High Carrier Mobility in Single Ultrathin Colloidal Lead SelenideNanowire Field Effect TransistorsRion Graham and Dong Yu*

Department of Physics, University of California, Davis, California 95616, United States

ABSTRACT: Ultrathin colloidal lead selenide (PbSe) nano-wires with continuous charge transport channels and tunablebandgap provide potential building blocks for solar cells andphotodetectors. Here, we demonstrate a room-temperaturehole mobility as high as 490 cm2/(V s) in field effecttransistors incorporating single colloidal PbSe nanowires withdiameters of 6−15 nm, coated with ammonium thiocyanateand a thin SiO2 layer. A long carrier diffusion length of 4.5 μmis obtained from scanning photocurrent microscopy (SPCM).The mobility is increased further at lower temperature,reaching 740 cm2/(V s) at 139 K.

KEYWORDS: Colloidal, nanowires, mobility, field effect transistors, scanning photocurrent microscopy, carrier diffusion length

Colloidal semiconductor quantum dots (QDs) have greatpotential for solar energy and photodetector applica-

tions.1 QDs of various sizes, shapes, and chemical compositionscan be synthesized inexpensively in large quantities2,3 and arereadily dispersed in solution, enabling photovoltaic cellproduction through roll-by-roll technology, similar to a printingpress for newspaper. One major challenge hindering theseapplications is the poor electric conductivity of the QD films inwhich charge needs to inefficiently hop between QDs.4,5

Compared to QDs, colloidal nanorods (NRs) and nanowires(NWs) provide more efficient charge transport channels.Ultrathin colloidal NWs have been synthesized6 and exhibitstrong quantum confinement effects, just like QDs.7−9

Recently, Alivisatos et al. have demonstrated device-scaleperpendicular alignment of colloidal NRs for solar cells.10

Despite their promising potential, much less work11−13 hasbeen done to understand the electronic structure and chargetransport in colloidal NWs compared to QDs. As a result, theirsuperior conductivity has not yet been demonstrated. Thecarrier mobilities in colloidal PbSe NWs have been measured inensemble devices previously to be ∼10 cm2/(V s),13 which islower than the recently reported 27 cm2/(V s) mobility in aCdSe QD thin film with short ligands.14 However, ensemblemeasurements only provide a lower boundary of the actualmobility due to contact barrier limitations and the fact that notall NWs bridge the electrodes in a device.13 By characterizingsingle NW devices, one can greatly reduce these uncertaintiesassociated with ensemble devices. Single colloidal PbSe NWshave been fabricated by directly growing NWs betweenelectrodes, but a low mobility is extracted, likely due toimperfect contacts and/or poor surface passivation.12 A muchhigher mobility of 180 cm2/(V s) has been demonstrated incolloidal PbSe NWs, but with a large diameter of 80 nm.15 It isnot clear if NWs with smaller diameters will maintain a highmobility as surface scattering may substantially reduce the

mobility in ultrathin NWs.16 In addition, the ultrathin channelmay be completely depleted, leading to conduction dominatedby a space charge limited current.17 A comprehensiveunderstanding of charge transport in ultrathin colloidal NWsrequires an experimental investigation of single NW devices.Nevertheless, detailed experimental results have rarely beenreported, partially because of the sensitivity of the colloidalNWs to the harsh fabrication procedure and ambientenvironment.Here, we demonstrate a hole mobility as high as 490 cm2/(V

s) in single ultrathin colloidal PbSe NW field effect transistors(FETs) measured at room temperature and in air. To passivatethe NW surface, we coat the colloidal NWs with a thin layer ofSiO2, which also protects the NWs from being damaged by thelithographic procedures. Devices incorporating single as-grownNWs coated with SiO2 exhibit an average mobility of 30 cm2/(V s). Furthermore, the postgrowth ligand replacement byammonium thiocyanate (NH4SCN) results in better surfacepassivation and leads to a factor of 10 improvement inconductivity and a high mobility of 490 cm2/(V s). Scanningphotocurrent microscopy (SPCM) also supports this enhance-ment with an increase in carrier diffusion length from 2.5 to 4.5μm after ligand exchange. The carrier mobility increases atlower temperature, reaching up to 740 cm2/(V s) at 139 K.PbSe NWs are synthesized following the colloidal method of

Murray et al.,18 isolated through high speed centrifugation (17000 g), dried under vacuum, and stored in a nitrogen glovebox.Selected area electron diffraction (SAED) shows that the NWsare single crystalline and grow along the ⟨100⟩ direction(Figure 1a), as previously reported.6 We have also shown high-

Received: June 7, 2012Revised: July 12, 2012Published: July 23, 2012

Letter

pubs.acs.org/NanoLett

© 2012 American Chemical Society 4360 dx.doi.org/10.1021/nl302161n | Nano Lett. 2012, 12, 4360−4365

resolution transmission electron microscopic (TEM) images ofthese NWs in a previous publication.19 The typical as-grownNWs have diameters of 6−15 nm and lengths of 5−10 μm,confirmed by TEM and atomic force microscopy (AFM). Tofabricate single NW devices, NWs are transferred and cleanedinside a N2 glovebox (H2O < 0.1 ppm, O2 < 0.1 ppm). First, as-grown NWs are dispersed in hexane and drop-cast onto a 300nm SiO2 covered, heavily B-doped Si wafer (Prime grade,University Wafer). The substrate and wires are cleaned ofexcess organics in a bath of anhydrous hexane (95%, Sigma-Aldrich) followed by ethanol (99%, extra dry, Acros Organics),each for 60 s with mild agitation, and dried by gently blowingwith N2. Without this cleaning procedure, the NWs are foundto be covered and surrounded by dark films presumably madeof organic residues such as oleic acid and trioctylphosphine. Forthose that undergo the ammonium thiocyanate ligandreplacement, the NWs are then treated in a 1 wt %/vol.solution of ammonium thiocyanate (>99%, Acros Organics) inanhydrous acetone (99.8%, Acros Organics) for 30 min,followed by soaking in anhydrous acetone for 1 min, and driedwith N2. Before usage, the ammonium thiocyanate is carefullydesiccated (dissolved in isopropanol at 90 °C followed byrecrystallization under N2). AFM (Veeco Dimension 3100)shows the deposited and cleaned wires are smooth andconfirms the NW diameters are between 6 and 15 nm (Figure1(b)).The harsh environment of the electron beam lithography

fabrication procedure leads to insulating devices unless a thinSiO2 layer is first deposited on the NWs for protection. Tominimize air exposure, the sample is transferred in an airtightcontainer (sealed in N2) to a CHA electron beam evaporator.The air exposure time is within 1 min before an immediatepumping of the CHA chamber. Then, a 30 nm layer of SiO2 isdeposited at a rate of 2 Å/s. This layer improves the NWsurface passivation, minimizes exposure to oxygen, and servesas a physical barrier to protect the wires during the fabricationprocess. Individual wires are then located and imaged in a high-resolution FEI 430 NanoSEM. Afterward, a layer of methylmethacrylate (MMA, Microchem) is spin-cast onto thesubstrate and cured on a hot plate, starting at roomtemperature and raised at ∼25 °C/min to 110 °C, for a totaltime of 5 min. This is followed by a similar process for poly(methyl methacrylate) (PMMA, Microchem). Ramping the hot

plate gradually to a maximum 110 °C is critical for preventingthermal damage to these ultrathin NWs. Devices baked at thenormal 180 °C PMMA crystallization temperature or withsudden temperature change do not conduct. Electrodes arethen patterned using electron beam lithography (FEI 430NanoSEM), developed, the SiO2 is etched off at the contactsfor 5 s in a 6:1 buffered HF, and metalized with 25/25 nm Cr/Au in the CHA electron beam evaporator, followed by liftoff inacetone (Figure 1c). A typical SEM image of a complete deviceis shown in Figure 1d.Electrical measurements are made using a National Instru-

ments data acquisition system and a DL Instruments 1211current amplifier. SPCM measurements are performed on asystem as previously described.20 Briefly, a diffraction-limitedlaser spot is raster scanned over a NW FET while reflection andphotocurrent are simultaneously measured. A Coherent CW532 nm laser is used as the illumination source. The beamintensity for SPCM is controlled to be around 100 W/cm2. Thefull width at half-maximum (fwhm) of the laser spot is 830 nm,determined from the photocurrent scan perpendicular to theaxis of the NW. Low-temperature measurements are performedin a Lakeshore optical cryostat and the temperature ismonitored by a Si diode temperature sensor.Both as-grown PbSe NW devices and ammonium

thiocyanate treated NW devices demonstrate nearly linearcurrent−voltage (I−V) characteristics with slight asymmetrybetween positive and negative drain−source voltages at roomtemperature. Four-probe measurements confirm a negligiblecontact resistance of <5% that of the NW resistance for bothtreated and untreated devices. The transparent contact isconsistent with the energy alignment between the Cr workfunction (4.5 eV) and the valence band of the colloidal PbSeNWs (∼4.6 eV).13 A histogram of conductance measured in 35devices (Figure 2c) shows the average conductance of treateddevices is 10 times that of as-grown devices. The ligandreplacement also results in greatly improved device stability.The conductivity of as-grown NW devices decreases substan-tially over a few days of measurements, while the treateddevices maintain their conductivity for over 30 days ofmeasurements and exposure to oxygen.Application of a negative gate voltage leads to an increased

conductivity, implying the wires are p-type (Figure 2(a,b)).From the I−V curves at different gate voltage (Vg), we extractthe hole field-effect mobility, defined as μFE = [L2/(CoxVDS)]-[(dI)/(dVg)], where L is the channel length, VDS is the drain−source voltage, and Cox is the oxide capacitance (Cox =(2πεε0L)/(ln(4h/d)), where h is the oxide thickness). In an as-grown NW device (D1), we obtain μFE = 30 cm2/(V s) andhole concentration p = 1.3 × 1018 cm−3. These values are 490cm2/(V s) and 2.25 × 1018 cm−3 for the ammoniumthiocyanate treated device (D2). In D2, the transconductancechanges impressively by 2 decades upon 1 V change in Vgacross a 300 nm gate oxide. We note the above field-effectmobility is extracted assuming a NW diameter of 10 nm. Theexact diameter of the NWs in the FETs cannot be accuratelymeasured because of the SiO2 coating. As the diameters of thecolloidal NWs range from 6 to 15 nm, the calculated mobilitiescan range from 587 to 434 cm2/(V s).The treated NW devices exhibit a greater than 10-fold

increase in mobility, indicating a suppression of carrierscattering. Recently, the use of ammonium thiocyanate hasled to improved QD thin film mobilities as a result of surfacepassivation and reduced interdot spacing.21 We have also

Figure 1. (a) TEM image of PbSe NWs. (Inset) Electron diffractionpattern taken in the area indicated by the red square. (b) AFM imageof a single PbSe NW on a SiO2 substrate. The height of the NW ismeasured to be 7.0 nm. (c) Schematic diagram of a single PbSe NWFET. (d) SEM image of a single PbSe NW FET with Cr/Auelectrodes.

Nano Letters Letter

dx.doi.org/10.1021/nl302161n | Nano Lett. 2012, 12, 4360−43654361

treated the NWs with hydrazine and ethane-dithiol, respec-tively, but these treatments are not as effective, presumably dueto the greater sensitivity of these chemicals to oxidation.Despite the narrow bandgap of PbSe (0.27 eV for bulk),ambipolar conduction is not seen, likely because the adsorptionof oxygen at the NW surface leads to p-doping. A previousreport on PbSe NWs also shows a unipolar hole conduction asa result of oxygen exposure.13 We will focus on the devicestreated by ammonium thiocyanate below, as they show muchhigher carrier mobilities.In D2, the conductance changes linearly with Vg when Vg is

between 0 and 0.8 V. At Vg < −0.1 V, the conductanceincreases much more slowly (Figure 2b). This conductancesaturation with Vg can be caused by the filling of the one-dimensional (1D) energy states of the ultrathin PbSe NWs, asthe density of states of 1D semiconductors is expected to bediscontinuous (Van Hove singularities). An intraband energyseparation of >100 meV is predicted22 between the two lowestenergy states in the valence band for PbSe NWs with diametersmaller than 10 nm. A similar saturation has also been observedin ultrathin InAs NW FETs.16 Also, interestingly, we observethat the current noise level increases in the linear region (0 < Vg< 0.8 V) (Figure 2b). This noise is not from theinstrumentation but seems to originate from the intrinsicfluctuation of the NW conductance. As evidence, after the NWsare depleted or saturated, the noise level is greatly reduced.This current fluctuation may be related to the charge trapping/detrapping at the surface of the NWs.As we fix the bias at 0.1 V and scan the gate voltage at a rate

of 100 mV/s, the current changes by more than 3 orders ofmagnitude (Figure 2b inset). The conductance of the depleted

NWs (Vg > 2 V) is beyond the instrumental sensitivity. Thefield-effect mobility is a function of the gate voltage sweep ratewith 100 mV/s yielding a field-effect mobility of 216 cm2/(V s),which is significantly lower than the value extracted from the I−V curves at various gate voltages (blue dots in Figure 2b inset).We attribute this gate sweep rate dependent field-effectmobility to a slow charging time, likely caused by the chargetraps at the interface between the NW and the SiO2 surface.

23

As the gate voltage changes quickly, interfacial traps are notcompletely filled, leading to a reduced current, and hence areduced apparent mobility.24 For this reason, the field-effectmobility taken from pseudostatic I−V curves at various Vg is abetter measure of the actual NW mobility. It should be notedthat even this value is a lower bound and can be a factor of 2lower than the true mobility.16 Better surface passivation suchas by coating the PbSe with a high bandgap shell may furtherimprove the FET switching speed.The elimination of uncertainties associated with ensemble

devices in single NW FETs and the improved surfacepassivation allow for more accurate extraction of carriermobilities in colloidal ultrathin PbSe NWs. Our best deviceexhibits a mobility about one-half of the Hall mobility (1000cm2/(V s)) in bulk PbSe.25 Considering the underestimation ofthe mobility from the field effect measurements,16 the actualmobility in PbSe NWs can be approaching the bulk value. Thisindicates that surface scattering in NWs with diameters of ∼10nm may not substantially limit their mobility. Further work isnecessary to better understand the diameter dependence of thecarrier mobility. Generally, the reduction in mobility isattributable to a combination of Coulomb scattering fromimpurities, surface scattering, and phonon scattering. Phononscattering plays an important role and will be discussed belowin the temperature dependence measurements. Given that wedo not intentionally dope the PbSe NWs with impurities, alongwith the high dielectric constant of PbSe, it is likely thatCoulomb scattering does not dominate. It has been reportedthat ultrathin NWs may have an even higher mobility than thebulk, as doping can be achieved by heterostructures26 or surfaceligands so that no carrier-scattering impurities are required inthe NW conduction channels.To investigate the optoelectronic properties, we first measure

the photocurrent response time as we turn on and off theillumination for both as-grown and treated devices. A fastresponse time of <100 μs (limited by the instrument, Figure3b) is obtained, as compared to many seconds for on-chipgrown wires.12 The fast photoresponse indicates a betterpassivated NW surface. Photocurrent is sensitive to thepolarization of the laser beam with an anisotropy ratio of0.47 (Figure 3c), which is defined by (I∥ − I⊥)/(I∥ + I⊥), whereI∥ (I⊥) is the photocurrent when incident electric field is parallel(perpendicular) to the orientation of the NW. A higheranisotropy ratio is expected given the large dielectric constantof PbSe. The discrepancy is not well understood but may becaused by the reduced dielectric constant in ultrathin NWsand/or scattering of the laser beam. Also, the photocurrent islinear with intensity in the range of this study (<200 W/cm2,Figure 3d). At this intensity, we estimate a photoinduced carrierconcentration of ∼1017 cm−3, smaller than the dark carrierconcentration.We then perform SPCM to analyze the spatially resolved

photocurrent in the PbSe NW FETs. Briefly, in SPCM,photocurrent is measured as a function of photoinjectionposition (Figure 3a). SPCM has been used to measure carrier

Figure 2. Electrical characteristics of PbSe NW FETs. (a) I−V curvesof a single as-grown PbSe NW device as a function of gate voltage,demonstrating p-type behavior with a mobility of 30 cm2/(V s). (b) I−V curves as a function of Vg for a device treated with ammoniumthiocyanate. (Inset) Current at Vb = −0.1 V extracted from the I−Vcurves in (a) (blue dots) and current as the Vg sweeps at a rate of 100mV/s (red curve). (c) Histogram of conductance (conductivity) for 35devices, both untreated (black) and treated (red). Nominalconductivities on the top axis are calculated by assuming all NWshave a diameter of 10 nm and a length of 4 μm. The blue squarerepresents the device in (a), the blue circle the device in (b).

Nano Letters Letter

dx.doi.org/10.1021/nl302161n | Nano Lett. 2012, 12, 4360−43654362

diffusion lengths and local band structures.27−31 Photocurrentspots are localized near the electrodes in our as-grown NWdevice D1 (Figure 4a). The photocurrent as a function ofdistance from the electrode (x) can be fitted by an exponentialfunction (I = I0 exp(−x/lD)), giving lD = 2.5 μm for D1. Incomparison, the treated device D2 shows a more elongatedphotocurrent spot from which we extract an increased lD = 4.5μm. The photocurrent profile is highly asymmetric in D2,dominated by the photocurrent peak near the right electrode,which is caused by the different contact barrier heights.

The characteristic length of photocurrent decay lD is notlikely due to a Schottky electric field near the contact, as theestimated depletion width is small (121 nm for D1 and 29 nmfor D2, using εr = 173 for bulk PbSe, and built-in potential Vbi =100 meV). Pseudo 1D semiconductors may have a moreextended depletion length, but even using a modifiedexpression of the depletion width for thin NWs,32 wNW = dexp(8ε0wbulk

2/εd2), results in depletion lengths similar to theNW diameter. This is because the large relative dielectricconstant renders the argument in the exponent small. Areduced effective dielectric constant (εr* ∼ 20) in ultrathinPbSe NWs also produces a depletion width much shorter thanthe measured lD. As such, the extension of the photocurrent isdue to the carrier diffusion. It should be mentioned that wecannot distinguish between exciton diffusion and free carrierdiffusion from the experimental results. A high exciton bindingenergy of ∼200 meV22 has been predicted in d = 4 nm PbSeNWs because of the reduced screening in the ultrathin NWs.Thus, lD is a measure of the minority carrier (electron) orexciton diffusion length. The treated device has a longerdiffusion length due to the increased mobility. Furthertheoretical and experimental work is necessary to distinguishthe type of diffusion.We can calculate the carrier recombination lifetime from the

measured diffusion length as τ = qlD2/μkBT, where the symbols

have their usual meanings. If we plug in the hole mobilitiesobtained from the gate measurements, we can estimate τ = 8.3nS for D1 and 22.5 nS for D2. As we expect that minoritycarriers (electrons) or excitons have a lower diffusioncoefficient, the actual lifetimes can be longer. As a reference,the single exciton recombination lifetime of colloidal PbSe QDshas been measured to be around 1 μs at 80 K.3 In PbSe NWs,the carrier lifetime has not been measured to our knowledgeand is expected to be a strong function of surface passivation.Indeed, the recombination lifetime increases after surfacetreatment, consistent with the idea that the density of surfacestates is reduced.The SPCM profile is a strong function of gate voltage in the

ammonium thiocyanate treated device D3 (Figure 5a). In thisdevice, the photocurrent is also highly asymmetric as in D2.The photocurrent direction at Vg = 0 V is consistent with anupward band bending toward the electrode. The band bendingindicates that the Fermi level of the PbSe NW is above the

Figure 3. (a) Schematic diagram of SPCM setup. (b) Temporalresponse of photocurrent to modulated laser light. Arrows indicate thebeginning and end of exposure. (c) Photocurrent as a function oflinear polarization angle. The photocurrent is maximized when parallelto the NW, indicated by the blue line, and minimized whenperpendicular, leading to an anisotropy ratio of 0.47. (d) Photocurrentas a function of laser intensity. A linear fit shows the near-linearrelation of photocurrent with laser intensity over ∼2 decades.

Figure 4. SPCM images and cross sections of photocurrent along theNW for D1 (a) and D2 (b) at Vb = Vg = 0 V and laser intensity of 130W/cm2 (λ = 532 nm). The reflectance images of the electrodes areoverlay with the SPCM image. The photocurrent curve is fit with anexponential (red curve) near the contact to extract the carrier diffusionlength.

Figure 5. Gate dependence of SPCM for D3 at laser intensity of 130W/cm2. (a) Photocurrent cross sections along the NW as a function ofgate voltage. The red dot indicates the photocurrent peak measuredfor (b). (b) Peak photocurrent as a function of gate voltage. Inset:Band bending diagrams for Vg = −2.0, −0.75, and 2.0 V, representingthe change of band bending at the contacts due to the applied gatefield. Note: the positive current direction is to the left.

Nano Letters Letter

dx.doi.org/10.1021/nl302161n | Nano Lett. 2012, 12, 4360−43654363

work function of Cr (4.5 eV). The photocurrent direction isreversed at Vg < −1 V, indicating a flip of the band bending. AtVg = −0.75 V, a flat band leads to a negligible photocurrent.From the peak photocurrent as a function of gate voltage(Figure 5b), we calculate the charge separation efficiency, ηcc =I/(qF), where F is the absorbed photon flux.33 ηcc increaseswith Vg and saturates at ∼100% at Vg = 2 V (Figure 5b). Byincreasing the gate voltage, the band bending at the contacts isenhanced, leading to a higher charge separation efficiency.However, the charge separation efficiency is saturated at unity,even as the band bending continues to increase with gatevoltage. A near unity ηcc implies that the charge separation issignificantly faster than the charge recombination at the NW/metal interface.Finally, we perform temperature dependent charge transport

measurements of the colloidal PbSe NW FET treated byammonium thiocyanate (device D3). At 79 K, the I−V curvesare much more asymmetric than at 300 K (Figure 6a). The

suppression of current at positive bias voltage is due to theasymmetric contact barrier heights, also observed in the SPCM(Figure 5a). At room temperature, the asymmetry is notobvious because the thermal activation is sufficient to overcomethe contact barrier height. In addition, the intrinsic currentfluctuation observed at room temperature is greatly reduced at79 K. We associate this with a “freezing out” of trap-sites, thatis, thermal energy is too small to lead to trapping/detrapping ofcharge carriers. From the I−V curves at different gate voltages,we extract the hole mobility as a function of temperature(Figure 6b). As temperature is decreased, the hole mobility firstincreases from 300 cm2/(V s) at 300 K to a maximum of 740cm2/(V s) at 139 K and then decreases to 410 cm2/(V s) at 79K. The mobility increase with decreasing temperature is due toreduced phonon scattering, consistent with bulk PbSe behavior.It has been reported that the bulk mobility increases by a factorof 10 upon cooling from 300 to 100 K.25 The reduction in the

measured mobility below 139 K is likely caused by the chargeblocking contact barrier, which is more pronounced at lowertemperature.At a fixed bias (Vb = −0.1 V), the current largely increases

with temperature (Figure 6c). From the slope, we can extractthe effective activation energy to be Ea = 90 meV at Vg = 0 Vand Ea = 60 meV at Vg = −5 V. This effective activation energyreflects both the change of carrier mobility and carrierconcentration as a function of temperature. The smaller Ea atnegative Vg is consistent with the valence band edge shiftingcloser to the Fermi level. Surprisingly, we find that the currentcurve at Vg = 0 shows a dip at ∼160 K (observed in all fivedevices measured). For Vg = −5 V, the dip at 160 K vanishes. Inorder to investigate the cause of the current dip, we measurecurrent at fixed Vb while sweeping the gate voltage (Figure 6d).We notice a sudden shift in threshold gate voltage (VT) from+1 V to −1 V at 160 K, which results in a reduction in currentat Vg = 0 V. At Vg = −5 V, there is not much change in current,consistent with the disappearance of the dip. While furtherinvestigations are necessary to clarify the mechanism of thisinteresting behavior, the dip is most likely related to surfacetraps, which can substantially influence the gate threshold.In summary, we have successfully fabricated FETs incorpo-

rating single ultrathin colloidal PbSe NWs treated byammonium thiocyanate and demonstrated a record-high holemobility of 490 cm2/(V s) at room temperature. The carrierdiffusion length is measured to be 4.5 μm and a close-to-unitycharge separation efficiency at the metal contact has beenachieved. Carrier mobility increases at lower temperature andpeaks at 740 cm2/(V s) at 139 K. A shift in gate thresholdvoltage has also been observed at 160 K. The demonstratedhigh carrier mobilities in ultrathin colloidal NWs encourage theoptoelectronic applications of these materials in devices such asphotodetectors and solar cells. The study identifies that surfacepassivation is key to improving performance of such devices.This work also provides a method for fabricating high quality,air-stable single colloidal NW FETs and opens up new excitingopportunities for understanding charge transport in lowdimensional colloidal semiconductors.

■ AUTHOR INFORMATIONCorresponding Author*E-mail: yu@physics.ucdavis.edu.NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThis work was supported by the U.C. Davis Startup Fund. Wethank P. A. Baeza and I. Arslan for assistance with TEM. Workat the Molecular Foundry was supported by the Office ofScience, Office of Basic Energy Sciences, of the U.S.Department of Energy under Contract No. DE-AC02-05CH11231.

■ REFERENCES(1) Murray, C. B.; Kagan, C. R.; Bawendi, M. G. Annu. Rev. Mater.Res. 2000, 30, 545.(2) Hines, M. A.; Scholes, G. D. Adv. Mater. 2003, 15, 1844.(3) Wehrenberg, B. L.; Wang, C. J.; Guyot-Sionnest, P. J. Phys. Chem.B 2002, 106, 10634.(4) Yu, D.; Wang, C. J.; Guyot-Sionnest, P. Science 2003, 300, 1277.(5) Yu, D.; Wang, C. J.; Wehrenberg, B. L.; Guyot-Sionnest, P. Phys.Rev. Lett. 2004, 92, 216802.

Figure 6. Temperature-dependent conductance of D3. (a) I−V curvesas a function of gate voltage at 79 K. (b) Mobility as a function oftemperature. (c) Semilog plot of current (at Vb = −0.1 V) vs 1/T forVg = 0 V (black) and Vg = 5 V (red). (d) Current vs gate voltage atvarious temperatures (Vg sweeps at 100 mV/s). Inset: voltagethreshold (VT) as a function of temperature (T).

Nano Letters Letter

dx.doi.org/10.1021/nl302161n | Nano Lett. 2012, 12, 4360−43654364

(6) Cho, K. S.; Talapin, D. V.; Gaschler, W.; Murray, C. B. J. Am.Chem. Soc. 2005, 127, 7140.(7) Yu, H.; Li, J. B.; Loomis, R. A.; Wang, L. W.; Buhro, W. E. Nat.Mater. 2003, 2, 517.(8) Boercker, J. E.; Clifton, E. M.; Tischler, J. G.; Foos, E. E.; Zega, T.J.; Twigg, M. E.; Stroud, R. M. J. Phys. Chem. Lett. 2011, 2, 527.(9) Giblin, J.; Vietmeyer, F.; McDonald, M. P.; Kuno, M. Nano Lett.2011, 11, 3307.(10) Baker, J. L.; Widmer-Cooper, A.; Toney, M. F.; Geissler, P. L.;Alivisatos, A. P. Nano Lett. 2010, 10, 195.(11) Khandelwal, A.; Jena, D.; Grebinski, J. W.; Hull, K. L.; Kuno, M.K. J. Electron. Mater. 2006, 35, 170.(12) Talapin, D. V.; Black, C. T.; Kagan, C. R.; Shevchenko, E. V.;Afzali, A.; Murray, C. B. J. Phys. Chem. C 2007, 111, 13244.(13) Kim, D. K.; Vemulkar, T. R.; Oh, S. J.; Koh, W. K.; Murray, C.B.; Kagan, C. R. ACS Nano 2011, 5, 3230.(14) Choi, J.-H.; Fafarman, A. T.; Oh, S. J.; Ko, D.-K.; Kim, D. K.;Diroll, B. T.; Muramoto, S.; Gillen, J. G.; Murray, C. B.; Kagan, C. R.Nano Lett. 2012, 12, 2631.(15) Liang, W. J.; Hochbaum, A. I.; Fardy, M.; Rabin, O.; Zhang, M.J.; Yang, P. D. Nano Lett. 2009, 9, 1689.(16) Ford, A. C.; Ho, J. C.; Chueh, Y. L.; Tseng, Y. C.; Fan, Z. Y.;Guo, J.; Bokor, J.; Javey, A. Nano Lett. 2009, 9, 360.(17) Talin, A. A.; Leonard, F.; Swartzentruber, B. S.; Wang, X.;Hersee, S. D. Phys. Rev. Lett. 2008, 101, 076802.(18) Cho, K. S.; Talapin, D. V.; Gaschler, W.; Murray, C. B. J. Am.Chem. Soc. 2005, 127, 7140.(19) Graham, R. M., C.; Triplett, M.; Yu, D. Proc. SPIE 2011, 8106.(20) Graham, R.; Miller, C.; Oh, E.; Yu, D. Nano Lett. 2011, 11, 717.(21) Fafarman, A. T.; Koh, W. K.; Diroll, B. T.; Kim, D. K.; Ko, D.K.; Oh, S. J.; Ye, X. C.; Doan-Nguyen, V.; Crump, M. R.; Reifsnyder,D. C.; Murray, C. B.; Kagan, C. R. J. Am. Chem. Soc. 2011, 133, 15753.(22) Bartnik, A. C.; Efros, A. L.; Koh, W. K.; Murray, C. B.; Wise, F.W. Phys. Rev. B 2010, 82, 195313.(23) Sun, S. C.; Plummer, J. D. IEEE J. Solid-State Circuit 1980, 15,562.(24) Dayeh, S. A.; Soci, C.; Yu, P. K. L.; Yu, E. T.; Wang, D. L. Appl.Phys. Lett. 2007, 90, 162112.(25) Dalven, R. Infrared Phys. 1969, 9, 141.(26) Xiang, J.; Lu, W.; Hu, Y. J.; Wu, Y.; Yan, H.; Lieber, C. M.Nature 2006, 441, 489.(27) Ahn, Y. H.; Tsen, A. W.; Kim, B.; Park, Y. W.; Park, J. Nano Lett.2007, 7, 3320.(28) Kelzenberg, M. D.; Turner-Evans, D. B.; Kayes, B. M.; Filler, M.A.; Putnam, M. C.; Lewis, N. S.; Atwater, H. A. Nano Lett. 2008, 8,710.(29) Miller, C.; Triplett, M.; Lammatao, J.; Suh, J.; Fu, D. Y.; Wu, J.Q.; Yu, D. Phys. Rev. B 2012, 85, 085111.(30) Strasfeld, D. B.; Dorn, A.; Wanger, D. D.; Bawendi, M. G. NanoLett. 2012, 12, 569.(31) Fu, D. Y.; Zou, J. J.; Wang, K.; Zhang, R.; Yu, D.; Wu, J. Q.Nano Lett. 2011, 11, 3809.(32) Leonard, F.; Talin, A. A.; Swartzentruber, B. S.; Picraux, S. T.Phys. Rev. Lett. 2009, 102, 106805.(33) The flux is calculated by integrating the Gaussian shape of thelaser beam over the NW, giving F = (π/2)1/2 (Pαr2)/(σhν), where P isthe total laser power, α = 105 cm−1 is the absoprtion coefficient ofPbSe at 532 nm, r is the radius of the NW, hν is the photon energy,and σ is the standard deviation.

Nano Letters Letter

dx.doi.org/10.1021/nl302161n | Nano Lett. 2012, 12, 4360−43654365