Nuclear Instruments and Methods in Physics Research A 467–468 (2001) 1481–1484

Surface science station of the infrared beamline at SPring-8

M. Sakuraia,*, T. Moriwakib, H. Kimurab, S. Nishidac, T. Nanbac

aPhysics Department, Kobe University, 1-1 Rokko-dai, Nada, Kobe 657-8501, JapanbSPring-8/JASRI, Mikazuki, Hyogo 679-5198, Japan

cGraduate School of Science and Technology, Kobe University, Kobe 657-8501, Japan

Abstract

An experimental station for surface science has been constructed at the infrared beamline (BL43IR) of SPring-8,Japan. The station utilizes synchrotron radiation in the energy range of 100–20000 cm�1 to perform infrared reflectionabsorption spectroscopy (IRAS) of surfaces. It consists of an experimental section, a preparation chamber, gas

handling equipment and a pair of focusing optics. In situ observation of vibrational spectra is possible using both IRASand high-resolution electron energy loss spectroscopy. # 2001 Elsevier Science B.V. All rights reserved.

PACS: 07.85.Qe; 42.82.Bq; 68.35.Bs; 68.35.Ja

Keywords: Infrared synchrotron radiation; Optical system; Infrared reflection absorption spectroscopy; Surface vibration; SPring-8

1. Introduction

Vibrational spectroscopy is a useful technique insurface characterization. Methods of vibrationalspectroscopy on surfaces are restricted to opticalmethods and an electron spectroscopy; high-resolution electron energy loss spectroscopy(HREELS). Among various optical techniquesinfrared reflection absorption spectroscopy(IRAS) is a standard technique applicable to theanalysis of adsorption system on well-definedsubstrates. Ordinary IRAS uses an FT-IR inter-ferometer and utilizes light sources installed in theinterferometer. IRAS is superior in resolutionwhile HREELS has better sensitivity and a wider

spectral range. The drawback of ordinary IRAS isthat the observable vibrational modes are limitedbecause of dipole selection rules and the spectraldistribution of the light source. Utilization ofsynchrotron radiation (SR) for the light source ofIRAS partly resolves the latter limitation. IRASwith an SR source is superior to ordinary IRAS inthe effective photon flux, low emittance of the lightsource, and especially in the far-infrared (FIR)region where ordinary light sources are not useful.IRAS measurements in the FIR region revealedsoft vibrational modes never observed with ordin-ary IRAS [1,2]. Since IRAS and EELS havecomplementary characteristics as techniquesfor surface vibrational spectroscopy, it is veryinformative to have spectroscopic data obtainedby in situ observation with both IRAS andHREELS techniques over whole spectral rangecorresponding to possible vibrational modes.

*Corresponding author. Tel.: +81-78-803-5648; fax: +81-

78-803-5770.

E-mail address: sakurai@phys.sci.kobe-u.ac.jp (M. Sakurai).

0168-9002/01/$ - see front matter # 2001 Elsevier Science B.V. All rights reserved.

PII: S 0 1 6 8 - 9 0 0 2 ( 0 1 ) 0 0 7 2 7 - 6

IRAS with an SR light source together with EELSin the same vacuum system has great advantages.A few reports on IRAS measurements in the FIRregion with an SR source in the world exist [3],however, the present apparatus for surface scienceemploys both IRAS and EELS techniques for thefirst time. In the present paper, we describe thedesign of experimental station for vibrationalspectroscopy on the infrared beamline 43IR ofSPring-8.

2. Design

The beamline consists of a Fourier-transforminterferometer (Bruker 120HR) and four experi-mental stations. The characteristics of upstream ofthe beamline and other stations will be describedelsewhere in this issue [4–7]. The present station isdesigned to accommodate to various kinds ofsamples. It consists of an experimental section, apreparation one, gas handling one and a pair offocusing ones. The detailed drawing of the stationis given in Fig. 1. The station has characteristicsdescribed below:

1. In situ observation of vibrational spectra ispossible using both IRAS and HREELSapparatus.

2. Two kinds of windows (diamond and Si) areused to cover a wider energy range, and areinterchangeable in vacuum without majoradjustment of the optical components of thefocusing system.

3. A sample in a pre-evacuated load-lock chambercould be introduced to the experimental sectionvia a preparation chamber in order to shortenthe time necessary to reach UHV.

2.1. Optics design for IRAS

Arrangement of optical elements for IRAS isillustrated in Fig. 2. The source of the IR radiationhas the acceptance angle of 36.5(horizon-tal)� 12.5(vertical) mrad2 from the full field re-gions of the bending magnet [4]. Upstream of thebeamline is designed to supply vertically polarizedparallel beam with rectangular cross section

(10mm in vertical scale and 3mm in horizontalone) [5]. The input focusing optics focuses thebeam on the sample surface, and the outputfocusing optics focuses the diverging beam re-flected from the sample on a detector. The inputfocusing optics consists of a plane mirror and anoff-axis parabolic mirror with the focal length of300mm, while the output focusing optics consistsof two off-axis parabolic mirrors with the focal

Fig. 1. Detailed drawing of whole system of the surface science

station at BL43IR of SPring-8.

Fig. 2. Arrangement of optical elements for IRAS measure-

ment.

M. Sakurai et al. / Nuclear Instruments and Methods in Physics Research A 467–468 (2001) 1481–14841482

length of 300 and 150mm. Since the polarizationdirection of the incoming beam on the sampleshould be horizontal, a pair of mirrors of the inputfocusing optics are directed to rotate the polariza-tion direction by 908.Vacuum windows are located to separate the

vacuum between the experimental section andboth focusing optics. A diamond and Si windowsare mounted on an exchange mechanism of thewindows. We used CVD-grown diamond win-dows, which are 0.3mm in thickness and have a1.28 wedge in order to avoid oscillatory modula-tion of power spectrum due to interference effect.The Si window is used as a window for the FIRregions. They are also made into wedged shape,and the wedge angle is so chosen that thedeflection angles coincide. This two-window sys-tem is expected to avoid the interference over theenergy range of 100–20000 cm�1.

2.2. Detail of the experimental section

The whole apparatus except commercial parts isdesigned and assembled by the Vacuum Products,Co. The experimental section is the major part ofthe station and miscellaneous apparatus areattached on three levels: IRAS, LEED and EELSstages. The diameter of the IRAS stage is100mmf. The LEED and EELS stages are atthe lower part of the experimental section(400mmf). The IRAS stage comprises an ion-gun, a mass spectrometer, a gas doser, a pumpingport, and input and output windows that intersectUHV region from the focusing optics sections inlow-vacuum. A LEED optics is installed on theLEED stage. Samples could be introduced via thepreparation section into the LEED stage ofthe experimental section. An EEL spectrometer isinstalled on the EELS stage. To avoid contamina-tion to the LEED and EELS stages during gasdosing, a gate valve is located between the IRASand LEED stages. The valve is integrated directlyin the vacuum chamber at between the IRAS andLEED stages to shorten the traveling distance ofthe sample between the IRAS and EELS stages. Inorder to investigate an adsorption system at lowtemperature, the manipulator of sample isdesigned to accommodate either of a liquid He

cryostat or a closed cycle refrigerator, which couldbe mounted on a rotary motion feed through atthe top of a XYZ translator. The spans oftranslational motion in X-, Y- and Z-directionsare 12.5, 12.5 and 410mm, respectively. Therotational motion and translational one in Z-direction are driven by stepping motors. At theend of the cryostat or refrigerator, a sample-mountstage and a carbon heater are attached. A samplemounted on a sample holder, which is transferablebetween the three chambers, could be heated up to1100K and cooled down to 40K. A sampledirectly attached on the cryostat could be cooleddown to near liquid helium temperature, and isalso heated up to 1300K.

3. Experimental results

Spectral distribution of the SR, which wastransmitted to the surface science station asFig. 2 illustrates, was measured using a HgCdTedetector (curve a in Fig. 3). The intensity corre-sponds to the ADC counts of the FT-IR. The SRlight is transmitted in vacuum through whole ofoptical passes, and is reflected by the surface ofPt (1 1 1) attached on the sample holder of the

Fig. 3. Spectral distribution of effective photon intensity at the

surface science station. Curves a and c were measured at the

resolution of 1 cm�1 and with 100 times scan, and curve b was

measured at the resolution of 4 cm�1 and with 400 times scan.

Curves a and b correspond to the spectra of SR, while curve c is

the spectrum of the globar source in the interferometer.

M. Sakurai et al. / Nuclear Instruments and Methods in Physics Research A 467–468 (2001) 1481–1484 1483

experimental section. The surface of sample ismirror-polished by the manufacturer, and hasenough quality for using it as a mirror for IRlight. A spectrum measured with a black bodysource (globar) in the FT-IR interferometer is alsoshown for comparison (curve c in Fig. 3). Spectraat the FIR region have not been measured yetbecause of the problem described later. Thespectra a and c were measured at the resolutionof 1 cm�1 and with 100 times scans. The curve b inFig. 3 was measured at the resolution of 4 cm�1

and with 400 times scans to check the reduction ofthe noise. Since the geometric configuration of theIRAS optics require low emittance of the lightsource, the photon flux of the SR available to theIRAS measurement is 5–6 times higher than thatof the globar source. Several dips exist in thespectra around 2000–2500 cm�1. They correspondto the absorption by two diamond windows in thisstation.As for the photon flux, the optical system of

whole the beamline including focusing optics forIRAS works well, however, the signal-to-noiseratio is not enough for the IRAS measurementthat requires the detection of the difference of less

than 1% in the reflected intensity between thesample and background spectra. This noiseproblem, which is common for this infraredbeamline, seems that comes from electric noise atthe circuit system around the IR detector, andcould be resolved without any serious reconstruc-tion of the beamline.

References

[1] C.J. Hirschmugel, G.P. Williams, F.M. Hoffmann,

Y.J. Chabal, Phys. Rev. Lett. 65 (1990) 480.

[2] C.J. Hirschmugel, G.P. Williams, Phys. Rev. 52 (1995)

14 177.

[3] D.A. Slater, P. Hollins, M.A. Chesters, J. Pritchard,

D.H. Martin, M. Surman, D.A. Shaw, I.H. Munro, Rev.

Sci. Instrum. 63 (1992) 1547.

[4] H. Kimura et al., Nucl. Instr. and Meth. A 467–468 (2001),

these proceedings.

[5] S. Kimura, H. Kimura et al., Nucl. Instr. and Meth. A 467–

468 (2001), these proceedings.

[6] H. Okamura et al., Nucl. Instr. and Meth. A 467–468

(2001), these proceedings.

[7] S. Kimura, T. Nanba et al., Nucl. Instr. and Meth. A 467–

468 (2001), these proceedings.

M. Sakurai et al. / Nuclear Instruments and Methods in Physics Research A 467–468 (2001) 1481–14841484