4 Setup

This section describes the setup with which I performed my experiments. I thought it would be helpful to split this up into three sections. First, a part about signal generation: how is the light shining onto the sample generated? Second: what devices help getting the best results and where are they in relation to the sample? Lastly, there is a section about how the output is measured and interpreted.

4.1 Generating Input

It is desirable to be able to excite the nanocrystals with a variety of wavelengths and at a variety of beam fluxes. In order to do this, half the setup consists of a box dedicated to the generation of photons. This box consists of four important parts: a source laser, OPO’s, frequency doublers and polarization filters. For extra convenience, it outputs its light through a single hole, in a single direction. In this box a lot of components are based on nonlinear optical materials1. These materials are a science in themselves, but the end result is that photons can be split and merged in arbitrary ratios and quantities.

The laser is an off the shelve Nd:YAG laser, which lases at 1065nm. It has a nonlinear optical crystal that triples the energy to 354nm. This process is known as third harmonic generation. I’ve used it in pulsed mode (100Hz), because sustained high fluxes can damage the optics in the rest of the setup.

Secondly the box has two OPO’s, or optical parametric oscillators. Very roughly, it consists of a nonlinear optical crystal and an optical resonator that can split a photon beam into two. The crystal ’cuts’ the beam into two other frequencies, depending on the beam incident angle. The optical resonator then deamplifies the original frequency, causing you to end up with only the two new frequencies. Rotating the crystal causes the exact output frequencies to change, allowing for precise selection of the desired frequency, exactly what we needed.

Because the crystals in the OPO’s can only cover so much in the frequency spectrum, there a also frequency doublers present in the box. They add up the energies of two photons. Again, a nonlinear optical crystal is used here.

Lastly, there’s a set of polarization filters. With these, the intensity of the output beam can be tuned.

4.2 The Sample

Now that we have a hole that outputs a photon beam of a frequency and intensity we choose, it’s time to discuss how the beam interacts with the sample. The sample is the collection of nanocrystals. They’re prepared by sputtering SiOx, with x < 2, onto a slide of quartz, the substrate. This takes place in a vacuum to prevent the silicon from oxidizing. Then, the slide is inserted into an oven and is then ’baked’, or annealed. The sputtered SiOx will settle, and depending on x, Si will clump together and integrate into the surrounding Si and O, which forms an SiO2 matrix. The nanocrystal size, spacing and shape depend on x, the length and temperature of the baking process. How exactly is a matter of chemistry, and very difficult to model.

Nonetheless, the sample looks like a slice a glass, typically 10x5x1mm. This, and the fact that the distribution of Si NCs in the slice is not uniform, causes the photons coming off of the sample to have strong anisotropic behavior. To compensate, there is an instrument called the integrating sphere6. This is simply a sphere with a highly reflective scattering surface on the inside. There are two holes to allow light to enter and exit the sphere. The sample is put on a frame, that suspends the sample roughly in the middle of the sphere. The effect of the sphere is that any anisotropies are smoothed out as light gets reflected. The two holes obviously constitute a loss of photons, but these losses are small (the holes are small) and isotropic (the ratio of photon energies are not changed by it). With this sphere, the measurement of an absolute quantum efficiency is possible.

4.3 Capturing Output

Converting a bundle of photons is done with either a charge-coupled device (CCD) or a photomultiplier tube (PMT). A fiber is attached to the integrating sphere, leading the light into a monochromator. This device is a box in which there’s a grating, that splits the photon beam up into it’s constituent frequencies. These frequencies are spatially spread over a certain angle. The PMT and CCD are attached to the monochromator, and via a mirror you can select which instrument to use. The CCD allows for the simultaneous measurement of the entire spectrum the grating reflects onto it, with a time resolution down to 13ms. The PMT can measure only one frequency, but with a much higher time resolution: down to 250ps. There are a few different gratings to choose from, which differ mostly in output window and operating frequency range.

This new setup, a laser box that can emit photons at higher energies (up to 6 eV, the bandgap of a NC is typically around 1.5 eV, see 12) and an integrating sphere that smooths out anisotropies, allows the research group to build graphs of the absolute quantum efficiency as a function of excitation wavelength. These images show if exiting the NCs at multiples of the bandgap energy actually results in quantum cutting. After all, quantum cutting would manifest itself as an increase in quantum efficiency.

Figure 6: An example of a measurement of the absolute Quantum Efficiency done with the new setup.