Suggested Sampling Protocol for OC and EC
Outlined by Barb Turpin (turpin@aesop.Rutgers.edu)
- Suggested Sampling Protocol
- Collection method: Two parallel samplers. Port one contains a quartz filter.
Port two contains a Teflon and quartz filter in series.
Quartz filters should be pre-baked and both should be from the same lot.
Face velocity: Between 20 and 40 cm/sec (ideally, value should be similar for
all samplers used in ACE Asia)
Duration: Between 12 and 24 hours
Data reported: OC and EC concentrations from both quartz filters.
Note on the use of denuders: Use of a denuder has the potential to provide more
accurate particulate carbon values but a denuder must be used with caution as denuder
collection efficiency varies substantially with the organic composition of the
gas-particle mixture, temperature, and the presence of competing species such as
water vapor. The denuder collection efficiency must be determined with real gas-particle
mixtures at the sampling location. This can be done by sampling particle-free ambient
air with a denuder-adsorbent combination in Port 1 and an adsorbent only in Port 2.
The denuder collection efficiency is the difference between the OC loading on adsorbents
in Port 2 and Port 1divided by the OC measured on the adsorbent in Port 2.
(See Turpin et al., Atmos. Environ., 34, 2983 – 3013 for further information on
the use of denuders and other issues in sampling and analyzing for OC and EC).
- Information submitted to UCAR data management system. All researchers measuring OC
and EC should include the following information when submitting their data to the UCAR
data management system:
- Face velocity
- Sample duration
- Type of substrate(s)
- Analytical method and manufacturer (e.g., TOR, TOT)
- Artifact correction approach, if any
- Comparability of data. As data sets are integrated after the experiment, the
comparability of the data collected should be evaluated based on face velocity,
sample duration, and substrate material. If these parameters are similar, the
OC and EC measurements should be comparable but not necessarily artifact-free.