- Email: [email protected]
Determination of species yield of ion sources used for intense neutral beam injector
153
Journal of Nuclear Materials Ill & 112 (1982) 153-158 North-Holland Publishing Company
DETERMINATION OF SPECIES YIELD OF ION SOURCES USED FOR INTENSE NEUTRAL BEAM INJECTOR * CC.
TSAI,
CF.
BARNETT,
H.H.
HASELTON,
R.A.
LANGLEY
and W.L. STIRLING
Fusion Energ): Division, Oak Ridge Notional Laboraiorf. P.0. Box Y, Oak Ridge, TN 37830, USA
For efficient plasma heating, ion sources of neutral beam injectors should be capable of producing ion beams with an atomic fraction of 90% or higher. Diagnostic techniques for quantitatively determining source species yield have been developed and evaluated. These indude a magnetic momentum analysis of the unneutralized ions passing through the neutralizer, an energy analysis of the atomic beam by electrostatic separation of ions emanating from a stripping cell, and a quantity versus implantation depth analysis of hydrogen implanted in a crystal by the Secondary Ion Mass Spectrometer (SIMS) technique. The operational features and the advantages and disadvantages of each technique will be discussed. If the effects of beamlet optics. energy straggling in the accelerator, and neutralizer gas scattering are taken into account. the results of the measurements using the three techniques are shown to be mutually consistent within experimental error.
The injection of energetic neutral beams of Do (or H”) is an effective heating technique for plasmas in magnetically confined fusion devices [l-6]. The development of the neutral beam technology has been aggressively pursued during the past decade [4,5]. In the United States, short-pulse neutral beam injectors have been developed for ATC, ORMAK, PLT, ISX-B, PDX, DIII, TFTR, ZXIIB, and TMX. The ongoing experiments on PLT, ISX-B, and PDX utilize low energy (
Carbide
Corporation.
0022-3 115/82/0~0-0000/$02.75
Multi-megawatt neutral beams are produced by efficient and powerful neutral beam injectors that are positive-ion-based injection systems. The main components of the neutral beam injectors are an ion source, a neutralizer, a bending magnet, ion dumps, and neutral beam calorimeter as shown in fig. 1. The ion source consists of a plasma generator and an ion accelerator. The plasma generator supplies positive ions that are extracted and accelerated to form an ion beam by a set of multi-aperture (or multi-slot) grids of the ion accelerator. The ion beams contain atomic and molecular ions of D” (H+), D: (Hz ), and D: (Hz ). Passing through the D, (H,) neutralizer, the ion beams are partially converted into neutral beams by dissociative and electron capture collisions. The energetic D” particles possess a kinetic energy that is equal, one-half, or one-third the ion acceleration energy. After removing the unneutralized ions at the neutralizer exit, the powerful neutral beams are injected into the target plasma of a fusion device. In the ion accelerator of a neutral beam injector, the gas density can be on the order of 3 X lOI cmp3. During acceleration, the primary ions of the source plasma will collide with gas molecules, creating secondary ions and neutrals via charge exchange and ionization processes. With their energies below the beam energy [ 161, these secondary particles will be measured as the molecular ions. Because of different trajectories [ 171 and additional angular and energy dispersions [ 181, the molecular ions measured may posses divergent angles that are different from those of the atomic ions.
0 1982 North-Holland
154
C. C. Tsar et al. / Determination
of species yield of ion sources ORNL-DWG
nm rLHS ^‘MA GENERATOR
ION
ACCELERATOR
( llil NEUTRALIZER
n
+D.,
III: I!b
$: VII:
VOLUME REGION TRANSITION REGION SURFACE REGION ENERGY STRAGGLING REGION NEUTRALIZATION REGION NEUTRAL-TARGET FRONT NEUTRAL-TARGET BACK
FE@
NEUTRAL PARTICLE ANALYZER
NEUTRAL TARGET
BENDING MAGNET
6
82-2483
/
/\
.DO _.._wD 0
‘I ION DUMPS
ETER ENERGY
ANALYZER
f
Fig. 1. Schematic of a conventional neutral beam injector and beam species analyzers.
Consequently, the spatial distribution of the beam species mix will vary axially and transversely. Different analyzers may estimate different percentages of the source species mix, even under the same beam conditions, because they sample either the entire beam or a part of the beam at different locations and because they have different acceptance angles. In fact, the source atomic ion yield determined from various diagnostics in various fusion research centers varies from below 70% to above 90% [ 19-261. This paper describes three separate diagnostic techniques for quantitatively estimating the source species yield that have been used at ORNL. These techniques include a magnetic momentum analysis of the unneutralized ions passing through the neutralizer, an energy analysis of the atomic beam by electrostatic separation of ions emanating from a stripping cell, and a quantity versus implantation depth analysis of hydrogen implanted into a crystal by the Secondary Ion Mass Spectrometer (SIMS) technique. In addition, advantages, disadvantages, and limitations of the techniques are given and preliminary results of certain beam characteristics are presented. At ORNL, the principal diagnostic used in the past has been the magnetic momentum analyzer. To improve the accuracy of the species mix estimation, other diagnostics, including both the particle energy analyzer and the implanted particle analyzer, are being studied. The operating features and an analysis of these three diagnostics will be described and discussed.
2. Magnetic momentum analyzer (MMA) The magnetic momentum analyzer (MMA) consists of a bending magnet at the neutralizer exit and a water-cooled calorimetric tube across the image center of the one-half energy ion dump (see fig. 1). Normally, the unneutralized ions, which emanate from an equilibrium neutralizer and pass through the bending magnet, will be separated into full, one-half, and one-third energy components and impinge upon their corresponding ion dumps. The power loadings of the ion dumps provide a rough indication of ion species mix in the source. Some of the ions will impinge on the calorimetric tube, dissipate their energy, and raise the cooling water temperature. The temperature peak of the cooling water is used to represent the relative amount of ions intercepted by the tube. Thus, the ion composition of the entire beam can be measured by changing the exciting current of the bending magnet. A typical curve obtained for an 80-keV beam is shown in fig. 2. The three peaks denoted by E, E/2, and E/3 are for ions which passed through the neutralizer but remained ions and originated from the H: , H: , and H: components, respectively, of the ion beam emanating from the source. The solid curves of fig. 3 show the E/2 and E/3 components obtained from the data of fig. 2. The width and location of the E/2 and E/3 components are adjusted to maintain the relationship such that the required magnetic field strength to deflect the ions to the calorimetric tube is proportional to the square root of the ion
CC. ORNL-
BEAM ENERGY 80 keV
Tsai et al. / Derermina,ion
DWG
FED
82-2484
= 4
CURRENT DENSITY= 126 mA/cm2
’ I
-! 10
0 IBM,
COIL
20 CURRENT
Fig. 2. Typical scan calorimetric probe.
30 40 50 OF BENDING MAGNET (A)
of ions from
the neutralizer
ORNL-DWG
1.4
across
the
FED
82-2484
,
E/3
H+
IONS
of species yield of ion sources
155
energy. The species ratio at the ion dump, as determined from the area of the E peak of fig. 2 and the E/2 and E/3 peaks of fig. 3, is J, : J2 : J3 = 84.5 : 7.0 : 8.5. The estimated species ratio at the exit of the ion accelerator, derived from the above values, is J,:J,:J,=72:11:17.
The cross sections for this estimation are given in ref. 27. In general, the beam properties of H+ , Hi, and Hz emanating from the source plasma should be the same; therefore, the shape of the peaks in the calorimetric scan for each energy component should also be the same if there are no other effects. If the calorimetric probe is located at or near the focal point of system, then it is possible to extract additional information from the experimental results. As an example of this the E compoent is shown as a dashed line in fig. 3 so that the difference in shape between the curves can be observed. The peak heights have been normalized. The difference in shape between the solid and dashed curves represents mainly straggling effects in the accelerator. Thus, the estimated species ratio determined by the area analysis represents the ion beam composition downstream from the ion accelerator and is referred to as the beam species ratio, whereas the species ratio upstream of the accelerator or adjacent to the source plasma is called the source species ratio. As given elsewhere [ 161, these effects which occur in the accelerator not only degrade the full energy component but also enrich the partial energy components (from near zero to near full energy) of the beam. These effects can be minimized in the data analysis if only the center portion of each component, represented by the peak heights of the curve in fig. 2, is used in the data analysis. Using this peak height analysis the estimated source species ratio is J,: J,: J,=79:9:12.
5
4.2
-
2
1.0
-
0 1
0.8
-
v) + 0.4
-
a
(6) E/2
H+,IONS
I o 28
30 I&,,
32
34
36
38
40
NORMALIZED COIL CURRENT OF BENDING MAGNET (A) Fig. 3. Unnormalized (solid) and normalized (dashed) curves for E/2 and E/3 unneutralized Hf ions as shown in fig. 2.
As an example of the type of study which can be done using the above described analysis the following measurements are described. Variations of the source atomic ion yield with respect to the beam current density were studied for PLT/ISX-B sources, and the results are shown in fig.4. From the curve denoted by ORNL, we notice that the source atomic ion yield. which is estimated from the peak height analysis, increases from - 70 to - 809, as the beam current density is increased from - 0.2 A/cm2 to 0.43 A/cm2. The other curve in the figure depicts the experimental results of the source atomic fraction that were obtained independently at Princeton Plasma Physics Laboratory (PPPL) by utilizing a particle energy analyzer [25]. The agreement between these two measurements is excellent.
C. C. Tsat ef al. / Determination ORNL-DW6
-
-
PPPL
--A--
ORNL
82-2480
FED
I
I
0.4
0.5
- -_---i 0.4
02
03
BEAM CURRENT DENSITY (A/cm2)
Fig. 4. Variation of source beam current
density
atomic ion yield with respect measured at PPPL and ORNL.
to
In fact, at PPPL, the maximum source atomic ion yield of - 87% was measured by a Doppler broadening scanning spectrometer 1261. The accuracy of the MMA is degraded by the uncertainty of the equilibrium gas-line density in the neutralizer, the reionization and neutralization effect in the bending magnet, the turbulence of the cooling water, and the reflection of energetic ions from the calorimetric tubes and the ion dump. The uncertainty in the measurement is about 5% and the uncertainty in estimating the source species yield is about 10%. The measured results seem to be independent of beam optics and neutralizer dispersion effects.
3. Particle Energy Analyzer (PEA) Downstream from the neutral target (fig. 1). a particle energy analyzer is located to study ion or neutral beamlets passing through a hole 6.4 mm in diameter in the target. This analyzer consits of a calibrated stripping cell and a parallel plate electrostatic energy analyzer [23]. The designed acceptance angles of the Faraday cups are sufficiently large for minimizing the effect of energy dispersion in the stripping cell and in the energy analyzer. The uncertainty in the measurement is about 3%, whereas the uncertainty in estimating the source atomic ion yield is about 5%. If the bending magnet
of species
yield
of ion sources
field is set equal to zero, then the resulting beam will consist of both ions and neutrals and the results obtained by the PEA can provide an independent measurement of the source species mix because: (1) the small acceptance angle (- i- 1.2”) of the stripping cell tends to screen out the secondary particles created in the accelerator due to their poor optics. (2) The corresponding detectors only accepts E/2 and E/3 particles with a small energy dispersion of 2.8 keV and tend to screen out the secondary particles with other energies. The study of the composition of ion beams produced by a PDX/ISX source elaborated below. At a beam current density of -0.1 A/cm’. the estimated source atomic ion yield as determined by PEA is about 79% of the total beam including ions and neutrals, whereas the estimated source atomic ion yield determined by MMA is about 73%. The 6% discrepancy is probably due to the fact that the former measures only the center of the beam whereas the latter measures the entire beam. The following discussion shows that such a difference is within the established uncertainties of the two techniques. The particle energy analyzer was installed on a movable frame. It could be moved in three directions: along the beam axis, up-down, and left-right across the beam. After the analyzer is aligned with the beam, the data for the above results were collected. This method only sampled a small fraction of the beam. During this experiment, we observed the following interesting features: (1) The estimaed source atomic ion yield increases with the beam density, for instance. from 79% at 0.1 A/cm2 to 82% at 0.13 A/cm’. This fact is consistent with the results shown in fig. 4. (2) The estimated source atomic ion yield is a sensitive function of beam optics. For instance. the measured species ratio for 22-keV beams with varius beam current densities is given as J, : Jz : J3 = 70 : 8 : 22 at 0.075 A/cm2 underdense J, : J,: J3 = 73: 8: 19 at 0.092 A/cm’ minimum divergence J, : J, : J3 = 67 : 10 : 22 at 0.103 A/cm2 overdense (3) The full energy atomic yield is - 80% at the beam center but drops below 70% at the beam edge. Considering the fact that the energy and angular dispersion in the neutralizer region is worse for molecular ions than that for atomic ions, this observed phenomenon is understandable. (4) The estimated source atomic ion yield is higher for the center beamlets (e.g., > 80% for the center beamlets and ~70% for the edge beamlets). This fact
C.C. Tsai et al. / Determination of species yield of ion sources
could be caused by the variation of (i) the source atomic ion yield across the grid, (ii) the beamlet optics due to the nonuniform source plasma, or (iii) the energy straggling due to the nonuniform distribution of the gas density in the accelerator and/or in the neutralizer. The exact cause of this feature needs further study. (5) The estimated source atomic ion yield is higher for an overdense gas density in the neutralizer, in the stripping cell, and/or in the region between them. This can be understood because the partial energy particles suffer serious energy and angular dispersion under such low energy, 20 to 30 keV of the beams under study, and are inefficiently detected. In addition to the valuable information just mentioned, this analyzer is capable of studying temporal variation of beam composition. It is an on-line in situ instrument. Obviously, it can be used for on-line optimization of the full energy fraction in neutral beams and for studying the effect of the source operating parameters on the species variations. The approach to equilibrium in the neutralizer can be studied by changing the neutralizer gas line density.
4. Implanted Particle Analyzer (IPA) Silicon crystals were located near the center of the neutral beam, either in front of or behind the neutral species target, and were irradiated by the neutral beam. The depth profiles of implanted hydrogen were then determined by SIMS analysis. As reported elsewhere [22], the estimated beam species ratio for 40-keV, 75-A beams with a current density of 0.25 A/cm2 is J,:J*:J,=70:15:15. The uncertainty in the measurement of peak areas as determined by the SIMS analysis is about 3%, but the uncertainty in the estimated source atomic ion yield is about 10%. Under the same beam conditions, the magnetic momentum analyzer was used to measure the unneutralized ions. The estimated beam species ratio from the three peak areas is J,:J,:J,=70:18:12. By detecting all energy components of the beam, the IPA will properly estimate the atomic fraction of the beam. Thus, the difference in the estimated beam atomic ion yield between the SIMS analyzer and the magnetic momentum analyzer is negligible and within the experimental uncertainty. This analyzer only measures a small sample of the beam. With many samples of irradiated crystals, it is
157
capable of studying species composition across the beam, of the center beamlets, and of edge beamlets. It can study source species mix from both ion and neutral beams. One of the outstanding features is that the accuracy of the implanted particle measurement is essentially independent of the beam optics and the angular dispersion in the neutralizer because the deposition profile depends mainly upon the particle energy. However, the measurement accuracy does depend upon crystal temperature, particle flux saturation, and the profile distribution assumption. The resolution between the E/2 and E/3 components can be substantially improved if the particle energy is increased and beryllium target is used. As stated above, the IPA measures the beam species ratio and estimates a lower source atomic ion yield than the previusly described techniques. This technique does not lend itself to on-line measurements.
5. Discussion From the above three analyzers, the estimated source atomic ion yield for duoPIGatron ion sources is between 70 and 80% at - 0.15 A/cm2. The variation can be the result of measurement errors and different beam conditions. In general, these analyzers can be used to estimate the source species mix and the neutral beam composition. Their merits are summarized below: (1) The magnetic momentum analyzer can estimate both the beam and the source species mix. Detailed analysis of the peak shapes can reveal energy straggling effects in the ion accelerator. (2) The particle energy analyzer can study the temporal variation of source species mix as a function of beam divergence, beam current density, gas feed in the source and in the neutralizer, beam radius, and grid radius. It is portable and can be used in various test facilities. It can be used for optimizing full energy neutrals for neutral beam applications. (3) The implanted particle analyzer can study the species of neutral and ion beams. Its capabilities are similar to that of the particle energy analyzer; particularly, its particle measurement is immune from the effect of beam optics and energy dispersion. It is not an on-line instrument. The implanted particle profile in crystals is analyzed at another facility; however, it is very convenient for comparison purposes of various ion sources. Preliminary measurements have been made to extend this technique to studies of low, medium, and high Z impurities and their energy distributions. Under neutral beam conditions, the gas density in
158
C. C. Tsai et al. / Determination
the accelerator can be on the order of 3 X lOi crnd3. With such high gas density, the energy straggling effect can seriously degrade the source atomic ion yields. The ion sources being developed for neutral beam application in various fusion research centers can be operated at different source gas efficiencies, grid temperatures. beam energy, and gas line densities in the neutralizer. The variation of the gas line density in the accelerator of these ion sources can be as large as 2 50%. Even with the same ion source and the same species analyzer, such a large fluctuation in the gas density may lead to at least k 10% fluctuation of the source atomic ion yield. This may explain claims of > 90% source atomic ion yield. Moreover, different analyzers in different test facilities have different acceptance angles. For instance, for an analyzer with kO.2” acceptance angle, the atomic ions will be the dominant particles measured because the partial energy particles with a larger energy dispersion will not be measured as efficiently as the full energy component. Consequently, the measured source atomic ion yield will be unusually high, say > 95%. Thus, the source atomic ion yield should be determined under standardized beam conditions by a few reliable analyzers.
Acknowledgements The authors would like to thank their colleagues for their assistance in this work: G.C. Barber, N.S. Ponte, D.O. Sparks, and R.E. Wright for electrical and electronics support; C.W. Blue, R.R. Feezell, S.C. Forrester, and J.A. Moeller for the test stand maintenance; D.E. Schechter and F. Sluss for the preparation of ion sources; and W.K. Dagenhart, W.L. Gardner, G.G. Kelley. M.M. Menon, P.M. Ryan, and J.H. Whealton for the fruitful discussions and the participation in experiments.
of species yield of ion sources
References [II J.F. Clarke, Proc. IEEE 69 (1981) 867. VI J. Sheffield, Proc. IEEE 69 (1981) 885. [31 T.C. Simonen, Proc. IEEE 69 (1981) 935. I41 M.M. Menon, Proc. IEEE 69 (1981) 1012. Berkeley Laboratory I51 W.B. Kunkel. Lawrence
Report #LBL-10447 (1980). [61 H. Eubank et al., Phys. Rev. Letters 43 (1979) 270. I71 Proc. 3rd IAEA Neutral Beam Workshop, to be published. PI INTOR Group, in: Rep. Intern. Tokamak Workshop. IAEA, Vienna, 1980. [91 D.H. Metzler and L.D. Stewart. in: 9th Symp. Eng. Prob. Fusion Research (1981) p. 748. [lOI W.A. Houlberg et al.. in: 9th Symp. Prob. Fusion Research (1981) p. 688. IIll L.M. Hively et al., in: 9th Symp. Eng. Prob. Fusion Research (1981) p. 715. I121 D.H. Metzler, in: 9th Symp. Eng. Prob. Fusion Research (1981) p. 1286. [I31 D.E. Post et al., Princeton Plasma Physics Laboratory Report MATT- 1262 ( 1976). 1141 D.L. Jassby. Nucl. Fusion 17 (1977) 309. [I51 G. Porter, private communication. [161 C.C. Tsai, in: Proc. 3rd IAEA Neutral Beam Workshop, to be published. iI71 Y. Ohara et al.. J. Appl. Phys. 51 (1980) 3614. [I81 H.H. Fleischmann. C.F. Barnett and J.A. Ray, Phys. Rev. A10 (1974) 569. 1191 Y. Okumura et al., Rev. Sci. Instrum. 52 (1981) I. WI T. Obiki et al., Rev. Sci. Instrum. 52 (1981) 1445. Pll C.F. Burrell et al., Rev. Sci. Instrum. 51 (1980) 1451. P21 R.A. Langley and C.W. Magee, J. Nucl. Mater. 93/94 (1981) 390. [231 C.F. Barnett and J.A. Ray, Oak Ridge National Laboratory Report ORNL/TM-7656 (1980). Report [241 C.C. Tsai et al., Oak Ridge National Laboratory ORNL/TM-6360 (1978). [251 L.R. Grisham et al., Princeton Plasma Physics Laboratory Report PPPL-1484 (1978). WI M. Ulrickson et al.. Bull. Am. Phys. Sot. 23 (1978) 796. Re[271 C.F. Barnett et al.. Oak Ridge National Laboratory port ORNL-5206-07 (1977).
Journal of Nuclear Materials Ill & 112 (1982) 153-158 North-Holland Publishing Company
DETERMINATION OF SPECIES YIELD OF ION SOURCES USED FOR INTENSE NEUTRAL BEAM INJECTOR * CC.
TSAI,
CF.
BARNETT,
H.H.
HASELTON,
R.A.
LANGLEY
and W.L. STIRLING
Fusion Energ): Division, Oak Ridge Notional Laboraiorf. P.0. Box Y, Oak Ridge, TN 37830, USA
For efficient plasma heating, ion sources of neutral beam injectors should be capable of producing ion beams with an atomic fraction of 90% or higher. Diagnostic techniques for quantitatively determining source species yield have been developed and evaluated. These indude a magnetic momentum analysis of the unneutralized ions passing through the neutralizer, an energy analysis of the atomic beam by electrostatic separation of ions emanating from a stripping cell, and a quantity versus implantation depth analysis of hydrogen implanted in a crystal by the Secondary Ion Mass Spectrometer (SIMS) technique. The operational features and the advantages and disadvantages of each technique will be discussed. If the effects of beamlet optics. energy straggling in the accelerator, and neutralizer gas scattering are taken into account. the results of the measurements using the three techniques are shown to be mutually consistent within experimental error.
The injection of energetic neutral beams of Do (or H”) is an effective heating technique for plasmas in magnetically confined fusion devices [l-6]. The development of the neutral beam technology has been aggressively pursued during the past decade [4,5]. In the United States, short-pulse neutral beam injectors have been developed for ATC, ORMAK, PLT, ISX-B, PDX, DIII, TFTR, ZXIIB, and TMX. The ongoing experiments on PLT, ISX-B, and PDX utilize low energy (
Carbide
Corporation.
0022-3 115/82/0~0-0000/$02.75
Multi-megawatt neutral beams are produced by efficient and powerful neutral beam injectors that are positive-ion-based injection systems. The main components of the neutral beam injectors are an ion source, a neutralizer, a bending magnet, ion dumps, and neutral beam calorimeter as shown in fig. 1. The ion source consists of a plasma generator and an ion accelerator. The plasma generator supplies positive ions that are extracted and accelerated to form an ion beam by a set of multi-aperture (or multi-slot) grids of the ion accelerator. The ion beams contain atomic and molecular ions of D” (H+), D: (Hz ), and D: (Hz ). Passing through the D, (H,) neutralizer, the ion beams are partially converted into neutral beams by dissociative and electron capture collisions. The energetic D” particles possess a kinetic energy that is equal, one-half, or one-third the ion acceleration energy. After removing the unneutralized ions at the neutralizer exit, the powerful neutral beams are injected into the target plasma of a fusion device. In the ion accelerator of a neutral beam injector, the gas density can be on the order of 3 X lOI cmp3. During acceleration, the primary ions of the source plasma will collide with gas molecules, creating secondary ions and neutrals via charge exchange and ionization processes. With their energies below the beam energy [ 161, these secondary particles will be measured as the molecular ions. Because of different trajectories [ 171 and additional angular and energy dispersions [ 181, the molecular ions measured may posses divergent angles that are different from those of the atomic ions.
0 1982 North-Holland
154
C. C. Tsar et al. / Determination
of species yield of ion sources ORNL-DWG
nm rLHS ^‘MA GENERATOR
ION
ACCELERATOR
( llil NEUTRALIZER
n
+D.,
III: I!b
$: VII:
VOLUME REGION TRANSITION REGION SURFACE REGION ENERGY STRAGGLING REGION NEUTRALIZATION REGION NEUTRAL-TARGET FRONT NEUTRAL-TARGET BACK
FE@
NEUTRAL PARTICLE ANALYZER
NEUTRAL TARGET
BENDING MAGNET
6
82-2483
/
/\
.DO _.._wD 0
‘I ION DUMPS
ETER ENERGY
ANALYZER
f
Fig. 1. Schematic of a conventional neutral beam injector and beam species analyzers.
Consequently, the spatial distribution of the beam species mix will vary axially and transversely. Different analyzers may estimate different percentages of the source species mix, even under the same beam conditions, because they sample either the entire beam or a part of the beam at different locations and because they have different acceptance angles. In fact, the source atomic ion yield determined from various diagnostics in various fusion research centers varies from below 70% to above 90% [ 19-261. This paper describes three separate diagnostic techniques for quantitatively estimating the source species yield that have been used at ORNL. These techniques include a magnetic momentum analysis of the unneutralized ions passing through the neutralizer, an energy analysis of the atomic beam by electrostatic separation of ions emanating from a stripping cell, and a quantity versus implantation depth analysis of hydrogen implanted into a crystal by the Secondary Ion Mass Spectrometer (SIMS) technique. In addition, advantages, disadvantages, and limitations of the techniques are given and preliminary results of certain beam characteristics are presented. At ORNL, the principal diagnostic used in the past has been the magnetic momentum analyzer. To improve the accuracy of the species mix estimation, other diagnostics, including both the particle energy analyzer and the implanted particle analyzer, are being studied. The operating features and an analysis of these three diagnostics will be described and discussed.
2. Magnetic momentum analyzer (MMA) The magnetic momentum analyzer (MMA) consists of a bending magnet at the neutralizer exit and a water-cooled calorimetric tube across the image center of the one-half energy ion dump (see fig. 1). Normally, the unneutralized ions, which emanate from an equilibrium neutralizer and pass through the bending magnet, will be separated into full, one-half, and one-third energy components and impinge upon their corresponding ion dumps. The power loadings of the ion dumps provide a rough indication of ion species mix in the source. Some of the ions will impinge on the calorimetric tube, dissipate their energy, and raise the cooling water temperature. The temperature peak of the cooling water is used to represent the relative amount of ions intercepted by the tube. Thus, the ion composition of the entire beam can be measured by changing the exciting current of the bending magnet. A typical curve obtained for an 80-keV beam is shown in fig. 2. The three peaks denoted by E, E/2, and E/3 are for ions which passed through the neutralizer but remained ions and originated from the H: , H: , and H: components, respectively, of the ion beam emanating from the source. The solid curves of fig. 3 show the E/2 and E/3 components obtained from the data of fig. 2. The width and location of the E/2 and E/3 components are adjusted to maintain the relationship such that the required magnetic field strength to deflect the ions to the calorimetric tube is proportional to the square root of the ion
CC. ORNL-
BEAM ENERGY 80 keV
Tsai et al. / Derermina,ion
DWG
FED
82-2484
= 4
CURRENT DENSITY= 126 mA/cm2
’ I
-! 10
0 IBM,
COIL
20 CURRENT
Fig. 2. Typical scan calorimetric probe.
30 40 50 OF BENDING MAGNET (A)
of ions from
the neutralizer
ORNL-DWG
1.4
across
the
FED
82-2484
,
E/3
H+
IONS
of species yield of ion sources
155
energy. The species ratio at the ion dump, as determined from the area of the E peak of fig. 2 and the E/2 and E/3 peaks of fig. 3, is J, : J2 : J3 = 84.5 : 7.0 : 8.5. The estimated species ratio at the exit of the ion accelerator, derived from the above values, is J,:J,:J,=72:11:17.
The cross sections for this estimation are given in ref. 27. In general, the beam properties of H+ , Hi, and Hz emanating from the source plasma should be the same; therefore, the shape of the peaks in the calorimetric scan for each energy component should also be the same if there are no other effects. If the calorimetric probe is located at or near the focal point of system, then it is possible to extract additional information from the experimental results. As an example of this the E compoent is shown as a dashed line in fig. 3 so that the difference in shape between the curves can be observed. The peak heights have been normalized. The difference in shape between the solid and dashed curves represents mainly straggling effects in the accelerator. Thus, the estimated species ratio determined by the area analysis represents the ion beam composition downstream from the ion accelerator and is referred to as the beam species ratio, whereas the species ratio upstream of the accelerator or adjacent to the source plasma is called the source species ratio. As given elsewhere [ 161, these effects which occur in the accelerator not only degrade the full energy component but also enrich the partial energy components (from near zero to near full energy) of the beam. These effects can be minimized in the data analysis if only the center portion of each component, represented by the peak heights of the curve in fig. 2, is used in the data analysis. Using this peak height analysis the estimated source species ratio is J,: J,: J,=79:9:12.
5
4.2
-
2
1.0
-
0 1
0.8
-
v) + 0.4
-
a
(6) E/2
H+,IONS
I o 28
30 I&,,
32
34
36
38
40
NORMALIZED COIL CURRENT OF BENDING MAGNET (A) Fig. 3. Unnormalized (solid) and normalized (dashed) curves for E/2 and E/3 unneutralized Hf ions as shown in fig. 2.
As an example of the type of study which can be done using the above described analysis the following measurements are described. Variations of the source atomic ion yield with respect to the beam current density were studied for PLT/ISX-B sources, and the results are shown in fig.4. From the curve denoted by ORNL, we notice that the source atomic ion yield. which is estimated from the peak height analysis, increases from - 70 to - 809, as the beam current density is increased from - 0.2 A/cm2 to 0.43 A/cm2. The other curve in the figure depicts the experimental results of the source atomic fraction that were obtained independently at Princeton Plasma Physics Laboratory (PPPL) by utilizing a particle energy analyzer [25]. The agreement between these two measurements is excellent.
C. C. Tsat ef al. / Determination ORNL-DW6
-
-
PPPL
--A--
ORNL
82-2480
FED
I
I
0.4
0.5
- -_---i 0.4
02
03
BEAM CURRENT DENSITY (A/cm2)
Fig. 4. Variation of source beam current
density
atomic ion yield with respect measured at PPPL and ORNL.
to
In fact, at PPPL, the maximum source atomic ion yield of - 87% was measured by a Doppler broadening scanning spectrometer 1261. The accuracy of the MMA is degraded by the uncertainty of the equilibrium gas-line density in the neutralizer, the reionization and neutralization effect in the bending magnet, the turbulence of the cooling water, and the reflection of energetic ions from the calorimetric tubes and the ion dump. The uncertainty in the measurement is about 5% and the uncertainty in estimating the source species yield is about 10%. The measured results seem to be independent of beam optics and neutralizer dispersion effects.
3. Particle Energy Analyzer (PEA) Downstream from the neutral target (fig. 1). a particle energy analyzer is located to study ion or neutral beamlets passing through a hole 6.4 mm in diameter in the target. This analyzer consits of a calibrated stripping cell and a parallel plate electrostatic energy analyzer [23]. The designed acceptance angles of the Faraday cups are sufficiently large for minimizing the effect of energy dispersion in the stripping cell and in the energy analyzer. The uncertainty in the measurement is about 3%, whereas the uncertainty in estimating the source atomic ion yield is about 5%. If the bending magnet
of species
yield
of ion sources
field is set equal to zero, then the resulting beam will consist of both ions and neutrals and the results obtained by the PEA can provide an independent measurement of the source species mix because: (1) the small acceptance angle (- i- 1.2”) of the stripping cell tends to screen out the secondary particles created in the accelerator due to their poor optics. (2) The corresponding detectors only accepts E/2 and E/3 particles with a small energy dispersion of 2.8 keV and tend to screen out the secondary particles with other energies. The study of the composition of ion beams produced by a PDX/ISX source elaborated below. At a beam current density of -0.1 A/cm’. the estimated source atomic ion yield as determined by PEA is about 79% of the total beam including ions and neutrals, whereas the estimated source atomic ion yield determined by MMA is about 73%. The 6% discrepancy is probably due to the fact that the former measures only the center of the beam whereas the latter measures the entire beam. The following discussion shows that such a difference is within the established uncertainties of the two techniques. The particle energy analyzer was installed on a movable frame. It could be moved in three directions: along the beam axis, up-down, and left-right across the beam. After the analyzer is aligned with the beam, the data for the above results were collected. This method only sampled a small fraction of the beam. During this experiment, we observed the following interesting features: (1) The estimaed source atomic ion yield increases with the beam density, for instance. from 79% at 0.1 A/cm2 to 82% at 0.13 A/cm’. This fact is consistent with the results shown in fig. 4. (2) The estimated source atomic ion yield is a sensitive function of beam optics. For instance. the measured species ratio for 22-keV beams with varius beam current densities is given as J, : Jz : J3 = 70 : 8 : 22 at 0.075 A/cm2 underdense J, : J,: J3 = 73: 8: 19 at 0.092 A/cm’ minimum divergence J, : J, : J3 = 67 : 10 : 22 at 0.103 A/cm2 overdense (3) The full energy atomic yield is - 80% at the beam center but drops below 70% at the beam edge. Considering the fact that the energy and angular dispersion in the neutralizer region is worse for molecular ions than that for atomic ions, this observed phenomenon is understandable. (4) The estimated source atomic ion yield is higher for the center beamlets (e.g., > 80% for the center beamlets and ~70% for the edge beamlets). This fact
C.C. Tsai et al. / Determination of species yield of ion sources
could be caused by the variation of (i) the source atomic ion yield across the grid, (ii) the beamlet optics due to the nonuniform source plasma, or (iii) the energy straggling due to the nonuniform distribution of the gas density in the accelerator and/or in the neutralizer. The exact cause of this feature needs further study. (5) The estimated source atomic ion yield is higher for an overdense gas density in the neutralizer, in the stripping cell, and/or in the region between them. This can be understood because the partial energy particles suffer serious energy and angular dispersion under such low energy, 20 to 30 keV of the beams under study, and are inefficiently detected. In addition to the valuable information just mentioned, this analyzer is capable of studying temporal variation of beam composition. It is an on-line in situ instrument. Obviously, it can be used for on-line optimization of the full energy fraction in neutral beams and for studying the effect of the source operating parameters on the species variations. The approach to equilibrium in the neutralizer can be studied by changing the neutralizer gas line density.
4. Implanted Particle Analyzer (IPA) Silicon crystals were located near the center of the neutral beam, either in front of or behind the neutral species target, and were irradiated by the neutral beam. The depth profiles of implanted hydrogen were then determined by SIMS analysis. As reported elsewhere [22], the estimated beam species ratio for 40-keV, 75-A beams with a current density of 0.25 A/cm2 is J,:J*:J,=70:15:15. The uncertainty in the measurement of peak areas as determined by the SIMS analysis is about 3%, but the uncertainty in the estimated source atomic ion yield is about 10%. Under the same beam conditions, the magnetic momentum analyzer was used to measure the unneutralized ions. The estimated beam species ratio from the three peak areas is J,:J,:J,=70:18:12. By detecting all energy components of the beam, the IPA will properly estimate the atomic fraction of the beam. Thus, the difference in the estimated beam atomic ion yield between the SIMS analyzer and the magnetic momentum analyzer is negligible and within the experimental uncertainty. This analyzer only measures a small sample of the beam. With many samples of irradiated crystals, it is
157
capable of studying species composition across the beam, of the center beamlets, and of edge beamlets. It can study source species mix from both ion and neutral beams. One of the outstanding features is that the accuracy of the implanted particle measurement is essentially independent of the beam optics and the angular dispersion in the neutralizer because the deposition profile depends mainly upon the particle energy. However, the measurement accuracy does depend upon crystal temperature, particle flux saturation, and the profile distribution assumption. The resolution between the E/2 and E/3 components can be substantially improved if the particle energy is increased and beryllium target is used. As stated above, the IPA measures the beam species ratio and estimates a lower source atomic ion yield than the previusly described techniques. This technique does not lend itself to on-line measurements.
5. Discussion From the above three analyzers, the estimated source atomic ion yield for duoPIGatron ion sources is between 70 and 80% at - 0.15 A/cm2. The variation can be the result of measurement errors and different beam conditions. In general, these analyzers can be used to estimate the source species mix and the neutral beam composition. Their merits are summarized below: (1) The magnetic momentum analyzer can estimate both the beam and the source species mix. Detailed analysis of the peak shapes can reveal energy straggling effects in the ion accelerator. (2) The particle energy analyzer can study the temporal variation of source species mix as a function of beam divergence, beam current density, gas feed in the source and in the neutralizer, beam radius, and grid radius. It is portable and can be used in various test facilities. It can be used for optimizing full energy neutrals for neutral beam applications. (3) The implanted particle analyzer can study the species of neutral and ion beams. Its capabilities are similar to that of the particle energy analyzer; particularly, its particle measurement is immune from the effect of beam optics and energy dispersion. It is not an on-line instrument. The implanted particle profile in crystals is analyzed at another facility; however, it is very convenient for comparison purposes of various ion sources. Preliminary measurements have been made to extend this technique to studies of low, medium, and high Z impurities and their energy distributions. Under neutral beam conditions, the gas density in
158
C. C. Tsai et al. / Determination
the accelerator can be on the order of 3 X lOi crnd3. With such high gas density, the energy straggling effect can seriously degrade the source atomic ion yields. The ion sources being developed for neutral beam application in various fusion research centers can be operated at different source gas efficiencies, grid temperatures. beam energy, and gas line densities in the neutralizer. The variation of the gas line density in the accelerator of these ion sources can be as large as 2 50%. Even with the same ion source and the same species analyzer, such a large fluctuation in the gas density may lead to at least k 10% fluctuation of the source atomic ion yield. This may explain claims of > 90% source atomic ion yield. Moreover, different analyzers in different test facilities have different acceptance angles. For instance, for an analyzer with kO.2” acceptance angle, the atomic ions will be the dominant particles measured because the partial energy particles with a larger energy dispersion will not be measured as efficiently as the full energy component. Consequently, the measured source atomic ion yield will be unusually high, say > 95%. Thus, the source atomic ion yield should be determined under standardized beam conditions by a few reliable analyzers.
Acknowledgements The authors would like to thank their colleagues for their assistance in this work: G.C. Barber, N.S. Ponte, D.O. Sparks, and R.E. Wright for electrical and electronics support; C.W. Blue, R.R. Feezell, S.C. Forrester, and J.A. Moeller for the test stand maintenance; D.E. Schechter and F. Sluss for the preparation of ion sources; and W.K. Dagenhart, W.L. Gardner, G.G. Kelley. M.M. Menon, P.M. Ryan, and J.H. Whealton for the fruitful discussions and the participation in experiments.
of species yield of ion sources
References [II J.F. Clarke, Proc. IEEE 69 (1981) 867. VI J. Sheffield, Proc. IEEE 69 (1981) 885. [31 T.C. Simonen, Proc. IEEE 69 (1981) 935. I41 M.M. Menon, Proc. IEEE 69 (1981) 1012. Berkeley Laboratory I51 W.B. Kunkel. Lawrence
Report #LBL-10447 (1980). [61 H. Eubank et al., Phys. Rev. Letters 43 (1979) 270. I71 Proc. 3rd IAEA Neutral Beam Workshop, to be published. PI INTOR Group, in: Rep. Intern. Tokamak Workshop. IAEA, Vienna, 1980. [91 D.H. Metzler and L.D. Stewart. in: 9th Symp. Eng. Prob. Fusion Research (1981) p. 748. [lOI W.A. Houlberg et al.. in: 9th Symp. Prob. Fusion Research (1981) p. 688. IIll L.M. Hively et al., in: 9th Symp. Eng. Prob. Fusion Research (1981) p. 715. I121 D.H. Metzler, in: 9th Symp. Eng. Prob. Fusion Research (1981) p. 1286. [I31 D.E. Post et al., Princeton Plasma Physics Laboratory Report MATT- 1262 ( 1976). 1141 D.L. Jassby. Nucl. Fusion 17 (1977) 309. [I51 G. Porter, private communication. [161 C.C. Tsai, in: Proc. 3rd IAEA Neutral Beam Workshop, to be published. iI71 Y. Ohara et al.. J. Appl. Phys. 51 (1980) 3614. [I81 H.H. Fleischmann. C.F. Barnett and J.A. Ray, Phys. Rev. A10 (1974) 569. 1191 Y. Okumura et al., Rev. Sci. Instrum. 52 (1981) I. WI T. Obiki et al., Rev. Sci. Instrum. 52 (1981) 1445. Pll C.F. Burrell et al., Rev. Sci. Instrum. 51 (1980) 1451. P21 R.A. Langley and C.W. Magee, J. Nucl. Mater. 93/94 (1981) 390. [231 C.F. Barnett and J.A. Ray, Oak Ridge National Laboratory Report ORNL/TM-7656 (1980). Report [241 C.C. Tsai et al., Oak Ridge National Laboratory ORNL/TM-6360 (1978). [251 L.R. Grisham et al., Princeton Plasma Physics Laboratory Report PPPL-1484 (1978). WI M. Ulrickson et al.. Bull. Am. Phys. Sot. 23 (1978) 796. Re[271 C.F. Barnett et al.. Oak Ridge National Laboratory port ORNL-5206-07 (1977).