CARBON-NANOTUBE FIELD EMISSION X-RAY TUBE FOR SPACE EXPLORATION XRD/XRF INSTRUMENT.

Transcription

1 Copyright JCPDS - International Centre for Diffraction Data 2004, Advances in X-ray Analysis, Volume CARBON-NANOTUBE FIELD EMISSION X-RAY TUBE FOR SPACE EXPLORATION XRD/XRF INSTRUMENT. P. Sarrazin 1, D. Blake 1, L. Delzeit 1, M. Meyyappan 1, B. Boyer 2, S. Snyder 2, B. Espinosa 3 1 NASA Ames Research Center, Moffett Field CA Oxford Instruments X-Ray Technology Inc., 275 Technology Circle, Scotts Valley, CA Microwave Power Technology, 1280 Theresa Avenue, Campbell, CA ABSTRACT The deployment of in-situ miniature XRD instruments for planetary exploration requires small focus bright X-ray sources that are physically small, lightweight, robust, and power efficient. Existing miniature X-ray tubes are limited by inefficient thermionic electron sources. We have investigated the potential of carbon nanotube (CNT) field emitters for improving the efficiency and durability of miniature X-ray tubes. CNT field emitters were fabricated using a thermal CVD growth process. High, sustainable current densities of over 1A/cm 2 were routinely measured from these emitters. CNT cathodes implemented in miniature X-ray tubes have shown improved efficiency and robustness. Focusing optics are being designed to yield the small focus required for XRD. The resulting miniature microfocused X-ray tube will be part of the CheMin XRD/XRF instrument, proposed for the Mars 2009 Mars Science Laboratory mission. INTRODUCTION CheMin is a miniature XRD/XRF instrument designed at NASA Ames Research Center to perform remote mineralogical analyses on solid bodies of the solar system such as Mars, Venus, Europa, the Moon, asteroids and Kuiper belt objects [1,2]. CheMin is based on a 2-dimensional CCD detector capable of both spatial and energy resolution of X-ray photons. The instrument offers the unique capability of simultaneous XRD and XRF characterization with a single detector in a package expected to weigh under 1 kg with a volume smaller than 1 liter. Prototype instruments built at NASA-ARC have demonstrated the capabilities of the CheMin concept [2,3]. XRD and XRF data were simultaneously acquired from a variety of minerals and rocks. Complex mixtures could be quantitatively analyzed. This development has shown that an XRD/XRF instrument can be miniaturized to the point that it meets the size and mass constraints of landed spacecraft instruments. While most technologies required for the deployment of CheMin in space are now available, a challenge remains with regard to the X-ray source. Radioactive sources are not bright enough for XRD and currently available miniature X-ray tubes are limited by one of their most critical components: the electron source. Thermionic emitters are not efficient enough for low power X- ray tubes due to the waste heat radiated from the filament. An alternate means of generating electrons is field emission, which presents many advantages over thermionic emission. As field emitters do not require heat to generate electrons, they are

2 This document was presented at the Denver X-ray Conference (DXC) on Applications of X-ray Analysis. Sponsored by the International Centre for Diffraction Data (ICDD). This document is provided by ICDD in cooperation with the authors and presenters of the DXC for the express purpose of educating the scientific community. All copyrights for the document are retained by ICDD. Usage is restricted for the purposes of education and scientific research. DXC Website ICDD Website -

3 Copyright JCPDS - International Centre for Diffraction Data 2004, Advances in X-ray Analysis, Volume more energy efficient and less likely to outgas species that would deteriorate the device and contaminate the target. The cold nature of the emission also prevents thermal drift of the cathode which allows better and more stable electron focusing. Field emitters can be rapidly switched on and off during the operation of the tube, eliminating the need for a mechanical shutter. Most prior attempts to use field emission in X-ray tubes have been unsuccessful because the emitters were rapidly destroyed by the arcing and cation sputtering that inevitably occur in an X-ray tube. Carbon nanotubes have been shown to be extremely good field emitters and are among the most robust materials in terms of their mechanical, thermal and chemical properties (in non oxidizing environments). Successful applications of carbon nanotube emitters in X-ray tubes have been recently reported [4, 5]. The work presented herein is the first step toward development of a miniature high efficiency microfocused X-ray tube for planetary exploration. PRINCIPLE OF FIELD EMISSION Field emission consists in the extraction of electrons from a conducting solid by an electric field. Unlike thermionic emission, no heat is required for obtaining field emission. Very high electric fields are however required for the electrons to tunnel through the surface potential barrier. When the solid is shaped as a tip, the electric field lines are concentrated around the tip and the local electric field is enhanced. This geometrical enhancement of the electric field is used in field emitters to allow extraction of electrons from sharp tips at relatively low macroscopic electric fields. The Fowler-Nordheim model used to describe field emission is illustrated on Figure 1. Maximizing the current output for a given applied voltage requires a material with a low workfunction that is shaped as sharp as possible to offer the highest field enhancement factor. From a technical standpoint, low voltage field emitters are often presented in the form of micromachined array of Si or Mo of tips, usually gated to apply the extraction field [6,7]. The emission characteristics of such emitters are strongly dependent on the tip radii, and preservation of the emission characteristics requires the tip sharpness remain unaltered. Figure 1; Fowler Nordheim model of field emission from a metallic tip; typical Fowler-Nordheim plot in which field emission appears as a straight line with a negative slope. I emitted current (A), V applied voltage (V), Φ workfunction of the material (V), β field enhancement factor (non-dimensional). Until recently, field emitters had not found practical application in X-ray tubes because the emitters were rapidly damaged by arcing or cation sputtering during operation. The potential benefits of field emission for X-ray tubes can only be achieved with field emitters that would be mechanically, chemically and thermally very robust.

4 Copyright JCPDS - International Centre for Diffraction Data 2004, Advances in X-ray Analysis, Volume CARBON NANOTUBE FIELD EMITTERS Carbon nanotubes are among the sharpest and strongest materials known. As a results, they have proven to be very interesting materials for field emission cathodes [8-10]. Various designs based on these materials have shown high current densities and low turn on voltages. Carbon nanotubes are mechanically, chemically and electrically very robust. It has been shown that operating a carbon nanotube emitter under poor vacuum conditions or subsequent to electrical arcing between emitter and the gate doesn t destroy the emitter but only lowers its performance [11]. Due to the very high aspect ratio of nanotubes, the field enhancement reduction consequent to physical damage of the tip is much less likely to destroy the emitter than with micromachined tips, as is illustrated on Figure 2. Figure 2; Comparison of the damage by arcing of a micromachined tip and a nanotube ; the field enhancement at the nanotube tip is not as dramatically reduced if the emitter is physically damaged. The mechanical strength, high current density, resistance to arcing, and lower vacuum requirements of carbon nanotubes make these materials particularly interesting candidates for field emission electron sources in miniature low-power X-ray tubes. The robustness and long life expectancy required for spacecraft instruments could be achieved with this technology. Cathode fabrication Much work has been done towards the growth of carbon nanotubes [12-15]. Multiwalled carbon nanotubes (MWCNT) cathodes were fabricated using a thermal chemical vapor deposition process developed at NASA-ARC [12]. This process consists in the growth of a MWCNT film on a substrate on which metal catalysts have been deposited. Ion beam sputtering was used to deposit a 50Å thick underlayer of Al followed by a 100Å thick layer of Fe, the active catalyst for MWCNT growth. The growth of MWCNTs has been studied on a variety of substrates such as Si wafers, fused quartz, mica, highly oriented pyrolytic graphite (HOPG) and copper. Molybdenum substrates were chosen for the X-ray tube cathode application. Mo substrates were machined to shape, polished to 600 grid SiC abrasive to obtain appropriate surface roughness and vacuum annealed prior to catalyst deposition. The shape of the film is controlled during the catalyst deposition using a mask. Cathodes were grown as circular film with diameters varying from 2mm to 75µm. Catalyst-covered substrates are inserted into a CVD reactor consisting of a quartz tube held within a high temperature furnace [13,14]. Argon is used to purge the reactor while the furnace is heated. At 750 C, the gas flow is switched to 1000 sccm of ethylene (99.999% pure) for 10 minutes. The gas flow is then switched back to argon and the furnace is cooled.

5 Copyright JCPDS - International Centre for Diffraction Data 2004, Advances in X-ray Analysis, Volume The CNT films were characterized by electron microscopy. Transmission Electron Microscopy (TEM) was used to characterize the internal structure of the nanotubes which were multiwalled with a good continuity of the graphene layers along the tube walls as shown in Figure 3. Scanning Electron Microscopy (SEM) was used to verify that nanotubes were grown and to assess their diameter, length and density, as well as the microscopic and macroscopic structures of the film. Figure 4.a-b show the typical microstructure of the films: randomly oriented tubes, several µm in length and presenting various curvatures. The CNTs run along the surface of the substrate, forming hoops and in some cases pointing tips away from the surface. The nanotubes are well attached to the substrate at their base (the catalyst particles). Figure 4.c-d show two different sizes of CNT film grown by using appropriately sized masks during catalyst deposition. Figure 3; Transmission electron microscopy image of CVD grown MWCNT. Figure 4; Scanning electron microscope images of CNT cathodes grown on Mo substrate. Films density is controlled by the catalyst formulation and growth conditions (a: high density, b: low density), cathode diameter is controlled by masking during catalyst deposition (c: 2mm diameter, d: 200µm diameter) Cathode characterization The field emission characteristics of the CNT cathodes were evaluated using several instruments specifically developed for this project. In an apparatus built at NASA ARC, a single CNT cathode is mounted in a holder, facing a flat anode with adjustable gap (Figure 5). The test assembly is installed in an ultra high vacuum (UHV) chamber pumped by a dry turbo-molecular pump backed with a dry diaphragm pump. All field emission measurements are performed in the 10-8 mb range. A Keithley 237 sourcemeasure unit is used for application of the high voltage (up to 1100V) and measurement of the emission current with pa sensitivity. Voltage cycles and data collection are controlled by a desktop computer using Testpoint software. The cathode - anode gap is adjusted with micrometric precision and controlled optically through view-ports. Typical gap values during field emission measurements were 200, 150 and 100 µm depending on the electric field requirement.

6 Copyright JCPDS - International Centre for Diffraction Data 2004, Advances in X-ray Analysis, Volume MWCNT films having a wide range of density and structure were characterized. Dense films showed high turn-on voltages due to field screening from high densities of CNTs (high emission site density but poor field enhancement), while very low-density films showed very low turn on voltage but poor current density (good field enhancement but very low site density). A range of moderately low film densities was found that provides desirable emission characteristics. Typical data recorded with such cathodes are shown on Figure 6. A rapid increase in the current is observed when the electric field is higher than 2V/µm. When plotted in a Fowler-Nordheim representation, the data show field emission behavior. Two different regimes of emission are observed. 2mm diameter cathodes can reach currents of a few ma before showing permanent degradation. When permanent degradation of the cathode does occur, the field emission properties are not completely lost, but rather the voltage required to achieve a given current is merely increased. This suggests that either the emitting sites are not completely destroyed or they are replaced by other nanotubes in the film. The requirement for initial break-in of new cathodes and the hysteresis often observed in field emission data suggest a CNT reorganization within the film when an electric field is applied, giving the film a configuration more favorable for field emission. This reorganization is also interpreted as a cause of the robust behavior of the cathodes when run in severe conditions such as repetitive arcing. Smaller diameter cathodes showed a significant improvement of the current density. Sustainable currents of over 1A/cm 2 were routinely measured with 100 µm diameter cathodes. This phenomenon requires further investigation but is believed to be the consequence of better field penetration in smaller diameter films which results in higher emitting site density. Figure 5; instrumental setup for field emission characterization. Figure 6; Field emission data of a 2mm diam. MWCNT cathode; left: applied electric field vs current density with typical operating current in X-ray tube; right: Fowler-Nordheim plot showing two emission regimes. Reproducibility of the emission characteristics was evaluated at Oxford Instruments X-ray Technology Inc. using a specially designed apparatus allowing characterization of several cathodes in parallel in a UHV chamber. These tests were run in a diode mode with a cathode to anode distance of 475 µm. Figure 7 shows the distribution of electric fields required to draw a current of 100 µa (after progressive run-in at 500 µa) measured on a series of 125 cathodes with 2 mm diameter films. Most cathodes require a low electric field to emit 100 µa, the typical operating current of a low power X-ray tube. The dispersion of values is large, however. While

7 Copyright JCPDS - International Centre for Diffraction Data 2004, Advances in X-ray Analysis, Volume this poor reproducibility only has a minor consequence on the fabrication of large focus triode tubes (since the extraction field can be independently controlled by a gate voltage), it is a significant drawback for the fabrication of commercial diode geometry tubes or microfocused tube. It is expected that further refinement of substrate preparation and catalyst deposition procedures will improve the reproducibility of cathode performance. This problem has, however, minimal consequences on the development of space deployable tubes as thorough selection allows one to choose cathodes for optimum performance % of Total Field (V/um) Figure 7; variation of the electric field required to extract 100 µa from 2mm diam. MWCNT films measured on 125 cathodes in diode geometry with 475µm gap; cathodes were progressively ramped up to 500 µa prior to measurements. MINIATURE FIELD EMISSION X-RAY TUBE MWNT field emitters were installed in miniature low power X-ray tubes designed by Oxford Instruments X-Ray Technologies Inc. (Figure 8). The tube design is based on a small ceramic enclosure in which is installed a CNT emitter gated with a transmission grid for emission current regulation. Although not designed as microfocused sources, these tubes were fitted with electron optics to limit the size of the X-ray spot (Figure 9). Both reflection and transmission anode geometries were produced using this design. The efficiency of these tubes was measured to be >80%, as compared to ~50% for filament tubes. Lifetime tests of these tubes are in progress and no failures have yet occurred. One tube has been operated at 1.5 W for >100,000 10s pulses (66% duty cycle) illustrating the robust operation of these CNT sources. This technology has been implemented in Oxford Instruments Eclipse II miniature X-ray sources, which integrate a 3 W CNT field emission tube and power supply in a small package (160x38mm, 300g) operated from batteries. These sources are commercially available in transmission and reflection anode configurations for X-ray spectroscopy applications.

8 Copyright JCPDS - International Centre for Diffraction Data 2004, Advances in X-ray Analysis, Volume Figure 8; Oxford Instruments X-ray Technology Inc. miniature field emission X-ray tube with gated MWNT 2mm diam. cathode. Figure 9; image of the X-ray emitting spot (1276x602µm) While CNT field emission cathodes are a significant leap forward in miniature X-ray tube technology, these sources do not address all the needs of the planetary XRD/XRF application. Further improvements of efficiency are expected from new gate designs and diode geometries under development. Reduction of the X-ray spot size to an appropriate focus of <50µm is expected from smaller cathodes combined with redesigned focusing optics. Microfocused miniature X-ray tube prototypes will be tested in early CONCLUSIONS Field emission cathodes based on multiwalled carbon nanotube films were fabricated and tested in miniature X-ray tubes. A thermal CVD carbon nanotube growth process was adapted for the cathode fabrication. Durable emission and high current densities were measured from these CNT cathodes. Miniature X-ray tubes implemented with such emitters have shown robust operation and improved efficiency. While this technology has been applied to miniature commercial tubes for portable XRF applications, further development is necessary to make these X-ray sources suitable for a planetary X-ray diffraction instrument. The resulting microfocused tube will be proposed for Mars Science Laboratory (2009) mission as part of the CheMin XRD/XRF mineralogical instrument. REFERENCES [1] Blake D. et al.. Lunar and Planetary Science XXIII, (1992) [2] Vaniman, D., D. Bish, D.F. Blake, S.T. Elliott, P. Sarrazin, S.A. Collins and S. Chipera, J. Geophys. Res. 103(E13), 31,477-31,489 (1998) [3] Sarrazin, P., D. Blake, D. Bish, D. Vaniman, S.A Collins, S. Chipera and T. Elliott, J. Phys. IV France 8: (1998) [4] Sugie, H., M. Tanemura, V. Filip, K. Iwata, K. Takahashi, and F. Okuyama l., Appl. Phys. Lett. 78(17), (2001) [5] Yue G. Z., Q. Qiu, Bo Gao, Y. Cheng, J. Zhang, H. Shimoda, S. Chang, J. P. Lu, and O. Zhou. Appl. Phys. Lett. 81(2), (2002).

9 Copyright JCPDS - International Centre for Diffraction Data 2004, Advances in X-ray Analysis, Volume [6] Spindt, C. A.. INSPEC Abstract Number A (1972) [7] Brodie, I. and Schwoebel, P.R. Proceedings of the IEEE, Vol 82 Number 7, (1994). [8] De Heer W.A., A. Châtelain and D. Ugarte, Science 270, 1179 (1995) [9] Saito Y. and S. Uemura, Carbon 38, (2000) [10] Bonard J-M., H. Kind, T. Stöckli and L-O. Nilsson, Solid-State Electronics 45, (2001) [11] Wang, Q.H., T.D. Corrigan, J.Y. Dai, and R.P.H. Chang, Appl. Phys. Lett 70(24), (1997) [12] Delzeit L., C. Nguyen, B. Chen, R. Stevens, A. Cassell, J. Han and M. Meyyappan, J. Phys. Chem. B, 106, 5629 (2002) [13] Cassell, A.M.; Verma, S.; Delzeit, L.; Meyyappan, M. Langmuir 17, 266 (2001) [14] Delzeit L., B. Chen, A. Cassell, R. Stevens, C. Nguyen, M. Meyyappan, Chem. Phys. Lett., (2001) [15] Nguyen, C., K. Chao, R. Stevens, L. Delzeit, A. Cassell, J. Han, M. Meyyappan, Nanotechnology 12, 363 (2001)