Selection of Low Ozone Depleting Oxygen System Cleaning Solvents

KEYWORDS

Solvents; Gas Turbine Oils; Hydraulic Fluids; Greases; Seals; Static; O-Ring

ABSTRACT

CFC113 (1,1,2-trichlorotrifluoroethane) was widely used in the military for removing dirt and lubricants from equipment. Users sought substitute "environmentally friendly" solvents due to banning of the ozone-depleting compounds. In this work, for wipe and liquid cleaning of liquid and gaseous oxygen systems, seven solvents, plus CFC113 as a baseline, were evaluated as CFC113 replacements. First, the cleaning ability was evaluated with common military lubricants and other common contaminants anticipated to be found on oxygen system components, MIL-PRF-7808 (ester-based gas turbine engine oil), MIL-PRF-83282 (hydraulic fluid), MIL-PRF- 27617 (perfluoropolyalkylether-based PTFE thickened grease), MIL-PRF-27617 sprinkled with Arizona road dust, SAE 20W-50 tube bending oil, and 3M 250 tape residue. These solvents were also tested for compatibility with oxygen systems using the liquid oxygen mechanical impact test and by determining the autogenous ignition temperature in pure oxygen. Candidate solvents must also be compatible with elastomer seals. None of the solvents caused permanent damage to the seals. No single solvent performed as well as nor was as universal as CFC113 but several were close.

BACKGROUND

This wipe solvent program was initiated to find a replacement for 1,1,2-trichlorotrifluoroethane, which will be referred to as CFC113 for the remainder of this paper, for cleaning liquid and gaseous oxygen systems on board aircraft and ground servicing equipment (Davis and Sioux-Lhyn1). CFC113 has been widely used for wipe cleaning because of its excellent cleaning abilities with a wide variety of soils and compatibility with oxygen systems. For details of a solvent replacement study concerned with non-oxygen system cleaners (see Roberts, et al.2). Production of CFC113 has been banned due to the Montreal Protocol and the US Clean Air Act because it is a Class 1 ozone-depleting compound. Initially, performance characteristics, other requirements, and the target cost of a new solvent were established. Along with the properties tested in these studies the candidate solvents needed to be environmentally friendly and non-toxic at user exposure levels. To initiate this program, a kick-off meeting was held to which Air Force users, Air Force systems and laboratory personnel, and potential solvent suppliers were invited (Davis and Sioux-Lhyn1).

EXPERIMENTAL

Candidate Solvents

A large number of solvents were included in the original testing that was narrowed to the primary candidates listed in Table 1 with their trade names, chemical classes, and reference names for this paper.

Performance Requirements

Performance requirements were determined through discussions with technicians that use CFC113 and will be using the replacements identified through this study (Davis and SiouxLhyn1). Studies were conducted based on the important solvent properties identified and listed below. The first two properties reveal the compatibility of the candidate solvents with liquid and gaseous oxygen systems, LOX and GOX, respectively, which is critical to the safety of working with the solvents. These were conducted at the NASA White Sands Test Facility (WSTF). The seal compatibility and immersion cleaning ability were conducted at the Air Force Research Laboratory (AFRL) at the Wright-Patterson Air Force Base (MLBT).Wipe cleaning experiments were conducted at both the Phoenix Chemical Laboratory and the WSTF Field evaluations were conducted at AFRL at the Edwards Air Force Base (PRSP).

* Liquid oxygen (LOX) mechanical impact

* Autogenous ignition temperature (AIT)

* Seal compatibility

* Immersion cleaning ability

* Wipe cleaning ability, automated

* Wipe cleaning ability, manual

* Field cleaning validation

Liquid Oxygen (LOX) Mechanical Impact

LOX mechanical impact tests were performed at NASAWSTF according to ASTM G86. Sample weight, thickness, and the diameter could not be measured due to the evaporation rate of the candidate solvents at room temperature. Approximately 0.3 ml solvent was used in each test. The test cup was 1.783 cm in diameter. A test medium of 100% liquid oxygen was used at a temperature of -183°C and a pressure of 85.5 kPa.

Twenty impacts were run initially. If there was one reaction the test was extended to 60 impacts at an energy of 72 ft-lb. WSTF recommended energy levels are 72, 65, 60, 55, 50, 45, 40, 35, 30, 25, 20, 15, and 10 ft-lb.

Autogenous Ignition Temperature (AIT)

The ASTM G 72 procedure was used for the AIT testing. Solvents were tested in 100% oxygen three times at a pressure of 0.0345 MPa (50 psia) and 5 times at a pressure of 13.8 MPa (2000 psia). The average sample weight was 0.22 ± 0.01 g and the heating rate was 5±1°C. The maximum vessel temperature was 450°C (842°F).

Seal Compatibility

The equipment used for the compatibility study was: 4-oz. wide-mouthed jars with PTFE-lined caps, wire stands, balance, tweezers, and a Shore A Durometer (Roberts, et al.3). Three of each type of o-ring were weighed in water and in air to determine their volume. The rubber o-rings were also measured for hardness. Duplicate tests were conducted. The three o-rings were hung on a wire stand and placed in a wide-mouthed jar (see Figure 1). Enough solvent was added to completely cover the o-rings. This resulted in approximately 50 ml of solvent being used per trial. One o-ring was removed from the solvent after 30, 60, and 90 days. The seal weight was measured in air and in water immediately after being removed to determine the shrinkage or swell experienced by each seal. The hardness was also measured for the rubber o-rings. Upon removal from the solvent, the weight of the o-rings was not stable. The weight decreased as the solvent evaporated from the surface of the elastomer. After the weight stabilized the weight in air and water and the hardnesses were again measured.

Immersion Cleaning

The equipment used for the cleaning study was: two 400-ml beakers, tweezers, analytical balance, a stainless steel wire stand with 4 prongs, oven, 5.08 cm by 1.27 cm (2 inch by l/i inch) ANSI 1010 steel coupons, and an ultrasonic bath (Roberts, et al.3). Metal coupons were cleaned by successive washings in hexane and acetone in an ultrasonic bath prior to using them in the cleaning study and between the tests. An identification number was imprinted on each coupon and they were washed again in the ultrasonic bath. After the second washing, the coupons were dried in an oven for 10 minutes and cooled to room temperature. Four coupons were weighed.

The contaminants for this program were based on input from the field personnel who work with the LOX and the GOX systems. For oil contaminants, MIL-PRF-7808 military engine oil, MIL-PRF-83282 hydraulic fluid, and SAE 20W50 tube bending oil, one drop of contaminant was placed on each coupon. The drop was spread to an area of 1.27 cm by 1.27 cm (V2 inch by V2 inch). This resulted in approximately 10 mg of contaminant. Twenty milligrams of MIL-PRF-27617 was spread to an area of 1.27 cm by 1.27 cm (l/2 inch by lh inch). The grease was smoothed to create a uniform thickness of the contaminated area. Since the density of MIL-PRF-27616 is approximately twice that of the other contaminants, similar volumes of liquid lubricants and grease were tested.

The Arizona road dust would not stick to the metal surface on its own. Ten milligrams of the MIL-PRF-27617 was spread to an area of 1.27 cm by 1.27 cm (V2 inch by Vi inch) and the Arizona road dust was sprinkled onto it. The coupon was tapped so that excess dust would fall off. More dust was sprinkled and the coupon tapped until approximately 10 mg of the dust stuck to the grease.

Three to five milligrams of 3M tape 250 residue was applied to the coupons. To apply the tape residue a piece of 3M tape 250 was placed onto the coupons. The coupons with the tape were placed in an oven for 24 h at 110°C. After 24 h the coupons were cooled to room temperature and the backing of the tape was removed. Three to five milligrams tape residue remained on the coupons.

The coupons were weighed with the contaminant. The four coupons were hung on a wire stand in an empty 400-ml beaker. Another 400-ml beaker was filled with 250 ml solvent. The wire stand with the four coupons was transferred to the beaker containing solvent (see Figure 2). The solvent was applied to the coupons in this manner to minimize agitation. A timer was started. After 30 s, the first coupon was removed from the solvent. The others were removed from the solvent at 1, 2, and 5 min. The coupons were allowed to dry and then were reweighed to determine the mass of the contaminant removed. The tests were conducted in duplicate. If there were inconsistencies in the data, extra trials were run.

Automated Wipe Cleaning

Since one of the primary cleaning methods to be used with the solvent was as a wipe solvent, it was important to determine the extent to which the mechanical action of wiping increased the cleaning efficiency of the solvents. To minimize the variability associated with different pressures and speeds of the wiper by hand, a modified Gardner washability test apparatus was used (see Figure 3).

Manual Wipe Cleaning

Additional manual wipe testing was done on CFC113, HFE1, HFE2, and HFC with MIL-PRF-27617, and SAE 20W-50 tube bending oil as the contaminants. Coupons measuring 5.08 cm by 10.16 cm (2 by 4 inches) were cleaned and weighed. A contaminating solution of 1 mg contaminant per 1 ml CFC113 was prepared. One milliliter of the contaminating solution was dripped onto the coupon and allowed to volatilize. Coupons were reweighed. This method applied 0.4 mg to 1 mg contaminant to the coupon. A 10.16 cm square (4 inch square) of a lint free, cleanroom wiper was loaded with 2 ml solvent. Each edge of the wiper was held between the thumb and the index finger of one hand. The center of the wiper was moved across the coupon. The coupon was allowed to dry for 5 min and then reweighed. The final weight was taken after 30 min.

For low temperature tests the contaminated coupon and the solvent were placed in a refrigerator at 5°C for 2 h prior to running the procedure listed above. For the high temperature test the contaminated coupon was placed in an oven at 50°C for 2 h prior to running the cleaning procedure.

Field Cleaning Validation

At the conclusion of the AIT, LOX compatibility, seal compatibility, static cleaning, and wipe cleaning tests, 5 solvents were chosen for field evaluations. They were PFBI, HCFC225, HFC, HFE1, and HFE2. In this work FIFE2 replaced HFCI because of HFCI failing performance in the NASA White Sands Test Facility evaluation. Typical field level cleaning procedures were selected and conducted by LOX and GOX maintenance personnel. Phase 1 evaluated the cleaning ability of each solvent. Phase 2 determined the amount of each solvent necessary to insure cleanliness of an oxygen line. Phase 3 tested cleanliness by visual inspection.

Phase 1

Solutions of 0, 1, 6, 12, and 150 ppm contaminant in CFC113 were prepared. The contaminants used were MIL-PRF-5606 hydraulic fluid, MIL-PRF-83282 hydraulic fluid, MIL-PRF-7808 engine oil, and SAE20W-50 tube bending oil. A Fourier transform infrared spectroscopy analysis was run on each sample to determine absorbance. For an example of spectra of clean CFC113 and contaminated CFC113 (see Figures 4 and 5). These absorbencies were graphed so that the amount of the contaminant remaining after cleaning could be determined.

Four parts were cleaned with CFC113. The parts used were an AN cap (an aircraft fitting), a T-fitting, a 6 inch metal flex line, and a wrench. Parts were contaminated with a known amount of hydrocarbon contaminant. The cap, Tfitting, and line were immersed in solvent and sonicated for 1 minute. The wrench was wiped with a cotton wipe that had been folded in quarters and soaked in 2.5 ml solvent. After cleaning, the part was dried in a vacuum oven for 5 minutes. The parts were immersed in CFC113 and sonicated for 5 minutes, except the line, which was sonicated for 7 minutes.

The CFC113 was collected and analyzed by Fourier transform infrared (FTIR) spectroscopy to determine how much contaminant had not been removed by the candidate replacement solvent. For an example standardization curve (see Figure 6).

Phase 2

A 33.91 cm by 0.64 cm (13.35 inch x V4 inch) ID metal line was cleaned with CFC113. The CFC113 was then analyzed by FTIR to insure cleanliness prior to the start of the test. The line was weighed. The line was contaminated with 10 mg of SAE 20W- 50 tube bending oil or MIL-PRF-83282 hydraulic fluid. A measured amount of solvent was poured into the line that had one end capped. The other end then was capped. The line was swished with a rocking motion. The line was drained and another aliquot of solvent was poured in and swished. The line was drained and dried with nitrogen for 30 s. Four milliliters of CFC113 was placed in the line and was swished 50 times. The CFC113 rinse was analyzed by FTIR to determine how much of the contaminant was left in the line after cleaning with the other solvents. The allowable contaminant in the line is 3 mg/foot for the size line used.

Phase 3

A coating of MIL-PRF-83282 was applied to the exterior of an oxygen fill valve nozzle using a saturated cotton swab. Desert sand/dirt was applied to the nozzle by holding it over the desert sand/dirt while nitrogen was blown over the desert sand/dirt. The nozzle was sprayed with a solvent using a new Type A spray bottle for each test until the part appeared clean. It was important to use this bottle because it is compatible with HCFC225, whereas the "Sure Shot Container" that was previously used is not. The "Sure Shot Container" was used previously because it is compatible with CFC113. Both spray bottles are manufactured by Milwaukee Sprayer Co.

RESULTS

LOX Mechanical Impact Performance Standards

The results are listed in Table 2. To pass the test, the solvent must have no reaction in 20 impacts at any energy level or no more than one reaction in 60 impacts at an energy of 72 ft-lb. The PFBI, although it failed this test, was still included in the program for a number of reasons. First, although it did not pass the 72 ft-lb energy level tests, it did pass the next level down; i.e., the 65 ft-lb energy tests. Second, the energies involved in the ignitions were quite low compared to typical ignitions, so it was deemed that there was an acceptable risk for its use with LOX/GOX system maintenance. Third, its overall performance as a direct replacement solvent for CFC113, e.g., cleaning ability, compatibility with elastomers, boiling point, etc., made it very attractive to keep in the program.

AIT Performance Standards

The following standards were deter- mined through discussion between WSTF and Wright Patterson Air Force Base.

Category A AIT < 250°F Not recommended for oxygen system use

Category B 250°F < AIT Caution when used in oxygen < 400°F systems

Category C AIT > 400°F Recommended for oxygen system use

All solvents tested rated in category C except PFBI which rated in category B. The detailed results are listed in Table 3.

Seal Compatibility

All solvents performed similarly to CFC113. All solvents swelled rubber o-rings initially (Roberts, et al.3). All o-rings volumes and hardnesses ended close to their original values after the solvents evaporated. None caused long term adverse effects to the rubber or plastic o-ring samples. Complete details of the elastomeric results are in Roberts, et al.3.

Immersion Cleaning

The 30 s cleaning data is shown in Figure 7. The 30-s test was the most discriminating (Roberts, et al.3). CFC113 cleaned nearly 100% of MIL-PRF-83282 and MIL-PRF-7808, about 45 and 55% of the MIL-PRF-27617 and Arizona Road Dust, respectively, in 30 s. HCFC225 and PFBI performed similarly to CFC113 in 30 s. The SAE 20W-50 tube bending oil was not cleaned well by any of the solvents. The other solvents lagged behind and caught up somewhat during the longer cleaning times. Therefore, the decisions on the cleaning ability were based on a 30 s time period.

The calculations for all cleaning studies were as follows:

(ProQuest: ... denotes formulae omitted.)

Data for 3M tape is not included because none of the solvents, including CFC113, cleaned the tape residue consistently.

Automated and Manual Wipe Cleaning

Wipe cleaning improved the effectiveness of all the solvents. The relative order of the cleaning ability remained the same. Detailed data are available in Roberts, et al.3.

Field Cleaning Results

Phase 1

Data is in Figure 8. HCFC225 cleaned all contaminated parts to nearly 100% clean. PFBI was not tested with the line but cleaned the other three contaminated parts to nearly 100%. HFE 1 and HFE 2 cleaned the wrench and the fitting to nearly 100% but only cleaned the line and the cap between 53% and 75%. HFC was not tested with the wrench or the fitting. HFC cleaned the line to 35.1% and the cap to 83%. HCFC225 performed the best of all the solvents tested.

Phase 2

Data is in Table 4. The ability of each solvent to clean the tube blending oil was tested in this phase. The HFE solvents were unable to satisfactorily clean the SAE 20W-50 tube bending oil. Two, three, and ten washes of 4 or 5 ml of the HFE solvents were used. HCFC225 and PFBI only required 2 washes of 2 to 4 ml to clean the line to less than 0.22 mg contaminant. These performed similarly to CFC113 which was tested as a standard. The total maximum allowable contaminant was calculated to be 0.22 mg based on 0.003 mg contaminant per cm2 (3 mg/ft2) of tube surface area.

Phase 3

Data is in Table 5. The nozzle was visually clean after it was sprayed with ~50 m of HCFC225 or PFBI. For HFC, HFE1, and HFE2 the nozzle never attained visual cleanliness even though it was sprayed with excessive amounts of each solvent.

CONCLUSIONS

HFEIPA and HFCIPA were determined to be unsuitable for further testing. HCFC225 and PFBI were found to be the best replacements for CFC113 of the solvents tested. Based on this work the military technical order TO. 15X-1-1 has been changed to require the use of HCFC225 for LOX and GOX cleaning. Work is currently underway to commercialize PFBI as a cleaning solvent because HCFC225 is a Class II ozonedepleting compound scheduled to be banned in the future while PFBI is not classified as an ozone depleting compound.

HFEIPA and HFCIPA severely failed the LOX mechanical impact test. PFBI slightly failed but the nature of the impact was mild enough to consider the risk of using PFBI in LOX systems to be acceptable.

* All solvents have an acceptable AIT.

* All solvents performed similarly to CFC113 in seal compatibility.

* HCFC225 and PFBI performed the best in immersion cleaning. HFCIPA was next. Wipe cleaning did not affect cleaning ability ranking as determined by the immersion cleaning test results of this investigation.

* Field cleaning tests confirmed that HCFC225 and PFBI have the best cleaning ability in the field conditions as well as lab conditions.

ACKNOWLEDGMENTS

The authors wish to thank Dennis Schroll of ASC/YTJV, Wright-Patterson Air Force Base who was the overall program manager, Harold Beeson of theWhite Sands Test Facility (WSTF) for conducting the AIT and LOX/GOX compatibility testing, Sioux-Lhyn Shutt of WSTF for conducting the manual wipe cleaning tests, Arthur Krawetz of the Phoenix Chemical Laboratory for conducting the automated wipe cleaning tests, and Joanne LaRue, Amanda Schoettmer, and Mervyn James of AFRL/PRSP for conducting the field cleaning validation study.

Presented at the STLE Annual Meeting in Las Vegas, Nevada

May 15-19,2005

Manuscript approved June 25,2006

Review led by Paul Bessette

©STLE

Editor's Note: This month's Editor's Choice paper highlights the process utilized by the U.S. military to evaluate potential candidate solvents to replace CFC113 cleaning solvent. The thoroughness of the evaluation is a testament to the due diligence required in evaluating any replacement product. Far too often these decisions are made from reviewing only written technical specifications, or perhaps only listening to verbal assurances from suppliers. A multifaceted trial using quantifiable measurements is seldom conceived, let alone executed. The results obtained within this particular study showcase how deceiving the evaluation process can be when interpreting singular benchmarks.

Evan Zabawski, CLS

Editor

REFERENCES

1. Davis, D.D. and Sioux-Lhyn, S. (2002), "Alternative Wipe-Solvent Testing," NASA WSTF-IR-0171-001-02. July 23, 2002.

2. Roberts, M.B., Gschwender, L.J. and Snyder, C.E. (2004), "Lubricant Cleaning and Compatibility Study for Candidate Chlorofluorocarbon and Hydrochlorofluorocarbon Solvent Replacements," TLT, 59 (2), pp 48-55.

3. Roberts, M.B., Gschwender, L.J. and Snyder,C.E., Jr. (Feb. 2003), "The Wipe Solvent Program," AFRL-ML-WP-TR-2003-4040.

Marcie B. Roberts

University of Dayton Research Institute, Dayton, OH

and

Lois J. Gschwender and Carl E. Snyder Jr.

Air Force Research Laboratory

Materials and Manufacturing Directorate

Wright Patterson AFB, Ohio