Cleaning process design for high reliability printed circuit assemblies

In the era of no clean, original equipment manufacturers outsourced Class 1 and Class 2 electronic assemblies that supported a true no clean process. Class 3 and some Class 2 high reliability electronic assemblies required removal of flux residue and other residual contaminates, even while many other assemblies support a true no-clean process. Additionally, miniaturization and reliability concerns drive more customers to specify cleaning as a required step in the manufacturing process. As OEMs outsource Class 3 high reliability contracts, electronic assembly cleaning becomes an additional process that EMS must incorporate within manufacturing operations. To support this need, contract assembly houses require cleaning technologies that work well on a wide range of soldering materials and cleaning equipment designs.

Additional constraints imposed by the transition from eutectic Sn/Pb (tin-lead) to Pb-free alloys have creased the cleaning challenges for EMS assembly operations. Assemblers may view Pb-free assembly as a future challenge, but in actuality, there are current challenges cleaning Sn/Pb assemblies. The conversion to Pb-free and new component Pb finishes, such as tin plated or gold over nickel/palladium, are harder to solder, leading to a need for more aggressive flux materials.

Introducing Class 3 customers to a no-clean EMS manufacturing shop presents processing challenges. High reliability contracts may require the introduction of new solder pastes, wave and touch up fluxes, which must be integrated into the existing operations. Class 3 soldering materials are rosin-based materials (type ROL0 per J-STD-001) and marketed by the material suppliers as no-clean materials. Cleaning performance may change as new soldering materials are introduced. Processing highly dense substrates creates challenges for removing flux residue from under low standoff components and fine pitch devices.3 Classes of assemblies
IPC association connecting electronic industries places electronic circuit assemblies into three classes:
• Class I assemblies includes products suitable for applications where the major requirement is function of the completed assembly.
• Class II assemblies includes products where continued performance and extended life is required, and for which uninterrupted service is desired but not critical. The typical end-use environment would not cause failures.
• Class III high performance electronic products includes assemblies where continued high performance or performance-on-demand is critical, equipment downtime cannot be tolerated, end-use environment may be uncommonly harsh, and the equipment must function when required, such as life support or other critical systems.

Cleaning process design
Electronic manufacturing assemblers require cleaning technologies that work well on a wide range of flux technologies. Best case scenario is a true drop in to existing process cleaning equipment. The manufacturing process must control variability, meet stringent environmental concerns, and provide a low cost of ownership. The methodology for developing an optimized cleaning process considers the following:
• Evaluation of the substrate
• Evaluation of the contaminant
• Application lab development
• Evaluation of available cleaning technologies
• Evaluation of the manufacturing process
• Implementing the process
• Addressing process issues
• Validating the process

Evaluation of the substrate
The first step in developing a cleaning methodology starts with closer review of the electronic assembly.
• Part composition
• Function
• Components
• Size/geometry/configuration
• Handling issues
• Part unique considerations/restrictions
• Compatibility review for both cleaning effects and requirements

Part composition, size and geometry can create low clearances, sandwiched parts with little egress, and residues that are very difficult to remove (Figure 1). Small and light weight parts typically must be in a fixture or basket as they track through the cleaning process. Cleaning fluid compatibility to components, board surface, metallization, anodized aluminum and marking labels and inks must be tested and considered. Part unique restrictions may limit some components from being subjected to the cleaning process.

Evaluation of the contaminant
With a clear understanding of the unique part consideration and restrictions, the next step in designing a cleaning process considers the evaluation of the contaminant. The process starts by testing the soil:
• Composition
• Physical properties
• Quantity
• Process characteristics
• Upstream/downstream process conditions for both cleaning affects and requirements
• Subsequent process for both cleaning effects and requirements

Solder pastes, paste fluxes, and wave fluxes perform a number of important functions. Print life, tackiness, solder balling, voiding, wetting and appearance are a few of the more important material considerations. Transfer efficiency, relaxation, stencil release and slump must be considered. The static cleaning rate is an excellent measure for gauging the speed of dissolution or removal.

Flux residues clean at different rates based on the flux make-up, time after reflow, reflow temperature, mechanical energy and the cleaning fluid design. Water-soluble flux residues typically clean at a faster rate than do rosin flux residues, which typically clean at a faster rate than low solids synthetic flux residues. Flux residue becomes more difficult to clean with the passage of time after reflow. Higher reflow temperatures allow the lower molecular weight solvent molecules to evaporate at a faster rate, leaving higher molecular weight resin molecules, which increases the difficulty of cleaning the residue. Cleaning fluid designs either dissolve or react with the flux soil, which influences the static cleaning rate.

As a result of these factors, the process cleaning rate will vary depending on residue and upstream process consideration (Figure 2). Soak reflow profiles subject the solder paste to longer time at, above, or near liquidus. The number of reflow cycles before cleaning increase cleaning difficulty. Hand soldering and repair may influence cleaning effects. Each of these processing decisions must be tested to understand cleaning efficacy.

Application lab development
Application testing labs are fully equipped for product development, analytical analyses, performance analyses and subsequent manufacturing quality assurance (Figure 3). Leveraging the capabilities of outside application labs takes advantage of the supplier’s process knowledge. Leveraging cleaning chemistry and equipment supplier's process development capabilities and knowledge provides EMS engineers with a higher assurance that the process will work.

Several cleaning equipments and cleaning fluids are available options. Selecting the proper equipment and cleaning fluid depend on the number and characteristics of the parts. Compatibility issues should be studied before locking in a process. Compatibility constraints can present difficult issues once the equipment and cleaning chemistry are running on the process floor. In the summary, application lab testing provides process options for addressing:
• Throughput requirements
• Soil characteristics
• Cleaning chemistry options
• Compatibility constraints
• Process recommendation

Cleaning fluid evaluation
Cleaning fluid selection can improve or take away from the cleaning process. Cleaning fluids vary in their design based on solvency, saponification, wetting (surfactancy), inhibition, and defoaming characteristics. The best cleaning fluids optimize and build performance characteristics that effectively accomplish several tasks in combination. The challenge is to create a universal cleaning fluid that works well on many flux residue types (Figure 4). Cleaning chemistry firms who study the many soil types design universal cleaning fluids that work to open the process window, which allow users to select different flux types without a major impact on the cleaning process.

Under optimal conditions, process engineers designing a cleaning process first identify the cleaning requirements, the cleaning solvent, and the selection of mechanical equipment followed by the design of the cleaning and statistical process controls. Many EMS contractors faced the limitation of implementing the cleaning process using existing cleaning equipment while being constrained on cleaning time and temperatures.

Class 1 and 2 circuit assembler’s who clean commonly use water soluble solder pastes and fluxes (type ORM0 and ORH0 per J-STD-001). Low standoff SMT components, such as TSOPS, area array packages and QFN packages represent common cleaning challenges when cleaning water soluble flux materials. Some assemblers use aqueous cleaning fluids at low concentration to reduce surface tension and improve cleaning efficacy.

Process constraints within the EMS may eliminate a number of commercial cleaning products from consideration. Either the materials are not compatible with the existing equipment (primarily pump and filter seals) or required modifications to the existing equipment. Cleaning process optimization requires an integrated approach aimed at understanding the substrate, contaminate, manufacturing process and unique processing conditions. An optimized cleaning process must address:
1. Cosmetics: No visible residues
2. Corrosion and/or leakage currents: Assemblies must be free of ionic or hydrophilic residues that could eventually react with moisture in the field and cause degraded performance or failures.
3. Electrical performance: No degradation of RF systems operating at high clock speeds through RF skin effects and mismatched impedances in input and outputs lines.
4. Testability: Residues on board must be penetrated by the existing in circuit test equipment with a minimum of changes to the test pins or fixtures.
5. Adhesion: No adverse impact on conformal coat adhesion or potting materials.

Evaluation of the manufacturing process
Developing a clear definition of the substrate and contaminant, the manufacturing engineer's next step is to evaluate the manufacturing process. Equipment designs are influence by throughput requirements. Factors worth consideration are as follows:
• Batch or continuous machine designs
• Process requirements/limitations
• Equipment/facility considerations
• Evaluate tank life characteristics

The timing and sequence of events in a cleaning process are critical. Each section or step in the process requires careful thought and understanding. Inline cleaning processes use a pre-wash that thoroughly wet the parts with the wash solution chemistry. Sufficient flow and contact time help elevate the assembly to wash temperature. This facilitates the full static-cleaning rate, from the cleaning chemistry. Residues are softened and dissolved in the soak zone between the pre-wash and the wash segment.

In the wash zone, the part should see several high impingement scourings, punctuated by brief soak periods. This optimizes the static rate by maintaining fresh cleaning fluid and optimizes the dynamic rate by focusing the maximum physical energy at the part surfaces. In the chemical isolation section there should be ample impingement force in the first air jet manifold to strip the wash chemistry from the assembly so that it can be returned to the wash tank. A second jet air and/or water manifold should thoroughly remove any remaining residue to drain. In the power rinse section, a series of high-pressure nozzles removes and dilutes the remaining ions using de-ionized water. A low flow/pressure final pure rinse of DI water removes trace ions following by displacement air drying. Figures 5-10 illustrate process requirement and limitations of aqueous inline cleaner processes.

Note the overflow line illustrated in (Figure 5). Excessive losses occur when the wash tank sensor fails and the wash tank overflows to drain. Wash tanks have a tendency to foam at low operating temperatures, low wash concentrations, or heavy soil load. When a wash tank foams, air entrainment increases the wash level by dropping out the high level sensor (Figure 6). This condition forces wash fluid to drain causing high chemical loss down the overflow.

(Figure 7) illustrates high chemical loss due to misdirected spray nozzles. Entrance and exit to the wash section is also contained from curtains fabricated out of elastomers. With time the curtain material will become brittle to a point that the curtain losses the ability to deflect and maintain the wash solution within the wash chamber. When boards track through the wash section, cleaning fluid is carried forward and backward causing high chemical losses.

(Figure 8) illustrates high chemical loss from excessive exhaust flow. This issue is magnified by the temperature of the wash tank. Higher wash tank temperatures elevate the water mist nearer to a gaseous state, which increases loss up the stack. High wash tank temperatures see less condensation from cooling effects.

(Figure 9) illustrates corrective actions to address high wash tank consumption losses. To address wash tank overflow conditions, check level controls and concentration of the cleaning fluid. Foaming is typically the root cause for this condition. Additionally, check for leaky drain values. To correct high chemical losses due to misdirected spray nozzles, re-align manifolds.

To address high consumption through the ventilation stack, three corrective actions consist of optimizing ventilation flow, reducing wash temperature and installing mist arrestors (Figure 10). Mist arrestors cool and condense saturated air vapor, with the larger water droplets flowing back into the wash section. Placing umbrellas over the top spray manifolds reduces saturated mist in the wash space, which reduces losses from ventilation.

Batch style cleaners are different from planar type batch and inline cleaners. Dishwasher style cleaners are most efficient when designed for maximum flow, not impingement, as potential energy and capillary forces must be relied upon for cleaning surface energy in the shadowed areas (Figure 11). Using too high pressure tends to atomize the spray reducing jet velocity much faster over distance and increases the splash interference with other jets.

Bath life is a function of critical soil load in the wash tank. Inline cleaning processes lose liquid from ventilation and dragout into the rinse sections. Liquid losses are made up with fresh cleaning chemistry and water. In a dynamic process inputs minus outputs commonly form a steady state condition that is less than the critical soil load. Operating under the critical soil load extends the bath life (Figure 12).

Hand-held and automated processes are available for controlling wash bath concentration. Hand-held refractometers measure the light refraction, which can be correlated to the concentration of the cleaning fluid in the wash tank, Alkaline cleaning fluids measure alkalinity to an end point, which is commonly referred to as titration. Titration correlates alkalinity to the end point with concentration.

Automated programmable control units continuously measure the wash bath concentration using refractive index. These units add both water and cleaning chemistry, as required, maintaining wash tank concentration. Automated process control units hold the concentration at tight tolerances, which supports statistical process control (Figure 13).

Process implementation
Process engineers who follow the design methodology suggested have completed the following:
• Evaluation of the substrate
• Evaluation of the contaminant
• Process application lab development
• Cleaning fluid selection
• Cleaning equipment and control equipment selection

Implementation following proper design methodology reduces risk. Process engineers who buy the cleaning equipment first and then work to integrate a cleaning chemistry may end up fighting nagging process issues over time. Once capital is allocated, the engineer may be faced with making a bad process work. Following design rules assure that the process works when implemented. Instead of wasting countless hours fighting process issues, the engineer's valuable time can be allocated elsewhere.

Addressing process issues
When process issues occur, the first step to resolving the problem is a root cause analysis. An operator may be reporting foaming in the rinse tank. Proper root cause may isolate the issue to the air movement through the machine. An operator may report flux residue on a part. Proper root cause analysis may reveal multiple reflow cycles. An operator may report excessive wash consumption. Root cause analysis may identify boards placed on the conveyor flush to each other, which carries the wash solution to the chemical isolation section. Any number of issues may occur during normal processing conditions. By performing proper root cause analysis, the correct problem can be addressed and resolved.

Validating the process
IPC J-STD Requirements for Soldered Electrical and Electronic Assemblies provides the design rules for meeting the standards for Class I, Class II, and Class III assemblies. Test boards processed through the designed cleaning process are validated using outside testing labs.

Conclusions and recommendations
Class 3 and many Class 2 products require cleaning for functionality and reliability concerns. This is forcing a reexamination of no-clean soldering materials and cleaning processes. Just as solder pastes continuously evolve, advanced cleaning materials are needed with improved environmental properties, better effectiveness on a wide range of contaminants, and flux residues. Engineered cleaning solutions must work in conjunction with customer constraints, existing equipment designs, and new solder materials to provide a total solution with in-line, batch and bench cleaning systems and methods.

The methodology presented identifies the metrics needed to validate the cleaning process. Cleaning material and cleaning equipment suppliers provide EMS companies with a valuable resource. These vendors provide application testing labs, analytical testing, and qualification testing. Their engineering staff can be leveraged to optimize process conditions. EM

About the Author
Mike Bixenman is the CTO of Kyzen Corporation. For more information, please email:
mikeb@kyzen.com