Sciencemadness Discussion Board

Build your own Gas Chromatograph

Harristotle - 28-4-2013 at 08:29

Hi to all.
Here is a picture of my home-built gas chromatography setup.
<a href="http://www.screencast.com/t/waz6AdEPccsD">Arduino GC setup-labelled</a>

I have published further details, including circuit diagram and code on the arduino forums.
http://arduino.cc/forum/index.php/topic,163305.0.html

I hope this is of some interest to someone, and I thank the members of Science Madness for their interest, passion and stimulating posts, and all that I have learned here over the last few years.

Cheers,
Harristotle.

neptunium - 28-4-2013 at 18:07

this is a nice little set up....congratulations! do you have any result to share?
why GC ? have you seen the Raman spectrometer here?

http://www.sciencemadness.org/talk/viewthread.php?tid=23422&...

always a pleaseure to talk with other motivated home scientist!

Harristotle - 29-4-2013 at 03:29

Hi.
Results are awaiting a new power supply (I cooked the old one, a modified atx, and I don't know how).

Will post when acquired and cleaned up (and I get some more DCM to run).

Thanks for the link to the raman - it looks like a great project. I went GC for a couple of reasons - as an ex biochemist, I am very comfortable with chromatography. Also, once I add a Figaro 822 sensor, I will be able to measure a lot of other things. (It produces a voltage, arduinos measure voltages, easy to add). I also can see how to make a good FID, using a sealed speck of americium from a smoke detector and some nickel plates in a butane flame. I may or may not do this - while I am happy that the americium in gold film is stable, of so negligible activity that they are ok to dispose of in the trash, I probably can't think of much I would need it for.

Have fun now !

smaerd - 29-4-2013 at 12:55

How is the nichrome insulated from the copper tube?

Harristotle - 30-4-2013 at 08:33

Astute question!

I bought a necklace of glass/ceramic beads from the flea market, and then I bogged the whole lot into the copper with JB weld.
JB weld also was used to glue the fan onto the heating pipe. The heating pipe was ordinary 1/2 inch copper pipe used for plumbing, available at Bunnings in Australia. The horizontal part of the oven was a 40cm length of such pipe, with two copper end caps, drilled out so that the borosilicate tube could fit through with a bit of clearance (1mm) - I was too chicken, and not engineer enough to check out the expansion coefficients of copper and borosilicate, so I erred cautiously.

update: still waiting on new power supply, will post chromatographs later (didn't save my first 2 trials and am now regretting it).

smaerd - 30-4-2013 at 10:06

Thank you. Now I understand and appreciate your design more fully. Are you using any column packing for running your trials? As that seems like quite a wide diameter column.

radagast - 30-4-2013 at 18:50

This is terrific. Looking forward to seeing those chromatograms and more details!

Harristotle - 3-5-2013 at 06:30

Well, an interesting week. I cooked a power supply and an arduino. My explanation is that I have a polished concrete floor, which is slightly damp due to winter. My power supply is double insulated, but not earthed. As I touched the floor in my bare feet and the circuitry as I adjusted it, I supplied a ground path which killed the arduino and the power supply. Since replaced, and now all working.

A picture of a chromatogram made in openoffice of dichloromethane, using silica gel as the matrix in the column can be seen here:
http://screencast.com/t/i5y7MFSh3MG
The run was conducted by injecting 0.5 ml of the vapour over a bottle of dichloromethane, into a silicone tubing fed by an aquarium pump. To make things consistent, a texta mark was made on the silicone so that both runs had the same void volume.

This figure shows that DCM sticks to the silica gel at low temperatures, but at higher temperatures (55 degrees C) it is flushed off. Note the tail or bleed on the chromatograph suggesting significant column interaction even at 60 degrees: if I were to do this again I would use a higher column temperature. Sadly, this is the end of my holidays, so the whole shebang will have to go into a box in the shed.

A few words on the construction: haloalkanes make beautiful colours when burnt over a copper wire, and you need exceptionally small amounts to produce a spectacular light. The sensor is just a light sensitive resistor, covered with enough heatshrink tubing to make it very directional and have a small angular opening. I did this because it is fast and cheap, and it generates a voltage signal which the a/d on the arduino can measure. To realise it, I just adjusted the level of the column output so that it fed into the air inlet of my MAPP gas blowtorch.
You can see all of this here http://www.screencast.com/t/iPuce3g2R
hopefully this makes it all clear.

Haloalkanes don't really float my boat - my next big improvement will be to include a Figaro 822 tin oxide gas sensor, which will detect a huge range of organics.

Also, I will be on the lookout to make a flowrate monitor for the aquarium pump - there are at least a dozen improvements that can be done here. However, I don't feel too bad - I have spent over 2 years playing with this, on and off - hey, improvements take time!

Do enjoy, I hope this is enough detail for anyone who wants to have a go at one. I will probably write up an Instructable over the Christmas holidays.

Cheers,
Harristotle

Harristotle - 3-5-2013 at 06:39

Just one last point:

I would like to take some silica gel, hydroxylate the surface of it by rinsing in H2SO4/H2O2, and then add a fatty acid to it, say Octadecanoic acid, and esterify the two together with an H2SO4 catalyst. Voila, C18 column. Much more interesting!

I reckon there is a whole load of good chemistry to be done here!

Also, didn't mention properly in last post - the silica gel was in sphere form 2-3mm, from a dessicant pack. Not a huge number of "plates" in that column - so there is plenty of room to play and to optimise!

but alas, now I must hibernate in my work cave for the next few months!

Cheers,
H.

smaerd - 3-5-2013 at 12:05

Well I would really like to thank you for sharing your work and being open about questions. Don't feel too bad about frying your PSU and arduino, all of us who've tried to build apparatii(is that the right plural) have done similar things hehee. I'm really impressed with the results even with suboptimal conditions. Really looking forward to you getting more time to doing this.

radagast - 21-5-2013 at 10:00

Quote: Originally posted by Harristotle  
Just one last point:

I would like to take some silica gel, hydroxylate the surface of it by rinsing in H2SO4/H2O2, and then add a fatty acid to it, say Octadecanoic acid, and esterify the two together with an H2SO4 catalyst. Voila, C18 column. Much more interesting!

I reckon there is a whole load of good chemistry to be done here!

Also, didn't mention properly in last post - the silica gel was in sphere form 2-3mm, from a dessicant pack. Not a huge number of "plates" in that column - so there is plenty of room to play and to optimise!

but alas, now I must hibernate in my work cave for the next few months!

Cheers,
H.


That's amazing, Harristotle. I can't believe you were able to get a usable chromatogram out of purely DIY components. I ordered an Arduino and time-permitting, will try to build a GC around your design.

Harristotle - 22-5-2013 at 03:18

Well that is funny, radagast.

Because I have just bought a 540nm lowpass filter (also called an OG54 fitler by photographers who like to photograph skies, and costing $12), a 2048 pixel linear ccd TCD1201 for $1 !!!, and the biggest -ever- lego baseplate 40cmx40cm !
(some time next year I'll even get to build it!)

Good luck with my GC, and keep contributing to the knowledge that is available to amateurs. Not since the 70s and CL Stong has the amateur science scene been so vibrant. I have the feeling that I will look back in this time in my old age, and be amazed and grateful to have lived through it.

Harristotle

Mildronate - 22-5-2013 at 22:39

Better get some old HP GC

sonogashira - 23-5-2013 at 12:31

What gas are you using? You will get a much greater separation if you use a liquid stationary phase. There are good descriptions of detector-designs (including simple but very sensitive ones made from light-bulb filaments etc.) in Clifton Meloan's book "Chemical Separations". I can scan some pages if it would help. Without a liquid phase, TiO2 would be a better solid phase than silica since it can be packed much more densely. I have seen it used without liquid coating for preparative GSC in industry, where a high temperature was needed. (This was an 8-metre-long tube, and the "separation" was actually still only a 'collection of fractions' when analyzed further, but this is what was desired).

It may be easier to buy a column from ebay and attach it to a home-made detector? If your area of interest is electronics (judging from your diagrams) then it will be possible to get some very accurate measurements from even a simple detector design. With your system at the moment I doubt that you will get very good separation though. Commercial analytical columns can be found for little expense, but the same cannot be said of sensitive detectors, which are nevertheless easy to make, and nanogram detection is possible for little expense (but the same could be said of TLC, and the expense is even less).

What do you intend to separate?

[Edited on 23-5-2013 by sonogashira]

smaerd - 23-5-2013 at 13:39

sonagashira I see you are referring to a thermal conductivity detector. If you could post scans that include circuit diagrams for such please do.

Sonagashira the only real technical issue I see with using a commercial capillary column is the sample loading size. One would likely need to devise a split-injection system. Not to say it's impossible, in fact it sounds very possible.

Mildronate - 23-5-2013 at 21:48

You need column, injector and detector from ebay all other can DIY.

Harristotle - 24-5-2013 at 01:43

Do you really think so Mildronate?

injector=septum(replaceable), port and microswitch.
Push your own start button and use a piece of silicone (same stuff as septum) tubing with a marked spot.

Column: who made columns before ebay? labs = us !

detector: FID detectors are simply a flame, a radiation source, a biasing voltage and a current meter, well within arduino reach. I have made crude FID out of a candle, a milo tin, and a high gain dralington transistor. This simple photodetector works well for haologenated substances. The fact that Vernier make a school GC for under $2000 should alert you to the fact that organic sensors (eg tin oxide based Figaro 822 type) are cheap and readily available.

@ sonogashira: the gas is simply air, from an aquarium pump.

@smaerd check out The Bell jar for circuits for TC sensors - there is really not much more to them than an opamp, a thermistor and a heating resistor


I have attached some references to stimulate discussion on this


Have fun now!

Harristotle


Making a halogenated alkane GC in schools
http://www.pathfinderscience.net/teachers/flask/act/pdf/chap...

Themal conductivity gauge : http://www.belljar.net/tcgauge.htm


822 detector for alcohol and acetone
Home-made Detection Device for a Mixture of Ethanol and
Acetone. Sensors, 2007, 7, 202-213

sonogashira - 25-5-2013 at 05:40

Quote: Originally posted by Harristotle  
Column: who made columns before ebay? labs = us !

They did, but they made them 6 feet long:
http://www.ebay.com/itm/TekLab-Glass-Column-for-Gas-Chromato...
http://www.ebay.com/itm/2-Supelco-Custom-Glass-Column-for-Ga...
If you expect any chromatographic separation from your 40 cm column without a liquid stationary phase then I wish you luck.
smaerd, there aren't detailed circuit diagrams in that book, but I have a collection of papers which I can send to you when I have time to scan them (perhaps a few weeks time).


[Edited on 25-5-2013 by sonogashira]

radagast - 27-5-2013 at 20:20

Quote: Originally posted by Harristotle  
Well that is funny, radagast.

Because I have just bought a 540nm lowpass filter (also called an OG54 fitler by photographers who like to photograph skies, and costing $12), a 2048 pixel linear ccd TCD1201 for $1 !!!, and the biggest -ever- lego baseplate 40cmx40cm !
(some time next year I'll even get to build it!)

Good luck with my GC, and keep contributing to the knowledge that is available to amateurs. Not since the 70s and CL Stong has the amateur science scene been so vibrant. I have the feeling that I will look back in this time in my old age, and be amazed and grateful to have lived through it.

Harristotle


Go Team GC/Raman!! (Is that what's the next step for us?) I'm really excited that you and other people are extending the raman spectrometer (especially since you know much more about electronics than I do -- maybe you can get a high-resolution spectra out of a CCD), and I'll definitely keep you posted on my progress on your GC design.

dubfactor - 18-12-2013 at 11:00

Did this stop back in June? I just came up w/ the same idea (homemade GC) ans was excited to find to site/page/topic!

What happened?!? I love this!

GC and Raman

Harristotle - 19-12-2013 at 05:05

The last teaching week for the year ended last friday.
Both the GC and the Raman are in my plans for resurrection over the next few weeks. The Raman will take longer, my lowpass filter is no good, I will need to get another.


I am happy to share what I have so far, ask and ye shall recieve!

Cheers,
H.