Flow Rate does not Match Sample Volume

[kunicki]

Hello,

I am curious if anyone else has had a similar experience as I have here:

The Helios Software (updated this year) reports a Neb/MUG of 0.17/0.42 respectively (HT Injector), and our PSI unit runs at "30 uL/min" using 13-20 PSI. When running a long (>2hr) sample, our team noticed 4 mL of sample was collected over 1 hour (3600 seconds), which would be 60+ uL/min. This has reoccurred for all of our samples throughout the day. What we observe is a typical sample acquisition yield (~60%), suggesting no abnormal cell loss. Further, the sample data looks normal, with little to no background signal (Ba, Pb, I). To my understanding, the data should have changed if the sample flow rate were doubled, due to the fluid dynamics of the nebulizer, spray chamber, and torch generating the plasma, along with the impact on sample ionization efficiency.

So my question to the audience is two parts I guess:
1) Has this happened to you before? What instrument & settings were you running? Software version?
2) Is there a mistake I'm overlooking here that could have lead to the increased loss of sample volume?

I'd be very interested in your general thoughts, even if it's not a proposed solution to our "problem"(?).

Regards,

[mleipold]

Hi Matthew,

We have had a case where a PSI was intermittently pumping too rapidly. As far as we could tell, it was a case where the flow sensor went bad and we had to have Fluidigm send a replacement.

However, there are a couple things to check first.

1. Have you checked to make sure that the output rate from the nebulizer capillary matches the input rate from the tube in the PSI? By this, I mean putting 2-3mL of water into the PSI, and making sure you get 2-3mL out of the nebulizer capillary.

If you have a cracked silica line or some other kind of leak, you could be losing part of your sample prior to it entering the machine. Depending on where the crack is, it could be fooling the flow sensor, causing it to register slower than actual flow.


2. Have you tried backflushing all the parts of the PSI? This includes all silica lines and the flow sensor, systematically connecting and disconnecting until you have freely flushed all parts. We have had cases where we had a clog in the flow sensor, which constricted the diameter of the liquid stream and caused the pressure and flow rate to fluctuate a lot. Depending on whether it was immediately before, directly in, or immediately after the flow sensor, the pressure and/or flow rate went up or down.

In some cases, backflushing and foward flushing were sufficient. In other cases, we had to run 3% nitric or Wash solution with the parts disconnected to finally remove the clog. In another case, we ran about 1min of undiluted bleach through the PSI, which cleared it (followed by a LOT of MilliQ). And in another case, the flow sensor was irreversibly clogged and the PSI had to be replaced.


Mike

[kunicki]

Hello Mike,

Thank you for the response! There was in fact a clog in the Flow Sensor, and I was able to declog using 50% contrad at room temperature, applied normally through the sample probe (appx. 50 PSI at 5 uL/min) with the sample line disconnected. It took about 15-25 minutes run time for the blockage to come through, and the pressure returned back to 8 PSI with all parts assembled, suggesting the sample and capillary lines were already clean.

Yes, my SOP is to backflush the lines in these cases. But none of these flushes could reduce the PSI.

Unfortunately, we were in urgent need to continue a study, so I did not collect sample volume from the nebulizer to test the output flow rate. After declogging the Flow Sensor, the sample volume continued to collect at a normal rate (appx. 30 uL/min). This means I may have lost the chance to see where that extra volume was going at the higher rate (60uL/min). At the end of the day, but prior to the declog, I could confirm no excess salt or sample build up was observed in the spray chamber, injector, cone, or torch. I'm really still at a loss how the data was unaffected. Although there was much less 'debris' in the sample after cleaning, the cell marker MFIs were practically identical.

Sorry I could not answer your question, Mike! That was a great point.

Best,
Matthew