Light-harvesting and ultrafast energy migration in porphyrin-based metal-organic frameworks

Article abstract of DOI:10.1021/ja310596a

Given that energy (exciton) migration in natural photosynthesis primarily occurs in highly ordered porphyrin-like pigments (chlorophylls), equally highly ordered porphyrin-based metal-organic frameworks (MOFs) might be expected to exhibit similar behavior, thereby facilitating antenna-like light-harvesting and positioning such materials for use in solar energy conversion schemes. Herein, we report the first example of directional, long-distance energy migration within a MOF. Two MOFs, namely F-MOF and DA-MOF that are composed of two Zn(II) porphyrin struts [5,15-dipyridyl-10,20-bis(pentafluorophenyl)porphinato]zinc(II) and [5,15-bis[4-(pyridyl)ethynyl]-10,20-diphenylporphinato]zinc(II), respectively, were investigated. From fluorescence quenching experiments and theoretical calculations, we find that the photogenerated exciton migrates over a net distance of up to ～45 porphyrin struts within its lifetime in DA-MOF (but only ～3 in F-MOF), with a high anisotropy along a specific direction. The remarkably efficient exciton migration in DA-MOF is attributed to enhanced π-conjugation through the addition of two acetylene moieties in the porphyrin molecule, which leads to greater Q-band absorption intensity and much faster exciton-hopping (energy transfer between adjacent porphyrin struts). The long distance and directional energy migration in DA-MOF suggests promising applications of this compound or related compounds in solar energy conversion schemes as an efficient light-harvesting and energy-transport component.

Full text of DOI:10.1021/ja310596a

Article
pubs.acs.org/JACS
Light-Harvesting and Ultrafast Energy Migration in Porphyrin-Based
Metal−Organic Frameworks
Ho-Jin Son,†,∥ Shengye Jin,‡,∥ Sameer Patwardhan,† Sander J. Wezenberg,†,# Nak Cheon Jeong,†,§
Monica So,† Christopher E. Wilmer,⊥ Amy A. Sarjeant,† George C. Schatz,† Randall Q. Snurr,⊥
Omar K. Farha,*,† Gary P. Wiederrecht,*,‡ and Joseph T. Hupp*,†
⊥
†Department of Chemistry, and Chemical and Biological Engineering, Northwestern University, Evanston, Illinois 60208, United
States
‡Center for Nanoscale Materials, Argonne National Laboratory, Argonne, Illinois 60439, United States
§Department of Emerging Materials Science, DGIST, Daegu, Korea
S
* Supporting Information
ABSTRACT: Given that energy (exciton) migration in natural
photosynthesis primarily occurs in highly ordered porphyrin-like
pigments (chlorophylls), equally highly ordered porphyrin-based
metal−organic frameworks (MOFs) might be expected to exhibit
similar behavior, thereby facilitating antenna-like light-harvesting and
positioning such materials for use in solar energy conversion
schemes. Herein, we report the first example of directional, long-
distance energy migration within a MOF. Two MOFs, namely F-
MOF and DA-MOF that are composed of two Zn(II) porphyrin
struts [5,15-dipyridyl-10,20-bis(pentafluorophenyl)porphinato]zinc-
(II) and [5,15-bis[4-(pyridyl)ethynyl]-10,20-diphenylporphinato]-
zinc(II), respectively, were investigated. From fluorescence quench-
ing experiments and theoretical calculations, we find that the
photogenerated exciton migrates over a net distance of up to ∼45
porphyrin struts within its lifetime in DA-MOF (but only ∼3 in F-MOF), with a high anisotropy along a specific direction. The
remarkably efficient exciton migration in DA-MOF is attributed to enhanced π-conjugation through the addition of two acetylene
moieties in the porphyrin molecule, which leads to greater Q-band absorption intensity and much faster exciton-hopping (energy
transfer between adjacent porphyrin struts). The long distance and directional energy migration in DA-MOF suggests promising
applications of this compound or related compounds in solar energy conversion schemes as an efficient light-harvesting and
energy-transport component.
crystalline MOFs may provide a basis for systematically relating
structure and composition to photon capture and energy
transport and delivery. At the same time, the ability with
frameworks to position chromophores proximal to each other,
yet not in direct physical contact, can prevent self-quenching
and deleterious attenuation of exciton lifetimes. A recent study
by Lin, Meyer, and co-workers reported facile intracrystal site-
to-site energy migration dynamics in Ru(II)/Os(II) (2,2′-
bipyridine)3-derived MOFs through luminescence quenching
measurements.14−17 Their work has clearly demonstrated the
potential of appropriately designed MOFs to function as light-
harvesting and energy-delivery structures. Key to achieving
facile migration in the Ru-based MOFs has been long excited-
state lifetimesa feature absent in natural photosynthetic
systems
INTRODUCTION
■
In natural photosynthesis, a core function is the sequential
migration of photoinduced energy (exciton) in light-harvesting
pigments (i.e., chlorophylls and carotenoids).1 Highly ordered
structures in antenna are composed of chlorophyll and
carotenoid pigments. For example, in purple photosynthetic
bacteria, highly symmetric wheel-like supramolecular architec-
tures play an essential role in the absorption of light and the
subsequent energy funneling to the reaction centers.2−4 These
nanometer-sized and three-dimensional structures have moti-
vated chemists to design artificial light-harvesting antenna
assemblies such as chromophoric polymers,5 dendrimers,6−8
covalently linked porphyrin arrays,9 and self-assembled supra-
molecular systems.10−13
Recently, metal−organic frameworks (MOFs), which are
hybrid materials made from organic linkers and inorganic
nodes, have begun to receive attention as alternative
compounds for configuration as light-harvesting systems. The
high degree of chromophore ordering achievable within
Received: October 27, 2012
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja310596a | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Given that energy migration in natural photosynthesis
occurs, in part, through highly ordered porphyrin-like pigments
(chlorophylls), we reasoned that crystalline MOFs assembled
from photoactive porphyrins might exhibit light-harvesting
behavior reminiscent of that of natural photosynthetic systems.
However, the use of porphyrin-based MOFs as functional light
harvesters (not just absorbers) has not yet been significantly
explored. We and our co-workers, as well as others, have
previously reported the successful development of porous
metalloporphyrin frameworks that can function as cata-
lysts.18−26 We have also recently reported that efficient,
single-step excitonic energy transfer can be observed in a
MOF containing a porphyrin and a complementary chromo-
phore as linkers.27 However, important questions, such as how
far molecular excitons can migrate, what factors determine the
efficiency of the exciton migration, and to what extent excitons
can migrate directionally, have not been answered, or even
examined, for porphyrinic MOFs. Exploration of these issues is
critical for designing and utilizing porphyrin-based MOFs for
efficient light-harvesting and energy deliveryfor example, to
redox-active chemical catalysts.17
the extent of quenching with various levels of FcPy
incorporation.28
Based on quenching measurements, we find that the exciton
can visit ∼90 porphyrin linkers in DA-MOF and ∼6 in F-MOF
within the brief (ca. nanosecond) exciton lifetime. Consistent
with the comparatively low symmetry of both the chromo-
phoric and non-chromophoric linkers, detailed computational
studies reveal, for both MOFs, that exciton migration is
substantially anisotropic. To the extent that exciton transport
can be channeled along a single direction, the net distance for
energy transport (exciton displacement) can be increased.
Directional and long-distance exciton migration that can be
rationally modulated is likely to be an important element in the
design of functional molecule-derived solar energy conversion
systems. The results of the present study provide useful MOF
and linker design guidelines for engendering such behavior.
EXPERIMENTAL SECTION
■
Structure Characterization of DA-MOF and F-MOF. Several
attempts were made to determine the single-crystal X-ray structure of
DA-MOF. Although consistent results, based on several syntheses,
were obtained for unit cell parameters, we were unable to obtain
sufficient data for full structure determination. The unit-cell
dimensions (11.47(3) × 15.589(3) × 27.170(6) Å) are appropriate
for the proposed framework. With these findings in hand, the
hypothetical MOF generator described by Wilmer and co-workers29
was used to construct a MOF structure. Three building blocks were
used as input: the dipyridylporphyrin linker (DA-H2P), inorganic ZnII-
paddlewheel nodes, and Zn(NO3)2·6H2O, 1,2,4,5-tetrakis(4-
carboxyphenyl)benzene (TCPB). In the actual MOF, the porphyrin
is likely capable of rotating on its dipyridyl axis; its equilibrium
orientation, however, is unknown to us.30 For the purpose of
generating the hypothetical structure, the plane of the porphyrin was
chosen to be at 43.8° to the plane defined by the Zn atoms and four of
the oxygen atoms of the metal corner. This angle is typical of MOFs
with dipyridyl ligands and ZnII-paddlewheel nodes that have been
characterized by single-crystal X-ray diffraction.18 The unit cell of the
generated structure (11.59 × 15.01 × 27.51 Å) is in an excellent
agreement with the experimentally obtained unit cell. Additionally, the
simulated powder X-ray diffraction (PXRD) pattern is in excellent
agreement with the experimental pattern (see Figure S5). Single-
crystal X-ray structural characterization of F-MOF has been described
previously.18 PXRD data for F-MOF are shown in Figure S6.
Herein, we report the design, synthesis and characterization
of two MOFs (F-MOF and DA-MOF) constructed from zinc
metalated porphyrin (ZnP) molecules. The efficacy of singlet
exciton migration in the MOFs is investigated via fluorescence
quenching measurements and the migration anisotropy is
examined by theoretical calculations. In Figure 1 we show a
Figure 1. (a) Photograph of a DA-MOF particle from which
fluorescence is recorded based on laser excitation at 446 nm.
Luminescence quenching measurements are conducted to probe the
energy (exciton) migration dynamics. (b,c) Schematic representation
of the exciton migration and quenching processes. Upon illumination,
the photogenerated exciton can hop between adjacent porphyrin units
until it reaches a quenching site. In the red ellipse is an expanded view
of the quenching site where the quencher (FcPy) is bound to the
porphyrinic Zn center; the exciton is quenched via electron transfer
from FcPy to the porphyrin entity (ZnP).
FcPy Doping in MOFs. In non-ligating solvents, pyridine (Py) or
FcPy readily binds to porphyrinic Zn(II) sites. (The binding constants
for FcPy and Py to free porphyrin molecules were found to be 4.5 ×
104 and 1.96 × 104 M−1, respectively, based on the UV−vis titration
measurements (See Figure S7).) To incorporate FcPy in a controlled
fashion, MOF samples were soaked in various FcPy-containing
solutions where the sum of the FcPy and Py concentrations is 0.01
M (CFcPy + CPy = 0.01 M). Thus, all porphyrinic Zn(II) sites are
ligated. (Note that zinc ions constituting nodes are already
coordinatively saturated, so are incapable of functioning as secondary
FcPy binding sites.) Since the energies of porphyrin excited-states can
be altered slightly via axial coordination of the metal center, this
protocol ensures that the energies of the fluorescing states of all
porphyrin subunits are essentially identical. Consequently, exciton
traps should be absent from these crystalline ensembles. The amount
of doped FcPy in the MOF was adjusted by changing the fraction of
FcPy in the soaking solution. To quantify the FcPy loading, we
determined, via inductively coupled plasma-optical emission spectros-
copy (ICP-OES), the mole ratio of Fe to Zn. Estimates of FcPy
loading based on the binding constants and on competitive binding
were found to agree well with values obtained directly from ICP-OES.
(Additionally, the ICP data serve to rule out significant binding to
alternative sites such as coordinatively defective MOF node sites.)
Fluorescence Quenching and Lifetime Measurements. MOF
fluorescence quenching measurements were carried out with a home-
photo of a DA-MOF crystallite and a schematic representation
of the fluorescence quenching measurements. Illumination of
the MOF generates a singlet excition that can migrate by site-
to-site (linker-to-linker) hopping within the MOF until it
decays (for example, by fluorescence) or is quenched via an
encounter with another species. In our experiments, the exciton
can be quenched by electron transfer to the excited linker from
an axially ligated donor molecule (pyridyl-ferrocene, FcPy). If
the donor is present at only a fraction of the available sites,
rapid exciton migration results in amplified fluorescence
quenching. The efficacy of exciton migration relative to
nonreactive exciton decay can be determined by measuring
B
dx.doi.org/10.1021/ja310596a | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
built scanning confocal microscope based on an Olympus IX-71
inverted microscope. Samples were excited with a CW laser (Coherent
CUBE) at 446 nm. Samples, from which solvent had first been allowed
to evaporate, were placed on glass coverslips on a piezo scanner and
the excitation beam (<15 nW) was focused through an objective (60×,
NA 1.35, oil immersion, Olympus). The resulting epifluorescence was
collected by the same objective. A band-pass filter (630 to 730 nm)
was placed in the collection path to block stray pump light, with the
fluorescence ultimately being detected by a fiber-coupled single-
photon avalanche photodiode (SPAD, Micro Photon Devices). The
output of the SPAD was recorded and analyzed by using a time-
correlated single photon counting (TCSPC, PicoQuant) module as
counts of photons per second. Low excitation powers (<15 nW) were
used so as to avoid exciton−exciton annihilation (self-quenching; see
Figure S8 in Supporting Information for the details).
Fluorescence lifetime measurements were carried out with the same
confocal microscope. The samples were excited by frequency-doubled
pulses from a Ti:sapphire laser (Coherent Micra) at 400 nm with a
repetition rate of 5 MHz (after a pulse picker) and an excitation power
of ∼300 nW. The output of the SPAD detector traveled through the
same optical filter and recorded and analyzed by a PicoQuantTCSPC
module.
strongly allowed than in the electronically more symmetrical F-
H2P molecule. Details of the synthesis of DA-H2P are
presented in Scheme S1 in the Supporting Information.
As shown in Scheme 1, the DA-H2P-based MOF (DA-
MOF) was synthesized in high yield via standard solvothermal
methods in diethylformamide (DEF) using TCPB, and DA-
H2P in the presence of HNO3. The F-H2P-based MOF (F-
MOF) was assembled using a recently reported method.18 The
MOF nodes consist of pairs of paddlewheel (carboxylate)
coordinated Zn(II) ions that are additionally axially ligated by
pyridyl functionalities. Crystallites for both MOFs are plate-like,
with lateral dimensions on the order of 100−200 μm. As shown
in the photographs in Scheme 1, F-MOF and DA-MOF are red
and green, respectively.
Shown in Figure 2 are electronic absorption and emission
spectra of free F-ZnP and DA-ZnP molecules in dimethylfor-
RESULTS AND DISCUSSION
Initial Characterization. The synthesis of the studied
MOFs is presented in Scheme 1. The construction of the
■
Scheme 1. Synthesis Routes of the Isostructural DA-MOF
and F-MOF Compounds and Photographs of Resulting
Crystals
MOFs starts from one of two free-base porphyrin molecules:
unsymmetrical [5,15-dipyridyl-10,20-bis(pentafluorophenyl)]-
porphyrin (designated F-H2P) or [5,15-di(4-pyridylacetyl)-
10,20-diphenyl]porphyrin (designated DA-H2P). These por-
phyrins were chosen, in part, because they yield non-catenated
and fully reactant-accessible active site MOFs, and also because
they offer structural and chromophoric similarity to various
natural chlorophylls. The synthesis of F-H2P followed a
previous report,18 while DA-H2P was newly designed for this
work. Compared with F-H2P, the structure of DA-H2P is
extended by addition of acetylene moieties at the 5- and 15-
positions. The acetylenes enable the pendant pyridine groups
to achieve coplanarity with the tetrapyrrole core and thereby
extend the molecule’s conjugation. In contrast, the phenyl
groups at the 10 and 20 positions are not strongly electronically
coupled to the porphyrin core. A consequence of the coupling
differences is a substantial reduction in electronic symmetry,
such that absorption in the Q-band region becomes more
Figure 2. UV−vis absorption (blue) and emission (red) spectra of F-
ZnP (a) and DA-ZnP (b) molecules in DMF. Their chemical
structures are shown in the inset.
mamide (DMF). Notably, absorption in the Q-band region
(long wavelength region) is substantially red-shifted, as well as
intensified, for DA-ZnP, relative to F-ZnP. Specifically, the DA-
ZnP absorption coefficient (ε) is 376 000 at 448 nm and 51
000 at 658 nm, while the photoluminescence quantum yield
(ΦDA) is 0.096. For F-ZnP, ε = 273 000 at 448 nm and ε =
8400 at 555 nm, with ΦF = 0.031. Similar energy and intensity
effects, attributable to differences in conjugation and extent of
electronic asymmetry, have been observed for related
porphyrins.31 Importantly, DA-ZnP is characterized by a
much larger overlap between emission and absorption spectra
than is F-ZnP. As one might anticipate from Forster theory,
̈
C
dx.doi.org/10.1021/ja310596a | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
and as discussed further below, the differences in spectral
overlap translate into sizable differences in energy migration
dynamics.
Thermogravimetric analysis (TGA) data for the as-
synthesized sample of DA-MOF revealed 35−40% weight
loss due to solvent molecules, thus implying substantial
porosity (Figure S9 in the Supporting Information). The two
MOFs are similar, with TCPB species coordinated to pairs of
ZnII ions to form two-dimensional sheets that are pillared by
either F-ZnP or DA-ZnP. Figure 3 shows space-filling models
Figure 3. Space-filled (left) and capped stick (right) representations of
(a) DA-MOF and (b) F-MOF showing channels in three directions
(yellow = Zn, red = O, green = F, blue = N, gray = C, black = H). The
DA-MOF model was constructed by using the hypothetical MOF
generator described by Wilmer and co-workers.29.
of the structure of DA-MOF and F-MOF along the various
channels. The structures of both reveal large channels in three
directions. The Zn-to-Zn distance between cofacial ZnPs is
11.60 Å for DA-MOF and 11.57 Å for F-MOF, while the
distance between in-plane (collinear) ZnPs is 27.51 Å for DA-
MOF and 22.12 Å for F-MOF. DA-MOF features a slightly
larger channel height due to the use of a longer porphyrin pillar.
Energy Migration Efficiency in MOFs. Amplified photo-
luminescence quenching, due to comparatively rapid exciton
migration, has been previously reported in MOFs,17 porphyrin
clay systems32 and conjugated polymers.33,34 Thus, sufficiently
rapid exciton migration within a MOF should enable dramatic
fluorescence quenching to be achieved using only small
amounts of quencher (see Figure 1). Correlations between
the extent of fluorescence quenching and the concentration or
amount of quencher can therefore be used to quantify exciton
migration efficiency. In previously reported studies, quenching
measurements have typically been carried out in the presence of
a surrounding solvent, with the quencher present only on the
external surfaces of the MOF or conjugated polymer. In this
study, FcPy quenchers were incorporated directly inside the
MOF structures.35,36 A list of DA-MOF and F-MOF samples
with different FcPy doping ratios ranging from ∼0.005 to 1 for
fluorescence quenching measurements is provided in Table S1.
Figure 4a shows the steady-state fluorescence intensity of
DA-MOF and F-MOF as a function of FcPy doping ratios. For
comparison, a simulated fluorescence quenching curve without
exciton migration is also plotted. As shown in Figure 4a, both
Figure 4. (a) Steady-state fluorescence intensities of F-MOF (closed
blue circles) and DA-MOF (closed red squares) as a function of extent
of FcPy loading (from 0 to 1). The linear gray plot shows the behavior
expected when exciton migration is absent and quenching is
exceptionally fast relative to the exciton lifetime. (b) The
corresponding modified Stern−Volmer plots, based on eq 3. Presented
in the inset are Stern−Volmer plots based on eq 2.
F-MOF and DA-MOF exhibit amplified fluorescence quench-
ing, implying significant exciton migration. The amplification
behavior, however, is substantially greater for DA-MOF.
Remarkably, ca. 50% of the fluorescence of DA-MOF is
quenched when only 0.5% of the porphryin linkers are ligated
by FcPy. For F-MOF, 50% quenching requires ligation ca. 20%
of the porphyrin units by FcPy to achieve a similar extent of
quenching.
The fluorescence quenching in the two MOF samples
plateaus after reaching a certain ratio of FcPy, as shown in
Figure 4a. This flat level is attributed to the saturation of
fluorescence quenching that starts at the ratio of ∼0.15 for DA-
MOF and ∼0.60 for F-MOF. This phenomenon is similar to
the luminescence quenching behavior of phosphorescent
MOFs.16 The saturation of quenching occurs when the
quenching process (electron transfer from FcPy to the excited
porphyrin) is not fast enough to completely eliminate the
fluorescence. Based on the saturation fluorescence intensity the
D
dx.doi.org/10.1021/ja310596a | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
electron transfer rate from FcPy to porphyrin can be estimated
according to eq 1,
Energy Migration Anisotropy in MOFs. The exciton
migration anisotropy in DA-MOF and F-MOF can be
evaluated through theoretical calculations. The four possible
exciton directions (from A to B, C, D, or E) with corresponding
intermolecular (Zn−Zn) distances have been depicted in
Figure 5 for both MOFs. Nominally one would expect the
I0
Is
Φ0
=
= 1 + τ0ke
Φ
(1)
s
where I0, Φ0, and τ0 are the intrinsic fluorescence intensity,
quantum yield, and lifetime without quenchers, respectively; Is
and Φs are the saturated fluorescence intensity and quantum
yield. ke is the electron transfer (quenching) rate from FcPy to
porphyrin, which was estimated to be ∼4 × 109 s−1 in DA-
MOF and ∼1.6 × 109 s−1 in F-MOF.
To quantify the exciton migration efficiency, quenching data
for DA-MOF and F-MOF were analyzed using the Stern−
Volmer (SV) relationship,
I0
Is
= 1 + KSV[Q]
(2)
where I is the fluorescence intensity at at a particular level of
quencher doping, KSV is the SV quenching constant, and [Q] is
the quencher doping ratio (i.e., the fraction of linkers featuring
an axially ligated FcPy molecule). The SV plots, shown in the
inset in Figure 4b, present a roughly linear dependence in the
region with FcPy ratio less than 0.16 for DA-MOF and less
than 0.6 for F-MOF, but deviate from this dependence and
reach a flat level due to the saturation of fluorescence
quenching at higher FcPy ratios. The presence of residual
fluorescence at the saturation level reduces the value of
fluorescence amplification (I0/I),and therefore we modified the
SV relationship by subtracting the saturated fluorescence
intensity Is, as in eq 3:
Figure 5. Schematic presentations of the structural configurations for
F-MOF (a-1) and DA-MOF (b-1). The letters on the porphyrin
blocks indicates the 4 nearest neighbors B, C, D and E around A. The
capped stick representations of the crystal structure of F-MOF (a-2)
and DA-MOF (b-2) with arrows indicating the four energy transfer
directions from A to B, C, D and E between the nearest neighboring
porphyrin blocks. The center-to-center distances between pairs of
porphyrin molecules along each direction are also indicated beside the
arrows.
I0 − Is
I − Is
= 1 + KSV[Q]
(3)
Modified SV curves, with only the linear section plotted for
the two MOFs, are shown in Figure 4b. Since the quencher is
directly doped into the MOF structure, the fluorescence
quenching should not follow a diffusional controlled mecha-
nism, but a quenching mechanism with KSV = Nt, where Nt is
the number of the porphyrin sites that the exciton visits within
its lifetime. Fits of the plots with eq 3 yield KSV values of 91 for
DA-MOF and 5.6 for F-MOF, indicating that an exciton visits,
on average, about 90 distinct porphyrin sites in DA-MOF and 6
sites in F-MOF within its lifetime.
direction with the highest exciton-hopping rate to be the
preferred exciton migrating direction. The rate of exciton
hopping between adjacent porphyrin linkers along each
direction is calculated using the Forster energy transfer
̈
equation (eq 4).
The above quenching measurements indicate that the energy
migration efficiency in DA-MOF is much higher than that in F-
MOF. As discussed further below, in the low-power limit the
fluorescence lifetimes of the two photoexcited MOFs are
similar; consequently, the differences in migration efficiency can
be largely understood in terms of differences in the dynamics of
migration. Energy migration in MOFs or conjugated polymers
involves sequential hopping of excitons between adjacent
chromophores, the rate of which (for singlet excitons) is
2π
kEnT
=
J2OI
(4)
ℏ
According to this equation the rate constant (kEnT) for exciton
hopping is governed by two key parameters: the overlap
integral (OI) between the normalized absorption and emission
spectra and the exciton coupling (J).
The overlap integrals for DA-MOF and F-MOF, calculated
from the corresponding experimental absorption and fluo-
rescence spectra of non-aggregated monomers in solution, are
0.57 and 0.19 eV−1, respectively. Two low-lying excited states
(ES1 and ES2) overlap in the Q-band region and both excited
states will participate in the energy transfer process. The
oscillator strength (os) of the Q-band, which is indirectly
related to the exciton coupling of eq 4 (os ∝ μ2 ∝ J), is higher
for the porphyrin strut in DA-MOF (0.21) than for the strut in
F-MOF (0.06).
describable by the Forster energy transfer model17,37 (albeit,
̈
sometimes with extensions beyond the point-dipole model; see
Supporting Information for details of the approach used here).
The hopping rate for singlet excitons depends on the spectral
overlap between emission and lowest-energy (singlet)
absorption bands of the molecular building blocks, and on
the separation distance (r) between the energy donor and
acceptor. In the point-dipole limit, the hopping rate scales as 1/
r6.
E
dx.doi.org/10.1021/ja310596a | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Exciton couplings were evaluated using atomic transition
densities.38−40 (See Supporting Information for details of the
approach used.) When chromophores are poorly approximated
by spheres, this approach yields much more reliable estimates
than does a point-dipole approximation. Couplings associated
with the exciton transitions ES1−ES1, ES2−ES2, and ES1−ES2
(cross) have been calculated. Because internal relaxation is fast,
coupling involving higher energy states is unimportant. The
effects of disorder of porphyrin units around the dipyridyl axes
have been taken into account by considering and then summing
(with equal population weighting) contributions from each of
several rotational conformations. The calculated exciton
hopping rates along different directions are assembled in
Table 1. In DA-MOF the hopping rates for the second excited
average, be 452 = 2025. For F-MOF, the total number of hops
will be 2.82 or about 8.
The average number of hops (both forward and backward)
made along each of four possible directions (AB, AC, AD, and
AE) can be estimated from the hopping rate in a designated
direction relative to the sum of hopping rates in all possible
directions, times the total number of hops made. For example,
for DA-MOF the fraction of hops occurring along AB is
k
EnT(AB)/{kEnT(AB) + kEnT(AD) + kEnT(AE) + kEnT(AC)} =
0.55. The absolute number of AB-directed hops is 0.55 × 2025
= 1114 hops. The root-mean-square displacement of the
exciton due to hops along the AB direction is (1114)1/2 = 33
hops. Table 2 lists the full set of estimates of directional
Table 2. Net Exciton Migration Displacement in the Number
of Hops and Distances (Å) along the Four Directions AB,
AD, AE, and AC in F-MOF and DA-MOF
Table 1. Calculated Forster Energy Transfer (Exciton
Hopping) Rate Constants kEnT along the Four Directions
̈
AB, AD, AE, and AC in F-MOF and DA-MOF Based on Eq
fraction of
hops
net displacement in net displacement in distance
a
4
hops
(Å)
F-MOF
ES1 + ES2 + ES3
ES1 (109 s−1
)
ES2 (109 s−1
)
ES3 (109 s−1
)
(109 s−1
)
AB
0.38
0.52
0.03
0.07
−
1.8
2.2
21
34
11
15
23
36
F-MOF
AD
AE
AB
AD
AE
AC
0.4
7.2
0.5
1.4
0.28
0.47
0.07
0.01
6.70
2.60
0.02
0.01
7.4
10.2
0.6
0.5
AC
AB′
AD′
0.8
2.0
1.5
−
2.3
DA-MOF
4.09
DA-MOF
33
AB
AD
AE
AC
99
30
39
11
0.04
0.08
0.01
0.00
103
33
AB
0.55
0.18
0.21
0.06
−
380
290
580
210
410
330
2.54
0.08
0.79
AD
AE
19
39
21
12
AC
AB′
AD′
11
a
35
Rates of energy transfer from ES1 to ES1, ES2 to ES2, and ES1 to
−
22
ES2 have been considered. Each rate is averaged over 25 rotational
conformations (the rotation angles of −30°, −15°, 0°, 15°, and 30°
around the bipyridyl axis of a porphyrin molecule).
displacements for excitons in both MOFs. As one expects based
on estimates of hopping rate, the displacements, in terms of
numbers of hops, are significantly anisotropic. If we recognize
that a hop in the AC direction is equivalent to one hop in the
AB direction + one hop in the AD direction, then the AB- and
AD-directed displacements will be slightly greater than
obtained otherwise. The modified displacement results are
listed in Table 2 as AB′ and AD′ displacements.
state (ES2) are comparatively small; they are negligible for
inter-excited-state (cross) transfer. The rate of energy transfer is
greatest along the AB direction, exceeding by ca. 3- to 9-fold
the total rates in each of the other directions. Qualitatively this
is expected, since the distance between nearest neighbor
chromophores is smallest in the AB direction. Within F-MOF,
however, energy transfer rate is greatest in the AD direction, an
effect that appears to arise from framework tilting.
Notably, the fastest rate in Table 1 for DA-MOF is about an
order of magnitude larger than the fastest for F-MOF. This
difference is attributed to the increased conjugation along the
dipyridyl axis of the porphyrin strut in DA-MOF compared to
F-MOF. The enhanced conjugation, in turn, leads to decreased
electronic symmetry, increased Q-band oscillator strength
(increased J values), and better absorption/emission spectral
overlap (Figure 2).
In addition to quantifying directional dynamics, we also are
interested in obtaining estimates of exciton transport distance
in various directions. As a starting point, we considered the case
of strictly one-dimensional transport. If an exciton in DA-MOF
visits 90 distinct fluorophores, its absolute displacement at the
end of its lifetime will, on average, be 45 linkers (fluorophores).
Because each successive hop can place the exciton either one
site (linker) closer or one site further from its point of origin
(i.e., the linker that absorbed the photon that created the
exciton), the total number of hops made by the exciton will, on
It is important to recall that the distance covered in a hop
depends strongly on the direction of the hop. For example, for
DA-MOF the step sizes in the AB and AE directions are ca. 11
and 27 Å, respectively. As a consequence, and as shown in
Table 2, the exciton displacement distance in DA-MOF is
greatest in the AE direction (ca. 580 Å) even though the
number of exciton hops along the AE direction is less than 40%
of the number made along the AB direction. The AE direction
is the direction defined by porphyrin−porphyrin collinear
alignment. In contrast, for F-MOF the largest displacement
distance is attained along the AD direction, i.e. in a direction
parallel to the plane defined by the non-chromophoric linker.
Given that the fluorescence lifetimes of DA-MOF and F-
MOF in the absence of FcPy are 2.9 and 4.9 ns (see Figure S10
and Table S2 in Supporting Information for the lifetime
measurements), respectively, the estimated exciton hopping
times are ∼2900/2025 = 1.4 ps for DA-MOF and ∼4900/7.8 =
620 ps for F-MOF, respectively. The exciton hopping times
derived from the experimental data are broadly consistent with
those listed in Table 1, suggesting that the Forster energy
̈
F
dx.doi.org/10.1021/ja310596a | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
transfer model (eq 4) as implemented here (see Supporting
Information for details of calculations of J) is reasonably reliable
for estimating the exciton hopping rate and its migration
anisotropy.
ASSOCIATED CONTENT
* Supporting Information
Experimental details, synthetic procedures and characterization
(NMR and mass spectra) of chromophores, PXRD of samples,
results of fluorescence decay, TGA, and detailed descriptions of
the theoretical calculations. This material is available free of
charge via the Internet at http://pubs.acs.org.
■
S
CONCLUSIONS
■
We have studied molecular-exciton migration within two Zn-
porphyrin-based MOFs (DA-MOF and F-MOF). In both
cases, transport is significantly anisotropic. Because of the
strong sensitivity of coupling constants (J values) to separation
distance, modest increases in the AB- and AC-defining
dimensions of the non-chromophoric, tetra-carboxylate linker
should lead to predominantly AE-directed transport of excitons
in DA-MOF. Conversely, shrinking the dimensions of the non-
chromophoric linker, and thereby positioning chromophores
closer in a lateral, but not vertical, sense, should disfavor
transport along the AE direction (i.e., the chromophore end-to-
end direction).
AUTHOR INFORMATION
Corresponding Author
o-farha@northwestern.edu; wiederrecht@anl.gov; j-hupp@
northwestern.edu
■
Present Address
#Laboratory of Organic Chemistry, Department of Chemistry
and Applied Biosciences, ETH Zurich, CH-8093 Zurich,
Switzerland
Author Contributions
Eliminating Zn(II) from the porphyrin centers would likely
increase exciton lifetimes. All else being equal, a 4-fold increase
in singlet exciton lifetime would yield 2-fold increases in exciton
displacements. In reality, elimination of the zinc centers will
reduce the symmetry of each of the porphyrins and split the Q
bands. In all likelihood, the splitting will alter spectral overlaps
and thereby further affect hopping dynamics and transport
distances. If the goal is to increase exciton migration distances,
a more effective strategy would likely be to further increase the
electronic asymmetry of the porphyrin linker in DA-MOF,
thereby increasing oscillator strengths in the Q-band region and
boosting both spectral overlap and coupling constants.
Porphyrin dimerization (via ethyne or butadiyne connectors)
is known to elicit such increases in Q-band absorption
intensity.41−43 Alternatively, triplet exciton migration might
prove effective. Often the primary decay pathway from
fluorescent excited states of zinc porphyrins is intersystem
conversion to a triplet excited state. In the absence of molecular
oxygen, such states often persist for several orders of magnitude
longer than the corresponding singlet states. Although rates of
triplet exciton migration are likely to be many times slower than
singlet migration rates, the enhanced exciton lifetime could
conceivably result in longer displacement distances. Regardless,
because the triplet exciton is formed at the final destination of
the singlet exciton, rather than at the site of photon absorption,
the exciton displacement distance (singlet + triplet contribu-
tions) should exceed the displacement attained by singlet
migration alone.
Returning to DA-MOF and F-MOF, we find that the average
numbers of chromophore-to-chromophore hops by singlet
excitons are about 2000 and 8, respectively. These translate into
a roughly 16-fold difference in the number of potential
fluorophores visited by excitons in the two MOFs. The
differences arise from the reduced electronic symmetry and
enhanced linear conjugation obtained with DA-MOF by
incorporating an acetylene unit between the tetrapyrrole core
and each of the two pyridyl anchoring groups of the
porphyrinic linker. These differences, in turn, engender greater
Q-band oscillator strength and better absorption/emission
spectral overlap, culminating, for DA-MOF in greater dipolar
coupling and a larger overlap integral.
∥H.-J.S. and S.J. contributed equally.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
S.P. thanks Dr. F. C. Grozema from Delft University of
Technology for providing assistance with semiempirical
INDO/CIS calculations. G.C.S. gratefully acknowledges the
AFOSR (grant no. AFOSR FA9550-11-1-0275) for support of
computational work. J.T.H. gratefully acknowledges support at
Northwestern from the U.S. Department of Energy, Office of
Science, Basic Energy Sciences, through grant no. DE-FG02-
87ER13808 for studies of molecular light-harvester and MOF
synthesis, MOF characterization, and quencher incorporation.
S.J.W. thanks the Spanish Ministry of Education for a pre-
doctoral FPU fellowship. S.J. acknowledges support at Argonne
from the ANSER Center, an Energy Frontier Research Center
funded by the U.S. Department of Energy, Office of Science,
Office of Basic Energy Sciences, under award number DE-
SC0001059 (fluorescence spectroscopy). Use of the Center for
Nanoscale Materials, and support of G.P.W., is funded by the
U.S. Department of Energy, Office of Science, Office of Basic
Energy Sciences through contract no. DE-AC02- 06CH11357.
REFERENCES
■
(1) Scholes, G. D.; Fleming, G. R.; Olaya-Castro, A.; van Grondelle,
R. Nat. Chem. 2011, 3, 763.
(2) Barber, J.; Andersson, B. Nature 1994, 370, 31.
(3) Kuhlbrandt, W.; Wang, D. N.; Fujiyoshi, Y. Nature 1994, 367,
614.
(4) McDermott, G.; Prince, S. M.; Freer, A. A.; Hawthornthwaite-
Lawless, A. M.; Papiz, M. Z.; Cogdell, R. J.; Isaacs, N. W. Nature 1995,
374, 517.
(5) Webber, S. E. Chem. Rev. 1990, 90, 1469.
(6) Sykora, M.; Maxwell, K. A.; DeSimone, J. M.; Meyer, T. J. Proc.
Natl. Acad. Sci. U.S.A. 2000, 97, 7687.
(7) Frec
3713.
́
het, J. M. J. J. Polym. Sci., Part A: Polym. Chem. 2003, 41,
(8) Balzani, V.; Bergamini, G.; Ceroni, P.; Vogtle, F. Coord. Chem.
Rev. 2007, 251, 525.
̈
(9) Aratani, N.; Kim, D.; Osuka, A. Acc. Chem. Res. 2009, 42, 1922.
(10) Li, X.; Sinks, L. E.; Rybtchinski, B.; Wasielewski, M. R. J. Am.
Chem. Soc. 2004, 126, 10810.
(11) Kelley, R. F.; Lee, S. J.; Wilson, T. M.; Nakamura, Y.; Tiede, D.
M.; Osuka, A.; Hupp, J. T.; Wasielewski, M. R. J. Am. Chem. Soc. 2008,
130, 4277.
The substantial distances and sizable anisotropy of singlet
exciton migration in DA-MOF suggest that it or related
framework compounds could prove useful as antenna-type
light-harvesters in solar energy conversion systems.
G
dx.doi.org/10.1021/ja310596a | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(12) Sengupta, S.; Ebeling, D.; Patwardhan, S.; Zhang, X.; von
Berlepsch, H.; Bottcher, C.; Stepanenko, V.; Uemura, S.; Hentschel,
̈
C.; Fuchs, H.; Grozema, F. C.; Siebbeles, L. D. A.; Holzwarth, A. R.;
Chi, L.; Wurthner, F. Angew. Chem., Int. Ed. 2012, 51, 6378.
̈
(13) Patwardhan, S.; Sengupta, S.; Siebbeles, L. D. A.; Wurthner, F.;
̈
Grozema, F. C. J. Am. Chem. Soc. 2012, 134, 16147.
(14) Kent, C. A.; Mehl, B. P.; Ma, L.; Papanikolas, J. M.; Meyer, T. J.;
Lin, W. J. J. Am. Chem. Soc. 2010, 132, 12767.
(15) Wang, C.; Lin, W. J. Am. Chem. Soc. 2011, 133, 4232.
(16) Kent, C. A.; Liu, D.; Ma, L.; Papanikolas, J. M.; Meyer, T. J.; Lin,
W. J. J. Am. Chem. Soc. 2011, 133, 12940.
(17) Kent, C. A.; Liu, D.; Meyer, T. J.; Lin, W. J. J. Am. Chem. Soc.
2012, 134, 3991.
(18) Shultz, A. M.; Farha, O. K.; Hupp, J. T.; Nguyen, S. T. J. Am.
Chem. Soc. 2009, 131, 4204.
(19) Farha, O. K.; Shultz, A. M.; Sargeant, A. A.; Nguyen, S. T.;
Hupp, J. T. J. Am. Chem. Soc. 2011, 133, 5652.
(20) Zhang, Z.; Zhang, L.; Wojtas, L.; Nugent, P.; Eddaoudi, M.;
Zaworotko, M. J. J. Am. Chem. Soc. 2012, 134, 924.
(21) Zhang, Z.; Zhang, L.; Wojtas, L.; Eddaoudi, M.; Zaworotko, M.
J. J. Am. Chem. Soc. 2012, 134, 928.
(22) Wang, X.-S.; Meng, L.; Cheng, Q.; Kim, C.; Wojtas, L.;
Chrzanowski, M.; Chen, Y.-S.; Zhang, X. P.; Ma, S. J. Am. Chem. Soc.
2011, 133, 16322.
(23) Suslick, K. S.; Bhyrappa, P.; Chou, J.-H.; Kosal, M. E.; Nakagaki,
S.; Smithenry, D. W.; Wilson, S. R. Acc. Chem. Res. 2005, 38, 283.
(24) Feng, D.; Gu, Z.-Y.; Li, J.-R.; Jiang, H.-L.; Wei, Z.; Zhou, H.-C.
Angew. Chem., Int. Ed. 2012, 51, 10307.
(25) Morris, W.; Volosskiy, B.; Demir, S.; Gandara, F.; McGrier, P.
́
L.; Furukawa, H.; Cascio, D.; Stoddart, J. F.; Yaghi, O. M. Inorg. Chem.
2012, 51, 6443.
(26) Burnett, B. J.; Barron, P. M.; Choe, W. CrystEngComm 2012, 14,
3839.
(27) Lee, C. Y.; Farha, O. K.; Hong, B. J.; Sargeant, A. A.; Nguyen, S.
T.; Hupp, J. T. J. Am. Chem. Soc. 2011, 133, 15858.
(28) To ensure that the experiments were not complicated by small
differences in excited-state energy for zinc(II)-ligated versus non-
ligated porphyrin linkers, pyridine was bound (axially) to all of the
ZnP units that were not coordinated by FcPy.
(29) Wilmer, C. E.; Leaf, M.; Lee, C. Y.; Farha, O. K.; Hauser, B. G.;
Hupp, J. T.; Snurr, R. Q. Nat. Chem. 2012, 4, 83.
(30) Gould, S. L.; Tranchemontagne, D.; Yaghi, O. M.; Garcia-
Garibay, M. A. J. Am. Chem. Soc. 2008, 130, 3246.
(31) LeCours, S. M.; DiMagno, S. G.; Therien, M. J. J. Am. Chem.
Soc. 1996, 118, 11854.
(32) Ishida, Y.; Shimada, T.; Masui, D.; Tachibana, H.; Inoue, H.;
Takagi, S. J. Am. Chem. Soc. 2011, 133, 14280.
(33) Thomas, S. W., III; Joly, G. D.; Swager, T. M. Chem. Rev. 2007,
107, 1339.
(34) Liu, Y.; Ogawa, K.; Schanze, K. S. J. Photochem. Photobiol. C−
Photochem. Rev. 2009, 10, 173.
(35) Mata, J.; Uriel, S.; Peris, E.; Llusar, R.; Houbrechts, S.; Persoons,
A. J. Organomet. Chem. 1998, 562, 197.
(36) Bucher, C.; Devillers, C. H.; Moutet, J.-C.; Royal, G.; Saint-
Aman, E. Coord. Chem. Rev. 2009, 253, 21.
(37) Muller, J. G.; Atas, E.; Tan, C.; Schanze, K. S.; Kleiman, V. D. J.
̈
Am. Chem. Soc. 2006, 128, 4007.
(38) Patwardhan, S.; Tonzani, S.; Lewis, F. D.; Siebbeles, L. D. A.;
Schatz, G. C.; Grozema, F. C. J. Phys. Chem. B 2012, 116, 11447.
(39) Beenken, W. J. D.; Pullerits, T. J. Chem. Phys. 2004, 120, 2490.
(40) Patwardhan, S.; Sengupta, S.; Wurthner, F.; Siebbeles, L. D. A.;
̈
Grozema, F. J. Phys. Chem. C 2010, 114, 20834.
(41) Angiolillo, P. J.; Uyeda, H. T.; Duncan, T. V.; Therien, M. J. J.
Phys. Chem. B 2004, 108, 11893.
(42) Zhang, T.-G.; Zhao, Y.; Asselberghs, I.; Persoons, A.; Clays, K.;
Therien, M. J. J. Am. Chem. Soc. 2005, 127, 9710.
(43) Lovett, J. E.; Hoffmann, M.; Cnossen, A.; Shutter, A. T. J.;
Hogben, H. J.; Warren, J. E.; Pascu, S. I.; Kay, C. W. M.; Timmel, C.
R.; Anderson, H. L. J. Am. Chem. Soc. 2009, 131, 13852.
H
dx.doi.org/10.1021/ja310596a | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX