Laserfiche WebLink
ce��tn PJB - - <br />"•e�' e�cn PJC. <br />Cate: <br />1 T S <br />1,,.o GETECTIOtr �;M Fa ; of L <br />Before any a n a I tic4� SAY s� �1 — --te a- fie 4inalysi-s of a <br />sample, Multi -Tech analyt eel laboratory personnel analyze <br />sufficient spiked and blank samples to statistically establish <br />the lowest concentration which may be distinc,ished from zero. <br />The detection li-its are determined using 9:'% confidence interval <br />determinations. data utilized in these determinations are from <br />s:•iced standard samples. Tr,e lowest calcu'iated concentration is <br />reported e5 t---=evezti� m,on l .it of r; .. ...e <br />1 east cr,e cf ..P t':c,: provided that et <br />st•ip.e i?ar7et:Cnce-.traticns is below t^e <br />calculated detection limit. Ot.nerwise, the lowest spike <br />concentration is the minimum level that can be reported as t�:e <br />detection 1!T.'t,hichest spike concentration represent,, he <br />ii � ntrie <br />upper limit of t•e reportaola data. <br />If the calculated detection limit is higher than the lowest <br />tested spike concentration and is higher tr"an desired for the <br />analysis, the cetection limit is recalculated after discarding <br />the data for the highest tareet concentration. Detection limits <br />may be rece l ::; i ated of to r successively ci sca rdi ng one <br />concentration level at a time in decreasing order of <br />concentration I e v e 1 5 until a minimum detection is obtained. At <br />least three posit=ve tenet levels must be used in calculating <br />the detection 14rn4t, and in no case will the reported detection <br />limit be less then the lowest tested concentration or higher then <br />the highest tested concentration used to calculate the detection <br />limit. <br />The detection limit is derived from the assumptions that the <br />relationship between the found concentration and the target <br />concentration is linear, that the variance about the least <br />squares linear regression line is homogeneous over the tested <br />concentration range, and that the distribution of found <br />concentrations for a given target concentration is a normal <br />distributed. <br />Based on these assumptions, the least squares linear regression <br />line of t h e <br />Ys Yo + bx <br />is determined,whepe <br />Yo s Y axis (found concentration) intercept <br />b s slope of the line <br />x s target concentration <br />A measure of t -e linearity of the Cate is the correlation <br />coefficient, P. L -r Ce1cL,Imt-.4 r,.. 4ke-- <br />