REVIEW OF SCIENTIFIC INSTRUMENTS 78, 035103 共2007兲 High performance temperature controlled UHV sample holder Hugo P. Marques,a兲 David C. Alves, Ana R. Canário, Augusto M. C. Moutinho, and Orlando M. N. D. Teodoro CEFITEC, Physics Department, New University of Lisbon, 2829-516 Caparica, Portugal 共Received 10 October 2006; accepted 29 January 2007; published online 9 March 2007兲 A requirement of many surface science studies is the capability to alter a sample temperature in a controlled mode. Sample preparation procedures such as heating or cooling ramps, high temperature spikes, fast annealing, or simply maintaining a sample at a very high, or very low, temperature are common. To address these issues, we describe the design and the construction of a multipurpose sample holder. Key points of this design are operation in an extended temperature range from liquid nitrogen 共LN2兲 temperature to ⬃1300 K, temperature control during heating and cooling, low thermal inertia with rates up to 50 K s−1 共heating兲 and −20 K s−1 共cooling兲, and small heated volume to minimize background problems in thermal desorption spectroscopy 共TDS兲 spectra. With this design the sample can be flash heated from LN2 temperature to 1300 K and cooled down again in less than 100 s. This sample holder was mounted and tested in a multitechnique apparatus and adds a large number of sample preparation procedures as well as TDS to the list of already available surface analysis techniques. © 2007 American Institute of Physics. 关DOI: 10.1063/1.2712892兴 INTRODUCTION DESIGN AND CONSTRUCTION Control of the sample temperature is essential in many surface science techniques. For example, thermal induced desorption is used to study the binding energies of adsorbates in surfaces and induced decomposition products 共samples for heterogeneous catalysis experiments also require controlled surface temperature兲1,2 and induced structural changes in thin films by rapid thermal annealing.3–5 Furthermore, since many cleaning procedures are based in inert gas sputtering, annealing at high temperatures is required to release the resulting stress in the crystalline structure and to promote surface reconstruction. These processes require that the sample be heated and cooled in a controlled mode, and often, in very short times. Rapid heating or cooling will depend on the thermal inertia of the holder; to achieve high performances the heated volume should be as small as possible. Small heating volumes are of great importance for thermal desorption spectroscopy 共TDS兲, since when the holder is at its lowest temperature 共⬃100 K兲 all the surfaces may adsorb residual or other introduced gases. If the heated volume is larger than the sample, then adsorbed gases will be released from the surrounding surfaces, thus masking the TDS signal or even leading to sample contamination. In this work we describe the design details and performance of a multipurpose sample holder, which addresses the following requirements: operation in an extended temperature range, temperature control while heating and cooling, short heating and cooling times, and a small heated volume. This holder is currently being used in a multitechnique surface analysis system, also developed in this research center.6 The prototype is an evolution of a previously reported holder.7 Its design goals were to minimize the duration of the heating and cooling processes and to keep the desorption area as close to the sample size as possible. To reduce desorption from unwanted surfaces when the sample is being heated, the rest of sample holder is kept at the cold source temperature. This improved design introduces a smaller sample mounting part, a new thermal connection, and an evolved heating assembly. Liquid nitrogen cooling was selected for the cold source since it is inexpensive, easily available in the laboratory, and most gases will adsorb at such temperatures. The upper temperature limit was set to ⬃1300 K as this is the temperature required to anneal silicon—one of the highest annealing temperatures. Heating by electron bombardment was selected over resistive heating so that the heating power may be concentrated in the smallest volume. Figure 1 shows a cross section of the sample holder. The sample has the shape of a short cylinder of 10 mm in diameter and 5 mm in height. Adapters may be used if other sample shapes or sizes need to be used. The sample is mounted in a molybdenum ringlike part. This part is grooved in order to be mounted using sapphire balls in the space between the ring and copper support part. The sapphires are mounted as in ball bearings. Sapphire provides good electrical insulation and good thermal conductivity at low temperature but poor thermal conduction at high temperatures. The sample is heated from the back side by electron bombardment. Therefore, the sample is placed on the top of a molybdenum disk to avoid sample damage and allow indirect heating. Temperature is measured using a type K thermocouple, in contact with the sample, mounted through a hole drilled in the side of the copper part to the groove. The space of a missing sapphire ball is used to connect the ther- a兲 Electronic mail: hpm@fct.unl.pt 0034-6748/2007/78共3兲/035103/3/$23.00 78, 035103-1 © 2007 American Institute of Physics Downloaded 12 Jun 2007 to 193.136.125.65. Redistribution subject to AIP license or copyright, see http://rsi.aip.org/rsi/copyright.jsp 035103-2 Rev. Sci. Instrum. 78, 035103 共2007兲 Marques et al. FIG. 1. 共Color online兲 Cross section of the sample assembly. The sample is held by compressing the sapphire spheres against the molybdenum ring. mocouple to the molybdenum ring. An extra wire is used to measure the sample current or to bias the sample in case of need. As the molybdenum ring, the external copper part has a side cut 共see Fig. 3兲. When an external bolt is fastened the copper part is pushed against the sapphire balls, tightening the molybdenum ring all around the sample. For improved thermal contact a thin gold foil 共⬇50 ␮m兲 is pressed between the sample and the molybdenum parts. The copper part is brazed to a long and thin stainless tube—the liquid nitrogen 共LN2兲 reservoir. The heater assembly is mounted below the copper part thermally and electrically insulated by a Macor™ ceramics piece. Electrons are produced through Ohmic heating of a thoriated tungsten filament. Being a design option to maintain all the parts exposed to the analysis side at a well defined ground potential, the accelerating voltage is provided by biasing the filament. Thus, during the heating operation, the filament is biased to −1000 V and the sample kept at ground potential. A special electrode was designed with the help of a SIMION simulation to confine the electrical potential in the filament side. The back of the molybdenum holder was shaped to hide the sapphire balls from the electron beam and to focus the electrons. This heater can provide up to 500 W of heating power. The emission current is controlled by the filament current. This current or the accelerating voltage may be used in a proportional-integral-derivative 共PID兲 loop to control the temperature. Care was taken to prevent electrons coming from the heater assembly to reach the upper surface, thus producing interference in the analytical techniques. The cut in the molybdenum ring has a V shape so that there is no straight line for electrons, or any other particles, to cross the holder from the heater side to the analysis side 共see Fig. 2兲. A copper washer is fastened over the copper plate to protect the sapphires from charging up and to avoid direct exposure during depositions. Additionally, since the washer remains cold even during sample heating, it will act as a trap for gases released from the sapphire balls, preventing these from reaching the sample surface. The sample surface is po- FIG. 2. Detailed view of the molybdenum ring. The V shaped cut, as shown in the side view, allows closing and opening for sample exchange. Additional partial cuts were made to assist continuous bending. sitioned slightly above the washer plane 共⬃0.5 mm兲, allowing irradiation at any incident angle, even at grazing angles. All materials were chosen to meet the requirements for UHV operation and to withstand the temperature range 共100– 1300 K兲. The holder is mounted in a manipulator, which allows linear movement in x, y, and z directions and rotational movement around the z axis. Analysis can be performed even at grazing angles. The complete holder is shown in Figs. 3共a兲 and 3共b兲. PERFORMANCE All tests were performed using a graphite sample. Initially, it takes about 10 min to cool the sample down from room temperature to 100 K. The minimum recorded temperature was 89 K after 2 h. The sample was submitted to several heating/cooling cycles using 100 W of heating power for gradually longer periods. Results are shown in Fig. 4. Heating rates in excess of 100 K s−1 were obtained at the beginning of the cycle, although lowering at higher temperatures. These results show FIG. 3. 共Color online兲共a兲 Exploded view of the complete setup 共without electrical connections兲. 共b兲 Fully assembled sample holder including the LN2 reservoir and heating assembly. Downloaded 12 Jun 2007 to 193.136.125.65. Redistribution subject to AIP license or copyright, see http://rsi.aip.org/rsi/copyright.jsp 035103-3 Rev. Sci. Instrum. 78, 035103 共2007兲 UHV sample holder FIG. 4. 共Color online兲 Heating and cooling cycles of a graphite sample with peak temperatures ranging from 375 to 1273 K. The heating power was kept constant at 100 W. that a constant heating ramp of 50 K s−1 is easily achieved as well as a cooling ramp of about 20 K s−1, considering the full temperature range. Temperature was measured on the copper piece with a second thermocouple. It remained low 共⬍150 K兲, confirming that the release of gas from surfaces other than the sample should be negligible. These performance details seem to meet or even exceed most of the requirements for thermal treatments and TDS studies. Furthermore, when comparing with some key aspects of other sample holders8–10 the achieved temperature range and cooling speed are excellent. Other holder designs do not address all of the requirements imposed here, but key aspects such as temperature range and cooling 共or heating兲 speed can be individually comparable. With this prototype the cool down times were greatly improved when compared with the previous version.6 The cool down time from 1300 to 100 K was reduced from 10 min to less than 100 s. A complete cycle of flash heating up to 1000 K now requires less than 60 s. DISCUSSIONS This sample holder meets challenging requirements in surface science: extended temperature range from 100 to 1300 K, very fast heating rates of ⬃50 K s−1, fast cooling times, small heated volume for reduced thermal inertia, and a well defined region for gas desorption. Its performance is achieved by a novel solution to fit the sample through the sapphire spheres in a ball-bearing-like arrangement. The main drawback of this holder is the difficulty to exchange the sample. Therefore, this holder is best suited when a sample can remain in place for extended periods of time. This is the case of many fundamental studies on adsorption and thin film growth. An improved version of this holder is being planned to allow exchanging of the subassembly formed by the copper plate, the molybdenum ring, and the sample. Then the sample could be exchanged ex situ and remounted in the holder using a fast entry air lock. This holder is being used in our current experiments concerning the properties of noble metal growth on TiO2共110兲 surfaces.11 These studies are being performed at controlled temperatures using techniques such as Auger electron spectroscopy 共AES兲, x-ray photoelectron spectroscopy 共XPS兲, low energy ion spectroscopy 共LEIS兲, and work function 共WF兲. Due to the catalytic relevance of these nanostructured surfaces, gas adsorption studies as a function of the surface temperature are planned in the near future. ACKNOWLEDGMENTS The financial support of the Portuguese Foundation for Science and Technology, POCTI, and FEDER is gratefully acknowledged. J. A. Rodriguez and D. W. Goodman, Surf. Sci. Rep. 14, 1 共1991兲. C. R. Henry, Surf. Sci. Rep. 31, 235 共1998兲. 3 E. K. F. Dang and R. J. Gooding, Phys. Rev. Lett. 74, 3848 共1995兲. 4 Y. Shao, M. L. Yan, and D. J. Sellmyer, J. Appl. Phys. 93, 8152 共2003兲. 5 M. Flores, V. Fuenzalida, and P. Haberle, Phys. Status Solidi A 202, 1959 共2005兲. 6 O. M. N. D. Teodoro, J. A. M. C. Silva, and A. M. C. Moutinho, Vacuum 46, 1205 共1995兲. 7 C. M. M. Leitão, T. A. Gasche, G. Bonfait, O. M. N. D. Teodoro, and A. M. C. Moutinho, Vacuum 52, 23 共1999兲. 8 U. Leist, A. Winkler, J. Bussow, and K. Al-Shamery, Rev. Sci. Instrum. 74, 4772 共2003兲. 9 J. S. G. Taylor and C. Norris, Rev. Sci. Instrum. 68, 3256 共1997兲. 10 H. Tajiri, K. Sumitani, S. Nakatani, T. Takahashi, K. Akimoto, H. Sugiyama, X. Zhang, and H. Kawata, Appl. Surf. Sci. 237, 641 共2004兲. 11 H. P. Marques, A. R. Canário, A. M. C. Moutinho, and O. M. N. D. Teodoro, J. Phys.: Conf. Ser. 共accepted for publication兲. 1 2 Downloaded 12 Jun 2007 to 193.136.125.65. Redistribution subject to AIP license or copyright, see http://rsi.aip.org/rsi/copyright.jsp