Noise after a post-column reactor

[DeimerlyC]

OK, I've posted a few times already trying to figure out why my peaks wouldn't resolve they way they would on a second system. I'm happy to report that I have successfully traced the problem to a back-pressure regulator, but as a result I have a new problem.

I have to place the regulator between the post-column reactor and the detector. If I don't, my baseline is poor (a lot of heavy noise, the up-and-down oscillating pump-noise pattern type) and I can't integrate well. If I do, my baseline is nice and flat but my system efficiency is gone (peaks merge in with each other).

I do have a 100psi BPR after the detector. A thought I had was to try lower and higher regulators to see if they have any effect.

When I had a very short connection between the reactor and the detector I also had a lot of noise, that's why I added the BPR. I have noticed that the stream coming out of my reactor is bubbling, which makes sense since the temperature is >120°C and venting to atmosphere. So I concluded that with the short connection reaction bubbles were getting into the detector. On my other system the problem seems to be solved by having extra tubing (about 1.2m) between the reactor and the detector in order to give the system time to cool (just speculating). I have since added a similar about of dead tubing between the reactor and detector.

My detector is actually two self-contained UV-VIS preset to 570nm and 440nm (yes, amino acid detection with ninhydrin). I have also discovered by switching the order they occur in the stream will change the noise found in each detector (second one has less noise). As a result I'm pretty sure something needs to be in the system to bring down the noise.

I'm a bit hesitant to take this to Pickering since I'm not currently using one of their products, but I'm pretty sure they would certainly be a good souce for information. If anyone has any ideas on how I can bring down my noise without sacrificing my efficiency I'd appreciate it.

Thanks in advance

[skunked_once]

What type of back-pressure regulator are you using for your system? Some that I have seen have a void volume and allow some mixing which is good for baseline stability but is very bad for chromatographic separation.

[DeimerlyC]

http://www.discoverysciences.com/litera ... sories.pdf

See upper right picture. I had a 29059 10psi located post-column and a have a 29064 100psi located post-detector.

Yes, I was also thinking if I could find a zero-volume BPR that would solve my problem.

...


Now that I'm looking at the picture over and over again, it appears that the one to the upper left would be a better choice. Its listed internal volume is 0.6 µL and the one I'm using has 100 µL. What would be your opinion?

Thanks for the reply btw!

[Uwe Neue]

To add pressure behind the detector causes the event that the detector sees the pump pulsation from the reagent pump. Therefore, this is NOT a good idea. To solve the problem with the bubbles, it is probably best to follow the successful strategy at the other system, where you have a length of tubing BEFORE the detector that allows the reagent to cool at least a bit BEFORE it enters the detector.

[skunked_once]

I would definitely try the 0.6 uL back-pressure regulator.

[Mark Tracy]

Just go the the Pickering website and download their hardware and application manuals. Lots of good advice. (I used to work for them, and I supported that application. I even wrote some of the stuff in the manuals.) Then imitate their setup as best you can.

[jamesque]

It sounds like the problem is twofold: 1) the high temp of eluent out of the derivatizing coil (~120C) is increasing the system noise and 2) when you add a BPR you lose resolution. You may want to try increasing the length of tubing between the reaction vesssel and the detector. But use tubing with a diameter of 5/1000 in and coil it. This will add length to cool prior to entering the detector but will not result in significant band broadening. Also the tubing will act as a BPR. The issue with the BPR is not total volume, but diameter (long and narrow is much better than short and fat). We have added as much as 20 feet (yes, 20 feet!) of 5/1000 ID tubing with minimal loss of resolution due to band broadening if the tubing is properly coiled. Hope this helps.

[Patrik Appelblad]

I also believe that sufficiently long coiled tubing would solve your problem, but the id of the tubing will depend on the flow rate of your application. If the tubing is properly knitted, i.e. into a tortuous path to force the fast moving centre of the liquid stream to mix radially with the slower moving boundary liquid layer, axial dispersion is minimized. The shape of a chromatographic peak entering the reactor is thus preserved very efficiently. To ensure a stable radial mixing, a linear flow rate of 10 cm/s or more is recommended.
You can read more at http://www.sequant.com/sn/p_intro.php?id=12
The longest knitted open tubular reactor we offer is 20 meters, and it is available by default in three different id’s.