Varying Injection Volume

[kingm3]

I am currently transferring an in-process method to our production department. The method states the injection volume is 5µl and the sample should be diluted to approx. 1 mg/ml of the active peak. However, there is a practice in the department to vary injection volume rather than diluting the sample correctly.

I do not agree with this practice but I'm finding it difficult to come up with a convincing arguement . The instruments are calibrated once a year for accuracy at 50µl. Their arguement is that if I'm using 5µl and not 50µl than why can't they use varying injection volumes from 5-50µl.

If the injection volume is incorrect at 5µl at least it will have the same bias for all injections, not so with varying the volume. Is there any other reason why the injection volume should not be varied during a sample set?

Thanks,

Marion

[unmgvar]

kingm3

does your department performs a linearity check on the autosamplers as well?
if yes then in my point of view you can do variation in the injection volumes.
you might need as well to perform for some of the applications a multilevel calibration and not only a single level. up to your judgement.

in general an autosampler is more accurate then human hands. you will have a wider range of possible error with manual dilution, then by using a different injection volume (given you show that the linearity of the autosampler is higher to at least 0.97).
i stand exactely at an oposite point of view then yours.
if you can tell the autosampler to dilute for you, and do most of the sample preparation steps, then you should do it because it will in most cases be more accurate then a human, and will also free a lot of time of the analyst.

[kingm3]

unmgvar, thank you for your response.

I found this in another post "Changing injection volume is considered bad practice (CPG was being polite!), so injection volume accuracy is generally not relevant to HPLC performance (i.e., since you will inject the same volume for the calibrators and the samples, you don't care too much what that volume is, you just want it to be the same everytime)."

If you are varying injection volume with a sample set than how can you compare like with like unless you calculate concentrations exactly?

[MK]

i've also been thinking this after the thread referenced by kingm3 above.

Keeping the injection volume constant and by pre-diluting the samples (manually or by autosampler) generally gives better results. But why? Could you gurus give a simple answer

Either way, you can inject the same amount of active. Is it simply that, in varying the injection volume you'll load the column with different amounts of diluent?

[tom jupille]

Either way, you can inject the same amount of active. Is it simply that, in varying the injection volume you'll load the column with different amounts of diluent?

That's it. The fewer variables you have on the chromatographic system, the less the chance of problems.

That said, there is a difference between "acceptable" and "best practice". If you can demonstrate acceptable results by changing injection volume, then that approach is acceptable (I know that's a tautology; that's the point! ). I do see more potential for screw-ups with the varying injection volume, and would still maintain that it's not good practice.

[Roy]

We almost never change injection volume in a specified method for In Process Analysis and Its essentially a V V V BAD PRACTICE. I completely agree with Tom.

However It depends how you specify this in the method. If your method says that "Inject 5 uL if 10 uL doesnt produce a peak area of X..." then I believe agencies will be fine. If you method says inject 5 uL and you are injecting 10 uL then you will have to defend that.

Another issue is if your method validated/qualified with 5 uL injection then you ought to follow the same.

Thanks.

[Peggsy]

my thoughts........

if i was to calibrate the LC system with increasing injection volumes from a set concentration standard, wouldn't the regression equation correlation be directly related to injector accuracy??

If you have to stay with one set injection volume, then sample dilution may be required to get an acceptable analyte loading onto the column. What figures are thus available to show that sample dilution is of a different accuracy than varying injection volume with an autosampler???



Ciao
Greg