Spurious peak

[Amar]

Hi
I am trying to do reverse phase HPLC using C18 column and gradient elution. Buffer A is 5% Acetonitrile 0.18%TFA. Buffer B is 90% acetonitrile 0.15%TFA. I degas them using vaccuum with continuous stirring for 15-30min. Detector -214nm UV.
I have two problems
1) I always see spurious peak at 52% BufferB.
2) I see sawtooth baseline during initial phase pf gradient i.e. 5%to 25% B
I have cleaned column and flushed system, also tried by changing nanopure water from two different sources and also acetonitrile. Still i have these two problems.
Can anyone suggest some solutions? Thanks.
Amar

[Cliff Mitchell]

What kind of pump are you using? One thing to be aware of is which line is the primary channel (I believe this only applies to low pressure mixing quaternary, tertiary, and binary pumps). I know this is the case with the HP 1050 and may exist in the 1100.

If you are pumping a gradient of A to B, during the time frame when A is the dominant component of the mobile phase, line A will be drawn from first during the pump's cycle, then B. This switches when B becomes the dominant mobile phase component. So at 49 and 50% A, A is drawn first and B is drawn second. Then at 51%, B is drawn first, and A is drawn second, and there is a segment of mobile phase richer in B than should exist (according to the gradient profile). This is usually seen in absorbing mobile phases as a wave or peak in the 50-55% region.

Also, if your TFA is old (more than a week) it may have gone bad and you are seeing the elution of degredants. Between runs, equilibrate for twice as long as normal and see if the peak is twice as big. In retrospect, this seems more likely....

As for the second problem, air in the pump? Anyone else?

[Mark Tracy]

I have seen this sawtooth at the start of a chromatogram under similar conditions. This occurred with a high-pressure mixing pump where the B-channel was pumping a very low percentage. The B-cam cycle was longer than the residence time in the dynamic mixer. I reduced the problem two ways: one was to use a larger mixer. Two was to make B with only 70% acetonitrile since that was all I needed for the elution; I reworked the initial %B to compensate. I would also suggest that you reduce the %TFA as much as possible without ruining the selectivity of your method; the sawtooth amplitude will be proportional to %TFA. (I am assuming that your pump is clean and properly functioning.)

[Amar]

Thanks for the very valuable sugestions from both of you.. I will try to implement.
Amar

[putnam]

I have seen your second problem (sawtooth baseline at the beginning of the run) before. This became evident when we transferred a method to a European partner. We use Shimadzu with a very large mixing volume ... they were using Agilent/HP equipment with much smaller mixing volumes. Both utilize high pressure mixing. The gradients went from 0 to 100% B in twenty minutes.

I believe the problem is related to whether two solvents with very different compressibilities have an adequate time to mix and form a homogenous mobile phase. The differences in solvent compressibilities and the ensuing volume changes upon mixing causes refractive index changes, hence the increased baseline noise. As the ratios of these solvents becomes less drastic, the difference in refractive index is less noticeable. In a system with a larger mixing chamber, this process is allowed to come to completion before the mobile phase enters the column and detector. In a system with a smaller mixing chamber, this process takes place continously while the mobile phase progresses through the system.

I try to develop methods that begin with 10% B ... and premix the A/B mobile phases to mimic the starting conditions I want. This is because most of the noise present is due to the initial mixing of 100% A and 100% B mobile phases. Premixing can alleviate much of this.

Our solution, though, for this particular case was to have the receiving laboratory increase the dwell volume of their system. This worked for them.

Luckily, I can reduce the dwell volume on our systems and future methods will be developed with a dwell volume I believe will be more easily transferred.

[DR]

the sawtooth amplitude will be proportional to %TFA. (I am assuming that your pump is clean and properly functioning.)

?
Even if the % TFA is essentially constant between the pumps?
I would be more inclined to attribute the wave pattern to differences in RI than to TFA. I have been seeing this type of problem with a Shimadzu 10 series (xL) system, high pressure binary, lowest mixing chamber volume. Increasing the volume would help, but it would also add to the dwell volume of the system.
Best solution? Get better pumps (pairs of smaller plungers moving faster to acheive the same flow rate).

[Mark Tracy]

The absorbance for 0.1%TFA is about 0.35 at 214 nm for a 10 mm cell. Any mismatch between the A and B will be multiplied by the average background absorbance. Close matching of A and B helps a lot, but chemists are only human.

Looking at the data on which my advice was based, I see a sawtooth of about 2.2 mAU for 0.1%TFA, 0.7 mAU for 0.03%TFA and 0.2 mAU for 0.01%TFA.

The rule about noise proportional to background also applies to post-column mixing.

BTW, my new low-pressure mixing pump does not show this sawtooth.