Upload
buiphuc
View
217
Download
1
Embed Size (px)
Citation preview
Integration of Power-Supply Capacitors with Ultrahigh Density on Silicon Using Particulate Electrodes
POWERSOC 2012
P M Raj, Himani Sharma, Kanika Sethi, Prof. Rao Tummala
3D Systems Packaging Research Center
Georgia Institute of Technology, Atlanta, GA 30332-0560
2nd GIT Workshop, November 14-16, 2012
Slide 1
Outline – Component integration using 3D IPAC Concept – 3D IPAC goals – Applications
o VRM and decoupling § Thinfilm decoupling capacitors
o Power convertor § High-density capacitors
2nd GIT Workshop, November 14-16, 2012
Slide 2
Passives Trend
Silicon and Glass IPDs
400-500 microns 500 microns (0402)
Discretes
Thick Devices 500-800 microns
Array of Discretes
500 microns (0402)
Thick Devices 500-800 microns 200-800 microns
Courtesy: Kemet, Fraunhofer IZM
3D IPAC
40 microns
40-100 microns
2nd GIT Workshop, November 14-16, 2012
Slide 3
Passives Passives
What is 3D IPAC Concept and Why? Active IC Active IC
40 µm
10 µm
1. Ultra-thin Glass Core ~ 30 µm
2. Low Loss
3. Fine-pitch and coarse-pitch through-package vias
4. Passive components on both sides of glass. • RF capacitors, inductors • Power supply capacitors and inductors • Decoupling Capacitors • Precision components
5. Active components for power, digital, RF and Analog
6. Interconnections
3D IPAC is a functional module
30 µm
2nd GIT Workshop, November 14-16, 2012
Slide 4
3D IPAC on Interposer
Passives Passives
Active IC Active IC
40 µm
10 µm
30 µm
3D IC
Si/Glass Core
3D IPAC
Thin 3D IPAC Capacitor
IC
10 µm
30 µm
3D IPAC
IC
2nd GIT Workshop, November 14-16, 2012
Slide 5
3D IPAC in the 3D POP
Passives Passives
Active IC Active IC
40 µm
10 µm
30 µm
2nd GIT Workshop, November 14-16, 2012
Slide 6
3D IPAC on PWB
Passives
Active IC Active IC
40 µm
10 µm
30 µm
PWB
2nd GIT Workshop, November 14-16, 2012
Slide 7
Goals of 3D IPAC § Miniaturization:
– 5X reduction in thickness
§ Cost: – 2—5X reduction in cost
§ Performance: – Higher bandwidth at lower power compared to today’s systems
§ Functionality: – Double-side component integration for heterogeneous functions in an
ultrathin module
2nd GIT Workshop, November 14-16, 2012
Slide 8
3D IPAC – Applications
RF MODULE Power Convertor Module
PDN Module
Glass
Active MOSFET Switches
High-density Capacitors
High-density Inductors
• RF ICs • High Q and Precision
Components: • Capacitors • Resistors • Inductors
• Voltage regulator module
• VRM output capacitors • Decoupling capacitors
• Power switches • High-density capacitor • High-density inductor
High-density Output Capacitors
Glass
High Frequency Decoupling
POWER IC
80 µm Glass
10 µm
Precision RF Capacitors
Precision Resistors
RF Inductors
By-pass capacitors
RF IC
2nd GIT Workshop, November 14-16, 2012
Slide 9
3D IPAC – PDN Module
POWER
GROUND FR4
Case 3
3D IPAC VRM
Glass core
Logic IC
30 µm
• Improved power integrity for 3D ICs and packages with thinner packages
• VRM and decoupling functions in the interposer, close to the package
• Reduces the number of decoupling capacitors
• Thin power ICs (ex. MOSFET switches)
• High-density inductors
• High-density capacitors
Glass
High Frequency Decoupling
POWER IC
2nd GIT Workshop, November 14-16, 2012
Slide 10
§ Thinfilms (100-300 nm): – Thinner films (Barium Strontium Titanate); – Faster crystallization; – Higher capacitance density of 14 nF/mm2
Key Result BST on Si/SiN/Ta/Ni
BST sputtering Ar/O2 (60% O2); 100 W; 2.5 hrs
Annealing conditions
750 C, 30 min in N2
Capacitance density 9 – 11 nF/mm2
Higher yield with large area electrodes 220 nF achieved on 0.28 cm2
Leakage current of the order µA/cm2 achieved up to 3 V (4 nF/mm2)
Yield 85% (30/35 devices yielded)
Sputtered BST Thinfilm Capacitors
2nd GIT Workshop, November 14-16, 2012
Slide 11
Solgel PZT –Thinfilm Capacitors
1.2 x 2.5mm2: 100-110 nF,
loss 0.1-0.3
Yield 51% (56/110)
yield 14% (8/57)
225 nF,
2.9 x 2.5mm2: loss 0.4-1.2
2.5 µF/cm2 (1 V) 1 µF/cm2 (10 V)
1.0E-‐8
1.0E-‐7
1.0E-‐6
1.0E-‐5
1.0E-‐4
1.0E-‐3
1.0E-‐2
1.0E-‐1
1.0E+0
0 5 10 15
Leakage Cu
rren
t (A/cm
2)
Volts (V)
PZT/LNO/Pt/Ta/Si
PZT/LNO/Si
PZT/LNO/ZrO2/Si
2nd GIT Workshop, November 14-16, 2012
Slide 12
3D IPAC-Based PDN • Three decoupling capacitor
integration scenarios were modeled
• SMD on board shows highest parasitics and lowest performance;
• Requires excessive on-chip decoupling
• ThinFilm integration in today’s packages can achieve higher frequency stability, but has limited manufacturability
• IPAC micro-assembled close to IC shows best performance because of thin interposer and short bump parasitics;
0.01$
0.1$
1$
10$
100$
1000$
0.00E+00$ 5.00E+08$ 1.00E+09$ 1.50E+09$ 2.00E+09$
Impedance))
(Ω))
Frequency)(Hz))
SMD
Integrated
Thin IPAC capacitor
5X improvement in frequency performance with the proposed 3D IPAC strategy
2nd GIT Workshop, November 14-16, 2012
Slide 13
3D IPAC – Power Converter Module Power Module:
• Switching regulator or charge-pump based • Thin power ICs • High-density inductors • High-density capacitors
Glass
Active MOSFET Switches
High-density Capacitors
High-density Inductors
2nd GIT Workshop, November 14-16, 2012
Slide 15
Area enhancement for 100 µm film
Pla
nar c
apac
itanc
e µ
F/cm
2
1.4
0.2
0.6
1.0
1.8
2.2
5 20 50 200 500
MLCC
2 µF/cm2
10 µF/cm2
40 µF/cm2
100 µF/cm2
Tantalum
PRC Focus
2nd GIT Workshop, November 14-16, 2012
Slide 16
Adhesion layer
Silicon Carrier
Rou3ng Layer
Bo7om Electrode
Dielectric
Top Electrode
Process flow for High-Density Capacitors
2nd GIT Workshop, November 14-16, 2012
Slide 17
Silicon Trench vs. Copper Par3culate Electrode
Silicon Compatible 100 X enhancement in area Expensive silicon micromachining tools
Silicon Compatible Potential for >2000 X enhancement in surface area from BET measurements ( 100 µm) Low-cost paste processing
Porous Electrode with Highest Surface Area Efficiency
2nd GIT Workshop, November 14-16, 2012
Slide 18
0
5
10
15
20
25
30
35
5 10 15 20 25 30 0
100
200
300
400
500
600
700
800
900 0.2 0.5 1 3 5 10
500 - 700 !!
ASPECT RATIO (TENCH)
PARTICLE SIZE (NANOELECTRODE)
AR
EA
EN
HA
NC
EM
EN
T
25 - 30
0
5
10
15
20
25
30
35
5 10 15 20 25 30 0
100
200
300
400
500
600
700
800
900 0.2 0.5 1 3 5 10
500 - 700
ASPECT RATIO (TENCH)
PARTICLE SIZE (NANOELECTRODE) A
RE
A E
NH
AN
CE
ME
NT (
TRE
NC
H)
AR
EA
EN
HA
NC
EM
EN
T (
NA
NO
ELE
CTR
OD
E)
25 - 30
Area Enhancement: Nanoelectrodes Vs Trenches
2nd GIT Workshop, November 14-16, 2012
Slide 19
PRC’s Background High-Density Capacitor Research
High surface area electrodes on silicon trench;
High surface area electrodes on organic substrate with etched foil
High surface area electrodes on organic substrate with sintered particles Adhe
sion layer
Adhesion layer
US 8,084,841 B2
US 8,174,017
2nd GIT Workshop, November 14-16, 2012
Slide 20
Silicon Bottom Cu
Sintered particles
Nanoelectrodes on Si
Step 1: § Copper powder § Dispersant (Phosphate ester) § Binder: Propylene carbonate Step 2: § Printing on Silicon Step 3 § Sintering – 400o C – 600o C
§ Copper particle size : 1-2 µm § High surface area enhancement
due to open pores § 30-40X surface area enhancement
for 25 µm film
2nd GIT Workshop, November 14-16, 2012
Slide 21
Porous electrode with Sacrificial Polymers • Using a sacrificial polymer (polypropylene carbonate) that would decompose easily at
sintering temperatures in reducing atmosphere
Porosity in place of polymer
HEAT
Polymer granules
Copper particles
Polypropylene carbonate Vola3le by-‐product
2nd GIT Workshop, November 14-16, 2012
Slide 22
ALD Conformality Studies
§ Increase diffusion time of the ALD vapor
§ For aspect ratio of 1000 – 1 torr: 100 seconds – 0.2 torr: 500 seconds – Larger than ALD pumpdown
time; – ALD time can start becoming
larger;
§ For aspect ratio of 10: – Micro to milli seconds; – Very small fraction of ALD
pumpdown time; – no affect on ALD process time;
1
100
10000 10 100 1000
10
0.001
0.1
0.01
0.0001
Aspect Ratio
Torr
X
s
econ
ds
Based on modeling by: Gordon, Haussman and Kim (Harvard); Shephard (IBM)
2nd GIT Workshop, November 14-16, 2012
Slide 23
Conformality Studies using EDS EDS Scan at the bo7om of 200 um thick electrode
0 20 40 60 80 1006
7
8
9
10
11
12
Ato
mic
Wei
ght %
Electrode Thickness (depth) (µm)
Al O
ALD conformality as a function of electrode depth
2nd GIT Workshop, November 14-16, 2012
Slide 24
Cross-section of sintered bottom electrode showing ALD alumina layer
105 nm
§ 700 cycles of ALD alumina (~ 100 nm)
§ FESEM showed 105 nm thick Alumina film on copper particle necks
Conformal Dielectric on Porous Electrodes
2nd GIT Workshop, November 14-16, 2012
Slide 25
Key Challenge: To access the high surface area by using a conformal top electrode Approach : Use of a liquid conducting polymer– polyethylene dioxythiophene (PEDT) Polymerization: EDT+ oxidizer+ inhibitor – 1:25 (in-situ)
§ Conductivity ~ 100 S/cm § Fluid nature with low viscosity for
easy infiltration of particulate electrode
§ Low temperature processability ( 50o C)
§ Self-healing characteristics
Conformal Top Electrode
2nd GIT Workshop, November 14-16, 2012
Slide 26
Demonstration of Nanoelectrode Capacitors
Device Thickness Area
Enhancement
Capacitance
Density
Planar (0.6mm x 0.6mm) -- 0.12µF/cm2
<5 µm 10X 1.1 µF/cm2
10 µm 25 X 3.9 µF/cm2
25 µm 150-200X 30 µF/cm2
>75 µm 400-500X 55-60 µF/cm2
2nd GIT Workshop, November 14-16, 2012
Slide 27
Particulate electrodes using PEDT show very high leakage as compared to planar capacitors. Two possible reasons identified: (I) PEDT/ Al2O3 interaction (II) Inherent Defects in Al2O3 film
Nanoelectrode Capacitors : Leakage Current
1.00E-‐09
1.00E-‐08
1.00E-‐07
1.00E-‐06
1.00E-‐05
1.00E-‐04
1.00E-‐03
1.00E-‐02
1.00E-‐01
1.00E+00
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
Leakage Cu
rren
t (A/
uF)
Voltage (V)
I-‐V plot of planar and par9culate electrode capacitor
Par3culate Capacitor with PEDT
Planar Capacitor with PEDT
Planar Capacitor with Au
2nd GIT Workshop, November 14-16, 2012
Slide 28
GT PRC: HDCAP Vs State-of-the-Art
28
Permittivity X Electrode Area Enhancement
Capacitance Volum
etric D
ensity µF/cm
2
1
100
100 1,000
Si Trench
PRC’s Si and Glass IPD
10
10,000
Thin Film Capacitors
Conformal dielectric on a nanoporous electrode
2nd GIT Workshop, November 14-16, 2012
Slide 29
Conclusions
Slide 29
• 3D IPAC shown as a compelling new technology for active and passive component integration in ultrathin functional modules
• Better miniaturization • Lower cost • Higher performance
• Novel silicon or glass-compatible high-density capacitor technologies investigated to meet 3D IPAC goals