ANGULAR RESOLVED ENERGY ANALYSIS OF 69Ga+ IONS FROM A GALLIUM LIQUID METAL ION SOURCE

HAL Id: jpa-00226835

https://hal.archives-ouvertes.fr/jpa-00226835

Submitted on 1 Jan 1987

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

ANGULAR RESOLVED ENERGY ANALYSIS OF 69Ga+ IONS FROM A GALLIUM LIQUID METAL

ION SOURCE

P. Marriott

To cite this version:

P. Marriott. ANGULAR RESOLVED ENERGY ANALYSIS OF 69Ga+ IONS FROM A GALLIUM LIQUID METAL ION SOURCE. Journal de Physique Colloques, 1987, 48 (C6), pp.C6-189-C6-194.

�10.1051/jphyscol:1987631�. �jpa-00226835�

ANGULAR RESOLVED ENERGY ANALYSIS OF 6 9 ~ a + IONS FROM A GALLIUM LIQUID METAL ION SOURCE

P. Marriott

Materials and Surface Science Group, Materials Development Division, Harwell Laboratory, United Kingdom Atomic Energy Authority, Didcot, Oxon, OX1 1 ORA, U.K.

Abstract - An analysis system has been designed and built to characterise liquid metal ion source beams. Both mass and angular resolved energy distribution measurements can be made, from which both FWHM energy spreads and energy deficits can be obtained. This paper briefly describes the system and presents and discusses the first off-axis results taken with a gallium liquid metal ion source.

Liquid metal ion (LMI) sources are finding increasing applications in the semiconductor and microanalysis fields. Despite considerable source development much is still to be understood about the ionisation mechanisms and complex internal beam interaction processes occurring. A system for characterising LnI source beams has been developed to study these phenomena.

I - ^{m E} ANALYSIS SYsm

The system consists of a 180° electrostatic hemispherical energy analyser and retarding lens coupled with a quadrupole mass filter. Ions are detected using either a Faraday cup or a channeltron electron multiplier. Angular variation is achieved by rotating the ion source about an axis at the needle tip (the ionisation region). The source and analysers are held in a UHV chamber routinely held at

.r. 1 x mbar; a schematic diagram is shown in figure 1.

The LMI source is supported in a purpose built stage attached to an XIZ translator.

This translator enables the source to be positioned on the axis of the retarding lens and is used to rotate both the source and the extraction electrode with respect to the analyser system. The stage holding the source was designed to facilitate alignment of the source needle with the axis of rotation of the translator and to position accurately the extractor electrode with respect to the source. Extractor electrode alignment was important as the emitted beam angle was strongly affected by the axial alignment of the needle tip with the electrode aperture. Alignment of both source needle and extractor electrode could be made to % 20 p.

Angular alignment of the needle with the axis of the retarding lens was made t o

%

0.3O and the extractor electrode was aligned perpendicular to this axis to < 0 . 8 ' . The emitted ion beam axis was aligned with the axis of the retarding lens using the rotational variation provided on the translator. A beam angular misalignment in the plane perpendicular to this rotational plane of % ' 3 was estimated from a set of angular intensity distributions. The entrance aperture of the retarding lens when used for LMI source measurements was 0.1 mm in diameter, allowing a cone of ions with a 0. lo half angle to enter the lens.

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1987631

C6-190 JOURNAL DE _PHYSIQUE

The purpose built stage also supported a tungsten wire heater which was electrically isolated from the source needle, and a Chromel-Alumel thermo- couple in contact with the source reservoir case. A gallium LMI source on loan from 1.B.T.-Dubilier Ltd. was used for the measurements presented in this paper.

The performance of the energy analyser and retarding lens was monitored using two indium oxide thermal ionisation sources, as previously described 111.

This enabled the change in transmission of the lens to be estimated, the energy resolution of the analyser to be measured under various operating conditions and the voltage deficit scale to be established. No change in transmission of .the lens was observed when operating under conditions similar to those used for LMI source energy spread measurements. For the following results, the analyser resolution was

< 0.4 eV for most energy distributions

taken and < 0.9 eV for some selected measurements above 10 VA beam current.

LE/E for the analyser was measured to be 0.018.

Foradoy Electron

m u l t o p l ~ ~ r High vacuum

/

I

' 1 '

Pump

Fig. 1 Sectional schematic diagram of the LMI source analysis system.

For all the following measurements made with this analyser system the 69 Dalton isotope of gallium was used and the source reservoir was maintained at room temperature.

11 - ^RESULTS

On axis, over 70 energy distribution measurements were made between 0.26 and 40 PA.

The FWHM measurements from a subset of these spectra were compared /I/ with those published by Mair et al. 121, being in very close agreement, and confirming the change in FWHM behaviour at beam currents below

2 p A . The peak position of ~ a * energy distributions has been found to vary with emitted beam current by many workers, but there are great differences between published results. The peak position shift with beam current observed with this analyser system is in very good agreement with values reported by Mair et al. 131, these two data sets being the only two even closely related. These comparisons have shown that the analyser system was well behaved, giving reproducible results largely in agreement with the detailed measurements of Mair et al.

Energy distribution measurements were made at 11 different beam currents off-axis, allowing measurement of FWHM and voltage deficit values. Us g the areas under the energy distributions, the relative intensity of the emitted " G a ' species could be plotted against the source angular rotation, as is shown in figure 2(a). The shapes of these angular distribution profiles are similar to those reported by Swanson et al. /4/ and are in good agreement with the prediction of the model of Kingham and Swanson 151.

Off-axis, remarkable variation in the FWHM and voltage deficit behaviour was

observed, as shown in figures 2(b) and 3. Above

1.8 PA, an increase in FWHM can

be seen with increasing angle, followed by a decrease at higher emitted currents. A

voltage deficit shift occurred at angles just greater than those where the decrease

in FWHM occurred. FWHM changes up to 15 eV and deficit shifts of over 8 V were

found at 14 PA.

- 3 1

-2 0 2 L 6 8

El

Angle l d e ~ l e e s l

(a)

Angle Iaegrresl (b)

Fig. 3 - Voltage deficit (peak position) versus angle of rotation at (a) 0.55 to 6.0 VA and (b) 6.0 to 14 PA beam currents.

These variations in FWHM and deficit can be seen to be the result of energy

distribution peak shape changes and shifts with angle. Figures 4(a) and (b) show

the progression of change in the distributions at 10 PA, where the spectra are shown

with normalised intensities; figure 4(c) shows the same distributions plotted with

their relative intensities. At angles < 16O, there was little change in the

C6-192 JOURNAL DE PHYSIQUE

distributions. At larger angles there was a relative increase in the low energy ion intensity and a smaller increase in the high energy intensity. A long low energy tail can be seen at angles up to

20°. At 22O, where the FWHM is at its greatest, this tail begins to disappear, (figure 4(a)). At larger angles the low energy intensity reduced, thus reducing the FWHM, the peak position remaining invariant at angles < 2 5 ' . At 2S0, the peak position began to shift to higher ion energy. This shift onset coincided with the peak shape becoming Gaussian, the asymmetry present at all smaller angles having disappeared.

At angles > 24O, most of the ions emitted

G . 1 > 4

-

^-

- ..*

had energies in excess of that attainable "...,. - ^sP .-,.

from the electrostatic field alone. From -.

c-s

figure 4(c) it can be seen that the number --

- - - > I

of ions emitted at such angles was small.

It can also be seen that there was an

increase in the absolute intensity of low c

energy ions emitted at % 18-22O over those

emitted on axis. Similar effects were seen at other beam currents. The long low c

energy tail was only observed at currents above

7 lA and the FWHM and deficit variations described above occurred at

- 2 8 - 1 8 B I 8 2 8 1 8 4 B

slightly different angles at the different

beam currents. > <.

I11 - DISCUSSION

The remarkable changes in FWHM and voltage deficit show no correlation or anti- correlation with the angular intensity profiles of figure 2(a). This is in contrast to the suggestion of Papadopoulos et al. /6/

that AU+ FWHM followed an anticorrelation with the angular intensity distributions.

For AU+ ions, a reduction in FWHn was observed with increasing angle, reaching a minimum at 15-20° off-axis, and showing some signs of increasing again at greater angles.

The minima were observed at similar angles at the 4 beam currents at which the

experiment was performed. However the small maxima in the angular intensity profiles occurred at

8, 10, 18 and 20° at 2, 4, 10 and 20 pA respectively. Due to this angular discrepancy, this suggestion of anti- correlation by Papadopoulos et al. cannot be justified.

The energy distribution peak shape changes off-axis result in the FWHM variations shown in figure 2(b). The on-axis distribution shapes show an asymmetry, and this appears to increase with angle and then decrease so that the distributions become Gaussian. An understanding of the processes causing the on-axis asymmetry would appear to be necessary before an explanation of the off- axis changes can be offered.

The on-axis distribution shapes recorded with this analyser system are similar to those reported by Swanson et al. /4/, Anazawa

Fig. 4 Energy distributions at

10 PA beam current at

various source angular

rotations. (a) and (b) with

normalised intensities, (c)

with relative intensities.

reasons for the asynanetry in the shape have not been satisfactorily explained. The asyrmPetry has been observed in distributions recorded at beam currents below 1 pA in the present work, and up to many tens of FA. At beam currents above

10 FA, a long low energy tail also becomes apparent for G a ' ; it appears independently of the peak asymmetry and some tens of volts below the peak position.

Processes which could result in ions of greater deficit are those of charge exchange collisions, the creation of low energy ions away from the emitter surface as a result of droplet fragmentation, and free space field ionisation of neutral atoms.

Neutral gallium atoms are very unlikely to be released by thermal evaporation in significant quantities / 9 / , particularly at low beam currents, but they might be produced by the break-up of charged droplets.

Charge exchange collisions, in which a neutral atom gains the charge from a moving ion can result in an ion with a voltage deficit. For neutral gallium atoms to participate in this process, the collisions must occur away from the high field region, otherwise free space field ionisation would occur and there would be no neutral atoms available. From computer modelling of the field strength in the vicinity of a gallium LMI source emitter by Castilho and Kingham /lo/, including space charge fffects, at 5 FA emission current the field strength does not drop below 10 Vnm- within 3 nm of the surface, showing that neutral atoms are unlikely to exist at distances much closer than this. As such, charge exchange collisions will produce ions with deficits of some tens of volts.

The creation of the low energy molecular ions observed in caesium and gold LMI sources, for example, has been thought to be due to the break-up of larger charged ions or charged droplets /11,12,13/. From the ion energy deficits of these molecular species, ions or droplets must have travelled some tens of A from the emitter before fragmentation. Monomer ions produced directly from this process, or neutral atoms released in this way and then field ionised, must also exhibit energy deficits similar to the low energy molecular ion peaks, some tens of volts in deficit, and in the case of gallium, > 30 V. In other metal sources, the higher energy molecular ion peaks observed might be due to 'overcharged' droplet

fragmentation at the emitter surface, as postulated by Joyes and Van de Walle /13/, but for gallium there is no evidenqe of this, all molecular ions having tens of volts deficit.

Kuk et al. /14/ and Vankatesan et al. 1151 measured photon yields which increased rapidly with ion current, indicating that collisional processes became significant only at high currents. Droplet emission was also found to increase rapidly when beam currents of some tens of pA were drawn from gallium /16,17,18/ and indium sources /19/, so any processes involving the break-up of charged droplets will also be more notable at these higher currents.

It is therefore reasonable to conclude that the low energy tail seen at high currents is due to one or more of the processes given above, but that these processes cannot account for the asymmetry of the Ga* energy distribution peak.

The peak asymmetry at present cannot be explained. Whatever processes cause the asymmetry on-axis are presumably also involved in the peak shape changes off-axis.

and hence the FWHM changes. These processes might also be responsible for the notable peak shape changes observed with gallium and galliwindium alloy sources at elevated temperatures /20,8/. Understanding the cause of these peak shape changes would greatly increase our knowledge of the emission and beam interaction processes in liquid metal ion sources. A further useful experiment would involve measuring energy distributions off-axis of ions emitted from a heated gallium source.

The large shift in peak position of the distributions at angles > 20' raises the

question from where do such off-axis ions gain their energy? An interaction process

involving the high space-charge region on and near the beam axis is proposed 1211,

C6-194 JOURNAL DE PHYSIQUE

whereby the ions off-axis are repelled further from the beam axis than would be the case in the absence of space-charge. These ions gain kinetic energy in this interaction from the ions near the beam axis, the electrostatic field then further accelerating the ions in a direction more parallel to the axis. The energy lost from the axial ions is small and would not easily be detected given the

uncertainties in the appearance energy calculations.

IV - ^SUMMARY

An analysis system has been built capable of being used for angular and mass

resolved energy distribution measurements, enabling voltage deficits to be recorded, and as such represents a unique facility. On-axis distribution measurements with a gallium LMI source were in excellent agreement with other workers. Remarkable peak shape variations were observed off-axis. Processes which could cause these

variations have been considered, but no satisfactory mechanism found. The deficit shifts observed off-axis have been attributed to a beam interaction process involving the high space-charge region on and near the beam axis.

V - ACKNOWLEDGEMENTS

I am grateful to R. Bartram for technical assistance. Thanks also to

Dr. R.G. Forbes '&d Dr. G.L.R. Mair for helpful discussions and for their interest in the work, and to Dr. P.D. Prewett for supplying the ion source. Supervision by Prof. D.G. Armour and Dr. J.C. Rivihre is gratefully acknowledged. This work has been carried out as part of a collaboration between the University of Salford and Harwell Laboratory.

REFERENCES

/1/ Marriott, P., J. Phys. D: Appl. Phys. 19 (1986) L115-L119.

/2/ Mair, G.L.R., Grindrod, D.C., Mousa, M.S. and Latham, R.V., J. Phys. D.: Appl.

Phys. 16 (1983) L209-L213.

/3/ Mair, G.L.R. and Mulvey, T., Scanning Electron Microscopy (1985) 111 959-971.

/4/ Swanson, L.W., Schwind, G.A., Bell, A.E. and Brady, J.E., J. Vac. Sci. Technol., 16(6) (1979) 1864-1867.

- /5/ Kingham, D.R. and Swanson, L.W., Appl. Phys. A34 (1984) 123-132.

/6/ Papadopoulos, S., Barr, D., Brown, W.L. and Wagner, A., J. Physique 45 C9 (1984) 217-222.

/7/ Anazawa, N., Aihara, R., Okunuki, M. and Shinizu, R., Scanning Electron Microscopy (1982) IV 1443-1451.

/8/ Ishitani, T., Umemura, K. and Tamura, H., Jap. J. Appl. Phys. g ( 6 ) (1985) L451-L454.

/9/ Kingham, D.R. and Swanson, L.W., Appl. Phys. (1986) 157-169.

/lo/ de Castilho, C.M.C. and Kingham, D.R., J. Phys. D: Appl. Phys. 2 (1986) 147-156.

/11/ Helm, H. and Moller, R., Rev. Sci. Inst. z(7) (1983) 837-840.

/12/ Papadopoulos, S., Barr, D.L. and Brown, W.L., J. Physique 67 C2 (1986) 101-106.

/13/ Joyes, P. and Van de Walle, J., J. Physique (1986) 821-827.

/14/ Kuk, Y. and Sakurai, T., Proc. 27th Int. Field Emission Symp., Tokyo, Japan (1980) 74-80.

/15/ Venkatesan, T., Wagner, A. and Barr, D., Appl. Phys. Lett. 2 (1981) 943-945.

/16/ Mair, G.L.R. and von Engel, A., J. Phys. D: Appl. Phys. 17 (1984) L13-L17.

/17/ Wagner, A., Venkatesan, T., Petroff, P.M. and Barr, D., J. Vac. Sci. Technol.

19(4) (1981) 1186-1189.

- /18/ Thompson, S.P., D. Phil. Thesis, University of Oxford, England (1982).

/19/ Papadopoulos, S., Submitted to J. Phys. D: Appl. Phys. (1986).

/20/ Swanson, L.W., Schwind, G.A. and Bell, A.E., J. Appl. Phys. 51(7) (1980) 3453-3455.

/21/ Private discussions with R.G. Forbes.