Effects of thermal expansion of crystal-lattice on x-ray imaging crystal spectrometers

Similar documents
In-situ wavelength calibration and temperature control for the C-Mod high-resolution X-ray crystal imaging spectrometer

The Instrumental Function of the X-ray Imaging Crystal Spectrometer on Alcator C-Mod

Princeton Plasma Physics Laboratory

Formation and stability of impurity snakes in tokamak plasmas

Spontaneous Core Toroidal Rotation in Alcator C- Mod L-Mode, H-Mode and ITB Plasmas.

Observations of Counter-Current Toroidal Rotation in Alcator C-Mod LHCD Plasmas

ICRF Mode Conversion Flow Drive on Alcator C-Mod and Projections to Other Tokamaks

This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under contract

Reduction of Turbulence and Transport in the Alcator C-Mod Tokamak by Dilution of Deuterium Ions with Nitrogen and Neon Injection

Validation of spectral MSE for Alcator C-Mod and for ITER

Plasma impurity composition in Alcator C-Mod tokamak.

Recent Results in Impurity Transport in Tokamaks. D R Mikkelsen PPPL. 2nd Coordinated Working Group Meeting Greifswald, June 4-6, 2007

PFC/JA MEASUREMENT OF THE D-D FUSION NEUTRON ENERGY SPECTRUM AND VARIATION OF THE PEAK WIDTH WITH PLASMA ION TEMPERATURE

Temporal Evolution of Temperature and Argon Impurity Density Profiles Observed by X-ray Imaging Spectrometer Measurements at Wendelstein 7-X

EXD/P3-13. Dependences of the divertor and midplane heat flux widths in NSTX

PFC/JA Precision Measurements of the Wavelengths of Emission Lines of Mg-like and Na-like Kv in Alcator C Plasmas

Observation of Co- and Counter Rotation Produced by Lower Hybrid Waves in Alcator C-Mod*

Enhanced Energy Confinement Discharges with L-mode-like Edge Particle Transport*

X-ray observations of medium Z H- and He-like ions with satellites from C-Mod tokamak plasmas

ICRF Induced Argon Pumpout in H-D Plasmas in Alcator C-Mod

Introduction to the Diagnosis of Magnetically Confined Thermonuclear Plasma

Cauchois Johansson x-ray spectrograph for kev energy range

Bayesian inference of impurity transport coefficient profiles

Spectroscopic studies of impurities in the LHD plasmas

Survey of EUV Impurity Line Spectra and EUV Bremsstrahlung Continuum in LHD )

Thin Faraday foil collectors as a lost ion diagnostic

Measurements of rotational transform due to noninductive toroidal current using motional Stark effect spectroscopy in the Large Helical Device

Physics of fusion power. Lecture 14: Anomalous transport / ITER

The Neutron Diagnostic Experiment for Alcator C-Mod

Stigmatic X-ray imaging using a single spherical Laue crystal

Dipartimento di Fisica, Università degli Studi di Milano-Bicocca, Milano 2. Istituto di Fisica del Plasma, CNR, Milano

Blob sizes and velocities in the Alcator C-Mod scrapeoff

Impurity transport analysis & preparation of W injection experiments on KSTAR

The Magnetic Recoil Spectrometer (MRSt) for time-resolved measurements of the neutron spectrum at the National Ignition Facility (NIF)

Noninductive Formation of Spherical Tokamak at 7 Times the Plasma Cutoff Density by Electron Bernstein Wave Heating and Current Drive on LATE

Lower Hybrid Current Drive Experiments on Alcator C-Mod: Comparison with Theory and Simulation

ICRF Mode Conversion Flow Drive on the Alcator C Mod Tokamak

L-to-H power threshold comparisons between NBI and RF heated plasmas in NSTX

X-ray Spectroscopy on Fusion Plasmas

ITB Transport Studies in Alcator C-Mod. Catherine Fiore MIT Plasma Science and Fusion Center Transport Task Force March 26th Boulder, Co

Plasma Spectroscopy in ISTTOK

0 Magnetically Confined Plasma

GA A22722 CENTRAL THOMSON SCATTERING UPGRADE ON DIII D

Scaling of divertor heat flux profile widths in DIII-D

Application of atomic data to quantitative analysis of tungsten spectra on EAST tokamak

High Beta Discharges with Hydrogen Storage Electrode Biasing in the Tohoku University Heliac

1 EX/P7-35. Spectroscopic Studies on GLAST-III Varying the Inductance and Charging Voltage of Vertical Field Coils

1 EX/P4-8. Hydrogen Concentration of Co-deposited Carbon Films Produced in the Vicinity of Local Island Divertor in Large Helical Device

Improved RF Actuator Schemes for the Lower Hybrid and the Ion Cyclotron Range of Frequencies in Reactor-Relevant Plasmas

Ohmic and RF Heated ITBs in Alcator C-Mod

Abstract. PEGASUS Toroidal Experiment University of Wisconsin-Madison

Power Balance and Scaling of the Radiated Power in the Divertor and Main Plasma of Alcator C-Mod

Full-wave Simulations of Lower Hybrid Wave Propagation in the EAST Tokamak

Comparison of Divertor Heat Flux Splitting by 3D Fields with Field Line Tracing Simulation in KSTAR

Non-local Heat Transport, Core Rotation Reversals and Energy Confinement Saturation in Alcator C-Mod Ohmic L-mode Plasmas

Ion Temperature Measurements in the

Hard Xray Diagnostic for Lower Hybrid Current Drive on Alcator C- Mod

Scaling of divertor heat flux profile widths in DIII-D

X-Ray Measurements of the Levitated Dipole Experiment

A Study of Directly Launched Ion Bernstein Waves in a Tokamak

Analyses of Visible Images of the Plasma Periphery Observed with Tangentially Viewing CCD Cameras in the Large Helical Device

Flow measurements in the Scrape-Off Layer of Alcator C-Mod using Impurity Plumes

Characterization of neo-classical tearing modes in high-performance I- mode plasmas with ICRF mode conversion flow drive on Alcator C-Mod

Implementation of a long leg X-point target divertor in the ARC fusion pilot plant

Study of Current drive efficiency and its correlation with photon temperature in the HT-7 tokomak

Gyrokine.c Analysis of the Linear Ohmic Confinement Regime in Alcator C- Mod *

Impurity accumulation in the main plasma and radiation processes in the divetor plasma of JT-60U

A Motional Stark Effect Instrument to Measure q(r) on C-Mod

Observation of Neo-Classical Ion Pinch in the Electric Tokamak*

Effect of secondary beam neutrals on MSE: theory

Propagation of Radio Frequency Waves Through Fluctuations in Plasmas

TARGET PLATE CONDITIONS DURING STOCHASTIC BOUNDARY OPERATION ON DIII D

Design of structural components and radial-build for FFHR-d1

University of Washington

Measurements of Plasma Turbulence in Tokamaks

TH/P6-08. CompX, Del Mar, CA 92014, USA; 2Xcel Engineering, Oak Ridge, TN 37830, USA; 3PSFC-MIT,Cambridge, MA 02139

Investigations on warm dense plasma with PHELIX facility

Driving Mechanism of SOL Plasma Flow and Effects on the Divertor Performance in JT-60U

Wolter Imaging On Z. Chris Bourdon, Manager Z Imaging and Spectroscopy Julia Vogel, LLNL; Ming Wu, SNL ICF Diagnostics Workshop, October 5 th 2015

Observations of the collapse of asymmetrically driven convergent shocks. 26 June 2009

Introduction to Nuclear Fusion. Prof. Dr. Yong-Su Na

The Path to Fusion Energy creating a star on earth. S. Prager Princeton Plasma Physics Laboratory

CCD readout of GEM-based neutron detectors

Comparison of Ion Internal Transport Barrier Formation between Hydrogen and Helium Dominated Plasmas )

Bolometry. H. Kroegler Assciazione Euratom-ENEA sulla Fusione, Frascati (Italy)

Investigation of Intrinsic Rotation Dependencies in Alcator C-Mod

NEW CORRECTION PROCEDURE FOR X-RAY SPECTROSCOPIC FLUORESCENCE DATA: SIMULATIONS AND EXPERIMENT

Damage to Molecular Solids Irradiated by X-ray Laser Beam

Line ratios and wavelengths of helium-like argon n = 2 satellite transitions and resonance lines

Diagnostics for Burning Plasma Physics Studies: A Status Report.

Flow dynamics and plasma heating of spheromaks in SSX

GA A27290 CALCULATION OF IMPURITY POLOIDAL ROTATION FROM MEASURED POLOIDAL ASYMMETRIES IN THE TOROIDAL ROTATION OF A TOKAMAK PLASMA

Observations of Rotation Reversal and Fluctuation Hysteresis in Alcator C-Mod Plasmas

ICRF Induced Argon Pumpout in H-D Plasmas at Alcator C-Mod

Simulation of Double-Null Divertor Plasmas with the UEDGE Code

Ion energy balance during fast wave heating in TORE SUPRA

Measurements of the Deuterium-Tritium Branching Ratio Using ICF Implosions

Magnetic Flux Surface Measurements at Wendelstein 7-X

Formation of High-b ECH Plasma and Inward Particle Diffusion in RT-1

Neutral Beam-Ion Prompt Loss Induced by Alfvén Eigenmodes in DIII-D

Transcription:

PSFC/JA-11-27 Effects of thermal expansion of crystal-lattice on x-ray imaging crystal spectrometers L. Delgado-Aparicio 1, M. Bitter 1, Y. Podpaly 2, J. Rice 2, W. Burke 2, M. Sanchez del Rio 3, K. Hill 1, N. Pablant 1, P. Beiersdorfer 4, R. Bell 1, R. Feder 1, C. Gao 2, D. Johnson 1, S. G. Lee 5, E. Marmar 2, M. L. Reinke 2, S. Scott 1 and R. Wilson 1 1 Princeton Plasma Physics Laboratory, Princeton, NJ, USA. 2 Massachusetts Institute of Technology, Cambridge, MA, USA. 3 European Synchrotron Radiation Facility, Grenoble, France. 4 Lawrence Livermore, National Laboratory, Livermore, CA, USA. 5 National Fusion Research Institute, Daejon, Korea November 2011 Plasma Science and Fusion Center Massachusetts Institute of Technology Cambridge MA 02139 USA This work was supported by US DoE awards DE-FC02-99ER54512 at MIT and DE- FC02-99ER54512 and DE-AC02-09CH11466 at PPPL and DE-AC52-07NA27344 at LLNL. Reproduction, translation, publication, use and disposal, in whole or in part, by or for the United States government is permitted. To be submitted for publication to Review of Scientific Instruments, 2011.

Effects of thermal expansion of crystal-lattice on x-ray imaging crystal spectrometers L. Delgado-Aparicio 1, M. Bitter 1, Y. Podpaly 2, J. Rice 2, W. Burke 2, M. Sanchez del Rio 3, K. Hill 1, N. Pablant 1, P. Beiersdorfer 4, R. Bell 1, R. Feder 1, C. Gao 2, D. Johnson 1, S. G. Lee 5, E. Marmar 2, M. L. Reinke 2, S. Scott 1 and R. Wilson 1 1 Princeton Plasma Physics Laboratory, Princeton, New Jersey 08543, USA 2 Massachusetts Institute of Technology - Plasma Science Fusion Center, Cambridge, Massachusetts 02139, USA 3 European Synchrotron Radiation Facility, Grenoble, France 4 Lawrence Livermore National Laboratory, Livermore, California 94550, USA 5 National Fusion Research Institute, Daejon, Korea December 8, 2011 Abstract X-ray imaging crystal spectrometers with high spectral and spatial resolution are currently being used on magnetically confined fusion schemes to infer the time history of profiles of the ion and electron temperatures as well as plasma flow velocities. Recent studies indicate that the crystal temperature must be kept constant within a fraction of a degree to avoid changes of the interplanar 2d-spacing by thermal expansion that cause changes in the Bragg angle, which could be misinterpreted as Doppler shifts. For the instrumental parameters of the x-ray crystal spectrometer on C-Mod, where those thermal effects were investigated, a change of the crystal temperature by one degree o C causes a change of the lattice spacing of the order of d = 1 10 5 Å. This effect must also be considered for x-ray imaging crystals spectrometers installed on LHD, KSTAR, EAST, NSTX and, in the future, ITER. ldelgado@pppl.gov Presently at the Massachusetts Institute of Technology - Plasma Science Fusion Center. 1

1 Introduction. High-resolution single-chord x-ray crystal spectrometers have been used to diagnose the core of tokamak plasmas for more than three decades (see ref. [1]-[2] and references therein). But thanks to the development of a new type of x-ray imaging crystal spectrometer which makes use of the imaging properties of spherically bent crystals [3]-[4], and due to important advances in the x-ray detector technology, especially, the manufacturing of two-dimensional pixilated x-ray detectors of large areas and high count rate capabilities [5], it is now possible to record spatially resolved x-ray spectra of highly charged ions from tokamak plasmas. These spectral data can be tomographically inverted to extract information on the local ion and electron temperatures (T i,e ) as well as toroidal and poloidal plasma flow velocity components (v φ,θ ). The ion temperature and flow measurements are based on measurements of the Doppler broadening and Doppler shift of the spectral lines, while the electron temperature is usually determined from the intensity ratios of the resonance line and the associated dielectronic satellites. At the heart of the x-ray imaging crystal spectrometers is a spherically bent crystal, which serves as the x-ray diffracting element. This spherically bent crystal Bragg reflects x-ray radiation emitted from the plasma onto a two-dimensional, position-sensitive x-ray detector, such that the chord-integrated spectra from multiple sightlines are spatially resolved and observed simultaneously (see references [3]-[14]). The spectrometer resolving power can be as high λ/dλ 2 10 4 while the spatial and temporal resolutions are typically, about 1 cm and 5-20 ms, respectively. This new type of x-ray imaging crystal spectrometer is currently being used on several fusion devices, including the National Spherical Torus Experiment (NSTX, see ref. [3]) and Alcator C-Mod (see ref. [4]- [8]) in the USA, the Korean Superconducting Tokamak Advanced Reactor (KSTAR, see ref. [9]), the Experimental Advanced Superconducting Tokamak (EAST, see ref. [10]-[12]) in China, and the Large Helical Device (LHD, see ref. [11]) in Japan. This type of x-ray imaging crystal spectrometer has also been selected by the International Thermonuclear Experimental Reactor (ITER) organization to be a primary diagnostic for profile measurements of ion temperature (T i ) and toroidal velocity (v φ ), and to be a secondary diagnostic for electron temperature (T e ) and poloidal velocity (v θ ) [13]-[14]; at least four coreviewing spectrometers will be used in the near future at ITER. The main objective of this paper is to investigate the long term thermal stability of this kind of crystal spectrometer in light of the harsh environment associated with a burning plasma in ITER; the dimensions and components of their future x-ray crystal imaging spectrometers (e.g. crystals, Bragg angles and x-ray detectors) will be very similar to those of the spectrometers presently used on C-Mod. This paper is organized as follows: The theoretically expected changes in crystal lattice spacing due to a thermal expansion of the crystal and resulting effects on wavelength measurements are described in Section 2. Experimental results from an investigation of these effects, using the x-ray imaging crystal spectrometer on C-Mod, are presented in Section 3; and implications for the ITER x-ray crystal spectrometers are discussed in Section 4. 2 Theoretical considerations. The basic equation for wavelength measurements with an x-ray crystal spectrometer is the Bragg condition: 2d sin θ = λ (1) where λ, d, and θ are the wavelength of an x-ray line, the interplanar spacing of the Bragg reflecting crys- Vacuum side Plasma He atmosphere 2D SXR detector cooling (Δθ>0) Incident λ ΔθΔθ θ heating (Δθ<0) Crystal (2d) Figure 1: (Color-online) Variation of Bragg angle with change of 2d spacing as a result of thermal expansion assuming a constant wavelength λ (not to scale). 2

a 1 Figure 2: (Color-online) Hexagonal structure of quartz crystals used on Alcator C-Mod. tal lattice planes, and the Bragg angle, respectively. One of the main tasks of the high resolution x-ray crystal spectrometers on tokamaks is to measure the velocity of an x-ray emitting ion from the observed wavelength shift (or Doppler shift) of a spectral line. The relationship between the Doppler shift, λ, and an ion velocity, v, is given by the Doppler equation: λ λ c = v c a 2 a 3 (2) From equation (1), we find - assuming that the lattice spacing, d, is constant - that the observed relative wavelength shift is given by λ λ = θ tan θ (3) On the other hand, we find from equation (1) that, for a constant wavelength, λ, a change, d, of the lattice spacing, e.g. by thermal expansion, leads to a change, θ, of the observed of the Bragg angle, given by d d = θ (4) tan θ The changes, θ, expected for a heating or cooling of the crystal lattice are illustrated in Fig. 1. Comparing equations (3) and (4), one finds that the observed changes θ of the Bragg angle due to a relative wavelength shift, λ/λ, and a relative change of the crystal lattice spacing, d/d, which may result from a thermal expansion of the crystal, have exactly the same functional dependence. The changes of the Bragg angle due to these very different causes are, therefore, experimentally indistinguishable. The relative changes of the crystal lattice spacing by thermal expansion are given by d d = α eff T (5) where, T is the change in temperature and α eff is an effective thermal expansion coefficient of the crystal which will be explained below. According to equations (3)-(5) and Figure 1, the expansion of the crystal due to an increases in temperature of T > 0 will result in a change θ < 0. This effect can easily be misinterpreted as a Doppler shift if the temperature variations of the crystal and its thermal expansion are unknown. Since the high-resolution crystal spectrometers on Alcator C-Mod as well as other tokamaks are designed to measure plasma flow velocities as low as 5-10 km/s, the minimum still observable Doppler shifts λ/λ must be of the order of 1 10 5. Thus it follows from equations (3) and (4) that errors due to the relative changes of the crystal lattice spacing d/d must be less than 10 5. Therefore, for our quartz crystals with d = 2.28112Å, the d-spacing must be kept constant to within 2 10 5 Å. The crystals most commonly used in x-ray spectrometers on tokamaks, are quartz crystals which have a hexagonal lattice structure. The interplanar 2d-spacing is given by [15], r h,k,l hex = 1 d hex h,k,l = 4 3a 2 0 [h 2 + k 2 + hk] + l2 c 2 0 (6) where, (h, k, l) are the Miller indices, and a 0 = 4913.04 må and c 0 = 5404.63 må are the lattice constants at 2d-interplanar spacing [ma] Hexagonal quartz (1,0,2) crystal o 4563.00 4562.75 4562.50 4562.25 For (1,0,2) quartz crystal installed at Alcator C-Mod 4562.00 0 5 10 15 ΔT=temperature change (T 0 =25 C) Figure 3: 2d-interplanar spacing as a function of temperature change ( T ) for the (1, 0, 2)-quartz crystal used at Alcator C-Mod. 3

Quartz crystals and mounts Thermal pad Be window Gate valve B-port He-like detectors HiReX-Sr housing Pilatus 2D detectors a) therefore given by: ( [ 4 h 2 + k 2 + hk ] d hex h,k,l(t ) = 3a 2 0 (1 + α T ) 2 + l 2 /c 2 0 (1 + α T ) 2 ) 1/2 (7) A plot of the 2d-spacing versus temperature excursion ( T ) using equation (7) is shown in Figure 3 for the quartz crystal cut with Miller indices (1, 0, 2) used at C-Mod. This function is well approximated by the linear expression: [2d] (T ) = [2d] 0 (1 + α eff T ) with an effective thermal expansion coefficient, α eff = 9.51021 10 6 / o C. L-shapped Al frame and rectangular He-like crystal He-like detectors H-like detector Annular-shapped Al frame and circular H-like crystal Readout electronics b) He-like arrangement Crystal Thermocouple Figure 4: (Color-online) Schematic views of HiReX- Sr spectrometer installed in Alcator C-Mod. a) The thermal pad used in the temperature-scan was placed on the outside wall next to the two quartz crystal mounts. b) The L- and annular-shaped aluminum frames support the glass substrates that hold the rectangular and circular He- and H-like crystals, respectively. 25 o C of the crystal unit cell as shown in Figure 2 [15]- [16]. Materials with anisotropic structures, such as quartz, have distinct thermal expansion coefficients in different directions. For quartz crystals the thermal expansion coefficients are α = 7.97 10 6o C 1 and α = 13.37 10 6 C 1 in the directions parallel to the c 0 and a 0, respectively [15]-[16]. The unit cell values may change from crystal to crystal and from reference to reference (e.g. synthetic quartz vs. natural quartz), which actually sheds light on the importance of having a crystal characterization and calibration prior to its installation and use. The lattice spacing for a given temperature excursion T is 3 Experimental results. Measurements of d/d as a function of temperature of the crystal were made by heating the outer wall of the spectrometer housing with a single thermal-pad as shown in the top-left insert of Figure 4-a). The thermal pad was located behind the crystals to selectively increase their temperature, and at the same time, minimize the temperature excursion of the other components installed throughout the spectrometer housing; this temperature measurements were done with two K-type thermocouples [17]. One probe was attached to the mount of the first crystal, which is used for measurements of He-like argon spectra and which is next to the heat source as shown in the bottomright insert of Figure 4-b), while the second thermocouple was installed in between the crystals and the x-ray detectors (at the center of the spectrometer Temperature ( o C) 40 35 30 25 Constant temperature during first ten ohmic plasmas Heating Heating Cooling Crystal mount 20 9 11 13 15 17 Time (hrs) Cooling Center spec. housing Figure 5: (Color-online) Time history of the temperature of the crystal mount during the heating and cooling phases. 4

Normalized intensity [counts] 1.0 0.8 0.6 0.4 0.2 Argon He-like spectra a) w-line n 4 10 n=3 140 160 180 C-Mod shot #s: 1100708001-10 Position of w-line [pixels] 163 b) =±4.26 10-3 pixel 5.5 5.0 =±1.79 10-3 pixel 162 4.5 2 4 6 8 Shot number Width of w-line [pixels] Normalized intensity [counts] 1.0 0.8 0.6 0.4 0.2 Argon He-like spectra a) w-line n 4 15 n=3 140 160 180 C-Mod shot #s: 1100708010-13,17-27 zoom Zoom of w-line heating cooling 15 C-Mod shot #s: 1100708010-13,17-27 145 150 160 Normalized intensity [counts] 1.0 0.8 0.6 0.4 0.2 Argon H-like spectra c) Lyα 1 -line 10 Lyα 2 -line Mo 32+ 0 20 40 C-Mod shot #s: 1100708001-10 Position of Ly α1 -line [pixels] 23 22 =±3.55 10-3 pixel =±2.17 10-3 pixel d) 2 4 6 8 Shot number 5.5 5.0 4.5 Width of Ly α1 -line [pixels] Normalized intensity [counts] 1.0 0.8 0.6 0.4 0.2 Argon H-like spectra b) Lyα 1 -line 15 Lyα 2 -line Mo 32+ 0 20 40 C-Mod shot #s: 1100708010-13,17-27 zoom Zoom of Ly α1 -line heating cooling 15 C-Mod shot #s: 1100708010-13,17-27 15 20 30 Figure 6: (Color-online) Overlay of He- and H-like argon spectra from 10- indicating high reproducibility of the spectral line positions during ohmic plasma conditions. The error bars from the multigaussian non-linear least-squares fit are well inside the data plot symbols and of the order of 10 3 pixels. housing). The spectrometer chamber was filled with a helium atmosphere at 1 atm pressure to avoid attenuation in air of the low energy ( 3 4 kev) x- rays of the observed argon spectra. The time histories of the temperature measurements obtained during a work day are shown in Figure 5. At the beginning of the experiment the spectrometer housing was kept at a constant room temperature of 25 o C (see Figure 5) and the exact positions of the spectral lines on the detector were determined by using ten consecutive and reproducible ohmic discharges. The temperature of the heating pad was then set to 80 o C using a remotely adjusted temperature controller. As is evident from Figure 5, the time-constant for the observed temperature rise in the crystal mount is of the order of 1 hr, contrary to that of the thermal pad which was of only 2-4 minutes. The heating phase was interrupted in order to secure the integrity of the crystal during a temporal malfunction of the magnetic coils of Alcator C-Mod, causing a drop in the temperature of the crystal mount shown in Figure 5. The experi- Figure 7: (Color-online) Overlay of spectral data from 15 during the temperature scan showing line shifts observed during the heating and cooling phases. mental conditions were later restored and the heating process begun after one hour. The heating was terminated when the crystal mount reached 40 o C which is the maximum operating temperature for the Pilatus detectors; the temperature excursion at the center of the spectrometer housing was less than 6 o C so we estimate a temperature increase of the detectors of only 1-2 o C. The measurements of the line positions were continued during the cooling process. Although the data were collected through several hours, we decided to base our conclusions on the data gathered during this last cooling period which was not interrupted by any impairments of the machine operation [see blue symbols in Figure 5 and Figures 8-a), b) and c)]. An overlay of the multi-gaussian fits of the resonance line w and its associated dielectronic satellites, as well as the Lyman-α doublet and the neon-like molybdenum line (Mo 32+ ) - which falls next to the Lyman-α doublet, are shown in Figures 6 and 7. The latter spectra were simultaneously recorded with the second crystal in our spectrometer. The pixel size on our detector is 172 µm. However, the Doppler-broadened line spreads over 5

10 pixels, so that it is possible to determined the position of the line with an accuracy much better that the pixel size. The line intensities, widths and centroid locations for the initial 10 consecutive did not change on a shot-to-shot basis (see Figs. 6) and were within the small error bars of (1.8-2.2) 10 3 and (3.5-4.3) 10 3 pixels, for the centroid position and line-width, respectively. The centroid positions, given in pixel numbers, for lines w, Lyman-α 1 and Lyman-α 2 at 25 o C are 162.45, 22.15 and 35.7 in pixel, respectively. Position of w-line [pixels] 163 162 (26.36,162.387) cooling the crystal a) ΔT~11.8 o C Δp~1.007 pixels (38.16,161.38) The same ohmic plasma conditions were used throughout the entire experiment. An overlay of the raw data showing the spectral features during the heating and cooling phases is depicted in Figure 7. Shifts of the centroid positions, as expected for the heating and cooling phases from the discussion that followed Figure 1, are already evident from these raw data. The changes in the measured centroid of the lines as a function of temperature are shown in Figure 8, where the two distinct branches correspond to the heating and cooling phases. The data inferred from the w-line and the Lyman-α doublet for the cooling process from 38 to 26 o C is summarized in Tables 1 and 2; R c is the radius of curvature of the crystal, while p is the experimentally inferred pixel shift and ds is the length of the arc subtended by the change θ in the Bragg angle due to the shrinkage of the 2d-spacing. The values of d/d inferred from the observed centroid shifts of the lines w and Lymanα doublet, listed in Tables 1 and 2 are 8.02 10 5, 12.59 10 5 and 12.47 10 5, respectively. On the other hand, the theoretically expected value of d/d according to equation (5) is 11.2 10 5 for the measured change in crystal mount temperature of 11.8 o C. The experimental data are therefore in reasonable agreement with our theoretical prediction. In the following we compare the effects of the thermal expansion of the crystal lattice with those produced by a thermal expansion of the metallic spectrometer housing, especially, the change of the distance between the crystals and the SXR detectors. The thermal expansion of the crystal lattice, for an increase of the crystal temperature by as much as 10 o C, has two effects: It causes (a) a shift of the w-line on the detector of the order of 1 pixel and shifts of similar magnitude for all the observed spectral lines, and (b) it reduces the separation between the w-line and the z-line on the detector (at the other end of the He-like Ar spectrum) by approximately 10 microns. Of these two effects (a) is by far the most significant. A change of the crystal-detector distance by thermal expansion will primarily change the sep- Position of Ly α1 -line [pixels] 22 21 (26.36,21.8946) cooling the crystal 161 ΔT~11.8 o C Δp~1.179 pixels (38.16,20.7149) 25 30 35 40 Temperature ( o C) 25 30 35 40 Temperature ( o C) b) c) Position of Ly α2 -line [pixels] 36 35 (26.36,35.4528) cooling the crystal ΔT~11.8 o C Δp~1.175 pixels (38.16,34.2775) 25 30 35 40 Temperature ( o C) Figure 8: (Color-online) Line-shifts from temperature-dependent changes of 2d-spacing. Parameters He-like (w line) λ (Å) 3.9492 θ B = sin 1 (λ/2d) 59.9542 o R c (mm) 1443 R c sin θ B 1249.10 p (pixels) 1.007 ds = p 0.172 0.1732 mm ds θ B R c sin θ B 1.387 10 4 d d = θ B tan θ B 8.02 10 5 Table 1: Spectroscopically measured parameters from the w-line of the He-like spectrometer used in Alcator C-Mod. The (1, 0, 2) quartz crystals have a 2dspacing of the order of 4562.25 må at 25 o C. aration between w- and z-lines and to a much lesser extent cause a shift of the entire spectrum, so that 6

Parameters Ly α,1 Ly α,2 λ (Å) 3.7311 3.7365 θ B = sin 1 (λ/2d) 54.8671 o 54.9852 o R c (mm) 1385 1385 R c sin θ B 1132.68 1134.32 p (pixels) 1.179 1.175 ds = p 0.172 0.2028 mm 0.2021 θ B d d ds R c sin θ B 1.79 10 4 1.781 10 4 = θ B tan θ B 12.59 10 5 12.47 10 5 Table 2: Spectroscopically measured parameters from the H-like spectrometer used in Alcator C-Mod. The (1, 0, 2) quartz crystals have a 2d-spacing of the order of 4562.25 må at 25 o C. with respect to line shifts caused by a change of the crystal-detector distance effect (b) is much more significant than effect (a). In order to reduce the separation between the w- and z-lines on the detector by the same amount of 10 microns, it would be necessary to change the crystal-detector distance of the C-Mod spectrometer by 0.5 mm, which based on the thermal expansion coefficient of 16.7 10 6 / o C for steel would require to reduce the temperature of the spectrometer housing by 25 o C. The edge and core x-ray imaging crystal spectrometers to be built for the International Thermonuclear Experimental Reactor (ITER) will have radial and partially tangential views to measure toroidal and poloidal velocities in the range from 1-100 km/s for a normalized radius of r/a < 0.9. According to the original specifications, the toroidal and poloidal velocities should both be measured with an accuracy below 30% [18]. For a typical core toroidal rotation measurement of the order of several tens to 100-200 km/s, the overall profile accuracy should be better than 10 km/s. A change of the temperature of the germanium crystals, which are presently being considered for the ITER spectrometers, by 1.0 degree o C would result in a d/d that is four times larger than that observed on C-Mod due to the fact that the typical thermal expansion coefficient for germanium is α eff = 36 10 6 / o C, so that the observed change in Bragg angle could be mistaken as a Doppler shift which corresponds to a velocity of 10 km/s. The measurements of the toroidal rotation for the outer plasma (r/a 0.5 0.9) will have an accuracy limited not only by signal-to-noise ratio constraints but also by the effect of changing the crystal temperature as discussed above. For a poloidal rotation measurement, with velocities typically an order of magnitude smaller, the problem is even more difficult. A peak poloidal velocity of 10-20 km/s implies a measurement accuracy that requires the crystal temperature excursion be smaller than 1/5 of a degree o C. Because the ITER spectrometer housing is kept at a constant temperature throughout the plasma, temperature excursions and spurious drifts in the inferred velocity estimates would likely arise from the x-ray flux heating the crystal surface. In summary, the thermal excursion of the crystals used in x-ray imaging spectrometers can affect the 2d interplanar spacing introducing changes in the Bragg angle that can be experimentally misinterpreted as Doppler shifts; such thermal drifts can be due to non-reproducible and varying ambient temperatures and possibly to the x-ray flux impinging on the crystal. It is therefore imperative to monitor the time-history of the crystal temperature with an accuracy better than one degree. This work was performed under US DoE contracts DE-FC02-99ER54512 at MIT, DE-AC02-09CH11466 at PPPL and DE-AC52-07NA27344 at LLNL. 4 Implications for ITER References [1] M. Bitter, et al., Can. J. Phys., 86, 291, (2008). [2] B. C. Stratton, et al., Fusion Science and Technology, 53, 431, (2008). [3] M. Bitter, et al., Rev. Sci. Instrum., 70, 292, (1999). [4] M. Bitter, et al., Rev. Sci. Instrum., 75, 3660, (2004). [5] See for instance DECTRIS website at: http://www.dectris.ch/sites/products.html [6] A. Ince-Cushman, Rotation Studies in Fusion Plasmas via Imaging X-ray Crystal Spectroscopy, Ph. D. thesis, MIT, (2008). [7] A. Ince-Cushman, et al., Rev. Sci. Instrum. 79, 10E302, (2008). [8] K. W. Hill, et al., Rev. Sci. Instrum., 79, 10E320, (2008). [9] S. G. Lee, et al., Rev. Sci. Instrum., 81, 10E506, (2010). 7

[10] Y. Shi, et al., Plasma Phys. Control. Fusion, 52, 085014, (2010). [11] M. Bitter, et al., PPPL report 2514, (2010). [12] N. Pablant, private communication, (2011). To be submitted to Rev. Sci. Instrum., (2011). [13] K. W. Hill, et al., Rev. Sci. Instrum., 81, 10E322, (2010). [14] P. Beiersdorfer, et al., J. Phys. B: At. Mol. Opt. Phys., 43, 144008, (2010). [15] B. E. Warren, X-ray difraction, Dover, NY, (1990). [16] R. W. G. Wyckoff, Crystal structures, 2nd edition, Vol. 1, John Wiley, (1963). [17] L. Delgado-Aparicio, et al., 52 nd Annual Meeting of the APS Division of Plasma Physics, Volume 55, Number 15, November 8-12, 2010. [18] A. J. H. Donne, et al., Nucl. Fusion, 47, S337- S384, (2007). 8