This is an open access article published under an ACS AuthorChoice License, which permits copying and redistribution of the article or any adaptations for non-commercial purposes.
Article pubs.acs.org/JACS
One-Dimensional Carrier Confinement in “Giant” CdS/CdSe Excitonic Nanoshells Natalia Razgoniaeva,†,‡ Pavel Moroz,†,‡ Mingrui Yang,†,‡ Darya S. Budkina,†,§ Holly Eckard,†,‡ Marissa Augspurger,∥ Dmitriy Khon,∥ Alexander N. Tarnovsky,†,§ and Mikhail Zamkov*,†,‡ †
The Center for Photochemical Sciences, ‡Department of Physics, and §Department of Chemistry, Bowling Green State University, Bowling Green, Ohio 43403, United States ∥ Department of Chemistry and Biochemistry, St. Mary’s University, San Antonio, Texas 78228, United States S Supporting Information *
ABSTRACT: The emerging generation of quantum dot optoelectronic devices offers an appealing prospect of a sizetunable band gap. The confinement-enabled control over electronic properties, however, requires nanoparticles to be sufficiently small, which leads to a large area of interparticle boundaries in a film. Such interfaces lead to a high density of surface traps which ultimately increase the electrical resistance of a solid. To address this issue, we have developed an inverse energy-gradient core/shell architecture supporting the quantum confinement in nanoparticles larger than the exciton Bohr radius. The assembly of such nanostructures exhibits a relatively low surface-to-volume ratio, which was manifested in this work through the enhanced conductance of solutionprocessed films. The reported core/shell geometry was realized by growing a narrow gap semiconductor layer (CdSe) on the surface of a wide-gap core material (CdS) promoting the localization of excitons in the shell domain, as was confirmed by ultrafast transient absorption and emission lifetime measurements. The band gap emission of fabricated nanoshells, ranging from 15 to 30 nm in diameter, has revealed a characteristic size-dependent behavior tunable via the shell thickness with associated quantum yields in the 4.4−16.0% range.
■
INTRODUCTION Colloidal processing of quantum dot (QD) solids represents a promising strategy for solution deposition of inorganic semiconductor devices featuring tunable optoelectronic properties.1−14 The task of transitioning from bulk semiconductors to assemblies of many nanoparticles, however, poses significant challenges. The key issue concerns a high density of defect states associated with nanocrystal boundaries and interfaces15−18 that introduce potential energy minima. These lattice sites tend to trap photoinduced charges or induce their recombination,19 which dramatically shortens the excited state lifetime in a solid causing a corresponding reduction in the carrier mobility and diffusion length. Improving the electrical conductivity of QD solids is often attempted through the use of large-size monocrystalline nanoparticles that show some degree of quantum confinement. In this regard, semiconductor nanotubes,20 nanosheets,21−23 nanorods,24−26 and atomically coherent superlatices27 are promising architectures that extend single-crystalline conduction channels beyond the size of an exciton Bohr radius. Incorporating these materials into devices, presently, is complicated by the challenging assembly protocols which require nontrivial processing steps. For instance, the alignment of nanorods or nanotubes perpendicular to the substrate © 2017 American Chemical Society
improves the charge transport characteristics of the device but is increasingly difficult to achieve for large aspect-ratio structures. Likewise, nanosheets or tetrapods can be challenging to assemble into an ordered pattern that assimilates the advantages of these geometries. As a result, one- or twodimensional colloidal nanocrystals (NCs) still rarely appear in record-efficiency devices.28 Here, we demonstrate an inverse energy-gradient nanoparticle architecture that supports the formation of twodimensional excitons in the shell domain. The developed geometry places a wide-gap semiconductor (CdS) at the core of the composite nanoparticle in order to funnel the photoinduced energy into the low-gap CdSe shell layer (Figure 1). As a result, the quantum confinement is achieved in 15−30 nm nanoparticles which dimensions exceed the exciton Bohr radius for CdSe (Figure 1b,c). The formation of excitons in the CdSe shell was manifested through a size-tunable emission and the characteristic step-like absorption profile. On the basis of ultrafast transient absorption measurements, we conclude that the excitation of a bulk-like core domain leads to a ∼ 2 ps recovery of the CdS bleach, attributed to electron cooling, Received: February 28, 2017 Published: May 23, 2017 7815
DOI: 10.1021/jacs.7b02054 J. Am. Chem. Soc. 2017, 139, 7815−7822
Article
Journal of the American Chemical Society
functions tend to delocalize over the entire heterostructure leading to the loss of the quantum confinement if the total nanocrystal size exceeds the exciton Bohr radius.29−33 To suppress such delocalization, the shell layer was designed to be sufficiently thick to exhibit an inverse type I band edge alignment (Figure 1a), which was evident through the onset of a characteristic CdSe band gap emission. Furthermore, the deposition of a thin layer of a wide-gap semiconductor (ZnS) was attempted at the interface of the core and the shell domains. Such an interstitial domain was expected to create a potential energy barrier to photoinduced charges in CdS and CdSe. The experimental deposition of the “sandwiched” ZnS phase, however, proved to be a challenging task due to Cd2+ to Zn2+ cation exchange, leading to only a miniscule amount of Zn in fully grown nanocrystals. At the final stage, a thin layer of either ZnS or CdS semiconductor was grown on the surface of the CdSe light-emitting shell to suppress carrier trapping at dangling bonds. CdS/CdSe nanoshell heterostructures were synthesized in a single pot reaction using a dropwise injection of precursors. Starting with small diameter CdS seeds, sequential layers of CdS1, ZnS1, CdS2, CdSe, and ZnS2 were grown at an approximate rate of 1 monolayer per 20 min. The first shell of CdS1 was designed to augment the size of the CdS core domain to 8−14 nm, resulting in the loss of the quantum confinement. The subsequent 1−2 monolayers (MLs) of the ZnS1 shell were further overcoated with additional 1−2 MLs of CdS2 that served as a stress relief phase between ZnS1 and CdSe domains (lattice mismatch ∼11.6%). In the next step, a CdSe light-emitting layer was grown until the excitonic feature corresponding to the 1S(e) − 1S3/2(h) transition was clearly pronounced in the absorption spectra. During this stage, the precursor injection speed was kept sufficiently slow to prevent the formation of isolated CdSe NCs. Under these conditions, stopping the injection of precursors at any time during the reaction lead to an almost simultaneous cessation of changes in the evolution of the nanoshell absorption profile. Such “adiabatic” growth conditions were found to be important for the uninterrupted growth of consecutive shells (CdS1→ ZnS1 → CdS2 → CdSe → passivating layer: CdS3 or ZnS2). An important benefit of the slow growth conditions used during the shell deposition stages lied in the ability to maintain thermodynamic focusing of nanocrystal shapes throughout the entire procedure. This trend is manifested in the transmission electron microscope (TEM) images of nanoshell structures corresponding to the sequential growth stages (Figure 2). According to Figure 2a, CdS/CdS1 core nanoparticles (d = 8.6 nm) grown in the first step of the procedure exhibited a relatively low size dispersion (∼5.5%), which is notoriously difficult to achieve in the case of large-diameter NCs.34 The subsequent overcoating of these nanoparticles with ZnS1 and finally CdS2 layers increased the total particle diameter to 11.6 nm and the corresponding size dispersion to 9.5%. A lattice strain at the ZnS/CdS interface and the instability of Zn2+derived surface complexes were believed to be responsible for a reduced morphology control during this stage. The subsequent addition of the Cd and Se precursors resulted in a slight transformation of nanoparticle shapes from spherical to faceted with an average diameter of 19.6 nm (size dispersion ∼12%). Nevertheless, CdS/CdSe nanoshells appeared to be nearly spherical even for particles approaching 30 nm in diameter (see Figures SF2 and SF3). After the addition of a thin passivating shell (ZnS2), the product was examined using a low-resolution
Figure 1. (a) Schematic illustration of the CdS/CdSe nanoshell geometry. The potential energy minima of the conduction and valence bands associated with the CdSe semiconductor promote the shelllocalization of both photoinduced charges. In this geometry, the core dimensions are allowed to exceed the exciton Bohr radius, leading to the quantum confinement in nanostructures approaching 30 nm in diameter. (b) A high-resolution transmission electron microscope (TEM) image of a typical CdS/CdSe nanoshell heterostructure. (c) The corresponding high angle annular dark field (HAADF)-STEM image of a CdS/CdSe nanoshell. (d) Atomic element mapping using a STEM energy dispersive X-ray detector.
followed by a much slower nanosecond bleach recovery in the CdSe layer. Notably, solution-processed solids of CdS/CdSe nanoshells showed a 7-fold enhancement in the device photocurrent relative to similarly processed films of spherical CdSe nanocrystals. The spatial confinement of photoinduced charges in the shell domain of a composite nanoparticle can lead to several interesting properties that reflect the unique dimensionality of such architecture. For instance, the shell-localization of excitons should promote an enhanced surface reactivity, which is useful for applications in photocatalysis, light sensitization, and biological/chemical sensing. Another appealing aspect of the demonstrated architecture is an intrinsically greater conductivity of disordered nanoshell solids compared to analogous films of spherical QDs (Figure 6), which was tentatively attributed to relatively lower interfacial area associated with nanoshell assemblies (see Figure SF1). Along these lines, we expect that nanoshell solids could potentially represent a useful geometry for improving charge transport characteristics in quantum dot solar cells and field effect transistors.
■
RESULTS AND DISCUSSION The surface localization of excitons is generally expected for inverse type I core/shell heterostructures, where both charge types find potential minima in the shell. Several previous works on the synthesis and spectroscopy of CdS/CdSe colloidal nanocrystals, however, have indicated that electronic wave 7816
DOI: 10.1021/jacs.7b02054 J. Am. Chem. Soc. 2017, 139, 7815−7822
Article
Journal of the American Chemical Society
The onset of the quantum confinement in a CdSe shell was observed for domain sizes as low as 2 nm in thickness and persisted until a 6 nm CdSe layer was reached (the size-tunable CdSe emission is plotted in Figure SF5). The volumes of CdS and CdSe semiconductor phases in a CdS1/ZnS1/CdS2/CdSe nanoshell structure, determined from TEM images of CdS2 and CdS2/CdSe samples, were consistent with the elemental energy dispersive X-ray (EDX) analysis of grown nanoparticles. According to Figure 3b, the Se to S elemental fraction in nanoshells composed of a 11.6 nm CdS and 4 nm CdSe shell was approximately 2:5, corroborating the expected volume ratio. The increased density of Se anions at the surface of nanoshells was also seen in energy dispersive X-ray STEM images, which show an atomic element distribution for several specimens (Figure 1d). Surprisingly, only a small amount of the interstitial Zn (ZnS1) was detected (∼1%, by Zn:Cd atomic ratio) in the ensemble-averaged EDX spectra (Figures 3b and SF7c) despite a noticeable onset of the ZnS band gap absorption observed during the CdS/ZnS1 growth stage (Figure 3a, blue arrow) as well as the substantial amount of Zn in CdS/ZnS1 nanoparticles (Figure SF7a). We speculate that Zn2+ cations could be partially displaced from the surface of growing nanoparticles during the CdS2 and CdSe overcoat stage. Cd2+ → Zn2+ cation exchange, facilitated by the 260−300 °C growth temperature,35 is also likely to diminish the fraction of Zn in growing nanoshells (Figure SF7), which appears to be correlated with the reduction in the fluorescence quantum yield of Zn-deprived CdS/CdSe. The expulsion of Zn is likely to be driven by the affinity of CdS/CdSe nanoshells to reduce the overall lattice strain associated with the ZnS1 interstitial “channel” and is facilitated by the slow “adiabatic” growth conditions. Multiple tests, however, have shown that the presence of even a small amount of ZnS1 leads to a brighter emission of the CdSe shell (see Figure 4). The X-ray powder diffraction (XRD) measurements of the crystal phase corresponding to 19.6 nm CdS/CdSe samples are summarized in Figure 3c. Based on the observed Bragg peak positions, we conclude that multiple sublayers of a complete nanoshell heterstructure (CdS1/ZnS1/CdS2/CdSe) appear to grow in a wurtzite crystallographic phase with an average lattice
Figure 2. TEM images of nanoshells corresponding to the three successive synthetic stages. (a,d) 8.6 nm CdS core NCs obtained by overcoating the original 4.5 nm CdS NC seeds with the additional layer of CdS1. (b,e) 11.6−nm nanostructures resulting from the addition of successive ZnS1 and CdS2 layers onto 8.6 nm CdS/CdS1 core NCs, which is accompanied by a thermally controlled diffusion of Zn2+ cations into the shell layer. (c,f) The transformation of nanoparticle shapes upon the addition of ∼4 nm CdSe shell (total nanocrystal diameter ∼19.6 nm). The core/shell morphology can be seen in the high-resolution TEM image. (g) Statistical diameter distribution of CdS core NCs. (h) Statistical size distribution of CdS/ ZnS1/CdS NCs. (i) Statistical size distribution of CdS/CdSe/ZnS2 nanoshells.
TEM to confirm the absence of isolated nanocrystals (see Figure SF4).
Figure 3. (a) Evolution of the absorption and emission spectra associated with the successive stages of the nanoshell growth. The corresponding nanoparticle diameter is indicated at the top of the figure. (b) The elemental analysis of fabricated nanoshells (CdS1/ZnS1/CdS2/CdSe) using SEM energy dispersive X-ray. (c) The X-ray powder diffraction (XRD) analysis of the crystal phase corresponding to 19.6 nm CdS/CdSe samples. Multiple sublayers of a complete nanoshell heterstructure (CdS1/ZnS1/CdS2/CdSe) appear to grow in a wurtzite crystallographic phase with an average lattice spacing lying between those of CdS and CdSe semiconductors (panel (c), red and yellow lines, respectively). 7817
DOI: 10.1021/jacs.7b02054 J. Am. Chem. Soc. 2017, 139, 7815−7822
Article
Journal of the American Chemical Society
thickness dispersion or an uneven shell growth within a single structure that results in funneling of excitons to the low-energy side. In the final step of the synthesis, nanoshells were overcoated with either ZnS or CdS “passivating” shell aimed to increase the emission quantum yield. In both treatment strategies, the emission enhancement was apparent. A giant CdS shell was found to be particularly effective in this regard, leading to the corresponding QY enhancement of up to 2-fold (reaching QY = 16% for CdS-capped 16.3 nm CdS/CdSe nanoshells). The dynamics of two-dimensional excitons in the shell domain was inferred from the lifetime of the CdSe band gap emission. The two nanoshell geometries featuring different amounts of the interstitial ZnS grown onto same-sized CdS1 core NCs have been employed to discern the effect of the intermediate layer. The emissive CdSe shell in both cases was approximately 2.5−2.8 nm in thickness. The respective sizes of the ZnS1 intermediate shells for the two investigated geometries were initially determined from the TEM analysis of CdS/ZnS1 nanocrystals as
