CLUSTERS AS PROJECTILES FOR SIMS

(1)

HAL Id: jpa-00229413

https://hal.archives-ouvertes.fr/jpa-00229413

Submitted on 1 Jan 1989

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are published 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.

CLUSTERS AS PROJECTILES FOR SIMS

M. Blain, E. Schweikert, E. da Silveira

To cite this version:

M. Blain, E. Schweikert, E. da Silveira. CLUSTERS AS PROJECTILES FOR SIMS. Journal de Physique Colloques, 1989, 50 (C2), pp.C2-85-C2-92. �10.1051/jphyscol:1989216�. �jpa-00229413�

(2)

JOURNAL DE PHYSIQUE

Colloque C2, supplhment au n02, Tome ^50, f b v r i e r 1989

CLUSTERS AS PROJECTILES FOR SIMS

M.G. BLAIN, E.A. SCHWEIKERT and E.F. DA SILVEIRA

Center for Chemical Characterization, Texas A and M University, College Station, TX 77843-3144, U.S.A.

RdsumC Des agregats polyatomiques et ions moliculaires ont i t 6 produits par desorption induite par des fragments de fission. Aprks accCleration i des energies de quelques keV, ces agregats ont it6 utilisis

ir

leur tour c o m e projectiles pour induire des emissions d'ions secondaires. Des spectres de masse en temps de vol indiquent pour des ions secondaires positifs emis ii partir de LiF, une possibiliti d'effets chirniques dans l e processus de desorption. Les possibilitis d'utilisation de la source d'agregats dicrite sont evidentes. Des rendements rChaussis en ions secondaues peuvent itre escomptis, les rendements experirnentau restent cependant ir determine.

Abstract Atomic and molecular cluster ions have been generated by fission fragment induced desorption and then accelerated to keV energies for use as projectiles inducing secondary ion emission. Time of flight mass spectra of positive secondary ions emitted from LiF are presented which show possible chemical effects in the desorption process. Although experimental verification of enhanced secondary ion yields remain to be completed, the potential benefits of using this type of cluster ion source are apparent.

Introduction

During the past few years there has been an increasing interest in the chemistry and physics of cluster induced secondary ion emission. The energies of the clusters range from keV's to MeV's. The motivation for these studies is application and under- standing of the enhanced secondary ion yields observed for both fast and slow incident cluster ions. Fast ions are those which have velocities greater than the Bohr velocity

( Y , = 0.22 cmlnsec) whereas slow ions refer to those with velocities well below v,.

The increased yields using clusters, as compared with single atom ions at the same velocity, persist independent of the velocity range/1,2,3/. The nature of the clusters include polyatomic hydrocarbons, metal dirners and trirners, and very large hydrogen clusters.

An important aspect in the study of cluster induced desorption is the ability to generate and control clusters of well defined composition and energy. In practice, the clusters are typically in an ionized state prior to interaction with the surface.

The two most common methods of producing beams of slow cluster ions are adiabatic expansion of a vapor into a vacuum after which the neutral clusters are ionized (usually by electrons) and direct formation of cluster ions by laser induced desorption of a solid sample/4/. In addition, electron impact/5/ and cold cathode161 ionization sources have been used. It would be desirable, however, to have a single ion source which is fle&le enough to provide a variety of cluster ion types for use as primary projectiles.

The purpose of this paper is to report on some preliminary results obtained using MeV, heavy ion induced desorption as a means of producing slow organic and inorganic cluster ions which are then used as projectiles.

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1989216

(3)

JOURNAL DE PHYSIQUE

E x p e r i m e n t a l

The schematic of the experimental arrangement is shown in Figure 1 and is analo- gous to that used by I<amenski, et a1./7/ for comparing fission fragment and keV single atom ion desorption. The set-up .is a modification of a standard particle desorption mass spectrometer (PDMS)/8/ consisting of a 2 5 2 ~ f fission fragment source, a "start"

detector, a target, a field-free region, and a "stop" detector. The associated electronics determine the time differences between the arrival of one of the complementary fission fragments at the start detector and the arrival of desorbed target ions at the stop detector, thereby allowing mass analysis. By replacing the stop detector in this standard set-up with a sample surface a t 45 degrees and adding a second field free region (with the stop detector now a t its end), one now observes ion emission from the sample due to impact by species desorbed by the fission fragment. We refer to the source of clust.er ions as "source", the surface a t which the cluster ions strike as "sample" and the ions desorbed by keV atomic and cluster primary ions as "secondary" ions.

Experiments were performed at 2 x 1 0 - ~ torr in a vacuum system having a base ptessure of 5x10-~ torr. The target used was either a LiF(200) single crystal or LiF vapor deposited on a Ni foil. The cluster sources were polystyrene, NaC1, CsC1, and CaF2, on aluminized mylar. Total primary ion (i.e. cluster and atomic ion) flux on the sample surface was 1-100 ions per second giving acquisition times which ranged & o m several to over 1 2 hours. Because the (time) correlation of detected positive secondary ions is with the fission fragment induced desorption process in the source, the resulting mass spectrum is a convolution of each spectrum from the sample due to each impact by a primary ion (which is just one in a train of ions from the ion source). That is, the location of any peak in this "total" mass spectrum is the result of the sum of two different flight times: that of the primary ion along the first field region and that of the secondary ion along the second field free leg. For peak assignment, the fight times of the secondary ions (as observed in a separately calibrated spectrum of the sample) are added to the flight times of the primary ions. This results in an estimate of the total flight t h e for a number of primary ion/secondary ion combinations. By

Stalt CEMA

Sample Stop CEMA

Gate & Delay

1

FH-~

^PDP^11/23

Figure 1. Modified particle desorption mass spectrometer (PDMS) set-up with a sample surface and second time-of-flight leg added.

(4)

then matching these estimated total flight times with the those associated with the peaks in the 'Ltotal" mass spectrum, peak assignment is achieved. The assignment can then be verified by changing the accelerating voltages in the mass spectrometer and monitoring the changes in peak positions. Where ambiguity occurs, the peak is either labelled with two or more sources (if more than one combination of primary and secondary ion is contributing) or is left unassigned (if no apparent combination contributes to the peak.)

Results a n d Discussion

Classification and Properties of PDMS Clusters as Projectiles

Cluster ions produced by heavy ion bombardment of a solid (in this case, by 252Cf fission fragments) can be of several different types but are generally in the form of ionized heteroclusters. These consist of one or more atoms (or molecules) with an atom and/or an electron removed (or a proton attached) in the case of a positive cluster, or with an atom removed and an electron attached in the case of a negative cluster. The type a1 clusters observed are of the form (MX)Mt or (MX)X- for alkali- halides, (MX2)*%Xt and (MX,)?X- alkaline earth-halides, and (M+H)t or (M-H)- (plus fragment ions generally of the type C:'- and C,H:*-) for organic, hydrocarbon based (bio)molecules. Quasi-molecular h e r ions and complex ions (for example (H20),HS/9/ and organic dyes/lO/) are also observed. Of course, the two most commonly observed simple cluster ions in the PDMS technique are

H$

and H,S/9/.

Because of the +bundance of different types of clusters which can be generated via plasma desorption, a knowledge of their basic differences becomes important for their use as projectiles. The &st important cvnsideration for use of a cluster as a projectile is its binding energy. Consequently, molecular and fragment ions are referred to a3 clusters since their translational energies as projectiles, ZlkeV in this work, are orders of magnitude larger than their binding energies, typically a few eV's/4/. In this sense, the binding energy is unimportant. However, it is important if the cluster is to remain intact until impact with the target. At the other extreme of very low impact energies (on the order of 10's and 100's of eV's) the binding energy becomes an increasingly larger portion of the total energy available upon collision and is again of some importance.

Employing the classification given by Mark/4/, of the clusters which are commonly seen with the PDMS technique all of them (except quasi-molecular dirners) are either ionic clusters with average (neutral) ionic bond energies of about 2 to 4 eV or valence clusters having conventional chemical bond energies of about 1 to 4 eV. Typically not seen with PDMS are van der Waals, molecular, and hydrogen- bonded clusters because of their low average binding energies ( l e v maximum) and metallic clusters because of the electronic nature of the desorption process.

The second potentially important factor, considering the actual impact and desorption process being induced, is the number of constituent atoms per cluster ion. The number of particles per cluster for a hydrocarbon molecular ion versus an alkali-halide cluster at the same approximate mass can easily be larger by-an order of magnitude.

If

coherence effects are present for desorption induced by slow ions, as has been exper- imentally suggested for fast clusters/2/ and those with velocities around v , / l / , then the ability to control and vary the particle number via cluster type would be useful.

(5)

C2-88 JOURNAL DE PHYSIQUE

Valence Clusters as Projectiles

Valence clusters have some interesting possibilities as desorption projectiles since, as noted, both high mass, intact quasimolecular ions and lower mass, simple carbon clusters or hydrocarbon fragments are easily produced by electronic energy deposition in an insulating material. For example, a good source of positive and negative

C,

and

C,H,

ions is a polystyrene solution spun cast on aluminized mylar and then exposed to fission fragments. The trend for the negative carbon cluster ions follows the well known odd-even abundance rule established for carbon clusters obtained with laser desorption methods/ll,l2/ which shows relative maxima at n-even. However, for positive carbon clusters, there is no well defined odd-even trend. This is not expected if stability and electronic structure are considered/l3/ and consequently, it is unclear as to what controls their formation. However, in the context of using the clusters as desorption probes the mechanism of cluster formation is not so much the issue as are the clues as to whether or not the species will be stable enough to stay intact until collision.

The best approach would be to understand what fraction of the species of interest is metastable, as has been demonstrated with the time-of-flight technique/l4,15/.

The positive ion desorption spectrum of LiF due to bombarding negative ion species from polystyrene is shown in Figure 2. The secondary ion emission of Lit due to impact at 2 keV of the first several C,H- and C, clusters can be seen to the left of

'a.

⁰⁰

CHANNELS + 10

'

Figure 2. Positive ion desorption spectrum of LiF due to incident negative carbon and hydrocarbon clusters at 2keV impact energy.

(6)

the large, broad "high background" area which dominates the spectrum. This high intensity background area is probably due to LiF desorption by higher mass (and poorly resolved) C, and C,H- species (n>8), decayed metastables, and uncorrelated electrons and ions. Desorption induced by the high energy electrons can also be seen.

The fact that discernable peaks are present in the spectra at bombarding energies as low as 2keV and that no sharp threshold for peak appearance was noted suggests the possibility of chemical effects at work in the desorption process/l6/.

Ionic Clusters as Projectiles

The formation of ionic clusters, in particular alkali-halides, with the PDMS technique is well documented (see, for example, 1171.) Positive ion spectra fiom an LiF(200) single crystal with CsCl and NaCl as sources for clusters are shown in Figure 3. The high background region is again present; however the appearance of cluster and atomic ion induced peaks in the spectrum is apparent. An interesting observation is the presence of some peaks which are assignable only if one assumes a desorption product which is a Li cluster containing an element which is also a component of the bombarding species. See for example the (LiF)CsS peak in Figure 3a. The two possible ways which this could occur are either ejection of a preformed (LiF)Cs+ species which already exists on the surface or incorporation of one of the incident Cs atoms upon collision of the cluster with the surface. Given the very low total fluence on the sample and the bombarding conditions, the of the latter ocurring may not be so remote. For the CsCl spectrum, the bombarding energy is 4 keV and the impact angle is 75 degrees from the surface normal. The bombarding conditions in the NaCl spectrum give an impact energy of 7.6 keV and an impact angle of 60 degrees.

A second type of source for ionic clusters is calcium floride. The CaF2 induced desorption spectrum of LiF is shown in Figure 4. For this spectrum, the impact energy of the primary ions was 2 keV and the impact angle was 75 degrees. There is an abundance of peaks in the entire mass spectrum (not shown), many of which could not be assigned due to uncertainty. However, the presence of ~i~ ~ e a k s due to the major CaF species is apparent. A more sophisticated experimental approach will be required to elucidate the other peaks in the spectrum.

Although it is true that the binding energy of LiF is rather low, as in the case with the hydrocarbon species no apparent sharp thresholds for the onset of sputtering were observed with any of the three sources of ionic clusters. This, coupled with the number of distinct, positive ion peaks for the incident ionic species (with some incorporating projectile species) tends to suggest the possibility of chemical effects in the collision/desorption process.

Because positive secondary ion spectra were acquired, the number of incident primary ions per secondary ion peak could not be determined. For generating quantitative yield measurements of positive secondary ions, a true start signal (possibly from light emission upon impact of the cluster) is necessary. Of course negative ion yields should be measurable by using electrons as the start signal, as has been shown/2/. This would allow the experiment to be performed in an event-by-event fashion.

(7)

C2-90 JOURNAL DE PHYSIQUE

0

0 1000 2000 3000 4000 5000 6000 7000 8000 CHANNELS

U

4L

oo ibo. oo I o. oo nbo. oo 2ho. oo ebo. 00

:HhNNELs 1

o

Figure 3. Positive ion desorption spectra of LiF(2OO) with incident positive alkali-halide atomic and cluster ions. a) CsCl cluster ion source. Potentials:

source +8500V, target +3300V. b) NaCl cluster ion source. Potentials: source +lOkV, target +2400V.

(8)

80.00 120.00 1ko. 00 CHANNELS * 10

'

Figure 4. Positive ion desorption spectra of LiF due to positive CaFz atomic and cluster ions. Source +3kV, sample +lkV.

Conclusion

The feasibility of performing a SIMS experiment/analysis with clusters generated by fission fragment induced desorption is demonstrated with the first spectra presented here. The ion source described is simple, yet produces a variety of atomic, molecular and cluster ions. Moreover, different projectiles causing secondary ion emission can be compared simultaneously without changing the experimental conditions. The use of a laser in the cluster ion source using the same approach could have some advantages.

Many of the positive secondary ion spectra have an abundance of well defined peaks indicating the potentially large gain in information by using a variety of primary ions. Peak assignment, however, (using the method described) becomes prohibitively complex for ion sources and samples with more than just a few major mass l i e s . This points out the necessity of knowing when th'e primary ion/sample surface cbllision occurs. Once this is accomplished, the tailoring of the cluster in terms of size and translational energy, coupled with the ease of directing and focusing keV projectiles, could be well suited to applications in microanalysis.

Acknowledgements: Support for this work in part by the National Science Foun- dation Grant INT-8602288 is gratefully acknowledged.

(9)

JOURNAL DE PHYSIQUE

References

/1/ J.P. Thomas, P.E. Filpus-Luyckx, M. Fallavier, a n d E.A. Schweikert, Phys. Rev.

Lett., 55(1985)103.

/ 2 / M. Salehpour, D.L. Fishel, J.E. Hunt, Int. J. Mass Spec. a n d Ion Proc., 84(1988)103.

/ 3 / H.H. Anderson and H.L. Bay, Rad. Eff., 19(1973)91.

/4/ T.D. Mark, Int. J . Mass Spec. a n d Ion Proc., 79(1987)1.

/5/ W . Reuter, Anal. Chem., 59(1987)2081.

/ 6 / T. Ast, Md.A. M a b u d , a n d R.G. Cooks, Int. J. Mass Spec. and Ion Proc., 82(1988)131.

/7/ I. Kamenski, P. Hakansson, B. Sundqvist, C.J. McNeal, R.D. Macfarlane, Nucl.

Inst. Meth., 198(1982)65.

/8/ W.R. Summers and E.A. Schweikert, Rev. of Sci. Inst. 57(1986)692.

/9/ E.F. d a Silveira, E.A. Schweikert, J. Chem. Phys., i n press (1988).

/lo/ J.A. Bennett, L. Van Vaeck, F. Adams, E.A. Schweikert, Analyt. Chirn. Acta, in press(1988).

/11/

G.

Seifert, S. Becker, a n d H.-J. Dietze, Int. J. Mass Spec. a n d Ion Proc., 84(1988)121.

/ 1 2 / N. Furstenau, F. Hillenkamp, a n d R. Nitsche, Int. J. Mass Spec. a n d Ion Proc., 31(1979)85.

/13/ K.S. Pitzer a n d E. Clementi, J. Am. Chem. Soc., 81(1959)4477.

/14/ W. Ens,

R.

Beavis, G. Bolbach, D. Main, B. Schueler, a n d 1C.G. Standing, Nucl.

Instr. Meth. A , 245(1986)146.

/15/ S. Della-Negra a n d Y. Le Beyec, Anal. Chem., 57(1985)2035.

/ 1 6 / J. Roth, in: Sputtering by Particle Bombardment 11, ed., R. Behrisch (Springer, Berlin, 1983).

/17/ W.R. Summers,

M.U.D.

Beug-Deeb, E.A. Schweikert, Anal. Chem., Anal.

Chem., 60(1988)1944.