Assay and impurities in the same injection

[dcobice]

Hi all,

My purpose for this e-mail is because I need any help about a HPLC method development.

I worked with a C18 HPLC platform using a common 5um X 4.6mm X150mm, my question is about the LOD and LOQ.

I tried to quantify (in the same injection) assay and impurities and recently I have problems with the LOD.

My sample preparation was 20mg / 50ml of sample diluent (80/20 methanol/water +0.055 of TFA) and the abundance of the detector was about 2.1Au in a water alliance HPLC (what do you think about this?) and then I want to have my LOD close to 0.01% of the drug substance (assuming that the response factor were the same) and I couldn’t see anything or very tinny peaks (the relation of signal/noise was under 3).

I thought that it was a compromise relation about the saturation of the detector (2.1 AU) and the LOD.What do you think about??? Will be possible to do assay and decomposition in the same injection???.

See the attachments below (chromatograms of LOD and LOQ and sample (25mg/50ml) 100%.



The HPLC parameters are:



Chromatographic Conditions:

Guard Column: None

Analytical Column: Inertsil ODS-3, 150 mm x 4.6 mm, 5-micron particles

Flow Rate: 1.50 mL/minute

Column Temperature: 35° C

Detector Wavelength: 220 nm

Injection Volume: 10 mL

Analysis Time: 25.00 minutes

Equilibration Time: 10 minutes

Needle Wash: 10/90 (v/v) water: methanol with 0.05 % trifluoroacetic acid

Mobile Phase: A: 75/25 (v/v) water/methanol with 0.05% trifluoroacetic acid

B: 20/80 (v/v) water/methanol with 0.05% trifluoroacetic acid

Sample Diluent: 50/50 (v/v) water/methanol with 0.05% trifluoroacetic acid




Gradient Profile
Minutes Curve Mobile Phase
%A %B
0.00 Linear 100.0 0.0
Hold time-Dwell time Linear 100.0 0.0
25.00-Dwell time Linear 0.0 100.0
25.01-Dwell time Linear 100.0 0.0
35.01-Dwell time Linear 100.0 0.0





Thank you in advance



Regards

[unmgvar]

dcobice
i believe that first you must check the linearity of your UV detector. from what i remember for most detector the linearity range ends between 1.5 to 1.8 AU. higher then that and your result are questionnable.

after you have established that you need to find what concentration of your STD gives you an apex ABS that is close to your detectors's maximum spec of the linearity range.

for this example let's say that you can safely work at 1.5 AU.
this will mean that your LOD will be around 0.00015 AU= 0.15mAU.

now to see if this can indeed be your LOD you need to see your S/N ratio.

if 0.15/noise of chromatogram>= 3 then you are OK. you can set it as your detection limit.

this means that the noise levels of your aplication cannot be more then 0.05 mAU basically.
you will need to have an SST check all the time to make sure of that.
you must inject a Detection Limit sample, and make sure that the S/N you get there is always >=3.

in the end use the peak height or area to disregard all "peaks" in the chromatogram which are smaller then your settings.

if in your combime method you still see no imp. in a satisafactory way then you will need to inject your sample in overload to inhence the imps signals, and ultimately won't be able to combine the 2 tests together.

[yaoguocan]

In my opinion,accuracy of LOD depends on detector ,mobile phase.take HAC whose noise is louder than H3PO4 buffer .
but I dont know whether so-called noise includes whole baseline or peak around only? and how can we make difference bewteen min-peak and noise?

[tom jupille]

In my opinion,accuracy of LOD depends on detector ,mobile phase.take HAC whose noise is louder than H3PO4 buffer .

It also depends on the age of your lamp and the pulsations of your pump.

but I dont know whether so-called noise includes whole baseline or peak around only?

In principle, the noise in question is the average noise on both sides of the peak, but obviously not including the peak itself. You can sidestep the entire issue by using the slope and standard error of the calibration line.

and how can we make difference bewteen min-peak and noise?

You measure the retention time and peak width using higher level standards, then ignore the region containing the peak when you measure the noise. Or, as suggested above, use the statistical definion of LOD.