A 30 μm Coaxial Nanowire Photoconductor Enabling Orthogonal

Jul 30, 2015 - Department of Physics, Xiamen University, Xiamen 361005, China. § Department ... Department of Chemical Engineering and Materials Scie...
2 downloads 11 Views 8MB Size
Letter pubs.acs.org/NanoLett

A 30 μm Coaxial Nanowire Photoconductor Enabling Orthogonal Carrier Collection Qiang Xu,†,‡ Shaopeng Qiao,§ Rajen Dutta,§ Mya Le Thai,† Xiaowei Li,∥ Crystin J. Eggers,† Girija Thesma Chandran,† Zhengyun Wu,‡ and Reginald M. Penner*,†,∥ †

Department Department § Department ∥ Department ‡

of of of of

Chemistry, University of California, Irvine, California 92697, United States Physics, Xiamen University, Xiamen 361005, China Physics and Astronomy, University of California, Irvine, California 92697, United States Chemical Engineering and Materials Sciences, University of California, Irvine, California 92697, United States

ABSTRACT: We describe the preparation and properties of a coaxial, three-layer, gold-CdSe-gold nanowire 30 μm in length that functions as a monolithic photodetector. The gold (Au) electrode core of this sandwich structure is prepared using the lithographically patterned nanowire electrodeposition (LPNE) method on a glass surface. A CdSe shell of defined thickness, dCdSe, from 200 to 280 nm is then electrodeposited on this Au nanowire. Finally, a conformal gold layer is electrodeposited on top of the CdSe shell. The two concentric gold electrodes within this architecture measure the photoconductivity of the ultrathin CdSe absorbing layer in the direction orthogonal to the nanowire axis. This architecture enables accelerated response/recovery of the nanowire to light while simultaneously maximizing the photoconductive gain without relinquishing any of the photoresponsive area of a ”bare” nanowire. Characterization by scanning electron microscopy (SEM) of focused ion beam (FIB) cross sections together with electron dispersive X-ray spectroscopy (EDS) reveal the distinct core−multishell nanostructure, layer thicknesses, and layer compositions. The positiondependent photoresponse along the axis of the nanowire, probed using a laser spot, shows that the Au nanoshell significantly enhances the photocurrent. The performance of Au−CdSe−Au core−multishell nanowire photodetectors depend sensitively on the thickness of CdSe nanoshell over the range of from 200 nm < dCdSe < 280 nm. The highest performance was obtained for the dCdSe = 250 nm this device, which showed a photoconductive gain of 2172, a responsivity of 209 A·W−1, a response time of 17 μs, and a recovery time of 96 μs. KEYWORDS: Multishell, cadmium selenide, gain, responsivity, photoconductor, photodetector, electrodeposition

S

ince the demonstration by Lieber and co-workers of photodetection using an indium phosphide nanowire in 2001,1,2 the development of sensitive and fast nanowire based photodetectors has been a goal of intense interest.3−5 Two types of nanowire photodetectors have been studied: photodiodes6−12 and photoconductors,13−17 with the majority of the work focusing on the latter. The performance of photoconductive detector can be characterized by two parameters: the photoconductive gain, G, and the bandwidth, f 3dB. G, the number of electrons collected in the external circuit for each absorbed photon, is given by the ratio between the free carrier lifetime, τ, and the transit time for carriers traveling between the two electrical contacts, τtr:18,19 τ G= τtr

τtr =

L2 (μn + μp )V

(2)

where μn and μp are the mobilities of electrons and holes, respectively. So G is inversely proportional to L2: G=τ

(μn + μp )V L2

(3)

f 3dB, however, is inversely proportional to τ:18,19

f3dB =

1 2πτ

(4)

The gain-bandwidth product, G f 3dB, is therefore independent of τ and equal to19

(1) Received: May 18, 2015 Revised: July 17, 2015

The value of τtr depends upon the applied voltage, V, and the channel length, L, according to19 © XXXX American Chemical Society

A

DOI: 10.1021/acs.nanolett.5b01941 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Table 1. Summary of Photoconductivity Metrics for CdSe Nanostructure-Based Photodetectors and Coaxial, Three-Layer, Gold−CdSe−Gold Nanowire Photodetectors devicea CdSe NRib CdSe/ZnS QD CdSe NR CdSe NW CdSe NW CdSe NW CdSe in NG concentric Au−CdSe−Au core-multishell nanowires: wCdSe = 200 (±10) nm 226 (±10) nm 250 (±10) nm 280 (±10) nm

Rb (A·W−1)

Sc

5.88 × 10−3 1.5 × 104

Gd

1.30 × 10−2 1.70 9.1−31

7 1.5−4 270 20 10−100 500

3 × 104 10−1000 4.9 0.043 20−45

21 ± 3 43 ± 4 210 ± 10 57 ± 4

71 ± 5 110 ± 16 500 ± 20 130 ± 20

140 ± 20 360 ± 28 2200 ± 100 400 ± 30

f 3dBe (Hz)

area (est)f (nm2)

11 125 × 103 1.7 × 103

175 × 103

7 × 10 3 × 103 1.4 × 107 3 × 105 9.5 × 105 9.5 × 105 1 × 105

39 21 38 40 22 23 13

720 180 2400 200

1.2 × 107 1.35 × 107 1.5 × 107 1.7 × 107

this this this this

44 × 103

6

ref

work work work work

a Abbreviations: QD = quantum dot, NR = nanorod, NRib = nanoribbon, NW = nanowire, NG = nanogap. bR = responsivity = Iphoto/Poptical. cS = photosensitivity = (Iphoto − Idark)/Idark. dG = photoconductive gain at 532 nm. ef 3dB = 1/(2πτ) where τ = τresp + τrec. fEstimated area = total photoactive cross-section of the device, e.g., diameter × length for a cylinder.

Gf3dB =

second conformal gold layer is electrodeposited on top of the CdSe shell (step vi). An electrical contact is established to the outermost gold shell using a final photolithography step (step vii). The evaporated nickel layer, which is partially removed in step iv, provides a contact to the gold nanowire at the core of the Au−CdSe−Au nanowire. As already indicated, just the gold nanowire at the core of this structure is prepared by LPNE (Figure 2e). The CdSe second layer is synthesized by operating this gold nanowire as a working electrode and plating CdSe potentiostatically from aqueous 0.30 M CdSO4, 0.70 mM SeO2, and 0.25 M H2SO4 at pH 1−2. Shown in Figure 2a is the cyclic voltammogram acquired at a 200 nm (w) × 80 nm (h) × 30 μm (l) gold nanowire. The reduction peak at −0.75 V is associated with the deposition of both cadmium metal and CdSe.28 The oxidation wave at −0.65 V is associated with the removal of cadmium metal from this CdSe/Cdo mixed layer as Cd2+.28 We found that the slow, potentiostatic deposition of CdSe at −0.60 V (Figure 2c) produced a smooth, dense CdSe layer (Figure 2f), that was also stoichiometric based upon EDX analysis. A deposition duration of 70 s produced a CdSe layer of thickness ∼250 nm. In the last step of this process (Figure 1, step vi), the gold top layer was electrodeposited on the Au−CdSe core− shell nanowire from a commercial Au plating solution (GreenEarth Solutions, Inc.) at −0.90 V. The cyclic voltammogram of this solution at the Au−CdSe core−shell nanowire (Figure 2b) shows a peak at ∼−1.0 V that is associated with irreversible gold electrodeposition. A conformal gold layer was electrodeposited atop the CdSe shell potentiostatically at −0.90 V. Deposition for 70 s produced a gold layer of 20−40 nm thickness, estimated from SEM images (Figure 2g). The increasing current seen in Figure 2c,d is associated with the increasing diameter of the nanowire as CdSe (c) and gold (d) are deposited under conditions of kinetic control. A low magnification SEM image of a typical concentric Au− CdSe−Au nanowire device (Figure 3a) shows the gold nanowire core extends 40 μm between two rectangular nickel pads. A coating of CdSe covers the gold nanowire over virtually this entire length, but the topmost gold layer is confined to the center 30 μm segment of the nanowire. An SEM image of a cross-section of the nanowire (Figure 3b) shows the central

(μn + μp )V 2

2πL

(5)

Equation 5 provides two options for increasing the gainbandwidth product of a photoconductive detector at constant V: First, the presence of defects in the light absorbing layer can be reduced to increase the carrier mobilities, μn and μp. Second, L can be reduced. The second strategy has been exploited by assembling ZnO nanocrystals20 and CdSe/ZnS core−shell QDs21 into 50−60 nm gold nanogaps, and an enhancement of G f 3dB has been demonstrated for both systems. We have also verified the dramatic influence of L on Gf 3dB, predicted by eqs 1−4, for nanocrystalline CdSe nanowires22−24 and Au−CdSe− Au nanogap13,24 structures prepared by electrodeposition. However, while reducing L can increase G f 3dB, the problem is that in all of the photoconductive detectors so far described the optically responsive area of the nanodetector is reduced in direct proportion to the reduction in L. This results in photoresponsive areas of 107 nm2, as demonstrated below. Coaxial, three-layer Au−CdSe−Au nanowires were synthesized entirely by electrodeposition using the process flow illustrated in Figure 1b. Briefly, a gold nanowire 50 μm in length is first prepared using the lithographically patterned nanowire electrodeposition (LPNE) method25−27 (steps i−iv). Then a conformal layer of CdSe, having a thickness varying from 200 to 280 nm, is electrodeposited on top of this gold nanowire (step v), as previously described.22,23,28 Finally, a B

DOI: 10.1021/acs.nanolett.5b01941 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 1. Design and process flow for the fabrication of a Au−CdSe− Au single nanowire photodetector. a) Rendering of the three-layer, Au−CdSe−Au single nanowire photodetector prepared in this study. The inset shows the cross-sectional diagram with Au core (yellow), CdSe shell (orange), and Au shell (yellow). (b) Seven-step process flow for the fabrication of the device shown in (a). This process begins with a glass slide coated with 80 nm of nickel using thermal evaporation; step (i) a layer of positive photoresist (PR) is spin-coated onto the nickel-coated glass slide; step (ii) the PR layer is patterned with a contact mask; after development, exposed Ni is removed by chemical etching; step (iii) a single Au nanowire is electrodeposited under the PR trench using the nickel edge as an electrode; step (iv) the patterned PR is removed, and a new PR layer is spin-coated to create a 40 μm wide window; step (v) a CdSe shell is then electrodeposited onto the Au nanowire; step (vi) the PR layer is again removed and a third PR layer is coated on the clean surface to produce a narrower, 30 μm width window, centered on the Au−CdSe core− shell nanowire. The third gold layer is then electrodeposited to complete the formation of the Au−CdSe−Au structure; step (vii) a 10 μm width Au contact is prepared using photolithography and metal lift-off method.

Figure 2. Electrodeposition of layers 2 (CdSe) and 3 (Au): (a) Cyclic voltammetry at 50 mV/s for a gold nanowire electrode in Cd2+ and SeO32− plating solution and (b) Cyclic voltammetry at 50 mV/s for a CdSe-coated Au nanowire in commercial gold plating solution. (c,d) Current vs time for the potentiostatic electrodeposition of (c) CdSe onto a gold nanowire at −0.6 V vs SCE, according to Cd2+ + H2SeO3 + 4H+ + 6e− → CdSe(s) + 3H2O. (d) Au onto a CdSe-coated Au nanowire at a −0.9 V vs SCE. (e−g) SEMs of wire sections before CdSe electrodeposition (e), after CdSe electrodeposition (f), and after electrodeposition of the Au top layer (g).

the nanowire in these images can be verified using EDX elemental maps (Figure 3c−h). For example, an interface between the central gold nanowire and the CdSe absorbing layer is seen in Figure 3c−e, while an interface between the CdSe layer and the outermost gold layer is shown in Figure 3f− h. Since the outermost gold layer is just 20−40 nm in thickness, excitation and X-ray emission from the CdSe is still observed through this layer (Figure 3g) whereas the 250 nm thickness of the CdSe layer strongly attenuates gold EDX signal from the central gold electrode (Figure 3e). The position-dependent photocurrent response for the concentric Au−CdSe−Au nanowire can be probed using a laser spot (Figure 4). The 532 nm laser of a Renishaw Raman Microscope was used for this purpose. This laser produced a spot size of 1 μm and a power density at the nanowire of ∼2200 mW/cm2. Seven spots, shown in the optical micrograph of Figure 4b, were illuminated, and Iphoto and Idark were measured at each spot. The quantity Iphoto/Idark is plotted in Figure 4a. Values of Iphoto/Idark between 900 and 1000 are seen at spots 5, 6, 9, and 10 where the three-layer, Au−CdSe−Au

gold nanowire, the surrounding CdSe absorbing layer, and the outermost gold layer. This SEM image also shows that the central gold nanowire is lifted off of the glass surface and embedded within the CdSe coating as this coating is electrodeposited, suggesting that the adhesion of the gold nanowire to the glass is weak. This image also shows that the gold overlayer is thickest at the top of the nanowire and thinner on its sides. This nonuniform distribution of the gold overlayer may be caused by the strong Ohmic drop imposed by the CdSe layer between the central gold nanowire and the surface of the CdSe. If this Ohmic drop is present, then the applied potential at the CdSe surface could be more negative in the direction of the counter electrode (the top) and somewhat more positive along its sides, reducing the growth rate of gold. The composition of the gold and CdSe regions of C

DOI: 10.1021/acs.nanolett.5b01941 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 3. SEM and EDX characterization of Au−CdSe−Au core−multishell nanowire. (a) Low magnification SEM image of a single Au−CdSe−Au core multishell nanowire with gold and nickel contacts. (b) SEM image of focused ion beam (FIB) cross-section Au−CdSe−Au multishell nanowire. A clear interface between the core (false color orange) and the shell region (yellow) is observed. A platinum embedding layer was overlaid to enable a high quality FIB cut. (c) High magnification SEM image of the Au nanowire−CdSe shell interface region. Gold and CdSe regions are labeled using false color. (d,e) EDX elemental maps for Cd (d) and Au (e) of the same region shown in (c). (f) High magnification SEM image of the CdSe shell−Au overlayer interface region. Gold and CdSe regions are labeled using false color. (g,h) EDX elemental maps for Cd (g) and Au (h) of the same region shown in (f). In (d) and (g), selenium EDX maps reproduce those of cadmium.

outermost gold layer to 70−90 nm. It is apparent (Figure 4a) that an increase in the gold thickness at spots 7 and 8 depresses Iphoto/Idark by 10−15% relative to spots 5, 6, 9, and 10 where a thinner gold top layer is present. This loss in photocurrent is attributed to a combination of scattering and absorption of the thicker Au film. At spots 1−4 and 11−14, the outermost gold layer is completely missing because these two segments of the nanowire, located adjacent to the nickel contacts, were masked prior to the electrodeposition of the outermost gold layer (Figure 1b, step vi). So at spots 1−4 and 11−14, photogenerated carriers must be collected between the central gold nanowire and the outermost gold layer that is ∼10 μm distant from the point of illumination. At both spots 1−4 and 11−14, we observe values of Iphoto/Idark below 50, showing that the outermost gold layer dramatically increases the efficiency of carrier collection. Photocurrent action spectra for concentric Au−CdSe−Au nanowire devices having four CdSe shell thicknesses, wCdSe, are shown in Figure 5a. All four devices show a sharp spectral threshold for photocurrent at λex = 820 nm that is red-shifted by ∼100 nm from the bulk bandgap of CdSe (1.75 eV or λex = 708 nm). The photocurrent threshold is also shifted by this same 100 nm interval in other photodetectors we have recently described,13,22,23 also based upon electrodeposited CdSe. If the outermost gold layer is omitted (but the central gold current collector is retained) the responsivity is reduced by an order of magnitude relative to the concentric Au−CdSe−Au nanowire having the same CdSe thickness, and the spectral response is dramatically altered, showing a gradual increase in the photocurrent from 750 to 600 nm (Figure 5b). Prior work has examined the influence of gold nanoparticles and nanostructures immobilized on the surface of a semiconductor absorber.17,29−34 Two effects of these metal structures are observed: First, the absorption edge is red-shifted and the magnitude of this shift increases with the diameter of the gold

Figure 4. Position-dependent photoresponse, Iphoto/Idark, of a coaxial, three-layer, Au−CdSe−Au nanowire. (a) Iphoto/Idark at 14 locations along the axis of the nanowire device shown in (b). The light source is a 532 nm laser spot with diameter 1 μm and a power at the nanowire of ∼2200 mW/cm2. At the four data points at far left, and the four at far right, the outermost gold layer is absent. At positions 7 and 8, the outermost gold layer has an augmented thickness due to the evaporation of the vertical gold electrical contact. (b) Optical micrograph of the coaxial, three-layer, Au−CdSe−Au nanowire probed in this experiment.

nanowire is probed. The thickness of the outermost gold layer is estimated to be 20−40 nm at these points. Spots 7 and 8, near the center of the 40 μm nanowire length, are located within the region where the vertical gold stripe contact has been evaporated, increasing the overall thickness of the D

DOI: 10.1021/acs.nanolett.5b01941 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 5. Photocurrent action spectra. (a) Photocurrent responsivity versus wavelength, λex, for three concentric Au−CdSe−Au nanowire devices: wCdSe = 200, 226, 250, and 280 nm. (b) Photocurrent responsivity versus wavelength comparison for two devices: wCdSe = 250 nm with the gold overlayer (red trace) and without the gold overlayer (gray trace). Vertical scale has been magnified by ×8 factor. All spectra were obtained using a xenon arc lamp-monochromator source.

nanoparticle absorbers. Yu et al.,34 for example, reported a shift of the absorption edge by 100−150 nm for 100 nm diameter gold nanoparticles dispersed on Si(100). Second, the absorption coefficient is increased by a factor of 2−20% across all wavelengths. We observe both of these effects in the data of Figure 5b. The strong, 80−90% attenuation of the current in the absence of the gold overlayer can be attributed, at least in part, to the longer path length for current in these devices. Any contribution of plasmonic effects on the current amplitude are obscured by the increased resistance of these devices, but the red-shifted spectral onset for photocurrent produced by the gold overlayer should not be strongly influenced by these electrical considerations, and it is likely that its plasmonic coupling to the CdSe layer contributes to this shift. A concentric Au−CdSe−Au nanowire photodetector enables the rapid and highly sensitive detection of light at wavelengths below 800 nm, across essentially the entire visible spectrum. The photoconductive gain, G, the photocurrent rise time, τresp, and the recovery time for detecting light, τrec, were measured at 532 nm (2.33 eV, 100 mW/cm2) using light chopped at 3 kHz (Figure 6). Light of this energy generates a strong, reproducible photocurrent response (Figure 6a) that is strongly dependent upon the thickness of the CdSe layer, wCdSe (Figure 6b−e). Devices with wCdSe values of 200, 226, 250, and 280 nm were investigated in this study. The response time of this current, τresp, is defined as the time interval between 10% and 90% of maximum photocurrent, I0, whereas τrec is the time interval from 90%I0 to 10%I0 in the dark. These were measured by fitting the rising (τresp) or trailing (τrec) edge of the normalized photocurrent transient to the equations:18,35,36 ⎡ ⎛ − t ⎞⎤ ⎟⎟⎥ + Idark I(t ) = I0⎢1 − exp⎜⎜ ⎢⎣ ⎝ τresp ⎠⎥⎦

(6)

⎛ −t ⎞ I(t ) = I0exp⎜ ⎟ + Idark ⎝ τrec ⎠

(7)

Figure 6. Photodetector metrics and performance. (a) Current versus time for the exposure of a wCdSe = 250 nm Au−CdSe−Au nanowire to 532 nm light (100 mW/cm2) modulated at 3 kHz. (b) Same experiment conducted for wCdSe = 280, 250, 226, and 280 nm. From these data, the response and recovery times, τrest and τrec, can be determined for each device. (c) Plot of τrest and τrec for concentric Au− CdSe−Au photodetectors as a function of wCdSe. (d) Current versus voltage plots acquired using unmodulated 532 nm light (350 mW/ cm2). Shown for comparison is the dark current for these same devices. (e) Photoconductive gain versus a function of wCdSe for the same devices. In (c) and (e), vertical error bars corresponding to ±1σ are smaller than the data points in some cases. The statistics represent N = 4 devices (wCdSe = 200 nm), 3 devices (226 nm), 5 devices (250 nm), and 4 devices (280 nm).

where Idark is the dark current at a particular applied bias and I0 is the photocurrent at this bias for specified illumination conditions. The values for τresp are in the 18−66 μs range, while τrec is 100−150 μs (Figure 6c). These translate into f 3dB bandwidths ranging from 200 Hz to 2.4 kHz (Table 1) with the highest f 3dB bandwidth produced by wCdSe = 250 nm devices. These devices are as fast as single nanowire, nanorod,

or nanogap structures based upon CdSe (Table 1), but the responsive area is a factor of 100 larger, on average. This is a E

DOI: 10.1021/acs.nanolett.5b01941 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters Notes

consequence of the relatively thick CdSe layer we have employed here, and the 30 μm length of the nanowire, made possible by the low-resistance, concentric Au−CdSe−Au architecture. In principle, there is no reason that the length of this nanowire could not be extended into the millimeter regime. An even stronger influence of the CdSe shell thickness is seen in the photocurrent magnitude (Figure 6d) and the resulting photoconductive gain, G (Figure 6e). The largest photocurrents are observed for wCdSe = 250 nm, corresponding to G = 2200 ± 100, followed by wCdSe = 280 nm (G = 400 ± 30), wCdSe = 226 nm (G = 290 ± 30), and wCdSe = 200 nm (G = 140 ± 20) (Table 1). Responsivity (R = 210) and photosensitivity (S = 500) are also optimized for wCdSe = 250 nm. These sensitivity metrics are competitive with, or superior to, those obtained in prior work involving a variety of different CdSe nanostructure (Table 1). Equation 3 can not account for this nonmonotonic dependence of G on wCdSe and optimum performance seen for wCdSe, 250 nm, and we continue to work to understand the physics responsible for this. A similar absorber thickness effect has been reported for GaN nanowire photoconductors.37 In that case, it is attributed to the influence of surface band bending on the kinetics of electron−hole recombination, but this model requires that τrec is directly correlated with G; that is, larger τrec correlated with higher G. We observed exactly the opposite in the present system (Figure 6c,e). The performance of the wCdSe = 250 nm devices described here does not exceed that of all prior efforts involving CdSe nanostructure emitters. The best CdSe nanoribbon devices described by Dai and workers,38 in particular, show an exceptionally high gain of 30,000 (Table 1), but even in this case, we achieve higher photosensitivity (500 vs 1.5−4.0) and comparable higher bandwidth (f 3dB = 2400 vs 1700) for the wCdSe = 250 nm devices. The complexity of the fabrication process for the CdSe nanoribbon devices is comparable to that described in Figure 1 when the application of electrical contacts using FIB or EBL is considered. We have described a simple process for producing ultralong (30 μm) coaxial, three-layer, gold−CdSe−gold nanowires that function as monolithic photodetectors. Our process relies on the ability to pattern a gold nanowire onto a dielectric using LPNE, followed by two additional electrodeposition steps: one to plate the CdSe absorber layer and a second to form the gold top electrode. At the end of this process, the gold nanowire becomes the center electrode in the concentric Au−CdSe−Au sandwich. In this device photoexcited carriers are collected in the direction orthogonal to the nanowire axis and across a CdSe shell layer that is 200−280 nm in total thickness. The optically active area of this photodetector is decoupled from the electrical path length through the absorber, enabling larger areas, corresponding to longer nanowires, to be accessed while maintaining the absorber thickness at a fixed value that maximizes both G and f 3dB. Here, a gain value of 2172, a responsivity of 209 A·W−1, a response time of 17 μs, and a recovery time of 96 μs have been achieved for a coaxial, threelayer, gold−CdSe−gold nanowire 30 μm in length with a CdSe thickness of ∼250 nm.



The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors gratefully acknowledge the financial support of this work by the National Science Foundation, Division of Materials Research through contract DMR-1206867. FE-SEM data were acquired using instrumentation of the LEXI facility (lexi.eng.uci.edu/) at UCI. Q.X. acknowledges the support of the Chinese Scholars Council and the Natural Science Foundation of China (Grants No. 61176049 and 61307047). The authors also thank Dr. Dima Fishman of the UCI Chemistry Laser Spectroscopy Facility for his expert assistance.



REFERENCES

(1) Duan, X.; Huang, Y.; Cui, Y.; Wang, J.; Lieber, C. M. Nature 2001, 409, 66−69. (2) Wang, J.; Gudiksen, M. S.; Duan, X.; Cui, Y.; Lieber, C. M. Science 2001, 293, 1455−1457. (3) Soci, C.; Zhang, A.; Bao, X.-Y.; Kim, H.; Lo, Y.; Wang, D. J. Nanosci. Nanotechnol. 2010, 10, 1430−1449. (4) Sang, L.; Liao, M.; Sumiya, M. Sensors 2013, 13, 10482−10518. (5) Zhai, T.; Fang, X.; Liao, M.; Xu, X.; Zeng, H.; Yoshio, B.; Golberg, D. Sensors 2009, 9, 6504−6529. (6) Wang, S.-Y.; Tan, M. R. Nanowire-based photodiode. 2008; US Patent Appl. 12/990,939. (7) Novotny, C. J.; Yu, E. T.; Yu, P. K. Nano Lett. 2008, 8, 775−779. (8) Luo, L.; Zhang, Y.; Mao, S. S.; Lin, L. Sens. Actuators, A 2006, 127, 201−206. (9) Matei, E.; Ion, L.; Antohe, S.; Neumann, R.; Enculescu, I. Nanotechnology 2010, 21, 105202. (10) Hayden, O.; Agarwal, R.; Lieber, C. M. Nat. Mater. 2006, 5, 352−356. (11) Qian, F.; Gradecak, S.; Li, Y.; Wen, C.-Y.; Lieber, C. M. Nano Lett. 2005, 5, 2287−2291. (12) He, J. H.; Te Ho, S.; Wu, T. B.; Chen, L. J.; Wang, Z. L. Chem. Phys. Lett. 2007, 435, 119−122. (13) Xing, W.; Kung, S.-C.; van der Veer, W. E.; Yan, W.; Ayvazian, T.; Kim, J. Y.; Penner, R. M. ACS Nano 2012, 6, 5627−5634. (14) Gao, T.; Li, Q.; Wang, T. Appl. Phys. Lett. 2005, 86, 173105. (15) Kim, C.-J.; Lee, H.-S.; Cho, Y.-J.; Kang, K.; Jo, M.-H. Nano Lett. 2010, 10, 2043−2048. (16) Soci, C.; Zhang, A.; Xiang, B.; Dayeh, S. A.; Aplin, D. P. R.; Park, J.; Bao, X. Y.; Lo, Y. H.; Wang, D. Nano Lett. 2007, 7, 1003− 1009. (17) Xu, Q.; Cheng, Q.; Zhong, J.; Cai, W.; Zhang, Z.; Wu, Z.; Zhang, F. Nanotechnology 2014, 25, 055501. (18) Rosencher, E.; Vinter, B. Optoelectronics; Cambridge University Press: Cambridge, U.K., 2002. (19) Beneking, H. IEEE Trans. Electron Devices 1982, 29, 1420−1431. (20) Seo, Y.-K.; Kumar, S.; Kim, G.-H. Phys. E 2010, 42, 1163−1166. (21) Hegg, M. C.; Horning, M. P.; Baehr-Jones, T.; Hochberg, M.; Lin, L. Y. Appl. Phys. Lett. 2010, 96, 101118. (22) Kung, S.-C.; Xing, W.; van der Veer, W. E.; Yang, F.; Donavan, K. C.; Cheng, M.; Hemminger, J. C.; Penner, R. M. ACS Nano 2011, 5, 7627−7639. (23) Kung, S.-C.; van der Veer, W. E.; Yang, F.; Donavan, K. C.; Penner, R. M. Nano Lett. 2010, 10, 1481−1485. (24) Penner, R. M. J. Phys. Chem. C 2014, 118, 17179−17192. (25) Xiang, C.; Yang, Y.; Penner, R. M. Chem. Commun. 2009, 859− 873. (26) Menke, E.; Thompson, M.; Xiang, C.; Yang, L.; Penner, R. Nat. Mater. 2006, 5, 914−919. (27) Xiang, C.; Kung, S.-C.; Taggart, D. K.; Yang, F.; Thompson, M. A.; Guell, A. G.; Yang, Y.; Penner, R. M. ACS Nano 2008, 2, 1939− 1949.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. F

DOI: 10.1021/acs.nanolett.5b01941 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters (28) Li, Q.; Brown, M. A.; Hemminger, J. C.; Penner, R. M. Chem. Mater. 2006, 18, 3432−3441. (29) Akimov, Y. A.; Ostrikov, K.; Li, E. Plasmonics 2009, 4, 107−113. (30) Atwater, H. A.; Polman, A. Nat. Mater. 2010, 9, 205−213. (31) Lee, J.-S.; Shevchenko, E. V.; Talapin, D. V. J. Am. Chem. Soc. 2008, 130, 9673−9675. (32) Ferry, V. E.; Sweatlock, L. A.; Pacifici, D.; Atwater, H. A. Nano Lett. 2008, 8, 4391−4397. (33) Lim, S.; Mar, W.; Matheu, P.; Derkacs, D.; Yu, E. J. Appl. Phys. 2007, 101, 104309. (34) Schaadt, D.; Feng, B.; Yu, E. Appl. Phys. Lett. 2005, 86, 063106. (35) Petritz, R. L. Phys. Rev. 1956, 104, 1508. (36) Sze, S. M.; Ng, K. K. Physics of Semiconductor Devices, 3rd ed.; Wiley-Interscience: Hoboken, NJ, 2007. (37) Chen, H.-Y.; Chen, R.-S.; Rajan, N. K.; Chang, F.-C.; Chen, L.C.; Chen, K.-H.; Yang, Y.-J.; Reed, M. A. Phys. Rev. B: Condens. Matter Mater. Phys. 2011, 84, 205443. (38) Wu, P.; Dai, Y.; Sun, T.; Ye, Y.; Meng, H.; Fang, X.; Yu, B.; Dai, L. ACS Appl. Mater. Interfaces 2011, 3, 1859−1864. (39) Jiang, Y.; Zhang, W.; Jie, J.; Meng, X.; Fan, X.; Lee, S.-T. Adv. Funct. Mater. 2007, 17, 1795−1800. (40) Fan, Z.; Ho, J. C.; Jacobson, Z. A.; Razavi, H.; Javey, A. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 11066−11070.

G

DOI: 10.1021/acs.nanolett.5b01941 Nano Lett. XXXX, XXX, XXX−XXX