Secrets of flow splitting?

Whether you are building your own mass spec, modifying an existing one, or just need help with instrument settings post about it in here.
Toxic
Angiotensin Member
Angiotensin Member
Posts: 39
Joined: Sun Sep 25, 2011 9:53 pm

Secrets of flow splitting?

Postby Toxic » Tue May 22, 2012 12:03 am

Who would like to share their secrets for getting flow splitting to work?

After many years of running a splitless nanoflow (Tempo), I have been trying to use our Agilent 1200 CapLC for doing LC-MALDI. So I'd like to split the flow down to 300-500nl/min before a manual injector to then run through a 75um ID IntegraFrit packed with Magic C18AQ. I've got a 25um ID capillary for the waste, but I'm getting very high back pressure at 20ul/min (150+ bar). I still have some mods to try, but proven suggestions would be greatly appreciated.

Cheers.

JDRCHEM
Phosphoserine Member
Phosphoserine Member
Posts: 24
Joined: Wed Jun 29, 2011 5:35 am

Postby JDRCHEM » Tue May 22, 2012 8:24 am

Unfortunately, I do not have any secrets for split nanoflow. Using split nanoflow was always a huge pain, but I was able to get it to work with lots of patience and probably more than a little dumb luck.

Do you know the internal volume of your injector? I have found that many injectors, unless rated for nanoflow applications, have large internal volumes (10 or more uL in some cases) that make splitting before the injector problematic unless your are willing to increase the time dedicated to loading sample onto the column and accept long gradient delays. I have also found that connections with non-zero dead volume fittings post split can cause frustratingly long delays. One trouble spot for me was the column connection to the injection valve.

I prefer to use a precolumn and split after the precolumn that way I can load the sample onto the column at several uL/min. Check out SI. figure 2 is this Ficarro/Marto paper. This is the way I configured my system with an autosampler. This requires two splits with trapping/analytical flow splitting controlled by the position of the 6-way valve on the mass spectrometer. I like to use long pieces of 50 um i.d. capillary and then gradually trim off the excess until I get the desired flow. I think switching from 25 to 50 um might give you a little more fine control of the flow split.

I personally don't think 150 bar is very high pressure especially if you are using 3 um particles. I used to run at about 200 bar, but the flow had to be ramped up and down because the switching valve on the MS system did not work very well above 100 bar.

Toxic
Angiotensin Member
Angiotensin Member
Posts: 39
Joined: Sun Sep 25, 2011 9:53 pm

Postby Toxic » Tue May 22, 2012 2:30 pm

I was in a rush when I made the post yesterday, so there's some stuff I left out. I'm thinking of mounting a CapTrap on the manual injector (no autosampler) and loading the sample into that before switching it into line with the nanoflow. At the moment I'm just flowing through the injector in the load position so nothing is going through the loop. As for the volume of that path, I have no idea.

How long is "long" for the capillary? The length of 25um I have on at the moment is 40-50cm.

I'm getting 150bar with no column fitted. There's a restriction somewhere that I'm trying to track down.

Post update: Now that I stare at that diagram, that IS the much better way to set the columns up. I have a spare manual injector that I could use to do this.

JDRCHEM
Phosphoserine Member
Phosphoserine Member
Posts: 24
Joined: Wed Jun 29, 2011 5:35 am

Postby JDRCHEM » Wed May 23, 2012 4:07 am

On my split-flow system, 40-50 cm would have provided too much backpressure for 25 um i.d. capillary. I probably used 30-40 cm of 50 um i.d. capillary, but, I was splitting the flow from 250 uL/min down to 250 nL/min. You might be able to get away with long lengths of smaller capillary if you are splitting from 20 uL/min.

As for your pressure problems, have you looked at the rotor seal in the injector? I have found that sometimes those grooves get compressed or partially clogged leading to artificially high backpressures. Less likely is the possibility that the load/inject positions on the valve are a little misaligned. That is unlikely if you are using a Rheodyne valve, but it never hurts to take apart the valve and inspect the stator and rotor seal.

Toxic
Angiotensin Member
Angiotensin Member
Posts: 39
Joined: Sun Sep 25, 2011 9:53 pm

Postby Toxic » Thu Jun 21, 2012 11:14 pm

Just to update this thread a little and hopefully get some more advice. I took JDRCHEM's advice and increased the diameter of the split waste line to 50umID. The length is about 1 metre and with a flow of 75ul/min, I get a split flow out of the end of the analytical column (10cm 75umID packed with 5um particles) of about 300nl/min with about 140bar of stable back pressure. So that problem is sorted.

The next problem has been the vented column set up. Using the diagram in the paper mentioned above and some other examples, I set up the vented configuration using a manual rheodyne valve. So the flow from the pump goes into another rheodyne with a sample loop and from there into the vented column set up. When in the trap loading position (split waste line blocked and all flow directed through trap and then out secondary waste line at tee with analytical column) pumping at 5-10ul/min through the sample loop and then the trap, the back pressure just keeps rising (to 300bar) and I'm not getting much flow through and out the second waste line. The trap column is also 75umID and packed to 5cm with the same particles. Even leaving the flow directed through for 30 mins and then switching the valve to run the gradient (where the pressure returns to 140bar), I'm not seeing any of the GluFib I load when the plate ends up in the MALDI. In desperation I tried loading the GluFib with my column packing bomb and then eluting, but I still didn't see anything.

So my next idea is to pack the trap with some 20um POROS R2 particles, which have far lower back pressure. Hopefully this might fix things, but if anyone has any suggestions, I'd welcome them.

Toxic
Angiotensin Member
Angiotensin Member
Posts: 39
Joined: Sun Sep 25, 2011 9:53 pm

Postby Toxic » Thu Jun 21, 2012 11:14 pm

Just to update this thread a little and hopefully get some more advice. I took JDRCHEM's advice and increased the diameter of the split waste line to 50umID. The length is about 1 metre and with a flow of 75ul/min, I get a split flow out of the end of the analytical column (10cm 75umID packed with 5um particles) of about 300nl/min with about 140bar of stable back pressure. So that problem is sorted.

The next problem has been the vented column set up. Using the diagram in the paper mentioned above and some other examples, I set up the vented configuration using a manual rheodyne valve. So the flow from the pump goes into another rheodyne with a sample loop and from there into the vented column set up. When in the trap loading position (split waste line blocked and all flow directed through trap and then out secondary waste line at tee with analytical column) pumping at 5-10ul/min through the sample loop and then the trap, the back pressure just keeps rising (to 300bar) and I'm not getting much flow through and out the second waste line. The trap column is also 75umID and packed to 5cm with the same particles. Even leaving the flow directed through for 30 mins and then switching the valve to run the gradient (where the pressure returns to 140bar), I'm not seeing any of the GluFib I load when the plate ends up in the MALDI. In desperation I tried loading the GluFib with my column packing bomb and then eluting, but I still didn't see anything.

So my next idea is to pack the trap with some 20um POROS R2 particles, which have far lower back pressure. Hopefully this might fix things, but if anyone has any suggestions, I'd welcome them.

Toxic
Angiotensin Member
Angiotensin Member
Posts: 39
Joined: Sun Sep 25, 2011 9:53 pm

Postby Toxic » Tue Jun 26, 2012 11:56 pm

The POROS R2 particles fixed the pressure problem on trap loading, but I had a bastard of a day with leaks with other parts. Hopefully tomorrow will prove more fruitful.

Toxic
Angiotensin Member
Angiotensin Member
Posts: 39
Joined: Sun Sep 25, 2011 9:53 pm

Postby Toxic » Fri Sep 21, 2012 1:18 pm

I have to drag this thread up from the dead because we are STILL having problems getting things to work. We've been trying to assess whether or gradient is being formed properly using 0.1% acetone with no column and get poor results. Without the splitter, things work fine and the gradient looks good.

So has anyone got a detailed protocol or SOP about flow splitting HPLC that they would be willing to share so that we can get this right?

The other request is has anyone tried perfoming gradient nanoflow using two syringe pumps? With such a system, how is the high back pressure dealt with? I've got two New Era pumps that will deliver nanoflow rates and can control each other for different pumping speeds and thus make a gradient at a set flow rate at the outlet of a tee. But I'm worried about pressure, syringe leakage and killing the pumps.

Cheers.

proteovations
Phosphoserine Member
Phosphoserine Member
Posts: 11
Joined: Sun Sep 02, 2012 12:47 pm

Postby proteovations » Sun Sep 23, 2012 8:32 am

Do not use capillary tubing for making the restriction. Use an old HPLC column or something else packed with c18 media. The reason is that the old column mimics the change in back pressure that changing from organic to aqueous causes. 20 um capillary peak tubing does not and typically I always saw my flow rate drop off during the gradient to organic. Quick changes from aqueous to organic make the problem worse.

As for the syringe pumps, I would contact the company and find out what pressure they can work at continuously. Must pumps can only work at the high end of their pressure range for a sort time. Found this out using a 1100 near 400 bar and having it break repeatedly even though it was rated to 400 bar.

Use stainless steel syringes with the syringe pumps and expected to work at half the max pressure. I do not know much about New Era pumps specifically though.

Also, make sure you are not running your pump to close to the low end of the flow rate because the innate dead volume in the pump will be a huge factor. So if the flow range is 10 ml /min - 50 ul/min, try to use atleast 500 ul/min.

I can offer more tips if you want. Would prefer to talk by phone to get more details.
Benjamin J Cargile
"Let me know how I can help."
Proteovations
Proteomics Services | Protein Identification


Return to “Instrumentation”

Who is online

Users browsing this forum: No registered users and 1 guest