The new Surrey ion beam analysis facility

A.Simon [1], C.Jeynes, R.P.Webb, R.Finnis, Z.Tabatabaian, P.J.Sellin [a], M.B.H.Breese [b], D.F.Fellows [c], R. van den Broek [d], R.M.Gwilliam,
University of Surrey Ion Beam Centre, Guildford, England, GU2 7XH a) b) c) d) 1) Department of Physics, University of Surrey, Guildford, England now at: Department of Physics, National University of Singapore Arun Microelectronics Ltd., Fitzalan Road, Arundel, West Sussex, England High Voltage Engineering B.V., Amersfoort, The Netherlands on leave from: Institute of Nuclear Research, Hungarian Academy of Sciences (ATOMKI), Debrecen, Hungary

We present the characteristics of an ion beam analysis facility based on a fully computer controlled 2MV Tandetron commissioned in 2002. The terminal voltage temperature stability is measured using the 3036keV 16O( , )16O resonance.

A microbeam beamline with an Oxford triplet lens and a coolable sample holder has been built. We report new temperature dependent time-resolved IBIC measurements from CZT detectors that illustrate the capability of the system. A millibeam beamline has been built with a 6-axis goniometer capable of handling 100mm wafers through an airlock. This instrument is designed for routine RBS/ERD and channelling applications, and incorporates a versatile controller networked to the machine computer and capable of fully automated data collection. A z-axis permits precise location of the sample surface at the eucentric point for work with glancing beam incidence.

1. Introduction
A new 2MV Tandetron accelerator from High Voltage Engineering Europe was installed at the University of Surrey in a new building with a very stable reinforced concrete floor and a separate local earth for the instrumentation. It was commissioned in Spring 2002. The old 2MV van de Graaff [1,2] has been replaced. The main purpose of this machine is to provide reliable light ion beams of 1H+, 3He+, 4 He+ and 4He++ to users of the Ion Beam Centre for a variety of ion beam applications, including both routine millibeam depth profile and channelling analyses on large batches of samples, and detailed microbeam analyses using IBIC as well as the standard PIXE, RBS and NRA. The machine is installed in an open lab which precludes the use of deuterium and high energy 3He beams due to radiation hazard.

2. The accelerator
The accelerator and its current beam lines are shown in Fig.1. The accelerator and injection assembly have been described by Mous et al [3]. A negatively charged He beam is provided by a duoplasmatron source together with a lithium charge exchange canal. A negatively charged H beam is extracted directly from the duoplasmatron source. We also have a caesium-sputter type ion source, and the accelerator has produced useful beams of 10MeV Au 4+ which we may use in the future for high resolution ERD analysis. This source may also be used in the future for high energy heavy ion implantation applications. For analytical convenience we will install two duoplasmatron sources to minimise the time to switch between H and He beams. The computer control of the accelerator [3] has been enhanced to allow unattended overnight running, and some communication between it and the target chamber controller. The accelerator provides an extremely stable beam, with beam current on target varying by less than 1% over 5 minutes and drifting by less than 5% over several hours. Energy stabilisation using feedback from either the low energy or the high energy magnet slits is not required to achieve this. However, the system is sensitive to the calibration of the generating voltmeter (GVM) which provides the system energy reference. We have observed thermal drift in this instrument by monitoring the yield of the 3036keV 16O( , ) 16 O resonance (Fig.2) and a thick silicon dioxide sample using non-Rutherford elastic backscattering into a detector with scattering angle 167. The laboratory temperature was measured at the GVM. At the resonance energy the yield is 50% of the maximum, and it is very sensitive to the precise energy of the beam. Fig.2 shows that the temperature coefficient of the terminal high voltage is about 1kV/C or about 0.07%/C, presumably due to the thermal drift of the mechanical dimensions of the GVM and its position relative to the high voltage terminal.

A beam stable in both position and energy is crucial for a nuclear microprobe. The Tandetron is fitted with the anti-ripple instrumentation described by Mous et al [4] giving a voltage stability expected to be better than 50V.

3. Microprobe beam line

3.1 Technical details
The nuclear microprobe line is installed on a 30 port of the magnet. The excellent energy stability of the accelerator (with a stable temperature) means that slit stabilisation and the 90 magnet deflection often used are unnecessary. Two sets of manual micrometer driven slits are installed: the object aperture close to the magnet and the divergence limiting aperture 6.3 m from the object slits and immediately before the scanning coils. The object aperture assembly is mounted on a massive antivibration concrete block, as is the lens assembly which includes the beam divergence aperture, the scanning coils and the target chamber. Both of these assemblies are separated from the magnet and the long flight tube with anti-vibration bellows. The target chamber is fitted with a Leybold Turbovac 340M (magnetic levitation) pump, which is separated from the vacuum system with anti-vibration bellows. The typical vacuum in the chamber is 1.5.10 -6 mbar, and in the beam line is 3.10-7mbar. The focussing system is an quadrupole triplet [5] from Oxford Microbeams Ltd. The object distance is 6.3m while the image distance is 16 cm, giving demagnifications of Dx=74 in the horizontal plane, and Dy=22 in the vertical one, as calculated using the PRAM computer code [6]. Typically 2 micron sized beam have been reached easily and routinely. The alignment of the system was done within one millimetre precision. There is no need to use the installed electrostatic steering plates after the object slits to get the beam into the target chamber. Improvements are scheduled to the object slit mechanism which will allow reliable focussing of the beam to a 1 micron diameter. A wide range of hydrogen and helium ions have been focussed so far. These include: 1.03.0 MeV H+, 0.9-1.4MeV 3He and 1.4-4.5 MeV 4He2+. The beam brightness of the accelerator is typically is B=1 pA/(m2 mrad2 MeV) for 2 MeV H+ and B= 0.5 pA/(m2 mrad2 MeV) for 2 MeV He2+ at a divergence of 0.08 mrad. These values are comparable to other reports of duoplasmatron sources (see references in [7]). We should note that these values are for guidance only: we have seen a much higher value of 8 pA/(m2 mrad2 MeV) for 2 MeV H+. The octagonal target chamber from Oxford Microbeams Ltd. Allows a small image distance and hence high demagnification. A 3-axes Huntington PM-600 TRC Precision XYZ sample stage with 2 microns movement precision is mounted onto the lid. A 80 mm2 Si(Li) detector with a 12.5 micron thick Be window is installed at a backward angle of 45 for PIXE spectroscopy. The calibrated distance of the Si crystal can be varied between 25mm and 70mm. Beside the charge calibration, for RBS and NRA analysis, a 50mm2 Ortec Ultra detector is mounted in Cornell-geometry (=165 , =16.5 msrad) above the entrance of the beam. In transmission geometry a Hamamatsu S1223 Si photodiode is mounted onto a rotating flange, making available both on-axis and off-axis transmission measurements. The data acquisition (OM-DAQ) system uses the OM 1000 Interface Unit (ADC) data analyser from Oxford Microbeams Ltd. The data collection on an event by event basis (listmode) is fully exploited with OM-DAQ software.

The beam current is measured directly from the isolated sample stage. This is a useful monitor but for accurate work the well known Q-factor method is used [8]. The beam current can be measured absolutely (by moving the sample out of the beam) on a carbon Faraday-cup situated on-axis at 0 on a rotating flange at the back of the target chamber. The microbeam chamber is fitted with a coolable sample stage, which is automatically controlled over the temperature range 80 - 300 K. The sample plate is attached to a coldfinger filled with liquid nitrogen, with an associated heater block. Temperature monitoring is carried out using two PT100 platinum resistance thermometer connected to a Scientific Instruments type 9700 temperature controller, which achieves a temperature stability of better than 5 K. A 1.3 litre liquid nitrogen dewar is mounted directly above the sample chamber, which requires refilling every 2 hours. The sample plate is mechanically connected to the chamber lid via a plastic rod to minimise thermal displacement and vibration of the sample during cooling, which is typically 100 m vertical displacement of the sample when cooling from 300 K to 100 K.

3.2 Research
A significant area of research on the microbeam line has been in the area of Ion Beam Induced Charge (IBIC) microscopy of charge transport in compound semiconductor materials. This work has been carried out as part of the ongoing research programme developing new materials for radiation imaging detectors which is carried out in the Department of Physics. The excellent spatial resolution and stability of the microbeam provides a powerful tool for IBIC imaging of bulk semiconductor materials, and measurements have been carried out on a range of materials including cadmium zinc telluride (CZT), cadmium telluride (CdTe), polycrystalline CVD diamond [9], and semiinsulating epitaxial gallium nitride (GaN) and silicon carbide (SiC). Conventional IBIC imaging [6] of pulse amplitude in these materials provides quantitative high resolution maps of charge collection efficiency (CCE) and charge carrier drift lengths. This data gives insights into the distribution and nature of charge traps in bulk semiconductor materials. With a 1 micron resolution beam, IBIC imaging of charge transport inside individual crystallites of CVD diamond has been used to measure intra-crystallite CCE and to show the concentration of trapping centres at the crystallite boundaries [10]. The use of the coolable sample stage allows IBIC imaging as a function of temperature, which allows the role of individual charge traps to be investigated. Figure 3 shows an example of an IBIC image acquired at 250 K, which images the CCE distribution across a corner of a 2 mm thick CZT detector [11]. The data demonstrates the excellent charge transport uniformity of this device, and also shows some additional features due to a surface scratch (two vertical lines) and mechanical probing marks (regular grid of small dots). A vertically-aligned region of higher CCE, reaching 80%, is observed on the right hand edge of the image at this temperature, which was not observed at 300 K, due to the partial filling of electrical traps by charge produced from the ion beam. Digital IBIC is a new extension of the regular IBIC technique that has been recently pioneered at the Surrey microbeam [12]. In this method, the analogue pulse processing electronics, ADC and multichannel analyser is replaced by a high-speed waveform

digitiser. This allows complete capture of the IBIC preamplifier output signal, allowing a digital pulse shape analysis to be performed in software. In this way digital IBIC performs high resolution ion beam imaging measurements of charge drift velocity and mobility in bulk semiconductors, in a way analogous to laserinduced time of flight measurements. Figure 4 shows the first data acquired with the digital IBIC system where a 2.5 MeV He beam was used to laterally scan across the edge of a 2 mm thick CZT detector, between the cathode and the anode. In this device, where there is significant hole trapping, the charge signal amplitude induced on the contacts is strongly dependent on the distance of the ion beam interaction from the cathode. Close to the cathode the induced signal is predominantly due to electron transport, and has a high magnitude. Conversely, close to the anode the induced signal is due to hole transport which produced a very small signal amplitude. Lateral IBIC imaging is a powerful technique to investigate these types of 'depth' dependent variations in device response. Figure 4(a) clearly shows the decrease in signal amplitude as a function of distance from the cathode, and regions of different amplitude response due to material non-uniformities. Figure 4(b) shows the corresponding increase in signal risetime due to the increasing influence of the slow hole component of the total signal. Analysis of this data gives a room temperature electron drift mobility value of 1.1 x 103 cm2/Vs. Other work in progress includes PIXE of blood samples [13], where the cooled sample stage is important to reduce beam damage artefacts, and work preliminary to implementing PIXE/STIM tomography [14].

4. Millibeam line
4.1. General
The purpose of this beam line is to facilitate unattended automatic data collection on large batches of samples, including depth profiling and multiaxis channelling on single crystals. The goniometer is designed to hold many or large samples and to allow access to the beam in glancing incident or exit directions. The instrument is large and is accomodated by a 600mm diameter target chamber. Depth profiles can be extracted automatically from the collected spectra using our DataFurnace [15] software. Two backscattering particle detectors are installed at different scattering angles to help resolve ambiguities and to provide two independent data channels for validation of accurate analysis. One of these is in the Cornell geometry with provision for extra collimation: this is for high depth resolution work. A further particle detector with a range foil is mounted in a forward recoil direction for H profiling by ERD with a 4He beam. Other instrumentation includes a chopper with backscattered particle detector for independent beam current monitoring, and an alignment laser giving a beam coincident with the ion beam and mounted behind the high energy magnet. A fixed video camera allows the samples to be viewed directly. Electrostatic deflection plates just after the aperture by the high energy magnet allow the beam to be switched off.

4.2. 6 motion goniometer

Fig.5 shows the new 6 movement goniometer which has been designed and built by Arun Microelectronics Ltd and which has a similar construction to that reported by Hollnder et al [16]. It has a vertical rotation axis with 363 of motion and 0.001 per step: stacked on top of this motion is a horizontal tilt axis perpendicular to the beam with 30 of motion and 0.0005 per step, and a rotation axis perpendicular to the tilt axis and normal to the sample stage surface with 210 of motion and 0.00025 per step. The axial motions have a maximum speed of 2000 Hz. The whole mechanism is behind the sample stage so that glancing beam incident and beam exit geometries can be used. In the centre of the sample stage (Fig.6a) is a 25mm hole which can be aligned with a corresponding hole in the goniometer mechanism so that the beam can be passed through the instrument to the back of the chamber allowing a Faraday cup beam stop (or further target chambers) to be installed.. The sample stage is 105*153mm and slides into the mechanism, with push-fit electrical connections being made to allow current integration from the sample, or to allow the installation of sample heating or other instrumentation. Every part of the sample stage can be brought under the beam with X and Y linear motions stacked on top of the rotation axis. A Z motion of 5.6mm perpendicular to the sample stage is used to make the plane of the sample surface intersect the eucentric point (the point of intersection of the three rotation axes) for glancing beam incidence applications. All the linear motions have a resolution of 0.004mm per step and a maximum speed of 1500 Hz. The sample stage exchange mechanism is through an airlock allowing the samples to be changed within 10 minutes.

4.3 Goniometer controller software

The six goniometer stepper motors are controlled by three SMD2 dual controllers from Arun, interfaced to the PC with RS232, which control the slew rate, acceleration behaviour and holding torque parameters, and which monitor the motor temperature. We have modified the controller firmware to allow current position and motor temperature to be read by the PC while the motor is moving. It is important for the control program to monitor motor temperature continuously since motor overheat protection will cause it to stop, losing steps. Continous position readout is important to implement our DualTrace program for channelling [1]where count rate is plotted against angular position. The program is named for its implementation of a dual trace chart recorder simultaneously displaying beam current and count rate. This chart recorder function also has a logging tool, which allows the whole session to be stored and replayed. DualTrace is a multipurpose data collection program written in VisualBasic which controls the goniometer, the video camera, three ADCs cards in the PC, digital inputs from the current integrator and chopper, and digital outputs for the chamber light, the alignment laser and the beam switch. There are tools for setting the goniometer hardware limits of motion, and also the sample exchange and beam-normal positions. Other tools set various parameters of the ADCs, display the current spectra, display the video image, determine ADC energy calibration with the procedure reported previously [2], and display the dual trace for channelling. The scripting tool is described below.

Fig.6a shows the stored video image of a sample plate taken when the sample plate is square to the camera and with pre-set X,Y position. This image has a standard calibration of the X,Y position so that the beam position on the plate can be displayed continuously for any orientation of the plate, provided that the beam intersects the eucentric point and provided that the eucentric point is intersected by the sample surface. Because of the standard calibration, the stored image can be used to move any part of the plate under the beam by point and click. We can then set up a series of pre-set Positions with point and click, and in principle we could use image processing software to determine the positions of the sample edges automatically. Fig.6b illustrates our scripting language implementation allowing unattended data collection. Looping structures are included, with various looping conditions. Nested loops are allowed. Arithmetical and logical operations are allowed, and an automatic channelling algorithm can also be scripted. Goto commands can be to named Positions, to absolute positions, relative to current position, and specified in steps or millimetres/degrees. All the ADC functions are available, including stopping at a specified collected charge or after a specified collection time. Automatic unique filenaming is implemented when the data is stored, where filename postfixes can be specified. Filename prefixes are A, B or C for the specific detectors used.

Acknowledgements
The Ion Beam Centre and PKS acknowledge the financial support of EPSRC under grants GR/M94434/01, GR/R50097/01 and GR/R34486/01. We are very grateful for the contribution of Mark Browton in the beam line installation and Adrian Cansell for the machine operation.

Figure Captions
Fig.1: Schematic of building with 2MV Tandem installation (approximately to scale). S = double source assembly; L = low energy magnet; T = 2MV Tandetron accelerator; H = high energy magnet; M = microbeam line with object aperture (o); B = broadbeam (millibeam) line with apertures (a). Fig.2: Temperature coefficient of Tandetron terminal voltage using a 4He+ beam. Fig.3: IBIC image acquired at 250 K imaging the charge collection efficiency (between 50% and 80%) distribution across a corner of a 2 mm thick CZT detector. 50V bias. Scan size ??? Fig.4: Digital IBIC images acquired laterally through the side of a CZT detector, showing (a) pulse amplitude and (b) pulse risetime. The device cathode is indicated towards the left of the image, the anode is beyond the right edge. Fig.5: 6 movement UHV goniometer from AML. The sample stage is removed in this image. The gear, stepper motor and anti-backlash constant force spring for the 210 horizontal rotation axis can be seen. Fig.6a: The DualTrace program, video utility. The cross over `Position 14` marks the current beam position. The absolute X,Y position in steps is given in the two text windows. There are three numbered Positions on each of four samples (see Fig.6b). The rectangle framing the samples gives the calibration in mm of the

video image. The hole in the centre of the sample plate allows beam transmission through the instrument. Fig.6b: The DualTrace program, scripting language. Portions of the `Script` window are shown. Right shows a portion of the script generator, left shows the script generated. Each While loop collects data at three Positions for one sample (see Fig.6a), where each spectrum has 200 counts in region of interest 0 and is stored with automatic unique filenaming. The MCBs are multichannel buffers: ORTEC TRUMP cards for the PC implementing the ADCs.

Fig.1: Schematic of building with 2MV Tandem installation (approximately to scale). S = double source assembly; L = low energy magnet; T = 2MV Tandetron accelerator; H = high energy magnet; M = microbeam line with object aperture ("o"); B = broadbeam (millibeam) line with apertures ("a").

Variation of energy with temperature

3036.5 3036.0 True energy (keV) 3035.5 3035.0 3034.5 3034.0 3033.5 3033.0 3032.5 22.5

23.0

23.5

24.0

24.5

Monitored Temperature (C)

Fig.2 Temperature coefficient of Tandetron terminal voltage using a 4He+ beam.

Fig.3: IBIC image acquired at 250 K from CZT (charge collection efficiency between 50% and 80%). 50V bias.

-50 V

10

Fig.4: Digital IBIC images acquired laterally through the side of a CZT detector, showing (a) pulse amplitude and (b) pulse risetime. The device cathode is indicated towards the left of the image, the anode is beyond the right edge.

(a)

(b)

11

Fig.5: 6 movement UHV goniometer from AML. The sample stage is removed in this image. The gear, stepper motor and anti-backlash constant force spring for the 210 horizontal rotation axis can be seen.

12

Fig.6a: The DualTrace program, video utility. The cross over `Position 14` marks the current beam position. The absolute X,Y position in steps is given in the two text windows. There are three numbered Positions on each of four samples (see Fig.6b). The rectangle framing the samples gives the calibration in mm of the video image. The hole in the centre of the sample plate allows beam transmission through the instrument.

13

Fig.6b: The DualTrace program, scripting language. Right shows the script generator, left shows the script generated. Each While loop collects data at three Positions for one sample (see Fig.6a), where each spectrum has 200 counts in region of interest 0 and is stored with automatic unique filenaming. The MCBs are multichannel buffers: ORTEC TRUMP cards for the PC implementing the ADCs.

14

References
1 J E Mynard, C Jeynes, J Thornton, A Way, R P Webb, P L F Hemment, K G Stephens, Improved facilities for ion beam analysis at the University of Surrey.Nuclear Instr. & Methods B6, 264-269 (1985) C.Jeynes, N.P.Barradas, M.J.Blewett, R.P.Webb, Improved ion beam analysis facilities at the University of Surrey, Nucl. Instr. and Methods B136-138 (1998) 1229-1234 D. J. W. Mous, A. Gottdanga R. Van den Broek, R. G. Haitsma, Recent developments at HVEE, Nucl. Instr. and Methods B99 (1995) 697-700 D. J. W. Mous, R. G. Haitsma, T. Butz, R. -H. Flagmeyer, D. Lehmann., J. Vogt, The novel ultrastable HVEE 3.5 MV Singletron accelerator for nanoprobe application, Nucl. Instr. and Methods B130 (1997) 31-36 G.W. Grime, M. Dawson, M. Marsh, I.C. McArthur, F. Watt, The Oxford sub-micron nuclear microscopy facility, Nucl. Instr. and Methods B54 (1991) 52. M.B.H. Breese, D.N. Jamieson, P.J.C. King, Materials Analysis Using a Nuclear Microprobe, Wiley, New York, 1996; code from: http://www.ph.unimelb.edu.au/~dnj/research/pram/pramdist.htm courtesy of Prof. Jamieson R. Szymanski, D.N. Jamieson, Ion source brightness and nuclear microprobe applications, Nucl. Instr. and Methods B130 (1997) 80-85 G.W. Grime, The ''Q factor'' method: Quantitative microPIXE analysis using RBS normalisation, Nucl. Instr. and Methods B109/110 (1996) 170-174 A. Simon, P. Sellin, A. Lohstroh, IBIC imaging of intra-crystallite charge transport in CVD diamond at reduced temperature, to be published in these proceedings: Abstract 318 P.J. Sellin, M.B.H. Breese, A. Simon, A. Galbiati, A. Lohstroh, S. Rath, L.C. Alves, Intracrystallite charge transport in CVD diamond using micron resolution Ion Beam Induced Charge imaging, submitted to Applied Physics Letters A. Lohstroh, P.J. Sellin, D. Hoxley, A. Simon, IBIC imaging of Cadmium Zinc Telluride, submitted to Nuclear Instruments and Methods in Physics Research A P.J. Sellin, A. Lohstroh, A. Simon, M.B.H. Breese, Digital IBIC - new spectroscopic modalities for Ion Beam Induced Charge imaging , submitted to Nuclear Instruments and Methods in Physics Research A L. L. Admans and N.M. Spyrou, Elemental composition of erythrocytes from Coronary Artery Bypass Grafting (CABG) patients with and without type II diabetes mellitus using PIXE analysis, Presented at Methods and Applications of RadioChemistry VI (Hawaii, May 2003), submitted to J.Nucl.Radioanalytical Chem. D. Beasley, N. M. Spyrou, The University of Surreys new proton microprobe and the feasibility of PIXE-tomography for trace element analysis of biological tissues, Presented at Methods and Applications of RadioChemistry VI (Hawaii, May 2003), submitted to J.Nucl.Radioanalytical Chem. Jeynes C, Barradas NP, Marriott PK, Boudreault G, Jenkin M, Wendler E, Webb RP, Elemental thin film depth profiles by ion beam analysis using simulated annealing - a new tool, J.Phys.D 36 (2003) R97-R126 Hollnder B, Heer H, Wagener M, Halling H, Mantl S, New high-precision 5-axes RBS/channeling goniometer for ion beam analysis of 150 mm diameter wafers, Nucl. Instr. and Methods B161 (2000) 227-230

2 3 4

5 6

7 8 9 10

11 12

13

14

15

16

15