I'm currenlty looking at a couple of ELDS: The SEDEX 55 or65 and the Polymerlabs PL-ELS-1000. I need to detect a degradation peak w/ no UV chromophore (main peak absorbs strongly @ 280nm, strange,I know. Just trust me on this). What are your opinions about the two brands listed above? Is there a superior brand that I didn't list? What's the 'real world' linear range of ELSDs? How reliable are ELSD's? Thanks in advance for your input.

The detection limits are between 5 - 200 µg, but don't take that for granted, it depends on structure and chromastography. I've tested only one compound on a PL-ELS-1000 and its calibration function was quadratic, so you might be able to work with linear calibration if it is within one order of magnitude. Please post your results, I would be interested.

Strange. Detection limits with glucose are touted at nanograms. A quadratic cal curve sounds as if the evaporation efficiency is concentration dependent or the buffer mix is not right.

Gregg,

There are two other ELSD brands, though not superior in my opinon. From my experience SEDEX ELSDs are the best.

ELSD is not a linear detector (at least not yet...), that is not a problem though as you may obtain "linear" curves by using bilogarithmical equations. In this way I obtain typically "linear dymamic ranges" of at least 2.5 orders of magnitude. Typical limits of detection are about 1 mg/L. An interesting recent paper from Cobb et al. demonstrated that by using capillary columns (0.3 mm i.d.) with ELSD you increase your mass sensitivity more than the theoretical one.

For your problem which is the detection of an impurity in the presence of the main component you may want to check for the software which is available from SEDEX guys which allows you to change the sensitivity of the detector during data aquisition allowing you to quantify both even if they have more than 3 orders of magnitude difference in their concentration.

ELSDs are reliable and very easy to use. Some people becomes frustating because they can not obtain good signal to noise ratios or good LOD for certain compounds by using ELSDs. This is mainly due to the lack of an ELSD "education" (for most of them).

Even today when people searching about organic solvents to use with the ELSD, evaluates them according to their UV cut-off and not to the solid residue concentration (the best thing of course is to test the ELSD background noise of the different solvents). Attention must be paid also in the mobile phase additives as well as their concentration. For a quick evaluation in these additives volatility you may see Petritis et al. LC-GC Europe 15 (2002) 98-102. Finally boiling point and molecular weight are not enough to determine solutes volatility. The vapor pressure of the compounds must be taken under consideration as well.

Hope the above helps,

Kostas

You should be able to see 5-10 ng on column easily. Of course this is compound and mobile phase dependant. As for the claibration curves, these dectectors are ALL non-linear, you will need to use a quadratic equation which should really not bother you, any decent data system can calculate quadratic curves without problem. Also, dont forget, no non-volitile buffers can be used.

Where was this paper by Cobb published?

This elsd from eisa seems to be very similar to the sedex. Halogen Light Source - long drift chamber. But I don't know. Does anyone have experience with them?
http://www.esainc.com/products/HPLC/Optical_Detectors/esa_model301.html

I have developed an assay for quantitation of DPPC related impurities. I use an Alltech 2000 ELSD. We tested Sedex, and Polymerlabs and found the Alltech to be more sensitive, stable, and versatile (offers an impactor mode where you can run at very low temperatures allowing for the analysis of volatile analytes). If you have the standard of the impurity then your all set. You will not be able to quantitate this impurity based on %area of the main peak like you would in UV since ELSD is not linear, you will need a std. curve. If you do not have standard material for this impurity you can try something else. I was able to demonstrate fairly consistent recoveries for multiple impurities when using solely a curve of the parent drug vs. a curve for each impurity. Since it is a light scattering detector response factors are not going to change as drastically as you may see with UV detectors. It's something you can test if you have standards for structurally similar compounds. Contact me if you want anymore info. on this.

-Daren

Anon. II

Its Cobb et al. The journal of microcolumn separations 2001 13 4 pp 169-175.

Kostas

Dave,

I think that in your rush to support the Alltech ELSD you made some false statements. The statement more “Sensitive, Versatile, and Stable” is an Alltech commercial statement that can be found here (http://www.alltechweb.com/elsd.asp) in the third line, however it refers to the comparison of the evaporative light scattering with RI and UV. All the ELSD are more sensitive, stable and versatile than RI and UV for non-volatile weak/no chromophore containing compounds. What stable would refer in relation to the other ELSDs? Furthermore, the statement that with the use of the impactor on/off the Alltech ELSD can analyse volatile compounds is false too. No ELSD can analyse volatile compounds (if it could it wouldn’t be able to operate as it would detect compounds like CH3CN, water , MeOH which are volatile and used as the mobile phase).

The impactor has been developed to compete with the nebulizer chamber the other ELSD are using. Personally I do not think that the impactor is an advantage. What option do you choose when working with gradients from low to high organic (i.e. combinatorial chemistry) and/or when you analyse both non and semi-volatile compounds and/or you work from low salt in the mobile phase to high salt concentrations (ion-exchange) and/or when you change the flow rate during the analysis.

Finally, (for Gregg) and I think/hope that everyone agrees at that point, the best way to do it, is to ask from each constructor a demo for your application. This is the best way to evaluate them and choose the one that fits better in your analytical problem(s).


Kostas

Kostas,

I actually hadn't read the commercial you linked, it is ironic that the same words were used. I was refering to direct comparisons my group did when we tested the three different ELSD's I had mentioned. Stability was refering to baseline stability, sensitivity (direct comparison for analyte sensitivity), and versatile was refering to the Impactor "on" "off" options, which I do feel offers you greater flexability than other models. You are correct though, I meant to say semi-volatile not volatile.

I have tested the analysis of semi-volatile analytes such as palmitic acid methyl ester (M.P. 31 -34 degrees C) using a reversed phase method. Most detectors would require a 70 + degree drift tube for this assay, with the impactor "On" mode I was able to drop the temp down to 39 degrees and show accurate, reproducible recovery (I'm not aware of other models capable of running a 30% aqueous M.P. at that temp). With Impactor "Off" I had to run at 80 degrees and wouldn't even see the peak most of the time. So there are advantages depending on the type of assays you are interested in running. There are also some disadvantages, by running in Impactor "On" you decrease your sensitivity because only a fraction of the elluent is reaching the drift tube.

So you're right, the best thing to do is bring them all in house and see what works best for the type of applications you are interested in running.

-Daren

Kostas,

I have used the impactor "On" to analyze both semi and non-volatile compounds in the same assay and it has worked well, loss in sensitivity though for the non-volatile. Low to high aqeuous could potentially work in both modes, the manual would suggest impactor "On" for high aqeuous, but by optimizing the nebulizer flow rate I'm sure you can run in the "Off" mode. I'm not sure about changing the flow rate, would think in either mode you're going to disrupt your baseline noise/drift. I've never heard of being able to use salts for ion exchange chromatography with an ELSD. Do you have any links or literature references on this, is it discussed in your LC-GC article?

Thanks,
Daren

Daren,

First of all I am sorry for butchering your name, it happens to me all the time (other butchering my name) so I sould be more careful in the spelling of other names.

I think that we have a very interesting discussion here, so I'll try to reply in detail to everything you state or ask.

1) ELSD operation in high aqueous mobile phase for the analysis semi-volatile compounds.

Some interesting semi-volatile compounds to work with are the lipids, urea, Cl- and NO3- etc... I will based mostly in the analysis of Cl- and NO3 in order to reply in your question about ion-exchange chromatography (and all my examples will be taken out of the bibliography).

I think the best reference for this case is in Haddad et al. J. Sep. Sci. 2002, 25, 23-28. This article describes the ion exchange chromatography of inorganic anions by using ELSD (SEDEX one). As the ions Cl- and NO3- are semi-volatile their ELSD response drops dramatically with the increase of the drift tube temperature (there is a figure for this). As a result the operating temperature in that article was 40 C, and the mobile phase was 100% water containing 60-100mM NH4COOH (!!!). Furthermore, you may notice that the ELSD used is a Sedex 45 which is a 2-3 generation old Sedex ELSD (there exist now 55, 65 and 75).
My point here is that the Sedex ELSD has solved the problem of high aqueous mobile phase and analysis of semi-volatile compounds way before, the competition just try to reach what has been already succeded.
You said that you are unaware "of other models capable of running a 30% aqueous M.P. at that temp (39 C). I gave you an example with 100% aqueous, 40 C and high salt.
This is not an isolated case. You may see also:
Chaimbault et al, J. Chromatogr. A 829, (1998)193-199 (again for the analysis of inorganic anions etc.
I have personally operated it at 40 C, 100% water for other applications. When the solutes are not semi-volatile I use 50-60 C.

You may see:
Petritis et al:
J. Chromatogr. A 833, 1999, 147-155,
J. Chromatogr. A 855, 1999, 191-202,
J. Chromatogr. A 870, 2000, 245-254,
J. Sep. Sci. 25, 2002, 593-600
J. Chromatogr. A, 957, 2002, 173-185
J. Chromatogr. A 961, 2002, 9-21

In all the above examples the ELSD was used isocratically at 100% water (containing acids as additives) or in gradient elution which started at 100% water for the analysis mainly of amino acids and peptides.

2) Ion-exchange chromatography and ELSD.

You asked about references in the coupling of ion-exchange chromatography with the ELSD. I sent them in my next message. The article in LC-GC in fact uses the ELSD to define the volatility of acids, bases and their corresponding salts (in 100 % water) in order to make a list of additives that can be used with mass spectrometry and most importantly up to which concentration. It does not covers any ion-exchange issues.

The references that I send you in my next message are from my personal reference database relative to the ELSD which I try to keep updated (from 1966 which was the first article in the ELSD up to now) and that recently reached the 500 articles. I have included only the references that had the word ion-exchange in the titles, but I am sure that have been used in other applications as well (especially for sugars) it just does not appear in the title.

Having (lets say almost) the entire bibliography in ELSD I have also noticed that the crashing majority of articles using the Sedex ELSD which implies that either is the most popular, either the easiest to use, either is superior which will justify the whealth of results either whatever which Sedex is doing well.

I had some more thoughts to develop but lets keep it not too long.

I hope the above helps, I am ready to continue the discussion if necessary.

Best regards,
Kostas

Daren,

Here are the references I mentioned in my previous message.

Kostas

El Haddad, M.; Mouchere, F.; Elfakir, C.; Dreux, M. Conditions for rapid determination of inorganic cations in water by ion-exchange chromatography and evaporative light scattering detection. J. Sep. Sci., 2001, 24, 669.

El Haddad, M.; Lazar, S.; Akssira, M.; Dreux, M. Isocratic conditions for analysis of inorganic anions by ion exchange chromatography with evaporative light scattering detection. J. Sep. Sci., 2002, 25, 23.

Grard, S.; Elfakir, C.; Dreux, M.; Characterization of sulfobutyl ether-b-cyclodextrins mixtures by anion-exchange chromatography using evaporative light scattering detection. J. Chromatogr. A, 2000, 897, 185

Mercier, J. P.; Elfakir, C.; Dreux, M.; Lazar, S.; El Haddad, M.; Assira, M. Simultaneous determination of inorganic anions and methyl and ethylposphonic acids by anion exchange chromatography using evaporative light scattering detection. J. Liq. Chrom. & Rel. Technol., 2000, 23, 2345.

Mouchere, F.; El Gaddad, M.; Elfakir, C.; Dreux, M. Determination of inorganic cations and anions by ion-exchange chromatography with evaporative light-scattering detection. J. Chromatogr. A, 2001, 914, 167.

Qin, W. S.; Sadan, S.; Koropchak, J.A. Investigation of the condensation nucleation light scattering detector (CNLSD) for carbohydrates with ion-exchange hydrophilic interaction separations. Am. Biotechnol. Lab., 2001, 19, ?.

Stredge, M.A.; Stevenson, S.; Lawrence, S. M. Mixed-mode anion-cation exchange/hydrophilic interaction liquid chromatography-electrospray mass spectrometry as an alternative to reversed phase for small molecule drug discovery. Anal. Chem. 2000, 19, 4629.

Kostas,

Thank you for the references, I will check some of them out. I am interested in learning more about ion-exchange with ELSD. I was unaware of Sedex's capabilities at low temperatures. I knew the Sedex is the most popular, and wasn't sure why, our sales rep. probably should have shown us more of it's capabilites than he did, but at the time we were demoing we weren't focusing on semi-volatile compounds.

I was wondering if you have any information or references regarding the equivalent mass response for all compounds with ELSD's. I have found some and they state that you will generally see up to + - 15% difference in response factors for all compounds. I have tested this for same class compounds (phospholipids and their by-products) and found the statement to be accurate. Have you found this to be accurate, are there any cases with which this statement is inaccurate?

Thanks,
Daren

Daren,

I thought also that it might be the sales rep. that didn't handle the demo very well.

Now about the equivalent mass response:

The answer is yes and this capability is very popular in combinatorial chemistry. I include some references in my next message, again these are the one with combinatorial in the title, but maybe there are more. There are different statement for the ELSD equivelent response varying from 15-25 % depending on the library

The first one that used this ELSD capability was Kibey and it is published in the Molecular Diversity journal. A nice high-troughput application has been released in the ASAP articles of Cremin and Zeng. They used eight channel autosampler and 8 analytical columns, 8 ELSDs, a MUX eight way multiple channel electrospray and a TOF MS. The ELSD used at that study are Alltech (I do not want to know how much has been cost to use 8 of them).
Even here the Sedex guys have been shown viosionnaires. In order to face the combinatorial challenges, they came with the 754 model which is four detectors in one. They have four nebulizer chambers, 4 drift tubes, and 4 photomultipliers under the same instrument and (in my surprice) it is even more compact than their one channel detector (I think they must compact their one channel as well).

The problem in the equivelent response will be again the volatility (including vapor pressure) of the compounds of interest. A nice example is between fructose and anthracene.
Fructose has MW: 180, Melting point (MP): 130 C and boiling point (BP) 301 C. Anthracene has MW: 178, MP 218 and BP 271. Fructose ELSD response remains the same when you increase the temperature of the drift tube (i.e. 40, 60, 80, 100 C). However, anthracene while having about the same characteristic with fructose have a higher vapor pressure and its ELSD response drops when increasing the drift tube temperature.

Speaking about compounds equivalent response I think I should mention the chemiluminescence nitrogen detector from ANTEK with which you obtain nitrogen equivalent responses. The RSD with this detector is even better that the ELSD (less than 10%) but of course you have to have nitrogen in your molecule and you can not use the popular water:ACN LC gradients as you must not use any nitrogen containing solvents or additives in the mobile phase. Furthermore, it costs maybe 3 or 4 times the price of an ELSD (well maybe only two times more than an Alltech one ;-). If you are working though in combinatorial chemistry this instrument should be in your laboratory as well.

Hope the above are helpful to you,

Best regards,

Kostas

Daren,

Here are the ELSD-combinatorial chemistry references:

Kostas

Cremin, P.A.; Zeng, L. Anal Chem. High-throughput analysis of natural product compound libraries by parallel LC-MS evaporative light scattering detection. 2002, in press.

Goetzinger, W.K.; Kyranos, J.N. Am. Labor., Fast gradient RP-HPLC for high-throughput quality control analysis of spatially addressable combinatorial libraries. 1998, April, 27.

Hsu, B.H.; Orton, E.; Tang, S.-Y.; Carlton, R.A. J. Chromatogr. B, Application of evaporative light scattering detection to the characterisation of combinatorial and parallel synthesis libraries for pharmaceutical drug discovery. 1999, 725, 103.

Kibey, C.E. Mol. Div., Quantitation of combinatorial libraries of small organic molecules by normal-phase HPLC with evaporative light-scattering detection. 1996, 1 (4) 247.

Liling, F.; Jianmin, P.; Bing, Y.; High-throughput determination of identity, purity and quantity of combinatorial library members using LC/MS/UV/ELSD. Biothechnol. Bioeng., 2001, 71, 162.

Stromgaard, K.; Brier, T. J.; Andersen, K.; Mellor, I.R. Saghyan, A.; Tikhonov, D.; Usherwood, P.N.R. Krogsbaard-Larsen, P.; Jaroszewski J. W. Solid-phase synthesis and biological evaluation of a combinatorial library of philanthotoxin analogues. J. Med. Chem. 2000, 43, 4526.

Kostas,

Thank you again for all the references and your discussion on equivalent response, it is very informative. I'm not doing any combinatorial chemistry but the applications you've referenced on utilizing equivalent ELSD response will be helpful in seeing the big picture regarding this characteristic of ELSD's. I'm interested in and have worked on impurity profile and stability indicating assays with ELSD and use them to quantify unknown impurities or degradants. I have been working mainly with phospholipids and have found the response factors to be fairly close, and am curious to see how other compounds respond. I really want to find out if it's worthwhile (accurate enough) to use an ELSD method to compliment a typical UV impurities/stability method when you have an unknown impurity reaching significant levels. It could allow you to determine if the amount your calculating is actually high or low compared to the true value (determined on the ELSD). Could be very helpful when having to determine which unknown impurities to identify/characterize first during product development.

I had looked at the ANTEK nitrogen detector a few years ago and tried to influence my group director to shell out the bucks for but it was too costly. I still think it can be a very useful instrument in most labs, would've made my life easier on many methods I had to develop.

Take care,

Daren

I forgot to say something about the equivalent response...

If you are using a gradient elution from high aqueous to high organic you will propably have some difference in your response (higher ELSD response for the high organic due to the better evaporation of the mobile phase-more solutes arriving in the photomultiplicator).

I remember that Pfiezer has simulated this effect and had corrected formulas to apply in the obtain responses of the solutes eluted in different times). Another solution (that I have not seen been published yet) it could be to use another pump which you can use post-column with an inverse gradient from the one you use (from high organic to high aqueous). This will decrease the sensitivity (which is not a problem in combinatorial chemistry) but it will apply a 50:50 ratio of water:ACN and potentially improving the equivalent response.

Kostas

Hi Daren,

Just got your last message. I think it is worthwhile to use the ELSD for these purposes. Furthermore there is always the possibility (as mentioned from Gregg in the first message) that you might have impurities from UV absorbing compounds that they do not longer absorb on UV.

It would be interesting to have some feedback from you once you finish your study. I would guess that ELSD it will be good for most of the applications. The problem is that you can never be sure about the volatily/vapor pressure of an unknown impurity. For example from your phospholipids you might have as impurities some lipids that would be potentially semi-volatile, so your best bet would be always to operate in the lowest possible temperature.

For example you could make a run at 40 C, you have a number of peaks, you make the same analysis at 50, 60 and 70 C. The compounds which ELSD response will decrease they are semi-volatile and you can descard them from being quantified by ELSD, for the rest you can be pretty confident that the ELSD method would work.

I have something in mind that would maybe homogenise the response of such compounds but I'll keep it for the moment for myself.

Again, for the problems you mention an chemiluminescence nitrogen detector (CLND) would be a good idea. I have worked with it and have tested its equimolar response that I found exchellent - at least for compounds of the same familly (RSD of 6.3%, you may see LC GC Europe 14, 2001, 389-395.) As I prepared the solutions myself maybe there is a small error intoduced by manual handling of the samples. I also investigated the possibility where a non-nitrogen compound is co-eluted with a nitrogen one in chromatography and found that you do not have any changes in the CLND response.

Again let me know, if you come up with some interesting results.

Regards,

Kostas

Hmm, just thought about it...

You could also quantify the semi-volatile compounds by using a different standard (semi volatile) which will have the same or similar slope of ELSD decreased response when you make the study at 40, 50, 60, 70 C. You could make a library of these compounds and each time use the appropriate one.

Yes... that should work...

Kostas

Kostas,

Am I understanding your last comment to say that you could standardize one somewhat volatile component versus another, by matching volatilities of the two?

That could be very useful if true, but in my experience with ELSD there is also a "response factor" (to keep it simple) to consider. Different types of compounds respond differently. Perhaps this is true only for crystals, not for droplets? (In my experience, closely related compounds have the same responses.)

Bruce,

In reply to Daren I mentioned two factors that affect the ELSD response which is the volatility as defined when you include vapor pressure as well and the mobile phase composition (there are people out there that are working in the enhancement of ELSD response by using the right additives in the mobile phase). Some of the cases of the different "response factor" may be volatility issues as well.

I proposed a potential solution that can be useful when handling with compounds that are semi-volatile or that can give semi-volatile impurities/metabolites. I have not test it myself. The semi-volatile compounds could be from the same family. But as you will much the slope of the ELSD decreased response you would maybe have the same "response factor".

I agree with you that closely related compounds (as in the case of a combinatorial library) have the same responses. I agree also that totally different non-volatile compounds might have difference ELSD responses.

It seems that there are no clear rules of what it would be the response of a compound of totally different structures. They might be the same, or may not. In this area I have an idea or two (and it could potentially used to solve the problem of the response difference of semi-volatile compounds in combination with low temperature) but as I said I would like to keep them for me for the moment. After all, (as everything in research might work or not).

It would be interesting to hear from you which class or families of compounds have you checked and how different are these response factors.

Thanks

Kostas

I smell vendors afoot.

Anonymous,

I gone through my (and others) messages again, and they may appear biased but I am not personally a commercial guy. You may investigate it or you may contact me personally for this.

Kostas

PS: In the same way I wonder if your message is not from a commercial in order to decrease the value of what has been said above (after all your message is an anonymous one...)

Kostas,
A humorous comment on my part. Anonymous is easy because I lost my password. I depend quite heavily on vendors for informed opinions. No big deal. Vendors are very welcome in my world.
Aside from that, I am very interested in ELSD because of it is is more sensitive than ri and allows gradient chromatography. Fatty acids can be determined without derivitization (if the volatilities of the acids are low enough), same with amino acids, proteins, etc.
Back to the original string. There seem to be a couple of analytical philosophies for ELSD - regular research and high throughput for the pharmaceutics. I do not know if something is lost from one end in order for there to be a gain with the other - i.e., quick turnaround versus sensitivity and accuracy which has been a trade-off in other technologies. That can be best addressed by knokwing your specific needs and asking pointed questions of "the vendors" or squeeze it out of specs and app pubs.
ves shirley - aka "anonymous".