The general answer is that detectability is ultimately limited by signal/noise ratio. There are only two ways to improve that:
- more signal
- less noise.
Under "more signal":
- inject a 10 x larger volume (this works only if they define detectability in terms of concentration)
- get a 10 x smaller column volume
- get a 10 x longer flow cell
- decrease retention (will help, but not 10 x)
- use smaller particles (will help, but not 10 x)
- find a better wavelength (maybe)
- make a more detectable derivative (maybe)
- find a more sensitive detection technique
Under "less noise":
- improve instrument performance (lamp age, degassing, temperature control) if system not meeting specifications.
- detector time constant (works for short-term noise only)
All of the above may have ramifications in terms of affecting other characteristics of the method, and may not be permissable in the case of a validated method.
