                      EPA HCl Performance Specification 
                         Draft Conference Call Summary
                     Thursday July, 21 from 1:30-2:30 EDT.
        Call in Conference Number 1-866-299-3188, code is 9195415511.  
                         Stakeholder Discussion Topic:
      What are the best approaches for in the field QC/recovery checks. 
                                       
1) Welcome  -  Introduction 
   a) EPA plans to have stakeholder calls on various topics at a frequency of once every two weeks
   b) We will work from the list of issues previously sent to the stakeholder.  The specific topic from the list will be announced by email prior to the meeting to allow everyone to prepare.  We will try to limit discussion to the announce topic so stakeholders can choose which conference calls to attend.  We plan to cover all of the topics in our original summary and may add topics that arise from our ongoing discussions with stakeholders.
   c) EPA will start working on a draft of the performance standard in August based in part on the information from the stakeholder calls.  Over the next six weeks stakeholders (Table 1) participating in this call will start seeing draft sections of the Performance specification and is associated procedure.  Comments on the drafts are welcome so we can get your input well before we have a complete draft.
   d) Our timing remains to have the first full draft for Stakeholder review by the end of the calendar year (2011).

2) Field QC Topics Discussion - In field hot wet through the probe spiking

When area source tests were performed for cement plants, Jim Peeler - Emissions Monitoring Inc. used Method 321Measurement of Gaseous Hydrogen Chloride Emissions At Portland Cement Kilns by Fourier Transform Infrared (FTIR) Spectroscopy.  This method requires through the probe spikes, including specifications for the maximum amount of spike introduced through the probe, and the relative concentration of spike that can be used. This QC approach provides an evaluation of the measurement system since it demonstrates that you can successfully transport a known concentration of gaseous HCl through the probe and filter cake into the heated instrument cell.  For cement kiln HCl measurement, the filtered particulate matter is comprised solids that react with HCl.  Through the probe spiking provides an assessment of both the sampling/analysis system and it tests the mathematical algorithms used in some equipment to correct for interferences in the sample matrix.  

   a) The spike is limited to a maximum of 10% of the sample gas stream.  The concentration target for through the probe spiking is the approximate concentration of the gas stream or 5 ppm is the gas tested is less than 5 ppmv.  For example, spiking a gas stream of 5 ppm involves using a certified gas at 50 ppmv HCl which is diluted 10/1 resulting in the addition of a spike of 5 ppm in addition to the stack gas HCl. The measured concentration of native plus spike is approximately double the concentration of the native HCl. In this approach the spike gas dilutes the native concentration by approximately 10%. 
   b) The sample gas flow rate and the spike gas flow rate must be measured very accurately or the tester can inject a known concentration of tracer such as SF6 to determine the exact dilution.
   c) The calibration gas standard is also delivered directly to the measurement device to determine response of the standard gas without sampling system or sample matrix bias.  This QC check is the relative concentration or relative indicator used to determine spike recovery.  Direct instrument measurement of the spike gas provides the exact response of the spike gas regardless of the absolute concentration of the certified gas cylinder.  
   d) Spike recovery can be determined in different ways as described in the ASTM FTIR measurement procedure for Stationary Sources.  Recovery based requirements permit the measurement procedure to be performance based rather than the more restrictive performance requirements of a prescriptive procedure. 
      i) Spike recovery can be based on the total expected response for the native plus the spike or
      ii) Spike recovery can be based on the incremental response resulting from the spike addition to the native concentration.
   e) For sources with highly variable HCl emissions there is a question of whether two analysis units are required so the instantaneous native HCl can be separated from the through the probe spike response. In both the ASTM procedure and Method 321, there are cautions about spiking when the process is not stable.  One alternative useful for highly variable sources is calculating the average of native concentration just prior and just after the spike interval.  Furthermore, if a the QC spike check fails and the tester believes the cause is temporal variation of the source, the spike can be repeated when the source is relatively constant.  Since the measurements are continuous, a simple statistical test on process variability may be useful to identify when the measurements are out of control or the failure is due to process variability.
   f) EPA has a guideline document for monitoring non-criteria pollutants that addresses these issues.
   g) Allen Rilling - ABB: The spiking system through the probe identifies sampling system bias.  What is the issue spiking at the analyzer once the initial performance of the system is verified.  Jim Peeler: Through the probe spiking is valid for a period of time until the accumulation of material on the filter from the process biases the results.  For cement kilns the filter collects calcium oxides and other material that reacts with HCl.  Sulfur Oxides compete for adsorption sites on the filtered material so through the probe spiking needs to be done periodically to confirm the measurement system meets bias requirements.  During measurement programs for HCl, testers are required to verify the instrument performance before and after each test run. The measurements requirement seems excessive for a continuous monitoring system but is standard for periodic measurement programs.  However, periodic through the probe is necessary to determine when sufficient material was accumulated on the filter to significantly bias the monitoring results.
   h) Robin Segall - EPA: suggests that if daily calibration checks are required as system calibration versus an analyzer calibration.  A daily calibration check through the filter will provide information on the bias caused by the filtered solids as opposed to an instrument only calibration check that provides information on instrument drift and span.  System calibration through the filter should be done daily but there's not enough information on cement kiln continuous monitoring to evaluate how often the analyte spike should be performed. Lime plants, downstream of limestone scrubbers and cement kilns are probably the worst case examples for needing through the probe analyte spikes.
   i) Jim Peeler: If daily system calibration checks are required then the job of the analyte spike technique performs is the same as the function of a RATA.  That is to detect excessive sampling system bias and measurement interferences.  Currently he does accuracy assessment on a quarterly basis.  He suggests doing initial certification including a series of through the probe spikes during a seven day certification period and then quarterly thereafter.  
   j) Robin Segall:  Question, are you proposing to substitute analyte spiking for a RATA against a reference method?  Jim Peeler: Yes, thus avoiding the issue of using reference Method 321 for HCl based on FTIR to evaluate an FTIR CEM. If a facility failed the quarterly through the probe analyte spike then there could be a requirement to do the spiking more frequently.
   k) Jim Peeler:  Agreed that calibration checks include the entire system for regulatory programs and the requirement in this PS should be consistent with that practice. If a system calibration check fails, testers would immediately run an analyzer calibration check to eliminate drift or response changes as the cause of the system calibration check failure.
   l) Allen Rilling : shared his concern about the amount of gas required to perform system calibration checks through the probe.  If quarterly checks verified no system bias, he asked if direct instrument calibration checks would be adequate.
   m) Laura Kenner: For cement kilns, quarterly through the probe QC checks seem too infrequent.  Allen Rilling: Once you've built up a history of passed QC checks, then less frequent through the probe QC checks seem to make sense and would reduce costs for calibration gas as well as increase uptime for source monitoring.
   n) There was some discussion of needing to do a direct instrument and a system calibration check sequentially to ensure that changes in the calibration gas concentration did not bias the QC check.  This assumes that the calibration is not stable and that issue will be discussed on a separate call.
   o) Robin Segall:  Using HCl CEMS in a compliance mode will require performing system calibration checks on a relatively consistent basis.  The starting point would probably be a daily calibration check since EPA established this frequency for our other compliance programs.  
   p) Allen Rilling:  Once you have an analyzer that demonstrates unbiased measurement for a year then you have support for performing less frequent calibration checks.  He asked if it was possible to base the frequency of system calibration checks on a past experience history for a specific monitor installation.
   q) Robin Segall:  If we wrote such a Performance Specification, the data from an interval between a failure and the previous passed system calibration check would be invalid data.
   r)  Jim Peeler:  Appendix F invalidates data from the time a QC check or audit is failed forward and does not retroactively invalidate data because a source operator might be encourage to fail a test to invalidate data the facility did not "like".  For systems like FTIR that measure multi-components, if the system is measuring one component correctly the instrument is probably measuring the other components that get to the analyzer correctly.  That would mean a calibration check for one component would assure that the instrument was operating correctly for all components.  The most reactive component being measured might serve as a surrogate for the rest of the multi-component measurements and that component will probably be HCl.
3) Alternatives to hot  -  wet  -  through the probe QC checks
   a) Rich Brown  -  Altech: Is the performance specification we are discussing an update of the FTIR procedure or is it a general HCl monitoring performance specification?  Candace Sorrell - EPA: The performance specification will be generally applicable to all technologies measuring HCl including all current HCl stack monitors and future technologies that serve this purpose.  Altech uses Perma-Pure nafion dryer technology at the probe location followed by dry delivery to a multigas NDIR filter based monitor.  This system is not hot wet since the sample is dried at the sampling probe/sampling location.  Altech does not use an all heated system after the dryer.  The system is an extractive system that brings dry cool sample to the analyzer.  The stack gas hot (180 C and filtered prior to the dryer.  There is no issue with spiking in the hot wet zone in front of the filter.  That means this sampling design would not be excluded from using a hot wet HCl performance spike.  Jim Peeler: We've successfully performed hot, wet field spikes with HCl using the Altech system. 
   b) Laura Kenner - Emission Monitoring Inc.:  It seems that hot wet spiking should work for extractive systems.  In-situ monitoring systems may require a different approach.
   c) Keith Crab - Cemtech Environmental: Cemtech offers an in-situ tunable diode laser cross stack monitoring system. Cemtech uses a flow through cell in the optical measurement path that can be spiked with calibration gas to check instrument performance.  The cell functions in a similar manner to the hot wet spike since it is an additive spike to the measurement path.  The cell can be mounted on the stack or at the instrument. The flow through cell seems to be a valid test of the instrument performance that can be practically tested.  Higher concentration gases appropriate to the path length would be required for flow through cell spiking. This technology lends itself to multipoint QC or calibration checks.  At the end of a QC cycle the cell is flushed with nitrogen until the baseline before spiking is reached.  The cell remains in place during measurements and calibration checks.  There is no practical way of doing a cross stack zero measurement insitu.  Zero and span checks will be discussed in a separate stakeholder call.  The temperature and pressure of the gas cell and the stack gas must be included in the calculation of gas concentration.  Cemtech's flow through cell is not heated and equilibrates to the temperature at its location.  Lorentzian line shape is used in the calculation of concentration.
   d) Cemtech is also investigating an extractive dynamic probe TDL system that could be spiked with HCl through the probe similar to other extractive systems.
4) Concentration of spikes
   a) Jim Peeler: With emission standards approaching 3 ppm for cement kilns there are issues because plants that include inline raw mills will cycle between mill off high concentrations and mill on low concentrations.  What's the best concentration to spike and is it reasonable to assume that if spikes at low concentrations are recovered successfully then higher spike concentrations will also be recovered successfully?  Periodic testing supports this conclusion.
   b) Low concentration spikes are more a concern during calibration checks when 100 % gas (flooding the system with calibration gas) is being tested.  Dynamic, hot wet spiking is not performed at no more than 10% of the volume of gas sampled which means the spike gas can be a higher concentration.  The spike gas should not be the same gas as the calibration check standard gas.  This hot wet dynamic spike is similar to quarterly cylinder gas audits used in other programs.
   c) Luc Sevrette  -  Gasmet: Are we talking about a representative daily check, and if so a calibration check going through the system dry would take more than an hour to equilibrate.  He recommends a daily hot wet spike through the probe but not a daily calibration check with low level standards that must equilibrate through the system dry.
   d) Allen Rilling:  We typically spike at 75% of the measure emission level.  The new compliance limits present additional bias opportunities since the concentration is so low that the delivery systems may affect how accurately the HCl can be delivered to the probe.  Options to consider are daily zero and span or daily through the probe spike.  
   e) Also to be considered is the possibility of using a tracer gas to verify the dilution ratio of the spike gas delivery system in place of relying on mass flow controllers or rotameters for flow measurements.  This may be an option for multi-component monitors but may be a issue for narrow wavelength or single component systems that only measure HCl.  If the definition of tracer gas is left flexible, then manufacturers will have to solve this issue on an instrument specific basis.
      Table 1: Conference Call Attendees and Contact Information
Stakeholder Name
Affiliation
Interest
Email
Present
Rachael Agnew
EPA/OAQPS/SPPD/MPG
EPA
agnew.rachael@epa.gov

Joe Aldina 
Covanta Energy
Facility
jaldina@covantaenergy.com

Collin Boswell
EPA/OAQPS/SPPD/MPG
EPA
boswell.collin @epa.gov

Rich Brown
Altech 
Equip Vendor

Y
Gary Catchatory
Tyco Environmental Systems
Equip Vendor


Kieth Crab
Cemtech Environmental
Equip Vendor

Y
Edul Chikhliwala
EcoChem Analytics
Equip Vendor
chikli@ecochem.biz

Michael Corvese
Thermofisher
Equip Vendor
michael.corvese@thermofisher.com

Bob Davis 
AirGas
Gas Vendor
bob.davis@airgas.com

Jason Dewees
EPA/OAQPS/AQAD/MTG
EPA
dewees.jason@ epa.gov

Frank Duckett
Thermofisher
Equip Vendor
frank.duckett@thermofisher.com

Andy Edwards
Ash Grove Cement
Facility


Gerri Garwood
EPA/OAQPS/SPPD/MPG
EPA
garwood.gerri@epa.gov

Fred Grunewald
MKS Instruments Inc.
Equip Vendor
fred.grunewald@mksinst.com

Desirea Haggard
Cal Portland
Facility
dhaggard@calportland.com
Y
Jeff Harrington
Tyco Environmental Systems
Equip Vendor


Malik Hatar
Altech
Equip Vendor


Mike Hayes
Linde Gas
Gas Vendor


Andre Hegre
Gas Technologies, Finland
Equip Vendor


Stephen Johnson
Thermofishter Inc.
Equip Vendor
Stephen.Johnson@thermosisher.com

Dan Kietzer
Sick-Maihak
Equip Vendor
dan.kietzer@sickmaihak.com

Laura Kenner   
Emission Monitoring Inc.
Tester
lkinner@mindspring.com
Y
Dieter Kita 
Thermofisher
Equip Vendor
dieter.kita@thermofisher.com

Bill McMaster
PermaPure
Equip Add-on

Y
Barbara Marshick
MKS Instruments
Equip Vendor
Barbara_Marshik@mksinst.com

Ray Merrill
EPA/OAQPS/AQAD/MTG
EPA
merrill.raymond@epa.gov
Y
Barrett Parker
EPA/OAQPS/SPPD/MPG
EPA
parker.barrett@epa.gov

Ashmud Patel
Ethyl Smith Automation



Jim Peeler
Emission Monitoring Inc.
Tester
jimpeeler@mindspring.com
Y
Allan Rilling 
ABB
Equip Vendor
allan.j.rilling@ca.abb.com
Y
Ralph Roberson
RMB Consulting
Tester
Roberson@mb-consulting.com

Matthew Rutledge
Lafarge
Facility


Robin Segall
EPA/OAQPS/AQAD/MTG
EPA
segall.robin@@epa.gov

Luc Sevrette      
Gasmet
Equip Vendor
luc.sevrette@gasmet.com

Ty Smith
Cemtech Environmental
Equip Vendor


Candace Sorrell
EPA/OAQPS/AQAD/MTG
Regulatory Agency
sorrell.candace@epa.gov
Y
Martin Spartz 
Prism Analytical technology
Tester
m.spartz@pati-air.com

Bob Vantuyl 
Ash Grove Cement
Facility
bob.vantuyl@ashgrove

Henry Veger 
CEM Specialties
Equip Vendor


Dave  Vigowax
Marantee Tile
Facility


Bethany White
RMB Consulting
Tester

Y
James Willis
Titan America
Facility
jwillis@titanamerica.com

Hector Ybanez, 
Holcim Inc.
Facility
hybanez@gmail.com

Musa  Zada
Altech
Equip Vendor


Peter Zemek
MKS Instruments
Equip vendor
Peter_Zemek@mksinst.com
Y
         Y= Higlighted attendance indicates first time participants.
