Technical Issues and No-Clean Approaches

P2 Tech Technical Issues and
No-Clean Approaches

TECHNICAL FEASIBILITY OF NO-CLEAN APPROACHES

VISIBLE RESIDUES: NOT NECESSARILY DETRIMENTAL

NO-CLEAN FLUX DOES NOT MEAN NO CLEANING AT ALL

AN ALTERNATIVE ATMOSPHERE, SUCH AS NITROGEN, CAN HELP

Return to the first page of the No-Clean Cleaning Technology Review

TECHNICAL FEASIBILITY OF NO-CLEAN APPROACHES

No-clean approaches for soldering in electronics manufacturing are already used successfully by a number of manufacturers (primarily the larger ones, such as IBM, AT&T, etc.) in full-scale production. In addition, the United States government, especially the Department of Defense, has funded a number of tests and investigations of the no-clean fluxes. Both laboratory, pilot-scale, and full-scale implementation studies have been reported on in the literature. It is clear from these reports that no-clean technology is a technically viable alternative for many soldering applications, using either wave soldering or oven reflow soldering. It is also clear from these reports that the technology is rapidly changing and evolving.

The need for alternatives to be available quickly once the Montreal Protocol went into effect forced many no-clean fluxes into the marketplace before their use was completely understood or optimized. This lead to some collaborative efforts between vendors and manufacturers where vendors have adjusted no-clean flux formulations while working with a manufacturer on no-clean implementation. It also lead to a major effort by the Department of Defense to test no-clean fluxes and prove their acceptability for military applications.

Some examples from the literature include:

Department of Defense-funded tests were conducted at Binghamton University on 10 different no-clean solder pastes for use in military applications. First, the pastes were all chemically analyzed, tested for corrosivity, screen printed onto standardized test boards, and processed (reflowed) in a five-zone infrared (IR) oven. Next, tests were conducted on four types of boards similar to those actually used in military assemblies made at Martin Marietta (Lockheed Martin). Processing with no-clean pastes resulted in boards that passed both electromigration and SIR tests, even though some quantity of residue was visible on all boards processed using the no-clean pastes. The researchers observed that the most frequently observed defects found in no-clean processed boards were bridges, coplanarity, insufficient solder, no-joint made and misregistration. However, the researchers observed that most of these defects stemmed from printing, as opposed to IR reflow. (Ref. 5)

Solectron Corporation of Milpitas, California reported on making a successful transition to no-clean solder pastes from aqueous-based solder pastes for both oven reflow and wave soldering processes. It found that yields with the no-clean pastes used were equal to those achieved with the aqueous pastes, and that only a minimal visual residue was left by the pastes. Electrical performance was not compromised with the no-clean paste. (Ref. 6)

IBM successfully converted from rosin-based flux to a no-clean flux for the flip-chip process (a method used to attach dies to ceramic carriers), resulting in the elimination of perchloroethylene and xylene in the process. This was done by combining a mildly activated, no-clean fluxing material in combination with a hydrogen reflow reducing atmosphere. The flux and hydrogen both clean solder surfaces and reduce oxidation in this system. The flux also almost completely decomposes after processing in the hydrogen atmosphere, thereby leaving minimal residue. This hydrogen-flux combination system was used because the standard no-clean fluxes will not withstand the high temperature reflow cycle required for flip-chip on ceramic assembly. The change to the no-clean system allowed IBM to eliminate half of the process steps previously required for joining, while maintaining yields and product quality comparable to the old process (defect rates were 0.05% with the no-clean process versus 0.03-0.06% for the rosin-based process). Defects that did occur were primarily due to incoming chip/ceramic carriers being contaminated and process or tooling errors. A specific contaminant that caused solderability problems was a silicone residue left behind from a tape used during dicing of the ceramic assemblies to their proper size. Eliminating this tape from the dicing process solved the problem. (Ref. 7)

Researchers at Motorola’s Government Electronics Group, in conjunction with Sandia National Laboratories and Los Alamos National Laboratory, evaluated a no-clean flux for potential application in military electronic assemblies as a replacement for conventional rosin-based flux. A dilute adipic acid flux coupled with a formic acid vapor in a nitrogen cover blanket and a wave soldering process was used in all tests, which were performed on three different test boards. Flux was applied to the boards using an ultrasonic spray process. The experimental design included testing using a number of different settings for key process parameters. For many of the settings used for tests, the no-clean process was comparable to the rosin-based control with regards to number of solder defects, ionic cleanliness, solder joint contact resistance, surface insulation resistance (SIR), solder joint mechanical strength, and long-term storage. (Ref. 8)

These and other (Ref. 9-14) references describe actual testing of no-clean fluxes. These and other investigations of cleaning has lead the military to abandon military specifications (MILSPECs) for cleaning of electronic assemblies, which require the use of ozone-depleting solvents. Instead, the military has adopted industry performance standards for all new electronics manufacturing contracts.

Some important issues that have been raised in the literature with regards to no-clean fluxes are: the need to accept visible residues when using many no-clean fluxes; the fact that not all cleaning in the soldering area is eliminated by using no-clean fluxes; that use of a alternative atmosphere, such as nitrogen, can help when using no-clean fluxes; and that even through conversion to no-clean fluxes does not require major process changes, it still is a major endeavor because of the up-front testing that is currently required. Each of these topics is addressed below.

VISIBLE RESIDUES: NOT NECESSARILY DETRIMENTAL

Many of the tests and case studies in the literature report that visible residues were left on the electronic assembly after soldering with a no-clean flux. A number of these references also report the assemblies passed other testing, such as rate of defects, SIR, etc. Examples include:

Binghamton University reported that, based on testing completed for military electronic assemblies, visible residues left when using no-clean fluxes "are not necessarily detrimental to board performance." (Ref. 5)

IBM reported that at flux residues were visible on boards processed using several no-clean fluxes as a "white haze" or "brown flux bubbles." Even though the boards had residues, they passed SIR reliability tests. (Ref. 11)

The recommendation that is made based on these results is that the manufacturer (or customer) may need to change its specification so that greater visible residues are accepted. In cases where a visible residue is unacceptable but a change to no-clean fluxes is still of interest, manufacturers have to consider using an alternative atmosphere, such as the hydrogen atmosphere used by IBM (discussed above) or a nitrogen atmosphere (discussed below) to try to decrease visual residues.

NO-CLEAN FLUX DOES NOT MEAN NO CLEANING AT ALL

Several researchers made an important clarification about using no-clean fluxes: cleaning is still required in the soldering process area. IBM reported that to implement a no-clean flux at a memory card manufacturing plant, cleaning is still required at the screening stage for the stencil and for mis-screened panels. (Ref. 11) Compaq Computer Corp. reported that localized solvent cleaning operations are often completed in no-clean environments after manual rework to improve the electronic assembly's appearance and to enhance automated contact testing. (Ref. 14) In addition, for companies with wave soldering, cleaning of the flux spray or foam applicators is still needed.

This means the soldering area likely will require cleaning equipment to accomodate these ancillary cleaning requirements.

AN ALTERNATIVE ATMOSPHERE, SUCH AS NITROGEN, CAN HELP

Several researchers who tested or implemented a no-clean flux used atmospheres other than air during the soldering step (wave or oven reflow). Most commonly, inert atmospheres, primarily nitrogen, were used, although IBM reported on an innovative approach using hydrogen, as discussed above.

Some of the major advantages attributed to inert atmosphere soldering, as reported in Printed Circuits Handbook (Ref. 1) include:

the prevention of further oxidation of the metals component of the soldering system,
the elimination of oxidation of the flux,
better fluxing action resulting in cleaner parts,
more thorough fluxing and dissolution of metal oxide rafts on the surface of the solder,
lower level of residues when using no-clean solder paste formulations,
reduction of solder ball formation, and
fewer voids in the solder joint.

Researchers at Solectron were cautionary with regards to what type of benefits could be realized using a nitrogen atmosphere. They tested a nitrogen atmosphere versus air for wave soldering with no-clean fluxes, and found that the most obvious and consistent benefit in using nitrogen was drastic cosmetic improvements. No-clean solder joints made using an inert nitrogen atmosphere were found to be "illustrious" and "virtually residue free." (Ref. 13) The authors stated that "when it comes to claims of such and such a process reducing defects, be wary. Defects are truly dependent on the layout design and how this interacts with a very dynamic soldering process."

A CHANGE TO NO-CLEAN: TESTING REQUIRED

While it may be different in the future, a change to no-clean still requires extensive testing by the manufacturer. There were no cases found where a change to no-clean was made where the new flux was used as a "drop-in" replacement for the current flux and little or no testing was done. This means that the company considering no-clean fluxes needs to be able to set aside the resources to conduct such tests. Requirements for testing are probably why many smaller manufacturers have not yet seriously considered no-clean fluxes. Hopefully, continued testing by industry and the government will allow smaller manufacturers to implement a no-clean solution in the future without the extensive testing required at present.

Continue on to the economics page of the No-Clean Technology Review.
or
Return to the introduction of the No-Clean Technology Review