Sensitivity of a UV coupled to an HPLC

[aparna]

Our UV-Vis spectrophotometer has a sensitivity of 0.6 ug/mL of a compound of interest. For lower concentrations, we are considering using an HPLC which provides a sensitivity of 1-10 ng/mL.

How is a UV coupled to an HPLC able to give more sensitivity than a stand-alone UV?

Thanks in advance.

Aparna

[danko]

Which solvent is the compound of interest dissolved in when measuring with the spectrophotometer?
Type of compound? Are there any other compounds in the solution that might absorb light at approximately the same wavelength? Which wavelength are you using by the way? The same on both equipments?

Best Regards

[Uwe Neue]

Here is a reference worth reading: Encyclopedia of Separation Science, 2007, Pages 1-15, Author: A.C. Gilby (CHROMATOGRAPHY: LIQUID | Detectors: Ultraviolet and Visible Detection).

Some aspects: noise is determined by a bunch of different things such as light intensity, lamp noise, electronic noise, tools to suppress noise. For HPLC detectors, the most important task was the reduction in noise combined with a reduction in cell volume, while keeping path length constant and response time fast. I do not think that classical spectrophotometers were subject to the same natural selection

[tom jupille]

HPLC which provides a sensitivity of 1-10 ng/mL

Are you sure that wasn't 1-10 ng? (not ng/mL).

And when you say "sensitivity", do you mean "limit of detection" (signal/noise = 3), or "limit of quantitation" (signal/noise = 10)?

And are you referring to the sensitivity of the HPLC detector itself or to the sensitivity of the HPLC method?

I hate to sound pedantic, but all of those factors come into play when doing a comparison.

[aparna]

Thanks to everyone for the replies. The HPLC-UV allows injectable sample volume as low as 50uL, but a stand alone UV needs a lot more volume. Is it mainly due to the dimensions of the flow cell? Or is it the integration of absorbance over time at a specific wavelength that gives a stronger signal?

Thanks,
Aparna

[mardexis]

I can't think of any reason that an HPLC is going to give better sensitivity than a stand-alone UV. The HPLC will be much smaller volume so the absolute mass of substance detectable might be better with an HPLC but not the concentration based detection limit.

For a good HPLC UV unit you can expect an Absorbance of about 2x10^-4 at the LOD and for some sort of aromatic a molar abs of about 1 x 10^4 so you might expect to see an LOD of around 2 ug/L, neglecting dilution on-column. In real life you should expect dilution of about ten fold or so on-column depending upon capacity factor.

I believe a sensitivity of low ng/mL with UV-Vis is not possible. Is it possibly a fluorescence detector?

[aparna]

HPLC which provides a sensitivity of 1-10 ng/mL

Are you sure that wasn't 1-10 ng? (not ng/mL).

---I guess 1-10ng

And when you say "sensitivity", do you mean "limit of detection" (signal/noise = 3), or "limit of quantitation" (signal/noise = 10)?

---I mean limit of quantification.

And are you referring to the sensitivity of the HPLC detector itself or to the sensitivity of the HPLC method?

---I am referring to the HPLC method.

[tom jupille]

Okay, now your question can be answered.

Limit of quantitation is defined as the concentration (or amount) of analyte that will generate a signal 10x the noise.

Look at the noise specification for the spectrophotometer and the UV detector (probably something like 10^-5 absorbance units. I would bet that the spectrophotometer has a lower noise level. Because absorbance is a function of concentration, that means that the spectrophotometer will have a lower LOQ in concentration units (nanograms per mL).

However -- as marc (mardexis) pointed out, the spectrophotomer takes a very large volume of solution (a standard 1 x 1 x 5 cm cell holds 5 mL), so the LOQ in mass terms is not that great.

The HPLC detector is less sensitive in concentration terms (more noise) in part because it uses a very small flow cell (8 microliters is typical). That means it can be very sensitive in mass terms.

[cyanogen78]

[HW Mueller]

Some of this has been discussed some time ago. It is somewhat surprising that all the UV spectrometers I ever used go down to 0.001 absorbance units whereas my HPLC UV detectors go to 0.00001 absorbance units. There really seems to be an evolutionary cause behind this (Uwe´s statement).
But: There must be further reasons. The difference in light absorbance due to scattering of 1x1 cm Q cells (I am talking about high grade quartz cells, cleaned very well) has, in our hands, been on the order of 0.045 abs. units. This is scattering due to the quartz-water interface, if one has air in the cell the scattering appears to be at least double. Now one (we) has enormous problems getting a believable absorbance with a HPLC detector when one uses it in a stationary mode, for instance by placing a solution into the cell via a syringe, even if one gets all air bubbles removed. The reading will be quite different if you cause a flow. I suspect that scattering by dissolved air plays a role here. That is, I wouldn´t be surprised that one would need careful degassing and close control of interface scattering + ?? to get a meaningful use out of a UV-sprectrometer that had a higher sensitivity than what is normally seen in labs,

[danko]

This is scattering due to the quartz-water interface……

Hans, this is not scattering, but refraction. Scattering is a completely different phenomenon, involving stimulation of molecule vibration.

Best Regards

[HW Mueller]

Ugh, almost missed this. I suppose that refraction could be included in my "+??" above? I was thinking of scattering at the "microscopically" rough quartz surfaces (like scattering by snowflakes). One can see this if one opens the cell compartment of spectrometers while the light beam is on (don´t do this with UV!).

[Bruce Hamilton]

I used to use a Cary 19? UV/Vis ( 1980s ) that displayed absorbance to 4 decimal places.

One problem was that even clean, new, cuvette variations could affect the 3rd place, and the manager eventually taped black paper over the 4th figure so users wouldn't spend excess time trying to stabilise the display for blanks.

For UV/Vis quantitative work, absorbance to 3 decimal places was adequate for most method accuracy, as 0.1 mg balances, glass pipettes and volumetric flasks were standard techniques.

I suspect that 3 decimal place absornace specifications also allowed much cheaper optical systems, as that was a monster instrument.

I assume I'm wrong, but the fixed wavelength Zn ( 214 ), Hg ( 254 ) Waters 440, 441 detectors seemed to be much more stable and sensitive than the variable wavelength models that replaced them.

Please keep having fun,

Bruce Hamilton

[HW Mueller]

Bruce, that´s interesting, I guess one manages to adapt to the apparatus at hand. As mentioned above we had trouble with cells going into the 2nd place.
The trouble with single wavelength machines is that they are really usually more sensitive (I think they have a higher light intensity), but only very close to the specified wavelength.