1) Calculate the mass of chloroform actually injected onto the GC column when you injected pure chloroform at the beginning. To do this you will need the volume injected, the split ratio, and the density of the compound. The split ratio is the volume injected that does not go to the column (and is vented instead) for every volume that does go onto the column. For example, if the split ratio is 100:1, then only 1/101 of the amount injected from the syringe actually goes onto the column (one part to GC, 100 parts vented). Express the answer in micrograms. 2) Even if you successfully inject the exact same volume of each compound onto the GC column, the peak heights and/or areas do not necessarily (and often are not) equal, because the detector has different sensitivities to different compounds. Since 1-propanol and 2-propanol are so similar (same chemical formula), the sensitivities of detection are probably also very similar. However, the sensitivity to chloroform should be significantly different. Which compound has the lowest FID sensitivity of the three compounds? Note that peak height decreases with retention time because peaks broaden while they are on the column. Peak area, however, is conserved. Look at the peak areas from the pure compound injections at the beginning of the experiment to answer this question. The greater the peak area (for the same amount injected), the more sensitive the detector is to that compound. Hypothesize why the sensitivities are different. To answer this question, consider the amount of carbon and hydrogen that burns in the molecules. 3) Calculate the confidence interval of the result computed from a least-squares line. It is C.I. = tsx. Convert with the dilution factor as was done for the final value of % (w/v) for the sample. Note for only 3 points on a calibration graph, t = 12.7 because there is only one degree of freedom (n-2), and the confidence is therefore poor. How well the three points lie on the line obviously also affect this confidence. The "fit" is quantified by two least-squares parameters in the third row of the LINEST output: the correlation coefficient , and the standard deviation about the regression, sr. The correlation coefficient is normalized to one. The closer to one, the better. At least "two nines" (i.e. 0.99 or better) is considered acceptable for well calibrated data. s is a value in the formula for the standard deviation of the computed result, sx. It is also called sy for the standard deviation of the y measurements, see Eq. 4-19.

pur-new-sol

Related Questions