PHYSICS AND MODELING OF THERMOELECTRIC MICROGENERATORS

1
Al 5-lea Seminar National de Nano, Academia
Româna, 2 februarie 2006
Thermoelectric Microgenerators with Nanometric
Films
1Gh.V.Cimpoca, 1I.Bancuta, 2Gh.Brezeanu, 3Ileana
Cernica, 3Maria Cimpoca, 4I.Grozescu
1Valahia University of Targoviste 2Politehnica
University of Bucharest 3National Institute for
RD Microtechnology, Bucharest 4National
Institute for RD Electrochemical Materials,
Timisoara
Al
Proiect finantat prin programul MATNANTECH,
contract 250/2004
2
INTRODUCTION
A thermoelectric in-plane
micro-generator with nanometric films has been
fabricated using compatible standard
semiconductor technologies (MEMS). The active
material is a nanolayer polycrystalline silicon
material laid on a dielectric membrane sustained
by a silicon frame. Hicks and Dresselhouse
predicted a huge increase of figure of merit ZT
if the dimensionality of the electron system in
thermoelectric materials is redused from 3D
behaviour in bulk materials to 2D behaviour via
nanoscaled layers. Reduced dimensionality offers
one strategy for increasing ZT relative to bulk
values 1-2. The use of
low-dimensional systems for thermoelectric
applications is of interest because low
dimensionality provides (1) - a method for
enhancing the density of states near EF, leading
to an enhancement of the Seebeck coefficient
(2) - opportunities to take advantage of the
anisotropic Fermi surfaces in multi-valley cubic
semiconductors (3) - opportunities to increase
the boundary scattering of phonons at the
barrier-well interfaces, without as large an
increase in electron scattering at the interface,
(4) - opportunities for increased carrier
mobilities at a given carrier concentration when
quantum confinement conditions are satisfied 3.
3
What makes a good Thermoelectric Material?

• Figure of Merit ZT (a2s/K)T
• T Absolute Temperature
• a2 (Seebeckcoefficient)2
• Tells how much average thermal energy is
  transported by each carrier
• s electrical conductivity
• Tells how much the carriers can transport energy
  without Joule loss
• K thermal conductivity
• Tells how small is the reverse flow of heat from
  the cold-side to the hot-side, opposing the
  electron-transport of heat
• Minimize thermal conductivity and maximize
  electrical conductivity
• Has been the biggest dilemma for the last 40
  years
• Can the conflicting requirements be met by
  nano-scale material design?

Efficiency versus ZT and DT
4
Big Jump in ZT with the Phonon-Blocking,
Electron-Transmitting Structures
With Advanced Semiconductor Materials?

• ZT need to improve over 1.3 at 300K for a major
  impact in electronics cooling and around 2.5 for
  a revolutionary impact in air conditioning, and
  power from waste - heat

5
Some of Approaches

• New Bulk Materials
• Skutterudites (Rensselaer, Oak Ridge, JPL, 1992)
• Cage structures with ratting atoms to scatter
  phonons
• Novel Chalcogenides and Clathrates (Michigan
  State and Arizona, 1994)
• Complex Variations of Bi2Te3 to reduce phonon
  mean-free paths
• Nano-scale Materials
• Low-Dimensional Structures (MIT, Mit Lincon Labs,
  1992)
• Quantum confinement to Enhance Density of
  states which increase Seebeck coefficient
• Nano-scale Superlattice (RTI, 1992)
• Phonon blocking from acoustic mismatch between
  superlattice components but electron-transmitting
  due to negligible electron-energy offsets
• Heterostructure Thermionics (UCSB, Oak Ridge,
  1996)
• Thermionic-like effects using energy barriers
  that can be controlled in hetero-structures

6
Some Bulk Material and Nano-Material Progress

• Cs Bi4Te6(Michigan State University)
• Bulk Materials with a ZT 0.8 at 225K but less
  than 0.8 at 300K (Science 287, 1024-1027, 2000)
• Filled Skuterrudites (JPL)
• Bulk materials with a ZT 1.35 at 900K(Proc. Of
  15th International Conf. On Thermoelectrics,
  1996)
• PbTe/PbTeSe Quantum-dots (Harman, MIT Lincoln
  Labs.)
• ZT 1.6 at 300K based on cooling data (Science
  297, Sep. 2002)
• Bi2Te3/Sb2Te3Superlattices(RTI)
• ZT2.4 at 300K in devices with all properties
  measured at the same place, same time, with
  current flowing and verified by two independent
  techniques (Nature, 597-602, Oct. 2001)

7
New Bulk Materials

• The skutterudite structure was originally
  attributed to a mineral from Skutterud (Norway)
  with a general formula (Fe, Co, Ni) As3. The
  skutterudite structure (cubic space group Im3,
  prototype CoAs3) is illustrated in figure. The
  unit cell contains 8 AB3 groups. The unit cell is
  relatively large and contains 32 atoms which
  indicates that a low lattice thermal
  conductivity might be possible. For the state of
  the art thermoelectric materials such as PbTe and
  Bi2Te3 alloys, the number of isostructural
  compounds is limited and the possibilities to
  optimize their properties for maximum performance
  in different temperature ranges of operation are
  also very limited.

The skutterudite unit cell of formula TPn3 (T-
transition metal, Pn - pnicogen).
Crystal type Chevrel
8
MICROGENERATORS WITH NANOMETRIC FILMS
Experimental techniques and sample preparation
9
Modes Of Working

• Two modes of working are anticipated for the
  in-plane thermoelectric micro-generator. The
  first mode of working (i.e. mRTG) is when the
  heat source is on the membrane (Fig. 2a). The
  silicon frame that sustains the membrane is the
  cold side. A large temperature difference along
  the thermoelectric legs should be created with
  small heat sources because the thickness of the
  area covered by the thermoelectric leg is thin
  (1250 nm) and its thermal conductivity is low
  (3.9 W.m-1.K-1). This large temperature
  difference is interesting to get high efficiency.

Fig.2a
Fig.2b
Fig.2b
The second mode of working (i.e. BHPW) takes
advantage of the large surface-to-volume ratio of
the membrane to use it as a radiator, the hot
side being the silicon frame (Fig. 2b). The heat
source may be the heat generated by a living
creature while the coolant could be simply air.
10
The fabrication method

• Low stress-silicon nitride and silicon dioxide
  sandwich layers were deposited on a
  lt100gt-oriented silicon wafer by low pressure
  chemical vapor deposition (LPCVD). A window was
  etched in the dielectric multilayer on the back
  of the silicon wafer by plasma etching. A
  polycrystalline silicon layer was deposited by
  LPCVD at 600C and patterned by wet etching, to
  define the position of the thermoelectric legs on
  the front side of the wafer. Selected legs were
  implanted with boron (p-type) while other legs
  were implanted with phosphorus (n-type).

Top view of a silicon-based thermoelectric
micro-generator.
Figure 2. Section of microgenerator
11
Experiment

• Using the model proposed by Koslov, assuming a
  onedimensional heat transfer along the
  thermoelectric legs, an active material with a
  low figure-of-merit and neglecting the heat
  losses by radiation and convection, it can be
  easily demonstrated that the maximum electrical
  power produced by a mRTG, at a given heating
  power is obtained for a thermoelectric leg
  thickness calculated from
• K1 d1 K2 d2
• where K1, K2 and d1, d2 are the thermal
  conductivities and thicknesses of the dielectric
  membrane and of the thermoelectric material,
  respectively
• The optimum leg length is calculated from
• l2/l11/3
• where l1 is half of the self-standing membrane
  width and l2 is the thermoelectric leg length.

12
Results
Polisilicon e KT mW/mK dT nm ZTm ?T K ?V V W mW
L 1,6 x 1,6 mm 50 couples P 1 mW vacuum 180 150 0,014 8,0 0,13 0,090
L 1,6 x 1,6 mm 50 couples P 1 mW air 150 190 0,016 5,1 0,084 0,058
L 1,6 x 1,6 mm 500 couples P 5 mW vacuum 270 140 0,014 9,9 1,6 0,58
L 1,6 x 1,6 mm 500 couples P 5 mW air 200 240 0,018 4,2 0,70 0,24
L 1,6 x 1,6 mm 500 couples P 10 mW vacuum 270 140 0,014 20 3,3 2,3
L 1,6 x 1,6 mm 500 couples P 10 mW air 200 240 0,018 8,4 1,4 0,98
13
Advantages of Superlattice Thermoelectric
Technology

• Enhanced efficiency
• Super-fast cooling and heating
• Enhanced power density
• Localized cooling/ heating technology

1/40,000th the actual TE material requirement
of bulk technology for same functionality low
recycle costs Eco-friendly technology
14
CONCLUSION

• A new family of promising thermoelectric
  materials with the skutterudite crystal
  structure has been presented. The possibilities
  of finding candidates for a particular operating
  temperature are great in such a large family of
  materials. Initial results obtained on some of
  their representatives demonstrate the great
  potential of skutterudites for high ZT values as
  very high mobility and very low lattice thermal
  conductivity can be obtained with materials of
  the same crystal structure. In particular, that
  if the high mobility of the binary skutterudite
  compounds can be somewhat preserved, there are
  several approaches for large reductions in
  thermal could lead to ZT values substantially
  larger than 1.
• In-plane thermoelectric micro-generators are
  very promising for powering micro-systems. A
  heating power of about 100 mW may be enough to
  produce 1 mW of useful electrical power in
  vacuum, using thin film technology.
• Thermoelectric micro-generators based on
  thick-film technology will be able to work in
  air. They will take advantage of their large
  surface-to-volume ratios to improve the coupling
  between the heat reservoirs and the thermo
  elements. This makes it a very promising device
  to efficiently convert heat wasted by our body to
  electrical power. A compact thermoelectric device
  may be able to produce as much as 60 mW with an
  output voltage of about 1.5 Volt. Nevertheless,
  the electrical contact resistances have to be
  lowered to a satisfactory level, good
  thermoelectric materials have to be used and
  thermoelectric thick-film technology needs to be
  improved or developed to get films with good
  thermoelectric properties at an acceptable
  economical cost.

15
REFERENCES
1 Hicks, L.D et al., Effect of quantum-well
structures on thermoelectronic figure of merit
Phys.Rev.B.Vol 47, No 19 (1993),
pp.12727-12731 2 Dresselhouse, M.S. et
al., Low Dimensional Thermoelectrics, Proc.16
th International Conference on termoelectrics,
Dresden, Germany, August 1997, pp 92-99 3
D.-J. YAO, C.-J. KIM, and G. CHEN Design of
thin-film thermoelectric microcoolers, in
HTD-Vol. 366-2, Proceedings of the ASME Heat
Transfer Division 2000, Volume 2, ASME
2000. 4 Gh.V.Cimpoca et all, Physics and
Modeling of Thermoelectric Microgenerators, 6 th
International Balkan Worckshop on Applied
Physics, Constanta, Romania, 5-7 July, 2005.