Dear all,

I have a little physical/chemical question for you. A column (C18) has a maximum temperature limit for example 80°C and I will perform my analysis at 70°C (respectively my column thermostat is set up for 70 °C) with a simple mobile phase ACN/water.
1. 70 °C is rather high temperature. I suppose, it can probably contribute to column life time shortening. Should I fear a lot? What is your experience?
2. Is a combination of high temperature and buffer/ACN mobile phase (instead of water/ACN)much more fatal for column?
3. Some people say, that at such a high temperature this mobile phase must be almost gas. But ACN has a boiling point 81°C and I suppose, water/ACN mixture has higher boiling point. I also suppose, that there must be some heat loss during heat transfer in column thermostat. When the set temperature is 70 °C, the column itself has probably lower temperature. Maybe it is not so dangerous to set the temp. for 70 or even 80°C. What do you think?

I look forward to your discussion. Thank you.

Jitka

Silica column life will be shortened by use of elevated temperature but there isn't any specific threshold or anything like that regarding product life. The higher the temperature, the shorter the product life. Certainly, some eluents are more corrosive than others. For example, alcohols are more corrosive than acetonitrile and phosphate buffers are generally relatively corrosive. The further you steer clear of the limits of pH stability, the better (i.e. < pH 2 or greater than pH 6.5).

I wouldn't worry about "gas" in your column. Even if you set the temperature at 81 degrees C and the internal column temperature was actually 81 degrees C there is still no risk of formation of gas in the column. The boiling point refers to atmospheric pressure. Even a modest level of back pressure will prevent the eluent from reaching the boiling point in the column. 3 or 4 atmospheres of back pressure should be sufficient to prevent this problem in your case.

A single temperature limit of a column is plainly nonsense. Column stability depends primarily on the temperature and the pH, second on the buffer type and of course on the type of packing used. For silica-based packings, the best stability is probably around pH 4. The stability at pH 2 or less depends on the type of surface coating. The stability at pH 8 or higher on the protection of the silica. Hybrid packings are stable at still higher pH.

Since you are asking a question, column lifetime is propably not a disaster. Observe the changes in retention and selectivity! If they are too much, lower the temperature. If the lifetime is long enough, be happy and continue doing what you are doing.

In general, if column lifetime is limited by the dissolution of the packing, it will decrease by roughly a factor of 2 to 3 for every 10 degrees Celsius.

I would be curious as why you would want to run at such high temperaturers?
In general you would see an adverse effect on resolution, although I accept for some "odd ball" applications temperature can improve resolution.
Run time should be quite short and you could run at higher flow rates due to lower backpressure. Which might be an advantage.
Would you not be encouraging thermally labile material to start reacting with the column/ mobile phase at such high temps?

I'll argue with the "in general you would see an adverse effect on resolution" statement.

As you increase temperature, for most types of HPLC (yes, there are exceptions):
- retention (k') decreases, which will tend to decrease resolution, since resolution is proportion to k'/(1+k')
- efficiency (N) increases (higher temperature = faster diffusion = a smaller C-term), which will tend to increase resolution
- selectivity (alpha) *may* change, and this change has a roughly equal probability of either increasing or decreasing resolution.

However, when you are *developing* a method, you usually vary mobile phase strength to keep retention in a reasonable range, so the first part of the above paragraph really doesn't apply. Resolution is a function of the square root of the plate number, so even a relatively large change in N doesn't affect resolution very strongly. This leaves selectivity as the dominant effect; as noted above, this can go either way.

Of course, temperature is only one way of controlling selectivity.

I'll echo and expand Uwe's comment: If running at a higher temperature *decreases* resolution, don't run at a higher temperature. If running at a higher temperature *increases* resolution but decreases column lifetime, then you have to decide which is more important. I'm not sure that I'd want to tell a client: "The resolution is inadequate, but the column sure lasts a long time!" :)

-- Tom Jupille