Nuclear Instruments and Methods in Physics Research B 85 (1994) 37-41 North-Holland
Beam Interactions with Materials 8 Atoms
Quantification of b~~~~iurn in thin films using proton backscattering J.A. Leavitt *, L.C. McIntyre Jr., R.S. Champlin, J.O. Stoner Jr., 2;. Lin, M.D. Ashbaugh, R.P. CQX,J.D. Frank Department of Physics, University of Arizona, Tucson, AZ 85721, USA
The strong resonance in the ‘H-‘Be backscattering cross section near ‘H laboratory energy 2525 keV has been used to determine 9Be area1 densities in thin films with accuracies of about 6%. We report measured cross sections (for a 170.5” laboratory backseat&ring angle) for ‘H on ‘Be for ‘H energies from 2400 to 2700 keV. The area1 densities of the self-supporting Be foils used for the iH-‘Be measurements were determined from 4He-9Be back~atte~ng measurements. We also report measured cross sections for 170.5” back~attering of 4He by ‘Be for 4He laboratory energies of 500 to 4200 keV. Results of channeling measurements in epitaxial Be films on crystalline substrates are also discussed.
1. Introduction Recent interest in the use of Be as a low-absorption spacer material in multilayer X-ray mirrors [l] has stimulated efforts to develop better ion beam analysis (IBA) techniques for quantifying Be in thin films. Rutherford backscattering (RBS) with 4He beams with energies < 2 MeV may be used to estimate the area1 density of Be in these multilaye~ by comparison of simulated and experimental spectra. The backscattering (BS) signa from the Be itself is usually not observed; Be is added to the simulation until the signals for the other film elements and substrate elements are in agreement on the two spectra. The accuracy of these Be area1 density determinations is difficult to estimate. Energy loss cross sections for 4He in materials  may be known to f5%, but slight amounts of impurity atoms (typically 0) in the Be strongly affect the Be area1 density determinations; signals from these impurities are themselves frequently unobservable. Consequently, we have chosen to use techniques that produce observable signals from Be itself. We considered use of the strong 5.1 MeV nonRutherford resonance in the 411e-9Be cross section , but the energy required is just beyond the convenient range of our accelerator. So we chose to use the 2.55 MeV non-Rutherford resonance in the 1H-9Be cross section (about 40 times Rutherford at tiLAB= 142.4”) . Preliminary experiments on Be films yielded
* Corresponding author, phone + 1602 6216793, fax + 1602 621 4721. Old-583X/93/$07.~ d
strong Be backscattering signals. Since there is disagreement [4,5] concerning the magnitude and energy of the resonance maximum, we felt it was advisable to measure the ‘II-‘Be cross section in the vicinity of the resonance. These cross section measurements required the use of targets with known Be area1 densities. We chose to measure the Be area1 densities of our targets using 4He backscattering; in the process, we measured 4He-9Be cross sections for 4He lab energies 500 to 4200 keV at the laboratory backscattering angle, @,a = 170.5”. In subsequent sections, we report results of the measurements of *H-9Be and 4He-9Be cross sections; we also discuss results of area1 density and proton channeling measurements in epitaxial Be films on crystalline substrates.
2. Experimental Beams of ‘H+ and 4He+ from our 5.5 MV Van de Graaff accelerator were used with a standard backscattering setup  to obtain BS spectra of several self-supporting bilayer films (Be/Al, Be/Au, Be/Mm Arizona Carbon Foil Company, Inc.); also examined were single-crystal (0001) hcpBe films grown on (111)Si and (OOOl)o-Al,O, substrates by molecular beam epitaxy (11. See Table 1 for film characteristics. Relevant experimental parameters were: detector (Ortec BA-1425-200) resolution = 14 keV for 4He, 5 keV for “t-I, solid angle subtended by the detector, $2 = (0.78 f 0.02) msr; backscattering angle = 8,, = 170.5” f 0.5”, with
0 1993 - Elsevier Science B.V. All rights reserved
J.A. Leavitt et al. /Nucl. Instr. and Meth. in Phys. Res. B 85 (1994) 37-41
. . . . .. .. :
Fig. 1. A typical 4He backscattering spectrum of the self-supporting bilayer BeAu 1, showing C and 0 impurities on both sides of the film.
BS angles in the range 170.5” f 0.9” accepted by the detector. The beam energy was uncertain by less than 0.08%; the beam energy spread was less than 2 keV . 2.1. 4He-9Be
cross section measurement
Fig. 1 shows a typical 4He BS spectrum of the Be/Au bilayer BeAu 1 with clean Be and Au peaks. The 4He-9Be cross section, [email protected]
,,), at 4He laboratory  in terms of the energy, EBe, may be expressed Rutherford cross section, uR(EBe), at the same energy, and measured quantities by
~~e(&d/41u(LJ u(E,c) ‘+dEne)= 4,(W/KdE~u)
Fig. 2. Measured 170.5” cross sections, (T, for 4He scattered by ‘Be for 4He laboratory energies between 500 and 3000 keV expressed in terms of the Rutherford cross sections ou, at the same energies (typical uncertainties +3%). The solid circles are our data; the open triangles are data from ref.  at 0= 145.4”.
by comparison with the Rutherford 4He-Au cross section at the same energy. Results of the 4He-9Be cross section measurements are plotted in Figs. 2 and 3 and compared with previous measurements [3,9]. Justification for the assignment of u/us = 1 to the E, region 500-1300 keV is provided by the data shown in Fig. 4. Comparison [lo] of the heights of BS signals corresponding to backscat-
F(E.dF(Eid F(%) ’
Alpha Lab Energy (keV)
where A&&E;,) and Ai, (Ei,) are integrated Be and Au BS peak counts obtained from a BS spectrum taken with incident 4He beam energy ET (where scattering from both elements is Rutherford; EL, and EAu are mean 4He energies in the Be and Au layers), and ABe(EBe), AAu(EAu), E, and EAu are corresponding quantities for incident beam energy E. It will be assumed that except for small corrections  due to the electron shells (the F factors), the 4He-9Be and 4HeAu cross sections are Rutherford for beam energies 500-1300 keV, justification for this assumption will be provided later. Use of Eq. (1) effectively allows determination of the non-Rutherford 4He-9Be cross section
-?= b 6
Alpha Lab Energy (keV)
Fig. 3. Measured 170.5” cross sections for 4He scattered by ‘Be for 4He energies between 3000 and 4200 keV (uncertainties - *3%). The dashed curve represents data from ref.  for 0, = 157.3”.
The ’ W BS spectrum of BeAu 1 shown in Fig. 5 exhibits peaks due to elastic scattering of protons from Be, Au, C and 0 but also peaks due to products (a, d and 6Li) af proton-induced nuclear reactions in the Be 1111. The lH-“Be cross sections, (+a=, were obtained from the data using D21
E = 1200 keV LI
Fig. 4. This backscattering spectrum of the self-supporting bilayer BeAl 6 provides data which indicate that the measured 4He-YBe cross section near I?, = 1200 keV is (1.03+0.07) times the Rutherford cross section.
and At atoms near the Be,‘At interface
z!z ) I wAl 1Ruth.
that is, the 4He-9Be cross section is Rutherford at ECU= 1200 keV. The Be area1 densities for BeAu 1 and BeAl 6 specified in Table 1 were determined using the cross sections shown in Fig. 2.
where A,, is the integrated Be proton peak count, KZ is the detector solid angle, Q = proton dose and (Nt), is the Be area1 density measured using 4He BS. Measured values of the 1H-9Be cross section ratios for proton lab energies between 2400 and 2700 keV are plotted in Fig. 6 and compared with Mozer’s values ; agreement is not expected since the backscattering angles are different. The maximum cross section ratio, u/us = 60 f 3, occurs for E, = 2525 keV. The uncertainty in (Nt),, is f 3%, and that of the (QCI) product (monitored with a Bi-implanted-in-Si RBS standard ) is f 3%, so the unce~ain~es in the 1H-9Be cross sections of Fig. 5 are &5%. Ubsemations and cafculations indicate that the d, peak Qabeled 9Be (p, d,) ‘Be in Fig. 5) may partially overlap the Be proton peak at E, = 2525 keV; at this energy the d,, peak count is 3% of the Be proton peak count. The two peaks occur at the same energy for EP - 2700 keV. This d, peak interference may be eliminated by placing a Mylar film (12 to 25 km) over the detector; the 0~s and 6Li are stopped and the d, energy is reduced
Be Y28.0 nm
Au 2.5 ram
E = 2475 keV P
= 170.5” LAB Q=3pC
Fig. 5. A typical ‘H backscattering spectrum of the self-supporting bilayer BeAu 1 showing (a> all peaks observed (the symbol “Li refers to 6Li from ‘Be(p, 6Li)4We), and (b) an expanded view of the energy region just above the Be peak. I. ELEMENT/ISOTOPIC
J.A. Leavittet al. /Nucl. Instr. and Meth. in Phys.Rex B 85 (1994) 37-41
E = 2525 keV
Proton Lab Energy (keV)
Fig. 6. The points are measured 170.5” cross sections, (T, for ‘H scattered by ‘Be for lH laboratory energies between 2400 and 2700 keV expressed in terms of the Rutherford cross sections, (TV, at the same energies; the solid curve serves to guide the eye; at E, = 2525 keV, u/oR = 6Ok3. The dashed curve (for I3LAB= 142.4”) is from ref. .
ence. Should the analyst wish to run without the Mylar film, the effective should be increased by 3% (to 62 f for the d, contribution; the ‘Be(p,
tion is non-resonant [141.
at E, = 2525 keV cross section ratio 3) to compensate
d,18Be cross secin the vicinity of E, = 2525 keV
2.3. Proton channeling measurements Fig. 7 shows proton channeling spectra for BeSi 18 (see Table l), a single-crystal (0001) hcp-Be film grown on a (111) Si substrate . Subtraction of the Si backgrounds from the Be peaks yields Abe = 5909 rf: 113 and AL = 2242 f 85 for the random and channeled cases, respectively. The Be area1 density, (Nt),,, may be calculated using Eq. (2) with Ah, divided by (62 L- 3) and (~a~ taken as the Rutherford ‘H-9Be cross section. The uncertainty in the ([email protected]
),, result is *6%, with the 3t5% cross section uncertainty added in quadrature to the normal f3% uncertainty associ-
Table 1 Characteristics of films used for measurement of ‘H-‘Be ‘H at 2525 keV and 4He at 1200 keV, respectively Film
1 2 1 2 1
Au Be Al Be Be
14.7f 1560 f 820 + 1780 f 1990 f
BeAl 6 BeSi 18
Fig. 7. Random and channeled proton Packscattering spectra of a single-crystal Be film (- 1650 A) on a single-crystal Si(ll1) substrate yield a Be area1 density = (199Ok 120)x 10” atoms/cm2, x:in = (0.26 f 0.01) and ,y9 = (0.38 f 0.02).
ated with the (Q0) product. The ratio of channeledto-random counts, jmin = 0.38 f 0.02 indicates moderate epitaxial growth for the whole Be film. We have also observed epitaxial growth of Be on CXA,O, substrates. For these films it is usually easier to align the substrate with a 4He beam rather than a ‘H beam because the ‘H-Al cross section is so small; channeled and random ‘H spectra are subsequently obtained.
We have reported measured 4He-9Be cross sections for 500
and 4He-9Be cross sections; AE, and AE, refer to film energy losses by
1015atoms/cm21 0.4 50 25 60 120
tnml - 2.5 129 137 146 165
A E, [keV]
A E, [keVl
0.2 2.7 3.5 3.1 3.4
1.7 36 43 42 46
J.A. Leavitt et al. / Nucl. Instr. and Meth. in Phw Res. B 85 (1994) 37-41
We have demonstrated that Be area1 density of thin films may be determined with - +6% accuracy using proton backscattering. We have demonstrated that proton backscattering may be used to evaluate the crystal structure of Be films.
AcknowIedgements This work is partially supported by the Optical Data Storage Center at the University of Arizona. We thank J.A. Ruffner, J.M. Slaughter, J. Eickmann and C.M. Falco for providing the crystalline Be films.
References [l] J.A. Ruffner, J.M. Slaughter and CM. Falco, Appl. Phys. I.&t. 60 (1992) 2995. [Z] J.F. Ziegler, J.P. Biersack and U. Littmark, The Stopping and Range of Ions in Solids, vol. 1 (Pergamon, New York, 1985).  J.D. Goss, S.L. Blat& D.R. Parsignault, CD. Porterfield and F.L. Riffle, Phys. Rev. C 7 (1973) 1837.
141 F.S. Mozer, Phys. Rev. 104 (1956) i386. PI R.A. Langley, M. Lewis and R.A. Zuhr, Nucl. Instr. and Meth. B 29 (1988) 599. bl J.A. Leavitt, Nucl. Instr. and Meth. B 24/25 (1987) 717. 171 J.A. Leavitt, L.C. McIntyre, Jr., M.D. Ashbaugh, J.G. Oder, Z. Lin and B. Dezfouly-~omandy, Nucl. Instr. and Meth. B 44 (1990) 260. 181J. L’Ecuyer, J.A. Davies and N. Matsunami, Nucl. Instr. and Meth. 160 (1979) 337. [91 Z.A. Saleh, F. Machali, 1.1. Bondouk and D.A. Datwish, Ann. Phys. 31 (1974) 76. DOI G. Foti, J.W. Mayer and E. Rimini, in: Ion Bean Handbook for Material Analysis, eds. J.W. Mayer and E. Rimini (Academic Press, New York, 1977) p. 53. [ill A. Kiss, E. Koltay, Gy. Szabo and L. V6gh, Nucl. Phys. A 282 (1977) 44; M. Ahab, A. Boucenna and M. Haddad, J. Phys. (Paris) 44 (1983) 579. ml W.K. Chu, J.W. Mayer and M.-A. Nicolet, Backscattering Spectrometry (Academic Press, New York, 1978). 1131Supplied by H.L. Eshbach, Centrat Bureau of Nuclear Measurements, Steenweg Op Retie, 2440 Geel, Belgium.  G. Weber, L.W. Davis and J.B. Marion, Phys. Rev. 104 (1956) 1307.  J.A. Ruffner, J.M. Slaughter, J. Eickmann and CM. Falco, private communication, 1993.