Evaluation of the TiSP3

The purpose of this project was to evaluate the performance and capabilities of the TiSP3, a titanium sublimation pump located at three meters from the interaction point of CESR. Maximum pumping speed obtainable, the optimal number of filament flashes, and the pumping capabilities of a nitrogen-saturated chamber were investigated. Introduction The Cornell Electron Storage Ring (CESR) stores a clockwise positron beam and a counter-clockwise electron beam and allows the beams to collide at the interaction point within the CLEO detector. Once the beams are transferred to CESR, they must be stored there for several hours before the collision can take place. This puts very strict demands on the quality of the vacuum. The residual gas molecules in the paths of the beam particles must be many billion times less dense than in the normal atmosphere. Vacuum pumps distributed around the ring create and maintain this ultra high vacuum. TiSP3 is the titanium sublimation pump located at 3 meters from the interaction point of CESR. It is the closest pump to the interaction point, so it is responsible for maintaining the ultra high vacuum needed throughout the detector and surrounding beam pipe. By capturing gas molecules on its inner surfaces, the TiSP3 maintains the vacuum by preventing the molecules from circulating back into the beam chamber. The Apparatus The TiSP3 is a circular stainless steel chamber around the central beam line tunnel. It houses five cartridges of three titanium filaments each and gains its pumping capabilities only after these filaments are flashed. Flashing is the process of heating up the filaments by applying 50 amps for a duration of two minutes. The filaments sublimate titanium molecules, which then cover the inner surfaces of the chamber to form a sticky surface. Gas molecules are then adsorbed onto this surface, thus removing them from circulation and improving the vacuum. For the duration of testing, several additional components were attached to the chamber. At the top of the chamber, a Residual Gas Analyzer (RGA) ionized gas molecules and then sorted the resulting ions by mass in order to determine the identity and quantity of gases present. Partial pressures were found by identifying molecules with dominant atomic masses. Two cold cathode gauges (CCG) were also connected to the chamber, one located at the upper part of the chamber and one near the beam line to get a more relevant pressure reading. They measured voltage, from which pressure could be obtained using equation 1 P = 10 , (1) where a and b were constants determined from initial data fits. The chamber CCG gave a different pressure reading than the beam line CCG because the central beam line tunnel was open to the rest of the chamber only through a series of pumping slots. The conductance of these slots limited the minimum pressure attainable at the beam line. One end of the chamber was attached to a mechanical turbo molecular pump. After accounting for the hoses that connected the turbo pump to the chamber, it had an effective pumping speed of 3 l/s. The other end of the chamber was connected to a gas manifold containing two calibrated leaks, N2 and CO. The known leak rates of these bottles were then used to study the pumping characteristics of titanium subject to differing conditions. Experiment Chronology The chamber was initially baked-out at 150 C in order to remove any residual gases, mainly nitrogen and water. After the chamber had cooled, the filaments were flashed once and the N2 leak was introduced, primarily to determine the RGA sensitivity calibration. In order to investigate the pumping characteristics of nitrogen-saturated titanium, the first CO leak (CO-2) was introduced without flashing. After this saturation, a local power failure caused the system to be exposed to air. Another bake-out was omitted in order to observe the effects of a contaminated system on pumping characteristics. The filaments were then flashed once before the next CO leak (CO-3) was introduced. In order to investigate the effects of increasing the number of flashes, the filaments were flashed twice before introducing the fourth CO leak (CO4), and then flashed three times before the fifth CO leak (CO-5). Results Pumping Speeds Evaluation of the pumping characteristics of TiSP3 was best done using pumping speeds. Pumping speed was calculated by taking the constant calibrated gas leak rate and dividing by the pressure, as shown in equation 2. S = (dQ/dt)/P (2) Figure 1 shows pumping speeds as each of the CO saturations proceeded. As the titanium saturated and pressures rose, pumping speeds dropped to that of the turbo pump alone. A large initial difference between chamber and beam line pumping speeds was also observable. Slot conductance limited the maximum beam line pumping speed, but the maximum pumping speed at the chamber was not affected. Figure 1. Comparison of Pumping Speeds from CO Saturations CO-2 through CO-5 The differences between maximum pumping speeds are quantized in Table I. The maximum beam line pumping speed of 1701.9 l/s is notable. Table I. Maximum Pumping Speeds CO-2 Smax (l/s) CO-3 Smax (l/s) CO-4 Smax (l/s) CO-5 Smax (l/s) Chamber 4334.5 2462.1 4048.9 5139.5 Beam Line 1663.1 1320.2 1651.6 1701.9 Gas Load Values at Saturation Another interesting property of titanium pumping we wished to investigate was the gas load at which the titanium became saturated. Gas load was defined as the product of the gas leak rate and the elapsed time since the gas leak was opened, as shown in equation 3. Q = (dQ/dt)*t (3) The definition of where Qsat occurred was quite ambiguous, so the method of calculation varied. The first method used stated that the titanium saturated at ten times the initial (or minimum in some cases) pressure. This resulted in two values of Qsat for each saturation, one that used the chamber pressure and another that used the pressure at the beam line. RGA data could also be used to calculate Qsat as ten times the initial partial pressure calculated from 28 AMU, or the mass of a CO molecule. Another method to calculate the gas load at which the system saturated was to determine where the ratio of beam line to chamber pressure equaled one. 6000