научная статья по теме RELAXATION TIME MAPPING OF SINGLE QUANTUM DOTS AND SUBSTRATE BACKGROUND FLUORESCENCE Физика

Текст научной статьи на тему «RELAXATION TIME MAPPING OF SINGLE QUANTUM DOTS AND SUBSTRATE BACKGROUND FLUORESCENCE»

Pis'ma v ZhETF, vol. 102, iss. 3, pp. 186-191 © 2015 August 10

Relaxation time mapping of single quantum dots and substrate

background fluorescence

E. Pshenay-Severm+*1\ I. Mukhinx°, S. Fasold+, R. Geiss+, A. Steinhrck+, R. Grange+, A. Chipouline+,

T. Pertsch+

+Institute of Applied Physics, Abbe Center of Photonics, Friedrich—Schiller— Universitt Jena, 07743 Jena, Germany

* Nonlinear Physics Center, Research School of Physics and Engineering, Australian National University,

ACT 0200 Canberra, Australia

x St. Petersburg Academic University, 194021 St. Petersburg, Russia

°St. Petersburg National Research University of Information Technologies, Mechanics and Optics,

197101 St. Petersburg, Russia

Submitted 22 June 2015

We experimentally investigated the role ol background signal in time resolved photoluminescence experiments with single quantum dots on substrates. We show that the background fluorescence signal from thin gold films labricated by electron-beam evaporation and from AI2O3 layers labricated by atomic layer deposition have to be taken into consideration in experiments on the single photon level. Though all investigated components can be distinguished by their photoluminescence decay rates, the presence of the background signal prevents the observation of photon anti-bunching from single quantum dots. Moreover, a single quantum dot acts as a hot-spot enabling the plasmon supported fluorescence enhancement of gold.

DOI: 10.7868/S0370274X15150060

1. Introduction. With the rapid development of nanotechnology it has become possible to engineer and study hybrid quantum-classical systems at the nanometer scale. Of particular interest is the research field of plasmonic hybrid systems consisting of a quantum emitter, e.g. a quantum dot (QD), dye molecule or nitrogen-vacancy center, coupled to a plasmonic nanostructure supporting surface plasmon-polaritons [1-3]. For example, coupling to a plasmonic nanoantenna changes the radiative properties of a quantum system by enhancement of spontaneous emission rate [4, 5] and improvement of its directionality [6-8]. Moreover, in the strong coupling regime, when the distance between a nanoantenna and a single quantum emitter is merely several angstroms, quantum effects in the nanoantennas become relevant [9, 10]. The strong coupling regime is of great interest for fundamental science and its realization promises benefits for signal processing applications at the single-photon level [11].

Obviously, the conduction of experiments in the strong coupling regime on the single photon level requires a critical evaluation of experimental conditions. Nowadays, time resolved photoluminescence (PL) measurements are broadly used clS cl powerful method for

-^e-mail: katja.severin@uni-jena.de

the investigation of the interaction between quantum emitters and nanoantennas. With this technique a relaxation rate enhancement is routinely observed for quantum emitters in the vicinity of the nanoantennas [6]. At the same time an appropriate processing of the experimental data and their correct interpretation is the main problem. The first step in such complex experiments is supposed to be done on unstructured substrates in order to provide a reference. In spite of the many published experimental papers devoted to the topic of optical nanoantennas (see for example references in Ref. [12]), scant attention is paid to the accurate differentiation between effects caused by the structured and unstructured components such as substrates, for example. Additionally, only a few authors address possible pitfalls in such experiments associated with background fluorescence of the used materials [13].

In this contribution we present our experimental results on time resolved fluorescence measurements of single CdSe/ZnS quantum dots spin-coated on substrates to be used in single nanoantenna experiments. We targeted on a system consisting of a nanoantenna fabricated using focused ion beam lithography applied to an electron-beam evaporated thin gold film. The main goal of the performed test was twofold: first, to quantify the luminescence from the substrates in order to extract

186

IlHCbMa b >K3TO tom 102 Bbin. 3-4 2015

this luminescence from the following tests with single QDs; and second, investigate relaxation enhancement due to the substrates using the developed methodology. As quenching of quantum dots placed directly on a metallic surface can occur [14], we investigated whether an isolating layer made of AI2O3 can be used in such systems. As the dimension of the system is a critical parameter for the experiments in the strong coupling regime we worked with CdSe/ZnS QDs with outer diameter of about 5 nm. Additionally, we experimentally verified, whether the observation of photon anti-bunching from QDs in correlation measurements can be applied in the systems under investigation for the identification of a single quantum dot [15]. Obviously, if the background fluorescence signal is strong enough and spectrally overlaps the observation of photon anti-bunching fails.

A. Sample preparation. In our experiments we investigated CdSe/ZnS core shell QDs (maximum emission wavelength around A = 620 nm) commercially available from PlasmaChem GmbH [16]. The core diameter of these QDs was about 4.4 nm [17], the ZnS shell thickness varies between 0.5 and 0.7 nm.

Time resolved photoluminescence experiments on CdSe/ZnS QDs were conducted on the following substrates: commercially available borosilicate glass, glass covered with a 15 nm Au layer, and glass covered with a 15 nm Au and 5 nm thin AI2O3 layer further referred to as AU/AI2O3 layer stack. The substrates were prepared as follows. The cover slips were coated with 3 nm titanium as adhesive layer and 15 nm gold by an electron beam evaporator with a rate of O.lnm/s for titanium and 1 nm/s for gold. After this step some samples were coated with a thin AI2O3 layer produced by atomic layer deposition (ALD). For this process trimethylaluminium and water were used as reactants at a process temperature of 200 °C. The size of the gold grains was estimated to be about 20 nm.

For the photoluminescence experiments CdSe/ZnS QDs were spin-coated on cleaned substrates. Spinning was performed in two steps: 1) 15 s with 500 r.p.m. and 2) 30 s with 2000 r.p.m. speed. Substrates were cleaned first with acetone for 5 min, then in methanol for 5 minutes, rinsed with deionized distilled water and blown with pure nitrogen. QDs were suspended in toluol with different concentrations from 10~7 to 10~8ppm. In order to ensure the suspension of QDs in clusters the suspensions were ultrasonicated for 10 min before spin-coating on substrates.

B. Measurement technique and data processing. The PL measurements were performed with a commercial PicoQuant MicroTime 200 time-resolved confocal fluorescence microscope. For the excitation a picosecond

diode laser (FWHM about 70 ps) with 5 MHz repetition rate and a wavelength of 530 nm was used. The excitation and collection of the optical signal was performed with a water immersion 60x objective with NA = 0.6. The power in the center of the excitation beam was estimated to be 0.35kW/cm2. Detection of the optical signal was performed with single photon avalanche diodes. Data acquisition was based on the Time-Correlated Single Photon Counting (TCSPC) method [18]. The wavelength of the excitation laser was filtered out by a long pass filter with transmission only for wavelengths higher than 540 nm. The time resolution of the measurements is defined by the instrument response function (IRF) of the MicroTime 200 and is about 80 ps. Photon anti-bunching tests for the verification of single QDs were performed by Hanbury Brown-Twiss setup [19] incorporated into the MicroTime 200.

The data processing was performed using the recon-volution fitting method, while the IRF was measured using the reflection from a gold mirror. The number of exponents used to fit the experimental data depends on the number of physically different luminescence sources, which is a priori unknown. Thus, the number of the exponents was selected to provide the optimal fit of the original curves.

2. Experimental results. First, in the case of the borosilicate glass no detectable signal has been observed.

Second, we conducted time-resolved PL measurements on a pure gold layer and an AU/AI2O3 layer stack without QDs (see Fig. la). For the pure gold sample, fitting of the measured PL decay curves was performed using single exponential model that gave the relaxation time tau = 0.08±0.01 ns. At the pump intensity level of 0.35 kW/cm2 the origin of this relaxation curve could be both pump signal replica or a luminescence with a relaxation time less or comparable with the IRF. In the case of the AU/AI2O3 layer stack a two exponential fitting function was used, which resulted in two decay times ti,au = 0.08 ± 0.01ns and t2iai2o3 = 3 ± 0.2 ns. The first time is identified, as in the case of pure gold, as a pump replica plus and possible fast luminescence from gold, while the second time is attributed to the luminescence of the AI2O3 layer.

Third, a series of experiments were performed with QDs on substrates. Positions of single QDs were first tested by means of confocal scanning of the optical signal in a scanning area of 80 x 80 /um2. At concentrations of 10~7ppm and below we observed well separated single QDs. In Figs. 2a and 3a a scan of the PL signal for a single CdSe/ZnS QD on a pure borosilicate glass and a typical time trace are shown, respectivly. It is notice-

micbMa b >K3TO tom 102 Htm. 3-4 2015

3*

t

t (ns)

Fig. 1. (a) - Time-resolved photoluminescence for a gold layer and AU/AI2O3 layer stack, (b) - PL decay dynamic measured from a hot spot on a gold layer formed by a bleached QD; PL decay dynamics from a gold substrate without a hot spot

-1.0 -0.5 0 0.5 t (ms)

Fig. 2. Results for a single CdSe/ZnS QD on a pure borosilicate glass, (a) - Scan of PL signal

Для дальнейшего прочтения статьи необходимо приобрести полный текст. Статьи высылаются в формате PDF на указанную при оплате почту. Время доставки составляет менее 10 минут. Стоимость одной статьи — 150 рублей.

Показать целиком