The pressure sensitivity of wrinkled B-doped nanocrystalline diamond membranes
At a Glance
Section titled “At a Glance”| Metadata | Details |
|---|---|
| Publication Date | 2016-10-21 |
| Journal | Scientific Reports |
| Authors | Sien Drijkoningen, Stoffel D. Janssens, Paulius Pobedinskas, Satoshi Koizumi, Marlies K. Van Bael |
| Institutions | National Institute for Materials Science, Hasselt University |
| Citations | 22 |
| Analysis | Full AI Review Included |
Executive Summary
Section titled “Executive Summary”- The study demonstrates 200 nm boron-doped nanocrystalline diamond (B:NCD) membranes as piezoresistive pressure sensors for operation up to a 0.8 bar differential pressure.
- Sensitivity improves with larger membrane diameter, higher boron concentration, and a substrate whose coefficient of thermal expansion closely matches diamond.
- Membranes fabricated on Corning Eagle 2000 glass showed 55% higher sensitivity than nominally similar membranes on Schott AF45 glass, because lower residual compression produced less wrinkling and more efficient stress-to-resistance conversion.
- The sensor architecture integrates the piezoresistive diamond membrane with a Hall-bar structure, simplifying fabrication relative to attaching discrete piezoresistive elements to a non-piezoresistive diaphragm.
- Diamond’s corrosion resistance, radiation hardness, thermal conductivity, and biocompatibility make these membranes suitable for harsh, electrochemical, biomedical, and radiation-exposed environments.
- A further sensitivity increase is expected by selecting a substrate with a thermal expansion coefficient close to, or below, that of diamond while maintaining adequate film adhesion and avoiding delamination.
Technical Specifications
Section titled “Technical Specifications”| Parameter | Value | Unit | Context |
|---|---|---|---|
| Diamond material | Boron-doped nanocrystalline diamond | — | Piezoresistive membrane and Hall-bar sensing element |
| Film thickness | 200 | nm | B:NCD active membrane |
| NCD grain size | 180 ± 10 | nm | Estimated from SEM |
| Pressure type | Positive differential | — | Membrane inflation against ambient conditions |
| Maximum differential pressure | 0.8 | bar | Reported wide-range operation |
| Typical B:NCD gauge factor | Up to 30 | — | Literature value for NCD thin films |
| Boron process-gas ratio | 10,000 | ppm | TMB-to-methane ratio |
| Process-gas methane content | 3 | % | In hydrogen plasma |
| Microwave power | 2,100 | W | ASTeX 6500 series MWPECVD system |
| Process pressure | 11 | Torr | Growth atmosphere |
| Substrate temperature | 550 ± 50 | °C | Measured by infrared pyrometry |
| Substrate dimensions | 10 × 10 × 0.710 | mm | Schott AF45 and Corning Eagle 2000 |
| Substrate thickness | 710 | µm | Glass coupon |
| S-AF45 glass transition temperature | 663 | °C | Growth remained below substrate transition |
| CE2000 glass transition temperature | 722 | °C | Growth remained below substrate transition |
| Seed particles | 6-7 | nm | Detonation nanodiamond suspension |
| Seed-solution zeta potential | 45 ± 5 | mV | Reported colloidal property |
| Raman reference diamond line | 1,332 | cm−1 | Intrinsic diamond optical phonon line |
| B:NCD Raman diamond line | Approximately 1,320 | cm−1 | Shift reflects boron incorporation and film stress |
| Raman shift after membrane release | 2.8 ± 0.8 | cm−1 | Shift toward lower wavenumber indicates stress release |
| Boron detection threshold | Greater than 1020 | cm−3 | Indicated by Raman boron-related component |
| Estimated graphitic phase | 2-3 | % | Estimated from diamond-to-graphitic peak intensity |
| Graphitic Raman feature | Approximately 1,550 | cm−1 | Used for diamond-quality assessment |
| SEM accelerating voltage | 15 | kV | FEI Quanta 200 FEG-SEM |
| Raman excitation wavelength | 488 | nm | Lexell SHG laser |
| Spectral resolution | 0.64 | cm−1 | Horiba Jobin Yvon T64000 Raman system |
| Surface-imaging lateral resolution | 0.69 | µm/pixel | Keyence VK-9700 laser microscope |
| Surface height resolution | 1 | nm | Keyence VK-9700 laser microscope |
| Four-point resistance method | Fixed DC current | — | Keithley 2400 multimeter; measurement after resistance stabilization |
| Elevated observation temperature | 400 | °C | Optical observation of membrane flattening |
| Hole-diameter thermal increase | Approximately 0.1 | % | At 400 °C relative to room temperature |
| Estimated compressive-stress reduction | 0.3 ± 0.1 | GPa | Attributed to differential thermal expansion |
| CE2000 sensitivity improvement | 55 | % | Compared with S-AF45-supported devices |
Membrane Geometry and Doping
Section titled “Membrane Geometry and Doping”| Membrane diameter or condition | Substrate | Resistivity | Reported influence |
|---|---|---|---|
| 280 | S-AF45 | — | Smaller response than larger membrane |
| 550 | S-AF45 | — | Higher response than 280 µm membrane |
| 130 | S-AF45 | 5 × 10−2 Ω·cm | More heavily doped than the 400 µm comparison |
| 400 | S-AF45 | 2 × 10−2 Ω·cm | Higher response associated with heavier doping |
| 280 | S-AF45 | 8 × 10−3 Ω·cm | High-doping, higher-sensitivity condition |
| 550 | S-AF45 | 8 × 10−3 Ω·cm | Largest response in the compared S-AF45 series |
| 560 | CE2000 | — | Wrinkling observed at room temperature and 400 °C |
| 800 | CE2000 | — | Largest plotted membrane on CE2000 |
| Approximately 360 | CE2000 | — | Included in the CE2000 sensitivity comparison |
Key Methodologies
Section titled “Key Methodologies”-
Prepare and clean glass substrates
- Use 10 × 10 × 0.710 mm Schott AF45 or Corning Eagle 2000 glass.
- Apply standard RCA-1 and RCA-2 cleaning procedures.
-
Deposit nanodiamond seed particles
- Apply a water-based colloidal suspension of ultradispersed detonation diamond by drop casting.
- Spin-dry the substrates.
- The suspension contains 6-7 nm particles and has a zeta potential of 45 ± 5 mV.
-
Grow the B:NCD layer by MWPECVD
- Deposit a 200 nm boron-doped nanocrystalline diamond film.
- Use a hydrogen plasma containing 3% methane.
- Set the trimethylboron-to-methane ratio to 10,000 ppm.
- Operate the reactor at 2,100 W microwave power and 11 Torr.
- Maintain a substrate temperature of 550 ± 50 °C, below the glass-transition temperatures of both substrate types.
- Monitor thickness using in-situ laser reflection interferometry.
-
Pattern and release the membrane
- Fabricate a B:NCD Hall-bar structure with the membrane located at its center.
- Suspend the membrane by locally removing the underlying glass in the active area.
- Preserve the peripheral B:NCD film for electrical connection and four-point measurement.
-
Condition and measure the electrical response
- Load the membrane into a purpose-built inflation fixture.
- Apply positive differential pressure at ambient conditions.
- Measure resistance by a four-point technique using a Keithley 2400.
- Apply a fixed DC current and begin data collection only after the resistance has stabilized, limiting self-heating drift.
-
Characterize morphology and stress
- Use SEM to verify a closed, faceted NCD layer and estimate grain size.
- Use Raman spectroscopy to identify diamond, evaluate boron incorporation, estimate graphitic content, and compare residual stress.
- Use laser-scanning microscopy to quantify three-dimensional wrinkling and surface morphology.
-
Compare design variables
- Evaluate membrane diameters between approximately 130 and 800 µm.
- Compare resistivities from approximately 5 × 10−2 to 8 × 10−3 Ω·cm.
- Compare Schott AF45 and Corning Eagle 2000 substrates.
- Relate membrane resistance change, differential pressure, membrane size, boron concentration, and wrinkle density.
6CCVD Solutions & Capabilities
Section titled “6CCVD Solutions & Capabilities”Recommended Material Configuration
Section titled “Recommended Material Configuration”| Application requirement | Recommended 6CCVD material | Engineering rationale |
|---|---|---|
| Highest sensitivity at approximately 200 nm | Heavy boron-doped PCD | PCD supports conductive B:NCD behavior while 6CCVD can provide a 0.1-500 µm thickness range |
| Robust suspended piezoresistive membrane | Heavy boron-doped PCD wafer or membrane | High modulus and fracture strength support thin, low-pressure-deflection diaphragms |
| Lower electrical resistance and stronger piezoresistive response | Heavy boron-doped PCD | The paper reports improved sensitivity with increased boron concentration |
| Alternative for reduced grain-boundary noise or enhanced optical behavior | Boron-doped SCD, where process-compatible | SCD can support thinner membranes down to 0.1 µm, but boron incorporation and conductivity must be validated for the intended resistance range |
| Development of flatter, lower-stress devices | Boron-doped PCD on a thermally matched substrate | Substrate selection should minimize diamond/substrate thermal-expansion mismatch |
- The closest direct analogue to the reported B:NCD film is heavy boron-doped PCD, not undoped optical-grade SCD.
- For replication, 6CCVD should define the boron level using the required sheet resistance or bulk resistivity—not only a nominal precursor ratio—because the paper links conductivity directly to pressure sensitivity.
- The target film thickness should initially be approximately 200 nm, with a controlled development range near 0.1-1 µm to optimize sensitivity, handling strength, and process yield.
Customization and Process Development
Section titled “Customization and Process Development”| 6CCVD capability | Application to this research |
|---|---|
| PCD plates and wafers up to 125 mm | Supports multiple membrane layouts, coupon arrays, and process-development wafers |
| Thicknesses of 0.1-500 µm for PCD | Covers the reported 200 nm regime and allows thicker carrier-supported process development |
| Substrates up to 10 mm | Enables integration studies involving thicker handling or structural substrates |
| Au, Pt, Pd, Ti, W, and Cu metallization | Supports Hall-bar contacts, bonding pads, multilayer adhesion stacks, and low-resistance electrical routing |
| Custom dimensions and laser cutting | Enables array-specific membrane windows, chip singulation, alignment features, and packaged sensor formats |
| PCD polishing below 5 nm Ra for inch-size material | Useful for smooth contact surfaces, bonding interfaces, optical inspection, and consistent lithography |
| SCD polishing below 1 nm Ra | Relevant if the project transitions from PCD to boron-doped SCD membranes or hybrid sensor architectures |
- A practical first-pass metallization stack could be evaluated as Ti/Au for adhesion and low-resistance contacts, with Ti/Pt/Au available when inertness, diffusion control, or wire-bond compatibility is required.
- For high-temperature operation, Ti/W/Au or another refractory adhesion layer may be assessed, but adhesion, contact resistance, thermal mismatch, and boron-doped diamond surface termination must be validated together.
- Laser cutting can define chip boundaries and support openings, although the active membrane release should use a separately qualified process to minimize edge defects and leakage.
- 6CCVD can support DoE splits across thickness, boron concentration, membrane diameter, substrate type, and contact metallization.
Recommended Development Matrix
Section titled “Recommended Development Matrix”| Variable | Suggested starting range | Expected effect |
|---|---|---|
| PCD thickness | 0.2, 0.5, 1.0 µm | Lower thickness increases deflection sensitivity but reduces manufacturing margin |
| Boron level | 3-4 targeted sheet-resistance or resistivity levels | Higher conductivity generally increases piezoresistive response |
| Membrane diameter | 0.2-1.0 mm, subject to strength testing | Larger diameter increases mechanical response |
| Substrate | CE2000, S-AF45, and a lower-expansion alternative | Better thermal matching should reduce compression and wrinkling |
| Contact metal | Ti/Au and Ti/Pt/Au | Separates contact-process effects from membrane-performance effects |
| Surface condition | As-grown and polished | Assesses roughness and thickness nonuniformity effects |
Engineering Support
Section titled “Engineering Support”- 6CCVD’s in-house PhD team can assist with material selection for harsh-environment piezoresistive pressure-sensor projects, including B:NCD conductivity, grain structure, stress, thickness control, and contact integration.
- Recommended project inputs include target pressure range, nominal resistance, membrane diameter, operating temperature, pressure medium, electrical current, and packaging constraints.
- Recommended validation outputs include four-point resistance versus pressure, gauge factor, hysteresis, repeatability, temperature coefficient of resistance, burst pressure, membrane flatness, Raman stress, and long-term chemical stability.
- Initial evaluation should include both CE2000 and a lower-expansion substrate because the paper indicates that closer matching to diamond can improve both flatness and sensitivity.
- For conductive PCD, “gauge factor” should be reported together with the resistance or resistivity range, operating current, temperature, and membrane geometry; otherwise, results from different boron levels cannot be compared reliably.
Application Considerations
Section titled “Application Considerations”- Operating medium: Diamond’s chemical resistance is valuable in corrosive and electrochemical environments, but seals, adhesives, exposed metals, and contamination must be qualified for the specific medium.
- Pressure range: The reported maximum differential pressure is 0.8 bar; the initial replication study should target approximately 0-0.8 bar before extending the design.
- Temperature: The growth process used 550 ± 50 °C on glass, while device behavior was optically examined at 400 °C. Packaged sensors require separate thermal-cycle and pressure-drift testing.
- Electrical stability: Fixed-current four-point testing should be retained, and self-heating should be quantified at each current level.
- Stress control: Raman peak position, membrane wrinkling, and surface topography should be tracked because residual stress governs both baseline resistance and pressure response.
- Delamination risk: A lower-expansion substrate can reduce compressive stress, but a substrate with a substantially lower expansion coefficient than diamond may introduce tensile stress and delamination. The adhesion limit must therefore be included in substrate selection.
- Geometry: Membrane-edge quality is critical. Laser-cut or etched openings should be evaluated for chipping, crack initiation, nonuniform clamping, and leakage.
For custom specifications or material consultation, visit 6ccvd.com or contact our engineering team directly.
View Original Abstract
Nanocrystalline diamond (NCD) membranes are promising candidates for use as sensitive pressure sensors. NCD membranes are able to withstand harsh conditions and are easily fabricated on glass. In this study the sensitivity of heavily boron doped NCD (B:NCD) pressure sensors is evaluated with respect to different types of supporting glass substrates, doping levels and membrane sizes. Higher pressure sensing sensitivities are obtained for membranes on Corning Eagle 2000 glass, which have a better match in thermal expansion coefficient with diamond compared to those on Schott AF45 glass. In addition, it is shown that larger and more heavily doped membranes are more sensitive. After fabrication of the membranes, the stress in the B:NCD films is released by the emergence of wrinkles. A better match between the thermal expansion coefficient of the NCD layer and the underlying substrate results in less stress and a smaller amount of wrinkles as confirmed by Raman spectroscopy and 3D surface imaging.