1. Manufacturing and Characterization of Piezoelectric generators
R. Boichot1, Alaa Faid ,Joel Omale, Edgar bautista, Belraj Arunachalam
1Science et Ingénierie des Matériaux et des Procédés, SIMAP, Grenoble INP‐CNRS‐UJF, BP 75, Saint Martin d’Hères, France.

2. Outline 1.Introduction : AlN 2. Thermodynamics Modeling
High temperature Hydride Vapor
Phase Epitaxy ( °C)
2. Thermodynamics Modeling
3. Experimental set‐up
4. AlN coating in Sic /si
5. Thin Film Stress Model
6. Conclusions

3. Aluminum nitride : Properties
Hexagonal Würtzite structure
a = 3,11 Å and c = 4,98 Å => AlN 2H polytype
Intrinsic properties :
Thermal expansion coefficient (ca. 4 10‐6 K‐1) close to that of other semiconductors
High thermal conductivity ca. 200 W.m‐1.K‐1: 80% compared to Cu
High electrical resistivity (insulating : 108 to 1013 Ω.cm)
Piezoelectric (along the c‐axis)
AlN is one of the few materials with electrical insulating properties together with high thermal conductivity.
Polar plane: c (0001)
Non polar planes: m (1100) et a (1120)
Semi polar planes: (hkjl) avec l ≠ 0
Intrinsic properties :
Refractory:
high decomposition temperature around 2400°C
Good chemical stability
for instance stability versus carbon and metals
High resistance to wear
High resistance to oxidation
High resistance to etching
Wide bandgap III‐V semiconductor
Eg th. = 6.2 eV
Optically transparent throughout UV, visible and near infra‐red regions

4. Aluminum nitride : Applications
Aluminium nitride : Processing techniques
Metallurgical coatings: Protective coatings, diffusion barriers
Microelectronics: Diffusion barriers, Packaging
Optoelectronic devices (medical use, dissociation of pollutant material i.e. water purification, semiconductor illumination, laser for High Density Optical Recording); UV Light Emitting Diode LED, UV LD, white LED
High temperature, voltage, frequency Transistors (HEMT)
Piezoelectric devices : Sensors‐ Surface Acoustic Wave (SAW)
Objectives: designing the function by controlling structure, purity, defects, thickness, growth mode on different foreign substrates
Originality: temperatures higher than 1200 °C
PVT : Physical Vapor Transport
 Growth rate up to 1 mm/h
 Crystalline quality
 Oxygen pollution: Crucible materials
/ powder purification
 Process control
MOCVD : Metal‐Organic Chemical Vapor Deposition
 High purity crystals
 Epitaxial or polycrystalline (depending on supersaturation)
 Use of gas (toxic , …)
 Low growth rates (0.1 to 10 µm/h)
 Aluminum condensation for low N/Al ratio
HTCVD, HVPE : High Temperature CVD, Hydride Vapor Phase Epitaxy
 Epitaxial or polycrystalline (depending on supersaturation)
 Control of N/Al ratio
 Use of gas (corrosive , …)
 Higher growth rates
(50 µm/h epitaxial, 300 µm/h polycrystalline)
H2+NH3
AlCl3
Induction
coils 7

5. Thermodynamic calculations
Global chemical reaction
N H 3 +AlC l 3 →AlN+3HCl
Temperature
Pressure
Composition

6. Data analysis Operation conditions
Define a set of variables T, P, X (operation condition)
Analyze the relevant possible reaction, besides the main reaction for the different steps, in FactSage software.
- Chlorination system
Reaction zone for AlN formation
Consider also stability of thermodynamic stability of graphite, silicon and sapphire under high temperature hydrogen atmosphere as a function of pressure and temperature.
Operation conditions
Chlorination system: Temperature not higher than 700 ºC
The main chemical reaction is greatly dependent on the temperature, and this should not exceed 1300ºC as some decomposition reactions may appear and threat stability.
Under these condition both the chlorination system reactor and the silicon wafer do not have remarkable influence AlN formation.

8. EXPERIMENTAL Cleaning
Done to remove contaminants from previous experiments or whilst idle.
Heating
HTCVD! To attain required ‘high’ temperature for reaction.
Deposition
Main reaction step.
Cooling & Purging
Cooling down to RT at 25oC/min, and purging lines of used reactants.
Characterization
SEM
TEM
Raman
XRD

9. EXPERIMENTAL—DEPOSITION

10. Orientation / crystalline quality
Results
SEM
Morphology
Raman
Stress /strain
TEM
Cross section
X ray Diffraction
Orientation / crystalline quality
Tensile
Energy (eV)

11. Source of stress in thin films
Epitaxial:
In epitaxial growth of single crystal thin films, the strain misfit is calculated as follows.
Where ɛmisfit is the strain induced due to misfit of the lattice, dfilm is the lattice parameter (inter planar spacing) of the thin film deposited and dsubstrate is the lattice parameter of the substrate.
Thermal:
When there is a differential thermal expansion between the thin film and the substrate, there will be thermal stress in the film and substrate. It is calculated as follows:
Where ɛmisfit is the strain induced due to misfit of the thermal expansion of the thin film and the substrate, αfilm and αsubstrate are the thermal expansion co-efficient of the film and substrate and T is the thin film deposition temperature, T0 is the room temperature.
Growth stress:
The thin film growth induces stress that are carried out by surface stress effects, phase transformation, impurities effects and coalescence of grain boundaries.
ɛ is the strain induced due to coalescence of grain boundaries, L is the grain size, Δ is the distance between each grain, E is the young modulus νis the poisson ratio and νsb, νgb are the surface free energy of the crystallites and that of the grain
boundaries.

12. Multilayer thin film model for stress analysis
The thermal stress is the dominant of all in thin film layers due to heterogeneous material.
A multilayer thin film model is studied and the stress is evaluated. It is further simulated in a Matlab code for supporting experimental data.
Schematics showing the incremental steps of bending of a multilayer strip due to thermal stresses: (a) stress-free condition, (b) unconstrained strains, ɛi=αiΔT and ɛs=αsΔT, due to a temperature change, ΔT (e.g., cooling), (c) constrained strain, ɛi=ɛs=c, to maintain displacement compatibility, and (d) bending induced by asymmetric stresses.

13. Stoney Formula
The stoneys formula is used to calculate the stress induced in the thin film deposited over the substrate. Here R is the final curvature, R0 is the initial curvature tf, ts are thin film and substrate thickness and the poisson ratio is used. By convention when σ<0 then it is compression and when σ>0 it is tension.

14. Matlab Model and Conclusion
A Matlab code was programmed for the multilayer thin film stress analysis model. Here for this case it is a trilayer code.
Matlab Model was fed with input data with parameters of AlN , SiC, Si [ i.e., thickness , misfit AlN/SiC , misfit SiC/Si , Temperature of deposition , Thermal expansion coefficient , youngs modulus , poisson ratio and Island thickness of 120 nm ]
Conclusion
The multilayer thin film stress model is studied and a Matlab code was executed.
We get the results that are consistent with the SEM and Raman spectroscopy.
We find that AlN film is in Tension and SiC is in compression.
Raman peaks simulated has a shift of 0.1 cm-1 which is very close t measured at 655 and 520 respectively for AlN and Si.