WPCo  23 ZBg0[FAXPRESS PCLFPPCLXx P7XP"T gp@6pSmall Circledg!=(!*!ßT gp.E"#Xx P7XP#X0Í ÍX0Í ÍҫXx P7XP* `(CG Times (Scalable)XXx P7XP* `(CG Times (Scalable)XXx P7XP* `(CG Times (Scalable)Xm*0 x7* `(CG TimesScalableXx P7XP* `(CG Times (Scalable)Xx*0 x7* `(CG TimesScalableXx P7XP* `(CG Times (Scalable)Xx*0 x7* `(CG TimesScalableA\  PP(x-  Z 6Times New Roman Regularx*0 x 7* `(CG TimesScalableXx P 7XP* `(CG Times (Scalable)X2e m |S      = Gary Lindgren OZ Technology N. 10900 Howell Road Rathdrum, Idaho 83858 Registered Mail; receipt requested Dear Mr. Lindgren:  Thank you for your petition, dated December 5, 1995, requesting that 1) EPA change the status of HC12a under the SNAP program from unacceptable to acceptable, and 2) EPA change the status of HFC134a under the SNAP program from acceptable to unacceptable. In reviewing this petition, EPA considered the documents to which it responds in this petition, as well as all information previously submitted regarding HC12a, OZ12 and HFC134a.  For the reasons stated in the enclosed attachment, EPA denies this petition. In summary, none of the documents submitted as part of this petition adequately addresses the use of HC12a as a CFC12 substitute. While some documents answer a limited number of questions, they all neglect one or more of the following aspects of flammability risk: 1. Risk assessment is a necessary tool  The key measure of risk posed by flammable refrigerants is a scientifically valid, comprehensive risk assessment. Such an assessment must accurately reflect potential leak scenarios, potential ignition sources, the likelihood of ignition, the consequences of ignition or explosion, and potential measures to mitigate the risk. None of the documents submitted with this or earlier petitions represents such a risk assessment. Without such an assessment, no positive statements can be made about the actual risk flammable refrigerants pose to people using them. Subjective measures, experiments that do not reflect actual systems, and questionable science do not allow for a credible estimate of risk posed by the flammability of HC12a. 2. Flammability risk differs with new system design versus existing systems  New systems can be designed to adequately address the risk of using flammable refrigerants. Existing systems were never intended to use such refrigerants, and therefore do not specifically protect consumers from that flammability. EPA requires a risk assessment for all use of flammable refrigerants, but their use in existing, unmodified systems requires special vigilance. No refrigeration systems (outside a limited number of industrial refrigeration systems), and no air conditioning applications, widely use flammable refrigerants in the US. Therefore, system owners (including car owners), service and disposal personnel, and fire fighting personnel do not expect to   find flammable refrigerants in these systems. EPA must ensure that such use is safe, including the widespread dissemination of safe handling practices, before listing HC12a or any flammable refrigerant as acceptable. Risk assessments for existing uses must reflect conditions found in the US today, rather than systems used in Europe or elsewhere in the world. For example, US refrigerators are much larger than the hydrocarbonbased systems found in Europe. A risk assessment for this enduse must address the larger system. 3. Flammability risk depends on the enduse and the refrigerant  It is inappropriate to extrapolate from one enduse to another, or from one refrigerant to another. Enduses differ greatly in operating conditions. For example, automobiles are driven at high speeds and frequently collide with each other and other objects. The condenser on automobiles is immediately behind the grille, where it is highly susceptible to puncture during a frontend collision. Refrigerators, in contrast, operate in an environment that features numerous internal and external ignition sources, including door light switches, automatic defrost heaters, potentially faulty wiring, gas ranges, and cigarette lighters.  In addition, flammable refrigerants differ substantially. As explained in detail in the attachment, HFC152a and propane are both flammable, but it takes more HFC152a to create a flammable concentration and the energy released from burning HFC152a is significantly less than that released from burning propane. These characteristics mean that the risks from using HFC152a are likely to be much lower than those from using HC12a.  Thus, the 1991 risk assessment on HFC152a in refrigerators is not relevant to the use of HC12a in refrigerators because of the second reason, and it is not relevant to the use of HC12a in automobiles for both reasons. 4. Flammability risk depends on the charge size  Large cars can contain twice as much refrigerant as small cars. Similarly, large US refrigerators contain much more refrigerant than small European models. A larger refrigerant charge means that a higher concentration in air is created by a given leak, and the leak will last longer. In addition, for a given volume, like the inside of a car, it is possible that a small charge would not create a flammable mixture, but a typical or larger charge could. A risk assessment must address these differences. 5. Flammability risk depends on system specifics  Even within a specific enduse, differences in design can yield differences in risk. For example, the smaller size of European refrigerators means not only that the refrigerant charge is smaller, but also that the physical size of the system is smaller, which could have an effect on where it is placed in a kitchen. Also, US refrigerators generally have automatic defrost heaters (a potential ignition source), whereas many European models do not. In cars, certain models may provide for continuous fresh air flowing through the passenger compartment, whereas others may not. A secondary loop can greatly reduce flammability risk in a supermarket refrigeration system, even though the system still falls into the retail food refrigeration enduse. Such design differences can have large impacts on overall flammability risk. Therefore, risk assessments must reflect system specifics. 6. Efficiency is not safety  EPA is aware that hydrocarbons offer potential energy efficiency gains over the use of other refrigerants. However, EPA's primary concern is the flammability of HC12a. Papers that only discuss the system performance characteristics of HC12a do not address this safety issue at all.  If you have any questions about this response or EPA's determinations, please contact Jeffrey Levy, the Refrigerants Analyst for the Significant New Alternatives Policy Program. Mr. Levy may be reached at (202) 2339727.  pp  xx!Sincerely,  pp  xx!&((+0  pp  xx!Mary D. Nichols  pp  xx!Assistant Administrator  pp  xx! for Air and Radiation enclosure  XX  XX   ӊ=$ ATTACHMENT ă  This Attachment responds to several submissions made by OZ Technology. Section I is a response to the documents submitted with OZ Technology's December 5, 1995 "Petition Pursuant to Section 612(d) of the 1990 Amendments to the Clean Air Act to List HC12a as an Acceptable Substitutes and to List HFC134a as an Acceptable Substitute" (December Petition). Section II is a response to four additional documents that OZ Technology has submitted and to which EPA had not previously responded. These four documents are: (1) "Supplemental Risk Assessment Regarding the Use of HC12a Refrigerant in Automotive Air Conditioners," submitted by Charles Lempesis on June 5, 1995; (2) "HC12a Supplemental Risk Assessment for Home Refrigerators," submitted by OZ Technology on May 31, 1995; (3) Videotape, "Hydrocarbon Refrigerants/Supervised Safety Tests," submitted by Bob Small, Executive Director, OZ Technology on November 25, 1995; and (4) "A CHEMICAL DISASTER, Why HFCS Have No Future," Greenpeace, submitted by Gary M. Lindgren, CEO, OZ Technology on July 16, 1996.  For the purposes of this response, any discussion of the HC12a determination should be taken in the context that, under the March 18, 1994 SNAP rule, EPA does not review substitutes for nonozonedepleting refrigerants like HFC134a, that the determination was only made within the refrigeration and air conditioning sector, and that the unacceptability determination did not include industrial process refrigeration. However, in the interest of clarity and brevity in this response, HC12a may simply be referred to as "unacceptable" without repeatedly stating the qualifications mentioned above. The term "unacceptable" should be taken to mean "HC12a is unacceptable as a CFC12 substitute for all refrigeration and air conditioning enduses other than industrial process refrigeration." Similarly, statements that HFC134a is an acceptable substitute should be taken to mean "HFC134a is an acceptable substitute for CFC12 in refrigeration and air conditioning enduses." Since the SNAP rule does not regulate the legitimate substitution of HC12a for first generation nonozonedepleting substances, the terms "acceptable" and "unacceptable" have no bearing on such use of HC12a. SECTION I RESPONSE TO DECEMBER 1995 PETITION Paragraphs 1-5  These paragraphs simply describe HC-12a and the purpose of the petition, and therefore require no response. Paragraphs 67  EPA disagrees that HC12a has been marketed "exclusively as a second generation substitute in refrigeration and air conditioning systems." Although HC12a has been marketed as a substitute for HFC134a, HC12a has also been marketed as a CFC12 substitute for nearly two years, as evidenced by numerous telephone calls and letters. In addition, OZ Technology distributes a brochure titled "HC12a: Natural Organic Refrigerant" that includes references to CFC12. Specifically, in the original brochure's stepbystep guide for charging a system with HC12a, step one referred to removing CFC12 from the system. In addition, the current version lists CFC12 properties in a table titled "Technical Summary of Refrigerant Properties," and a graph labeled "HC12a: Vapor Pressure vs. Temperature" includes a curve for CFC12. On the basis of this evidence, EPA concluded that HC12a was being marketed as a CFC12 substitute and notified OZ Technology on July 31, 1994 that it was required to submit information on HC12a for review under EPA's Significant New Alternatives Policy (SNAP) Program.  It is true that on September 26, 1994 (59 FR 49108), EPA proposed listing HC12a as an unacceptable substitute for CFC12 in all refrigeration and air conditioning enduses other than industrial process refrigeration. On June 13, 1995 (60 FR 31092), EPA promulgated a final rule listing HC12a as unacceptable. The petition also omits the fact that HFC134a was found to be an acceptable substitute for CFC12 in all enduses in the original SNAP rule, published on March 18, 1994 (59 FR 13044). Paragraph 8  EPA agrees that HC12a is non-ozone-depleting and would characterize it as low in toxicity. It consists of hydrocarbons, so it also has a relatively low global warming potential. EPA disagrees, however, with the statement that it has "a limited potential for flammability." There is no reason to expect that HC12a's flammability characteristics differ from its components, including propane, butane, or other hydrocarbons. All of these substances have been designated as "flammable" by Underwriters Laboratories (UL) and the American Society of Heating, Refrigerating, and Air Conditioning Engineers (ASHRAE). In addition, the U.S. Department of Transportation's (DOT) Office of Defects Investigation conducted an initial study of the flammability risk associated with using OZ-12, a similar hydrocarbon blend manufactured by OZ Technology, in unmodified, existing automobiles. The result of this study was a December 30, 1993 letter from DOT to OZ Technology that stated: "ODI believes there is an unacceptable fire risk associated with the use of OZ12 in a conventional motor vehicle. Accordingly, we request that you initiate a safety recall concerning this matter." Again, no evidence suggests that HC12a is less flammable than OZ12, or that using HC12a in a system would pose less risk than using OZ12. The rest of this response will address the statement that HC12a is "now entitled to a designation as an acceptable first generation substitute." Paragraph 9  This paragraph accurately reflects EPA's basis for determining that HC12a is an unacceptable substitute that OZ Technology has not submitted a scientifically valid risk assessment. EPA has made statements regarding the potential repercussions for failing to comply with the SNAP rule; these statements reflect the enforcement provisions in section 113 of the Clean Air Act (CAA). In accordance with section 113, a violation of EPA's SNAP rule is potentially subject to civil or criminal enforcement proceedings. Specifically, section 113(a)(3) of the CAA provides for administrative, civil or criminal enforcement for, among other things, violation of Title VI or rules promulgated pursuant to that title. Under the civil judicial enforcement provisions of section 113(b), EPA may, among other things, seek to recover a "civil penalty of not more than $25,000 per day for each violation." Under the criminal enforcement provisions of section 113(c), in certain circumstances, a person who "knowingly violates any requirement or prohibition" shall be punished by a fine, imprisonment or both upon conviction. Paragraph 10  EPA maintains that it is incumbent upon OZ Technology to demonstrate, for each proposed end-use, that any submitted refrigerant may be used safely. The primary concern for HC12a is risk to human health posed by the use of a flammable refrigerant in systems not designed to address the risk of flammability.  This approach is based on section 612(c) of the Clean Air Act, which states "Within two years after enactment of the Clean Air Act Amendments of 1990, the Administrator shall promulgate rules under this section providing that it shall be unlawful to replace any class I or class II substance with any substitute substance which the Administrator determines may present adverse effects to human health or the environment, where the Administrator has identified an alternative to such replacement that- (1) reduces the overall risk to human health and the environment; and  (2) is currently or potentially available. The Administrator shall publish a list of (A) the substitutes prohibited under this subsection for specific uses and (B) the safe alternatives identified under this subsection for specific uses."  To list a given substitute as unacceptable, EPA must determine that a substitute "may present adverse effects to human health or the environment." (Emphasis added.) It is incumbent upon anyone who disagrees with the basis for such a finding to produce evidence controverting such a determination. Given the flammability of HC12a and the DOT conclusions, it is clear that a potential risk exists. Note that this potential does not presuppose the level of the actual risk. It is sufficient reason, however, to require a risk assessment.  As stated in the SNAP rule: "If a substitute is flammable, the submitter must analyze the risk of fire resulting from the use of such a substitute and assess the effectiveness of measures to minimize such risk." 40 CFR 82.178(a)(9). This statement requires that a valid, comprehensive risk assessment be conducted for each proposed enduse of a flammable substitute. This risk assessment must analyze the potential risk to consumers, service technicians, and personnel in both manufacturing and disposal facilities.  EPA disagrees with the assertion that HFC134a has not been subject to the same standards as HC12a. In response to federal requirements and producers' concerns for product liability, HFC134a has undergone extremely rigorous testing, precisely because it was to be used so widely. Through toxicological studies required by EPA under the Clean Air Act and the Toxic Substances Control Act, HFC134a has been determined to be less toxic than CFC12, which was used safely for 60 years. In addition, ASHRAE found HFC134a to be nonflammable and to be low in toxicity. In fact, ASHRAE issued an "A1" safety rating for HFC134a. This rating is reserved for nonflammable refrigerants that are low in toxicity. UL also found HFC134a to be "practically nonflammable." This designation means that HFC134a cannot be made to ignite under standard atmospheric pressure, but it does have an autoignition temperature below 750 degrees Celsius.  As stated above, a risk assessment addresses the incremental risk posed by using a flammable refrigerant in a system not designed to use a flammable refrigerant. HFC134a is not flammable, and, therefore, its use does not require such an assessment. Paragraph 11  This paragraph refers to several additional documents, each of which is responded to below.  A.Exhibit A: Relative Risks of HC12a and HFC134a 1.` ` Flammability Definition and Quantification  The summary begins with a description of terms used by organizations such as ASHRAE, UL, and the Society of Automotive Engineers (SAE). However, rather than being "generic terms," the terms "flammable," "practically flammable," and "nonflammable" have precise meanings as defined by the relevant bodies.  All definitions of flammability depend on the concept of flammability limits. The concentration in air below which the mixture will not ignite is the lower flammability limit, while the concentration in air above which the mixture will not ignite is the upper flammability limit. For example, propane's lower flammability limit is 2.1% and its upper flammability limit is 9.6%.  The discussion makes the incorrect claim that under UL's definition, HC12a would be considered nonflammable. The UL definition of flammability relies on two factors: flammability limits in air and autoignition temperature. If a refrigerant exhibits flammability limits in air at normal atmospheric pressure, it is deemed "flammable." If there is no concentration in air at which the refrigerant becomes flammable, UL then examines the autoignition temperature, which is the temperature at which a sample will spontaneously ignite with no external ignition sources. Note that the entire sample is heated to this temperature. If the autoignition temperature is above 750 degrees Celsius, the refrigerant is deemed "nonflammable." If the autoignition temperature is below 750 degrees Celsius, the refrigerant is deemed "practically nonflammable."  Although HC12a has not been tested by UL or in accordance with UL procedures, the claim in this paragraph ignores a key part of the definition of "nonflammable," which is that nonflammable refrigerants do not exhibit flammability limits in air. If a refrigerant exhibits such limits, it cannot be considered "nonflammable." Furthermore, "practically nonflammable" refrigerants also do not exhibit flammability limits in air, but have an autoignition temperature below 750 degrees C. This is the category that includes HFC134a and HCFC22. It should be noted that a) HFC134a's autoignition temperature is 743 degrees C, and b) most of the substitutes listed as acceptable are classified as practically nonflammable. HC12a's components exhibit flammability limits in air, and HC12a would most likely be classified as "flammable" by UL if it were tested. OZ Technology appears to recognize this fact, given that the sales brochure and the label both state that HC12a is flammable.  Both the UL testing method and ASTM E681 (relied upon by both SAE and ASHRAE) specify that refrigerants be tested at atmospheric pressure. This is not an accident, but related to engineering experience and judgment. While it is true that refrigerants are pressurized within containers and refrigeration and air conditioning systems, they are not mixed with air in large quantities in those situations.  2.` ` Product of Combustion  The Southwestern Laboratories report demonstrated that even a 100% HFC134a atmosphere produced only 10 ppm of hydrogen fluoride (HF) (a thermal decomposition product of HFC134a) when passed through tubing heated to approximately 1100 degrees F (assumed to be the temperature of a burning cigarette). However, no car passenger would ever be exposed to a pure HFC134a environment. In fact, even if the entire charge of a car air conditioning system were released into the passenger compartment, the concentration of HFC134a would not rise above approximately 10% (90% air), which proportionally would yield an HF level of only 1 ppm. Contrary to the assertion that this is dangerous, 1 ppm is only 3% of the Immediately Dangerous to Life and Health (IDLH) limit of 30 ppm, recommended by the US Institute for Occupational Safety and Health as a maximum limit from which one could escape within 30 minutes, without a respirator and without experiencing any escapeimpairing or irreversible health effects. In addition, note that no car is truly airtight; therefore the concentration of HF would decrease over time. Finally, 1 ppm is only one third of even the longterm exposure threshold limit value (TLV) of 3 ppm. Keep in mind that 3 ppm of HF is not an instantaneous danger, but is the level at which continuous exposure for 8 hours/day over an indefinite period will not result in a health risk. Thus, although the study did show measurable quantities of HF, the quantified risk from inhalation through a cigarette is too small to be of concern.  The material safety data sheet (MSDS) for HFC134a is similar to that used for many refrigerants. In general, it is certainly recommended to avoid exposing refrigerants to open flames and excessively high temperatures. Smoke inhalation is dangerous, regardless of the extent of HF contained in the decomposition products. However, the fact that these recommendations are appropriate does not render HFC134a unacceptably hazardous. It simply means that HFC134a requires the same handling as CFC12 and all other refrigerants. It is true that the use of HFC134a and many other alternative refrigerants in refrigeration and air conditioning systems requires the use of synthetic lubricating oils. These oils may be flammable, but they do not differ in this regard from the traditional mineral oils used with CFC12. Lubricating oils are generally flammable, so the incremental risk from the synthetic oils is negligible. A final point is that all oils are liquids at room temperature, and leaking oil will simply pool on the floor of the vehicle. HC12a, on the other hand, is a gas at room temperature, and will mix with the air in the automobile, potentially forming an explosive mixture.  EPA is currently gathering information about the characteristics of these lubricating oils. Although it must be noted that EPA lacks authority under section 612 to regulate the new lubricants, an informal evaluation of their safety similar to that for refrigerants under SNAP would follow guidelines such as comparing new lubricants with existing products.  3.` ` Velocity of Combustion  OZ relies on an article from an unspecified newspaper to support its claim that HFC134a is flammable. However, an article, standing alone, cannot controvert the numerous tests that have demonstrated that HFC134a is nonflammable at atmospheric pressures in accordance with ASHRAE standards. In addition, it is unclear what is meant by the claim that the velocity of combustion is "very high." A more relevant issue would be how the implied velocity of combustion for HFC134a compares to that actually measured for flammable hydrocarbons such as HC12a. See the response to item f. of Additional Supporting Documentation, below, for a more detailed discussion of this article.  4.` ` Global Warming Potential  It is true that HFC134a has a 100year Global Warming Potential (GWP) of 1300, while that for most hydrocarbons is small. However, the GWP of several other alternatives is similar to that of HFC134a. While GWP is one factor to consider, EPA must balance the risks associated with available refrigerants. At this time, EPA believes the potential flammability risk posed by using HC12a in unmodified systems not designed to use a flammable refrigerant to be a more significant risk than the unknown effects of global warming that will be spread over large areas and lengthy amount of time.  5.` ` Costs  The discussion on costs misrepresents the relevant considerations for SNAP review. EPA does not believe it is appropriate to prohibit a substitute's use because of cost. The market will decide whether costly substitutes have other merits that justify additional expense. Rather, cost data are relevant only when a new substitute is so much better than an existing one that the latter would more appropriately be unacceptable; in that case, EPA will consider whether the new substitute is too costly to justify moving the already acceptable substitute to the unacceptable list.  All retrofitting entails cost. However, note that the vast majority of centrifugal chillers, for which a cost estimate is given, use CFC11, not CFC12. OZ Technology only submitted HC12a for use as a CFC12 substitute, and, therefore, this cost estimate is irrelevant. However, given that the discussion mentions this CFC11based enduse, the cost data appear to be inaccurate since building codes would likely require costly improvements to use flammable refrigerants, if in fact their use would be allowed at all. Under ASHRAE Standard 15, which is widely referenced in state laws and building codes, severe restrictions are placed on the use of flammable refrigerants in large air conditioning systems.   6.` ` Additional Supporting Documentation to Exhibit A Č A response for each document follows. a. Refrigeration News Notes  This UL information release, dated October 1995, includes UL's classification of HFC134a as practically nonflammable. Note that the majority of new refrigerants are also listed as practically nonflammable. As described above, "practically nonflammable" means that a refrigerant does not exhibit flammability limits in air, but does have an autoignition temperature below 750 degrees C. b. HC12a Autoignition Temperature Test Results  This document, dated November 21, 1995, states that the autoignition temperature of HC12a is 1627 degrees F. The document appears to be a summary prepared by OZ Technology, not a report from an independent laboratory. EPA is unaware of OZ Technology's qualifications to perform laboratory testing in accordance with accepted standards. Therefore, OZ's report must be supported by independent laboratory testing. Aside from the quality of the data, however, is the relevance of autoignition temperature to flammability classification.  As discussed above, both ASHRAE and UL determine whether a substance is flammable by using the ASTM E681 testing procedure to determine flammability limits in air at normal atmospheric pressure. If the refrigerant does not exhibit flammability limits when tested according to ASTM E681, then UL uses the autoignition temperature to determine whether it is "nonflammable" or "practically nonflammable." However, as discussed above, no evidence indicates that, given HC12a's flammable hydrocarbon components, it would be found "nonflammable" or "practically nonflammable" by either ASHRAE or UL. Finally, as mentioned above, both OZ Technology's sales brochure and label describe HC12a as flammable. A more thorough discussion of flammability categories is included in the response to paragraph 16 below. For purposes of SNAP review, OZ Technology should submit the original report and a contact number for the laboratory that performed the testing and the report or OZ should identify the standard followed while testing, since testing standards may vary. c. SUVA HP Refrigerants  It should be noted that this onepage document was clearly removed from an unidentified, larger document; therefore, it is impossible to determine the context of the information it presents. This undated document specifies several handling practices for refrigerants, including HFC134a. These practices apply to most refrigerants, and therefore do not demonstrate that HFC134a is especially hazardous. For instance, any refrigerant will displace oxygen and pose a health hazard in large enough concentrations, including hydrocarbons. The combustibility of HFC134a at elevated temperatures and pressures is discussed elsewhere in this response. d.  CFCFree Freon Samples for Potentially Hazardous Recombination Products  These reports, dated March 11, 1994 and May 12, 1994, state that a pure HFC134a sample produced 10 ppm of HF. This result is discussed in depth above, under number 3, Product of Combustion. e. MSDS  This material safety data sheet, dated May 21, 1992, discusses R502 (a mixture of CFC115 and HCFC22), not HFC134a. It is unclear how this MSDS for R502 was intended to support this petition. As EPA has maintained, however, many of the handling practices are similar to those recommended for HFC134a, again demonstrating that they are typical for all fluorocarbon refrigerants. Given that R502 has been used safely for decades, recommending these handling practices does not indicate an unusual hazard. f. Hoechst Chlorofluorocarbon Plant Explosion: Cause Unknown (several German articles and one English article)  The undated English article discusses an explosion in a Frankfurt, Germany HFC134a plant. (Note: from the German articles, it appears that the explosion took place on March 18, 1994.) Two possible causes are mentioned: excessive pressure in a container, and excessive mixing of air into a pressurized container that yielded a combustible mixture. The first possibility has no bearing on the SNAP determination for HFC134a, since exceeding pressure limits in any container with any product is a dangerous practice. In addition, automobile air conditioning systems generally have a safety valve that prevents pressures from exceeding a preset value that is significantly below the structural capabilities of system components. Therefore, it is unlikely that an explosive rupture would occur. In addition, note that all air conditioning refrigerants, including CFC12, HFC134a, and HC12a, function at pressures higher than atmospheric pressure; therefore, ruptures do not represent a hazard particular to HFC134a. Finally, improper or faulty containers pose a hazard regardless of the particular pressurized gas they contain, so this cause of the Hoechst explosion would not demonstrate a hazard specific to the use of HFC134a.  The second possibility relates to the potential combustibility of HFC134a under pressures above atmospheric pressure and in the presence of both large volumes of air and an ignition source. This explanation bears little relevance to the safety of using HFC134a in a system normally pressurized above atmospheric pressure that does not contain air, such as a car air conditioning system. In addition to the lack of air, there would be no ignition sources inside an air conditioning system. Once released by a leak into the passenger compartment, HFC134a is at atmospheric pressure, under which it exhibits no flammability limits and is, therefore, nonflammable. g. Ozone Depletion Today  This document, dated Nov. 20, 1995, is a daily compendium of ozonerelated news items. It contains an article that quotes Trane Company officials estimating the cost of retrofitting centrifugal chillers in the US. As stated above, the costs of retrofitting CFC11 chillers bears no relevance to the costs of replacing CFC12 with HC12a. Also, as stated above, cost has only limited relevance for SNAP determinations. h. Risk Assessment of Flammable Refrigerants, Part 3: Car AirConditioners  This document attempts to estimate increased risk from using hydrocarbons in MVACs. This study was conducted by Arthur D. Little on behalf of Calor Gas, Ltd. Calor is a British company that has distributed hydrocarbon (HC) fuels for several decades. It has recently introduced a series of hydrocarbon blends for use in newly designed refrigeration and air conditioning systems in the UK. This risk assessment covers CARE 30, one such blend. CARE 30 does not have the same composition as HC12a. As discussed in detail below, certain flammability characteristics may differ.  However, before responding in detail, EPA notes that this paper was originally submitted to EPA for comment by Calor Gas, and it is a preliminary study only. The risk assessment has not been released as a final document. Furthermore, in a letter dated April 23, 1996, Calor Gas stated that the purpose of this study was not to justify the use of HCs in existing mobile AC systems. In fact, you will not be able to find any instance where we have recommended their use in this way. Therefore, this document does not support OZs claim that HC12a is safe for use in unmodified vehicles. EPA has repeatedly stressed the difference between a newly designed system that accounts for a flammable refrigerant, and using such a refrigerant in a system without any modifications that protect vehicle occupants, service technicians, and disposal personnel. Calor agrees with this point, and continued in its letter that it does not doubt that design changes will be necessary if manufacturers are to adopt HC refrigerants. EPA has always welcomed interest in HC refrigerants from manufacturers who investigate such design changes.  In addition, EPA notes that the risk assessment focuses solely on risk to passengers. It does not even attempt to estimate the incremental risk to service and disposal personnel posed by using a highly flammable refrigerant. Technicians have over 40 years of experience with nonflammable CFC12, and have not been trained to handle a flammable refrigerant. This important risk vector has never been quantified in any document submitted to EPA by OZ Technology.  Even lacking quantitative risk assessment for technicians, however, a comprehensive training program can mitigate whatever risk does exist. In its April 23, 1996 letter, Calor Gas stated that it fully supports...the paramount importance of ensuring that flammable refrigerants are handled safely....We have instituted a comprehensive training programme aimed at the service engineer to ensure our products safe use. In fact our sales policy restricts sales only to people who have attended one of our training workshops. However, OZ Technology has never submitted information about a detailed training program, nor has it indicated an interest in mandatory training before selling HC12a or OZ12. In contrast to Calors statement that it is not seeking to unleash these products on an...untrained...retrofit market, OZ has sold its products to a market that has absolutely no training and no experience handling, charging and disposing of flammable refrigerants. Given this lack of training, EPA does not believe that OZ Technology has adequately addressed risks to and mitigation measures for service and disposal personnel.  This risk assessment does have some limited merit. For instance, it accurately describes the premise of a risk assessment: to quantify, using existing data and supportable assumptions, the risk to human health and property posed by various combinations of leak scenarios, ignition sources, and their respective probabilities. However, it is not complete and it relies on many unsupported assumptions to draw its conclusions. It is not necessary to respond in detail to every section; rather, this response will focus primarily on deficiencies and will point out information that is needed to strengthen the report's conclusions.  (1) Chapter 4: Introduction to the Study . This section discusses the 1991 preliminary study performed by Arthur D. Little. This study was addressed in detail in EPA's July 25, 1995, response to the November, 1994 OZ12 petition. The conclusions were preliminary, and the author of the study is on record stating that using it to claim that hydrocarbon refrigerants are safe is premature.  In particular, EPA questions the conclusion that injury is not possible either because of venting in the case of an engine compartment explosion or because of the windows blowing out in the case of a passenger compartment explosion. EPA has not received any medical evidence indicating that a person caught in such an explosion would not be injured. This statement is repeated several times throughout this study, and EPA questions its validity.  In addition, EPA notes that the statistical estimates for the likelihood of system damage caused by a collision is unlikely to sustain scrutiny because the sample size was so small. Nine cars do not comprise a statistically significant sample. Despite EPAs pointing out this deficiency in the July 1995 response, the current study does not include any additional wrecking yard surveys.  (2) Chapter 5: Hazard Identification . EPA notes that the 400g charge of refrigerant considered in this study more closely matches a typical charge size for a car in the US than other studies included in the petition, such as those written by Dr. Maclainecross. However, it should also be noted that charge sizes of over 800g would exist in large cars, and vans with dual systems could contain nearly 1000g. This difference in charge size is significant. Even if EPA were to determine that a 400g charge could be safely used, additional testing and analysis would be required to demonstrate the safety of a larger charge in an MVAC system. Note, however, that EPA still does not believe sufficient evidence exists to demonstrate the safety of using even a 400g charge in unmodified vehicles, for the reasons stated in this response, numerous letters to OZ Technology, and in the July 25, 1995 responses to OZ's prior petitions.  Section 5.3 describes various failure scenarios and conclusions about the likelihood of system leaks in the engine compartment versus in the passenger compartment. These assumptions should be checked against survey data such as those collected by the Mobile Air Conditioning Society.  (3) Chapter 6: Consequence Analysis . This section seems to give scant attention to details that may prove quite significant. First the report states that all electrical contacts will be sealed in a car which is using A3' type refrigerants in its air conditioner. This statement may be accurate for a newly designed car, but it most likely does not apply to current, unmodified models. This issue must be resolved before ignition from unsealed switches can be dismissed in existing cars. In addition, this claim does not address normally isolated contacts that over time may be exposed through normal use.  Next, the report refers to testing performed to determine ignition sources for methane. The report concludes that the characteristics of CARE 30 make it slightly more flammable than methane. EPA questions that characterization of the difference. The lower flammability limit of methane is more than 2.5 times higher than that for CARE 30. Therefore, over 2.5 times as much methane must be leaked into a given area as CARE 30 to yield a flammable concentration. In addition, the ignition energy (i.e., the minimum energy an ignition source must deliver to ignite a given gas) is 15% higher for methane than for CARE 30. Therefore, an ignition source that fails to ignite methane may well ignite CARE 30. Even given these differences, the report concludes that matches and cigarette lighters would ignite CARE 30. Finally, there is no comparison of the flammability limit or ignition energies of CARE 30 and HC12a, so it is difficult to extrapolate any conclusions to HC12a.  Section 6.6.2 discusses overpressure, the maximum pressure rise created during an explosion, which causes effects such as blowing out windows. Overpressure causes damage that is distinct from heat effects from burning refrigerant. Section 6.6.2 concludes that an engine compartment explosion would cause a maximum overpressure of 0.4 psi. It appears that this is a calculated value, rather than the result of testing. In addition, EPA questions the validity of the conclusion that even this overpressure will cause no damage, especially given the table at the bottom of the page that shows that overpressures of as little as 0.3 psi can break windows. Testing would clarify whether such pressures would cause damage to other engine compartment systems. One essential test would evaluate whether the gasoline system was strong enough to sustain an hydrocarbon refrigerant explosion and not cause a secondary fire.  This section also makes several assumptions about damage to passengers caught in an explosion in the passenger compartment, as opposed to an engine compartment explosion, discussed in the previous paragraph. The report concludes in Chapter 9 that "explosions would however blow out the windows rather than harm the occupants." However, section 6.6.2 concludes that maximum overpressure from a passenger compartment explosion could be as high as 2.1 psi. Given that, according to table 6.5, overpressure of as little as 1.5 psi can cause repairable damage to buildings, it seems quite possible that pressure alone could cause injury to passengers, completely apart from fire. In addition, no testing was actually conducted to determine likely overpressures caused by a passenger compartment explosion, and no sources were cited as to the effects of various overpressures on actual occupants.  Section 6.6.3 discusses the potential for injury caused by explosions and/or flame jets in the passenger compartment. Again, it seems that many of the conclusions are simply assumptions. Section 6.6.1 points out that flame jets could reach 2.8m in length. Given that conclusion, it is not clear why the report concludes that such jets emanating from the evaporator would cause burns to passengers only 1% of the time, or why it is very unlikely that flame jet will extend into the passenger compartment...because the evaporator is contained within a partially sealed unit in the dashboard. This conclusion is not based on any testing, but rather seems to be an unsupported assumption. In addition, there is no estimate of the damage to other engine components, such as fuel or brake lines. Another question is whether car interiors in the US meet similar standards to British Standard FNVSS302, related to flame resistance in upholstery. Again, testing is necessary to validate the assumption that flame jets will not ignite materials on the inside of the car, including upholstery and passengers hair and clothing.  Finally, this section concludes that explosions in the passenger compartment will not cause any harm to occupants other than very minor burns. As mentioned earlier, EPA has not received any evidence of such minor injury. This assumption plays a major role in the small calculated risk, and should be confirmed by evidence.  (4) Chapter 7: Frequency Analysis . Section 7.2 refers to the 1991 ADL study, which included a survey of 200 cars from a wrecking yard. It then states an extrapolated risk of refrigerant leakage from a collision, based on UK statistics. This initial calculation must be performed using US statistics in order to make the analysis relevant to US collisions because model designs and safety standards in the US may differ from those in the UK. These differences would make accurate extrapolations from a small UK sample to the US car population difficult. In addition, the report makes no distinction between evaporator, condenser, and other leaks. Such data would yield more meaningful estimates of the likelihood of passenger compartment leaks versus those in the engine compartment.  Instead of using such data, however, the report simply assumes that only 1% of collisionrelated leaks occur in the passenger compartment. In addition, statistics for intrinsic (not related to collisions) leaks were based on refrigerated transport systems. These systems do not use flexible hoses, and can be made much less leaky than those in automobiles. In addition, the report assumes that car air conditioning systems leak 25% less often than refrigerated transport systems. The result is that the estimated leak frequency is only about 0.8% for any given car.##Xx P7XP#  Note that this value is not mentioned in the text, but according to the risk tree pages, is the one used. Apparently the 1.04% leak frequency cited in the text is for refrigerated transport systems themselves.# In contrast, however, a recent survey showed that 75% of all vehicles needing air conditioning service had leaked refrigerant. Detailed estimates of leak rates in the US are necessary; the calculated value of 0.8% is extremely low and most likely does not reflect reality.  Section 7.3 discusses additional questionable assumptions. To begin, the report assumes away all harmful consequence from an engine compartment leak. However, there is no discussion of the potential for secondary fires caused by burning wiring, broken gas or brake lines, etc. If data exist that support such conclusions, they should be submitted or referenced. For example, the authors may wish to consult with the US National Highway Traffic Safety Administration (NHTSA), which conducted testing on OZ12 in 1994. NHTSA may be able to confirm or refute the assumption that engine compartment fires or explosions will have no effect. The report also makes no mention of simple damage to engine components. While such damage may not pose a direct threat to passengers, it may increase the likelihood of loss of control and a subsequent crash.  This section continues with the conclusion that the most serious potential outcome from an ignited event in the passenger cabin arises from the event itself distracting, or through minor injury distressing the driver, sufficient to cause escalation into a car crash. Yet this conclusion is directly related to two factors: 1) the separation of localized and very minor burns, and 2) the very small values assigned to the likelihood of either type of burn. In fact, all of the probabilities in table 7.1 are quite low. It is difficult to imagine, for instance, that a sudden jet of flame erupting from the air conditioning vents would only cause a car crash 1% of the time, or that a passenger compartment explosion would only cause an accident 5% of the time. Similarly, there is no evidence that a flame jet would only cause localized burns 1% of the time. These estimates may be based on the references cited at the end of the report, but no specific citation is used. It therefore appears the estimates are based on assumptions by the authors. Again, EPA recommends that the authors search existing databases on US cars, collisions, and car fires.  (5) Chapter 8: Risk Assessment . Section 8.1 mentions statistics for the UK that describe how many people died because of car fires. Interestingly, no mention is made of secondary fires. Again, EPA questions the assumption that no secondary fires would occur, and requests that statistical evidence be submitted to support this claim.  Section 8.2 makes several assumptions that may be appropriate for the UK, but do not represent the US. For instance, the report estimates that only 10% of cars have air conditioning. In the US, this figure is approximately 90%. And again, this section mentions the assumption that no secondary fires occur.  (6) Chapter 9: Conclusions . This chapter simply repeats conclusions drawn throughout the report. Rather than simply repeat objections and questions detailed above, EPA summarizes several pertinent principles here.  " "EPA does not agree that where information has been lacking, the analysis has erred on the side of conservatism. See the discussion above for several examples.  " "An enormous difference exists between using hydrocarbon refrigerants in existing, unmodified cars, and in newly designed cars. EPA believes it may well be possible to design a hydrocarbonbased system that is safe. The letter from the report's author, Calor Gas clearly states that the company does not believe it is appropriate to use this report to support using hydrocarbon refrigerants in cars not designed for their use.  " "There is no source given for the claim that passenger compartment explosions will not cause serious injury to occupants.  " "As the report notes, it has been beyond the scope of this study to examine the detailed design and structural integrity of the air conditioning system components. Further study would be needed. EPA wholeheartedly agrees. In particular, data are needed to address the question of secondary fires.  " "Data appropriate to the US must be used when calculating risk. Although there are some similarities between these data and those for the UK, there are also significant differences, such as the difference in the percentage of cars that have air conditioning.  " "This study does not address training at all, thereby ignoring the safety of service and disposal personnel.  " "The report does not present any crash test data to verify any assumptions. See the discussion above for more detailed examples. i. Hydrocarbon Risk in Car AirConditioners  This 1995 report by Dr. Ian MaclaineCross and E. Leonardi does not respond to the concerns raised by EPA in response to Dr. Maclainecross' first paper, which was submitted to support OZ Technology's November 1994 petition to find OZ12 acceptable. It is not a risk assessment, and it does not include any data that allow for an estimate of the risks from flammability to passengers, service personnel, or disposal personnel. Rather, it attempts to estimate the expected value of insurance claims from using hydrocarbon refrigerants. Furthermore, it does not focus on HC12a, but rather on hydrocarbons in general.  First, EPA notes that nowhere does Dr. Maclaine-cross correct his earlier dismissal of potential transportation and servicing risks. Automobile service personnel have received no training whatsoever in the handling of highly flammable materials, unlike professionals in the liquid petroleum gas industry. A risk assessment must address how service technicians would be trained to minimize the flammability risks of using hydrocarbons. Section 4, despite being titled "Car mechanics' accident scenarios," does not address these data needs at all. Rather, it attempts to respond, without any scientific data, to various claims made about the hazards of using flammable refrigerants.  In section 5, as in the 1994 paper, the ignition test on a hot engine is inconclusive given the data presented. The autoignition temperature can only be compared to a known engine temperature. It is unclear what temperature is created by running the engine for 10 minutes in "fine weather." Furthermore, this test did not address other driving scenarios, such as extensive stop-and-go traffic, highway driving, or starting a car after sitting in the sun for several hours. Such tests would be necessary to assess whether typical car engines can reach the autoignition temperature of hydrocarbons.  In addition, the authors claim that open relays, which could provide an ignition source, are uncommon in cars. However, the authors are based in Australia and they present no information about typical manufacturing practices for US models. Such details would be necessary to allow a reasonable evaluation of the potential for these relays to represent an ignition source. Finally, the authors should address the typical wear of such components over the life of a typical car, paying special attention to potential exposure of originally sealed relays.  In section 7, the authors claim that hydrocarbons are flammable in the same manner as paint, plastic, antifreeze, engine oil, and petrol (gasoline). However, all of these substances are liquids at room temperature, and none are pressurized. The primary flammability concern for liquids is the ignition of vapors above open containers. In contrast, HC12a is a gas at room temperature and is used in a pressurized system, and several scenarios could pose flammability risk to people using HC12a, including the sudden release of HC12a into the passenger compartment of a car because of a collision. Therefore, this comparison does not shed any light on the question of fire risk from flammable gaseous refrigerants like HC12a.  In section 8, the authors describe a test conducted in Florida in which OZ12 was ignited after being released into the passenger compartment of a car. The explosion caused considerable damage to the vehicle. The test itself did not yield estimated flammability risk, but it certainly demonstrated the potential for risk. In addition, the test allegedly used 5 1/2 ounces of OZ12. The normal charge is likely much larger for larger automobiles. According to OZ Technology's sales brochure, a more typical charge of HC12a would be 15 ounces. Another test, performed in North Carolina and using two cans, or 12 ounces of HC12a, also resulted in an explosion that caused considerable damage to the car.  The discussion of the accident scenario is precisely the type of occurrence for which a risk assessment is necessary. However, the authors present neither an estimate of the probability of ignition caused by a collision nor an estimate of the risk of explosion. Although the authors presume that opening a window would dissipate a flammable cloud, they present no estimate of the likelihood that the passengers would be unable to open the windows because they were stunned or unconscious. Simply assuming that they would open them does not provide an accurate estimate of the risk of a fire or explosion. Finally, the authors' claims that there would be little or no bodily damage to the passengers from being caught in such an explosion is entirely unsubstantiated.  The author estimates system failures due to fatigue, but presents no experimental evidence to support those estimates. For instance, on what basis does he estimate that fatigue fractures occur every million years? What is the estimated probability of a sudden failure specifically because of a collision? Finally, when assessing the risks during a collision, the author uses charge sizes unlikely to be used in the US. Despite the authors' assumption that charges would be limited to 300g, it is likely that a more typical charge in US models would be at least 400g, as discussed under the response to Appendix A to "Supplemental Risk Assessment Regarding the Use of HC12a Refrigerant in Automotive Air Conditioners." Because of these assumptions, it is difficult to assess the risks of fire. As Dr. Maclainecross stated in his 1994 paper: Collisions are very important in determining insurance risk since they may create a source of ignition and a flammable mixture at the same time. If these are in the same place a fire results....crash testing is necessary to measure this risk with greater accuracy.  Yet this report does not reflect such testing. In sum, like the 1994 report by the same author, this paper does not represent a valid risk assessment. j. Hydrocarbons and Other Progressive Answers to Refrigeration This document, dated October 1995, is a Greenpeace compendium and includes over 20 papers on the use of hydrocarbons in a variety of newly designed refrigeration and air conditioning systems. It also contains several paper that discuss nonhydrocarbon technologies, which are clearly irrelevant to the safety of HC12a. The hydrocarbon papersfall into a few broad classes: - reports on performance of hydrocarbons - reports on the use of hydrocarbons in new equipment- reports on environmental regulations and characteristics of various alternatives - report by Dr. Ian Maclaine-cross, which is addressed in depth above paper by Gary Lindgren, CEO of OZ Technology  Clearly, the first three classes of papers are not relevant to the use of hydrocarbons in existing systems like motor vehicle air conditioning. Dr. Maclainecross' paper is addressed in depth above, and Mr. Lindgren's paper contains statements addressed throughout this response.  None of these papers is a risk assessment. Aside from the papers by Dr. Maclainecross and Gary Lindgren, none discusses safety except to mention specific measures related to new manufacturing. Only one abstract discusses retrofitting refrigeration systems to use hydrocarbons, and it also is not a risk assessment (the full paper is not included in the compendium). Rather, the abstract identifies several concerns associated with hydrocarbon refrigerants, including the identification of safety precautions.  As EPA has maintained for over two years, hydrocarbons offer potential gains in energy efficiency and GWP versus both the ozonedepleting CFCs they replace and other alternatives. EPA has consistently supported responsible development of hydrocarbon refrigerants. EPA agrees with Greenpeace that hydrocarbons offer promise as CFC alternatives. However, EPA notes that in a letter to OZ Technology, Greenpeace specifically denied any endorsement of OZ Technology's products.  Responsible development includes designing systems with the flammability of these refrigerants in mind. Several standards exist in other countries that govern charge size and require certain system components. In addition, other countries' legal systems recognize efforts by manufacturers to reduce risks in systems they design and build. In the US, however, such standards are only beginning to be explored, and section 612 of the Clean Air Act requires EPA to prohibit the use of more hazardous refrigerants when other, safer alternatives are available. Furthermore, OZ Technology proposes to use HC12a in existing systems, with no modifications, and it is not at all clear that such use is safe or appropriate. The fact that hydrocarbons are being used safely in new systems in other countries bears little or no relevance to their use in existing systems in the US. For example, one writer recommends the use of an indirect system on the evaporator side of a refrigerated truck or a car air conditioning system. However, this design is not used in current US models. Finally, refrigeration systems differ from each other and from motor vehicle air conditioning, so safety in one type of system does not imply safety in the other. EPA has, however, repeatedly requested that Greenpeace submit risk assessments for the GreenFreeze refrigerator. To date, EPA has yet to receive such assessments.  Interestingly, one paper refutes OZ Technology's contention that adding an odorant to HC12a removes the flammability risk because auto passengers or people in a kitchen would immediately open the windows when they notice the smell. The paper, titled "Breakthroughs in Mobile Hydrocarbon Refrigeration and Cooling", by Dr.Ing. Michael Kauffeld, reports that "Norwegian studies, however, have shown that the normally used tracer ethyl mercaptan disappears during operation of the plant. It probably migrates into the oil." Therefore, additional testing is required to demonstrate that the odorant does, in fact, remain active even after prolonged use of HC12a. k. Health and Safety Issues Related to the Use of HFC134a as an Auto AirConditioning Refrigerant  This undated document is discussed under paragraph 17 below. Paragraph 12  As noted above, HFC134a and all other substitutes for CFC12 have been reviewed in accordance with the same standards applied to HC12a. Also, as noted above, based on this thorough review, HFC134a was found to be an acceptable substitute. In response to this petition, EPA is reaffirming that acceptability listing. Paragraph 13  This paragraph simply reiterates the purpose of the petition. Paragraph 14  As stated above, numerous studies have demonstrated the relative safety of HFC134a. OZ Technology has submitted no further tests or studies to controvert those that were the basis of EPA's determination that HFC134a is an acceptable substitute.  Given the nonflammable, nonozonedepleting, low toxicity nature of HFC134a, it poses acceptably low risk within those vectors. It also does not contain chlorine, so it contributes no chlorine loading. The global warming impacts of HFC134a are quantified by the GWP, which is 1300. While this value is higher than that for hydrocarbons, it is within the range of values for other acceptable alternatives. EPA compared all of these factors both to those of CFC12 and to those of other available alternatives. Furthermore, as parties submit new refrigerants for review, EPA compares each substitute's characteristics with those of the existing set of acceptable alternatives.  The primary risk factor of concern for HC12a is flammability. As described above, it is clear there is a potential risk in using HC12a that is not present when using any nonflammable or practically nonflammable refrigerant, such as HFC134a. Because OZ Technology has not submitted adequate risk assessment data, EPA is unable to quantify this risk. However, it is clear that the other alternatives found acceptable to replace CFC12 do not pose this risk. In addition, the other characteristics of the other acceptable alternatives, including HFC134a, are comparable (see the Table, below). Paragraph 15  EPA cannot conduct independent analyses of every substitute submitted for review under SNAP. Neither the resources nor the legal mandate exist for such analyses. Rather, as explained in the final SNAP rule, EPA requires that companies interested in manufacturing or distributing substitutes must submit specific information for review. EPA then compares the data for a given substitute with those for the original ozonedepleting substance and the set of acceptable substitutes. These comparisons do not represent testing of individual substitutes, although flammability is one of the risk characteristics included in the comparisons.  In response to both this paragraph and paragraph 14, a table comparing the various characteristics of all of the refrigerants submitted to date follows. EPA has long accepted the approach of using HCFCs with low ozone depletion potential (ODP) as shortterm replacements for CFCs. The consequences of using HCFCs have been quantified in the Regulatory Impact Analysis (RIA) conducted for the December 10, 1993 CFC phaseout rule (58 FR 65018). The latest addendum to the full RIA, which is several hundred pages in length, may be obtained from the USEPA Air Docket (2022607548) as file number IIA15 within docket A9213. While the HCFCs and HFCs do exhibit GWPs, EPA has determined that none are so much higher than the other acceptable substitutes that they pose additional risk. As is clearly shown, the vast majority of substitutes reviewed under SNAP pose neither flammability nor toxicity risks. Ammonia, although flammable and slightly toxic, is only used in systems designed to minimize the risk posed by its toxicity. HFC152a was shown by a risk assessment to be safe for use in new household refrigerators, which is the only refrigeration and air conditioning enduse for which it is acceptable. Furthermore, note that both ammonia and HFC152a have significantly higher flammability limits than hydrocarbons (i.e. more of these refrigerants is required to create a flammable mixture in air), and the energy released when they burn is much lower. All of the acceptable substitutes, therefore, have been compared to one another and found to be acceptable for specifically identified enduses. #m*0 x7# r44aFr 44aFr r &BBBBBRBBBBBR&#Xx P7XP#Substitutes for CFC12 in Air Conditioning & Refrigeration under the Significant New Alternatives Policy (SNAP) Program as of February 8, 1996&BRBBBRBRBBBR&#x*0 x7#Substitutes (Name Used in the Federal Register) Trade Name (contents)Ozone Depletion Potential*Global Warming Potential*Flammable?**Toxicity#Xx P7XP#***&TIQIIIQTIQIIIQ&#x*0 x7#HCFC2222 (pure)0.051700nolower&TIQIIIQTIQIIIQ&HFC134a134a (pure)!01300nolower&TIQIIIQTIQIIIQ&HFC152a152a (pure)!0140lowerlower&TIQIIIQTIQIIIQ&HFC227ea227ea (pure)!03300nolower&IQIIIQIQIIIQ&R401A, R401B, R401CMP39,MP66, MP52 (22/124/152a)0.05/0.03/01700/480/140nolower&TIQIIIQTIQIIIQ&R406AGHG (22/142b/600a)0.05/0.06/01700/2000/?no#A\  PP#lower&IQIIIQIQIIIQ&R409A (HCFC Blend Gamma)R409A (22/124/142b)0.05/0.03/0.061700/480/2000nolower&IQIIIQIQIIIQ&R411A, R411BR411A, R411B (22/152a/1270)0.05/0/01700/140/?nolower&TIQIIIQTIQIIIQ&(HCFC Blend Beta)FRIGC(124/ 134a/600)0.03/0/0480/1300/?nolower&TIQIIIQTIQIIIQ&(HCFC Blend Lambda)GHGHP (22/142b/600a)0.05/0.06/01700/2000/?no#x*0 x 7#lower&IQIIIQIQIIIQ&Ammonia Vapor Compression!0?lowerhigher&TIQIIIQTIQIIIQ&Evaporative Cooling!0NAnolower&TIQIIIQTIQIIIQ&Desiccant Cooling!0NAnolower&TIQIIIQTIQIIIQ&Ammonia/water Absorption!0NAlowerhigher&IQIIIQIQIIIQ&Water/Lithium Bromide Absorption!0NAnolower&IQIIIQIQIIIQ&OZ12, HC12a, other hydrocarbons (Hydrocarbon Blend A, Hydrocarbon Blend B)OZ12, HC12a (hydrocarbons)!0?higherlower&TIQIIIQTIQIIIQ&R176(12/22/142b)0.9/0.05/0.068500/1700/2000nolower&<JRJJJR<JRJJJR&R405AR405A (22/142b/152a/c318)0.05/0.06/0/01700/2000/140/9100nolower#Xx P 7XP#* Source: Scientific Assessment of Ozone Depletion: 1994. Note that GWPs for the hydrocarbons are not available and thus are listed a "?". However, it is expected that they will be less than 10. Alternative technologies marked "NA" do not have a direct global warming potential. ** As determined according to the ASTM E681 testing procedure, which tests for flammability limits at room temperature and normal atmospheric pressure. This test forms the basis for ratings by UL and ASHRAE. "Lower" flammability indicates that the lower flammability limit is above 0.10 kg/m3 and the heat of combustion is below 19,000kJ/kg. "Higher" flammability indicates that either the lower flammability limit is less than or equal to 0.10 kg/m3, or the heat of combustion is greater than 19,000 kJ/kg. "No" indicates that the refrigerant does not exhibit flammability limits and is, therefore, nonflammable. The hydrocarbons are all rated in the higher flammability category, although HC12a has not been submitted to either UL or ASHRAE for a flammability rating. *** As rated by ASHRAE. Lower toxicity indicates that "toxicity has not been identified at concentrations less than or equal to 400 ppm, based on data used to determine Threshold Limit Value TimeWeighted Average or consistent indices." Higher toxicity indicates that there is evidence of toxicity below 400 ppm. Note that ammonia is the only refrigerant in the higher toxicity category, and it is only used in specially designed systems and in amounts that are consistent with this toxicity, as set out in ASHRAE standard 15. Paragraph 16  EPA is unaware of any studies that it has not already analyzed regarding safety issues related to the use of HFC134a. Although OZ Technology maintains that HFC134a can be lethal if inhaled through a cigarette, the mere statement of such a concern is not sufficient demonstration of its validity in the face of years of research that demonstrate the low toxicity of HFC134a. As discussed above, nothing submitted to EPA by OZ Technology or anyone else to date demonstrates that the dose of breakdown products would be of any concern. If you have such documentation, please submit it immediately.  In addition, the focus should be on flammability, not on the autoignition temperature of HFC134a. Steel, at the proper temperatures, can also be made to burn. Yet no one would mistake steel for a flammable substance. The term "flammable" is not an arbitrary one. Engineering societies such as the SAE and ASHRAE use a standard testing procedure, ASTM E681, to determine the flammability of refrigerants. If a product exhibits flammability limits, it is determined to be flammable. Under that test procedure, HFC134a has repeatedly been demonstrated to be nonflammable.  UL lists three flammability categories. "Nonflammable" means the refrigerant exhibits no flammability limits at normal atmospheric pressure and has an autoignition temperature above 750 degrees Celsius. "Practically nonflammable" means the refrigerant exhibits no flammability limits at normal atmospheric pressure and has an autoignition temperature below 750 degrees Celsius. HCFC22 and HFC134a are both in this category. Again, no one claims HCFC22 is flammable simply because its autoignition temperature is below 750 degrees Celsius. Flammable refrigerants exhibit flammability limits in air at normal atmospheric pressure, regardless of their autoignition temperature. The alkanes, such as propane and butane, are classified as flammable by UL. Again, however, if you possess scientifically valid test results that show HFC134a is flammable, please submit them. Paragraph 17  The petition did not include an Exhibit F. The only attachment other than Exhibit A is a paper titled "Health and Safety Issues Related to the Use of HFC134a as an Auto AirConditioning Refrigerant." Since there was no attached Exhibit F, I presume this document is what you intended to include, and I will respond to it as such. First, however, note that a discussion of the refrigerant alternatives reviewed in the final SNAP rule is included in the background document for the refrigeration and air conditioning sector. This document may be found in USEPA air docket A9142, number VB5, titled Risk Screen on the Use of Substitutes for Class I OzoneDepleting Substances: Refrigeration and Air Conditioning. An addendum to this document is found at number VB5a.  The flammability of HFC134a at atmospheric pressure is addressed above. When released to the atmosphere, HFC134a is by definition at atmospheric pressure, and it does not exhibit flammability limits. When HFC134a is pressurized, and when mixed with air in volumes greater than 60%, it may become combustible. However, it is impossible for a refrigeration or air conditioning system to contain that proportion of air and still operate. As stated in the paper itself, even the low side of an auto air conditioning system is significantly above atmospheric pressure. Therefore, it is impossible for large amounts of air to enter the system. If you have evidence that it is possible for such a mixture to occur, please submit it immediately.  As stated above, the concern that passengers or fire fighters might inhale significant amounts of hydrogen fluoride (HF) is unsubstantiated. Please submit any evidence that demonstrates that the concentrations would approach a level that is hazardous. As discussed under paragraph 11, the Southwestern Laboratories report demonstrated that even if the entire charge of a car air conditioning system were to leak into the passenger compartment, exposure to HF would be limited to at most 1 ppm. This exposure is only 3% of 30 ppm, the emergency level set by the US National Institute of Occupational Safety and Health as immediately dangerous to life and health. In fact, it is even below the longterm exposure TLV of 3 ppm. Keep in mind that 3 ppm of HF does not pose an instantaneous danger; rather, it is the level at which continuous exposure will not result in a health risk.  It is not sufficient to simply claim that a hazard exists with respect to HFC134a. There must be a reasonable expectation that the risk is real. EPA had ample evidence to suggest that a flammability risk could be posed by HC12a. The Department of Transportation concluded there was sufficient risk posed by OZ12 to request its recall. In addition, tests conducted by the state of Florida demonstrated that hydrocarbons in a car can explode. The Florida test did not comprise a valid risk assessment, but it did demonstrate the potential for risk. Finally, OZ Technology's brochure and label both describe HC12a as flammable. Paragraph 18  As explained in response to paragraph 15, EPA does not have the resources or the mandate to test every alternative. Rather, the manufacturer must submit appropriate data, as specified in the final SNAP rule. Furthermore, as stated above, EPA believes that adequate data have been submitted and analyzed regarding HFC134a. Paragraph 19  This letter constitutes the response described in the final SNAP rule. Paragraph 20  See the table included in the discussion under paragraph 15. Paragraph 21  EPA will add this response to the public docket for the SNAP rule. In addition, EPA will publish a Notice alerting the public that the response exists and explaining how to obtain the response from the docket. The Notice will also explain how to obtain a copy of the response from EPA's World Wide Web site. (a) As stated numerous times, the SNAP rule does not regulate secondgeneration substitutes. It is not EPA's role to promote a specific product, nor to make declarations concerning the regulation of HC12a by other jurisdictions. For example, as you are well aware, several states, including OZ Technology's home state of Idaho, prohibit the use of HC12a and all other flammable refrigerants in car air conditioning. (b) and (c)See the discussion of the data submitted as part of this submission under paragraph 11. In addition, EPA has articulated the flammability concerns related to the use of HC12a above, in numerous letters to OZ Technology, members of Congress, and the general public, and in public presentations. It is incumbent upon OZ Technology to submit comprehensive, scientifically valid risk assessments to demonstrate that it is safe to use a highly flammable refrigerant in unmodified, retrofitted systems that were designed to use CFC12, a nonflammable refrigerant. This risk assessment must include specific data related to the use of HC12a under several scenarios, including installation, use, accidents in the case of car air conditioning, leaks in all cases, recovery, and disposal. In addition, if OZ Technology wishes to use HC12a in new equipment, a risk assessment must address safety risks associated with charging, packaging, transporting, and installing such equipment. For each scenario, a risk assessment must analyze in a detailed manner the probability of various types of release, the types of ignition sources and their likelihood, the probabilities of ignition occurring near a flammable mixture of HC12a and air, and the results of an ignition or explosion. Finally, these risk assessments must explain measures to reduce any risks identified in the previous analysis. Such risk assessments are extremely dependent on enduse and refrigerant. (d) Section 612 of the Clean Air Act, which is implemented by the SNAP rule (40 CFR Part 82, Subpart G), provides the legal authority for making SNAP listing decisions. (e) EPA has repeatedly stated that it does not favor certain companies or individuals over others. In fact, numerous small companies are now marketing substitutes found acceptable under the SNAP program in many use sectors. The SNAP final rule stated quite clearly that it is not EPA's role to determine which refrigerants, among those found acceptable, will find success in the market.  Several documents in the docket demonstrate that EPA is not biased against OZ Technology or the use of its products. Memoranda from the Stratospheric Protection Division, within EPA's Office of Air and Radiation, to the EPA Stratospheric Hotline and to Regional staff, clearly state that these personnel should rely on EPA's HC12a fact sheet when responding to queries. In addition, in numerous public presentations, EPA personnel have stated the position that hydrocarbons show promise as refrigerants, with the understanding that responsible development includes assessing and minimizing flammability risks. (f) See the discussion above, in particular under paragraphs 1517, the final SNAP rule, and documents such as the refrigeration and air conditioning background document in the EPA Air Docket. (g) See the discussion above, in particular under paragraphs 1517, the final SNAP rule, and documents such as the refrigeration and air conditioning background document in the EPA Air Docket. SECTION II RESPONSE TO OTHER DOCUMENTS  In addition to responding to the documents included with this petition as Exhibit A, EPA will reply to several documents submitted by OZ Technology over the past year, including: A."Supplemental Risk Assessment Regarding the Use of HC12a Refrigerant in Automotive Air Conditioners," submitted by Charles Lempesis on June 5, 1995. B."HC12a Supplemental Risk Assessment for Home Refrigerators," submitted by OZ Technology on May 31, 1995. C.Videotape, "Hydrocarbon Refrigerants/Supervised Safety Tests," submitted by Bob Small, Executive Director, OZ Technology on November 25, 1995. D."A CHEMICAL DISASTER, Why HFCS Have No Future," Greenpeace, submitted by Gary M. Lindgren, CEO, OZ Technology on July 16, 1996. The response to each document follows. A.  Supplemental Risk Assessment Regarding the Use of HC12a Refrigerant in Automotive Air Conditioners  1.Preface  The preface to this document quotes large portions of the 1991 Arthur D. Little risk assessment. In July, 1995, EPA responded to this document as part of the response denying OZ Technology's November and December 1994 petitions to list OZ12 and HC12a as acceptable. In addition to the section repeated below, it should be noted that the sample size for this report was only nine cars, far too small to represent a statistically valid portion of the population. Part of that response stated: "EPA received a letter from John Dieckmann, a primary author of the A. D. Little Report, that stated: The study documented in this report was a preliminary investigation...The study addressed applications to new vehicles, as opposed to retrofit to existing vehicles...the preliminary conclusion of the study is that fire risks could prove to be acceptably low; however, extensive testing and data collection will be required to verify this conclusion and identify and validate any design changes needed to reduce the fire probabilities... The bottom line: The study indicates that certain flammable refrigerants are a promising option for "environmentally friendly" and safe automobile air conditioning. However, based solely on the results of this study, it is premature to conclude that hydrocarbons can be used as a retrofit refrigerant in automobile air conditioning systems with an acceptable level of safety. [Dieckmann's emphasis in each case] Thus, EPA does not believe the A. D. Little report demonstrates the safety of using hydrocarbon refrigerants in existing, unmodified MVACS."  The OZ risk assessment then discusses testing by OZ Technology regarding normal leakage. Normal leakage, however, is not the most pressing concern. EPA agrees that during operation of a vehicle, normal leakage will normally be too slow to pose a significant fire risk. However, this assessment does not address the likelihood of sudden release following a collision, a catastrophic failure of a system component, or potential hazards to service or disposal personnel.  The tests to determine whether a lit cigarette will ignite HC12a are inconclusive. In order for ignition to occur, it is first necessary that a flammable concentration exist. If either too much fuel or too little fuel is mixed with air, the mixture will not ignite. A jet of pure refrigerant is most likely considerably above the upper flammability limit. Tests should be conducted to show that in the presence of measured flammable concentrations, a cigarette will not ignite the mixture.  As mentioned above under document 10, the addition of an odorant may reduce the risks to passengers by alerting them to a strange smell, therefore increasing the likelihood they will open a window. However, also as described above, a Norwegian study concluded that ethyl mercaptan was most likely absorbed in lubricating oil. OZ must address this possibility through testing with typical mineral oils used in existing systems.  Similarly, the hot plate test did not address what happens when a collision allows a large amount of flammable material to mix with air under the hood after a collision. Simply asserting that air flow will quickly dissipate a flammable cloud does not respond to the aftermath of such a collision, when the vehicle is no longer moving. Playing a jet of pure gas over a hot plate simply demonstrates once again that pure refrigerant will not readily burn because it is above the upper flammability limit.  This document next discusses a report by NHTSA's Office of Defects Investigation. This report concluded that the use of OZ12 poses an unacceptable risk of fire when used in automobile air conditioning. NHTSA requested in a letter that OZ Technology voluntarily recall OZ12. OZ Technology refused to recall OZ12, but did agree to stop selling it. As OZ is aware, the contention that the report resulted in a finding of "No Safety Defect Trend" is simply false. EPA has entered into the public record a letter from NHTSA to OZ alerting OZ of an erroneous NHTSA summary table that has since been corrected. Despite this correction, OZ continues to repeat the error in brochures, this petition, and letters to EPA and others.  This document next discusses testing by OZ Technology, including firing bullets at cans of OZ12 and releasing cans of OZ12 under a running engine. It also discusses a paper by Dr. Ian Maclainecross. Both of these items were responded to as part of the July 1995 EPA response denying the petition to find OZ12 acceptable. Neither document represents a risk assessment or a relevant test of the flammability risk posed by the use of OZ12 or HC12a.  Finally, the preface to this document discusses a second paper, addressed above, by Dr. Maclainecross and a test performed in Florida. To briefly recap, Dr. Maclainecross failed in this second report to respond to issues raised by EPA in regard to his first paper. The test in Florida no more represented an accurate risk assessment than did firing bullets at cans of OZ12. However, the former did demonstrate that OZ12 will ignite when exposed to a spark, and the latter demonstrated that bullets penetrating a can of OZ12 will not ignite the gas. The relevance of the first item is that the first showed the potential for risk, thereby justifying EPA's insistence that OZ demonstrate the safety of using HC12a in unmodified, existing systems. It is difficult to identify the relevance of the second item.  2.Discussion  The initial discussion section of this document accurately describes the placement of the major system components. It does not, however, mention that the condenser is typically located immediately behind the grille. This placement puts it in danger of puncture during a frontend collision.  a. Evaporator Rupture Risks  Although evaporators are somewhat shielded from collisions, the Arthur D. Little study examined a very limited number of cars. To be specific, only 9 cars were evaluated. Given the fact that potentially 90 million cars are still using CFC12, and typical risks are likely in the several per million car range, this sample size is far too small to yield a meaningful result. As stated above, furthermore, this study was only a preliminary analysis. John Dieckmann, the author, has explicitly stated in a letter that EPA has placed in the public record that he does not believe it is appropriate to draw conclusions from this study regarding the use of flammable refrigerants in existing air conditioning systems.  b. Underhood Component Risks  Again, this discussion relies heavily on the 1991 Arthur D. Little study, which should not be used as conclusive evidence of the safety of hydrocarbon refrigerants. However, it should also be noted that the study stated that engine compartment fires were assumed not to be a threat to injury. It is not at all clear what was the basis for this assumption. Finally, although the study recommends several design modifications to sharply reduce the dangers of engine fires, OZ Technology proposes to use HC12a in unmodified car air conditioning systems. Risk from fire is extremely sensitive to system design; claiming low risk based on assumed modifications that are not actually made is an inappropriate use of even these preliminary conclusions.  c. Risks from Slow Refrigerant Leaks  EPA believes that typical slow leaks probably pose extremely low risk of fire. Typical leaks on the order of one pound per year will likely never create a flammable concentration inside the passenger compartment. However, sudden loss from failed evaporators does occur, and this event can certainly create a flammable mixture. In a typical car, with an interior volume of 120 cubic feet, the sudden release of a typical 400g charge will create a concentration well over the lower flammability limit. A risk assessment must use available data (for example, from the Mobile Air Conditioning Society's annual survey) to assess the potential risk of evaporator failure.  3.Conclusion  Several of the claims responded to above are repeated in the conclusion section of this document. A few specific points are worth mentioning. First, despite the addition of an odorant, at least one study shows that the odorant may be removed from circulation by the oil. In addition, in the event of an evaporator failure, especially after a collision, it is not at all obvious that the passengers would be conscious or coherent enough to roll down the windows. Furthermore, sparking from broken wires could ignite a flammable mixture before passengers have time to escape. These scenarios must be addressed before a final conclusion can be drawn regarding the risk from the flammability of HC12a. They cannot simply be assumed to be zero.  Second, the fact that burning hydrocarbons do not produce toxic byproducts is beside the point. What are the consequences of being in a car when hydrocarbons burn or explode? EPA has never received evidence that reddened skin is the primary result, despite numerous claims to this effect.  Third, as explained in detail above, the risk from inhaling HFC134a through a burning cigarette is infinitesimal. The dose received would be well below the hazard threshold. Similarly, those in the vicinity of a fire would be in no danger, whereas they may well be threatened by a hydrocarbonfueled fire.  Finally, the fact that other flammable fluids are present in a car is irrelevant. The only relevant factors are those that address the added risk posed by using a flammable refrigerant like HC12a in a system never intended to use flammable substances. The risks posed by a gasoline tank are inherent in a vehicle, and certain design modifications mitigate potential risks. The same cannot be said for unmodified CFC12 based automobile air conditioning systems. 4.Appendix A: Study: Safety of Hydrocarbon Refrigerants for Car Air Conditioning Systems  This study accurately identifies many of the scenarios of concern when using hydrocarbon refrigerants, especially in unmodified cars. It does not, however, then address those concerns.  Chapters 13 simply discuss ozone depletion and refrigerant characteristics. EPA agrees that hydrocarbons pose no ozone depletion potential and low global warming potential. Furthermore, they are low in toxicity, inexpensive to produce, and they may offer energy efficiency gains over the use of CFC12 or HFC134a. However, they are highly flammable. Chapter 4 identifies several potential hazards posed by the flammability of hydrocarbons. Chapter 5 discusses components of an automobile air conditioning system. None of these chapters, however, provides a risk assessment.  Chapter 6 describes a test wherein a torch was blown out and then played over various engine components. As described in the response to the November 1994 petition, however, this test is inadequate to demonstrate that engine components cannot ignite hydrocarbon refrigerants. The autoignition temperature can only be compared to a known engine temperature. It is unclear what temperature is created by running the engine at idle in a garage. It also does not address the results of a sudden leak as a result of a collision that creates a cloud of refrigerant under the hood. Furthermore, this test did not address other driving scenarios, such as extensive stop-and-go traffic, highway driving, or starting a car after sitting in the sun for several hours. Such tests would be necessary to assess whether typical car engines can reach the autoignition temperature of hydrocarbons. Finally, the test does not address faulty switches or sparks caused by friction during a collision. The remainder of this chapter describes the test of refrigerant flow through the passenger compartment and does not address these issues.  Chapter 7 discusses ignition of hydrocarbon refrigerants. In section 7.4, and in subsequent chapters, reference is made to a 300g charge. However, engineering experience in the US indicates that a typical 2 pound CFC12 charge would be replaced by about 1.8 pounds, or 30 ounces, of HFC134a. According to OZ's own HC12a sales brochure, 30 oz. of HFC134a would be replaced by about 2.5 cans, or 15 oz. of HC12a. Fifteen oz. is over 400g of HC12a. Note also that this is a typical charge, and by no means representative of the charge found in larger automobiles, which could exceed 800g. The remainder of section 7.4 makes various unsupported assumptions that were addressed in the response to the November 1994 petition. For example, there is no explanation of the assumption that a fatigue fracture in the engine bay during operation may occur once in ten thousand operating years, or the assumption that less than 1% of the vehicle population may contain ignition sources.  Chapter 8 describes modeling and experiments that resulted in calculations of concentrations of CO2 in various cars. Note that the largest charge was slightly less than 300g, and that test resulted in a "safe time" of 12.5 minutes when the air vent was closed, as would be common during a hot day. As mentioned above, however, it is quite likely that a charge at least 33% larger will exist in most cars, with even larger charges quite common. Even with the fresh air switch on, the 300g charge created a flammable concentration for nearly 30 seconds. Note that these tests all assume normal operation of the car and normal air flows, and still do not address collisions. It is unclear how these tests demonstrate that HC12a is safe for use in automobile air conditioners. In fact, they show that there is a potential for a hazard.  Section 8.4 discusses the results of the air flow tests. As mentioned above, at least one test shows that the odorant is absorbed by the oil and loses its potency. Therefore, occupants may not be aware that a flammable concentration exists in the passenger compartment. Furthermore, the occupants may not be conscious or sufficiently coherent following a collision to recognize the odor and to roll down the windows. Several additional unsupported assumptions are made in the calculations about the dangers faced by a smoker in a car with a flammable refrigerant. For example, there is no basis for the assumption that the probability of ignition of a flammable mixture is 1 in ten million. Again, rather than demonstrating the safety of hydrocarbon refrigerants, this section merely raises additional questions.  Chapter 9 simply discusses conclusions drawn in the rest of the paper. It is not necessary to repeat the objections described in detail above. 5.Appendix B: Study: Field Trials of Propane/Butane in Automotive Air Conditioning  This study minimally addresses the safety of using hydrocarbon refrigerants in unmodified automobile air conditioning systems. There are only two sections that mention safety in the entire paper. The first cites a paper by James Calm as evidence that all refrigerants carry some safety hazard. EPA agrees that all refrigerants have the potential to cause cardiac sensitization and asphyxiation, including hydrocarbons. However, that statement does not negate the potential risk posed by hydrocarbons' flammability. This section then refers to Dr. Maclainecross' papers, which have been addressed above and in the response to the November 1994 petition. The second section referring to safety and flammability assumes that because hydrocarbons were released without incident during performance testing, they are therefore somehow safe in all situations. This paper contains no information relevant to the question of the flammability risk posed by hydrocarbon refrigerants, but rather refers to other papers addressed previously by EPA. B. HC12a Supplemental Risk Assessment for Home Refrigerators  This document purports to demonstrate the safety of using HC12a in household refrigerators by evaluating potential ignition sources and the probability of ignition of HC12a. It is directly based on the 1991 Arthur D. Little (ADL) risk assessment that examined the risks of using HFC152a in such systems and was derived from testing by Underwriters Laboratories (UL). Neither UL nor ADL is in any way connected to the manufacture or sale of HFC152a. Note that in section 2.1, Introduction , the assessment incorrectly implies that EPA contracted with Arthur D. Little to conduct a risk assessment because of interest in HC12a. As mentioned above, the 1991 risk assessment concerned HFC152a, not HC12a.  The majority of this document appears to be taken word for word from the 1991 ADL risk assessment. OZ performed supplemental testing of various ignition sources in the context of using HC12a in a household refrigerator. There is no discussion of the qualification of OZ Technology personnel to conduct such testing. In cases where OZ's results conflict with those of UL, EPA considered the reputation and capabilities of the relevant labs.  Sections 13 contain background information about the CFC phaseout, the need for substitutes, and refrigerator system layout and components. However, throughout the remainder of the document, there are several errors in calculations and references to other sections. For example, the numbering scheme for tables and figures is inconsistent, and is nonsequential, leaving the reader to wonder whether something was omitted.  The first part of section 4.2 describes potential leaks. The rest of section 4 purports to discuss ignition sources and the probability of their igniting various leak sources. However, despite the fact that both tables 4.4 and 4.5a seem to indicate positive probability of ignition for various ignition sources, table 4.5b dismisses all ignition sources and apparently concludes the probability of ignition is zero in each case.  Despite virtually no discussion of these tables, the summary at the end of the document contains no actual estimates of risk posed by using HC12a, and simply claims that HC12a "poses very little risk." Such claims have been made consistently by OZ Technology. Unfortunately, this document does not represent a valid risk assessment, and does not support such broad conclusions.  The remainder of the response to this document explains in more detail why the testing of ignition sources performed by OZ Technology was faulty.Source 1.0: Arcing in door switch  The assessment claims that OZ Technology's test 2 demonstrated that this switch is not an ignition source. Even taking OZ Technology's testing at face value, the determination that the door switch is not an ignition source is premature for three reasons. First, OZ only tested a fully functioning door switch, and did not address a faulty switch. A short circuit may cause an arc that could ignite the refrigerant, which is why in the 1991 ADL study, the door switch was assigned a positive probability for ignition. In addition, testing at Underwriters Laboratories (see docket file VID139) showed that even normally functioning door switches are a potential ignition source.  Second, it appears that the concentration of HC12a was above the upper flammability limits during the test. This conclusion is based on the fact that, according to OZ, this test used "a liquid volume of HC12a equivalent to 200 grams of CFC12." According to OZ Technology advertising materials included in the HC12a SNAP petition, systems use 43% as much HC12a as CFC12, which means that 200g of CFC12 is equivalent to 86g of HC12a. In a typical 5 cubic foot freezer, 86g of HC12a will create a concentration of over 25%. Given the components of HC12a, it is not likely that the upper flammability limit of HC12a exceeds 10%. Therefore, the initial concentration of HC12a was substantially higher than the upper flammability limit. However, as refrigerant leaked out of the system, the concentration would have dropped into the flammable range. It is unclear what the results would have been had the test continued to this point.  Finally, a typical refrigerator uses more than 300 grams of CFC12, not 200 grams. Smaller charges are less likely to create flammable concentrations. EPA must consider the effects of using hydrocarbon refrigerants in all systems used in the US, not simply the smallest ones. Sources 2.0, 6.1: Lighted cigarette in mouth; lighted cigarette in room as refrigerator door is opened  The assessment claims that previous testing demonstrated that a cigarette is not an ignition source. OZ's result concurs with the testing by UL mentioned above that was performed on a different refrigerant. In several trials performed by UL, a lit cigarette did not cause ignition. Source 2.1: Cigarette being lit as refrigerator door is opened  The assessment claims that test 1 proves this is not an ignition source. However, test 1 actually investigated the potential for automatic defrost heaters to ignite HC12a. Apparently, OZ intended to list test 3 in support of this claim, which is discussed under source 3.0, below. Source 3.0: Gas range top operating while refrigerator door is open  The assessment claims that test 3 demonstrated that a gas range is not an ignition source. Note that a typical refrigerator will contain more than 300g of CFC12, not 200g. Therefore, as in test 2, discussed under source 1.0, this test does not represent typical US refrigerators. From the test report, it appears that OZ conducted only one trial. In contrast, UL conducted several trials similar to this experiment, and in nearly every case, a pilot light flame on a nearby gas range ignited the hydrocarbon refrigerant. Given UL's qualifications and independent status, EPA believes this result to be more accurate. Therefore, it is inappropriate to simply set the probability of a pilot light flame igniting HC12a to zero for purposes of a risk assessment. Since this report did, indeed, assume that such ignition is impossible, it includes no discussion of the consequences of such an event. Source 4.2: Broiler door open as refrigerator door is opened  OZ simply assumes that no one would open the broiler door because the odorant in HC12a would alert the person to ventilate the room. Interestingly, OZ does not indicate that the odorant would prevent the other scenarios from occurring. At any rate, EPA believes this conclusion to be inaccurate. It is not likely that people would associate the smell with a refrigerant leak because refrigerators have not contained odorized refrigerants for several decades. Furthermore, nonflammable CFC12 has been used for this period, and people do not associate a flammability hazard with their refrigerators. Finally, as discussed elsewhere, at least one study has found that the odorant may be absorbed by the lubricating oil. Again, since OZ assumed this ignition source does not exist, there is no further discussion. Source 5.1: Short produced by another appliance  OZ assumes such a short circuit would not ignite HC12a because the height of another appliance would mean any sparking would take place above the leaked refrigerant, which is heavier than air and would pool on the ground. However, appliances have wiring ranging from essentially floor level to several feet above the floor. Therefore, EPA questions the basis for making this assumption. Since OZ assumes this is not an ignition source, there is no further discussion of the risks from short circuits. Unlisted source: Automatic defrost heaters  One additional source not listed in Table 4.5b, but addressed in testing, is an automatic defrost system. Test 1 purported to demonstrate that a defrost heater could not ignite HC12a. This test has some of the same flaws as test 2: a) the lack of concentration monitoring means that there is no way to determine whether the refrigerant was at a flammable concentration or substantially above the upper flammability limit, and b) too little refrigerant was used, since the typical refrigerator in the US uses over 300g of CFC12. In addition, the actual temperature of the heating coils was not measured. Finally, no assessment was conducted of a faulty heater.  In summary, OZ Technology successfully identified several ignition sources in this risk assessment. However, through faulty analysis and testing, the risk assessment assumes that the probability of ignition is zero for each source. Testing by UL has demonstrated that several sources do ignite hydrocarbons, sometimes with spectacular results. The flammability testing was clearly inadequate. EPA therefore concludes that this risk assessment is invalid and incomplete. C.Videotape: Hydrocarbon Refrigerants/Supervised Safety Tests  This tape shows two tests performed in Australia. The first test purported to demonstrate that hydrocarbons would not reach a flammable concentration inside a car's passenger compartment even in the event of a catastrophic leak. The purpose of the second test was unclear.  In the first test, two occupants sat in a car while 300 grams of refrigerant supposed to be a hydrocarbon blend flowed from the air conditioning system through a hose directly into the passenger compartment. One of the occupants held a candle several inches below the ceiling. The second occupant held a hydrocarbon detector several inches from the ceiling. A second gauge was placed on top of the dashboard, immediately under the upper slope of the windshield. The detectors were calibrated to show the percentage of the lower flammability limit of the refrigerant. It appeared that the vents were not on, so there was no air flow into the car.  No ignition occurred, and the maximum concentration was approximately 60% of the lower flammability limit. The lower flammability limit is the minimum percentage of a given gas in air at which ignition will occur in the presence of a flame or a spark. Since the concentration did not reach this limit, it is not surprising that no ignition occurred. The tester concluded that the test demonstrated the safety of using hydrocarbon refrigerants.  This test is flawed for several reasons. First, the detector and the candle were held several inches from the ceiling. Hydrocarbons are heavier than air, and it is likely that the gas simply pooled near the floor of the car. No measurement was made of the hydrocarbon concentration near the floor. Second, the vents were not on, so there was little or no mixing of the air and hydrocarbons. A flammable substance will only ignite when its concentration is between the lower and upper flammability limits, as explained above. Hydrocarbon gas pooled near the floor, therefore, may not ignite, while wellmixed gas could. The test therefore does not represent what happens during a sudden release of refrigerant from the evaporator into the car when the air conditioning is operating. Air would be blowing out of the vents along with the hydrocarbon refrigerant, and they would be wellmixed. There was no test of what happens when the hydrocarbon is evenly mixed throughout the interior volume of the car. Third, as mentioned elsewhere in this response, the typical American car will contain 400 grams of hydrocarbon refrigerant, and possibly much more, rather than 300 grams. There is no measurement of such a larger charge. Finally, since the composition of the refrigerant is unknown, it is unclear whether its characteristics match that of HC12a. For instance, methane has a lower flammability limit more than twice as high as propane and more than three times as high as butane. Therefore, it would take much less of these gases to reach a flammable composition. For these reasons, the test is of dubious relevance to the question of the safety of using HC12a in unmodified motor vehicle air conditioning systems.  In addition to the flaws in the test itself, it should be noted that at least two other tests have resulted in explosions. In the more recent test, two cans of HC12a were injected into a car and ignited with a spark plug. The resulting explosion blew out the windows and the hatchback glass and caused body damage to the car. This test did not represent a risk assessment, but it demonstrated a scenario that must be included in such an assessment.  The purpose of the second test was unclear. An unspecified amount of HFC134a was released into the passenger compartment after passing over heating coils. The temperature of the coils was not stated. A person sitting in the passenger compartment took samples of the air. The person wore a suit designed to prevent exposure to hazardous substances. Apparently, the test was intended to show that HFC134a degrades to produce toxic byproducts. However, no measurements were presented. In fact, no results were discussed at all.  Prior to the test, the person read a warning label for HFC134a that included several specific cautions: contents under pressure, HFC134a can cause oxygen deprivation, and misuse can cause death. None of these warnings is unusual, and in fact all three apply equally well to hydrocarbons and every other refrigerant, including CFC12 and HC12a. All refrigerants are pressurized, and suddenly puncturing a can could cause injury. All refrigerants, when released in large quantities, can reduce the concentration of oxygen in the air and cause asphyxiation. All refrigerants can cause injury or death if used improperly. For instance, spraying refrigerant directly onto exposed skin can freeze it because all refrigerants emerge from a pressurized container at very cold temperatures. This characteristic is, in fact, what allows air conditioning systems to work. It is the release of pressure as the refrigerant travels from the compressor through the expansion valve to the evaporator that allows the refrigerant to evaporate at cold temperatures and therefore absorb heat from the air in the car. All refrigerants must be handled carefully, and all refrigerants carry warnings. These potential hazards are well understood, and were fully considered in making the acceptability determination for HFC134a. In contrast to these common refrigerant hazards, however, the label on a container of HC12a states "Warning. Contents under pressure. Extremely flammable." HC12a's flammability poses a risk that HFC12a does not. D. Greenpeace: A Chemical Disaster, Why HFCs Have No Future  This supplement included two documents. The first is the last page of an undated letter to an unspecified recipient from Michael Russell of "Ozone Campaign." The second is a report titled "A Chemical Disaster," written by Greenpeace and dated September 1994. Because the letter is incomplete, EPA cannot adequately respond to it.  "A Chemical Disaster" claims that several environmental problems associated with HFCs render them inappropriate for use as CFC substitutes. These issues include global warming, toxic breakdown products when released to the atmosphere, and manufacturing hazards. Finally, the report supports the use of hydrocarbon refrigerants.  It is true that HFC134a has a 100year Global Warming Potential (GWP) of 1300, while that for most hydrocarbons is small. However, the GWP of several other alternatives is similar to that of HFC134a; all are much lower than CFC12's 100year GWP of 8,500. While GWP is one factor to consider, EPA must balance the risks associated with available refrigerants. At this time, EPA believes the potential flammability risk posed by using HC12a in unmodified systems not designed to use a flammable refrigerant to be a more significant risk than the unknown effects of global warming that will be spread over large areas and lengthy amount of time.  When HFC134a is released to the atmosphere, it eventually breaks down. The report's discussion of breakdown products lacks detail, however. While mentioning several potential problematic chemicals, the report repeatedly uses qualifiers like "may" or "could." In addition, the report makes no estimate as to actual damage caused by such products. If more detailed information is available, EPA will consider it. However, EPA does not have evidence that indicates that atmospheric breakdown products from HFC134a pose a significant risk to ecosystems.  The report claims that HFC production results in the production of a significant amount of toxic wastes. However, it again neglects to mention specific details as to volumes or specific waste products. It also does not address how such wastes compare to the environmental impact of manufacturing other alternatives, including hydrocarbons. In addition, it does not reflect the controls placed on such wastes through other regulations, such as the Toxic Substances Control Act or the Resource Conservation and Recovery Act. The report does not provide sufficient evidence that production of HFC134a is a serious environmental hazard.  The report also discusses the explosion at a HFC134a plant. This explosion is discussed under article 6 of Exhibit A: Relative Risks of HC12a and HFC134a.  Finally, the report supports the use of hydrocarbons in newly designed equipment. As stated numerous times in this response and in other documents, EPA believes hydrocarbons may be viable alternative refrigerants for new systems. Even for such systems, however, significant design changes may be required to safely use flammable refrigerants such as hydrocarbons. Furthermore, since no auto manufacturer uses hydrocarbon refrigerants, OZ Technology advertises and sells HC12a for use in existing systems with no modification. This report does not address the potential risks of such use, and therefore does not support the use of HC12a as a CFC12 substitute in existing cars and trucks.