Creating negative ion beams from neutral gases using a negative hydrogen ion source

A typical method to produce negative ion beams uses alkali vapour as a medium for a double charge exchange to convert incident positive (1+) beams to negative (1-) beams. Alkali vapours pose a problem in ion production as they are flammable, explosive and cause vacuum surface contamination. Thus, it is advantageous to use non-metallic vapours for charge exchange as it will prevent hazards and the contamination of vacuum surfaces and targets in which the negative ion beams are incident upon. In this paper we will describe and demonstrate the process of creating negative ion beams by impinging a 15 keV to 30 keV beam from D-Pace's TRIUMF licenced H- volume-cusp ion source (1 mA to 15 mA) onto a volume of non-metallic neutral gas (X) resulting in a single or two-step charge exchange: H- + X → H + X- or H- + X → H + X + e → H + X-. The newly created X- ion beams will be accelerated by a (1 to 20) kV electrostatic accelerator and passed through a mass spectrometer system to separate the primary H- beam from the X- beam. The two gases studied are He and H2 and we will present the magnitude of the resultant beam currents after being separated from the incident H- beam by the mass spectrometer system.


Introduction
Negative ion beams are valuable for applications where tandem accelerators are used for ion injection [1], such as university research centers in the area of surface analysis using RBS (Rutherford Backscattering Spectrometry) and PIXE (Particle Induced X-Ray Emission) [2] and for high energy, light ion implantation in semiconductor devices [3].A typical method for negative ion production uses a charge exchange method where positive (1+) ions are incident upon a vacuum region of alkali or other metallic vapour at an energy of a few tens of keV [1], such that a double charge exchange occurs to produce negative (1-) ions.This technique results in alkali or other metal contamination of vacuum surfaces and is difficult to maintain, and, with regards to ion implantation, the metallic vapour can contaminate the silicon wafers being processed.Negative helium ions may be useful in the semiconductor industry for light ion implantation and are typically formed using the alkali vapour charge exchange method described above [4].Injecting helium into D-Pace's TRIUMF volume-cusp ion source to produce He − would be useful for the semiconductor industry as it would not require metallic vapours.However, this source does not produce He − , as He − is metastable and only has lifetimes of up to ∼350 μs [4], thus, it is unable to escape the plasma and reach the detector.Nevertheless, it is possible that the H − volume-cusp source could be utilized to produce He − or other negative ions from neutral vapour by the following charge exchange: Y − + X → Y + X − , where Y represents the incident beam and X represents the neutral gas target.This would be useful for semiconductor production as it eliminates the need for metallic vapours.This charge exchange process follows the example set by Doupé and Litherland [5] where a Cu − (gas) beam (Y) is incident on neutral ICl (gas) (X) for a single-step (near-resonant) charge exchange: Cu − + ICl → Cu + ICl − * → Cu + I + Cl − , where the newly created Cl − along with any remaining Cu − are accelerated with an electrostatic accelerator.Near-resonant charge exchange occurs when the difference in electron affinities (  ) between the colliding species is within 1 eV [6].In our case, the incident negative ion beam is H − (Y) and the target neutral vapour is He or H 2 (X). Figure 1 displays this charge exchange process (H − + X → H + X − ) throughout the beamline of the apparatus used for this experiment with example beam energies for the H − and X − beams.The test setup is described in more detail in section 2.

Experimental setup & method
D-Pace's TRIUMF licenced H − ion source [7] is leveraged to provide the incident beam into a vacuum box containing an electrostatic accelerator in which the neutral gases are injected.The incident beam energy and current can be adjusted between 15 keV and 30 keV and 1 mA and 15 mA respectively.The gas flow rate (measured in standard cubic centimeters per minute (sccm)) into the accelerator unit is adjusted using a mass flow controller to alter the pressure within the system and the potential of the accelerator is determined by a high voltage lead connection to the first isolated electrode, which can be swept between −1 kV to −20 kV, and is shown in figure 2.
The first four electrodes in the accelerator are isolated using high-density polyethylene (HDPE) rings which create a small, semi-contained region for higher neutral gas pressure at the upstream side of the accelerator.The remaining electrodes are isolated using ceramic standoffs which allows the gas to disperse downstream of the gas injection point so that newly created negative ions travel through a more evacuated beam path.The potential between all of the electrodes is determined by a series of resistors to divide the voltage applied to the first electrode throughout the accelerator.The first electrode is negatively biased at a potential   so that the incident H − beam at a potential   − is slowed to a potential   (eq.(2.1)).The H − beam that does not collide with the inert gas is re-accelerated to   − as it passes through the accelerator and the newly created ions are accelerated to the potential   .
It is anticipated that the majority of the collisions between the incident H − beam and the target gas will occur in the first few gaps of the accelerator.This region has a higher background gas density due to the semi-containment by the Teflon ring insulators and small electrode apertures.The electric field in this region is small in order to reduce the energy spread of newly formed X − ions.Most of -2 -  the acceleration will occur downstream of the target injection point.A circuit schematic and visual representation of the potential versus position in the accelerator can be found in figure 3.
The incident beam is collimated by a water cooled copper plate with a 5 mm aperture.The system is evacuated using three turbomolecular pumps, two on the TRIUMF licenced ion source and the third downstream of the accelerator unit, to ensure proper pumping of the gas injected.Pressures in the range of 10 −7 Torr are achieved when the H − beam is not being produced.Downstream of the vacuum box containing the accelerator unit is a 1:500 resolution mass spectrometer system (meaning that two ions with the same mass and energies  0 and (1 + 1/500) 0 can be resolved) with a 90 • bending electromagnet and two slits.The magnetic field strength can be adjusted using a variable current power supply to separate and detect ion beams of different masses and energies.After the mass spectrometer system, a 600 W capacity fixed Faraday cup is used to collect the separated ion beams and measure their corresponding beam current.The collected beam current is read using a digital picoammeter (Keithley 6485) with a resolution of 0.1 nA.A top-view of the full beamline for the experimental setup and ion species path is displayed in figure 1.
The incident H − beam of mass   − at an energy   − of 30 keV is detectable at the Faraday cup when 8.3 A of current is applied to the mass spectrometer dipole magnet, this corresponds to a magnetic field strength of 0.13 T. The current   at which negative ions with different mass   and energy   are detected is deduced from the relation between centripetal force and the Lorentz force law.
2) gives the maximum current at which newly created ions (X) are detectable by the Faraday cup.Ions created further downstream within the accelerator unit would not be exposed to the full accelerating potential and would be detected at a lower magnet current.

H − incident on He
The primary H − beam (  = 0.75 eV [8]) from the D-Pace TRIUMF licenced ion source was incident upon the accelerating unit/gas cell in order to carry out the charge exchange method outlined in section 1.The energy of the incident H − beam was varied for multiple experiments from 15 keV to 30 keV in 1 keV intervals, as this was the operating region for producing a well-focused beam.The incident beam current was also varied from 1 mA to 12 mA.The accelerator potential was swept between −1 kV and its limit of −20 kV which was expected to produce a negative helium ion beam at an energy corresponding to the respective potentials.Varying the accelerator potential and incident beam energy also produced collisions of varying energies which could result in an optimal cross-section for charge exchange.The flow of He (  = 0.08 eV [4]) into the gas cell was also adjusted in an attempt to find the optimal vessel pressure to produce a secondary negative ion beam (He − ).For a 30 keV incident H − beam with an accelerator potential of −18 kV it was expected to detect a peak of H − beam current (1 amu) from the source at a mass spectrometer current of 8.3 A and a secondary beam current peak of He − (4 amu) at a spectrometer current up to 12.85 A (calculated by eq.(2.2)).It would also be possible to detect beam current from an He − beam at a lower spectrometer current if the He − ions were created further downstream in the accelerating unit.No He − beams -4 -were detected on the picoammeter with a resolution of 0.1 nA for any of the experimental parameters outlined above.One of the null test results is displayed in figure 4 showing the expected region for He − detection with respect to the incident H − beam.It should be noted that there is an inconsistent measurement of positive current that is observed after the incident beam peak.The reason for this positive current measurement is unknown but could be due to instrument error.Thus, it is more likely to observe the He − beam at a higher accelerator potential corresponding to an He − beam energy that would be detected above 11 A on the mass spectrometer.It should also be noted that because the magnitude of the incident H − beam is not measurable on the picoammeter, the plots in figure 4 are from two different measurements with the same experimental parameters.
These experiments were performed based on the possibility of a near-resonant charge transfer between H − and He (Δ  < 1 eV) since He − has important industrial applications.After the experiments produced no measurable signal of He − , further literature research showed that He − can only be produced from an electronic transition from the excited neutral triplet state (1s2s) 3 S [9].The target helium is likely to only occupy the ground state (1s 2 ) 1 S which could cause no He − formation which would be measurable at a resolution of 0.1 nA for this experimental setup.

H − incident on H 2
An incident H − beam from the D-Pace TRIUMF licenced ion source was impinged upon inert H 2 with the expectation of producing a secondary beam of H − 2 through the method described by Doupé and Litherland.H 2 (  = 2 eV [10]) is on the threshold of producing near-resonant charge transfer.It is also known to be a metastable ion which could result in electron detachment before measurement [10].The expected charge transfer reaction is shown in eq.(3.1).

JINST 19 C05053
A beam of H − 2 was not measured, likely due to electron detachment before measurement or possible off-resonance from the incident H − beam.For the observation of a secondary H − beam, the expectation was that the (15 to 30) keV beam would excite the inert gas (H 2 ) in a negatively ionized state which could decay into a negative H − beam.The chemical process outlined in eq.(3.2) was expected to produce H − at energies governed by the accelerator potential.
It is important to note that this section of the experiment is performed to test this method of charge transfer and explore the interaction between H − and H 2 .The resulting secondary beam does not necessarily have practical applications as it would be an H − beam of lower beam current which could be created more easily using the D-Pace TRIUMF licenced ion source.A newly produced H − ion by this method (eq.(3.2)) would only be distinguishable from the primary beam if it is re-accelerated to an energy differing from the incident beam energy.Initially, a 20 keV H − beam was used with a potential of −10 kV on the accelerating unit and a varying gas flow.It was expected that a secondary beam of H − ions would be detectable at a mass spectrometer current of 4.66 A or lower (by eq. ( 2.2)).It was found that 6 sccm was the optimal flow of H 2 to produce the largest secondary beam current of newly produced negative ions.The initial results showed that a secondary beam of negative ions was detectable at a spectrometer current greater than 4.66 A as displayed in figure 5(a).This result would not be possible for a secondary beam of H − created in the gas cell as it would need to be exposed to a potential greater than −10 kV.This result was only produced when gas was flowing into the gas cell and potential was applied to the accelerator unit as shown by the control plot lines in figure 5(a).A second experiment was performed with a lower accelerator potential of −7.3 kV at the same incident beam energy of 20 keV.For a resonant charge exchange, a secondary beam current peak would be detected at a lower spectrometer current than the experiment performed with −10 kV applied to the accelerator.As displayed in figure 5(b), this was not the case.Figure 5(b) also shows the location of the incident H − beam, but it should be noted that the full magnitude of the incident beam is not displayed as it is not within the range of the picoammeter.The shaded sections of figure 5 show the regions that the secondary H − beam would be detected if it were created by means of the charge exchange outlined in eq.(3.2).Further experiments were performed with varying accelerator potentials (−1 to −17) kV and incident beam energies (15 keV, 20 keV, 25 keV) and it was found that as the accelerator potential was increased, the magnetic field strength (spectrometer current) required to collect the negative beam peaks decreased, contrary to what was expected.A theory that was hypothesized to explain these unexpected results was that the primary H − beam was slowed to the potential   (eq.(2.1)) and some amount was neutralized upon collision with H 2 .This neutralized hydrogen beam would then pass through the beamline and collect a free electron downstream of the accelerator, negatively ionizing it, before reaching the spectrometer so that it is detected at the Faraday cup.The attachment of a free electron to a neutral hydrogen atom is not typically found to be a likely reaction, but Tata et al. suggest that it is a possible mechanism [11].The neutral hydrogen beam would pass through the gas cell without acceleration at the energy   and collect a free electron in between the end of the accelerator unit and the mass spectrometer, a distance of approximately 80 cm.Some of the beam will not be stripped at the first electrode and receive some acceleration before the -6 -  electron stripping from the target gas, this would explain the energy spread shown in figure 5.For this theory, a higher potential applied to the accelerating unit would produce a lower energy H − beam by means of the following charge exchange hypothesis which is consistent with the data collected.
The current applied to the mass spectrometer at which the negative beam peak maxima would occur for both charge exchange methods described (eq.(3.2) and eq.(3.3)), was calculated for different accelerator potentials by eq.(2.2).The actual spectrometer current at which these negative beam peak maxima were detected was measured for varying accelerator potential values at three different incident beam energies and the comparison is displayed in figure 6.
The 15 keV, 20 keV, and 25 keV incident H − beams show peaks at spectrometer currents of 5.8 A, 6.6 A, and 7.5 A respectively.Thus, all the spectrometer current values at which peaks occur that are displayed in figure 6 have a smaller energy than their respective incident beams when potential is applied to the accelerator unit.This is consistent with the theory that the beam is being slowed by the first electrode to a potential   .It should be noted that a secondary beam of H − was also produced in the same mass spectrometer current range when using helium and argon as a target, confirming that the peaks measured are not a product of the H 2 target.A secondary beam of H − created by the single-step (resonant) charge exchange, outlined by Doupé and Litherland, is likely not being observed due to the off resonance of H − impinged onto H 2 instead of H − incident on H.This off resonance collision would result in a much smaller cross-section that may not produce enough beam current to be detectable by a 0.1 nA resolution current meter.

keV incident beam
Accelerator potential (kV) Mass spectrometer current (A) (a) Spectrometer current at which beam peak maxima are measured for different accelerator potential values and the expected location of the maxima if charge exchange (3.2) is produced.

Conclusion
The goal of this study was to show a novel method for creating negative ion beams from neutral gases using a negative hydrogen ion source and to measure the magnitude of the resulting beam currents from the preliminary tests performed.In the case of H − + He → H + He − , a He − beam was not observed at the Faraday cup using a picoammeter with a resolution of 0.1 nA.The expected single-step charge exchange outlined by Doupé and Litherland was not observed for an incident H − beam on a target gas of H 2 , likely due to decay of H − 2 or off-resonance resulting in no charge exchange.The other single-step charge exchange process outlined in eq.(3.2) was also not observed, again likely due to no creation of H − 2 resulting in no emission of H − .However, the neutralization and electron recapture of the incident beam from collisions with the inert H 2 was recorded.The magnitude of the beam current of the slowed incident H − beam was measured for different incident beam energies (15 keV, 20 keV, and 25 keV) and accelerator potential values.Future experiments will further investigate the single-step (resonant) charge exchange method in an attempt to create negative ions from O 2 and CO 2 .Experiments using a helium target could be performed by exciting the gas target to enter the (1s2s) 3 S state which may allow for nearresonant charge transfer.Attempts to improve the resolution of the current meter used for measuring beam current on the Faraday cup will be made in order to measure smaller beam current values.Plans to shorten the beamline of the system are in place in order to avoid potential issues with beam divergence and electron self detachment which may be limiting negative ion detection opportunities.

Figure 1 .
Figure 1.Beam-line diagram of the Ion Source Test Facility (ISTF) apparatus in which the experiment is conducted (all units of measurement are in mm).
(a) Model of the cross-section of the accelerating unit labelling major components and length.(b)Accelerating unit with the beam entrance/exit locations specified as well as the gas cell and variable high voltage lead connection.

Figure 2 .
Figure 2. A diagram of the charge exchange gas cell/accelerating unit including both a cross-section model of the design and the constructed unit placed in the beamline of the apparatus.

Figure 3 .
Figure 3.A circuit schematic of the electrostatic accelerator design with the potential gradient plotted below to show the acceleration profile.The resistance values used in the diagram are calculated in order to allow for 20 kV to be applied to the accelerator from the DC power supply.The diagram also shows the path of the incident H − beam and the location of target gas injection (X 0 ) as well as the extraction of newly formed negative ions (X − ).The potential of the first electrode is labeled as   , the same notation as eq.(2.1).

Figure 4 .
Figure 4. 30 keV (V  − ) incident H − beam (1.5 mA) with 6 sccm of He flow into gas the cell and −18 kV applied to the accelerating unit.
(a) 0 kV and −10 kV (  ) applied to the accelerating unit with varying flow of H 2 .(b) −7.3 kV (  ) applied to the accelerating unit showing the beam current peaks of the incident H − beam and a secondary H − beam when 6 sccm of H 2 is injected into the accelerator.

Figure 5 .
Figure 5. 20 keV, 3.5 mA H − beam incident on the gas cell showing measured data and the area in which beam current peaks are expected.
Spectrometer current at which beam peak maxima are measured for different   values (eq.(2.1)) and the expected location of the maxima if charge exchange (3.3) is produced.

Figure 6 .
Figure 6.Expected (calculated) spectrometer current to find H − beam peaks compared to the measured current at which the peaks occur over a range of different slowed beam potential values (  ).  is determined by the accelerator potential and these measurements are recorded using 15 keV, 20 keV, and 25 keV H − beams of 3 mA with 6 sccm of H 2 flow.