Before suggesting any levels of uncertainty for anything, you should familiarize yourself with the analytical technique(s) and specific method(s) being used. Students new to these techniques should first review some of the classic texts on the topic (HPLC & GC analysis) , then have someone local, knowledgeable in these areas explain to you the immense complexity of using these techniques with a specific method to acquire data from actual samples. Only after understanding all of the variables will you appreciate how difficult it is to assign any uncertainty values to anything. In this case, such values can not be assumed.

A key point that you may not have learned yet in your classes is that MS detection methods (used with HPLC-MS or GC-MS) are not universal detection methods and do not "identify" anything on their own. If set up properly, they can suggest things. In fact, due to the many variables of the method which is employed to use them, sometimes they detect nothing at all or output data which is completely erroneous (you can easily choose settings and methods which result in invalid data). It is the use of multiple, orthogonal analytical techniques plus standards, coupled to many years of professional training that allow us to sometimes be sure of the results.

Any scientific measure of "uncertainty" needs to be described in detailed terms, not generalizations as you have. A 'brand' is not an instrument (it matters not that one system is an Agilent and another is from Shimadzu, or any other vendor). A 'technique' alone provides no context or information related to measurement uncertainty. Any quantitative values must be based on a specific detailed method plus real world experimental data. Values used to describe anything should not be assigned without such information.

It doesn't matter! The uncertainty in the method (extraction and analysis) will DWARF the uncertainty of the instrument. The question you should be asking is whether the instrument can do the job! Don't compare apples to oranges.

it is not comparing apples to oranges. but the software asking to set the uncertainty. However, we estimate our uncertainty annually using QC reading. Thank you for your comment

In evaluating the uncertainty of analytical results, it is necessary to evaluate everything from the point of sample collection to the interpretation of the results. At every stage in between sample collection and interpretation of results that can and will affect the results. It is not just the instrument or software. Evan in the case of measuring the mass of a certified standard on a high precision balance or mass comparator that has a nominal performance of one part per million or lower uncertainty, there are aspects of the measurement that can lead to unacceptable uncertainty. The impact of the whole measurement process needs to be evaluated. Further, it should be evaluated using a reasonable statistical control process that will provide some level of certainty that the uncertainty in the process is under control. If there is no evidence that the whole process is under control, then the uncertainty is meaninless.

Dear John, you are right if I want to use bottom-up uncertainty, which is more accurate. However, at the moment, the up-down approach is adequate for our purpose. thank you

All the comments mentioned by others are on point. As a forensic scientist we addressed this last year and published a paper directed at the GC-MS. We directed it toward the overall functions of the instrument and focused on retention time and its reproducibility. The paper is posted here on Research Gate. We took data from 3 different laboratories and gathered all the points with the bottom line is the tolerance within all three laboratories were extremely low. Much less than our previously assessed numbers. Now difference you might exhibit could be dependent on the column you use or the method you are employing . As mentioned by John and Bruce, the method will give you the highest uncertainty as opposed to the instrument. You will have minor differences between instruments but I would expand my uncertainty to incorporate this so that you will only have to give one value.

All answer are on point. I only add that the calibration of this instruments should also be carried out before use.

I agree with Benny Lum .

https://books.google.az/books?id=n1zEDQAAQBAJ&pg=PA311&lpg=PA311&dq=Why+there+is+difference+in+instrumental+uncertainties?&source=bl&ots=jPFNoacVAJ&sig=ACfU3U2ISG3j0EA7cUDqP72WY_fDoPdfDQ&hl=ru&sa=X&ved=2ahUKEwiTm6XQw6LqAhWLGuwKHaPCDKgQ6AEwAXoECAkQAQ#v=onepage&q=Why%20there%20is%20difference%20in%20instrumental%20uncertainties%3F&f=false