Components baked at 70˚C for 24 hr. in a vacuum passed IPC and EIA testing.

Baking surface-mount moisture-sensitive components is about the most loathsome task a board assembler performs. It’s time-consuming, expensive and not value-added. It encourages tin whisker generation, promotes intermetallics and metal migration, and degrades solderability.

Baking is performed because moisture absorbed in surface-mount packages expands into steam at solder reflow temperatures.¹ At a given moisture content level, the force of the expanding moisture is sufficient to fracture the component. Baking remains the most efficient way to remove the damaging moisture.

Assemblers may be facing even more component baking. This is because J-STD-020D addresses higher temperatures required by Pb-free solders.² Some experts feel this could raise the moisture sensitivity level (MSL) of specific SMT packages by one or more levels. The higher the MSL rating of a component, the shorter its factory life (the time a package can be exposed to 30˚C/60% RH before it must be dried).

Most surface-mount packages arrive for placement in tape-and-reel or in plastic shipping tubes. Because commercial tapes, reels and tubes are rated at maximum temperatures of 60˚C, J-STD-033B lists just three drying processes for taped surface-mount packages.³ Assume taped, MSL-3 devices that have exceeded their floor life by ≤72 hr. The processes are:

1. Remove packages from the tape; place in Jedec trays and bake at 125˚C for 48 hr.; return to new tapes. Because of its short duration, this is the most common process. However, it has several major disadvantages:

• Scraps the original tapes.
• Needs the correct Jedec trays.
• Requires new tapes and a taping service.
• Risks mechanical or ESD package damage.

2. Purchase all moisture-sensitive surface-mount devices in polycarbonate tapes and reels that can withstand 90˚C. Bake for eight days. This option has disadvantages, too:

• Polycarbonate tape and reel materials cost much more than commercial tapes and reels.
• Limited number of suppliers.
• Additional costs not justified if packages never exceed their floor life.
• Baking must be performed at ≤ 5% RH. This excludes standard convection ovens.

3. Leave packages in polystyrene/polyester tapes and polystyrene reels or polyvinyl tubes and bake at 40˚C for 67 days.

• Baking must be done at ≤ 5% RH.
• Baking duration is unworkable for most PCB assemblers.

J-STD-033B places a costly baking burden on the assembler. To mitigate this, it permits “alternative drying processes” with the following note: “Table 4-1 is based on worst case molded leadframe SMD packages. Users may reduce the actual bake time if technically justified (e.g., absorption/desorption data, etc.).” For lack of a better term, these alternative drying processes “conform” to J-STD-033B.

In 1994, a major cellphone maker internally reported success in driving sufficient moisture from taped MSD packages to prevent delamination at solder reflow.⁴ The company baked components at 70˚C in a vacuum for 24 hr. The firm claims to have dried more than 10 million packages without experiencing package delamination using this method. If true, this process is a vast improvement over the J-STD-033B drying processes.

The report made several suspect conclusions. Yet, in light of today’s more demanding requirements, we investigated a similar process. In our design of experiments, we worked closely with key members of the IPC/JEDEC subcommittee who authored J-STD-033B. We wanted to ensure specification conformance.

Vendor-generated data are commonly viewed with suspicion. For this reason, we limited involvement in the experiment to creating the DoE, supplying sample materials, funding transportation and some services. Independent industry organizations performed all experimental procedures. Participants were Freescale Semiconductor, Advantek Inc., Solectron-Texas, Priority Labs, ST Microelectronics and Electro-Comp Taping Systems.

‘Proof of Process’

Experiment goals. We established the following goals:

1. The experimental process must conform to J-STD-033B criteria for an alternative package drying process.

2. The novel drying process need not fully reset the J-STD-033B specified package “floor-life clock.” However, it must remove sufficient moisture from the taped or tubed packages to prevent package delamination during Pb-free solder reflow.

3. The experimental process must not:

• Significantly exceed cover tape peel strength specifications.⁵
• Functionally deform carrier or cover tapes.
• Deform commercial high-impact polystyrene reels.

Experimental samples. Tables 1 and 2 identify the experimental components and carriers.



DoE.

1. All sample packages were initially inspected by a C-mode acoustical microscope (reflective mode) to ensure they were free of delamination.
   
   
2. All samples were baked in a convection oven at 125°C for 72 hr. to drive moisture from the packages.
   
   
3. The “dry-weight” of each sample was determined in grams to the fourth decimal place.
   
   
4. All sample packages were placed in 44 mm Advantek carrier tape and cover tape.
   
   
5. Taped samples were moisture conditioned at 60°C ± 2° at a relative humidity of 60% RH ± 3% for 168 hr., to simulate 30 days exposure to 30°C/60% RH.
   
   
6. One lot of taped samples was removed from the tape and weighed to determine its “wet-weight” in grams to the fourth decimal point.
   
   
7. All taped samples were baked in the novel apparatus for 24 hr. at 70°C ±0.1°C and ≤ 5%RH.
   
   
8. Packages from one sample were removed from the tape. The total dry weight of the packages was determined in grams to the fourth decimal place, to assess the effectiveness of the 70°C novel process.
   
   
9. The remaining six samples of taped packages were placed in a temperature/humidity chamber maintained at 30°C ± 2°/60% RH ± 3%.
   
   
10. Samples were removed from chamber and subjected to Pb-free reflow per J-STD-020C, one sample per 24 hr. This spanned the six-day factory floor life of the samples.
    
    
11. After Pb-free reflow, samples were inspected by C-SAM (reflective mode) to detect any internal package delamination.
    
    
12. The longest period of 30°C/60% RH storage showed no sign of internal package delamination after Pb-free reflow was reported as the number of solder reflow safe days gained by the novel baking process.
    
    
13. Empty 44 mm tape assemblies were baked in the 70°C novel apparatus for 24 hr., then sent to Advantek for testing, analysis and review. A sample of unbaked, assembled cover and carrier tape was submitted as a control.
    
    
14. High-impact polystyrene reels were subjected to 70°C for 24 hr. of convection baking. Reel R-1 was placed vertically, supported on its rims in a 70°C oven. Reel R-2 was horizontally arrayed and axially supported, as it would be in the novel oven.

Experimental findings. Temperature is the prime factor in determining the rate that moisture is extracted from a plastic molding compound. The higher the temperature, the more rapidly moisture is expelled from the material.

Increased bake duration at a lower baking temperature is not effective in removing moisture from a package. For example: Reducing the temperature of the novel baking cycle to 60°C and increasing the duration to 48 hr. resulted in negligible weight reduction and 100% package delamination after Pb-free reflow.

The major portion of the ingested moisture is removed from a package during the first 24 hr. of baking. Moisture loss beyond 24 hrs. of baking continues at a very slow and declining rate.

Figure 1 compares the weight of ingested moisture with the weight of moisture removed by the experimental process. It is important to note that the samples were required to be “moisture conditioned” to the level expected for 30 days’ exposure to 30˚C/60% RH. Less ingested moisture would be seen if the samples had been exposed for only six days of floor-life at this atmosphere.


The rate of a drying process slows as the package’s moisture content nears the relative humidity within an oven.

The novel process uses a vacuum pump to reduce the experimental chamber to ≤100 Torr. Vacuum is the most economical way to lower chamber humidity.

Contrary to some claims, vacuum was not found to accelerate the rate that moisture is withdrawn from a plastic package. Also, the amount of moisture removed from a package was consistent, regardless of its position within a reel.

Sufficient moisture was driven from the taped packages so they could be exposed to 30°C/60% RH for three days and be Pb-free reflowed without suffering internal delamination (Figure 2).


The number of safe reflow soldering days could be extended by placing dried packages in a desiccant cabinet at ≤5% RH immediately after drying with the novel process.

A simple and inexpensive experiment would be required to determine the additional solder reflow safe days that the novel process adds to a specific package.

Taped assemblies were processed with the novel process and submitted to Advantek to determine how the process affected their function (Figure 3).

• Cover tape deformation. No functionally adverse effect resulted from the novel bake. Advantek observed: “The cover tape had taken on some minor wave conditions, visible along the edges of the sealing flanges of the carrier tape. In addition, the cover tape does not appear to be lying as flat on the carrier, as compared with the pre-bake sample. No pocket sidewall warpage was evident and no changes appeared to have occurred that prevent normal removal of a device from a pocket.”⁶
  
  
• Carrier tape deformation. Advantek submitted the 70°C novel baked carrier tape to AOI inspection. They reported: “Key dimensions of the post-bake tape samples were inspected and compared with the (Advantek manufacturing specification). In the case of this square pocket, pre-bake Ao and Bo dimensions are both 25.6 mm. Post-bake measurements ranged from 25.3 mm to 25.35 mm, reflecting shrinkage of 1% to 1.2%, which should not prevent easy extraction of a package.”⁶
  
  
• Adhesive bond strength between the cover and carrier tapes. Excessive temperature can raise the strength of the adhesive between the carrier and cover tape to a point where the chipshooter can’t easily pull the cover tape free from the carrier tape. EIA-481C specifies the peel strength between the 44 mm cover and carrier tapes used in the experiment be no less than 10 g and no more than 130 g.⁷
  
  
• Cover tape peel strength testing. Taped assemblies exposed to novel baking were cover tape peel-strength tested by Advantek (Figure 4). Advantek reported, “The two spikes above 130 g can be attributed to minor waves that developed along the edges of the sealing flanges of the carrier. If the carrier is not relatively flat, the peel force test … can have the peeled cover tape come into contact with cover tape that is still adhered. This friction can cause the spiking seen on the chart and would not represent a true increase in the adhesive bond between the cover and carrier.”⁸
  
  
  
  
  
  
• High-impact polystyrene reel deformation. 70°C is ≈ 10°C above the stated softening point of the high-impact polystyrene material used in commercial reels. Figure 5 shows a 13" reel that was placed vertically, resting on its rims, in an oven at 70°C for 24 hr. During baking, it deformed to the point where it could no longer support itself and fell to a horizontal position. Obviously, this was a reel - deformation failure. Figure 6 shows a high-impact polystyrene reel that was axially supported and baked at 70°C for 24 hr. No significant warpage was seen in either 13" or 7" reels.⁸
  
  
   
  
  
  
• Process temperature control is vital to the success of the novel process. The novel oven temperature is held to 70°C ±0.1°. Figure 7 illustrates the “wetting” adhesive failure that resulted from baking tapes at 72°C. The adhesive migrated over the entire underside of the cover tape.7 A key to achieving this rigidly controlled temperature is a radical vertical, low thermal mass cylindrical chamber design.
• Purge gases. The experiments proved that purge gases, CDA or nitrogen, were not required.

References

1. IPC-SM-786, Impact of Moisture on Plastic IC Package Cracking, December 1990.
2. Tom Adams and Steve Martell, “The Moisture Sensitivity Standard Goes Pb-Free,” Circuits Assembly, January 2007.
3. Ibid.
4. Rae Levy and Ishayau Perelman, Vacuum Oven Study, Aug. 25, 1997.
5. EIA-481C, 8 mm through 200 mm Embossed Carrier Taping and 8 mm & 12 mm Punched Carrier Taping of Surface Mount Components for Automatic Handling, November 2003.
6. Peter A. Swanson, Advantek Inc.
7. EIA-481C, 8 mm through 200 mm Embossed Carrier Taping and 8 mm & 12 mm Punched Carrier Taping of Surface Mount Components for Automatic Handling, November 2003.
8. Peter A. Swanson, Advantek Inc.

Charles S. (Stu) Leech Jr. is director of engineering at Innovative Drying Co. LLC; sleech@qwest.net.