Last time the discussion was about throughput and conductance in vacuum systems. This time we will look at the pressure profile throughout the vacuum system in a slightly different way than it was shown last time. The first thought might be that once the vacuum system is under vacuum carrying out the process, the lowest pressure will be in the vacuum chamber and that the highest pressure will be at the primary pump exhaust which will be atmospheric pressure. As we see from Fig. 1, this is not quite correct.
 |
| Fig. 1. Pressure & Throughput in a vacuum system. |
Fig. 1 shows how the pressure changes through the system and actual values of P pressure and S speed are given in the table, Fig. 2. The pressures shown assume that the chamber has been evacuated (pumped down) to the process pressure needed and conditions are stable.
The graph shows values of throughput Q, pressure P and pumping speed S at four points in the system. The dotted lines indicate the point where the values occur in this example. As was stated last time, throughput Q is constant at any point in the system. That means that as pressure P changes in one direction pumping speed S must change in the other direction as Q is constant.
Although the graph shows P1 at the chamber outlet to the vacuum pumps, let’s start at P2 the inlet to the diffusion pump. In North America the American Vacuum Society (AVS) measures diffusion pump pumping speeds at the inlet flange in liters per second (l/sec or l sec-1) , in Europe these speeds are measured one radius above the pump inlet and as a result are slightly lower. For a vacuum system manufacturer it means that they should be cautious when comparing the rated speed of a USA made diffusion pump against one made in Europe.
At Q2 = S2 P2, the inlet to the diffusion pump, the pumping speed is shown as 600 l/sec and the pressure is indicated as 2.0 x 10-6 Torr. As we discussed last time, conductance through piping and accessories reduces the effective pumping speed so the effective pumping speed in the vacuum chamber will be lower than at the pump inlet.
When we look at the table for values at Q1 = S1 P1 at the outlet of the chamber we see that the effective pumping speed has dropped to 200 l/sec due to conductance losses through the piping and the high vacuum valve, and the pressure is higher at 6 x 10-6 Torr.
This shows why the vacuum piping should be as short as possible, to minimize conductance losses. In an efficient design the high vacuum valve and diffusion pump would be mounted as close as possible to the chamber outlet flange. In the case of chambers where the diffusion pump can’t be mounted underneath, the right angle high vacuum valve is mounted directly to the chamber outlet flange, and the diffusion pump directly below the valve. In some case a cold trap is mounted between the high vacuum valve and the diffusion pump. (Cold traps will be discussed in a month or two.)
Returning to position 2, the inlet to the diffusion pump, and following the pressure line into the diffusion pump we see the following. There is a slight pressure drop towards the top jet of the jet assembly, which is the lowest pressure shown in this representation. Then, as the gas stream passes the top jet, it is compressed to a higher pressure between the two jets. At the second jet there is another small pressure drop as the gas molecules become entrained in the oil vapor jet and then the vapor jet compresses the gas stream to the pressure at which it is exhausted towards the primary pump. At this point the graph shows Q3 = S3 P3. From the table the value of S3 is shown 0.06 l/sec (or 3.6 l/min) and the pressure is now up to 2.0 x 10-2 Torr. Pressure P3 is about what would be expected in the foreline of the primary vacuum pump.
 |
| Fig. 2. Pressure and Pumping Speed table. |
The next section of the pressure line indicates a gradual pressure drop in the foreline until the gas stream reaches the inlet of the primary pump, position 4. In most heat treating furnaces the primary pump will be either an oil sealed rotary piston or rotary vane pump. Depending on the size of the furnace these pumps may also have a Roots design vacuum booster mounted on the inlet. The vacuum booster pump develops a very high pumping speed in the pressure range from about 30 Torr to 10-2 Torr. This is the pressure range where most of the water vapor is released from the chamber, hot zone and product surfaces.
At position 4, where Q4 = S4 P4, the pumping speed S4 is shown in the table as 1.2 l/sec (72 l/min) and pressure P4 is shown as 1.0 x 10-3 Torr. This would indicate that the pump used in this example is a two stage rotary vane pump rather than a single stage rotary piston pump that has an ultimate vacuum of about 1 x 10-2 Torr (10 microns).
The final part of the pressure line then shows an initial pressure drop on the inlet side of the rotary vane pump as the gas expands into the void between the rotor and stator, and then a pressure rise as the gas is isolated, compressed up to atmospheric pressure and expelled from the pump on the outlet side of the mechanism.
Following through these steps shows that there are a number of pressure changes through the system as the gas molecules are pumped from the vacuum chamber to atmospheric pressure at the primary pump exhaust.
References:
The two figures used in this discussion are taken from the textbook “Modern Vacuum Practice” (3rd edition, page 70) written and published in the UK by Nigel Harris. Both have been slightly modified to show Torr units.
Copyright Howard Tring, Tring Enterprises LLC Vacuum & Low-Pressure Consulting.