Article

pubs.acs.org/JPCC

Magnitude of the Förster Radius in Colloidal Quantum Dot Solids

A. Jolene Mork,† Mark C. Weidman,‡ Ferry Prins,‡ and William A. Tisdale*,‡
†
Department of Chemistry and ‡Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge,
Massachusetts 02139, United States

ABSTRACT: Excitonic energy transfer among colloidal nanocrystal quantum dots (QDs) is
responsible for exciton transport in many QD optoelectronic devices. While Förster theory has
successfully accounted for the distance scaling of energy transfer in many QD systems, the
overall magnitude of the Förster radius in close-packed QD solids remains an open question.
Here, we use spectrally resolved transient photoluminescence quenching to measure the
magnitude of the Förster radius in blended donor−acceptor QD assemblies. For blends of
CdSe/CdZnS core/shell QDs consisting of 4.0 nm diameter donors (λmax,em ≈ 550 nm) and
5.5 nm acceptors (λmax,abs ≈ 590 nm), we measure energy transfer rates per donor−acceptor
pair that are 10−100 times faster than the predictions of Förster theory. These rates
correspond to an effective Förster radius of 8−9 nm, compared to a theoretical Förster radius
of 5−6 nm. We discuss possible sources for the discrepancy between theory and experiment
including the magnitude of the absorption cross section, dipole orientation, and dipole−multipole couplingand suggest that
several common assumptions should be considered carefully before applying Förster theory to solid QD films.

■ INTRODUCTION
The unique size-dependent optical characteristics of semi-
conductor nanocrystal quantum dots (QDs) have motivated
research in next-generation technologies ranging from solid-
state optoelectronic devices, such as LEDs,1−3 solar cells,4,5
lasers,6,7 photodetectors,8 and luminescent solar concentrators,9
to biomedical technologies utilizing imaging, sensing, and
labeling with QDs.10,11 The physical properties of colloidal
QDs, such as broadband absorption, narrow emission line
width, band gap tunability, and high quantum yield, have
allowed commercialization of QD technologies;3,12 for many of
these technologies, energy transfer between QDs must be
controlled to achieve optimum performance. For example,
white-light-emitting QD-LEDs require an excess of blue
emitters relative to green and red in order to compensate for
energy transfer to, and emission from, smaller band gap
(redder) dots.13 Similarly, control of exciton diffusion14 via
cascaded energy transfer structures allows for efficient exciton
funneling to acceptor sites and dramatically increases the
fluorescence intensity of the acceptor.15,16 Several biological
detection assays rely on energy transfer from QDs to organic
chromophores, where the proximity of a donor and acceptor
can be inferred from the energy transfer efficiency.17,18 Förster
resonance energy transfer (FRET) theory has been well
established as a theoretical and practical framework for
understanding energy transfer in QD systems, in particular
between QDs and organic chromophores in solution.19−22
However, recent reports of anomalously large polarization
effects in QD solids23,24 raise the important question of how
energy transfer in close-packed QD solids may differ from that
between QDs and dyes in dilute solution.
Energy transfer within colloidal QD films25,26 is often
understood in the context of Förster theory, which describes
the coupling between two point dipoles. However, the
comparatively large size of QDs has led previous authors to
question whether QDs behave as point dipoles separated by a
distance equal to the center-to-center spacing, as the model
suggests.19−21,27 Prior experiments in solid films have measured
the energy transfer rate, kET, between donor and acceptor QDs
through monitoring the donor fluorescence lifetime in either
donor−acceptor blended films25,26,28,29 or bilayer struc-
tures,15,26,30,31 and have calculated values for kET in both
geometries on the order of (1−10) × 108 s−1. Changes in
deposition technique (drop-casting vs Langmuir−Blodgett)15
as well as changes in the spacing between donor and acceptor
layers in bilayer structures30 or with a ligand series32 have been
shown to affect kET by altering the proximity of nearest acceptor
neighbors. Similarly, for energy transfer within an inhomoge-
neous distribution of QDs, the line width and Stokes shift of
QDs affect the energy transfer rate from large band gap to small
band gap QDs, consistent with the expected dependence of
spectral overlap on the energy transfer rate predicted by Förster
theory.33,34 These experiments have revealed that Förster
theory correctly predicts some scaling parameters that affect the
energy transfer rate in QD solids, such as overlap integral and
interparticle separation. However, few have made quantitatively
rigorous comparison to numerical predictions based on spectral
properties in order to test the limits of Förster theory as it
applies to solid QD films.
In this work, we measure energy transfer rates from 4.0 nm
diameter donor QDs (λmax,em ≈ 550 nm) to 5.5 nm acceptor
QDs (λmax,abs ≈ 590 nm) in homogeneously blended solid-state
films and compare the results to predicted energy transfer rates
using Förster theory, paying particularly careful attention to the
Received: February 28, 2014
Revised: May 30, 2014

© XXXX American Chemical Society A dx.doi.org/10.1021/jp502123n | J. Phys. Chem. C XXXX, XXX, XXX−XXX
The Journal of Physical Chemistry C Article

Figure 1. TEM and GISAXS measurements of donor−acceptor separation. (A−C) Short-ligand acceptor. (D−F) Long-ligand acceptor. (A, D) TEM
micrographs of donor−acceptor blended samples, with insets illustrating the corresponding donor−acceptor pair and size histogram revealing the
approximately equimolar bimodal distribution in QD sizes. Scale bar = 40 nm. (B, E) GISAXS patterns for blended 3D films. (C, F) Histograms of
the measured nearest-neighbor interparticle distance (Gaussian fit to TEM data indicated with dotted gray line), demonstrating good agreement
between TEM and GISAXS measurements.

propagation of experimental uncertainty. We obtain simulta-
neous spectral and temporal resolution using spectrally resolved
transient photoluminescence spectroscopy, where the fluo-
rescence lifetime is measured as a function of emission
wavelength. We observe changes in the donor lifetime when
mixed with a short-ligand acceptor or a long-ligand acceptor
and compare to the no-energy-transfer case of a pure donor in
solution to calculate quenching efficiencies and energy transfer
rates. The energy transfer rates measured from donor
quenching experiments are more than an order of magnitude
larger than those predicted, using standard assumptions, by
Förster theory. We discuss several factors that may be
responsible for this discrepancy, including nonrandom dipole
alignment, contributions from higher-order multipoles, and an
enhanced absorption cross section of QDs in close-packed
solids. These results indicate that care must be taken when
using Förster theory to make quantitative predictions of energy
transfer in colloidal QD solids, which has implications for the
design of QD optoelectronic devices such as LEDs and solar
cells where excitonic energy transfer in the solid state plays a
role in device performance.
phonic acid), and the other had a shorter aromatic ligand
(benzylphosphonic acid). The acceptor QDs were synthesized
identically except for the ligand introduced during overcoating;
see Dang et al. for further details.6 From TEM images of pure
close-packed films, the short-ligand acceptor had a diameter of
5.5 ± 0.5 nm with mean center-to-center separation of 8 ± 1
nm, while the long-ligand acceptor had a diameter of 5.3 ± 0.5
nm with mean center-to-center separation of 9 ± 1 nm.
Sample Preparation. Molar concentrations of quantum
dots were estimated using the method of Yu et al., based on the
band-edge absorption wavelength and intensity.35 Donor/
acceptor mixtures were prepared by adding approximately
equal moles of each QD species, as determined by absorption
measurements, and diluting to the desired concentration. QD
films used for this study were prepared in a nitrogen-filled
glovebox by spin-casting ∼5 mg/mL solutions of QDs in
toluene onto a cleaned glass slide that was previously treated
with mercaptopropyltrimethoxysilane, resulting in films ap-
proximately 20 nm thick. After spin-coating, the samples were
packaged in an inert atmosphere for removal from the glovebox
by placing the QD slide face-down onto a clean glass slide and

■ EXPERIMENTAL METHODS
Materials. CdSe/CdZnS nanocrystals were supplied by QD
Vision, Inc., and were used as received. QD materials were kept
under inert atmosphere during all handling and experimenta-
tion. QD solutions were prepared with anhydrous solvents
purchased from Sigma-Aldrich. All QD cores are CdSe with an
overcoated CdZnS shell. Additional details pertaining to these
QD materials, including the synthesis procedure, can be found
in the previous publication by Dang et al.6 TEM analysis
revealed that the donor QDs used for this study were 4.0 ± 0.5
nm in diameter, with a center-to-center distance in pure close-
packed monolayers of 6.0 ± 0.9 nm. Two types of acceptor
QDs were paired with the donor in energy transfer studies: one
acceptor had a long-chain aliphatic ligand (octadecylphos-
B dx.doi.org/10.1021/jp502123n | J. Phys. Chem. C XXXX, XXX, XXX−XXX
sealing the edges with epoxy.
TEM Analysis of Blended Samples. Transmission
electron microscopy (TEM) was used to determine the average
center-to-center distance between donors and acceptors in
blended QD films. Images were taken on a JEOL 2011
operated at 200 kV and using an objective aperture to increase
mass−thickness contrast. Samples for TEM analysis were
prepared by drop-casting the donor/acceptor mixture onto a
copper TEM grid with amorphous carbon support film. The
donor/acceptor mixtures were diluted such that a monolayer
was formed on the TEM grid, allowing for identification of
individual QDs.
TEM images were processed using ImageJ and then analyzed
in Matlab. First, an FFT bandpass filter was applied to the
image, which served to better distinguish the edges of the
The Journal of Physical Chemistry C
nanocrystals from the carbon support.36 The brightness/
contrast of the image was then maximized, followed by
conversion from grayscale to black and white. Next, particles
were indexed and tabulated according to the coordinates of
their center of mass, and the area was used to calculate an
effective diameter for each particle. Artifacts of the image
processing algorithm that were clearly not individual QDs were
rejected based on size or shape. By plotting the histogram of all
diameters in an image, two peaks could be resolved which
corresponded to the two different sizes of QDs present in the
mixture (see insets to Figures 1A,D). We assigned a threshold
diameter between these two peaks, creating a population of
“small” QDs (diameters less than the threshold) and “large”
QDs (diameters greater than the threshold). Using the centroid
location of each particle in the image, it was then possible to
compute center-to-center distances between every pair of QDs,
including the case of interest here, which is the distance
between “small” (donor) and “large” (acceptor) QDs. The first
peak in the histogram of donor−acceptor distances (Figures
1C,F) represents the average nearest-neighbor center-to-center
distance between each donor and acceptor. The width of this
peak represents the distribution of different center-to-center
distances present in the ensemble.
GISAXS Analysis. The determination of the QD center-to-
center distance by TEM was complemented by grazing-
incidence small-angle X-ray scattering (GISAXS) analysis of
spin-cast 3D films. GISAXS measurements were performed at
the undulator-based X9 beamline at the National Synchrotron
Light Source at Brookhaven National Laboratory. The X-ray
energy was set to 14.1 keV (wavelength 0.0879 nm), and the
beam size was focused to 50 μm tall by 100 μm wide at the
sample position using a KB mirror system. Because of the
grazing incidence, beam projection along the beam direction
was ∼1 cm. The grazing-incident angle of the X-ray beam was
varied from 0.07° to 0.3°. The presented results are for 0.15°,
which were found to be representative. 2D scattering data were
collected using a Pilatus 1M detector which was calibrated
using a silver behenate standard. The QD films used for
GISAXS were prepared in a nitrogen-filled glovebox by spin-
casting ∼5 mg/mL solutions of QDs in toluene onto a cleaned
glass slide that was previously treated with mercaptopropyl-
trimethoxysilane (the same procedure used to prepare samples
for Förster radius measurements). The interparticle spacing
(Figures 1C,F) was extracted from a given scattering pattern by
taking a line cut along the in-plane scattering direction (qy) at
the Yoneda peak (streak of high intensity just above the sample
horizon in Figures 1B,E).37
Spectrally Resolved Transient Photoluminescence
Spectroscopy. Transient photoluminescence measurements
were performed by spectrally resolved time-correlated single-
photon counting in a modified optical microscope. Briefly, the
prepackaged QD samples were loaded onto an inverted
microscope with a 60×, 1.4 NA, oil-immersion objective and
excited at 405 nm with a pulsed diode laser (PicoQuant)
operating at 5 MHz. Epifluorescence was collected by the
objective and reimaged onto the entrance slit of a spectrograph
(Princeton Instruments) where the light was dispersed by a
grating. The dispersed fluorescence was either projected onto a
CCD array (Pixis, Princeton Instruments) for collection of a
time-integrated spectrum, or a small wavelength band (∼4 nm)
was passed through an alternate exit slit for time-resolved
detection by an avalanche photodiode (τ-SPAD, PicoQuant).
All spectra were corrected for the spectral sensitivity of the
C dx.doi.org/10.1021/jp502123n | J. Phys. Chem. C XXXX, XXX, XXX−XXX
Article
detection system using a calibrated tungsten halogen lamp.
Data were collected from multiple locations on each sample
and averaged together to obtain statistically representative
results.
■ RESULTS AND DISCUSSION
Measurement of Donor−Acceptor Separation. To
carefully modulate the donor−acceptor interparticle separation
for quantification of energy transfer rates, we first verified that
the length of the ligand bound to the acceptor QD indeed alters
the donor−acceptor spacing. Approximately equimolar sol-
utions of donor and acceptor QDs were prepared and used
both for TEM and GISAXS analysis as well as fluorescence
lifetime measurements. TEM analysis revealed that the donor
and acceptor species are mostly homogeneously mixed (see the
later section on Experimental Uncertainty for a discussion of
the effect of incomplete mixing on our Fö rster radius
measurements) and that changing the acceptor ligand from
benzylphosphonic acid to octadecylphosphonic acid does
indeed alter the interparticle separation (Figure 1; see
Experimental Methods for a detailed description of the image
processing procedure). A statistical analysis of large-area TEM
micrographs revealed donor−acceptor distances that were
normally distributed around a mean value. In the short-ligand
sample, the average donor−acceptor center-to-center distance
was 7.0 ± 0.9 nm (Figure 1C), whereas in the long-ligand
sample the distance was 7.8 ± 1.1 nm (Figure 1F).
We note that TEM analysis was performed on monolayer
samples, whereas spectroscopic measurements were performed
on multilayers. Therefore, to verify that the TEM interparticle
separation measurements were representative of the inter-
particle separation in a 3D film, we measured the grazing-
incidence small-angle X-ray scattering (GISAXS) pattern for
samples prepared in the same way as fluorescence lifetime
samples. Figures 1B,E show the GISAXS patterns for the two
types of mixed films (donor QD blended with each of the two
acceptor QDs). From these scattering images, a histogram of
nearest-neighbor spacing in the 3D blended films was obtained
(see Experimental Methods for detailed procedure). The
GISAXS interparticle distances and the TEM-measured
interparticle distances are plotted together in Figures 1C,F.
The measured GISAXS distribution is broader than the TEM
distribution because it includes donor−donor and acceptor−
acceptor distances (in addition to the donor−acceptor distance,
which is of interest here), whereas our TEM analysis extracts
just the donor−acceptor distance. The excellent agreement
between TEM and GISAXS data inspires confidence in our
quantification of the donor−acceptor spacing as well as our
understanding of its statistical variation within the sample. We
emphasize the importance of obtaining a statistically
representative distribution of separation distances before
performing quantitative analysis of energy transfer data.
Characterization methods that isolate just a few donor−
acceptor pairs via high-resolution TEM or AFM38 introduce
observer bias into the measurement and paint an incomplete
picture of energy transfer.
Donor Fluorescence Quenching. To begin the study of
energy transfer between donor and acceptor QDs, the first
experimental observation is the quenching of donor fluo-
rescence in the presence of the acceptor. For the materials used
here, the emission of the donor overlaps with the second
absorption feature of both acceptors (Figures 2A,B), indicating
that energy transfer from donor to acceptor should be possible.
The Journal of Physical Chemistry C
Figure 2. (A, B) Emission spectrum of donor (blue) and absorption
spectrum of acceptor (red) QDs. The acceptor dots differ only in the
length of the surface-bound ligand, which is either short (A) or long
(B) as illustrated above each column. (C, D) Emission spectra of the
mixed solution (dotted) and film (solid), each normalized to the same
number of counts, of nearly equimolar blends of donor and acceptor
QDs, showing quenching of donor fluorescence in film compared to
solution. The large difference in donor and acceptor emission
intensity, despite similar molar concentrations, results from the
difference in absorption cross section between the large acceptor
and small donor QDs at the laser excitation wavelength.
Indeed, the ratio of donor to acceptor emission intensity
decreased upon deposition from solution to film for both
donor−acceptor blends studied here (Figures 2C,D). In
particular, the reduction in relative donor fluorescence is
more significant for donors in the presence of the short-ligand
acceptor than the long-ligand, qualitatively confirming initial
expectations that ligand length may be used to control energy
transfer efficiency between donors and acceptors in otherwise
identical materials. These observations suggesting the occur-
rence of energy transfer are further strengthened and quantified
by measurements of changes in the donor lifetime in the
presence versus absence of acceptors (described in detail
below). Changes in donor lifetime conclusively indicate near-
field energy transfer and rule out other possible mechanisms to
explain changes in the donor intensity, such as emission/
reabsorption. In the following sections, we first calculate the
expected energy transfer rates using the overlap integral and the
measured distance between donors and acceptors and then
compare these theoretical predictions to the experimentally
measured energy transfer rates based on changes in the donor
lifetime in the absence or presence of acceptors.
Calculation of Expected Energy Transfer Dynamics
Using Fö rster Theory. Theoretical energy transfer rates may
be calculated using the physical properties of the donor−
acceptor blend.39,40 In this case, the energy transfer rate (kET) is
inversely proportional to the center-to-center distance between
QDs (d) and the total lifetime of the donor (τD)
1 ⎛ R0 ⎞
6
kET = ⎜ ⎟
τD ⎝ d ⎠ (1)
D dx.doi.org/10.1021/jp502123n | J. Phys. Chem. C XXXX, XXX, XXX−XXX
Article
where R0 is the Förster radius (defined to include the donor
quantum efficiency), given by
9ηPLκ 2
6
R0 = 5 4
128π n
∫ λ 4FD(λ)σA(λ) dλ (2)
where ηPL is the quantum efficiency of the donor (0.8 ± 0.1 for
samples used here), κ2 is an orientation factor typically assumed
to be 2/3 for randomly oriented dipoles, n is the refractive
index of the material (typically assumed to be the volume-
weighted sum of the refractive indices of the cores and organic
ligands, equal to about 1.8),25 FD(λ) is the donor emission
spectrum normalized to an integrated value of 1, and σA(λ) is
the acceptor absorption spectrum in units of cross-sectional
area. The absorption cross section for the core−shell QDs used
here was calculated according to the method described by De
Geyter and Hens using core and shell diameters (see expanded
discussion of absorption cross section below).41 From eq 2, the
theoretical Förster radii for the donor/short-ligand-acceptor
and donor/long-ligand-acceptor pairs are 5.2 ± 0.2 and 5.1 ±
0.2 nm, respectively, resulting in expected energy transfer rates
of 10 (±8) × 106 and 5 (±4) × 106 s−1 per donor−acceptor
pair. The uncertainty in R0 arises from an estimated 10%
uncertainty in both the quantum yield and the overlap integral,
while the uncertainty in the associated energy transfer rate
stems from the errors in R0, the donor lifetime τ, and (most
significantly) the experimentally measured interparticle separa-
tion (Figure 1). Based on these calculations, the energy transfer
rate is expected to differ between the two blends by about a
factor of 2 due to the difference in interparticle separation
imparted by the different ligands.
Experimental Measurement of the Energy Transfer
Rate. To quantify energy transfer dynamics between donor and
acceptor QDs, we measured the exciton lifetime as a function of
emission wavelength (Figure 3). Plots of emission intensity as a
function of emission wavelength and time (Figures 3A,B) show
donor quenching and, in particular, reveal more significant
donor quenching in the case where the donor and acceptor
QDs are closer together (Figure 3A). Spectral and temporal
resolution allows for clear distinction between donor and
acceptor fluorescence. To obtain population decay curves
(panels C−F of Figure 3), the partially overlapping emission
spectra from donor and acceptor QDs were separated by fitting
two Gaussians to the emission spectrum recorded at each time
slice of the data. The intensities in the donor components of
the double Gaussian fit were then spectrally integrated to
calculate a donor lifetime in the presence of acceptors. The
exciton population decay of the donor in the presence of the
acceptor in solid films is plotted in Figure 3, panels C−F, and
compared to the donor lifetime in solution where no energy
transfer occurs. The dramatic change in lifetime results from
energy transfer to neighboring acceptors, and the per-pair
energy transfer rate may be extracted from the lifetime change
as described below in order to compare to theoretical
predictions presented in the previous section. The presence
of the QD acceptors in solid film reduces donor fluorescence
lifetime as well as intensity, indicating near-field energy transfer
between donor and acceptor QDs.
The change in donor lifetime observed by spectrally resolved
transient photoluminescence (Figure 3) was used to calculate
the energy transfer rate between donor−acceptor pairs in the
blended solid films. In the absence of energy transfer, the total
The Journal of Physical Chemistry C
Figure 3. Spectrally resolved transient photoluminescence measure-
ments of donor−acceptor blended films. The left and right columns
correspond to the short-ligand and long-ligand acceptor, respectively.
(A, B) Normalized fluorescence intensity, indicated by the color bar,
plotted as a function of time and emission wavelength. Spectrally
integrated donor counts are plotted on both linear (C, D) and
semilogarithmic (E, F) axes. Solid black lines are fits to the
experimental data (open circles).
fluorescence decay rate (kpure) is the sum of the intrinsic
radiative (kr) and nonradiative (knr) decay rates
k pure = k r + k nr
In the mixed donor−acceptor film, the total fluorescence decay
rate (kmix) is
k mix = k r + k nr + ∑ kET
where the summation is over all possible FRET acceptors.
Therefore, the total energy transfer rate may be calculated by
subtracting eq 3 from eq 4. Because of the 1/d6 distance scaling,
only nearest neighbors are expected to contribute substantially
to the total energy transfer rate.26 Therefore, the per-pair
energy transfer rate (kET,pp) can be calculated from
1 distances (Figure 1) is expected to result in a distribution of
kET,pp = (k mix − k pure)
N (5)
where N is the number of nearest neighbors. Danisch et al.
previously determined the approximate number of nearest
neighbors of a given size, given a random close packed lattice of
two different sized spheres with size ratio 1:1.4 (the ratio of
donor and acceptor QD diameters in our study).42 For an
equimolar mixture of donor and acceptor QDs with this size
ratio, N equals 3 (see expanded discussion of experimental
uncertainty below). Using eq 5, the lifetime comparison yields a
direct experimental measurement of the energy transfer rate,
Article
which can be used, in turn, to calculate the effective Förster
radius.
Correctly applying eq 5 requires comparison between the
donor lifetime in the absence and presence of energy transfer.
For this purpose, it is essential to use the solution decay rate for
kpure rather than the pure film lifetime because energy transfer
within the inhomogeneously broadened pure donor ensemble26
and exciton diffusion to quenching sites3,43 obfuscates the true
intrinsic decay rate of a QD in the solid film. We verify that
energy transfer occurs within the pure donor film by
observation of a spectral red-shift over the course of the
exciton lifetime (Figures 4B,D), resulting in an emission-
Figure 4. Spectrally resolved transient photoluminescence of (A) pure
donor solution and (B) pure donor film. Population decay at selected
wavelengths are shown in panels C and D for the donor solution and
film, respectively. Decay traces range in wavelength from 520 nm
(purple) to 590 nm (red), in increments of 10 nm.
energy-dependent lifetime not observed in donor solutions
(3) (Figures 4A,C).26 We note that the lifetime of the reddest QDs
in the pure donor film approaches the lifetime of the donor
QDs in solution. This observation further confirms that the
solution lifetime is the inherent no-energy-transfer lifetime of
the QDs in the ensemble, as these reddest dots have no
(4)
neighbors capable of accepting a band-edge exciton and are
therefore effectively isolated. Noninteracting QDs in solution
and film have the same inherent lifetime, so the solution lifetime
and quantum efficiency (rather than the pure donor film
lifetime and quantum efficiency) must be used to predict
energy transfer dynamics in blended QD assemblies.
In the blended QD films, the distribution of donor−acceptor
energy transfer rates and, accordingly, a distribution of donor
fluorescence lifetimes. To obtain a statistically averaged lifetime,
we first fit the donor fluorescence decay to a multiexponential
function to deconvolve the time-dependent exciton population
n(t) from the instrument response function (determined almost
entirely by the temporal width of the laser pulse). The
statistically averaged lifetime, τavg, is then obtained via
∞
∫0 tn(t ) dt
τavg = ∞
∫0 n(t ) dt (6)
E dx.doi.org/10.1021/jp502123n | J. Phys. Chem. C XXXX, XXX, XXX−XXX
The Journal of Physical Chemistry C
Table 1. Summary of Theoretical and Experimental Energy Transfer Rates (per Donor−Acceptor pair, pp)
measured donor−acceptor distance
(nm) R0 (nm) from theory
short ligand 7.0 (±0.9) 5.2 (±0.2)
long ligand 7.8 (±1.1) 5.1 (±0.2)
The lifetime of the donor in solution is 15.5 ± 0.3 ns (kpure =
0.065 ± 0.001 ns−1), while the average lifetime of the donor in
the short-ligand acceptor blend is 1.2 ± 0.2 ns (kmix = 0.8 ± 0.1
ns−1) and the average lifetime of the donor in the long-ligand
acceptor blend is 4.1 ± 0.2 ns (kmix = 0.24 ± 0.01 ns−1) (Figure
3). Using these values and eq 5, we find that the energy transfer
rate per donor−acceptor pair in the short-ligand blend is kET,pp
= 2.6 (±0.6) × 108 s−1, while for the long-ligand blend the
energy transfer rate is kET,pp = 0.6 (±0.1) × 108 s−1. An
experimental R0 can be calculated from these values by
rearranging eq 1
R 0 = d(τpurekET,pp)1/6 (7)
The experimentally determined values of R0 are 9 ± 2 nm for
the donor-short-ligand acceptor pair and 8 ± 1 nm for the
donor-long ligand acceptor pair. These numbers are
summarized in Table 1.
Discrepancy between Theory and Experiment. The
theoretical Förster radii and energy transfer rates calculated
from materials properties are substantially smaller than those
measured experimentally (see Table 1). The expected
theoretical per-pair energy transfer rates based on the overlap
integral and interparticle separation are 10 (±8) × 106 and 5
(±4) × 106 s−1 for the donor paired with the short-ligand
acceptor and long-ligand acceptor, respectively. However, the
measured energy transfer rates are more than an order of
magnitude larger, 260 (±60) × 106 and 60 (±10) × 106 s−1 for
the two cases, respectively (Table 1). Calculated Förster energy
transfer rates to second nearest neighbors are on the order of
104 s−1 and therefore contribute negligibly to the total per-pair
energy transfer rate. The discrepancy between theoretical
predictions and experimental measurements in our model
donor−acceptor system suggests that application of Förster’s
equation to solid state QD systems requires further
consideration, as we now address.
Experimental Uncertainty. Obvious experimental uncer-
tainty cannot account for the magnitude of the difference
between experimentally measured and theoretically predicted
energy transfer rates and Förster radii. There is some systematic
error in the measurement of donor−acceptor separation
distance due to an inability of the image processing algorithm
to distinguish QDs that appear very close together in the TEM
micrograph. This leads to a systematic overestimation of the
interparticle separation on the order of a few tenths of
nanometers, but this error is small compared to the propagated
uncertainty of ∼1.0 nm for all calculations and cannot explain
the discrepancy between theory and experiment. Moreover, the
excellent agreement between TEM and GISAXS data suggests
that the propagated experimental uncertainty captures any error
in interparticle distance.
Experimental determination of kET and R0 depends on
knowledge of the number of nearest neighbors, as specified in
eq 5. However, even if there are twice as many nearest acceptor
neighbors as we expect (N = 6 instead of N = 3), the
experimentally measured energy transfer rate remains ∼5−15
times larger than the predicted value. Further, because our QD
F dx.doi.org/10.1021/jp502123n | J. Phys. Chem. C XXXX, XXX, XXX−XXX
Article
kET,pp (s−1) from
kET,pp (s−1) from theory experiment R0 (nm) from experiment
10 (±8) × 106 260 (±60) × 106 9 (±2)
5 (±4) × 106 60 (±10) × 106 8 (±1)
films are only ∼3 QD layers thick and surface QDs have fewer
nearest neighbors, we have probably underestimated the per-pair
energy transfer rate by choosing a bulk value for N.
Additionally, the TEM images shown in Figure 1 indicate
that the QDs are not entirely randomly distributed in a
homogeneously blended film. If slight aggregation of similar
types of QDs occurs within our 3D films, this would also
decrease the true value of N to less than 3. A smaller number of
nearest acceptor neighbors per donor QD further exaggerates
the discrepancy between theoretical and measured energy
transfer rates and reinforces our claim that actual energy
transfer rates are faster than those predicted by standard theory.
These factors, as well as errors in measurement of quantum
efficiency, lifetime, and other experimental parameters, have
been quantitatively accounted for in the propagated uncertainty
and are insufficient in explaining the difference between
expected and measured energy transfer rates. We believe that
other factors play a role beyond those usually considered in the
calculation of the Förster radius. We consider some possibilities
below.
Dipole Orientation. As seen in the TEM images of Figure 1,
the QDs used in this study are not spherical but instead have a
nearly tetrahedral geometry, as discussed in a previous study
that utilized these same materials.6 This may lead to some
oriented packing and alignment of the QDs in thin film,
resulting in a transition dipole that is not random but instead
aligned. Furthermore, the observed 2D character of the
emission transition dipole in CdSe QDs makes better-than-
random alignment between donor emission and acceptor
absorption transition dipole moments even more plausible.44
Murray et al. found that QDs within superlattices tend to align
their crystal axes,45 suggesting ready alignment of transition
dipoles in some QD assemblies. If the transition dipoles were
perfectly aligned (κ2 = 4),27 this would increase the calculated
R0 from 5.1 nm to about 7.0 nm and thereby increase the
theoretical energy transfer rate by about a factor of 7 for both
films. This cannot completely account for the discrepancy
between theory and experiment, and furthermore, it is unlikely
that all of the transition dipoles in our mixed-size films are
perfectly aligned because of geometric packing restrictions.
Consequently, dipole orientation is likely a contributing cause
of the observed discrepancy, but only a minority factor.
QD Shape. In addition to the QD shape affecting packing
and transition dipole alignment, it is also possible that shape
affects Coulombic and dipole−dipole coupling between
neighboring nanocrystals.46 Several reports in the literature
have demonstrated different energy transfer behavior in systems
of QDs compared to quantum rods or dot-in-rod struc-
tures.46,47 On the basis of published reports of the magnitude of
this affect, we estimate that shape could account for up to 10−
20% of the discrepancy.
Higher Order Multipoles. In addition to these physical
factors that may be incorrectly accounted for in the usual
calculations of R 0 and k ET , the discrepancy may be
fundamentally due to the assumption that QD−QD energy
transfer in thin film is mediated by point dipole−point dipole
The Journal of Physical Chemistry C
coupling. Baer and Rabani showed that for small surface-to-
surface separations dipole−quadrupole coupling may contrib-
ute non-negligibly to the total energy transfer rate.48 In
particular, for surface-to-surface distances less than about 20%
of the total center-to-center distance, dipole−quadrupole
coupling may contribute more than a third to the combined
energy transfer rate. For this model system, the expected
difference in energy transfer rates from donor to acceptor for
the two test cases was a factor of 2, as described above;
however, a factor of 4 difference was observed experimentally.
The relative change in energy transfer rate from the donor to
the short-ligand acceptor vs long-ligand acceptor may be due in
part to dipole−multipole interactions as the surface-to-surface
separation is 25% of the total center-to-center separation in the
case of the short-ligand acceptor (putting it into the realm of
more significant dipole−quadrupole coupling) but more than
40% of the center-to-center distance in the case of the long
ligand acceptor. As a result, the larger than expected difference
between the measured energy transfer rates with the two
acceptors could be due in part to dipole−quadrupole coupling,
which contributes to the energy transfer rate in the case of the
shorter-ligand acceptor but has a negligible effect for the longer-
ligand acceptor. Contributions from dipole−quadrupole
coupling may account for part of the magnitude difference
between the measured and theoretical energy transfer rates, but
the exact contribution is difficult to quantify.
Absorption Cross Section. The absorption cross section of
the acceptor QDs sensitively affects the calculated overlap
integral, and therefore the Förster radius, for a donor−acceptor
pair. However, there are inconsistencies in the literature as to
the magnitude of the absorption cross section as a function of
QD size. In this work, the calculated Förster radius changes
between 4.8 and 6.2 nm depending on the value of the
absorption cross section used. CdZnS shells surround the
nanoparticles used in this study, which complicates the
determination of absorption cross section. As a result, the
work of De Geyter et al., which directly addresses this issue, was
used for calculating absorption cross section.41 However, other
methods35,49,50 for calculating the absorption cross section can
yield slightly different values that may change the absorption
cross section and therefore the Förster radius. The lack of
experimental consensus renders exact calculation of the
absorption cross section difficult and highlights the need for
improved data in this important area of QD research.
In addition to variation in the calculated absorption cross
section depending on method, a further likely contribution to
the larger than expected energy transfer rate is an enhanced
absorption cross section of the acceptor QDs in film compared
to solution. Recently, Geiregat et al. showed that the absorption
cross section per QD in close-packed films is enhanced relative
to the solution value due to dipole-induced-dipole effects in the
solid.23 For CdSe QDs of the sizes used here, the absorption
cross section in thin film was measured to be 3−4 times larger
than the absorption cross section in solution.23 If the measured
absorption spectrum is scaled to this new, larger magnitude
value (and possible dipole orientation effects ignored), the
theoretical value for R0 increases from 5.1−5.2 to 6.2 nm for
both donor−acceptor pairs. Furthermore, these polarization
effects should play a larger role in short-ligand QD assemblies
due to closer proximity of neighboring QDs.23 We indeed
observe a larger deviation from the predicted theoretical energy
transfer rate in the system with the shorter ligand acceptor than
that with the longer ligand acceptor, consistent with stronger
G dx.doi.org/10.1021/jp502123n | J. Phys. Chem. C XXXX, XXX, XXX−XXX
Article
absorption enhancement in the denser QD array. These factors
suggest that the results of a full ligand series to incrementally
change the interparticle separation may be difficult to interpret.
Even if Förster theory is the correct theoretical framework, such
a study would likely not result in the expected 1/d6 scaling
because the polarizability (acceptor absorptivity) of the
nanocrystal assembly, its refractive index, and the interparticle
separation all change simultaneously in different directions.
■ CONCLUSIONS
We have measured the energy transfer rate between donor and
acceptor QDs in homogeneously blended films, where the
interparticle separation was tuned through altering the ligand
length of the acceptor. The measured energy transfer rates were
found to be more than an order of magnitude larger than the
theoretical energy transfer rates calculated based on Förster
theory, using standard assumptions. Several possible sources of
the discrepancy were identified and quantified, including QD
shape, preferential orientation of transition dipoles, enhanced
absorption cross section, and dipole−multipole coupling; these
factors may result in different behavior for QDs in solid films
than for QDs in solution. To further elucidate the mechanisms
underlying energy transfer in CdSe QD assemblies, further
experimentation may focus on the direct comparison of energy
transfer between isolated donor−acceptor pairs in solution to
that in solid film. This work suggests that predicting energy
transfer between QDs at small distances within a close-packed
film requires careful consideration and informs our under-
standing of excitonic energy transfer in QD films relevant for
optoelectronic devices such as solar cells and LEDs.
■ AUTHOR INFORMATION
Corresponding Author
*E-mail tisdale@mit.edu (W.A.T.).
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS
The authors thank Liza Lee for helpful discussions, Kevin Yager
for performing the GISAXS measurements, and Jonathan
Steckel and Seth Coe-Sullivan of QD Vision, Inc., for supplying
QD materials. This work was supported as part of the Center
for Excitonics, an Energy Frontier Research Center funded by
the U.S. Department of Energy, Office of Science, Office of
Basic Energy Sciences, under Award DE-SC0001088 (MIT).
A.J.M. and M.C.W. gratefully acknowledge fellowship support
from the National Science Foundation Graduate Research
Fellowship Program under Grant No. 1122374. GISAXS
measurements were performed at the Center for Functional
Nanomaterials and the National Synchrotron Light Source,
Brookhaven National Laboratory, which is supported by the
U.S. Department of Energy, Office of Basic Energy Sciences,
under Contract DE-AC02-98CH10886.
■ ABBREVIATIONS
QD, quantum dot; FRET, Förster resonant energy transfer;
CdSe, cadmium selenide; CdZnS, cadmium zinc sulfide.
■ REFERENCES
(1) Colvin, V. L.; Schlamp, M. C.; Alivisatos, A. P. Light-Emitting
Diodes Made from Cadmium Selenide Nanocrystals and a Semi-
conducting Polymer. Nature 1994, 370, 354−357.
The Journal of Physical Chemistry C
(2) Coe, S.; Woo, W. K.; Bawendi, M.; Bulovic, V. Electro-
luminescence from Single Monolayers of Nanocrystals in Molecular
Organic Devices. Nature 2002, 420, 800−803.
(3) Shirasaki, Y.; Supran, G. J.; Bawendi, M. G.; Bulovic, V.
Emergence of Colloidal Quantum-Dot Light-Emitting Technologies.
Nat. Photonics 2013, 7, 13−23.
(4) Kamat, P. V. Quantum Dot Solar Cells. Semiconductor
Nanocrystals as Light Harvesters. J. Phys. Chem. C 2008, 112,
18737−18753.
(5) Pattantyus-Abraham, A. G.; Kramer, I. J.; Barkhouse, A. R.; Wang,
X.; Konstantatos, G.; Debnath, R.; Levina, L.; Raabe, I.; Nazeeruddin,
M. K.; Sargent, E. H. Depleted-Heterojunction Colloidal Quantum
Dot Solar Cells. ACS Nano 2010, 4, 3374−3380.
(6) Dang, C.; Lee, J.; Breen, C.; Steckel, J. S.; Coe-Sullivan, S.;
Nurmikko, A. Red, Green and Blue Lasing Enabled by Single-Exciton
Gain in Colloidal Quantum Dot Films. Nat. Nanotechnol. 2012, 335−
339.
(7) Klimov, V. I.; Ivanov, S. A.; Nanda, J.; Achermann, M.; Bezel, I.;
McGuire, J. A.; Piryatinski, A. Single-Exciton Optical Gain in
Semiconductor Nanocrystals. Nature 2007, 447, 441−446.
(8) Konstantatos, G.; Howard, I.; Fischer, A.; Hoogland, S.; Clifford,
J.; Klem, E.; Levina, L.; Sargent, E. H. Ultrasensitive Solution-Cast
Quantum Dot Photodetectors. Nature 2006, 442, 180−183.
(9) Shcherbatyuk, G. V.; Inman, R. H.; Wang, C.; Winston, R.;
Ghosh, S. Viability of Using Near Infrared PbS Quantum Dots as
Active Materials in Luminescent Solar Concentrators. Appl. Phys. Lett.
2010, 96, 191901.
(10) Bruchez, M.; Moronne, M.; Gin, P.; Weiss, S.; Alivisatos, A. P.
Semiconductor Nanocrystals as Fluorescent Biological Labels. Science
1998, 281, 2013−2016.
(11) Medintz, I. L.; Uyeda, H. T.; Goldman, E. R.; Mattoussi, H.
Quantum Dot Bioconjugates for Imaging, Labelling and Sensing. Nat.
Mater. 2005, 4, 435−46.
(12) Bourzac, K. Quantum Dots Go on Display. Nature 2013, 493,
283−283.
(13) Anikeeva, P. O.; Halpert, J. E.; Bawendi, M. G.; Bulović, V.
Electroluminescence from a Mixed Red-Green-Blue Colloidal
Quantum Dot Monolayer. Nano Lett. 2007, 7, 2196−200.
(14) Akselrod, G. M.; Prins, F.; Poulikakos, L. V.; Lee, E. M. Y.;
Weidman, M. C.; Mork, A. J.; Willard, A. P.; Bulović, V.; Tisdale, W. A.
Subdiffusive Exciton Transport in Quantum Dot Solids. Nano Lett.
2014, in press.
(15) Achermann, M.; Petruska, M. A.; Crooker, S. A.; Klimov, V. I.
Picosecond Energy Transfer in Quantum Dot Langmuir-Blodgett
Nanoassemblies. J. Phys. Chem. B 2003, 107, 13782−13787.
(16) Xu, F.; Ma, X.; Haughn, C. R.; Benavides, J.; Doty, M. F.;
Cloutier, S. G. Efficient Exciton Funneling in Cascaded PbS Quantum
Dot Superstructures. ACS Nano 2011, 5, 9950−9957.
(17) Medintz, I. L.; Clapp, A. R.; Brunel, F. M.; Tiefenbrunn, T.;
Uyeda, H. T.; Chang, E. L.; Deschamps, J. R.; Dawson, P. E.;
Mattoussi, H. Proteolytic Activity Monitored by Fluorescence
Resonance Energy Transfer through Quantum-Dot-Peptide Con-
jugates. Nat. Mater. 2006, 5, 581−9.
(18) Medintz, I. L.; Clapp, A. R.; Mattoussi, H.; Goldman, E. R.;
Fisher, B.; Mauro, J. M. Self-Assembled Nanoscale Biosensors Based
on Quantum Dot FRET Donors. Nat. Mater. 2003, 2, 630−8.
(19) Allan, G.; Delerue, C. Energy Transfer between Semiconductor
Nanocrystals: Validity of Förster’s Theory. Phys. Rev. B 2007, 75,
195311.
(20) Curutchet, C.; Franceschetti, A.; Zunger, A.; Scholes, G. D.
Examining Forster Energy Transfer for Semiconductor Nanocrystalline
Quantum Dot Donors and Acceptors. J. Phys. Chem. C 2008, 112,
13336−13341.
(21) Medintz, I. L.; Mattoussi, H. Quantum Dot-Based Resonance
Energy Transfer and Its Growing Application in Biology. Phys. Chem.
Chem. Phys. 2009, 11, 17−45.
(22) Pons, T.; Medintz, I. L.; Sapsford, K. E.; Higashiya, S.; Grimes,
A. F.; English, D. S.; Mattoussi, H. On the Quenching of
H dx.doi.org/10.1021/jp502123n | J. Phys. Chem. C XXXX, XXX, XXX−XXX
Article
Semiconductor Quantum Dot Photoluminescence by Proximal Gold
Nanoparticles. Nano Lett. 2007, 7, 3157−3164.
(23) Geiregat, P.; Justo, Y.; Abe, S.; Flamee, S.; Hens, Z. Giant and
Broad-Band Absorption Enhancement in Colloidal Quantum Dot
Monolayers through Dipolar Coupling. ACS Nano 2013, 7, 987−993.
(24) Wolcott, A.; Doyeux, V.; Nelson, C. A.; Gearba, R.; Lei, K. W.;
Yager, K. G.; Dolocan, A. D.; Williams, K.; Daniel, N.; Zhu, X. Y.
Anomalously Large Polarization Effect Responsible for Excitonic Red-
Shifts in PbSe Quantum Dot Solids. J. Phys. Chem. Lett. 2011, 2, 795−
800.
(25) Kagan, C.; Murray, C.; Nirmal, M.; Bawendi, M. Electronic
Energy Transfer in CdSe Quantum Dot Solids. Phys. Rev. Lett. 1996,
76, 1517−1520.
(26) Crooker, S.; Hollingsworth, J.; Tretiak, S.; Klimov, V. Spectrally
Resolved Dynamics of Energy Transfer in Quantum-Dot Assemblies:
Towards Engineered Energy Flows in Artificial Materials. Phys. Rev.
Lett. 2002, 89, 186802.
(27) Medintz, I.; Hildebrandt, N. FRET - Förster Resonance Energy
Transfer: From Theory to Applications; Wiley-VCH: Weinheim,
Germany, 2013.
(28) Kagan, C.; Murray, C.; Bawendi, M. Long-Range Resonance
Transfer of Electronic Excitations in Close-Packed CdSe Quantum-
Dot Solids. Phys. Rev. B 1996, 54, 8633−8643.
(29) Lunz, M.; Bradley, A. L.; Chen, W.-y.; Gun’ko, Y. K. Two-
Dimensional Forster Resonant Energy Transfer in a Mixed Quantum
Dot Monolayer: Experiment and Theory. J. Phys. Chem. C 2009, 113,
3084−3088.
(30) Kim, D.; Okahara, S.; Nakayama, M.; Shim, Y. Experimental
Verification of Förster Energy Transfer between Semiconductor
Quantum Dots. Phys. Rev. B 2008, 78, 153301.
(31) Moreels, I.; Justo, Y.; Rainò, G.; Stöferle, T.; Hens, Z.; Mahrt, R.
F. Impact of the Band-Edge Fine Structure on the Energy Transfer
between Colloidal Quantum Dots. Adv. Opt. Mater. 2013, 2, 126−130.
(32) Lingley, Z.; Lu, S.; Madhukar, A. A High Quantum Efficiency
Preserving Approach to Ligand Exchange on Lead Sulfide Quantum
Dots and Interdot Resonant Energy Transfer. Nano Lett. 2011, 11,
2887−2891.
(33) Lunz, M.; Bradley, A. L.; Chen, W.-Y.; Gerard, V. A.; Byrne, S.
J.; Gun’ko, Y. K.; Lesnyak, V.; Gaponik, N. Influence of Quantum Dot
Concentration on Förster Resonant Energy Transfer in Monodis-
persed Nanocrystal Quantum Dot Monolayers. Phys. Rev. B 2010, 81,
205316.
(34) Poulikakos, L. V.; Prins, F.; Tisdale, W. A. Transition from
Thermodynamic to Kinetic-Limited Excitonic Energy Migration in
Colloidal Quantum Dot Solids. J. Phys. Chem. C 2014, 118, 7894−
7900.
(35) Yu, W. W.; Qu, L.; Guo, W.; Peng, X. Experimental
Determination of the Extinction Coefficient of CdTe, CdSe, and
CdS Nanocrystals. Chem. Mater. 2003, 125, 2854−2860.
(36) Segets, D.; Lucas, J. M.; Klupp Taylor, R. N.; Scheele, M.;
Zheng, H.; Alivisatos, A. P.; Peukert, W. Determination of the
Quantum Dot Band Gap Dependence on Particle Size from Optical
Absorbance and Transmission Electron Microscopy Measurements.
ACS Nano 2012, 6, 9021−9032.
(37) Yoneda, Y. Anomalous Surface Reflection of X Rays. Phys. Rev.
1963, 131, 2010−2013.
(38) Zheng, K.; Ž ídek, K.; Abdellah, M.; Zhu, N.; Chábera, P.;
Lenngren, N.; Chi, Q.; Pullerits, T. Directed Energy Transfer in Films
of CdSe Quantum Dots: Beyond the Point Dipole Approximation. J.
Am. Chem. Soc. 2014, 136, 6259−6268.
(39) Fö rster, T. Zwischenmolekulare Energiewanderung Und
Fluoreszenz. Ann. Phys. 1948, 437, 55−75.
(40) Powell, R. C.; Soos, Z. G. Singlet Exciton Energy Transfer in
Organic Solids. J. Lumin. 1975, 11, 1−45.
(41) De Geyter, B.; Hens, Z. The Absorption Coefficient of PbSe/
CdSe Core/Shell Colloidal Quantum Dots. Appl. Phys. Lett. 2010, 97,
161908.
(42) Danisch, M.; Jin, Y.; Makse, H. A. Model of Random Packings
of Different Size Balls. Phys. Rev. E 2010, 81, 051303.
The Journal of Physical Chemistry C Article

(43) Miyazaki, J.; Kinoshita, S. Site-Selective Spectroscopic Study on

the Dynamics of Exciton Hopping in an Array of Inhomogeneously
Broadened Quantum Dots. Phys. Rev. B 2012, 86.
(44) Chung, I.; Shimizu, K. T.; Bawendi, M. G. Room Temperature
Measurements of the 3D Orientation of Single CdSe Quantum Dots
Using Polarization Microscopy. Proc. Natl. Acad. Sci. U. S. A. 2003,
100, 405−408.
(45) Murray, C. B.; Kagan, C. R.; Bawendi, M. G. Self-Organization
of CdSe Nanocrystallites into Three-Dimensional Quantum Dot
Superlattices. Science 1995, 270, 1335−1338.
(46) Schrier, J.; Wang, L.-W. Shape Dependence of Resonant Energy
Transfer between Semiconductor Nanocrystals. J. Phys. Chem. C 2008,
112, 11158−11161.
(47) Halivni, S.; Sitt, A.; Hadar, I.; Banin, U. Effect of Nanoparticle
Dimensionality on Fluorescence Resonance Energy Transfer in
Nanoparticle−Dye Conjugated Systems. ACS Nano 2012, 6, 2758−
2765.
(48) Baer, R.; Rabani, E. Theory of Resonance Energy Transfer
Involving Nanocrystals: The Role of High Multipoles. J. Chem. Phys.
2008, 128, 184710.
(49) Jasieniak, J.; Smith, L.; Embden, J. V.; Mulvaney, P.; Califano,
M. Re-Examination of the Size-Dependent Absorption Properties of
CdSe Quantum Dots. J. Phys. Chem. C 2009, 113, 19468−19474.
(50) Leatherdale, C. A.; Mikulec, F. V.; Bawendi, M. G. On the
Absorption Cross Section of CdSe Nanocrystal Quantum Dots. J. Phys.
Chem. B 2002, 106, 7619−7622.

I dx.doi.org/10.1021/jp502123n | J. Phys. Chem. C XXXX, XXX, XXX−XXX