Evaluation of mesoporous silicon thermal conductivity by electrothermal finite element simulation
© Siegert et al.; licensee Springer. 2012
Received: 30 April 2012
Accepted: 14 July 2012
Published: 31 July 2012
The aim of this work is to determine the thermal conductivity of mesoporous silicon (PoSi) by fitting the experimental results with simulated ones. The electrothermal response (resistance versus applied current) of differently designed test lines integrated onto PoSi/silicon substrates and the bulk were compared to the simulations. The PoSi thermal conductivity was the single parameter used to fit the experimental results. The obtained thermal conductivity values were compared with those determined from Raman scattering measurements, and a good agreement between both methods was found. This methodology can be used to easily determine the thermal conductivity value for various porous silicon morphologies.
KeywordsMesoporous silicon Thermal conductivity Electrothermal simulation Calibration
Application of porous silicon for electrical isolation in radiofrequency devices is widely studied [1, 2]. However, characterization of its thermal properties is very important, too, in order to anticipate any thermal risk due to self-heating effects. Thermal conductivity of porous silicon depends mainly on its own morphological parameters at nanoscale level, such as pore morphology and average size of Si nanocrystallites .
There are several existing methods to measure thermal conductivity which include scanning thermal microscopy, that maps the local temperature and thermal conductivity of an interface , the 3ω method, which determines the thermal conductivity by applying an AC signal to metal strip , or the modeling approach , which is used in modeling the effect of pore size, pore arrangement, and porosity. The principle of comparing electrothermal simulations and experiments in order to determine thermal parameters of a material has already been applied successfully [7, 8].
In this study, we propose to apply this method in order to evaluate the thermal conductivity of a porous silicon nanomaterial (mesoporous in our case) by electrical characterization of metal strips with subsequent electrothermal finite element simulations. These metal strips are integrated with the dielectric layers on top of bulk silicon and mesoporous silicon PoSi/silicon substrates. The electrical tests have been performed with a probe station on various test lines with different widths of 10, 30, 50 μm and a constant length of 1,100 μm. The electrical resistance of the test line is measured versus the applied current. This resistance is not constant with the current and varies due to the resistance temperature dependence. Firstly, the model parameters have been calibrated by comparing experimental and simulation results for test lines integrated onto bulk silicon substrate. Secondly, thermal conductivity of the mesoporous silicon has been determined by fitting the experimental and simulation results for test lines integrated onto PoSi/Si substrate. To validate this approach, a micro-Raman thermal conductivity measurement is performed.
Six inch porous silicon samples were fabricated by anodization of highly n-doped (100) silicon wafers (ρ = 10 mΩ·cm). The anodization was run in a HF (50%)-acetic acid-water solution (4.63:2.14:1.43). Acetic acid has been used as a surfactant which improves the electrolyte penetration into the pores . Moreover, acetic acid is known to improve the roughness and the uniformity of porous layers as well as to increase the etch rate .
Test line patterns
Test lines with various designs were integrated on the porous silicon samples of highly n-doped silicon (ρ = 10 mΩ·cm). A 500-nm plasma enhanced chemical vapor deposition (PECVD) oxide has been deposited as a cap layer on the porous silicon. Test lines were made with 1-μm aluminum deposited at 350°C by physical vapor deposition (PVD). A 500-nm PECVD oxide was then deposited to passivate the aluminum. Finally, the electrical contact to the first metal level has been made with 1-μm PVD aluminum. The patterns were defined by standard photolithography and dry etching with chloride-based plasma by reactive ion etching.
Results and discussion
where R(T) is the resistance depending to the temperature; R0, the resistance at 0°C; and TCR0, the temperature coefficient of resistance at 0°C.
Alpha coefficient calculation for the different dimension strips for both substrates (Si and PoSi)
The simulator used is the Sentaurus Sdevice tool from Synopsis, Inc. (Mountain View, CA, USA). The structure is created and meshed. Electrothermal simulation is then performed. The model uses Poisson's equation coupled with temperature. The temperature induced by the current is calculated, and the dissipation is computed by using Fourier's law. Both effects are calculated at the same time (coupled simulation).
This model depends on many parameters necessary to be implemented in order to make the simulation reliable.
In order to model the experiments, different parameters (dimensional, material properties, boundaries conditions) have to be investigated, determined, and implemented in the simulation. Electrothermal simulation requires a good knowledge of the materials' properties; for this reason, the model simulation has to be calibrated.
Dimensions (μm), resistance (Ω), and resistivity (Ω·m) of the three patterns
R(Ω) at 30°C
ρ(Ω·m) at 30°C
3.26 × 10−8
3.30 × 10−8
3.33 × 10−8
with a = 0.03 K·m·W−1b = 1.56 × 10−3 m·W−1, and c = 1.65 × 10−6 m·K−1·W−1.
The silicon thermal conductivity (TC) varies with a number of parameters such as impurities, doping level, and carrier concentration. In this work, we have taken the general equation of silicon TC and make the assumption that the difference between both values was too weak to influence the final result (metal heating).
· Boundary conditions. The meaning of the boundary conditions (BC) is to replicate the experimental setup . There are two kinds of BCs: thermal and electrical. The input of the electrical BC is rather trivial; it is just a reference potential, and the input/output terminal (electrical contact) has to be set. The specification of the thermal BCs, however, is more complex. Two thermal conditions may be defined: at the top surface, to take into accounts the thermal exchange with air, and at the bottom surface, at the chuck/substrate interface. The chuck temperature is at 30°C; this temperature is applied to the bottom surface. At the top surface, there is natural convection with air. We use the traditional value of natural convection of 26 W/(K·m²).
Model simulator calibration on silicon substrate
Two conclusions can be made. Firstly, values of simulated resistance and experimental resistances are matched at low current. This means that electrical and dimensional parameters for the aluminum are correct. Secondly, for the high current, the simulation diverges from experiments. Simulated DC resistance evolution is lower than the experimental DC resistance. The simulation computes a higher dissipation of the heat generated in the line than what actually occurs in the experiments.
Two parameters have been chosen: the thermal conductivity of the oxide and the BC at the bottom surface. The goal is to vary these parameters in the simulation in order to reach a fit of experimental data. A design experiment is performed on these two parameters defined as the factors. The response is the alpha coefficient. Simulations are performed for each pattern. This makes a total of 27 simulations (3 × 9).
Comparison of the alpha coefficients between the experiments and the simulations after calibration
α Experimental (Ω·W−1)
α Simulated (Ω·W−1)
The parameters found and the conditions in the probe station are fixed and used for the PoSi thermal conductivity determination.
PoSi thermal conductivity determination
The thermal conductivity of the porous silicon is largely dependent on many parameters, intrinsically to porous silicon: pore size, depth, density, substrate P or N .
Electrical tests are performed on the same test line integrated with PoSi. The same methodology is used as described above: the resistance versus current is measured; the resistance versus dissipated power is plotted; and then the alpha coefficient (slope) is computed.
By multi-response optimization, a thermal conductivity is deduced to fit the experimental alpha value for each pattern. The best results (with the lowest error with experimental values) give 4.2 W/(K·m).
Thermal conductivity micro-Raman measurement
Micro-Raman scattering spectrometry  has been used to measure directly thermal conductivity value of our porous Si layers. From this measurement, we have obtained a value of 5 W/(K·m), confirming the validity of our simulation methodology.
The difference between these methodologies can explain the small gap between both values.
In this work, we have evaluated by electrical measurement coupled with electrothermal simulation the mesoporous silicon thermal conductivity. In order to validate the determined value, we also performed a micro-Raman thermal conductivity measurement. We found a good agreement between both values.
In the model, some assumptions have been taken: mesoporous silicon is assumed to be a bulk material. The pore size and the morphology are not considered. Therefore, the thermal-determined conductivity value is an average value.
Moreover, to apply this methodology, the thermal conductivity value of the PS needs to be much lower than that for the silicon substrate in order to characterize the temperature rise difference of the test line between the silicon and the porous silicon. The thickness has also to be high enough in order to influence the thermal dissipation and hence the thermal characterization.
In this study, we have characterized test lines under a probe station by measuring the resistance rise versus the applied current. This characterization has been performed on two different substrates: a silicon substrate and a mesoporous silicon/silicon substrate. Firstly, the model simulator has been calibrated in order to replicate the experiments, and secondly, the thermal conductivity of the porous silicon layer has been evaluated. The determined value of 4.2 W/(K·m) is in agreement with the value obtained from micro-Raman scattering measurements. Our method appears to be efficient and useful in order to determine thermal conductivity of porous silicon morphology provided that its thermal impact is significant.
- Nam C-M, Kwon Y-S: High-performance planar inductor on thick oxidized porous silicon (OPS) substrate. Microwave and Guided Wave Letters, IEEE 1997, 7(8):236–238. 10.1109/75.605489View ArticleGoogle Scholar
- Capelle M, Billoué J, Poveda P, Gautier G: N-type porous silicon substrates for integrated RF inductors. Electron Devices, IEEE Transactions on 2011, 58(11):4111–4114.View ArticleGoogle Scholar
- Korotcenkov G, Cho B: Silicon porosification: state of the art. Critical Reviews in Solid State and Materials Sciences 2010, 35(3):153–260. 10.1080/10408436.2010.495446View ArticleGoogle Scholar
- David L, Gomes S, Galland P, Vassort B, Raynaud M: Characterisation of thin films using scanning thermal microscopy. In 17th European Conference on Thermophysical Properties (ECTP). , Bratislava; 2005.Google Scholar
- Jou R: Heat transfer investigations of a nano-porous silicon film deposited on a flexible cyclic olefin copolymers substrate. Key Engineering Materials 2008, 364: 931–936.View ArticleGoogle Scholar
- Randrianalisoa J, Baillis D: Thermal conductivity across nanostructured porous silicon films. Journal of Physics: Conference Series 2007, 92: 012081.Google Scholar
- Crowell C, Shih C, Tyree V: Simulation and testing of temperature distribution and resistance versus power for SWEAT and related joule-heated metal-on-insulator structures. In 28th Annual Proceedings International Reliability Physics Symposium: 27–29 March 1990; New Orleans. IEEE, New York; 1990:37–43.Google Scholar
- Feng B, Ma W, Li Z, Zang X: Simultaneous measurements of the specific heat and thermal conductivity of suspended thin samples by transient electrothermal method. Review of Scientific Instruments 2009, 80: 064901. 10.1063/1.3153464View ArticleGoogle Scholar
- Canham L, Bayliss S: Properties of porous silicon: EMIS datareviews series no. 18 [book review]. Engineering Science and Education Journal 1997, 8: 58.Google Scholar
- Cojocaru A, Carstensen J, Ossei-Wusu E, Leisner M, Riemenschneider O, Föll H: Fast macropore growth in n-type silicon. Physica Status Solidi (C) 2009, 6: 1571–1574. 10.1002/pssc.200881031View ArticleGoogle Scholar
- Grove AS: Physics and Technology of semiconductor devices. Wiley, New York; 1967.Google Scholar
- Glassbrenner CJ, Slack Glen A: Thermal conductivity of silicon and germanium from 3 K to the melting point. Phys Rev 1964, 134: 1058–1069.View ArticleGoogle Scholar
- Rzepka S, Banerjee K, Meusel E, Hu C: Characterization of self-heating in advanced VLSI interconnect lines based on thermal finite element simulation. Components, Packaging, and Manufacturing Technology, Part A, IEEE Transactions 1998, 21(3):406–411. 10.1109/95.725203View ArticleGoogle Scholar
- Kleiner M, Kuhn S, Weber W: Thermal conductivity measurements of thin silicon dioxide films in integrated circuits. Electron Devices, IEEE Transactions 1996, 43(9):1602–1609. 10.1109/16.535354View ArticleGoogle Scholar
- Lysenko V, Perichon S, Remaki B, Barbier D: Thermal isolation in microsystems with porous silicon. Sensors and Actuators A: Physical 2002, 99: 13–24. 10.1016/S0924-4247(01)00881-0View ArticleGoogle Scholar
- Gomes S, David L, Lysenko V, Descamps A, Nychyporuk T, Raynaud M: Application of scanning thermal microscopy for thermal conductivity measurements on meso-porous silicon thin films. J Phys D: Appl Phys 2007, 40: 6677. 10.1088/0022-3727/40/21/029View ArticleGoogle Scholar
- Lysenko V, Perichon S, Remaki B, Barbier D, Champagnon B: Thermal conductivity of thick meso-porous silicon layers by micro-Raman scattering. J Appl Phys 1999, 86: 6841. 10.1063/1.371760View ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.