ULTRA-HIGH VACUUM SOFT X-RAY REFLECTOMETER

ULTRA-HIGH VACUUM SOFT X-RAY REFLECTOMETER Maurizio Sacchi, Carlo Spezzani and Piero Torelli Laboratoire pour l'Utilisation du Rayonnement Electromag...
Author: Dinah Carr
3 downloads 1 Views 202KB Size
ULTRA-HIGH VACUUM SOFT X-RAY REFLECTOMETER

Maurizio Sacchi, Carlo Spezzani and Piero Torelli Laboratoire pour l'Utilisation du Rayonnement Electromagnétique (LURE) Centre Universitaire Paris-Sud, Bât. 209d B.P. 34, 91898 Orsay (France)

Antoine Avila, Renaud Delaunay and Coryn F. Hague Laboratoire de Chimie Physique - Matière et Rayonnement (LCP-MR) Université Pierre et Marie Curie 11, rue Pierre et Marie Curie, 75005 Paris (France)

ABSTRACT We have designed, built and tested a new instrument for soft x-ray scattering experiments. The reflectometer works under ultra-high vacuum and permits in situ preparation and characterization of the samples. In particular, deposition and sputtering operations can be performed while measuring x-ray scattering. We report the results of test measurements performed using synchrotron radiation. The precision of the combined positioning of sample and detector angles is better than 0.01°. Separately, sample and detector rotations have a repeatibility that is better than 0.005°. Applications will be in the field of surface physics, with emphasis on magnetic properties of surfaces, thin films and multilayered structures.

1

I.

INTRODUCTION

X-ray reflectometers are designed to work under vacuum either to avoid x-ray absorption by air, or because a windowless connection to a high vacuum x-ray source is required. In the former case they can function in the 10-1 - 10-5 mbar range, in the latter a pressure of 10-7 -10-8 mbar is needed. It means that the two-axis goniometer required to perform x-ray scattering measurements must operate under vacuum and various technical solutions have been reported in the literature (see, e.g., Refs. 1 to 5). To our knowledge, vacuum constraints have seldom been dictated by the in situ clean sample preparation standards commonly in force for surface science experiments i.e., pressures in the order of 10-10 mbar [3]. The instrument described here is intended for resonant x-ray scattering studies under such UHV conditions, using tuneable, polarized soft x-ray synchrotron radiation. By resonant scattering it is meant that the x-ray photon energy is finely tuned to an inner shell absorption edge. Following on the pioneering work of Kao et al. [6], many applications for resonant scattering of polarized soft x-rays have been found in the study of magnetic materials [7-12]. Work has mainly centered on artificially modulated structures such as multilayers [8,9,11] or striped domains [12]. The activity of our group in Orsay started in 1994, with a series of studies on in situ-prepared clean surfaces [13] and thin films [14] aimed at extending x-ray resonant magnetic scattering (XRMS) techniques to surface science. The lack of an adequate ultra-high vacuum (UHV) goniometer limited our studies to photon-energy dependent measurements at fixed scattering angles. In this paper we report on the construction, technical performance, and first results of a new two-circle goniometer mounted on a UHV chamber. The instrument is hereafter referred to as IRMA [15].

2

I Fig. 1. Schematic drawing of the two-axes goniometer. The photon beam is normal to the page, at the crossing of the two dashed lines.

I. INSTRUMENT DESIGN A Goniometer The two-circle goniometer

design is shown schematically in Fig. 1. In the following,

numbers in parenthesis refer to the notation in Fig. 1. The rotation of the central shaft (1) defines the angular position of the sample, that of the external hollow shaft (2) the position of the detector. The shafts are made of 316L stainless-steel. The rotation of each shaft about its common axis is actuated by two commercial rotary-stages (6) driven by high-precision encoded stepper-motors [16]. The rotary stages are rigidly connected by means of five struts (8) to a CF160 UHV flange (7) that serves to connect the goniometer to the vacuum chamber. Two dry sleeve-bearings (3) and (3'), set 350 mm apart, maintain the central shaft (1) coaxial with respect to (2). In this way, the two shafts can rotate independently and (1) can also slide along the rotation axis relative to (2). The working surfaces of (1), (2) and (3) are ground to an RMS roughness of 0.8 µm and a 5 µm mechanical tolerance. The shafts were subjected to an anti-sticking treatment to avoid seizing under vacuum [17]. Without altering the

3

mechanical finish, this procedure results in hardened, low friction surfaces that require no lubricant and can be baked to 150 °C even in UHV. Two double-stage differentially-pumped feedthroughs maintain UHV conditions during rotation. One (4) is mounted between (2) and (7), the other (5) between (1) and (2). It is worth underlining that these feedthroughs ensure that the system is vacuum-tight upon rotation, but the precision of the coaxial angular motions depends on the high precision rotary stages only.

B Sample holder Three orthogonal translations along x, y and z (x is the rotation axis, y lies in the plane of the sample surface, and z is normal to the surface) are provided for aligning the sample in addition to the rotation θS. The displacement along x is achieved by translating the entire sample shaft with respect to the detector arm. A standard motorized UHV linear motion manipulator (9) with an accuracy of ±10 µm over a range of 150 mm is used for this purpose. y and z translations are obtained by using UHV compatible motors and translation tables (12) mounted between the sample shaft (1) and the sample holder (11). Two set ups are available. One uses an ensemble of stepper-motors [18] and custom built vacuum tables capable of positioning a sample to within ±1 µm (backlash corrected) over a range of 12 mm and sustaining an off center load of 0.5 N·m. The other is a commercial two-axis table (encoded piezoelectric device) with an accuracy of 0.1 µm over a range of 20 mm compatible with an off-center load of 0.1 N·m. The stepper-motor precision is sufficient for most applications. The higher precision of the piezoelectric system is required for future applications using micron-sized photon beams to analyze small magnetic objects [19]. Custom-designed sample holders can be attached to the translation tables according to specific needs, such as an electromagnet capable of a applying a magnetic field of about 700

4

Oe (2 kOe pulsed) along the sample surface either parallel or perpendicular to the scattering plane.

C Detector holder A detector arm (10) is attached to the external shaft (2). It can carry up to three diodes and a channel electron multiplier (CEM) at a radial distance of 150 mm from the rotation axis. We employ either 10x10 mm2 Si diode-detector [20], capped with 400Å of aluminum to reduce sensitivity to visible light, or GaAs diodes with an active surface of 4x4 mm2 or 2x2 mm2 [21]. Diodes and CEM can be fitted with slits of variable aperture between 0.1 and 1 mm (vertical accepted angle between 0.04° and 0.4°). The CEM [22] can be used either in positive or negative front voltage mode to measure electrons or photons, respectively. For photon detection, the slit in front of the CEM can be biased in order to reduce possible contributions from positively charged ions in the chamber. For complete visible light and electron tightness, a thin Al foil (500 nm) can be mounted on the slit aperture.

D. UHV chambers Because of its relatively small size and CF160 mounting flange, the goniometer unit can be installed or dismounted readily from an experimental chamber, thus it can be used in connection with existing UHV systems. As part of the IRMA project, we have developed two dedicated UHV chambers that complete the final set up of the reflectometer. Scattering measurements and some in situ treatments of the sample take place in the main chamber, where the goniometer is mounted. A second chamber, also under UHV, is connected to the main chamber. It has more versatile sample preparation capabilities, involving standard surface science techniques. Within the main chamber, several operations can take place at the sample position:

5

- Sample sputtering, using a differentially pumped ion gun with precise positioning, focussing and raster of the ion beam. The sputtering angle can be varied continuously between 0 and 45°. - Two evaporation sources point at the sample surface, 30° and 60° relative to the axis of the incoming photons. Two types of sources are available: a tungsten basket, for use with ingots, heated by joule effect or electron bombardment and a tungsten crucible, also heated by electron bombardment. The latter is water cooled to dissipate the radiated heat. Evaporation is collimated (less than 5° spread) by a small aperture placed between the source and the chamber. To limit the pressure rise during evaporation, both sources are differentially pumped. A gate-valve separates the sources from the chamber so reloading can take place without breaking the vacuum. - Two view ports are available in the scattering plane, pointing at the sample, and positioned 90° apart, making it possible to perform in situ Kerr rotation analysis. The laser source, polarizers, photoelastic modulator and detector are thus located outside the vacuum chamber. - An external ccd camera provides a view of the sample (the field of view is variable from 15x20 mm to 90x120 mm).

Thin film deposition and ion sputtering can also take place in the preparation chamber, with in addition the possibility of heating the sample up to 1300° C and of performing Auger and LEED analysis for the characterization of the sample surface. The preparation chamber can be mounted in-line with the incoming x-ray beam, making it possible to characterize the sample by x-ray absorption and photoemission spectroscopies. The preparation chamber is equipped with a fast entry lock system for loading new samples without breaking the vacuum. Via a transfer system consisting of wobble sticks and transfer rods, the sample can be moved between the two chambers, while maintaining UHV conditions.

6

III. PERFORMANCE A. Vacuum It is planned to operate with a combination of turbo, cryogenic, and ion pumps. The values reported below refer to measurements with a 550 l/s turbomolecular pump in use on the main chamber and with the goniometer fully operational, i.e., including motors, detectors, and electromagnets. A base pressure of 1x10-8 mbar was attained in 24 h without bakeout. A pressure of 2x10-10 mbar was reached following a 48 h differential bakeout i.e., different bakeout temperature settings for various parts of the reflectometer. Our goal of 1x10-10 mbar should be reached quite straightforwardly with the full complement of pumping units. No increase in pressure was observed during normal θS and/or θD scans (maximum speed of 10 °/s) after slight initial outgassing. The pressure of Ar required to sputter with 1 keV ions and a 1-2 µA sample current leads to a total pressure of 1x10-8 mbar in the main chamber (5x10-9 mbar for a 0.3-0.5 µA sputtering current). This is an important feature because it means that Ar-sputtering may be performed with the system opened onto the x-ray beam while recording reflectivity data. The evaporators can be used with hardly any observable increase in the base pressure thanks to preliminary independent outgassing, water cooling, and differential pumping. However the large distance to the sample (250 mm) limits the effective evaporation rate to a maximum of about 5 Å/min (see below). A new mounting is being developed to move the source closer to the sample via a linear translator for use when higher rates are required.

B. Mechanical

7

The encoded stepper motors that drive the rotations of both sample and detectors have a resolution of 0.001° and a repeatability of 0.002°. Axial misalignment is < 10 µrad for a complete rotation, corresponding to a vertical positioning error of approximately ±2 µm at the sample position. Taken alone, these values surpass the performance required. What matters in reflectivity measurements, however, is the simultaneous control of both incident and scattered angles. This depends on several parameters: 1. Coaxial rotation of the two axes. This is defined by the precision of the machined parts and assembly procedures. 2. Mechanical coupling between rotations. The sample and detector shafts are in contact via the two bearings (3) and (3'). This is not a problem as long as both motors move in the same direction, but their backlash is in effect coupled. To solve this problem the software that controls the rotations corrects the backlash on both axes regardless of which one moves back. 3. Intersection of the photon beam and rotation axes. The incoming photon beam must intersect the goniometer axis exactly and be normal to it. It means that either the goniometer or the photon beam axes must be adjustable. In practice the goniometer is attached to the main chamber via a CF160 bellow fitted with three micrometer screw adjustments. The goniometer axis may thus be positioned precisely with respect to the photon beam after a rough alignment of the entire chamber. 4. The scattering surface must be situated at the exact crossing of the beam and rotation axes. This depends on the precision of the in-vacuum z-adjustment.

High performance is therefore the result of both accurate construction and careful alignment. Test experiments using synchrotron radiation have been carried out at the SuperACO beamline SU-7 (LURE, Orsay) and at the X-ray Circular Polarization beamline at ELETTRA (Trieste).

8

θS (max) - θD / 2 (°)

Fig. 2.

0.03 0.02 0.01 0.00 -0.01 -0.02 -0.03

intensity around the specular condition (θS = θD / 2 ± 0.2°) for θD in the range 10-150°. For each θD, θS 0

10

20

30

40

θD / 2

50

60

70

(b)

scans are normalized between 0 and 1 for better

80

comparison. b) : Plot of the deviation of θS from

(°)

the expected specular position, as a function of θD.

140

Filled circles refer to the data shown in (a). Open

120

circles are obtained after lowering the rotation axis of 1 mm with respect to the beam axis. Open

(°)

100

θD

a) : Gray scale map of the scattered

squares are obtained after reoptimizing the

80

alignment. 60 40

(a)

20 -0.15 -0.10 -0.05

0.00

0.05

0.10

0.15

θS - θD / 2 (°)

Fig. 2a presents a series of measurements where, for each position of the detector angle θD between 2° and 150°, the sample angle θS is rotated around the nominal value for specular reflectivity. For perfect alignment, the specular condition (i.e. maximum scattered intensity) is obtained at θS (max) = θD / 2. Fig. 2b (filled circles) shows the deviation θS (max) - θD / 2 from this ideal value, as a function of θD: the maximum deviation over the measurable range of the instrument is within ±0.01°. To illustrate the importance of point (3) above, we deliberately introduced a misalignment between the photon beam and the rotation axis by moving the latter down by 1 mm. This leads to an increased spread of ±0.03° in θS (max) - θD / 2 (open circles in fig. 2b), with a clear trend from the beginning to the end of the scan. Realigning the rotation axis led to the results denoted by open squares in Fig. 2b. Twenty repetitions of the same scan of a single axis (either θS or θD) always led to equivalent curves within 0.005°, i.e. within the tolerance for angular positioning imposed by the software.

9

These results show that the goniometer fully complies with the designed mechanical performance.

Difference

1.0

Fig. 3. Bottom: specular reflectivity from a

x 100

Co/Cu multilayer exhibiting antiferromagnetic

0.5

coupling between adjacent Co layers. Photon

0.0

energy is 777 eV (Co 2p edge). Line and dots

Reflectivity (arb. units)

-0.5 10

2

10

1

10

refer to opposite helicities of the circularly Charge Bragg peaks

st

1 order

10 10

-2

10

-3

10

-4

10

-5

multiplied by 10 for better display. The

nd

0

-1

polarized photons. The continuous line is

2 order

difference curve is reported in the top panel.

st

1 order

rd Antiferro 3 order Bragg peaks

0

5

10

15

20

25

30

35

θS (°)

IV. RESULTS A. Samples prepared ex situ Our first test was concerned with metallic Co/Cu multilayers displaying giant magnetoresistivity and already analyzed on other reflectometer endstations at ALS [11] and BESSY. Fig. 3 shows an example of specular scans measured with our new reflectometer mounted on the Circular Polarization beamline at ELETTRA. The photon energy (777 eV) is set at the Co 2p absorption edge in order to have maximum sensitivity to the magnetic properties. The beamline is equipped with an electromagnetic helical undulator and the helicity of the circularly polarized x-rays can be switched easily and without affecting the beam characteristics. The bottom panel of figure 3 shows two specular reflectivity curves obtained for opposite helicities. The angular location of charge and antiferromagnetic peaks is indicated (see ref. [11] for a more detailed discussion of this sample). The top panel shows the corresponding difference curve. The quality of these data compares well with results 10

previously obtained on beamline 6.3.2 of the Advanced Light Source [4] and on the UE56PGM1 polarimeter at BESSY [5].

(a)

θS = 45°

(b)

Reflectivity (arb. units)

TEY (arb. units)

4' 18' 25'

150

160

Fig. 4. Absorption spectrum (a) and 45° specular reflectivity (b) at the Er-N4,5 edge after three different evaporation times.

170

180

190

200

Photon energy (eV)

B. samples modified in situ As we mentioned before, the UHV system has been designed in such a way that scattering measurements can be performed under deposition or sputtering conditions. As an example, we deposited in situ an Er thin films on a silicon substrate. We first calibrated the deposition rate using a quartz balance mounted behind the sample. The maximum rate attainable in optimum vacuum was 5 Å/min, with a total dissipated power of 75 watts on the W basket. Fig. 4a shows the Er N4,5 absorption spectrum recorded in the total electron yield (TEY) mode for three deposition times of 4 min, 18 min, and 25 min. Specular reflectivity curves at θS = 45° were simultaneously recorded, and are displayed in Fig. 4b. At 175.5 eV photon energy (maximum of the giant resonance), the absorption monotonically increases with the Er thickness, which is not the case for the reflectivity curves. Fig. 5 shows the reflectivity at θS =

11

45° and hν = 175.5 eV as a function of deposition time. Starting from t = 8 min, these data have been recorded continuously during the evaporation, with only a break at 18 min for taking an energy dependent spectrum. The oscillating reflectivity versus deposition time originates from the oscillating phase difference between the waves reflected at the vacuum/Er and Er/Si interfaces, as a function of the Er layer thickness. A quantitative analysis of the reflectivity curves estimated the deposition rate at about 5.7 Å/min.

Estimated thickness (angs.) 0

25

50

75

Scattered intensity (pA)

100

100

125

Fig. 5. Specular reflectivity at 45° and 175.5

hν = 175.5 eV

eV as a function of the evaporation time. 80

The three arrows indicate where the spectra of Fig. 4 were taken.

60

40

20

θS = 45° ; θD = 90°

0 0

5

10

15

20

25

Evaporation time (minutes)

C. Scattered intensity Although this point is not specific to our new instrument, the important issue of the range of fractional intensity that can be measured in a scattering experiment should also be discussed. Using our GaAs photodiodes, we can reliably measure currents down to about 50 fA, which, based on the diode characteristics, correspond to an estimated photon flux of about 105 - 104 photons/s at energies between 100 eV and 1000 eV. Considering 1011 photons/s as a typical incoming flux at an undulator synchrotron beamline, this means that we can cover about 6 to 7 orders of magnitude of the intensity in a scattering experiment, from direct beam monitoring to wide angle diffusion analysis. It is in order to extend the minimum flux that can be detected that we used a CEM in combination with the diodes. Our CEM gives very good performances

12

from a few counts per second (cps) up to 106 cps. Considering that its efficiency is in the order of 10% for soft x-rays, it allows photon fluxes as low as 10-100 photons/s to be analyzed and it offers a large overlap of about two decades with the GaAs diode (fig. 6). In this way, the measurable range of scattered photon flux can span 9 orders of magnitude for an incoming flux of 1011 photons/s.

6

10

1

10

θD = 15° 5

hν = 777 eV

0

10

10

-1

10 4

10

-2

10 3

10

-3

10

CEM counts 2

10

-2

Photodiode current (pA)

CEM (counts per second)

Fig. 6. Rocking scan of the sample for a Diode current

detector angle of 15°. Dots: diode measurement. Line: CEM measurement. The direct beam current on the diode was ~100 nA at 777 eV.

-4

0

2

4

6

8

10

12

14

16

10

θS (°)

Finally, it should be stressed that the IRMA project was developed primarily for resonant magnetic scattering experiments on clean surfaces and in situ prepared thin films. The "magnetic" and "in situ" aspects, though, are not strict requirements, and the instrument will be open as a synchrotron users' facility for general soft x-ray scattering experiments. The reflectometer is currently operational at the SU-7 beamline of the SuperACO storage ring at LURE and it will be installed in the near future at one of the beamlines planned for the new synchrotron source SOLEIL

ACKNOWLEDGEMENTS We acknowledge the expert technical assistance of LURE and LCP-MR staff. In particular, we wish to thank Jean-Maurice Berset, André Szwec, Jean Walocha (LURE) and Hugues

13

Ringuenet (LCP-MR). We also wish to thank Stefano Turchini and the staff of the Circular Polarization beamline at ELETTRA (Sincrotrone Trieste), where part of the instrumental tests was carried out. The project IRMA has received the financial support of the Centre National de la Recherche Scientifique (France).

REFERENCES 1. H. Hogrefe, D. Giesenberg, R.-P. Haelbich, and C. Kunz, Nucl. Instrum. and Methods 208, 415 (1983) 2. W.R. Hunter and J.C. Rife, Nucl. Instrum. and Methods A 246, 465 (1986) 3. Y. Hirai, I. Waki, A. Momose, and K. Hayakawa, Rev. Sci. Instrum. 60, 2219 (1989) 4. E.M. Gullikson, J.H. Underwood, P.C. Batson, and V. Nikitin, J. X-Ray Sci. Tech. 3, 283 (1992); J. H. Underwood and E. M. Gullikson, J. Electron Spectrosc. Relat. Phenom. 92, 265 (1998) 5. F. Schäfers, H.-Ch. Mertins, A. Gaupp, W. Gudat, M. Mertin, I. Packe, F. Schmolla, S. Di Fonzo, G. Soullié, W. Jark, R. Walker, X. Le Cann, M. Eriksson, R. Nyholm, Appl. Opt. 38, 4074 (1999). 6. C.C.Kao, J.B.Hastings, E.D.Johnson, D.P.Siddons, G.C.Smith, and G.A.Prinz, Phys. Rev. Lett. 65, 373 (1990) 7. C.C.Kao, C.T.Chen, E.D.Johnson, J.B.Hastings, H.J.Lin, G.H.Ho, G.Meigs, J.M.Brot, S.L.Hubert, Y.U.Idzerda, and C.Vettier, Phys. Rev. B50, 9599 (1994) 8. J.-M. Tonnerre, L. Sève, D. Raoux, G. Soullié,B. Rodmacq, and P. Wolfers, Phys. Rev. Lett. 75, 740 (1995) 9. M. Sacchi, C.F. Hague, A.Mirone, L.Pasquali, P.Isberg, J-M.Mariot, E.M.Gullikson, and J.H.Underwood, Phys. Rev. Lett. 81, 1521 (1998) 10. Y.U.Idzerda, V. Chakarian, and J.W. Freeland, Phys. Rev. Lett. 82, 1562 (1999)

14

11. C. Spezzani, P. Torelli, M. Sacchi, R. Delaunay, C.F. Hague, F. Salmassi, and E.M. Gullikson, Phys. Rev. B 66, 052408 (2002) 12. H.A. Durr, E. Dudzik, S.S. Dhesi, J.B. Goedkoop, G. van der Laan, M.Belakhovsky, C. Mocuta, A. Marty, and Y. Samson, Science 284, 2166 (1999) 13. M.Sacchi, J.Vogel, and S.Iacobucci, J. Mag. Mag. Mater. 147, L11 (1995) 14. M.Sacchi, A.Mirone, and S.Iacobucci, Surface Sci. 442, 349 (1999) 15. IRMA (Instrument pour la Réflectivité MAgnétique) has been designed and built jointly by the Laboratoire pour l'Utilisation du Rayonnement Electromagnétique (LURE, Université Paris-Sud, Orsay) and the Laboratoire de Chimie Physique - Matière et Rayonnement (LCP-MR, Université Paris VI, Paris). 16. MicroContrôle-Newport RV160PP. 17. Kolsterizing ® 18. Phytron VSS-25 19. The reflectometer, as an endstation, is part of the project of a new beamline with micronsize x-ray spot to be developed on the SOLEIL synchrotron source. 20. International Radiation Detectors, model AXUV-100-Al2 photodiode 21. Hammamatsu Photonics, models G1126-2 and G1127-2 GaAsP photodiodes 22. Dr.Sjuts Optotechnik, KBL 5RS/R channel electron multiplier

15

Suggest Documents