Measurement of Air Toxics in Engine Exaust Using Fourier Transform Infrared (FTIR) Spectroscopy

Page 12

concentrations measured by FTIR in the spiked gas stream. Lower native concentrations were observed for acrolein than for acetaldehyde. Since both were in the same spike gas cylinder at almost the same concentration, it was not possible to adjust the amount of acrolein spiked without affecting the acetaldehyde concentration. Therefore, acetaldehyde was spiked at concentrations which roughly doubled the native concentrations resulting in acrolein spike concentrations approximately six times the native concentrations. Concentration data in Table 3 were entered into spreadsheets designed to calculate the proposed method's acceptability under Method 301, Section 6.3, Analyte Spiking. Relative standard deviation (RSD) for spiked and unspiked concentrations must be less than or equal to 50% for the proposed method to be acceptable. As shown in Table 4, all RSD values were less than 10%. The following equation was used to calculate the method bias: where: B Sm ~ DF CS = = = = = B = Sm - (~ * DF) - CS Bias at the spike level Measured mean of the spike of samples Measured mean of the unspiked samples Dilution factor for the spiked sample stream Calculated spike level (1) In determining if the method bias was statistically significant, the t-statistic was calculated as follows: t -statis tic = --;:::==I=B==I === ./SD 2 + sn2 V s u (2) where I B I is the absolute value of bias, and SDs and SDu are the standard deviation of the spiked and unspiked values, respectively. For formaldehyde and acetaldehyde, the FfIR method bias calculations indicate that the results meet the Method 301 validation criteria with no further action. For acrolein, the results showed significant bias with a correction factor of 1.14 calculated. The correction factor (CF) was calculated according to the following equation: CF = 1 1 +B-CS 11 (3)