The Sphol'skii fluorescence spectrum of cyclopenta[c,d]pyrene. Novel evidence for anomalous S2-emission

Article abstract of DOI:10.1016/S1386-1425(98)00045-6

The high resolution Sphol'skii fluorescence spectrum of the non-alternant polycyclic aromatic hydrocarbon cyclopenta[c,d]pyrene (CPP, 1) in an n-hexane matrix at cryogenic temperature is reported. Two distinct narrow spectral lines at 378.0 and 379.6 nm representing two sites were found, whereas in n-octane no Sphol'skii effect was observed. In accordance with reported data on the fluorescence of CPP (1) these lines have to be assigned to anomalous S₂-emission.

Full text of DOI:10.1016/S1386-1425(98)00045-6

Spectrochimica Acta Part A 54 (1998) 1443–1449
The Sphol’skii fluorescence spectrum of cyclopenta[c,d]pyrene.
Novel evidence for anomalous S -emission
2
a,
1,a
a
b
Cees Gooijer *, Igor Kozin , Nel H. Velthorst , Martin Sarobe ,
2
,b
b
Leonardus W. Jenneskens , Edward J. Vlietstra
a
Department of General and Analytical Chemistry, Free Uni6ersity, De Boelelaan 1083, 1081 HV Amsterdam, The Netherlands
b
Debye Institute, Department of Physical Organic Chemistry, Utrecht Uni6ersity, Padualaan 8, 3584 CH Utrecht, The Netherlands
Received 16 October 1997; accepted 17 February 1998
Abstract
The high resolution Sphol’skii fluorescence spectrum of the non-alternant polycyclic aromatic hydrocarbon
cyclopenta[c,d]pyrene (CPP, 1) in an n-hexane matrix at cryogenic temperature is reported. Two distinct narrow
spectral lines at 378.0 and 379.6 nm representing two sites were found, whereas in n-octane no Sphol’skii effect was
observed. In accordance with reported data on the fluorescence of CPP (1) these lines have to be assigned to
anomalous S -emission. © 1998 Elsevier Science B.V. All rights reserved.
2
Keywords: Sphol’skii fluorescence; Cyclopenta[c,d]pyrene; S -emission
2
1
. Introduction
the large energy difference between the S - and
1
−
1
S -states (E –E =DE
ꢀ10000 cm ) [1].
2
S
S
gap
2
1
Among the plethora of (non)-alternant poly-
cyclic aromatic hydrocarbons (PAH) several rep-
resentatives exhibit anomalous fluorescence, i.e.
However for other representatives, the situation is
less clear. It is not always evident whether anoma-
lous fluorescence or fluorescence of a trace-level
amount impurity with a high fluorescence quan-
emission from the higher excited singlet S -state,
2
tum yield (F ) is responsible for the emission. For
F
instead of (or in addition to) emission from the
S1-state. A seminal example is the non-alternant
PAH azulene. Upon excitation it gives strong
S2-emission, which is rationalized on the basis of
instance, the long-standing controversy concern-
ing the emissive properties of fluoranthene for
which earlier two emissions have been reported at
3
[
50 and 410 nm was put to an end by some of us
2]. Sphol’skii spectroscopy—a cryogenic high-
resolution molecular luminescence technique
3]—unambiguously showed that the 350 nm
*
Corresponding author. Tel.: +31 20 4447540; fax: +31
2
0 4447543; e-mail: gooijer@chem.vu.nl/aac@chem.vu.nl
[
1
Present address: Department of Chemistry, Queen’s Uni-
emission is not due to anomalous S -emission, but
2
versity, Kingston, K7LN6 Ontario, Canada.
2
a consequence of a strongly fluorescent impu-
rity—most probably acephenanthrene. This ex-
Tel.: +31 30 2533128; fax: +31 30 2534533; e-mail:
jennesk@chem.run.ul
1
386-1425/98/$19.00 © 1998 Elsevier Science B.V. All rights reserved.
PII S1386-1425(98)00045-6

1
444
C. Gooijer et al. / Spectrochimica Acta Part A 54 (1998) 1443–1449
Scheme 1.
ample underlines the potential of Sphol’skii
fluorescence spectroscopy for the identification of
PAH and interfering fluorescent impurities.
for a very weak though measurable anomalous
−
1
S2-state fluorescence near 25000 cm
(400 nm;
quantum yield F 0.003 and lifetime ~ of ꢀ3 ns)
F
In addition, high-resolution PAH Sphol’skii
fluorescence data are of interest from an analytical
point of view. It is well documented that PAH are
invariably generated during incomplete combus-
tion of hydrocarbon containing fuels and that
specific representatives are primarily responsible
for the genotoxicity of these combustion samples
has been reported [9]. Unfortunately, cryogenic
absorption spectra of CPP (1) provide little addi-
tional information compared with spectra
recorded at room temperature. Upon cooling a
−
4
10
M solution of CPP (1) in n-octane to 77 K,
there is hardly any improvement of spectral reso-
lution [10]. At 5 K, weak Sphol’skii-type features
in the 460–550 nm spectral range can be observed.
It should be noted, however, that the latter spectra
were obtained for saturated solutions of CPP (1)
in n-octane [10]. Under these conditions, compli-
cating factors like aggregation of CPP (1), which is
detrimental for the Sphol’skii effect, and the possi-
ble presence of impurities in the sample showing
the Sphol’skii effect cannot be excluded.
[4]. Hence, in situ PAH identification in complex
samples using Sphol’skii fluorescence spectroscopy
has become an important research subject [5]. A
drawback, however, is the availability of pure
PAH for the determination of reference data.
Among the ubiquitous PAH generated during
combustion
the
non-alternant
cyclopen-
ta[c,d]pyrene (CPP, 1) has been recognized to
contribute substantially to the genotoxicity of the
samples [6]. With regard to the origin of CPP (1)
under high temperature conditions in the gas
phase the intermediacy of 1-ethynylpyrene (3) ob-
tained via either C or ethyne (C H ) addition to
Here we report the Sphol’skii fluorescence spec-
trum of CPP (1). It is shown that the anomalous
emission occurs from the S -state.
2
2
2
2
pyrene has been invoked [7]. This proposal was
supported by flash vacuum thermolysis (FVT) of
2. Experimental
1
-chloroethenylpyrene (4) being a ‘masked’ 1-
2.1. Chemicals
ethynylpyrene. Under unimolecular FVT condi-
−
2
tions (1000°C, 10
(
Torr) 1-choroethenylpyrene
By recording fluorescence emission-excitation
spectra the purity of the solvents (n-octane, n-hex-
ane and dichloromethane) was checked. No sig-
nificant response was observed and the solvents
were used without further purification.
4) is efficiently converted into CPP (1, mass
recovery 88%) via a consecutive process, i.e. facile
hydrogen chloride elimination (4“3) followed by
ethynyl ethylidene carbene equilibration (3“2)
and ultimately carbene C–H insertion (2“1,
Scheme 1) [8]. The accessibility of CPP (1) on a
gram scale prompted us to reinvestigate its anoma-
lous emissive properties using Sphol’skii fluores-
cence spectroscopy. In the case of CPP (1), instead
2.2. Flash 6acuum thermolysis of
1-chloroethenylpyrene (4)
FVT (unpacked quartz tube (length 40 cm,
−
2
of S -state fluorescence which is absent, evidence
diameter 2.5 cm), 1273 K and 10
Torr, subli-
1

C. Gooijer et al. / Spectrochimica Acta Part A 54 (1998) 1443–1449
1445
o
mation temperature 130 C) of 1-chloroethynyl-
Applied Biosystems Model ABI1000S Diode Ar-
ray Detector (5 nm bandpass).
pyrene (4, 1.5 g, 5.8 mmol) gave CPP (1, 1.16 g,
5
.1 mmol, mass recovery 88%). Satisfactory ana-
Isocratic elution was carried out with a mixture
of n-hexane and dichloromethane (9:1 v/v) at a
1
13
lytical data ( H and C NMR, GC-MS, IR),
which were in line with those previously reported
−
1
flow rate of 1.5 ml min . The effluent was
usually monitored at 254 nm. Chromatographic
fractions containing PAH were collected into sep-
arate vials and submitted either to reversed-phase
LC or Sphol’skii fluorescence measurements.
Reversed-phase LC was carried out using a
Varian Star LC system (Star 9012 Solvent Deliv-
ery System), a Varian Mistral Column Ther-
mostat, a Star 9100 Autosampler, a Separations
GT-103 Degasser, a Vydac 250×4.6 mm i.d. RP
C18 column and a Varian 9065 Polychrom Diode
Array Detector (4 nm bandpass). The level of
acetonitrile was increased stepwise gradually from
[8], were obtained. For the fluorescence measure-
ments crude CPP (1) was further purified by
preparative column chromatography (silica, elu-
ens n-hexane:chloroform 1:1 v/v). The UV/Vis
absorption spectrum of CPP (1, solvent n-hexane,
−
4
1
0
M, room temperature) was recorded on a
Cary 1 spectrophotometer (Table 1).
2
.3. Liquid chromatography separation
For normal-phase liquid chromatography (LC),
a Sigma-Aldrich Nucleosil NH2 column (5 mm
particle size, 250×4.6 mm i.d.) equipped with a
6
5% in the beginning of separation up to 100% at
the end of the run. The effluent was monitored at
u 254 nm.
1
00 ml injection loop was used. The hardware
consisted of two Applied Biosystems Model 400
Solvent Delivery Systems, and a multi-wavelength
2
.4. Sphol’skii fluorescence detection
Table 1
For cryogenic fluorescence measurements, the
fractions collected after carrying out normal-
PPP/SCF CI vertical excitation energies and oscillator
strengths of CPP (1)
a
phase LC were dried in a gentle flow of N , and
2
PPP u [nm]
Oscillator
strength f
Experimental u [nm]
{log m}
redissolved in either n-hexane or n-octane. A 450
W Xenon arc lamp coupled with a Bausch &
Lomb monochromator (5 nm bandpass) provided
lamp excitation.
A closed-cycle helium CTI cryogenics refrigera-
tor was employed to simultaneously cool down to
b
b
b
452.9 (505)
369.9 (386)
354.8 (352)
0.086
0.360
0.365
450 {2.77}
386 {3.37}
373 {4.06}
3
3
3
63 {3.99}
51 {4.20}
41 {4.06}
1
5–25 K 4×10 ml sample solutions and the cold
b
b
b
b
b
2
2
2
2
2
2
2
2
2
69.7 (308)
66.2 (295)
53.8 (276)
26.2 (247)
19.2 (217)
13.5
12.0
08.9
00.6
0.189
0.439
0.338
0.138
0.717
0.314
0.636
0.432
0.324
291 {4.22}
279 {4.16}
237 {4.66}
221 {4.81}
tip of a cryogenerator was provided by using a
gilded copper sample holder covered with sap-
phire windows [3]. For luminescence detection,
the emission signal from a front-face illuminated
sample was focused on the entrance slit of a
Jobin-Yvon HR1000 monochromator (0.08 nm
resolution) and measured by an EG&G (Prince-
ton, NJ) Model 1421B blue-enhanced intensified
linear photodiode array. This enables simulta-
neous data acquisition from about 20 nm wide
spectral regions. The spectral data were collected
and processed by employing a Model 1463 detec-
tor controller and a Model 1460 OMA III control
system.
a
For the PPP/SCF CI calculation of CPP (1) a standard was
˚
used (pyrene part: C–C 1.40 A, C–C–C-valence angles 120°,
fused cyclopentene moiety: C–C 1.45 A, C=C 1.35 A and
C–C–C-valence angles 108 ). For the calculation of the opti-
cal properties full single CI was carried out [11]. Vertical
excitation energies of CPP (1) from Ref [9] (no numerical
values of the oscillator strengths f are reported).
˚
˚
o
b

1
446
C. Gooijer et al. / Spectrochimica Acta Part A 54 (1998) 1443–1449
3
. Results and discussion
To determine if strongly fluorescent PAH-like
impurities in trace-level amounts are still present
in our CPP (1) sample both normal-phase and
reversed-phase column liquid chromatography
(LC, UV absorption detection u 254 nm) at high
sensitivity was used. The expanded normal-phase
chromatogram of CPP (1) obtained after prepara-
tive colum chromatography, is depicted in Fig.
1(A). Clearly, besides the main peak at 6.5 min,
which corresponds to CPP (1), trace-level
amounts of impurities, which also absorb at u 254
nm, are present. Therefore, three fractions (1, 2
and 3) (Fig. 1(A)), which represent the effluent
parts from 4.5 to 6.25 min, 6.45 to 6.70 min and
7.25 to 15.0 min, respectively, were isolated and
subsequently subjected to Sphol’skii fluorescence
analysis. For the normal-phase effluent fraction 2
only a single chromatographic peak assigned to
CPP (1) was also found under reversed-phase LC
conditions; this strongly corroborates the absence
of additional impurities (Fig. 1(B)).
The Sphol’skii fluorescence spectra obtained in
n-octane for fractions 1, 2 and 3, respectively, are
shown in Fig. 2. From the data it can be con-
cluded that fraction 1 contains pyrene and substi-
tuted pyrene derivatives (uexc 339, 344 and 349
nm, Fig. 2(A)) [15]. Note that under Sphol’skii
conditions pyrene is efficiently excited at 339 nm
(providing a 0–0 fluorescence transition at 372
nm in n-octane), whereas it can be hardly excited
at 349 nm—a wavelength adequate for the excita-
tion of ethylpyrene (0–0 transition in n-octane at
ꢀ378 nm and a vibronic transition at ꢀ383 nm)
[15]. Hence, the Sphol’skii fluorescence spectrum
does not allow discrimination between various
ethylpyrene isomers. Furthermore, the presence of
substituted pyrene containing either an ethynyl or
a 1-chloroethenyl moiety cannot be excluded
(Scheme 1).
Evidence for anomalous S -emission from CPP
1) after excitation was previously obtained from
steady-state and time-resolved fluorescence spec-
troscopy on liquid-state samples at room tempera-
ture. The steady-state excitation spectrum showed
reasonable agreement with the absorption spec-
2
(
trum in the S region [9]. The possible occurrence
2
of S -emission is further supported by PPP/SCF
2
CI calculations [11]. Vertical excitation energies
and oscillator strengths f are found, which are in
line with available experimental data (Table 1).
Furthermore, the PPP/SCF CI results show sig-
nificant configuration interaction for all calculated
excitations. Note, however, that our PPP/SCF CI
version, which was parameterized for the calcula-
tion of vertical excitation energies and oscillator
strengths f of dye molecules, gave results in better
agreement with experiment than those previously
obtained with another PPP/SCF CI version [9].
The considerable DE (=E –E ) of ꢀ5000–
gap
S
S
1
2
−
1
6
000 cm
is in line with that required for
anomalous S -emission [12].
2
A prerequisite for the unambiguous assignment
of CPP (1) S -emission is to ascertain, that trace-
2
level amounts of strongly fluorescent impurities
are absent. The latter, if present, may produce
fluorescence signals of comparable intensities with
that assigned to CPP (1) S -emission. This is of
2
special importance for CPP (1), since it is pre-
pared from a substituted pyrene (4, Scheme 1).
Note that pyrene possesses a fluorescence quan-
tum yield FF 0.10 for S -emission [13]. Fortu-
1
nately, Sphol’skii fluorescence spectroscopy
possesses additional selectivity [3]. Whereas the
spectra are highly resolved due to the Sphol’skii
effect, this effect will only be observed for
fluorophores that fit well in the n-alkane matrix.
For example a survey of available atlas of refer-
ence spectra [10,14], reveals that polar
fluorophoric molecules do not meet this require-
ment. In fact, in view of the demands involved in
the present study, most probably only parent
PAH and their alkylated (possibly including their
ethenylated and ethynylated derivatives) have to
be accounted for.
In fraction 3, among others, benz[a]pyrene was
unambiguously identified—note the 0–0 emission
band at 403 nm and a vibronic band at 408 nm
(uexc 348 and 298 nm, respectively, Fig. 2(B)) [15].
We refrain from a structural elucidation of the
constituents in fractions 1 and 3, since it is beyond
the scope of this paper. Of interest are the Sphol’-
skii fluorescence spectra obtained for fraction 2

C. Gooijer et al. / Spectrochimica Acta Part A 54 (1998) 1443–1449
1447
Fig. 1. Normal-phase chromatogram of CPP (1) and related impurities (A) and reversed-phase chromatogram of fraction 2
containing CPP (1; B, see Section 2).
(
3
Fig. 2(C)). In the n-octane Sphol’skii matrix (uexc
33 nm) no sharp features are discernible and the
normal-phase chromatographic separation proce-
dure to purify CPP (1) has been adequate. Re-
markably, however, upon changing the Sphol’skii
matrix from n-octane to n-hexane distinct narrow
possible evidence for the presence of fluorescent
solutes is only suggested by the irregular baseline.
The absence of distinct narrow lines, such as those
observed for fractions 1 and 3, unequivocally
shows that in fraction 2 no related strongly
fluorescent impurities—alkylated PAH—are
present. In agreement with the HPLC data the
−
1
spectral lines positioned at 378.0 (26459 cm
)
−
1
and 379.6 nm (26346 cm ) were recorded for
fraction 2 (uexc 333 nm; Fig. 2(C)). Note that the
blank n-hexane Sphol’skii fluorescence spectrum
did not show any narrow-banded emissions. Such

1
448
C. Gooijer et al. / Spectrochimica Acta Part A 54 (1998) 1443–1449
Fig. 2. Sphol’skii fluorescence spectra of PAH impurities contained in the normal-phase LC fractions 1 (A) and 3 (B; matrix
n-octane), and fraction 2 (C; matrix either n-octane or n-hexane).
an extreme solvent influence is exceptional; for
regular’ PAH (type A molecules), exchange from
n-octane to n-hexane typically induces only
changes in (multiple) site structures in the emis-
sion spectrum, while narrow spectral features re-
main discernible. Only for compounds denoted as
type B molecules in Ref. [16], the appearance of
the Sphol’skii spectrum depends critically on vari-
ous experimental parameters, such as the choice
of matrix and the rate of matrix solidification
upon cooling. This is attributed to the fact that
type B molecules possess molecular dimensions
‘

C. Gooijer et al. / Spectrochimica Acta Part A 54 (1998) 1443–1449
1449
that are incompatible with the crystalline (n-
Acknowledgements
alkane) host matrix. In a simplified explanation
part of them occupy positional sites in the solid
matrix exhibiting quasi-lines, while the other part
are isolated molecules present in intercrystalline
amorphous positions, thus giving rise to broad
bands. The ratio of both fractions is strongly
influenced by the matrix choice. A well known
example of a type B molecule is acenaphthene,
whose Sphol’skii spectra have been thoroughly
studied [17]. Thus under the assumption that CPP
Gerard J. Stroomberg (Institute of Environmen-
tal Studies, Vrije Universiteit, Amsterdam, The
Netherlands) for carrying out reversed-phase chro-
matography and financial support from the Basque
Government (Beca para Formacion de Investi-
gadores (M.S.)) are gratefully acknowledged.
References
(
1) behaves as a type B molecule, the results
[
1] J. Griesser, U. Wild, Chem. Phys. Lett. 52 (1980) 117.
presented in Fig. 2(C) can be rationalized. At any
rate, fraction 2 does not contain any ‘regular’ type
A molecule.
[2] J.W. Hofstraat, G.P. Hoornweg, C. Gooijer, N.H.
Velthorst, Spectrochim. Acta 41A (1985) 801.
[
3] C. Gooijer, F. Ariese, J.W. Hofstraat, N.H. Velthorst,
Trends Anal. Chem. 13 (1994) 53.
Hence, the lines observed at 378.0 and 379.6
nm in Fig. 2(C) have to be ascribed to CPP (1)
emission. In view of the available absorption and
fluorescence data or CPP (1, Refs. [9,10] and
Table 1), they unequivocally reflect anomalous
[4] J.J. Jacob, Pure Appl. Chem. 68 (1996) 301.
[
5] J.C. Fetzer, W.R. Biggs, Polycycl. Arom. Compd. 4
1994) 3.
(
[
6] J.B. Howard, J.P. Longwell, J.A. Marr, C.J. Pope, W.F.
Busby Jr., A.L. Lafleur, K. Taghizadeh, Combust. Flame
101 (1995) 262.
[
[
7] H. Bockhorn, F. Fetting, H.W. Wenz, Ber. Bunsenges.
Phys. Chem. 87 (1993) 1067.
8] M. Sarobe, J.W. Zwikker, J.D. Snoeijer, U.E. Wiersum,
L.W. Jenneskens, J. Chem. Soc. Chem. Commun. (1994)
emission coming from the S -state. The two lines
2
represent 0–0 transitions from two sites. This is in
accordance with the type B character of CPP (1).
In comparison with the fluorescence spectra
recorded for fractions 1 and 3 the line intensities
are weak, despite the fact that CPP (1) strongly
dominates (Fig. 1(A)). These results are not to-
tally unexpected in view of the small quantum
8
9.
9] B.F. Plummer, Z.Y. Al-Saigh, J. Phys. Chem. 87 (1983)
579.
[
1
[10] L.A. Nakhimovskii, M. Lamotte, J. Joussot-Dubien,
Handbook of Low Temperature Electronic Spectra of
Polycyclic Aromatic Hydrocarbons, Elsevier, Amsterdam,
The Netherlands, 1989.
[11] J.E.M.J. Raaymakers, L.W., Jenneskens, PPP-MO Calcu-
lations for IBM-PC and Compatibles using OS/2, Quant.
Chem. Program Exch. Bull. (1995) 15, 68; QCPE 156.
yield (FF 0.003) for S -fluorescence reported [9].
2
Finally, the observation that the 378.0 and 379.6
nm lines are rather sharp is in agreement with
literature data. Because of the relatively long
emission lifetime ~ ꢀ3 ns [9], spectral lifetime
broadening effects, which are usually important
for emission from higher excited states, are negli-
gible for CPP (1).
[
12] G. Eber, S. Schneider, R. Dorr, Chem. Phys. Lett. 52
1977) 59.
[13] J.B. Birks, Chem. Phys. Lett. 25 (1974) 315.
(
[
14] W. Karcher (Ed.), Spectral Atlas of Polycyclic Aromatic
Compounds, vol. 1, Reidel, Dordrecht, 1985; vol. 2,
Reidel, Dordrecht, 1988; vol. 3, Reidel, Dordrecht, 1991.
15] I.S. Kozin, C. Gooijer, N.H. Velthorst, J. Harmsen, R.
Wiegers, Intern. J. Environ. Anal. Chem. 61 (1995) 285.
16] J. Rima, L. Nakimovsky, M. Lamotte, J. Joussot-Dubien,
J. Phys. Chem. 83 (1984) 4302.
[
[
In summary, our results substantiate the previ-
ously reported data concerning the occurrence of
anomalous S -state emission in the case of CPP
2
[17] J.W. Hofstraat, I.L. Freriks, M.E.J. de Vreeze, C. Gooi-
jer, N.H. Velthorst, J. Phys. Chem. 93 (1989) 184.
(
1) [9].
.