Photolysis mechanism of a squarylium dye

Article abstract of DOI:10.1016/j.jphotochem.2010.07.009

The photodegradation mechanism of a squarylium dye was investigated by analyzing the photoproducts using a UV-Vis spectrometer, MALDI-TOF-MASS, HPLC and GC-MS. We found that a squarylium dye was likely decomposed into smaller species by way of an intermediate species which was an adduct of an original dye with a solvent molecule or solvent radical. The reaction mechanism was confirmed using the quantum chemical calculation by estimating the structure of the intermediate species. Since the solvent molecules are involved in the reaction, the reaction rate depended on the used solvent, and the photodegradation rate was much faster in a chloromethane than in an alcohol.

Full text of DOI:10.1016/j.jphotochem.2010.07.009

Journal of Photochemistry and Photobiology A: Chemistry 214 (2010) 264–268
Contents lists available at ScienceDirect
Journal of Photochemistry and Photobiology A:
Chemistry
journal homepage: www.elsevier.com/locate/jphotochem
Photolysis mechanism of a squarylium dye
Kenji Katayamaa,∗, Takafumi Shojia, Kyohei Naitoa, Takeshi Eitokua, Akira Nakayamab
a Department of Applied Chemistry, Faculty of Science and Engineering, Chuo University, 1-13-27 Kasuga, Bunkyo, Tokyo 112-8551, Japan
b Division of Chemistry, Graduate School of Science, Hokkaido University, Sapporo 060-0810, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 2 April 2010
Received in revised form 7 July 2010
Accepted 9 July 2010
Available online 15 July 2010
The photodegradation mechanism of a squarylium dye was investigated by analyzing the photoproducts
using a UV–Vis spectrometer, MALDI-TOF-MASS, HPLC and GC–MS. We found that a squarylium dye was
likely decomposed into smaller species by way of an intermediate species which was an adduct of an
original dye with a solvent molecule or solvent radical. The reaction mechanism was confirmed using
the quantum chemical calculation by estimating the structure of the intermediate species. Since the
solvent molecules are involved in the reaction, the reaction rate depended on the used solvent, and the
photodegradation rate was much faster in a chloromethane than in an alcohol.
Keywords:
Squarylium dye
Photodegradation mechanism
Decomposition
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
2. Experimental setup
The squarylium dye is one of the dye categories, and known as
a functional dye used for the light absorbing material for solar cells
[1,2], optical recording media [3,4], nonlinear optical material [5,6],
bio-labeling [7], singlet oxygen generation [8], etc. Squarylium
dyes, which are 1,3-disubstituted compounds were synthesized by
condensation of one equivalent of squaric acid with two molar
equivalents of electron rich aromatic or heterocyclic methylene
bases [9,10], or by stepwisely the introduction to squaric acid skel-
ton of aromatic or heterocyclic compounds via dialkyl squarate or
squaryl dichloride. The feature of this dye category is that they have
a sharp and intense absorption and fluorescence in the red to near-
infrared region [11,12]. Furthermore, the absorption wavelength
can be controlled by the functional group connected to each side
of the squaric acid, and series of dyes have been investigated and
synthesized [13,14].
In the practical applications used as functional dyes, we need
to know the mechanism of the photodegradation and light resis-
tance of the dye molecule, however there are only a few reports on
the degradation of the squalyrium dyes; reports on a photodegra-
dation process [11], a thermal decomposition process [15,16] and
photocatalytic decomposition process [17] can be found. Here we
investigated the mechanism of the photodegradation processes of
a squarilium dye using various analytical methods and we found
that characteristic intermediate species are involved in the pho-
todegradation reaction and that the reaction rate is dependent on
the solvent.
The sample dye was one of the squalyrium dyes (SQ)
(C28H26ClN3O3, MW: 487, ꢀmax 577.5 nm (CHCl3)) (Kyowa Hakko
Chemicals, Co., Ltd.), and the structure is shown in Fig. 1. On each
side of the squaric acid, it has indolenine and pirazole frameworks.
The SQ was dissolved in a methanol, ethanol, propanol, chloro-
form, and dichloromethane. In the photodegradation monitoring
by an absorption spectrum, the samples with a concentration of
0.05 mM were irradiated with a nanosecond pulse laser with a
wavelength of 532 nm, a pulse energy of 3 mJ/pulse, a repetition
rate of 20 Hz, a beam diameter of 8 mm (GAIA, Rayture systems).
The sample was put in a 10 mm × 45 mm optical cell with a screw
cap with a thickness of 1 mm, and the laser light was shone with-
out focusing. The sample was kept moving during the irradiation.
The absorption spectrum was measured by a spectrometer (V-650,
JASCO) every time after the irradiation of time. The photodegra-
dation speed was accelerated about 10 times using a nanosecond
pulse laser instead of a continuous wave laser (532 nm, 100 mW,
Photop). It was confirmed that the photoproduct species generated
by both the lasers were similar, which was confirmed by MALDI-
TOF-MASS measurements. It is considered that the photoproducts
(intermediate species) are involved in the reaction and they are
concentrated in case of a short optical pulse because the generated
species cannot diffuse away from the irradiated area. However, it
is supposed that the reaction mechanism is similar because the
generated (intermediate) species are similar.
For the structure analysis of the photoproduct species, MALDI-
TOF-MS (AXIMA-CFR, SHIMADZU), GC–MS (6890N Network GC
System, Agilent Technologies) were mainly used. For the MALDI-
TOF-MS measurements, 1,8,9-trihydroxyanthracene was used as
a matrix material. For these analyses, the sample solutions were
∗
Corresponding author. Tel.: +81 3 3817 1913; fax: +81 3 3817 1913.
E-mail address: kkata@kc.chuo-u.ac.jp (K. Katayama).
1010-6030/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.jphotochem.2010.07.009

K. Katayama et al. / Journal of Photochemistry and Photobiology A: Chemistry 214 (2010) 264–268
265
Fig. 1. The molecular structure of the used squalirium dye. The photoproduct
species observed by GC–MS measurement was also shown.
irradiated by a nanosecond pulse laser for 120 min for alcohol
solutions, and for 3 min for chloromethane solutions.
3. Results and discussion
The absorption spectra during the light irradiation are shown
in Fig. 2 for methanol, ethanol, and chloroform solvents. In alco-
hol solutions, it took a few hours for the decay of the main
peak, namely photodegradation, while it took only several min-
utes for chloromethane solvents. The degradation time until the
original molecule was reduced to half was 120, 150, 150, 20,
4 min for methanol, ethanol, propanol, dichloromethane, chloro-
form, respectively. The insets in each figure show the normalized
spectrum for various irradiation times at the intensities of the
main peaks because it can be easily seen that the intensities of the
shoulder peaks around 530 nm gradually increased relatively to the
main peaks. At the same time, a small absorption profile gradually
increased in the UV region.
The photodegradation rate was much faster in chloromethane
solvents than in alcohol solvents, and it implies that solvent
molecules are involved in the reaction mechanism, causing the
difference in the reaction rate. Since the shoulder peak around
530 nm decreased with time but increased its ratio to the main
peak intensity, it is considered that it corresponds to the peak for an
intermediate species in the photodegradation process. It is assumed
that the intermediate species remains its molecular structure sim-
ilar to the original molecule because of the visible absorption close
to the absorption for the original molecule. The monotonic increase
in the UV absorption indicates photoproducts with a smaller molec-
ular size were generated as final products.
To confirm the assumption that the shoulder peak corresponds
to the intermediate speices, 3D-HPLC was measured for the ethanol
solutions and the result was shown in Fig. 3. The eluent was
THF/0.1% formic acid in water (50:50). The original molecule cor-
responds to the peak at 14.3 min, which was confirmed by the
absorption spectrum shown in Fig. 3b. The other peaks except for
the one at 9.7 min showed no absorption in the visible wavelength
region. For the peak at 9.7 min, the absorption around 520 nm
was confirmed. Since the absorption peak is close to the shoulder
peak observed in Fig. 2, this corresponds to the main intermediate
species.
MS spectra before and after irradiation of light were shown in
Fig. 4 for methanol and chloroform solvents. Light was irradiated
for 120 min for alcohol solvents and for 3 min for chloromethane
solvents. The original molecular peak of SQ showed a typical peak
profile including a chloride isotope. In alcohol solutions, besides
Fig. 2. Temporal change in the absorption spectra of the squalirium dye in (a)
methanol, (b) ethanol, and (c) chloroform solvents by irradiation of a pulse laser
(wavelength: 532 nm, pulse energy: 3 mJ/pulse, repetition rate: 20 Hz). In (a) and
(b), spectra for every 30 min are shown, while those for every 1 min are shown in
(c). In (c), absorption spectra for the wavelength less than 320 nm were not mea-
sured due to fluorescence. The insets show the spectra normalized by the intensity of
the main absorption peak. The absorption profiles in the UV region were expanded
for clarity.

266
K. Katayama et al. / Journal of Photochemistry and Photobiology A: Chemistry 214 (2010) 264–268
Fig. 4. Mass spectra (MS) for the squalirium dye before and after the irradiation of
the pulse laser. The MS spectrum for the original molecule is shown in (a), and the
followings correspond to those after irradiation in (b) methanol for 120 min and in
(c) chloroform for 3 min.
Fig. 3. A 3D-HPLC result for a sample after 4 h irradiation of light in 0.05 mM EtOH
solution; the LC chart (a) and absorption spectra for the retention time 14.3 min (b)
and 9.7 min (c), respectively.
adduct of the SQ molecule with a solvent molecule is formed. In
chloromethane solvents, it is considered that the excited state of
a SQ dye was interacted with a solvent molecule, which induces
the dissociation of C–H or C–Cl bonds, causing formation of solvent
radicals. A hydrogen position was assumed to be substituted by the
radicals. Radical formation is supported by the fact that the pho-
todegradation rate of SQ is much faster in chloromethane solvents
and that various adduct species (CH2CH2, CHCl and CCl2, etc.) were
observed in the MS spectra. Since it is considered that the addi-
tion of such solvents or solvent radicals does not cause the total
disappearance of the absorption in the visible wavelength region,
it is safely concluded that the adduct species are the intermediate
species corresponding to the shoulder peak in Fig. 2 and the peak
at 9.7 min in Fig. 3.
The subsequent process of the photodegradation should be con-
sidered because the solution color was finally lost in the process.
Since the final photoproducts corresponding to the UV absorption
observed in Fig. 2 could not be detected by MALDI-TOF-MS, the
photoproducts were analyzed by GC–MS, expecting the detection
of smaller photoproducts, although SQ itself cannot be detected
the peak for SQ molecule (m/z = 488 as [M+H]+), another peak at
m/z = 519 for methanol, at m/z = 533 for ethanol was observed.
The molecular weights for the newly observed peaks correspond
to those for SQ + CH3O (or CH3OH) and CH3CH2O (or CH3CH2OH),
respectively. In chloromethane solvents, the peak for SQ molecule
decreased again by light irradiation, and a few other peaks were
observed. For a chloroform solvent, peaks at m/z = 516, 536, and
571 were found. They correspond to the molecular weights for
SQ + CH2CH2, CHCl and CCl2, respectively. We could not find new
peaks in the region of the smaller molecular number than 487, pos-
sibly because they were not ionized or buried in the matrix spectra.
To confirm the above results, a sample after 4 h irradiation of the
light in 0.05 mM EtOH solution was also analyzed by LC–ESI-MS.
In the LC–ESI-MS measurement, a component corresponding to
SQ + CH3CH2O (CH3CH2OH) was found from the observed peaks at
m/z 534 (ESI+) and m/z 532 (ESI−).
From the above results, it is considered that addition of solvents
or their radicals to a SQ molecule is the initial step for the pho-
todegradation process. In alcohol solvents, it is supposed that an

K. Katayama et al. / Journal of Photochemistry and Photobiology A: Chemistry 214 (2010) 264–268
267
structure using the BLYP functionals with the long-range correction
(LC) [20] and a 6-31G(d,p) basis set. All calculations are performed
using the GAMESS package [21].
The S0 → S1 transition in the SQ dye is viewed as HOMO → LUMO
transition, and as seen in Fig. 5a and b, it involves a charge-transfer
from the oxygen atoms at the both sides of the squaric ring to the
carbon atoms in the ring. The excitation energy obtained by TD-DFT
calculation was calculated to be 2.46 eV (504 nm), which is about
0.3 eV more than the experimental value. This discrepancy is mainly
due to the neglect of solvent molecules.
Considering the final photoproduct of indolenine framework
and the observation that the electronic excitation energy is local-
ized around the squaric ring, we speculate that a solvent molecule
attack on the left side of the squaric ring. Thus, the following adduct,
where CH3O– add on the left carbon atom in the squaric ring (Fig. 5c
and d), was examined in the calculation. After the geometry opti-
mization, the electronic excitation energy was calculated to be
2.61 eV (475 nm), which is about 30 nm blue-shifted from the orig-
inal SQ molecule. This excitation is also HOMO → LUMO transition
and it involves a charge transfer from the central squaric ring to the
left indolenine framework. The oscillation strength for this transi-
tion was given as 0.2, which is relatively small due to the nature of
long-range charge-transfer transition. The observed blue-shift is in
agreement with the experimental findings and it is partly due to the
breaking of the electronic conjugation between the two moieties
attached to the squaric ring. It is assumed that the addition of sol-
vent molecules weaken the bond strength between the indolenine
and squaric ring part, leading to the dissociation. Furthermore, it
was predicted from the NMR spectrum of the intermediate species
that the double bond peak corresponding to the indolenine and
squaric ring bond was weakend.
Finally, it is commented that the above reaction is not a ther-
mally induced reaction. It was checked by the change in the
absorption spectrum of SQ solutions at 323 K, but no change was
observed for hours. Since the temperature rise due to light irra-
diation is only a few degrees in our experiment, the degradation
observed is due to light irradiation.
4. Conclusion
The photodegradation process of a squarilium dye was inves-
tigated by several analytical methods. It was clarified that the
processes depended on the solvent and the solvent molecules were
involved in the reaction mechanism. Since there have been only a
few reports on the molecular mechanism of the photodegradation
of squarilium dyes, the mechanism would be informative for those
who design the molecular structure or utilize squarilium dyes as
photosensitizer, recording media, etc.
Fig. 5. Highest occupied molecular orbitals (HOMO) and lowest unoccupied molec-
ular orbitals (LUMO) for the squarylium dye (a, b) and its adduct (c, d) with a
methanol molecule at the equilibrium structure in the electronic ground state.
Acknowledgements
by GC–MS due to high evaporation temperature. The indolenine
framework was found only for the samples after light irradia-
tion, and the molecular structure is shown in Fig. 1. This result
strongly suggests that the single bond connecting the indoleneine
framework and the squarine ring was dissociated in the pho-
todegradation process.
This research was financially supported by a Grant-in-Aid for
Scientific Research from the Japan Society for the Promotion of Sci-
ence (No. 18686063). The sample was provided by Kyowa Hakko
Chemical Co., Ltd.
References
In order to obtain the detailed understanding of the reaction
mechanism involving the intermediate states, we have performed
quantum chemical calculations for the original SQ dye and its
adduct species with solvent molecules. The geometry optimization
in the electronic ground state (S0) is carried out using density func-
tional theory (DFT) with the B3LYP combination of exchange and
correlation functionals and a 6-31G(d,p) basis set. Subsequently,
the time-dependent (TD) DFT calculations [18,19] are performed
for calculating the vertical excitation energy at the equilibrium
[1] A.P. Piechowski, G.R. Bird, D.L. Morel, E.L. Stogryn, J. Phys. Chem. 88 (1984) 934.
[2] D.L. Morel, A.K. Ghosh, T. Feng, E.L. Stogyn, P.E. Purwin, R.F. Shaw, C. Fishman,
Appl. Phys. Lett. 32 (1978) 495.
[3] K.Y. Law, J. Phys. Chem. 92 (1988) 4226.
[4] K.Y. Law, Chem. Rev. 93 (1993) 449.
[5] C.W. Dirk, W.C. Herndon, F. Cervantes-Lee, H. Selnau, S. Martizez, P.
Kalamegham, A. Tan, G. Campos, M. Velez, J. Zyss, I. Ledoux, L.-T. Cheng, J. Am.
Chem. Soc. 117 (1995) 2214.
[6] M. Furuki, M. Tian, Y. Sato, L.S. Pu, S. Tatsuura, O. Wada, Appl. Phys. Lett. 78
(2001) 2634.

268
K. Katayama et al. / Journal of Photochemistry and Photobiology A: Chemistry 214 (2010) 264–268
[7] F. Welder, B. Paul, H. Nakazumi, S. Yagi, C.L. Colyer, J. Chromatogr. B 793 (2003)
93.
[15] M. Tian, M. Furuki, I. Iwasa, Y. Sato, L.S. Pu, S. Tatsuura, J. Phys. Chem. B 106
(2002) 4370.
[8] P.F. Santos, L.V. Reis, P. Almeida, A.S. Oliveira, L.F.V. Ferreira, J. Photochem.
Photobiol. A 160 (2003) 159.
[9] H.E. Sprenger, W. Ziegeubein, Angew. Chem. Int. Ed. Engl. 5 (1966) 894.
[10] H.E. Sprenger, W. Ziegenbein, Angew. Chem. Int. Ed. Engl. 6 (1967) 553.
[11] S.H. Kim, S.H. Hwang, Dyes Pigments 35 (1997) 111.
[12] B. He, X. Cai, M. Lu, D. Huang, Dyes Pigments 41 (1999) 31.
[13] Ch. Prabhaker, K. Yesudas, G.K. Chaitanya, S. Sitha, K. Bhanuprakash, V.J. Rao, J.
Phys. Chem. A 109 (2005) 8604.
[16] T. Wu, G. Liu, J. Zhao, H. Hidaka, N. Serpone, New J. Chem. 24 (2000) 93.
[17] J. Li, C. Chen, J. Zhao, H. Zhu, J. Orthaman, Appl. Catal. B 37 (2002) 331.
[18] A. Dreuw, M. Head-Gordon, Chem. Rev. 105 (2005) 4009.
[19] P. Elliott, F. Furche, K. Burke, Rev. Comp. Chem. 26 (2009) 91.
[20] H. Iikura, T. Tsuneda, T. Yanai, K. Hirao, J. Chem. Phys. 115 (2001) 3540.
[21] M.W. Schmidt, K.K. Baldridge, J.A. Boatz, S.T. Elbert, M.S. Gordon, J.H. Jensen, S.
Koseki, N. Matsunaga, K.A. Nguyen, S.J. Su, T.L. Windus, M. Dupuis, J.A. Mont-
gomery, J. Comput. Chem. 14 (1993) 1347.
[14] R.A. Negres, O.V. Przhonska, D.J. Hagan, E.W. Van Stryland, M.V. Bondar, Y.L.
Slominsky, A.D. Kachkovski, IEEE J. Sel. Top. Quantum Electron. 7 (2001) 849.