The interest in high intensity ion source is increasing rapidly with the development of the worldwide high intensity proton cyclotron and D⁺ linac. For instance, it is an effective way to accelerate in place of H⁺ in high current cyclotrons to reduce the space charge effect during beam transportation.^[1] Moreover, deuteron accelerators are widely used in neutron source such as Sprial 2,^[2] Soreq applied research accelerator facility (SARAF),^[3] etc. As neutron radiation is induced by D(d, n) reaction along the beam line,^[4] it is proposed to use as a pilot beam in place of D⁺ to be accelerated during the commissioning phase in order to reduce the radiation problem and to guarantee the safety of the facility and staff.

At Peking University (PKU), all permanent magnet 2.45-GHz ECR ion sources for singly charged and lowly charged ion generation have been developed in the past decades.^[5,6] Recently, 40-mA and 20-mA ion beams were successfully generated by PMECR II — an SS ECR ion source in the plasma chambers with different diameters. Experimental results show that a plasma chamber with a diameter of 64 mm (no liner) and another one with a diameter of 30 mm (stainless steel (SS) liner) seem to be advantageous for molecular ions generation over those with other dimensions.^[7]

Recently, special attention was paid to the physics of hydrogen generation processes within a plasma chamber with different liner materials. Waldmann and Ludewigt mounted different liners on the front and end plates of the plasma chamber in a 2.45-GHz microwave ion source.^[8] It indicated that with BN or Al₂O₃, a higher proton fraction could be achieved. But with SS, the ratio of protons was relatively low. Jung et al. also investigated the suitable material of the plasma electrode for proton generation in a helicon ion source, and alumina that had a very low recombination possibility was concluded to be very effective.^[9] However, there is no particular investigation concerning the dependence of hydrogen molecular ion on the plasma chamber surface. Here, SS, Ta, quartz, and Al are selected in our study. Experiments indicate that SS and Ta are much more suitable for the generation of than others. A 42.3-mA ion beam is obtained with a fraction of 52.9%.

An obvious enhancement of fraction is witnessed after the SS plasma chamber operating with oxygen discharge for several hours during our experiments. This may indicate something changed on the SS surface after O₂ discharge. Possibly, except for the category of materials, the surface morphology also has an influence on the generation of molecular ions. Comparison experiment between a native SS surface and as SS surface after O₂ discharge are also carried out.

The rest of this paper is organized as follows. In Section 2, some physical analyses of hydrogen plasma recombination processes are given. In Section 3, the function of material on the generation is displayed, showing the O₂ influence on the ion fractions. Moreover, SS samples are placed on the chamber surface to simulate the bombardments of O⁺ and H⁺ on the wall. The analysis results of atomic force microscopy (AFM) and x-ray photoelectron spectroscopy (XPS) of the chamber surface are presented. More details are reported in this paper.

In hydrogen plasmas, ions are produced inevitably by direct ionization of hydrogen molecular H₂:

Simultaneously, a proton is generated mainly by a two-step process as the electron temperature in 2.45-GHz ECR ion source at PKU is normally lower than 15 eV:^[10]

If the number of H atoms in the plasma increases, it will be advantageous for proton, vice versa. So for an ion source, one should reduce the concentration of hydrogen atoms to some extent.

In addition, in ion sources, the recombination of hydrogen atoms tends to happen on the surface of the chamber wall:^[11]

This recombination process will make the number of H atoms somehow decrease in the discharge volume. Practically, the possibility of recombination on the wall is determined mainly by the category of material, its morphology and temperature.^[11] This leads to a big difference in recombination coefficient (γ) among numerous materials as shown in Table 1.^[11] It can be seen that metals such as stainless steel, tantalum (Ta) have a relative high recombination coefficient with a magnitude of ; but for dielectric materials like quartz and oxide, their recombination coefficients are only around in magnitude. Obviously, for yield, a higher recombination possibility is preferred to reduce the number of H atoms.

Table 1.

Table 1.

Table 1. Atomic hydrogen recombination coefficients for various materials.[11,12] . Material γ Pure SS 0.1 Pure Ta Pure Al Quartz Native oxide-covered SS Native oxide-covered Al

Table 1. Atomic hydrogen recombination coefficients for various materials.[11,12] .

The surface morphology will also have a great influence on the recombination possibility. So, even with identical material, the recombination coefficient will change accompanied with varying of the surface morphology.

The experimental ion source (PMECR II) and test bench were already described in detail in Ref. [13]. Briefly, the ion source was a 2.45-GHz permanent magnet ECR ion source. The magnetic field of the ion source was generated by three NdFeB rings. Then, a mixed beam with H⁺, , and was extracted through a three-electrode system with a Φ6-mm hole. The operation pressure was measured through the vacuum chamber connected to the extraction system. The total current was measured by a water-cooled Faraday cup, and the fraction of each species was analyzed by a 90° dipole magnet afterwards. The inner diameter of the experimental source was 64 mm, and the original chamber material of this source was SS. Generally, the material of the chamber surface was varied by inserting a cylindrical tube liner as shown in Fig. 1.

Fig. 1.

	Figure Option ViewDownloadNew Window
	Fig. 1. (color online) Schematic diagram of ECR ion source at PKU.[5]

3.1. Investigations of liner material

Experiments were carried out under pulsed mode with a duty factor of 10% (100 Hz/1 ms). The extraction voltage in our experiment was fixed at 45 kV. Tube liners made of pure SS, pure Ta, pure Al, and quartz with the same dimension were used to change the surface of the plasma chamber. In Fig. 2, the pure current of varying with rf power is compared by putting different liner cylinders inside the ion source with other operation parameters remaining the same.

Fig. 2.

	Figure Option ViewDownloadNew Window
	Fig. 2. (color online) current versus rf power for different materials.

It is obvious in Fig. 2 that SS and Ta materials seem to have high potentials for producing more compared with quartz and Al. This may indicate that a high recombination coefficient material is advantageous for producing .

Actually, the optimized operation parameters for production were always different with different materials. So, the maximum currents of that can be obtained with different materials are illustrated in Fig. 3. For SS and Ta, nearly 40-mA pure beam was extracted from the ion source. For quartz, whose γ is , a lower production was shown. But for the Al liner, which actually should have very high recombination possibility, only 20-mA could be generated here. This was because a stable alumina layer could exist on the outer surface of metal Al even under hydrogen plasma bombardments which meant that the recombination coefficient of Al liner was actually very low compared with pure metal as shown in Table 1.^[12] Furthermore, alumina also had a high secondary electron emission ability which was advantageous for the dissociation of and formation of proton.^[9] This may explain why the current of was the lowest with Al liner in experimental materials.

Fig. 3.

	Figure Option ViewDownloadNew Window
	Fig. 3. (color online) Optimized currents varying with category of material.

3.2. Effect of O₂ discharge

It was found in our experiment that after the plasma discharging with O₂ in the chamber, there was an increasing trend for the fraction working with H₂ again. To investigate the effect of O₂ discharge, the fraction of was compared with those in the cases of two kinds of chamber surfaces: native SS surface directly from factory (S1) and SS surface after O₂ discharge (S2). For S2, the O₂ discharge was sustained for 5 hours with power 1400 W and pressure , and the energy of O⁺ was several eV.^[14] After that, the ion source vessel was pumped continuously for 2 days to clean impurities before discharging with H₂ flow. Hydrogen discharge results with S1 and S2 were compared in Table 2. It is easy to find out that the fraction of increased from 37.5% to 51.2% after O⁺ ion bombardments on SS surface under the same operation parameters (1400 W and ). Even after 20-h hydrogen discharge, the fraction would not decrease according to our experiment.

Table 2.

Table 2.

Table 2. Ion fractions measured with native SS surface and SS surface after O2 discharge. . SS surface H+ S1 58.4% 37.5% 4.1% S2 41.5% 51.2% 7.3%

Table 2. Ion fractions measured with native SS surface and SS surface after O2 discharge. .

3.3. Surface investigations

The reason for the effect of O₂ discharge was not so clear. For analyzing the changes on the chamber wall, several square SS samples each with a dimension of 1 × 1 cm and thickness of 1.5 mm were fabricated. These pieces were placed on the surface of the discharge chamber to simulate the environment under O⁺ or H⁺ bombardments, and the pieces were all polished by sand paper to keep the roughness the same prior to treatment. The discharge parameters were shown in Table 3. Then an Agilent 5500 atomic force microscope was used to analyze the morphology and an x-ray photoelectron spectroscope on an AXIS-Ultra instrument from Kratos Analytical was used to analyze the constitutes.

Table 3.

Table 3.

Table 3. Treatment conditions for SS samples. . Sample series Treat condition 1 untreated 2 O+discharge for 5 hours 3 H+ discharge for 20 hours 4 Firstly, O+ discharge for 5 hours; then H+ discharge for 20 hours

Table 3. Treatment conditions for SS samples. .

Figure 4 shows the typical three-dimensional (3D) AFM images, and figure 5 shows the surface roughness measured by AFM. The surface of sample 2 becomes obviously uneven after O⁺ discharge as shown in Fig. 4(b). The roughness increased from 55 nm to 73 nm due to O⁺ bombardments. It was in accordance with the result from Zhao et al.^[15] Interestingly, it is also shown in Fig. 5 that the roughness values of SS do not vary obviously after 20-h H₂ discharge by comparing roughness values of samples 3 and 4 with those of samples 1 and 2 separately.

Fig. 4.

	Figure Option ViewDownloadNew Window
	Fig. 4. (color online) AFM 3D images: (a) untreated 1, (b) O2 discharge, (c) H2 discharge. (d) first O2, then H2 discharge.

Fig. 5.

	Figure Option ViewDownloadNew Window
	Fig. 5. (color online) Roughness of sample surfaces under different discharge conditions.

Aleksander Drenik et al. found the relationship between roughness and the recombination possibility of hydrogen atoms on fine-grain graphite surface,^[16] indicating that the recombination coefficient γ increases with the roughness increasing. This can be explained by two facts: firstly, the effective surface for recombination has actually increased due to the increase of roughness; meanwhile, more than one collision between atoms and the surface will happen on an uneven surface compared with on a flat surface. For metal, the recombination coefficient could also increase due to the increase of roughness. So the increasing of roughness caused by O⁺ bombardments, which leads to a higher recombination coefficient, may be mainly responsible for the increasing of yield after O₂ discharge.

As an oxide layer could also exist during O₂ discharge, the effect of the oxide layer should be estimated. Actually, unlike the oxide layer of Al, the oxide layer of SS can only exist for a very short time (on the order of second) when exposed to hydrogen plasma because it will be cleaned by hydrogen plasma.^[17] Therefore, the influence of oxide should be minimum in hydrogen discharge. In our experiments, the XPS analysis also shows that the relative concentration of O increases after O₂ discharge, but it is recovered to the original value soon after H₂ discharge. So, oxide is indeed not the main reason for the enhancement of ions as it cannot exist in hydrogen plasma.

In conclusion, O⁺ discharge with the SS surface is advantageous for the yield. As the effect of the oxide layer is excluded, it might mainly be caused by the increasing of roughness on SS surface.

After the above study and improvements, a high current ion beam is generated with stainless steel discharge chamber. It can be seen in Fig. 6 that ion has already become the main ion species in the mixing beam. The pure current of ions is improved to 42.3 mA with a fraction of 52.9%.

Fig. 6.

	Figure Option ViewDownloadNew Window
	Fig. 6. (color online) Profile of H+ and and .

A high current ion source is under development at PKU. To obtain more ions, the operation parameters and ion source geometry should be designed thoroughly. In our study, the choosing of wall material is found to have an influence on generation as the recombination possibility varies with material type. Moreover, O₂ plasma discharge can be used to improve the recombination possibility as well as fraction. In the future, an ion source working in CW mode will be designed, and more results will be reported.