Kinetics of Au-Ga droplet mediated decomposition of GaAs nanowires

Subscriber access provided by ALBRIGHT COLLEGE

Communication

Kinetics of Au-Ga droplet mediated decomposition of GaAs nanowires Marcus Ulf Tornberg, Daniel Jacobsson, Axel R. Persson, Lars Reine Wallenberg, Kimberly A. Dick, and Suneel Kodambaka Nano Lett., Just Accepted Manuscript • DOI: 10.1021/acs.nanolett.9b00321 • Publication Date (Web): 30 Apr 2019 Downloaded from http://pubs.acs.org on May 1, 2019

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 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 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.

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 23 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

Kinetics of Au-Ga droplet mediated decomposition of GaAs nanowires ∗,†

Marcus Tornberg,

‡,¶

Daniel Jacobsson,

†,‡,¶

Kimberly A. Dick,

†Solid ‡Centre ¶National

‡,¶

Axel R. Persson,

‡,¶

Reine Wallenberg,

§

and Suneel Kodambaka

State Physics, Lund University, Box 118, 22100, Lund, Sweden

for Analysis and Synthesis, Lund University, Box 124, 22100, Lund, Sweden

Center for High Resolution Microscopy, Lund University, Box 124, 22100, Lund, Sweden

§Department

of Materials Science and Engineering, University of California Los Angeles, 410 Westwood Plaza, Los Angeles, CA 90095

E-mail: [email protected]

Abstract Particle-assisted III-V semiconductor nanowire growth and applications thereof have been studied extensively. However, the stability of nanowire in contact with the particle as well as the particle chemical composition as a function of temperature remain largely unknown. In this work we use in-situ transmission electron microscopy to investigate the interface between a Au-Ga particle and the top facet of an h1 1 1i-oriented GaAs nanowire grown via vapor-liquid-solid process. We observed a thermally activated bilayer-by-bilayer removal of the GaAs facet in contact with the liquid particle/droplet during annealing between 300 and 420 ◦ C in vacuum. Interestingly, the GaAs removal rates initially depend on the thermal history of the sample and are time-invariant at later times. In-situ X-ray energy dispersive spectroscopy was also used to determine 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

that the Ga content in the particle at any given temperature remains constant over extended periods of time, and increases with increasing temperature from 300 ◦ C to 400 ◦ C. We attribute the observed phenomena to droplet-assisted decomposition of GaAs at a rate that is controlled by the amount of Ga in the droplet. We suggest that the observed transients in removal rates are a direct consequence of time-dependent changes in the Ga content. Our results provide new insights into the role of droplet composition on the thermal stability of GaAs nanowires and complement the existing knowledge on the factors inuencing nanowire growth. Moreover, understanding the nanowire stability and decomposition is important for improving processing protocols for the successful fabrication and sustained operation of nanowire based devices.

Keywords Nanowire, In-situ, Transmission Electron Microscopy, X-ray Energy Dispersive Spectroscopy, GaAs, Annealing Semiconductor nanowires have attracted considerable interest due to their geometry, 1 their ability to form metastable phases, 2 and the potential for multicomponent structures with tunable composition. 36 These attributes, together with the possibility of creating atomically sharp interfaces in heterostructures, 710 make nanowires suited for applications within electronics and optoelectronics. Most research eorts have focused on understanding and controlling the nanowire formation and fabrication of prototypical devices. However, relatively few studies have considered the stability of nanowires or the processes by which they decompose, both of which are important to understand in order to determining acceptable conditions for fabrication and operation of nanowire-based devices. Moreover, understanding the stability of the materials is critical to preserving the nanowire crystal structure 11,12 and composition during annealing, 13 fabrication and operation of devices. 14 Semiconductor nanowires are most commonly grown using the vapor-liquid-solid (VLS) process, often using Au-based liquid alloys. 15,16 It is therefore critical to understand how the 2


Page 2 of 23

Page 3 of 23 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

presence of Au impacts the stability of the grown material. For III-V semiconductors, previous studies have consistently shown that Au-particles alloyed with group III elements strongly aect the nanowire decomposition. 11,12,17 For example, Persson et al. 17 observed assisted decomposition of nanowires in the presence of Au-Ga alloy at temperatures between 350 and 600 ◦ C. 17 Recently, Fauske et al. 18 investigated the eects of annealing GaAs nanowires in contact with large Au reservoirs and found that the GaAs nanowire gets replaced with Au by a thermally activated process that depends on the relative amounts of Au and Ga. The thermal stability of single-crystalline GaAs surfaces has been extensively investigated and both Ga- and Au-droplets have been reported to lower the activation energy of evaporation from low index facets of GaAs and GaP. 1923 This particle-assisted decomposition happens at temperatures signicantly lower than the congruent evaporation temperature of approximately 625 ◦ C for GaAs. 24,25 A mechanistic understanding of the processes that occur at the particle-nanowire interface at elevated temperature is therefore needed for Au-seeded GaAs nanowires to be broadly useful. In particular, it is necessary to determine conditions in which the decomposition can be minimized, or ideally, even be controlled. In addition to informing protocols for successful device fabrication and operating conditions, knowledge gained from investigation of the thermochemical stability of nanowires is also likely to provide important insights into the nanowire crystal growth itself. Of particular interest are the particle's state, composition, and shape with respect to temperature and environment, such as partial and total pressure of gases, as these characteristics have been suggested to aect the resulting III-V nanowire crystal structure. 2629 Furthermore, the composition of the particle has been suggested to impact the growth direction 30,31 as well as the diameter of the resulting nanowire. 32,33 In addition, theoretical studies have predicted that the composition of the alloy inuences the nanowire growth kinetics. 3436 Despite the increased interest in engineering the morphology, composition and crystal structure of III-V semiconductor nanowires, direct investigations of the particle are rare. Recent in-situ experiments using transmission electron microscopy (TEM) have provided some insights into 3


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 eect of Au-alloy particle morphology on the III-V semiconductor crystal phase evolution. 11,12,27 However, to-date, there are no reports of direct measurements of the Au-alloy particle composition during the growth or decomposition of III-V nanowires, and theoretical growth models have been forced to rely on indirect or post-growth ex-situ estimates of composition. In this letter, we report on the kinetics of Au-Ga droplet mediated decomposition of h1 1 1ioriented GaAs nanowires at set temperatures between 300 and 420 ◦ C. Using an imagecorrected Hitachi HF-3300S environmental transmission electron microscope (ETEM) operated at 200 kV, we investigated the layer-by-layer removal of GaAs at its interface to the Au-Ga droplet (presented as supporting movie SM-I) as a function of time and temperature during annealing in vacuum. In addition, using X-ray energy dispersive spectroscopy (EDS), we measured the Au-Ga droplet composition as a function of temperature and correlate the compositional changes with the removal rate of material from the particle-nanowire interface. We present the removal rate of GaAs bilayers with time at a given temperature, both for the initial stage before the process attains a steady state and at the later stages where the layer removal rate has stabilized. The insights provided here are important not only for understanding the nanowire stability, but also provide a rst step towards understanding the role of Au-Ga-alloy and its temperature-dependent changes in particle composition during GaAs nanowire growth. In addition, the very regular layer-by-layer removal process may be interesting in itself as an 'etch-back' process for designing highly controlled nanowire interfaces. All experiments were carried out using pre-grown zincblende h1 1 1i-oriented GaAs nanowires mechanically transferred to TEM-compatible micro-electro-mechanical system (MEMS) heating devices with electron transparent SiNx windows. These devices allowed for resistive heating of the sample to study the stability of GaAs nanowires during annealing in vacuum, by controlling the resistance through a W-coil embedded in the surrounding SiNx . The details 4


Page 4 of 23

Page 5 of 23 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

Figure 1: Conventional TEM image of a typical GaAs nanowire (a) with an assisting AuGa particle on top acquired at 150 ◦ C along with selected area electron diraction patterns from a portion of the particle as highlighted by the circle (∼ 300 nm2 ) in (a) at 300 ◦ C (b) and 350 ◦ C (c). Reections from the particle are visible at 300 ◦ Cwhile they are absent at the higher temperature (c), indicating a change from a solid to an amorphous state with increasing temperature. Representative higher magnication conventional TEM images of the particle-nanowire interface obtained during annealing at: 320 ◦ C (d), 360 ◦ C (e), 400 ◦ C (f), and 420 ◦ C (g). Gallium content (cGa ) of the particle, extracted from in-situ energy dispersive spectroscopy measurements, is presented as a function of temperature (h) and shows a monotonic increase of Ga with temperature.The uncertainty interval indicated is the combined uncertainty between individual acquisitions and the uncertainty of quantication using set k-factors for Ga K-lines (kK,Ga = 1.603) and Au L-lines (kL,Au = 2.721).

5


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

of the nanowire growth procedure, crystal structure, MEMS device and the sample transfer are presented as supplementary information (SI-I). At 300 ◦ C and a base pole-piece pressure of 1e-5 Pa, the particle-nanowire system appears as presented in gure 1a, where a representative Au-GaAs nanowire has been imaged by conventional TEM. The state (solid or liquid) and composition of the Au-Ga particle were evaluated from selected area electron diraction (SAED) patterns and 60 s of EDS acquisition from the circled region of interest in gure 1a. The electron diraction patterns of the particle prior to and during annealing at temperatures up to 300 ◦ C, as shown for example in gure 1b, reveal reections indicative of a crystalline phase. Complementary EDS analysis of the particle prior to annealing revealed 8 ± 0.8 atomic percent (at.%) Ga in the particle. This low Ga concentration is a result of cooling the nanowires in an AsH3 /H2 environment after the metal-organic vapor phase epitaxy (MOVPE) growth. 9 An initial annealing of the nanowires at 350 ◦ C for two minutes resulted in the disappearance of a few spots in the diraction pattern as shown in gure 1c. We attribute the change in the diraction pattern to melting of the particle at the higher temperature. Note that the rings and reections visible in the pattern at 350 ◦ C stem from the supporting SiNx -membrane and the GaAs nanowires, respectively. After the initial annealing of the nanowires at 350 ◦ C, the temperature was lowered to 300 ◦

C, and subsequently increased to 420 ◦ C with increments of 20 ◦ C. At each temperature, the

particle phase, shape, and wetting characteristics were determined from a series of conventional TEM images (Audio Video Interleave [avi] format) within the rst two to four minutes from reaching the new temperature. During the two-hour experiment, the particle-nanowire system maintained a stable morphology, dierent for each temperature investigated, as shown by the representative conventional TEM images, gures 1d-g, acquired at 320 ◦ C, 360 ◦ C, 400 ◦

C, and 420 ◦ C. Interestingly, the particle shape appears to be smoothly curved, commonly

associated with a liquid state, even at temperatures as low as 320 ◦ C as shown in see gure 1d. The SAED pattern (not shown) of the particle at this temperature revealed diraction spots, indicative of solid phase(s). Based on these two observations, we suggest that this 6


Page 6 of 23

Page 7 of 23 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

particle consists of both solid and liquid phases at this temperature. From the binary Au-Ga and pseudo-binary Au0.98 As0.02 −Ga0.98 As0.02 phase diagram, 37,38 a liquid phase Au-Ga alloy containing approximately 33 at.% Ga can form at temperatures as low as 339 ◦ C. The formation of a liquid phase below the solidication temperature is plausible, for example due to thermal hysteresis, 39 as a result of the initial annealing at 350 ◦ C. At 420 ◦ C, we observe increased contrast across a segment of the nanowire at its interface to the particle, highlighted in gure 1g. This observation can be interpreted as either a partial wetting of the nanowire sidewalls or a truncation of the top facet. This change in wetting was associated with a dynamical change in particle shape and contact angle, a process which was observed during imaging and is provided as a supplementary movie (SM-II). For consistency and a quantitative description of our observations, we limit our analysis and discussion of the steady-state process in the following sections to the experimental data obtained at temperatures between 300 and 400 ◦ C as the particle-nanowire interface appears to be similar and comparable at these temperatures. In order to study the composition of the particle as a function of temperature, X-ray EDS was acquired for 60 s at each temperature after the particle-nanowire system had stabilized, which in our experiment occurred within six minutes at a new temperature. The chemical analysis did not reveal the presence of arsenic in the particle at any temperature between 300 and 400 ◦ C; however, small (< 2 at.%) amounts of arsenic, not distinguishable from the background signal of the EDS, are likely to be present in the Au-Ga alloy particles. 38 We observe Ga enrichment in the Au-Ga alloy with increasing temperature as shown in gure 1h. At lower temperatures, 300 and 320 ◦ C, the Ga concentration in the particle is nearly constant at 22.6 ± 0.9 at.%, corresponding to the solid Au7 Ga2 -phase that is expected to be stable in contact with GaAs. 40 Further increases of the set temperature resulted in an increased Ga concentration within the particle, up to 27.8 ± 1.0 at.% at 400 ◦ C for the temperature interval presented here in gure 1h. The gradual increase in Ga concentration in the particle at temperatures above 340 ◦ C can be related to the liquid state of the particle 7


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 its increasing solubility for Ga. While increased solubility of Ga in the liquid has been predicted by calculated phase diagrams, direct measurement of the concentration within the particles assisting nanowire growth has not previously been reported.

Figure 2: Real-time conventional TEM images (extracted from SM-I) acquired during annealing at 360 ◦ C in vacuum showing the layer-by-layer removal of GaAs at the particle-nanowire interface highlighted in the inset. Each of the parallel lines visible in the images corresponds to a bilayer of GaAs. The arrows indicate the position of a step formed by the partiallyremoved bilayer of GaAs. The extracted cumulative number of bilayers (N ) removed is graphically visualized (b) as a function of time (t) since the temperature was reached. Each set of symbols correspond to the data obtained at a given temperature. In the following sections, we focus on the dynamics of the Au-Ga/GaAs nanowire interfaces observed during annealing. In our experiments, in-situ TEM images reveal layer-by-layer removal of GaAs at the particle-nanowire interfaces at temperatures above 300 ◦ C. This process is presented in gure 2a, which shows lattice-resolved conventional TEM images of a portion of the Au-Ga/GaAs nanowire interface obtained at one-second intervals during annealing of the sample at 360 ◦ C. The images are part of a longer measurement sequence obtained at 360 ◦ C and is provided as a supplementary movie (SM-I). Through imaging of the nanowire close to the h1 1 2i zone axis, the Au-Ga alloy on the top of the nanowire appears dark while the GaAs nanowire appears brighter in comparison. The arrows in the images highlight the bilayer ledge and its propagation across the Au-Ga/GaAs interface from right to left. Typically, we observe the ledge to form at one of the two vapor-liquid-solid triple points visible in the TEM images. Occasionally, we also observe the appearance of 8


Page 8 of 23

Page 9 of 23 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

the ledge within the interface, presumably due to its formation at one of the triple points not visible in the TEM images. At this temperature, the formation and propagation of a ledge across the interface takes approximately seven seconds once the system has stabilized. We nd that this process of formation and propagation results in removal of GaAs bilayers, occurring (nearly) periodically at all temperatures above 300 ◦ C. While GaAs removal is observed at the particle/nanowire interface, the rest of the GaAs nanowire did not show loss of material during the TEM observation. In addition, the presented results did not show observable changes with changes of the electron dose, 5 electrons/Å2 s during imaging and 6-15 electrons/Å2 s during diraction, or as a result of intermittent sample exposure to the beam. Using our recorded images, we extracted the times at which consecutive GaAs layers have been removed in order to determine the GaAs removal rates at dierent temperatures. In gure 2b the number of bilayers removed (N ) is presented as a function of time (t) following a change in temperature. Each set of symbols in the graph corresponds to the data obtained at a specic temperature. The rate of GaAs bilayer removal, dened as the number of bilayers removed per unit time (dN/dt), is initially higher within the rst 120 s and decreases gradually to a stable rate. As we show later, the exact time taken to reach this quasi steadystate rate decreases with increasing temperature. The temperature dependence of the GaAs bilayer removal process was evaluated based on the quasi steady-state removal rates and is presented as an Arrhenius plot in gure 3. We nd that dN/dt increases exponentially with increasing temperature (T ). Assuming that the process is thermally activated, the rate of removal of GaAs bilayers can be expressed using the activation energy (∆E ), the molar gas constant (R), a pre-exponential factor (ν ) and the number of atoms in an bilayer (N0 ) as follows,

dN/dt = νN0 e−∆E/RT .

9


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

Figure 3: The GaAs bilayer removal rate (dN/dt) is presented as a function of the inverse temperature (1/T ). Solid circles are the dN/dt values are extracted from the data in gure 2b and the dashed line is the linear least square t to the data for temperatures above 340 ◦ C, where the removal occurs in the presence of a liquid particle. The linear regression provides the activation energy (∆E ) associated with the bilayer removal process. From the data presented in gure 3, using a linear least-squares t for the temperatures above 340◦ C, we obtain an activation energy of 116 ± 4 kJ/mol (1.21 ± 0.04 eV). The corresponding pre-factor is obtained from the t as 7.6e12 s−1 with a deviation within the interval [3.8e12, 1.5e13] s−1 . The kinetics at low temperature (below 340◦ C) was excluded from this analysis as they appeared to be both solid and liquid. Details of the calculation are provided as supporting information, SI-II. The extracted activation energy is comparable to previously reported value of 124 kJ/mol associated with the dissolution of GaAs nanowire in contact with a large reservoir of Au. 18 While our observations of GaAs decomposition at the Au-Ga alloy interface are similar to those reported in the literature, the striking dierence between our data and the previous reports is our observation of the Au-Ga alloy reaching a stable composition over time and its inuence on GaAs decomposition rate. Earlier studies showed that the rates of decomposition of GaAs nanowires and single crystal GaAs surfaces depend on the Au-Ga alloy composition and that the decomposition stops when the alloy composition reaches the thermodynamic equilibrium value at a given temperature. 18,24 In our experiments, however, we nd that both the GaAs removal rate and the Ga concentration in the Au-Ga particles vary little with time, after the initial changes immediately following the 10


Page 10 of 23

Page 11 of 23 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

change in temperature, at all temperatures above 340 ◦ C, (see supporting information (SI-III) for more details). That is, both Ga and As must leave the particle-nanowire system at the same rate as they are produced from the decomposition of GaAs (note that Ga remaining in the particle would cause its volume to increase with time). At 400 ◦ C, where the removal rate of GaAs is the highest in the presence of a stable particle, we do not observe any signicant changes in the particle shape or size, suggesting that the volume changes associated with Ga incorporation are negligible. Detailed analysis of the particle shape and composition over time is provided as supporting information (SI-III). In order to understand the chemical composition of Au-Ga droplets and the kinetics of GaAs removal, we monitored the rate of GaAs bilayer removal at times immediately following a heating or cooling of the sample. The time needed to remove a layer (∆tlayer ) as the experiment progresses is shown in gure 4a, for data obtained immediately after increasing the temperature by 20 ◦ C. It is possible to see uctuations in ∆tlayer as a function of time since the new set temperature was reached. Despite these rate uctuations, we nd that the removal rate of a bilayer increases gradually from the time of reaching the new temperature, eventually saturating at the steady-state removal rate. The time needed for rate saturation (tsaturation ) decreases from approximately 140 s when reaching 340 ◦ C to less than 30 s when the sample was heated to 380 ◦ C. At temperatures above 380 ◦ C, any changes in ∆tlayer are too small and could not be resolved in our experiments. In a complementary set of experiments, we monitored the GaAs removal rates after cooling to a desired temperature (instead of heating). The measured ∆tlayer data as a function of time are shown in gure 4b, obtained after lowering the temperature (i) from 400 ◦ C to 380 ◦

C (open triangles) and (ii) from 420 ◦ C to 380 ◦ C (solid circles). In both cases, prior to

lowering the temperature, the bilayer removal times were constant at 3 and 2 s per layer at 400 and 420 ◦ C, respectively. Upon cooling to 380 ◦ C, ∆tlayer temporarily increases to a higher value than the steady-state time of the same temperature. Interestingly the time 11


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

Figure 4: Transient changes in the GaAs removal rates following a temperature change. Here, we presented the time needed to remove a GaAs bilayer (∆tlayer = 1/(dN/dt)) as a function of time after increasing (a) or decreasing (b) the temperature. The data set presented in (a) was obtained after increasing the temperature in increments of 20 ◦ C from 320 ◦ C to 340 ◦ C (2), from 340 ◦ C to 360 ◦ C (#) and from 360 ◦ C to 380◦ C (3). For the data presented in (b), the particle-nanowire sample was cooled, following the the inset of temperature over time, from 400 ◦ C to 380 ◦ C (4) and from 420 ◦ C to 380◦ C (#). In both panels, the time is the relative time from when sample temperature reached the desired value upon heating/cooling. temperature was obtained. At times immediately following the increase (decrease) in temperature, ∆tlayer is lower (higher), in other words faster (slower) removal rate compared to removal rates later times.

12


Page 12 of 23

Page 13 of 23 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

required to reach this steady-state value, henceforth referred to as tdepletion as the Ga content decreases with temperature, upon cooling is constant at ∼45 s at 380 ◦ C, independent of initial temperature. In comparison, the transient time (tsaturation ) for reaching steady-state after increasing the temperature to 380 ◦ C was ∼25 s. To explain our observations of GaAs removal rates dependent on both the temperature and the thermal history of the sample together with seemingly time-independent Ga content of the Au-Ga-alloy at a xed temperature, we propose (and justify) the following mechanism. The removal of GaAs bilayers at the Au-Ga/GaAs interfaces occurs via decomposition of GaAs to Ga and As, a rate that depends on the particle composition, followed by desorption of the adspecies. Direct evaporation of GaAs into the vacuum is not considered as it is expected to occur at temperatures well above those used during our experiment. 24 Upon decomposition of GaAs to Ga and As, arsenic desorbs readily due to its high vapor pressure (up to 100 Pa) and low solubility (below 2 at. %) in the alloy particles at the temperatures used in the experiments, while Ga is less likely to evaporate at the same rate as As. 41 As a result, we assume that the rate of evaporation of Ga from the Au-Ga particles into vacuum is negligible for this experiment. We hypothesize that Ga is removed from the GaAs nanowire by (I) rst dissolving in the particle followed by (II) diusion along the nanowire sidewall. We expect that the activation barriers associated with As desorption, Ga diusion (both through the Au-Ga alloy and along the nanowire), and dissolution into the alloy are all relatively small and hence these processes occur readily. This out-diusion of Ga was not observed to result in accumulation of Ga on the SiNx lm and the lm most likely contributed to transporting the Ga away from the nanowires. Based on our nding of slower Ga depletion of the particle in comparison to the saturation (for example at 380 ◦ C as shown in gure 4); we suggest that the out-ux of Ga from the droplet (step II) is the rate limiting process for the steady-state particle-assisted GaAs decomposition, as follows,

dN/dt ∝ k1 [µAu−Ga − µGa,As ], 13


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

where k1 is a thermally-activated eective rate constant and µAu−Ga and µGa,As are the steady-state chemical potentials of Ga in Au-Ga alloy particle and in the ambient, respectively. This is relation holds under the assumption that the chemical potential of the solid GaAs is constant. For the steady-state case, the chemical potentials are time-independent and hence results in a constant rate of removal. However, as has been shown in this letter, dN/dt is not constant, especially not at initial times after changing the temperature;

tsaturation is larger than tdepletion when increasing or decreasing the temperature to 380 ◦ C as shown in gure 4. We interpret this behavior as a consequence of the time-dependent Ga accumulation (depletion), and thus a time-dependent µ∗Au−Ga , until tsaturation (tdepletion ) when the droplet attains a steady-state composition and chemical potential. The transient removal rate can therefore be expressed as,

dN/dt ∝ µAu−Ga − µ∗Au−Ga (cGa ). In our experiments, steady-state removal rates are observed two minutes after the temperature has been reached for temperatures above 340 ◦ C, based upon which we suggest that the Ga concentration in the particles reaches the steady-state value within this time. The controlled bilayer removal of GaAs driven by the chemical potentials of the included components, as demonstrated here, shares some important similarities with epitaxial nanowire growth. In both cases the process happens for one layer at a time, with rates that are determined by, temperatures, gas- and particle-compositions, 3436 yielding highly controlled growth 42,43 or, as presented here, removal of crystalline material. Previous in-situ TEM investigations have shown that extremely precise layer-by-layer control of the growth is possible under the right conditions. 44 The results here show that similar precision is possible for removal of layers, which indicates the possibility of performing both process in the same system to for example limit or reduce the axial length of a complex nanostructure. In conclusion, we investigated layer-by-layer removal of GaAs from the Au-Ga/GaAs inter14


Page 14 of 23

Page 15 of 23 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

faces formed after Au-Ga-assisted growth of GaAs nanowires via vapor-liquid-solid process. Using in-situ TEM and EDS, we measured the time- and temperature-dependent rates of GaAs removal and the composition of Au-Ga particles on top of GaAs {1 1 1}-type facets at temperatures up to 420 ◦ C. We observed a liquid-phase Au-Ga alloy at temperatures as low as 340 ◦ C, whose Ga content increases with increasing temperature, from the solid 22.6 at. % below 340 ◦ C to 27.8 at. % at 400 ◦ C. We show that GaAs removal rates depend on the thermal history of the sample; at any given temperature, the rates vary initially and become constant at later times. We nd that the steady-state removal rates increase exponentially with temperature. Based on all of these observations, we suggest that the removal of GaAs bilayers at the Au-Ga/GaAs interfaces occurs via decomposition of GaAs and that the rate is determined by the time-dependent depletion of Ga accumulated in the Au-Ga droplet. Our results provide direct evidence for the inuence of particle composition on the stability of nanowires, suggesting that compositional tuning may be a means to control this stability. In addition, our in-situ observations of the thermal stability of the metal-nanowire interface provide essential knowledge of operation lifetime for nanowire-based devices. Moreover, the controlled bilayer removal of III-V semiconductor material may provide an additional method of investigating epitaxial semiconductor growth due to their similar mechanisms for formation and removal.

Methods The experiments were conducted using GaAs nanowires mechanically transferred, using a dry lint-free cleanroom wiper, to MEMS devices (Norcada Inc), designed for in-situ heating, with transparent SiNx windows for TEM compatibility. The zincblende h1 1 1i-oriented GaAs nanowires were grown using metal-organic vapor phase epitaxy by supplying tri-methyl gallium and arsine for 5 minutes at 475◦ C at molar fractions of 2.53e-5 and 1.88e-3, respectively. Further details of the growth are presented as supporting information (SI-I). The

15


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

data presented here were obtained using an image corrected (B-Cor) Hitachi HF3300S TEM equipped with a cold eld emission gun operated at 200 keV with an emission current of 6-10 µA. The transferred nanowires were initially heated to a nominally low temperature of 200 ◦ C and held at this temperature until the base pressure in the TEM column was below 5e-5 Pa. The purpose of this annealing was to degas the TEM holder-sample assembly and minimize residual gases in the TEM column and hence electron beam induced deposition of contaminants on the sample. While degassing at this temperature, the electron beam was aligned, after which the sample was rapidly (2 ◦ C/s) heated to 300 ◦ C. At the start of each experiment, following the degassing and alignment of the microscope, additional annealing of the nanowires was conducted at 350 ◦ C to assist the removal of physisorbed species on the nanowire as well as for evaluation of the droplet shape and possible eects of SiNx . After these two annealing steps, the temperature was lowered to 300 ◦ C and kept for 90 minutes, the in-situ observations of particle/nanowire interfacial dynamics were monitored at regular temperature intervals while heating the sample to 420 ◦ C in steps of 20 ◦ C at a rate of 2 ◦

C/s. The temperature curve for the experiment is presented in gure S2. At each temper-

ature, chemical composition of the Au-Ga particle and the GaAs nanowire etching behavior were determined as a function of time. Chemical characterization of the particle was performed by condensing the beam on the particle away from its interface to the nanowire, at each temperature using EDS typically six minutes after reaching the set temperature, by which time the particle-nanowire system was at a steady-state. The X-ray signal was collected for 60 s using a silicon drift detector (X-MaxN 80T, Oxford Instruments) protected by a Moxtek AP window and quantied using virtual standards based on kα,Si -factors and the Cli-Lorimer thin-foil approach using the K- and L-lines of Ga (kK,Ga = 1.603) and Au (kL,Au = 2.721) respectively. 45 In this letter, the uncertainty interval presented for the X-ray EDS measurement is the combined experimental measurement uncertainty and the quantication uncertainty provided by the internal quantication using the Aztec software (Oxford Instruments). Further details of the estimated errors are presented as supplemen16


Page 16 of 23

Page 17 of 23 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

tary information (SI-I). In total, seven nanowires were investigated during the experiments to validate the trends shown, the quantitative data presented here are observations of three separate nanowires. The temporal changes in Au-Ga/GaAs nanowire interfaces were monitored in conventional mode and the TEM images, acquired by a GATAN Orius SC1000B CCD camera at 1-4 frames/s, were saved as a sequence of images. Particle shapes and decomposition rates were extracted from the TEM images using ImageJ. 46 Presented videos have been compressed (Xvid format) using the open source software VirtualDub. 47

Acknowledgments The authors acknowledge nancial support from the Knut and Alice Wallenberg Foundation (KAW) and NanoLund.

Supporting Information Available Supporting information is available for; videos of continuous removal of GaAs at the droplet interface at (SM-I) 360 ◦ C and (SM-II) 420 ◦ C, (SI-I) details of the pre-grown GaAs nanowires and the chemical analysis, a typical experimental set temperature curve over time and possible eect of the supporting SiNX membrane on particle-nanowire stability, (SI-II) analysis of the GaAs bilayer removal rates and extraction of pre-factor, and (SI-III) analysis of particle shape and composition data over time.

References (1) Glas, F. Critical Dimensions for the Plastic Relaxation of Strained Axial Heterostructures in Free-Standing Nanowires. Phys. Rev. B - Condens. Matter Mater. Phys. 2006, 74,

25.

17


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

(2) Johansson, J.; Zanolli, Z.; Dick, K. A. Polytype Attainability in III-V Semiconductor Nanowires. Cryst. Growth Des. 2016, 16, 371379. (3) Wu, Z. H.; Sun, M.; Mei, X. Y.; Ruda, H. E. Growth and Photoluminescence Characteristics of AlGaAs Nanowires. Appl. Phys. Lett. 2004, 85, 657659. (4) Lim, S. K.; Tambe, M. J.; Brewster, M. M.; GradecÌak, S. Controlled Growth of Ternary Alloy Nanowires Using Metalorganic Chemical Vapor Deposition. Nano Lett.

2008, 8, 13861392. (5) Jung, C. S.; Kim, H. S.; Jung, G. B.; Gong, K. J.; Cho, Y. J.; Jang, S. Y.; Kim, C. H.; Lee, C. W.; Park, J. Composition and Phase Tuned InGaAs Alloy Nanowires. J. Phys. Chem. C

2011, 115, 78437850.

(6) Berg, A.; Heurlin, M.; Tsopanidis, S.; Pistol, M.-E.; Borgström, M. T. Growth of Wurtzite Al(x)Ga(1-x)P Nanowire Shells and Characterization by Raman Spectroscopy. Nanotechnology

2017, 28, 035706.

(7) Björk, M. T.; Ohlsson, B. J.; Sass, T.; Persson, A. I.; Thelander, C.; Magnusson, M. H.; Deppert, K.; Wallenberg, L. R.; Samuelson, L. One-dimensional Heterostructures in Semiconductor Nanowhiskers. Appl. Phys. Lett. 2002, 80, 10581060. (8) Larsson, M. W.; Wagner, J. B.; Wallin, M.; Håkansson, P.; Fröberg, L. E.; Samuelson, L.; Wallenberg, L. R. Strain Mapping in Free-Standing Heterostructured Wurtzite InAs/InP Nanowires. Nanotechnology 2007, 18, 015504. (9) Lehmann, S.; Wallentin, J.; Jacobsson, D.; Deppert, K.; Dick, K. A. A General Approach for Sharp Crystal Phase Switching in InAs, GaAs, InP, and GaP Nanowires Using Only Group V Flow. Nano Lett. 2013, 13, 40994105. (10) Assali, S.; Gagliano, L.; Oliveira, D. S.; Verheijen, M. A.; Plissard, S. R.; Feiner, L. F.;

18


Page 18 of 23

Page 19 of 23 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

Bakkers, E. P. A. M. Exploring Crystal Phase Switching in GaP Nanowires. Nano Lett.

2015, 15, 80628069. (11) Zheng, H.; Wang, J.; Huang, J. Y.; Wang, J.; Zhang, Z.; Mao, S. X. Dynamic Process of Phase Transition from Wurtzite to Zinc Blende Structure in InAs Nanowires. Nano Lett.

2013, 13, 60236027.

(12) Zhang, Z.; Liu, N.; Li, L.; Su, J.; Chen, P.-p.; Lu, W.; Gao, Y.; Zou, J. In Situ TEM Observation of Crystal Structure Transformation in InAs Nanowires on Atomic Scale. Nano Lett.

2018, 18, 65976603.

(13) Tornberg, M.; Dick, K. A.; Lehmann, S. Thermodynamic Stability of Gold-Assisted InAs Nanowire Growth. J. Phys. Chem. C 2017, 121, 2167821684. (14) Orrù, M.; Rubini, S.; Roddaro, S. Formation of Axial Metal-Semiconductor Junctions in GaAs Nanowires by Thermal Annealing. Semicond. Sci. Technol. 2014, 29, 054001. (15) Yazawa, M.; Koguchi, M.; Muto, A.; Ozawa, M.; Hiruma, K. Eect of one monolayer of surface gold atoms on the epitaxial growth of InAs nanowhiskers. Appl. Phys. Lett.

1992, 61, 20512053. (16) Dick, K. A.; Deppert, K.; Karlsson, L. S.; Wallenberg, L. R.; Samuelson, L.; Seifert, W. Role of the Au/III-V Interaction in the Au-assisted Growth of III-V Branched Nanostructures. Conf. Proc. - Int. Conf. Indium Phosphide Relat. Mater. 2005, 2005, 487 490. (17) Persson, A. I.; Larsson, M. W.; Stenstrom, S.; Ohlsson, B. J.; Samuelson, L.; Wallenberg, L. R. Solid-Phase Diusion Mechanism for GaAs Nanowire Growth. Nat Mater

2004, 3, 677681. (18) Fauske, V. T.; Huh, J.; Divitini, G.; Dheeraj, D. L.; Munshi, A. M.; Ducati, C.; We-

19


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

man, H.; Fimland, B. O.; Van Helvoort, A. T. J. In Situ Heat-induced Replacement of GaAs Nanowires by Au. Nano Lett. 2016, 16, 3051. (19) Lou, C. Y.; Somorjai, G. A. Studies of the Vaporization Mechanism of GaAs Single Crystals. J. Chem. Phys. 1971, 55, 3624. (20) Bauer, C. L. Investigation of Interfacial Reactions Between Thin Films of Gold and Substrates of Gallium Arsenide by Transmission Electron Microscopy. Surf. Sci. 1986, 168,

395403.

(21) Holloway, P. H.; Mueller, C. H. Chemical Reactions at Metal/Compound Semiconductor Interfaces: Au and GaAs. Thin Solid Films 1992, 221, 254261. (22) Chatillon, C.; Chatain, D. Congruent Vaporization of GaAs (s) and Stability of Ga (l) Droplets at the GaAs (s) Surface. J. Cryst. Growth 1995, 151, 91. (23) Heyn, C.; Jesson, D. E. Congruent evaporation temperature of molecular beam epitaxy grown GaAs (001) determined by local droplet etching. Appl. Phys. Lett. 2015, 107, 161601. (24) Kinsbron, E.; Gallagher, P. K.; English, A. T. Dissociation of GaAs and Ga0.7Al0.3As During Alloying of Gold Contact Films. Solid. State. Electron. 1979, 22, 517. (25) Goldstein, B.; Szostak, D. J.; Ban, V. S. Langmuir Evaporation from the (100), (111A), and (111B) Faces of GaAs. Surf. Sci. 1976, 57, 733740. (26) Wen, C. Y.; Reuter, M. C.; Terso, J.; Stach, E. A.; Ross, F. M. Structure, Growth Kinetics, and Ledge Flow During Vapor-solid-solid Growth of Copper-catalyzed Silicon Nanowires. Nano Lett. 2010, 10, 514519. (27) Jacobsson, D.; Panciera, F.; Terso, J.; Reuter, M. C.; Lehmann, S.; Hofmann, S.; Dick, K. A.; Ross, F. M. Interface Dynamics and Crystal Phase Switching in GaAs Nanowires. Nature 2016, 531, 317. 20


Page 20 of 23

Page 21 of 23 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

(28) Joyce, H. J.; Wong-Leung, J.; Gao, Q.; Tan, H. H.; Jagadish, C. Phase Perfection in Zinc Blende and Wurtzite III-V Nanowires Using Basic Growth Parameters. Nano Lett.

2010, 10, 908. (29) Lin, P. A.; Liang, D.; Reeves, S.; Gao, X. P. A.; Sankaran, R. M. Shape-controlled Au Particles for InAs Nanowire Growth. Nano Lett. 2012, 12, 315320. (30) Wang, J.; Plissard, S. R.; Verheijen, M. A.; Feiner, L. F.; Cavalli, A.; Bakkers, E. P. A. M. Reversible Switching of InP Nanowire Growth Direction by Catalyst Engineering. Nano Lett.

2013, 13, 3802.

(31) Kelrich, A.; Sorias, O.; Calahorra, Y.; Kaumann, Y.; Gladstone, R.; Cohen, S.; Orenstein, M.; Ritter, D. InP Nanoag Growth from a Nanowire Template by in Situ Catalyst Manipulation. Nano Lett. 2016, 16, 28372844. (32) Crawford, S.; Lim, S. K.; Grade£ak, S. Fundamental insights into nanowire diameter modulation and the liquid/solid interface. Nano Lett. 2013, 13, 226232. (33) Crawford, S. C.; Ermez, S.; Haberfehlner, G.; Jones, E. J.; Grade£ak, S. Impact of nucleation conditions on diameter modulation of GaAs nanowires. Nanotechnology 2015, 26,

225604.

(34) Glas, F. Chemical potentials for Au-assisted vapor-liquid-solid growth of III-V nanowires. J. Appl. Phys. 2010, 108, 073506. (35) Dubrovskii, V. G. Inuence of the Group V Element on the Chemical Potential and Crystal Structure of Au-catalyzed III-V Nanowires. Appl. Phys. Lett. 2014, 104, 053110. (36) Grecenkov, J.; Dubrovskii, V. G.; Ghasemi, M.; Johansson, J. Quaternary Chemical Potentials for Gold-Catalyzed Growth of Ternary InGaAs Nanowires. Cryst. Growth Des.

2016, 16, 45294530. 21


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

(37) Elliott, R. P.; Shunk, F. A. The Au-Ga (Gold-Gallium) system. Bull. Alloy Phase Diagrams

1981, 2, 356358.

(38) Ghasemi, M.; Johansson, J. Phase Diagrams for Understanding Gold-Seeded Growth of GaAs and InAs Nanowires. J. Phys. D. Appl. Phys. 2017, 50, 134002. (39) Kodambaka, S.; Terso, J.; Reuter, M. C.; Ross, F. M. Germanium Nanowire Growth Below the Eutectic Temperature. Science (80-. ). 2007, 316, 729732. (40) Prince, A.A.; Raynor, G.V.; Evans, D.S. Phase Diagrams of Ternary Gold Alloys, Inst. of Materials,

London, 123, 1990.

(41) Alcock, C.B.; Itkin, V.P.; and Horrigan, M.K. Vapour Pressure Equations for the Metallic Elements: 298-2500K, Canadian Metallurgical Quarterly, 1984, 23 309. (42) Harmand, J. C.; Tchernycheva, M.; Patriarche, G.; Travers, L.; Glas, F.; Cirlin, G. GaAs Nanowires Formed by Au-assisted Molecular Beam Epitaxy: Eect of Growth Temperature. J. Cryst. Growth 2007, 301-302, 853. (43) Fontcuberta i Morral, A.; Colombo, C.; Abstreiter, G.; Arbiol, J.; Morante, J. R. Nucleation mechanism of gallium-assisted molecular beam epitaxy growth of gallium arsenide nanowires. Appl. Phys. Lett. 2008, 92, 063112. (44) Chou, Y.-C.; Hillerich, K.; Terso, J.; Reuter, M. C.; Dick, K. a.; Ross, F. M. AtomicScale Variability and Control of III-V Nanowire Growth Kinetics. Science 2014, 343, 2814. (45) Cli, G. ;Lorimer, G.W. (1975) The quantitative analysis of thin specimens. J. Microsc.

1975, 103, 203-207. (46) https://imagej.nih.gov/ij/index.html. (47) http://www.virtualdub.org. 22


Page 22 of 23

Page 23 of 23 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

Graphical TOC Entry k2(T, y) AuxGay

k1(T, y)

k’2

k’1

t + ∆t

GaAs

23


Kinetics of Au-Ga droplet mediated decomposition of GaAs nanowires

Recommend Documents