New Thermoelectric Components Using Microsystem ... - IEEE Xplore

414

JOURNAL OF MICROELECTROMECHANICAL SYSTEMS, VOL. 13, NO. 3, JUNE 2004

New Thermoelectric Components Using Microsystem Technologies Harald Böttner, Joachim Nurnus, Alexander Gavrikov, Gerd Kühner, Martin Jägle, Christa Künzel, Dietmar Eberhard, Gerd Plescher, Axel Schubert, and Karl-Heinz Schlereth

Abstract—This paper describes the first thermoelectric devices based on the V–VI-compounds Bi2 Te3 and (Bi Sb)2 Te3 which can be manufactured by means of regular thin film technology in combination with microsystem technology. Fabrication concept, material deposition for some 10- m-thick layers and the properties of the deposited thermoelectric materials will be reported. First device properties for Peltier-coolers and thermogenerators will be shown as well as investigations on long term and cycling stability. Data on metal/semiconductor contact resistance were extracted form device data. Device characteristics like response time for a Peltier-cooler and power output for a thermogenerator will be compared to commercial devices. [859] Index Terms—Bi2Te3-compounds, semiconductor-device design, thermoelectric devices, thin-film structure, thin-film morphology.

I. INTRODUCTION

M

ICRO-PELTIER coolers with efficient cooling capacity, small area—down to parts of a millimeter—and short response time, are in high demand on the telecommunication markets. Also, microthermoelectric generators can be used in a lot of small low-power devices such as hearing aids or wrist watches. This has been shown recently by Seiko and Citizen with their commercialized thermoelectrically driven low-power wrist watches. Nevertheless only few approaches to manufacture thermoelectric devices in small dimensions, which makes them suited especially for applications mentioned above, are known up to now. Fleurial [1], [2] showed by design calculation the potential of these devices and figured out the main technological problems inherent like the necessity of high thermal conductive substrates and the impact of low ohmic contacts between electrodes and semiconductor material. Focussing on applications at room temperature he based his development on V–VI-compounds. The thermoelectric compounds were grown using electrochemical methods. The vertical structures were defined during the electrochemical growth by several 10- -thick photolithographic masks. Up to now no device properties were reported. Manuscript received May 6, 2002; revised January 13, 2003. Subject Editor G. B. Hocker. H. Böttner, J. Nurnus, A. Gavrikov, G. Kühner, M. Jägle, C. Künzel, D. Eberhard, and G. Plescher are with the Fraunhofer Institut Physikalische Messtechnik, Freiburg D-79110, Germany (e-mail: harald.boettner@ ipm.fhg.de). A. Schubert is with the Infineon Technologies AG, München D-81541, Germany. K. H. Schlereth is with the Osram Opto Semiconductors, Regensburg D-93049, Germany. Digital Object Identifier 10.1109/JMEMS.2004.828740

Fan [3], [4].developed microcoolers based on SiGe/Si superlattices. He reached about 2–3 K net cooling at room temperdevice temperature for ature and up to about 10 K at 200 p-type and n-type coolers. Venkatasubramanian reported on net cooling of 32 K at around room temperature for a p-type superlattice device. The [5]. The thermofigure of merit ZT was determined to [7] with electric figure of merit ZT is defined by Seebeck coefficient, electrical conductivity, ( power factor), thermal conductivity. Unfortunately no technological information was given for the preparation of this semi-Peltier device. Thus up to now no complete microthermoelectric thin-film device has been presented which could reach the requirements such as laser-cooling and temperature control in telecommunication applications: a net cooling of more than 40 K with some 100 mW cooling power at temperatures of several 10 . We decided to design microstructured thermoelectric devices with the objective of meeting both the telecommunication requirements as well as possible applications for highly effective thermogenerators. Thus it was necessary to develop a completely new technological route to fabricate these devices with techniques taken from typical microelectronic fabrication and microsystem technology. Here we will report on the fabrication principles, the preparation of the p- and n-type thermoelectric material, their structural and thermoelectric properties as well as material properties, some details about the device fabrication including information about contact resistances, and about the preliminary device properties. II. EXPERIMENT A. Fabrication Principles The fabrication technology is based on a two-wafer process that finally leads to the devices by chip-to-chip, chip-to-wafer or wafer-to-wafer soldering. The initial wafer size is , but all processes are developed to be easily transferable to other common sizes like . This “two-wafer” fabrication principle requires suitable thickness matched p- and n- material deposition, sticking over-growth of the thermoelectric material over their contact electrodes as well as etching techniques for the several 10- -thick thermoelectric layers. Fig. 1 depicts a schematic drawing of the two wafer process in the left part and an also schematic drawing of the resulting device together with the application for telecommunication purposes in the right part. To estimate the possible performance of such a device like Fig. 1 we assume for the thermoelectric

1057-7157/04$20.00 © 2004 IEEE

BÖTTNER et al.: NEW THERMOELECTRIC COMPONENTS USING MICROSYSTEM TECHNOLOGIES

415

Fig. 1. Schematic drawing of the developed two wafer (I,II) concept (left) and a schematic drawing of an application of the investigated Micro-Peltier coolers as a possible telecommunication device. Suitable dimension are 550 m 3 650 m for the cold- (laser) side, a wafer thickness of 200 m and a height of the thermoelectric material (TE-material) of 20 m.

performance of n- and p-type material a power factor of 40 , which corresponds to the data of common comrelated materials and for the contact resistance mercial . All other design details are kept under proprietary. The following performance may be achieved for an optimum design of the electrode structure and thickness, the thickness of thermoelectric materials, the coverage of the thermocouples on the two substrates: maximum temperature difference results in a net cooling of as much as 70 K; at a reduced cooling of 60 K a cooling power of 0.3 W should be possible. It should be menwere tioned, that in this case a contact resistance of assumed. This is in accordance to the remarks for the devices described in [1]–[4]. Venkatasubramanian reported in [5] on conbut without any tact resistances, achievable down to details about contact layers and preparation techniques. The asis nevertheless much lower than sumed resistance of reported for standard ohmic contacts on n- and p-type V–VI-material [6]. B. Growth of Thermoelectric Material and pmaterials were grown Both nby cosputtering from 99.995% element targets (Bi, Sb, Te) onto prestructured electrodes. In spite of the well known fact of the superior thermoelectric properties of the alloys, these alloys were not grown due to delivery problems of Se-target suppliers. -passivated -Si The electrodes were structured on wafers. Calculations show, that due to the chosen thinness of the substrate and insulating material, the thermal resistance of the substrate is sufficiently small. Adhesion problems arise for the needed layer thickness due to huge differences in the thermal expansion coefficients. Also performance problems arise from the known anisotropy of the V–VI-compounds. To illustrate this, Table I shows important material properties of perpendicular to the -axis and for the ratio nto the -axis (data taken from perpendicular/parallel [7]). It is obvious, that a material growth with the -axis perpendicular to the heat flux (see Fig. 1) would result in the best device performance, if exclusively these material parameters are taken into account.

TABLE I SOME MATERIAL PROPERTIES OF N-Bi Te (S : SEEBECK-COEFFICIENT, : SPECIFIC CONDUCTIVITY, : THERMAL CONDUCTIVITY, : LINEAR THERMAL EXPANSION COEFFICIENT, Z : FIGURE OF MERIT Z = S , P F : POWER FACTOR P F = S , DATA TAKEN FROM [7]

III. RESULTS AND DISCUSSION A. Structural Properties of Thermoelectric Material Both materials were successfully grown up to layer thickness . The growth rates are similar for both materials of about 20 and in the range of 5 . This is similar to the rate for [8]. SEM analysis was used to electrochemically grown visualize roughly the growth direction and overgrowth quality. EDX analysis was taken to verify the metal/chalcogen ratio for at%) and to control the Bi/Sb all grown layers (accuracy ratio of the p-type material, X-ray was taken for phase analysis. The stoichiometry was adjusted by the individual power applied to the sputter targets during deposition. As reported by Nurnus [9], [10] the properties of MBE-grown epitaxial -layers depend quite sensitive on the growth temperature. In particular the Te-sticking coefficient changes drastically with temperature. The quality control was performed by the determination of the metal/chalcogen ratio via EDX-analysis and Seebeck coefficient measurements After deposition an for some hours is carried additional annealing around 300 out in order to further optimize the material properties. Thermoelectric layers were sputtered on heated (hot sputtered) as well as on nonheated (cold sputtered) substrates. Fig. 2 shows on the left a possible electrode design (lightgray) together with the design of the two types of thermoelectric materials (oppositely hatched bars) and in its right part a cross layer section SEM picture of the overgrowth of a ponto a contact electrode. The electrode materials were deduced from metals used in [6].

416


Fig. 2. Left part: a possible electrode design (light-gray) together with the design of the two types of thermoelectric materials (dark and light bars). Right part: a cross section SEM picture of the overgrowth of a 5 thick p- ; layer over a contact electrode.

m

(Bi Sb) Te

2 22

(Bi; Sb) Te layer. The peak identification (Bi; Sb) Te .

Fig. 4. = -analysis for a pproves untextured single phase p-

2 22

Bi Te Bi Te

Fig. 3. = -analysis for a nuntextured single phase n.

layer. The peak identification proves

A random nontextured growth can be seen in the SEM picture. Inside the left inserted circle a direct crack-free contact with the thermoelectric material can be seen. The cracks inside the right inserted circle may be caused by the rupture due to the SEM sample preparation. The structural performance was also analyzed by X-ray -analysis). Figs. 3 and 4 show the diffraction pattern ( results of the X-ray analysis for both materials. . Fig. 4 Fig. 3 proves the existence of stoichiometric nalso proves the growth of the desired p-compound within the detection limit of this analysis. Additional peaks can be assigned to insulating and contact layers. EDX analysis was performed to verify the metal/chalcogen ratio and, in case of p-type materials, to control the Bi/Sb ratio at%). (accuracy B. Thermoelectric Properties The thermoelectric properties for n- and p- type material were determined by van der Pauw Hall- (specific conductivity) and Seebeck-coefficient measurement at room temperature. The power factors were calculated using these data, they are shown in Figs. 5 and 6, respectively. The increase of the Seebeck-coefficient from “cold sputtered” compound formation to “hot sputtered” shows, that the only takes place when using a heated substrate. Furthermore it is obvious, that as expected the maximum of the Seebeck-coefficients appear near the stoichiometric composition and has a

Fig. 5. (a) Seebeck coefficients versus Te- content for the n-type (Bi,Te) materials. (b) Power factor versus Te content for different n-(Bi,Te) materials. Dashed lines are as a guide for eyes.

significant smaller decrease with increasing Te-content as compared to decreasing Te-content, aside of the existence range of this compound. The resulting power factor is plotted in Fig. 5(b). In Fig. 5(a) and (b) it is denoted with “after annealing,” that a post growth annealing is performed to enhance the material performance. Due to peculiar growth conditions to stabilize a certain native doping of Te we expect an improvement of the uniformity of the thermoelectric material by this annealing process.


417

Fig. 7. Nearly completely dry etched wafer, the insert shows a cleaved bar of also etched thermoelectric material.

Fig. 6. (a) Seebeck coefficient versus Te content for different p-(Bi,Sb,Te) materials. (b) Power factor versus Te content for different p-(Bi,Sb,Te) materials. Dashed lines are inserted as a guide for eyes.

Also a lowering of the possible grain boundary scattering is expected. At least, as indicated in Fig. 5(a) and (b), an improvement of the Seebeck-coefficient or power factor by post growth annealing is achieved. In contrary to the S(eebeck-coefficient) vs. Te-content for the n-type material Fig. 5(a), from Fig. 6(a) a significant weaker dependence in the p-material system can be seen. The enhancement in the power factor, Fig. 6(b), results here from an improvement of the specific electrical conductivity (not shown here). The highest power factor we achieved up to now were 15.7 for n-material and 25.3 for p-material. C. Device Fabrication wafer covered with either Devices were processed from n-type or p-type material. To define the geometry of the n- and p-type material, both materials were etched using reactive dry etching using photoresist as an etching mask. Before etching the solder material was deposited and structured on top of the thermoelectric material. Fig. 7 shows a SEM picture of a nearly completely etched wafer. The step halfway down the thermoelectric bar, Fig. 7, is due to an interruption of the dry etching process. A cleaved bar of etched thermoelectric material is shown in the insert. It can be seen, that the contact pads for a pn-pair are insulated . According to a twoby small gaps in the range of some wafer concept, each pad is occupied in this stage with only one

Fig. 8. Soldered device: the upper part shows the accuracy of positioning, the lower part a soldered device with 3 n-p-pairs; the insert shows the adjustment of n- to p-part of the device just before soldering.

material type. Furthermore it is obvious, that an etching angle steeper than was achieved. From the etched n- and p-wafers, n-chips and p-chips were cut. The p-single chips were aligned to the complementary n-side and then soldered together. A soldered device is shown in Fig. 8. The upper part of Fig. 8 shows the accuracy of positioning. In the lower part, a soldered device with 3 pn-pairs is shown. In the insert, the adjustment of n- to p-part of a device just before soldering can be seen. To verify the technology, e.g., the critical parameter contact resistance and also our modeling, net cooling and the device resistance was measured for different device geometries. Fig. 9 shows net cooling versus current for a micro-Peltier device with 3 p/n-junctions. A net cooling of about 11 K was achieved using

418

Fig. 9. Dependence of the achieved net cooling of a 3-n-p-pair device at 60 on the applied current I .


C

Fig. 10.

Device stability under constant current during about one week.

TABLE II NET COOLING AND DEVICE RESISTANCE FOR DIFFERENT DEVICE GEOMETRY TYPES (I, II, III). UNDERLINED ARE THOSE DATA FOR NET COOLING AND DEVICE RESISTANCES WHICH WERE MEASURED WITH THE SAME DEVICE

a current of . Table II summarizes some measurements for three different device geometries differing in width, length and distance of the thermoelectric bars. The height was . measured to 20–21 One very important but unknown device parameter is the contact resistance after processing the complete device. For an estimation we measured the overall resistance of the processed devices and compared this data with those we calculated ones , using the design assuming a contact resistance of parameters (electrode materials and their dimensions), the electrical properties of those p- and n-layers, which were used for device fabrication. Despite that the resistance data scatters by a factor of about two, the measured data are quite close to the calculated ones. According to our calculations, nearly no cooling should be observed with a contact resistance in the range of . Based on Table II, we conclude, that we achieved or even better for these contact resistances around first devices. The calculations of maximum net cooling based on the same set of data as used for the device resistance including the influence of the substrate. The formula were taken from [7] and adapted to our Peltier designs. As compared to the predicted net cooling of 15.5 K (geometry II) we obtained about 70% (10.9 K) of the expected net cooling. One important parameter for the reliability of these devices is their stability under current. Therefore a device was driven under a continuous current of 250 mA for about six days. Fig. 10

Fig. 11. Thermovoltage vs. time for different currents I . The voltage rise is directly related to the temperature difference across the thermoelectric material. The specified T values were measured over the whole device.

1

supports strongly the assumption, that these devices will keep reliable also for a long term continuous operation. The measurements according to Fig. 10 were performed keeping the device continuously under current without measuring the net cooling. The net cooling was determined at the indicated points (time) switching the current off/on and measuring the net cooling as noticed. The net cooling was stable over the whole period of 6 days. Another important hint for the long term stability is the constancy after cycling tests. and a After 200 cycles with a maximum temperature of 55 after an overall cycle time of minimum temperature of about 90 h, no significant deterioration was observed. Another important feature for the intended applications is a short rise time for those microdevices. The rise time was evaluated by a Harman-like measurement [11]. Fig. 11 shows the rise of the thermovoltage after current impact in 20 ms. This is much shorter compared to the rise time of about 8 sec measured for the smallest commercial Peltier device. This is shown clearly by the comparison of the measured time constants of a Micro-Peltier and a small commercial device- seven n/p - for a net cooling of about 6 K in pairs, area of 4 4 both cases (Fig. 12). The measured rise time is in accordance


419

TABLE III COMPARISON OF MINIATURIZED THERMOELECTRIC GENERATORS. ALL VALUES ARE FOR THE CASE OF A MATCHED LOAD RESISTANCE AND A TEMPERATURE DIFFERENCE OF 5

C

Fig. 12. Comparison of the time constants of a micro Peltier cooler and the . smallest commercially available device for T of about 6

1

C

factor of 10 smaller and possess only 12 thermocouples the power output of a single thermocouple is a factor of 5 higher than reported by Seiko and even more the 50 times larger than the values reported by D.T.S. Since all devices use the same active thermoelectric materials this result once more demonstrates the good material deposition, processing and soldering techniques developed for the production of miniaturized MicroPelt thermoelectric generators and MicroPelt coolers. IV. CONCLUSION

Fig. 13. Dependence of the power output of a miniaturized thermoelectric generator consisting of 12 p-n-junctions for different load resistances R as a function of the external temperature difference.

to calculated data, from which about 20 ms were expected. It is obvious, that the huge reduction of the rise time is caused by the tiny thermal mass of this microdevice. The technology for the fabrication of thick film MicroPelt’ier coolers as described above of course also can also be applied for the fabrication of miniaturized thermoelectric generators. The fabricated generators have the same overall dimensions as the Peltier coolers shown in Fig. 8, the area of the single p- and n-type elements was reduced in order to increase the number of elements per unit area. In Fig. 13 the measured electric power output of a miniaturized thermoelectric generator is plotted for different external load resistances against the temperature difference between the hot and the cold side. In the case of a matched load resistance we measured a maximum power output at a temperature difference of 5 . Conservatively of 800 ) assuming the overall device geometry (1400 as active area results in a minimum power density of 0.60 . The comparison of the results mentioned above with the published performance data of D.T.S. [12] and Seiko [13] is shown in Table III. Although the MicroPelt thermogenerators are a

It was shown, that a successful production of micro-Peltier coolers and micro-thermogenerators is feasible by means of regular thin film technology in combination with micro-system technology. Main features of the manufacturing on standard silicon wafers are: the control of stoichiometric composition of the sputtered n- and p-type thermoelectric materials of the bis, the muth-chalcogenide family in thicknesses of a few 10 vertical structuring of all functional layers, and the reasonable low ohmic contact resistance. The net cooling of 11 K around room temperature, which has already been achieved, and the reliability of the components prove the chosen technology route. Due to their small overall dimensions these microdevices have a quick rise time down to a few msec. Also their small dimensions open the possibility for new designs of optoelectronic devices. The successful production of these devices opens the way to new prototyping with optimized steps in the processes. Further progress may be expected, e.g., by switching to ndue to a better control of native and/or foreign imperfections. In summary, we are convinced that micro-Peltier coolers and microthermogenerators can be manufactured cost effectively in serial production. ACKNOWLEDGMENT The authors would like to thank B. Acklin for encouraging ideas and fruitful discussions and also H. Beyer for critical reading the manuscript as well as N. Herres (FhG-IAF) for performing the X-ray measurements. REFERENCES [1] J. P. Fleurial, G. J. Snyder, J. A. Herman, P. H. Giauque, W. M. Phillips, M. A. Ryan, P. Shakkottai, E. A. Kolawa, and M. A. Nicolet, “Thick-film thermoelectric microdevices,” in Proc. 18th Int. Conf. on Thermoelectrics, Baltimore, MD, Aug.–Sept. 29–02, 1999, pp. 294–300. [2] J. P. Fleurial, A. Borshchevski, W. Phillips, E. Kolawa, G. Snyder, T. Caillat, T. Kascich, and P. Mueller, “Microfabricated Thermoelectric Power-Generation Devices,” U.S. Pat. 6 388 185 B1 14.05.2002, International publication number WO00/08693, 17.02.2000.

420


[3] X. Fan and G. Robinson, Physics of Low-Dimensional Structures. Moskow: VSV Co Ltd., 2000, vol. 5–6. [4] G. Zeng, A. Shakouri, C. La Bounty, G. Robinson, E. Croke, P. Abrahams, X. Fan, H. Reese, and J. E. Bowers, “SiGe micro-cooler,” Electron. Lett., vol. 35, no. 24, pp. 2146–2147, Nov. 1999. [5] R. Venkatasubramanian, E. Siivola, T. Colpitts, and B. O’Quinn, “Thin-film thermoelectric devices with high room-temperature figures of merit,” Nature, vol. 413, no. 6856, pp. 597–602, 2001. [6] U. Birkholz, R. Fettig, and J. Rosenzweig, “Fast semiconductor thermoelectric devices,” Sens. Actuators, vol. 12, no. 2, pp. 179–184, Aug.–Sept. 1987. [7] D. M. Rowe and C. M. Bhandari, CRC Handbook of Thermoelectrics. Boca Raton, FL: CRC, 1995. [8] P. Magri, C. Boulanger, and J. M. Lecuire, “Synthesis, properties and performances of electrodeposited bismuth telluride films,” J. Mater. Chem., vol. 6, pp. 773–773, 1996. [9] J. Nurnus, H. Böttner, H. Beyer, and A. Lambrecht, “Epitaxial bismuth telluride layers grown on (111) barium fluoride substrates suitable for MQW-growth,” in Proc. 18th Int. Conf. on Thermoelectrics, Baltimore, MD, Aug.–Sept. 29–02, 1999, pp. 696–699. , “Layered (IV–VI)–(V–VI)-materials for low dimensional thermo[10] electric structures,” in Proc. 18th International Conference on Thermoelectrics, Baltimore, MD, Aug.–Sept. 29–02, 1999, pp. 704–708. [11] T. C. Harman, J. H. Cahn, and J. Logan, “Measurement of thermal conductivity by utilization of the Peltier-effect,” J. Appl. Phys., vol. 30, no. 9, pp. 1351–1359, 1959. [12] M. Stordeur and I. Stark, “Low power thermoelectric generator-self-sufficient energy supply for micro systems,” in Proc. 16th Int. Conf. on Thermoelectrics, Dresden, Germany, Aug. 26–29, 1997, pp. 575–577. [13] Data Taken [Online]. Available: www.zts.com

Harald Böttner received the Dr.rer.nat. from WWUMünster, Germany, in 1977. In 1978, he joined the Fraunhofer Society working from 1978 to 1980 at the Fraunhofer-Institut für Silicatforschung, Würzburg, Germany, in the field of ORMOCERE. Since 1980, he has been with the Fraunhofer-Institut für Physikalische Meßtechnik, Freiburg, Germany. Until 1993, he developed IV–VI semiconductor infrared lasers. Currently, he works on metal–oxide–semiconductor gas sensors as well as on the field of thermoelectrics.

Joachim Nurnus received the Dr.rer.nat. from Albert-Ludwigs Universität Freiburg, Germany, in 2001. In 1997, he joined the Fraunhofer Society, working since then at the Fraunhofer-Institut für Physikalische Messtechnik Freiburg, Germany, in the fields of IV–VI- and V–VI-based narrow-gap semiconductors in particular quantum wells and superlattices. Currently, he works on V–VI thin-film thermoelectric materials and devices as well as in the field of IV–VI-infrared emitters.

Alexander Gavrikov received the Diploma degree in electrical engineering from the Fachhochschule Offenburg, Germany, in 1998. He then joined the Fraunhofer-Institut für Physikalische Messtechnik Freiburg, Germany, working in the field of thermoelectrics.

Gerd Kühner received the Diploma degree in physical techniques from the Fachhochschule Heilbronn, Germany, in 1988. In 1988, he joined the Fraunhofer-Institut für Physikalische Messtechnik Freiburg, Germany. Currently, he is engaged in development of thin-film technology in particular semiconductor gas sensors and thermoelectric devices.

Martin Jägle received the Diploma degree in physics from the Albert-Ludwigs Universität Freiburg, Germany, in 1995. Since 1997, he has been working at the Fraunhofer-Institut für Physikalische Messtechnik Freiburg, Germany. He is involved in thermoelectrics, micromechanics and sensors for gases and liquids.

Christa Künzel, photograph and biography not available at the time of publication.

Dietmar Eberhard received the Diploma degree in electrical and electronics engineering from the University of Bochum, Germany, in 1981 and the Dr.-Ing. degree from the University of Dortmund, Germany, in 1987. In 1986, he joined the Fraunhofer-Institut für Physikalische Messtechnik Freiburg, Germany, where he first was engaged in the fields of integrated-optics, thin-film techniques, thermoelectrics, and microsystem engineering. Currently, he is working on laser imaging systems and space science. His current interests also include optical, mechanical, and electrical simulations.

Gerd Plescher received the Diploma degree in physical techniques from the Westsächsische Hochschule Zwickau, Germany, in 1998. Since then, he has been with the Fraunhofer-Institut für Physikalische Messtechnik Freiburg, Germany, where he has been engaged in the development of laser spectroscopy, gas sensor systems, and thermoelectric devices.

Axel Schubert received the Diploma degree in physics from the Technische Universität München, Germany, in 1985. Since then, he first worked at the semiconductor division of Siemens and is now with Infineon. His main topics are silicon applications for optoelectronics and functional layers on silicon.

Karl-Heinz Schlereth received the Ph.D. and Diploma degrees from University of Würzburg in 1987 and 1990, respectively. He has 15 years experience in various technical and management positions in the optoelectronic semiconductor business at Fraunhofer Institute for Physical Measurement Technique IPM, Freiburg, Germany, Alcatel-SEL Research Center, Stuttgart, Germany, Alcatel-Optronics at Villarceau, France, and Siemens/OSRAM Opto Semiconductors. He currently holds the position of a Director of Epitaxy Manufacturing at OSRAM Opto Semiconductors, Regensburg, Germany.