Structural and electrical characterisation of semiconductor materials using a nuclear microprobe

Beam Interactions with Materials 8 Atoms

ELSEVIER

Nuckar

Instruments

and Methods

in Physics Research

B 136 138 (1998) l-13

Structural and electrical characterisation of semiconductor materials using a nuclear microprobe

Abstract High-resolution imaging techniques (submicron) have not traditionally been the domain of MeV ions. Instead, this domain is occupied by a vast array of techniques that utilise scanned low-energy ion beams (keV ion microprobe), electrons (transmission or scanning electron microscopy), light (near field microscopy) or all variants of scanning probe microscopies. Techniques that utilise focused X-rays are also developing. Now, with a nuclear microprobe, high-energy ion beams have begun to enter this domain, bringing a range of unique techniques for making images. With a MeV ion beam focused to a probe smaller than 1 pm, conventional techniques like Rutherford (and non-Rutherford) backscattering spectrometry and particle induced X-ray emission may be used to image and analyse the composition and structure of microscopic regions of inhomogeneous specimens. New ion beam analysis techniques have been developed specifically for use with microbeams. These include ion beam induced charge for imaging charge transport characteristics in materials to ion microtomography for making non-destructive 3-D maps of mass density and elemental composition. Another novel technique is ionoluminescence, which may be used to map various electronic properties of the material. Presented here are some examples of these imaging techniques in a wide variety of semiconductor materials including diamond, HgCdTe, polycrystalline silicon and a mineral semiconductor crystal: pyrite. In all these examples, the specimens display structural inhomogeneities on the scale of 10 pm. making it essential to employ a focused beam. Present limitations to taking imaging with focused MeV ions into the nanometre resolution domain are also discussed. 0 1998 Elsevier Science B.V.

PACS: 01.30.R: 61.83; 41.75.A; 61.72 K~~+~ords; Nuclear microprobe; Nuclear microscopy; charge; Crystal

defects; Trace elements;

Charge

Semiconductor

crystals; Ion channeling;

Ion beam induced

trapping

1. Introduction Problems abound at the micrometer scale in both nature and technology. So far these problems

’ Tel.: +61 (0)3 9344 5376; fax: +61 (0)3 9347 4783: e-mail: [email protected]. 0168-583X/98/$19.00 0 1998 Elsevier Science B.V. All rights reserved. PIISO168-583X(97)00657-5

have been mainly addressed by analytical techniques involving electron beams, photons and fine probes. With a focused beam, ion beam analysis also has a role in addressing these problems. While ion beam analysis, through Rutherford Backscattering Spectrometry (RBS), has excellent depth resolution on a scale well matched to problems in semiconductor device technology, the

problem of spatial resolution is lagging well behind. The traditional role for ion beam analysis has been in areas where lack of spatial resolution is not a serious problem. The areas include: problems involving thin films, interfacial reactions, regrowth of amorphous layers, diffusion of dopants, phenomena involving ion implantation, measurement of surface contaminants, and so on. In short, phenomena in the “semiconductor industry where thin-film and ion implantation problems abound” [l]. However use of an unfocused beam in these situations imposes stringent demands on the specimen. In fact, for a 1 mm beam spot, “if a range of depth analysis is 2000 A, the width of the beam spot is a factor of 5000 larger than the thickness of the layers” [l]. So “the lateral uniformity of the sample must therefore be assured on the surface as well as in depth” [l]. This is a very stringent requirement. It is the purpose of this paper to review the technology and applications of focused ion beams where the width of the beam spot is reduced to 5 urn or smaller. This is now less than ?5 times the thickness of the representative 2000 A layer. This review first briefly discusses the history and technology of focused MeV ion beams, then presents some applications to the analysis of a wide variety of semiconductor materials. A sequence of earlier reviews of this expanding field is also available [2-lo] and complete details of all analytical techniques discussed here can be found in a recent book [I 11.

2. History of focused beams A history of the development of the nuclear microprobe has recently been compiled by Legge [ 121. Briefly, the development of focused MeV ion beams arose out of work at Harwell, UK, where collimated beams were used to make positional measurements. The low current available made for long measurement times and limited spatial resolution. Cookson and his co-workers drew on theoretical work on strong focusing lens systems designed for objective lenses for MeV electron microscopes [13] and they constructed the first MeV proton microprobe system. This used standard

beam-line magnetic quadrupole lenses that allowed a 20 urn probe to be obtained with a practical beam current. Later, purpose-built magnetic quadrupole lenses, with superior machining tolerances and better mounting arrangements to allow more precise lens alignment, reduced this to 3 urn [14]. Many early nuclear microprobe systems were built after this time, mostly involving lenses of large bore diameter (-50 mm). The importance of a large demagnification factor in the production of small probes was soon recognised. The maximum practical pole tip field attainable was around 3 kG, so the lenses became smaller, with bore diameters N 15 mm. This allowed the lens system to be made more compact with the consequence of increased demagnification and smaller probes. At present the Oxford high excitation triplet, with a demagnification of about 80 and a probe current of 100 pA has demonstrated comfortably submicron resolution of around 300 nm [15]. Of the 50 laboratories in the world equipped with focusing lenses, most use magnetic lenses, although one uses electrostatic quadrupole lenses and several use superconducting solenoids.

3. Nuclear microprobe analytical techniques

system

configuration

and

The typical nuclear microprobe configuration is shown in Figs. 1 and 2, with the physical parameters of the Melbourne system listed in Table 1. It differs from the standard set-up for ion beam analysis in the presence of the object and aperture collimators, the probe forming lens system, probe system and the more elaborately scanning equipped specimen chamber which includes an optical microscope for positioning the region of interest under the beam. In these respects the system resembles a scanning electron microscope. The focused beam is scanned over the sample and the yield of induced or scattered particles is used to form an image. Analogues of the transmission electron microscope, have been proposed [ 161, but practical difficulties have not seen the development of such an instrument. This is most likely because of the fact that the deBroglie wavelength of a MeV ion beam is small so that the rich phenomena

D. N. Jumieson I Nutl. hstr.

urld Meth. in Phys. Rer. B 136-138 (1998)

Fig. 1. Schematic diagram of the essential features of a nuclear microprobe system (not to scale). A: Beam steerer. B: Object collimators. C: Aperture collimators. D: X-ray detector, E: backscattered particle detector array. F: Probe forming lens system, G: Optical microscope, H: vacuum chamber. I: Ion pumps, J: sample goniometer, K: Beam scanner, L: low vibration girder, S: Specimen location. Physical parameters of the Melbourne system (MPZ) appear in Table 1.

Fig. 2. Photograph of the probe forming lens system and sample chamber of the Melbourne nuclear microprobe system (MP2).

associated with diffraction effects are generally absent and that the scattering cross sections are relatively small compared to those for electron scattering. The techniques applied to the analysis of semiconductors with a nuclear microprobe may be divided into two categories: high current techniques and low current techniques. The high current techniques are those of traditional Ion Beam Analysis: Rutherford Backscattering, Parti-

I-13

3

cle induced X-ray Emission, Ion channeling, and, less frequently, Nuclear Reaction Analysis. The extension of these traditional ion beam analysis techniques into the micrometre spatial resolution domain with a nuclear microprobe is shown in Fig. 3. This figure also shows, approximately, the elemental sensitivity and spatial resolution of various complimentary analytical techniques for semiconductors that involve electrons, photons and neutrons. The figure is drawn from data provided by Charles Evans and Associates [17] and also includes minimum detectable limits for neutron activation analysis [l&19], inductively coupled plasma mass spectrometry [ 18,2%23] and nuclear reaction analysis [24,25,27]. In addition, more efficient techniques, discussed below, have been developed that can extend the resolution of the nuclear micro-

Fig. 3. The extension of Ion Beam Analysis into the micrometre and sub-micrometre regime by use of a nuclear microprobe. Adapted. with permission. from the home page of Charles Evans and Associates. [17]. Acronyms: AFM: Atomic Force MiAES: Auger Electron Spectrometry, ESCA: croscopy, Electron Spectrometry for Chemical Analysis. EDS: Energy Dispersive Spectrometry. IBA: Ion Beam Analysis, IBIC: Ion Beam Induced Charge. ICPMS: Inductively Coupled Plasma Mass Spectrometry, NAA: Neutron Activation Analysis, NMP: Nuclear Microprobe (MeV Ion Microprobe). NRA: Nuclear Reaction Analysis. PIXE: Particle Induced X-ray Emission, RBS: Rutherford Backscattering Spectrometry. SEM: Scanning Electron Microscopy, SIMS: Secondary Ion Mass Spectrometry. STM: Scanning Tunneling Microscopy, TEM: Transmission Electron Microscopy. ToF: Time of Flight, XPS: X-ray Photoelectron Spectroscopy. XRF: X-ray Fluorescence. Additional techniques most often applied with a nuclear microprobe system and frequently cited in previous Nuclear Microprobe Technology and Applications Conference Proceedings: CCM: Channeling contrast Microscopy (which can also include (C)STIM: (Channeling) Scanning Transmission Ion Microscopy) and IL: Ionoluminescence.

4

Physical

D.N. Jumirson / Nucl. instr. und Mrth. in Phys. Rrs. B 136-138 (19981 1.~13

parameters

of the Melbourne

nuclear

Accelerator Object diaphragm Aperture diaphragm Object - aperture distance Aperture specimen distance Aperture scan coils distance Probe forming lens system Lens bore radius Turns per pole Lens 1 and 4 length Lens 2 and 3 length Lens spacing Working distance (last lens to specimen). Demagnificdtion at shortest WD X-ray detector (A, 9. 0) Particle detectors (A. R, I))

Typical

parametera

microprobe

system (MP2)

WD

for analysis

Object diameter Aperture diameter Beam convergence angle Beam current for 7 MeV He’ Beam current for 3 MeV H+

probe into the submicrometre domain. For convenience, acronyms associated with all these techniques are listed in the caption of Fig. 3. The major difference when the traditional ion beam analysis techniques are employed with a nuclear microprobe is that the total beam current in the focused probe is typically from 5 nA to 100 pA for probe sizes from 10 to 1 pm in diameter. Consequently, the total integrated charge used is typically of the order of 1 PC for a reasonable scan time up to a few hours with detector solid angles up to 25 msr. This is about two orders of magnitude less integrated charge than the situation for unfocused beams used for high sensitivity analysis in the examples presented by Maisch et al. [28] who also used detectors with solid angles between 10 and 150 msr. Hence the strength of the nuclear microprobe is mainly in the ability to map structures. However high sensitivity, high resolution, trace element analysis is also possible on robust samples. The traditional ion beam analysis techniques are mainly aimed at analysis of the elemental

NEC 5U pelletron with RF ion source 5 200 pm diameter SO pm to 3 mm diameter 7250 mm I200 mm 300 mm Russian antisymmetric quadruplet of magnetic quadrupole lenses 6mm I5 30 mm 60 mm 35 mm 150 to 500 mm 16 Si(Li) 30 mm?. 33 msr. 135” PIPS 450 mm:. 126 msr: 135” PIPS 100 mm’, 30 msr: 145” Annular 300 mm:. 97 msr: 175” Additional forward detectors as required High current 100 pm I mm 0.16” 700 pA 360 pA

techniques

Low current

techniques

5 pm 50 pm 0.013” -2 fA -5 l-A

structure of the sample. However in most semiconductor applications the electrical characteristics are of primary importance. These characteristics can be measured by the other major collection of ion beam analysis techniques employed with a nuclear microprobe. These typically involve low beam currents, that is less than 100 pA down to single particles. Each of these low current techniques is usually applied with a nuclear microprobe rather than a broad beam. One of the most widely applied is ion beam induced charge, where the charge induced in a sample by the beam is collected and regions of differing collection efficiency form contrast in images. The sample itself becomes the detector and so every beam particle provides information. Consequently, the number of incident beam particles required to form images is considerably reduced compared to traditional techniques, allowing the diameter of the beam collimators to be reduced with a consequent improvement in resolution, often well into the submicron domain.

D.N. Jamieson I Nucl. Instr. and Merh. in PhFs. Rex B 136-138 (19981 l-13

Although the beam current available for focused beams is lower than that generally used for unfocused beams for traditional ion beam analysis, the current density is considerably higher. For 100 pA focused in to a 1 urn probe, the current density is more than 1000 times higher compared to a notional 40 nA unfocused beam 1 mm in diameter. It is therefore of importance to understand the effect of the intense focused probe on the damage in the sample under analysis. Several stages in the build up of damage with increasing dose are shown in Table 2. Although the electrical characteristics of the crystal are rapidly changed by ion beam irradiation. the consequences for ion channeling are far less stringent. The first effect seen in the channeling spectrum from Si or C is dechanneling as a result of crystal swelling from the effects of the ion implantation. This dechanneling is first seen around the border of a scan, with a width of about 30 urn. This becomes visible when the edges of the scanned region tilt beyond the channeling critical angle. An example of the effects observed in diamond is shown in Fig. 4. In this case a (100) oriented diamond slab was implanted with 5 x 1017 He/cm? and the implanted region was then analysed with a 1.4 MeV H beam in the channeling orientation. The channeling contrast image clearly shows the dechanneling from the swollen edges of the original implant as well as a signal from the trapped He at the end of range of the original implant. Analysis of the peak area of the He signal allows the amount of trapped He to be

2.5x1 0170.6. MeV He/cm*

Fig. 4. Channeling contrast images of the build-up of damage in a nuclear microprobe irradiated region in a diamond. An initial implant of 2.5 x 10” 0.6 MeV He/cm’ was done into a 150 x 150 pm square region which was then imaged with a 200 x 200 pm square scan with a 2.25 MeV H beam in the channeling orientation. The image shows the dechanneling from the edges of the original scan region.

measured. This was used to study the cracking of the diamond as a result of the He implant, since this allowed the He to escape [38]. In summary, the main requirements of a nuclear microprobe are: (i) a stable accelerator with beam spread preferably less than 100 eV per MeV and beam brightness above 0.1 pA/(nm’ mrad’ MeV) [39], (ii) a beam scanning system that

Table 2 Onset of damage

with microprobe

irradiation

Dose

Sample

Single ion 2 x 10” 2 MeV H/cm’ 5 x lO’5 2 MeV He/cm? 3.7 x 10lh 1.4 MeV H/cm?

Semiconductor microelectronic Si devices GaAs crystal P implanted diamond

4 8 2 5 5

x x x x x

10’h lOI 10” IO” IO’*

2 MeV He/cm’ 1.4 MeV H/cm’ 2 MeV He/cm’ 2 MeV He/cm’ 0.4 MeV He/cm’

GaAs crystal Diamond Silicon crystal Silicon, diamond Silicon

devices

Effect

References

Single event upset, current transient Decrease in charge pulse height Dechanneling from lattice damage 15’%~improvement in channeling by ion beam annealing Swelling and dechanneling Swelling and dechanneling Swelling and dechanneling Dechanneling from lattice damage Ablation of surface

[6.1 I.291 (30-331 ]341 ]351 ]341 ]351 [341 ]361 ]371

6

D.N. Jumicwm I Nwl. Instr. and Mcth. in Phys. Rr.r. B 136 13X

can provide 10-1000 urn scans, (iii) a precision probe forming lens system with less than 0.5% parasitic multipole field contamination and a demagnification of 20 or more, (iv) a vacuum system of better that 1 x 1O-7 Torr to minimise beam halo, (v) a low vibration mounting system with an environment free from electromagnetic fields and (vi) a data acquisition system able to log the (x,y,E) coordinates from one or more detectors at count rates in excess of 5 kHz per detector. These requirements have now been met in the more than 40 laboratories worldwide equipped with nuclear microprobe systems. In these laboratories, approximately 25% of the work is directed to the study of problems involving semiconductor materials [40].

(149811 13

Optical image

CCM image Depth (ml)

4. Applications The examples of applications presented here are divided into two categories as described above: high current and low current techniques. The high current techniques are mainly applied to the study of sample structure. The low current techniques are mainly applied to the study of sample electrical characteristics. 4. I. High current techniques One of the first applications of the nuclear microprobe to the analysis of semiconductor materials was to the study of laser annealing and epitaxial regrowth of amorphous silicon layers. Fig. 5 shows a channeling contrast image of a 50 urn laser spot annealed region in a specimen that was amorphised and then implanted with In into the amorphous layer. Since each event from the backscattered particle detector is tagged with the (XJ) coordinates, the RBS spectrum from the inside of the laser spot could be extracted and compared with the spectrum from the as-implanted material, as also shown in Fig. 5. These spectra revealed that most of the amorphous layer regrew epitaxially, but a residual thin polycrystalline layer formed which contained most of the In following annealing [41]. An extension of the method involved the extraction of the energy spectra from

Fig. 5. Laser spot annealed homoepitaxy silicon images with an early application of channeling contrast microscopy. Top left: optical micrograph of a 50 mm laser spot annealed amorphous layer in Si. Lower left: Channeling contrast image of same region. Top right: Spectra from within the centrc of a laser spot. W designates the widow used to produce the image on lower left. Lower right: The In signal from within the laser spot region. Images and spectra obtained with a 2 MeV He beam.

annuli around the laser spot showing decreasing regrowth as a function of distance from the centre because of decreasing temperature [2]. Another application of the nuclear micropobe to the study of epitaxial growth was the growth of HgCdTe on GaAs. In this case both PIXE and RBS were used. The growth can involve the formation of growth defects in the form of 10 urn wide pyramids. Channeling contrast images of these pyramids revealed increased induced M shell X-ray signals from Hg compared to induced L shell X-ray signals from Cd and Te, as shown in Fig. 6. This suggested that excess Hg was involved in their formation. Additional growth defects were also observed, not visible optically, that appeared to have nucleated misaligned material from scratches in the original surface of the GaAs substrates 1431. Epitaxial growth of diamond on large area synthetic diamond substrates is important to many potential applications in the microelectronics device industry. A method for the production of

D. N. Jumieson I Nucl. Insrr. wui Meth. in Phys. Rex B 136-138

(199X/ I-13

mercury cadmium telluride grown on GaAs substrate imaged with a 7 MeV He ’ beam. CCM Fig. 6I. Growth defects in heteroepitaxial backscattering signal showing thicker hillocks. CCM surface: deep: image from energy window at lower energy end of Cd+Te+Hg image from top 500 nm showing defects, CdTeL: L X-rays. HgM: M X-rays showing enhanced yield from Hg over Cd and Te.

large area single crystal has been devised by Pehrsson et al. [43J. It involves 175 keV C implantation into cheap synthetic diamond substrates to create a shallow damaged layer buried beneath a relatively undamaged surface. Chemical vapour deposition (CVD) is then used to grow thick (>40 urn) homoepitaxial films on the surface. Finally chemical etching is used to attack the buried layer resulting in the lift off of a large area slab. A key question is whether the implantation has an effect on the quality of the homoepitaxial layer. Also, numerous growth defects are evident in the surface of the as-grown layers. A photomicrograph of a growth defect is shown in Fig. 7(a) along with channeling contrast images of this defect that revealed its cause. A spectrum from the film grown on the unimplanted areas of the substrate was compared to a spectrum from the film grown on the implanted areas (Fig. 7(b)). This revealed that the implantation did not change the quality of the film. A spectrum extracted from within the loop of the growth defect showed no signal from the bur-

ied damaged layer. This suggests that this region was masked from the original implant, perhaps by surface dust. The edges of the masked region are tilted due to implantation induced swelling and hence nucleated misaligned crystal during the CVD growth phase. It is interesting that this appears to be associated with a linear defect extending across the film. In the case of many polycrystalline materials, lattice location studies of dopants is of importance. However the size of the grains makes conventional ion beam analysis impossible unless the grains are all oriented in the same direction and it is possible to channel simultaneously into many grains. With a focused microprobe, many classes of polycrystalline material are accessible to study. In the case of pyrite (Fe&, a semiconducting mineral), epitaxial growth in the natural environment plays a role in the formation of the mineral crystals. These can be as large as several centimetres in size and incorporate a range of lattice substituted elements. An important element often found in association with

D.N. Jumicwt~ I Nucl. lwtr.

and Mrth. in Phys. Res. B 136. 13X (1498) l-13

Surface Yield

c/c >

C: Misaligned growth B: CVD growth on implant

Energy (keV1 Fig. 7. (a): Growth defect in homoepitaxial ion implanted diamond imaged with 1.4MeV H channeling contrast microscopy. Top row: channeling contrast microscopy images from masked edge of implanted region. Bottom row: channeling contrast microscopy image of loop defect. (b): The energy spectra extracted from various regions of interest from the sample in (a). A detector of solid angle 25 msr at a scattering angle of 150” and a deposited charge of 0.44 PC were used to obtain these data.

D.N. Jumieson I Nucl. Instr. and Meth. in Phy.

Fig. 8. Semiconductor mineral pyrite crystal Images are 500 x 500 urn square.

Rex B 136-138 (1998) l-13

images with 3 MeV H- induced

pyrite is Au. Fig. 8 shows elemental maps from a 300 urn wide crystal of pyrite from the Emperor mine in Fiji, obtained by proton induced X-ray emission with a 3 MeV proton beam. The zoning of As, Cu, Au, Pb and MO is striking. The concentration of these elements in the centre of the crystal ranged from 0.19% for Au down to 150 ppm for Pb and 1 ppm for Zn. Ion channeling can be used to probe the lattice location of Au in pyrite crystals. With a 2 MeV He microbeam it was possible to perform channeling angular scans from the centre of the crystal allowing the nearest high order axis to the surface normal to be identified. As shown by Fig. 9, this revealed that more than 35% of the Au was occupying lattice substituted sites in the crystal. This information can be used to distinguish between different models for the growth of these ores [44]. This technique could obviously be extended to synthetic polycrystalline materials. 4.2. Low current techniques

1.2

emission

.

,

.

showing

,

.

the zoning

(

.

,

of trace elements.

(

1 .o $ F

8 r: 0.6

0.4 -3

Presented here are some examples of electrical characterisation of several different semiconductor materials and devices. In each case the sample itself is the detector and the microprobe is used to induce charge which is then collected. Polycrystalline silicon in photovoltaic cells offers interesting challenges to nuclear microprobe

X-ray

-2

-1

0

1

2

3

Tilt angle (degrees) Fig. 9. Crystal orientation of 300 urn crystal of Fe& probed with I.4 MeV He+ channeling contrast microscopy. Top left: Optical image of single crystal insitu in a quartz matrix. Top right: channeling angular scan image showing the symmetry of the crystal planes about the highest order axis closest in orientation to the normal. Bottom: channeling angular yield curves from the backscattered particles.

analysis [45]. Induced charge is collected owing to the presence of a depletion layer at the p-n junction. Grain boundaries and other defects trap charge reducing the efficiency of the device. The efficiency of the cell can be imaged with IBIC as shown in Fig. 10. In addition to the obvious grain boundaries, additional structure is visible within individual grains that appears to be correlated with surface structure visible in SEM images. A potential advantage of this work with MeV ion beams is that the range of a MeV ion is comparable to longer wavelengths of light in single crystal silicon. In addition PIXE has sufficient sensitivity to detect some trace contaminants that perhaps contribute to the degradation of the charge collection efficiency [46]. Diamond has been proposed for many possible applications in microelectronic devices. particularly as a radiation detector. High quality polycrystalline diamond films manufactured by the Norton company have been the subject of extensive study by IBIC (see list of references in [lo]). This work has been done to investigate the charge transport properties of the material by using the focused ion beam to create electron-hole pairs within the material which are then collected by biased electrodes applied to the diamond surface (the “energy signal”). On av-

sutfac? contact

1000

erage, the diamond films are about 8% efficient. with localised regions in the centres of grains rising to an efficiency of 24% [47]. A problem with the application of IBIC to diamond is that the ion beam creates space charge within the highly insulating diamond. This causes a significant amount of hysteresis in the signals collected from the sample. An example of this is shown in Fig. 11. Here three electrodes have been deposited on the surface of a Norton diamond. As the bias voltage is increased, the energy signal increases as the charge collection efficiency increases. Just before electrical breakdown, signals from the centres of the diamond grains show a peak in the energy spectrum. Decreasing the bias voltage allows energy signals to be collected for lower voltages than are possible from an initially un-irradiated sample. High-resolution IBIC images from within single diamond grains reveal significant structure as shown in Fig. 12. Analysis of spectra extracted from these different regions will allow insights to be gained into the charge collection mechanisms as a function of diamond growth direction. The charge collection efficiency from the centres of the most efficient grain in this specimen was about 5 times greater than the average efficiency of the whole film.

old ,scan

pm2

Fig. 10. Grain boundaries in a polycrystalline MeV He’ beam was used for these images.

500 pm2

silicon solar cell imaged

by ion beam induced

250 pm2

charge

at increasing

magnification.

A 2

I Energy Channel

I

Fig. I I. Charge collection efficiency mapped in a Norton diamond film by ? McV He’ ion beam induced charge. The hysteresis in the energy spectra from this material is evident from the differing bias voltages applied for each image.

5. Future of microbeams A variety of new techniques are emerging for new imaging methods employing focused MeV ion beams. One of these exploits the deeply penetration of the beam to form tomographic images of microscopic objects smaller than the range of the ion beam. For example, a microcrystal of diamond was probed with a 3 MeV H beam transmitted through a crystal and detected. From the energy loss as a function of (s,_t,) and the rotation angle of the sample, a three dimensional map of the density variations within the crystal could be obtained [48]. This provided insights into the growth mechanism of the crystal. An extension of this technique involves PIXE tomography [49] where the characteristic X-rays are detected instead of the energy loss of the transmitted ions. This will potentially allow three-dimensional maps of trace elements in microscopic objects to be obtained. For the future, it would be desirable to have much brighter ion sources to allow smaller probes to be obtained with a similar beam current to those available now. Apart from this. further advances

225pm2

increasing charge collection efficiency Fig. I?. Extreme close up image of a single diamond grain in a polycrystalline diamond tilm image with 3 MeV H ’ ion beam induced charge. Left to right: images of increasing charge collection efficiency. The images show structure within the single gmin. The grain is also visible in the top right of the large scan (215 pm’).

D.N. Jamieson

12

I Nucl. Instr. and Meth. in Phys. Rex B 136-138 (1998) l-13

will come from probe forming lens systems with even greater demagnification compared to the present systems that are more closely matched to the characterisitics of the ion source and accelerator where further optimisation is possible. At present the main barrier to further reduction in probe size appears to be mainly mechanical vibrations and parasitic magnetic fields which limit the resolution to 100 nm. Also, more sophisticated instrumentation would be valuable to more fully exploit limited beam time. As shown by the many applications here, the ability to make images opens provides a large range of valuable information. In fact a key common factor in all of these applications presented here is that crucial information about the structure and electrical characteristics of the samples would have been lost had unfocused beams been employed.

Acknowledgements

I am grateful to the organisers of IBA-13 for inviting me to present this paper. The work at Melbourne has been supported by the Australian Research Council and the University of Melbourne and the Australian Department of Science, Industry and Tourism. I acknowledge useful discussions with Steven Prawer, contributions to Fig. 4 by Julius Orwa, Fig. 5 by Jeff McCallum and Jim Williams, Fig. 6 by Sean Dooley and Patrick Leech, Fig. 7 by Steven Prawer and Jim Butler, Figs. 8 and 9 by Jacinta den Besten and Chris Ryan, Fig. 10 by Lachlan Witham, Fig. 11 by Andrew Saint and Deborah Beckman and Fig. 12 by Mark Breese.

References [I] W.-K. Chu, J.W. Mayer, [2] [3] [4] [5] [6]

M.-A. Nicolet. Backscattering Spectrometry, Academic Press, New York, 1978. J.S. Williams, J.C. McCallum, R.A. Brow, Nucl. Instr. and Meth. B 30 (1988) 480. M. Takai, K. Hirai. K. Ishibashi, A. Kinomura, S. Namba. Nucl. Instr. and Meth. B 54 (1991) 209. M. Takai, R. Mimura, H. Sawaragi, R. Aihara, Scanning Microscopy 7 (1993) 815. D.N. Jamieson, M.B.H. Breese, A. Saint, Nucl. Instr. and Meth. B 85 (1994) 676. M. Takai, Nucl. Instr. and Meth. B 85 (1994) 664.

I71 D.N. Jamieson, Nucl. Instr. and Meth. B 104 (1995) 533. PI F.Watt. G.W. Grime Eds., Principles and Applications of

t91 PO1

Ull

[121

P31 [I41

High Energy Ion Microbeams, Adam Hilger, Bristol. 1987. M. Takai, Nucl. lnstr. and Meth. B I 13 (1996) 330. D.N. Jamieson. Proceedings of the Fifth International Conference on Nuclear Microprobe Technology and Applications, Santa Fe, NM. USA. 1996 (Nucl. Instr. and Meth. B 130 (1997) 706). M.B.H. Breese, D.N. Jamieson. P.J.C. King, Materials Analysis with a Nuclear Microprobe. Wiley, New York, 1996. G.J.F. Legge, Proceedings of the Fifth International Conference on Nuclear Microprobe Technology and Applications, Santa Fe. NM, USA. 1996 (Nucl. Instr. and Meth. B 130 (1997) 9). A.D. Dymnikov. S.Ya. Yavor. Sov. Phys.-Tech. Phys. 8 (1964) 639. J.A. Cookson, A.T.G. Piling, J. Radioanal. Chem 12

(1972) 639. Proceedings of the Fifth International Conference on Nuclear Microprobe Technology and Applications. Santa Fe, NM, USA. 1996 (Nucl. Instr. and Meth. B 130 (1997) I). [I61 S.S. Klein, P.H.A. Mutsaers. Nucl. Instr. and Meth. B 30 (1988) 349. [I71 Charles Evans and Associates web page. http://www.cea.-

s51 F. Watt,

corn.

[I81 S.J. Parry, J. Anal. Chem. 51 (1996) 1164. P91 M.M. Musa, W.M. Markus. A.A. Elghondi,

R.H. Etwir. A.H. Hannan, E.A. Arfa, J. Radioanalytical Chemistry & Nuclear Chemistry 198 (1995) 17. PO1 J.K. Aggarwal. M.B. Shabani, M.R. Palmer, K.V. Ragnarsdottir, Anal. Chem. 68 (1996) 4418. VI M.D. Norman, N.J. Pearson, A. Sharma, W.L. Griffin, Geostandards Newsletter 20 (1996) 247. PI G.S. Walker, M.J. Ridd. G.J. Brunskill, Rapid Comm. Mass Spec. 10 (1996) 96. [231 W. Hub. H. Amphlet. Fresenius J. Analytical Chem. 350 (1994) 587.

P41 R. Samlenski,

C. Haug, R. Brenn. C. Wild, R. Lecher. P. Koidl. Appl. Phys. Lett. 67 (19) (1995) 2798. P51 J. Krauser. F. Wulf, D. Braunig. J. Non-Cryst. Solids 187 (1995) 264. F. Rauch, Nucl. Instr. and WI D. Endisch, H. Sturmand. Meth. B 84 (1994) 380. ~271 G.J.F. Legge. in: J.R. Bird, J.S. Williams (Eds.). Microprobe Analysis, Ion Beams for Materials Analysis, ch. 10. Academic Press, New York. 1989. P81 Th. Maisch, V. Schule. R. Gunzler, P. Oberschachtsiek, M. Weiser. S. Jans, K. Izsak, S. Kalbitzer. Nucl. Instr. and Meth. B 50 (1990) 1. 1291 F.W. Sexton. IEEE Trans. Nucl. Sci. 43 (1996) 687. [301 M.B.H. Breese. P.J.C. King, G.W. Grime. F. Watt, J. Appl. Phys. 72 (6) (1992) 2097. [31] M.B.H. Breese, G.W. Grime, F. Watt. Nucl. Instr. and Meth. B 77 (1993) 301.

[32] M.B.H. [33] [34] [35] [36] [37] [38] [39]

[40]

Breese. C.H. Sow. D.N. Jamieson.

F. Watt. Nucl.

Instr. and Meth. B 85 (1994) 790. M.B.H. Breese. J.S. Laird, G.R. Moloney, A. Saint, D.N. Jamieson. Appl. Phys. Lett. 64 (15) (1994) 1962. S.P. Dooley. D.N. Jamieson, Nucl. Instr. and Meth. B 66 (1992) 369. S.P. Dooley, D.N. Jamieson. S. Prawer. Nucl. Instr. and Meth. B 77 (1993) 484. M. Takai. Scanning Microscopy 6 (1992) 147. M. Takai, K. Hirai, K. Ishibashi. A. Kinomura. S. Namba. Nucl. Instr. and Meth. B 54 (1991) 209. J. Orwa. D.N. Jamieson. K.W. Nugent. S. Prawer, R. Kalish, Nucl. Instr. and Meth. B 124 (1997) 515. R. Szymanski, D.N. Jamieson. Proceedings of the Fifth International Conference on Nuclear Microprobe Technology and Applications, Santa Fe. NM. USA. 1996 (Nucl. Instr. and Meth. B 130 (1997) 80). Proceedings of the Fifth International Conference on Nuclear Microprobe Technology and Applications. Santa Fe, NM, USA, 1996 (Nucl. Instr. and Meth. B 130 (1997)).

[41] J.C. McCallum.

R.A.

Brown.

E. Nygren,

J.S. Williams.

G.L. Olson. Mater. Res. Sot. Symp. Proc. 69 ( 1986) 305. [42] P.W. Leech, L.C. Witham, S.P. Dooley. D.N. Jamieson. J. Vat. Sci. Tech. A 13 (1995) 21. [43] P.E. Pehrsson, T. McCormick. W.B. Alexander, M. Marchywka, D. Black, J.E. Butler, S. Prawer, Mater. Res. Sot. Symp. Proc. 416 (1996) 61. [44] J. den Besten. D.N. Jamieson. C.G. Ryan. submitted to J. Appl. Phys., 1997. [45] C. Donolato. R. Nipoti, D. Govoni. G.P. Egeni. V. Rudello. P. Rossi, Materials Science & Engineering B 42 (1996) 306310. [46] L.C. Witham, D.N. Jamieson. R.C. Bardos. these proceedings. [47] C. Manfredotti. F. Fizzotti. P. Polesello. E. Vittone, M. Jaksic. I. Bogdanovic. S. Fazinic. Mat. Res. Sot. Proc. 416 (1996) 193. [48] S.A. Stuart. M. Cholewa. A. Saint, S. Prawer. G.J.F. Legge, D. Weirup. Nucl. Inst. and Meth. B 77 (1993) 234. [49] A. Saint, M. Cholewa. G.J.F. Legge. Nucl. Instr. and Meth. B 75 (1993) 504.