Transcript Nanopore Multilayer Isotope Batteries

Nanopore/Multilayer Isotope Batteries
Using Radioisotopes from Nuclear
Waste
N. Luo[1], B. Ulmen[2] and George H. Miley[3] (Speaker)
University of Illinois at Urbana-Champaign, Urbana, IL 61801
[1] Visiting Research Assistant Professor, Department of Nuclear, Plasma and Radiological
Engineering, 104 S. Wright, AIAA Member.
[2] Graduate Student, Department of Nuclear, Plasma and Radiological Engineering, 104 S.
Wright.
[3] Professor, Department of Nuclear, Plasma and Radiological Engineering, 104 S. Wright,
Associate AIAA Fellow.
Outline
•
•
•
•
•
•
•
•
Introduction
Space applications
Operational model
Design philosophy
Planar Geometry
Porous Geometry
New Multi p-n Junction designs
Future directions
Types of nuclear batteries
P-N Junction
Betavoltaic
Thermoelectric
Conversion of thermal energy of
an isotope using the Seebeck
effect
Utilize direct conversion of
beta emission in a PN
junction
Miley, G. H., Direct Conversion of Nuclear Radiation Energy,
American Nuclear Society, La Grange Park, 1970.
Beta source
Solid-State Si Converter P
(P-N junction)
N
Vacuum collector
RL
Work like a capacitor with a
vacuum dielectric where alpha or
beta particles charge the plates
http://saturn.jpl.nasa.gov/spacecraft/safety/
P-N junction betavoltaic battery
Basic attributes
•
•
•
•
•
Long-lived (with suitable isotope choice)
Completely solid-state
No maintenance
Low power but high power density
Direct nuclear-to-electric conversion
Advantages over thermoelectric
•
•
No plutonium
Potential for higher efficiency
Advantages over vacuum collector
•
•
No high voltages
Higher energy density
Space applications
•
•
•
•
•
Low power but extremely long lived electrical power
source
Long duration missions where solar power is not an
option (outer planetary missions, landers, etc.)
Independent power source that could have its output
current tailored to the application
On-chip nanopower source
Trickle charge for Li-ion
Operation model
•
•
Similar principle as a solar cell
Uses nuclear beta decay to produce electron-hole pairs in
semiconductor P-N junction.
Ni-63 N-type Depletion region
collected
not collected
some collected
Minority carrier
diffusion lengths
P-type
Ohmic contact
Design philosophy
Semiconductor selection
• Isotope selection
• Depletion depth
• Maximum radioisotope thickness
•
Semiconductor - Silicon
Silicon
•Relatively high energy band gap
•High lattice damage threshold ~200-250 keV
•No environmental sensitivity
•Widely used, relatively low cost
•Well known microfabrication techniques
GaAs, Ge, and SiC also have attractive properties.
More difficult manufacturing problems
Choosing the radioisotope
•Long lifetime
•High specific power
•Below radiation damage threshold of Silicon (200-250 keV)
•Beta emitters – keV of energy, low momentum transfer
•No alpha emission – MeV of energy, high momentum transfer
•Minimum gamma emission – requires more shielding, reduced total device
energy density
Element
Half-Life (yr)
Specific Power (W/g)
Average Beta
Energy (keV)
Mode of Decay
12.3
0.0009
5.7
Beta
204
81Tl
3.78
0.6700
255
Beta
63
28Ni
100.1
0.0060
17.43
Beta
147
61Pm
2.62
0.3450
224.14
Beta, Gamma (few)
29.1
0.9210
200
Beta, Gamma
(when decays to 39Y90)
1H
3
38Sr
90
Reference planar /nano-pore design uses Ni 63
New multi-layer design is an extension of this work to
allow the use of higher energy beta emitters such as Sr
90 obtained from nuclear waste
We discuss the basic Ni 63 design first.
Determining depletion depth
The starting point for depletion depth is found through the range of
the beta particle in silicon.
Kanaya-Okayama Range Equation
Range of a 17.6 keV beta in Si:
R (um) = 0.0276 x 28.1 x 17.61.67 = 3.99
140.89 x 2.33
um
Self-absorption of beta particles
There is a maximum thickness where additional
radioisotope produces negligible energy flux from the
fuel surface.
µ = self absorption coefficient = 1.60 cm2/mg
x = thickness (mg/cm2)
A∞ = experimentally determined constant
Self-absorption dominates for thicknesses greater than ~2.2 µm.1,2.
Typical thickness: a few hundred nanometers
1. Schweitzer, G. K.; Stein, B. R.; Nehls, J. W. J. Phys. Chem. 1952, 56, 692–693.
2. Sims, G. H. E.; Juhnke, D. G. Int. J. Appl. Radiat. Isot. 1967, 18, 727.
Conversion Efficiency of Thermal Power Density to
Device Power
•
Reported Efficiencies – <6%
– Duggirala et al.: 1.8%
– Chandrashekar et al: 6%
•
Assuming 5% conversion efficiency for all
Geometry
Power Density
(µW/cm3)
Energy Density
100,000 hr (W·hr/L)
Energy Density
50 yr (W·hr/L)
Near-term Planar
0.75
70
280
Optimized Planar
2.1
190
780
Near-term Porous
9.8
900
3620
Conventional chemical batteries & fuel cells yield 200-300 W·hr/L
R. Duggirala et al. 2006 IEEE International Solid-State Circuits Conference. Abstract No. 1-4244-0079-1/06.
M. V. S. Chandrashekar et al. “Demonstration of a 4H Si betavoltaic cell.” Applied Physics Letters 2006, 88, 0330506.
Preparation of planar geometry device
P<100> Si, res.
160-240 Ω·cm
N-type spinon dopant
Cr/Au
SiO2
Al
Ti/Ni
A thermal oxide film is grown
Oxide film removed from one side
Exposed Si is coated with N-type
spin-on dopant
Dopant is thermally diffused into
the Si wafer
Unreduced glass and oxide layers
are removed
Al is sputtered and annealed on the
backside
Cr/Au is sputtered over the Al, Ti/Ni
sputtered on P-N junction side
Electrodepositing
Electrical contact pads area = 1 cm2
area = 0.09 cm2
The wafer is cut into several cells.
Each cell is 12 x 17 mm.
Area for Ni-63 electrodeposition is
1 cm2.
Preliminary Results – Planar Battery
VOC = 0.84 mV
Is = -11 nA
Pmax = 2.5 pW @ 0.4 mV
Conversion Efficiency: 0.0005%
Electron Beam Induced Current
EBIC is a method to test our PN junction response to beta emission
without needing to physically deposit nuclear material.
Physical Setup
Circuit Schematic
SEM e-beam irradiates P-N
junction contact
P
N
e- beam
current
A
Internal SEM ground
External
picoammeter
A
Direction
of e- flow
Internal SEM
ground
Faraday cup used to
calibrate the beam
current
JEOL 6060LV SEM @ Frederick
Seitz Materials Research Laboratory
– U of I
EBIC results
Responsivity is amps of
device current per watt
incident beam power
Dead layer loss results from
metallization seed layers for
electroplating
Dead layer loss
Porous design
With a planar battery, half the particles are going the “wrong way”.
A porous design allows a higher surface area per device volume
increasing energy and power densities.
Ni-63 solid metallic beta source
P-N junction
Ohmic contacts
Aspect ratio = 0
Assume 100% conversion efficiency
50% betas travel into silicon, 50% lost
Aspect ratio = 50
Assume 100% conversion efficiency
90% betas travel into silicon, 10% lost
Example of porous battery
Tritium battery using porous silicon (University of
Rochester, Rochester, NY)
– Increased energy conversion efficiencies of 3D vs. 2D
diodes.
η2D = 0.023% η3D = 0.22%
η3D / η2D = 9.9 ± 2.2
3H
W. Sun et al. “A Three-Dimensional Porous Silicon p-n Diode for Betavoltaics and Photovoltaics.” Advanced Materials 2005, 17,
1230-1233.
Fabrication of Porous Silicon
8 µm thick layer plasma-resistant photoresist spun on
Photoresist is patterned into a series of trenches
ICP-DRIE is used to anisotropically etch down
Wafer is cleaned of photoresist and
fluorinated polymer, a by-product of DRIE
~2584 trenches per 1 cm2 area
Fabrication of Porous Battery
Electroless deposition of non-radioactive Ni seed layer on Si for
electrodeposition of Ni-63.
Aqueous Solution 3.
1 M NiSO4
2.5 M NH4F
0.7 mM Sodium dodecyl sulfate
pH 5.6
3. C. Xu et al. Journal of The Electrochemical Society 2007, 154, D170-D174.
Recent Studies of New Multi-p-n-junction
designs
•
•
•
Objectives
Increase efficiency
Allow use of other isotopes such as Sr
Multi-p-n-junction Options, (1) with pores
-
beta
emitter
} multilayers
Planarized
Al w/ reflow
+
•
•
Ni
n-Si
p-Si
Al
Benefits: potentially high efficiency
Drawbacks: difficult to materialize; many steps in silicon
processing; relatively low yields.
Multi-p-n-junction Options, (2) crystalline silicon, without pores
beta
emitter
} multilayers
Simple Al
+
Ni
n-Si
p-Si
Al
• Benefits: high efficiency with simpler implementation vs pored type.
• Drawbacks: crystal silicon processing; planarization, temperature
compatibility (Al, Si, etc.)
Multi-p-n-junction Options, (3) amorphous silicon without pores.
-
beta
emitter
Ni top
cover
n a-Si:H
i a-Si:H
p a-Si:H
+
Al
• Benefits: simplest implementation, minimal steps in silicon processing,
radiation resistant.
• Drawbacks: may not get the highest efficiency vs crystalline Si.
A Prototype Multi-p-n-junction Betavoltaic Device
implemented with the type (3) geometry
• Implemented with discharge enhanced CVD of a-Si:H
• Stack of 5 p-n junctions
• The 2 pictures are the same device from different angles
A Prototype Multi-p-n-junction Betavoltaic Device:
Results
Output Voltage-Current of 5-P-N-Junction Beta Cell
0.9
0.8
Voltage (V)
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0
0.5
1
1.5
2
2.5
3
Current (nA)
• Note the open circuit voltage is > 0.7V, the theoretical
maximal for a single silicon p-n junction.
A Prototype Multi-p-n-junction Betavoltaic Device: Results (cont’d)
• Beta source: Ni-63
• Total activity: 0.00032 Ci
• Maximum power: 1.15nW
• Maximum conversion efficiency: 3.57%.
• Verified the scaling of power versus the number of pn junction layers
Summary
•
P-N junction nuclear batteries produce low power but have a long
lifetime
•
There is a potential for their use in several important space
applications requiring low power, long lifetime, rugged power sources
•
Designs have ben developed for both low and high energy beta emitter
sources.
The multi-layer designs are intended to open the door to use of
isotopes separated from nuclear wastes, giving a plentiful inexpensive
supply
•
Assumptions for optimized planar device:
75% isotopic enrichment – 42.5 Ci/g – 39 mW/cm3
28 mCi/cm2 deposited – 2150 nm
Wafer thickness 350 µm, area is 1 cm2
Available thermal power = 1.48 µW
Thermal power density (BoL) = 42.3 µW/cm3
5% efficiency thermal-to-electric conversion = 2.11 µW/cm3
100,000 hr mission = 190 W.hr/L
50 yr mission = 780 W.hr/L
Work in Progress
•
Porous battery operation
•
Parametric study of the effect of depletion
region depth on voltage and current
•
Develop rigorous model of device operation
Acknowledgements
This work was made possible by the DOE-NEER program.
For more information, please
contact:
George H. Miley
Department of Nuclear, Plasma &
Radiological Engineering
[email protected]
217-333-3772
Calculation of energy densities
Assumptions for near-term achievable (physical properties):
27% isotopic enrichment 56.72 Ci/g * 0.27 = 14.78 Ci/g
17.4 keV * 1.60E-16 J/keV *3.7E10 s-1.Ci-1 * 14.78 Ci/g * 8.9 g/cm3 = 13.6 mW/cm3
5 mCi/cm2 deposited
5E-3 Ci/cm2 /(14.78 Ci/g * 8.9 g/cm3) = 380 nm
Available thermal power = 13.6 mW/cm3 * 380E-7 cm * 1 cm2 = 0.52 µW
Wafer thickness 350 µm, area is 1 cm2 – device volume is 3.5E-2 cm3
Thermal power density (BoL) = 0.52 µW / 3.5E-2 cm3 = 14.9 µW/cm3
5% efficiency thermal-to-electric conversion = 0.75 µW/cm3
Energy density =
tEoL
0
P0 e-λt dt
100,000 hr
3.6E8 s
Energy density =
0.75 µW/cm3 * exp(- ln(2) * t / 3.2E9 s) dt = 250 J/cm3
0
250 J/cm3 * (1 W.hr / 3600 J)(1 cm3/.001 L) = 70 W.hr/L
If mission life is 50 yrs: 1000 J/cm3 = 280 W.hr/L
Assumptions for near-term porous device:
27% isotopic enrichment – 14.78 Ci/g – 13.6 mW/cm3
10 mCi/cm2 deposited – 760 nm
Wafer thickness 350 µm, area is 1 cm2
Porous increase in surface area per volume = 6.6x
Available thermal power = 6.8 µW
Thermal power density (BoL) = 194 µW/cm3
5% efficiency thermal-to-electric conversion = 9.8 µW/cm3
100,000 hr mission = 900 W.hr/L
50 yr mission = 3620 W.hr/L