30
ObservAir
Operating Manual
5.
Best Practices
5.1.
Filter replacement
As particulate matter deposits on the filter, the ObservAir uses a simple
mathematical relationship to calculate black carbon (BC) concentrations in
the sample flow as a function of the light attenuation rate through the filter
(Section 1.1.1). As the aerosol filter becomes overly saturated with BC
deposits, however, this underlying relationship degrades and the
ObservAir’s BC concentration measurements are underreported (lower
than the true value). This measurement degradation is known as the ‘filter
loading artifact’, and can be largely avoided by changing the aerosol filter
tab when the optical attenuation (ATN) values exceeds 80: a commonly
accepted ATN threshold for aerosol photometry at 880 nm. While some
measurement error persists below this threshold, it is usually small and
grows larger for higher ATN values. Therefore, it is best practice to
only
operate the ObservAir while the optical ATN is < 80, and replace the filter
when this threshold value is exceeded.
Given this optical attenuation limit, the filter’s operational life before
requiring replacement depends only on the average BC concentration and
flow rate of air sampled through the sensor. For an average BC
concentration of 1
µ
g/m
3
,
Table 2 shows that the effective filter life ranges from 6.3 to 1.6 days as the
sample flow rate settings increases from 50 to 200 ccm. Filter life is
inversely proportional to both average BC and flow rate, and may be
calculated using the equation below. The equation approximates the total
sampling time required for the aerosol filter to reach an ATN of 80 for the
given input conditions.
𝐹𝐿 =
313.5
𝐵𝐶
)-&
∙ 𝐹𝑅 (5)
FL
= Filter life (days)
BC
avg
= Average BC concentration (
µ
g/m
3
)
FR
= Flow rate (ccm)
5.2.
Filter loading correction
While sensor operation at low optical attenuation levels largely eliminates
the filter loading artifact, this may not always be practical and convenient,
and some BC measurement error necessarily remains. As a result, many
filter loading correction algorithms have been developed to compensate
aerosol absorption photometers’ BC measurements as a function of optical
attenuation. These empirical corrections are well documented and
31
ObservAir
®
Operating Manual
generally straightforward to implement. However, the filter loading artifact
is not static – it depends on the emissions source (e.g. biomass vs. diesel),
atmospheric conditions, seasonality, and other factors. Therefore, filter
loading correction algorithms should be calibrated and validated for each
particular application of the ObservAir, and periodically updated over long-
term deployments. Standard filter loading correction procedures for the
ObservAir are under development, and will be released in future versions of
this manual. Helpful resources on the filter loading artifact are provided
below.
Filter loading correction resources:
1.
Good, N.; Mölter, A.; Peel, J. L.; Volckens, J. An Accurate Filter Loading
Correction Is Essential for Assessing Personal Exposure to Black Carbon
Using an Aethalometer. J. Expo. Sci. Environ. Epidemiol. 2017, 27 (4),
409–416.
https://doi.org/10.1038/jes.2016.71
.
2.
Jimenez, J.; Claiborn, C.; Larson, T.; Gould, T.; Kirchstetter, T. W.;
Gundel, L. Loading Effect Correction for Real-Time Aethalometer
Measurements of Fresh Diesel Soot. J. Air Waste Manag. Assoc. 2007,
57 (7), 868–873.
https://doi.org/10.3155/1047-3289.57.7.868
.
3.
Virkkula, A.; Mäkelä, T.; Hillamo, R.; Yli-Tuomi, T.; Hirsikko, A.; Hämeri,
K.; Koponen, I. K. A Simple Procedure for Correcting Loading Effects of
Aethalometer Data. J. Air Waste Manag. Assoc. 2007, 57 (10), 1214–
1222.
https://doi.org/10.3155/1047-3289.57.10.1214
.
5.3.
Flow rate setting: Filter life vs. BC resolution
The filter’s operational life decreases at higher flow rates, so it may be
tempting to set flow rate at the lowest possible option (25 ccm) and
minimize sensor maintenance. However, effective BC measurement
resolution also depends on flow rate: Since the filter loads up with BC more
rapidly at higher flows, the time rate of light attenuation is more readily
detectable, and BC baseline noise decreases. BC baseline noise represents
the sensor’s effective measurement resolution and is shown for various
sampling intervals in Table 4 for a flow rate of 100 ccm. BC baseline noise
is inversely proportional to the sample flow rate. For example, 1-minute BC
noise is around 0.025
µ
g/m
3
at 200 ccm (half that shown in Table 4 for
100 ccm). In this way, higher flow rates provide measurements with higher
temporal resolution, but at the expense of filter life. The optimum flow rate
setting maintains adequate BC measurement resolution for the monitoring
application while maximizing filter life such that sensor maintenance
remains convenient and practical. The procedure below outlines the
calculation and selection of an appropriate ObservAir flow rate setting.