                                       
                                       
                                       
                                       
                      SNAP TECHNICAL BACKGROUND DOCUMENT:
                                       
                          RISK SCREEN ON THE USE OF 
                            SUBSTITUTES FOR CLASS I
                          OZONE-DEPLETING SUBSTANCES:
                                       
                      REFRIGERATION AND AIR CONDITIONING
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                     U.S. Environmental Protection Agency
                          Office of Air and Radiation
                       Stratospheric Protection Division
                                Mail Code 6205J
                              401 M Street, S.W.
                            Washington, D.C.  20460
                                       
                                       
                                       
                                       
                                  March 1994
                                       
	NOTICE


      This background document supports the U.S. Environmental Protection Agency's (EPA) rulemaking for controlling the replacement of Class I ozone-depleting substances (ODSs) with any substitute that presents unacceptable risk to human health or the environment where safer alternatives are available.  Specifically, this document examines the risk from the use of various refrigeration and air conditioning substitutes.  Similar documents address other industry sectors.

      ICF Incorporated provided support for the preparation of this document at the technical direction of the Stratospheric Protection Division, Office of Air and Radiation, EPA, Contract No. 68-D9-0068,  Work Assignments 3-23 and 3-69; and Contract No. 68-C2-0107, Work Assignment 39; and Contract No. 68-D3-0021, Work Assignments 1-11 and 2-18.  Meridian Research Incorporated also provided technical support under Contract No. 68-D9-0068, Work Assignment 3-68 and Contract No. 68-D3-0021 Work Assignment 2-18, particularly for the occupational evaluation.  ZBA, under Work Assignment 3-74, provided technical support for evaluating the feasibility of substitutes.


	TABLE OF CONTENTS


Section	Page

Notice		ii

Executive Summary	vi
      ES.1	The Refrigeration and Air Conditioning Sectors	vi
      ES.2	Risk Screening Methodology	viii
      ES.3	Risk Screening Results	x

1.	INTRODUCTION	1-1
      1.1	Background	1-1
      1.2	Evaluation of Substitutes	1-2
      1.3	Risk Screening Reports for Refrigeration and 
            Air Conditioning	1-3

2.	OVERVIEW OF SUBSTITUTES AND END USES	2-1
      2.1	End Uses	2-1
      2.2	Alternative Refrigerants	2-2
            2.2.1	HCFCs	2-2
            2.2.2	HFCs	2-3
            2.2.3	Hydrocarbons	2-3
            2.2.4	Ammonia	2-3
            2.2.5	Chlorine	2-3
            2.2.6	Potential Substitutes Manufacturers Have Submitted as
                  Confidential Business Information	2-4
      2.3	Alternative Technologies	2-4
            2.3.1	Absorption Refrigeration Systems	2-4
            2.3.2	Evaporative Cooling	2-4

3.	TOXICITY REFERENCE VALUES FOR SUBSTITUTES	3-1
      3.1	Occupational (Inhalation) Exposure Limits	3-1
      3.2	General Population/Consumer (Inhalation) Toxicity Reference Values	3-6
            3.2.1	Reference Concentrations (RfCs)	3-6
            3.2.2	Cancer Slope Factors 	3-8
      Chapter 3 References	3-9

4.	ATMOSPHERICS MODELING	4-1
      4.1	Introduction	4-1
      4.2	Description of Refrigerator and Air Conditioning End Uses	4-1
      4.3	Control Options	4-4
      4.4	The Baseline for the Analysis	4-4
            4.4.1	Uniform Assumptions Made Throughout the Analysis	4-4
            4.4.2	Assumptions Made for All Other End Uses	4-5
            4.4.3	Assumptions Made Within the End Use of Concern	4-5
                  4.4.3.1  Ideal Substitute Baseline	4-5
                  4.4.3.2  No Substitution Baseline	4-7
      4.5	Overview of the Modeling Approach	4-8
            4.5.1	Chemical Use	4-8
            4.5.2	Chemical Emissions	4-8
            4.5.3	Atmospheric Effects	4-8
            4.5.4	Human Health Effects	4-9
      4.6	Model Results	4-10
            4.6.1	Description of the Results	4-10
            4.6.2	Interpreting the Results	4-10
            
      4.7	Global Warming 	4-92
            4.7.1	The Nature of the Global Warming Problem	4-93
            4.7.2	Properties of Greenhouse Gases	4-93
            4.7.3	Properties of CFCs and CFC Substitutes	4-95
            4.7.4	Indirect Effects from Changes in Energy Efficiency	4-98
                  4.7.4.1	Refrigeration and Air Conditioning End Uses	4-98
                  4.7.4.2	Foam Insulation	4-99
      Chapter 4 References	4-100

5.	OCCUPATIONAL EXPOSURE AND HAZARD ANALYSIS	5-1
      5.1	Introduction	5-1
      5.2	Description of the Model	5-1
      5.3	Discussion of the Results	5-4
            5.3.1	Presentation of Results	5-4
            5.3.2	Interpretation and Comparison of Model Results with Sampling Data	5-5
      5.4	Potential Workplace Flammability and Explosivity Risks	5-8
      Chapter 5 References 	5-10

6.	CONSUMER EXPOSURE AND RISK SCREENING ANALYSIS	6-1
      6.1	Home Appliances	6-2
            6.1.1	Risks from Long-term Exposure to Leakage and Servicing Events  Previous Estimates	6-2
            6.1.2	Risks form Short-term Exposure During Servicing and Accidental
                  Releases	6-3
            6.1.3	Summary of Modeling Results	6-10
            6.1.4	Safety Concerns:  HFC-152a in Household Refrigerators	6-10
      6.2	Mobile Air Conditioners (MACs)	6-11
            6.2.1	Approach Used in a Previous Consumer Exposure Report	6-11
            6.2.2	Comparison of Parameters Contained in Consumer Exposure Report 
                  with Data from Other Sources	6-12
            6.2.3	Summary	6-13
      Chapter 6 References	6-14

7.	GENERAL POPULATION EXPOSURE AND RISK ANALYSIS	7-1
      7.1	HCFCs and HFCs	7-1
            7.1.1	Approach	7-1
            7.1.2	Results	7-4
            7.1.3	Caveats and Limitations	7-4
            7.1.4	Conclusions	7-7
      7.2	Other Alternative Refrigerants	7-7
            7.2.1	Background	7-8
            7.2.2	Approach	7-8
            7.2.3	Results 	7-9
      7.3	Chlorine	7-9
      7.4	Lithium Bromide and Ammonia in Absorption Systems	7-11
      Chapter 7 References	7-12
      Attachment 7-A:  Detailed Categorization of Refrigeration End Uses and 
                      Equipment Types	7-A-1
      Attachment 7-B:  Assumptions for Ambient Air Release Calculations	7-B-1
      Attachment 7-C:  Methodology for Estimating General Population Exposures to Ambient  		    Air Releases of Substitutes for Class I Substances for SNAP Risk     Screens 	7-C-1
      Attachment 7-D:  Detailed Results	7-D-1
      Attachment 7-E:  Percent of Equipment with Assumed Charge Size	7-E-1

8.	VOLATILE ORGANIC COMPOUND ANALYSIS	8-1
      8.1	VOC Emissions from Substitutes to Class I ODSs	8-1
      
      8.2	VOC Emissions from All Sources	8-1
      8.3	Comparison of VOCs from Substitutes to VOCs from All Sources	8-1
      8.4	Caveats and Limitations	8-2
      Chapter 8 References	8-3
      Attachment 8-A:  Estimation of Total VOC Emissions from All Sources in 1995	8-A-1
      Attachment 8-B:  VOC Controls	8-B-1



                               EXECUTIVE SUMMARY


      This report presents a screening assessment of human health and environmental risks associated with the use of potential substitutes for ozone depleting substances (ODSs) in the refrigeration and air conditioning sectors.   ODSs are being phased out of production in response to a series of diplomatic and legislative efforts that have taken place in the past few years, including the Montreal Protocol and, more recently, the Clean Air Act Amendments of 1990.  The analyses presented in this report support the United States Environmental Protection Agency's Significant New Alternatives Policy (SNAP) Program, which is authorized by Section 612 of the Clean Air Act Amendments.  Implementation of this program involves: (1) developing, promulgating, and administering a regulatory program for identifying and evaluating substitutes, and (2) publishing a list of acceptable and unacceptable substitutes for specific end uses.

      The remainder of this summary provides an overview of ODS use in the refrigeration and air conditioning sectors, the methodology used to conduct the risk screen, and the key results.

ES.1	THE REFRIGERATION AND AIR CONDITIONING SECTORS

      ODS use in the refrigeration and air conditioning sectors accounted for almost 22 percent of the total use of Class I substances in the United States in 1990.  CFC-12 is the most widely used refrigerant, with applications in mobile air conditioners (MACS), household refrigerators and freezers, various appliances, chillers, retail food refrigeration equipment, cold storage warehouses, refrigerated transport systems, and industrial equipment.  CFC-11 is most commonly used to provide cooling for large buildings, while CFC-115, as a component in the refrigerant blend R-502, is used for low temperature applications in a variety of sectors.  CFC-113 and CFC-114 are used in special applications chillers.

      Of the Class II controlled substances, HCFC-22 is the refrigerant of choice in small to medium size air conditioning systems, and some types of retail food and industrial process refrigeration systems.

      EPA has divided the refrigeration and air conditioning sector into the following general end uses:

      	Retail Food Systems:  These include stand alone refrigeration cases found in small markets, convenience stores, restaurants and other food establishments, large systems found in supermarkets, commonly referred to as parallel systems, and HCFC-22 systems found in a wide variety of retail and service establishments;

      	Cold Storage Systems:  Public and private facilities used to store meat, produce, dairy products, frozen foods, and other perishable goods.  These systems refrigerate warehouses of varying sizes and include high pressure and low pressure systems;

      	Chillers:  These are large commercial air conditioning systems used to cool office buildings, shopping malls, and other large indoor spaces.  End uses include CFC-11 (low pressure) chillers, CFC-12 centrifugal chillers, CFC-12 reciprocating chillers, HCFC-22 chillers, CFC-114 centrifugal chillers, and CFC-500 centrifugal chillers;


      	Industrial Process Refrigeration:  Refrigeration systems used in the chemical, pharmaceutical, petrochemical, and other manufacturing and food processing industries.  These include both packaged and built-up systems used in manufacturing processes as well as extra-large ice makers;

      	Ice Skating Rinks:  There are several different types of ice rink systems, but all are grouped into one end use in this analysis;

      	Ice Makers:  These include small, medium, and large ice makers used by a number of entities, including restaurants and hotels;

      	Household and Light Commercial Air Conditioning:  These include window units, packaged terminal air conditioners, central air conditioners, direct expansion commercial air conditioners, and heat pumps used for space cooling and heating in homes and businesses.  All systems in this end use currently use HCFC-22;

      	Refrigerated Appliances:  These include household refrigerators and freezers, water coolers, vending machines, and dehumidifiers.  Most currently use CFC-12 as a refrigerant, but there are some CFC-502 units in existence;

      	Refrigerated Transport Systems:  These include systems designed to refrigerate materials in transit.  They are found in trucks, trains, and ships; and

      	Mobile Air Conditioning:  This end use includes systems installed in motor vehicles for the comfort of occupants.

      The following substitutes to Class I and Class II substances are examined in this report: 

      	HCFCs:  HCFCs are chemically similar to CFCs except that they contain hydrogen in addition to chlorine and fluorine.  HCFCs analyzed as possible substitutes to the use of Class I chemicals in the refrigeration sector include HCFC-22, HCFC-123, HCFC-124, HCFC-141b, and HCFC-142b.

      	HFCs:  HFCs do not contain chlorine and do not contribute to destruction of stratospheric ozone.  They can, however, have high greenhouse gas potentials and may contribute to global warming.  HFCs analyzed as possible substitutes to Class I and Class II substances in the refrigeration sector include HFC-134a, HFC-152a, HFC-143a, HFC-125, HFC-227ea, and HFC-23.

      	Hydrocarbons:  Propane, ethane, propylene, and (to some extent) butane are used as refrigerants in specialized industrial applications, primarily in oil refineries and chemical plants.  Perfluoropropane and other perfluorocarbons (PFCs) also have potential applications in the refrigeration sector.

      	Ammonia:  Ammonia has been used as a refrigerant in vapor compression cycles for more than 100 years.  Its main application is in moderate to low temperatures systems.

      	Chlorine:  Chlorine has been proposed as a Class I substitute refrigerant for use in chlorine liquification, a processing step in the manufacture of the chemical.

      	Potential Substitutes Submitted as Confidential Business Information.  A few substitutes included in this analysis were submitted by their manufacturers as confidential business information.  As a result the identity of these substances are not revealed.  The results of the use analyses for these compounds are presented under the following identities:  R200b through 200i; and HCFC/HFC/Fluoroalkane.
      
      In addition to these chemical substitutes, there are a number of alternative technologies considered in this risk screening analysis:

      	Absorption Refrigeration Systems:  Absorption refrigeration systems are a major existing alternative to systems based on vapor compression cycles.  Ammonia and lithium bromide are used in absorption refrigeration and air conditioning systems.

      	Evaporative Cooling:  Evaporative cooling is another technology that could be an alternative to CFC- and HCFC-based vapor compression systems in some applications.  Contrary to vapor compression systems, evaporative coolers do not require compressors and consequently can offer considerable energy savings.

ES.2  RISK SCREENING METHODOLOGY

      To examine risks associated with the use of substitutes for ODSs in each sector, EPA examined many different types of exposure routes, receptors, and effects.  As discussed in more detail later, EPA used an atmospherics modeling framework to assess global ozone depletion and associated health effects.  To examine risks to workers, consumers, and the general population, EPA first conducted screening-level assessments using generally conservative assumptions to identify scenarios potentially of concern, and then conducted more in-depth assessments of those scenarios.  These risk screening assessments were conducted in three basic steps:

      (1)	Assess exposure.  EPA estimated levels of substitutes to which workers, consumers, the general population, and environmental receptors may be exposed, and over what period of time.

      (2)	Evaluate toxicity.  EPA compiled information on the toxicity of each substitute, including occupational exposure limits, reference concentrations, and cancer slope factors.  These values were identified from an appropriate source (e.g., Permissible Exposure Limits [PELs] from the Occupational Safety and Health Administration [OSHA]) or were derived for this report using extrapolation and other techniques.

      (3)	Estimate risk.  Exposure assessment results were combined with toxicity values to estimate risks.  Uncertainties in the risk estimates were analyzed qualitatively.

      EPA did not estimate risks for every possible combination of substitute, end use, and type of effect; combinations that clearly were not of concern were not analyzed.  For example, substitutes that volatilize quickly when exposed to the atmosphere were not examined for potential releases to streams and effects on aquatic life.  EPA believes that all of the significant human health and environmental exposure pathways and effects were covered in the risk screen.

      Below are brief descriptions of how the general approach discussed above was applied to the risk screen for the refrigeration sector.
      
      Worker Exposure and Risk.  EPA used a mass-balance model to estimate worker exposure to HCFC and HFC refrigerant and air conditioning substitutes for ODSs used in this sector.  Potential exposure levels to the substitutes were assessed during the manufacture, installation, and servicing of refrigeration and air conditioning equipment.  Estimated exposures were compared to OSHA and EPA occupational exposure limits to screen the potential risks to workers.


      Consumer Exposure.  Consumers may be exposed to ODS substitutes that could potentially be used in household refrigeration and air conditioning equipment and mobile air conditioners.  Mass-balance models were used to estimate the risks from short term and long term exposures to ODS substitutes used in these end uses.  To assess noncarcinogenic effects, estimated long term exposures were compared to reference concentrations.  Cancer risks were calculated using standard risk equations.  The estimated short term exposures were compared to interim emergency guidance levels (EGLs) developed by EPA for each substitute.

      General Population Exposure.   To screen risks to the general population, EPA considered releases from the following types of facilities:  manufacturing facilities that produce refrigeration or air conditioning equipment, operating facilities using refrigeration or air conditioning equipment, service stations (mobile air conditioners only), and recycling centers (household refrigeration and air conditioning end uses only) or salvage yards (mobile air conditioners only).  Model facilities were developed for each refrigeration and air conditioning end use.  For each type of facility, conservative release estimates were multiplied by dispersion factors to predict fenceline concentrations of HCFC, HFC, and other chemical substitutes downwind of these facilities.  These concentrations were then compared to estimated reference concentrations to screen noncancer health risks.  Cancer risks were calculated using EPA's standard risk equation.

      Atmospherics.  In this analysis, two different baselines are employed to present the atmospheric impacts associated with each of the potential substitutes in a given end use:  an "ideal" or "no-risk" baseline, and a "no-substitution" baseline.  These two baselines bound the range of health effects, since the first examines impacts from a risk-free scenario and the second presents effects assuming continued use of all CFCs.  These baselines can be summarized as follows:

      	"Ideal" Baseline:  This baseline presents the incremental ODP risks associated with the use each of the potential substitutes in a given end use relative to a hypothetical substitute with an ODP of 0.0.  In essence, this baseline presents the increase in ODP risks that could result from the use of the substitutes relative to a no risk (or ideal) baseline.

      	No Substitution Baseline:  Under this baseline, the incremental performance of each substitute within an end-use is judged against continued use of Class I and Class II substances.  In essence, this baseline presents the reduction in ODP risks associated with phasing out the use of the Class I compound in a given end use and switching to each of the potential substitutes.

      EPA used the Atmospheric and Health Effects Framework (AHEF) to model the effects of substitutes for ODSs in the refrigeration and air conditioning sector.  The framework employs the Vintaging Model for forecasting consumption and emissions of refrigeration and air conditioning substitutes.  The Atmospheric Stabilization Framework (ASF) uses these emissions as a basis for predicting changes in ozone levels.  Changes in UV-B flux are then estimated from the ASF output.  Finally, AHEF utilizes a human health effects model to calculate skin cancer cases and fatalities.

      VOCs.  VOC emissions from the refrigeration and air conditioning end uses were estimated under the conservative assumption that all ODSs currently used in refrigeration and air conditioning equipment would be replaced with VOCs.  These emissions were then compared to total VOC emissions from all sources at the national level.

      The use of substitutes for ODSs in the refrigeration and air conditioning sector were assumed to result in no releases of chemicals to surface waters that could result in potential risks to the general population or aquatic life, nor were the substitutes expected to result in the generation of solid waste that could potentially result in risks to the general population.

ES.3	RISK SCREENING RESULTS

      The results of EPA's risk screens are summarized briefly below and discussed in greater detail in the body of this report.
      
      Atmospherics.  Exhibits ES-1 and ES-2 summarize the impacts on skin cancers, fatalities, and cumulative Clx resulting from the use of the potential ODS substitutes in the refrigeration and air conditioning sectors relative to the ideal and no substitution baselines, respectively.  Exhibit ES-1 (ideal baseline) indicates that for the vast majority of end use and substitute combinations the increase in skin cancers is less than 10,000 cases and the increase in fatalities is less than 1,000, relative to the ideal baseline.  The greatest increase in skin cancers and fatalities results from certain substitutes examined in the MACs sector.  This is due to (1) the significant quantities of chemical use in the MACs sector; (2) the ODPs of some of the alternatives examined; and importantly (3) the timing of the availability of the chemical substitutes.  Exhibit ES-2 (no substitution baseline) indicates that virtually all of the substitutes analyzed are substantially less harmful to the ozone layer than continued use of CFCs.

      Worker exposure.  The results of EPA's analysis indicates that for the vast majority of end use, substitute, and operation combinations workers will not be exposed to potentially hazardous airborne gas or vapor concentrations.  Exhibit ES-3 summarizes the results for those scenarios where the model suggests that worker exposures exceed either the 8-hour or 30-minute occupational exposure limit (OEL) for the alternative refrigerant (Exhibit ES-4 summarizes the OELs for the potential alternative refrigerants).

      In considering the implications of the results presented in Exhibit ES-3, it is important to note the following observations that suggest that the analytical results will overstate likely occupational exposures to alternative refrigerants:

      	The emission estimates from the Vintaging Model used in the analysis are considered to be worst-case estimates.

      	The analysis does not assume the use of any local exhaust ventilation, which is likely to be used in manufacturing and disposal operations involving any significant releases.  The use of properly designed local exhaust ventilation would likely prevent gas or vapor concentrations from exceeding the OELs most of the time.

      	The analysis assumes for each release scenario that the entire daily release occurs over a single 8-hour workday at a single location.  This is not the case for many scenarios, and particularly for manufacturing operations, where the estimated daily release may occur over two or three workshifts or where the release may be spread over two or three production lines.

      	The vast majority of the exceedances over the OELs are the result of assuming that there will be no recycling of HFC refrigerants.  Given that such recycling is
      
es2

	EXHIBIT ES-3
	SUMMARY OF ESTIMATED WORKER EXPOSURE TO ALTERNATIVE REFRIGERANTS
	FOR SCENARIOS WHERE THE OCCUPATIONAL EXPOSURE LIMIT IS EXCEEDED[a]

                                       
                                 8-Hour TWA[b]
                                 30-Minute TWA
                                End-Use Sector
                                   Operation
                                  Substitute
                                  Material[c]
                                   Predicted
                                Exposure (ppm)
                                   Operation
                                  Substitute
                                  Material[c]
                                   Predicted
                                Exposure (ppm)
Retail Food (Parallel System)
Disposal



HFC-134a
HFC-125
HFC-143a
HFC-23
                                                                          1,579
                                                                          1,342
                                                                          1,917
                                                                          2,300
Disposal
HFC-134a
HFC-125
HFC-143a
HFC-23
                                                                          4,742
                                                                          4,030
                                                                          5,757
                                                                          6,909
Cold Storage Warehouses
Disposal
HFC-152a
HFC-23
                                                                          1,309
                                                                          1,234


                                                                               
Centrifugal Chillers (Low Pressure)
Manufacture
HCFC-123
                                                                          73[d]


                                                                               
Centrifugal Chillers (High Pressure)
Disposal
HFC-124
HFC-134a
HFC-227ea
                                                                          2,271
                                                                          3,050
                                                                          1,647
Disposal


HFC-124
HFC-134a
HFC-227ea
                                                                          6,819
                                                                          9,159
                                                                          4,928
Industrial Process (Low Pressure)
Manufacture
Repair
HCFC-123
HCFC-123
                                                                             54
                                                                             49


                                                                               

	EXHIBIT ES-3 (continued)
	SUMMARY OF ESTIMATED WORKER EXPOSURE TO ALTERNATIVE REFRIGERANTS
	FOR SCENARIOS WHERE THE OCCUPATIONAL EXPOSURE LIMIT IS EXCEEDED[a]

                                       
                                 8-Hour TWA[b]
                                 30-Minute TWA
                                End-Use Sector
                                   Operation
                                  Substitute
                                  Material[c]
                                   Predicted
                                Exposure (ppm)
                                   Operation
                                  Substitute
                                  Material[c]
                                   Predicted
                                Exposure (ppm)
Industrial Process
  (High Pressure          Build-Up)
Manufacture

Disposal
HFC-143a
HFC-152a

HFC-134a
HFC-152a
HFC-125
HFC-143a
HFC-23
                                                                          1,293
                                                                          1,016
                                                                               
                                                                         30,840
                                                                         47,661
                                                                         15,740
                                                                         18,375
                                                                         13,490
Manufacture


Disposal
HFC-134a
HFC-143a
HFC-152a
HFC-134a
HFC-152a
HFC-125
HFC-143a
HFC-23
                                                                          3,941
                                                                          7,751
                                                                          6,090
                                                                         92,608
                                                                        143,122
                                                                         47,266
                                                                         56,260
                                                                         40,509
Large Ice Makers
Repair
HFC-152a
                                                                          1,076


                                                                               
Ice Rinks
Disposal
HFC-134a
HFC-152a
HFC-23
                                                                          4,609
                                                                          7,123
                                                                          2,016
Disposal
HFC-134a
HFC-152a
HFC-23
                                                                         13,840
                                                                         21,390
                                                                          6,054
Central A/C and Home Heat Pumps
Manufacture
HFC-152a
                                                                          1,332


                                                                               

[a]	Exposure levels are rounded to the nearest whole number.
[b]	Time-weighted average.
[c]	One manufacturer's representative states that worker exposures to HCFC-123 during centrifugal chiller manufacture are typically 1 ppm or less, due to the use of exhaust ventilation.
[d]	The refrigerant blends analyzed for these sectors may be considered confidential business information (CBI) by the manufacturer.  As a result, the identification of the constituents of any blend and the proportion of each constituent contained in any blend are not presented in this exhibit.


	EXHIBIT ES-4
	OCCUPATIONAL EXPOSURE LIMITS FOR
	POTENTIAL ALTERNATIVE REFRIGERANTS*

                                  Substitute
                                   Material
                                       
                                   WGL (ppm)
                                       
                                   EGL (ppm)
HCFC-22
                                     1,000
                                     5,000
HFC-23
                                     1,000
                                     3,000
HCFC-123
                                        30
                                     2,500
HCFC-124
                                     1,000
                                     3,000
HFC-125
                                     1,000
                                     3,000
HFC-134a
                                     1,000
                                     3,000
HFC-143a
                                     1,000
                                     3,000
HCFC-142b
                                     1,000
                                     5,000
HFC-152a
                                     1,000
                                     4,000
HFC-227ea
                                     1,000
                                     3,000



            required under the Clean Air Act beginning in 1995, exposures should be below the OELs for these compounds.

      	The model's tendency to overstate worker exposures is further supported by sampling obtained for HCFC-123 refrigerants during repair operations conducted on low-pressure centrifugal chillers.

      Finally, the analysis did not consider the use of ammonia, butane, propane, or perfluoropropane as alternative refrigerants.  However, existing OSHA and ASHRAE standards governing the handling and use of these materials should serve to limit occupational exposures to them.

      Consumer exposure.  EPA's consumer exposure analysis of household refrigeration, household air conditioning, and mobile air conditioning equipment indicates that unless the entire refrigerant charge is released over a short period of time, none of the substitutes exceed EPA's threshold of concern for human health risk in any piece of equipment.  Routine leakage and servicing with recycling simply will not release enough refrigerant to pose significant short-term or long-term risks to human health.

      If the entire charge is released (due to an accident or during servicing without recycling of HFC-containing equipment), then the 30-minute average concentrations exceed the EGLs for substitutes in household air conditioning equipment only.  The following factors should be considered when judging whether these exceedances are cause for concern:

      	Recycling at servicing will be required for the HFCs by 1995.  Until then, EPA believes that most HFCs will be recycled anyway, because of their high cost relative to CFCs.

      	Accidental releases of an entire refrigerant charge occur infrequently.  The UNEP (1992) reports that less than 5 percent of the refrigerators and freezers have accidental releases of the entire charge.

      	Exposures are below levels that would raise concern for cardiotoxic effects.  This would be true even if higher, shorter-term TWA exposures (e.g., 15-minute TWAs) had been used as the basis for the assessment.


      General population exposure to ambient air releases.  EPA's general population exposure analysis indicates that even with conservative screening assumptions there are no end uses within the refrigeration and air conditioning sectors where the use of HFCs or HCFCs as substitutes for CFCs will result in exposure concentrations above the chemicals' health-based reference concentrations for non-cancer effects.  Similarly, there are no areas for which predicted cancer risks would be above the upper-bound cancer risk level of concern.  The analysis also indicates that general population exposures are not of concern for ammonia, butane, propane, and perfluoropropane.

      VOCs.  EPA's analysis indicates that potential VOC emissions from all substitutes for all end uses in the refrigeration and air conditioning sector are likely to be insignificant relative to VOC emissions from all other sources.

	1.  INTRODUCTION


      This report documents the results of the examination of human health and environmental risks associated with the use of substitutes for ozone depleting substances (ODSs) in the refrigeration and air conditioning sectors.  This introduction provides a brief background on the history and philosophy of the Significant New Alternatives Policy (SNAP) program, discusses in general terms the types of analyses performed to screen the risks from the substitutes, and provides a list of the chapters that follow.

1.1	BACKGROUND

      Over the past decade, a series of diplomatic and legislative efforts have taken place in response to growing concerns over the progressive depletion of the Earth's stratospheric ozone layer and its attendant health and environmental problems.  The first major step was in September 1987, when the United States and 23 other nations signed the Montreal Protocol.  The original agreement established a schedule for reducing by half the production and consumption of eight specific ODSs, including CFC-11, CFC-12, CFC-113, CFC-114, CFC-115, Halon-1211, Halon-1301 and Halon 2402.  When the parties to the Protocol met again in London in June 1990, they agreed to fully phase out the original eight ODSs, as well as methyl chloroform, carbon tetrachloride, and other fully halogenated CFCs.

      Congress incorporated the requirements of the London Amendments into the Clean Air Act (CAA) Amendments of November 15, 1990, and in some cases went beyond the London requirements.  Title VI of the CAA differs from the London Amendments by mandating a faster phaseout of methyl chloroform, a restriction on the use of hydrochlorofluorocarbons (HCFCs) after 2015, and a ban on the production of HCFCs after 2030.  When the Parties to the Montreal Protocol met in Copenhagen in November 1992, they further accelerated the phaseout dates for many controlled substances.  To implement the Copenhagen amendments and adjustments, EPA promulgated regulations in December 1993 that set the phaseout date at January 1, 1996 for Class I substances (as defined in the CAA), with the exception of halons which were subject to a phaseout by January 1, 1994.  The regulation also includes a phaseout of methyl bromide by the year 2001.

      Section 612 of the CAA requires EPA to develop a program to evaluate the risks to human health and the environment posed by alternatives to ODSs.  EPA is referring to this new program as the Significant New Alternatives Policy (SNAP) program.  The SNAP program is directed at fulfilling the general mandate in Section 612 of identifying acceptable and unacceptable substitutes for the Class I ODSs.  Initial SNAP implementation involves two key activities: (1) developing, promulgating, and administering a regulatory program for identifying and evaluating substitutes, and (2) publishing a list of acceptable and unacceptable substitutes for specific end uses.  

      EPA took the first major step in implementing Section 612 of the CAA on January 16, 1992, with publication of a Request for Data and Advance Notice of Proposed Rulemaking (ANPRM; 57 Federal Register, p. 1984).  This notice, developed with the help of the Stratospheric Ozone Protection Advisory Committee (STOPAC), presented the initial plan for implementing the SNAP program and requested producers and formulators of substitutes for ODSs to provide EPA with information to facilitate review of the substitutes.  A Notice of Proposed Rulemaking was published in the Federal Register on May 12, 1993.  The proposed rule described the proposed structure and process for administering the SNAP program and presented initial determinations on the acceptability of key substitutes.  This Final Rule provides both the SNAP process and additional SNAP determinations on the acceptability of substitutes.  In the future, substitutes will be addressed on a case-by-case basis through individual notices.  In addition, EPA will accept petitions requesting that a substance be added or deleted from either of the lists.
      
1.2	EVALUATION OF SUBSTITUTES

      To develop the lists of unacceptable and acceptable substitutes, EPA conducted a comprehensive screening assessment of the health and environmental risks posed by various substitutes in the context in which they are used.  Substitutes are grouped into one or more of eight general sectors:  refrigeration and air conditioning; solvent cleaning; foam blowing; fire extinguishing; aerosols; sterilization; adhesives, coatings, and inks; and tobacco expansion.  Within each sector, substitutes are evaluated in the context of particular end uses.  Examples of end uses within the refrigeration and air conditioning sectors include cold storage warehouse refrigeration systems, household refrigerators, and mobile air conditioners.  EPA also determined the feasibility of using each substitute by assessing its production capacity and projected market share.  Based on these analyses, EPA will identify as "unacceptable" only those substitutes that pose significantly higher human health and environmental risks than other available substitutes.  EPA does not intend to restrict substitutes that are only marginally worse based on some criteria.

      EPA's evaluation of each substitute in each end use is based on the following types of information and analyses:

      	Atmospheric effects are assessed by using models to predict stratospheric ozone depletion.  Ozone depletion is measured in terms of cumulative Clx loadings and increased incidence of skin cancer cases and skin cancer mortalities.  Changes in global temperatures may result from releases of the substitutes themselves or from changes in fossil fuel use due to increases or decreases in energy efficiency.  These impacts are discussed qualitatively.

      	Exposure assessments are used to estimate levels of substitutes to which workers, consumers, the general population, and environmental receptors may be exposed, and over what period of time.  These assessments are based on personal monitoring data or area sampling data if available.  Otherwise, exposures are assessed using measured or estimated releases as inputs to mathematical models.  Exposure assessments may be conducted for many types of releases, including releases in the workplace and in peoples' homes; releases to ambient air and surface water; and releases from the management of solid wastes.

      	Toxicity data are used to assess the possible health and environmental effects from exposure to the substitutes.  If Occupational Safety and Health Administration (OSHA)-approved or EPA-wide health-based criteria such as Permissible Exposure Limits (PELs; for occupational exposure), reference concentrations (RfCs; for noncarcinogenic effects), or cancer slope factors (for carcinogenic risk) are available for a substitute, exposure information is combined with this toxicity information to determine whether there is potential for concern.  Otherwise, toxicity data are used in conjunction with existing EPA guidelines to develop health-based criteria for interim use in these risk screens.

      	Flammability is examined as a possible safety concern for workers and consumers.  EPA screens flammability risk using data on flash point and flammability limits (e.g., OSHA flammability/combustibility classifications), test data on flammability in consumer applications conducted by independent laboratories (e.g., Underwriters Laboratories), and information on flammability risk minimization techniques.

      	Some of the proposed substitutes are volatile organic compounds (VOCs), which are chemicals that increase tropospheric air pollution by contributing to ground-level ozone formation.  Local and nationwide increases in VOC loadings from the use of substitutes are also evaluated.


      In conducting these assessments, EPA made full use of previous analyses, including the 1990 interim assessments prepared by EPA's Office of Pesticides and Toxic Substances (OPTS) and the supporting documentation (full citations are given in appropriate chapters of this report).  These analyses were modified in some cases to incorporate more recent data or new field experience.  Where possible, EPA incorporated data submitted in response to the ANPRM and NPRM.  Finally, these analyses assume that the regulated community is in compliance with applicable requirements of other statutes and regulations administered by EPA (e.g., recycling requirements promulgated under the CAA) and other federal agencies (e.g., enforceable workplace standards set by OSHA).

1.3	RISK SCREENING REPORTS FOR REFRIGERATION AND AIR CONDITIONING

      The remainder of this package includes a series of chapters on the following topics:

      	Overview of substitutes and end uses;

      	Human health and environmental concern levels for substitutes;

      	Atmospherics modeling (ozone depletion, global warming);

      	Occupational exposure and risk screening analysis;

      	Consumer exposure and risk screening analysis;

      	General population exposure and risk screening analysis (ambient air); and

      	VOC analysis.

Flammability is addressed within several of the chapters.  Compliance costs are addressed in conjunction with other SNAP decisions in a separate package, as are regulatory burdens and effects on small businesses.

      Finally, three of the substitutes analyzed in this report, HCFC-123, HCFC-124, and HFC-134a contribute to or generate trifluoroacetic acid (TFA) as an atmospheric transformation product.  The potential ecological effects on aqueous and terrestrial ecosystems are not addressed in this report because of insufficient data on the toxicity of TFA and the level to which ecosystems may be exposed.  Research efforts of the Agency in cooperation with the Alternative Fluorocarbons Environmental Acceptability Study (AFEAS) are underway to define the potential ecological risks associated with the formation of TFA in the environment.

      Although there are significant uncertainties about the potential risks associated with TFA, and research is continuing to address these risks, a number of preliminary statements can be made based on the research completed to date:

      	Research indicates that virtually 100% of the HCFC-123 and HCFC-124 released to the environment will degrade to TFA, and suggests that approximately 25% to 40% of HFC-134a will also degrade to TFA.


      	TFA is a highly soluble acid that will exist primarily in aqueous environments (rain, rivers, lakes, ocean, and groundwater).  Tests have shown that it is very resistant to decomposition and will therefore reside in the environment for long periods of time.

      	Given conservative estimates of global HCFC and HFC use and emissions, the environmental concentrations of TFA will be low over the next few decades.  

      	Experimental results show at high concentrations TFA has low to moderate acute toxicity to animals and initial tests indicate that relatively high concentrations of TFA cause toxic effects on some plants and algae.  Research to evaluate these effects is continuing.

      	The expected environmental concentrations of TFA over the next few decades are unlikely to pose a significant hazard to animals or plants.

      	There has been concern that TFA might degrade to MFA (monofluoroacetic acid) and that MFA could exist at high enough concentrations to pose health and environmental risks.  Research to date has identified no environmental process that would convert TFA to MFA.  Research to elucidate all the degradation pathways for TFA, including potential pathways to MFA is continuing.

      EPA is closely monitoring the research efforts underway with respect to TFA, and will assess the long term ecological risks associated with this compound when the research is completed.  As noted above, research completed to date has not uncovered any risks that warrant the implementation of measures to control atmospheric concentrations of TFA.  If such risks are identified in the future, the Agency will re-assess its position on TFA and evaluate measures to mitigate such risks.



	2.  OVERVIEW OF SUBSTITUTES AND END USES


2.1	END USES

      The refrigeration industry was the first to make widespread use of CFCs after this class of chemical compounds was discovered in the 1930's.  In 1990, refrigeration and air conditioning accounted for almost 22 percent of the total use of Class I substances in the United States.  There are over 500 million pieces of refrigeration and air conditioning equipment that use these chemicals as the working fluids in a vapor compression cycle.

      Many Class I substances exhibit desirable thermophysical properties for use in refrigeration cycles.  They are relatively nontoxic, nonflammable, and inexpensive to produce  characteristics that have contributed to their appeal as refrigerants.  CFC-12 is the most widely used refrigerant, with applications in mobile air conditioners (MACs), household refrigerators and freezers, various appliances, chillers, retail food refrigeration equipment, cold storage warehouses, refrigerated transport systems, and industrial equipment.  CFC-11 is most commonly used to provide cooling for large buildings, while CFC-115, as a component in the refrigerant blend R-502, is used for low temperature applications in a variety of sectors.  CFC-113 and CFC-114 are used in special applications chillers.

      Of the Class II controlled substances, HCFC-22 is the most common refrigerant in small to medium size air conditioning systems, and some types of retail food and industrial process refrigeration systems.

      EPA has divided the refrigeration and air conditioning sector into the following general end uses:

      	Retail Food Systems:  These include stand alone refrigeration cases found in small markets, convenience stores, restaurants and other food establishments, large parallel systems found in supermarkets, and HCFC-22 systems found in a wide variety of retail and service establishments;

      	Cold Storage Systems:  Public and private facilities used to store meat, produce, dairy products, frozen foods, and other perishable goods.  These systems refrigerate warehouses of varying sizes and include high pressure and low pressure systems;

      	Chillers:  These are large commercial air conditioning systems used to cool office buildings, shopping malls, and other large indoor spaces.  End uses modeled are CFC-11 (low pressure) chillers, CFC-12 centrifugal chillers, CFC-12 reciprocating chillers, CFC-114 centrifugal chillers, and CFC-500 centrifugal chillers;

      	Industrial Process Refrigeration:  Refrigeration systems used in the chemical, pharmaceutical, petrochemical, and other manufacturing and food processing industries.  These include both packaged and built-up systems used in manufacturing processes as well as extra-large ice makers;

      	Ice Skating Rinks:  There are several different types of ice rink systems, but all are grouped into one end use in this analysis;

      	Ice Makers:  These include small, medium, and large ice makers used by a number of entities, including restaurants and hotels;


      	Household and Light Commercial Air Conditioning:  These include window units, packaged terminal air conditioners, central air conditioners, direct expansion commercial air conditioners, and heat pumps used for space cooling and heating in homes and businesses.  All systems in this end use currently use HCFC-22;

      	Refrigerated Appliances:  These include household refrigerators and freezers, water coolers, vending machines, and dehumidifiers.  Most currently use CFC-12 as a refrigerant, but there are some CFC-502 units in existence;

      	Refrigerated Transport Systems:  These include systems designed to refrigerate materials in transit.  They are found in trucks, trains, and ships; and

      	Mobile Air Conditioning:  This end use includes systems installed in motor vehicles for the comfort of occupants.

      Industry has invested heavily in the search for suitable alternative refrigerants that exhibit the favorable characteristics of the controlled substances, but that do not contribute to stratospheric ozone depletion or global warming.  The hydrochlorofluorocarbons (HCFCs) and hydrofluorocarbons (HFCs) have received the most attention, along with expanded use of traditional refrigerants such as ammonia and hydrocarbons.  In many cases the most promising solution appears to be a blend of refrigerants.  Despite great progress, the suitability of many replacements is still uncertain.  Toxicity concerns, high ozone depletion potentials, or flammability may limit the ultimate attractiveness of potential substitutes in many applications when testing is completed.  The 1991 report by UNEP's Refrigeration, Air Conditioning, and Heat Pumps Technical Options Committee contains detailed information about the status of alternative refrigerants in various applications.

      EPA believes that HCFCs have an important role to play as transitional refrigerants, both in retrofit applications and in new equipment.  HCFCs have the disadvantage that they do contribute to the destruction of stratospheric ozone, although to a much lesser extent than CFCs.  HFCs have ozone depletion potentials (ODPs) of zero, but their practical application is years away in many cases.  The interim use of HCFCs, until such time as safer alternatives are available, will allow industry to move away from CFC refrigerants more rapidly.  EPA believes that this approach will have greater environmental benefits than permitting the continued use of CFCs until equipment that uses HFCs becomes available.

2.2	ALTERNATIVE REFRIGERANTS

2.2.1	HCFCs

      HCFCs are chemically similar to CFCs except that they contain hydrogen in addition to chlorine and fluorine.  HCFC-22 has been used as a refrigerant for many years.  It is the primary refrigerant used in small to medium sized air conditioners and chillers, and has found increasing application in medium temperature retail food refrigeration systems.  HCFC-123 holds promise as the primary retrofit option for CFC-11 in low pressure centrifugal chillers.  HCFC-124 has potential applications in blends for many types of refrigeration equipment and, as a pure refrigerant, has been suggested as a replacement for CFC-114 in high pressure centrifugal chillers.

      Because they contain hydrogen, the HCFCs break down more easily in the atmosphere, and therefore have lower ODPs.  Their global warming potentials are lower than either the CFCs or the HFCs.  Production of HCFCs is controlled by the Clean Air Act Amendments of 1990.  Beginning in 2015, it will be unlawful to produce HCFCs for uses other than in equipment manufactured before 2020.  Beginning in 2030, it will be unlawful to produce HCFCs, except for certain purposes that are exempted.

      Because their thermophysical properties are, in many cases, similar to CFCs, equipment designed to use CFCs can sometimes be retrofitted to operate with HCFCs.  As noted above, EPA believes that HCFCs will play an important role as transitional refrigerants.  There are clear environmental benefits to be gained by allowing their use until better substitutes are developed.  
      
2.2.2	HFCs

      HFCs do not contain chlorine and do not contribute to destruction of stratospheric ozone.  Many, however, have relatively high greenhouse gas potentials and contribute to global warming.  Although a few HFCs have been in use for some time (HFC-152a is a component in the azeotropic blend R-500, used in smaller tonnage reciprocating equipment and large tonnage centrifugal equipment), the potential for HFCs as a replacement for CFCs has only recently been investigated.  HFC-134a and HFC-152a hold the most promise as single-substance replacements for Class I and Class II refrigerants.  HFC-143a and HFC-125, as parts of blends, have been considered as substitutes for R-502 in low temperature refrigeration applications.  HFC-227ea has possible applications as a refrigerant in mixtures used in high pressure centrifugal chillers.  HFC-23 and HFC-32, when used in blends with other chemicals, have been proposed as a potential substitute for certain HCFC-22, CFC-13, and R-503 end-uses.

2.2.3	Hydrocarbons

      Propane, ethane, propylene, and, to some extent, butane are used as refrigerants in specialized industrial applications, primarily oil refineries and chemical plants, where they are frequently available as part of the process stream and where their use contributes only slightly to the incremental risk of fire or explosion.  These industrial systems are specially designed to meet rigid requirements for reliability, durability, and safety.  ASHRAE Standard 15, Safety Code for Mechanical Refrigeration, and Standard 34, Refrigerants, are incorporated into building codes in most of the U.S.  These standards limit the use of flammable refrigerants in many applications.  Perfluoropropane and other perfluorocarbons (PFCs) have potential applications in the refrigeration sector as well.

2.2.4	Ammonia

      Ammonia has been used as a refrigerant in vapor compression cycles for more than 100 years.  It is by far the refrigerant of choice in many food processing and industrial applications, such as the meat packing, chicken processing, dairy, frozen juice, brewery, and cold storage industries.  It is also widely used to refrigerate holds in fishing vessels.  Industrial process refrigeration equipment uses rotary screw or reciprocating compressors.  Ammonia's main application is when moderate to low temperatures are required.  Ammonia has a characteristic pungent odor, excellent refrigerant properties, is low in cost, and has no long-term environmental drawbacks.  It is moderately flammable and toxic, however, but it is not a cumulative poison.  Occupational Safety and Health Administration (OSHA) standards specify that the maximum allowable concentration of ammonia for an eight-hour working exposure is 50 ppm.

2.2.5	Chlorine

      Chlorine has been proposed as a Class I substitute refrigerant for use in chlorine liquification, a processing step in the manufacture of the chemical.  When chilled below its boiling point, chlorine can be stored as a liquid at atmospheric pressure, a method that for safety reasons is much preferred to storing the chemical as a pressured gas at ambient temperatures.  Compatibility of the refrigerant with liquid chlorine is critical because of chlorine's high reactivity; CFC-12 has been widely used because it is nonreactive with chlorine.

      Chlorine compressors would be specialized units made to resist chemical attack by liquid and gaseous chlorine.  Because a chlorine refrigeration system would use part of the process stream as the refrigerant, the proposed use of chlorine as a refrigerant is analogous to that of hydrocarbon refrigerants in the oil and gas industry.  EPA has determined that, if the refrigeration system is placed so that any leakage or losses of chlorine would be contained and neutralized by the process safety mechanisms, chlorine can be used safely in these specialized applications.

2.2.6	Potential Substitutes that Manufacturers have Submitted as Confidential Business Information


      A few substitutes that may be used in certain refrigeration and air conditioning end uses have been submitted by their manufacturers as confidential business information.  As a result, the identity of these substitutes are not revealed in this analysis.  The results of the risk analyses for these compounds are presented under the following identities:  R200b through R200i and HCFC/HFC/Fluoroalkane Blend A.

2.3	ALTERNATIVE TECHNOLOGIES

2.3.1	  Absorption Refrigeration Systems

      Absorption refrigeration systems are the only major existing alternative to systems based on vapor compression cycles.  Ammonia is also used in absorption refrigeration and air conditioning systems.  Small ammonia refrigeration units are popular in recreational vehicles and in some household applications as they need no electrically driven mechanical compressor, relying instead on a propane flame as an energy source.  Small refrigerators using absorption technology are produced for use in hotel rooms, where the focus is on their silent operation rather than the lack of a suitable supply of electricity.  Small absorption type systems use hydrogen to maintain a system pressure high enough to allow the ammonia refrigerant to evaporate at low pressure and temperature (and condense at room temperature), and are constructed to withstand high internal operating pressures.  The absorption mechanism itself is a sealed unit, which usually needs no servicing over its operating life.

      Commercial ammonia absorption systems are used for air conditioning comfort cooling, particularly in instances where waste heat is available.  As with all chillers, these produce chilled water, which is circulated to the space being cooled.  Lithium bromide is also used in commercial absorption systems, where it serves as an absorbent.  Such systems operate at very low pressure to allow water to act as a refrigerant.  Lithium bromide is a relatively nontoxic, nonflammable, nonexplosive, chemically stable compound.  Both types of absorption chiller systems have traditionally been competitors to electrically driven CFC chillers.

2.3.2   Evaporative Cooling

     Evaporative cooling is another technology that could be an alternative to CFC- and HCFC-based vapor compression systems in some applications.  Three types of EC systems exist: direct, indirect, and desiccant-assisted.  In direct EC, the airstream temperature is decreased by bringing the air into direct contact with liquid water.  Because it humidifies when it cools, direct evaporative cooling can give customer comfort only where the ambient air is hot and dry, such as in the southwestern parts of the United States.  Indirect systems use a heat exchanger between the water and the airstream, and desiccant-assisted systems use a desiccant to dry the air before it is exposed to water.  These new developments have greatly expanded the applicability of EC to include the entire U.S.

      Contrary to vapor compression systems, evaporative coolers do not require compressors and consequently can offer considerable energy savings.  Evaporative coolers have been used for many years to provide cooling for comfort and for a variety of industrial and agricultural applications.  They are energy-efficient, inexpensive to purchase, and have low operating costs.




	3.  TOXICITY REFERENCE VALUES FOR SUBSTITUTES


      To assess potential health risks from exposure to substitutes for ozone-depleting substances (ODSs) in the refrigerants sector, EPA identified or developed toxicity reference values for each substitute.  These values include occupational exposure limits, reference concentrations, and cancer slope factors.  The toxicity reference values were compared to or combined with predicted exposure concentrations to estimate risks to workers, consumers, and the general population, as described in subsequent chapters of this report.

      This chapter first discusses EPA's general approach for identifying or developing the toxicity reference values for substitutes examined in the SNAP background documents.  Section 3.1 discusses occupational exposure limits, and Section 3.2 discusses general population toxicity reference values.  Both of these sections pertain to the inhalation route of exposure.

      This chapter concludes by presenting the most recent toxicity reference values available for the substitutes covered in the refrigerants sector at the time of publication of this background document (Exhibit 3-1).  The list of substitutes presented in Exhibit 3-1 is not all inclusive; however, the Agency believes that the substitutes selected for this review are representative of the types of agents that could replace ODSs in this sector.  The Agency has also included in this screen replacement chemicals that may have high toxicity potential to ensure that the assumptions used in the subsequent hazard assessments are valid and conservative.

3.1	OCCUPATIONAL (INHALATION) EXPOSURE LIMITS

      Allowable occupational exposure limits (OELs) for continuous and short-term exposure, denoted as the workplace guidance level (WGL) and emergency guidance level (EGL), respectively, are used to examine potential risks to workers.  Whenever available, the WGLs and EGLs are based on OELs developed by the Occupational Safety and Health Administration (OSHA), the National Institute for Occupational Safety and Health (NIOSH), the American Conference of Governmental Industrial Hygienists (ACGIH), or the American Industrial Hygiene Association (AIHA).  

      The Occupational Safety and Health Administration (OSHA) is a governmental agency within the Department of Labor that sets enforceable occupational standards.  Limits set by OSHA include permissible exposure limits (PELs) and short-term exposure limits (STELs).  The PEL is an 8-hour time-weighted average (TWA) airborne exposure limit that is designed, to the extent feasible, to reduce a significant risk of material impairment of health or functional capacity associated with exposure to a hazardous substance.  The STEL is the employee's 15-minute TWA exposure that should not be exceeded at any time during a work day.  It is designed to protect workers from experiencing adverse acute health effects such as respiratory irritation, eye irritation, and narcosis.  It must be noted that because OSHA's exposure limits are legally enforceable, OSHA must consider a variety of factors including technologic and economic feasibility in establishing them.  Thus, the limits set by OSHA do not solely reflect the toxicological properties of the material.
      
Exhibit 3-1

Exhibit 3-1 (cont.)

      The National Institute for Occupational Safety and Health (NIOSH), a governmental institute under the jurisdiction of the Department of Health and Human Services, develops recommended exposure limits (RELs) that are non-enforceable.  Unlike OSHA, NIOSH is not constrained by issues of technical or economic feasibility or significance of risk when deriving RELs.  NIOSH also estimates concentrations of chemicals that are immediately dangerous to life and health (IDLH), which represents "the maximum concentration from which. . . one could escape within 30 minutes . . . without experiencing any escape-impairing (e.g., severe eye irritation) or irreversible health effects" (NIOSH 1990).

      The American Conference of Governmental Industrial Hygienists (ACGIH) is a non-governmental limit setting body.  ACGIH's non-enforceable TWA threshold limit values (TLV-TWAs) are similar to OSHA's PELs.  ACGIH does not formally analyze for technical feasibility, but there appears to be some implicit recognition of feasibility since TLVs are set to be protective for most workers but not so low as to be "unduly restrictive."  ACGIH TLVs provided much of the basis for OSHA's 1989 air contaminant standards.

      The American Industrial Hygiene Association (AIHA) is also non-governmental.   The AIHA workplace environmental exposure levels (WEELs) are like the TLVs in that the justification of a particular WEEL is based predominantly on toxicological information rather than feasibility.  However, it is likely that feasibility is considered implicitly since the WEELs are limits that are protective and practical.

      The long-term occupational exposure limits (WGLs) used for the SNAP risk screening assessments are based on OSHA PELs, NIOSH RELs, ACGIH TLV-TWAs, or AIHA WEELs, if available.  The short-term occupational exposure limits (EGLs) are based on OSHA STELs or NIOSH IDLHs, if available.  EPA recognizes that in an industrial setting OSHA STELs are typically much lower than IDLHs and other emergency guidance levels, and that exposure at the OSHA STEL is not a situation requiring emergency corrective action.  Therefore, using OSHA STELs as EGLs is a conservative way to screen risks from short-term exposure to chemicals.

      If OSHA, ACGIH, NIOSH, or AIHA OELs were not available for a substitute, and WGLs or EGLs were considered important for the risk screening assessment, they were estimated by EPA.  Although the estimated WGLs are intended to be analogous to OSHA PELs, they are different in that PELs are developed by a regulatory process and incorporate complex decision criteria such as technological feasibility.  In estimating WGLs, EPA took into consideration the characteristics of workers who are healthier and younger than the general population, and experience intermittent rather than continuous exposure.  Where appropriate, WGLs were estimated to be ten times greater than the suggested reference concentration (see Section 3.2.1).  If the resulting estimate was greater than the 1,000 ppm TLV-TWA set by ACGIH (1986) ". . . as a guide for good hygiene practice for vapors of low toxicity" (i.e., the "good housekeeping" limit), then the latter was used for the risk screening analysis.

      The EGL estimate is similar to the NIOSH IDLH (i.e., the maximum concentration from which one could escape within 30 minutes without experiencing escape-impairing or irreversible health effects).  It also resembles the emergency exposure guidance level (EEGL) and the short-term public emergency guidance level (SPEGL) suggested by NRC (1986) and the emergency response planning guideline suggested by AIHA (1989).  In instances when a substitute has an OEL but the available toxicological data were not sufficient to develop a STEL, the ACGIH excursion limit recommendation as stated below was used to estimate the EGL:

      "Excursions in worker exposure levels may exceed 3 times the TLV-TWA for no more         than a total of 30 minutes during a work-day, and under no circumstances should they        exceed 5 times the TLV-TWA, provided that the TLV-TWA is not exceeded" (ACGIH,     1992).

Thus, EGLs can range from three to five times the OEL.

In the event of a fire, explosion, or catastrophic emission, the EGL estimates for most of the HCFCs and HFCs and several other chemicals in this sector are based on the lowest-observed-adverse-effect level (LOAEL) or on the no-observed-adverse-effect level (NOAEL) for cardiotoxicity in epinephrine-sensitized dogs because the primary concern for acute high-level exposure to these compounds is cardiotoxicity.  Cardiac sensitization is of particular interest in these acute, episodic exposures because human heart arrhythmias and sudden death resulting from overexposure to halons, CFCs, and other halogenated hydrocarbons have been documented in workplace settings and in volatile substance abuse (e.g., glue sniffing).  As defined by the Agency, cardiotoxicity is the ability of a compound to cause serious and sometimes fatal cardiac arrhythmia, and is best evaluated by exposing the appropriate species, usually the dog, to an agent by inhalation in the presence of epinephrine (Rubenstein and Bellin 1993).  The Agency is using these values in the SNAP risk screen to ensure protection of the worker population.  

      The protocols used to determine the cardiotoxic LOAEL and NOAEL concentrations for each agent are conservative; they entail measurement of cardiotoxic effects in animals made more sensitive to these effects by the administration of epinephrine.  The administered doses of epinephrine are just below the concentrations at which epinephrine alone would cause cardiotoxicity in the experimental animal, and are approximately ten times greater than the concentration a human being would likely secrete under stress.  Thus, the estimated LOAELs and NOAELs would be conservative for humans even in high-stress situations.

      EPA is adopting OSHA standard (29 CFR 1910, subpart L) section 1910.162 arising from the cardiac sensitization induced by Halon 1301, and is basing its risk assessment decisions for the substitutes that induce cardiotoxicity upon this standard.  Under safe use conditions for Halon 1301 (29 CFR 1910.162), the exposure limits, based upon the length of time it takes to evacuate, can be summarized as follows:

      	Where egress takes longer than 30 seconds but less than one minute, the employer shall not use the agent in a concentration greater than its LOAEL for cardiotoxicity (e.g., 100,000 ppm for Halon 1301).

      	Agent concentrations greater than the LOAEL for cardiotoxicity are only permitted in areas not normally occupied by employees provided that any employee in the area can escape within 30 seconds.  The employer shall assure that no unprotected employees enter the area during agent discharge.

      	Where egress from an area cannot be accomplished within one minute, the employer shall not use this agent in concentrations exceeding its NOAEL for cardiotoxicity (e.g., 75,000 ppm for Halon 1301).

      In addition, EPA recognizes that agents should not be used at a concentration that would significantly displace oxygen in the lungs.  Since most of the refrigerant substitutes are gaseous, heavier-than-air compounds, following a leak or catastrophic emission, they may tend to pool near the ground, (i.e. in the breathing zone).  These agents are, in the main, colorless with minimal odor and little toxicity or irritant effect, therefore, they can lead to asphyxiation by oxygen displacement, if the unwary person inadvertently walked into an area of oxygen depletion.


      Under the SNAP risk screen, EPA is only intending to set conditions for the safe use of refrigerant substitutes in the workplace until OSHA incorporates specific language addressing gaseous agents into the OSHA regulation.  These general conditions will no longer apply once OSHA establishes applicable work place requirements.

      Where appropriate, the Agency also based its analysis of short-term exposures to consumers and the general population on these cardiotoxic levels.

3.2	GENERAL POPULATION/CONSUMER (INHALATION) TOXICITY REFERENCE VALUES
      
      For the refrigerants sector, risks to the general population and consumers were screened for short-term and/or long-term exposure scenarios.  Risks from short-term exposure were screened using the EGLs and cardiotoxic LOAELs and NOAELs discussed in the previous section.  This section discusses EPA's general approach for identifying and developing toxicity reference values applicable to chronic (i.e., lifetime) exposure:  verified or interim reference concentrations (RfCs) for noncarcinogens and cancer slope factors (SFs) for carcinogens.

3.2.1	Reference Concentrations (RfCs)

      When possible, the RfC verified by the Agency's RfD/RfC Work Group, whether or not it has been listed on the Agency's Integrated Risk Information System (IRIS) database, was used to assess the risk from long-term exposure to ODS substitutes in the SNAP background assessments.  The RfC is designed to protect the general population against adverse systemic (i.e., noncancer) effects, and is defined as:

      "An estimate (with uncertainty spanning perhaps an order of magnitude) of a daily exposure to the human population (including sensitive subgroups) that is likely to be without appreciable risk of deleterious effects during a lifetime.  The inhalation reference concentration is for continuous inhalation exposures and is appropriately expressed in units of mg/m[3]" (U.S. EPA 1990).

      However, when verified RfCs were not available, interim RfCs were used to evaluate risks associated with chronic exposure to ODS (listed in decreasing order of preference):

      	Currently undergoing verification by the Work Group;

      	Not yet undergoing verification but contained in the Agency's Health Effects Assessment Summary Tables (HEAST; U.S. EPA 1992);

      	Estimated from other data for the substitute; or

      	Estimated based on surrogate limits.

The remainder of this section discusses the interim RfCs derived from other data for the substitute or estimated based on surrogate limits.

RfCs Estimated From Other Data


      In the absence of a verified RfC or an RfC pending verification by the Agency's RfD/RfC Work Group, EPA calculated an interim RfC using the methods specified in EPA's Interim Methods for Development of Inhalation Reference Concentrations (U.S. EPA 1990).  Since adequate human data were not available to calculate RfCs for the proposed substitutes that did not already have RfCs, data obtained from animal inhalation toxicity experiments served as the basis for estimating the RfCs.  This approach was adopted in developing an interim RfC for HFC-143a.  For the purposes of the SNAP risk screen, the interim RfC that has been estimated for HFC-143a was based on limited animal data.  The Agency's RfD/RfC Work Group reviewed the quality and quantity of the animal data and determined that the available information was inadequate to derive an RfC.  The compound's interim RfC is, therefore, currently listed as "not verifiable".  In general, the verification process requires the existence of a sizeable body of information.  Ideally, this should include two well-conducted chronic inhalation studies, one reproductive study, and two developmental toxicity studies in different species.  The term "not verifiable" does not imply that if only limited data are available that these data are technically deficient.  The Agency will develop the verified RfC for HFC-143a when these data become available; use of the interim value should be restricted solely to the SNAP risk screen.

      If adequate animal inhalation toxicity data were unavailable for a substitute, the interim general population RfC was estimated by extrapolation from oral references doses (RfDs) or oral NOAELs or LOAELs.  According to the Interim Methods for Development of Reference Concentrations, such route-to-route extrapolation can be performed if pharmacokinetic data, including measurements of absorption efficiency and comparative excretion data (when the associated metabolic pathways are equivalent) are available for both routes of interest, and comparative systemic toxicity data indicate equivalent effects by each route of interest.  However, the Interim Methods stipulate that oral data should not be used to estimate inhalation exposure limits when the following conditions exist:

      (1)	When groups of chemicals are expected to have different toxicity by the two routes of exposure (e.g., metals, irritants, and sensitizers);

      (2)	When a first-pass effect is expected by the liver, or when the pulmonary system was not adequately studied in the oral studies;

      (3)	When a pulmonary effect is established, but dosimetry comparison cannot be clearly established between the two routes; and

      (4)	When short-term inhalation studies or in vitro studies indicate potential portal-of-entry effects at the lung, but studies themselves are not adequate for RfC development.

The appropriate data to assess these conditions were not available for substitutes requiring a route-to-route extrapolation.  Therefore, the interim RfCs that were estimated by extrapolation from oral data would most likely be considered "not verifiable" by the Agency's RfD/RfC Work Group.  However, in the absence of adequate data and in need of a chronic reference toxicity value for the purpose of the SNAP risk screens, extrapolation from oral data was nevertheless performed according to the following procedures:

      (1)	Oral RfDs (expressed in mg/kg/day) were converted to inhalation RfCs (expressed in mg/m[3]) by multiplying the RfD by the default assumption for average human body weight (70 kg) and dividing by the default assumption for average human breathing rate (20 m[3]/day) (EPA 1988).

      (2)	Alternatively, in the absence of adequate experimental inhalation toxicity data, RfCs for some substitutes were estimated by converting the OEL for that substitute (or a suitable surrogate, as discussed below) to a general population exposure limit.  For example, the aliphatic hydrocarbon isobutane is best represented by pentane.  The OSHA PEL for pentane is 1,800 mg/m[3].  This value is based on an 8-hour TWA worker exposure over a 40-hour work week.  To extrapolate to a threshold for long-term continuous general population exposure, the PEL was multiplied by a factor of 10 m[3]/20 m[3] (assuming that 10 m[3] are inhaled in an 8-hour day by a healthy worker, extrapolated to 20 m[3] in a 24-hour day by the general population) (Jarabek and Hasselblad 1991), and then divided by an uncertainty factor of 1,000 to account for sensitive subpopulations and other uncertainties.


Interim RfCs Based on Surrogate Limits

      For some of the substitutes, there are neither RfCs nor adequate experimental inhalation toxicity data, oral RfDs, oral data, or OELs from which to estimate an RfC.  In these cases, exposure limits for a surrogate such as HFC-134a, the surrogate for several HFCs in this sector, was used.  Surrogates were chosen by structure-activity relationship (SAR) analysis to provide the worst-case/most conservative estimate for the exposure limit.  SAR is used to predict the biological activity of an untested compound based on an examination of analogous chemicals with well-documented health or environmental effects data.

      SAR has been an important tool in guiding the development of new chemicals and drugs for many years.  EPA has used SAR in assessing new chemicals prior to their manufacture or importation under the Premanufacture Notice Program.  Also, the pharmaceutical industry has used SAR to assist in the design and synthesis of new drugs.  

3.2.2	Cancer Slope Factors

      The cancer slope factor (SF) is used to estimate the upper-bound risk for cancer.  Because chemicals that produce cancer are assumed not to show a threshold for their effects, the observed threshold level (i.e., the NOAEL) is not used in the calculation of the cancer SF.  Rather, the slope of the dose-response for the effect is used to extrapolate levels at which the risk of the effect is small.  It has been argued that the dose-response function for carcinogenicity could be linear and that it is unlikely to exceed linearity in the low-dose region.  Thus the model chosen to extrapolate low-dose effects from the much higher doses at which effects are observed assumes low-dose linearity for cancer production.  EPA guidelines (U.S. EPA 1986) recommend that, in absence of adequate information to the contrary, the linearized multistage procedure should be used for the extrapolation.  This model expresses upper confidence limits on cancer risk as a linear function of dose in the low-dose range.  Cancer SFs are estimated by fitting the model to experimental animal carcinogenicity data using the maximum likelihood method.  In addition, an upper bound on the dose-response curve is calculated, reflecting uncertainty of extrapolating the curve to low doses at which human exposures are anticipated to occur.  This upper bound can be considered to represent the largest reasonable linear extrapolation to low doses consistent with the data.  The 95 percent statistical upper limit on the linear term, q1, is referred to as the q1*, or the SF.

      The SNAP cancer risk estimates were based on cancer SFs that have been verified by EPA's Cancer Risk Assessment Verification Endeavor (CRAVE) Work Group and are contained in IRIS.  If these were unavailable, a second choice was SFs that have not yet been verified but are contained in HEAST (U.S. EPA 1992).  The interim SFs for HCFC-22 and HCFC-123 were estimated using raw data from recently completed chronic inhalation cancer bioassays, according to the methods recommended in the EPA Risk Assessment Guidelines (U.S. EPA 1986).
      
                                  REFERENCES


ACGIH.  1986.  Documentation of the Threshold Limit Values and Biological Exposure Indices.  Fifth Edition.  

ACGIH.  1992.  1992-1993 Threshold Limit Values for Chemical Substances and Physical Agents and Biological Exposure Indices.  

AIHA.  1989.  Concepts and Procedures for the Development of Emergency Response Planning Guidelines (ERPGs).  ERPG Committee.  

Jarabek, A.M. and V. Hasselblad.  1991.  Inhalation Reference Concentration Methodology:  Impact of Dosimetric Adjustments and Future Directions Using the Confidence Profile Method.  Presented at 84th Annual Meeting and Exhibition, Vancouver, B.C.  June 16-21, 1991.

NIOSH.  1990.  NIOSH Pocket Guide to Chemical Hazards.  NIOSH Pub. No. 90-117.  US DHHS.

NRC.  1986.  Criteria and Methods for Preparing Emergency Exposure Guidance Level (EEGL), Short-term Public Emergency Guidance Level (SPEGL), and Continuous Exposure Guidance Level (CEGL) Documents.  Committee on Toxicology.  National Academy Press, Washington, DC.

Rubenstein, R. and J.S. Bellin.  1993.  Short Duration High-Level Exposure to Halon Substitutes: 
      Potential Cardiovascular Effects.  Presented at Halon Alternatives Technical Working Conference, Albuquerque, NM, May 1113, 1993.

U.S. EPA.  1992.  Health Effects Assessment Summary Tables.  Annual Update.  Environmental Criteria and Assessment Office, Office of Health and Environmental Assessment.  Cincinnati, OH.  OHEA ECAO-CIN-821.

U.S. EPA.  1990.  Interim Methods for Development of Inhalation Reference Concentrations.  Review Draft.  August 1990.  Environmental Criteria and Assessment Office, Office of Health and Environmental Assessment, Research Triangle Park, NC.  EPA 600/8-90/066A.

U.S. EPA.  1988.  Recommendations for and Documentation of Biological Values for Use in Risk Assessment.  February 1988.  Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati, OH.

U.S. EPA.  1986.  The Risk Assessment Guidelines of 1986.  Office of Health and Environmental Assessment.  Washington, DC.  EPA/600/8-87/045.



	4.  ATMOSPHERICS MODELING


4.1	INTRODUCTION

      EPA has conducted a risk screen of possible alternatives to Class I ozone depleting substances (ODSs) as part of its implementation of Section 612 of the Clean Air Act Amendments of 1990.  Class I substances include chlorofluorocarbons (CFCs), halons, carbon tetrachloride, and methyl chloroform (MCF).  Class I ODSs deplete stratospheric ozone and may contribute to global warming.  These atmospheric changes have potentially serious consequences for human health and the environment.  The depletion of stratospheric ozone increases the amount of biologically active ultraviolet radiation (UV-B) at the earth's surface.  Among the consequences of increased UV-B are the increased incidence of cataracts and skin cancers, damage to plants and animals, and damage to polymers (chemical compounds often used in plastics, paints, and coatings).  This chapter presents estimates of the potential atmospheric and human health effects of the alternatives available to replace Class I compounds in the refrigeration and air conditioning sector.

      The remainder of this chapter is divided into the six following sections:  Section 4.2 presents an overview of the refrigeration and air conditioning sector, including a description of the major end uses; Section 4.3 summarizes the potential replacements for Class I compounds in the refrigeration and air conditioning sector; Section 4.4 describes the baseline for the analysis; Section 4.5 details the modeling framework used to quantify the atmospheric and human health impacts of the alternatives; Section 4.6 presents the results of the analysis; and Section 4.7 presents a qualitative discussion of global warming and global warming potentials.

4.2	DESCRIPTION OF REFRIGERATION AND AIR CONDITIONING END USES

      CFCs are used as refrigerants in numerous refrigeration and air conditioning applications.  For the purposes of this analysis, the refrigeration and air conditioning sector was divided into several end uses to facilitate appropriate comparisons between possible substitutes.  Two general principles define how equipment types were classified into the various end uses:  (1) all equipment types in an end use employ roughly comparable refrigeration technologies (e.g., low and high pressure cold storage warehouse systems are separated ); and (2) all equipment types in an end use serve the same general consumer or industrial purpose (e.g., ice skating rinks and parallel retail food systems form separate end uses).

      The advantage of this end use classification is that substitutes applicable to a given end use will typically be able to be used in all segments of that end use.  A consequence, however, is that some end uses are considerably larger than others.  For example, all high-pressure industrial process refrigeration systems form one end use, while another end use consists only of CFC-12 reciprocating chillers.

      While the principal focus of this analysis is to evaluate alternatives to Class I substances, many end uses in the refrigeration and air conditioning sector primarily use HCFC-22 (a Class II compound).  In order to comprehensively examine the health and environmental effects of substitutes throughout this sector, HCFC end uses were analyzed in the same manner as CFC end uses. 

      The end use categories are as follows:

      	Retail Food Systems:  These include stand alone refrigeration cases found in small markets, convenience stores, restaurants and other food establishments, large parallel systems found in supermarkets, and HCFC-22 systems found in a wide variety of retail and service establishments.


      	Cold Storage Systems:  These systems refrigerate warehouses of varying sizes and include high pressure and low pressure systems.

      	Chillers:  These are large commercial air conditioning systems used to cool office buildings, shopping malls, and other large indoor spaces.  End uses modeled are CFC-11 (low pressure) chillers, CFC-12 centrifugal chillers, CFC-12 reciprocating chillers, CFC-114 centrifugal chillers, and CFC-500 centrifugal chillers.

      	Industrial Process Refrigeration:  These include both packaged and built-up systems used in manufacturing processes as well as extra-large ice makers.  High and low pressure systems are modeled as separate end uses, although the overwhelming majority of industrial process systems are high pressure units.

      	Ice Skating Rinks:  There are several different types of ice rink systems, but all are grouped into one end use in this analysis.

      	Ice Makers:  These include small, medium, and large ice makers used by a number of entities, including restaurants and hotels.

      	Household and Light Commercial Air Conditioning:  These include window units, unitary air conditioning systems, and heat pumps.  All systems in this end use currently use HCFC-22.

      	Refrigerated Appliances:  These include household refrigerators and freezers, water coolers, vending machines, and dehumidifiers.  Most currently use CFC-12 as a refrigerant, but there are some CFC-502 units in existence.

      	Refrigerated Transport Systems:  These include systems designed to refrigerate materials in transit.  They are found in trucks, trains, and ships.

      	Mobile Air Conditioning:  This end use includes systems installed in motor vehicles for the comfort of occupants.

      Exhibits 4-1 and 4-2 display the consumption of ODSs in the refrigeration and air conditioning sectors as a fraction of total U.S. ODS consumption.  The consumption amounts presented for ODSs in Exhibit 4-1 are weighted according to the chemicals' ozone depletion potentials (ODPs).

ex 4-1  &  4-2

4.3	CONTROL OPTIONS

      The alternatives to CFC refrigerants with the widest applicability are HCFC and HFC refrigerants, although other alternatives such as ammonia, propane, and butane are possibilities in some end uses.  Whether or not a chemical substitute is feasible in a given end use depends on a number of factors, including the required operating temperatures and pressures that need to be attained in that end use.

      Substitutes can replace CFCs in both new and existing equipment in the refrigeration and air conditioning sector.  In general, existing equipment will have to be modified  or retrofitted  to be capable of using the substitutes.  In this analysis, the effects of substitutes in new equipment were distinguished from the effects of the same substitutes as retrofits.  This allows one to study the implications of substitutes in new equipment and existing equipment independently from one another.

      Additional information on possible substitutes, their availability, their potential end use market penetration, and their impact on energy consumption is presented in Section 4.6 with the results.

4.4	THE BASELINE FOR THE ANALYSIS

      The baseline provides a benchmark against which the effects of possible substitution scenarios can be measured.  An appropriate baseline incorporates assumptions regarding chemical emissions in the particular end use being examined, but at the same time it must account for all other sources of emissions of relevant compounds both in the U.S. and in the rest of the world (ROW).  A global baseline is required because the incremental atmospheric effects of emissions in an end use depend on the quantity emitted from all other end uses worldwide.

      The baseline is described in three steps.  First, the uniform assumptions made throughout the analysis are specified.  Second, the assumptions that are applied to all of the end uses (including end uses in other sectors such as solvents, fire extinguishing, foams, and sterilants) other than the one in which multiple substitutes are being compared are specified.  Finally, the baseline assumptions used within an end use when evaluating multiple substitutes are specified.

4.4.1	Uniform Assumptions Made Throughout the Analysis

      This section describes the uniform phaseout and recycling assumptions made throughout the analysis.  The U.S. phaseout assumptions are listed in Exhibit 4-3.  The exhibit lists the last year in which a given chemical can be produced for use in either new equipment or existing equipment.  Other developed countries (Canada, Japan, Australia, New Zealand, and the nations of Western Europe) are assumed to follow the same schedule.  The phaseout of ODSs in developing countries is assumed to lag that of the developed world by a decade.  In addition, this analysis assumes that CFCs and HCFCs are recycled beginning in 1992 during the servicing and disposal of certain refrigeration equipment and that HFCs are likewise recycled starting in 1995.  These recycling options are implemented in both the U.S. and the rest of the world (with a ten-year lag for developing countries).




                                  EXHIBIT 4-3
             PRODUCTION PHASEOUT ASSUMPTIONS FOR THE UNITED STATES

                                       
                               Chemical Compound
                                 Phaseout Date
                                    for New
                                   Equipment
                                 Phaseout Date
                                 for Existing
                                   Equipment
               CFCs, methyl chloroform and carbon tetrachloride
                                       
                                     1996
                                       
                                     1996
                                    halons
                                     1994
                                     1994
                                methyl bromide
                                     2001
                                     2001
                                   HCFC-141b
                                     2003
                                     2003
                             HCFC-22 and HCFC-142b
                                     2010
                                     2020
                                All other HCFCs
                                     2015
                                     2030


4.4.2	 Assumptions Made for All Other End Uses

      This section summarizes the substances industry is assumed to choose to replace CFCs, halons and MCF in all end uses other than the one in which multiple substitutes are being compared.  Although in a few sectors it remains uncertain which substitutes will be chosen, a "best estimate" substitute or combination of substitutes was assumed for each end use under the 1996 phaseout in order to best approximate total worldwide emissions of Class I and Class II substitutes.  These substitution scenarios were developed based on EPA's best judgment of the likely substitutes to be used in each end use under the 1996 phaseout.

4.4.3	 Assumptions Made Within the End Use of Concern

      This section describes the baseline assumptions used within the end use in which multiple substitutes are to be compared.  In this analysis, two different baselines are used to present the results of the substitution scenarios.  Before detailing these baselines, it is important to note that the choice of the baseline affects the presentation, and not the substance, of the results.  Under either of the baselines described below, the relative differences between the substitutes with respect to ozone depletion will be the same.  The absolute results presented for each substitute will, however, differ between the two baselines.

      Each of the baselines used with the end use of concern are discussed below.  One is a "no-risk baseline"; the other is a "no-substitution" baseline.  These two baselines bound the range of health effects, since the first examines impacts from a risk-free scenario and the second presents effects assuming continued use of all CFCs.

4.4.3.1	Ideal Substitute Baseline

      The "ideal" substitute is a hypothetical alternative that has an ozone depletion potential (ODP) of zero.  The concept of the "ideal" substitute was developed to provide a "no-risk" point of reference from which actual alternatives could be compared against one another.

      In this case, the difference between substitute and baseline scenarios isolates the effects of implementing the substitute in the U.S. relative to an ideal or "no-risk" substitute.  As presented schematically in Exhibit 4-4, the reported effects of a substitute are the difference between the estimated effects of the substitution scenario and the "ideal" substitute baseline.  The "ideal" substitute baseline consists of the "ideal" substitute in the end use of concern in the U.S. only and the "best estimate" substitute in all other end uses  both in the U.S. and in the rest of the world.  The substitution scenario consists of the substitute in the end use of concern in the U.S., and the "best estimate" scenario in all other end uses.  
      
      In the refrigeration and air conditioning sector, special care also has to be taken to distinguish between new and existing equipment, especially regarding baseline issues.  As mentioned previously, this analysis examines separately the effects of substitutes in new equipment and in retrofits (existing equipment).  As a result, for new equipment scenarios, the analysis measures the impact of replacing the ideal substitute with the substitution scenario.  The retrofit scenarios assess the impact of replacing the ideal substitute in existing equipment with the substitution scenario.


                                  EXHIBIT 4-4
               SUMMARY OF METHODOLOGY USED IN ATMOSPHERICS RUNS
               TO CALCULATE IMPACTS RELATIVE TO "IDEAL" BASELINE
                                       
                                       

	United States

	Rest-of-World

	End Use
	of Concern
	All Other
	End Uses

	End Use
	of Concern
	All Other
	End Uses

BASELINE SCENARIO
	"Ideal"
	Substitute
	Best
	Estimate
	Substitute

	Best
	Estimate
	Substitute
	Best
	Estimate
	Substitute

SUBSTITUTION SCENARIO

	Substitute

	Best
	Estimate
	Substitute

	Best
	Estimate
	Substitute
	Best
	Estimate
	Substitute
                                       


      The model includes consideration of speed of market penetration.  In all analyses, the "ideal" substitute fully penetrates the new equipment market in 1992 and begins to penetrate the existing equipment market in 1992, reaching a maximum market penetration in 2000.  The slower rate of penetration and the smaller final market share for retrofits reflects the behavior that would be expected in practice for most possible retrofit scenarios.  As a result, a substitute that has low ODP but only penetrates that market slowly shows adverse health effects that are an artifact of the modeling, since until the time the substitute becomes available, Class I compounds must continue to be used.

      The advantage of the "ideal" substitute baseline is that it allows for a "no-risk" point of reference for presentation purposes but accounts for changes in atmospheric conditions caused by activity in other end uses.  It must be emphasized that the "ideal" substitute is only used as a reference point.  In all end uses, the relevant comparisons are those between one substitute and another, not between the substitute and the ideal.

4.4.3.2  No Substitution Baseline

      Under this baseline, the incremental performance of each substitute within an end-use is judged against a no substitution baseline (i.e., continued use of Class I substances within the end use).  This option clearly indicates the reduction in ODP risks associated with replacing the use of Class I compounds with each of the potential substitutes and allows for direct comparisons in risk reduction achievable through each substitute.


      In this case, the difference between substitute and baseline scenarios isolates the effects of implementing the substitute in the U.S. relative to a scenario that assumes the continued use of the Class I compound.  As presented schematically in Exhibit 4-5, the reported effects of a substitute under this baseline are the difference between the estimated effects of the substitution scenario and the no substitution baseline.  The no substitution baseline consists of the continued use of the Class I compound in the end use of concern in the U.S. only and the "best estimate" substitute in all other end uses  both in the U.S. and in the rest of the world.  The substitution scenario consists of the substitute in the end use of concern in the U.S., and the "best estimate" scenario in all other end uses.  


                                  EXHIBIT 4-5
               SUMMARY OF METHODOLOGY USED IN ATMOSPHERICS RUNS
          TO CALCULATE IMPACTS RELATIVE TO "NO SUBSTITUTION" BASELINE
                                       


	United States

	Rest-of-World

	End Use
	of Concern
	All Other
	End Uses

	End Use
	of Concern
	All Other
	End Uses

BASELINE SCENARIO

	Continued
	CFC Use
	Best
	Estimate
	Substitute

	Best
	Estimate
	Substitute
	Best
	Estimate
	Substitute

SUBSTITUTION SCENARIO

	Substitute

	Best
	Estimate
	Substitute

	Best
	Estimate
	Substitute
	Best
	Estimate
	Substitute


      As discussed above, this analysis examines separately the effects of substitutes in new equipment and in retrofits (existing equipment).  As a result, for new equipment scenarios, the analysis measures the impact of phasing out the use of the Class I compound in new equipment in a given end use and replacing it with the substitution scenario.  The retrofit scenarios assess the impact of replacing the use of the Class I compound in existing equipment in a given end use and replacing it with the substitution scenario.

      The intent of the Agency in presenting the resulting of the tradeoff analysis relative to the no substitution baseline is to indicate that all of the substitutes analyzed are substantially less harmful to the ozone layer than continued use of CFCs.  In presenting the results in this fashion, however, EPA is cognizant of the fact that the reduction in ODP risks due to the phaseout of Class I compounds are attributable to the Class I Phaseout Rule and not the Safe Alternatives Program.

4.5	OVERVIEW OF THE MODELING APPROACH

      This section provides a brief description of the Atmospheric and Health Effects Framework (AHEF), the modeling framework used to assess the global environmental impacts of alternatives to Class I substances in the United States.

4.5.1	Chemical Use


      The starting point of an AHEF analysis is to forecast the consumption of the relevant chemical compounds in each of the end uses in each sector.  For U.S. estimates, this task is usually performed by the Vintaging Model, which simulates the chemical requirements of the various equipment types in each end use over their lifetimes.  The Vintaging Model has been used to support several significant Agency rulemakings including the phaseout of Class I and II compounds under Section 604 of the Clean Air Act.  It is documented in Appendix A of EPA's Regulatory Impact Analysis: Compliance with Section 604 of the Clean Air Act for the Phaseout of Ozone Depleting Substances (EPA 1992), which was included in the Phaseout rulemaking docket and was subject to public review and comment.

      For other regions of the world, chemical use estimates are typically derived by applying the London Amendment reduction schedule to estimates of what demand would be in the rest of the world had this phaseout never occurred.

4.5.2	Chemical Emissions

      Emissions of a compound into the atmosphere are not necessarily simultaneous with use.  The delay between use and emission of a compound depends on the kind of equipment in which the compound is employed.  For example, hermetically sealed refrigerators emit CFCs over a very long period of time, whereas packaging foams emit similar compounds in less than one year.  AHEF simulates compound emissions by means of release functions which specify the time delay between use and emissions for each particular end use.

4.5.3	Atmospheric Effects

      The output from the emissions module are used in an atmospheric model, the Atmospheric Stabilization Framework (ASF), that predicts changes in stratospheric ozone associated with changes in atmospheric composition.  ASF computes the globally and annually averaged concentrations of climatically important atmospheric constituents, taking into account various feedbacks between climate parameters and the constituents themselves.  Examples of feedbacks accounted for in ASF are:

      	The dependence of a compound's atmospheric lifetime on column (stratospheric) ozone and temperature;
      	Radiative and chemical feedbacks due to water vapor;
      	Ocean absorption;
      	Atmospheric circulation effects; and
      	Chemical interactions between compounds.

      ASF relies on simple empirical relations based either on observations or on the results of more complex models in order to quantify physical feedbacks.  In particular, the dependence of model ozone on stratospheric chlorine is adjusted so that it predicts a historical ozone depletion equal to that observed for the current atmosphere.  In addition, ASF input parameters are adjusted to give column ozone changes that are consistent with consensus ODP estimates.

      The ASF is a "consensus" model developed by a committee of prominent atmospheric scientists in government, academia, and private consulting firms.  It was designed to approximate the behavior of more sophisticated two and three-dimensional physical models without requiring the computing power needed for physical simulations.  Additional information on the ASF can be found in National Aeronautics and Space Administration Conference Publication 3023, An Assessment Model for Atmospheric Composition (NASA 1988).

4.5.4	Human Health Effects

      The ozone column changes predicted by the ASF are inputs to a health effects model that estimates the number of skin cancer cases and fatalities that result from the scenario modeled.  Estimates of skin cancers include both melanoma and non-melanoma cancers and represent all cases experienced by individuals born in the United States before 2030.


      In order to quantify the human health impacts of ozone depletion in the U.S., the globally averaged ozone depletion predicted by ASF is converted to estimates of depletion by latitude.  This latitudinal variation is estimated using simple regression techniques from a prior two-dimensional modeling study. (Isaksen, 1986).  Results from a separate UV model are in turn used to estimate changes in UV-B intensity associated with the latitudinally-varying ozone depletion. 

      Human health effects are computed by AHEF as a function of UV-B exposure.  The effective exposure of various population cohorts, representing different sexes and age groups, is estimated using appropriate factors that weight exposure at various stages during a person's lifetime.  Empirical "dose-response" relationships for each cohort are then used to predict changes in skin cancer and cataract incidence over time.

      The human health effects model was reviewed in 1987 as part of an EPA Science Advisory Board review of Assessing the Risks of Trace Gases that Can Modify the Stratosphere (EPA, 1987).

4.6	MODEL RESULTS

      This section presents the results of the analysis and lists several considerations that are helpful in interpreting them.

4.6.1	Description of Results

      The results are presented in Exhibits 4-6 through 4-45, one for each end use sector.  Each exhibit contains two tables and two graphs.  Parts A and B of each exhibit present the atmospheric results relative to the "ideal" baseline.  Parts C and D of each exhibit present the results relative to the no substitution baseline.  The table lists the assumptions regarding the availability and replacement potential of each alternative and estimated atmospheric and health effects resulting from those assumptions.  Note that numbers in parenthesis represent decreases from the baseline (e.g., negative numbers).  The graph illustrates the change in cumulative Clx in 2075 for each of the alternatives.

      More specifically, the following information is provided:

Start Date:  The year in which the alternative option is expected to be commercially available (i.e., after all research and development is complete).  A start date of 1992 indicates that the alternative is currently available.

Years to Maximum Market Penetration:  The number of years it would take for an alternative, once it became commercially available, to reach its maximum possible use in the end use market.

Maximum Market Penetration:  The maximum, technically feasible percentage of total ODS consumption in an end use that can be replaced by the alternative.

Cumulative Clx:  This is a measure of the change in the total quantity of chlorine in the stratosphere over time.  More technically, it is the integral of the Clx curve above two parts per billion.

Skin Cancer Cases:  The estimated change in the number of cases of melanoma and non-melanoma skin cancers experienced by individuals in the U.S. population born before 2030.  The estimate of skin cancers includes fatalities that may result from these cancers.

Skin Cancer Fatalities:  The estimated change in the number of fatalities due to melanoma and non-melanoma skin cancers of individuals in the U.S. population born before 2030.  Skin cancer fatalities are a small portion of the total number of skin cancer cases.

4.6.2	Interpreting the Results

      When interpreting the results, one should be aware of the following:

      	For the results presented relative to the ideal baseline, the baseline for the substitution scenarios for new and existing equipment assumes that the "ideal" substitute is fully implemented in new equipment in 1992 and is retrofit in 50% of existing equipment between 1992 and 2000.  As a result, the numbers presented for each alternative in new equipment scenarios account for:  (1) the effects of that alternative; (2) the effects of any continued CFC use occurring in the period after the implementation of the ideal substitute in the baseline and before the implementation of the alternative; and (3) the continued CFC use attributable to the lack of a retrofit.  Accordingly, the later the availability of the alternative, the greater the quantity of continued CFC use that is attributed to that alternative.  Even an ozone-safe alternative can have cancers indirectly attributed to it if it penetrates the market after the "ideal" substitute.

      	Under the ideal baseline, a "no retrofit" scenario (i.e. "ideal" in new equipment and continued CFC use in existing equipment) is included with the retrofit results for each end use to indicate the effects of continued CFC use in existing equipment that are potentially avoidable through retrofitting.  The results of this scenario are helpful in interpreting both retrofit and non-retrofit graphs because they demonstrate the benefits of retrofitting and show how much of the effects in new equipment scenarios can be attributed to the lack of a retrofit.

      	Since different HCFCs have different phaseout dates, the total quantity of substitute chemical ultimately released into the atmosphere may differ depending on the alternative.  For example, the results of scenarios using HCFCs with early phaseout dates reflect the fact that less of the chemical is ultimately used.

      	Blends containing HCFCs are phased out as soon as their earliest constituents are phased out.

      	Modeling reflects the market penetration potential for each substitute.  This means that health effects are dependent on potential market size, so that a low ozone-depleting substitute with large market potential could look worse than a high ozone-depleting substitute with only small uses.

Exhibits 4-6 through 4-45.
80 pages

4.7.	Global Warming 

      There is no doubt that human activity is increasing the concentration of greenhouse gases in the atmosphere.  The sources and industrial sectors contributing these gases are extremely varied.  No single economic sector can be held entirely responsible for the greenhouse effect, nor can a single source, because it contributes only a small amount to the problem, be exempt from accountability.  Additional research is needed to better characterize the effects of greenhouse gas emissions on the environment and to evaluate the benefit of various reduction strategies.  However, as in the case of ozone depletion, there is growing international consensus among scientists and policymakers that immediate efforts are necessary to curb the rate of growth of atmospheric concentrations of these gases.

      Based on total emissions, the major greenhouse gases are carbon dioxide (CO2) and methane (CH4), and it is on these major contributors that EPA has focused much of the effort of its nascent voluntary programs such as Green Lights.  However, a substantial contribution is made in the aggregate by other gases including CFCs, halons, and other halocarbons.  These minor or trace gases are important for two reasons.  First, although these gases are currently emitted in smaller quantities than CO2 and CH4, their effectiveness in influencing climate tends to be greater.  Second, while individual use of these gases may add little to the overall burden, the rising concentration of greenhouse gases in the atmosphere cannot be controlled without addressing the aggregate effect of the trace gases.

      As the transition away from the ozone-depleting compounds proceeds, a key policy goal of the SNAP program in reducing overall risk is to be sure that this shift does not also result in a significant rise in contributions to global warming due to an increased reliance on greenhouse gas substitutes.  Alternatives reviewed under SNAP to date have varying potential to contribute to global warming.  Their contribution depends on many factors evaluated under SNAP, including radiative properties and persistence in the atmosphere.  It is difficult to make relative judgements among those which do contribute to global warming, however, because their contributions, when viewed through the SNAP lens of individual end-uses, are uniformly small.

      Recognizing this, SNAP avoids evaluating the viability of greenhouse gases as substitutes based solely on a single measure, such as an end-use by end-use contribution to actual warming.  Instead, a broader evaluation of a substitute's inherent characteristics, such as its absorption spectra and atmospheric lifetime, is used to help clarify whether a substitute could contribute significantly to the problem relative to other substitutes available.  For example, a compound which might be acceptable in its contribution to global warming in limited uses might become a problem in widespread emissive uses.  Thus, a careful review of the volume likely to be used in a particular end-use, and how much is likely to be emitted from that end-use, also helps.  Finally, because SNAP's primary goal is to facilitate the transition away from the ozone-depleting chemicals, a substitute which is also a significant global warming gas might be allowed with controls if there exist few other alternatives in a particular end-use.

      This section reviews the risk screen for global warming under SNAP.  It includes a discussion of the nature of the global warming problem, general properties of greenhouse gases, and global warming resulting from changes in the energy efficiency of equipment or processes using the alternative chemicals.

4.7.1	The Nature of the Global Warming Problem


      The Earth's stable average temperature is maintained by a radiative balance in which the amount of energy absorbed from the sun is equal to the amount of energy radiated back into space by the Earth and atmosphere.  If a factor is introduced into the atmosphere that changes this balance, the climate must warm or cool until the radiative fluxes are equal and the balance is restored.  A change of this kind is known as a change in radiative forcing (or warming commitment).  Radiative forcing, expressed in watts per square meter (W/m[2]), is a measure of the downward radiation flux received by the climate system.  Concerns about global warming stem from observed increases in the concentrations of certain gases that may be able to alter the climate by causing an increase in radiative forcing.

      A characteristic that these gases share is that they absorb radiation in the infrared (or thermal) part of the electromagnetic spectrum.  Most of the energy that the Earth receives from the sun is in the form of shortwave radiation in the visible and ultraviolet parts of the spectrum.  The energy emitted from the Earth to space is in the form of longwave radiation in the infrared part of the spectrum.  These gases absorb some of the outgoing infrared (or thermal) radiation and re-emit it both up into space and back down to the Earth.  The net result is a decrease in the heat radiation escaping to space and an increase in the heat radiation falling to the Earth (an increase in radiative forcing).  Because this process tends to concentrate heat in the troposphere, it is known as the greenhouse effect.  The gases responsible for the greenhouse effect are known as greenhouse gases.  Greenhouse gases include carbon dioxide, methane, and nitrous oxide, as well as chlorofluorocarbons (CFCs) and some CFC substitutes.

      The greenhouse effect is a natural phenomenon which plays an important role in keeping the Earth warm.  Concerns about the greenhouse effect and global warming arise from indications that human activities have been increasing the concentrations of both naturally occurring and anthropogenic greenhouse gases.  Increasing concentrations of these gases may lead to an enhanced greenhouse effect, which may cause the Earth to warm at a rate much faster than natural rates of climate change.  This accelerated climate change may have a number of adverse consequences including damage to ecosystems, changes in climatic zones, changes in weather patterns, and sea level rise.  Adjusting to these changes quickly may be both difficult and expensive.

4.7.2	Properties of Greenhouse Gases

      All organic compounds that can volatilize into the atmosphere share the property that they absorb energy in the infrared portion of the electromagnetic spectrum.  However, these gases do not all absorb infrared radiation equally well, and thus some have greater impacts than others on global warming.  The following factors determine the global warming impact of emissions of a particular gas:


      	Atmospheric lifetime.  Various physical and chemical processes tend to remove and break down chemicals in the atmosphere.  Atmospheric lifetime is a measure of how long a gas stays in the atmosphere before it is removed by these processes.  The lifetimes of the greenhouse gases are determined by their sources and sinks in the oceans, atmosphere, and biosphere.  In general, the most effective greenhouse gases are those with long lifetimes since the impact of long-lived gases is more persistent than the impact of short-lived gases.  In some cases this persistence may be cause for concern.  With very-long-lived gases it may take thousands of years for concentrations to decline significantly after emissions have stopped.  Thus, the impacts these emissions have are essentially irreversible.

      	Molecular weight.  Emissions are typically measured according to their mass (in kilograms).  However, the increase in atmospheric concentration (by volume) resulting from an emission is proportional to the number of molecules in the emission, not the total mass of the emission.  The number of molecules in an emission of a given mass is inversely proportional to the molecular weight of the gas emitted.  As a result, the molecular weight is needed to determine the increase in concentration resulting from an emission of a given mass.  The lower the molecular weight of the gas, the greater the increase in concentration resulting from a given emission.

      	Radiative forcing per molecule.  This is a measure of the effectiveness of the gas at absorbing infrared radiation emitted by the Earth.  It is proportional to the increase in radiative forcing resulting from a given increase in concentration.

      Of these factors, only molecular weight is an irreducible characteristic of a gas.  Atmospheric lifetime depends on other characteristics such as solubility in water, rates of reaction with other components of the atmosphere, and susceptibility to photolysis.  Radiative forcing per molecule depends on a number of characteristics that are relevant to greenhouse gas behavior.  These include the following:

      	Integrated infrared band strength.  Each kind of molecule absorbs energy at a unique set of wavelengths.  The wavelengths at which a molecule absorbs energy are in patterns of contiguous regions in the electromagnetic spectrum which are referred to as absorption bands.  A molecule's absorption bands together make up its absorption spectrum.  The strength of the gas's infrared absorption spectrum is a measure of how well molecules of the gas absorb radiation at their characteristic wavelengths (i.e., the stronger the absorption band, the more energy that is absorbed by the molecule).

      	Location of infrared absorption bands.  The location of the gas's infrared absorption bands is important for two reasons:  First, the energy radiated from the earth is not evenly distributed across the infrared spectrum.  More energy is radiated at some wavelengths than at others.  Molecules with absorption bands at wavelengths of high infrared radiation will tend to be more effective greenhouse gases than molecules with absorption bands at wavelengths of low infrared radiation.  Second, the impact of a greenhouse gas may be reduced if its absorption bands overlap with those of other gases present in the atmosphere in significant quantities.  For example, if a gas has absorption bands that overlap with those of an abundant greenhouse gas like carbon dioxide, the relative impact of the gas on radiative forcing is reduced since much of the energy at these wavelengths is already absorbed.

      	Initial concentration of the gas.  If the initial atmospheric concentration of a gas is high enough so that a significant portion of the energy at relevant wavelengths is already absorbed, the impact of additional increases in concentration is reduced.  This effect is significant for carbon dioxide.  It is not significant for CFCs and CFC substitutes since they are present in much smaller quantities than carbon dioxide.  Radiative forcing scales linearly with concentration for these chemicals.


      A useful index for comparing emissions that incorporates all of these factors is the Global Warming Potential (GWP).  The GWP depends on the position and strength of the absorption bands of the gas, its lifetime in the atmosphere, its molecular weight, and the time period over which the climate effects are of concern.  Specifically, the GWP is defined as the time-integrated change in radiative forcing resulting from a kilogram of emissions of a given chemical relative to the time-integrated change in radiative forcing resulting from a kilogram of emissions of a reference gas, typically carbon dioxide.

      Because different gases have different lifetimes, the period of integration has an important impact on relative GWPs.  This point is illustrated by the example of two different gases, each with an initial concentration of 1 ppb.  If the lifetime of the first gas is 10 years, its concentration will be 1/e[10] (0.00005) ppb after 100 years (ten lifetimes).  A 100-year GWP captures the entire impact of this gas since the concentration of the gas is effectively zero at the end of the period of integration, and it can no longer influence radiative forcing.  If the lifetime of the second gas is 100 years, its concentration will be 1/e (about 0.37) ppb after 100 years (one lifetime).  A 100-year GWP does not capture the entire impact of this gas since the gas is still present in significant quantities at the end of the period of integration, and it still affects radiative forcing.  If the period of integration were lengthened, the ratio of GWPs between the long-lived and the short-lived gases would increase because the value of the integral in the numerator would increase, while the value of the integral in the denominator would remain the same.

4.7.3	Properties of ODSs and ODS Substitutes

      Exhibit 4-46 presents atmospheric lifetime, molecular weight, integrated infrared band strength and GWPs for a number of ODSs and ODS substitutes.  Although there is overlap between categories, in general the PFCs have the highest GWPs among ODS substitutes.  HFCs are next, followed by HCFCs, which have the lowest GWPs.  PFCs have high GWPs because of their high integrated band strengths and their extremely long atmospheric lifetimes.  HCFCs generally have low GWPs because of their short atmospheric lifetimes (see Exhibit 4-46).  HFCs typically fall in the middle because of strong integrated band strengths and moderate atmospheric lifetimes.  In examining this table the following issues should be considered:

      	Although the GWPs of ODSs and ODS substitutes may be thousands of times the GWP of the reference gas, carbon dioxide, this is offset by the fact that emissions of these compounds are orders of magnitude smaller than emissions of carbon dioxide.  Nevertheless, it is estimated that CFCs accounted for about 23 percent of the direct radiative forcing from US emissions of greenhouse gases in 1990.

      	If the lifetime of a chemical is short compared to the lifetime of the reference gas, the GWP of the chemical declines as the period of integration increases.  If the lifetime of the chemical is long compared to the lifetime of the reference gas, the GWP of the chemical rises as the period of integration increases.
      
      	In practice a high GWP does not necessarily mean a large impact on warming.  If chemicals are never emitted they cannot cause a direct contribution to global warming even if they have high GWPs.  For example, if good service practices are in place, emissions from hermetically sealed household refrigerators are very small.  If the material is recovered for recycling or destruction after the useful life of the equipment, the refrigerant is never released to affect the atmosphere.  In contrast, mobile air conditioners tend to be leaky.  Substitutes used in this end use are more likely to be released to the atmosphere and, thus, impact global warming.


4.7.4	Indirect Effects from Changes in Energy Efficiency

      In addition to direct and indirect atmospheric effects, ODS substitutes may also have indirect energy effects that can influence global warming.  Use of alternative chemicals may change the energy efficiency of equipment and products currently using CFCs.  Because most energy is derived from burning fossil fuels, changes in energy efficiency may result in changes in carbon dioxide, nitrous oxide, and methane emissions, which will impact radiative forcing and global warming.  The two major end use categories in which the use of alternatives can impact the energy efficiency of equipment or products are refrigeration and air conditioning applications and foam insulation applications.  The key factors affecting energy efficiency for each of these applications are discussed below.

4.7.4.1		Refrigeration and Air Conditioning End Uses

      Refrigeration and air conditioning equipment can consume significant quantities of energy during normal operation.  Whether used in new or existing equipment, a new refrigerant can impact both the energy efficiency and cooling capacity of the refrigeration system.  The specific factors that impact the potential change in energy consumption associated with the use of a new refrigerant are as follows:

      	Thermodynamic Properties of the Alternative Refrigerant:  The thermodynamic properties of the refrigerant affect the amount of energy used by the compressor.  This is because these properties determine the evaporating and condensing pressures at which the equipment can operate given the temperature requirements of the application.  In practice, the greater the pressure difference between the evaporator and the condenser, the more the compressor must work to achieve a given cooling capacity.

      	Design of Equipment Components:  New refrigeration equipment is always designed around the refrigerant to be used.  System components such as the condenser, evaporator, or other heat exchangers are selected to minimize the work that a refrigeration system needs to do given a particular refrigerant.  For example, switching from a CFC to an alternative refrigerant might require a different type of heat exchanger tubing in the evaporator to enhance heat transfer or different sized expansion devices to better control the amount of refrigerant metered to the evaporator.

      	Retrofit Optimization:  The use of an alternative refrigerant in existing equipment (retrofits) often results in a decrease in the energy efficiency or refrigerating capacity of the equipment.  This is because the equipment was originally designed to operate most efficiently given the thermodynamic properties of the CFC refrigerant.  A number of steps can be taken to alter existing equipment so that it operates more efficiently with a given alternative.  Such steps include changing the motor drive train, increasing the motor speed, changing the heat exchangers, and replacing the impeller (for chillers).  The gains in energy efficiency associated with such measures must be weighed against cost and possible impact on refrigeration capacity.


      A widely used measure for the energy efficiency of refrigeration equipment is the coefficient of performance (COP).  This is defined as the total amount of refrigeration delivered by the equipment divided by the total amount of energy (i.e., work) input into the system.  The COP can be calculated for an ideal cycle.  The COP for an actual piece of equipment can be measured through field testing of the actual energy input and refrigerating effect of the equipment.  Field measurements of COP may vary from the ideal COP because of energy losses not accounted for in the model, or variations from the ideal cycle caused by friction losses or pressure drops.

      Finally, it is important to note that actual changes in energy efficiency are equipment and application specific.  As a result, it can be difficult to generalize about the potential impacts of a given substitute on energy efficiency in a given refrigeration or air conditioning application.

4.7.4.2		Foam Insulation

      Several factors combine to limit heat flow in thermal insulating foams including the volume fraction of the solid phase; the cell size which suppresses convection and radiation through cell walls; and the conductivity of the enclosed cell gases.  In addition to the conductivity per unit volume of the foam, the foam thickness also affects the thermal insulating capacity of foam products.

      CFCs were highly suitable as blowing agents in part because their low thermal conductivity contributed to the insulating capacity of the foam product.  Substitute blowing agents with higher thermal conductivity contribute less.  Higher thermal conductivity of the substitute blowing agent, however, does not necessarily translate into lower energy efficiency in foam products if formulation or thickness changes are made to compensate for the relative change in substitute blowing agents.
      

	CHAPTER 4 REFERENCES


Isaksen, I.S.A.  1986.  Ozone Perturbation Studies in a Two Dimensional Model with Temperature 	Feedbacks in the Stratosphere Included.  Presented at the UNEP Workshop on the Control of Chlorofluorocarbons, Leesburg, Virginia.  September 1986.

National Aeronautics and Space Administration (NASA).  1989.  An Assessment Model for 	Atmospheric Composition.  National Aeronautics and Space Administration Conference Publication 3023.

U.S. EPA.  1992.  Regulatory Impact Analysis: Compliance with Section 604 of the Clean Air Act 	for the Phaseout of Ozone Depleting Substances, Appendix A.  Office of Air and Radiation.  March 12, 1992.

U.S. EPA.  1987.  Assessing the Risks of Trace Gases that Can Modify the Stratosphere.  Office of 	Air and Radiation.  December 1987.

	5.  OCCUPATIONAL EXPOSURE AND HAZARD ANALYSIS


5.1	INTRODUCTION

      This section presents an analysis of potential occupational exposures to alternative refrigerant materials during the manufacture, installation, repair, and disposal of refrigeration equipment in several end-use sectors.  Evaluation of potential occupational exposures to these materials was based primarily on the use of a modelling approach that is similar to the approach used by EPA in their previous analyses of potential occupational exposure to refrigerant vapors (EPA 1990).  Where possible, the results obtained from the model were compared to available worker exposure data to determine the direction and magnitude of any uncertainty resulting from assumptions used in the model.  The following discussion describes the model used to evaluate potential occupational exposures to alternative refrigerants; this is followed by a discussion of the results obtained.

5.2	DESCRIPTION OF THE MODEL

      Occupational exposure to a gas or vapor is the result of two events:  a release of the material to the air and the presence of a worker within the contaminated space.  The duration and magnitude of the resulting exposure is influenced by (1) the duration and intensity of the release, (2) the rate at which contaminated air is diluted with uncontaminated air in the space affected by the release, (3) the proximity of the worker to the source of the release, and (4) the length of time that the worker remains in the space affected by the release.

      The first two factors are addressed directly by the mathematics of the "box" model used in this analysis.  This model has been widely used for many years to estimate probable exposures of workers to hazardous airborne materials, and has been described in detail by the National Institute for Occupational Safety and Health (NIOSH 1973).  Starting with a fundamental material balance, assuming that contaminated air is diluted with uncontaminated air:

	Rate of Accumulation = Rate of Generation - Rate of Removal 

	
	Where: 
            C = Concentration of gas or vapor at time t
            G = Rate of generation of the contaminant (cubic feet per minute (cfm))
            Q'= Effective rate of ventilation that is 
                    corrected for incomplete mixing (cfm)
            V = Volume of room or enclosure (ft[3])

Converting units and rearranging the equation yields the mass balance equation



	

      Where:


            	V = Room Volume (m[3])		 		 
            	Ci = Indoor Concentration (g/m[3])
			S = Source Generation Rate (g/hr)
            	rt = Residence Time of the indoor air (hr)
            	m = Mixing Factor (dimensionless)
            	Co = Outdoor Concentration (g/m[3])

      The integrated form of this equation can be used to calculate the average concentration of a contaminant in a room during the release of a chemical substance:

	

      Where:

            	Cb = Concentration already existing in the room when 				the emission starts (i.e., at time = 0)

      Assuming that the concentration of contaminant at time = 0 is zero (i.e., Cb = 0), that contaminated air is diluted with uncontaminated air (i.e., Co = 0), and that the concentration has reached a steady-state (equilibrium) level, the integrated equation above can be further simplified: 

 
	

A dimension analysis reveals that:

	

      Since it is customary to measure occupational exposures to gasses and vapors in units of parts per million, the equation above can be modified to:

	

where MW is the molecular weight of the contaminant.

      Because the residence time r is the inverse of the air exchange rate in air changes per hour (ACH), the following can be derived, which is the working equation used to estimate worker exposure to alternative refrigerants:    

	


      However, as described above, the degree to which a worker is exposed to a gas or vapor also depends on the location of the worker at the time the contaminant is released.  To address this aspect, the use of the box model was modified to recognize the fact that workers move around an area where a release occurs.  Specifically, the workspace was divided into four areas or zones.  These zones are defined as follows:

      	Zone 1 represents a space measuring 3' x 3' x 3' centered around an emission source.  A worker is said to be in Zone 1 whenever his/her breathing zone is within 18 inches of the emission source.  Thus, this zone represents an area occupied by a worker performing operations within an arm's reach of the emission source;

      	Zone 2 represents a space measuring 12' x 12' x 12' centered around an emission source.  A worker is said to be in Zone 2 while working within 6 feet of the source;

      	Zone 3 represents a space measuring 30' x 30' x 30' centered around an emission source.  A worker is considered to be in Zone 3 while working within 15 feet of the emission source; and

      	Zone 4 represents a space sufficiently far from the emission source that no measurable exposure occurs.

Thus, refrigerant vapor concentrations will be highest in Zone 1 and will decrease as the worker moves through each zone away from the emission source.  Use of this approach captures the idea that there is a vapor concentration gradient surrounding the emission source, and that a worker, in the course of performing normal work duties, will move between these zones.

      A second important assumption concerns the rate at which air is exchanged within each zone.  In the implementation of the box model, air within a small space is exchanged at a faster rate than is air within a larger space.  This is due to the presence of local air currents and eddies.  In Zone 1, which is the smallest space defined in the model, the motion of the worker's hands and tools will serve to increase the rate of air exchange; however, in a larger space, local air currents will have a lesser effect on the overall rate of air exchange within that space.  As such, the air within a 3' x 3' x 3' space may be exchanging with surrounding air at a rate of 30 air changes per hour (ACH) even though the overall rate of air exchange within a 30' x 30' x 30' room is only 0.5 ACH.  In the model, it is assumed that air within Zone 1 is moving through the zone at a rate of about 2 mph.  This rate is cut in half for Zone 2, and by 75 percent in Zone 3.  The rate of air movement within Zone 3 is equivalent to about 50 feet per minute, which represents a background rate of air movement frequently seen in industrial facilities having moderate general dilution ventilation.  A mixing factor of 0.5 is assumed, which cuts the effective air exchange rate in half.

      Thus, to implement the box model, the variables in the working equation above were substituted as follows.  The variable S is the release rate in g/hr.  The value of S was derived from the Vintaging model used to estimate refrigerant release rates for various release scenarios.  The release rate per event (from the Vintaging model) was multiplied by the number of events estimated to occur over a workday.  For equipment manufacturing, the number of events per workday was assumed to equal the number of units produced per plant per year divided by 365 days per year.  For scenarios where fewer than 365 units are manufactured per year, it is assumed that the event causing the emission occurs over a single 8-hour workday, rather than over several days.  For installation and repair operations, it is assumed that one event occurs per workday for large refrigeration systems and four events per workday for small systems.  For disposal operations involving refrigerators and freezers, other refrigeration appliances, and mobile air conditioners, it is assumed that 100 units are disposed of over an 8-hour shift.  For all other end-use sectors, it is assumed that one unit is disposed of during a single shift.


      The variable V represents the volume of space in which the worker is performing his or her work duties.  For each exposure scenario, it is assumed that a worker spends a certain percentage of time within each of the four zones, as follows.  For manufacturing and installation operations, it is assumed that a worker spends 10 percent of his or her time over an 8-hour workday in Zone 1, 50 percent in Zone 2, 30 percent in Zone 3, and 10 percent in Zone 4.  For repair operations, it is assumed that a worker spends 20 percent of the workday in Zone 1, 40 percent in Zone 2, 20 percent in Zone 3, and 20 percent in Zone 4; this reflects the likely possibility that, in repair operations as compared to manufacturing or installation operations, a worker will spend a larger proportion of the workday in close proximity to the emission source, but will also spend a greater proportion of the day away from the emission source while travelling to the repair site.

      After using the box model to estimate the average vapor concentration within each zone, a worker's 8-hour time-weighted-average exposure is calculated as follows:

	

where 
      Cx = Concentration of vapor within Zone x; and
      Px = Percent of time spent in Zone x by the worker.

The term representing Zone 4 drops out because there is no exposure to refrigerant vapors in Zone 4 (by definition).

      To estimate short-term (i.e., 30-minute) worker exposures to refrigerant vapors and gasses, it is assumed that 25 percent of the daily emission is released over a 30-minute period.  

5.3	DISCUSSION OF RESULTS

5.3.1	Presentation of Results

      The attached tables (Attachment 5-A) show the modelling results for each end-use sector, work operation, and alternative refrigerant.  For each scenario, the tables report the average 8-hour and 30-minute time-weighted average (TWA) concentrations of refrigerant present in each of the zones, and the estimated 8-hour and 30-minute TWA levels to which workers may be exposed.

      Exhibit 5-1 summarizes the occupational exposure limits for the alternative refrigerants that were used to determine whether workers may be exposed to potentially hazardous airborne gas or vapor concentrations.  Exhibit 5-2 summarizes the results for those scenarios where the model suggests that worker exposures may exceed either the 8-hour or 30-minute occupational exposure limit for the alternative refrigerant.  The highest exposures are predicted to occur during the manufacturing and disposing of industrial process (high pressure build-up) refrigeration systems.

5.3.2	Interpretation and Comparison of Model Results with Sampling Data

      Several observations suggest that the model will tend to overstate likely occupational exposures to alternative refrigerant vapors and gasses.  First, the emission estimates from the 
      
	EXHIBIT 5-1
	OCCUPATIONAL EXPOSURE LIMITS FOR
	POTENTIAL ALTERNATIVE REFRIGERANTS*

                                  Substitute
                                   Material
                                       
                                   WGL (ppm)
                                       
                                   EGL (ppm)
HCFC-22
                                     1,000
                                     5,000
HFC-23
                                     1,000
                                     3,000
HCFC-123
                                        30
                                     2,500
HCFC-124
                                     1,000
                                     3,000
HFC-125
                                     1,000
                                     3,000
HFC-134a
                                     1,000
                                     3,000
HFC-143a
                                     1,000
                                     3,000
HCFC-142b
                                     1,000
                                     5,000
HFC-152a
                                     1,000
                                     4,000
HFC-227ea
                                     1,000
                                     3,000

*  This table may contain or utilize exposure limits that are not Agency verified or that were updated after this chapter was written.  Chapter 3 presents the most recent exposure limits for each chemical, identifies which are verified versus estimated values, identifies methods used to derive estimated values, and provides information sources.



Vintaging model are considered to be worst-case estimates.  Second, the assumptions regarding air exchange rates for each of the work zones do not reflect the use of any local exhaust ventilation, which is likely to be used in manufacturing and disposal operations involving any significant releases.

      The use of properly designed local exhaust ventilation or other means of isolating the worker from emission sources would likely prevent gas or vapor concentrations from exceeding the occupational exposure limits most of the time.  Given the assumed sizes of many of the potential releases (some of them exceeding a few hundred pounds), it would appear unlikely that an employer would conduct such operations without providing some means of mechanical ventilation.  It is for this reason that the exposure estimates for high-pressure industrial process refrigeration systems are certain to be unrealistically high.  In addition, for each release scenario, it is assumed that the entire estimated daily release occurs over a single 8-hour workday and at a single location.  This is not the case for many scenarios, and particularly for manufacturing operations, where the estimated daily release may occur over two or three workshifts or where the release may be spread out over two or more production lines.  Finally, most of the predicted exceedances over the occupational exposure limits are the result of assuming that there will be no recycling of HFC refrigerants.  To the extent that recycling technologies are used, exposures to HFCs should be below the occupational exposure limits for these compounds.

      The model's tendency to overstate worker exposures is further supported by sampling data obtained for HCFC-123 refrigerant during repair operations conducted on low-pressure
      
	EXHIBIT 5-2
	SUMMARY OF ESTIMATED WORKER EXPOSURE TO ALTERNATIVE REFRIGERANTS
	FOR SCENARIOS WHERE THE OCCUPATIONAL EXPOSURE LIMIT IS EXCEEDED[a]

                                       
                                 8-Hour TWA[b]
                                 30-Minute TWA
                                End-Use Sector
                                   Operation
                                  Substitute
                                  Material[c]
                                   Predicted
                                Exposure (ppm)
                                   Operation
                                  Substitute
                                  Material[c]
                                   Predicted
                                Exposure (ppm)
Retail Food (Parallel System)
Disposal



HFC-134a
HFC-125
HFC-143a
HFC-23
                                                                          1,579
                                                                          1,342
                                                                          1,917
                                                                          2,300
Disposal
HFC134a
HFC-125
HFC143a
HFC-23
                                                                          4,742
                                                                          4,030
                                                                          5,757
                                                                          6,909
Cold Storage Warehouses
Disposal
HFC-152a
HFC-23
                                                                          1,309
                                                                          1,234


                                                                               
Centrifugal Chillers (Low Pressure)
Manufacture
HCFC-123
                                                                          73[d]


                                                                               
Centrifugal Chillers (High Pressure)
Disposal
HFC-124
HFC-134a
HFC-227ea
                                                                          2,271
                                                                          3,050
                                                                          1,647
Disposal


HFC-124
HFC-134a
HFC-227ea
                                                                          6,819
                                                                          9,159
                                                                          4,928
Industrial Process (Low Pressure)
Manufacture
Repair
HCFC-123
HCFC-123
                                                                             54
                                                                             49


                                                                               

	EXHIBIT 5-2 (continued)
	SUMMARY OF ESTIMATED WORKER EXPOSURE TO ALTERNATIVE REFRIGERANTS
	FOR SCENARIOS WHERE THE OCCUPATIONAL EXPOSURE LIMIT IS EXCEEDED[a]

                                       
                                 8-Hour TWA[b]
                                 30-Minute TWA
                                End-Use Sector
                                   Operation
                                  Substitute
                                  Material[c]
                                   Predicted
                                Exposure (ppm)
                                   Operation
                                  Substitute
                                  Material[c]
                                   Predicted
                                Exposure (ppm)
Industrial Process
  (High Pressure          Build-Up)
Manufacture

Disposal
HFC-143a
HFC-152a

HFC-134a
HFC-152a
HFC-125
HFC-143a
HFC-23
                                                                          1,293
                                                                          1,016
                                                                               
                                                                         30,840
                                                                         47,661
                                                                         15,740
                                                                         18,375
                                                                         13,490
Manufacture


Disposal
HFC-134a
HFC-143a
HFC-152a
HFC-134a
HFC-152a
HFC-125
HFC-143a
HFC-23
                                                                          3,941
                                                                          7,751
                                                                          6,090
                                                                         92,608
                                                                        143,122
                                                                         47,266
                                                                         56,260
                                                                         40,509
Large Ice Makers
Repair
HFC-152a
                                                                          1,076


                                                                               
Ice Rinks
Disposal
HFC-134a
HFC-152a
HFC-23
                                                                          4,609
                                                                          7,123
                                                                          2,016
Disposal
HFC-134a
HFC-152a
HFC-23
                                                                         13,840
                                                                         21,390
                                                                          6,054
Central A/C and Home Heat Pumps
Manufacture
HFC-152a
                                                                          1,332


                                                                               

[a]	Exposure levels are rounded to the nearest whole number.
[b]	Time-weighted average.
[c]	The refrigerant blends analyzed for these sectors may be considered confidential business information (CBI) by the manufacturer.  As a result, the identification of the constituents of any blend and the proportion of each constituent contained in any blend are not presented in this exhibit.
[d]	One manufacturer's representative states that worker exposures to HCFC-123 during centrifugal chiller manufacture are typically 1 ppm or less, due to the use of exhaust ventilation.

centrifugal chillers (EPA 1991).  This study showed that worker exposures were generally well below 1 ppm as an 8-hour TWA during repair operations involving the use of recycling equipment; however, results from the model suggest that workers may be exposed to about 4 ppm HCFC-123 during these repair operations, even when proper recycling is used.  For the manufacture of low-pressure centrifugal chillers, the model suggests that workers could be exposed to HCFC-123 concentrations as high as 73 ppm as an 8-hour TWA; however, a representative of one manufacturer indicated that, due to the use of local exhaust ventilation, worker exposures were typically in the range of 1 ppm (personal communication).

      The analyses presented above did not consider the use of ammonia, butane, propane, or perfluoropropane as alternative refrigerants.  However, existing OSHA and ASHRAE standards governing the handling and use of these materials should serve to limit occupational exposures to them.  Currently, OSHA enforces permissible exposure limits (PELs) for two of these materials: ammonia (50 ppm as an 8-hour TWA), and propane (1000 ppm as an 8-hour TWA).  In addition, ASHRAE classifies butane and propane as "highly flammable" refrigerants, and ammonia as a "moderately flammable" refrigerant having high toxicity (ASHRAE Standard 34); as such, special requirements set forth in ASHRAE Standard 15 apply to the use of these materials as refrigerants.  These requirements include:

      	Prohibiting the use of any of these refrigerants in systems designed for human comfort;

      	Limiting the amount of refrigerant that can be used in systems installed in occupied areas;

      	Requiring the use of vapor detection alarms that activate when concentrations of refrigerant exceed toxic or explosive levels; and

      	Prohibiting the use of butane or propane in any system installed in commercial or institutional buildings (i.e., only industrial use is permitted).

5.4	POTENTIAL WORKPLACE FLAMMABILITY AND EXPLOSIVITY RISKS

      One of the alternative refrigerants considered in this analysis, HFC-152a, has been classified as a moderately flammable material under proposed ASHRAE Standard 34.  It has a lower flammability limit (LFL) of 4.4 percent in air (44,000 ppm) and a heat of combustion of less than 19,000 kJ/kg.  As such, it represents a lesser fire and explosion risk than propane or butane, which have LFL's that are less than half that of HFC-152a and heats of combustion greater than 19,000 kJ/kg.

      The risk of fire or explosion in an industrial setting arises when local concentrations of a flammable or explosive gas or vapor are allowed to exceed the lower flammable limit for the gas or vapor in the presence of an ignition source (i.e., flame, heat, electrical energy, or static sparks).  In general, when handling flammable or explosive materials, most industrial facilities consider that accumulations of gas or vapor concentrations of from 10 to 25 percent of the LFL or higher for that substance represent an emergency situation that requires corrective action.  For HFC-152a, an emergency concentration would therefore be considered to exist if concentrations exceeded 4,400 to 11,000 ppm (i.e., 10 to 25 percent of the LFL).


      The exposure assessment analysis presented above shows that, under worst-case assumptions and assuming no recycling, local short-term concentrations of HFC-152a would not be expected to exceed 4,400 ppm under normal operating conditions in all but one end-use sector (high-pressure industrial process refrigeration).  As stated above, EPA believes that the exposure estimates for this end-use sector are unrealistically high because of the assumption that no mechanical ventilation would be used during operations involving the release of such large quantities of refrigerant.  Thus, with the appropriate use of equipment designed for the safe use of flammable or explosive gases and with the use of mechanical ventilation systems, concentrations of HFC-152a are not likely to exceed 10 percent of the LFL and should therefore not present a substantial fire or explosion risk in the workplace.  Isolation of potential ignition sources (such as welding operations conducted on the assembly line) from processes involving the handling of HFC-152a, and the use of combustible gas detectors and alarm systems, will serve to further reduce this potential risk.

	CHAPTER 5 REFERENCES

NIOSH (Mutchler, J.E.)  1973.  "Principles of Ventilation."  The Industrial Environment--It's 	Evaluation and Control.  NIOSH, Center for Disease Control, U.S. Department of Health and Human Services.  ch. 39.

U.S. EPA.  1990.  Revised Draft:  Occupational Exposure and Environmental Release Data for 	Chlorofluorocarbons (CFCs) and Their Substitutes.  Office of Toxic Substances.

U.S. EPA.  1991.  Results of Employee Exposure Monitoring for HCFC-123 at Centrifugal Chiller 	Installations.  Final Report.  November, 1991.

	ATTACHMENT 5-A




	6.  CONSUMER EXPOSURE AND RISK SCREENING ANALYSIS


      This chapter presents a screening level assessment of the risks to consumers from releases of CFC substitutes used in home appliances (refrigerators and freezers, central air conditioners, window air conditioners, and miscellaneous refrigerated appliances) and automobile air conditioners.  These are the only applications in the refrigeration and air conditioning sectors that have consumer end uses.  

      The end uses and substitutes evaluated in this chapter are presented below.  The composition of one -- R200b -- is confidential:

      	Refrigerators and Freezers:

            	HCFC-22
            	HFC-134a
            	HFC-152a
            	HCFC-22/HCFC-142b
            	HCFC-22/HFC-152a/HCFC-124
            	R200b

      	Dehumidifiers: Same as refrigerators and freezers.

      	Central Air Conditioners and Home Heat Pumps:

            	HFC-125
            	HFC-134a
            	HFC-125/HFC-143a
            	HFC-23/HFC-134a
            	HFC-125/HFC-143a/HFC-134a
            	HCFC-22/propane/HFC-125
      
      	Window Air Conditioners:  

            	HFC-125
            	HFC-134a
            	HFC-125/HFC-143a
            	HFC-23/HFC-134a

      	Mobile Air Conditioners:

            	HFC-134a
            	HCFC-22/HFC-152a/HCFC-124

      This chapter first presents an assessment of consumer exposure to releases from home appliances, followed by an analysis for mobile air conditioners.  The analyses performed are screening analyses whose purpose is to identify the proposed substitutes for all end uses that may pose any potential concern for human health or environmental risk.  As a result, the analyses employ conservative assumptions that tend to overestimate potential risk.  Considerable uncertainty exists in these analyses.

6.1	HOME APPLIANCES

      Three types of releases from home appliances are of interest for consumer exposure and risk:  (1) routine operating leakage, (2) releases during servicing, and (3) accidental releases.  

      A report on consumer exposure to CFCs and their substitutes prepared for EPA (Versar 1990) examined exposures from leakage and servicing releases from refrigerators, freezers, and dehumidifiers.  Using engineering estimates of releases as input to an equation given on page 4-5 of this report, average annual concentrations of CFC substitutes were calculated in median-sized and 10th percentile-sized homes.  Because this consumer exposure analysis spreads the releases and resulting concentrations over a year's time, it assesses only long-term, chronic exposures.  It does not address potential risks from short-term exposure to higher concentrations that result from servicing or accidental releases.

      Below, potential risks from chronic exposures are examined by comparing estimated doses from the consumer exposure report to toxicity values for the CFC substitutes.  Subsequently, a screening analysis of potential risks associated with higher, but shorter-term, exposures is presented.  Finally, potential safety concerns from the use of HFC-152a in household refrigerators are discussed.

6.1.1	Risks from Long-term Exposure to Leakage and Servicing Events  Previous Estimates

      The consumer exposure report estimated releases from refrigerators, freezers, and dehumidifiers with refrigerant charge sizes ranging from about 140 to 570 g.  The annual leakage rate was assumed to be 0.2 percent of the charge (0.28 to 1.1 g), and the servicing emission was assumed to be 85 g.

      The estimated average annual concentrations in two different house sizes  a median size of 411 m[3] (equivalent to 1,808 ft[2], assuming an 8-ft ceiling height) and a 10th percentile size of 174 m[3] (766 ft[2], assuming an 8-ft ceiling)  ranged from 0.4 μg/m[3] to 3.7 μg/m[3] for annual leakage.  To estimate potential risk from chronic exposure to leakage, the consumer exposure report's exposure information was combined with available toxicity information as follows:

      	To evaluate noncarcinogenic effects, the modeled exposure concentrations were compared to reference concentrations (RfCs).

      	To calculate cancer risks, a lifetime average daily dose of the substitute was estimated using the exposure concentration, a 20 m[3]/day inhalation rate, and a 70 kg body weight.

The RfCs and cancer slope factor used in this assessment are given in Chapter 3 of this document.

      The highest estimated average annual exposure for routine leakage from the consumer exposure report  3.7 μg/m[3]  is well below the RfCs for all substitutes.  Only one of the substitutes  HCFC-22  is a possible carcinogen.  Assuming that people are exposed to an HCFC-22 concentration of 3.7 μg/m[3] 24 hours per day, 365 days per year, over a 70-year lifetime, the cancer risk for HCFC-22 is 2.1E-07.  Based on EPA guidelines regarding levels of risk, chronic health effects from long-term exposure to routine leakage of CFC substitutes from home appliances are not of concern.
      
      The consumer exposure report estimated the typical servicing emissions from all pieces of equipment to be 85 grams per service.  During a year in which the equipment is serviced once, the report calculated average annual concentrations ranged from 118 μg/m3 for a median-sized house to 279 μg/m[3]  for a house sized in the 10th percentile.  These are also well below the RfCs for the substitutes.  If it is assumed that servicing occurs once every year and that people spend all of their time at home, the cancer risk associated with lifetime exposure to HCFC-22 would be 6.7E-06 in the median house and 1.6E-05 in the smaller house.  However, these values overstate potential cancer risks because (1) home appliances such as refrigerators, freezers, and humidifiers are serviced much less frequently than once a year; and (2) people typically do not spend all of their time at home.  Given these factors, EPA believes that actual cancer risks are well below 10[-5], the threshold level of concern.

6.1.2	Risks from Short-term Exposure During Servicing and Accidental Releases

      The next step was to examine potential health risks from short-term exposure to temporarily higher concentrations associated with servicing emissions and accidental releases.  Four situations were examined:

      	Exposure to vapors from refrigerators or freezers located in the kitchen;

      	Exposure to vapors from a dehumidifier located in a room;

      	Exposure to vapors from a central air conditioner or heat pump, where the release occurs in the basement; and

      	Exposure to vapors from a window air conditioner located in a room.

Approach

      To simulate the initial increase and subsequent decrease in airborne concentrations resulting from refrigerant releases, the same "box model" approach was used as in the previous analysis of occupational exposure (see Chapter 5).  This model was derived from a differential material balance that, when integrated, provides a basis for relating the air concentration of a contaminant to the generation and removal rates of the contaminant.  

      To calculate releases, it was assumed that the amount of substitute released was equal to the amount of CFC released.  CFC release information was derived from input files to the Vintaging Model.  The following assumptions concerning releases were made for this analysis:

      	Servicing releases occur over 15 minutes.  HCFCs are recycled at servicing (under Section 608 of the Clean Air Act, HCFC recycling became mandatory in 1992).  Because HFC recycling will not be mandatory until 1995, servicing releases for HFCs are calculated both with and without recycling.  If there is no recycling, the entire charge is released.  The releases are as follows:


                                                Recycling	No Recycling

            Refrigerators and Freezers		   	   15 g	   	  300 g
            Dehumidifiers					   20 g	   	  400 g
            Central A/Cs and Heat Pumps		   80 g	 	3,840 g
            Window Air Conditioners			   40 g	 	1,820 g


      	Accidental releases are modeled as leakage of the entire charge over two different time frames:  15 minutes and four hours.  The releases are the same as those listed above under the "no recycling" scenario.

      The following inputs were made to the box model:

      	Room sizes:

            	411 m[3] for house volume (equivalent to 1,808 ft[2], assuming an 8-ft ceiling height);
            	18 m[3] (79 ft[2]) for the kitchen;
            	102 m[3] (449 ft[2]) for the basement; and
            	41 m[3] (180 ft[2]) for the room housing the window A/C or miscellaneous unit.

      	Air exchange rate:

            	0.25 air changes per hour (ACH) for the house, basement, and room; and
            	2.3 ACH for the kitchen.
 
      Thirty minute time weighted average (TWA) exposures were calculated and compared to the interim emergency guidance levels (EGLs) shown in Chapter 3.

Results

      Exhibits 6-1 through 6-4 show the 30 minute average concentrations and their ratios to the interim EGLs for the relevant substitutes.  (The values shown in these tables are based on the substitute being composed of 100 percent of the named chemical.  In practice, some substitutes would be used only in a blend.)  The results are discussed below.

      The results show that servicing with recycling never results in concentrations above the EGL.  All of the other scenarios involve release of the entire charge.  This can result from an accident or  for the HFCs only  from a servicing event that does not involve recycling.  The modeling indicates that, unless otherwise noted, the concentrations of the following substitutes or combinations of substitutes never exceed the EGL even if the entire charge is released:

      	Refrigerators and freezers:

            	HCFC-22;

            	HFC-134a;

            	HFC-152a;
            
                                       
                                       
                                  EXHIBIT 6-1
                30-MINUTE AVERAGE CONCENTRATIONS FROM SERVICING
                     AND ACCIDENTAL RELEASES, AND RATIO TO
                   INTERIM EMERGENCY GUIDANCE LEVELS (EGLs):
                          REFRIGERATORS AND FREEZERS


                            Service, With Recycling
                          Service, Without Recycling
                              Accidental 15-Min.
                                    Release
                            Accidental 4-Hr Release
KITCHEN
30-Min. TWA, mg/m[3]

Ratio to EGL:
        HCFC-22
        HCFC-124
        HCFC-142b
        HFC-134a
        HFC-152a        
        HFC-227ea
     430     


     0.025
     0.025
     0.022
     0.034
     0.032     
     0.020
   10,000


    NA
    NA
    NA
    0.79
    0.74 
    0.48
   10,000


    0.57
    0.58
    0.50
    0.79
    0.74
    0.48
  3,300


   0.19
   0.19
   0.17
   0.26
   0.24
   0.16
HOUSE
30-Min. TWA, mg/m[3]

Ratio to EGL:
        HCFC-22
        HCFC-124
        HCFC-142b
        HFC-134a
        HFC-152a
        HFC-227ea
      24


     0.0014
     0.0014
     0.0012
     0.0019
     0.0018
     0.0011
     550


     NA
     NA
     NA
    0.044
    0.041
    0.026
     550


    0.031
    0.032
    0.028
    0.044
    0.041
    0.026
    570


   0.033
   0.033
   0.029
   0.045
   0.042
   0.027

                                       
                                       
                                  EXHIBIT 6-2
                30-MINUTE AVERAGE CONCENTRATIONS FROM SERVICING
                     AND ACCIDENTAL RELEASES, AND RATIO TO
                   INTERIM EMERGENCY GUIDANCE LEVELS (EGLs):
                                 DEHUMIDIFIERS


                            Service, With Recycling
                          Service, Without Recycling
                              Accidental 15-Min.
                                    Release
                            Accidental 4-Hr Release
ROOM
30-Min. TWA, mg/m[3]

Ratio to EGL:
        HCFC-22
        HCFC-124
        HCFC-142b
        HFC-134a
        HFC-152a
        HFC-227ea
     440     


     0.025
     0.026
     0.022
     0.035
     0.033
     0.021
   6,100


    NA
    NA
    NA
    0.48
    0.45
    0.29
   6,100


    0.35
    0.36
    0.31
    0.48
    0.45
    0.29
  6,400


   0.37
   0.37
   0.32
   0.51
   0.47
   0.30
HOUSE
30-Min. TWA, mg/m[3]

Ratio to EGL:
        HCFC-22
        HCFC-124
        HCFC-142b
        HFC-134a
        HFC-152a
        HFC-227ea
      42


     0.0024
     0.0025
     0.0021
     0.0033
     0.0031
     0.0020
     610


    NA
    NA
    NA
    0.048
    0.045
    0.029
     610


    0.035
    0.036
    0.031
    0.048
    0.045
    0.029
     630


   0.036
   0.037
   0.032
   0.050
   0.047
   0.030




                                  EXHIBIT 6-3
                30-MINUTE AVERAGE CONCENTRATIONS FROM SERVICING
                     AND ACCIDENTAL RELEASES, AND RATIO TO
                   INTERIM EMERGENCY GUIDANCE LEVELS (EGLs):
                 CENTRAL AIR CONDITIONERS AND HOME HEAT PUMPS


                            Service, With Recycling
                          Service, Without Recycling
                              Accidental 15-Min.
                                    Release
                            Accidental 4-Hr Release
BASEMENT
30-Min. TWA, mg/m[3]

Ratio to EGL:
        HCFC-22 
        HFC-23
        HFC-125
        HFC-134a
        HFC-143a
     590


     0.034
     0.069
     0.040
     0.047
     0.058       
  28,000


    NA
    3.3
    1.9
    2.2
    2.7 
   28,000


     1.6
     3.3
     1.9
     2.2
     2.7
  29,000


    1.7
    3.4
    2.0
    2.3
    2.8   
HOUSE
30-Min. TWA, mg/m[3]

Ratio to EGL:
        HCFC-22
        HFC-23
        HFC-125
        HFC-134a
        HFC-143a
     150


     0.0086
     0.018
     0.010
     0.012
     0.015     
   7,000


    NA
    0.82
    0.48
    0.56
    0.69    
   7,000


    0.40
    0.82
    0.48
    0.56
    0.69 
   7,200


    0.41
    0.84
    0.49
    0.57
    0.71 



                                  EXHIBIT 6-4
                30-MINUTE AVERAGE CONCENTRATIONS FROM SERVICING
                     AND ACCIDENTAL RELEASES, AND RATIO TO
                   INTERIM EMERGENCY GUIDANCE LEVELS (EGLs):
                            WINDOW AIR CONDITIONERS


                            Service, With Recycling
                          Service, Without Recycling
                              Accidental 15-Min.
                                    Release
                            Accidental 4-Hr Release
ROOM
30-Min. TWA, mg/m[3]

Ratio to EGL:
        HFC-23
        HFC-125
        HFC-134a
        HFC-143a
     740


     0.087
     0.050
     0.059
     0.073
  34,000


    4.0 
    2.3
    2.7
    3.3
  34,000


    4.0
    2.3
    2.7
    3.3
  34,000


    4.0
    2.3
    2.7
    3.3
HOUSE
30-Min. TWA, mg/m[3]

Ratio to EGL:
        HFC-23
        HFC-125
        HFC-134a
        HFC-143a
      74


     0.0087
     0.0050
     0.0059
     0.0073
   3,300


    0.39
    0.22
    0.26
    0.32
   3,300


    0.39
    0.22
    0.26
    0.32
   3,400


    0.40
    0.23
    0.27
    0.33

            	R200b.  The weighted average of the ratios is below 1.0, indicating that this blend is of no concern even if the health hazards are additive.

            	HCFC-22/HCFC-142b.  The weighted average of the ratios is below 1.0, indicating that this blend is of no concern even if the health hazards are additive.

            	HCFC-22/HFC-152a/HCFC-124.  The weighted average of the ratios is below 1.0, indicating that this blend is of no concern even if the health hazards are additive.

            Releases from refrigerators and freezers are based on a charge size of 300 g.  A conservative estimate of charge sizes for refrigerators and freezers is 240 g for refrigerators and 340 g for freezers.  Since more people own refrigerators than freezers, the estimated charge size of 300 g used in this analysis actually overstates releases and exposures for most of the population.

      	Dehumidifiers:

            	HCFC-22;

            	HFC-134a;

            	HFC-152a;

            	R200b.  The weighted average of the ratios is below 1.0, indicating that this blend is of no concern even if the health hazards are additive.

            	HCFC-22/HCFC-142b.  The weighted average of the ratios is below 1.0, indicating that this blend is of no concern even if the health hazards are additive.

            	HCFC-22/HFC-152a/HCFC-124.  The weighted average of the ratios is below 1.0, indicating that this blend is of no concern even if the health hazards are additive.

      Under the assumptions used in this analysis, all of the other substitute/end use combinations result in exposures above the EGL if the entire charge is released:

      	Central air conditioners and home heat pumps:  None of these compounds or blends exceed their EGLs in the house.  In the basement, however, all alternatives (taken either individually or as blends) exceed their EGLs.  These figures are based on the assumption that about 20 percent of central air conditioners and home heat pumps have a charge size of 3,800 g, about 75 percent have a charge size of 3,000 g, and 5 percent have a charge size of 700 g.

      	Window air conditioners:  All alternatives exceed their EGLs in the room where the air conditioner is located.  All window air conditioners are assumed to have a charge size of 1,800 g.

As discussed above, these results are based on a comparison of 30-minute TWA exposures with interim EGLs.  Another way to approach this assessment would be to calculate shorter-term exposures and compare them to the cardiotoxic LOAELs or NOAELs given in Chapter 3.  If EPA were to calculate 15 minute TWA exposures for the scenarios with 15-minute releases (i.e., servicing without recycling, 15-minute accidental release), the TWA exposures would be well below the cardiotoxic NOAELs and LOAELs for every substitute.

6.1.3	Summary of Modeling Results

      Analysis has shown that unless the entire refrigerant charge is released over a short period of time, none of the substitutes exceed EPA's threshold of concern for human health risk in any piece of equipment.  Routine leakage and servicing with recycling simply will not release enough refrigerant to pose significant short-term or long-term risks to human health.

      If the entire charge is released, the calculated 30-minute average concentrations exceed the EGLs for substitutes in air conditioning equipment.  The following factors should be considered when judging whether these exceedances are cause for concern:

      	Recycling at servicing will be required for the HFCs by 1995.  Until then, EPA believes that most HFCs will be recycled anyway, because of their high cost relative to CFCs.

      	Accidental releases of an entire refrigerant charge occur infrequently.  The UNEP (1992) reports that less than 5 percent of the refrigerators and freezers have accidental releases of the entire charge.

      	The analysis is based on conservative assumptions:

            	Exposures are generally calculated for the equipment with the largest charge size.

            	All releases are assumed to occur indoors.  This is not necessarily the case for central air conditioners, home heat pumps, and window air conditioners.

            	It was assumed that people remain in the room in which the release occurs for 30 minutes.

            	Air exchange rates estimated are conservative; they reflect an energy efficient house with no windows or doors open.

            	The EGLs that are used as toxicity benchmarks are conservative, and incorporate a margin of safety.  A similar approach based on cardiotoxic LOAELs or NOAELs would show that exposures are below the concern level.

6.1.4  Safety Concerns:  HFC-152a in Household Refrigerators

      The release of HFC-152a in a household kitchen can potentially cause personal injury in a few ways:

      	injury from sudden overpressure (i.e., shock wave);
      	injury from shrapnel resulting from an explosion; and
      	injury from thermal radiation.

The likelihood and severity of these risks were studied in a report on risk assessment of flammable refrigerants for home appliances.  (EPA 1991).  Using a mathematical model to relate vapor concentration to peak overpressure, the report found that the sudden ignition of a typical charge of HFC-152a would be sufficient to cause eardrum rupture of a person who was in the kitchen at the time of ignition.  The explosion would also likely cause structural damage to the house.  However, the report also found that sudden ignition of a full charge was an extremely unlikely event; rather, there would be some dispersion of HFC-152a vapor below the lower flammable limit of 4.4 percent, and thus the subsequent event would be limited to a vapor fire rather than an explosion.

      Using information on historical leakage rates coupled with a fault tree analysis, the report predicts that the probability of a fire or explosion due to an external refrigerant leak is
      
2.0 x 10[-8]/yr (5.0 x 10[-7] under worst-case assumptions).  The probability of an explosion due to leakage of HFC-152a inside the refrigerator is considerably smaller (2.7 x 10[-13]/yr under average-case assumptions and 7.0 x 10[-12]/year under worst-case assumptions).

6.2	MOBILE AIR CONDITIONERS (MACs)

6.2.1	Approach Used in a Previous Consumer Exposure Report

      The consumer exposure report (Versar 1990) estimated MAC emissions to which consumers could be exposed as the product of four factors: (1) total charge of refrigerant in the system; (2) fraction of refrigerant that leaks from the system before it is noticed and repaired; (3) fraction of refrigerant entering the vehicle passenger compartment; and (4) frequency of leak episodes.

      	Total charge:  Total refrigerant charge is based on averages from a previous study (Chilton 1985) for four vehicle class sizes: subcompact, 27 ozs.;  compact, 38 ozs.; midsize, 47 ozs.; and large, 53 ozs.

      	Total refrigerant leaked before being noticed/repaired:  In newer vehicles with protection devices and in warmer climates, the consumer exposure report assumed that leaks would be noticed and fixed quickly.  In cooler climates and older cars without protection devices, a MAC leak would probably go unnoticed until most of the refrigerant had leaked.  Overall, the report assumed an average leak of 50 percent of charge capacity.

      	Fraction of leak entering passenger compartment:  This parameter depends on the location of the leak (evaporator case or engine compartment) and whether the vehicle is in motion or stationary.  The consumer exposure report assumed that the vehicle is stationary 15 percent of the time that it is in use; that the air conditioning or heat are on 2/3 of the time a vehicle is in operation; and that the average annual consumer vehicle operation time is 526 hours.

      	Leak episodes per vehicle:  The consumer exposure report assumed that there are 0.76 leaks/yr/vehicle.

      Leaks from the evaporator case result in infiltration of the entire release into the passenger compartment, and therefore the highest consumer exposures.  Therefore, the remainder of this discussion focuses on evaporator case leaks.  The consumer exposure report's emission calculations for evaporator case leaks are as follows:

            Emissions, g/yr = (oz. charge/vehicle class) x (0.50 fraction emitted) x (28.35 g/oz) x (0.76 leaks/yr)

                  Size class			Emissions, g/yr
                  Subcompact				290
                  Compact				409
                  Midsize				506
                  Large					602


      When a leak occurs in the evaporator case, the rate of air intake affects the concentration of the refrigerant in the passenger compartment by either circulating the compartment air out of the vehicle (high ventilation rate) or letting the air stand in the compartment (low ventilation rate), allowing refrigerant concentrations to build up.  The consumer exposure report assumed that the MAC is working at maximum level while it is on, minimizing ventilation.  The average speed of travel for a vehicle, according to this report, is 30 mph, yielding a 27 ACH (27 full compartment volumes/hour) ventilation rate.  Concentrations are then calculated in terms of buildup of refrigerants in the recirculated air.  Refrigerant concentrations also depend on passenger compartment sizes, which are broken down into the same four units by a previous study (Chilton 1985):  subcompact vehicles, 2.3 m[3];  compact, 2.5 m[3];  midsize, 2.8 m[3];  and large, 3.1 m[3].  The consumer exposure report's calculated passenger compartment concentrations are as follows:

            Concentration, mg/m[3] = (emissions, g/yr) x (1 yr/8760 hrs) x (ventilation rate, 
            1 hr/27) x (1000 mg/g) x (1/passenger compartment volume, m[3]) =

                  Size class			Concentration, mg/m[3]							Subcompact				 0.53
                  Compact				 0.69
                  Midsize				 0.76
                  Large					 0.82

      Comparing these concentrations to reference concentrations (RfCs) for the substitutes (see Chapter 3) shows that the highest estimated passenger compartment concentration (0.82 mg/m[3]) is far below all of the RfCs.

      HCFC-22 is the only possible carcinogen of the four substitutes analyzed in the study.  If the consumer exposure report's exposure assumptions are used (inhalation rate is 0.6 m[3]/hr for a driving adult; average consumer vehicle operation time is 526 hours per year), the lifetime cancer risk is 2.0 x 10[-6].

6.2.2	Comparison of Parameters Contained in Consumer Exposure Report with Data from Other Sources

      Some of the parameters built into the consumer exposure report's MAC analysis (Versar 1990) are also included in EPA's Vintaging Model and the Market Vision Research Report.  Parameters addressed by both the consumer exposure report and one of these other sources include: (1) total MAC charge; (2) total charge lost before leak is detected and fixed; (3) frequency of operational leak events; and (4) emissions to the vehicle passenger compartment.  Below, the values used in these other sources are compared to those used by the consumer exposure report. 

      	Total MAC charge:  While the consumer exposure report assumed 53 oz. maximum for total refrigerant charges, the Vintaging Model uses a less conservative 42.3 oz. upper bound.

      	Total charge lost before the leak is detected and fixed:  The Vintaging Model assumes a loss of 30 percent of the initial charge, substantially less than the more conservative average of 50 percent loss of initial charge used by the consumer exposure report.

      	Frequency of operational leak events:  According to the consumer exposure report, the frequency of leak events is 0.76 leaks/car/year.  This value is based on an Automotive Refrigeration Products Institute study in which 384 recharges were performed on 506 air-conditioned vehicles in one year.  According to the Vintaging Model, 37 percent of cars with MACs have their MACs serviced in any given year.  Thus, the servicing (and therefore, leakage) frequency in the consumer exposure report is about two times higher than the Vintaging Model.

      	Total emissions to the passenger compartment:  The Vintaging Model assumes that upper end operating releases from MACs are about 100 grams per year.  This is much lower than the upper end of 602 grams per year assumed by the consumer exposure report.


      All else being equal, a decrease in the value of any of these parameters would result in reduced risk.  As shown above, the values assigned to these parameters by the other sources are always lower than those used by the consumer exposure report.  Therefore, any adjustments that might be made to the consumer exposure analysis to be consistent with these other sources would only serve to reduce the already-low exposures and associated health risks.

6.2.3	Summary

      The consumer exposure report estimated consumer exposure to emissions of CFC substitutes from MACs.  These exposure levels are below those that would raise concern for human health risks.  An examination of this report's inputs and assumptions indicates that the analysis is conservative, confirming that consumer exposure is not of concern.


	CHAPTER 6 REFERENCES


Automotive Refrigeration Products Institute.  1988.  1988 Private Sector Vehicle Refrigerant Consumption Study.  Published by the Automotive Refrigeration Products Institute, Lanham, MD.

Chilton.  1985.  Chilton's Guide to Air Conditioning Repair and Service.  Chilton Book Co., Radnor, PA.

Market Vision Research Report.  Results from survey work performed for the Automotive Refrigerant Products Institute.

U.S. EPA.  1991.  Risk Assessment of Flammable Refrigerants for Use in Home Appliances, Revised Draft Report.  Global Change Division.  September 10, 1991.

Versar Inc.  1990.  June 19.  Revised Draft Final Report, Consumer Exposure to Chlorinated Fluorocarbons (CFCs) and CFC Substitutes.  



	7.  GENERAL POPULATION EXPOSURE AND RISK ANALYSIS


      This chapter presents a screening level assessment of the risks to the general population from exposure to ambient air releases of substitutes for CFCs in the refrigeration and air conditioning sectors.  EPA previously performed a similar study (EPA 1990a) using data from another report (EPA 1990c).  Since then, more recent data on CFC emissions from refrigeration and air conditioning equipment have been collected and a new analysis conducted.  The evaluation of substitutes has also been expanded to include compounds not previously included in EPA reports.

      This analysis assesses each of 20 main end uses in the refrigeration and air conditioning sectors.  To develop this list of 20 end uses, the 49 end uses covered by the Vintaging Model were subdivided into equipment types with similar charge sizes and emission rates.  For each end use, this report examines compounds identified by EPA as possible substitutes for present CFC use.  The list of end uses and substitutes is presented in Attachment 7-A.

      The substitutes are grouped into four categories for this analysis:  (1) the HCFCs and HFCs; (2) other alternative refrigerants, including ammonia, butane, isobutane, propane, perfluoropropane, cyclopropane, and dimethyl ether; (3) chlorine; and (4) lithium bromide/water absorption and ammonia/water absorption.  Each category is discussed in a separate section below.  Alternatives to mechanical refrigeration (e.g., evaporative coolers and turbochillers) are not analyzed, since these alternatives do not use chemical refrigerants and therefore pose no exposure risks to the general population.

      The four analyses below are presented as screening analyses.  Their purpose is to identify the proposed substitutes for all end uses that may pose any potential concern for human health or environmental risk.  As a result, the analyses employ conservative assumptions that tend to overestimate potential risk.  Considerable uncertainty exists in these analyses.  The reader should recognize the caveats and limitations identified throughout the chapter.

7.1	HCFCs AND HFCs

7.1.1	Approach

      This analysis of the HCFCs and HFCs uses conservative screening assumptions to estimate potential risks.  The basic approach is to:  (1) estimate high-end annual releases of substitutes from equipment manufacturing plants, operating sites, service stations, and recycling centers or salvage yards; (2) use generic dispersion factors to calculate concentrations of the substitutes in air near the release site; and (3) compare these concentrations to health-based reference concentrations (RfCs) and, for carcinogens, calculate cancer risk.  These three steps are discussed in more detail below.

A.	Estimating Ambient Air Releases

      To estimate releases to ambient air, the first step was to identify the activities that lead to emissions and the types of sites at which such activities occur.  Activities through which CFC substitutes are released to the environment include equipment manufacture, installation, operation, servicing, and disposal.  The types of sites at which these activities occur include factories, operating sites, service stations, and recycling centers or salvage yards.


      The second step was to calculate the amount of CFC substitute released during each activity.  For the screening-level analysis, it was assumed that in all cases the amount of substitute released would equal the amount of CFC released presently.  Most CFC release information was derived from input files to the Vintaging Model; some data were received directly from industry associations and individual manufacturers.

      The emissions values used in the exposure assessment for each end use represent the highest values that would be expected to occur as a result of typical refrigeration practices.  Release estimates for each of the 20 end uses were based on the equipment type with the highest emission rate within that end use (generally the equipment with the largest charge size).  Catastrophic scenarios, such as a factory explosion that could cause temporarily high exposures in a nearby town, were excluded.

      Emissions during servicing and disposal are subject to recycling, which can greatly decrease the amount of refrigerant released to ambient air.  Under Section 608 of the Clean Air Act, the recycling of HCFCs became mandatory in 1992.  However, HFC recycling will not be required until 1995.  Therefore, scenarios both with and without recycling were examined for HFCs, but only scenarios with recycling for HCFCs.

      The third step in estimating releases involved using the information generated in the first two steps to calculate the amount of CFC substitute released to the atmosphere, in a year's time, from each type of release site.  The methods used to calculate these releases are summarized below.

      	Releases from factories:  Annual air releases of HCFCs and HFCs from factories were calculated by multiplying the number of units manufactured per factory per year by the amount of substitute released per unit at the factory.  One end use, low pressure centrifugal chillers, was an exception.  On the recommendation of a leading manufacturer, it was assumed that only 20 percent of the units were tested (and therefore charged and recovered) at the factory.  Data derived from the Vintaging Model, and originally from comprehensive industry surveys, were used to estimate the total number of pieces of equipment manufactured per year in the U.S.  In calculating the number of units manufactured per plant per year, unequal market share was assumed (i.e., that some manufacturers are larger and produce more units per year).  More accurate data were used if verifiable information was readily accessible.  Manufacturing emissions include emissions associated with equipment and component testing.  Where applicable, charging and leak testing emissions were also included.

      	Releases from equipment operating sites:  To estimate air releases from operating sites, the highest possible amount of CFC substitute that could be released from an operating site in any given year was calculated, which would be either the year of installation or disposal.  To do this, the releases from equipment operation, servicing, and either installation or disposal were added together, depending upon which of the latter two activities' releases was higher.  ("Disposal" releases occur at the equipment operating site if the charge is drained or removed at the site before the equipment is dismantled, which has been assumed in all cases except for automobile air conditioners (MACs) and household appliances.)


      	Releases from service stations:  This release site applied only to the MACs end use.  Two types of activities that cause emissions typically occur at service stations: the installation of after-market MACs and the servicing of MACs.  Annual air releases associated with each activity were calculated by multiplying the number of MACs installed and serviced annually at service stations by the amount of substitute released during each installation or servicing.  Since Section 609 of the CAA mandates the recycling of refrigerants during the servicing of MACs, emission estimates for servicing assumed that recycling would occur.  The annual installation and servicing releases were added together to derive the total annual releases from the service station.

      	Releases from recycling centers/salvage yards releases:  A centralized equipment disposal scenario was created for each of two types of end uses  household appliances and MACs:

            	The disposal scenario for household appliances was based on an actual site  an Appliances Recycling Centers of America (ARCA) facility in Minneapolis, Minnesota  which processes refrigerators, freezers, room air conditioners, and dehumidifiers.  Releases were calculated by multiplying disposal emissions by the number of appliances processed annually at that site.

            	The disposal scenario for MACs is a large automobile salvage yard.  Releases were estimated by multiplying disposal emissions by the number of cars processed annually that contain MACs which still possess their charge.  Information on these parameters was obtained from a New Jersey Department of Environmental Protection study (Aucott 1990).

      Attachment 7-B details all of the assumptions used for estimating ambient air releases for each end use.

B.	Estimating Downwind Concentrations

      Downwind concentrations were estimated using the Industrial Source Complex Long Term (ISCLT) Dispersion Model, provided by Versar, an Agency contractor, and summarized in Attachment 7-C.  This approach consisted of converting the release estimates from kilograms per year to grams per second, and then multiplying by normalized dispersion factors (in μg/m[3] per g/sec).  These calculations yielded fence-line concentrations in μg/m[3].  Professional judgment was used to determine which factors to apply to each type of site.  

C.	Screening Potential Risks

      To estimate potential risk, exposure information was merged with available toxicity data, as follows:

      	To evaluate noncarcinogenic effects, the modeled exposure concentrations were compared to reference concentrations (RfCs).


      	To calculate cancer risks, EPA's standard risk equation was used.  A lifetime average daily dose of the substitute was estimated using an exposure concentration, an average 20 m[3]/day inhalation rate, and an average body weight of 70 kg.  For factory releases, the calculated cancer risks were adjusted to account for the phaseout of HCFCs.  HCFC-22 (a long-lived HCFC) will be phased out of new equipment production by 2010; therefore, a person living downwind of a factory producing equipment using HCFC-22 would be exposed to the substitute for only 18 years (1992 to 2010), instead of the full 70 years which the cancer risk equations assume.  Thus, a correction factor of 18/70 was used to calculate cancer risks for HCFC-22 from factories.  Similarly, for HCFC-123, a short-lived HCFC scheduled to be out of production by 2015, a correction factor of 23/70 was applied to the calculated cancer risks.  The reader should note that the projected dates for these accelerated phaseouts are tentative and subject to change.

The reference concentrations and cancer potency factors used in this assessment are given in Chapter 3 of this document.

7.1.2	Results

      All end uses have calculated exposure concentrations well below the reference concentrations for each HCFC and HFC substitute.  Estimated cancer risks from HCFC-22 and HCFC-123 were highest during factory releases, but still below 1E-05, and are therefore not considered to be of concern.  Exhibit 7-1 summarizes the results, showing the highest exposure concentration for each substitute and the end use responsible.  Complete results for each end use and potential substitute can be found in Attachment 7-D.

7.1.3	Caveats and Limitations

      This screening analysis of risks to the general population is a simple screening analysis that incorporates a variety of simplifying assumptions.  The approach was intended to be conservative, i.e., to overestimate exposure point concentrations and potential human health risks.  The major caveats and limitations of the analysis are listed below.

A.	Estimating Releases.  It was necessary to make assumptions to characterize releases, and the assumptions that would lead to the most conservative risk estimates were generally chosen.

      	Within each end use grouping, there may have been several types of equipment with slightly varying charge sizes.  It was assumed that the charge size of all equipment was equal to the largest charge of a piece of equipment in that category.  In about half of the end uses, this assumption is true, but for the other half this is a very conservative assumption.  For example, only 10 percent of commercial direct air conditioning equipment has the assumed charge size of 33 kg.  The other 90 percent has a charge of only 10.2 kg, and so the charge size (and thus releases) for 90 percent of commercial direct air conditioning equipment has
      



	EXHIBIT 7-1
	RESULTS SUMMARY FOR HCFCs AND HFCs

                                  Substitute
                        Highest Exposure Concentration/
                                 RfC Ratio[a]
                              Highest Cancer Risk
                            Associated End Use/Site
HCFC-22
                                   9.7 E-03
                                   7.1 E-06
Central A/C and Home Heat Pumps/Factory
HCFC-123
                                   1.0 E-02
                                   4.3 E-06
High Pressure Centrifugal Chillers/Factory
HCFC-124
                                   1.1 E-03
                                      N/A
Retail Food Stand Alone/Factory
HCFC-141b
                                   2.1 E-04
                                      N/A
Low Pressure Centrifugal Chillers/Factory
HCFC-142b
                                   3.2 E-03
                                      N/A
Refrigerators and Freezers/Factory
HFC-23
                                   8.1 E-02
                                      N/A
Central A/C and Home Heat Pumps/Factory
HFC-125
                                   4.9 E-02
                                      N/A
Central A/C and Home Heat Pumps/Factory
HFC-134
                                   1.0 E-02
                                      N/A
High Pressure Centrifugal Chillers/Factory
HFC-134a
                                   4.9 E-02
                                      N/A
Central A/C and Home Heat Pumps/Factory
HFC-143
                                   1.4 E-03
                                      N/A
High Pressure Centrifugal Chillers/Factory
HFC-143a
                                   6.9 E-03
                                      N/A
Central A/C and Home Heat Pumps/Factory
HFC-152a
                                   9.5 E-03
                                      N/A
Window A/C Units/Salvage Yard
HFC-227ea
                                   1.1 E-02
                                      N/A
Refrigerators and Freezers/Factory

N/A =  Not Applicable

[a]	Some of the RfCs are Agency-verified while others are not.  See Exhibit 3-1 for further information.

            been overestimated.  See Attachment 7-E for a table detailing the charge sizes for the actual equipment versus the assumed charge size for each end use.

      	The number of units manufactured per factory is not tracked by the Vintaging Model, so estimates for this analysis had to be developed.  Data for production rates were based on estimated market shares for the largest manufacturer.  For example, if there are three manufacturers of chillers, it was assumed, based on information from these manufacturers, that the largest manufacturer has 50 percent of the market share and of total annual production.  It was also assumed that this manufacturer's total annual production was manufactured at one factory.  These assumptions will, in many cases, result in an overestimation of the true production (and thus releases) at the actual largest factory.

      	With one exception (low pressure chillers), it was assumed that every piece of equipment is charged (and recovered) at least once in the factory for run testing, although some manufacturers' quality control may include testing only a portion of all equipment sold.

      	It was assumed that the equipment contains its full original charge at the point of service.  In practice, this would rarely be the case since most servicing calls are made after leaking of refrigerant causes the equipment to malfunction.  The effect of this assumption is to overstate releases, since the amount released is a percentage of the charge at the time of servicing.

      	Disposal emissions also assume that a percentage of the total charge is vented at disposal, although this is rarely the case.  As in servicing, a portion of the charge will have leaked before disposal.

      	Reports from manufacturers of certain types of equipment, such as chillers, indicate that refrigerant leakage from operating equipment varies significantly depending on whether the equipment's level of maintenance is poor, average, or above average.  Poor maintenance was assumed in all cases.  An example of the amount of overestimation of releases which this approach could yield can be seen in chillers, of which only five percent to ten percent are generally poorly maintained.

      	The amount of substitute used may not equal the amount of CFC used, due to the use of blends, increased efficiency in newly-built equipment, and the increased cost of the substitutes and the concomitant smaller equipment charge sizes.  Also, some substitutes, such as HCFC-123, are already in use today and are known to be handled very differently from CFCs (i.e., they are handled more carefully).  In the case of HCFC-123, a possible carcinogen, the actual risks will be smaller than those predicted.

B.	Estimating Downwind Concentrations.  Uncertainties associated with derivation of the dispersion factors are discussed by EPA (EPA 1990b).  They include the following:

      	EPA used a generic site location to model downwind concentrations.  The meteorological conditions at this generic site were selected to cause conservative estimates of downwind concentrations.

      	EPA's selection of model parameters characterizing the release (e.g., release height, size of the release facility) were based on professional judgment.


      	Facility fenceline distance estimates (where exposure was simulated to occur) were also assumed by EPA.  For some types of release sites, the model could not estimate concentrations at the assumed fenceline distance because in some cases the distances were as close as 10 meters.  In these cases, the modeled concentrations at the exposure point were based on the highest concentrations predicted, but could be underestimates.

      	The uncertainty faced in selecting dispersion factors to apply to each type of release site is discussed in detail in Attachment 7-C.

C.	Screening Potential Risks.  There are three important caveats associated with the risk calculations:

      	Most of the calculations assume that an exposed individual lives and works in the same zone of concentration 24 hours per day, year-round, throughout a 70-year lifetime.  Although this conservative assumption is often used in risk assessments, it is overly conservative for the HCFCs, which are scheduled to be phased out of production.  To account for the phaseout in HCFC production, it was assumed that equipment manufacturing facilities would stop placing HCFC-22 and HCFC-123 in equipment by 2010 and 2015, respectively.  Therefore, when calculating cancer risk, the duration of exposure to factory releases of these substitutes was shortened to 18 and 23 years, respectively.  A similar adjustment was not made in exposures to operating site releases because the phaseout will not preclude HCFCs from being used in existing equipment for the length of the equipment's lifetime.  However, because the equipment is likely to replaced before 2062  the end of the exposed individual's 70-year lifetime  cancer risks from releases of HCFC-22 and HCFC-123 from operating sites are overestimated in this analysis.

      	Risks from exposure to operating site releases were based on the highest amount of substitute that would be released in any one year (i.e., the sum of releases from equipment operation, servicing, and either installation or disposal).  Because installation and disposal each happen only once over the equipment's lifetime, and servicing typically occurs less frequently than once a year, exposures and risks from operating sites are overestimated.

7.1.4	Conclusions

      Even employing conservative screening assumptions, EPA found that there are no end uses within the refrigeration or air conditioning sectors where the use of HFCs or HCFCs as substitutes for CFCs will result in exposure concentrations above the chemicals' health-based reference concentrations for non-cancer effects.  Similarly, there are no areas for which predicted cancer risks would be above the upper-bound cancer risk level of concern, 10[-5].

7.2	OTHER ALTERNATIVE REFRIGERANTS

      This section discusses the following refrigerant substitutes:  ammonia, butane, isobutane, propane, cyclopropane, perfluoropropane, and dimethyl ether (DME).  After a background discussion, the approach for the analysis and the results are discussed.

7.2.1	Background

      Many of these alternative refrigerants currently find wide use in industrial applications.  Ammonia was one of the first refrigerants to gain acceptance, and is used almost exclusively in the meat packing, dairy, frozen juice, brewery, cold storage, and other food industries.  The hydrocarbon refrigerants are used primarily in oil refineries and chemical plants, where they are frequently available as part of the process stream.  The expanded use of many of the refrigerants in this section in other refrigeration and air conditioning equipment has been proposed, particularly as components in blends with the HFCs and HCFCs.
      
      Most industrial process refrigeration systems are designed by consulting firms that specialize in the field.  The consultants typically work with the client to determine the exact specifications (cooling loads, for example), then design the system using off-the-shelf or specially- fabricated components.  Smaller, more standard designs, such as those for a cold storage warehouse, are more likely to use off-the-shelf components as these are cheaper and more readily available.  The consultants sometimes supervise construction, installation, and start-up of the refrigeration system.

      Compressors for large industrial process refrigeration systems are built by a handful of companies.  Because they are large, complicated pieces of machinery, compressors are tested at the factory, but for safety purposes testing is not done with the refrigerant that will ultimately be used.  For hydrocarbon refrigerants, run testing is typically done with dry air, and a correlation made with the refrigerant to determine the compressor's operating performance (Shephard 1992).  Ammonia compressors are tested with air and a soap solution or tested under water to see if bubbles appear at joints.  Ammonia package systems (those that have the evaporating and condensing units attached) are similarly tested with an air/soap solution, then charged with dry nitrogen before being shipped (Anderson 1992; Pillis 1992).

      Industrial systems are erected in the field, leak tested again, then charged with the refrigerant.  Care is taken during the charging operation to ensure that vapor concentrations are kept below hazardous levels.  Ammonia systems usually have a charge size that makes them subject to OSHA PEL limits for ammonia (Anderson 1992), while ASHRAE standards for refrigeration systems recognize that worker exposure due to routine operating leakage from the system, or that occurs during servicing, should be kept below threshold levels.  These codes are adopted into local and state building codes in most of the U.S. and are used as guidelines by the system designers.  In all cases, it is unlikely that emissions from a process refrigeration system operating on an industrial refrigerant would exceed those of a similar system operating on CFCs or HCFCs. 

7.2.2	Approach

      This analysis of ammonia, butane, isobutane, propane, cyclopropane, perfluoropropane, and DME is a simple screening analysis to determine whether general population exposure is potentially of concern in specific end uses.  The analysis used the fenceline exposure concentrations predicted above for HCFCs and HFCs and assumed that the fenceline concentrations of the alternative refrigerants noted above would be the same.  This is a conservative approach because (1) the HCFC and HFC fenceline concentrations were calculated using conservative assumptions that are likely to overestimate exposure levels, and (2) releases of these alternative refrigerants are likely to be even lower because of the OSHA and ASHRAE standards to which their refrigerant uses are subject.

      Conclusive data on the significance of factory releases were unavailable.  Consequently, to be conservative in the analysis for these alternative refrigerants, factory releases were evaluated in addition to releases on-site during operation and servicing and those at disposal.  The only exception to this approach was for ammonia.  Although compressors for large industrial process refrigeration systems are tested at the factory, for safety purposes this testing is not done with ammonia.  Similarly, charging of ammonia systems is typically performed at the operating site rather than the factory.  Therefore, ammonia is unlikely to be released from a factory setting.

      These fenceline concentrations were then compared to reference concentrations for each of the substitutes.  From Chapter 3, the following reference concentrations were employed:


      Ammonia:		0.1 mg/m[3]		Perfluoropropane:	4.5 mg/m[3]

      Butane:		0.95 mg/m[3] 		Cyclopropane:		0.9 mg/m[3]

      Isobutane:		0.95 mg/m[3]		Dimethyl ether:	0.4 mg/m[3]

      Propane:		0.9 mg/m[3]

7.2.3	Results

      Exhibit 7-2 summarizes the results of the approach described in 7.2.2.  For each end use, the highest exposure concentration (from Attachment 7-D) is shown.  The location of this exposure (i.e., at the factory, on-site, or at the disposal site) is also reported.  For the applicable refrigerants, the ratio of the concentration to the substitutes' reference concentrations is calculated.  None of these ratios exceeded the health-based thresholds.  

      Caveats and limitations associated with this analysis are the same as those discussed above for HCFCs and HFCs.  In addition, there is considerable uncertainty associated with the health-based thresholds used for butane, propane, and perfluoropropane, which were extrapolated from workplace exposure limits rather than calculated from laboratory studies.  EPA believes that the exposure limits used in this analysis are conservative.

      Given the conservative nature of the release estimates and interim health-based levels used in this analysis, EPA concludes that general population exposure is not of concern for these seven substitutes.

7.3	CHLORINE

      Chlorine has been proposed as a Class I substitute refrigerant for use in chlorine liquification, a processing step in the manufacture of the chemical.  When chilled below its boiling point, chlorine can be stored as a liquid at atmospheric pressure, a method that for safety reasons is much preferred to storing the chemical as a pressurized gas at ambient temperatures.  Compatibility of the refrigerant with liquid chlorine is critical because of chlorine's high reactivity; CFC-12 is widely used because it is nonreactive.

	EXHIBIT 7-2:  RESULTS SUMMARY FOR OTHER ALTERNATIVE REFRIGERANTS[a]




































a Releases are based on the assumption that the refrigerant is not recycled when the equipment is serviced or dismantled for disposal.  This does not reflect common operating practices and therefore overstates true risk.

      Chlorine refrigeration systems generally do not exchange chlorine with the process stream.  However, the system would be placed so that any leakage or losses of chlorine would be contained and neutralized by the process safety mechanisms.  Charge sizes in the refrigeration system would probably be less than 1000 kg, as compared to a bulk storage container havingseveral hundred times that capacity.  Chlorine emissions from the refrigeration system are likely to be dwarfed by those emanating from the process and storage systems, which are already kept small for reasons of worker safety.

      Based on these considerations, EPA does not believe that the use of chlorine as a refrigerant in this limited application  the manufacture of chlorine  poses significant incremental risk to the general population.

7.4	LITHIUM BROMIDE AND AMMONIA IN ABSORPTION SYSTEMS

      Lithium bromide is used in commercial absorption systems, where it serves as an absorbent.  Such systems operate at very low pressure to allow water to act as a refrigerant.  The lithium bromide is in the form of an ionic salt having a negligible vapor pressure, so there is no possibility of a gaseous release in other than catastrophic circumstances.  Lithium is also a relatively nontoxic, nonflammable, nonexplosive, chemically stable chemical, and poses little incremental risk to the general population.

      Absorption refrigeration systems employing a mixture of ammonia and water have been used for many years.  Small refrigeration units are popular in recreational vehicles and hotel refrigerators as they need no electrically-driven mechanical pump.  Domestic absorption type systems use hydrogen to maintain a system pressure high enough to allow the ammonia refrigerant to evaporate at low pressure and temperature (and condense at room temperature), and are  ruggedly constructed; the absorption mechanism itself is a sealed unit, which usually needs no servicing over its operating life.  Commercial ammonia absorption systems are used for air conditioning comfort cooling, particularly in instances where waste heat is available.  These produce chilled water, which is circulated to the space being cooled.  The same standards and guidelines that apply to ammonia vapor compression equipment also apply to commercial ammonia absorption systems.  In a well-maintained system, it is unlikely that ammonia emissions would pose a problem.

	CHAPTER 7 REFERENCES


Anderson, Kent.  1992.  May 5. Phone conversation between Mark Radka, ICF Incorporated, and Kent Anderson, Chairman of the International Institute of Ammonia Refrigeration.

Aucott, Michael.  1990.  December 14.  Feasibility of CFC-12 Reclamation from Salvaged Vehicles.  Department of Environmental Protection, Division of Solid Waste Management, State of New Jersey.

Pillis, J.  1992.  May 5.  Phone conversation between Mark Radka, ICF Incorporated, and J. Pillis, Chief Engineer, The Frick Company.

Shephard, James.  1992.  May 4.  Phone conversation between Mark Radka, ICF Incorporated, and James Shephard, Vice President of Engineering, Lewis Energy Systems, and Chair of the ASHRAE Technical Committee on Custom Refrigeration Systems.

U.S. EPA.  1992.  Regulatory Impact Analysis:  Compliance with Section 604 of the Clean Air Act for the Phaseout of Ozone Depleting Chemicals.  Office of Air and Radiation.  March 12, 1992.

U.S. EPA.  1990a.  External Review Draft, Hydrofluorocarbons and Hydrochlorofluorocarbons, Interim Report.  Office of Toxic Substances.  November, 1990.

U.S. EPA.  1990b.  Potential Ambient Inhalation Exposures for Chlorofluorocarbon Substitutes.  	Office of Toxic Substances.  Exposure Evaluation Division, Exposure Assessment Branch.  November 7, 1990.

U.S. EPA.  1990c.  Revised Draft: Occupational Exposure and Environmental Release Data for Chlorofluorocarbons (CFCs) and Their Substitutes.  Office of Toxic Substances.

	ATTACHMENT 7-A

	DETAILED CATEGORIZATION OF REFRIGERATION END USES
	AND EQUIPMENT TYPES
	("Near-Term" Substitutes In Italics) 


1.	Retail Food - Parallel Systems

   	CFC-12 (medium temperature)  	HFC-134a 
                                          HCFC-22 
                                          HCFC-22/HFC-152a/HCFC-124
                                          HFC-125/HFC-143a/HFC-134a
                                          ammonia

   	CFC-502 (low temperature)   	HFC-125 
                                          HFC-143a
                                          HCFC-22 
                                          HCFC-22/propane/perfluoropropane
                                          HCFC-22/propane/HFC-125

   	HCFC-22 (medium and low temperature)  	HFC-125
                                          HFC-134a
                                          HFC-23/HFC-134a
                                          HFC-125/HFC-143a
                                          HCFC-22/propane/HFC-125


2.	Retail Food - Stand Alone Systems

   	CFC-12 (medium temperature)   	HFC-134a 
                                          HCFC-22 
                                          HCFC-22/HFC-152a/HCFC-124
                                          HFC-125/HFC-143a/HFC-134a
                                          ammonia

   	CFC-502 (low temperature)   	HFC-125 
                                          HFC-143a
                                          HCFC-22 
                                          HCFC-22/propane/perfluoropropane
                                          HCFC-22/propane/HFC-125


3.	Cold Storage Warehouses

   	CFC-12 (medium/high temperature)   	HFC-134a 
                                          HFC-152a 
                                          HCFC-22
                                          HCFC-22/HFC-152a/HCFC-124 
                                          ammonia
                                          turbochill



   	HCFC-22 (medium/high temperature)   	HFC-134a
                                          HFC-152a
                                          HFC-23/HFC-134a
                                          
HCFC-22/HFC-152a/HCFC-124 
                                          ammonia
                                          turbochill

   	CFC-502 (low temperature)   	HFC-125 
                                          HCFC-22
                                          HCFC-22/propane/perfluoropropane        						ammonia


4.	Reciprocating Chillersa

   	CFC-12  	HFC-134a 
                                          HFC-152a 
                                          HCFC-22
                                          HCFC-123
                                          HFC-125/HFC-143a/HFC-134a

   	HCFC-22 	HFC-134a 
                                          HFC-23/HFC-134a
                                          HFC-125/HFC-143a
                                          HCFC-22/propane/HFC-125
                                          ammonia


5.	Screw Chillers[a]

   	HCFC-22   	HFC-134a
                                          HFC-23/HFC-134a
                                          HCFC-22/propane/HFC-125
                                          ammonia 


6.	Centrifugal Low Pressure Chillers[a]

   	CFC-11  	HFC-134a
      	HCFC-22
                                          HCFC-123
                                          HCFC-141b
                                          lithium bromide/water absorption
                                          ammonia/water absorption



7.	Centrifugal High Pressure Chillers[a]

   	CFC-12   	HFC-134a
                                          HCFC-22
                                          HCFC-123
                                          lithium bromide/water absorption

   	CFC-114   	HCFC-124
                                          HCFC-142b
                                          R200b
                                          R200c
                                          R200d
                                          R200e
                                          R200f
                                          R200g
                                          R200i
                                          R200j

   	CFC-500   	HFC-134a
                                          HCFC-22/HFC-152a/HCFC-124 
                                          HFC-134a/HFC-152a
                                          ammonia

   	HCFC-22   	HFC-134a 
                                          ammonia 


8.	Industrial Process Refrigeration - Low Pressure Systems

   	Centrifugal Package Systems (CFC-11)   	HCFC-123
                                          HCFC-141b

   	Centrifugal Built-Up Systems (CFC-11)   	HCFC-123
                                          HCFC-141b


9.	Industrial Process Refrigeration - High Pressure Package Systems

   	CFC-12 Reciprocating  	HFC-134a 
                                          HFC-143a
   	HCFC-22 Reciprocating	HFC-152a
                                          HCFC-22 
   	CFC-502 Reciprocatinga	HFC-23/HFC-134a
                                          HFC-125/HFC-143a
                                          HCFC-22/HFC-152a/HCFC-124
                                          HFC-125/HFC-143a/HFC-134a
                                          HCFC-22/propane/HFC-125
                                          propane 
    		butane 
                                          ammonia 


10.	Industrial Process Refrigeration - High Pressure Built-Up Systems


   	CFC-12 Reciprocating	HFC-134a 
                                          HFC-152a
   	HCFC-22 Centrifugal	HCFC-22
                                          HFC-23/HFC-134a
   	CFC-500 Centrifugal	HFC-125/HFC-143a
                                          HCFC-22/HFC-152a/HCFC-124
   	CFC-12 Reciprocating	HFC-125/HFC-143a/HFC-134a
                                          HCFC-22/propane/HFC-125
   	HCFC-22 Reciprocating 	propane 
butane 
                                          ammonia 
                                          chlorine


11.	Ice Skating Rinks

   	CFC-12	HFC-134a
                                          HFC-152a
   	HCFC-22	HCFC-22
                                          HFC-23/HFC-134a
   	CFC-502	HCFC-22/HFC-152a/HCFC-124
                                          ammonia



12.	Medium Ice Makers

   	CFC-12	HCFC-22
                                          HFC-134a
                                          HFC-152a
                                          HCFC-22/HFC-152a/HCFC-124
                                          ammonia

13.	Large Ice Makers

   	CFC-12	HCFC-22
                                          HFC-134a
                                          HFC-152a
                                          HCFC-22/HFC-152a/HCFC-124
                                          ammonia


14.	Refrigerated Transport Units

   	CFC-12   	HFC-134a 
                                          HCFC-22
HCFC-22/HFC-152a/HCFC-124
                                          HFC-125/HFC-143a/HFC-134a

   	CFC-500   	HFC-134a
                                          HCFC-22
                                          HCFC-22/HFC-152a/HCFC-124 
                                          HCFC-22/propane/HFC-125

   	CFC-502   	HFC-125
   
                                          HFC-134a
                                          HCFC-22
                                          HCFC-22/propane/HFC-125


15.	Window Air Conditioners

   	HCFC-22   	HFC-125
                                          HFC-134a 
                                          HFC-23/HFC-134a
                                          HFC-125/HFC-143a


16.	Central Air Conditioners and Home Heat Pumps

   	Package Terminal Air 	HFC-125
     Conditioners (HCFC-22)			HFC-134a
                                          HFC-23/HFC-134a	
   	Residential Unitary Air 	HFC-125/HFC-143a
     Conditioners (HCFC-22)	HFC-125/HFC-143a/HFC-134a
                                          HCFC-22/propane/HFC-125
   	Unitary Heat Pumps (HCFC-22)	

   	Water Source Heat Pumps (HCFC-22)


17.	Commercial Direct Expansion Air Conditioning Systems 

   	Small Commercial Air 	HFC-125 
     Conditioners (HCFC-22)			HFC-134a
                                          HFC-23/HFC-134a
   	Large Commercial Air	HFC-125/HFC-143a 
     Conditioners (HCFC-22)	


18.	Refrigerators and Freezers 

   	Refrigerators (CFC-12)   	HFC-134a 
                                          HFC-152a
   	Freezers (CFC-12)	HCFC-22
R200b
                                      	HCFC-22/HFC-152a/HCFC-124 
                                          HCFC-22/HCFC-142b 
                                          ammonia/water absorption


19.	Other Refrigerated Appliances

   	Dehumidifiers (CFC-12) 	HFC-134a 
                                          HFC-152a 
   	Vending Machines (CFC-12)	HCFC-22/HFC-152a/HCFC-124

   	Small Ice Makers (CFC-12)	HCFC-22 

   	Water Coolers (CFC-12)

   	Small Ice Makers (CFC-502)   	HCFC-22 


20.	Mobile Air Conditioners

   	CFC-12   	HFC-134a 
                                          HCFC-22/HFC-152a/HCFC-124




	ATTACHMENT 7-B:
	ASSUMPTIONS FOR AMBIENT AIR RELEASE CALCULATIONSa

                              RETAIL FOOD SYSTEMS
                             Type/Site of Release
                               Parallel Systems
                              Stand-Alone Systems
MANUFACTURING RELEASES - FACTORY
                                                                               
                                                                               
   No. of units manufactured per plant per year
                                                                          2,900
       30,000b
   Release per unit manufactured (kg)
                                                                            6.0
                                                                            2.0
ON-SITE RELEASES
                                                                               
                                                                               
   No. of units per site

                                                                              1
                                                                              2
   Release per unit installed (kg)

                                                                             93
                                                                            3.0
   Release per unit serviced (kg)
                                                                               
                                                                               
     without recycling
                                                                            7.7
                                                                             20
     with recycling
                                                                            3.1
                                                                            1.2
     recycling efficiency

                                                                            99%
                                                                            99%
   Release per unit operating (kg/yr)

                                                                            130
                                                                            1.9
   Release per unit disposed (kg)
                                                                               
                                                                               
     without recycling
                                                                            470
                                                                             19
     with recycling
                                                                            4.7
                                                                           0.19
     recycling efficiency
                                                                            99%
                                                                            99%


                            COLD STORAGE WAREHOUSES
                             Type/Site of Release
                            Cold Storage Warehouses
MANUFACTURING RELEASES - FACTORY
                                                                               
   No. of units manufactured per plant per year
                       57a
   Release per unit manufactured (kg)
                                                                             15
ON-SITE RELEASES
                                                                               
   No. of units per site

                        4b
   Release per unit installed (kg)

                                                                            3.5
   Release per unit serviced (kg)
                                                                               
     without recycling
                                                                             52
     with recycling
                                                                            5.0
     recycling efficiency

                                                                            99%
   Release per unit operating (kg/yr)

                                                                             50
   Release per unit disposed (kg)
                                                                               
     without recycling
                                                                            250
     with recycling
                                                                            2.5
     recycling efficiency
                                                                            99%

                                   CHILLERS
                                       
                                       
                             Type/Site of Release
                                   Recipro-
                                cating Chillers
                                       
                                Screw Chillers
                             Centrifugal Chillers

                                                                               
                                                                               
                                 Low Pressure
                                 High Pressure
MANUFACTURING RELEASES - FACTORY
                                                                               
                                                                               
                                                                               
                                                                               
   No. of units manufactured per plant per year
 3,000a
    230b
1,200c
  370[c]
   Release per unit manufactured (kg)
                                                                            5.0
                                                                            5.1
                                                                            16d
                                                                             52
ON-SITE RELEASES
                                                                               
                                                                               
                                                                               
                                                                               
   No. of units per site

                                                                              1
                                                                              1
                                                                              1
                                                                              1
   Release per unit installed (kg)

                                                                            1.1
                                                                            1.1
                                                                            2.0
                                                                            2.0
   Release per unit serviced (kg)
                                                                               
                                                                               
                                                                               
                                                                               
     without recycling
      5.1e
      11[e]
   28[e]
   37[e]
     with recycling
                                                                           0.15
                                                                           0.21
                                                                            1.3
                                                                            1.4
     recycling efficiency

                                                                            99%
                                                                            99%
                                                                            99%
                                                                            99%
   Release per unit operating (kg/yr)

    23f
      54[f]
  110[f]
  260[f]
   Release per unit disposed (kg)
                                                                               
                                                                               
                                                                               
                                                                               
     without recycling
                                                                             73
                                                                            230
                                                                            450
                                                                            900
     with recycling
                                                                           0.73
                                                                            2.3
                                                                            4.5
                                                                            9.0
     recycling efficiency
                                                                            99%
                                                                            99%
                                                                            99%
                                                                            99%

                       INDUSTRIAL PROCESS REFRIGERATION
                                       
                             Type/Site of Release
                                 Low Pressure
                             High Pressure Package
                            High Pressure Built Up
                               Ice Skating Rinks
                               Medium Ice Makers
                                       
                               Large Ice Makers
MANUFACTURING RELEASES - FACTORY
                                                                               
                                                                               
                                                                               
                                                                               
                                                                               
                                                                               
   No. of units manufactured per plant per year
    32a
      130[a]
  130[a]
     41[a]
50,000b
                                                                           8.0c
   Release per unit manufactured (kg)
                                                                             24
                                                                             13
                                                                            190
                                                                            1.0
                                                                           0.19
                                                                            2.1
ON-SITE RELEASES
                                                                               
                                                                               
                                                                               
                                                                               
                                                                               
                                                                               
   No. of units per site

                                                                              1
                                                                              1
                                                                              1
                                                                              1
                                                                              1
                                                                              1
   Release per unit installed (kg)

                                                                             11
                                                                             11
                                                                             11
                                                                             11
                                                                              0
                                                                           0.68
   Release per unit serviced (kg)
                                                                               
                                                                               
                                                                               
                                                                               
                                                                               
                                                                               
     without recycling
                                                                            140
                                                                             19
                                                                            910
                                                                             28
                                                                            2.3
                                                                             34
     with recycling
                                                                             15
                                                                            2.8
                                                                             10
                                                                            1.3
                                                                           0.05
                                                                           0.68
     recycling efficiency

                                                                            99%
                                                                            99%
                                                                            99%
                                                                            99%
                                                                            99%
                                                                            99%
   Release per unit operating (kg/yr)

                                                                             68
                                                                             27
                                                                          1,800
                                                                             57
                                                                           0.15
                                                                            2.2
   Release per unit disposed (kg)
                                                                               
                                                                               
                                                                               
                                                                               
                                                                               
                                                                               
     without recycling
                                                                            680
                                                                            180
                                                                          9,100
                                                                          1,400
                                                                            2.3
                                                                             34
     with recycling
                                                                            6.8
                                                                            1.8
                                                                             91
                                                                             14
                                                                           0.02
                                                                           0.34
     recycling efficiency
                                                                            99%
                                                                            99%
                                                                            99%
                                                                            99%
                                                                            99%
                                                                            99%

                            TRANSPORT REFRIGERATION
                                       
                             Type/Site of Release
                            Transport Refrigeration
MANUFACTURING RELEASES - FACTORY
                                                                               
   No. of units manufactured per plant per year
                14,100a
   Release per unit manufactured (kg)
                                                                            1.0
ON-SITE RELEASES
                                                                               
   No. of units per site

                                                                              1
   Release per unit installed (kg)

                                                                           0.25
   Release per unit serviced (kg)
                                                                               
     without recycling
                                                                            2.0
     with recycling
                                                                           0.02
     recycling efficiency

                                                                            99%
   Release per unit operating (kg/yr)

                                                                           0.20
   Release per unit disposed (kg)
                                                                               
     without recycling
                                                                            8.1
     with recycling
                                                                           0.08
     recycling efficiency
                                                                            99%


                  HOUSEHOLD AND OTHER REFRIGERATED APPLIANCES
                                       
                                       
                             Type/Site of Release
                                       
                               Window A/C Units
                                       
                       Central A/C & Home Heat Pumps
                        Commercial Direct Expansion A/C
MANUFACTURING RELEASES - FACTORY
                                                                               
                                                                               
                                                                               
   No. of units manufactured per plant per year
300,000a
  400,000[a]
   38,600b
   Release per unit manufactured (kg)
                                                                           0.05
                                                                           0.23
                                                                           0.33
ON-SITE RELEASES
                                                                               
                                                                               
                                                                               
   No. of units per site

                                                                              1
                                                                              1
                                                                              1
   Release per unit installed (kg)

                                                                              0
                                                                              0
                                                                             11
   Release per unit serviced (kg)
                                                                               
                                                                               
                                                                               
     without recycling
                                                                            1.8
                                                                            3.8
                                                                             33
     with recycling
                                                                           0.04
                                                                           0.08
                                                                           0.66
     recycling efficiency

                                                                            99%
                                                                            99%
                                                                            99%
   Release per unit operating (kg/yr)

                                                                           0.38
                                                                           0.80
                                                                            6.9
   Release per unit disposed (kg)
                                                                               
                                                                               
                                                                               
     without recycling
                                                                             NA
                                                                            3.8
                                                                             33
     with recycling
                                                                             NA
                                                                           0.04
                                                                           0.33
     recycling efficiency
                                                                             NA
                                                                            99%
                                                                            99%
DISPOSAL RELEASES - RECYCLING/SALVAGING CENTER
                                                                               
                                                                               
                                                                               
   No. of units processed per site per year
  26,000c
                                                                             NA
                                                                             NA
   Release per unit processed (kg)
                                                                               
                                                                               
                                                                               
     without recycling
                                                                            1.8
                                                                             NA
                                                                             NA
     with recycling
                                                                           0.02
                                                                             NA
                                                                             NA
     recycling efficiency
                                                                            99%
                                                                             NA
                                                                             NA

NA = Not applicable

                  HOUSEHOLD AND OTHER REFRIGERATED APPLIANCES
                                       
                             Type/Site of Release
                                       
                          Refrigerators and Freezers
                         Other Refrigerated Appliances
MANUFACTURING RELEASES - FACTORY
                                                                               
                                                                               
   No. of units manufactured per plant per year
 750,000a
 170,000b
   Release per unit manufactured (kg)
                                                                           0.04
                                                                           0.04
ON-SITE RELEASES
                                                                               
                                                                               
   No. of units per site

                                                                              1
                                                                              1
   Release per unit installed (kg)

                                                                              0
                                                                              0
   Release per unit serviced (kg)
                                                                               
                                                                               
     without recycling
                                                                           0.41
                                                                           0.50
     with recycling
                                                                           0.02
                                                                           0.03
     recycling efficiency

                                                                            95%
                                                                            95%
   Release per unit operating (kg/yr)

         0.40c
         0.50[c]
   Release per unit disposed (kg)
                                                                               
                                                                               
     without recycling
                                                                             NA
                                                                           0.50
     with recycling
                                                                             NA
                                                                           0.03
     recycling efficiency
                                                                             NA
                                                                            95%
DISPOSAL RELEASES - RECYCLING/SALVAGING CENTER
                                                                               
                                                                               
   No. of units processed per site per year
   17,000d
                                                                         2,400e
   Release per unit processed (kg)
                                                                               
                                                                               
     without recycling
                                                                           0.40
                                                                           0.41
     with recycling
                                                                           0.02
                                                                           0.02
     recycling efficiency
                                                                            95%
                                                                            95%

NA = Not applicable


                            MOBILE AIR CONDITIONERS
                                       
                             Type/Site of Release
                            Mobile Air Conditioners
MANUFACTURING RELEASES - FACTORY
                                                                               
   No. of units manufactured per plant per year
              325,000a
   Release per unit manufactured (kg)
                                                                           0.11
INSTALLATION RELEASES - SERVICE STATION
                                                                               
   No. of units per site
                     63b
   Release per unit installed (kg)
                                                                           0.11
SERVICING RELEASES - SERVICE STATION
                                                                               
   No. of units serviced per site per year
                   165c
   Release per unit serviced (kg)
                                                                               
     without recycling
                                                                            1.2
     with recycling
                                                                           0.06
     recycling efficiency
                                                                            95%
DISPOSAL RELEASES - RECYCLING/SALVAGING CENTER
                                                                               
   No. of units processed per site per year
                 4,250d
   Release per unit processed (kg)
                                                                               
     without recycling
                                                                            1.2
     with recycling
                                                                           0.06
     recycling efficiency
                                                                            95%



                                ATTACHMENT 7-C
                                       
METHODOLOGY FOR ESTIMATING GENERAL POPULATION EXPOSURES TO AMBIENT AIR RELEASES OF SUBSTITUTES FOR CLASS I SUBSTANCES
                             FOR SNAP RISK SCREENS


   The approach used to estimate general population exposures to ambient air releases of substitutes for Class I ozone-depleting substances (ODSs) for the SNAP risk screens is patterned after the approach EPA used previously (EPA 1990a) for the interim report on HFCs and HCFCs (EPA 1990b).  The sections below summarize EPA's earlier approach and discuss how it was applied to the SNAP risk screens. 

SUMMARY OF EPA'S EARLIER APPROACH FOR ESTIMATING GENERAL POPULATION EXPOSURES

   EPA's approach for estimating general population exposures for the interim report on HFCs and HCFCs is summarized in the two documents referred to above.  EPA used information developed by PEI (EPA 1990c), combined with engineering judgment, to develop estimates of ambient air releases from each of the activities shown in Exhibit 7-C-1.  EPA then used the Industrial Source Complex Long Term (ISCLT) Dispersion Model to estimate ambient air concentrations at an assumed exposure point.

   The ISCLT model is a steady-state Gaussian plume model which can be used to assess pollutant concentrations from a wide variety of sources associated with an industrial source complex.  Additional information on the model can be found in EPA's Industrial Source Complex (ISC) Dispersion Model User's Guide (EPA 1986) and addendum (EPA 1987).  The model has not been through Science Advisory Board or formal external peer review, but it has been reviewed and recommended for general use by EPA's Office of Air Quality Planning and Standards for applications similar to this.  

   To run the ISCLT model, it was necessary for EPA to develop inputs for source characteristics and meteorological factors.

   	EPA used engineering judgment to develop assumptions about the characteristics of the facilities (e.g., stack height, surface area) from which the ODSs are released.

   	EPA used meteorological conditions for a "generic" site, which was selected based on an analysis of maximum exposed individual concentrations calculated for an identical release from each of 392 sites across the U.S.  The meteorological conditions at this generic site were known to cause conservative estimates of exposure point concentrations.

EPA then used the ISCLT model outputs to estimate concentrations at the fenceline of the release site.  Distances to the fenceline varied depending on the type of release site, and were based on professional judgment.  EPA points out that for some types of release sites, the model was unable to predict concentrations as close as the assumed fenceline distance.  In these cases, the exposure estimates were based on the highest concentrations predicted by the model.

   EPA documents several areas of uncertainty associated with the release, concentration, and exposure estimates; these are not repeated here.

APPLYING EPA'S EARLIER APPROACH TO THE SNAP RISK SCREENS

   For the SNAP risk screens, general population exposures to ambient air releases of Class I substances were estimated using three basic steps:


   (1)	The first step was to update the ambient air release information presented in the EPA report (EPA 1990c) using more recent data.  Conservative estimates were made of annual ambient air releases of ODS substitutes from different types of facilities (e.g., chemical manufacturing plants, factories that produce equipment containing the substitutes, sites at which the substitutes are used, service stations, centralized recycling or disposal facilities);

   (2)	Each type of release site was "matched up" with one that was modeled earlier by EPA; and

   (3)	The results of the previous ISCLT runs were used to determine the relationship between the release rate and exposure point concentration for each type of release site; that factor was then applied to the revised release estimates to derive a new fenceline concentration.

Steps (1) and (3) were straightforward.  The approach used to calculate releases from different types of facilities (Step 1) is described in the section documenting the general population exposure analysis for each use sector.  To determine the relationship between releases and fenceline concentrations (Step 3), EPA's contractor for the earlier analysis (Versar) provided a table identifying the "normalized maximum concentration at the fenceline, g/m[3]" for each of the runs conducted for the interim report.  These normalized maximum concentrations, shown in Exhibit 7-C-1, are the maximum fenceline concentrations that result from the release of one gram per second (g/sec) of ODS from a source.  Using them is simply a matter of converting estimated annual releases to g/sec and multiplying them by the normalized concentration, yielding maximum fenceline concentrations in g/m[3].

   Some difficulties were encountered in the second step, however, which add to the uncertainties associated with the ambient air modeling.  These difficulties arise from two factors, discussed below.

   	First, it was sometimes difficult to identify the type of facility associated with each normalized concentration listed in Exhibit 7-C-1.  This is because the normalized concentrations are presented by type of activity (e.g., manufacturing, servicing) rather than type of source (e.g., factory, operating site, service station).  To cross-walk the normalized concentrations provided in Exhibit 7-C-1 with different types of release sites for the SNAP risk screens, EPA examined the model input parameters and patterns among the normalized concentrations themselves.  EPA believes that the types of release sites associated with each normalized concentration were identified accurately using this approach.

   	Second, and more importantly, the SNAP risk screens cover certain end uses that were not covered in the general population exposure portion of the interim analysis because the end uses were believed to consume relatively small amounts of ODSs.  Because the SNAP risk screens attempt to cover all end uses, normalized concentrations were needed for source types not previously modeled.  Professional judgment was used to select which normalized concentrations (among those listed in Exhibit 7-C-1) to use for source types that were not modeled earlier.  

   Exhibit 7-C-2 lists the normalized concentrations that EPA used for each source type in the SNAP analyses of general population exposure to ambient air releases.

	EXHIBIT 7-C-1
	NORMALIZED FENCELINE CONCENTRATIONS (EPA/VERSAR)
	SECTOR
	ACTIVITY
                                TYPE OF RELEASE
                   NORMALIZED MAX. FENCELINE CONC. (g/m[3])
                            CHEMICAL MANUFACTURING
Manufacture of CFCs
Stack
     8.9

Manufacture of CFCs
Area
    60.4
                                 REFRIGERATION
Retail Food: Manuf./Install.
Volume
   166.3

Retail Food: Servicing	
Volume
   864.8

CSW: Manuf./Install.
Volume
   102.5

CSW: Servicing
Volume
   102.5

Chillers: Manuf./Install.
Volume
   166.3

Chillers: Servicing
Volume
    40.7

Ice Makers: Manuf.
Volume
   166.3

Ice Makers: Servicing
Volume
   864.8

Ice Skate Rinks: Manuf.
Volume
   254.8

Ice Skate Rinks: Servicing
Volume
   254.8

MACs: Manuf.
Volume
    58.8

MACs: Servicing
Volume
   864.8

Chem. Proc. & Refineries: Manuf.
Area
    56.7

Chem. Proc. & Refin.: Servicing
Area
    56.7
FOAM BLOWING
Foam Blowing: Rigid & Flexible
Stack
     6.2

Foam Blowing: Rigid & Flexible
Volume
    62.4
SOLVENTS
PC Board Cleaning
Stack
     9.6[a]

PC Board Cleaning
Volume
   166.3

Other Electronics Cleaning
Stack
     9.6[a]

Other Electronics Cleaning
Volume
   472.7

Metal Cleaning
Stack
     9.6[a]

Metal Cleaning
Volume
   166.3
STERILIZATION
Sterilization
Volume
    40.7

[a] Normalized maximum concentrations for stack releases for the solvents sector were not provided by Versar.  These values were back-calculated based on the releases and fenceline concentrations provided in the earlier EPA reports.

                                 EXHIBIT 7-C-2
NORMALIZED MAXIMUM CONCENTRATIONS (g/m[3]) APPLIED TO DIFFERENT TYPES OF RELEASE SITES FOR SNAP RISK SCREENS
	SECTOR
	END USE
                               FACTORY RELEASES
                                OPERATING SITE 
                                   RELEASES
                           SERVICE STATION RELEASES
                           DISPOSAL SITE RELEASES[b]
CHEMICAL 
MANUFACTURE
Manufacture of CFCs
Stack: 8.9
Area: 60.4
   NA
   NA
   NA
REFRIGERATION
Retail Food
 166.3
 864.8
   NA
   NA

CSW
 166.3
 102.5
   NA
   NA

Chillers
 166.3
  40.7
   NA
   NA

Ice Makers
 166.3
 864.8
   NA
   NA

Ice Rinks
 166.3
 254.8
   NA
   NA

MACs
  58.8
   NA
  864.8
  56.7

ChProc & Ref
 166.3
  56.7
   NA
   NA

Transport Ref
 166.3
 864.8
   NA
   NA

Home 
Appliances
 166.3
 864.8
   NA
 254.8

Commercial A/C
 166.3
 254.8
   NA
   NA
FOAM BLOWING
Flexible Foam
   6.2
   NA
   NA
   NA

Other End Uses
  62
   NA
   NA
   NA
SOLVENT 
CLEANING
Cleaner 
Formulation
Stack: 8.9
Area: 60.4
   NA
   NA
   NA

PC Board Cleaning
  NA
Stack: 9.6
Volume: 166.3
   NA
   NA

Other Electron. Cleaning
  NA
Stack: 9.6
Volume: 472.7
   NA
   NA

Metal 
Cleaning
  NA
Stack: 9.6
Volume: 166.3
   NA
   NA
STERILIZATION
Sterilization
  NA
  40.7
   NA
   NA
AEROSOLS
Propellants
 166.3
   NA
   NA
   NA

Solvents
 166.3
  472.7
   NA
   NA
HALONS
Processing and Equipment
 173
   NA
   NA
   NA
ADHESIVES, COATINGS, AND INKS
All End Uses
 166.3
  166.3
   NA
   NA

[a]	Entries in bold were developed for the SNAP risk screens based on a review of EPA and Versar documents; those not in bold were taken directly from Versar (see Exhibit 7-C-1).
[b]	Applies to cases in which equipment is transported to a centralized location (e.g., recycling center or salvaging yard) before the ODS substitute is removed.


	REFERENCES FOR ATTACHMENT 7-C


U.S. EPA.  1990a.  Potential Ambient Inhalation Exposures for Chlorofluorocarbon Substitutes.  Office of Toxic Substances, Exposure Evaluation Division, Exposure Assessment Branch.  November 7, 1990.

U.S. EPA.  1990b.  External Review Draft, Hydrofluorocarbons and Hydrochlorofluorocarbons, Interim Report.  Office of Toxic Substances.  November, 1990.

U.S. EPA.  1990c.  Revised Draft: Occupational Exposure and Environmental Release Data for 	Chlorofluorocarbons (CFCs) and Their Substitutes.  Office of Toxic Substances.

U.S. EPA.  1986.  Industrial Source Complex (ISC) Dispersion Model User's Guide, Second Edition, 	Volumes 1 and 2.  Publication Nos. EPA-450/4-86-005a, and -005b.  U.S. Environmental Protection Agency, Research Triangle Park, NC. (NTIS PB86 234259 and PB86 234267).

U.S. EPA.  1987.  Industrial Source Complex (ISC) Dispersion Model.  Addendum to the User's Guide.  	U.S. Environmental Protection Agency, Research Triangle Park, NC.


	ATTACHMENT 7-D
	NOTE


   The disposal scenario for household and other refrigerated appliances included emissions from the refrigerators and freezers, dehumidifiers (part of the "other refrigerated appliances" category), and window air conditioners sectors.  Since all three kinds of equipment could conceivably be disposed of at the same salvage location, for the purposes of maintaining the conservative nature of this analysis, ratios should be summed for any substitutes that are common across these sectors.  The resulting ratios are:

Substitute		Ratio w/Recycling		Ratio w/o Recycling		Cancer Risk

HCFC-22		    9.7E-05			      N/A			  3.8E-08
HCFC-124		    2.2E-06			      N/A			    N/A
HFC-134a		    4.9E-04			     4.4E-02			    N/A
HFC-152a		    1.2E-04			     1.1E-02			    N/A

None of the resulting summed ratios are high enough to be considered of concern.

	ATTACHMENT 7-E

	Percent of Equipment with Assumed Charge Size

	END USE
	Charge Size
	(Kgs)
	Percent of
	Total in End
	Use Category
	Comments
Retail Food - Parallel
                                      466
                                      100

Retail Food - Stand alone
                                      19
                                      50
other 50% are 10 kg
Cold Storage Warehouses
                                      250
                                      100

Reciprocating Chillers
                                      73
                                      100

Screw Chillers
                                      230
                                      100

Centrifugal Chillers - LP
                                      450
                                      100

Centrifugal Chillers - HP
                                      900
                                      70
25% are 820 kg,
5% are 540 kg
Industrial Process - LP
                                      680
                                      100

Industrial Process
High Pressure Package
                                      180
                                      89
11% are 145 kg
Industrial Process
High Pressure Built Up
                                     9100
                                       7
58% are 4550 kg
6% are 2200 kg
29% are 1800 kg
Ice Skating Rinks
                                     1360
                                      52
48% are approx. 350 kg
Medium Ice Makers
                                      2.3
                                      100

Large Ice Makers
                                     34.1
                                      100

Transport Refrigeration
                                      8.1
                                      100

Window A/C
                                      1.8
                                      100

Central A/C and Heat Pumps
                                      3.8
                                      20
75% are approx. 3 kg
5% are 0.7 kg
Commercial Direct A/C
                                      33
                                      10
90% are 10.2 kg
Refrigerators and Freezers
                                      0.4
                                       8
92% are between 0.19 kg and 0.33 kg
Other Refrigerated Appliances
                                      0.5
                                      10
42% are 0.4 kg
14% are 0.2 kg
34% are 0.1 kg
Mobile Air Conditioners
                                      1.2
                                      40
60% ar 0.88 kg

	8.  VOLATILE ORGANIC COMPOUND ANALYSIS


   This chapter examines volatile organic compound (VOC) emissions from substitutes to Class I ozone-depleting substances (ODSs) that may be used for some end uses within the refrigeration and air conditioning sector. The VOCs of concern for this sector are the hydrocarbons.  VOCs are important to investigate as they can lead to the formation of ground-level ozone and, thus, degradation of ambient air quality and adverse health effects.

   This is a screening-level assessment where overly conservative, worst-case scenarios are presented.  The general approach used in this analysis was to compare total nationwide potential emissions from sources using VOC substitutes in the refrigeration and air conditioning industry with total VOCs from all other sources (e.g., other industries, mobile sources, and biogenic sources).  Next, a hypothetical sample area (the Trane facility in the city of Indianapolis) was chosen to demonstrate a worst-case air emission situation.  The estimated potential emissions from a facility that had switched to VOCs were added to a per-capita emissions amount (to account for releases from refrigeration and air conditioning units already in operation) and compared with the total VOC emissions from all other sources. 

   These comparisons give an indication of the potential impact of Class I substitutes in this sector on VOC emissions and the resulting formation of ground-level ozone.

8.1	VOC EMISSIONS FROM SUBSTITUTES TO CLASS I ODSs

   To estimate VOC releases from ODS substitutes in the refrigerant and air conditioning industry, the estimated 1992 nationwide release of ODSs were examined in all refrigerant and air conditioning end uses for which hydrocarbons have been suggested as substitutes.  The year 1992 was chosen based on the assumption that the effects of the CAA phaseout mandate would not begin until after 1991 or 1992.

   The release estimates are from the Vintaging Model.w  For the purposes of this screening-level analysis, it was assumed that hydrocarbons would capture 100 percent of the market share for substitutes to Class I ODSs in 1992, and that the replacement factor for substituting a Class I substance with a VOC alternative is one.  EPA believes that these assumptions grossly overestimate the amount of VOCs that could potentially be released, which is consistent with the objective of conducting a screening-level assessment.

8.2	VOC EMISSIONS FROM ALL SOURCES

   The VOCs from both anthropogenic and biogenic sources are considered for purposes of this analysis as emissions of VOCs from all sources (i.e., background).  Attachment 8-A provides details on the calculation of total (background) VOCs from all sources.  Refer to Attachment 8-A for the breakdown of total emissions for the nation, for each of the six areas as taken from the Urban Airshed Model (UAM), and for a description of the UAM.



8.3	COMPARISON OF VOCS FROM SUBSTITUTES TO VOCS FROM ALL SOURCES

   The comparison between total nationwide VOC emissions from all sources (21,070,000 MT) and VOC emissions from ODS substitutes in the refrigeration and air conditioning industry in 1992 (67,295 MT) results in a national increase in total VOCs of no more than about 0.32 percent.  


   EPA does not expect to see any significant difference between the national ratio and the regional ratios, and therefore has not estimated VOC increases in detail for each region in  Attachment 8-A (as was done for several other sectors).  However, releases from refrigerant and air conditioning use in a concentrated area were roughly evaluated by placing several of the "worst" possible hypothetical factories in that area, assessing emissions, adding this to a per-capita emissions amount (to account for releases from units in operation), and comparing the potential VOC emissions to the total VOC emissions from all sources in that area.

   In the sample region, Trane was chosen as the factory with the highest estimated potential air emissions (because of its size).  It was assumed that Trane manufactures all industrial process equipment in one factory.  (Industrial process equipment factories account for some of the highest VOC releases of any equipment-producing factories; see the general population analysis for more details.)  From the analysis on general population risk for the refrigeration and air conditioning sectors, VOC releases from three of these facilities were estimated to be 74 MT/yr.  Next, the per-capita emissions were added in to account for non-factory sources.  To do this, the 67,295 MT/yr emissions estimated from the vintaging model were used and it was assumed that there are 250 million people in the United States.  Indianapolisa has a population of 780,000 people, yielding a total annual release of 210 MT/yr for non-factory sources.  Total VOCs from hydrocarbon substitutes in the refrigerant and air conditioning sector in Indianapolis thus is estimated at 284 MT/yr.  This increase is 0.43 percent of the total VOCs from all sources in Indianapolis (66,640 tons).  Using data from Systems Applications International (SAI),b and the assumption that one month of VOC emissions (23.67 MT) occurs instantaneously and enters a 3,000 foot high column of air over Indianapolis (which has an area of approximately 10[12] m[2]), ozone levels would increase by about 12.8 ppt.  When added to Indianapolis' current ozone level of 0.121 ppm, the increase would be insignificant, especially compared to the amount needed to change Indianapolis' ozone classification to the next worse nonattainment level (i.e., moderate at 0.138 ppm ozone). 

8.4	CAVEATS AND LIMITATIONS

   EPA believes that actual VOC increases due to refrigerants will be less than presented here for several reasons, including:

   	VOC emissions controls required by EPA and the states will likely be used (Attachment 8-B describes the expected impact of VOC controls);

   	Improved waste minimization and other housekeeping techniques will likely be used;

   	VOCs will not have 100 percent of the market share.
   
	CHAPTER 8 REFERENCES


Reid, Steven.  Systems Applications International (SAI).  1990.  May-June.  San Rafael, CA.  Personal communication.

	ATTACHMENT 8-A
	ESTIMATION OF TOTAL VOC EMISSIONS FROM ALL SOURCES IN 1995


   The estimates for total VOC emissions from all sources incorporate both anthropogenic (including mobile sources) and biogenic sources.  For Los Angeles, Philadelphia, Atlanta, and Dallas, estimates of VOC emissions for 1995 were extracted directly from the latest Urban Airshed Modela results (June 1990).  For Indianapolis and Hennepin, VOC emissions estimates for 1995 were not available (nor were estimates for any other marginal nonattainment or attainment area).  Similarly, anthropogenic and biogenic estimates were not readily available for nationwide 1995 VOC emissions.  Therefore, anthropogenic emissions for 1985 were first obtained from the National Acid Precipitation Assessment Program (NAPAP) emissions inventory.  These 1985 figures then were adjusted for 1995 using economic indicator data from the Bureau of Economic Analysis (BEA) Regional Projections Data System (June 1990).  The biogenic emissions were obtained by averaging seasonal biogenic VOC data from the Regional Biogenic Emissions Inventory System (1990)b.  These biogenic emissions were assumed to remain the same for 1995.  The figures obtained from this analysis are presented in Exhibit 8-A-1.


	EXHIBIT 8-A-1
	TOTAL VOC EMISSIONS FROM ALL SOURCES IN 1995 (TONS)


	Area
	Current
	Status
	Anthropogenic
	Emissions
	Biogenic
	Emissions
	Total VOC
	Emissions
Los Angeles, CA
Extreme
Nonattainment
                                    762,485
                                    295,285
                                   1,057,770
Philadelphia, PA
Severe
Nonattainment
                                    760,295
                                    230,315
                                    990,610
Atlanta, GA
Serious
Nonattainment
                                    244,550
                                    466,105
                                    710,655
Dallas, TX
Moderate
Nonattainment
                                    360,620
                                    253,310
                                    613,930
Indianapolis, IN
Marginal
Nonattainment
                                    71,588
                                     1,718
                                    73,306
Hennepin, MN
Attainment
                                    82,022
                                     2,078
                                    84,100
United States
	--
                                      --
                                      --
                                  23,176,700

	ATTACHMENT 8-B
	VOC CONTROLS


   Facilities in ozone nonattainment areas switching from ODSs to VOCs may be required to offset the VOC emissions from their facility so that net VOC emissions decrease within the nonattainment area.  Within nonattainment areas, major sources of emissions are subject to the provisions of the New Source Review (NSR) program.  This program requires affected sources to reduce emissions to the lowest achievable emissions rate (LAER), and to offset new VOC emissions increases with reductions obtained within the facility or at other facilities located within the nonattainment area.

   The NSR program affects "major sources" in nonattainment areas, and the  definition of major source depends primarily upon the potential of the source to emit pollutants (potential emissions).  Potential emissions are the amount of pollutant the source is capable of emitting in the absence of any federally enforceable emission limitation.  These amounts are similar to those estimated for individual facilities in the SNAP analyses.  The potential emissions necessary for a source to be defined as major is a function of the classification of the ozone nonattainment area, with the amount of emissions necessary to be defined as a major source becoming smaller as ozone nonattainment problems of the area become greater (Exhibit 8-B-1). 


	EXHIBIT 8-B-1
	MAJOR SOURCE DEFINITION AND OFFSET REQUIREMENTS

                      Nonattainment Area Classification 
	Major Source Definition
	(Tons per Year)
	Offset Ratios
	(VOC reduct.:VOC incr.)
	Marginal
	100 TPY 
	1.1 : 1
	Moderate
	100 TPY[a]
	1.15 : 1
	Serious
	 50 TPY[a]
	1.2 : 1
	Severe
	 25 TPY[a]
	1.3 : 1
	Extreme
	 10 TPY[a]
	1.5 : 1

[a] States have the option of electing to define major source at 5 tons/yr (and accepting other conditions) as opposed to demonstrating further reasonable progress.  See text.


   It should be noted that some states may elect to opt out of the reasonable further progress (RFP) demonstration required by the CAA, which requires states with some classifications of nonattainment areas to demonstrate a 15 percent reduction in VOC emissions within the first six years of enactment (i.e., by November 1996).  States electing to opt out of the RFP demonstration must define major source at the 5 ton/yr level and accept other conditions.  

   The CAA also requires that major sources in nonattainment areas any offset any new VOC emissions with reductions in VOC emissions from within their facility or from other facilities located within the nonattainment area.  This offset requirement results in a net decrease in VOC emissions for sources affected by the NSR program.  As shown in Exhibit 8-B-1, the required offsets are a function of the classification of the nonattainment area and increase with increasing ozone nonattainment status.

   Sources affected by the NSR program are also required to reduce emissions to LAER.  LAER determinations are typically based on the minimum expected performance of air pollution control devices.  The reductions achievable by control devices are well established and as shown in Exhibit 8-B-2 exceed 95 percent.


	EXHIBIT 8-B-2
	EFFECTIVENESS OF COMMON VOC CONTROL DEVICES

	Control Device
	Minimum Expected Reduction 
                             Efficiency (Percent)
	Thermal Oxidizer
	98%
	Carbon Absorber
	95%
	Condenser
	95%


   Because facilities switching to VOCs would be required to install LAER, and controls are capable of reducing emissions by 95 percent, EPA would expect the projected increases in VOC from NSR affected facilities to be reduced by at least 95 percent.  These facilities must offset any residual emissions (i.e., the remaining 5 percent) by the amount specified in Exhibit 8-B-2, thus mitigating any impacts from switching from ozone depleting substances to VOCs.


