The determination of absolute oxygen coverage by nuclear reaction analysis

The determination of absolute oxygen coverage by nuclear reaction analysis

Nuclear Instruments North-Holland and Methods THE DETERMINATION REACTION ANALYSIS in Physics Research OF ABSOLUTE B45 (1990) 107-109 OXYGEN CO...

312KB Sizes 0 Downloads 41 Views

Nuclear Instruments North-Holland

and Methods

THE DETERMINATION REACTION ANALYSIS

in Physics

Research

OF ABSOLUTE

B45 (1990) 107-109

OXYGEN

COVERAGE

I.V. MITCHELL, G.R. MASSOUMI, W.N. LENNARD, K. GRIFFITH& S.J. BUSHBY and P.R. NORTON Interface

107

BY NUCLEAR

S.Y. TONG, P.F.A. ALKEMADE,

Science Western, The University of Western Ontario, London, Ontario, N6A 3K7 Canada

The saturation

of a Ni(lOO) surface

with oxygen

is demonstrated

by NRA

for both I60 and l8O. A value of 2.5 monolayers

is

found. The same saturation value is obtained for a surface first converted to the chemisorbed c(2 X 2) phase in “0, then taken to saturation in 160. This simple demonstration suggests that surface reaction mechanisms may be studied through isotopic labelling in -conjuction with NRA methods.

1. Introduction The study of adsorption and chemisorption of molecular gases - H, (D2), N,, CO, CO,, HZ0 (D,O), etc. - at surfaces is an important aspect of basic research into oxidation, corrosion, and catalysis. For a detailed uderstanding of reaction kinetics and surface dynamics, including structural phase changes at singlecrystal surfaces, it is important that the absolute coverage of a sorbate be determined. This is of particular importance when testing models for LEED calculations, for example. It is therefore desirable that the coverage be measured directly rather than inferred as is usually done via surface exposures. The usefulness of methods based upon MeV energy ion beam analysis is readily apparent. An extensive literature on nuclear reaction product energies (spectroscopy) of cross sections and of angular distributions [l] has made it possible to apply nuclear reaction analysis (NRA) techniques to surface interaction studies. Of all sorbate species, oxygen is perhaps the best supported by a variety of suitable nuclear reactions [2] for quantitative coverage measurements with sensitivities extending to one monolayer (ML) and below. This is the region where the most dramatic changes in surface structure occur. The reaction of oxygen with clean nickel surfaces has been the subject of intensive study [2]. The reaction sequence for oxygen with the Ni(lOO) surface, from clean (1 X l), via the c(2 X 2) chemisorbed state to the saturated oxide, has been reported by Alkemade et al. [3] using NRA. A value of 2.4 + 0.2 ML is given for saturated coverage in oxygen of ‘*O nominal 99% enrichment, the coverage being derived from the ‘*O(p, o)15N yield at Ep = 1766 keV. There is good agreement between the values obtained for the c(2 x 2) coverage in “0 reported by this group [3] and in 160 0168-583X/90/$03.50 0 Elsevier Science Publishers B.V. (North-Holland)

reported by Frenken et al. [4] using medium-energy ion scattering, viz. 0.42 _t 0.04 ML and 0.46 f 0.04 ML respectively. We have chosen the Ni(lOO)-0 system as a candidate for measuring the absolute oxygen isotopic composition at saturation for different ‘*O/i60 loading, using the 160(d,p)170 and l*O(p, a)15N nuclear reactions for the assay. In the absence of isotope effects (which are not expected to be significant) this constitutes a very useful test of the NRA methods and anticipates their application with labelled isotopes to surface reaction studies. The experiment was carried out in an UHV chamber connected via a doubly differential pumping system to one of the beam-lines of the UWO 2.5 MV Van de Graaff positive-ion accelerator. Chamber bakeout to 400 K for 12 h allows a base pressure of better than 5 X lo-i1 Torr to be realized. The chamber is equipped with a bakeable gas manifold for admitting gases of interest, a LEED apparatus together with a CMA Auger system, a Kelvin probe and thermal desorption spectroscopy system. Samples are mounted on a 4-axis (x, y, z, 0) manipulator. The sample stage can be cooled to - 82 K via LN, cooling lines. Two bakeable solid-state detectors are mounted in the chamber, one equipped with an Al filter of thickness 8.3 Pm to reduce the flux of elastically scattered particles. A beam of - 2 mm diameter was delivered to the target with currents variable to - 60 nA. Current integration was performed via a rotating Faraday cup, with a calibrated duty cycle, located upstream of the target. The target consisted of a nickel single crystal prepared with a (100) surface. Heating of this surface to 870 K in D, at a pressure of 7 X lop7 Torr for 15 min results in removal of all surface oxygen, i.e. less than 0.1 ML (1 ML = 1.61 X 1015 atoms/cm’) as measured by I. NRA

108

I. L? Mitchell et al. / Determination

of absolute oxygen coverage

NRA data as a function of 1602 exposure at 293 K. The loading procedure described for the mixed layer should therefore produce - 0.4 ML of 180 and - 2.1 ML of 160.

2. Results

0.0

J

0

50

100 Exposure

150

200

250

(Langmuirs)

Fig. 1. Oxygen (160) coverage exposure of a Ni(lOO)surface at 293 K. Coverageshave been determinedby 160(d,p)‘70 nuclear reaction yield measurements.

the oxygen KLL Auger signal and by the appearance of a characteristic and sharp (1 x 1) LEED pattern. Reference samples of I60 and “0 anodically grown oxides of tantalum were mounted to one side of the nickel specimen. These samples could be brought into the target geometry by appropriate rotations and translations of the goniometer. These oxide targets consisted of a 26.7 nm thick oxide (1.43 x 1017 160 atoms/cm2) and an oxide grown in 180-enriched electrolyte on top of an oxide of natural isotopic composition (3.99 x 1017 “0 atoms/cm2 and 1.85 X 1Ol7 160 atoms/cm*). The anodic standards were previously calibrated via 160(d,p)170 and ‘*O(p, a) ‘sN nuclear reaction yields at 972 keV and 750 keV, respectively, using absolute ( + 3%) “0 and 160 anodic oxide standards supplied from the Universite Paris VII laboratory [5]. Cross-comparison of independently grown anodic 160 standards from CRNL, Chalk River, gave consistency to within + 2% absolute. Three sets of measurements were made. In the first, 1602 gas was admitted to the chamber to allow the surface oxide to saturate. This proceeds via the c(2 x 2)-O phase which appears after an exposure of - 2L (where 1 Langmuir = 10m6 Torrs) at 293 K, to the saturated oxide at an exposure of > 300 L 0,. Measurements of the total oxygen coverage were made using the 160(d,p)170 and “O(p, a)15N reactions on the clean and the 160-saturated surface using both reference oxide standards. Yield ratios are corrected for finite target thickness. The surface oxide was then cleaned. In the second set of measurements, the procedure was repeated for exposures to nominally 97% pure 1802 with the saturated surface being assayed for 160 and “0. Finally, one isotopically mixed surface was produced through exposure of clean Ni(lOO) to 300 L O,, comprising 10 L of “0, followed by 290 L of 1602 both at 293 K with 160 and “0 NRA assays as before. Fig. 1 shows a plot of oxygen coverage derived from the

The nuclear reaction products of interest are readily resolved, free from background interference. A typical spectrum for the l8 O(p, 01)~~ N reaction on a “O-covered Ni(100) surface is shown in fig. 2. The intense elastically scattered proton flux from Ni is heavily attenuated by the 8.3 urn Al filter. The 160(d, p)170 reaction spectrum for I60 analysis is, likewise, free from interferences. Background oxygen (bulk) levels on the cleaned Ni(lOO) surface were always less than 0.09 ML. The value for saturation coverage in 160 is found to be 2.45 + 0.10 ML. The value for “0 saturation is 2.35 ML, the measured 160 content raising the total to 2.49 f 0.08 ML. This implies an enrichment of - 94% for ‘* 0 in the source gas, but within the precision of our measurement it is not in disagreement with the nominal figure of 97% nor with a residual gas analysis figure of 98%. (We cannot exclude the possibility of isotope exchange with e.g. residual water vapour on the inner walls of the gas delivery system or vacuum chamber, but if present this dilution is a small effect). The two-step loading, first with l8 O,, then to saturation with 1602, results in a 2.42 + 0.10 ML saturation value. The measured fraction of 180, 0.53 + 0.04 ML, is consistent with the value we expected, based on the isotope loading procedure we used.

J

l10 -

ESZ6-

0’

042-

Or

I

1

25

I

I

35

,

I

125

CHANNEL

I,,

,

175

,

225

No.

Fig. 2. Spectrum for l’O(p, u)‘~N at E, = 750 keV and 19= 150° for the Ni(lOO) surface saturated (2.5 ML) with oxygen enriched in “0. A filter of 8.3 pin Al had been used to attenuate

the elastic scattering of protons from ordinate scale is logarithmic.

the Ni. The

I. VT Mitchell et al. / Determination

of absolute oxygen coverage

109

3. Discussion

4. Summary

One result of the present study has been to provide an independent measurement of the oxygen content in a saturated Ni(lOO) surface oxide (table 1). This value, 2.5 ML, is in agreement with the value reported by Alkemade et al. [3]. It may be noted that the value we report is obtained by reference to a standard oxide while Alkemade et al. used an absolute value for the “O(p, a)15N cross section. Further, if with the sequential exposure experiment we associate the lx0 coverage with the c(2 x 2) chemisorbed phase, the magnitude 0.53 ML is in reasonable agreement with values determined by Alkemade et al. [3] and Frenken et al. [4] for this phase. It must be pointed out, however, that we did not check for the phase of this surface prior to completing saturation in 160,, and, judged from the data in fig. 1, we may have exceeded the optimum coverage for this phase. Also, this experiment lacks a check of the “0 content in the c(2 X 2) phase prior to oxide completion in 1602_ Despite these imperfections, these results suggest that no exchange reaction between the chemisorbed ix 0 and the gaseous i602 has occurred. Experiments on isotopic exchange to CO, H,O, etc. are amenable to these NRA methods and preliminary work has begun in this laboratory on 0 exchange between physisorbed and surface oxides. A broad class of isotopically labelled surface studies is accessible through NRA nuclear reactions specific to (D, T), (12C, 13C) and (14N, “N). New reaction data are now becoming available for 160(3He,p)1xF and ‘60(3He, a)150 reactions [6] which are useful alternatives to the 160(d,p)170 reaction.

The oxygen content of the saturated oxide on a Ni(lOO) surface has been measured by NRA. The same saturation value, 2.5 ML, is found for 160, for “0 and for (“0 +16 0), where the latter is formed by growing the chemisorbed (2 x 2) phase in “0 and completing oxidation in 160. Applications of NRA methods to surface reaction studies of 1 ML and below are anticipated.

Table 1 Saturated surface coverages of Ni(100) exposed to (160,, lx 0,) at 293 K

Clean 160 sat. “0 sat. 160+ l8 sat.

I60 [ML] a)

“0

[ML] a)

i 0.09 2.45 + 0.10 0.14 + 0.02 1.89 +0.09

< 0.02 2.35 i 0.08 0.53 + 0.04

Total 0 [ML] a) < 0.1 2.45 f 0.10 2.49 * 0.08 2.42 * 0.10

The authors wish to acknowledge the generosity and the advice of C. Ortega and G. Amsel of the Univerisite Paris VII. This work was supported with funding from the Natural Sciences and Engineering Research Council, Canada.

Postscript Our attention has been drawn to an independent measurement of the 160( 3He, a)150 reaction cross section, reported in these Proceedings by F. Abel et al. [7]. Very good agreement is found with the data of ref. [6].

References VI Ion Beam Handbook for Material Analysis, eds. J.W. Mayer and E. Rimini (Academic Press, New York, (1977). PI K. Wandelt, Surf. Sci. Rep. 2 (1982) 1. [31 P.F.A. AIkemade, S. Deckers, F.H.P.M. Habraken and W.F. Van der Weg, Surf. Sci. 189/190 (1987) 161. 141J.W.M. Frenken, R.G. Smeenk and J.F. Van der Veen, Surf. Sci. 135 (1983) 147. [51 G. Amsel, G.P. Nadai, C. Ortega and J. Siejka, Nucl. Instr. and Meth. 149 (1978) 705 and 713. t61 W.N. Lemmrd, S.Y. Tong, I.V. Mitchell and G.R. Massoumi, Nucl. Instr. and Meth. B43 (1989) 187. [71 F. Abel, G. Amsel, E. d’Artemare, C. Ortega, J. Siejka and G. Vizkelethy, these Proceedings (Ninth Int. Conf. on Ion Beam Analysis, Kingston, Ontario, Canada, 1989) Nucl. Instr. and Meth. B45 (1990) 100.

“) 1 monolayer = 1 ML = 1.61 X 1Ol5 cm-*.

I. NRA