Communication
On the optimization of current injection in AlGaInP core-shell nanowire light-emitting diodes Pyry Kivisaari, Alexander Berg, Mohammad Karimi, Kristian Storm, Steven Limpert, Jani Oksanen, Lars Samuelson, Håkan Jan Pettersson, and Magnus T Borgström Nano Lett., Just Accepted Manuscript • Publication Date (Web): 23 May 2017 Downloaded from http://pubs.acs.org on May 27, 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 32
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
On the optimization of current injection in AlGaInP core-shell nanowire light-emitting diodes Pyry Kivisaari,† Alexander Berg,† Mohammad Karimi,†,‡ Kristian Storm,† Steven Limpert,† Jani Oksanen,§ Lars Samuelson,† Håkan Pettersson,†,‡ and M. T. Borgström*,† †
Solid State Physics and NanoLund, Lund University, P.O. Box 118, SE-221 00, Lund, Sweden
‡
Laboratory of Mathematics, Physics and Electrical Engineering, Halmstad University, P.O. Box 823, SE-301 18 Halmstad, Sweden §
Engineered Nanosystems Group, Aalto University, P.O. Box 12200, FI-00076 Aalto, Finland
Keywords: Nanowires, light-emitting diodes, ideality factor, leakage current, full device simulation, emission enhancement
Core-shell nanowires offer great potential to enhance the efficiency of light-emitting diodes (LEDs) and expand the attainable wavelength range of LEDs over the whole visible
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
spectrum. Additionally, nanowire LEDs can offer both improved light extraction and emission enhancement if the diameter of the wires is not larger than half the emission wavelength (λ/2). However, AlGaInP nanowire LEDs have so far failed to match the high efficiencies of traditional planar technologies, and the parameters limiting the efficiency remain unidentified. In this work, we show by experimental and theoretical studies that the small nanowire dimensions required for efficient light extraction and emission enhancement facilitate significant loss currents which result in a low efficiency in radial NW LEDs in particular. To this end, we fabricate AlGaInP core-shell nanowire LEDs where the nanowire diameter is roughly equal to λ/2, and we find that both a large loss current and a large contact resistance are present in the samples. To investigate the significant loss current observed in the experiments in more detail, we carry out device simulations accounting for the full 3D nanowire geometry. According to the simulations, the low efficiency of radial AlGaInP nanowire LEDs can be explained by a substantial hole leakage to the outer barrier layer due to the small layer thicknesses and the close proximity of the shell contact. Using further simulations, we propose modifications to the epitaxial structure to eliminate such leakage currents and to increase the efficiency to near unity without sacrificing the λ/2 upper limit of the nanowire diameter. To gain a better insight of the device physics, we introduce an optical output measurement technique to estimate an ideality factor that is only dependent on the quasi-Fermi level separation in the LED. The results show ideality factors in the range of 1-2 around the maximum LED efficiency even in the presence of a very large voltage loss, indicating that the technique is especially attractive for measuring nanowire LEDs at an early stage of development before electrical contacts have been optimized. The presented results and characterization techniques form a 2 ACS Paragon Plus Environment
Page 2 of 32
Page 3 of 32
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
basis of how to simultaneously optimize the electrical and optical efficiency of core-shell nanowire LEDs, paving way to nanowire light emitters that make true use of larger-thanunity Purcell factors and the consequently enhanced spontaneous emission.
Nanowires (NWs) have raised considerable interest to be used for efficient and functional next-generation light-emitting diodes (LEDs).1 Compared to planar structures, NWs allow much more freedom in bandgap tuning. Additionally, they can be grown on substrates with different lattice constants because of their small footprint,2,3 avoiding the necessity of using costly techniques like wafer-bonding.4 Furthermore, particularly the use of a radial core-shell geometry paves the way to higher efficiencies in NW array LEDs due to a larger integrated volume of the light-emitting layer as compared to equally large planar structures with a similarly thick active layer. As the increased volume leads to a larger current density for a given carrier density, the use of NWs could alleviate the efficiency droop of typical planar LEDs,5 occurring essentially as a consequence of the increasing carrier density. Most importantly, however, NW LEDs can be designed to utilize nanophotonic effects for optimized extraction efficiency and emission enhancement over planar LEDs, and this is especially promising if the diameter of the individual NWs is less than half the wavelength of the emitted light (λ/2).6 The special properties and new emission wavelengths enabled by NWs make them very attractive for improving present LEDs and creating new LED applications, but for such goals it is crucial to reach a high NW LED device efficiency. However, all radial AlGaInP NW LEDs reported so far have shown relatively poor efficiency. Gutsche et al. fabricated n-GaAs/InGaP/p-GaAs core-multishell NW nip-structures and obtained an EL 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
peak at 1.4 eV.7 Svensson et al. reported electrolumincence (EL) measurements on radial GaP/GaAs/InGaP NW LEDs with a peak at around 1.5 eV.8 They estimated the external quantum efficiency (EQE) of their devices to be between 5×10−5 and 8×10−4, and attributed this to possible surface defects on the GaAs NW core as well as to rotational twins in the NW.8 However, previous reports have not analyzed the low efficiency and the physical processes behind it in detail. Therefore, to enable improved NW LEDs and new applications, in-depth studies are needed to understand the physics behind their poor efficiency and how it can be improved by identifying the efficiency-limiting loss processes in a combination of experimental and theoretical efforts. We recently reported on the epitaxial growth and materials characterization of radial AlGaInP core-shell NW LED structures emitting in the red wavelength range with dimensions matching the requirements for high light extraction and enhanced visible light emission.9 In this work we fabricate red-emitting AlGaInP core-shell heterostructure LEDs and study their electro-optical properties using experiments and simulations. Based on our results, we identify large loss currents that can explain the so far poor efficiencies reported for core-shell NW LEDs and propose how to make large improvements in their device performance. To explain the large loss current observed in the experiments, we carry out device modeling accounting for the full 3D radial NW geometry. The modeling shows that the small dimensions of the NWs and the close proximity of the shell contact to the active region (AR) facilitate a notable hole leakage current, even if all the foreseeable parasitic loss processes are neglected. Finally, we analyze how large-bandgap capping layers affect the device performance and how they can be used to significantly improve the efficiency of radial core-shell NW LEDs. We also demonstrate how EL measurements can be used to 4 ACS Paragon Plus Environment
Page 4 of 32
Page 5 of 32
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
estimate the ideality factor that only depends on the local biasing conditions of the LED even in the presence of large contact resistances. Our ideality factor measurement method is especially attractive for the study and development of nanoscale LED prototypes at their early development stages, when the unoptimized electrical contacts can still entail significant voltage losses that prevent a conventional ideality factor measurement. The AlGaInP core-shell NWs (see Figure 1a) were grown under similar growth conditions as previously reported,9 with the difference that the radial p-GaInP buffer layer growth time was extended from 2 min to 4 min in order to completely fill the hole in the SiNx growth mask before the growth of the other layers. To study if the presently used AlGaInP barriers provide sufficient carrier confinement to the i-GaInP QW, a series of five growth runs was made with growth times of 5, 15, 30, 60, and 300 s to control the QW thickness. An SEM image of the NWs before processing is shown in Figure 1a, and Figures 1b-c show the schematics of the processed NW device. Note that the diameter of the NW light emitters corresponds to recommendations given in Ref.6, i.e., it is slightly smaller than λ/2 required for both high extraction efficiency and emission enhancement, particularly if also the emission towards the substrate can be exploited by using the bottom contact as a mirror. After growing the NW arrays, we processed 100 µm × 100 µm large LED devices comprising 10,000 NWs (Figures 1b-c). Atomic layer deposition (ALD) was used to deposit 50 nm silicon dioxide (SiO2) and 5 nm aluminum oxide (Al2O3) on the samples to avoid a short-circuit between the NWs due to the minor planar growth on the mask. The Al2O3 layer was used to improve resist adhesion.10 The resist was spin-coated and etched back down to about half of the NW length to expose the oxide at the upper part of the NWs, 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
after which they were etched in HF:H2O (1:10). To remove any residual native oxide on the NW sidewalls, the samples were etched in a H2SO4:H2O solution (1:10) directly before sputtering of 150 nm indium-tin-oxide (ITO) on the NWs to fabricate transparent electrical contacts to the NW shells. The device area was defined by UV lithography and resist development techniques. The bond pads facilitating electrical probe measurements were defined by an additional UV lithography step and subsequent metal evaporation of 15 nm titanium (Ti) and 400 nm gold (Au) directly on the ITO layer covering the resist next to the device area, followed by lift-off. The p-side of the device was contacted via the backside of the substrate attached with silver glue to a copper coin.9-12 A Cascade 11000B probe station with a Keithley 4200 semiconductor characterization system was used for measuring room temperature IV characteristics of the devices. The voltage was swept from -2 V to +2 V. For room temperature EL measurements the emitted light was collected by a lens, fed into a fiber and finally dispersed by a spectrometer and recorded by a charge-coupled device (CCD). The relative position between the lens and the sample was identical for different samples in order to compare the light intensity. Note that for most of the light to be emitted by the QW, the AlGaInP barriers should prevent the flow of electrons towards the p-type GaInP layer and the flow of holes towards the n-type GaInP layer. As the doped GaInP layers have the same bandgap as the QW, carriers diffusing over the AlGaInP barriers can recombine and produce photons with roughly the same photon energy (with the quantization energy subtracted) as carriers in the QW. Therefore to also account for light emitted by the doped layers, we define the AR here so that it includes the QW and the doped GaInP layers of the same composition. In planar LEDs, a design with the same composition in the AR as in the doped layers would be 6 ACS Paragon Plus Environment
Page 6 of 32
Page 7 of 32
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
problematic predominantly because the thicker doped layers would absorb a large fraction of the emitted photons. In radial NWs, the doped layers are much thinner than in typical planar LEDs, and therefore we do not expect significant absorption in the NW. Nevertheless, advanced optoelectronic models accounting also for photon recycling are needed to study in detail how many photons would be lost due to reabsorption in the NWs.
Figure 1. (a) SEM image of the as-grown radial GaInP/AlGaInP NWs. The scale bar is 500 nm and the image was recorded at 30° tilt. (b) Schematics of a fully processed device showing only one NW (dimensions not to scale). (c) 3D illustration of a fully processed
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
Page 8 of 32
device showing the NW array. The NWs are about 1.8 µm long and about 300 nm in diameter. The oxide consists of 50 nm SiO2 and 5 nm Al2O3. The GaInP layers and AlGaInP
barriers
were
grown
with
nominal
compositions
Ga0.47In0.53P
and
(Al0.4Ga0.6)0.47In0.53P, respectively.
To lay the foundations for the analysis performed in the coming sections, we present the theoretical models used in the analysis and simulations. The device operation of coreshell NW LEDs can be modeled using the drift-diffusion model given by (see, e.g., Ref.13,14 for details) ∇ ∙ −∇ = − + − ∇ ∙ = ∇ ∙ − ∇ = ∇ ∙ = ∇ ∙ − ∇ = − ,
(1)
where ε is the static permittivity, φ is the electrostatic field, q is the elementary charge, n and p are the electron and hole densities, Nd and Na are the ionized donor and acceptor densities, Jn and Jp are the electron and hole current densities, µn and µp are the electron and hole mobilities. φn and φp are the quasi-Fermi potentials of the conduction and valence bands, and R is the net recombination rate density. Setting the boundary conditions for free surfaces and contacts similarly as in Ref.14, we solve these equations for the core-shell structure from which device characteristics and light output power can be extracted and compared to experiments. The net recombination rate density R in Eq. (1) consists of radiative recombination (spontaneous emission), non-radiative Shockley-Read-Hall (SRH) recombination, Auger recombination, and surface recombination in the case of unpassivated surfaces.
8 ACS Paragon Plus Environment
Page 9 of 32
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
Conventionally, the net recombination rate density is expressed as (see, e.g., Ref.15 for details) = + + =
!!"#
+ $ + % + & − '" ,
(2)
where A, B and C are the SRH, radiative and Auger recombination coefficients, respectively, and ni is the intrinsic carrier concentration. Here, we have assumed that the SRH lifetimes are equal for electrons and holes and that the electron and hole densities are equal to ni in the case where the Fermi level coincides with the trap level responsible for SRH recombination. This corresponds to the dominant trap level to be located roughly in the middle of the bandgap. Furthermore, we have assumed that the radiative recombination coefficient B only includes photons that do not generate new electron-hole pairs through reabsorption in the AR (photon recycling), and that the Auger coefficient is the same for eeh and ehh Auger processes. These assumptions are generally used since they are simple and typically allow a reasonable fitting of the simulations to experiments. In this work, we use the model described above to simulate current transport in a single NW. Since the NW is practically axially symmetric, we solve Eqs. (1)-(2) in axially symmetric form by using 3D cylindrical coordinates where all the derivatives are zero along the angular coordinate. The full list of all parameters is displayed in Table S1 (Supporting Information). Some parameters such as permittivity, effective masses, and recombination parameters were taken from literature that describes planar (AlxGa1-x)yIn1-yP structures lattice-matched to GaAs where y is typically about 0.5.16 We assumed relatively low ionized dopant densities of 1017 cm-3 in the simulations for all the layers except for the strongly doped p-GaP substrate because of preliminary Hall measurements (not shown),
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
Page 10 of 32
performed as described in Refs.17-19 Ideal ohmic contacts were assumed in the simulations for the substrate and the outer shell of the NW, and a fairly high extraction efficiency of 80 % was also assumed. Due to the inevitable axial growth on top of the NWs during shell growth (see Figure 1b), we assumed no contact between ITO and the top of the NW and treated the NW top surface as insulated. Before analyzing the experimental current-voltage characteristics of our diodes, we note that most of the mechanisms i (i referring to a single process, e.g., SRH, radiative or Auger recombination, or leakage current) that typically generate current in LEDs can be approximated by an exponential current law, and therefore the total current is given by20 = ∑' ',) *exp *
./
0,# 12
4 − 14 ≈ ) 7exp 7 3
./
0 12 3
8 − 18
(3)
where J is the current density, Ji,0 a constant factor which depends on the process i generating the current, V the applied bias, nV,i the ideality factor which also depends on the process i generating the current, kB the Boltzmann constant and T the temperature. To enable easier manipulation of the equations, we ignore the factor 1 in the equation from now on as it only has an effect at extremely low current densities. Based on Eq. (2), ideality factors nV,i of 2, 1, and 2/3 can be derived for recombination currents generated by SRH, radiative and Auger recombination in the AR, respectively. Furthermore, the ideality factor for diffusive diode current is 1. To enable an estimation of the ideality factor from measurements, the sum is typically approximated by a single exponential term as in Eq. (3), where J0 and nV in general depend on the applied bias. The value measured for nV is essentially determined by the dominant process(es) in the sum, and therefore the value measured for nV can be used to help identifying the dominant current-generating process(es)
10 ACS Paragon Plus Environment
Page 11 of 32
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
at a specific applied bias. The ideality factor in Eq. (3) as a function of current density can be estimated by20 / =
.
12 3
9 :;
7 8 /
=
.
12 3
7
?$ '" exp 71 38 2
(5)
where χ is the number of extracted photons divided by the total number of emitted photons (or more accurately, the fraction of net bimolecular recombination events resulting in a detected photon), d the effective thickness of the AR, B the net radiative recombination coefficient and U the effective voltage, or more precisely the quasi-Fermi level separation in the AR. U can be solved from the equation above as
11 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
A=
12 3
D
Page 12 of 32
ln *E.FG 4 .
.
#
(6)
Using this estimated AR voltage instead of the applied bias in Eq. (4), the ideality factor can be estimated from U following a similar procedure as in Eq. (4), resulting in @ =
.
12
9 :;
7 8 3 @
.
(7)
To show that the ideality factor in Eq. (7) does not depend on any fitting parameters and does not require an absolute measurement of the light output power, we first manipulate the derivative dJ/dU using the chain rule as 9
@
9 D
9
.
= D @ = D = 1 3. 2
(8)
If we plug this derivative into Eq. (7), we get the ideality factor as 9
9 :;
@ = D 7D8
