STACK SAMPLING 1119
The EPA Test Method 104 write-up contains detailed
instructions, along with a lists of references. It should be
read in detail before the Method is attempted. As with all of
these methods, testing should be performed only by trained
and experienced personnel using equipment and materi-
als designed for this purpose. This is especially true for this
Method, considering the hazard associated with Be exposure,
as well as the perchlorate danger.Test Method 105Test Method 105 is not a stack testing method. Instead, it is
the method used by EPA for estimating the maximum pos-
sible emissions of mercury from sewage sludge treatment
plants. This is accomplished by measuring the mercury con-
tent of the feed sludge, multiplying this by the maximum
feed rate to calculate the maximum emissions if all of the
mercury in the sludge were to go up the stack. Method 105
thus serves as an inexpensive screening method for use at
plants that expect their emissions to be well below the emis-
sion standard.
The Method requires the collection of about one liter
of sludge each half hour for 8 hours. The samples are then
combined into one, weighed, digested in potassium per-
manganate, and analyzed as in Method 101A, using atomic
spectrophotometry.
Details, including the range, limitations, reagents, and
equipment, may be found in the method write-up. Even
though the field portion of this test is relatively straightfor-
ward, it should be performed only by trained personnel.Test Method 106Test Method 106 is used to measure the emissions of vinyl
chloride monomer (VCM) in stack gas. It does not measure
VCM in particulate matter. Stack gas is withdrawn from the
centroid of the stack into a tedlar bag using the bag-in-a-box
technique that isolates the sample from the pump. The sample
is then analyzed directly using a gas chromatograph-flame
ionization detector (GC-FID).
The sampling probe for this method is a standard stain-
less steel probe. The rest of the sampling train is unusual,
however, because most of it is actually never seen by the
sample. The probe is connected directly by new teflon
tubing to an empty tedlar bag (usually between 50 and 100
liters capacity). The bag is sealed in a large box that is, in
turn, connected to a needle valve, pump, carbon tube, and
rotameter for measuring the flow. As the pump evacuates the
air in the box, the bag sucks gas from the stack. Through the
use of this indirect pumping procedure, there is no need for
concern for contamination from the pump, or loss of VCM
in the pump. While elaborate box configurations are avail-
able, a 55 gallon drum with two fittings and a sealable top
works fine.
Prior to sampling, the bags and the box should be leak
checked. Just before sampling, the bag should be partially
filled with stack gas to condition the bag and purge the
sample lines. It is then emptied by switching the pump to thebag fitting. Sampling then proceeds at a fixed rate propor-
tional to the stack flow rate.
The GC-FID is calibrated by injection of 3 VCM stan-
dard gases prepared from a single certified 99.9% cylinder
of VCM gas or cylinder standards appropriately certified
by the manufacturer. During calibration and subsequent
sample analysis, the VCM peaks must be of sufficient size
and must not overlap with interfering peaks (such as aceta-
dehyde). Procedures for calculating overlap are provided in
Appendix C to 40 CFR Part 61 (right after the NESHAPS
Test Methods). Procedures for minimizing overlap and inter-
ference are mentioned in the Method but are not specified.
This is why an experienced GC operator is required for this
analysis.
Immediately before analyzing samples, the analyst must
analyse 2 audit gas cylinders supplied by EPA or another
independent party. These audit cylinders contain concentra-
tion of VCM in nitrogen in 2 ranges to ensure the accuracy
of the analytical procedures.
The sample is then injected directly into the GC-FID.
The ratio of the peak height to peak area for the VCM peak
is compared to the ratio for the nearest peak standard peak.
If they differ by more than 10%, interference is probably
present and an alternate GC column is needed. It is strongly
recommended that the laboratory be given actual samples of
stack gas for analysis at least a day before on official test is
to take place. This will allow time to select appropriate levels
for the standard gas concentrations and to determine whether
alternate GC columns are required.
The concentration of VCM in the bag is then calculated
using the GC calibration curve and the temperature, pressure,
and humidity in the stack. This process is usually repeated
for three test runs.
After the sample is extracted from the bag, the bag must
be leak checked by filling it with air and connecting it to a
water manometer. Any displacement of the manometer after
10 minutes indicates a leak. The box must also be checked
for leaks. This is accomplished by placing the bag in the
box-and evacuating the bag with a rotameter in line before
the pump. Any displacement of the rotameter after the bag
appears empty indicates a leak.
During the entire sampling and analytical procedure,
extreme can must be taken to avoid exposure to the VCM
gas, which is carcinogenic. The carbon tube included in
the sampling train is for the purpose of absorbing VCM.
Nevertheless, the tubing existing the pump should be aimed
well away from the sampling team. Similarly, all labora-
tory work should be performed in appropriate vented or
hooded areas. This is one more reason why this test for
VCM should be performed only by experienced personnel
who are extremely familiar with their equipment and with
the procedure.Test Method 107Test Method 107 is not a stack testing method. It is used for
determining the Vinyl Chloride Monometer (VCM) content
of water or slurry samples associated with the manufacture ofC019_003_r03.indd 1119C019_003_r03.indd 1119 11/18/2005 11:07:17 AM11/18/2005 11:07:17 AM