The Molecular Air Pump by Wolfgang Gaede
On the 80th Anniversary of Wolfgang Gaede’s Death By Guido Pfefferle and Gerhard Voss
The inner workings of a molecular air pump in historic images
Wolfgang Gaede died on June 24, 1945, in Munich. With this article, the authors commemorate the 80th anniversary of his death and, for the first time, publish images revealing the inner workings of a molecular air pump by E. Leybold’s successors.
The Patent of Gaede’s Molecular Vacuum Pump
Imperial Patent Office [Berlin] – Patent No. 239213 Dr. Wolfgang Gaede in Freiburg im Breisgau – Rotating Vacuum Pump – Patented in the German Empire as of January 3, 1909
The title of Patent No. 239213 may not sound particularly spectacular at first. However, behind the name "Rotating Vacuum Pump" lies a true milestone in vacuum technology: the molecular air pump.
In the above-mentioned patent specification, Wolfgang Gaede writes: "The molecular air pump utilizes solely the friction between the gas being pumped and a rapidly moving solid surface to transport the gas. The use of a 'sealing fluid', such as mercury or oil, is not required." In today’s terminology, this means: Wolfgang Gaede’s molecular air pump was the world’s first dry-compressing vacuum pump.
Figure 1 [3] shows, on the left, the molecular air pump built by Leybold according to Gaede’s patent, here shown evacuating an X-ray tube. Naturally, a rotary vane pump from Leybold is used as the fore-vacuum pump.
The Operating Principle of the Molecular Air Pump
In his habilitation thesis [1], Wolfgang Gaede introduced the term “external friction of gases”, describing the interaction of gas molecules with a rapidly moving solid surface. His molecular air pump operates based on this very principle. A schematic representation of the pump’s operating principle can be found in the Preliminary Communication on a New High-Vacuum Pump, published in 1912 by E. Leybold’s Nachfolger [2].
The “Figure 2” included in [2] is reproduced in this article as Figure 2. It is accompanied by the following original text: “Grooves of depth b and width a are cut into the cylinder A, which rotates around axis a. At a distance h', A is enclosed by a cylindrical housing B. On one side, a lamella comb C, which is attached to housing B, protrudes into the grooves.” [In Figure 1 (left), gas is transported from n to m when rotor A rotates clockwise at high speed around axis a. An undesired loss in gas flow occurs when gas flows back from m to n through the gap between C and A. In the technical realization of the pump, this gap must therefore be no more than a few hundredths of a millimeter wide.] [To achieve the best possible high-vacuum, the gas must be significantly compressed between the pump inlet (the high-vacuum side) and the outlet (the fore-vacuum side). This is accomplished by the following principle:] “The individual grooves are connected in series, such that opening m connects to n₁, m₁ to n₂, and so on. As a result, the gas pressure continuously decreases from the ends of the rotor toward the center.”
The Technical Design by Leybold
Figure 3 [3] shows the technical design of Leybold’s molecular air pump in a longitudinal section along the rotor axis a. The housing B, shown hatched in Figure 3, supports the upper assembly K and is fastened to it in an “airtight” manner. The rotor A, made from a solid brass cylinder, is rigidly connected to the axis a. Grooves D are milled into the brass cylinder, into which the lamella comb C (dark hatching) extends. Additionally, S indicates the pump’s inlet on the high-vacuum side, while H marks the pulley used to drive the axis a. It is worth noting that the true secret of the pump lies within the upper assembly K. It contains a complex system of gas distribution channels—of which no drawings or photographs exist.
A Look Inside the Leybold Pump
To explore the inner workings of the molecular air pump, we first removed the four screws securing the upper assembly K. Once loosened, the assembly could be lifted off the housing B—and we were truly surprised. There was no gasket between the assembly and the housing, just brass on brass with a bit of grease.
Figure 4 shows both the underside of the flipped-back upper assembly K and the top side of housing B. We will take a closer look at the underside of K in Figure 6. On the top side of B, one can see a series of slots that connect to the interior of B—and thus also to the grooves in the rotor. Additionally, the lamella comb C is mounted on the top side of B, aligned parallel to the longitudinal axis of the pump. The holes for the mounting screws are also visible in Figure 4. After removing the components E, F, G, and H shown in Figure 3, we were able to extract and measure the rotor A. According to our measurements, it has a diameter of 100.00 + 0.01 mm. To demonstrate the interaction between the lamella comb and the rotor, we inserted the comb—milled from a single piece of brass—into the grooves of the rotor. This can be seen in more detail in Figure 5.
For the mechanism to function as shown in Figure 2, fine mechanical precision and reproducibility within hundredths of a millimeter are absolutely essential. Leybold was already capable of achieving this level of precision back in 1912.
With Figure 6, which shows the underside of the upper assembly K in detail, we move closer to the “secret” inner workings of the molecular air pump. Using a metal cleaning fluid, we traced the internal channels.
This allowed us to follow the complex path of the gas between the high-vacuum (HV) and fore-vacuum (FV) sides in detail: The inlet on the high-vacuum side (HV connector S, top left in Fig. 6) is connected to position 1 in Fig. 6. From there, the rotor transports the gas to position 2 on the right. This means that the rotor—when viewed from the labeled side of K (E. Leybold’s Nachfolger, Coeln and Berlin, German Imperial Patent)—must rotate counterclockwise.
Through the visible channel filled with soft solder, the gas moves from 2 right to 2 left, then via the rotor to 3 right, through a channel in K to 3 left, through the rotor to 6 right, via a channel in K to 6 left, through the rotor to 4 right, via a channel in K to 4 left, through the rotor to 7 right, via a channel in K to 7 left, through the rotor to 5 right, via a channel in K to 5 left, through the rotor to 8 right, and finally through a channel in K to 8 left, from where it reaches the annular groove (FV position) that is connected to the fore-vacuum pump.
Apologies for the lengthy explanation—but after 113 years, it had to be written down and documented. Thus, we were able to experimentally confirm Wolfgang Gaede’s statement that the gas is extracted from the center of the pump. The exact structure of the internal channel system in K can likely only be determined non-destructively using X-rays. It’s important to note that the groove connected to the fore-vacuum pump surrounds the inner high-vacuum area in a ring.
This design ensures that any air leakage from the environment (1000 mbar) to the groove (0.1 mbar) is intercepted by the fore-vacuum pump. The pressure difference between the annular groove and position 1 is typically 10,000 times smaller than the pressure difference between the ambient atmosphere and the groove. As a result, leakage from the groove to position 1 is far smaller than from the environment into the groove. In summary: The annular groove protects the high-vacuum area from air leakage coming from the ambient atmosphere.
Epilogue
Leybold can consider itself fortunate that two original molecular air pumps are preserved in the Gaede Archive. This fortunate circumstance inspired the idea to restore one of the pumps to working condition. This has been successfully achieved—but there is still work to be done to fully optimize the pump’s operation.
References
[1] Wolfgang Gaede Habilitation Thesis: The External Friction of Gases University of Freiburg im Breisgau, 1912 [2] Preliminary Communication on a New High-Vacuum Pump (Molecular Air Pump) according to Dr. Gaede E. Leybold’s Nachfolger, Cöln a[m] Rh[ein], 1912 [3] Special Price List No. VI on Molecular Air Pumps according to Dr. Gaede E. Leybold’s Nachfolger, Cöln a[m] Rh[ein], 1912
Authors:
Guido Pfefferle
Prototyping and Tooling E-Mail: [email protected]
Dr. Gerhard Voss
Gaede Archive Cologne
E-Mail: [email protected]
