Quantum dots (QDs) are colloidal, semiconductor nanocrystals (Empedocles and Bawendi, 1997; Prasad, 2004; Alivisatos, 1996) which are at the forefront of nanoscience. They are spherical (Law et al., 2009), contain roughly 200-10 000 atoms (Smith et al., 2008) and range in size from 1 to 10nm (Zhu et al., 2013). QDs are composed of groups II-VI (e.g. CdTe) or III-V elements (e.g. InP; Chan et al., 2002). They have received considerable interest for their contribution to both fundamental materials research and industrial needs (Talapin et al., 2010; Cozzoli et al., 2006; S. Scalbi et al., 2016). QDs have many unique optical and electronic properties (Smith et al., 2008; Alivisatos, 2004; Ferrari, 2005; Niemeyer, 2001) which can be used for various applications such as fluorescent labelling (Smith et al., 2008), biological imaging (Santra et al., 2005), and LED lighting (Wood and Bulovic, 2010).
Semiconductors each have a unique band gap between their valence and conduction bands. A semiconductor can only be activated by a photon with more energy than that band gap (Smith et al., 2008). This causes an electron to be excited from the valence to the conduction band which leaves a hole in the valence band. An electron-hole pair, called an exciton, is formed by the attraction of the opposite charges between the electron and the hole. Radiative recombination occurs when the electron returning to the hole is accompanied by the fluorescent emission of a photon (Chan et al., 2002).
When the physical size of the semiconductor is reduced to a critical size, called the Bohr exciton radius, the electron energy levels should be treated as discrete, like the discrete energy levels of electrons in an atom. That is called the quantum confinement effect (Algar and Krull, 2009; Efros and Efros, 1982; Brus, 1986; Wang and Herron, 1991; Alivaisatos, 1996) and it gives rise to QDs outstanding optical properties including high quantum yield (QY; ratio of emitted to absorbed photons), long fluorescence lifetime, broad absorption spectra, large effective Stokes shift (frequency difference between the first absorption band and the maximum emission), narrow and symmetric size-tunable emission, and strong resistance to photobleaching and chemical degradation (Li and Zhu, 2013; Resch-Genger et al., 2008). Decreasing the QD radius (R) results in a blueshift in absorption and photoluminescence (PL) via an increase in the band gap energy, which is roughly proportional to 1/R2 (Wang and Herron, 1991; Alivisatos, 1996). This is how QDs of the same composition but of different sizes can generate different fluorescent wavelengths (Michalet et al., 2005). QD optical properties are controlled by the constituent material, particle size, dispersity, surface chemistry and shell (Dabbousi, 1997).
SYNTHESIS AND SURFACE CHEMISTRY
The synthesis of QDs can be generally classified into two categories: organometallic (Talapin et al., 2001; Qu, Peng and Peng, 2001; 2006; Peng and Peng, 2001) or aqueous (Rogach et al., 1996; Zou et al., 2008; Kalasad, Rabinal and Mulimani, 2009). The organometallic synthesis is costly and time-consuming, it also requires high temperatures and toxic chemical reagents (Kumar and Nann, 2006; Peng et al., 2001; Chen et al., 2003; Talapin et al., 2001). The solvent selected during synthesis must have both a high boiling point and high coordinating capabilities to both the metal and chalcogen elements, such as trioctylphosphine oxide (TOPO; Zhu et al., 2013). However, their crystal structures are nearly perfect, so they have high QYs and a narrow size distribution (Zhu et al., 2013). This route produces hydrophobic QDs, and since solubilisation is required for all biological purposes, complicated steps must be taken to phase transfer the QDs (Wuister et al., 2003; Fischer et al., 2006; Lei et al., 2008; Liu et al., 2007; Gerion et al., 2001; Smith et al., 2008). The QD properties will be affected by the coating used for solibilization because the charge and size will change among other things. Therefore, the solubilisation strategy should be tailored to the biological system being used (Luccardini et al., 2006). The most common methods used are ligand-exchange (Gerion et al., 2001; Bruchez et al., 1998; Kim and Bawendi, 2003; Smith et al., 2008; Chan and Nie, 1998) or encapsulation with an amphiphilic polymer (Pellegrino et al., 2004; Smith et al., 2008). Ligand exchange involves mixing a suspension of TOPO-coated QDs with a solution containing an excess of a heterobifunctional ligand. It must have a functional group to bind to the QD and another that is hydrophilic. This allows for the displacement of TOPO as the bifunctional ligand adsorbs and renders the QD hydrophilic. This method leads to smaller overall QDs than encapsulation (Smith et al., 2008), but it tends to decrease fluorescence efficiency and aggregate in biological and other high ionic strength buffers (Verwey and Overbeek, 1948). The encapsulation method yields QDs surrounded by phospholipid micelles (Dubertret et al., 2002). The QDs can be dispersed in aqueous solution and remain stable for long periods of time through hydrophobic interactions (Smith et al., 2008). Although this method is simpler, the result is usually too large to enter cells and they disassemble more readily than the ligand-coated QDs, however they are less sensitive to ionic strength.
The aqueous synthesis uses short-chain thiols as surfactants and leads to hydrophilic QDs (Weng et al., 2006; Green et al., 2007; Zheng, Gao and Ying, 2007; Gaponik et al., 2002; Qian et al., 2006; He et al., 2006; Weiyong et al., 2007), making it the generally preferred option despite lower quantum yield, broader size distribution, and lower stability (Law et al., 2009; Zhu et al., 2013, Smith et al., 2008) because it is simpler, reproducible and the reaction is less toxic (Law et al., 2009; Zhu et al., 2013). The resulting QDs are ready to use in biological environments (Gaponik et al., 2002).
The core, shell, and coating of the QD all affect the photochemical properties (Jamieson et al., 2007), so the right selection for the desired use is essential. The composition, size, and shape of QD core are essential to their PL emission range, which can be tuned from ultraviolet to the near infrared (NIR) range (Zhu et al., 2013). QD cores on their own have been proven impractical because they may have surface imperfections and they are highly reactive (Kirchner et al., 2005). Imperfections in the crystalline structure may act as temporary traps for the electron or hole preventing their radiative recombination. The alternation between trapping and untrapping can result in blinking (intermittent fluorescence; Nirmal et al., 1996; Raab et al., 2004; Kuno et al., 2003) and reduces overall quantum yield. The cores are reactive because of their large surface area to volume ratio which can lend itself easily to photochemical degradation (Jamieson et al., 2007).
To protect surface atoms from oxidation and other chemical reactions, cores are now often coated with another semiconductor with a larger band gap to create a shell (Talapin et al., 2004; Marchi-Artzner et al., 2008). If the shell is designed carefully, it can even improve the PL, QY and stability as well as negate the surface imperfections and reduce toxicity (Kirchner et al., 2005; Manna et al., 2002; Hines and Guyot-Sionnest, 1996; Jamieson, et al., 2007; Reiss and Bleuse, 2002). A poor design can lead to strain between the core and shell which causes misfit dislocations, relaxing the nanocrystal structure which adversely affect QY and colloidal stability (Talapin et al., 2004; Dabbousi et al., 1997). The addition of a shell results in a slight red shift (Dabbousi, 1997). Core/shell-structured QDs are preferable for their use in biomedical applications because they exhibit strong luminescence with low toxicity.
The coating is what allows the QD to be used for specific applications (Jamieson et al., 2007). QDs can be used for targeting once coated with ligands with biomolecular affinities, such as peptides, proteins or nucleotides (Zhu et al., 2013). The large surface area to volume ratio of QDs can be beneficially used to attach multiple agents (Smith et al., 2008). For QDs to interact with the target and therefore perform non-invasive imaging, they need to be coated with ligands that interact with the target (Curnis et al., 2008; Goldman et al., 2002). The surface properties of QDs and the functional groups of the selected ligands must be considered when choosing a conjugation strategy. As mentioned above, QDs with hydrophobic surfaces need to be rendered hydrophilic through surface ligand exchange (Li et al., 2009; Park et al., 2011) or interaction with amphiphilic polymers (Muro et al., 2012). Once they have hydrophilic surfaces, they share similar conjugation strategies with the QDs synthesized via the aqueous route.
Based on the functional groups on the QD surface, different reactions can be used for surface conjugation of QDs. It can be done electrostatically, via biotin-avidin interactions, by covalent cross-linking or by binding to polyhistidine tags (Parak et al., 2002; Xing et al., 2007; Medintz et al., 2005; Mason et al., 2005). Ethyl-dimethylaminopropyl carbodiimide (EDC), is widely used as a coupling reagent between carboxlic acid and amino groups to form amide bonds (Hua et al., 2006). EDC coupling can cause non-specific crosslinking and lead to loss of bioactivity. The thiol-maleimide reaction is more specific, because usually the thiol, but not the maleimide, are present on the biomolecule. The (strept)avidin-biotin interaction is attractive because of its high affinity and specificity (Maldiney et al., 2012; Chen et al., 2012; Allen et al., 2010). It relies on either direct binding between streptavidin-functionalized QDs with biotinylated proteins, or the use of avidin as a bridge between biotinylated QDs and biomolecules. Metal-histidine binding introduced nickel nitrilotriacetic acid groups (Ni-NTA; Bae et al., 2009; Park et al., 2010; Susumu et al., 2010), which is a compound widely used for isolation and purification of proteins that contain histidine tags.
There is no technique which consistently allows preparation of QDs with control over the ratio of biomolecules per QD and their orientation on the surface. Current strategy based on modifying COOH groups on the QD surface for covalent attachment of amine groups is limited by problems of reproducibility and aggregation (Mattoussi et al., 2000; Jamieson et al., 2007).
Accurate and sensitive detection of water-soluble targets is highly sought due to the potential uses for medicine and defense. (Hermanson, 1996; Iqbal et al. 2000; Ligler and Rowe, 2002; Clapp et al., 2003). Organic dyes are well established as fluorescent labels and have been valuable in the study of biological phenomena due to their inherent high sensitivity of detection and ease of use (Frangioni, 2003; Bremer and Weissleder, 2001). However, they have several shortcomings including narrow absorption spectra, broad emission spectra with long tailing, allowing only small Stokes shifts, short lifetimes and sensitivity to photobleaching (Hermanson, 1996; Lakowicz, 1999). Fluorescence techniques are well suited to detect the interactions of biomolecules quickly, sensitively, reliably and reproducibly (Mason, 1999; Lakowicz, 2006; Zhang et al., 2002). The main problems associated with fluorescence microscopy are cell auto-fluorescence, which can mask signals from labeled molecules, the requirement of long observation times and low tissue tissue penetration (Michalet et al., 2005; Frey et al., 2009; Zrazhevskiy, Sena and Gao, 2010). Imaging in the near infrared (NIR) region can overcome some of those problems. The combined drawbacks of reduced quantum yield (QY) and photostability at NIR wavelengths hampers the use of organic dyes for those imaging applications (Levene et al., 2004; Rosenthal et al., 2002; Kim et al., 2004).
QDs have attracted significant attention for fluorescent labelling, because of their exceptional properties (Resch-Genger et al., 2008). With size-tunable emission, QDs can be generated for any specific wavelength, from UV through to the NIR region (Wang et al., 2003; Talapin, 2004; Kumar and Nann, 2006; Xu, Ziegler and Nann, 2008; Jiang et al., 2006; Shavel, Gaponik and Eychmuller, 2006; Hinds et al., 2007; Lefshitz et al., 2007; Delehanty et al., 2009; Murphy et al., 2006). QD emission peaks are narrow and symmetric, and the broad absorption spectra allows for the excitation of multiple fluorophores with a single light source, called multiplexed imaging (Rosenthal, 2001; Gao and Nie, 2003; Gao and Nie, 2004; Smith and Nie, 2004; Wang et al., 2006; Chan et al., 2002; Han et al., 2001). QDs have molar extinction coefficients several orders of magnitude higher than organic dyes, and high quantum yields (QYs), resulting in bright fluorescent probes (Chan and Nie, 2016; Dubertret et al., 2002; Ballou et al., 2004). QDs have a large effective Stokes shift (Medintz et al., 2005) and they are highly resistant to photobleaching with long fluorescence lifetimes, which allows them to be distinguished from background and other fluorophores for increased sensitivity and time-gated detection, as well as continuous monitoring (Wu et al., 2003; Bruchez et al., 2013; Smith and Nie, 2004; Dahan et al., 2001). QDs have large two-photon action cross-sectional efficiency, which makes them suitable for in vivo deep-tissue imaging (Larson et al, 2003; He et al., 2007; Padilha et al., 2005; Clapp et al., 2007; Voura et al., 2004; Larson et al., 2003).
Suitable fluorescent labels should not aggregate or precipitate under relevant conditions. For organic dyes, solubility can be tuned via substituents, if the optical properties are not affected by them. There are plenty of organic dyes available that are soluble in relevant medias. QD dispersibility is controlled by the chemical nature of the surface coating as discussed above (Wang et al., 2007).
The spectroscopic properties of fluorophores are sensitive to the microenvironment, including polarity, viscosity, pH, ionic strength and the presence of surfactants, fluorescence quenchers or conjugated molecules. The effect depends on the dye class, nature of the emitting state, excited state redox potential, charge and hydrophilicity (Gruber et al., 2000; Buschmann, Weston and Sauer, 2003; Randolph and Waggoner, 1997; Panchuk-Voloshina et al., 1999; Berlier et al., 2003; Seidal, Schulz and Sauer, 1996). The spectral characteristics of dyes with a resonant emission are only moderately affected, but the QY and lifetime can change considerably and they can undergo aggregation-induced fluorescence quenching (Mujumdar et al., 1993; Gruber et al., 2000; Buschmann, Weston and Sauer, 2003; Panchuk-Voloshina et al., 1999; Berlier et al., 2003). QDs are also less prone to aggregation-induced fluorescence quenching. Dyes with an emission from an excited state that has a considerable dipole moment tend to have notable spectral changes with changes in polarity. For QDs, the microenvironmental effect on spectroscopic features is mainly governed by the accessibility of the core surface (Medintz et al., 2005). Typically, properly shelled QDs are minimally sensitive to microenvironment polarity (Ji et al., 2008). QD emission is barely responsive to viscosity, unlike many organic dyes. However, electrostatically stabilized QDs tend to aggregate with increasing ionic strength.
Bioconjugation often leads to a decrease in QY for both label types. The parameters that can affect label fluorescence are the chemical nature, the length of the spacer and, at least for organic dyes, the type of neighboring oligonucleotides or amino acids in the bioconjugated form (Buschmann, Weston and Sauer, 2003; Randolph and Waggoner, 1997; Seidel, Schultz and Sauer, 1996). The knowledge of such microenvironmental effects greatly simplifies label choice. This is generally an advantage of organic dyes. Only a few systematic studies have been performed so far on the effect of the microenvironment on QD spectroscopic properties.
Photostability is a critical feature in most fluorescence applications. A fluorescent label must be stable under relevant conditions, and in the presence of typical reagents at common temperatures and under a typical excitation light flux over routinely used detection times. Label stability is of crucial importance for detection sensitivity, especially in single molecule experiments, and for contrast.
Organic dyes and most NIR fluorophores have poor photostability (Dahne, Resch-Genger and Wolfbeis, 1998; Mason et al., 2005; Panchuk-Voloshina et al., 1999; Berlier et al., 2003; Eggeling, Volkmer and Seidal, 2005). Many NIR dyes also suffer from poor thermal stability in aqueous solutions (Soper and Mattingly, 1994). Some organic dyes have been designed to display enhanced photostability, and readout times have decreased (Panchuk-Voloshina et al., 1999; Berlier et al., 2003). Despite these improvements, limited dye photostability can hamper microscopic applications requiring high excitation light intensities or long-term imaging. In contrast, adequately surface-passivated QDs display excellent thermal and photochemical stability and photo-oxidation is almost completely suppressed for relevant time intervals (Mason et al., 2005; Sun et al., 2006; Ziegler et al., 2007; Nida et al., 2008). This is a considerable advantage over organic fluorophores for high intensity and long-term imaging applications (Mason et al., 2005; Han et al., 2001; Alvivisatos, 1996; Kang et al., 2012). However, photooxidation of QDs has been observed (Zhang et al., 2006), as has the QD-specific phenomenon of photobrightening (Parak et al., 2002), and undesired aggregation of QDs can contribute to reduced stability (Hoshino et al., 2004).
QDs emit light with a decay time in the order of a few tens of nanoseconds at room temperature, which is slower than the auto-fluorescence background decay, but fast enough to maintain a high photon turnover rate (Dahan et al., 2003; Gerion et al., 2001; Pinaud et al., 2006). In time-gated analysis, photons hitting in the first few nanoseconds are disregarded to decrease background noise and increase sensitivity. The fast fluorescence emission of organic dyes upon excitation coincides closely with short-lived autofluorescence background from many naturally occurring species, reducing the signal-to-noise ratio. However, the typically mono-exponential decay kinetics of organic dyes enable straightforward dye identification from measurements of fluorescence lifetimes, making dyes suitable for applications involving lifetime measurements. Conversly, the complication size, surface, wavelength, and time dependent, bi or multiexponential QD decay behaviour (Schlegel et al., 2002; Zhang et al., 2006) renders species identification from time-resolved fluorescence measurements very difficult.
Target quantification using fluorescence is affected by both the stability of the label and the sensitivity of its spectroscopic properties to the environment. Organic dyes have been successfully applied for quantification in a broad variety of in vitro fluorescence applications, but use of QD quantification is rare. QD photobrightening can hamper direct quantification and may render the use of reference standards necessary (Parak et al., 2002). For single-molecule spectroscopic applications, blinking of QDs and of organic dyes can be a considerable disadvantage (Zhang et al., 2006; Gomez, Califano and Mulvaney, 2006). Another aspect that may influence the usability of QDs for quantification is the fact that not all QDs in a preparation are luminescent; some exist in permanently nonfluorescent states (Ebenstein, Mokari and Banin, 2002).
For single-molecule imaging and tracking applications, QDs are superior to most organic fluorescent dyes owing to their photostability, which would allow tracking for longer. However, QD blinking, the causes and mechanism of which are not yet completely understood, needs to be overcome. Blinking may be exploited for superresolution microscopy by analyzing the intermittent fluorescence to allow identification of the light emitted by each individual label and to localize it accurately with a resolution of a few tens of nanometers (Lidke et al., 2005).
Labelling of biomolecules with fluorophores requires suitable functional groups for attachment or binding. The advantage of organic dyes in this regard is the commercial availability of functionalized purification and characterization techniques for dye bioconjugates (Waggoner, 2006). Several strategies for site-specific covalent and noncovalent labeling of proteins in living cells are available including enzyme-catalyzed labeling by post-translational modification, which may be used to label proteins at the cell surface. Both intracellular and surface labeling have also been achieved by specific chelation of membrane-permeant fluorescent ligands or by self-labeling (Miller et al., 2005). Acetomethoxymethyl (AM)-ester derivatization, microinjection, cationic liposomes, controlled cell volume or cell membrane manipulation and endocytosis are established methods for delivery of organic dye labels into cells (Waggoner, 2006; Torchilin et al., 2003). The small size of organic dyes minimizes possible steric hinderance, which can interfere with biomolecule function, and allows the attachment of several fluorophores to a single biomolecule to maximize the fluorescence signal. However, high label densities can result in steric hinderance which could lead to fluorescence quenching (Dahne, Resch-Genger and Wolfbeis, 1998; Mujumdar et al., 1993; Gruber et al., 2000) and may also influence biomolecule function.
There are no consensus methods for labeling biomolecules with QDs (Xing et al., 2007). The general principle for QD biofunctionalization is that they are first made water-soluble and then they are bound to the biomolecules. There are only a few standard protocols available for labelling biomolecuels with QDs and the choice of suitable coupling chemistries depends on surface functionalization. It is difficult to define general principles because QD surfaces are unique. In contrast to labeling with small organic fluorophores, several biomolecules are typically attached to a single QD, which could sterically hamper access to cellular target, and it is difficult to control biomolecule orientation. This can affect the spectroscopic properties and colloidal stability of the QD as well as the biomolecule function. QD bioconjugates have been found to be powerful imaging agents for specific recognition and tracking of plasma membrane antigens on living cells. The bright, stable fluorescence emitted from these QDs allowed the continuous observation of protein diffusion on the cellular membrane, and could even be visualized after the proteins were internalized. These conjugates showed superior photostability, lateral resolution, and sensitivity relative to organic dyes.
Fluorescence resonance energy transfer (FRET) involves the energy transfer from a photoexcited donor particle to an acceptor particle whenever the distance between them is smaller than the Forster radius (Riegler and Nann, 2004). That transfer reduces the donor’s emission intensity and length while increasing the acceptor’s emission intensity. FRET can be used to measure changes in distance on the 1 to 10 nm scale (Selvin, 2000; Miyawaki, 2003). For example, it can be used to measure protein conformational changes (Heyduk, 2002; Day, Periasamy and Schaufele, 2001; Li and Bugg, 2004; Miyawaki, 2003; Mason, 1999; Sapsford, Berti and Medintz, 2006; Hermanson, 1996; Lakowicz, 1999). To detect the change in distance effectively, good spectral overlap and well resolved donor and acceptor PL are required. The excitation line should coincide with the minimum of the acceptor absorption spectrum to reduce direct excitation of the acceptor.
While FRET has already been used extensively for biological applications, accurate measurements of time-dependent conformational change, such as monitoring protein dynamics over an extended period and continuous monitoring of target toxins or small molecule analytes under realistic constraints remain difficult when using conventional organic dyes as the interacting donor-acceptor pair. While the overall energy transfer may be efficient, there is often significant emission overlap that obscures the individual behavior of the donor and acceptor (cross-talk) which leads to complications in the data analysis (Lakowicz, 1999; Turro, 1991).
Recently QDs have been successfully exploited as FRET donors with organic dyes as acceptors (Grecco et al., 2005; Sapsford, Berti and Medintz, 2006). However, there are several issues that affect their use. The physical dimensions of QDs, particularly after adding a shell and capping, make close approach to the QD core difficult, reducing FRET efficiency (Clapp et al., 2004; Medintz et al., 2004; Sapsford, Berti and Medintz, 2006). By attaching acceptor dye-labelling proteins onto QDs, two advantages over organic fluorophores became apparent. First, the QD donor’s narrow photoemission could be size-tuned to improve spectral overlap with the acceptor dye, which circumvents cross-talk (Clapp et al., 2004). Second, having several acceptor dyes interact with a single QD donor substantially improves the efficiency (Clapp et al., 2004; Sapsford, Berti and Medintz, 2006; Medintz et al., 2003, Medintz et al., 2004). However, steric hinderance is a concern for subtrate accessibility (Chang et al., 2005). Displacement of peptide-dye conjugates from a central QD has also been reported, particularly when larger biomolecules are being used (Shi et al., 2007).
The application of QDs as FRET acceptors is not recommended because of their broad absorption bands, which favour cross-talk. Generally, FRET applications of QDs should only be considered if there is another QD-specific advantage for the system in question, such as the possibility to avoid excitation cross-talk, their longer fluorescence lifetimes or their very large two-photon action cross-sections (Clapp et al., 2007; Medintz et al., 2003). In most cases, fluorescent proteins or organic dyes are to be favoured for FRET.
QD-FRET is being used to develop QD biosensors which would detect biomolecular interactions in real-time (Smith and Nie, 2004). FRET has been the major proposed mechanism to render QDs switchable from a quenched state to a fluorescent state (Tran et al., 2002). That ability would be a very powerful tool. Biosensors based on organic fluorophores have already shown promise in vivo (Weissleder et al., 1999).
Spectral multiplexing or multicolour detection is typically performed at a single excitation wavelength and discriminates between different fluorescent labels based on their emission wavelength. A tunable Stokes shift and very narrow, preferably well-separated emission bands of simple shape are the desirable optical properties of a suitable fluorophore for this application. The suitability of organic dyes is limited. However, an increasingly common multiplexing approach uses donor-acceptor dye combinations that make use of FRET from the donor to the acceptor fluorophore to increase the spectral separation of the absorption and emission (Sapsford, Berti and Medintz, 2006). QDs are ideal candidates for spectral multiplexing at a single excitation wavelength because of their optical properties (Medintz et al., 2005; Jaiswal et al., 2003; Han et al., 2001).
Despite many superior optical properties as compared to established organic dyes, the solutions for using QDs have so far been individual ones. The fact that QDs behave not as molecules but as nanocolloids complicates their application in biological environments. The costs of finding a solution to the challenges of their system must be weighed against the benefits of the advanced spectroscopic features of QDs. Organic dyes still have an advantage for some applications due to their smaller size. Future improvements in QDs will provide increased benefit for areas in which long-term luminescence stability, high brightness or multi-colour detection are crucial. QDs could have a bright future in NIR fluorescence for in vivo imaging. QDs are unlikely to replace organic dyes, but they will be used to complement dye deficiencies in certain applications.
QDs can offer unprecedented interactions with biomolecules both on and inside cells which may revolutionize biological imaging (Gao et al., 2004; Gao et al., 2010; Cai and Hong, 2012; Sun et al., 2012; Rhyner et al., 2006; Xing et al., 2007; Wu et al., 2003; Kim et al., 2004). There have been improvements in the stability and biocompatibility of QDs, so their use for in vivo imaging has become more realistic (Li and Zhu, 2013; Tandon and Nordstrom, 2011; Tavares et al., 2011). The most frequently used QDs for in vivo imaging are CdSe/ZnS (Diagaradjane et al., 2008; Yang et al., 2009; Kato et al., 2010; Lu et al., 2011; Mukthavaram et al., 2011; Wang et al., 2012) because they are commercially available, well-studied and characterized (Hering et al., 2007; Dembski et al., 2008; Yong et al., 2008; Kirchner et al., 2005; Clarke et al., 2008).
For in vivo imaging applications, the following factors need to be considered when designing the probes: potential toxicity, interference in or of physiology, circulation lifetime, and optimal emission wavelength for sufficient tissue penetration of signal (Zhu et al., 2013). Surface modifications can lead to aqueous solubility, protect against degradation and fluorescent quenching, reduce non-specific uptake into organs, and aid in further bioconjugation. PEG coating is one of the most commonly used strategies for surface modification because of its well-known biocompatibility and mature chemistry. PEG coating allows for a longer half-life once administered and reduces the uptake into recticuloendothelial organs (Al-Jamal et al., 2009; Schipper et al., 2009; Ballou et al., 2004). Other polymers, liposomes and inorganic silica have also been used as coatings (Loginova et al., 2012; Nicolas et al., 2011; Jayagopal, Russ and Haselton, 2007). Inorganic silica can provide a hydrophilic surface and facilitate the incorporation of various functional groups such as carboxyl, amine, and thiol groups for further bioconjugation (Bruchez et al., 1998; Szabo and Vollath, 1999). The silica coating reduces the uptake of QDs into the liver and spleen, redirecting the QDs to the kidney, blood and urine.
QDs have been successfully used in several ways including in vitro bioassays, labelling fixed cells (dead cells) and tissue specimens as well as imaging membrane proteins on living cells (Smith et al., 2008). Theoretically, QDs can be used for real time imaging, multiplexing, single molecule detection, and biological sensing on the nanoscale, which would help answer a broad range of analytical problems and biological questions. However, owing to their size, the intracellular delivery of QDs is challenging (Parak, Pellegrino and Plank, 2005). There is no general protocol, so individual solutions need to be established. Several methods have been used to deliver QDs to the cytoplasm including non-specific uptake by endocytosis, direct microinjection of nanolitre volumes (tedious and limits the number of cells labelled), electroportation (uses increased permeability of membrane under pulsed electric fields to deliver QDs) and receptor-mediated uptake (exploits the cells propensity to recognize and internalize certain peptides, and labelling of QDs with peptides so they are internalized together; Chen and Gerion, 2004; Derfus, Chan and Bhatia, 2004; Voura et al., 2004; Jaiswal et al., 2003; Parak, Pellegrino and Plank, 2005; Rozenzhak et al., 2005; Chen and Gerion, 2004; Jaiswal et al., 2003; Sun et al., 2006), but they have not been particularly successful. Peptide-targeted uptake looks to be the most specific for delivering dispersed QDs into cells, although it may be limited to cells expressing the appropriate receptors.
Once delivered inside the cytoplasm of cells, dispersion of the QDs depends strongly on surface coating and pH stability. QDs tend to aggregate, especially QDs capped with COOH-terminated groups due to their poor stability in acidic conditions, or they are trapped in endocytotic vesicles, even if uptake was through a non-endocytic pathway (Jamieson et al., 2007; Smith et al., 2008). Advanced delivery methods are needed to deliver QD probes into living cells.
Cell-penetrating peptides are chemical transfectants that have gained widespread interest due to their high transfection efficiency, the versatility of conjugation, and their low toxicity. Therefore, cell-penetrating peptides such as Tat are being investigated for their capacity to deliver QDs into living cells. The Tat peptide-mediated delivery occurs via macropinocytosis, a fluid-phase endocytosis process that is initiated by the binding of Tat-QD to the cell surface. An important finding is that the QD-loaded vesicles are actively transported by molecular machines along microtubule tracks to an asymmetric perinuclear region called the microtubule organizing center (MTOC). Tat-QDs strongly bind to cellular membrane structures like filopodia, and large QD containing vesicles can pinch off from the tips of filopodia.
A new class of cell-penetrating QDs was developed based on the use of multivalent and endosome-disrupting surface coatings (Duan and Nie, 2001). Hyperbranched copolymer ligands such as PEG-grafted polyethylenimine (PEI-g-PEG) encapsulated and stabilized QDs in aqueous solution through direct ligand binding to the QD surface. Due to the cationic charges and a ‘proton sponge effect’ associated with multivalent amine groups, these QDs penetrated the cell membrane and disrupted endosomal organelles in living cells. In comparison with previous QDs encapsulated with amphiphilic polymers, the cell-penetrating QDs were smaller and very stable in acidic environments. The QDs were rapidly internalized by endocytosis, and the pathways of the QDs inside the cells showed dependence on the number of PEG grafts of the polymer ligands. While higher PEG content led to QD sequestration in vesicles, the QDs coated by PEI-g-PEG with fewer PEG grafts could escape from endosomes and be released into the cytoplasm. PEI and other polycations are known to be cytotoxic, however the grafted PEG segment was found to significantly less so.
Another problem for QD intracellular labelling is the impossibility of removing probes that did not find their target, because it is not possible to distinguish bound QDs from unbound QDs. Without the ability to wash away unbound probes, there is a need for activateable probes that are ‘off’ until they reach their intended target.
Extracellular targeting with QDs is typically accomplished through QD functionalization with specific antibodies in order to image cell-surface receptors (Xing et al., 2007; Medintz et al., 2005) or via biotin ligase–catalyzed biotinylation in conjunction with streptavidin-functionalized QDs (Chen et al., 2005; Howarth et al., 2005).
For most in vivo imaging applications using QDs, systemic intravenous delivery into the bloodstream will be the main mode of administration. For this reason, the interaction of the nanoparticles with the components of plasma, the specific and nonspecific adsorption to blood cells and the vascular endothelium, and the eventual biodistribution in various organs are of great interest. It has been consistently reported that QDs are taken up nonspecifically by the reticuloendothelial system (RES), including the liver and spleen, as well as the lymphatic system (Akerman et al., 2002; Dubertret et al., 2002). These findings are not necessarily intrinsic to QDs, but are strictly predicated by thier size (Frangioni et al.). To achieve efficient urinary excretion and elimination of QDs, the overall size should be under 5.5 nm. The use of biodegradable QDs, which can be broken down into components that can be cleared by the renal system, is another possibility which may be developed in the future.
One of the most immediately successful applications of QDs in vivo has been their use as contrast agents for the two major circulatory systems of mammals, the cardiovascular system and the lymphatic system. For imaging of the lymphatic system, the overall size of the probe was an important parameter for determining biodistribution and clearance. QDs of 9 nm were found to migrate further into the lymphatic system, with up to 5 nodes showing fluorescence at once (Kim et al., 2004). This technique could have great clinical impact due to the quick speed of lymphatic drainage and the ease of identification of lymph nodes, enabling surgeons to fluorescently identify and excise nodes draining from primary metastatic tumors for the staging of cancer. A current problem is that a major fraction of the QDs remain at the site of injection for an unknown length of time.
Imaging of tumors presents a unique challenge because of the unique biological attributes inherent to cancerous tissue. Blood vessels are abnormally formed during tumor induced angiogenesis, having erratic architectures and wide endothelial pores. These pores are large enough to allow the extravasation of large macromolecules, which accumulate in the tumor microenvironment due to a lack of effective lymphatic drainage. Because cancerous cells are effectively exposed to the constituents of the bloodstream, their surface receptors may also be used as active targets of bioaffinity molecules. In the case of imaging probes, active targeting of cancer antigens has become an area of tremendous interest (Jayagopal et al., 2009). QDs were conjugated to peptides with affinity for various tumor cells and their vasculatures to actively target a tumor (Akerman et al., 2002). Tumor targeting with QDs generated tumor contrast on the scale of whole-animal imaging (Gao et a., 2004). Tumor fluorescence was significantly greater for the actively targeted conjugates compared to nonconjugated QDs, which also accumulated passively.
There have been many other new discoveries in biological imaging. Aptamers have recently emerged as a new class of targeting ligands, with many advantages including small size, versatile chemistry, ease of synthesis, and lack of immunogenicity (Hong et al., 2011; Zhang et al., 2011). RGD peptide-conjugated Cd-free QDs (Gao et al., 2010; Erogbogbo et al., 2011) exhibit highly specific tumor targeting and reduced accumulation in the lung, kidney, and heart in mice (Han et al., 2012; Yong, 2010; Yong, Roy and Law, 2010; Li et al., 2012; Gao et al., 2012). After conjugation to RGD peptides, significantly higher tumor uptake and longer retention at the tumor sites was observed compared to non-targeted dendron-coated QDs. Captopril is a drug for treating hypertension since it can inhibit the activity of angiotensin-converting enzyme. The in vivo behavior of captopril-conjugated QDs was investigated after intraperitoneal injection (Kato et al., 2010). Strikingly, it could be delivered into the brain via systemic circulation, suggesting that it may be a potential platform to break the blood-brain barrier.
Compared to the study of living cells in culture, different challenges arise with the increase in complexity to a multicellular organism, and with the accompanying increase in size. Unlike monolayers of cultured cells and thin tissue sections, tissue thickness becomes a major concern because biological tissue attenuates most signals used for imaging. Optical imaging has been used in living animal models, but it is still limited by the poor transmission of visible light through biological tissue. It has been suggested that there is a NIR optical window in most biological tissue that is the key to deep-tissue optical imaging. Rayleigh scattering decreases with increasing wavelength, and that the major chromophores in mammals, hemoglobin and water, have local minima in absorption in this window, so biological tissue is generally more transparent at wavelengths in the NIR region (Lim et al., 2003; Weissleder, 2001). Therefore, for deep tissue penetration, QDs that emit in the NIR will be required (Bailey and Nie, 2003; Kim et al., 2003; Hines and Scholes, 2003; Murray et al., 2001; Kershaw et al., 2000). Light wavelengths in this range could improve the tumor imaging sensitivity at least ten-fold, allowing the sensitive detection of 10-100 cancer cells. The NIR fluorescence of Ag2Se QDs was shown to penetrate the abdominal cavity of mice, demonstrating their potential application in in vivo imaging. NIR QDs were used to image coronary vasculature in a rat model (Lim et al., 2003) and to visualize sentinel lymph nodes in a pig, in 1 cm deep tissue (Kim et al., 2004). They will be powerful tools for molecular imaging because they can be used to image in real time with multiplexed detection to monitor biomolecular phenomena in vivo.
There are many uses for advanced imaging in medicine at all stages. Each of the following common imaging modalities has its limitations, including magnetic resonance imaging (MRI; James and Gambhir, 2012; Sosnovik and Weissleder, 2007; Balyasnikova et al., 2012), computed tomography (CT) imaging (Bhargava et al., 2012; Tejwani et al., 2012), positron emission tomography (PET) imaging (Gambhir et al., 2001; Gambhir, 2002; Iagaru, 2011; Cai and Hong, 2011; Li et al., 2012), single-photon emission computed tomography (SPECT) imaging (Fakhri, 2012), optical imaging (Huang, Lee and Chen, 2011; Gaikwad and Ray, 2012; Cai, Zhang and Kamp, 2011; Chin et al., 2012), and ultrasound imaging (Dayton and Rychak, 2007). Optical imaging is highly sensitive, easy and non-damaging, but it has low spatial resolution and depth penetration (Lee et al., 2012; Niedre and Ntziachristos, 2008). MRI has high resolution, soft-tissue contrast and penetration, but it is high cost, low sensitivity and has a long imaging time (Cheon and Lee, 2008; Lee et al., 2012). SPECT imaging has unlimited depth penetration, but has limited spatial resolution and requires radioactivity (Lee et al., 2012). PET imaging is high sensitivity with unlimited depth penetration, but is high cost and has radiation risk (Lee et al., 2012). CT imaging has high spatial resolution, but has poor soft-tissue contrast and is not targeted with radiation risk (Lee et al., 2012). Ultrasound imaging has high spatial and temporal resolution and low cost, but it is operator dependant with targeted imaging limited to vascular compartment (Lee et al., 2012).
More precise biological data could be acquired by combining two or more imaging modalities (Nolting et al., 2012; Thorek et al., 2012; Wang et al., 2011; Zeman and Scott, 2012). Therefore, manufacturing of hybrid medical equipment and studies related to multimodal imaging probes have been on the rise (Kim, Piao and Hyeon, 2009). Several types of probes have been under development (Cai et al., 2007; Fernandez et al., 2008; Wang et al., 2007). Tremendous advances have been made in many imaging techniques over the last decade, however, no single imaging modality is perfect to obtain all the necessary information for a given study. Combination of multiple imaging techniques using a single probe can potentially overcome the disadvantages and provide synergistic information. Many QD-based nanoprobes have been designed and evaluated for such applications.
QDs have large surfaces that can be modified through versatile chemistry. This makes them convenient scaffolds to accommodate multiple imaging and therapeutic agents. Not only would this allow tracking of pharmacokinetics, but diseased tissue could be treated and monitored simultaneously and in real time (Samia, Chen and Burda, 2003). QDs are fluorescent contrast agents, but can also be used as markers for electron microscopy due to their high electron density (Nisman et al., 2004; Dahan et al., 2003). Dual-modality QDs for both optical imaging and MRI were developed by chemically incorporating paramagnetic gadolinium (Gd) complexes in the lipid coating layer of QDs (Mulder et al.; Jin et al., 2008; Prinzen et al., 2007; Oostendorp et al., 2008; Mandal et al., 2005; Zhang et al., 2014). The Gd-doped QD can be prepared in one step (Li and Yeh, 2010; Liang et al., 2013; Liu et al., 2011), so it is simple and low cost. In vitro experiments showed that labeling of cultured cells with these QDs led to significant T1 contrast enhancement with a brightening effect in MRI, as well as an easily detectable fluorescence signal from QDs (Mulder et al.). By correlating the deep imaging capabilities of magnetic resonance imaging (MRI) with ultrasensitive optical fluorescence, a surgeon may be able to visually identify tiny tumors or other small lesions during an operation and remove diseased cells and tissue completely. QDs coated with paramagnetic gadolinium complexes and PEGylated lipids could be used as a dual-modality probe for both optical and magnetic resonance imaging, using the RGD peptides as ligands for tumor targeting (Mulder et al., 2006).
Traditional clinical contrast agents for CT scans are based on iodinated molecules and compounds with high X-ray absorption coefficient. However, these contrast agents are typically cleared very rapidly, which is a major disadvantage. Cardiovascular disease is a leading cause of death worldwide, and unstable atherosclerotic plaques represent important diagnostic targets in clinical settings for improving patient management (Rudd et al., 2009; Temma and Saji, 2012). QDs were combined with iodinated molecules to create a dual-modal contrast agent, composed of a hydrophobic iodinated oil core with QDs embedded inside (Ding et al., 2013).
64Cu was used to label QDs through the DOTA chelator and the RGD peptide was employed as the targeting ligand to label a tumor (Cai et al., 2007) using dual-modality fluorescence and PET imaging. The quantification ability and ultrahigh sensitivity of PET imaging enabled the quantitative analysis of the biodistribution and targeting efficacy of this dual-modality imaging probe. This imbalance in sensitivity is fundamental to the differences in the physics of these imaging modalities, and points to an inherent difficulty in designing useful multimodal imaging probes. These new materials are of great interest, although they are still in development, they could vastly increase accuracy of imaging results.
LIGHT-EMITTING DIODES (LEDS)
QDs have emerged as a competitive choice for the emissive component of light-emitting devices (QD-LEDs; Colvin et al., 1994; Tessler et al., 2002; Klein et al., 1997). The size and composition of the QDs can be easily-tuned to change the wavelength produced, without modifying the device structure (Sun et al., 2007; Wood and Bulovic, 2010). QD-LEDs offer great potential as they enable the optimization of both photometric and colorimetric properties of the device, which make them promising for improving warm white lighting and for flat-panel displays (FPDs; Coe-Sullivan et al., 2005), due to their superior colour purity (Li et al., 2005; Xu et al., 2005; Shirasaki et al., 2012). However, it is difficult to maintain the initial optical properties of the QDs during device fabrication and over long-term operation.
The main advantages of QDs, compared to organic molecules, are the chemical and optical stability, tunable emission wavelength and high colour saturation (He et al., 2013). The QDs-LED device represents the next frontier of the organic LED (OLED) technology, which used small organic molecules and semiconducting polymers (Tang and Van Slyke, 1987; Braun and Heeger, 1991), because it combines the optical performances of the QDs with the possibility to use laser technology for patterning. The QDs light emission is stimulated only when the energy gap of QDs is included in the energy band gap of the polymer (Bansal et al., 2014). The incorporation of QDs as light emitting centres improves the device performance in terms of colour ageing, device lifetime and ease of their industrial utilisation (Yang et al., 2015; Supran et al., 2013; Li et al., 2006; Colvin et al., 1994; Coe et al., 2002). The key advantages of OLEDs for flat-panel display (FPD) applications are their self-emitting property, high luminous efficiency, full-colour capability, wide viewing angle, high contrast, low power consumption, low weight, potential for large area colour displays and flexibility (Geffroy et al., 2006).
For solid state lighting applications, warm white is preferred over cold white. Commercial GaN blue LEDs have achieved over 65% wall plug efficiency (WPE). However, green LEDs which are more sensitive to the lumen eyes are suffering from low efficiency and red LEDs face long-term thermal stability issues. Thus, down-conversion became the industry standard. It involves a high-energy pumping source, usually a blue LED, and a down-converting luminophore as the emitter. Currently, widely adopted down-converting materials are micro-sized phosphors which have high quantum efficiency. For such an LED light source, the evaluation of light quality depends on the optical performance of the phosphors. Though a combination of red and green phosphors a warmer white LED can be generated, at the cost of efficiency.
QDs are resistant to photobleaching. Overcoating the QD core with a wider bandgap shell of a material such as CdS or ZnS can further improve the stability (Hines and Guyot-Sionnest, 1996). Recent developments in QD synthesis techniques have given rise to extremely stable QDs. For example, ‘thick’ shell QDs, where CdSe cores are overcoated with CdS such that the thickness of the shell is more than double the core diameter (Mahler et al., 2008; Chen et al., 2008), have been reported to sustain a high degree of thermal stability and maintain their luminescence even when the passivating ligands are removed (Chen et al., 2008). They also suppress blinking (Chen et al., 2008; Mahler et al., 2008), a key source of inefficiency in electrically excited QDs (Wood et al., 2009). Placing organic aliphatic ligands on QDs with metal chalcogenide ligands has enabled QD films that are entirely inorganic and exhibit record electronic transport properties (Kovalenko, Scheele and Talapin, 2009).
The device contains a hole-transport layer (HTL), an anode, a QD layer, an electron-transport layer (ETL), and a cathode (Coe et al., 2002). A buffer layer is also used to obtain stable, hole-free electrical conduction across the device (Lim et al., 2007; Jang et al., 2008; Ziegler et al., 2008; Ali et al., 2007; Nizamoglu et al., 2007; Nizamoglu, Zengin and Demir, 2008). Thickness of the QD layer played a critical role in colour purity, EL efficiency and maximum luminance of the devices. The optimal thickness of the QD layer in different coloured QD-LEDs is dependent on the size and structure of the QDs. HTL and ETL thicknesses were both optimized to confine the injected electrons and holes to recombine predominantly within the QD layer and provide optimal hole and electron transportation.
Understanding what limits efficiency is critical for the systematic development of QD-LEDs. Efficient electron and hole injection, balance of charge carriers arriving at the QD active layer, and minimization of the electric field across the QDs are important design criteria for ensuring high-performance QD-LEDs. (Mashford et al., 2013; Anikeeva et al., 2008; Wood and Bulovi, 2009). However, these design guidelines are highly device specific and difficult to achieve in the same device for different colour emitters with various chemistries and sizes (Anikeeva et al., 2009; Wood and Bulovi, 2009).
QY is probability that an exciton will recombine to emit a photon. In a QD-LED, External quantum efficiency (EQE), the number of photons emitted from the device per injected electron (Bozyigit and Wood, 2013), is proportional to QY, which depends on the exciton nonradiative (nr) and radiative recombination (r) rates, the two major contributors to the nr rate are electronic trap states and free-charge carriers. Electronic trap states result in an increase in nr, which decreases the QY. The presence of charge on a QD, can increase the Auger non-radiative recombination, where energy is dissipated as kinetic energy to a charge carrier instead of as a photon (Klimov, 2000; Galland et al., 2011. Again, this increases nr and decreases the QY.
Chemistry advances can be broadly categorized into two trends: overcoating of the QD core with a shell material and grading the QD core composition, which is referred to as alloying (Talapin et al., 2010). These modifications to the QD can result in a change in the degree of confinement of an exciton or in the extent of surface passivation or a combination of these two effects. Starting with a CdSe core, the addition of a shell passivates the surface of a QD core and offers physical separation of the exciton from defect states on the surface of the QD. This results in a decrease in the trap-assisted nr rate, thereby improving the QY (Hines and Guyot-Sionnest, 1996; Dabbousi et al., 1997; Chen et al., 2013). This is highlighted by the ‘giant’ shell QD as mentioned above. While the addition of a shell tends to improve passivation of the QD, depending on the energy levels of the shell material relative to those of the core, the shell can either increase or decrease confinement of the exciton. For example, if a ZnS shell is added to a CdSe core, the wave functions of the exciton are more strongly spatially confined in the CdSe core due to the large energy difference between the valence and conduction bands of the CdSe and ZnS. This increase in exciton confinement can be observed as a shift of the emission in the QD to higher energies. This spatial confinement promotes strong carrier-carrier interactions, which results in the QDs retaining the high Auger non-radiative recombination (ka) rate of the core (Klimov, 2000). QDs with alloyed composition are also understood to have low ka rates (Wang et al., 2009). This has been explained theoretically by a smoothing of the shape of the confinement potential that is believed to occur in these alloyed QDs (Cragg and Efros, 2010). Alloying is also thought to be present at the core–shell interface in the ‘giant’ shell QDs, which can explain the observation of extremely low ka rates in these QDs (Cragg and Efros, 2010).
‘Giant’ shell QDs, which offer the optimal passivation and low kas, would be ideal choices for an emitter in a QD-LED. However, record efficiencies in QD-LEDs have not been reported with ‘giant’ shell CdSe/CdS QD materials (Pal et al., 2012). Instead, QD-LEDs with multilayered alloyed structures exhibit the highest EQEs (Anikeeva et al., 2009; Steckel et al., 2006; Caruge et al., 2008). This is understandable when considering the impact of the electric field on the QY. Increasing the electric field across CdSe/CdS QDs decreases the QY (Bozyigit, Yarema and Wood, 2013; Kraus et al., 2007; Empedocles, 1997; Park et al., 2007). The extent of the decrease is dictated by the thickness of the CdS shell. Significantly less quenching is observed for a ZnS shell (Bozyigit et al., 2012). In most QD-LEDs, the QD active layer also experiences an electric field on the order of around 1 MV/cm. In a QD, while nr is not affected by the electric field in the device, the electric field can induce a spatial separation of the electron and hole wave functions that is sufficient to significantly reduce r. (Bozyigit, Yamera and Wood, 2013). In the limit where r is smaller or on the order of nr (r≤nr), a reduction of r can cause a significant reduction of EQE.
The challenge in developing a QD for use as an emitter in a QD-LED, where the QDs are subject to both charge carriers and electric fields is as follows: a QD with minimal electronic confinement will likely not suffer from reduced QY even as charge accumulates on it during LED operation, however, this QD will experience increased electric field-induced PL quenching. Design of the optimal QD emitter for a QD-LED is therefore non-trivial. One strategy to improve upon QD emitters is to minimize field-induced luminescence quenching in QDs already exhibiting low Auger non-radiative recombination rates and high QYs. This could be done through the rapid transfer of an exciton formed in an alloyed or thick-shelled QD to a localized state such that exciton polarization in the presence of an electric field is reduced or through minimizing the exposure of the QDs to high electric fields (Anikeeva et al., 2008).
Since QDs are synthesized from organometallic precursors, they retain a passivating layer of ligands, making them solution processable. This facilitates a variety of low cost, large-area deposition techniques, such as phase separation, inkjet printing, and microcontact printing (Coe et al., 2002; Tekin et al., 2007; Wood et al., 2009; Kim et al, 2008). These high efficiency, multicolour QD-LEDs were made possible by the development of QD contact printing, in which single monolayers of QDs can be transferred, solvent free, from a parylene-C coated poly(dimethylsiloxane) (PDMS) stamp onto an organic charge transport layer (Kim et al., 2008).
While QD-LEDs used for down-conversion have made it possible to begin the commercialization of QD-containing products, electrically excited QD films would eliminate the need for multicomponent integration of QD-LEDs. Planar, electrically excited QD-LED technology has so many different application including novel-format lighting applications at high brightness or simple paper-thin flat panel emissive display with high colour quality. The first demonstration of optical down-conversion with QDs used blue GaN LEDs to excite QDs in polylaurylmethacrylate and generate point sources of saturated colour light (Lee et al., 2000). More recently, a full colour, flexible EL display was achieved using inkjet printed thin films of QDs in polyisobutylene, optically excited by the blue EL from a commercial phosphor powder (Wood et al., 2009).
Using optical excitation of QDs, both challenges of QD charging and luminescence quenching can be avoided. Electrical current need not pass through the QD emissive layer, so the QDs do not experience charging and can be dispersed in a polymer to preserve their high luminescent efficiencies. Several products are being created to take advantage of this new technology. Evident Corporation developed strings of Christmas lights where blue LEDs excited QDs in an epoxy to achieve monochromatic colour emission across the visible spectrum. QD Vision Inc. and Nexxus Lighting are commercializing a solid-state lighting solution based on optical excitation of QDs. QDs in a transparent matrix are dispersed onto a substrate. They are excited by an array of efficient blue LEDs to provide mixed colour emission across the visible spectrum.
The device structure is completed with two electrodes. A pulsed applied electric field enables the sustained EL of the QDs in the film. When a voltage is applied across the device, no charge is injected from the contacts into the QD layer. Once the voltage exceeds the bandgap energy of a QD, an electron will transfer from its valence band to its neighbour’s conduction band. This creates a spatioally sparated electron and hole in the QD film that can subsequentially radiatively recombine (Wood et al., 2011). It is also possible to achieve field-driven electroluminescence in relatively low-voltage, constant current-driven devices by incorporation of thin (∼15 nm thick) electron blocking layers into the device, which permit sufficient buildup of an electric field to allow for the QD ionization process (Wood et al., 2010).Therefore, despite the high electric fields on the order of 5 MV/cm needed to generate the free charge, luminescence in field-driven QD-LEDs occurs under lower electric field conditions (∼1 MV/cm; Empedocles, 1997). Furthermore, because electric field-driven luminescence is inherently a local process, the emissive layer need not be a continuous QD film, but could consist of clusters of QDs embedded within an insulating matrix. Indeed, field-driven electroluminescence has been demonstrated for high QY, QD-insulating polymer blends, which had previously been restricted to applications involving optical excitation of colloidal QDs (Wood et al., 2011).
The field-driven QD-LED alleviates the band alignment considerations that typically dictate which emissive materials can be electrically excited using charge transport layers. Luminescent materials that have different chemistries and absolute energy level positions, and whose peak emission wavelengths span the visible to NIR regions, can all be excited within the same device structure (Wood et al., 2009).
We recently reported the development of a unipolar light-emitting device architecture, which highlights the possibility for a paradigm shift away from direct charge injection into QDs as a means for EL in inorganic-based QD-LED structures (Wood et al., 2010). These devices are the first reported example of DC-field-driven EL in QDs, and we demonstrate that we can eliminate charging of the QD film, showing constant luminance from the device over 20 h of continuous operation in air, unpackaged.
QDs consisting of alloyed cores with a smoothed confinement potential shape, which are further overcoated with shells, offer a solution to the design trade-off. Additionally, selection of QDs exhibiting localized luminescence, placement of the QDs in a high dielectric host material, or adoption of field driven QD-LED architectures serve as examples of potential innovations that address the challenges facing the realization of high-efficiency QD-LEDs.
There are a wide variety of QDs that exist. There are different cores, different shells and many different potential coatings. Their unique optical properties lend them well to a whole variety of applications including those described briefly in this review: fluorescence labelling, biological imaging and LED lighting.