GUIDELINES FOR
             COMBUSTION  SOURCE
SULFURICACID EMISSION  MEASUREMENTS
                    FEBRUARY 1977
                 * R. F. Maddalone
        EPA PROJECT OFFICER:  Dr. R. M. Statnick
                     Prepared for
    INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
          ENVIRONMENTAL PROTECTION AGENCY
     RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711

              CONTRACT NO. 68-02-2165
                 TASK ORDER NO.  13
                                           TRW DOCUMENT NO.
                                           28055-6005-RU-OO
                     TRW
                  DfFffVSf A/VDSPACI SYS TIMS GROUP
              P A H K •  REDONDO

-------
                                       TRW  Document  No.
                                       28055-6005-RU-OO
          GUIDELINES  FOR COMBUSTION

  SOURCE  SULFURIC  ACID  EMISSION  MEASUREMENTS


                February 1977


            By:  R. F.  Maddalone

  EPA Project  Officer:   Dr.  R. M.  Statnick
                Prepared for
Industrial  Environmental  Research Laboratory
     Office of Research and Development
      Environmental  Protection Agency
    Research Triangle  Park, N. C.  27711
          Contract  No.  68-02-2165


              Task  Number  13
              Approved  by:
                           R.  F.yMaddalone
                           Task Order Manager
             Approved  by
:    CJ4LJ
  C.  A.  Flegalff
                           Program  Manager

-------
                             ACKNOWLEDGMENT
     This document summarizes the procedures developed on Task 13, Under-
standing S03 Data on EPA Contract No.  68-02-2165, Sampling and Analysis of
"Reduced" and "Oxidized" Species in Process Streams.  The Chemistry and
Materials Laboratory Applied Technology Division was responsible for the
work performed on this task.  The work was conducted under the EPA Project
Officer Dr. R. M. Statnick,  Environmental  Research Center, Research Triangle
Park, North Carolina.  Dr.  C. A. Flegal was the Program Manager and the
Task Order Manager was Dr.  R. F. Maddalone.  The laboratory tests during
the development of these procedures were done by Mr. Morton L. Kraft,
Mr. David R. Moore, and Mr.  Maynard D. Cole.  The overall review and
support, during the field test program, from Mr. Richard G. Rhudy of Bechtel,
Mr. Steven Newton, and Mr.  John Lawton of the TVA has been greatly appreciated.

-------
1.0  INTRODUCTION
     The purpose of this document is to summarize the controlled condensa-
tion system  (CCS) approach to HgSO^ (S03) measurements in gaseous process
streams.   Included is  information on the scope of the CCS, hardware and
methodology, and labor and cost estimates to implement this procedure.
This document is not meant to be a detailed procedure document.  TRW has
published  (Reference 1) a detailed procedure manual describing this SO.,
sampling and analysis methodology.  Both of these were prepared for the
Industrial and Environmental Research Laboratory of the Environmental
Protection Agency, Research Triangle Park, North Carolina, as part of
Task 13 of Contract No. 68-02-2165.
     The technical objective of that task was to develop a stack sampling
procedure for the measurement of the mass emission rate of sulfur trioxide
(HpSO* vapor) within a precision of tlO%, but not to exceed a precision of
+20%.  The method chosen on the basis of previous experience (Task 02  of
that program) was the Controlled Condensation System (CCS).  A test program
was then established to evaluate the CCS under simulated stack conditions.
     As a result of this intensive laboratory and field test program,  we
feel that this ^SO* measurement system offers the best method available
to monitor H2SO» emissions from combustion sources.

2.0  SCOPE
     The CCS is the manual sampling and analysis approach to measure HgSO*
(SO.,) in process gas streams.  This approach is applicable to process  streams
                     3                               3
under high (> 1.2 g/m  -0.5 gr/cfm) or low (< 1.2 g/m ) mass loadings, gas
temperatures as high as 300°C (572°F) and S02 concentrations of up to  6000 ppm,
The nominal measurement range is 0 to 50 ppm using the recommended flowrates.
This range can be expanded by manipulation of sampling times or flowrates.

3.0  DEFINITIONS
3.1  Particulate Matter
     For the purpose of this document, particulate matter is any material
that can be collected by a heated (250°C) quartz filter.

-------
3.2  Sulfuric Acid/S03
     Since under most process gas conditions S03 will  exist  as  a liquid
aerosol  of  H2SO,, the two terms SCL and
in this  document.
                                             will  be  used  interchangeably
4.0  METHODOLOGY
     The development of this method grew out of a basic  gap  in measurement
methodology for HUSO,  in particle-laden gas streams.   Currently  there are
two EPA compliance methods for SO  measurements.   In  both  EPA Method 6 and
                                A
Method 8 (References 2 and 3), the HUSO, is separated from S02 by passing
the gas stream through an impinger of 80% isopropanol (IPA)  followed by
two impingers  of  3% HLOp.  The IPA collects the H2SO. and  passes the S02
into the Hp02  impingers where it is oxidized and  collected as H2SO».  While
these methods  are the  EPA compliance methods, problems have  been found in
trying to measure HUSO, in particle-laden streams using  these systems.  In
both procedures inefficient particle removal prior to the  first  impinger
leads to variable positive interferences from sulfate containing fly ash,
and catalytic  oxidation of S02 in the IPA impinger.   The precision of Method
8 has been studied (Reference 4), and it was shown to be in  excess of ±50%.
Because of these  problems, an alternate H2SO^ method  is  needed.
      THERMOCOUPLE
         WELL
                      60 MM MEDIUM
                       FRIT
       18/9
      3 CM
                                                         18/9
                           ywwwwwwv
                                                                GAS
                                                                "FLOW
              FIGURE 1  -  CONTROLLED CONDENSATION COIL (CCC)
                                   3

-------
     The controlled condensation procedure (Reference 5) for H^SO. is a
likely candidate because of its simplicity and clean separation of particulate
matter, S02, and H2$04.  This procedure is based on the separation of HgSO^
from S02 by cooling the gas stream below the dewpoint of H2SO, but above the
H20 dewpoint.  Cooling is accomplished in a water-jacketed coil (Figure 1)
where the H2SCK is collected.  Any aerosol not collected in the coils is
collected on the glass frit.  Particulate matter is removed from the gas
stream prior to reaching the cooling coil by means of a quartz wool plug
placed in the heated glass probe.  While the basic idea of a controlled
condensation to separate S03 from S02 had merit, the system design and
operating parameters had several built-in difficulties.  Consequently, a
laboratory study was initiated by TRW to optimize:
     •  Probe, filter, and coil heating temperature
     •  Particulate removal system
     •  Filter selection
     The result was a  method  capable  of measuring  H2S(h  with  a laboratory
precision of ±7%,  using  an  improved filtration  system  and  an  inert filter
material.  The following sections  will describe the  operation of the modified
controlled condensation  system.
              ADAPTER FOR CONNECTING HOSE

                       TC WELL
                                                       RUBBER VACUUM
                                                         HOSE
        FIGURE 2 - SCHEMATIC OF CONTROLLED CONDENSATION SYSTEM (CCS)

                                      4

-------
4.1  Hardware
     The  hardware for the CCS is  normally available  in  any lab that  is
periodically  involved in making any  pollution measurements, but several
components will  have to be fabricated.   Figure 2 is  a view of the overall
layout of the  CCS system.  The following sections describe each component.
4.1.1  Probe
     The  probe should be made of  Vycor  or quartz for the best chemical  and
thermal stability.   Commercially  available probes can be used if they  can
maintain  a temperature of 315°C (600°F) for extended periods of time.
Figure 3  is an alternate probe design with the main  components labeled.
                                                   PROBE T.C.
                                                              18/9
                                         SILICONE
                                         STOPPER
                SILICONS
                STOPPER
                                                                 GLASS HEATING
                                                                 TAPE LEAD
        VYCOR TUBE
     TEFLON UNION
                                                         STACK
                                                         T.C.
     GAS FLOW
(7) STOPPERS SHOULD BE AWAY FROM HEATING TAPE
(T) ASBESTOS COVER SHOWS SLIGHT OVERLAP
      FIGURE 3  -  PROBE DESIGN FOR CONTROLLED CONDENSATION SYSTEM  (CCS)
                                      5

-------
4.1.2   Filtration System
     Figure  4  details  the recommended design for the quartz filter holder.
This filter  holder consists of a modified 40/50 standard taper quartz joint.
The modifications included adding a coarse quartz frit and an extension
tube to the  male  joint to act as a pressure seal when a Tissuequartz filter
(Pallflex Corp.,  Pulman, Conn.) pad is in place.  Tissuequartz filters are
recommended  because of their proven inertness to HpSO-.  Ball-and-socket
joints, which  provide  a degree of flexibility, are used to connect the
filter holder  to  the probe and controlled condensation coil.
     The filter system is heated by a heating mantle (Glass-Col, Terre Haute,
Ind.) so that  a £as_ out, temperature of 288°C (550°F) is maintained.  This
temperature  is required to ensure that none of the H^SO, will condense in
the filter holder or on the Tissuequartz filter.
4.1.3  Controlled Condensation Coil  (CCC)
     The CCC consists  of a Graham condenser modified to accept a medium frit
(Figure 1).  This device is not commercially available, but requires
fabrication.  The water jacket is maintained at 60°C (140°F) by a heater/
recirculator.  This temperature is adequate to reduce the flue gas to below
the dewpoint of H$0.
      18/9
      BALL
                        SPRING
                        ATTACHMENT
                        HOOKS
TISSUE QUARTZ
FILTER
                      STANDARD
                      TAPER QUARTZ
                      40/50
                         18/9
                        SOCKET
SEAL
EXTENTION   EXTRA COARSE
TO STD.      QUARTZ FRIT
TAPER JOINT
                       FIGURE  4.    QUARTZ FILTER HOLDER

-------
4.1.4  Impingers, Pump, and Dry Test Meter
     The two impingers shown in Figure 2 are filled with 3% H202 and silica
gel to remove S02 and HpO, respectively.  In some systems because of the
high S02 levels, a second H2CL impinger might be required to remove all of
the S02.
     The pump should be capable of 28 &pm gas flow and be leak-free.  Pumps
larger than this tend to be hard to control, and small sizes will not have
the reserve capacity to overcome a pressure drop that might occur in the
system.  For similar reasons the dry test meter should be sized so that
accurate readings are possible at the 8 &pm flowrate.  Stainless steel
construction with a capacity of approximately 680 Jiph is recommended,
4.1.5  Miscellaneous Equipment
     Besides the above specific equipment requirements, additional apparatus
is associated with the CCS.  Variacs, thermocouples and a temperature
readout are also necessary.  The Variacs will be used to control the probe
and heating mantles according to the temperatures read by the thermocouples.
The most important point in considering the selection of readout is that it
has an internal cold junction and that the thermocouple junction be matched
to the readout junction (i.e., iron/constantan with iron/constantan).
Laboratory equipment is minimal and requires nothing more than a standard
quality control laboratory would have.
4.2  Sampling Procedures
     The sampling time is normally set at one hour at 8 Jipm.  This time and
rate will normally provide enough sample for accurate titrations.  The best
indicator, however, is to watch the H2S04 fog creep along the coil.  When
1/2 to 2/3 of the coil  is fogged, sampling can stop.
     Normally, when a particle stream is sampled with a probe, the sampling
velocity at the nozzle is matched to the velocity in the duct or stack
(isokinetic conditions).  Isokinetic sampling prevents larger or smaller
particles from being preferentially collected depending upon whether the
velocity at the nozzle is respectively less or greater than stack velocity.
Under most conditions,  the H2SO. will exist as a vapor or a fine aerosol
which will  behave as a  gas, thus the collection of the H,,S04 should not be

-------
affected by nozzle flowrates.  In fact, it is advised that the probe nozzle
be turned downstream during the sampling period.  Positioning the nozzle  in
this fashion will reduce the quantity of large particles (^Sy) reaching
the filter section.  The lesser amount of material on the filter, the
better the overall recovery of the HgSO^ by the CCS.
     Though a gas or fine aerosol is being sampled, it is possible that
stratification can occur in the duct or stack.  While studies (Reference  6)
have shown that gas stratification in ducts is on the average low (±15%),
it is recommended that the probe be placed in the center of the duct after
an area of high turbulence to ensure adequate mixing  of the gas stream.
     Finally it is possible to predict the expected variance in the average
HgSO, value.  From HpSCh sampling tests performed at  a coal-fired utility,
data on the hourly and daily fluctuations in HpSCh concentrations has been'
compiled.  These results indicate a 10-hour variance  of ±32.4% and a daily
variance of ±57.9%.  Using these values, Figure 5 is  generated.  Thus, if it
is desired to estimate the average H^SO, within 20 percent, a sampling plan
of five days with five samples per day, or six days with two samples per  day,
or seven days with one sample per day is required.
4.3  Analysis Procedures
     The amount of H^SO* in the condensate can be measured either by sulfate
or H  titrations.  Because of the simplicity and sensitivity of the acid/base
titration, it is the preferred procedure.  The recommended acid/base titra-
tion uses Bromophenol Blue as the indicator, since the endpoint of the NaOH
and HgSO^ titration falls near the pH range (3-4.6) of Bromophenol Blue's
color change.
     In practice the probe and coil are rinsed with D.I. H^O and the solution
returned to the lab for analysis.  The filter is also saved and extracted
with D.I. HpO.  In most cases, if the correct temperatures are maintained,
no acid should be found on the filter or probe.  It is possible that some
particulates might have a strong absorptive quality toward H2SO», without
reacting with the HgSO^.  In this case any acid found on the filter should
be included in the ppm H^SO, calculations along with  the condensate rinse.
                                    8

-------
     o
     I/O
      CN
     X
     O
     LU
     oo
     LLJ
      _
     O


     u
     u
     LU
     Q_

     X
10
                              234


                               SAMPLES PER DAY



                                FIGURE 5.


   EXPECTED COEFFICIENT OF VARIANCE (CV) OF THE H2S04 MEASUREMENT BASED


                      ON THE NUMBER OF SAMPLES TAKEN
     Studies (Reference 7) have shown that 0.3 g of fly ash on the filter

can reduce the H2S04 recovered by 12%.  Following the above sampling

procedures will reduce the fly ash collected and minimize the loss of HS

-------
4.4  Quality Control
     Estimation of the error in H2$04 value obtained by the CCS has  been
discussed in Section 4.2.  While a test is being carried out,  there  must  be
some activities devoted to insuring the quality of the data collected.  In
order to check the accuracy of the titrations performed on the G/R samples,
an independent check of the NaOH solution and titration method is  required.  .
A blind standardized sample of H2SO^ approximately 0.01 N should be  analyzed
by the test personnel every two weeks.  The analysis of the sample should be
completed in triplicate and reported to 3 places (O.X Y Z). Analysis of
this sample will provide information on the precision of the titrations and
accuracy of the results.
     In most analytical tests it is possible to spike the unknown  with a
known amount of the substance under analysis to determine the  accuracy of
the measurement.  This type of quality assurance is not possible when one
is dealing with a 5-ft. stack.  Consequently, alternative schemes  must be
devised to monitor the quality of the data being collected.
     In many cases where inlet and outlet HUSO, values are measured, it was
possible to monitor the HUSO, results by plotting the inlet and outlet HUSO*
values against each other.  Since there is a direct correlation between
outlet and inlet concentrations, a simple control chart using  regression
analysis can be used (Figure 6) to evaluate the data.  The area between the
2 and 3 limits is the warning zone.  A point falling in this area  indicates
that the measurement system may be out of control.  The region between the
2a and +2a limits should contain, in the long run, 95 percent  of all future
paired measurements.  A point falling outside of the 3a limits indicates
that the measurement system is out of control.  The region between the -3a
and the +3o limits should, essentially, contain all future paired  measure-
ments of inlet and outlet contamination.
     The a limits should be based on a "large sample", say * 30, of  paired
measurements.  If, for a particular environmental situation, the sample size
is less than 30, interim charts should be established using tolerance limits.
                                    10

-------
                                                       •3(7 LIMIT
                                                       '2
-------
                 TABLE 1

       EQUIPMENT/FABRICATION COSTS
FOR CONTROLLED CONDENSATION SYSTEM (CCS)
COMPONENT
1. Probe (3 Ft)
a. Vycor Liner with Joint
b. Heating Tape
c. 304 SS-1" Tubing
d. Thermocouples (2)
e. Thermocouple Connectors
f. Teflon Union
g. Asbestos Cloth
h. Miscellaneous
Probe Subtotal
2. Filter Holder
a. Standard Taper Quartz
Joint
b. Quartz Coarse Frit
c. Quartz Ball & Socket
Joints
Filter Holder Subtotal
3. Controlled Condensation
Coil
a. Graham Condenser
b. Ball & Socket Joints
c. Medium Frit
Controlled Condensation
Coil Subtotal
4. 10-Position Digital
Temperature Readout
5. Heater/Recirculator
6. Impingers (3)
7 . Pump
8. Dry Test Meter
9. Heating Mantle
10. Variacs (2)
11. Tissuequartz Filters
12. Miscellaneous (Hoses,
Clamps, etc.)
TOTAL CONTROLLED
CONDENSATION SYSTEM COST
SUPPLIER
-
Ace Glass
Fisher Sci.
Omega Engineering
Omega Engineering
Swagelok
Fisher Sci .
Fisher Sci.
.
Ace Glass
Ace Glass
Ace Glass
—
Ace Glass .
Ace Glass
Ace Glass
Omega Engineering
Techne
Ace Glass
Gast Pump
Fisher Sci.
Glass-Col.
Fisher Sci.
Pall -Flex
Fisher Sci.
C
MATERIAL
COST
-
$ 80
$ 15
$ 20
$ 17
$ 10
$ 25
$ 20
$ 30
$ 217
-
$ 50
$ 15
$ 30
$ 95
—
$ 20
$ 10
$ 5
$ 35
$ 500
.$ 340
$ 140
$ 250
$ 500
$ 150
$ 100
$ 50
$ 75
$2452
ST
FABRICATION
HOURS
4
1
0
0
0
0
0
0
0
5
4
0
0
0
4
4
0
0
0
4
_
-
.-
-
-
-
-
-

13
                    12

-------
                                TABLE 2
              MAN-HOUR ESTIMATES FOR TYPICAL  CCS  TEST RUN
       ACTIVITY
TIME (HOUR)
           REMARKS
TEST RUN
  - Site Set-up
  - Run Time
  - Sample Recovery
  - Clean-up
             TOTAL
   1.5
   1.0
   0.5
   0.25
   3.25
Turnaround time between runs
is less
More or less depending on
HSO, concentration
^ 2.00 hours per test after
initial set-up
LABORATORY
  - Analysis
  - Data Reduction
  - Quality Assurance
             TOTAL
   1.0
   0.25
   0.25
   1.5
Triplicate analysis plus blank
determination
H2S04 concentration
      calculations
Plotting of data
                                   13

-------
6.0  REFERENCES

1.   R. F. Maddalone, and N. Garner, "Process Measurements  Procedures:
     Sulfuric Acid .Emissions,"  TRW Defense and Space Systems,  EPA Contract
     No. 68-02-21,65, Task Number 1^. February 1977
2.   Federar Register, 41_ (111), 23087 0976).

3.   Federal Register, 41_ (111), 23083 (1976).

4.   H. F. Hamil, R. E. Camann, and D. Thomas,  "Collaborative Study of
     Methods for the Determination of Sulfuric  Acid Mist and Sulfur Dioxide
     Emissions from Stationary Sources," S.  W.  Research Institute,
     PB-240-752/6, November 1974.

5.   Annual Book of ASTM Standards, Part 26, Method D-3226-73T,  American
     Society for Testing and Materials, 1916 Race Street, Philadelphia,
     Pennsylvania, 19103 (1975).

6.   E. F. Brooks, and R. L. Williams, "Process Stream Volumetric Flow
     Measurement and Gas Sample Extraction Methodology," TRW Defense and   :
     Space Systems, EPA Contract 68-02-1412, Task 13, November 1975.

7.   R. F. Maddalone, S. Newton, D. Rhudy, and  R. M. Statnick, "Laboratory
     and Field Evaluation of the Controlled Condensation System  (Goksoyr-
     Ross Coil) for S03 Measurements in Flue Gas Streams," to be presented
     at the June, 1977 APCA Meeting in Toronto, Ontario, Canada.
                                    14

-------