NANO LETTERS
Scanning Tunneling Microscope Excited Cathodoluminescence from ZnS Nanowires
2006 Vol. 6, No. 5 926-929
Dorothy Duo Duo Ma and Shuit-Tong Lee* Center of Super-Diamond and AdVanced Films (COSDAF) and Department of Physics and Materials Science, City UniVersity of Hong Kong, Hong Kong SAR, (China)
Peter Mueller and Santos F. Alvarado* IBM Research GmbH, Zurich Research Laboratory, Sa¨umerstr. 4, 8803 Ru¨schlikon, Switzerland Received December 20, 2005; Revised Manuscript Received February 15, 2006
ABSTRACT The electronic excitation spectrum of ZnS nanowires (NWs) is characterized by cathodoluminescence (CL) excited using the tip of a scanning tunneling microscope as a highly localized and bright source of low-energy (150−350 eV) electrons with impinging power densities in the range of up to 60 kW/cm2 or more. The CL spectra reveal significant differences when compared with the photoluminescence spectra of ZnS NWs. The differences can be associated with the properties of the surface region of the NWs, which are preferentially emphasized in the CL spectra owing to the small probing depth of low-energy electrons.
Semiconductor nanowires (NWs) are emerging as fundamental building blocks for the fabrication of next-generation nanoscale optoelectronic devices. Much effort has been devoted to their use in applications such as optical switches,1 nanophotonic devices,2 and lasers.3-10 To achieve progress in this field requires a detailed understanding of the behavior of photons as well as of charge-carrier transport and injection phenomena under the strong physical confinement conditions given in these systems. Here we report on the use of a scanning tunneling microscope (STM) to excite cathodoluminescence (CL) in ZnS NWs, arising from the recombination of electron-hole pairs created by impinging electrons of moderate energy field-emitted from the tip of the STM. A comparison between the spectra of the STM-excited emission and photoluminescence reveals differences that could be useful to characterize the properties of the surface region of NWs and, in addition, yield insight into the injection of charge carriers for electrical laser excitation. The ZnS nanowires used in the experiments were synthesized by hydrogen-assisted thermal evaporation of high-purity ZnS powder onto a Si substrate coated with Au catalyst. The synthetic reaction was carried out in a horizontal quartz tube furnace using high-purity Ar:5% H2 (20 sccm) as the carrier gas at 1100 °C for 2 h, which yielded a white woollike * Corresponding authors. E-mail: zurich.ibm.com. 10.1021/nl052507j CCC: $33.50 Published on Web 04/25/2006
[email protected]; alv@
© 2006 American Chemical Society
product covering the substrate. Under these growth conditions, ZnS NWs with wurtzite structure are produced. Further details are described in refs 11 and 12. The experiments were performed on two different samples, which were prepared on graphite (highly oriented pyrolytic graphite (HOPG)) substrates by drop-coating (a) a low-density and (b) a highdensity ZnS-nanowire/ethanol suspension prepared by ultrasonication in ambient air under normal conditions. Immediately after preparation, the samples were inserted into an ultrahigh vacuum (UHV) chamber by means of a load lock. All STM as well as CL excitation measurements were performed under UHV conditions, p ) 10-10 mbar, at room temperature. The UHV-STM apparatus with optical excitation/detection capability is described in refs 13 and 14. STM images were collected at tunneling currents as low as 2 pA. They show NWs in the sample, although it should be mentioned that the topographs also reveal regions that exhibit structures that do not resemble NWs; see Figure 1a. An scanning electron microscopy (SEM) micrograph of the NWs deposited on HOPG is shown in Figure 1b. From this image, we estimate the diameter of the NWs to be in the range of 20-50 nm and an NW length of up to a few micrometers. It can also be seen that a fraction of the NWs has a Au-Si alloy nanoparticle, approximately 25 nm in diameter,3 attached at one end. An analysis by energy-dispersive X-ray spectroscopy confirmed the chemical composition of the sample to be ZnS.
Figure 2. Normalized PL and CL spectra of the ZnS nanowire sample. PL excitation at λ ) 266 nm.
Figure 1. (a) STM topograph and (b) SEM micrograph of a ZnS NW on HOPG sample.
Cathodoluminescence was excited by means of low-energy electrons field-emitted from the tip of the STM14 at voltages in the range of 150-350 V and emission currents of a few nanoamperes up to 32 µΑ. Τhe typical tip-sample distance was in the range of a few micrometers, resulting in an impinging power density as high as 60 kW/cm2 and more on the sample. The STM and SEM images show that the NWs lie flat on the substrate. Because of “light piping” along the wire axis, we therefore expect emission from the NWs to occur with a predominant orientation in that plane. Accordingly, a special sample holder was constructed that allows one to align the plane of the substrate within the acceptance cone and close to the axis of the optical detection system. Photoluminescence (PL) spectra were taken in the UHV chamber by excitation with a diode laser at λ ) 266 nm, 1 mW power, and a beam diameter of 0.5 mm. The spectra were collected in the wavelength range of 290-1000 nm using an optical multichannel analyzer (OMA) and are corrected for the response of the detection system. The polarization of the emission in the 400-1000 nm range was analyzed by rotating a linear polarizer placed in the path of the emitted light. CL measurements on bare HOPG substrates and on HOPG exposed to ethanol show that any contributions from the substrate to the spectra are negligible compared with the emission intensity of the ZnS NW samples. Nano Lett., Vol. 6, No. 5, 2006
Regarding the stability of the NWs upon electron irradiation, we note that in some exceptional cases the emission spectrum exhibited changes upon prolonged exposure of a given spot on the sample to the electron beam. This could be associated with current density peaks that occur during some particular measurements and which can induce modifications in the surface region of the sample due to Coulombic loading. We note that at higher energies, e.g., above 1 kV, where the penetration of the electrons is higher, degradation of ZnS samples can take place.15-17 Figure 2 shows PL and CL spectra collected on overlapping spots of the sample. The typical PL spectrum exhibits a single peak centered at ca. λ ) 380 nm (3.26 eV) due to near band-edge emission. The PL cutoff wavelength is approximately 316 nm (3.92 eV), which agrees with the cutoff observed in the PL spectrum previously reported.12 This cutoff is blue-shifted with respect to that of the highpurity ZnS source powder used for synthesizing NWs: 344 nm (3.6 eV).12 The PL peak of our sample is much broader and lies at a lower energy than that of a different ZnS NW sample.12 In addition, the spectrum exhibits a longwavelength tail extending up to about 650 nm. A possible explanation for the broad PL spectrum is the presence of a size distribution of NWs in the sample. We note that a broad PL emission band centered at 394 nm with a half-width at half-maximum (hwhm) of 75 nm has recently been reported by Kar et al. for a ZnS sample whose NW diameters vary within 15-30 nm.18 As can be seen in Figure 2, the CL exhibits exactly the same peak found in PL, but also additional emission bands. A strong emission line, peaking at approximately 450 nm, appears in the spectrum, which is believed to arise from surface states (see ref 18 and refs 23-25 therein). In addition a red emission line, peaking at 610 nm (2.03 eV), can be seen. Analysis of several hundred CL spectra reveals that the energy and relative intensity of these additional emission bands can differ at different spots of the sample. Figure 3 compares spectra in which it can be seen that the intensity of the transitions at 450 and 535 nm, for example, change upon lateral displacement of the tip by about 15 µm. Remarkably, some of these emission bands coincide in energy with those observed in PL spectra of doped ZnS bulk material.19,20 This suggests the presence of optically active 927
Figure 3. Illustration of changes in the relative intensity of the optical transitions observed for CL spectra (A) and (B) collected at different spots on the ZnS NW sample.
impurities21 in the surface region of the ZnS NWs, such as oxygen,22 and/or native lattice defects, such as S- or Zn-ion vacancies (see ref 19 and references therein). Zn-ion vacancies, for instance, are known to give rise to emission in the blue, whereas oxygen causes emission in the red. Another characteristic of these ZnS transitions is that they are polarized.19-22 We note that emission in the range about ca. 500 to 550 nm is also expected due to plasmon excitation in the Au-Si alloy nanoparticles23,24 attached to the tip of the NWs. The fact that the CL spectra reveal emission bands that do not appear in the PL spectrum can be explained by the high sensitivity of the STM-CL technique to surface properties. This is so because the inelastic mean free path of low-energy electrons in solids is on the order of a few nanometers.25 On the other hand PL can also probe the core of the NWs because the probing depth of photons can be several orders of magnitude larger. Measurements of the polarization of the CL radiation are shown in Figure 4a. We note that because of the difference in reflectance of s- and p-polarized light one would expect light emitted from the NWs to become s-polarized (electric field oscillating parallel to plane of substrate); i.e., we expect maximum CL intensity when the analyzing linear polarizer is oriented parallel to the plane of the NWs substrate, θ ) 90°. This is indeed the general trend revealed by the data. We find, however, indications of a shift of the intensity minimum, expected at θ ) 0° (parallel to the STM tip axis), by approximately -15° to -20° for the 527- and 785-nm bands. In addition, the intensity of the peak at 650 nm seems to be maximum near θ ) 0° and decreasessdisappearing in the backgroundsat higher angles. It can also be seen that the relative change of the CL with increasing angle is much smaller that for the 527- and 785-nm bands. These observations suggest that the polarization of this band is crossed relative to that of other transitions. Measurements of CL intensity vs bias voltage, measured at constant tip-sample distance, indicate a linear dependence of CL vs power of the electron beam; see Figure 5. The fieldemitted current varies from several nanoamperes up to 32 µΑ with increasing voltage. Close to the maximum current, CL was observable by the naked eye through a telescope in the darkened laboratory. Figure 6 shows CL spectra collected 928
Figure 4. (a) Polarization of CL spectra seen upon rotation of the linear polarizer by an angle θ, where θ ) 0° corresponds to the linear polarizer parallel to the axis defined by the tip of the STM. (b) Intensity vs angle of the linear polarizer. The error bars in Figure 4b represent intensity variations observed upon repeated measurements. The lines connecting the points are meant as a guide to the eye.
Figure 5. CL intensity vs power of the exciting low-energy electron beam impinging on the ZnS NW sample, measured at λ = 500 nm. Tip-sample distance is kept constant during the measurement.
at different emission currents at constant tip-sample distance. Owing to focusing, the diameter of the impinging beam is smaller than the tip-sample distance. A conservative estimate yields power densities of up to 60 kW/cm2, but the actual maximum could be an order of magnitude higher. For high injection efficiency, this might well be within the range of several 10 kW/cm2 required to excite laser emission by Nano Lett., Vol. 6, No. 5, 2006
Acknowledgment. We thank R. Allenspach for his support in the realization of this project and R. Schlittler for technical advice. D. D. D. Ma is grateful to S. F. Alvarado, P. Mu¨ller, and R. Allenspach for their hospitality and support extended to her at the IBM Zurich Research Laboratory. S. T. Lee acknowledges the financial support of the Research Grants Council of Hong Kong SAR (Project No. CityU 101504 and N_CityU125/05), China. References
Figure 6. CL spectra collected at different excitation voltages of the electron beam emitted from the tip of the STM.
optical pumping in NW materials such as ZnS, ZnO, and CdS.2-4,7 If, however, the injection efficiency is low, for example because of surface defects, contaminants, etc., the actual injected power density into the core of the NWs could be much smaller and thus may not be sufficient for laser excitation. Note that the spectral changes in Figure 6 that occur upon increasing the current up to 32 µA, i.e., the intensity of the NW-core emission peak at approximately 390 nm, decrease markedly. This suggests that high-density electron irradiation changes the surface properties, causing a reduction of the excitation efficiency of the core of the NWs. In summary, we have described the use of an STM to excite cathodoluminescence in ZnS nanowires with high spatial resolution. Because of the high surface sensitivity of this technique, it is possible to characterize the surface region of the NWs by probing its optical excitations. We thus find blue, green, and red emission bands in the CL spectra, which are much weaker or absent in the bulk-sensitive PL spectra obtained on the same sample. In addition some of these transitions appear to be polarized. This information can be useful to identify modifications of the chemical composition, the stoichiometry, and the presence of crystal defects at the surface region of the nanostructures. The STM-based CL technique allows one to obtain excitation energy densities in the range of several 10 kW/cm2 using electrons of energy in the range of a few 100 eV. This should be sufficient to excite laser emission in a single wire, provided the excitation energy can be efficiently transported into the core of the wires.
Nano Lett., Vol. 6, No. 5, 2006
(1) Kind, H.; Yan, H.; Messer, B.; Law, M.; Yang, P. D. AdV. Mater. 2002, 14, 158-160. (2) Huang, Y.; Duan, X.; Lieber, C. M. Small 2005, 1, 142-147. (3) Ding, J. X.; Zapien, J. A.; Chen, W. W.; Lifshitz, Y.; Lee, S. T.; Meng, X. M. Appl. Phys. Lett. 2004, 85, 2361-2363. (4) Zapien, J. A.; Jiang, Y.; Meng, X. M.; Chen, W.; Au, F. C. K.; Lifshitz, Y.; Lee, S. T. Appl. Phys. Lett. 2004, 84, 1189-1191. (5) Duan, X.; Huang, Y.; Agarwal, R.; Lieber, C. M. Nature 2003, 421, 241-245. (6) Agarwal, R.; Barrelet, C. J.; Lieber, C. M. Nano Lett. 2005, 5, 917920. (7) Johnson, J. C.; Yan, H.; Schaller, R. D.; Haber, L. H.; Saykally, R. J.; Yang, P. D. J. Phys. Chem. B 2001, 105, 11387-11390. (8) Maslov, A. V.; Ning, C. Z. Appl. Phys. Lett. 2003, 83, 1247-1249. (9) Huang, M. H.; Mao, S.; Feick, H.; Yan, H.; Wu, Y.; Kind, H.; Weber, E.; Russo, R.; Yang, P. Science 2001, 292, 1897-1899. (10) Ko¨nenkamp, R.; Word, R. C.; Schlegel, C. Appl. Phys. Lett. 2004, 85, 6004-6006. (11) Meng, X. M.; Liu, J.; Jiang, Y.; Chen, W. W.; Lee, C. S.; Bello, I.; Lee, S. T. Chem. Phys. Lett. 2003, 382, 434-438. (12) Jiang, Y.; Meng, X.-M.; Liu, J.; Hong, Z.-R.; Lee, Ch-S.; Lee, S. T. AdV. Mater. 2003, 15, 1195-1198. (13) Alvarado, S. F.; Riess, W.; Jandke, M.; Strohriegl, P. Org. Electron. 2001, 2, 75-82. (14) Alvarado, S. F.; Riess, W.; Seidler, P. F.; Strohriegl, P. Phys. ReV. B 1997, 56, 1269-1278. (15) Swart, H.; Sebastian, J. S.; Trottier, T. A.; Jones, S. L.; Holloway, P. H. J. Vac. Sci. Technol., A 1996, 14, 1697-1703. (16) Rodrı´guez-Viejo, J.; Mattousi, H.; Heine, J. R.; Kuno, M. K.; Michel, J.; Bawendi, M. G.; Jensen, K. F. J. Appl. Phys. 2000, 87, 85268534. (17) Lee, D. C.; Bekusov, S. A.; Lee, S.; Kang, J. H.; Yeon, D. Y.; Park, D. H.; Kim, J. Y. J. Electrochem. Soc. 2004, 151, H227-H231. (18) Kar, S.; Biswas, S.; Chaudhuri, S. Nanotechnology 2005, 16, 737740. (19) Koda, T.; Shionoya, S. Phys. ReV. 1964, 136, A541-A555. (20) Birman, J. L. Phys. ReV. Lett. 1959, 2, 157-159. (21) Suzuki, A.; Shionoya, S. J. Phys. Soc. Jpn. 1971, 31, 1719-1726. (22) Curie, D. Luminescence in Crystals; Methuen & Co. Ltd.: London, 1960, p 116. (23) Spadavecchia, J.; Prete, P.; Lovergine, N.; Rella, N. J. Phys. Chem. B 2005, 17347-17349. (24) Schaadt, D. M.; Feng, B.; Yu, E. T. Appl. Phys. Lett. 2005, 063106063109. (25) Tanuma, S.; Powell, C. J.; Penn, D. R. Surf. Interface Anal. 1991, 17, 927-939.
NL052507J
929
