DIY GC-FID system

[ChristofferBR]

Hi! I'm new here! I'm a chemistry student from Denmark, and I'm really into scientific instrument/apparatus design, and I've gotten into my head that I want to build a GC-FID system on a student-friendly budget.

There are obviously a "few" obstacles, but the largest one has just been overcome: I now have a column! Drawing quartz capillary myself might be a bit outside the realm of possibility...

Anyway, I'm still getting the design together, but the biggest hurdle right now is column 'plumbing'.

I want to build a simple splitless injector, see image below, but the route I've chosen has the (maybe grave!) problem of having the first ~2-3 cm of column outside the oven. Will this mess things up completely?
Ideally, I'd get a piece of inactive/guard column, feed that through and join them inside the oven, but that adds cost I'd prefer to avoid.

Obviously one major issue is the fact that it's not heated.. But really, my first quest is just to get sample through a system, separation comes second.

I do see Agilent/HP injectors on ebay occasionally, so I'll replace cobbled-together stuff with proffesional spare parts when they come around cheaply.


The other issues that I've sortof figured out, but not made yet are:
-Oven from a kitchen 'mini oven' - with computer control/temp feedback
-FID built after 1960's early GC articles, output to pen plotter
-Carrier gas N or He, don't want a H2 cylinder in my home!
-FID Hydrogen supply perhaps from small electrolysis rig?

Thanks for your interest!

[Peter Apps]

It's good to see someone swimming against the everything automated and off the shelf tide.

Everything will be much easier if you use a packed column - the volume flow rates are higher and so easier to control, you do not need a splitter or even a vaporizing inlet because you can inject liquid directly onto the column, and the whole thing will be much more robust to contamination. The column can be made from 1/8 or 1/4 inch metal tubing and connections using ordinary stainless compression fittings. Your injection port can be a T-connector with a septum on one port, the column on the opposite one and the carrier gas coming in from the side. The septum is on the outside of the oven and the injection is made into the heated zone of the column. What packing to use ? You need solid, thermally stable inert particles with a narrow size range - you can buy packings ready made but if you want to go strictly DIY you could try something like swimming pool filter sand. I'm sure there will be better suggestions form other contributors.

Peter

[GOM]

Fascinating - I love it !

Perhaps better posting in the student section?

I once considered building an FID out of a Bunsen burner to demonstrate the principle

What equipment/budget do you have and what is your ultimate aim in terms of analysis and separating sample type?

As Peter said - consider a packed column

As for electrolysis - you may not be able to generate the amount of hydrogen required

Would you happy to have a nitrogen cylinder in your home?

With some thinking outside the box there may be a sample/ column system that may enable using just pumped air as a carrier gas at a low temperature for demonstration purposes

Same for the FID using an alternative fuel

As an academic exercise building a GC on a desert island I really like this - that is not meant in a derogatory sense

Kind regards

Ralph

[ChristofferBR]

Thanks for the replies and the interest! Yes, a packed column would be way less involved, and maybe I should've started with that, but I really want a capillary system - something that can take advantage of surplus HP/agilent etc. components.

something that could atleast come near-ish a professional instrument in separation power - if you keep swapping DIY components with surplus ones, at some point it'll be a complete non-diy system. Certainly an analytically operational system, is the goal.

I don't mind having an N2 or He cylinder - it removes the whole Hindenburg scenario. Think I'll look for a surplus welding N2 bottle. Can always get it filled with lab-grade nitrogen. Party balloon helium cylinders might be worthwhile looking into, if properly filtered (activated carbon or conc. sulphuric acid).

Electrolysis does pose a challenge. I'm actually considering using a bunsen burner for a very rough first FID - with a fine stainless capillary introduced through the air vent.

I don't really have a set budget - this is likely a project that'll run over some years, but suffice to say I don't really have the resources to just BUY a GC.

Flow control of carrier gas is another issue I'm looking into, and where I don't think anything but an off-of-ebay flow valve would do, but who knows?

[GOM]

Hi

Your comments

Yes, a packed column would be way less involved, and maybe I should've started with that, but I really want a capillary system - something that can take advantage of surplus HP/agilent etc. components.

Party balloon helium cylinders might be worthwhile looking into, if properly filtered (activated carbon or conc. sulphuric acid).

and

I'm actually considering using a bunsen burner for a very rough first FID - with a fine stainless capillary introduced through the air vent.

something that could at least come near-ish a professional instrument in separation power

really concerns me and is worrying

What is your experience with GC?



For this you want you really need to think again and consider buying a near-ish professional instrument


As I said before - as an academic exercise for a student on how to build a GC on a desert island it would be an interesting challenge
Regards

Ralph

[Peter Apps]

If it's the end rather than the means that matters then I would trawl the web and make personal contacts with labs and universities for someone who has a scrap GC languishing in a back room. The swap out the various bits until you have something that works.

You can't make an FID out of a bunsen burner, but you can make a flame photometric detector; use ordinary LPG as the fuel, poke the end of the column into the air inlet, put a coil of copper wire into the flame and you have a detector that will show a green colour when a chlorinated compound goes through it.

Peter

[ChristofferBR]

Gom; sorry, that was bad phrasing, I of course don't mean to get anything super reliable with balloon gas and a bunsen burner, I just meant I might as well dimension and size the components similarly, so that I can upgrade various parts in the future, some day maybe getting decent results. The helium thing was just implying that you can get cheap He specifically FOR balloons, but of course this is pretty impure.

Peter Apps;

It's both, really. I don't have any need, aside from a general interest in analyzing stuff around me, for a GC, but the construction of a functioning instrument (that isn't just a demonstration thing) is exciting.

An FID is literally just two platinum electrodes with a giant DC bias in a hydrogen flame. That config. with a bunsen burner, was even used as diodes in very early radio receivers. The question remains how sensitive it can get, though. The electronics are as simple as coupling out the signal from one of the electrodes, have a separate linear amplifier and integrating amplifier, then outputting to two channels of a pen plotter or ADC.

I'm likely not going to use an actual bunsen burner either, due to the relative large volume and large flame, but a small FID made from glass and stainless hypodermic tubing has been described in the litterature.

Again, thanks for the interest, but my question remains, though; would it be catastrophic for a reverse-phase capillary GC column to have the first ~2cm where the injection is made being not heated? And if so, could a short lenght of inactive column be used for feedthrough with a capillary nut coupling on the inside?

[osp001]

Again, thanks for the interest, but my question remains, though; would it be catastrophic for a reverse-phase capillary GC column to have the first ~2cm where the injection is made being not heated? And if so, could a short lenght of inactive column be used for feedthrough with a capillary nut coupling on the inside?

In my estimation, this much un-heated column might lead to undesirable condensation and cause some weirdness in your chromatographic results.

If you were able to sleeve the column through some 1/8" copper tubing, and leave enough of the copper tubing inside the oven, that tubing would get pretty darned hot- enough to keep that ~2cm of capillary column nice and toasty, particularly if it were wrapped in glass wool or another insulator.

[Rndirk]

Just my two cents.

For an easier challenge, to visually separate mixtures for demonstration or fun, i think it would be much easier to focus on liquid column chromatography..

[ChristofferBR]

Again, thanks for the interest, but my question remains, though; would it be catastrophic for a reverse-phase capillary GC column to have the first ~2cm where the injection is made being not heated? And if so, could a short lenght of inactive column be used for feedthrough with a capillary nut coupling on the inside?

In my estimation, this much un-heated column might lead to undesirable condensation and cause some weirdness in your chromatographic results.

If you were able to sleeve the column through some 1/8" copper tubing, and leave enough of the copper tubing inside the oven, that tubing would get pretty darned hot- enough to keep that ~2cm of capillary column nice and toasty, particularly if it were wrapped in glass wool or another insulator.

That's a great idea, I might try that! My column->injector coupling seems to be mostly fused quartz, so it'd probably withstand the heat:
http://www.ebay.com/itm/Syringe-to-GC-C ... xygPtSsKN6

Just my two cents.

For an easier challenge, to visually separate mixtures for demonstration or fun, i think it would be much easier to focus on liquid column chromatography..

I'm mostly an organic chemist, I run silica columns way too often! The challenge is that it's a challenge; like that guy who built his own electron microscope some years back.

[Peter Apps]

A cold spot in the column just after the inlet will be a disaster. Rather drill a wider hole through the oven wall and have your inlet going through it with just the septum sticking out.

The old literature on GC from the 1950s and 1960s has covered pretty much all of this, and it's fun to read.

Peter

[ChristofferBR]

Roger. I'll keep that in mind, thanks!

Another thing with regards to 'capillary plumbing':
- Mixing capillary with different ID's. Does a guard column/ sample loop etc. have to have the same diameter as the analytical column? We're talking 0.18mm and 0.25mm, specificly.

- Capillary nuts and ferrules: I've found some manuals specifying that a 0.4mm ferrule can be used for a 0.18mm column - what are the size limits of ferrules?

[James_Ball]

Roger. I'll keep that in mind, thanks!

Another thing with regards to 'capillary plumbing':
- Mixing capillary with different ID's. Does a guard column/ sample loop etc. have to have the same diameter as the analytical column? We're talking 0.18mm and 0.25mm, specificly.

- Capillary nuts and ferrules: I've found some manuals specifying that a 0.4mm ferrule can be used for a 0.18mm column - what are the size limits of ferrules?

0.4mmm for 0.18 and 0.25 column
0.5mm for 0.32 column
0.8mm for 0.53 column

If you want to go capillary I have an even easier way to make the injector and keep the column in the heated zone

Use a standard Swagelok 1/8" Tee fitting inside the oven. The carrier gas comes in the side of the fitting. On the bottom use a 1/8' to capillary sized vespel/graphite ferrule, this will allow direct coupling of the column to the Tee fitting. On the top of the fitting use a short piece of 1/8" copper tubing that passes through the oven wall then use a coupling fitting and place a septum under the nut to seal the opening and make an injection port. If you want the flow coming in higher above the column head, flip the assembly so the Tee is outside the oven and the coupling fitting is inside. Actually that would probably be more like a conventional direct injection inlet with the Tee outside the oven.

If you really want to be creative, use two Tee fitting, one outside for the incoming carrier gas and one inside the oven used at a splitter. To do this you would need to add a backpressure regulator to the split vent line to set the column head pressure, and a needle valve on the incoming line to control total flow and set the split ratio.

[ChristofferBR]

Thanks! that's really good advice! Split injection could be fun implementing at a later stage, absolutely! I may have sourced a surplus HP FID too, so this is shaping up! I'll be posting about the progress on this project here:

https://chrisbnstuff.tumblr.com/

For those who might be interested in my progress. Not much on there at the moment, though.

[GOM]

Thinking about it

At the end of the day the injection port is just a way of transferring your sample into the column, with all its inherent problems and post queries

There is another way

Since you are determined to use capillary columns

There is a way of making a relatively crude manual on column injection with a method that could enhance or even replace a septum based split /splitless injection port

Not my original idea but was taught to me during the early days of glass capillary columns when on-column injectors had not been not invented

I found that it also worked for a fused silica column as well on a modern instrument on odd occasions with a slight modification and worked well

It totally by-passes the injection port for introducing the sample solution into the column and is a form of what would be described as on-column, so in the extreme could replace the need for an injection port

I almost didn't mention it because it is so crude and simple but it does work

Regards

Ralph