Reproducibility of the Selectivity

We have pointed out above that the reproducibility of a stationary phase has some limitations. How wide the range is depends on the type of stationary phase, the synthesis conditions, and the manufacturer. We do not want to end this chapter without saying a few words about the reproducibility of methods and stationary 

The reproducibility of a stationary phase depends on the experience of the manufacturer and the quality of the test method for the stationary phase. By the latter, we mean the quality of a test of the selectivity of the stationary phase, not a column test method, which is commonly used for measuring the column plate coimt. The method that we have described here for the characterization of the stationary phases was based on a very sensitive test method for the reproducibility of a packing material. High-quality manufacturers of modern stationary phases now use such highly sensitive methods for the measurement of the batch-to-batch reproducibility of such packings. At the same time, one has learned that batch-to-batch differences are measurable and can be quantified. 

If one observes difficulties with the reproducibility of a method, one should not, however, immediately blame the column manufacturer. If the method in question is a new method, i.e., a method that has not yet been used on many different columns, it is not impossible that the column-to-column differences observed are simply related to the age and use of these columns. Often, a brand new column wiU give different results compared to an old, much used column. In order to test this, it is worthwhile to check whether the two columns in question have been prepared from the same batch of packing material. This information may be contained in the Certificate of Analysis that came with the column, or it can be obtained from the manufacturer. If the batch of packing material in both columns is identical, the selectivity of the separation must be identical. If the selectivity is different, one can be sure that the differences are due to the age and use of the older column. This difficult situation is not the fault of the column manufacturer, and cannot be resolved by the column supplier. There is no other choice but to redevelop the separation from scratch. Unfortunately, many reproducibility problems are of this type. If, on the other hand, the new column is from a different batch of packing material than the older column, then one should inquire whether the manufacturer can supply a new column from the old batch of packing. This is usually possible, unless the old column really is several years old. 

If this comparison of the selectivity of different columns does indeed demonstrate that the observed difference correlates with a difference in the batch of packing material, then there are two possible solutions to the problem. One possibility is to reserve a quantity of bulk packing or packed columns specifically for the customer or even at the customer s site. This is the best solution if the separation is needed only for a limited amount of time. If, on the other hand, the separation will be needed for many years to come, as is often the case with quality control methods, it is better to redevelop and harden the method. Often, only a small change in the mobile phase conditions is necessary to solve the problem. N ow that everybody is aware of the potential differences in the method on different batches of packing, one should check the reproducibility on several new batches of packing. Several column manufacturers offer reproducibility test kits that contain several different batches of packing and enable such studies. 

The most difficult situation for everybody involved occurs if a separation that used to run smoothly for several years on many different batches of the packing suddenly does not work anymore. This problem occurs very rarely with modem. 

---

Centrifuge down the culture at 10,800g for 10 min, and immediately precipitate the phage from the supernatant using PEG (see Section 3.4.). The phage, representing the library, should be used within 1 wk, or can be stored in aliquots at -20°C These aliquots will be all of the same quality and will ensure the reproducibility of the selection procedure. 

Two-component mixtures can be quantitatively analysed without a calibration chromatogram or calibration graph. Solutions of identical concentration are prepared of each pure compound and the UV spectra are recorded. If the wavelength at which the spectra lines intersect is chosen for detection (isosbestic point), then the peak area is identical with the mixing ratio of the sample, on condition that the UV detector used shows high stabiUty and excellent reproducibility of the selected wavelength (at least to 0.2 nm). 

---