Removal of lead-based solders as mandated by RoHS initiatives has created a need for PCB materials that withstand high processing temperatures. A great deal of attention has been given to the traditional PCB market where suppliers of FR4-type products compete to introduce material systems that are completely “lead-free capable.” New lead-free capable products are introduced on a frequent basis--almost as frequent as explanations why previously launched products failed to meet advertised expectations. The explanations have touched upon the effect of curing agents and other additives, applicability of test procedures, new rules proposed for designers, tightened tolerances during circuit fabrication, and additional processing requirements forced upon contract manufacturers.

For decades, suppliers in the high frequency circuit market have relied upon ceramic powders and specialized resin systems to manufacture base materials for use in wireless and, more recently, high-speed digital applications. Finely sized ceramic powders help to maintain stable electrical properties through a range of thermal conditions. Non-polar resin systems possess very low electrical loss and moisture absorption characteristics. The same filler and resin components that define the attractive electrical and thermo-electrical characteristics of these high-end composites also place many of these materials at the peak of the thermal reliability pyramid. This article compares the thermal properties, copper bond retention capabilities, and multi-layer board reliabilities of one family of high performance materials against the expectations of a lead-free capable material system. The RO4000® family, manufactured by Rogers Corporation (Rogers, CT), was chosen for this evaluation because it represents the most widely adopted high frequency material in the market, due to its electrical properties and ease of fabrication. In particular, much of the analysis used RO4350B™ material, a high frequency PCB laminate commonly used in many power amplifiers and high speed digital boards for the telecommunication market.

Thermal Properties

The thermal properties currently considered as predictive of a material’s compatibility with lead-free processing conditions include glass transition temperature, decomposition temperature, coefficient of thermal expansion in the Z-axis, and time to delaminate at a given temperature.

Glass Transition Temperature: The glass transition temperature (Tg) is the temperature or, more accurately, the range of temperatures where a thermoset material transitions from being rigid and glass-like (below Tg) toward becoming rubbery and more compliant (above Tg). Surpassing a material’s Tg is believed to result in decreased copper bond, increased risk of measling, and reduced reliability of plated-through holes. Tg can be measured using differential scanning calorimetry (DSC), thermal-mechanical analysis (TMA), or dynamic-mechanical analysis (DMA).

The Tg of materials supported as being lead-free compatible ranges from 175^oC to 220^oC. When measured by DSC, the high performance material being evaluated didn’t experience a glass transition over the entire test range of –50^oC to 300^oC.

Decomposition Temperature: The decomposition temperature (Td) of base PWB materials has been used to predict the temperature at which a material may be at increased risk of blistering and/or delaminating as pressures are created by the out-gassing of a decomposing resin system. The Td may also be interpreted as the temperature at which a thermosetting resin system permanently degrades as cross-links are thermally destroyed.

Td is measured using thermo-gravimetric analysis (TGA) and is set at the temperature where the material’s mass has been reduced by 5%. The Td of FR4-type products supported as being lead-free ranges from 300ºC to 350ºC, while the measured Td of the high performance material was greater than 390ºC.

Z-axis coefficient of thermal expansion (Z-CTE): Z-axis coefficient of thermal expansion (Z-CTE), as measured using thermo-mechanical analysis (TMA), is the rate at which a material’s thickness changes with changes to temperature. The expansion rate of most circuit materials increases remarkably when measured at temperatures in excess of Tg. To demonstrate this point, the data in Figure 1 was generated by testing the Z-CTE of four materials (high performance material, 150ºC Tg FR4, 175ºC Tg FR4, epoxy/PPO material) over a –50ºC to 250ºC temperature range. The Z-CTE of each material was calculated over 50ºC increments beginning at –50ºC to 0ºC and ending at 200ºC to 250ºC. The resulting data points were charted, connected by curves, and offered in chart form against the expansion characteristics of copper. Ideally, the expansion of a circuit board material and copper would not differ significantly. PTH failures during exposures to thermal extremes are a significant risk when the substrate material vs. copper Z-axis CTE differential is large.

Enlarge this picture Figure 1. CTE vs. Temperature range for various PCB material sets and copper.

As observed in Figure 1, the Z-CTE of the FR4 and epoxy/PPO materials increased dramatically as the test temperature surpassed the Tg of each material. The Z-CTE of these three materials through the 200ºC to 250ºC temperature range was close to 300ppm/C greater than the Z-CTE of copper. Obviously, significant stresses would result when multilayer boards (MLBs) made using these materials were exposed to lead-free reflow temperatures. In contrast, the Z-CTE of the high performance material was a relatively flat line (i.e., no glass transitions) and, through the entire temperature range, remained relatively close to the expansion characteristics of copper. As will be demonstrated later, a design engineer using this information to predict outstanding PTH reliability would find an excellent correlation.

A time to delaminate (Td) test is run using TMA equipment. A test sample, preferably a multi-layer construction, is held at a constant temperature until a delamination occurs. Delamination is detected as a sharp and sudden increase to the thickness of the sample. Materials are considered lead-free compatible should they survive 30 minutes at 260ºC (T-260) and 10 minutes at 288ºC (T-288). While it is rare to find an FR4-like material that survives ten minutes, the high performance material survived T-288 test conditions for longer than three hours.

Enlarge this picture Figure 2. T-288 testing for the high performance material.

T-288 testing was performed on a 125 mil thick MLB made using alternating core and prepreg layers of the high performance material. A slight and gradual reduction to thickness was observed during an initial 90-minute test exposure. The slight reduction to thickness was the product of a cure reaction that resulted when the material was exposed to the high temperature for the first time. The thickness of the material remained stable when exposed to a second 90-minute T-288 test (Figure 2).

Retention of Copper Adhesion

Reduction to copper bond can be permanent or reversible. Permanent changes can result when a substrate material is exposed to temperatures in excess of its Tg and/or Td. The effect can be cumulative and is affected by the extremity of the temperature, time at temperature, and the number of exposures. Reversible changes occur when a material is at elevated temperatures. Upon cooling, the peel strength may or may not return all the way to nominal depending upon the severity of the thermal exposure. Both types of changes can impact the reliability of surface mounted components at room temperature, during rework operations, and while the circuit boards are at operating temperatures.

Permanent Changes to Copper Bond

Enlarge this picture Figure 3. Target time/temperature profile used for reflow conditioning.

The peel strength to lead-free compatible materials is expected to survive multiple exposures to soldering temperatures. To determine the resistance to permanent changes, copper peel strength was measured on the high performance materials after each of ten exposures to molten solder and reflow cycles. The solder temperature was 288ºC and the exposure time was one minute. The target time/temperature reflow profile is provided in Figure 3.

Enlarge this picture Figure 4. Peel strength vs. reflow cycles and solder floats.

As can be seen in Figure 4, the peel strength of the high performance materials remained stable through the ten solder floats and the lead-free reflow cycles. In contrast, a 175ºC Tg, phenolic cured epoxy glass material lost copper bond during each exposure to 288ºC solder.

Reversible Changes to Copper Bond

The freedom to replace defective components on finished assemblies offers a significant cost savings to a contract manufacturer. Unfortunately, there aren’t many indirect tests that can accurately predict the re-workability of a material. Tg can be used to determine when a material will soften and experience a loss to copper bond, but can’t be used to measure actual copper bond at elevated temperatures. Procedures such as those outlined by IPC-TM-650 2.4.8.2 Peel Strength of Metallic Clad Laminates at Elevated Temperatures (Hot Fluid Method) can be used to measure peel strength at elevated temperatures, but the results can be misleading. These tests, typically run at 150ºC, almost always ignore the reduction to peel strength that occurs when a material is exposed to temperatures above its glass transition.

Enlarge this picture Figure 5. 175 Tb FR4 Cu peel vs. test temperature.

As an example, Figure 5 shows the peel strength of a 175 Tg, phenolic cured FR4 material when measured at 25ºC increments from 0ºC (ice bath) to 225ºC (reflow oil). At 150ºC, the copper bond has been reduced by only 25%. However, the peel strength is reduced by 97% when measured at 225ºC (up to 100ºC cooler than lead-free re-work conditions).

Figure 6. Test vehicle for SMT re-work simulation testing.

As bond at elevated temperatures can be misleading, the ability of materials to survive the thermal rigors of re-work in a lead-free environment was determined through re-work simulations. The test vehicle (pictured in Figure 6) was a standard SMT design that is used to provide certification training to re-work technicians. The design provided a variety of component types and connection styles.

Test boards were processed onto both the high performance material and a 175ºC Tg FR4 control material. Boards were finished with electroless nickel/immersion gold, immersion silver, and OSP. All boards were populated using Sn/Ag/Cu (SAC) solder paste and the reflow profile pictured in Figure 3. Three certified technicians were given three samples of each material and final finish type and asked to complete three rework cycles on each of the components. One rework cycle was defined as component extraction, removal of excess solder, and device replacement.

Extraction techniques evaluated included solder tip, solder tip fixture, hot air, IR, and conduction oven. Solder wick and vacuum were the methods used to remove excess solder. All components were replaced using the tip of a soldering iron. Where applicable (all extraction, solder removal, and component replacement techniques not using IR), operating temperatures were set at 371ºC (700ºF).

The high performance material survived all re-work cycles regardless of final finish, rework technician, extraction technique, method of removing excess solder, or re-work cycle count. One technician experienced a lifting of a few SMT pads during the first re-work cycle of the FR4 control. All three technicians reported pad lifting and blistering of the FR4 material during the second re-work cycle with most losses occurring during exposures requiring hot air.

Reliability of Multi-Layer Boards

Lead-free processing conditions increase the reliability risk to multilayer constructions in many ways. The risk of delamination, measling, and blister formation are increased due to the thermal degradation of resin systems and the increased pressure of water vapors at lead-free reflow temperatures. The risk of PTH failures is increased by the greater Z-CTE differential that exists between copper and most substrate materials at the very high temperatures (Figure 1). The shock of the high temperature exposures can result in immediate PTH failures (shock failures) or an accelerated rate of fatigue failures during thermal cycling.

PTH reliability MLBs were made using alternating 0.004" core and prepreg layers of the high performance material. MLBs were made to thicknesses of 0.060", 0.090", 0.125", and 0.175". Each MLB provided sixteen 5"x4" coupons with three clusters of 220 daisy-chained PTHs per coupon (Figure 7). The drilled hole diameters in one of each of the three clusters per coupon were 9.8, 13.5, and 19.8 mils. In considering the thickness of the MLBs and the diameters of the holes, this study considered aspect ratios ranging from 3:1 to 18:1.

Shock Conditioning

The resistance of daisy-chained PTH clusters was measured before and after exposure to the following shock conditions:

• Ten 60-second floats on 288^oC solder
  
• 60 minutes of pressure pot conditioning followed by a 288^oC solder float
  
• 100 hours conditioning at 85^oC/85% R.H. followed by a 288^oC solder float
  
• 24 hours submersion in 50^oC water followed by a 288^oC solder float

Fatigue Conditioning: Coupons were sent to an outside contractor where they were exposed to 10 reflow cycles using the profile pictured in Figure 3 and then exposed to the following list of thermal shock cycles:

• 1000 air-air thermal shock cycles from –55^oC to 125^oC
  
• 500 air-air thermal shock cycles from –55^oC to 150^oC
  
• 50 liquid-liquid thermal shock cycles from 0^oC to 225^oC

In the case of air-air cycling, the time at each temperature was 20 minutes and the elapsed transition time between temperatures was 60 seconds. The dwell time at each temperature during liquid-liquid cycling was ten minutes and the transition time between temperatures was less than 30 seconds. Resistance was measured constantly during air-air cycling and after every fifth cycle during liquid-liquid cycling.

Using a 5% increase over initial high temperature resistance readings as the threshold limit, no failures were detected on any daisy-chained clusters regardless of aspect ratio or conditions of shock or fatigue (electrical pass). In addition, there were no signs of blistering, measling, or delamination observed during inspections of cross-sections (mechanical pass). The surface of materials exposed to air-air cycling did darken in color significantly.

Conclusion

Materials once chosen for electrical performance may be among the most reliable choices for designs requiring lead-free solders. Table 1 shows that at least one of these materials meets and exceeds the requirements of the lead-free world.

Enlarge this picture Table 1.

Art Aguayo is the Marketing Development Manager, Rogers Corporation.
Michael Kuszaj is a Sr. Technical Engineer, Rogers Corporation.

For more information, email Cynthia Kiss, Marketing Communications, Rogers Corporation, at cynthia.kiss@rogerscorporation.com.