1.0 Introduction

The Electronic Packaging and Fabrication Section at the Jet Propulsion Laboratory currently produces various types of CCAs and cable harness assemblies. In many cases, components are still manually soldered to the PWB surface to create the completed CCA, and cable harness assembly often entails manual soldering. The compound 1,1,1-trichloroethane (methyl chloroform, TCA) has been used for a number of years as the cleaning agent of choice for removing flux residues. This material has very good cleaning power, a not too overpowering odor, and its Permissible Exposure Limit (PEL), being set at 350, is acceptable for use in the work environment. However, the Ozone Depleting Potential (ODP) of TCA is 0.12. Although this number is relatively low, compared to 1.0 for CFC-11 and 1.10 for CFC-113, because such large quantities were consumed, it was placed in Category I Group V and its production was slated to be phased out at the end of 1995. Because hydro- chlorofluorocarbon (HCFC) 141-b has about the same ODP as TCA (HCFC 141-b ODP = 0.11), it was considered unacceptable. In addition, production of HCFC 141-b will cease as of January 1, 2003.

In early 1994, the Jet Propulsion Laboratoryís Electronic Packaging and Fabrication Section initiated a project to find an acceptable alternative to TCA as a cleaner in the manual electronics assembly operation.

2.0 Goal for JPLís Electronic Packaging and Fabrication Section

The goals of JPLís Electronic Packaging and Fabrication section were:

• Replace 1,1,1-trichloroethane in cleaning CCAs that have been manually assembled.
• Replace 1,1,1-trichloroethane in the hand cleaning of cable harnesses.
• If feasible, find one cleaning agent that meets all of the sectionís cleaning goals.

Finding a new cleaning agent can be challenging. In particular, at JPL there were a number of both internal and external criteria that any new agent first had to meet in order to be acceptable.

The internal criteria were:

• The new cleaning agent must clean as well or better than TCA using both visual inspection criteria and ionic contamination testing criteria.
• The new cleaning agent must have an ozone depletion potential (ODP) less than 0.05 (TCAís ODP = 0.12).
• The new cleaning agent must be deemed safe to use, especially for hand cleaning operations.
• The new cleaning agent must exhibit good material compatibility with a wide variety of materials used in electronics fabrication and assembly.
• The new cleaning agent should be capable of being used in a suitable piece of cleaning equipment and neither present a hazard, e.g flammability, nor emit undue losses to the atmosphere.
• The new cleaning agent must not have a disagreeable odor, if being used in a hand cleaning operation.
• The new cleaning agent must evaporate at an acceptable rate.
• The new cleaning agent must not exhibit unacceptable outgassing characteristics using typical flight connectors and PWAs.

• The new cleaning agent must be granted, or already have been granted, SNAP (Significant New Alternatives Program) approval by the U.S. EPA.
• The new cleaning agent should not be considered a volatile organic compound (VOC) by Californiaís South Coast Air Quality Management District (SCAQMD).
• If the new cleaning agent does contain some VOCs, it must be granted exemption by SCAQMD.

Back in the summer of 1994, an initial investigation was conducted to find a suitable replacement for 1,1,1-trichloroethane. At that time, a number of different chemistries were on the market, many of them classified as semi-aqueous. That is, they consisted of either a terpene-based or nonterpene-based material, such as various alcohol mixtures, but the final rinsing agent was generally water or a low molecular weight alcohol. At JPLís Electronics Packaging and Fabrication facility, at least fifteen (15) different materials were examined. Most of the materials used in the initial evaluation had a zero (0) ODP.

4.1 Results of First Set of Candidates. The first stage eliminated the cleaning agents whose boards showed an unacceptably high level of ionic contamination and/or visually exhibited a large visible flux residue after cleaning. Also, those solvents having an objectionable odor, caused dizziness, or shortness of breath were eliminated. At this first stage, ten (10) of the fifteen candidates were eliminated. One candidate, a solvent, was not eliminated; rather, it was put on the back-burner, so to speak, because of its cost (see below: Solvent #1). The cleaning agent chosen was a material designated as Semi-aqueous Agent #1. This material was considered, at first cut, to show good cleaning results based on both the ionic contamination test and visual inspection. In addition, it exhibited a mild odor reminiscent of baby powder and was not found objectionable in this regard. Its ODP was zero, and it was not overly expensive.

Semi-aqueous Agent #1 did, however, have several drawbacks. Although it was a good cleaner, it did not readily evaporate. Also, it had a low flash point (F.P.), namely, 44°C (111°F), so it definitely could not be used in a traditional vapor degreaser. Equipment that could hold Semi-aqueous Agent #1 was carefully examined; however, the equipment was not deemed acceptable by several of the manufacturing engineers. In addition, an outgassing test performed using a fluorosilicone grommet/connector exhibited very high outgassing characteristics. Gas chromatographic (GC) analysis revealed that Semi-aqueous Agent #1 was a complex mixture of different chemical ingredients, containing approximately 20 different chemical species.

5.0 Second Set of Candidate Cleaning Agents

Because of the problems associated with Semi-aqueous Agent #1, it was decided to reexamine the issue and determine whether a different sort of cleaning agent would suffice. In the meanwhile, a number of new cleaning agents also emerged in the marketplace, and several of these appeared promising as potential candidates for replacing TCA in the electronics hand cleaning assembly operations. In addition, Solvent #1, which was mentioned above, had given good results. It was eliminated initially because of its high cost. However, it was reconsidered, along with some of the newer solvent materials that were beginning to emerge.

In 1995-96, several other cleaning options were available and were investigated:

Solvent #1. This material is classified as a hydrochlorofluorocarbon (HCFC). That is, the molecule contains the following elements: hydrogen (H), chlorine (Cl), fluorine (F), and carbon (C). The original material investigated was the ethyl alcohol (EtOH) azeotrope of two closely related HCFC isomers. This material has a small, but finite, ODP, generally given as 0.03. The presence of the chlorine in the molecule causes it to have a finite ODP. Under current EPA rules, it is due to be phased out of production in 2015.

Solvent #2. This material is classified as a hydrofluorocarbon (HFC). That is, the molecule contains the following elements: hydrogen (H), fluorine (F), and carbon (C). Since the molecule contains no chlorine, it has a zero (0) ODP. The original material investigated was an azeotrope-like formulation containing the following ingredients: HFC 50.5 wt.%, trans-1,2-dichloroethylene (t-DCE) 43.0 wt.%, methanol (MeOH) 6.0 wt.%, nitromethane (MeNO₂) 0.5 wt.%.

Solvent #3. This material was chlorobromomethane.

Solvent #4. This material was 1-bromopropane (n-propyl bromide).

Solvent #5. This material was a volatile methyl siloxane (VMS) designed for hand cleaning operations.

5.1 Results of Second Set of Candidates. Solvent #1. Although in general this material, the stabilized azeotrope of the HCFC isomers and ethyl alcohol, rendered good results, it often left an unacceptable white residue (WR). The manufacturer was consulted, and a modified solvent was recommended. In addition to the two HCFC isomers and ethyl alcohol, the modified solvent also contained cyclohexane. In this paper, this modified solvent is designated as Solvent #1-M. This material worked quite well in all respects and was considered to be a candidate for final consideration. The modified material (Solvent #1-M) investigated was an azeotrope-like formulation containing the following ingredients: HCFC isomer #1 38.3 wt.%, HCFC isomer #2 46.8 wt.%, cyclohexane 10.0 wt.%, ethyl alcohol (EtOH) 4.5 wt.%, stabilizers 0.4 wt.%.

Solvent #2. Although in general this material, the stabilized azeotrope-like solvent consisting of the HFC, t-DCE, MeOH, and MeNO₂, rendered good results, it also often left an unacceptable white residue (WR). The manufacturer was consulted, and a modified solvent was recommended. In addition to the cited ingredients, the modified solvent also contained cyclopentane. In this paper, this modified solvent is designated as Solvent #2-M. This material worked quite well in all respects and was also considered to be a candidate for final consideration. The modified material (Solvent #2-M) investigated was an azeotrope-like formulation containing the following ingredients: HFC 53.5 wt.%, trans-1,2-dichloroethylene (t-DCE) 25.0 wt.%, cyclopentane 15.0 wt.%, methanol (MeOH) 6.0 wt.%, nitromethane (MeNO₂) 0.5 wt.%.

Solvent #3. When a 3-D model was run to arrive at an ODP value for this material, it emerged unacceptably high, that is, Solvent #3ís ODP > 0.1. Hence, it was withdrawn from any further consideration. It did, however, clean in an acceptable fashion.

Solvent #4. This material did clean in an acceptable fashion, but it had, in the opinion, of many of the technicians, a powerful and unpleasant odor.

Solvent #5. Upon evaluation, this material did not remove flux any better than IPA. Hence, it was withdrawn from any further consideration.

5.2 Odor Test. Because odor can be an important factor, especially for situations calling for manual cleaning, it was decided to run a test utilizing five different technicians to determine the extent of how offensive the odor actually was of the various cleaning agents under consideration. Admittedly this test was qualitative. The scale used was: 10 = Worst case; 1 = No effective odor.

The following ranking was obtained: Solvent #6 (not tested) >> Solvent #4 (n-propyl bromide) >> TCA > Solvent #1-M (HCFC/EtOH/cyclohexane) > Solvent #5 (VMS) > Solvent #2-M (HFC/t-DCE/MeOH/ cyclopentane) where > indicates the solvent exhibited a more obnoxious odor.

Based on all the results up to this point, it was decided to drop Solvents #4 and #5, and proceed with Solvents #1-M and #2-M.

8-Hr. Time Weighted Average Testing

The following two solvents were tested to determine their 8-hr. time weighted average (TWA): Solvent #1-M and Solvent #2-M.

For Solvent #2-M, the PEL is 200 ppm; for Solvent #1-M it is 50 ppm. This latter figure was set by Solvent #1-Mís manufacturer; it may be revised up to 100 ppm.

The Solvent #1-M test was performed in the morning. Each technician had attached to her an activated charcoal tube and a small pump that drew room air through the tube. The tubes were analyzed using gas chromatography to determine the amount of solvent to which each technician was exposed. In the case of Solvent #1-M, the test was set up and conducted by JPL Safety Operations. The results are presented in Table 1 below; the figures given in Table 1 (Row 1) are only for the amount of HCFC to which each technician was exposed.

6.0 Conclusions

The two solvents that were closely scrutinized as suitable replacements for 1,1,1-trichloroethane in manual electronics assembly are:

1. Solvent #1-M: HCFC Isomer #1 38.3 wt.%, HCFC Isomer #2 46.8 wt.%, Cyclohexane 10.0 wt.%, Ethyl alcohol (EtOH) 4.5 wt.%, Stabilizers 0.4 wt.%.
2. Solvent #2-M: HFC 53.5 wt.%, Trans-1,2-dichloroethylene (t-DCE) 25.0 wt.%, Cyclopentane 15.0 wt.%, Methanol (MeOH) 6.0 wt.%, Nitromethane (MeNO₂) 0.5 wt.%.

1. In the case of Solvent #1-M, the following advantages pertain:
• Regarding cleaning power, Solvent #1-M Å 1,1,1-trichloroethane, whereas Solvent #2-M < 1,1,1-trichloroethane.
• Easy to use. It is a direct replacement for solvents based on CFC-113 and for TCA.
• Very good material compatibility with a number of different polymerics and metals.
• No significant global warming potential.
• Can be used in a conventional vapor degreaser. However, a degreaser having a minimum of 100% freeboard, extra chilling coils, and a suitable rolltop cover are highly recommended to keep solvent losses at a minimum. There is a caveat regarding using this material in a degreaser. See below.

• Low PEL. During the 8-hr TWA test, two of the five technicians exceeded the PEL which is currently set at 50 ppm. However, proper training could probably reduce the exposure so that it could be kept under 50 ppm. This is evident since three of the technicians managed to stay well under the 50 ppm limit.
• Finite ODP of 0.03ñtherefore, production will eventually cease (the current phaseout date is currently set at January 1, 2015).
• Expensiveñest. cost ~$125ñ$135/gal. (5-gal pail quantities). Note: Solvents are sold on a weight basis, not on a volume basis. A 5-gal pail contains 44.1 lb.; the manufacturerís recommended price for this quantity is $14.15/lb.

• Although it can be used in a conventional vapor degreaser, the particular formulation is not azeotropic. That is, it may be subject to fractionation, leading to an unacceptable buildup of its flammable ingredients, especially the cyclohexane. However, precautions can easily be taken to prevent this from occurring.

1. In the case of Solvent #2-M, the following advantages pertain:
• Also a good cleaner, but probably not quite as good as Solvent #1-M.
• Zero (0) ODP. Therefore, there is no phaseout date.
• Easy to use. It is a direct replacement for solvents based on CFC-113 and for TCA.
• Very good material compatibility with a number of different polymerics and metals.
• No significant global warming potential.
• Can be used in a conventional vapor degreaser. However, a degreaser having a minimum of 100% freeboard, extra chilling coils, and a suitable rolltop cover are highly recommended to keep solvent losses at a minimum.

• Very high VOC loading, approximately 47 wt.% of the formulation is VOC. This may pose a problem in some locations.
• Expensiveñest. cost ~$140ñ$145/gal. (5-gal pail quantities). Note: Solvents are sold on a weight basis, not on a volume basis. A 5-gal pail contains 51.0 lb.; the manufacturerís recommended price for this quantity is $13.95/lb.