Plasma impurity composition in Alcator C-Mod tokamak. I. O. Bespamyatnov a, W. L. Rowan a, K. T. Liao a, M. Brookman a, M. L. Reinke b, E. S. Marmar b, M. J. Greenwald b a Institute for Fusion Studies, The University of Texas, Austin, TX b MIT Plasma Science and Fusion Center, Cambridge, MA bespam@physics.utexas.edu
Abstract Accurate characterization of the impurities in the tokamak plasma is an important and complex task. Impurities may lead to fuel dilution and enhanced radiative losses from plasma. Moreover, density of impurities relate to the effective ion charge Z eff, which greatly influences main-ion and electron transport. The task of quantifying the impurities in Alcator C-Mod tokamak is even more critical for few reasons. C-Mod s plasma contains a vast variety and amount of intrinsic or seeded impurities during regular operations. The most typical ones are: B, Ar, Mo, W, Ca, Ne, N, He. Some of the impurities are seeded for the diagnostic purposes, some for other reasons. Z eff of well confined C- Mod plasma may reach values more than 5. C-Mod s impurities dilute the main ions, influence confinement and particle and thermal transport. In order to properly characterize these impurity effects, density of each impurity should be accurately quantified. Impurities are typically diagnosed spectroscopically. Measurement of the absolute density requires proper radiometric diagnostic calibration, which is a difficult and sometimes impossible task. Some of the calibrations may be acquired from contrasting the specific impurity measurement with independent measurement of Z eff and neutron rate. Some of the diagnostics may be cross-calibrated. Here we attempt to perform a constrained analysis of the impurity content for a variety of the C-Mod plasma discharges. The goal here is to develop a technique to combine results from C-Mod impurity diagnostics into a joint analysis. The reliable estimate of impurity content and Z eff for most of the C-Mod plasma discharges is sought as the result. Few examples of the impurity composition analysis are discussed in detail. *Work supported by USDOE Awards DE-FG03-96ER54373 and DE-FC02-99-ER54512
Plasma composition Alcator C-Mod contains variety of intrinsic and seeded impurities (and main ions): 1. Mo and W from tiles 2. B- boronization 3. D, H, He main ions 4. Ar, He seeded for diagnostic purposes 5. Ne, N 2 puff decreases the Mo levels 6. Ca laser blow-off injection (transport studies) It is critical to know the relative composition of the plasma (Z eff, f imp, <Z> ion ) 1. High impurity content will lead to fuel dilution and enhanced radiative losses from a the plasma. 2. Z eff is a important parameter for plasma transport modeling 3. Z eff and f ion are needed for calculation of the neutral beam attenuation. 4. Z eff and <Z> ion profiles are also needed, not only <Z eff > nimp ne = nionzion = nh + DZH, D ( = 1) + nimpzimp fimp = n n Z Z = = 1 + f Z ( Z 1) 2 ion ion eff imp imp imp nionzion Z ion e fimpzimp ( Zimp 1) = 1+ fimpzimp
Visible bremsstrahlung Bremsstrahlung radiation broadband emission from electron-ion collision. Frequently used for Z eff measurement and calibrations of spectral diagnostics. Two measurements at C-Mod. 1. Z ave - single chord measurement of the chord-average Z eff. Chord location: F-port, midplane, R tan = 0.675m (λ=536nm) Uses: Interferometer measured average ne (nl 04 ) and GPC core temperature Te(0). Advantages: Fast data-analysis Disadvantage. No Z eff profile, Molecular pseudo-continuum + possible impurity lines 2. DALSA visible continuum imaging diagnostic*. Chord location: K-port, midplane, coverage (full LFS profile) (λ=536nm) Uses: TS ne and Te measurements. Abel inversion. Advantages: Full Z eff profile. Disadvantage. Molecular pseudo-continuum, mirror reflectivity changes in time. ε λ 28 2 1 1.89 10 nz e eff g ff 12400 W = exp 12 2 3 4π Te λ Teλ cm A sr * *E.S. Marmar et al, RSI, 72,1,(2001)
Z eff from neoclassical conductivity Z eff (neo) External and bootstrap plasma currents are calculated based on neoclassical theory and compared to measured plasma current and loop voltage: (σ neo (Z eff ) and j BS (Z eff ) are from*) Z eff is being iterated until calculated and measured currents agree. 1. Standard approach (current profile is unknown): Vloop ( measured ) Itot ( measured) = Iext + I BS ( neo) = σ neoda jbs ( neo) da 2π R + 0 Assumptions: Electric field E is at equilibrium and the constant across the profile. Z eff profile is flat 2. Extended approach: V ( measured ) j ( EFIT ) = j + j = E σ + j = σ + j loop tot ext BS ( neo) neo BS ( neo) neo BS ( neo) 2π R0 Assumptions: Electric field E is at equilibrium and the constant across the profile. EFIT J tot is well known (options: Kinetic EFIT+ MSE constraint) Advantages: full Z eff profile, less ambiguity in local convergence *O. Sauter et al, POP, 6,7 (1999)
Z eff from neoclassical conductivity Examples on Z eff (neo) profiles (extended approach) Based on two EFIT runs (Standard EFIT and MSE-constrained EFIT with SIR correction) EFIT runs by S. Shiraiwa and M. Greenwald Calculated Z eff profile is very sensitive to current profile shape.
Comparison of Z ave and Z eff(neo) Comparison of Z ave and chord averaged Z eff (neo) (extended approach) (left) Comparison of brightness (measured by Z ave ) and (calculated using TS and Z eff (neo) )(right) Zbright ( brightness) = ε λdl ε λ 28 2 2 1 1.89 10 nz e eff g ff 12400 W = exp 12 2 3 4π Te λ Teλ cm A sr
Impurity measurements at C-Mod There are many diagnostics which can provide some information on specific impurities. 1. CXRS can measure density profiles of a particular light impurities (like B 5+ and He 2+ ). 2. HIREX-Sr (if calibrated) can provide density profiles of some excitation states of seeded high-z impurities like Ar and Ca. 3. XEUS and VUV (not calibrated) average densities of many different impurities. Can serve as a measure of relative density change of an impurity in time * 4. H/D ratio Measurements of the integrated radiation. (Consistency check or calibration for a dominant impurity) 1. XTOMO- 3D structure and dynamics of SXR radiation in Alcator C-Mod. 2. Bolometers- Power loss from the plasma 3. AXUV - 3D structure and dynamics of XUV radiation in Alcator C-Mod. 4. Measurement of the neutron flux *M.L.Reinke et al, RSI 81, (2010)
Boron and He density profiles measured by CXRS Two CXRS systems. 1. Core-CXRS. Poloidal and toroidal array: Ion: B or He Location: F-port Resolution: 1 cm and 20 msec. Coverage: ρ Ψ =[0-1.0] 2. Edge-CXRS. A set of gas puff and beam CXRS systems Ion: boron Location. B, K and F-port. Resolution: 1.2mm, 5 msec. Coverage: ρ Ψ =[0.8-1.0] Both systems provide 2D(r, t) profiles. Examples for Core-CXRS system B 5+ as trace impurity He 2+ as main ion
Historical calibration of the HIREX-Sr Selected Ohmic shot with strong Ar puff (1110125006) Measured (non-calibrated) emissivity of Ar 16+ (Ly α ) and Ar 17+ (Z) lines compared to calculated emissivity. Emission rate coefficients are from FAC-based modeling (THACO). Calibration coefficients are calculated by comparing Ar densities to Z eff contributions. rec exc ion exc Ar Ar Ar Ar εly = CLy ne n 18+ σly n 17+ σly n 16+ σly ne n 17+ σ α α + + α α α Lyα rec exc ion exc Ar Ar Ar Ar εz = CZ ne n 17+ σz n 16+ σz n 15+ σz ne n 16+ + + σz f Ar n Ly 17+ Ar α 17+ = 2 n C n σ exc e Ly e Ly α ε α f Ar n ε n C n 16+ Ar 16+ = 2 Z exc e Z e σ Z 1 ( slope) = 1.27 10 C α Ly 7 1 ( slope) = 1.33 10 C Z 7
Neutron rate for consistency check Total fusion yield is measured by two neutron detectors: Fusion yield can also be calculate based on: 1. n e (from TS) 2. T i (from consolidated TFIT (HIREX+CXRS) 3. Z eff from neoclassical conductivity (flat) 4. n D (from calculated impurity composition(z eff, <Z> ion ) 5. Cross-sections from Bosch et al, NF 32, 611 (1992) Comparison of the measured and calculated yields may serve as a cross-check for impurity composition* Z Z n n R ( ) ion eff D= e /1+ H/ D Z 1 ion C.L.Fiore et al, RSI, 63,10 (1992) Measure of dilution *M.L.Reinke NSE fall seminar 9/14/2011
Impurity composition (Three H-modes) Shot 1110119005. Three H-modes (marked in red). Clear H to L transition. B T =4.8T. The high confinement, or H-mode regime, is characterized by edge transport barriers in both energy and particle channels. Argon was injected for HIREX-Sr. DNB was injected for CXRS. Boron is intrinsic.
Impurity composition (Three H-modes) Calculated Ar and B densities (from Hirex-Sr and CXRS) contribute to average plasma Z eff (neo). a b c The remainder on Z eff is a considered to be a contribution from Mo (a) A calculated contribution from Mo correlates with Mo line emission measured by McPherson (VUV). (b) The calculated impurity fractions for Ar, B and Mo. (c)
Impurity composition (Three H-modes) A measured and calculated neutron rates are compared on the graph. T i profiles: T e profile corrected by central HIREX-Sr T i (0) are used for calculation of the neutron yield. Results from two calculations are plotted. Z Z ion eff 1. n <Z> ion calculated from f imp. (blue) D= ne /1 ( + RH/ D) Z 1 ion 2. <Z> ion =5 (pure boron) (purple) Both calculations well agrees with measurement. Although, purple line is a bit below. Less agreement in the Ohmic phase
Impurity composition (Long steady H-modes) Shot 1110119007. Long steady EDA H-mode (marked in red). B T =4.4T. The high confinement, or H-mode regime, is characterized by edge transport barriers in both energy and particle channels. Argon was injected for HIREX-Sr. DNB was injected for CXRS. Boron is intrinsic.
Impurity composition (Long steady H-mode) Calculated Ar and B densities (from Hirex-Sr and CXRS) contribute to average plasma Z eff (neo). a b c The remainder on Z eff is a considered to be a contribution from Mo (a) A calculated contribution from Mo correlates with Mo line emission measured by McPherson (VUV). (b) The calculated impurity fractions for Ar, B and Mo. (c)
Impurity composition (Long steady H-mode) A measured and calculated neutron rates are compared on the graph. T i profiles: Te profiles corrected by central HIREX-Sr T i (0) are used for calculation of the neutron yield. Results from two calculations are plotted. Z Z ion eff 1. n <Z> ion calculated from f imp. (blue) D= ne ( 1 RH/ D) Z 1 ion 2. <Z> ion =5 (pure boron) (purple) Both calculations well agrees with measurement. Although, purple line is a bit below. Less agreement in the Ohmic phase.
Impurity composition (I-mode) Shot 1101209010. Long steady I-mode (marked in red). A new discovered I-mode regime is characterized by edge thermal transport barrier without a significant reduction in particle transport. Argon was injected for HIREX-Sr. DNB was injected for CXRS. Boron is intrinsic.
Impurity composition (I-mode) Calculated Ar and B densities (from Hirex-Sr and CXRS) contribute to average plasma Z eff (neo). a b c The remainder on Z eff is a considered to be a contribution from Mo (a) A calculated contribution from Mo correlates with Mo line emission measured by McPherson (VUV). (b) The calculated impurity fractions for Ar, B and Mo. (c)
Impurity composition (I-mode) A measured and calculated neutron rates are compared on the graph. T i profiles measured by HIREX-Sr are used for calculation of neutron yield. Results from two calculations are plotted. 1. <Z> ion calculated from f imp. (blue) Z Z ion eff n = n 1 R 2. <Z> ion =5 (pure boron) (purple) Z 1 ( ) D e H/ D ion Good example, where calculated impurity composition (blue) agrees with measurement, but <Z> ion =5 (purple) does not. If Z eff constitutes only from boron, then calculated neutron yield is twice smaller than measured. Less agreement during the second half of the I-mode. (possibly that T i is underestimated or some additional very high Z contributor to plasma (W?)
Summary and conclusions 1. A historical calibration analysis of the HIREX-SR was performed. In the result the Ar 17+ and Ar 16+ density profiles were be extracted from HIREX emissivity profiles of Ly-α and Z lines. 2. These Ar densities and CXRS measured B 5+ densities were used to analyze their contributions to Z eff. Average Z eff (neo) was calculated from measured I p and V loop and assuming that plasma conductivity is neoclassical. 3. Z eff (neo) was compared to Z ave, calculated from measured visible bremsstrahlung. 4. Three shots (Two H-mode and one I-mode) were analyzed. 5. For each shot the Ar and Mo densities and their contributions to Z eff were identified. 6. The remaining Z eff contributions were associated with Mo. The calculated Mo contributions well correlate with brightnesses of the Mo XXXII (127.81) line. 7. The results of the plasma impurity composition analysis were cross-checked by the comparison of the measured neutron yield and calculated neutron yield using measured T i, Z eff and calculated <Z> ion. Boron is not the main contributor to Z eff (for H-mode and I-mode). It seems Mo is the main contributor. Ar is also an important contributor for typical levels of Ar puff needed for HIREX-Sr analysis.
Proposed consolidated analysis and plans Create a new plasma composition widget The code collects data from different impurity spectroscopy diagnostics. All data will be split among several categories like: 1. Calibrated absolute 2D (r,t) profile of density of a specific impurity (or state). 2. Full calibrated 1D (t) average density of a specific impurity (or state). 3. Non-calibrated 2D (r,t) profile of density of a specific impurity (or state). 4. Non calibrated 1D (t) average density of a specific impurity (or state). User selects which data to use and selects the list of impurities for composition analysis. Neutron yield (as well as P rad, Axuv) serves as constrains for composition. Z eff : 2D and 1D profiles serve as another constraint. Some calibrations of non-calibrated diagnostics may be acquired as historical calibrations for dominant impurity. Plans for the near future. Apply the analysis for shots with Ne puffs. Apply the analysis for shots with He as minority and He as the main ion. (He measured by CXRS Try to use Z eff profiles (DALSA, Z eff(neo) ) to get a spatial resolved impurity composition.