Characterization of Thin Films

7 downloads 0 Views 15MB Size Report
ron gun. Sz Sr r. To current digitizer lon pump. Gote volve. Ht-. Solid stote detector. X-roy ...... Implantation ln Semiconductors, €d. by F. Chernow. Plenum Press,.
Report of Investigations

8455

Characterization of Thin Films and Solid Surfaces Using Proton-fnduced X-Ray Emission By Bruce D. Sartwell and Arthur B. Campbell III

UNITED STATES DEPARTMENT OF THE INTERIOR Cecil D. Andrus, Secretary BUREAU OF MINES

Lindsay D. Norman, Acting Director

This publication has been cataloged as follows:

Sartwell, Bruce

D

Chqrqcterizqtion

of thin films and solid surfcces

using

proton-induced X-rcy emission. (Report of investigations

.

Bureau of Mines

Bibliography: p" 2l-22" Supt. of Docs.

no,: |

;

8455)

28.23:8455"

1" Thin films, Effect of radiation. one 2. Solids-..Surfaces" I. Camp. ioint author" II. Title. III. Series: United States" Bureau of Mines" Report of investigations ; 8455"

bell, Arthur B" III L943-

TN23.U43 [QC176.84.R3] 622s [SgO.a'r] 79.607938

CONTENTS

Page

Abstract... .................... r.'...................... ................. Introductlon . . . . . . . . . . . . . . . . . . . . . o . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . Experimental work. ........... ......................... .............. ... . Theory and appllcatlons..r.. ...... ......................... ............. . Results and discussion.... ........ ............................ ...... Cross sections and calibration measurements . ....................... Thln-film analysis. . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bulk analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ......................... Depth Profiling ... . .................

L5 18

.

L9 2L

Summary and conclusions ........ ........ ................................. References, . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . .. . .. . .. . . . ... . . . . . . . ..

1 L

2

4 10 10

t4

ITTUSTMTIONS

1. 2, 3. 4, 5. 6, 7

,

8. 9, 10.

Schematic of ultra-high-vacuum target chamber in whlch the PI)G analyses werg conducted.. ................................... PLot of the oxygen K X-ray yield I(Eo ) as a function of Fe3On film Ehtckness. . .. . .. . . . . . . . . . . . , . . . . . . . . . . .. . .. . . . . .. . .. . . . . . . .. Plot of the equlvalent surface coverage of carbon determined by PIXE as a functton of the acEual bulk carbon impurity COneentfatiOn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CorrecEion factors used in the normalizatilorL of indivldual X-ray yields to obtaln atomi.e percent concentratton for bulk analysls. . X-ray producLion cross sections for K-shell, L-sheLL, and M-shell

..... ... exCitatiOn by PfOtOnS. .. . ..... ... ........................ Plot of the iron L X-ray yield as a function of iron film thlckness. . . . . . .. . . . . . . . . . . . . . . . . .. . . . . . . . . . .. . . . . .. . . . . . . . . . . . . . Oxide thickness plotted as a function of exposure time for iron o..... and three lon-implanted iron samples ........ ............... 0xidation klnetics of high-purity iron for Ewo dlfferent initial sulfur surface coverag€s.. ....... ........................ Integral dlstribution and depth proflles as determined by the Pl)G-sputtering technique for 25-kev Ni- lmplantatLon lnto lron. . Platlnum M and iron K experimenLal X-ray yields for L5O-kev proton bombardment as a function of sputtering charge for an elgctrodeposited platlnum fiLm. ... ..... ..........................

11 L2 L3 15 L7

L9

TABTES

1.

2.

3. l+.

Thlckness of oxlde fl1m measured usLng three different protonbeam angles of tncidenC€.... . . ................................... Ratlo of the substrate alumlnum K ytelds obtalned before and after evaporation as compared to the ratio predlcted from the evaporated iron flIm thlckness measured using the thinfllm and integral approximattons-.. .......-......................... PI,)G results for thg A-286 turbing whegl........................... PDCE resutts for the Nlmonic-1O5 turbine blade.... .................

5

13 L6 L7

CHARACTERIZATION OF TH IN FILMS AND SOLID SURFACES USING PROTON.INDUCED X.RAY EMISSION by

Bruce D. Sortwell

I ond Arthur B. Compbell lll2

ABSTMCT

the Bureau of Mines is usi.ng characteristlc X-rays produced by proton of a solid surface to provide quantiEati.ve compositlonal analyses of surface layers of metals. An integral X-ray yield equation has been developed that quantitatively relates Lhe measured X-ray yield to the thtckness of thin fllms ranging from less than a monolayer to severat thousand angstroms. The use of the integral- yield equation is demonstrated for the growth of a thln fllm by vacuum evaporatlon. Results are presented that show hot^z proton-induced X-ray emission (PIXE) has been used Eo determlne the oxidation kineEics of iron that had been implanted with several different species of heavy ions. the sensitivity of PIXE for the measurement of very-Lhin-fllm growth klnetics is demonstrated from studies of the effect of fractional monolayer coverages of sulfur on the initial oxidation kinetics of iron. Results are also presented in which PIXE has been used to study the redistrlbution of alloying elements in components of a gas turbine that had fractured due Eo corros ion fatigue . By combining PIXE with Iow-energy i.on sputtering, quantitative composition depth profiles of surface layers are obtained. Examples are presented for the profillng of iron samples implanted with 25-kev nicket ions and for a platinum coating on an Fe-5Cr substrate. bombardment

INTRODUCTION

A principaL objeettve of the Bureau of Mines is to conserve the Nati.onrs mineral resources. In keeping with this goal, the Bureaurs Avondale Research Center has been conducLing sEudies to determine the mechanlsms responsible for the oxidation and corrosi.on of metallic materials that can be used in metallurgical processing operations and thus increase the useful life of materials in process components. Although metallic corrosion has been extensively studi-ed for many decades, i.t was not untit the early 1970rs thaL the powerful techniques of surface sclence began to be fully utilized in this.area.

]Research physicist, Avondale Research Center, Bureau of Mines, Avondale, oFormerly research physicist, Avondale Research Center, Bureau of Mines. Avondale, I\,[d., now with Naval Research Laboratoryr Washington, D.C.

Md.

2

Iuletallic corrosion can be defined as the interaction of a metaL surface with its environment. In many cases, the initiatlon of a corrosion process lnvoLves the adsorption of a layer of a foreign specte on the metal surface, w:ith subsequent chemi.caL combination and then gro,.hrth of a thin fi1m. In aL1oys containing two or more metals, different chemical free energies for the formation of oxides can result ln the depletion of a particular element or elements at the f i1m-meta1 interface. This can also resul-t in the formation of a protective oxide fiLm by the element with the most negative free energy of formation. The final phase of corrosion can invoLve the formation of thick corrosion products resultlng in changes in the substrate a1Loy composition over a significant depth. Therefore, a general study of corrosion processes should include (1) surface analyses that include only the first few atomic layers and are sensitive to sma1l fractions of an atomic layer , (2) thin-film and thick-film analyses, and (3) substrate bulk analyses.

In this paper, the use of proton-induced X-ray emission (PIXE) ln each of these three areas wilL be demonstrated. The emphasis w11L be on presenti.ng ilLus trati.ve examples rather than on detailed interpretive anaLyses of the results. These examples wilL incLude (1) monitoring the thickness of vacuumevaporated fl1ms , (2) determining the oxide film groqTth kinetics for iron and for iron that has been implanted with 25-kev F"* , Ar* , or Cr* , (3) determining the effects of different sulfur surface coverages on the tnitial oxi.dation kinetics of iron, (4) determinlng the redistrlbutlon of alloying eLements in superalloy turblne components subjected to corrosion fatigue, and (5) obtaining composition depLh profiles of ion-implanted iron samples and electrodeposlted platinum coatings.

EXPERIMENTAL

WORK

The PI)(E anaLyses described in within an energy range of 100

this paper were accomplished using proton to 2O0 kev that were obtained from a Cockroft-Walton electrostatic accelerator equipped with a radio-frequency ion source. the ion beam was mass-analyzed to separate the proton beam from the moLecular hydrogen ion beam. The ultra-high-vacuum chamber in which the pI)(E analyses were performed, shonm schematically in figure 1, had a base pressure 1 x lfle torr; for most measurements reported herel the ir.rsure in the chamber ranged from 4 to 8 x LdB torr. The proton beam was collimated to 0.2-cm diameter by apertures 51 and S2 with aperture SB biased for secondary electron suppressi.on. The characteristic X-rays emitted from the sample were detected using two types of detectors: a gas-flow proportional counter containing a beams

90% €Lrgon-lO% methane gas mixture and having two 2-p*-thlck aluminized mylat entrance windours W1 and Wp ; and a Si (Li) soLid state detector wlth Z-g^-thick aluminized MyLar and 25.O-Ltrm-thick beryllium entrance windor^rs, We and InIn , respectively. The proportional counter was used only for detecting the very low energy C K and O K X-rays in order to determine surface carbon coverages and oxide film thicknesses. The solid state detector was used to measure a1L other el-ements. Due to attenuation in the beryllium wi.nd.ow, X-1.ays with an

Sopphire leok volve

r

To

current digitizer

lon pump

Sputtering ron gun Gote volve

Sz Sr

Ht-

f sl

Somple

Solid stote detector

X-roy f ilters

Proportiono counter FIGURE

l. - Schemotic

I

of ultro-high-vocuum torget chomber in which the PIXE onolyses were

conducted.

energy Less than 0.8 kev couldonot be detected. ltre Si(Li) detectorrs cross-sectlonal area riras 80 mri and lt had a resolutlon of zLO ev for I4n Kcr X-rays, The X-ray fllter attached to the proporEional counter was utiLlzed in the oxide film thlckness measurements with high-purity oxygen belng bled lnto the fitter to a pressure of L00 torr. The oxygen attenuated the lnterfering C K and Fe L X-ray lines to a much greater degree than the 0 K X-rays and, therefore, permLtted very sensltive measurements of surface oxygen. For all measurements made wtth the St (Li) detector , the X-ray fllter was maintalned under vacuum, The sputter lon gun was used in the depth-profiling studies. The procedure used for sputtering was to inject ultra-high-purity argon into the chamber to a pressure of 5 x LdE torr and then bombard the sampl-e with L,O-kev Ar+ at normal incldence for a preselected amount of charge as determLned by a digttal current integrator. ALso shown Ln flgure 1 is a sapphlre leak valve lrhich was used for injectlng elther ultra-hlgh-purlty argon (for sputtert*g) or oxygen (for oxidation studles) lnto the chamber.

4

Ttre X-ray spectra, obtained from anaLyses of sampLes , were accumulated ln a muLtlchannel putse helght anaLyzet (PHA). Proton beam current was measured on the sample using a digltal currenL lntegrator. A preset scaler, connected to the output of the lntegrator, gated the PIIA; this provided for preselection of the amount of proton beam charge to be accumuLated on the target. Each peak in the X-ray spectrum was lntegrated and, following correction for background, the experimental X-ray yield, Y, expressed as the number of detected characteristic X-rays per tncident microcoulomb of proton beam charge, was calculated.

Ihe factor Y could be converted to the absolute X-ray yield I (Eu ) , where Eo is the incident proton beam^energy and I(Eo ) is expressed as X-rays per proton, using the equation (7)'

r(Eol=ffi,

(1)

dCI is the solid angl-e subtended by the detector and I is the detector efficlency vrhich is determined from the transmlssion of the X-rays through the entrance windorrs of the detectors and, for the Sf Gi) detector, the transmission of the X-rays through the slllcon dead Layer on the detector.

where

THEORY AIID APPTICATIONS

The excLtation of characteristic X-rays by proton material is generally descrlbed by the equation (4)

r(Eo) = "l:

o"[E(Eo

bombardment

of a solld

,r) lexp (-prr)dr,

(2)

n is the numbg:r of target atoms/g, R" is the totaL projected range of thg protons tn g/e? t ox[E (Eo" rr) ] ls d; X-ray productlon cross section Ln crrf /atom for prototrs with instantaneous energy E at a distance r Ln g/ctf along the projected proton path, and Lt is the absoqption eoefflcient in crrf /g for the characteristic X-ray ln the target material. If the element belng detected is present as a surface ftlm of thlckness T, then equatton 2 can be rewritten with a change of varlabie from the projected proton path Length , t, to a depth x in cm normal to the sampLe surface, where

e

I(Eo) =

No

J:'"" o* [u

(Eo

,x/cos 0lexp (-pp x/cos 0 )d (x/cos 0 ) ,

(3)

where 0 is both the angle between the lncident proton beam and the sample normaL and the angle between the detector and the sample normal, No is the atomic denslty of the film ln atoms/cm3, and p is the bulk density of the film in g/ c*3 .

Underlined numbers in parentheses at the end of this report.

refer to i

es

If a film is sufficiently Lhin such chat the energy loss of Lhe Protons can be negtected, ox is a constant and equation 3 can be integrated to give r(Eo

)

.

= No o* LEo

-(1

lt:-

exp (-pr,p

T/cos e \ 1

(4)

u,p

In addition , Lf the film is thin enough that absorption of Lhe X-r€tys is negligible, the exponential in equation 4 can be expanded in a Taylor series to give (s) I exp (-U,P T/ cos e ) : pP T/ cos e I(Eo

and

) = Ns 0x [u" ] (T/cos e) rF

or

--E

- I(Eo)cos No

o* Lgo

(6)

e

(7)

J

The cos e term takes into account the lncrease ln X-ray yield for increase