Shell Nanocrystals with “Permanent

LETTER pubs.acs.org/NanoLett

Copper-Doped Inverted Core/Shell Nanocrystals with “Permanent” Optically Active Holes Ranjani Viswanatha,†,|| Sergio Brovelli,†,|| Anshu Pandey,†,§ Scott A. Crooker,‡ and Victor I. Klimov*,†,§ †

Chemistry Division, ‡National High Magnetic Field Laboratory, §Center for Advanced Solar Photophysics, Los Alamos National Laboratory, Los Alamos, New Mexico 87545, United States

bS Supporting Information ABSTRACT: We have developed a new class of colloidal nanocrystals composed of Cudoped ZnSe cores overcoated with CdSe shells. Via spectroscopic and magneto-optical studies, we conclusively demonstrate that Cu impurities represent paramagnetic +2 species and serve as a source of permanent optically active holes. This implies that the Fermi level is located below the Cu2+/Cu1+ state, that is, in the lower half of the forbidden gap, which is a signature of a p-doped material. It further suggests that the activation of optical emission due to the Cu level requires injection of only an electron without a need for a valence-band hole. This peculiar electron-only emission mechanism is confirmed by experiments in which the titration of the nanocrystals with hole-withdrawing molecules leads to enhancement of Cu-related photoluminescence while simultaneously suppressing the intrinsic, band-edge exciton emission. In addition to containing permanent optically active holes, these newly developed materials show unprecedented emission tunability from nearinfrared (1.2 eV) to the blue (3.1 eV) and reduced losses from reabsorption due to a large Stokes shift (up to 0.7 eV). These properties make them very attractive for applications in light-emission and lasing technologies and especially for the realization of novel device concepts such as “zero-threshold” optical gain. KEYWORDS: Nanocrystal quantum dot, copper doped nanocrystal, ZnSe/CdSe core/shell, p-type, copper oxidation state

C

olloidal semiconductor nanocrystals (NCs) are an emerging class of tunable materials with a large promise for applications in low-cost, solution-processable devices. At present, however, their applicability is greatly limited by the difficulty of controlled introduction of permanent electrically and/or optically active charges. The resolution of this problem would greatly benefit applications of NCs in areas such as photovoltaics1,2 and light-emitting diodes (LEDs)3 and could also enable new advanced device concepts such as “zero-threshold” optical gain.4,5 The concept of NC doping with optically and electrically active impurities is illustrated in Figure 1. In the example in panel a, the deep acceptor impurity creates an optically active intragap hole state, which can capture an electron from the conduction band via a radiative transition and thus produce optical emission without injection of a hole. An interesting feature of this transition is that it has an “emission-only”, one-directional character that does not produce competing optical absorption. This property can be utilized to realize a highly efficient four-level optical gain scheme (Figure 1a, right). Such a scheme would be especially beneficial for NC materials, as it would allow for optical amplification in the regime of a vanishingly small concentration of excitons (hence, the concept of zero-threshold gain) before the onset of multiexciton Auger recombination.4 This would help to overcome a major problem in NC lasing associated with ultrafast optical gain decay via the Auger process.4,6 In Figure 1b, we show a different example of doping when a shallow acceptor impurity state is located in close proximity to the valence band and thus can thermally inject holes into the r 2011 American Chemical Society

extended states of the host semiconductor. In the case of colloidal nanostructures, such doping with electrically active charges would facilitate charge transport (Figure 1b, right) and could help to resolve the problem of extremely low intrinsic conductivity of NC assemblies. It will also benefit practical applications of NCs in electronic and optoelectronic devices by allowing for a facile engineering of built-in electric fields via controlled placement of p
n junctions.7 So far, most successful NC doping efforts have focused on magnetically active Mn ions.8
13 Temporary introduction of active charge carriers using strongly reducing chemical agents has also been demonstrated previously.14 Permanent incorporation of charges into NCs, however, would require placement of active impurities into the lattice. Doping of NCs with transition-metal ions that exhibit variable valence is a promising strategy to achieve this goal, and an increasing number of efforts have been devoted to incorporation of ions such as Cu, Ag, and Au into various NCs.10,13,15,16 Intragap luminescence centers associated with Cu dopants in bulk II
VI semiconductors have been known since the early 1960s.17
19 In materials such as ZnSe and ZnS, these earlier studies have demonstrated that Cu replaces the cations and that transitions to and from the Cu 3d electron shell correspond to energies that lie within the forbidden gap of the semiconductor Received: July 27, 2011 Revised: September 22, 2011 Published: September 23, 2011 4753

dx.doi.org/10.1021/nl202572c | Nano Lett. 2011, 11, 4753–4758

Nano Letters

Figure 1. Concepts of NC doping with optically and electrically active charges. (a) A pre-existing “permanent” hole associated with a deep impurity acceptor state (left), which is not thermally accessible by valence-band electrons, is optically active and can participate in radiative transitions involving conduction-band electrons. Since this transition is “unidirectional” (i.e., does not have associated absorption), it can be used to realize the concept of zero-threshold optical gain and a highly efficient 4-level scheme of NC-lasing (right). We note that the photophysical cycle still requires an external hole to ensure charge balance after optical emission. Importantly, such charge-conserving hole can be located in a localized surface state or defect site and not necessarily in the valence band of the semiconductor. The process of charge compensation is indicated by a blue arrow. (b) A hole residing in a shallow acceptor state (left) is electrically active as it can be thermally excited into the valence band and can then participate in charge transport through delocalized quantized states of the NCs (right). (c) The difference in optical behaviors of Cu1+ and Cu2+ impurities. The Fermi level is indicated by a green line and labeled “FL”. The Cu1+ state corresponds to the completely filled 3d10 shell (the Fermi level is above it) and is optically passive (left). It can participate in radiative transitions only after capturing a hole from the valence band (a sequence of events leading to emission of a photon is indicated by numbers 1 and 2); in our diagrams we show only the t-level of the Cu impurity. On the other hand, the Cu2+ state corresponds to the 3d9 configuration (right) and is optically active as it contains a hole which can radiatively capture a conduction band electron (step 1); the electron is then transferred nonradiatively to the valence band (step 2). Importantly, since Cu1+ dopants have a completely filled 3d electron shell, they represent nonmagnetic spin-zero species (s = 0), while Cu2+ dopants have one unpaired electron and, therefore, are paramagnetic spin-1/2 species.

host. Within a simple one-electron picture, the 3d electrons of the Cu dopant are split by the tetrahedral crystal field of the host lattice, resulting in two distinct levels within the gap. These levels, typically labeled as t (higher in energy, 6-fold degenerate) and e (lower in energy, 4-fold degenerate),19 reside near the valence band edge. They are predominantly 3d-like in character, but are also thought to contain some admixture of 3p and 4s states from the host lattice. These levels are optically and electronically active and can exchange electrons with conduction and valence bands

LETTER

via radiative and nonradiative transitions. When the Fermi level is above the t-state, both levels are fully occupied. This corresponds to the +1 oxidation state of Cu with the completely filled 3d shell (electronic configuration (Ar)3d10). When the Fermi level is below the t-state, one electron is removed and the t level can be considered as having a hole, which corresponds to electronic configuration (Ar)3d9 and the +2 oxidation state of Cu. The sum of the onsets of optical absorption due to transitions from the valence band to the t-level (unoccupied in the Cu2+ state) and from the t-level (occupied in the Cu1+ state) to the conduction band is equal to the band gap energy Eg (ref 19). Therefore, in the single-electron picture, the Cu1+ and Cu2+ states can be shown by the same level as we do in Figure 1c. In bulk ZnSe, the t- and e-levels are ca. 0.72 and 0.35 eV above the valence band edge, respectively. The Cu1+ state is optically passive (i.e., it does not contain an optically active hole) but can be activated through capture of a hole from the valence band, which then can participate in radiative transitions involving conduction-band electrons (Figure 1c, left). On the other hand, a Cu2+ state is optically active and can be treated as a state with a pre-existing hole. Specifically, following photoexcitation, the radiative transition of a conduction band electron to the Cu2+ level can generate PL without participation of the valence-band hole (Figure 1c, right). Recently, several publications reported the development of a characteristic intragap Cu-related emission feature in copper-doped II
VI NCs.15,16,20
23 Several studies have also attempted to characterize the oxidation state of Cu upon incorporation into NCs.20,23,24 The results of these investigations, however, have been controversial as both +1 (refs 21, 25, and 26) and +2 (refs 22 and 27) oxidation states have been ascribed to copper dopants. Distinguishing between these two possibilities is important for understanding the electronic, optical, and magnetic properties of Cu impurities. Specifically, the +1 state has a filled 3d10 shell and is nonmagnetic and cannot contribute to emission without first capturing an external hole. On the other hand, the +2 state has an unpaired electron in its 3d9 shell and is therefore paramagnetic and can be considered as a state with a permanent optically active hole, which can participate in emission without the injection of a hole. Furthermore, if the Cu ions are preferentially in the +2 state, this implies that the Fermi level is below the Cu impurity t-level and hence is located near the valenceband edge, which is formally a signature of a p-type material. In this work, we address the nature of the state associated with Cu impurities in II
VI NCs by conducting comprehensive spectroscopic and magneto-optical studies of novel NC structures composed of Cu-doped ZnSe cores overcoated with CdSe shells. The use of core/shell architectures serves two critical purposes: (1) it helps to retain the dopants in the core during the chemical synthesis and thus overcome the problem of “self purification”;10 (2) further, it allows for the extension of the range of spectral tunability into the near-infrared region through control of the spatial distribution of the electronic wave functions between the core and the shell regions. While the core/shell motif has been widely applied to control optical properties of undoped NCs,4,28
30 this work provides the first example of the use of such a strategy for tuning the impurity-related emission in doped NC materials. Using these structures, we conclusively demonstrate that Cu impurities represent paramagnetic +2 species, and hence, can serve as a source of optically active permanent holes. This peculiar electron-only emission mechanism is confirmed by experiments in which the titration of the NCs with hole-withdrawing molecules leads to enhancement 4754

dx.doi.org/10.1021/nl202572c |Nano Lett. 2011, 11, 4753–4758

Nano Letters

Figure 2. Synthesis and optical properties of Cu:ZnSe/CdSe NCs. (a) Schematic diagram illustrating various steps of the synthesis of core/ shell Cu:ZnSe/CdSe NCs (see Supporting Information for details and notations). (b) A band alignment diagram of core/shell NCs with a thin (left) and a thick (right) CdSe shell. The intragap state introduced by the Cu impurity is shown by a dashed line; its oxidation state (Cu1+ vs Cu2+) is determined by the position of the Fermi level with regard to the impurity level. Schematic representations of the radial probability distributions for electrons and holes are shown by black lines. The band-edge (BE) and Curelated (D) radiative transitions are shown by the green and the red arrow, respectively. (c) Optical absorption (dashed lines) and PL (solid lines; excited at 3.1 eV) spectra of undoped (blue) and Cu-doped (red) ZnSe/ CdSe NCs with a nominal shell thickness H = 0.35 nm. (d) PL decay curves of the undoped (blue; band-edge peak) and Cu-doped (red; Cu-related peak) NCs (same samples as in c). We note that, similarly to the undoped NCs, the lifetime of the residual band-edge PL from the doped samples is of the order of a few ns, which is significantly faster than that of the Cu-related emission. Inset: optical absorption (black) and PLE (red; collected at 1.6 eV) spectra of the Cu-doped NCs.

of Cu-related PL while simultaneously suppressing the intrinsic, band-edge exciton emission. A typical synthetic scheme is shown in Figure 2a. Cu-doped ZnSe cores were synthesized using a three-step procedure, which includes the growth of small ZnSe NCs, followed by incorporation of copper ions into the surface layer and subsequent overcoating with ZnSe (ref 15). We complete the structure with a CdSe shell of variable thickness, which results in a so-called “inverted” core/shell NC, where a wider-band gap semiconductor is encapsulated within a lower band-gap material.28 Elemental analysis using inductively coupled plasma atomic emission spectroscopy indicates that these NCs contain ∼1% of copper ions with respect to the other cations in the ZnSe core. Further, the amount of copper is preserved throughout the process of CdSe shell growth. The shell thickness is consistent with the ratio of Zn to Cd precursors used in the reaction, suggesting efficient epitaxial growth. More details on the synthesis and spectroscopic characterization of the Cu-doped and the reference undoped structures are provided in the Supporting Information.

LETTER

Figure 3. Spectral tunability achievable with Cu-doped core/shell NCs. (a) PL spectra of undoped (top) and Cu-doped (bottom) ZnSe/CdSe core/shell NCs with progressively increasing shell thickness, H, starting from core-only ZnSe NCs (R0 = 1.5 nm, purple curve) to ZnSe/CdSe NCs with R0 + H = 2.8 nm (dark red curve). Inset: the chromaticity diagram coordinates (x = 0.36; y = 0.36) of the PL spectrum of Cu:ZnSe/CdSe NCs (R0 + H = 1.6 nm), which correspond to a mildly “warm” white light (color temperature ∼4400 K), indicate that these nanostructures can be applied in white-light LEDs and solid state lighting. (b) Spectral positions of the 1S absorption feature (red circles) and the Cu-related PL band (green triangles) as a function of particle size (R0 + H). All NCs have the same core radius R0 = 1.5 nm. All measurements were carried out at room temperature. The inset shows schematically the positions of the conduction (CB) and valence (VB) band edges with respect to the Cu impurity level as a function of NC size. The change in the energy of the Cu-related PL (green arrows) is solely due to the shift of the conduction band-edge state, while the change in the absorption onset (i.e., Eg, red arrows) is contributed by the shifts of both the conduction- and valence-band levels. Black arrows show the Stokes shift.

Figure 2b shows the band alignment diagrams of the ZnSe/ CdSe NCs for the case of a thin (Figure 2b, left) and a thick (Figure 2b, right) CdSe shell. Upon growing a thicker CdSe shell, both electron and hole wave functions experience progressive delocalization into the shell region,28 which allows for wide spectral tunability of the intrinsic band-edge PL (Figure 3a; upper spectra). As discussed above, copper ions introduce localized intragap states in II
VI semiconductors (the t-state is shown by the dashed line in Figure 2b) that serve as radiative decay centers for conduction band electrons.17,18,31,32 Following the introduction of copper into our NCs, we indeed detect a new intragap PL band, which is observed simultaneously with residual band-edge PL (Figure 2c). This copper-related PL exhibits a large Stokes shift of ∼0.65 eV relative to the bandedge absorption feature (Figure 2c). This value is close to the separation of the Cu impurity t-level from the valence-band edge in bulk ZnSe,19 which confirms that the observed emission is indeed due to recombination of the conduction band electron with the hole in the Cu state (due to a large hole effective mass in II
VI materials, a confinement-induced shift of the valence band edge is small). The Cu PL exhibits a long lifetime of ca. 390 ns versus ∼10 ns for the band-edge feature (see Figure 2d), which is consistent with the localized nature of the impurity state having a reduced spatial overlap with the extended conduction-band state. 4755

dx.doi.org/10.1021/nl202572c |Nano Lett. 2011, 11, 4753–4758

Nano Letters The PL excitation (PLE) spectrum of Cu-emission is essentially identical to the NC absorption spectrum (inset of Figure 2d), indicating that it is excited via the host semiconductor. The fact that we observe simultaneously both the band-edge and Curelated PL bands characterized by two different lifetimes does not necessarily imply that some of the NCs in the ensemble are undoped. As we discuss later in the paper, this double-band PL spectrum can develop due to the existence of two NC subensembles with distinct relaxation channels for photoexcited holes (surface trapping vs radiative recombination with a conduction-band electron) even if all of the NCs contain Cu ions. Because of its mixed impurity-host character, the Cu emission can be tuned spectrally by shifting the conduction band state of the NCs. In our “inverted” structures, this can be accomplished by changing the dimensions of both the core and/or the shell.28 As illustrated in Figure 3a, just by changing the shell thickness (H), from core-only ZnSe NC (R0 = 1.5 nm) to core/shell structures with H = 1.3 nm, we can achieve a tunability of Cu related emission from ∼1.3 to ∼2.7 eV. The corresponding tunability of the band-edge absorption feature, which represents an approximate measure of Eg, is from ∼2.0 to ∼3.3 eV (Figure 3b). As illustrated in the inset of Figure 3b, the change in the energy of the Cu-related PL band is solely due to the shift of the conduction band-edge state, while the change in the absorption onset (i.e., Eg) is in principle contributed by the shifts of both the conduction- and valence-band levels. The fact, however, that the emission Stokes shift is almost size independent (0.6
0.7 eV) indicates that the relative contribution of hole confinement to the band gap energy is fairly small, as expected based on the relationship between the electron and hole effective masses. The PL quantum yield progressively increases from ∼0.5 to ∼7% with increasing CdSe shell thickness up to H = 1.3 nm; this suggests a progressive improvement in electronic passivation of interfacial defects in thicker-shell NCs. By lowering the temperature, the emission quantum yield can be increased further to ∼40%. The combination of wide spectral tunability with a large emission Stokes shift, which reduces losses due to reabsorption, makes these novel NC materials useful candidates for applications in light-emission and lasing technologies including white light LEDs (inset of Figure 3a) and novel four-level-based lasing schemes (Figure 1a, right). To explore the full potential of these new tunable materials, it is important to understand the nature of the Cu emission, and specifically, the oxidation state of the copper ion in the absence of optical excitation. There are a number of distinctions between Cu1+ and Cu2+ impurities that can be detected spectroscopically. Cu1+ centers have a filled 3d10 electronic configuration and are therefore expected to be nonmagnetic. In marked contrast, Cu2+ centers possess a single unpaired spin-1/2 electron in their 3d9 shell and are therefore expected to behave paramagnetically. The absence or presence of enhanced magneto-optical effects in these NCs can therefore identify whether copper ions incorporate as Cu1+ or Cu2+, respectively. To this end, we measure the circularly polarized magnetoabsorption of these NCs in high magnetic fields. Owing to the optical selection rules in these II
VI semiconductors, right- and left-circularly polarized light couples preferentially to the spin-up and spin-down states of the nanocrystals 0 1S exciton, which become Zeeman-split in an applied magnetic field (B). In agreement with prior studies of nonmagnetic NCs,8,12,33 our undoped ZnSe/CdSe NCs exhibit a small Zeeman splitting of the 1S exciton (∼0.1 meV/T, or exciton g-factors of order 2).

LETTER

Figure 4. Circularly polarized magneto-absorption of Cu-doped ZnSe/ CdSe nanocrystals. (a) The absorption spectrum at zero applied magnetic field. (b) Absorption spectra of right- and left-circularly polarized light (RCP and LCP, respectively) at B = 6T. A clear enhancement of the 1S band-edge Zeeman splitting is observed, suggesting the influence of magnetically active (i.e., Cu2+) dopants.

In contrast, copper-doped (but otherwise identical) NCs clearly reveal a markedly enhanced Zeeman splitting (∼2.5 meV at B = 6 T; see Figure 4), commensurate with an effective exciton g-factor of order 7. These results strongly suggest that the Cu dopants are incorporated as magnetically active Cu2+ ions, in agreement with recent studies of EPR-active Cu2+ dopants in ZnSe nanocrystals.22 This further implies that the Fermi level in these NCs is located below the Cu impurity t-state and, hence, is within ∼0.7 eV from the valence-band edge (see Figure 1c; right panel). Since Eg in our NCs is from ∼2.0 to ∼3.3 eV, this corresponds to the lower half of the forbidden gap, which is formally a signature of p-doping. We would like to point out, however, that because of the large depth of the Cu acceptor state (0.6
0.7 eV) the fraction of holes thermally activated into the valence band is extremely small and is not expected to affect electrical conductance except, perhaps, the case of very high doping levels when the Cu states merge into a conductive impurity band. As was mentioned earlier, an interesting feature expected for emission due to the Cu2+ centers is that it can be activated by injecting conduction-band electrons without the need for valence-band holes. In fact, the presence of a band-edge hole in the NC can diminish the Cu-related PL as this hole will compete for the same conduction-band electron with a permanent hole associated with the Cu2+ impurity (Figure 5). Specifically, if the photogenerated (i.e., transient) hole is present in the NC for sufficiently long time, the emission is dominated by band-toband recombination (Figure 5a, left) as its rate is significantly faster than that of the Cu-related transition (see PL dynamics in Figure 2d). On the other hand, if the photogenerated hole is quickly removed from the NCs (due, e.g., to trapping at a surface defect), then the PL should be dominated by the transition to the Cu state (Figure 5a, right). Given this recombination scenario, the band-edge emission can be observed even if all of the NCs are doped with Cu ions. In this case, the relative intensities of the Curelated and intrinsic excitonic PL features are determined by the fractions of the NCs with and without hole trapping sites, respectively. 4756

dx.doi.org/10.1021/nl202572c |Nano Lett. 2011, 11, 4753–4758

Nano Letters

Considering the relatively low room-temperature PL quantum yield of the undoped NCs (e5%), the fraction of dots with trap sites in the ensemble is significant. These considerations further suggest that one could attempt to shift the balance in favor of the Cu-related PL channel by intentionally introducing hole traps (Figure 5a, right). To test this hypothesis, we treat NCs with dodecane thiols (DDT) that are known to act as hole-withdrawing species when attached to CdSe surfaces. In the case of undoped NCs, the treatment with thiols leads to rapid quenching of emission.34,35 The situation, however, is different with Cu-doped NCs. In Figure 5b, we show the PL spectra of Cu:ZnSe/CdSe NCs titrated with increasing amounts of DDT. The untreated sample shows a band-edge feature at ∼2.3 eV and a copperrelated band at ∼1.85 eV. The apparent color of the sample’s fluorescence is green (inset of Figure 5b; left vial) due to the strong contribution from the high-energy band-edge PL. Treatment of the samples with DDT leads to a gradual quenching of band-edge PL, accompanied by a complementary growth of the Cu-related PL (Figure 5c). We also observe a dramatic change in color, which becomes orange due to increasing contribution from Cu PL (inset of Figure 5b; right vial). These observations are consistent with the model of the Cu2+ emission center (Figure 5a) where the permanent and photogenerated (transient) holes compete for the same conduction-band electron; when the transient hole is quickly removed from the NC, the

emission occurs solely through the Cu centers. In the case of the Cu1+ impurities, the Cu-related band would be suppressed together with the band-edge emission. Therefore, the results of the thiol-titration experiments provide very strong evidence for the Cu2+ nature of the dopants in our structures. Interestingly the treatment with thiols also leads to an overall increase of the PL quantum yield (from ∼3.5 to ∼6% in the example in Figure 5c; red triangles). This can be rationalized in the following way. Since the Cu emission involves a permanent hole, its intensity is mainly limited by electron surface trapping, which is usually a slower process than trapping of band-edge transient holes.36 The latter, however, is the main carrier-loss mechanism in the case of band-edge PL. Therefore, by “forcing” the emission through the Cu channel, we eliminate losses due to hole surface-trapping and thus increase the overall efficiency of PL. These surface treatment experiments demonstrate the crucial importance of understanding of the mechanism underlying the optical activity of impurities for controlling the properties of doped materials. In the above example, such understanding helped us to develop an unusual and somewhat counterintuitive strategy for controlling both the emission color and its efficiency via intentional introduction of surface hole traps. To summarize, we have demonstrated a new class of tunable nanostructures consisting of Cu-doped ZnSe cores overcoated with CdSe shells. These novel NCs feature broad tunability of emission wavelength from near-IR (1.2 eV) to the blue (3.1 eV) and a large emission Stokes shift up to 0.7 eV, which greatly reduces losses due to photon reabsorption. Importantly, comprehensive spectroscopic and magneto-optical studies demonstrate unambiguously that copper is incorporated into these structures as a magnetically active +2 species, and hence, represents a source of optically active permanent holes. This assessment is confirmed by experiments in which NC treatment with hole-withdrawing thiols enhances Curelated emission at the expense of intrinsic exciton emission. The +2 nature of the Cu impurity indicates that the Fermi level in these NCs is located in the lower half of the forbidden gap (within less than ∼0.7 eV from the valence band), which is a signature of p-doping. The effect of doping is achieved by introducing permanent substitutional impurities into the NC core rather than surface modification, most often used in previous NC studies. These novel materials are promising candidates for applications in various light-emitting technologies and advanced device concepts that exploit permanent optically active dopants, such as zero-threshold optical gain.

’ ASSOCIATED CONTENT

bS

Supporting Information. Synthesis and material characterization, spectroscopic studies, magneto-absorption measurements, and thiol titration experiments. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]; Phone: 505-665-8284. Author Contributions

)

Figure 5. Interplay between band-to-band and Cu-related recombination channels during titration of Cu-doped NCs with hole-withdrawing molecules. (a) If a photogenerated hole is present in the NC for sufficiently long time, radiative recombination is primarily due to a band-to-band transition as it is much faster than the band-to-Cu transition; this results in the PL spectrum dominated by band-edge emission (left). If a photogenerated hole is instead quickly removed from the NC (using, e.g., hole-withdrawing species) emission becomes dominated by the Cu-related band (right). (b) Titration of Cu:ZnSe/ CdSe NCs (R0 + H = 1.6 nm) with an increasing amount of hole accepting molecules of DDT (dodecane thiol; varied from 0
16 μL) leads to a progressive increase of the Cu-related band accompanied by the suppression of band-edge emission; this results in a pronounced change of the emission color (from green to orange) as shown in the inset. (c) During NC treatment with increasing amounts of DDT the total PL quantum yield increases (triangles), while the normalized spectrally integrated intensities of the band-edge (solid circles) and Cu-emission (open circles) vary in a complementary way.

LETTER

These authors contributed equally to this work.

’ ACKNOWLEDGMENT R.V. and S.A.C. acknowledge support by the Chemical Sciences, Biosciences, and Geosciences Division of the Office of Basic Energy Sciences (BES), Office of Science, U.S. Department of 4757

dx.doi.org/10.1021/nl202572c |Nano Lett. 2011, 11, 4753–4758

Nano Letters Energy (DOE). V.I.K. is supported by the Center for Advanced Solar Photophysics, an Energy Frontier Research Center funded by the Office of BES, Office of Science, U.S. DOE. S.B. and A.P. are supported by the Los Alamos National Laboratory Directed Research and Development Program.

’ REFERENCES (1) Greenham, N. C.; Peng, X. G.; Alivisatos, A. P. Phys. Rev. B 1996, 54 (24), 17628–17637. (2) Luther, J. M.; Law, M.; Beard, M. C.; Song, Q.; Reese, M. O.; Ellingson, R. J.; Nozik, A. J. Nano Lett. 2008, 8 (10), 3488–3492. (3) Kim, T.-H.; Cho, K.-S.; Lee, E. K.; Lee, S. J.; Chae, J.; Kim, J. W.; Kim, D. H.; Kwon, J.-Y.; Amaratunga, G.; Lee, S. Y.; Choi, B. L.; Kuk, Y.; Kim, J. M.; Kim, K. Nat. Photonics 2011, 5 (3), 176–182. (4) Klimov, V. I.; Ivanov, S. A.; Nanda, J.; Achermann, M.; Bezel, I.; McGuire, J. A.; Piryatinski, A. Nature 2007, 447 (7143), 441–446. (5) Wang, C. J.; Wehrenberg, B. L.; Woo, C. Y.; Guyot-Sionnest, P. J. Phys. Chem. B 2004, 108 (26), 9027–9031. (6) Klimov, V. I.; Mikhailovsky, A. A.; Xu, S.; Malko, A.; Hollingsworth, J. A.; Leatherdale, C. A.; Eisler, H. J.; Bawendi, M. G. Science 2000, 290 (5490), 314–317. (7) Paleari, A.; Brovelli, S.; Lorenzi, R.; Giussani, M.; Lauria, A.; Mochenova, N.; Chiodini, N. Adv. Funct. Mater. 2010, 20 (20), 3511–3518. (8) Bussian, D. A.; Crooker, S. A.; Yin, M.; Brynda, M.; Efros, Al. L.; Klimov, V. I. Nat. Mater. 2008, 8 (1), 35–40. (9) Erwin, S. C.; Zu, L. J.; Haftel, M. I.; Efros, Al. L.; Kennedy, T. A.; Norris, D. J. Nature 2005, 436 (7047), 91–94. (10) Norris, D. J.; Efros, Al. L.; Erwin, S. C. Science 2008, 319 (5871), 1776–1779. (11) Ochsenbein, S. T.; Feng, Y.; Whitaker, K. M.; Badaeva, E.; Liu, W. K.; Li, X.; Gamelin, D. R. Nat. Nanotechnol. 2009, 4 (10), 681–687. (12) Norris, D. J.; Yao, N.; Charnock, F. T.; Kennedy, T. A. Nano Lett. 2001, 1 (1), 3–7. (13) Bryan, J. D.; Gamelin, D. R. Prog. Inorg. Chem. 2005, 54, 47–126. (14) Shim, M.; Guyot-Sionnest, P. Nature 2000, 407 (6807), 981–983. (15) Chen, D.; Viswanatha, R.; Ong, G. L.; Xie, R.; Balasubramaninan, M.; Peng, X. J. Am. Chem. Soc. 2009, 131 (26), 9333–9339. (16) Mocatta, D.; Cohen, G.; Schattner, J.; Millo, O.; Rabani, E.; Banin, U. Science 2011, 332 (6025), 77–81. (17) Broser, I.; Schulz, H. J. J. Electrochem. Soc. 1961, 108 (6), 545–548. (18) Peka, P.; Schulz, H. J. Physica B 1994, 193 (1), 57–65. (19) Stringfellow, G. B.; Bube, R. H. Phys. Rev. 1968, 171 (3), 903–915. (20) Pradhan, N.; Goorskey, D.; Thessing, J.; Peng, X. J. Am. Chem. Soc. 2005, 127 (50), 17586–17587. (21) Srivastava, B. B.; Jana, S.; Pradhan, N. J. Am. Chem. Soc. 2011, 133 (4), 1007–1015. (22) Jana, S.; Srivastava, B. B.; Acharya, S.; Santra, P. K.; Jana, N. R.; Sarma, D. D.; Pradhan, N. Chem. Commun. 2010, 46 (16), 2853–2855. (23) Xie, R.; Peng, X. J. Am. Chem. Soc. 2009, 131 (30), 10645– 10651. (24) Stouwdam, J. W.; Janssen, R. A. J. Adv. Mater. 2009, 21 (28), 2916–2920. (25) Corrado, C.; Jiang, Y.; Oba, F.; Kozina, M.; Bridges, F.; Zhang, J. Z. J. Phys. Chem. A 2009, 113 (16), 3830–3839. (26) Meulenberg, R. W.; van Buuren, T.; Hanif, K. M.; Willey, T. M.; Strouse, G. F.; Terminello, L. J. Nano Lett. 2004, 4 (11), 2277–2285. (27) Suyver, J. F.; Beek, T. v. d.; Wuister, S. F.; Kelly, J. J.; Meijerink, A. Appl. Phys. Lett. 2001, 79 (25), 4222–4224. (28) Ivanov, S. A.; Nanda, J.; Piryatinski, A.; Achermann, M.; Balet, L. P.; Bezel, I. V.; Anikeeva, P. O.; Tretiak, S.; Klimov, V. I. J. Phys. Chem. B 2004, 108 (30), 10625–10630.

LETTER

(29) Brovelli, S.; Schaller, R. D.; Crooker, S. A.; Garcia-Santamaria, F.; Chen, Y.; Viswanatha, R.; Hollingsworth, J. A.; Htoon, H.; Klimov, V. I. Nat. Commun. 2011, 2. (30) Zavelani-Rossi, M.; Lupo, M. G.; Tassone, F.; Manna, L.; Lanzani, G. Nano Lett. 2010, 10 (8), 3142–3150. (31) Grimmeiss, H. G.; Kullendorff, N.; Broser, R. J. Appl. Phys. 1981, 52 (5), 3405–3412. (32) Grimmeiss, H. G.; Ovren, C.; Ludwig, W.; Mach, R. J. Appl. Phys. 1977, 48, 5122. (33) Kuno, M.; Nirmal, M.; Bawendi, M. G.; Efros, A.; Rosen, M. J. Chem. Phys. 1998, 108 (10), 4242–4247. (34) Pandey, A.; Guyot-Sionnest, P. Science 2008, 322 (5903), 929–932. (35) Wuister, S. F.; de Mello Donega, C.; Meijerink, A. J. Phys. Chem. B 2004, 108 (45), 17393–17397. (36) Klimov, V. I.; McBranch, D. W.; Leatherdale, C. A.; Bawendi, M. G. Phys. Rev. B 1999, 60 (19), 13740.

4758

dx.doi.org/10.1021/nl202572c |Nano Lett. 2011, 11, 4753–4758