Column Efficiency and Assay %

5 posts / 0 new
Last post
Lysander
Lysander's picture
Column Efficiency and Assay %

Hello there guys
Hope all is well
Had a question-
Is there any correlation between Assay% and higher efficiency ( Theoretical Plate numbers)
I'm asking this question because recently I bought a Phenomenex 100 mm x 2.6 micron x 4.6 mm Hilic column and it gave me plate count in ballpark of 10000. Same analysis on a different column was giving me a plate count in 2500s-
the manufacturer people say; more efficiency correlates with more precise results.
Why I did raise this question?
because; I'm getting a 0.1% higher number on my assay-

Dr. Analytical
Dr. Analytical's picture
Lysander:

Lysander:
No, there should not be a direct correlation between assay and efficiency.  The efficiency will only determine the peak width.  Sometimes there is a hidden peak which is not resolved at lower efficiency.  However, in this case the new assay will be a lower value.

Precision might be better if the resolution is better, but in general precision is determined by the injector and detector reproducibility.  With more narrow peaks sometimes the integration is easier, but that is the only other factor that could influence precision.

What you have is an accuracy difference.  First I need to ask if this is a significant difference, based on your known standard deviation of analysis?  Have you analyzed the same sample using both columns on the same instrument?  If this is significant, then the first place to look is the integration baselines for both the old and new columns.  There may be some very small changes.

If this doesn't help, write back and we can try some other ideas.
 

Lysander
Lysander's picture
Dr. A

Dr. A
Like always; thank you so much for replying sir

I ran the analysis on the same column in two different mobile phases

One assay was giving me  2.84 % average with ST. DEV of 0.05 @ 1.6 minutes -210 nm @1ml/min
The other assay was giving me 2.96% average with ST.DEV of 0.03 @ 4.3 minutes @210nm@ 1ml/min

Old 30 cm x 5 micron x 4.6 mm  column was giving me a 2.8% average as well- But the peaks were tailing. So, I thought perhaps the column is aging. Hence; I bought the new 10 cm column with 2.6 micron x 4.6 mm  dimensions from Phenomenex- This is the one which I ran the above two analysis described above.
 
In these two analysis, I ran the same sample on the same  phenomenex column on the same instrument-
Outside lab reports 2.8% - which is about what I'm looking for.
Calibration curves are
y = 1.2462x + 1.08 for 2.84% average assay and R square value of 0.9999
and
y= 1.2944x + 2.32 for 2.96% average assay and R square value of 0.9999

I don't think it's a significant difference, I'm justt curious whether the data is reliable on the new column.
The SST for second test is passing-The first one gave me very very low Theoretical Plate counts
That's why I decided to go to second method with retention time of 4.3 minutes.

Both baseline integration seems to be ok. The instrument goes to Autozero mode every time before the injection.

Thank you in advance Dr.A

Dr. Analytical
Dr. Analytical's picture
First, you should be able to

First, you should be able to run your new 2.6 um column at  a flow rate of about 2 mL/min.  When you decrease the particle size, you increase the flow rate.

The 1.1 minute peak is very close to the void volume, and analysis is always difficult.  I would try a faster flow rate using the second set of conditions.

For the 2.96% data, your slope is slightly larger and the intercept is twice as large.  This curve will always generate larger values for a certain peak area.  So, first you should look at why the curves are so different. 

Lysander
Lysander's picture
Dr.A

Dr.A

I would've gone to 2 ml/min- but @ 1 ml/min  the back pressure jumps to 360 bars- I guess, there is buffer being accumulated at the top of the column- So I will give it a good wash.
My machine can only go up to 400 bars max. It's a Dionex P680 Low Gradient Pump.
As far as the difference between two slopes and intercepts- I will try to investigate the problem

Thanks Dr.A