negative peak on REZEX ROA

[SandraAlbrecht]

Hi, I'm new in this forum. I hope I can find an answer to my problem. I'm running a REZEX ROA H+ column with 0.5% TFA. I always get a negative peak (-0.1 AU)for the UV detector at around 10 min. Even if I inject only water. Does anyone have an idea why this happens?

[tom jupille]

It's called a "system peak". Comes from the slight equilibrium upset generated by the injection. In principle, you can get one system peak for every component of your mobile phase if you inject a sample that does not contain that component.

[SandraAlbrecht]

Thanks! I didn't expect it to be so big though

[tom jupille]

How big it is depends on how sensitive your detector is to that component. An RI detector can give a huge peak, as can a conductivity detector. UV would be more variable, depending on the wavelength.

One way to check would be to inject your mobile phase. You should see very little (not necessarily zero because the pressure pulse from the injector can shift equilibria a bit). Then inject water with a *higher* concentration of TFA and see if that peak doesn't turn positive.

[SandraAlbrecht]

It must be the void peak, even if I can't really explain the nature of it. I have all my substances dissolved in the mobile phase, but it still gives quite a big response (as it does injecting mobile phase only) for UV detection as well as for RI. Also, the calculated retention time for it should be approx 14 min (7.8 x 300 mm column, 0.5 mL/min) and it comes at approx 9.5 min, which I'm grateful for, since it shortens the run. The material is packed extremely well, otherwise I don't have an idea why the retention time differs so much.

Nonetheless, I'm having it up and running nicely and it's reproducible between runs on this machine and this and another one. I'm just trying to understand why what happens

[M Farooq]

It must be the void peak, even if I can't really explain the nature of it. I have all my substances dissolved in the mobile phase, but it still gives quite a big response (as it does injecting mobile phase only) for UV detection as well as for RI. Also, the calculated retention time for it should be approx 14 min (7.8 x 300 mm column, 0.5 mL/min) and it comes at approx 9.5 min, which I'm grateful for, since it shortens the run. The material is packed extremely well, otherwise I don't have an idea why the retention time differs so much.

Nonetheless, I'm having it up and running nicely and it's reproducible between runs on this machine and this and another one. I'm just trying to understand why what happens

How did you calculate the retention time of the system peak? At times when the injection valve is worn out, one also sees a wave or disturbance in the chromatogram (somewhere near the dead time). Since your dip is far away, it is most likely a system peak. Ideally, a system peak should vanish if a mobile phase is injected because there is no disturbance in the mobile-phase stationary phase equilibrium. It is surprising that you still see a large dip upon injecting the same mobile phase.

[SandraAlbrecht]

t0 = ½ Pi x r-squared x l/flow rate
r = 3.9 mm
l = 300 mm
flow rate = 0.5 mL/min
mobile phase = 0.5% TFA (if that helps to explain the issue)

I thought that maybe some of the syringe wash solution (50% methanol) was injected because, the injector might not be as tight as it used to be, but we got the inlet changed very recently (which helped to reduce the peak, but did not make it disappear). I also compared it with a different system. There the peak is much smaller (5-10 times), but at the same time (as it should)

[M Farooq]

t0 = ½ Pi x r-squared x l/flow rate
r = 3.9 mm
l = 300 mm
flow rate = 0.5 mL/min
mobile phase = 0.5% TFA (if that helps to explain the issue)

I thought that maybe some of the syringe wash solution (50% methanol) was injected because, the injector might not be as tight as it used to be, but we got the inlet changed very recently (which helped to reduce the peak, but did not make it disappear). I also compared it with a different system. There the peak is much smaller (5-10 times), but at the same time (as it should)

Sandra, I am afraid about the validity of the above relation, if something is missing there. I get a different number; typically one would assume that the dead volume of the column is 70% of the geometrical column volume (pi*r^2*length). With the dimensions of your column e.g. 30 cm, and r=0.39 cm, one gets a void volume as =0.70*3.14*(0.39x0.39)*30 =10 cm^3, and with a flow rate of 0.5 cm^3/min, the dead time of this column is 20 minutes! Does this number make sense?

[tom jupille]

Does this number make sense?

Actually, for that type of column it does not. First of all, the packing is a homogeneous PS-DVB resin; it's non-porous (effectively, a gel) which means that there is no mobile phase inside the particles as there would be in a silica or macroporous polymer based material. Vm is therefore simply the extra-particle volume, so somewhere around 25 -30% of the empty cylinder volume, depending on how tightly it's packed.

In effect, this corresponds the the total exlusion volume.

[M Farooq]

The packing is a homogeneous PS-DVB resin; it's non-porous (effectively, a gel) which means that there is no mobile phase inside the particles as there would be in a silica or macroporous polymer based material. Vm is therefore simply the extra-particle volume, so somewhere around 25 -30% of the empty cylinder volume, depending on how tightly it's packed. In effect, this corresponds the the total exlusion volume.

If the Rezex packing is non-porous, that further simplifies the relative prediction of the void volume (assuming a cubic close packing). I am afraid there is still something missing in the following relation because it assumes the column is 50% empty.

t0 = ½ Pi x r-squared x l/flow rate

[tom jupille]

it assumes the column is 50% empty

Where did the 50% empty come from? A better estimate for total exclusion volume is around 27% (again, assuming cubic closest packed; it might actually be a bit less because these columns can be "overpacked"; the resin particles are somewhat deformable).

[M Farooq]

Where did the 50% empty come from? A better estimate for total exclusion volume is around 27% (again, assuming cubic closest packed; it might actually be a bit less because these columns can be "overpacked"; the resin particles are somewhat deformable).

True. I am not sure of the cross link of the substrate and its compressibility. The 50% assumption comes from the incorrect relation used by the original poster to estimate the dead time of the Rezex column t0 = ½ Pi x r-squared x l/flow rate

[SandraAlbrecht]

I was given this formula as a rough estimate for calculating the void peak.

Anyways, I managed to get the peak very small by changing the syringe wash solution to the mobile phase. I guess there's a bit of a leak when injecting, so wash solution gets into the system. I checked this with RI and got a peak when washing with 50% MeOH or isopropanol.

I'm all good now, thanks for the help!