REVIEW OF SCIENTIFIC INSTRUMENTS

VOLUME 74, NUMBER 7

JULY 2003

Scanning Auger microscopy study of W tips for scanning tunneling microscopy L. Ottaviano,a) L. Lozzi, and S. Santucci Unita` INFM & Dipartimento di Fisica, Universita` degli Studi dell’Aquila, Via Vetoio 10, I-67010 Coppito L’Aquila, Italy

共Received 6 January 2003; accepted 6 April 2003兲 Tungsten tips used in scanning tunneling microscopy 共STM兲 共prepared via electrochemical etching with a 2 N KOH or NaOH solution兲 have been studied with state of the art scanning Auger microscopy 共SAM兲 with chemical lateral resolution of 10 nm. The experiments were focused on the investigation of the W tips’ apex shape and surface composition, for tips as etched, or after various postetching treatments performed for cleaning, sharpening, and surface oxide removal purposes. Ultrasonic cleaning likely bend the tip apex. Hydrofluoride etching successfully removes the native WO3 oxide layer, but this happens at the expense of the tip sharpness. Ion sputtering in ultrahigh vacuum is not always effective in sharpening and cleaning the tungsten tip apex, and we sometimes observed the formation of needle like nanotips, mostly composed of WO3 . Direct resistive annealing of the tip 共operated in the STM at 10 V, 50 nA set-point sample bias voltage and current, respectively兲 to remove the oxide layer, produces a coiling of the tip apex. In this case, atom transfer from the sample to the tip is directly demonstrated with Auger spectra taken at the tip apex. © 2003 American Institute of Physics. 关DOI: 10.1063/1.1581392兴

I. INTRODUCTION

This layer has to be removed in UHV one way or another to ensure good tunneling. Thus, among postprocessing procedures of the W tips, those leading to an oxide free W surface on the tip apex after insertion in UHV should routinely be adopted. Various methods have been proposed to this purpose: ex situ HF etching,13 or annealing at high temperatures,12 or ion milling.14 –18 The successful effects of these procedures are often simply verified a posteriori from the quality and reliability of the STM images. The W tip shape and surface composition can sometime be determined indirectly from deconvolution of STM images 共like in Ref. 19兲 and spectra 共like in Ref. 21兲. Quite rarely the modification of its shape and surface composition has been investigated directly with electron microscopy techniques like scanning electron microscopy 共SEM兲, and transmission electron microscopy 共TEM兲12,14,15,20,21 or field ion microscopy.22 Moreover, very little indeed is known about the W tip apex surface composition immediately after etching, or after postetching treatments.23,24 From this point of view, the W tip surface is ‘‘the dark side of the moon’’ of the nano world. Effectively, there are no studies that correlate high resolution images of the W tips’ apex, and a detailed knowledge of the chemical composition of its surface, with the various treatment methods used for cleaning, sharpening, and removing the native oxide layer. State of the art scanning Auger microscopy 共SAM兲 matches the possibility of acquiring SEM images with quite good lateral resolution 共10 nm兲 and simultaneously 共with the same lateral spatial resolution, via Auger spectroscopy of the secondary electrons兲 elemental information of the surface of the samples under investigation. These features make SAM an ideal tool for the systematic quantitative investigation and characterization of the shape and

Since the invention of scanning tunneling microscopy 共STM兲1 the family of scanning probe microscopies 共SPMs兲, that readily followed up, has allowed us to turn early fascinating speculations on the world of ‘‘nano’’2 into the stunning reality of the experimental evidence.3 SPMs allow investigation, control, and modification of the properties of surfaces at the atomic level. The probe tips of the SPMs are the real gates to this fascinating ‘‘nanoworld.’’ In this respect, it is evident that a control and knowledge of their shape, and 共bulk and surface兲 composition is of extreme importance. In the case of STM metal tips their shape is determined via electrochemical etching.4 – 6 W tips, prepared either in a solution of 2 N KOH or NaOH,4 are those most commonly used for STM in ultrahigh vacuum 共UHV兲. This is essentially due to W hardness, which prevents deformation and erosion of the tip during imaging, and ensures a relative tip stability during the experiments. There are systematical studies of the parameters that influence the W tip aspect ratio, and its apex radius during the etching procedure.7,8 Other studies are devoted to postetching treatments like 共to briefly mention some outstanding examples兲 the deposition of a magnetic layer with a well defined easy axis of magnetization for spin polarized STM studies with atomic resolution,9 or the controlled growth of pyramidal nanotips at the tip apex in UHV,10 or the operation of the tip in field emission regime in order to melt the tip apex to obtain reliable and reproducible scanning tunneling spectroscopy results.11 After etching in the electrolyte solution, W tips inevitably have a native WO3 oxide layer, that is a few nm thick.12 a兲

Electronic mail: [email protected]

0034-6748/2003/74(7)/3368/11/$20.00

3368

© 2003 American Institute of Physics

Downloaded 31 Oct 2003 to 192.84.136.174. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/rsio/rsicr.jsp

Rev. Sci. Instrum., Vol. 74, No. 7, July 2003

Auger microscopy of W tips for STM

3369

TABLE I. Summary table of the preparation and post preparation treatments 共‘‘ex’’ and in situ兲 of the W tips. The chronological order for each tip treatment goes from top to bottom of the table. W tip Preparation and treatment KOH 共2 N兲 dc etching NaOH 共2 N兲 dc etching Ultrasonic cleaning 共acetone兲 Cleaning with HF Ar⫹ ion sputtering 共2 keV兲 in situ High field in STM 共ex situ兲

KOH共a兲

KOH共b兲

NaOH共a兲

NaOH共b兲

NaOH共c兲

NaOH共d兲

䊉 — — — 䊉 䊉

䊉 — 䊉共3⫹15 min兲 — 䊉 —

— 䊉 䊉15 min — 䊉 —

— 䊉 — — 䊉 䊉

— 䊉 — 䊉 — —

— 䊉 — 䊉 — 䊉

composition of the apex surface of W tips before and after the above mentioned postetching procedures. This work is devoted to fill in the above mentioned lack of experimental information, with the use of SAM with 10 nm nominal spatial and chemical resolution. II. EXPERIMENT

The SEM/SAM experiments have been performed using a model 15-680 scanning Auger nanoprobe system from Physical Electronics. The system operates in UHV with a base pressure of 2⫻10⫺10 Torr. The SEM images and the Auger spectra have been all acquired using an electron beam 共20 kV, 0.5 nA兲 routinely producing SEM images with 10 nm ultimate lateral resolution 共as measured on a gold on carbon standard sample at the analytical position on the stage兲. The energy resolution of the Auger spectra is better than 0.5% 关as measured by the full width at half maximum of a 1 keV elastic peak from a copper on allumina sample after subtraction of a linear background兴. The system can acquire, either every point Auger maps 共SAM maps兲 on the SEM images, or single point Auger spectra at selected points of interest individuated on the SEM images. The kinetic energy range spanned, during acquisition of the single point Auger spectra, is between 50 and 2000 eV. The ‘‘single point’’ operating mode was chosen for the present experiment. During acquisition of the Auger spectra on the selected points of the SEM images, an alignment software routine ensures that the beam is always actually pointing at the selected location of the sample and compensates for thermal drifts of the sample. These drifts, on the order of a few nm/min, if not compensated, can dramatically compromise the reliability and the actual lateral resolution of the Auger spectra at very high (105 ⫻) magnifications. The overall acquisition time of the Auger spectra was chosen as the best compromise between low noise level and high spatial resolution of the spectra themselves. SEM images are presented as taken, and Auger spectra are presented after a postacquisition numerical differentiation 共three points, third order Savitzky Golay algorithm兲. Table I summarizes the various treatments performed on the six W tips whose images and Auger spectra are presented in this work. Hereafter they are labeled KOH共a兲, KOH共b兲, NaOH共a兲, NaOH共b兲, NaOH共c兲, and NaOH共d兲. They are representative of a larger set of tips investigated in order to validate with statistical significance the results presented. All the W tips investigated were prepared by standard electro-

chemical dc etching using a solution of de-ionized water with 2 N of either KOH or NAOH. The tip drop-off time was electronically controlled and minimized as in Ref. 8. This method routinely produces W tips with similar aspect ratio 共width/length ratio兲, and with an apex radius of curvature in the 10–50 nm range 共as verified in this work兲. These tips routinely give atomic resolution on the Si共111兲7⫻7 standard surface as verified in a UHV system described elsewhere25 and hosting a VT-100 Omicron scanning tunneling microscope. Immediately after the dc etching the W tips have been thoroughly rinsed with de-ionized water. W tips are then mounted on standard Omicron stainless steel tip holders. Recalling what is summarized in Table I W tips have been studied with SEM/SAM: 共i兲 共ii兲 共iii兲

共iv兲

共v兲

as chemically etched 共after rinsing in de-ionized water兲; after 3 min 共or 15 min兲 acetone cleaning in an ultrasonic bath; after 48% 共or 25%兲 HF etching 共operation done ex situ, samples are then rinsed with distilled water, dried with ultrapure nitrogen flux, and immediately inserted in the UHV SAM system兲; after in situ argon sputtering 共2 kV beam voltage, 0.4 ␮A average ion beam current, 2⫻2 mm2 raster scan, 30 min cycles, ion gun-sample distance 20 cm, ion beam coincident with the tip axis兲; after operating the tips in the STM system25 at a sample to tip bias of ⫹10 V, 50 nA tunneling current, at tunneling distance from clean Si共111兲7⫻7, and at a fixed location of the sample for 60 s.

After every ex situ tip treatment procedure the samples have been immediately inserted in the UHV SAM system. Typical oxygen and carbon contamination of the tip surface could not be avoided. III. RESULTS AND DISCUSSION A. W tips as etched and after ex situ cleaning treatments

Figure 1 shows the SEM images 共at various magnifications兲 of the KOH共a兲 sample tip as inserted in the UHV system after dc etching with KOH solution. The image at lower magnification 关Fig. 1共a兲兴 is representative of the typical looking of all the dc etched tungsten tips of this work. Their aspect ratio is ranging from 0.16 to 0.21 共see Table II兲.

Downloaded 31 Oct 2003 to 192.84.136.174. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/rsio/rsicr.jsp

3370

Rev. Sci. Instrum., Vol. 74, No. 7, July 2003

Ottaviano, Lozzi, and Santucci

FIG. 2. Panel 共a兲 and 共b兲: single point Auger spectra 共10 nm lateral resolution during acquisition兲 taken at the A and B locations of Fig. 1共c兲, respectively. The main Auger lines for K, C, N, O, and W are labeled in panel 共a兲. The experimental data into an energy window containing the K (LM M ) 共at 252 eV兲 and the C KLL 共at 275 eV兲 Auger lines are redrawn 共see the insets in the middle of each panel, pointed by arrows兲 in the two panels with a three times magnified energy scale. Panel 共c兲 Auger spectrum of clean (Ar⫹ sputtered兲 W 共same primary electron beam energy of all the other spectra presented in this work, 20 keV兲.

FIG. 1. SEM images 共20 kV, 0.5 nA兲 of the KOH共a兲 W tip as inserted in the SEM/SAM system immediately after dc etching 共and rinsing with deionized water兲. Panel 共a兲: image at 2600⫻, panel 共b兲 image of the tip apex at 20 400⫻, panel 共c兲 image of the tip apex at 100 000⫻. Single point Auger spectra 共reported in Fig. 2兲 have been taken at the points marked A and B in this image.

Despite the rinsing treatment of each tip immediately after the dc etching, higher magnification images typically show, as in Fig. 1 关panels 共b兲 and 共c兲兴 on the KOH prepared tips, the presence of aggregates of residual contaminants. These aggregates can 关Fig. 1共c兲兴 contaminate even the very apex of the tip. Figure 2 关panels 共a兲 and 共b兲兴 reports the Auger spectra acquired on the tip of Fig. 1 at the 共A兲 and 共B兲 points marked in Fig. 1共c兲. As in the rest of this work, these spectra are really typical of the points marked on the figure with an

accuracy of 10 nm. In Fig. 1共c兲 the 共A兲 point is chosen at the tip apex, while 共B兲 is about 800 nm away from the apex along the tip axis. The peak energies 共and sensitivity factors at 20 kV primary beam energy兲 of the principal Auger lines of the elements of interest in this work are reported in Table III. Due to exposure to air 共and to the electrolyte兲 the Auger spectra presented in Figs. 2共a兲 and 2共b兲 共as in many other spectra reported in the following兲 typically present the O KLL signal 共peak energy at 510 eV in the derivative spectrum兲 accompanied by two smaller peaks at 473 and 489 eV kinetic energies, respectively. These minor features 共especially the one at 473 eV兲 are hardly distinguishable from the noise level in the Auger spectra of this work. Carbon, also almost ubiquitously present on the samples investigated, is detectable via the KLL Auger line at 275 eV. Nitrogen, if present as in Fig. 2 关panels 共a兲 and 共b兲兴, has a single typical

TABLE II. Summary of the geometrical parameters of the investigated W tips as prepared after electrochemical etching. Aspect ratios 共width/height兲 are calculated from the final 3 ␮m length portion of the various W tips estimated from SEM images at 20 000⫻. Tip apex radii are calculated from the very ending portion of each tip 关for example from the small protrusion on the rounded shape of the tip apex of NaOH共d兲 see Fig. 7共c兲兴.

Aspect ratio 共width/height兲 Tip apex radius 共nm兲

KOH共a兲

KOH共b兲

NaOH共a兲

NaOH共b兲

NaOH共c兲

NaOH共d兲

0.19 50

0.16 25

0.21 45

0.17 30

0.21 65

0.19 30

Downloaded 31 Oct 2003 to 192.84.136.174. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/rsio/rsicr.jsp

Rev. Sci. Instrum., Vol. 74, No. 7, July 2003

Auger microscopy of W tips for STM

TABLE III. Kinetic energies and sensitivity factors of the principal Auger lines of the elements of interest in the present work. The kinetic energy values are those calculated from the peak position in the Auger derivative spectra reported in Ref. 26. The sensitivity factors are reported normalized to the most intense W Auger line 关the W M 5 N 6,7N 6,7 (W1) at 1737 eV兴. They are calculated with a third order polynomial extrapolation to 20 keV primary electron beam energy from tabulated Auger sensitivity factors at 3.0, 5.0, and 10.0 kV in Ref. 26. The relative amplitudes for the W Auger lines 关and the corresponding sensitivity factors marked with 共 兲 in the table兴 * are estimated experimentally from an Auger spectrum taken with a 20 kV primary beam energy on clean Ar⫹ sputtered polycrystalline W 关see Fig. 2共c兲兴. Auger line C (KLL) N (KLL) O (KLL) Na (KLL) K (LM M ) Si (KLL) W (N 4,5N 6,7N 7 ) 关 W5 兴 W (M NN) 关 W6 – W7 兴 W (M NN) 关 W3 – W4 兴 W (M NN) 关 W1 – W2 兴

Kinetic energy 共eV兲

Sensitivity factor

275 389 510 996 252 1626 182 1312–1362 1524 –1571 1737–1798

0.29 共0.10兲 1.03 共0.10兲 2.03 共0.10兲 1.07 共0.10兲 1.67 共0.10兲 0.95 共0.10兲 0.85共*兲 共0.10兲 0.20 ( * ) – 0.11 ( * ) 共0.05兲 0.34 ( * ) – 0.22 ( * ) 共0.04兲 1.00 – 0.59 ( * )

fingerprint at 389 eV kinetic energy. The K (KLL) Auger line lies at 252 eV, close to the 275 eV peak typical of carbon. It is observed in Fig. 2共a兲. This is evident by comparison of the two insets of Fig. 2 关共a兲 and 共b兲兴, that report, on a magnified kinetic energy scale, the region around the C KLL Auger line. The spectrum acquired at point 共A兲 关panel 共a兲兴 at the tip apex of Fig. 1 shows a significant presence of K, which is completely absent in the spectrum of point 共B兲 关panel 共b兲兴. The substantial differences in the two spectra of Fig. 2 indicate that the elemental concentration of the W tip contaminants is not homogeneous over the tip surface: there are potassium conglomerates 共presumably crystallites of KOH兲 over the regular W tip surface that is essentially free from K. A 50 nm size cluster of residual contaminants with potassium is covering the tip apex. We note that spectra like the one presented in Fig. 2共a兲, showing a significant signal from K, always relate, with statistical significance, to aggregates observed on the tip surface in the SEM images. Obviously, W Auger lines are also detected in Fig. 2 关panels 共a兲 and 共b兲兴 and a reference Auger spectrum for clean W obtained by Ar⫹ ion sputtering onto polycrystalline W using a 20 keV electron beam is reported in Fig. 2共c兲. W has a complex Auger spectrum. In particular, the W Auger line of highest intensity for beam energy of 20 keV is the M 5 N 6,7N 6,7 line at 1737 eV. For brevity, we will hereafter refer to this line as the W1 line 共as marked in Fig. 2兲. We also label W2 and W3 the other M NN Auger lines at 1798 and 1524 eV. The triplet W1–W2–W3 is the W fingerprint in the Auger spectra at the high kinetic energy side. On the low kinetic energy side there is another rather intense Auger line, namely the N 4,5N 6,7N 7 共W5 in Table III兲 at 182 eV. As can be directly estimated in Fig. 2共c兲 共and as reported in Table III兲 the relative intensity of this line to the W1 one is 0.85. May we remark that this work is devoted mainly to the elemental chemical characterization of the W tips with high spatial resolution. A precise elemental quantitative analysis at the W

3371

tips’ surface is beyond the purposes of the present investigation, as well as a clear determination of the W chemical state. Nonetheless, to this respect, some remarks can be made. WO3 can be likely expected to be present as the native oxide layer that develops at the W-tip surface immediately after the etching procedure.12 Its thickness should be on the order of 3– 4 nm.12 From Table III the relative O/W Auger sensitivity can be estimated to be about 2 at 20 keV primary beam energy. Thus, the expectation O/W signal ratio for a pure WO3 layer should be about 6. Moreover, the O signal should be further enhanced by the presence of typical air contaminants 共like CO and CO2 as testified by the presence of the C Auger peak兲 and by the residual presence of KOH on the tip surface. On the other hand the measured O/W intensity ratio is 0.9 关both in panels 共a兲 and 共b兲 in Fig. 2兴, and more in general, an O/W 共⬇1兲 Auger intensity ratio is observed in all the W tips as etched. Evidently, due to the high kinetic energy typical of the secondary Auger electrons from the W M NN lines, the W Auger lines 共on the high kinetic energy side兲 are mostly probing the bulk metallic W underneath the native oxide layer on the W tip surface. Instead, an indirect, though clear, sign of the presence of WO3 at the surface of the as etched tips is the observation in Figs. 2共a兲 and 2共b兲 of a strongly reduced W5/W1 intensity ratio 关compared to that one of Fig. 2共c兲兴. In fact, the inelastic mean free path associated with the W1 共W5兲 electrons is 2.2 nm 共0.5 nm兲 for a clean tungsten matrix27 共these values increase to 3.2 and 0.6 nm for the W1 and W5 electrons, respectively, whether they escape from a WO3 matrix兲.27 Thus the W5/W1 ratio of 0.85 关as in Fig. 2共c兲兴 is observed only if W is homogeneously distributed in depth from the topmost layer of the surface investigated. Bulk metallic W, or bulk oxidized WO3 will show the same W5/W1 intensity ratio. Nonetheless if a WO3 layer is covering bulk metallic W, because of the reduced volume concentration of W in WO3 , this ratio is expectedly reduced. Moreover, a contaminant layer of hydrocarbon molecules and residual aggregates from the electrolyte should further decrease this ratio. Figure 3 is a summary of the SEM images taken at the highest magnification used 共100 000⫻兲 for the KOH共b兲, NaOH共a兲, NaOH共b兲, and NaOH共c兲, W tips. It summarizes the typical shape of the apex of the W tips as etched. SEM images are taken with the highest possible magnification compatible with reliable SAM measurements. From the original figures we estimate that the lateral resolution in the images of the tip apex is approximately 15 nm. This value is comparable with the highest nominal resolution of the 680 SAM from Physical Electronics 共10 nm兲. Obtaining SEM images with higher resolution, close to the instrumental limit of our system, was not possible because of mechanical vibration of the tip end. These vibrations are the source of the slight reduced resolution 共mainly lines in the scanning directions兲 that can be observed in Fig. 3 关panels 共b兲 and 共c兲兴. The high magnification SEM images presented in this work are comparable in quality and resolution with those presented recently in Ref. 18, while other work shows images with definitely lower resolution. Table II reports the measured values for each tip apex radius. A tip apex radius of 30 nm 共determined out of a set of ten tips兲 is observed in the aver-

Downloaded 31 Oct 2003 to 192.84.136.174. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/rsio/rsicr.jsp

3372

Rev. Sci. Instrum., Vol. 74, No. 7, July 2003

Ottaviano, Lozzi, and Santucci

FIG. 3. SEM images 共20 kV, 0.5 nA, 100 000⫻兲 of the apexes of: 共a兲 KOH共b兲; 共b兲 NaOH共a兲; 共c兲 NaOH共b兲; 共d兲 NaOH共c兲. In panels 共a兲 and 共c兲 the points are marked where the Auger spectra were taken for spatially resolved elemental chemical analysis. The spectrum acquired at point A in panel 共a兲 is identical to the one reported in Fig. 6共a兲. The spectrum acquired at point A in panel 共c兲 is reported in Fig. 4.

age, almost regardless of the electrolyte used, provided that the other details 共molar concentration of the electrolyte, length of the immersed W wire in the solution,7 and threshold voltage in the control circuit of the drop-off time7兲 during the etching procedure are not varied. Surface contamination with K has been also clearly detected on the KOH共b兲 tip, again with nonhomogeneous distribution on the scale length of 10 nm. For example, there is a remarkable K signal 共three times higher that the C one兲 detectable when pointing the electron primary beam to the flag shaped feature 关point A in Fig. 3共a兲兴 observed on the right side of the tip apex of Fig. 3共a兲. This spectrum is identical to the one reported in Fig. 6共a兲 关taken on the same location of the tip after ultrasonic cleaning, see also Fig. 5共a兲兴. A similar spectrum 共not shown for brevity兲 is acquired if pointing at 共B兲 in the same image, while no K signal is detected on point 共C兲 共also not shown for brevity兲. Microscopic aggregates of contaminants are instead absent when using NaOH as an electrolyte 共six tips investigated兲. We could never detect a significant Na Auger signal 共at 996 eV兲 on the W samples prepared with NaOH solution. Using a 20 keV primary electron beam, whether present in aggregates, sodium should be clearly observable with a well defined Auger line at 996 eV, which is characterized by an Auger sensitivity factor comparable to the one of W1 共see Table III兲. Instead, only a weak signal, with intensity hardly distinguishable from the noise level, can be discerned at 996 eV in Fig. 4 关acquired at point A in Fig. 3共c兲兴.28 The W tips of Fig. 1 and Fig. 3 were all identically processed. Thus the fact that residual aggregates of KOH are evidenced in Figs. 1 and 2 关and Fig. 3共a兲兴 and, correspondingly, similar aggregates are absent on the tips prepared with NaOH is rather intriguing. Ultimately, the only difference in the electrolytic solution is the size of the alkaline metal used: K 共220 pm atomic radius兲 is larger than Na 共180 pm atomic

radius兲. We speculate that this difference produces the different Auger spectra observed. Effectively, Na can diffuse in the WO3 surface layer, forming a Na–WO3 compound having the perovskite structure with the Na atoms in the interstitial position of the corner sharing network of the WO6 octahedra 共which are typical of the various distorted phases of cubic WO3 ).29 K, due to its size, cannot be accommodated in a similar interstitial position and K–WO3 only exists in the hexagonal phase.30 Thus, K can hardly be expected to diffuse into the WO3 surface layer. This can produce in one case 共Na兲 the homogeneous diffusion of the alkaline metal into the surface of the as prepared tip, and in the other case 共K兲 its surface coalescence. It remains to be explained why Na, which we conjecture to have homogeneously diffused into the surface of the tips, is hardly observed. This can be explained by the observation that the high kinetic energy Na Auger signal as the W 1,2,3 line is not particularly sensitive to the very topmost layers of the surface under investigation 共as similarly testified by the observation of an understoichiometric O/W ratio for the WO3 that is at the surface layer兲, and

FIG. 4. Single point Auger spectrum acquired at point A in Fig. 3共c兲 关tip NaOH共b兲兴.

Downloaded 31 Oct 2003 to 192.84.136.174. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/rsio/rsicr.jsp

Rev. Sci. Instrum., Vol. 74, No. 7, July 2003

Auger microscopy of W tips for STM

3373

FIG. 6. Effects of acetone cleaning in ultrasonic bath: 共a兲 single point Auger spectrum taken at point 共A兲 of Fig. 5共a兲 共tip after 3⬘ acetone cleaning兲; 共b兲 single point Auger spectrum taken at point 共A兲 of Fig. 5共b兲 共after 3⫹15 min cleaning兲. The spectra are reported in the two energy windows of the K, C, and main W lines.

FIG. 5. SEM 共20 kV, 0.5 nA兲 images of the effects of ultrasonic cleaning 共in acetone兲 on the W tips. Main frames 共insets兲 are taken with 100 000⫻ 共10 000⫻兲 magnification: 共a兲 tip KOH共b兲 after 3⬘ acetone cleaning in ultrasonic bath: the upper 共lower兲 inset shows 共with lower magnification兲 the appearance of the tip before 共after兲 cleaning. A single point Auger spectrum 关reported in Fig. 6共a兲兴 has been taken in 共A兲. The transversal white line on the tip apex mark approximately the break point due the subsequent ultrasonic cleaning 关shown in 共b兲兴. 共b兲 Close-up and 共upper right inset兲 zoomed out view of the KOH共b兲 tip appearance after a subsequent 15⬘ ultrasonic cleaning in acetone. A single point Auger spectrum 关reported in Fig. 6共b兲兴 has been acquired in A. 共c兲 Close-up and 共lower right inset兲 zoomed out view of the NaOH共a兲 tip appearance after 15⬘ ultrasonic cleaning in acetone 关the same tip apex is shown in Fig. 3共c兲 prior to the ultrasonic cleaning treatment兴.

accordingly only a very weak Na signal should be detected in the spectra 共see Fig. 4兲. Figure 5 is a collection of SEM images showing the effects, on a microscopic scale, of cleaning the as prepared tips in acetone using an ultrasonic bath. Two W tips 关KOH共b兲 and NaOH共a兲兴 have been submitted to this cleaning treat-

ment. A 3 min cleaning in ultrasonic bath in acetone partially removes the KOH conglomerates on the tip surface. This occurrence is evident by comparing the two low magnification SEM insets of Fig. 5共a兲. The upper one shows the as etched KOH共b兲 tip, and the lower one shows the same tip after the first ultrasonic cleaning cycle. Indeed, K is not completely removed. The flag shaped conglomerate of contaminants at the tip apex 关characterized by the Auger spectrum of Fig. 6共a兲兴 essentially survives the cleaning cycle. A subsequent longer 共15 min兲 ultrasonic cleaning of the same tip indeed removes the large aggregates with K contamination, as evidenced in the SEM image of Fig. 5共b兲 at higher magnification 共100 000⫻兲. Actually, the very apex of the tip 共approximately a portion of 250 nm length兲 has been cut 关approximately at the line artificially marked on the tip of Fig. 5共a兲兴 after the mechanical stress owed to the prolonged ultrasonic bath. The tip apex looks clean and sharp 共approximately 20 nm apex radius兲. A lower magnification image of the same tip is given in the upper right inset of Fig. 5共b兲. The last portion 共of approximately 1.5 ␮m length兲 is bent. K contamination is not completely removed after this procedure. Auger investigation on point 共A兲 of Fig. 5共b兲 reported in Fig. 6共b兲 still indicates a marked presence of potassium at the tip apex. The strong mechanical stress induced on the tip apex of the as prepared tips can also cause the complete blunt of the W tip as in the case of the NaOH共a兲 tip submitted to 10⬘ acetone ultrasonic cleaning, as shown in Fig. 5共c兲. HF etching after the tip fabrication can be likely believed to aggressively remove the contaminants layer on the tips’ surface. This treatment has been proposed in Ref. 13. It is based on the idea that in fact all the oxides of tungsten are soluble in concentrated HF31 while metal tungsten is inert to its attack.32 The quality of HF treated W tips was certified in Ref. 13 by routine production of atomically resolved STM images from as treated HF tips, and directly with SAM by Paparazzo et al.24 though using a very low 共1 ␮m兲 spatial

Downloaded 31 Oct 2003 to 192.84.136.174. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/rsio/rsicr.jsp

3374

Rev. Sci. Instrum., Vol. 74, No. 7, July 2003

Ottaviano, Lozzi, and Santucci

FIG. 7. SEM 共20 kV, 0.5 nA兲 images of the effects of ex situ HF etching on the W tips. Panel 共a兲: 共left兲 NaOH共c兲 tip as inserted 关100 000⫻, like Fig. 3共d兲兴, and 共right兲 after 30 s etching in HF 共48%兲, the tip diameter is 150 nm. Panel 共b兲 NaOH共c兲 tip 共100 000⫻兲 after 30⫹30 s etching in HF 共48%兲, the tip diameter is now 340 nm. Panel 共c兲: main frame NaOH共d兲 tip as inserted 共100 000⫻兲 共upper right inset, same tip at 10 000⫻兲. Panel 共d兲: main frame NaOH共d兲 tip after 60 s etching in HF 共25%兲 共100 000⫻兲 共upper right inset, same tip at 10 000⫻兲.

resolution. In our work, the microscopic effects of such treatment on the tip apex are reported in Fig. 7. Two KOH and two NaOH prepared tips were submitted to HF treatments. The removal of the oxide layer, and its etching time, does not depend on the electrolyte used. Thus only results of the NaOH prepared tips are presented. Figure 7共a兲 共left portion兲 shows the NaOH共c兲 tip apex as prepared after the electrochemical etching and 共right portion兲 after 30 HF 共48%兲 etching. The tip apex shape is substantially unaltered. Auger spectra taken at the tip apex 共not reported for brevity兲 indicate that the C and oxygen intensity 共relative to W兲 is also unaltered in the two spectra. Figure 7共b兲 shows the apex of the same NaOH共c兲 tip after a subsequent 30 s HF 共48%兲 etching ex situ. The tip is blunt, and comparison of the Auger spectra after and before such second treatment indicate that the W/C 共W/O兲 ratio increases by a 5.4 共2.3兲 factor 共similarly to the results reported in Ref. 24兲. Thus prolonged HF etching is effective in removing the C contaminants and reducing the thickness of the oxide layer, but significantly blunts the tip apex. May we note that this result is consistent with those presented in Ref. 13. There, it has been demonstrated that HF treated tips routinely produce STM images with atomic resolution on highly oriented pyrolitic graphite 共HOPG兲. Effectively, an oxide free tip apex surface like the one of Fig. 7共b兲, although quite dull, can very likely produce atomic resolution on typically 10⫻10 ␮m2 flat surfaces like HOPG. A more gentle treatment, with a 25% HF solution 共60 s etching time兲 still changes the tip apex 关see Fig. 7共c兲 before, and Fig. 7共d兲 after the HF treatment兴 but does not effectively remove the carbon and oxygen contaminants 共with respect to the as-etched tips兲 from the tip apex as verified by a substantially unvaried W/O and W/C line intensity ratio 共estimated from spectra taken at the apex of the tip, not shown for brevity兲. Thus HF is substantially not effective in removing the native

oxide layer at the W tips’ apex surface without altering and dulling the tip apex shape. B. In situ „UHV… tip treatments

Very interesting results are obtained when the W tips are submitted to argon sputtering in UVH in the SAM system. The ion beam direction and the tip axis were coincident with the ion gun facing the tip at a 20 cm distance. The ion guntip axis angle, as predicted in Ref. 18, can produce variations only of the aspect ratio of the very apex of the tip, so it is not a parameter of pivotal importance. Previous studies presented in Refs. 14 and 15 have been performed with SEM and TEM to observe the effects of ion milling on the tip apex shape. Indeed, the tip used had a fairly large apex radius in comparison to the ones routinely studied in the present work. SAM has also been used to this type of study with a much lower 共1 ␮m兲 lateral resolution in the Auger spectra.23,24 In the case of W tips prepared using KOH as electrolyte, Ar⫹ sputtering usually carves a needle like tip apex whose length almost linearly increases with time exposure to the ion beam. This was observed in three out of four KOH prepared tips. This is evidenced by the two high resolution images of the tip apex of the KOH共a兲 tip shown in Fig. 8共a兲: 60 min ion sputtering produces a cylindrical tip of about 380 nm length 关right side of Fig. 8共a兲兴; as the sputtering proceeds 关see the left side of Fig. 8共a兲兴 after another 120 min, the length of the needle increases to 1300 nm 共this length has been estimated from a 50 000⫻image not shown for brevity兲. A similar nanotip is formed by ion sputtering the apex of the KOH共b兲 tip 关see Fig. 8共b兲 and the lower magnification image of the same tip in the upper right inset of Fig. 8共b兲兴. We verified that the formation of such nanotips after ion sputtering is not peculiar of the W tips etched with KOH, although more likely in this case. This occurrence was found in just one out of three

Downloaded 31 Oct 2003 to 192.84.136.174. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/rsio/rsicr.jsp

Rev. Sci. Instrum., Vol. 74, No. 7, July 2003

Auger microscopy of W tips for STM

3375

FIG. 9. Single point Auger spectra 共10 nm lateral resolution兲 taken on various W tips after Ar⫹ sputtering in situ in UHV. Panel 共a兲: spectrum acquired at the black square dot of Fig. 8共a兲 关tip KOH共a兲, 180 min sputtering兴. Panel 共b兲: spectrum acquired at the black square dot of Fig. 8共b兲 关tip KOH共b兲, 120 min sputtering兴. Panel 共c兲: spectrum acquired at the black square dot of Fig. 8共c兲 关tip NaOH共b兲, 90 min sputtering兴.

FIG. 8. SEM 共20 kV, 0.5 nA兲 images of the effects of in situ Ar⫹ sputtering on the W tips 共the points artificially marked with a black square dot on the W tips in the images indicate the locations where the single point Auger spectra, presented in Fig. 9, were taken兲. Panel 共a兲: Tip KOH共a兲 共right兲 after 60 min sputtering and 共left兲 after another 120 min sputtering 共both images at 100 000⫻兲; panel 共b兲 tip KOH共b兲 after 120 min sputtering: main frame at 100 000⫻ and upper right inset at 10 000⫻; panel 共c兲 tip NaOH共b兲 after 90 min sputtering 共main frame兲, the upper left inset reports the same tip after only 30 min sputtering, the black line in the inset indicates approximately the etching front after 90 min sputtering 共the two images are both taken with 100 000⫻兲.

NaOH prepared tips. Argon ion sputtering on the NaOH共b兲 sample tip does not produce a nanotip at its apex. The tip apex is etched away, and the radius of curvature of the tip front is substantially unchanged. The inset in Fig. 8共c兲 shows the appearance of the NaOH共b兲 after 30 min sputtering. It

appears similar to the shape of the same tip as inserted 关see Fig. 3共c兲兴. The portion of the tip apex on the left side of the line artificially superimposed to the SEM image in the inset marks approximately the front of the tip apex after the sputtering 共90 min兲 共shown with higher magnification in the same panel兲. Even more interesting are the Auger spectra relative to the sputtered tips reported in Fig. 9. In fact, the Auger spectra taken on the surface of the nanotips of Fig. 8 关panels 共a兲 and 共b兲兴 indicate that these apex protrusions are not composed by clean tungsten. The spectrum reported in Fig. 9共a兲 is representative of many spectra acquired along the nanotip axis. Spectra were taken close to the very apex of the tip 关see for example the black square dot marked in Fig. 8共a兲兴 and at 150 nm intervals along the nanotip axis. It is noteworthy that no significant systematic variations of the relative intensities of the Auger lines were observed along the nanotip axis 共and at its apex兲 and between spectra taken after 60 min sputtering 关relative to Fig. 8共a兲 right兴 and spectra acquired after a successive cycle of another 120 min sputtering 关relative to Fig. 8共a兲 left兴. As a matter of fact, sputtering is only increasing the length of the needle-like nanotip, as observed in Fig. 8共a兲, and is not effectively altering its chemical composition. All the spectra, as in the representative one of Fig. 9共a兲, show a significant decrease of the C Auger line intensity 共if compared to the spectra of the as etched W tips兲. Indeed the carbon contamination is the only one effectively removed from the nanotip surface by the sputtering procedure. There

Downloaded 31 Oct 2003 to 192.84.136.174. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/rsio/rsicr.jsp

3376

Rev. Sci. Instrum., Vol. 74, No. 7, July 2003

is a strong 共the most intense line in the spectrum兲 K signal that, as opposed to the as etched tip surfaces, is almost homogeneously found with similar intensity along the nanotip axis. A significant signal from Fe is also found in the spectrum of Fig. 9共a兲 关as well as in panels 共b兲 and 共c兲兴. Fe is coming very likely from secondary Fe ions sputtered from the stainless steel tip holder 共typically 1–1.5 mm away from the W tip apex兲 and backscattered toward the tip apex. Oxygen seems not to be effectively removed from the nanotip surface. Indeed, the O/W intensity ratio after ion sputtering increases to an average value of 2.3, significantly and substantially higher than the 1.0 ratio observed in the average at the tip apex before the ion sputtering. Indeed, in some point spectra, as in Fig. 9共a兲, this ratio approaches the expectation value of 6 that should be typical of pure WO3 . The W M NN Auger kinetic energy in Fig. 9共a兲 is 4.5 eV lower than the one observed in the other spectra presented in this work. This indicates the presence of oxidized W in agreement with previous Auger studies onto clean and oxidized surfaces of W.23–24,33 Moreover, in the spectrum of Fig. 9共a兲 the W5/W1 intensity ratio approaches the expectation value of 0.9 as in the case of homogeneous in depth distribution of W in the surface under investigation. Similar experimental evidence 共increased O/W intensity ratio at the tip apex after sputtering, and chemical shift of the W1 line toward lower kinetic energies兲 is found for the chemical composition of the needle like tip of Fig. 8共b兲 whose typical Auger spectrum 共taken at the tip apex兲 is reported in Fig. 9共b兲 关in this case the K/W and O/W intensity ratios are significantly different from those observed in Fig. 9共a兲兴. On the other hand argon sputtering 共90 min兲 of the NaOH共b兲 tip 关Fig. 8共c兲兴 produces a surface whose chemistry at the very apex of the tip is given by the spectrum of Fig. 9共c兲. In this case the O signal is reduced, Fe ions are still implanted on the tip surface, and the surface is mostly composed by metallic tungsten. The results found in the latter case for the NaOH共b兲 tip are consistent and in good agreement with other works reported in the literature on the effects of Argon sputtering on W tips.18,24 Effectively, in agreement with our findings, Paparazzo et al. 共although with significantly lower spatial resolution in their analysis兲 have reported that, the W/O Auger intensity ratio at the apex of W tips 共similarly prepared with NaOH兲 is systematically and monotonically increased as a function of Argon sputtering time.24 Moreover, our observation of the etching of the front of the NaOH共b兲 tip without significant changes of the tip apex shape, is in agreement with that observed experimentally, and predicted by numerical Monte Carlo simulations in Ref. 18. After a bombardment time, which depends on the specific conditions used, the tip reaches its final shape, and further bombardment does not change the foremost part of the tip. An enlightening analogy is that one with the macroscopic effect of a pencil sharpener. Once the apex of the pencil is formed, further sharpening does not change the shape of the pencil apex, rather its overall length is decreased. The formation of the nanotip apex after ion sputtering has never been reported previously. The ‘‘pencil sharpener’’ effect is not observed in this case: the nanotip length instead linearly increases with ion etching time. This effect can be

Ottaviano, Lozzi, and Santucci

explained either assuming that the sputtering yield of the material forming the nanotip is lower that the one of bulk W, or assuming that the very apex of the tip is somewhat shielding the portion of the tip underneath. This latter explanation seems more likely. As noticed in Ref. 18 carbon contaminant of the very apex of the tip could play a role in the ion shield, due to the very low sputter yield of carbon. This happens, as we observed once, also in the case of NaOH prepared tips provided that there is significant carbon contamination at the tip apex. In the case of tips prepared with KOH, the survival of potassium 共rather that C兲 to the ion etching and its uniform distribution along the nanotip axis is a rather surprising observation. Effectively, potassium, as the other alkali metals, has a very low sublimation energy,34 and, accordingly, should be very easily sputtered away. One possible explanation for this observation, which would also justify the observed uniform distribution of K 共not anymore present in the form of aggregates like on the as etched W tip surface兲, might be K diffusion into the nanotip. As we already mentioned K does not likely form a stable compound with pseudocubic WO3 , nonetheless another stable structure is possible, K–WO3 in the hexagonal phase with the K intercalated into the hexagonal channels of this peculiar phase.30 The formation of hexagonal K–WO3 promoted by Ar sputtering remains the only likely explanation for our experimental evidences. Another method to remove the WO3 oxide layer covering the tip surface in UHV is annealing the tip above 725 °C. The WO3 phase diagram indicates that, above this temperature, the following reaction takes place: 2 WO3 ⫹W →3 WO2 . Because WO2 is volatile a metal W surface is generated.35 Annealing the W tip can be implemented via direct thermal contact with a W filament resistively heated,12 or by electron bombardment in UHV,23 or by direct resistive heating the tip by flowing elevated currents through the tip apex.11 The latter method can be easily used during STM measurements in UHV just increasing the tunneling current between the tip and a clean surface. This method was routinely adopted by our group to obtain high quality STM images with atomic resolution of metal induced reconstructions onto Si共111兲.36 We have examined the actual effect of operating the tip in tunneling conditions under high bias 共10 V positive sample–tip bias voltage兲 and elevated tunneling currents 共50 nA兲 onto reconstructed Si共111兲7⫻7. Figure 10 shows the final appearance of some of the as treated tips 关namely KOH共a兲, NaOH共b兲, and NaOH共d兲兴. Evidently, due to the high current density 共that can be estimated on the order of 104 A/cm2 ) on the thin apex of the tips, which induces a strong temperature gradient on the tip apex, they reproducibly assume a coiled shape. SAM spectra of the W tip prepared with NaOH and annealed in UHV with electron bombardment are presented in Ref. 23 showing the removal of the oxide layer. Also in our case the high temperatures (T ⬎725 °C) certainly reached by the tip apex after the abovementioned treatment, must have very likely induced the sublimation of the WO3 layer. We have demonstrated that this happens at the expense of a dramatic worsening of the tip apex shape. The observation of improved tunneling condition

Downloaded 31 Oct 2003 to 192.84.136.174. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/rsio/rsicr.jsp

Rev. Sci. Instrum., Vol. 74, No. 7, July 2003

Auger microscopy of W tips for STM

3377

FIG. 11. Single point Auger spectrum acquired at the point A of Fig. 10共c兲 共upper right inset兲 on the NaOH共d兲 tip apex after operating the W tip in tunneling conditions at high bias 共⫹10 V sample–tip bias兲 and high tunneling current 共50 nA兲. The Si signal is detectable in the spectrum. Si is owed to field evaporated atoms from the Si共111兲 sample used for scanning in the STM system 共Ref. 25兲.

10共c兲, is particularly interesting. There is a significant Si signal at the tip apex. This is an interesting finding that directly indicates that migration of matter has occurred from the sample to the tip during the operation at elevated bias and tunneling current in the UHV STM system. This matter transfer can be originated from field evaporation of Si from the Si surface to the W tip 共due to the high bias applied兲 or by direct contact of the tip apex with the sample as speculated in Ref. 20. In this respect the sample–tip bias polarity seems unimportant. This is an experimental demonstration of transfer of atoms from the sample to the tip apex during an STM experiment. ACKNOWLEDGMENTS

Dr. M. Renzetti and Tiziana Lepidi are acknowledged for valuable experimental support. Financial support from the European Union 共INFM Progetto Sud兲 is also acknowledged. G. Binning and H. Rohrer, Helv. Phys. Acta 55, 726 共1982兲. R. P. Feynman, California Institute of Technology 共CALTECH兲 Engineering Science; February 共1960兲 具http://www.zyvex.com/nanotech/ feynman.html典 3 G. A. Fiete, J. S. Hersch, E. J. Heller, H. C. Manoharan, C. P. Lutz, and D. M. Eigler, Phys. Rev. Lett. 86, 2392 共2001兲. 4 A description of the various tip etching procedures can be found in Sec. 3.2.1 of: T. T. Tsong, Atom Probe Field Ion Microscopy 共Cambridge University Press, Cambridge, 1990兲. 5 M. Cavallini and F. Biscarini, Rev. Sci. Instrum. 71, 4457 共2000兲. 6 C. S. Chang, W. B. Su, and T. T. Tsong, Phys. Rev. Lett. 72, 574 共1994兲. 7 A. J. Melmed, J. Vac. Sci. Technol. B 9, 601 共1991兲. 8 J. P. Ibe, P. P. Bey, Jr., S. L. Branbow, R. A. Brizzolara, N. A. Burnham, D. P. Di Lella, K. P. Lee, C. R. K. Marrian, and M. J. Colton, J. Vac. Sci. Technol. A 8, 3570 共1990兲. 9 M. Bode, O. Pietzsch, A. Kubetzka, and R. Wiesendanger, J. Electron Spectrosc. Relat. Phenom. 111, 1055 共2001兲. 10 V. T. Binh and N. Garcia, Ultramicroscopy 42, 80 共1992兲. 11 R. M. Feenstra, J. A. Stroscio, and A. P. Fein, Surf. Sci. 181, 295 共1987兲. 12 O. Albrektsen, H. W. Salemink, K. A. Mørch, and A. R. Tho¨le´n, J. Vac. Sci. Technol. B 12, 3187 共1994兲. 13 L. A. Hockett and S. E. Creager, Rev. Sci. Instrum. 64, 263 共1993兲. 14 D. K. Biegelsen, F. A. Ponce, J. C. Tramontanta, and S. M. Koch, Appl. Phys. Lett. 50, 696 共1987兲. 15 D. K. Biegelsen, F. A. Ponce, and J. C. Tramontanta, Appl. Phys. Lett. 54, 1223 共1989兲. 16 W. Hauffe, and E. Paneff, Proceedings of the 11th European Conference on Electron Microscopy, Dublin, 1996, Vol. 1, p. 1. 17 R. Zhang and D. G. Ivey, J. Vac. Sci. Technol. B 14, 1 共1996兲. 1 2

FIG. 10. SEM 共20 kV, 0.5 nA兲 images of the effects of operating the W tips in a STM system at high voltages 共⫹10 V sample–tip bias兲 and high tunneling currents 共50 nA兲. Panel 共a兲 tip KOH共a兲 at 100 000⫻ 关previous tip apex shape shown in Fig. 8共a兲兴. Panel 共b兲 tip NaOH共b兲 at 100 000 关previous tip apex shape shown in Fig. 8共c兲 main frame兴. Panel 共c兲 tip NaOH共d兲 at 10 000⫻ 关previous tip apex shape shown in Fig. 7共d兲兴 and, upper right inset, image of the tip apex at a different view angle at 50 000⫻. The point A marks the location where the single point Auger spectrum of Fig. 11 has been taken.

that we verified after such violent treatment36 is only justified, after the direct visualization of the tip apex shape, hypothesizing an occasional formation of clean oxide free Si nanotips still suitable for giving atomic resolution images on atomically flat surfaces like those investigated in Ref. 36. From this respect, although in our case we could not avoid C and O recontamination of the W tip treated in the STM chamber, because of the sample transfer, the spectrum shown in Fig. 11, taken at the end point of the curly apex of Fig.

Downloaded 31 Oct 2003 to 192.84.136.174. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/rsio/rsicr.jsp

3378

Rev. Sci. Instrum., Vol. 74, No. 7, July 2003

P. Hoffrogge, H. Kopf, and R. Reichelt, J. Appl. Phys. 90, 5322 共2001兲. S. Heike, T. Hashizume, and Y. Wada, J. Vac. Sci. Technol. B 14, 1522 共1996兲. 20 J. Garneas, F. Kragh, K. A. Mørch, and R. A. Tho¨le´n, J. Vac. Sci. Technol. A 8, 441 共1990兲. 21 C. J. Chen, Ultramicroscopy 42, 1653 共1992兲. 22 Y. Kuk and P. J. Silverman, Appl. Phys. Lett. 48, 1597 共1986兲. 23 A. Cricenti, E. Paparazzo, M. A. Scarselli, L. Moretto, and S. Selci, Rev. Sci. Instrum. 65, 1558 共1994兲. 24 E. Paparazzo, L. Moretto, S. Selci, M. Righini, and I. Farne`, Vacuum 52, 421 共1999兲. 25 L. Ottaviano, G. Profeta, A. Continenza, S. Santucci, A. J. Freeman, and S. Modesti, Surf. Sci. 464, 57 共2000兲. 26 K. D. Childs, B. A. Carlson, L. A. LaVanier, J. F. Moulder, D. F. Paul, W. F. Stickle, and D. G. Watson, in Handbook of Auger Electron Spectroscopy, 3rd ed., edited by C. L. Hedberg 共Physical Electronics, Eden Prairie, 1995兲. 27 S. Tanuma, C. J. Powell, and D. R. Penn, J. Vac. Sci. Technol. A 8, 2213 共1990兲. 18 19

Ottaviano, Lozzi, and Santucci 28

We remark that point Auger spectra with acquisition times longer then those typical of the spectra presented in this work could not be taken without a significant worsening of the spatial resolution. 29 T. Vogt, P. M. Woodward, and B. H. Hunter, J. Solid State Chem. 144, 209 共1999兲. 30 J. Bludska and I. Jakubec, Z. Phys. Chem. 共Munich兲 194, 69 共1996兲. 31 CRC Handbook of Chemistry and Physics, edited by D. R. Lide 共CRC, Boston, MA, 1990兲. 32 C. A. Hample, The Encyclopedia of the Chemical Elements 共Reinhold, New York, 1968兲, p. 768. 33 Y. Goldstein, A. Many, S. Z. Weisz, M. Gomez, O. Resto, and M. H. Farias, J. Electron Spectrosc. Relat. Phenom. 67, 511 共1994兲. 34 G. P. Chambers and J. Fine, in Practical Surface Analysis, 2nd ed., Ion and Neutral Spectroscopy, Vol. 2, edited by D. Briggs and M. P. Seah 共Wiley, New York, 1992兲. 35 N. E. Promisel, Tungsten, Tantalum, Molybdenum, Niobium, and Their Alloys 共Pergamon, New York, 1964兲, p. 293. 36 L. Ottaviano, M. Crivellari, L. Lozzi, and S. Santucci, Surf. Sci. 445, L41 共2000兲.

Downloaded 31 Oct 2003 to 192.84.136.174. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/rsio/rsicr.jsp