How to approach photonics related research works?

Question

I am a student in sophomore year in a college pursuing electrical and electronics engineering. My college offers chance to major in any of the 3 sectors of this department which are:

1)Electronics and photonics 2)Communication and signal processing 3)Power.

I was fascinated on knowing applications of photonics and wanted to study more on this especially on bio-photonics. Moreover I'm bit interested in learning ML. I have already started ML from coursera and know basic neural networks as well. Although it is kind of stereotypical in our place that ML is mostly required in signal processing. But I was looking forward to any topic under which I could merge both ML and photonics. Is there any such topic?

Also I would like to know about a constructive pathway to learn photonics. I'm not much familiar with quantum mechanics either. What could be the course of action in this case so that I would be able to study about photonics and might involve in research works on photonics?

Thank you.

Answer 1

First and foremost, talk to your advisor. Second, try to get an internship at a company that employs electro-optical engineers.

You don't need to know much about quantum mechanics to use semiconductor photonics devices effectively -- it's mostly classical optics between the device and the rest of the world, electronics once you get onto the device, and then it's only quantum-mechanics on the device.

Electro-optical engineers are in short supply right now, and probably will be for the rest of the decade. I'd recommend that you look at what such people do and see if it overlaps with your interests. It'll be heavily math and physics based, with a need for a wide-based knowledge of both physics and electronics. It'll be hard, jobs will be few and you'll need to move to them when they open up. However, as long as EO engineers are in short supply you'll be essential and that'll translate into good pay and job security.

Answer 2

A general pathway could be:

1. Learn electronics and electromagnetism.
2. Learn fundamental optics, both ray- and wave-optics as well as optics applications such as waveguides and lasers.
3. Learn physics of semiconductors and semiconductor-based devices, general solid-state physics.
4. Some photonic effects, such as Purcell effect, may require some familiarity with quantum mechanics.
5. Yes, you need to be also comfortable with applying ML algorithms (classical and neural-network-based ones) to various problems.
6. At some point, you have to complete a project showcasing the knowledge of aspects of the above fields. An easier approach is to find a theoretical problem and solve it with ML. A more difficult one will be to find an experimental problem
   Here, you might want to consider projects involving Raspberry Pi or Arduino. The goal is to gain some experience and to enrich a CV.
7. Networking may be very useful here, perhaps, it is possible to connect with somebody who would facilitate the involvement with photonics and ML. There are quite a bit of research groups out there.

Remarks: In the word "Photonics", there is the reference to "electronics", implying that instead of electrons, photons (particles of light) are used for communication. Nowadays, a popular theme is communication using entangled photons / quantum light, so that, again, some familiarity with quantum mechanics could be beneficial.

Answer 3

Learning about Photonics

Since you are still in college, and your college offers photonics courses, those photonics courses are by far your best way to learn about photonics. For example, take a summer course of the prerequisites (electromagnetics, as others mentioned for example).

If you are most interested in finding out what the Applications of photonics are, to see how interested you are, try

1. Looking up the research web pages of Profs who teach the photonics courses, their web pages, articles and research papers will often start with possible applications. Those Profs may also have suggestions.
2. Check out some Photonics magazines, such as the following: Photonics Spectra, Optica, SPIE open-access (free) journals. Your university may have subscriptions to the non-open journals, so access through your univ. library or on-campus network (a very expensive and valuable service, I might add!).
3. See if your university has a student group such as IEEE Photonics Society, Optica Student Chapter, or SPIE Student Chapter. They may host talks, events and have magazines etc. for free.
4. If you are nearby, the free Expo at Photonics West in San Fransisco is a pretty amazing place to see many applications in-person, from laser cutting to holographic displays, and all the supporting companies that enable those apps. It is free, anyone can show up and go to the Expo (showcase of 1000 photonics-related companies), only the talks etc. cost money.

ML & Photonics

I believe there is some significant application of ML + photonics. Especially for photonic systems that acquire huge amounts of data, fast ML and analysis algorithms can be very important. Some applications I can think of off the top of my head are:

1. Multi-spectral imaging, where different colors (wavelengths) are used to identify materials. Sophisticated ML is used in industry to try to automatically identify materials quickly - for example to ID different plastics quickly by their optical spectrum, or ID good/bad fruits or underwatered farm fields.
2. In Biophotonics, about half the papers I see are on computational work, a lot of image analysis and automatic classifications, faster processing algorithms etc. For example, having an algorithm automatically classify some biological image as diseased or normal.
3. In other chemical identification applications (such as optical glucose monitoring, to name one), the algorithms to quickly ID chemicals using optical spectral signatures is important. I have seen a number of startup companies making optical devices to detect specific molecules (such as cell types, chemicals in blood, gasses like Methane), where the analysis after the optical detection is perhaps the more difficult part of the technology.
4. LIDAR is a huge application for photonics right now - another where large amounts of 3D spatial data need to be quickly analyzed, for example in self-driving cars or military/other apps.