Is no-clean truly a cleaning challenge? A new study dispels long-held assumptions about cleaning at lower temperatures.
[TEXT NOT REPRODUCIBLE IN ASCII]
Three factors with the most effect on both droplet size and velocity are spray pressures, fluid viscosity and temperature. Of these three, only increases in pressure and temperature yield an increase on impact energy. These energies are typically delivered by nozzles that are placed about 4 to 5" from the PCB (Figure 1).
[FIGURE 1 OMITTED]
The need for thermal energy (i.e., temperature) has played a pivotal role in permitting the solvation process of residues left on assemblies. Until recently the industry blindly assumed that temperature could not be reduced further to more acceptable values. That was partially the case due to the widespread use of surfactant-based products and their need (140-160[degrees]F) for activation. (2)
This study was undertaken to evaluate a recent trend in cleaning-related problems with current solutions--surfactants trying to clean no-clean. In particular, this evaluation was to develop cleaning process protocols that particularly took lead-based and eutectic no-clean paste/flux formulation (most challenging) into account and at the same time evaluate whether cleaning at lower temperatures would be feasible. (When a high-temperature object is placed in contact with a low-temperature object, energy will flow from the high-temperature object to the lower temperature object, and they will approach an equilibrium temperature; Figure 3, online.)
The assumption of energy input and output has been largely ignored in electronics cleaning. Recent cost-saving initiatives have challenged this idea. A general estimation of energy consumption per hour and equipment costs was found to be at least as high as $100. A fraction of that cost is related to the heating modules of the cleaning agent holding tank. It was therefore imperative to evaluate and benchmark the product technology that was designed and proven to effectively remove all types of flux-based residues even at ambient temperatures. This was of particular interest because it had been assumed that temperature was not a variable that could be amendable. With new cleaning technologies (i.e., cleaning agents) this viewpoint is subject to change.
Surfactant-based product technologies are known to require sufficient heating to work effectively during cleaning. This is the explanation, for example, why all currently available batch and in-line cleaning equipment is sold with heating units. At the time, alternative cleaning products were simply not available. It is foreseeable that such product innovations will lead to even more cost-effective process solutions, if one can effectively establish that thermal energy can be neglected to a large extent.
It is estimated that most surfactant-based formulations lose up to 60% of the cleaning agent through continuous evaporation. In addition, the evaporation rate doubles with every 10[degrees]C (18[degrees]F) increase in temperature. For example, a reduction from 160[degrees]F to 140[degrees]F reduces evaporative losses by 50%. A product used at ambient temperature will provide even more significant process savings (Figure 4).
Besides the savings on cleaning agent, it is furthermore established that operating at 160[degrees]F, the current temperature limit of polypropylene cleaners, harshly impacts the life of the equipment. Reducing wear and tear means reduced equipment cost and significantly less process maintenance. In summary, the process-related advantages realized with cleaning at lower temperature:
* Minimizing evaporative losses.
* Lower risk of corrosion and oxidation.
* Minimize wear and tear on cleaning equipment (body, pump, etc.).
* Minimizes component-related MOC issues (coatings, masks, heat sensitive components, etc.).
* Brighter solder joints.
* Wider process window (option remains to increase temperature).
* Less fluctuations in cleaning agent concentration leads to better cleaning stability.
Apart from mechanical and thermal energy, the basic chemical interaction between the cleaning agent and particle is the most important variable. Chemical reactions typically involve breaking and making some (or even all) of the bonds that hold the atoms of reactant and product molecules. Energy is required to break bonds, and since the strengths of different kinds of bonds differ, often a significant overall energy change occurs in the course of a reaction. With the latest cleaning technologies, bonds are not being broken or formed, and chemical interactions are solely based on aspects such as solubility, physical and electrostatic properties. This runs contrary to older, surfactant-based technologies for which dissolution was based on acid base reactions, and active ingredients (surfactant molecules) were depleted during cleaning.
As indicated, one important aspect toward achieving cleanliness is that the cleaning agent (chemical energy) has to be fully "compatible" with residues that are being removed. In other words, equipment-related efforts to remove residues from PCBs are avoided by conducting evaluations on how well an assembly can be cleaned. These are aimed at the general ability to remove contamination with an appropriate cleaning agent. Depending on the respective cleaning technology, they can either lift, dissolve or temporarily suspended various particles and contamination.
Based on previous tests conducted with proprietary MPC-based cleaning technologies, the initial experiments were begun at 122[degrees]F with assemblies soldered under regular atmosphere (Table 1, online). Variation in belt speed did not have any significant impact on the cleaning results. Also, variation in spraying technology proved no different. The authors believe that a further increase in the belt speed at 122[degrees]F would have been possible while maintaining an excellent cleaning outcome. With all cleaning results being satisfactory (visually clean, bright solder joints), the authors further-more argue that such "process reserve" might be helpful for the removal of more challenging flux residues, e.g., lead-free no-clean formulations (Figure 5, online).
Based on the encouraging results obtained at elevated cleaning agent temperature, the soldered boards (regular atmosphere) were subjected to the standardized cleaning process, with the cleaning agent in an unheated state. Initially, the belt speed was set to 1 ft./min. to ensure sufficient soaking and cleaning intervals. The flat spray nozzles were used and upon achieving positive results the nozzles were substituted for solid stream to measure any perceived difference. For all pastes tested, the results were satisfactory and no visible difference between the different spray technologies was seen.
Subsequently, the same parameters were used to evaluate whether higher belt speeds (shorter soaking/spraying) might have a greater influence on the ability to clean these eutectic pastes and fluxes with different nozzles.
[FIGURE 4 OMITTED]
The belt speed was increased to 2 ft./min. and the tests repeated. Still no change in cleaning behavior was noticed, and both spraying technologies again afforded excellent results. The very bright and clean solder joints were particularly noteworthy. The cleaning agent at this point was still in an unheated state. Increasing the belt speed to 3 ft./min. did not show further limitation and it was therefore concluded that for the products evaluated, the difference in spraying technology had no effect on the effectiveness of ambient defluxing of eutectic no-clean formulations (Table 2, online).
For eutectic no-clean formulations the use of an inert atmosphere (nitrogen) for soldering was deemed unnecessary because further cleaning improvements were not required. This also widened the process window, added process flexibility and exemplified an unprecedented ease of removability.
Most important, these results illustrated for the first time that process variable "temperature" could be considered a variable to be eliminated from cleaning processes. This, of course, requires that the appropriate cleaning technology (agent) is combined with the most suitable process parameters (equipment, nozzles, pressure, etc.). (3)
Industry and associations are investigating solder paste alternatives to traditional SnPb pastes. A number of alternative pastes have been developed. One current problem is that new flux formulas are necessary to guarantee reliable soldering at higher temperatures. That includes solvents with higher boiling points, increased rosin content (solids content), and in particular more aggressive activators to inhibit oxidation of the solder at higher temperatures and to provide adequate wetting characteristics. (4) These factors are expected to increase the amount of remaining flux residues and, at the same time, demands on the cleaning process. The high affinity of silver to form hydroxides and sulphides is the major reason for the observed electrochemical behavior. The larger the quantity of activators required by the higher reflow temperatures, the higher the amount of moisture that is adsorbed by these hygroscopic residues. Coinciding with the increased amount of fluxes used is the question of fully reliable encapsulation. (5) Various companies have acknowledged the limited reliability of lead-free soldered components. Especially due to the aforementioned higher reflow temperatures required for most lead-free solder pastes (up to 40[degrees]C higher), additional oxidation and polymerization reactions of the fluxes in use will occur (Figure 6, online). These reactions cause these flux residues to be firmly baked-on during soldering, thus making them significantly more difficult to remove with chemically based cleaning agents.
Such behavior can certainly be depressed through the use of nitrogen, which provides a generally inert atmosphere during soldering. Thus, various oxidation-induced or -based reactions are prohibited. The immediate result is that flux residues are simply not baked-on as much as when in an oxygen-containing environment (Tables 3 and 4, online).
Regular atmosphere was initially used with all lead-free formulations. Also, the cleaning agent was used at 122[degrees]F for the first set of experiments. It was observed that belt speed could be increased to 2 ft./min. in some cases, at 1 ft./min. All lead-free no-clean formulations provided satisfactory cleaning results. Interestingly, the change from flat spray to solid stream nozzles did not change the results at all. Hence, it was concluded that for standard process parameters, the difference in cleaning results was mainly attributable to thermal and chemical energies.
As with the eutectic products, the cleaning agent was subsequently cooled (only pump agitation) to further establish the limits of this process. With unheated medium under standard process parameters (1 ft./min.), it was observed that only 30% of the results were satisfactory. Nevertheless, the amount and nature of the remaining residues were so miniscule, the authors argued their full removal upon further process optimization (judged as "0"). Simply changing to solid stream nozzles provided for the second successful cleaning result. Here the amount of cleaning agent required played an important role, all other process parameters remained unchanged (Figure 6, online). Further alteration of belt speed or spraying technology did not provide any additional positive test results. In total, only one ("0") result remained for the removal of lead-free no-clean formulations at ambient temperature.
For each product tested, favorable process parameters were achieved to fulfill the goals of this study. The authors concluded that for the no-clean lead-free products, the required chemical compatibility between contamination and cleaning agent is of utmost importance. Given an adequate compatibility, process parameters can be easily optimized to provide for a more cost-effective cleaning process.
Subjecting the soldering of all test vehicles to a fully inert atmosphere provided for a very different outcome however. In all cases, the unheated cleaning agent was able to remove contamination effectively. Changing the belt speed proved of no significance. The authors assert that an increase in belt speed even beyond 3 ft./min. would have been feasible (Table 5, online).
With the higher soldering temperature of lead-free, the baked-on flux residues polymerized to a higher extent, permitting full removal under standard process conditions. These findings stand in strict contrast to previously published findings on cleaning lead-free formulations. (1)
These results raise valuable questions of general cleanability versus the overall economics of conveyorized in-line processes. Using inert atmospheres in some cases might prove advantageous to meet certain throughput objectives, while adversely impacting the overall economics. Most cost-effective processes are generally achieved with optimized process conditions such as spraying technologies, belt speeds or unheated cleaning technologies.
From the results in this study the extent of strain exerted on flux has become noticeable. As mentioned, additional oxidation, polymerization and more aggressive flux residues are more thoroughly baked-on and therefore harder to remove. Modern cleaning technologies have therefore been designed to cope with 95% of no-clean formulations, and with adequate equipment optimization this goal remains realistic.
(1.) M. Bixenman, "Lead-Free Soldering: DOE Study to Understand its Effect on Electronic Assembly Defluxing," IPC Apex Proceedings, March 2004.
(2.) B.N. Ellis, Cleaning and Contamination of Electronics Components and Assemblies, ISBN 0 901150 20 7, 1997, and Dr. C. Lea, After CFCS?: Options for Electronics Cleaning Assemblies, ISBN 0901150 258, 1990.
(3.) Thus, various oxidation-induced reactions can be circumvented by an inert atmosphere. The immediate result is that flux residues are simply not baked-on as much as with an oxygen-containing environment.
(4.) For an excellent review on current lead-free trends and materials see: J. Lau, Electronics Manufacturing: with Lead-Free, Halogen-Free, and Conductive-Adhesive Materials, ISBN 0-07-138624-6, 2003.
(5.) H. Schweigart, "Lead-Free and No-Clean: A Contradiction in Terms?" SMT, October 2002.
Umut Tosun is an application technology manager at Zestron America (zestronusa.com); firstname.lastname@example.org. Dirk Ellis is an application engineer at Speedline Technologies (speedlinetech.com); email@example.com.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Cover Story|
|Date:||Apr 1, 2005|
|Previous Article:||Remember when surface-mount was the 'next big thing?' Reminisces on the scary "new" technology.|
|Next Article:||The future of boundary scan: combining JTAG with methods such as flying probe can improve test coverage and speed.|