My boss wants me to look into ELS (Evaporative Light Scattering Detector) over UV detector for the prep system that we have.
As we all know, UV detectors are compounds biased, which means a small impurity having a strong chromophore could have strong UV detection. The manufacture of the ELS claims that the response is proportional to the mass passing through the detector. Is it true that the ELS detectors are independent of what the compounds are, and only is a factor of the mass amount passing through it?
Has anyone had experience replacing UV detectors with ELS and has a good opinion on it? I appreciate any input. Thanks.

ELSD is an excellent compliment to a UV detector but is not a replacement. Many compounds have a much higher sensitivity using UV relative to ELSD and, of course, the opposite is also true (Sucrose for example). As well an ELSD will only detect compounds less volitile than the mobil phase. This generally means that these detectors are not sutible for volitile compounds. As well, they do not support the use of phosphate buffer (basicly for the same reason). Although the response is proportional to the mass passing through the detector, each compound can have very different response factors. Many labs find it very usefull to run UV or PDA in series with ELSD. This is also a destructive technique, so for your use (prep), you will have to split the flow.

A.A.

Shirley;

I have done some evaluations of ELSD units. I agree with AA in the sense that these detectors are not really a replacement for UV detection. Keep in mind that the response is in most cases nonlinear, the sensitivity is somewhat low, and that work with highly aqueous mobile phases is difficult.

I would say that if you are working with compounds that, have little or no UV absorption, can be easily separated without buffers, in mobile phases that are moderately high in organic solvents, and at relatively high levels, then the ELSD can be an attractive option.

Benjamin

I have ELSD 2000 from Alltech, a replacement for early Varex model. I have found it useful for stuff without chromophores that have low refractive index responses. I have used it for triglyceride assays in consumer products and for nonionic surfactants; for those, signal to noise is substantially greater than RI.

hi shirley:
Based on our experience with a Varex Mark II (re: hplc of neutral lipids & marine steroids, the ELS detector has improved S/N relative to RI & UV @low wavelengths. In addition this detector is gradient compatible cf to RI. As said by others: sufficient volatisation of solvent & not solute is critical. Volatile buffers such as ACOH, TFA can be used at conc less than 0.1%. Optimisation of the gas & flow rate is mandatory as the droplet size affects the scattered light and this in turn will change the solute peak area.

For prep work, stream splitters are effective however this model cannot handle more than 1.0 mL/min max flow. Also, aqueous mobile phases can corrode the drift tube (heating chamber) in particular if the user has left mobile phase flowing when the nebulizer is off or the temp is not at T evap.

We routinely, clean the nebuliser & drift to minimise potential build up of non volatiles from the mobile phase.

Hi Shirley,

In addition to all the above, ELSD has been used a lot in combinatorial chemistry where people in general claiming having equal response with a 25% CV. That will depend of course from the volatility of your compounds of interest. If some of them are semi-volatile or if they can easily pass in the gas phase (I don't remember the English term) they will give decreased responses.

If you are also concerned about the majority/minority compounds, you may ask as I know that at least one constructor has automatic gain control in order 1)do not loose any compounds that are in smaller concentrations and/or 2) not saturate the detector (has a chromatographic quantifiable peak). In other words you can in the run automatically change the gain.

About volatile mobile phase additives you might see one of our last publications LC-GC Europe 15 (2002)98-102 entitled volatility evaluation of mobile/phase electrolyte additives for mass spectrometry. The volatility of the additives has been tested in relation to their ELSD response (or more precisly in the background noise they create). From all the compounds tested we kept 10 acids and 10 bases which are volatile up to 5 mM and tested them in mixtures in different concentrations. The results are presented in a semi-pyramidal colour table while depending their volatile the buffers has been divided in: 1)non volatile, volatile up to 5 mM, volatile up to 25 mM and volatile at least up to 100 mM.

You may find the article for free (pdf form) in the following cite:

http://lcgceurope.adv100.com/lcgceurope/data/articlestandard/lcgceurope/062002/9090/article.pdf

If you have any problems you may contact me I can send you by e-mail the pdf (much better quality than the one provided on line).

I can also give you the e-mail of someone that has done prep-ELSD in order to have some further tips about the coupling.

Hope the above helps,

Kostas

I reread my message and saw that I missed something, the 10 volatile acids and bases selected were volatile "at least" 5 mM and not "up to".

I apologise for that.

Kostas

Dear everyone:

I'm so grateful to all of you for the responses that I've got. Seems like everyone has a good experience with ELSD! I'll get all these together when consedering buying one. Thanks again.

Shirley

I have some experience in using ELSD for sugars analysis. The sensitivity is much better than that of RI detector. However, the linear range of the ELSD is much poorer that that of UV or PDA.

I would agree with the comment that an ELS is complimentary to UV (preferably a PDA). We have been using a Polymer Labs ELS after a PDA for 9 months now and have found it to work well from 100% water to 100% methanol. For optimum sensitivity, temperatures and gas flowrates need to be adjusted at the lower organic solvent concentrations. We routinely use 2.1 mm id columns, and one problem we have found is that, to get linearity in response over a wider concentration range, we have had to drop the gas flow rate to 0.3 to 0.4 SLM. This is below the minimum recommended for the ELS, but it works fine (our N2 cylinders last for much longer too!).

ELSD is a universal detector. The good thing about ELSD is that it can see everything. The bad thing is that it can see everything. They're not very sensitive (since they can see everything).

Think of it as a replacement for an RI detector. RI detectors are not amenable to gradients. ELSD detectors are. They are good complements to systems where there is no chromophore (UV) or the analyte(s) do not ionize (MS).

It is common now to see LC/UV/MS/ELSD systems.