Troubleshooting no signal from FPD

[TestRig]

I purchased a Carlo Erba HRGC 5300 for which I could not find manuals. The unit was not working. At least the oven did not start to heat up. I tried my best to fix the original electronics but I am not skilled enough. So I replaced the oven control system with an industrial controller. The unit was also missing the column so I purchased a used one online. Unfortunately there are no ferrules available for 250um columns for this type and my column is of course just that. I ended up overtightening 500um ferrules to be snug when pulling. Can you do that?

When I finally got the oven set up I made an arduino based measurement setup for the FPD unit. I am able to record the voltage from the FPD. I have now tried to inject 2ul samples of DMSO diluted to isopropyl alcohol. To my understanding I should be able to get some kind of signal but I don't. I am using approximately 1:100 000 by weight dilution.

Is there a way to test the FPD separately? I am wondering whether it works. The unit is close to 40 years old so it is perfectly possible. Nothing seems to last these days...

[TestRig]

I tried removing the photomultiplier while measuring and this resulted in signal saturation. So at least the thing is somehow working and connected.

[uzman]

Did you check if the flame is lit.
FPD detectors use the flame emission of sulphur or phosphorus ; depending on the filter installed.

In order to get a signal from the photomultiplier , the flame must be lit and a sulphur or phosphorus containing compound must be injected.

[TestRig]

Pushing the flame ignition button makes a pop sound come out from the FPD. This happens only once. If I press the button again it only saturates the signal output. If I turn off the hydrogen and restart the system again one pop and so fourth. I assume the flame is ignited.

[TestRig]

The flame apparently is not lit. I took the detector to pieces and cleaned the windows. The exhaust cap on the top was jammed but with violence I was able to get it out. I saw a video on youtube showing how to place a polished wrench on top of the exhaust to see whether water condensate comes on the surface. The condensate only appears for a short while after ignition. Is there some procedure to get good parameters for a stable flame? The unit has kPa dials to adjust the flows. Somebody has drawn lines to 50kPa H2 and 100kPa air with a black marker pen.

[TestRig]

The flame started to stay on when I put 60kPa to Make-up. I can now see condensate forming on a room temperature polished allen key. Still no signal from the sample. Now I am wondering how difficult it is to see DMSO with the FPD? I have let myself to believe that in order to get a signal you need to reduce the sulfur to S2- radicals which emit the photons when excited. Maybe I should increase the make-up hydrogen until flame turns off and go back a notch to ensure that I am not burning lean?

[uzman]

Are you sure that the detector has sulphur filter ?

Since it is a second hand GC , other filters such as phosphorus filter may be present, so you cannot get a response for sulphur compounds.

[TestRig]

There were some papers that came with the unit that said something about sulphur filter. Also the filter has a blue tint. Is it possible to tell somehow?

[uzman]

You may also check that there is carrier gas flow.

Turn off the flame , disconnect the detector side of the column and immerse the column end in a vial filled with a solvent ; such as methanol.

You must see stream of bubbles coming out of the column end .

Then connect the column again to the detector ;with a new ferrule.

Also the column insertion depth to the detector is very important , you may google to find some information about the model of your GC.

[TestRig]

I am now thinking that the problem is related to the inlet. I seem to get spiking with positive signals after injection for about an hour or so. I do get some occasional negative spikes as well I don't know where they originate. Anyway silly me has connected the 250um column straight to the septumless input configuration that this unit has. I think it is a cool on column type since it is unheated. I had some struggle with the inlet since I think someone had snapped a needle inside. With a help of a skilled machinist we tried drilling the way clear which ended with first good drilling chips then a jam and a snapped drill. Apparently the needle material differs significantly from normal 316l stainless. The part was then remanufactured with a bore to fit a 0.67mm (23gauge) needle.

My theory at the moment is that I need a retention gap in between and a smaller needle to be able to inject into the column. Today I took the original part for EDM drilling and ordered missing parts. Let's see. My reasoning is that the positive spiking is a result of intermittent sulfur droplet fragments finding their way to the flame/detector.

I did check the flow with a glass of water.

[nipe]

PFD is very sensitive to relation carrier/H2/air. For example, for classic FID this relation is about 1:1:10, but for PFD - 1:1:2, and ignition is very hard, but flame is unstable. Because you must measuring flows in detector, but not pressure! You can set carrier flow (out of column) 2-4 ml/min, H2 30-40 ml/min, air - 200-250 ml/min for ignition(!) and decrease to 60-100 ml for stable job. Sulfur filter is about 395 nm (blue/violet) , but as your GC is very old, PMT tube may be not sensitive, and power supply of PMT have not enough high voltage also. You must check it primary. Burning of flame you can control by mirror, retention times for your column and components you can find in column manufacturing catalogs or on their site (and best working condition also!). You can check detector also by replacing your column by empty part (20-30 cm length) of any column, you must see total (high) sulfur peak. And check the injector of couse! Good luck!
P.S. Hydrocarbons decrease PFD signal (for example, at the same time as your components)!