14397641, 2003, 3, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cphc.200390042 by CAPES, Wiley Online Library on [20/10/2022].

See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Photoluminescence Intensity Fluctuations and
Electric-Field-Induced Photoluminescence
Quenching in Individual Nanoclusters of
Poly(phenylenevinylene)
Ivan Scheblykin,[a, b] George Zoriniants,[a] Johan Hofkens,[a] Steven De Feyter,[a]
Mark Van der Auweraer,*[a] and Frans C. De Schryver[a]

Substantial fluctuations of the fluorescence intensity have been ™on∫ to ™off∫ state does not explain the character of the fluctuations
detected for single clusters of poly(phenylenevinylene) containing observed. We suggest that the quenching of the light-emitting
more than 75 polymer chains or 30 000 monomer units. To the best exciton by a long-lived species, such as, for example, polarons,
of our knowledge, this is the first time such fluctuations (which plays an important role in these unusual fluctuations. Electric field
resemble the ™blinking∫ effect in single-molecule fluorescence) have induced fluorescence quenching differs significantly for different
been reported for such a large molecular ensemble containing clusters. It is proposed that this fluorescence was mainly quenched
several macromolecules. Together with the distinct jumps, smooth by polarons injected from the electrodes in the presence of an
fluctuations of the fluorescence intensity, with characteristic times electric field. The specific behaviour of each cluster is explained by
from milliseconds to seconds, were observed. This fact distinguishes suggesting a different position of the clusters with respect to the
the fluorescence behaviour of the polymer clusters from that of electrodes.
other multichromophoric systems such as the single chains of
conjugated polymers reported in the literature. The consecutive or KEYWORDS:
simultaneous switching of one or several emitting sites from the
conjugation ¥ fluorescence ¥ FRET (fluorescence resonance
transfer) ¥ nanostructures ¥ polymers ¥ single-molecule
studies

Introduction
Since the discovery of electroluminescence from conducting large number of chromophores, the fluorescence of the polymer
polymers,[1] these organic materials have attracted substantial molecule exhibits an on/off behaviour (or a ™blinking∫ effect)
attention due to their possible applications as, for example, well-known in single-molecule spectroscopy.[9, 10, 12] This effect
organic light-emitting diodes (OLEDs).[2±4] However, despite has been attributed to efficient exciton migration within the
great effort,[5] some basic physical, chemical and morphological polymer chain to a trapping site where the exciton could be
properties of these systems have not yet been completely quenched. It has been suggested that the quenching occurs on
characterised. Recently, in addition to the conventional optic photochemically generated defects, or on a pair of separated
methods allowing the investigation of the dynamics of excited charges.[6]
states in dilute solutions and in bulk polymer films[2, 3] it has In the current contribution, we study objects which have an
become possible to study the physics and chemistry of a single intermediate size between the single polymer molecule and the
polymer chain using single-molecule spectroscopy.[6±10] bulk polymer film. These are clusters of poly(phenylenevinylene)
To understand the behaviour of conducting polymers, it is (PPV) separated from each other in a very thin film of 30 nm of an
important to study the individual properties of nanosized
polymer entities[11] (clusters, aggregates and so on), which play [a] Prof. M. Van der Auweraer, Dr. I. Scheblykin, Dr. G. Zoriniants, Prof. J. Hofkens,
a very important, and sometimes unclear, role in the operation of Dr. S. De Feyter, Prof. F. C. De Schryver
Laboratory for Molecular Dynamics and Spectroscopy
OLEDs,[3] especially in OLEDs made of polymer blends.[4] K.U. Leuven, Celestijnenlaan 200 F
Generally, most optic studies deal with ensemble-averaged, Leuven 3001 (Belgium)
bulk properties of materials. Efficient OLEDs with a nanosize Fax: (‡ 32)16-327990
active area have been fabricated by using the effect of phase E-mail: [email protected]

separation in a blend of two different polymers.[11] [b] Dr. I. Scheblykin

P.N. Lebedev Physics Institute
Over the past few years, fluorescence properties of single RAS, Leninsky pr. 53, 119991 GSP-1 Moscow (Russia)
chains of conjugated polymers have attracted the attention of Supporting information for this article is available on the WWW under http://
several research groups.[6±8] It has been found that despite the www.chemphyschem.org or from the author.

260 ¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 1439-4235/03/04/03 $ 20.00+.50/0 CHEMPHYSCHEM 2003, 4, 260 ± 267
 14397641, 2003, 3, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cphc.200390042 by CAPES, Wiley Online Library on [20/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Photoluminescence Quenching in Nanoclusters

inert polymer matrix. To the best of our knowledge, this is the means that the lateral size of the emitting entities was smaller
first time that random variations in the fluorescence intensity for than the spatial resolution of the microscope, which was
these large molecular ensembles, containing at least 30 000 determined by diffraction. Fluorescence spectra (Figure 1 b) of
monomer units or 75 polymer chains, have been observed.[13] these spots, as well as the spectrum of the background, suggest
The features of the fluorescence fluctuations were more that in both cases the fluorescence originates from PPV.[3]
complicated compared to those observed in the fluorescence However, the spectrum of the background fluorescence is
of single polymer chains.[6±8] broader and reveals almost no vibronic structure, contrary to the
In order to understand the mechanism of exciton quenching observations made for the spectrum of a spot.
in clusters, we studied fluorescence in the presence of an Although the spots are diffraction limited, we can estimate the
external electric field. It is known that an electric field can number of chromophores per spot in two ways. The first method
significantly quench the fluorescence of bulk films of conjugated (see Supporting Information, Section A for details) gives us a
polymers by increasing the rate of exciton dissociation to value of 3  104 monomers corresponding to a molecular weight
polarons.[14±16] In addition, the electric field enhances the injection Mcluster of 3  106 g mol 1, which is a lower limit of the spot size. It
of polarons into the film. The long-lived polarons can efficiently is based on a comparison of the fluorescence intensity of a spot
quench the excitation by resonant energy transfer (FRET).[17, 18] with that of a model system of known optic and photophysical
parameters. The upper limit (1  106 monomers corresponding
to a molecular weight Mcluster of 1  108 g mol 1) was obtained by
the second method (Supporting Information, Section B), in
Results and Discussion which we took into account the concentration of the PPV in
1. Origin of the Fluorescent Spots the poly(vinyl alcohol) (PVA) matrix, the number of spots per unit
area of the film and the contribution of the spots to the total
A fluorescence image of the film obtained by a scanning
fluorescence of the film. As the molecular weight of the
confocal microscope (described in the Experimental Section) is
precursor polymer is about 105 g mol 1, which corresponds to
presented in Figure 1 a. The fluorescence image consists of
400 monomers per chain, the spots must be clusters containing
bright spots appearing on a uniform, less intense background.
from 75 to 2500 polymer chains.
We analysed numerous spots of different intensities and found
Figure 1 a shows that the clusters reveal a certain distribution
that the average diameter at half height was 326  27 nm and
of fluorescence intensities. For clusters smaller than the confocal
that the diameter did not correlate with the peak intensity. This
spot, this distribution of intensity could be attributed to the
distribution of the cluster size. The estimations of the cluster size
mentioned above were performed for clusters of average
fluorescence intensity.
The presence of PVA in the mixture possibly plays an important
role in the formation of clusters. It is known that the addition of
polymers can dramatically affect the particle formation process.[19]

2. Fluorescence Intensity Fluctuations

The fluorescence intensity of the polymer clusters was large

enough to observe the clusters visually using the wide-field
fluorescence microscope described in the Experimental Section.
The fluorescence of the sample could be monitored under a total
magnification of 400. It was possible to observe not only the
separated clusters but also their fluorescence intensity fluctua-
tions. At an excitation power of 200 W cm 2, the fluorescence
intensity of most of the clusters (except the very bright ones)
exhibited substantial fluctuations. The simple method of visual
monitoring of the fluorescence fluctuations was found to be very
convenient for a quick, preliminary, check of the samples.
Examples of fluorescence intensity transients recorded using
the scanning confocal microscope are shown in Figure 2 and
Figure 3. The fluorescence intensity transients exhibit a compli-
cated behaviour showing jumps, short spikes and smooth
changes of intensity. To prove that the observed fluctuations
(especially the smooth ones) were not experimental artefacts, we
also measured the jitter of the laser intensity as a function of
Figure 1. a) Fluorescence image of the film recorded by a confocal scanning
fluorescence microscope. b) Fluorescence spectra of a cluster (thick line) and of time (Figure 2 a), and the background luminescence (by collect-
the background (thin line). Excitation wavelength 488 nm. ing the emission from a site between the clusters: Figure 2 e).

CHEMPHYSCHEM 2003, 4, 260 ± 267 261

 14397641, 2003, 3, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cphc.200390042 by CAPES, Wiley Online Library on [20/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
M. Van der Auweraer et al.

The jitter of the laser intensity was recorded as scattered light

from the sample using the same experimental setup in which the
notch filter (which normally completely blocks the excitation
laser light) was replaced by a dense neutral density filter. One can
see that no fluctuations comparable to those of the cluster's
fluorescence were present in signals a) and e). Furthermore, for
single 100-nm polymer beads (Molecular Probes, Inc) containing
hundreds of dye molecules[10] no fluorescence intensity fluctua-
tions were observed under the same excitation and detection
conditions (Figure 2 b) and similar count rates. The observation
of abrupt (within a bin time) jumps of fluorescence intensity
along with smooth ones, as well as the previous experiments
with single molecules[20] strongly suggest that the smooth
fluctuations were not caused by fluctuations of the response
function of the setup.
The presence of large, smooth variations of the fluorescence
intensity is well illustrated in Figure 3. The initial rise of the signal
in Figure 3 a from around 70 counts (after correction for the
background luminescence) at times less than 2 s to about
600 counts (after correction for the background luminescence)
at 5 s seems to be abrupt. However, upon closer inspection, it is
smooth with a rise time of about 0.2 seconds, as shown in
Figure 3 b with an extended time scale. In contrast to those
™slow∫ fluctuations, jumps of the intensity occurring within the
bin time were also observed (Figure 3 d). It should be noted that
we consider the jumps ™sharp∫ or ™distinct∫ if they are faster than
Figure 2. Fluorescence intensity transients of a multichromophore polymer bead the detection bin time (5 ms). Fluorescence fluctuations of some
(b), PPV clusters (c,d), signal from the background fluorescence (from a place clusters showed a jumplike blinking superimposed on smooth
between clusters) (e). Plot a) is the jitter of the scattered light from the excitation fluctuations of the intensity occurring on a timescale of seconds
laser. The time scale is the same for all graphs. The dashed line shows the level of
the background luminescence (see the text). The data (a, b, c) were averaged over (see Figure 2 d).
10 points, d) over 20 points for better representation. The clusters were excited The fluorescence spectra of a PPV cluster recorded simulta-
with an excitation power of 200 W cm 2, at 488 nm. neously with a fluorescence transient are shown in Figure 4. The

Figure 3. Fluorescence intensity transient of a PPV cluster excited at 488 nm with a power of 200 W cm 2. Two parts (plots a and c) of the total transient are shown with
the same time scale. The magnified parts of these plots (b and d) illustrate the smooth and sharp features of the fluctuations. The dashed line corresponds to the
background luminescence measured between the clusters.

262 CHEMPHYSCHEM 2003, 4, 260 ± 267

 14397641, 2003, 3, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cphc.200390042 by CAPES, Wiley Online Library on [20/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Photoluminescence Quenching in Nanoclusters

However, the current/voltage (I/V) curve was found to be

symmetrically related to the sign of the voltage applied. This
indicates that the main part of the measured current was a
leakage current flowing through the PVA matrix.[14] At small fields
(corresponding to a bias between 5 and ‡ 3 V), quenching was
observed only for a fraction of the clusters. At larger applied
fields, more clusters showed quenching.
Figure 5 illustrates the effect in a more detailed way. It
represents two series of photographs of two different areas, A
and B, on the film. In each series, the images 1, 2 and 3 were
obtained sequentially: 1st image–without the voltage, 2nd–
with the voltage and 3rd–without the voltage again. The
exposure time and excitation intensity were the same for all
images. Thus, the brightness directly corresponds to the
fluorescence efficiency.

Figure 4. a) Fluorescence time transient and b) fluorescence spectra recorded

simultaneously. The numbered bars indicate time intervals during which the
corresponding fluorescence spectra 1, 2, 3, 4 were collected. Fluorescence spectra
were normalised to the collection time of 0.5 s. The dashed line in panel a) shows
the level of the background luminescence.

acquisition time intervals for each spectrum (1...3 s) are shown in

the figure by numbered bars. It can be seen that the spectrum of
the cluster did not change substantially during the fluorescence
fluctuations. The spectra coincide (within experimental error)
with the spectrum of a bright cluster obtained with a much
longer acquisition time (Figure 1 b). This suggests that spectral
diffusion cannot be the reason for the fluorescence intensity
fluctuations.

3. Electric Field Induced Quenching of the PPV Cluster

Fluorescence

To study the electric field induced photoluminescence (PL)

quenching, the films were prepared on an indium tin oxide (ITO)
covered glass slide and then covered by a vapour deposited
aluminium contact. Thus, the sample had a diodelike arrange-
ment, as used in electroluminescence devices.[1] The voltage was
considered positive when the ITO was connected to the positive
output of the power supply (forward bias). Only after increasing
the voltage to 20 V (which corresponds to an electric field of
Figure 5. Fluorescence microphotographs of the film used in the experiments
about 6.7 MV cm 1) the films could be burned by electrical with electric field. Series A and B were recorded at different sites on the sample.
breakthrough. This observation indicates that the electric The subsequent images in each series were taken under the same exposure
stability of the films was quite high and they did not contain conditions except for the voltages, with the electric field being off, then on, etc.
The voltages applied are denoted on the microphotographs.
thin spots, pinholes or other similar defects. A typical value of the
excitation intensity at 488 nm was about 40 W cm 2. The electric
field was varied between 0.3 and 4.5 MV cm 1 by applying 1 to
14 V to the sample. One should also note that no electro- It can be seen that the PL of various individual clusters showed
luminescence from the clusters was observed while the voltage a specific and reproducible response to the applied voltage. A
was being applied to the sample. The photoluminescence substantial quenching of the clusters b, g and d (marked on
intensity was monitored visually using the fluorescence micro- Figure 5) was easily observed in the series. However, the cluster a
scope and then microphotographs were taken. The fluorescence and other features seen in the images did not exhibit any
showed substantial, easily observable quenching when a voltage notable change in PL intensity. So far, we have not observed any
larger than 5 V was applied. The quenching was more pro- evident correlation between the cluster brightness and its
nounced under a forward bias than under a reverse bias. sensitivity to the voltage applied.

CHEMPHYSCHEM 2003, 4, 260 ± 267 263

 14397641, 2003, 3, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cphc.200390042 by CAPES, Wiley Online Library on [20/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
M. Van der Auweraer et al.

It can be seen that the clusters on the images A3 and B3 are dissociation, play an important role in quenching.[14] Therefore, in
less bright than those on images A1 and B1, which were our case, the effect of the PL quenching has a different origin
obtained before the electric field was applied. This is due to from that explained by direct field-induced exciton dissociation
photodegradation of the polymer. The decrease of the fluo- into positive and negative polarons.[15, 16] We assumed that the
rescence intensity of the polymer at the site exposed to the laser main part of the PL quenching was due to the quenching of
excitation was more substantial when an electric field was excitons by polarons, which were injected, or photo-injected,
applied. Moreover, although the electric field was applied to a into the cluster from the electrodes. Polaron quenching has been
large area of film between electrodes, the decrease of the proposed to be responsible for the photoluminescence modu-
fluorescence intensity occurred only at the illuminated area. In lation of submicrometer sized domains of poly(2-methoxy-5-(2'-
other words, the nonincidental exposure of the clusters to light ethylhexoxy)-1,4-phenylenevinylene (MEH-PPV) under near-field
or electric current destroyed the conjugated polymers much optical excitation.[21, 22]
more slowly than a simultaneous exposure. This suggests that The individual ™sensitivity∫ of a cluster to an electric field can
the photogeneration of charges plays an important role in the easily be explained by the specific position of the cluster within
photodegradation of PPV clusters. the film (Figure 6). Namely, the strongest quenching effectshould
be observed for case a), which has the best junction. Also, case b)
may yield efficient quenching when hole injection is the pre-
4. Mechanism of the Electric Field Induced Fluorescence dominant quenching process. The clusters c), d) and e) may
Quenching also undergo quenching, however, at a much higher applied
The atomic force microscopy (AFM) study showed that the films voltage.
containing clusters possessed quite a uniform surface. Peak to
peak variations of the thickness were less then 10 nm, the root 5. Evaluation of the Fluorescence Intensity Fluctuations
mean square was 1.1 nm and the total thickness of the film was
30 nm. The fluctuations of the cluster fluorescence intensity (Section 2)
The volume of the cluster (Vcluster) can be derived from its mass resemble the so-called ™on/off∫ or ™blinking∫ behaviour observed
estimated in Section 1: Vcluster ˆ Mcluster/1 ˆ (0.05 to 1.5)  10 16 g/ for the fluorescence of single polymer molecules and other
(1 g cm 3) ˆ (0.05 to 1.5)  10 16 cm3 (where 1 g cm 3 is assumed multichromophoric systems which are single molecules or
for the density of the polymer material). Therefore, the clusters molecular ensembles exhibiting effective energy transfer over
can be imagined to be spheres with a radius Rcluster ˆ (3V/4p)1/3 ˆ the chromophores.[6, 7, 20, 23±26]Until now, the largest multichro-
(11 to 33)  10 7 cm ˆ 11 to 33 nm. This estimation shows that mophoric system known to show the ™blinking∫ effect was a
the clusters are large enough (regardless of their shape) to be MEH-PPV single polymer chain,[7, 27] about 2000 ± 4000 monomer
able to touch the top or the bottom of the polymer layer, or even units long. For single PPV oligomers containing tens of monomer
both. units, this ™blinking∫ has been studied by several research
In experiments where an electric field was applied, the film groups.[6, 28] Huser et al. have reported that self-aggregation of a
was sandwiched between two electrodes. Therefore, these two long, single MEH-PPV chain led to a more distinct and
electrodes could be electrically connected through the polymer pronounced fluorescence blinking behaviour than that observed
cluster. Of course, the resistance of such a nanojunction depends for a stretched chain of the same polymer.[8, 27] This increase
drastically on the particular morphology of the cluster, its occurs as the ability of an exciton to migrate over a chain is
position with respect to the electrodes and the area of contact enhanced in the case of a collapsed coil, which allows excitation
with the electrodes. Some possible arrangements are shown in transfer in three dimensions.[8]
Figure 6. Evidently, the injection of polarons can vary in a drastic One can consider a PPV cluster to be an aggregate formed by
way, depending on the way the clusters are connected to the many polymer chains. In this way, clusters are different from the
electrodes. other systems investigated by methods of single-molecule
As mentioned, the electric-field-induced quenching was more spectroscopy. A typical value of the exciton delocalization
pronounced under a forward bias. The injection of charge length in PPV is estimated to be six monomer units.[16] Thus,
carriers is also more efficient under a forward bias when ITO can one can consider the PPV cluster to be a system consisting of at
very efficiently inject holes.[2] This certainly means that charges least 5000 chromophores.
generated at the interface, either by dark injection or exciton As we have already mentioned, the clusters did not show well
distinguished ™on∫ and ™off∫ states corresponding to particular
levels of the fluorescence intensity. Together with distinct
™jumps∫[7, 8] (Figure 3 c, Figure 2 c) smooth variations of the
fluorescence intensity were observed with a characteristic time
up to seconds (Figure 2 d and 3 b). To the best of our knowledge,
such large, smooth fluorescence intensity fluctuations in
single nanometer scale objects have not previously been de-
scribed.[20]
Figure 6. A sketch of the possible positions of the clusters with respect to the The model[7, 10] that describes the ™blinking∫ behaviour in
electrodes. molecular ensembles is based on the idea that the excitation

264 CHEMPHYSCHEM 2003, 4, 260 ± 267

 14397641, 2003, 3, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cphc.200390042 by CAPES, Wiley Online Library on [20/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Photoluminescence Quenching in Nanoclusters

generated on a site in the ensemble can migrate over the system spectra from a similar MEH-PPV polymer where the latter was
until it reaches the site (denoted by A) of minimum energy measured by photoinduced absorption-detected magnetic
(™energy funnel∫[7] ) or a trap site. The emission always occurs resonance[37] (see Figures 3(a) and Figure 1(a) in ref. [37])
from site A. If there is a quencher Q, for example, a triplet We performed an estimation of the Fˆrster radius (R0)[38] of
exciton[20, 29] or a charged radical[20, 30] close to site A, the polaron ± exciton interaction based on the transition dipole
fluorescence is suppressed. The stochastic appearance and moment of the polaron excitation (3.33  10 29 C m),[35] the
disappearance of the quencher results in distinct, sharp, random features of the absorption spectrum of the polaron, the
changes of the fluorescence intensity of the ensemble. This fluorescence quantum yield and the fluorescence spectrum of
explanation has been applied to MEH ± PPV single chains and the exciton. For PPV films, the fluorescence quantum yield
oligomers.[7] reported in the literature[1, 32, 39] is about 10 % (in some cases up
As we have already seen, in our case, the transients are very to 30 %), so we took 10 % as a conservative estimate. A value of
complicated. A closer inspection of the fluctuation shown in the Fˆrster radius R0 ˆ 3.8 nm was thus obtained.
Figure 3 a indicates that the fluorescence intensity increased The radius of a single collapsed polymer chain of 2000 mono-
continuously over approximately 0.2 s from a level of 200 counts, mers[7] is of the order of 4.3 nm when the density of this material
which is close to that of the background (140 counts), up to is assumed to be 1 g cm 3. This means that if a quencher
600 counts. Such a rise cannot be explained in the framework of (polaron) with R0 ˆ 3.8 nm appears in the system the fluores-
a model considering a small number of independent sites cence is quenched, regardless of the mutual positions of A and
exhibiting ™on/off∫ behaviour. It is very unlikely that many Q, which will lead to a distinct on/off behaviour.[6, 8] In this
independent sites will go to the ™on∫ state consecutively, one-by- system, the motions of the emitting site and the quencher are
one. A large amplitude of the fluctuations would require a low not essential for the fluorescence blinking. The key difference of
number of independent emitting sites. On the other hand, to the system studied here from single polymer chains is that the
have smooth fluctuations the system must contain many cluster radius (R  11 nm) exceeds the Fˆrster radius by a factor
fluorescence intensity levels and therefore a large number of of three. Even if one considers exciton hopping, preceding
independently emitting sites. This obvious contradiction proves energy transfer to the polaron,[40] the probability of quenching
that ensemble averaging alone cannot account for the features depends not only on the presence of the quencher but also on
of those ™smooth∫ fluctuations. the distance RAQ between the emitting site A and the quen-
Barbara and co-workers have proposed some mechanisms for cher Q. Thus, in clusters, the mutual position of A and Q must be
exciton quenching responsible for fluorescence blinking in MEH- taken into account. If one assumes that RAQ changes smoothly
PPV single chains[7] and conjugated oligomers.[10] They have over time due to hopping of the polarons, the proportionality of
suggested energy transfer to photochemically formed charged the quenching rate to RAQ 6, will lead to smooth changes of the
defects (polarons). The present experiments with an electric field fluorescence yield (or the smooth type of fluorescence intensity
show (Section 3), and it is known from the literature,[18] that fluctuations observed in Figure 4 A).
polarons can quench the fluorescence of excitons in conjugated The changes of RAQ can be caused by a number of factors. Both
polymers by an energy transfer. Electric-field-induced fluores- sites A and Q can be mobile. On the one hand, the quencher Q
cence quenching in conjugated polymers has also been studied can hop slowly over the cluster. For example, it may be slow
by McNeill et al.[21, 22] (see also references therein) on the tunnelling or hopping of a polaron from one deep trap to
nanometer-size scale and the role of polarons was discussed. another. Such a slow motion can be both random (diffusion
Therefore, it is possible that polarons are also responsible for the controlled) or unidirectional (drift motion) due to a local electric
fluctuations of the PL quantum yield in clusters. field. On the other hand, the emitting site A may also move due
Polarons can be photogenerated by the intense excitation to gradual photodegradation of the polymer. After the degra-
light used in the experiments. The photoinduced formation of dation of the light-emitting site A, another site A' could start to
polarons is enhanced by a partial photooxidation of the act as an emitting site. Such a process would lead to the moving
polymer.[31] Furthermore, in the present experiments, the film of the emitting site A over the cluster. This process can be quite
thickness (30 nm) was much smaller than that of most of the slow, depending on the photostability of the sample, and can
previously studied conjugated polymer films. In such a thin film, result in fluctuations on a time scale of seconds. If both sites A
interfacial effects, and especially the interaction with the and A' are spatially close together, the latter process could lead
aluminium layer which is present in our samples, could play an to smooth fluctuations. All these effects can cause the smooth
important role.[32±34] type of fluorescence intensity fluctuations observed in the
The quenching of an exciton by a polaron may be explained in experiments (Figure 4 A).
terms of Fˆrster-type resonant energy transfer. The value of the In addition to the ™smooth∫ fluctuations, abrupt fluctuations
transition dipole moment calculated for polarons in PPV were also observed. If, for example, the distance between the
oligomers is about 3.33  10 29 C m (10 D).[35] This value is close sites A and A' is large, on the scale of R0 , photodegradation
to that calculated (4.43  10 29 C m) for an allowed transition in a (compare with above) could lead to abrupt fluctuations of the
cyanine dye.[36] The polaron absorption spectrum overlaps with emission intensity. Abrupt fluctuations could also be related to
the low-energy tail of the emission spectrum of the excitons.[37] the generation of triplet excitons. The latter absorb at 700 ±
Because of the lack of information about the unsubstituted PPV, 900 nm[41±44] and hence can also quench[45] the PPV fluorescence.
we obtained the fluorescence emission and polaron absorption The photoinduced absorption spectrum of triplet excitons is red-

CHEMPHYSCHEM 2003, 4, 260 ± 267 265

 14397641, 2003, 3, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cphc.200390042 by CAPES, Wiley Online Library on [20/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
M. Van der Auweraer et al.

shifted with respect to that of polarons, and overlaps only with Experimental Section
the long wavelength tail of the photoluminescence spectrum
of MEH-PPV.[37] So, the energy transfer to triplet excitons should Films containing clusters of PPV were prepared by the following
be less efficient than, but still comparable to, the energy transfer procedure starting from two stock solutions: a 1 wt % aqueous
to polarons. The shorter lifetime of triplet excitons (1 ms)[46] solution of the tetrahydrothiophenium PPV precursor (pre-PPV)
purchased from Lark Enterprises (USA)[32] (molecular weight more
compared to the bin time can explain why their stochastic
than 1  105 g mol 1 according to the specifications of the supplier)
formation and decay induces abrupt fluctuations in the
and a 1 wt % aqueous solution of PVA (molecular weight 1.3 
fluorescence intensity. Furthermore, a triplet lifetime of 105 g mol 1). The two solutions were mixed in a ratio of 1/100 to
1 ms,[46] combined with the high power density of the focused obtain a 1 wt % mixture of pre-PPV in PVA. The mixed solution was
excitation light, can lead to appreciable local triplet concen- stored at 5 8C for two weeks. The solution was then spin-coated onto
trations. substrates (either a thin cover glass or an ITO covered glass
However, we cannot exclude a different reason for the smooth purchased from Merck), at room temperature, at 1500 rpm. To
fluctuations. Our system is essentially nonuniform and consists convert the precursor to PPV, the films were heated at 120 8C in a
vacuum (better than 10 3 mbar) for 10 h. The film thickness was
of a dense cluster of PPV containing thousands of PPV
estimated to be about 30 nm from topographic profiles of scratched
monomers residing in a matrix consisting of electrically neutral
samples measured with an atomic force scanning microscope
PVA with a small content of ™free∫ PPV chains. Thus, the local (Topometrix, Discoverer). The appearance of PPV clusters was
concentration of polarons will be much higher in the clusters strongly dependent on the preparation conditions, such as temper-
than in the surroundings. This may be the primary origin of the ature, storage time of the polymer mixture and so on. Films prepared
nonequilibrium in the system, which may lead to irregular from solutions kept at 5 8C for several months contained very large
fluorescence changes. clusters with sizes up to a few microns and showed no fluorescence
intensity fluctuations. Spin-coating a freshly prepared solution led to
homogeneous films without clusters.
To limit the fast photodegradation of the polymer under intensive
6. Conclusions laser irradiation, the films were covered by an aluminium layer using
Fluorescent films of PPV/PVA blends consisting of well-dispersed vacuum deposition.[7] For the covered films, no substantial decrease
of fluorescence was observed, even after several minutes of
PPV and bright PPV clusters were prepared from aqueous
irradiation at excitation powers of 10 W cm 2 under ambient
solutions using the PPV precursor route. The clusters were conditions.
estimated to contain at least 3  104 monomer units of PPV,
The fluorescence of the films was monitored visually through an
corresponding to approximately 75 single polymer chains, or
optical microscope OPTIPHOT-2 (Nikon) with an episcopic fluores-
5  103 chromophores. cence attachment Nikon EFD-3. The microscope could also be
The clusters exhibited fluorescence intensity fluctuations coupled to a photo-camera Nikon FX-35WA. A CW Ar-ion laser
resembling the ™blinking∫ effect in single-molecule fluorescence. (Spectrophysics, model 2025) was used for fluorescence excitation at
A characteristic time of the fluctuations was found from 1 ms 488 nm or 458 nm. A dichroic mirror (working range 440 ± 500 nm)
(resolution of the instrument) to 20 s when illuminated by laser was used to guide the excitation light to the sample. A notch filter
radiation of about 100 W cm 2 at 488 nm. It was demonstrated prevented the scattered light from passing into the registration port
that the fluorescence intensity also showed smooth, large of the microscope. The diameter of the laser beam in the sample
plane could be tuned from 200 mm to 20 mm by replacing the optic
variations of the intensity. To the best of our knowledge, such
elements (lenses and diaphragms) introduced into the beam. In most
smooth changes of the fluorescence yield of a single fluorescent experiments, a MPlan 40  N. A. 0.5 dry objective lens (Nikon) was
nanoobject of this size have not been observed before. We used. Photographs of the samples were taken using a Kodak
propose that exciton quenching, possibly by photogenerated 5054TMZ 3200ASA black and white film with an exposure time of
polarons, is responsible for the smooth fluctuations of the 1.5 s.
fluorescence intensity of the cluster. Whatever the details of the Fluorescence images, spectra and fluorescence intensity transients
mechanism are, it is necessary to combine a long-lived quencher were recorded on a scanning confocal fluorescence microscope
with a cluster size that exceeds the exciton diffusion length and which was constructed on the basis of a modified LUMINA (Top-
the Fˆrster radius in order to obtain the smooth fluctuations. ometrix) and adapted for single-molecule spectroscopy.[20] A Nikon
The effect of fluorescence quenching by an applied electric immersion objective 100  N.A. 1.40 was used. The laser mentioned
field was studied for films sandwiched between aluminium and above was used for the excitation. The scanning confocal micro-
scope was used to record the fluorescence images, fluorescence
ITO electrodes. Essentially specific electric-field-induced quench-
spectra and fluorescence intensity transients with bin times of 1 to
ing of the fluorescence of nanoclusters is shown. Generally, the
5 ms, at a particular location of the sample with a confocal spatial
larger the applied voltage, the greater the number of clusters resolution of 0.33 mm.
that exhibited quenching. This effect was more prominent when
the ITO electrode was positively biased, which was better for
charge injection into the film. We assume that the PL is The authors are grateful to Prof. Igor Osad'ko (Lebedev Physical
quenched by polarons injected from the electrodes. The Institute, Russia), Vladimir Arkhipov (IMEC, Belgium), David
variation of the effect among the different clusters is explained Beljonne (University of Mons, Belgium), Prof. Christian von Borczy-
by a distribution of the positions of the clusters with respect to skowski and Dr. Frank Cichos (TU Chemnitz, Germany) for useful
the electrodes. discussions and some preliminary experiments. This work was

266 CHEMPHYSCHEM 2003, 4, 260 ± 267

 14397641, 2003, 3, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cphc.200390042 by CAPES, Wiley Online Library on [20/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Photoluminescence Quenching in Nanoclusters

supported by DWTC (Belgium) through grant IUAP-IV-11 and IUAP- [23] T. Gensch, J. Hofkens, A. Hermann, K. Tsuda, W. Verheijen, T. Vosch, T.
V-03, the F.W.O.-Vlaanderen, the Nationale Loterij, the research Christ, T. Basche¬, K. M¸llen, F. C. De Schryver, Angew. Chem. 1999, 111,
3970; Angew. Chem. Int. Ed. 1999, 38, 3752.
Council of the K.U. Leuven through GOA 2001/2, COST D14. I.G.Sch. [24] M. A. Bopp, Y. W. Jia, L. Q. Li, R. J. Cogdell, R. M. Hochstrasser, Proc. Natl.
thanks RFBR for the grant 00 ± 02 ± 16607, INTAS grant YSF2001/2 ± Acad. Sci. U.S.A. 1997, 94, 10 630.
111, the F.W.O.-Vlaanderen for a ™Visiting Postdoctoral Fellowship∫ [25] M. A. Bopp, A. Sytnik, T. D. Howard, R. J. Cogdell, R. M. Hochstrasser, Proc.
and the K.U. Leuven for a ™Junior Research Fellowship∫. Natl. Acad. Sci. U.S.A. 1999, 96, 11 271.
[26] M. Cotlet, J. Hofkens, S. Habuchi, G. Dirix, M. Van Guyse, J. Michiels, J.
Vanderleyden, F. C. De Schryver, Proc. Natl. Acad. Sci. U.S.A. 2001, 98,
14 398.
[1] J. H. Burroughes, D. D. C. Bradley, A. R. Brown, R. N. Marks, K. Mackay, R. H. [27] T. Huser, M. Yan, L. J. Rothberg, Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 11187.
Friend, P. L. Burns, A. B. Holmes, Nature 1990, 347, 539. [28] J. D. White, J. H. Hsu, S.-C. Yang, W. S. Fann, G. Y. Pern, S. A. Chen J. Chem.
[2] R. H. Friend, R. W. Gymer, B. A. Holmes, J. H. Burroughes, C. Taliani, D. D. C. Phys. 2001, 114, 3848.
Bradley, D. A. Dos Santos, J. L. Bre¬das, M. Lˆgdlund, W. R. Salaneck, Nature [29] J. Hofkens, W. Schroeyers, D. Loos, M. Cotlet, F. Kˆhn, T. Vosch, M. Maus, A.
1999, 397, 121. Herrman, K. M¸llen, T. Gensch, F. C. De Schryver, Spectrochim. Acta, Part A
[3] Handbook of Conducting Polymers, 2nd Ed. (Eds.: T. A. Skotheim, R. L. 2001, 57, 2093.
Elsenbaumer, J. R. Reynolds), Marcel Dekker, New York, 1998. [30] Th. Christ, F. Petzke, P. Bordat, A. Herrmann, E. Reuther, K. M¸llen, Th. Basche¬,
[4] A. Kraft, A. C. Grimsdale, A. B. Holmes, Angew. Chem. 1998, 110, 416; J. Lumin. 2002, 98, 23.
Angew. Chem. Int. Ed. 1998, 37, 402. [31] E. J. W. List, J. Partee, J. Shinar, U. Scherf, K. M¸llen, E. Zojer, K. Petritsch, G.
[5] A. J. Heeger, Angew. Chem. 2001, 113, 2660; Angew. Chem. Int. Ed. 2001, Leising, W. Graupner, Phys. Rev. B: Condens. Matter Mater. Phys. 2000, 61,
40, 2591. 10 807.
[6] D. A. Vanden Bout, W.-T. Yip, D. Hu, D. K. Fu, T. M. Swager, P. F. Barbara, [32] I. G. Scheblykin, V. I. Arkhipov, M. Van der Auweraer, F. C. De Schryver, Helv.
Science 1997, 277, 1074. Chim. Acta 2001, 84, 3616.
[7] J. Yu, D. Hu, P. F. Barbara, Science 2000, 289, 1327. [33] H. Killesreiter, H. B‰ssler, Chem. Phys. Lett. 1971, 11, 411.
[8] T. Huser, M. Yan, J. Photochem. Photobiol., A 2001, 144, 43. [34] M. Pope, C. E. Swenberg, Electronic Processes in Organic Crystals and
[9] Ph. Tamarat, A. Maali, B. Lounis, M. Orrit, J. Phys. Chem. A 2000, 104, 1. Polymers, 2nd Ed., Oxford University Press, New York, 1999, Ch. 2, p. 320.
[10] W.-T. Yip, D. Hu, J. Yu, D. A. Vanden Bout, P. F. Barbara, J. Phys. Chem. A [35] A. Ye, Ph. D. Thesis, University of Mons, Belgium, 2002.
1998, 102, 7564. [36] M. Van der Auweraer, I. Scheblykin, Chem. Phys. 2001, 275, 285.
[11] M. Granstrˆm, O. Ingan‰s, Adv. Mater. 1995, 7, 1012. [37] X. Wei, Z. V. Vardeny, N. S. Sariciftci, A. J. Heeger, Phys. Rev. B 1996, 53,
[12] X. S. Xie, J. K. Trautman, Annu. Rev. Phys. Chem. 1998, 49, 441. 2187.
[13] I. Scheblykin, G. Zoriniants, J. Hofkens, S. De Feyter, M. Van der Auweraer, [38] V. M. Arganovich, M. D. Galanin, Electronic Excitation Energy Transfer in
F. C. De Schryver, 5th International Symposium on Functional p-Electron Condensed Mater; North Holland, Amsterdam, 1982, Ch. 1.
Systems (Fp5), Ulm/Neu-Ulm, Germany, 2002, Book of Abstracts, p. 211. [39] N. C. Greenham, I. D. W. Samuel, G. R. Hayes, R. T. Phillips, Y. A. R. R.
[14] M. Deussen, M. Scheidler, H. B‰ssler, Synth. Met. 1995, 73, 123. Kessener, S. C. Moratti, A. B. Holmes, R. H. Friend, Chem. Phys. Lett. 1995,
[15] R. Kersting, U. Lemmer, M. Deussen, H. J. Bakker, R. F. Mahrt, H. Kurz, V. I. 241, 89.
Arkhipov, H. B‰ssler, Phys. Rev. Lett. 1994, 73, 1440. [40] F. Willig, A. Blumen, G. Zumofen, Chem. Phys. Letters 1984, 108, 222.
[16] M. Pope, C. E. Swenberg, Electronic Processes in Organic Crystals and [41] N. F. Colaneri, D. D. C. Bradley, R. H. Friend, P. L. Burn, A. B. Holmes, C. W.
Polymers, 2nd Ed., Oxford University Press, New York, 1999, Ch. 9, Spangler, Phys. Rev. B 1990, 42, 11 671.
pp. 825 ± 854. [42] X. Wei, B. C. Hess, Z. V. Vardeny, F. Wudl, Phys. Rev. Lett. 1992, 68, 666.
[17] G. Leising, S. Tasch, W. Graupner in Handbook of Conducting Polymers, 2nd [43] A. R. Brown, K. Pichler, N. C. Greenham, D. D. C. Bradley, R. H. Friend, A. B.
Ed. (Eds.: T. A. Skotheim, R. L. Elsenbaumer, J. R. Reynolds), Marcel Dekker, Holmes, Chem. Phys. Lett. 1993, 210, 61.
New York, 1998, p. 847. [44] D. Beljonne, Z, Shuai, R. H. Friend, J. L. Bre¬das, J. Chem. Phys. 1994, 102,
[18] E. J. W. List, C. H. Kim, W. Graupner, G. Leising, J. Shinar, Mater. Sci. Eng. B 2042.
2001, 85, 218. [45] J. Hofkens, W. Schroeyers, D. Loos, M. Cotlet, F. Kˆhn, T. Vosch, M. Maus, A.
[19] J. Rieger, D. Horn, Angew. Chem. 2001, 113, 4460; Angew. Chem. Int. Ed. Herrmann, K. M¸llen, T. Gensch, F. C. De Schryver, Spectrochimica Acta Part
2001, 40, 4330. A 2001, 57, 2093.
[20] J. Hofkens, M. Maus, T. Gensch, T. Vosch, M. Cotlet, F. Kˆhn, A. Herrmann, K. [46] N. C. Greenham, R. H. Friend in Solid State Physics (Ed.: H. Ehrenreich),
M¸llen, F. C. De Schryver, J. Am. Chem. Soc. 2000, 122, 9278. Academic Press, New York, 1995, Vol. 49, p. 56.
[21] J. D. McNeill, D. B. O'Connor, D. M. Adams, P. F. Barbara, J. Phys. Chem. B
2001, 105, 76. Received: March 14, 2002 [Z 386]
[22] J. D. McNeill, P. F. Barbara, J. Phys. Chem. B 2002, 106, 4632. Revised: October 30, 2002

CHEMPHYSCHEM 2003, 4, 260 ± 267 267