Application of the U.S. EPA procedure for determining method detection limits to EDXRF measurement of filter-based aerosol samples

Figures & data

Table 1. Epsilon 5 application used to analyze samples.

Secondary Target	Analysis Time, s	kV	mA	Detector Setting	Reported Elements
CaF2	600	40	15	High Resolution	Na, Mg, Al, Si, P, S, Cl, K
Fe	400	75	8	Standard	Ca, Ti, V, Cr
Ge	400	75	8	Standard	Mn, Fe, Co, Ni, Cu, Zn
SrF2	500	100	6	Standard	As, Se, Br
Zr	500	100	6	Standard	Pb
Mo	500	100	6	Standard	Rb, Sr
Al2O3	500	100	6	Standard	Zr Ag, Cd, In, Sn, Sb, Cs, Ba, Ce

Figure 1. The variance of deposit normalized to average loadings (Var(dep)/C_ME) of selected elements on the ME set.

Table 2. The comparison of MDL_LB and MDL_ME along with current reported MDL (MDL_FB, will be discussed later) and MDL estimated by 3-sigma-criterion from the literature (all MDL values have units of ng cm-2, n.a. refers to not available). Elements with asterisks (*) have C_ME/MDL_LB outside of the recommended 2–10 × MDL range. Note MDL_LB and MDL_ME were determined in 2018, while MDL_FB was determined in 2020. The greater than (>) signs indicate elements where MDL_LB exceeded MDL_ME.

Element	MDLLB	MDLME	MDLFB	MDL^a	MDL^b	MDL^c
Na*	76	140	66	105.0	n.a.	n.a.
Mg*	45	94	37	85.0	n.a.	6.7
Al	17	46	19	55.0	115.0	0.6
Si*	15 >	13	11	20.0	93.0	75.3
P*	0.0	8.6	1.6	10.0	14.0	n.a.
S*	0.0	13	0.75	7.0	32.0	5.0
K	5.9 >	5.6	4.4	2.5	7.0	10.0
Ca	11	18	8.0	4.5	7.0	42.4
Ti	2.5	8.4	2.4	4.0	5.0	3.5
V*	0.55	8.8	0.58	1.5	n.a.	3.3
Cr	1.8	2.8	1.9	1.5	3.0	1.6
Mn	3.8	5.4	2.4	4.0	14.0	2.5
Fe	14 >	11	6.9	3.5	11.0	15.3
Co	1.3	2.1	1.3	4.0	n.a.	4.3
Ni	0.85	3.0	0.99	1.5	3.0	0.4
Cu	6.9	8.0	3.5	1.5	10.0	7.0
Zn	2.4	5.3	1.4	2.0	8.0	4.1
As	1.0	6.0	0.03	8.0	0.5	12.6
Se	2.8	3.8	2.0	7.0	n.a.	n.a.
Rb	2.8	5.1	2.6	n.a.	n.a.	n.a.
Sr	3.8	6.3	2.4	3.0	n.a.	1.9
Zr	18	36	11	5.0	n.a.	n.a.
Ag	9.1	16	10	n.a.	n.a.	n.a.
Cd	12	14	11	n.a.	n.a.	18.5
In	16	27	12	n.a.	n.a.	n.a.
Sn	13	26	13	n.a.	n.a.	27.7
Sb	22	35	13	n.a.	33.0	39.1
Cs	24	59	22	n.a.	n.a.	n.a.
Ba	34	66	23	1.5	52.0	n.a.
Ce	48	78	30	n.a.	n.a.	n.a.
Pb	7.3	9.2	5.4	9.0	15.0	6.9

^a(Gunchin et al. 2019)

^b(Zhang et al. 2012)

^c(Yatkin, Gerboles, and Borowiak 2012)

Figure 2. The average elemental loadings (C_ME) normalized to MDL_LB ratios (C_ME/MDL_LB) for the ME set. Note: P and S have MDL_LB of 0, so their C_ME/MDL_LB values are not shown. Horizontal lines are upper and lower limits recommended in the EPA procedure.

Figure 3. Ratios of MDLs acquired in 2021 for annual verifications to those acquired in 2018 right after generation. Elements marked with triangles have C_ME/MDL_LB outside of 2–10.

Figure 4. Comparison of MDL_EPA(2021) and MDL_FB(2020). Elements are in order of low to high atomic number (Z).

Figure 5. Comparison of three different approaches for estimating MDL: MDL_EPA(2021), MDL_rep-blk(2019) and MDL_rep-ME(2021).

Gunchin, G., S. C., D. I., A. G. Germanos Karydas, K. Eleftheriadis, S. Lodoysamba, D. Shagjjamba, A. Migliori, R. Padilla-Alvarez, M. Manousakas, et al. 2019. Three-year long source apportionment study of airborne particles in Ulaanbaatar using x-ray fluorescence and positive matrix factorization. Aerosol Air Qual. Res. 19 (5):1056–67. doi:https://doi.org/10.4209/aaqr.2018.09.0351.

 Web of Science ®Google Scholar

Zhang, N., J. Cao, H. Kinfai, and H. Yuanqing. 2012. Chemical characterization of aerosol collected at Mt. Yulong in Wintertime on the Southeastern Tibetan Plateau. Atmos. Res. 107 (April):76–85. doi:https://doi.org/10.1016/j.atmosres.2011.12.012.

 Google Scholar

Yatkin, S., M. Gerboles, and A. Borowiak. 2012. Evaluation of standardless EDXRF analysis for the determination of elements on PM10 loaded filters. Atmos. Environ. 54 (July):568–82. doi:https://doi.org/10.1016/j.atmosenv.2012.02.062.

 Google Scholar

Data availability statement

The data that support the findings of this study are available from the corresponding author, NPH, upon reasonable request.