| Created: 10/24/95 |
Updated: 8/05/98 |
Moisture Absorption in Printed Wiring Boards
Project Number: C96-10
Point of Contact:
Dr. Michael Pecht
CALCE EPSC
Email: pecht@calce.umd.edu
Phone: (301)-405-5323
Fax: (301)-314-9269
Objective
To investigate the electronic industry's most common laminates (FR-4, HTFR-4,
PI, CE, and BT) for their equilibrium levels of moisture and diffusion
rates and to develop a method which allows in-situ monitoring of moisture
absorption in printed wiring boards.
Background
Since their invention in the 1930s, printed wiring boards (PWBs) have become
the back bone for component interconnection in products ranging from computers,
to telecommunication products and to consumer electronics. PWBs are
generally comprised of several layers of laminates upon which electrical
connections are fabricated.
Laminate manufacturing is a process consisting of encapsulating a ply
(or several ply) of glass fabric within a polymeric resin. The fabric-resin
combination provides dielectric as well as mechanical and thermomechanical
properties. However, since the resins used in laminates are hydrophilic,
the assembly can be susceptible to degradation and performance failure
mechanisms driven by environmental moisture.
Moisture can reduce the quality of lamination, metalization, solder
masking, and manufacturing steps associated with board fab and assembly.
Moisture reduces the glass-transition temperature (Tg) so that excess thermal
stresses can cause damage and increases the dielectric constant, leading
to a reduction in circuit switching speeds and an increase in propagation
delay times. Moisture ingress can also facilitate ionic corrosion
leading to both open and short circuits. Additionally, moisture
that accumulates at the interfaces of the resin and fiberglass can cause
interfacial degradation resulting in conductive filament formation (CFF).
In this study, laminates with resin systems of FR-4, HTFR-4, PI,
CE, and BT (TABLE II) were obtained from two suppliers (A & B).
All of the laminates were woven E-glass fabric in either 0.038 or 0.053
cm approximate thickness. Tests were conducted to evaluate and model
the equilibrium moisture content of a laminate under various isothermal
conditions. The tests covered relative humidities of 40, 50, 60,
70, and 85% and temperatures of 40, 50, 60, 70, and 85oC.
Two coupons of each laminate type were placed in isothermal environments
at a constant relative humidity until equilibrium was reached, at which
time the laminate mass was measured and returned to the environmental chamber.
In order to measure the moisture content of a PWB, the Rockwell Science
Center provided a capacitance monitoring method to allow for in-situ evaluation
of a laminate’s moisture content.
Work Accomplished
-
The electronic industry’s most common laminates (FR-4, HTFR-4, PI, CE,
and BT) were investigated for their equilibrium levels of moisture and
diffusion rates and monitored for their moisture content as a function
of electrical capacitance. Isothermal sorption tests were performed
on the laminates to find the equilibrium moisture content in each laminate
at various environmental conditions. A model was developed to assess
the equilibrium content as a function of both temperature and humidity.
-
Each of the laminates were characterized for their diffusivities, through
sorption-time experiments. Data from the experiments was used to
evaluate a diffusion coefficient. The Fickian model was used to facilitate
the evaluation.
-
The moisture content of each laminate was evaluated as a function of electrical
capacitance via a capacitance monitoring plate attached to each laminate.
Since the data indicated that the two parameters were linearly proportional,
a factor was calculated to convert the change in electrical capacitance
to moisture content.
-
The results of the moisture content calibration were subsequently applied
to a PWB with a capacitor plate on each of its laminates. The capacitance
method of moisture measurement in PWB shows slower moisture ingress than
the theoretical moisture diffusion based on laminate experiments.
This is due to impermeable barrier to moisture ingress in the metal copper
foil of the capacitor plate which is about 80% of the plate area.
It can be concluded from this that the presence of copper plane areas in
real circuit boards forms barriers to moisture movement and therefore,
circuit boards with large copper planes may require longer bake-outs than
standard boards.