VACUUM SYSTEM OF A 300 MeV SYNCHROTRON

A. G. VLASOV, V. P. PONOMAREV

The problem of constructing vacuum systems for electron accelerators and obtaining a sufficiently low pressure in the vacuum chamber is a rather complex task. The operating characteristics and efficiency of the accelerator depend on the solution of this problem.

The requirements for vacuum systems in general and for accelerator vacuum systems in particular—for the material of accelerator vacuum chambers, for their design, and for the required rarefaction created in the chamber—are described quite fully in works [1—4].

The 300 MeV electron synchrotron built at the institute is an accelerator with weak focusing and straight sections. The air gap of the electromagnet was chosen to be small and has dimensions of \(80 \times 200\) mm. All this imposes specific requirements on the vacuum system of this synchrotron. The main elements of the synchrotron vacuum system are: the vacuum chamber, the pumping system, and the control, interlock, and signaling system (CIS).

1. Vacuum chamber

The dimensions of the cross section and the radius of the chamber are derivatives of the interpole space of the electromagnet and of the radius of the equilibrium orbit of the accelerator. Therefore, in designing the vacuum chamber of the synchrotron it was specified that: the outer axial dimension of the chamber cross section \(2v\) should be no more than 70 mm; the inner radial dimension of the chamber cross section \(2a_{\mathrm{v}}\) should be not less than 185 mm; the mean radius of the chamber \(r_{\mathrm{av}} = 967.5\) mm; the length of the straight sections of the chamber \(L_{\mathrm{pr}} = 600\) mm at a quadrant angle \(\gamma_{\mathrm{qv}} = 83^\circ 30'\).

On the basis of the requirements imposed on the material of accelerator chambers [3], it was proposed to make the synchrotron chamber of porcelain. A calculation of the mechanical strength of the chamber showed that at the most highly stressed point of the chamber cross section the thickness of the porcelain should be not less than 17 mm, and at the least stressed point, 7 mm [5]. If the chamber were made with a uniform wall thickness of 17 mm, the azimuthal internal dimension of the cross section would be only 36 mm, and the interpole space of the electromagnet would be used very inefficiently. Therefore, a chamber cross section was designed with different wall thicknesses—7.5 mm along the axial axis and 18 mm along the radial axis. Ta-

Thus, the section of the chamber shown in Fig. 1 is an oval with internal dimensions of \(55 \times 185\) and external dimensions of \(70 \times 221\) mm. Calculation of such a section, and subsequently testing, showed sufficient mechanical strength of the designed and fabricated porcelain chamber.

Fig. 1. Cross section of the vacuum chamber

It is not possible to manufacture the chamber as a single piece because of its considerable dimensions. Therefore the chamber is made of separate porcelain sectors. The sectors are glued into quadrants, which in turn are connected to one another by means of flanges. Each quadrant of the chamber contains 5 curvilinear sectors with branch pipes for accommodating devices for indicating the beam of accelerated electrons, targets, etc. In the first quadrant there is a curvilinear sector with a branch pipe tangential to the orbit, through which the radiation from the target is brought out. The straight sections of the chamber are made of porcelain sections with branch pipes for accommodating indicating devices and for connection to the pumping system. The injection section is glued together from 2 sections of special shape and is a separate part of the chamber located on the straight section of the electromagnet. The second straight section also constitutes a separate part of the chamber, is provided with a section with electrodes, and serves as the accelerating section connected to the high-frequency generator. The two other straight sections are each divided into two parts, and these parts are glued to the adjacent quadrants. This distribution of the sectors made it possible to provide a total of 6 flange joints of the chamber instead of 8. The general appearance of the synchrotron chamber is shown in Fig. 2.

The chamber sectors are glued to one another with BF-4 glue according to the technology described in [3]. The flange joints of the quadrants and straight sections, shown in Fig. 3, are made of duralumin and are sealed during assembly with vacuum rubber. Duralumin flanges are likewise glued to all branch pipes of the sectors and sections; they are closed either by covers, on which the indicating devices are mounted, or by plugs. The covers, the indicator rods, and the light guides are sealed with vacuum rubber. The seals of the rods are sliding seals, and they can move in the radial direction of the chamber.

The inner surface of the chamber sectors is coated with a conducting layer. The requirements for the conducting coating of electron-accelerator chambers are described in detail in [3, 6].

Fig. 2. Vacuum chamber on the test stand

In our case the sectors are coated with Aquadag having a specific resistance of 30–50 ohm/cm². The conducting layer is brought out by strips of fired silver from both ends of the sectors. The strips are connected by copper foil soldered with soft solder to the silver strips.

In order that the electrons not be lost in the course of acceleration due to scattering by the residual gas, a high vacuum must be produced in the chamber.

The operating pressure in the chamber of the synchrotron being described, according to calculations carried out by the method described in [7, 8], should be no more than \(4–5 \cdot 10^{-6}\) torr.

2. Pumping System

To obtain a high vacuum in accelerator chambers, diffusion oil-vapor pumps operating in series with fore-vacuum pumps are usually used. Recently, instead of vapor-jet pumps, ion-sorption pumps have increasingly begun to be used, since they do not contaminate the chamber with vapors of the working fluid

Fig. 3. Flange connection of quadrants and straight sections. 1 — porcelain sections; 2 and 4 — duralumin flanges; 3 — vacuum-rubber ring.

quality, and this is very important during prolonged operation of the accelerator. A more complete comparison of diffusion and ion-sorption pumps, their disadvantages and advantages when used for pumping accelerator chambers, is described in [9].

At the time when the pumping system of the 300 MeV synchrotron was being designed and built, we did not have suitable ion-sorption pumps. Therefore the main pumping system of the synchrotron was designed and manufactured using diffusion oil-vapor pumps.

The synchrotron chamber has straight sections, two of which are occupied by the injection system and the resonator, while two are free. Pumping sections are located on the two free straight sections; diffusion pumps are connected to these sections through intermediate elements (a branch pipe, a valve, a trap, etc.).

Calculations of the synchrotron pumping system, carried out by the method described in [3], showed that the required pumping speed of the diffusion pumps must be no less than 340 l/s.

In order to provide, at the points of the chamber farthest from the inlets of the diffusion pumps, an operating vacuum of \(5 \cdot 10^{-6}\) torr, it is necessary to have a pump ultimate vacuum of \(1.9 \cdot 10^{-6}\) torr, since the pressure drop from the pumps to these points, according to the calculation, is \(3.1 \cdot 10^{-6}\) torr.

For pumping the chamber we selected VA-05-1 high-vacuum units, having a pumping speed at the pump inlet of 500 l/s. However, these units produce an ultimate vacuum at the pump inlet of \(3 \cdot 10^{-6}\) torr, which is insufficient for pumping the synchrotron chamber. When pumping with these units alone, at the farthest point of the chamber we would obtain a vacuum of \(6.1 \cdot 10^{-6}\) torr, which is inadmissible.

To increase the ultimate vacuum of the units and to prevent oil vapors from the diffusion pumps from entering the chamber, freezing traps are usually used. The use of freezing traps cooled with liquid nitrogen makes it possible to obtain, at the outlet of the VA-05-1 unit, an ultimate vacuum of \(5 \cdot 10^{-7}\) torr. Then the operating vacuum would be \(3.6 \cdot 10^{-6}\) torr, which satisfies the specifications, but operation of the vacuum system when traps are used becomes considerably more complicated.

We used thermoelectric traps of the TVL-500 type as freezing traps. The use of these traps prevents oil vapors from entering the chamber, but does not increase the ultimate vacuum of the units. Therefore, near the injection and resonator straight sections, i.e., near the points of the chamber farthest from the pumping locations, two titanium ion-sorption pumps, each with an ultimate vacuum of \(7\text{–}8 \cdot 10^{-7}\) torr and a capacity of 40 l/s, were connected to the sector branch pipes. These pumps were developed and manufactured in our laboratory, and their design is described in [9, 10].

The installation of titanium ion-sorption pumps made it possible to obtain, at the most remote point of the chamber, a pressure of \(3\text{–}4 \cdot 10^{-6}\) torr. The added complication in operating the vacuum system when using the additional titanium pumps is very small.

To create the preliminary vacuum in the chamber and at the inlets of the diffusion pumps, booster pumps are usually used in combination with fore-vacuum pumps, or fore-vacuum pumps alone. The arrangement and number of fore-vacuum pumps are selected on the basis of design considerations and convenience of operation. But in doing so it is necessary that the fore-

the vacuum system ensured, at the inlet of the VA-05.1 units, a pressure of no more than \(7.5 \cdot 10^{-2}\) torr.

In our case, a single fore-vacuum system, made of stainless-steel pipes 50 mm in diameter, is used for pumping down the two high-vacuum units. It is connected to the pumps through valves. Vacuum rubber serves as the seal. One pump of type VN-1 is used as the fore-vacuum pump; it makes it possible to create, in the chamber and at the inlet of the diffusion pumps, a preliminary rarefaction no worse than \(2 \cdot 10^{-2}\) torr within 15–20 minutes.

The complete layout of the synchrotron vacuum system is shown in Fig. 4.

Fig. 4. General layout of the synchrotron vacuum system. 1 — chamber quadrants; 2 — resonator section; 3 — injection section; 4 — high-vacuum units VA-0.5-1; 5 — fore-vacuum pipeline; 6 — fore-vacuum pump VN-1; 7 — gauges for measuring vacuum; 8 — flanged connections of the quadrants and straight sections

The stringent requirements imposed on the reliability of operation of all synchrotron assemblies, in particular on the vacuum system, make it necessary to provide automatic control of the pump power supply, as well as continuous monitoring of their operation and of the operating pressure in the chamber. For these purposes a power-supply and control circuit for the vacuum system was designed and manufactured; a detailed description of it is given in a separate article.

Here we shall only note that, in implementing the control system, a vacuum relay with an operating pressure range of \(10^{-1}\)—\(10^{-2}\) torr was developed and manufactured, as described in [11], since relays for the indicated pressure range are not produced by our industry.

Vacuum is measured at several points of the chamber and at the inlet of all pumps by LM-2 and LT-2 manometric lamps, mounted in special mushroom-type vacuum seals.

3. Adjustment of the accelerating chamber and the vacuum system

The vacuum accelerating chamber creates a space, partially evacuated of gas, in which the process of electron acceleration takes place. The sources of residual gas and vapor in the accelerating chamber are leakage through non-tight joints and gas evolution from the surface of the chamber and the various fittings placed in it.

The purpose of adjusting the accelerating chamber is to ensure the tightness of the permanent and detachable joints of the chamber, and also to reduce the internal surface area of the chamber.

Testing of the accelerating chamber is divided into bench testing and testing directly on the accelerator. In preparing the chamber elements for testing, the following operations are carried out:

a) the internal surfaces of the quadrants and straight sections of the chamber, coated with Aquadag, are wiped with a soft cloth and blown with a jet of air;

b) the chamber flanges, plugs, and indication devices are washed with a solvent (alcohol, gasoline, acetone).

A significant source of gas evolution in the chamber is the gaskets made of vacuum rubber. Reduction of their gas evolution is achieved by boiling in a 5% KOH solution [2] for 30 minutes, followed by washing with distilled water and alcohol. Better results are obtained by vacuum treatment of rubber gaskets. In this treatment the gaskets are heated to 100°C and pumped through a nitrogen trap for 20 hours [12].

After the indicated preparation, the vacuum chamber is assembled on the test stand. To test the vacuum chamber for tightness and gas evolution, a standard vacuum test station [13] is assembled, consisting of a VA-05-1 high-vacuum unit with an RVN-20 backing pump and a PTI-6 helium leak detector.

Testing of the accelerating chamber is carried out in two stages:

1. Leaks are detected with the leak detector while the chamber is being pumped on the stand. Leaks at the glued joints are eliminated by coating the seam with an adhesive of the following composition: ED6-100 epoxy resin, parts by weight; porcelain powder—40–50 parts by weight; dibutyl phthalate—10 parts by weight; polyethylene polyamine—8 parts by weight. Leaks in detachable assemblies are eliminated by the appropriate tightening of screws. It is useful, before installation, to lubricate the rubber gaskets lightly with castor oil. After this the chamber is tested for leakage and gas evolution by the method described in [13]. Testing of the chamber for leakage and gas evolution must be carried out during the first pumping, when gas evolution is at its maximum. A vacuum chamber prepared and tested on the stand may be mounted on the accelerator.

2. The second stage of chamber testing consists in checking all chamber assembly units with the leak detector. Detected leaks must likewise be eliminated.

The measures for preparing the pumping system are similar to the adjustment of the vacuum accelerating chamber. Leaks in the pumping system are likewise impermissible. In the pumping system one need not determine the total magnitude of gas evolution and leakage. But before connecting the chamber to the pumping system, its principal operating para-

the ultimate vacuum, in meters, that it can provide over a prolonged period of continuous operation.

For normal operation of the accelerator this pressure should be \(4—5\cdot 10^{-7}\) torr. Preparation of the high-vacuum and fore-vacuum pumps is carried out in accordance with the instructions of the manufacturing plants.

General questions concerning the preparation and operation of vacuum pumps are fully covered in the literature \([1, 2, 14, 17]\), and therefore we do not present them here.

A convenient criterion for checking the normal operation of the vacuum system on the accelerator is the residual pressure remaining in the chamber after a certain interruption in operation. Thus, after an interruption in operation of \(\sim 15\) hours, the residual pressure in the chamber, with normal leakage and gas evolution, should be no higher than \(3—6\cdot 10^{-2}\) torr.

Long-term operation of the accelerator has shown the good reliability of the vacuum system. The operating pressure in the chamber is maintained at the level of \(3—4\cdot 10^{-6}\) torr.

REFERENCES

1. Vacuum Equipment and Vacuum Technology, ed. by A. Guthrie and R. Wakerling, IL, 1955.

2. B. I. Korolev. Fundamentals of Vacuum Technology. Gosenergoizdat, 1958.

3. A. G. Vlasov, V. P. Ponomarev, M. T. Shivyrtalov, P. M. Shanin. Izv. TPI, vol. 122, 99, 1962.

4. I. F. Malyshev, A. V. Popkovich et al. PTE, No. 4, 46, 1962.

5. A. G. Vlasov. On the Mechanical Calculation of the Strength of Vacuum Chambers. Report NII-0207, Tomsk, 1958.

6. V. P. Ponomarev. Conducting Coatings of Accelerator Chambers. Report NII-01522/273, Tomsk, 1961.

7. A. G. Vlasov. Izv. vuzov “Fizika”, No. 1, 20, 1961.

8. A. G. Vlasov. ZhTF, XXXI, 613, 1961.

9. Electron Accelerators. Proceedings of the Fourth Interuniversity Conference, 384, Moscow, 1964.

10. V. P. Ponomarev. PTE, 6, 196, 1963.

11. V. P. Ponomarev. Vacuum Relay with a Range of \(10^{-1}—5\cdot 10^{-3}\) torr. Report NII-01520/274, Tomsk, 1961.

12. V. A. Karetnikov et al. Linear Ion Accelerators. Gosatomizdat, 1962.

13. V. A. Lanis, L. E. Levin. Practical Foundations of Vacuum Tests. GEI, 1955.

14. J. Yarwood. High Vacuum Technique, GEI, 1960.

15. R. Jaeckel. The Production and Measurement of Vacuum. IIL, 1952.

16. V. I. Kuznetsov. Mechanical Vacuum Pumps. GEI, 1959.

17. Ya. Groshkovskii. High-Vacuum Technology. IIL, 1957.