59 



pyrene recoveries exceeded these limits in the MSD analysis. Because all other analytes met 

 both accuracy and precision limits, the analytical system was assessed to be in control and no 

 further corrective action was implemented. Fluoranthene and pyrene were reported without 

 laboratory qualification. 



Using NLDS 15 A for the 1998 MS/MSD analyses, recoveries ranged from 84 to 

 109% and were acceptable. The laboratory QC sample had recoveries that varied from 76 to 

 92%. The accuracy tests indicated that the PAHs analysis met the QC criteria. 



Surrogate Recovery. Each sediment sample for PAH analysis was spiked with 

 surrogate compounds as a measure of accuracy. Surrogate samples are analyzed as a check 

 on the laboratory's ability to extract known concentrations of compounds not normally found 

 in the sample, but having similar characteristics. Surrogate compounds (generally 

 compounds labeled with stable isotopes) are the only means of checking method 

 performance on a sample-by-sample basis. Recoveries of surrogate spikes must fall within a 

 range of 30 to 130%, depending upon the surrogate compound, as stated in the WHG 

 laboratory standard operating procedure. Measured recoveries of surrogate spikes for this 

 data set ranged from 46 to 119%, indicating acceptable recovery. 



Method Blanks. Method blanks are laboratory QC samples that are processed with 

 the samples but contain only reagents. Method blanks test for contamination that may be 

 contributed by the laboratory during sample preparation. The method blanks for PAHs, 

 TOC, and metals were free from contamination and below the instrument detection limit. 



2.4.4.3 Assessment of Analytical Precision 



Analytical precision can be expressed as the relative percent difference (RPD) 

 between two results, or the relative standard deviation (RSD) between three or more results. 

 To prepare analytical replicates, a sample is homogenized by the laboratory and then divided 

 into two or more subsamples. The subsamples are analyzed independently. The closer the 

 numerical values of the measurements are to each other, the lower the RPD or RSD. Low 

 RPD or RSD values indicate a high degree of analytical precision. 



The relative percent difference (RPD) between two sample results was calculated 

 using the following equation: 



(sample result - duplicate result) 

 KrD — X 1UU 



(sample result + duplicate result) / 2 



Monitoring Cruise at the New London Disposal Site, Seawolf Mound 1995 - 7998 



