Hot injection synthesis of colloidal nanocrystals
Chemically synthesized colloidal nanocrystals have become an exciting class of materials with precise control over a wide range of size, shape, and composition. The solution-processibility of colloidal nanocrystals generates tremendous technological opportunities as they can be deposited with simple methods such as spin-coating, spray-coating and dip-coating, or patterning techniques including inkjet printing and nanoimprint lithography. As a result, nanocrystals are ideal building blocks for future low-cost electronics,13 optoelectronics,14 and photonics.15
Synthesis of colloidal nanocrystals with controlled size distribution is the first step for the utilization of their size dependent properties. The milestone for the synthesis of high quality colloidal nanocrystals was achieved in the early 1990s by the group of Brus16 and Bawendi2. Using a “hot injection” method, they introduced organometallic precursors into high boiling point coordinating solvents. It was the first time when II-VI CdE (E = S, Se or Te) QDs were obtained with nearly monodisperse size and shape, and resultant sharp optical (absorption and emission) features at room temperature. This finding has set the basis for the rational synthetic schemes of colloidal nanocrystals over the past four decades.
The cation (or anion) precursor, ligands, and high boiling point coordinating solvent are mixed in a three-neck flask connected to Schlenk line. The mixture is kept under vacuum at an elevated temperature to remove the oxygen, water, and impurities.
Hot injection. After degassing, the atmosphere is switched to an inert gas (Ar or N2) to further increase the temperature. When the temperature is stabilized at the reaction temperature, the anion (or cation) precursor is rapidly injected into the reaction bath. The nucleation is initiated upon rapid injection due to the saturation, and soon terminated due to the temperature drop after the addition of room temperature precursor. The temporal separation between nucleation and growth is thus achieved and ensures the monodisperse size and shape of the nanocrystals.
According to the target material system and size, as well as the reactivity of the precursor, the reaction temperature varies. For example, the synthesis of cadmium chalcogenides requires relatively high reaction temperature of around 250 °C, whereas for the synthesis of mercury chalcogenides, the temperature is usually below 100 °C. The size of the nanocrystal can also be tuned by the duration of reaction.
It is also worth noting that, in general, the cation precursor is injected prior to the anion precursor, whereas in some cases the sequence is reversed. For instance, during the synthesis of silver chalcogenides, the silver precursor is injected into a flask containing chalcogenides precursor. Because, in the absence of any chalcogenide, silver nanocrystals will be formed.17
Quenching of the reaction. To stop the growth of the nanocrystals, excessive of ligands are generally injected into the system to avoid the precursors from accessing to the nanocrystals, and the flask is cooled by air flux or by a water or ice bath.
To extract the nanocrystals from the reaction products, a purification process is conducted. Typically, a polar solvent such as ethanol, methanol, or acetone, is added to the reaction mixture. The change of the polarity of the solvent leads to the flocculation of the nanocrystals, leaving the excess ligands and unreacted precursors in the solvent, enabling the precipitation of nanocrystals by centrifugation. The precipitated nanocrystals are then dispersed in non-polar solvents such as hexane, toluene, or chloroform. The purification is conducted at least twice before storing the clean nanocrystals in a nonpolar solvent.
Ligands of nanocrystals
A typical CQD consists of an inorganic nanocrystal core capped by long-alkyl-chain ligands, see Figure I.4 a. Figure I.4 b shows a TEM image of CdSe/CdS CQDs with a diameter of around 9 nm, revealing a high degree of monodispersity, which is mandatory to resolve the discrete energy levels of a CQD assembly and resultant optical properties.
Although the nanocrystalline core drives the target optoelectronic properties of the CQDs, their surface ligands are paid substantial attention owing to the large surface /volume ratio of nanomaterials. The commonly used ligands for colloidal nanocrystals are hydrocarbon molecules with anchoring end groups. The chain of the molecules is usually long with 12 to 18 carbons, and the end functional groups are usually thiols, amines, carboxylic acid, and phosphines. There are multiple roles played by the ligands throughout the synthesis, processing, and application:
The nonpolar nature of the long alkane ligands ensures the colloidal stability of the nanocrystals in a nonpolar solvent, which is critical for solution-processable fabrication.
During the growth, the ligands rapidly adsorb and desorb from the surface of the growing nanocrystals, which regulate the addition and the removal of the atoms to the crystalline surface, slowing down the growth rate and in turn circumventing the aggregation and enabling fine control over the growth.3
rthermore, some capping ligands can selectively bond to certain facets of the nanocrystals and in turn terminate the growth of the facet18. It is widely used for the synthesis of anisotropic shapes such as nanorods3 and nanoplatelets4.
The capping ligand can passivate the electronic trap states induced by the dangling bonds on the QD surface, preserving the photoluminescence and exciton lifetime.
The length of the ligands determines the interparticle spacing in a nanocrystal solid. The native long-chain ligand massively hindering their electrical coupling. Ligand exchange toward shorter species is needed for application-targeted high-mobility NC films.
Heterostructure of nanocrystals
Along with the study of the single-component nanocrystals, the synthesis of heterostructure nanocrystals, achieved by epitaxially growing an inorganic shell onto the core nanocrystal, has been developed.5,19 This is first motivated by the potential that the heterostructures, beyond single-component counterparts, can bring novel electrical and optical properties and carrier behaviors. From the application point of view, the inorganic shell can better passivate the unsaturated bonds of the core nanocrystal surface than the ligands, enhancing the photoluminescence; it can also work as a physical barrier, making the optical-active core less sensitive to the environmental changes and surface chemistry. This robustness brought by shelling is critical for their integration to a device, where the materials will be inevitably exposed to harsh conditions such as strong electric field, significant heat, strong photo-irradiation and the solvents.20 All the above conditions can be challenging for single-component nanocrystals to maintain their optical properties, especially emission properties.
The carrier localization and the optoelectronic properties depend heavily on the relative position of the conduction and valence band of the core/shell nanocrystal components. Thus the core/shell nanocrystals can be categorized into type I and type II band configurations according to their band offset.11
As in a typical type I CdSe/ZnS QD, the band edge of the CdSe core is fully included in that of the ZnS shell, confining the electrons and holes in the core (Figure I.5 a), thus facilitating the radiative recombination of the excitons. Indeed, after Hines and Guyot-Sionnest published reports of ZnS-capped CdSe QDs showing luminescence up to 50% PLQY,5 considerable efforts were devoted to optimizing QD structures to achieve intense and narrow emissions 21–23, which has paved the way for the synthesis of nanocrystals with near-unity efficiency and size-tunable emission for display. It is worth noting that the lattice mismatch of the heterostructure materials should be treated carefully to avoid the interfacial defects. The common solution to release the strain and to maintain high PLQY of the heterostructure is to alloy the shell or to build a chemical gradient of core/shell materials.24
In a typical type II band alignment (Figure I.5 c), the offset of conduction bands and valence bands are of the same sign. CdTe/CdSe core/shell nanocrystals,25 for example, its electrons are localized in the shell materials with a lower conduction band while the hole wave function is restrained in the CdTe core. This spatial separation of electron and hole wave function reduces the recombination rates of the excitons, thereby prolonging the PL decay time. This type II heterostructure also allows access to optical transition energies that are not restricted to band-gap energies and can emit at lower energies than the band gaps of comprising materials.25
In between the carrier-localized type I and carrier-separated type II, there is a situation defined as quasi type II, where one carrier is confined in one material whereas the other can move across the entire heterostructure. This requires the conduction band or the valence band of the constituent materials close in energy. The most popular manifestation of quasi type II band alignment is CdSe/CdS nanocrystals,26 in which the holes are localized in the CdSe core and the electrons delocalizing in the two materials. This band alignment can provide high photoluminescence quantum yield as type I heterostructure for light emission. One of the interesting characterizations of this materials is that the emission red shifts with the increase of shell thickness, demonstrating the delocalization of the electrons.
Regarding the different carrier localization and the consequent interesting optical properties brought by different band alignment of the heterostructures, it is critical to engineer the core/shell materials for the design of optoelectronic devices. For example, the type I core/shell heterostructure is beneficial for high PLQY and stability required by light emission applications, whereas the type II structure can facilitate the charge carrier separation, which is preferable for photovoltaic devices and photodiodes.
Shape control of Nanocrystals
Ever since the establishment of hot injection synthesis of near-spherical CQDs, the synthesis of anisotropic nanocrystals with various shapes and dimensionality has been followed. In 2000, Peng et al. reported the first anisotropic “quantum rods” and demonstrated their directional emission.3 It was until 2008 that the zinc blende CdSe nanoplatelets (NPLs), with one dimension much smaller than the other two, was discovered, which presents extremely narrow thickness-dependent emission.27
In this part, I will focus on colloidal nanoplates, also called colloidal quantum well (CQW), as they have the electronic properties of the 2D quantum well grown by molecular beam epitaxy. Colloidal NPLs with their unique growth mechanism, allows a large lateral extension, while the confinement direction is flat with atomic precision, see the scheme of a CdSe NPL with 3 monolayers of Se atoms and 4 monolayers of Cd covering the surface in Figure I.6 a, and a TEM of CdSe NPL in Figure I.6 b. As a result, the excitons in the NPLs can move freely in the 2D plane but are strongly confined in the thickness direction (1-2 nm). For sphere QDs, their size is more or less continuous, which inevitably induces inhomogeneous broadening of the optical features, whereas the atomically flat NPL presents extremely narrow linewidth of absorption and emission (below 10 nm) due to high monodispersity. It has been reported that the linewidth of NPLs in the solution is the same as that of the single NPL, demonstrating that there is no inhomogeneous broadening of the NPL batch.28
So far, among all known nanomaterials, NPLs provide the highest color purity owing to the thickness control down to the atomic level, making them the most promising nanocrystal for high-quality light emitting applications. Moreover, the 2D structure with an in-plane 2D optical dipole can result in polarized emission, which can be harnessed to improve the extraction efficiency in a planar LED device .29,30
Although the single content nanoplatelets has already been interesting for light emission device, there are several drawbacks: (1) NPLs with only integral monolayers of atoms can be obtained, limiting their size-tunability in the confinement direction and consequently their emission wavelengths. (2) Like all the single component colloidal nanocrystals, the dangling bonds on the surface of NPLs generate trap states and damages their PLQY. Also, they are sensitive to the change in the environment medium and photo-oxidation and show poor optoelectronic stability, hindering their integration to operational devices.
To overcome the above-mentioned limitation, the core/shell heterostructures of NPLs have also been developed.31 Figure I.6 c and d respectively show the cartoon and the TEM image of a core/shell CdSe/CdS NPLs. It can be clearly observed that both CdSe core and ZnS shell can be grown with atom-resolved thickness, and there is a flat interface between the core and the shell, as indicated by the image contrast due to their different atomic density. In the early stage, the core/shell structure is obtained via colloidal atomic layer deposition (c-ALD) 32. In a typical c-ALD process, the cationic and anionic precursors are successively added into the reaction system. Each reaction step is self-limited by the surface binding sites, which allows the removal of excess precursors after each reaction, preventing their reaction with the following regent. In this manner, CdS shell with a thickness of up to 7 monolayers is grown on the CdSe NPLs core, emitting at 665 nm with a narrow linewidth of 20 nm.
The c-ALD approach, conducted at room temperature, can only provide core/shell NPLs with moderate PLQY (typically below 50%) and stability. The exploration of high-PLQY core/shell NPLs for high-performance light emitting devices has been continued and a breakthrough was reported in 201633 by adopting hot injection to the shell growth. CdSe/ZnS NPLs with reproducible near-unity PLQY were reported by Demir’s group in 2019.34 Soon after, the same group demonstrated a NPL-based LED with the record-high external quantum efficiency of 19.2% 35, close to what has been achieved by the QDs, although the stability of the LED device needs to be further improved.
The transport and doping of colloidal nanocrystal arrays
In the preceding part, I have been focusing on the synthesis, composition, heterostructures and shape control of colloidal nanocrystals. In this part, I will emphasize the electronic properties of colloidal nanocrystals, including transport properties, doping and their energy levels in the absolute energy scale, which are critical for the design of a complex device and for the understanding of their performance.
The hopping transport in nanocrystal solids and ligand exchange
The emergence of high-mobility nanocrystal solids37 up to 400 cm2 V-1 s-1 has open up interesting opportunities for high-performance electronic and optoelectronic devices, such as field effect transistors (FET)38–40, photodetectors41–43, solar cells44, and light emitting diodes (LEDs)45,46.
At room temperature and under an electric field, the transport of a charge carrier in a nanocrystal solid occurred through hopping from one nanocrystal to its nearest neighbor, as shown in Figure I.7 a. The mobility of the carriers in a nanocrystal array is determined by the electronic coupling strength between nanocrystals, and more directly speaking, by the hopping rate. During a hopping event, the carrier needs to overcome the tunnel barrier. As illustrated in Figure I.7 b and Figure I.7 c, the height of the tunnel barrier is controlled by the nature of the barrier material (typical value for organic ligands is probed to be 2 eV),40 while the tunnel width is the inter-dot spacing (i.e., the length of ligands). In a nanocrystal solid with native ligands , the tunnel width is around 1.5 nm the typical length of alkyl chains (Figure I.7 b), which makes the nanocrystal film insulating with mobility of around 10-8 cm2 V-1 s-1. 40
To boost the conductance of the nanocrystal arrays while preserving the quantum confinement, ligand exchange of native long ligands toward shorter organic molecules or small inorganic ions47 by either solid-state method40 or solution phase transfer47 has been well established. It is reported that, in the absence of other changes, the mobility is supposed to increase exponentially with decreasing ligand length.40 To change the original long ligands with their short counterparts, the most common strategy is solid-state ligand exchange. For example, by immersing a nanocrystal films capped with long ligands into a solution of ethanethiol (EDT) in ethanol. The EDT of higher concentration will occupy the surface and replace the origin ligands. After ligand exchange towards EDT, the tunneling barrier width can be decreased to 0.3 nm (Figure I.7 c), and the mobility of the carriers improved by about 6 orders of magnitudes.40 The loss of ligand volume during this solid-state ligand exchange leads to voids in the thin film, which can be filled by multilayer deposition and exchange. It is worth noting that, for the target of light emission, the ligand species as well as the solvents for ligand exchange need to be carefully chosen to preserve the PL efficiency while enhancing the mobility.
The replacement of original ligands with compact inorganic ions such as S2- ions48 and halides44, can simultaneously reduce the tunnel barrier and height, leading to the stronger electronic coupling of the nanocrystals. Interestingly, small ions such as OH– and Cl–, helps to match the density of ligands to the number of surface atoms, in turn benefits the electronic passivation of surface dangling bonds49. For the ligand exchange toward small molecules or ions, phase-transfer ligand exchange is usually adopted. Typically, a polar solvent with the small ligands is mixed with the original nanocrystal solution. After sonication or vigorous agitation, the small molecular ligands are attached to the nanocrystal surface and bring the nanocrystal in the polar phase, whereas the original long-chain ligands are stripped from the nanocrystal and remain in the original nonpolar phase. In this phase-transfer manner, we can obtain an “ink” of nanocrystals with short ligands, which enables the deposition of thick films with good quality in one step, instead of the conventional layer-by-layer solid-state ligand exchange.
Along with the development of nanocrystal based optoelectronic devices, more complex ligand exchange strategies have been adopted. For example, the use of hybrid organic-inorganic (ammonium acetate-metal halide) inks of PbS nanocrystals has provided a nanocrystal solid with both high carrier mobility and good surface passivation, leading to high power efficiency of PbS nanocrystals-based solar cells50. In a recent report, a HgSe film with a hybrid ligand of amine/halogen and thiols using solution phase transfer demonstrates mobility of 1 cm2 V-1 s-1, two orders of magnitude higher than that of the film obtained with traditional solid-state ligand exchange using EDT solution.51
Field effect transistor
For the design of semiconductor optoelectronic devices, the insight of the electronic properties of the materials including doping, carrier mobility and energy levels are required. Since the Hall effect, which is commonly used in solid state physics, is not easy to conduct in a nanocrystal solid due to its low carrier mobility,52 field effect transistor (FET) has become the most straightforward way to identify their majority carrier and carrier mobility. Although FET itself has been one of the target applications for the next generation of low-cost nanocrystals-based electronics, here the goal of the FET configuration is not application-oriented but rather probing the transport characteristics of the nanocrystal solids and providing insights on the nature of majority carrier, and the mobility of the carriers in a nanocrystal solid.
Table of contents :
I.1 Colloidal semiconductor nanocrystals
I.2 The transport and doping of colloidal nanocrystal arrays
II.1 Introduction to nanocrystal-based infrared photodetection
II.2 Ag2Se nanocrystals for mid-infrared photodetection
II.3 Degenerately doped ITO nanocrystals for mid-infrared detection
I.1.1 Quantum confinement effect
I.1.2 Hot injection synthesis of colloidal nanocrystals
I.1.3 Ligands of nanocrystals
I.1.4 Heterostructure of nanocrystals
I.1.5 Shape control of Nanocrystals
I.2.1 The hopping transport in nanocrystal solids and ligand exchange
I.2.2 Field effect transistor
I.2.3 X-ray photoemission to build energy diagrams of nanocrystal arrays
Part II Heavy-metal-free nanocrystals for mid-infrared photodetection
II.1.1 Infrared photodetection
II.1.2 Photoconductors and photodiodes
II.1.3 Figures of merit for infrared photodetection
II.1.4 Infrared-active nanocrystals
II.1.5 The state-of-the-art of nanocrystal-based photodetectors
II.1.6 Challenges of nanocrystal based infrared photodetectors
II.2.1 Tunable mid-infrared intraband transitions of Ag2Se
II.2.2 The origin of doping for Ag2Se nanocrystals
II.2.3 Transport properties of Ag2Se nanocrystal arrays in dark conditions
II.2.4 Photoconductance of Ag2Se nanocrystal arrays
II.2.5 Conclusions and perspectives
III.1 Introduction to nanocrystal light emitters
III.2 Nanoplatelet-based LEDs for all-nanocrystal LiFi-like communication
III.3 HgTe nanocrystals for infrared electroluminescence and active imaging
II.3.1 LSPR in conducting nanostructures
II.3.2 Synthesis and optical properties of ITO nanocrystals
II.3.3 Transport properties of ITO nanocrystals
II.3.4 Photoconductance in ITO nanocrystal films
II.3.5 Conclusions and perspectives
Part III Nanocrystal-based LEDs and their applications
III.1.1 Colloidal nanocrystals for display with large gamut
III.1.2 Nanocrystals as down converters for QD-LCD display
III.1.3 QLED for future display
III.1.4 Nanocrystal-based LEDs beyond QD and visible
III.2.1 Synthesis and characterization of CdSe/CdZnS NPLs
III.2.2 Fabrication and characterization of NPL based LED
III.2.3 Characterization of the LED devices based on different CdSe/CdZnS NPLs
III.2.4 The origin of efficiency droop: beyond Auger recombination in emitting layer
III.2.5 Toward all-nanocrystal-based LiFi-like communication
III.2.6 Conclusions and perspectives
III.3.1 The design of the new-generation HgTe nanocrystal-based LED
III.3.2 Synthesis and characterization of the building-block nanocrystals
III.3.3 The investigation of HgTe/ZnO heterojunction as light emitter
III.3.4 Fabrication and characterization of the SWIR HgTe based LEDs
III.3.5 Toward narrower and brighter LED using sphere HgTe seeds
III.3.6 Conclusions and perspectives