Characterizing Quantum Dots and Color Centers in Nanodiamonds as Single Emitters

Transcription

1 University of Rochester OPT253 Lab 3-4 Report Characterizing Quantum Dots and Color Centers in Nanodiamonds as Single Emitters Author: Nicholas Cothard Peter Heuer Professor: Dr. Svetlana Lukishova November 20th 2013

2 List of Figures 1 Bandgap samples A Diagram of a confocal microscope (Image from Wikipedia) Picture of our Hanbury-Brown and Twiss interferometer Picture of TTL pulse as observed on oscilloscope Wiring diagram for calibrating time delay Wiring diagram for antibunching measurement Screen capture of quantum dot raster scan Screen captures of antibunching curves Diagram of AFM (Image from Wikipedia) Test samples AFM images of a quantum dot sample Wiring diagram for fluorescence lifetime measurement Pulse from laser Quantum Dot Fluorescence Images from Confocal Microscope Quantum Dot Fluorescence Curves Nanodiamond Sample Sites Nanodiamond Fluorescence Curves Inverted image of the laser through the confocal microscope nm spectral line of the excitation laser Spectrum of single emitters Quantum dots blinking Screen capture showing a bleached signal flow List of Tables 1 Quantum Dot Fluorescence Lifetimes (ns) Nanodiamond Fluorescenec Lifetimes (ns) Abstract Single photon sources utilizing single emitters are essential for the realization of quantum technology. We produced single photons (antibunched light) with a quantum dot source, and measured the fluorescence lifetimes of quantum dots and nanodiamond color centers to be approximately 4ns and 3ns respectively. A spectrometer, atomic force microscope, and electron-multiplying CCD camera were also used to further characterize the samples. This investigation suggests that nanodiamond color centers are a superior single emitter because they are less susceptible to blinking and bleaching phenomena that can destroy quantum dots. 2 Introduction Modern quantum technology increasingly relies on the ability to produce a weak beam of spatially separated photons, known as a single photon source. This is complicated by the fact that photons are bosons, which statistically tend to clump together. Simply attenuating a laser beam to the single photon level will create a much weaker beam, but the photons will naturally distribute themselves into groups rather than spread out evenly. This phenomenon is called photon bunching. Photon bunching can be most easily circumvented by producing each photon in the beam with a single emitter, such as an individual atom or molecule. When a single emitter is stimulated, it emits a single photon. The emitter must then wait a characteristic fluorescence lifetime before emitting a second photon. If the fluorescence lifetime is long enough, the first photon will have 1

3 moved too far away to interact with the second photon by the time it is produced. The resulting beam of evenly spatially spaced photons is said to be antibunched. In order to stimulate only one emitter, a very dilute solution of single emitters is deposited onto a slide. As the solution is spread out, the single emitters will be left on the slide relatively far away from one another. A single emitter can then be stimulated with a highly focused laser beam through a confocal microscope. A wide range of single emitters, from atoms and ions to molecules, have been used to produce antibunched light. In this lab we studied the properties of two promising single emitters; colloidal quantum dots and nanodiamond color centers. Colloidal quantum dots are small pockets of semiconductor with a small band gap embedded within a larger semiconductor with a larger band gap. These two structures combine to create discrete energy levels, which then absorb and emit light much like a single atom. Color centers in nanodiamonds are common defects in the diamond lattice where two adjacent carbon atoms are missing and one has been replaced by a nitrogen atom. The resulting color center also has an energy level structure that makes it behave much like an atom, just like a quantum dot. Determining which single emitters are right for which applications requires a thorough understanding of their physical and optical properties. We prepared samples of quantum dots and nanodiamonds, suspended in toluene and chiral nematic liquid crystal respectively. We physically examined these samples with an atomic force microscope, demonstrated that they produced antibunched light, and collected data on their fluorescence lifetimes and spectral properties. We also investigated several drawbacks of quantum dots as compared to nanodiamonds, such as their tendency to blink on and off and eventually bleach. 3 Preparing Quantum Dot and Nanodiamond Samples In Lab 3-4 we created two different types of single emitters. First, we created a sample from a diluted quantum dot solution. To do this we placed 10µL of the prepared diluted quntum dot solution onto a microscope glass cover and placed it on a spin coating machine. The platform was spun at 3000rpm for approximately 30s while the slide was held onto the platform with suction. The sample was dry and evenly distributed so it was ready to be looked at with the confocal microscope. The second type of single emitter that we prepared were single emitters in a photonic bandgap cholesteric liquid crystal host. To prepare this, we used capillary tubes to deposit a drop of the chiral nematic liquid crystal blend onto the microscope slide. Then, we deposited a drop of the nanodiamond solution onto the slide next to the liquid crystal blend. We waited for the water in the solution to evaporate and then mixed the liquid crystal blend into the quantum dot remains. We mixed for about five minutes and then placed a glass cover slip on top of the sample. After application of pressure on the slides, the sample turned red (Figure 1a) and blue depending on the viewing angle. Another method of creating a photonic bandgap host is to begin with a solid form of the liquid crystal on a glass slide. When placed on a hotplate and heated to about 200 C, the solid enters a liquid crystal state becomes transparent. Once it reaches a temperature about about 240 C, the liquid crystal enters a fully liquid state and becomes colored again. Placing a slide on top of the sample and applying pressure, we saw a blue-green colored photonic bandgap material (Figure 1b). 2

4 (a) Red bandgap material from cholesteric liquid crystal host (b) Blue-green bandgap material from solid liquid crystal Figure 1: Bandgap samples 4 Demonstrating Antibunching To measure the anti-bunching properties of our quantum dot sample, we used a confocal microscope to control and drive the emissions of our single emitter. A confocal microscope uses an laser to excite our single emitter sample which to release signal photons of a different wavelength. For the antibunching measurements, we used a HeNe 633nm excitation laser which produced 800nm photons from the single emitter. Figure 2: A Diagram of a confocal microscope (Image from Wikipedia) We measured the continuous wave HeNe laser power to be 0.52mW before the confocal microscope and so, we applied two orders of magnitude attenuation infront of the confocal microscope input. The excitation laser enters the microscope setup and reflects off the dichroic mirror. A dichroic mirror is one which reflects a specific range of wavelengths and transmits others. The beam passes through the objective and onto the sample which through the process of fluorescence lifetime, produces anti-bunched 800nm photons. The 800nm photons pass through the dichroic mirror and out of the confocal microscope. To measure antibunching of the single emitters, we sent the single photons to a Hanbury-Brown and Twiss interferomter (Figure 3) to measure the coincidence timing of consecutive photons. If 3

5 the photons are completely antibunched, we expect that no photons will arrive at the same time. Figure 3: Picture of our Hanbury-Brown and Twiss interferometer The photons produced in the confocal microscope are sent into the interferomter which consists of a beam splitter, two photodetectors, and a computer which can measure the time between two photons arrival. In our setup, we used two APDs (avalanche photodiodes) to detect the photons and send electrical signals to the TimeHarp computer card which measures the time between detector firings. One APD is connected to the start channel of the TimeHarp while the other APD is connected to the stop channel. We need two APDs because each APD has a deadtime in which it must rest itself after detecting a photon. During the deadtime, the detector may miss a photon so a second detector is used in attempt to capture any photons that may arrive during deadtime. When the APDs absorb a photon, they emit a TTL (Transistor-Transistor Logic) pulse to the TimeHarp card. The TTL pulse is a step function pulse (Figure 4) which begins the time integration of the TimeHarp card. When a pulse hits the start channel of the card, capacitors in the card begin charging. When a pulse hits the stop channel of the card, the capacitor is released and time between start and stop pulses can be determined by the amount of charge that collected in the capacitor. From this, we can measure the time between consecutive photons. The TimeHarp card requires that the stop signal must be inverted with respect to the start signal, and that both signals be attenuatedl. Figure 4: Picture of TTL pulse as observed on oscilloscope 4

6 Using the start and stop pulses, we can build a histogram of time differences between consecutive photons. If our photons are anti-bunched, we would expect to see no data points for t = 0 but in our setup, we used an electronic time delay to delay the pulse of the stop channel so we expect to see no data points at t = t delay. To measure the value of t delay, we connected a single APD to both the start and the stop channels with an inverter in the stop channel to satisfy the TimeHarp card (Figure 5). The electrical pulses from the APDs were divided between the channels and the coincidence time measured between start and stop was the time that the stop signal was delayed. This was measured to be 57.34ns. Figure 5: Wiring diagram for calibrating time delay Figure 6: Wiring diagram for antibunching measurement To observe anti-bunching, both APDs were connected to the TimeHarp (Figure 6), the excitation laser was turned on, and a raster scan of the sample was created. To do this, we used a Piezo translation stage which moved the sample in the x and y directions over the microscope objective. Doing this, we were able to form an image of our quantum dot samples using a raster scan method. The raster scan image takes over a minute to complete so the scan is not only a spatial scan, it is also a temporal scan. Figure 7 is an example of a raster scan of the quantum dot sample. It is clear that the sample changes over time since we can see that some dots seem to dissappear in some rows. This is due to a phenomenon known as blinking which will be discussed later in this report. 5

7 Figure 7: Screen capture of quantum dot raster scan With the raster scan completed, we were able to move the Piezo stage to a specific spot on the sample and collect single photons from a single quantum dot. To do so, we used LabView software which controlled the Piezo stage to move a single emitter over the microscope objective. We then began collecting time differences from the TimeHarp card. Figure 8 show the data from two trials. Figure 8a shows collection from a quatum dot that appears to have died before we could collect data. The low amounts of photons and the uniform distribution over time do not show any signs of anti-bunching. This may have been the result of a phenomenon known as bleaching which will be dicussed later in the report. Figure 8b has a dip at approximately t = 52ns. Very few photons arrived for this value of t and this is approximately the delay time that we measured. The absence of photons around t = 52ns suggests that the photons are antibunched and our sample is a sample of true single emitters. (a) Flat curve that does not demonstrate antibunching (b) V curve demonstrating antibunching around t = 52ns Figure 8: Screen captures of antibunching curves 6

8 5 Atomic Force Microscopy of Sample Surface Once the samples were prepared they were examined with an atomic force microscope. The resulting images allowed us to study how evenly and densely the single emitters were distributed over the substrate. In its simplest form, an atomic force microscope works by dragging a needle with an atomically sharp point along the surface of the sample. As the needle moves along the sample, it moves up and down, maintaining a constant force between the tip and the sample. This change is measured by bouncing a laser beam off of the top of the needle and recording its deflection with a photo sensor. Hookes law then allows us to calculate the force on the needle to infer its distance from the sample. More complicated modes vibrate the tip of the needle near its resonance frequency as it moves over the sample. When the needle is closer to the surface, the forces of the atoms in the sample dampen the tips oscillation. This change in frequency can be used to infer the height of the tip above the sample. We utilized a combination of these techniques, known as tapping mode. The tip is vibrated above the sample with relatively high amplitude such that the tip taps the surface. Changes in the inter-atomic forces felt by the needle at different locations on the sample can be used to infer its height. This method has the advantage of minimally disturbing the sample, so that we can be sure not to alter the position of the single emitters on the slide while performing the measurement. Figure 9: Diagram of AFM (Image from Wikipedia) We first tested the AFM by examining the surface of a CD-ROM (Figure 10a) and a calibration sample (Figure 10c). We then observed a sample of quantum dots, zooming in to image a single dot (Figure 11c). The spot size of our laser when passed through our microscope objective can be calculated as: x = 0.61 λ = 240nm (1) NA We can see from the image in Figure 11a that the single emitters on the sample are sufficiently separated to be targeted individually by a beam of this size. 7

9 (a) CD-ROM Surface. (b) Calibration Sample Surface. Figure 10: Test samples (c) Test sample. (a) Initial image of sample. (b) Zoomed in on smaller area. (c) Single dot. Figure 11: AFM images of a quantum dot sample 6 Measuring Fluorescence Lifetimes Figure 12: Wiring diagram for fluorescence lifetime measurement In order to measure the fluorescence lifetime of single emitters, we first performed a raster scan with the confocal microscope to locate a single emitter to examine. Once we found an emitter, we positioned it in the focus of the confocal microscope with the translation stage.we then excited the single emitter with a picosecond scale laser pulse. The resulting fluorescence photon was captured by an avalanche photo diode (APD), sending a TTL logic pulse (Figure 4) to start a timer on a TimeHarp computer DAQ card. When the next laser pulse was emitted, the laser sent another pulse (Figure 13) to the computer, stopping the timer. 8

10 Figure 13: Pulse from laser Over many cycles, a histogram of the time between the laser pulses and the photon detections was created. These curves (Figure 15) show that the probability of a fluorescence photon being emitted decays exponentially. We measured the time constant (the fluorescence lifetime) of the sample by fitting these curves with decaying exponential function in IgorPro. Data was collected for both quantum dots in toluene and nanodiamonds in chiral nematic liquid crystals at different sites and with varying pump beam intensitys. (a) Site 2 (Poor data) (b) Site 3 (c) Site 4 Figure 14: Quantum Dot Fluorescence Images from Confocal Microscope Three quantum dot sites (Fig. 14a, Fig. 14b, and Fig. 14c) were tested. Two sites (Fig. 14b and Fig. 14c) demonstrated exponential decay, with lifetimes of about 4ns (Fig. 15b and Fig. 15c). However, the first site exhibited a linear decay (see Fig. 15a). Since the sample in question was taken on the very edge of a quantum dot cluster (see Fig. 14a), it is possible that no quantum dots were actually within the sampling region. Site 2 Site 3 Site µW Bad data µW Bad data Table 1: Quantum Dot Fluorescence Lifetimes (ns) Five nanodiamond sites were examined (Figure 16). When fit with an exponential curve, all exhibited lifetimes ranging from 3ns to 3.5ns. 9

11 (a) Site 2 (Poor data) (b) Site 4, 76.5µW (c) Site 4, 35.2µW Figure 15: Quantum Dot Fluorescence Curves (a) Site 1 (b) Site 2 (c) Site 3 (d) Site 4 (e) Site 5 Figure 16: Nanodiamond Sample Sites 76.5µW 35.2µW Site Site Site Site Site Table 2: Nanodiamond Fluorescenec Lifetimes (ns) In all cases, data was taken with an excitation beam intensity of 76.5 µw, and with an attenuated beam of (0.46) 76.5µW = 35.2µW. For all the samples tested, attenuating the pump laser by 0.46x increased the fluorescence lifetime by approximately 10 percent (see tables 1 and 2). 10

12 (a) Site 1, 76.5µW (b) Site 1, 35.2µW Figure 17: Nanodiamond Fluorescence Curves 7 Spectra of Single Emitter Samples In order to further characterize our single emitter sample, we attempted to collect its spectrum. For this analysis to be conclusive, more investigation is required but unfortunately we were unable to collect the appropriate measurments due to time constraints. From the viewing port on the confocal microscope, it is possible to see the interference patern of the excitation laser. Figure 18 is an image taken with a cooled CCD camera of the interference patern due to abberations and internal reflection. We were able to focus the confocal microscope by hand using this image. The microscope was in focus when the image was most crisp. Figure 18: Inverted image of the laser through the confocal microscope With the laser focused, we switched the port of the confocal microscope to send the image into the spectrometer. Using a difraction grating spectrometer (which separates a light source into its individual wavelengths) the spectrum of the image was sent to an EM-CCD. Figure 19 shows the spectrum as observed by the CCD and the analyzed spectrum which was created by calibrated software that decomposes the CCD image into a spectrum. The spectral line of the excitation laser is clear to see at 531nm (actually 532nm). 11

13 (a) Inverted CCD image of laser spectrum (b) Spectrum of excitation laser Figure 19: 532nm spectral line of the excitation laser Placing the interference filters into the confocal microscope, we obtain the spectrum in Figure 20. The 532nm line is still visible but the other peaks represent the spectral lines of the sample. We cannot be certain which lines belong to the single emitter. To be certain, it would be necessary to capture the spectrum of a blank slide without the single emitter sample. Such a spectrum could be subtracted from a spectrum of the single emitters. This would reveal the spectral lines of the single emitter sample. If we had more time for this lab, we would have collected such a spectrum but unfortunately, we did not. (a) Inverted CCD image of single emitter spectrum (b) Spectrum of our single emitters Figure 20: Spectrum of single emitters 8 Disadvantages of Quantum Dots In Figure 7, we saw qunatum dots turn on and off in time. This phenomenon is known as blinking and is a random event where a single emitter swtiches to an off state where it does not emit photons. Eventually, the emitter will turn back on. This causes a problem because it prevents us from collecting data. Blinking can be seen in videos taken of the quantum dot sample. Figure 21 shows this by presenting two frames from the video, one second apart. Notice that new dots have appeared while some dots have disappeared. Another phenomenon can be seen in Figure 22 which shows a screen capture of intensity vs. 12

14 (a) t= 0 (b) t 1s Figure 21: Quantum dots blinking time from the LabView software used to control data collection. The plot shows the number of APD counts per millisecond. The phenomenon known as bleaching can be seen where the signal abruptly drops. Bleaching occurs because of the fluorescent nature of our sample. The laser used to excite the sample eventually destorys the fluorescent molecules and bleaches. Figure 22: Screen capture showing a bleached signal flow These phenomena are problematic for single photon sources because they limit the ability to create a constant, controlled beam of antibunched photons. These are problems that will need to be overcome if single photon sources, such as the quantum dots that we test here, are to be used for quantum cryptography in which controlled antibunched photon sources are a requirement. Blinking and bleaching were not observed with the nanodiamond sample. For this reason, nanodiamonds are a much better candidate for single emitters to be used in quantum cryptography. 9 Conclusion In this lab we succesfully created single emitter sources and, with the use of a confocal microscope, we created an antibunched photon source using quantum dots. By examining the time between consecutive photons that were emitted from an excited quantum dot, we successfully demonstrated antibunching of our sample (Figure 8b). An atomic force microscope was used to determine the physical size of the quantum dots which was found to be on the order of 200nm. The fluorescence lifetime of both samples was measured to be on the order of 4ns and 3ns for the quantum dot and nanodiamond samples respectively. We attempted to capture the spectrum of our samples but due to time constraints, this measurement was inconclusive. With respect to the application of quantum cryptography which has been discussed thouroughly in this course, it was found that nanodiamonds are superior candidates for efficient single emitters because quantum dots are unreliable due to the blinking and bleaching phenomena. Nick contributed to this report by writing the sample preparation, antibunching, spectra, 13

15 disadvantages, and conclusion sections of this paper. Peter contributed by writing the abstract, introduction, AFM, and fluorescence lifetime sections of this paper as well as calculating the fluorescent lifetimes of all of our samples. The images and figures were collected and compiled by both of us, and together we typeset the report in L A TEX. 10 References AFM Diagram: svg Confocal Microscope Diagram: block_diagram.svg Lab 3-4: Single Photon Source, by Svetlana Lukishova. workgroups/lukishova/quantumopticslab/homepage/opt253_labs_3_4_manual_08.pdf 14