fahmi_dry_cyc_impr.eps


Acta Polytechnica Vol. 52 No. 2/2012

FMEA and FTA Analyses of the Adhesive Joining Process using
Electrically Conductive Adhesives

E. Povolotskaya, P. Mach

Abstract

This paper introduces a formulation of appropriate risk estimation methods that can be used for improving of processes
in the electronics area. Two risk assessment methods have been chosen with regard to the specifics of adhesive joining
based on electrically conductive adhesives. The paper provides a combination of a failure mode and effect analysis
(FMEA) and fault tree analysis (FTA) for optimizing of the joining process. Typical features and failures of the process
are identified. Critical operations are found and actions for avoiding failures in these actions are proposed. A fault tree
has been applied to the process in order to get more precise information about the steps and operations in the process,
and the relations between these operations. The fault tree identifies potential failures of the process. Then the effects of
the failures have been estimated by the failure mode and effect analysis method. All major differences between failure
mode and effect analysis and fault tree analysis are defined and there is a discussion about how to use the two techniques
complement each other and achieve more efficient results.

Keywords: failure mode and effect analysis, fault tree analysis, adhesive joining, electrically conductive adhesives.

1 Introduction
Electrically conductive adhesives (ECAs) are becom-
ing increasingly important in the electronics indus-
try. These materials are used in two main areas of
electronics packaging – in mounting of heat sensitive
components such as LCDs, and in mounting ultra-
fine pitch electronic packages [1].
ECAs create a permanent electrical and mechan-

ical connection between the pad and the component
lead. Adhesives on an epoxy basis filled with sil-
ver conductive particles aremainly used. The curing
temperature is lower than the soldering temperature
of lead-free solders. The electrical conductivity de-
pends on the concentration of conductive particles
in the resin, on the shape and the material of these
particles [2].
Anappropriate surfacepretreatmentof the joined

parts and detailed control of the filler through an
analysis of the grains arenecessary to achieve of good
electrical,mechanicalandthermalpropertiesof adhe-
sive joints [3]. To achieve parameters adhesive joints
that are comparable with soldered joints it is neces-
sary to optimize the process of adhesive joining.
The electrical resistivity of adhesives, electrical

noise and the nonlinearity of the current vs. volt-
age characteristic are higher than these parameters
for lead-free solders. The mechanical properties and
climate resistivity of ECAs are also worse than these
of solders [4].
There are many parameters that influence the

quality of adhesive joints in the process of adhesive

joining. Optimization of this process requires the use
of proper quality control tools such as failure mode
and effect analysis (FMEA) and fault tree analysis
(FTA). These analyses make an examination of the
process critical parameters possible.
FMEA is a technique for analyzing the occur-

rence of process failures and their effect on the re-
sult of a process [7]. Currently, FMEA is a widely
used method for risk assessment in industrial pro-
cesses. This method is primarily adapted for ma-
terial and equipment failures. There are four basic
types of FMEA: process FMEA, system FMEA, de-
signFMEAand serviceFMEA.FMEAproduces risk
priority numbers (RPNs) as outcomes. RPNs are ob-
tained for each failure that can occur in a process.
AnRPN is amultiplication of the severity, occur-

rence and detection numbers for each failure. Sever-
ity of failure shows the level of seriousness of the fail-
ure. Theoccurrencenumber representshowoften the
failure occurs and the detection number indicates the
levelofvisibilityof the failure. ThepurposeofFMEA
is to reduce theRPNbyreducingone, twoorall three
numbers in order to improve the process and ensure
the non-appearance of such errors subsequently. The
appearance of the failure can be reduced by improv-
ing the technical documentation requirements in the
process to eliminate the causes of failures or reduce
their frequency. Detection canbe reducedby offering
new or improved assessment methods or by offering
additional equipment fordetection. Several examples
ofFMEA implementation for industrial processes are
presented in [12–15].

48



Acta Polytechnica Vol. 52 No. 2/2012

A fault treeforms the basis for logical-probabilis-
ticmodels of systemfailure causality, failures of its el-
ements and other events or impacts [8]. Thismethod
is based on sequences and combinations of distur-
bances and faults [9]. Thus it is amultilevel structure
or diagramof causal relationships [10]. FTAprovides
a common vision of the process, components, and
howthese componentsare related. Thismakes it easy
to identify the defects arising in the process. It also
provides a way for proposing step-by-step improve-
ments to prevent defects and errors and for making
a troubleproof process. Several examples of a suc-
cessful combination of fault tree analysis and failure
mode and effect analysis methods for application in
industrial processes are shown in [10,16–19].
This paper presents the use of FTA and FMEA

for optimizing the joining process when electrically
conductive adhesives (ECA) are used.

2 Theoretical background

2.1 Basic risk analysis methods

Generally, risk is the possibility of the occurrence of
certain undesirable events that initiate various types
of failures. Risk analysis is used to find causes of fail-
ures and to prevent the occurrence of these failures
in the future. The results of risk analysis canbe used
for process optimization.
Risk analysis is divided into two complementary

types:
1. Qualitative.
2. Quantitative.
The task of qualitative analysis is to identify the

risk areas in a process, types of risks, and the factors
causing the risks. This is done in various ways, for
example, by an expert, by brainstorming and so on.
Quantitative analysis enables the level of effect to be
quantified for each type of risk.
Basic methods for risk analysis are as follows:
1. Analogies.
2. Expert methods.
3. Statistical methods.
4. Modeling, etc.
The analogy approach is focused on an examina-

tionofanalogiesamongdataobtained fromarangeof
sources. Expertmethods areused to collect the opin-
ions of qualified specialists. A statistical approach to
risk analysis uses various types of statisticalmethods
to process data that has been obtained experimen-
tally. The simulation is based on calculating various
types of models and on testing or these models in
various situations.
The followings are some of the most commonly

used risk analysis methodologies [5]:
1. Structured What-If Technique (SWIFT).
2. Fault tree analysis (FTA).

3. Event tree analysis (ETA).
4. Failure modes and effects analysis (FMEA).
Two expertmethods for risk analysis— fault tree

analysis (FTA) and failure mode and effect analysis
(FMEA)—were used for an analysis of a conductive
adhesive joining process [22].

2.2 Fault tree analysis

FTA is a very powerful systematic way which is
widely used for estimating process quality. Start-
ing from the top event, the fault-tree method uses
a Boolean algebra and logical modeling to make a
graphical representation of the relations among var-
ious failure events at different levels of the process
(Figure 1) [21].

Fig. 1: Typical fault tree

In this technique, deductive logic is used. It en-
ables the root causes of the failure events of a process
to be found. This type of logic helps to establish a
clear and detailed scheme of relationships between
steps or events in the process that can affect their
quality.
The contribution of the fault tree is as follows:
• Allows potential failure parts of the process to
be seen in detail.

• Helps identify failures deductively.
• Enables a qualitative or quantitative analysis of
the process to be made.

• Themethod can focus on individual parts of the
process, and can extract specific failures.

• It clearly represents the behavior of the process.
Themain advantage of the fault tree (in compari-

sonwithothermethods) is that the analysis is limited
to identifying only those events of the process which
lead to a specific process failure.
The disadvantages of fault trees are as follows:
• Implementation of the method requires consid-
erable inputs, becausemoreprocess details leads

49



Acta Polytechnica Vol. 52 No. 2/2012

to a geometric increase in the analyzed area, and
the number of influencing events grows corre-
spondingly.

• A fault tree is a Boolean logic diagram, which
shows only two states: working and failed.

• It is difficult to estimate the state of partial
failure of the process parts, because use of the
method generally indicates that the process is
either in good condition or in a faulty state.

• It requires a reliability specialist with deep
knowledge of the process.

2.3 Failure mode and effect analysis

TheFMEAmethod is applied in addition to theFTA
technique. The Failure Mode and Effects Analysis
(FMEA) is a widely used analytical tool. It is espe-
cially useful in connection with reliability, maintain-
ability and safety analyses.
Themain goals of the technique are to determine:

• Possible failures (defects) of the process, their
causes and consequences;

• The criticality of the effects on the process (S),
the probability of occurrence (defects) (O) and
their detectability (D).

• Generalized assessments of the functionality of
the process — calculation of RPN. Ten-point or
five-point rating scales are often used for occur-
rence, detection and severity numbers. A rule of
thumb is usually used for the risk priority num-
ber. This means that a serious look has to be
taken at RPNs higher than 125. When a ten-
point scale is used [3].
A special team is set up to conduct FMEA. The

values of S,D,O, andRPNaredeterminedby expert
estimates [22].
FMEAof the production process covers the stage

of technical preparation of equipment and materials
for the process tobe started. It endsbefore thedirect
work begins [22].

3 Experimental part

Before the risk analysis is started, it is necessary to
define the main steps in the process. A flow-chart of
the process of electrically conductive adhesive joining
is shown in Figure 2.
Failures of adhesive joints are mostly connected

with theirmechanical and/or electrical properties. A
table of failure resistance and nonlinearity of the cur-
rentvs. thevoltagecharacteristicof anadhesive joint
is shown in Table 1. The structure of the total resis-
tance of an adhesive joint is shown in Figure 3.
Here R1 represents the resistances between the

component lead and the adhesive, R2 represents the
resistance of adhesive, and R3 represents the resis-
tance between the pad and the adhesive.

Fig. 2: Flow-chart for a joining process based on ECA

Fig. 3: Total resistance of an adhesive joint

Table 1: Adhesive joining characteristics (the values are
valid for joining components of dimension type 1206)

Typical value Failure value

Resistance (R) 20mΩ ≥ 40mΩ
Nonlinearity(U) 10μV ≥ 25μV

Parts of the assembly processwhich influence the
valuesof these resistancesare examinedandanalyzed
for the risk of the occurrence of potential failures.
Deductive approach (FTA) and inductive ap-

proach (FMEA) are reviewed.
The first step in the process of an examination

of adhesive joining using a fault tree analysis is to
identify the main undesirable events. To define such
an event, it is necessary to define still acceptable
values for joint resistance, nonlinearity of the cur-
rent vs. voltage characteristic of the joint, shear
strength, tensile strength, etc. Typical and failure
values of electrical adhesive joining characteristics,
such as joint resistance and joint nonlinearity of the
current vs. voltage characteristic, are shown in Ta-
ble 1.
These events become the top-event of a fault tree.

50



Acta Polytechnica Vol. 52 No. 2/2012

Table 2: Influence of basic fault events on the properties of the adhesive

FAULTY ECA JOINT
Low

mechanical
resistivity

High
electrical
resistance

High
nonlinearity
of adhesive

High
noise

FTA code

Improper material of a lead × × FCU1
Improper material of a pad × FCU2
Improper surface finish of a lead × × × FCU3
Improper surface finish of a pad × × × × FCU4
Inappropriate curing × × × FCU5
Inappropriate storing × × FCU6
Inappropriate type of resin × × FCU7
Inappropriate concentration of filler particles × × × FCU8
Inappropriate viscosity × × × FCU9

Fig. 4: FTA for joints formed by ECA

The second step is to identify of events directly
related to the top-event. This is a repeatable process
and can be continued until we reach the basic events
that cause the top-event. Fault tree analysis (FTA) is
generally performed using a logical structure ofAND
andORgates. In the caseof the joiningprocessbased
on electrically conductive adhesive (ECA), each ba-
sic faulty event, alone, can cause one ormore failures
of an adhesive joint, so instead of grouping them un-
der gates, we used a tabular representation of FTA
(Table 2).
As a result of applying the FTA method to adhe-

sive joining we found the weakest parts of the pro-
cess. Toobtainabetterunderstandingof the failures,
we applied FTA to each type of typical joint failure.
Fault trees are presented in Figures 4, 5 and 6.
Varianceof the electrical resistance sometimesap-

pears in joint of this type. It can be caused by im-
proper surface finish of the pad and the component
lead by faulty placing of a component or by using an
adhesive with faulty consistency. Figure 7 shows a
more detailed representation of the joint.

When the FTA has been performed, an induc-
tive method such as failure mode and effect analysis
(FMEA) is applied to the joining process in order to
analyze the significance of various types of failures.
With the help of FMEA, a potential failure mode in
the process is analyzed to define the effect on the
result of the process and to classify each potential
failure mode according to severity. In the process
considered here, we used Process FMEA in an origi-
nal functional approach [20].
In this approach, each step in the process per-

forms a number of events which can be determined
as outputs. The outputs are listed and analyzed.

Fig. 5: FTA for joints formed by ECA

51



Acta Polytechnica Vol. 52 No. 2/2012

Fig. 6: FTA for joints formed by ECA

Fig. 7: FTA for joints formed by ECA

In our approach, we used the list of failures, i.e.
events which cause the top-event, defined during the
FTA analysis and for each undesirable event we de-
fine:
• Basic causes of failures,
• Specific features of the process,
• Severity numbers of failures(S): an assessmentof
the seriousness of the effects on the failure pro-
cess,

• Number of occurrences of failures (O): an as-
sessment of the likelihood that a failure will oc-
cur,

• Detection number of failures (D): assessment
of whether current control methods detect the
causes of failures on an appropriate level,

• Risk priority numbers (RPN): multiplication
of detection, occurrence and severity numbers.
This is used to set priorities for failures on pro-

52



Acta Polytechnica Vol. 52 No. 2/2012

Table 3: Part of the failure mode and effect analysis table for joints formed by ECA

F
M
E
A

F
M
E
A
P
ro
ce
ss
o
r

D
a
te
o
f
F
M
E
A

F
M
E
A
fo
r
jo
in
in
g
fo
rm
ed
by
el
ec
tr
ic
a
ll
y
co
n
d
u
ct
iv
e
a
d
h
es
iv
es

D
ep
a
rt
m
en
t
o
f
E
le
ct
ro
te
ch
n
o
lo
gy

S
ep
te
m
be
r,
2
0
1
1

O
bj
ec
t
o
f
F
M
E
A

P
ro
ce
ss
ed
A
re
a

F
M
E
A
-s
ta
tu
s

A
ss
em
bl
y
te
ch
n
o
lo
gy

A
d
h
es
iv
e
jo
in
in
g

R
u
n
n
in
g

P
o
te
n
ti
a
l

F
a
il
u
re
M
o
d
e
F
T
A
co
d
e
F
u
n
ct
io
n

P
o
te
n
ti
a
l

F
a
il
u
re

E
ff
ec
ts

S
P
o
te
n
ti
a
l

C
a
u
se
s

O
C
u
rr
en
t

C
o
n
tr
o
ls

D
R P N

A
ct
io
n
s

R
ec
om
m
en
d
ed

R
es
p
o
n
si
b
il
it
y

F
a
u
lt
y
E
C
A

jo
in
t

F
C
U
3

P
re
p
a
ra
ti
o
n
o
f

co
m
p
on
en
ts

su
rf
a
ce
s

H
ig
h

el
ec
tr
ic
al

re
si
st
a
n
ce

8
Im
p
ro
p
er

su
rf
a
ce
fi
n
is
h

o
f
a
le
a
d

6
V
is
u
a
l

co
n
tr
o
l

4
1
9
2
C
a
re
fu
l

cl
ea
n
in
g
o
f

su
rf
a
ce
le
a
d

F
C
U
2

P
re
p
a
ra
ti
o
n
o
f

co
m
p
on
en
ts

su
rf
a
ce
s

H
ig
h

el
ec
tr
ic
al

re
si
st
a
n
ce

8
Im
p
ro
p
er

su
rf
a
ce
fi
n
is
h

o
f
a
p
a
d

6
V
is
u
a
l

co
n
tr
o
l

4
1
9
2
C
a
re
fu
l

cl
ea
n
in
g
o
f

su
rf
a
ce
p
a
d

F
C
U
5

C
u
ri
n
g

L
ow
m
ec
h
a
n
ic
a
l

re
si
st
iv
it
y

8
In
a
p
p
ro
p
ri
a
te

cu
ri
n
g

3
M
ea
su
ri
n
g
o
f

a te
m
p
er
a
tu
re

p
ro
fi
le

3
7
2
C
a
re
fu
se
tt
in
g

u
p
o
f
cu
rr
in
g

p
ro
fi
le

F
C
U
6

S
to
ri
n
g
o
f

a
d
h
es
iv
e

L
ow
m
ec
h
a
n
ic
a
l

re
si
st
iv
it
y

8
In
a
p
p
ro
p
ri
a
te

st
o
ri
n
g

3
C
o
n
tr
o
l
o
f

st
o
ri
n
g

2
4
8
T
ra
ck
in
g
o
f

st
o
ri
n
g

A
p
p
ro
v
a
l
si
g
n
a
tu
re
s

C
o
n
cu
rr
in
g
si
g
n
a
tu
re
s

53



Acta Polytechnica Vol. 52 No. 2/2012

cess levels, and to establish what requires addi-
tional quality planning,

• Corrective actions,
• Checks on corrective actions.
Parts of the output of this analysis of the Faulty

ECA Joint failure mode are shown in Table 3.

4 Conclusion
The outcome of this approach is a reliability anal-
ysis realized through the interaction of the FMEA
andFTA reliability tools. Each of these risk analysis
methods has advantages, which enable the techno-
logical process to be investigated and help to observe
theprocessmore clearly fromdifferentpoints of view.
TheFMEAmethod in general is a libraryof all possi-
ble potential failures and their consequences,whereas
FTAenables adetailed analysis of logical and tempo-
ral relationships that lead to a failure, taken over the
top of the tree. The application of these twomethods
to the process, complementing each other, provides
deeper information than applying the methods sep-
arately. As a consequence of this approach, more
efficient results have been achieved. The most sig-
nificant steps in forming high-quality adhesive joints
are the preparation of the pad and the lead surfaces.
Preventive measures to avoid failures have been pro-
posed.

References

[1] John, H. L., Wong, C. P., Ning, C. L., Ricky
Lee, S.W.: Electronicsmanufacturing with lead-
free, halogen-free and conductive-adhesive ma-
terials. New-York : McGraw-Hill, Steve Chap-
man, 2003.

[2] Duraj, A.: Using of Electrically Conductive Ad-
hesives for Bonding in Electrical Engineering.
Prague : Czech technical university in Prague,
2001.

[3] Charles, A., Happer, M., Miller, M.: Electronic
packaging, microelectronics, and interconnec-
tion dictionary. New-York : McGraw-Hill, 1993.

[4] Ageeva, N. D., Vinakovskaya, Y. U., Li-
fanov,V.N.: Electrotechnical material engineer-
ing. Vladivostok : DVGTU, 2006.

[5] Vose, D.: Risk Analysis: A Quantitative Guide.
3rd edition. England : John Wiley and songs,
Ltd, 2008.

[6] Stamatis, D. H.: Failure mode and effect anal-
ysis: FMEA from theory to execution. 2nd ed.
p.cm. Visconsin : ASQ Quality Press, 2003.

[7] McDermott, R. E., Mikulak, R. J., Beaure-
gard, M. R.: The basics of FMEA. 2nd edition.
NewYork : Taylor&FrancisGroup, LLC, 2009.

[8] Bilal, M. A.: Risk analysis in engineering and
economics. Florida : CRC Press LLC, 2003.

[9] ICS 03.100.01, Czech technical standard ISO
31000:2009. Riskmanagement—Principles and
guidelines, 2010.

[10] Queiroz, S. R., Álvares, A. J.: FMEA and FTA
Analysis for Application of the Reliability Cen-
tered Maintenance Methodology: Case Study
on Hydraulic Turbines. Symposium Series in
Mechatronics, 3, 2008, p. 803–812.

[11] Tichy, M.: Risk control. Analysis and manage-
ment. Prague : C. H. Beck, 2006.

[12] Cassanelli, G., Mura, G., Cesaretti, F., Van-
zi, M., Fantini, F.: Reliability Predictions in
Electronic Industrial Applications. Microelec-
tronics Reliability, 45, 2005, p. 1321–1326.

[13] Douglas, W. B., Patrick, W. K., Carl, S. B.,
Michael, J.R.: ElectronicPrognostics—ACase
Study Using Global Positioning. Microelectron-
ics Reliability, 47, 2007, p. 1874–1881.

[14] Abdul-nour, G., Beaudoin, H., Ouellet, P., Ro-
chetie, R., Lambert, S.: A Reliability Based
Maintenance Policy; a Case Study. Comput-
ers and Industrial Engineering, 35 (3–4), 1998,
p. 591–594.

[15] Scipioni, A., Saccarola, G., Centazzo, A.,
Arena, F.: FMEA Methodology Design, Imple-
mentation and IntegrationwithHACCPSystem
in a Food Company. Food Control, 13, 2002,
p. 495–501.

[16] Han-Xiong, L., Ming, J. Z.: AHybridApproach
for Identification of Root Causes and Reliabil-
ity Improvement of a Die Bonding Process — a
Case Study. Reliability Engineering and System
Safety, 64, 1999, p. 43–48.

[17] Ortmeier, F., Reif, W., Schellhorn, G.: For-
mal Safety Analysis of a Radio-Based Railroad
Crossing Using Deductive Cause-Consequence
Analysis (DCCA). Report D-86135. Augsburg :
Lehrstuhl Für Softwaretechnik und Program-
miersprachen, Universität Augsburg, 2005.

[18] Gasanelli, G., Fantini, F., Serra, G., Sgatti, S.:
Reliability in Automotive Electronics; a Case
Study Applied to Diesel Engine Control.Micro-
electronics Reliability, 43, 2003, p. 1411–1416.

54



Acta Polytechnica Vol. 52 No. 2/2012

[19] Moser, T., Melik-Merkumians, M., Zoitl, A.:
Ontology-based Fault Diagnosis for Industrial
Control Applications. Report 11662353.Bilbao :
Vienna University of Technology, 2010.

[20] Vesely, W. E., Goldberg, F. F., Roberts, N. H.,
Haasl,D. F.: Fault TreeHandbook.Washington,
D.C. : Mail Stop SSOP, 1981.

[21] Ching-Torng,L.,Mao-Jiun, J.W.: HybridFault
TreeAnalysisUsingFuzzySets.Reliability Engi-
neering andSystemSafety,58, 1997,p. 205–213.

[22] Kane, M. M., Ivanov, B. V., Koreshkov, V. N.,
Skchirtladze,A.G.: Systems, methods and tools
of quality management. Saint Petersburg : Piter
Press, 2009.

Evgenia Povolotskaya
Pavel Mach
E-mail: povolevg@fel.cvut.cz
Czech Technical University in Prague

55