Installation and first operation of the negative ion optimization experiment

Fusion Engineering and Design 96–97 (2015) 249–252

Contents lists available at ScienceDirect

Fusion Engineering and Design journal homepage: www.elsevier.com/locate/fusengdes

Installation and first operation of the negative ion optimization experiment Michela De Muri a,b,∗ , Marco Cavenago a , Gianluigi Serianni b , Pierluigi Veltri b , Marco Bigi b , Roberto Pasqualotto b , Marco Barbisan b , Mauro Recchia b , Barbara Zaniol b , Timour Kulevoy c , Sergey Petrenko c , Lucio Baseggio b , Vannino Cervaro b , Fabio Degli Agostini b , Luca Franchin b , Bruno Laterza b , Alessandro Minarello a , Federico Rossetto b , Manuele Sattin a , Simone Zucchetti b a b c

INFN-LNL, v.le dell’Università 2, I-35020 Legnaro, PD, Italy Consorzio RFX, CNR, ENEA, INFN, Università di Padova, A cciaierie Venete SpA – Corso Stati Uniti 4, 35127 Padova, Italy ITEP, B. Cheremushkinskaya 25, 117218 Moscow, Russia

h i g h l i g h t s • • • •

Negative ion sources are key components of the neutral beam injectors. The NIO1 experiment is a RF ion source, 60 kV–135 mA hydrogen negative ion beam. NIO1 can contribute to beam extraction and optics thanks to quick replacement and upgrading of parts. This work presents installation, status and first experiments results of NIO1.

a r t i c l e

i n f o

Article history: Received 26 September 2014 Received in revised form 22 June 2015 Accepted 23 June 2015 Available online 9 July 2015 Keywords: Negative ion source Neutral beam injection NIO1 installation NIO1 experiments

a b s t r a c t Negative ion sources are key components of the neutral beam injectors for thermonuclear fusion experiments. The NIO1 experiment is a radio frequency ion source generating a 60 kV–135 mA hydrogen negative ion beam. The beam is composed of nine beamlets over an area of about 40 × 40 mm2 . This experiment is jointly developed by Consorzio RFX and INFN-LNL, with the purpose of providing and optimizing a test ion source, capable of working in continuous mode and in conditions similar to those foreseen for the larger ion sources of the ITER neutral beam injectors. At present research and development activities on these ion sources still address several important issues related to beam extraction and optics optimization, to which the NIO1 test facility can contribute thanks to its modular design, which allows for quick replacement and upgrading of components. This contribution presents the installation phases, the status of the test facility and the results of the first experiments, which have demonstrated that the source can operate in continuous mode. © 2015 Consorzio RFX. Published by Elsevier B.V. All rights reserved.

1. Introduction Negative ion sources are key components of the neutral beam injectors for thermonuclear fusion machines. NIO1 (negative ion optimization-phase 1) [1] source is being developed in Padova, Italy, by Consorzio RFX and INFN-LNL. It is a flexible test facility that allows testing several configurations and magnet layout, with quick replacement and upgrading of components. The work is

∗ Corresponding author. E-mail address: [email protected] (M. De Muri). http://dx.doi.org/10.1016/j.fusengdes.2015.06.107 0920-3796/© 2015 Consorzio RFX. Published by Elsevier B.V. All rights reserved.

carried out in the framework of accompanying activities in support to the ITER Neutral Beam Test Facility [2] that is under construction in Padova (Italy). NIO1 will produce a 60 kV–135 mA hydrogen negative ion beam composed of 3 × 3 beamlets over an area of about 40 × 40 mm2 and the frequency of the RF source is 2 MHz. It relies on an inductively coupled plasma (ICP), with negative ion production enhanced by adding cesium vapour in the plasma. The installation of the service plants of NIO1 experiment is nearing completion and their commissioning has started. This contribution presents the latest activities on NIO1 and the status of the test facility, together with the first experiments. These preliminary experiments demonstrated that the source can operate reliably in continuous mode.

250

M. De Muri et al. / Fusion Engineering and Design 96–97 (2015) 249–252

Fig. 1. NIO1 experiment layout. Fig. 3. Connections diagram.

2. Status of the experiment At the end of 2013 the vessel support was assembled, fixed to the floor and leveled, and the vacuum vessel was installed on it. The source was assembled, aligned and mounted on the vessel. Fig. 1 shows the status of the NIO1 test facility. The alignment of the source grids, satisfying beam optics requirements, was performed using a specific tool, provided with alignment bars: the tool allows also rotation of the source for final positioning (Fig. 2). During the installation the yielding of the cantilevered source was measured and found negligible. However, the source was installed on a cradle, to guarantee the source alignment over time. The yielding of the cradle was verified: under the source weight of about 40 kg, a deformation of 0.1 mm was measured. The source was fixed to the vessel taking particular care to the parallelism of the mating flanges; then also the cradle was fixed to the source

and to the floor. The pumping system, realized with a pre-vacuum scroll pump and a turbo-molecular pump, was installed and a base pressure of 2 × 10−5 Pa was achieved. The box for the ion source insulator transformer [3] and the transformer were initially installed. Then the electric distribution board for the power supply of the experiment was installed and tested. A diagram of the high voltage connections and earthing is shown in Fig. 3: Connections diagram. The high voltage deck (HVD) was assembled and installed on site, equipped with: matching box (MB), RF generator (RFPS, 2.5 kW), EGPS (extraction grid power supply, 4 kW 10 kV dc), high current power supplies (polarization of bias plate with respect to plasma grid – BPP – 50 A–30 V, polarization of source with respect to plasma grid – BIAS – 76 A–20 V and magnetic filter field – PGFPS – 400 A–8 V), feedback control of gas injection system and acquisition board of source thermocouple signals. To design the radiation shield, the maximum acceleration voltage of 60 kV and 800 h/y expected operation time are considered. The acceleration of negative hydrogen ions also produces parasitic particles (electrons and positive ions), whose expected parameters are detailed in Table 1 [4]. All these particles are blocked by the 10 mm thick stainless steel vessel. The main source of radiation is due to bremsstrahlung (electron–wall interaction). The X-rays are shielded with a safety factor of 2 by 2.5 mm thick lead walls at 0.5 m from the source (the enclosed area is about 1.5 × 3.5 m2 ). The lead

Table 1 Particle producing radiation in the NIO source.

Fig. 2. Alignment of the source.

Accelerated particle

Max energy (keV)

Max current (mA)

Negative hydrogen ion Electrons Electrons Positive hydrogen ion

60 10 60 60

150 300 10 10

M. De Muri et al. / Fusion Engineering and Design 96–97 (2015) 249–252

251

Fig. 5. Cesium oven.

Fig. 4. Scheme of the NIO1 cooling system.

panels stiffness is increased by 3 and 1 mm thick aluminum sheets on either side of the lead panels. The whole shielding is composed of easily removable panels and allows entrance of cooling pipes and signal cables and has three access doors, which also guarantee multiple escape routes. The gas injection system is now installed for hydrogen gas, but a modification could be implemented to allow the injection of hydrogen–argon mixture to better estimate the electron temperature [5]. The inlet gas line is directly on the source and requires a ceramic insulator (Fig. 3). This insulator was tested in air at atmospheric pressure by gradually increasing the applied voltage up to 89 kV, when voltage discharge occurred. This insulator will then be used in hydrogen only at low power, while full power pulses will required an identical insulator connected in series, while the central point voltage will be defined by a voltage divider (not shown in the figure). The cooling system is a closed circuit (Fig. 4). The glycol water of the secondary circuit flows in a heat exchanger and cools the deionized water that flows in the experiment cooling circuits (primary circuit). The total required cooling power is estimated to be up to 20 kW. A variable speed pump provides a flow up to 2 m3 /h and up to 8 × 105 Pa pressure, which is adjustable and depends on load condition. A 50 L stainless steel accumulation tank, a 24 L expansion tank of and a 7 ␮m water filter are included. A three-ways valve can regulate the coolant temperature in the experiment circuits. The MAIN manifold is at ground potential and is installed outside of the shielding. It feeds six on board circuits: post acceleration grid, calorimeter, fast emittance scanner diagnostic, a separate manifold for the cooling circuits at −60 kV (source potential, A–F in Fig. 4) and two spare circuits (G–M in Fig. 4). The calorimeter has its own manifold that serves 13 cooling segments. To guarantee the electrical insulation, the piping from the MAIN manifold to the −60 kV manifold includes 7 m Rilsan® coils (one at inlet and one at outlet), equipped with measurement of the electrical current flowing in the deionized water. The cooling circuit will be re-filled with clean deionized water when the current exceeds a preset maximum limit (0.1 mA/coil corresponding to a negligible H2 production of 0.08 cm3 /h). The −60 kV manifold serves the circuits for: the source (divided into two blocks); the series of RF generator, RF coil and the

matching box so that a water leak in the coil can block its power supply; the series of bias plate and plasma grid circuits; the circuits for the extraction grid (EG) and one spare circuit. The EG at +9 kV from the source has three different cooling circuits which also include two Rilsan® coils, 1.7 m long. Data acquisition and interlock systems are under development, and some part related to safety is designed and constructed to carry out experiments. During daily operation, an earthing sequence is required to access the lead shield, after the operator has switched off all power supplies of the experiment. By pushing the “aperture request” button, the output of all power supplies (60 kV, 10 kV, 400 A, 50 A and 76 A, RF) will be disabled and a timer (10 s) will start. After this time all power supplies will have discharged through suitable parallel resistances, the red “high voltage on” sign will switch off and an orange “high voltage off” sign will light up. At this point a visual inspection of the zero voltage of both 60 and 10 kV power supplies will be performed, looking at voltage meters visible outside of the shielding. Then the operator manually closes the earthing switch, so that the mechanical closing of the contacts can be surely and easily verified; the orange light “high voltage off” will switch off and a green light “grounded” will light up. The earthing switch is then locked and only at this stage can the key be removed. To enter in the shielded area, the main door near the HVD area can be opened with the same key and the switch hook can be put in the safety position (in the HVD). A second door will be also opened as escape route. The safety interlock connection is realized as a single loop, including optical transmission between ground potential and −60 kV. To restart the experiment the power supplies have to be powered on, but an interlock protection system prevents this operation if one of the following conditions is verified: one of the shielding doors is open; the earthing switch is closed; the switch hook is not in the stowed position (off of the HVD). An emergency button is used to power off the power supplies of all the experiment. The operation with cesium is planned during this year [6]. The cesium oven (Fig. 5) has two parts that can be maintained at different temperatures: the oven reservoir (at temperature T1 ) and the thermally insulated pipe (at T2 ), connecting the ion source to the reservoir with a regulation valve (rated to 300 ◦ C). By keeping T2 greater than T1 cesium condensation in the pipe is avoided, so in principle the time integral of cesium delivered to the source depends only on T1 regulation. The pipe, the valve and the reservoir are enclosed in copper shells housing four heaters. In a separate test bed T1 and T2 were independently regulated and stabilized (160 and 180 ◦ C, respectively) within 1 ◦ C, which is well satisfying the requirements for operation. Remote control with RS232 links is under development.

252

M. De Muri et al. / Fusion Engineering and Design 96–97 (2015) 249–252

Fig. 6. Pressure calibration.

Fig. 8. Frequency dependence of photomultiplier (red), transmitted power (blue), ratio between reflected and foreword powers (green) and voltage to the RF coil (violet) (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.).

In another test with pressure fixed to 5.8 Pa, the generator was replaced by a broad band amplifier with a fixed drive voltage at 440 mV, and the frequency was scanned in the range 2060–2088 kHz range. Fig. 8 shows that there is a minimum in the fraction of the reflected power over the total RF power at 2074 kHz (RR/RF), moreover, the PMT signal strongly increases with frequency. 4. Conclusions and perspectives Fig. 7. Plasma luminosity as a function of filling pressure (RF amplifier with fixed input, output power rises with pressure from 128 to 224 W).

3. First experiments First experiments were performed with air as filling gas. Pressure is measured in two different points: one is in the source with a Pirani gauge that will be substituted by a baratron and the ITR90 (Pirani and ionization combined sensor) on the opposite side, at the end of the vessel. During the commissioning, the Pirani was disconnected to avoid damages, so a calibration between the two measurements was done in stationary conditions with the RF off and constant gas flow. At low pressure (2.8 × 10−3 –2.3 × 10−1 Pa) the relation is linear; at higher pressure (2.3 × 10−1 –4 Pa) it is logarithmic (Fig. 6). The 2500 W RF generator is connected to RF coil through a matching box based on two adjustable vacuum capacitor banks connected in step-up configuration [7]. This equipment was previously tested with a similar coil in another experiment where plasma shape was monitored by CCDs; the plasma ignition and its transition from capacitive to inductive coupling were observed under 200 W, depending on gas filling pressure, ranging from 0.3 to 10 Pa (air). Thanks to water cooling of all RF circuits, a very stable operation was observed at 2050 kHz with few percent RF reflection during the plasma-on phase. In NIO1 the RF generator is installed in the HVD just over the matching box with a 0.8 m long connection to the RF coil. The operation frequency was tuned to 2074 kHz, verified at low power level. Plasma ignition in air was observed at about 50 W power at relatively high pressure (6 Pa). At a convenient power level Pf = 200 W, plasma was maintained at p1 > 0.7 Pa and plasma luminosity, as observed by a PMT (photomultiplier), increased with p1 and saturated at 11 Pa (Fig. 7).

The NIO1 source was successfully operated for hours, without extraction voltage. The planning until the end of 2014 involves activities in three areas: completion of installation and commissioning of the plant units as well as operation at low RF power both in air (negative oxygen ions extracted) and in hydrogen.

Acknowledgments This work has been partially carried out within the framework of the EUROfusion Consortium and has received funding from the European Union’s Horizon 2020 Research and Innovation Programme under grant agreement number 633053. The views and opinions expressed herein do not necessarily reflect those of the European Commission. Precious help by M. Cazzador and A. Mimo during the experiments is thankfully acknowledged.

References [1] M. Cavenago, et al., Development of small multiaperture negative beam sources and relater simulation tools, AIP Conf. Proc. 1097 (2009) 149. [2] P. Sonato, et al., Status of PRIMA, the test facility for ITER neutral beam injection, AIP Conf. Proc. 1515 (2013) 549. [3] M. Recchia, et al., Conceptual design and circuit analyses for the power supplies of the NIO1 experiment, Fusion Eng. Des. 86 (2011) 1545–1548. [4] A. Coniglio, Relazione tecnica di radioprotezione, unpublished. [5] U. Fantz, et al., Spectroscopy – a powerful diagnostic tool in source development, Nucl. Fusion 46 (2006) S297–S306. [6] M. Cavenago, et al., Status of NIO1 construction, AIP Conf. Proc. 1390 (2011) 640. [7] M. Cavenago, et al., Design of a versatile multiaperture negative ion source, Rev. Sci. Instrum. 81 (2010) 02A713.