                      EPA HCl Performance Specification 
                         Draft Conference Call Summary
                 Thursday August 18, 2011 from 1:30-2:00 EDT.
        Call in Conference Number 1-866-299-3188, code is 9195415511.  
                         Stakeholder Discussion Topic:
                         Calibration  -  Requirements?
                                       
1) Welcome  -  Introduction 
   a) Our timing remains to start the first draft by the middle of August and have the first full draft for Stakeholder review by the end of the calendar year (2011).  EPA has started an internal draft of the Performance Specification for HCl CEMS.  That draft requires internal review before it can be distributed to you stakeholders. We (EPA) ask that recipients of the notes from these calls review them and get back to us with corrections or comments.
   b) Today's agenda includes: 
      i) How should Calibration be performed
         (1) How should the range of Span be set and checked, multiple point calibration?
         (2) How should instrument calibration checks be performed in the field?
         (3) Should Instrument Vendors check instrument calibration against certified Protocol  HCl Standards?
            
2) How should the range of Span be set and checked, multiple point calibration?
Allen Rilling, ABB: Asked for clarification on the meaning of the number of points to check linearity.  Is this how many calibration points need to be provided to check linearity?
Ray Merrill, EPA: How many different concentrations do you use to perform or demonstrate the initial manufacturer's calibration?

Allen Rilling:  ABB calibrates IR units using one point calibrations for each component being measured. This approach is equivalent to a zero and span point calibration.  ABB also does a multiple concentration of water to compensate for water interference in addition to the instrument calibration.  After initial certification, calibration has been demonstrated in Europe to be valid for 6 months to a year.   Initial calibration and initial in field calibration checks are performed at the instrument and not through the probe.  In field, recertification or post maintenance calibration checks are performed in the field through the probe to verify performance.

Ray Merrill: Asked for comments on the linearity of FTIR over the range of 1 to 150 ppm.

Laura Kenner, Emissions Monitoring Inc.: HCl is not a linear absorber in the IR, so various manufacturers' algorithms deal with this issue in different ways.

Allen Rilling: there are many factors that go into the linearity of the output of an FTIR unit.  Partly the linearity depends on the resolution of the FTIR.  Reproducible line shapes allow manufacturers to predict what the nonlinear function of their instrument will be over a range of HCl concentrations.  ABB units are demonstrated consistent from unit to unit, so the calibration algorithm for HCl can be transferred from unit to unit.  ABB uses a single spectrum for calibration, and a known line shape and nonlinear function that are transferable from unit to unit.

Ray Merrill: The challenge for EPA is to have a means to ensure that whatever is required by a manufacturer to generate linear HCl results over a range of concentrations can be checked in the field as part of the performance of the instrument.  Allen offered to send a reference to the European standard wording requiring demonstration of linear response and certification that the response function is identical for all of a manufactures' similar technology units.

Ralph Roberson, RMB Consulting: Asked about the need for a low concentration component for the performance specification.  Do we need to demonstrate calibration linearity between 0  -  1 ppm and another demonstration for 0 to 150 ppm?  It doesn't make sense to require three calibration points between 0  -  1 pm but it does make sense to demonstrate the calibration at 1 ppm and/or 150 ppm as required by the compliance limit.

Rich Brown, Altech: The Altech HCl NDIR using gas filter correlation meet the Pennsylvania performance specification and daily zero, span checks and linearity checks.  Factory calibration is done using higher concentrations certified gas cylinders that are diluted to the calibration range for the instrument.  Ten calibration points are used between zero and 200 ppm span to factory calibrate the NDIR units.  Altech demonstrated RATA results for their NDIR unit at or below 10 ppm that meet the Pennsylvania PS for HCl CEMS. Calibration and calibration checks are done through the probe.  Altech NDIR units include a permeation dryer at the sample probe to remove the moisture interference. 

Frank Duckett, Thermofisher: Thermofisher has the same experience as Altech with their NDIR-Gas Filter Coorelation HCl CEMS.

Rich Brown: What's the range of concentration that will be covered in the new HCl Performance specification?  

Ray Merrill: EPA's preference is to write an HCl CEM performance specification that would be applicable to rules that require consistently low level HCl measurements and those rules that require a range (e.g., Cement MACT) into the 100s of ppm.
Jim Peeler: Repeated his comments from a previous conference call that preheater, precalciner cement kilns with in line raw mills have two different operating modes, mill on and mill off.  During mill on operation the HCl release is anticipated at fractional parts of ppm and during mill off the HCl concentration may reach 20 ppm.  For the Cement rule testers must be able to measure both low level, (<1 ppm) and mid level (~20 ppm) HCl concentration with the same instrument at any time during the testing.  Both low and mid concentration HCl must be measured accurately to meet rule requirements to report a 30 day rolling average.  The range of concentration for calibration and calibration checks may be industry specific based on the range of HCl emissions expected during normal operation.
Ray Merrill; It may be possible to write the Performance specification general enough to cover the differences across source categories by referring the tester to a span specified in each rule.

3) How should instrument calibration checks be performed in the field 21:38
   
   Ray Merrill: We're seriously considering periodic (daily) field calibration checks at the instrument and periodic system recovery checks through the probe and sampling system.
   
   Frank Dukett: In the Field, Thermfisher treats calibration checks on HCl monitors the same as any other CEM system (i.e., Thermo does calibration checks the same for HCl, SOx, and NOx). Calibration checks are performed through the probe with excess cal gas to flood the system displacing all other gas in the sample transport and instrument system allowing the system to pull 100% calibration gas.  
   
   Rich Brown: Altech checks calibration checks the same way as Thermofisher.  This approach can take additional time and the requisite amount of gas to equilibrate the perma pure dryer and sampling line.  
   
   Allen Rilling: Substituting a calibration transfer standard (CTS) for the certified HCl gas would free up the need to use so much HCl gas.  If the analyzer is reliably reading the correct concentration for the CTS then the instrument calibration has been validated (for multi-component systems). A single CTS standard may not provide spectral coverage for all the measured components. This is due to the CTS perhaps having absorbance in single spectral region while other components of interest are in different spectral range. One solution, if this proves to be an issue or concern, may be to have 2 CTS covering lower and upper frequency range.   For single component systems, calibration checks require spiking the target gas (HCl). Injecting dry HCl calibration gas in ABB systems would require 30 to 40 minutes for equilibration. 
   
   Ray Merrill: Our thoughts on calibration and recovery includes at the instrument 100% HCl standard spiking to demonstrate instrument performance and through the probe spiking while sampling stack gas for HCl recovery demonstration. 
   
   Jim Peeler:  How often are these QC checks performed?  You (Ray) indicated daily instrument calibration checks, but you haven't indicated how often through the probe spikes are required. It's hard to comment on these two QC checks until the testers know how often the checks are done. 
   
   Ray Merrill: EPA is taking input on how often through the probe spiking should occur.  To determine how often through the probe spike recovery QC occurs, we need information on how quickly an interfering filter cake builds up in the sample conditioning system.  We have considered more frequent through the probe QC checks initially until the system stability is demonstrated.  
   
   Rich Brown: Asked about the possibility of leaving both option in the performance specification. The options include flooding the probe with a calibration standard on a daily basis and calibrations at a less frequent interval at the instrument. Altech experiences a longer equilibration time (than SOx or NOx calibration)  is required for HCl through the probe checks at incineration units using 100% dry calibration gas, but not an unreasonable extra time.  Typical calibration checks spiked through the probe are done at or above 25 ppm.  Altech does linearity checks down to high single digit ppm concentration on a quarterly basis. 
   
   Mike Hayes, Linde: Certified gas cylinders below 10 ppm are accurate to 10% at best for periods of 6 month.  When NIST standards become available the gas supplier certification should improve to 5%.
   
   Frank Dukett: Another practice in the field involves accurately diluting a higher concentration certified standard.  In principle this approach could be more accurate and the standards may be stable longer.  Calibration standards could be diluted with ambient air containing ambient humidity to reduce the equilibration time of the monitoring system.  This may not help systems that dry the sample before transport to the instrument. 
   
   Laura Kenner, Emission Testing Inc.: FTIR systems use 15 to 20 HCl reference spectra that cover a range of concentration that are used as the fundamental calibration of the analyzer.  The daily calibration check is a check of how well the FTIR is using the reference spectra to generate HCl concentration.  Does EPA anticipate adjusting the analyzer based on the calibration checks?  Since the fundamental FTIR calibration is factory set by the reference spectra, how will testers use the results of the calibration and system integrity checks?
   
   Jim Peeler:  Its clear that we can't tolerate an uncertainty that is large compared to the emission standard.  For example a 1 ppm uncertainty will not work if the regulatory standard is 1 ppm.  It's critical to test the sample conditioning system to ensure the filter build up and reactive sites do not compromise the data validity.  The technologies and sample conditioning systems are sufficiently different for HCl CEMS for a one size fits all instrument check.  The initial instrument or calibration should be left to the instrument vendors.  Vendors must provide ways to determine if the instrument is operating properly with appropriate diagnostics and corrective actions. The common denominator for the HCl CEMS is the sampling system and testers need a performance based check of the entire system to verify to the regulated user that the data from the HCl CEMS is trustworthy. For stand alone compliance tests, the system challenge is required before and after each test run. 
   
   Lara Kenner: For continuous monitoring, the system challenge needs to be done every day until the tester demonstrates the total measurement system is stable for extended periods of time.  
   
   Jim Peeler: there are good parallels to HCl CEMS that involve dilution of a known gas standard and humidification of the spike gas.  The HOVACAL approach for HCl mirrors the approach taken for mercury.  There is an extensive document by EPA on the requirements and guidance to certify the HOVACAL for oxidized mercury.  This document also contains requirements for ongoing QA of the HOVACAL because it's not an NIST traceable standard.
   
   Unidentified Call Participant:  Recommended a performance specification with sections that call out the different procedures for NDIR-GFC, FTIR, TDL, CRDS, extractive and in-situ since some requirements won't make sense for all of the technologies we have now or in the future.
   
   Siv Jensen, SICK: Regarding the HOVACAL, in Europe ammonia, HCl, and HF can only be qualified by spiking wet gases.  SICK uses the HOVACAL to qualify instruments installed on waste incinerators these measurement targets.  They perform system checks every three to six months.  He is concerned that requiring less than 10 ppm system checks using dry calibration gas will not work for systems because that low level of calibration standard will never equilibrate if it is added dry.  SICK currently pursues getting a NIST traceability certification for their automated HOVACAL for mercury. SICK is demonstrating a new HOVACAL for through the probe mercury calibration at a facility in the U.S. 
   
   
4) Comments received offline
   a) Roy Vinckik, M&C Products Analysis Technology, Inc:
   As a sample conditioning manufacturer and supplier it is evident that a system check (one that flows gas from the 1st primary filter to the analyzer) is a must to validate that the HCl system is operating per specification and expectations.  In the case of calibration, it will be necessary to have the cal gas transport in a humidified state.  It is possible that on a Source that has a 4% by volume moisture content that their issues are far different than a similar site with 10% by volume moisture sample concentration.  It may require that the sample itself be humidified for transport in some applications to achieve reasonable response times and accuracy.
   
   As discussed the difference between the use of an FTIR, IR, or Laser based unit can possibly create its own special system issues and a calibration method that is both suitable and realistically reliable and user friendly.  
   
   The cost of a calibration system (Hovacal, dilution, or high breed of some sort) will be a part of the operating system and its cost will be a part of ownership and calculated over time.  The system cost of ownership will include the degree of maintenance to have it meet the operating criteria dictated by the EPA.   This will include the cost of cal gases or cal system upkeep per year.
   
   As Laura indicated, a daily cal would almost be a foregone conclusion.  I would have to think that a frequency that was less than 24 hours would be dictated by the instrument manufacturer or the system supplier to ensure maximum data capture, and they should not be punished as long as they are able to achieve the required number of data sets per hour.  This maybe dependent upon the technique that is selected by the facility and may be dictated by their concentration of HCl.   
   
   In (I think) most cases a big issue of the sample transport will be the umbilical issue.  Many facilities find themselves doing standard criteria pollutant monitoring and or Hg and they want to incorporate as much of the original systems as possible into the addition of the upcoming monitoring requirement.  When HCl, NH3, or Chlorine is the adder, to ensure minimum data loss of the other gases of interest, it will be in the best interest of all that they have a dedicated line for the transport of the HCl (in this case) sample and a dedicated line for the transport of the cal gas to ensure a reasonable response to sample and cal gas.  Obviously, it would be highly recommended that both lines be heated.
   
   As Jim and others stated, it is going to be difficult to have a standard that cannot be matched by a validation standard that can prove or disprove the accuracy of the system in operation.  And as Ralph indicated, it is a very costly process to prove or disprove the utilization of a standard and
   its variability,  dependent upon individual application sites.
   Comments Received After the Conference Call:
   
   Gary Cacciatore, Tyco Env Systems: 
   Based on the fact that the measurement of HCL is very dependent upon the amount of moisture I think the rule should be to make it mandatory to not only calibrate with a moisture background in the calibration gas but also to have to measure and pass a calibration on the H2O.  IR's technology, especially FTIR is very affected by the water content in the gas sample.
    
   I'm not sure I have seen the HCL reporting requirements but the measurement should be reported on a wet bases because of the moisture interference issues.
    
   Calibration:  I see that there will be 2 extreme ranges, a 0-5 ppm and a 0-50 ppm.  The analyzer will need a dual range and have to be calibrated with a zero gas and two span gases, one for each range.  
   Calibration needs to be performed once every 24 hours.  Calibration needs to be at the probe and must include every and all components that will touch the wet stack gas. 
   A daily local calibration that only passes thru the analyzer needs to be performed daily either immediately before or after the probe calibration.
   Any deviation of 5% drift between the local and probe calibration will fail the calibration.
   Typical part 60 regulations on pass fail criteria should apply.  One day 10% drift, 5 consecutive days over 5% drift are failures.  Drift shall be calculated on the range of the analyzer.
    
   Linearity:  There should be a 3 point linearity test performed for each of the ranges to confirm accuracy:  1.5, 2.5 and 4.5 for the 0-5ppm range and 15, 25 & 45ppm for the 0-50ppm range.  The 3 point linearity check needs to be performed every quarter.
   
   A 3 point probe spike check also needs to be performed every 30 days to coincide with the 30 day rolling average requirement.  This needs to be done because it is a fact that cement plants scrubbing media that builds up on the probe and filters will inevitably scrub HCL thus making the measurement biased. 
    
      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
rbrown@altechusa.com
Y
Gary Cacciatore
Tyco Environmental Systems
Goyen Valve Corporation
Equip Vendor
gcacciat@aol.com or gcacciatore@goyen.com

Kieth Crab
Cemtech Environmental
Equip Vendor
keith@cemtecks.com

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
Y
Andy Edwards
Ash Grove Cement
Facility
andy.edwards@ashgrove.com
Y
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
jsharrington@goyen.com

Malik Hatar
Altech
Equip Vendor


Mike Hayes
Linde Gas
Gas Vendor
mike.hayes@linde.com
Y
Andre Hegre
Gas Technologies, Finland
Equip Vendor


Pat Hennessy, 
CEMEX


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

Siv Jensen
SICK- Maihak
Equip Vendor

Y
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

Phil McMaster
PermaPure
Equip Add-on
pmcmaster@permapure.com

Barbara Marshick
MKS Instruments
Equip Vendor
Barbara_Marshik@mksinst.com

Ray Merrill
EPA/OAQPS/AQAD/MTG
EPA
merrill.raymond@epa.gov

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

Allan Rilling 
ABB
Equip Vendor
allan.j.rilling@ca.abb.com
Y
Ralph Roberson
RMB Consulting
EGU Consultant
Roberson@mb-consulting.com
Y
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
ty@cemteks.com

Candace Sorrell
EPA/OAQPS/AQAD/MTG
EPA
sorrell.candace@epa.gov

Martin Spartz 
Prism Analytical technology
Tester
m.spartz@pati-air.com

Bob Vantuyl 
Ash Grove Cement
Facility
bob.vantuyl@ashgrove
Y
Henry Vergeer
CEM Specialties
Equip Vendor
vergeer@cemsi.on.ca

Dave  Vigowax
Marantee Tile
Facility


Bethany White
RMB Consulting
EGU Consultant
white@rmb-consulting.com

James Willis
Titan America
Facility
jwillis@titanamerica.com
Y
Hector Ybanez, 
Holcim Inc.
Facility
hybanez@gmail.com

Musa  Zada
Altech
Equip Vendor


Peter Zemek
MKS Instruments
Equip vendor
Peter_Zemek@mksinst.com

         Y= Higlighted attendance indicates first time participants.
