The Effects of Moisture in PCBA Manufacture
BY MIKE CUMMINGS, TECHNICAL DIRECTOR, TSI CONSULTANTS, MOTHERWELL, UK
Moisture plays a key role in manufacturing too little and things dry out, ESD increases, dust levels are higher, stencil apertures block easier, the stencil wear increases, and throughput has been shown to go down. Too much and materials absorb moisture, causing delamination, pop corning, solder balls. Moisture also lowers the Tg value of materials increasing dynamic warp during reflow.
Surface moisture introduction
MOISTURE LAYERS ON METALS, ETC.
Nearly all solid surfaces (Like metal glass ceramic silicon, etc.) have a moisture layer (monolayer or layers), this moisture layer becomes visible when the surface temperature is at the dew point temperature for th e ambient air conditions (temperature, humidity and air pressure). Friction between metal on metal increases as humidity goes down, @ 20%RH and below friction is increased 1.5 times that of friction @ 80%RH conditions.
MOISTURE LAYER ON ORGANICS PLASTICS, ETC.
Porous or hygroscopic surfaces (Epoxy, plastics flux etc) tend to absorb these moisture layers, and as such these surfaces lower the visible dew point (condensation), none the less the moisture layers are present in the bulk of the material surface.
It is these surface monolayer’s of water that permeate through plastic devices (MSD) when the monolayer’s approach 20 monolayer’s in thickness, the moisture absorbed by these monolayer’s of water ultimately result in pop corning issues during reflow soldering.
Moisture exposure of the plastic devices should be controlled IAW IPC-STD-020.
MOISTURE EFFECTS DURING MANUFACTURE
Moisture causes various manufacturing effects; in general the moisture is not visible, detectable (with the exception of weight gain), however the consequences are:
- Blow holes, voids, solder splatter, solder balls and voiding in underfill to name a few.
The worst kind of moisture for any process would be condensation; the aim should be to ensure, surface moisture is controlled within allowable limits that do not adversely affect the materials or the process.
What is a controlled allowable limit?
The dew point versus the substrate temperature levels are well recognised measures in nearly all coating processes (Spin coating on silicon semiconductor fabrication, mask films and coating on metals), however the board assembly manufacturing sector has never considered environmental issues to be a cause for concern (although we in the global consumer team have released environmental control guidance and specified various parameters that should be controlled).
As technology drives processes to finer features, smaller parts and higher density boards we come ever closer to the microelectronics and semiconductor industry critical environmental requirements.
We already understand the issues of dust control and the problems this brings with machines and processes; we know that High RH levels allow moisture to build up in components (IPC-STD-020) and boards causing material degradation, process and reliability failures.
We have driven some equipment manufacturers to control environments within machines and material suppliers to formulate materials that can operate in more hostile environments, we have up till now blamed moisture issues on solder paste, stencils underfill materials etc.
Typically coatings like solder paste are solids suspended in some form of solvents, water or solvent mixtures, the main features of these liquids on application to the metal substrate is to form a tacky bond to the metal surface, however if the metal surface approaches the ambient dew point, water could partially condense, trapping moisture under the solder paste, causing adhesion issues (blisters etc under the coating).
In the metal coating industry dew meters are used to ensure adherence of the coating to the metal substrate.
Basically this instrument measures accurately the surrounding RH levels on or around the metal substrate and calculates the dew point, it compares this result with the measured substrate surface temperature of the part and then calculates the ΔT between the substrate temperature and the dew point, if the ΔT is less than 3 to 5°C of the surface temperature, the part cannot be coated as poor adhesion and voids will result.
How moisture builds up relative to RH and the dew point.
At around 20% RH the board substrate and pad has a monolayer of water molecules hydrogen bonded to the surface (not visible). The water molecules are not mobile and in this state the water is benign, even electrically. Some drying out may occur depending on the cards moisture history on the shop floor but this moisture exchange evaporates out maintaining a constant single monolayer on the surface.
Further monolayer’s layers form, dependant on the moisture uptake from the substrate surface. Epoxy, flux and OSP have high water uptake – metal surfaces have none.
The metal pad (copper) will draw more moisture, even through OSP, as the RH levels increase towards the dew point, more monolayer’s (Multi-layers) are formed. The critical point is 20 monolayer’s and above where bulk water can form and electrons can flow, dendrites and or CAF can form, although contamination would need to be present, as the dew point temperature is approached (dew point/condensation) porous surface like boards absorb the bulk water readily, effectively lowering the dew point temperature or the temperature where bulk water can be present on hygroscopic surfaces, these moisture hugging surfaces absorb critical amounts relative to our processes causing reduced flux efficiency, outgassing during underfill and reflow as well as poor paste release issues at screen printing.
Solder paste in effect is a similar process to coating materials such as paint, in as much as the flux must adhere to the substrate surface to effect good paste release from the stencil aperture. Solder that approaches the ambient dew point will have reduced tack strength and therefore poor paste release.
The ECU unit air temperature should where possible adhere to the metal coating rules for dew point, namely the substrate temperature should not exceed 4°C ±1°C for metal coatings like gold or tin of the dew point temperature, and for porous/hygroscopic surfaces like OSP, the minimum temperature for our requirements should be ≥5°C
DEK Printer settings
DEK ECU actual set temperature on the shop floor is 26°C. With a machine humidity of 45%RH the substrate calculated dew point within the machine environment would be 15°C. The coldest substrate temperature recorded prior to entering the screen printer that day was 19°C, the Delta T (Difference between the substrate temperature and the Dew point) of the coldest substrate measurement of the card prior to entering the printer is (19-15) 4°C from the Dew point and the substrate temperature, and would just pass ASTM and ISO coating specs for safe coating on metal (a minimum of 4°C ±1°C) but likely fail in the field. The porous surface temperature (upper limit of the coating specification 5°C) and therefore we can assume the card will absorb moisture.
If we placed a cold (19°C card) into other machines such as Fuji with a shop floor humidity >60%RH we would have a Delta T of 2°C which would fail all the ASTM/ISO coating specs because the board is too wet. An optimum good setting would be ≥5°C above dew point.
Shop floor measurements
Surface moisture is dependant on the surface temperatures vs the ambient air temperature and relative humidity (dew point) as we approach the dew point thicker multilayer’s of water are formed making adhesion (tackiness) of paste etc. less efficient resulting in poor paste release from the damp pads.
Below are critical temperatures calculated from various temperature and humidity ranges experience on the shop floor. There was three recorded substrate temperatures 19°C, 20°C and 21°C, the charts below indicate safe facility humidity’s vs facility temperature ranges to avoid moisture absorption, (internal machine environments need to be measured).
The warmer the card the less critical is the facility environment.
Dew point surface tests
All substrate surfaces have poor wetting as humidity increases (>50%RH) and as it approaches the dew point to within 4 to 5°C. We set up a trial with ambient room RH levels around 43%RH, basically too low for a real measure of shop floor worst case (60 to 65%RH), where the impact on processes are more prevalent, none the less we conducted the test by placing a clean card in the shop floor fridge for ½ hour till cold (required for dew point temperature with low humidity shop floor values), when tested with Dyne pens, the dyne value had dropped from >40 Dynes to 37 Dyne-cm, sufficient to demonstrate an impact, although measurements at higher humidity’s and room temperature would be more relevant and the drop I feel sure would be more dramatic.
Plastic devices moisture sensitivity
Hydrophilic materials (FR4, plastic components) can absorb water from the atmosphere or desorbs (sweat) it out in an effort to stabilise or approach equilibrium with the surrounding atmosphere.
Plastic components are pre treated by the manufacturers before they are sealed in a foil bag. For example they can be baked for 24 hours at 125°C. These very dry parts (<10%RH in the bag) are not in equilibrium with the shop floor humidity (typically 45%RH), and they rapidly (over exposure periods rated on the foil bags) adsorb moisture from the shop floor atmosphere.
Therefore, the more humid the environment, the more liquid can be adsorbed. As the temperature increases, desorption starts to prevail over adsorption. The higher the temperature is, the more energetic the water molecules are and the easier they leave the surface and return to it.
It’s not easy to control humidity on a shop floor, but understanding the critical measures like the PCB surface dew point, gives you the power to control the impact.
For example keeping your printer ECU close to room temperature will reduce the possible dew point issues for paste release, being aware that dry cabinets can produce cold devices in hot humid rooms, or that low humidity increases metal on metal friction producing stencil wear.
Mike has been in manufacturing related industries for over 40 years he started in Radar system test before moving into Defect analysis, He started TSI in 1992 and has trouble shot everything from open cast mining through to Subsea electronic failures. He was appointed by the Pentagon Certification Board as a Cat A MIL-STD-2000 Senior Instructor Examiner in the 1980’s.