Anisotropic-Strain-Induced Band Gap Engineering ... - ACS Publications

Mar 26, 2018 - deposited at 100 °C lacks the necessary solidity to transfer any significant strain to the NWs. Through the grain structure, one can t...
1 downloads 0 Views 6MB Size
Subscriber access provided by Service des bibliothèques | Université de Sherbrooke

Anisotropic-strain-induced bandgap engineering in nanowire-based quantum dots Luca Francaviglia, Andrea Giunto, Wonjong Kim, Pablo Romero, Jelena Vukajlovic-Plestina, Martin Friedl, Heidi Potts, Lucas Güniat, Gözde Tütüncüoglu, and Anna Fontcuberta i Morral Nano Lett., Just Accepted Manuscript • Publication Date (Web): 26 Mar 2018 Downloaded from http://pubs.acs.org on March 26, 2018

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 16 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

Anisotropic-strain-induced bandgap engineering in nanowire-based quantum dots Luca Francaviglia,

†,‡

†,‡

Andrea Giunto,



Jelena Vukajlovic-Plestina,

Wonjong Kim,



Martin Friedl,

Tütüncüoglu,





Pablo Romero-Gomez,

Heidi Potts,





Lucas Güniat,

and Anna Fontcuberta i Morral



Gözde

∗,†

†Laboratoire des Matériaux Semiconducteurs, Institut des Matériaux, Ecole Polytechnique

Fédérale de Lausanne, 1015 Lausanne, Switzerland ‡equal contribution E-mail: anna.fontcuberta-morral@ep.ch

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

Abstract

28

Tuning light emission in bulk and quantum structures by strain constitutes a complementary method to engineer functional properties of semiconductors. Here, we demonstrate the tuning of light emission of GaAs nanowires and their quantum dots up to 115 meV by applying strain through an oxide envelope. We prove that the strain is highly anisotropic and clearly results in a component along the NW longitudinal axis, showing good agreement with the equations of uniaxial stress. We further demonstrate that the strain strongly depends on the oxide thickness, the oxide intrinsic strain, and the oxide microstructure. We also show that ensemble measurements are fully consistent with characterizations at the single-NW level, further elucidating the general character of the ndings. This work provides the basic elements for strain-induced bandgap engineering and opens new avenues in applications where a band-edge shift is necessary.

29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51

23 24 25

Keywords

52

Nanowires, quantum dots, strain, oxide, GaAs, PECVD, Raman, photoluminescence.

54

Semiconductor nanowires (NWs) are very ver-

57

53

26 27

55 56

58

satile building blocks for optoelectronic devices. As an example, NWs can host material or phase combinations otherwise dicult to obtain in the bulk or in thin lms. 3,4 In this way, a plethora of heterostructures including quantum dots (QDs) can be obtained within NWs. The optical performance of the embedded QDs benets from the tailored shape and size of a NW, e.g. by an enhanced photon extraction 58 or detection 9,10 and the potential for an ecient electrical excitation. 11 Several research groups have tried to tune the emission energy of QDs for different purposes. Tuning energy levels involves tuning the absorption/emission which may be used for sensing, for the storage of information, as well as to facilitate the coupling to cavity modes. 1215 Strain has shown to be a valuable approach to largely and reversibly tune the emission of QDs. 16 The particular structure and mechanical properties of NWs provide an extended elastic regime in which strain can be applied. 1720 The approaches used to apply strain to NW structures include a exural tensile setup, the use of nanowires as cantilevers, the application of surface acoustic waves, and the deposition of an oxide envelope. 1823 So far, the latest approach has produced the largest tuning of the emission energy, but the origin of the shift is not clear yet. Here, we report on a NW-oxide system with QDs embedded in the shell of core-shell GaAs-

ACS Paragon Plus Environment

1

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

a)

QDs

SiO2

b)

c)

Page 2 of 16 d)

e)

GaAs AlGaAs

f)

Ga, As, Al Si, O

Figure 1: (a) Sketch of the NW on growth substrate, uncoated (left) and coated by SiO2 (right). The cross sections

above show the QDs in the shell of the NW. 1,2 (b) TEM X-ray energy-dispersive-spectroscopy (EDS) map of the top of a GaAs-AlGaAs NW coated with SiO2 . (c) Low-magnication TEM micrograph of a NW coated with SiO2 deposited at 300◦ C. (d) Low-magnication TEM micrograph of a NW coated with SiO2 deposited at 100◦ C.(e)-(f) SEM images at 20◦ tilting of the same NWs in an array, respectively before and after the coating with 180 nm of SiO2 . Scale bars: (b) 100 nm, (c)-(f) 2µm. 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89

AlGaAs NWs. 1,2 We provide the rst experi- 90 mental evidence for the presence of strain inde- 91 pendently from the redshift of the semiconduc- 92 tor bandgap. To do so, we use non-resonant Ra- 93 man spectroscopy. We also provide experimen- 94 tal support to the fact that the applied strain 95 mainly results in a uniaxial component along 96 the main axis of the NW, in agreement with 97 previous simulations. 22,23 We univocally corre- 98 late the strain and the redshift on large and 99 variable ensembles of NWs and QDs and at the 100 single nanowire level, in order to prove the re- 101 liability and reproducibility of the chosen tech- 102 nique. We provide understanding on the role of 103 the microstructure and deposition temperature 104 of the oxide. This last result, together with the 105 reproducibility of the straining method, opens 106 the possibility to engineer the bandgap and the 107 surface properties of NWs and QDs in NWs. 108 The combination of the two degrees of freedom 109 and the application of dierent materials sug- 110 gests the potential of this technique in photo- 111 electrochemical or optoelectronic applications 112 where a bandgap modulation and surface pro- 113 tection by an oxide is needed. 114 The NW-oxide straining system. We 115 start by explaining how to perform strain en- 116 gineering in NWs and core-shell QDs by apply- 117 ing an external amorphous shell. Figure 1a de- 118 picts the schematics of the experiment and the 119 system studied. We can grow NWs both in a 120

self-ordered way 24,25 and in patterned arrays. 26 In both cases we grow the NWs on silicon to ensure there is no light emission from the substrate. We begin with a GaAs NW obtained by the Ga-assisted method 2628 and then follow up with the growth of an AlGaAs shell that intrinsically contains self-assembled QDs. 1,2 The QD emission is then tuned by applying a static straining device (Figure 1a), which is a layer of silicon dioxide obtained by plasma enhanced chemical vapor deposition (PECVD). Figure 1b corresponds to the chemical analysis of the III-V/oxide NW-coating structure, indicating a sharp interface between the semiconductor and the oxide. Figure 1c and d correspond to typical transmission electron microscopy (TEM) micrographs of NWs respectively coated with SiO2 at 300◦ C (Figure 1c) and 100◦ C (Figure 1d) as a substrate temperature. One observes that PECVD is highly conformal and the oxide layer follows the geometry of the NW (more details in Supporting Information). However, a microstructure formed by longitudinal grains is visible in Figure 1d and pronounced tapering is also observed in Figure 1c: the oxide thickness gradually decreases from the top to the base of the NW. This gradient is due to both electriceld enhancement at the NW tip and the mutual shadowing of the precursors by neighbour NWs during the deposition. The thicker the

ACS Paragon Plus Environment

2

123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169

a)

Uncoated NW

Uncoated NW

b)

4K

12K

GaAs 3

AlGaAs 5

2

4

1

c)

180nm SiO2

e)

180nm SiO2

d) Intensity (a.u.)

122

nominal oxide thickness is, the more important the tapering becomes. For this reason, we have calibrated the real-vs-nominal oxide thickness along the NW length (more details in Supporting Information). In the following, we will refer to the SiO2 thickness as the average value measured at the top of the NWs, unless dierently specied. The arrows in Figure 1c indicate what we consider as the NW top. We could recognize and address the single NWs before and after the oxide deposition when core-shell GaAs-AlGaAs NWs 1 were grown on site-selected positions on a silicon substrate. 26 Figures 1e-f show scanning electron microscopy (SEM) micrographs, from exactly the same NWs at the corner of an array, before and after the deposition of 180 nm of oxide. It is interesting to notice that the tapering is absent in the array NWs, as demonstrated by the one in Figure 1d. This is most probably due to the fact that the length of these NWs is shorter (4 µm) than in the case of the self-assembled NWs (10 µm), like the one shown in Figure 1c. For a given inter-NW distance, a smaller NW length causes less mutual shadowing and the resulting tapering is reduced or even eliminated. We characterized the light emission from the QDs by micro-photoluminescence (µ-PL) spectroscopy performed on each individual NW, while we studied the oxide-induced strain by means of Raman spectroscopy. Unless dierently stated, both PL and Raman spectra were always taken at the same top position indicated in Figure 1c in order to discard any spread in the results related to the variation of the oxide thickness along the NW axis. This choice equally allows to discard any variations of the QD emission energy along the NW axis. 25 Impact of strain on NWs and QDs. First, we elucidate the controlled strain by Raman spectroscopy. In particular, we demonstrate that the SiO2 provides a tensile strain along the NW direction. It is well known that an applied stress eld results in a change of the phonon energies of an (Al)GaAs crystal, which can be assessed by Raman spectroscopy. 19,20,2931 The Raman measurements shown in Figure 2a, c and e were performed at 12K, on NWs transferred on a Si substrate in the same backscat-

Intensity (a.u.)

121

360nm SiO2

3

SiO2 2 1

4 5

360nm SiO2

f)

GaAs-like TO GaAs-like LO AlAs-like TO AlAs-like LO GaAs TO

3

12

4

SiO2 5

Emission energy (eV)

Raman shift (cm-1) h)

g) Normalized intensity

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

12K

250 300 350 400 Raman shift (cm-1)

Norm. intensity

Page 3 of 16

12K

1

2 34

12 4 3

1.4 1.65 1.85 Emission energy (eV)

Figure 2: From top to bottom: optical measurements

on single NWs respectively uncoated (a, b), coated with 180 nm of SiO2 (c, d) and with additional 180 nm of SiO2 (e, f), as represented in the sketches in the right corners of (b, d, f). (a, c, e) Micro-Raman at 12K of single NWs. The peaks are tted by Lorentzian curves. Color-coding legend in (e). The pentagon in (a) labels the 2nd -order Raman scattering of the Si substrate. The dashed vertical line corresponds to the position of the GaAs TO peak in the uncoated NW as a reference. (b, d, f) µ-PL spectra at 4.2K of exactly the same array NW. The dashed vertical lines correspond to the emission from the core and the QD line 3 in the uncoated NW. Raman (g) and PL (h) spectra of the very same vertical array NW before and after SiO2 coating. The star in (g) indicates the GaAs LO mode. The color coding is the same as in (e). The dashed vertical line corresponds to the GaAs TO mode in the uncoated NW (g). Four QD PL lines are labelled in (h).

ACS Paragon Plus Environment

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

170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218

tering conguration as in Ref. 19. We consider 219 the Lorentzian t of the peaks in the spectrum 220 of the uncoated NW as a reference. The spectra 221 are composed of two groups of peaks. At high 222 wavenumbers we nd the AlAs-like peaks from 223 the AlGaAs shell (TO 359.5 ± 0.3 cm−1 and 224 LO 376.7 ± 0.2 cm−1 ). At lower wavenumbers, 225 we nd the GaAs TO mode at 266.4±0.1−1 30 226 and the GaAs-like modes from the AlGaAs- 227 shell (TO 260.3 ± 0.5 and LO 277.9 ± 0.1 cm−1 ). 32,33 The GaAs LO mode is not present, due to the selection rules. 30,32 The position of the AlAs and GaAs-like LO modes of 5 NWs is consistent with an average Al composition of 29±8%. 34 All the Raman modes downshift upon both the oxide depositions shown in Figure 2a. In particular, the GaAs TO mode downshifts by 228 2.31 ± 0.21 cm−1 after the deposition of 180 229 nm of SiO2 and by 4.41 ± 0.14 cm−1 after the 230 deposition of 360 nm of SiO2 . This trend fol- 231 lows a linear correlation between downshift and 232 oxide thickness. We thus deduce that the oxide 233 has a clear impact on the strain and that it can 234 235 be assessed by Raman spectroscopy. Previous works suggested that the oxide pro- 236 vides anisotropic response along the longitudi- 237 nal axis of the NW, even under the assumption 238 of isotropic properties of the oxide. 22,23 Here, 239 we provide additional arguments of why the 240 strain should be uniaxial. The TEM micro- 241 graph of the NW coated with SiO2 shown in 242 Figure 1d provides some insight. The micro- 243 graph clearly shows a granular structure, with 244 oriented grains tilted towards the NW axis. The 245 TEM micrographs of all the NWs show this fea- 246 ture, even when the grains are less pronounced. 247 We believe that this structure results from the 248 impingement direction of atoms and ions dur- 249 ing the oxide deposition (Figure 1d and more 250 details in Supporting Information). Similar 251 structures have been observed before on NWs 252 coated with permalloy by sputtering. 35 Given 253 this structural anisotropy of the oxide shell, 254 we deduce that the exerted strain is probably 255 anisotropic, as well. In particular, a net longi- 256 tudinal contribution should be considered along 257 258 the NW growth axis. We proceed now with the quantication of the 259

Page 4 of 16

strain by evaluating the Raman shifts. For this we use the model from Ref. 19. In this work, Signorello et al. applied a controlled uniaxial stress along the longitudinal axis of GaAs NWs, which corresponds to the (111) direction of the zinc-blende (ZB) crystal. This direction is the same as in our case. Eq. 1 relates the shift of the GaAs TO mode (∆ωT O ) with the strain along the NW longitudinal axis (zz ):

zz =

with H =

(1)

1 − 2ν 3

ωT O is the relaxed phonon frequency in cm−1 for the GaAs TO mode, γT = 1.35 and rT0 = −0.88 are respectively the hydrostatic and deviatoric mode Grüneisen parameters 19 and ν = 0.16 is the Poisson ratio for GaAs (111). We use the same Grüneisen parameters as for bulk GaAs because a minor dierence was reported, 19,31 while we utilize the Poisson ratio found for GaAs NWs in the (111) direction. 19 We use the same parameters for all temperatures, as no substantial dierence has been reported for measurements at lower temperatures. 19 The shift of the GaAs TO mode is consistent with a tensile strain of 0.54±0.06% and 1.04±0.08% upon, respectively, the rst and second deposition of 180 nm and 360 nm of SiO2 . Given the brittle behavior of GaAs at low temperature, it is remarkable that the NWs do not break under such a high strain. 36 Probably, the high surface-to-volume ratio and the absence of bulk and surface defects increase the fracture stress in the NWs. 17 Further support to the calculation of strain by means of eq. 1 will be provided in the following by the measurements on NW ensembles. We address now the impact of strain on the optical properties of the NWs and QDs. Figures 2b, d anf f report the µ-PL spectra performed exactly on the same NW upon successive coatings with SiO2 . The top spectrum (Figure 2b) corresponds to the NW-QD structure before any oxide deposition. The two spectra below

ACS Paragon Plus Environment

4

1 ∆ωT O ωT O [−3γT H + rT0 (1 − H)]

Page 5 of 16 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

260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308

Nano Letters

(Figures 2d and f) were taken after depositions 309 of 180 nm each time, for comparison with the 310 same oxide thickness as in the Raman spectra 311 in Figures 2a, c and e. The PL spectra were 312 acquired at 4.2K, with the laser incident on the 313 top of vertical NWs. 314 In each µ-PL spectrum, we distinguish emis- 315 sion of two dierent origins. The broader emis- 316 sion around 1.5 eV corresponds to the free ex- 317 citon emission from the GaAs NW core. It 318 mainly originates from GaAs in the ZB phase. 319 In the top spectrum of the uncoated NW, the 320 peak exhibits a small shoulder at lower energies, 321 which may be due to the presence of few crystal 322 twins and few wurtzite (WZ) segments. The in- 323 tensity modulation is attributed to Fabry-Pérot 324 resonances 37 and it is particularly visible in the 325 samples coated with SiO2 . The emission clearly 326 redshifts for an increasing oxide thickness. We 327 have tted the peaks with a Gaussian curve 328 to assess the energy shift upon SiO2 deposi- 329 tion. In particular, we tted the two peaks in 330 the PL of the uncoated NW with two Gaus- 331 sian curves, while in the other PL spectra the 332 Gaussian curves t the maxima of the Fabry- 333 Pérot resonances. The rst deposition results 334 in a redshift of about 68 meV, while the sec- 335 ond deposition brings an additional redshift of 336 about 9 meV. One should note that the second 337 deposition results in the oxide completely lling 338 the space between the NWs. We think that this 339 changes the straining conditions with respect to 340 the case in which the single NWs are enveloped 341 by independent oxide shells. In turn, we think 342 that this can explain the smaller redshift of the 343 core emission after the second deposition. 344 The group of narrow lines at higher energy in 345 the spectra in Figure 2b, d and e originates from 346 the QD emission in the NW AlGaAs shell. Due 347 to the excitation depth of the laser we excite 348 several QDs simultaneously. 1,2 One advantage 349 of measuring the same single NW is that we 350 can follow the evolution of individual emission 351 lines. Some illustrative peaks at dierent en- 352 ergies are labelled as 1, 2, 3, 4 and 5. Also in 353 this case a redshift is qualitatively visible. As 354 quantitative examples, the peak ts of the lines 355 1, 2, 3, 4 and 5 all give a redshift between about 356 60 and 73 meV upon the rst deposition. The 357

second deposition brings and additional redshift between about 7 and 21 meV. There is no apparent correlation between the redshift and the initial QD emission energy. On the contrary, both the core and all the measured QD lines, redshift more after the rst deposition. As already mentioned, the second deposition is less eective probably because the oxide completely lls the space between the NWs. Interestingly, we notice an overall increase in both the Raman and PL intensity after coating the NWs, which is already visible in the spectra in Figures 2a to f. (more details in Supporting Information). We believe that this positive effect depends on the smoother transition in refractive index between the vacuum and the NW in presence of SiO2 . 38 Figures 2a to f show the cumulative eect of the SiO2 deposition on the Raman and PL spectra of the NWs. In order to corroborate the link between the Raman downshift and the PL redshift, Figures 2g and h correspond to the same vertical NW. We acquired Raman and PL spectra at the same temperature (12K) before and after the deposition of 110 nm of SiO2 . We observe that the GaAs TO peak downshifts by slightly more than 1 cm−1 and the GaAs LO by about 0.6 cm−1 ; the PL emission of the QD 3 in Figure 2h redshifts by 23 meV and the core PL by about 37 meV. The PL redshift is in reasonable agreement with the Raman downshift measured on the same NW. As expected for 110 nm of SiO2 , these shifts are lower than those obtained for 180 nm of SiO2 and shown in Figures 2c and d. No relevant dierence is present between the PL spectra acquired at 4K (Figure 2b, d, f) and those acquired at 12K (Figure 2g, h). Nature of strain. In order to provide more precise evidence of the nature of strain we turn to the systematic study of larger ensembles of NWs. Measurements on large NW ensembles provide a statistically robust support to the results obtained on single NWs and conrm the reproducibility of the results. For this, the NW structures were grown in a self-organized manner on the Si substrates as in Ref. 2,24,25. A large-area sample was divided into nine pieces. Each piece underwent the deposition of a SiO2

ACS Paragon Plus Environment

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

a)

Page 6 of 16

b)

c)

12K

4K

QDs 90 meV/%

QDs AlAs-like TO

12K GaAs TO

GaAs core

Core 63 meV/%

Figure 3: (a) Average GaAs TO (red circles) and AlAs-like TO (blue triangles) Raman shift of horizontal NWs at 12K vs oxide thickness. All the error bars are the standard deviations of the distributions. (b) QD emission energy acquired by PL on ensembles of horizontal NWs at 4.2 K vs oxide thickness. The horizontal error bars are the standard deviations of the oxide thickness distributions. Average of the emission energy of the QDs (green circles) and GaAs core (black triangles) distributions with the standard deviation as a vertical error bar. The open black circles in a and b are the values of the corresponding quantities derived from the single-NW spectra in Figures 2a and b. (c) Average QD (green circles) and core (black triangles) PL redshift vs strain. The dashed green and black lines are a linear t of, respectively, the QD and the core PL redshift. The dash-dotted red line corresponds to the theoretical redshift of the heavy-hole bandgap recombination in GaAs NWs. 19 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381

coating of dierent thickness, excluding one 382 kept as an uncoated reference. The oxide thick- 383 ness ranged between 75 nm and 527 nm, as 384 characterized by TEM. For these Raman and 385 PL measurements on large ensembles, we dis- 386 persed the NWs onto silicon substrates. 387 We chose 5 NWs for each oxide thickness to 388 perform Raman measurements at 12K. Figure 389 3a shows the average peak position of two TO 390 modes as a function of the thickness of the ox- 391 ide in each subsample. The peak position was 392 derived from Lorentzian ts, like in Figure 2a 393 (more details in Supporting Information). The 394 error bars represent the standard deviations. 395 Here, we show only the evolution of the GaAs 396 TO and AlAs-like TO modes for clarity. The 397 GaAs TO mode is independent from the per- 398 centage of Al in the shell. Similarly, the AlAs- 399 like TO mode has an almost at dependence on 400 the Al content in the AlGaAs alloy. 34 Therefore, 401 we expect that the peak position of these modes 402 depends only on the strain in the NW and we 403 use the GaAs TO mode as a gauge of the ap- 404 plied strain. Figure 3a shows that the Raman 405

peaks downshift for increasing oxide thickness in a linear way. The Raman downshift conrms the increase in tensile stress applied by an increasing thickness of the oxide, which, in turn, redshifts the PL emission energy. We notice a broadening of the vertical error bar in the graph of Figure 3a. This broadening is a consequence of the larger variability in the oxide properties for thicker depositions, like local changes in thickness and density. Indeed, it actually highlights the large impact of the oxide shell against minor variations in the properties of the oxide and the NW. Considering the average GaAs TO peak, we measure a maximum downshift of 6.45 ± 2.22 cm−1 between the uncoated NWs and those coated with the thickest oxide (measured thickness of 527 nm on average). We performed µ-PL measurements at 4.2K on 25 NWs from the same ensembles prepared for Raman spectroscopy. Figure 3b shows the average emission energy of the NW GaAs core and shell QDs as a function of the measured oxide thickness. The error bars are the standard de-

ACS Paragon Plus Environment

6

Page 7 of 16 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

406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454

Nano Letters

viations of the emission distributions. Both the 455 QD and the core PL redshifts are clearly visible 456 and maximized by the second-to-last deposition 457 with respect to the uncoated NWs (measured 458 oxide thickness of 365 nm). The average QD 459 position shifts by ∼ 115 meV, while the GaAs 460 core by only ∼ 94 meV. The shifts are signi- 461 cantly larger than the dispersion given by the 462 error bars. Both the horizontal and vertical er- 463 ror bars in Figure 3b increase for thicker ox- 464 ides. This observation reects the increase in 465 the variability of the oxide properties in thicker 466 depositions. Yet, the overall trend of the graph 467 is clear, and it supports the fact that the ap- 468 plied strain is large enough to overcome minor 469 variations in the properties of the oxide and the 470 NWQDs. 3941 Theoretical predictions and ex- 471 periments show that the pressure coecient of 472 AlGaAs at the Γ point increases with the Al 473 content in the composition range of our inter- 474 est. 34,42 As a consequence, one can deduce that, 475 for a given strain, the AlGaAs bandgap red- 476 shifts more than the one of pure GaAs. This 477 fact can account at least in part for the dier- 478 ence in the overall redshift of the core and the 479 QDs, which is visible in Figure 3b. Moreover, 480 one should take into account that, in the case 481 of the QDs, strain does not only redshift the 482 AlGaAs band gap in which the QD potential 483 well is set. Strain can also decrease the conne- 484 ment by changing the QD shape and decreas- 485 ing the QD potential barrier. This second con- 486 sideration is supported by the aforementioned 487 increase in pressure coecient in presence of 488 Al. 34,42 The QD barriers redshift more than the 489 Al-poor QD well. Therefore the connement 490 on the excitons trapped in the QDs is smaller, 491 which provides an additional redshift. 492 It is also interesting to focus on the onset 493 of the PL redshift of the QDs and the core. 494 The core PL shows a rst slope that seems to 495 get steeper after the rst two depositions. Ac- 496 cording to Ref. 19, uniaxial stress along [111] 497 lifts the GaAs crystal symmetry and light- and 498 heavy-hole bands split; in particular, the heavy- 499 hole band redshifts more than the light-hole 500 one. Under small strain, the two bands are still 501 close enough in energy to contribute to the PL 502 emission. 19 Under larger strain, only the heavy 503

holes contribute. This transition may be visible as the initial change in slope in the core redshift in Figure 3b. On the contrary, because of connement, we can assume that QDs only emit by recombination of electron and heavy-hole excitons, that is with a unique slope. We turn now to the validation of the strain as mostly uniaxial. For each oxide thickness of the NW ensembles, we calculate the corresponding average strain from the shift of the Raman peaks. Eq. 1 is valid for the GaAs TO mode. In Figure 3c we plot the redshift of the median of the QD and core PL in function of the corresponding strain. We use eq. 1 to calculate the strain from the Raman GaAs TO shifts in Figure 3a. In Ref. 23, the nature and value of the strain is uniform along the NW radial direction. In our case, the data do not seem to indicate otherwise. Therefore, we assume that the strain calculated from the downshift of the GaAs TO peaks corresponds to the strain applied to the QDs as well as to the core. We can calculate a linear t of the redshift in function of the strain (the dash-dotted green and black lines in Figure 3c respectively for the QDs and the core). We obtain that the QD PL shifts by 90 meV/%, while the core by 63 meV/%. The QD redshift is consistent with those already reported for a very similar NWoxide system. 23 It also agrees with theoretical calculations 19 and with the value found in Ref. 19, by experiments with a straining system that guarantees the uniaxial nature of the applied stress. Our QD redshift is slightly more pronounced than the experimental one for GaAs. 19 This dierence may further support our previous considerations on the loss of connement in strained QDs and on the increase in pressure coecient in the AlGaAs matrix. 34,42 Calculations only based on hydrostatic stress 31 lead to inconsistently larger slopes. This comparison discards purely hydrostatic strain in our case and, together with the agreement with the work of Signorello et al., further validates the conclusion that, eectively, also in our system the stress is mainly applied as a component along the NW longitudinal axis. On the other hand, the core PL redshifts less than the theoretical expected value (red dash-

ACS Paragon Plus Environment

7

Nano Letters a) Uncoated

c)

517

//

518 519 520



521 522 523

Raman shift (cm ) -1

b) 527nm SiO 2 -3.8cm-1



524 525

-1

//

Oxide thickness (nm)

d) Downshift (cm )

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

TO modes

526 527 528 529

LO modes

530 531

Raman shift (cm ) -1

532

Figure 4: (a) - (b) Polarization-dependent Raman

533

spectra of uncoated (a) and coated (b) horizontal NWs 534 at 12K. In each panel, the upper curve (red) corresponds 535 to a spectrum taken with both collection and excitation 536 with linear polarization parallel to the NW longitudinal 537 axis, while in the lower curve (black) only the excita- 538 tion polarization is parallel to the NW longitudinal axis and the collection is perpendicular. The vertical dashed 539 lines correspond to the position of the TO peak in the 540 case of parallel collection. (c) Average full width at half 541 maximum (FWHM) of the GaAs TO mode in the Ra- 542 man spectra taken at 12K on horizontal NWs vs oxide 543 thickness. The linear polarization of these spectra was not selected. The sketch shows that the mode linewidth 544 broadens as a consequence of the convolution of the two 545 peaks revealed by polarization-dependent Raman spec- 546 troscopy. (d) Average Raman shift with respect to the 547 uncoated NWs for all the modes at 12K vs oxide thick- 548 ness. The color coding follows the one introduced in Figure 2e. The Raman shift of each mode is referred to 549 550 those of the uncoated NWs as zero. 551 504 505 506 507 508 509 510 511 512 513 514 515 516

552

dotted line in Figure 3c). We mainly ascribe 553 this discrepancy to the large inuence (more 554 details in the Supporting Information) that few 555 twins and WZ segments have on the PL emis556 sion energy. 43 In particular, they are respon557 sible for uncontrolled shifts of the PL emis558 sion energy from the NW core. Together with 559 the PL shift that may be induced by surface 560 charges trapped at the NW-oxide interface dur561 ing PECVD, we prefer to rely on Raman spec562 troscopy to estimate the applied strain. 563 We provide now further support to the highly 564 anisotropic nature of the strain by polarization565

dependent Raman measurements. In Figures 4a and b we respectively show the Raman spectra of an uncoated and a coated NW lying horizontally. We excite the NWs with light linearly polarized along the longitudinal NW axis and alternatively collect the scattered light with a linear-polarization lter aligned parallel or perpendicular to this axis, as sketched in each panel in Figures 4a and b. With this method it is possible to distinguish the dierent contributions of the two GaAs TO modes, because their degeneracy is lifted under anisotropic stress. The two modes are usually named TOS and TOD , 44,45 respectively detected for scattered light with parallel and perpendicular polarization with respect to the NW axis. 19 No splitting is observed between the two congurations for the uncoated NW in Figure 4a. In the case of the coated NWs, the spectrum collected in the perpendicular conguration is clearly shifted with respect to the one collected in the parallel conguration (Figure 4b): the GaAs TO mode in the parallel conguration is at 260.6 cm−1 , while it is at 264.4 cm−1 under perpendicular collection. The dierence (3.8 cm−1 ) indicates a smaller strain in the perpendicular direction. One can therefore state that the applied stress is highly anisotropic and it is mainly applied as uniaxial tension in the direction parallel to the NW longitudinal axis (see the Supporting Information for more measurements). The TO splitting is manifested as broadening in non-polarized Raman spectra. The GaAs TO peak is thus the convolution of the two contributions. The splitting, and thus the full width at half maximum (FWHM) of the convoluted spectra, increases with increasing strain. In Figure 4c, we plot the average FWHM of the GaAs TO modes from non-polarized Raman spectra at 12K in function of the sample oxide thickness. The linewidth broadens with the oxide thickness, in agreement with an increase in anisotropy. In Figure 4d, we plot the Raman shift of all modes from spectra acquired at 12K on horizontal NWs. We use the position of the modes in the uncoated NWs as zero, in order to compare the downshift of the dierent modes in the same plot. As expected, one can observe that

ACS Paragon Plus Environment

8

Page 8 of 16

Page 9 of 16 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

566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614

Nano Letters

in the thickest depositions, the downshift of the 615 LO modes is less and less pronounced with re- 616 spect to the one of the TO modes. 617 In Ref. 19, the authors estimate the Poisson 618 ratio for GaAs NWs, under the assumption that 619 the deformation potentials are the same as in 620 the bulk. They report a slightly smaller value 621 for NWs (0.16±0.04) with respect to the bulk 622 (0.186), although their result is aected by a 623 large experimental error. We use the results 624 from the polarization-dependent Raman spec- 625 tra to deduce the Poisson ratio of GaAs and 626 analyze if it changes with the envelope-induced 627 strain or the nanoscale size of the NWs. We ob- 628 tain an average value of ν = 0.19±0.02, which 629 highlights no major dierences with the bulk 630 (more details on the derivation in the Support- 631 ing Information). 632 Importance of oxide nature. At the origin 633 of strain, the literature 47 typically distinguishes 634 between intrinsic and extrinsic contributions. 635 Among several cases, the rst category includes 636 the structure of the deposited lm, while a typ- 637 ical example of the second category is the ther- 638 mal strain that results from the dierence be- 639 tween the thermal expansion coecient (TEC) 640 of the oxide and the NW. 641 We now provide considerations on the struc- 642 ture of the oxide and its correlation with strain. 643 In Figure 5, we present the data about four dif- 644 ferent depositions as illustrative examples out 645 of ten NWs for each case. In particular, Fig- 646 ures 5a-c show SEM images of array NWs of 647 purely GaAs coated with SiO2 at dierent sub- 648 strate temperatures during the PECVD: 100◦ C 649 in Figure 5a, 300◦ C in Figure 5b and 400◦ C 650 in Figure 5c. By increasing the substrate tem- 651 perature, the surface morphology of the SiO2 652 becomes smoother; as expected, 47 the grain 653 structure is less pronounced and the size of 654 the grains decreases. The TEM micrographs in 655 Figure 1c and d of, respectively, a NW coated 656 at 300◦ C and 100◦ C, conrm this observation. 657 Figures 5e-h report the Raman spectra of the 658 corresponding vertical NWs in Figures 5a-d. 659 For each case we acquired the Raman spec- 660 tra at room temperature (RT) before and after 661 the SiO2 PECVD. In this scattering congura- 662 tion, both the GaAs TO and LO are allowed. 663

In addition, a surface optical (SO) mode appears 46,4850 and splits after the SiO2 deposition. In general, the SO mode shifts because of a change in the dielectric constant at the NW surface. 46,4850 Once coated, in our NWs both the strain and the change in dielectric constant control the SO position. However, there is a unique and continuous NW-oxide interface. Therefore, there is also a unique dielectric constant at the NW surface after coating and the SO split may rather depend on the anisotropic nature of the applied strain. Moving to the analysis of the TO and LO modes, we observe that they all downshift in all the three SiO2 depositions. We report the downshifts of the TO modes in blue in the graphs. The downshifts increase from left (+0.7 cm−1 in Figure 5e) to right (+4.1 cm−1 in Figure 5g). Therefore the tensile strain increases together with the deposition temperature. The downshifts are coherent for all the ten NWs measured in each sample, except for the deposition at 100◦ C. In this case, we observed uctuations and even upshifts of the Raman modes. We highlight that this effect is not related with the oxide thickness. On purpose, the SiO2 thickness of the depositions at 300◦ C and 400◦ C are both about 300 nm and the corresponding Raman shifts similar. The deposition at 100◦ C shows the smallest shift in spite of a larger oxide thickness of about 480 nm. The evolution of the tensile strain with the SiO2 structure indicates a correlation between the two. In particular, the stiness of the SiO2 deposited at lower temperature reduces. As one can infer from the low-density gaps between two grains in Figure 1d, the large grains obtained at lower temperature are loosely connected to each other. In this case, the oxide is easily damaged, for instance during the transfer of the NWs on TEM grids (more details in the Supporting Information). On the contrary, the SiO2 deposited at higher temperatures is more robust. In other words, the SiO2 deposited at 100◦ C lacks the necessary solidity to transfer any signicant strain to the NWs. Through the grain structure, one can therefore control the stiness of the oxide and, in turn, the application of the tensile strain. This possibility demonstrates

ACS Paragon Plus Environment

9

Nano Letters a)

b)

c)

d)

e) 100°C SiO2

f) 300°C SiO2 TO -3.9cm-1 SO

g) 400°C SiO2

h) 300°C TiO2

Uncoated Coated

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

Page 10 of 16

-0.7cm-1 TO SO LO

-4.1cm-1-1 -3.9cm

LO

TO SO

LO

+0.7cm-1 TO

LO

E2H

Figure 5: (a) - (d) From left to right: SEM images of GaAs array NWs respectively coated with SiO2 at 100◦ C,

300◦ C and 400◦ C and TiO2 at 300◦ C. Scale bars are 500 nm. (e) - (h) Raman spectra of the NWs from (a) to (d) in the same order. Black: uncoated NWs. Red: coated NWs. The dashed curves are Lorentzian ts. The TO, LO, 30 EH and surface optical (SO) 46 modes are labeled. The dashed vertical lines in blue correspond to the TO GaAs 2 mode of the uncoated NWs. In blue in the upper left of each plot is the shift of the Raman TO upon coating. 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689

a further degree of freedom in SiO2 PECVD: 690 high-temperature depositions can maximize the 691 applied strain, while a low-temperature depo- 692 sition may be useful for protecting NWs from 693 the environment without changing their optical 694 properties. 695 At the atomic level, the importance of the ox- 696 ide structure for the application of strain gets 697 further support in Ref. 23. Here the authors 698 consider the atom and ion impingement during 699 PECVD and how it forces the atomic bonds of 700 SiO2 in a non-equilibrium conguration. This 701 builds up strain in the oxide itself, that is even- 702 tually transferred to the NW. We found in the 703 literature that this can be controlled through 704 the deposition parameters, such as the plasma 705 frequency during PECVD, 51 but we have not 706 further investigated this technical aspect. The 707 authors of Ref. 23 nd that the built-in strain 708 is as important as the strain derived from the 709 thermal dilation. In particular, the TEC of 710 SiO2 52,53 is smaller than the one of GaAs and 711 AlGaAs 34,54,55 and this agrees with the applica- 712 tion of tensile strain to the NW. However, also 713 in our case the thermal strain cannot account 714 for all the strain we observe. For instance, there 715

is an almost linear correlation between the thermal strain and the temperature gradient from the deposition temperature to RT, but this is not the trend that we observe from the Raman data in Figure 5. Furthermore, we do not record a larger downshift of the Raman peaks when we further decrease the measurement temperature down to 12K (more details in the Supporting Information). Therefore, we conclude that the main working principle of our straining device is the atomicscale internal strain of the oxide, modulated by the oxide stiness at the level of its microstructure. Yet, when the internal strain is set to a smaller contribution by adjusting the parameters of the deposition, the thermal strain may play a major role. We also deposited TiO2 by sputtering in order to further demonstrate the versatility of the same device concept for another deposition technique and another material. We studied an ensemble of 10 NWs before and after the sputtering. As an example, in Figure 5d we show the SEM image of an array NW that we sputtered with TiO2 at 300◦ C and in Figure 5h we compare the Raman spectra acquired at RT

ACS Paragon Plus Environment

10

Page 11 of 16 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

716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762

Nano Letters

before and after the deposition. The Raman 763 modes upshift as a consequence of the deposition, which corresponds to compressive strain. 764 In the Supporting Information, we show the µ- 765 PL response of the NW at RT before and af- 766 ter the TiO2 deposition: the blueshift of the 767 GaAs PL signal further supports the compres- 768 sion of the NW by means of the TiO2 coating. 769 Therefore, we can change the strain from tensile 770 to compressive and we demonstrate this change 771 with a more general approach than in Ref. 22. 772 In conclusion, we presented an ecient and 773 yet simple system to shift the emission of QDs 774 embedded in NWs. The strain can be quantita- 775 tively tuned by the thickness of the deposited 776 oxide. We give robust evidence of the repro- 777 ducibility of the system in spite of the variabil- 778 ity of the emission of the NWs and the QDs. 779 By means of Raman spectroscopy, we provide 780 the rst experimental proof of the presence 781 of tensile strain independently from the mea- 782 surements of the redshift of the NW bandgap. 783 Raman spectroscopy provides quantitative ev- 784 idence to evaluate the amount and nature of 785 the applied strain. This would not be possible 786 only with PL measurements, because the PL 787 emission energy is sensitive to many more ef- 788 fects, such as surface states and polytypism. 43 789 We give an experimental estimation of the ap- 790 plied strain in agreement with its uniaxial na- 791 ture, that we further support by the analy- 792 sis of the polarization-dependent Raman spec- 793 tra. We have validated this method for dier- 794 ent materials and techniques. Concomitantly, 795 the temperature of the deposition brings a fur- 796 ther degree of freedom to the functionalization 797 of the NW-oxide system. This opens the pos- 798 sibility to functionalize the properties of the 799 NWs at the surface while tuning their bandgap. 800 This could open new avenues for instance, in 801 photo-electrochemical applications in which the 802 bandgap of the semiconductor is related to the 803 light absorption but also to the overpotential. 804 Here, the oxide shell would protect the semi- 805 conductor from corrosion and at the same time 806 it would enable the tuning of the overpotential 807 808 of the device by strain. 809 810

Experimental

Nanowire growth We grew the NWs in a DCA

P600 MBE machine through the Ga-assisted method 26,27 on Si (111) substrates. More details on the growth of the self-assembled NWs are available in Ref. 25,27, while the detailed growth process of the arrays NWs is available in Ref. 26. Oxide deposition We coated the NWs with SiO2 by PECVD in a PlasmaLab system 100 PECVD by Oxford Instruments. We used a 400 sccm of N2 98% - SiH4 and 710 sccm of N2 O at a radio frequency power of 20 W. We deposited the TiO2 coating in a Pfeier SPIDER 600 sputtering machine. Morphology and structure We acquired the SEM images in Zeiss MERLIN and Zeiss GEMINI 300 microscopes operated at 3 kV to study the morphology of the NW on their growth substrates. We used FEI Tecnai OSIRIS and FEI Talos microscopes operated at 200 kV on NWs transferred on carbon-coated TEM grids for studying the structure and measuring the thickness of the oxide. By switching to the STEM mode on the FEI Tecnai OSIRIS microscope, we acquired the X-EDS maps. Photoluminescence For the low-temperature µ-PL spectra, we put the NWs in a close-loop liquid-helium cooled cryostat at 4.2K. We used a continuous-wave HeNe laser (emission wavelength = 632.8 nm) as an excitation source. We focused the laser light in a 1µm-large spot on the sample by means of an objective with NA = 0.85. The power density on the samples was in the order of 100W/cm2 . We collected the PL signal through the same objective, dispersed it with a 300l/mm grating and detected it by means of a nitrogen-cooled charge-coupled device (CCD). For the QD PL spectra at RT and the NW-core PL spectra at 12K we used a single-frequency optically-pumped semiconductor laser at 532 nm wavelength (continuous wave) as an excitation source. We used a microscope objective (NA = 0.75) to focus the laser on the NWs in a spot with a diameter of 1µm and a power density of about 103 W/cm2 . We collected the PL signal through the same objective and dispersed it with a 300l/mm grating

ACS Paragon Plus Environment

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

811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847

onto a Peltier-cooled CCD. 857 We analyzed the PL signal from the NW en- 858 sembles by means of an automatized code. We 859 acquired the spectra from 25 NWs for each ox- 860 ide thickness. From each PL spectrum, we cal- 861 culated through the code the median of the 862 emission energy weighted on the intensity of the 863 signal at the dierent energies. We plotted the 864 average of the medians of the 25 spectra. The 865 standard deviation of the average gave the error bars. Raman spectroscopy We collected the lowtemperature Raman spectra by putting the NW on the cold nger of a helium-cooled cryostat at 12K. We used the same setup as the one for the RT PL, with the dierence that we used a power density of about 104 W/cm2 and a triple spectrometer in order to separate the Raman 866 emission from the Rayleigh scattering. A nal 1800l/mm grating dispersed the light on a 867 Nitrogen-cooled CCD. We realized the measure868 ments in backscattering conguration with the 869 NWs transferred in horizontal position on a Si 870 wafer. For the polarization-dependent Raman 871 spectra, we ltered the laser light with a linear 872 polarizer and aligned the selected polarization 873 to the NW main axis by means of a rotator. We used a second polarizer to lter the scattered light. 874 We collected the RT Raman spectra by means of a commercial Renishaw inVia Raman micro- 875 scope using the same excitation source as in 876 the low-temperature measurements. We used a grating with 1800l/mm to disperse the light, 877 a notch lter to discard the Rayleigh scattering 878 of the laser and a Peltier-cooled CCD to detect 879 880 the signal. 881

848 849 850 851 852 853 854 855 856

Author information

882 883

PL spectroscopy on 884 single NWs was performed by AG and LF. PL 885 spectroscopy on NW ensembles along with the 886 statistical analysis was realized by AG. Ra887 man spectroscopy on NW ensembles and single 888 NWs along with the analysis was realized by 889 LF. TEM measurements and analysis were performed by LF. PRG performed the deposition Author contributions

by sputtering, while AG and LF by PECVD. Strain analysis and modelling was performed by LF. NW ordered arrays were obtained by WK, JVP MF, HP and LG, while the ensembles by GT. AFiM conceived and designed the experiments and supervised the project. AG and LF made the gures. AG, LF and AFiM wrote the manuscript. All authors discussed the results and commented on the manuscript. The authors thank funding through SNF by ERAnet-Russia grant nr IZLRZ2_163861, by the NCCR QSIT and project nr 200021_169908 and the H2020 program through the project INDEED. The authors declare no competing nancial interest. Acknowledgement

Supporting Information Available The Supporting Information is available free of charge on the ACS Publications website. It contains the TEM micrographs of several depositions in dierent conditions, the calibration of the measured-vs-nominal oxide thickness and additional Raman and PL data.

References (1) Heiss, M. et al. Nature Mater. 439444.

2013,

12,

(2) Mancini, L.; Fontana, Y.; Conesa-Boj, S.; Blum, I.; Vurpillot, F.; Francaviglia, L.; Russo-Averchi, E.; Heiss, M.; Arbiol, J.; Morral, A. F. i.; Rigutti, L. Appl. Phys. Lett. 2014, 105, . (3) Messing, M. E.; Wong-Leung, J.; Zanolli, Z.; Joyce, H. J.; Tan, H. H.; Gao, Q.; Wallenberg, L. R.; Johansson, J.; Jagadish, C. Nano Letters 2011, 11, 38993905. (4) Dick, K. A.; Thelander, C.; Samuelson, L.; Caro, P. Nano Letters 2010, 10, 3494 3499.

ACS Paragon Plus Environment

12

Page 12 of 16

Page 13 of 16 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

890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930

Nano Letters

(5) Borgström, M. T.; Zwiller, V.; Müller, E.; 931 (17) Wang, Y.-B.; Wang, L.-F.; Joyce, H. J.; Imamoglu, A. Nano Lett. 2005, 5, 1439 932 Gao, Q.; Liao, X.-Z.; Mai, Y.-W.; 1443. 933 Tan, H. H.; Zou, J.; Ringer, S. P.; Gao, H.934 J.; Jagadish, C. Advanced Materials 2011, (6) Reimer, M. E.; Bulgarini, G.; Akopian, N.; 935 23, 13561360. Hocevar, M.; Bavinck, M. B.; Verheijen, M. A.; Bakkers, E. P. A. M.; Kouwen- 936 (18) Montinaro, M.; Wüst, G.; Munsch, M.; Fontana, Y.; Russo-Averchi, E.; Heiss, M.; hoven, L. P.; Zwiller, V. Nature Commu- 937 938 Fontcuberta i Morral, A.; Warburnications 2012, 737. 939 ton, R. J.; Poggio, M. Nano Lett. 2014, (7) Claudon, J.; Gregersen, N.; Lalanne, P.; 940 14, 44544460. Gérard, J.-M. ChemPhysChem 2013, 14, 941 (19) Signorello, G.; Karg, S.; Björk, M. T.; 23932402. 942 Gotsmann, B.; Riel, H. Nano Letters (8) van Dam, D.; Abujetas, D. R.; Paniagua- 943 2013, 13, 917924. Domínguez, R.; Sánchez-Gil, J. A.; G.; Lörtscher, E.; Bakkers, E. P. A. M.; Haverkort, J. E. M.; 944 (20) Signorello, Khomyakov, P.; Karg, S.; Dheeraj, D.; Gómez Rivas, J. Nano Letters 2015, 15, 945 946 Gotsmann, B.; Weman, H.; Riel, H. 45574563. 947 Nature Communications 2014, 5 . (9) Bulgarini, G.; Reimer, M. E.; Hocevar, M.; Bakkers, E. P. A. M.; Kouwenhoven, L. P.; 948 (21) Singh, U.; Adenwalla, S. Nanotechnology 949 2015, 26, 255707. Zwiller, V. Nature Photonics 2012, 6 . (10) Mokkapati, S.; Saxena, D.; Tan, H. H.; Ja- 950 (22) Bouwes Bavinck, M.; Zieli«ski, M.; 951 Witek, B. J.; Zehender, T.; Bakkers, E. gadish, C. Scientic Reports 2015, 5 . 952 P. A. M.; Zwiller, V. Nano Letters 2012, 12, 62066211. (11) Minot, E. D.; Kelkensberg, F.; 953 van Kouwen, M.; van Dam, J. A.; 954 (23) Stepanov, P.; Elzo-Aizarna, M.; Bleuse, J.; Kouwenhoven, L. P.; Zwiller, V.; 955 Malik, N. S.; Curé, Y.; Gautier, E.; FavreBorgström, M. T.; Wunnicke, O.; 956 Nicolin, V.; Gérard, J.-M.; Claudon, J. Verheijen, M. A.; Bakkers, E. P. A. M. 957 Nano Letters 2016, 16, 32153220. Nano Lett. 2007, 7, 367371. 958 (24) Matteini, F.; Tütüncüo§lu, G.; Potts, H.; (12) Yan, R.; Gargas, D.; Yang, P. Nature Pho959 Jabeen, F.; Fontcuberta i Morral, A. Crystonics 2009, 3, 569. 960 tal Growth & Design 2015, 15, 31053109. (13) Hennessy, K.; Högerle, C.; Hu, E.; 961 (25) Francaviglia, L.; Fontana, Y.; ConesaBadolato, A.; Imamo§lu, A. Applied 962 Boj, S.; Tütüncüo§lu, G.; Duchêne, L.; Physics Letters 2006, 89, 041118. 963 Tanasescu, M. B.; Matteini, F.; i Morral, A. F. Applied Physics Letters 2015, (14) Freeman, R.; Willner, I. Chem. Soc. Rev. 964 965 107, 033106. 2012, 41, 40674085. Kim, W.; (15) Akopian, N.; Wang, L.; Rastelli, A.; 966 (26) Vukajlovic-Plestina, J.; Dubrovski, V. G.; Tütüncüo§lu, G.; Schmidt, O. G.; Zwiller, V. Nature Pho- 967 968 Lagier, M.; Potts, H.; Friedl, M.; Fontcutonics 2011, 5, 230. 969 berta i Morral, A. Nano Letters 2017, (16) Kumar, S.; Trotta, R.; Zallo, E.; 970 17, 41014108. Plumhof, J. D.; Atkinson, P.; Rastelli, A.; Schmidt, O. G. Applied Physics Letters 971 (27) Colombo, C.; Spirkoska, D.; Frimmer, M.; 972 Abstreiter, G.; Fontcuberta i Morral, A. 2011, 99, 161118. 973 Phys. Rev. B 2008, 77, 155326. ACS Paragon Plus Environment

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

974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015

Page 14 of 16

(28) Uccelli, E.; Arbiol, J.; Magen, C.;1016 (38) Zhu, J.; Yu, Z.; Burkhard, G. F.; Hsu, C.Krogstrup, P.; Russo-Averchi, E.;1017 M.; Connor, S. T.; Xu, Y.; Wang, Q.; Heiss, M.; Mugny, G.; Morier-Genoud, F.;1018 McGehee, M.; Fan, S.; Cui, Y. Nano LetNygård, J.; Morante, J. R.; Fontcu-1019 ters 2009, 9, 279282. berta i Morral, A. Nano Lett. 2011, 11, 1020 (39) Schliwa, A.; Winkelnkemper, M.; Bim38273832. 1021 berg, D. Phys. Rev. B 2007, 76, 205324. (29) Welber, B.; Cardona, M.; Kim, C. K.; Rodriguez, S. Phys. Rev. B 1975, 12, 57291022 (40) Jöns, K. D.; Hafenbrak, R.; Singh, R.; 1023 Ding, F.; Plumhof, J. D.; Rastelli, A.; 5738. 1024 Schmidt, O. G.; Bester, G.; Michler, P. (30) Zardo, I.; Conesa-Boj, S.; Peiro, F.;1025 Phys. Rev. Lett. 2011, 107, 217402. Morante, J. R.; Arbiol, J.; Uccelli, E.; Abstreiter, G.; Fontcuberta i Morral, A.1026 (41) Kuklewicz, C. E.; Malein, R. N. E.; 1027 Petro, P. M.; Gerardot, B. D. Nano LetPhys. Rev. B 2009, 80, 245324. 1028 ters 2012, 12, 37613765. (31) Zardo, I.; Yazji, S.; Marini, C.; Uccelli, E.; Fontcuberta i Morral, A.; Abstreiter, G.;1029 (42) Lifshitz, N.; Jayaraman, A.; Logan, R. A.; Maines, R. G. Phys. Rev. B 1979, 20, Postorino, P. ACS Nano 2012, 6, 32841030 1031 23982400. 3291. (32) Yu, P. Y.; Cardona, M. Fundamentals of1032 (43) Heiss, M.; Conesa-Boj, S.; Ren, J.; Tseng, H. H.; Gali, A.; Rudolph, A.; semiconductors : physics and materials1033 1034 Uccelli, E.; Peiró, F.; Morante, J. R.; properties ; Springer, 2010. 1035 Schuh, D.; Reiger, E.; Kaxiras, E.; Ar(33) Amaduzzi, F.; Alarcón-Lladó, E.; Russo-1036 biol, J.; Fontcuberta I Morral, A. Physical Averchi, E.; Matteini, F.; Heiss, M.;1037 Review B - Condensed Matter and MateTütüncüo§lu, G.; Conesa-Boj, S.; de la1038 rials Physics 2011, 83, 110. Mata, M.; Arbiol, J.; i Morral, A. F. JourAnastassakis, E.; nal of Applied Physics 2014, 116, 184303.1039 (44) Wickboldt, P.; 1040 Sauer, R.; Cardona, M. Physical Re(34) Adachi, S. Journal of Applied Physics1041 view B 1987, 35, 13621368. 1985, 58, R1R29. 1042 (45) Cerdeira, F.; Buchenauer, C. J.; Pol(35) Rüer, D.; Slot, M.; Huber, R.;1043 lak, F. H.; Cardona, M. Physical Review Schwarze, T.; Heimbach, F.;1044 B 1972, 5, 580593. Tütüncüo§lu, G.; Matteini, F.; RussoAverchi, E.; Kovács, A.; Dunin-1045 (46) Adu, K.; Xiong, Q.; Gutierrez, H.; Chen, G.; Eklund, P. Applied Physics A Borkowski, R.; Zamani, R. R.;1046 1047 2006, 85, 287. Morante, J. R.; Arbiol, J.; i Morral, A. F.; Grundler, D. APL Materials1048 (47) Freund, L. B.; Suresh, S. Thin Film 2014, 2, 076112. 1049 Materials: Stress, Defect Formation and

Surface Evolution ; Cambridge University (36) Hutchinson, W.; Suo, Z. Mixed mode1050 1051 Press, 2004. cracking in layered materials ; Advances in Applied Mechanics; Academic Press,1052 (48) Gupta, R.; Xiong, Q.; Mahan, G. D.; Ek1991; Vol. 29. 1053 lund, P. C. Nano Letters 2003, 3, 1745 1750. (37) Hua, B.; Motohisa, J.; Ding, Y.; Hara, S.;1054 Fukui, T. Applied Physics Letters 91, 131112.

2007,

1055 1056

(49) Krahne, R.; Chilla, G.; Schüller, C.; Carbone, L.; Kudera, S.; Mannarini, G.;

ACS Paragon Plus Environment

14

Page 15 of 16 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

1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076

Nano Letters

Manna, L.; Heitmann, D.; Cingolani, R. Nano Letters 2006, 6, 478482. (50) Spirkoska, D.; Abstreiter, G.; Fontcuberta I Morral, A. Nanotechnology 2008, 19 . (51) Tarraf, A.; Daleiden, J.; Irmer, S.; Prasai, D.; Hillmer, H. Journal of Micromechanics and Microengineering 2004, 14, 317. (52) Tada, H.; Kumpel, A. E.; Lathrop, R. E.; Slanina, J. B.; Nieva, P.; Zavracky, P.; Miaoulis, I. N.; Wong, P. Y. Journal of Applied Physics 2000, 87, 41894193. (53) Jansen, F.; Machonkin, M. A.; Palmieri, N.; Kuhman, D. Journal of Applied Physics 1987, 62, 47324736. (54) Straumanis, M. E.; Krumme, J. P.; Rubenstein, M. Journal of The Electrochemical Society 1967, 114, 640641. (55) Ettenberg, M.; Pa, R. J. Journal of Applied Physics 1970, 41, 39263927.

ACS Paragon Plus Environment

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

1077

Graphical TOC Entry QDs

AlGaAs

SiO2

//

GaAs

1078



Raman shift (cm-1)

ACS Paragon Plus Environment

16

Page 16 of 16