Home
Search
Collections
Journals
About
Contact us
My IOPscience
In situ etching for control over axial and radial III-V nanowire growth rates using HBr
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2014 Nanotechnology 25 505601 (http://iopscience.iop.org/0957-4484/25/50/505601) View the table of contents for this issue, or go to the journal homepage for more Download details: IP Address: 132.239.1.231 This content was downloaded on 15/06/2017 at 15:43 Please note that terms and conditions apply.
You may also be interested in: Control of morphology and crystal purity of InP nanowires by variation of phosphine flux during selective area MOMBE A Kelrich, V G Dubrovskii, Y Calahorra et al. InxGa1xAs nanowires with uniform composition, pure wurtzite crystal phase and taper-free morphology Amira S Ameruddin, H Aruni Fonseka, Philippe Caroff et al. Particle-assisted GaxIn1-xP nanowire growth for designed bandgap structures D Jacobsson, J M Persson, D Kriegner et al. Sn-seeded GaAs nanowires grown by MOVPE Rong Sun, Neimantas Vainorius, Daniel Jacobsson et al. Synthesis and properties of ultra-long InP nanowires on glass Veer Dhaka, Ville Pale, Vladislav Khayrudinov et al. Growth parameter design for homogeneous material composition in ternary GaxIn1xP nanowires Alexander Berg, Filip Lenrick, Neimantas Vainorius et al. Direct nucleation, morphology and compositional tuning of InAs1xSbx nanowires on InAs (111) B substrates Luna Namazi, Sepideh Gorji Ghalamestani, Sebastian Lehmann et al. Recent progress in metal-organic chemical vapor deposition of (0001) N-polar group-III nitrides Stacia Keller, Haoran Li, Matthew Laurent et al.
Nanotechnology Nanotechnology 25 (2014) 505601 (9pp)
doi:10.1088/0957-4484/25/50/505601
In situ etching for control over axial and radial III-V nanowire growth rates using HBr Alexander Berg1, Kilian Mergenthaler1, Martin Ek2, Mats-Erik Pistol1, L Reine Wallenberg2 and Magnus T Borgström1 1 2
Solid State Physics, Lund University, Box 118, S-221 00, Lund, Sweden Polymer and Materials Chemistry/nCHREM, Lund University, Box 124, S-221 00 Lund, Sweden
E-mail: [email protected] Received 2 August 2014, revised 6 October 2014 Accepted for publication 10 October 2014 Published 25 November 2014 Abstract
We report on the influence of hydrogen bromide (HBr) in situ etching on the growth of InP, GaP and GaAs nanowires. We find that HBr can be used to impede undesired radial growth during axial growth for all three material systems. The use of HBr opens a window for optimizing the growth parameters with respect to the materials’ quality rather than only their morphology. Transmission electron microscopy (TEM) characterization reveals a partial transition from a wurtzite crystal structure to a zincblende upon the use of HBr during growth. For InP, defect-related luminescence due to parasitic radial growth is removed by use of HBr. For GaP, the etching with HBr reduced the defect-related luminescence, but no change in peak emission energy was observed. For GaAs, the HBr etching resulted in a shift to lower photon emission energies due to a shift in the crystal structure, which reduced the wurtzite segments. Keywords: nanowire, HBr, in situ etching, MOVPE (Some figures may appear in colour only in the online journal)
was reported [14, 15], opening the field for optimizing NWs with respect to the material’s quality rather than compromising the growth parameters to optimize the morphological properties. In addition to InP, this method has been successfully applied to GaP [16], GaInP [17] and InAsP [18] NWs. However, until now, HCl has not been reported to eliminate the tapering of GaAs NWs. Although our group has tried to grow taper-free GaAs NWs, we were not able to remove the radial growth by HCl [19]. Here, we report on the effects of in situ etching by use of hydrogen bromide (HBr) on InP, GaP and GaAs NWs. HBr was chosen as an alternative halide to HCl, as it is known to be a versatile etchant of III-V materials [20–24]. We show that in situ etching by use of HBr can prevent radial growth of InP, GaP and GaAs NWs, despite growth conditions that otherwise lead to strong tapering. This allows control over the design and growth of axially defined NW materials. However, the use of HBr for impeding the radial growth on GaAs NWs is limited to a smaller parameter space than that for phosphides.
1. Introduction Nanowires (NWs) have recently attracted attention as one of the most promising ways to combine high performance III-V materials with new functionality [1–4] with highly mature Si technology [5, 6]. For device applications, however, uncontrolled parasitic radial growth on the NW side’s facets during the remainder of the synthesis typically occurs. Such unintentionally grown NW shells can lead to short circuiting of the axially designed components and can become a surrounding and competing recombination centre for charge carriers [7]. In order to affect the ratios between the axial and radial NW growth, parameters such as the growth temperature [8–10] and catalyst supersaturation [11, 12] can be used. For thin film growth, however, parameter tuning, such as the tuning used to favour axial NW growth, typically leads to poor crystalline quality [13] and inhibits the use of growth parameter variations for the optimization of crystal quality and doping. Recently, a method using in situ etching by use of HCl in order to take control over axial and radial growth rates 0957-4484/14/505601+09$33.00
1
© 2014 IOP Publishing Ltd Printed in the UK
A Berg et al
Nanotechnology 25 (2014) 505601
2. Experimental
recorded on the [110] zone axis in sequence along the length of a few NWs. Photoluminescence (PL) measurements were performed at a temperature of about 5 K. The NWs were broken off from native substrates and were spread out on a gold-covered silicon surface. The NW under study was evenly illuminated throughout its length using a CW laser emitting at 532 nm for the InP and GaAs and a 325 nm laser for the GaP. The PL was collected by an optical microscope, dispersed through a spectrometer and detected by a thermoelectrically cooled CCD camera. At least 5 NWs per sample were studied.
The samples were prepared by depositing 40 nm diameter Au particles on (1¯ 1¯ 1¯)B InP, (1¯ 1¯ 1¯)B GaP and (1¯ 1¯ 1¯)B GaAs substrates via an aerosol technique [25], resulting in randomly distributed monodisperse particles with a homogeneous density of 1 × 106 cm−2. The particle-assisted NW growth was performed by use of a low-pressure (100 mbar) metal organic vapour phase epitaxy (MOVPE) system. Trimethylindium (TMI), trimethylgallium (TMG), phosphine (PH3) and arsine (AsH3) were used as precursors, and hydrogen bromide (HBr) was used as an etching agent in a total flow of 6.0 l min−1 using hydrogen (H2) as the carrier gas. For the InP NWs, the PH3 and TMI molar fractions (χPH3, χTMI) were set to χPH3 = 6.25 × 10−3 and χTMI = 3.5 × 10−6. For the GaP, the TMG molar fraction (χTMG) was set to 37.8 × 10−6, and the χPH3 was set to 6.25 × 10−3. For InP and GaP, prior to growth, the samples were heated to 600 °C under a PH3/H2 gas mixture for 10 min in order to desorb any surface oxides. Then, the reactor temperature was set to a growth temperature of 450 °C (InP) or 600 °C (GaP) for a direct comparison of the NW growth in the presence of HBr with the results of in situ removal of tapering by use of HCl for InP [14] and GaP [16], respectively. At these elevated temperatures, strong tapering is expected to occur [9, 26]. GaP has a larger parameter space for growth with respect to temperature compared to InP [26]; therefore, a temperature-dependent study from 440 °C to 600 °C in steps of 20 °C while keeping a constant HBr molar fraction (χHBr) of 72 × 10−6 was carried out. For GaAs, a growth temperature of 470 °C was chosen for the evaluation. At this temperature, GaAs NWs grow under optimal conditions with respect to the material’s use but are tapered [8]. The samples were annealed at 650 °C for 10 min under an AsH3/ H2 gas mixture. χTMG was set to 6.4 × 10−6, and χAsH3 was set to 1.1 × 10−4. In order to assess the growth dynamics of GaAs NWs, an χAsH3 variation of 0.55 × 10−4 < χAsH3 < 8.8 × 10−4 was carried out using χHBr = 0 and 4.9 × 10−6. The NW growth was initiated by adding the group III precursor to the flow. HBr was added to the flow after a 15 s nucleation step for InP and a 30 s nucleation step for GaAs. For GaP, the HBr etchant was introduced into the reactor simultaneously as TMG. χHBr was varied between 0 and 140 × 10−6 for growth of the NWs. The growth was terminated after 10 min and 15 s for the InP after 3 min for the GaP and after 10 min and 30 s for the GaAs by switching off group III and the HBr simultaneously, after which the samples were cooled down to room temperature under a mixture of the respective group V precursor and H2. Scanning electron microscopy (SEM) was used to characterize the NW morphology. At least 10 NWs at each of the three different locations on the samples were evaluated with respect to length and tapering. The samples were prepared for high resolution transmission electron microscopy (HRTEM) by direct transfer of the NWs to a Cu grid with a lacey carbon film by gently pressing the grid onto the substrate. For all three material systems, two samples, each with different χHBr, were characterized by HRTEM. For each sample, the images were
3. Results and discussion We first discuss the general trends that are valid for all three materials systems. Figures 1–3 show SEM images of the InP, GaP and GaAs NWs, respectively, as well as the effect on the NW morphology of using different χHBr during growth. Figure 1(a) shows a reference InP NW grown intentionally with parameters that give strong tapering when HBr is not present. The rough InP (1¯ 1¯ 1¯)B and GaP (1¯ 1¯ 1¯)B substrate surfaces, seen in figure 2(a), indicate competing InP and GaP surface growth. By increasing χHBr, the NW tapering gradually decreases, and the InP and GaP substrates’ roughness diminishes (figures 1(b), (c) and 2(b), (c)). The GaAs (1¯ 1¯ 1¯)B (figure 3(a)) substrate does not show a clear roughness before in situ etching; consequently, no strong effect of the etching on the substrate growth is observed (figures 3(b), (c)). The tapering is reduced by increasing χHBr for all three materials; thus, HBr is more versatile than HCl in the use of in situ etching of NWs in order to take full control over axial and radial growth rates. To assess the effect of using χHBr during NW growth, we evaluated the NW volume, NW length and the ratio between the bottom and top NW diameter (Db/Dt) as a function of χHBr by SEM after growth. The data is shown in figure 4 (not including the NW volume, which was estimated by approximating the NW geometry as a truncated cone. It decreases with increasing χHBr for all three materials, which means that the NW surface is not simply passivated by HBr). Figures 4(a)–(c) show the effect of in situ etching by χHBr on the NW length and the ratio Db/Dt for InP, GaP and GaAs NWs, respectively. The general trend for using HBr during growth is comparable for InP, GaP and GaAs in the respect that the length of the NWs with respect to the reference NWs (χHBr = 0) increases (up to a factor of about 1.6 for the GaP) through use of small amounts of HBr, which is similar to in situ etching of InP and GaP NWs by the use of HCl [14, 16]. The reaction products contribute to the NW growth. For the use of HBr above a certain χHBr, the trend is reversed; for all three binary materials, a seemingly linear decrease of the NW length with increased χHBr occurs (figures 1(d), 2(d) and 3(d)). At sufficiently high χHBr, no NWs could be observed. The ratio between the NW’s bottom and top diameter decreases with χHBr for all of the materials until radial growth is impeded (Db/Dt = 1). The NW diameter is unaffected by the increasing χHBr once tapering is impeded. We 2
A Berg et al
Nanotechnology 25 (2014) 505601
Figure 1. SEM images of in situ etched InP NWs. The InP NWs were grown at 450 °C with different amounts of HBr in the gas phase: (a) reference InP NW, χHBr = 0; (b) χHBr = 7.8 × 10−6; (c) χHBr = 14.1 × 10−6; (d) χHBr = 19.3 × 10−6. The images are recorded at the same magnification and at an angle of 30° toward the normal range of the substrate. The scale bar is 1 μm.
Figure 2. SEM images of in situ etched GaP NWs. The GaP NWs were grown at 600 °C with different amounts of HBr in the gas phase: (a)
reference GaP NW, χHBr = 0; (b) χHBr = 4.9 × 10−6; (c) χHBr = 37.7 × 10−6; (d) χHBr = 72.9 × 10−6. The images were recorded at the same magnification and at an angle of 30° toward the normal range of the substrate. The scale bar is 1 μm.
evaluated the temperature-dependent growth rate in the kinetically limited growth regime for GaP NWs in the presence of HBr, assuming that the growth rate is constant throughout the growth process. The rate-limiting process was found to have an activation energy of 118 kJ mol−1, similar to that of GaP NW growth in the presence of HCl (120 kJ mol−1
[16]), which was attributed to the group V dissociation [9]. In other words, HBr does not affect the rate-limiting process for GaP NW growth. For InP NWs grown at 450 °C, χHBr = 14.1 × 10−6 was necessary to eliminate radial growth completely (figure 1(c)), whereas χHCl = 25.0 × 10−6 was necessary to reach the same 3
A Berg et al
Nanotechnology 25 (2014) 505601
Figure 3. SEM images of in situ etched GaAs NWs. The GaAs NWs were grown at 470 °C with different amounts of HBr in the gas phase:
(a) reference GaAs NW, χHBr = 0; (b) χHBr = 2.5 × 10−6; (c) χHBr = 4.9 × 10−6; (d) χHBr = 9.8 × 10−6. The images were recorded at the same magnification and at an angle of 30° toward the normal range of the substrate. The scale bar is 1 μm.
Figure 4. Effect of HBr on (a) InP, (b) GaP and (c) GaAs NW lengths and tapering. The NW length is represented by squares, and the ratio between the bottom and top NW diameter (Db/Dt) is represented by triangles. The dashed line is a guideline for the eye and shows the NW length. The solid line is a guideline for the eye, indicating a Db/Dt of one, i.e. a non-tapered NW geometry. (d) Length of the GaAs NWs grown at 470 °C with different χAsH3, both with χHBr = 4.9 × 10−6 and without HBr. (e) Corresponding bottom/top diameter ratio. The error bars represent the standard deviation.
4
A Berg et al
Nanotechnology 25 (2014) 505601
reducing χAsH3 with respect to non-tapered growth. For high χAsH3, the NWs are similar in length and diameter, with and without HBr in the gas phase (figures 4(d), (e)). Here, the influence of HBr on GaAs NW growth with respect to impeding the radial growth is limited. Interestingly, for GaAs NWs not using HBr, the NWs are much shorter for low than for high χAsH3, indicating a change in NW growth from being group III to being group V limited. In this regime, the NWs grown using HBr are longer than the NWs grown without HBr and are taper-free. In particular, for our lowest χAsH3 = 0.55 × 10−4, the use of HBr was necessary for the growth of NWs. This suggests that HBr does not only react with the group III species, it also reacts with the group V species, i.e. for the case of growth of GaAs to form AsBr, which has a dissociation bond energy of 2.26 eV [37] and can interact with the Au alloy particle to assist NW growth. We cannot rule out that the bromination of solid GaAs occurs, contributing to NW growth by GaBr formation and As release. In order to assess the effect of HBr on the crystalline phase formed in the NWs, which could affect radial growth [12], a TEM characterization was carried out. Figure 5 shows, as an example, overview TEM images from samples grown with low and high etchant fractions for all three materials systems. The high etchant fraction sample shows no measurable tapering for InP, GaP or GaAs. No change in the crystal structure with the increasing etchant fractions could be detected for the GaP NWs, which were pure WZ with only a few stacking faults (
Search
Collections
Journals
About
Contact us
My IOPscience
In situ etching for control over axial and radial III-V nanowire growth rates using HBr
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2014 Nanotechnology 25 505601 (http://iopscience.iop.org/0957-4484/25/50/505601) View the table of contents for this issue, or go to the journal homepage for more Download details: IP Address: 132.239.1.231 This content was downloaded on 15/06/2017 at 15:43 Please note that terms and conditions apply.
You may also be interested in: Control of morphology and crystal purity of InP nanowires by variation of phosphine flux during selective area MOMBE A Kelrich, V G Dubrovskii, Y Calahorra et al. InxGa1xAs nanowires with uniform composition, pure wurtzite crystal phase and taper-free morphology Amira S Ameruddin, H Aruni Fonseka, Philippe Caroff et al. Particle-assisted GaxIn1-xP nanowire growth for designed bandgap structures D Jacobsson, J M Persson, D Kriegner et al. Sn-seeded GaAs nanowires grown by MOVPE Rong Sun, Neimantas Vainorius, Daniel Jacobsson et al. Synthesis and properties of ultra-long InP nanowires on glass Veer Dhaka, Ville Pale, Vladislav Khayrudinov et al. Growth parameter design for homogeneous material composition in ternary GaxIn1xP nanowires Alexander Berg, Filip Lenrick, Neimantas Vainorius et al. Direct nucleation, morphology and compositional tuning of InAs1xSbx nanowires on InAs (111) B substrates Luna Namazi, Sepideh Gorji Ghalamestani, Sebastian Lehmann et al. Recent progress in metal-organic chemical vapor deposition of (0001) N-polar group-III nitrides Stacia Keller, Haoran Li, Matthew Laurent et al.
Nanotechnology Nanotechnology 25 (2014) 505601 (9pp)
doi:10.1088/0957-4484/25/50/505601
In situ etching for control over axial and radial III-V nanowire growth rates using HBr Alexander Berg1, Kilian Mergenthaler1, Martin Ek2, Mats-Erik Pistol1, L Reine Wallenberg2 and Magnus T Borgström1 1 2
Solid State Physics, Lund University, Box 118, S-221 00, Lund, Sweden Polymer and Materials Chemistry/nCHREM, Lund University, Box 124, S-221 00 Lund, Sweden
E-mail: [email protected] Received 2 August 2014, revised 6 October 2014 Accepted for publication 10 October 2014 Published 25 November 2014 Abstract
We report on the influence of hydrogen bromide (HBr) in situ etching on the growth of InP, GaP and GaAs nanowires. We find that HBr can be used to impede undesired radial growth during axial growth for all three material systems. The use of HBr opens a window for optimizing the growth parameters with respect to the materials’ quality rather than only their morphology. Transmission electron microscopy (TEM) characterization reveals a partial transition from a wurtzite crystal structure to a zincblende upon the use of HBr during growth. For InP, defect-related luminescence due to parasitic radial growth is removed by use of HBr. For GaP, the etching with HBr reduced the defect-related luminescence, but no change in peak emission energy was observed. For GaAs, the HBr etching resulted in a shift to lower photon emission energies due to a shift in the crystal structure, which reduced the wurtzite segments. Keywords: nanowire, HBr, in situ etching, MOVPE (Some figures may appear in colour only in the online journal)
was reported [14, 15], opening the field for optimizing NWs with respect to the material’s quality rather than compromising the growth parameters to optimize the morphological properties. In addition to InP, this method has been successfully applied to GaP [16], GaInP [17] and InAsP [18] NWs. However, until now, HCl has not been reported to eliminate the tapering of GaAs NWs. Although our group has tried to grow taper-free GaAs NWs, we were not able to remove the radial growth by HCl [19]. Here, we report on the effects of in situ etching by use of hydrogen bromide (HBr) on InP, GaP and GaAs NWs. HBr was chosen as an alternative halide to HCl, as it is known to be a versatile etchant of III-V materials [20–24]. We show that in situ etching by use of HBr can prevent radial growth of InP, GaP and GaAs NWs, despite growth conditions that otherwise lead to strong tapering. This allows control over the design and growth of axially defined NW materials. However, the use of HBr for impeding the radial growth on GaAs NWs is limited to a smaller parameter space than that for phosphides.
1. Introduction Nanowires (NWs) have recently attracted attention as one of the most promising ways to combine high performance III-V materials with new functionality [1–4] with highly mature Si technology [5, 6]. For device applications, however, uncontrolled parasitic radial growth on the NW side’s facets during the remainder of the synthesis typically occurs. Such unintentionally grown NW shells can lead to short circuiting of the axially designed components and can become a surrounding and competing recombination centre for charge carriers [7]. In order to affect the ratios between the axial and radial NW growth, parameters such as the growth temperature [8–10] and catalyst supersaturation [11, 12] can be used. For thin film growth, however, parameter tuning, such as the tuning used to favour axial NW growth, typically leads to poor crystalline quality [13] and inhibits the use of growth parameter variations for the optimization of crystal quality and doping. Recently, a method using in situ etching by use of HCl in order to take control over axial and radial growth rates 0957-4484/14/505601+09$33.00
1
© 2014 IOP Publishing Ltd Printed in the UK
A Berg et al
Nanotechnology 25 (2014) 505601
2. Experimental
recorded on the [110] zone axis in sequence along the length of a few NWs. Photoluminescence (PL) measurements were performed at a temperature of about 5 K. The NWs were broken off from native substrates and were spread out on a gold-covered silicon surface. The NW under study was evenly illuminated throughout its length using a CW laser emitting at 532 nm for the InP and GaAs and a 325 nm laser for the GaP. The PL was collected by an optical microscope, dispersed through a spectrometer and detected by a thermoelectrically cooled CCD camera. At least 5 NWs per sample were studied.
The samples were prepared by depositing 40 nm diameter Au particles on (1¯ 1¯ 1¯)B InP, (1¯ 1¯ 1¯)B GaP and (1¯ 1¯ 1¯)B GaAs substrates via an aerosol technique [25], resulting in randomly distributed monodisperse particles with a homogeneous density of 1 × 106 cm−2. The particle-assisted NW growth was performed by use of a low-pressure (100 mbar) metal organic vapour phase epitaxy (MOVPE) system. Trimethylindium (TMI), trimethylgallium (TMG), phosphine (PH3) and arsine (AsH3) were used as precursors, and hydrogen bromide (HBr) was used as an etching agent in a total flow of 6.0 l min−1 using hydrogen (H2) as the carrier gas. For the InP NWs, the PH3 and TMI molar fractions (χPH3, χTMI) were set to χPH3 = 6.25 × 10−3 and χTMI = 3.5 × 10−6. For the GaP, the TMG molar fraction (χTMG) was set to 37.8 × 10−6, and the χPH3 was set to 6.25 × 10−3. For InP and GaP, prior to growth, the samples were heated to 600 °C under a PH3/H2 gas mixture for 10 min in order to desorb any surface oxides. Then, the reactor temperature was set to a growth temperature of 450 °C (InP) or 600 °C (GaP) for a direct comparison of the NW growth in the presence of HBr with the results of in situ removal of tapering by use of HCl for InP [14] and GaP [16], respectively. At these elevated temperatures, strong tapering is expected to occur [9, 26]. GaP has a larger parameter space for growth with respect to temperature compared to InP [26]; therefore, a temperature-dependent study from 440 °C to 600 °C in steps of 20 °C while keeping a constant HBr molar fraction (χHBr) of 72 × 10−6 was carried out. For GaAs, a growth temperature of 470 °C was chosen for the evaluation. At this temperature, GaAs NWs grow under optimal conditions with respect to the material’s use but are tapered [8]. The samples were annealed at 650 °C for 10 min under an AsH3/ H2 gas mixture. χTMG was set to 6.4 × 10−6, and χAsH3 was set to 1.1 × 10−4. In order to assess the growth dynamics of GaAs NWs, an χAsH3 variation of 0.55 × 10−4 < χAsH3 < 8.8 × 10−4 was carried out using χHBr = 0 and 4.9 × 10−6. The NW growth was initiated by adding the group III precursor to the flow. HBr was added to the flow after a 15 s nucleation step for InP and a 30 s nucleation step for GaAs. For GaP, the HBr etchant was introduced into the reactor simultaneously as TMG. χHBr was varied between 0 and 140 × 10−6 for growth of the NWs. The growth was terminated after 10 min and 15 s for the InP after 3 min for the GaP and after 10 min and 30 s for the GaAs by switching off group III and the HBr simultaneously, after which the samples were cooled down to room temperature under a mixture of the respective group V precursor and H2. Scanning electron microscopy (SEM) was used to characterize the NW morphology. At least 10 NWs at each of the three different locations on the samples were evaluated with respect to length and tapering. The samples were prepared for high resolution transmission electron microscopy (HRTEM) by direct transfer of the NWs to a Cu grid with a lacey carbon film by gently pressing the grid onto the substrate. For all three material systems, two samples, each with different χHBr, were characterized by HRTEM. For each sample, the images were
3. Results and discussion We first discuss the general trends that are valid for all three materials systems. Figures 1–3 show SEM images of the InP, GaP and GaAs NWs, respectively, as well as the effect on the NW morphology of using different χHBr during growth. Figure 1(a) shows a reference InP NW grown intentionally with parameters that give strong tapering when HBr is not present. The rough InP (1¯ 1¯ 1¯)B and GaP (1¯ 1¯ 1¯)B substrate surfaces, seen in figure 2(a), indicate competing InP and GaP surface growth. By increasing χHBr, the NW tapering gradually decreases, and the InP and GaP substrates’ roughness diminishes (figures 1(b), (c) and 2(b), (c)). The GaAs (1¯ 1¯ 1¯)B (figure 3(a)) substrate does not show a clear roughness before in situ etching; consequently, no strong effect of the etching on the substrate growth is observed (figures 3(b), (c)). The tapering is reduced by increasing χHBr for all three materials; thus, HBr is more versatile than HCl in the use of in situ etching of NWs in order to take full control over axial and radial growth rates. To assess the effect of using χHBr during NW growth, we evaluated the NW volume, NW length and the ratio between the bottom and top NW diameter (Db/Dt) as a function of χHBr by SEM after growth. The data is shown in figure 4 (not including the NW volume, which was estimated by approximating the NW geometry as a truncated cone. It decreases with increasing χHBr for all three materials, which means that the NW surface is not simply passivated by HBr). Figures 4(a)–(c) show the effect of in situ etching by χHBr on the NW length and the ratio Db/Dt for InP, GaP and GaAs NWs, respectively. The general trend for using HBr during growth is comparable for InP, GaP and GaAs in the respect that the length of the NWs with respect to the reference NWs (χHBr = 0) increases (up to a factor of about 1.6 for the GaP) through use of small amounts of HBr, which is similar to in situ etching of InP and GaP NWs by the use of HCl [14, 16]. The reaction products contribute to the NW growth. For the use of HBr above a certain χHBr, the trend is reversed; for all three binary materials, a seemingly linear decrease of the NW length with increased χHBr occurs (figures 1(d), 2(d) and 3(d)). At sufficiently high χHBr, no NWs could be observed. The ratio between the NW’s bottom and top diameter decreases with χHBr for all of the materials until radial growth is impeded (Db/Dt = 1). The NW diameter is unaffected by the increasing χHBr once tapering is impeded. We 2
A Berg et al
Nanotechnology 25 (2014) 505601
Figure 1. SEM images of in situ etched InP NWs. The InP NWs were grown at 450 °C with different amounts of HBr in the gas phase: (a) reference InP NW, χHBr = 0; (b) χHBr = 7.8 × 10−6; (c) χHBr = 14.1 × 10−6; (d) χHBr = 19.3 × 10−6. The images are recorded at the same magnification and at an angle of 30° toward the normal range of the substrate. The scale bar is 1 μm.
Figure 2. SEM images of in situ etched GaP NWs. The GaP NWs were grown at 600 °C with different amounts of HBr in the gas phase: (a)
reference GaP NW, χHBr = 0; (b) χHBr = 4.9 × 10−6; (c) χHBr = 37.7 × 10−6; (d) χHBr = 72.9 × 10−6. The images were recorded at the same magnification and at an angle of 30° toward the normal range of the substrate. The scale bar is 1 μm.
evaluated the temperature-dependent growth rate in the kinetically limited growth regime for GaP NWs in the presence of HBr, assuming that the growth rate is constant throughout the growth process. The rate-limiting process was found to have an activation energy of 118 kJ mol−1, similar to that of GaP NW growth in the presence of HCl (120 kJ mol−1
[16]), which was attributed to the group V dissociation [9]. In other words, HBr does not affect the rate-limiting process for GaP NW growth. For InP NWs grown at 450 °C, χHBr = 14.1 × 10−6 was necessary to eliminate radial growth completely (figure 1(c)), whereas χHCl = 25.0 × 10−6 was necessary to reach the same 3
A Berg et al
Nanotechnology 25 (2014) 505601
Figure 3. SEM images of in situ etched GaAs NWs. The GaAs NWs were grown at 470 °C with different amounts of HBr in the gas phase:
(a) reference GaAs NW, χHBr = 0; (b) χHBr = 2.5 × 10−6; (c) χHBr = 4.9 × 10−6; (d) χHBr = 9.8 × 10−6. The images were recorded at the same magnification and at an angle of 30° toward the normal range of the substrate. The scale bar is 1 μm.
Figure 4. Effect of HBr on (a) InP, (b) GaP and (c) GaAs NW lengths and tapering. The NW length is represented by squares, and the ratio between the bottom and top NW diameter (Db/Dt) is represented by triangles. The dashed line is a guideline for the eye and shows the NW length. The solid line is a guideline for the eye, indicating a Db/Dt of one, i.e. a non-tapered NW geometry. (d) Length of the GaAs NWs grown at 470 °C with different χAsH3, both with χHBr = 4.9 × 10−6 and without HBr. (e) Corresponding bottom/top diameter ratio. The error bars represent the standard deviation.
4
A Berg et al
Nanotechnology 25 (2014) 505601
reducing χAsH3 with respect to non-tapered growth. For high χAsH3, the NWs are similar in length and diameter, with and without HBr in the gas phase (figures 4(d), (e)). Here, the influence of HBr on GaAs NW growth with respect to impeding the radial growth is limited. Interestingly, for GaAs NWs not using HBr, the NWs are much shorter for low than for high χAsH3, indicating a change in NW growth from being group III to being group V limited. In this regime, the NWs grown using HBr are longer than the NWs grown without HBr and are taper-free. In particular, for our lowest χAsH3 = 0.55 × 10−4, the use of HBr was necessary for the growth of NWs. This suggests that HBr does not only react with the group III species, it also reacts with the group V species, i.e. for the case of growth of GaAs to form AsBr, which has a dissociation bond energy of 2.26 eV [37] and can interact with the Au alloy particle to assist NW growth. We cannot rule out that the bromination of solid GaAs occurs, contributing to NW growth by GaBr formation and As release. In order to assess the effect of HBr on the crystalline phase formed in the NWs, which could affect radial growth [12], a TEM characterization was carried out. Figure 5 shows, as an example, overview TEM images from samples grown with low and high etchant fractions for all three materials systems. The high etchant fraction sample shows no measurable tapering for InP, GaP or GaAs. No change in the crystal structure with the increasing etchant fractions could be detected for the GaP NWs, which were pure WZ with only a few stacking faults (
