UDK 543.428.2:541.183 Izvirni znanstveni članek ISSN 1580-2949 MTAEC 9, 36(6)307(2002) J. T. GRANT: ANALYSIS OF NANOWEAR AND THIN FILMS USING AES AND XPS ANALYSIS OF NANOWEAR AND THIN FILMS USING AES AND XPS ANALIZA NANOOBRABE IN TANKIH PLASTI Z UPORABO AES IN XPS John T. Grant Research Institute, University of Dayton, 300 College Park, Dayton, Ohio, 45469-0168, USA j.grantŽieee.org Prejem rokopisa - received: 2002-10-11; sprejem za objavo - accepted for publication: 2002-12-16 Auger electronspectroscopy (AES) and X-ray photoelectronspectroscopy (XPS) have beenused to characterize the nanowear of ZnO, the interdiffusion of Al/Ti films grown on SiC, contamination formed during the fabrication of MEMS (Micro-Electro-Mechanical Systems) switches, and the surface composition along a crack in an aluminum alloy. The relatively high spatial resolution of AES is very useful for the surface compositional analysis of MEMS devices and can often provide additional informationabout surface chemistry. Whenspatial resolutionis not a problem, XPS usually provides superior quantitative analysis as well as more chemical information about the surface composition. Key words: Auger electron spectroscopy, X-ray photoelectron spectroscopy, MEMS, wear, diffusion, contamination Spektroskopija Augerjevih elektronov (AES) in rentgenska fotoelektronska spektroskopija (XPS) sta bili uporabljeni za karakterizacijo nanoobrabe ZnO, interdifuzije v Al/Ti plasteh na podlagi SiC, kontaminacije med izdelavo MEMS (mikro-mehanskih sistemov) in sestave površine vzdolž razpoke v aluminijevi zlitini. Relativno velika prostorska ločljivost AES je zelo uporabna za analizo površine MEMS-naprav in lahko prinese dodatne informacije o sestavi površine. Ko ni problem prostorska ločljivost, XPS prinaša boljše kvantitativne podatke in več podatkov o kemijski sestavi površine. Ključne besede: spektroskopija Augerjevih elektronov, rentgenska fotoelektronska spektroskopija, MEMS, obraba, difuzija, kontaminacija površine 1 INTRODUCTION Auger electronspectroscopy (AES) and X-ray photoelectronspectroscopy (XPS), (XPS is also called ESCA, electronspectroscopy for chemical analysis), are commonly used techniques for determining the surface compositionof materials. InAES and XPS applications, there is often a need for high spatial resolution, for example instudying MEMS-related devices. The spatial resolutions of AES and XPS instruments have improved over the years, with modern commercial instruments being able to acquire spectra from areas of the order of 10 nm (AES) and 10 µm (XPS). With AES, the regions to be analyzed are determined by operating the instrument in the SEM mode and measuring the secondary electrons or backscattered electrons leaving the sample, or by measuring changes in the current to ground, to form an image of the surface as the electron beam is rastered onthe sample. With XPS, regions to be analyzed are located by (previously calibrated) optical microscopes, video cameras, X-ray induced electron emission, or by imaging the sample through the analyzer. Higher spatial resolutions with XPS can be achieved by conducting the analysis at a synchrotron, where spatial resolutions of the order of 100 nm can be achieved. This paper describes the applicationof AES to study the nanowear of ZnO, the interdiffusion of Al/Ti films grown on SiC, and to identify contamination formed MATERIALI IN TEHNOLOGIJE 36 (2002) 6 during the fabrication of MEMS switches. XPS was used to determine the surface composition along a crack in an aluminum aircraft alloy, and also to study the contamination formed during the fabrication of the MEMS switches. 2 EQUIPMENT For the AES analysis, a Physical Electronics Inc (PHI) model 5700 AES/XPS system was used. For the XPS analysis, a Surface Science Instruments (SSI) M-Probe was used. The electrongunonthis PHI system has a minimum electron beam diameter of 100 nm at 10 keV. Secondary electrons were used for imaging the sample surface. The highest spatial resolutionof the SSI M-probe is 150 µm indiameter, with the maximum analysis area being a rectangle 400 µm × 1000 µm. In the SSI system, the analysis area is determined by the size of the monochromatic Al X-ray beam on the sample, and the sample is positioned in the analysis chamber using an optical microscope. Both of these systems have fast sample insertion capabilities, so analysis can be started within ˝ hour after mounting the sample on a holder. The SSI system has a larger sample introduction port allowing samples of 25 mm height to be inserted, whereas this PHI system is limited to samples 15 mm high. 307 J. T. GRANT: ANALYSIS OF NANOWEAR AND THIN FILMS USING AES AND XPS 3 AES STUDY OF THE NANOWEAR OF ZnO In ambient conditions, zinc oxide thin films can show low friction and long wear life, and this is thought to be due to extra defects introduced into the ZnO by preferred doping. This hypothesis was examined by J. Nainapa-rampil 1, by studying the nanowear of ZnO crystal faces by prolonged scanning (tribological stressing) with the Si tip of an atomic force microscope (AFM). Continuous scanning with the Si tip would be expected to form transfer films at the nano-wear scar and produce a reductioninfriction, as Si is one of the preferred dopants. The pyramidal tip of the AFM had a height of 3 µm, an apex angle of 15°, and a circular base with a radius of 1 µm. Lateral force measurements showed that the nanowear scar had a lower friction coefficient than the surrounding areas 1, so AES was used to examine the wear scar for the presence of Si. A nanowear scar is showninFigure 1. Auger spectra were measured inside the nanowear scar, and outside on the surrounding area. The Auger spectra are showninFigure 2. As the scarred area could not be seen in the AES system using SEM imaging, the area scarred by the AFM was referenced to the edge of a mesh (a cut TEM grid) so it could be located inthe AES system. Figure 2a shows the three Zn transitions (40-110 eV) from ZnO in the un-worn surrounding region(lower spectrum). The upper spectrum inFigure 2a was taken from the nanowear scar and shows the additional presence of Si. The presence of Si in the nanowear scar was confirmed using the high energy Si KL2,3L2,3 transitionas shownby the upper spectrum in Figure 2b; the lower spectrum is from the surrounding un-wornarea. The spectra shownhere are inderivative form and were obtained from the raw, direct spectra using differentiation in the computer. Auger signals were detected from Zn, O, C, and Si in the nanowear scar, and their atomic concentrations were calculated from the peak-to-peak intensities of the derivative Auger spectra, giving atomic concentrations of 20, 13, 65, and 2%, respectively. The corresponding atomic concentrations for Zn, O, and C from the un-worn surrounding area were 19, 11, and 70%, respectively. Si was not detected inthe un-wornarea. These calculations assume that the analyzed regions are homogeneous. Anenergy dispersive spectroscopy (EDS) study of the worn tip showed the presence of Zn, which came from the ZnO crystal 1. In summary, prolonged scanning of a Si AFM tip on a ZnO (0001) surface resulted in a lubricious surface, and AES proved there was a transfer of Si from the tip to the ZnO. 4 THE INTERDIFFUSION OF Al/Ti FILMS GROWN ON SiC A number of Al/Ti films grownonSiC were examined after being annealed at different temperatures. The goal was to determine if and when interdiffusion betweenthe deposited layers occurred. AES was ideally suited for this study, as sputter depth profiles through the films will show whendiffusionoccurred. Linear least squares (LLS) fitting of the spectra also allowed Auger peak overlap problems to be removed, to determine if different chemical states of the elements exist in the layered structure, and to improve the signal-to-noise in si /"""V/"- - NanowearScar ČT Zn Outside Nanowear Scar Zn : / Î J Zn Kinetic Energy (eV) 1580 1585 1590 1595 Figure 1: Nanowear scar on a ZnO crystal produced by prolonged scanning with the Si tip of an atomic force microscope (AFM) Slika 1: Raza zaradi nanoobrabe na kristalu ZnO, ki je nastala pri podaljšanem skeniranju s Si-konico v mikroskopu na atomsko silo Kinetic Energy ( eV ) Figure 2: Auger spectra from inside the nanowear scar (upper spectrum), and outside on the surrounding area (lower spectrum); (a) low kinetic energy region (40 - 110 eV), and (b) high kinetic energy region (1580 - 1630 eV) Slika 2: AES-spektri iz notranjosti raze (zgornji spekter) in iz njene zunanjosti na površini v okolici (spodnji spekter); (a) nizka kinetična energija (40-110) eV in (b) področje z visoko kinetično energijo (1580-1630) eV a b 308 MATERIALI IN TEHNOLOGIJE 36 (2002) 6 J. T. GRANT: ANALYSIS OF NANOWEAR AND THIN FILMS USING AES AND XPS 30 J ČrČ \ / Tia. SO \ / 70 \ / 60 ° \/ 50 \ \ 40 iV A 30 w / \ 20 cA\ / V 10 Ä Vt* \ v . 30 AM \ e SO 70 30 50 / Ti2 40 30 20 \ 01 10 01 — C1 -2 0 2 4 6 8 10 12 14 16 18 Sputter Time (min) 15 20 30 b 30 C1.IS1 ČIslČvVV Sitisi . " " jl-/ \ Aitisi / Ctls2 20 /; N. Ctlsl / ,0 brisi -ȱŮČ>iyčČ Na. J/K. 30 f SO 70 . C1 ¦ 30 50 \r\ s Ti2 \ r- -Čv_- Si1 _ \i y \ / 40 30 20 10 I V AI1 Č5c Z ¦ -20 0 20 40 60 80 100 120 140 160 180 Sputter Time (min) Sputter Time (min) ¦ c C1.ls1 Ti2.ls1 \ Ctlsl Sitisi y / Č , Sitls2/ / ', /\ Aitisi / X il / ,s* s. /A Otls1 7Č> •• ?V *» / \ .* j-iČV-vaLČČČ .Čk '' / /\\ ' *Č" ČM g C1 ah Ti2 ,Č-Č— Si1 yČČČ C1 I . 01 ČČ-Č_ —VČ-yČ AM -20 0 20 40 60 80 100 120 140 160 18 Sputter Time (min) Sputter Time (min) 30 d BO ™ 60 . Olisi Ctlsl Ctls2' - . C1 Is1 j*",*AvV \rN - y J \ Aitisi /VČČviJ / Sills' 20 A Čv ,.-' siiis2 yj F\ 10 / -JLtČTČ ;'- Č / l\\ ¦rfČw.-ČxA-ys/c. ČŁ>Čw/Č 100 120 140 16 Sputter Time (min) h C1 ' Ti2 \ /Č \/v. Si1 -\r C1 -1 7Č---------------'--------------------'— >, au Ti2 f\ 01 / \Š / .......i « ü -50 0 50 100 150 200 Sputter Time (min) Figure 4: Sputter depth profile of unannealed Al/Ti on SiC using AES, with a 3 keV Ar+ ionbeam onduring Auger data acquisition. The ion-excited signals from Al and Si dominate the profile and give an erroneous quantitative analysis Slika 4: Profilni AES-diagram nežarjenega Al/Ti na SiC, določen z ionskim jedkanjem Ar+, 3 keV. Ionsko vzbujena signala Al in Si prevladujeta in dajeta nepravilno kvantitativno analizo. 00 70 D 310 MATERIALI IN TEHNOLOGIJE 36 (2002) 6 J. T. GRANT: ANALYSIS OF NANOWEAR AND THIN FILMS USING AES AND XPS b Figure 5: (a) SEM secondary electron image of part of a MEMS switch being fabricated onGaAs. The regionmarked 1 is contaminating residue left after stripping the photoresist; (b) Auger spectra obtained from the regions marked 1, 2, and 3 in part (a) Slika 5: (a) SEM-slika dela MEMS-stikala, ki je bilo izdelano na GaAs. Področje, označeno z 1, je kontaminacijski ostanek po jedkanju z fotorezistom, (b) AES-spektri področij, označenih z 1, 2, 3 in deloma (a) In summary, this analysis shows the usefulness of AES sputter depth profiling to monitor the interdiffusion of Al/Ti layers on SiC after annealing, and how the information in, and the quality of, the profiles were enhanced by linear least squares fitting of the data. 5 CONTAMINATION DURING THE FABRICATION OF MEMS SWITCHES An unknown white residue was observed after stripping the photoresist in RF MEMS switch fabrication onGaAs substrates. Several pieces of residue could be seen, the largest being about 500 µm in size. When this sample was received, the PHI system (for AES analysis) was not available so some XPS work was performed with the limited spatial resolutionof the SSI system. The XPS analysis of the largest piece of white residue gave surface concentrations of 75 at. % C, 20 at. % O, with the remainder being N (this assumed that the surface compositionwas homogeneous inthe analyzed region). A region away from the white residue gave similar relative concentrations of C, O and N, but with about 0.7 at. % each of Ga and As. This region was known to be inhomogeneous, due to the patterning made onthe device. This XPS analysis showed that the deposit was thick enough to prevent signal being detected from the GaAs substrate (as expected) and was probably due to photoresist. However, with the small size of the residue there was some uncertainty regarding the positionof the 150 µm X-ray beam onthe sample using the microscope for alignment, as this system is used by many people for various projects, and an alignment calibrationof the microscope was not performed just prior to analysis. AES analysis in the PHI system was therefore planned for later. The residue could be easily located using the electron beam inthe PHI system. AnSEM image from part of the sample is showninFigure 5a, where the deposit appears as the large region(labeled 1) at the right. Note also the patterned areas onthe sample. Auger spectra takenfrom the three areas marked in Figure 5a are showninFigure 5b. Note that the patterned area (labeled 3) shows Ga and As signals from the substrate, whereas the other two areas do not. The C signal from region 1 is broader than that from regions 2 and 3 and indicates electrical charging inregion1. This AES analysis showed that the residue was organic with a composition similar to photoresist. The white residue occurred due to the apparent interaction of photoresist and a low power RF bias plasma sputter cleaning process. The plasma cleaning altered the state of the photoresist insome areas such that standard photoresist stripping solutions did not remove the residue that formed. Based onthis analysis, the photolithography processing steps were modified to eliminate this problem. 6 THE SURFACE COMPOSITION ALONG A CRACK IN AN ALUMINUM AIRCRAFT ALLOY XPS was used to determine the composition at different places along a crack in an aluminum structural part from anaircraft. The crack was long, approximately 25 cm, so the part had been cut into five sections along its length, and then cut from behind so each part was about Figure 6: A photograph of a crack inanaluminum structural part from anaircraft. The 19 places marked show where XPS analysis was performed. Slika 6: Posnetek razpoke na delu letala iz aluminija. Na 19 označenih mestih je bila izvršena XPS-analiza. MATERIALI IN TEHNOLOGIJE 36 (2002) 6 311 J. T. GRANT: ANALYSIS OF NANOWEAR AND THIN FILMS USING AES AND XPS Table 1: Atomic concentrations of elements detected at the 19 regions examined along the crack shown in Figure 6. The concentrations are in atomic percent, and assume that the surface is homogeneous within each 400 µm x 1000 µm analysis area. Region 14 was analyzed twice, the second analysis being for a much longer time (improved signal to noise and detectability). Tabela 1: Atomska koncentracija elementov na 19 mestih vzdolž razpoke na sliki 6. Predpostavljena je homogena sestava površine na področju 400×1000 µm. Točka 14 je bila analizirana dvakrat, trajanje druge analize je bilo daljše (izboljšano je razmerje med signalom in ozadjem ter občutljivost) Location Zn O Ca C Al Cd N Mg Na Si S F P Cu Cl 1 1.1 25 1.1 71 0.7 1.0 0.4 2 1.0 23 1.5 70 2 0.7 0.8 1.4 3 0.9 23 1.5 68 2 0.5 0.3 3.0 1.1 4 1.1 30 1.0 59 5 0.1 0.5 1.5 0.8 0.1 5 1.4 28 1.9 61 2 0.8 0.1 0.4 3.2 1.7 6 1.2 24 1.7 68 0.6 0.4 2.5 1.5 7 1.2 32 2.2 54 4 0.6 0.9 3.5 2.3 8 1.4 33 2.6 54 4 0.8 0.1 2.1 2.8 9 1.4 32 1.8 57 4 0.7 0.3 2.7 10 1.4 31 60 5 0.4 0.6 1.7 11 0.5 43 1.2 39 15 0.7 12 0.7 46 2.1 38 13 0.03 0.1 0.0 13 1.0 48 1.3 36 12 0.09 0.3 1.4 0.4 14 1.1 39 2.5 47 6 0.3 0.4 0.8 0.9 1.2 14 1.1 39 2.6 47 7 0.3 0.4 1.3 0.2 0.9 0.3 1.1 15 0.7 54 17 22 5.9 16 0.6 43 37 11 0.7 3.8 1.8 1.5 0.6 17 0.4 42 35 16 0.7 4.9 0.5 18 0.2 34 49 10 1.3 3.2 0.9 1.5 0.2 19 30 46 14 1.5 4.3 3.2 0.5 2 cm thick. Cutting was required to make samples small enough so they could be inserted into the SSI system using the fast insertion system. Each section was mounted and analyzed separately. A photograph of the crack is showninFigure 6. Analysis was performed at 19 places along the crack, corresponding to parts with different optical appearances. The elements detected, and their concentrations, are listed in Table 1. As canbe seenfrom Table 1, Cd, P and Ca were generally detected from the top of the crack, left side in Figure 6, analysis location1, downto the top of section number 4 (location14). Mg was detected onthe lower half of the crack, and Al increased (generally) down the crack. About 1 atomic percent Zn was detected almost everywhere onthe crack surface (Mg and Znare present inthe aluminum alloy). Sectionnumber 2 showed a few atomic percent F at all locations (locations 3 through 8), but as none was found on the other specimens, it was most likely contamination from handling. This compositional information was given to the customer, and it was determined that the Cd detected was from a fastener at the crack initiation point. 7 SUMMARY The usefulness of AES for studying nanowear of ZnO, the interdiffusion of Al/Ti films on SiC, and identifying contamination formed during the fabrication of MEMS switches has beenshown. The applicationof XPS to determine the surface compositiondowna long crack inanaluminum aircraft alloy has beenshown, and it was also used to examine contamination formed during the fabricationof MEMS switches. Linear least squares fitting of sputter depth profile data was used to improve the interpretation of the data. Acknowledgements The author would like to thank his colleagues for providing these samples for analysis: Jose Nainapa-rampil for the nanowear of ZnO; Mike Capano for the Al/Ti films onSiC; KevinLeedy for the MEMS switch fabricationonGaAs; and, Deb Peeler for the crack inthe Al aircraft alloy. The support of the U.S. Air Force Research Laboratory in conducting these analyses is also gratefully acknowledged. 8 REFERENCES 1 J. J. Nainaparampil, J. T. Grant, and J. S. Zabinski, in submission 2 W. F. Stickle, and D. G. Watson, J. Vac. Sci. Technol. A, 10 (1992) 2806 3 J. T. Grant, M. P. Hooker, R. W. Springer, and T. W. Haas, J. Vac. Sci. Technol. 12 (1975) 481 4 T. W. Haas, J. T. Grant, and G. J. Dooley, J. Appl. Phys., 43 (1972) 1853 5 PHI - Multipak, Physical Electronics, Inc., Eden Prairie, Minnesota, USA 312 MATERIALI IN TEHNOLOGIJE 36 (2002) 6