Webinar Highlights: “Developing, Testing, and Troubleshooting Chromatography Method Transfer”

By February 1, 2017

tomwheatHi, I’m Tom Wheat, Principal Scientist at Waters, and I recently hosted a webinar entitled, “Developing, Testing and Troubleshooting Chromatography Method Transfer.”

Over the course of the webinar, we discussed the transfer of chromatographic methods that occurs at many stages in the development and production of a product. In its simplest form, the method is moved from one laboratory to another within the same organization, holding all instrumental and chemical parameters constant.

More commonly, however, it is necessary to accommodate changes in available instrumentation. We also reviewed the steps in moving to new instrumentation as well as the role of information rich detectors in confirming a successful transfer that preserves analytical quality and accuracy as well as the more challenging process of migrating a method to more modern instruments and column chemistry.

If you missed the webinar, you can watch the webinar on demand.

I addressed several questions at the conclusion of the webinar.  Here are a few more questions that we received but did not have time to answer during the information-rich hour.


Waters Webinar by Tom Wheat: Developing, Testing, and Troubleshooting Chromatography Method Transfer

Q1: What is the best way to determine the dwell volume? Will the same approach work for Waters or Agilent or other vendor’s HPLC systems?

The best way to measure dwell volume is use a strong solvent that is spiked with a UV absorbing material and to run a gradient from 0% to 100% over 10-20 minutes.  The offset from the programmed time corresponds to the dwell volume.  This approach will work for any manufacturer’s LC or UHPLC system.

The details of making this measurement may be found in the following application note:  Dwell Volume and Extra-Column Volume: What Are They and How Do They Impact Method Transfer.

Q2: How do you use MS with classic method mobile phases containing, for example, phosphate?

Many classic mobile phases include non-volatile buffers and other additives.  Such components interfere with the processes of ionization that are required for MS analysis. We have had good success using a small trapping cartridge or column to isolate the peak of interest.  That isolated segment is washed with an LC-MS mobile phase to eliminate the salt.  The material is then eluted with a standard LC-MS gradient into the MS detector.

For a detailed description of this configuration and process, please consult the following application note: Routine MS Detection for USP Chromatographic Methods.

Q3: Do you have to measure the system volume each time you start a new method?

It is not necessary to re-measure the system volume unless some major component of the fluid path has been changed.  By major component, we mean more than tubing, since that is only a few microliters per foot.  Realistically, a change in the mixer or perhaps sample loop, a change greater than 75 to 100 µL or more, would represent a major change that would justify re-measuring the dwell volume.  It is however, still true that each system configuration would always be constant.  So a laboratory might have a short list of dwell volumes for the instrument configurations that they use.

Q4: Do I have to worry about the dwell volume if I use RRTs?

Relative Retention Times do not perfectly correct for the differences among systems, for several reasons. In the best case, the analyte peaks elute in the center part of the gradient where the rate of solvent change is at a steady state.  This region of the chromatogram is then shifted as a unit to an earlier or later time, depending on the difference in system volume.

If the relative retention times are calculated relative to the column void volume, the shift in RRT can be noticeable.  If the times are adjusted to a reference peak that is also in the steady state part of the gradient, the shift will be much less significant. If the analyte peaks happen to elute near the beginning or end of the gradient, there will be deviation from ideal spacing and will reflect differences in the shape of the mixing volume.

Q5: Is it possible to emulate a flow cell of a different instrument? How does USP approach this detail in method transfer?

Detector cell properties are seldom explicitly described in validated methods or in regulations.  Most often, detector effects are implicit in the required characteristics of the chromatographic result.  The most common effects of the detector cell originate with the path length and with the dispersion inherent in the cell design that is mostly related to cell volume.  Path length, in and of itself, mostly affects signal intensity, or peak height.  That parameter may be specified as part of a method requirement, either directly or indirectly with signal-to-noise.

Post-column dispersion in the detector cell can also be quite significant in affecting resolution and peak shape, as well as signal intensity.

There is no straightforward way to emulate detector behavior.  If the detector is to be replaced in the method transfer process, first try to keep the path length the same, then match the cell volume as closely as possible.

Q6: You said it is allowed to use different size particles by method transfers according to USP. How about the different packing materials of modern columns?

Allowable changes in particle size presume that the base particle and the surface chemistry are the same across the different particles.  Any change in column chemistry will be accompanied by changes in the selectivity of the separation for most sample mixtures and methods.  Most column manufacturers make some attempt to provide consistently scalable.  For Waters, all of the modern packings are available with identical base and surface chemistry from 1.7-1.8 mm up to 5-10 mm.

Q7: What affects the baseline noise when transferring method to same type instruments?

More: How does the different type of mixers (100 mL vs. 350 mL) on the same type of instrument system affect the baseline noise? Any comments on the size of dwell volume for ensuring best mixing for best quality baselines, especially with gradients? On a number of LC / UHPLC systems we have had to add additional mixers for TFA containing mobile phases.

The details of the mixing device used on any gradient system tend to unique and specific to that particular solvent delivery system.  They are always evaluated in two different ways.

  • The first to be considered is the reproducibility and control of retention time as well the deviation of the delivered gradient from the programmed ideal.
  • The second criterion of excellence is the quality of the baseline as it affects peak detection and integration.

These qualities may have somewhat conflicting solutions, but in general, stable retention is easier to establish than flat smooth baseline.  It is, therefore, wise to establish an adequate baseline for detection at the required sensitivity. This will usually give sufficiently stable retention times.  For making adjustments, mixers are often described in an absolute volume, but the shape of this volume is often more important to blending.

The devices may be loosely described as small diameter/high length or large diameter/short length.

  • The former, also called a low common volume mixer, primarily blends by providing time for passive diffusion as the solvents pass down the length of the tubing. This design minimizes any distortion of the programmed gradient, but is the least efficient mixer.
  • The latter design, also called as large common volume mixer, provides more turbulent or chaotic solvent blending.  This mixer design gives the most efficient mixing but has the potential to alter the shape of the gradient, particularly near the beginning and the end.

To decide what changes to the mixing may be required, examine the baseline for evidence of a periodic ripple.  If the frequency of the ripple changes over the course of the gradient, there is almost certainly incomplete solvent blending.  The period of the mixing ripple gives an indication of how far separated the solvent pulses are.  That, in turn, gives a measure of the magnitude of the common volume required to bring those pulses together for complete mixing.

(As an added factor to observe, if the period of the ripples does not change with the gradient changes, the ripple is not likely a result of incomplete mixing, but it is rather a pulsation in the flow of liquid, possibly indicating a check valve or seal leak or the need for some compliant pulse dampening.)

Q8: Will the use of premixed solvents give a significantly different retention time compared to the solvents mixed using the UPLC system mixer?

In general, pre-mixed solvents will be very similar to blends generated by the instrument from stocks of pure solvents.  There will be differences observed for some solvent combinations and for some samples.  Remember that there are at least four reasonable ways to prepare 50:50 Methanol: Water.  Because solvent mixing is seldom perfectly additive and because solvents differ in density and viscosity, the effective solvent strength of each alternative will be slightly different.  This can be reflected in changes in retention time.  Such changes are likely to affect retention more than selectivity.  The performance of the method may not be affected, but some level of verification is required.

Q9: My question is specific for UPLC. I have 50 mm BEH and HSS columns, and during applying the sequence the pressure always increases and often gives over pressure although I apply a washing well.

This problem always occurs when I use any percent of methanol. What can I do?

This phenomenon is most often observed when the sample contains material that accumulates at the head of the column.  Such material may be particulate or it may be high concentrations of strongly retained material.

Similar effects may be observed when some component of the sample or diluents  precipitates when it comes in contact with the mobile phase, as for example a high concentration of acetonitrile meeting a phosphate buffer, or perhaps a very hydrophobic analyte in contact with a highly aqueous mobile phase.

If these possibilities can be excluded by, for example, running a series of blank injections containing just initial strength mobile phase, we might suspect that the system is contaminated or shedding particulates.  Sometimes there are damaged components, such as pump seals that are contributing small particles.  But somewhat more commonly, systems can be contaminated with growing microorganisms in aqueous solvent lines.

Q10: What is the allowable variation in peak retention time acceptable during method transfer?

Would you please address the statistical criteria to determine if LC results are demonstrated to be similar and transferred?

For this webinar discussion, we were focused on the transfer of established chromatographic methods.  In this case, the source method includes operating criteria, or system suitability criteria, that are established during the development and validation of the method.  Such measures include retention time windows, resolution specifications, quantitative precision and linearity, limits of detection and quantitation, and so on.

The transferred method must meet these criteria with the same rigor as expected for just another series of runs in the originator laboratory and instrument.  For results outside those criteria, some level of verification and revalidation is required.


Additional resources:

Categories: Technologies