VACUUM CHAMBERS AND VACUUM SYSTEM OF 15–25 MeV BETATRONS

A. G. VLASOV

(Presented by the scientific seminar of the Physico-Technical Faculty)

One of the main elements of a betatron is the vacuum accelerating chamber.

The vacuum accelerating chamber must meet the following requirements:

1. The chamber must have sufficient mechanical strength to withstand atmospheric pressure equal to \(1 \text{ kg}/\text{cm}^2\).

2. The chamber material must be vacuum-tight, i.e., gas-impermeable.

3. The chamber material must have a low sorption capacity and must degas well during pumping.

4. The chamber material must also be a dielectric permitting a strong coating of the inner surface with a conducting layer.

Fig. 1

For betatrons with accelerated-electron energies up to 25–30 MeV, the greatest use has been made of vacuum chambers manufactured from glass or porcelain. They satisfy the requirements listed above quite fully.

The chambers for betatrons from 5 to 25 MeV used by us are manufactured from molybdenum glass of grade ZS-5 (Fig. 1).

The transverse section of the chamber is usually chosen to be egg-shaped and repeats the section of the interpolar space.

For the selected cross section and diametrical dimensions of the chamber, a cast-iron mold split along the central plane is cast. Depending on the chamber diameter, the mold is cut either into 2 parts, or one third of the mold is cut out. For chambers with an outside diameter up to 400–450 mm, one third is cut out of the mold; with an outside diameter above 450 mm, the mold is cut into 2 parts.

Thus, chambers with diameters up to 450 mm are made from two halves, and those with diameters above 450–500 mm—from three parts.

The upper and lower halves of the mold are fastened with hinges (Fig. 2).

Fig. 2. Mold for blowing chamber sectors for a 15 MeV betatron

In this mold, chamber sectors are blown from molten glass; after thorough annealing, the necessary number of branch pipes are fused onto them for connecting the injector, vacuum pumps, etc.

After another thorough annealing, the sectors are sealed together with the open flame of gas burners, and the finished chamber is annealed again.

When a chamber is made from three sectors, annealing is carried out after each seam has been welded.

The cost of a chamber made by this method for 15 MeV betatrons does not exceed 600–650 rubles, and for a 25 MeV betatron, 800 rubles.

The manufacture of glass chambers for extracting the electron beam is somewhat more complicated, since it is necessary to weld on branch pipes of elliptical cross section, of fairly large dimensions along the major axis of the ellipse and directed tangentially to the chamber. Welding on such branch pipes is rather difficult. Therefore chambers for extracting the electron beam are more expediently made of porcelain by casting. The manufacturing technology for such chambers has been developed quite well at the Leningrad Porcelain Factory named after Lomonosov.

Chambers made by the above method for 15–25 MeV betatrons are subjected to the following treatment:

1. Thorough washing of the inner surface of the chamber.
2. Soldering on a previously prepared branch pipe with a prepared lead-out of the layer.
3. Deposition of a conducting layer on the inner surface of the chamber.

The purpose of washing the chamber is thorough degreasing of its inner surface before applying the conducting layer; this is done by various methods depending on the degree of contamination of the surface. Most often we use washing with concentrated nitric acid, 40–50% alkali, chromic mixture, and, if this is insufficient, also with alcohol.

Especially important in preparing the chamber for operation is arranging a reliable lead-out of the layer from the chamber. The lead-out must have good electrical contact with the layer, allowing currents of up to 1–2 amperes to pass in a pulse, and in betatrons for higher energies even more, 3–4 A.

We carried out the following investigations of various arrangements for the lead-out of the layer:

a) Molybdenum leads were sealed into the branch pipe for pumping; after cleaning them of oxide electrolytically, their connection to the layer was made by applying Wood’s alloy, followed by melting it;

b) Nickel strips were glued to the layer with powdered Wood’s alloy mixed with BF-4 adhesive;

c) A spring contact with the layer was made, connected by electric welding to a molybdenum lead;

d) A layer of silver was sprayed under vacuum onto molybdenum leads that had first been nickel-plated galvanically, since a sprayed silver layer without nickel-plating of the leads does not give good electrical contact with molybdenum. Nickel serves as an intermediate layer between the silver and the molybdenum.

At present, the most widely used arrangement for the lead-out is by applying silver by the method of firing silver paste into the glass. We use this method in two ways.

For chambers with continuous pumping, a small ring of paste is applied to the inner surface of a previously prepared branch pipe for pumping. Leads from this ring are made with the same paste over the edge of the branch pipe, which is connected to the vacuum system. After this the branch pipe is soldered to the chamber.

The second modification of the firing method is most convenient for sealed-off chambers.

Into a glass tube with a diameter equal to the diameter of the chamber pumping branch pipe, 30–35 mm, and a length of 15–18 cm, 2–3 molybdenum leads are sealed. The diameter of the molybdenum is 1–1.2 mm. The ends of the leads protruding inside the tube

by 3–4 mm, are cleaned of oxide and nickel-plated by an electrolytic method. At the place where the leads are sealed in, a paste is applied in a ring 2–2.5 cm wide to the inner surface of the glass tube and to the ends of the leads. After firing, a reliable electrical contact is obtained between the layer of fired silver and the molybdenum leads; it permits an alternating current of the order of 10–15 amperes to be passed for 10–15 minutes, and only after this, owing to heating of the layer and the glass, is the contact broken.

The first three above-described methods of making the layer lead-out give a much poorer contact with the layer and are destroyed when a current of 0.5–0.8 A is passed for a short time.

The method of depositing silver under vacuum provides a sufficiently strong contact, withstanding a current of up to 8–10 A for 5–10 minutes; but when the layer is applied it partly separates from the glass, which leads to deterioration of the contact. Thus the best method proved to be the method of firing silver.

The tube with the completed lead-out of the layer is soldered to the pump-out branch tube of the chamber, and the chamber is coated with a conducting layer. The conducting layer applied to the inner surface of the chamber must satisfy the following requirements:

1. A sufficiently strong bond with the walls of the chamber.
2. Low sorption capacity and good degassing during the pumping process.
3. Low emission capacity when bombarded by electrons.
4. A sufficiently high specific resistance.

There are no definite recommendations in the literature for the specific resistance of the layer, and the data given are very different [1–8].

The conducting layer in our chambers has a resistance of 50–100 ohms between diametrically opposite points in chambers for 15 MeV betatrons.

Application of the conducting layer is carried out by a wide variety of methods.

For chambers of 15–25 MeV betatrons, the most widely used methods have been coating the inner surface of the chamber with aquadag or by chemical silvering. These are the simplest methods of applying the layer, requiring no special equipment and more or less satisfying the above requirements for the layer.

These methods also have substantial disadvantages. A layer applied by chemical silvering adheres poorly to the surface of the chamber and, when moisture gets in, separates from the surface. This drawback makes it impossible to use silvering for making a sealed-off chamber, since when the injector is soldered in, nonuniform heating of the chamber is unavoidable, and in the coldest places the condensing water vapor destroys the layer.

A layer applied with aquadag has a very considerable sorption capacity, is difficult to degas—which is especially inconvenient in chambers with continuous pumping—and requires prolonged pumping. The aquadag layer also releases a significant quantity of gases when bombarded by electrons. The use of an aquadag coating for sealed-off chambers is inconvenient because it is opaque and makes it difficult to observe the correctness of the welding-on of the injector. Other methods of applying the conducting layer are also known, such as silver sputtering, coating with liquid palladium, tin oxide, etc., but we did not use them [5, 6].

After the layer has been applied, the chamber is ready for installation in the inter-pole space of the betatron and for connection to the vacuum system.

The vacuum system we use consists of a fore-vacuum pump of the RVN-20 type, a high-vacuum pump of the TsVL-100 type

and a “trap” in the form of an elbow with measuring tubes LT-2 and LM-2 placed on it.

The “traps” are made either of glass, connected to the high-vacuum pump by means of a ground joint, or of metal, soldered directly to the pump flange. If the “trap” is glass, the chamber is attached with picein cement. In the case of a metal trap, the chamber and the manometric tubes are attached by means of a rubber seal.

The vacuum system we use provides a vacuum in the chamber on the order of \(2\text{--}5 \cdot 10^{-6}\) mm Hg, which is quite sufficient for stable operation of the betatron. The use of traps cooled with liquid nitrogen for pumping glass or porcelain betatron chambers up to 25 MeV is inexpedient, since it complicates the system and its operation.

Control of the vacuum system is carried out directly from the control panel, with the exception of closing off the hose connecting the high-vacuum pump with the roughing pump.

One of the important improvements in betatrons is the replacement of continuously pumped accelerating chambers by sealed-off chambers, which do not require bulky vacuum systems and roughing and high-vacuum pumps for their operation.

Fig. 3

When sealed-off vacuum chambers are used, the betatron can be prepared for operation within several minutes, whereas when operating with a continuously pumped chamber, about 1–1.5 hours are spent preparing the betatron for operation, this time being necessary to attain the required degree of vacuum.

Another advantage of using a sealed-off vacuum chamber is the possibility of constructing a mobile betatron, which is especially important for medicine in irradiating patients.

A sealed-off vacuum chamber makes it possible to arrange the injector most rationally in azimuth, which is not always possible in chambers with continuous pumping. Thus, the use of a sealed-off chamber has a great advantage over chambers with continuous pumping.

Despite the fact that the modern literature contains references to the operation of such chambers, the technology for manufacturing them has not yet been published anywhere.

Together with B. N. Gulko we manufactured a sealed-off chamber for a 15 MeV betatron (Fig. 3).

When the chamber intended for sealing off was manufactured at the factory, the branch pipe for the injector was welded on somewhat longer than in ordinary chambers and with a thinner wall. This is necessary so that, when the injector is soldered on, the heating of the joint between the branch pipe and the chamber will not be too great. Under strong heating this joint may crack.

The fabrication of a sealed-off chamber for a 15 MeV betatron showed that, as a result of a slight complication in the manufacturing technology of the vacuum chamber, a considerable simplification in the operation of the betatron is achieved.

REFERENCES

1. Grinberg A. P. Methods of Accelerating Charged Particles, State Publishing House of Technical-Theoretical Literature, 1950.
2. Vorob'ev A. A. Charged-Particle Accelerators, GEI, 1950.
3. Resonance Cyclic Accelerators of Charged Particles (collection of articles), I. L., 1951.
4. Kerst D. W., Rev. Sci. Instr., 1942, 13, 384–394.
5. Kerst D. W., Abams I. D. and Jr., Phys. Rev., 1950, 78, 297.
6. Jregg E. C., Rev. Sci Instr., 1951, 22, 176–182.
7. Kerst D. W., Phys. Rev., 1941, 60, 47–52.
8. Jeward F. R. Lawson I. D. and Jr., Proc. I E E, 1950, 97, 1, 320–325.