Improved vacuum system for high-power proton beam operation of the rapid cycling synchrotron

The vacuum system in the rapid-cycling synchrotron of the Japan Proton Accelerator Research Complex has been operated for more than 10 years. It becomes evident that the high-power beam operation has more powerful effects on the vacuum system than expected at the time of the design. Those effects of the high-power beam are categorized into two types of events, the malfunction of vacuum equipment and the large pressure rise. A specific example of the former event is the failure of the turbomolecular pump (TMP) controller by the full loss of a single-shot high-intensity beam. The TMP itself is also damaged by a bearing crush due to a touch down without braking the rotor. We have developed a TMP controller that can connect with long power and control cables of more than 200 m lengths. This length makes the controller be installed in a control room where there is no radiation influence. The TMP with high-strength bearing has been also developed. The TMP system becomes tolerant to a large loss of the high-intensity beam by using such a controller and a TMP. The latter event is an extreme pressure rise with increasing the beam power. The dynamic pressure behaviors during the beam operations are investigated to understand the outgassing mechanism by the beam. It is indicated that the pressure rise mechanism is a result of the ion-stimulated gas desorption. The analytical calculation based on the ion-stimulated gas desorption model well reproduces the measured dynamic pressure. The calculation also shows that, among the parameters, a larger pumping speed per unit length $s$ and a smaller initial surface density of the adsorbed molecules $q_0$ are required to suppress the extreme pressure rise. The nonevaporable getter (NEG) pumps are additionally installed to obtain the larger $s$ and smaller $q_0$. It is confirmed by simulating the beam line pressure distribution that the additional NEG pumps are effective to obtain the larger $s$ in every position along the beam line. The ability of the NEG pump to keep a low pressure during the vacuum system shutdown, which can contribute to maintain the small $q_0$, is examined by the buildup test. It is finally confirmed that the dynamic pressure during the high-power beam is effectually suppressed.

Figure 1

Layout of the RCS vacuum system. The RCS consists of three straight sections and three arc sections, between which all-metal gate valves are installed. The beam line is divided into 27 cells for convenience. Vacuum gauges are not drawn in the figure. The ion pumps have not been used because of the intermittent outgassing problem (Fig. 4). The beam pipes and bellows are made of titanium. The beam pipes in the magnets are made of alumina ceramics, whose inner surface is coated by TiN.

Figure 2

Typical example of the beam line pressure distribution in the RCS. The pressure at about one month from the start of evacuation after the vacuum system shutdown.

Figure 3

Configuration of a TMP system. The radiation-resistant TMP is installed in the accelerator tunnel. The TMP controller and the backing pump, which are not radiation resistant, are installed in the utility tunnel. The accelerator tunnel and utility tunnel are separated by the concrete floor with a thickness of more than 1 m. Thus, the radiation is much lower in the utility tunnel compared to the accelerator tunnel. There are three types of valves between the TMP and the backing pump, which are metal manual valve (MV), pneumatic valve (PV), and solenoid valve (SV). The SV closes when the electrical voltage is not applied to the backing pump.

Figure 4

Pressure instability caused by the ion pumps. The pressure measured by the CCG in C19 is plotted as a representative.

Figure 5

Example of pressure rise during high-power beam operation. Pressure evolution during the beam commissioning with an output beam power of 600 kW is plotted. The pressure measured by a CCG in C15 of the second arc section is shown in this figure as a typical case.

Figure 6

Example of residual gas spectra in the beam line (a) without and (b) with beam measured by the RGA in the second arc section. Each spectrum corresponds to points $\alpha$ and $\beta$ in Fig. 5, respectively.

Figure 7

(a) Example of the correlation between the beam line pressure and the beam loss monitor signal in the 600 kW beam operation of Fig. 5. (b) Example of the correlation between the pressure and the number of protons measured by the beam current monitor in the same period as panel (a).

Figure 8

Beam loss monitor signal distribution when a single-shot beam with the power equivalent to 1.2 MW is fully lost. The signal is integrated for one cycle from the beam injection to the extraction, 20 ms. The position of the failed TMP system is also shown. The beam loss signal distribution in the normal 1 MW equivalent beam acceleration case is also shown for reference.

Figure 9

Damaged rotor and stator blade of the TMP. The diameter of the rotor and stator is about 220 mm. (a) Chipped edges of the rotor. (b) Deformed stator blade.

Figure 10

Damaged bearing parts of the TMP. (a) Inner bearing rail with impressions by the balls. (b) Bearing ball with scuffs.

Figure 11

Configuration diagram of the improved TMP system using the controller with the long cable specification. The controller is 220 m away from the TMP in maximum. There is no radiation influence in the control room.

Figure 12

Measured roundness distribution of the inner bearing rail after the FRTD test for the (a) normal and (b) improved bearings.

Figure 13

Change of the rotor rotation speed against the time in the FRTD test for the TMP with the normal bearing and improved one. The origin of the horizontal axis is the time of the rotor touch down.

Figure 14

History of the output beam power to the MLF from the RCS. Denotations A–G, X, and XX represent points, for which the dynamic pressures are analyzed in the present work.

Figure 15

Beam line pressure during the high-power beam operation. Panels (a) to (e) correspond to the beam line pressure in points A to E in Fig. 14, respectively. The pressure measured by the representative CCGs in each section is plotted. C02: Injection section, C06: first arc section, C12: Extraction section, C15: second arc section, C20: RF section, and C24: third arc section. (a) (Point A) Pressure in the beam commissioning after position alignment work of the beam line equipment. The beam lines in all sections have been purged by argon gas and opened to air during the work. (b) (Point B) pressure during the user operation at the 200–300 kW beam power after point A. (c) (Point C) pressure in the beam commissioning after suspension of the beam operation. The beam line has been kept to be evacuated by the TMPs during the suspension. (d) (Point D) pressure in the beam commissioning after vacuum system shutdown for maintenance. A few flanges in some sections have been opened to air for a short time. (e) (Point E) pressure during the user operation at the 500 kW beam power after point D.

Figure 16

Relation of the maximum pressure to each beam power in points A, C, and D of Fig. 14 [Figs. 15, 15, and 15]. The CCG values in C06 and C24 are plotted as representatives. Each pressure value is normalized by the pressure without beam.

Figure 17

Beam line pressure in points F and G in Fig. 14. The pressure measured by the same CCGs as Fig. 15 is plotted. (a) [Point F] Pressure during the user operation with the 500 kW beam power. (b) [Point G] Pressure when high-power beam commissioning is performed after the vacuum system shutdown. The beam line has been kept at a vacuum with the TMPs stopped except in the third arc section, where the NEG pumps are additionally installed in the third arc section during the shutdown (Sec. 5a). The pressure value of $5\times {10}^{-8}\mathrm{Pa}$ is the minimum limit of CCG detection.

Figure 18

Dynamic pressure during the 1 MW beam power operation in points X and XX in Fig. 14. The pressure measured by the same CCGs as Fig. 15 and the CCG in C26 is plotted. (a) [Point X] Pressure in the first 1 MW beam test. The TMP in C26 has been stopped because of the FRTD trouble. The beam operation is interrupted three times due to the beam loss monitor signal increase caused by the pressure rise. (b) [Point XX] Pressure in the second 1 MW beam test. All TMPs have been uneventfully under operation. The beam operation is interrupted four times due to the trip of the acceleration tube in the linac, etc.

Figure 19

Comparison of the calculated pressure with the one measured by the CCGs. The origin of the horizontal axis is the time when the beam operation started. The adjusted $q_0$ value is also presented in each panel. (a) Pressure in C15 at point G of Fig. 14 during the beam operation with 500 kW power [Fig. 17]. (b) Pressure in C26 at point X of Fig. 14 during the beam operation with 1 MW power [Fig. 18]. The beam interruptions during the beam operation are not considered in the calculation.

Figure 20

Dependence of the pressure on the beam power at point G in Fig. 14 [Fig. 17]. In the calculation, the $q_0$ value is fixed to $4\times {10}^{20}{\mathrm{m}}^{-2}$, which is the value obtained for the case of 500 kW beam power [Fig. 19]. (a) Calculated time variation of the pressure. The origin of the horizontal axis is set to the time when the beam operation started. (b) Comparison of the calculated maximum pressure with the one measured by the CCG in C15 at each beam power.

Figure 21

Calculated dependence of the maximum pressure on the initial surface density of the adsorbed molecules for some pumping speed per unit length values. The beam power is fixed to 1 MW. The initial pressure is set to $5\times {10}^{-7}\mathrm{Pa}$.

Figure 22

Calculated dependence of the maximum pressure on the beam power for the combinations of two realistic $s$ and $q_0$ values. The initial pressure is set to $5\times {10}^{-7}\mathrm{Pa}$. The open circles represent the critical beam power $P_{\mathrm{beam}\_\mathrm{cr}}$ in each combination of $s$ and $q_0$ values. The $P_{\mathrm{beam}\_\mathrm{cr}}$ for the combination of $s$: $0.06{\mathrm{m}}^3{\mathrm{s}}^{-1}{\mathrm{m}}^{-1}$ and $q_0$: $5\times {10}^{19}{\mathrm{m}}^{-2}$ is out of range, which is 2.6 MW.

Figure 23

Layout of the vacuum system for the third arc section, where the NEG pumps are additionally installed. Three NEG pumps are installed in the position of the ion pumps.

Figure 24

Model of the molflow+ calculation for the third arc section. The positions of the gauges and the pumps are also depicted.

Figure 25

Pressure distribution in the third arc section before and after the NEG pump installation. The dashed and solid lines denote the calculated pressure by molflow+ without and with NEG pumps, respectively. The open triangles and circles represent the pressure without the NEG pumps measured by the CCGs and the EXGs, respectively. The closed triangles and circles represent those with NEG pumps. The pressure value of $5\times {10}^{-8}\mathrm{Pa}$ is the minimum limit of CCG detection.

Figure 26

Result of the buildup test. Time variation of the pressure measured by the CCGs in the second arc section without NEG pumps and the third arc section with NEG pumps. The origin of the horizontal axis is set to the start of the buildup test.

Figure 27

Time variation of the RGA ion current for typical gas mass numbers during the buildup test. The origin of the horizontal axis is set to the time when the TMPs are stopped. (a) Result of the second arc section without NEG pumps. (b) Result of the third arc section with NEG pumps. In the second arc, the RGA is switched off at around 94 h by the interlock which is set for the total pressure to $2\times {10}^{-3}\mathrm{Pa}$. It should be noted that the absolute ion current between the different RGAs cannot be compared because their sensitivities are different.

Figure 28

Pressure measured by the CCGs during the beam operation after the installation of the additional NEG pumps in the second arc section. Other conditions are the same as point G in Fig. 14 [Fig. 17].

Figure 29

Beam power dependence of the pressure measured by the CCG in C15 of the second arc section without and with NEG pumps during the beam operation. Each data is obtained in 2019 and 2020, respectively. The maximum pressure in each beam power operation is plotted. The values are obtained from Figs. 17 and 28 for the case without and with NEG pumps, respectively.

Figure 30

Pumping curve measured by the EXG in C24 of the third arc section before and after NEG installation. The origin of the horizontal axis is the time after one day of pumping by the TMPs.