EPA/600/A-96/035
DEGREASING ME TAL PARTS WITH LIQUID CARBON DIOXIDE
Liz Hill. Research Triangle Institute, Research Triangle Park, NC
Charles H. Darvin, Air Pollution Prevention and Control Division, U.S. EPA, Research Triangle Park, NC
Introduction
This paper describes a demonstration of
liquid carbon dioxide (LC02) degreasing of metal
parts from aircraft and vehicles cleaned during routine
maintenance procedures. The objective of this
research was to find and demonstrate innovative parts
cleaning technologies to replace environmentally
damaging chemicals with more benign processes.
The process chosen to be replaced was vapor
degreasing in 1,1,1-trichloroethane (TCA). This
solvent needs to be replaced because it is an ozone-
depleting compound, a hazardous air pollutant (HAP),
and one of the 17 chemicals on the U.S.
Environmental Protection Agency (EPA) 33/50 list of
priority pollutants.1 Carbon dioxide (C02) degreasing
was chosen for demonstration as an alternative to
TCA because it is non-ozone depleting, nontoxic,
nonflammable, and on the EPA Significant New
Alternatives Program (SNAP) list of approved
cleaning alternatives.2 This project was sponsored by
the EPA under Cooperative Agreement No.
CR818419.
Prior to the on-site demonstration, Research
Triangle Institute (RTI) performed preliminary
nonvolatile residue (NVR) analyses on metal parte
cleaned with the LCOz process at the equipment
manufacturer's facility. Based on favorable results
from this preliminary testing, full-scale degreasing
equipment was moved to the Warner Robins Air
Logistics Center (ALC) at Robins Air Force Base
(RAFB), GA. The ALC maintains and repairs F-15
fighters, C-130 and C-141 transport aircraft, and all
Air Force helicopters, as well as electronics for
airborne avionics, communications, vehicles, and
radar equipment. The equipment was operated on-site •
for 2 weeks. During this time, samples of each type
of part were cleaned and inspected, and some were
sent back to RTI for NVR testing. Dirty parts and
parts cleaned by current processes were also collected
and tested for comparison.
The primary parts tested were fuel system
tubing, brass filters, miscellaneous machined metal
parts, and rags from the maintenance processes.
Preliminary soaking in a warm mixture of
hydrocarbon oil (spindle oil) and surfactants (hot oil
process or HOP) was necessary to remove difficult
contaminants from many parts. This high-boiling,
nonflammable mixture was then removed by the
LC02 process and could be recovered for reuse.
Background oil C02 Chemistry
co2 is a gas at normal room temperature and
pressure. It is odorless, colorless, nonflammable, and
nontoxic. It is not an ozone-depleting material,
volatile organic compound (VOC), or HAP. It is a
greenhouse warming gas, but commercial supplies of
COa are obtained by recycling by-products from other
processes. Therefore, its use results in no net addition
of C02 to the atmosphere.
By increasing pressure on C02 and
controlling the temperature, if can be changed from
the gas phase to solid, liquid, or supercritical fluid
phases. When the pressure is reduced, it reverts to a
gas. As a liquid or supercritical fluid, C02 has
excellent solvent properties for many oils, greases,
and other common machining contaminants.
Cleaning with liquid C02 is generally performed
within the temperature range of 60-80 °F (16-27 °C)
and pressures of 700-1500 psi (48-102 bar).
Liqaid C02 is distinct from the better known
supercritical CO* because H can be maintained at
lower pressures and temperatures than supercritical,
so cleaning equipment may be less expensive. Both
liquid and supercritical COz have the advantage of
permeating into tiny holes like a gas and have good
solvency for many oils, greases, and other
contaminants. Liquid C02 will not remove paint, rust,
conversion coatings, or most adhesives. Liquid COz
is not applicable for degreasing parts that will be
damaged by pressure or by leaching of plasticizers
from elastomers.
There is no solvent waste or other media to
dispose of because C02 reverts to a gas after cleaning.
The contaminants can be collected for reuse or
disposal. Even though it can be released to the
atmosphere, the C02 was recycled between cleaning
cycles during the demonstration, so material costs
also were reduced.
Process Description
For this demonstration, most parts were
precleaned by immersion in HOP oil, sometimes with
hand scrubbing or ultrasonics. The oil was healed in
a tank to about 140 °F (60 °C). The HOP's purpose
was to solubilize or mechanically remove difficult
soils and materials that were not soluble in the IX02,
such as waxes and soaps fronj greases. After the parts
were soaked in the HOP, they were drained or spun in
a centrifugal spinner loaned by Nobles Manufacturing
1

-------
to remove most of the oil. Then they were degreased
in LCOj to remove the solubilized contaminants and
particles carried in the HOP oil. Upon completion of
the cleaning cycle, the C02 was automatically
transferred from the cleaning vessel to the recycler,
and the parts were removed clean and dry.
A typical C02 degreasing cycle takes about
25 minutes including loading and unloading parts.
'Hie cycle is divided into five steps: filling, cleaning,
draining, purging, and venting. Parameters that can
be changed by the operator are the duration of die
cleaning step and the speed and direction of basket
rotation. These values are chosen based on the type
and amount of contamination and the complexity and
fragility of the part to be cleaned. Once the unit has
been programmed, these parameters are controlled by
microprocessor. The operator needs only to load the
parts, close the vessel, and push the start button.
Projected Costs
Equipment - LC02 equipment cost depends
on the size of the cleaning vessel. The unit used at
RAFB (supplied by DEFLEX Corporation) costs
about $100,000 with the recycler. A centrifugal
spinner that holds a similar sized basket (12 In.
diameter x 12 in. deep) costs $3,000-5,000. A tank
with a heater for HOP oil costs less than $2,000.
Materials - COz costs range from about
$0.20/gaI for bulk to a high of about $0.85/gal for
C02 in bottles. Each bottle used in this demonstration
cost $28 and held 40 gal of COz, so cost about
$0.70/gal. The average recovery rate of C02
recycled during the demonstration was 93.8%. Total
C02 lost during the 37 cleaning batches run during the
demonstration was about 24 gal. At $0.70/gaI, this
would mean a cost of about $17.
Spindle oil is $130/gal in 55 gal drums. It is
not very volatile, so the major losses are the material
dragged out on the parts. If the oil removed by the
C02 degreasing were returned to the HOP tank, these
losses would be minimal.
Parts Tested
Parts cleaned during the demonstration included;
•	Aluminum and titanium tubing for F-15 fuel
and breathing oxygen systems contaminated
with drawing compound (tallow, fatty acids,
and alkaline salts), metal chips, and shavings,
•	Brass filters contaminated with black grime
and used hydraulic fluid,
•	Small pieces of aluminum honeycomb core
contaminated with light oils,
•	Steel bolts contaminated with heavy grease,
•	Gears and other machined parts contaminated
with oil and black carbon residue,
•	Flap jack screw yokes (assemblies of
machined steel parts containing a sealed
bearing, used to repair F-15 aircraft) heavily
crusted with dirt, grime, and oil.
•	Large used aircraft bearings packed with
heavy grease,
•	Small precision bearing assemblies for
avionics,
•	An electronic test circuit board, and
•	Rags contaminated with shop dirt, drawing
compound, grease, and solvents.
The current cleaning processes for a few of
these parts did not involve TCA, but the parts' owners
were not satisfied with the current cleaners or results
of the current process, and requested testing in LCOz.
Results
Tnbes - Visual inspection showed the tubes
to be completely free of drawing compound after the
HOP/LC02 process. It removed large metal chips and
debris but did not remove all the fine particles. In all
cases, a haze of particles or small piles like dried
puddles could be seen inside the tubes. In some, these
particles could be seen only with bright light pointing
through the tube towards the viewer.
In one test, the HOP oil was deliberately
skipped. The LC02 used without HOP oil removed
metal chips and the oily portion of the grease, but left
visible residues of white, soap-like material.
Five aluminum tubes cleaned by the
HOP/LCOj process were sealed in a nylon bag and
returned to the plating shop to be anodized. The parts
were anodized inside and out and were inspected by
shop personnel. They verified that the tubes were
sufficiently clean for the anodizing process. The
tubes also passed visual inspection after anodizing.
Four of them were kept for further tests by RAFB.
Several sets of tubes were sent to RTI for
NVR testing (Table 1). This included parts cleaned
by the current TCA process, parts contaminated with
drawing compound and cutting debris, and several
batches of tabes cleaned in the HOP/LC02 process.
Tubes from the HOP/LC02 batches had loweT NVR
than the TCA tubes and slightly less variability in
standard deviations, which suggests that H0P/LCO2
is more effective for removing this NVR.
This establishes that the HOP/LC02 process
is a feasible alternative to the TCA process for
removing drawing compound, metal chips, and shop
2

-------
debris from tubes. Further testing would be necessary
to determine whether the process would be capable of
removing small particles to the levels required for
breathing oxygen equipment.
Table 1. Average Values and Standard Deviations
for NVR Tests on Tubes (g/ ft2)

Average
NVR
Standard
Deviation
Dirty tubes
5.462
2.471
TCA cleaned
0.009
0.009
CO„ batch 10
0.00$
0.005
CO* batch 36
0.008
0.004
C02> batch 37
0.003
0.002
Steel Bolts - Most of the heavy grease was
removed by soaking in HOP oil with stirring. The
most difficult areas to clean were inside the cap heads
of the bolts, where grease was packed into the comere
and would not come out with stirring alone. Once
most of the grease was removed with the HOP, the
bolts were cleaned in LC02. This completely
removed all visible traces of oil, but some grease
remained in the bottom of the cap heads. Rust and
paint were not removed, but some paint was flaking
off. These parts were returned for inspection by
RAFB personnel who reported that the parts were
cleaned as well as those from the current process and
would be acceptable for use in the next process.
Brass filters - The owners were interested In
a new process because the current one takes too long
and does not clean well, leaving visible black residue
on the surface of the filter. They are considered clean
enough when light can be seen through the mesh.
Filters cleaned with H0P/LC02 without hand
scrubbing were still covered with crusty black residue
left behind when the oil and grease were removed,
and light could not be seen through the mesh. Once
the process was changed to include hand scrubbing in
HOP oil (with a 0-5 minute presoak), the filters came
out of the LC02 visually clean and shiny, and light
could be seen through all open portions of the mesh.
Parts cleaned by the current process were brought to
the demonstration site for comparison. These filters
had black residue blotches on the surfaces; the
H0P/LC02 cleaned parts did not.
Additional filters were cleaned and returned
to RTT for NVR testing (Table 2). Seven Filters were
blown out with shop air between the HOP and LCOa
processes to see if this would dislodge any of the
contamination between the mesh screens. Seven
others were not blown out. Samples of dirty parts and
parts cleaned by the current process were sent to RTI
for comparison. Hie amount of contamination on the
dirty parts varied widely, from 55 to 559 mg/filter.
The NVR results for the parts blown out with air were
as low as for the current process. The NVR of parts
not blown out was not as low as those cleaned by the
current process, but were still much cleaner than the
dirty parts. The results could be further improved
with some minor process engineering.
Table 2. Averages and Standard Deviations for
Brass Filter NVR Tests (mg/filter)

Average NVR
Standard
Deviation
Dirty filters
289.02
210.13
Current process
17.13
4.68
CQj cleaned,
blown out
16.80
3.18
C02 cleaned,
not blown out
24.61
9.73
C02 cleaned,
not blown out
(another batch)
25.65
15.09
Large C-130 Aircraft Bearings - This
grease was not very soluble in HOP oil. After LC02,
all of the heavy grease was gone, but a thin film of
soap-like material was left under the roller bearings.
The assembler said it looked cleain enough to be used,
but was not as clean as the current process.
Aluminum Honeycomb Core - The main
concern was whether rotation of the parts in the
machine would damage the core edges. There were
no visible signs of contamination on the parts either
before or after cleaning. There was no visual sign of
damage to the core after cleaning. Further testing is
in progress for these parts to determine whether the
process will clean the honeycomb core well enough
for bonding to aluminum skins.
Gears and Machined Parts - The hydraulic
fluid and grease were completely removed, but black
carbon residue remained on some pieces. The owners
of this part said it did not meet their final cleanliness
criterion of "no moveable contamination," but their
current process includes a subsequent abrasive step to
remove the black carbon.
3

-------
Flap Jack Screw Yokes - The oil and grease
were removed from all parts. Parts that were not
hand-scrubbed had areas of black carbon residue in a
few places. Soap from the grease in the bearings was
visible on the outside of the seals. The owner of the
parts said they were clean enough for use.
Small Avionic Bearing Assemblies - The
first bearing had no signs of oil or other
contamination after the HQP/LCOa cleaning. The
engineers in charge of the avionic bearing assembly
visually determined it to be cleaner than was obtained
with the CFC-113 degreaser. A bearing that had
already been cleaned once using a hydrofluorinated
ether (HFE) solvent in a vapor degreaser was then
cleaned with H0P/LC02, RAFB personnel inspecting
the part said it was cleaner than after the HFE
process.
Electronic Test Circuit Board - This was an
initial test to determine if the components on a circuit
board would withstand the 800-900 psi (54-61 bar)
pressure of the LC02 process. Avionics provided a
test board specifically designed to measure the effect
of processes on circuitry operations. The board had
both through hole and surface mount components. It
was cleaned with the same HOP/LCO, procedure as
the most ragged parts. The board was visually
determined to be clean and showed no signs of
damage to the components or the inks used to mark
them. It was then taken back to avionics for
electronic component testing where it was determined
that all components still were functioning properly.
Rags - If HOP oil was used, drawing
compounds were completely removed, leaving no oily
feel or smell. When HOP oil "was not used, the soap-
like portion of the drawing compound was left
behind. The H0P/LC02 process removed grease and
oils, but not ground-in black carbon grime. There
was no fraying or other visible damage, even to paper
towels. If the batch was filled too full, clumps of C02
ice formed that left oily spots behind as they
evaporated. Once cleaned in HOP/LCOa, the rags
were considered clean enough for general shop use.
Summary
The LC02 is similar in effectiveness to TCA
vapor degreasing for many contaminants. It is
capable of cleaning debris from a wide variety of
substrates, including precision and nonprecision
applications. It can dissolve oils and most greases,
but does not have enough mechanical force to remove
fine particles or carbonaceous wear products on its
own. The HOP is necessary to remove fine particles
and assists greatly with removal of other
contaminants.
The H0P/LC02 process in this feasibility
demonstration proved to be very effective for general
cleaning applications. It removed drawing
compounds from aluminum and titanium tubes used
in fuel systems, heavy grease from bolts, hydraulic
fluid from brass filters, and general shop dirt from
aluminum, brass, and stainless steel parts. The
aluminum tubes were cleaned to a lower NVR level
than with the current TCA process. Brass filters were
cleaned equally as well as the current process if an air
blowout was used. Components from breathing
oxygen systems and aluminum honeycomb core also
were cleaned, but additional testing would be required
to validate the LC02 process for these applications.
This work was the first step in establishing
feasibility of the H0P/LC02 method for replacing
chlorinated solvent vapor degreasers. As with all
processes, it needs further refinement to optimize the
cleaning effectiveness for specific applications.
Notices
The information described in this paper has
been funded wholly by the Environmental Protection
Agency under Cooperative Agreement No. CR818419
to Research Triangle Institute. It has been subjected
to Agency review and approved for publication.
The use of trade names and company names
in this paper does not signify recommendation for use
or endorsement by either the EPA or Research
Triangle Institute.
The LCOj cleaning and recycling equipment,
the hot oil process (HOP) material, and related
processes, designs, and operations used in this
investigation and described in this paper are
proprietary to DEFLEX Corporation and are the
subject matter of issued patents and pending patents
and applications.
References
1.	U.S. EPA Pollution Prevention Fact Sheet; EPA's
33/50 Program. Office of Pollution Prevention,
Washington, D.C., August 1991.
2.	Final Rulemaking (59 FR 13044), published in the
Federal Register on March 18,1994.
4

-------
WR MR I -RTP—P-lfl"? TECHNICAL REPORT DATA
iNftiVliXij ftlr 1 iUO (Please read Instructions on the reverse before completing)
-
1. RETORT NO. 2.
EPA/600/A-96/035
3. RECIPI
4, TITLE AND SUBTITLE
Degreasing Metal Parts with Liquid Carbon Dioxide
5. REPORT DATE
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Liz Hill (RTI) and C. H. Darvin (EPA)
S, PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Research Triangle Institute
P. 0. Box 12194
Research Triangle Park, North Carolina 27709
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
CR818419
12. SPONSORING AGENCY NAME ANO ADDRESS
EPA, Office of Research and Development
Air Pollution Prevention and Control Division
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Published paper; 1/95-3/96
14. SPONSORING AGENCY CODE
EPA/600/13
^supplementarynotesAPPCD oject officer is Charles H. Darvin, Mail Drop 61, 919/
541-7633. Presented at National Pollution Prevention Roundtable Conference,
Washington, DC, 4/10-12/96.
is,abstract Xhe paper summarizes the demonstration of an innovative parts cleaning
technology with a more benign process to replace environmentally damaging chemi-
cals. The process has the potential to replace 1,1,1-trichloroethane (TCA), an ozone-
depleting compound, a hazardous air pollutant, and 1 of the 17 chemicals on EPA's
33/50 list of priority pollutants. Carbon dioxide (C02), in a liquid or supercritical
form, has significant surface cleaning properties. Liquid CC2 (LC02) is distinct
from supercritical C02 because it can be maintained at lower pressures and tem-
peratures than supercritical. Both have the advantage of permeating tiny holes (e. g.,
a gas) and have good solvency for many oils, greases, and other contaminants.
There is no solvent waste to dispose of because C02 returns to a gas phase after
cleaning. Contaminants can be collected for reuse or disposal. In the demonstrated
system, CG2 was recycled between cleaning cycles, so material costs also were
reduced. The technology was demonstrated at Warner Robins Air Logistics Center at
Robins Air Force Base, GA. Preliminary soaking in a high-boiling, nonflammable
mixture of hydrocarbon oil and surfactants (HOP) was necessary to remove difficult
contaminants from many parts. The HOP was then removed by LC02 and could be
recovered for reuse.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b.lOENTIFIERS/OPEN ENDED TERMS
c. COSATI Fiekl/Group
Pollution
Degreasing
Metal Products
Carbon Dioxide
Solvents
Pollution Prevention
Stationary Sources
Metal Parts
13 B
13 H
11F
07B
11K
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Reportf
Unclassified
21. NO. OF PAGES
20. SECURITY CLASS (This page}
Unclassified
22, PRICE
EPA Form 2220-1 (9-73)

-------