UNITED
STATES
ENVIRONMENTAL
PROTECTION
AGENCY
NATIONAL
RISK
MANAGEMENT
RESEARCH
LABORATORY
Air
Pollution
Prevention
and
Control
Division
Research
Triangle
Park,
NC
27711
Office
of
Research
and
Development
MEMORANDUM:

DATE:
March
13,
2005
SUBJECT:
Results
of
Recent
Testing
to
Collect
Data
for
Development
of
a
Potential
Instrumental
Reference
Method
for
Hg
CEMS
FROM:
Jeff
Ryan
Office
of
Research
and
Development,
Air
Pollution
Prevention
and
Control
Division
TO:
Bill
Maxwell
Office
of
Air
Quality
Planning
and
Standards,
Emissions
Standards
Division
The
following
data
package
contains
results
from
recent
testing
to
evaluate
and
demonstrate
procedures
and
techniques
that
may
be
suitable
for
potential
development
of
an
Instrumental
Reference
Method
for
Hg
CEMS.
In
support
of
OAR's
efforts
to
develop
procedures
and
guidance
for
monitoring
Hg
emissions
from
coal­
fired
power
plants,
EPA's
Office
of
Research
and
Development
has
provided
technical
support
through
laboratory,
pilot­
plant
and
field
testing.
Recently,
ORD
has
conducted
field
testing
to
identify
and
demonstrate
procedures
that
may
be
suitable
for
the
potential
development
of
an
Instrumental
Reference
Method
for
Hg
CEMS.

A
comprehensive
series
of
tests
were
conducted
in
conjunction
with
field
test
activities
to
evaluate
the
performance
of
Hg
CEMS
with
respect
to
recently
proposed
procedures
and
requirements
for
monitoring
Hg
emissions
from
coal­
fired
power
plants,
including
Hg
CEMs.
The
tests
performed
have
been
described
in
a
separate
Test
Plan.
In
summary,
the
experiments
were
designed
to
assess
the
accuracy
and
precision
of
candidate
reference
Hg
CEMs,
including
the
assessment
of
site
specific
spectral
and
matrix
interference
effects.
The
latter
of
which
are
accomplished
using
the
dynamic
spiking
technique.
The
following
information
contains
preliminary
data
from
recent
dynamic
spiking
testing.
The
quality
of
the
candidate
reference
analyzer
was
also
subjected
to
an
OH
RATA
test.
Data
are
not
yet
available.

Calibration
Error/
Linearity/
System
Bias
Test
Data
Known
concentrations
of
Hg0
and
HgCl2
were
injected
at
multiple
locations
to
demonstrate
system
performance.

Table
1.
HgCl2
at
Orifice
Reference
Tekran
Value
Average
Difference
Error
µ
g/
m3
µ
g/
m3
µ
g/
m3
0.0
0.4
0.4
N/
A
2.8
2.8
0.0
0%
5.5
5.5
0.0
0%
11.1
10.9
­
0.2
­
2%

Table
2.
HgCl2
at
Probe
Reference
Tekran
Value
Average
Difference
Error
µ
g/
m3
µ
g/
m3
µ
g/
m3
0.0
0.3
0.3
N/
A
2.8
2.6
­
0.2
­
7.1%
5.5
5.1
­
0.4
­
7.3%
11.1
N/
A
N/
A
N/
A
Spectral
Interference
Test
Data
Dynamic
spiking
spectral
interference
tests
were
performed
at
levels
or
concentrations
representing
increases
of
25,
50
and
100%
of
native
stack
levels
for
SO2,
NO,
and
HCl.
Results
are
presented
in
Table
3.

Table
3.
Spectral
Interference
Dynamic
Spike
Data
Condition
Expected
Concentration
Observed
Concentration
Difference
Recovery
µ
g/
m3
µ
g/
m3
µ
g/
m3
Cal
Drift
(
Hg0)
@
orifice
7.4
7.9
0.6
107.6%
Cal
Drift
(
HgCl2)
@
probe
5.3
4.5
­
0.8
84.6%
Zero
Drift
@
probe
0.0
0.2
0.2
Stack
6.7
Stack
+
N2
7.8
6.4
­
1.3
82.9%
Stack
+
25%
SO2
8.1
7.4
­
0.7
90.8%
Stack
+
50%
SO2
7.8
8.8
1.0
112.9%
Stack
10.1
Stack
+
25
%
NO
10.0
10.1
0.1
101.1%
Stack
+
50
%
NO
9.5
10.0
0.4
104.7%
Stack
10.5
Stack
+
25
%
HCl
10.1
10.7
0.6
105.8%
Stack
+
50
%
HCl
9.8
9.6
­
0.2
97.7%
Stack
10.3
Cal
Drift
(
Hg0)
@
orifice
7.4
7.9
0.6
107.6%
Cal
Drift
(
HgCl2)
@
probe
5.3
4.6
­
0.7
86.3%
Zero
Drift
@
probe
0.0
0.3
0.3
Matrix
Interference
Tests
Data
The
Hg0
matrix
interference
tests
were
conducted
using
the
Tekran
3310
Hg
°
generator.
Sequences
were
set
up
for
the
Tekran
to
spike
approximately
25%,
50%,
and
100%
on
top
of
stack
concentrations.
Hg0
dynamic
spiking
results
are
presented
in
Table
4.

Table
4.
Hg0
Dynamic
Spike
Test
Data
Baseline
Expected
Observed
Spike
Condition
Stack
Concentration
Concentration
Recovery
ug/
m3
ug/
m3
ug/
m3
%
Stack
Conc.
+
25%
Hg0
7.0
9.1
9.2
103%
8.1
10.1
10.5
111%
8.5
13.3
13.6
105%
8.8
13.6
13.8
104%
8.9
16.5
16.7
103%
9.0
16.6
16.8
103%

Stack
Conc.
+
50%
Hg0
6.8
8.9
9.0
101%
6.8
9.0
9.2
109%
7.9
12.8
12.8
100%
7.8
12.7
13.1
107%
8.2
15.8
15.6
98%
7.5
15.2
15.6
105%

Stack
Conc.
+
100%
Hg0
7.7
9.7
9.8
102%
7.6
9.6
9.8
107%
7.7
12.6
12.9
105%
8.1
12.9
13.1
104%
7.7
15.4
15.4
100%
7.6
15.3
15.4
101%
The
HgCl2
matrix
interference
tests
were
conducted
using
the
HOVACAL
oxidized
Hg
gas
generator.
HgCl2
was
spikeed
to
add
approximately
25%,
50%,
and
100%
on
top
of
stack
concentrations.
HgCl2
dynamic
spiking
results
are
presented
in
Table
5.

Table
5.
HgCl2
Dynamic
Spike
Test
Data
Condition
Expected
Concentration
Observed
Concentration
Difference
Recovery
µ
g/
m3
µ
g/
m3
µ
g/
m3
Cal
Drift
(
HgCl2)
@
probe
5.3
4.7
­
0.7
87.7%
Zero
Drift
@
probe
0.3
0.3
Cal
Drift
(
Hg0)
@
orifice
4.8
5.1
0.3
106.0%
Stack
7.4
Stack
Values
+
N2
6.7
6.7
0.0
100.7%
Stack
Values
+
50%
Hg0
8.3
7.9
­
0.4
95.3%
Stack
6.8
Stack
Values
+
100%
Hg0
9.9
9.3
­
0.6
93.9%
Stack
6.9
Stack
Values
+
200%
Hg0
13.5
12.5
­
1.0
92.3%
Stack
Values
+
N2
6.4
6.3
­
0.1
99.1%
Stack
6.7
Cal
Drift
(
HgCl2)
@
probe
5.3
5.4
0.1
101.9%
Zero
Drift
@
probe
0.0
0.6
0.6
Cal
Drift
(
Hg0)
@
orifice
4.9
5.2
0.3
106.1%

Summary
and
Observations
The
data
presented
have
only
recently
been
collected
and
as
such
are
preliminary
and
by
no
means
fully
interpreted.
As
can
be
seen
from
the
data,
all
dynamic
spiking
recoveries
are
within
90
­
110%
of
expected
concentrations.
At
1st
interpretation,
these
recoveries
should
be
considered
excellent
and
appear
to
indicate
the
absence
of
spectral
or
matrix
effects.
During
the
conduct
of
these
tests,
it
was
observed
that
relatively
rapid
changes
in
stack
Hg
concentrations
are
possible
and
cause
difficulties
in
establishing
an
accurate
stack
baseline
concentration.
An
accurate
baseline
concentration
is
needed
to
accurately
calculate
expected
concentrations.
Existing
expected
concentration
are
based
on
the
average
of
baseline
stack
concentrations
before
and
after
each
dynamic
spiking
data
point.
The
existing
data
will
be
examined
further
in
order
to
optimize
further
testing.
Shorter
periods
between
baseline
measurements
will
greatly
alleviate
this
problem.
