Uncertainty in Light Scattering Measurements by TSI Nephelometer: Results from Laboratory Studies and Implications for Ambient Measurements

Figures & data

TABLE 1 Property summary of sub-1 micron aerosols generated for the study. For polydisperse particles, the size is the number mean diameter obtained from size distribution tests. DOS is a Fluka, Inc. product; the other compounds are commercially available at Sigma-Aldrich Inc.

	Type, Size
	(nm) for
Compound	1% vol.	Optical	Refractive
name	solutions	characteristics	index (RI)
Ammonium sulphate (AS)	Polydisperse, d = 100 nm	Non-absorbing	1.53+0.0i
Sodium chloride (NaCl)	Polydisperse, d = 110 nm	Non-absorbing	1.54+0.0i
Dioctyl sebacate (DOS)	Polydisperse, d = 120 nm	Non-absorbing	1.45+0.0i
Di-carboxylic acids (ORG)	Polydisperse, d = 110 nm	Non-absorbing	n/a
Nigrosin Dye (NSN)	Polydisperse, d = 90 nm	Absorbing	1.7+0.3i
MIXPO	Polydisperse, d = 100 nm	Slightly absorbing	n/a
Humic Acid	Polydisperse, d = 100 nm	Slightly absorbing	n/a

FIG. 1 Schematic diagram of the experimental setup used in this study.

FIG. 2 (a) σ_sp measured by the TSI nephelometer (corrected using the AO98 expression, and converted to 532 nm) displayed versus the σ_ep measured by the CRD for sub-1 micron dry AS particles (slope = 0.998, R²= 0.999). (b) TSI nephelometer submicron correction factors determined by the AO98 correction plus wavelength conversion (C_λ,AO98, solid line) and fit to the experimental data plus wavelength conversion (C_λ,EXP, dashed line) for polydisperse AS, NaCl, and DOS particles. Error bars are ±3.0% (1-σ uncertainty).

FIG. 3 Ratio (R) of scattering values obtained by difference (σ_ep minus σ_ap) and directly measured by TSI nephelometer (σ_sp) plotted as a function of ω₁. Data are for AS/NSN internal mixtures. The σ_sp coefficients are corrected using the AO98 scheme and converted to 532 nm. Error bars are calculated by propagating the absolute uncertainties of the optical coefficients (δσ_ep= 1%; δσ_ap= 5%; δσ_sp= 3% following normalization with the CRD).

TABLE 2 Input parameters for Mie code calculations to estimate ω values for the internally mixed AS/NSN particles. From left to right columns are: NSN percentage in the mixture; RI calculated by the linear mixing rule (Equation [3]); size calculated for the internally mixed particles; single scattering albedo values from Mie code (ω_Mie); and the experimental ω data, i.e., ω₁, ω₂, and ω₃

% NSN	RI	D (nm)	ωMie	ω1	ω2	ω3
0%	1.53+0.0i	100	1	0.99	1.002	0.99
8%	1.548+0.03i	95	0.82	0.89	0.90	0.90
15%	1.56+0.06i	90	0.70	0.83	0.84	0.83
27%	1.57+0.09i	90.3	0.58	0.68	0.71	0.71
54%	1.63+0.18i	100	0.49	0.55	0.60	0.58
75%	1.66+0.24i	95	0.43	0.47	0.52	0.51
100%	1.7+0.3i	90	0.39	0.41	0.48	0.47

FIG. 4 Experimental (ω₁, ω₂, ω₃) and theoretical (ω_Mie) single scattering albedo values for internal mixtures AS/NSN, reported versus the mass percentage of AS in the mixture.

FIG. 5 TSI 3563 nephelometer correction factors C calculated using Mie theory for various RI values, for submicron aerosols (panels a and b) and for sizes up to 2 μm (panels c and d). Points are color coded either by n (panels a and c) or ω (panels b and d). The thin solid lines of panel a show a correction calculated as a function of n only. The dashed line is the C from AO98; the solid lines are the C values from Equations (4) and (5). Circles are calculations for NSN (n= 1.7, k=0.3).

TABLE 3 Coefficients ν₀, ν₁, ν₂, and ν₃ for Equations (4) and (5) given for the TSI 3563 nephelometer center wavelengths of 453, 554, and 698 nm for situations with and without submicron impactor

	ν0	ν1	ν2	ν3
698 nm sub-μm	0.8627	0.1423	0.1816	0.0306
554 nm sub-μm	0.8511	0.1589	0.2153	0.0439
453 nm sub-μm	0.8863	0.1327	0.2758	0.0610
698 nm all	0.9868	0.0182	0.7980
554 nm all	0.9948	0.0152	0.8951
453 nm all	1.0072	0.0118	1.0036

FIG. 6 Ratio (R) of scattering values obtained by difference (σ_ep minus σ_ap) and directly measured by TSI nephelometer (σ_sp) plotted as a function of ω₁ for the AS/NSN mixtures with 27%, 54%, 75%, and 100% NSN. Theσ_sp coefficients are corrected using the real-refractive index dependent correction term C from Equations (4) and (5).

FIG. 7 (a) Percentage uncertainty inσ_sp from TSI nephelometer (δσ_sp) calculated using the AO98 scheme (triangles) and corrected using Equations (4) and (5) (squares), and (b) percentage uncertainties of single scattering albedo δω₁ (CRD and PAS) and δω₂ (CRD and TSI nephelometer). δω₁ is calculated using δσ_ep= 1% and δσ_ap = 5%. δω₂ is calculated combining δσ_ep=1% with either δσ_sp AO98 (triangles) or δσ_sp Equations (4) and (5) (squares).