CdTe Nanowires by Au-Catalyzed Metalorganic Vapor Phase Epitaxy

Jun 14, 2017 - Nano Letters ..... narrowly monodispersed around an average contact angle (θave = 130.5° in ... As the energy of stacking fault (SF) ...
0 downloads 0 Views 1MB Size
Subscriber access provided by CORNELL UNIVERSITY LIBRARY

Communication

CdTe nanowires by Au-catalyzed metalorganic vapor phase epitaxy Virginia Di Carlo, Paola Prete, Vladimir G. Dubrovskii, Yury Berdnikov, and Nico Lovergine Nano Lett., Just Accepted Manuscript • Publication Date (Web): 14 Jun 2017 Downloaded from http://pubs.acs.org on June 14, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Nano Letters is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

CdTe nanowires by Au-catalyzed metalorganic vapor phase epitaxy Virginia Di Carlo,† Paola Prete,‡ Vladimir G. Dubrovskii,§,#,⊥ Yury Berdnikov,§,⊥ and Nico Lovergine*,† † Dipartimento di Ingegneria dell’Innovazione, Università del Salento, Via Monteroni, I-73100 Lecce, Italy ‡ Istituto per la Microelettronica e Microsistemi, SS Lecce, Via Monteroni, I-73100 Lecce, Italy § St. Petersburg Academic University, Khlopina 8/3, 194021 St. Petersburg, Russia # Ioffe Physical Technical Institute RAS, Politekhnicheskaya 26, 194021 St. Petersburg, Russia ⊥ ITMO University, Kronverkskiy pr. 49, 197101 St. Petersburg, Russia

KEYWORDS. CdTe nanowires, Au-catalyzed growth, metalorganic vapor phase epitaxy, VLS growth modelling, cathodoluminescence. ABSTRACT: We report on the first Au-catalyzed growth of CdTe nanowires by metalorganic vapor phase epitaxy. The nanowires were obtained by a separate precursors flow process, in which (i) di-isopropyl-telluride (iPr2Te) was first flowed through the reactor to ensure the formation of liquid Au-Te alloy droplets, and (ii) after purging with pure H2 to remove unreacted iPr2Te molecules from the vapor and the growth surface, (iii) dimethyl-cadmium (Me2Cd) was supplied to the vapor so that Cd atoms could enter the catalyst droplets, leading to nanowire self-assembly. CdTe nanowires were grown between 485°C and 515°C on (111)B-GaAs substrates, the latter preliminary deposited with a 2-μm thick (111)-oriented CdTe buffer layer onto which Au nanoparticles were provided. As-grown CdTe nanowires were vertical ([111]-aligned) straight segments of constant diameter and showed an Au-rich nano-droplet at their tips, the contact angle between the droplets and the nanowires being ~130°. The nanowire axial growth rate appeared kinetics-limited with an activation energy ~57 kcal/mol. However, the growth rate turned independent from the nanowire diameter. Present data are interpreted by a theoretical model explaining the nanowire growth through the diffusion transport of Te adatoms, under the assumption that their growth occurs during the Me2Cd-flow process step. Low temperature cathodoluminescence spectra recorded from single nanowires showed a well resolved band-edge emission typical of zincblend CdTe, along with a dominant band peaked at 1.539 eV. nanostructures of the relevant semiconducting materials].6,7 In particular, the synthesis/fabrication of nano-crystalline semiconductors in the form of nanowires offers significant physical advantages over single- or polycrystalline PV devices: (i) the shortened photo-generated carrier transit lengths in p-n junction core-shell nanowires are predicted to reduce recombination losses,8 (ii) the antireflection and light-trapping effects associated with dense nanowire arrays9,10 would allow high performance cells with minimal materials usage, and (iii) the materials lattice mismatch in heterostructured nanowires can be accommodated more easily than in planar structures,11 allowing for a greater flexibility in heterostructure device design and tailoring of the PV cell performances.

CdTe is the most widely studied material among II-VI compound semiconductors due to its great potentials in a variety of solid-state device applications, especially in the areas of photovoltaic (PV) energy conversion,1 and X-/γray radiation detection.2 In particular, its near-infrared direct band gap (Eg=1.49 eV at 300 K, Ref. 3) matches the maximum of solar energy spectrum, leading to high optical absorption coefficients for photons in the visible portion of the sunlight. After several decades of intense research efforts and slow progresses, laboratory-scale CdS/CdTe poly-crystalline thin films PV cells have recently achieved power conversion efficiency (PCE) figures close to 21%,4 which remain however, far below theoretical expected values (28-30%). The origin of such discrepancy has been ascribed to enhanced carrier recombination loss at CdTe grain boundaries and associated point defects in current poly-crystalline cells.5

Recently, radial p-n junction nanowire solar cells based on dense arrays of CdTe nanowires coated with indium tin oxide (ITO)/ZnO/CdS triple shells were demonstrated,12 although their reported PCE figures were limited to ∼2.5%. Clearly, to further improve these figures

A widespread strategy to improve the efficiency of PV cells further is to implement cell architectures containing

1 ACS Paragon Plus Environment

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

and fully exploit the physical advantages of CdTe-based nanowires to the realization of high efficiency cells, requires developing suitable techniques for the synthesis of free-standing nanowires with a large degree of control over materials purity/doping, crystalline properties and morphology, along with the necessary scalability (largearea growth) for practical application to the field of photovoltaics. Metal-catalyzed (so-called vapor–liquid–solid, VLS) growth13 has been proven a relatively straightforward technique for growing a large variety of inorganic semiconductor nanowires.14-17 The method relies on the ability of a suitable metal-catalyst nanoparticle (NP) (usually Au for III-V compound semiconductors) to react and alloy at relatively low temperatures with one or more metal elements constituting the semiconductor crystal. The catalyst NP acts as a solvent for the metal(s) similarly to the case of liquid phase epitaxy. It is assumed that, when in contact with a nutrient vapor phase (referred to as the Vapor in the VLS acronym) and for temperatures above the eutectic melting point of the intermetallic alloy, the NP forms a supersaturated liquid droplet (the Liquid), leading to the nucleation and subsequent growth of the semiconductor nanowire (the Solid). Contrary to the case of III-V semiconductors, only a limited number of reports have appeared to date on the self-assembly of CdTe nanowires by the VLS mechanism. This may be explained by the limited availability of metal catalysts that, when alloyed with Cd and Te, form eutectic compositions with melting temperatures comparable to or below those (300-500°C) required for growing devicequality CdTe from the vapor. Bi-catalyzed growth of wurtzite phase CdTe nanowires by pulsed laser deposition has been demonstrated in the 365-390°C interval,18 although Bi evaporation proclivity at these temperatures makes difficult to nucleate/control the amount of catalyst seeds onto the substrate surface. Au-catalyzed growth of zincblend CdTe nanowires well above 450°C has been reported by molecular beam epitaxy,19 physical vapor deposition,20 close-spaced sublimation (CSS),21 and thermal evaporation.22 No report exists yet in the literature on CdTe nanowires grown by Au-catalyzed metalorganic vapor phase epitaxy (MOVPE).

Page 2 of 10

In this work we demonstrate for the first time the Aucatalyzed self-assembly of zinc-blend CdTe nanowires by a specially devised MOVPE process. The nanowires were vertically grown on 2-µm thick CdTe epilayers deposited on (111)B-GaAs substrates. The morphology of as-grown nanowires, the value of the Au-NP/nanowire contact angle, and their size (diameter, length) distribution is here analyzed, along with the dependence of the nanowire axial growth rate on relevant process parameters. Present experimental data are further analyzed on the basis of a nanowire growth model based on the surface diffusion of Te adatoms. Finally, we report the results of low temperature cathodoluminescence (CL) measurements performed in a field emission scanning electron microscope (FE-SEM) on single CdTe nanowires. CdTe nanowire growth by the separate precursor flow process. The VLS self-assembly of semiconductor nanowires by the MOVPE process relies on the existence of a kinetic hindrance to (planar) epitaxial growth,23 a condition satisfied by growing at sufficiently low temperatures, but still above the eutectic melting point of the intermetallic alloy seed driving the VLS process. For Au-catalyzed CdTe nanowires, it would require temperatures above the Au-AuTe2 eutectic melting point (447°C). However, the MOVPE growth of CdTe at these temperatures is quite efficient, and can easily reach planar growth rates of several µm/h. In order to suppress planar growth in favor of the VLS growth, we devised a novel MOVPE process in which Te and Cd precursors were supplied at different times to the CdTe/GaAs substrate surface, onto which Au NPs were suitably prepared by an annealing step (see Experimental methods); after raising the temperature of the Au-NP/CdTe/(111)B-GaAs samples to the final growth value Tg (varied between 470°C and 515°C) diisopropyl-telluride (iPr2Te) + H2 was flowed through the reactor chamber for a time ∆ (varied in the 1-10 min interval). This step ensures the formation of liquid Au-Te alloy droplets onto the surface of the CdTe buffer layer. The chamber was then purged with pure H2 for a time ∆tH2=4 min to remove any unreacted iPr2Te molecule from vapor and the growth surface, after which dimethyl-cadmium (Me2Cd) + H2 was admitted to the

Figure 1. Schematic of process steps adopted for the Au-catalyzed MOVPE growth of CdTe nanowires by the separate precursors flow process.

2 ACS Paragon Plus Environment

Page 3 of 10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

reactor and flowed for a time ∆ =10 min. It is assumed that, during this step, Cd atoms enter the liquid Au-Te droplet and, upon reaching saturation, the growth of CdTe nanowires by the VLS mechanism is initiated. The VLS character of the growth is supported by the fact that our growth temperatures (in the 485−515 °C range) are higher than the eutectic melting temperature for the binary Au-AuTe2 alloy (see above) from which the nanowire growth starts. A liquid state of the catalyst droplet is thus guaranteed for low Cd concentrations in the droplet. However, as adding Cd can increase the melting temperature of the ternary alloy, we do not exclude the vapor-solid-solid (VSS) growth mechanism,24

or a combined regime, where droplets are partly solidified for higher Cd concentrations. A schematic of the main process steps is reported in Figure 1. Figure 2(a) shows a FE-SEM micrograph of an as-grown nanowire sample ( =507°C, ∆ =5 min), evidencing the formation of an array of vertically-standing nanowire structures. As the underlying CdTe buffer layer is oriented onto the GaAs substrate (see Supporting Information), this implies that the CdTe nanowires grew (homoepitaxially) with their major axis along the vertical direction, as commonly found for most Au-catalyzed III-V and II-VI nanowires.23,25,26

Figure 2. (a) FE-SEM micrograph (30,000x magnification, 45° tilt view) of [111]-aligned CdTe nanowires grown at 507°C (Au NPs by thermal de-wetting of a 1-nm thick Au film); (b) Count histogram of the Au-NP/nanowire contact angle (θ) for a sample grown at Tg=500°C. The solid (red) curve in (b) represents the Gaussian distribution function best-fitting the experimental data with average value θave=130.5°±0.1° (standard deviation σ=3.6°). Inset in (b): geometrical definition of the nanowire radius (R), the Au-NP radius (RNP) and the Au-NP/nanowire contact angle (θ).

Figure 3. (a) FE-SEM micrographs showing the typical nanowire morphology as a function of growth temperature (nanowires grown using Au NPs from colloidal solutions). (b) Arrhenius plot of the nanowire mean length (symbols) as a function of growth temperature (∆ =1 min). Error bars represent standard deviations. The dashed (red) line is the Arrhenius exponential function best-fitting the experimental data above 490°C, and corresponding to an activation energy EA=(56.8±5.6) kcal/mol. The solid curve is only a guide for the eyes.

1 ACS Paragon Plus Environment

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The nanowires have an almost constant diameter and show a nanodroplet at their upper tips. Spatially-resolved energy dispersive x-ray spectroscopy (EDXS) analyses performed on as-grown samples (see Supporting Information) showed a relatively intense Au signal within the catalyst droplets, confirming that the latter are made of an Au-alloy and that the nanowires were indeed grown by the Au-catalyzed mechanism. We further estimated the contact angle (θ) between the Au-alloy NP and the CdTe nanowire. Assuming a planar Au-NP/nanowire interface, the contact angle for each nanowire can be estimated by FE-SEM measurements of the Au NP (RNP) and nanowire (R) radius (Inset of Figure 2(b)), such that θ=arcsin(R/RNP). Figure 2(b) reports a count histogram of as-obtained θ values for a nanowire sample grown at 500°C. For all nanowires θ appears narrowly monodispersed around an average contact angle (θave = 130.5° in Figure 2(b)), whose value does not appreciably change with the growth temperature. Interestingly, it closely matches the values reported in the literature for Au-catalyzed GaAs nanowires.27

Page 4 of 10

sections and SFs is expected in present CdTe nanowires. The VLS process appears to quench below ~485°C. Under these conditions, the process leads to the formation of flat hexagonal hillocks onto the sample surface, with the Aualloy NPs still occurring at their tops. Figure 3(b) reports the nanowire mean length as function of inverse growth temperature (Arrhenius plot): above 490°C the nanowire axial growth appears kinetically activated, with the apparent activation energy around EA=57 kcal/mol. To further understand the growth dynamics of CdTe nanowires by the separate precursors flow process, we analyzed the influence of Te pre-deposition time ∆ on the nanowire length. Figure 4 shows that, for a constant growth temperature of 507°C, increasing ∆ results in a rapid increase of the mean nanowire lengths for ∆tTe in the 1−5 min interval, beyond which their length seems to remain limited to around 800 nm. In addition, while our data did not evidence any noticeable length-diameter correlation for present CdTe nanowires (see Supporting Information), they do show the occurrence of rather broad and asymmetric length distributions (LDs) (Figure 5). Growth model of CdTe nanowires by the separate precursors flow process. Our growth model tries to explain the above experimental observations under the assumption that CdTe nanowires grow during the Cdflow process step only, i.e. without any supply of Te from the vapor. We speculate that under these conditions the flux of Te to the Au-Cd-Te droplet occurs through the diffusion of Te adatoms from the substrate surface to the nanowire base, with a certain flux  0, and then along the nanowire sidewalls up to its top. The growth is assumed to be limited by the diffusion transport of Te in the Cd-rich conditions, as in the VLS growth of III-V nanowires under excess As flux.30 The nanowire axial growth rate is then determined by π R 2dL/ dt = Ω jdiff (L) where   is the diffusion flux to the top of a nanowire having the length L, Ω is the elementary volume of solid CdTe and the nanowire radius R is assumed independent of time (a condition supported by our experimental data). The fraction of the flux that reaches the catalyst droplet out of the base flux  0 is given by cosh-1(L/λ),31,32 where λ is the diffusion length of Te adatoms on the nanowire sidewalls. Hence, we have the axial nanowire growth rate in the form

Figure 4. Experimental mean lengths of CdTe nanowires (symbols) versus Te pre-deposition time ∆tTe (with  =507°C and ∆ =10 min), and best-fitting values (red solid curve) using Eq. (3) (see discussion in the main text). Error bars represent standard deviations.

We systematically investigated the morphology of CdTe nanowires and their size (diameter, length) as function of growth temperature and Te pre-deposition time ∆ . Asgrown nanowires showed diameters ranging from 60 to 350 nm and lengths between 40 nm and 1.7 μm. It appears that higher (above 490°C) temperatures promoted the nanowire axial (i.e., in the 〈111〉-direction) growth. Noteworthy is that many CdTe nanowires showed zigzagged sidewalls (see Figure 3(a)), as also often observed for III-V NWs,28,29 an effect ascribed to the presence of twin defects along the nanowire trunk. As the energy of stacking fault (SF) formation in CdTe is fairly low (31-34 erg/cm2), the nucleation of twinned-crystal

j diff (0) dL Ω = 2 dt πR cosh(L / λ ) (1) without any direct flux of Te atoms from the vapor. Our data (see Supporting Information) do not evidence any marked decrease of the nanowire lengths on their diameters, as usual in the diffusion-induced growth of IIIV (Refs. 33,34) and CdTe (Ref. 35) nanowires. From Eq. (1), the radius-independent growth rate requires that the diffusion flux  0 is proportional to  , that is, to the

2 ACS Paragon Plus Environment

Page 5 of 10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

surface area of the nanowire base. Such a situation is typical for surface growth of point islands at high diffusivities where the neighboring islands compete for the diffusion flux.36-38 On the other hand, the mean length of nanowires grown from the colloidal Au NPs at 507oC with ∆ =1 min is 912 nm, and is reduced to only 219 nm when grown from thermally annealed Au film (Table 1). This effect is not associated with a larger size of the Au colloids as the nanowire length is almost radiusindependent (see Supporting Information). The main difference between the two series of samples is a higher (several times) surface density of NPs self-assembled from thermally annealed Au thin films with respect to what is obtained from colloidal solutions (see Supporting Information). Hence, we assume that the Te diffusion flux received by a nanowire is inversely proportional to the Au NP density N. These considerations yield 2 with a certain pre-factor A. The almost jdiff (0) = A(πR / N)

a plausible estimate of λ=260 nm. Clearly, the longer length of nanowires grown from Au colloids and its increase with the growth temperature (Tg) are explained by (i) the smaller density of Au NPs that increases ν in Eq. (2) and (ii) the increase of ν with  due to the temperature-activated mechanism of Te adatom diffusion towards the nanowire bases. Considering the length histograms in Figure 5, we first note that all nanowire LDs are asymmetric (with a long tail on the left) and broader than for a Poissonian distribution (the Poissonian LDs would yield the variance   = <  > , Refs. 39,40). The broad LDs seen for our nanowire ensembles should not be due to the diffusion-induced spreading in the regimes where the axial growth rate ⁄∆ is proportional to L (Ref. 41) because our nanowires are fed from the substrate and their elongation is independent of L. Rather, the observed asymmetric broadening with its large representation of shorter nanowires is caused by difficult nucleation of the very first nanowire monolayers, as in Ref. 42 for Au-catalyzed InAs nanowires. Ignoring in the

linear increase of the mean nanowire length with the Te pre-deposition time ∆ for short (