consolidated comments of legal & safety employer … · have an open process of comprehensive...

Consolidated Comments of

Legal & Safety Employer Research, Inc. (LASER), Sierra Club Michigan Chapter

& Concerned Citizens of Shiawassee County

Regarding Draft Air Quality Permit forE85, Inc. – Corunna, MI Ethanol Production Plant

Presented to

Michigan Department of Environmental QualityAir Quality Division

&U.S. Environmental Protection Agency, Region V,

Air & Radiation Division, Permits & Grants Section & Air Enforcement Section

Legal & Safety Employer Research, Inc. (LASER)Sierra Club Michigan Chapter (SCMC)

Concerned Citizens of Shiawassee County (CCSC)Tracy J. Andrews, Counsel

Olson, Bzdok & Howard, P.C420 East Front St., Traverse City, MI 49686

(231)946-0044; http://www.envlaw.com

May 16, 2007

Prepared byAlexander J. Sagady, Environmental Consultant

657 Spartan Avenue, East Lansing, MI 48823(517)332-6971; http://www.sagady.com [email protected]

This document available on the web at:http://www.sagady.com/workproduct/LASERCommentE85IncCorunnaMI.pdf

Table of Contents

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

2 The Proposed Issuance of the Air Discharge Permit for E85 - Corunna is a MajorState Decision Which Requires a Process of Comprehensive EnvironmentalReview of Pollution, Impairment and Destruction of Michigan’s NaturalResources, Consideration of Reasonable Alternatives to the Proposed Action and aMDEQ Final Finding and Determination as to these Matters in Order to Complywith the Michigan Constitution and the Natural Resources and EnvironmentalProtection Act . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

2.1 The Michigan Department of Environmental Quality (MDEQ) Does NotHave an Open Process of Comprehensive Environmental Review in Placeas a Systematic Decisionmaking Process and MDEQ Has Not Provided anAd Hoc Equivalent of Such a Process for the E85 - Corunna FacilityDecision that Meets the Statutory Requirements for Environmental Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

2.2 The Decision to Issue an Air Permit to E85 - Corunna is a Major ActionHaving Significant Consequences for Management of Natural Resourcesand Protection of the Environment; Such Consequences Have Not BeenSubjected to an Open Public Process of Comprehensive Review and MDEQDoes Not Plan to Make Final Determinations Involving Such Consequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

3 Applicant is Subject to a 100 Ton Major Stationary Source New Source ReviewThreshold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

4 The Draft Permit Must Not Issue Since Applicant’s Facility is a Major StationarySource of One or More New Source Review Regulated Pollutants and a HazardousAir Pollutant Under the Clean Air Act . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

5 Comments Applicable to Multiple Emission Units . . . . . . . . . . . . . . . . . . . . . . . . 9

5.1 Various Emission Unit Sections of the Draft Permit Do Not ContainSufficient Federally Enforceable Physical Production Rate and/orThroughput Limitations on the Potential to Emit to Ensure the Facility DoesNot Exceed Major Stationary Source Thresholds and Individual EmissionUnit Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

5.2 Draft Permit Provisions Addressing Hazardous Air Pollutants (HAP) . . . 11

5.2.1 Applicant Has Not Quantified Hazardous Air Pollutants FromFugitive Emissions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

5.2.2 The Draft Permit Should be Amended to Include SpecificAcetaldehyde, Acrolein and Formaldehyde Limits . . . . . . . . . . . . 12

5.3 Neither the Draft Permit, Nor MDEQ’s Current Policies on ApprovingStack Testing Practices, Contain Any Assurances that the Total Mass Rateof Volatile Organic Compound Emissions is Measured for EmissionLimitations Compliance Purposes, for Evaluation of the VOC BACTStringency of VOC Emission Limitations and For Purposes of Determiningthe Major Source Status of the Proposed Facility . . . . . . . . . . . . . . . . . . . 12

5.3.1 EPA Policy is Clear that the Clean Air Act New Source ReviewPrograms Must Ensure Accountability for the Total Mass Rate ofVolatile Organic Compound Emissions . . . . . . . . . . . . . . . . . . . . . 12

5.3.2 As Proposed, the Draft Permit Does Not Specify Any Test Methodsor Compliance with EPA’s Midwest Protocol for Total VOC MassRate Emissions Determination from Ethanol Plants . . . . . . . . . . . 14

5.3.3 MDEQ’s Apparent Practice with Ethanol Facilities is to Unlawfullyand Impermissibly Use Unadjusted EPA Test Method 25 and 25ADeterminations to Stand for VOC Emission Limitation Compliance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

5.3.4 MDEQ-AQD Does Not Have Rules or Firm Policies About“Department Requirements” on Ethanol Plant Testing . . . . . . . . . 16

5.3.5 Compliance Testing During Maximum Emissions Potential . . . . . 17

5.3.6 Chemical Speciation Listing for EPA Method 18 DeterminationsShould be Extended . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

5.4 The Draft Permit Does Not Have a Realistic, Practical Enforcement Methodfor FG-FACILITY Emission Limitations . . . . . . . . . . . . . . . . . . . . . . . . . 17

5.5 Compliance Monitoring of Fabric Filter Controlled Emission Units . . . . 18

5.6 Testing Requirements for Condensible Particulate Emissions . . . . . . . . . 19

5.7 Applicant’s Emission Characterizations for All Fabric Filter ControlledEmission Units Do Not Incorporate Actual Cubic Feet Per Minute ProcessGas Flows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

5.8 In Violation of Michigan New Source VOC BACT and Air Use RuleRequirements, the Applicant Has Not Provided an Appropriate BestAvailable Control Technology Determination for All Volatile OrganicCompound Emission Sources and Has Not Properly Quantified Emissionsat All Such Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

5.9 The Applicant Has Failed to Disclose or Evaluate its Emissions, PollutionControls and Ambient Health Impacts from of an Important HazardousMaterial at the Planned Corunna Facility . . . . . . . . . . . . . . . . . . . . . . . . . 23

5.10 The Applicant Has Not Properly Characterized Acrolein and UrethaneEmissions for the Subject Facility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

5.11 Other Comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

6 Discussion of Permit Regulatory Sections and Emission Calculations by IndividualEmission Unit and Process Groupings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

6.1 Natural Gas-Fired Boilers #1 and #2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

6.1.1 Applicant Has Erroneously Characterized the Stack Gas ExhaustVolume Based on the Wrong Natural Gas F-Factor . . . . . . . . . . . . 25

6.1.2 Applicant Will Not Be Able to Comply with the NOX EmissionLimitations for Natural Gas Fired Boilers #1 and #2 with the PlannedEquipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

6.1.3 The Emission Limitations of the Draft Permit Should Not Place aMass per Unit Time Emission Limit Bubble Over the Two Boilers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

6.1.4 The Draft Permit Should Require Testing for both Filterable andCondensible Particulate Matter . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

6.1.5 The Applicant Hasn’t Used AP-42 Emission Factors forCharacterizing Volatile Organic Compound Emissions from theBoilers and MDEQ has Not Set Volatile Organic CompoundEmission Limitations for the Natural Gas Fired Boiler Units . . . . 27

6.1.6 Boiler Monitoring Conditions Require Revision . . . . . . . . . . . . . . 27

6.2 Site Roads Fugitive Emission Unit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

6.2.1 E85, Inc. Submitted False Information in its Air Permit ApplicationInvolving Fugitive Road Dust Emission Calculations . . . . . . . . . . 28

6.2.2 MDEQ’s Fact Sheet Made Erroneous Assumptions AboutApplicant’s Assumed Silt Loading Factors and Fugitive RoadEmission Calculations for PM and PM-10 . . . . . . . . . . . . . . . . . . . 31

6.2.3 The Draft Permit Doesn’t Contain Any Requirements to VerifyCompliance with Silt Loading Rates Assumed in the EmissionCalculation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

6.2.4 Applicant Fugitive Road Emission Calculation Failed to Account forVMT from Truck Delivery of Process-Related Chemicals . . . . . . 32

6.2.5 Condition 5.3 Should Require Improved Recordkeeping andContemporaneous Collection of Actual Truck Traffic Data . . . . . 33

6.2.6 The Draft Permit Should be Amended to Require that the ApplicantMaintain a Working Street Sweeper On-site at All Times . . . . . . . 33

6.2.7 Section 5 of the Draft Permit for EU-TRKTRAFFIC Must beAmended to Incorporate an Annual and an Hourly Numerical PMand PM-10 Emission Limitations in Addition to the PhysicalLimitation on the Potential to Emit . . . . . . . . . . . . . . . . . . . . . . . . 33

6.2.8 MDEQ-AQD Must Not Remove Responsibility from the Applicantby Merely Assuming that Rail Shipments Will Always EnableApplicant to Avoid Paved Road Fugitive Emissions . . . . . . . . . . . 34

6.2.9 The Applicant’s Intentions to Actually Comply with their Permit toInstall Must be Questioned in Light of their Public Statements . . . 34

6.2.10 The Draft Permit Fails to Require that Applicant Maintain AllRoadways and Parking Lots in a Paved Condition . . . . . . . . . . . . . 35

6.2.11 Monitoring and Reporting Requirements Must EmbraceEnforcement of Road Sweeping Requirements . . . . . . . . . . . . . . . 36

6.2.12 Applicant’s Fugitive Road Emission Characterization ContainsAssumptions Which Cause Significant Underestimation of FugitiveRoad Emissions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

6.2.13 Commenter’s Fugitive Road Particulate Emission CalculationFactors for Applicant’s E85, Inc. Corunna, MI Facility . . . . . . . . . 37

6.2.14 Revised Vehicle Miles Traveled Information and SubsequentRevised Particulate Emission Calculations for E85 - Corunna . . . 37

6.3 Loading Rack Emissions from Truck and Railcar Tanker Loadout . . . . . 39

6.3.1 The Draft Permit Must be Amended to Incorporate EnforceableVOC, NOX and CO Emission Limitations . . . . . . . . . . . . . . . . . . 39

6.3.2 The Draft Permit Should be Amended to Require Compliance withAll Features of Flare Control and Monitoring RequirementsSpecified in 40 C.F.R. §60.18, Including Testing and MonitoringRequirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

6.3.3 No Practically Enforceable Provisions in the Draft Permit ProvideFederally Enforceable Requirements Ensuring That All Truck andRailcar Tankers to be Unloaded at the Facility Meet Appropriate On-Board Vapor System Collection Efficiency and Vapor “Tightness”Performance Standards and That the Facility Ensure Compliancewith Such Requirements Through Monitoring, Recordkeeping andReporting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

6.3.4 If MDEQ Fails to Provide a Federally Enforceable Requirements for the Owner/Operator to Verify the Leak Testing Status of TruckTankers Loaded at the Facility, the Loading Rack Potential to EmitVOC and HAP Calculation Must be Rejected and Revised Upward41

6.4 Rail and Truck Grain Receiving . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

6.4.1 The Draft Permit Contains No Federally Enforceable Requirementsfor Emission Limitations and Fugitive Emissions Control for Grain Receiving . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

6.4.2 The Draft Permit Does Not Maintain Sufficient ConditionsPhysically Limiting the Fugitive PM and PM-10 Potential to Emit 42

6.4.3 Applicant’s Truck Grain Receiving Emission Characterization is Erroneous . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

6.5 Cooling Tower . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

6.5.1 The Draft Permit Fails to Provide Federally Enforceable PM/PM-10Emission Limitations, Monitoring Requirements and PhysicalLimitations on the Potential to Emit . . . . . . . . . . . . . . . . . . . . . . . . 43

6.6 Dryers #1, #2, Product Cooling and Thermal Oxidizer . . . . . . . . . . . . . . . 44

6.6.1 The Emission Limitations of the Draft Permit Should Not Place aMass per Unit Time Emission Limit Bubble Over the Two Dryers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

6.6.2 The Draft Permit Should Require Testing for both Filterable andCondensible Particulate Matter . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

6.6.3 Startup Emissions for the Dryers . . . . . . . . . . . . . . . . . . . . . . . . . . 45

6.6.4 Limits on Dryer Natural Gas Usage are Not Sufficient to Limit thePotential to Emit of Dryer Process Unit Emissions of CO, PM-10and VOC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

6.6.5 The Application and the Draft Permit Do Not Include SufficientParameter Monitoring and Other Testing Provisions to AssureCompliance with Volatile Organic Compound and Hazardous AirPollutant Emission Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

6.6.6 Continuous Monitoring Provisions of the Draft Permit ShouldRequire Stack Gas Flow Monitoring at the Stack Location WhereContinuous Monitors are Located . . . . . . . . . . . . . . . . . . . . . . . . . 46

6.6.7 The Draft Permit Should Be Amended to Clearly Allow NOX andCO Emission Limitation Violation Enforcement at the Dryer #1 and#2 Emission Units Through Use of Continuous NOX and COContinuous Emission Monitoring Results . . . . . . . . . . . . . . . . . . . 47

6.6.8 Applicant’s Acetaldehyde Emission Characterization for the DryerUnits Lacks a Supporting Basis and Is Not Backed by a VendorGuarantee or Demonstrated Compliance Stack Testing . . . . . . . . . 47

6.7 Fermentation Scrubber Emission Unit . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

6.7.1 Applicant Has Failed to Properly Support Their EmissionCharacterization of the Fermentation Scrubber Emission Unit . . . 48

6.7.2 Applicant’s Acetaldehyde Emission Characterization for theScrubber Controlled Units is Subject to Challenge as UnrealisticallyLow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

6.7.3 Compliance Testing and Parameter Monitoring in the Draft Permitfor Scrubber-Controlled Emission Units is Not Sufficient to EnsureCompliance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

6.7.4 The Draft Permit Should Be Amended to Require Continuous VOCEmission Monitoring for the Fermentation Scrubber Emission Unit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

6.8 The Applicant Failed to Quantify Hazardous Air Pollutants from PlantFugitive Components for FG-NSPSVV . . . . . . . . . . . . . . . . . . . . . . . . . . 50

Comments of LASER, SIERRA CLUB MICHIGAN CHAPTER& CCSC on Draft Air Permit for E85, Inc. - Corunna, MI Page 1

1 Introduction

Legal & Safety Employer Research, Inc. (LASER), Sierra Club Michigan Chapter andConcerned Citizens of Shiawassee County (CCSC) have produced these comments as anindependent review of the air permit application and draft permit for the proposed E85,Inc. facility at Corunna, MI.

We submit these comments for filing with the Michigan Department of EnvironmentalQuality - Air Quality Division and the U.S. Environmental Protection Agency, Region 5,Air & Radiation Division.

2 The Proposed Issuance of the Air Discharge Permit for E85 - Corunna is aMajor State Decision Which Requires a Process of ComprehensiveEnvironmental Review of Pollution, Impairment and Destruction ofMichigan’s Natural Resources, Consideration of Reasonable Alternatives tothe Proposed Action and a MDEQ Final Finding and Determination as tothese Matters in Order to Comply with the Michigan Constitution and theNatural Resources and Environmental Protection Act

Michigan Law requires that major state decisions on matters such as the granting of an airdischarge permit to install for the proposed E85 - Corunna facility properly consider theconsequences of such decisions for pollution, impairment and destruction of Michigan’snatural resources, the public’s trust placed on these natural resources and theconsideration of feasible and prudent alternatives that would limit such pollution,impairment and destruction.

The Michigan Constitution of 1963 provides:

“The conservation and development of the natural resources of the state are herebydeclared to be of paramount public concern in the interest of the health, safety andgeneral welfare of the people. The legislature shall provide for the protection of theair, water and other natural resources of the state from pollution, impairment anddestruction.” (Michigan Constitution, 1963; Article IV, Section 52)

The Michigan Natural Resources and Environmental Protection Act (NREPA) provides:

“In administrative, licensing, or other proceedings, and in any judicial review ofsuch a proceeding, the alleged pollution, impairment, or destruction of the air,water, or other natural resources, or the public trust in these resources, shall bedetermined, and conduct shall not be authorized or approved that has or is likely tohave such an effect if there is a feasible and prudent alternative consistent with thereasonable requirements of the public health, safety, and welfare.” (MCL324.1705(2)) (emphasis added)


NREPA provides that any citizen may commence an action to gain declaratory andequitable relief to protect the natural resources of this state and the public trust therein:

“(1) The attorney general or any person may maintain an action in the circuit courthaving jurisdiction where the alleged violation occurred or is likely to occur fordeclaratory and equitable relief against any person for the protection of the air,water, and other natural resources and the public trust in these resources frompollution, impairment, or destruction.

(2) In granting relief provided by subsection (1), if there is a standard for pollutionor for an antipollution device or procedure, fixed by rule or otherwise, by the stateor an instrumentality, agency, or political subdivision of the state, the court may:

(a) Determine the validity, applicability, and reasonableness of the standard.

(b) If a court finds a standard to be deficient, direct the adoption of astandard approved and specified by the court.” (MCL 324.1701)

Finally, decisions affecting Michigan’s natural resources and the public trust therein thattake place in a permit process and which represent agency decisions must be subject toprior public notice and public comment under multiple NREPA provisions and under theMichigan Administrative Procedures Act.

2.1 The Michigan Department of Environmental Quality (MDEQ) Does Not Havean Open Process of Comprehensive Environmental Review in Place as aSystematic Decisionmaking Process and MDEQ Has Not Provided an Ad HocEquivalent of Such a Process for the E85 - Corunna Facility Decision thatMeets the Statutory Requirements for Environmental Review

Michigan formerly had a process for environmental and natural resources impact reviewthat featured mandatory consideration of major state actions before the MichiganEnvironmental Review Board (MERB). That process was abolished by Executive Orderin the last days of the Blanchard Administration with the intent on replacing it with analternative and substitute process. However, a replacement process was neverimplemented and major state decisions since the abolition of MERB have not receivedcomprehensive environmental review commensurate with the MEPA provisions ofNREPA which were previously in effect as a stand alone statute prior to the 1994codification of natural resource and environmental protection laws by the Legislature.

As such, there is no regular and organized system of comprehensive environmental andnatural resources impact review that reaches the non-discretionary requirements of MCL324.1705(2) for consideration of pollution, impairment and destruction of Michigan


natural resources and consideration of alternatives. This chronic non-compliance withMichigan statutory requirements must end.

It is conceivable that an ad hoc system of review could be afforded to decisionmaking onsuch matters as the E85 - Corunna facility. However, there has been no such ad hocprocess set up to reach and satisfy the non-discretionary requirements of MCL324.1705(2) for purposes of this subject facility.

In fact, the present air permit process has discouraged comprehensive review withstatements in “Public Participation Documents” such as:

“The Air Quality Division does not have the authority to regulate noise, localzoning, property values, off-site truck traffic, or lighting.”

While it may be true that the AQD does not have such authorities for direct regulation ofthe matters listed, MDEQ nevertheless has authority to consider and evaluatecomprehensive environmental and natural resources impacts from the proposed facilityunder MCL 324.1705(2). In partial recognition of this authority, Michigan air use rulesprovide that air permit to install applications provide:

“Information, in a form prescribed by the department, that is necessary for thepreparation of an environmental impact statement if, in the judgment of thedepartment, the equipment for which a permit is sought may have a significanteffect on the environment.” (MAC R 336.1203(1)(g))

However, in the present case, MDEQ-AQD has committed error by failing to require thesubmittal of any such information as provided by its rules and has further failed toconduct a comprehensive environmental impact statement and environmental review onthe proposed E85 - Corunna facility.

2.2 The Decision to Issue an Air Permit to E85 - Corunna is a Major ActionHaving Significant Consequences for Management of Natural Resources andProtection of the Environment; Such Consequences Have Not BeenSubjected to an Open Public Process of Comprehensive Review and MDEQDoes Not Plan to Make Final Determinations Involving Such Consequences

Although the present proceeding involves consideration of air pollution impacts of thefacility, the review process so far has failed to properly characterize airborne emissionsfrom the plant and to provide for appropriate permitting and technology-based controls asoutlined in sections of this comment below.

Moreover, the existing air pollution permitting process has failed to considerenvironmental impacts from the large amounts of greenhouse gases – on the order of


500-600 thousand tons per year of carbon dioxide from fermentation processes at theproposed E85 - Corunna facility.

The proposed plant is expected to mine groundwater using either nearby production wellsand/or City of Corunna wells at a rate exceeding 620 million gallons per year. To thebest of our knowledge, MDEQ has not conducted a detailed review of the consequencesof such groundwater mining on aquatic flora and fauna, stream flows, wetlands, publicand private wells and other land/water features in the area..

This groundwater mining of 620 million gallons per year is actually an expropriation atzero cost of a public trust resources by E85 - Corunna and further represents a singularuse of a public resource with no review and determination of the consequences byMDEQ. This “free” groundwater is being provided to a facility that is already going toreceive public financial subsidies. If alternate means of energy dissipation wereprovided, or a mix of alternate means were considered, the massive demands forgroundwater might be significantly reduced. MDEQ has not determined the public trustinterests in area groundwater and has made no public trust determinations applicable tothis situation.

Finally, an estimated half of this Great Lakes Watershed groundwater will be lost toevaporation, representing a large out of Great-Lakes basin transfer of this Great Lakesresource. This latter fact poses significant questions of Great Lakes water managementpublic policy which have not been addressed or answered by MDEQ.

Other social and community impacts are posed by the subject facility, including thepotential effects of the facility operations on grain markets, animal agriculture utilizationin Shiawassee County, water supply for adjacent property owners, cooling tower icing onadjacent property owners and on area roads.

The comprehensive environmental, natural resources and community effects of theproposed facility have not been adequately considered (or considered at all) by MDEQ onthe proposed major state decision to grant the E85 - Corunna air permit.

In summary, MDEQ has failed to carry out its mandatory and non-discretionary duties forenvironmental and natural resources impact review, consideration of alternatives andprotection of the public trust in Michigan’s natural resources in making the decision toissue the E85 - Corunna air permit to install. This failure gives rise to a cause of actionby any person to have adjudications of these issues by the applicable circuit court ofjurisdiction.


1 42 U.S.C. §7479(1)

2 40 C.F.R. §52.21(b)(1)(i)(a)

3 Corn is a bio-mass fuel used for heat input through combustion. Corn biomass fuelcontains 8000 to 8500 BTUs per pound on a dry basis. See:

3 Applicant is Subject to a 100 Ton Major Stationary Source New SourceReview Threshold

The Applicant has proposed a process which has traditionally been regulated as a“chemical process unit” for purposes of major stationary source new source review airpermitting. EPA has published a final rule attempting to revise EPA new source reviewprocedure to remove ethanol production facilities from being considered as “chemicalprocess units.”

Notwithstanding the content of the recent EPA Federal Register notice of final rule, sucha decision does not affect Michigan air permitting unless and until such a change has beenvetted in the Michigan Administrative Procedure Act rulemaking process after sufficientpublic notice and opportunity for public comment. In addition, such a change constitutesan change in the Michigan State Implementation Plan under the Clean Air Act whichmust be properly noticed for adoption as a revised federal requirement for Michigan-specific air permitting. In addition, such a change can be considered as impermissibleSIP backsliding which may be prohibited under Clean Air Act Sections 110(i) and 193 asany such emissions may affect a non-attainment area or air quality maintenance plans foran existing attainment area.

Even if EPA’s final rule disallowing consideration of ethanol plants as chemical processunits were construed to have immediate effect, such an action would not have retroactiveapplication to the subject facility which both applied for a permit and had publication of adraft permit before EPA’s final rule was published in the Federal Register. Applicationof the federal rule relaxing the major stationary source thresholds for ethanol plants isbarred in Michigan air permit proceedings as an arbitrary action that doesn’t conform toMDEQ-AQD’s duties for the public trust protection in the air resources of this state toprevent pollution, impairment and destruction of such public trust air resources andMichigan constitutional provisions holding as paramount the protection of public healthand natural resources of this state.

EPA’s final rule does not alter the triggering status of fossil fuel steam generation units(or combinations thereof) that are 250 MMbtu/hour heat input or greater from mandatorydesignation as major stationary sources. The 100 ton major stationary source status ofsuch steam generation units is provided by statute1 and by EPA rule.2 Similarly, ethanolplants are “fuel conversion plants” under the same statutory provisions and regulationsthat convert biomass-fuels3 to alcohol fuels and trigger 100 ton major stationary source


http://energy.cas.psu.edu/energycontent.html

status in the same manner as the 250 MMbtu/hour heat input fossil fueled boiler unitsapplicability criterion.

Applicant’s facility will still have a major stationary source new source review thresholdof 100 tons even after publication of EPA’s final notice because the Applicant plansoperations of two boilers fired by natural gas as a fossil fuel with a combined heat inputcapacity of 300 MMbtu/hour and Applicant’s facility will be a biomass energy fuelconversion facility.

Setting aside the question of whether the facility is or is not a “chemical process unit”because of the new final federal rule, ethanol plants of the size proposed by the Applicanthave necessarily been defined as having the primary pollutant-generating activity at thesite as being the generation of steam from the combustion of fossil fuels. The associatedupstream and downstream process activities with operations of the fossil fuels steamgeneration systems on site cannot practically be isolated or distinguished as separatestationary sources from the steam generation function on site as these terms are definedby the Clean Air Act. There is no functional or economic purpose for an ethanolproduction site fossil fuel-fired steam generation unit without its need to generate steamfor beneficial use elsewhere on site. All of the other emission units on site are integral tothe pollutant generating aspects of this particular implementation of a fossil fuel firedsteam generating unit and all emission units associated with it under the Clean Air Act’srubric of a major stationary emission source.

Under the prevention of significant deterioration regulation and apart from “chemicalprocess unit” applicability, the following primary pollutant-emitting source activitiestrigger the major stationary source threshold of 100 tons, as well as source-wideaccountability for fugitive emissions in contributing to such a total:

“....fossil fuel boilers (or combinations thereof) totaling more than 250 millionBritish thermal units per hour heat input...”

“....fuel conversion plants....” (40 C.F.R. §52.21(b)(1)(i)(a); 40 C.F.R.§52.21(b)(1)(iii)(u) and 40 C.F.R. §52.21(b)(1)(iii)(q))

Applicant’s proposed facility will be a stationary source as defined in the regulations:

“Stationary source means any building, structure, facility, or installation whichemits or may emit a regulated NSR pollutant.” 40 C.F.R. §52.21(b)(5) (emphasisadded)


(6) Building, structure, facility, or installation means all of the pollutant emittingactivities which belong to the same industrial grouping, are located on one ormore contiguous or adjacent properties, and are under the control of the sameperson (or persons under common control) except the activities of any vessel.Pollutant-emitting activities shall be considered as part of the same industrialgrouping if they belong to the same ‘‘Major Group’’ (i.e., which have the samefirst two digit code) as described in the Standard Industrial Classification Manual,1972, as amended by the 1977 Supplement (U. S. Government Printing Officestock numbers 4101–0066 and 003–005–00176–0, respectively)” 40 C.F.R.§52.21(b)(5)

Applicant’s fossil fueled boilers are necessarily in the same “industrial grouping” as theother portions of the single stationary source at the site because the production of ethanolis inextricably intertwined with the need to produce steam at the site, because all processequipment is under common control and all such processes are adjacent to each other onthe same site.

The fossil fueled steam generation/boiler process as articulated by the Applicant in theirsubmittal consists of two 150 MMbtu/hour natural gas fired boilers whose physicalnatural gas throughputs are limited by enforceable permit condition.

In addition, Applicant’s entire facility and associated process constitutes a “...fuelconversion plant...” transforming biomass corn fuels into ethanol fuels.

The provisions of the Act and applicable regulations cited above also similarly ensure thatfugitive emissions from such a major source will be counted towards the 100 tonthreshold.

4 The Draft Permit Must Not Issue Since Applicant’s Facility is a MajorStationary Source of One or More New Source Review Regulated Pollutantsand a Hazardous Air Pollutant Under the Clean Air Act

The comments in subsequent sections identify a number of serious problems showing Applicant’s and MDEQ-AQD’s underestimation of the potential to emit from specificemission units at the proposed facility and from the entire emission source as a whole. Given that the potential to emit for one of more New Source Review Regulated pollutantsor more criteria pollutants exceeds 100 tons per year from the entire source, Applicant’spermit as proposed may not be approved since the facility would not have undergone therequired Prevention of Significant Deterioration review, including a determination of BestAvailable Control Technology for all criteria pollutants emitted in significant amountsand an air quality impact analysis. The latter analysis must necessarily include a reviewof compliance with PSD ambient increments and a demonstration that attainment andmaintenance of the National Ambient Air Quality Standards will not be jeopardized. The


Applicant has not submitted a Best Available Control Technology determination nor anair quality impact assessment meeting PSD requirements; the required PSD-relateddeterminations have not been made by MDEQ-AQD for the subject facility.

The margins between the Applicant’s claimed potential to emit emissions and the 100 tonthreshold for the subject facility are small. In subsequent subsections of this comment,we identify a number of emission unit/process areas subject to the following problems:

The potential to emit is underestimated.

Emission units are not listed or characterized.

There are no emission limitations to enforce on individual emission units needed toensure that the whole source total potential to emit does not exceed majorstationary source thresholds.

There are no short term emission limitations and/or limitations on the potential toemit that can ensure that air quality modeling predictions will actually be valid.

There are either insufficient or no federally enforceable physical limitations on thepotential to emit.

Testing and monitoring methodologies are either not specified, are insufficient orare not capable of assuring compliance with prior characterizations of the potentialto emit.

There are no compliance assurance measures at some of the emission units.

The facility will be unable to comply with its permit as published in draft version.

When the margins from the amount of some of the clear underestimations can bequantified and summed, it is clear that the subject facility will have NSR regulatedpollutant emissions exceeding the 100 ton major stationary source threshold. Other unitshave emission projections but the terms of the draft permit do not provide sufficientmonitoring measures to assure compliance with the emission limitations on an continualbasis. Under these circumstances, the permit should not be issued because of failure toassure compliance in conforming to major stationary source permitting requirements.

Notwithstanding the major stationary threshold issue, the individual process unit emissioncharacterizations constitute error in cases where we identify underestimation of expectedemissions.


4 http://www.epa.gov/ttn/atw/pte/june13_89.pdf

Finally, the draft permit should not be issued because Applicant’s facility is a majorsource exceeding ten tons for acetaldehyde emissions subject to case by case MaximumAchievable Control Technology Requirements for major hazardous air pollutant sources.

5 Comments Applicable to Multiple Emission Units

5.1 Various Emission Unit Sections of the Draft Permit Do Not Contain SufficientFederally Enforceable Physical Production Rate and/or ThroughputLimitations on the Potential to Emit to Ensure the Facility Does Not ExceedMajor Stationary Source Thresholds and Individual Emission UnitLimitations

Michigan DEQ-AQD must incorporate emission unit specific physical limitations on theproduction rate and/or process throughput rate in order to properly limit the potential toemit to reflect and validate potential to emit emission calculations in the application. Such limits must be stated on an hourly and an annual basis. Since air quality modelingdemonstrations depend on short term emission rates, hourly emission limitation are alsonecessary to protect ambient significant deterioration increments.

Michigan DEQ-AQD must also require enforceable requirements for monitoring,recordkeeping and reporting on such physical limitations on the potential to emit in orderto assure compliance and practical enforceability.

Such physical limitations are also necessary to ensure that E85-Corunna remains belowthe 100 ton potential to emit threshold for major stationary source status.

Michigan DEQ-AQD is required to run its new source permitting activity pursuant toClean Air Act requirements. Under the Court’s holding in the case of U.S. v. Lousiana-Pacific Corporation, D. Colo., blanket emission limitations cannot be considered asprovisions which limit the potential to emit of an emission unit. This court decision isdescribed by EPA in its June 13, 1989 Guidance on Limiting the Potential to Emit:4

“In United States v. Louisiana-Pacific Corporation, 682 F. Supp. 1122 (D. Colo.Oct. 30, 1987) and 682 F. Supp. 1141 (D. Colo. March 22, 1988), Judge AlfredArraj discussed the type of permit restrictions which can be used to limit a source'spotential to emit. The Judge concluded that:

... not all federally enforceable restrictions are properly considered in thecalculation of a source's potential to emit. While restrictions on hours ofoperation and on the amount of materials combusted or produced are


properly included, blanket restrictions on actual emissions are not. (682 F.Supp. at 1133)

The Court held that Louisiana-Pacific's permit conditions which limited carbonmonoxide emissions to 78 tons per year and volatile organic compounds to 101.5tons per year should not be considered in determining "potential to emit" becausethese blanket emission limits did not reflect the type of permit conditions whichrestricted operations or production such as limits on hours of operation, fuelconsumption, or final product.

The Louisiana-Pacific court was guided in its reasoning by the D.C. Circuit'sholding in Alabama Power v. Costle, 636 F. 2d 323 (D.C. Circuit 1979). BeforeAlabama Power, EPA regulations required potential to emit to be calculatedaccording to a source's maximum uncontrolled emissions. In Alabama Power, theD. C. Circuit remanded those regulations to EPA with instructions that the Agencyinclude the effect of in-place control equipment in defining potential to emit. EPAwent beyond the minimum dictates of the D.C. Circuit in promulgating revisedregulations in 1980 to include, in addition to control equipment, any federallyenforceable physical or operational limitation. The Louisiana-Pacific court foundthat blanket limits on emissions did not fit within the concept of proper restrictionson potential to emit as set forth by Alabama Power.

Moreover, Judge Arraj found that:

...a fundamental distinction can be drawn between the federally enforceablelimitations which are expressly included in the definition of potential toemit and (emission) limitations.... Restrictions on hours of operation or onthe amount of material which may be combusted or produced ... are,relatively speaking, much easier to "federally enforce." Compliance withsuch conditions could be easily verified through the testimony of officers,all manner of internal correspondence and accounting, purchasing andproduction records. In contrast, compliance with blanket restrictions onactual emissions would be virtually impossible to verify or enforce.

Thus, Judge Arraj found that blanket emission limits were not enforceable as apractical matter.

The Louisiana-Pacific court was guided in its reasoning by the D.C. Circuit'sholding in Alabama Power v. Costle, 636 F. 2d 323 (D.C. Circuit 1979). BeforeAlabama Power, EPA regulations required potential to emit to be calculatedaccording to a source's maximum uncontrolled emissions. In Alabama Power, theD. C. Circuit remanded those regulations to EPA with instructions that the Agencyinclude the effect of in-place control equipment in defining potential to emit. EPAwent beyond the minimum dictates of the D.C. Circuit in promulgating revised


regulations in 1980 to include, in addition to control equipment, any federallyenforceable physical or operational limitation. The Louisiana-Pacific court foundthat blanket limits on emissions did not fit within the concept of proper restrictionson potential to emit as set forth by Alabama Power.

Moreover, Judge Arraj found that:

...a fundamental distinction can be drawn between the federally enforceablelimitations which are expressly included in the definition of potential toemit and (emission) limitations.... Restrictions on hours of operation or onthe amount of material which may be combusted or produced ... are,relatively speaking, much easier to "federally enforce." Compliance withsuch conditions could be easily verified through the testimony of officers,all manner of internal correspondence and accounting, purchasing andproduction records. In contrast, compliance with blanket restrictions onactual emissions would be virtually impossible to verify or enforce.

Thus, Judge Arraj found that blanket emission limits were not enforceable as apractical matter.” (EPA memo at p 8-10)

The draft permit features a physical limit on maximum annual grain to be loaded at asingle emission unit, another limit on the maximum annual product to be run throughloading racks at another emission unit and a source-wide natural gas combustion limit.. However, these physical limitations do not necessarily sufficiently limit the potential toemit on all of the other individual emission units on site because of pollutant-specificpotential to emit and process issues. Examples of such circumstances are discussed inthe specific emission unit sections of this Comment.

5.2 Draft Permit Provisions Addressing Hazardous Air Pollutants (HAP)

5.2.1 Applicant Has Not Quantified Hazardous Air Pollutants From FugitiveEmissions

Commenters can find no mention in Applicant’s of the presence of hazardous airpollutants (HAPs) contained in plantwide fugitive emissions from plant components. TheApplicant has identified 8.22 tons per year of fugitive plant-component-related volatileorganic compounds. While HAPs may be a small portion of the component fugitives,such emissions are not zero and must be quantified in the Application to determine theircontribution to plant-wide totals, both for individual HAPs and for total HAPs.


5.2.2 The Draft Permit Should be Amended to Include Specific Acetaldehyde,Acrolein and Formaldehyde Limits

The draft permit provisions for the distillation, prefermentation and fermentationscrubbers and the two dryer/thermal oxidizer units should each be amended to includehourly and annual numerical emission limitations for acetaldehyde, acrolein andformaldehyde.

For the fermentation scrubber exhaust in particular, the Applicant should be heldaccountable on these three HAPs for the very high control efficiency assumed for thatemission unit in Applicant’s emission characterization.

If the Applicant insists that scrubber units will control VOC and HAP emissions with acontrol efficiency of 99.5% to 99.7%, then the draft permit should enact a controlefficiency emission limitation for the scrubber controlled units. These are very highcontrol efficiency claims and they should be backed up with emission limitation andcompliance testing requirements.

5.3 Neither the Draft Permit, Nor MDEQ’s Current Policies on Approving StackTesting Practices, Contain Any Assurances that the Total Mass Rate ofVolatile Organic Compound Emissions is Measured for Emission LimitationsCompliance Purposes, for Evaluation of the VOC BACT Stringency of VOCEmission Limitations and For Purposes of Determining the Major SourceStatus of the Proposed Facility

5.3.1 EPA Policy is Clear that the Clean Air Act New Source Review ProgramsMust Ensure Accountability for the Total Mass Rate of Volatile OrganicCompound Emissions

EPA directives are clear that direct or unaltered use of “as carbon” or “as propane” measurements for purposes of new source review and Title V applicability andcompliance are not permissible:

“For the other regulated pollutants that you listed, with the exception of VOC,calculation of the actual or potential emissions for purposes of NSR and title Vapplicability should follow the EPA principles for developing emission factors,inventories and test methods for the subject pollutant. For VOC emissions,however, it is recognized that the EPA’s test methods do not measure the pollutantmass exactly or only measure a subset of the pollutant mass. Nevertheless, for thepurposes of both NSR and title V applicability, our policy has been that VOCemissions should be calculated as the total mass of VOCs. That is, a value for eachvolatile organic compound known to be emitted should be calculated separatelyand the sum of the individual values should be reported as total VOCs (e.g., 20 tpy


5 June 5, 2001 letter from John Seitz, Director, EPA Office of Air Quality Planning andStandards, to D. Edward Settle, Manager, Air Quality, ThermoRetec Corporation, Golden, COavailable on EPA’s Region 7 NSR website or from Commentors.

of toluene and 26 tpy of methyl ethyl ketone should be calculated separately andthen reported as 46 tpy of VOC). This follows our guidance in the document titled“Procedures for Preparing Emission Factor Documents,” where we indicate thatemission factors for VOCs should be reported “in terms of actual weight of theemitted substance.” Those organic substances which are specifically excludedfrom EPA’s definition of VOC at 40 CFR § 51.100(s), because they have“negligible photochemical reactivity,” should not be included in the total VOCemission calculation for NSR and title V applicability. The document also providesan exception in the case of unknown species by stating that such emissions shouldbe calculated using an “educated guess” or a molecular weight of 44 (for reportingas propane). Where necessary, this procedure should be used to calculateemissions of those volatile organic compounds that cannot otherwise bequantified.”

“It is the EPA’s intent that a consistent approach be taken, wherever possible, toquantify and report pollutant emissions for its various air programs. Thus, themethods described above for quantifying pollutant emissions would also apply toour procedures for such things as NSR netting, emission trading and offsets, aswell as for other SIP-related programs for criteria pollutants.”5

This is clear articulation of EPA policy for new source review air permitting proceedings. As a result, reliance on emission estimation methods reflecting VOCs measured only ascarbon or propane that understate the total mass of VOC species emitted cannot be usedto compare to emission limitations and to evaluate the source status as to the 100tons/year major stationary source threshold.

The draft permit should be amended to cite EPA’s guidance document concerning themeasurement of volatile organic compounds from ethanol plants entitled “MidwestScaling Protocol for the Measurement of ‘VOC Mass Emissions’ VOC Sampling at Wetand Dry Grain Mills and Ethanol Production Facilities,” August, 2004 (See Attachment#9). Permit language should be added to clarify that all Method 25/25A determinationsshould be subject to EPA’s current generic scalar of 2.2, or to be otherwise in compliancewith the protocols contained in the Midwest Scaling Protocol. The draft permit shouldbe amended to ensure that the result measured for volatile organic compound emissionsbe appropriately scaled if EPA Methods 25/25A are used before the VOC results arecompared with the legally enforceable volatile organic compound emission limitations forthe two emission units mentioned in the prior paragraph.


5.3.2 As Proposed, the Draft Permit Does Not Specify Any Test Methods orCompliance with EPA’s Midwest Protocol for Total VOC Mass RateEmissions Determination from Ethanol Plants

The draft permit contains no information or requirements on how the total mass rate ofvolatile organic compounds will be determined from the dryer and the scrubber emissionpoints. Failure to specify such information means there is no way to evaluated thestringency of VOC BACT determinations since the compliance test method chosen willhave a large effect on the ultimate stringency of the compliance test.

MDEQ-AQD must not place sole reliance on unmodified and/or unadjusted compliancedeterminations by the Applicant’s use of EPA Methods 25 and 25a for VOC emissionlimitation compliance purposes. However, the lack of any specific test methodinformation and the failure to embrace EPA’s Midwest Protocol for dry mill ethanol planttesting virtually ensures that MDEQ-AQD will allow the facility to used methods whichdo not measure the total mass rate of oxygenated VOC compounds expected to be emittedby the facility.

EPA volatile organic compound emission test Methods 25 and 25A do not account for thefull mass rate of all volatile organic compound emissions in a manner that accounts forthe full molecular weight of VOC compounds emitted by ethanol production operations. EPA Method 25 (as carbon) and 25A (as propane) completely discount the effect on thereported VOC mass emission rate from the presence of oxygenated compounds. Most ofthe VOC compound species emitted by ethanol production operations are oxygenates,including alcohols, aldehydes, acids, ketones, glycerols, etc.

Without a clearly delineated test method and means to determine compliance set forth inthe Draft Permit, it is impossible to have a context to understand the stringency of anyparticular VOC emission limitation. EPA Method 25 and 25A can be expected to yieldresults which are less than 45% of the true value of the actual VOC mass emission ratefrom ethanol production emission units. As a result, sole reliance on reported results ofEPA Method 25 and 25A test methods will allow the facility to evade actual compliancewith VOC BACT requirements in Michigan’s State Implementation Plan and a verifiablecompliance determination with major stationary source thresholds for volatile organiccompounds.

Without a clearly delineated compliance procedure, emission limitations for VOC in theproposed permit are not practically and federally enforceable.

The Draft Permit should be amended to embrace EPA’s Midwest Scaling Protocol (SeeAttachments #9 & #10) or to otherwise force the subject facility to ensure that itscompliance stack tests for volatile organic compounds reflects the full molecular weightof all VOC species emitted. Care must be taken to ensure that any interferences andweakness in any test method used in detecting certain categories of compounds, such as


aldehydes, does not interfere with accurate VOC determination. EPA has also stated thatthe scalar of 2.2 may be applied to EPA Method 25/25A determinations. However, EPAnotes in the protocol that FID detectors used in Method 25A determination are hinderedin their ability to detect oxygenates even as propane, so care is required in test methodselection.

5.3.3 MDEQ’s Apparent Practice with Ethanol Facilities is to Unlawfully andImpermissibly Use Unadjusted EPA Test Method 25 and 25A Determinationsto Stand for VOC Emission Limitation Compliance

Attachment #11is an MDEQ approval letter on a test method protocol for MichiganEthanol. This protocol indicates that Michigan DEQ-AQD is approving VOCdeterminations using Method 25A that embraces only one small part of EPA’srecommended Midwest Protocol for determination of total mass rate VOC emissions. Inthe Michigan Ethanol approval letter, the Method 25A analyzer response is compared to aknown ethanol standard and then MDEQ-AQD simply lets the source reports an “asethanol” result for VOC emissions.

This approach, which doesn’t come close to reaching a determination on the total massrate of volatile organic compound species for compliance purposes, cannot be considereda valid compliance determination for ethanol plant dryer and scrubber emissions determination. for the following reasons.

First, the method doesn’t recognize EPA’s finding that Method 25A FID units have adiminished response to oxygenates and that a response curve against a known compoundstandard for each oxygenate component must be determined. Doing such a responsedetermination against ethanol only neglects the multiplicity of other oxygenatedcompounds found in typical exhausts. For dryer units, ethanol will not be the mostimportant or even the largest ethanol species present. As a result of this Michigan-approved practice, Method 25A FID units are never properly calibrated to account fortheir problems in detecting oxygenates.

Second, reporting volatile organic compounds “as ethanol” means that the mass ratecontributions from multiple oxygenates (e.g. acetic acid - 2O; lactic acid – 30; glycerol -3O) which will be predominate species in some flows will not be reflected in the reportedresults. Reporting VOCs “as ethanol” means the full effect of such multiple oxygenatesis neglected and inaccuracies in determining higher molecular weight VOC species.

Michigan DEQ should fully embrace the EPA Midwest Protocol in its ethanol planttesting results and follow the example of states like Minnesota who have previouslyconfronted these problems. See Attachment #12, pages 3-6 on “Quantifying VOCEmissions” and “Data Summary and Interpretation.” Embrace of the EPA protocolshould be incorporated into permits as Illinois EPA has done in order to give notice to all


parties about the problematic VOC emission determination problems associated withthese facilities.

5.3.4 MDEQ-AQD Does Not Have Rules or Firm Policies About “DepartmentRequirements” on Ethanol Plant Testing

Each of the testing sections in the draft permit contains language saying that testing willbe required “....in accordance with Department requirements...” However, MDEQ-AQDdoes not have any rules for testing ethanol plants. MDEQ-AQD ad hoc determinationsin the compliance testing protocol approval process cannot ensure uniform stringencyacross multiple situations and ethanol plant compliance testing programs. MDEQ-AQDshould be able to articulate the fundamentals of what should be required in the text of thetesting requirements of its permits, but apparently chooses not do so by stripping all suchdiscussions from its permits. No other state in the Midwest has chosen this indefiniteand arbitrary path on specificity for ethanol compliance plant testing requirements.

At this writing, Michigan’s air pollution control rules provide the following; first, thedefinition of “volatile organic compound”:

“(f) "Volatile organic compound" means any compound of carbon or mixture ofcompounds of carbon that participates in photochemical reactions, excluding thefollowing materials, all of which have been determined by the United Statesenvironmental protection agency to have negligible photochemical reactivity:

[text omitted]

The methods described in R 336.2004 and R 336.2040 shall be used formeasuring volatile organic compounds for purposes of determining compliancewith emission limits. Where such a method also measures compounds withnegligible photochemical reactivity, these negligibly-photochemical reactivecompounds may be excluded as volatile organic compounds if the amount of suchcompounds is accurately quantified and such exclusion is approved by thedepartment.” R 336.1122(f) (emphasis added)

Then, in reviewing R 336.2004 and R336.2040 there are no rules or methods listed whichexplicitly address the measurement of the total mass rate of emissions from ethanolplants. The effect of the rule in the definition of “volatile organic compounds” is torequire use of methods which are not appropriate when taken alone in doing compliancetesting with such plants. Until Michigan enacts appropriate rules to address thissituation, clear articulation of testing methods and requirements must be put into permitsin order to properly enforce and carry out Clean Air Act new source review andcompliance requirements. Placing the matter into the realm of post-permit-processnegotiated informal agreements between sources and MDEQ compliance test evaluation


staff denies opportunities for public comment and accountability and undermines theenforcement process by failing to enact clearly known and publicly vetted enforceablecompliance testing requirements.

5.3.5 Compliance Testing During Maximum Emissions Potential

Although Michigan rules require that compliance testing occur during maximumemissions from maximum process rates, this requirement has not been written into thetesting provisions of the draft permit and such a change should be made.

During recent ethanol plant emission testing of scrubbers at the US BioEnergy plant atWoodbury, MI, MDEQ-AQD field staff allowed VOC/acetaldehyde testing to occurduring a time when one of three fermenters was being emptied to a beer well. Thisprocess mode interval is not one when maximum VOC emissions would be generated.

5.3.6 Chemical Speciation Listing for EPA Method 18 Determinations Should beExtended

EPA Method 18 or 320 determinations (or determinations using altered methodssuggested by NCASI) should include all of the following specific speciated volatileorganic compound emissions:

acetaldehyde, acetic acid, ethanol, formaldehyde, formic acid, 2-furaldehyde, methanol, butyric acid, glycerol, pyruvic acid, lactic acid, propionic acid, butanol, acrylamide, acrolein, isoamyl alcohol, ethyl acetate, succinic acid, butanediol, isoamyl acetate, acetone and urethane

Several of these compounds are well known yeast fermentation byproducts with higherboiling points (100 -300 Deg C) that will be present in “syrup” that is evaporated productfrom thin stillage and which is introduced to DGS dryers where such material is eitherdirectly volatilized or subject to thermal decomposition and incomplete combustion.

Acrolein, in particular, is a carcinogen which is a thermal breakdown product of glycerol,a principle fermentation byproduct present in syrup.

5.4 The Draft Permit Does Not Have a Realistic, Practical Enforcement Methodfor FG-FACILITY Emission Limitations

Although the Draft Permit has facility-wide emission limitations, there is no practicalmethod provided in that section of the permit to actually enforce such facility-widelimitations. Any such compliance method would necessarily involve testing and


compliance evaluation determinations at individual emission units and flexible groups. However, many such sections in the draft permit are written without annual emissionlimitations and without any limitations or compliance measures relating to fugitiveemission sources. This is an unacceptable, not practically enforceable situation as itrelates to FG-FACILITY emission limitations.

The Draft Permit must be rewritten to include annual emission limitations on all point andfugitive emission sources in addition to the hourly limitations provided. Each suchsection must include a specific compliance determination method for annual emissionlimitation compliance evaluation. Each emission limitation, either short or long term,must have an accompanying physical throughput or production rate limitation on thepotential to emit in order to ensure that there will be actual emission limitationcompliance.

Failure to provide enforceable provisions and procedures to ensure the FG-FACILITYemission limitations are met is an arbitrary decision by MDEQ-AQD that will allow theApplicant to evade major stationary source thresholds for PM and criteria pollutants andfor hazardous air pollutants. In addition, such failure to include sufficient physicalpotential to emit limitations and numerical emission limitations means that compliancewith prevention of significant deterioration will be jeopardized.

5.5 Compliance Monitoring of Fabric Filter Controlled Emission Units

As presently written, there are no requirements that test ongoing fabric filter performanceafter a single, initial stack test. Fabric filter controls can deteriorate from wear and agingeffects on equipment. There is no way that the single stack test conducted shortly afterthe commencement of operations is capable of detecting the future current performance ofthe fabric filter emission control units potentially many years later. The draft permitshould be amended to provide for fabric filter leak detection monitoring on all fabric filtercontrolled emission units.

Although the permit does require EPA Method 9 determinations and pressure dropmonitoring to check on fabric filter emissions, such determinations cannot be used toassure compliance with numerical emission limitations because they are not sensitiveenough to detect more subtle fabric filter defects, such as pinholes. Additional leakdetection, physical inspection requirements and periodic testing requirements should beplaced in the permit for all fabric filter controlled units.


6 See the note in EPA Method 5 at the top of page 378 prior to section 6.1.1.9 of themethod.

7 See the note in EPA Method 5 at the top of page 378 prior to section 6.1.1.9 of themethod.

5.6 Testing Requirements for Condensible Particulate Emissions

The Fact Sheet claims:

“By definition, testing for PM10 includes condensible particulate.”

However, the terms “condensible particulate” and “PM10" are not defined in Michiganair rules. The testing methods for “particulate matter” do not list EPA Methods 201,201A and 202 in Michigan’s rules on compliance testing. Rule 336.1116(c) defines“particulate matter:”

“(c) "Particulate matter" means any air contaminant existing as a finely dividedliquid or solid, other than uncombined water, as measured by a reference testspecified in R 336.2004(5) or by an equivalent or alternative method.”

Unfortunately, condensible particulate matter can be emitted as a gas consisting of highmolecular weight hydrocarbons so it doesn’t necessarily exist as a “liquid or solid” at thepoint of emission. Moreover, the referenced EPA tests at rule R 336.2004(5) do notrequire the collection and reporting of ‘back half’ sampling and collection trainparticulate, leaving such reporting solely with the discretion of the regulated party.

EPA Method 5 leaves the matter of whether condensible particulate matter in the “backhalf” of the PM sampling train is collected for purposes of reporting along with thefilterable PM at the sole discretion of the stack testing operator and/or Applicant in theabsence of a legally enforceable requirement to incorporate the back half catch in thereported results.6

As a result, MDEQ-AQD is asking for a “faith-based” determination that a test methodwhich is not specified in the draft permit will always result in reporting of both filterableand condensible particulate matter for all future tests conducted and that suchrequirements not be made federally enforceable. This is highly objectionable and doesnot ensure accountability of the regulated party for proper emissions determination afterthe commencement of source operation. EPA Method 5 leaves the matter of whethercondensible particulate matter in the “back half” of the PM sampling train is collected forpurposes of reporting along with the filterable PM at the sole discretion of the stacktesting operator and/or Applicant in the absence of a legally enforceable requirement toincorporate the back half catch in the reported results.7


8 55 Fed. Reg. 12426 (March 17, 1990). See also 55 Fed. Reg. 14246 (April 17, 1990)(“emissions that contribute to ambient PM10 concentrations are the sum of in-stack [non-Condensible] PM10 . . . and Condensible emissions.”); 55 Fed. Reg. 41546 (October 12, 1990)(“Condensible particulate matter (CPM) emissions form very fine particles in the PM10 sizerange and are considered PM10 emissions”); 56 Fed. Reg. 65433 (December 17, 1991) (same).

9 March 31, 1994 letter from Thompson Pace, SO2/Particulate Matter Program Branch,EPA Office of Air Quality Planning and Standards to Sean Fitzsimmons, Iowa Department ofNatural Resources

The draft permit should be amended to require with affirmative language in all testingprovisions involving PM and PM-10 compliance stack tests that any EPA Method 5 determinations require the “back half” catch of the Method 5 sampling train to beincorporated in the reported PM/PM-10 emission totals, or that the source shouldotherwise be required to use EPA methods 201/201A and 202.

Commenters remind Michigan DEQ-AQD that compliance determinations for ensuringthat modeled PM10 air quality demonstrations are representative of actual emissionsrequire that filterable and condensible PM stack test results must be added together toevaluate compliance with PM10 emission limitations and this is a matter of considerablefederal interest. In fact, EPA does not approve the approach of setting PM 10compliance only to filterable “front half” PM stack test determinations. EPA hasrecognized that....

“...condensible emissions are also PM10, and that emissions that contribute toambient PM10 concentrations are the sum of in-stack PM10 and condensibleemissions.”8

Similarly, EPA’s Office of Air Quality Planning and Standards has statedunequivocally that....

“[s]ince CPM is considered PM-10 and, when emitted, can contribute to ambientPM-10 levels, applicants for PSD permits must address CPM if the proposedemission unit is a potential CPM emitter.”9

EPA has repeatedly required permitting authorities to include condensible PM10 limitsand testing methods in permits.

In retrospect, all emission units where PM-10 testing is to be accomplished should ensurethat condensible PM is reported in PM emission totals.


5.7 Applicant’s Emission Characterizations for All Fabric Filter ControlledEmission Units Do Not Incorporate Actual Cubic Feet Per Minute ProcessGas Flows

Most/all of the emission calculations done for grain receiving, handling and hammermillparticulate emission control are written on the basis of “air flow” clearly labeled as “DSCFM” and “PM loading” clearly labeled as “gr/dscf.”

Fabric filter performance guarantees for the types of control units to be used on grainhandling/receiving/hammermill sources provide for guarantees on the basis of actualcubic feet per minute rather than dry standard cubic feet per minute. Use of dry standardcubic feet per minute rather than actual cubic feet per minute means that all of the airflows for grain/material handling emission units are understated. As a result, all of theemission estimates might be similarly understated; because actual emissions would begreater than provided in the emission calculation, the Owner/Operator would be unable tocomply with emission limitations.

The Application should either provide actual cubic feet per minute air flows or certify onthe record that all of the flow indicated as “DSCFM” are the same as actual cubic feet perminute.

5.8 In Violation of Michigan New Source VOC BACT and Air Use RuleRequirements, the Applicant Has Not Provided an Appropriate BestAvailable Control Technology Determination for All Volatile OrganicCompound Emission Sources and Has Not Properly Quantified Emissions atAll Such Sources

Michigan rule 336.1702(a) requires that new sources of volatile organic compounds, evenfor minor sources, apply “best available control technology” (BACT) which is a federallyapproved Michigan State Implementation Plan rule. Michigan BACT for new sources ofvolatile organic compounds is fully as stringent as, and equivalent to, BACT as appliedto volatile organic compound emissions from major stationary sources in federalprevention of significant deterioration permits.

Ohio EPA has a much weaker rule for control of such new, minor sources called OhioBest Available Technology, which is defined:

“(T) "Best available technology" or "BAT" means any combination of workpractices, raw material specifications, throughput limitations, source designcharacteristics, an evaluation of the annualized cost per ton of air pollutantremoved, and air pollution control devices that have been previously demonstratedto the director of environmental protection to operate satisfactorily in this state orother states with similar air quality on substantially similar air pollution sources.”


The Applicant has recently gone through air permitting for nearly identical plants of thesame size, technology and configuration as the planned Corunna facility at both Newark,OH and Lancaster, OH. In making it draft air permit to install decision at the beginningof March, 2007, Ohio EPA nevertheless insisted that control of several additional processunits through use of the thermal oxidizer constituted Ohio BAT and was required underits weak VOC control rule.

Prior to the issuance of the draft permits for E85-Newark, OH and E85-Lancaster, OH,the Applicant’s corporate entity accepted these additional control requirements for VOCcontrol from several of these sources. I have placed and marked the relevant pages fromthe draft air permit to install permit for E85-Lancaster, OH at Attachment #13.

Most/all of the volatile organic compound process sources that will be controlled with athermal oxidizer at the two planned E85 plants in Ohio will be uncontrolled (or poorlycontrolled as at the distillation vent scrubber) as proposed with the draft permit for E85 -Corunna:

Cooking and mash formation – mixer, slurry tanks, cook tubes, flash tanks,liquifaction tanks

Pre-fermenters – yeast conditioning vessels

Distillation process – beer column, side stripper, rectifier column[controlled by a vent scrubber but not bythermal oxidizer at E85 - Corunna]

Stillage handling – surge tank, centrifuge, evaporators/vaporizers,syrup tank, conveyors, thin stillage tank, wholestillage tank

In addition, the Applicant did not properly characterize all of the VOC emissions from allof its planned uncontrolled emission units at E85 - Corunna for the same emission unitsthat will be controlled with a thermal oxidizer at the two planned E85 Ohio plants.

Because of these failures, the full potential to emit for VOC emissions is not known in asituation where admitted VOC emissions are very close to 100 tons. In addition, by notproperly characterizing all VOC emissions at all of its emission units, the Applicant failsto conform to Michigan Part 2 air use approval requirements.

In the Liberty Renewable Fuels proceeding, that Applicant indicated that VOC emissionswere significant from the “centrate tank” – See Attachment #15. The calculateduncontrolled emission from tank in the Liberty Renewable Fuels proceeding was 2.4lbs/hr using the same method as was used for the other miscellaneous VOC sources in


10 Seehttp://www.corm.us/images/stories/Ethanol/townmeetingpresentation--corunna.pdfslide 11.

that application. The resulting 10+ ton/year VOC emission was ultimately controlled bya thermal oxidizer. In the E85 - Corunna plant, the equivalent units would be EU-FEEDTANK and possibly portions of EU-DECANTER. These are presently shown asuncontrolled sources for which the Applicant has performed no emission characterization. Because of their uncontrolled potential to emit as demonstrated in the Liberty RenewableFuels application for that planned facility, the entire E85-Corunna facility would beplaced well over 100 tons for VOC emissions. Assuming most/all of the uncontrolledunits have zero emissions that don’t need to be characterized, tested and monitored is anarbitrary decision by MDEQ that defies facts in the record and common sense.

The Applicant has displayed a remarkable lack of candor in not revealing the decision ofOhio EPA and E85's corporate acceptance of that decision for VOC control at severalemission units at the planned Corunna facility. As part of a BACT demonstration, theApplicant should have revealed that the emission units it sought to leave uncontrolledwere being required for control in another jurisdiction at identical process plannedfacilities.

5.9 The Applicant Has Failed to Disclose or Evaluate its Emissions, PollutionControls and Ambient Health Impacts from of an Important HazardousMaterial at the Planned Corunna Facility

The Applicant has disclosed its planned use of 2.8 million pounds of “enzymes” in thepower point presentation provided to the community at the January 24, 2007 publicmeeting about the planned Corunna facility.10 The Applicant’s submittals indicate thatalpha-amylase is used to break down starches to dextrins and gluco-amylase is added toconvert dextrins to glucose.

Process information from Figure F-3 in the Application indicates the use of 159 lbs perhour of alpha-amylase and 191 lbs per hour of gluco-amylase for a total usage of 350 lbsper hour of these enzymes. That suggests annual enzyme usage over 3 million pounds peryear.

The application contains absolutely no information about methods of delivery of theseenzyme powders to the facility, how they are delivered for loading, their storage, transferand other pertinent facts. There is no information on emissions in the application ofthese materials or arrangements for emission control from loading and transfer operationsof these powders. No information is provided on possible ambient impacts or healtheffects of these materials.


Alpha-amylase is a known airborne toxicant in the State of Michigan. MDEQ-AQD AirToxics section has developed an Initial Threshold Screening Level for alpha-amylase of avery, very low 0.02 micrograms/cubic meter for a 1 hour average. Alpha-amylase is amember of a substances class known as a “subtilisin” and is also known as a proteolyticenzyme.

This class of compounds are known as respiratory sensitizers and can induce humanrespiratory asthma, which is the reason for the very low Michigan Initial ThresholdScreening Level.

The Applicant must disclose details about its use of this material, including what form itwill be receiving it in, what concentration alpha-amylase will be in with the materials asreceived and what arrangements will be used for receipt of alpha-amylase andglucoamylase if they are received as dry powders, including emissions, emission controlsand predicted ambient impacts.

5.10 The Applicant Has Not Properly Characterized Acrolein and UrethaneEmissions for the Subject Facility

Acrolein and urethane are federally regulated hazardous air pollutants whose emissionsmust be properly characterized in proposed facilities. Such characterization is necessaryto make a determination whether a facility is major or minor for hazardous air pollutantsand to determine the health and environmental impact of the facility under Michigan Part2 air use rules.

The Applicant has admitted to acrolein emissions only from the emergency firewaterdiesel engine and no other emission source at the site.

Acrolein emissions are known to occur in natural gas fired combustion units. The draftpermit should not issue as long as the application is incomplete for failure to properlycharacterize acrolein emissions from natural gas fired combustion units. In addition,acrolein is a known thermal degradation product of glycerol which will be present inabundance in wet DGS charge to the dryers and subject to thermal degradation reactionsin the dryer hot gas loop.

Urethane is a known fermentation byproduct and is also a reaction product betweenalcohols present and urea that is used in Applicant’s process. The Applicant has failed tocharacterize urethane emissions associated with the planned process equipment.


5.11 Other Comments

Appendix A does not require determination and reporting of “out of control” periods forcontinuous emission monitoring equipment in addition to the requirement to determineand report full scan exceedances. Such “out of control” periods must be determined,quantified and reported as per 40 C.F.R. Part 60, Appendix F – Quality AssuranceProcedures.

6 Discussion of Permit Regulatory Sections and Emission Calculations byIndividual Emission Unit and Process Groupings

6.1 Natural Gas-Fired Boilers #1 and #2

6.1.1 Applicant Has Erroneously Characterized the Stack Gas Exhaust VolumeBased on the Wrong Natural Gas F-Factor

Table 16-A shows a stack gas volume of 75,605 acfm based on “f Factor, wet, 1040BTU/scf, HHV.” This flue gas characterization must be incorrect since the EPA Method19 F-factor on a wet flue gas basis is 10,610 wet scf per million BTU at standardconditions of 20 deg C (68 deg F) and 760 mm Hg (29.92 in Hg). At 150 million BTUper boiler and standard conditions, flue gas discharge would be about 26,500 scf/minuteor an equivalent of about 38,200 acfm at the actual 300 degree F stack temperature. Thislower volumetric discharge rate would mean a lower stack exit velocity, less plume riseand a greater ambient air quality impact of emissions from the subject emission unit. Themodeling analysis assumed over 75,000 SCFM per stack, which is greater than a Method19 F-factor based exhaust volume.

6.1.2 Applicant Will Not Be Able to Comply with the NOX Emission Limitationsfor Natural Gas Fired Boilers #1 and #2 with the Planned Equipment

The Applicant has indicated a very low emission factor of 0.011 lbs NOX per MMbtu asthe basis of emission characterizations for the two natural gas fired boilers. At the sametime, the Applicant has indicated its plans for use of Low NOX Burners (LNB) for theseunits.

Emission control performance in the range of 0.011 lbs NOX per MMbtu is characteristicof the use of Ultra Low NOX Burners (ULNB) and Flue Gas Recirculation (FGR) and/orULNB/FGR with Selective Catalytic Reduction (SCR). In the last several years, therehave been few BACT determinations outside of California that have embraced emissionsperformance at the 0.011 lbs NOX per MMbtu level, and these have incorporatedULNB/FGR or ULNB/FGR/SCR. If all the Applicant is intending to install is LNBrather than ULNB/FGR or ULNB/FGR with SCR, emission control performance would


more likely be in the neighborhood of 0.03 lbs NOX per MMbtu heat input and theApplicant would violate permit limits for heat input and mass rate emission limitations. That would also put the Applicant over 100 tons of NOX per rolling 12 month period forpotential to emit.

Applicant is not allowed to claim a very high level of emission control performance intheir application that they are not intending to carry out as such a practice wouldconstitute sham permitting. Proposing such a high level of performance in order to meetmajor stationary source emission thresholds with a view to later gaining an emissionlimitation relaxation would also be impermissible permitting practice.

As part of a pre-construction review of a minor source, MDEQ-AQD must verify that theApplicant is fully able to comply with the expected provisions and permit emissionlimitations using the claimed LNB approach. MDEQ-AQD must not make an arbitrarydecision allowing the Applicant to proceed even as Applicant’s process and technologycontrol submittals for the two onsite boilers do not pass muster for assuring compliancewith claimed levels of emission control performance. Such an arbitrary decision ignoresreasonably available information fully accessible to MDEQ-AQD and the Applicant. The Applicant should either be required to submit vendor guarantees they can complywith 0.011 lbs NOX per million btu with LNB alone. Otherwise they should be requiredto commit to ULNB/FGR or ULNB/FGR/SCR as a permit-enforceable requirement Such a requirement is not in the draft permit as presently proposed and the Applicationcontains no commitment for ULNB and/or ULNB/FGR/SCR at the present time.

6.1.3 The Emission Limitations of the Draft Permit Should Not Place a Mass perUnit Time Emission Limit Bubble Over the Two Boilers

Mass per unit time emission limitations in the draft permit should not allow a bubble forboth boiler units. Each boiler has a separate stack and they should be able to complywith mass rate emission limitations at each unit. Declaring a mass rate emissionlimitation over two separate emission units with different stacks creates an enforcementvulnerability associated with proving simultaneous emission totals. Each boiler shouldbe subject to an emission limitation that is half of the mass rate per hour emissionlimitations presently included in the permit.

6.1.4 The Draft Permit Should Require Testing for both Filterable and CondensibleParticulate Matter

Merely saying that a stack should be tested for PM-10 doesn’t mean that there is anenforceable requirement to test for both filterable and condensible particulate matter. Astack can be tested for PM-10 by Method 201 and the facility can legitimately claim ittested for PM-10. The draft permit should explicitly require that EPA Method 201 and


202 be used to test for PM-10, or that EPA Method 5 be used provided the back half ofthe PM catch is reported in with total test results.

6.1.5 The Applicant Hasn’t Used AP-42 Emission Factors for CharacterizingVolatile Organic Compound Emissions from the Boilers and MDEQ has NotSet Volatile Organic Compound Emission Limitations for the Natural GasFired Boiler Units

The Applicant claimed to be using AP-42 emission factors for characterizing the volatileorganic compound emission limitations from the two natural gas fired boilers. However,the potential to emit of 2.2 tons per year for boilers does not conform to an AP-42estimate. The Applicant used 0.0017 lbs VOC per MMbtu heat input.

Emission characterization by AP-42 factor at 1020 BTU per scf for natural gas is for afactor of 0.0054 lb VOC per million BTU. At 150 MMbtu/hr per boiler, each boiler willemit 3.5 tons VOC/year and not 1.1 tons per year as depicted in the draft permit.

When characterizing volatile organic compound emissions using AP-42 from these twoemission units, the Applicant will not be able to comply with their draft permit and theoverall facility potential to emit will exceed the 100 ton major stationary source thresholdfor volatile organic compounds.

MDEQ-AQD’s decision to allow Applicant the use of the wrong emissions factor forVOC emissions from the boiler and to ignore the impact of this decision rendering thefacility a major source for volatile organic compounds emissions is arbitrary.

Finally, as proposed the draft permit contains no annual or hourly volatile organiccompound emission limitations. Therefor there is no way to enforce volatile organiccompound limitations on the two natural gas fired boiler units. When there is no methodto enforce the potential to emit limitation at a single emission unit, then the entireapproach for limiting annual facility-wide emissions under FG-FACILTY is jeopardized.

The draft permit should be amended to incorporate an hourly and an annual volatileorganic compound emission limitation, VOC testing and other compliance requirements.

6.1.6 Boiler Monitoring Conditions Require Revision

Condition 15.11 is written to require only a single device to monitor natural gas usage forthe two boilers. This is not sufficient to allow the natural gas usage limit in condition15.4 placed on each boiler to be monitored and enforced. Condition 15.11 should beclarified to ensure that each boiler is separately monitored for natural gas usage.


11 Applicant submitted a March, 2007 Malcolm-Pirnie drawing, Revised Figure F-2,“Corunna, MI Facility Layout”

Similarly, Condition 15.12, 15.13 and 15.14 are written for monitoring devices in thesingular and not the plural for boilers operating with two different stack emission points.

6.2 Site Roads Fugitive Emission Unit

6.2.1 E85, Inc. Submitted False Information in its Air Permit Application InvolvingFugitive Road Dust Emission Calculations

In its February, 2007 submittals, Applicant submitted emission calculations for fugitiveroad emissions assuming a 0.68 mil truck trip length. MDEQ responded in a February15, 2007 data request:

“In looking at the plant layout, the 0.68 mile truck trip length used for road PMemission estimates looks a little low. Please verify the trip length. The permit willlimit the total miles traveled, so if the trip length is underestimated, the totalnumber of truck trips would have to be correspondingly reduced.”

Resolution/Action: “We have added a factor of safety and consequently increasedthe truck trip lengths to 0.9 miles. The revised calculations are attached.”

In March, 2007, the Applicant submitted a revised site plan, including site road network,11

for the subject facility as part of its further air quality modeling activities. The additionalsubmittal included a revised Table 14 (See Attachment #5) showing a truck trip length of0.9 mile per trip for all truck shipments and deliveries at the facility.

Thus having a second chance to provide accurate information about truck trip lengths forits facility Applicant slightly revised upward its projected truck trip lengths. However,even this revised information was false and misleading and has directly led to erroneousemission characterization of the subject facility.

Attachment #6 is a reduction of the final Site Facility maps provided in large sheet formatwith some limited cleanup to allow annotations and point lettering and arrows to showspecific points on the map for trip length calculation. We assume, based on the needs ofsafety planning, that all site roads other than the segment A-B-C are one way roads, thatthe north-south segment C-D is one way south and the Segments F-H-M and I-J-L are oneway north. The following table shows nominal trip lengths for all types of shipments anddeliveries expected at the facility:


Activity – E85 - Corunna,MI site

Route MapInches

ActualDistance(ft)

Miles

Truck Grain Delivery ABCDEFHMCBA 72.4 7240 1.4

Truck Ethanol ShippingLoadout

ABCDEFGIJLMCBA

81.3 8130 1.5

Truck Gasoline Delivery ABCDEFGIJLMCBA

81.3 8130 1.5

Truck DDGS ShippingLoadout

ABCDEFGIJKLMCBA

84.9 8490 1.6

Truck Wet DGS ShippingLoadout

ABCDEFGIJKLMCBA

84.9 8490 1.6

Truck Chemical Deliveries ABCDEFHMCBA 72.4 7240 1.4

All of the truck trip lengths in the table above that will actually take place at the facility are considerably longer than the 0.90 miles assumed by Applicant when they submittedtheir revised materials and two times or more than their original 0.68 miles per truck tripFebruary, 2007 application information. Although the Applicant claimed their truck tripinformation was “conservative,” in fact, the Applicant submitted false and misleadinginformation concerning truck trip lengths for site traffic and for calculation of thepotential to emit for fugitive emissions from roads.

Michigan law provides that a permit application can be rejected on the basis of submittal of false information:

“In accordance with this part and rules promulgated under this part, the departmentmay, after notice and opportunity for public hearing, deny or revoke a permitissued under this part if any of the following circumstances exist.....

(c) The person applying for the permit makes a false representation or providesfalse information during the permit review process.” MCL 324.5510(c)

At the very least, the Applicant was under an obligation to correct their error when theMDEQ air permit engineer questioned the original erroneous 0.68 miles per truck basisfor fugitive emission calculations. The Applicant did not make such an error correctionand instead termed their revised truck trip length as “conservative.” The failure toprovide accurate information in their application is clearly grounds for MDEQ to denypermit issuance to the Applicant under MCL 324.5510(c).


12 Liberty Renewable Fuels - Response to Comments Document, March 5, 2007, Page 20

If there was a knowing submittal of a false material statement about roads and on-sitetruck trip length, or a knowing failure to supply the correct information about roads andon-site trip length, then Michigan law provides the following criminal penalty:

“A person who knowingly makes a false material statement, representation, orcertification in, or omits material information from, or knowingly alters, conceals,or fails to file any notice, application, record, report, plan, or other documentrequired to be submitted pursuant to this part or a rule promulgated under this part,or who knowingly fails to notify or report information required to be submittedunder this part or a rule promulgated under this part, or who knowingly falsifies,tampers with, renders inaccurate, or knowingly fails to install any monitoringdevice or method required under this part or a rule promulgated under this part, isguilty of a misdemeanor punishable by imprisonment for not more than 1 year anda fine of not more than $10,000.00 per day, for each violation.” MCL 324.5531(2)

Because of Applicant’s egregious false information submittal, the permit application aspresently constituted should have never have been proposed for approval as a draft permitby MDEQ-AQD because it reflect arbitrary and capricious decisionmaking by the agencyand a situation in which the Applicant will not be able to comply with emissionlimitations and other applicable requirements for the facility.

Finally, in the Liberty Renewable Fuels proceeding, a similar issue came up in regard tothat applicant’s underestimation of road length, vehicle miles traveled and thus fugitiveroad emissions. MDEQ answered comments objecting to this situation by saying:

“Review of the site plans, and additional site plans submitted in response to thiscomment, indicates the trip length may be more than one mile. The actual triplength, to be used in calculating emissions due to truck traffic, will be determinedonce the facility is constructed.”12

MDEQ-AQD’s response must not provide a similar response in the present proceedinginvolving Applicant’s site road network for the Corunna facility and Commenter’sobjections above. MDEQ-AQD’s response to this issue in the Liberty Renewable Fuelsproceeding was arbitrary and capricious; MDEQ-AQD’s response to such matters directlyabdicates the agency’s duty to conduct a pre-construction review based on the Applicant’ssubmittals and information that is required under the pre-construction review provisionsof 42 U.S.C. §7475 when the subject facility is a major stationary source as is the case forthe subject facility as per these comments. Such an approach to delaying thedetermination of what potential emissions will be occurring also violates Michigan’s Part2 air use rules.


13 MDEQ Fact Sheet on E85-Corunna, dated May 10, 2007, but issued earlier with thepublic notice.

14 The Liberty Renewable Fuels comments are available at:http://www.sagady.com/workproduct/LASERCommentLibertyRenewableFuels.pdf

6.2.2 MDEQ’s Fact Sheet Made Erroneous Assumptions About Applicant’sAssumed Silt Loading Factors and Fugitive Road Emission Calculations forPM and PM-10

MDEQ’s Fact Sheet for the issuance of the draft permit for the E85, Inc.-Corunna facilitycontained the following paragraph:

“Fugitive dust from plant roadways: The silt loading factor for the roadways (theamount of dirt that can cause particulate matter (PM) emissions) was assumed to be 0.5 g/m2. Using 0.6 g/m2, as suggested by a commenter on another ethanolplant, would increase PM emissions from the roadways by about half a ton. Sincethe proposed PM emission limit for the facility is 26.9 tons per year below themajor source level, testing to determine the actual silt loading of the plantroadways is not required to ensure the facility is a minor source. The total milesallowed to be traveled on the plant roadways are limited, which serves to limit thePM emissions from the plant roadways. In addition, a fugitive dust controlprogram for the plant roadways is required.”13

MDEQ’s “Fact Sheet” paragraph is erroneous. Applicant’s uncontrolled fugitiveemission factors of 2.85 lbs PM per vehicle mile traveled and 0.55 lbs PM-10 per vehiclemile traveled are the results of using a silt loading factor of 5.0 grams per square meterand Applicant’s other assumptions with the applicable AP-42 equation. E85, Inc.’spresent use of the 5.0 gram per square meter silt loading factor is more reasonable thanthe 0.4 gram per square meter silt loading factor used in the Liberty Renewable Fuelsproceeding (note however, we still contest their uncontrolled emission per VMT factorson other grounds explained in another subsection). Commenters reserved the right tocontest any decision by MDEQ to allow E85, Inc. to use a smaller silt loading factor than5.0 grams per square meter in any subsequent final decision, during any potential to emitenforcement action and/or in any annual emission inventory report. Commentersincorporate by reference Section 8.8.3 of the LASER comments (and associatedAttachments #1-#4 shown in the present document) on Liberty Renewable Fuels ofJanuary 29, 2007 objecting to use of a factor of 0.4 grams per square meter for the Libertypermit in this proceeding.14

Commenters note that E85, Inc. has proposed use of a silt loading factor of 5.0grams/square meter to Ohio EPA for identical plants to the one sited for Corunna,Michigan at Lancaster, OH and Newark, OH.


15 Seehttp://www.corm.us/images/stories/Ethanol/townmeetingpresentation--corunna.pdf slide 11. This material is available from Commenters if it is no longer posted.

Commenters assert that E85's “sample calculations” in Table 14 of the application leavesan erroneous impression that Applicant used a 0.5 gram per square meter silt loadingfactor when, in fact, Applicant actually used a 5.0 gram per square meter silt loadingfactor.

6.2.3 The Draft Permit Doesn’t Contain Any Requirements to Verify Compliancewith Silt Loading Rates Assumed in the Emission Calculation

The Draft Permit should be amended to require quarterly testing of road silt loading andspecification of a recognized test method for such silt loading monitoring. TheOwner/Operator should be put under a burden of proving through silt loading testing thatthe assumptions made during potential to emit characterizations remain reflective offacility operations during the life of the operation. Determination of actual emissionsshould incorporate a real world determination of the silt loading rates and vehicle milestraveled at the site.

6.2.4 Applicant Fugitive Road Emission Calculation Failed to Account for VMTfrom Truck Delivery of Process-Related Chemicals

Applicant has given power point presentations concerning its facilities to localgovernments involved.15 One slide from these presentations indicates that Applicant willbe getting the following chemical deliveries:

2.8 MM lbs of enzymes per year7.5 MM lbs of urea per year5.9 MM lbs of sulfuric acid per year3.9 MM lbs of ammonia per year

This is a total of 20.1 MM lbs of cargo deliveries or just over 10,000 tons of deliverables. The site plan indicates such chemical deliveries will be at a truck-accessible location atarea 1900 indicated as “chemical unloading and storage.” If we assume that thesedeliveries will be roughly in 25 ton cargo increments based on 80,000 lb truck loads, thismeans 400 additional truck trips that Applicant has not accounted for in their fugitiveemission calculations.


6.2.5 Condition 5.3 Should Require Improved Recordkeeping andContemporaneous Collection of Actual Truck Traffic Data

Condition 5.3 should be amended to more specifically set forth Applicant’s recordkeepingand reporting duties to physically limit the potential to emit for fugitive road emissionsfrom the facility. Applicant must not rely on erroneous truck trip length assumptions thatthey have attempted to foist upon the MDEQ in their application. Because the trip lengthdepends on the type of shipment or delivery being made, recordkeeping must take placecontemporaneously with vehicles arriving. Recordkeeping to properly supportappropriate physical limits on the potential to emit must indicate the date of each truckarriving and its purpose for shipments and/or deliveries. The recordkeeping must reflectactual data collected with contemporaneous recordkeeping of actual truck traffic, notiterations of assumptions about what traffic has occurred, as the Applicant has attemptedto negotiate in air permits its seeks at two Ohio facilities.

Finally, the Applicant must be put under an obligation of reporting any exception to therequirements to conform to Condition 5.2 in allowing excessive truck traffic on site orother violation of fugitive emission control requirements.

6.2.6 The Draft Permit Should be Amended to Require that the Applicant Maintaina Working Street Sweeper On-site at All Times

Identification in the emission unit table for EU-TRKTRAFFIC doesn’t contain anyidentification of the emission control for this emission unit. The Owner/Operator shouldbe placed under a condition requiring that the facility purchase or lease and maintainonsite at all time a working street sweeper to support the claim Applicant has made for50% control efficiency in the fugitive emission calculation. Merely having a fugitiveemissions control plan in place is not sufficient to ensure that the emission limitation willbe maintained in the absence of (or failure to maintain in working condition) such aphysical piece of equipment.

6.2.7 Section 5 of the Draft Permit for EU-TRKTRAFFIC Must be Amended toIncorporate an Annual and an Hourly Numerical PM and PM-10 EmissionLimitations in Addition to the Physical Limitation on the Potential to Emit

There is no way to enforce Section 18 of the permit for EU-FACILITY for PM and PM-10 annual emission limitations unless there is a means to make PM and PM-10 emissionlimitations enforceable at each emission unit at the facility. If there are no emissionlimitations for PM and PM-10 applicable to fugitive road emissions at the site, there isn’tany effective way of enforcing an overall EU-FACILITY PM and PM-10 emissionlimitation.


EU-TRKTRAFFIC must have hourly emission limitations (or other short term limit not toexceed a 24 hour averaging time) in order to ensure the facility does not cause a nuisanceand does not jeopardize compliance with the PM-10 National Ambient Air QualityStandards and the 24 hour PM-10 PSD increments. If there is no emission limitation,there is no way to ensure that the facility will comply with a level of emissions reflectingwhat was used in the modeling study. In addition, there will be no way to use annualemission inventory reports to enforce against PM and PM-10 emissions violations at thisemission unit.

6.2.8 MDEQ-AQD Must Not Remove Responsibility from the Applicant by MerelyAssuming that Rail Shipments Will Always Enable Applicant to Avoid PavedRoad Fugitive Emissions

While Applicant may have considerable ability to ship grain to the site via railroad fromits cooperating elevator partners, Applicant may find that rail shipments of product,DDGS and WDGS via railcar will be considerably more problematic.

Rapid expansion of the ethanol production industry may cause short and long termshortages of tank cars in dedicated ethanol product delivery service. Some railroads haverefused DDGS shippers the use of their covered gondola rail cars for DDGS service asthis sometimes sticky material can cause significant bridging, offloading and cleaningproblems in actual rail car services. Applicant may have to acquire rail cars fordedicated DDGS service and if that occurs, increased truck deliveries of DDGS, modifiedwest DGS and WDGS may become attractive and necessary. For these reasons, MDEQmust not let Applicant escape the truck traffic/paved road/fugitive dust issue without afinal and definitive solution and permit limitation.

6.2.9 The Applicant’s Intentions to Actually Comply with their Permit to InstallMust be Questioned in Light of their Public Statements

Applicant’s fugitive road emission characterization assumes that only 10% of the graindeliveries to the facility will be in trucks, with the rest being made by railcar. However,the Applicant has also intimated that it will be using local corn at its facility from localproviders who would bring dried corn in grain trucks, not in rail cars. Applicant’ssubmittal does not show the worst case fugitive emissions from roads when graindeliveries exceed the assumed 10% proportion on grain deliveries as claimed by theApplicant in its public statements.

The Applicant has circulated the following Q & A document to local officials in Corunnaand to the public:


“35. Exactly how much extra traffic (train or road) will be produced due to thisplant's operations, and what time of day will this traffic be occurring?

Corn is expected to be delivered by truck and rail. If all com is delivered bytruck, this will amount to 136 trucks per day. If all com is delivered by rail,this will amount to about 30 rail cars per day.

E85 plans to acquire as much corn locally as possible to benefit the localeconomy. This could conservatively account for 1/4 of the corn beingdelivered via truck or 34 trucks per day with the remainder being deliveredvia rail

The distillers grains produced would be shipped via rail predominatelyresulting in about 9 railcars per day.

Ethanol will be shipped via rail resulting in about 11 railcars per day.”

The Applicant is thus telling the public and local officials one thing and telling MDEQ airpermit engineers something completely different as to its intentions for transportationmodal dispatch to and from its facility.

With a maximum capacity of 900 tons per hour, it is within the physical design capacityof the proposed facilities to receive 100% of grain deliveries via truck. As a result, theApplicant understated the potential to emit of the proposed facility by understating the VMT for grain truck deliveries at 10% of annual deliveries at the fugitive roads EU-TRKTRAFFIC emission unit. The Applicant clearly states their intent to emphasizelocal corn deliveries that “conservatively” could be 25% of grain deliveries.

6.2.10 The Draft Permit Fails to Require that Applicant Maintain All Roadways andParking Lots in a Paved Condition

Nothing in the draft permit actually requires the Applicant to maintain all of its roadways,parking lots and staging areas in a paved condition.

The Applicant has submitted facility layout plans showing Plant Area 0100 identified as a “truck staging area.” This area is shown as a park/staging area for a total of 24 trucks inan island in the truck road entrance to the facility. In E85's application for identical plantsto the present one at Corunna at Lancaster and Newark, Ohio, E85 showed its intent tohave the truck staging area be unpaved.

Applicant has not quantified the fugitive emissions from such an unpaved area and hasnot provided clear, quantified area information for this staging area.


The draft permit should be amended to require that the Owner/Operator maintain allroadways, parking lots and staging areas onsite in a paved condition.

6.2.11 Monitoring and Reporting Requirements Must Embrace Enforcement ofRoad Sweeping Requirements

The draft permit must be amended to provide accountability for fugitive emission controlrequirements assumed to provide a 50% reduction over uncontrolled emission rates. Otherwise, the uncontrolled emission rate should be assumed as the potential to emit.

Recordkeeping requirements must emphasis daily sweeping and recordkeeping to showsuch measures have been completed. In addition, recordkeeping and exception reportingmust address any downtime on the functioning of the roadway sweeper or other controlequipment.

6.2.12 Applicant’s Fugitive Road Emission Characterization Contains AssumptionsWhich Cause Significant Underestimation of Fugitive Road Emissions

The Applicant assumed that the payload for all truck deliveries was 25 tons, either bydirect statement or by implication from “notes.”. Applicant noted truck trips were halffull (35 tons) and half empty (10 tons). Applicant then when on to assume that the28,953,750 gallons of denatured ethanol that would be shipped via truck would be intrucks holding 8200 gallons each. However, 8200 gallons of ethanol weighs 27 tons, soeither applicant’s upper bound truck weight for ethanol deliveries is wrong or more tripswill be made to haul the same amount of ethanol shipments at fewer gallons per truck.

As noted in the prior paragraph, Applicant assumed empty trucks were 10 tons and fulltrucks were 35 tons. However, several other state air permitting jurisdictions, including Illinois and Indiana, have issued air permits with fugitive emissions based on 40 ton fullweight trucks and 15 tons empty weight trucks. The Liberty Renewable Fuels proceedingassumed for emission calculations 40 tons full and 15 tons empty. MDEQ is nowconsidering the Marysville Ethanol, which assumes 40 tons full and 15 tons empty forsite shipments and deliveries. Gasoline marketing delivery vehicles that may take backloads of ethanol to refineries and bulk gasoline plants can commonly carry as much as13,500 gallons with weights well over 80,000 lbs on six to seven trailer axles. Truckscapable of hauling 8200 gallons with limited trailer axles will commonly be loaded at80,000 lbs for interstate travel. E85's emission calculation allowance for 10 ton emptyand 35 ton full for truck tankers is too low and the emission calculations should berevised for heavier empty and loaded vehicles in the same manner as other Great Lakesstates are figuring such fugitive emissions in ethanol permits. At the very least, MDEQshould explain why special treatment is being afford to E85 on the truck weight fugitiveemissions calculation basis issue.


6.2.13 Commenter’s Fugitive Road Particulate Emission Calculation Factors forApplicant’s E85, Inc. Corunna, MI Facility

Correction of the problems raised above with the fugitive emission factors yields thefollowing particulate emission VMT factor calculations using the following assumptions:

Empty tanker and grain trucks – 15 tons

Full tanker and grain trucks – 40 tons

Full tankers contain 7500 gallons of ethanol at 6.58 lbs/gallon at a weight of 24.7tons

Commenters accept Applicant’s assumption of 5.0 grams per meter squared for siltloading and 50% control for sweeping provided the draft permit is amended toensure that the Owner/Operator is accountable for daily road sweeping measures.

Mean vehicle weight = 27.5 tons

E = [(k[sL/2]0.65 * [W/3]1.5) - C] * [1 - (P/4N)]

EPM = [(0.082 * [5/2]0.65 * [27.5/3]1.5 ) - 0.00047] * [1 - [100/(4 * 365)]]

EPM = [ ( 0.082 * 1.814 * 27.8 ) - 0.00047 ] * [ 1 - 0.0685 ]

EPM = [ 4.134 ] * 0.9315 = 3.85 lb PM / VMT

EPM-10 = [(0.016 * [5/2]0.65 * [27.5/3]1.5 ) - 0.00047] * [1 - [100/(4 * 365)]]

EPM-10 = [( 0.016 * 1.814 * 27.8 ) - 0.00047 ] * [ 1 - 0.0685 ]

EPM-10 = [ 0.8064 ] * 0.9315 = 0.751 lb PM-10 / VMT

6.2.14 Revised Vehicle Miles Traveled Information and Subsequent RevisedParticulate Emission Calculations for E85 - Corunna

Attachment #7 shows revised vehicle mile traveled calculations for E85 - Corunna underthree different set of assumptions using the revised, corrected onsite truck trip length dataand revised VMT emission factors shown in this section.

These new VMT data are incorporated in the following revised emission calculations forE85 - Corunna:


Revised E85 – Corunna Fugitive PM Emission Calculations Using Revised VMTScenario VMT Uncontrolled

PM EmissionWith Revised3.85 lb/VMTFactor (tons)

PM With 50%Control (tons)

UncontrolledPM-10Emission withRevised 0.751lb/VMTFactor (tons)

PM-10 With50%Control(tons)

Original ApplicantAssumptions withCorrect Truck TripLength, HeavyTankers

41492 79.9 40.0 15.6 7.8

As per E85 PublicStatement,“Conservative”25% Grain TruckDeliveries plusMisc ChemicalDeliveries, HeavyTankers

51695 99.5 49.8 19.4 9.7

As per E85 PublicStatement,“Conservative”25% Grain TruckDeliveries plusMisc ChemicalDeliveries, 40 TonTankers

52301 100.7 50.4 19.6 9.8

Note that all calculations above assume trucks that are 40 tons loaded and 15 tons empty,even though “heavy” tankers will be above 40 tons.

Because the Applicant has falsified its fugitive road emission characterization byfalsifying its expected vehicle miles traveled as a result of falsifying its truck trip lengthcharacterizations, there is serious question whether the subject facility can operatewithout violating the draft permit conditions.

Under all VMT scenarios above, the Applicant will violate the VMT restriction and emitsignificantly greater PM and PM-10 emissions from the fugitive road emission unit thanwas depicted in the application. Under the last two scenarios, PM emissions will exceedthe 100 ton major stationary source threshold, and come quite close to the threshold forthe first scenario from the revised emission calculation alone. Commenters assert that thelast two emission scenarios reflect actual potential to emit from the planned facility giventhe company’s announced intentions.

MDEQ should not merely assume that Applicant will live with an 18,125 mile annuallimitation on truck VMT in light of the comments above. In addition, it should be noted


that 18,125 miles of VMT was in the emission calculations that are obsolete withApplicants false original submission of 0.68 miles per truck trip. This was replaced byApplicant’s second false VMT characterization of a total of 23,978.6 miles of VMT at 0.9miles per truck trip.

MDEQ should not assume that the Applicant can readily divert all/most of the plannedtruck deliveries and shipments to rail transportation without clear evidence from theApplicant in the form of proven ownership records or contractual agreements for thenecessary dedicated rail equipment to make such a diversion. Because of the rapidexpansion of ethanol production, demand for ethanol rail tankers is dedicated service isvery high. Commenters previously mentioned that some railroad lines are presentlydisallowing shipments of DDGS in their equipment because of handling/cleaning problems associated with the sticky DDGS product. In addition to the problem of havingadequate rail equipment available, there is also the market problem. The Applicant maybe subject to marketing forces that ensure that 100% rail transportation solutions may notbe available for site operations. DDGS markets are expected to saturate and theApplicant may not have much capability to market DDGS outside of the local area. Local area customers for DDGS means use of trucks for delivery, not rail transportation.

6.3 Loading Rack Emissions from Truck and Railcar Tanker Loadout

6.3.1 The Draft Permit Must be Amended to Incorporate Enforceable VOC, NOXand CO Emission Limitations

The present draft permit contains no hourly or annual emission limitations for VOC,NOX and CO. Failure to provide for enforceable emission limitations for both flarestack and fugitive emissions associated with loading rack operations means there can beno effective enforceable of facility wide emission limitations.

A potential to emit calculation must reflect the highest amounts of emissions that canoccur consistent with equipment design constraints and federally enforceable physicallimitations on the potential to emit.

6.3.2 The Draft Permit Should be Amended to Require Compliance with AllFeatures of Flare Control and Monitoring Requirements Specified in 40C.F.R. §60.18, Including Testing and Monitoring Requirements

The draft permit presently contains no effective requirements for monitoring, testing andreporting to ensure that all features of 40 C.F.R. §60.18 flare requirements are met. Thereis no requirement to test by a date certain the BTU content of flare gas and the maximumflare gas exit velocity during times when the maximum gas flow is expected to the flare.


These requirements should be incorporated into the permit. There is no requirement thatthe Owner/Operator physically install and maintain a continuous pilot flame monitor asprovided in 40 C.F.R. §60.18(f)(2), for example. There is no requirement that theOwner/Operator report all times when a pilot flame monitor was not working or when theapplicant operated the loading rack system when there was no pilot flame. These arecritical omissions since the uncontrolled rate of emissions from the loading rack will bevery high during a flare pilot outage.

There is no requirement that the Owner/Operator file quarterly reports showing eithercontinuous proper flare operation or environmental exception reports showing improperflare operation, loss of flare pilot or loss of flare pilot monitoring capability.

6.3.3 No Practically Enforceable Provisions in the Draft Permit Provide FederallyEnforceable Requirements Ensuring That All Truck and Railcar Tankers tobe Unloaded at the Facility Meet Appropriate On-Board Vapor SystemCollection Efficiency and Vapor “Tightness” Performance Standards andThat the Facility Ensure Compliance with Such Requirements ThroughMonitoring, Recordkeeping and Reporting

Nothing in the draft permit requires that the Applicant not load a truck or railcar unlessthe owner/operator obtains evidence from the tanker owner that such transportationequipment has passed an annual leak test within the past year for vapor tightness and thatany conveyance portions of vapor control systems on the tanker itself are properlyinstalled and operating before the facility loads that tanker. On-board tanker vaporcollection systems include piping, hatch opening seals, block valves, vapor controlvalves, vacuum breakers, etc.

Proper operation and collection efficiency of vapor collection systems depends on boththe fixed elements at the ethanol rack process area as well as mobile elements on thetransportation equipment. Failure to hold Applicant responsible for not loading non-compliant and leaky truck and railcar tankers will significantly increase fugitive volatileorganic compound emissions from product loading operations. Such operations may alsopose a safety risk at the subject facility.

If the permit were amended to include a requirement that the Owner/Operator beprohibited from loading tankers that did not have certified evidence of passing a vaportightness testing in the last year, the sheet found in the file as Attachment #8 would showthe correct emission calculation for VOC at a total of 2.86 tons VOC/year for truckloading and 0.67 tons VOC/year for rail loading for a total of 3.53 tons per year for bothpost-controlled loading rack flare and fugitive VOC emissions, combined.

The 3.53 ton per year total is higher than shown on the Applicant’s March 12, 2007version of Table 12-A and the March 9, 2007 version of Table 1-A for total facility-wide


potential to emit (listed erroneously as 1.59 tons/year total for truck and rail loadingcombined. These latter submissions by the Applicant reflect, in part, going back to theuse of a 0.6 saturation factor for truck uncontrolled emission calculation after they weredirected to use a factor of 1.0 by MDEQ. Commenters reserve the right to object to anyallowance by MDEQ providing that a saturation factor of 0.6 can be used to calculateemissions from truck loading operations at the proposed facility.

6.3.4 If MDEQ Fails to Provide a Federally Enforceable Requirements for theOwner/Operator to Verify the Leak Testing Status of Truck Tankers Loadedat the Facility, the Loading Rack Potential to Emit VOC and HAPCalculation Must be Rejected and Revised Upward

Both the existing AP-42 and the 1995 draft revised AP-42 5.2 sections indicate that thecollection efficiency for truck tankers that have not passed an annual leak test should beassessed at 70%. As a result, if MDEQ fails to provide for an enforceable requirement toensure that the Owner/Operator checks the annual leak testing certification, then worstcase fugitive emissions could be as high as 28.8 tons ov VOC per year. Otherwise,MDEQ should recalculate the fugitive emissions based on the proportion of tankers in thestate which do not receiving annual vapor tightness testing.

6.4 Rail and Truck Grain Receiving

6.4.1 The Draft Permit Contains No Federally Enforceable Requirements forEmission Limitations and Fugitive Emissions Control for Grain Receiving

Nothing in the draft permit requires the Owner/Operator to construct, maintain or usefugitive dust emission controls and practices and to observe numerical emissionlimitations for grain receiving fugitive PM and PM-10. Without such measures there canbe no assurance that the facility grain receiving fugitive emissions will be limited to the0.73 ton PM/year and 0.21 ton PM-10/year.

The Applicant’s emission characterization is predicated on a 98% collection efficiencyfor both PM and PM-10, including the use of choke flow conditions for unloading ofhopper cars and building enclosures for both truck and railcar unloading. There is nofederally enforceable provisions in the permit to ensure choke flow work practices(including grate manipulation) are carried out and that the enclosures are constructed andmaintained to ensure the claimed 98% control efficiency. There is no accountabilityrequired in the draft permit to maintain numerical emission limitations for the 0.73 tonPM/year and 0.21 ton PM-10/year level of grain receiving fugitive emissions. Thepermit should not issue without amendments to correct these problems. There can be nofirm assurances that PSD increments for PM 10 will be protected without such federallyenforceable conditions.


6.4.2 The Draft Permit Does Not Maintain Sufficient Conditions PhysicallyLimiting the Fugitive PM and PM-10 Potential to Emit

While Condition 6.2 of the draft permit does contain a provision limiting grain deliveries to 1,148,000 tons per rolling 12 month year, this condition alone cannot physically limitthe potential to emit for fugitive PM and PM-10 emissions from truck grain receivingwhen both hopper trucks and straight trucks used the facility.

EPA AP-42 emission factors from grain receiving are shown in the table below:

AP-42 Grain Receiving Uncontrolled Emission Factors (lbs of emissions per ton received)

Emission Source PM Emission Factor PM-10 Emission Factor

Straight Truck (SCC 3-02-005-05) 0.18 0.059

Hopper Truck (SCC 3-02-005-51) 0.035 0.0078

As can be seen from the table, the uncontrolled PM emission factor for straight truckgrain receiving is over 5 times higher, and the PM-10 uncontrolled factor is over 7.5times higher, than corresponding uncontrolled emission factors for hopper truckunloading. Applicant’s submittal contained a composite emission factor of 0.064 lbs PM/ ton and 0.018 lbs PM-10/ ton for truck delivery of grain. These factors were based onan 80/20 split between hopper grain trucks and straight grain trucks.

The problem is that nothing in the draft permit limits straight truck deliveries to no morethan 20% of actual truck deliveries. A federally enforceable provision making such alimitation enforceable by permit requirement is necessary to limit the potential to emit forfugitive emissions from truck grain receiving operations; the draft permit must beamended to incorporate such a requirement along with the necessary recordkeeping inCondition 6.9 as to straight truck deliveries.

6.4.3 Applicant’s Truck Grain Receiving Emission Characterization is Erroneous

While the Applicant’s method of calculating uncontrolled emissions from truck grainreceiving based on a defined 80/20 split for hopper vs. straight truck deliveries isacceptable, it is not acceptable for the Applicant to apply a 98% control efficiency to theaggregate calculated uncontrolled emission rate to arrive at a single controlled emissionnumber for both types of trucks.

Choke flow conditions of loading can be used to achieve an approximate 95% controlefficiency with a properly operated system of grate unloading of a bottom hopper methodof grain transfer. Choke flow conditions can be used to drastically minimize grain flow-air interface, drop heights and grain bin air entrainment. Adding another 50-60% control


associated with maintaining an enclosure can probably allow the facility to achieve a 98%overall collection efficiency factor for fugitive grain receiving emissions control.

However, straight truck dumping cannot be done in a manner to maintain choke flowconditions and minimizing grain-air interactions. As a result, it is erroneous in theemission calculation to take straight truck dumping and pretend that the same controlefficiency can be applied to straight truck dumping operations to achieve the same levelof emission control as can be achieved by hopper truck choke flow grain receiving.

At best, the straight truck dumping in a shed might be at 50-60% control. This leads tothe following fugitive dust emissions associated with straight truck dumping:

Scenario #1 – 10% grain receiving by truck, with 20% of this amount by straighttruck with 50% fugitive emission control:

1,148,000 ton/yr * 0.1 * 0.2 * 0.18 lb PM/ton * .5 / 2000 lb/ton = 1.03 tons/year





These additional fugitive PM emissions, combined with other additional emissions fromApplicant’s underestimations, could help push the entire facility over 100 tons/year forPM emissions.

The permit should not issue without rectifying problems associated with straight graintruck loading articulated in this and prior subsections.

6.5 Cooling Tower

6.5.1 The Draft Permit Fails to Provide Federally Enforceable PM/PM-10 EmissionLimitations, Monitoring Requirements and Physical Limitations on thePotential to Emit

Although the cooling tower is depicted as having a potential to emit of 10.8 tons PM peryear and 2.0 tons PM-10 per year, nothing in the draft permit provides for anything at allthat is practically enforceable about such emissions. There are no PM/PM10 emission


limitations, no testing requirements on the drift eliminator, no monitoring requirementsfor cooling water total dissolved solids, no ceiling requirements on maximum coolingwater total dissolved solids and no physical limitations on the cooling tower recirculationrate to ensure that there are physical limitations on the potential to emit for this emissionunit.

The absence of enforceable PM/PM-10 emission limitations and accepted levels forpotential to emit renders any attempt to enforce facility-wide PM/PM-10 emissionrequirements as being impractical.

The permit must no issue until these problems with the regulation of this emission unit areaddressed.

6.6 Dryers #1, #2, Product Cooling and Thermal Oxidizer

6.6.1 The Emission Limitations of the Draft Permit Should Not Place a Mass perUnit Time Emission Limit Bubble Over the Two Dryers

Mass per unit time emission limitations in the draft permit should not allow a bubble forboth dryer units. Each dryer has a separate stack and they should be able to comply withmass rate emission limitations appropriate for each unit. Declaring a mass rate emissionlimitation over two separate emission units with different stacks creates an enforcementvulnerability associated with proving simultaneous emission totals. Each boiler shouldbe subject to an emission limitation that is half of the mass rate per hour emissionlimitations presently included in the permit. Failure to require accountability foremissions at each stack by using a bubble emission limitation applicable to both stacksalso undermine the binding requirement for volatile organic compound best availablecontrol technology at these dryer process units.

6.6.2 The Draft Permit Should Require Testing for both Filterable and CondensibleParticulate Matter

Merely saying that a stack should be tested for PM-10 doesn’t mean that there is anenforceable requirement to test for both filterable and condensible particulate matter. Astack can be tested for PM-10 by Method 201 and the facility can legitimately claim ittested for PM-10. The draft permit should explicitly require that EPA Method 201 and202 be used to test for PM-10, or that EPA Method 5 be used provided the back half ofthe PM catch is reported in with total test results.


6.6.3 Startup Emissions for the Dryers

Section 12 of the permit should be amended to include a startup condition for processoperation of the dryers that prohibits introduction of WDGS into the dryer units duringstartup until the recirculating hot air/steam loop has achieved the required operatingtemperature shown during the last stack test to indicate compliance with emissionlimitations.

6.6.4 Limits on Dryer Natural Gas Usage are Not Sufficient to Limit the Potentialto Emit of Dryer Process Unit Emissions of CO, PM-10 and VOC

The draft permit contains a provision at Condition 12.4 limiting the natural gas charged tothe dryer thermal oxidizer to 789 million standard cubic feet per 12 month rolling timeperiod per dryer. This limitation will only partially limit the potential to emit of thisprocess unit. In addition to limitations on natural gas charged to the dryer thermaloxidizers, addition physical limitations must be place on either the DDGS production rate,the rate of thin stillage fed to the centrifuge or the rate of WDGS charged to the dryers. The carbon monoxide, VOC , PM and NOX emission rates are not solely surrogates ofthe dryer thermal oxidizer heat input rate. Dryer NOX, for example, consists of thermalNOX formed from dryer thermal oxidizers plus NOX derived from nitrogen containingdryer exhaust PM burned in the thermal oxidizers. Derivation of dryer exhaust PM,VOC and NOX are primarily a function of dryer process material throughputs rather thanheat input rates.

Maximum production rates and emission rates of the dryers for DDGS can take place atheat input rates less than the maximum burner natural gas firing rate. As a result, thedraft permit should be amended to physically limit the potential to emit of the two dryersby limiting the DDGS production rate to 19.5 tons per hour per dryer, provided there is aclear method for measuring the actual dryer DDGS production rate. If there is not a clearmethod for measuring the DDGS hourly production rate at the output conveyor of theDDGS dryers, then the physical potential to emit can be limited by the whole stillagepumping rate to the decanter/centrifuge at a rate of 580,838 lbs of whole stillage per hour.

It is essential to physically limit the potential to emit in the manner described in thissection to conform the potential to emit to the emission characterization (which assumes a39 ton/hour DDGS production rate) since the block flow diagram for the facilitycontemplates the physical capability of a 43.87 ton/hour DDGS rate and a 653,369lb/hour whole stillage pumping rate. This is a case where the Applicant did not disclosethe full magnitude of the design production rate for the dryer process equipment andupstream processes.

Moreover, note that the Applicant assumes the maximum DDGS throughput of 345,889tons/year of DDGS in Table 6-A, but then notes a “normal production” rate of 384,322


tons/year of DDGS in the general process information. This supports our comment thatthe Applicant didn’t use the maximum design basis of the process equipment for DDGSproduction and that the potential to emit of the dryer process units will be higher thanclaimed in the PTE emission characterization unless and until MDEQ-AQD placesfederally enforceable DDGS and/or whole stillage production rate limitations in the draftpermit to limit the PTE to the Table 6-A characterizations. Any such PTE limitation forthe dryer process units must also have sufficient monitoring, recordkeeping and reportingfunctions enacted to make such physical limitations to limit potential to emit federally andpractically enforceable.

6.6.5 The Application and the Draft Permit Do Not Include Sufficient Parameter

Monitoring and Other Testing Provisions to Assure Compliance with VolatileOrganic Compound and Hazardous Air Pollutant Emission Limitations

Apart from a single stack test and ongoing thermal oxidizer temperature monitoring, noother monitoring is conducted to ensure that all criteria and hazardous air pollutantemissions remain within emission limitations. There is no volatile organic compoundcontinuous monitoring.

In order to comply with the Clean Air Act, at the very least, the draft permit should beamended to include a requirement for a continuous oxygen monitor at the thermaloxidizer exhaust. Both oxidizer temperature and oxygen monitoring are necessary toensure proper combustion conditions in the oxidizer as a parameter monitoring surrogatefor control of volatile organic compounds and hazardous air pollutants.

6.6.6 Continuous Monitoring Provisions of the Draft Permit Should Require StackGas Flow Monitoring at the Stack Location Where Continuous Monitors areLocated

Unlike simple combustion systems, the Swiss Combi dryer units do not allow a singlesource parameter to be used in conjunction with an F-factor in order to accuratelydetermine stack gas flow. The draft permit should be amended to include a requirementfor a stack flow monitor to be used in association with continuous emission monitoring toensure accurate flue gas volume flow rate determination with integration of thisinformation with the continuous monitoring system for carbon monoxide and nitrogenoxides.


6.6.7 The Draft Permit Should Be Amended to Clearly Allow NOX and COEmission Limitation Violation Enforcement at the Dryer #1 and #2 EmissionUnits Through Use of Continuous NOX and CO Continuous EmissionMonitoring Results

The draft permit should be amended to clearly indicate that continuous emissionmonitoring information for NOX and CO can be used for compliance evaluation andenforcement purposes.

6.6.8 Applicant’s Acetaldehyde Emission Characterization for the Dryer UnitsLacks a Supporting Basis and Is Not Backed by a Vendor Guarantee orDemonstrated Compliance Stack Testing

Attachment #18 is the vendor guarantee on the dryer units for emission controlperformance. Note that the vendor is not guaranteeing any claimed emission controlperformance for acetaldehyde, but just indicates an “expected value.” The expectedvalue is qualified with a note saying that it is based on “Vogelbusch mass balance forethanol in the dryer feed of a 100 MGPY nameplace ETOh plant.” The Applicant’s massbalance approach for emission estimation of acetaldehyde in dryer emissions lackscredibility for process dynamic reasons. Mass balance techniques for emissionestimation are not amenable to the type of drying process that Applicant plans to usebecause destruction of volatile organic compounds in the dryer will likely involvesignificant products of incomplete combustion and thermal degradation products. In sucha circumstance it would be too difficult to adapt mass balance techniques to an emissiondetermination because of the complex chemical transformation dynamics within thesystem.

MDEQ-AQD should require the Applicant to submit acetaldehyde stack tests from othersimilar dryer units. Such results must certainly exist for the MGP Pekin IL plant cited inarticles attached to Applicant’s submittals. However, the Applicant didn’t submit anysuch acetaldehyde stack testing results. MDEQ-AQD’s decision in failing to requireApplicant’s accountability with real data for their dryer emission acetaldehyde emissionestimate represents an arbitrary decision by the agency. It is also a harmful decisionbecause it involves an attempt by the Application to escape major hazardous air pollutantstatus on a facility which almost certainly exceeds 10 tons per year of acetaldehydeemissions.


6.7 Fermentation Scrubber Emission Unit

6.7.1 Applicant Has Failed to Properly Support Their Emission Characterization ofthe Fermentation Scrubber Emission Unit

During permitting in Ohio, Ohio EPA asked the E85 to support its data submittal on VOCand HAP emission characterization from the fermentation scrubber with stack test data,but the Applicant’s submittal was inadequate. In Ohio, the Applicant submitted a singlestack test generally showing low VOC emissions, but there was no information sufficientto determine an emission factor and the stack test did not identify uncontrolled emissionrates and control efficiencies. As a result, Applicant’s claim that VOC and several HAPand non-HAP airborne toxicants will all be controlled with 99.50% control efficiency wasnot supported in the record of that Ohio proceeding.

Here in Michigan, MDEQ-AQD has made no demand on the Applicant to justify its claimed of 99.5%+ control efficiencies for VOC and HAPs from the vent, pre-fermentation and fermentation scrubber control units. Failure to require suchdocumentation from the Applicant constitutes an arbitrary decision by MDEQ-AQD.

Applicant’s HAP emission characterization for scrubber-controlled emission unitsprovided HAP data but lumped acetaldehyde, acrolein and formaldehyde together. Suchtreatment interferes with proper HAP emission characterization and ambient assessmentof airborne toxicant impacts..

6.7.2 Applicant’s Acetaldehyde Emission Characterization for the ScrubberControlled Units is Subject to Challenge as Unrealistically Low

The Applicant assumed approximately 99.5% control efficiency for acetaldehyde controlfrom the scrubber-controlled emission units. This is an unrealistic emission controlefficiency for this hazardous air pollutant and the Applicant has submitted no stack testinformation showing that such control efficiency can be consistently maintained forcontrolling acetaldehyde.

The Applicant must be placed under a duty to submit information to justify theiracetaldehyde emission rate projections rather than requiring MDEQ-AQD and the publicto accept them as an article of faith. Failure to require the Applicant to submit evidencethat their scrubber technology is likely to comply with 99.5% control for acetaldehydeconstitutes an arbitrary decision by MDEQ-AQD.

In information submitted by the State of Nebraska to US EPA in a rulemaking proceeding(see Attachment #16), that agency cited information indicating that ethanol facilities inthat state have had a great deal of difficulty achieving evan a 98% control efficiency onscrubber-related acetaldehyde emissions. Controlled acetaldehyde hourly emission rates


in some of the Nebraska examples of fermentation scrubbers cited significantly exceededthe proposed MDEQ-AQD draft permit combined acetaldehyde emission rate of 0.99 lbsper hour.

A recent fermentation scrubber acetaldehyde stack test done at US Bio-EnergyWoodbury, a 53 MMgal/year facility, indicates a violation of a 1.9 lb/hour acetaldehydeemission limitation (See Attachment #17), which is greater than the allowed emissions forApplicant’s proposed Corunna facility.

Ohio EPA staff had information in their files assuming a control efficiency as low as 68%for acetaldehyde.

Applicant’s assumed 99.5% control assumption and resulting mass emission limitationsprovide lower predicted emissions than many recent permit determinations made acrossseveral states. The Applicant should not be allow to mis-characterize their emissions withthe hope of avoiding major stationary source thresholds for volatile organic compoundsafter the fact of permit to install issuance. Any decision by MDEQ-AQD allowing suchtreatment by Applicant would constitute sham permitting and an arbitrary exercise ofauthority by the agency.

Given Applicant’s predicted acetaldehyde potential to emit of 3.7 for the twodryer/thermal oxidizer emission unit, the Applicant would have to show that the total acetaldehyde emissions from the three scrubber controlled units would not exceed 6.3tons per year. The Applicant admits to a total of 1170 tons per year for an uncontrolledemission rate from the scrubber controlled units. If the scrubber control efficiency waseven just 99.4% instead of 99.5%, the potential to emit for acetaldehyde would render theentire facility to be a major HAP source subject to case by case MACT and otherrequirements. The draft permit contains no emission limitation requirement that theApplicant actually demonstrate 99.5% control efficiency for acetaldehyde.

6.7.3 Compliance Testing and Parameter Monitoring in the Draft Permit for Scrubber-Controlled Emission Units is Not Sufficient to Ensure Compliance

Only a single compliance stack test is required under the draft permit. The parametermonitoring provided is not sufficient to assure compliance with the acetaldehyde emissionlimitation. There is no requirement to maintain required parameters on an hourlyintegrated basis. There is no requirement to address scrubber inlet temperature andscrubber liquid flow temperatures and verification of the molar concentration of bisulfiteliquid injection to the scrubbers.

For a wet, packed tower scrubber attempting to control gaseous pollutants, the parametermonitoring provided in the draft permit is not sufficient to ensure compliance and to


ensure that control efficiency is maintained at least to the level demonstrated by the laststack test.

For all the parameters cited in the draft permit and in this section of the comment, stacktesting must be used to establish suitable floors or ceilings on parameters demonstratingcompliance with emission limitations. In addition, thresholds and time intervals ofmaximum deviation before a malfunction condition is declared must be established. Once the performance test is completed, the draft permit should provide for a process toestablish all needed levels of parameter performance and acceptable conditions, and theseparameter monitoring protocols should be subject to MDEQ-AQD approval after proposalby the Owner/Operator.

6.7.4 The Draft Permit Should Be Amended to Require Continuous VOC EmissionMonitoring for the Fermentation Scrubber Emission Unit

Because the uncontrolled emission rates are high and the Applicant has made a claim forvery high control efficiency on the fermentation scrubber emission unit, continuousemission monitoring for VOC should be required for this emission unit. Continuousmonitoring is also justified in circumstances where the process is subject to variabilitybecause of stages of fermentation in the units the fermentation scrubber controls.

6.8 The Applicant Failed to Quantify Hazardous Air Pollutants from PlantFugitive Components for FG-NSPSVV

The Applicant has produced an emission characterization with zero hazardous airpollutants from plant fugitive components. This failure must not be allowed to stand.

Attachment #14 shows the hazardous air pollutant characterization for Marysville Ethanoland Liberty Renewable Fuels, with acetaldehyde emissions shown as 1.5 tons/year and1.4 tons/year, respectively.

With Applicant’s admission of potential to emit of 9.22 tons of acetaldehyde per yearfrom sources other than component leaks, realistic appraisal of Applicant’s acetaldehydeemissions, shown by Applicant’s competitors, would put Applicant’s facility over the tenton major stationary source hazardous air pollutant threshold for acetaldehyde withsignificant regulatory consequences not considered in the present permit.

The Applicant must be made to quantify all hazardous air pollutants associated with itscomponent leaks for the application to be considered complete. Given other plantpredictions of fugitive component emissions, Applicant’s facility air permit should beconsidered non-approvable because it is a major HAP source for acetaldehyde.

Attachment #1

Attachment #2

Attachment #3

Attachment #4

Attachment #5

Attachment #6

Administrator

Line

Administrator

Line

Administrator

Line

Administrator

Line

Administrator

Line

Administrator

Line

Administrator

Line

Administrator

Line

Administrator

Line

Administrator

Line

Administrator

Line

Administrator

Line

Administrator

Line

Administrator

Text Box

B

Administrator

Text Box

C

Administrator

Text Box

D

Administrator

Text Box

E - Grain Unloading

Administrator

Text Box

Administrator

Text Box

F

Administrator

Text Box

G - Ethanol Truck Loadout

Administrator

Text Box

H - Chemical Deliveries

Administrator

Text Box

I

Administrator

Text Box

J

Administrator

Text Box

K -- DDGS/WDGS Truck Shipping

Administrator

Text Box

L

Administrator

Text Box

M

Administrator

Text Box

A - Plant Site Entrance

Map Segment Map Inches Actual Feeton 100% Original at 1"=100'

A - B 22.5 2250B - C 2.2 220C - D 8.5 850D - E - F 3.2 320F - H - M 8.1 810F - G - I 4.6 460I - J 3.4 340J - K - L 8.0 800L - M 4.6 460M - C 3.2 320J - L 4.4 440

Attachment #7

E85, Inc. Corunna, MI VMT under original PTE assumptions only with revised trip length

Activity Truck % Thoughput Units Quan/trucktruck tips/year

onsite miles/trip

Annual milage

Truck Grain Delivery 10 114,800 tons 25 4592 1.4 6429Truck Ethanol Loadout 25 29,953,750 gallons 8200 3653 1.5 5479Truck Gasoline Delivery 100 5,515,000 gallons 8200 673 1.5 1009Truck DDGS Loadout 100 344,756 tons 25 13790 1.6 22064Truck Wet Cake Loadout 100 101,732 wet tons 25 4069 1.6 6511Truck Chemical Deliveries 100 0 tons 25 0 1.4 0

VMT 41492

E85, Inc. Corunna, MI VMT under "conservative" corn deliveries at 25%


onsite miles/trip

Annual milage

Truck Grain Delivery 25 287,000 tons 25 11480 1.4 16072Truck Ethanol Loadout 25 29,953,750 gallons 8200 3653 1.5 5479Truck Gasoline Delivery 100 5,515,000 gallons 8200 673 1.5 1009Truck DDGS Loadout 100 344,756 tons 25 13790 1.6 22064Truck Wet Cake Loadout 100 101,732 wet tons 25 4069 1.6 6511Truck Chemical Deliveries 100 10,000 tons 25 400 1.4 560

VMT 51695


onsite miles/trip

Annual milage

Truck Grain Delivery 25 287,000 tons 25 11480 1.4 16072Truck Ethanol Loadout 25 29,953,750 gallons 7500 3994 1.5 5991Truck Gasoline Delivery 100 5,515,000 gallons 7500 735 1.5 1103Truck DDGS Loadout 100 344,756 tons 25 13790 1.6 22064Truck Wet Cake Loadout 100 101,732 wet tons 25 4069 1.6 6511Truck Chemical Deliveries 100 10,000 tons 25 400 1.4 560

VMT 52301

E85, Inc. Corunna MI VMT under "conservative" corn deliveries and 80,000 lbs tankers, and misc chemical deliveries

Attachment #8

Attachment #9

Midwest Scaling Protocol for the Measurement of “VOC Mass Emissions”

VOC Sampling at Wet and Dry Grain Mills and Ethanol Production Facilities

U.S. Environmental Protection AgencyOffice of Air Quality Planing and Standards

Office of Regulatory Enforcement

August 2004

Version 1.6August 2004

Page 2 of 28

VOC Sampling from Wet and Dry Grain Mills and Ethanol Production Facilities

Introduction

This protocol is designed to determine the actual volatile organic compound (VOC) massemission rates from sources where significant amounts of oxygen-containing organic compoundsare emitted. Either U.S. EPA Method 25 or Method 25A is used to determine the total organiccompound concentration of the emission samples. The concentration data are then converted tocarbon mass (or propane mass) emission rates. Simultaneously, the concentrations of the mostsignificant individual organic compounds in the emission sample are measured with Method 18.

This protocol is designed to be used in conjunction with Methods 25 or 25A to provideaccurate VOC mass emission measurements from most air emission units at grain mills andethanol production facilities. VOC mass emissions based on concentration measurements withMethods 25 or 25A reported “as carbon” or “as propane” results in reported VOC emission ratesless than the actual emissions of the VOC pollutants. The Midwest Scaling Protocol (MSP)provides a way to convert the VOC results from “as carbon,” when Method 25 is used, or from“as propane,” when Method 25A is used, to “as VOC ” emission rates.

Sources in this industry may opt to use a standard scaling factor (SF) of 2.2 pounds ofVOC per pound of VOC as carbon instead of performing quantitative measurements ofindividual volatile organic compounds in order to derive individual scaling factors for eachsource. Alternatively, the MSP provides an acceptable means for the quantitative measurementsof air emissions of individual volatile organic compounds from sources at grain mills and ethanolproduction facilities. The MSP also serves as a reference for equations used to convert VOCconcentration measurements reported “as carbon” or “as propane” to actual VOC massemissions.

The decision to use Method 25 or Method 25A to measure total VOC concentrations issource dependent. In general, Method 25 is applicable to all sources with total VOCconcentrations >50 ppmC (parts per million carbon). Methane and carbon monoxideconcentrations are also measured with Method 25. However, referring to Method 25A, section1.1 of Method 25 states:

“Direct measurement of an effluent with a flame ionization detector (FID) analyzer maybe appropriate with prior characterization of the gas stream and knowledge that thedetector responds predictably to the organic compounds in the stream. If present,methane (CH4) will, of course, also be measured. The FID can be applied to thedetermination of the mass concentration of the total molecular structure of the organicemissions under any of the following limited conditions: (1) Where only one compound is known to exist; (2) when the organic compounds consist only of hydrogen and carbon;


Page 3 of 28

(3) where the relative percentages of the compounds are known or can be determined,and the FID response to the compounds are known;

(4) where a consistent mixture of the compounds exist before and after emission controland only the relative concentrations are to be assessed; or

(5) where the FID can be calibrated against mass standards of the compounds emitted(solvent emissions, for example).”

The FID used in Method 25A has a depressed response to organic compounds thatcontain oxygen. The tester must determine, for the specific FID unit used for each test, theresponse factor for each organic compound that constitutes 5% or more of the total mass of theindividual VOC species analyzed. A weighted average of these response factors shall be used toadjust the FID’s response to the actual emission samples. The tester shall adjust the analyzer’sresponse prior to converting the response to a mass emission rate.

If the tester uses Method 25A to measure VOC from a source where the moisture contentis greater than 10%, then the tester must normally dilute the sample using the procedures inMethod 205 to reduce the water content of the sample to less than 10%. The tester shall use aheated sample line to transport the sample from the stack to the analyzer to prevent condensationof water and organic compounds. At the time of this writing, at least one FID analyzer has atolerance for moisture content up to 40%. The moisture content for which Method 205 dilutionis required is analyzer-dependent.

One specific application of Method 18 for measuring the kinds of oxygen containingcompounds that are most common in the emissions from grain mills is the impinger methoddeveloped by the National Council for Air and Stream Improvement, Inc. (NCASI). NCASI hasdesignated this method as NCASI CI/SG/PULP-94.02, Chilled Impinger/Silica Gel Tube TestMethod at Pulp Mill Sources for Methanol, Acetone, Acetaldehyde, Methyl Ethyl Ketone andFormaldehyde (NCASI 94.02). Water soluble organic compounds are collected in impingersfilled with chilled laboratory grade water. Any target compounds that break through the chilledwater are collected on organic grade silica gel. Method 18 analytical procedures, gaschromatography with flame ionization detection or mass spectrometric detection, are used tomeasure the target organic compounds listed in Table 1.1 that are collected in the sampling train,except for formaldehyde which is measured by a colorimetric procedure. The sample collection,recovery and preservation procedures for this specific application of Method 18 are described inAppendix B along with recommended GC/FID procedures for most target compounds. Theanalytical procedures to measure formaldehyde are described in Appendix C. Additional GCoperating conditions may be necessary to quantify all of the water-soluble volatile organiccompounds on the target list.

These data are used to calculate the weighted average ratio of the VOC molecular weightdivided by the VOC carbon mass (or VOC propane mass). This SF is then used to convert thetotal organic carbon mass emission rate to the total VOC mass emission rate (i.e., the results areconverted from “as carbon” or “as propane” to “as VOC”).


Page 4 of 28

It should be noted that the VOC mass emission calculation based on a conversion ofMethod 25 or Method 25A data using Method 18 measurements of a specific list of oxygenatedorganic compounds may slightly bias the true total mass VOC emission rate compared to the useof a complete set of organic compound concentrations, including all non-oxygenatedhydrocarbons. The source is allowed, at its discretion and with the EPA’s approval, to performadditional sampling and analysis to quantify the concentrations of other hydrocarbon compounds,including non-oxygenated compounds, and use the overall average molecular weight for allquantified organic compounds in the calculations discussed below. Failure to conduct additionaltesting indicates that the source accepts the oxygenated organics weighted average molecularweight to carbon weight ratio as representative of the actual average molecular weight to carbonweight ratio of all organic compounds present in the emissions from the specific unit beingtested.

1.0 Scope and Applicability.

1.1 Analytes. The analytes in Table 1.1 must be measured from each source being tested. These compounds have been found to comprise the bulk of the identified VOC emitted fromsources at grain mills and ethanol production facilities.

Analyte CAS Number Interference-Free AnalyticalSensitivity

Total Organic Compounds NA M25A ~3 ppmC, M25 ~50 ppmC

Acetaldehyde 75070 ~ 1 ug/ml

Acetic Acid 64197 ~ 1 ug/ml

Ethanol 64175 ~ 1 ug/ml

Formaldehyde 50000 ~ 1 ug/ml

Formic Acid 64186 ~ 1 ug/ml

2-Furaldehyde 98011 ~ 1 ug/ml

Methanol 67561 ~ 1 ug/ml

1.2 Applicability. This protocol is applicable to determining the actual VOC massemission rates from sources at grain mills and ethanol production facilities.

1.3 Data Quality Objectives. The quality of the data needed is determined by the needs ofthe data user. If the test using this protocol is required as part of a regulatory process and if the


Page 5 of 28

tester follows and meets the performance criteria in the protocol, including all Method 18 spikerequirements, it is presumed that the MSP produces data of suitable quality to determinecompliance with that regulation. The performance criteria in the protocol are set at levels that anoperator properly using well designed equipment will consistently attain or exceed. However,because the protocol allows different options to comply with some of the performance criteria, itis the responsibility of the owner or operator of the emission unit, as the data provider, to identifythe specific requirements in the protocol that were followed and document that the protocol’sperformance criteria were met, or to identify deviations as an exception to the protocol. Theregulatory agency is considered the data user and, therefore, is entitled to make the finalassessment of data quality.

For the purpose of determining only the SF to be used in calculating VOC massemissions, the spike requirements of Method 18 may be replaced with an analytical spike setconsisting of one low concentration and one high concentration spike sample. These alternatespike samples shall be prepared in the field by spiking the first impinger of the sample collectiontrain and drawing a measured amount of hydrocarbon-free air through the impinger trainequivalent to the nominal sample volume. The spike samples shall be recovered and analyzedusing the same procedures as those used to recover and analyze the source samples.

2.0 Summary of Protocol. Total organic emissions are measured based on the carbon contentof the sample. The list of individual organic compounds that are present in significant quantitiesare measured individually by Method 18 (using the specific application described in Appendix B)and used to convert the total carbon based measurements to a true VOC mass.

3.0 Definitions. Use the definitions as specified in the following methods.

3.1 EPA Methods. These are methods found in 40 CFR Part 60, Appendix A, and 40CFR Part 51, Appendix M.

3.1.1 Method 25 — Determination Of Total Gaseous Non-methane Organic EmissionsAs Carbon

3.1.2 Method 25A — Determination Of Total Gaseous Organic Concentration Using aFlame Ionization Analyzer

3.1.3 Method 18 — Measurement Of Gaseous Organic Compound Emissions By Gas

Chromatography

3.1.4 Method 205 —Verification of Gas Dilution Systems for Field InstrumentCalibrations

3.1.5 Method 5 —Determination Of Particulate Emissions From Stationary Sources


Page 6 of 28

3.1.6 Method 1 — Sample And Velocity Traverses for Stationary Sources

3.1.7 Method 2 — Determination Of Stack Gas Velocity And Volumetric Flow Rate

3.1.8 Method 3A—Determination of Oxygen and Carbon Dioxide Concentrations inEmissions from Stationary Sources (Instrumental Analyzer Procedure).

3.1.9 Method 4—Determination of Moisture Content in Stack Gases.

3.1.10 Method 10 —Determination Of Carbon Monoxide Emissions From StationarySources

3.1.11 Method 10B —Determination Of Carbon Monoxide Emissions From StationarySources

4.0 Interference. Interference as specified in the methods in Section 3 and Appendix B.

5.0 Safety. Follow the safety precautions as specified in the methods in Section 3 andAppendix B. Note that some sources and some areas of grain processing and ethanol production facilitiesmay be fire or explosion hazards. Use appropriate caution and selection of sample collectionprocedures.

6.0 Equipment and Supplies. Equipment and supplies as specified in the methods in Section 3and Appendix B.

7.0 Reagents and Standards. Reagents and standards as specified in the methods in Section 3and Appendix B, with the following exception:

7.1 For Method 25A, obtain a calibration standard of all individual target analytes inSection 1.1 and/or other target analytes found in screening tests at significant levels (>5% of thetotal VOC). The standards shall be within the range of 25 % to 200 % of the expectedconcentration of the individual compound. These calibration standards will be used to developresponse factors for each individual compound. These gases shall meet the specifications ofSection 7.1 of Method 25A.

8.0 Sample Collection, Preservation, Storage and Transport.

8.1 Test Protocol (TP). The procedures in Appendix A, entitled “A Guide for StackTest Protocol Development and Submittal For VOC Emission Tests at Grain Processing andEthanol Production Facilities,” shall be used to assure consistency and adequacy. Failure tosubmit a complete TP could add cost and time due to postponements or additional submittals ofthe TP.


Page 7 of 28

8.2 Operating Conditions. For the entire period of its performance test, each affectedsource shall operate at 90% to 100% of its maximum achievable capacity or itsallowable/permitted capacity under representative conditions while maintaining safe and stableload conditions using the highest emitting fuel (normal power sources) and processing typicalmaterial resulting in normal product. Operational parameters shall be recorded at 15-minuteintervals during the test to substantiate the load. The inlet and outlet gas temperatures of thedryers, syrup addition feed rate and solids content, wet cake feed rate (e.g. tons/hour) shall berecorded during the test.

8.3 The samples shall be collected using the following parameters:

8.3.1 Sources Without Entrained Water Droplets or Aerosols. If the tester intends to useprocedures for sources that do not have entrained water droplets, the tester shall conduct a visualinspection and a saturation test of the exhaust gases immediately prior to testing to demonstratethe stack gas is not saturated. A saturation test consists of measuring the moisture content of theexhaust gases using Method 4 and comparing the measured moisture results to tabulated valuesfor moisture content at 100 % relative humidity at the average temperature of the stack gas. Ifthe measured moisture content exceeds the moisture content from the tabulated values, then thestack gas shall be considered to be saturated and to contain water droplets. If the stack gas doesnot contain water droplets or visible aerosols, collect the samples directly from the stack gasusing the procedures in Method 25 or Method 25A and Method 18 as described in Appendix B. Use appropriate caution and unheated sample trains when collecting samples from explosion orfire hazard rated sources regardless of aerosol or water droplet content.

The need for unheated sample trains may dictate the requirement for using Method 25 ifthe Method 25A sample train would be subject to sample condensation.

8.3.2 Sources That Contain Entrained Water Droplets. If the stack gas contains entrainedwater droplets, the sample shall be extracted directly from it using the isokinetic samplingprocedures described in Method 5 with the exception that the sample shall be drawn from asingle representative point, preferably near the center of the stack or duct. Use Method 1 todetermine the appropriate sampling location. The tester shall maintain the probe and filter of theMethod 5 sampling train at 250° F + 25° F. Between 20 and 30 dry standard cubic feet (dscf)shall be drawn through the Method 5 sampling train over a one-hour period for each of the threeruns.

Use two stainless steel compression fittings behind the filter in the heated filter box of theMethod 5 sampling train to withdraw the sample for the total organic compound quantificationtest (Method 25 or 25A) and for the individual organic compound analysis (Method 18 asdescribed in Appendix B). Place a valve between the Method 5 and the Method 25 or 25Asampling system, and between the Method 5 and the Method 18 sampling systems to isolate eachof the sampling systems for leak checks. The tester shall account for the amount of samplediverted to the total organic quantification test and to the Method 18 sampling trains when


Page 8 of 28

calculating the isokinetic sampling rates. Method 25 samples require ~5 dry standard liters (~0.2dscf) per sample, and the Method 18 train requires ~30 dry standard liters (~1.1 dscf) per sample. Method 25A analyzers have different flow requirements and must be determined individually.

8.4.3 All Sources. Measure stack gas velocity according to the procedures in Methods 2and 3 at the beginning and the end of each test. Measure the moisture content of the stack duringeach test according to the procedures in Method 4. If the moisture content of the sample streamis greater than 10% (or as otherwise specified for the specific FID used) and if the tester ismeasuring total organic compounds by Method 25A, the tester shall use the procedures inMethod 205 to dilute the sample to reduce the moisture content to within the linear and unbiasedoperating range of the FID. The tester shall conduct cyclonic flow tests prior to thecommencement of testing at all sampling locations. If cyclonic flow is determined, appropriatecorrections must be conducted.

8.4.4 Dryers and Combustion Sources. Measure the carbon monoxide content ofemissions from dryers and combustion sources using the procedures in Method 10 orMethod10B.

8.5 Sample Recovery.

8.5.1 If using Method 25 for the total organic compound quantification test, follow theprocedures in that method to recover the sample, store it and transport it to the laboratory.

8.5.2 Follow the recovery procedures in Method 18 as described in Appendix B with thefollowing exception: If the tester uses an empty impinger as the final impinger in the sampletrain to collect any carryover impinger solution due to high moisture content in the stack, thetester shall recover any liquid in the final impinger and treat it as part of the sample. The testermay combine this recovered liquid with the sample from the impinger immediately in front of thefinal impinger or may recover it in a separate container.

9.0 Quality Control. Follow the quality control procedures as specified in the methods inSection 3 and Appendix B.

10.0 Calibration and Standardization. Follow the procedures for calibration andstandardization as specified in the methods in Section 3 and Appendix B with the followingexceptions:

10.1 For Method 25A, the tester shall determine the response factor of the actualinstrument used for measuring the total organic compound concentration to each of theindividual compounds in Section 1.1 that comprise >5% of the identifiable VOC in the sample. The response factor shall be determined by the instrument’s response to the calibration gas usedduring the emissions test. The tester may determine the response factor in the laboratory, at thetest site prior to the testing, or in the laboratory within 45 days after the first day of the testing


Page 9 of 28

provided that the instrument has not been modified or repaired in the interim. The responsefactor shall not be acceptable if the instrument is modified, repaired or adjusted between the testdate and the date that the response factors are determined. After the tester has determined theresponse factor for an individual instrument, the tester may use this response factor for othertests on the same emission unit using the same instrument until the instrument is modified orrepaired.

Immediately prior to determining the response factors, the tester must introduce zero gasand high-level calibration gas at the calibration valve assembly. The analyzer output shall beadjusted to the appropriate levels, if necessary. The predicted response as carbon shall becalculated for the compound for which a response factor is being determined by multiplying theconcentration of the compound by the number of carbon atoms in each molecule of thecompound. Then, the tester shall introduce the calibration gas to the measurement system, recordthe analyzer response, and calculate the response factor using the equation in Section 12.7.

11.0 Analytical Procedure. Follow the analytical procedures as specified in the methods inSection 3 and Appendix B.

12.0 Calculations and Data Analysis. Follow the calculation and data analysis procedures asspecified in the methods in Section 3 and Appendix B with the following additions:

12.1 Scaling Factor, SF. Calculate the scaling factor using the following equation.

Equation 1

Where SF = Factor used to correct mass as carbon to “as VOC” or actual mass (expectedto be 1.9-2.6)N = Total number of compoundsMWi = Molecular weight of compound iMWCi = Molecular weight of carbon per mole of compound i MFCi = Mole fraction of carbon contributed by compound i

12.2 Mole Fraction of Carbon.


Page 10 of 28

Equation 2

Where mCi = Milligrams of carbon contributed by compound i in the Method 18 sample.

12.3 Mass of Carbon Contributed by Each Compound.

Equation 3

Where mi = Milligrams of compound i in the Method 18 sample

12.4 Actual Mass Concentration VOC in the Sample Gas. Calculate the actual massconcentration of VOC in the sample gas from the measured VOC concentration as carbon usingthe following equation.

Equation 4

Where ma = Actual mass concentration of VOC in the samplemc = Measured carbon mass concentration of VOC in the sample, mg/dscm.

12.5 Carbon Mass in the Sample Based on Method 25A Measurement. For Method 25A,calculate the carbon mass in the Method 25A measured sample using the following equation.

Equation 5

Wheremc = Organic concentration as carbon, ppmv from Method 25A.RFave = Weighted average response factor from Equation 6.

12.6 Average Response Factor for Method 25A. Calculate the weighted averageresponse factor, RFave, for Method 25A using the following equation.


Page 11 of 28

Equation 6

Where Ci = Concentration in ppm carbon of organic compound iRFi = Response Factor of organic compound i

12.7 Response Factor for Individual Compounds. Calculate the response factor forindividual organic, RFi, compounds using the following equation.

Equation 7

Where Cci = Concentration in ppmv carbon of organic compound i as certified by themanufacturer of the standardCmi = Measured concentration in ppm carbon of organic compound i fromSection 10.1


Page 12 of 28

Appendix A

A Guide forStack Test Protocol Development and Submittal

For VOC Emission Tests at Grain Processing and Ethanol Production Facilities

PROTOCOL DEVELOPMENT

A detailed protocol, describing all test equipment, procedures, and quality assurance (QA)measures to be utilized, will help ensure that a complete and representative stack test isperformed. The protocol must be specific for the test, facility, operating conditions, andparameters to be measured. Adherence to the protocol should eliminate unnecessarydelays and costs in the performance of the test, whether the work is done in-house or by aconsultant.

The term "tester" will be used to refer to the individual(s) performing the emission test,whether in-house or a consultant. The tester should make at least one on-site inspectionof the emission point(s), testing ports, stack access and other parameters in order toprepare the protocol.

The following provides specific guidance pertinent to the major elements of the stack testprotocol.

1. Project Description

Provides a general description of the project. This should include sufficient detail toallow those individuals responsible for review and approval to perform their tasks. Where appropriate, the following shall be included:

a. Intended end use of the acquired data.

b. Dates anticipated for the beginning and the completion of testing.

c. Description of plant processes and control equipment, including flow diagramsand permitted, or maximum achievable, process rates.

d. Description of plant operating conditions, including but not limited to productionrate, fuel rate, process data (including relevant temperatures and/or flow rates),and pollution control operational data.


Page 13 of 28

e. Proposed operation during the stack test . Unless approved or specified by U.S.EPA or the applicable state agency, the test will be deemed unacceptable if therelevant process(es) are operated at less than 90% of maximum capacity.

f. List of operating and emission parameters to be measured during the test, typicaloperating ranges for these parameters, and the maximum ranges for theseparameters.

2. Project Organization and Responsibility

Include a table or chart showing the project organization and line of authority. List thekey individuals, including the Quality Assurance Officer (QAO), who are responsible forensuring the collection of valid measurement data and the routine assessment ofmeasurement systems for precision and accuracy.

3. QA Objectives for Measuring Data

All measurements must be made to ensure that results are representative of the normal, orpermitted, maximum operating conditions of the facility. Data quality objectives will bedetermined for each measurement and compared with the requirements for the specificproject. This will ensure that the data collected will be appropriate for their intended use.

4. Sampling Procedure

For each major measurement parameter, provide a description of the sampling proceduresto be used. Officially approved EPA procedures and Reference Methods should be usedwhere applicable. The protocol should include the following:

a. A stack diagram showing test ports, their distances from upstream anddownstream disturbances, the stack diameter, planned sampling equipment andmonitoring locations.

b. The proposed method for the determination of the presence and quantification ofcyclonic flow.

c. The proposed number of sample flow measurement points and the total samplevolume.

d. A detailed description of all sampling, sample recovery, and analytical procedures. In the case of non-standard procedures or modifications to standard procedures,the entire procedure should be described with justifications and necessary data forbackup. Options offered by the Reference Method should be selected andjustified.


Page 14 of 28

e. Any special conditions for the preparation of the sampling equipment andcontainers to avoid sample contamination.

f. Samples of forms to be used to record sample history, sampling conditions, andplant operating conditions.

g. The methodology for measurement of plant and pollution control device operatingconditions.

h. If more than one sampling train is to be used, a detailed description of the relevantsequencing and logistics.

i. If Continuous Emission Monitors (CEMs) are to be used, a detailed description ofthe operating and data logging procedures.

5. Sampling Procedures for Ethanol Production Facility Dryers

The protocol for the emission test should include the following test methods to accuratelycharacterize the VOC emissions from dryers:

Test Methods -

USEPA Method 1: Sampling Location and Cyclonic Flow DeterminationUSEPA Method 2: Stack Gas Velocity and Volumetric Flow RateUSEPA Method 3: Stack Gas Molecular WeightUSEPA Method 4: Stack Gas Moisture ContentUSEPA Method 18: Gas Chromatography

The preferred application of Method 18 based on similar sources is the NCASIMethod CI/SG/PULP-94.02: Chilled Impinger/Silica Gel Tube Test Method atPulp Mill Sources for Methanol, Acetone, Acetaldehyde, Methyl Ethyl Ketoneand Formaldehyde

USEPA Method 25: Determination of Total Gaseous Non-Methane Emissions asCarbon

USEPA Method 25A: Determination Of Total Gaseous Organic Concentration UsingA Flame Ionization Analyzer

Location - Sampling shall be performed at the exit of each stack. If the stack has acontrol device for VOC emissions, sampling shall occur before and after the controldevice where applicable and consistent with the Project Description listed above.

Isokinetics - Sample shall be drawn isokinetically from a single representative point forall methods in any stack that contains uncombined water or organic aerosols.


Page 15 of 28

Detection Limits - The limits of detection for each targeted compound and for total VOCshall be calculated in Kg/hr and/or lbs/hr.

6. Sample Custody

Sample custody is a part of any good laboratory or field operation. At a minimum, thefollowing sample custody procedures shall be addressed in the protocol:

a. Documentation of procedures for preparation of reagents or supplies that becomean integral part of the sample (e.g., filters and absorbing reagents).

b. Procedures and forms for recording the exact location and specific considerationsassociated with sample acquisitions. As samples are transferred betweenindividuals, the individuals should sign and date their relinquishing of, or receiptof, the samples on the Chain of Custody form.

c. Prepared sample labels containing all information necessary for effective sampletracking. Labels or custody seals should cover the sample container cap such thatit would be evident if the sample was opened by a person other than the laboratoryanalyst.

7. Calibration Procedures and Frequency

Include calibration procedures and information for each major measurement device,including coefficients, by reference to a standard method or by providing writtendescription. Provide the frequency planned for recalibration during the test and a list ofall calibration standards, including their source and traceability. Equipment to becalibrated would include, for example, dry gas meters, orifice meters, pitot tubes,thermometers/thermocouples, nozzles, flow meters as well as all process parametermonitors. Also include a detailed description of spike preparation procedures.

8. Documentation

Include sample copies of all data log sheets and examples of any calculations that will beperformed on the raw data. Note: copies of all raw data sheets, including manually andautomatically recorded data (strip charts and data logger or computer printouts) will besubmitted with the test report and copies must be available at the end of the day's testing.


Page 16 of 28

Appendix BMethod 18 for Oxygenated Organics Other Than Formaldehyde

Introduction.This appendix describes a specific application of the general Method 18 procedures to measurethe individual oxygenated organic compounds other than formaldehyde that are required by theMidwest Scaling Protocol. Formaldehyde is collected in the same Method 18 sample, but isanalyzed by a separate procedure found in Appendix C. Both this specific application ofMethod 18 and the formaldehyde procedure in Appendix C were developed by the NCASI andvalidated for their use at pulp mills. The NCASI identifies the procedure as NCASI MethodCI/SG/PULP-94.02, Chilled Impinger/Silica Gel Tube Test Method at Pulp Mill Sources forMethanol, Acetone, Acetaldehyde, Methyl Ethyl Ketone and Formaldehyde.

AcknowledgmentThis method was prepared by Dr. MaryAnn Gunshefski, Senior Research Scientist, andWard Dickens, Research Associate, at the NCASI Southern Regional Center. Otherassistance was provided by Terry Bousquet, Senior Research Scientist, with the NCASIWest Coast Regional Center.

This specific application follows the general Method 18 procedure with the following additionsto Method 18 taken directly from the NCASI Method CI/SG/PULP-94.02.

1.0 Scope and Application. Same as Method 18 with the following addition: Stability - The stability of acetaldehyde in the impinger catch was found to be 10 days, withrefrigeration at approximately 4°C. The stability of acetone, methyl ethyl ketone, and methanolwas found to be 21 days, with refrigeration at approximately 4°C. The stability of acetaldehyde,acetone, methyl ethyl ketone, and methanol on the silica gel sorbent tubes was found to beapproximately 10 days, with refrigeration at approximately 4°C. Once desorbed in 3% n-propanol, these same compounds are stable for up to 21 days, with refrigeration at approximately 4°C.

2.0 Summary of Method. Same as Method 18 with the following addition:This method involves collection of an air sample by drawing it through a midget impinger, whichis filled with water, and then through two 2-section silica gel sorbent tubes. The impinger is keptin an ice water bath during sampling to enhance collection efficiency. The impinger catch isanalyzed for methanol, acetaldehyde, ethanol, formic acid, acetic acid, 2-furaldehyde, by directinjection into a gas chromatograph equipped with a flame ionization detector (GC/FID). Thesilica gel sorbent is desorbed with a 3% (v/v) solution of n-propanol. The desorbate is injecteddirectly into the GC/FID for analysis of methanol, acetaldehyde, ethanol, formic acid, acetic acid,and 2-furaldehyde. Alternative GC procedures may be used with prior approval.

3.0 Definitions. Same as Method 18.


Page 17 of 28

4.0 Interferences. Same as Method 18 with the following addition: method interferences maybe caused by contaminants in solvents, reagents, glassware and other sample processinghardware. Clean all glassware by detergent washing with hot water and rinsing with tap water. The glassware should then be drained dry and baked at greater than 100°C for over 2 hours.

5.0 Safety. Same as Method 18.

6.0 Equipment and Supplies. Same as Method 18 with the following additions:

6.1.1 Sampling apparatus. A diagram of the sampling train is shown in Figure 1 (see below).

6.1.1.1 Probe/sampling line. The probe is made from Teflon tubing or stainless steel, which isthen attached to the first impinger.

6.1.1.2 Impinger train. Three 30 mL capacity midget impingers are connected in series to thesampling probe. The impingers should have regular tapered stems. All impinger trainconnectors should be glass and/or Teflon.

6.1.1.3 Sorbent tubes Two 2-section silica gel sorbent tubes (SKC #226-15 GWS) are placed inline after the impingers.

6.1.1.4 Rotameter. A 1000 mL/min capacity rotameter should be placed in line after the silicagel sorbent tubes for a visual flow check during sampling and leak checking. The rotameter isnot used to determine the actual flow rate through the impingers.

6.1.1.5 Critical orifice. A 400 ± 50 mL/min critical orifice should be used for flow control.

6.1.1.6 Vacuum pump - The critical orifice is followed by a pump capable of providing avacuum of about 18 inches of Hg. Pump capacity should be sufficient to obtain and maintaincritical conditions at the orifice.

6.1.1.7 Pressure gauges. One pressure gauge is placed before the critical orifice, and onepressure gauge is placed before the pump, and both are used when leak checking the sampletrain. The pressure gauge downstream of the critical orifice provides a check for critical flowconditions at the orifice.

6.1.1.8 On/off valve. An on/off valve is placed between the critical orifice and the secondpressure gauge, and is used when leak checking the sample train.

6.1.1.9 Flowmeter. A bubble tube flowmeter is used to measure flow at the sampling line tipprior to and after sampling. Alternatively, a dry gas meter may be used.


Page 18 of 28

Alternative Sampling Apparatus. An equivalent sample gas collection system may be proposedby the tester (e.g., use of a volumetrically calibrated evacuated vessel and controller consisting ofa needle valve and rotameter along with pre- and post-tank temperature and absolute pressuremeasurements, or use of a Volatile Organic Sampling Train [VOST] console with its low-flowcalibrated dry gas meter.)

6.1.1.10 Thermometer - An accurate thermometer is used to measure ambient temperature.

6.1.1.11 Barometer - A barometer is used to measure barometric pressure.

6.1.1.12 Sample storage bottles. Glass (i.e., 40 mL VOA vials) or polyethylene bottles can beused to store the impinger catch sample after stack sampling is complete.

6.1.2 GC/FID analysis apparatus

6.1.2.1 Laboratory glassware. Volumetric pipets, volumetric flasks, autosampler vials, syringes,and cuvettes necessary for standards preparation and analysis.

6.1.2.2 NCASI-recommended gas chromatography system. Gas chromatography/flameionization detector system, complete with a temperature-programmable gas chromatographsuitable for splitless injection and all required accessories including syringes, analytical columnsand gases. Note that we suspect systems with EPC are not designed to handle aqueousinjections, and as a result the FID flame may begin to go out during the runs. This could be dueto the water which builds up in the GC system after several injections on any type of GC. Bakeouts are necessary for any type of GC system, but more frequent bakeouts of a system withEPC may need to be performed.

6.1.2.3 Column - 30 m x 0.53 mm x 1 :m bonded phase DB-WAX fused silica capillary column(J&W Scientific or equivalent); 30 m x 0.32 mm x 0.25 :m bonded phase DB-WAX fused silicacapillary column (J&W Scientific or equivalent); 30 m x 0.53 mm x 3 :m bonded phase DB-624fused silica capillary column (J&W Scientific or equivalent); or other column shown to becapable of separating methanol, acetone, acetaldehyde, methyl ethyl ketone and n-propanol.

6.1.2.4 GC detector - Flame ionization detector with appropriate data system.

7.0 Reagents and Standards

7.1 Water - Deionized water is to be used as the impinger collection liquid, and in thepreparation of all standard and spike solutions.

7.2 Pure compounds - Reagent grade methanol, acetaldehyde, ethanol, formic acid, acetic acid,2-furaldehyde, formaldehyde solution in water (stabilized with methanol) for preparation of


Page 19 of 28

standard and spike solutions. Be sure to account for the methanol in the formaldehyde solutionwhen calculating spike concentrations.

7.3 GC/FID calibration primary stock solution - Prepare stock solution by diluting 0.126 mL ofpure methanol, 0.128 mL of pure acetaldehyde, 0.073 mL of glacial acetic acid, 0.127 mL of pureethanol, 0.082 mL of pure formic acid, 0.086 mL of pure 2-furaldehyde, and 0.270 ml of 37%formaldehyde solution in 100 ml volumetric flask with DI water (1000 mg/L plus the methanol inthe formaldehyde solution).

7.4 GC/FID calibration and matrix spike solutions - Prepare standard solutions by serialdilutions of the stock solution. The recommended calibration range is 0.5 to 1000 mg/L. It hasbeen found that the linear range can be extended up to 10,000 mg/L. Prepare matrix spikesolutions by calculating the concentration of analytes desired and diluting the primary stocksolution.

7.5 GC/FID internal standard primary spiking solution (if used) - Prepare primary stock solutionby adding 0.312 mL cyclohexanol and diluting to 100 mL with DI water in a 100 mL volumetricflask (3 mg/mL cyclohexanol). Another internal standard material could be used if it isdemonstrated that it does not interfere with the analyte peaks in the chromatogram.

7.6 n-propanol - Prepare a 3% (v/v) n-propanol/water solution for desorption of the analytesfrom the silica gel sorbent tubes.

8.0 Sample Collection, Preservation, Storage, and Transport. Same as Method 18, Sections8.2.4, 8.3, and 8.4.3 with the following additions:

8.1.1 Sample bottle preparation - Determine the number of sample bottles required for thesampling trip. Weigh each bottle and record the pre-sampling weight on the bottle.

8.1.2 Sampling.

8.1.2.1 Measure and record ambient temperature and barometric pressure.

8.1.2.2 Preparation of collection train. Measure 20 mL of DI water into each of the first andsecond impingers and assemble the sampling train.

8.1.2.3 Leak and flow check procedure. Make sure that the on/off valve is in the on position,close the valve to the M-5 train and turn on pump to draw a vacuum. When the vacuum readingis approximately 25 inches of Hg, turn the pump on/off valve to the off position, then record timeand pressure reading on first pressure gauge. A leak is indicated by a flow of bubbles in theimpinger, liquid being drawn into the stem of the impinger or a loss of vacuum. If a leak ispresent, tighten fittings, connections and impingers, and restart the leak check procedure. After 2minutes, record the pressure reading on the first pressure gauge again. The leakage rate should


Page 20 of 28

not be in excess of 1 inch Hg (vacuum) in 2 minutes. Slowly and carefully open the valve to theM-5 train, and turn the on/off valve back to the on position. If using the critical orificeprocedure, check the flow rate at the probe inlet with a bubble flowmeter. The flow rate shouldbe comparable to the flow rate of the critical orifice with the impingers off-line. Record fivemeasurements of the flow rate and turn off the pump.

8.1.2.4 Sample collection - Insert the probe into the stack and secure it. Start the pump,recording the time and the flow reading on the rotameter. End the sampling after 60 minutes.Record the time and remove the tubing from the vent. Recheck the sample flow rate at the probeinlet and turn off the pump. If the flow rate has changed significantly, redo sampling with freshcapture water. A slight variation (< 5%) in flow can be averaged. With the probe inlet end of theline elevated above the impinges, add about 5 mL of water into the inlet tip to rinse the line intothe first impinger.

8.1.3 Sample recovery - Transfer the contents of the impingers into an appropriately labeled andpre-weighed sample storage bottle. The contents of both impingers can be combined into onebottle. If a large amount of water was collected in the dropout impinger, two bottles can be used. Remove the silica gel tubes from the sampling train, cap ends (tape caps on if necessary), andlabel. Store both impinger and sorbent tube samples in a cooler with ice until they can be storedin a laboratory refrigerator at approximately 4°C.

9.0 Quality Control. Same as Method 18 with the following exception: for the purpose ofdetermining only the Scaling Factor to be used in calculating VOC mass emissions, the spikerequirements of Method 18 may be replaced with an analytical spike set consisting of one lowconcentration and one high concentration spike sample. These alternate spike samples shall beprepared in the field by spiking the first impinger of the sample collection train and drawing ameasured amount of filtered air through the impinger train equivalent to the nominal samplevolume. The spike samples shall be recovered and analyzed using the same procedures as thoseused to recover and analyze the source samples.

10.0 Calibration and Standardization. Obtain calibration standards for each target compoundto be analyzed. Prepare or obtain enough calibration standards so that there are three differentconcentrations of each organic compound expected to be measured in the source sample. Foreach organic compound, select those concentrations that bracket the concentrations expected inthe source samples. A calibration standard may contain more than one organic compound. Prepare or obtain standards in the same solvent used for the sample extraction procedure. Verifythe stability of all standards for the time periods they are used. Analyze each standard intriplicate.

10.1 GC/FID analysis of calibration standards.

10.1.1 Internal standard calibration.


Page 21 of 28

10.1.1.1 Inject 1 :L of a methanol, acetaldehyde, ethanol, formic acid, acetic acid, and 2-furaldehyde calibration solution containing the internal standard and determine the retention timeof the analytes relative to the internal standard. Each analyst should optimize the temperatureprogram or instrument conditions, as necessary, to establish distinct separate peaks.

10.1.1.2 Calculate the relative response factor for the analytes (RRFM) using Equation 1 (section12.1, below). If the average of the relative response factor for the analytes is constant, i.e.,exhibits a coefficient of variation less than 20%, the calibration is acceptable and the averageRRFM can be used in all subsequent calculations; otherwise, the calibration curve solutions mustbe reanalyzed and reevaluated. It may be necessary to perform instrument maintenance prior toreanalysis. If reanalysis also fails to produce a linear curve, new calibration standards must beprepared and analyzed.

10.1.1.3 Analyze and calculate the relative response factor of a midrange calibration standarddaily, prior to each sample set, using Equation 2 (section 12.2, below) to verify the calibration. The relative response factors must be within an acceptable range. If they are not, either prepare anew standard or perform instrument maintenance. If necessary, re-calibrate the instrument.

10.2.2 External standard calibration

10.2.2.1 Inject 1 :L of a methanol, acetaldehyde, ethanol, formic acid, acetic acid, and 2-furaldehyde calibration solution and determine the retention time of each analyte. Each analystshould optimize the temperature program or instrument conditions, as necessary, to establishdistinct separate peaks.

10.2.2.2 Measure and plot the response of each analyte vs. concentration. If the correlationcoefficient of the graph is greater than 0.99, the calibration is acceptable and the equation of thecurve can be used in all subsequent calculations; otherwise, the calibration curve solutions mustbe reanalyzed and reevaluated. It may be necessary to perform instrument maintenance prior toreanalysis. If reanalysis also fails to produce a correlated curve, new calibration standards mustbe prepared and analyzed.

10.2.2.3 Analyze and calculate the concentration of a mid-range calibration standard daily, priorto each sample set, to verify the calibration. The recovery should be between 70 and 130%. If itis not, either prepare a new standard or perform instrument maintenance. If necessary, re-calibrate the instrument.

10.3 Analytical range and minimum calibration level

10.3.1 Demonstrate that the calibration curve is acceptable (relative response factors exhibit acoefficient of variation less than 20%, or correlation coefficient greater than 0.99) throughout therange of the calibration curve.


Page 22 of 28

10.3.2 Demonstrate that the analytes are detectable at the minimum levels using the lowest levelcalibration curve solution.

11.0 Analytical Procedures.

11.1 Preparation of impinger samples. Remove bottles from refrigerator. Weigh the samplebottles and record weights on the bottle. Transcribe initial and final bottle weight to sample fielddata sheet. Bottles do not need to be at room temperature before weighing. Remove an aliquotof sample and place in the sampler vial, add 10 :L of internal standard solution (if using internalstandard calibration curve), and cap vial.

11.2 Preparation of sorbent tube samples. Remove sorbent tubes from refrigerator. Remove endcaps and score glass to remove the silica gel from one section. All sections of the silica gel tubescan be combined and analyzed together. This is considered the “back half” of the samplecollection train. Pour into a 4.0 mL screw-capped vial and add 3.0 mL of a 3% (v/v) n-propanol/water desorption solution. Allow to sit for 30 minutes, with occasional light shaking. Vigorous shaking causes the silica gel particles to adhere to the cap and walls of the vial. Remove an aliquot of the desorption solution and place in an autosampler vial. Add 10 :L ofinternal standard solution (if using internal standard calibration curve) and cap vial.

11.3 GC/FID analysis. Analysis is performed by direct aqueous injection into the GC/FID.Representative conditions for the GC/FID analysis are given in Tables 1, 2 and 3 (section 18,below). Other chromatographic columns and conditions may be used if it has been establishedthat the compounds are separated and quality control parameters are met. Once the GC/FIDsystem is optimized for analytical separation and sensitivity, the sample operating conditionsmust be used to analyze all samples, blanks, calibration standards and quality assurance samples. Note that constant injections of aqueous samples can cause water to build up in the system. Thiswill cause the retention times to shift, and the peaks to broaden. It is recommended that afterapproximately 50 injections a bakeout of the system be performed. This should consist ofheating the injector to 250°C, the oven to over 200°C and the detector to 275°C for at leastseveral hours.

12.0 Data Analysis and Calculations. Same as Method 18 Sections 12.7 -12.9 with thefollowing additions:

12.1 Relative Response Factor. Calculate the relative response factor (RRFM using the followingequation.

Equation 1

Where:AM = area of analyte peak


Page 23 of 28

AIS = area of internal standard peakCM = concentration of analyte injected CIS = concentration of internal standard injected

12.2 Calibration Verification. Calculate the concentration of the midrange standard using thefollowing equation.

Equation 2

Where:AM = Area of the analyte peakCIS = Concentration of the internal standard (mg/L)AIS = Area of the internal standard peakRRFM = Relative response factor of analyte

13.0 Method Performance. Same as Method 18.

14.0 Pollution Prevention. [Reserved]

15.0 Waste Management. [Reserved]

16.0 Alternative Procedures. [Reserved]

17.0 References. Same as Method 18 with the following addition:17.1 National Council for Air and Stream Improvement, Inc. (NCASI). Methods Manual -NCASI Method CI/SG/PULP-94.02 Chilled Impinger/silica Gel Tube Test Method at PulpMill Sources for Methanol, Acetone, Acetaldehyde, Methyl Ethyl Ketone and Formaldehyde. National Council for Air and Stream Improvement, Inc.. Research Triangle Park, N.C. 1998.

18.0 Tables, Diagrams, Flowcharts, and Validation Data.

Table 1: GC/FID Operating Conditions for Methanol, Acetaldehyde, Acetone and MethylEthyl Ketone Analysis-DB-WAX ColumnInjection: DirectInjector Temperature: 150°CInjection Volume: 1 :LFID Detector Temperature: 250°CCarrier Gas: HeliumColumn: DB-WAX, 30 m x 0.53 mm id x 1 micronfused silica capillary columnTemperature Program °C:


Page 24 of 28

Initial: 18vC for 8 minRamp 1: 3°C/min to 20°C for 2 minutesRamp 2: 50°C/min to 220°CRamp 3: [deliberately blank]Final Hold Time: 5 minutesRetention Time Order: acetaldehyde, acetone, methyl ethyl ketone,methanol, n-propanol, cyclohexanol

Table 2: GC/FID Operating Conditions for Methanol, Acetaldehyde, Acetone and MethylEthyl Ketone Analysis-DB-WAX ColumnInjection: DirectInjector Temperature: 170°CInjection Volume: 1 :LFID Detector Temperature: 275°CCarrier Gas: HeliumColumn: DB-WAX, 30 m x 0.32 mm id x 0.25 micronfused silica capillary columnTemperature Program °C:Initial: 0°C for 3 minRamp 1: 5°C/min to 50°C for 4 minutesRamp 2: 70°C/min to 100°C for 10 minRamp 3: 70°C/min to 200°CFinal Hold Time: 4 minutesRetention Time Order: acetaldehyde, acetone, methyl ethyl ketone,methanol, n-propanol, cyclohexanol

Table 3: GC/FID Operating Conditions for Methanol, Acetaldehyde, Acetone and MethylEthyl Ketone Analysis-DB-624 ColumnInjection: DirectInjector Temperature: 170°CInjection Volume: 1 :LFID Detector Temperature: 275°CCarrier Gas: HeliumColumn: DB-624, 30 m x 0.53 mm id x 3 micronfused silica capillary columnTemperature Program °C:Initial: 0°C for 3 minRamp 1: 5°C/min to 50°C for 0 minutesRamp 2: 70°C/min to 105°C for 17 minRamp 3: 70°C/min to 220°CFinal Hold Time: 3 minutesRetention Time Order: acetaldehyde, methanol, acetone, n-propanol,methyl ethyl ketone, cyclohexanol


Page 25 of 28


Page 26 of 28

Appendix CAnalysis of Method 18 Samples for Formaldehyde

1.0 Scope and Application. Same as Appendix B with the following addition: The stability of formaldehyde was found to be 21 days, with refrigeration at approximately 4°C.

2.0 Summary of Method. This method contains procedures for analyzing the samples collected by the Method 18procedure described in Appendix B for formaldehyde. To analyze for formaldehyde, theacetylacetone derivatization/spectrophotometric analysis method is used on an aliquot of theimpinger solution collected according to Appendix B.

3.0 Definitions. Same as Appendix B.

4. Interferences. Same as Appendix B with the following addition:

Interferences with the formaldehyde analysis can be caused by the presence of sulfur compounds(i.e. SO2) in the source gas.

5.0 Safety. Same as Appendix B.

6.0 Equipment and Supplies. Same as Appendix B with the following addition:

6.1. Formaldehyde analysis apparatus

6.1.1 Spectrophotometer - A spectrophotometer capable ofmeasuring absorbance at 412 nm.

7.0 Reagents and Standards.

7.1 Water. Deionized water is to be used as the impinger collection liquid, and in thepreparation of all standard and spike solutions.

7.2 Pure compound. Reagent grade 37% formaldehyde solution (formalin) for preparation ofstandard and spike solutions.

7.3 Acetylacetone reagent. Prepare by dissolving 15.4 g of ammonium acetate in about 50 mLof DI water in a 100 mL volumetric flask. Add 0.20 mL of acetylacetone to this solution, alongwith 0.30 mL of glacial acetic acid. Mix thoroughly and dilute to 100 mL with DI water. Storereagent in a brown glass bottle in the refrigerator. Reagent is stable for at least two weeks.

7.4 Formaldehyde analysis primary stock solution. Prepare stock solution by diluting 2.7 mL offormalin in a 1000 mL volumetric flask with DI water (1000 mg/L formaldehyde).


Page 27 of 28

7.4.1 Formaldehyde analysis calibration standard solution. Prepare standard solution by diluting1.0 mL of primary stock solution in a 100 mL volumetric flask with DI water (10 mg/Lformaldehyde).

8.0 Sample Collection, Preservation, Storage, and Transport. The sample is collectedaccording to the procedures in Appendix B.

9.0 Quality Control. Each field sampling program or laboratory that uses this method isrequired to operate a formal quality assurance program. Laboratory or field performance iscompared to established criteria to determine if the results of analyses meet the performancecriteria of the method.

9.1 Field blank samples. A field blank sample of water must be prepared to assure that the waterbeing used in the impingers is not contaminated. It is made in the field by filling a 40 mLVOA vial or polyethylene bottle with the same water being used to fill the impingers.

9.2 Field spike sample. A field spike sample should be prepared by spiking the impinger with aknown amount of analyte before sampling. The spike solution described in Appendix A shouldbe used for this purpose. After the impinger is spiked, a sample bottlecontaining DI water should also be spiked. This provides a check of the spiking solution andspiking procedure. The impinger spiking may be done on a duplicate sampling train ifthe equipment is available or may be done during a normal sampling run. This type of spiking isperformed when a check of the complete sampling procedure, sample storage and sampleanalysis is desired.

9.3 Laboratory blank sample. A laboratory blank sample should be analyzed with each batch ofsamples. A batch is considered no more than 10 samples of similar matrix type.

9.4 Laboratory duplicates. A replicate injection of one sample in the analytical batch should beperformed. The results of the duplicate analysis should be within 10% of the mean of theoriginal and duplicate sample analysis.

9.5 Laboratory matrix spike samples. A laboratory matrix spike sample may be prepared witheach group of similar matrix type. Using the mean concentration determined by the replicateanalyses or the background level determined from a single measurement, determine the spikinglevel which will give one to four times the background. If the background sample doesnot have detectable levels of analytes, spike the sample at approximately five times the lowestcalibration level of the instrument. Spike the sample with the determined amount ofthe calibration standard/matrix spike solution and proceed to analyze the sample in the normalmanner. The results can be considered acceptable if the calculated spike recovery is 70 to130%. In cases where multiple analytes are present, the analyte with the highest concentrationshould govern the acceptance criteria.


Page 28 of 28

10.0 Calibration and Standardization.

10.1 Formaldehyde analysis calibration solutions. A series of calibration standards are madefrom the standard solution (Section 7.1.4.1) by adding 0, 0.1, 0.2, 0.4, 1.0 and 1.5 mL of thestandard solution to individual screw-capped vials. The volume in each vial is adjusted to 2.0mL with DI water. This corresponds to 0, 0.5, 1, 2, 5 and 7.5 mg/L calibration solutions. Toeach vial, 2.0 mL of the acetylacetone reagent is added, and the procedure described in Section11.1 is then followed.

11.0 Analytical Procedures.

11.1 Formaldehyde sample analysis. Remove a 2.0 mL aliquot of the impinger sample andtransfer to a screw-capped vial. Add 2.0 mL of the acetylacetone reagent and mix thoroughly.Place vial in a water bath at 60/C for 10 minutes. Allow vials to cool to room temperature. Transfer the solution to a cuvette and measure the absorbance at 412 nm. If the sample solutionconcentration is above the calibration curve, dilute original sample and repeat entire procedure. Do not dilute colored (derivitized) samples.12.0 Data Analysis and Calculations.

13.0 Method Performance. [Reserved]

14.0 Pollution Prevention. [Reserved]

15.0 Waste Management. [Reserved]

16.0 Alternative Procedures. [Reserved]

17.0 References. Same as Appendix B.

18.0 Tables, Diagrams, Flowcharts, and Validation Data. [Reserved]

Attachment #10

Ms. Mary A. GadeSonnenschein Nath and Rosenthal LLP8000 Sears Tower238 South Wacker DriveChicago, IL 60606

Dear Ms. Gade:

Thank you for your letter of December 10, 2003, to Ms. Walker Smith, Director of Officeof Regulatory Enforcement (ORE), concerning the use of volatile organic compound (VOC)measurement methods for the corn wet milling industry. You requested the EnvironmentalProtection Agency’s (EPA) position on the status of the draft Midwest Scaling Method (whichwe now refer to as the Midwest Scaling Protocol, as described below), and also a statement onEPA’s policy regarding compliance with permit emissions limits that are based on EPAreference methods and the use of promulgated EPA methods specified in permits.

First and foremost, to comply with the requirements of the New Source Review (NSR)program or the Title V operating permitting program under the Clean Air Act, theowner/operator of an emissions source must calculate the total emissions of criteria andhazardous air pollutants to determine the applicability of these programs. It is the responsibilityof the owner/operator of the source to do so and to use a protocol that accurately measures themass emissions. With regard to VOC’s, emissions must be calculated on a total VOC mass basis(“as VOC basis”), not on the basis of a surrogate such as “mass as carbon” or “mass as propane”[see definition of VOC at 40 CFR 51.100(s)]. One cannot determine the applicability of aregulation under NSR and Title V unless the emissions are reported as VOC(s). In addition, forthe purposes of reporting an annual total VOC emissions to the State or local agencies and theEPA, the total VOC mass must be identified.

The test methods in 40 CFR 60, Appendix A for measuring VOC emissions (e.g., Method 25/25A) do not directly address the issue of reporting VOC emissions “as VOC.”Furthermore, these methods must be modified or an additional method must be added to measurethe actual mass of VOC’s emitted. Such modifications or alternative methods are allowed asspecified in the excerpt of the regulation below:

Minor changes in the test methods should not necessarily affect the validity of the results

2

1 This procedure, which we originally called the “Midwest Scaling Method,” is now referred toas the Midwest Scaling Protocol (MSP). To avoid confusion some have expressed regarding the term“method.”

and it is recognized that alternative and equivalent methods exist. Section 60.8 providesauthority for the Administrator to specify or approve (1) equivalent methods, (2) alternative methods, and (3) minor changes in the methodology of the test methods.

We developed the draft Midwest Scaling Protocol (MSP)1 to help members of the ethanolproducing industry determine their VOC mass emissions. The draft MSP is a generallyacceptable protocol based on data and experience from the ethanol producing industry andcontains the steps necessary to convert VOC emissions measured using Method 25 or 25A to “asVOC” mass emissions. It is comprised largely of the EPA methods with appropriatemodifications in accordance with 40 CFR Part 60, Appendix A. The draft MSP is currentlyunder public review and comment. It is intended to be advisory in nature. Owner/operators ofsources for which the draft MSP is applicable may propose to use it when conducting a test forcompliance and applicability determinations. Assuming a successful test, EPA will generallyaccept that the emissions are adequately quantified for regulatory purposes.

There might be other procedures, including scaling methods, that can be used to quantifythe total mass of VOCs emitted from your industry that could also be approved by the EPA. Oneexample might be Test Method 320, Measurement of Vapor Phase Organic and InorganicEmissions by Extractive Fourier Transform Infrared Spectroscopy (FTIR) (see 40 CFR 63,Appendix A). This method uses FTIR to measure certain organic compounds on an individualbasis. The individually measured organic compounds can then be summed to produce the totalVOC mass, assuming that each VOC in the emissions stream can be measured with FTIR withinadequate limits of detection. Similarly, EPA Method 18 (40 CFR 60, Appendix A), in which gaschromatography is used to isolate individual organic compounds prior to measuring them, can beused to measure total VOCs if the quality assurance steps are sufficient to quantitatively measureall of the individual organic compounds. As you are aware, we are currently working with theCorn Refiners Association (CRA) to investigate the feasibility of a method for that industry. Aswe have suggested with the MSP, a facility may also propose to modify existing test methods ordevelop new test methods as alternatives to existing test methods. If a facility submits analternative test method request to your office, you should forward the request to the Office of AirQuality Planning and Standards for review and approval.

If the owner/operator of a source has used only Method 25 or 25A to determine theapplicability of NSR or Title V without appropriate modifications or use of an additionalapproved method to determine the “as VOC” mass emissions, then the VOC emissions may besubstantially underestimated and the owner/operator may not be in compliance with thoseprograms. Owner/operators who are in this situation should estimate their “as VOC” massemissions as soon as possible. To the degree that an owner/operator identifies that he or she maynot be in compliance with NSR and/or Title V, then contact should be made with EPA’sappropriate Regional Office or ORE to resolve any such issues. As of the date of this letter,

3

owners/operators who disclose this information expeditiously will have the same opportunitiesas are described in the recent settlements with the ethanol producing industry (see for example,www.epa.gov/compliance/civil/programs/caa/ethanol).

Regarding the existing emission limits in permits related to Federal or State and localstandards, as noted in 51.100(s)(2), it is appropriate to base those limits and the compliancemethod in the permit on the reference method that was used in setting the standard. Title 40CFR section 51.100(s)(2) states: "For purposes of determining compliance with emission limits,VOC will be measured by the test methods in the approved [SIP] . . . ." For example, if astandard specified that Method 25 or 25A should be used to establish the percent reduction of acontrol device, it is appropriate to use only Method 25 or 25A as the compliance method in thepermit because estimating the total mass of the VOC emissions is not necessary to judgecompliance with such a standard. If a compliance method was not specified in the standard, thenan appropriate one must be developed and approved by the permitting authority during theprocess of obtaining a permit. In addition, when permit limits are designed to maintain afacility’s emissions below an applicable threshold (e.g., NSR), the limits should require the useof methods that quantify the actual mass of the VOC’s emitted. Where such information isneeded for applicability determinations and compliance, and an existing permit requires a testprotocol that does not accurately identify mass emissions, the permit-holder should notify thepermit authority and request a change in its test protocol.

I appreciate the ongoing work of the CRA with us on investigating alternative methodsfor that industry. I also appreciate this opportunity to respond to your questions, and I hope thisresponse answers your questions. If you have additional questions on the appropriate use ofdifferent test methods, please contact Gary McAlister at (919) 541-1062. For questions relatedto potential compliance issues with NSR or Title V, please contact Mr. Cary Secrest of ORE at(202) 564-8661.

Sincerely,

Stephen D. Page DirectorOffice of Air Quality Planning

and Standards

/s/ 12/30/2003

Attachment #11

Attachment #12

ENVIRONMENTAL BULLETIN - No. 8 -3- Minnesota Pollution Control Agency

All of Minnesota’s ethanol facilities currently inoperation fire natural gas to run their process and controlequipment with propane often used as a back-up fuel. Oneplant under construction and another proposed facility willburn coal as a primary fuel source. The major sources ofNOx emissions from dry mill plants are the boilers used forfacility operations, the dryer, and the control equipmentultimately placed on the dryers. In almost every case, thiscontrol equipment was a thermal oxidizer (TO) or a regen-erative thermal oxidizer (RTO). Both are referred to here assimply thermal oxidizers (TOs). VOCs proved more elusivecoming from multiple units at the plants including the dryer,distillation columns, fermentation tanks and scrubbersassociated with these sources.

Quantifying VOC EmissionsBoth regulators and the fuel ethanol industry agree

that there are at least 8 to 10 VOC species that can bequantified from the gas stream of ethanol plant emissionunits. These species include acrolein, acetaldehyde, ethanol,furfuraldehyde, formaldehyde, acetic acid, lactic acid, andformic acid with some substitutions to this list at specificplants (EPA, 2004). EPA estimates that this list comprisesapproximately 60 to 90 percent of the gas stream whileindustry representatives maintain that these constituentsaccount for 90 percent or more of the gas stream. There islittle quantitative evidence to substantiate either position.

The effluent gas from ethanol plants can have amoisture content as high as 50 percent. The in-stack gascontains a large volume of moisture droplets that entrainorganic chemicals and act like particles in the gas stream.Conventional stack testing for total VOCs (EPA method25A) relies on instruments that draw gas through samplingequipment at a constant rate. The moisture in the gas isremoved and the sample then passes through a flameionization detector where volatiles are ignited and radiative

Figure 2: Schematic of the Dry Mill Ethanol Production Process.

Source: Renewable Fuels Association


energy at specific wavelenghts is recorded. Because theequipment used in this method does not adjust forisokinetic sampling (modifying the sampling rate tomatch the uneven flow across the stack diameter) as isgenerally done for particulate testing, the method islikely to underquantify any VOC that may be containedin the water droplets.

VOC testing currently accepted under method 25Ais designed to report total VOC as “carbon” meaningit assigns a mass to the sample based on the amount ofcarbon present. It does not account for unevenlydistributed species and larger oxygenated molecules.For example, in a gas stream where ethanol(CH3CH2OH) contributes more to the total mass thanacetaldehyde (CH3CHO, i.e., both having two carbonatoms), this test method would be unable to differenti-ate between the two and therefore fail to accuratelycharacterize VOC emissions.

The chemicals listed earlier were only a startingpoint, and subsequent testing has indicated the presenceof other species in the gas streams at most plants.Many compounds have been reported as “non-detect”values or “below detection limit” because the equip-ment used to sample the gas stream was not sensitiveenough to quantify the actual amount present in the gasstream. This does not necessarily mean that thechemical is not present but rather that the chemical islikely present in some quantity between zero and thedetection limit. Since there are possibly many com-pounds present below detection limits, the total VOCemissions represented by the sum of these unquantifiedemissions could be significant.

EPA attempted to resolve some of themeasurement uncertainties by developing a testingprotocol specifically for ethanol plants. This proce-dure, entitled the “Midwest Scaling Protocol,” em-ployed a combination of existing testing methods.Since the major obstacle to traditional VOC testing wasthe entrainment of organic chemicals in water droplets,the solution was to sample the stack gases as if thesource were emitting a particulate gas stream. Thismethod required isokinetic testing and collection ofsamples in glass impingers, some containing sodiumbisulfite solutions in order to precipitate out individualchemicals (EPA, 2004).

A “scaling factor” was developed to increase the totalmass emissions quantified under Method 25A in order toaccount for the mass of chemicals not measurable. Thisfactor was calculated from the speciated VOC data andapplied to the total organic carbon mass measured byMethod 25A which was performed in conjunction with themethodology as implemented by the Midwest ScalingProtocol. EPA developed a default scaling factor of 2.2 thatcould be applied to total VOC results from Method 25A inlieu of conducting speciated testing. In other words, thetotal VOC would be increased by a factor of more thantwo to account for the portion of chemicals not measuredby the stack test but still contributing to the total mass ofVOC. This scaling factor was designed to be “conserva-tive,” that is, designed so that any error would be on the sideof overestimating emissions.

Table 1 presents the averaged results from these testsbetween December 2002 and August 2004. For all facilities,these tests were performed after control equipment wasinstalled as required by the consent decree. All tests wereconducted by an independent testing consultant. Pre-testmeetings were held with facility personnel, testing consult-ants, and MPCA staff to ensure that the test would complywith the combined methodologies described above. Thesemeetings also served to discuss how the data would bereported (speciated, total mass, etc.). MPCA staff witnessedparts of most tests. Prior to submittal to the MPCA, reportswere sent to the ethanol facility’s management or consultantsfor review and approval. Reports were then reviewed bythe MPCA for accuracy, compliance with approved testmethodologies, calibration of equipment, any problematicoccurrences in the field, and chain of custody samplehandliing parameters, and ultimate verification of results.

ENVIRONMENTAL BULLETIN - No. 8 -5 Minnesota Pollution Control Agency

Data Summary andInterpretation

Summary statistics for the ethanol plant emissionstest data are presented in Table 1. Figure 3 is a graphicalrepresentation of the distribution of emission rates for eachchemical from each type of emission unit. In general, mostfacilities consistently reported detectable levels of acetalde-hyde, acetic acid, ethyl acetate, formaldehyde, ethanol andmethanol at one or more emissions units. The amount ofeach species and that which is the predominant constituentvaries across sources. Acetic acid and ethanol were thepredominant emissions from cooling cyclones. From thesingle fluid bed cooler tested, acetic acid was the dominantemission. Acetaldehyde and acetic acid accounted for mostof the emissions from thermal oxidizers. Fermentationscrubbers had the most diverse and highest overall emis-sions with significant amounts of acetaldehyde, acetic acid,ethanol, ethyl acetate, and isoamyl alcohol. Emissions of allsubstances were low from distillation scrubbers.

Several limitations in the data should be stated clearly.First, not all emissions units were tested for the same analytes,confounding comparisons among units and facilities. There isalso considerable variation in the quantities of analytes acrossfacilities, which we believe represents actual differences amongfacilities. The data are also left-censored, including valuesreported as zero and values reported as below detectionlimits. The proper treatment of values reported as zero orbelow detection is a matter of discussion. Ignoring suchvalues will bias the data and lead to incorrect estimates ofemissions.

In some cases the total speciated emissions (Method 18)did not match well with the total VOC measurement (Method25A). The discrepancies were greatest for the thermaloxidizer test results at the Agri-Energy and Al-Corn facilities,along with one of the test runs at the EXOL facility. Forthese tests, the speciated VOC emissions were much largerthan the total VOC measurement, a finding that appeared tobe due to suspiciously low total VOC measurements andperhaps a better accounting of larger oxygenated compounds

Table 1: Summary Statistics of VOC Measurements at Ethanol Production Facilities. Units are Pounds per Hour (lb/hr).

Emission Unit Statisticacetalde-

hyde

2,3-butane-

dionacetic acid acetoin acetone acrolein ethanol

ethyl acetate

formalde-hyde furfural

isoamyl alcohol methanol

total speciated

total VOCs

percent speciated

n 9 9 9 9 6 9 9 9 9n<dl 0 0 0 3 3 6 0 0 0min 0.02 0.17 0.12 <0.01 <0.02 <0.01 0.65 0.55 54%max 0.10 0.82 4.01 0.01 0.03 0.03 4.44 3.29 145%

mean 0.05 0.44 1.28 0.01 0.02 0.02 1.83 1.78 105%95% UCL 0.08 0.65 2.48 NA 0.03 0.03 2.94 2.4 127%

n 3 3 3 3 3 3 3 3 3 3n<dl 0 0 1 0 0 0 0 0 0 0min 0.02 0.01 0.00 0.01 0.01 0.00a 0.00a 0.04 0.03 120%max 0.06 0.01 <0.0005 0.02 0.01 0.00a 0.00a 0.1 0.05 194%

mean 0.04 0.01 NA 0.01 0.01 NA NA 0.07 0.04 163%95% UCL NA NA NA NA NA NA NA NA NA NA

n 17 3 17 3 17 17 14 17 6 6 14 17 17 17n<dl 0 0 6 3 15 0 0 6 5 0 10 0 0 0min 0.01 0.02 <0.01 <0.01 <0.005 0.01 0.11 0.00a <0.005 0.02 <0.005 0.15 0.54 15%max 1.97 0.07 0.26 <0.01 0.02 18.27 1.28 0.01 0.01 0.45 0.02 20.56 13.67 183%

mean 0.66 0.05 0.08 NA 0.006 4.20 0.39 0.005 0.01 0.22 0.007 5.62 5.9 88%95% UCL 1.19 NA 0.14 NA 0.009 9.41 0.73 0.007 0.01 0.45 0.012 10.11 7.76 127%

n 3 3 3 3 3 3 3n<dl 0 0 0 0 0 0 0min 0.14 1.54 0.16 0.03 1.87 2.86 44%max 0.22 2.24 0.16 0.04 2.66 4.47 93%

mean 0.17 1.89 0.16 0.04 2.26 3.87 62%95% UCL NA NA NA NA NA NA NA

n 18 6 18 6 6 15 18 12 18 18 18 18 18n<dl 6 3 3 3 3 13 9 7 6 12 0 0 0min <0.03 <0.05 <0.05 <0.05 <0.05 <0.05 <0.03 <0.03 <0.03 <0.03 0.25 0.01 25%max 0.73 0.06 1.46 0.17 0.03 0.07 2.28 0.11 0.17 0.20 4.43 3.56 9900%

mean 0.21 0.03 0.73 0.08 0.03 0.05 0.28 0.04 0.08 0.05 1.46 1.1 877%95% UCL 0.36 0.07 1.00 0.11 0.11 0.06 0.68 0.07 0.11 0.09 1.97 1.88 3352%

a - values reported as zero on test report

Thermal Oxidizer

Cooling Cyclone

Distillation Scrubber

Fermentation Scrubber

Fluid Bed Cooler


through speciated testing than provided in Method 25Aalone. The total VOC measurements also appeared anoma-lously low at the one distillation scrubber tested at the Al-Corn facility. At other emission units and other facilities thetwo test methods were in better agreement, although thevariance among the test results was large.

While this small data set is far from perfect, it is still themost extensive available that we’re aware of. It is clear thatfurther, systematic testing is necessary to thoroughly charac-terize the complex gas stream from these facilities. EPAestimates that this testing captures approximately 60 to 90percent of the constituents based on testing at one facility in2001 comparing Method 18 results as carbon to Method25A results. The ethanol industry believes the true percentagecaptured to be at the higher end of that range. The availabledata are insufficient to strongly support either contention.The percentage quantified may not be important to the typeof controls installed or how the facility routinely operates,but imprecision in the information complicates accuratehealth risk analysis of the emissions from these facilities.

Use of Data in Risk AnalysisThe MPCA has developed some experience in

permitting ethanol plants. Recently, newly proposed plantshave been asked to perform a risk analysis prior to permit-ting. Evaluating potential health effects from a facility issomething new to the permitting process for ethanolproduction facilities and, as such, presents challenges. VOCsare usually risk drivers at ethanol plants, so scarcity andimprecision in VOC emissions data complicates the AirEmission Risk Analysis (AERA) process. The MPCA usesthe AERA tool to identify potentially unacceptable healthrisks at new and expanding facilities. For most ethanolproduction facilities, the chemicals that pose the greatest riskare NOx, acetaldehyde, formaldehyde, and acrolein. Withthe exception of NOx, which can usually be well controlledand characterized, these chemicals are part of the complexVOC gas stream emitted by the plants.

In order to quantify potential risk, the analysis requiresboth reliable emission data and chemical-specific toxicityvalues that can be applied to the emission data. As dis-cussed, VOC emissions at fuel ethanol production facilities

Figure 3: Bar Chart of Speciated VOC Emission Rates by Emissions Unit.


The MPCA’s Environmental Bulletin Series is designedto highlight environmental outcomes and results ofscientific studies the MPCA conducts in air, water andwaste management. The bulletin is availableelectronically on the MPCA’s web site at: http://www.pca.state.mn.us.

Correspondence with the author about this bulletin canbe directed to Daniel Brady [email protected] or (651) 282-6144 orGregory Pratt at [email protected] or(651) 296-7664. For more information about theEnvironmental Bulletin Series, contact either of thefollowing MPCA staff of the EnvironmentalInformation and Reporting Unit.

Patricia Engelking (651) 297-3847Tom Clark (651) 296-8580

Printed on recycled paper with at least 30 percentpost-consumer waste.

Minnesota Department of Health. 2005. Ethanol SectorSpecific Interim Exposure Values for Air, available at http://www.health.state.mn.us/divs/eh/risk/guidance/essievs.html

Renewable Fuels Association. Ethanol Production Process,available at www.ethanolrfa.org/prod_process.html.

U.S. Environmental Protection Agency. 2004. MidwestScaling Protocol for the Measurement of “VOC MassEmissions” VOC Sampling at Wet and Dry Grain Mills andEthanol Production Facilities. Office of Air Quality Planningand Standards.

U.S. Environmental Protection Agency. 2000. FederalReference Method 5, Particulate Gas Stream Sampling.Emission Measurement Center.

USEPA, OSWER, 2002)Calculating Upper ConfidenceLimits for Exposure Point Concentrations at HazardousWaste Sites (http://www.hanford.gov/dqo/training/ucl.pdf)).

Yacobucci, B., Womach, J. 2004. Fuel Ethanol: Backgroundand Public Policy Issues. Congressional Research ServiceReport for Congress.

ReferencesArmstrong, S. 1999. Ethanol: Brief Report on its Use inGasoline. Cambridge Environmental, Inc. sponsored by theRenewable Fuels Association. Cambridge, MA.

Interpoll Laboratories, Inc. 2003. Results of the December4-5, 2002 Air Emission Compliance Test at the Ethanol 2000Facility Located in Bingham Lake, Minnesota. Circle Pines,Minnesota.

Interpoll Laboratories, Inc. 2003. Results of the January 28-30, 2003 Air Emission Compliance Test at the Agri EnergyFacility Located in Luverne, Minnesota. Circle Pines,Minnesota.

Interpoll Laboratories, Inc. 2003. Results of the January 21-23, 2003 Air Emission Compliance Test at the Al-CornClean Fuel Facility Located in Claremont, Minnesota. CirclePines, Minnesota.

Interpoll Laboratories, Inc. 2003. Results of theAugust 3, 2004 Air Emission Compliance Test at theAl-Corn Clean Fuel Facility Located in Claremont, Minne-sota. Circle Pines, Minnesota.

Interpoll Laboratories, Inc. 2003. Results of theApril 1-2, 2003 Air Emission Compliance Test at thePro-Corn Facility Located in Preston, Minnesota.Circle Pines, Minnesota.

Interpoll Laboratories, Inc. 2003. Results of theAugust 26-29, 2003 Air Emission Compliance Test at theEXOL [Agra Resources] Facility Located in Albert Lea,Minnesota. Circle Pines, Minnesota.

Interpoll Laboratories, Inc. 2004. Results of theJanuary 20-29, 2004 Air Emission Compliance Test at theDiversified Energy Company Facility Located in Morris,Minnesota. Circle Pines, Minnesota.

Interpoll Laboratories, Inc. 2004. Results of the June 2-4,2004 Air Emission Compliance Test at the Chippewa ValleyEthanol Company Facility Located in Benson, Minnesota.Circle Pines, Minnesota.

Knapp, K., Stump F., Tejada, S. 1998. The Effect of EthanolFuel on Emissions of Vehicles over a Wide Range ofTemperatures. Journal of the Air and Waste ManagementAssociation. Vol. 48, no. 1047-3289, pp. 646-653. 1.

Minnesota Department of Agriculture. 2005. Ethanol Plantsin Minnesota, available at www.mda.state.mn.us.

Attachment #13

Administrator

Line

Administrator

Pencil

Administrator

Pencil

Administrator

Line

Administrator

Pencil

Administrator

Pencil

Administrator

Pencil

Administrator

Line

Administrator

Pencil

Administrator

Pencil

Administrator

Line

Administrator

Pencil

Attachment #14

Attachment #15

Attachment #16

Attachment #17

Attachment #18