- Email: [email protected]
The new LLNL AMS spectrometer
Nuclear Instruments North-Holland
and Methods
in Physics
Research
B52 (1990) 301-305
The new LLNL AMS spectrometer
301
*
J.R. Southon, M.W. Caffee, J.C. Davis, T.L. Moore, I.D. Proctor, B. Schumacher and J.S. Vogel University of California, Lawrence Livermore National Laboratory,
Livermore
CA 945.50, USA
The multi-user tandem laboratory at Lawrence Livermore is a new facility analysis techniques. The AMS spectrometer design aims and implementation planned improvements are discussed.
1. Introduction
dedicated to AMS and a variety of other ion-beam are presented here, and present performance and
ion microprobe studies and AMS. This paper describes
The multi-user tandem facility at Lawrence Livermore National Laboratory is a new laboratoty built around an upgraded FN accelerator and dedicated to a variety of ion-beam analysis techniques, including PIXE,
the design, implementation, and performance of the actual AMS spectrometer. The programs and the overall design of the laboratory are discussed elsewhere [l].
2. Beam optics design
* Work performed
under the auspices of the USDOE by the Lawrence Livermore National Laboratory under contract W-7405-Eng-48. Reference herein to any specific commercial products, processes or service by trade name, trademark, manufacturer or otherwise, does not necessarily constitute or imply endorsement, recommendation, or favoring by the US Government or the University of California.
2.1. Design aims The requirements of AMS were a major consideration in the laboratory design, and the spectrometer
Fig. 1. The overall layout of the laboratory 0168-583X/90/$03.50
0 1990 - Elsevier
Science Publishers
and the AMS spectrometer.
B.V. (North-Holland)
II. NEW Br FUTURE
FACILITIES
302
J.R. Southon et al. / The new LLNL AiUS spectrometer
(cm) (cm)
source
stripper canal
grid
dipole
dipole
Wlen filter
detectors
Fig. 2. Beam-optics traces for the spectrometer.Critical apertures are shown. A source emittance of 0.5 mm by 50 mrad was used, and sufficient scattering at the stripper to fill the high-energy tubes was assumed. Note the good transmission predicted by this simulation.
includes relatively few compromises other than those imposed by space limitations. As fig. 1. shows, the overall layout of the laboratory is effectively that of a dedicated AMS spectrometer with other ion sources and beam lines added. The spectrometer was designed using the beam optics code OF’TRYK [T. Greenway, personal communication] and optics traces are presented in fig. 2. An important design aim was to achieve high sample throughput through use of a high-intensity multisample ion source. Operation at high current raised the possibility of intensity-dependent beam losses in the system arising from space-charge effects near the source [2]. The system was therefore designed for the best possible beam transmission to minimize such effects. Fig. 2 shows the extent to which this was achieved. Other important considerations included ease of tuning and operation through provision of adequate beam diagnostics and corrective steerers, and computer control of the transport system to allow eventual unattended operation.
2.2. Ion source and injection The present AMS ion source is a Genus Model 846 sputter source, based on a design by Middleton [3] and equipped with a 60-sample changer. A detailed discussion of the source appears elsewhere [4]. Operating voltages are 8 kV on the source cathode plus 25-30 kV on the extraction, for a total injection energy of just 35 kV. The decision was taken not to place the source on a high-voltage deck, since the large calculated acceptance of the modified FN indicated that beam transmission for typical sputter source emittances would be excellent, and the design of the injection beam line is simplified. Beams from the source are focus& by an einzel lens on to the object slits of a 90° double-focussing Danfysik injection magnet (r = 50 cm, ME/Z* = 7.5). The magnet is provided with rotatable downstream pole tips so that a second ion-source leg opposite the present one can be implemented. The vacuum box is insulated to f 5 kV. By means of high-voltage switches connecting
J. R. Southon et al. / The new LLNL A MS spectrometer the box to dc power supplies or to ground, different isotopes are switched into the accelerator under control of the data acquisition system. A large magnet gap of 5 cm was specified to ensure good transmission, and the accel-deccel gaps placed near the magnet object and image positions were also made large (10 cm diameter) to increase the lens focal lengths and thus minimize differential focussing for different bias voltages. The present ion-source line is a temporary setup which has remained in place far longer than anticipated, but will be modified shortly. The line is not properly matched to the downstream optics and beam must be severely scraped near the source to reduce the divergence from 40 mrad down to about 20 mrad, to avoid losses in the magnet. Nevertheless, injected “C- currents of up to 0.5 PA have been attained, and it is already clear that mass resolutions well in excess of M/AM = 100 will be available. When the system is properly tuned, we see no visible change in the beam shape or transmission for different injector bias voltages. A differential vertical shift of l-2 mm at the downstream of the magnet has persisted in spite of numerous attempts to identify the cause, but the remainder of the system has a sufficiently large acceptance that this has little effect. 2.3. Accelerator The FN accelerator was obtained second-hand from the University of Washington, Seattle, and has been substantially upgraded. Dowlish titanium spiral inclined-field accelerator tubes and a Pelletron conversion were installed to increase the beam transmission and the terminal-voltage range, and to improve energy stability and reliability. Although the entrance grid of the #I Dowlish tube intercepts 10% of the injected beam, this loss is insensitive to changes in the beam size or position and so can be tolerated. As fig. 2 shows, the beam transmission with these tubes is otherwise excellent. Other improvements to the FN included installation of a large (1 cm) diameter stripper canal and a turbomolecular terminal pump, for increased transmission with gas or foil strippers. Systems in the terminal are presently controlled via plastic rods, but an infrared light link control system will be installed shortly. The terminal-voltage stabilization system currently relies on generating voltmeter and capacitive pickoff inputs, and voltage stabilities of about 1 part in lo4 are achieved. A slit-stabilization system is under construction. The accelerator has generally performed well. showing the expected good stability and beam transmission. Operating voltages were limited to 5.5 MV by the use of N/CO, insulating gas until the gas handling system had been thoroughly exercised and made leak-tight. The gas was recently replaced with SF, and the FN has been conditioned to 7.5 MV. Some initial problems with
303
failures of column grading resistors and tracking of control rods were encountered, and conditioning took noticeably longer than with aluminum accelerator tubes [J.W. Stark, personal communication]. However, the FN now operates uneventfully at 7 MV, so that optimum running conditions for the C4+ charge state are attainable, and we anticipate that further voltage increases to lo-11 MV will be routine. 2.5. High energy spectrometer Scattering in the stripper in the accelerator terminal inevitably increases the emittance of the transmitted beam sufficiently to cause some losses in the high-energy accelerator tubes, particularly for heavy isotopes and low energies [5]. While we wanted the best possible transmission, we had neither the finances nor the foolhardiness to experiment with nonstandard large-diameter tubes. The high-energy spectrometer was therefore designed to pass the maximum beam divergence that could emerge from the Dowlish tubes to the AMS detectors without loss. This proved unexpectedly easy to accomplish using surplus large quadrupoles and analyzing magnets from LLNL and from the HEPL laboratory at Stanford. A design consisting of two identical 90° magnets and a Wien filter was chosen for its good optical properties and compact layout. Beams from the FN are focussed by a 10 cm diameter magnetic-quadrupole triplet to the object point for the first analyzing magnet. The choice of magnetic quadrupoles is a compromise, since correct focussing is achieved for just one isotope at a time. This decision was dictated by cost. the availability of a suitable triplet, and the difficulties of building large electrostatic quadrupoles of sufficient focussing strength to handle the high-energy H, D and T beams required by other users. The triplet misfocussing is accommodated by opening the object slits and was taken into account in the positioning of the Faraday cups which detect analyzed stable isotope beams. Since the analysis magnet gap is large, beam losses are avoided. The two identical ex-HEPL dipoles (single-focussing, r = 139 cm, gap = 6.4 cm, ME/Z2 = 150) together form a first-order achromat, leading to a small beam waist at the start of the final beam-line leg and reducing any jitter from energy shifts. The momentum dispersion at the image slits of the first magnet is about 1 in 800. Stable beams are detected in Faraday cups in a large vacuum chamber downstream of the magnet. and these cups are equipped with internal slits for beam-position monitoring and (eventually) for terminal-voltage stabilization [6]. The pole width is sufficiently wide that masses 12-14 could be accommodated without changing the magnet field, but this would require a new vacuum tank, and only 13C and 14C are accelerated at present. 11. NEW & FUTURE FACILITIES
304
J. R. Southon et al. / The new LLNL A MS spectrometer
The final beam-line leg contains a second magnetic triplet, a Wien filter and the AMS detectors. The Wien filter (length = I m, gap = 5 cm, 3 kG, f60 kV) was built by Danfysik and is a scaled-up version of one used previously by the Simon Fraser AMS group [D.E. Nelson personal communication], with a velocity resolution Au/o for 35 MeV C4+ of about 1%. The filter was preferred over electrostatic deflectors for the versatility provided by its variable dispersion and for the ease of alignment arising from the straight-line beam path. The optical magnification of the final leg was detiberately made large to provide a small final beam divergence. Longitudinal detector positions are less critical and time-of-flight detectors of modes! length can be implemented without using refocussing quadrupoles. Ample space is available to extend the line if long flight paths and refocussing prove necessary.
3. Data acquisition Particle detection is currently by means of a multianode transverse gas ionization detector. A longitudinal gas ionization detector for Be and two carbon foil-channel plate time-of-flight detectors for heavy isotopes are currently under construction. Data acquisition is based on NIM and CAMAC electronics feeding CAMAC ADC’s and scalers, with HP9000 workstations running under UNIX and in-house acquisition software written in C.
5. Ehrthquake The magnitude 7.1 Loma Prieta earthquake of Detober 17 1989 was felt in Livermore and seriously affected the spectrometer. The horizontal motion during the quake was perpendicular to the axis of the FN, and beam transmission dropped almost to zero. Alignment checks soon revealed a sideways displacement of almost 2 cm at the terminal, with no direct path for light from one end of the accelerator to the other: the accelerator column was kinked. After the terminal was jacked back into position and the FN was repressurized, the terminal again moved sideways, to about half the initial displacement. A second treatment reduced this pressurization displacement to about 2 mm. This is still of concern, since it indicates that there is residual transverse stress in the column, but the displacement is sufficiently small that the transmission is back to normal. An electrical short to ground through a protruding bolt in a cooling manifold of one of the 90° magnet coils was probably also caused by the earthquake: we suspect that the coil shifted slightly. This short caused serious problems of instability, both in the magnet and elsewhere, as the shorted magnet current caused massive ground loop problems and many failures in the control system electronics, and took several frustrating weeks to track down.
6. Results 4. Control system The accelerator and beam-line elements are computer-controlled through CAMAC and HP9000 workstations, using a control system developed at LLNL and the CEBAF laboratory 1781. Local computers each coruscate via GPIB with a single CAMAC crate which controls a cluster of beam-line elements. A supervisory compuIel. -i iue operator console scans the local computers via a LAN, with a systemwide data update rate of about 10 Hz. The supervisor also controls the main graphics display, flags errors, and receives operator input through the keyboard or via nine reassignable knobs. All definitions of signal connections and control algorithms are set up by manipulating icons with a graphics editor, so that control functions can be changed without writing new code. The system controls the entire spectrometer, with the exceptions of the source sample changer (see ref. [4]), and the injection bounce timing which is driven by the data acquisition system. Some failures of CAMAC ADC’s and DAC’s are still encountered, but the computers themselves have proven robust.
The transmission through the accelerator and the high-energy spectrometer are close to the theoretical values, and tuning of the beam transport system and scaling from one isotope to another have proven to be straightforward. Analyzed 13C4+ currents and modem 14C count rates of up to 1 gA and 130 s-‘, respectively, have been measured. Although our first AMS measurements were made some time ago, the spectrometer only really achieved semi-routine operation near the end of 1989, after the problems from the earthquake had been resolved. Accuracies of l-2% for radiocarbon dates on known-age samples and backgrounds equivalent to 4550 kyr were immediately achieved using prepared graphite samples [9]. We have seen no indication of beam intensity dependence in the “C/‘3C ratios for ionsource currents of up to 50 PA of ‘*C-. Preliminary measurements on “Be have also been carried out and an experimental beryllium program will shortly get under way, with work on 36Cl and 41Ca to follow. Though numerous minor problems remain to be ironed out and many planned improvements are still pending, the spectrometer is now a successful working instrument.
J. R. Southon et al. / The new LLNL
Acknowledgements Support from programmatic and Institutional Research and Development funding of the Lawrence Livermore Laboratory and from the Regents of the University of California is gratefully acknowledged.
References
111J.C. Davis et al., these Proceedings (AMS 5) Nucl. Instr. and Meth. B52 (1990) 269. I21 C. Bronk and R.E.M. Hedges, Proe. Workshop on Teehniques in AMS, Oxford, 1986, eds. R.E.M. Hedges and E.T. Hall. p. 60.
AM.7 spectrometer
305
[3] R. Middleton, ibid.. p. 82. [4] I.D. Proctor et al., these Proceedings (AMS 5) Nucl. Instr. and Meth. B52 (1990) 334. [S] T. Joy, Nucl. Instr. and Meth. 106 (1973) 237. [6] J.R. Southon, J.S. Vogel and D.E. Nelson, Proc. Workshop on Techniques in AMS, eds. R.E.M. Hedges and E.T. Hall, Oxford, 1986. p. 40. [7] T.L. Moore, Nucl. Instr. and Meth., B40/41 (1989) 984. [S] M.A. Roberts, R.S. Homady, T.L. Moore and J.C. Davis, Prcc. 2nd Int. Conf. on Ion Microprobe Technology. Melbourne, 1990, Nucl. Instr. and Meth., to be published. [9] J.S. Vogel, D.E. Nelson and J.R. Southon, Radiocarbon 29 (1987) 323.
II. NEW & FUTURE FACILITIES
and Methods
in Physics
Research
B52 (1990) 301-305
The new LLNL AMS spectrometer
301
*
J.R. Southon, M.W. Caffee, J.C. Davis, T.L. Moore, I.D. Proctor, B. Schumacher and J.S. Vogel University of California, Lawrence Livermore National Laboratory,
Livermore
CA 945.50, USA
The multi-user tandem laboratory at Lawrence Livermore is a new facility analysis techniques. The AMS spectrometer design aims and implementation planned improvements are discussed.
1. Introduction
dedicated to AMS and a variety of other ion-beam are presented here, and present performance and
ion microprobe studies and AMS. This paper describes
The multi-user tandem facility at Lawrence Livermore National Laboratory is a new laboratoty built around an upgraded FN accelerator and dedicated to a variety of ion-beam analysis techniques, including PIXE,
the design, implementation, and performance of the actual AMS spectrometer. The programs and the overall design of the laboratory are discussed elsewhere [l].
2. Beam optics design
* Work performed
under the auspices of the USDOE by the Lawrence Livermore National Laboratory under contract W-7405-Eng-48. Reference herein to any specific commercial products, processes or service by trade name, trademark, manufacturer or otherwise, does not necessarily constitute or imply endorsement, recommendation, or favoring by the US Government or the University of California.
2.1. Design aims The requirements of AMS were a major consideration in the laboratory design, and the spectrometer
Fig. 1. The overall layout of the laboratory 0168-583X/90/$03.50
0 1990 - Elsevier
Science Publishers
and the AMS spectrometer.
B.V. (North-Holland)
II. NEW Br FUTURE
FACILITIES
302
J.R. Southon et al. / The new LLNL AiUS spectrometer
(cm) (cm)
source
stripper canal
grid
dipole
dipole
Wlen filter
detectors
Fig. 2. Beam-optics traces for the spectrometer.Critical apertures are shown. A source emittance of 0.5 mm by 50 mrad was used, and sufficient scattering at the stripper to fill the high-energy tubes was assumed. Note the good transmission predicted by this simulation.
includes relatively few compromises other than those imposed by space limitations. As fig. 1. shows, the overall layout of the laboratory is effectively that of a dedicated AMS spectrometer with other ion sources and beam lines added. The spectrometer was designed using the beam optics code OF’TRYK [T. Greenway, personal communication] and optics traces are presented in fig. 2. An important design aim was to achieve high sample throughput through use of a high-intensity multisample ion source. Operation at high current raised the possibility of intensity-dependent beam losses in the system arising from space-charge effects near the source [2]. The system was therefore designed for the best possible beam transmission to minimize such effects. Fig. 2 shows the extent to which this was achieved. Other important considerations included ease of tuning and operation through provision of adequate beam diagnostics and corrective steerers, and computer control of the transport system to allow eventual unattended operation.
2.2. Ion source and injection The present AMS ion source is a Genus Model 846 sputter source, based on a design by Middleton [3] and equipped with a 60-sample changer. A detailed discussion of the source appears elsewhere [4]. Operating voltages are 8 kV on the source cathode plus 25-30 kV on the extraction, for a total injection energy of just 35 kV. The decision was taken not to place the source on a high-voltage deck, since the large calculated acceptance of the modified FN indicated that beam transmission for typical sputter source emittances would be excellent, and the design of the injection beam line is simplified. Beams from the source are focus& by an einzel lens on to the object slits of a 90° double-focussing Danfysik injection magnet (r = 50 cm, ME/Z* = 7.5). The magnet is provided with rotatable downstream pole tips so that a second ion-source leg opposite the present one can be implemented. The vacuum box is insulated to f 5 kV. By means of high-voltage switches connecting
J. R. Southon et al. / The new LLNL A MS spectrometer the box to dc power supplies or to ground, different isotopes are switched into the accelerator under control of the data acquisition system. A large magnet gap of 5 cm was specified to ensure good transmission, and the accel-deccel gaps placed near the magnet object and image positions were also made large (10 cm diameter) to increase the lens focal lengths and thus minimize differential focussing for different bias voltages. The present ion-source line is a temporary setup which has remained in place far longer than anticipated, but will be modified shortly. The line is not properly matched to the downstream optics and beam must be severely scraped near the source to reduce the divergence from 40 mrad down to about 20 mrad, to avoid losses in the magnet. Nevertheless, injected “C- currents of up to 0.5 PA have been attained, and it is already clear that mass resolutions well in excess of M/AM = 100 will be available. When the system is properly tuned, we see no visible change in the beam shape or transmission for different injector bias voltages. A differential vertical shift of l-2 mm at the downstream of the magnet has persisted in spite of numerous attempts to identify the cause, but the remainder of the system has a sufficiently large acceptance that this has little effect. 2.3. Accelerator The FN accelerator was obtained second-hand from the University of Washington, Seattle, and has been substantially upgraded. Dowlish titanium spiral inclined-field accelerator tubes and a Pelletron conversion were installed to increase the beam transmission and the terminal-voltage range, and to improve energy stability and reliability. Although the entrance grid of the #I Dowlish tube intercepts 10% of the injected beam, this loss is insensitive to changes in the beam size or position and so can be tolerated. As fig. 2 shows, the beam transmission with these tubes is otherwise excellent. Other improvements to the FN included installation of a large (1 cm) diameter stripper canal and a turbomolecular terminal pump, for increased transmission with gas or foil strippers. Systems in the terminal are presently controlled via plastic rods, but an infrared light link control system will be installed shortly. The terminal-voltage stabilization system currently relies on generating voltmeter and capacitive pickoff inputs, and voltage stabilities of about 1 part in lo4 are achieved. A slit-stabilization system is under construction. The accelerator has generally performed well. showing the expected good stability and beam transmission. Operating voltages were limited to 5.5 MV by the use of N/CO, insulating gas until the gas handling system had been thoroughly exercised and made leak-tight. The gas was recently replaced with SF, and the FN has been conditioned to 7.5 MV. Some initial problems with
303
failures of column grading resistors and tracking of control rods were encountered, and conditioning took noticeably longer than with aluminum accelerator tubes [J.W. Stark, personal communication]. However, the FN now operates uneventfully at 7 MV, so that optimum running conditions for the C4+ charge state are attainable, and we anticipate that further voltage increases to lo-11 MV will be routine. 2.5. High energy spectrometer Scattering in the stripper in the accelerator terminal inevitably increases the emittance of the transmitted beam sufficiently to cause some losses in the high-energy accelerator tubes, particularly for heavy isotopes and low energies [5]. While we wanted the best possible transmission, we had neither the finances nor the foolhardiness to experiment with nonstandard large-diameter tubes. The high-energy spectrometer was therefore designed to pass the maximum beam divergence that could emerge from the Dowlish tubes to the AMS detectors without loss. This proved unexpectedly easy to accomplish using surplus large quadrupoles and analyzing magnets from LLNL and from the HEPL laboratory at Stanford. A design consisting of two identical 90° magnets and a Wien filter was chosen for its good optical properties and compact layout. Beams from the FN are focussed by a 10 cm diameter magnetic-quadrupole triplet to the object point for the first analyzing magnet. The choice of magnetic quadrupoles is a compromise, since correct focussing is achieved for just one isotope at a time. This decision was dictated by cost. the availability of a suitable triplet, and the difficulties of building large electrostatic quadrupoles of sufficient focussing strength to handle the high-energy H, D and T beams required by other users. The triplet misfocussing is accommodated by opening the object slits and was taken into account in the positioning of the Faraday cups which detect analyzed stable isotope beams. Since the analysis magnet gap is large, beam losses are avoided. The two identical ex-HEPL dipoles (single-focussing, r = 139 cm, gap = 6.4 cm, ME/Z2 = 150) together form a first-order achromat, leading to a small beam waist at the start of the final beam-line leg and reducing any jitter from energy shifts. The momentum dispersion at the image slits of the first magnet is about 1 in 800. Stable beams are detected in Faraday cups in a large vacuum chamber downstream of the magnet. and these cups are equipped with internal slits for beam-position monitoring and (eventually) for terminal-voltage stabilization [6]. The pole width is sufficiently wide that masses 12-14 could be accommodated without changing the magnet field, but this would require a new vacuum tank, and only 13C and 14C are accelerated at present. 11. NEW & FUTURE FACILITIES
304
J. R. Southon et al. / The new LLNL A MS spectrometer
The final beam-line leg contains a second magnetic triplet, a Wien filter and the AMS detectors. The Wien filter (length = I m, gap = 5 cm, 3 kG, f60 kV) was built by Danfysik and is a scaled-up version of one used previously by the Simon Fraser AMS group [D.E. Nelson personal communication], with a velocity resolution Au/o for 35 MeV C4+ of about 1%. The filter was preferred over electrostatic deflectors for the versatility provided by its variable dispersion and for the ease of alignment arising from the straight-line beam path. The optical magnification of the final leg was detiberately made large to provide a small final beam divergence. Longitudinal detector positions are less critical and time-of-flight detectors of modes! length can be implemented without using refocussing quadrupoles. Ample space is available to extend the line if long flight paths and refocussing prove necessary.
3. Data acquisition Particle detection is currently by means of a multianode transverse gas ionization detector. A longitudinal gas ionization detector for Be and two carbon foil-channel plate time-of-flight detectors for heavy isotopes are currently under construction. Data acquisition is based on NIM and CAMAC electronics feeding CAMAC ADC’s and scalers, with HP9000 workstations running under UNIX and in-house acquisition software written in C.
5. Ehrthquake The magnitude 7.1 Loma Prieta earthquake of Detober 17 1989 was felt in Livermore and seriously affected the spectrometer. The horizontal motion during the quake was perpendicular to the axis of the FN, and beam transmission dropped almost to zero. Alignment checks soon revealed a sideways displacement of almost 2 cm at the terminal, with no direct path for light from one end of the accelerator to the other: the accelerator column was kinked. After the terminal was jacked back into position and the FN was repressurized, the terminal again moved sideways, to about half the initial displacement. A second treatment reduced this pressurization displacement to about 2 mm. This is still of concern, since it indicates that there is residual transverse stress in the column, but the displacement is sufficiently small that the transmission is back to normal. An electrical short to ground through a protruding bolt in a cooling manifold of one of the 90° magnet coils was probably also caused by the earthquake: we suspect that the coil shifted slightly. This short caused serious problems of instability, both in the magnet and elsewhere, as the shorted magnet current caused massive ground loop problems and many failures in the control system electronics, and took several frustrating weeks to track down.
6. Results 4. Control system The accelerator and beam-line elements are computer-controlled through CAMAC and HP9000 workstations, using a control system developed at LLNL and the CEBAF laboratory 1781. Local computers each coruscate via GPIB with a single CAMAC crate which controls a cluster of beam-line elements. A supervisory compuIel. -i iue operator console scans the local computers via a LAN, with a systemwide data update rate of about 10 Hz. The supervisor also controls the main graphics display, flags errors, and receives operator input through the keyboard or via nine reassignable knobs. All definitions of signal connections and control algorithms are set up by manipulating icons with a graphics editor, so that control functions can be changed without writing new code. The system controls the entire spectrometer, with the exceptions of the source sample changer (see ref. [4]), and the injection bounce timing which is driven by the data acquisition system. Some failures of CAMAC ADC’s and DAC’s are still encountered, but the computers themselves have proven robust.
The transmission through the accelerator and the high-energy spectrometer are close to the theoretical values, and tuning of the beam transport system and scaling from one isotope to another have proven to be straightforward. Analyzed 13C4+ currents and modem 14C count rates of up to 1 gA and 130 s-‘, respectively, have been measured. Although our first AMS measurements were made some time ago, the spectrometer only really achieved semi-routine operation near the end of 1989, after the problems from the earthquake had been resolved. Accuracies of l-2% for radiocarbon dates on known-age samples and backgrounds equivalent to 4550 kyr were immediately achieved using prepared graphite samples [9]. We have seen no indication of beam intensity dependence in the “C/‘3C ratios for ionsource currents of up to 50 PA of ‘*C-. Preliminary measurements on “Be have also been carried out and an experimental beryllium program will shortly get under way, with work on 36Cl and 41Ca to follow. Though numerous minor problems remain to be ironed out and many planned improvements are still pending, the spectrometer is now a successful working instrument.
J. R. Southon et al. / The new LLNL
Acknowledgements Support from programmatic and Institutional Research and Development funding of the Lawrence Livermore Laboratory and from the Regents of the University of California is gratefully acknowledged.
References
111J.C. Davis et al., these Proceedings (AMS 5) Nucl. Instr. and Meth. B52 (1990) 269. I21 C. Bronk and R.E.M. Hedges, Proe. Workshop on Teehniques in AMS, Oxford, 1986, eds. R.E.M. Hedges and E.T. Hall. p. 60.
AM.7 spectrometer
305
[3] R. Middleton, ibid.. p. 82. [4] I.D. Proctor et al., these Proceedings (AMS 5) Nucl. Instr. and Meth. B52 (1990) 334. [S] T. Joy, Nucl. Instr. and Meth. 106 (1973) 237. [6] J.R. Southon, J.S. Vogel and D.E. Nelson, Proc. Workshop on Techniques in AMS, eds. R.E.M. Hedges and E.T. Hall, Oxford, 1986. p. 40. [7] T.L. Moore, Nucl. Instr. and Meth., B40/41 (1989) 984. [S] M.A. Roberts, R.S. Homady, T.L. Moore and J.C. Davis, Prcc. 2nd Int. Conf. on Ion Microprobe Technology. Melbourne, 1990, Nucl. Instr. and Meth., to be published. [9] J.S. Vogel, D.E. Nelson and J.R. Southon, Radiocarbon 29 (1987) 323.
II. NEW & FUTURE FACILITIES