(ICP/AE) and atomic absorption (AA) methods, both of which are conducted in a laboratory 
and typically take two to four days to get results. 4 
During Phase I, Syracuse sent all samples from the first 15 homes to the laboratory for analysis 
using these methods. In addition, Syracuse provided XRF data for these samples. The laborato¬ 
ry performed a statistical comparison between samples analyzed by XRF and the same samples 
analyzed by ICP/AE + AA, yielding a percent difference for each set of values. The EPA Region 
2 laboratory reviewed these values and helped establish performance criteria to be used by the 
field XRF operators. In conducting this evaluation, Syracuse took into account the fact that 
XRF technology tends to have a bias to the low side of laboratory determined values. To pro¬ 
tect against false negative results due to instrument bias, Syracuse reviewed results from tests 
where both XRF and laboratory methods were used. Results where the XRF reading taken was 
above the laboratory result for the same sample were disqualified from the analysis as outliers. 
The remaining data were separated by location type (i.e., floors, window sills or window wells); 
the difference between the XRF and laboratory methods were taken for each set of samples; and 
a standard deviation calculated for each location type. Results within one standard deviation 
below the acceptable level are also considered positive results as a worst case estimate . To pro¬ 
tect against false positive results, where the 
worst case estimates are within 2 times the 
limit of detection, samples are sent to the 
laboratory to confirm results. 
The project team used these findings during 
Phase II to determine which new samples 
would be sent for laboratory confirmation. 
After establishing a statistical correlation, 
Syracuse started sending only those XRF 
samples falling within a specific range to the 
laboratory for confirmation (See table adja¬ 
cent as well as the post mitigation report 
entitled Settled Dust Sample Results 
included at the end of Chapter 8). Syracuse 
expects that the laboratory will continue to 
refine this statistical analysis as more data 
become available. Current results, however, 
show an acceptable correlation between XRF 
and laboratory data. 
Data Management 
In Syracuse, the XRF instrument is the main data management tool used by the field sampling 
technician. The XRF has pre-loaded software that can read and store up to 3,000 entries 
before data is downloaded to alternative storage. As explained in Chapter 5, each sample bag is 
given a unique number designating the sampling location within the house. Upon completion 
of sampling and analysis, Syracuse downloads the data from the XRF to the City of Syracuse 
computers. 
XRF Readings Requiring Laboratory 
Confirmation 
Sample Type XRF Reading Lab 
Confirmation? 
Floor >40 g/ft 2 (MDL) No 
Floor <40 g/ft 2 Yes 
Window Sill <100 g/ft 2 No 
Window Sill 100 g/ft 2 and 250 g/ft 2 Yes 
Window Sill >250 g/ft 2 No 
Window Well <180 g/ft 2 No 
Window Well 180 g/ft 2 and 400 g/ft 2 Yes 
Window Well >400 g/ft 2 No 
4 It should be noted that the samplings conducted by the Syracuse project are not regulatory compliance 
tests and therefore do not require the use of an EPA-approved method. 
Analyzing Lead 
Dust Samples Using XRF Technology 
4 5 
