The overall goal of this procedure is to fabricate and characterize a photonic crystal slow light wave guide or cavity. This is accomplished by first preparing the silicon on insulator chip for electron beam lithography by cleaning and spin coating the chip with ZEP five 20 a resist. The second step is to expose and develop the photonic crystal pattern using electron beam lithography for pattern definition and xylene for development.
Next, transfer the photonic crystal pattern into the top silicon layer of the chip by plasma etching the sample in the reactive ion etcher. The final step is to remove the buried oxide layer by etching away the silica in hydrofluoric acid. This creates membrane photonic crystal devices, in other words, devices with air both above and below the silicon photonic crystal layer.
Ultimately, the device is optically characterized if the sample contains photonic crystal wave guides amongst under interferometer based slow light experiment is performed if the sample consists of photonic crystal cavities. Instead, a resonance scattering technique is used for characterization. The goal of our work is to study light matter interaction in photonic nanostructures and to make novel devices and explore interesting concepts in such structures.
So in order to concentrate and be able to really focus on the physics of of this work, we need a reliable fabrication protocol. We have developed a protocol based on silicon on insulator, which we find very reliable and convenient, and by making this protocol available to a wider range of users and other researchers, we hope to stimulate more activity in this area. As for the characterization of our slow life of Tory crystal wave guides, our technique stands out for its simplicity and power as it requires no on-chip interferometric components or movable Parts.
The main advantage of this resonance scattering technique over existing methods is that it allows a characterization of photonic crystal cavities and an out of plane arrangement, which would usually require an internal light source To start clean a silicon on insulator wafer by placing it in an ultrasonic bath of acetone for one to two minutes. Then transfer the sample into isopropanol for 30 seconds to remove any residual acetone and dry it using a clean, dry nitrogen gun. Next, place the wafer into a spin coder and pipette on enough CEP 5 28 Photoresist to completely cover the sample.
Spin the sample at 3, 200 RPM for 60 seconds to produce a 350 nanometer thick film. Remove the wafer from the spin coder and bake on a hot plate at 180 degrees Celsius for 10 minutes. Once the chip is cooled, place it into the pattern exposure chamber of the electron beam lithography system and pump down when full vacuum has been achieved.
Set the acceleration voltage to 30 kilovolts and leave the system to equilibrate for one hour. Then expose the sample with an area dose of 55 micro ampers per centimeter squared. Using a step size of two nanometers.
Develop the sample by placing it into xylene at 23 degrees Celsius for 45 seconds. Then rinse away residual xylene with isopropanol. After cleaning the reactive ion etching chamber, load the sample and pump down the system to less than three times 10 to the negative six.
Millibar Then precondition the chamber by flowing in Fluor and sulfur hexa fluoride in a one-to-one ratio at 100 SECM, and allow the gases to flow for at least 10 minutes. Next, set the RF power to 20 watts and ignite a plasma. Etch the sample for approximately two minutes while ensuring that a chamber pressure of five times 10 to the negative two milli bar is maintained.
Then remove the sample and clean it by rinsing. In 1165, remove with ultrasonic agitation for one to two minutes, followed by acetone and isopropanol. To begin membrane isolation spin coat the sample with UV sensitive photoresist micros S 1818 G two.
Using an appropriate photo mask. Define windows within the resist above the photonic crystal patterns using the UV mask aligner. Expose the sample for approximately 30 to 45 seconds.
Develop the photo resist in micro posit developer MF 3 1 9 for 30 to 45 seconds. Rinsing afterwards and deionized water. Next, etch the sample in one part hydrofluoric acid to five parts deionized water for 15 minutes.
After etching, rinse the sample thoroughly and deionized water. Remove the remaining photoresist using acetone for five minutes or until the sample appears clean. Then rinse, isopropanol, and blow dry with nitrogen.
Next, clean the sample in a piana bath of three part sulfuric acid to one part hydrogen peroxide for five minutes. Then rinse the sample in deionized water acetone and isopropanol. Finally, cleve the sample by making small scratches on each side of the chip lining the scratches up with the edge of a glass slide and pressure to cleave along the scratch lines.
In order to measure transmission and group index curves of slow light photonic crystal wave guides, set up a mock zender interferometric detection system shown here as described in detail in the accompanying protocol. Next, insert the chip into place and couple light into a blank ridge waveguide. Adjust the delay stage until the reference arm length is shorter than the sample arm.
Continue to adjust the delay stage and observe the transmitted power through a continuous skin of the optical spectrum analyzer. Adjust until four to 10 fringes are observed in a 10 nanometer wavelength range. Once adjusted, keep the delay line fixed throughout the measurements.
Next, calibrate the setup by running three scans of the blank wave guide on the optical spectrum analyzer using a resolution of 0.05 to 0.1 nanometers. One scan for the interference spectrum and one scan for each of the two arms separately. Record each measured spectrum.
Then run the same three scans on each photonic crystal wave guide on the chip. Finally, calculate the transmission curve by normalizing the sample spectrum of a photonic crystal wave guide to that at the blank wave guide. Also, calculate the group index curve from the interference spectra as described in the protocol.
Begin by mounting the sample vertically with a 45 degree orientation to the axis of the polarizer on a differential driven XY, Z micro block, and adjust the micro block so that the sample is in focus and a cavity can be seen with the camera using an amplified spontaneous emission source. Align the beam with the center of the cavity. Next, start a broad scan with a load to moderate resolution.
In order to identify the cavity peaks, obtain the course wavelength of the resonance in the amplified spontaneous emission scan with an accuracy of plus or minus one nanometer. Once the cavity peaks have been identified, perform high resolution scans with the tunable laser source attenuated to micro watt levels. Scan with a resolution of one pico meter for a two nanometer range centered at the previously found resonance wavelength.
Change the XY, Z position of the micro block and rerun the scan until the signal to noise ratio is maximized and the line shape is close to that of a lian shown here. Successful etching with the reactive ion etching system produces a sample with vertical sidewalls and no widening of the holes at either surface of the silicon. Pressure, power, and time settings have all been optimized for ideal structure creation.
When the reactive ion etching pressure is above optimum, the increase in pressure causes an angle in the photonic crystal wall. Increases in the RF power can also cause irregularities in the crystal, such as the widening of the photonic crystal holes shown on the right. Too much exposure time causes a combination of effects due to the breakdown of photoresist resulting in widening of the photonic crystal holes and angled sidewalls.
A blank waveguide results in uniformly distributed fringes over the entire wavelength range. When the blank is replaced with an 80 micrometer long engineered slow light photonic crystal wave guide, the fringes become denser at wavelengths above 1, 575 nanometers marking the transition from the fast to the slow light regime. This method provides measured group indices in excessive 100 for an 80 micrometer long wave guide, and almost 90 for a 300 micrometer long guide.
High resolution scans with a T Noble laser source will ideally produce a Lian line shape for accurate Q factor analysis of 43, 855 at a wavelength of around 1, 562 nanometers. The protocol we've just described has been optimized for photonic crystal wave guides and cavities in the important telecommunications wavelength range of 1550 nanometers. Clearly, we can also make other types of wave guides that are equally demanding.
We can also do it at different wavelengths, and by providing all this detail to the community, we hope that other people will be inspired to also try this out in different materials and for other applications. With our measurement technique, based on spectral interferometry and foer transform analysis, we're able to measure group indices in excess of 100 with a very Simple optical setup. The cavity characterization technique allows a fast, straightforward, and non-intrusive characterization of the spectral properties of photonic crystal nano cavities, such as the determination of the quality factor of the resonant mos.
