Energy funneling of IR photons captured by dendritic antennae and acceptor mode specificity: Anti-stokes resonance raman studies on iron(III) porphyrin complexes with a poly(aryl ether) dendrimer framework

Article abstract of DOI:10.1021/ja042196z

A series of poly(aryl ether) dendrimer chloroiron(III) porphyrin complexes (L_nTPP)Fe(III)Cl (number of aryl layers [n] = 3 to 5) were synthesized, and their Boltzmann temperatures under IR irradiation were evaluated from ratios of Stokes to anti-Stokes intensities of resonance Raman bands. While the Boltzmann temperature of neat solvent was unaltered by IR irradiation (L_nTPP)Fe(III)Cl (n = 3 to 5), all showed a temperature rise that was larger than that of the solvent and greater as the dendrimer framework was larger. Among vibrational modes of the metalloporphyrin core, the temperature rise of an axial Fe-Cl stretching mode at 355 cm^-1 was larger than that for a porphyrin in-plane mode at 390 cm^-1. Although most of the IR energy is captured by the phenyl ν₈ mode at 1597 cm^-1 of the dendrimer framework, an anti-Stokes Raman band of the phenyl ν₈ mode was not detected, suggesting the extremely fast vibrational relaxation of the phenyl mode. From these observations, it is proposed that the energy of IR photons captured by the aryl dendrimer framework is transferred to the axial Fe-Cl bond of the iron porphyrin core and then relaxed to the porphyrin macrocycle.

Full text of DOI:10.1021/ja042196z

Published on Web 06/25/2005
Energy Funneling of IR Photons Captured by Dendritic
Antennae and Acceptor Mode Specificity: Anti-Stokes
Resonance Raman Studies on Iron(III) Porphyrin Complexes
with a Poly(aryl ether) Dendrimer Framework
Yu-Jun Mo,†,§ Dong-Lin Jiang,‡ Makoto Uyemura,‡ Takuzo Aida,‡ and
Teizo Kitagawa*,†
Contribution from the Okazaki Institute for IntegratiVe Bioscience, National Institutes of Natural
Sciences, Myodaiji, Okazaki 444-8787, Japan, and Department of Chemistry and Biotechnology,
School of Engineering, The UniVersity of Tokyo, Hongo, Bunkyo-ku, Tokyo 113-8656, Japan
Received December 28, 2004; E-mail: teizo@ims.ac.jp
Abstract: A series of poly(aryl ether) dendrimer chloroiron(III) porphyrin complexes (LnTPP)Fe(III)Cl (number
of aryl layers [n] ) 3 to 5) were synthesized, and their Boltzmann temperatures under IR irradiation were
evaluated from ratios of Stokes to anti-Stokes intensities of resonance Raman bands. While the Boltzmann
temperature of neat solvent was unaltered by IR irradiation (LnTPP)Fe(III)Cl (n ) 3 to 5), all showed a
temperature rise that was larger than that of the solvent and greater as the dendrimer framework was
larger. Among vibrational modes of the metalloporphyrin core, the temperature rise of an axial Fe-Cl
stretching mode at 355 cm-1 was larger than that for a porphyrin in-plane mode at 390 cm-1. Although
most of the IR energy is captured by the phenyl ν8 mode at 1597 cm-1 of the dendrimer framework, an
anti-Stokes Raman band of the phenyl ν8 mode was not detected, suggesting the extremely fast vibrational
relaxation of the phenyl mode. From these observations, it is proposed that the energy of IR photons captured
by the aryl dendrimer framework is transferred to the axial Fe-Cl bond of the iron porphyrin core and then
relaxed to the porphyrin macrocycle.
Introduction
allow site-specific positioning of multiple chromophore units
in a three-dimensional nanospace. Examples of pioneering works
Biological photosynthesis, which converts solar energy into
chemical potentials, is a critical photochemical event in plants.
In purple photosynthetic bacteria, for example, wheel-like arrays
of bacteriochlorophyll units play a key role in the efficient
capture of light energy and its subsequent funneling into the
reaction center.1 This structural feature has motivated chemists
to design artificial light-harvesting antennae and to explore their
photochemical properties for the purpose of realizing long-range
vectorial energy transfers. Since the first report of Balzani and
co-workers on photophysical properties of ruthenium(II)/os-
mium(II) multinuclear supramolecular dendrimers,2 dye-func-
tionalized dendritic macromolecules have attracted great atten-
tion as light-harvesting antennae for energy transduction.3 Due
to the flexibility of molecular design, dendritic architectures
include photochemical studies on dendritic macromolecules such
as self-assembled lanthanide-core dendrimers,4 perylene-
terminated phenylacetylene conjugated dendrons,5 and porphy-
rin-incorporated dendrimers,6 along with some theoretical
calculations on energy transfer events in dendritic macromol-
ecules.7
Through studies on a series of porphyrin-cored poly(benzyl
ether) dendrimers and dendrons, we have found that a large,
spherical fourth-generation dendrimer porphyrin, upon exposure
to UV light, shows an exceptionally high energy transfer
efficiency from the peripheral aromatic units to the porphyrin
core.6c An analogous trend has been observed for star and
conically shaped dendritic zinc porphyrin arrays having a free-
base porphyrin core, where the former shows a much more
† National Institutes of Natural Sciences.
(4) Kawa, M.; Fre´chet, J. M. J. Chem. Mater. 1998, 10, 286-296.
(5) (a) Devadoss, C.; Bharathi, P.; Moore, J. S. J. Am. Chem. Soc. 1996, 118,
9635-9644. (b) Kopelman, P.; Shortreed, M.; Shi, Z.-Y.; Tan, W.; Xu,
Z.; Moore, J. S.; Klafter, J. Phys. ReV. Lett. 1997, 78, 1239-1242.
(6) (a) Aida, T.; Jiang, D.-L. In The porphyrin handbook Vol. 3. Inorganic,
Organometallic, and Coordination Chemistry; Kadish, K. M., Smith, K.
M., Guilard, R., Eds.; Academic Press: New York, 2000; pp 369-384.
(b) Jin, R.-H.; Aida, T.; Inoue, S. J. Chem. Soc., Chem. Commun. 1993,
1260-1261. (c) Jiang, D.-L.; Aida, T. J. Am. Chem. Soc. 1998, 120, 10895-
10901. (d) Choi, M.-S.; Aida, T.; Yamazaki, T.; Yamazaki, I. Angew.
Chem., Int. Ed. 2001, 40, 3194-3198. (e) Choi, M.-S.; Aida, T.; Yamazaki,
T.; Yamazaki, I. Chem.-Eur. J. 2002, 8, 2668-2673.
‡ The University of Tokyo.
§ Present address: Physics Department, Henan University, Kaifeng
475001, China. E-mail: yujunmo@henu.edu.cn.
(1) Mcdermott, G.; Prince, S. M.; Freer, A. A.; Hawthornthwaite-Lawless, A.
M.; Rapiz, M. Z.; Cogdell, R. J.; Isaacs, N. W. Nature 1995, 374, 517-
521.
(2) Campagna, S.; Denti, G.; Serroni, S.; Ciano, M.; Juris, A.; Balzani, V.
Inorg. Chem. 1992, 31, 2982-2984.
(3) For selected reviews, see: (a) Aida, T.; Jiang, D.-L. In Dendrimers and
Dendritic Polymers; Fre´chet, J. M. J., Tomalia, D. A., Eds.; Wiley-VCH:
Weinheim, Germany, 2000; Chapter 17, pp 425-438. (b) Adronov, A.;
Fre´chet, J. M. J. Chem. Commun. 2000, 1701-1710. (c) Choi, M.-S.;
Yamazaki, T.; Yamazaki, I.; Aida, T. Angew. Chem., Int. Ed. 2004, 43,
150-158.
(7) (a) Bar-Haim, A.; Klafter, J.; Kopelman, R. J. Am. Chem. Soc. 1997, 119,
6197-6198. (b) Tretiak, S.; Chernyak, V.; Mukamel, S. J. Phys. Chem. B
1998, 102, 3310-3315.
9
10020
J. AM. CHEM. SOC. 2005, 127, 10020-10027
10.1021/ja042196z CCC: $30.25 © 2005 American Chemical Society

Energy Funneling of IR Photons
A R T I C L E S
Figure 1. Schematic representations of (TPP)Fe(III)X (X ) Cl or Br) and (LnTPP)Fe(III)Cl (n ) 3 to 5).
efficient energy transfer than the latter upon visible-light
excitation of the peripheral zinc porphyrin units.6d,e These
observations, along with fluorescence anisotropy studies, have
indicated a morphology effect on the light-harvesting functions
of dendritic macromolecules. In relation to such antenna effects
of dendrimers for UV and visible lights, we have reported an
intriguing phenomenon that large third- and fourth-generation
poly(benzyl ether) dendrimers with an azobenzene core, upon
exposure to infrared (IR) light, undergo cis-to-trans isomeriza-
tion of the azobenzene unit.8 A detailed study with monochro-
matized IR lights has indicated that the energy of IR photons
absorbed by the aromatic building units in the dendrimer
framework is responsible for this isomerization. Among three
different wavenumbers used for IR irradiation [2500 cm-1
(transparent), 1597 cm-1 (ν8), and 1155 cm-1 (CH2-O str.)],
only the 1597 cm-1 radiation facilitated the isomerization
reaction, and the observed dependence of the isomerization rate
constant on the IR irradiation intensity indicated that 4.9 photons
are involved in this photochemical process of the n ) 5
azodendrimer.8 Under otherwise identical conditions, lower-
generation homologues as well as nondendritic azobenzene did
not show any response to IR irradiation. We have tentatively
suggested that a large poly(benzyl ether) dendrimer framework
around the azobenzene core could prevent the vibrational energy
from dissipation, whereas such an energy dissipation in ordinary
systems takes place instantaneously.
enabled us to repeat the phenomenon for an identical sample
and thus to accumulate the spectra to raise the signal-to-noise
ratios. It is well established that the intensity of anti-Stokes
Raman scattering reflects the Boltzmann temperature because
the scattering intensity is proportional to the population at
vibrationally excited levels. Even if an absolute value of a
scattering cross section is not available, the Boltzmann tem-
perature can be determined from the ratio of the Stokes Raman
intensity (IS) to the anti-Stokes Raman intensity (IAS) of a given
vibrational mode. By means of picosecond time-resolved anti-
Stokes Raman scattering spectroscopy, we have successfully
analyzed the cooling process of heme following photodissocia-
tion of CO from carbonmonoxy-myoglobin.9
Given the validity of this method, we herein applied continu-
ous wave (CW) anti-Stokes Raman spectroscopy to (LnTPP)-
Fe(III)Cl (n ) 3 to 5) for the investigation of energy funneling
of IR photons captured by their dendritic antennae. (LnTPP)-
Fe(III)Cl (n ) 3 to 5) under IR illumination could enable
monitoring of the Boltzmann temperatures of their individual
building units. One may argue against this possibility because
the vibrational relaxation is so fast that stationary-state measure-
ments would not provide any correct information. This would
be true if the number of IR capturing units is small. When this
number is sufficiently large, IR energy captured by different
antenna units at the same time increases and may exceed the
relaxation rate at the porphyrin, although a nonlinear relaxation
process such as annihilation may provide additional pathways
compared to lower generation dendrimers. As a result, it may
practically raise the temperature of the porphyrin core during
IR irradiation. Indeed, the results with higher-generation (Ln-
TPP)Fe(III)Cl clearly indicated that the energy acquired from
IR light is localized to a great extent at the focal Fe-Cl bond.
In the present article, we report effects of IR irradiation on
In the present work, we synthesized poly(benzyl ether)
dendrimers (LnTPP)Fe(III)Cl (Figure 1; number of benzyl ether
layers; n ) 3 to 5) bearing a chloroiron(III) porphyrin
functionality at the core and investigated their Raman spectral
features in dioxane under irradiation with IR light that can excite
the aromatic units in the dendrimers. Since the porphyrin core
undergoes no photochemical reaction under IR irradiation, it
(8) Jiang, D.-L.; Aida, T. Nature 1997, 388, 454-456.
(9) Mizutani, Y.; Kitagawa, T. Science 1997, 278, 443-445.
9
J. AM. CHEM. SOC. VOL. 127, NO. 28, 2005 10021

A R T I C L E S
Mo et al.
(L2) into a polychromator (2) equipped with liquid nitrogen-cooled
charge-coupled device (CCD) as a detector. The laser power at the
sample point was 3 mW, unless otherwise stated. The measurements
of the ratios of Stokes and anti-Stokes Raman intensities of a solvent
band of sample solutions with the laser power of 3 mW indicated that
the temperature was 1.5 K higher than the case with 0.6 mW (lowest
power) even when water at 21 °C was circulated through the cell holder.
Accordingly, the surrounding temperature for the dendrimer samples
in the practical measurements was 22.5 °C.
The source of IR irradiation (3), which is a ceramic heater of 60 W
(Jasco, model LS-600), is expected to emit radiation close to the
blackbody at 1300 °C at 100 V of applied voltage. The real emission
pattern from the ceramic heater in the 4000-1000 cm-1 region
measured with an FTIR spectrometer is given in Figure S1 in the
Supporting Information. The temperature of the heater, but not the IR
intensity, is linearly proportional to the applied AC voltage between
75 and 110 V. This IR source was placed at one of the focal points of
an ellipsoidal mirror (4; Hamamatsu Photonics, model E4169-01), and
the sample was placed at the other focal point using a reflection mirror
(R2) to conduct the IR radiation effectively onto the sample. To see
qualitative effects of IR irradiation intensity on the Boltzmann
temperatures, the output intensity of the IR source was changed by
varying the input electric (AC) voltage to the ceramic heater.
Figure 2. Schematic diagram of the system for Raman measurements under
IR illumination. (a) Scheme for the whole measurement system. 1, laser;
2, polychromator for Raman spectra; 3, ceramic heater as an IR source; 4,
ellipsoidal reflecting mirror; 5, IR band-pass filter; L1 and L2, focusing
lenses; R1; laser reflection mirror; R2, gold-coated IR reflection mirror; S,
CaF2 sample cell; NF, notch filter; PS, polarization scrambler. (b)
Transmission spectrum of the IR band-pass filter in the 1300-1900 cm-1
range, while more details are given in Figure S3. (c) The rectangular CaF2
sample cell and cell holder. A, inlet of circulating water from the
temperature-controlling equipment; B, outlet for circulating water; C,
magnetic mixer driver; D, power supply for the magnetic mixer; E, magnetic
stir bar; F, rectangular CaF2 cell; G, glass transfer; H, Teflon stopper; I,
copper block.
The band-pass filter (5; Jasco, model 6.275 µm) was placed when
necessary. This filter transmits IR light between 1730 and 1480 cm-1
with a transmittance of 50-100% but <1% in the 8000-1900 and
1400-1100 cm-1 regions. Its transmittance spectrum observed in this
study is illustrated by Figure 2b, and the more detailed spectrum
between 8000 and 350 cm-1 is given in the Supporting Information
(Figure S2). The transmittance at 1597 cm-1 is 70.05%.
As illustrated in Figure 2c, the sample cell, which was designed
specifically for this study, is a rectangular cuvette (F) made of an optical
CaF2 crystal with a cross section of 8 × 3 mm2 (inside) and a height
of 45 mm. The cuvette is held by a copper holder (I), and its temperature
is maintained by circulating temperature-controlled water from A to B
through I as shown in Figure 2c. A magnetic stirring bar (E) was placed
in the cuvette and rotated by a mixer (C) during the Raman
measurements. To examine the performance of this IR irradiation
system, we applied it to an azobenzene-containing dendrimer8 and
measured its UV-vis spectrum instead of its Raman spectrum. Its
performance was confirmed when we observed that the absorption
spectral change due to the cis-to-trans isomerization of azobenzene
occurred even at the lowest voltage (55 V) applied to the ceramic heater.
Accordingly, the present IR irradiation power at the lowest applied
voltage is sufficient to induce the reaction requiring 4.9 IR photons,8
and it can be increased by raising the applied voltage.
intensity ratios of Stokes and anti-Stokes Raman bands of (Ln-
TPP)Fe(III)Cl (n ) 3 to 5) and compare the Boltzmann
temperatures of the Fe-Cl bond, porphyrin macrocycle, and
dendritic aromatic units, in comparison with that of dioxane as
the solvent.
Experimental Section
Preparation of Samples. (LnTPP)Fe(III)Br (n ) 3 to 5) was
synthesized as previously described.10 Briefly, a tetrahydrofurane (THF)
solution containing free-base (LnTPP)H2 and FeBr2 (2-10 equiv) was
refluxed under Ar for 18 h. Completion of metal insertion was assessed
by following the change in the UV-visible spectrum for the disap-
pearance of fluorescence from (LnTPP)H2. The reaction mixture was
chromatographically separated on silica gel using CHCl3 as the eluent.
The second fraction collected was put into methanolic HCl (0.1 M).
The (LnTPP)Fe(III)Cl thus prepared was recrystallized from a CHCl3/
MeOH mixture. Dioxane solutions of (LnTPP)Fe(III)Cl were used at a
concentration of approximately 2 × 10-4 M for Raman spectroscopy.
(TPP)Fe(III)Cl was purchased from Aldrich Chemical Co. (TPP)-
Fe(III)Br was synthesized from (TPP)Fe(III)Cl via [(TPP)Fe(III)]2O,
by mixing 5 mg of (TPP)Fe(III)Cl in 50 mL of CH2Cl2 with 50 mL of
1 M aqueous NaOH solution, followed by addition of 50 mL of 1 M
aqueous HBr. The CH2Cl2 layer of the resulting mixture was collected
by a separatory funnel and evaporated to give (TPP)Fe(III)Br.
Raman Measurements. Figure 2a schematically illustrates our
Raman measurement-IR irradiation system. Raman scattering was
excited by the 413.1-nm line of a Kr+ ion laser (Figure 2a; 1), and the
scattered light from the sample (S) was collected by an objective lens
Results
To identify the axial Fe-Cl vibrational mode of (LnTPP)-
Fe(III)Cl, we utilized two nondendritic (TPP)Fe(III)X complexes
(Figure 1; X ) Cl or Br) as model compounds. Figure 3 displays
the low-frequency anti-Stokes resonance Raman spectra of
(TPP)Fe(III)X (X ) Cl or Br) in CH2Cl2 obtained with
excitation at 413.1 nm. The band marked with an asterisk arose
from the solvent. (TPP)Fe(III)Cl and (TPP)Fe(III)Br did not
show any difference in the high-frequency region from 1200 to
1700 cm-1 of the Stokes spectra (data not shown) but exhibited
a distinct difference in the low-frequency region. The 355-cm-1
band of (TPP)Fe(III)Cl (Figure 3a) disappears in the spectrum
of (TPP)Fe(III)Br (Figure 3b), but a new band appears at 237
cm-1. These two frequencies are similar to the previously
reported Fe-X stretching modes of (OEP)Fe(III)X (OEP )
octaethylporphyrin) at 606, 364, and 279 cm-1 for X ) F, Cl,
(10) (a) Jiang, D.-L.; Aida, T. Chem. Commun. 1996, 1523-1524. (b) Jiang,
D.-L.; Aida, T. J. Macromol. Sci., Pure Appl. Chem. 1997, A34, 2047-
2055.
9
10022 J. AM. CHEM. SOC. VOL. 127, NO. 28, 2005

Energy Funneling of IR Photons
A R T I C L E S
Figure 3. Anti-Stokes resonance Raman spectra of (TPP)Fe(III)X (for a,
X ) Cl and for b, X ) Br) in CH2Cl2 at an environmental temperature of
294.2 K and excitation by a Kr+ ion laser at 413.1 nm with a power of 1
mW.
Figure 5. Stokes and anti-Stokes resonance Raman spectra of L4(TPP)-
Fe(III)Cl in dioxane in the absence of an IR band-pass filter with a voltage
applied to the IR source at (a) 0 V and (b) 80 V (295.7 K, 413.1 nm, 3
mW).
observed at 390 cm-1 (porphyrin ν8) with relatively high
intensity. This mode includes primarily the Fe-N stretching
vibration and some contributions from the CR-Cm-CR defor-
mation and Cm-CR stretching.12 Virtually identical Stokes
Raman spectra were observed for lower-generation L3(TPP)-
Fe(III)Cl and higher-generation L5(TPP)Fe(III)Cl.
Figure 5 shows the simultaneously observed anti-Stokes and
Stokes resonance Raman spectra between -950 and +950 cm-1
of L4(TPP)Fe(III)Cl measured with the same experimental
conditions as those for Figure 4. (Hereafter, the anti-Stokes
region is designated by negative wavenumbers.) The anti-Stokes
bands at -355 and -390 cm-1, corresponding to the axial Fe-
Cl stretching and porphyrin in-plane vibrational modes, respec-
tively, are clearly observed but with much weaker intensity than
those at +355 and +390 cm-1. Spectra a and b reflect the
Raman intensities in the absence and presence of IR irradiation,
respectively. Their intensities apparently changed little with IR
irradiation, but precise analysis demonstrated systematic changes
as described later. The anti-Stokes solvent band at -835 cm-1
is also discernible. The extended anti-Stokes resonance Raman
spectra of L4(TPP)Fe(III)Cl obtained with the same instrumental
conditions are given in the Supporting Information (Figure S3).
Anti-Stokes Raman bands corresponding to the high-frequency
modes of porphyrin above 1000 cm-1 and the aryl ν8 mode
around 1600 cm-1 were not observed even when a voltage of
80 V was applied to the ceramic heater.
To determine the vibrational temperature, it is essential to
evaluate the area intensities of individual Raman bands correctly.
Its adopted procedure is explained in Figure 6, in which the
solid and broken lines denote the observed and fitted spectra,
respectively. The spectral fitting was performed with the
program Grams/32, version 4.14 (Spectral Notebase, Galactic
Industries Corp.). When an appropriate spectral range and the
number of bands contained therein are given, this program
makes a linear base line and fits the spectra with the given
number of Lorentzian-Gaussian mixed functions. In practice,
as shown in Figure 6, we assumed the presence of a single band
in the frequency range between -910 and -760 cm-1, five
bands between -540 and -280 cm-1, five bands between +280
and +530 cm-1, and five bands between +700 and +875 cm-1
Figure 4. Stokes resonance Raman spectra of L4(TPP)Fe(III)Cl in dioxane
in the absence of IR band-pass filter with a voltage applied to the IR source
at (a) 0 V and (b) 80 V at an environmental temperature of 295.7 K and
excitation by a Kr+ ion laser at 413.1 nm with a power of 3 mW. The
spectral background was taken out.
and Br, respectively.11 Accordingly, the 355-cm-1 band observed
in Figure 3a can be assigned to the axial Fe-Cl stretching mode
of (TPP)Fe(III)Cl.
Figure 4 shows the typical Stokes resonance Raman spectra
of L4(TPP)Fe(III)Cl in dioxane in the presence and absence of
IR irradiation. The bands marked with an asterisk originate from
the solvent, whereas the band at 322 cm-1 is due to the CaF2
cell. Other bands are assigned to modes of (TPP)Fe(III); the
detailed analysis of vibrational modes was carried out as
described by Spiro and co-workers.12 The number and positions
of peaks in the resonance Raman spectrum remained unchanged
after prolonged IR irradiation, even at an applied voltage of 80
V. In good agreement with the assignment established in the
model study, in both cases we observed a band at 355 cm-1
due to the axial Fe-Cl stretching vibrational mode. On the other
hand, a representative porphyrin in-plane vibrational band was
(11) Kitagawa, T.; Abe, M.; Kyogoku, Y.; Ogoshi, H.; Watanabe, E.; Yoshida,
Z. J. Phys. Chem. 1976, 80, 1181-1186.
(12) Li, X.-Y.; Czernuszewicz, R. S.; Kincaid, J.; Su, Y.-O.; Spiro, T. G. J.
Phys. Chem. 1990, 94, 31-47.
9
J. AM. CHEM. SOC. VOL. 127, NO. 28, 2005 10023

A R T I C L E S
Mo et al.
Figure 6. Spectral fitting for Figure 5b with Grams/32, version 4.14 (Galactic Industries Corp.). Solid lines denote the observed spectra (Figure 5b) and
broken lines indicate the fitted spectra after subtraction of the computer defined linear base line. The left and right panels represent the anti-Stokes and
Stokes spectra, respectively.
to evaluate the band intensities of the bands at -835, -390
and -355, +355 and +390, and +835 cm-1, respectively. The
solid lines in Figure 6 denote the corresponding regions of
Figure 5b in a discontinuous way, and the broken lines denote
the simulated spectra which are represented as baseline sub-
tracted spectra. The area intensities were calculated for three
bands thus derived at 355 cm-1 (Fe-Cl stretching), 390 cm-1
(porphyrin ν8 mode), and 835 cm-1 (solvent). The intensities
will be utilized later for the evaluation of their Boltzmann
temperatures. For neat solvent, only the 835 cm-1 bands were
measured under IR irradiation and demonstrated no temperature
increase under IR irradiation within the uncertainty of 0.6 °C.
obtained experimentally from the IS/IAS measurement as follows:
Tw ) (hνi/k)/ln[(ν0 + νi)3IS/(ν0 - νi)3IAS]
(2)
Although the vibrational mode dependence would not be large
in the measurements using a CW laser due to rapid vibrational
relaxation, some trends are expected to appear if the Fe-Cl
bond is specifically heated.
In the case of nonresonance Raman scattering, absorption for
scattered light is negligible and, accordingly, the observed
intensities would correspond to true intensities. In the case of
resonance Raman scattering, however, the absorption of scat-
tered light by the solution is different at ν0 + ν and ν0 - ν in
the envelope of absorption band of the solution. This effect is
larger as the separation between ν0 + ν and ν0 - ν becomes
larger and thus depends on vibrational frequency ν. Furthermore,
the scattering cross section under resonance conditions at an
excitation frequency ν0 for Stokes and anti-Stokes Raman
scattering are, strictly speaking, different (difference in Ei in
the denominator of the scattering cross section). To reduce such
effects, we selected smaller values of ν among many Raman
bands and two different modes whose ν were close to each other.
Since amounts of real absorption for the excitation light and
the scattered light by the sample solution depend on the optical
pass length through the solution (a distance between the cell
surface and scattering point) in each measurement and thus on
the geometrical arrangement of the cell, quantitative estimation
of these terms is difficult. However, to precisely determine the
Boltzmann temperature from the ratio IS/IAS, such effects should
be corrected experimentally as much as possible.
Discussion
Theoretical Base and Practical Data Treatment. The ratio
of Stokes (IS) to anti-Stokes (IAS) Raman intensities for a
nondegenerate vibrational mode (νi) in a thermal equilibrium
at a particular absolute temperature (T) can be represented by
the following equation:13
IS/IAS ) [(ν0 - νi)/(ν0 + νi)]3 exp(hνi/kT)
(1)
where ν0 is the frequency of Raman excitation light, ν0 - νi
and ν0 + νi are the absolute frequencies of Stokes and anti-
Stokes Raman bands, respectively, and h and k are Planck’s
and Boltzmann’s constants, respectively. In the physical sense,
the intensity ratio of Stokes to anti-Stokes Raman bands depends
only on the population ratio of the vibrationally excited to the
ground levels. If thermal equilibrium has not been established,
vibrational populations among different modes cannot be
determined with this equation. However, even in such a case,
the vibrational temperature of a given mode can be defined by
eq 1, and, therefore, cooling of photoexcited metalloporphyrin
was successfully estimated from the temporal dependence of
the anti-Stokes Raman intensities.14 For the sake of convenience,
we call such a temperature the Boltzmann temperature, because
it may differ from the macroscopic temperature. Accordingly,
the Boltzmann temperature (Tw) is defined for each vibrational
mode, where w denotes a vibrational frequency (νi) and can be
These effects would be the same at different temperatures
for a given scattering setting of the sample and for a given
vibrational mode. To circumvent these problems, we tried to
incorporate these effects into IS/IAS using a correction factor.
The basic assumption is that the experimental IS/IAS can be
represented by the true (IS/IAS true
)
multiplied with a correction
factor, Ci × (IS/IAS true
)
. The value of Ci depends on vibrational
mode and scattering geometry but not on temperature (applied
voltage for IR irradiation). In other words, Ci is specific to a
given experimental geometry and concentrations of sample and
therefore can be used for all temperatures under the same
(13) Kip, B. J.; Meier, R. J. Appl. Spectrosc. 1990, 44, 707-711.
(14) Mizutani, Y.; Kitagawa, T. Bull. Chem. Soc. Jpn. 2002, 75, 623-639.
9
10024 J. AM. CHEM. SOC. VOL. 127, NO. 28, 2005

Energy Funneling of IR Photons
A R T I C L E S
Figure 7. Dependence of the Boltzmann temperatures of the solvent T835
in a solution of (LnTPP)Fe(III)Cl/dioxane on the voltage applied to the IR
source. The experiment was performed in the absence (open markers) and
presence (filled markers) of an IR band-pass filter at an environmental
temperature of 295.7 K and excitation by a Kr+ ion laser at 413.1 nm with
a power of 3 mW.
Figure 8. Dependence of the Boltzmann temperatures ∆T390 (T390 - T835
)
and ∆T355 (T390 - T835) of a (LnTTP)Fe(III)Cl/dioxane solution on the
voltage applied to the IR source in the presence of IR band-pass filter at an
environmental temperature of 295.7 K. The complete sets of independent
measurements were carried out three, three, and two times for L3, L4, and
L5 samples, respectively. The error bars were determined from the scattering
of the data.
scattering geometry when Ci is determined at a certain temper-
ature. The true value, (IS/IAS true, would be obtainable for ideal
)
independent measurements for L3, L4, and L5 samples, respec-
tively, and are expected to involve the contributions from errors
in Ci. The temperature thus obtained in the absence of IR
illumination (at 0 V of applied voltage) was the same as the
environment temperature (295.7 K). It is clear that, at a
designated voltage (55-90 V) applied to the ceramic heater
and under otherwise identical conditions, the Boltzmann tem-
perature of the solvent in the absence of an IR band-pass filter
is slightly higher than that in the presence of the filter. On the
other hand, in both cases, the Boltzmann temperature of the
solvent showed a similar tendency to increase with the intensity
of IR illumination, and the maximum temperature increase of
the solvent upon IR illumination was 5 K at a voltage of 90 V
in the absence of the IR band-pass filter. Since the temperature
increase of neat solvent under the same irradiation condition
was negligible, the temperature increase of the solvent is mainly
due to the vibrational energy relaxation of the excited antenna
molecules in the solution. This inference is reasonable because
the temperature increments are diminished in the presence of
the IR band-pass filter compared to its absence. The temperature
increments of the solvent did not differ greatly among solutions
of (LnTPP)Fe(III)Cl with different numbers of aryl ether layers.
However, the temperature rise of the solvent observed in the
presence of the IR band-pass filter supports the concept of IR
energy capturing by vibrational modes of the aryl units, which
could also transfer the captured energy to the solvent.
experimental values of IS and IAS and therefore should be the
same value as that obtained from eq 1 when T ) T0 and ν ) νi.
We determined the value of Ci using the experimental value
of IS/IAS [) (IS/IAS)0] obtained from each Raman measurement
of (LnTPP)Fe(III)Cl at room temperature without IR irradiation
and the ideal true value, (IS/IAS true, calculated from eq 1 with
)
νi and T0. Accordingly, Ci was determined for each νi:
Ci ) (IS/IAS)0/(IS/IAS)true
(3)
For example, the (IS/IAS true
K (22.5 °C) are 47.41, 6.064, and 5.158 for the bands at 835,
390, and 355 cm-1, respectively, and the experimental (IS/IAS 0
)
values for T ) 294.2 + 1.5 ) 295.7
)
values for (LnTPP)Fe(III)Cl in dioxane in the absence of IR
irradiation were 15.13, 5.252, and 2.458 for bands at 835, 390,
and 355 cm-1, respectively. Accordingly, the values of Ci are
0.3192, 0.8661, and 0.4765 for the 835, 390, and 355 cm-1
bands, respectively. In the practical measurements under IR
irradiation, the true value of IS/IAS is (IS/IAS)T and the experi-
mentally obtained value of IS/IAS is (IS/IAS obs
obtained from (IS/IAS obs using the following equation for three
kinds of νi.
) . (IS/IAS)T was
)
(IS/IAS)T ) (IS/IAS)obs/Ci
(4)
Stokes and Anti-Stokes Raman Spectral Profiles. Using
these assumptions and considerations, we determined the
Boltzmann temperatures of the solvent (T835) using eq 2 for
dioxane solutions of (LnTPP)Fe(III)Cl (n ) 3 to 5) as well as
neat solvent after calculating the (IS/IAS)T from the observed
Evaluation of Boltzmann Temperatures. To determine
whether the captured energy is transferred to the chloroiron-
(III) porphyrin unit located at the center of (LnTPP)Fe(III)Cl
(n ) 3 to 5), the Boltzmann temperatures of two porphyrin bands
[390 (T390) and 355 cm-1 (T355)] were determined in the same
way as for the solvent. To eliminate a possible contribution from
heating effects by other sources, the results were examined in
the form of differential temperatures by subtracting the Boltz-
mann temperature of the solvent (T835) from T390 and T355: ∆T355
) T355 - T835 and ∆T390 ) T390 - T835. Figures 8 and 9 show
the plots of ∆T355 and ∆T390 versus the voltage applied to the
ceramic heater, which were measured in the presence and
(IS/IAS obs
values of the 835-cm-1 band in the presence and
)
absence of the IR band-pass filter. The Boltzmann temperature
for the neat solvent with IR irradiation at 80 V without the IR
filter was 294.1 ( 0.6 K, indicating no direct heating effect by
the ceramic heater. Figure 7 shows the Boltzmann temperatures
of solvent plotted against the voltage applied to the ceramic
heater. The values contain the contribution of heating (1.5 K)
by the Raman excitation light at 3 mW. The error bars represent
the magnitude of data scattering in three, three, and two
9
J. AM. CHEM. SOC. VOL. 127, NO. 28, 2005 10025

A R T I C L E S
Mo et al.
aryl units may be transferred to the porphyrin ring mode with
the closest frequency, such as the one at 1555 cm-1 (see Figure
4). However, as noted for Figures 5 and S3, there was no anti-
Stokes band observed in the region from 1500 to 1650 cm-1
.
Instead, we observed the anti-Stokes bands for the axial Fe-
Cl stretching and porphyrin in-plane vibrational modes at 355
and 390 cm-1, respectively. The mode dependency of the
temperature increase (∆T355 > ∆T390) is compatible with the
idea that the vibrational energy captured by the dendrimer
framework is transferred to the axial position of the core
porphyrin unit first and then promptly relaxed to in-plane
vibrational modes. Because the vibrational energy redistribution
finishes within a few picoseconds for metalloporphyrins,14 the
temperatures of the Fe-Cl stretching and porphyrin in-plane
vibrational modes are usually the same when the IR pumping
rate is slower than the rate of vibrational energy redistribution
in the photo steady-state measurements. Furthermore, the
Boltzmann temperatures of the solute should be the same as
that of the solvent if intermolecular vibrational energy transfer
is faster than the IR pumping rate (∆T ) 0). Interestingly, both
∆T355 and ∆T390 are positive, meaning that the observed
temperature changes are caused by some systematic process,
and the ∆T355 thus obtained is still larger than ∆T390 (7.5 versus
3.8 K in the presence of the IR filter and 11.5 versus 5.7 K in
the absence of the IR filter). The difference between ∆T355 and
∆T390 would be much larger if picosecond time-resolved Raman
measurements were performed using a picosecond IR pump
source.
Figure 9. Dependence of the Boltzmann temperatures ∆T390 (T390 - T835
)
and ∆T355 (T390 - T835) of (LnTTP)Fe(III)Cl/dioxane solution on the voltage
applied to the IR source in the absence of an IR band-pass filter at an
environmental temperature of 295.7 K. The complete sets of independent
measurements were carried out three, three, and two times for L3, L4, and
L5 samples, respectively. The error bars were determined from the scattering
of the data.
absence of the IR band-pass filter, respectively. We found the
following general trends:
(1) All of the ∆T355 and ∆T390 values are positive, suggesting
that the Boltzmann temperatures of the solute (T390 and T355
)
are higher than that of the solvent (T835).
(2) The ∆T355 and ∆T390 values show a weak but clear
tendency to increase with the voltage applied to the ceramic
heater.
(3) At a designated voltage applied to the ceramic heater,
the ∆T355 and ∆T390 values increase with the number of the
aryl ether layers (n) of the dendrimer framework.
The proportion of reduction of ∆T by the IR filter is 65%
() 7.5/11.5) and 67% () 3.8/5.7) for the 355 and 390 cm-1
bands, respectively. These values happen to be very close to
the value of transmittance of the IR filter at 1597 cm-1 (70.05%).
Since the ν8 IR absorption of the dendrimer framework has some
width and the transmittance of the filter changes with the
frequency, the precise estimation of the expected value is not
straightforward. However, it can be concluded that other
dendrimer modes (such as CH2-O str. at 1155 cm-1) do not
play a larger role in the heating. If the CH2-O mode were
involved, the measurements with the filter should include the
contribution of removing this large absorber, and thus the ∆T
change should be much greater than 65%. Consequently, the
present observations strongly suggest that IR energy contributing
to temperature rise of the core porphyrin is mainly captured by
the ν8 mode of the dendrimer faramework. Such a trend is
observed for all members of the (LnTPP)Fe(III)Cl family.
On the other hand, if the phenyl ν8 mode (1597 cm-1) of the
dendrimer framework is specifically excited by IR irradiation
and its vibrational relaxation is relatively slow, its anti-Stokes
Raman band is expected to appear at 1597 cm-1. Therefore,
we attempted to find an anti-Stokes band under IR irradiation
in the region between 1500 and 1650 cm-1. However, as
described earlier (Figures 5 and S3), we failed to detect a trace
of such a peak in this frequency region, even upon IR irradiation
at a voltage of 80 V applied to the ceramic heater. These results
indicate that the vibrational energy transfer from the aryl
dendritic antenna to the core porphyrin unit must be extremely
fast.
(4) For each dendrimer chloroiron(III) porphyrin complex,
the temperature rise for the Fe-Cl stretching mode is always
higher than that for the porphyrin ν8 mode (∆T355 > ∆T390),
regardless of whether it is measured in the presence or absence
of the IR band-pass filter.
(5) The difference between ∆T355 and ∆T390 enlarges when
the generation of the dendrimer framework is increased.
(6) The ∆T355 and ∆T390 values in the presence of the band-
pass filter are smaller than those evaluated in the absence of
the filter. However, the proportion of reduction is compatible
with the idea that IR energy is primarily captured by the aryl
ν8 vibration of the dendrimer framework as discussed below
quantitatively. The maximum values of ∆T355 observed for (L5-
TPP)Fe(III)Cl in the presence and absence of the filter were
7.3 and 11.6 K, respectively.
Mechanistic Aspects. These findings indicate that the aryl
ether dendrimer framework in (LnTPP)Fe(III)Cl absorbs IR
photons through the phenyl ν8 mode at 1597 cm-1 and transfers
their energy to the central porphyrin unit. As shown in Figure
1, larger dendrimers contain an increased number of aryl units
in their framework. Therefore, this property is in good agreement
with the results from the harvesting of IR photons through the
aryl units, as demonstrated by the dendrimer size effect;
specifically, (LnTPP)Fe(III)Cl with a large dendrimer framework
has a large temperature increase that is even greater under high-
intensity IR illumination.
The fact that ∆T355 exceeded ∆T390 for all independent
measurements leads us to conclude that the IR energy captured
by the aryl dendritic antenna is accumulated on the axial ligand
From the general viewpoint of vibration-to-vibration reso-
nance energy transfer, the vibrational quantum excited in the
9
10026 J. AM. CHEM. SOC. VOL. 127, NO. 28, 2005

Energy Funneling of IR Photons
A R T I C L E S
position of the central chloroiron(III) porphyrin unit. At present,
we cannot explain why this occurs, although we note that this
observation is consistent with the IR-induced cleavage of the
axial ligand of oxygenated dendrimer iron(II) porphyrin com-
plex.10 Further study is necessary to unravel this process. Such
an accumulation feature was thought unusual at that time, but
as demonstrated in this study, it was confirmed by a quite
different technique, that is, the direct measurement of temper-
ature at the reaction point under IR irradiation. The difference
between ∆T355 and ∆T390 increases with increasing dendrimer
size while both the ∆T355 and ∆T390 values become larger,
presumably because the more confining environment around
the axial Fe-Cl bond in a larger dendrimer would protect the
energy transfer from collision scattering and would insulate the
Fe-Cl bond from thermal perturbation. This, in turn, would
suppress the vibrational energy relaxation to the porphyrin
macrocycle.
was in agreement with the transmittance of IR light at 1597
cm-1. This indicates that the energy of IR photons captured
mainly by the ν8 mode of the aryl dendritic antenna is transferred
to the axial position of iron and subsequently relaxed to the
porphyrin macrocycle. Thus, the present study not only dem-
onstrates a new IR illumination effect on dendrimer metallopor-
phyrins but also provides a novel approach to evaluate IR energy
funneling in light-harvesting dendritic macromolecules.
Acknowledgment. We thank Prof. Hiroshi Fujii of the
Okazaki Institute for Integrative Bioscience, Japan, for his help
in the preparation of (TPP)Fe(III)Br, and Dr. Hirotsugu Hira-
matsu of this laboratory for determining the transmittance
spectrum of the IR filter. This work was supported by a Grant-
in-Aid for Specifically Promoted Research from the Ministry
of Education, Culture, Sports, Science, and Technology, Japan,
to T.K. (14001004).
Supporting Information Available: Relative emission inten-
sity of the ceramic heater in the 4000 to 0 cm-1 region when
AC 100 V is applied, which was observed with a Jasco FT/IR-
6000 (vacuum type) (Figure S1), the transmission spectrum of
the IR filter in the wavenumber range between 8000 and 350
cm-1, in which the inset represents the expanded spectrum
between 2000 and 400 cm-1 (Figure S2), and anti-Stokes
resonance Raman spectra of L4(TPP)Fe(III)Cl in dioxane in the
absence of an IR band-pass filter with a voltage applied to the
IR source at (a) 0 V and (b) 80 V at an environmental
temperature of 295.7 K and excitation by a Kr+ ion laser at
413.1 nm with a power of 3 mW (Figure S3). This material is
available free of charge via the Internet at http://pubs.acs.org.
Conclusions
By using anti-Stokes Raman spectroscopy, we determined
the Boltzmann temperatures of (LnTPP)Fe(III)Cl in dioxane for
two vibrational bands of the central chloroiron(III) porphyrin
functionality under variable intensity IR illumination. The
Boltzmann temperatures of (LnTPP)Fe(III)Cl modes (390 and
355 cm-1) were higher than that of the solvent (835 cm-1),
while the temperature of neat solvent (835 cm-1) was unaffected,
and the temperature rise was greater for dendrimers with a large
framework and increased with increasing IR intensity. Between
the two porphyrin vibrational modes with similar frequencies,
the Fe-Cl stretching mode was more heated than an in-plane
mode of the porphyrin macrocycle. On the other hand, the
proportion of reduction of the temperature rise by the IR filter
JA042196Z
9
J. AM. CHEM. SOC. VOL. 127, NO. 28, 2005 10027