Access Official Technical PDF

Share Summary:LinkedInEmail

Re-entrant nanostructures offer a way to control water behaviour on III-V semiconductor surfaces without applying a hydrophobic polymer coating. In a reported InP/InGaAs/InP multilayer demonstration, carefully designed, mushroom-shaped features changed an untreated, water-attracting surface into one that repelled water or allowed a droplet to be lifted away. The work also highlights a manufacturing consideration for device teams: patterning throughput and feature uniformity both matter when translating nanoscale surface designs into useful areas.

Why control wettability on III-V semiconductors?

Materials such as indium phosphide (InP) and gallium arsenide are used in optoelectronic and high-frequency microelectronic components, including lasers, photodetectors and sensors. Some devices must function in environments where water exposure is a concern. A protective polymer such as polydimethylsiloxane (PDMS) can provide waterproofing, but polymer coatings may have limited thermal conductivity and can be undesirable where device processing requires compatibility with semiconductor fabrication steps.

Surface structuring provides a different approach: instead of adding a separate protective layer, it changes the geometry of the semiconductor surface. The important point is that a surface’s interaction with water depends not only on the material itself, but also on the shape and spacing of its features.

How re-entrant nanostructures change water behaviour

A flat InP surface in the reported work had a water contact angle of about 22.4 degrees, indicating that it was hydrophilic. The researchers formed T-shaped, re-entrant features—structures with an overhanging top—on the surface. These features supported a droplet on a reduced fraction of solid area, leaving space beneath much of the droplet rather than allowing it to contact the entire underlying surface.

This arrangement can produce a Cassie-Baxter wetting state. In this model, the apparent contact angle depends on the material’s inherent contact angle and the fraction of the surface that is actually in contact with the liquid. Reducing that contact fraction can increase the apparent contact angle even when the underlying semiconductor is hydrophilic.

However, a high contact angle alone does not determine whether a droplet will detach. A textured surface can repel water while still holding a droplet in place. That distinction is important when the desired outcome is not simply water resistance, but the ability to remove water from the surface.

Fast electron beam lithography for patterned surfaces

The team used fast electron beam lithography (FEBL), a single-shot patterning approach that specifies an array’s coverage area and spacing rather than drawing each circular feature as a collection of polygons. The beam rasters across the resist and dwells at defined points. Adjusting dwell time changes the exposed dot size, while the spacing between dwell points sets the array pitch.

Compared with conventional electron beam lithography, which can spend substantial time defining and fracturing individual shapes, this approach reduced the time required for large arrays. For example, the reported comparison for a 1 mm by 1 mm array of 300 nm dots on a 500 nm pitch was more than 24 hours using conventional EBL versus about one hour using FEBL. For 95 nm dots at the same pitch, the reported FEBL patterning time was under 20 minutes. The study also reported throughput above 106 µm² per hour for the demonstrated dot-patterning process.

These figures describe the reported patterning process, not the complete manufacturing time for a finished device. The resulting re-entrant structures also required resist processing, metal-mask formation, reactive ion etching and selective wet etching. The method’s value is its combination of direct-write flexibility, faster array definition and demonstrated feature uniformity.

Designing for droplet release, not just repellency

Two InP surface designs illustrate why geometry must be selected for the intended droplet behaviour. Sample A1 used features with a 4.5 µm diameter, 8 µm pitch and 2.5 µm height. Its measured contact angle was 124.5 degrees, close to the modelled value of 121.4 degrees. The droplet was pinned to the surface when lifted with a capillary needle.

Sample A2 used a 4 µm diameter, 10 µm pitch and 2 µm height. Its lower solid contact fraction reduced the droplet’s adhesion to the textured surface. In the demonstration, the droplet could be removed completely using the needle. The practical lesson is that “water-repellent” and “low-adhesion” are not interchangeable design objectives. Feature dimensions should be chosen in relation to how water must behave in the application.

Using adhesion-energy modelling

The researchers compared the energy holding the droplet to the dispensing needle with the energy associated with contact between the droplet and the structured semiconductor, plus the energy needed to lift the droplet against gravity. In simplified terms, detachment is expected when needle-to-water adhesion exceeds the combined surface adhesion and lifting requirement.

For A1, the calculated water-to-surface adhesion plus lifting energy was greater than the water-to-needle adhesion, consistent with the observed pinning. For A2, the balance reversed, consistent with complete removal. This modelling approach links a measurable outcome to design variables such as feature spacing and contact fraction, helping distinguish surface geometries that merely increase apparent contact angle from those that promote release.

What the demonstration means for device development

The structures were fabricated on an InP/InGaAs multilayer, with the top InP layer patterned and the underlying InGaAs selectively under-etched to create the overhanging profile. The arrays covered a 3 mm by 3 mm area, and related FEBL array demonstrations showed high uniformity. These results establish a laboratory route for controlling wettability on a semiconductor surface; they do not by themselves establish long-term durability or performance in a specific operating environment.

For engineering teams considering this approach, the key evaluation questions include:

  • Does the device need water to bead, or must droplets detach under a particular handling condition?
  • Can the surface pattern and etch sequence fit the existing semiconductor process flow?
  • What feature size, pitch and patterned area are needed, and can they be produced with the required uniformity?
  • How should wetting and droplet release be validated under the device’s intended conditions?

Conclusion

Re-entrant nanostructures can tune the wetting behaviour of intrinsically hydrophilic III-V semiconductors without relying on a polymer coating. Fast electron beam lithography makes large, flexible arrays more practical to pattern, while adhesion-energy modelling helps identify geometries that encourage droplet pinning or release. For device designers, the central implication is to treat surface geometry, fabrication compatibility and the required water-removal behaviour as connected design decisions—not to rely on contact angle alone.

Access Official Technical PDF

About The Author