Strong Size Dependency on the Carrier and Photon Dynamics in

(4, 7, 10, 14, 21, 29, 30) The degree of carrier and optical confinement, ... peak transition energy shift, full-width at half-maxima in photoluminesc...
0 downloads 0 Views 2MB Size
Subscriber access provided by UNIV OF NEWCASTLE

Communication

Strong size dependency on the carrier and photon dynamics in InGaN/GaN single nanowalls determined using photoluminescence and ultrafast transient absorption spectroscopy Shonal Chouksey, Sandeep Sankaranarayanan, Vikas Pendem, Pratim K Saha, Swaroop Ganguly, and Dipankar Saha Nano Lett., Just Accepted Manuscript • DOI: 10.1021/acs.nanolett.7b00970 • Publication Date (Web): 22 Jul 2017 Downloaded from http://pubs.acs.org on July 23, 2017

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

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

Page 1 of 23

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

Nano Letters

Strong size dependency on the carrier and photon dynamics in InGaN/GaN single nanowalls determined using photoluminescence and ultrafast transient absorption spectroscopy S. Chouksey, S. Sankaranarayanan, V. Pendem, P. K. Saha, S. Ganguly, and D. Saha Applied Quantum Mechanics Laboratory, Indian Institute of Technology Bombay, Powai, Mumbai – 400076, India. Phone: +91-2225767443, Email: [email protected] KEYWORDS: nanowall, carrier/photon dynamics, photoluminescence, transient absorption spectroscopy, exciton binding energy, quantum confinement Abstract: Here, we have demonstrated strong size dependency of quasi-equilibrium and nonequilibrium carrier and photon dynamics in InGaN/GaN single nanowalls using photoluminescence and transient absorption spectroscopy. We demonstrate that twodimensional carrier confinement, strain relaxation and modified density of states lead to a reduced Stokes shift, smaller full width at half-maxima, increased exciton binding energy, and reduced non-radiative recombination. The ultrafast transient spectroscopy shows that carrier capture is a two-step process dominated by optical phonons, and carrier-carrier scattering in succession. The carrier capture is a strongly size dependent process and becomes slower due to modulation of the density of available states for progressively decreasing nanowall sizes. The slowest process is the electron-hole recombination, which is also extremely size-dependent and the rate increases by almost an order of magnitude in comparison to that of quantum-well structures. Electron-hole wavefunction overlap and

1 ACS Paragon Plus Environment

Nano Letters

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

modified density of states are among the key aspects in determining all the properties of these structures.

2 ACS Paragon Plus Environment

Page 2 of 23

Page 3 of 23

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

Nano Letters

Light emitting diodes (LEDs) with InGaN/GaN as the active layer has been the subject of immense studies for various applications [1-6]. Nanowire based LEDs are among the primary choices for various advantages including better light extraction efficiency due to larger surface to volume ratio, additional confinement supporting stronger electron-hole wavefunction overlap, and strain relaxation resulting in reduced quantum confined Stark effect [714]. Nanowires may have higher photon reabsorption reducing extraction efficiency, however, the deficiency can be overcome at higher excitation power and carrier injection [15]. Moreover, these structures are relatively free from defects, which further improves the performance of these devices [16-19]. Lateral nanowires and nanowalls, in particular, belong to an important class due to the ease of fabrication and a precise process control of their positions and dimensions [20-22]. Furthermore, nanowalls, in general, are very interesting as a host device to observe various quantum mechanical phenomena at room temperature [2328]. It is necessary to understand the basic physics of these structures in order to maximize the efficiency of the nanowall based optoelectronic devices. Quasi-equilibrium and nonequilibrium ultra-fast transient spectroscopy of carriers and photons have been proved to be essential for the same. Several studies have been performed for the quasi-equilibrium case using temperature dependent photoluminescence measurements [4,7,10,14,21,29,30]. The degree of carrier and optical confinement, temperature dependent radiative and non-radiative processes, and exciton transitions have been explored through experiments under quasiequilibrium condition. Reports on the ultrafast spectroscopy under non-equilibrium condition for the InGaN/GaN nanowalls are limited and the understanding is at the nascent stage [31,32]. The fundamental mechanism of carrier capture into the bound states followed by the decay through various mechanisms is at the heart of this study. Here, we report twodimensional size dependent carrier and photon dynamics in InGaN/GaN lateral nanowalls

3 ACS Paragon Plus Environment

Nano Letters

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

both for the quasi-equilibrium and highly non-equilibrium conditions in a common framework. The macroscopic manifestation of the evolving one dimensional (1-D) density of states is found in terms of the quasi-equilibrium peak transition energy shift, full-width at half-maxima in photoluminescence, and Arrhenius activation energy, which are then correlated to common microscopic phenomena. Further, ultrafast transient spectroscopy with ~50 fs resolution indicates that the perceived understanding of electron-electron (e-e) carrier scattering as the dominant mechanism for carrier capture with the hot phonon effect prolonging the process [32,33] does not provide adequate picture at very small dimensions. It is found that the carrier capture is actually a two-step process, which is uniquely carried out by polar optical phonon and fluence dependent e-e scattering process, which is strongly dependent on the nanowall dimension. While larger surface to volume ratio increases extraction efficiency, surface recombination is found to produce a major leakage path in the non-radiative carrier decay process, particularly at very small dimensions. The carrier decay rate through radiative recombination is found to improve by an order of magnitude due to increasing carrier confinement. All the quasi-equilibrium and non-equilibrium processes are found to be strongly dependent on the nanowall dimensions. Figures 1(a)-(d) show schematics of the metal-organic chemical vapor deposited (MOCVD) InGaN/GaN single quantum-well heterostructure used in this work, along with the fabricated nanowalls. A 25 nm thick low temperature buffer layer is grown first on a 430 µm thick cplane sapphire substrate followed by a 1.24 µm thick un-doped GaN layer. A 3.64 µm thick Si-doped (2×1018 cm−3) n-type GaN layer is deposited on this structure followed by the growth of the active region well layers. The well consists of a 3 nm thick InGaN (14%) layer sandwiched between two 12 nm thick GaN layers. The structure is covered with a 40 nm thick layer consisting of a graded AlxGa1−xN:Mg, layer, where x varies from 0.2 to 0 from bottom to top. The graded AlGaN/AlGaN superlattice suppresses the electron leakage from

4 ACS Paragon Plus Environment

Page 4 of 23

Page 5 of 23

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

Nano Letters

the active region. This is followed by the deposition of a 100 nm thick Mg doped p-type GaN layer (6×1019 cm−3). A heavily doped p-type GaN layer (1.2×1020 cm−3) of thickness 10 nm is finally deposited on this heterostructure. The n and p dopant densities were determined by Hall measurements on control samples grown under the same condition as that of the n and p regions of the heterostructure. The energy-band diagram of the heterostructure as determined by self-consistently solving Schrodinger and Poisson equations is shown in Fig. 1(e). The nanowalls of various dimensions are fabricated by a combination of dry and wet chemical etching processes [22,34,35]. An oxide mask is patterned on the heterostructure using electron beam lithography (EBL) as shown in Fig. 1(b). The sample is then dry etched by Inductively Coupled Plasma Reactive Ion Etching (ICP-RIE) with chlorine and argon gas plasma at 70W (Fig. 1(c)) [36]. The resultant profile height is confirmed to be 300 nm by profilometer ensuring that the active quantum-well region is completely etched away. This is followed by wet etching in boiling phosphoric acid for 5 s to remove the RIE induced damages as well as reducing the device dimensions further. The device is finally treated in buffered hydrofluoric acid for 1 minute to remove any residual oxide (Fig. 1(d)). A scanning electron microscope (SEM) of a 20 nm nanowall is shown in Fig. 1(f). Lateral nanowalls with 20 and 50 nm width and 5 µm length are fabricated for this work. The sidewalls of the nanowalls are found to be extremely smooth. We have also considered a quantum-well sample as the control heterostructure. While single nanowalls are used in this study, we have also fabricated an array of nanowalls, where we have used a 500 µm × 500 µm oxide mask with areal density 1.6×107 cm-2 for dry etching. An SEM image of such an array is shown in Fig. 1(g) indicating the reproducibility and repeatability of our process and also its technological importance. All the nanowalls are found to be identical in nature without any noticeable size variation. The same individual nanowalls were used for all the measurements.

5 ACS Paragon Plus Environment

Nano Letters

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

The quasi-equilibrium carrier and photons dynamics are studied through temperature and power dependent photoluminescence (PL) measurements on a single nanowall. The PL measurements are performed with a 325 nm He-Cd laser source and an air cooled CCD detector (see supplementary document). Figure 2(a) shows the typical room temperature PL spectra of a 20, 50 nm nanowall and quantum well samples. The peak positions are found to be at 455, 459 and 465 nm for the 20 nm, 50 nm nanowalls, and quantum-well samples, respectively by fitting with Lorentzian curve. It is apparent that the position of the peak suffers a blue shift for the increasing confinement of the nanowalls, which is also shown in the inset. The increase in the bound-state energy for both electrons and holes due to confinement, thus, overcomes the increasing exciton binding energy owing to larger electron and hole wave-function overlap, which is later validated through modeling. The PL peak wavelength and FWHM as a function of temperature are shown in Figs. 2(b) and (c), respectively, for all the samples. The carrier localization due to non-uniformity, which is prevalent in quantum-well structures, gives rise to the non-monotonic change in the peak wavelength. The peak positon initially shows a blue shift of ~2 nm, which is follows by a red-shift of ~3 nm. This is found to be largely absent for the nanowall samples indicating better uniformity and lesser defects. The highest quantum confined sample is also found to show the least FWHM of 14 nm at room temperature. The lowest FWHM is consistently observed for entire temperature range of 10 to 300 K. The activation energy is determined from the Arrhenius plot as shown in Fig. 2(d). Two activation energies at high (EA1) and low (EA2) temperatures match experimental data for all the samples. The activation energy increases with increasing confinement, with the highest activation energy observed for 20 nm nanowall as EA1 = 134 and EA2 =10 meV. This is also a signature of the higher exciton binding energy and better confinement for decreasing dimensionality of the nanowall. This confirms that the non-radiative centers are mostly removed in the nanowalls and larger carrier

6 ACS Paragon Plus Environment

Page 6 of 23

Page 7 of 23

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

Nano Letters

confinement deters parallel non-radiative recombination of the electron-hole pairs in the active region. The evolution of FWHM is understood from the density of states for the 1-D structure, which has a singularity at the effective band-edge and it is accommodating a large number of electrons in comparison to that for higher energy levels. This leads to the progressively smaller FWHM for 1-D like nanowalls. The simulated electron distribution over energy levels for ideal 3-D, 2-D and 1-D cases is shown in Fig. 2(f) for a fixed electron density of 1017 cm-3 in the active region. Lesser number of defects and better uniformity for In in the active region further augments to this cause. The shift in the PL peak position with confinement is largely determined by the effective bandgap and the exciton binding energy, where both the parameters significantly depend on the quantum confinement. We have determined the effective transition energy by self-consistently solving two-dimensional Schrodinger and Poisson equations and estimating the exciton binding energy using the hydrogenic model as [37] ቌ−

ħ2



߂2‫ ݖ‬+ ܸ݂݂݁ ሺ‫ݖ‬ሻቍ ߶ሺ‫ݖ‬ሻ = ሺ‫ ݐ݋ݐܧ‬− ‫ ݁ܧ‬− ‫ܧ‬ℎ ሻ߶ሺ‫ݖ‬ሻ

where, ‫ܧ‬௘ and ‫ܧ‬௛ are single electron and hole energies, ħ is the reduced Planck’s constant,

Etot is the total energy of the system, µ is the reduced effective mass, and ܸ௘௙௙ is the effective exciton potential expressed as ܸ݊,݈ ݂݂݁ ሺ‫ݖ‬ሻ =

−݁2 4ߨɛ

න ݀‫ݔ݀ ݁ݔ‬ℎ ݀‫ݕ ݀ ݁ݕ‬ℎ

2

ቚ߰݊,݈,݇‫ ݖ‬ሺ‫ ݁ݔ‬, ‫ ݁ݕ‬ሻ߰݊,݈,݇‫ ݖ‬ሺ‫ݔ‬ℎ , ‫ݕ‬ℎ ሻቚ

ටሺ‫ ݁ݔ‬− ‫ݔ‬ℎ ሻ2 + ሺ‫ ݁ݕ‬− ‫ݕ‬ℎ ሻ2 + ‫ݖ‬2

The estimated effective potential, which in essence determines the exciton binding energy, is shown in Fig. 2(f) along with the 1s exciton levels. It is interesting to note that there is a significant increase in the binding energy due to 2-D confinement in spite of the fact that the lateral confinement is larger than the effective Bohr radius of ~3 nm. This theoretical 7 ACS Paragon Plus Environment

Nano Letters

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

observation is in harmony with the higher activation energy observed earlier for the nanowalls [4]. Barring the small discrepancy, the experimental PL characteristics matches well with the theoretical estimation for all the cases both qualitatively and quantitatively. The power dependent PL characteristics of the nanowalls and the quantum well sample are shown in Figs. 3(a)-(c). The shift in the PL peak position and FWHM with power density are shown in Figs. 3(d) and (e), respectively. The peak wavelength is found to be consistently shorter and the FWHM smaller for increasing confinement of the nanowall. We observe that the quantum-well sample does not show any significant change in the peak position with the increase in power from 1 to 10 kW/cm2. The nanowall samples show a very small red-shift (~2 nm) in addition to a slight decrease in FWHM. It may be noted that the nanowall samples show a sharp cut-off for longer wavelengths in comparison to that of quantum-well samples. While a longer tail in PL for the quantum-well samples is primarily due to the defects and In composition variation, the absence of it for the nanowalls further corroborates our earlier observations. In contrast to the low energy tail, the high energy tail for the nanowalls is much longer, which is again due to easy band-filling for higher energy states. Figure 3(f) shows the measured internal quantum efficiency (IQE) at 300 K determined as the ratio of the room temperature and 10 K PL intensities. The peak IQE is found to be the highest for the 20 nm nanowall sample. To ensure that the non-radiative centers are frozen out at low temperature, integrated PL intensity is plotted as a function of excitation energy at 10 K for all the samples in Fig. 3(g). A near unity slope indicates that non-radiative recombination is minimal at this low temperature [38]. The transient carrier and photon dynamics is investigated using transient absorption spectroscopy. A high intensity femtosecond duration (~50 fs) laser pump-pulse is made incident on the sample, which takes the system to a highly non-equilibrium state. From the same source, a coherent low intensity femtosecond duration laser white probe-pulse is

8 ACS Paragon Plus Environment

Page 8 of 23

Page 9 of 23

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

Nano Letters

generated, which is used to probe the dynamics. A tunable delay stage is used to delay the probe-pulse with respect to the pump-pulse. The absorption is captured as a function of time and wavelength though differential measurements. The detailed experimental set-up is illustrated in the supplementary document. We have used an excitation wavelength of 325 nm, which is the same as used in PL. Figures 4(a)-(c) show the differential absorption characteristics as a function of probe-wavelength for various delay times for the 20 nm, 50 nm nanowalls, and quantum well samples. A negative value indicates band-filling as the dominant mechanism for differential absorption spectra. A significant contrast is be noted that the peak absorption edge suffers a red shift of ~25 nm with increasing delay time for the QW sample, whereas both the nanowall samples peak position do not show any significant shift. The pump laser reduces the polarization in the quantum well due to screening effect of the generated carriers. This, in effect, reduces the quantum confined Stark effect and the bands become flatter leading to higher transition energy. However, as the carriers decay with time the polarization field takes over and the transition energy again reduces. However, the nanowalls samples do not show significant change in the differential absorption spectra with time and the change is the least significant for the 20 nm nanowall. This indicates that the nanowalls must be significantly strain relaxed. In addition, the absorption spectral width is observed to be extremely narrow (10 nm) for the nanowall samples in comparison to that of quantum well (30 nm), indicating nanowall samples have less number of defects, which corroborates our earlier observation for lower FWHM in PL. The smaller number of defects may be ascribed to the wet-etching by boiling phosphoric acid used for the fabrication of the nanowalls . A careful observation of both the nanowall samples indicate additional absorption peaks for shorter wavelengths at 438 and 435 nm for 50 nm and 20 nm samples, respectively, which is completely absent for the quantum well samples. The peak is more prominent and energetically separated at higher energy for the 20 nm sample. While 2s excitons can give

9 ACS Paragon Plus Environment

Nano Letters

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

rise to another absorption peak, it is less likely to form at high pump density due to manybody effects [39]. Hence, the secondary peak is probably due to the second sub-band filling at high pump power. The separation of peaks is a direct measure of the approximate sub-band separations, which are found to be 38 meV and 25 meV for 20 nm and 50 nm samples, respectively. The time evolution of transient absorption is found to be associated with one rise time (t0) and two decay time constants as shown for a typical wavelength of 450 nm in Fig. 5(a). The risetime constant, which is associated with the carrier capture, is very small (t0