Sensors and Actuators A xxx (2005) xxx–xxx

Mechanical characterization of thermal flow sensors membranes N. Sabat´e a,∗ , I. Gr`acia b , J. Santander b , L. Fonseca b , E. Figueras b , C. Can´e b , J.R. Morante a a

Departament d’Electr`onica, Universitat de Barcelona, Mart´ı i Franqu´es 1, 08028 Barcelona, Spain b Centro Nacional de Microelectr´ onica(IMB-CSIC), Campus UAB, 08193 Barcelona, Spain Received 1 March 2005; received in revised form 22 June 2005; accepted 14 July 2005

Abstract The aim of this paper is to determine the mechanical properties of the films that typically compose a membrane based thermal flow sensor. Departing from a sample composed of a single layer of Si3 N4 deposited by low pressure chemical vapour deposition, used as a basic mechanical support of a great variety of micromechanical sensors, the residual stress introduced by two different passivating layers deposited by plasma enhanced chemical vapour deposition has been determined. The characterization has been performed by means of the bulge test. This technique allows to obtain not only the residual stress of the samples under test, but also the Young modulus of the stacked layers. In addition it has to be emphasized that up to now, this test has been done only in bare membranes and at ambient temperature; in this work, bulge test has been carried out directly on the thermal flow sensors. As these devices operate in the range of 100–200 ◦ C, three different temperatures within this range have been tested, providing a direct measurement of the evolution of the average residual stress onto biased sensors membranes. © 2005 Elsevier B.V. All rights reserved. Keywords: Residual stress; Young modulus; Thermal stress; Micromachined membranes

1. Introduction Thermal microsensors are widely used in multiple applications such as gas and flow detection, inclinometers and bolometers. Generally, these structures are based on micromachined membranes fabricated with microelectronics technology. Up to now, most of the studies of this kind of devices have been focused on power consumption and fast response achieved by the implementation of the device over dielectric membranes [1,2], but often mechanical issues have been neglected. Regarding this, it has to be considered that thin films used in micromachined structures exhibit residual mechanical stress strongly dependent on the layer composition, the deposition technique used, and the process parameters used for their obtaining. Moreover, a deposition of a multilayer is often required, adding factors like abrupt transitions in thermal, elastic and plastic mismatch across the interfaces that have a direct effect on the resultant stress of the fabricated structures. Due to this, the resultant stresses induce sometimes undesirable con-

∗

Corresponding author. E-mail address: [email protected] (N. Sabat´e).

sequences; severe stress intensities may cause delamination and microstructural changes in the material that can lead to the breaking of the structure during the fabrication stage or affect the behaviour of the final device as well as its reliability [3]. In this sense, Rossi et al. proposed to build a smart combination of the thin dielectric layers that composed a micromachined gas sensor as an attempt to reduce the average residual stress of the membrane to safe stress intensity values [4]. However, even if the long-term mechanical stability of the membrane is ensured after fabrication, other undesired effects like out-ofplane deformation can still disturb the operating features of some kinds of devices (i.e. a non-desired out-of-plane deformation of a few microns can modify the flow regime over a thermal flow sensor membrane or entail adherence problems between a gas sensing material and its supporting sensor membrane). Moreover, in operating conditions, an additional thermal stress originated due to the difference in the thermal expansion coefficient (CTE) between the materials that compose the membrane and the silicon rim adds to the residual stress of the structure. This generally results in an increase of the outof-plane deformation in membranes already buckled and, if a certain temperature limit is exceeded, can cause the membrane fracture. Therefore, the intensity of this stress has to be

0924-4247/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.sna.2005.07.024 SNA-4892;

No. of Pages 7

2

N. Sabat´e et al. / Sensors and Actuators A xxx (2005) xxx–xxx

maintained within a certain range during the sensor service [5]. In recent approaches to these concerns, finite element methods have proved to be a powerful tool that allows to predict and interpret microsensors mechanical behaviour [6]. However, these works usually adapt mechanical material properties reported by other authors, which can sometimes restrict the accuracy of the mechanical analysis. In this paper, we present a new approach to the evaluation of stress in micromachined devices by means of a bulge test. Applied to a membrane, this test allows to determine both its Young modulus and its residual stress. Up to now, this mechanical test has been performed to bare monolayered membranes [7,8]. However, in this case, the analysis has been performed directly onto thermal flow sensors based on mono and bilayered membranes, in this way the mechanical properties of the different stacked materials that typically compose such a device have been extracted by comparison between the different available samples. Moreover, unlike previous works, the multilayered membrane structures incorporate the metallic microrresistors that appear in a typical configuration of a calorimetric flow sensor. As it will be shown, the benefit of this approach consists on the possibility of measuring not only the mechanical properties of such membranes at room temperature, but also to perform the test to biased samples and to obtain, for the first time, a direct measurement of the average stress induced in a membrane based sensor membrane under operating conditions.

Fig. 1. schematic view of the flow sensor membrane.

Table 1 Membrane composition and values for coefficients E/(1 − ν2 ) and σ 0 obtained from data fitting σ o (MPa)

E/1 − ν2 (GPa)

Sample

Composition

A

300 nm Si3 N4 LPCVD

177 ± 10

410 ± 21

B

300 nm Si3 N4 LPCVD 800 nm SiO2 PECVD

−20 ± 5

187 ± 12

C

300 nm Si3 N4 LPCVD 350 nm Si3 N4 PECVD

−204 ± 23

313 ± 30

2. Thermal flow sensor structure A typical calorimetric flow sensor consists of three resistors placed on a micromachined membrane which isolates them thermally from the silicon substrate. The central resistor, which plays the role of a heater, is biased properly to reach an operating temperature that depends on the sensitivity to be obtained from the device (generally in a range of 100–200 ◦ C). At the same time, the heating of this resistor causes the temperature of the other two resistive elements at both sides to rise. As they are placed at the same distance from the heater, they have the same temperature at zero flow. When flow passes over the sensor, it modifies the temperature gradient created by the heater and due to that one of the sensing elements increases its temperature whereas the other cools down. Flow rate is determined by the measurement of the difference between the two resistance values. Fig. 1 shows schematically the basic structure of the fabricated samples which are labelled in this paper as type A, B and C. In type A, the flow sensor resistors are placed on a LPCVD Si3 N4 membrane that plays the role of mechanical support. In type B and C samples a layer of PECVD silicon oxide or silicon nitride has been added, respectively, these layers are often necessary to protect the sensors of degradation in harsh environments. Heater and sensing resistors have been implemented in platinum, a material that in comparison to the widely used Ni/Cr or polysilicon presents higher stability, linearity and a value of TCR of 1.9 × 10−3 K−1 [9].

A 100 mm diameter, 350 ␮m thick, 100 p-type silicon double side polished wafer has been used as the starting material for the three types of fabricated flow sensors. Over this substrate a 300 nm thick LPCVD silicon nitride layer is deposited on a thin pad oxide. The silicon nitride deposition is performed at 800 ◦ C with a residual pressure between 25 and 28 Pa. Because of the high tensile stresses showed by this type of nitride after deposition, the layer is implanted with boron ions with a dose of 4.1015 at/cm2 [10]. Then, a layer of 250 nm of sputtered platinum is deposited at room temperature. Platinum needs a previous titanium layer of 25 nm to improve its adherence to the nitride. After the patterning of platinum resistors, the process ends with the silicon micromachining of the substrate, which is performed by means of a back side KOH etching. As presented in Table 1, an additional layer has been included in samples B and C. A 800 nm PECVD SiO2 layer deposited at 380 ◦ C and a pressure of 200 Pa after platinum patterning was deposited in sample B. In the case of sample C, this SiO2 passivation layer has been substituted by a PECVD Si3 N4 , layer 350 nm thick deposited at 380 ◦ C and 200 Pa of residual pressure. 3. Characterization of the residual stress In this section, the main mechanical properties of the three different stacked membranes that compose the thermal flow sensors are characterized. In order to quantify the values of residual

N. Sabat´e et al. / Sensors and Actuators A xxx (2005) xxx–xxx

3

stress of the membranes the bulge test has been performed directly onto the sensors and the average values of residual stress and the Young modulus of the stacked membranes have been determined. In addition, departing from data obtained from sample A, the residual stress and Young modulus of the PECVD layers deposited on samples B and C have been extracted. After fabrication, samples B and C type showed a buckled profile thus, providing a clear evidence about the compressive nature of the membrane stress. On the contrary, sample A had a complete flat appearance which probably was caused by a tensile residual stress (see Fig. 2). As an additional contribution to this work, we have taken advantage of this particular circumstance to calculate the prestrain of the three different samples and check the validity of the model for single bare membranes reported by Ziebart et al. [11] when it is applied to multilayered membranes. 3.1. Bulge test Load-deflection methods are particularly well-suited when applied to membrane structures as they allow to evaluate both the Young modulus E of the thin film membrane and its residual stress σ 0 . The load can be applied either locally [12], electrostatically [13] or by means of a hydrostatical pressure, known as bulge test [7,14]. In the present case, a homogeneous pressure has been applied at the lower side on the membrane and the resulting deflection has been measured by means of a microinterferometer. The relation between applied pressure and displacement of the central point of a square membrane in the large-deflection regime is expressed in (1), being d, the membrane thickness; a, the lateral dimension and λ, the perpendicular displacement of the central point [13]. σo d 2 P = 12.18 2 a



3 λ Ed 4 λ + 27.43 4 d a (1 − ν2 ) d

(1)

Values of the applied pressure extended up to 150 kPa; beyond that point, some of the samples experienced fracture. Fig. 3 shows some interference images obtained for one of the membranes labelled as A when four different values of differential pressures are applied. The main factors that affect the accuracy of the resultant values of E and σ 0 are basically, the accuracy in the measurement of the deflection (since this parameter is determined by counting interference fringes the accuracy is equal to half of the wavelength), the accuracy in the pressure measurements, the accuracy of the membrane thickness and the process dispersion [15]. Fig. 4 shows the displacement of the three types of samples for different pressures. Data have been fitted to Eq. (1) within a factor of R2 = 0.999. Table 1 shows the resulting values for coefficients E/(1 − ν2 ) and σ 0 for the three membrane types. Type A sample shows a high tensile stress, type B sample is in a weakly compressive state whereas type C sample has yielded a high compressive stress. Once obtained the mechanical parameters of sample A, the intrinsic properties of the PECVD layers deposited in samples B and C can be computed independently using Eqs. (2) and (3), where h accounts for the total thickness of each membrane and

Fig. 2. Micrographs of the micromachined membranes labelled as type A, B and C.

4

N. Sabat´e et al. / Sensors and Actuators A xxx (2005) xxx–xxx

Fig. 3. Interference image of a A-type sensor membrane for different values of applied pressure: 0 Pa (a); 2 kPa (b); 6 kPa (c) and 8 kPa (d).

di corresponds to the thickness of the different layers: σmembrane h =



σi di

(2)

i

Emembrane h  Ei di = 1 − ν2 1 − νi2

(3)

i

Thus, starting from the value of 177 ± 10 MPa found for the ensemble LPCVD Si3 N4 –Pt microrresistors, the PECVD SiO2 and Si3 N4 layers have yielded compressive stress values of −94 ± 10 MPa and −526 ± 22 MPa, respectively. These values lay within the wide range of residual stress values reported in literature for low stressed LPCVD nitrides [16] and PECVD silicon dioxides [17]. Moreover, the high compressive stress encountered in the PECVD nitride seems to be characteristic of Si H bond rich samples [17]. Values for the Young modulus of these two layers where also computed yielding a value of 101 ± 10 GPa for the PECVD oxide and 219 ± 12 GPa for the PECVD nitride. For the calculus, Poisson coefficient values of 0.17 and 0.24 were used [17]. Data is summarized in Table 2. From this results it can be concluded that, the application of bulge test to flow sensors based on stacked membranes allows to determinate their intrinsic stress and Young modulus. Moreover, the comparison between passivated (samples B and C) and Table 2 Residual stress and Young modulus of the PECVD layers typically used in passivation Thickness (nm)

Fig. 4. Out-of-plane displacement of the three types of membranes when a pressure in the range between 0 and 150 kPa is applied at the back side. Continuous lines correspond to the polynomial fitting of experimental data.

SiO2 PECVD Si3 N4 PECVD *

800 350

σ o (MPa)

E (GPa)

ν*

−94 ± 10 −526 ± 22

101 ± 10 219 ± 12

0.17 0.24

Poisson ratio used to extract Young’s modulus value.

N. Sabat´e et al. / Sensors and Actuators A xxx (2005) xxx–xxx

non-passivated samples (sample type A) has made possible the calculus of the mechanical properties of the PECVD oxide and nitride. 3.2. Optical inspection of the buckling profiles The physical appearance of a micromachined membrane provides information about the nature of its residual stress. At tensile residual stresses the membrane remains in a flat position. In contrast, at sufficient negative stresses, the membrane is no longer stabilized by its flexural rigidity and its total elastic energy is reduced by an out-of-plane deflection. Ziebart et al. distinguish in their work three different regions of residual stress depending on the reflection symmetries hold by the square membranes [11]. The factor that delimitates these regions is the reduced prestrain, a parameter independent of the membrane dimensions and related to the residual stress through Eq. (4), where a accounts for the lateral dimensions of the membrane and h corresponds to its thickness. a2 (1 − ν)σ0 (4) h2 E The first region corresponds to tensile or weakly compressive stresses. In this region the membrane is stable in a flat position and shows all the symmetries of a square. The second region starts when the reduced prestrain of the membrane overpasses the critical value ε¯ cr1 that depends on the value of the Poisson’s ratio ν as shown in Eq. (5). For the usual insulating layers used in MEMS the value of the prestrain yields around −3.5 (that is, considering a value of the Poisson ration of 0.25). In this regime of stress the membrane deflects either upwards or downwards and loses its σ z symmetry. ε¯ o =

4.363 (5) 1+ν Finally, in the third region, the membrane surpasses a second mechanical stability at ε¯ cr2 and minimizes its strain energy by sacrificing its σ x and σ y reflection symmetries. This second value of the reduced prestrain is weakly influenced by Poisson’s ratio. Whatever the value of ν, ε¯ cr2 ranges between −206 and −226. In the present case, different values of residual stresses have been obtained by the deposition of a passivation layer. Fig. 2a–c show the micrographs obtained with a confocal microscope corresponding to type A, B and C samples, respectively. It can be seen that the flat appearance of sample A corresponds to a tensile membrane whereas the out-of-plane deflection of samples B and C reveals a post-buckling state caused by a compressive stress induced by the deposition of PECVD passivation layers. From these results a first idea of the sign and the intensity of the internal stress of the membrane can be obtained. It has to be noted that, despite the non-symmetric tensile effect exerted on the membrane by the platinum tracks of the sensor resistors, the buckled profile of sample B keeps its σ x and σ y symmetries characteristic of the second region of compressive stress. At the same time, the less symmetric profile shown by sample C indicates that it has overpassed the second critical value ε¯ cr2 , what means that the stress of the membrane has become more compressive. ε¯ cr1 = −

5

As the mechanical parameters of the membranes have been already determined in the last section, the corresponding prestrain values can be computed with Eq. (4). A ε¯ o value of +1985 has been found for the high tensile type-A membranes whereas compressive type B and C have yielded a value of −32 and −695, respectively. As predicted by optical inspection, the reduced prestrain of these samples correspond to the second and third region of compressive residual stress discussed earlier. These results extend the validity of the three prestrain regions and their relation to their symmetry to multilayered sensor membranes, that is, multilayered sensor membranes of different materials and dimensions prestrained to the same εo show identical buckling profiles. 4. Membrane average stress under sensor operating conditions The paper addresses now the study of the temperature effect on the mechanical behaviour of the device. Under operating conditions a temperature distribution is created in the sensor membrane due to the Joule heating of the central microrresistor. For that reason, the materials that compose the membrane expand in relation to the silicon rim, which due to its high thermal conductivity remains at room temperature. The thermal expansion of the membrane induces an additional compressive stress that causes a measurable increase on the out-of-plane deformation of those membranes already bended. At the same time, the mismatch of the coefficient of temperature expansion (CTE) between the membrane materials and those used in the implementation of the active elements such as heaters or resistors builds up local stresses that, at high temperature operation of the sensor, can cause the membrane rupture. Effects of local stresses can not be observed until they cause irreversible damage to the structure. Generally, a FEM model is required to extract both average and local stress evolution on an operating thermal sensor. The deformation of the central point of the membrane is taken as an input value to fit the model to the average thermal stress induced by the change in temperature of the membrane. This approach requires the previous knowledge of the CTE of the different membrane materials and allows us to evaluate indirectly the induced stress only on compressive samples. In the present approach, we propose to determine the average intensity of the thermal stress by performing the bulge test to the biased samples. In this way, a direct measurement of the average stress is obtained and even induced stress in samples with an initial tensile stress-state can be determined. In this way, the stress introduced by temperature will be detected as a change on the value of the membrane residual stress measured at room temperature. The microrresistors of the samples have been heated to 75 ◦ C and 150 ◦ C above room temperature, that correspond to typical flow sensor operating values, and the deformation of the membranes have been measured with a confocal microscope. Fig. 5 shows the evolution of the membrane profiles for the three different compositions. It has to be noted that the thermal stress induced in sample A membranes is not sufficient to cause an

6

N. Sabat´e et al. / Sensors and Actuators A xxx (2005) xxx–xxx

ical operating temperature range the membrane does not reach the critical prestrain εcr1 that would cause its buckling. On the contrary, sample C increases its initial compressive stress up to a −335 MPa. According to [4,6] this value could affect the long-term mechanical stability of the membrane. 5. Conclusions

Fig. 5. Modification of the confocal membrane profiles at different temperatures of the microheater.

out-plane deflection. In samples B and C, the value of membrane deformation is directly proportional to the mean temperature of the microheater. Moreover, in the range of temperatures applied to the microheaters, the deformation of the membrane does not change its orientation under the effect of the bimetallic effect in the zone where the heater is placed. Fig. 6 shows the evolution of the average stress in the three types of membranes. This parameter corresponds to σ o computed from Eq. (1). The change of residual stress depends on the membrane composition. It can be seen that the compressive stresses originated in membranes B-type are lower that in the other cases. This is due to the presence of silicon dioxide, which has a CTE significantly low compared to the other materials that integrate the structure [18]. In this way, when the temperature of the membrane is risen, the thermal expansion of the platinum tracks and the nitride layer drags the oxide inducing tensile stresses in it. This compensates partially the compressive effect exerted by the silicon rim on the rest of the membrane. On the contrary, samples A and C show a very similar increase of compressive stress. In sample A, the initial tensile stress is partially compensated by the extrinsic stress originated thermally and within this typ-

Fig. 6. Evolution of the average residual stress versus temperature for the different types of membranes.

In this paper, the determination of the average stress of three different stacked membranes typically used in the fabrication of micromachined thermal-based sensors has been presented. The studied combinations at room temperature have yielded residual stresses that place the membranes in the three different representative prestrain regions. These regions can be identified by the inspection of the physical appearance of the membrane. In this way, the observation of the symmetries held by the square membranes has proved to be a tool to provide guiding information about the sign and intensity of the residual stress. A more accurate evaluation of the stress intensity of the membranes has been performed by means of the bulge test. This test, generally applied to bare membranes, has been applied in this work to thermal flow sensor structures. The comparison of the results obtained in monolayered and bilayered membranes has allowed us to determine the mechanical properties (Young modulus and residual stress) at room temperature of every single deposited layer. Moreover, the test has been also performed at the typical sensor operating conditions. This new approach permits to perform for the first time, a direct measurement of average thermal stresses induced in an operative membrane-based sensor and thus, to provide a new input in reliability studies of thermal sensors based on micromachined membranes. Acknowledgments This work has been financed by the Spanish CICYT projects DPI-2001-3213-CO2-01 and Ref.: TIC2001-0554-C03-02 References [1] I. Simon, N. Bˆarsan, M. Bauer, U. Weimar, Micromachined metal oxide gas sensors: opportunities to improve sensor performance, Sens. Actuators B 73 (2001) 1–26. [2] H. Baltes, O. Paul, O. Brand, Micromachined thermally based CMOS microsensors, Proc. IEEE 86 (8) (1998) 1660–1677. [3] G. Bitko, A.C.Mc. Neil, D.J. Monk, Mater. Res. Soc. Symp. Proc. 444 (1997) 221–226. [4] C. Rossi, P. Temple-Boyer, D. Esteve, Realization and performance of thin SiO2 /SiNx membrane for microheater applications, Sens. Actuators A 64 (1998) 241–245. [5] R. Kazinczi, J.R. Mollinger, A. Bossche, New failure mechanism in silicon nitride resonators, Microelectromech. Syst. (2000) 229–234. [6] J. Puigcorb´e, D. Vogel, B. Michel, A. Vil`a, I. Gr`acia, C. Can´e, J.R. Morante, Thermal and mechanical analysis of micromachined gas sensors, J. Micromech. Microeng. 13 (2003) 548–556. [7] C. Poilane, P. Delobelle, C. Lexcellent, S. Hayashi, H. Tobushi, Analysis of the mechanical behavior of shape memory polymer membranes by nanoindentation, bulging and point membrane deflection tests, Thin Solid Films 379 (2000) 156–165.

N. Sabat´e et al. / Sensors and Actuators A xxx (2005) xxx–xxx [8] V. Ziebart, O. Paul, U. M¨unch, J. Schwizer, H. Baltes, Mechanical properties of thin film from the deflection of long clamped membranes, J. Microelectromech. Syst. 7 (3) (1998) 320–328. [9] N. Sabat´e, J. Cerd`a, I. Gr`acia, J. Berganzo, C. Can´e, J.R. Morante, Evaluation of sensitive materials for integrated thermal flow sensors, in: Proceedings of the IEEE Industrial Electronics Society Conference, Piscataway, NJ IEEE, 2002, pp. 2681–2685. [10] W. Lang, G. M¨uck, J. Bausells, K. K¨uhl, N. Moldovan, S. Suski, Stress compensation techniques in thin layers applied to silicon micromachining, in: Proceedings of the Materials Research Society Fall Meeting, Boston, USA, 1992. [11] V. Ziebart, O. Paul, H. Baltes, Strongly buckled square micromachined membranes, J. Microelectromech. Syst. 8 (4) (1999) 423–432. [12] N. Tang, R.L. Engelstad, E.G. Lovell, Pooint-deflection method for in situ stress determination of advanced lithographic masks, Microelectron. Eng. 61–62 (2002) 271–276. [13] T. Bourouina, C. Vauge, H. Mekki, Variation method for tensile stress evaluation and application to heavily boron-doped square-shaped silicon diaphragms, Sens. and Actuators A A49 (1995) 21–27. [14] M. J´ozwik, P. Delobelle, C. Gorecki, A. Sabac, L. Nieradko, C. Meunier, F. Munnik, Optomechanical characterisation of compressively prestressed silicon oxynitride films deposited by plasma-enhanced chemical vapour deposition on silicon membranes, Thin Solid Films 468 (1–2) (2004) 84–92. [15] X.Y. Ye, J.H. Zhang, Z.Y. Zhou, Y. Yang, Measurement of Young’s modulus and residual stress of micromembranes, in: IEEE Seventh International Symposium on Micromachine and Human Science, 1996, pp. 125–129. [16] I. George, P. Cemeli, B. Bonvalot, M. Wagener, A. Girard, A. Zarudiansky, J. Suski, Thin membranes with optimised thermomechanical properties for microsystem applications, Sens. Actuators A 46–47 (1995) 38–42. [17] A. Tarraf, J. Daleiden, S. Irmer, D. Prasai, H. Hilmer, Stress investigation of PECVD dielectric layers for advanced optical MEMS, J. Micromech. Microeng. 14 (2004) 317–323. [18] A.K. Shinha, H.J. Levinstein, T.E. Smith, Thermal stresses and cracking resistance of dielectric films on Si substrates, J. Appl. Phys. 49 (1978).

Biographies Neus Sabat´e was born in Tarragona, Spain, in 1975. She received her BSc degree on physics from Barcelona University (Spain) in 1998. In 1999, she joined the Microsystems Department of CNM and she obtained her PhD in physics in 2003, working on the development of gas and flow sensor devices and microsystems. He is currently working for the Electronics Department of the University of Barcelona in MEMS applications for the gas sensing field.

7

Isabel Gr`acia received the PhD degree in physics in 1993 from the Autonomous University of Barcelona, Spain, working on chemical sensors. She joined the National Microelectronics Center (CNM) working on photolithography, she is currently in the microsystems department working in the gas sensing field. Joaquin Santander received his PhD degree in physics from the Autonomous University of Barcelona, Spain, in 1996. He is currently working at the Microelectronics National Center in Barcelona, as a responsible of the electrical characterization laboratory. His main research areas are related to different microelectronic technologies (CMOS, MCM, sensors, microsystems), and electrical parametric characterization using mainly test structures. Luis Fonseca was born in Barcelona, Spain, in 1966. He received his BS and PhD degrees in physics from the Autonomous University of Barcelona in 1988 and 1992, respectively. In 1989, he joined the National Center of Microelectronics as a post-graduate student, working till 1992 on the growth and characterization of thin dielectric films for VLSI and ULSI applications. After this first research period he has worked as a process engineer, currently leading the diffusion and deposition areas of the CNM production facilities. Eduard Figueras received his PhD in Physics from the Universitat Autonoma de Barcelona, Spain, in 1988. Until 1998 as Clean Room Manager at CNM he supervised all technological process developed at CNM and was responsible of the standardization of all new fabrication process. He developed a data base of standard fabrication process. His main activity is focused on microsystems fabrication process. Since 1999, he works at the Microsystems and Silicon Technologies Department in the field of gas sensors and surface and bulk micromachining. He is currently working on the developing of micromachining resonant devices to be used as gas sensors. Carles Can´e received the BSc degree in telecomunication engineering in 1986 and the PhD degree in 1989 from the Universitat Polit`ecnica de Catalunya in Barcelona, Spain. Since 1990 he is permanent researcher at the National Microelectronics Center (CNM) in Barcelona. He is currently working in the fields of sensors and microsystems and their compatibility with standard CMOS technologies. Joan Ramon Morante received the PhD in physics from the University of Barcelona in 1980. He joined the Department of Applied Physics and Electronics of the same university in 1977, and in 1986, he was appointed full professor of electronics in this department. The same year, he founded the Laboratory of Characterization of Materials for Microelectronics (LCMM). In 1991 was founded Engineering and Electronic Materials Research Group (EME), now Department of Electronics, which leads. His current research activities and projects are focused on the fields of characterization of electronic materials and processes, micromachined Si-based sensors and actuators, gas sensors and semiconductors devices.