Gas Chromatography - packing a 1/8'' tube with adsorbent?

[ChristofferB]

Hi! This is a little bit off topic, but I swear i'll get to the electronics of this project eventually.

Analytical scientific instrumentation is a strange hobby of mine, and I'm trying to build a simple-as-possible gas chromatograph - as simple as can be done while NOT just making a demo tool, and as you may know, this requires a column.

I'm settling on 1/8'' (3.1mm) packed columns. These show up used on ebay, but i don't want to risk them having been used previously for dangerous stuff, and they're PRETTY pricy new.

So making my own it is. the stainless tubing and swage-lok fittings are easy enough, and i'm going to pack it with either alumina or alumina coated with wax/fat/etc to make them unpolar.

But here's my dilemma: the adsorbent powder in these columns need to be packed so good that there aren't any air pockets, and so that it doesn't fly out the other end when pressure is applied. I can find no references about how to do this, and i assume the proffesional companies that makes them has some very advanced vibrators or shakers to pack them.

So how would I go about packing? Do packed columns have a frit or plug on the output side? I think i'll start by just trying to pack a 30CM lenght of straight tubing, seeing how nice it packs but i'd love a little less -from the bottom- solution.


I know none of this screams electrical engineering, but I'm hoping someone will know something regardless. 

thanks in advance!

[ChristofferB]

Sorry to answer my own question, but it seems the preferred method is to put a glass wool plug in one end, attach a vacuum pump, and then pack by sucking adsorbent from a reservoir.
Ingenious!

When i get any actual progress down on this project i'm planning to do a big catch-all thread over the entire project! Hope someone will find that interesting!

[Hamrod]

Analytical chemist here, I think your best option is to go with a used column.  Any residual gunk is likely to be in the first foot or two of column, which can just be cut off and discarded.  Also, many stationary phases can be solvent rinsed to clean out any trace amounts of stuff.

As long as a used column hasn't been totally trashed, it ought to give at least halfway decent results, so you can focus on making sure all of the other hardware (pressure/flow/temperature control, detector, etc.) is working correctly.

Another thing to keep in mind for any column with an organic stationary phase is that water/oxygen in your carrier gas + high temp will kill a column very quickly (more true for the 0.25 mm I.D. capillary columns I'm used to, but still applicable).

[vindoline]

Professional chemist with 35 yrs experience here. Hamrod is right. Go with a used column. The real question is what do you want to analyze? Packed GC columns were obsolete technology back in the early 80's when I first encountered them. The sad truth is that there is a world of pain, price wise, if you want to play amature chemist. It makes ee look cheap and simple by comparison.

[ChristofferB]

Damn, I must have missed a notification! Sorry, didn't see the replies.

I know packed columns have great limitations over modern ones, but the big advantage i want to exploit is the significantly lower required carrier gas pressure, and greater column capacity.

So, I made a packed column myself, 40cm, just packed with alumina. It probably won't give much (any) analytically significant separation, but it will do the trick of restricting the flow through the system.

I finished my room temperature prototype, whether it does anything sciency is TBD, but atleast it looks fairly proffessional.





What is shown is the entire setup: the swage-lok tee has a septum under the third nut, and the straight tubing holds a diode sinking current and heating itself, used as a crude unbalanced TCD.

I have seen packed columns on ebay for reasonable prices but I've always had reservations about possible unknown nasty residues in them, but I guess I could look out for NOS ones too.

I don't have any specific applications in mind, I just think there might be an interest in an idiot proof, higher capacity instrument to analyze undiluted samples from reaction mixtures and such.

Thanks for your replies and sorry for the previous lack of mine.

[jpanhalt]

During my days in basic chemistry (circa 1966-1970) I made a few, short ( 12") packed columns (alumina) and used them in Varian GC's with flame detector.  Never had a problem making them.  Just be sure they are carefully packed with plenty of tapping/vibration.   Used them to separate hydrocarbons (e.g., 1,2-diplenylcyclobutene, diphenlyacetylene, isomers of 1,2-diphenyl-3,4-dimethylcyclobutene ). 

Was able to analyze reaction mixtures at low temps (80° to 85°C) without thermal degradation of the compounds.  Injection volumes were very small (much smaller than anyone else in the lab used).  You are limited by the vapor pressure of those compounds at those temperatures.  Never used a commercial column; although, I see little reason that wouldn't work too.  I know that is archaic, but was the best we had pre-LC -MS days.

[ChristofferB]

That's interesting to hear! Maybe there's hope for my column yet!

[vindoline]

Great to see you've got something built  . It will be interesting to see if you can get it to give you a signal. Can you give us some details about your TCD? Do you have any experience with a commercial GC?

[ChristofferB]

Great to see you've got something built  . It will be interesting to see if you can get it to give you a signal. Can you give us some details about your TCD? Do you have any experience with a commercial GC?

Thanks! My TCD is based on an article in jounal of chemical education, but I found it quite simplistic and clever. An LM317 current source is set up to put some current through an 1n4148 diode, negative terminal spot welded (with two screwdrivers and a car battery) inside a lenght if 1/8" 304 tubing, putting the diode in the gas flow path. Diode heats itself, and a steady output (with carrier gas on) is reached. The diode will then either heat up or cool down depending on if it encounters a gas component with less or more thermal capacity. It's balanced to a battery with a pot to tune the voltage, this should probably be improved in future designs though.

I don't have a chart recorder, so I added a instrumentation amp to output to a microcontroller.

Here's the schematic.

I've had a decent amount of experience with the HP/agilent 58XX series instruments, both -MS and -FID during my undergrad. I got really bitten by it, and want to explore what one could do in that direction, designing an instrument myself. I'm debating whether i should make a datalogging system with an arduino and a pushbutton to be activated when the sample is injected, or I should get one of the old HP integrators.

thanks for the interest !

[vindoline]

 or I should get one of the old HP integrators.

  on an old hp integrator! I hate those things. Even if you get one for free, replacing the crusty old ink/print head will probably cost more than a diy solution. A simple Arduino or Raspberry Pi based datalogger will serve you much better.

[ChristofferB]

Yeah that's kinda why I've hesitated as well. In general, relying on consumables that might disappear is a mess.

I've just completed the carrier gas flow control system! mass flow controller and panel mount gauge, connected to a disposable He-cylinder.

Only bodge is that the balloon He-adapter has an 1/8'' SAE fitting and I couldn't find a nice fitting for that so erm... a piece of high pressure hose with two gear clamps holds it together.. Probably not a solution i'd dare if my carrier were hydrogen!

here's a few pics.

[vindoline]

Clever idea about the He ballon tank. I hadn't thought of that. I have no idea about the purity or dryness of those though. It might be worth putting it through a diy Drierite (CaSO4) drying tube to remove water. Now that I  think of it, it's probably a good idea to have a drying tube on your diy alumina column when it's not on n use to try and keep it dry.

[ChristofferB]

And I have first measured flow through column! 20 PSI at the injector gives 3.8 ml/min!

you're right, a moisture filter would be nice. I don't care much about oxygen impurities, since in a packed column I think i'll have so much stationary phase compared to oxygen, degradation will be VERY slow.
Right now my biggest problem is that I don't have a proper injector septum, its just a hard rubber disc that i'm fairly sure will just mangle my Hamilton microliter syringes.

maybe a foot or so of 1" ss pipe with two swagelok reducing fittings to 1/8" - filled with calcium chloride or gypsum? I'll need some sort of particle filter though.

[vindoline]

Honestly, for this application and pressures, I'd look for an acrylic tube unit on eBay. I'd also use the blue CoCl2 doped Drierite to monitor the moisture level. When it turns pink (wet) you can regenerate it by baking in your oven.

[vindoline]

You can just use a packing of glass wool at either end as a particle filter. Also, I don't think plain alumina is O2 sensitive. Water will definitely deactivate it though!

[ChristofferB]

Yeah that would probably work well enough, but honestly those cheap ebay knockoff swagelok fittings are so cheap i think it'll be about equally expensive.

I'm actually a bit surprized they aren't leaking. I'd have thought surface finish on the ferrules would be a major issue.
 

[ChristofferB]

I gotta share this, it's running now! Here is seen samples of atmospheric air at different amounts injected. I haven't measured the area but they look proportional.

There's no separation of components, as would be expected, on 25 cm of Al2O3 column - next column will probably be 60 cm alumina wetted in paraffin. I have also read that solid aluminum stearate is a good alternative (1960's litterature, the definition of good gc phases has probably changed slightly).

anyway here's a pic! it takes a bit for my data aquisition to start, so i'm considering putting a small pushbutton connecting the arduino input with 5V through a resistor so when I inject i can briefly press it to get a very sharp line in my diagram marking the time.

for now i'm happy! the peaks are even reasonably symmetric!!

 

[vindoline]

Congratulations! Very nice proof of concept 

[coppercone2]

Unless you are worried about inorganic contamination then heating it to 400C will destroy pretty much everything organic

[ChristofferB]

Congratulations! Very nice proof of concept 

Thanks! I'm going to try and format the output data so OpenChrom will read it, i think it accepts a file format called .xy that's just a text list of "time,value".

There's a lot of clearance around the diode in the detector, it's a thinner-walled 1/8'' pipe, i'm going to mess around with a new detector. An important point, the diode's spot welded to the tube, and the leads is a large thermal mass in a small diode, so the sensor as shown is almost as sensitive to airflow on the outside! I wrapped the entire thing in cotton wool and tinfoil to mitigate that.

I also wanna try installing the column in an oven. maybe just a small toaster oven with built-in thermostat. I think, though, that the detector has to be hotter than the column to avoid your sample from solidifying/condensing in it.

This ended up being a kindof catch-all thread for this project. Maybe it should've been in the projects section.

Thanks for all the input!

--Chris

[coppercone2]

Speaking of cheap swagelok fittings, did you ever notice Hylok?   
How about hoke? almost sounds like your broke

What exactly do they do to compact it? Ultrasonic? Very high shock?

Swagelok says their fittings are supposed to polish themselves when you tighten them.

[vindoline]

I wasn't aware of OpenChrom. Looks very interesting! I'll have to take a look at this for work. Do you have any experience using this software?

[ChristofferB]

Not much! I just loaded a test file in it to see if it would do something like the format i described. It looks as if it has most of the features one would expect, it opens MS spectra too, i think.

I tried optimizing the detector, but couldn't really get anything stable out of it. I think i'll work towards a full wheatstone bridge detector with NTC's as feelers, see attached.

the Tee fittings have completely square bodies so i'm going to sandwich aluminum sheet inbetween them, making one block that can be heated uniformly.

[ChristofferB]

I've run into a strange problem.

Whatever I inject, regardless of gas content, i get the same peak, same-ish area, retention time and shape.
Then I tried sucking up a syringeful of Helium from just inside the injector septum and inject that - same peak!

The only explaination I have is either my syringe or septum is leaky, enough that the entire sample is just squirted out before hitting the column, and the peak i see is just from the rise in pressure.
but that still leaves out the strangest: why do my TCD detect the carrier gas? is it the pressure differential?

hm. I think step 1 is to replace the septum.

--Chris