The Search for the Ideal Detector

When developing impurity-profiling methods for pharma applications, detection and sensitivity can be challenging. On one hand, the high variability of physico-chemical properties of the analytes requires that we use universal detection systems, while on the other hand the expectations of regulators drive the need for higher sensitivity and detectability (for trace analysis, degradation by-products, and so on).

The pharmaceutical industry has been basing most of its release methods on liquid chromatography (LC) with UV detectors and gas chromatography with flame ionization detectors, with the occasional use of fluorescence and mass spectrometry. There is still a reluctance to introduce other “universal detectors”. Moreover, pharma’s “ideal” detector should be sensitive and robust, ideally showing a universal response independently of the properties of the sample. It should be able to detect and quantify all components, including unknowns for which no reference standards are available (which is typically the case for early phase projects). Ease-of-use, the ability to interface with high-performance liquid chromatography (HPLC) and ultra HPLC instruments, cost effectiveness, as well as a good understanding of the response curves, are fundamental for the purchase and implementation of a new detector.

The truth is, although we are promised great things, we are still waiting for a commercially available and truly universal detector!

The next best thing to a universal detector, according to instrument manufacturers, is an aerosol-based detector, but, in reality, this only delivers what it promises for detecting non-volatile components under isocratic elution conditions. While there have been significant advances in this field, there is still a lot to be understood in the responses obtained with light scattering and charged aerosol detectors – for example, gradient effects, detector settings, influence of compound properties on the detection. Interestingly, a few recent studies highlight that volatility of a compound is not sufficient to explain differences observed in detection. For non-volatiles and semi-volatiles, parameters such as molar volume and diffusivity, as well as net charge of the compound seem to play a role in detection. These new aspects still need further investigation and explanation by theoreticians and instrument manufacturers.

Understanding the response model of aerosol-based detectors under hydrophilic interaction LC (HILIC) conditions can frequently prove even more challenging. Analytes eluted in HILIC often require an alternative detection technique due to low UV absorption (sugars, aliphatic amines, lipids, amino acids). There is ongoing research in the field but so far, most publications describe work done under mainly isocratic elution conditions.

In my opinion, we still have a long way to go in understanding aerosol-based detection under HILIC conditions. A HILIC expert once told me “we never use gradients in HILIC, they are too complex”, showing that there is still a big gap to cover, model and understand, both in terms of separation and of detection. A universal response model for non-volatile molecules in HILIC is inherently more complex and needs to consider not only the gradient effects, but also the detector settings, the mobile phase interference, and the polarity of the analytes.

Manufacturers need to investigate the fundamentals of nebulization and particle formation further to design detectors with increased sensitivity and greater uniformity in their response. In some cases, it might not be enough to widen the dynamic range of the detector by data interpolation through algorithms introduced between the analog and the digital output of the signal. We need to detect more and better – and I don’t think that, in the case of aerosol-based detection, an algorithm (as refined as it may be) will save the day.

I am convinced, however, that when there is a better understanding of the response model for aerosol-based detection under a variety of analytical conditions, the pharma industry will be very happy to introduce them routinely. The need is there. In the meantime, we are still looking for that ideal detector...

This article was originally published in The Analytical Scientist (www.theanalyticalscientist.com), a sister publication to The Medicine Maker.