Sciencemadness Discussion Board
Not logged in [Login ]
Go To Bottom

Printable Version  
Author: Subject: Open-source Raman spectrometer looking for academic partners
Luc
Harmless
*




Posts: 4
Registered: 23-12-2017
Member Is Offline

Mood: No Mood

[*] posted on 25-4-2020 at 03:39
Open-source Raman spectrometer looking for academic partners


Hi all,

Before I start, I know that there has been an extensive post about DIY Raman spectroscopy in the past but I decided to create a new post because what I would like to discuss here is not related to the instrument itself but on a different matter that I will detail below just after a short intro.

Long things short, I spent the last few years working almost exclusively on developing a DIY Raman spectrometer and got very nice results that I decided to share as an open-source & modular platform. I opened a website and very recently a first fully functional version.

Here is a picture of the setup:

starteredition.jpg - 82kB

All the assembly instructions & CAD files to create copies are detailed on the website : http://www.open-raman.org/build/

I have foreseen two versions at the moment that I call "Starter Edition" and "Performance Edition". Only the Starter Edition is available at the moment. It is essentially designed for teaching applications as it features a low power laser and a reduced resolution (<35 cm-1 tested). The Performance Edition will come shortly and has a 12 cm-1 tested resolution but it is still too expensive and I'm trying to take the cost down. This last version does work but it's about ~4500€ at the moment and I would like to make it go below 3000€. The Starter/EDU version is about 2500€. Please note that I am not selling these instruments and am NOT making any money on this; people who would like to create their own version have to download the BOM and place orders themselves at the various suppliers.

Here are typical spectra obtained with iso-propanol:

i-PrOH.jpg - 58kB

There are more spectra here for those who are interested and you can even download them in .csv: http://www.open-raman.org/performances/

I have already performed the following reactions that you can read on the website:
- Quantitative analysis of MeOH in EtOH with a detection threshold of 0.6%
- Tracking of Ethyl Acetate kinetics over 35 hours where I evaluated every minute the concentration of ethyl acetate, ethanol, and acetic acid to find the kinetic equation of the reaction
- Proof of the addition of the bisulfite ion to formaldehyde; the aim was to show that Raman spectroscopy can unambiguously show that a reaction took place when your eyes see nothing

There are many more reactions I'm still working on but you know what it is...

So I'm writing this post because the thing works great and I'm confident that it can find applications in Universities so I would like to find academics who would be excited in a partnership. Non-commercial applications only, I'm looking for people who like science and would like to share something with others.

I'm looking for both partners for teaching applications and for research applications although I'm currently focusing on the teaching applications first since the Performance Edition is not available yet.

The idea is that the academics would develop a concept such as a lab session for their students and work that with them. They would own 100% of the application and I will not interfere with what they are doing. I will provide support in assembling and operating the instrument to my best as a trained professional optical engineer with a strong background in chemistry. In exchange, I would like to advertise on my website that "XYZ University" is using the spectrometer for "ABC" type of activity and provide a link to the webpage of the teacher/lab facility. If a paper is issued, I would like to be credited (either as an author or just cited in the references depending on the amount of work done -- it is obvious that I'm not asking to be a co-author if I just exchanged a few e-mails but that I would be if I start procuring custom CAD files to extend the instrument for that particular application; but this is common sense I guess).

Since the idea is to create an "open" framework, I would strongly encourage people to share notes if they create a lab that uses the instrument (like protocols, what teachers should focus on etc.) but it is in no way compulsory. Again, academics remain 100% owners of what they produce and I'm not interefering with that.

The spectrometer can be used with optics or chemistry students as well. It can be a low-cost alternative to expensive lab hardware (nobody like students to "accidentally" drop conc HCl on a $20,000 spectrometer), a possibility for the students to align the spectrometer themselves or even let them develop custom parts for the spectrometer.

This last possibility (upgrade) is the "modular" part of the design. By default the spectrometer comes naked and you have to add a "cuvette" to it to analyze a certain type of sample. By default I propose what I call the "standard cuvette" which allows to test liquids in test tubes. But one can imagine custom cuvettes to study solids, liquids flow of reaction, and so on! You just plug off the type of cuvette and replace it with the new one :) So I strongly encourage academics to design their own custom cuvettes and I will provide all the info to help them do that.

So if there are people here who are interested in the idea, just reply to the post or contact me so that we see together how we can partner. Again, I would like to stress that this is not like a "prototype version" or anything else, it's working for real and it's complete. I'm ready to offer support to my best to anybody who would like to use the spectrometer for his own applications :)

Also, even though it is not mentioned so far, if you are just an enthusiast and would like to do Raman at home for youtube videos or anything else I will gladly help you too :)
View user's profile View All Posts By User
Metacelsus
International Hazard
*****




Posts: 2394
Registered: 26-12-2012
Location: Cambridge, MA
Member Is Offline

Mood: Double, double, toil and trouble

[*] posted on 25-4-2020 at 05:24


Very cool!

I noticed the most expensive item on the Bill of Materials was the camera. In high school I built a Raman spectroscope using this guide: https://hackaday.io/project/1279-ramanpi-raman-spectrometer

which used a linear CCD array. The CCD array (TCD1304DG) was much cheaper and performed reasonably well. Do you think it would be possible to substitute this for the camera?




As below, so above.
View user's profile View All Posts By User
Luc
Harmless
*




Posts: 4
Registered: 23-12-2017
Member Is Offline

Mood: No Mood

[*] posted on 25-4-2020 at 06:20


Quote: Originally posted by Metacelsus  
Very cool!

I noticed the most expensive item on the Bill of Materials was the camera. In high school I built a Raman spectroscope using this guide: https://hackaday.io/project/1279-ramanpi-raman-spectrometer

which used a linear CCD array. The CCD array (TCD1304DG) was much cheaper and performed reasonably well. Do you think it would be possible to substitute this for the camera?


No, unfortunately it is not possible.

These CCD arrays are too noisy and require you to build custom ultra-low-noise read-out electronics which are extremely complex to design. It is true that professional Raman spectrometers use linear CCD array but these (CCDs) are in the 1-2k€ price range without the readout electronic (typically TEC cooled expensive Hamamatsu sensors). The spectrometer that uses cheap sensor like the ILX511 in the USB2000 spectrometer from OceanOptics do not work in Raman applications due to their high noise level.

When I designed the spectrometer, I knew that the noise of the sensor would be a critical feature (as well as light collection efficiency) so I made an educated choice for the sensor. I chose to use CMOS machine vision camera because these sensors are produced at high volumes and therefore achieves very good "bang 4 bucks" ratio *plus* you get the electronics enclosed and the SW to operate them. To give you an idea, I tested the FLIR camera I'm using on a test bench at the office and I experimentally confirmed that the sensor was able to detect as low as 4 photons! This is a truly impressive performance for less than 400€ (all-included).

Experimentally, I'm reaching SNR on the order of 1:1000 for only 3-5 seconds exposures for common lab solvents so the spectra are really clean even with short exposures snapshots. Even at 10 Hz / 100 ms exposure you get good real-time Raman spectra.

I also wrote an article about how to choose the best camera sensor for spectroscopy based on decision-aid methods. You can check it here: http://www.thepulsar.be/article/choosing-the-best-camera-sen...
View user's profile View All Posts By User
earpain
Hazard to Self
**




Posts: 53
Registered: 11-9-2019
Member Is Offline


[*] posted on 4-5-2020 at 12:02


Quote: Originally posted by Luc  
Quote: Originally posted by Metacelsus  
Very cool!

I noticed the most expensive item on the Bill of Materials was the camera. In high school I built a Raman spectroscope using this guide: https://hackaday.io/project/1279-ramanpi-raman-spectrometer

which used a linear CCD array. The CCD array (TCD1304DG) was much cheaper and performed reasonably well. Do you think it would be possible to substitute this for the camera?


No, unfortunately it is not possible.

These CCD arrays are too noisy and require you to build custom ultra-low-noise read-out electronics which are extremely complex to design. It is true that professional Raman spectrometers use linear CCD array but these (CCDs) are in the 1-2k€ price range without the readout electronic (typically TEC cooled expensive Hamamatsu sensors). The spectrometer that uses cheap sensor like the ILX511 in the USB2000 spectrometer from OceanOptics do not work in Raman applications due to their high noise level.

When I designed the spectrometer, I knew that the noise of the sensor would be a critical feature (as well as light collection efficiency) so I made an educated choice for the sensor. I chose to use CMOS machine vision camera because these sensors are produced at high volumes and therefore achieves very good "bang 4 bucks" ratio *plus* you get the electronics enclosed and the SW to operate them. To give you an idea, I tested the FLIR camera I'm using on a test bench at the office and I experimentally confirmed that the sensor was able to detect as low as 4 photons! This is a truly impressive performance for less than 400€ (all-included).

Experimentally, I'm reaching SNR on the order of 1:1000 for only 3-5 seconds exposures for common lab solvents so the spectra are really clean even with short exposures snapshots. Even at 10 Hz / 100 ms exposure you get good real-time Raman spectra.

I also wrote an article about how to choose the best camera sensor for spectroscopy based on decision-aid methods. You can check it here: http://www.thepulsar.be/article/choosing-the-best-camera-sen...


I salute your work! My background is in Information Technology and I am open-source to the bone. I see that you are offering CAD files and source code, you are the real thing.

I am no where near your level so I doubt any collaboration would be of interest.
But I do plan to build a spectrophotometer via a Raspberry PI NOIR camera.
I'm pretty sure this would not even be in the same universe of resolution and precision as your device, but people have had impressive results. Especially considering the cost of the parts is ~$50

Initially I was inspired by this professor teaching first year Chemistry, and authoring a lesson which many other teachers are doing: He uses an iphone to determine the exact concentration of sugar in kool-aid:
http://faculty.ccbcmd.edu/~cyau/124%2003%20Concentration%20o...

Initial talks of PI Noir: https://www.raspberrypi.org/forums/viewtopic.php?t=40614

@Luc you probably know all this already, and I didn't mean to derail your very specific request, but just wanted to share my excitement about your project, and similar projects i've seen. And I believe I can get the pi noir device to work and hopefully will be posting about it soon.
Thank you!
View user's profile View All Posts By User
andy1988
Hazard to Others
***




Posts: 111
Registered: 11-2-2018
Location: NW Americus ([i]in re[/i] Amerigo Vespucci)
Member Is Offline

Mood: No Mood

[*] posted on 4-5-2020 at 14:07


Here is possible application if it can be packaged in a hand held form factor, with this research group:
https://phys.org/news/2019-11-quickly-accurately-scan-nutrie...
https://kurouskilab.com/

Their work brings to mind Gregor Mendel, the "Father of Genetics".




View user's profile View All Posts By User

  Go To Top