Revised 6-Jan-97
Paper E-1 presented at the Fifth International Conference on Nuclear Microprobe Technology and Applications, Santa Fe and Albuquerque, NM, USA, November 11-15 1996.
You can also jump straight into the slide show here.
RECENT APPLICATIONS OF NUCLEAR MICROPROBE ANALYSIS TO FRONTIER MATERIALS, David N. Jamieson, School of Physics, Microanalytical Research Centre, University of Melbourne, Parkville, 3052, AUSTRALIA.
Frontier materials, with many potential applications in modern technology, continue to provide challenges for the nuclear microprobe community. Over the past two years, new applications for the nuclear microprobe in the study of frontier materials have been found in diverse areas. These have ranged from investigations of the material composition and purity, through probing lattice locations of possible dopants, to the study of the charge transport properties of single and polycrystalline radiation detectors. The techniques applied ranged from traditional Backscattering and Particle Induced X-ray Emission, through emerging techniques such as Ion Beam Induced Charge, to the experimental technique of Ionoluminescence. This paper reviews some of these recent applications of the nuclear microprobe to the study of frontier materials.
Introduction
Many aspects of modern technology depend on materials that have a composition or structure that gives them unique and useful properties. New materials are introduced from the frontier. At any given time, a number of materials are at the frontier. Some materials have been at the frontier for many years. Several stages can be identified in the development of a frontier material from its synthesis to eventual applications, as shown in Fig. 1. At each stage, the nuclear microprobe can play a role and some of these roles are identified in table 1 where the applications from the past two to three years are also listed. It is the purpose of this paper to review some of these roles. Reviews of applications that occurred earlier are available [1,2]. For a review of the application of the nuclear microprobe to the study of materials fabricated into complete microelectronic devices and elsewhere in the semiconductor industry, see Takai [3]. Extensive details about nuclear microprobe technology and many applications to the analysis of materials are described in a recent book [4].
Most nuclear microprobe studies can be divided in to two categories on the basis of the analytical technique employed: High beam current (PIXE, BS, IL) or Low beam current (IBIC, STIM). These and most other acronyms used here are defined in table 1. The high current techniques typically employ beam currents in excess of 100 pA focused into a probe 1 mm. Values for low current techniques are typically more than an order of magnitude smaller in both dimensions. Phenomena investigated range from the fundamental arrangement and proportion of atoms in the material to the electronic properties of the electron energy level bands built on the underlying atomic arrangement. Examples presented here are mostly from the Melbourne group, examples of work elsewhere can be found in the references in table 1.
Applications
Diamond
Since the last conference in this series, there has been an explosion of activity in the applications of nuclear microprobe analysis to the study of diamond. Diamond has been a promising frontier material for a number of years, based largely on its superlative electrical and structural properties that make it attractive for use in microelectronic devices and radiation hard detectors for adverse environments including high energy physics experiments [4-6]. With a wide band gap (5.5 eV) it is relatively insensitive to thermal noise or light and with excellent thermal conductivity it can operate at high temperatures. Also, undoped diamond has extremely high resistivity and high breakdown field strength offering excellent electrical isolation together with high electron and hole mobilities (see table 2 where these quantities are compared to Si and GaAs). Furthermore, diamond is regarded as an inert, bio-compatible material with a low atomic number that is closely matched to human tissue making it ideal for radiation dosimetry applications. These and other applications of diamond radiation detectors are discussed in the excellent review of Kania et al. [7].
The applications of the nuclear microprobe to the study of diamond have occurred in three main areas. The first is in the synthesis of Chemical Vapour Deposition (CVD) films. The second is in attempts to make n-type diamond by ion implantation and annealing. The third is in the investigation of the charge transport in natural and poly-crystalline CVD grown diamond films.
Diamond Synthesis
In the synthesis of diamond films, attention must be paid to purity of the diamond film, since contaminant atoms can degrade the performance of detectors fabricated from the material. Residual impurities consisting of Al, Si, Ti and Fe, and to a lesser extent Na, Mg, S, K, Ca, Cr and Pd have been mapped by PIXE [8]. Additional electron microprobe analysis revealed that some of these impurities were concentrated at grain boundaries and were not detected within grains. This is expected, since grain boundaries act as a favourable path for the diffusion of impurities. Further studies in Melbourne by combined BS and PIXE with a 1.4 MeV H+ microbeam on 6 mm thick homoepitaxial hot-filament CVD-grown diamond films revealed W contamination from the filament at the interface with the substrate, with a concentration of 0.5% [9]. This may limit the purity of films grown by this method.
Diamond Ion Implantation and Annealing
Most applications of diamond to the fabrication of microelectronic devices will require both p-type and n-type diamond. P-type diamond can be made, but the n-type doping is much more problematic. The high atomic density of the diamond lattice and the metastability of the sp3 bonds eliminates traditional techniques that are applicable to silicon. However doping with boron during CVD growth has been demonstrated, although has limitations (see the review by Kalish [10]) and has not been extended to n-type dopants.
Ion implantation offers a possible solution and previous work in Melbourne had shown that a focused laser beam, diameter 20 mm, could partially regrow the damage produced by deeply buried 2.8 MeV C+ implantation (Rp ±DRp = 1.48 ±0.06 mm) to a dose of 3x1015 C+/cm2 [11,12]. In that early study, Channeling Contrast Microscopy (CCM) was used to obtain channeling spectra from the annealed regions to demonstrate the success of the annealing scheme. A new laser annealing scheme has now been devised that uses multiple laser pulses of increasing energy to more fully anneal the diamond [13]. In this case a diamond was implanted with 4.0 MeV P+ (Rp ±DRp = 1.34 ±0.07 mm) to a dose of 1x1015 P+/cm2. The effectiveness of the multiple pulse annealing scheme was clearly shown by CCM analysis which revealed a close to pristine diamond min in the laser annealed regions. A 1.4 MeV H+ microbeam was used to take advantage of the relatively large scattering cross section for C [14] which is about factor of 10 greater than that for a more traditional 2 MeV He+ beam (see data in table 2).
The P concentration at the end of range was about 18 ppm which is too low to be detected by backscattering spectrometry. However, with the nuclear microprobe system in Singapore, optimised for high sensitivity trace element analysis in organic materials by PIXE, angular yield curves could be obtained from the P Ka X-ray line. Sufficient sensitivity was obtained with a microprobe of 1 MeV H+, as this produced the lowest intensity background signal under the P Ka X-ray, with a large area X-ray detector (crystal total area 80 mm2, collimated to 60 mm2) in close geometry (~15 mm). This allowed a P channeling angular yield curve to be obtained from the centre of a laser annealed spot in the P implanted diamond, as shown in Fig. 2. This shows a reduction in the yield of P X-rays in the channeling orientation indicating that at least 50% of the P atoms are substitutional on the lattice sites [15]. Owing to dechanneling of the incident beam at the depth of Rp, the true substitutional fraction is almost certainly higher. Electrical measurements are now in progress to determine the electrical activity of the substituted atoms [16]. More detailed studies of the residual defects have been done with Raman spectroscopy which revealed that several types of defects exist, not all of which are removed by the laser annealing [17].
Charge transport in diamond
To successfully use diamond as a radiation detector, it is necessary to fully characterise the charge transport, recombination and collection mechanisms. Recently, high quality synthetic diamonds have become available. These are polycrystalline with a grain size of ~50 mm so the nuclear microprobe is ideal for performing Ion Beam Induced Current (IBIC) analysis on these materials. Considerable IBIC work has been done by the group of Manfredotti et al. [8,18-22] in collaboration with the nuclear microprobe group in Zagreb. Diamond films in these studies, as well as those for the examples presented here from the Melbourne group, were provided by the Norton Diamond Film Company. The specimens can be up to 400 mm thick, with columns of diamond crystals extending through the layer. It is usual to polish off the layer of the film originally in contact with the substrate owing to its large number of defects.
The mechanism for the production of IBIC signals from diamond films is illustrated in Fig. 3. The diamond film, typically 250 to 400 mm thick, is configured in the Metal-Insulator-Metal (MIM) configuration. Ion beams may be incident from the surface (normal IBIC), which is also the growth direction, or from the transverse direction (lateral IBIC). A bias voltage to provide an electric field is necessary for the detector to function. The mechanism for charge generation in the external circuit is by induction due to the movement of the electron-hole pairs under the influence of the electric field. This is described by Ramo's theorem and its generalisations [23-25]. Defects and other inhomogeneities can rapidly trap charge allowing it to recombine. This reduces the amount of charge otherwise available in the external circuit. The IBIC images are very sensitive to distribution of these devective regions in the diamond.
Some representative normal IBIC images obtained in Melbourne are shown in Fig. 4. For more details of this work see Beckman et alia [26]. Complete understanding of the IBIC images is complicated by a wide variety of phenomena including transient polarisation of the matrix by the beam or the bias voltage and effects arising from properties of the metal-diamond contact. However, the IBIC images display considerable structure indicating about three times more efficient charge collection from the centre of the diamond grains compared to the grain boundaries. Very low efficiency (~1/10 of grain centres) was observed from regions not covered by the metal contacts. This last result is quite different to the situation for most IBIC work to date on semicondutor materials and devices [4] where the specimen usually incorporated a depletion region.
The ionoluminescence (IL) [27,28] spectrum (Fig. 5) of nearby uncovered regions displays features typical of high quality CVD diamond films [29]. Images (Fig. 4e,f) made from the two peaks seen in the spectrum show strong luminescence at 440 nm from the centres of the grains, with more uniformly distributed luminescence at 500 nm. We attribute the different luminescence from grain boundaries to the presence of defects that provide some non-radiative recombination pathways. Detailed comparison of the IL and IBIC spectra from the same regions of the diamond films may provide additional insights into the charge transport and recombination mechanisms.
Solar cells
Silicon as a material for the fabrication of large area photovoltaic panels of solar cells has been studied for several decades [30,31]. Economic fabrication techniques for efficient cells need to be devised. Cells must be able to function efficiently when made from low grade silicon that contains contaminant trace elements or crystal defects. Hence such low grade silicon can be defined as a frontier material for use in solar cells.
The highest efficiency, inexpensive solar cells are fabricated from poly-crystalline silicon. Contaminant trace elements, grain boundaries and other defects trap electron-hole pairs allowing them to recombine. This is typically about 16% of the collectable charge [32] and can be considerably higher [33]. Traditional probes of light or electron beams can be used to investigate the recombination, however, surprisingly, both of these probes suffer from considerable disadvantages which could be overcome with MeV focused ions. A summary of the relevant properties of the three types of probes, with nominal parameters, are in table 3. MeV ions are the only ones which probe to high resolution the same volume of the solar cells in which incident sunlight creates electron-hole pairs.
In solar cells, a layer of n-type material (the collector) is formed upon a layer of p-type material (the absorber). At the junction, a depletion layer is formed, deficient in electrons and holes. Incident photons (or electrons or protons) create electron-hole pairs which diffuse into the depletion region where they drift to the electrical contacts under the influence of the intrinsic electric field and a current flows in the external circuit. Fig. 6 shows these processes in a polycrystalline solar cell.
Previous work on imaging grain boundaries using the conventional technique of Electron Beam Induced Current (EBIC) has been used to study the effect of various passivation techniques aimed to reduce recombination at grain boundaries [34]. However, these studies can only probe the top few micrometres of the specimen, owing to the relatively shallow penetration depth of the electron beams. Apart from EBIC, the other conventional technique for the study of recombination in solar cells has been Light Beam Induced Current (OBIC or LBIC). In a recent study [35], this technique was used with an IR laser to probe the recombination rate at grain boundaries in solar cells that had been passivated by oxidation. The recombination rate was measured to a spatial resolution of 200 micrometres using a focused IR laser.
Considerably higher resolution images can be obtained with IBIC. Some examples of work done in Melbourne [36] are shown in Fig. 7. These images show significant charge recombination at grain boundaries, including evidence for horizontal grain boundaries below the surface which trap charge diffusing from below, and evidence for grain boundaries that are visible in optical images but do not appear to trap charge. Spectra extracted from the regions of interest within the different features in the images, Fig. 8, provide a measure of the relative charge collection efficiency of the different structures in the cell.
The effect of trace elements on solar cell performance has been measured by Davis et al. [37]. Those measurements showed that trace elements with concentrations in bulk of 500ppm (for Cu) to 1ppm (for Cr) were sufficient to degrade the efficiency of some types of cell by about 50%. Other elements have a much more severe effect at even lower concentrations. However it may be possible to usefully map some of these trace elements with PIXE, since, in bulk samples, it is possible to detect some transition metals to a sensitivity of about 2 ppm [38,39]. The spatial distribution of the trace elements could then be compared to the IBIC images. We have already discovered that some trace elements in solar cells, like Zn, appear to segregate so that the microscopic concentrations are several orders of magnitude higher than the bulk concentration. A representative PIXE spectrum from within a grain of a commercially available solar cell is shown in Fig. 9 where the Zn signal is clearly visible.
Other Applications
Detailed investigations have been made of the transmission channeling patterns produced by microbeams, with an aperture mask, passing through thin (~1 mm) crystals. The use of a microbeam for this work allows convenient control over the beam angle relative to the crystal axis. An ingenious experiment, where the beam was channeled through two stacked crystals, with one rotated relative to the other, provided an analogue of a defective crystal [40] in which defects could be imaged by CSTIM. The same method has been used to directly view the phenomenon of channeling oscillations in the trajectories of ion beams in the channels of crystals [41,42] which may lead to the development of new methods for the characterisation of crystal defects.
Many additional applications of the nuclear microprobe are listed in table 1.
Conclusion
Future applications of the nuclear microprobe in the analysis of frontier materials will most likely see an expansion in the role of IBIC, particularly as the feature size of microelectronic devices shrinks to 0.25 mm and below [3]. It is also possible that more applications for IL may develop, particularly for the analysis of diamond and other wide band-gap materials, such as AlN or GaN [43]. More exotic applications for the nuclear microprobe could also be proposed, such as for micromachining of diamond. A recent paper has shown how small diamond gear wheels can be cut from CVD films using a focused laser and special etching techniques [44]. It is possible a scanned microbeam could be used for the same purpose, by exploiting the very large current densities in the focused spot to selectively graphitise outlines of desired shapes allowing the shape to be freed from the diamond matrix. This would also lead to the possibility of the fabrication of fine gratings or other optical components.
From trace elements to electron transport properties, the nuclear microprobe has demonstrated its remarkable versatility for the analysis of frontier materials.
Acknowledgments
The generosity of the MARC staff and students who provided data for this paper is gratefully acknowledged: Deborah Beckman and Andrew Saint for their IBIC work on CVD diamond, Lachlan Witham for the IBIC work on polycrystalline silicon solar cells and Andrew Bettiol for the IL data on the CVD diamond films. Without the excellent work by Roland Szymanski in maintaining the laboratory facilities, none of the Melbourne work would have been possible. The essential participation of Steven Prawer for many valuable discussions is also gratefully acknowledged. This work has been supported by grants from the Australian Research Council.
References
[1] D.N. Jamieson, M.B.H. Breese and A. Saint, Nucl. Instr. Meth. B85 (1994) 676.
[2] D.N. Jamieson, Nucl. Instr. Meth. B104, (1995) 533.
[3] M. Takai, Nucl. Instr. Meth. B113 (1996) 330.
[4] M.B.H. Breese, D.N. Jamieson and P.J.C. King, Materials Analysis with a Nuclear Microprobe, Wiley, New York (1996).
[5] C. Bauer et al. (the RD42 collaboration) Nucl. Instr. Meth. A367 (1995) 202 & 207.
[6] C. White, Nucl. Instr. Meth. A352 (1994) 217.
[7] D.R. Kania, M.I. Landstrass, M.A. Plano, L.S. Pan and S. Han, Diamond and Related Materials, 2 (1993) 1012.
[8] C. Manfredotti, F. Fizzotti, E. Vittone, M. Boero, P. Polesello, S. Glassini, M. Jakšic, S. Fazinic and I. Bodganovic, Nucl. Instr. Meth. B109 (1996) 555.
[9] P.W. Weiser, S. Prawer, K.W. Nugent, A.A. Bettiol, L.I. Kostidis and D.N. Jamieson, Proc. 9th Australian Conference on Nuclear Techniques of Analysis, Newcastle, 1995 (ISBN 1325-1694).
[10] R. Kalish, Presented at the Second International Symposium on Control of Semiconductor Interfaces, Karuizawa, Japan, Oct-Nov. 1996, to be published in The Journal of Applied Surface Sciences.
[11] S. Prawer, D.N. Jamieson and R. Kalish, Phys. Rev. Let. 69 (1992) 2991.
[12] D.N. Jamieson, S. Prawer, S.P. Dooley and R. Kalish, Nucl. Instr. Meth. B77 (1993) 457.
[13] M.G. Allen, S. Prawer, D.N. Jamieson and R. Kalish, Appl. Phys. Lett 63 (1993) 2062.
[14] R. Amirikas, D.N. Jamieson and S.P. Dooley, Nucl. Instr. Meth. B77 (1992) 110.
[15] S. Prawer, D.N. Jamieson, R.J. Walker, K.K. Lee, F. Watt and R. Kalish, Diamond Films and Technology, (1996) in press.
[16] R.J. Walker, S. Prawer, D.N. Jamieson and R. Kalish, in Ion Beam Modifications of Materials, J.S. Williams, R.G. Elliman and M.C. Ridgway (eds), Elsevier Science B.V. (1996).
[17] S. Prawer, D.N. Jamieson and K.W. Nugent, to be published (1997).
[18] M. Jakšic, I. Bodganovic, M. Bogovac, S. Fazinic, S. Glassini, K. Kovacevic, C. Manfredotti and E. Vittone, Nucl. Instr. Meth. B113 (1996) 378.
[19] C. Manfredotti, F. Fizzotti, E. Vittone, M. Boero, P. Polesello, S. Glassini, M. Jakšic, S. Fazinic and I. Bodganovic, Nucl. Instr. Meth. B100 (1995) 133.
[20] C. Manfredotti, F. Fizzotti, P. Polesello, E. Vittone, M. Jakšic, I. Bodganovic and S. Fazinic, Mat. Res. Soc. Symp. Proc., Vol 416, "Diamond for Electronic Applications, L. Dreifas, Ed.,
[21] C. Manfredotti, F. Fizzotti, E. Vittone, P. Polesello, M. Jakšic, S. Fazinic and I. Bodganovic, presented at the Eurodiamond Conference, Torino, Italy January (1996).
[22] C. Manfredotti, F. Fizzotti, K. Mirri, P. Polesello, E. Vittone, M. Jakšic, T. Tadic, I. Bodganovic and T. Pochet, Presented at Diamond 96, 7th European Conference on Diamond, Diamond-like and Related Materials, Tours, France September (1996).
[23] S. Ramo, Proc. I.R.E., September (1938) 584.
[24] G. Cavalleri, G. Fabri, E. Gatti and V. Svelto, Nucl. Instr. Meth. 21 (1963) 177.
[25] G. Cavalleri, E. Gatti, G. Fabri, and V. Svelto, Nucl. Instr. Meth. 92 (1971) 137.
[26] D.R. Beckman, D.N. Jamieson, S. Prawer, A. Saint and P. Gonon, these proceedings.
[27] K.G. Malmqvist, M. Elfman, G. Remond and C. Yang, Nucl. Instr. Meth. B109 (1996) 227.
[28] A.A. Bettiol, K.W. Nugent and D.N. Jamieson, these proceedings.
[29] A.M. Zaitsev, Ch XIII in Optical Properties of Diamond, FernUniversitat GHS in Hagen (1995).
[30] C.R.M. Grovenor, Microelectronic Materials, IOP Press (1989).
[31] M.A. Green, Silicon Solar Cells, Advanced Principles and Practice, Centre fo Photovoltaic Devices and Systems (1995).
[32] Y. Hamakawa, Photovoltaic Power, Scientific American, April 1987.
[33] R.H. Bube, Ch 6 in Photoelectronic Properties of Semiconductors, Cambridge, 1992.
[34] Z. Chen, R.J. Matson and L.C. Burton, Phys. Stat. Sol. (a) 132 51 (1992).
[35] S.W. Glunz, Ch. Hebling, W. Warta, W. Wettling, First World Conference on Photovoltaic Energy Conversion, Hawaii, Dec 5-9 (1994).
[36] L. Witham, D.N. Jamieson, to be published (1997).
[37] J.R. Davis, A. Rohatgi, R.H. Hopkins, P.D. Blais, P. Rai-Choudhury, J.R. McCormick and H.C. Mollenkopt, Impurities in Solar Cells, IEEE Trans. Elec. Dev. ED-27 (4) 337 (1990).
[38] S.A.E. Johansson and J.L Campbell, PIXE: A novel technique for elemental analysis, Wiley, 1988.
[39] S.A.E. Johansson, J.L. Campbell and K.G. Malmqvist, eds., Particle Induced X-ray emission spectroscopy (PIXE), Wiley, 1995.
[40] M.B.H. Breese, Nucl. Instr. Meth. B114 (1996) 25.
[41] M.B.H. Breese, P.J.C. King, G.W. Grime, P.J.M. Smulders, L.E. Seiberling and M.A. Boshart, Phys. Rev. B53 (1996) 8267.
[42] P.J.C. King, M.B.H. Breese, P.R. Wilshaw, P.J.M. Smulders and G.W. Grime, Nucl. Instr. Meth. B99 (1995) 419.
[43] S. Strite, M.E. Lin and H. Morkoç, Thin Solid Films, 231 (1993) 197.
[44] J.D. Hun, S.P. Withrow, C.W. White, R.E. Clausing, L. Heatherly, C. Paul Christensen and N.R. Parikh, Nucl. Instr. Meth. B99 (1995) 602.
[45] D. Dujmic, M. Jakšic, N. Soic, T. Tadic and I. Bogdanovic, Nucl. Instr. Meth. B111 (1996) 126.
[46] P.J.C. King, M.B.H. Breese, D. Meekeson, P.J.M. Smulders, P.R. Wilshaw and G.W. Grime, J. Appl. Phys. 80 (1996) 1.
[47] D.H. Morse, G.S. Bench, S.P.H.T. Freeman and A.E. Pontau, Nucl. Instr. Meth. B99 (1995) 427.
[48] A. Bensaada, R.W. Cochrane, R.A. Masut, R. Leonelli and G. Kajrys, J. of Crystal Growth, 130 (1993) 433.
[49] M.G. Bowler and N.A. Jelly, Physics Letters B 331 (1994) 193.
[50] M.G. Bowler and N.A. Jelly, Zeitschrift Für Physik C 68 (1995) 391.
[51] P.J.C. King, M.B.H. Breese, P.J.M. Smulders, A.J. Wilkinson, G.R. Booker, E.H.C. Parker and G.W. Grime, Appl. Phys. Lett., 67 (1995) 3566.
[52] R. Šandrik and G.W. Grime, Nucl. Instr. Meth. B109 (1996) 675.
[53] S.R. Walker, P.N. Johnston, I.F. Bubb, W.B. Stannard, D.N. Jamieson, S.P. Dooley, D.D. Cohen, N. Dytlewski and J.W. Martin, Nucl. Instr. Meth. B113, (1996) 312.
[54] D.N. Jamieson, S. Prawer, K.W. Nugent and S.P. Dooley, Nucl. Instr. B106 (1995) 641.
[55] P. Gonon, S. Prawer, K.W. Nugent, D.N. Jamieson, J. Appl. Phys., 60, (1996).
[56] A. Saint, M.B.H. Breese and G.J.F. Legge, Rev. Sci. instr. 67 (1996) 2940.
[57] H. Bakhru, E. Nickles and A.A. Haberl, Nucl. Instr. Meth. B99 (1995) 410.
Table 1: Applications of the nuclear microprobe to the study of frontier materials
Stage Nuclear Material References
Important issues Microprobe
Technique
(i) Initial Fabrication
Impurities ERD Mylar, ZrCo alloy [45]
Impurities PIXE CVD diamond, CdTe, Si [8]
Impurities PIXE, IL Synthetic zircons doped [27]
with Gd, Ho
Impurities PIXE, BS W in CVD diamond [9]
Impurities CSTIM Cu precipitates in thin Si [46]
Composition & uniformity PIXE with WDS Stainless steel [47]
Composition & uniformity PIXE GaInP epitaxial layers on [48]
InP
Composition & uniformity PIXE Ba35SO4 radioactive source [49,50]
for neutrino mass
experiments
Defect distribution CSTIM Misfit dislocations in [42]
epitaxial films of Si1-x
Gex on Si
Defect distribution CSTIM Elastic relaxation of [51]
epitaxial films of Si1-x
Gex on Si
Study of phase diagram BS, PIXE Ag-Co thin film alloys [52]
Basic science of CSTIM Planar oscillations in [41]
materials 0.5mm thick Si
Charge transport IBIC & Lateral CVD diamond, CdTe, Si [8,18]
properties IBIC diodes
Charge transport IBIC CVD diamond [19]
properties
Charge transport Lateral IBIC CVD diamond [20,21]
properties
Charge transport Lateral IBIC CVD and natural diamond [22]
properties
(ii) Modification
Ion implantation CCM I implanted Si and GaAs [53]
Ion Implantation & CCM P implanted into diamond [13]
Annealing
Ion beam damage Ion-implantatio Raman spectrum of damaged [54]
n diamond
Resistivity Ion Effect of localised ion [55]
implantation damage on the resistivity
of CVD diamond films
Micromachining & Mixed beams, Ion Irradiation of [56]
analysis IBIC & STIM perspex, circuits
(iii) Device Fabrication
Multilayer deposition & RBS Metallization and [57]
Interconnects superconductor films
Explanation of Acronyms:
(R)BS (Rutherford) Backscattering Spectrometry
CCM Channeling Contrast Microscopy
CSTIM Channeling Scanning Transmission Ion Microscopy
CVD Chemical Vapour Deposition
ERD Elastic Recoil Detection
IBIC Ion Beam Induced Charge
IL Ionoluminescence (Ion Beam Induced Luminescence)
PIXE Particle Induced X-ray Emission
WDS Wavelength Dispersive Spectrometry
Table 2: Properties of diamond, silicon and gallium arsenide at 293 K
Property Diamond Silicon Gallium
Arsenide
Band Gap (eV) 5.5 1.12 1.43
Dielectric strength (Vcm-1) 107 3x105 5x105
Resistivity (-cm) >1011 2.3x105 1.0x108
Electron mobility (cm2V-1s-1) 1800 1350 8500
Hole Mobility (cm2V-1s-1) 1200 480 400
Thermal conductivity (Wm-1K-1) 1000-2000 150 45
Energy to create e-h pair (eV) 13 3.6 4.2
Matrix scattering height* for 1.4 620 45 208
MeV H+
Matrix scattering height* for 2 5.2 18 64
MeV He+
*Detector angle 150o, units Counts.mC-1keV-1msr-1 , the values listed
for GaAs are the average for Ga and As.
Table 3: A comparison of photon, electron, proton and alpha particle beams for the analysis of crystalline silicon solar cells.
Probe Probe Probe Penetration Sensitivity Depth
Resolution Technique Deptha to Trace Information
(micron) (micron) Elements
Blue light (400 ~1 (good) OBIC 0.1 (poor) no no
nm)
Red light (630 ~10 (poor) LBIC 3 (good) no no
nm)
IR (1000 nm) >10 (poor) OBIC 160 no no
(excellent)
20 keV electrons 1 (good) EBIC 1 (poor) see EDS no
20 keV electrons 1 (good) EDS 1 (poor) down to 1 % no
level
3 MeV protons 1 (good) PIXE 90 (excellent) ~ 100 ppm or yes (with BS)
better
3 MeV protons 0.1 IBIC 90 (excellent) see PIXE yesb
(excellent)
2 MeV alphas 0.1 IBIC 8 (good) possible yesb
(excellent)
OBIC Optical beam induced current
LBIC Laser beam induced current
EBIC Electron beam induced current
EDS Energy dispersive spectrometry (electron
induced x-rays)
See table 1 for an explanation of
additional acronyms in this table.
aPenetration depth for the photon probes
(OBIC and LBIC) are from ref. [31].
bWith different beam energies.
FIGURE CAPTIONS
Figure 1: Stages in the development of a frontier material used to classify the applications of nuclear microprobes in table 1.
Figure 2: Channeling angular yield curves from the centre of the laser annealed and regrown regions shown in the CCM images on the right. The CCM images cover and area of 100100 mm2. A 1 MeV H+ beam was used to obtain these data with an X-ray detector and a particle detector at scattering angles of 90o and 150o respectively.
Figure 3: IBIC analysis of CVD diamond films showing the two geometries used. A bias voltage (not shown) is also necessary to collect the charge.
Figure 4: IBIC (a-d) median energy images and IL (e-f) intensity images of CVD diamond films with 500500 mm2 contact pads. All obtained with 2 MeV H+ beams scanned over 1.21.2 mm2 except (c) which used a 2 MeV He+ beam scanned over 200200 mm2. Initially no signals are seen until the bias voltage is increased to 400 V (a). Efficiency improves as the bias is raised to 700 V after which breakdown occurs (b). A close-up reveals high relative efficiency from the grain centres (c). After irradiation, the bias may be lowered to 100 V before the signals disappear (d). IL images from uncovered diamond near the contact pads shows 440 nm luminescence from the centre of the grains (e) and 500 nm luminescence from the centre of the grains and elsewhere (f).
Figure 5: IL spectrum for the sample in Fig. 4 (e-f).
Figure 6: IBIC analysis of polycrystalline silicon solar cells. No bias voltage is necessary to collect signals.
Figure 7: IBIC images of polycrystalline silicon solar cells using a 3 MeV H+ beam scanned over an area of 2x2 mm2. Corresponding energy spectra from several regions of interest are shown in Fig. 8. Windows (in channels) for the energy spectra in figure 8 used to produce the intensity maps in this figure were (a) 2-6 , (b,c) 13-75, (d) 104-151. (e) is a photograph of the same region and the median map (f) shows the same information in (a-d), with a different grey level for each region of interest. In (a-d) white to black indicates high to low efficiency. In (f) white to black indicates high to low efficiency.
Figure 8: Energy spectra (equivalent to the efficiency of the solar cell) from each of the regions of interest from Fig. 7. The peak at channel 300 is from (d) and is probably signal pile-up.
Figure 9: A PIXE spectrum from within a grain of a commercially available solar cell. The Ti signal is from the surface layer, as verified by a BS spectrum (not shown) and maps of the Zn peak revealed a clumpy distribution at the scale of 100 mm. "Si P" and "Si & Ti" designate pile-up peaks.