Semiconductor nanocrystallites (quantum dots, QDs) whose radii are smaller than the bulk exciton Bohr radius constitute a class of materials intermediate between molecular and bulk forms of matter. Quantum confinement of both the electron and hole in all three dimensions leads to an increase in the effective band gap of the material with decreasing crystallite size (Figure 1). Consequently, both the optical absorption and emission of quantum dots shift to the blue (higher energies) as the size of the dots gets smaller. Although nanocrystallites have not yet completed their evolution into bulk solids, structural studies indicate that they retain the bulk crystal structure and lattice parameter. Recent advances in the synthesis of highly monodisperse nanocrystallites have paved the way for numerous spectroscopic studies assigning the quantum dot electronic states and mapping out their evolution as a function of size.
In 1993 the Bawendi group developed a synthetic method for producing monodisperse cadmium chalcogenide nanocrystals. This synthetic method was based upon the classic work of La Mer and Dinegar in 1950 on colloid formation . The injection of precursors above a critical temperature creates a nucleation event and this is followed by rapid cooling to a growth temperature where no further nucleation is favorable.
Since 1993, the Bawendi group has pursued the synthesis of novel semiconductor nanocrystals (Figure 1). The unique c axis of the wurtzite CdSe structure allows for the growth of nanoscale heterostructures upon the terminal faces of the c axis. This has been exploited in the synthesis of CdSe/CdTe nanobarbells to spatially separate excitons for the potential purpose of creating more efficient solar cell materials.
In this lab, we design and fabricate microfluidic reactors as a synthesis tool and a mean to systematically study the mechanism of QD formation (Figure 1). Microfluidic reactors enable a number of advantages over conventional chemical processes including enhanced control of heat and mass transfer, lower reagent consumption during optimization, and sensor integration for in-situ reaction monitoring.,
We make great efforts in improving the reactor design, such as a change from single-phase to segmented flow in our design work (Figure 2). In the single-phase laminar flow, diffusion is the only means of mixing. As a result of the parabolic fluid-velocity profile, particles near the wall spend more time in the reactor than those in the center, resulting in broad Residence Time Distributions (RTDs). In the two-phase case, recirculation within each liquid slug brings material from the wall to the center of the channel. This facilitates mixing, which narrows the RTD, and results in narrower size distributions.
Precursor solutions are delivered separately into the heated section and an Ar gas stream introduced further downstream results in a segmented gas-liquid flow. Recirculation within the liquid slugs rapidly mixes the heated precursors, thereby initiating the reaction. The fluids initially pass through a meandering section to induce good mixing across the channel before reaching a longer straight-channel section where the majority of the particle growth occurs. The segment lengths are very uniform during conditions of a typical QD synthesis and the resulting monodisperse samples exhibit excellent optical properties. The reaction is stopped when the fluids enter the cooled outlet region of the device.
The development of monodisperse magnetic nanocrystals (NC) has been intensively pursued due to their fundamental and technological interest. Magnetic NC often exhibit very interesting physical and chemical properties, which are significantly different from those of their bulk counterparts. Recently, it has been reported that when sample containing an interface between a ferromagnet and an antiferromagnet is cooled in a magnetic field, it may exhibit an additional unidirectional anisotropy due to magnetic coupling at the interface.
We investigated systematically the magnetic properties of colloidal Co NPs after three extents of oxidation (Figure 1). The native sample has a thin (1.0 nm) CoO shell and exhibits no exchange biasing. The partially oxidized sample has a thicker CoO shell (3.2 nm), and is exchange biased. The sample fully oxidized to CoO loses exchange biasing. We observe three distinct magnetic properties that result from the finite-thickness antiferromagnet shell exchange coupled to a finite-size ferromagnet core, and from crystal and stoichiometric defects: (1) an enhancement of the thermal stability of the orientation of the magnetic moment due to exchange biasing in the partially oxidized sample, (2) a low-temperature paramagnetic response in the partially and fully oxidized samples due to defects in the CoO shell, and (3) an asymmetry in the field-dependent magnetization for the partially oxidized sample at low temperature due to small clusters of Co in a diffusion layer around the Co core (Figure 2).
Despite great efforts directed toward the development of adaptable chemistries for synthesis of nanocrystals, the exact reaction mechanisms leading to crystal growth in these systems are still poorly understood. Furthermore, since experimental methods to study reactive processes occurring on crystal surfaces are limited by the extreme difficulty of isolating elementary reaction steps, we decided to exploit first principles methods. In particular, we have employed periodic density functional theory (DFT) calculations to investigate the mechanistic details leading to nanocrystalline growth of CdSe. Some of our key findings are summarized below (Figure 1).,
Collaborator: Dr. John V. Frangioni (Beth Israel Deaconess Medical Center)
For in vivo biological imaging applications, QD materials are chosen based on size, optical properties, and toxicity. The emission wavelength should ideally be in a region of the spectrum where blood and tissue absorb minimally but detectors are still efficient, approximately 700-900 nm in the near-infrared (NIR). In addition, the hydrodynamic size of the QD should be appropriately matched to the biological experiment of interest.
Quantum dots composed of CdTe core enclosed in a shell of CdSe have been developed to extend the fluorescence wavelength into the NIR range. A type-II energy structure is formed by the two semiconductors, the conduction and valence bands of which are positioned such that the holes and electrons are quantum-confined to the core and the shell, respectively (Figure 1); the nanocrystals thus behave very similarly to indirect semiconductors near the band edge. Charge carriers must cross the core-shell interface for radiative recombination, emitting photons with energies dependent on the band offset and hence can be smaller than that the bandgap of either material. Type II CdTe/CdSe quantum dots therefore exhibit widely tunable fluorescence, and wavelengths between 700 to 1000nm have been reported by our group by varying the core size and shell thickness of the nanostructures.
The size and NIR emission of type II CdTe/CdSe quantum dots have been applied in bioimaging studies after being rendered water soluble by coating the nanocrystals using oligomeric phosphine.  Having a final hydrodynamic diameter (HD) of 15.8-18.8 nm, these QDs were injected into live animals and were successfully used for selectively mapping the sentinel lymph node (SLN) in rat and pig models(Figure 2). Emission is observed directly through skin, providing a visual surgical guide for SLN resection.
Recently, we synthesized a size series of (InAs)ZnSe (core)shell QDs that emit in the near-infrared and exhibit HD < 10 nm (Figure 3). We have demonstrated their utility in vivo by imaging multiple, sequential lymph nodes and showing extravasation from the vasculature in rat models, neither of which has been achieved before with QDs. Longer emission wavelengths ranging from 750 to 920 nm can be achieved by increasing (1) the core size or (2) the shell thickness or (3) by altering the band offsets between core and shell such as by adding a small amount of Cd to the ZnSe shell. Due to the perceived undesirability of Cd for in vivo imaging, however, we generally obtained longer wavelengths by the first two options. The biological utility of these fluorescent probes resulted from our intentional choice to match the semiconductor material and water-soluble ligand with a desired final HD and emission wavelength.
Most recently, we have been able to reduce the size of QDs further to <6 nm HD using (CdSe)ZnS (core)shell QDs ligand exchanged with DL-cysteine (QD-cys), and observed some interesting new in-vivo behavior. When injected into rats intravenously, QD-cys mainly accumulated in the bladder 4 hr post-injection, demonstrating for the first time renal clearance of semiconductor nanocrystals in rat models (Figure 4).
By radioactively labeling QD-Cys using Tc99m and quantitatively tracking biodistribution as a function of nanocrystal size, a HD of ~5.5 nm was established as the threshold for renal clearance of QDs (Figure 5). Thus far, few nanometer-sized objects are being actively translated to the clinic. Our study suggests that to satisfy both patient safety and regulatory review, nanoparticle biodistribution and clearance must be carefully considered. This study provides a foundation for the design and development of biologically targeted nanoparticles for biomedical applications.
Collaborator: Alice Ting (Chemistry)
Labeling of biological molecules using fluorescent tags is a common and very useful practice in biological science. However, the intrinsic photophysical properties of organic and genetically encoded fluorophores, which generally have broad absorption/emission profiles and low photobleaching thresholds, have limited their effectiveness in long-term imaging and 'multiplexing' (simultaneous detection of multiple signals) without complex instrumentation and processing. Colloidal semiconductor nanocrystals can overcome these problems by combining the advantages of high photobleaching threshold, good chemical stability, and readily tunable spectral properties.
Recently, we report a family of water-soluble quantum dots (QDs) that exhibit low nonspecific binding to cells, small hydrodynamic diameter, tunable surface charge, high quantum yield, and good solution stability across a wide pH range. These QDs are half the size of commercial QDs and were applied to a number of live cell labeling applications, including single molecule imaging. The favorable properties of these QDs for cell labeling were enabled by a new class of surface ligands incorporating dihydrolipoic acid, a short poly(ethylene glycol) (PEG) spacer, and an amine or carboxylate terminus, which allowed for further covalent modification of the QDs.
High-affinity labeling was demonstrated by covalent attachment of streptavidin, thus enabling the tracking of biotinylated epidermal growth factor (EGF) bound to EGF receptor on live cells (Figure 1). The covalent attachment of molecules was demonstrated by appending a rhodamine dye to form a QD-dye conjugate exhibiting fluorescence resonance energy transfer (FRET). In addition, QDs solubilized with the heterobifunctional ligands retain their metal-affinity driven conjugation chemistry with polyhistidine-tagged proteins. This dual functionality was demonstrated by simultaneous covalent attachment of a rhodamine FRET acceptor and binding of polyhistidine-tagged streptavidin on the same nanocrystal to create a targeted QD that also exhibited dual-wavelength emission (Figure 2). Such emission properties could serve as the basis for ratiometric sensing of the cellular receptor's local chemical environment.
Collaborator: Nocera Group (Chemistry)
Semiconductor NC quantum dots are typically isolated both chemically and electronically from their surroundings by inorganic and organic passivating layers, and these features are responsible in part for the photostability that has made them an increasingly popular choice for biological imaging and tracking studies. However, these coatings tend to make NCs unresponsive to their environment. A sensing function may be derived by overlaying the NC with a coating containing a lumophore whose emission is environmentally responsive. This type of construct takes advantage of the NC’s robustness for NC-based sensors in which the unchanging fluorescence behavior of the NC core is compared against the environmentally-responsive coating. However, additional advantages may be achieved if the NC becomes actively involved in the response of the coating via a resonant energy transfer relationship established between the QD donor and the appended molecules. The continuous excitation spectrum of the NC is conferred upon the construct, obviating the need for independent excitation of the sensing fluorophore. NCs are also well-suited as FRET donors for sensor applications because of the very high valency presented by a suitably functionalized NC surface: that is, a large and tunable number of acceptors can be associated with each NC so as to achieve the optimal overall FRET efficiency to maximize sensitivity.
The sensor depicted at right is constructed from CdSe/ZnS core/shell NCs linked to squaraine-derived acceptor molecules (Figure 1, top). Sensing action results from the engineered overlap of the pH-sensitive dye absorption spectrum with the (pH-insensitive) quantum dot emission spectrum (Figure 1, middle). Whereas strong overlap leads to efficient FRET at low pH, deprotonation of the phenolic proton(s) at basic pH (pKa=8.8) reduces the overlap and thus lowers the FRET efficiency. The solution pH can be quantified by taking the ratio of each emission peak intensity (dot and dye) to the intensity at the isosbestic point, which functions as an internal reference (Figure 1, bottom). Self-referencing is preserved under a varying excitation power and the imposition of a strongly scattering medium, making the method a robust sensing approach for a variety of biological applications.
This work is being pursued through an ongoing collaboration with the Nocera group.
Collaborators Rakesh Jain (Mass. General Hospital)
Microspheres containing chromophores have been used in various applications including photonic crystals, biological labeling, and flow visualization in microfluidic channel. In our group, we prepared silica microspheres incorporated quantum dots. These microspheres process silica surfaces that are biocompatible and can be functionalized with various silane coupling chemistry. Moreover, advantages of using QDs as fluorophores are their continuous adsorption spectra, narrow emission bandwidths, and large two photon absorption cross-section, which are needed in many applications.
The microspheres were prepared using modified Stöber process. CdSe/CdZnS and CdS/CdZnS NCs were first chemically modified with ligand processing alkoxysilane groups making them compatible with the sol-gel process. Then, these NCs were incorporated into silica shell during the step of shell formation onto preformed silica microsphere cores. Using this method, we obtained the fluorescent silica micropheres with narrow size distribution. Also, color and size of the resulting microspheres can be tuned independently (Figure 1).
The applications of these microspheres in biological systems have been demonstrated in collaboration with Professor Jain’s group from MGH. For example, we used these microspheres for blood flow profile study in the brain of a mouse (Figure 2). Also, these microspheres can be used as a probe for study of the role of particle size in the distribution within the extracellular matrix of a human melanoma.
We also applied this method to synthesizing similar materials such as fluorescent microspheres with titania shell. We are currently working on applying this process to other nanoparticle materials to expand the function of microspheres. We are also exploring the use of these microspheres in other biological applications.
To help guide applications, many of the relevant aspects of the single exciton state structure and fluorescence have been studied extensively. However, a handful of potential applications rely on the less well-understood multiexcitonic states in which two or more electrons and holes have been excited. For example, the biexciton-exciton transition provides the optical gain in nanocrystal-based lasers.
Our group was among the first to study the optical properties of multiexciton complexes. Time-resolved photoluminescence spectroscopy of nanocrystal ensembles reveals new features with distinct dynamical and spectral signatures growing in when the sample is excited with strong laser pulses. First a fast feature due to the biexciton appears superimposed on the slow single exciton decay and then a further blueshifted feature is seen at higher power corresponding to emission from higher multiexcitons (Figures 1 and 2). The fast nonradiative decay of these multiexcitons is thought due to an enhanced “Auger”-like process whereby an electron-hole pair recombines by simultaneously promoting another one of the available free electrons or holes in a single energy-conserving step (Figure 3).
Multiexciton states are of interest also for their potential use as nonclassical light sources because electrons and holes recombine pairwise so that multiple, possibly correlated, photons are emitted within a short time from a single quantum object. Such light sources have applications for fundamental quantum optics studies and possibly in quantum computation and cryptography. Our group has experimentally demonstrated photon pairs due to this cascaded emission of multiexcitons and excitons from a single nanocrystal. Using two single-photon detectors in a Hanbury-Brown-Twiss geometry we have observed pairs of photons emitted from a NC after a single excitation pulse (see Figure 4). This constitutes unambiguous proof of the creation of multiexcitons, and could open the door for future applications of multiexcitons in NCs.
In recent years multiexcitons in NCs have become interesting for a key role they could play in light harvesting. In a conventional solar cell with a single active layer, photons with energy higher than the bandgap create a single electron-hole pair and any excess energy is lost to heat by rapid relaxation via phonon emission. However, if those initial highly energetic charge carriers are able to relax instead by collision with valence electrons to create additional e-h pairs a large part of the otherwise wasted energy could be recouped. The process, known as carrier multiplication, is very inefficient in the bulk, but recent experimental studies by other researchers have shown evidence for very efficient, ultrafast multiplication in a variety of semiconductor nanocrystal materials including PbX (X=S,Se,Te), InAs, Si and CdSe.
By looking at individual quantum dots, we have learned about phenomena never expected from ensemble measurements. Blinking, the tendency for dots to turn on and off with peculiar statistics, was discovered in our group in 1996  That same year, we saw spectral diffusion, where the narrow emission spectra of a dot changes position over the course of a measurement. 
These two properties are related. Spectral diffusion is believed to be caused by fluctuations in the local electric field and can be induced by placing a single dot between two electrodes (Figure 1). Often times, these spectral changes will be preceded by a blinking event (Figure 2). Such a blinking event could be caused by the dot becoming charged and decaying nonradiatively. When the charge leaves, the dot returns on but the local electric fields have changed, causing a Stark effect in the dot. Despite over 10 years of study and many phenomenological studies showing the universality of blinking, the underlying mechanism remains an open question. Blinking has important implications for the utility of dots in applications such as single particle tracking and single photon sources. Even though blinking was not expected from ensemble studies, future research in our group demonstrated that this property does in fact also influence ensemble luminescence. 
We continue to do spectroscopic studies at the single dot level to better understand properties such as multiexciton formation, radiative decay, and ordered photon emission ,  Such studies allow us to study fundamental physics of nanocrystals without the inhomogeneities inherent to ensemble measurements.
We have resolved luminescence from individual colloidal CdSe/ZnS core/shell nanocrystals embedded in organic light emitting devices under laser and electrical excitation at room temperature. This work demonstrates that individual semiconductor nanocrystals can serve as emissive probes in organic optoelectronic devices and might be of interest as single photon sources.
Photoluminescence studies at low laser excitation power have shown that quantum dots behave as non-classical light sources., Electroluminescence from single quantum dots, which is more appealing for practical applications, has only been demonstrated recently at low temperature  using self-assembled quantum dots grown by molecular beam epitaxy. In the recent study, we reported the demonstration of electroluminescence from single colloidal CdSe/ZnS (core/shell) nanocrystals embedded in organic light emitting device structures at room temperature. Spectral diffusion and blinking from individual quantum dots, key features of single quantum dots, were observed both in electro- and photoluminescence. We proposed a model in which the nanocrystals act as seeds for the formation of current channels that lead to enhanced exciton recombination in the vicinity of the quantum dots. Atomic force microscopy (AFM) studies on different layers of the device also support the model.
By using small-bandgap semiconductors (Eg ~ 0.4 eV), we are able to synthesize quantum dots that absorb photons all the way from the UV to the visible and into the infrared regions of the solar spectrum. These materials are solution-processable and can function on a flexible electrode, so the final solar cell product could be printed like newspaper, which would be much less expensive than the fabrication of solar cells on the market today. While quantum dot solar cells have been successfully made in a number of labs, the field of quantum dot solar cells is relatively new, and the fundamental mechanisms for charge transport and photoconductivity are not fully understood. Much of our work focuses on understanding how current flows through a solid-state quantum dot system and how quantum dots interface with other materials to absorb sunlight and output electricity.
Characterization of charge transport in NC films
The ability to develop NC based devices depends on a strong understanding of the nature of charge transport in the NC films. Carrier transport occurs through hopping between localized states in the semiconductor nanocrystal core. The levels of trap states, type of charge doping, interdot barriers and exciton lifetimes vary not only with the type of NC used, but also with the ligand, annealing temperature and exposure to air and light. For example, by replacing the long surface ligand (a crucial component of the nanocrystal synthesis and the reason the nanocrystals can be suspended in solution) with a shorter ligand, the nanocrystals can get closer together in the film. An electron moving between nanocrystals can do so much more easily when the ligand barrier is shorter. In a solar cell that translates to a higher current at a given voltage, or equivalently, a higher power.
We study these fundamental properties both in our lab and in collaboration with Kastner group in the physics department at MIT.
Quantum dots in photovoltaic architectures
The behavior of a photovoltaic cell as a whole depends on each of the materials in the architecture as well as the interfaces between them. We fabricate solar cells with various architectures and materials to investigate how the energy levels and transport properties affect the overall solar cell performance.
Semiconductor nanocrystals have features of both molecules and bulk semiconductors, which gives them versatility in device architecture. They can be mixed into a solution with a polymer in a bulk-heterojunction architecture as a carrier separation and transport network, or deposited into a nanocrystal-only film to create a junction with an adjacent device layer.
We fabricate and study these devices both in our lab and in collaboration with the Bulovic group in the electrical engineering department at MIT.
Electrical photodetection of IR light using QDs
QD-based short-wave infrared (SWIR) detectors show promise as a low cost, solution processable alternative to InGaAs focal plane arrays. Our design involves QDs deposited in a photoconductive geometry on a modified integrated circuit, which allows for complicate fabrication of the circuit and electrodes prior to QD deposition. Using a single-pixel test structure, we have established a method to controllably tune the resistance of the QD film to match the capacitance of the circuitry.
We fabricate and study these devices both in our lab and in collaboration with the Bulovic group in the electrical engineering department at MIT.
Luminescent UV/IR dual-band photodetectors
There is significant interest in developing dual band UV-IR photodetectors for communications, fire detection and other applications, however p-n junction photodetectors suffer from inherent limitations to sensitivity. The bright and tunable optical properties of IR emitting QDs can be using in a passive yet robust manner to improve the UV efficiencies of IR detectors via luminescent downshifting. By embedding QDs which absorb in the UV and emit in the IR regions of detector efficiency, we have effectively sensitized an InGaAs detector to UV radiation without hindering its IR performance, thereby creating a dual band detector. Infrared light incident on the bare detector is collected in the active region of the detector (gray box). Ultraviolet light is collected near the surface of the detector (yellow box) at low efficiency. With the addition of the QD layer, ultraviolet light is absorbed by the QDs and re-emitted in the infrared where it is collected in the active region of the detector. Due to the absorption profile of the QDs, the LDC layer absorbs most of the UV light, but is nearly transparent to infrared light.