig. 9. Infrared spectrum of 3-methyl-[1(2,3)4]pentamantane. A comparison with the 3-lm spectrum of its non-methylated counterpart is shown in the
inset.
Table 2
Infrared active modes of cyclohexamantane, experimental versus calculated
Calc freq (cm 1)
Intensity (km/mol)
Sym
Exp freq (cm 1)
Description
668.2
758.9
774.4
867.1
873.5
929.9
943.0
952.1
1027.9
1063.9
1073.0
1084.0
1118.2
1181.8
1210.3
1280.8
1285.8
1326.1
1366.8
1368.1
1477.7
1478.7
2834.6
2841.5
2853.2
2858.5
2860.4
2870.1
2877.9
2904.4
2905.3
0.2
0.2
4
0.1
0.02
2
2
1
1
4
19
0.1
0.1
0.7
1
0.1
2
6
2
0.3
13
13
8
150
17
2
36
293
418
62
176
Eu
A2u
Eu
A2u
Eu
Eu
Eu
A2u
Eu
Eu
A2u
Eu
Eu
A2u
Eu
Eu
Eu
A2u
Eu
A2u
Eu
A2u
Eu
A2u
Eu
A2u
Eu
A2u
Eu
Eu
A2u
665
760
775
865
874
931
939
—
1026
1058
1074
1088
—
1178
—
1276
1297
1311
1337
1352
1445
—
—
2847
—
—
2858
—
2874
—
2906
Deform
Deform
Deform
CH2 rock
Deform/CH2 rock
Deform/CH2 rock
CC str/CH bend
CC str/CH bend
CC str/CH bend
CH2 wag/CH bend
CH2 rock/CH bend
CH bend
CH2 twist/CH bend
CC str/CH bend
CH bend
CH bend/CH2 twist
CH bend
CH bend
CH bend
CH bend
CH2 scissor
CH2 scissor
CH str
CH str
CH str
CH str
CH str
CH str
CH str
CH str
CH str
character. The experimental band at 1445 cm 1 indeed
appears to be somewhat asymmetrically shaped suggesting
the presence of two unresolved bands. When the molecule
is viewed along the (1 0 0) crystal lattice plane, the modes correspond to asymmetric scissoring of the CH2 moieties on
diagonally opposing corners of the molecule.
J. Oomens et al. / Journal of Molecular Spectroscopy 238 (2006) 158–167
As impressive as the match for the frequencies is, the
intensity predictions are not all that excellent. Particularly
the relative intensities of the modes in the 1250–1400 cm 1
range are underestimated. Moreover, the intensity ratio
between the CH stretching range and the rest of the spectrum is not so well reproduced.
Analogous to for instance olefins and acetylenes, the
strongest CH stretching bands roughly correspond to collective asymmetric stretching modes, though now along
one of the crystal lattice planes. For instance in cyclohexamantane, the band calculated at 2877.9 cm 1 corresponds
to the asymmetric CH stretching band in the (1 1 0) lattice
plane, the band at 2905.3 cm 1 oscillates asymmetrically
in the (1 1 1) plane, and the 2870.1 and 2841.5 cm 1 bands
in the (1 0 0) lattice plane.
3.2. General discussion of diamondoid spectra
Upon global inspection of all spectra, one clearly
observes large similarities between them. Many parallels
with the cyclohexamantane spectrum can be observed,
although the lower symmetry of most other species yield
a richer vibrational pattern. The general appearance of
the spectra could be considered as the spectral fingerprint
of the diamondoid class of molecules. Inspecting the spectra more closely, four distinct groups of bands can be recognized in all diamondoid spectra. Visualization of the
calculated vibrational displacements allows one to roughly
assign a general mode description pertinent to the four
groups (see Table 3).
The modes around 2900 cm 1 (Group A) are of course
the CH stretching modes. Due to the very strong and overlapping bands and the broader line widths in the 3 lm
range, it is harder to recognize the quality of the match.
In some cases, it appears that the experimental bands are
more widely spaced than their computed counterparts. A
possible explanation could be strong anharmonic interactions between closely spaced modes of equal symmetry,
which all have very similar CH stretching character.
A very recognizable group of modes in all spectra, mostly appearing as a doublet around 1450 cm 1 is due to two
or more CH2 scissoring modes. These bands are clearly
identifiable throughout the spectra, mostly appearing as a
doublet. In [123]tetramantane, the low symmetry causes
further splitting into about four bands. In [12312]hexamantane, where only a single band is observed, the computation predicts three bands with non-vanishing intensity
Table 3
General assignment pertinent to all molecules studied in four distinct
spectral ranges
Group
Group
Group
Group
A
B
C
D
Region (cm 1)
General description
2900
1450
1000–1400
61000
CH stretch
CH2 scissor
CH2 rock, wag, twist
Skeletal deformation
165
within 1 cm 1. In all cases the bands are well reproduced
by theory, except for a consistent blue shift of the calculated frequencies of around 35 cm 1. A slightly smaller scaling factor of about 0.966 on average appears to be more
appropriate for these modes. Compared to the average
scaling factor used for the mid-infrared range of 0.986
(see Table 1) this amounts to a discrepancy of 2 %.
Despite this discrepancy, the assignment appears to be
fairly secure on account of the consistency of the shift
and the recurring pattern of these bands throughout the
spectra (see Figs. 2–9).
The bands falling roughly between 1000 and 1400 cm 1
have mostly CH bending and CH2 rocking, wagging, and
twisting character. Toward the red end of this group, some
CC stretching character is found. Finally, in the long wavelength part of the spectra, below 1000 cm 1, mostly skeletal
deformation modes are found and these modes have been
classified as belonging to Group D in Table 3. Comparing
the bands in Groups C and D to the DFT calculation, a
striking observation is the underestimation of their intensities (except some around 1050 cm 1) by the DFT calculations. The reason for this discrepancy, that appears to be
a general feature in all diamondoid spectra, is at present
unclear. A possible influence of the basis set used was
investigated for the [121]tetramantane molecule (see
Fig. 10). However, the underestimation of intensities in
Groups C and D is consistently reproduced upon using
basis sets ranging in size from 3-21G to cc-pVDZ.
As to the computed frequencies, one notices in Fig. 10
that the frequency scaling factor becomes closer to unity
as the basis set is increased, which is a well-known effect.
Interestingly, however, the CH2-scissoring modes (Group
B) appear to shift more than the other bands in the spectrum, such that the anomalous scaling factor mentioned
for these modes above, is almost completely corrected for
when using the cc-pVDZ basis set.
In conclusion, a global inspection of the spectra shows
that the match between theoretical and experimental frequencies is reasonable for most of the bands in the mid-infrared range. This does, however, not apply to the relative
intensities, which are not so well reproduced. Although
intensities were corrected for the wavelength dependency
of ATR, there remains a significant discrepancy (a factor
of 3 approximatively) between the experimental and computed intensity ratios for the mid-IR versus the 3 lm spectral ranges. Judging from the increased linewidth and the
more Lorentzian-like lineshape, it seems that this discrepancy results at least partly from saturation in the 3 lm
range. It is, however, also known that while DFT usually
provides a good prediction of the frequencies, intensities
are not so reliably predicted. In addition, interaction of
the molecules with their solid environment may affect the
intensities differently for different vibrational modes. In
fact, for a number of polyaromatic molecules, Joblin
et al. [24] found that the CH stretching modes were attenuated by a factor of 3 in the solid-phase with respect to the
gas-phase. Although such an effect would indeed explain
166
J. Oomens et al. / Journal of Molecular Spectroscopy 238 (2006) 158–167
3-21G
6-31G
6-31G(d,p)
D95(d,p)
cc-pVDZ
4
30
EXP
20
2
0
600
10
800
1000
1200
1400
0
1600 2700
2900
3100
Wavenumbers (cm-1)
Fig. 10. Infrared spectrum of [121]tetramantane compared to (unscaled) calculated spectra using different basis sets.
here the observed discrepancies, only when gas-phase diamondoid spectra become available can such a conclusion
be truly substantiated.
3.3. Methyl substitution at the surface of the cage
In addition to the pure diamondoid molecules, the infrared spectrum has been recorded for a methyl substituted
diamondoid: 3-methyl-[1(2,3)4]pentamantane (see Fig. 9),
where the H atom at the position indicated with an arrow
in Fig. 1 is substituted by a CH3 group (thereby causing a
reduction in symmetry from Td to C3v).
With respect to the general appearance of diamondoid
spectra as described above, significant differences are
found. In the CH stretching region around 3 lm, the
calculation produces CH stretching modes localized on
the methyl group at 2930.1 and 2860.5 cm 1, corresponding to the CH3 stretching modes parallel and perpendicular to the threefold symmetry axis, respectively. These
modes can be well observed in the experimental spectrum,
particularly when it is compared with the experimental
spectrum of the non-methylated [1(2,3)4]pentamantane
(see inset in Fig. 9). Though weak, additional bands are
observed for the methylated species at 2857 cm 1 and
around 2945 cm 1 (we cannot explain the observation of
what appears to be a doublet at present). At the red
end of the CH stretching manifold, there appears to be
another additional band at about 2820 cm 1, but the calculation indicates that this band is due to the three degenerate CH stretching modes in the (1 1 1) lattice plane,
J. Oomens et al. / Journal of Molecular Spectroscopy 238 (2006) 158–167
which are red-shifted by about 10 cm 1 in the methylated
molecule with respect to the non-methylated one.
In the remainder of the spectrum, it is not so easy to pinpoint the direct influence of the methyl group on the spectrum. Comparison of the spectrum with that of
unsubstituted [1(2,3)4]pentamantane (see Fig. 6) clearly
shows that additional bands appear, but inspection of the
corresponding normal modes indicates that they are due
to non-localized modes. It therefore appears that these differences are mainly induced by the reduction of symmetry
from Td for [1(2,3)4]pentamantane to C3v for 3-methyl[1(2,3)4]pentamantane.
Some local modes of the methyl group are computed
at 1369.7 cm 1 (CH3 umbrella) and 1457.0 cm 1 (CH3
scissor), but they are not easily recognized in the experimental spectrum, probably due to overlapping diamondoid bands.
4. Conclusion
We report in this paper the first infrared spectroscopic
characterization of several higher diamondoid molecules.
These experimental spectra, supported with density functional theory calculations provide the spectroscopic fingerprint for this class of molecules. Among others, this may
strengthen a possible detection of this family of compounds
in different astrophysical objects, which has thus far been
based only on the observation of CH stretching bands
around 2850 cm 1. In general, a good agreement between
experiments and DFT computations was found, which
allowed us to identify several groups of bands corresponding to certain classes of molecular vibrations. Finally,
methyl substitution on the diamondoid cage induced significant changes particularly in the 3 lm range of the spectrum, which may help to evaluate their astrophysical
relevance.
Appendix A. Supplementary data
Supplementary material. Optimized structures at the
B3LYP/D95(d,p) level for all molecules studied here as
well as experimental frequencies.
Supplementary data for this article are available on
ScienceDirect (www.sciencedirect.com) and as part of the
Ohio State University Molecular Spectroscopy Archives
(http://msa.lib.ohio-state.edu/jmsa_hp.htm).
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
Acknowledgments
We gratefully acknowledge the SARA Supercomputer
center in Amsterdam, The Netherlands, for providing the
CPU time. This work is part of the research program
of FOM (Project No. 03PR2218), which is financially
supported by the ‘Nederlandse Organisatie voor
Wetenschappelijk Onderzoek (NWO)’.
167
[19]
[20]
[21]
[22]
[23]
[24]
P.V.R. Schleyer, J. Am. Chem. Soc. 79 (1957) 3292.
D.V. Korolkov, O.V. Sizova, Int. J. Quantum Chem. 88 (2002) 606.
E. Anders, E. Zinner, Meteorites 28 (1993) 490.
L.J. Allamandola, S.A. Sandford, A.G.G.M. Tielens, T.M. Herbst,
Astrophys. J. 399 (1992) 134.
O. Guillois, G. Ledoux, C. Renaud, Astrophys. J. 521 (1999) L133.
C. van Kerckhoven, A.G.G.M. Tielens, C. Waelkens, Astron.
Astrophys. 384 (2002) 568–584.
H.C. Chang, J.C. Lin, J.Y. Wu, K.H. Chen, J. Phys. Chem. 99 (1995)
11081–11088.
J.C. Blades, D.C.B. Whittet, Mon. Not. R. Astron. Soc. 191 (1980) 701.
W.A. Schutte, A.G.G.M. Tielens, J.L. Allamandola, M. Cohen, D.H.
Wooden, Astrophys. J. 360 (1990) 577.
K. Sellgren, Spectrochim. Acta A 57 (2001) 627.
J.E. Dahl, S.G. Liu, R.M.K. Carlson, Science 299 (2003) 96.
J.E.P. Dahl, J.M. Moldowan, T.M. Peakman, J.C. Clardy, E.
Lobkovsky, M.M. Olmstead, P.W. May, T.J. Davis, J.W. Steeds,
K.E. Peters, A. Pepper, A. Ekuan, R.M.K. Carlson, Angew. Chem.,
Int. Ed. 42 (2003) 2040–2044.
J. Filik, J.N. Harvey, N.L. Allan, P.W. May, J.E.P. Dahl, S. Liu,
R.M.K. Carlson, Spectrochim. Acta A 64 (2006) 681–692.
G. Yan, N.R. Brinkmann, H.F. Schaefer, J. Phys. Chem. A 107
(2003) 9479–9485.
N. Polfer, B.G. Sartakov, J. Oomens, Chem. Phys. Lett. 400 (2004)
201–205.
S.L. Richardson, T. Baruah, M.J. Mehl, M.R. Pederson, Chem. Phys.
Lett. 403 (2005) 83–88.
Y.R. Chen, H.C. Chang, C.L. Cheng, C.C. Wang, J.C. Jiang, J.
Chem. Phys. 119 (2003) 10626–10632.
A.T. Balaban, P. von Rague Schleyer, Tetrahedron 34 (1978)
3599–3609.
A.D. Becke, J. Chem. Phys. 98 (1993) 5648–5652.
C.W. Bauschlicher, S.R. Langhoff, Spectrochim. Acta A 53 (1997)
1225–1240.
M. J. Frisch et al., Gaussian 98 Revision A.11.4 Gaussian, Inc.,
Pittsburgh PA.
A.P. Scott, L. Radom, J. Phys. Chem. 100 (1996) 16502–16513.
http://webbook.nist.gov.
C. Joblin, L. d’Hendecourt, A. Léger, D. Défourneau, Astron.
Astrophys. 281 (1994) 923–936.