Capacitance-based absolute pressure gauge for beamline vacuum measurement applications

For low-pressure measurements, capacitance diaphragm-based absolute pressure gauges are used owing to their high accuracy, media independence, and simplistic design. Catering to the general requirements of synchrotron beamline vacuum measurements, a capacitance diaphragm-based absolute pressure gauge is designed, developed, tested, and calibrated. The sensor is having electrode assembly which forms a dual capacitor with the gauge body. The two capacitances are used in post-processing to reduce the noise levels and to enhance the precision and accuracy of the measurements. The reference side is at high vacuum of the level 5 × 10-5 Pa. The vacuum of the reference side is generated using a Turbo Molecular Pump and is sustained by pinching-off the reference side at this vacuum. A getter pump is provided and energized for improving the vacuum which otherwise gets deteriorated due to outgassing in the reference region of the gauge. The fabrication methods adopted for fabricating the gauge are novel. The diaphragm is fabricated from bulk material using the electrical discharge method (EDM). This helped in achieving a perfect diaphragm with linear deflection with pressure variation and attained a 100% success rate. Unlike the fabrication by welding of plate, this method is robust, economical, and helped in achieving similar properties across the gauges. The second innovation is in the fabrication of electrode assembly. Unlike commonly used electrodeposited electrodes; the presented technique used copper plates which sandwich the Alumina dielectric. The plates are bolted/brazed/welded across the ceramic disc with the help of pins which also help in the parallel connection of the dual capacitors are described later. This made this method of fabrication largely rejection free with the potential of repairs and testing with different plate configurations. This paper is presenting the design, production, and calibration of the gauge using these fabrication techniques for 1 × 103 Pa to 1 × 105 Pa.


Introduction
Low-pressure measurement sensors have numerous applications in industry and research laboratories, specially in beamlines.Even high vacuum systems need low vacuum level gauges for generating TMP (Turbo Molecular Pump) switch-on signals.Diaphragm-based capacitive pressure sensors are simple, robust, accurate, and media-independent and measure only the pressure exerted by residual gas particles [1].
Figure 1 shows the basic sketch of the capacitance-based diaphragm gauge.It consists of two hermetically sealed and isolated regions called the process side (one having media to be measured) and the reference side (maintained at low pressure ∼ 1 × 10 −5 Pa).A thin diaphragm isolates these two regions.The diameter and thickness of the diaphragm determine the key parameters of the gauge like range, sensitivity, and dimensions.The process side is connected to the process under measurement via KF (Klein Flansche), or CF (Con-flat) flange.The reference side consists of the capacitance electrode assembly, and getter assembly and is connected to the TMP by a pinch-off tube for evacuation to pressure of the order of 1 × 10 −5 Pa.
The electrode assembly forms two sets of capacitors with respect to the diaphragm (gauge body) as shown in figure 2.
The two capacitances change when the diaphragm, as shown in figure 2 deflects under the influence of differential pressure.This causes the average distance between the diaphragm and electrode to decrease.Reduction in distance increases the capacitance as per the following formula  C is the capacitance between the diaphragm and the electrode, A is the area of the electrode, and d is the distance between the electrode and the diaphragm.This formula is applicable to parallel plate capacitance with hard boundary approximation.However, is sufficient to explain the behavior of the capacitance diaphragm gauge.
Capacitance is a monotonous function of applied pressure.As seen in the above figure, each side of the electrode is having two copper plates in the form of a central disc and an outer ring that sandwiches the ceramic disc.The outer ring facing the process side is connected to the central disc facing the reference side and the outer ring facing the reference side is connected to the central disc facing the process side.These connections are made using copper pins through the ceramic discs.Refer to figure 2 for more clarity.The two capacitances are named C x and C y .The capacitance of both these capacitors increases as pressure on the process side increases.The rate of change of these two capacitances is different and the pressure (P) on the process side is measured in terms of -2 -the ratio (A) of these two capacitors as given in equation (1.2) 2) The measurement of the ratio ensures the minimization of ambient noise in the measurement.The working of the absolute pressure gauge needs to be discussed from a slightly different perspective.
Since the reference side is at high vacuum, when the process side is exposed to the ambient (atmospheric pressure), the diaphragm is deflected proportionally to the ambient ∼ 1 × 10 5 Pa (may defer depending on geographic location).When the process side is attached to the media for measurement, the diaphragm further deflects if the applied pressure is more than ambient or relaxes if the applied pressure is less than ambient.The two capacitances are electronically measured.The very nature of capacitance ratio A with respect to applied pressure is not a linear function.Thus, the electronics need to hold the calibration data and carry out interpolation based on measured ratio A, to display and/or retransmit the measured pressure in terms of the current signal or voltage signal.Low pressure of the order of 1 × 10 −5 Pa is generated by Turbo Molecular Pump via the pinchoff tube.The pinch-off tube is a soft annealed copper tube with is welded to the reference side of the gauge on one side and connected to TMP on the other side via the CF flange.Subsequent to evacuation and required bake-off the pinch-off tube is cold welded using a special pinch-off tool.The capacitance electrodes connections are taken to the ambient for connection to measurement PCB (Printed Circuit Board) through glass metal seals.The heater-embedded nonevaporable getters are also taken out to ambient via glass metal seals.The getter is activated, after pinch-off, after allowing sufficient outgassing by passing the required electrical current.A serious limitation imposed by such gauges is the range versus sensitivity trade-off.The optimization of diaphragm thickness, radius, and electrode (radius r 1 , r 2 , and r 3 in figure 2) is not in the scope of this paper.
Two main challenges of fabricating CDG (Capacitance Diaphragm Gauge)-based vacuum gauges, discussed in this paper, are the fabrication of a hermetically sealed diaphragm with better than 1 × 10 −12 Pa m 3 /sec leak rate; linear deflection with applied pressure and the fabrication of the electrode assembly.Both these problems are resolved in the present design.A 100% acceptance rate is achieved in about ten gauges developed so far.

Building blocks
The sketch of the capacitance diaphragm gauge showing the basic building block is shown in figure 1.These include Process side sub-assembly, Electrode sub-assembly, diaphragm sub-assembly, and interface plate.PCB for data acquisition, processing, and re-transmission is not shown in the figure but is an important sub-system of the gauge.a. Process side sub-assembly.This is the interface between the media being measured and the vacuum gauge and has a cavity where the diaphragm is exposed to the media.The diaphragm is pulled away from the electrode depending if the pressure is less than ambient and vice-versa.It has KF or CF flange for connection to the process. -

Fabrication of diaphragm
In conventional capacitance gauges, the diaphragm is realized using a thin sheet/foil of the material which is welded to the bulk material.Ensuring a linear relationship without abrupt step deflection is a big challenge for such constructions.The success rate is limited, and the expertise of the welder plays an important role in the success of the gauge.Maintaining equal pressure along the azimuthal symmetry is a non-trivial task.
In the proposed method, the diaphragm is fabricated from bulk material using the EDM fabrication technique.Diaphragm was successfully fabricated using both EDM wire cut and spark erosion process.Both processes attained 100% success and acceptance.The limitation of this method is the inability to make a thinner diaphragm (≲ 250 microns).This however is not a handicap as the usage of thicker diaphragm have high sensitivity.Figure 3 and figure 4 show the photograph of the diaphragm developed using the spark erosion method and figure 5 shows fabrication using EDM wire cut.Hermetic sealing of the gauge on the process side and reference side is achieved using laser welding with better than 1 × 10 −12 Pa m 3 /sec helium leak rates.
The qualification of the diaphragm is based on its geometric measurements, MSLD leak test, and deflection versus pressure response.Geometric qualification includes the measurement of diaphragm thickness, the parallelism of two opposite surfaces of the diaphragm (process side and reference side), and the parallelism of the diaphragm reference side plane to the sitting surface of the electrode.Thickness measurement is done using an ultrasonic tester and Coordinate Measuring Machine while the other two geometric parameters are measured using CMM.The thickness variation of the diaphragm is less than 20 microns.Leak tightness is measured by a Helium MSLD (Mass Spectrometer Leak Detection) leak detector.The acceptance value of leak detection better -4 -   The smooth deflection of the diaphragm as applied pressure is varied is the most crucial achievement of this method of diaphragm fabrication.All developed gauges showed similar behavior.Figure 6 shows the deflection versus pressure variation of the gauge.-5 -

Fabrication of electrode assembly
Conventionally, electrodes of such gauges are realized using electro-deposition of metal over the ceramic disk.Masks of required shapes are prepared and electrodeposition is usually carried out in multiple steps.This process is a controlled process and needs expertise in executing it.Another challenge in realizing it is the interconnection of the metal deposition across the ceramic disk.
The proposed method has simplified this to a problem of routine precision fabrication with a 100% success rate, versatility, and reliability.Figure 7 shows the sketch of the proposed electrode assembly.The proposed design uses a ceramic disk sandwiched by copper plates in the form of a disc and ring, one type each on either side, as an electrode.Figure 7 shows the top side and bottom sides of the electrode assembly.Each side is having one outer ring and an inner disc.Overall four capacitances are formed with the body of the gauge.The inner disc on the top side is connected to the outer ring on the bottom side.This is like a parallel connection of two capacitances.This -6 -overall capacitance is termed C x .Similarly, the outer ring on the top side is connected to the inner disc on the bottom side.This capacitance is termed C y .These two capacitances are brought out in the ambient with the help of glass metal seals as shown in figure 1.This has simplified the design, removed dependency on the sensitive physics process of electrodeposition, eased bulk production, gave freedom for design trials and design optimization, and helped in achieving 100% acceptance rates.
Figure 8 and figure 9 show the photograph of the ceramic disk and fully assembled electrode assembly respectively.

Electronics
The capacitance measurement, data acquisition, processing, linearization, and re-transmission are carried out using an electronic circuit, designed specifically for this purpose.The block diagram of the electronics system is shown in figure 10.Electronics is developed for high dynamic range using capacitance measurement IC (CAV444) in the analog domain.Two-channel capacitance measurement was carried out, followed by digitization using ADC (Analog to Digital Converter) and processing in the microcontroller.The calibration curve was fed into a microcontroller for a direct display and retransmission of the pressure being measured.Re-transmission was provided in 4-20 mA and 0-10 V corresponding to the full range of measurement.The PCB board is having a temperature sensor which is also read by the microcontroller and was used for temperature calibration of the pressure gauge.The CAV ICs also generate voltage proportional to the IC temperature, it was used to correct the drift in capacitance measurement that arose due to temperature variation.Figure 11 shows the developed electronics.The PCB has an onboard display that is configured for the display of absolute pressure.The electronics is having 4 user-configurable digital I/O pins which can be used to generate interrupts based on the pressure reading.

Noise floor analysis
The measurement of electronics noise floor was carried out at fixed process pressure.The two capacitances C x and C y were measured for about 20 minutes duration in a controlled environment.Figure 12 shows the time domain C x , C y , and C y /C x measurement in this duration.The value of capacitance for both channels shows a spike at around 4000 data locations.By virtue of using the ratio of these two capacitances, this spike is absent in the ratio data.This demonstrates the usefulness of using the ratio as a measure of the pressure.The usage of capacitance for pressure measurements would have resulted in a false spike at this location.During this measurement, the pressure was also measured using a standard gauge which doesn't display any spike.
Figure 13 shows frequency domain noise floor of deviation from the mean for C x , C y , and C y /C x .The noise floor of ratio is significantly lower than the noise floor of the individual capacitances.

Calibration
Figure 14 shows a snap of the developed gauge.The gauge needs calibration for correct pressure display and retransmission.The applied pressure is measured in terms of the ratio of two capacitances.A closed-form expression is not possible to get the output pressure as a function of the capacitance ratio, as capacitance is highly sensitive to the gap and area of the plates.Prior to calibration the reference side is evacuated and maintained at a vacuum better than 1 × 10 −5 Pa.The gauge is baked for enhancing the vacuum performance.After maintaining the vacuum for about 15 days, the pinch-off tube is cold welded while maintaining the required vacuum levels.The cold-welded portion is suitably embedded in vacuum-compatible epoxy for safe keeping the cold weld from mechanical damage.After the pinch-off, the gauge is calibrated.Calibration is carried out under controlled temperature and humidity.A standard absolute gauge for measuring the true value and needle valves for varying the vacuum levels on the process side was used for calibrating the vacuum gauge.
Figure 15, figure 16, and figure 17 show the values of C x , C y , C y /C x as a function of applied pressure for the gauge having an upper range limit (URL) of 750 mbar.
-8 -     the microcontroller of the electronics.Pressure is calculated by spline interpolation of calibration data.The response of duly temperature compensated and calibrated gauge with respect to the standard gauge for the range of applied pressure is shown in figure 18.

Figure 2 .
Figure 2. Electrode with reference to the diaphragm.

Figure 3 .
Figure 3. Full view of the Diaphragm sub-assembly fabricated using EDM spark erosion.

Figure 4 .
Figure 4. Close-up of the diaphragm fabricated using EDM spark erosion.

Figure 5 .
Figure 5. Diaphragm (seen from the process side) fabricated using EDM wire cut.

Figure 6 .
Figure 6.Deflection of the diaphragm as a function of applied pressure.

Figure 7 .
Figure 7. Sketch of the electrode sub-assembly.

Figure 8 .
Figure 8. Ceramic disk with grooves for holding copper electrodes.

Figure 10 .
Figure 10.Block diagram of the vacuum gauge's electronics.

Figure 11 .
Figure 11.Developed electronics with onboard pressure display.

Figure 12 .
Figure 12.Noise floor of C y (top), C x (middle) and C y /C x (bottom).

Figure 13 .Figure 14 .
Figure 13.Frequency domain noise floor of C x , C y and C y /C x .

Figure 15 .
Figure 15.C x , its coefficient of variation, and maximum deviation from mean.

Figure 16 .Figure 17 .
Figure 16.C y , its coefficient of variation and maximum deviation from the mean.

Figure 18 .
Figure 18.Measurement of developed gauge with standard gauge for range of applied pressure.
3 -b.Diaphragm sub-assembly.This contains the diaphragm and associated mechanical structures.It contains the required interfaces for the placement of electrode sub-assembly.The mechanical interfaces ensure the required gap between the diaphragm and electrode.The gap determines the range of the gauge.c.Interface plate.It contains the required interfaces between the gauge and the world.The interfaces include pinch-off tube connections, vacuum feed-throughs for capacitance measurements, and vacuum feed-through for getter pump heaters.It also contains required tapped holes for mounting the PCB.d.Electrode assembly.It contains copper electrodes in the form of circular annulus rings and central discs sandwiching the ceramic disc.It makes dual capacitors with the diaphragm body and the ratio of these capacitors is taken as the measure of applied pressure.
e. Electronics unit.All measurements, acquisition, signal processing, interpolation, and generation of re-transmission signals are carried out in the electronics unit of the gauge.