Chapter 7. X-ray Nano-probe. 7.1 Introduction

Transcription

1 Chapter 7 X-ray Nano-probe 7.1 Introduction The X-ray Nanoprobe (XNP) Beamline Project was approved as one of the first seven beamlines at the Taiwan Photon Source (TPS). The XNP beamline and the associated instruments are designed to utilize the high brilliance of the TPS light source to resolve atomic, chemical and electronic structures of semiconductor-based devices with a spatial resolution of tens of nanometers in a tomographic and nondestructive manner. The beamline optics are designed to deliver a focal spot of ca. 40 nm in diameter with photon fluxes around photons/sec in the energy range of 4 15 kev. The major scientific projects planned to be performed at the XNP beamline are mostly pioneering works closely related to practical semiconductorbased industrial applications. Some of the proposed projects are listed below: Characterizations of 22 nm node and beyond in semiconductor devices using nano-exafs and nano-scattering/diffraction Probing local structures, electronic structures, and X-ray-to-visible downconversion mechanism in nanocrystals/quantum dots/quantum wires of semiconductors, oxides, and phosphors Nanoprobe beam for material characterization of MEMS/NEMS devices Based on the existing X-ray techniques, the proposed beamline is expected to provide structural probes with atomic, elemental and chemical resolutions. These X-ray techniques include: 321

2 322 CHAPTER 7. X-RA NANO-PROBE nano-xrf (X-ray fluorescence) for element-specific nanoimaging nano-xafs (X-ray absorption fine structures) element-specific for local electronic structure and chemical environments, averaged over nanoscale area nano-xeol (X-ray excited optical luminescence) for X-ray-to-visible down-conversion efficiency in nano-phosphors In order to fulfill the diverse requirements of various X-ray techniques, several design criteria are set for a modern XNP beamline: (1) minimizing the number of beamline optical elements to preserve the source coherence; (2) nano-focusing optics with a diffraction-limited spot; (3) high throughput optics; (4) achromatic optics are preferred; (5) a large working distance; (6) minimizing the mechanical and ground instabilities. The beamline will consist of one horizontal focusing mirror (HFM), a horizontally diffracting monochromator before the nano-focusing optics. It will operate without X-ray windows, maintaining a vacuum environment from the upstream beamline optics to the samples. The nested Montel mirrors are chosen as the nano-focusing optics for reasons of easy alignment, being achromatic in a wide energy range, and a high photon throughput. The ultimate focal spot is expected to be around 40 nm at the sample, with photon fluxes in the neighborhood of photons/sec with an energy resolution of The end station will provide conventional X-ray probes such as X-ray fluorescence (XRF), X-ray absorption fine structures (XAFS), and X-ray excited optical luminescence (XEOL). Emerging techniques, such as Braggptychography (BP), will be developed for study of the strain dynamics of nanodevices. An off-line SEM will be provided as a complementary surface probe. The scientific merits of this beamline will be outlined in Section 7.2. The characteristic parameters of the photon source, IU22, will be described in Section 7.3. The optical design of the beamline along with the expected beamline performance simulated by ray-tracing programs will be shown in Section 7.4. The major optical components such as a DCM and mirrors, mechanical considerations, and vacuum considerations will be detailed in Section 7.5. The design of the end station, due to its complexity, is still ongoing. What is presented in Section 7.6 is only preliminary and conceptual. However, some progress in design will be given. Radiation safety issues follow

3 7.2. SCIENTIFIC OPPORTUNITIES 323 in Section 7.7, although related designs are still conceptual at this moment. The construction timeline will be outlined in Section 7.8, followed by a discussion of the commissioning plan in Section 7.9. A list of the construction team will end the chapter in Section The following is a brief summary of the design features of the beamline: Focusing optics (nested Montel mirrors) Spatial resolution (40 nm with focusing optics, less than 10 nm with coherent technique) X-ray flux (energy range 4 15 kev, flux > photons/s at the sample position) Beamline design (two-stage horizontal focus with deflective optics) Horizontal deflection geometry for the DCM Multi-axis sample stage, with sub-10 nm precision or better for its translational degrees of freedom (possible as long as the temperature is uniform and well controlled) Probes (n-xrf, n-xafs, n-xeol, Bragg-Ptychography, SEM) Bragg Ptychography for strain mapping Designs of the detectors are being deliberated and not yet specified Data reduction strategy (an important issue under further analysis) 7.2 Scientific Opportunities The major scientific projects planned to be performed at the proposed beamline are mostly pioneering works closely related to practical industrial applications. Some of them are listed below: 1. Characterizations of IC circuits of 22 nm node and beyond using nano- EXAFS and nano-scattering/diffraction The semiconductor industry is facing unprecedented challenges in materials and physics due to the limitations of Si CMOS transistor scaling arising from non-scaling of matters, namely gate dielectrics and channel

4 324 CHAPTER 7. X-RA NANO-PROBE mobility. The drive for alternative high κ dielectrics (to replace SiO 2 ) and metal gates initiated a decade ago has resulted in a 2007 Intels news announcement of high κ+ metal gate transistor breakthrough on 45 nm microprocessors, which went into production lines in However, even with very aggressive adoption of new methods for performance enhancements, the industry may have run out of technologies. As driven by continual demands for a faster speed of enhanced transport in channels and reducing power dissipation beyond the 22 nm node (see below), to extend the Moores Law, the current consensus is that we urgently need to employ Ge and III-V semiconductors as high-mobility channels integrated with high κ gate dielectrics for future CMOS technology. Realization of these new MOSFETs has presented great challenges not only for material scientists and processing engineers, but also for condensed matter and device physicists as a whole. The proposed research covers a broad, but essential spectrum of research activities ranging from fundamental nano-science, nano-materials, to nano electronic devices. 2. Probing local structures, electronic structures, and X-ray-to-visible downconversion mechanism in nanocrystals/quantum dots/quantum wires of Intel Transistor Scaling and Research Roadmap Figure 7.1: Intel transistor scaling and research road map

5 7.2. SCIENTIFIC OPPORTUNITIES 325 semiconductors, oxides, and phosphors The nanocrystals/quantum dots/quantum wires of a wide variety of material systems such as semiconductors, oxides, and phosphors exhibit many unique physical properties with great potentials in industrial applications. For example, nanocrystals of semiconductors in the shapes of dots, wires, and rings have been demonstrated to possess widened bandgaps, ultra fast transitions, and improved photoluminescence efficiencies due to the quantum confinement effects. Oxides such as the high-hardness cubic zirconia system have also shown much enhanced structural stability in the form of nanocrystals compared with that of the bulk. To understand the underlying mechanism leading to these favorable properties, it is of essential importance to obtain the information pertaining to the local structures and electronic structures surrounding the dopants and constituent atoms in these nanoscale structures. For the sake of distinction between the signals from nanoparticles under investigation and those from their surfactants or substrates, nano-exafs, nano-xanes and nano-xeol measurements using focused X-ray beam on samples with sparsely distributed nanoparticles will be employed. The full-field X-ray microscope will also be very useful for directly characterizing these material systems. 3. Microprobe beam for material characterization of MEMS/NEMS devices Microbeam is an indispensable tool to characterize the structure of microsensors or micro-actuators nondestructively. For recently developed MEMS /NEMS devices, the inspection tool usually is SEM/TEM. Although the SEM/TEM offers a better lateral resolution for select area diffraction than X-ray microscopy, the X-ray microprobe offers a better penetration into the sample and a less tedious sample preparation procedure, especially for non-conductive samples. The microprobe can offer a microbeam which is able to probe the atomic structure by X-ray diffraction and scattering, to obtain the chemical information by micro-xas, and to do imaging by scanning the sample in micro-steps. With undulators of a high brilliance, TPS will offers a much better lateral resolution and intensity than the current TLS. As such time resolved micro-beam investigation will become feasible. In the applications of MEMS/NEMS devices, such as micro-sensors and micro-actuators, problem of lifetime estimation and diagnosis is an important factor to improve the quality of the devices. For example, corrosion of a micro sensor can be monitored over the lifetime nondestructively. A

6 326 CHAPTER 7. X-RA NANO-PROBE protection layer can be added to prevent the device from degradation. How the protection layer acts as time evolves can be tested, and material diffusion problem and compound formation can be studied. To give as a real example, the development of a solid oxide fuel cell (SOFC), which could be an important energy saving device in a small town or village in the future, faced a serious corrosion problem due to its operation at a high temperature and cycling all the time. Micro-channels are also cut to transport the fuel (hydrogen) and oxygen efficiently in the SOFC, which present a complex geometry to be scrutinized. Microprobe X-rays can be quite useful in observing the detailed structure change in the SOFC, even when it is operated at a high temperature. For a micro-actuator, the stress causing fatigue always limits its service lifetime. How the stress and cracks develop during its lifetime is an interesting question. When the size of a device is close to the single grain size, the macroscopic picture with traditional metallurgical knowledge is sometimes not applicable. This is where microprobe X-rays can come in to reveal the intricacy of the problem. 7.3 Photon Source The beamline utilizes the light source generated from an in-vacuum undulator IU22 with a nominal period length of 22 mm. The characteristic parameters of the IU22 are tabulated in Table 7.1. The IU22 is positioned at the center of a 7 m straight section of the TPS. With a minimum gap of 5 mm, the IU22 is able to emit X-rays in the energy range of kev by varying the undulator gap and switching among the harmonic number The spectra of the IU22 in terms of brilliance and flux are calculated by SPECTRA[1] and shown in Figure 7.1. The horizontal and vertical photon source sizes are calculated to be 282 µm and µm (in full-width at half-maximum, FWHM), respectively, and the horizontal and vertical source divergences, µrad and µrad (FWHM), respectively. Source Parameters The electron beam in an undulator is modulated by a periodic magnetic field defined by the magnetic gap and these parameters determine the wavelength of the emitted light. The deflection parameter K and the photon

7 7.3. PHOTON SOURCE 327 wavelength of the nth harmonic λ n are given by and K = eb 0λ u 2πmc = λ ub 0, (7.1) λ n = λ u(1 + K2 ) 2 E 2 n where K : deflection parameter e : electronic charge, Coulomb B 0 : peak magnetic field (Tesla) λ u : length of the magnet period (cm) λ n : photon wavelength of the n th harmonic (Å) K: electron energy (GeV) n : number of odd harmonics (n =1, 3, 5...), (7.2) The above equations show that K is proportional to the magnetic field and a large K value results in extended range at the lower energy end. Taking into consideration of the initial operating parameters of the TPS and the required energy range, the minimum acceptable vertical gap is 5 mm and a period of 22 mm is chosen. The in-vacuum undulator IU22, named after its period length, will be placed in a short straight section of the storage ring to provide the required hard X-rays for the Nanoprobe Beamline. The undulator parameters are shown in Table 7.1. Source Brilliance and Flux Due to the low emittance of the TPS ring, at 500 ma storage ring current and a magnetic gap of 5 mm, the brilliance in the range of 5 20 kev calculated by SPECTRA[1], using harmonics #3 to #15, is greater than photons s 1 mr 2 mm 2 (0.1%bw) 1, as shown in Figure 7.2(a). The photon flux is above photons s 1 (0.1%bw) 1, as shown in Figure 7.2(b). Source Size and Divergence The effective beam size (σ eff ) and divergence (σ eff ) are obtained by convolving the diffraction-limited beam size (σ photon ) and divergence (σ photon ), with the electron beam source size, and divergence, as shown in the following equations:

8 328 CHAPTER 7. X-RA NANO-PROBE Table 7.1: Source Parameters of the IU22 at the TPS In-vacuum undulator IU22 Photon energy (harmonic numbers 3 15) kev Period length, λ u 22 mm Number of period, N period 140 Peak field 1.05 T Deflection parameter, Ky max 2.15 Total magnetic length 3.1 m Minimum magnet gap 5 mm Total power at 500 ma 9.7 kw Power density at 500 ma 65 kw/mrad 2 σ photon = 1 2Lλ, 4π (7.3) λ σ photon = 2L, (7.4) σ eff = σelectron 2 + σ2 photon, (7.5) σ eff = where σ electron : electron beam source size (σ x,y ) σ electron : electron beam divergence (σ x,y) σ 2 electron + σ 2 photon, (7.6) Figure 7.3 and Figure 7.4 show the source size and divergence of the IU22, in the unit of FWHM. The horizontal source size varies insignificantly around 282 µm while the vertical size varies from 14.8 to 12.5 µm in the energy range of 4 20 kev. The change in vertical divergence over this energy range is from 27.5 to 21.9 µrad, and the horizontal divergence is from 48.2 to 45.1 µrad. Both source size and divergence are superior to the current TLS ring. In particular, the vertical divergence is very small and collimation of the photon beam is excellent, so that a collimating mirror is not necessary

9 7.3. PHOTON SOURCE Brilliance (phs/s/mr 2 /mm 2 /0.1%bw) Brilliance (phs/s/mr 2 /mm 2 /0.1%bw) IU22 IU22 mm, L = 3 m, 22 Kmm, max = L 2.15 = 3 with m, gap 5 mm K max = 2.15 with gap 5 mm Photon 10 Energy 15 (kev) Photon Energy (kev) Flux (ph/s/0.1%bw) Flux (ph/s/0.1%bw) IU22 IU22 mm, L = 3 m, 22 Kmm, max = L 2.15 = 3 with m, gap 5 mm K max = 2.15 with gap 5 mm Photon 10 Energy 15 (kev) Photon Energy (kev) Figure 7.2: The brilliance (a) and the flux (b) emitted by a 3 m-iu22 at a ring current of 500 ma. while still maintaining the energy resolving power of this beamline.

10 330 CHAPTER 7. X-RA NANO-PROBE Horizontcal, FWHM Source size ( m) Vertical, FWHM Photon Energy (kev) Figure 7.3: Horizontal (upper) and vertical (lower) source sizes of the IU22 as a function of energy.

11 7.3. PHOTON SOURCE 331 Source angular divergence ( rad) Horizontcal, FWHM Vertical, FWHM Photon Energy (kev) Figure 7.4: Horizontal (upper) and vertical (lower) source divergences of the IU22 as a function of energy.

12 332 CHAPTER 7. X-RA NANO-PROBE 7.4 Beamline Optical Design Diverse experimental requirements on the XNP beamline lead to a complex and interrelated set of optical design considerations. We have optimized each design parameter such as flux, energy resolution, beam size, and highorder harmonics, etc., to reach an optimum set of beam parameters that will provide the end users an excellent platform to conduct state-of-the-art experiments. In this section, the optical consideration as well as the optical performance of the beamline will be addressed, whereas the details of the beamline components will be described in the next section. The beamline design targets are listed below: 1. Energy range: 4 15 kev 2. Photon flux: photons/sec 3. Energy resolution E/E: with a Si (111) crystal 4. Beam size at the focal point: 50 nm 50 nm at 10 kev (H V, FWHM) 5. High-order harmonics: Working distance (from the end of focusing optics to the focal point): 20 mm The beamline, the layout of which is shown in Figure 7.5, is designed based on a two-stage horizontal focusing strategy. Starting from the source, the first beamline optic is a water-cooled 700 mm-long horizontal focusing mirror (HFM) located at z = 24 m. With a magnification ratio 1:3, the HFM images the source downstream at z = 32 m where a secondary source is defined by a water-cooled beam-defining aperture (slit set 2). The HFM made of Si serves to suppress the high harmonic and takes up most of the heat load. The surface of the HFM is half-coated with Rh and the other half uncoated. With a 4 mrad incident angle, the HFM reflects X-rays up to 16.9 kev. The HFM deflects the beam horizontally to provide an offset of 360 mm at the sample position (z = 69 m), sufficient to keep the source out of the direct line of sight as a safety requirement. The optical considerations of the major optical components, the DCM and the nested Montel mirrors, will be described in the next sections, followed by simulation results, including the estimated flux at samples, the ultimate focal sizes, the energy resolution, and the higher harmonic suppression ratio.

13 7.4. BEAMLINE OPTICAL DESIGN 333 Fig 7-5 Figure 7.5: Optical layout of the X-ray Nanoprobe beamline at the TPS Double Crystal Monochromator The double crystal monochromator (DCM) with two Si(111) crystals is located at z = 66 m. The location of the DCM is a trade-off between the power density impinging on the DCM and the angular stability of the first crystal with respect to the second. Two strategies are taken to ensure the mechanical stability while scanning the DCM at a fixed exit condition: (1) To count on the cause of gravity, the Si crystals are placed in horizontally diffracting geometry in which the rotational axes of the crystals are perpendicular to the ground surface; (2) An 8 mm gap between the two Si crystals is adapted to further reduce one translational stage alone z-direction. The total heat-load and power density impinging on the DCM can be much reduced with helps of front-end aperture and beamline slits. By locating DCM at Z = 66 m, the maximum total heat-load and power density (assumed in normal incidence) are estimated to be 0.2 W and 0.62 W/mm 2, respectively. For comparison, the power density on DCM is 68 times higher when the DCM is located Fig 7-11 at z = 33 m than that at 66 m. Despite that a water-cooling strategy seems feasible under such power density, the liquid nitrogen-cooling strategy is preferable at this stage of design. The energy resolution of DCM is a resultant quantity determined by the Darwin width of Si (111) crystals, the angular size of the secondary source subtended at the monochromator and the angular acceptance practically set by the downstream nano-optics[2]. In the present design, the sourcesubtending angle is 0.2 µrad (FWHM). The horizontal angular acceptance of the DCM is calculated to be 16 µrad (FWHM), set by the location (z = 69 m), the length (150 mm) and the incident angle (4 mrad) of the nested Montel mirrors. The resultant energy resolution, despite not as fine as intrinsic

14 334 CHAPTER 7. X-RA NANO-PROBE Darwin width of Si(111) ( ), is calculated less than in 4 20 kev Nano-Optics The X-rays can be focused to tens of nanometers by means of optical components such like K-B mirrors, Fresnel zone plate, compound refractive lens, etc. The characteristics of these three focusing methods and their applicability to different experiments are compared in Table 7.2. This beamline requires energy scanning which mandates constant focal position at all energies, i.e. no chromatic aberration. The nested Montel mirrors[3] are chosen following the double crystal monochromator (DCM) to focus the monochromatic beam to the required size at the end-station. The nano-focusing optics, a pair of nested Montel Rh-coated mirrors, is located at z = 69 m. With mirror length of 150 mm and an edge-clear working distance of 20 mm, the resultant demagnification ratio upon combining the HFM and the nested Montel mirrors is calculated to be 1/726 and 1/389 in vertical and horizontal direction, respectively. By assuming a 7 µm opening of beam-defining aperture, one can directly estimate the geometrically demagnified focal spot at the sample position to be around 20 nm (FWHM) in both directions. The image burring caused by the diffraction limit of the nest mirrors will be discussed in next section Diffraction-Limited Optics The image capability of an optical system will reach a certain limit when the wave property of light starts to dominate and thus the system is diffraction-limited. This is described by the following 1-dimension equation[4], ε = λq/d, (7.7) where ε is the resolvable image size, λ the wavelength, q the image distance, and D the projection of focusing mirror at z-axis. As can be seen from the equation, at fixed wavelength and incident angle, decreasing the image distance and increasing the mirror length will reduce the identifiable image size. But larger mirror length is accompanied by larger slope error effect which will increase the image size. Therefore a compromise is reached among these factors to achieve an optimum image.

15 7.4. BEAMLINE OPTICAL DESIGN 335 Table 7.2: Characteristics and comparisons of three X-ray focusing systems chromatic dependence K-B mirrors Fresnel zone plate Compound refractive lens low-pass 1/λ 1/λ 2 flux delivery highest high low energy range critical angle limited < 10 kev > 10 kev focal size 20 nm nm 30 nm nm 50 nm nm theoretical limit < 20 nm < 10 nm 5 nm geometry reflective in-line in-line diffraction application fluorescent analysis imaging application spectroscopy ideal for polychromatic and monochromatic micro/nanodiffraction high flux using KB multilayer OK allowing large and rapid energy changes limited access to reciprocal space high resolution, low cost, easy operation highest resolution, commercialized for high energy applications Although reducing the image distance, and hence the working distance, will lessen the effect of diffraction limit, the working distance needs to accommodate a reasonably sized sample holder and in our case we choose this distance in the range of 20 to 40 mm. For the image distances and mirror sizes of nested Montel mirrors, first we decide these values of the mirrors by considering the projection of light on them, slope error, and diffraction limit.

16 336 CHAPTER 7. X-RA NANO-PROBE As will be described in next section that the mirror serving as a low-pass energy filter will only reflect photons with energy smaller that the critical energy. The relationship between impinging angle θ and cut-off energy E c is simply that the product of E c and θ equals constant, 67.4 for the case of Rh-coating mirror. Part of the nested Montel mirrors surfaces determined this way exhibits an incident angle larger than total reflection angle due to very high demagnification ratio. Figure 7.6 shows the dependence of the impinging angle and the cut-off energy on the mirror surface deviating from the nominal center of mirror. With the above design considerations, Figure 7.7 shows the diffraction limited spot size as a function of photon energy at three working distances. To minimize the effect of diffraction limit we choose a working distance of 20 mm, for which the length of mirrors will still maintain a reasonable slope error. The nominal mirror size and the image distances are 150 mm and 95 mm, respectively. As a reminder, the image distance q and the projection aperture D are not constant, accordingly Cut-off energy Impinging angle (mrad) Impinging angle Cut-off energy (kev) Deviation from mirror center (mm) Figure 7.6: Dependence of the impinging angle and the cut-off energy on the mirror surface deviating from the nominal center of mirror.

17 7.4. BEAMLINE OPTICAL DESIGN Diffraction-limit spot size (nm) mm 30 mm 40 mm Energy (kev) Figure 7.7: The diffraction limited spot size as a function of photon energy at three working distances. It is remarkable to find that the diffraction-limited spot size is not a monotonic function with photon energy. A minimum spot size occurs in the energy 8 15 kev, relying on adapting the nominal incident angle 4 mrad Simulations The performance of beamline such as photon flux, energy resolution, beam size, high-order harmonics, and heat load, etc., are calculated using SHADOW, XOP, and Spectra. Image Size The final focal size is a convolving result of the geometrical demagnification, the slope error of mirrors and the effect of diffraction limit. The diffraction-limited spot size ε is estimated by ε 0.88λq/D; where λ is the wavelength, q the image distance (95 mm), and D (0.6 mm) the projection aperture of nano-focusing mirrors. Figure 7.8 shows the simulated focal spot sizes at sample position by setting slits2 a 7 µm horizontal opening, with

18 338 CHAPTER 7. X-RA NANO-PROBE three various slope errors of nested mirrors, 0 µrad (square), 0.05 µrad (triangle) and 0.1 µrad (circle), along with the contributions from geometrical demagnification (solid line) and diffraction-limit (dash line). It is noticed that in the energy range 4 15 kev the focal spots of nm can be obtained by employing mirrors with slope errors less than 0.05 µrad. The degree of coherence inside the focal spot can be approximately estimated by rationing the fractional area of the diffraction-limited spot occupied. In the case of nested mirrors with slope error of 0.05 µrad, the coherent fraction is 69% and 13% at energy 4 kev and 15 kev, respectively, resulting a coherent flux photons/sec in the energy range 4 15 kev. 100 Focal spot (nm)( FWHM) diffraction-limit geometrical slope error = 0 rad slope error = 0.05 rad slope error = 0.10 rad Energy (kev) Figure 7.8: Simulated focal spot sizes at the sample for various slope errors of nested Montel mirrors 0.00 µrad (squre), 0.05 µrad (triangle) and 0.10 µrad (circle), along with the contributions from geometrical demagnification (solid line) and diffraction-limit (dash line).

19 7.4. BEAMLINE OPTICAL DESIGN 339 Photon Flux The calculated photon flux at the focal points are shown in Figure 7.9, which is obtained by the ray tracing program by including the HFM (700 mm long) with slit2 openings (7 µm), the DCM, focusing optics, and various parameter of the constituent optics. For comparison, three alternative nano-focusing optics are simulated and shown in Figure 7.9, namely, the nested Montel mirrors (Montel, circle), the conventional KB mirrors (KB, triangle) and the Fresnel zone plates (ZP, square). While a two-stage focusing is taken for the cases of Montel and ZP, the KB directly images the source at the sample. A horizontally diffracting DCM is employed for the three cases. The edge-clear working distance for the three cases is kept 20 mm, and the diffraction efficiency of ZP is assumed 10%. The ultimate focal spot sizes are assumed same as the case of the Montel with slope error 0.05 µrad. The mirror length is 150 mm for Montel Flux (Photons/sec) Photon Energy (kev) Figure 7.9: Comparison of the expected flux delivered at the sample with three alternative nano-focusing optics, the nested Montel mirrors (circle), the conventional K-B mirror (triangle) and the Fresnel zone-plate (square), with the flux spectra from IU22 (solid line).

20 340 CHAPTER 7. X-RA NANO-PROBE and 40 mm (H) 100 mm (V) for KB. The incident angle for the mirrors is kept 4 mrad. The diameters of zone plates vary from µm in energy range 4 15 kev. The higher flux throughput in the cases of Montel and KB than that of ZP is a direct result of optical efficiency of the nano-optics used. Advantaging by the geometrical arrangement, the acceptant aperture for Montel is 0.6 mm (H) 0.6 mm (V) over that of KB 0.16 mm (H) 0.4 mm (V). However, the up-to-date optical efficiency of Montel is much lower than 70%, and the slope error of Montel mirrors is yet not reaching 0.05 µrad as expected in the simulation[3]. Energy Resolution The calculated energy resolutions are shown in Fig The major two terms that determine the resolution are the intrinsic Darwin width of the crystal and the horizontal divergence of white light entering the DCM. This translates into energy broadening via the Braggs Law, and dominates the total energy resolution ( E/E) of the Nano-probe beamline, which lies in the range from to This value can be further improved by limiting vertical acceptance angle, however, with the attendant loss in photon flux. High-order Harmonics Table 7.3 shows the high order harmonics contents of this beamline. As will be detailed in Sec , Rh reflective surface is used to reduce harmonics. Only the ratio at the lower energy end of the range is calculated, as the value is the highest at that point. The ratio at 6 kev is , which will meet the demand of end-station users. Table 7.3: High-order harmonics content of the Nanoprobe Beamline. n Energy (kev) K Source Flux End-station Flux Ratio (photons/sec) (photons/sec) (Flux 18 kev /Flux 6 kev ) x x x x x 10 7

21 7.5. BEAMLINE DETAILED DESIGN 341 Energy Resolution ( E/E ) 2.0x x x x10-5 Total Si (111) Darwin width Source divergence (with 7 m slits at 32 m) Source size Photon Energy ( kev ) Figure 7.10: Calculated energy resolution delivered by the beamline. 7.5 Beamline Detailed Design Optical Layout Summing up the parameters described in last sections, the overall layout of the Nano-probe Beamline is shown in Figure The space allocation around this beamline is shown in Figure This beamline will use port 23 of the TPS ring, with a total length of 75 m. As shown in the Figure 7.11, this beamline consists of three major optical components, a HFM generates a virtual horizontal source at 32 m, a DCM at 66 m positioned in horizontal geometry for energy-selection and a nested Montel mirrors at 69 m that creates the final focal spot of nm at the sample Double Crystal Monochromator The double crystal monochromator is often used in synchrotron X-ray beamlines to facilitate energy scanning and adequate energy resolution while maintaining constant beam position when appropriate compensation mechanism is applied in conjunction with the parallel surface of the double crystal.

22 342 CHAPTER 7. X-RA NANO-PROBE Fig 7-5 Fig 7-11 Figure 7.11: Optical layout of the Nanoprobe Beamline.

23 7.5. BEAMLINE DETAILED DESIGN 343 DCM hv Slits 2 HFM hv Slits 2 Nested K-B (Motel) mirrors DCM horizontal type HFM Figure 7.12: Space allocation of the Nanoprobe Beamline. Si(111) is used in this beamline for the energy range of 4 15 kev at resolution better than , which will satisfy most of the absorption experiments. In the realm of nano-focusing, accuracy of the compensation mechanism of DCM, and temperature fluctuations and vibrations inside and outside the DCM, will all influence the quality of focusing. These factors and the energy resolution are discussed in the following sections. Energy Resolution The configuration of double crystals used in X-ray beamlines usually adopts the non-dispersive type with parallel diffraction surfaces to keep the outgoing beam in a fixed path. The diffraction from the crystal lattice is defined according to the Braggs Law, nλ = 2d sin θ B, (7.8) where λ is the wavelength, d the lattice spacing, θ B the Bragg angle, and n an integer(1, 2, 3,...).

24 344 CHAPTER 7. X-RA NANO-PROBE The energy resolution for the crystal is then given by[2] E E = λ λ = θ(cot θ B ) = {( E E )2 crystal + [( θ) source size + ( θ) slit ] 2 cot 2 θ B } 1/2, (7.9) where ( θ) source size = v/r 1 and ( θ) slit = s 1 /r 1, v is the size of the photon source, r 1 is the distance between the crystal and photon source, s 1 is the entrance slit opening, and r 1 is the distance between slit and photon source. The energy resolution for the crystal rocking width is given by ( E E ) crystal = ( θ) Darwin (cot θ B ) = 4r e πv d 2 b C F hr e M, (7.10) where θ B is the Bragg angle, r e = e 2 /mc 2 the classical electron radius, V the volume of the unit cell, d the Bragg plane spacing, b the asymmetry, C the polarization factor (which equals 1 or cos 2θ B ), F hr the crystal structural factor, and e M the temperature factor. Synchrotron X-rays are linearly polarized in the plane of electron orbit with the electric vector oriented in this plane. Since F hr varies little with energy, ( E ) E crystal can be viewed as a constant while operating the DCM in the vertical deflection geometry (C = 1). Commonly used crystals such as Si(111), Si(311), and Ge(111) have values of , , and While operating the DCM in the horizontal deflection geometry (C = cos 2θ B ) and using Si(111) crystals, the resultant energy resolution, as simulated in the last section, will still be better than The integral reflecting power for a particular state of polarization is given by[5], I = R(θ)dθ. (7.11) For weakly absorbing crystals, this can be well approximated by, I = 1 b 8 3 sin 2θ B r e λ 2 πv C F hr e M, (7.12) where R(θ) is the reflectivity. Bragg reflections dispersing in the orbital plane (horizontal type DCM) suffer degraded reflectivity and reduced angular width because of the unfavorable polarization of the incident wave. We

25 7.5. BEAMLINE DETAILED DESIGN 345 have C = cos 2θ B for such Bragg reflections, which renders the reflectivity and angular width very small when 2θ B approaches 90. The total flux collected by the horizontal type DCM is slightly lower than that by the vertical type. The energy resolution is comparable by using Slit 2 to limit the beam in the horizontal direction. The advantage of using the horizontal type is reduction of gravitational effect on the mechanical mechanisms, hence the much greater stability compared to the vertical type. For the above reasons, we choose the horizontal type DCM for the monochromator. Mechanical Design The mechanical stability of the DCM is a critical factor in the overall beam stability and we will use the most stringent specifications for the fabrication and assembly of the DCM. In addition, we stress the following three requirements: (1) maintaining the ambient temperature variation of DCM within 0.5 C. (2) positioning the DCM as close to the focusing optics as possible. (3) using a quadrant beam position monitor (QBPM) to provide feedback to the DCM to compensate fluctuation in beam position, and they are discussed below. The stability of the ambient temperature of DCM contributes to the deformation of the whole system and thus influences the outgoing beam. The distance between DCM and focusing optics should be kept small to reduce the amplification effect of any angular uncertainties from upstream. In the present case this distance is reduced to minimum of 3 m, which is required to accommodate a pair of high precision slits, screen, QBPM, and shutter. The QBPM tracks the beam and provides feedback to the DCM to compensate any beam displacements introduced by the scanning mechanism of DCM during energy scan. With a set of slits upstream of the focusing optics to further define the beam, the focused beam after focusing optics will have a high degree of spatial stability. The design parameters of DCM are listed in Table Focusing System The first phase experiment of this beamline requires energy scanning which mandates constant focal position at all energies, i.e. no chromatic aberration. Therefore we choose nested K-B (Montel) mirrors for nano-focusing

26 346 CHAPTER 7. X-RA NANO-PROBE Table 7.4: Design parameters of the DCM Crystal 1 Crystal 2 position (mm) 66,000 - crystal orientation horizontal type crystal offset D (mm) 8 energy range (kev) 4-15 Bragg angle (degree) shape plane plane crystal size (mm 2 ) 40 x x 60 substrate material Si (111) Si (111) at this beamline to satisfy user s experimental requirements. In this section we will discuss the detail parameters of the nested K-B (Montel) mirrors, including the reflective surface, incident angle, demagnification ratio, and the role of diffraction limits in the optical design. Geometrical optics is insufficient to describe the optical phenomenon when the scale of focused photon beam reaches the realm of nanometer. A more complete understanding of the relation between object and image will be obtained only when diffraction effects are included. The overall image equation should be[6]: where h (x i, y i ) = U i (x i, y i ) = h (x i, y i ) U g (x i, y i ), (7.13) P (λdx, λdy )e j2π(x ix +y i y ) dx dy. (7.14) Equations (7.13) and (7.14) represent the final results of image analysis. They indicate that when diffraction effects are included, the image is no longer a perfect replica of the object. Rather, the actual image obtained is a smoothed version of the object, a consequence of the nonzero width of the impulse response h. This smoothing operation can strongly attenuate the final detail of the object, with a corresponding loss of image fidelity resulting. In the sections below we will start with the discussion on energy range and ratio of high-order harmonics that defines the reflective surface and incidence

27 7.5. BEAMLINE DETAILED DESIGN 347 angle, then determining the working distance and mirror size with diffraction limits, and followed by the demagnification ratio that decides the parameters of nested K-B (Montel) mirrors and horizontal slits. Finally the mechanical design is described. Incidence Angle and Coating Materials The reflectivity of a mirror is a function of both photon energy and incident angle and is determined by the type of metal(s) used for reflecting surface and its micro-roughness. For X-ray photons, the critical angle at total reflection, Θ c, can be used for selecting the materials for reflecting surface and the incidence angle. The critical angle for metals is given by[7], Θ c ( n ar e λ 2 Z ) 1/2, (7.15) π where n a is the atomic density of metal, r e the classical electron radius, λ the wavelength, and Z the atomic number. From equation (7.15) at a given photon energy metals of higher atomic density has larger critical angle, and thus are often used for reflection of X- rays. In the case of Au or Pt, while they are often chosen for this purpose, their absorption structures between kev limit the application for high resolution spectroscopy in this region. Figure 7.13 (a), (b), (c), and (d) show the reflectivity of six types of surfaces at four incident angles, and Table 7.5 list the related parameters of six elements used for X-ray reflection. With these data it can be shown that at 4 mrad incident angle Rh has a critical energy of 16.9 kev that is sufficient to cover the 4 15 kev range. Overall it will satisfy the allowed percentage of high-order harmonics for this beamline. Demagnification Ratio The source size of IU22 has a FWHM of 282 µm 12.5 µm (H V). One stage vertical focusing can focus the beam to the required vertical size, but the same is not sufficient in the horizontal direction. We thus choose to add a HFM with a set of horizontal slits to limit the source size for the nested-hfm to reach the desired size while maintaining the simplicity of the mechanical design.

28 348 CHAPTER 7. X-RA NANO-PROBE mrad mrad 0.8 Si Pt 0.8 Si Pt 0.7 Au Rh 0.7 Au Rh Reflectivity Pt 25 nm + Rh 8 nm Pt 25 nm + Rh 5 nm Reflectivity Pt 25 nm + Rh 8 nm Pt 25 nm + Rh 5 nm Photon Energy ( kev ) (a) mrad 0.9 Si 0.8 Pt Photon Energy ( kev ) (b) mrad 0.9 Si 0.8 Pt 0.7 Au Rh 0.7 Au Rh Reflectivity Pt 25 nm + Rh 8 nm Pt 25 nm + Rh 5 nm Reflectivity Pt 25 nm + Rh 8 nm Pt 25 nm + Rh 5 nm Photon Energy ( kev ) (c) Photon Energy ( kev ) (d) Figure 7.13: Reflectivity of six types of surfaces at various incident angles: (a) 3.9 mrad, (b) 4.0 mrad, (c) 4.1 mrad, and (d) 4.2 mrad. Table 7.5: Elemental properties and Θ c E for six commonly used materials for X-ray mirrors. Element Atomic Atomic Density Θ c E Number Weight ρ (g/cm 3 ) (mrad x kev) Bare Si Ni Rh Ag Pt Au

29 7.5. BEAMLINE DETAILED DESIGN 349 In following sections we will discuss the detailed optical parameters of the nested-vfm and nested-hfm. A. Nested-VFM The total length of the beamline is 75 m, including the 3 m downstream of the focal point reserved for end-station use. The nested-vfm is positioned at 69,000 mm from the source, and 95 mm from the focal point, with 150 mm mirror length. The image size without considering diffraction limit is the square root of the sum of two terms, one from the geometrical demagnification and the other from the slope error. Our calculation shows that a typical slope error of 0.1 µrad is not low enough for the nano-focusing, while 0.05 µrad is the least required value. The trial data (at 8 kev) are listed below. Beam size: a. demagnification of beam 12.5 µm 95 / 69,000 = nm b. beam size contribution from slope error µrad m = nm (Slope error 0.05 µrad) µrad m = nm (Slope error 0.1 µrad) c. overall beam size nm (Slope Error 0 µrad) ( ) = nm (Slope Error 0.05 µrad) ( ) = nm (Slope Error 0.1 µrad) B. Nested-HFM As discussed above, the horizontal focusing is assisted by a HFM with a set of slits to limit the source size, which introduces an additional variable compared with the vertical focusing. The nested-hfm is 69,000 from the source of IU22 and 95 mm from the focal point, with 150 mm mirror length. The source size of nested-hfm is defined by the slit set and these are

30 350 CHAPTER 7. X-RA NANO-PROBE included in the equation for calculating the image size: (s 2 q H p H ) 2 + ( q H ) 2 = s 2 H, (7.16) where s 2 is the opening of the horizontal slit set 2, q H the image distance, p H the distance from the slit to the nested-hfm, the slope error of the nested-hfm, and s H the image size. The design parameters of the nested Montel mirrors are listed in Table 7.6. Mechanical Design of the nested Montel mirrors Both the nested K-B (Montel) mirrors and the sample holder and its adjustment mechanism are attached to a single granite piece weighing a few tons to gain sufficient mechanical stability and to reduce the relative motion between the mirror set and sample. The supporting structure and adjustment mechanism will use materials with low expansion coefficient such as invar, and the variation of local ambient temperature will be kept to within Table 7.6: Design parameters of mirrors. HFM nested-vfm nested-hfm position (mm) 24,000 69,000 69,000 object distance, p(mm) image distance, q(mm) grazing angle (mrad) 24,000 69,000 37,000 8, shape Cylindrical Plane-ellipse Plane-ellipse clearance aperture (mm 2 ) 700 x x x 20 substrate material Silicon single crystal Silicon single crystal Silicon single crystal coating material Rh (600 Å thickness) Rh (600 Å thickness) Rh (600 Å thickness) slope error (µrad, RMS) roughness (Å, RMS) < 1 < 0.05 < 0.05 < 2 < 0.5 < 0.5

31 7.5. BEAMLINE DETAILED DESIGN C to reduce external effects. Because the nested K-B (Montel) mirrors and sample holder sit on the same granite, the effect of thermal expansion on the relative displacement of these two comes from the mirrors and holders and their respective supporting structures. Assuming the mirrors (Si) are 15 mm thick with a stand (invar) high of 60 mm, a temperature change of 0.03 C will induce a height change of 5 nm, about one tenth of the image size. This illustrates the importance of a uniform and constant temperature environment. In addition, a suitably designed sample load-lock system will minimize external effects on this environment Heat Load Analysis The high flux density and the accompanying high heat load from undulator IU22 necessitate a comprehensive heat load analysis of the beamline optics and effective mechanical and cooling design to ensure stable performance. The TPS ring will run at 500 ma and we add 20% safety margin to run IU22 at 5 mm gap (equivalent to k = 2.15) for the design of this beamline. IU22 outputs 9.67 kw in the worst case scenario. For the high flux density mode, most of the power is absorbed by a water-cooled aperture at 18 m in the front end, such that the water-cooled Slits 1 outside the shielding wall receives much less heat load. The input/output and absorbed power, and power density on each optical element calculated with the 3.08 m IU22 operating at 600 ma is shown in Table 7.7. The diamond filer/window is chosen for the materials excellent thermal conductivity. With the water-cooled aperture in the front end taking much of the heat load, a water cooling scheme is sufficient for the DCM. This removes much of the maintenance issues associated with a liquid nitrogen cooling system. Heat loads on the focusing optics are negligible Mechanical Consideration The mechanical design of optical system is crucial to the success of nanofocusing. In addition to the list of requirements, other factors such as tolerances on the fabrication of optics, stability of ambient temperature, and effect of vibrations, etc. are all critical to the quality of nano-focusing, and can severely degrade the beam quality if not properly addressed. Therefore we focus our design on two major criteria, temperature stability and simple mechanical design, in order to attain high stability and reproducibility. The

32 352 CHAPTER 7. X-RA NANO-PROBE Table 7.7: Heat load data on the optical components. Input Power (W) Output Power (W) Front End Absorbed Power (W) Diamond Window Slits DCM 1.86 << Conditions: 3.0 m ma, max. k of 1.85, with a 5 mm gap design specifics are detailed below. The ambient temperature of critical optics is kept within 0.03 C to 0.5 C to maintain beam stability. The DCM is positioned as close to the focusing optics as possible to reduce angular uncertainties. A quadrant beam position monitor (QBPM) is positioned upstream of focusing optics to track the beam and provides feedback to the DCM to compensate fluctuation in beam position. A pair of horizontal and vertical slit system is positioned upstream of focusing optics to provide a fixed source point Vacuum Consideration In order to prevent phase degrading caused by optical components, such like beamline windows, filter etc, the whole beamline, including the beamline components and end station chamber, will be operated under vacuum condition. There are several locations deserve specific care. The HFM chamber at 32 m housing the horizontal focusing mirror is the first major beamline optical component. The HFM reflecting half of the incident photons will be kept at low 10 9 Torr by employing a high capacity ionic pump ( > 500 l/s). The DCM at 66 m, accompanying with multi-axis translation stages, is usually operating under 10 7 Torr. Although a turbo-molecular pump (500 l/s)

33 7.5. BEAMLINE DETAILED DESIGN 353 will be able to keep the DCM under 10 7 Torr, the ionic pump is preferred as the ultimate pumping agent for its quiet mechanical operation. The end station chamber housing with SEM will be kept under at least 10 6 Torr. An ultra quiet turbo-molecular pump is highly demanded for the station chamber. The head and vibration caused by the turbo-molecular pump will be subject to further engineering consideration. The vacuum strategy for the beamline is depicted in Figure Other Optical Elements Optical elements in addition to DCM and focusing optics include slit sets, screens, BPMs, etc. As mentioned in the above sections, there is a high precision slit set before DCM, and a set of QBPM for feedback purpose. There is also a slit set to in the white light path to define the beam, and screens in combination with two sets of BPMs to monitor the beam path. The high precision slits set downstream of the DCM are made in house by NSRRC. These are followed by screens in the light path for monitoring purposes, and the compensating QBPM.

34 354 CHAPTER 7. X-RA NANO-PROBE 10-5 Nested K-B (Montel) Mi HFM Screen 1 Long Pipes Center DCM Vacuum Pressure( torr) Fig ( 下半圖 ) Figure 7.14: The beamline vacuum strategy.

35 7.6. END STATIONS End Stations Conceptual Design The main purpose for the nanoprobe end station is to solve practical issues of domestic semiconductor-based high technology companies. This end-station will integrate multi-diagnostic probes (Figure 7.15) for resolving the atomic, chemical and electronic structures of semiconductor-base devices with sub-ten nm spatial resolution in tomographic and nondestructive manners. However, limited by the total length of the beamline (< 75 m), the beamline optics is not able to deliver useful focal spot small than 30 nm, which may not be fine enough for resolving practical issues from the next generation IC devices (channel length < 30 nm). The development of new X-ray techniques becomes unavoidable and in fact challenging. At the present stage of station design, we plan to include conventional probes mature in present techniques, like X-ray fluorescence (XRF), X-ray absorption fine structures (XAFS), X-ray excited optical luminescence (XEOL), and emerging ones that will need intensive development, such like Bragg-ptychography. The ptychography (or scanning coherent diffraction image) has recently been proven experimentally for reconstructing images of extended samples. While operating in Bragg diffraction, the Bragg-ptychography (demonstrated at CNM Nanoprobe at APS) is expected to provide the strain information of a crystalline specimen with several nm resolution. The geometry of Bragg- Surface Multiple Diffraction (BSMD) (Figure 7.16) enables interaction of three diffractions along the interface[8]. While incorporated with BP, it is expected and subject to intensive development to provide a novel interfacial scanning mechanism and phasing algorithm for 3D strain mappings. The nano-focusing optics and sample stage will be installed in a vacuum chamber stationed on a vibration dampened supporting frame. The BP detector frame, less influence by ground vibration will be separated from the sample frame. The relative movement between sample and focusing optics will be accomplished by ultra precise linear translational stages and controlled by laser interferometer with precision better than 2 nm. At least two major detectors are planned: A multi-element energy-resolved detector for XRF; and a photon counting detector for coherent scattering. An off-lined field emission SEM connected to the sample chamber via a load-lock system will provide complementary analysis other than X-rays.

36 356 CHAPTER 7. X-RA NANO-PROBE Laser Interferometer hv SEM Fluorescence Detector CCD Figure 7.15: The conceptual design drawing of the Nanoprobe end station.

37 7.6. END STATIONS 357 Figure 7.16: The Bragg Surface Multiple Diffraction (BSMD) distinguishes the strained and unstrained interfaces via diffraction pattern Strategy for End Station Development We describe the system for two stages. The first part of the end station is the basic function for which we are able to conduct the elemental mapping (XRF) and X-ray absorption spectroscopy with high spatial resolution. The second part is the capability doing the Bragg-ptychography. Basic Function The basic function is the scanning probe X-ray microscope with detectors, which includes: Support frame, optical table and vibration dampening Vacuum system (including the Beryllium window and pump) Beam define aperture with 2-axis motorized stages Optics module with 6-axis motorized precision stage with 6-axis piezoelectric actuators. (X,,Z, Pitch, Roll, aw)

38 358 CHAPTER 7. X-RA NANO-PROBE Sample module with 8-axis motorized precision stage.(coarse X, Coarse, Coarse Z, Fine X, Fine, Fine Z, Rz, Rx) (3 laser interferometer at 5 nm ) Fly-scan high speed scanning capability. (scaler and analog addressing signal) In-vacuum detector with 3-axis motorized stage EPICS interface Optical align camera The focus optics for X-ray energy from 5 to 15 kev Additional Functions The additional functions are SEM and the Bragg-Ptychography. part includes: This Diffraction/goniometer module with 3 axis (Theta, Phi, Chi) SEM column integration Diffraction detector (CCD and PILATUS detector) Computer Clusters and the reconstruction algorithms for coherent Bragg X-ray diffraction Image Control software and Interfaces for the end station The vibration less than 10 nm by active vibration control In-situ package Progress of the development A test system is designed and manufactured to study the feasibility of the nano-probe end station. This system includes the sample stages, interferometer, a mirror stage, an SEM column, a secondary electron detector, and a multi-element x-ray fluorescence detector. An online SEM will be installed in the main chamber for locating the sample and further precise positioning[9]. The system is shown in Figure 7.17.

39 7.6. END STATIONS 359 Figure 7.17: Top: A preliminary design drawing of the test system of the nano-probe end station. Bottom: A photograph of the test system.

40 360 CHAPTER 7. X-RA NANO-PROBE Development of the fly scan system A control board is designed to coordinate different functionalities which are needed to complete a spectrum scan process during the fly scan. The fly scan is to monitor the position by laser interferometer with a high update rate, generally higher than 100k samples/second. Within the time slot, the movement is less than the desired spatial resolution. For instance, the highest scan speed at 15 Hz for a 30 µm field of view is around 1 mm/s. If we require 1000 points per line, the movement is 30 nm per point. The minimum spatial sampling point is 1 mm/s times 10 µs (sampling time for 100 khz), which is 10 nm less than 30 nm. In this case, the high sampling rate from the interferometer ensures the accuracy of the position change during the scan without stopping the movement. In the following we will describe the operation of the board as shown in Figure When the board accepts a start scan command from a host controller, it initiates the scan process which will output a control signal to move the stage via a DAC port depending on the address feedback information from the interferometer. At the same time, the board will get the spectrum data generated by the SDD detector via the FPGA digital input port and store them to the internal memory. The collected spectrum data are then transferred to a host PC through the Ethernet for analysis and display. The control board will repeat the process until the scan is completed or interrupted by the host controller. The control board can only be configured to operate by the host PCs user interface. The architecture of the control board is shown in Figure The fly-scan circuit is tested by scanning a 650 nm laser focused on a #100 copper grid (not well focused). The result is shown in Figure Despite the weak laser signal and the poor focus, this result shows the fly-scan circuit is functioning properly. We will test the signal of a multi-element fluorescence detector using electron beam in the near future. Design of the Montel K-B focusing optics The Montel K-B mirror is chosen as the focusing optic, which has the advantages of a small size, a short working distance, and ease of manipulation. A Montel K-B mirror capable of focusing to tens of nanometer is hard to make, due to a high requirement of the slope error of the optics and the crack problem in the center parts. These problems will be addressed in

41 7.6. END STATIONS 361 Figure 7.18: The system architecture of the fly-scan system. Figure 7.19: The architecture of the control board. our system. This mirror will be fabricated by JTEC, which has the EEM polishing technique and is able to produce a slope error smaller than 0.05 µrad. A simulation of the Montel K-B mirror is done to determine the parameters of the K-B mirror. There are several constraints when deciding on the optical parameters of the Montel optics. They are:

42 362 CHAPTER 7. X-RA NANO-PROBE Figure 7.20: A preliminary test image of the fly-scan system by an optical laser system. 1. The objective distance: this is mainly determined by the length of the beamline, and the required demagnification factor. 2. The working distance: this is mainly determined by the requirements of the experimental condition, the sample size, and the required demagnification factor. 3. The reflection angle: this is determined by the energy and the coating material. 4. The mirror length: this is mainly due to the limitation of the manufacturing process and the required demagnification factor. After taking these constraints into consideration, the suitable parameters for the Montel optics can be determined. The objective distance is 69 meter, and other optical parameters of the Montel K-B mirror pair is listed in Table 7.8. Simulation of the Montel K-B focusing optics The focal spot size and divergence of the beam are simulated using the parameters above and the figure error provided the JTEC company. The layout of simulations of the Montel mirror is succinctly illustrated in Figure 7.22 The simulation results are shown in Figure The left part of the figure is the simulated spot size of X-rays at the sample position. It shows a focual spot around 40 nm in diameter, and a divergence of ca 7 mrad in the

43 7.6. END STATIONS 363 Table 7.8: Parameters of the Montel KB mirrors. Half Major Axis (m) Half Minor Axis (mm) Mirror Profile (Tangential/ Sagittal) (Ellipse/Plane) Mirror Length (mm) 110/100, useful Mirror Width (mm) 20/5, useful Slope Error (µrad) < 0.05 Coating Material Rh Reflection angle (mrad) B % (?) % (?X) Figure 7.21: The mirror profile of montel K-B mirror. horizontal direction. The average reflection is around 80% at 10 kev after two reflections. The simulation parameters of the X-ray source are shown in Table 7.9. The percentages of different modes of the Montel K-B mirror are shown in Table 7.10.

44 % (?X) Fig 6.7 the mirror profile of montel KB mirror The simulation of the Montel KB focusing optics The simulation result of the focus spot size and divergence of the beam is simulated 364 by using the parameter above and figure CHAPTER error from JTEC company. 7. X-RA The simulation NANO-PROBE of the Montel mirror is briefly described as Fig Focus beam 2. Only one mirror reflection HM only Focus beam 39 No cross VM only 3. No cross Figure 7.22: Illustration of the simulation layout of the K-B mirror pair. Figure 7.23: Simulated X-ray spot size and divergence at 10 KeV Specifications of stages Montel Mirror stage The Montel K-B mirror will be placed on a specially designed holder, the design of which is not finalized yet. Figure 7.24 is a conceptual drawing of the mirror mounting method and its associated alignment axes. The requirements for the manipulating mechanisms of these axes are listed in Table 7.11.

45 7.6. END STATIONS 365 Table 7.9: The simulation parameters of the X-ray source. Source Position 69 meters from the Montel mirror size X (µm) 13 size (µm) 13 Divergence X (µrad) 26 Divergence (µrad) 6 Table 7.10: The simulation results of the Montel mirror, showing the percentages of different modes of the beam. Percentage No Cross 40.83% HM only 0.38% VM only 0.42% Focus, HM first, and VM later 29.12% Focus, VM first, and HM later 29.23% At Gap* 0% *This simulation does not include the gap effect. Table 7.11: Specifications of the manipulating mechanisms of the Montel mirror. Montel Mirror Stage Axis Parameter Remark Specification X Range ± 5 mm Resolution 0.1 µm Repeatability ± 1 µm Stability* 5 hours (rms) 10 nm Continued on next page

46 366 CHAPTER 7. X-RA NANO-PROBE Table 7.11 Continued from previous page Axis Parameter Remark Specification Limits/Home Mark Encoder Range ± 12 mm Resolution 0.1 µm Z Pitch Roll Repeatability ± 1 µm Stability* 5 hours (rms) Limits/Home Mark Encoder Range ± 5 mm Resolution 0.1 µm Repeatability ± 1 µm Stability* 5 hours (rms) 10 nm Limits/Home Mark Encoder Range ± 20 mrad Resolution 0.01 µrad Repeatability 0.1 µrad Stability* 5 hours (rms) 0.1 µrad Limits/Home Mark Encoder Range ± 20 mrad Resolution 0.1 µrad Repeatability 1 µrad Stability* 5 hours (rms) 0.1 µrad Limits/Home Mark Encoder Continued on next page

47 7.6. END STATIONS 367 Table 7.11 Continued from previous page Axis Parameter Remark Specification Range ± 20 mrad Resolution 0.01 µrad aw Repeatability 0.1 µrad Stability* 5 hours (rms) 0.1 µrad Limits/Home Mark Encoder Sample stage Table 7.12 lists the specifications of the sample stage manipulating mechanisms. Table 7.12: Specifications of the sample stage manipulating mechanisms. Sample Stage Axis Parameter Remark Specification X Range Resolution Repeatability Stability* Limits/Home Mark Encoder Range Resolution Repeatability Stability* Limits/Home Mark Encoder ± 15 mm 1 nm ± 5 nm 5 nm ± 15 mm 1 nm ± 5 nm 5 nm Continued on next page

48 368 CHAPTER 7. X-RA NANO-PROBE Table 7.12 Continued from previous page Axis Parameter Remark Specification Range ± 10 mm Resolution 1 nm Z Pitch aw Flexure X Flexure Repeatability ± 5 nm Stability* 5 nm Limits/Home Mark Encoder Range No Limit Resolution 0.1 mrad Repeatability Stability* Limits/Home Mark Encoder Range No Limit Resolution 0.05 µrad Repeatability Stability* Limits/Home Mark Encoder Range +± 40 µm Resolution Open loop 0.2 nm Repeatability < ± 5 nm Stability* with feedback(rms) 5 nm Limits/Home Mark Encoder Range ± 40 µm Resolution Open loop 0.2 nm Repeatability < ± 5 nm Stability* with feedback(rms) 5 nm Continued on next page

49 7.6. END STATIONS 369 Table 7.12 Continued from previous page Axis Parameter Remark Specification Flexure Z Limits/Home Mark Encoder Range +± 40 µm Resolution Open loop 0.2 nm Repeatability < ± 5 nm Stability* with feedback(rms) 5 nm Limits/Home Mark Encoder The relative position between the mirror stage and the sample stage will be monitored by a laser interferometer and will be actively controlled by a PID feedback circuit with a flexure stage. The low frequency vibration will be suppressed by a factor of 2, according to our recent experimental result. A detailed experiment regarding the vibration and drift with a closely monitored and controlled temperature will be carried out. Fluorescence detector stage Table 7.13 lists the specifications of the manipulating mechanisms of the fluorescence detector stage. Table 7.13: Specifications of the manipulating mechanisms of the fluorescence detector stage. Fluorescence Detector Stage Axis Parameter Remark Specification X Range ± 200 mm Resolution 2 µm Repeatability ± 5 µm Stability* 1 µm Limits/Home Mark Continued on next page

50 370 CHAPTER 7. X-RA NANO-PROBE Table 7.13 Continued from previous page Axis Parameter Remark Specification Encoder Diffraction detector stage The diffraction detector stage will be placed around 1 to 2 meters always from the sample stage. The detector stage is currently in the conceptual design phase. The normal of the detector plane will always be aligned to the sample. A strong robotic arm with six degrees of freedom will be designed to hold the detector The specification of detectors Fluorescence detector Brand name/type: Hitachi / Vortex-Me4 Description: the fluorescence detector is a multi-element, high countrate SDD (silicon drift detector). Beryllium window thickness: 12.5 µm. Element number: 4. Detector size for each channel: 42.5 mm 2. (Total: 170 mm 2 ) Energy resolution: Fe (Mn K α ) 130 ev at a peaking time of µs. Maximum count rate for each channel: 250k cps, at a peaking time of 0.25 µs. Weight: Projection detector* *Not purchased yet. Brand name/type: Princeton Instruments / PIXIS-XF-2048B Description: A CCD with a scintillator, indirect detection. Pixel Size: 13.5 µm 13.5 µm

51 6.4 The specification of stages Montel Mirror stage The Montel KB mirror will be placed on a special designed holder, of which design is not finalized 7.6. yet. ENDA conceptual STATIONSdrawing as Fig 6.10 shows the method of mounting the 371 mirror and moving axis. The requirement of each axis is listed in table 6.2. Fig 6.10 The illustration of the Montel optics, conceptual drawing. Figure 7.24: Illustration of the Montel optics (a conceptual drawing). Montel Mirror Stage Remarks Range +/- 5 mm Resolution 0.1 µm X Repeatability +/- 1 µm Stability* 5 hours (RMS) 10 nm Limits/Home Mark Encoder Range +/- 12 mm Resolution 0.1 µm Repeatability +/- 1 µm Stability* 5 hours (RMS) Limits/Home Mark Encoder Range +/- 5 mm Resolution 0.1 µm Z Repeatability +/- 1 µm Stability* 5 hours (RMS) 10 nm Limits/Home Mark Encoder Pitch Range Figure 7.25: The preliminary design of the sample +/- 20 stage. mrad 41

52 372 CHAPTER 7. X-RA NANO-PROBE Pixel number: Detector Area: 27.6 mm 27.6 mm Dark -30 C: e-/(p sec) Full Well: 125k e- Read out noise: e- (rms) Frame scale: 0.5 Hz Weight : 2.75 Kg XEOL detector* *Not determined yet. Diffraction detector* *Not purchased yet. Brand name/type: Dectris / PILATUS 1M Description: Active pixel area detector Pixel Size: 172 µm 172 µm Pixel number: Detector Area: 169 mm 179 mm Dynamic range: 20 bits (1: ) Frame Full scale: 30 Hz Count Rate: > photons/s Weight: 25 Kg SE detector The SE detector is an Everhart-Thornley type detector which is integrated with the SEM column.

53 7.7. RADIATION SAFET 373 Photon detector* *Not determined yet Specifications of the load lock chamber (optional, to be defined) 7.7 Radiation Safety The beamline radiation safety follows the protocol set by the Radiation Safety Division at the NSRRC. Two main aspects are of concerns here, the radiation shielding and the bremsstrahlung shielding. While the radiation shielding is subject to detailed engineering design including the design of radiation hutches, the bremsstrahlung shielding can be defined versus a preliminary sketch. Fig shows the bremsstrahlung shieldings of the beamline. A 300 mm thick Pb bremsstrahlung collimator is placed at 22.6 m to limit and define the viewing angle into the undulator source. Because the HFM at 32 m deflects the incident beam by 8 mrad, the W-made bremsstrahlung stopper located at 34 m will effectively block any direct viewing angle from downstream and assure the safety requirement.

54 374 CHAPTER 7. X-RA NANO-PROBE Top view Front view Shield wall Bremsstrahlung collimator (Pb) Bremsstrahlung collimator (Pb) Bremsstrahlung stopper (W) Figure 7.26: Bremsstrahlung shieldings for the beamline.

55 7.8. SCHEDULE Schedule The X-ray Nanoprobe facility at TPS is scheduled to take the first synchrotron light by the end of the 3rd quarter of The construction schedules for the beamline and the end stations are shown in Figures 7.27 and 7.28, respectively. Fig 7.21 Construction timeline for the beamline Figure 7.27: Construction timeline for the beamline.

56 376 CHAPTER 7. X-RA NANO-PROBE Fig 7.22 Construction timeline for the end stations Figure 7.28: Construction timeline for the end stations. 7.9 Commissioning Plan There will be two stages in commissioning for the XNP in The beamline construction will be completed by the second quarter of By then the beamline will be well aligned and all the components are in place. After completing the vacuum test for each beamline components, the beamline will be allowed for interlock test followed by radiation shielding survey. In parallel, a preliminary end station will be set up for commissioning of the photon flux, energy resolution, mechanical stability, etc. The end station will be settled by the third quarter of The XAF will be the first station ready for commissioning, while the commissioning of CDI and BP will take place in For endstation: Off-line test: All control function, actuators and electronics by input signal test Vacuum test