Evaluation of Case Size 0603 BME Ceramic Capacitors
Alexander TeverovskyAS&D, Inc.
NASA Electronic Parts and Packaging (NEPP) Program
1Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov presented by Alexander Teverovsky at the Components for Military and Space Electronics, Conference and Exhibition (CMSE), Los Angeles, CA, March 1-3, 2015.
https://ntrs.nasa.gov/search.jsp?R=20150002857 2020-05-30T16:11:08+00:00Z
List of Acronyms and Symbols
2Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov presented by Alexander Teverovsky at the Components for Military and Space Electronics, Conference and Exhibition (CMSE), Los Angeles, CA, March 1-3, 2015.
A area of cross-section PME precious metal electrode BME base metal electrode PWB printed wiring board
C capacitance RT room temperature CTE coefficient of thermal expansion STD standard deviation DCL direct current leakage t thickness of the dielectric DWV dielectric withstanding voltage telectr electrification time
E Young’s modulus Tg glass transition temperature Ea activation energy TSD terminal solder dip
HALT highly accelerated life testing Tsold melting temperature of solder HT high temperature VBR breakdown voltage HV high voltage VBR75 third quartile of VBR distribution IR insulation resistance VBRmin minimal VBR in the group LV low voltage VO
++ charged oxygen vacancy MLCC multilayer ceramic capacitor VR rated voltage
N number of layers XRF X-Ray Fluorescence
Abstract
3
High volumetric efficiency of commercial base metal electrode (BME) ceramiccapacitors allows for a substantial reduction of weight and sizes of the parts comparedto currently used military grade precious metal electrode (PME) capacitors. Insertionof BME capacitors in space applications requires a thorough analysis of theirperformance and reliability. In this work, six types of cases size 0603 BME capacitorsfrom three vendors have been evaluated. Three types of multilayer ceramiccapacitors (MLCCs) were designed for automotive industry and three types forgeneral purposes. Leakage currents in the capacitors have been measured in a widerange of voltages and temperatures, and measurements of breakdown voltages (VBR)have been used to assess the proportion and severity of defects in the parts. Theeffect of soldering-related thermal shock stresses was evaluated by analysis ofdistributions of VBR for parts in “as is” condition and after terminal solder dip testing at350°C. Highly Accelerated Life Testing (HALT) at different temperatures was used toassess the activation energy of degradation of leakage currents and predict behaviorof the parts at life test and normal operating conditions. To address issues related torework and manual soldering, capacitors were soldered onto different substrates atdifferent soldering conditions. The results show that contrary to a commonassumption that large-size capacitors are mostly vulnerable to soldering stresses,cracking in small size capacitors does happen unless special measures are takenduring assembly processes.
Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov presented by Alexander Teverovsky at the Components for Military and Space Electronics, Conference and Exhibition (CMSE), Los Angeles, CA, March 1-3, 2015.
Outline
4
Introduction. Leakage currents and insulation resistance.
Absorption and leakage currents. Temperature and voltage dependence of leakage currents.
Breakdown voltages. Initial evaluation. Effect of TSD at 350°C.
Degradation of leakage currents during HALT. Activation energy of degradation. Effect of HALT on breakdown voltages.
Effect of soldering. Manual and reflow soldering. Effect of substrate temperature for manual soldering.
Conclusion.Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov presented by Alexander Teverovsky at the Components for Military and Space Electronics, Conference and Exhibition (CMSE), Los Angeles, CA, March 1-3, 2015.
Introduction
5Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov presented by Alexander Teverovsky at the Components for Military and Space Electronics, Conference and Exhibition (CMSE), Los Angeles, CA, March 1-3, 2015.
Two major issues with MLCC: Insulation Resistance (IR) degradation related to oxygen vacancies. Failures related to soldering induced cracking.
Due to the difference in electro-chemical behavior of Ni and Ag/Pd and formed products, the probability of low-voltage failures for BME is less than for PME capacitors.
BME PME
Intrinsic wear-out failures caused by oxygen vacancies typically do not cause failures during applications.Concentration of VO
++ should be under control to reduce the probability of failures in the presence of defects.
n
i
dVBR
VBRTTFTTF
×= 0
V0++
Degradation in the presence of defects
Results of electro-chemical migration in PME and BME capacitors
The Significance of Breakdown Voltages
6Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov presented by Alexander Teverovsky at the Components for Military and Space Electronics, Conference and Exhibition (CMSE), Los Angeles, CA, March 1-3, 2015.
In most cases distributions of VBR for BME capacitors are bimodal. The high voltage (HV) mode has tight distributions (STD/Mean
~4%) indicating intrinsic breakdown. The presence of low voltage (LV)
subgroup is due to defects.
BME capacitors with bimodal distributions
breakdown voltage, V
cumu
lative
prob
abilit
y, %
200 2200600 1000 1400 18001
5
10
50
99
1812 1uF 50V
0805 0.1uF 25V
0805 0.12uF 50V
1210 0.1uF 50V12 um, floating electrode
15 um
13 um
14 um
The interception point of VBR distribution indicates the proportion of defects, and the spread of VBR towards low voltages indicates the significance of defects.
Lot acceptance criterion: VBRmin/VBR75 > 0.5. Migration of VO
++ in capacitors with defects results in IM failures.
VBR intrinsic
VBR defect
Part Types Used in This Study
7Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov presented by Alexander Teverovsky at the Components for Military and Space Electronics, Conference and Exhibition (CMSE), Los Angeles, CA, March 1-3, 2015.
X-Ray Fluorescence (XRF) analysis showed that barium titanate ceramics (BaTiO3) doped with different elements (mostly Zr, Y, W) is used in all parts.Same size and nominal auto and general purpose MLCCs
have similar design and materials.
application Mfr. C, µF VR, V t, µm N plates
Margins, µmEnd Cover Side
auto C 0.1 50 8 60 180 110 170
auto M 0.1 50 9 62 125 70 120
auto A 0.1 50 10 46 115 140 110
general C 0.1 50 8 120 150
general M 0.01 25 18 19 170 140 180
general C 0.01 50 15 17 160 220 160
Absorption and Leakage Currents
8Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov presented by Alexander Teverovsky at the Components for Military and Space Electronics, Conference and Exhibition (CMSE), Los Angeles, CA, March 1-3, 2015.
Direct Current Leakage (DCL) at room temperature (RT) decreases with the electrification time, as ~1/telectr.
Absorption currents at VR prevail during first hours of electrification and depend mostly on the value of capacitance.
Absorption currents are reproducible, increase with voltage, and stabilize at V > ~2VR.
The larger telectr. the better the sensitivity of the DCL/IR testing to the presence of defects.
No significant difference in DCL for generic and auto capacitors.
Relaxation of leakage currents at room temperature
y = 6E-08x-0.801
1.E-11
1.E-10
1.E-09
1.E-08
1.E-07
1.E+1 1.E+2 1.E+3 1.E+4
curre
nt, A
time, sec
BME 0603 0.1uF 50V at 22C, 50V
Mfr.C autoMfr.M autoMfr.A autoMfr.C gen
y = 1E-08x-0.569
1.E-11
1.E-10
1.E-09
1.E-08
1.E-07
1.E+1 1.E+2 1.E+3 1.E+4
curre
nt, A
time, sec
BME 0603 0.1uF 50V Mfr.M
50V0_50V100V0_100V150V0_150V y = 2E-08x-0.766
1.E-11
1.E-10
1.E-09
1.E-08
1.E-07
1.E+1 1.E+2 1.E+3 1.E+4
curre
nt, A
time, sec
BME 0603 0.1uF 50V Mfr.C
25V0_25V50V0_50V100V0_100V150V0_150V
Insulation Resistance
9Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov presented by Alexander Teverovsky at the Components for Military and Space Electronics, Conference and Exhibition (CMSE), Los Angeles, CA, March 1-3, 2015.
y = 6E-08x-0.695
y = 1E-07x-0.893
1.E-11
1.E-10
1.E-09
1.E-08
1.E-07
1.E-06
1.E+1 1.E+2 1.E+3 1.E+4
curre
nt, A
time, sec
0603 BME 0.1uF 50V
RT_Ca RT_Ma RT_A125C_Ca 125C_Ma 125C_AMIL_RT MIL_125 MUR_RTMUR_125
125C
22C
At 125°C intrinsic currents prevail after ~ 100 sec of electrification. IR measurements in LV capacitors during mass production is a
challenge. The test voltage should be increased. IR values for 0603 capacitors are within the range of values typical
for different types of BME MLCCs, but some are out of the MIL limit. MIL requirements for IR do not allow sufficient margin for high
volumetric efficiency BME capacitors. Murata auto limits are more reasonable.
1.E-10
1.E-09
1.E-08
1.E-07
1.E-06
1.E+1 1.E+2 1.E+3 1.E+4cu
rrent
, A
time, sec
Mfr.M auto 0603 0.1uF 50V at 85C
25V50V75V100V150V
depolarization
IR_avr = 5E9/C
1.E+07
1.E+08
1.E+09
1.E+10
1.E+11
1.E+12
1.E+13
0.001 0.01 0.1 1 10 100
IR, O
hm
capacitance, uF
BME IR_120 sec
all BMEBME 0603murata autoMIL limit
Effect of capacitance on IR in different types of BME capacitors
Relaxation of leakage currents at different temperatures and voltages
Effects of Temperature and Voltage on Leakage Currents
10Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov presented by Alexander Teverovsky at the Components for Military and Space Electronics, Conference and Exhibition (CMSE), Los Angeles, CA, March 1-3, 2015.
m=2.4m=1.9
m=1.6
m=1.5
1.E-08
1.E-07
1.E-06
1.E-05
1.E-04
3 30 300
curre
nt, A
voltage, V
Mfr.C auto 0.1uF 50V
85C_Ca
125C_Ca
150C_Ca
175C_Ca
m=2.6m=2.3
m=2.1
m=1.8
m=1.5
1.E-11
1.E-10
1.E-09
1.E-08
1.E-07
1.E-06
10 100
curre
nt, A
voltage, V
Mfr.M 0.01uF 25V
22C85C125C150C175C
m=3.2m=2.3
m=1.8
m=1.5
1.E-09
1.E-08
1.E-07
1.E-06
1.E-05
3 30 300
curre
nt, A
voltage, V
Mfr.A auto 0.1uF 50V
85C_Aa125C_Aa150C_Aa175C_Aa
Variations of leakage currents with voltage at different temperatures
1
1.5
2
2.5
3
3.5
75 100 125 150 175 200
m
temperature, deg.C
Variation of exponent m with temperature
Mfr.Ca 0.1uF 50VMfr.Ma 0.1uF 50VMfr.Aa 0.1uF 50VMfr.Mg 0.01uF 25VMfr.Cg 0.1uF 50VMfr.Cg 0.01uF 50V
At high temperatures I-V characteristics in all parts can be described with a power law, I ~Vm.The exponent m decreases with temperature.At 125°C 1.8 < m < 2.2 and at 175°C m ~ 1.5.DCL at HT can be explained based on the
space charge limiting model.
Effects of Temperature and Voltage on Leakage Currents, Cont’d
11Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov presented by Alexander Teverovsky at the Components for Military and Space Electronics, Conference and Exhibition (CMSE), Los Angeles, CA, March 1-3, 2015.
For BME capacitors activation energy of intrinsic leakage currents decreases with voltage.
Ea for BME is less than for PME capacitors. Voltage and temperature dependencies for
auto and general type capacitors are similar.
0
0.3
0.6
0.9
1.2
1.5
0 50 100 150 200 250
activ
atio
n en
ergy
, eV
voltage, VMfr.Ca 0.1uF 50V Mfr.Ma 0.1uF 50VMfr.Aa 0.1uF 50V M123 1uF 50VMfr.Mg 0.01uF 25V Mfr.Cg 0.1uF 50VMfr.Cg 0.01uF 50V M123 0.1uF 100V
PME
BME
1eV
0.61eV
0.55eV
0.47eV
0.42eV
1.E-09
1.E-08
1.E-07
1.E-06
1.E-05
1.E-04
0.002 0.0022 0.0024 0.0026 0.0028 0.003
curre
nt, A
1/T, 1/K
Mfr.A auto, 0603, 0.1uF, 50V25V50V100V150V200V
0.87eV0.73eV
0.6eV 0.55eV
0.52eV
1.E-09
1.E-08
1.E-07
1.E-06
1.E-05
1.E-04
0.002 0.0022 0.0024 0.0026 0.0028 0.003
curre
nt, A
1/T, 1/K
Mfr.M auto, 0603, 0.1uF, 50V
25V50V100V150V200V
0.9eV0.69eV
0.57eV
0.51eV
0.47eV
1.E-09
1.E-08
1.E-07
1.E-06
1.E-05
1.E-04
0.002 0.0022 0.0024 0.0026 0.0028 0.003
curre
nt, A
1/T, 1/K
Mfr.C auto, 0603, 0.1uF, 50V
25V50V100V150V200V
Variations of leakage currents with temperature at different voltages
Breakdown and Rated Voltages
12Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov presented by Alexander Teverovsky at the Components for Military and Space Electronics, Conference and Exhibition (CMSE), Los Angeles, CA, March 1-3, 2015.
Tested lots did not have gross defects: VBRmin/VBR75 > 0.5.Thickness of the dielectric is not the only factor determining VBR.No significant difference between automotive and general types of
BME capacitors.VR is 20 to 30 times less than VBR. One of the limiting factors for
VR is voltage dependence of capacitance.
Distributions of VBR for 0603 BME MLCCsCase size 0603 general type BME capacitors
breakdown voltage, V
cumu
lative
prob
abilit
y, %
500 2000800 1100 1400 17001.E-1
5.E-11
510
50
100
1.E-1
0.01uF, 50V
0.01uF, 25V
0.1uF, 50VMfr.C, 8um
Mfr.M, 18um
Mfr.C, 15um
BME 0603 0.1uF 50V capacitors
breakdown voltage, V
cumu
lative
prob
abilit
y, %
700 1200800 900 1000 11001
5
10
50
99
Mfr.C
Mfr.MMfr.A
autogeneral
0
0.02
0.04
0.06
0.08
0.1
0.12
0 10 20 30 40
capa
cita
nce,
uF
voltage, V
BME 0.1uF 50V capacitors
0603 Mfr.A, 10um0603 Mfr.C, 8um1210 Mfr.C, 23um1210 Mfr.A, 24um
C-V characteristics
HALT, General Type BME Capacitors
13Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov presented by Alexander Teverovsky at the Components for Military and Space Electronics, Conference and Exhibition (CMSE), Los Angeles, CA, March 1-3, 2015.
1.E-09
1.E-08
1.E-07
1.E-06
1.E+2 1.E+3 1.E+4 1.E+5 1.E+6
curre
nt, A
time, sec
BME MLCCs at 125C 200V
0.01uF 50V Mfr.C
0.01uF 25V Mfr.M
0.1uF 50V Mfr.C0.1uF 50V Mfr.C
0.01uF 25V Mfr.M
0.01uF 50V Mfr.C
1.E-11
1.E-10
1.E-09
1.E-08
1.E-07
1.E+2 1.E+3 1.E+4 1.E+5
curre
nt, A
time, sec
BME 0603 at 22C, 200V
1.E-08
1.E-07
1.E-06
1.E-05
1.E+2 1.E+3 1.E+4 1.E+5 1.E+6
curre
nt, A
time, sec
0603 BME MLCCs at 150C 200V
0.01uF 50V Mfr.C
0.01uF 25V Mfr.M
0.1uF 50V Mfr.C
1.E-08
1.E-07
1.E-06
1.E-05
1.E-04
1.0E+05 1.5E+05 2.0E+05 2.5E+05
curre
nt, A
time, sec
0603 BME MLCCs at 175C 200V
0.01uF 50V Mfr.C
0.01uF 25V Mfr.M
0.1uF 50V Mfr.C
At 125°C some degradation was observed in 0.1µF 50V capacitors only.At 150°C degradation in 0.1µF 50V capacitors
was noticeable and some parts failed.All 0.1µF 50V capacitors failed at 175°C, but
0.01 µF capacitors, both 25V and 50V, increased currents ~ 50% only.
Step stress HALT was carried out at T = 22°C, 125°C, 150°C, and 175°C, for 100hr and 200V at each step.
Leakage currents were monitored through the testing.
HALT, Automotive Grade BME Capacitors
14Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov presented by Alexander Teverovsky at the Components for Military and Space Electronics, Conference and Exhibition (CMSE), Los Angeles, CA, March 1-3, 2015.
1.E-07
1.E-06
1.E-05
1.E-2 1.E-1 1.E+0 1.E+1 1.E+2
curre
nt, A
time, hr
BME auto 0603 0.1uF 50V at 125C 200V
Mfr.A
Mfr.C
Mfr.M
1.E-07
1.E-06
1.E-05
1.E-2 1.E-1 1.E+0 1.E+1 1.E+2
curre
nt, A
time, hr
BME auto 0603 0.1uF 50V at 150C 200V
Mfr.A
Mfr.C
Mfr.M
No current increase for capacitors from Mfr.M that had minimal IR. It is possible that large intrinsic currents mask degradation.Some degradation can be observed at 125°C for Mfr.A and C.Degradation and failures in parts from Mfr.C at 150°C are similar to
generic capacitors from the same manufacturer.Comparison of results for Mfr.C 0.1µF and 0.01µF capacitors
indicates the effect of dielectric thickness on degradation processes.
Variations of leakage currents in 0.1µF 50V size 0603 BME auto capacitors at 200V with time at 125°C and 150°C
Activation Energy of Degradation
15Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov presented by Alexander Teverovsky at the Components for Military and Space Electronics, Conference and Exhibition (CMSE), Los Angeles, CA, March 1-3, 2015.
y = 1E-13x + 2E-07
y = 3E-12x + 6E-07
y = 2E-11x + 2E-06
1.E-07
1.E-06
1.E-05
1.E+3 1.E+4 1.E+5 1.E+6
curre
nt, A
time, sec
0.1uF 50V BME_C at 200V SN5
125C
150C
175C
Ea=1.6eV
1.E-14
1.E-13
1.E-12
1.E-11
1.E-10
0.0022 0.0023 0.0024 0.0025 0.0026
rate
, A/s
ec
1/T, 1/K
0.1uF 50V BME_C at 200V
SN5
SN6
SN7
Activation energy of degradation is consistent with the migration of positively charged oxygen vacancies model.Degradation at life test and operating conditions is negligibly small.
At Ea ~1.6 eV, the predicted degradation rate at 125°C and 200V is ~ 1.3×10-13 A/sec. Using a conservative estimation for the voltage acceleration constant, n = 3, the rate at 2VR would be ~ 1.5×10-14 A/sec. At this rate it would take ~2×106 years for the current to increase by 1 µA during life testing.
Time dependence of DCL approximated with linear functions.
Temperature dependence of degradation rates for 3 samples.
Effect of HALT on VBR
16Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov presented by Alexander Teverovsky at the Components for Military and Space Electronics, Conference and Exhibition (CMSE), Los Angeles, CA, March 1-3, 2015.
Parts with higher rate of degradation had a more substantial decrease of VBR.Post-HALT degradation of VBR is consistent with the defect-related
failure model:An IM failure occurs when accumulation of VO
++ at a defect site would be sufficient to increase current density to a level necessary to initiate a local thermal run-away process.
V0++
defect
Effect of HALT on general type BME 0603 MLCC
breakdown voltage, V
cumu
lative
prob
abilit
y, %
0 2000400 800 1200 16001
5
10
50
99
Mfr.C0.01uF 50V
Mfr.M0.01uF 25V
initial
after HALT
Comparison of initial and post_HALT distributions of VBR.Effect of HALT on BME auto 0603 MLCC
breakdown voltage, Vcu
mulat
ive pr
obab
ility,
%400 1400600 800 1000 12001
5
10
50
99
Mfr.A
Mfr.M
initial
post HALT
Effect of Soldering Thermal Shock on Breakdown Voltages
17Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov presented by Alexander Teverovsky at the Components for Military and Space Electronics, Conference and Exhibition (CMSE), Los Angeles, CA, March 1-3, 2015.
BME MLCC 0603 0.1uF 50V Mfr.A
breakdown voltage, V
cumu
lative
prob
abilit
y, %
700 1200800 900 1000 11001
5
10
50
99
Mfr.CMfr.A
initial
initialafter TSD at 350C
after TSD at 350C 600
800
1000
1200
1400
1600
1800
600 800 1000 1200 1400 1600 1800
VB
R_T
SD
350,
V
VBR_init, V
BME 0603 capacitors
Mfr.Aa 0.1uF 50VMfr.Ma 0.1uF 50VMfr.Ca 0.1uF 50VMfr.Cg 0.01uF 50VMfr.Mg 0.01uF 25VMfr.Cg 0.1uF 50V
30 capacitors of each type were stressed by the terminal solder dip testing at a solder pot temperature of 350°C (10 cycles).
The effect was evaluated by visual examinations and by comparing initial and post-terminal solder dip (TSD)350 distributions.
No substantial variations in distributions of breakdown voltages after thermal shock testing.TSD_350 does not generate cracks on 0603 BME capacitors.
Effect of Manual Soldering
18Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov presented by Alexander Teverovsky at the Components for Military and Space Electronics, Conference and Exhibition (CMSE), Los Angeles, CA, March 1-3, 2015.
It is often assumed that large size MLCCs (1210 and above) are more susceptible to cracking.
Insertion of BME MLCCs in hi-rel applications means extensive use of small size capacitors ( below 0805). Are they less vulnerable to cracking?
Experiment: Groups with 14 to 12 samples of 0.1µF, 50V, size 0603 BME MLCCs were soldered manually onto FR4 PWBs using recommended precautions and a soldering iron set to 315°C.
MLCC Open circuit Short circuit Intermittent Total failures
Mfr. C 3/14 2/14 0/14 5/14Mfr. M 1/14 2/14 5/14 8/14Mfr. A 3/12 0/12 2/12 5/12
Manual soldering can cause failures of small size capacitors. Out of 40 samples of 0603 size MLCCs soldered with a soldering
iron at 315°C 70% were electrical failures.
Post-manual-soldering test results
Manual vs. Reflow Soldering
19Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov presented by Alexander Teverovsky at the Components for Military and Space Electronics, Conference and Exhibition (CMSE), Los Angeles, CA, March 1-3, 2015.
BME 0603 0.01uF 25V capacitors Mfr.M
breakdown voltage, V
cumu
lative
prob
abilit
y, %
100 1600400 700 1000 13001
5
10
50
99
initialFR4 PCB soldered by paste reflowFR4 PCB at 150C, manual sold., 0.08" tip at 350CFR4 PCB at 22C, manual sold., 0.03" tip at 350CFR4 PCB at 22C, manual sold., 0.08" tip at 350C
manual soldering, cold FR4 PWB
Effects of manual and reflow soldering were compared by measurements of VBR for case size 0603 BME 0.01uF, 25V capacitors from Mfr.Msoldered onto the same FR4 PWB.
Manual soldering was carried out with different tip sizes (0.03” and 0.08”) and different temperatures of the board (cold, 22°C, and hot, 150°C).
No defects were observed in MLCCs after solder reflow and after manual soldering onto a board preheated to 150°C.Soldering iron tips of larger size increase the probability of failures
from ~20% for 0.03” to ~70% for 0.08”.
0.01uF 25V capacitors failed after manual soldering
crack
Post Soldering Stresses
20Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov presented by Alexander Teverovsky at the Components for Military and Space Electronics, Conference and Exhibition (CMSE), Los Angeles, CA, March 1-3, 2015.
( ) ( )( )
11
−−
×+
−×−=
cap
PWBPWBcap
rgcapPWBreflow
AAEE
TTαασ
( )
( )1
1−
−
×+
−×=
cap
PWBPWBcap
rsoldcapcoldPWB
AAEE
TTασ
case size
reflow FR4
cold PWB FR4
reflow alumina
cold alumina
0603 32.6 -107 -45.8 -1812225 17.3 -56.7 -37.8 -149
Board
MLCCSolder
Compressive stress
CTEPWB > CTECAP
Tensile strength of X7R ceramics is 70 to 250 MPa. PWB with coefficient of thermal expansion (CTE)PWB > CTEcap
create compressive stresses (σ > 0) in MLCCs after reflow soldering.
Ceramic substrates with CTEPWB < CTEcap create tensile stresses (σ < 0) that are much more dangerous.
The level of stresses caused by solder reflow and manual soldering onto a cold board can be estimated using a one-dimensional model:
Material E, GPa CTE, ppm/oC Tg/Tsold
PWB (FR4) 17 15 150Alumina 360 7.7 230MLCC 100 10
Mechanical characteristics used for stress calculations
Mechanical stresses in MLCCs assembled by reflow and manual soldering (3mm FR4, 1mm alumina)
Reflow soldering onto a PWB creates relatively small compressive stresses in case of FR4 and tensile stresses in case of alumina board.Manual soldering onto cold PWB creates significant tensile stresses.Post manual soldering stresses are greater for small size MLCCs.
Effect of Substrate and Type of MLCC
21Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov presented by Alexander Teverovsky at the Components for Military and Space Electronics, Conference and Exhibition (CMSE), Los Angeles, CA, March 1-3, 2015.
Manual soldering onto alumina board resulted in less damage compared to FR4 PWB. A contradiction with the model is likely due to a much higher thermal conductivity of alumina (increase T of the board).All case size 0603 0.01 µF 50V capacitors passed DWV test after
reflow soldering. However, contrary to 0.1 µF 50V MLCCs, ~10% of capacitors had defects.
BME 0603 0.01uF 25V capacitors Mfr.M
breakdown voltage, V
cumu
lative
prob
abilit
y, %
100 1600400 700 1000 13001
5
10
50
99
initial
ceram PCB at 22C,
FR4 PCB at 22C, manual sold.,
manual sold.,
0.008" tip at 350C
0.08" tip at 350C
0.01uF 25V capacitors were soldered onto alumina and FR4 PWBs manually (350°C)
Mfr.C BME 0603 0.01uF 50V. Effect of soldering.
breakdown voltage, V
cumu
lative
prob
abilit
y, %
0 2000400 800 1200 16001
5
10
50
99
in oil
in still air
after reflow soldering loose MLCCs
loose MLCCs
0.01uF 50V capacitors were soldered onto an FR4 PWB by solder reflow
Failures Caused by Manual Soldering
22Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov presented by Alexander Teverovsky at the Components for Military and Space Electronics, Conference and Exhibition (CMSE), Los Angeles, CA, March 1-3, 2015.
cracks
Typical soldering thermal shock cracks in large MLCCs
Contrary to the annular thermal stress cracks in large size capacitors, cracks in 0603 capacitors occur along the terminations.Termination cracks are not specific for BME
MLCCs only. They were also observed in small-size PME capacitors.
Size 2225
PME 0402
Size 0603
Post soldering cracks in small size MLCCs
Conclusion
Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov presented by Alexander Teverovsky at the Components for Military and Space Electronics, Conference and Exhibition (CMSE), Los Angeles, CA, March 1-3, 2015. 23
Design, materials, leakage currents and breakdown voltages in automotive and general application capacitors are similar.Leakage currents: at RT absorption currents prevail over intrinsic conduction currents that are several
orders of magnitude less than currents measured within 2 min of electrification; Ea decreases with voltage from ~ 0.9 eV at 0.5VR to ~ 0.5 eV at 4VR. I-V characteristics at HT follow a power low with the exponent decreasing from ~3 at
85°C to ~ 1.5 at 175°C.Degradation of DCL and failures were observed during HALT in
parts having minimal thickness of the dielectric. Activation energy of degradation is large, ~1.6 eV, so no intrinsic wear-out failures are expected at life testing or normal operating conditions.0603 MLCCs are vulnerable to manual soldering and more than
50% of capacitors can fail in case of soldering onto a cold PWB.Board preheating is critical to reduce the probability of failures.Post-soldering fracturing along the terminals is a specific feature
of small-size capacitors.