Inter-comparison of low-cost sensors for measuring the mass concentration of occupational aerosols

Figures & data

Table 1. Aerosol measurement instruments specifications used in the current work.

Device	Cost category	Cost ($)	Size (m) (H × W × D)	Weight (g)	Sampling flow	Size range	Concentration range
DC 1700	Low	425	0.18 × 0.11 × 0.08	544	NA	>0.5 µm	0–10^6 #/cm^3
						>2.5 µm
Sharp GP	Low	12	0.046 × 0.03 × 0.02	15	NA	≥0.5 µm	0.5 V/ (0.1 mg/m^3)
Sharp DN	Low	21	0.05 × 0.044 × 0.02	52	NA	≥0.5 µm	1 V/ (0.1 mg/m^3)
pDR-1500	Medium	<10,000	0.181 × 0.143 × 0.08	1200	1.0–3.5 L/min	1.0–10 µm	0.001–400 mg/m^3
CPC 3007	Medium		0.14 × 0.14 × 0.292	1700	0.7 L/min	10–1000 nm	1–10^5 particles/cm^3
SMPS/CPC(GRIMM)	High	∼60,000			1 L/min	5–1000 nm	1–10^7 particles/cm^3
APS 3321	High	∼20,000	0.18 × 0.3 × 0.38	10,000	5 L/min	0.5–20 um	10,000 particles/cm^3

* Less than 10% coincidence loss at 10⁶ particles/cm³ based on the specifications provided directly from the manufacturer.

Figure 1. Experimental setup used to measure the detection efficiency of the DC1700.

Figure 2. Experimental setup used to determine the performance of low-cost sensors shown in panel (a). Schematic diagrams of aerosol generation systems shown in panel (b).

Figure 3. Detection efficiency of the DC1700 by aerodynamic diameter. Particles smaller than 0.5 µm were generated with a nebulizer followed by electrical classification, and the reference concentration was measured with the CPC. Particles larger than 1 µm were generated with the VOAG, and the reference concentration was measured with the APS. Total (fine + coarse) corresponds with the “small bin” of the DC1700 and coarse corresponds with the “large bin” of the Dylos. Fine was calculated as the small bin minus the large bin.

Figure 4. Raw output of the DC1700 relative to reference mass concentration for DC1700. Reference mass concentration was calculated by correcting SMPS + APS data with mass concentration measured with a gravimetric filter for each aerosol. The error bars represent one standard deviation.

Figure 5. Raw output of the low-cost sensors relative to reference mass concentration for: (a) Sharp DN; (b) Sharp GP. Reference mass concentration was calculated by correcting SMPS + APS data with mass concentration measured with a gravimetric filter for each aerosol. The error bars represent one standard deviation.

Table 2. Precision of low-cost sensors raw output (raw) and calculated mass concentration based on the regression model (mass) expressed as the mean coefficient of variation (CV,%) by aerosol type.

	DC1700
	Small bin		Large bin		Sharp DN		Sharp GP
Aerosol	Number concentration	Mass	Raw	Mass	Voltage	Mass	Voltage	Mass
0.9% salt aerosol	2.2	1.4	7.2	—	27	5.9	14	5.9
5% salt aerosol	2.2	1.4	15	—	24	5.4	51	4.8
Arizona road dust	4.3	3.1	5.0	—	27	1.7	5.0	1.9
Diesel fume	14	1.8	12	—	17	0.8	15	0.9
Welding fume	9.0	8.0	14	—	30	7.1	2.0	1.7
Overall mean	7.4	3.1	11	—	25	4.2	17	3.1

Table 3. The average calculated mass median diameter (MMD) and geometric standard deviation (GSD) for each aerosol based on the SMPS and APS data.

Aerosol	MMD + standard deviation (µm)	GSD + standard deviation
0.9% salt solution	0.58 ± 0.04	1.80 ± 0.04
5% salt solution	0.81 ± 0.01	1.60 ± 0.08
Arizona road dust	2.63 ± 0.09	1.00 ± 0.06
Diesel fume	Multimodal distribution, two modes (Figure S2 in the online supplemental information, SI)
Welding fume	Multimodal distribution, two modes (Figure S3 in the online SI)

* Multimodal MMD and CMD were not calculated for this work.

Table 4. Regression equations to estimate mass concentration (y in µg/m³) from DC1700 number concentration (x = small bin particles/cm³). For particle concentrations (x) less than 106 particles/cm³.

Aerosol	Equation	R^2
0.9% salt aerosol	y = 2.6 × x – 22	0.96
5% salt aerosol	y = 6.2 × x – 153	0.99
Arizona road dust	y = 8.6 × x + 98	0.98
Diesel fume	y = 54 × x + 511	0.91
Welding fume	y = 12 × x – 12	0.96

Table 5. Evaluation of mass concentrations measured with the pDR-1500 with reference to those measured with the SMPS and APS.

Aerosol	Data pairs	Slope ± std. error	Intercept ± std. error (µg/m^3)	r	R^2	% bias
0.9% salt aerosol	21	1.2 ± 0.02	−46 ± 12	0.99	0.99	−3.4
5% salt aerosol	21	0.9 ± 0.006	−5.0 ± 4.0^1	0.99	0.99	4.6
Arizona road dust	18	1.3 ± 0.04	−54 ± 14	0.99	0.98	−2.0
Diesel fume	21	1.1 ± 0.02^1,2	28 ± 66	0.99	0.99	−9.1
Welding fume	18	1.0 ± 0.02	−55 ± 24	0.99	0.99	−1.1

¹ Meets EPA criterion.

² Meets NIOSH criterion.

Table 6. Evaluation of mass concentrations estimated with the DC1700 with reference to those measured with the SMPS and APS. Based on DC1700 particle concentrations less than 106 particles/cm³.

Aerosol	Data pairs	Slope ± std. error	Intercept ± std. error (µg/m^3)	r	R^2	% bias
0.9% salt	21	0.9 ± 0.04	14 ± 7.0	0.99	0.98	4.0
5% salt	21	0.8 ± 0.05	21 ± 7.0	0.99	0.98	7.0
Arizona road dust	18	1.2 ± 0.04	10 ± 26	0.98	0.98	−18
Diesel fume	21	1.1 ± 0.05	−51 ± 23	0.95	0.91	−38
Welding fume	18	0.9 ± 0.07	71 ± 67	0.98	0.95	−3.0

¹ Meets EPA criterion.

² Meets NIOSH criterion.

Table 7. Evaluation of mass concentrations estimated with the Sharp sensors with reference to those measured with the SMPS and APS.

Aerosol	Data pairs	Slope ± std. error	Intercept ± std. error (µg/m^3)	r	R^2	% bias
A. Sharp DN
0.9% salt aerosol	21	1.2 ± 0.02	−34 ± 14	0.99	0.99	−3.7
5% salt aerosol	21	0.9 ± 0.006	−5.0 ± 5.0	0.99	0.99	5.2
Arizona road dust	18	1.3 ± 0.04	−55 ± 30	0.99	0.98	−5.8
Diesel fume	21	1.1 ± 0.05	3.3 ± 65	0.99	0.99	−9.8
Welding fume	18	1.0 ± 0.02	−57 ± 31	0.99	0.99	−1.3
B. Sharp GP
0.9% salt aerosol	21	1.2 ± 0.02	−33 ± 14	0.99	0.99	−3.8
5% salt aerosol	21	0.9 ± 0.006	−0.6 ± 5.0	0.99	0.99	4.6
Arizona road dust	18	1.3 ± 0.03	−56 ± 26	0.99	0.98	−6.3
Diesel fume	21	1.1 ± 0.025	8.2 ± 66	0.97	0.95	−9.3
Welding fume	18	1.0 ± 0.01	−55 ± 25	0.99	0.99	−2.4

¹ Meets EPA criterion.

² Meets NIOSH criterion.