The right MEMS glass wafer for anodic bonding must match the silicon wafer, bonding temperature, applied voltage, surface requirements, and device design. For many silicon-based MEMS assemblies, borosilicate glass is a practical starting point because its thermal expansion can be relatively close to silicon. However, the best choice depends on whether the priority is low thermal stress, optical performance, electrical insulation, chemical resistance, or cost. At Glass Circuit, I recommend selecting the glass and bonding process together rather than treating the wafer as a standalone component.
For more information, please visit our website.
Anodic bonding uses heat and an applied electric field to join an alkali-containing glass wafer to a conductive material, commonly silicon. During bonding, mobile ions in the glass migrate under the electric field, creating an interfacial electrostatic attraction and a permanent bond. The process is sensitive to material composition, cleanliness, temperature, voltage, surface contact, and wafer geometry.
I first define the complete device stack before choosing the glass. This includes the silicon type, cavity depth, metallization, electrodes, operating temperature, pressure requirements, optical path, and post-bonding processes. A wafer that bonds successfully in a simple test may still be unsuitable if it creates excessive stress, distorts a cavity, interferes with optical transmission, or reacts with later fabrication steps.
Thermal expansion compatibility is one of the most important selection factors. A typical borosilicate glass may have a coefficient of thermal expansion near 3.3 ppm/K, while silicon is commonly described at approximately 2.6 ppm/K near room-temperature conditions. This relatively close relationship can help reduce thermally induced stress, although the actual result depends on the temperature range, wafer geometry, bonding cycle, and device structure.
Soda-lime glass can offer economical processing and may support anodic bonding because it contains mobile alkali ions. Its thermal expansion is often much higher than silicon, commonly around 8 to 10 ppm/K, so it may create greater residual stress during cooling. Fused silica has a very low expansion value, often close to 0.5 ppm/K, but it is not automatically the best anodic-bonding choice because its composition and electrical behavior may require a different bonding strategy.
| Glass option | Typical selection reason | Important caution |
|---|---|---|
| Borosilicate glass | Good general-purpose balance of thermal compatibility, insulation, and process availability | Exact composition and expansion value must be confirmed with the supplier |
| Soda-lime glass | Potentially lower material cost and accessible anodic-bonding behavior | Higher expansion mismatch can increase stress or wafer deformation |
| Fused silica | Low thermal expansion and strong optical transparency in selected applications | May require a specialized bonding approach and is not a universal replacement |
These values are useful for preliminary comparison, not final design approval. Glass manufacturers may use different compositions, and thermal expansion can vary with temperature and measurement method. I recommend requesting the supplier’s technical data for the exact grade rather than relying only on a generic material name.
A suitable MEMS glass wafer must work within the available equipment window. Anodic bonding is commonly performed at elevated temperature, often within approximately 200 to 500°C, with an applied voltage that may range from about 200 to 1,500 V depending on the glass, wafer size, equipment, and process design. These figures are broad engineering ranges rather than universal recipes, so the supplier and bonding-equipment manufacturer should confirm the correct conditions for the selected material.
Bonding time is also important because longer exposure can increase thermal load and affect throughput. The target process should be established through controlled trials that measure bond integrity, leakage, wafer bow, particle-related defects, and post-bonding stress. I would avoid choosing a wafer solely because it can withstand a particular voltage or temperature; the useful question is whether it produces a stable, repeatable bond without damaging the device.
Material compatibility alone does not guarantee successful bonding. Wafer diameter, thickness, total thickness variation, flatness, bow, warp, and surface roughness all influence how evenly the glass contacts the silicon. For cavity and cap applications, thickness also affects mechanical stiffness, optical behavior, handling strength, and the final package profile.
I recommend defining the geometry using measurable requirements rather than general descriptions such as “high quality” or “polished.” The purchase specification should state the wafer size, thickness and tolerance, surface finish, edge profile, orientation if relevant, and inspection method. If the device has a narrow bonding ring, a wafer that appears acceptable at the center may still produce edge voids or nonuniform bonding when the overall flatness is not controlled.
Glass Circuit Product Page
A polished, clean surface is generally preferred at the bonding interface, but the required finish depends on the device design and supplier capability. Patterned glass may include cavities, alignment marks, holes, electrodes, or other features, and these structures can change local stress and contact behavior. I advise reviewing the drawing with the supplier before fabrication so that feature depth, edge quality, tolerances, and cleaning requirements are considered together.
The glass must perform its functional role after bonding, not only during the bonding step. If it is used as an electrical insulator, its resistivity, dielectric behavior, and thickness may matter. If it forms an optical window, transmission, internal defects, surface quality, and coating compatibility should be specified. If the package will contact gases, liquids, or cleaning chemicals, chemical stability and surface treatment may also affect the selection.
For through-glass vias or embedded conductors, the glass composition and thermal cycle must be compatible with the conductor material. The design team should evaluate potential thermal-expansion differences between glass, silicon, metal, and seal structures. This is especially important when the package will experience repeated temperature changes during operation or qualification.
A lower-cost wafer may increase total project cost if it requires extra process development, produces more voids, or creates unacceptable stress. Material price should be compared with yield risk, inspection requirements, packaging, minimum order quantity, and lead time. For early development, a small engineering batch can be more valuable than an immediate volume order.
“Borosilicate” or “glass wafer” is not always a complete technical specification. Different compositions can have different expansion values, ion mobility, softening behavior, optical properties, and chemical resistance. I recommend requesting the exact grade, material data, available thickness range, and applicable inspection information before approving the purchase.
Bonding may be successful at the interface while the assembled wafer later fails during cooling, dicing, coating, or operation. The complete thermal history should therefore be included in evaluation. A simple coupon test followed by visual inspection is useful, but it may not reveal long-term stress, hermeticity, or fine-pattern damage without additional testing.
For MEMS glass wafer sourcing, I look for a supplier that can discuss both wafer manufacturing and bonding requirements. Useful support may include material selection, custom thickness, wafer sizing, surface finishing, patterning coordination, inspection documentation, packaging guidance, and sample-batch planning. The available support should be confirmed for each project rather than assumed.
Glass Circuit can support technical discussions around custom MEMS glass wafer requirements for anodic-bonding applications. I recommend sharing the silicon material, wafer diameter, thickness, bonding target, surface requirements, pattern drawing, expected quantity, and delivery schedule during the initial inquiry. This information allows us to identify practical material options and clarify which specifications require sampling or process validation.
To choose the right MEMS glass wafer for anodic bonding, I would begin by matching thermal expansion and electrical bonding behavior to the silicon and equipment. I would then confirm wafer geometry, surface quality, optical or chemical requirements, and compatibility with the complete manufacturing flow. Borosilicate glass is often a sensible starting point for silicon-based designs, while soda-lime and fused silica may fit specific cost or performance goals when their limitations are addressed.
The next step is to prepare a technical inquiry with your wafer size, thickness, material stack, bonding temperature, voltage range, surface specification, pattern requirements, and target quantity. Glass Circuit can review these details, suggest suitable MEMS glass wafer options, and discuss sample evaluation before production sourcing. A controlled engineering trial is the most reliable way to confirm that the selected wafer delivers the required bond quality and device performance.
For more mems glass waferinformation, please contact us. We will provide professional answers.