Fast and Sensitive Colloidal Quantum Dot Mid ... - ACS Publications

Jul 5, 2018 - only 30% transparent in the MWIR,8 with a 50 nm indium tin .... diode equation:20. = −. −. +. −. I I. eV. RI nk T. V. R I. R exp. ...
1 downloads 0 Views 1MB Size
Subscriber access provided by TUFTS UNIV

Article

Fast and Sensitive Colloidal Quantum Dot Mid-Wave Infrared Photodetectors Matthew M Ackerman, Xin Tang, and Philippe Guyot-Sionnest ACS Nano, Just Accepted Manuscript • DOI: 10.1021/acsnano.8b03425 • Publication Date (Web): 05 Jul 2018 Downloaded from http://pubs.acs.org on July 7, 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 28 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

ACS Nano

Fast and Sensitive Colloidal Quantum Dot Mid-Wave Infrared Photodetectors Matthew M. Ackerman,† Xin Tang,† and Philippe Guyot-Sionnest* James Franck Institute, 929 E. 57th Street, The University of Chicago, Chicago, IL 60637 *[email protected] †Equally contributed to this work.

ACS Paragon Plus Environment

ACS Nano 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 2 of 28

ABSTRACT Colloidal quantum dots (CQDs) with a band gap tunable in the mid-wave infrared (MWIR) provide a cheap alternative to the epitaxial commercial photodetectors such as HgCdTe (MCT) and InSb. Photoconductive HgTe CQD devices have demonstrated the potential of CQDs for MWIR photodetection, but face limitations in speed and sensitivity. Recently, a proof-of-concept HgTe photovoltaic (PV) detector was realized, achieving background-limited infrared photodetection (BLIP) at cryogenic temperatures. Using a modified PV device architecture, we report up to two order of magnitude improvement in the sensitivity of the HgTe CQD photodetectors. A solid-state cation exchange method was introduced during device fabrication to chemically modify the interface potential, leading to an order of magnitude improvement of external quantum efficiency at room temperature. At 230 K, the HgTe CQD photodetectors reported here achieve a sensitivity of 109 Jones with a cutoff wavelength between 4-5 µm, which is comparable to the commercial photodetectors. In addition to the chemical treatment, a thinfilm interference structure was devised using an optical spacer to achieve near unity internal quantum efficiency (IQE) upon reducing the operating temperature. The enhanced sensitivity of the HgTe CQD photodetectors reported here should motivate interest in a cheap, solutionprocessed MWIR photodetector for applications extending beyond research and military defense. KEYWORDS quantum dots, mid-wave infrared, photodetector, HgTe, specific detectivity

2 ACS Paragon Plus Environment

Page 3 of 28 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

ACS Nano

Thermal imaging in the mid-wave infrared (MWIR) currently relies primarily on wafer grown materials like mercury cadmium telluride (MCT), indium antimonide (InSb), and superlattice structures; however, their high fabrication and processing costs limit applications mostly to defense and research.1 Yet, there are many opportunities for cameras with sensitive, fast, and high resolution thermal imaging, such as for large-scale pollution monitoring and autonomous vehicles. The challenge is to realize much cheaper materials with fast response and high sensitivity at temperatures easily accessible to thermoelectric cooling, while also being compatible with simple deposition processes on silicon integrated circuits. Over the past decade, much of the research on solution-processed colloidal quantum dots (CQDs) has been motivated by improving solar cell performance in competition with already cheap technologies.2 However, the tunable optical properties of CQDs offer advantages at longer wavelengths where disruptive advances are likely, given the very high cost of the existing midwave infrared (MWIR) technologies.1 In the past few years, HgTe CQDs have been explored for infrared detection in the near, mid-wave, and long-wave IR based on photoconductors,3–5 photovoltaics,6–8 and phototransistors.9,10 Fast MWIR video imaging was readily achieved using simple drop cast HgTe CQD solutions on commercial read-out integrated circuits, showing a clear path to cheap high resolution thermal imaging if the CQD performances were very significantly improved.5 As a method for improving sensitivity of the detectors, a photovoltaic (PV) architecture was devised to reduce the 1/f noise limit of the photoconductor devices.11 For comparing different devices, the specific detectivity given by Equation (1), is a measure of the signal-to-noise ratio for a given incident power and normalized to the detector area with units of √    , or Jones. 3 ACS Paragon Plus Environment

ACS Nano 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 4 of 28

(1)

N is the spectral noise density in A Hz-1/2, S is the photocurrent in A, P is the incident power in W, and Ad is the area of the detector in cm2. Prior HgTe CQD PV devices achieved backgroundlimited performance at 5 µm with specific detectivity D* of 3.8×1010 Jones at cryogenic temperatures and approximately 108 Jones at 230 K.8 However , the maximum external quantum efficiency, EQE, was only 2% at 80 K. The device structures were based on the concepts of electron (e.g. NiCr) and hole (i.e. Ag2Te nanoparticles with Ag paint) injecting electrodes with operation similar to a Schottky junction. This article demonstrates improvements in the sensitivity and efficiency of photovoltaic MWIR detectors by developing a pn junction with enhanced light collection. The introduction of a solidstate cation exchange method for the Ag2Te nanoparticle layer vastly improves the device efficiency and sensitivity. Chemical, electronic, and optical characterization provide insights into the source of the improvement by investigating the effect of doping concentration and junction built-in potential on the performance. Finally, the introduction of an optical spacer in the device stack leads to further improved quantum efficiency, and a comparison of the device performance with commercial MWIR detectors is provided.

RESULTS and DISCUSSION HgTe CQD Photodetector Fabrication The previous8 understanding of the HgTe CQD PV structures was based on the use of electron and hole injecting contacts, as for Schottky barrier CQD PV.12,13 Thus, we initially explored

4 ACS Paragon Plus Environment

Page 5 of 28 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

ACS Nano

different electrode materials, characterized by their resistance and work function. This led to replacing the 5 nm NiCr layer, which was only 30% transparent in the MWIR,8 by a 50 nm indium tin oxide (ITO) layer which was more transparent. The work function of ITO between 4.5-4.7 eV14 was compatible with the conduction band position measured by electrochemistry for HgTe CQDs at 4.5 eV,15 such that barriers to electron transport should be minimized, making ITO an appropriate electron contact. As the top metal contact, silver paint was replaced by evaporated gold, using a shadow mask to precisely define the device areas. This gold electrode provided better resistance to thermal cycling and functions as a back reflector to the infrared radiation. For the underlying p-type contact layer, the initial motivation to use Ag2Te nanocrystals was inspired by the use of Ag as a p-type dopant in epitaxial HgCdTe.8,16 In addition, the p-type layer acted as an hole transport layer. The construction of the HgTe CQDs infrared detector is illustrated in Fig.1a. On a sapphire substrate, ITO film is deposited by sputtering and annealed to increase conductivity. Then, HgTe CQDs are deposited layer-by-layer by drop-casting, and each layer is rinsed by a solution of 1,2ethanedithiol (EdT), hydrochloric acid (HCl), and isopropyl alcohol (IPA) (1:1:20 by volume) solution. The final thickness of the HgTe CQDs layer is around 400 nm. A solution of ~ 10 nm diameter Ag2Te nanoparticles is then spin-coated over the HgTe layer. Finally, a gold layer is evaporated through a shadow mask to pattern the top electrode. All device fabrication was performed under ambient conditions. The cross-sectional scanning electron microscopy (SEM) image of the fabricated detector is shown in Fig.1b. Using the fabrication strategy previously described, D* was increased to 1011 Jones and the responsivity was improved by a factor of ~ 10 at low temperature compared to previous devices reported in reference 8; this was largely due to replacing the NiCr contact by the ITO. However, 5 ACS Paragon Plus Environment

ACS Nano 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 6 of 28

a large fraction of these devices showed poor performances and no rectification even at low temperature. Initially, this was attributed to poor film fabrication thought to lead to shorting, but this was not supported by optical and electron microscopy. We then concluded that the Ag2Te layer provided a hole injection layer but was insufficient to form a reliable built-in potential due to the sample-to-sample variation in the doping level of the narrow gap HgTe CQDs. To overcome the challenge of an insufficient junction, a solid-state cation exchange step was added to the previously described sample fabrication, specifically, exposing the sample to a 10 mM HgCl2 methanol solution following the Ag2Te nanoparticle deposition. The expectation was that HgCl2 would partially convert Ag2Te to HgTe CQDs to create an adjacent p-doped HgTe CQD layer. The HgCl2 solution treatment led to immediate improvement of the devices with a nearly 100% reliable fabrication of rectifying devices with much improved performances. All detectors described in the following were made using this key step in the fabrication.

Fig.1 Photodetector Design. (a) Schematic of the detector structure. (b) Cross-sectional SEM of the fabricated detector.

6 ACS Paragon Plus Environment

Page 7 of 28 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

ACS Nano

The detectors were characterized between 85 K and 295 K, under vacuum in a cryostat. A typical current density J- bias voltage V curve of the HgCl2-treated HgTe CQDs MWIR detector at an operation temperature of 85 K is shown in Fig.2a. The devices exhibit background limited performance (BLIP) up to 140 K. Under illumination with a calibrated 600 ºC blackbody and operating temperature of 85 K, the responsivity, in units of A W-1,  = 

(2)

and specific detectivity were  = 0.38 A/W and D* =1.2 ×1011 Jones, respectively. The detector showed temperature-dependent cut-off wavelength from 3.8 µm at 290 K to 4.8 µm at 85 K, covering the MWIR range, as shown in Fig.2b.

Fig.2. Performance of the HgCl2-treated HgTe CQDs MWIR detectors. (a) Current density as a function of bias voltage V under dark condition, background radiation, and 600 ºC blackbody radiation at an operation temperature of 85 K. The inset figure zooms on the dark 7 ACS Paragon Plus Environment

ACS Nano 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 28

current (closed shield) and background current (open shield). (b) Photocurrent spectrum at 85 K, 235 K, and 290 K. (c) (T). (d) D* (T) from the HgCl2-treated HgTe CQDs MWIR detectors (●) and the HgTe PV MWIR detector (▼) in ref.8.

The temperature-dependent responsivity is shown in Fig.2c. The peak responsivity of 0.56 A/W at 160 K is about 7-fold larger than previously reported HgTe CQDs PV MWIR detectors.8 The temperature-dependent D* is shown in Fig 2d. Below 100 K, D* exceeds 1011 Jones and D* is 1010 Jones at 200 K. At 230 K, where thermoelectric cooling is practical, D* is 109 Jones, exceeding microbolometer performances.17 The D* of previously reported HgTe CQDs MWIR detectors without HgCl2 treatment is also shown for comparison. The exposure of the Ag2Te nanoparticle films to HgCl2 was therefore a key step and we attempted to clarify the process. Ag2Te nanoparticles are known to react extensively with HgCl2 in a cation exchange to form HgTe nanocrystals;18,19 however, small amount of Ag+ may remain as surface species or dopants. Infrared spectra of Ag2Te nanoparticles films before and after exposure to HgCl2 show an absorption edge attributable to HgTe CQDs, albeit with a long absorption tail (Supplementary Fig.S1b). Photoluminescence (PL) spectra of Ag2Te nanoparticles before and after HgCl2 treatment show the emergence of infrared PL not attributable to Ag2Te, further supporting conversion to HgTe CQDs (Supplementary Fig. S1c). X-ray photoelectron spectroscopy (XPS) shows that an Ag2Te CQDs film treated with HgCl2 solution has an atomic ratio of Ag:Cl:Hg:Te of ~4:4:1:1, which was expected for HgTe conversion and suggests AgCl resided on the surface of HgTe (Supplementary Fig.S3). However, the direct exposure of devices to soluble Ag ions from a solution of AgNO3 led to systematic device failure with ohmic I(V) response and no photocurrent at room temperature. We

8 ACS Paragon Plus Environment

Page 9 of 28 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

ACS Nano

concluded that soluble Ag ions diffusing through the entirety of the active layer would indeed remove any built-in potential. We therefore propose that the low solubility of AgCl that was released in the conversion of Ag2Te is important for limiting the diffusion of Ag ions such that only the most adjacent HgTe CQDs were affected. However, the low solubility may also leave unwanted excess AgCl on the device surface. The Ag2Te thickness was then optimized to reduce residual surface AgCl that, presumably acting as an insulator, hindered carrier transport and collection efficiency. The performance was then further improved with a best device having a responsivity of 0.74 A/W responsivity (Supplementary Fig.S5) with a cutoff wavelength of 5 µm. Secondary ion mass spectrometry (SIMS) was used to obtain a depth profile analysis of the elemental distribution within the device structure (Supplementary Fig. S4). However, thermal diffusion and irradiation diffusion of Ag in the porous HgTe CQD film during SIMS limited the amount of useful information. One conclusion was that the Ag concentration at the surface decreased with exposure to a HgCl2 solution, which was consistent with a reaction to HgTe and increased Ag diffusion into the HgTe film. The role of the Ag2Te nanoparticles preparation on device performance was then further explored. Details of the Ag2Te CQDs synthesis and characterization are available in the Methods and supporting information. Interestingly, we found that smaller Ag2Te nanoparticles of ~ 5 nm yield inactive devices. From many possibilities, we propose that this may be due to their conversion into smaller and more insulating HgTe nanoparticles. Additionally, we speculate that the easier diffusion of the smaller nanoparticles into the HgTe CQD films could produce an effect similar to the AgNO3 treatment previously discussed. Other information obtained by SIMS was a 10-fold increase of chloride in the HgTe films following the HgCl2 solution treatment. Direct exposure of the HgTe film to

9 ACS Paragon Plus Environment

ACS Nano 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 28

HgCl2 solution in the absence of Ag2Te was explored, but these devices failed, producing no photocurrent at room temperature. At this stage, we have no direct evidence for the depth or concentration of the Ag ions, but all tests were consistent with the hypothesis that the Ag2Te nanoparticle layer and the HgCl2 treatment provide a source of localized Ag ions that act as a p-doping diffusion layer for the HgTe CQD film, affording a well-defined junction potential. Importantly, the reliable device fabrication allowed us to make many further improvements and additional characterizations detailed below. To further characterize the electrical properties of the devices, the temperature-dependent dark currents were measured (Fig. 3a) and fitted to a modified Shockley diode equation:20  =   

 !"# $% &'( )

* − 1- +

!"# $ "/

(3)

I0 is the reverse saturation current, e is the elementary charge, Rs is the series resistance, n is the ideality factor, V is the external voltage, T is the operation temperature, kB is Boltzmann constant, and R0 is the shunt resistance. The temperature dependence of the R0A product is shown in Fig.3b. Activation energies, EA, of 143 ± 24 meV and 76 ± 18 meV for HgCl2-treated and untreated devices, respectively, were obtained from linear regression of the data between 200 K and 300 K. Typically, the dark current mechanism in a MWIR photodiode is dominated by either diffusion of minority carriers across the depletion region or generation-recombination of thermally generated carriers within the depletion region; the former varies as 1/ni2 and the latter 34

as 1/ni, where 01 2  − * is the intrinsic carrier concentration and Eg is the band gap.17 5' ) (

Since EA is approximately half the band gap (0.25 < Eg < 0.3 eV), the HgCl2-treated device appears dominated by generation-recombination current at high temperatures. At temperatures 10 ACS Paragon Plus Environment

Page 11 of 28 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

ACS Nano

below 70 K, for both the HgCl2-treated and untreated device, the R0A product becomes weakly temperature dependent. This is attributed to leakage currents and non-radiative pathways, also observed in bulk HgCdTe diodes17 and epitaxial quantum dot infrared photodetectors (QDIPs).21 For an ideal pn junction, Voc should be equivalent to Eg at low temperature with a linear decrease with temperature given by,20 678 =

34 



'( ) 

90

: ; &