United States
Environmental Protection
Agency
Environmental Monitoring
Systems Laboratory
Research Triangle Park NC 27711
Research and Development
EPA/600/S4-87/011 Sept. 1987
&EPA Project Summary
Sampling and Analytical
Methods Development for Dry
Deposition Monitoring
J. E. Sickles, II, William A. McClenny, and Richard J. Paur
The U.S. Environmental Protection
Agency (EPA) plans to implement a dry
deposition monitoring network. The
constituents of interest are HIM03, NO2,
SO2. NH3. IMH/, IMO3~, SO/", H +
(acidity), and O3. The objective of this
research was to identify the most
promising sampling and analysis
methods for network deployment in
relation to these constituents.
A phased approach consisting of
literature reviews, laboratory studies,
and field evaluations was employed.
Literature reviews were used to identify
those methods that appeared promis-
ing for direct application or for devel-
opment into methods suitable for field
application. The identified methods
were tested and refined in laboratory
studies. In those cases where the
methods remained potentially accepta-
ble after laboratory investigation, field
evaluations were performed.
A field study was then conducted in
the Research Triangle Park, North
Carolina, between August and
December of 1986 in which selected
airborne gaseous and paniculate acid-
ifying species were measured using
different types of samplers, and the
differences observed under field sam-
pling conditions were evaluated. The
bulk of the full report presents and
compares results collected in the 1986
study using the Annular Denuder
System (ADS) with those collected
using the Transition Flow Reactor
(TFR). Comparisons were also made
with results from the Filter Pack (FP),
Tunable Diode Laser Absorption Spec-
trometer (TDLAS) methods and
methods of other investigators. The
overall goal was to use the results of
the field evaluation along with other
information gathered during this pro-
ject to assess the current status of
available sampling and analysis
methods and identify method strengths
and weaknesses that must be consi-
dered prior to their field deployment.
This Project Summary was devel-
oped by EPA's Environmental Monitor-
ing Systems Laboratory, Research
Triangle Park, NC, to announce key
findings of the research project that is
fully documented in a separate report
of the same title (see Project Report
ordering information at back).
Introduction
The phenomenon of acid deposition
has received increasing attention in
recent years. Acid deposition is the
transfer of acidic substances in the
earth's surface by wet or dry deposition.
Dry deposition includes all processes by
which airborne contaminants are
removed from the atmosphere at the
earth's surface, excluding those pro-
cesses directly aided by precipitation. Dry
deposition contributes substantially to
the acidic deposition burden, at tirries
accounting for more than 50 percent of
the total. Efforts to monitor dry deposition
and to investigate its behavior will
provide insight into its role in the larger
problem of acidic deposition.
The U S. Environmental Protection
Agency (EPA) plans to implement a dry
deposition monitoring network. The
constituents of interest and the sampling
and analysis methods to be used, how-
ever, require definition. Many of the
airborne chemicals that are thought to
be important contributors to ecosystem
acidification are given in Table 1, along
with their nominal concentrations.
-------
Table 1. Contributors to Ecosystem Acidification
HN03
HN02
NO2
S02
NH3
NHt
N03~
so,2-
H* and Acidity
03
H,Oz
RCOOH
PAN
HCHO
Constituent
Nitric Acid
Nitrous Acid
Nitrogen Dioxide (an acidifying precursor}
Sulfur Dioxide (an acidifying precursor)
A rnmonia
Ammonium (particulate)
Nitrates (particulate)
Sulfates (particulate; may be H2SOt,
NH
-------
FP
The Filter Pack approach is employed
in networks operated by Environment
Canada and the Ontario Ministry of the
Environment The system consists of a
three-stage filter pack sampling at
approximately 20 Lpm An open-face 2
/urn pore size Teflon filter is the first
element, followed by a 1 /jm pore size
nylon filter, followed by an impregnated
(KjCOs and glycerin) Whatman-41 filter.
The filter size is approximately 47 mm
in diameter. Since open-face filters are
used, sampling is not size selective. The
extract from the Teflon filter is analyzed
for 1C for SCu2" and NO3~, by 1C in one
network and by colorimetry in the other
for NhU*, and by colorimetric titration for
acidity. The nylon filter is extracted in
0.003 NaOH and analyzed by 1C for N03"
and SOU2". The impregnated Whatman-
41 filters are analyzed for SC>42~ by 1C
which is added to that recovered from
the nylon filter and reported as SO2
The F P has the following limitations:
a. As a result of potential positive and
negative biases in HN03 and panic-
ulate N03~ sampling, the sampler
does not provide HN03 concentra-
tion data. The sum of the NOa" from
the Teflon and nylon filters repre-
sents Total Inorganic Nitrate (TIN).
b. The presence of acidic particles on
the Teflon filter offers the potential
for neutralization and change in
speciation of acidic sulfates when
NH3 is present in sampled air.
c. The system does not permit the
determination of gaseous NH3. As a
result, it also offers the potential for
high bias of paniculate NH4+ esti-
mates derived from Teflon filter
extracts.
d. The potential biases in nitrate sam-
pling, coupled with the potential
neutralization of acidity by gaseous
NH3, make interpretation of H*/
acidity data from the Teflon filter
extract difficult.
e. The system does not permit the
determination of HNOa. However, if
HNOz is transmitted by the Teflon
filter, then it should be retained on
the nylon and K2CO3-coated filters.
f. Sulfate collected on the nylon and
impregnated Whatman-41 filters
must be summed to infer SO2 con-
centration, since nylon can collect
SO2 as SO42".
g. Since the approach does not employ
size selective sampling, distinction
of the distribution of chemical con-
stituents between coarse and fine
particles is not possible
The advantages of FP are simplicity,
low cost of deployment, and high sen-
sitivity. The FP approach has been used
in different configurations for network
sampling in Canada since the late
1970's. Thus, of the systems considered,
the FP has the largest historical
database.
TFR
As the name suggests, the TFR incor-
porates Transition Flow Reactors and
filter collection media. As shown in
Figure 1, the TFR system is comprised
of modules arranged in series to collect
various constituents of interest. The first
module is a Teflon cyclone having a D5o
(p = 1.0) of 1.8 Aim at 33.2 SLPM and
of 2.5/jm at 17.3 SL PM The second and
third modules are two TFR for determi-
nation of HN03 and NH3, followed by a
filter pack containing a 2-/um pore size
Teflon filter, a 1 -/urn pore size nylon filter,
and an oxalic acid impregnated glass
fiber filter. The first TFR uses a nylon strip
as a partial denuder to remove a constant
fraction (i.e., 8.5%) of gaseous HN03
under transition flow conditions. The
second TFR uses a Nation strip as a
partial denuder to remove a constant
fraction (i.e., 17%) of gaseous NH3. The
subsequent Teflon and nylon filters
collect paniculate sulfate (PS), panicu-
late nitrate (PN), and HN03 vapor. PS
is determined from the analysis of sulfate
from the Teflon filter. The total gaseous
HNOs is determined from the analysis of
nitrate from the nylon strip, and the PN
is then determined algebraically, using
the nitrate found on the Teflon and nylon
filters. The total gaseous NH3 is deter-
mined similarly, using the analysis of
ammonium from the Nation strip, and the
paniculate ammonium is found algebrai-
cally, using the ammonium found on the
Teflon and oxalic acid coated filters.
Downstream of the filter pack, the flow
splits with 14.3 Lpm to a mass flow
controller and pump, while 1.8 Lpm
passes through two TEA-coated glass
fiber filters for collecting S02 and N02.
This low volume stream then passes to
a mass flow controller and pump. A third
flow stream at a nominal rate of 1.0 Lpm
is drawn from the base of the stagnation
zone of the cyclone to prevent particle
accumulation in the event of heavy
particle loading.
The TFR has the following limitations
a The system does not permit the
determination of HN02
b. Nitrates collected in the cyclone and
not removed during sampling may
volatilize to give artificially high
values of gaseous nitric acid.
c HN02 may be collected to some
extent on the nylon filter, while the
remainder is likely to be collected on
the TEA-coated filters, acting as a
positive bias.
d The collection of PAN and TEA may
provide bias for N02 determinations.
e Sulfur dioxide is likely to be collected
on currently available nylon filter
materials If this retention of SO2 is
significant, SOa determinations
using results from the TEA-coated
filters will require adjustment.
f. The potential for deposition and
partial retention of acidic and basic
gases and particles on the Teflon
filter makes interpretation of acidity
determinations difficult and may
prevent unambiguous inference of
atmospheric paniculate constit-
uents.
The TFR advantages include- it can be
used for sampling dry deposition constit-
uents for periods of 1 to 7 days; it is
modular, easy to install and ship; sample
analysis is by existing, accepted methods
(i.e., 1C and colorimetry); and it is
reasonably simple to operate.
ADS
The ADS collects gas samples with
annular denuders and paniculate matter
on filter collection media. As shown in
Figure 2, the ADS is comprised of
modules arranged in series to collect
various constituents of interest The first
module is a Teflon cyclone similar to that
employed by the TFR. The second and
third modules are two annular denuders
for collection of HN03, HN02, and SO2,
followed by a third annular denuder for
collection of NHa The fifth element is a
filter pack contain ing a Teflon and a nylon
filter.
The nominal sampling rate is 15 Lpm
Ambient air is passed under laminar flow
conditions through the annular space
3
-------
Nylon
Nylon
Oxalic Acid
Oxalic Acid
Nation Nylon
Nylon Nation
i I
Mass FLow
Controller
1 78 SLPM
Mass Flow
Controller
14 3 SLPM
Mass FLOW
Controller
1.78 SLPM
Mass Flow
Controller
14.3 SLPM
Critical
Orifice
1 l/min
Vacuum
Pump
Figure 1. Transition flow reactor sampling system.
Vacuum
Pump
between two concentric tubes. The
outside of the inner tube and the inside
of the outer tube are coated with a
specific gas-absorbing solution. For the
first two denuders, this coating contains
Na2CO3, for the third denuder, the NH3-
absorbmg coating contains citric acid
The annular design permits a reduction
in denuder tube length and an increase
in flow rates over the conventional open-
tube denuder design The third denuder
is followed by a Teflon filter to collect
PS, nonvolatilized nitrates, and particu-
late ammonium, while the nylon filter
collects HMOs that is passed due to
volatilization from the Teflon filter. A
third filter may be added downstream of
the nylon filter to collect any volatilized
ammonium
HMOs and S02 are determined from the
differences between 1C analyses of NCV
and SCu2 in the aqueous extracts of the
first and second denuders Analysis of
the second denuder extract confirms the
near quantitative collection of these
species and permits a quality control
check
HNO2 isdeposited nearlyquantitatively
on the first denuder Deposition of PAN
and NO2 ranged from 1 to 3 percent.
Since deposition of these two species is
very small on the first denuder, the
deposition should be approximately the
same on the second denuder. The
recovery of HNO2, PAN, and N02 as NO2~
on the first two denuders, and the low
deposition of the latter two species,
permits the determination of HNO2 by
using the difference in NO2~ between the
first two denuders. Extract analysis can
be performed by either 1C or colonmetry.
NH3 and NH/ are determined from
colon metric NhU* analysis of the aqueous
extract of the citric acid-coated third
denuder and Teflon filter, respectively
PS and nonvolatilized nitrates are deter-
mined by 1C analysis of SCu2 NOa~ in
the aqueous extract of the Teflon filter.
Nitrate analysis of the 0003 N NaOH
extract of the nylon filter permits deter-
mination of volatilized nitrates Fine PN
is determined by summing the nonvola-
tilized and volatilized nitrates.
The ADS has several limitations:
a. The system does not currently permit
the determination of N02.
b. Care must be taken at high humid-
ities to prevent dissolution and loss
of capacity of Na2COs denuder coat-
ings, and data must be inspected at
low humidities to ensure good col-
lection efficiencies.
c. The collected HNO2 may undergo in
situ oxidation to nitrate on the first
denuder, giving rise to artificially
high values for HNOs and low values
for HN02.
d. Nitrates collected in the cyclone and
not removed during sampling may
volatilize to give artificially high
values of gaseous nitric acid.
e. The loss of HNO3 by volatilization
from the Teflon filter makes inter-
pretation of the acidity determina-
tions difficult.
-------
1
j
1
N
G
n(
P
JFp
'^
x"
if
If*
--H
Tjss»T
[
1
f
+
f
— x
Y
,
i
y ;
h i
1 !
r •
JL
^n
si
3
^
=
r
f *
<
l\
C J
,6 f
Zi a
V /
n
A/
Cyclone ,
\
\
•^^
a. f£ Teflon cyclone
b Aluminum shield
c. NazCOz-coated denuder
d Teflon coupling
e NazCOz-coated denuder
f. Teflo
n rnnnlinn
g. Citnd acid-coated denuder
h. Support bracket
i Velcro fastener
j Filter pack
k. Flow line
1. Light bulb
m. Fan
n. Electrical connection
o Flow control module
p. Sampler box
\1
}o
V>^
^^•^u ^••^_
j
•
r
^
x Flow
Control
hM /.,/_
Figure 2. Annular denuder sampling system.
f. In the absence of an NH3 collection
medium such as a coated filter
downstream of the nylon filter, a
portion of the volatilized paniculate
NH4+ may be lost.
The ADS has advantages, it can be
used for sampling dry deposition constit-
uents for periods of hours to 7 days; it
is modular, easy to install and ship;
sample analysis is by existing, accepted
methods (i.e., 1C and colorimetry); and it
is reasonably simple to operate.
Conclusions and
Recommendations
The chemical-specific sampling and
analysis methodologies for many chem-
ical constituents of importance in eco-
system acidification have been reviewed
and evaluated. Their status is summar-
ized in Table 5 of the full report.
Recommended methods are given for
four species: HN03, NO3~, S042~, and 03.
The bulk of the methods are designated
as favored, i.e., the method, at its current
state of development, holds promise for
providing unambiguous concentration
data for the species of interest and for
achieving routine implementation in an
air monitoring network. Two(i.e., RCOOH
and HCHO) are in the research mode, and
no assessment can be made at this time
For some species, well-established
methodologies may exist (e.g., FTIR and
TOLAS), but their resource and/or
manpower requirements prevent them
from consideratoin in network
application.
The selection of a multiple constituent
sampling system is not clear. The F P can
provide estimates of concentrations of
total inorganic nitrate, total paniculate
sulfate, and SOz and biased estimates
of paniculate NH4* and acidity at a
relatively low potential cost. The TFR and
ADS offer estimates of a larger number
of constituents but at a larger potential
cost. The ADS can provide concentration
estimates of HN02, but not N02; whereas
the TFR can provide N02, but not HN02.
Although successful field trials have
been reported for the TFR, successful
deploymentfor measuring HN03andfine
paniculate NO3~ was not realized by RTI
in the field study described in the report.
As a result, the TFR cannot be endorsed
for HNO3 or fine particulate N03~ meas-
urements at this time. From an opera-
tional perspective, the ADS is marginally
preferable to the TFR. Thus, while the
available evidence does not provide
overwhelming support for the selection
of either the TFR or the ADS for deploy-
ment in an air monitoring network, on
balance the ADS does appear to have a
slight advantage over the TFR. Additional
development and field tests are recom-
mended to provide the comparisons
necessary to assess the merits of two
sampling systems more clearly. It is also
recommended that data quality objec-
tives (DQO's) be clearly established for
the species of interest. This will permit
selection of a sampling system that is
optimized to meet user needs.
The full report has focused on the
analysis and discussion of results of a
methods comparison study conducted in
the fall of 1986 by RTI and others at the
E PA dry deposition site at Research
Triangle Park, North Carolina. This study
consists of 13 daily samples; weekly data
were also collected with the ADS. The
term TFR normally refers to the TFR
system operated daily by RTI throughout
the study; TFR(EPA) refers to the TFR
operated by EPA/ASRL on 5 of the 13
study days.
Selected conclusions and findings
drawn from this study are given below.
• Detection limits for the ADS and TFR
are generally less than 0.15 ppb,
except for NH3 with the ADS at 0.4
ppb and for SO2, NO2, and NH3 with
the TFR at 0.4, 0.5, and 1.6 ppb.
• Both ADS and TFR showed good
precision for the measured species
except fine particle NO3~. In general,
paired daily samples had median CVs
of less than 20 percent.
• Quality control checks indicate
ambient TFR HNO3~ and fine partic-
ulate N03~ results to be suspect. In
24 of 26 cases, TFR Fine NO3~ results
were negative. This invalidated TFR
Fine NO3 results and prevented their
use in subsequent analyses. Results
of the current study do not permit the
endorsement of the TFR for the
determination of HMOs or Fine NO3~.
• Statistical comparisons of ADS and
TFR results show no difference for
Total N03~, particulate NO3~ on Teflon
filters. Total SO42~, and NH3; the ADS
estimate to exceed that of the TFR for
Fine S042~, SO2, and I-T; and the TFR
estimate to exceed that of the ADS
for HN03, Total NH3+NH4+, and Fine
NH4+.
• Total NOa" as measured by the ADS,
TFR, and FP are in good agreement.
Total SO42~ as measured by the ADS,
TFR, and FP
agreement.
are also in good
Nitric acid spiking experiments indi-
cate that appreciable amounts of
HN03 may be retained by Teflon
cyclones, and that HN03 transmission
approaches 100 percent for some
types of Teflon after a brief condition-
ing period.
A substantial difference in particle
collection efficiency by the ADS and
TFR cyclones resulted from sampling
at different flow rates. This is most
apparent for particulate nitrates
where the cyclone catch accounted for
59 percent of the total in the ADS and
85 percent in the TFR.
Denuding the sampled atmosphere of
all the gaseous HN03 and NH3 in the
ADS caused substantial volatilization
of collected fine particulate nitrates:
52 percent for daily samples and 93
percent for weekly samples. This and
the difference in cyclone efficiencies
prevent meaningful direct comparison
-------
of nitrates on the Teflon filter for the
ADSandTFR.
• Nitric acid, as measured by the ADS,
FP, and TFR(FP), a nonstandard
interpretation of TFR data, are in good
agreement with those of the TOLAS
(i.e., the differences between the
TOLAS results and those of each of
the three listed measures are not
significant at the 95 percent confi-
dence level). Two sets of TFR results
were collected by independent oper-
ators (RTI and EPA). Differences
between TOLAS results and one of the
TFR results are significant. The daily
TFR results are consistently much
higher than those of the TOLAS. The
five daily TFR(E PA) measures of HN03
are also consistently higher but are
in better agreement with TOLAS
results.
• Nitrogen dioxide as measured by the
TFR and TOLAS are highly correlated,
although differences between the two
sets of results are statistically
significant.
• Both nitrate and ammonium are
volatilized from the ADS Teflon filter.
A nylon filter is employed downstream
to collect volatilized nitrate. The ADS,
as operated with no provision to
collect volatilized NhU*, undersamples
particulate NH<+. It is recommended
that a citric acid-coated filter be
incorporated in the ADS to collect
volatilized ammonium.
• Total particulate NH/ as measured by
TFR and FP are in good agreement
and exceed measurements with the
ADS.
• (-T concentrations measured with the
ADS exceed those with the TFR
because undenuded NHa in the air
sample reaching the TFR Teflon filter
neutralizes a portion of the H* col-
lected on the Teflon filter with that
sampler.
• S02 concentrations measured with
the ADS exceed those measured with
the TFR by approximately 30%. Labo-
ratory tests suggest that the TFR
results require a correction for
reduced S042~ recovery efficiency (i e ,
80 percent) under humid conditions
• Field studies have demonstrated that
the ADS can be successfully deployed
for sampling periods of 4, 6, 10, 12,
and 22 hours, and up to 7 days. The
results are subject to a potential high
bias for HNO3 (from MONO oxidation)
and potential low biases from inlet
losses, oxidation of MONO and NH4+
volatilization (if a citric acid-coated
filter is not added). Field studies have
demonstrated that the TFR can be
successfully deployed for sampling
periods of 22 hours and 7 days. The
results are subject to a low potential
bias for SOz (if S02 collected on the
nylon filter and a recovery efficiency
correction are ignored). TFR results
may be unreliable for HNOs and Fine
NOa". These findings, although favor-
ing the ADS, do not permit the
complete endorsement of either the
ADS or the TFR for field deployment
without further development and field
testing.
This report was submitted in fulfill-
ment of Contract Number 68-02-4079 by
the Research Triangle Institute under the
sponsorship of the U.S. Environmental
Protection Agency. This report covers a
period from December 13, 1983, to June
30, 1987, and work was completed as
of March 31, 1987.
-------
J. E. Sickles, II, is with Research Triangle Institute, Research Triangle Park,
NC 27709; the EPA authors William A. McClenny and Richard J. Paurfalso
the EPA Project Officer, see below) are with the Environmental Monitoring
Systems Laboratory, Research Triangle Park, NC 2771 1.
The complete report, entitled "Samp/ing and Analytical Methods Development
for Dry Deposition Monitoring," (Order No. PB 87-233 318; Cost: $24.95,
subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
t, c
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
1 Of:"r,/ j ;. , ,
' - •""; 0 3 "--I-
' ' ' - -.'•:*> £
Official Business
Penalty for Private Use $300
EPA/600/S4-87/011
0000329
S
PS
PROTECTION AGENCY
------- |