We have had in the past problems with proportiong valves. In fact the service technician said the machines have problems getting the gradiant in the range of 1 to 10%. He said the machine could more accuratly measure out 48 and 52%. In order to reduce the effect of the proporting valves we decided to make a premix buffers for our gradiants. It has appeared to work for some products but we have seen a change in chromatography for another. It does not make sense to us why there would be any change.

We have a gradient that starts out and 73%buffer and 27%acetonitrile it ramps up linearly to 65%buffer and 35%acetonitrile over a 7 min and held there for 5 min. We use to have the machine mix this for us. We switched to have a premix of 73:27 and started with a 100% of this solution and ramped it up linearly to a pre mixed buffer of 65:35 over 7 min. The chromatography was different for this method with the premix. I would appreciate any imput on why this would be different Thank you for your assistance.

The problem is most likely the same as the one described by Veronika Meyer in her "Pitfalls and Errors of HPLC". The actual composition of a mixture of organic solvent and water depends on how it is mixed. If I add first water to a graduate and fill it with organic solvent, I will get a different composition than if I add first the organic solvent and then fill the container with water. In a somewhat more subtle way, the mixing of a gradient will follow the same difficulty.
I have treated the same story in an American Lab HPLC Troubleshooting article. If you give me your e-mail, I'll send you a copy.
Uwe_Neue@Waters.com

I am surprised that you thought it would be the same. But let me try and get this right, you had a gradient that went from 27% to 35% B. Your service rep says that your instrument cant do 1-10%. So, you make mixes that got from 0% to 100% B. Am I confused or does not this new gradient not have 2 portions that fall (1-10% B and 1 -10% A)within this 1-10 mixing problem zone.

Also, what kind of instrument are you using, you should easily be able to do 1-10%.

A long time ago a prof. (in Denver) kept telling us students that 2+2 is not = to 4 in chemistry. He got a fit if he saw us mix solvents in a graduadet cylinder, as the brew will be undefined. Firstly, there is usually a temp. change on mixing (thus a vol. change), secondly, the more exothermic or endothermic the mixing, the more the volumes may differ from the total premixed volume. So if you pour in a solvent on top of another you get some mixing right away, differently each time you do it.
Now we are sometimes(when composition is very critical) having repetition trouble, even when measuring the solvents separatly first, then mixing. The cylinders are just too inaccurate.
That´s why I have been looking for a mixer, guess have to look very carefully.
Also: hopefully authors of articles will specify more often in the future, how they mixed their moph.

Thank you for you quick response. Let me clarify how the mixing occured. In order to prevent the incorrect mixture a graduated cylinder was used first adding the buffer to the correct volume and then transfering to a beaker and than the appropriate volume of acetonitrile was added to the empty graduated cylinder. This in turn was added to the beaker, mixed, and filtered.

The machine is capable of making a mixture in the 1 to 10 percent range but it's accuracy is not as good. The idea of the premix has the concentrations that are not as critical at the 1 to 10 range and the criticle points will be at the premix concentration, which would be ran at the 100% mixture.

Let me see if I understand, by premixing the buffers the volume change so when I tell the pump to go from one percent to another I am not getting the true amount added. I am still having a problem with this. It will still be added at the same rate but the initial concentration of each would be greater? Hence, I supose the increase in buffer concentration would not have much effect but the increase in concentration of the acetonitrile would have a great effect. Is this what you are saying?

What type of gradient system are you using? A low-pressure gradient system (single-pump type), or a high-pressure gradient system (dual-pump type)?

If A solvent was initially buffer (conc.??) and B solvent was ACN, and the instrument was allowed to mix, you could be seeing a difference in ionic strength now that you're using premixed mobile phases. You also implied that you added the ACN to the same graduated cylinder in which you measured the buffer. Ideally, the solvents should be measured in separate grad. cylinders.

Sorry for the delay in the response. The system is a low pressure system. When I said I used the same cylinder it was after I poured out all of the buffer, leaving only a residual amount before adding the acetonitril

You still havn´t told us how your chrom. changed and whether the flow was the same at the column outlet for the two methods. Still, one can now agree with the 2nd anonymous: it´s not surprising that they would be different. Simple, and therefore attractive, possibilities: a. The measured volumes by the two systems is not identical, and/or b. the energy of mixing will be different (temp. of the MP is different), c. the mixing efficiency is different.

Considering how you mix the solvents, I don’t think that the solvent contraction of the MeCN/H2O mix is the problem. Anonymous II had a good point, but you need to weigh the statement of the service rep with a grain of salt. Unless your 1 to 10% mix is off by a hugh margin, you may be better off with the premix. Think about it: if a 1% mix is off by 1%, this is a giant error. If your 27% mix is off by 0.27% (1% of 27%), it is a small error. In my opinion, you might be better off with the premix, but you should get a clarification from the service rep as to the meaning of his statement.
I still don’t have a good answer on the difference (except for the possible issues mentioned above). Another possibility is cooling, but I think it is far fetched. If you mix MeCN with water/buffer, you get a significant amount of cooling. With all the lines in a standard HPLC, I would think that this is not important, but I could be wrong. Do you use temperature control for the column?

The chromatography had coeluting peaks with the main analyt when the premix was used, but when the machine mixed the mobile phase the peaks would resolve. We now have run several assays with out premixing and they have all worked. I doubt there is a change due to the temerature. There has been changes in teperature during the year that have never effected the chromatography in this way. We typically do not use a column heater for this.

Uwe Neue, the reason you give for the premix is exactly why we have been trying premix. Some methods will have organic of only 4%. If this changes any direction it would drastically change the chromatography.

Thank you for all your assistance.

Your problem may be dwell volume-related and also due to some non-robust behavior with respect to gradient steepness. Are these small molecules or peptides/proteins? Since your instrument uses low-pressure mixing, the mobile phase may not have been "well mixed" for your gradient generated originally with buffer in A and ACN in B. Now, with premixes in A (27:73) and B (35:65) you may be delivering a slightly different %ACN to the column at the same times as before. It might be worth trying an isocratic hold with the initial gradient program (buffer A, ACN in B) to see if the resolution deteriorates. You could also explore different gradient times from 27 to 35% ACN to see whether resolution changes significantly (e.g., 5 min, 7 min, 15 min) for your critical pair. If you want to stay with premixed endpoints you may have to redevelop your gradient program to accommodate the "new" mobile phases.

Hello, the compresability of the mobile phases will have change now they have been mixed. This may be causing slightly different volumes of mobile phase delivery by the pump.