Quantum dot optical devices with enhanced gain and sensitivity and methods of making same

Abstract

Various embodiment include optical and optoelectronic devices and methods of making same. Under one aspect, an optical device includes an integrated circuit having an array of conductive regions, and an optically sensitive material over at least a portion of the integrated circuit and in electrical communication with at least one conductive region of the array of conductive regions. Under another aspect, a film includes a network of fused nanocrystals, the nanocrystals having a core and an outer surface, wherein the core of at least a portion of the fused nanocrystals is in direct physical contact and electrical communication with the core of at least one adjacent fused nanocrystal, and wherein the film has substantially no defect states in the regions where the cores of the nanocrystals are fused. Additional devices and methods are described.

Claims

1. An apparatus, comprising: a focal plane array including at least one optically sensitive layer formed on an underlying circuit, the circuit being patterned to measure and relay electronic signals on a pixel-by-pixel basis, the signal being indicative of light absorbed in a medium from which the focal plane array is made, the focal plane array being sensitized to become responsive to wavelengths absorbed in the at least one optically sensitive layer.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) Embodiments of the invention will now be described, by way of example only, with reference to the accompanying drawings. The drawings are not necessarily to scale. For clarity and conciseness, certain features of the invention may be exaggerated and shown in schematic form. In the drawing:

(2) FIG. 1 shows a schematic of a known quantum dot nanocrystal.

(3) FIG. 2 shows a two-dimensional schematic of a layer of fused quantum dots.

(4) FIG. 3A shows an optical micrograph of a light sensitive layer formed on an electronic read-out chip.

(5) FIG. 3B shows side view of an optical device which includes an integrated circuit with an array of electrodes located on the top surface thereof.

(6) FIG. 4A is a side view of a portion of an optical device configured in a vertical sandwich structure.

(7) FIG. 4B is a side view of a portion of an optical device configured in a lateral planar structure.

(8) FIG. 4C is a plan view of a portion of an optical device configured in a lateral interdigitated structure.

(9) FIG. 5 shows an overview of steps in a method of making a QD optical device with enhanced gain and sensitivity.

(10) FIG. 6A shows absorption spectra and TEM images of lead sulfide QDs as ligands are exchanged from oleic acid to primary butylamine.

(11) FIG. 6B shows absorption spectra and TEM images of lead sulfide QDs as ligands are exchanged from oleic acid to primary butylamine.

(12) FIG. 6C shows FTIR spectra of lead sulfide QDs after ligands are exchanged to primary butylamine.

(13) FIG. 6D shows FTIR spectra of lead sulfide QDs with butylamine ligands before and after methanol wash.

(14) FIG. 6E shows XPS data of lead sulfide QDs through various processing stages.

(15) FIG. 6F shows FTIR data of lead sulfide QDs precipitated in inert and in oxidizing conditions.

(16) FIG. 7 A shows a device structure (inset) and plots of I-V characteristic for 100 nm and 500 nm QD layer devices.

(17) FIG. 7B shows the responsivity of a variety of QD layer devices.

(18) FIG. 7C shows the dark current density of a variety of QD layer devices.

(19) FIG. 7D shows the measured noise current as a function of measured dark current for a variety of QD layer devices.

(20) FIG. 7E shows the normalized detectivity of a variety of QD layer devices.

(21) FIG. 7F shows spectra of responsivity and normalized detectivity of a QD layer device.

(22) FIG. 7G shows the electrical frequency response of a QD layer device.

(23) FIG. 7H shows the temporal response of a QD layer device.

(24) FIG. 8 shows spectral dependence of responsivity and normalized detectivity D* at various levels of applied bias under 5 Hz optical modulation and 0.25 nW incident optical power.

(25) FIG. 9 shows the frequency dependence of responsivity and normalized detectivity at different levels of applied bias at 975 nm and 0.25 nW of incident optical power.

(26) FIG. 10 shows the noise equivalent exposure of a QD layer device as compared with conventional Si CCD and CMOS sensor devices.

DETAILED DESCRIPTION

(27) Overview

(28) The present invention provides quantum dot (QD) devices and methods of making devices. Many embodiments are optical devices with enhanced gain and sensitivity, and which can be used in optical and infrared (IR) imaging applications, photovoltaic applications, among other applications. The term quantum dot or QD is used interchangeably herein with the term nanocrystal, and it should be understood that the present invention is not limited exclusively to quantum dots but rather to any nanoscale crystalline material.

(29) Some embodiments of the QD optical devices are single image sensor chips that have a plurality of pixels, each of which includes a QD layer that is light sensitive, e.g., optically active, and at least two electrodes in electrical communication with the QD layer. The current and/or voltage between the electrodes is related to the amount of light received by the QD layer. Specifically, photons absorbed by the QD layer generate electron-hole pairs, generating a current and/or voltage. By determining the current and/or voltage for each pixel, the image across the chip can be reconstructed. The image sensor chips have a high sensitivity, which can be beneficial in low-light applications; a wide dynamic range allowing for excellent image detail; and a small pixel size, which to a large extent is limited by currently available CMOS techniques such as lithography. The responsivity of the sensor chips to different optical wavelengths is also tunable by changing the size of the QDs in the device, by taking advantage of the quantum size effects in QDs. The pixels can be made as small as 1 micron square or less.

(30) In many embodiments, the optically sensitive QD layer includes a plurality of QDs that have been specially processed to give the layer an enhanced gain and sensitivity as compared with conventional silicon-based layers as well as other kinds of QD layers, such as those described in the incorporated patent references. Specifically, a plurality of QDs are fabricated using well-known techniques, and typically include a core as well as an outer surface that includes a plurality of ligands. The ligands are exchanged for shorter, volatile ligands, and then the ligand-exchanged QDs are solution-deposited onto a substrate to form a QD precursor layer. The substrate itself may include one or more electrodes, or the electrodes may be deposited in a later step. Subsequently, the short ligands are removed from the QD precursor layer. This brings the QDs in the QD precursor layer into very close contact, so that at least some of the QDs contact their neighbors. This contact between QDs may be referred to as necking. Subsequently, the necked QD layer is annealed, which fuses the necked QDs together. The QD precursor layer is typically maintained in an inert atmosphere after ligand removal, so that the outer surfaces of the individual QDs do not oxidize until annealing is complete.

(31) While two given fused QDs in the annealed QD layer retain a large portion of their original shape, and thus remain individually recognizable, after annealing the QDs are no longer physically distinct from each other. Instead, the cores of the QDs together form a continuous electrical path. Thus, if many adjacent QDs neck, during annealing those necked QDs fuse to form an electrical network with a physical extent that is substantially greater than that of the individual QDs, and through which current will readily flow. For example, the fused QD film may have a macroscopic extent, though the QDs themselves are nanoscopic. In some embodiments, the finished QD layer after ligand removal, necking, and annealing, can essentially be considered a continuous inorganic film having nanoscale features. The general shapes of the individual QDs may still be recognizable, but their cores form a continuous electrical network that is mechanically robust. For example, a micrograph of the finished QD layer would show the general shape and size of the individual QDs from which the layer is formed, as well as robust joints between many adjacent QDs.

(32) In many embodiments, the fused QD layer is subsequently processed to modify its outer surfaces. For example, a material such as a semiconductor shell can be coated on the outer surfaces of the fused quantum dots. Or, for example, defect states can be formed on the exposed outer surfaces of the QDs, e.g., by oxidizing the fused QDs layer. These defect states effectively trap holes generated by photons, so that they recombine with electrons far less readily and thus greatly enhance the amount of current that a given photon generates in the finished QD layer, i.e., greatly enhance the photoconductive gain of the device. The fused QD cores, and the juncture between them, will generally not have defect states, so current will flow readily between them, in certain embodiments.

(33) FIG. 2 shows a two-dimensional representation of a portion of a QD layer. The layer includes a continuous network of fused QD cores 20, having outer surfaces 21 that are a different composition than that in the core, e.g., oxidized core material, or a different kind of semiconductor. The individual QD cores in the film are in intimate contact, but continue to exhibit many of the normal properties of individual quantum dots. For example, a lone (unfused) quantum dot has a well-characterized central absorbance wavelength that arises from quantum effects related to its tiny size, e.g., 1-10 nm. The central absorbance wavelength of the fused QDs in the film is not significantly shifted from their central absorbance wavelength before fusing. For example, their central absorbance wavelength may change by about 10% or less when fused. Thus, the QDs in the film retain their quantum effects, despite the fact that they may be an integral part of a macroscopic structure.

(34) Current is not generally though of as flowing through a lone (unfused) QD; instead, electrons simply occupy well-known quantum energy states in the CD core. If two lone (unfused) QDs are brought near each other, current can flow between them by electron hopping between the QDs, which has a well-known dynamic. In contrast, current does readily flow between fused QD cores, even though the cores themselves generally retain their quantum energy states. Because the cores are in contact, electrons can easily move between them. This aspect of QD fusing typically provides a carrier mobility of between about 0.001-10 cm.sup.2/Vs, in some embodiments between about 0.01-0.1 cm.sup.2/Vs, for example greater than 0.01 cm.sup.2/Vs, while at the same time not fusing the QDs to an extent that they lose their identity, namely their individual characteristics that provide quantum confinement. A film of fused QDs typically also exhibits a a relatively low electrical resistance pathway, e.g., having a resistance above about 25 k-Ohm/squareIt is also possible to overfuse QDs, in which case they no longer exhibit many of the normal properties of individual quantum dots. In the overfused case, the cores of the QDs do not generally have their own quantum energy levels, but the energy levels are instead distributed over multiple QD cores. This results in a film with a very low electrical resistance, e.g., less than about 25 k-Ohm, but which in many ways is effectively a bulk semiconductor material. Overfused QDs can also be recognized experimentally as a relatively large shift (e.g., greater than about 10%) in a shift to the red (longer wavelengths) in their absorption and/or emission spectra.

(35) In certain embodiments the QD layer is exceptionally light sensitive. This sensitivity is particularly useful for low-light imaging applications. At the same time, the gain of the device can be dynamically adjusted so that the device will not saturate, that is, additional photons continue to provide additional useful information. Tuning of gain can be conveniently achieved by changing the voltage bias, and thus the resultant electric field, across a given device, e.g., a pixel. As discussed in greater detail below, photoconductive gain, and correspondingly the responsivity in A/W, varies approximately linearly with bias and field. Thus, in a given device, a bias of about 0.1 V may result in a gain of about 10, while a bias of about 10 V may result in a gain of about 100.

(36) Some embodiments of QD devices include a QD layer and a custom-designed or pre-fabricated CCD or CMOS electronic read-out integrated circuit. CCD and CMOS electronic read-out circuits are readily commercially available at low cost. The QD layer is then formed directly onto the custom-designed or pre-fabricated CCD or CMOS electronic read-out integrated circuit. The QD layer may additionally be patterned so that it forms individual islands. Wherever the QD layer overlies the circuit, it continuously overlaps and contacts at least some of the features of the circuit. If the QD layer overlies three-dimensional features of the circuit, the QD layer conforms to those features. In other words, there is a substantially contiguous interface between the QD layer and the underlying CCD or CMOS electronic read-out integrated circuit. One or more electrodes in the CCD or CMOS circuit contact the QD layer and are capable of relaying information about the QD layer, e.g., the amount of light on it, to a readout circuit. The QD layer can be provided in a continuous manner to cover the entire underlying circuit, such as a readout circuit, or patterned. If in a continuous manner, the fill factor can approach about 100%, which is much greater than known CMOS pixels; with patterning, the fill factor is reduced, but can still be much greater than a typical 35% for a CMOS sensor.

(37) In many embodiments, the QD optical devices are readily fabricated using standard CMOS techniques. For example, a layer of QDs can be solution-coated onto a pre-fabricated CCD or CMOS electronic read-out circuit using, e.g., spin-coating, which is a standard CMOS process, and optionally further processed with other CMOS-compatible techniques to provide the final QD layer for use in the device. Details of QD deposition and further processing are provided below. Because the QD layer need not require exotic or difficult techniques to fabricate, but can instead be made using standard CMOS processes, the QD optical devices can be made in high volumes, and with no significant increase in capital cost (other than materials) over current CMOS process steps.

(38) Individual features and embodiments of QD devices, and methods of making same, will now be described in greater detail.

(39) Electronic Read-Out Integrated Circuit

(40) FIG. 3A shows an optical micrograph of a light sensitive layer formed on a commercially available electronic read-out integrated circuit (ROIC) chip. In many embodiments, the light sensitive layer includes a plurality of QDs, as described in greater detail below. The light sensitive layer, e.g., the QD layer, overlays and conforms to the features of the underlying chip. As can be seen in FIG. 3A, the electronic read-out chip, e.g., a CCD or CMOS integrated circuit, includes a two-dimensional array of rows 310 and columns of electrodes 320. The electronic readout chip also includes a two-dimensional array of square electrode pads 300 which together with the overlying QD layer and other circuitry form a two-dimensional array of pixels. The row electrodes 310 and column electrodes 320 allow each pixel (including square electrode pad 300 and overlying QD layer) to be electronically read-out by a read-out circuit (not shown) that is in electrical communication with the electrodes. The resulting sequence of information from the ROIC at the read-out circuit corresponds to an image, e.g., the intensity of light on the different regions of the chip during an exposure period, e.g., a frame. The local optical intensity on the chip is related to a current flow and/or voltage bias read or measured by the read-out circuit.

(41) FIG. 3B shows a schematic cross-section of the imaging system shown in FIG. 3A. The imaging system includes a read-out structure that includes a substrate 32, an optically sensitive layer 38, e.g., QD layer 38 and a transparent electrode 36. Substrate 32 includes a read-out integrated circuit (ROIC) having a top surface with an array of pixel electrodes 34 located in the top surface thereof with counter-electrodes 36 located outside the array, i.e., transparent electrode 36 overlaying QD layer 38. Electrodes 34 shown in FIG. 3B correspond to square electrode pads 300 shown in FIG. 3A. The array of electrodes 34, which together form focal plane array 30, provide voltages for biasing and current collection from a given pixel 34 of the array, and convey signals from the array to an input/output electrode 32 (connection not shown). Optically sensitive layer 38, e.g., QD layer, is formed on the top surface of the integrated circuit. More specifically, QD layer 38 overlays the array of pixel electrodes 34 on the top surface of the integrated circuit. The optically sensitive layer 38 defines an array of imaging pixels for collecting light incident thereon.

(42) In the imaging system of FIG. 3B, QD layer 38 is monolithically integrated directly onto the electronic read-out chip. In contrast, as previously mentioned, existing imaging systems are often made by separately fabricating 1) the read-out integrated circuit and 2) the sensitizing semiconductor array, and subsequently assembling the two, e.g., using a process such as microbump bonding.

(43) Referring now to FIG. 4A, there is shown at 40 a side view of a basic optical device structure, which in certain embodiments can be used as an individual pixel in the completed integrated array shown in FIGS. 3A-3B. Device 40 includes substrate 42, which may be glass or other compatible substrate; contact/electrode 44; optically sensitive layer, e.g., QD layer 48; and at least partially transparent contact 45 that overlays the QD layer. Contacts 44 and 45 may include, e.g., aluminum, gold, platinum, silver, magnesium, copper, indium tin oxide (ITO), tin oxide, tungsten oxide, and combinations and layer structures thereof, and may include band-pass or band-block filters that selectively transmit or attenuate particular regions of the spectrum which are appropriate to the end use of the device. The device has an overall vertical sandwich architecture, where different components of the device generally overlay other components. In operation, the amount of current that flows and/or the voltage between contact 45 and contact 44 is related to a number of photons received by QD layer 48. In operation, current generally flows in the vertical direction. The embodiment shown in FIG. 4A may also include one or more additional optional layers for electron/hole injection and/or blocking. The layer(s) allows at least one carrier to be transported, or blocked, from an electrode into the QD layer. Examples of suitable layers include a QD layer including QDs of a different size and/or composition, semiconducting polymers, and semiconductors such as ITO and Si.

(44) Referring now to FIG. 4B, there is shown at 40 a side view of a basic device structure which has a different configuration than each pixel in the completed integrated array shown in FIGS. 3A-3B, but which could be used to form a similarly functioning optical device. The configuration in FIG. 4B corresponds to a lateral planar structure in which the optically sensitive layer 48 is deposited across two spaced contacts/electrodes 44 and 46. Contacts 44 and 46 are deposited on a substrate, e.g., glass substrate 42. The integrated circuit, including contacts 44, 46, and substrate 42 may include any appropriate system with which the optically sensitive material is compatible (e.g., Si, glass, plastic, etc.). Contacts 44 and 46 may include aluminum, gold, platinum, silver, magnesium, copper, indium tin oxide (ITO), tin oxide, tungsten oxide, or combinations or layer structures thereof. The device has an overall lateral planar architecture, where at least some of the components of the device are generally laterally displaced from other components, forming a planar electrode structure. In operation, the amount of current that flows and/or the voltage between contact 44 and contact 46 is related to a number of photons received by the QD layer 48. In operation, current generally flows in the lateral direction.

(45) FIG. 4C shows a plan view of another basic device structure 40 that includes interdigitated electrodes, and which also can be used to form an optical device. The materials may be selected from those provided above regarding FIGS. 4A-4B.

(46) Each basic device 40, 40, and 40 as shown in FIGS. 4A-4C, among other possible architectures, can be thought of as representing a single device or an element in a larger device, such as in a linear array or a two-dimensional array. The basic devices can be used in many kinds of devices, such as detection and signal processing, as discussed above, as well as in emission and photovoltaic devices. Not all embodiments need be optical devices. Many QD layers have optical characteristics that can be useful for optical devices such as image sensors useful in one or more of the x-ray, ultraviolet, visible, and infrared parts of the spectrum; optical spectrometers including multispectral and hyperspectral; communications photodetecting optical receivers as well as free-space optical interconnection photoreceivers; and environmental sensors; Some QD layers also have electrical characteristics that may be useful for other kinds of devices, such as transistors used in signal processing, computing, power conversion, and communications.

(47) In one embodiment, the underlying electrodes on the integrated circuit define imaging pixels in an imaging device. The QD layers formed on the electrodes supply optical-to-electrical conversion of incident light.

(48) In another embodiment, in addition to the definition of pixels via electrodes on the integrated circuit, further patterning of the optically sensitive layers, e.g., QD layers, provides further definition of pixels, including of which pixel is read by which electrodes on the integrated circuit. This patterning may also be accomplished with well-known CMOS techniques such as photolithography. Other options include self-assembly of QD layers onto pre-patterned metal layers, such as Au, to which the QDs and/or their ligands have a known affinity. Patterning may also be achieved by depositing a conformal QD layer onto a topologically-variable surface, e.g., including hills (protrusions) and valleys (trenches) and subsequently planarizing the QD film to remove material accumulated on the hills while preserving that in the valleys.

(49) Further layers may be included in the layers atop the structure, such as electrical layers for making electrical contact (e.g. an at least partially transparent contact such as indium tin oxide, tin oxide, tungsten oxide, aluminum, gold, platinum, silver, magnesium, copper, or combinations or layer structures thereof), antireflection coatings (e.g. a series of dielectric layers), or the formation of a microcavity (e.g. two mirrors, at least one formed using nonabsorbing dielectric layers.), encapsulation (e.g. an epoxy or other material to protect various materials from environmental oxygen or humidity), or optical filtering (e.g. to allow visible light to pass and infrared light not to, or vice-versa.)

(50) The integrated circuit may include one or more semiconducting materials, such as, but not limited to silicon, silicon-on-insulator, silicon-germanium layers grown on a substrate, indium phosphide, indium gallium arsenide, gallium arsenide, or semiconducting polymers such as MEH-PPV, P3OT, and P3HT. The integrated circuit may also include one or more semiconducting organic molecules, non-limiting examples being end-substituted thiophene oligomers (e.g. alpha,w-dihexyl hexathiophene (DH6T)) and pentacene. Polymers and organic molecules can be useful as a substrate in the QD devices because they may be flexible, and thus allow bendable and conformable devices to be made that are thus nonplanar.

(51) Other appropriate substrates may include, e.g., plastic and glass.

(52) Optically Sensitive Layer

(53) The optically sensitive layer includes a material that is optically sensitive in any one or more of the infrared, visible and ultraviolet region of the electromagnetic spectrum. As discussed above, in many embodiments the optically sensitive layer includes one or more types of quantum dot nanocrystals (QDs), which may be fused together.

(54) In some embodiments, the optically sensitive layer includes a combination of two or more types of QDs, each including a distinct semiconductor material and/or having distinct properties. The different types of QDs may be separately synthesized and mixed prior to being applied to the surface of the integrated circuit or they may be synthesized in a one pot synthesisi.e. in a single vessel.

(55) In some embodiments, the optically sensitive layer includes an optically sensitive semiconducting polymer such as, but not limited to MEH-PPV, P3OT and P3HT. In other embodiments the optically sensitive layer includes a polymer-QD mixture having one or more types of QDs that sensitive to different parts of the electromagnetic spectrum.

(56) A. Quantum Dot Nanocrystals

(57) In many embodiments, the QDs are fabricated using known techniques, but in substantially inert, anhydrous environments, e.g., environments that are substantially free of water and oxygen. Syntheses may be performed using Schlenk line methods in which ambient gases such as oxygen and water in the air are excluded from the system, and the syntheses are instead performed in the presence of substantially inert gases such as nitrogen and/or argon, or in a vacuum.

(58) In some embodiments, the QDs include any one or combination of PbS, InAs, InP, PbSe, CdS, CdSe, ternary semiconductors, and a core-shell type semiconductors in which the shell is one type of semiconductor and the core is another type of semiconductor. For example, the ternary QDs may be In.sub.xGa.sub.1-xAs nanocrystals or (CdHg)Te nanocrystals. For example, the core-shell quantum dot nanocrystals may be ZnSe(PbS), ZnS(CdSe), ZnSe(CdS), PbO(PbS), or PbSO4(PbS).

(59) In some embodiments, before depositing the QD precursor layer on the integrated circuit or substrate, the QDs are ligand exchanged to substitute the as-fabricated ligands with pre-selected ligands, e.g., ligands that are considerably shorter than the as-fabricated ligands. The pre-selected ligands are selected to be sufficiently short to enable closer packing of the QDs in the precursor layer. Closer packing allows the QDs to fuse together in a subsequent step, thereby greatly increasing the electrical conductivity between the QDs. The pre-selected ligands may also be selected to be relatively volatile, so that they can be vaporized during a subsequent step to provide a film consisting mainly of QDs and being substantially free of ligands. This allows the QDs to get much closer to each other, which may enhance the conductivity in the final device. For example, the QDs may be fabricated with a first set of ligands with carbon chains that are more than 10 carbons long; the first set of ligands is then substituted with a second set of ligands with carbon chains that are between 1-10 carbons long. In some circumstances, the ligands of the second set of ligands is less than about 1 nm long. This can bring the QDs closer, e.g., more than 50% closer, more than 75% closer, or even more than 90% closer, than they could get before ligand exchange. The second set of ligands may generally have an affinity for attachment to the QDs that is at least competitive with the affinity of the first set of ligands to attach to the QDs, otherwise the first set of ligands may not sufficiently exchange with the first set of ligands. The second set of ligands may also generally have an affinity for attachment to the QDs which allows them to be removed during a later step. This affinity is related to the end functional group on the ligand, which is illustrated in FIG. 1. Amines, thiols, carboxylates, and sulfones, among other end functional groups, many of which will have free electron pairs, are generally suitable for use in the second (pre-selected) set of ligands.

(60) In some embodiments, the ligand exchange involves precipitating the as-synthesized QDs from their original solution, washing, and redispersing in a liquid that will dissolve and thus dissociate the original ligands from the outer surfaces of the QDs, and which either is or contains the ligands to be substituted onto the QDs. In some embodiments the liquid is or includes primary, secondary, or tertiary-butylamine, pyridine, allylamine, methylamine, ethylamine, propylamine, octylamine, or pyrrolidine or a combination of these organic solvents, which substitute the ligands previously on the QDs. In other embodiments, the liquid is or includes pyridine, which substitutes the ligands previously on the QDs. Leaving the QDs in this liquid for between 24 and 120 hours either at room temperature or at an elevated temperature is generally sufficient for ligand exchange, although in some circumstances longer or shorter times will be sufficient. In an illustrative example, the ligand exchange process was performed under an inert atmosphere to prevent the QDs from oxidation. QDs having oleate ligands and dissolved in methanol were precipitated, dried, and redispersed in n-butylamine at a concentration of 100 mg/ml (nanocrystals by weight/butylamine by volume). The solution was left for 3 days under inert conditions. The oleate ligands had a length of about 2.5 nm, and the exchanged butylamine ligands had a length of about 0.6 nm, bringing the QDs to about 25% of their original distance from each other.

(61) In some embodiments, two or more types of QDs are separately fabricated in coordinating solvents. Each kind of QD is then precipitated, washed, and dispersed in a liquid that is or contains the ligands to be substituted onto the QDs. Tills exchanges the ligands on the two or more types of QDs as discussed above. Then the two types of QDs are mixed in solution to create a heterogeneous QD mixture, which is spin-cast or otherwise deposited as thin films on a substrate to form a heterogeneous QD precursor layer. The order in the heterogeneous QD precursor layer is controlled through separate selection of QD size and ligand for each type of QD and additional treatment with solvents and heating.

(62) Examples of ligands include amine-terminated ligands, carboxyl-terminated ligands, phosphine-terminated ligands and polymeric ligands. The amine-terminated ligands may include any one or combination of pyridine, allylamine, methylamine, ethylamine, propylamine, butylamine, octylamine, and pyrrolidine. The carboxyl-terminated ligands may include any one or combination of oleic acid, stearic, capric and caproic acid. The phosphine-terminated ligands may include guanosine triphosphate. The ligand may be one or more of DNA, an oligonucleotide, a polymer such as polythiophene or MEH-PPV, or an oligomer such as oligothiophene. As mentioned above, it can be useful to substitute short and volatile ligands, e.g., pyridine, allylamine, methylamine, ethylamine, propylamine, butylamine, octylamine, or pyrrolidine, onto the QDs so that the QDs can be brought into closer proximity in later steps.

(63) B. Forming Precursor QD Layer on Integrated Circuit

(64) After the QDs are fabricated and ligand-exhanged, e.g., as described above, they may be deposited onto a substrate such as an integrated circuit. This forms a QD precursor layer, which may be subsequently processed to form a finished QD layer for use in a device.

(65) The QD precursor layer may be formed by solution-depositing it directly on the surface of a read-out integrated circuit or other substrate, for example using spray-coating, dip-casting, drop-casting, evaporating, or blade-casting. Another method of depositing the QD precursor layer is spin coating the QD precursor layer, which once spin-coated onto the surface may be further processed to form the optically sensitive QD layer as described below. In many embodiments, the QD layer has a thickness selected to absorb most or even substantially all of the light incident on it, in the wavelength region the device is intended to operate in. Typically this thickness will range between about 50 nm and 2 m, though thinner or thicker films can be used according to the desired functionality of the device. Spin-coating can allow the process of covering circuitry with a QD layer to be performed at lower temperatures without vacuum processing and alignment and bonding issues.

(66) C. Ligand Removal and Annealing of QD Precursor Layer

(67) After forming the QD precursor layer, the QDs may be fused together to produce a QD film with enhanced optical and electrical characteristics, and which is suitable for use in a finished electronic or optoelectronic device.

(68) In one embodiment, at least a portion of the QDs in the QD precursor layer are fused by annealing the layer at temperatures up to about 450 C., or between about 150 C. and 450 C. In other embodiments, the layer is treated at lower temperatures, for example between about room temperature up to about 150 C., or up to about 100 C., or up to about 80 C. In some embodiments, the QD precursor layer is not heated substantially above ambient (room) temperature. As mentioned above, the step of fusing brings the cores of adjacent QDs into direct physical and electrical contact. It is also possible to overfuse the QDs, in which case they may lose their individual characteristics and appear more like a bulk semiconductor material. It is desirable to prevent such overfusing through the parameters chosen for annealing or through monitoring to prevent an overfused condition. The annealing step will typically be performed in a vacuum or in an otherwise anhydrous environment to prevent the development of defect states (e.g., oxidation) on the outer surfaces of the QDs before the cores of the QDs fuse together. This way, there will be substantially no defect states in the regions where the QDs are joined together, but these regions instead will have a substantially homogeneous composition and crystalline structure. In other embodiments the fusing step may be performed in an oxygen-rich environment, or an oxygen environment in which the partial pressure of oxygen is regulated.

(69) The ligands in the QD precursor layer are also typically removed, either before or concurrently with the fusing step. For example, ifthe ligands in the QD precursor layer are volatile, they may easily be removed during annealing because they will simply volatilize from the heat. Or, for example, if the ligands in the QD precursor layer are not volatile, they can be removed from the QD precursor layer by soaking the layer in a solvent that dissolves and thus dissociates the ligands from the QDs but which does not generally disrupt the arrangement of QDs in the QD layer. In general, it is preferable that removing the ligands does not significantly change the volume of the QD layer, e.g., by less than about 30%; a large volume change may crack or otherwise damage the finished QD film.

(70) D. Creation of Defect States on Outer Surfaces of Fused QDs

(71) In many embodiments, particularly those suitable for optical applications, defect states are created on the outer surfaces of the fused QDs. By defect state it is meant a disruption in the otherwise substantially homogeneous crystal structure of the QD, for example, the presence of a dislocation or a foreign atom in the crystal lattice. In many cases this defect state will exist on the outer surface of the QDs. A defect state can be created by, e.g., oxidizing the QDs after fusing and ligand removal. During operation, if an electron-hole pair is generated within the QD film, one or more holes may be trapped by the defect state; this will preclude rapid recombination of holes with electrons, which will then allow the electrons to flow for a much longer time through the film. This can positively affect photoconductive gain, among other things.

(72) In general, the outer surface of the fused QDs can be coated or otherwise treated so it has a different composition than the cores of the fused QDs. For example, the outer surface can include a semiconductor or insulator shell.

(73) E. Summary of Steps in Fabricating QD Layer

(74) FIG. 5 shows a flow chart of steps in a method for creating various embodiments of QD layers for use in optical devices.

(75) First, the QDs are fabricated (500), e.g., using well-known techniques. The QDs will typically include a plurality of relatively long ligands attached to their outer surfaces.

(76) Then, the QDs are ligand-exchanged (510), e.g., by substituting shorter ligands for those used during fabrication of the QDs. This step may allow the QDs to pack more closely in subsequent processing steps.

(77) Then, the QDs are deposited on a suitable substrate (520), e.g., on an electronic read-out integrated circuit. This step may be accomplished with various solution-based methods, many of which are compatible with standard CMOS processes such as spin-coating.

(78) Then, the precursor layer is washed to remove the ligands on the QDs, and to cause necking (i.e. touching) between at least some adjacent QDs (540).

(79) Then, the necked QD layer is annealed, which fuses necked QDs together (540).

(80) Then, defect states are created in the fused QD layer (550), e.g., by oxidizing the layer.

(81) In general, when fabricating a device intended to have multiple pixels, the QD layer may then optionally be patterned, e.g., using photolithography, to separate the continuous layer into a plurality of pixels.

(82) The resulting QD layer can be incorporated into devices such as those described herein.

Examples

(83) An exemplary photoconductive detector was made using a single layer of PbS QD nanocrystals spin-cast directly from a chloroform solution onto an interdigitated electrode array. The device structure is illustrated in FIG. 7A, and is analogous to the basic device of FIG. 4B. The parallel gold electrodes are supported by a glass substrate and have a height, width, and separation of 100 nm, 3 mm, 5 m, respectively. The thickness of the QD layer was controlled through the concentration of the chloroform-QD solution and the spin-casting parameters. In studies carried out by the inventors the thickness ranged from 100 nm up to 500 nm.

(84) The treatment of the surfaces of the QDs was an important determinant of photodetector performance. Devices made directly from QDs capped with oleic acid, as synthesized through an organometallic route, did not exhibit any measurable conductance, as the 2 nm-long oleate ligand inhibits carrier transport among QDs. A post-synthesis ligand exchange was therefore used to replace the as-synthesized oleate ligands with much shorter butylamine ligands. To this end, the QDs were redispersed in butylamine for a period of three days. Butylamine is a four-carbon-atom chain with an amine head as the functional group to attach to the QD surface. The ligand exchange was monitored for blue shift in QD absorption resulting from a decrease in QD effective diameter as ligands remove Pb atoms during exchange

(85) FIG. 6A shows absorption spectra and TEM images of lead sulphide QD nanocrystals as the ligands are exchanged from oleic acid to primary butylamine. The TEM images illustrate the dramatic decrease of inter-QD spacing following ligand exchange and nonsolvent treatment. The absorption spectrum shifts steadily to the blue with increasing exchange time. When the shift is less than that associated with the removal of a monolayer of Pb atoms (roughly 170 nm), the size distribution remains roughly constant. After this point the polydispersity increases. In the example provided, the best device performance was obtained using QD nanocrystals shifted by 170 nm.

(86) The QDs were precipitated, washed using a nonsolvent, redispersed in CHCl.sub.3, and treated again using a nonsolvent (nonsolvent refers to a material that is not a solvent for the nanocrystals, but that may be a solvent for the ligands). The impact of ligand exchange and nonsolvent treatment on QDs is illustrated in the transmission electron micrographs of FIG. 6A. The as-grown (untreated) QD nanocrystals show well-ordered patterns with interdot spacing determined by ligand length. Exchanged and washed QDs exhibit a drastic reduction in interdot spacing and preferential formation of clusters instead of well-ordered arrays. Prior to treatment, the nanocrystal films can be redispersed using organic solvents, while after treatment, nanocrystal films can no longer be readily redispersed.

(87) The combination of ligand exchange, nonsolvent treatment, and thermal processing at temperatures such as up to about 150 C. (typically) and potentially as high as 450 C., removes at least a portion of the QDs' ligands, and enables the QDs to fuse, providing mechanically robust films with vastly increased electrical conductivity, as reported below.

(88) FIG. 6B shows the absorbance spectra of QD nanocrystals before ligand exchange (oleate-capped), after ligand exchange (butylamine-capped), and following soaking in methanol for 2 hours to remove the butylamine ligands. The progressive blueshift across these treatments is consistent with surface modification following exchange and partial surface oxidation (also confirmed by XPS and FTIR). The inset of FIG. 6B shows TEM micrographs of the nanocrystals before and after ligand exchange. The reduction in interparticle distance is attributed to the replacement of the oleic acid ligands with butylamine ligands.

(89) FIG. 6C shows the FTIR spectra of the neat solvent n-butylamine, the neat solvent chlorofonn, and n-butylamine-exchanged QDs dispersed in chloroform. NH stretching and bending vibrations are tabulated to lie between 3200-3600 cm.sup.1 and 1450-1650 cm.sup.1 respectively. Carbonyl stretching vibration of pure oleic acid is tabulated to be found at 1712 cm.sup.1. The results indicated that oleate ligands originally attached to the PbS QDs have been replaced by n-butylamine, indicated by the absence of carbonyl stretching vibration, a significant shift of the NH stretching vibrations after exchange from 3294 and 3367 cm.sup.1 (=73 cm.sup.1) for n-butylamine to 3610 and 3683 cm.sup.1 (=73 cm.sup.1), and the presence of NH bending vibrations for the n-butylamine exchanged sample.

(90) FIG. 6D shows the FTIR spectra of inert-exchanged ligand-exchanged QDs with butylamine ligands before and after methanol wash, which substantially removes the ligands from the QDs. Following methanol wash, features attributable to butylamine (1400, 1126, 989, 837, and 530 cm.sup.1) are much less pronounced. The inset also shows the NH stretching vibrations, which are again much less pronounced following methanol wash.

(91) FIG. 6E shows spectra obtained by X-ray photoelectron spectroscopy (XPS) to confirm the material modifications that occur to the PbS QDs throughout various processing steps. After background subtraction, the binding energy was reference to the Cls hydrocarbon line at 285.0 eV. The curves were fitted by applying Gaussian-Lorenzian functions and the atomic ratios were obtained by integrating the areas under the signals. The nanocrystals immediatedly after exchange to butylamine ligands demonstrate a S2-peak at 160.7 eV corresponding to lead sulfide. No lead sulfate (PbSO.sub.4) signal was detected. Nanocrystals that were precipitated in air exhibit an SO.sub.4.sup.2 at 167.5 eV characteristic of PbSO.sub.4 formation. This oxide may be associated with the role of barrier to conduction among nanocrystals. The ratio of PbS/PbSO4 for this case was found to be about 3.4:1. XPS of the inert-precipitated QDs after methanol soaking exhibits also formation of lead sulfate. The PbS/PbSO.sub.4 ratio in this case was 18.6:1. Further annealing of this film in air at 120 C. for 1 hour dramatically increased the amount of sulfate and the PbS/PbSO.sub.4 ratio was 2.44:1.

(92) FIG. 6F shows the FTIR spectra of ligand-exchanged QDs precipitated in inert conditions (butylamine-called QDs) and precipitated in air-ambient conditions (oxidize-then-neck QDs). The inert-precipitated exchanged QD layer after 2 hours of methanol wash (neck-then-oxidize QDs) are also shown. The broad feature around 1147 cm.sup.1 is attributed to PbSO.sub.4 (lead sufate). The spectra show that ligand-exchanged QDs precipitated in inert conditions do not show this feature; methanol wash introduces some oxidation; ligand-exchanged QDs precipitated under an air ambient show evidence of strong oxidation. These results agree with the XPS data above.

(93) Some performance characteristics of various representative devices having different kinds of QD nanocrystallayers (e.g., neck-then-oxidize, oxidize-then-neck, butylamine-capped, and neck-then-overoxidize) were measured. The general device structure also shown in the inset of FIG. 7A, and can be seen to generally be similar to that of FIG. 4B. The device included a transparent glass substrate; two gold electrodes having a length of about 3 m, a width of about 5 m, and being spaced from each other by about 5 m; and a QD nanocrystal of variable thickness between the electrodes.

(94) Photoconduction was studied with the aid of optical excitation through the glass substrate, with excitation light being transmitted through the space separating interdigitated electrodes, i.e., where the QD layer was formed. The current-voltage characteristics for two different QD nanocrystal layer thickness are depicted in FIG. 7A, specifically the I-V characteristic for a thin 100 nm and a thick 500 nm QD nanocrystal layer devices. Photocurrents and dark currents respond linearly to applied bias. The responsivity of the thick device reached 166 A/W. The linear I-V characteristic indicates an ohmic electrode-nanocrystal contact and suggests not a tunneling but a strong, direct conductive connection between QD nanocrystals. Photocurrent in the thick device is significantly higher than the photocurrent of the thin device by virtue of greater absorbance in the thick device.

(95) In order to determine optical power incident over the detector area and to calculate the responsivity R, a 2 mm radius beam from a 975 nm laser was incident, first through a series of optical attenuators of known transmittance, and thence through the glass substrate, onto the device from the back side. On the top surface, infrared-opaque interdigitated gold electrodes were separated by 5 m over a 3 mm path length. The optical power incident on the device was obtained by integrating the intensity profile of the laser over the unobstructed area of the device. Current-voltage characteristics were acquired using an Agilent 4155 semiconductor parameter analyzer. The optical power impinging on each device was about 80 pW.

(96) The responsivity as a function of applied bias of devices made with different kinds of QD nanocrystal layers is shown in FIG. 7B. Here, the nanocrystal layers were about 800 nm thick. The neck-then-oxidize QD device, corresponding to a device having a layer of fused QDs with defect states on their outer surfaces, can clearly be seen to have a significantly higher responsivity than the other devices. The oxidize-then-neck QD device, in which the ligands are removed from the QDs, and the QDs are fused, but in which the QDs are not maintained in an inert atmosphere between the steps of ligand removal and QD fusing, has defect states in the regions in which the QDs are joined that reduces their responsivity, as compared with the neck-then-oxidize device, in which the QDs are maintained in an inert atmosphere between the steps of ligand removal and QD fusing. All of the necked devices have a significantly higher responsivity than the device having butylamine capped QDs, in which the butylamine ligands block facile conduction of electrons between QDs.

(97) In general, the responsivity of QD devices (particularly the neck then oxidize QD devices) as measured in A/W is at least about 10 A/W, 100 A/W, 1000 A/W, or even more than about 10000 A/W. The responsivity is a function in part of the bias voltage applied, with a greater responsivity at higher bias. In some embodiments, the QD devices (particularly the neck then oxidize QD devices provide a substantially linear responsivity over 0-10 V with a bias applied across a distance of 0.2 to 2 m width or gap.

(98) FIG. 7C shows the dark current density for the devices described regarding FIG. 7B. As for the responsivity, necked devices have a significantly higher dark current density than the device having butylamine capped QDs. Devices made using QDs exposed to oxygen before necking (oxidize-then-neck) show a superlinear I-V behavior characteristic of field-assisted transport. In contrast, devices made using QDs fused before oxidation (neck-then-oxidize) exhibit linear (field-independent) behavior. Further oxidation of neck-then-oxidize devices (neck-then-overoxidize) leads to a decrease of conductivity owing to excessive oxide formation.

(99) FIG. 7D shows the measured noise current as a function of the measured dark current for the devices described regarding FIG. 7B. Neck-then-oxidize devices exhibited the lowest noise current, approaching within 3 dB the shot noise limit. Oxidize-then-neck devices had the highest noise current, consistent with multiplicative noise. Neck-then-overoxidize QD devices showed lower noise levels than the oxidize-then-neck QD devices although they contained larger amounts of oxide. This indicates the role of the oxidiation step in the fabrication process. The Johnson noise limit, the shot-noise limit, and the fundamental background-limited thermodynamic (BLIP) noise current of the best-performing device (neck-then-oxidize) are also plotted for comparison.

(100) FIG. 7E shows a plot of the normalized detectivity D* as a function of applied bias. The normalized detectivity D* is measured in units of Jones (cmHz.sup.1/2 W.sup.1). D* is given as (Af).sup.1/2R/I.sub.n, where A is the effective area of the detector in cm.sup.2, f is the electrical bandwidth in Hz, and R is the responsivity in AW.sup.1 measured under the same conditions as the noise current i.sub.n in A. The material figure of merit D* allows comparison among devices of different powers and geometries. The device figure of merit, noise equivalent power (NEP)the minimum impinging optical power that a detector can distinguish from noiseis related to D* by NEP=(Af).sup.1/2/D*. As can be seen in FIG. 7E, the normalized detectivity D* is the highest for the neck-then-oxidize device, and the lowest for the oxidize-then-neck device. In other words, allowing the QDs to be exposed to oxygen after ligand removal and before necking or fusing significantly affects the normalized detectivity of the finished device. In the example devices shown, the normalized detectivity of the neck-then-oxidize device is more than an order of magnitude higher than that for the oxidize-then-neck device. The highest detectivity was found at a modulation frequency of 30 Hz, and reached 1.310.sup.13 jones at 975 nm excitation wavelength.

(101) FIG. 7F shows the results of further measurements of the neck-then-oxidize device described above regarding FIG. 7B. The spectra of responsivity and the normalized detectivity D* is shown for the neck-then-oxidize device at an applied bias of 40 V and an electrical frequency of 10 Hz. D* was measured to be 1.810.sup.13 jones at the excitonic peak wavelength. FIG. 7G shows the electrical frequency response of the same device under 40 V bias. The 3-dB bandwidth of the detector is about 18 Hz, consistent with the longest excited-state carrier lifetime in the device. High sensitivity (D*>10.sup.13 jones) is retained at 20 imaging rates of about 30 frames per second.

(102) FIG. 7G shows the photocurrent temporal response following excitation of the QD layer of the neck-then-oxidize device of FIG. 7B, where the excitation was a 7 ns pulse centered at 1064 nm, at a frequency of 15 Hz. This allows investigation of the transit time and distribution of carrier lifetimes in the device. The response of the detector to the optical pulse was found to persist over tens of milliseconds, attributable to the longest-lived population of trap states introduced by oxidation. The response exhibits multiple lifetime components that extend from microseconds (though shorter components may exist they are not observable in this measurement) to several milliseconds. Decay components were found at about 20 s, about 200 s, about 2 ms, about 7 ms, and about 70 ms. A transit time of about 500 ns was obtained for a bias of about 100 V, revealing that transit time depends linearly on bias with a slope corresponding to a mobility of about 0.1 cm.sup.2V.sup.1 s.sup.1. The ratio of the longest component of carrier lifetime to the transit time was thus on the order of about 10,000. The observed responsivity of about 2700 A/W in this example could thus be explained by photoconductive gain, given the films absorbance of 0.3 at an optical wavelength of 975 nm. This high responsivity was observed under low-level optical power conditions relevant to ultrasensitive detection. In general, in some embodiments, as illumination intensity was increased, the longest-lived trap states became filled, and shorter lived, so lower-gain trap states began to account for a significant component of carrier lifetime. The devices of these embodiments were thus highly sensitive under low-light conditions, and exhibit intrinsic dynamic-range-enhancing gain compression under increasing illumination intensity.

(103) For determining the photocurrent spectral response, a bias of 50 V was applied to the sample connected in series with a 100 Ohm load resistor. Illumination was provided by a white light source dispersed by a Triax 320 monochromator and mechanically chopped at a frequency of 100 Hz. Filters were used to prevent overtones of the monochromator's grating from illuminating the sample. The voltage across the load resistor was measured using a Stanford Research Systems SR830 lock-in amplifier. The intensity through the monochromator at each wavelength was measured separately using a calibrated Ge photodetector. The photo current at each wavelength was subsequently scaled accordingly. After the photocurrent spectral shape was determined in this way, the absolute responsivity at 975 nm was used to obtain the absolute spectral response 800 nm-1600 nm, which is shown in FIG. 8.

(104) For measurement of noise current and calculation of NEP and D*, the photoconductive device was placed inside an electrically-shielded and optically-sealed probe station and connected in series with a Stanford Research SR830 lock-in amplifier. Alkaline batteries were used to bias the device for the measurement of the noise current in order to minimize noise components from the source. The lock-in amplifier measured the current in the photodetector and reported noise current in A/Hz.sup.1/2. Special care was taken in choosing an appropriate passband in order to acquire stable and meaningful measurements of noise current at various frequencies. This measurement revealed a significant increase in the noise current below 5 Hz which is attributed to 1/f noise, while white noise patterns are observed above 50 Hz. The noise current divided by the responsivity under the same measurement conditions of applied bias and frequency modulation yielded the noise equivalent power (NEP). The normalized detectivity, D*, was obtained as a function of wavelength, applied bias, and frequency by dividing the square root of the optically active area of the device by the NEP.

(105) To validate the NEP values obtained using this technique, the identical procedure was preformed using a commercial Si detector with known NEP. The system described above reported NEP values of the same order of magnitude, but typically somewhat larger than, the specified NEPs. The NEP and D* determination procedure used herein thus provides a conservative estimate of these figures of merit.

(106) FIG. 8 shows spectral dependence of responsivity and normalized detectivity D* for biases of 30, 50, and 100 V, under 5 Hz optical modulation and 0.25 nW incident optical power. The responsivity shows a local maximum near 1200 nm corresponding with the exciton absorption peak of the nanocrystal solid-state films shown in the inset of FIG. 8. The responsivity increases with voltage (but not as rapidly as does the noise current, resulting in higher D* at lower biases) and reaches 180 A/W at 800 nm. For 30 and 50 V of applied bias, D* is 210.sup.11 Jones and is more than double the detectivity of commercial polycrystalline PbS detectors which have benefited from 50 years of science and technology development. Though the responsivity is higher at 100 V, the bias-dependence of the measured noise current results in D* being maximized at the lower bias of 30 V.

(107) FIG. 9 shows the frequency dependence of responsivity and normalized detectivity for three values of applied bias at 975 nm and 0.25 nW of incident optical power. The 3 dB bandwidth of the device responsivity was 15 Hz for 100 V and 50 V and 12 Hz for 30 V. The measurements were taken with optical excitation from a 975 nm laser and incident optical power 0.2 nW. The noise current was also measured for the three different biases throughout the frequency range. The noise current was significantly higher at frequencies below 20 Hz, while frequency-independent white noise was observed at higher frequencies. Noise equivalent exposure, or NEE, is another way of expressing the lowest amount of light detectable by a detector. NEE is defined to be the number of joules of optical energy that can create a signal at the detector that is equivalent in magnitude to the noise on the detector, and is calculated to be the RMS noise on the detector, divided by the responsivity of the detector. FIG. 10 shows the NEE of a QD device having a layer of fused QDs with defect states (e.g., oxidation) on their outer surface, as compared with the NEE of a conventional Si CCD detector as well as a conventional Si CMOS detector. The QD device has an NEE of less than 10.sup.11 J/cm.sup.2 at wavelengths of 400 to 800 nm, and further less than 10.sup.10 J/cm.sup.2 at wavelengths of 400 to 1400 nm. The NEEs of the conventional Si devices are significantly higher than that of the QD device, in some cases more than an order of magnitude higher.

(108) The figures of merit obtained from the quantum dot detectors presented herein result from a combination of processing procedures. First, the shortening of the distance between QDs via exchange to a much shorter organic ligand provided enhanced inter-QD conduction. Post-deposition treatment using a nonsolvent and exposure to elevated temperatures in an oxygen-rich atmosphere enabled further ligand removal, QD fusing, and the formation of a native oxide on the QD surface. This oxide has previously been shown in polycrystalline PbS devices to be useful in achieving high D* in photoconductors. However, chemical bath-grown polycrystalline devices with 200 nm domain sizes do not allow refined control over interfaces. In contrast, using pre-fabricated, highly monodisperse, individually single-crystal QDs with highly-controlled ligand-passivated surfaces to fabricate optical devices allows exceptional control over interface effects compared with polycrystalline-based devices. The quantum dot optical devices described herein are superior across many figures of merit to conventional grown-crystal semiconductor optical devices. At the same time the fabrication of the devices is strikingly simple, while maintaining optical customizability based on the quantum size effects of quantum dots.

Alternative Embodiments

(109) Although the QDs are solution-deposited in the described embodiments, the QDs may deposited in other ways. As mentioned above, one motivation for using solution-deposition is its ready compatibility with existing CMOS processes. However, satisfactory devices can be made by vacuum-depositing or otherwise depositing the QDs.

(110) Other embodiments are within the following claims.

INCORPORATED REFERENCES

(111) The following references, in some cases referred to above as the incorporated references, are incorporated herein by reference in their entireties. Sargent, E. H., Infrared quantum dots. Advanced Materials 17, 515-522 (2005). Tessler, N., Medvedev, V., Kazes, M., Kan, S. & Banin, D., Efficient near-infrared polymer nanocrystal light-emitting diodes. Science 295, 1506-1508 (2002). Bakueva, L., Musikhin, S., Hines, M. A, Chang, T.-W. F., Tzolov, M., Scholes, G. D., Sargent, E. H., Size-tunable infrared (1000-1600 nm) electroluminescence from PbS quantumdot nanocrystals in a semiconducting polymer. Applied Physics Letters 82, 2895-2897 (2003). Rong., H., Jones, R., Liu, A, Cohen, o. Hak, D., Fang, A., Paniccia, M., A continuous wave Raman Silicon laser. Nature 433, 725-728 (2005). McDonald, S. A., Konstantatos, G., Zhang, S., Cyr, P. W., Klem, E. J. D., Levina, L., Sargent, E. H., Solution-processed PbS quantum dot infrared photodetectors and photovoltaics. Nature Materials 4, 138-142 (2005). Lim, Y. T., Kim, S., Nakayama, A., Stott, N. E., Bawendi, M. G., Frangioni, J. F., Selection of quantum dot wavelengths for biomedical assays and imaging. Molecular Imaging 2, 50-64 (2003). Kim, S., Lim, Y. T., Soltesz, E. G., De Grand, A. M., Lee, J., Nakayama, A., Parker, J. A., Mihaljevic, T., Laurence, R. G., Dor, D. M., Cohn, L. H., Bawendi, M. G., Frangioni, J. V., Near-infrared fluorescent type II quantum dots for sentinel lymph node mapping. Nature Biotechnology 22, 93-97 (2004). Ettenberg, M., A little night vision. Advanced Imaging 20, 29-32, 2005. Hines, M. A., Scholes, G. D., Colloidal PbS nanocrystals with size-tunable near-infrared emission: observation of post-synthesis self-narrowing of the particle size distribution. Advanced Matererials 15, 1844-1849 (2003). Yu, D., Wang, C., Guyot-Sionnest, P., n-Type Conducting CdSe Nanocrystal Solids. Science 300, 1277-1280, 2003. Wessels, J. M., Nothofer, H.-G., Ford, W. E., Von Wrochem, F., Scholz, F., Vossmeyer, T., Schroedter, A., Weller, H., Yasuda, A., Optical and electrical properties of three-dimensional interlinked gold nanoparticles assemblies. Journal of the Americal Chemical Society 126, 3349-3356 (2004). Photoconductivity Conference: held at Atlantic City, Nov. 4-6, 1954, Wiley. Konstantatos, G., Howard, I., Fischer, A., Hoogland, S., Clifford, J., Klem, E., Levina, L., Sargent, E. H., Ultrasensitive solution-cast quantum dot photodetectors. Nature 442, 180-183 (2006).