- Email: [email protected]
Thick film flow sensor for biological microsystems
Sensors and Actuators A 160 (2010) 109–115
Contents lists available at ScienceDirect
Sensors and Actuators A: Physical journal homepage: www.elsevier.com/locate/sna
Thick film flow sensor for biological microsystems H. Bartsch de Torres ∗ , C. Rensch, M. Fischer, A. Schober, M. Hoffmann, J. Müller Ilmenau University of Technology, Institute of Micro- and Nanotechnologies, G.-Kirchhoff-Str. 7, 98693 Ilmenau, Germany
a r t i c l e
i n f o
Article history: Received 13 October 2009 Received in revised form 31 March 2010 Accepted 3 April 2010 Available online 29 April 2010 Keywords: LTCC Flow measurement Thick film sensor Bioreactor Biocompatibility
a b s t r a c t An anemometer for the in situ control of the flow rates in fluidic systems is designed, manufactured and characterized. For the first time, a flow sensor according to the boundary layer principle is manufactured with exclusive use of thick film technologies. This principle enables the application of the sensor for low fluid temperatures as required in biological fluid systems. The sensor is integrated in a retention module consisting of Low Temperature Cofired Ceramics (LTCC), which allow the cost-effective realisation of complex fluidic microsystems with integrated electronics by only using thick film technologies. Thermistor compositions are printed on a free-standing bridge and encapsulated to ensure biological compatibility. The encapsulation becomes possible by using an adapted technology. At the same time the design facilitates a maximal heat-insulation of the sensor element from the substrate. The control of the stress influences on the free-standing sensor bridge due to shrinking mismatch, TCE mismatch, density gradients and deformation during the lamination is investigated using design of experiments (DoE), resulting in an adapted design and fabrication process. The presented anemometer has a linear sensor characteristic for flow rates up to 80 l/min. Compatibility investigations of LTCC with biological substances will be presented. Crown Copyright © 2010 Published by Elsevier B.V. All rights reserved.
1. Introduction Miniaturized systems on the meso scale level use Low Temperature Cofired Ceramics (LTCC) as a substrate material on account of the excellent dielectric and thermal properties offered and the fact that electrical and fluidic multilayers can be easily produced [1]. Additionally, their high chemical and thermal stability are valuable features for the construction of complex fluidic systems or the packaging of chemical sensors and BIO-MEMS. An increasing interest is therefore focused on the use of LTCC fluid systems for biological processes. Complex hybrid fluid systems have already been presented [2,3]. Previous work was focused on the realisation of a modular fluid system, as described in [4,5]. Several fluidic operations such as mixing and heating of microreaction devices are already demonstrated. Further developments are now in progress to enable handling of cell fluids and to mix reagents for enzymatic assays such as the polymerase chain reaction (PCR). A basic requirement for the integration of process control units into the fluid kit is the flow monitoring, which can be done by means of an integrated sensor in a retention module. Thus, in addition to routeing, the control unit can be situated on a fluidic component. Thick film flow
sensors have already been presented. A simple component, which consists of a free standing ceramic bridge with a printed negative temperature coefficient (NTC) thermistor composition, is described in [6]. The same principle is also utilized [7] for the fabrication of gas flow sensors. Another example of a flow sensor combines LTCC technologies for simple via formation and thin film techniques to achieve a fast sensor for application in the harsh environment of a common rail diesel injection system at high pressures up to 135 MPa [8]. All of these sensors use the hot-wire principle, which requires high sensor temperatures for an adequate signal. A more accurate principle at lower temperatures, but also more demanding, is the boundary layer method, which has a linear dependency of measure and flow at small flow rates. This work demonstrates the design of a flow sensor which uses this principle and is compatible with biological processes. The exclusive use of thick film processes simultaneously guarantees a cost-effective solution. The sensor is built using LTCC Green TapeTM 951 from DuPont Nemours, from here on called 951. 2. Sensor design 2.1. Function principle
∗ Corresponding author. Tel.: +49 3677 69 3440; fax: +49 3677 69 3360. E-mail addresses: [email protected] (H. Bartsch de Torres), [email protected] (C. Rensch), [email protected] (M. Fischer).
A basic advantage of the boundary layer principle for use in biological environments is that the heater is electrically decoupled from the sensor elements. In this way, the maximal temperature of the biological fluids can be separately limited. Simultaneously, the
0924-4247/$ – see front matter. Crown Copyright © 2010 Published by Elsevier B.V. All rights reserved. doi:10.1016/j.sna.2010.04.010
110
H. Bartsch de Torres et al. / Sensors and Actuators A 160 (2010) 109–115
Fig. 2. Sensor design.
Fig. 1. Functional principle of the flow sensor with thermistor resistances R1 , R2 for sensor S1 and S2 , respectively.
differential measurement guarantees a linear characteristic. The sensor arrangement is explained schematically in Fig. 1. The electric loss at the heater determines the temperature of the fluid. The thermal flow resulting from these losses can be divided into a convective part and an undesirable conductive part. The convective part generates a symmetric temperature profile on top of the heater. Without any forced flow there is no temperature difference between the two sensors. A flow deforms the temperature profile and creates a temperature difference DTS between the sensors, which is then measured. Its dependency on the flow rate can be approximated by a linear function for small mass flows [9]. The convective part depends on the heater surface area, the heater temperature, the dimensions of the fluid channel and the flow rate [10]. To minimize the conductive heat loss, the cross-sectional area of the bridge must be minimized. 2.2. Sensor dimensioning The fluid channel with a fired cross-section of approximately 0.7 mm2 is formed by means of a bottom and a top part of the 951 ceramic. Heater and sensor elements are situated on a free standing bridge, which is suspended between them to ensure good thermal coupling to the fluid. The three-dimensional arrangement is depicted in Fig. 2. A meander of the platinum thick film paste DP 9896 (available from DuPont) forms the heater. It is screen printed with a line width of 100 m and a pitch of 200 m and has a resistance of 5 . It covers an area of 0.77 mm2 and the dimension in the direction of flow is 1.1 mm. A maximum temperature of 60 ◦ C is allowed in order to avoid disturbing the handled biological substances, leading to an excess temperature Te of 40 ◦ C, which has to be generated through the heater element. The temperature characteristic of the heater is separately calibrated. The required excess temperature will not be exceeded, if the heater is supplied with a constant voltage of 2.3 V or less. The thermistors are printed with a positive temperature coefficient composition (PTC paste 5093 D, available from DuPont). They have a theoretical resistance of 1 k and a temperature resolution of approximately 2.75 /K. The distance between heater and sensors amounts to 500 m. All electrical elements are connected by means of 200 m vias.
The bridge consists of two layers. All functional elements are screen printed between them to ensure the necessary encapsulation for biocompatibility. Exhaust channels are embossed to enable the use of the thickfilm compositions inside the multilayer. The necessity of this approach will be discussed in the next section. The thinnest commercially available layer combination is used to minimize the cross-section of the bridge. Furthermore, cut-outs decrease the thermal cross-section of the bridge and thus minimize the conductive heat loss. Six modules are arranged on one tile with a dimension of 85 mm × 85 mm. The sensor is electrically connected by means of a standard plug. HPLC standard tubing is used for the fluidic supply. 3. Technological challenge resulting from the thick film technology 3.1. Biocompatibility of the LTCC tape The compatibility of 951 with biological processes is investigated for some applications [11], but for the demanded polymerase chain reaction (PRC) and the widely distributed mammalian cell lines HEK 293 FT and CHO-K1 were no data available. Hence the compatibility of the 951 base tape and affiliated function compositions was evaluated in some assays. Typical metal screen print compositions available from DuPont and low temperature sealing glasses were tested concerning their compatibility. From the DuPont Portfolio two cadmium-free sealing glasses, the QQ550 and QQ600 encapsulant were chosen. They are designed to cover thick film resistors. Additionally, the encapsulant FX 11-036 with a low thermal expansion coefficient available from Ferro Electronic Materials was investigated. Various tests were run in which the PCR was carried out in the presence of the 951 ceramic as well as screen print compositions, and subsequently compared with a reference sample. The screen print compositions listed in Table 1 were applied to the substrates. Both cell lines were grown on the ceramic surfaces, bare and printed with the pastes listed in Table 1, and their number was counted and compared with a parallel reference. The results are shown in Table 1. The tests demonstrate that the base tape is well tolerated by all investigated biological processes, while all conductor compositions essentially have a negative influence. Sealing glass compositions strongly inhibit the PCR assay as well as the cell growth on the surface. This fact is a sign of toxic ingredients in the layer. The resulting consequence is that all functional layers should be covered with the basis tape to avoid the direct contact with the reagents. Thus, the sensor has to be encapsulated using the 951 base tape.
H. Bartsch de Torres et al. / Sensors and Actuators A 160 (2010) 109–115
111
Fig. 3. Results of paste compatibility tests: blistering generated during the cofiring of functional pastes (a) and same arrangement with exhaust channels (b).
3.2. Process compatibility of pastes
3.3. Influence of the compression on the sinter shrink
The glass chemistry of thermistor compositions with a high temperature coefficient of resistance is designed for post-fire use only. Cofiring inside of a multilayer generates gases due to chemical interactions between tape and paste reagents, which causes blistering. The PTC composition 5093D and the platinum composition DP 9896 both demonstrate this behavior. The gases must be steadily exhausted to avoid delaminating. In order to keep the thickness of the bridge as small as possible, embossed fluid channels were employed and proved to be an effective option. The fabrication technology for such channels in 951 is described in [12]. To prove the approach, test samples were prepared with and without exhaust channels by utilization of these paste compositions. Exhaust channels with dimensions of 300 m lines and spaces with a depth of 50 m were embossed into a 113 m thick 951 tape. The paste compositions were printed on a 50 m thick tape and laminated with a pressure of 2 MPa on the prepared channels. A parallel reference without exhaust was prepared to compare both processes. Fig. 3 shows the results for the platinum composition DP 9896. This paste shows moderate gassing. While the sample without exhaust channels on the left show blisters, the one with exhaust channels on the right does not show any. In our case, channels with a width of 300 m and a depth of 50 m have an adequate crosssection to exhaust the gases. The metallizations must be in contact to the channels on some points, but must not be in direct contact to the exhaust over the whole surface. Patterns with a distance of 300 m to the channels remain without blisters. The PTC composition 5093D from DuPont shows very intensive gassing during cofiring. Test patterns, which were not in direct contact with the exhaust, were deformed during the sintering process. Here, we used 300 m-wide channels with a depth of 50 m which were completely spread over the screen printed patterns to avoid blistering.
The integration of the sensor arrangement with embossed exhaust channels in a free-standing bridge across the fluid channel requires a demanding process evaluation. This is due to the fact that free-standing elements in LTCC tend to deform during the sintering step due to gravity, shrink mismatch of pastes and tape, their TCE mismatch and the compression state of the multilayer. This fact has already been demonstrated in [13]. The compression state of a ceramic multilayer is often used to compensate shrink deviations. Therefore, shrink curves are available for typical lamination pressures. During the embossing step, the tape is subjected to high compressions up to 120 MPa. Since no data exist, the compression behavior at such high pressures is investigated here. Tapes with dimensions of 30 mm × 30 mm are prepared, stacked in multilayers with a height of 1 mm, compressed in a uniaxial press with various pressures and sintered. Their x, y and z dimensions are measured after each process step. Fig. 4 depicts the volume changes concerning the different states. The green body compression stands for the volume change between the untreated and compressed tape. It increases with increasing pressure due to the compression of voids. The sinter shrink is the volume difference with reference to the total volume between the compressed state and the fired state. During the sintering process the organic part of the composite is burned out. At the sintering temperature of 850 ◦ C, the final density of the glass ceramic composite develops independently from the density in the green state. Therefore the shrink decreases with increasing pressure. The total volume change refers to the volume of the fired ceramic with regard to the volume of the untreated material and does not depend on the compression. The density curve in Fig. 5 additionally
Table 1 Compatibility of basis tape and conductor pastes, available within the 951 material system, with polymerase chain reaction (PCR) and mammalian cell lines (HEK 293 FT and CHO-K1). Material
PCR
HEK 293FT
CHO-K1
951 DP 9896 (Pt) DP 5734 (Au) DP 6145 (Ag) FX 11-036 QQ 550 QQ 600
+ + ----
+ --+ ----
+ -+ + ----
+: not affected; -: inhibited; - -: strongly inhibited.
Fig. 4. Volume changes during compression, sintering and the whole process as a function of the compression pressure.
112
H. Bartsch de Torres et al. / Sensors and Actuators A 160 (2010) 109–115 Table 2 DoE setting matrix and effects. Factor
Fig. 5. Density of the composite material after sintering as a function of the compression pressure.
stresses this fact. Hence, the density is not influenced by the compression in the green state. A strongly compressed area therefore shrinks less than a slightly compressed one. A linear dependency can be approximated with small aberration. 3.4. Shape control of free standing elements Strongly compressed free standing elements tend to sag as a consequence of the sintering behaviour. A bridge, spread out over a channel, is also deformed by the shrink and TCE mismatch of printed patterns on it. Sacrificial carbon inlays support free-standing elements and thus influence the deformation. These contributions must be considered for the stress control in order to achieve an even bridge. An experimental setup using the design of experiments method (DoE) was carried out for the quantitative rating of the respective deformation influences. A bridge with a length of 1 mm is used in the test setup, which is depicted in Fig. 6. Two ceramic half-shells, consisting of three layers of 951 each, form the bottom and top piece with the fluid channel. The sensor bridge is clamped between them and consists of one 113 m thick layer of 951. The whole bridge is uniformly compressed with a pressure of 50 MPa to reproduce the embossing step in the test setup. Lines of the platinum paste DP 9896 with a width of 100 m and a pitch of 300 m act as a model for sensor and heater meanders. The DoE setting matrix is given in Table 2 and the factor effects shows Fig. 7. The bridge compression is the first factor A. Uncompressed bridge tapes were used for graduation 1 and those compressed for 8 min at 55 ◦ C with a pressure of 50 MPa in a uniaxial press for graduation 2. The influence of screen printed patterns represents the second factor B. Tapes without printing were used for graduation 1 and those with printing for graduation 2. The influence of carbon inlays was investigated as the third factor C. A high purity carbon black tape (TCS-CARB-1 from Harmonics Inc. Seat-
Fig. 6. Sample arrangement for the investigation of the stress influences.
Graduation 1
2
A = compression effect A1 /A2
0 MPa 42 m
50 MPa 22 m
B = screen print effect B1 /B2
0 (without) 1 m
1 (with) 64 m
C = carbon inlay effect C1/C2
0 (without) 21 m
1 (with) 44 m
tle) with adapted thickness was used. Graduation 1 corresponds to experiments without inlay and graduation 2 to experiments with inlay. The investigated factors are assumed to superpose without interactions, thus their contribution on the deformation can be evaluated by means of an L4-matrix using Taguchis method for three factors and two graduations [14]. Four experiments are necessary to determine the factor effects. Thus, four parts were prepared according to the experiment matrix in Table 2. The prepared halfshells and bridges were uniaxially laminated at a pressure of 2 MPa and a temperature of 70 ◦ C for 6 min. After the subsequent sintering with a standard profile at 850 ◦ C, the tile was sawed to prepare a cross-section of the bridge and measure the deformation. Fig. 8 depicts cross-sections of test bridges to illustrate the assessment. For each experiment four parts are analyzed and the factor effects are calculated from the analysis of means. These results illustrates Fig. 7. Using the experiment setup with factor graduations A2, B1 and C2 even bridges were achieved as shown in Fig. 8b. Here, the deformation of the compressed bridge is compensated by the use of a carbon inlay. It can be concluded, that the assumed superposition is given. 4. Sensor manufacturing and characterization Fig. 7 clearly shows that the printed patterns have the strongest influence on the deformation. The deformation caused by the use of carbon inlays acts in the same direction and that of the compression opposite. Hence, the influence of printed patterns in the test assembly in Fig. 6 cannot be compensated for and additional measures are necessary to manufacture an even sensor bridge. The restriction of the heater meanders to half of the bridge length reduces the stress influenced bridge area and, thus, the deformation. A further reduction of the buckling is achieved by using a reinforced bridge. Additionally, this reinforcement covers the active elements to ensure biocompatibility. The manufactured assembly is depicted in Fig. 9. To obtain a good thermal decoupling from the bridge to the substrate, cut-outs are provided as shown in the sketch in Fig. 2.
Fig. 7. Factor effects on the deformation.
H. Bartsch de Torres et al. / Sensors and Actuators A 160 (2010) 109–115
113
Fig. 10. Process flow for the manufacturing of the retention module.
Fig. 8. Measurement of the deformation on a test bridge: (a) buckling of a bridge with screen printed patterns and (b) even bridge with compression and carbon inlay.
The process flow for the assembly is depicted in Fig. 10. Top part, bottom part and bridge are separately prepared. The bridge consists of one embossed layer and the cover layer, on which the sensor elements are printed on. The blanked sheets are conditioned in a convection oven for volatilization of the solvents. Electrical vias and fluid channels are punched into all blanks and the electrical vias are filled with the composition DP 5738. The exhaust channels are embossed into the bridge tape with a 55 m (+5/−0 m) deep silicon tool at a temperature of 60 ◦ C and an embossing pressure of 100 MPa. Fig. 11 shows the profile scan of the embossed tape. The mould is reproduced with high accuracy. A detailed description of the embossing process for 951 tapes is presented in [15].
Fig. 9. Layer assembly for the retention module.
Heater and sensor elements are screen printed on the bottom of the 50 m thick cover layer. Subsequently, both bridge layers are uniaxially laminated at 70 ◦ C for 6 min with a reduced lamination pressure to maintain the exhaust channels [12]. Cut-outs and fluid channels are then created in this laminated bridge sheet by use of a UV-laser. The fluid channels of the top and bottom part are also laser cut. The outside tapes and two fluid channel tapes are uniaxially laminated at 70 ◦ C for 5 min using a lamination pressure of 2 MPa in order to form the bottom and top part. Afterwards, all parts are stacked and the whole module is laminated at a pressure of 2 MPa at 70 ◦ C for 5 min. Then the substrate is co-fired at 850◦ in a standard firing step. Some modules were cut to examine the bridge formation. The sensor bridges are evenly spread out over the channel. Fig. 12 shows a cross-section of a sensor bridge and the fluid channel. It is apparent from this picture that the process is adequate to produce even bridges without deformation. Six retention modules are separated from each ceramic tile by sawing. Finally, cable connectors are soldered on the designated pads. The sensor is fitted into a connector system with minimum death volume, which uses standard plugs to link the fluidic [5]. The assembly is shown in Fig. 13. The tolerance range of thick film resistors requires that the temperature characteristic for each sensor element is determined. This is done in an oil bath with constant temperature control in the range between 20 and 80 ◦ C for every thermistor. The characteristics are logged and used to calculate the temperature difference between the sensor elements. In future designs these values can be stored in an integrated memory unit on the module.
Fig. 11. Embossed bridge layer and profile scan of the embossed channels.
114
H. Bartsch de Torres et al. / Sensors and Actuators A 160 (2010) 109–115
5. Summary
Fig. 12. Cross-section through a sensor bridge.
A flow sensor has been monolithically integrated in a retention module, which is part of a modular LTCC fluidic system. The biocompatibility of the material system 951 from Du Pont was investigated and found to be poor for screen printed thick film compositions. Therefore, an encapsulated design of the sensor was chosen. To ensure the process compatibility, embossed exhaust channels are used to draw out the gases, which are generated during the sintering process. The stress control of the highly compressed sensor bridge was carried out by means of the Design of Experiments. An even bridge with embossed exhaust channels and screen printed sensor and heater elements was achieved by adapting the bridge design. The sensor was manufactured using exclusively thick film techniques and was integrated into a retention module of the fluid handling system. The temperature difference was measured as the sensor characteristic curve. For flow rates up to 80 l it can be approximated by a linear function. Utilizing this data, the concept for a thick film flow sensor working according to the boundary layer principle was proven, which allows the adequate measurement of small flow rates at low fluid temperatures.
Acknowledgements We thank the BMBF and the Thuringian Ministry of Culture for the financial support within the Centre for Innovation Competence MacroNano® .
References
Fig. 13. Retention module with connections.
Afterwards, the sensor arrangement is calibrated for each module by use of a precision analytical pump. The flow rate is forced with high accuracy and the respective temperature difference is calculated utilizing the data of the sensor element characteristics. Water at room temperature (22 ◦ C) is pumped through the sensor system at defined flow rates between 0.02 and 10 ml/min. Both sensor resistances are measured with a multi meter and their difference is applied to the former logged temperature characteristic. With this information the temperature difference DTS is calculated and the resulting graph is shown in Fig. 14. A linear characteristic is demonstrated for flow rates up to 80 l/min. The power consumption of the heater amounts to 455 mW.
Fig. 14. Sensor characteristic curve.
[1] M.R. Gongora-Rubio, M. Bariatto, A. Fontes, E. Wagner Simões, J. SantiagoAvilés, LTCC hybrid technology applied to water quality meso analytical system, Electrochemical Society Proceedings 8 (2002) 142–150. [2] L.J. Golonka, Technology and applications of low temperature cofired ceramic (LTCC) based sensors and Microsystems, Bulletin of the Polish Academy of Sciences: Technical Sciences 54 (2006) 221–231. [3] K.A. Peterson, K.D. Patel, C.K. Ho, S.B. Rohde, C.D. Nordquist, C.A. Walker, B.D. Wroblewski, M. Okandan, Novel Microsystem Applications with New Techniques in LTCC, CICMT 2005 Ceramic Interconnect and Ceramic Microsystems Technologies Proceedings. [4] G.A. Groß, T. Thelemann, S. Schneider, D. Boskovic, J.M. Köhler, Fabrication and fluidic characterization of static micromixers made of low temperature cofired ceramic (LTCC), Chemical Engineering Science 63 (2008) 2773–2784. [5] T. Thelemann, J. Müller, M. Fischer, A. Groß, LTCC based fluidic components for chemical applications, Journal of Microelectronics and Electronic Packaging 4 (2007) 167–172. [6] M. Gongora-Rubio, L.M. Solá-Laguna, P.J. Moffett, J.J. Santiago-Avilés, The utilization of low temperature co-fired ceramics (LTCC-ML) technology for meso-scale EMS, a simple thermistor based flow sensor, Sensors and Actuators 73 (1999) 215–221. [7] J. Kita, F. Rettig, R. Moos, K.-H. Drüe, H. Thust, Laser forming of LTCC ceramics for hot-plate gas sensors, in: Proceedings of XXVIII International Conference of IMAPS Poland Chapter, 2004. [8] U. Schmid, G. Krötz, D. Schmitt-Landsiedel, A volumetric flow sensor for automotive injection systems, Journal of Micromechanics and Microengineering 18 (2008) 045006. [9] O. Fiedler, Strömungs- und Durchflussmesstechnik, 1. Auflage München: R, Oldenburg Verlag (1992) 183. [10] K.W. Bonfig, Technische Durchflussmessung, 3, Vulkan-Verlag Essen (Germany), 2002, p. 186. [11] L. Rebenklau, Merkmale der Biokompatibilität von Industriellen Niedrigtemperatur-Glaskeramiken, 1st Workshop BioAVT, TU Dresden, 2003. [12] H. Bartsch de Torres, M. Hoffmann, Embossing of fluidic channels into ceramic multilayer, Proceedings of the Smart Systems Integration Conference (2007). [13] P. Espinoza-Vallejos, J. Zhong, M. Gongora-Rubio, L. Sola-Laguna, J.J. SantiagoAviles, The measurement and control of sagging in meso (intermediate scale) electromechanical LTCC structures and systems, MRS Conf. Proc., 518, 1998, pp. 73–79. [14] B. Klein, Versuchsplanung DoE, München, Oldenburg Wissenschaftsverlag, 2004, pp. 35–53. [15] H. Bartsch de Torres, A. Albrecht, M. Hoffmann, Systematic characterisation of embossing processes for LTCC-tapes, Journal of Microelectronics and Electronic Packaging 5 (2008) 142–149.
H. Bartsch de Torres et al. / Sensors and Actuators A 160 (2010) 109–115
Biographies
Heike Bartsch de Torres received her diploma degree in precision engineering from the Ilmenau University of Technology, Ilmenau, Germany, in 1994. Her research interest has developed along the lines of micromechanics and fluidic systems. After working in industry, she returned to the Ilmenau University of Technology in 2005. At the present she is a Ph.D. student at the Institute of Micro- and Nanotechnologies. Her current interests are focused on micro-structuring methods for ceramic green tapes, thick film sensor applications and microfluidics.
115
Andreas Schober received his physics diploma from the Ludwig Maximilians University (LMU) at the MPI of Quantum Optics in Munich (Professor Walther). Focusing 1990 his main interest towards biological problems, he worked for his PhD at the MPI of biophysical chemistry in the department of Professor Eigen on “strategies of evolutionary biotechnology”. Heading different groups in the fields of nanosystem integration, microfluidics and microreaction technology in industry and science, he became head of the research group “Microfluidics and Biosensors” 2006 in the ZIK MacroNano, the BMBF centre for innovation competence at the IMN and 2007 acting head of the nanotechnology group.
Martin Hoffmann worked from 1992 to 2003 as scientific staff member and chief engineer at Dortmund University, Germany at the High Frequency Institute. From 2003 to 2005 he held R&D positions in the MEMS companies HL Planartechnik, Dortmund, and Silicon Manufacturing Itzehoe, Germany. Since 2006 he is the Head of Department “Micromechanical Systems” at the Ilmenau University of Technology and since 2007 the Director of the Institute IMN MacroNano® Ilmenau University of Technology.
Christian Rensch graduated from the Ilmenau University of Technology in 2008 with a diploma degree in micromechatronics. During his studies he worked at the Institute of Micro- and Nanotechnologies (Ilmenau, Germany) on acoustic bubble detection techniques for print heads and Low Temperature Co-Fired Ceramic (LTCC) based microfluidic systems. At MicroFab Technologies Inc. (Dallas, USA) he investigated new actuation signals for microfluidic drop-on-demand systems. In 2008 he joined GE Global Research (Munich, Germany) and his activities there focus on the development of a chip-based microfluidic synthesizer for radiopharmaceuticals.
Jens Müller received his diploma degree in electrical engineering and the doctoral degree from Ilmenau University of Technology, Ilmenau, Germany, in 1992 and 1997, respectively. From 1997 to 2005, he held managing positions in development departments at MicroSystems Engineering GmbH, Berg, Germany. In 2005, he returned to Ilmenau University of Technology to establish the junior research group “Functionalised Peripherics”. In July 2008 he was assigned full professor for the department of Electronics Technology at the same university. His research interest covers functional integration for ceramic based System-in-Packages considering aspects of harsh environmental use, and high thermal/high-
Michael Fischer received the diploma degree in mechanical engineering from Ilmenau University of Technology, Ilmenau, Germany, in 1993. Currently, he is Ph.D. student at the Ilmenau University of Technology and member of the academic staff at the MacroNano® Centre for Innovation Competence focussing on the development of BIO-MEMS and MEMS-Packaging Technologies including new options for combining different material classes (silicon, ceramic, glass and polymer).
frequency requirements.
Contents lists available at ScienceDirect
Sensors and Actuators A: Physical journal homepage: www.elsevier.com/locate/sna
Thick film flow sensor for biological microsystems H. Bartsch de Torres ∗ , C. Rensch, M. Fischer, A. Schober, M. Hoffmann, J. Müller Ilmenau University of Technology, Institute of Micro- and Nanotechnologies, G.-Kirchhoff-Str. 7, 98693 Ilmenau, Germany
a r t i c l e
i n f o
Article history: Received 13 October 2009 Received in revised form 31 March 2010 Accepted 3 April 2010 Available online 29 April 2010 Keywords: LTCC Flow measurement Thick film sensor Bioreactor Biocompatibility
a b s t r a c t An anemometer for the in situ control of the flow rates in fluidic systems is designed, manufactured and characterized. For the first time, a flow sensor according to the boundary layer principle is manufactured with exclusive use of thick film technologies. This principle enables the application of the sensor for low fluid temperatures as required in biological fluid systems. The sensor is integrated in a retention module consisting of Low Temperature Cofired Ceramics (LTCC), which allow the cost-effective realisation of complex fluidic microsystems with integrated electronics by only using thick film technologies. Thermistor compositions are printed on a free-standing bridge and encapsulated to ensure biological compatibility. The encapsulation becomes possible by using an adapted technology. At the same time the design facilitates a maximal heat-insulation of the sensor element from the substrate. The control of the stress influences on the free-standing sensor bridge due to shrinking mismatch, TCE mismatch, density gradients and deformation during the lamination is investigated using design of experiments (DoE), resulting in an adapted design and fabrication process. The presented anemometer has a linear sensor characteristic for flow rates up to 80 l/min. Compatibility investigations of LTCC with biological substances will be presented. Crown Copyright © 2010 Published by Elsevier B.V. All rights reserved.
1. Introduction Miniaturized systems on the meso scale level use Low Temperature Cofired Ceramics (LTCC) as a substrate material on account of the excellent dielectric and thermal properties offered and the fact that electrical and fluidic multilayers can be easily produced [1]. Additionally, their high chemical and thermal stability are valuable features for the construction of complex fluidic systems or the packaging of chemical sensors and BIO-MEMS. An increasing interest is therefore focused on the use of LTCC fluid systems for biological processes. Complex hybrid fluid systems have already been presented [2,3]. Previous work was focused on the realisation of a modular fluid system, as described in [4,5]. Several fluidic operations such as mixing and heating of microreaction devices are already demonstrated. Further developments are now in progress to enable handling of cell fluids and to mix reagents for enzymatic assays such as the polymerase chain reaction (PCR). A basic requirement for the integration of process control units into the fluid kit is the flow monitoring, which can be done by means of an integrated sensor in a retention module. Thus, in addition to routeing, the control unit can be situated on a fluidic component. Thick film flow
sensors have already been presented. A simple component, which consists of a free standing ceramic bridge with a printed negative temperature coefficient (NTC) thermistor composition, is described in [6]. The same principle is also utilized [7] for the fabrication of gas flow sensors. Another example of a flow sensor combines LTCC technologies for simple via formation and thin film techniques to achieve a fast sensor for application in the harsh environment of a common rail diesel injection system at high pressures up to 135 MPa [8]. All of these sensors use the hot-wire principle, which requires high sensor temperatures for an adequate signal. A more accurate principle at lower temperatures, but also more demanding, is the boundary layer method, which has a linear dependency of measure and flow at small flow rates. This work demonstrates the design of a flow sensor which uses this principle and is compatible with biological processes. The exclusive use of thick film processes simultaneously guarantees a cost-effective solution. The sensor is built using LTCC Green TapeTM 951 from DuPont Nemours, from here on called 951. 2. Sensor design 2.1. Function principle
∗ Corresponding author. Tel.: +49 3677 69 3440; fax: +49 3677 69 3360. E-mail addresses: [email protected] (H. Bartsch de Torres), [email protected] (C. Rensch), [email protected] (M. Fischer).
A basic advantage of the boundary layer principle for use in biological environments is that the heater is electrically decoupled from the sensor elements. In this way, the maximal temperature of the biological fluids can be separately limited. Simultaneously, the
0924-4247/$ – see front matter. Crown Copyright © 2010 Published by Elsevier B.V. All rights reserved. doi:10.1016/j.sna.2010.04.010
110
H. Bartsch de Torres et al. / Sensors and Actuators A 160 (2010) 109–115
Fig. 2. Sensor design.
Fig. 1. Functional principle of the flow sensor with thermistor resistances R1 , R2 for sensor S1 and S2 , respectively.
differential measurement guarantees a linear characteristic. The sensor arrangement is explained schematically in Fig. 1. The electric loss at the heater determines the temperature of the fluid. The thermal flow resulting from these losses can be divided into a convective part and an undesirable conductive part. The convective part generates a symmetric temperature profile on top of the heater. Without any forced flow there is no temperature difference between the two sensors. A flow deforms the temperature profile and creates a temperature difference DTS between the sensors, which is then measured. Its dependency on the flow rate can be approximated by a linear function for small mass flows [9]. The convective part depends on the heater surface area, the heater temperature, the dimensions of the fluid channel and the flow rate [10]. To minimize the conductive heat loss, the cross-sectional area of the bridge must be minimized. 2.2. Sensor dimensioning The fluid channel with a fired cross-section of approximately 0.7 mm2 is formed by means of a bottom and a top part of the 951 ceramic. Heater and sensor elements are situated on a free standing bridge, which is suspended between them to ensure good thermal coupling to the fluid. The three-dimensional arrangement is depicted in Fig. 2. A meander of the platinum thick film paste DP 9896 (available from DuPont) forms the heater. It is screen printed with a line width of 100 m and a pitch of 200 m and has a resistance of 5 . It covers an area of 0.77 mm2 and the dimension in the direction of flow is 1.1 mm. A maximum temperature of 60 ◦ C is allowed in order to avoid disturbing the handled biological substances, leading to an excess temperature Te of 40 ◦ C, which has to be generated through the heater element. The temperature characteristic of the heater is separately calibrated. The required excess temperature will not be exceeded, if the heater is supplied with a constant voltage of 2.3 V or less. The thermistors are printed with a positive temperature coefficient composition (PTC paste 5093 D, available from DuPont). They have a theoretical resistance of 1 k and a temperature resolution of approximately 2.75 /K. The distance between heater and sensors amounts to 500 m. All electrical elements are connected by means of 200 m vias.
The bridge consists of two layers. All functional elements are screen printed between them to ensure the necessary encapsulation for biocompatibility. Exhaust channels are embossed to enable the use of the thickfilm compositions inside the multilayer. The necessity of this approach will be discussed in the next section. The thinnest commercially available layer combination is used to minimize the cross-section of the bridge. Furthermore, cut-outs decrease the thermal cross-section of the bridge and thus minimize the conductive heat loss. Six modules are arranged on one tile with a dimension of 85 mm × 85 mm. The sensor is electrically connected by means of a standard plug. HPLC standard tubing is used for the fluidic supply. 3. Technological challenge resulting from the thick film technology 3.1. Biocompatibility of the LTCC tape The compatibility of 951 with biological processes is investigated for some applications [11], but for the demanded polymerase chain reaction (PRC) and the widely distributed mammalian cell lines HEK 293 FT and CHO-K1 were no data available. Hence the compatibility of the 951 base tape and affiliated function compositions was evaluated in some assays. Typical metal screen print compositions available from DuPont and low temperature sealing glasses were tested concerning their compatibility. From the DuPont Portfolio two cadmium-free sealing glasses, the QQ550 and QQ600 encapsulant were chosen. They are designed to cover thick film resistors. Additionally, the encapsulant FX 11-036 with a low thermal expansion coefficient available from Ferro Electronic Materials was investigated. Various tests were run in which the PCR was carried out in the presence of the 951 ceramic as well as screen print compositions, and subsequently compared with a reference sample. The screen print compositions listed in Table 1 were applied to the substrates. Both cell lines were grown on the ceramic surfaces, bare and printed with the pastes listed in Table 1, and their number was counted and compared with a parallel reference. The results are shown in Table 1. The tests demonstrate that the base tape is well tolerated by all investigated biological processes, while all conductor compositions essentially have a negative influence. Sealing glass compositions strongly inhibit the PCR assay as well as the cell growth on the surface. This fact is a sign of toxic ingredients in the layer. The resulting consequence is that all functional layers should be covered with the basis tape to avoid the direct contact with the reagents. Thus, the sensor has to be encapsulated using the 951 base tape.
H. Bartsch de Torres et al. / Sensors and Actuators A 160 (2010) 109–115
111
Fig. 3. Results of paste compatibility tests: blistering generated during the cofiring of functional pastes (a) and same arrangement with exhaust channels (b).
3.2. Process compatibility of pastes
3.3. Influence of the compression on the sinter shrink
The glass chemistry of thermistor compositions with a high temperature coefficient of resistance is designed for post-fire use only. Cofiring inside of a multilayer generates gases due to chemical interactions between tape and paste reagents, which causes blistering. The PTC composition 5093D and the platinum composition DP 9896 both demonstrate this behavior. The gases must be steadily exhausted to avoid delaminating. In order to keep the thickness of the bridge as small as possible, embossed fluid channels were employed and proved to be an effective option. The fabrication technology for such channels in 951 is described in [12]. To prove the approach, test samples were prepared with and without exhaust channels by utilization of these paste compositions. Exhaust channels with dimensions of 300 m lines and spaces with a depth of 50 m were embossed into a 113 m thick 951 tape. The paste compositions were printed on a 50 m thick tape and laminated with a pressure of 2 MPa on the prepared channels. A parallel reference without exhaust was prepared to compare both processes. Fig. 3 shows the results for the platinum composition DP 9896. This paste shows moderate gassing. While the sample without exhaust channels on the left show blisters, the one with exhaust channels on the right does not show any. In our case, channels with a width of 300 m and a depth of 50 m have an adequate crosssection to exhaust the gases. The metallizations must be in contact to the channels on some points, but must not be in direct contact to the exhaust over the whole surface. Patterns with a distance of 300 m to the channels remain without blisters. The PTC composition 5093D from DuPont shows very intensive gassing during cofiring. Test patterns, which were not in direct contact with the exhaust, were deformed during the sintering process. Here, we used 300 m-wide channels with a depth of 50 m which were completely spread over the screen printed patterns to avoid blistering.
The integration of the sensor arrangement with embossed exhaust channels in a free-standing bridge across the fluid channel requires a demanding process evaluation. This is due to the fact that free-standing elements in LTCC tend to deform during the sintering step due to gravity, shrink mismatch of pastes and tape, their TCE mismatch and the compression state of the multilayer. This fact has already been demonstrated in [13]. The compression state of a ceramic multilayer is often used to compensate shrink deviations. Therefore, shrink curves are available for typical lamination pressures. During the embossing step, the tape is subjected to high compressions up to 120 MPa. Since no data exist, the compression behavior at such high pressures is investigated here. Tapes with dimensions of 30 mm × 30 mm are prepared, stacked in multilayers with a height of 1 mm, compressed in a uniaxial press with various pressures and sintered. Their x, y and z dimensions are measured after each process step. Fig. 4 depicts the volume changes concerning the different states. The green body compression stands for the volume change between the untreated and compressed tape. It increases with increasing pressure due to the compression of voids. The sinter shrink is the volume difference with reference to the total volume between the compressed state and the fired state. During the sintering process the organic part of the composite is burned out. At the sintering temperature of 850 ◦ C, the final density of the glass ceramic composite develops independently from the density in the green state. Therefore the shrink decreases with increasing pressure. The total volume change refers to the volume of the fired ceramic with regard to the volume of the untreated material and does not depend on the compression. The density curve in Fig. 5 additionally
Table 1 Compatibility of basis tape and conductor pastes, available within the 951 material system, with polymerase chain reaction (PCR) and mammalian cell lines (HEK 293 FT and CHO-K1). Material
PCR
HEK 293FT
CHO-K1
951 DP 9896 (Pt) DP 5734 (Au) DP 6145 (Ag) FX 11-036 QQ 550 QQ 600
+ + ----
+ --+ ----
+ -+ + ----
+: not affected; -: inhibited; - -: strongly inhibited.
Fig. 4. Volume changes during compression, sintering and the whole process as a function of the compression pressure.
112
H. Bartsch de Torres et al. / Sensors and Actuators A 160 (2010) 109–115 Table 2 DoE setting matrix and effects. Factor
Fig. 5. Density of the composite material after sintering as a function of the compression pressure.
stresses this fact. Hence, the density is not influenced by the compression in the green state. A strongly compressed area therefore shrinks less than a slightly compressed one. A linear dependency can be approximated with small aberration. 3.4. Shape control of free standing elements Strongly compressed free standing elements tend to sag as a consequence of the sintering behaviour. A bridge, spread out over a channel, is also deformed by the shrink and TCE mismatch of printed patterns on it. Sacrificial carbon inlays support free-standing elements and thus influence the deformation. These contributions must be considered for the stress control in order to achieve an even bridge. An experimental setup using the design of experiments method (DoE) was carried out for the quantitative rating of the respective deformation influences. A bridge with a length of 1 mm is used in the test setup, which is depicted in Fig. 6. Two ceramic half-shells, consisting of three layers of 951 each, form the bottom and top piece with the fluid channel. The sensor bridge is clamped between them and consists of one 113 m thick layer of 951. The whole bridge is uniformly compressed with a pressure of 50 MPa to reproduce the embossing step in the test setup. Lines of the platinum paste DP 9896 with a width of 100 m and a pitch of 300 m act as a model for sensor and heater meanders. The DoE setting matrix is given in Table 2 and the factor effects shows Fig. 7. The bridge compression is the first factor A. Uncompressed bridge tapes were used for graduation 1 and those compressed for 8 min at 55 ◦ C with a pressure of 50 MPa in a uniaxial press for graduation 2. The influence of screen printed patterns represents the second factor B. Tapes without printing were used for graduation 1 and those with printing for graduation 2. The influence of carbon inlays was investigated as the third factor C. A high purity carbon black tape (TCS-CARB-1 from Harmonics Inc. Seat-
Fig. 6. Sample arrangement for the investigation of the stress influences.
Graduation 1
2
A = compression effect A1 /A2
0 MPa 42 m
50 MPa 22 m
B = screen print effect B1 /B2
0 (without) 1 m
1 (with) 64 m
C = carbon inlay effect C1/C2
0 (without) 21 m
1 (with) 44 m
tle) with adapted thickness was used. Graduation 1 corresponds to experiments without inlay and graduation 2 to experiments with inlay. The investigated factors are assumed to superpose without interactions, thus their contribution on the deformation can be evaluated by means of an L4-matrix using Taguchis method for three factors and two graduations [14]. Four experiments are necessary to determine the factor effects. Thus, four parts were prepared according to the experiment matrix in Table 2. The prepared halfshells and bridges were uniaxially laminated at a pressure of 2 MPa and a temperature of 70 ◦ C for 6 min. After the subsequent sintering with a standard profile at 850 ◦ C, the tile was sawed to prepare a cross-section of the bridge and measure the deformation. Fig. 8 depicts cross-sections of test bridges to illustrate the assessment. For each experiment four parts are analyzed and the factor effects are calculated from the analysis of means. These results illustrates Fig. 7. Using the experiment setup with factor graduations A2, B1 and C2 even bridges were achieved as shown in Fig. 8b. Here, the deformation of the compressed bridge is compensated by the use of a carbon inlay. It can be concluded, that the assumed superposition is given. 4. Sensor manufacturing and characterization Fig. 7 clearly shows that the printed patterns have the strongest influence on the deformation. The deformation caused by the use of carbon inlays acts in the same direction and that of the compression opposite. Hence, the influence of printed patterns in the test assembly in Fig. 6 cannot be compensated for and additional measures are necessary to manufacture an even sensor bridge. The restriction of the heater meanders to half of the bridge length reduces the stress influenced bridge area and, thus, the deformation. A further reduction of the buckling is achieved by using a reinforced bridge. Additionally, this reinforcement covers the active elements to ensure biocompatibility. The manufactured assembly is depicted in Fig. 9. To obtain a good thermal decoupling from the bridge to the substrate, cut-outs are provided as shown in the sketch in Fig. 2.
Fig. 7. Factor effects on the deformation.
H. Bartsch de Torres et al. / Sensors and Actuators A 160 (2010) 109–115
113
Fig. 10. Process flow for the manufacturing of the retention module.
Fig. 8. Measurement of the deformation on a test bridge: (a) buckling of a bridge with screen printed patterns and (b) even bridge with compression and carbon inlay.
The process flow for the assembly is depicted in Fig. 10. Top part, bottom part and bridge are separately prepared. The bridge consists of one embossed layer and the cover layer, on which the sensor elements are printed on. The blanked sheets are conditioned in a convection oven for volatilization of the solvents. Electrical vias and fluid channels are punched into all blanks and the electrical vias are filled with the composition DP 5738. The exhaust channels are embossed into the bridge tape with a 55 m (+5/−0 m) deep silicon tool at a temperature of 60 ◦ C and an embossing pressure of 100 MPa. Fig. 11 shows the profile scan of the embossed tape. The mould is reproduced with high accuracy. A detailed description of the embossing process for 951 tapes is presented in [15].
Fig. 9. Layer assembly for the retention module.
Heater and sensor elements are screen printed on the bottom of the 50 m thick cover layer. Subsequently, both bridge layers are uniaxially laminated at 70 ◦ C for 6 min with a reduced lamination pressure to maintain the exhaust channels [12]. Cut-outs and fluid channels are then created in this laminated bridge sheet by use of a UV-laser. The fluid channels of the top and bottom part are also laser cut. The outside tapes and two fluid channel tapes are uniaxially laminated at 70 ◦ C for 5 min using a lamination pressure of 2 MPa in order to form the bottom and top part. Afterwards, all parts are stacked and the whole module is laminated at a pressure of 2 MPa at 70 ◦ C for 5 min. Then the substrate is co-fired at 850◦ in a standard firing step. Some modules were cut to examine the bridge formation. The sensor bridges are evenly spread out over the channel. Fig. 12 shows a cross-section of a sensor bridge and the fluid channel. It is apparent from this picture that the process is adequate to produce even bridges without deformation. Six retention modules are separated from each ceramic tile by sawing. Finally, cable connectors are soldered on the designated pads. The sensor is fitted into a connector system with minimum death volume, which uses standard plugs to link the fluidic [5]. The assembly is shown in Fig. 13. The tolerance range of thick film resistors requires that the temperature characteristic for each sensor element is determined. This is done in an oil bath with constant temperature control in the range between 20 and 80 ◦ C for every thermistor. The characteristics are logged and used to calculate the temperature difference between the sensor elements. In future designs these values can be stored in an integrated memory unit on the module.
Fig. 11. Embossed bridge layer and profile scan of the embossed channels.
114
H. Bartsch de Torres et al. / Sensors and Actuators A 160 (2010) 109–115
5. Summary
Fig. 12. Cross-section through a sensor bridge.
A flow sensor has been monolithically integrated in a retention module, which is part of a modular LTCC fluidic system. The biocompatibility of the material system 951 from Du Pont was investigated and found to be poor for screen printed thick film compositions. Therefore, an encapsulated design of the sensor was chosen. To ensure the process compatibility, embossed exhaust channels are used to draw out the gases, which are generated during the sintering process. The stress control of the highly compressed sensor bridge was carried out by means of the Design of Experiments. An even bridge with embossed exhaust channels and screen printed sensor and heater elements was achieved by adapting the bridge design. The sensor was manufactured using exclusively thick film techniques and was integrated into a retention module of the fluid handling system. The temperature difference was measured as the sensor characteristic curve. For flow rates up to 80 l it can be approximated by a linear function. Utilizing this data, the concept for a thick film flow sensor working according to the boundary layer principle was proven, which allows the adequate measurement of small flow rates at low fluid temperatures.
Acknowledgements We thank the BMBF and the Thuringian Ministry of Culture for the financial support within the Centre for Innovation Competence MacroNano® .
References
Fig. 13. Retention module with connections.
Afterwards, the sensor arrangement is calibrated for each module by use of a precision analytical pump. The flow rate is forced with high accuracy and the respective temperature difference is calculated utilizing the data of the sensor element characteristics. Water at room temperature (22 ◦ C) is pumped through the sensor system at defined flow rates between 0.02 and 10 ml/min. Both sensor resistances are measured with a multi meter and their difference is applied to the former logged temperature characteristic. With this information the temperature difference DTS is calculated and the resulting graph is shown in Fig. 14. A linear characteristic is demonstrated for flow rates up to 80 l/min. The power consumption of the heater amounts to 455 mW.
Fig. 14. Sensor characteristic curve.
[1] M.R. Gongora-Rubio, M. Bariatto, A. Fontes, E. Wagner Simões, J. SantiagoAvilés, LTCC hybrid technology applied to water quality meso analytical system, Electrochemical Society Proceedings 8 (2002) 142–150. [2] L.J. Golonka, Technology and applications of low temperature cofired ceramic (LTCC) based sensors and Microsystems, Bulletin of the Polish Academy of Sciences: Technical Sciences 54 (2006) 221–231. [3] K.A. Peterson, K.D. Patel, C.K. Ho, S.B. Rohde, C.D. Nordquist, C.A. Walker, B.D. Wroblewski, M. Okandan, Novel Microsystem Applications with New Techniques in LTCC, CICMT 2005 Ceramic Interconnect and Ceramic Microsystems Technologies Proceedings. [4] G.A. Groß, T. Thelemann, S. Schneider, D. Boskovic, J.M. Köhler, Fabrication and fluidic characterization of static micromixers made of low temperature cofired ceramic (LTCC), Chemical Engineering Science 63 (2008) 2773–2784. [5] T. Thelemann, J. Müller, M. Fischer, A. Groß, LTCC based fluidic components for chemical applications, Journal of Microelectronics and Electronic Packaging 4 (2007) 167–172. [6] M. Gongora-Rubio, L.M. Solá-Laguna, P.J. Moffett, J.J. Santiago-Avilés, The utilization of low temperature co-fired ceramics (LTCC-ML) technology for meso-scale EMS, a simple thermistor based flow sensor, Sensors and Actuators 73 (1999) 215–221. [7] J. Kita, F. Rettig, R. Moos, K.-H. Drüe, H. Thust, Laser forming of LTCC ceramics for hot-plate gas sensors, in: Proceedings of XXVIII International Conference of IMAPS Poland Chapter, 2004. [8] U. Schmid, G. Krötz, D. Schmitt-Landsiedel, A volumetric flow sensor for automotive injection systems, Journal of Micromechanics and Microengineering 18 (2008) 045006. [9] O. Fiedler, Strömungs- und Durchflussmesstechnik, 1. Auflage München: R, Oldenburg Verlag (1992) 183. [10] K.W. Bonfig, Technische Durchflussmessung, 3, Vulkan-Verlag Essen (Germany), 2002, p. 186. [11] L. Rebenklau, Merkmale der Biokompatibilität von Industriellen Niedrigtemperatur-Glaskeramiken, 1st Workshop BioAVT, TU Dresden, 2003. [12] H. Bartsch de Torres, M. Hoffmann, Embossing of fluidic channels into ceramic multilayer, Proceedings of the Smart Systems Integration Conference (2007). [13] P. Espinoza-Vallejos, J. Zhong, M. Gongora-Rubio, L. Sola-Laguna, J.J. SantiagoAviles, The measurement and control of sagging in meso (intermediate scale) electromechanical LTCC structures and systems, MRS Conf. Proc., 518, 1998, pp. 73–79. [14] B. Klein, Versuchsplanung DoE, München, Oldenburg Wissenschaftsverlag, 2004, pp. 35–53. [15] H. Bartsch de Torres, A. Albrecht, M. Hoffmann, Systematic characterisation of embossing processes for LTCC-tapes, Journal of Microelectronics and Electronic Packaging 5 (2008) 142–149.
H. Bartsch de Torres et al. / Sensors and Actuators A 160 (2010) 109–115
Biographies
Heike Bartsch de Torres received her diploma degree in precision engineering from the Ilmenau University of Technology, Ilmenau, Germany, in 1994. Her research interest has developed along the lines of micromechanics and fluidic systems. After working in industry, she returned to the Ilmenau University of Technology in 2005. At the present she is a Ph.D. student at the Institute of Micro- and Nanotechnologies. Her current interests are focused on micro-structuring methods for ceramic green tapes, thick film sensor applications and microfluidics.
115
Andreas Schober received his physics diploma from the Ludwig Maximilians University (LMU) at the MPI of Quantum Optics in Munich (Professor Walther). Focusing 1990 his main interest towards biological problems, he worked for his PhD at the MPI of biophysical chemistry in the department of Professor Eigen on “strategies of evolutionary biotechnology”. Heading different groups in the fields of nanosystem integration, microfluidics and microreaction technology in industry and science, he became head of the research group “Microfluidics and Biosensors” 2006 in the ZIK MacroNano, the BMBF centre for innovation competence at the IMN and 2007 acting head of the nanotechnology group.
Martin Hoffmann worked from 1992 to 2003 as scientific staff member and chief engineer at Dortmund University, Germany at the High Frequency Institute. From 2003 to 2005 he held R&D positions in the MEMS companies HL Planartechnik, Dortmund, and Silicon Manufacturing Itzehoe, Germany. Since 2006 he is the Head of Department “Micromechanical Systems” at the Ilmenau University of Technology and since 2007 the Director of the Institute IMN MacroNano® Ilmenau University of Technology.
Christian Rensch graduated from the Ilmenau University of Technology in 2008 with a diploma degree in micromechatronics. During his studies he worked at the Institute of Micro- and Nanotechnologies (Ilmenau, Germany) on acoustic bubble detection techniques for print heads and Low Temperature Co-Fired Ceramic (LTCC) based microfluidic systems. At MicroFab Technologies Inc. (Dallas, USA) he investigated new actuation signals for microfluidic drop-on-demand systems. In 2008 he joined GE Global Research (Munich, Germany) and his activities there focus on the development of a chip-based microfluidic synthesizer for radiopharmaceuticals.
Jens Müller received his diploma degree in electrical engineering and the doctoral degree from Ilmenau University of Technology, Ilmenau, Germany, in 1992 and 1997, respectively. From 1997 to 2005, he held managing positions in development departments at MicroSystems Engineering GmbH, Berg, Germany. In 2005, he returned to Ilmenau University of Technology to establish the junior research group “Functionalised Peripherics”. In July 2008 he was assigned full professor for the department of Electronics Technology at the same university. His research interest covers functional integration for ceramic based System-in-Packages considering aspects of harsh environmental use, and high thermal/high-
Michael Fischer received the diploma degree in mechanical engineering from Ilmenau University of Technology, Ilmenau, Germany, in 1993. Currently, he is Ph.D. student at the Ilmenau University of Technology and member of the academic staff at the MacroNano® Centre for Innovation Competence focussing on the development of BIO-MEMS and MEMS-Packaging Technologies including new options for combining different material classes (silicon, ceramic, glass and polymer).
frequency requirements.