EPA-600/2-89-009
February 1989
EVALUATION OF REFRIGERANT FROM MOBILE AIR CONDITIONERS
By:
Leo Weilzman
Acurex Corporation
Energy and Environmental Division
Southeasl Regional Office
Post Office Box 13109
Research Triangle PatK North Carolina 27709
EPA Contract 68-02-4285
Work Assignment No. 1/006
EPA Project Officer: Dale L. Harmon
U.S. Environmental Protection Agency
Air and Energy Engineering Research Laboratory
Research Triangle Park, North Carolina 27711
AIR AND ENERGY ENGINEERING RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NC 27711
-------
ACKNOWLEDGEMENT
This document was prepared for the U.S. Environmental Protection Agency by Acurex
Corporation's Environmental Systems Division. The Principal Investigator was Leo Weitzman and he was
assisted by Larry Hamel, both of Acurex. The EPA Project Officer was Dale Harmon. Guidance was
provided by Stephen O. Andersen, Chief of the Technology and Economics Branch, and Jean Lupinacci,
Economist, of EPA's Office of Air and Radiation.
We express our appreciation to the members of the ad hoc industry group comprised of
representatives of the following:
Mobile Air Conditioning Society (MACS)
Motor Vehicle Manufacturers Association of the United States(MVMA)
Manufacturers of Small Recovery/Reclamation/Recycle Devices
Society of Automotive Engineers (SAE)
American Society of Heating, Refrigerating, and Air Conditioning Engineers (ASHRAE)
Small air conditioning service shops
Other industry representatives and individuals
A special thanks is due to Jack Loos, Consultant, who helped set up the sampling program and to
the following members of the ad hoc working group who were especially helpful in setting up the program
and conducting the sampling:
Co-Chairs: Robert Bishop, Harrison Division, General Motors
Simon Oulouhojian, Mobile Air Conditioning Society
Technical Advisors: James Baker, Harrison Division, General Motors
Underwriters Laboratory
Sampling: Don Fournier of Acurex Corporation, MACS, Murray Corporation, Robinair, J&N Auto Air,
White Industries, Barney Gross Auto Air, Houston Auto Air, ARA of Houston, West End Auto Air, Triple L,
Hertz, Budget, and many more.
ii
-------
ABSTRACT
This project was initiated to provide a scientific basis tor choosing a reasonable standard of purity
tor recycled chloroflurocarbon (CFC) refrigerant in operating automobile air conditioners. It evaluated the
quality of the refrigerant from air conditioners in automobiles of different makes, ages, and mileage, from
different parts of the country, and with both failed and properly working air conditioners. The refrigerant,
CFC-12, was tested for water content, acidity, residue quantity, refrigerant purity, residue purity, inorganic
chloride, and inorganic fluoride. This work will be the basis for programs to reduce CFC emissions from
the servicing of automotive air conditioners.
Of the 227 cars sampled, neither the compressor oil nor the refrigerant showed any measurable
levels of acid (to 1 ppm), inorganic chlorides (to 0.1 ppm), or inorganic fluorides (to 0.1 ppm). One
possible explanation of these findings is that an automobile air conditioner is a relatively benign
environment for a material as chemically stable as CFC-12. A second explanation is that the acids that
might form are fully contained in the lubricant or are neutralized by the metal of the air conditioner
components. There was evidence that any small amount of tree acid that may have been in the sample
reacted with the material of the sampling system. This chemical reaction would result in deterioration of
metal, but would not degrade the refrigerant.
The gaseous refrigerant, in all but two samples was of higher purity than the specification for new
CFC-12. The two contaminated samples were analyzed to have 2- and 5-percent CFC-22. These levels
of CFC-22 did not reduce the air conditioners' performance to below the level acceptable to the owner.
The amount of residue measured in the CFC-12 was simply the compressor oil which was carried
over into the sampling container by the refrigerant. The amount of residue in each sample depended on
the amount of refrigerant in the air conditioner, the rate at which the sample was removed (the sampling
rate), and on how long since the air conditioner had been used before the sample was taken.
The residue (compressor oil) was also tested for purity. It was found to be very pure (>99 percent
in all but one or two samples). That impurity was found to consist of very small amounts (<1 ppm) of a
large number of different organic compounds. The concentration of any one of the compounds was too
low to identify. There was no statistically significant correlation between residue purity and car mileage,
whether the car's compressor was functioning, or with the area of the U.S. where the sample was taken.
The water content of the refrigerant was found to exceed the Federal Specification BB-F-1421A
(also known as "mil spec") of 10 ppm maximum. The mean for all of the samples was found to be
56 ppm. No statistically significant correlation was found between the water content of the refrigerant and
whether the compressor was working or failed nor with the area where the sample was taken; however, a
statistically significant correlation was found between the odometer reading of the car and the water
content. The mean water content for odometers registering up to 18,000 miles was 34 ppm. At higher
mileage ranges, the mean moisture content of the refrigerant was in the 56- to 94-ppm range.
iii
-------
TABLE OF CONTENTS
Section Page
Acknowledgement ii
Abstract iii
List of Figures . vi
List of Tables vii
1 Introduction 1
1.1 Operation ol a Mobile Air Conditioner 3
2 Conclusions 5
2.1 Acidity and Inorganic Halides 6
2.2 Residue Quantity and Purity 6
2.3 Refrigerant Gas Phase Purity 8
2.4 Water 8
2.5 Conclusion 8
3 Experimental Procedures Sampling and Analytical Program 10
3.1 Sampling Protocol 11
3.2 Sampling Equipment Description 11
3.3 Sample Equipment Preparation 13
3.4 Sampling Procedure 14
4 Analytical Procedures 17
4.1 Moisture Content 18
4.2 Acidity or Acid Number 19
4.3 Purity of the Refrigerant 20
4.4 Total Residue 20
4.5 Purity of Residue by Gas Chromatograph 21
4.6 Free Halide 22
5 Results 23
6 Quality Assurance/Quality Control 38
6.1 Data Quality Objectives for Critical Measurements 38
6.2 Sample Container Preparation 38
6.3 Data Reduction Methods 38
6.3.1 Moisture Content 40
6.3.2 Acidity or Acid Number 40
6.3.3 Purity of the Refrigerant 40
6.3.4 Total Residue 40
6.3.5 Purity of Residue by Gas Chromatograph 41
6.3.6 Free Halides 41
6.3.7 Quality Assurance Objectives for Clean Sample Containers 41
iv
-------
TABLE OF CONTENTS (concluded)
Section Page
6.4 Calculation of Data Quality Indicators 41
6.4.1 Accuracy 41
6.4.2 Precision 41
6.4.3 Completeness 42
6.5 Corrective Action Procedures 42
6.6 System and Performance Audits 42
6.7 Replicate Analyses and Standards 47
6.8 Internal QA Audit 48
7 References 51
Appendix A: Detailed Analytical Methods 52
Operating Procedure for Purity of Residue by Gas Chromatograph 53
Operating Procedure for Moisture Determination in CFC-12 54
Operating Procedure for Residue Analysis of Auto Air Conditioning Samples 59
Operating Procedure for Acid Number 62
Operating Procedure for Purity of R-12 in Automobile Air Conditioning Samples 65
Operating Procedure for Free Halide Analysis of R-12 Samples 66
NOTICE
This document has beer, reviewed in accordance with
U.S. Environmental Protection Agency policy and
approved for publication. Mention of trade names
or comercial products does not constitute endorse-
ment or recommendation for use.
v
-------
LIST OF FIGURES
Figure Page
1-1 Schematic of Mobile Air Conditioning System 4
3-1 Sampling Schematic 12
3-2 Vehicle Information Form 16
vi
-------
LIST OF TABLES
Table Page
2-1 Summary of Results (ppm) 7
3-1 Number of Cars Sampled n
5-1 Sampling Locations 23
5-2 Vehicle Information Listed by Sampling Location 24
5-3 Results of Analysis: Acids, Halides, Refrigerant Purity 31
5-4 Results of Analyses for Moisture, Residue, and Residue Purity 32
6-1 Quality Assurance/Quality ControlData Quality Objective 39
6-2 QA Objectives for Clean Sample Containers (ppm) 42
6-3 Results ol First Performance Evaluation Audit 43
6-4 Results of Second Performance Evaluation Audit 44
6-5 Results of Third Performance Evaluation Audit 44
6-6 Results of Fourth Performance Evaluation Audit 45
6-7 Results of Fifth Performance Evaluation Audit 45
6-8 Results of Analysis of QA Audit Samples 47
6-9 Results of Replicate Moisture Analyses 48
6-10 Standards and Replicates (Moisture) 49
vii
-------
SECTION 1
INTRODUCTION
Approximately 25 percent of all domestically consumed chlorofluorocarbons (CFC or CFC's) is
currently used in automobile air conditioners, the single largest use of these chemicals. Moreover,
current servicing practices result in substantial but unnecessary emissions of CFC-12
(dichlorodifluoromethane). During typical servicing, any CFC-12 remaining in the automobile air
conditioner is first vented to the air, a new charge of CFC-12 is sometimes used to test the system and
locate the leak, and finally the system is recharged wiih CFC-12 after repair.
In an effort to reduce the amount of CFC compounds released, automobile manufacturers,
servicing trade associations, and recycling equipment manufacturers are working together to develop a
standard for recycling CFC-12 for automobile air conditioners. Some equipment for draining, cleaning,
storing, and returning refrigerant to the system during servicing is presently available, and many
companies are working to introduce improved designs. The recycling equipment currently available is
often used during the servicing of fleet vehicles such as buses that hold large refrigerant charges, but it is
rarely used to reclaim the refrigerant during automobile servicing.
A reduction in the release of CFC to the atmosphere could be achieved by requiring the recovery
and reuse of the refrigerant from all automobile air conditioners serviced; however, there has been little
information available on the level of contamination in operating automobile air conditioners and the ability
of equipment to satisfactorily clean the CFC for reuse.
This project to evaluate CFC refrigerant from automobile air conditioners was initiated in
response to these questions. The quality of refrigerant present in vehicles of different makes, ages, and
mileage and from different parts of the country has now been assessed. The refrigerant from
227 vehicles with both failed and properly working air conditioners was collected and tested. The results
of the program have provided an understanding of not only the quality of the refrigerant found in
automobiles but also of how failure of the compressors and other equipment affects its contamination.
This work will be the basis for programs to reduce CFC emissions from the servicing of automotive air
conditioners.
The work was guided by and performed in cooperation with an ad hoc industry group comprised
of representatives of the following:
Environmental Protection Agency
Mobile Air Conditioner Society (MACS)
Motor Vehicle Manufacturers Association (MVMA)
Manufacturers of Small Recovery/Reclamation/Recycle Devices
American Society of Heating, Refrigeration, and Air Conditioning Engineers (ASHRAE)
Small air conditioning shops
i
-------
Underwriters Laboratory
Other industry representatives and individuals
The Ad Hoc committee, chaired by Robert Bishop/GM, and Simon Oulouhojian/MACS, composed
of representatives of automobile manufacturers, servicing trade association, and recycling equipment
manufacturers agreed that the recycled CFC would not have to meet specifications for virgin CFC.
Instead, they agreed that a standard of purity comparable to that of CFC in automobiles that have been in
use for 15,000 miles {+/- 3,000 miles) with properly working air conditioners would be adequate. The
committee decided to work cooperatively to define the acceptable standard of purity and to
simultaneously work toward development and certification of recycling equipment to satisfy this standard.
The first step in the program was to decide the parameters that needed to be measured to
determine the quality of refrigerant. The ad hoc industry group agreed that the following parameters
would fully describe possible refrigerant contaminants:
1.
Water content
2.
Acidity
3.
Residue
4.
Chloride ion
5.
Purity of the liquid phase
6.
Purity of the gas phase
The group also determined that the recycled refrigerant would be considered satisfactory for
reuse if recycling equipment could achieve a standard of purity comparable to that of the refrigerant in
properly working air conditioners in automobiles that have been driven for 15,000 ± 3,000 miles. Thus,
the two main objectives of the program were (1) to determine the purity of CFC using the six parameters
listed above for properly working air conditioners in cars at 15,000 ± 3,000 miles and (2) to determine the
maximum CFC contamination for cars that will seek service due to major component failure.
Manufacturers of refrigerant recycling equipment will then certify that their equipment can clean
the possible contamination to the levels of purity of the 15,000 mile standard. The workgroup will formally
recommend to automobile manufacturers that procedures using certified recycling equipment qualify for
new car warranty work.
The following information was required to meet the objectives for this program:
1. The amount of deterioration that the refrigerant suffers over the service life of the vehicle.
This degree of deterioration corresponds to a determination of the quality of the refrigerant
versus the vehicle mileage for various parts of the country and for both operating and
defective air conditioners.
2. The chemical nature of the contaminants that cause the deterioration of the refrigerant. This
information will identify the contaminants that need to be removed to recycle the refrigerant,
If it is unsatisfactory.
3. The conditions that cause contaminated refrigerant. It is unnecessary to require that
refrigerant removed from an air conditioner during servicing be subjected to chemical
2
-------
analysis if equipment can automatically clean it to a satisfactory purity. If necessary, service
technicians can also consider factors such as the vehicle's age and mileage or the reason
for failure of the air conditioner system when they service particular vehicles.
In response to the objectives, refrigerant in a population of vehicles from different parts of the
U.S. with a variety of mileage and automobile air conditioners was evaluated. The program comprised
the following main components:
1. Determine the level of impurity found in the refrigerant of vehicles presently on the road at
varying mileage and with both properly functioning and defective automobile air conditioning
systems.
2. Determine whether the degree of degradation (if any) is the same for each region of the U.S.
or whether variabilities such as climate and level of use result in a variation in the quality of
the refrigerant.
Prior to full-scale field sampling, the sampling and analytical methods used were tested by
sampling 12 cars from the Research Triangle Park, North Carolina, area. This step perfected the
sampling procedures and determined the variability likely to be encountered.
For the full-scale field sampling, the Mobile Air Conditioning Society identified automobile air
conditioner service centers in four areas of the United States. Refrigerant samples were taken and
analyzed from these areas.
1.1 OPERATION OF A MOBILE AIR CONDITIONER
Before further discussion of this program, a brief explanation of mobile refrigeration systems will
be presented for background. A typical mobile air conditioner, which is shown in Figure 1 -1, consists of a
compressor, evaporator, expansion valve, and condenser. Other valves and equipment, which are not
shown, are installed in a working system to scavenge trace amounts of moisture and dirt and to ensure
that the system functions properly. The condenser and evaporator are heat exchangers, similar in
appearance to automotive radiators, that contain the refrigerant. In an automobile, the condenser is
mounted in front of the engine fan under the hood where it is cooled by the outside air; the evaporator is
mounted in the ventilation system to cool the air inside the car.
The refrigeration cycle works by pumping the working fluid through the compressor to increase its
pressure. The refrigerant is a gas at that point; its temperature is well above ambient. The refrigerant
flows from the compressor to the condenser where outside air cools the refrigerant to nearly the ambient
(outside) temperature, causing the refrigerant to condense into a liquid and releases its heat to the
outside air.
The liquid refrigerant flows from the condenser through an expansion valve to the evaporator
where the pressure is low enough so that the refrigerant vaporizes. During vaporization, the temperature
of the refrigerant drops. Air is blown across the cooled refrigerant tubes of the evaporator and into the
passenger compartment. The refrigerant returns to the compressor though the suction throttling valve to
repeat the cycle.
3
-------
BmMi
SS&tfl
HIGH PRESSURE UCXJD
LOW PRESSURE UQUID
HIGH PRESSURE GAS
LOW PRESSURE GAS
Evaporator
Expansion
Valve
Figure 1-1. Schematic o< mobie air cond Honing system.
4
-------
SECTION 2
CONCLUSIONS
This project was initiated in response to a need for information on the level of contamination of
the refrigerant in operating automobile air conditioners. It evaluated the quality of chlorofluorocarbon
(CFC) refrigerant from air conditioners in automobiles of different makes, ages, and mileage, from
different parts of the country, and with both failed and properly working air conditioners. The Ad Hoc
committee composed of representatives of automobile manufacturers, servicing trade association, and
recycling equipment manufacturers agreed that the recycled CFC would not have to meet specilications
for virgin CFC. Instead, they agreed that a standard of purity comparable to that of CFC in automobiles
that have been in use for 15,000 miles (+/- 3,000 miles) with properly working air conditioners would be
adequate. The general goal of this project was to determine the following:
1. The amount of deterioration that the refrigerant suffers over the service life of the vehicle.
This degree of deterioration corresponds to a determination of the quality of the refrigerant
versus the vehicle mileage lor various parts of the country and for both operational and
defective air conditioners.
2. The chemical nature ol the contaminants that cause the deterioration of the refrigerant. This
information will identify the contaminants that need to be removed to recycle the refrigerant,
if it is unsatisfactory.
3. The conditions that cause contaminated refrigerant. It is unnecessary to require that
refrigerant removed from an air conditioner during servicing be subjected to chemical
analysis if equipment can clean it to a satisfactory purity automatically. II necessary, service
technicians can also consider factors such as the vehicle's age and mileage or the reason
for failure of the air conditioner system when they service particular vehicles.
In response to the objectives, refrigerant in 227 vehicles from different parts of the U.S. with a
variety of mileage and automobile air conditioners was evaluated. The program comprised the lollowing
main components:
1. Determine the level of impurity found in the refrigerant of vehicles presently on the road at
varying mileage and with both properly functioning and defective automobile air conditioners
systems.
2. Determine whether the degree of degradation (if any) is the same for each region of the U.S.
or whether variabilities such as climate and level of use result in a variation in the quality of
the refrigerant.
The refrigerant in the samples was evaluated on the basis of the following six parameters:
1. Acidity
2. Chloride and Fluoride ioninorganic halides
5
-------
3.
Residue
4. Purity of the liquid phase
5. Purity of the gas phase
6. Water content
Table 2-1 summarizes the results of the sampling and analytical effort. The following sections
discuss each of the above six parameters in greater detail.
2.1 ACIDITY AND INORGANIC HALIDES
Neither the refrigerant nor the residue (compressor oil) which came with it during the sampling
showed any measurable level of acid (to 1 ppm), inorganic chlorides (to 0.1 ppm), or inorganic fluorides
(to 0.1 ppm). The refrigerant in all the samples were better than the purity requirement for new CFC-12
by these criteria.
One possible explanation of the absence of acid is that an automobile air conditioner is a
relatively benign environment for a material as chemically stable as CFC-12. A second explanation is that
the acids that might form are fully contained in the lubricant or are neutralized by the metal content of the
air conditioner components. There was evidence that any small amount of free acid that may have been
in the sample reacted with the material of the sampling system. This chemical reaction would result in
deterioration of metal, but would not degrade refrigerant.
The finding on the lack of acid is good news for the program. Because the sampling system was
selected to closely duplicate the recover system that will be used to recycle the refrigerant, there is every
reason to believe that no significant quantity of acid will be removed from the MAC during
recycling/servicing. Furthermore, any acid present during normal capture and recycling of the refrigerant
can be removed by the recycling equipment. Acid can be neutralized by contact with metal components
or by the use of special absorbents which can be incorporated in the recycling equipment. Based on this
laboratory analysis, acidity in recycled refrigerants will not be a problem if recycling equipment is properly
designed.
2.2 RESIDUE QUANTITY AND PURITY
The level of residue in each sample depended on the amount of refrigerant in the air conditioner,
the rate at which the sample was removed (the sampling rate), and on how recently since the air
conditioner had been used before the sample was taken. The residue detected in the samples is
primarily the compressor oil which was carried over into the sampling container by the refrigerant. No
significant contamination, other than oil, was found in the CFC.
The residue (compressor oil) was also tested for purity. It was found to be very pure (>99 percent
in all but one or two samples). The impurity was found to consist of very small amounts (<1 ppm) of a
large number of different organic compounds. The concentration of any one compound was too low to
allow identification. The residue turned out to be a reasonably good quality compressor oil. Attempts
were made to correlate the residue purity with car mileage, with whether the compressor had failed or not,
and with the part of the country where the sample was taken. No correlation was found with any of these
three parameters.
6
-------
Table 2-1.
Summary of Results (ppm)
Moisture
Residue purity
Good compressors
(miles)
No.
Maximum
Mean
SD
No.
Maximum
Mean
SD
0-12,000
15
207
34
50
16
7,600
1,841
2,300
12,000-18,000
49
127
34
28
47
9,900
1,969
2,353
18,000-40,000
39
1,002
73
189
39
10,600
1,656
2,327
40,000-60,000
25
413
56
77
23
6,600
1,246
1,558
60,000-90.000
41
224
49
36
41
9,700
1,230
2,277
>90,000
23
755
94
147
22
4,700
785
1,232
Subtotal
192
188
Failed Compressors
24
515
58
100
26
5,700
B52
1,208
Total
216
214
Blanks
21
65
15
16
20
2,100
313
504
-------
2.3 REFRIGERANT GAS PHASE PURITY
The purity of the refrigerant itself was tested by withdrawing a sample of the gas phase from the sampling
container and analyzing it with a gas chromatograph/flame ionization detector. The purpose of this test
was to determine whether any of the refrigerant samples had been contaminated with other CFC's such
as HCFC-22. The test could also identify any gaseous products of decomposition of the refrigerant or of
the compressor oil. Except for two samples that showed some HCFC-22, no measurable extraneous
materials were found in the gas phase of the refrigerant. The level of detectability of the analytical
method was approximately 100 ppm for any chemical other than CFC-12.
Trace quantities of other CFCs and HCFCs are common contaminants In CFC-12 and are
allowed by the specifications for new CFC-12 to compose up to 0.5 percent of the product. Samples of
new CFC-12 from several suppliers were analyzed as part of this program and were found to contain up
to 0.1% HCFC-22 as well as of other volatile components. HCFC-22 contamination in operating
automobile air conditioners cannot remain very high because it quickly leaks out through the rubber hose
materials. Only two samples of refrigerant out of the 227 automobiles tested were found to contain more
than 0.5 percent HCFC-22 in the CFC-12 and neither level of contamination (up to 5 percent) caused the
air conditioner's performance to deteriorate to the point where the owner chose to have it repaired. This
is discussed further in Section 5.
2.4 WATER
The water content of the used refrigerant was found to exceed the Federal Specification
BB-F-1421A for new CFC (also known as "mil spec") of 10 ppm maximum [1], The mean for all of the
samples was found to be 56 ppm. No correlation was found between the water content of the refrigerant
and the area of the United States where the sample was taken; however, a correlation was found
between the odometer reading of the car and the water content. The mean water content for cars up to
18,000 miles was 34 ppm. Above this mileage, the mean moisture content of the refrigerant in different
mileage ranges remained in the 56- to 94-ppm range. No statistical difference was found between the
water content of systems having failed and functioning compressors.
2.5 CONCLUSION
This sampling and analytical program showed that the refrigerant in operating air conditioners is
very pure. Acids do not accumulate in the refrigerant. Any impurities that accumulate in the air
conditioning system are concentrated in the compressor oil. They are dissolved by the liquid phase of the
refrigerant but do not get carried over into the gas phase. The gas phase proved to be free of
contaminants and equivalent in purity (as measured by a gas chromatograph with a flame ionization
detector) to new CFC-12.
Water content was the only parameter which was highly dependant on vehicle mileage. In better
than 95 percent of the sample analysis, moisture was present above the mil. spec. It tended to be greater
in vehicles with higher mileage. However, even refrigerant in new vehicles had a moisture level greater
than the 10 ppm specification on new CFC-12. This may be due to the small amount of moisture that is
present on all manufactured parts such as the compressor, expansion valve, and hoses and to the
migration of moisture through hose material. As illustrated by the relatively small standard deviations
shown in Table 2-1, the moisture in the lower mileage ranges does not vary as much as it does in the
higher mileage ranges.
This increase in moisture at the 18,000-mile level could indicate the start of deterioration of the
drying agent in the air conditioner. As the drying agent is saturated, the moisture is at a higher
concentration and the variability increases.
-------
The moisture level in the refrigerant did not show any correlation with geographic location or
vehicle make.
Contamination of the CFC-12 with HCFC-22 is not widespread. Only two cars out of more than
200 tested contained more than ihe limit for new CFC-12 of 0.5 percent maximum. Even if it occurs, its
effect is limited since it quickly leaks out of the system through the hoses and has a very limited effect on
the air conditioner's performance.
In summary, the data gathered here indicate that the CFC-12 refrigerant does not degrade
significantly with use. Furthermore, while small amounts of contaminant are removed with the refrigerant
during servicing, the bulk of the contaminants remain with the compressor oil. Current servicing practices
do not require that the compressor oil be changed unless the compressor is replaced. The presence of
HCFC-22 in concentrations above the specification for new CFC-12 is rare, less than 1 percent of the
cars tested. HCFC-22 contaminant quickly leaks out of the automotive air conditioner through hoses and,
does not appear to cause operational problems while in the system.
9
-------
Intentionally Blank Page
-------
SECTION 3
EXPERIMENTAL PROCEDURES SAMPLING AND ANALYTICAL PROGRAM
The protocol lor this sampling and analysis program called for a statistically significant sample
both from cars with properly functioning air conditioners and those in the service center specifically for a
problem with the air conditioner. The goal was to select cars with the following mileage and/or
malfunction:
1. 15,000 ±3,000 (12,000 to 18,000) miles
2. 40,000 to 60,000 miles
3. 90,000 and greater miles
4. Cars of any mileage with defective compressors or other components that will likely cause
overheating and a resultant deterioration of the refrigerant.
Efforts were made also to select a percentage of cars that represents the share of the U.S.
market from the four categories of manufacturers listed below:
1. General Motors
2. Ford Motor Company
3. Chrysler Corporation
4. Foreign manufacturers
Because the sampling program was limited by availability of vehicles while sampling personnel
were at the automobile repair shops, it proved impossible, in spite ol many attempts to do so, to select
cars of different manufacture and restricted mileage ranges in the numbers specified. It was decided,
therefore, to relax these constraints and sample cars in the desired mileage ranges and not be as
restrictive on the category of manufacturer as that called for in the sampling plan. A good mix of cars
from the four categories of manufacture was achieved in this way. A subsequent statistical evaluation of
the data indicated that the objectives of the study were met under this less restrictive sampling program.
To determine whether geography is a factor in refrigerant deterioration, samples were taken from
service centers at the following four locations:
1. Texas-Louisiana area (Gulf Coast): hot and humid to hot and arid weather; mostly
long-distance driving
2. Maryland (Mid-Atlantic States): hot and humid weather; mostly short-distance, high-traffic
driving
10
-------
3. Northern Ohio (Midwest): hot and humid part ol the year and very cold weather
4. Denver, Colorado (Mountain States): low humidity weather; high altitude
3.1 SAMPLING PROTOCOL
The program was performed in three stages. The first was a preliminary program to test the
sampling procedure and obtain an estimate of the level of variability that was likely to be encountered in
the subsequent program. This preliminary estimate of the level of variability was necessary to determine
the number of samples that needed to be taken for a statistically valid sample size. For this stage,
12 automobiles in the Research Triangle Park, North Carolina, area were sampled. The second stage
called for approximately 120 samples to be taken at service centers in the Gulf Coast area. For the third
stage, approximately 40 samples would be taken at a single air conditioning service center in each of the
three other geographical areas.
Table 3-1 shows the number of cars that were sampled in each geographic area. As can be
seen, the 107 samples taken at the Gull Coast locations were lower than the goal of 120 because of the
limited number of cars available at any one site during the sampling program. The goal oi 40 samples
was achieved at each of the other three locations.
3.2 SAMPLING EQUIPMENT DESCRIPTION
The equipment used for sampling the automobile air conditioners, which is shown in Figure 3-1,
consists of the following three components:
1. Sampling cylinder
2. Sampling line
3. Manifold gauge set and vacuum pump
Three hundred sampling containers that were manufactured specifically for this purpose were
used for this program. Each container is an (approximately) 1-gal steel vessel equipped with a 2-way
valve suitable for refrigerant 12 (R-12). The vessel is rated for a minimum of 250 psi, and the valve has a
safety release which opens at this pressure. The valve opening is equipped with a metal screw cap to
protect the container during shipping and handling and to act as a secondary seal to reduce the likelihood
of sample loss.
Table 3-1. Number of Cars Sampled
Geographic area No. cars sampled
Gulf Coast 107
Northeast 40
Midwest 40
Mountain 40
Total 227
11
-------
Figure 3-1. Sampling schematic.
-------
The sampling line, which is shown in Figure 3-1 as the line within the broken rectangle and which
is made of 1/4-in flexible copper tubing, was used only once to prevent the risk ot cross-sample
contamination. One line was made up for each container and was shipped with it. Prior to shipment,
these lines were flushed with CFC-113, blown out with nitrogen, and capped. One in 10 of these lines
was tested by filling it with CFC-113 and analyzing the solvent in the same way as the containers were
tested.
The manifold gauge set and vacuum pump along with the ancillary tubing are standard equipment
at all automobile air conditioner repair shops and were provided by the service centers where the
sampling took place. Since no refrigerant sample flows through these lines, no special effort was
necessary to ensure their cleanliness. Nevertheless, after the sampling system was hooked up but
before sampling was started, a small amount of refrigerant was vented through the manifold gauge set to
purge the lines.
3.3 SAMPLE EQUIPMENT PREPARATION
Before being shipped to the sampling sites, ail sampling containers and sampling lines were dried
and then tested to ensure that they were clean and capable of holding pressure. The containers were
first cleaned, then tested for cleanliness. Each cylinder was also tested to ensure that It could hold the
sample of CFC-12.
It was assumed that all the containers would have moisture in them as a result of the
manufacturing process. Thus, all containers were first dried upon receipt by heating them to
approximately 110 °C (230 °F) in an oven overnight. The next day, each container was removed from the
oven, attached to a vacuum pump and a vacuum of less than 3.4 kPa (1 in Hg) was applied to it for a
minimum of 30 min. This evacuation removed other volatile impurities as well as water. Any container
that could not hold this level of vacuum was eliminated as defective.
The next step involved determining the cleanliness of the containers. One container in 10, for a
total of 30 containers, was randomly selected for this test. The containers selected were filled with clean
CFC-12 and were allowed to stand at least overnight. The contents were subsequently analyzed for
water, residue, purity, and acidity by the procedures given in Section 4 and in Appendix A. The
containers were found to satisfy the requirements for all parameters except residue. All the 30 containers
tested had an excessive amount ot residue, which probably accrued from the manufacturing process.
Since the purity of the residue in the sampled refrigerant was one of the parameters being measured, the
containers had to be cleaned prior to use.
The sampling containers were rinsed by partially filling each with clean CFC-12, then manually
shaking and rotating it to expose all internal surfaces to the solvent. Once all of the containers had been
cleaned in this way, they were again checked for purity by refilling one with CFC-12, agitating and rotating
it, and then, pouring the solvent into a second container. The same solvent was poured and then agitated
from one container to the next, for a total of 10 containers. The solvent was then analyzed for residue.
This procedure reduced the amount of solvent required by a factor of 10 and the number of analyses
needed from 300 to 30. All rinsed containers passed this test. The QA objectives for. the program, which
also describe those for the sample containers, are given in Section 6.
Once the sample containers were found to be clean, they were pressure tested to ensure that
they would hold the sample through shipment. This test was performed by filling the containers with dry
nitrogen to approximately 1400 kPa (200 psi) and letting them stand overnight. Any container which
showed a measurable pressure drop (to within 3.4 kPa |1 in Hg]) overnight was not used. After the
pressure was tested on the next day, the nitrogen in the containers was released. They were again filled
with nitrogen to 1400 kPa (200 psi), and soap solution was put on the junction of the valve and container
to test for leaks. If no leaks were detected, the cylinder was emptied again and filled with nitrogen to
70-140 kPa (10-20 psi) for shipping.
13
-------
The sampling lines used to transfer the sample from the automobile's air conditioning system to
the sampling container were also cleaned and tested. These lines had been assembled expressly for use
in this program from readily available lubing and fittings. The cleaning and testing procedure took
advantage of the mating fitting at each end of the sampling lines (identified by the dotted block on
Figure 3-1). Ten lines were connected using the fittings, and then clean CFC-113 was poured into the
lines that were hooked together. The lines were rotated and mixed, and the solvent was then poured out.
This procedure was repeated three times. The solvent from the third rinse was then tested in a manner
similar to that from the sampling tanks. All lines were found to be clean, had they not been, the rinse
would have been repeated until they were. The lines were then purged with dry air and evacuated to
remove the CFC-113 residue. They were disassembled, and the ends were covered with Teflon tape to
keep dirt out during handling and shipping.
Each sampling container was packed with a sampling line and shipped to the service centers to
be filled with samples.
3.4 SAMPLING PROCEDURE
At the sampling site, the sampling containers were filled with refrigerant from the automobiles by
placing the sampling container into a pan of dry ice and evacuating the line and container. The sample
was then drawn from the air conditioning system at the high pressure side so that lubricant would be
withdrawn with the refrigerant. Once the sampling was completed, the air conditioning system was
serviced, if needed, and then recharged by service center personnel following normal procedures.
The sampling system schematic is discussed in Section 3.1. The sampling procedure, which
refers to Figure 3-1, consists of the following steps:
1. Take the valve cover off the sampling container and set it aside for later use. Open the
valve on the sample container and listen for a slight hiss of released gas. The container was
filled to about 10 psi with nitrogen before shipping to keep air out. If you do not hear the
hiss, the container leaks. Do not use this sample container.
2. Loosely connect a new sampling line to the sample container and the vacuum line as shown.
3. Put the sampling container into a pan and cover about two-thirds of the way up with dry ice.
4. Make sure that valve V-1 is closed, and then connect the sampling line to the high pressure
side of the air conditioner (A/C). The A/C must not be running.
5. Open V-1 slightly for less than 1 s to purge the line. Close V-1.
6. Tighten all the connections and be sure that V-1 remains closed. Open V-2, V-3, and the
valve on the sample container. Turn on the vacuum pump.
7. Evacuate the system up to V-1 including the sample container to as low a pressure as
possible: a minimum 29 in vacuum is needed. Hold for 5 min.
8. Close valve V-3, (leave V-2 and the valve on the sample container open), and then shut off
the vacuum pump.
9. Open V-1 and allow the complete refrigerant charge to go into the sample container.
Monitor the pressure/vacuum gauge to ensure that the pressure does not exceed 200 psi. If
that pressure is exceeded, close the valve and stop the sampling. Otherwise, proceed with
the sampling.
14
-------
10. When the sampling is completed, close the valve on the sample container, disconnect the
system, and replace the cap on the sample container valve. Remove the sample container
from the dry-ice bath and fill out the vehicle information form.
11. Put the sample container, sampling line, and the vehicle information form back into the box,
seal the box, and set it aside for shipping back to the laboratory.
12. The air conditioner can now be serviced in the normal manner.
For each vehicle, the model, year, type ot engine and air conditioner, and other information will be
recorded on the Vehicle Information Form shown in Figure 3-2.
15
-------
PASSENGER CAR, VANS AND LIGHT DUTY TRUCKS ONLY
CFC (R-12) SAMPLING
DATE CONTAINER # SAMPLED BY (Name)
SERVICE CENTER
NAME PHONE #
ADDRESS (City, State)
VEHICLE MAKE & YEAR
VEHICLE MODEL
VEHICLE VIN
COMPRESSOR MANUFACTURER
FACTORY AIR CONDITIONING: YES NO TYPE OF COMPRESSOR
CURRENT MILEAGE
PREVIOUS SERVICE: YES PLEASE EXPLAIN
(Mileage at service)
NO UNKNOWN
HAS THE A/C BEEN CHARGED IN THE LAST YEAR: YES NO
ORIGINAL: EVAP CONDENSER HOSE COMPRESSOR RECEIVER/ACCUMULATOR
SERVICE DIAGNOSIS AND REPAIR, (this date)
DO NOT TAKE SAMPLES OF TEST CHARGE
Figure 3-2.
-------
SECTION 4
ANALYTICAL PROCEDURES
The sampling procedure described above resulted in samples of refrigerant containing significant
amounts of compressor oil. The contents of each sample container were analyzed as received by the
method indicated for the following:
1. Moisture content: Karl Fischer titration
2. Acidity or acid number: KOH titration
3. High boiling residue or oil content: Gravimetric analysis
4. Cleanliness or purity of the refrigerant: Determined by GC
5. Purity of the residue: By GC
6. Free halides: Ion chromatography
Standard methods for these tests are available and are listed in Section 7,
"References." [2,3,4,5,7,8* ] These methods were tried early in the program but found not to be
acceptable since they are designed to test pure CFC-12, not the oil/chlorofluorocarbon mixture that was
evaluated under this program. The analytical methods which were actually used are described in greater
detail below and in Appendix A. Also see Section 6 for additional information. The methods are
presented only for the purpose of documenting the procedure used. They are not being recommended
for use in all situations.
The first three tests of the six shown above are standard wet analyses commonly used
throughout the industry. The moisture content test is a Karl Fisher (K-F) titration procedure, with only
minor modifications. The acidity of the sample is determined by a potassium hydroxide (KOH) titration,
typically done to a bromothymol blue endpoint, although other indicators are sometimes used. The
high-boiling residue or oil content of the refrigerant is typically measured by slowly evaporating the
contents of a small weighed sample cylinder and rinsing the residue from the walls of the cylinder into a
tared dish, which is reweighed.
Two purity tests were conducted: one of the refrigerant gas phase and one of the residue. The
gas phase purity was tested to determine whether the CFC-12 was contaminated with other refrigerants
or volatile products of decomposition. It was hypothesized that this type of contamination could occur
because of previous improper recharging or because the refrigerant had decomposed to a lighter
material.
* ASTM Standards are available from American Society for Testing Materials, 1916 Race Street,
Philadelphia, PA.
17
-------
Both these measurements were made with a GC equipped with a flame ionization detector.
Attempts were made to identify the types of impurities found by the use of a GC equipped with a mass
spectrometer (GC/MS); however, the concentrations of impurities in all samples were too low to be
Identified.
The purity of the gas phase of the refrigerant was checked by injecting a sample Into the GC with
a gas sample loop. The purity of the residue was checked by evaporating a weighed quantity of liquid
refrigerant from a pressure bomb, rinsing the interior of the bomb with a solvent, and injecting the solvent
into the gas chromatograph with a flame ionization detector.
The test for inorganic chloride is intended to determine if the refrigerant has decomposed
because of severe thermal stress. Because chlorofluorocarbons are exceptionally resistant to hydrolysis,
chemical decomposition, especially under the mild conditions in an automobile air conditioner, is
uncommon. Indeed, no free chlorides or fluorides were found. This test is performed by extracting the
sample with a buffered aqueous solution and then analyzing the extract for chloride ion. The standard
analytical method for doing this is bubbling the refrigerant into silver nitrate solution. Free chlorides would
show up as a precipitate. This method was felt to be far too insensitive to be used in this program.
Rather, the buffered aqueous solution was analyzed using an Ion chromatograph, a procedure which is
sensitive to approximately 100 parts per billion (ppb) of free chloride or fluoride.
All the analyses were performed on the total samples as received. Because of the sampling
procedures that were used, the refrigerant samples were mixed with compressor oil from the air
conditioners. The analyses were performed on the combined refrigerant and compressor oil. The
following sections describe in greater detail the analytical procedures used. The operating procedures
which were followed by the analyst are given in Appendix A.
: The standard procedures for water, acidity, and halide analyses recommended by Allied Signal
Corporation and the DuPont Freon Products Laboratory specify that the sample of refrigerant to be
analyzed be transferred from the field sampling container to a clean and dry stainless steel 150-ml
laboratory sample cylinder. The analysis is performed by letting the CFC-12 bubble into an absorbing
solution which is analyzed for water acidity or halides by the appropriate analytical technique described
below.
The two-step procedure was tested early in the program, and it was found to increase the risk of
sample loss in the 150-mL sample cylinder. It was determined that the only reason for first transferring a
portion of the sample to the smaller vessel was to allow it to be weighed to the nearest 0.1 g. Newer
types of electronic balances have sufficient capacity to determine the weight of the sample to 0.1 g
without the need to transfer part to a smaller container. The analytical procedure used here and all
absorptions were made directly from the field sampling container.
The following subsections give a description of the methods that were used to perform the
analyses. Appendix A gives the Operating Procedures for each of the analyses.
4.1 MOISTURE CONTENT
Moisture content was determined by K-F titration using a Fisher Coulomatic K-F Titrimeter
System (Fisher Scientific Catalog Number 09-313-447). This apparatus has a coulometric cell filled with
K-F reagent. The water content of the CFC-12 was measured by bubbling a known weight of liquid phase
sample directly into the K-F reagent in the cell and then reading the amount of water that was captured.
The procedure is more direct and reproducible than the standard method which requires bubbling the
refrigerant into anhydrous methanol and manually titrating the methanol with K-F reagent.
The following procedure was used for the analysis:
18
-------
1. Attach the sampling container to one end of a Teflon tube. Attach the other end to a fritted
gas dispersion tube. Make sure that the Teflon tube is long enough to reach to the K-F
apparatus. Do not put the dispersion tube into the K-F reagent until step 3, below.
2. Agitate the sampling container and placed it valve side down in a stand. This is done so that
liquid flows from the sample container through the dispersion tube.
3. Slowly open the valve on the sampling container to blow a small amount of refrigerant
through the line to flush it and remove the entrained air. This should be done for less than
1 s to avoid cooling the line to the point where water may condense in it.
4. Tighten the fitting on the sampling container and weigh the assembly including the Teflon
tube and gas dispersion tube to the nearest 0.1 g.
5. Put the sampling cylinder into a rack upside down (to allow liquid to flow out of the valve)
and place the dispersion tube into the K-F apparatus.
6. Zero the apparatus to ensure that the solution has no water in it at the start.
7. Open the valve on the sampling container a small amount and let approximately 100 g of
CFC-12 flow into the K-F solution. The rack and sampling cylinder can be placed on a
balance to monitor the amount of sample withdrawn. The f lowrate should be adjusted to
allow a minimum of 10 min for the release of the 100 g of sample.
8. At the conclusion, read the K-F apparatus for the amount of water captured and record the
value.
9. Reweigh the sampling container to the nearest 0.1 g and record the weight.
Standards for this analysis were prepared to bracket the concentration of water in the sample by
injecting microliter amounts of deionized water into the solution to prepare standard curves. Sample
spikes were prepared by injecting microliter amounts of water into CFC-12.
A 4-point standard curve is run at the beginning of the day. A standard is run at midday and ai
the end of the day to ensure instrument response stability.
4.2 ACIDITY OR ACID NUMBER
The acidity was determined by alkalimetric titration using KOH by the following procedure. The
acidity was reported as "ppm as HCI."
1. Place 150 mL of double deionized water into a 250-mL Erlenmeyer flask, add 6 to 8 drops of
bromothymol blue indicator, and put in a magnetic stir bar.
2. Put the flask on a magnetic stirrer.
3. Titrate the deionized water and indicator to the green endpoint.
4. Repeat steps 1 through 8 of the procedure described in Section 4.1 with the exception that
the gas dispersion tube is put into the Erlenmeyer flask prepared in step 1 above.
5. Titrate the sample to the green endpoint with 0.005 N KOH to the nearest 0.001 mL.
19
-------
The normality of the 0.005 N KOH was validated daily by titration with NBS traceable KHP
(potassium acid phthalate).
4.3 PURITY OF THE REFRIGERANT
Each sample of refrigerant was analyzed by gas chromatograph with a flame ionization detector
to check for contaminants, primarily to determine whether other refrigerants, such as HCFC-22, may have
gotten into the system. These analyses were done by sampling the gas phase (as opposed to the liquid
phase, which was sampled for the water and acidity analyses) and injecting it directly into the gas
chromatograph with a flame ionization detector through a gas sampling loop.
The gas phase of the refrigerant was sampled by attaching a sample valve directly to the
sampling container. The sample was constructed so that it could be flushed by at least 10 volumes of
gas before a sample is taken. A 5-mL sampling loop was used to perform the injections of the gas into
the GC.
The GC conditions were as follows:
Column temperature: 50 °C
Injector temperature: 125 °C
Detector temperature: 250 °C
Flow rate: 30 mUmin He
Detector: FID
Column: 5 percent Fluorcol on Carbopack B 60/80 mesh 10-ft by 1/8-in SP alloy
Samples of a variety of high purity CFC's were used as standards for comparison against the
samples.
Although the protocol called for identifying unknown impurities via a GC/MS, this procedure
proved unnecessary because none were found.
4.4 TOTAL RESIDUE
The total residue analysis determines the amount of high-boiling compressor oil and degradation
products present in the refrigerant. The method is a gravimetric determination. The GC samples for the
residue purity analysis were generated during the following procedure.
1. Dry a 150-mL stainless steel sample cylinder by heating it overnight to 100 °C and purging
with dry air or nitrogen.
2. Weigh the sample cylinder to the nearest 0.1 g.
3. Attach a Teflon line to one end of the cylinder and to the sampling container.
4. Open both valves on the sample cylinder and briefly (less than 1 s) open the valve on the
sample container to purge the lines. Close the valves.
5. Turn the sample cylinder upside down and transfer approximately 100 g of sample as a
liquid into it by opening the valve on the sampling container and the valve on the sample
20
-------
cylinder which connects it to the sampling container. Keep closed the second valve on the
sample cylinder, which opens it to the atmosphere. When the sample has been transferred,
close the valves.
6. Reweigh the sample cylinder to determine sample weight to the nearest 0.1 g.
7. Put the sample cylinder in a rack in the hood so that one ol the valves points straight up.
8. Open the valve at the top on the gas phase of the cylinder a small amount to slowly release
the refrigerant over a period of 1 h.
9. When the refrigerant has completely evaporated, add 25 mL of CFC-113
(1,1,2 trlchloro-1,2,2 trifluoroethane) to it. Swirl It well to rinse the walls thoroughly and pour
the CFC-113 into to a 100-mL volumetric flask. Repeat the solvent rinse three times, and
combine the solvent from each rinse in the flask. Bring the flask to volume with CFC-113
and mix well.
10. Pour the solvent into a 50 mL volumetric flask to the mark. Transfer a small, measured
fraction of this solution into a vial with a Teflon lined cap for the residue purity analysis
discussed in Section 4.5.
11. Pour the remaining CFC-113 from the 100 mL volumetric flask into a tared aluminum dish;
rinse the flask with two 5-mL washes and add to the dish. Blank gravimetric analyses are
run by the addition of 100 mL of CFC-113 into a clean aluminum dish. The dishes are
placed on a hot plate in the hood and evaporated to near dryness. The hot plate is adjusted
so that it remains below the boiling point of the solvent. When the dishes are near dryness,
they are removed to an oven at 105 °C for 30 min. The dishes are removed from the oven
and allowed to cool in a desiccator. The dishes are reweighed to the nearest 0.0001 g to
determine the amount of high-boiling residue. One in 10 samples are done in duplicate, and
a blank is run every day of analysis.
4.5 PURITY OF RESIDUE BY GAS CHROMATOGRAPH
The method used to check the purity of the high-boiling residue is a variation of the Total
Chromatographable Organics (TCO) analysis, which is an interim procedure described in Document No,
AEERL/13, Revision 3, September 25,1986. The changes in this method include changing the solvent
from dichloromethane to CFC-113 and possible analysis by GC/MS to determine the identity of the oil
decomposition products. Blank GC analyses are run with 1,0-mL portions of clean solvent. The samples
for this analyses were generated as part of the total residue analysis (step 10) described in Section 4.4
Standard solutions were of selected volatile compounds in compressor oil. These solutions cover
the linear range of the GC and were used to quantitate the amount of volatile compounds in the oil of the
samples. In addition, the standards help to illustrate the impurities in the sample by providing a pattern
against which to match the samples.
The standards were run every day of analysis, and a control chart was maintained daily to help
identify problems in the stability of the instrument. Using a GC/MS, attempts were made to identify the
impurities that were observed. They could not be identified as no single impurity existed in a sufficient
concentration to be identifiable. The gas chromatograph with a flame ionization detector used for the
screening analyses is several orders of magnitude more sensitive than the GC/MS.
21
-------
4.6 FREE HALIDE
The method for analysis of the free halides is by ion chromatography (IC). A clean, dry sample
cylinder was weighed and attached to the sample container. The valve on the sample container was
cracked to purge the sample line, and then the fittings were tightened. The sample container was
positioned to sample the liquid phase, and a 100-g portion was transferred to the sample cylinder. The
gas phase was then bubbled through approximately 75 mL of buffered IC eluent solution
(0.0056 M NaHC03, 0.0045 M Na2C03) at a rate of 0.1 -1 Umin until the bubbling stopped. The sample
cylinder was reweighed to determine the amount of sample added.
The buffered eluent was brought to a volume of 100 mL, and a sample of it was injected onto the
IC, a Dionex 2110i. Standards for chloride and fluoride were made in IC eluent by using analytical-grade
sodium salts of both chlorine and fluorine to make a 1,000-ppm stock which was diluted with deionized
water as necessary. Spiked samples of known concentrations were made separately from the original
chemicals and analyzed with each batch of samples. All spikes satisfied the data quality
objectives (DQO).
Audit samples of chloride (HCI) in CFC-12 were provided but the results of these are shown in
Section 6, Table 6-7. The results confirmed those found for acidity. It showed that the analytical method
was satisfactory, it could detect less than 1 ppm chloride. The sampling method, however, could
qualitatively show the presence of chloride at levels between 5 and 10 ppm.
22
-------
SECTION 5
RESULTS
Two hundred and twenty-seven (227) automobile air conditioners were sampled as part of this
program. They were sampled in four different geographic locations and at service centers in different
cities within these locations. Table 3-1 in Section 3 lists the number of cars sampled in each geographic
location. The cities in which samples were taken at each geographic location are given in Table 5-1
which also shows the number of service centers in each city at which samples were taken. The
information on the types of automobiles, their mileage, and types of service they required are shown in
Table $-2.
All the samples were shipped to the EPA's Air and Energy Engineering Research Laboratory in
Research Triangle Park, North Carolina, for chemical analysis.
The results of the analyses for acidity, chloride, fluoride, and refrigerant purity are given in
Table 5-3. As can be seen, none of these parameters (except for two refrigerant purity samples which
are discussed below) exceeded the levels specified for new CFC-12. The fact that no acid nor inorganic
chloride and fluoride was found in any of the samples was of initial concern. The analytical methods were
checked using specially prepared samples of known concentrations of these analytes and were found to
be sensitive to the following levels:
Acid1 ppm as HCI
Inorganic chloride500 ppb
Inorganic fluoride500 ppb
Table 5-1. Sampling Locations
Geographic area
Location of service center
Number of sampling sites
at each location
Gulf Coast
Harlingen, Texas
1
Houston, Texas
3
Northeast
Cockeysville, Maryland
1
Midwest
Angola, Indiana
1
Montpelier, Ohio
1
Mountain
Denver, Colorado
2
23
-------
SAMPLE
NUMBER LOCATION
VEHICLE
DESCRIPTION
TX BLNK
15
HARUNGEN.TX
44
HOUSTON, TX
48
HARLINGEN.TX
53
HARLINGEN, TX
81
HOUSTON, TX
106
HOUSTON, TX
133
HOUSTON, TX
140
HOUSTON, TX
152
RTI
178
RTI
215
HARLINGEN.TX
226
RTI
248
RTI
269
HARLINGEN.TX
293
HARLINGEN.TX
TX OK COMPRESSORS
2
HOUSTON, TX
4
HOUSTON, TX
5
HOUSTON, TX
8
HOUSTON, TX
9
HOUSTON, TX
11
HOUSTON, TX
12
HOUSTON. TX
13
HOUSTON, TX
14
HOUSTON, TX
16
HOUSTON, TX
17
HOUSTON. TX
19
HOUSTON, TX
22
HOUSTON, TX
24
HOUSTON, TX
28
HOUSTON, TX
29
HARLINGEN, TX
31
HARLINGEN. TX
33
HOUSTON. TX
36
HOUSTON, TX
38
HOUSTON, TX
42
HOUSTON. TX
43
HARLINGEN, TX
46
HOUSTON. TX
NOT USED
NOT USED
FIELD BLANK
FIELD BLANK
NOT USED
FIELD BLANK
FIELD BLANK
NOT USED
QA SAMPLE FROM RTI
OA SAMPLE FROM RTI
TRIP BLANK
QA SAMPLE FROM RTI
QA SAMPLE FROM RTI
TRIP BLANK
TRIP BLANK
1988 FORD TAURUS
1984 BUICK PARK AVE
1988 FORDTHUNDERBIRD
1974 FORD VAN
1988 FORD MUSTANG
1968 FORD THUNDERBIRD
1987 TOYOTA COROLLA
1988 FORD ESCORT
1988 PONTIAC GRAND AM
1988 PONTIAC GRAND AM
1987TOYOTA COROLLA
1988 CHEVY CAVALIER
1972 DODGE DART SWINGER
1988 OLDS CUTLASS CIERA
1988 FORD THUNDERBIRD
1983 CADILLAC FLEETWOOD
1984 CHEVY CELEBRITY
(YR?) FORD ESCORT
1988 FORD TEMPO
1985 CHEVY SILVERADO
1987 MITSUBISH MIRAGE
1977 LINCOLN TOWN CAR
1968 MITSUBISHI MIRAGE
TABLE 5-2. VEHICLE INFORMATION
LISTED BY SAMPLING LOCATION
VEHICLE VEHICLE
MILEAGE A/C SYSTEM
PREVIOUS A/C SERVICE
AND COMMENTS
SERVICE
THIS VISIT
SAMPLE LINE RESTRICTED
NO NITROGEN IN SAMPLE CONTAINER
NEW REFRIGERANT
NEW REFRIGERANT
NO NITFIOGEN IN SAMPLE CONTAINER
NO NITROGEN IN SAMPLE CONTAINER
FD
13,755
FACTORY
GM
56,707
FACTORY
(SAMPLE FROM LO-SIDE)
FD
17.500
FACTORY
FD
191,231
FACTORY
FD
10,438
FACTORY 220 AND 9 ARE FROM SAME CAR, 220 USED FIRST
FD
15.095
FACTORY
IM
23.462
FACTORY
FD
17,385
FACTORY
GM
8,901
FACTORY
GM
14,768
FACTORY
IM
16,612
FACTORY
GM
12,228
FACTORY
CH
159,047
FACTORY
REPLACE EXPANSION VALVE AND DRYER
GM
11,038
FACTORY
FD
15,902
FACTORY
GM
80,087
FACTORY
GM
44,345
FACTORY
(SAMPLE FROM LO-SIDE)
FD
14,601
FACTORY
FD
9,729
FACTORY
GM
76,574
FACTORY
(SAMPLE FROM LO-SIDE)
IM
12,291
FACTORY
FD
69,947
FACTORY
NEW EVAPORATOR, 1986
IM
18,423
FACTORY
continued
-------
TABLE 5-2. VEHICLE INFORMATION (continued)
USTED BY SAMPLING LOCATION
SAMPLE
VEHICLE
VEHICLE
VEHICLE
PREVIOUS A/C SERVICE
SERVICE
NUMBER LOCATION
DESCRIPTION
MILEAGE A/C SYSTEM
AND COMMENTS
THIS VISIT
TX OK COMPRESSORS (continued)
47
HARLINGEN, TX
1978 FORD VAN CLUB WAGON
FD
111,990
ADD ON. 1984
RECHARGE, 1986
52
HARLINGEN, TX
1984 TOYOTA CELICA GT
IM
63,502
FACTORY
(SAMPLE VALVE SEALED W/ PUTTY)
LEAKATCOND
54
HARLINGEN, TX
1983 LINCOLN MARK VI
FD
69,420
FACTORY
PLUGGED
ORIFICE TUBE
55
HARLINGEN, TX
1986 PLYMOUTH VOYAGER VAN
CH
26.411
FACTORY
LEAKATCOND
56
HOUSTON, TX
(YR?) TORD TEMPO
FD
10,433
FACTORY
58
HARLINGEN, TX
1982 JEEP WAGONEER
CH
24,451
FACTORY
59
HOUSTON, TX
1988 FORD ESCORT
FD
12,682
FACTORY
60
HOUSTON, TX
1988 FORD TAURUS
FD
13,624
FACTORY
62
HOUSTON. TX
1987 BUICK SOMERSET
GM
21,595
FACTORY
63
HOUSTON, TX
1981 TOYOTA COROLLA
IM
49,797
FACTORY
EXPANSION VALVE
64
HOUSTON, TX
1988 FORD TEMPO
FD
12,646
FACTORY
65
HOUSTON, TX
1987 TOYOTA COROLLA
IM
19.544
FACTORY
67
HARUNGEN, TX
1988 PONTIAC 6000
GM
6,450
FACTORY
6S
HARLINGEN, TX
1987 CHEVY SCOTTSDALE TRUCK
GM
18,027
FACTORY
70
HOUSTON, TX
1982 FORD BRONCO
FD
92,606
FACTORY
REPLACE COMP, 77000 (LO-SIDE SAMPLE)
71
HARLINGEN, TX
1980 CADILLAC LIMO
GM
115,391
FACTORY
REBUILT COMP, EVAP(?)
LEAK REPAIR
72
HOUSTON, TX
1988 TOYOTA CAMRY
IM
10,147
FACTORY
73
HARUNGEN, TX
1972 FORD LTD
FD
128,000
FACTORY
NEW EVAP AND ACCUM, 1985
75
HOUSTON, TX
1988 FORD THUNDERBIRD
FD
13,358
FACTORY
76
HARLINGEN, TX
1981 FORD LTD
FD
87,479
FACTORY
REPLACE EVAP. 75000
77
HOUSTON, TX
1988 TOYOTA CAMRY
IM
12,029
FACTORY
7a
HARLINGEN, TX
1987 FORD LTD WAGON
FD
28.009
FACTORY
(SAMPLE FROM LOSIDE)
80
HOUSTON, TX
1988 FORD ESCORT
FD
11,019
FACTORY
83
HOUSTON, TX
1988 MERCURY SABLE
FD
14,768
FACTORY
84
HOUSTON, TX
1988 CHEVY CORSICA
GM
21,336
FACTORY
85
HOUSTON, TX
1987 MERCURY TRACER
FD
29,692
FACTORY
86
HARLINGEN, TX
1982 GMC HIGH SIERRA
GM
72,136
FACTORY
REPLACE EVAP, COMP, RECEIVER (?)
87
HOUSTON, TX
1988 FORD TAURUS
FD
12,100
FACTORY
89
HOUSTON. TX
1988 OLDS CUTLASS
GM
14,504
FACTORY
92
HARUNGEN, TX
1977 CHEVY SILVERADO TRUCK
GM
58,360
FACTORY
NEW EVAP AND ACCUM. 54000
93
HOUSTON, TX
1988 TOYOTA COROLLA
IM
16,879
FACTORY
94
HOUSTON, TX
1988 MERCURY SABLE
FD
17,520
FACTORY
95
HOUSTON, TX
1988 FORD THUNDERBIRD
FD
16,905
FACTORY
100
HARUNGEN, TX
1957 FORD FAIRLANE
FD
128,141
ADDON, 1985
(7000 MILES)
101
HOUSTON. TX
1988 FORD THUNDERBIRD
FD
12,701
FACTORY
103
HOUSTON, TX
1988 FORD THUNDERBIRD
FD
11,040
FACTORY
104
HOUSTON. TX
1985 TOYOTA PICKUP
IM
61,138
ADDON
105
HOUSTON. TX
1981 CADILLAC SEVILLE DIESEL
GM
81,412
FACTORY
109
HOUSTON. TX
1987 OLDS CUTLASS CIERA
GM
23.732
FACTORY
RECHARGE AT 3254
111
HOUSTON, TX
1985 RAM CHARGER TRUCK
CH
66,385
FACTORY
HIGH SIDE HOSE LEAK
REPLACE HOSE
115
HOUSTON. TX
1976 CHEVY SUBURBAN
GM
299,001
FACTORY
REPLACE COMP AND ACCUM. 270000
NOT COOLING
117
HOUSTON, TX
1987 MERCURY TOPAZ
FD
15,203
FACTORY
119
HOUSTON, TX
1988 FORD TEMPO
FD
15,194
FACTORY
continued
-------
TABLE 5-2. VEHICLE INFORMATION (continued)
USTED BY SAMPLING LOCATION
fo
-------
TABLE 5-2. VEHICLE INFORMATION (continued)
USTED BY SAMPLING LOCATION
SAMPLE
VEHICLE
VEHICLE
VEHICLE
NUMBER LOCATION
DESCRIPTION
MILEAGE
A/C SYSTEM
MD BLANK
137
BALTIMORE. MD
NOT USED
153
BALTIMORE. MD
FIELD BLANK
FACTORY
159
BALTIMORE, MD
FIELD BUNK
FACTORY
MDOK COMPRESSORS
135
BALTIMORE, MD
1983 FORD MUSTANG
FD
88.476
FACTORY
136
BALTIMORE, MD
1985 DODGE COLT
CH
47,494
FACTORY
141
BALTIMORE, MD
1987 FORD RANGER XLT
FD
17,358
FACTORY
142
BALTIMORE, MD
1984 FORD CROWN VICTORIA SW
FD
84,977
FACTORY
144
BALTIMORE, MD
1986 CHEVY CELEBRITY GL
GM
34,911
FACTORY
146
BALTIMORE, MD
1984 PLYMOUTH HORIZON
CH
45.551
FACTORY
151
BALTIMORE. MD
1986 PONTIAC 6000 LE
GM
45,619
FACTORY
155
BALTIMORE. MD
1987 FORD TEMPO
FD
23,452
FACTORY
157
BALTIMORE, MD
1983 NISSAN PULSAR
IM
68,779
FACTORY
158
BALTIMORE, MD
1987 MERCURY COUGAR
FD
21,687
FACTORY
161
BALTIMORE, MD
1987 MAZDA 626
IM
15,867
FACTORY
164
BALTIMORE. MD
1985 HONDA ACCORD
IM
54,083
FACTORY
165
BALTIMORE, MD
1983 CHEVY CITATION
GM
64,624
FACTORY
176
BALTIMORE, MD
1984 CHEVY CAMARO
GM
60,780
FACTORY
186
BALTIMORE, MD
1988 CHEVY CORSICA
GM
10.857
FACTORY
187
BALTIMORE. MD
1983 SUBARU GL WAGON
IM
52,161
FACTORY
196
BALTIMORE. MD
1987 DODGE DAKOTA
CH
15,188
FACTORY
198
BALTIMORE, MD
1987 DODGE SHADOW
CH
15,360
FACTORY
208
BALTIMORE, MD
1985 TOYOTA CAMRY
IM
62,552
FACTORY
214
BALTIMORE, MD
1982 CHRYSLER TOWN-COUNTRY
CH
98,506
FACTORY
219
BALTIMORE. MD
1985 BUICK LESABRE
GM
13,045
FACTORY
223
BALTIMORE. MD
1984 CHEVY CAVALIER
GN
86,445
FACTORY
225
BALTIMORE, MD
1987 FORD TEMPO
FD
26,027
FACTORY
234
BALTIMORE, MD
1987 NISSAN PULSAR NX
IM
16,929
FACTORY
242
BALTIMORE. MD
1986 CHEVY S-10 BLAZER
GM
45,262
FACTORY
243
BALTIMORE, MD
1986 CHRYSLER RELIANT WAGON
CH
28,540
FACTORY
245
BALTIMORE. MD
1985 CHEVY CAMARO
GM
41,724
FACTORY
247
BALTIMORE, MD
1978 MERCURY ZEPHYR
FD
61.469
ADDON (YR?)
256
BALTIMORE, MD
1987 NISSAN MAXIMA
IM
12.876
FACTORY
257
BALTIMORE, MD
1984 PONTIAC GRAND PRIX
GM
75,745
FACTORY
263
BALTIMORE, MD
1987 CHEVY Z24
GM
18,880
FACTORY
265
BALTIMORE, MD
1988 JEEP CHEROKEE
CH
14,143
FACTORY
273
BALTIMORE. MD
1986 CHRYSLER LASER
CH
31,539
FACTORY
275
BALTIMORE. MD
1983 TOYOTA COROLLA
IM
76,197
FACTORY
277
BALTIMORE, MD
1984 CHEVY CELEBRITY
GM
53,902
FACTORY
281
BALTIMORE, MD
1987 OLDS CUTLASS CIERA
GM
16,123
FACTORY
PREVIOUS A/C SERVICE
AND COMMENTS
SERVICE
THIS VISIT
NO NITROGEN IN SAMPLE CONTAINER
COMPRESSOR AT 2.000 MILES
continued
-------
TABLE 5-2. VEHICLE INFORMATION (continued)
USTED BY SAMPLING LOCATION
SAMPLE VEHICLE VEHICLE VEHICLE PREVIOUS A/C SERVICE SERVICE
NUMBER LOCATION DESCRIPTION MILEAGE A/C SYSTEM AND COMMENTS THIS VISIT
MD FAILED COMPRESSORS
147
BALTIMORE, MD
1984 PONTIAC SUNBIRD
GM
62,881
FACTORY
FAILED COMP
255
BALTIMORE. MD
1981 FORD FAIRMONT
TD
26,461
FACTORY
FAILED COMP
291
BALTIMORE, MD
1982 HONDA ACCORD
IM
77.367
FACTORY
FAILED COMP
295
BALTIMORE. MD
1981 CHEVYIMPALA
GM
64,628
FACTORY
FAILED COMP
296
BALTIMORE. MD
1983 CADILLAC DEVILLE
GM
73,997
FACTORY
FAILED COMP
OH BLANK
148
ANGOLA, IN
FIELD BLANK
166
MONTPELIER. OH
FIELD BLANK
FACTORY
254
MONTPEUER, OH
FIELD BLANK
FACTORY
268
ANGOLA, IN
FIELD BLANK
FACTORY
OH OK COMPRESORS
143
ANGOLA, IN
1980 BUICK CENTURY
GM
82,000
FACTORY
154
MONTPEUER, OH
1982 CHEVY SILVERADO
GM
55,819
FACTORY
156
MONTPELIER, OH
1986 CHEVROLET S 10
GM
13,083
FACTORY
163
ANGOLA, IN
1987 OLDS CALAIS
GM
32,436
FACTORY
171
MONTPELIER. OH
1987 MERCURY GRAND MARQUIS
FD
19,801
FACTORY
172
MONTPELIER, OH
1981 FORD MUSTANG
FD
41,975
FACTORY"
177
ANGOLA, IN
1984 LINCOLN TOWN CAR
FD
51,613
FACTORY
181
MONTPELIER, OH
1985 MERCURY TOPAZ
FD
24,482
FACTORY
182
MONTPELIER. OH
1987 MERCURY TRACER
FD
19,238
FACTORY
183
ANGOLA, IN
1983 DODGE CARAVAN
CH
52,215
FACTORY
185
MONTPELIER, OH
1985 PONTIAC 6000
GM
27.352
FACTORY
189
ANGOLA, IN
1985 FORD TEMPO
FD
72,401
FACTORY
191
MONTPELIER. OH
1982 CHRYSLER LEBARON
CH
78,178
FACTORY
194
MONTPELIER, OH
1978 FORD ECONOUNE VAN
FD
40,666
FACTORY
203
MONTPELIER. OH
1982 VW RABBIT
IM
106,765
FACTORY
204
MONTPELIER, OH
1983 DODGE DIPLOMAT
CH
113,987
FACTORY
207
MONTPELIER. OH
1986 VW J ETTA
IM
72,683
FACTORY
209
ANGOLA, IN
1983 FORD LTD CROWN VICTORIA
FD
42,442
FACTORY
212
MONTPELIER. OH
1988 HONDA CIVIC
IM
14,173
FACTORY
216
MONTPELIER. OH
1986 OLDS CALAIS
GM
22,810
FACTORY
217
ANGOLA, IN
1982 OLDS 98
GM
74,667
FACTORY
21B
MONTPEUER. OH
1977 CHRYSLER CORDOBA
CH
78,944
FACTORY
232
ANGOLA, IN
1974 OLDS CUTLESS
GM
76.962
FACTORY
233
MONTPELIER. OH
1985 FORD LTD WAGON
FD
76,767
FACTORY
237
MONTPELIER. OH
1987 CHEVROLET CAVAUER
GM
21,241
FACTORY
239
ANGOLA. IN
1984 PONTIAC SUNBIRD
GM
90,006
FACTORY
246
MONTPELIER. OH
t985 TOYOTA COROLLA
IM
19,905
FACTORY
259
MONTPELIER. OH
1988 CHEVY ASTRO VAN
GM
11.337
FACTORY
262
MONTPELIER, OH
1973 CADILLAC DEVILLE
GM
74.387
FACTORY
264
MONTPEUER. OH
1987 CHEVY ASTRO VAN
GM
10,883
FACTORY
continued
-------
TABLE 5-2. VEHICLE INFORMATION (continued)
USTED BY SAMPLING LOCATION
SAMPLE
NUMBER
LOCATION
VEHICLE
DESCRIPTION
VEHICLE VEHICLE
MILEAGE A/C SYSTEM
PREVIOUS A/C SERVICE
AND COMMENTS
SERVICE
THIS VISIT
I OK COMPRESORS (continued)
266
ANGOLA, IN
1981 MERCURY BROUGHAN
FD
66,567
FACTORY
271
ANGOLA, IN
1984 TOYOTO CELICA GT
IM
60.951
FACTORY
278
MONTPELIER, OH
1986 OLDS 98
GM
47,583
FACTORY
285
MONTPELIER, OH
1988 CHEVY CAVAUER
GM
13.328
FACTORY
288
MONTPELIER, OH
1987 CHRYSLER NEW YORKER
CH
12,128
FACTORY
289
MONTPELIER, OH
1987 MERCURY TOPAZ
FD
19,419
FACTORY
297
MONTPELIER, OH
1985 CADILLAC EL DORADO
GM
64.979
FACTORY
298
ANGOLA, IN
1981 CHEVROLET MONTE CARLO
GM
82,892
FACTORY
299
MONTPELIER, OH
1987 BUICK SKYLARK
GM
26.359
FACTORY
301
MONTPELIER, OH
1987 FORD TEMPO GL
FD
17,297
FACTORY
CO BLANK
199
DENVER. CO
FIELD BLANK
213
DENVER, CO
FIELD BLANK
CO OK COMPRESSORS
139
DENVER, CO
1984 FORD BRONCO XLS
FD
74,973
FACTORY
144
DENVER. CO
ISUZU TROOPER II
IM
18,562
FACTORY
145
DENVER, CO
1979 MAZDA GLC WAGON
IM
150,479
FACTORY
149
DENVER, CO
1986 FORD F150 PA/
FD
35.701
FACTORY
167
DENVER, CO
1980 TOYOTA CELICA SUPRA
IM
78,986
FACTORY
168
DENVER. CO
1985 FORD LTD CROWN VICTORIA
FD
92,347
FACTORY
169
DENVER, CO
1988 NISSAN MAXIMA
IM
11,865
FACTORY
173
DENVER, CO
1971 FORD LTD
FD
122,131
FACTORY
175
DENVER, CO
1987 PLYMOUTH VOYAGER VAN
CH
12,117
FACTORY
179
DENVER, CO
1979 FORD ECONDINE 150 VAN
FD
64,076
FACTORY
192
DENVER, CO
1981 FORD GRANADA
FD
41,463
FACTORY
193
DENVER, CO
1988 FORD TAURUS
FD
12,312
FACTORY
195
DENVER, CO
1979 BMW 320I
IM
122,881
FACTORY
201
DENVER. CO
1988 PLYMOUTH GRAND VOYAGER
CH
11,592
FACTORY
202
DENVER, CO
1982 BUICK SKYLARK LIMITED
GM
98,263
FACTORY
205
DENVER, CO
1934 SUBARU GL WAGON
IM
78,792
FACTORY
206
DENVER, CO
1988 TOYOTA CAMRY LE WAGON
IM
17,834
FACTORY
211
DENVER. CO
1985 NISSAN MAXIMA
IM
64,293
FACTORY
221
DENVER. CO
1986 FORD BRONCO II
FD
18,106
FACTORY
222
DENVER. CO
1988 NISSAN SENTFIA
IM
13,303
FACTORY
224
DENVER. CO
1988 DODGE RELIANT K
CH
19,798
FACTORY
231
DENVER. CO
1987 TOYOTA COROLLA FX
IM
11,222
FACTORY
235
DENVER. CO
1977 CHEVY CAPRICE LANDAU
GM
117,095
FACTORY
236
DENVER. CO
1988 PLYMOUTH RELIANT K
CH
17,234
FACTORY
238
DENVER. CO
CHEVY BLAZER 4X4
GM
73.035
FACTORY
241
DENVER. CO
1978 TOYOTA CELICA LIFTBACK
IM
80.466
FACTORY
244
DENVER. CO
1987 FORD TAURUS
FD
15.587
FACTORY
VAPOR ONLY? HIGH SIDE RESTRICTION
CAR OVERHEATS. NO BELT ON AC
REPLACE COMP FRONT SEAL COMP REBUILD
continued
-------
TABLE 5-2. VEHICLE INFORMATION (concluded)
USTED BY SAMPLING LOCATION
SAMPLE
VEHICLE
VEHICLE
VEHICLE
PREVIOUS A/C SERVICE
SERVICE
NUMBER
LOCATION
DESCRIPTION
MILEAGE A/C SYSTEM
ANDCOMMENTS
THIS VISIT
CO OK COMPRESSORS (continued)
249
DENVER, CO
CHEVY BLAZER 4X4
GM
24,505
FACTORY
251
DENVER, CO
19B7 NISSAN SENTRA
IM
19,843
FACTORY
252
DENVER, CO
1984 DODGE POWER RAM P/V
CH
54,059
FACTORY
253
DENVER. CO
1976 FORD ELITE
FD
77,599
FACTORY
258
DENVER. CO
1978 BUICK LIMITED
GM
101.021
FACTORY
261
DENVER, CO
1985 CHEVY CITATION II
GM
36,184
FACTORY
272
DENVER, CO
1986 HONDA CIVIC
IM
57,888
FACTORY
274
DENVER, CO
1985 FORD BRONCO II XLT
FD
60,227
FACTORY
284
DENVER, CO
1987 FORD BRONCO 4X4
FD
25,660
FACTORY
286
DENVER, CO
1985 CHEVY SILVERADO P/V
GM
56,652
FACTORY
207
DENVER, CO
1983 CHRYSLER ECLASS NY
CH
70,160
FACTORY
294
DENVER, CO
1977 PONTIAC GRAND PRIX
GM
50,922
FACTORY
CO FAILED COMPRESSORSFACTORY
188 DENVER, CO 1988 CHEVY CUTLESS CIERA
GM
33,324 FACTORY REPLACED ACCUMULATOR SYSTEM PLUGGED
FAILED COMP
-------
Table 5-3. Results of Analysis: Acids, Halides, Relrlgerant Purily
All samples
Analyte Concentration Analytical method
Acidity
<5 ppm as HCI
Titration
Chloride ion
<0.5 ppm
IC
Fluoride ion
<0.9 ppm
IC
Refrigerant contaminant level
£0 5%
GC/FID
all but two samples
#262
2%
#218
5%
Extensive testing of standards and check samples showed a high level of reliability and
reproducibility in the analytical method. A series of cocktails of propionic acid (CH3CH2COOH) was
prepared at concentrations ranging from 1 ppm to 100 ppm. It was found that above approximately
5 ppm the analyses matched the expected value. Below this limit, the analyses correlated very poorly
with the actual concentrations of the acid. Thus, the laboratory technique accurately confirmed that acid
in the CFC was less than 1 ppm and in the air conditioner, less than 5 ppm. Section 6 discusses this
QA/QC effort in greater detail.
The finding on the lack of acid is good news for the program. Because the sampling system was
selected to closely duplicate the recover system that will be used to recycle the refrigerant, there is every
reason to believe that no significant quantity of acid will be removed from the MAC during
recycling/servicing. Furthermore, any acid present during normal capture and recycling of the refrigerant
can be removed by the recycling equipment. Acid can be neutralized by contact with metal components
or by the use of special absorbents which can be incorporated into the recycling equipment. Based on
this laboratory analysis, acidity in recycled refrigerants will not be a problem if recycling equipment is
properly designed.
The lack of chloride or fluoride ions in the samples further reinlorces the above conclusion.
These ions would typically form by hydrolysis of the CFC-12, forming hydrochloric or hydrofluoric acid.
The lack of these ions, coupled with the high purity seen in the CFC-12 itself, indicates that refrigerant
breakdown does not occur under the conditions encountered in an automobile air conditioner.
The refrigerant purity analyses also showed very little contamination. Only two samples of
refrigerant out of the 227 tested were found to contain more than 0.5 percent HCFC-22. Sample Number
262 measured 2% and Sample number 218 measured 5% HCFC-22. Both cars were from Montpelier,
OH and had similar mileages (74,387 and 78,947 respectively). Sample 262 came from a 1973 model
whose air conditioner had been recharged at approximately 60,000. Sample 218 came from a 1977
vehicle. Both vehicles' air conditioners were performing to the satisfaction of the owners and were not
being serviced at the time the samples were taken. The source of this HCFC-22 contamination is
uncertain. The most likely source is that at some point in their service life the air conditioners in both of
these cars had been recharged by an inexperienced person with HCFC-22.
The results of the moisture, residue, and residue purity analyses are given in Table 5-4. Because
these results are crucial to the success of subsequent programs, they were discussed in Section 2.
31
-------
Table 5-4. Results ot Analyses for Moisture. Residue, and Residue Purity
SAMPLE RESIDUE
NUMBER MILEAGE MOISTURE RESIDUE PURITY
(ppm) (ppm) (ppm) LOCATION
OK COMPRESSORS
67
6,450
<10
26
<100
TX
14
8,901
16
22535
100
TX
36
9,729
52
109
6200
TX
72
10,147
207
69230
4800
TX
56
10,433
18
11294
800
TX
9
10,438
****
49509
3200
TX
220
10,438
21
288
500
TX
130
10,443
18
1430
1800
TX
186
10,857
13
2549
1300
MD
264
10,883
<10
4915
300
OH
80
11,019
14
-16
<100
TX
24
11,038
<10
19328
900
TX
103
11,040
***
***
TX
231
11,222
44
20886
7600
CO
259
11,337
23
745
1600
OH
201
11,592
20
42710
<100
CO
169
11,865
67
10027
200
CO
77
12,029
91863
2400
TX
87
12,100
<10
29
1000
TX
175
12,117
22
22541
100
CO
288
12,128
48
90
3200
OH
210
12,207
14
3207
600
TX
19
12,228
21
6100
TX
42
12,291
49
11963
3600
TX
193
12,312
34
4475
900
CO
64
12,646
51
13
TX
59
12,682
12
27
<100
TX
101
12,701
115
303
9900
TX
150
12,800
<10
14
<100
TX
256
12,876
29
14682
800
MD
219
13,045
13
12
4400
MD
156
13,083
16
17076
100
OH
222
13,303
42
100072
300
CO
285
13,328
40
22072
<100
OH
75
13,358
15
4994
600
TX
60
13,624
20
4572
1800
TX
2
13,755
15
129
3500
TX
280
13,770
21
67324
3300
TX
265
14,143
17
6506
300
MD
212
14,173
127
18703
500
OH
89
14,504
<10
20247
155
TX
33
14,601
28
272
1000
TX
16
14,768
<10
18341
400
TX
83
14,768
24
7739
1800
TX
290
15,025
31
717
6800
TX
11
15,095
12
22
<100
TX
196
15,188
70
41472
<100
MD
119
15,194
41
175
<100
TX
117
15,203
16
8318
1000
TX
198
15,360
70
45603
200
MD
continued
32
-------
Table 5-4. Results of Analyses lor Moisture, Residue, and Residue Purity
SAMPLE
RESIDUE
NUMBER
MILEAGE
MOISTURE
RESIDUE
PURITY
(ppm)
(ppm)
(ppm)
LOCAT
SSORS (continued)
244
15,587
57
11097
800
CO
125
15,812
14
51222
1600
TX
161
15,867
64
3304
7800
MD
28
15,902
20
29
500
TX
281
16,123
11
43576
200
MD
17
16,612
30
31007
TX
93
16,879
13
37337
6000
TX
95
16,905
23
42
4500
TX
234
16,929
34
61898
650
MD
236
17,234
11
40113
<100
CO
301
17,297
45
139549
6000
OH
141
17,358
34
1349
2500
MD
13
17,385
27
33
500
TX
5
17,500
110
99
2700
TX
94
17,520
68
849
1200
TX
206
17,834
68
35735
5900
CO
128
17,942
20
-2
2700
TX
68
18,027
<10
3234
100
TX
221
18,106
21
76
10600
CO
46
18,423
43
10681
1600
TX
144
18,562
CO
263
18,880
<10
14166
300
MD
120
18,921
48
140
4000
TX
182
19,238
36
4457
5300
OH
132
19,294
<10
51
<100
TX
289
19,419
49
193
3000
OH
134
19,461
41
3617
<100
TX
65
19,544
42
48220
3900
TX
224
19,798
<10
6504
<100
CO
171
19,801
65
21
<100
OH
251
19,843
68
59521
100
CO
246
19,905
40
10991
2900
OH
237
21,241
46
9026
500
OH
84
21,336
37
111
<100
TX
62
21,595
17
2080
<100
TX
158
21,687
34
2
<100
MD
216
22,810
<10
10214
300
OH
155
23,452
24
23
6500
MD
12
23,462
49
16869
3700
TX
109
23,732
16
63
2300
TX
58
24,451
744
34164
6000
TX
181
24,482
40
59
200
OH
249
24,505
15
86650
800
CO
284
25,660
24
2937
1300
CO
225
26,027
27
25
2000
MD
299
26,359
<10
884
100
OH
55
26,411
1002
50325
<100
TX
185
27,352
10
77
2000
OH
78
28,009
54
1332
100
TX
continued
33
-------
Table 5-4. Results of Analyses for Moisture, Residue, and I
SAMPLE
RESIDUE
NUMBER
MILEAGE
MOISTURE
RESIDUE
PURITY
(ppm)
(ppm)
(ppm)
LOCAT
SSORS (continued)
243
28,540
60
22963
300
MD
85
29,692
<10
148
<100
TX
273
31,539
48
3024
200
MD
163
32,436
37
7088
200
OH
144
34,911
<10
4739
600
MD
149
35,701
44
216
4100
CO
261
36,184
36
230
1100
CO
122
37,956
<10
636
<100
TX
194
40,666
58
79
6600
OH
192
41,463
29
7131
300
CO
245
41,724
23
28953
400
MD
172
41,975
52
2796
1400
OH
209
42,442
41
62
2000
OH
121
42,616
69
92
<100
TX
31
44,345
21
312
<100
TX
242
45,262
11
22200
100
MD
146
45,551
76
114
3700
MD
151
45,619
13
19386
400
MD
136
47,494
58
14420
1700
MD
278
47,583
12
12358
2100
OH
63
49,797
45
8773
200
TX
294
50,922
46
367
800
CO
177
51,613
59
51
3800
OH
187
52,161
413
43629
MD
183
52,215
51
42270
400
OH
277
53,902
<10
15443
1100
MD
252
54,059
60
2345
200
CO
164
54,083
117
25828
700
MD
154
55,819
75
8555
300
OH
286
56,652
<10
6912
1800
CO
4
56,707
<10
21
<100
TX
272
57,888
36
72
500
CO
92
58,360
74
36
TX
274
60,227
18
CO
176
60,780
84
16303
600
MD
271
60,951
42
3879
1300
OH
104
61,138
73
88718
400
TX
247
61,469
85
12080
400
MD
208
62,552
54
38667
3200
MD
52
63,502
63
48
<100
TX
179
64,076
32
3453
100
CO
211
64,293
28
41110
<100
CO
165
64,624
20
50
500
MD
297
64,979
224
10360
700
OH
123
65,354
54
39049
<100
TX
266
66,567
35
51
2200
OH
111
68,385
63
5060
<100
TX
157
68,779
63
107950
200
MD
54
69,420
81
1003
<100
TX
continued
34
-------
Table 5-4. Results of Analyses for Moisture, Residue, and Residue Purity
SAMPLE
RESIDUE
NUMBER
MILEAGE
MOISTURE
RESIDUE
PURITY
(ppm)
(ppm)
(ppm)
LOCAT
SSORS (continued)
43
69,947
50
58262
300
TX
287
70,160
48
400
2100
CO
86
72,136
13
6251
<100
TX
189
72,401
60
630
1100
OH
207
72,683
36
1096
1200
OH
238
73,035
61
35283
100
CO
262
74,387
66
1489
<100
OH
217
74,667
<10
99
9200
OH
139
74,973
33
16
2600
CO
257
75,745
11
671
200
MD
275
76,197
*
*** ~
**«#
MD
38
76,574
18
31
300
TX
233
76,767
29
193
7300
OH
232
76,962
26
32
9700
OH
253
77,599
50
1963
200
CO
191
78,178
64
50805
500
OH
205
78,792
32
40376
<100
CO
218
78,944
78
52185
400
OH
167
78,986
38
181814
700
CO
29
80,087
11
31
<100
TX
241
80,466
«**
398
2900
CO
105
81,412
42
20
<100
TX
143
82,000
*
*
OH
298
82,892
54
76695
100
OH
142
84,977
53
130
200
MD
223
86,445
13
6115
900
MD
76
87,479
13
36
<100
TX
135
88,476
77
48
300
MD
239
90,006
42
512
900
OH
168
92,347
13
227
<100
CO
70
92,606
24
5722
121
TX
202
98,263
<10
44
1100
CO
214
.98,506
87
35854
700
MD
258
101,021
45
36
2900
CO
260
103,374
86
27
<100
TX
203
106,765
161
56
500
OH
170
107,000
12
8
TX
47
111,990
78
36
<100
TX
204
113,987
79
104481
600
OH
71
115,391
48
4508
<100
TX
235
117,095
47
1182
<100
CO
126
118,377
755
TX
250
119,258
<10
60
<100
TX
173
122,131
285
700
CO
195
122,881
45
166
4700
CO
73
128,000
77
10892
<100
TX
100
128,141
78
18313
200
TX
160
134,643
173
125
<100
TX
145
150,479
96
139543
700
CO
22
159,047
77
28
200
TX
continued
35
-------
Table 5-4. Results of Analyses for Moisture, Residue, and Residue Purity
SAMPLE RESIDUE
NUMBER MILEAGE MOISTURE RESIDUE PURITY
(ppm) (ppm) (ppm) LOCATION
OK COMPRESSORS (continued)
8 191,231 57 50 3500 TX
115 299,001 72 13 <100 TX
FAILED COMPRESSORS
6
18
21
25
26
32
45
66
88
90
91
98
99
110
112
116
118
124
129
200
230
270
147
255
291
295
296
188
71,522
54,491
34,916
129,340
56,182
47,208
45,579
81,368
53,187
42,351
27,784
53,157
31,269
44,509
71,883
68,798
126,253
77,123
56,814
41,500
92,693
43,356
62,881
26,461
77,367
64,628
73,997
33,324
<10
*4**
<10
22
57
»*«
12
72
65
<10
87
19
515
15
22
<10
71
<10
51
54
110
**»»
67
15
54
19
32
143
51
18
29
4804
42
35
12213
13061
356
188
-142
28
191
40
17
11
20
83
1062
22717
«***
896
15696
59
5488
27646
17239
<100
300
200
<100
200
<100
<100
<100
<100
<100
100
1900
1900
1500
1900
1300
300
<100
100
1700
1100
2100
5700
900
500
300
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
MD
MD
MD
MD
MD
CO
BLANKS
15 TX
44 TX
48 <10 83 <100 TX
53 15 70 <100 TX
81 TX
106 19 50 <100 TX
133 14 35 <100 TX
140 TX
152 37 34 <100 TX
178 10 19 <100 TX
215 <10 66 <100 TX
continued
-------
Table 5-4. Results of Analyses for Moisture, Residue, and Residue Purity
SAMPLE RESIDUE
NUMBER MILEAGE MOISTURE RESIDUE PURITY
(ppm) (ppm) (ppm) LOCATION
BLANKS (continued)
226
12
2563
900
TX
248
<10
408
<100
TX
269
55
55
<100
TX
293
<10
165
TX
137
****
«** +
MD
153
11
93
200
MD
159
<10
66
200
MD
148
<10
98
<100
OH
166
65
96
2100
OH
254
14
254
<100
OH
268
14
44
500
OH
199
<10
191
<100
CO
213
13
122
<100
CO
*" = Not Enough Sample lo Conduct Test
Blank entries indicate that the sample leaked in transit or that it was otherwise lost during
analysis or handling
37
-------
SECTION 6
QUALITY ASSURANCE/QUALITY CONTROL
6.1 DATA QUALITY OBJECTIVES FOR CRITICAL MEASUREMENTS
Critical measurements for this study included moisture content, acidity, halide (chloride and
fluoride) ion concentration, purity of the CFC, and purity of the residue. The techniques used to measure
these criteria were moisture by K-F titration, acidity by titration with a base solution to a visual endpoint,
chloride ion by IC, quantitative purity by gas chromatograph with a flame ionization detector, and residue
by gravimetric analysis.
Data quality objectives were established on completeness, precision, and accuracy. See
Table 6-1 for the measurements that were applied to the DQO's. A number of the critical measurements
(i.e., moisture, chloride, acidity, CFC purity) were found to routinely approach the detection limit of the
respective procedure. The preparation of artificial samples was biased towards the region from the
detection limit to the quantifiable limit so that these numbers were adequately defined.
Completeness was calculated for each critical measurement. Completeness is defined as:
Percent Complete = 100 x (No. valid samples/No. needed for statistical power).
A sample was not analyzed only if the automobile air conditioner had lost so much of its
refrigerant charge that the sample obtained was too small to allow the analysis to be performed.
6.2 SAMPLE CONTAINER PREPARATION
When they were received, the sample containers were each assigned a unique sample number
that was painted onto the containers with indelible paint. The sample containers were new and
specifically manufactured for this purpose. They consisted of a (approximately) 1 -gal steel vessel
equipped with a 2-way valve suitable for refrigerant 12 (R-12). The vessel is rated for a minimum of
250 psi, and the valve has a safety release which opens at this pressure. The valve opening is equipped
with a metal screw cap to protect it during shipping and handling and which contains an "O" ring to
provide a secondary seal to reduce the likelihood of sample loss. The manufacturer has assured Acurex
that the interior of the vessel and valve is free of dirt and oil.
Upon receipt, 1 in 10 of these containers was tested to ensure that they did not contain any
impurities. The testing was performed by filling each of the containers selected with new R-12 which had
been previously analyzed. Each container was then allowed to stand overnight, and the contents were
analyzed for the four parameters listed in Table 3-2. Since more than 10 percent of these containers
measured a level of impurity greater than that shown in Table 3-1, all the containers were cleaned with
FC-113 and rechecked prior to being shipped.
6.3 DATA REDUCTION METHODS
The results of the analysis for each sample were reported on the analysis sheet for the method.
The analysis sheets were entered into the log books and on the computer at the end of each day of
38
-------
Table 6-1. Quality Assurance/Quality ControlData Quality Objective
Parameter
(method)
Dectection
limit
Mass
(Gravimetric)
Chloride
(Ion
chromatography)
Moisture
(Karl Fischer
titration)
Acidity
(Base titration)
0.1 g
10ppm
10 ppm
0.1 ppm
(HCI)
Completeness
Actual Objective
Reference
standards
95
94
95
95
90
90
90
90
Analytical Precision
(percent RSD)
Actual Objective
Standard weight 0.2
Synthetic standards 1.9
prepared in extraction
media
Synthetic standards 13
prepared in anhydrous
methanol
Synthetic standards 1.3
prepared in de-ionized
water
±5
<20
<20
<20
Accuracy
Actual Objective
0.25
2.8
10.5
3.2
±5
±20
±20
±20
CFC purity
(GC/FID)
Oil purity
(GC/FID)
Residue
(Gravimetric)
0.1%
100 ppm
(per compound)
0.1 mg
94
94
94
90 Synthetic standards 0.9 <20
of CFC
90 Synthetic standards 8.9 <20
prepared in FC-113
90 Synthetic standards 0.4 ±5
prepared in FC-113
2.6
17
0.8
±20
±20
±5
-------
analysis. All data were given to the lead chemist to track samples in the laboratory. The lead chemist
checked random calculations for each analysis. The lead chemist also tracked ail numbers and reported
the data to the project manager.
6-3-1 Moisture Content
The Karl Fischer coulimetric analyzer automatically reported the moisture content in micrograms.
Calculation by the following formula gave the amount of moisture as ppm water in the refrigerant:
ngr
ppm Moisture
Wi W2
where:
|xgr = micrograms reported by Karl Fischer Titrator
W1 = weight ol full cylinder (g)
W2 = weight of emptied cylinder (g)
6.3.2 Acidity or Acid Number
The acid number is determined by the following formula as "ppm acidity as HCI":
ppm - (mL KOH x Normality KOH x 36,460) / (A - B)
where:
A = initial weight of cylinder
B = final weight of cylinder
36,460 = 1000 x molecular weight of HCI
6-3.3 Purity of the Refrigerant
The purity of the refrigerant was calculated by use of the response factors for the components
found. The general formula is:
percent purity = 100- 100 (amount of impurity)/amount sampled
6.3.4 Total Residue
The total residue is a gravimetric determination with the correction factor accounting for the
amount removed for "purity by GC." The formula is:
grams residue/gram sample = (WD2 - WD1)(2)/(WC2 - WC1)
where:
WD1 = weight of empty dish
WD2 = weight of sample and dish
40
-------
WC2 = weight of cylinder and sample
WCi = weight of empty cylinder
2 = correction factor for solvent removed from residue
6.3.5 Purity of Residue bv Gas Chromatoaraph
The amount of residue was calculated by the gravimetric method. The GC analysis was used to
quantitate contaminants in the oil. The response factor for each component was used to determine the
level of contamination. The general formula is:
Response factor = amount/area unit
6.3.6 Free Halides
The halides were calculated by use of the Response Factor for each component. The general
formula is:
Response factor = 100 * (amount/area units)/(sample weight)
6.3.7 Quality Assurance Objectives for Clean Sample Containers
Table 6-2 gives the QA objectives for the cleaned sample containers, these objectives were met.
6.4 CALCULATION OF DATA QUALITY INDICATORS
6.4.1 Accuracy
Synthetic standards were prepared in the appropriate medium and submitted to the analytical
queue. Accuracy was calculated in terms of the deviance from the known value.
deviance = calculated - known
percent accuracy = 100 x deviance/known
6.4.2 Precision
Synthetic standards were prepared in the appropriate medium and submitted to the analytical
queue in replicate. Precision was calculated as percent relative standard deviation (requiring at least
3 measurements) in terms of the standard deviation and the calculated mean value.
(
mean >
n
£ calculated,
i=1
a = standard deviation =
/n
L caiculatedj2
i=1
n
I calculated]
i=l
^2
/n
n-1
percent relative standard deviation = (100 x cr)/mean
41
-------
Table 6-2. QA Objectives for Clean Sample Containers (ppm)
Federal
Maximum
specification
Analyte
value
BB-F-1421A
Water
10
10
Acidity
10
*
High boiling impurities
100
100
Chloride ion
10
<20 ppm by AgNOa solution
Purity
10
* No specification given.
6.4.3 Completeness
At each stage, the calculation for completeness was based upon the number of samples
attempted and the number of samples successfully completing that stage. An unsuccessful sample is
defined as one which is irretrievably lost or fails the associated quality control checks. Samples which
initially fail a stage but are successfully repeated were counted as successful attempts.
percent completion = 100 x (No. valid samples/No. needed for statistical significance)
Outliers were reported and used in all calculations unless they were shown to be statistically
invalid. All values found were reported.
6.5 CORRECTIVE ACTION PROCEDURES
Procedures for corrective action are straightforward. Sampling containers unable to hold
pressure upon testing were rejected prior to shipment to the service centers. Should greater than
10 percent of the sample containers tested prior to shipping fail the QC check, all the sample containers
were cleaned and 1 in 10 was rechecked. Prior to being filled at the service centers, containers would be
rejected by the service technician if, when opened, they failed to exhibit the characteristic hiss of escaping
nitrogen gas that would validate positive pressure.
During the analysis phase, the cylinders used in the analyses were recleaned should they fail QC
checks prior to use. Should a spiked sample fail to lall within the DQO listed in Table 6-2 after reanalysis,
standards for the analysis in question were reprepared and samples were rechecked. Corrective action
for instrumentation, if necessary, was made according to manufacturer's specifications.
The lead chemist takes corrective action if any analysis should fail to meet the DQO. In addition,
corrective action was taken in response to a QA audit. All matters requiring corrective action were
reported to the lead chemist.
6.6 SYSTEM AND PERFORMANCE AUDITS
During the project performance, the Acurex Quality Assurance Officer (QAO) conducted a
systems audit designed to assess compliance with the Quality Assurance Project Plan (QAPP). Items
evaluated included sampling procedures, sample tracking, QC checks of sample cylinders, calibration of
analysis techniques, frequency of spikes and replicates, and correspondence of data with established
DQO. The results of the analyses of the audit samples are given in Tables 6-3, 6-4,6-5,6-6, and 6-7. As
can be seen, few problems were encountered in any of the procedures with the exception of the acidity
and CI.
42
-------
Table 6-3. Results of First Performance Evaluation Audit
Cylinder
No.
Analytical
Test
Acurex
Result
True
Amount
Relative %
Difference
Detection
Limit (ppm)
152 Total CFC
Retrige rant M ass 1037.8 g
Acidity as HCI <1.0 ppm
Moisture 37.0 ppm
Total Residue 34.0 ppm
CFC Refrigerant
Purity 99.99%
178 Total CFC
Refrigerant Mass 1,412.1 g
Acidity as HCI <1.0 ppm
Moisture 10.0 ppm
Residue Purity 19.0 ppm
CFC Refrigerant
Purity >99.99%
226 Total CFC
Refrigerant Mass 651.6 g
Acidity as HCI 11.3 ppm
Moisture 12.0 ppm
Total Residue 2,563.0 ppm
CFC Refrigerant
Purity >99.99%
248 Total CFC
Refrigerant Mass 1,200.0 g
Acidity as HCI <1.0 ppm
Moisture <10.0 ppm
Total Residue 408.0 ppm
CFC Refrigerant
Purity >99.99%
1080 g
0
0
0
>99.99%
1,438 g
2.25 ppm
0
489.4
(hydrocarbon
compounds)
>99.99%
687 g
9.41 ppm
0
2.042 ppm
>99.99%
1,236 g
0
0
283.4 ppm
>99.99%
-3.9
-I.8
Acurex did
not detect
-96.1
-5.2
+20.1
+25.6
0
-2.9
0
0
+44.0
1.0
10.0
a{100.0)
1.0
10.0
3(100.0)
1.0
10.0
a(iOO.O)
1.0
10.0
a(i00.0)
a not actually a detection limit; however, each clean audit cylinder contains <100 ppm residue.
43
-------
Table 6-4, Results of Second Performance Evaluation Audit
Acurex
Sample
Analysis
Acurex
True
Percent
Accuracy
Detection
No.
Performed
Result
Concentration
Error
Objective
Limit
37
Acid
0
2.08 ppm
=20%
1 ppm as HCI
as HCI
Residue
Purity C12
*120-125 ppm
*43 ppm
+179%
NONE
100 ppm in
C15
*135-140 ppm
*43 ppm
+213%
NONE
compressor
oil
267
Acid
10 ppm as HCI
11 ppm as HCI
-9.1%
=20%
1 ppm as HCI
Residue
Purity C12
*650-700 ppm
*282 ppm
+130%
NONE
100 ppm in
Ci5
*650-700 ppm
*281 ppm
+131%
NONE
compressor
oil
ppm for this sample is calculated as percentage of total Refrigerant, not as a percentage of
compressor oil.
Table 6-5. Results of Third Performance Evaluation Audit
Acurex
Sample
Analysis
Acurex
True
Percent
Accuracy
Detection
No.
Performed
Result
Concentration
Error
Objective
Limit
276
Total Residue
728 ppm
864 ppm
-15.7
±20%a
100 ppm
Acid
Not detected
4.14 ppm
+20%
1 ppm as HCI
(as HCI)
Residue
Purityb
C12
989 ppm
796 ppm
+24.3
not
100 ppm
C15
1,058 ppm
795 ppm
+33.1
es-
100 ppm
Total
2,117 ppm
1,590 ppm
+33.1
tablished
100 ppm
279
Total Residue
438 ppm
481 ppm
-8.9
+20%a
100 ppm
Acid
Not detected
1.15 ppm
±20%
1 ppm as HCI
(as HCI)
Residue
Purityb
C12
62 ppm
79.6 ppm
-22.1
not
100 ppm
C15C
204 ppm
79 .5 ppm
+ 157
es-
100 ppm
Total
293 ppm
159 ppm
+84.3
tablished
100 ppm
a objective has been modified from original QAPP.
b Residue Purity is calculated as ppm of compressor oil.
c Interference seen in Acurex result.
44
-------
Table 6-6. Results of Fourth Performance Evaluation Audit
Acurex
Sample
Analysis
Acurex
True
Percent
Accuracy
Detection
No.
Performed
Result
Concentration
Error
Objective
Limit
229 Acid
0.17 ppm 4.17 ppm -96%
(as HCI)
±20%
1 ppm
197 Acid
#1 2.34 5.16 ppm -54.7%
#2 3.74* (as HCI) -27.5%
±20%
1 ppm
2nd analysis performed approximately two hours after the 1 st analysis.
Table 6-7. Results of the Fifth Performance Evaluation Audit,
Free Chloride Audit Samples
Acurex Analysis
Sample No.
and Analysis Acurex True
Description Performed Results Concentration
Percent
Error
Acurex
Accuracy
Objective
226 (audit
cylinder w/Refrigerant
and NaCi soln.)
Free 0.018 ppm 4.43 ppm
Chloride
-99.6
±20%
CL8258-2 (NaCI
(NaCI soln. in
water)
Free 6.59 ppm 6.23 ppm
Chloride
IC
+5.8
RTI Analysis
Sample No.
and
Description
Analysis RTI
Performed Results
True
Concentration
Percent
Error
178 (audit
cylinder w/Refrigerant
and NaCI soln.)
Free 0.58 ppm
Chloride
4.20 ppm
-86.2
Duplicate Sample
from 178
Free 0.50
Chloride ppm
4.20
ppm
-88.1
CL8258-2
(NaCI soln. in
water)
Free 6.29
Chloride ppm
IC
6.23
ppm
+1.0
45
-------
Audit samples of a known quantity of propionic acid and a high-boiling hydrocarbon in CFC-113
along with a sample ol the pure hydrocarbon were provided by an outside organization designated by the
AEERL QAO. These samples were analyzed for compliance with the established DQO using the acidity
and purity analyses.
For purposes of comparison, the third column in Table 6-2 gives the Federal specifications for
refrigerant. Note that the specifications for chloride ion do not identify what "none" means. Ten parts per
million for the chloride ion was selected as a reliable lower limit of delegability. See Section 4 for a
discussion of analytical methods.
As part of the Q/A Audit for the program, several samples of known acidity were prepared and
analyzed following the standard procedures. The samples were prepared in the following manner. A
known quantity of propionic acid (CH3CH2COOH) was dissolved in CFC-113. The valve was removed
from a cleaned sampling container, and a known amount of this solution was pipetted into it. The valve
was replaced, and the container was filled with a known amount of CFC-12.
The containers were submitted as normal samples to the laboratory which analyzed the contents
in the same manner as for the normal samples. The results of these analyses are shown in Table 6-8.
The results indicated that for low levels of acid (below approximately 10 ppm), recoveries were
nol as expected. Possible reasons for this were problems with the following:
1. The titration
2. The recovery of the acid from the sample container
3. Reaction or deposition of the acid with the sample container
Each of these possible reasons was investigated in turn. First, the analytical procedure (the
titration) was checked. Solutions of propionic acid in water and in CFC-113 were prepared and titrated by
the method used for the sample. The method was reproducible to levels of 2.6 micrograms of HCI and
5.3 micrograms of propionic acid; both solutions had standard deviations of less than 4 percent at these
levels. This level is equivalent to about 0.1 ppm in a 100 g sample. Whether the solution was water or
CFC-113 did not effect the recovery of the acids. Conversations with personnel at DuPont and Allied
Signal confirmed that the method was valid to 10 micrograms of HCI, although neither company had
attempted to titrate an organic acid. The titration was apparently acceptable. The procedures being used
were reviewed again with Allied and DuPont who confirmed that they were similar. The only exceptions
between our procedures and those used by allied and DuPont were:
a. We rinsed all lines used to transfer the samples and they normally do not require that this be
done.
b. We used a different indicator although neither group considered this to be significant.
The next possibility was that the sample was not being delivered from the sample container to the
solution. For several samples (No. 276, No. 279, and No. 37), the valves were removed, and the interior
of the sample container was rinsed with water to remove the acid. Titration of the resulting solution
showed no acid present.
The remaining possibility was that the acid reacted with or deposited onto the container wall. It
was noted that when a higher level of compressor oil was present in the containers, the recovery was
better. The laboratory made a sample by removing the valve and adding approximately 15 g of
compressor oil to the container. The container was rolled and shaken to coat the walls, then the acid was
46
t
-------
Table 6-8. Results of Analysis of QA Audit Samples
Sample No.
ppm Prepared
ppm Measured
178
226
37
267
276
279
197
197
229
2.37
9.92
2.08
11.0
4.14
1.15
5.16
5.16
4.17
<1
11.3
<1
9.52
<1
<1
2.35
3.74
<1
added, the valve was replaced, and the container was charged with CFC-12 to make a standard of
5.3 ppm acid. Triplicate analysis of the sample taken at successive two-hour intervals were 5.04 ppm,
5.2 ppm, and 5.41 ppm acid present.
This information was passed on to the QA contractor who submitted two additional samples, No.
197, and No. 229. For sample No. 197, 9 mL of oil were added to the container and mixed, and the acid
added; tor No. 229,10 mL of oil in CFC-113 were added before the acid. As is noted in Table 6-8, the
recovery was better for No. 197. These results indicate that recovery is improved if the samples are
mixed in oil (as are the samples taken from the field) rather than in a solvent not present in the refrigerant
samples, such as the CFC-113 used for the audit samples.
General Motors is doing similar research with its fleet of methanol-powered cars, and a phone
conversation with GM's chemist indicated that the recovery of chloride from their sample containers was
also poor (10 percent or less). The chloride was added as NaCI in water. They used stainless steel
rather than the mild steel containers which were used for this program. It is well known that most types of
stainless steel chemically react with chloride ions.
Some minor changes were made in the analytical procedure for acid number on the basis of this
audit. The first change is to rinse the delivery lines from the sample container to the flask with the
solution used in the titration. Second, the solution would be neutralized to the endpoint before the sample
is bubbled through it. Previously three blanks were run each day, and the average value was subtracted
from each sample titration. The field data was gathered following the revised procedure.
These tests indicate that the sampling and analytical procedure produces reliable results if the
acidity is greater than about 10 ppm. Below this level, quantitation is uncertain; however, if acidity were
present at a level above 5 ppm, its presence would be observed qualitatively.
6.7 REPLICATE ANALYSES AND STANDARDS
Because of the relatively small size of the samples removed from the auto air conditioners,
1,000 grams, maximum, compared to the amount of sample required to conduct the analyses it was not
possible to conduct replicates on many samples. As discussed earlier in the report, the only analyses
which proved significant was the moisture determination. As a result, when a sample of sufficient size
was found, it was reanalyzed for moisture. The results of these replicate analyses are given in Table 6-9.
As can be seen, the reproducibility of the analyses was, with the exception of a few samples, excellent. It
is surmised that the difference that did occur in several samples was due to inhomogeneity in the
samples. The water in a sampling container distributes itself between the liquid phase, vapor phase, and
free water floating on the CFC-12 liquid. This phenomenon introduces a potential uncertainty to the
47
-------
Table 6-9. Results of Replicate Moisture Analyses
ppm Moisture
Date
Sample Number
Repl. 1
Repl. 2
6/02
102 (QA)
33.9
36.5
6/01
79 (QA)
7.0 '
7.8
7/07
92
79.6
73.5
7/29
281
22.1
11.4
8/10
117
15.8
58.7
8/18
239
110.4
42.0
9/02
216
7.9
7.4
7/12
160
173.1
58.8
actual value of moisture measured for any one car; however, the overall effect on the analyses of the
CFC-12 from a large number of cars would be small.
In addition to performing replicates, the Karl-Fisher apparatus was tested daily against injections
of known volumes of water. These tests proved the equipment to be highly reliable. As a result, in the
interests of brevity, the daily results are not presented here. Table 6-10 gives the results of one analysis
per week over the weeks that the program was being performed, for illustrative purposes. As can be
seen, the equipment was tested over a wide range of moisture levels daily and it performed well.
6.8 INTERNAL QA AUDIT
A Technical Systems Audit of the CFC project was conducted on July 11, 1988, during the early
stages of the work by Mr. Kevin R. Bruce, the ERO Quality Assurance Officer. Mr. Bruce examined
several areas of the project activities and evaluated them for compliance with the pre-approved QAPP.
The following items were among those audited:
Sample tracking; sample custody sheets for randomly picked samples were requested and
examined and all tracking information was in place.
Balance QC checks; standard weights were presented and their use documented.
Laboratory blanks; blank runs for titrative analyses were performed and documented.
IC QC checks; spiked samples were run on the IC and results documented.
Calibration standards; calculation of concentration of standards made from reagents were
checked and verified.
Sampling methods; the sampling system was visually examined for obvious leakage.
Staff training; questioning of staff to ensure adequate job knowledge and training verified
that this was the case.
The audit results indicated that sampling and analysis methods did indeed conform to prescribed
techniques given in the QAPP.
48
-------
Table 6-10. Standards and Replicates (Moisture)
Amount Injected Amount Measured
Date (Mg/H20) (Mg/H20)
5/24 0 0
1.6 0.906
2.1 1.104
7.1 6.798
21.7 21.5
6/1 0.0 0.0
1.0 0.825
4.9 4.542
4.7 4.099
10.0 9.918
6/8 0.0 0.000
1.05 0.901
5.5 5.074
10.80 10.78
6/14 0 0.000
1.01 0.763
5.18 5.213
9.87 9.950
12.20 12.66
6/30 0 0.000
0.98 0.778
4.89 4.651
6.37 6.398
10.31 9.831
7/6 0.0 0.000
1.26 0.735
4.39 4.819
9.86 9.935
0.0 0.011
7/12 0.0 0.000
1.57 0.957
0.95 0.745
4.91 4.859
8.26 7.888
7/21 0.0 0.007
0.56 0.465
3.78 3.594
14.58 13.51
(continued)
49
-------
Table 6-10. Standards and Replicates (Moisture) (concluded)
Amount Injected
Amount Measured
Date
(Mg/H20)
(Mg/H20)
7/28
0.0
0.000
0.91
1.290
1.04
0.823
6.27
6.105
9.46
9.957
8/9
0.0
0.0002
1.12
1.124
4.90
4.487
9.77
10.30
8/16
0.0
0.0000
0.84
1.001
5.79
5.910
9.09
8.972
8/23
0.0
0.0000
0.81
0.6127
4.59
4.238
9.71
9.321
8/30
0.0
0.000
0.81
0.6803
4.87
4.730
10.80
10.27
9/6
0.0
0.000
1.0
1.025
5.89
6.307
10.92
10.80
9/12
0.0
0.000
1.17
1.323
6.43
6.591
9.97
9.809
50
-------
SECTION 7
REFERENCES
1. Federal Specification for "Fluorocarbon Refrigerants": BB-F-1421B, March 5,1982 (commonly
referred to as the "Mil-Spec").
2. ASTM D-3401-85: "Standard Test Method lor Water in Halogenated Organic Solvents and Their
Admixtures '' May 31,1985.
3. ASTM 2989-86 (reapproved 1981): "Standard Test Method for Acidity Alkalinity of Halogenated
Organic Solvents and Their Admixtures." April 25,1986.
4. Allied Chemical Company. Approved Procedure GP-GEN-2A: "Determination of Acidity In
Fluorocarbons." 1983.
5. DuPont Method F3200.037.01 CW(P): "Determination of Boiling Range, Residue, Particulates"
Octobers, 1985.
6. Parmelee, H.M. "Solubility of Air in Freon-12 and Freon-22." Refrigerating Engineering.
June 1951.
7. ASTM Standard E700-79: "Water in Gases Using Karl Fischer Reagent." September 28,1979.
8. ASTM Standard D-1533-86: "Waste in Insulating Liquid (Karl Fischer Reaction Method)."
January 31,1986.
51
-------
APPENDIX A
DETAILED ANALYTICAL METHODS
52
-------
OPERATING PROCEDURE FOR PURITY OF RESIDUE BY GAS CHROMATOGRAPH
The method used to check the purity of the high boiling residue is a variation of the Total
Chromatographable Organics (TCO) analysis. The changes trom the TCO include changing the solvent
from dichloromethane to FC-113, possible analysis by GCMS to determine the identity of the oil
decomposition products, and an adjustment to the GC conditions. The changes in the GC conditions are
listed below. Blank GC analyses will be run with 1.0-mL portions of clean solvent.
The samples are generated in the total residue analysis and are prepared for GC analysis at that
time. The samples are taken to 100 mL with FC-113, and 50 mL of the sample is taken for purity analysis
by GC. The sample is concentrated by evaporating a 50-mL portion of the solvent to 10 mL. The sample
is then made to volume with FC-113 to a concentration that is compatible to both GC and GC/MS
analysis.
Standard solutions of hydrocarbons in compressor oil will be prepared to cover the linear range of
the GC. The standards will help to illustrate the impurities in the sample by providing a pattern against
which to match the samples.
Standards of the impurities identified by mass spectrometer will be prepared to verify the identity
of the impurities. Calculations for the method will consist of ratioing the areas of the oil decomposition
peaks to the areas of the compressor oil peaks. The impurities will be expressed as parts per million
(ppm) of the compressor oil.
Below are the changes in the GC conditions from the TCO conditions:
Initial temperature: 60 °C
Hold for 3 min
Ramp temperature at 8 °C/min to a final temperature of 300 °C and hold for 45 min
Integrate the entire run after the solvent peak.
53
-------
OPERATING PROCEDURE FOR MOISTURE DETERMINATION IN CFC-12
1.0 PROCEDURAL
1.1 SCOPE AND APPLICATION
This method describes the procedure for the determination ol moisture content in samples of
R-12 (dichlorodifluoromethane) automobile air conditioning refrigerant. It also includes the preparation,
sampling, and quality assurance procedures involved with the analysis. The experiment uses the Karl-
Fischer (K-F) coulomatic titrimeter, which determines parts per million (ppm) moisture content of a given
sample. These data will be useful in determining the level of contamination likely to be found in the R-12
from various groups of cars.
1.2 SUMMARY OF METHOD
Calibration of the K-F titrimeter is determined by injecting a weighed amount of deionized (Dl)
water to the nearest 0.1 mg. Four points of approximately 0 mg, 1 mg, 5 mg, and 10 mg will be used. A
final calibration point will be run at the end of each day. Coulamat conditioner will be used for a control
sample. The analysis procedures are as follows.
To analyze a sample, first weigh sample containers and then suspend a sample container above
the K-F titrimeter, connect the needle valve (liquid phase is sampled from the bottom) to the injection
tube, and start the titration run. (Note: PERSIST must be set to 180 s, 3 min.) Carefully open the needle
valve and allow the refrigerant to bubble into the solution (Coulomat/A). At the end of persist time,
remove the sample container, plug the injection tube to prevent contamination, and weigh the container to
nearest 0.1 g. The difference will be input into the K-F titrimeter at the end of the titration run, which will
automatically display and print out ppm of moisture.
Moisture in the refrigerant is determined titrimetrically using the K-F titrimeter. The method is
based on the oxidation of sulfur dioxide by iodine in the presence of water to form sulfuric acid. The end
point Is achieved when free iodine appears and remains in the titration run solution. The end point is
sensed eledrometrically and Is achieved instrumentally with the K-F titrimeter. For more information,
consult the K-F titrimeter manual.
1.3 INTERFERENCES
1.3.1 Coulomat/A and /C
1.3.1.1 Coulomat/A
Coulomat/A is not easily contaminated by other organics of refrigerant such as oil, grease, and
dirt. However, it is recommended by the manufacture to replace the Coulomat/A solution periodically if
problems occur.
54
-------
1.3.1.2 Coulomat/C
Coulomat/C will easily be depleted if, during the set-up stage of operation when the cell is wet,
the Iodine solution is not used to dry the cell. Using this solution saves on time and titrant. Also, if the A
and C liquid levels reach an equilibrium, this will quickly diminish the titrant's ability.
1.3.1.3 Moisture Contamination
Any water will change the readout. During the entry of the refrigerant, the injection line may
freeze up. Since this is very difficult to prevent, it is advisable to check the O-ring at the base of the line.
If this is kept tight and secure, any moisture buildup will be kept out of the reaction vessel. At the end of
each run, use a cork stopper in the top of the injection line to prevent contamination.
1.3.2 Troubleshooting the Instrument
For more information on care and operation of the K-F titrimeter, refer to the manufacturer's
manual,
1.3.3 Calibration
Blanks and standards are run daily with the K-F titrimeter to ensure that it is properly calibrated.
Any other equipment such as the syringe or containers are to be kept clean and dry to discourage
contamination of any sample.
1.4 APPARATUS
1.4.1 Injection Tube
Th1 e injection tube takes the sample into the cell. The O-ring in the tube housing must be kept
secure to avoid moisture contamination. At the end of the tube is a bubbler, which should not be allowed
to drop lower than the detector probe. The tube is always stoppered when not in use.
1.4.2 Detector Probe
The detector probe, which plugs into the back of the unit, should be well clear ol the stir bar to
avoid damage. An error signal should be displayed if there are problems with this probe.
1.4.3 Cathode Cell
The C cell or cathode cell, which contains the Coulomat/C solution, has an internal cathode and
an external anode. The titrant is made and measured electronically through the fritted disc at the bottom
of the C cell. If the fritted disc becomes dirty or plugged, cleaning may necessary. The anode and
cathode plug into the back of the unit.
1.4.4 Anode Cell
The A cell or anode cell contains the Coulomat/A solution, and all other probes and leads are
entered into the system through the top of this cell. Be sure to secure all parts to avoid any
contamination.
1.4.5 Injection Port
The septum port on top of the A cell is used to inject calibration standards into the system.
55
-------
1.4.6 Exhaust Port
The exhaust port on the top o1 the A cell is used to vent any fumes into the exhaust hood with a
rubber hose. Care should be taken when initially opening the sample container to avoid excessive
overspill. Use of the rubber hose may prevent any personal harm in the event of overflow by the
Coulomat solution.
1-4.7 Syringes
A 30-ml and a 10-iu.l syringe will be needed to inject the iodine, conditioner, Dl water standard, or
the A and C solvents into the cells. It is advisable to assign one syringe for each of the four solutions
used and to store them in a desiccator between uses.
1.4.8 Stir bar
A Teflon-coated magnetic stir bar is needed in the bottom of the A cell to keep the titrant and
sample mixed evenly.
1.4.9 Titrimeter
The K-F titrimeter is a self-contained unit that has the stir plate, ROM program, and an internal
integrator for execution of analysis. Refer to the owner's manual for more operation information.
1.4.10 Sample Support
A laboratory stand with a cross bar and a clamp will be needed to suspend a container above the
K-F titrimeter for taking sample measurements.
1.4.11 Printer (optional)
A printer may be interfaced with the K-F titrimeter. The printer will make a hard copy of analysis
information. See owner's manual for setup and use.
1.4.12 Balances
Two balances will be needed. One should have an accuracy of ±0.1 g for sample weights This
balance may be interfaced with the K-F titrimeter to make data storage and operation more convenient,
but it is not necessary to the analysis. The other should have an accuracy of ±0.1 mg for standard
measurements.
1.5 REAGENTS AND MATERIALS
1.5.1 Anode Solution
Hydranai-Coulomat/A anode solution is the carrier for the moisture. All measurements and
additions are made in the A solution. The level of the A solution should at all times be higher than that of
the C solution.
1-5.2 Cathode Solution
Hydranal-Coulomat/C cathode solution is the titrant material. Extra care should be taken to
protect the C solution from prematurely exhausting itself. It may be necessary to remove the C solution
from the cell between uses.
56
-------
1,5.3 Drying the Cell
Hydranal Composite-5, the drying accelerator, is used to dry out the cell in the event ol excess
moisture during the setup stage. The use of this solution will expand the life expectancy of the
Coulomat/C solution. Caution should be taken not to over dry the cell by using too much iodine solution.
1.5.4 Conditioning the Cell
Coulomat conditioner solution is used in the event of excessive dryness, which usually results
from using too much of the iodine solution.
1.5.5 Standards
Dl water is used in the calibration of the K-F titrimeter. A1 CHiL syringe is gravimetrically
measured to ±0.1 mg and then injected into the titrimeter.
1.6 CALCULATIONS
The K-F titrimeter calculates the ppm content based on the input weight of the sample. The only
calculation is the determination of the sample weight. The formula is as follows:
Wl - W2 = Sco
where
Wl « Filled weight of the container with R-12
W2 = Weight of the container after the R-12 is added to the titrimeter
Sco = Sample weight analyzed
2.0 QUALITY ASSURANCE/QUALITY CONTROL (QA/QC)
2.1 PHASE ZERO (PRESCREEN)
Phase 0 QA involved a spot check of 30 canisters out of 300 for moisture content. Each canister
tested was filled with virgin R-12 and sampled according to method 1.2 to double check the
manufacturer's claim that all canisters are clean and dry.
2.2 PHASE ONE (DURING SAMPLING)
Phase I QA is a check of procedures and operations. Specific canisters will be filled in the field
with virgin R-12 that has known amounts of contamination. The laboratory technician will not be aware of
which canisters are the QA samples. This ensures proper operation of equipment by the laboratory
technician or shows any procedural problems.
2.3 CALIBRATION
During the operation of the K-F titrimeter, a 4-point standard curve will be prepared each day
before sample analysis. A blank will be run each day to confirm that there is no cross contamination. At
the end of each day, a final standard will be analyzed. Also, a QA chart will be maintained by running a
test on a known concentration of H2O in methanol.
57
/
-------
3.0 SAFETY
Laboratory personnel should be aware of facility safety rules and regulations. Gloves should be
worn when using solvents. Safety glasses and a laboratory coat are required. Fumes from the
refrigerant are nontoxic but will displace oxygen from the lungs. If exposed, personnel should go to a
well-ventilated area and bend over. The refrigerant Is heavier than air and will flow out. The solvents
contain methanol and chlorinated solvent and should be handled with respect. The solvent will be
disposed of by the recommended methods. A satellite container is provided in the laboratory.
Dichlorodifluoromethane (R-12) is a colorless gas with a characteristic ether-like odor at
concentrations above 20 percent by volume. It is incompatible with chemically active metals.
It is nontoxic but does displace oxygen. Concentrations above 10,000 ppm require a
contained breathing device. Rapid vaporization of liquid R-12 will freeze tissue. Because
containers are at about 50 psig pressure, care should be taken to release the pressure
slowly.
K-F reagents contain methanol, carbon tetrachloride, diethylamine, and sulfur dioxide, which
are RCRA-regulated solvents. The formula is proprietary. Disposal of this material should
be made according to regulations for halogenated solvents. The hazards for methanol
apply: flammability, pungent odor, incompatible with strong oxidizers, and poisonous. Skin
should be washed if exposed, and medical attention should be summoned if any is
swallowed. A hood, safety glasses, and latex gloves should be worn when handling these
chemicals.
58
-------
OPERATING PROCEDURE FOR RESIDUE ANALYSIS OF AUTO AIR CONDITIONING SAMPLES
1.0 PROCEDURE
1.1 SCOPE AND APPLICATION
The following method is a procedure for the determination of the amount of residue in a sample of
refrigerant from an automobile air conditioner. This information is useful in determining the amount of
high boiling compressor oil present in the refrigerant of a sampled automobile and will provide the
material for the purity sample. The purity sample will be analyzed by gas chromatograph (GC).
1.2 SUMMARY OF METHOD
A sample of the liquid phase of the refrigerant dichlorodiflouromethane (R-12) is taken by shaking
the sample container to mix the contents and then inverting the sample container so that the valve is on
the bottom. A sample cylinder is weighed to the nearest 0.1 g, the sample cylinder is connected to the
sample container, and an aliquot of the R-12 is transferred. The cylinder is reweighed to determine the
amount of sample taken. The cylinder is mounted in an upright position, and the gas phase of the sample
is slowly bled out. The cylinder is rinsed with CFC-113 (1,l,2-trichloro,1,2,2-lrifluoroethane) to remove
the residue. The CFC-113 solution is built to 100 mL in a volumetric flask. The solution is split into 50 mL
portions, and one portion is set aside for GC purity analysis. The remaining portion of the CFC-113 is
placed into a tared pan and evaporated. The residue is determined by the weight difference of the pan.
1.3 INTERFERENCES
There are no chemical interferences to the method; however, the CFC-113 will dissolve most
organic compounds. Thus, the cylinders must be very clean to avoid a false result in the residue analysis.
The GC residue sample is taken from this sample, and improper cleaning of the cylinder will result in
extraneous peaks In the GC chromatogram.
1.4 PERSONNEL REQUIREMENTS
Personnel must be familiar with standard laboratory techniques or be supervised by a chemist.
All personnel must be made aware of the safety guidelines for laboratory work and for the compounds
that will be used during the project.
1.5 FACILITIES AND LABORATORY REQUIREMENTS
The laboratory must be equipped with a fume hood to exhaust the R-12 and CFC-113 fumes.
Electrical outlets must be available (117 volt, 15 amp, minimum). Analytical balances are required for this
analysis. One balance must be accurate to +0.1 mg. The other balance must be accurate to +0.1 g.
1.6 SAFETY PRECAUTIONS
The CFC-113 and R-12 are not poisonous but will displace air in the lungs at concentrations
above 10,000 ppm. Exposure to the R-12 or CFC-113 may require relocating to a well-ventilated area
and bending over. The gas is heavier than air and will flow down out of the lungs.
59
-------
Safety glasses/goggles and gloves should be worn when handling the R-12 refrigerant because
of the possibility that flashing R-12 will freeze tissue. The container and cylinder are at about 50 psig
pressure. The valves should always be opened slowly to prevent flashing. The CFC-113 is an excellent
solvent and will dry oils rapidly from the skin. Contact should be avoided by wearing latex gloves.
1.7 APPARATUS AND MATERIAL
1.7.1 Glassware
Vials (GC autosampler with crimp tops); 4-dr size for storage of redissolved sample
Disposable pipets
1.7.2 Miscellaneous
Disposable pipet bulbs
Pasteur pipets
Aluminum pans
Hotplate
Labels
Digital scales
Desiccator
Oven
1.8 REAGENTS
CFC-113
1.9 CALIBRATION
Digital scales should be clean and tared to zero. S-class weights should be used to confirm
accuracy.
1.10 ANALYSIS PROCEDURE
The cylinder is filled with R-12 from the sample container in the usual manner. The cylinder is
then clamped vertically in the hood with the needle value pointing up. The needle value is then cracked,
and the gas is allowed to slowly escape. Gradually, the valve is fully opened, thus purging the cylinder of
gas. Each cylinder is then rinsed three times with approximately 25 mL of CFC-113 each time. The
volume of CFC-113 is brought to 100 mL in a volumetric flask, and the sample is split into 50-mL portions;
one of the portions is set aside for GC purity analysis. Addition of the CFC-113 takes place through an
attached Tygon tube which, in turn, is fastened tightly to a separatory funnel. The funnel, which has been
marked in increments of approximately 25 mL, is clamped in the hood. The second 50-mL portion is
poured into a tared (to the nearest 0.1 mg) aluminum pan and set on a hot plate to evaporate. The hot
plate must be adjusted such that the CFC-113 does not boil. The pans should always be handled with
gloves or tongs to avoid errors in weight from oils transferred from skin. When the liquid has
60
-------
disappeared, the pan is set in a 110 °F-oven for a baking time ot 30 min. after which time the pan is left to
cool in the desiccator to room temperature and finally reweighed.
1.11 CALCULATIONS
The following formula is used to determine the ppm of residue per sample:
grams residue (WD2 - WD:) x 2
= x 1Q6
grams sample {WC2 - WC^)
WDj = weight of empty dish
WD2 =» weight of sample and dish
WC2 = weight of cylinder and sample
WC1 = weight of empty cylinder
2.0 QUALITY CONTROL
This experiment adheres to a quality control procedure which includes analysis of "blank"
samples. Random containers are tested with pure refrigerant to determine initial contamination levels and
other potential problems. This analysis follows the same guidelines described above.
61
-------
OPERATING PROCEDURE FOR ACID NUMBER
1.0 PROCEDURAL ELEMENTS
1.1 Scope and Application
During the operation of an automobile air conditioner, any moisture or air that enters the system
may contribute to breakdown of the R-12 (dichlorodifluoromethane) refrigerant. The following method is
used to determine the acidity of R-12 refrigerant sampled from various automobile air conditioning
systems. The determination is made by titration to a visual endpoint using potassium hydroxide (KOH)
and Bromothymol blue indicator.
1.2 DEFINITIONS
1.2.1 Titrant
The titrant used will be =¦ 0.005N KOH. This will be standardized for each day's use with 0.005N
potassium biphthalate (KHP).
1.2.2 Container
Container refers to the can that is sent to the field for R-12 refrigerant sampling. Each container
is tagged with a number, which is retained throughout the sampling and analysis process, belore being
sent to the field. Each acid number test will be labeled with the following format: Container number - A
(e.g., 37-A refers to the acid number test of container number 37).
1.2.3 Parts Per Million
Part per million (ppm) is reported as HCI.
1.3 INTERFERENCES
1.3.1 Other Chemicals
CO2 will react with KOH, resulting in a reduction of normality in the titrant. Prior to each day's
analysis, the KOH will be standardized with 0.005N KHP, and the new normality for KOH will be used in
the calculations of the acid number. Also, the buret must be rinsed daily with 0.005N KOH to avoid both
contamination and change in titrant normality.
1.3.2 Glassware
The glassware should be washed with soap and water to be kept free of contamination, after
which it should be thoroughly rinsed with clean deionized (Dl) water. After the glassware is dried in an
oven at 100 °C for at least 15 min, the openings are covered with aluminum foil, and the glassware is
stored in the laboratory cabinet until ready for use.
62
-------
1.3.3 Introduction of Sample
The bubbling of the refrigerant into 150 mL of water should proceed at a moderate rate to avoid
any overspill of solution.
1-3.4 Digital Balance
The digital balance should be cleaned and zeroed before use. Periodically, the balance needs to
be checked with class-S weights to ensure proper operation. Calibration checks should be performed
once a week.
1.4 Apparatus and Glassware
200-mL graduated cylinder
Stir bars
Stir plate
Clamp
Ringstand
Plastic bubbling tube
Digital balance (±0.1 g)
5-mL buret (0.01 mL graduated)
2-mL pipet
500-mL and 1-L volumetric flasks
300-mL graduated cylinder
10 to 20 250-mL Erylenmeyer flasks
1.5 REAGENTS AND MATERIALS
Deionized (Dl) water
Bromothymol blue indicator
0.005N KOH (Potassium Hydroxide)
0.005N KHP (Potassium Biphthalate)
O.OO5NH2SO4
1.6 CALIBRATION
A 2-mL sample of 0.005N H2SO4/150 mL Dl water mixture is titrated daily as a control sample.
The KOH is standardized against a pure KHP solution of known normality.
63
-------
1.7 ANALYSIS PROCEDURES
After 150 mL of Dl water are placed in the flask, a stir bar and 8 drops of Bromothymol blue
indicator are added to the flask. The mixture is mixed at low speed on the stir plate. Then, KOH (the
exact normality of which has been determined during calibration) is slowly added from the buret just until
a green tinge remains in the mixture for 15 s without fading to yellow. The sample is taken directly from
the container. The container is weighed and the sample line, which contains a needle valve and bubbling
tube, is connected to the inverted container. The needle valve must be closed, and then the container
valve is opened. The needle valve is cracked open, and the refrigerant is allowed to bubble gently
through the water. After about 2-5 min, or approximately 50 g of refrigerant, the container valve is closed
and the refrigerant is allowed to flow from the line. The lines are rinsed into the solution using a Pasteur
pipette. The bubbling line is removed, and the container is reweighed to the nearest 0.1 g. The
Erlenmeyer flask with the Dl water/refrigerant mixture then undergoes the titration process. The solution
is titrated to the green endpoint, and the number of milliliters of KOH is determined (to the nearest
0.001 mL) from the buret markings and is recorded.
1.8 CALCULATIONS
The acid number in ppm as HCl is determined by the following formula:
(milliliters KOH) (normality KOH) (36, 460)
ppm as HCl =
(A-B)
where: A = initial weight of cylinder
B = final weight of cylinder
36,460 = 1000 x molecular weight ol HCl
2.0 SAFETY PROCEDURES
Laboratory personnel should be aware of facility safety rules and regulations. Gloves should be
worn when using solvents or handling R-12. Safety glasses and a laboratory coat are required. Fumes
from the refrigerant are nontoxic but will displace oxygen from the lungs. If exposed, personnel should go
to a well-ventilated area and bend over. The refrigerant is heavier than air and will flow out of the lungs.
Dichlorodilluoromethane R-12 is a colorless gas with a characteristic ether-like odor at concentrations
above 20 percent by volume. It is incompatible with chemically active metals. It is nontoxic but does
displace oxygen. Concentrations above 10,000 ppm require a contained breathing device. The
containers are at about 50 psig pressure. Because rapid vaporization of the R-12 liquid will freeze tissue,
care should be taken to release the pressure slowly. The low normality of the solutions used in the
analysis does not present a great health hazard; however, the preparation of the solutions requires
knowledge of the hazards involved for solid KOH and for KHP. The KOH is incompatible with acids and
flammable liquids. If skin comes in contact, it should be cleaned and rinsed thoroughly with water for
10 min or more. The KHP is relatively nontoxic, but care should be taken to avoid breathing the powder
and to get as little powder on the skin as possible.
3.0 QUALITY CONTROUQUALITY ASSURANCE
The KOH normality is standardized against a pure standard KHP solution on a daily basis. Two
milliliters of KHP solution is titrated with bromothymol blue indicator to verify the KOH solution. Samples
of sulfuric acid solution are analyzed as control samples.
64
-------
OPERATING PROCEDURE FOR PURITY OF R-12 IN AUTOMOBILE AIR CONDITIONING SAMPLES
1.0 PROCEDURE
This method follows guidelines of EPA Method 18, Measurement of Gaseous Organic Compound
Emissions by Gas Chromatography, CFR Part 60, Appendix A, Method 18. Method 18 does not define
the chromatographic columns or conditions in detail; however, the conditions are defined in an article in
The Supelco Reporter. Vol. V, No. 4, October 1986, entitled "New Packed GC Column for Fluorocarbons
is Unaffected by Reactive Gases." The column as defined is a 5 percent Fluorocol coating on 60/80
mesh Carttopack B in a 10 ft by 1/8 SP alloy.
The article defines the gas chromatographic conditions as follows:
Column temperature: 50 °C isothermal
Sample loop temperature: Ambient (Sample size 0.5 and 5.0 mL)
Injector temperature: 120 °C
Detector temperature: 250 °C
Detector: FID
Carrier gas: Helium 30 mL/min
2.0 CALCULATIONS
The R-12 (dichlorodifluoromethane) is expressed as a percent of the entire area. The integrator
will calculate the peaks on a percent-of-area basis. The number is reported as presented. A minimum
purity of 99.5 percent is expected.
3.0 SAFETY PRECAUTIONS
Method 18 and general gas chromatograph (GC) safety precautions apply: hot surfaces to be
aware of, possible explosion hazard from hydrogen gas, and possible replacement of oxygen in the lungs
by carrier gas. Make sure there are no leaks in the gas lines for the GC.
Dichlorodifluoromethane (R-12) is a colorless gas with a characteristic ether-like odor at
concentrations above 20 percent by volume. It is incompatible with chemically active metals. It is
nontoxic but does displace oxygen. Concentrations above 10,000 ppm require a contained breathing
device. Rapid vaporization of liquid R-12 will freeze tissue. Because containers are at about 50 psig
pressure, care should be taken to release the pressure slowly.
65
-------
OPERATING PROCEDURE FOR FREE HALIDE ANALYSIS OF R-12 SAMPLES
The method lor analysis of the free halides is ion chromatography. The valve on the sample
container is cracked to purge the sample line, and then the fittings are tightened. The sample container is
placed so that the liquid phase will be sampled, and a 100-g portion is bubbled through approximately
100 mL oi buffered eluent (0.0056 M NaHC03,0.0045 M Na2C03). The container is reweighed to
determine the amount of sample added. The buffered eluent is brought to a volume of 100 mL and is
ready for injection onto the ion chromatograph. The final concentration will be the reported amount
(ng/mL) times the volume of eluent divided by the weight of the sample.
Samples will be analyzed on a Dionex 2110i Ion Chromatograph (IC). Standards for chloride and
fluoride will be made in IC eluent by using analytical grade sodium salts of both halides to make a
1000-ppm stock and diluting as necessary. Spiked samples of known concentrations will be made
separately from the original chemicals and analyzed with each batch of samples. Should the spike fail to
be analyzed within the range given in the data quality objectives, the entire sample set will be repeated.
A lour-point standard curve will be generated for each day of analysis.
66
-------
TECHNICAL REPORT DATA
(Please reed huxruct1ons on the reverse before completing}
1. REPORT NO.
EPA-600/2-89-009
2.
J"pCB8TlCr§T8 2/AS
4. title andsubtitle
Evaluation of Refrigerant from Mobile Air
S. REPORT DATE
February 1989
Conditioners
8. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
Leo Weitzman
,
9. PERFORMING OROANIZATION NAME ANO ADORESS
Acurex Corporation
10. PROGRAM ELEMENT NO.
P. C. Box 13109
Research Triangle Park, North Carolina
27709
11. CiNYhAfiT/'SAAkf n6.
68-02-4285, Task 1/006
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Air and Energy Engineering Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT ANO PERIOD COVERED
Task final; 4/88-1/89
14. SPONSORING AGENCY CODE
EPA/600/13
15. supplementary notes AEERL project officer
541-2429.
is Dale L. Harmon, Mail Drop 62b, 919/
is. abstract The report gives results of a project to provide a scientific basis for
choosing a reasonable standard of purity for recycled chlorofluorocarbon (CFC) re-
frigerant in operating automobile air conditioners. The quality of refrigerant from
air conditioners in automobiles of different makes, ages, and mileages, from dif-
ferent parts of the U. S., and with both failed and properly working air conditioners
was measured. The refrigerant (CFC-12) was tested for water content, acidity,
residue (compressor oil) quantity, refrigerant purity, residue purity, inorganic
chloride, and inorganic fluoride. Cf the 227 cars sampled, neither the refrigerant
nor the residue showed measurable levels of acid or inorganic chlorides and fluor-
ides. The gaseous refrigerant, in all but two samples, was of higher purity than the
specification for new CFC-12. The residue was greater than 99% pure in all but two
samples. The mean water content for all samples (56 ppm) exceeded the Federal
Specification BB-F-1421A of 10 ppm maximum. This work will be the basis for pro-
grams to reduce CFC emissions from the servicing of automotive air conditioners.
17.
KEY WORDS AND OOCUMENT ANALYSIS
a. DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Pollution
Refrigerants
Air Conditioners
Automobiles
Chlorohydrocarbons
Pollution Control
Stationary Sources
Automotive Equipment
Chlorofluorocarbons
13 B
13 A
13F
07 C
13, DISTRIBUTION STATEMENT
IB. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
74
Release to Public
JO SECURITY CLASS (Thispage)
Unclassified
22 PRICE
EPA Form 2220-1 (1-71)
------- |