by Dirk A. Boonstra
All Rights Reserved


Review and historical development

During the 17th century vacuum investigation moved from a pure philosophical approach to the area of experimental physics. The start of the development of vacuum technology was advanced by genius explorers, like Otto von Guericke (1602-1686) and Evangelista Torrcelli (1608-1647) (He did not live very long, maybe because of the mercury?) Well known is Otto von Guericke's experiment: a ball, existing of two half spheres, which could be evacuated, was pulled apart by two teams of horses of 8 horses each. With this experiment for the first time the force of air pressure was proven. (Actually, the test with the 16 horses required several trials. For more than 6 times the test was repeated, only because it is impossible that one man can control all these horses at the same time. In fact it was a stunt, because two teams of two horses each can do the job much easier. You can control these 4 horses perfectly so that they are pulling all together at the same interval. Each time the test is done with only 4 horses, the two half spheres pull out of each other with an enormous big bang.)

The original Magdenburger half spheres together with the original pump from Otto von Guericke still exist in real life and you can see both in the Deutsches Museum in Munich.

In 1643 Torricelli invented the mercury manometer. A glass tube with one side closed, with a length of approx. 1 meter, is filled with mercury and is placed with the open side down into a mercury filled box. The height of the mercury column in the glass tube shows you the exact atmospheric pressure at that time. Blaise Pascal (1623-1662) proved, with this instrument, the fact that the air pressure at the top of high mountains is lower than in the valley just above sea level. He made his first test at the top of the Pyrenees in France on September 09,1648.

Robert Boyle (1627-1691) was experimenting over a long period of time with the mercury manometer, and he developed the barometer from it.

In The Netherlands Van Musschenbroek and Gravenzande were conducting several technical vacuum experiments and they were building several vacuum pumps. Examples of these pumps can be found in the Teylers Museum in Haarlem and in the Boerhave Museum in Leiden. Both cities are located in The Netherlands.

In the second part of the 19th century, before new incentive came for further developments in vacuum technology. Herr Bunsen, well known for the Bunsen torch (or burner), developed the waterjet pump, and McLeod invented the compression manometer in 1874. At the same time Topler and Sprengler were building their mercury filled vacuum pumps.

The first technical application of vacuum technology came at the end of the 19th century through the experiments of Roentgen, and by Edison due to the research of the incandescent lamp which he developed in 1879.

The researches of Gaede, Langmuir, Knudsen and Pirani at the beginning of the 20th century was the beginning of all modern vacuum technology. At that time the basis of present day vacuum physics and vacuum technology was founded.

In the Philips laboratories in Eindhoven Doctor Penning devised the well known Penning manometer. It is in fact the principal of this manometer which was further developed into the ion getter pump during the second world war in the USA.

In 1956 Gifford and McMahon were working toward the development of the cryopump - a pump that came to the market in 1974 because of a lack of interest for this pump type during the fifties.

The constantly advancing technology of electronics made it possible that in the year 1980 an absolute standard vacuum meter was built. The technical realisation of complete oil free vacuum plants was accomplished in 1987 by the introduction of the oil free fore (rough) vacuum pumps.

The developments of vacuum technology will be best illustrated by the following facts:

During the year 1654, it cost Von Guericke several days of hard labor with his water-tight pumps, (in fact he was using rebuilt fire department pumps), to achieve a pressure of 20 Torr.

In about 1900, only a few physicists, using very complicated hard to handle vacuum apparatus, were able to achieve a pressure of about 1x10E-4 Torr within one day.

In the year 1940, the technical limit of an end pressure was about 1x10E-6 Torr. One of the most important applications was the production of radio tubes.

Today it is possible to achieve within a few hours, with easy to handle apparatus, a pressure lower then 1x10E-11 Torr. And we can measure these low pressures as well.

These comparisons show you clearly the enormous progress of vacuum technology in this century. And the improvement, whether we are talking about diffusion or dry pumps, is still going on.

Fundamental principles of the vacuum technology.

In vacuum technology we concerned with gasses. These gasses bring about a sudden force to the walls of the receptacle. This pressure can be understood by noting that in a gas each separate gas particle can move freely. Therefore they are hitting very often and at different angles against the wall of the receptacle - depending on the density (number of particles) and their temperature (speed of the particles). The angle at which the particles are reflecting from the wall is random, so the particless are not following the rule of incoming angle equals outgoing angle. (It is possible, of course, that the outgoing angle can be the same as the incoming angle but that is a random situation as well).

The total result of all these separate collisions creates a sudden force to the area of the wall. The definition of the force can be given by the following equation:

force divided by the square = pressure, or:

P= F / A

The definition of vacuum is pressure that is below atmospheric pressure. This is important as you can see that the unit of force and pressure as well, are the same - namely the Pascal (Pa). The pressure of force per area is measured in Newtons per square meter. This unit is the Pascal (Pa).

1 x Pa = 1 x N / (m x m)

One atmosphere (bar) is 10E+5 Pa which equals 10E+5 x N / mxm

Other well known units are bar, mbar, Torr and mm Hg. We understand that 1 Torr is the pressure which causes the mercury column to rise 1 mm in the tube of Toriicelli. In older catalogues and publications people will find the unit of Torr mm Hg as well. Due to the fact of international normalization, the legal unit for pressure will be the Pa(scal). The Torr or mm Hg here in Europe is not used. We mostly use the unit of mbar while the progressive firms and schools are now using the Pa.

The calculation is as follows:

1 bar = 1000 mbar = 750 Torr = 10E+5 Pa therefore:
1 mbar = .750 Torr = 10E+2 Pa or:
1.33 mbar = 1 Torr = 133 Pa

The advantage of using the mbar in vacuum technology is in the fact that Torr and mbar for low pressures are the same. One times nearly nothing is the same as 1.33 times nearly nothing. So:

1 mbar = .750 Torr, or 1.33 mbar = 1 Torr but in practice, 10E-7 Torr is the same pressure as 10E-7 mbar. We cannot even measure the difference between these two. Only when the pressure is higher then 1x10E0 mbar does the difference begin to be important.

We will use from here on only the unit of mbar.

Today's technical working range of vacuum is between 1 bar and approx: 10E-13 mbar and spans therefore a range of 16 decades.

Based on pressure the different ranges are named:

from 1000 mbar to 1 mbar - low (fore) vacuum range from 1mbar to 10E-3 mbar - middle vacuum range from 10E-3 mbar to 10E-7 mbar - high vacuum range (HV) lower then 10E-7 mbar - ultra high vacuum range (UHV)

Another set of ranges which is important to know, is the based on the different pressures combined with conduit diameters. To help calculate the conductance C of a conduit (or tube), we distinguish 4 rates of gaseous flow according to the gas pressure and the tubing diameter.

These four rates of gaseous flows are:

1. The turbulent rate - this rate is close to atmospheric pressure and can reach 8 Beafort in the glass tube for a very short period of time.

(Note: The Beaufort Scale is a measure of wind force. The scale runs from 0-17, 0 being calm, 12-17 being hurricane and moderate being 4. A Beaufort Scale reading of 8 would be equivalent to approximately 39-46 mph wind speed.)

2. The viscous or laminar rate- this rate is at average pressure.
3. The intermediate rate.
4. The molecular rate - this rate is at low pressures.

Above rates are indicated as follows:

Turbulent rate: p x d > 100 mbar cm
Viscous rate: 100 > p x d > 1 mbar cm
Intermediate rate: 1 > p x d > 10E-2 mbar cm
Molecular rate: 10E-2 > p x d mbar cm

d = the inner diameter of the tube or conduit in cm p = the average pressure in the tube or conduit in mbar

During the turbulent rate there is a stream of moving particles, like dust and fluorescent powder parts, which is not desirable. In this case, pumping via a restriction is the solution.

For that same reason we never fully open the main stopcock at once when we are beginning to evacuate a neon system. Open the stopcock (or valve) slowly to prevent the above mentioned turbulence. Open the stopcock fully when the pressure is at the viscous rate.

Looking at the above, it is clear that using a so called high speed pump - which works on an enormous amount of liters at atmospheric pressure - is not preferable, as the coating can come loose from the wall or there is evaporation from pockets which were filled with air between the glass wall and the fluorescent layer. No one should or can do the job faster than is technically possible. Besides, opening the main stopcock partially will only add a few extra seconds a day. What will you do with these few seconds. Go for a holiday?

In practice, the final or end vacuum after degassing and activation consists of lowering the pressure in the neon tube from a few mbar to at least 1x10E-3 mbar. We prefer, for more than one reason, 1x10E-4 mbar especially when we are making the so-called cold cathode tubing, or contour lines with a diameter of 18 to 19 mm and a with a length that is minimally two meters for outdoors.

Taking into account the usual dimensions of the manifold (well, at least "my manifold") we can almost ignore the resistance of it when calculating gaseous flow.

For neon shops, who are working with their vacuum equipment in the intermediate range, we will produce the formula for that pressure range. It is:

C = 12.1 x ( (d x d x d)/L ) x J = liters/sec

In this formula:
C the conductivity
d the inner diameter of the tube in cm
L the length of the tube in cm
12.1 is a constant number
J is a correction factor according to the pressure varying from 1 to 16, but it is not linear with the pressure. At 1x10E-1 the number, for example, it is 4.

The conductivity is given in the units of liters per second.

However, as we will ignore this formula, because of the pressure range it represents, we will also ignore the neon shops who can deal or must calculate with it. The end pressure achieved in the intermediate range is far from sufficient to produce quality neon tubing. (At least that's what we think here in Europe).

So, we will calculate further on with the formula that is given for molecular gas flow. This formula is independent from pressure and is the most important and used formula in the low pressure mercury discharge industry, for evacuating luminous tubes.

C = 12.1 x (d x d x d / L) and is given in liters/sec.

Let's give an ordinary example for the bombarding method:

(C1) The overall length of a given manifold is 35 cm with an inner diameter of 20 mm.
(C2) The overall length of the tubulation pipe is 15 cm with an inner diameter of 4 mm.
(C3) The overall length of the two throats built to facilitate cutting are 1.5 cm long with an inner diameter of 2 to 2.5 mm. (C3)

(Note: In Europe, it is commonplace to utilize a single head hand torch for tipping off tubulation glass. The fact that they are using "capped" electrodes also means it is essential to tip off the tubulation rather close to the cap. Commonly, the tubulation glass from the unit is reduced in one or two places running to the manifold for ease of tipping off. One reduction or "throat" for neon filled units - near the cap, and two for argon/mercury filled units - one near the cap and one for the manifold side of the mercury trap).

The conductivity of the manifold is:

C1 = 12.1 x ( 8 / 35 ) = 12.1 x .23 = 2.77 liters/sec

We calculate for the conductivity of the tubulation pipe:

C2 = 12.1 x ( .06 / 15 ) = 12.1 x .004 = .05 liters/sec

We calculate for the conductivity of the two throats:

C3 = 12.1 x ( .01 / 1.5 ) = 12.1 x 0.007 = 0.08 liters/sec

Analogous to electricity laws, we can calculate the resulting conductivity if the several different conductivities are known.

The resulting formula is given below:

1 / Ctotal = ( 1 / C1 ) + ( 1 / C2 ) + ( 1 / C3 )

1/Ctotal = ( 1 / 2.77 ) + ( 1 /.05 ) + ( 1 /.08) resulting in:

1/Ctotal = .36 + 20 + 12.5 = 32.86

The conductivity C of the complete conduit is:

Ctotal = ( 1 / 32.9 ) = .0303951 l/s

So the total conductivity of the manifold with a tube mounted to it with all the rest is:

0.03 l/s

This is the best conductivity you can achieve with an excellent manifold in superb condition and it is the best that the system can do for you. Whatever sales people are telling you about their sophisticated pumping stations with an enormous high speed pump, the above mentioned 0.03 I/s is a fact of life. But we are not finished yet! The formula for calculating the effective pumping speed of the HV pump installed in the pumping station is analogous to the above:

1 / Seff = ( 1 / Spump ) + ( 1 / C )

This formula guarantees that the real pumping speed will always be less than that of the pump.

Let's say we buy a pump with a pumping speed of 500 liters/sec.

We can calculate as an end result:

1 / Seff = ( 1 / 500 ) + ( 1 /.03) resulting in:

1 / Seff = .002 + 33.33 = 33.332

At this time we can already see clearly that the effective pumping speed is only dependent on the conductivity i.e. the resistance of the system and the tubes with their capillaries (tubulations) mounted to the manifold.

Let's look at the end result:

Seff = 0.03 liters/sec and that is exactly the same as the conductivity of the system. So the installed pump capacity in liters/sec is not of any importance.

The pumping speed is of no importance, but on the other hand the pump must Be capable of reaching the desired end vacuum and that is of the highest importance. A given pump that can pull by PNEUROP regulations, for example, to 1x10E-4 mbar as an end pressure is not sufficient to quickly produce proper neon. It will cost you too long a time to reach that pressure (if you can at all reach it anyway). In practice, it not possible that anyone, in terms of how the neon industry evacuates, can reach the end pressure that the brochure is telling you that you can reach.

Let's calculate another pump type that is rated 50 liters/sec - a pump with a ten times slower pumping speed.

For this pump the effective pumping speed in the tube is:

Seff = 0.03 liters a second which is the same end result as a pump installed with a ten times bigger capacity.

For our last example we take what competitors are saying, is too small: A pump that can only do 7.5 liters/sec. That is a pump of almost 70 times lower capacity then the first one we did calculations for.

Installed, this pump's effective pumping speed is, and I am not surprising you if you have read the above:

Seff = 0.03 liters/sec


If we do the calculations again without the conductivity of the two throats we can see that our pumping speed is nearly doubled. So for that reason it is without doubt that it is very important that these throats must be made as wide, and short as possible. As we are making two throats with an overall length of 3 cm and an inner diameter of 1 mm, our effective pumping speed, whatever pump is installed, will only be....... (Come on folks, give me the correct pumping speed by fax or yell it into my ear over the phone.) Are you surprised?

The total pumping speed of a neon pumping station depends on the quality and length, with the amount of valves mounted in, of the manifold. It is the achieved conductivity and not the capacity of the installed high vacuum pump, whatever the brand is.

Pumps with a capacity of 70 times lower gives you the same end result in the bombarding method as we showed very clearly by these figures.

It is of more importance that you clean your manifold frequently, because of conductivity, and not use a mercury manometer (for different reasons).

So, the most preferable pump, in my opinion, to achieve high vacuum for luminous tubes is a dry pumping system that can operate without heaters or cooling water or even an air blower.

These systems can do their job only consuming 160 or 60 watts, depending on the pressure range the system is operating in, instead of the normally installed 1 kW for the rotary vane pump with the electrical heated boiler of the diffusion pumps. The pump I prefer is only drawing 55 watts when operating in the molecular pressure area. Still, the pump must be capable of 1x10E-6 mbar under lab conditions. These systems can operate at least 10 times cheaper for you, consuming less power, which is of importance especially with the prices we must pay for electricity and the cooling water here in Europe.

Dirk A. Boonstra owns de Blaecker in the Netherlands. He is also distributes for Tecnolux. He has been a frequent contributor to European Sign Magazine and he teaches neon at the school in Delft.

Editted by Kenny Greenberg
Special Thanks to Tom Unger for his explanatory notes.