rating and derating for low-voltage multilayer ceramic ... · low-voltage multilayer ceramic...
TRANSCRIPT
Rating and Derating for Low-Voltage Multilayer Ceramic
Capacitors (MLCCs)Alexander Teverovsky
Dell Services Federal Government, Inc.work performed for Code 562, GSFC
NASA Electronic Parts and Packaging (NEPP) Program
1To be published on nepp.nasa.gov.
List of Acronyms
2To be published on nepp.nasa.gov.
MLCC Multilayer ceramic capacitorBME Base metal electrodePME Precious metal electrodeDWV Dielectric withstanding voltage IR Insulation resistanceEMR Electromechanical resonance VBR Breakdown voltageCCS Constant current stressSTD Standard deviationHAST Highly accelerated stress testingHALT Highly accelerated life testingVR Rated voltage
Rated Voltages and Derating for MLCCs
3
Scope: Class II MLCCs rated to voltages ≤ 100V.
Outline: Breakdown voltages in BME and PME. Effect of voltage on capacitance. How performance of MLCCs is affected by voltage?
Dielectric withstanding voltage (DWV) Insulation resistance (IR) Electromechanical resonance (EMR) Voltage conditioning and life test.
How the parts are rated? Derating requirements. Where derating does not work? Conclusion
To be published on nepp.nasa.gov.
1
10
100
0 25 50 75 100 125
VBR
/VR
rated voltage, V
Ceramic and Tantalum Capacitors
Ta wet
1
10
100
0 25 50 75 100 125
VBR
/VR
rated voltage, V
Ceramic and Tantalum Capacitors
Ta wet
MLCC
Breakdown Voltage in Capacitors
4
VBR seems to be the most natural limit to VR:VR = VBR - margin
How much margin is necessary?
VBR does not limit VR for low-voltage MLCCs. Margin decreases as VR increases. There appears no difference in VBR for PME and BME.
0
20
40
60
80
100
120
140
160
0 10 20 30 40 50 60
E_B
R, V
/um
thickness, um
X7R MLCCs
PME
BME
To be published on nepp.nasa.gov.
Constant Current Stress Testing
5
CCS is a useful tool to test for VBR and quasi-static capacitance:
Non-linearity of V-t characteristics is due to C(V) and I(V) variations.
At V > VR capacitance drops substantially.
0
100
200
300
0 10 20 30 40
volta
ge, V
time, sec
1uF 50V at 5uA
Mfr.M 0805 SN3Mfr.A 1812 SN2Mfr.M 2220 SN1Mfr.C 1210 SN1Mfr.A 2220 SN1
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
0 25 50 75 100 125 150
capa
cita
nce,
uF
voltage, V
1uF 50V at 5uA
Mfr.M 0805 SN3Mfr.A 1812 SN2Mfr.M 2220 SN1Mfr.C 1210 SN1Mfr.A 2220 SN1
( ) ( )[ ] dttIIC
tVt
Lch∫ ×−×=0
1dtdVIVC ch=)(
0
50
100
150
200
250
300
350
400
0 5 10 15 20 25
volta
ge, V
time, sec
1uF 50V Mfr.M 0805
3 uA10 uA30 uA
To be published on nepp.nasa.gov.
Effect of Voltage on Capacitance
6
0.45
0.6
0.75
0.9
1.05
0 10 20 30 40
C(V
)/C(0
)
voltage, VPME Mfr.V 1825 0.47uF 50V PME Mfr.C 1825 0.47uF 50VBME Mfr.C 1825 0.47uF 50V BME Mfr.A 1825 0.47uF 50VBMEMfr.T 1210 10uF 25V PME Mfr.P 1825 1uF 50VBME 1206 22uF 6.3V
0.6
0.75
0.9
1.05
0 10 20 30 40
C(V
)/C(0
)
voltage, V
Mfr.T 1210 0.1uF 50VMfr.A 0805 0.1uF 25VMfr.M 1206 10uF 16VMfr.C 0402 0.01uF 6.3V
Decrease of C with voltage is greater for high volumetric efficiency parts.
Do we need limiting the voltage effect on C similar to MIL-PRF-55681 or leave the decision to the designer?
0
4
8
12
16
20
24
1.E+02 1.E+03 1.E+04 1.E+05
capa
cita
nce,
uF
freqiency, Hz
1206 Mfr.M 22uF 6.3V
0V 2V 4V 6V 9V 12V 20V 30V
Standard measurements of capacitance at 1 kHz.
To be published on nepp.nasa.gov.
Effectiveness of DWV Test
7
Gr.6 2225 Mfr.C 1uF 50V
breakdown voltage, V
cum
ulativ
e pr
obab
ility,
%
60 2000100 10001
5
10
50
90
99
virgin
TS
X-sect
MF
DWV test requires 2.5VR.Only ~20% of parts with gross
defects failed DWV test.19 out of 30 (63%) lots of parts
damaged by X-sect and TS had the probability of DWV test failure of less than 1%.
Some Mfr. testing parts at 5VR, JAXA - 30VR?
VBR is sensitive to the presence of defects and reflects quality of the lot.
The effectiveness of the existing DWV testing is low. Guidelines: - (5% level)/2 Average VBR and STD should be provided by manufacturers.
( )σ×−×= 25.0 avrcr VBRVBR
To be published on nepp.nasa.gov.
Effect of Voltage on IR
8
1uF 50V MLCCs Mfr.M
current, A
cum
ulativ
e pr
obab
ility,
%
1.E-10 1.E-71.E-9 1.E-81
5
10
50
99
22201206
0805
1000 sec60 sec
n = 0.76n = 0.71
n = 0.61
1.E-12
1.E-11
1.E-10
1.E-09
1.E-08
1.E-07
1.E+0 1.E+1 1.E+2 1.E+3 1.E+4
curr
ent,
A
time, sec
Mfr.A 0.12uF 50VMfr.C 1uF 6.3VMfr.N 0.12uF 16VMfr.C CDR31 0.018uF 50V
IR requirements for BMEs are relaxed vs.MIL parts. Does it mean they are worse?
Does voltage affect IR?According to spec, IR does not depend
H and V, but depends on the value of C:
However, ρ is not constant, and due to Schottky conduction IR should decrease exponentially with V/H.
HSC 0εε
=SHIR ρ
=C
IR 0ρεε=
ntItI −×= 0)(
Currents in MLCCs follow Curie von Schweidler law:
Will an increase in voltage reduce IR?What currents are measured during
first 120 sec of electrification?To be published on nepp.nasa.gov.
Leakage Currents in MLCCs
9
R = 5.6E8
R = 1.5E10
R = 3.7E8R = 2.4E8
0.0E+00
5.0E-08
1.0E-07
1.5E-07
0 20 40 60 80
curre
nt at
120
sec
, A
voltage, V
4.7uF 25V0.47uF 25V10uF 16V22uF 6.3V
n = 0.7
1.E-10
1.E-09
1.E-08
1.E-07
1.E-06
1.E+0 1.E+1 1.E+2 1.E+3 1.E+4
curre
nt, A
time, sec
virgin
crack
),,(),(),(),(),,( RHVTIVTIVtIVtIVTtI dlilabsch +++=
Absorption currents prevail up to several hours of electrification.
Iabs increases linearly with voltage, hence IR does not depend on how accurately VR is determined.
IR in high-C MLCCs is not sensitive to the presence of defects.
n = 0.83
1.E-11
1.E-10
1.E-09
1.E-08
1.E-07
1.E-06
1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06
curre
nt, A
time, s
Ct=1, rt=1e8Ct=2, rt=1e9Ct=3, rt=1e10Ct=4, rt=1e11sum
n = 2.02
1.E-11
1.E-10
1.E-09
1.E-08
1.E-07
1.E-06
1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06
curre
nt, A
time, s
Ct=1, rt=1e8Ct=0.33, rt=1e9Ct=0.1, rt=1e10Ct=0.033, rt=1e11sum
Capacitor with absorption
∑
−++=
i iiild
tr
VRV
RVtI τexp)( 000
To be published on nepp.nasa.gov.
Effect of Voltage on EMR
10
y = 4.09E+06x
0.E+00
1.E+06
2.E+06
3.E+06
4.E+06
5.E+06
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1
frequ
ency
, Hz
1/2W, 1/mm
Resonance frequencies in X7R MLCCs with different size
Inverse piezoelectric effect results in acoustic waives in polarized capacitors.
Several resonance frequencies. The amplitude increases with V (=>VR).EMR results in excessive noise and might
cause mechanical damage and failure.History case: a stacked capacitor failed in
PS systems. The major fres depends on the width,
W, of the capacitor: fres = v/(2W), v is the sound velocity.
For X7R materials v ~4.1×103 m/sec. Guidelines: warning about EMR effect
and impedance spectroscopy if necessary.
0.001
0.01
0.1
1
1.E+05 1.E+06 1.E+07
ES
R, O
hm
frequency, Hz
1uF 50V MLCC
Gr.5 0V
0.001
0.01
0.1
1
1.E+05 1.E+06 1.E+07
ES
R, O
hm
frequency, Hz
Mfr.C 1812 1uF 50V
Gr.5 0VGr.5 40V
To be published on nepp.nasa.gov.
Effect of Voltage on Life Test
11
1.E+02
1.E+03
1.E+04
1.E+05
1.E+06
1.E+07
1.E+08
2 4 6 8 10
time,
yea
rs
voltage acceleration constant, n
Time at 50C and 0.5VR that is equivalent to 96 hr of testing at 125C 2VR
0.7eV 0.9eV1.1eV 1.3eV1.5eV 1.7eV1.9eV
−×
==
211
2
222
111 11exp),(),(
TTkE
VV
TVTVAF a
n
ττ
Large VBR allows for extremely high voltage acceleration.
IR degradation and failures due to VO++
are observed mostly during HAST.
Voltage conditioning at 125C, 2VR for 100 hr corresponds to thousands of years of operation at 50C and 0.5VR.
Shiota (Murata, CARTS’13): - acceleration at HV might result in errors of extrapolation to operating conditions;- a stress by T acceleration instead of V.
Derating is an effective means for reducing risk of failures caused by VO
++.To be published on nepp.nasa.gov.
Early HALT Failures
12
1.E-06
1.E-05
1.E-04
1.E-03
0.001 0.01 0.1 1 10 100
curre
nt, A
time, hr
BME 1uF 50V at 182C 200V
Mfr.A 2220 SN1 Mfr.A 2220 SN2Mfr.A 2220 SN3 Mfr.C 2225 SN1Mfr.C 2225 SN2 Mfr.C 2225 SN3
1.E-07
1.E-06
1.E-05
1.E-04
0.001 0.01 0.1 1 10 100
curre
nt, A
time, hr
1825 BME fract. Mfr.A 1uF 50V at 125C 100V
fr. SN1 fr. SN2fr. SN3 fr. SN4contr
Cracks might appear as IR degradation
Cracks might cause early HALT failures
1.E-05
1.E-04
1.E-03
0 100 200 300 400 500 600
curr
ent,
A
time, hr
Mfr.M 1210 10uF 25V 125C at 50V
Are parts from Mfr.A better than from C? At E=1 eV, n=3 “failures” at 30hr 182C
200V correspond to >47,000 years at 25V and 55C.
Early failures might be more important than wear-out failures.
Early HALT failures and IR degradation might be due to cracks.To be published on nepp.nasa.gov.
How Voltage for Capacitors is Rated?
13
In ferroelectric materials the dielectric constant decreases with electric field.
For BaTiO3 type materials ε(E)function depends on the composition, structure of the dielectric, and is sensitive to the process conditions, and probably H => proprietary information.
Based on the acceptable level of the ε(E) variations, Emax can be selected.
For a given thickness of the dielectric, H, the rated voltage is VR = H × Emax – margin (?)
VR in low-voltage MLCCs is controlled by polarization processes in the dielectric, and is not related to breakdown voltages.
VR is a technical parameter chosen so that voltage dependent characteristics and reliability remain within the specified limits.
To be published on nepp.nasa.gov.
14
Guidelines for Derating Requirements
Part TypeVoltage
derating factor 1/
Maximum operating
temperature
Ripple current derating factor
2/
MIL 0.6 110 °C NA
BME 0.5 100 °C 0.75Notes:1/ The derating factor applies to the sum of peak AC ripple and DC polarizing voltage.2/ Ripple currents in power applications shall be derated to 75% of the manufacturers’ recommendations. The frequency of ripple current should be outside the electromechanical resonance frequency for the part.
Derating = stress reduction => increase of reliability and mitigation of unforeseen events (e.g. EMR).
Derating is based on history of applications and consensus. Danger of over-derating: larger parts weight more and might
be more susceptible to fracturing.To be published on nepp.nasa.gov.
15
Where Derating Does not Work?
Failures due to VO++ are observed mostly during HAST and can
be mitigated by derating. Failures due to the presence of cracks observed during box-
level testing and operation at low voltages. Bad news: derating does not help for low-voltage failures. Good news: BMEs are less susceptible to low-voltage failures.
BME, WLT at 0.8V, 23 min
To be published on nepp.nasa.gov.
16
Conclusion
Rating of capacitors is a prerogative of manufacturers. Manufacturers should assure and demonstrate reliability of
their product at rated conditions. (For Hi-Rel parts the requirements and test conditions are set by users.)
Manufacturers should also provide acceleration factors and methods used for their calculation. (Standardization of HAST?)
Derating is an efficient means for users to reduce risks of IR degradation and failures caused by Vo
++ ; it might also work for cracking related failures.
A 50% voltage derating seems reasonable. An open issue: should we limit a decrease of
capacitance at rated voltages. (Derating might be sufficient to take care of the problem if manufacturers specify C at VR.)
To be published on nepp.nasa.gov.