Chromatography Forum

Hello,

I am a novice in area of LC-MS/MS. For couple weeks I am trying to learn the technique without any specialized supervision. I have optimized my compound using flow injection (automated optimization). The system was tuned by a specialized personnel from sciex. I know my parent ion and most abundant daughter ions. However, I have not bought the internal standard yet. All I am trying to do is, making a calibration plot with my standard sample. My concerns are:

1. I am running MRM scan with dwell time of 0.51 msec and running it for 450 cycles. I am running sample with autosampler but I am using a zero volume connector (no column). For the quantification, I am using the automatic qunatification tool (Analyst) and it is giving me variable peak areas for my replicate samples. Moreover, I should have observed an increase in peak area with increasing standard but I am not observing any trend.

Is this beause I am not using any internal standard? Is it necessary to run internal standard with sample in LC-MS/MS?

No. It is more likely you have a 'very' unstable molecule. What are you trying to analyze? What is your sample matrix?

0.5 msec is just about the fastest you can run each transition, and probably part of the reason you have unstable results. You really should be trying this with a column, even a short 50mm column will give you better peak shapes and better reproducibility. If the direct inject without column is giving too narrow of a peak, you may not be seeing it at the peak maximum which would be the reason for the variability.

Ideally you should be at 5-10msec dwell times at least, and should have at least 6 scans across the peak 10-20 is even better to give good integration results. Internal standard is not needed unless you are trying to compensate for matrix effects.

Are you 100% sure you're actually monitoring the compound you think, and not some background contaminant? Flow injection analysis is fraught with traps because there is nothing to separate contaminants from the target compound, whether the contaminants come from your sample or the autosampler. If the peak area doesn't increase with concentration in a series of standards I'd be very sceptical. You mentioned that the optimisation was done by a specialized person from the manufacturer; if that person was a development chemist, then the MRMs are probably good. If it was an engineer providing training on how to use the system, it is possible they concentrated on demonstrating the principles of how to set it up, rather than thinking about what might go wrong (no offence intended, but my experience is that engineers are very, very good at diagnosing faults, but are aware of only very limited aspects of analytical work, while manufacturers' applications chemists and demo chemists are often very experienced and know every possible pitfall of setting up a new method). Good luck!

Good thoughts from others so far on increasing the dwell time and using a column.

Another thought: What's your sample concentration? You may be surprised that the linear working range is lower than you think. For example, I'm working on an assay for a compound right now with an upper linear concentration at about 200 ng/mL (2 ng/injection). At around 80,000 ng/mL, the response pretty much flattens out and is the same if I inject even more. If you're running at analyte concentrations typical for UV analysis in the hundreds of µg/mL range, this may be something to consider.