2019 AOAC Furans WG Book

41 Recovery . — The fraction or percentage of analyte that is measured when the test sample 42 is analyzed using the entire method.

43 44 45 46 47 48 49 50 51 53 54 55 60 61 62

5. Method Performance Requirements

Table 1. Limit of quantitation (LOQ)

Coffee (solid material)

≤ 20 µg/kg ≤ 5 µg/kg

Other matrices

Table 2. Recovery, repeatability and reproducibility parameters

Recovery, %

80-110

RSDr, % RSDR, % 0.66 times RSDR as derived from (modified) Horwitz equation a As derived from (modified) Horwitz equation a a Horwitz equation for predicted relative standard deviation of reproducibility: PRSD R = 2C -0.15 , 52 where C is analyte concentration expressed as mass fraction. 56 57 Suitable methods will include blanks and appropriate check standards. 58 Method (procedural) and solvent blanks should be below the limit of detection (LOD = 59 0.3 x LOQ). 63 64 Validation should be conducted at the target LOQ and 10xLOQ levels. The LOQ is 65 determined as the lowest spiking level that meets the recovery and repeatability 66 requirements. Suitable matrix blanks should be selected that do not contain more than 67 30% of the target LOQ level for each analyte. 68 69 For matrices that naturally contain higher levels of furan and alkyl furans ( e.g. ground 70 roasted coffee) and where suitable matrix blanks are not available (for all or certain 71 analytes), spiking experiments should be conducted for the affected analytes at two 72 concentration levels in the range of 3-10x the analyte level in the evaluated matrix. In 73 this case, the LOQ can be estimated based on extrapolation of signal-to-noise ratio (S/N) 74 obtained for a concentration level naturally present in the evaluated matrix to a 75 concentration level that would correspond to S/N = 10. 6. System Suitability Tests and/or Analytical Quality Control 7. Validation Guidance

76 77