Dear All,

I just have some general questions that hope everyone could advise me on:

1. When one start developing a certain method for a compound, do you basically repeat the published methods and try to see if it works? Or do you just start from scratch (using guidelines from books like "Practical HPLC Method Development")?

2. Have anyone tried repeating the methods with similar (or same) hardwares, but cannot perform as the published methods stated?

3. In regards to publications, can people submit an article that only small modifications have been made to a previous method?

4. During a method development procedure, do we all try by "trial and error" (like using methanol versus acetonitrile) or is there any ways that we could do it systemically? Because I have asked some other people, and all they said is "By experience".

5. When developing an LC/MS method, do we follow the same rules as developing an LC Method? (I know that we have to use volatile buffers). How about the ionization theories?

Thank you!

Whew! very good (and very deep) questions that require long, detailed answers. I'll give you my brief opinion (I'm sure others will chime in!)

1. Unless there is some obvious downside to the published method, you have little to lose by trying it. If it works, you have saved yourself considerable time. If it *doesn't* work, then trying to tweak it can be a black hole that absorbs time and effort; at that point, I suggest starting over using some sort of systematic approach.

2. Failure to reproduce published methods happens all the time. Reasons can range from hardware differences to column variations to differences in mobile phase preparation techniques to . . .

3. Are minor variations pubishable? Depends on the journal and on how minor the variation is. If you submit an improved method for dimethylchickenwire analysis to the "Journal of Applied Chickenwire Research", it might well be published as being very useful to the readers. The same paper would probably be rejected by a chromatography journal unless it contained information that could be generalized to other separations.

4. "Successive approximations" sounds more scientific than "trial and error". Method development should be done systematically, but the decision taken at each step should be based on the total body of information available -- including experimental data, previously published results, and, yes, experience with similar separations.

To take your specific example (ACN vs MeOH), there is no way to tell *a priori* which will provide the better selectivity for a specific sample. In many cases, ACN will be the first choice (lower UV cutoff provides more detection flexibility, and lower viscosity means lower back pressure and slightly higher efficiency), but MeOH is less expensive and provides somewhat better solubility for buffers. If you're clever about setting up your experiments and if you use the right tools (I'll put in a shameless plug for our DryLab software here), you can tell whether a given solvent will work with only two or three experiments. If it doesn't, don't hesitate to switch.

I'm obviously oversimplifying here -- we (LC Resources) teach a two-day course on method development that focuses on establishing and implementing strategies -- but I hope you see the general approach.

5. Same general approach: identify which parameters affect the separation, rank the parameters in order of expected effectiveness, and then vary each parameter in that order. The details change (generally less resolution in required for LC-MS versus LC-UV methods, buffer volatility and ionization issues put some pH ranges out of reach, etc.,) but the principles are the same

-- Tom Jupille / LC Resources Inc.

I have just a little elaboration to add to the
previous message. Typically, when you are
using positive ion LCMS, you should use and
an acidic buffer/mobile phase. When you use
negative ion, you should use a basic mobile
phase. This will increase ionization
effeciency. Of course, there are exceptions.
To find examples of this, look up papers that
discuss "wrong-way (round) electrospray".
Quite often, when people are trying to analyze
a compound in negative mode, but they have
an acidic mobile phase, they will use a pump
to introduce a base between the column and
the MS.

The other thing to consider when converting
from LC to LC/MS, aside from the buffer
volatility which you mentioned, is that too
much water in the mobile phase can
adversely effect ionization. This is largely due
to surface tension and conductivity.

Thank you for your elaboration. I am currently working on a method to analyse 2 drugs. One has a pKa of 6.6 and the other has a pKa of around 8.0. According to your experience, should I use a basic mobile phase (using ammonium bicarbonate as the buffer) to improve the peak tailing problem? It is because I had been using an acidic buffer (with ammonium acetate) and there was an immense tailing problem. (Our column could tolerate the high pH as well). But the MS engineers were telling me that it is hard to quantify stuff using the negative mode (because of the specs and stuff). Since I am a new MS user, is it true? I have tried infusing the standards (diluted in 70% acetonitrile) into the MS using both positive and negative mode, but there seems to have no big difference between the counts. Is that the right way in checking should I use the positive or negative mode?

Please also recommend some papers regarding what you said about the "exceptions" in the buffer and the polarity of the mode.

Thank you. Looking forward to your reply.

You did the right thing. If you are getting good response and little noise in either mode, either will be good to use. Never mind what the engineers are telling you ...

I don't know of any references for the
exceptions off hand, but I have seen them, and
a literature search from the last 3 years or so
should give you a few refs. If you think
ammonium bicarbonate will be your mobile
phase, inject your standards in that solution.
You can try this in both positive and negative
mode, in case your compounds are an
exception to the rule. I am not sure what your
analytes are, but don't listen to the engineers.
I have found that a lot a people don't know as
much about LC/MS as they think they do, and
they present word of mouth as if it were a fact.
The only way for you to find out how things
work and to learn to pick out patterns, is
through trial and error. Keep in mind that
whether your analyte is detectable in positive
or negative mode is dependent on the
functional groups, not the pKa. The pKa is
more of a factor for the chromatography,
although sensitivity can be influenced by the
pH of the mobile phase.

And don't worry. I know the learning process
can be a little daunting. Two years ago, I was
supposed to be trained in LC/MS ( I only had
LC experience), but the person who was
supposed to train me was fired before I had a
chance to learn anything. Now I know more
about LC/MS than most people. You'll get the
hang of it.