Hi, I am developing a GC/FID method for a 6890 that characterizes and quantitates halogenated analytes used in vapor degreasing. My problem is that my results are not precise and I do not have a calculation method to express my results as weight percent (

7/2/19

Hi,

I know I probably wont get immediate replies from the forum because of the holiday, but please after the holiday, any help or insight is always welcome. I am continuing on this thread because it relates to the original question of accuracy and precision in regards to the characterization and quantitation of solvents used in vapor degreasing. Previous entries from me have mentioned that my method is for four compounds, decafluoropentane, tetrafluoroethyl trifluoroethyl ether, 1,2 trans dichloroethylene, and 1,2 butylene oxide. My goal was to create a stock standard to run a calibration to analyze the compounds I mentioned. As said before, each compound is in liquid form and the concentration is in % purity. For each compound a percentage purity is described and can be converted to a unit that better reflects percent weight (w/w). For example decafluorpentane has 99.9% purity in liquid form. This can also be expressed as 999900 ug/g. Each compound can then be expressed as ug/g based on their respective purities. I weighed out each compound to 999900 ug, or 0.9999 g into a 2 oz glass bottle with a gasketed screw top lid to minimize evaporation. Once the compounds were combined in the mixture, I added an equal amount of diethyl ether as a diluent, thus bringing the concentration down to 499950 ug. I then call this my upper limit of 99.0% w/w. The preparation of the calibration curve is next with 7 levels, each level having a specific concentration in ug/g. Since my last entry, I have passed my calibration curve with all seven levels (does having more levels in a curve and having them pass make the overall method more "robust")?

When I run samples and calibration verification samples, I am still having a few issues. The CCVs perform very well, with very little variation, but when I run samples, I do not get the numbers I expect.

When running my calibration, I discovered that when I modify my calculation page to include a multiplier (in this case 0.0001) and a dilution factor of 2, I can generate a result that is read in percentage weight w/w. For example,  a LVL 4-25.0%w/w is expressed as 24.7834% which is what I expected.

My sample preparation is exactly the same as the preparation of my calibration standards and proceeds as follows: the samples analyzed are either neat raw product, or they are mixtures of the relevant analytes sampled from a working vapor degreaser. Regardless of neat or in process samples, the percent purity remains the same, so I can sample solvent, weight out exactly 0.9999g in a 2 mL vial, and dilute it by half to give a concentration of 499950 ug/g...however, each compound instead of expressing as a percent weight, my decimal point moves after I calculate and my results are off...for example instead of 14.8%w/w, it will show 198.62132 as my result. I cannot figure out what is going on. Would I have better results if I calculated using an internal standard, instead of an external standard? Like I said, any input is welcome.

7/2/19

Hi,

I know I probably wont get immediate replies from the forum because of the holiday, but please after the holiday, any help or insight is always welcome. I am continuing on this thread because it relates to the original question of accuracy and precision in regards to the characterization and quantitation of solvents used in vapor degreasing. Previous entries from me have mentioned that my method is for four compounds, decafluoropentane, tetrafluoroethyl trifluoroethyl ether, 1,2 trans dichloroethylene, and 1,2 butylene oxide. My goal was to create a stock standard to run a calibration to analyze the compounds I mentioned. As said before, each compound is in liquid form and the concentration is in % purity. For each compound a percentage purity is described and can be converted to a unit that better reflects percent weight (w/w). For example decafluorpentane has 99.9% purity in liquid form. This can also be expressed as 999900 ug/g. Each compound can then be expressed as ug/g based on their respective purities. I weighed out each compound to 999900 ug, or 0.9999 g into a 2 oz glass bottle with a gasketed screw top lid to minimize evaporation. Once the compounds were combined in the mixture, I added an equal amount of diethyl ether as a diluent, thus bringing the concentration down to 499950 ug. I then call this my upper limit of 99.0% w/w. The preparation of the calibration curve is next with 7 levels, each level having a specific concentration in ug/g. Since my last entry, I have passed my calibration curve with all seven levels (does having more levels in a curve and having them pass make the overall method more "robust")?

When I run samples and calibration verification samples, I am still having a few issues. The CCVs perform very well, with very little variation, but when I run samples, I do not get the numbers I expect.

When running my calibration, I discovered that when I modify my calculation page to include a multiplier (in this case 0.0001) and a dilution factor of 2, I can generate a result that is read in percentage weight w/w. For example,  a LVL 4-25.0%w/w is expressed as 24.7834% which is what I expected.

My sample preparation is exactly the same as the preparation of my calibration standards and proceeds as follows: the samples analyzed are either neat raw product, or they are mixtures of the relevant analytes sampled from a working vapor degreaser. Regardless of neat or in process samples, the percent purity remains the same, so I can sample solvent, weight out exactly 0.9999g in a 2 mL vial, and dilute it by half to give a concentration of 499950 ug/g...however, each compound instead of expressing as a percent weight, my decimal point moves after I calculate and my results are off...for example instead of 14.8%w/w, it will show 198.62132 as my result. I cannot figure out what is going on. Would I have better results if I calculated using an internal standard, instead of an external standard? Like I said, any input is welcome.