LETTER pubs.acs.org/JPCL
Anomalously Large Polarization Effect Responsible for Excitonic Red Shifts in PbSe Quantum Dot Solids Abraham Wolcott, Vincent Doyeux,† Cory A. Nelson, Raluca Gearba, Kin Wai Lei, Kevin G. Yager,§ Andrei D. Dolocan, Kenrick Williams, Daniel Nguyen, and X.-Y. Zhu* Department of Chemistry and Biochemistry, University of Texas, Austin, Texas 78712, United States
bS Supporting Information ABSTRACT: The formation of solid thin films from colloidal semiconductor quantum dots (QDs) is often accompanied by red shifts in excitonic transitions, but the mechanisms responsible for the red shifts are under debate. We quantitatively address this issue using optical absorption spectroscopy of two-dimensional (2D) and three-dimensional (3D) arrays of PbSe QDs with controlled inter-QD distance, which was determined by the length of alkanedithiol linking molecules. With decreasing inter-QD distance, the first and second exciton absorption peaks show increasing red shifts. Using thin films consisting of large and isolated QDs embedded in a matrix of small QDs, we determine that a dominant contribution to the observed red shift is due to changes in polarization of the dielectric environment surrounding each QD (∼88%), while electronic or transition dipole coupling plays a lesser role. However, the observed red shifts are more than 1 order of magnitude larger than theoretical predictions based on the dielectric polarization effect for spherical QDs. We attribute this anomalously large polarization effect to deviations of the exciton wave functions from eigenfunctions of the idealized spherical quantum well model. SECTION: Nanoparticles and Nanostructures
S
emiconductor quantum dots (QDs) are called artificial atoms because quantum confinement of charge carriers or excitons leads to an atomic orbital-like discrete electronic structure. These artificial atoms can also be assembled into artificial solids with unique electronic and optoelectronic properties determined not only by the properties of individual QDs, but also their mutual interactions. In principle, one can control the strength of QDQD interaction by varying the length (thus, the inter-QD distance) or electronic structure of linking or capping molecules. Colloidal QDs are usually synthesized with insulating capping molecules that lead to poor interQD coupling in QD solids. Removing or replacing these insulating capping molecules with short and/or conjugated ones in the thin film has been shown to dramatically increase electrical conductivity.1 It is commonly observed that when a close-packed QD solid is formed from a colloidal QD solution, there is a red shift in the first exciton transition in an optical absorption spectra.16 Further red shift of the excitonic absorption peak is also observed as the inter-QD distance decreases due to the exchange of long capping molecules by shorter ones.16 While there have been attempts at correlating such red shifts with electronic coupling strength, excitonic red shifts can also result from varying polarization effects (i.e., solvatochromism) due to changes in the local dielectric environment7,8 and to exciton delocalization among neighboring QDs;9 the latter is better known from studies of molecular J-aggregates.10,11 r 2011 American Chemical Society
There have been a number of studies on the excitonic red shift in QD solids. Vossmeyer et al. investigated CdS nanoclusters in the diameter range of 1.33.9 nm and showed large optical red shifts of up to 250 meV upon the formation of solid thin films from the colloidal solution; the magnitude of the red shift was found to decrease with increasing nanocluster size.12 These authors suggested inter-QD electronic coupling as the source of the red shift but also commented on the (unknown) role of polarization effects. Letherdale and Bawendi investigated CdSe QDs in solvents of different dielectric constants and in solid thin films; they attributed the small excitonic red shift (∼24 meV) in CdSe solids to solvatochromism based on a polarization model of a core shellmedium structure.13 Kim et al. reinvestigated this system as a function of hydrostatic pressure and found negligible difference between the CdSe QDs in the solution or solid state; optical red shift in the solid thin film was only observed with the small molecule pyridine acting as a ligand.14 In an examination of stoichiometric CdS clusters, they explained the excitonic red shift upon solid thin-film formation by a dipoleinduced dipole interaction, which was essentially a simplified version of the polarization Received: January 18, 2011 Accepted: March 9, 2011 Published: March 16, 2011 795
dx.doi.org/10.1021/jz200080d | J. Phys. Chem. Lett. 2011, 2, 795–800
The Journal of Physical Chemistry Letters
LETTER
model.15,16 Utilizing photoluminescence spectroscopy, it was shown that transition dipoletransition dipole interaction was responsible for long-range resonance energy transfer in CdSe QD solids, but this interaction did not lead to red shifts in optical absorption spectra.17,18 Most recently, Frederick and Weiss reported large optical shifts of up to 220 meV when they exchanged oleic acid (OA) ligands with phenyldithiocarbamate; they attributed the red shift to delocalization of the hole wave function on the QD into the ligand.19 A number of studies have focused on IVVI semiconductor QD thin films due to their potentially attractive applications in electronics and photovoltaics. Bulk Pb-chalcogenide semiconductors are characterized by narrow band gaps (Eg), large bulk exciton Bohr radii (rB), and large dielectric constants (εo and ε¥ for static and optical dielectric constants, respectively). For example, the values for PbSe are Eg = 0.28 eV, rB = 46 nm, εo = 250, and ε¥= 24.20 QDs of these materials with diameters in the nanometer range allow the exploration of quantum confinement effects much stronger than those in IIVI (e.g., CdSe) materials. Talapin and Murray studied thin films of PbSe QDs (8 nm) and showed an increase in electrical conductance by 10 orders of magnitude when they removed the insulating OA capping molecules by treatment with hydrazine.1 The drastic increase in electrical conductivity strongly suggested inter-QD electronic coupling for the closely contacting PbSe QDs; this conclusion was also corroborated by the 20 meV red shift in the first exciton absorption peak upon the removal of OA. Nozik and co-workers examined PbSe QD thin films upon a variety of chemical treatments and correlated increased electrical conductivity and red shift in the first exciton absorption peak with decreasing inter-QD distance.2,3 Williams et al. examined the optical red shift as a function of OA ligand removal using a hydrazine treatment in two-dimensional (2D) PbSe QD films; they observed maximum optical shifts of 82 and 157 meV for 6 and 4 nm QDs, respectively, and attributed these to inter-QD electronic coupling.4 A similar conclusion was reached by Liu et al., who replaced OA capping ligands in PbSe QD films with several alkanedithiols.5 The most recent investigation of PbSe QD thin films studied photoinduced exciton dissociation after ligand exchanges with short (2 carbon) and conjugated dithiols.6 With the shortest linker (1,2-ethanedithiol), these authors reported the largest red shifts (170 and 20 meV for QD diameters of 2.7 and 6.5 nm, respectively) in the first exciton absorption peak and concluded that the red shifts were due to strong electronic coupling; they argued that the electronic coupling was large enough to overcome the exciton binding energy, resulting in direct photocarrier formation. Despite all of these suggestions for strong electronic coupling, the polarization effect is always present and should contribute to the excitonic red shift. The key question is, how can one quantitatively separate the polarization effect from electronic coupling and exciton coupling? We attempt to quantitatively answer the above question using the model system of PbSe QD thin films of 0.52.5 monolayer thickness. We control the inter-QD distance using a series of alkanedithiol capping molecules, HS(CH2)nSH, and quantify this distance in real and reciprocal space using scanning electron microscope (SEM) imaging and grazing-incidence wide-angle X-ray scattering (GIWAXS), respectively. As expected, the inter-QD distance decreases with the length of the capping molecules, while the magnitude of the red shift in optical absorption increases. On the basis of a quantitative comparison of a thin film of monodispersed QDs (6.2 nm) with that
consisting of isolated QDs (6.2 nm) embedded in a matrix of smaller ones (4.8 nm) at a ratio of 1:35, we show that the dominant contribution to the optical red shift following the exchange of OA capping molecules by ethanedithiols comes from polarization effects (88%), with a minority (12%) attributed to resonant electronic and exciton coupling. More importantly, this polarization effect is more than one order of magnitude larger than theoretical predictions based on the dielectric polarization effect for QDs in an idealized spherical quantum well model. PbSe QD Films. We prepared PbSe QDs using the hot injection technique with the addition of diphenyl phosphine to increase nanocrystal yield, as described previously.1,21 The OA-capped QDs were dispersed in tetrachloroethylene (TCE) or hexane for further use. Detailed characterization of these QDs is summarized in the Supporting Information (Figure S1). The optical absorption spectrum for the solution sample showed the first excitonic peak at 0.727 eV with a full width at half-maximum (fwhm) of 45 meV. A combination of SEM and scanning transmission electron microscopy (STEM) was used to determine the size distribution, which corresponds to an average particle diameter (D) of 5.4 ( 0.3 nm. TEM also revealed the crystal lattice of the single crystal PbSe QDs. We prepared OA-capped PbSe QD thin films via a dip-coating method on oxide-terminated silicon waveguide surfaces. We obtained a desired QD coverage on the silicon oxide surface by varying the withdraw speed in the dipcoating process and verified the surface coverage by atomic force microscopy and SEM. To exchange the OA (C18) capping ligands, we immersed the QD thin film coated samples in 0.1 M solutions of the corresponding alkanedithiol molecules in anhydrous acetonitrile for 1 min, followed by acetonitrile rinsing and drying in an inert N2 environment. We used the following alkanedithiols: octanedithiol (HSC8H16SH, abbreviated as C8), hexanedithiol (HSC6H12SH, or C6), pentanedithiol (HSC5H10SH, or C5), butanedithiol (HSC4H8SH, or C4), propanedithiol (HSC3H6SH, or C3), and (HSC2H4 SH, or C2). As a control, we also completely removed capping molecules with hydrazine (Supporting Information, Figures S9 and S10). Structure and Inter-QD Distance. We determine the structure of PbSe QD thin films in real space by SEM and those in reciprocal space by GIWAXS. Figure 1 shows SEM images of approximately one monolayer (ML) coverage of PbSe QDs (5.4 ( 0.3 nm diameter) on SiO2 substrates before and after ligand exchange reactions with alkanedithiols. The as-deposited QD film with OA capping was highly ordered in a hexagonal close-pack arrangement. Replacing the OA capping molecules with alkanedithiols of decreasing length introduced more and more disorder and the breaking up of large QD islands into smaller domains. As expected, the inter-QD distance decreased as the length of the capping molecules became shorter. On the basis of autocorrelation function analysis on SEM images corresponding to the different treatments (see Supporting Information Figure S2 and S3), we quantitatively obtained inter-QD distance as a function of capping molecule length. The inter-QD distances were also obtained from GIWAXS analysis at both submonolayer (∼0.5 ML) and multilayer (∼2.5 ML) coverages (Supporting Information Figure S4). Figure 2a shows, as a function of capping molecule length, the measured inter-QD distance (d), which is defined as the edge-toedge nearest-neighbor distance based on a QD diameter of 5.4 ( 0.3 nm. Here, the data obtained from GIWAXS analysis (0.5 and 2.5 ML) and those from autocorrelation function analysis of SEM 796
dx.doi.org/10.1021/jz200080d |J. Phys. Chem. Lett. 2011, 2, 795–800
The Journal of Physical Chemistry Letters
LETTER
Figure 1. SEM images of ∼1 ML of PbSe QDs (D = 5.4 ( 0.3 nm) on SiO2 substrates. The QDs are capped with the following molecules: oleic acid (C18), octanedithiol (C8), hexanedithiol (C6), butanedithiol (C4), propanedithiol (C3), and ethanedithiol (C2). The SEM image for pentanedithiol (C5) is provided in Supporting Information Figure S2.
Figure 3. ATR-FTIR spectra for multilayer (∼2.5 ML, upper) and submonolayer (∼0.5 ML, lower) QD thin films with OA and various dithiol capping molecules. EX1 and EX2 denote the first and the second exciton transitions, respectively. CH: CH stretch vibrational transitions.
Red Shifts in Optical Absorption. Having established the structure and inter-QD distance in the QD thin films, we now turn to the red shift in excitonic transitions. Figure 3 shows the attenuated total reflectance Fourier transform infrared (ATRFTIR) absorption spectra of a multilayer (∼2.5 ML) and a submonolayer (∼0.5 ML) QD film with OA and dithiol capping molecules. The nearly complete exchange of OA molecules by alkanedithiols was verified by the decreasing CH stretch vibrational peak area with decreasing alkanedithiol molecular length; within experimental uncertainty, the integrated CH peak area was proportional to the number of CH2 repeating units in each molecule (Supporting Information Figure S5). While the first or second exciton peak of the OA-capped QDs showed little change going from the solution phase (in tetrachloroethylene or TCE) to the solid film, we saw an obvious red shift as the capping molecule length become shorter. Peak position and fwhm were obtained by fitting the first exciton absorption peak with a Gaussian function. As a further control, ligand exchange reactions were also carried out with 1,2-ethanedithiol and 1-ethanethiol in TCE solution and observed shifts (red and blue) of much smaller magnitudes (∼1/4, Supporting Information Figure S6). Note that there is no measurable broadening of the exciton peaks after the exchange reactions with the thin film samples (Supporting Information Figure S7). Figure 2b shows the measured red shift in the first exciton transition as a function of the length of the capping molecules. Within experimental uncertainty, we see no significant difference between the data set for the monolayer (open triangles) and that of the multilayer (solid triangles). We also used different substrate materials (SiO2 and TiO2) and found no measurable difference in the magnitudes of red shift. The dashed and dotted lines are theoretical predictions, as detailed later. The magnitude of the observed red shift in submonolayer films increased linearly from 22.3 ( 0.1 meV for 1,8-octanedithiol (C8) to 32.1 ( 1 meV for 1,2-ethanedithiol (C2). Note that for the same ligand exchange reaction, the magnitude of excitonic red shift increases with decreasing QD size (see below).
Figure 2. (a) Inter-QD distance (edge-to-edge) as a function of capping molecule length (number of C units) obtained from autocorrelation functions of SEM images (crosses, submonolayer) and GIWAXS analysis (open and solid circles for 0.5 and 2.5 ML, respectively). The solid line is a linear fit; (b) red shifts in the first exciton transition with respect to the OA-capped QDs as a function of capping molecule length (number of C units) for 0.5 (open triangles) and 2.5 ML (solid triangles) QD coverages. The solid and dashed lines (scaled by 20 and 35, respectively) are simulations, as detailed below.
images (0.5 ML) were in good agreement. For films of OA-capped QDs, the average inter-QD distance of d = 2.6 ( 0.3 nm was far less than the ∼4 nm value expected from the sum of two completely packed OA capping layers. In contrast, the OA insulatory shell on each QD are not believed to be close-packed, resulting in significant intercalation of the OA molecules between adjacent QDs. When the OA molecules were replaced by alkanedithiols, the inter-QD distance was consistently longer than the fully stretched molecular length in each case, indicating that intercalation rather than a single molecule bridge dominates. The solid line is a linear fit to the data and serves as guide to the eye. Optical red shifts as a function of capping molecule length in Figure 2b will be discussed later. 797
dx.doi.org/10.1021/jz200080d |J. Phys. Chem. Lett. 2011, 2, 795–800
The Journal of Physical Chemistry Letters
LETTER
each larger QD. In addition, inherent disorder of the QDs in terms of size, shape, and transition dipole orientation also minimizes any coupling between transition dipole moments. Kagan et al. showed that for monodispersed CdSe QD solids, transition dipoletransition dipole interaction was responsible for long-range resonance energy transfer and red shifts in fluorescence spectra, but not in optical absorption spectra.17,18 We conclude that the observed red shift of Δ = 21.4 ( 0.3 meV for the large QDs (D = 6.2 nm) in a bimodal film cannot come from inter-QD electronic or transition dipole coupling. Therefore this red shift must be attributed to changes in polarization effect when the C18 ligands are replaced by C2. In view of this conclusion from the bimodal sample, we can attribute the additional red shift of 3 meV in Figure 4b to resonant electronic interaction and/or exciton delocalization (transition dipoletransition dipole coupling) in the nearly monodispersed sample (D = 6.2 ( 0.4 nm). Thus, we attribute 88 ( 2% (=21.4/ 24.4) of the total excitonic red shift (Δ = 24.4 ( 0.3 meV) in Figure 4b to the polarization effect. The remaining 12 ( 2% in red shift maybe attributed to inter-QD electronic coupling and/or transition dipoletransition dipole coupling. Anomalously Large Polarization Effect. The red shift in optical absorption from the polarization effect (solvatochromism) can be understood as follows. Within the two-band nearly freeelectron model for a QD, the optical gap or the first exciton transition energy is given by7,22
Figure 4. (a) Optical absorption spectra for C18- and C2-capped QD films (∼2 ML film) of large QDs (D = 6.2 nm) embedded in a matrix of smaller ones (D = 4.8 nm) at a number ratio of 1:35. The first exciton peak of the large QDs red shifts by 21.4 ( 0.3 meV, while that of the small QDs red shifts by 50.5 ( 0.5 meV. The inset shows the SEM image of the mixed QD film; the minority of large QDs (bright spots) are completely isolated in the matrix of small QDs. (b) Optical absorption spectra for a ∼2 ML film of large QDs (D = 6.2 nm) before (blue) and after (red) ligand exchange reaction of C18 by C2. The first exciton red shifts by 21.4 ( 0.3 meV after the ligand exchange reaction.
kin dir Pol Pol Pol E1EX ¼ E0g þ Ekin e þ Eh þ ECoul þ Ee þ Eh þ Je;h
ð1Þ
kin where E0g is the bulk transport gap, Ekin e (Eh ) is the kinetic energy of the electron in the conduction band (hole in the valence band) due to quantum confinement, Edir Coul is the Coulomb attraction Pol between the electron and the hole within the QD, EPol e (Eh ) is the polarization stabilization energy, that is, the attraction between an electron (hole) in the QD and its image charge of opposite sign in the dielectric medium surrounding each QD, and EPol e,h is the polarization destabilization energy, that is, the repulsion between an electron (hole) and the image charge of the hole (electron). The first four terms are related only to the QD, while the last three terms are determined by the dielectric medium surrounding each QD. Previous studies have shown that the total polarization correction to the first exciton energy is given by23 Z 1 πe2 ¥ 2l þ 1 δ¼ a Al ½j0 ðπxÞ2 x2l þ 2 dx ð2Þ 2ε1 ε0 a l ¼ 1 0
Because contributions from electronic coupling, excitonic coupling (transition dipoletransition dipole coupling), and polarization effects are expected to increase with decreasing inter-QD distance, the question is, how do we distinguish these effects? To answer the above question, we measured the optical red shift for isolated QDs in a nonresonant matrix of smaller QDs. Figure 4a shows the optical absorption spectra for large QDs (D = 6.2 nm) embedded in a matrix of smaller ones (D = 4.8 nm) at a number ratio of 1:35 before and after the exchange reaction of OA by C2. Note that due to variations in QD growth time, the diameter of the large QDs used here is not the same as that in other figures. SEM imaging (inset to Figure 4a) of the film prepared from the cryo-evaporation procedure (see Supporting Information) clearly shows that the minority of large QDs are completely isolated in the majority of small QD matrix. While the optical red shift of 50.5 ( 0.5 meV is expected upon ligand exchange for the small QD matrix, it is surprising that a large red shift of 21.4 ( 0.3 meV is also seen for the large QDs that are completely isolated from each other. For comparison, the film of the same large QDs prepared in the pure form, Figure 4b, shows an optical red shift of Δ = 24.4 ( 0.3 meV upon the ligand exchange reaction. HRSEM was also used to examine the bimodal QD solid after the C2 treatment to ensure QD isolation (Supporting Information Figure S8). For each large QD embedded in a matrix of smaller ones, its 1Se and 1Sh electronic energy levels are not in resonance with those of the smaller QDs in the matrix. The optical transition dipoles are also not in resonance with those of the matrix, as clearly shown by the large separation between the first exciton transitions of the two different size QDs (Figure 4a). Any off-resonance transition dipoletransition dipole coupling is weak due to the lack of oscillator strength of the smaller QDs at the transition energy of
∑
where j0(z) = sin(z)/z is the spherical Bessel function, a is the QD radius, ε1 is the optical dielectric constant of the QD material, ε0 is the vacuum permittivity, and e is the fundamental charge of the electron. For a QD surrounded by a dielectric shell of thickness b (equal to half of the inter-QD distance, d) of dielectric constant ε2 and further embedded in a medium of dielectric constant ε3, that is, the core/shell/medium (C/S/M) model, the term Al is given by Al
¼
ðl þ 1Þ a2l þ 1
a2l þ 1 ðε2 ε3 Þ½ε1 þ lðε1 þ ε2 Þ þ b2l þ 1 ðε1 ε2 Þ½ε3 þ lðε2 þ ε3 Þ ε2 Þðε2 ε3 Þlðl þ 1Þ þ b2l þ 1 ½ε2 þ lðε1 þ ε2 Þ½ε3 þ lðε2 þ ε3 Þ
a2l þ 1 ðε1
ð3Þ Here, ε2 is taken to be the optical dielectric constant of the capping molecules in the condensed phase; when we set b to 0, this reduces to the simpler C/M model of a QD core embedded in a medium 798
dx.doi.org/10.1021/jz200080d |J. Phys. Chem. Lett. 2011, 2, 795–800
The Journal of Physical Chemistry Letters
LETTER
by the spherical quantum well model. The first is that the actual boundary condition is much smoother than the step-like potential profile used in the model; in a different context, a smooth boundary condition is believed to be essential in explaining carrier dynamics in QDs.27 The second cause may be the perturbation approach in deriving eqs 2 and 3; as an example, the magnitude of the optical red shift reaches almost 10% of the optical band gap for PbSe with hydrazine treatment, casting doubts on the adequacy of the perturbation treatment. The third cause may result from deviation of the QD shape from the spherical model, leading to partial localization of the electron (hole) wave function to one part of the QD, as seen previously in scanning tunneling spectroscopy measurements.28 All of these factors can lead to much larger polarization effects than what are predicted from the spherical quantum well approximation. In PbSe QD solid thin films, the replacement of OA ligands by shorter alkanedithiols molecules leads to increasing red shifts in the exciton peak position as ligand length and inter-QD distance decrease. A dominant contribution (∼88%) to the observed red shift is due to changes in polarization of the dielectric environment surrounding each QD, with only a minor effect (∼12%) attributed to inter-QD electronic and/or exciton (transition dipoletransition dipole) coupling. The observed red shifts are more than 1 order of magnitude larger than theoretical predictions based on the dielectric polarization effect for spherical QDs. The failure of the idealized spherical quantum well model calls for more advanced theories in treating the QD polarization problem.
Figure 5. Red shifts with respect to the OA-capped QDs in the first exciton transition calculated as a function of the diameter of the QD core for C2-capped (dashed) and bare (solid) PbSe QDs in the C/M model.
with the effective dielectric constant, ε3.24 We can calculate ε3 from the effective media approximation (EMA), assuming that the QD film is made of close-packed spheres, each consisting of a QD core (ε1) covered with a capping molecular shell (ε2) of thickness d/2. We can use fcc close packing with a space filling factor of f = 0.74 or random close packing with f = 0.63 in calculating ε3.25 The ligand exchange of C18 capping molecules by the shorter C2 decreases b and increases ε3, leading to a different stabilization energy. The shift in the first exciton transition is then given by Δ = δ0 δ, where δ and δ0 are values before and after ligand exchange, respectively. We calculate Δ values from eqs 2 and 3 using the inter-QD distances obtained from the fit to experiment data (solid line in Figure 2a) and fcc close packing based on the C/S/M and C/M models. The calculated optical red shifts are compared with experimental results in Figure 2b. While both models qualitatively describe the trend in the optical red shift as a function of inter-QD distance or capping molecule length, they fail to account for the magnitude of the optical red shifts. For example, the C/S/M model predicts optical red shifts that are ∼35 smaller than experimental values; for the C/M model, the calculated values are ∼20 smaller. Why Do the Polarization Models Fail so Dramatically in Quantitatively Accounting for the Observed Red Shifts? We believe the answer lies in the sensitivity of the polarization stabilization energy (i.e., the last three terms in eq 1) to the spatial distribution of the electron or hole wave function. For a point charge located from an interface between two phases with different dielectric constants, the polarization energy (chargeimage interaction energy) scales inversely with distance to the interface; for a planar interface, the scaling factor is (4z)1 (z is distance to the interface).26 The sensitivity of the polarization energy is also reflected in the calculated Δ as a function of QD size, Figure 5. For the same capping molecule shell thickness (or bare QDs with no molecular shell), the magnitude of the calculated first exciton red shift increases nearly exponentially with decreasing QD size. As the QD becomes smaller, the spatial distribution of the electron (hole) wave function lies closer to the interface with the dielectric medium, leading to a larger polarization effect. The calculated |Δ| for bare dots at D = 5.4 nm is 5.9 meV, which is 1 order of magnitude smaller than the experimental value obtained for thin films of PbSe QDs (D = 5.4 nm) with hydrazine treatment (Supporting Information Figures S9 and S10). Note that the insensitivity of the red shift to the QD film thickness (0.52.5 ML) and substrate material (SiO2 and TiO2) suggests little influence of the substrate polarizability on the observed red shifts. There are a number of possible causes for |ψ|2 of the electron (hole) to lie much closer to the QD surface than what is predicted
’ ASSOCIATED CONTENT
bS
Supporting Information. Materials and experimental procedures are included along with TEM images, solution-phase optical absorption spectra, HRSEM, ATR-FTIR spectroscopy, and GISAXS data. This material is available free of charge via the Internet http://pubs.acs.org.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected]. Present Addresses †
UFR de Physique, Universite Joseph Fourier, 38140 Grenoble CEDEX 9, France. § Center for Functional Nanomaterials, Brookhaven National Laboratory, Upton, NY 11973.
’ ACKNOWLEDGMENT This work was supported by the Department of Energy under Grant No. DE-FG02-07ER46468 and by the Welch Foundation. A.W. acknowledges an ACC fellowship supported by the National Science Foundation (CHE# 0937032). C.A.N. acknowledges a graduate fellowship supported by the Department of Energy. Research carried out at the Center for Functional Nanomaterials and the National Synchrotron Light Source, Brookhaven National Laboratory, was supported by the U.S. Department of Energy, Office of Basic Energy Sciences, under Contract No. DE-AC02-98CH10886. We thank Prof. Peter Rossky for helpful discussions. 799
dx.doi.org/10.1021/jz200080d |J. Phys. Chem. Lett. 2011, 2, 795–800
The Journal of Physical Chemistry Letters
LETTER
’ REFERENCES (1) Talapin, D. V.; Murray, C. B. PbSe Nanocrystal Solids for n- and p-Channel Thin Film Field-Effect Transistors. Science 2005, 310, 86–89. (2) Law, M.; et al. Structural, Optical, And Electrical Properties of PbSe Nanocrystal Solids Treated Thermally or with Simple Amines. J. Am. Chem. Soc. 2008, 130, 5974–5985. (3) Luther, J. M.; et al. Structural, Optical and Electrical Properties of Self-Assembled Films of PbSe Nanocrystals Treated with 1,2-Ethanedithiol. ACS Nano 2008, 2, 271–280. (4) Williams, K. J.; et al. Strong Electronic Coupling in TwoDimensional Assemblies of Colloidal PbSe Quantum Dots. ACS Nano 2009, 3, 1532–1538. (5) Liu, Y.; et al. Dependence of Carrier Mobility on Nanocrystal Size and Ligand Length in PbSe Nanocrystal Solids. Nano Lett. 2010, 10, 1960–1969. (6) Choi, J. J.; et al. Photogenerated Exciton Dissociation in Highly Coupled Lead Salt Nanocrystal Assemblies. Nano Lett. 2010, 10, 1805–1811. (7) Brus, L. E. Electron Electron and ElectronHole Interactions in Small Semiconductor Crsyallites — The Size Dependence of the Lowest Excited Electronic State. J. Chem. Phys. 1984, 80, 4403–4409. (8) Franceschetti, A.; Fu, H.; Wang, L. W; Zunger, A. Many-Body Pseudopotential Theory of Excitons in InP and CdSe Quantum Dots. Phys. Rev. B 1999, 60, 1819–1829. (9) Mirafzal, H.; Kelley, D. F. Singlet/Triplet Reversal in StronglyCoupled GaSe Nanoparticle Aggregates. J. Phys. Chem. C 2009, 113, 7139–7146. (10) Jelly, E. E. Molecular, Nematic and Crystal States of I: I-Diethyl--Cyanine Chloride. Nature 1937, 139, 631. (11) Kobayashi, T., Ed.; J-aggregates; World Scientific: Singapore, 1996. (12) Vossmeyer, T.; et al. CdS Nanoclusters — Synthesis, Characterization, Size-Dependent Oscillator Strength, Temperature Shift of The Excitonic-Transition Energy, and Reversible Absorbency Shift. J. Phys. Chem. 1994, 98, 7665–7673. (13) Leatherdale, C. A.; Bawendi, M. G. Observation of Solvatochromism in CdSe Colloidal Quantum Dots. Phys. Rev. B 2001, 63, 165315. (14) Kim, B. S.; Islam, M. A.; Brus, L. E.; Herman, I. P. Interdot Interactions and Band Gap Changes in CdSe Nanocrystal Arrays at Elevated Pressure. J. Appl. Phys. 2001, 89, 8127–8140. (15) Dollefeld, H.; Weller, H.; Eychmuller, A. Particleparticle Interactions in Semiconductor Nanocrystal Assemblies. Nano Lett. 2001, 1, 267–269. (16) Dollefeld, H.; Weller, H.; Eychmuller, A. Semiconductor Nanocrystal Assemblies: Experimental Pitfalls and a Simple Model of ParticleParticle Interaction. J. Phys. Chem. B 2002, 106, 5604–5608. (17) Kagan, C. R.; Murray, C. B.; Bawendi, M. G. Long-Range Resonance Transfer of Electronic Excitations in Close-Packed CdSe Quantum-Dot Solids. Phys. Rev. B 1996, 54, 8633–8643. (18) Kagan, C. R.; Murray, C. B.; Nirmal, M.; Bawendi, M. G. Electronic Energy Transfer in CdSe Quantum Dot Solids. Phys. Rev. Lett. 1996, 76, 1517–1520. (19) Frederick, M. T.; Weiss, E. A. Relaxation of Exciton Confinement in CdSe Quantum Dots by Modification with a Conjugated Dithiocarbamate Ligand. ACS Nano 2010, 4, 3195–3200. (20) Wise, F. W. Lead Salt Quantum Dots: The Limit of Strong Quantum Confinement. Acc. Chem. Res. 2000, 33, 773–780. (21) Leschkies, K. S.; Beatty, T. J.; Kang, M. S.; Norris, D. J.; Aydil, E. S. Solar Cells Based on Junctions between Colloidal PbSe Nanocrystals and Thin ZnO Films. ACS Nano 2009, 3, 3638–3648. (22) Babic, D.; Tsu, R.; Greene, R. F. Ground-State Energies of OneElectron and Two-Electron Silicon Dots in an Amorphous-Silicon Dioxide Matrix. Phys. Rev. B 1992, 45, 14150–14155. (23) Allan, G.; Delerue, C.; Lannoo, M.; Martin, E. Hydrogenic Impurity Levels, Dielectric-Constant, and Coulomb Charging Effects in Silicon Crystallites. Phys. Rev. B 1995, 52, 11982–11988. (24) Iwamatsu, M.; Fujiwara, M.; Happo, N.; Horii, K. Effects of Dielectric Discontinuity on the Ground-State Energy of Charged Si
Dots Covered with a SiO2 Layer. J. Phys.: Condens. Matter 1997, 9, 9881–9892. (25) Spanier, J. E.; Herman, I. P. Use of Hybrid Phenomenological and Statistical Effective-Medium Theories of Dielectric Functions to Model the Infrared Reflectance of Porous SiC Films. Phys. Rev. B 2000, 61, 10437–10450. (26) Echenique, P. M.; Pendry, J. B. Existence and Detection of Rydberg States at Surfaces. J. Phys. C: Solid State Phys. 1978, 11, 2065–2075. (27) Cragg, G. E.; Efros, A. L. Suppression of Auger Processes in Confined Structures. Nano Lett. 2010, 10, 313–317. (28) Millo, O.; Katz, D.; Cao, Y. W.; Banin, U. Imaging and Spectroscopy of Artificial-Atom States in Core/Shell Nanocrystal Quantum Dots. Phys. Rev. Lett. 2001, 86, 5751–5754.
’ NOTE ADDED AFTER ASAP PUBLICATION This paper was published ASAP on March 16, 2011. Various changes were made to the Supporting Information. The revised paper was reposted on March 21, 2011.
800
dx.doi.org/10.1021/jz200080d |J. Phys. Chem. Lett. 2011, 2, 795–800
