Creating a New HPLC Method

Sometimes a new method is required. This may be for a product which has never been analysed anywhere before, or simply for an application which has not been done in your particular lab. Either way, a method must be sourced or prepared, and then it must be implemented, optimised and validated. This can be very time-consuming and without a high level of competence in HPLC, can lead to a method which works now but gives problems later

We have quite large libraries of methods which have been done before. So the first way we may be able to help is to avoid re-inventing the wheel! If a method cannot be found, trials need to be undertaken to establish a starting point for method development and optimisation. It is helpful to have a diode array detector or LC-MS to assist in peak identification and tracking, and it may be necessary to try a number of different columns. For some applications (eg sugars by ligand exchange, or GPC) the columns can be around £900 each, so this can be a very expensive process.

Laserchrom has hundreds of columns and almost every type of detector available. So the expense of trying these options is not so great. In the main, new detectors and new columns do not need to be purchased for evaluation.

Method development can be very time-consuming. There are many columns to try, and an almost infinite combination of eluent, pH, temperature, etc to choose from. This can mean doing a very large number of analyses, and may still not find the optimum conditions. We have Drylab software, which allows us to optimise all of the above and more in the minimum of runs, and provides very accurate predictions for the otpimum conditions. So again it is most likely that we can provide not only a better method at the end, but in much less time, and for much less money.

At Laserchrom we have 27 different reversed phase columns from Dr Maisch, Germany, along with four different types of polymer-based reversed phase columns from Shodex. We can evaluate any or all of these for any method as necessary, along with many others for other applications. We have many normal phase columns, along with ion exchange, ion chromatography, ligand exchange for sugars, ion exclusion for organic acids, size exclusion, and mixed mode columns for more complex samples.

Method Optimisation will include the type of oprganic solvent, the %B, temperature, pH, buffer type, buffer concentration, column type, detection mode and detector settings (eg wavelength, voltage for ECD etc)

We are also able to identify peaks accurately, and optimise detector conditions for maximum sensitivity. Once a method is established, we can then carry out a validation exercise, before bringing the method to your lab to set up on your equipment, and we can provide on-site training to ensure that the results work for you too.

We are happy to discuss your requirements, and if some parts of the process can be done in your lab, we will be pleased to work alongside and fill in the gaps where you either don't have the time, columsn or equipment to do the work in-house. Or we could develop the method and hand it back to you for validation.

Sample pretreatment or extraction can be part of the work we do, or if you feel confident to undertake this, we can work with cleaned up samples ready for injection.

Once the method is transferred to your lab, we provide telephone technical support free of charge, and on-site support, training and troubleshooting for a daily rate.

Surfactants used to clean up contaminated soil.

The problem was that each of the two surfactants contained maybe 50 peaks. Under most elution conditions, these peaks had considerable overlap from one surfactant to another and there was two surfactants in the mix to give the required hydrophilic/lipophilic balance. One surfactant was not soluble at all in water. And in amongst it all was trichlorobenzene, which was being extracted from the contaminated land. We needed to be able to quantify each surfactant, so that the extracted surfactant mix could be cleaned up and re-used, with the right concentration of each solvent so as to achieve the correct balance for further extraction. We had five days to complete the method!

After trying various columns, we found that each gave a large number of peaks, and it was very difficult to be sure which peaks came from which surfactant  However we tried a mixed mode column which reduced the selectivity such that each surfactant gave only three peaks. It worked with enough acetonitrile in the elunt to keep both surfactants in solution. Of the three peaks for each component, two of each overlapped, but the first peak from one surfactant, and the last peak from the other we well resolved. We ran a trial calibration over the concentration range required and found we had a perfect linear calibration. By changing the wavelength we were also able to quantify the trichlorobenzene which eluted five minutes later.

Analysis of an additive to lubricating oil.

We were supplied with the additive, and quickly found a method for its analysis. But when we came to analyse the lubricant with the additive in, we found many peaks, eluting all over the time window where our additive eluted. The oil contained many non-polar species which would bind irreversibly to a C18 column so an extraction was necessary. We also discovered there were eight other additives in the lubricant.

We tried an acetonitrile extraction of the oil, and found that the additive was quantitatively extracted, but along with a large number of other species. Methanol hardly extracted any of our target material! So we went with the acetonitrile extract, and tried various columns to get a separation. It is normal with reversed phase analysis to get more selectivity from changing the eluent than from changing the column, but even so, after trying eigtht columns we were unable to find one that would separate our additive from the other components of the extract! Then we tried a phenyl column, and began to get separation. After 2-3 days of changing the gradient profile we found conditions which gave us a neat window amongst all the other peaks, and managed to slot the additive into the window. However even then, the column overloaded with the non-polar components, and even a guard column only lasted a few injections. A 30 minute hold at 100% acetonitrile managed to clear the column, but the run time was then over an hour. However we switched to a THF water mix, and because of the increased eluent strength of THF we were able to clear all components from the column without even getting to 100% THF, so the long hold was eliminated and we had a really neat separation offereing quantitative results down to very low levels of additive, and with good column life and no need for a guard column.

Analysis of Aspirin from a plastic material used as a surgical implant

The analysis of acetyl salicylic acid (aspirin) is well documented, and the customer had a method already. However it was giving very low and unreproducible results. Their run time was 33 minutes (much too long for just three peaks) and their extraction process included a step taking the material up into chloroform and drying off the chloroform.

First we changed the column and redeveloped the method to take 3 minutes. This was important because aspirin degrades noticeably over an hour or two. So the sample runs needed to be short, and the autosampler chilled to 4oC. We ran stability tests, injecting the same sample over and over for several hours to quantify the degradation. The second change was to use dichloromethane as extraction solvent, because its boiling point is so much lower than chloroform.  Finally we discovered that the customer's method used pH2.5, and at pH2.7 salicylic acid and acetyl salicylic acid co-eluted. Above pH 2.7 their positions were reversed. So we changed the pH to 2.0, well clear of the danger zone, and emphasized the importance that accurate pH measurements were taken.

Biogenic amines from foodstuffs

Much has been published for the separation of biogenic amines, but there seems to be little concensus, and the methods not only vary widely, but give different results! We tried post-column derivatisation with OPA, pre-column derivatisation with OPA, ion-exchange with post-column ninhydrin and with OPA, and on-column deriatisation using OPA.

The run times all ended up being about 30 minutes for standards, but longer for food extracts because of the gradient eluetion that was then necessary. The selectivity was best for the on-column method, but we found that because of such good selectivity of the biogenic amines, it made th sample extracts very long runs, and we were not confident of quantitative derivatisation. The pre-column method also gave excellent selectivity, but because the derivatisation was of the 'dirty' sample rather than of pure peaks, we felt less confident of the quantitative nature of the derivatisation, and we could show that the derivatised product started degradation within 10 minutes. After 30 minutes, only about 10% of the product was left, and this didn't fit well with a run time which may exceed 30 minutes for sample extracts. The ion exchange worked well, but to get reproducible results, a very expensive column was required, and although very stable and selective, we found the ninhydin lacked sensitivity for the levels typically found in the foodstuff in question. It also requires a long reaction time and 123oC, which takes the eluent over its boiling point, so a back pressure device was required after the detector. The preferred method was the post-column OPA, using an ion-pair reagent to reduce the polarity of the amines and give selectivity. At that point the acetonitrile shortage hit the world, so the meethod was evaluated using methanol or THF and combinations of the two, and optimum conditions found to givereally nice baseline resolution of putrescine, cadaverine, histamine and tyramine, and excellent linear quantitation.

Telephone:        +44 1634 294001 Ask for Dr Stuart Jones

Email:                 stuart@laserchrom.co.uk

Fax:                    +44 1634 297533