Novel Conductive Adhesives for Surface Mount

Applications

DAOQIANG LU, C. P. WONG

Materials Science and Engineering and Packaging Research Center, Georgia Tech, Atlanta, Georgia 30332-0245

Received 15 December 1998; accepted 26 February 1999

ABSTRACT: Electrically conductive adhesives (ECAs) have been explored as a tin/lead

(Sn/Pb) solder alternative for attaching encapsulated surface mount components on
rigid and flexible printed circuits. However, limited practical use of conductive adhe-
sives in surface mount applications is found because of the limitations and concerns of
current commercial ECAs. One critical limitation is the significant increase of joint
resistance with Sn/Pb finished components under 85°C/85% relative humidity (RH)
aging. Conductive adhesives with stable joint resistance are especially desirable. In this
study, a novel conductive adhesive system that is based on epoxy resins has been
developed. Conductive adhesives from this system show very stable joint resistance
with Sn/Pb-finished components during 85°C/85% RH aging. One ECA selected from
this system has been tested here and compared with two popular commercial surface
mount conductive adhesives. ECA properties studied included cure profile, glass tran-
sition temperature (T g ), bulk resistivity, moisture absorption, die shear adhesion
strength, and shift of joint resistance with Sn/Pb metallization under 85°C/85% RH
aging. It was found that, compared to the commercial conductive adhesives, our in-
house conductive adhesive had higher T g , comparable bulk resistivity, lower moisture
absorption, comparable adhesion strength, and most importantly, much more stable
joint resistance. Therefore, this conductive adhesive system should have better perfor-
mance for surface mount applications than current commercial surface mount conduc-
tive adhesives. © 1999 John Wiley & Sons, Inc. J Appl Polym Sci 74: 399 – 406, 1999

Key words: electrically conductive adhesives; stable joint resistance; tin/lead-fin-

ished components; surface mount applications

INTRODUCTION printed circuit board and the surface mount com-

ponents. Conductive fillers in the adhesive are
The continued evolution of the electronics indus- responsible for the electrical interconnection and
try has created a need for environment and user- the resin mainly provides the mechanical inter-
friendly alternatives to tin/lead (Sn/Pb) solders. connection (Fig. 1).
For many years, special efforts were made to re- Besides the environmental issue, conductive
place lead-containing solder materials with con- adhesives as one of the alternatives to solder have
ductive adhesives in surface mount technology. the following potential advantages: (i) lower sen-
Conductive adhesives can simultaneously estab- sitivity to thermomechanical stresses, due to
lish mechanical and electrical joints between higher flexibility than solder; (ii) lower curing
temperatures enabling the use of heat sensitive or
nonsolderable materials (chip on glass or surface
Correspondence to: C. P. Wong (e-mail: cp.wong@mse. mount device on polyester flex substrates); (iii)
gatech.edu).
Journal of Applied Polymer Science, Vol. 74, 399 – 406 (1999)
high-resolution capability for fine-pitch intercon-
© 1999 John Wiley & Sons, Inc. CCC 0021-8995/99/020399-08 nects due to smaller particle size than solder
399
400 LU AND WONG

used in conductive adhesive formulations were

obtained from Degussa Corporation. An adhesion
promoter was purchased from Dow Corning. All
the chemicals were used as received. Eutectic
Sn/Pb wires (0.25 mm diameter) were provided by
Hisco Company. Two commercial conductive ad-
hesives, ECA-A and ECA-B, were provided by the
same manufacturer.

Cure Study
Figure 1 Conductance of conductive adhesives in a
joint. Curing profiles of all adhesives were studied by
using a differential scanning calorimeter (DSC)
from TA Instruments, model 2923. An adhesive
sample of about 10 mg was placed into an alumi-
pastes; and (iv) simple processing if compared to num hermetic DSC pan. All samples were studied
wave soldering (less process steps).1–5 by both dynamic cure and isothermal cure. In the
On the other hand, compared to traditional dynamic cure, the sample was heated in the DSC
soldering technology, concerns and limitations cell from 25 to 250°C at a heating rate of 5°C/min.
about conductive adhesive technology are also In the isothermal cure studies, the sample was
present. One major limitation is that joint resis- quickly placed into the DSC cell, which had been
tance of conductive adhesive with Sn/Pb finished preheated to a prescribed temperature and the
components increased dramatically under 85°C/ DSC data was collected thereafter.
85% relative humidity (RH) aging. Although
Sn/Pb finishing of the components is not neces- Measurements of Glass Transition
sary for using conductive adhesives, nevertheless, Temperatures (Tg)
in the short term, Sn/Pb compatibility is a desir-
able quality for compatibility of this technology T g s were measured with a thermomechanical an-
with the existing component supply infrastruc- alyzer (TMA) from TA Instruments, model 2940.
ture. Therefore, conductive adhesives with stable Preparation of TMA specimens was based on the
joint resistance are especially needed.3 Besides following procedures: (a) placed an adhesive sam-
stable joint resistance, a desirable conductive ad- ple in an aluminum pan (1.5 in. diameter); (b)
hesive also should have the following properties: cured the sample in a convection oven; (c) re-
higher impact resistance, low moisture pickup, moved the aluminum pan away after the cured
low bulk resistivity, and high adhesion strength.4 sample cooled down to room temperature; and (d)
In this study, a conductive adhesive system cut the specimen into squares with dimensions of
based on epoxy resins has been developed. Shift of about 6 3 6 3 1.5 mm. A macroexpansion probe
joint resistance with Sn/Pb metallization under was used here and the static force applied on the
85°C/85% RH aging, and other properties includ- probe was set to 0.050 Newton. Temperature was
ing curing profile, glass transition temperature, ramped from 25 to 250°C at a heating rate of
bulk resistivity, moisture absorption, and die 5°C/min. The dimension change with tempera-
shear adhesion strength of one electrically con- ture was recorded. T g s were obtained from the
ductive adhesives (ECA) selected from this sys- TMA curves.
tem have been tested and compared with those of
two popular commercial surface mount conduc- Measurements of Bulk Resistivity
tive adhesives.
Resistivity of an ECA was calculated from the
bulk resistance of the ECA specimen with specific
dimensions. Two strips of an adhesive tape were
EXPERIMENTAL
applied onto a precleaned glass slide with a gap of
0.1 in. (0.254 cm) between these two strips. A
Materials
conductive adhesive paste was then spread
Epoxy resins used in this study were purchased within the space by means of a doctor blade, and
from Shell Chemical Company. Hardeners were then the tapes were removed. After cure, the bulk
supplied by Aldrich Chemical Company. Ag flakes resistance of this ECA strip was measured by
 NOVEL CONDUCTIVE ADHESIVES 401

struments, model 552. The size of the die was 2

3 2 mm. The die and substrate were both passi-
vated with Silicon Nitride (Si3N4). Ten specimens
were tested for each ECA sample. The average
adhesion strength and standard deviation for
each ECA sample were reported. The adhesion
strength data was not included in the calculation
of the average adhesion strength if the die was
Figure 2 Layout of bulk resistance measurement broken or fractured during die shear adhesion
setup. test.

Joint Resistance Variation

using a Keithley 2000 multimeter with a four- Joint resistance variations were tested by using
point probe. The layout of the measurement setup an in-house test device which is depicted in Fig-
is shown in Figure 2. The length of the specimen ure 3.6 This device consists of metal wire seg-
l is length of the glass slide (75.60 cm); its thick- ments (1 cm long) that were separated by small
ness h is that of the adhesive tape (0.007 cm), and gaps (1 mm). Conductive adhesives were applied
its width w is that of the gap between the two to the gaps between the wire segments with 1 mL
adhesive strips (0.254 cm). Resistivity was calcu- syringe and connected the metal wires. After the
lated from bulk resistance R by using following specimen was cured, joint resistance was mea-
equation: sured from the two wire ends with a Keithley
2000 mutlimeter. Eutectic Sn/Pb wire was used
w3h here to simulate Sn/Pb finished components. The
Resistivity (ohm-cm) 5 R 3 test devices were aged under 85°C/85% RH in a
l
temperature and humidity chamber from Lunaire
Five specimens for each ECA sample were tested. Environmental, model no. CEO932W-4. Joint re-
An average bulk resistivity and a standard devi- sistance data were collected periodically during
ation for each sample were calculated and re- aging.
ported.

Measurements of Moisture Absorption RESULTS AND DISCUSSION

Moisture absorption of all the cured ECAs were Conductive Adhesive Formulations
tested using a dynamic vapor sorption system
from Surface Measurement System, model DVS Our in-house conductive adhesive formulations
1000. Dimensions of all the specimens used in were based on epoxy systems. The epoxy could be
this study were 150 mm thick, 1.5 cm wide, and 15 either Bisphenol-F or Bisphenol-A type epoxies. A
cm long. The dimensions of the specimens were typical ECA formulation includes an epoxy resin,
closely controlled by the following procedures: (a) a hardener, a catalyst, an adhesion promoter, and
placed two strips of an adhesive tape (150 mm Ag flake fillers. Filler loading of these two com-
thick) on a Teflon coated aluminum plate with a mercial ECAs were measured by dissolving resin
1.5 cm space between them; (b) applied an con- parts with an acetone solvent. It was found that
ductive adhesive paste on the plate between the
two adhesive strips with a doctor blade; (c) re-
moved the tape strips, and (d) cured the ECA
samples. Then the specimens were aged under a
85°C/85% RH condition in the DVS instrument
and their mass changes with time were recorded
continuously.

Die Shear Adhesion Strength

Die shear adhesion strength was measured at
25°C by using an adhesion tester from Royce In- Figure 3 In-house joint resistance test device.
402 LU AND WONG

Figure 4 DSC curves of ECAs during a dynamic cure.

filler loading of both commercial adhesives was 80 Measurements of Glass Transition

wt %. To ensure comparability, the filler loading Temperatures (Tg)
of in-house formulation ECA-C was also kept Adhesives ECA-A, ECA-B, and ECA-C were cured
80%. at 150°C for 40, 20, and 20 min, respectively,
based on our DSC isothermal cure study results.
Cure Study The cured samples were studied by a TMA and
One of the outstanding features of conductive ad- results are given in Figure 6. As can be seen from
hesives is that they can cure at much lower tem- Figure 6, the in-house formulation ECA-C had a
peratures than Sn/Pb solder reflow temperature. higher T g (105.67°C) than both commercial ECAs,
Therefore, ECAs are especially valuable when at- ECA-A (63.61°C) and ECA-B (72.15°C). High T g s
taching heat-sensitive components and low-cost are desirable for conductive adhesives.4 ECAs
substrates. Another desirable property is that with higher T g s generally are less susceptible to
ECAs can be cured rapidly, which can save time creep, which is one of the possible mechanisms for
and lower cost. the unstable joint resistance during 85°C/85% RH
The cure behaviors of the in-house formulation aging.4
(ECA-C) during both dynamic cure and isother-
Bulk Resistivity
mal cures were studied by using a DSC and com-
pared to those of ECA-A and ECA-B. The cure Current commercial surface mount conductive
profiles of all the samples during dynamic and adhesives generally have bulk resistivity in the
isothermal cures are given in Figures 4 and 5, range of 1024 ohm-cm, which is higher than that
respectively. As can be seen in Figure 4, the ECA- of Sn/Pb solder, 1025 ohm-cm, in surface mount
A shows a broader cure peak at a higher temper- applications.7 The high resistivity is one of the
ature, compared to ECA-C and ECA-B. The in- limitations of current commercial ECAs. The re-
house adhesive, ECA-C, had similar cure profile sistivity of the three ECAs is shown in Figure 7.
as ECA-B. ECA-A and ECA-B were both recom- The in-house ECA-C had a lower bulk resistivity
mended by the manufacturers to cure at 150°C, as than ECA-B but with a similar bulk resistivity to
such isothermal cure profiles at 150°C of these ECA-A.
adhesives were studied and compared with our
in-house ECA-C (Figure 5). At 150°C, ECA-C had Moisture Absorption
similar cure kinetic as ECA-B but it was cured Moisture absorption is one of very important pa-
much faster than ECA-A. rameters of conductive adhesives. Condensed wa-
 NOVEL CONDUCTIVE ADHESIVES 403

Figure 5 DSC curves of ECAs during an isothermal cure (at 150°C).

ter from moisture can degrade mechanical prop- Low moisture pickup is a particularly desirable
erties through depression of the T g , giving rise to property of a surface mount conductive adhesive.
swelling stresses in the system, and creating Moisture absorption data, both dynamic and
voids or promoting the catastrophic growth of kinetic, of these three adhesives are given in Fig-
voids already present in the system.7 Water also ures 8 and 9, respectively. As can be seen from the
might affect electrical properties of ECAs by in- figures, the in-house formulation ECA-C reached
ducing the formation of oxide layers on pad metal an equilibrium moisture absorption faster than
resulting from corrosion and oxidation reactions.7 the commercial ECAs and it showed much lower

Figure 6 TMA curves of the conductive adhesives.

404 LU AND WONG

Figure 7 Bulk resistivity of the conductive adhesives.

moisture absorption, 0.25%, than both commer- Joint Resistance Variation

cial materials (0.60 and 0.75%).
At present, compatibility of ECAs with Sn/Pb fin-
ished surface mount components is an important
Die Shear Adhesion Strength
property although Sn/Pb finishing eventually
Die shear adhesion strength (at 25°C) of all the might not be necessary. Therefore, ECAs that
ECAs before and after 500-h 85°C/85% RH aging have stable joint resistance with Sn/Pb finished
were tested, and results are given in Figure 10. As components and comparable other properties are
can be seen from the figure, all of the three ECAs particularly needed. A convenient joint resistance
showed comparable adhesion strengths both be- test device was specially designed for this study.
fore and after aging. Eutectic Sn/Pb wire was selected and used in the

Figure 8 Dynamic moisture absorption of conductive adhesives.

 NOVEL CONDUCTIVE ADHESIVES 405

Figure 9 Kinetic moisture absorption of the conductive adhesives.

test device to simulate Sn/Pb finished surface much better compatibility with Sn/Pb finished
mount components. Changes of joint resistance of components.
these conductive adhesives with Sn/Pb solder
wire during 85°C/85% RH aging are given in Fig-
ure 11. It was found that our in-house conductive CONCLUSIONS
adhesive showed no significant joint resistance
change but the two commercial ECAs showed dra- An epoxy-based electrically conductive adhesive
matic joint resistance increase after 500-h aging. system that has improved compatibility with
Therefore, the in-house conductive adhesive had Sn/Pb finished components was developed. Com-

Figure 10 Adhesion strength of conductive adhesives before and after aging.

406 LU AND WONG

Figure 11 Shift of joint resistance of the ECAs during 85°C/85% RH aging.

pared to the two commercial surface mount con- 3. Zwolinski, M.; Hickman, J.; Rubin, H.; Zaks, Y.
ductive adhesives, this conductive adhesive has Proceedings of the 2nd International Conference
more stable joint resistance with Sn/Pb finished on Adhesive Joining & Coating Technology in
components, much lower moisture absorption, Electronics Manufacturing, Stockholm, Sweden,
June 3–5, 1996, 333.
faster cure, comparable bulk resistivity, compara-
4. Jagt, J. C. IEEE Trans Components, Packaging,
ble adhesion strength, and higher glass transition
and Manufacturing Technology 1998, 21(2), 215.
temperature. 5. Alphert, B. T.; Schoenberg, A. J. Electronic Pack-
aging and Production 1991, 130.
6. Wong, C. P.; Lu, D.; Vona, S.; Tong, Q. Proceedings
REFERENCES of the 1997 1st IEEE International Symposium on
Polymeric Electronics Packaging 1997, 80.
1. Jagt, J. C.; Beris, P. J. M.; Lijten, G. F. C. M. IEEE 7. Smith-Vargo, L. Electronic Packaging and Produc-
Trans Components, Packaging, and Manufacturing tion 1986, 48.
Technology 1995, 18(2), 292. 8. Khoo, C. G. L.; Liu, J. Proceedings of the 2nd
2. Luchs, R. Proceedings of the 2nd International International Conference on Adhesive Joining
Conference on Adhesive Joining & Coating Tech- & Coating Technology in Electronics Manufac-
nology in Electronics Manufacturing, Stockholm, turing, Stockholm, Sweden, June 3–5, 1996,
Sweden, June 3–5, 1996, 76. 395.