ZnSe Nanodumbbells - ACS

Heavy-Metal-Free Fluorescent ZnTe/ZnSe Nanodumbbells ... Publication Date (Web): June 27, 2017 ... Herein, we report a colloidal synthesis of fluoresc...
2 downloads 8 Views 3MB Size
Heavy-Metal-Free Fluorescent ZnTe/ZnSe Nanodumbbells Botao Ji, Yossef E. Panfil, and Uri Banin* The Institute of Chemistry and Center for Nanoscience and Nanotechnology, The Hebrew University of Jerusalem, Jerusalem 91904, Israel S Supporting Information *

ABSTRACT: For visible range emitting particles, which are relevant for display and additional applications, Cdchalcogenide nanocrystals have reached the highest degree of control and performance. Considering potential toxicity and regulatory limitations, there is a challenge to successfully develop Cd-free emitting nanocrystals and, in particular, heterostructures with desirable properties. Herein, we report a colloidal synthesis of fluorescent heavy-metal-free Znchalcogenide semiconductor nanodumbbells (NDBs), in which ZnSe tips were selectively grown on the apexes of ZnTe rods, as evidenced by a variety of methods. The fluorescence of the NDBs can be tuned between ∼500 and 585 nm by changing the ZnSe tip size. The emission quantum yield can be greatly increased through chloride surface treatment and reaches more than 30%. Simulations within an effective-mass-based model show that the hole wave function is spread over the ZnTe nanorods, while the electron wave function is localized on the ZnSe tips. Quantitative agreement for the red-shifted emission wavelength is obtained between the simulations and the experiments. Additionally, the changes in radiative lifetimes correlate well with the calculated decrease in electron−hole overlap upon growth of larger ZnSe tips. The heavy-metal-free ZnTe/ZnSe NDBs may be relevant for optoelectronic applications such as displays or light-emitting diodes. KEYWORDS: ZnTe/ZnSe, nanodumbbells, type II, heavy-metal-free, fluorescence

C

explore and develop heavy-metal-free semiconductor nanocrystals. In particular, consider a seemingly minor change from cadmium- to zinc-chalcogenides, which naively would appear as an easy translation of the chemical synthesis principles to a closely related element system. However, this modification is found already to be highly challenging, as reflected by the lack of strongly fluorescent well-controlled zinc-chalcogenide nanocrystals.9,10 To this end, we report here a colloidal synthesis of fluorescent heavy-metal-free ZnTe/ZnSe nanodumbbells (NDBs) in which ZnSe tips were grown on both apexes of ZnTe nanorods. The NDB structure was reported initially for selective growth of Au on the apexes of CdSe rods, exploiting the intrinsic anisotropic rod reactivity associated with its wurtzite crystal structure.11 This asymmetric chemical reactivity of the rods was later also employed for synthesis of semiconductor heterostructures including CdSe rods tipped by PbSe or CdTe, and analogous structures were realized also for CdS rods with CdSe or ZnSe tips.12−23 But so far, NDBs have been made only for

olloidal semiconductor nanostructures have reached a high level of synthetic control, yielding intense fluorescence with properties tuned by size, composition, and shape. 1−3 This has led to their successful implementation as components of back light systems in actual displays on the market, providing high color quality accompanied by energy saving characteristics.4 An additional area in which fluorescent semiconductor nanocrystals have found successful applications is as biological taggants, where their exceptional photophysical properties offer some advantages over other fluorescent tags such as more traditional fluorescent dyes and fluorescent proteins.5 However, tallying the successful materials systems in which these incredible achievements have been realized shows that they mostly contain heavy metal elements.6 For emission in the visible range, cadmium-chalcogenides are the nanocrystal materials of choice so far, providing well-controlled narrow fluorescence peaks with quantum yields approaching unity.7,8 Regulatory aspects and considerations on toxicity of materials unfortunately impede the widespread implementation of heavy metal and, in particular, cadmium-containing semiconductor nanocrystals. Considering this and the high interest in furthering the accomplishments of colloidal chemical control demonstrated for the Cd-chalcogenides, it is an important challenge to © 2017 American Chemical Society

Received: May 16, 2017 Accepted: June 27, 2017 Published: June 27, 2017 7312

DOI: 10.1021/acsnano.7b03407 ACS Nano 2017, 11, 7312−7320

Article

www.acsnano.org

Article

ACS Nano heavy-metal-containing semiconductors, specifically based on CdSe and CdS nanorods. In the more common core/shell systems, either both electron and hole (for type I band alignment) or one of the charge carriers (for type II band alignment) will be localized in the core, which can introduce some limitations for charge injection (e.g., in a light-emitting diode) or extraction purposes (e.g., for photocurrent generation).24 Conversely, the dumbbell architecture offers direct access to either part of the heterostructure.14,16 For example, when the two semiconductors have a type II band alignment where either the conduction band or valence band of one semiconductor is located in the band gap of a second semiconductor, the NDB morphology allows electron−hole charge separation into the two different parts, which both directly touch the surroundings. The interest and potential of NDBs were indeed most clearly demonstrated recently, where Oh et al. showed that CdS-CdSe-ZnSe doubleheterojunction NDBs (CdS rods with CdSe/ZnSe tips) allowed both electroluminescence and photocurrent generation upon light illumination.25 The fabricated LEDs from these NDBs were also responsive to external light and may be used in various interesting display modalities. In this report, colloidal heavy-metal-free ZnTe/ZnSe NDBs are synthesized. The NDBs show tunable emission from ∼500 to ∼585 nm, with quantum yields that reach more than 30% after chloride treatment. The relatively long fluorescence lifetime is in line with partial charge separation as expected for the type II band alignment and consistent with effectivemass-based modeling combining quantum confinement and Coulomb interaction effects. Quantitative comparison of the model calculations with the experimental lifetimes and wavelength shifts upon ZnSe tip growth provides a good understanding of the optoelectronic properties of the NDBs.

Figure 1. TEM images of ZnTe nanorods (A) and ZnTe/3ZnSe NDBs (B), along with schematic illustrations of the structures. (C) Absorption spectra (Abs) and photoluminescence spectra (PL) of ZnTe nanorods and ZnTe/3ZnSe NDBs; no emission was observed from bare ZnTe nanorods. (D) Scheme of band alignments in ZnTe/ZnSe NDBs presenting the spatially indirect charge recombination; bulk values of ZnTe and ZnSe are used.

polytelluride species, leading to formation of irregular shaped nanoparticles. The synthesized colloidal ZnTe nanorods did not show any photoluminescence (PL), consistent with previous reports.26,29−32 The absence of fluorescence likely results from surface trap states, and indeed the ZnTe rods are highly sensitive to oxidation. Exposing purified ZnTe nanorods (dispersed in nonpolar organic solvent) to air for several hours leads to the appearance of a black precipitate, indicating the production of Te from the oxidation of ZnTe. Thereby, all the manipulations in this work were performed strictly under conditions free of oxygen and water to avoid oxidation. In order to obtain fluorescence, a second semiconductor material can be grown on ZnTe nanorods to modify this nanostructure. The available heavy-metal-free material candidates are the other Zn-chalcogenides and their alloys. First, ZnS growth on ZnTe nanorods was attempted. However, only very weak band gap emission is obtained in these experiments (Figure S2). The poor optical properties of ZnTe/ZnS likely originate from incomplete shell growth of ZnS or from interfacial defects. Both behaviors can be caused by the large lattice mismatch between ZnTe and ZnS (11.4%).32 Compared with ZnS, ZnSe has smaller lattice mismatch with ZnTe (6.4%). The growth of ZnSe tips on ZnTe nanorods was performed via a layer-by-layer method, in which suitable calculated amounts of Zn and Se precursors were added sequentially. Purified ZnTe nanorods were dispersed in the mixture of trioctylphosphine (TOP) and oleylamine (OAm). Zinc and selenium precursors with the calculated amounts for growth of one complete shell monolayer on the rods were alternately added every 15 min at 240 °C under Ar. Hereafter, we use ZnTe/nZnSe to represent ZnTe/ZnSe NDBs with the addition of ZnSe precursors calculated for n monolayers. When the sequential additions of ZnSe precursors with desired

RESULTS AND DISCUSSION As a starting point for the NDB synthesis, short ZnTe nanorods were first synthesized according to a published method with modifications.26 Already at this step, achieving well-controlled quasi-1D nanorods of ZnTe is significantly more difficult compared to the Cd-chalcogenide nanorods, in which the surfactant-controlled growth is well developed.2,27,28 For the ZnTe nanorod synthesis, a highly reactive polytelluride solution, which was prepared by mixing superhydride solution and TOP-Te, was injected into a zinc oleate solution at 160 °C. The temperature was increased to 240 °C at a rate of 5 °C/min. Relatively monodispersed ZnTe nanorods with a diameter of 4.6 nm and length of 12 nm are obtained after 50 min of growth at 240 °C, as shown in the transmission electron microscopy (TEM) image, Figure 1A. The absorption spectrum exhibits the excitonic peak around 463 nm (Figure 1C), indicating a narrow diameter dispersion. An increased heating rate of 10 °C/min results in the formation of ill-defined ZnTe nanorods (Figure S1). It is therefore interesting to consider the growth mechanism of ZnTe rods to better understand the role of different control parameters in this reaction. The highly reactive polytellurides contain a mixture of reduced Te species including Te2−, Te22−, and Te32−, which have different reactivity.26 The most reactive Te2− ions react with the zinc precursor and nucleate in wurtzite phase at the lower temperature (160 °C). The other reduced Te species then react at elevated temperatures and grow on specific facets to form the ZnTe nanorods. Overly rapid heating can hamper the growth balance between the different 7313

DOI: 10.1021/acsnano.7b03407 ACS Nano 2017, 11, 7312−7320

Article

ACS Nano

Figure 2. Evolution of (A) absorption and (B) emission spectra in an exemplary synthesis of “ZnTe/3ZnSe” NDBs; the zinc and selenium precursors with the calculated amounts for growing one complete shell on the existing nanoparticles were alternately added every 15 min at 240 °C; after the third addition of selenium precursor, only zinc precursor was added every 30 min to further promote the growth of ZnSe as well as the surface passivation. (a) Bare ZnTe nanorods; (b−d) 1, 2, and 3 monolayers of ZnSe; (e and f) 2 and 4 more zinc additions. (C) Evolution of PL wavelength and PL QY as a function of reaction time.

Figure 3. (A) XRD of ZnTe NRs (black) and ZnTe/ZnSe NDBs with increased amounts of ZnSe precursors: 2 (blue) and 4 (red) monolayers. The standard XRD stick patterns of bulk wurtzite ZnTe and zinc-blende ZnSe are also shown for comparison. High-resolution TEM (HRTEM) images of ZnTe nanorods (B) and ZnTe/3ZnSe NDBs (C). (E) Elemental analysis of Se (red line) and Te (blue line) (EDX spectra line scan, smoothed; see Materials and Methods) along the long axis of a single ZnTe/3ZnSe NDB. Corresponding STEM image of the measured ZnSe/ZnTe NDBs is shown in (D), with the thick red line indicating the scan axis. All scale bars are equal to 5 nm.

amounts were completed, excess zinc stock solution was injected at 240 °C, promoting the reaction with residual selenium on the surface to improve the surface passivation (see details in the Materials and Methods). Figure 1B shows a TEM image of ZnTe/3ZnSe heterostructures with dumbbell morphology. The size histogram indicates that the average length of the NDBs is 16.1 nm with an average tip width of 6.3 nm (Figure S4C). Comparing with the ZnTe rod length, an average elongation of 2.1 nm along the rod direction on both apexes of the ZnTe nanorods is inferred. The absorption spectrum of the NDBs manifests a shoulder at ∼470 nm and a featureless tail out to ∼550 nm (Figure 1C). The NDBs exhibit bright PL at ∼580 nm with a photoluminescence quantum yield (PL QY) of ∼18% at room temperature. Both absorption and PL are reflecting the ZnTe/ ZnSe type II character, in which the holes are confined in the ZnTe rod, whereas the electrons are localized in the ZnSe tips according to the band alignments of the bulk semiconductors (Figure 1D). The PL originates from the radiative spatial indirect recombination of excitons, bound by the Coulomb interaction, with an effective band gap determined essentially by the valence band level of the ZnTe and the conduction band level of the ZnSe tips, and is smaller than the band gap of either ZnTe or ZnSe.29,33

To monitor the progress of the ZnSe tip growth, we measured the absorption and PL spectra during the synthesis of ZnTe/3ZnSe as a function of reaction time (Figure 2). After the first addition of Zn and Se precursors, a red shift of ∼7 nm of the excitonic peak (to ∼470 nm) is observed in the absorption spectra. Due to the ZnSe growth, the nanoparticles become emissive (at ∼497 nm) as shown in Figure 2B (b), although the PL QY is very low at this stage. As more Zn and Se precursors are introduced, the exciton signature of the ZnTe rods can be still discerned throughout the synthesis (marked in the dashed ellipse in Figure 2A). This is consistent with the formation of the NDB structure. Meanwhile, a small absorption tail develops and shifts to the red along with red shift of the emission, indicating the decrease of the effective band gap upon the ZnSe tip growth related to the type II structure. The PL shifts from ∼500 nm to ∼580 nm, and the PL QY steadily increases with addition of further Zn and Se precursors (Figure 2C). The PL continues to red-shift with time even after the last selenium precursor addition (spectra d to f in Figure 2B and Figure S5, black line), indicating incomplete reaction during the Se addition time intervals. Thereby, heating is continued until no more red shift is observed, during which more zinc oleate is added to promote ZnSe growth and improve the surface passivation. The obtained nanoparticles from adding additional 7314

DOI: 10.1021/acsnano.7b03407 ACS Nano 2017, 11, 7312−7320

Article

ACS Nano

Figure 4. (A) Calculated electron and hole wave functions for NDB samples of ZnTe/1ZnSe and ZnTe/3ZnSe from the simulations including Coulomb interaction. Shown are color-based maps of the probability distributions over a silhouette of the NDB structures, as well as the cross section of the wave functions along the c-axis overlaid on the potential diagram. (B) Measured PL emission spectra and (C) corresponding PL decay traces of ZnTe/nZnSe NDBs with increased tip sizes for n = 1−4. (D) Comparison of experimental (red triangles) and calculated PL emission energies (blue circles) of ZnTe/ZnSe NDBs as a function of ZnSe tip width along the c-axis of the ZnTe nanorod. An aspect ratio of 3 between the widths along and perpendicular to the c-axis of the ZnTe nanorod is used in the calculations based on the measured values. (E) Comparison of the calculated relative exciton overlap integrals (blue circles) of ZnTe/ZnSe NDBs as a function of the ZnSe tip width with the overlap integral of ZnTe/1ZnSe NDBs used as the reference. Experimental relative radiative transition rates are marked as red triangles.

zinc oleate display a larger red shift of emission as well as a higher PL QY (Figure S5). The PL QY increases from less than 1% to ∼18% at the end of the growth process. ZnSe tip growth on ZnTe nanorods can also be performed by injecting selenium precursor (TOP-Se) into the solution of ZnTe nanorods dispersed in the mixture of TOP, oleylamine, and zinc oleate at high temperature. A similar dumbbell structure, but less welldefined, is obtained, and the PL QY is much lower than that via the above layer-by-layer method (Figure S3). Powder X-ray diffraction (XRD) of bare ZnTe nanorods is consistent with a wurtzite structure of ZnTe, as shown in Figure 3A. The slightly sharper (002) peak at ∼25° indicates the favorable growth along the c-axis. Upon ZnSe growth, a slight shift to higher angle (