Wavelength Tunable Triggered Single-Photon Source from a Single

Dec 9, 2008 - A nearly perfect single-photon emission from the exciton lines is preserved even after energy tuning. The lifetime is also measured befo...
0 downloads 0 Views 106KB Size
NANO LETTERS

Wavelength Tunable Triggered Single-Photon Source from a Single CdTe Quantum Dot on Silicon Substrate

2009 Vol. 9, No. 1 304-307

Mohamed Benyoucef,*,†,‡ Hong Seok Lee,‡ Juliane Gabel,‡ Tae Whan Kim,| Hong Lee Park,§ Armando Rastelli,‡ and Oliver G. Schmidt‡ Institute for IntegratiVe Nanosciences, IFW Dresden, Helmholtzstrasse 20, D-01069 Dresden, Germany, Max-Planck-Institut fu¨r Festko¨rperforschung, Heisenbergstrasse 1, D-70569 Stuttgart, Germany, AdVanced Semiconductor Research Center, DiVision of Electronics and Computer Engineering, Hanyang UniVersity, 17 Haengdang-dong, Seongdong-gu, Seoul 133-791, Korea, and Institute of Physics and Applied Physics, Yonsei UniVersity, Seoul 120-749, Korea Received September 27, 2008; Revised Manuscript Received November 21, 2008

ABSTRACT Triggered single-photon emission from a single CdTe quantum dot (QD) grown on Si(001) substrate is demonstrated for the first time. The emission wavelength of QDs can be tuned in a wide spectral range (more than 8 meV) using a focused laser beam. A nearly perfect singlephoton emission from the exciton lines is preserved even after energy tuning. The lifetime is also measured before and after laser processing, and no appreciable change is observed.

Single quantum emitters have become an emerging area of fundamental research during recent years, driven by the need for nonclassical light sources delivering single photons on demand for future implementation in the field of quantum information including quantum cryptography1 and quantum computing.2 It has also been shown that the availability of single-photon sources able to generate indistinguishable photons enables the implementation of quantum computation using only linear optical elements and single-photon detectors.3 Other possible applications are imaging and lithography beyond the diffraction limit as well as quantum teleportation. Recent experiments have shown that self-assembled quantum dots (QDs) are good candidates for the production of triggered single photons,4-6 indistinguishable photons,7 and entangled photon pairs.8 Most of the work concerning photon statistics has been done on III-V materials such as (In,Ga)As QDs embedded in GaAs matrix or integrated in microcavities.4,6,9 In contrast to this, only few studies on II-VI compounds10-12 have been reported. II-VI QDs are interesting due to their large excitonic binding energies, shorter radiative lifetimes, and the strong Coulomb interaction which make them very * Corresponding author. E-mail: [email protected]. Phone: +49 (0)351 4659 670. Fax: +49 (0)351 4659 782. † Max-Planck-Institut fu¨r Festko¨rperforschung. ‡ Institute for Integrative Nanosciences, IFW Dresden. | Hanyang University. § Yonsei University. 10.1021/nl802948a CCC: $40.75 Published on Web 12/09/2008

 2009 American Chemical Society

attractive for single-photon generation and quantum optical experiments. The short lifetime could also allow operation at high repetition rates. From a technological point of view it would be desirable to obtain single photons on demand from single quantum dots grown on conventional Si(001) substrates. The integration of classical optoelectronic devices on Si is a long standing dream and keeps being a formidable challenge, as it is usually difficult to grow direct bandgap compound semiconductors on Si. Planar structures with QDs have been grown on Si,13 but single-photon emission has not been reported so far. In this Letter, we report on the observation of triggered single-photon emission from a single CdTe QD embedded in a ZnTe layer grown on Si(001). The emission wavelength of single dots can be locally tuned in a wide spectral range (more than 8 meV) by means of a focused laser beam. Moreover, we have carried out time-resolved photoluminescence (TRPL) measurements to investigate the lifetime of the single CdTe QDs. We find that the single-photon emission and the lifetime are preserved after energy tuning of the QD emissions. The studied sample was grown by molecular beam epitaxy (MBE) and atomic layer epitaxy (ALE) on a Si(001) substrate and consists of a single CdTe QD layer (with dot density of about 3 × 1010 cm-2) embedded in ZnTe barriers. The Si(001) substrate was etched in a mixture of NH4F and HF (7:1) at room temperature for 1 min and rinsed in deionized

water prior the growth, and subsequently, the substrate is mounted onto a molybdenum susceptor. A 900 nm ZnTe buffer layer was first grown on Si(001) substrate at 320° using MBE, followed by 2.5 monolayers of CdTe grown at the same temperature using ALE. The CdTe QDs were then capped with 100 nm ZnTe using MBE (see ref 14 for further details). For the laser-heating experiment, a part of the sample was coated with 200 nm of SiO2 using electron beam evaporation. The PL measurements are performed at 6 K; the sample is mounted in a helium flow cryostat which can be moved by computer-controlled xy-linear stages with spatial resolution of 50 nm. A microscope objective (NA ) 0.5) is used to focus a frequency-doubled Nd:YVO4 continuous wave (CW) laser with an excitation wavelength of 532 nm for PL and laser processing or a frequency-doubled Ti-sapphire laser operating at 450 nm (pulse width of ∼2 ps and repetition frequency of 76.2 MHz) for photon correlation and time-resolved measurements. The PL from the dots is then collected by the same microscope objective and sent to a 50/50 beamsplitter for correlation measurements. In each arm after the beamsplitter, the light is dispersed by a 75 cm focal length monochromator. Each spectrometer is equipped with a flip mirror to switch the output between a Si-CCD camera for the visualization of the spectrum or a single-photon counting avalanche photodiode (SAPD) for photon correlation measurements. The SAPDs output signals were used to trigger the start and stop channels of a time-to-amplitude converter (TAC) the output of which was stored in a PC-based multichannel analyzer (MCA). In this way, a histogram n(τ) of photon correlation events as a function of the time delay τ ) tstop - tstart is recorded. The TRPL measurements are performed by using a time-correlated single-photon counting technique. The signal is dispersed by a 75 cm monochromator and detected by a fast APD with temporal resolution of about 35 ps. To demonstrate triggered single-photon generation, we measured the second-order correlation function g(2)(τ) under pulsed excitation. Figure 1 shows the measured unnormalized correlation function n(τ) under pulsed excitation of the exciton (X) emission of a single QD before and after energy tuning by laser heating (see below) and the corresponding PL spectra. All displayed measurements were performed at low excitation power to prevent blue shift of the QD emission lines. The measured n(τ) consists of series of correlation peaks separated by the repetition period 13.12 ns. As expected, in the case of a coherent source (pulsed laser), all peak areas are identical (see Figure 1a) showing Poisson distributed statistics. In contrast to the mode-locked laser, the central peak at τ ) 0 ns of the QD X emission is significantly suppressed, an unambiguous signature of a single-photon source. Note that the width of the X photon correlation peaks is comparable to that of the laser. This is due to the short X lifetime (see below) so that the width is limited by the time resolution of the setup (700 ps). For perfect single-photon emission the central peak is absent indicating the generation of only one photon per pulse. For the investigated QD, a practically perfect single-photon emission is observed. The value of the second-order corNano Lett., Vol. 9, No. 1, 2009

Figure 1. Photon correlation measurements of (a) a pulsed laser and (b and c) the X line from a single CdTe QD in a sample coated with 200 nm SiO2 before and after laser heating. (d and e) The corresponding PL spectra before and after energy tuning by laser heating.

relation g(2) (0) is about 0.14 for the central peak, which does not reach its theoretical value of zero due to the presence of a weak uncorrelated background. A problem related to the use of self-assembled QDs as independent sources for indistinguishable single photons is that such QDs emit at different wavelengths due to unavoidable fluctuations in size/shape/composition during the fabrication processes. Therefore, it is important to find strategies to accurately tune the emission of single QDs into resonance ideally within the lifetime-limited dephasing time. We have already shown that with laser heating it is possible to tune into resonance spatially separated QDs.15 We applied a similar method here to shift the X lines and then verify the single-photon emission (Figure 1, panels c and e). As discussed below, using laser heating, we first blue shift the X line and then measure the autocorrelation again using the pulsed laser. Interestingly, the single photon emission from the same dot after laser processing is preserved (Figure 1c). Our findings show that it is feasible to use the in situ laser processing to tune a number of independent single-photon emitters into resonance with each other to possibly fabricate on-demand indistinguishable single-photon sources. Figure 2 demonstrates that the focused laser beam used for µ-PL spectroscopy can also be used to controllably blue shift the emission of a single CdTe/ZnTe QD. The approach is similar to that used in ref 15 for InGaAs/GaAs QDs. At low laser power (here 50 µW) we collect µ-PL spectra; at high laser powers (here up to 18 mW) we heat the structure and promote structural changes producing a blue shift of the emission. Parts a and b of Figure 2 compare the emission behavior for two different QDs in the as-grown sample (Figure 2a) and in the same sample, but coated with a 200 nm SiO2 layer (Figure 2b). After each heating step (of 5 s duration), the excitonic X line is gradually blue shifted. Figure 2c illustrates the peak position and the line width as 305

Figure 2. PL spectra of a CdTe QD anealed several times at different laser powers and collected at 50 µW from (a) as-grown sample and (b) sample coated with 200 nm SiO2. (c) Peak position and line width of exciton as a function of heating powers from the coated sample.

Figure 3. Exciton peak position from a single QD in the SiO2 coated sample during laser heating and cooling (a) and as a function of sample temperature (b). In (a) the laser power is constant in regions A and E (100 µW) and in region C (13 mW) while it is continuously increased in region B and decreased in region D. The X peak position measured while ramping down the laser power is plotted in the inset of (a). The time scale of the experiment is indicated in the horizontal axis.

a function of heating power of X from the SiO2 coated sample. At a laser power of 18 mW, we obtain a maximum energy shift of about 8 meV for the sample coated with SiO2 (Figure 2c). In contrast to this, only a small shift is observed for the QD in the uncoated sample (0.7 meV). The line width remains almost constant at reasonably low heating powers up to 13 mW and then increases gradually for the investigated QDs suggesting that no significant damage is produced by the laser heating. Interestingly, the single-photon emission from the same dot after “laser processing” is preserved (see Figure 1c). In similar experiments performed on InGaAs/GaAs QDs15 the blue shift was attributed to intermixing between QD and barrier material. In order to understand the mechanism responsible for the blue shift in the present experiment, we probe the local temperature at the QD position by following the QD emission while increasing and decreasing the laser power (see Figure 3a). At high power, the emission red shifts because of local temperature rise. This red shift is partially compensated by a gradual blue shift, as seen when the power is reduced again (see Figure 2b). In order to eliminate this 306

Figure 4. (a) Peak intensity as a function of excitation power and (b) lifetime of X line from a single CdTe QD before and after energy tuning. The inset shows the corresponding PL spectra of the X line. Data collected from the SiO2 coated sample.

effect and determine the shift due to only temperature increase, the laser power was first ramped up to 13 mW (region B), kept constant until the peak position reached a stable value (region C), and then lowered again (region D). The X peak position measured while ramping down the laser power is plotted in the inset of Figure 3a. To extract the temperature from the energy shift, we performed a “calibration measurement”, i.e., measured the QD emission energy when the whole sample is heated (Figure 3b). By comparing the maximum red shift (6 meV) caused by the laser heating and the red shift produced by controlled temperature increase, we extract a maximum temperature at the QD location of about 60 K. At such a temperature, intermixing (i.e., group II atom exchange) is completely unrealistic. Therefore, the X blue shift of single QD cannot be explained by the interdiffusion of Cd and Zn atoms at the interface between the CdTe QDs and the ZnTe. It has been shown that capping CdTe QDs with ZnTe of increasing thickness causes the peak position to shift to higher energy due to the compressive stress generated by the ZnTe cap layer.16 The small blue shift in the as-grown sample indicates that the compressive stress increases even further when the sample is heated with the laser beam. Coating the sample with a 200 nm SiO2 layer introduces tensile stress which results in a red shift of the QD emission by about 5 meV (not shown here). We argue that by heating the sample at different powers we blue shift the QD emission by gradually restoring the initial stress situation (before SiO2 coating), e.g., by the formation of a gap at the interface between ZnTe and SiO2. Although the exact mechanism remains unclear, it is important to note that the tuning of QD emissions is local (i.e., the emission energy of QDs a few micrometers away from the heating spot is not affected). When the peak shift reaches to the original energy position (before coating with SO2), the line width broadens and is comparable with the line width of the uncoated sample. This could explain the large energy shift of the X line from the SiO2 coated sample. Figure 4a presents the peak intensity of the X line as a function of power before (full symbols) and after (empty symbols) energy tuning, respectively. The saturation peak intensity is similar in both cases which indicates that no additional nonradiative recombination channels are created Nano Lett., Vol. 9, No. 1, 2009

after laser processing. Time-resolved measurements can help us to further understand the dynamics of the recombination process. The lifetimes of X lines before (full symbols) and after (empty symbols) laser processing are shown in Figure 4b. In both cases the maximum is reached at the same time. A biexponential decay is present in all recorded spectra, in agreement with previous observations on CdTe/ZnTe QDs.12 The fast component represents the excitonic radiative lifetime, while the weak slow component is usually attributed to dark exciton contribution in II-VI systems.17,18 The fitted data of different QDs provide radiative lifetime values ranging from 260 to 300 ps which are comparable to those observed in previous work on CdTe/ZnTe QDs grown GaAs substrate.19 This is an indication of equivalent crystalline quality (at least in the surrounding of the QD). The lifetime after the laser treatment is preserved which confirms that no additional nonradiative recombinations centers were introduced by the laser heating. In conclusion, we have demonstrated triggered singlephoton emission using a single CdTe quantum dot grown on a silicon substrate before and after energy tuning. We have also shown a large energy tuning from single quantum dots on a planar substrate. These results indicate that it may become feasible to fabricate indistinguishable single-photon sources. By use of a time-correlated single-photon counting technique, the lifetime of the single QD was measured. No change in the lifetime was observed after the laser processing. Acknowledgment. This work was supported by the BMBF (01BM459), and the DFG (FOR730). One of the authors (H. S. Lee) was supported by the Korea Research Foundation Grant funded by the Korean Government (KRF2008-357-C00035).

Nano Lett., Vol. 9, No. 1, 2009

References (1) Brassard, C. H.; Bennet, G. Proceedings of IEEE International Conference on Computers, Systems and Signal Processing, Bangalore, India, 1984, pp 175-179. (2) Loss, D.; DiVincenzo, D. P. Phys. ReV. A 1998, 57 (1), 120–126. (3) Knill, E.; Laflamme, R.; Milburn, G. J. Nature (London) 2001, 409, 46–52. (4) Ge´rard, J. M.; Gayral, B. J. LightwaVe Technol. 1999, 17 (11), 2089– 2095. (5) Michler, P.; Kiraz, A.; Becher, C.; Schoenfeld, W. V.; Petroff, P. M.; Zhang, L.; Hu, E.; Imamoglu, A. Science 2000, 290, 2282–2285. (6) Benyoucef, M.; Ulrich, S. M.; Michler, P.; Wiersig, J.; Jahnke, F.; Forchel, A. J. Appl. Phys. 2005, 97 (2), 023101. (7) Santori, C.; Fattal, D.; Vucˇkovic´, J.; Solomon, G. S.; Yamamoto, Y. Nature (London) 2002, 419, 594–597. (8) Stevenson, R. M.; Young, R. J.; Atkinson, P.; Cooper, K.; Ritchie, D. A.; Shields, A. J. Nature (London) 2006, 439, 179–182. (9) Moreau, E.; Robert, I.; Ge´rard, J. M.; Abram, I.; Manin, L.; ThierryMieg, V. Appl. Phys. Lett. 2001, 79 (18), 2865–2867. (10) Ulrich, S. M.; Strauf, S.; Michler, P.; Bacher, G.; Forchel, A. Appl. Phys. Lett. 2003, 83 (9), 1848–1850. (11) Aichele, T.; Zwiller, V.; Benson, O.; Akimov, I.; Henneberger, F. J. Opt. Soc. Am. B 2003, 20 (10), 2189–2192. (12) Couteau, C.; Moehl, S.; Tinjod, F.; Ge´rard, J. M.; Kheng, K.; Mariette, H. Appl. Phys. Lett. 2004, 85 (25), 6251–6253. (13) Li, L.; Guimard, D.; Rajesh, M.; Arakawa, Y. Appl. Phys. Lett. 2008, 92 (26), 263105. (14) Lee, H. S.; Park, H. L.; Lee, I.; Kim, T. W.; J., Appl. Phys. 2007, 102 (10), 103507. (15) Rastelli, A.; Ulhaq, A.; Kiravittaya, S.; Wang, L.; Zrenner, A.; Schmidt, O. G. Appl. Phys. Lett. 2007, 90 (7), 73120. (16) Lee, H. S.; Park, H. L.; Kim, T. W. Appl. Phys. Lett. 2008, 92 (5), 052108. (17) Labeau, O.; Tamarat, P.; Lounis, B. Phys. ReV. Lett. 2003, 90 (25), 257404. (18) Patton, B.; Langbein, W.; Woggon, U. Phys. ReV. B 2003, 68 (12), 125316. (19) Suffczys´ki, J.; Kazimierczuk, T.; Goryca, M.; Piechal, B.; Trajnerowicz, A.; Kowalik, K.; Kossacki, P.; Golnik, A.; Korona, K. P.; Nawrocki, M.; Gaj, J. A.; Karczewski, G. Phys. ReV. B 2006, 74 (8), 085319.

NL802948A

307