Significant New Alternatives Policy Program 
Refrigeration and Air Conditioning Sector
Risk Screen on Substitutes for CFC-12, HCFC-22 and R-502 in Vending Machines
                          Substitute: Propane (R-290)
This risk screen is restricted to vending machines meeting the requirements of UL 541: Refrigerated Vending Machines.
                                       
This risk screen does not contain Clean Air Act (CAA) Confidential Business Information (CBI) and, therefore, may be disclosed to the public.
1. 	INTRODUCTION
Ozone-depleting substances (ODS) are being phased out of production in response to a series of diplomatic and legislative efforts that have taken place over the past two decades, including the Montreal Protocol and the Clean Air Act Amendments of 1990 (CAAA).  The U.S. Environmental Protection Agency (EPA), as authorized by Section 612 of the CAAA, administers the Significant New Alternatives Policy (SNAP) Program, which identifies acceptable and unacceptable substitutes for ODS in specific end-uses based on assessment of their health and environmental impacts.  
EPA's decision on the acceptability of a substitute is based on the findings of a screening assessment of potential human health and environmental risks posed by the substitute in specific applications.  EPA has already screened a large number of substitutes in many end-use applications within all of the major ODS-using sectors including: refrigeration and air conditioning, solvent cleaning, foam blowing, aerosols, fire suppression, adhesives, coatings and inks, and sterilization. The results of these risk screens are presented in a series of Background Documents that are available in EPA's docket.
The purpose of this risk screen is to supplement EPA's Background Document on the refrigeration and air conditioning sector (EPA 1994) (hereinafter referred to as the Background Document). This risk screen evaluates the potential use of propane as a substitute for CFC-12, HCFC-22, and R-502 in the vending machine end-use. Table 1 presents the composition of the proposed substitute, including the estimated concentrations of impurities, which may be present in the proposed substitute. 
                       Table 1.  Composition of Propane
                                   Component
                               Chemical Formula
                                  CAS Number
                                 Concentration
                               (Weight Percent)
                              Proposed Substitute
                                    Propane
                                     C3H8
                                    74-98-6
                                     99.5%
                 Potential Impurities (maximum concentration)
                                    Ethane
                                     C6H14
                                    74-84-0
                                     0.17%
                                   Propylene
                                     C3H6
                                   115-07-1
                                     0.03%
                                   n-Butane
                                     C4H10
                                   106-97-8
                                     0.1%
                                    Sulfur
                                       S
                                   7704-34-9
                                    0.001%
                                     Water
                                      H2O
                                   7732-18-5
                                    0.0012%
          
Section 2 summarizes the results of the risk screen for the proposed substitute listed in Table 1. The remainder of the risk screen is organized into the following sections:

         * Section 3: Atmospheric Assessment
         * Section 4: Volatile Organic Compound Assessment 
         * Section 5: Discussion of End-Use Scenario Modeled
         * Section 6: Potential Health Effects 
         * Section 7: Flammability Assessment
         * Section 8: Asphyxiation Assessment
         * Section 9: End-Use Exposure Assessment 
         * Section 10: Occupational Exposure Assessment 
         * Section 11: General Population Exposure Assessment
         * Section 12: References
2.	 SUMMARY OF RESULTS						
Propane is recommended for SNAP approval for vending machines that comply with Underwriters Laboratory (UL) Standard 541: Refrigerated Vending Machines. EPA's risk screen indicates that the use of the proposed substitute will be less harmful to the atmosphere than the continued use of CFC-12, HCFC-22, and R-502 as it is less harmful to the ozone layer, has lower climate impact, and a shorter atmospheric lifetime. Propane is subject to volatile organic compound (VOC) regulations under the CAA (40 CFR 51.100(s)). Based on analysis of potential impacts of hydrocarbon refrigerant emissions on ground-level ozone concentrations, VOC emissions from the use of propane in vending machines are not anticipated to contribute significantly to ground level ozone concentrations in the United States. No significant asphyxiation risks to consumers or toxicity risks to workers, consumers, or the general population are expected. It is expected that procedures identified in the material data safety sheet (MSDS) for propane and good manufacturing practices will be adhered to, and that the appropriate safety and personal protective equipment (PPE) (e.g., protective gloves, tightly sealed goggles, protective work clothing, and suitable respiratory protection in case of leakage or insufficient ventilation) consistent with Occupational Safety and Health Administration (OSHA) guidelines will be used during manufacture, installation and servicing, and disposal of vending machines using propane. Because vending machines are to be installed in locations with adequate space and/or ventilation in accordance with EPA recommendations and the equipment maintenance manual for propane, as discussed in greater detail in Section 9, significant toxicity or flammability risk to consumers is also unlikely. Additional safeguards, including specified refrigerant concentration limits (RCL) for propane, are also provided by adherence to industry standards including American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE) Standards 15, 34, and 62 and Underwriters Laboratory (UL) Standard 541, Supplement SA.[4] 
3. 	ATMOSPHERIC ASSESSMENT
This section presents an assessment of the potential risks to atmospheric integrity posed by the use of propane in vending machines.  The ozone depletion potential (ODP), global warming potential (GWP), and atmospheric lifetime (ALT) of the proposed substitute are presented in Table 2.
The proposed substitute is substantially less harmful to the ozone layer, has lower climate impact, and a shorter atmospheric lifetime when compared to CFC-12, HCFC-22, and R-502. In addition, propane also has lower climate impact and a shorter atmospheric lifetime than those predicted for other substitutes examined in the Background Document, as well as commonly utilized substitutes R-404A and HFC-134a. Thus, EPA believes that the use of propane would result in substantially less harm to the climate and ozone layer than the continued use of ODS and commonly used ODS substitutes. 
Table 2.  Atmospheric Impacts of Propane Compared to Other Vending Machine Refrigerants
                                  Refrigerant
                                    ODP[a]
                                    GWP[b]
                                 ALT (Years) 
                                    Propane
                                       0
                                      3.3
                                    0.03[a]
                                    CFC-12
                                     0.82
                                    10,900
                                    100[b]
                                   R-502[c]
                                      0.2
                                     4,650
                                     NA[d]
                                    HCFC-22
                                     0.05
                                     1,810
                                     12[b]
                                   R-404A[e]
                                       0
                                     3,920
                                     NA[f]
                                   HFC-134a
                                       0
                                     1,430
                                     14[b]
NA= Not Available.
a WMO 2010 Scientific Assessment Report (2011).
[b] IPCC 4th Assessment Report (Forster et al. 2007).
[c] R-502 is a blend consisting of HCFC-22 (49%) and CFC-115 (51%). 
d Atmospheric lifetimes are not given for blends, because the components separate in the atmosphere. The ALT for HCFC-22 is 12 years and the ALT for CFC-115 is 1,700 years (IPCC 4th Assessment Report [Forster et al. 2007]).
[e] R-404A is a blend consisting of HFC-143a (52%), HFC-125 (44%) and HFC-134a (4%). 
[f] Atmospheric lifetimes are not given for blends, because the components separate in the atmosphere. The ALT for HFC-143a is 52 years, the ALT for HFC-125 is 29 years, and the ALT for HFC-134a is 14 years.
4.	VOLATILE ORGANIC COMPOUND ASSESSMENT
Propane is regulated as a VOC under CAA regulations (40 CFR 51.100 [s]). Through regulations and standard industry practices, VOC emissions should be controlled.  A separate analysis was prepared by EPA (2014) to evaluate the potential impact of the use of hydrocarbon refrigerants on ground-level ozone concentrations in the United States. The analysis estimated refrigerant emissions from refrigeration and air conditioning equipment, which were assumed to contain propylene, isobutane, and/or propane  under different scenarios. Under the most conservative scenario, it was assumed that propylene was used in all refrigeration and air conditioning equipment. In the most realistic scenario, all three hydrocarbons were assumed to be used in certain types of refrigeration and air conditioning equipment, depending on the proposed use of each alternative under submissions received by the SNAP Program at the time of the analysis. End-uses in the evaluation included a number of types of smaller, self-contained refrigeration and room air-conditioning units for which SNAP hydrocarbon applications have been received and/or UL Standards covering flammable refrigerants exist.  The hydrocarbon emissions from these scenarios were estimated based on U.S. EPA's Vintaging Model, and their potential contributions to ozone concentrations were assessed using U.S. EPA's Community Multiscale Air Quality (CMAQ) model. 

CMAQ modeling was performed for April through the end of September, as these months presented the largest releases of hydrocarbon refrigerant as well as weather conditions favorable for ozone formation. The ozone concentrations were estimated for the Atlanta, Houston and Los Angeles regions, due to their distinctive geographic setting and chronic high levels of ground level ozone, and then scaled for national emission estimates. The results of the CMAQ modeling indicated that under the most realistic scenario, hydrocarbon refrigerants could potentially increase the maximum 8-hour average ground level ozone by no more than 0.15 ppb in Los Angeles, the city with the greatest ozone problem. This is roughly 0.2 percent of the current National Ambient Air Quality Standard (NAAQS) for ozone of 75 ppb. In the most conservative case, where it was assumed that the most reactive hydrocarbon, propylene, was used in all refrigeration and air conditioning equipment, there could be an incremental increase of the 8-hour average as high as 6.61 ppb ozone, or an increase of up to 9 percent of the NAAQS. However, this upper bound level of increase is not likely, as most ozone nonattainment areas are not VOC-limited (i.e., the formation of ozone in these areas are not limited by VOC emissions, but by other compounds such as nitrogen oxide [NOx]). In addition, the analysis assumed no use of VOC-exempt refrigerants which may be used in the refrigeration and air conditioning end-uses. Based on the results of this analysis, VOC emissions from the use of propane in vending machines are not anticipated to contribute significantly to ground-level ozone concentrations in the United States.
5.	DISCUSSION OF END-USE SCENARIOS MODELED
Propane has been proposed for the vending machine refrigeration end-use. EPA expects that propane vending machines will have a charge size of less than 150 grams, consistent with UL Standard 541: Refrigerated Vending Machines, which states that vending machines containing flammable refrigerants may have a charge size of up to 150 grams of A3 refrigerant. 
To represent a reasonable worst-case scenario, it is assumed that a catastrophic leak of refrigerant occurs while the vending machine is installed at the end-use. Because vending machines can be installed in a wide range of locations with varying room volumes, the analysis in this risk screen conservatively assumes that the vending machine contains the maximum charge size indicated by UL 541 (150 grams), and is located in the minimum room size, an enclosed alcove of a commercial building that can be isolated by a door. Under the worst-case scenario, the alcove is assumed to have an effective floor area of 6.15 m[2] (i.e., excluding the space filled by the vending machine, furniture, boxes, etc.) and a height of 2.4 meters), equivalent to a volume of 15 m[3] (530 ft[3]). Under the worst-case scenario, the full charge of the unit is assumed to be emitted over the course of one minute, into the enclosed alcove with 0.5 air changes per hour (ACH).  A vertical concentration gradient is also assumed in the worst-case scenario since propane is denser than air (specific gravity of propane relative to air is 1.52 [air = 1]) and will settle in higher concentrations closer to the ground.  In order to simulate the vertical concentration gradient, it is assumed that 95 percent of the leaked refrigerant mixes evenly into the bottom 0.4 meter of the room, and the rest of the refrigerant mixes evenly in the remaining volume (Kataoka 2000). 
To represent a more typical scenario, this risk screen also evaluates potential risk associated with less conservative conditions. Based on ICF expert opinion, the enclosed alcove is assumed to have an effective floor area of 10m[2] (108ft[3]) with a height of 2.7 m (9 ft), equivalent to a volume of 27.5 m[3] (972 ft[3]), with an air exchange rate of 1.5 ACH (EPA 1997). Additionally, because the typical breathing zone is above 0.4 meters (1.3 feet) and mixing is likely to occur upon release of propane from the unit because the system pressure is higher than the surrounding atmospheric pressure, a vertical concentration gradient is not assumed. Table 3 details the end-use modeling assumptions used throughout the risk screen (i.e., in Sections 7, 8, and 9).
                  Table 3. End-Use Scenario Model Assumptions
Parameter
                                  Assumptions

                        Reasonable Worst-Case Scenario 
                               Typical Scenario
Room Type
                                Enclosed Alcove
                                Enclosed Alcove
 Effective Size (m[3])[a]
                                15 (530 ft[3])
                               27.5 (972 ft[3])
 Ventilation Rate (air changes per hour)
                                      0.5
                                      1.5
Refrigeration Unit
                                Vending Machine
                                Vending Machine
 Charge Size (g)
                                      150
                                      150
 Length of Release (minutes)
                                       1
                                       1
 Vertical Concentration Gradient
                                      Yes
                                      No
     a The volume of the room minus the volume occupied by the vending machine or other large pieces of equipment or furniture present.
   
EPA recognizes that while the majority of vending machines may be placed outdoors or in other well-ventilated areas, which would largely minimize the risks of exposure, there may a few instances in which the consumer chooses to place multiple vending machines in an enclosed space (thus further limiting the effective volume of the space) or chooses to place a single vending machine unit in a very small enclosed space (with the size of the vending machine being the limiting factor). A typical beverage vending machine measures 72 in x 39 in x 33 in, which is equal to 53.6 ft[3] or 1.5 m[3] (Vending Solutions 2009). To address these concerns, this risk screen incorporates threshold analyses in addition to the worst-case and typical scenario modeling. The results from the threshold analyses are used to establish guidelines for choosing an effective room size, charge size, and ventilation rate such that use of propane in vending machines does not present risk to consumers or servicing technicians. 
6.	POTENTIAL HEALTH EFFECTS
To assess potential health risks from exposure to the proposed substitute in vending machines, EPA identified the relevant toxicity threshold values for comparison to modeled exposure concentrations for different scenarios. To protect consumers from the potential dangers of a catastrophic leak from the refrigeration unit, ASHRAE Standard 34 determined refrigerant concentration limits (RCLs) to reduce the risks of acute toxicity, asphyxiation, and flammability hazards in occupied spaces (ASHRAE 2010b). ASHRAE Standard 15 implements the aforementioned standard, requiring that "the concentration of refrigerant in an enclosed space following a complete discharge of a high-probability system shall not exceed the RCL" (ASHRAE 2010a). As such, this risk screen references the RCL in addition to the lower flammability limit, hypoxia No Observed Adverse Effect Level (NOAEL), and exposure limits, as an additional, conservative limit to ensure that significant flammability, asphyxiation, and end-use exposure risks, respectively, do not occur.  
For the occupational exposure analysis, potential risks from chronic and acute worker exposure were evaluated by comparing exposure concentrations to available occupational exposure limits. Potential risks of chronic worker exposure were evaluated using workplace guidance levels (WGL). Risks from potential short-term consumer exposures were evaluated by comparing exposure concentrations to emergency guidance levels (EGL). In the absence of an established short-term exposure limit (STEL), acute exposure guidance levels (AEGLs), or an emergency response planning guideline (ERPG), potential short-term, end-use exposures of some of the impurities were compared to an excursion limit, which was calculated according to the methodology described in ACGIH (2004). Table 4 lists the relevant toxicity limits and RCLs of propane and its impurities, and is followed by Table 5, which provides an explanation of each toxicity limit. EPA's approach for identifying or developing these values is discussed in Chapter 3 of the Background Document. 
         Table 4. Exposure Limits of Propane and Potential Impurities
                                   Component
                         8-hr WGL (Long-term Exposure)
                                      ppm
                           EGL (Short-term Exposure)
                                      ppm
                     Refrigerant Concentration Limit (RCL)
                                      ppm
                                    Propane
                                    1,000a
                                  (OSHA PEL)
                           10,000b,c (10 min AEGL-1)
                            6,900c (30 min AEGL-1)
                                   5,300[d]
Potential Impurities
                                    Ethane
                             1,000[e]  (ACGIH TLV)
                    3,000f (30 min ACGIH Excursion Limit) 
                           2,900[g] (15-min TEEL-1)
                                   7,000[d]
                                   Propylene
                              500[e] (ACGIH TLV)
                     1,500f (30 min ACGIH Excursion Limit)
                                   1,000[d]
                                   n-Butane
                             1,000[e]  (ACGIH TLV)
                           6,900[h]  (30 min AEGL-1)
                                   4,000[d]
[a] OSHA PEL available at: http://www.cdc.gov/Niosh/npg/npgd0524.html
[b] An IDLH of 2,100 ppm has been established for propane.  However, NIOSH (1996) states that "[b]ased on acute inhalation toxicity data in humans (ACGIH 1991; Braker 1980), a value much greater than 10,000 ppm would have been appropriate. However, the revised IDLH for propane is 2,100 ppm based strictly on safety considerations (i.e., being 10% of the lower explosive limit of 2.1%)."  Therefore, as this IDLH value is based on flammability concerns and not toxicity concerns, it is not used in the evaluation of toxicity risks in this risk screen.
[c] AEGL-1 available at: http://www.epa.gov/opptintr/aegl/pubs/tsd96.pdf
[d] ASHRAE (2010b)
[e] ACGIH TLV for aliphatic hydrocarbon gases: alkane (C1 - C4) is 1,000 ppm TWA (ACGIH 2012) 
f Calculated as three times the TLV-TWA (ACGIH 2004).
[g] DOE (2008)
h AEGL-1 available at: http://www.epa.gov/oppt/aegl/pubs/rest102.htm
         Table 5. Explanation of Exposure Limit-Related Terminology[a]
Organization 
Definition
OSHA
Occupational Safety and Health Administration
NIOSH
National Institute for Occupational Safety and Health
ACGIH
American Conference of Governmental Industrial Hygienists
AIHA
American Industrial Hygiene Association
Exposure Limit
Definition
Explanation
Short-Term Exposure
RCL
Refrigerant Concentration Limit
The RCL for a refrigerant is intended to reduce the risks of acute toxicity, asphyxiation, and flammability hazards in normally occupied, enclosed spaces. The RCL for each refrigerant is the lowest of the Acute-Toxicity Exposure Limit (ATEL), Oxygen Deprivation Limit (ODL), and Flammable Concentration Limit (FCL). Determination assumes full vaporization with no removal by ventilation, dissolution, reaction, or decomposition and complete mixing of refrigerant in the space to which it is released.
STEL
Short-Term Exposure Limit
A 15-minute TWA exposure that should not be exceeded at any time during a workday, even if the 8-hour TWA is within the TLV - TWA, set by ACGIH. 
AEGL[b,c]
Acute Exposure Guideline Level 1
AEGL-1 is the airborne concentration of a substance above which it is predicted that the general population, including susceptible individuals, could experience notable discomfort, irritation, or certain asymptomatic nonsensory effects. However, the effects are not disabling and are transient and reversible upon cessation of exposure. 

Acute Exposure Guideline Level 2
AEGL-2 is the airborne concentration (expressed as ppm or mg/m3) of a substance above which it is predicted that the general population, including susceptible individuals, could experience irreversible or other serious, long-lasting adverse health effects or an impaired ability to escape.

Acute Exposure Guideline Level 3
AEGL-3 is the airborne concentration (expressed as ppm or mg/m3) of a substance above which it is predicted that the general population, including susceptible individuals, could experience life-threatening health effects or death.
ERPG
Emergency Response Planning Guideline 1  
The maximum airborne concentration below which it is believed that nearly all individuals could be exposed for up to 1 hour without experiencing other than mild transient health effects or perceiving a clearly defined, objectionable odor according to AIHA. 

Emergency Response Planning Guideline 2  
The maximum airborne concentration below which it is believed that nearly all individuals could be exposed for up to 1 hour without experiencing or developing irreversible or other serious health effects or symptoms which could impair an individual's ability to take protective action according to AIHA.

Emergency Response Planning Guideline 3  
The maximum airborne concentration below which it is believed nearly all individuals could be exposed for up to 1 hour without experiencing or developing life-threatening health effects according to AIHA.
Excursion Limit [d]
Excursion Limit 
Because methane (and other short-chain alkane gases through butane) does not have a TLV-short-term exposure limit, an excursion limit may be applied. Excursions in worker exposure levels may exceed 3 times the TLV-TWA for no more than a total of 30 minutes during a workday, and under no circumstances should they exceed 5 times the TLV-TWA, provided that the TLV-TWA is not exceeded according to ACGIH. 
IDLH
Immediately Dangerous to Life and Health
If exposed to this concentration, room occupants are expected to be able to escape the room within 30 minutes without experiencing escape-impairing or irreversible health effects.
TEEL[e]
Temporary Emergency Exposure Limit 1
TEEL-1 is the maximum concentration in air below which it is believed nearly all individuals could be exposed without experiencing other than mild transient adverse health effects or perceiving a clearly defined, objectionable odor. 

Temporary Emergency Exposure Limit 2
TEEL-2 is the maximum concentration in air below which it is believed nearly all individuals could be exposed without experiencing or developing irreversible or other serious health effects or symptoms that could impair their abilities to take protective action.

Temporary Emergency Exposure Limit 3
TEEL-3 is the maximum concentration in air below which it is believed nearly all individuals could be exposed without experiencing or developing life-threatening health effects. 
Long-Term Exposure
PEL
Permissible Exposure Limit
This is an 8-hour time-weighted average exposure limit set by OSHA. 
TLV-TWA
Threshold Limit Value  -  Time-Weighted Average
The TWA concentration for a conventional 8-hour workday and a 40-hour workweek, to which it is believed that nearly all workers may be repeatedly exposed, day after day, for a working lifetime without adverse effect according to ACGIH. 
[a] All information in this table taken from EPA (1994) except where otherwise noted.
b EPA (2012)
c Applicable to emergency exposure periods ranging from 10 minutes to 8 hours.
[d] ACGIH (2004)
[e] Chemical exposure guidelines to use for emergency planning if no AEGL or ERPG is available. TEELs are developed based on concentration limits or toxicology parameters (DOE 2008).

According to the MSDS, exposure to propane may be hazardous if inhalation, skin contact, or eye contact with propane occurs. The most likely pathway of exposure is through inhalation. Propane can cause symptoms of asphyxiation when present in concentrations high enough to significantly lower oxygen concentrations below 19.5 percent by volume, such as headaches, ringing in ears, dizziness, drowsiness, nausea, vomiting, depression of all senses, and also unconsciousness. Under some circumstances of over-exposure (i.e., oxygen levels fall below 6 percent by volume), death may occur. In addition, at high concentrations, propane can act as a narcotic and cause central nervous system depression, including dizziness, drowsiness, and headaches. 
If propane is inhaled, person(s) should be immediately removed and exposed to fresh air. In accordance with the MSDS, EPA further recommends that if breathing is difficult, person(s) be given oxygen, provided a qualified operator is present, and medical attention be sought.  Rescuers should not attempt to retrieve victims of exposure to propane without adequate PPE. At a minimum, a self-contained breathing apparatus (SCBA) should be worn. Exposures of propane to the skin may cause frostbite. In the case of dermal exposure, the MSDS for propane recommends that person(s) immediately wash the affected area with water and remove all contaminated clothing; if frostbite occurs, bathe (not rub) the affected area with lukewarm, not hot, water. If water is not available, cover the affected area with a clean, soft cloth. Alternatively, if the fingers or hands are frostbitten, warm the affected area by placing it in the armpit; gently exercise the affected part while being warmed, and seek medical attention immediately. Exposures of propane to the eyes could cause eye irritation. In case of ocular exposure, the MSDS for propane recommends that person(s) immediately flush the eyes, including under the eyelids, with copious amounts of water for 15 minutes. 
EPA's review of the human health impacts of this proposed substitute is contained in the public docket for this decision. The potential health effects of propane can be minimized by following the exposure guidelines and ventilation and PPE recommendations outlined in the MSDS for propane and this risk screen.
7.	FLAMMABILITY ASSESSMENT
ASHRAE Standard 34 classifies propane as a Class A3 refrigerant. Propane is flammable when its concentration in air is in the range of 2.1 percent and 9.5 percent by volume (21,000 ppm to 95,000 ppm). In the presence of an ignition source (e.g., static electricity, a spark resulting from a switch malfunction, or a cigarette), an explosion or a fire could occur if the concentration of propane exceeds its lower flammability limit (LFL) of 21,000 ppm, posing a significant safety concern for workers and consumers if it is not handled carefully. The remainder of this section assesses flammability risks and summarizes the recommended measures to ensure safe handling and use of the refrigerant during manufacture, servicing, and end-use. 

7.1	Flammability Risks at Manufacture
EPA expects manufacturing to occur in a closed system and that all units are completely sealed before delivery and installation. If the vending machines are not charged and sealed in a closed system, EPA expects personnel to wear proper PPE (see Section 2) to protect themselves from releases or exposure to propane In addition, manufacturing facilities are anticipated to maintain proper ventilation at all times during the manufacture of equipment containing propane and adhere to good manufacturing practices. As a result, releases of propane during these manufacturing and installation operations in the presence of an ignition source are not anticipated. 
All propane storage and transport equipment should be installed with safety devices that minimize the likelihood of catastrophic releases. For example, NFPA 58 Liquefied Petroleum Gas Code (NFPA 2014) for liquefied propane requires the use of overfill protection devices (OPD) on cylinders to minimize the likelihood of leaks.  The NPFA 58 Code also contains propane storage and transportation requirements/guidelines. Similar equipment safety and procedural requirements should be developed for propane and other flammable refrigerants. EPA believes that because relevant safety standards, the MSDS for propane, and OSHA requirements under 29 CFR 1910 (e.g., proper ventilation and storage practices within manufacturing facilities to prevent fire and explosion) are followed by workers, flammability during manufacture is not a concern.  

7.2	Flammability Risk at Servicing and End-Use
The risk of flammability during servicing and end-use for the reasonable worst-case scenario and typical scenario (see Section 5) was investigated for propane. Both servicing and end-use of propane vending machines are expected to take place in the same room. In order to determine the potential flammability risks during servicing or end-use in case of a catastrophic release of refrigerant in an enclosed alcove (see Section 5), concentrations of propane immediately following a complete release of refrigerant, either accidental or deliberate (e.g., through vandalism or theft), were compared to the LFL for propane. The reasonable worst-case scenario assumes an effective room size of 15 m[3] and air exchange rate of 0.5 ACH. In this worst-case scenario, the modeling results reflect the maximum concentration in the lower 0.4 meters of the room, as the upper portion of the room presents a much lower risk, with only 5 percent of the leaked refrigerant present in this area which also has a greater volume than the lower 0.4 meters of the room. In the typical scenario, an effective room size of 27.5 m[3] and air exchange rate of 1.5 ACH are assumed while the modeling results reflect the maximum concentration in the entire room, as a vertical concentration gradient is not assumed. Table 6 presents the results of the analysis.   

                      Table 6. Flammability Assessment[a]
                                   Scenario
                                Charge Size (g)
                          Effective Room Size (m[3])
                  Vertical Concentration Gradient  Height (m)
                      Maximum Instantaneous Concentration
                                  (ppm) b[,c]
                             Reasonable Worst-Case
                                      150
                                15 (530 ft[3])
                                      0.4
                                    30,900
                                    Typical
                                      150
                               27.5 (972 ft[3])
                                     None
                                     2,950
                      Threshold Analysis 1a: Charge Size
                                      102
                                15 (530 ft[3])
                                      0.4
                                    21,000
                      Threshold Analysis 1b: Charge Size
                                     1,120
                               27.5 (972 ft[3])
                                     None
                                       
                       Threshold Analysis 2a: Room Size
                                      150
                                22 (780 ft[3])
                                      0.4
                                       
                       Threshold Analysis 2b: Room Size
                                      150
                                3.9 (128 ft[3])
                                     None
                                       
 Bold font indicates modeling results.
 [a] Cells highlighted in green are the scenarios with acceptable exposure levels given various modeling assumption options. 
 [b] Lower Flammability Limit of propane is equal to 21,000 ppm.
 [c] Values provided in these columns refer to the concentration in the lower 0.4 meters of the room, which presents the most conservative risk as 95 percent of the leaked refrigerant in present in this space. 
  
In the reasonable worst-case scenario, at a charge size of 150 grams, the maximum instantaneous concentration of propane in the lower 0.4 meters of the room would be approximately 30,900 ppm, which is 150 percent of the LFL for propane (see Table 6). In the typical scenario, the maximum instantaneous concentration of propane in the entire room would be approximately 2,960 ppm, which is 14 percent of the LFL for propane. However, the accidental release of all refrigerant into the enclosed alcove is highly unlikely. This is because the refrigerant system is located in a protected compartment on the back side of the vending machine and accidents resulting in a puncture to the compressor, whether through attempted theft of the vending machine contents or vandalism, would be unlikely. In the case of accidental release, however, propane is likely to be dispersed into the air evenly and almost instantaneously upon release; the vending machine compressor is pressurized, which will cause propane to be released into the room (assumed to be at atmospheric pressure of approximately 14.7 psi) under pressure, resulting in the refrigerant's quick and turbulent release, and subsequent mixing into the air. However, it is also unlikely that 100 percent of the refrigerant charge will be released from the system because as the refrigerant leaks, the system pressure will decrease and eventually equilibrate with atmospheric pressure, leaving a small amount of refrigerant in the system. To completely remove the entire refrigerant charge from the system, a vacuum pump must be used.
Catastrophic releases of large quantities of refrigerant, especially in areas where refrigerant is stored, could cause an explosion. For the flammability threshold analysis, the conditions at which a flammability concern would exist (i.e., when the maximum instantaneous concentration in the lower 0.4 meters of the enclosed alcove equals the LFL for propane) was determined. Under the room size, ventilation rate, and vertical concentration gradient assumed within the reasonable worst-case scenario, the charge size of the propane refrigeration unit would have to be at least 102 grams for a flammability concern to exist, as shown in Threshold Analysis 1a. Under the room size and ventilation rate assumed within the typical scenario, the charge size of the propane refrigeration unit would have to be at least 1,120 grams for a flammability concern to exist, as shown in Threshold Analysis 1b. Using a 150-gram refrigeration unit, as shown under Threshold Analysis 2a and 2b, the volume of the room would have to be smaller than 22 m[3] (780 ft[3]) or 3.9 m[3] (138 ft[3]) for a flammability risk to occur under the reasonable worst-case and typical scenarios, respectively. EPA recommends that these guidelines be followed to further minimize risk of flammability.
According to the results of this flammability analysis, the risk of fire is minimal if a vending machine is installed in a room with volumes in accordance with standards and regulations. According to ASHRAE Standard 34, the RCL for propane (9.5 g/m[3] of propane or 5,300 ppm by volume) should not be exceeded in any room that contains a propane vending machine. EPA recommends that in addition to the RCL, these guidelines be followed to further minimize risk of flammability.  For this reason, it is important that only properly trained and certified technicians handle propane. As a further precaution, certification requirements and training programs for technicians that handle propane should be developed using these guidelines. During servicing operations, technicians should ensure that proper ventilation is in place through the use of fans (or other mechanical ventilation devices) and portable refrigerant detectors should be used to alert technicians to the presence of flammable gases in the area. Vending machines should not be installed in small, poorly ventilated spaces such as very small enclosed areas or storage closets (especially as other equipment or appliances in the space would reduce the effective volume of the room). For vending machines installed in larger areas, the risk of fire and explosion is minimal. Vending machines installed with propane should be clearly labeled as containing a flammable refrigerant and designed to prevent catastrophic leaks.  The installation of leak prevention devices would further protect against the very limited risk of explosion. EPA believes that the risk of fire is minimal if a propane vending machine is installed in a room with volumes and ventilation in accordance with standards, MSDS, and recommendations.  
8.	ASPHYXIATION ASSESSMENT	
The risk of asphyxiation for the reasonable worst-case scenario was investigated for propane. In this section, risk of asphyxiation is assessed in three ways: 1) modeling the oxygen concentration under the charge size and room size specified in the worst-case scenario, 2) performing a threshold analysis of minimum charge size needed to cause an asphyxiation risk in the room size specified in the worst-case scenario 3) performing a threshold analysis of the maximum room size needed to cause an asphyxiation risk with the charge size specified in the worst-case scenario. This analysis does not consider conditions that are likely to occur that would increase oxygen levels to which individuals would be exposed, such as open doors or windows, fans operating, conditioned airflow (either heated or cooled), or even openings at the bottom of doors that allow air to flow in and out. As specified in Section 5, the reasonable worst-case scenario in this analysis assumes a vertical concentration gradient. If the proposed substitute passes the screening analysis with these restrictive assumptions in place, it can be reasonably assumed that no risks of asphyxiation will be present under real-world conditions. The results of the asphyxiation assessment are summarized in Table 7 below.  



                      Table 7. Asphyxiation Assessment[a]
                                   Scenario
                                Charge Size (g)
                          Effective Room Size (m[3])
                  Vertical Concentration Gradient Height (m)
                       Percent Oxygen Concentration [b] 
                             Reasonable Worst-Case
                                      150
                                15 (530 ft[3])
                                      0.4
                                      21
                                    Typical
                                      150
                               27.5 (972 ft[3])
                                     None
                                      21
                      Threshold Analysis 1a: Charge Size
                                     2,710
                                15 (530 ft[3])
                                      0.4
                                      12
                      Threshold Analysis 1b: Charge Size
                                    29,800
                               27.5 (972 ft[3])
                                     None
                                      12
                       Threshold Analysis 2a: Room Size
                                      150
                               0.83 (29.3 ft[3])
                                      0.4
                                      12
                       Threshold Analysis 2a: Room Size
                                      150
                               0.14 (4.9 ft[3])
                                     None
                                      12
Bold font indicates modeling results.
s Cells highlighted in green are the scenarios with acceptable exposure levels given various modeling assumption options.
[b] The typical concentration of oxygen in air is considered to be 21 percent (Mackenzie & Mackenzie 1995).
   
In order for a risk of asphyxiation to occur, the normal concentration of oxygen in air (21 percent) in the lower 0.4 meters of the room must be reduced to 12 percent. Based on both the reasonable worst-case scenario and typical scenario modeling assumptions, propane in vending machines does not present a significant risk of asphyxiation. The concentration of propane in the air following the release of the 150-gram charge size does not exceed 2.4 percent or 0.2 percent in the reasonable worst-case and typical scenarios, respectively, which has an insignificant impact on the normal concentration of oxygen in air. Because this modeling does not take into account any ventilation which is likely to occur, such as conditioned airflow, open doors, or even openings at the bottom of doors that allow air to flow in and out, as mentioned above, the actual asphyxiation risk to personnel is likely to be even smaller than modeled. 
For the asphyxiation threshold analyses shown in Table 7, the conditions at which an asphyxiation concern would exist were determined. As shown in Threshold Analysis 1a, the minimum charge of propane necessary to reduce the oxygen levels to 12 percent in air in the lower 0.4 meters of a room of volume 15 m[3] (530 ft[3]) was calculated to be 2,710 grams, which is more than 18 times the maximum UL 541 charge size of 150 grams. As shown in Threshold Analysis 1b, the minimum charge of propane necessary to reduce the oxygen levels to 12 percent in the air of a room of volume 27.5 m[3] (972 ft[3]) was calculated to be 29,800 grams, which is approximately 200 times the maximum UL 541 charge size of 150 grams. These analyses, Threshold Analysis 1a and 1b, assumed that 1) nitrogen and oxygen retain the same relative volumes in the rooms with the balance composed entirely of propane, and 2) the pressure of the room does not increase significantly with the addition of the refrigerant. As shown in Threshold Analysis 2a, for asphyxiation to be of concern with the proposed maximum charge size of 150 grams, in a 15 m[3] room, and assuming a vertical concentration gradient height of 0.4 meters, the effective volume of the enclosed alcove would have to be about 0.83 m[3] (29.3 ft[3]) (see Table 7), which is smaller than the size of the vending machine itself. As shown in Threshold Analysis 2b, for asphyxiation to be of concern with the proposed maximum charge size of 150 grams and assuming no vertical concentration gradient in a 27.5 m[3] room, the effective volume of the enclosed alcove would have to about 0.14 m[3] (4.9 ft[3]), which is also smaller than the size of the vending machine itself. Based on the worst-case modeling results and threshold analysis, EPA does not believe that the use of propane in vending machines poses a significant risk of asphyxiation or impaired coordination to consumers. 
9.	END-USE EXPOSURE ASSESSMENT	
This section presents estimates of potential consumer exposures to propane and each of its impurities in vending machines. An exposure analysis was performed to examine potential catastrophic releases release of the proposed substitute under the reasonable worst-case scenario and typical scenario outlined in Section 5.  

For the end-use exposure assessment scenario, 30-min TWA exposures for the proposed substitute and each of its impurity constituents were calculated using the box model described in the Background Document, which was adapted to estimate concentrations on a minute-by-minute basis. The analysis was undertaken to determine the 30-minute TWA exposures of the proposed substitute because consumers are anticipated to only spend short amounts of time in the presence of vending machines while purchasing and consuming products. These exposures were then compared to the standard toxicity limits presented in Table 4 to assess the risk to end-users.  However, the TWA values are fairly conservative as the analysis does not consider opened windows, fans operating, conditioned airflow (either heated or cooled) and other variables that would reduce the levels to which individuals would be exposed. Modeling was performed for propane and all of its impurity constituents; however, only the modeling results for propane are presented in Table 8. As propane is approximately 99.5 percent pure, it presents the greatest risk of exceeding the STEL and thus is the most conservative assessment. 

                    Table 8. End-Use Exposure Assessment[a]
                                   Scenario
                                Charge Size (g)
                          Effective Room Size (m[3])
                            Ventilation Rate (ACH)
                  Vertical Concentration Gradient Height (m)
                  30-minute TWA End-Use Exposure[b][,c] (ppm)
                            Reasonable Worst-Case 
                                      150
                                15 (530 ft[3])
                                      0.5
                                      0.4
                                    27,420
                                    Typical
                                      150
                               27.5 (972 ft[3])
                                      1.5
                                     None
                                     2,098
                      Threshold Analysis 1a: Charge Size
                                     37.7 
                                15 (530 ft[3])
                                      0.5
                                      0.4
                                     6,900
                      Threshold Analysis 1b: Charge Size
                                      493
                               27.5 (972 ft[3])
                                      1.5
                                     None
                                       
                       Threshold Analysis 2a: Room Size
                                      150
                               60 (2,120 ft[3])
                                      0.5
                                      0.4
                                       
                       Threshold Analysis 2b: Room Size
                                      150
                                8.4 (297 ft[3])
                                      1.5
                                     None
                                       
                    Threshold Analysis 3a: Ventilation Rate
                                      150
                                15 (530 ft[3])
                                      8.9
                                      0.4
                                       
                    Threshold Analysis 3b: Ventilation Rate
                                      150
                               27.5 (972 ft[3])
                                       0
                                     None
                                       
Bold font indicates modeling results. 
a Cells highlighted in green are the scenarios with acceptable exposure levels given various modeling assumption options.
[b] 30-min AEGL-1 STEL for propane is 6,900 ppm; See Table 4 for more information.
[c] Values provided in this column refer to the concentration in the lower 0.4 meters of the room, which presents the most conservative risk as 95 percent of the leaked refrigerant in present in this space. 

In the reasonable worst-case scenario and typical scenario described in Section 5, catastrophic release of propane from a vending machine with a charge size of 150 grams was modeled. During the release event in the typical scenario, the modeling results indicate the maximum anticipated 30-min TWA end-use exposure does not exceed 30 percent of the 30-min AEGL-1 for propane. Under the plausible worst-case scenario, however, the 30-min AEGL-1 for propane of 6,900 ppm is likely to be exceeded (see Table 8). Additionally, the calculated 30-min TWA exposure also exceeds the AEGL-2 (17,000 ppm) for propane, but not the AEGL-3 (33,000 ppm). However, such consumer exposure is unlikely, as the reasonable worst-case scenario (see Section 5) assumes that 95 percent of the refrigerant is leaked into the lower 0.4 meters of the room, where it is unlikely that persons would be contained. In addition, these modeled exposures were derived using conservative assumptions (e.g., minimum room volume, low ventilation rate [0.5 ACH], maximum charge size) that do not necessarily reflect the typical location for propane vending machine installations (e.g., vending machines will be installed in commercial buildings, which have higher ventilation rates). 
Nonetheless, these potential exposures emphasize the importance associated with proper equipment installation (i.e. matching charge size with room dimensions). A threshold analysis was performed to determine the charge size, minimum room volume, and ventilation rates where vending machines containing propane could be installed to avoid the short-term exposure limit for propane. Under the worst-case scenario in Threshold Analysis 1a, the maximum charge size that should be installed in a room of 15 m[3] (530 ft[3]) volume is 37.7 grams. Under the typical scenario in Threshold Analysis 1b, the maximum charge size that should be installed in a room of 27.5 m[3] (972 ft[3]) volume is 493 grams. If a charge size of 150 grams is used in vending machines, the vending machines should be installed in larger spaces with higher ventilation rates (i.e., assuming the conditions in the reasonable worst-case scenario, vending machines should be installed in rooms that are at least 60 m[3] [2,120 ft[3]] at a ventilation rate of 0.5 ACH or vending machines installed in a space of 15 m[3] [530 ft[3]] must have a ventilation rate of at least 8.9 ACH [see Threshold Analysis 2a and 3a]). Assuming more conservative conditions in the typical scenario in Threshold Analysis 2b and 3b, vending machines should be installed in rooms that are at least 8.4 m[3] (297 ft[3]) at a ventilation rate of 1.5 ACH while vending machines installed in a large space of 27.5 m[3] (972 ft[3]) do not require ventilation. 
Because end-use exposure is only a concern under the most conservative scenario, EPA recommends these minimum room and ventilation rate guidelines to ensure that if a catastrophic leak of propane occurs, it is not anticipated to result in a significant risk to consumers at end-use. To prevent exposure and potential serious side effects during larger releases, EPA recommends that the charge size for a propane vending machine does not exceed the RCL (9.5 g/m[3]) and proposes to require compliance with UL 541 in any space containing a vending machine, unless proper leak protection devises are in place in order to prevent exposures of propane beyond the recommended limits. Additionally, these room volume and ventilation rate requirements simultaneously ensure that there is not a significant flammability or asphyxiation risk, as these requirements are more restrictive than the room sizes recommended in the flammability and asphyxiation assessments (see Sections 7 and 8). Proper leak protection devices and engineering control requirements and adherence to the MSDS will further ensure that exposure limits are maintained below those described in Table 8.
10.	OCCUPATIONAL EXPOSURE ASSESSMENT
This section assesses potential exposures to workers during manufacture, installation, servicing, and disposal of propane. EPA expects that the manufacture of propane refrigerant (i.e., formulation mixing) and the charging of propane vending machines (i.e., receiving, blending. and filling operations) occurs in a closed system; all units are completely sealed before delivery and installation. As a result, exposure during these manufacturing and installation operations is not anticipated. 
To ensure that use of the proposed substitute in vending machines does not pose an unacceptable risk to workers during servicing and disposal, occupational exposure modeling was performed using a box-model approach.  For a detailed description of the methodology used for this screening assessment, the reader is referred the occupational exposure and hazard analysis described in Chapter 5 of the Background Document. Estimates of refrigerant release per event for various release scenarios and data on number of events were obtained from the Vintaging Model. For the vending machine end use, the release per event was conservatively assumed to be 0.25 percent of the equipment charge during servicing and 35 percent of the equipment charge during disposal. Additionally, EPA is proposing to allow intentional venting or release of propane as a refrigerant during installation, servicing, maintenance and disposal from window AC units. If venting does occur, it should be done in a well-ventilated area (e.g., outdoors). Therefore, scenarios in which 100 percent of the equipment charge was released during servicing and disposal were also considered. The release rate per event was multiplied by the number of events estimated to occur over a workday.  For equipment servicing, the number of events per workday was assumed to equal the maximum number of units anticipated to be serviced in one day (eight units) divided by eight hours per workday.  These assumptions result in approximately 8 events per day.  For disposal, it was conservatively assumed that 10 units are disposed during an 8-hour work day. The modeled exposure concentrations were compared to short-term occupational exposure limits at installation and servicing and long-term exposure limits at disposal.

10.1	Occupational Exposure at Servicing
The potential occupational exposure during servicing was analyzed. The maximum 30-minute TWA exposures for propane and its impurities were estimated for the servicing exposure scenario and compared to their respective STELs. As summarized in Table 9, the modeling results indicate that the maximum anticipated occupational exposure concentrations at no point exceed 0.01 percent and 4.3 percent of the STEL for propane for vending machines with a 150-gram charge size for the 0.25 percent release and 100 percent release of refrigerant during servicing, respectively.  Because the exposure concentrations modeled under the conservative assumptions is significantly lower than the exposure limits for each constituent, occupational exposure to the proposed substitute during servicing is not considered a significant risk to workers. Further, the estimated exposures were derived using conservative assumptions, and represent a worst-case scenario with a low probability of occurrence. These types of systems are typically serviced by trained personnel using proper industrial hygiene techniques.  
     
             Table 9. Occupational Risk Assessment at Servicing[a]
                                   Component
                   30-minute TWA Occupational Exposure (ppm)
                            Charge size: 150 grams
                    30-min Short Term Exposure Limits (ppm)
                                       
                                 0.25% Release
                                 100% Release
                                       
                                    Propane
                                     0.74
                                      297
                                   6,900[b]
Potential Impurities
                                    Ethane
                                     0.002
                                     0.74
                                   3,000[c]
                                   Propylene
                                  2 x 10[-4] 
                                     0.09
                                   1,500[c]
                                   n-Butane
                                     0.001
                                     0.23
                                   6,900[b] 
   a Cells highlighted in green are the scenarios with acceptable exposure levels given various modeling assumption options.
	[b] AEGL-1
	[c] ACGIH (2004)
              
10.2	Occupational Exposure at Disposal
Disposal of propane vending machines is expected to occur with limited frequency (up to approximately one disposal event per day) at disposal facilities and with limited duration of exposure to the refrigerant charge. Potential exposures to the refrigerant during recovery and disposal are expected to occur during activities related to draining the refrigerant from the vending machine into cylinders (e.g., connecting of pipes). Such activities and related exposure is anticipated to occur within 15-30 minutes (per event/day).
Table 10 displays the maximum estimated 8-hour TWA occupational exposure levels of propane and its impurities during disposal. Based on the assumptions described in the beginning of Section 10, the modeling indicates that 8-hour worker exposure concentrations for propane and each of its impurities during the 35 percent release scenario and the 100 percent release scenario will at no point exceed 19 percent and 53 percent of the long-term exposure limits of any of the individual components during disposal of propane for vending machines with a 150-gram charge size. Table 10 displays the maximum estimated 8-hour TWA occupational exposure levels of propane and its potential impurities during disposal. Because each value is significantly lower than the exposure guidelines for each constituent, occupational exposure to the proposed substitute during disposal does not represent a true occupational health hazard.  Further, the estimated exposures were derived using conservative assumptions, and represent a worst-case scenario with a low probability of occurrence. These types of systems are typically disposed of by trained personnel using proper industrial hygiene techniques.  
             Table 10. Occupational Risk Assessment at Disposal[a]
                                   Component
                    8-Hour TWA Occupational Exposure (ppm)
                            Charge size: 150 grams
                    8-Hour Long Term Exposure Limits (ppm)
                                       
                                  35% Release
                                 100% Release
                                       
                                    Propane
                                      184
                                      526
                                    1,000b
Potential Impurities 
                                    Ethane
                                     0.46
                                     1.32
                                   1,000[c]
                                    Propene
                                     0.058
                                     0.17
                                    500[c]
                                   n-Butane
                                     0.14
                                     0.40
                                   1,000[c] 
        a Cells highlighted in green are the scenarios with exposure levels deemed to be acceptable given various modeling assumption options.
	   [b] OSHA PEL
          [c] ACGIH TLV 
       
Although anticipated occupational exposures are well below the exposure limits for each of the components of propane, the recommendations for proper engineering controls and PPE (e.g., safety glasses and gloves) in the MSDS for propane should be followed. Adequate ventilation should always be established during any use, handling, or storage of propane. Engineering controls should include vapor-in air detection systems and local exhaust ventilation during use of propane to prevent dispersion of propane throughout the work place. In addition, an eye wash and safety shower should be near the manufacturing facility and locations where propane is stored and ready for use. EPA believes that if proper handling and disposal guidelines are followed in accordance with good industrial hygiene and manufacturing practices and the MSDS for propane, there is no significant risk to workers during the manufacturing, installation, servicing, and disposal of propane in vending machines.
11.  	GENERAL POPULATION EXPOSURE ASSESSMENT
Propane is not expected to cause a concern for human health in the general population when manufactured for use and used as a refrigerant in vending machines. The proposed substitute will be manufactured in a closed process and is proposed for use in closed systems, and thus, significant releases are not anticipated. At room temperature, propane is a gas and, therefore, releases to ground or surface water are not anticipated, as propane is anticipated to dissipate into the atmosphere upon release to outside air (i.e., because natural ventilation rates would be higher and there is no enclosed space to keep propane concentrated). Should air releases during manufacturing operations occur, engineering controls should be used (e.g. carbon absorption scrubbers) to collect propane and prevent the release of propane to the atmosphere. EPA believes that by using proper engineering controls and by following disposal and containment recommendations outlined in the proposed substitute's MSDS, exposure to propane is not expected to pose a significant toxicity risk to the general population.
12.  	REFERENCES

ACGIH. 1991.  Propane. In: Documentation of the threshold limit values and biological exposure indices.
6th ed. Cincinnati, OH: American Conference of Governmental Industrial Hygienists, pp. 1286 - 1287. 

ACGIH. 2004. The Documentation of the Threshold Limit Values and Biological Exposure Indices. American Conference of Governmental Industrial Hygienists. Available online at: www.acgih.org. 

ACGIH. 2012. Guide to Occupational Exposure Values. American Conference of Governmental Industrial Hygienists.

ANSI/ASHRAE. 2010a. Standard 15: Safety Standard for Refrigeration Systems. American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.

ANSI/ASHRAE. 2010b. Standard 34: Designation and Safety Classification of Refrigerants. American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. 
ANSI/ASHRAE. 2013a. Standard 62: Ventilation for Acceptable Indoor Air Quality. American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.
Braker W, Mossman AL. 1980. Matheson gas data book. 6th ed. Secaucus, NJ: Matheson Gas Products, pp. 615 - 623. 
	
Forster, P., V. Ramaswamy, P. Artaxo, T. Berntsen, R. Betts, D.W. Fahey, J. Haywood, J. Lean, D.C. Lowe, G. Myhre, J. Nganga, R. Prinn, G. Raga, M. Schulz and R. Van Dorland. 2007.  Changes in Atmospheric Constituents and in Radiative Forcing. In: Climate Change 2007:The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M.Tignor and H.L. Miller (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.
ICF. 1997. Physiological Effects of Alternative Fire Protection Agents - Hypoxic Atmospheres Conference. Stephanie Skaggs prepared the proceedings of the conference held May 22, 1997 in New London, CT.

IPCC/TEAP. 2005. Bert Metz, Lambert Kuijpers, Susan Solomon, Stephen O. Andersen, Ogunlade Davidson, José Pons, David de Jager, Tahl Kestin, Martin Manning, and Leo Meyer (Eds).  Safeguarding the Ozone Layer and the Global Climate System.  Intergovernmental Panel on climate Change Technology and Economic Assessment Panel. Cambridge University Press, UK. pp 478.
Kataoka, O., M. Yoshizawa, & T. Hirakawa  2000.  "Allowable Charge Limit of Flammable Refrigerants and Ventilation Requirements."  Daikin Industries. International Refrigeration and Air Conditioning Conference. Paper 506. Available online at: http://docs.lib.purdue.edu/iracc/506.

Mackenzie, F.T. and J.A. Mackenzie. 1995. Our changing planet. Prentice-Hall.

NFPA. 2014.  NFPA Liquified Petroleum Gas Code. 2014 Edition.	

NIOSH. 1996.  Propane: IDLH Documentation.  August 1996.  Accessed 17 Feburary 2009.  Available online at: < http://www.cdc.gov/niosh/idlh/74986.html>.

OSHA. 2004.  "Safety and Health Topics: Isobutane."  February 2004.  Available online at: http://www.osha.gov/dts/chemicalsampling/data/CH_247840.html. 

UL. 2011. UL 541: Standard for Refrigerated Vending Machines. Underwriters Laboratory.

U.S. Department of Energy (DOE). 2008. DOE Handbook  -  Temporary Emergency Exposure Limits For Chemicals: Methods And Practice. Available online at: http://orise.orau.gov/emi/scapa/files/doe-hdbk-1046-2008_ac.pdf.

U.S. EPA 1994.  Significant New Alternatives Policy Technical Background Document:  Risk Screen on the Use of Substitutes for Class I Ozone-depleting Substances: Refrigeration and Air Conditioning.  Stratospheric Protection Division.  March, 1994.

U.S. EPA. 1997. Exposure Factors Handbook. National Center for Environmental Assessment, Office of Research and Development. Available online at: http://www.epa.gov/ncea/pdfs/efh/front.pdf.

U.S. EPA. 2012. Acute Exposure Guidelines (AEGLs) Definitions. Available online at: http://www.epa.gov/oppt/aegl/pubs/define.htm.

U.S. EPA. 2014. Draft Assessment of the Potential Impact of Hydrocarbon Refrigerants on Ground Level Ozone Concentrations. Prepared for the U.S. EPA Stratospheric Protection Division by ICF International. February 24, 2014.

Vending Solutions. 2009. "Soda Machines." Available online at:  http://www.vendingsolutions.com/vending-machines/. 

WMO (World Meteorological Organization). 2011. Scientific Assessment of Ozone Depletion: 2010, Global Ozone Research and Monitoring Project -- Report No. 52, 516 pp., Geneva, Switzerland, 2011.

