Chromatography Forum

Most modern pumps use complicated mathematical algorithms (are there other kinds?) to offer some level of compensation of changes during the run to the mobile (temperature, compressibility, viscosity and others) to give some level of both good flow accuracy and precision. Precision is more important (imho), consistent retention times are something I like.

Compensate for what? How does the pump know, whether I am running acetonitrile or water or methanol or hexane? If it measures something at high pressure and needs to compensate, I doubt it is doing the right thing from the standpoint of the analysis.

There are some exceptions in the case of high-pressure solvent mixing, but all what this means is that high-pressure mixing is fundamentally flawed from the standpoint of flow rate precision.

Please do not misunderstand me: I am a fan of high-pressure mixing, but everything has its advantages and disadvantages.

In some cases the pump (1200SL) does indeed know (because you tell it) what solvent(s) you are running. Exactly how the pump (actually the pumps firmware) uses this information, I do not know.

Hi there,

Does anyone care to explain to me, what the rational is for multiplying the area by the flow rate, in order to “recoverâ€

Danko, I explained it already, but here it is again:

The peak height is proportional to concentration. When you integrate, you get the peak area in the form of concentration C times time t.

area = C*t = m/V *t

with concentration being m/V = mass (or weight) divided by volume. This is per se not very useful, but if you multiply with flow rate, you get mass:

area * F = m / (V/t) *F = m / ((V/t) * (V/t) = m

because the flow rate F is volume divided by time.

Thus the peak area multiplied by the flow rate is the mass injected.

Elementary, my dear Danko!

AA: if a pump on purpose increases the flow rate to have the "correct" flow rate at the column inlet, it is doing the wrong thing with respect to having a flow-rate independent peak area at the column outlet.

I know that I may very well be a lone voice in the wilderness. But I do not understand why anybody would want to compensate such that one gets a constant flow rate at the column inlet, when due to solvent compressibility it will change over the length of the column and result in an arbitrary flow rate at the column outlet, where quantitation is done by the detector.

Of ourse, in real life, such subtleties become irrelevant, because you create your calibration curve under the same conditions as you do your analysis. If you would use my approach though, you could calibrate at one flow rate and measure at another, provided you have a concentration-proportional detector such as a UV-Vis detector.

Uwe,

I fully agree that, from a practical user point of view, it is of little interest what goes on inside the pump, as long as one get some reasonable level of reproducibility. And, of course, all pumps have some specification that defines flow accuracy (1 mL/min +/- X% at 1000 psi backpressure using methanol at 25 deg. C or something like that) and flow precision (usually a %RSD of some retained peak under a defined set of conditions). The point I was trying to make (and I am not sure why), is that much more goes into the design of LC pumps that one might think and there are algorithms making subtle adjustments to the pumps speed, stroke length and probably other factors that most of use are highly unaware, mostly because we don’t need to be. I would say that a reasonable analogy would be the average car. Most understand how an internal combustion engine works, but few could tell you how the air/fuel mixture is constantly adjusted for peak performance.

But, again, as you say it really doesn’t matter, samples go in, peaks come out and all is well

Danko,

Just some points for you to ponder based on your last post.

- System volume is not necessarily constant (1100 and 1200 system volumes change with backpressure due to the pulse dampener). Backpressure in these systems is constant under isocratic conditions but changes when running gradients and also would be different using different isocratic conditions (water/MeOH compared to water/ACN)
- Both "time constant" (or what ever you choose to call it, i.e. Filter constant) and sampling rate (which you describe well enough) will change the peak height, and therefore the peak area
- The amount of change that these 2 parameters have is somewhat dependant on the peak width, fast chromatography generating narrow peaks, say less than 2 seconds, the effect can be quite dramatic, wide peaks (30 seconds or more), The effect can still be seen but has much less of an influence.

AA

I have not crawled inside a data system in recent times, so things may have changed. But here it is how it used to be (and still may be).

Data are collected by the detector at the fastest rate possible. The data are accumulated (which reduces the noise with low sampling rates) and then passed on. Alternatively, you can do the exact same thing with the data system, i.e. the data system is collecting the data at the highest rate and then accumulates them for storage. If you do that, the peak area is not affected by the sampling rate, only the peak height will be if the sampling rate is too low compared to the peak width.

Hi Uwe,

Not that elementary I would say
By the way, I had red your previous posting, but thanks for the repetition
You see, I was referring to Mags second post:

I should have clarified more, not only does the height decrease, but the area also decreases, so I would think that it is not an efficiency issue

So how does your formula (the one with the volume/time) address the height issue?

To AA, you can see the same effect also on systems without pulse damper. Besides, you’re looking at very marginal changes.
Regarding the sampling rate, I wrote my self that it can influence the peak definition, so there’s no need to remind me of that. But seeing the peak widths jzt shared with us being in the range of 0.02 – 0.04 min @ ½ H, I concluded that 20 Hz was more than fast enough to define these peaks.

Best Regards [quote][/quote]

Danko, look at the posts again: the area times the flow rate remained constant, as predicted in the equations that I showed.

The height of the peaks decreases due to a change in column performance. As the plate count decreases, the peaks get (proportionally) wider and the height (at constant absolute area) decreases, when you increase the flow rate.

Start off with the idea that the peak area times the flow rate is constant, then all the rest makes complete sense and is completely normal.

Hi Uwe,

You're right. Mags doesn't mention the peak widths. They might as well be staying constant. And then our dear van Deemter is in the picture again.
So, unless other facts come up, I’m going in pause mode

Best Regards

Good! We are now ~ back to Uwe´s first contribution and mine from 9th Nov.

hi,

Read about Van demetre equation, answer to your question lies in that

Hi 4ugirish,

Do you answer someone’s question, or are you just spamming the forum?