Access Official Technical PDF

Share Summary:LinkedInEmail

Ultra-pure water verification is essential when water is used to clean semiconductor wafers. Particles or microorganisms left behind during rinsing can compromise wafer surfaces, while increasingly small circuit features make contamination control more demanding. A study of air gap membrane distillation (AGMD) investigated whether water produced by a prototype module could meet this need, using particle-size measurements and wafer-based tests to assess its purity.

The work also considered how the module’s geometry affects its temperature profile and examined several materials for storing ultra-pure water (UPW). The findings offer a practical view of both water-quality testing and the engineering trade-offs involved in developing an AGMD system.

Why ultra-pure water matters in wafer cleaning

In the process described in the study, a silicon wafer’s protective oxide layer is removed using a 5% hydrofluoric acid solution. The wafer is then rinsed. If the rinse water carries particles or other contaminants, residue may remain on the surface and interfere with subsequent semiconductor manufacturing.

As circuit line widths become smaller, the acceptable particle size in cleaning water also becomes more stringent. The study identifies 20 nanometres as an upper limit for particles in the water, with a preference for sizes below 10 nanometres. These targets make it important to assess water quality not only through a general measure of purity, but also by examining particle presence and the results of actual wafer-rinsing tests.

How air gap membrane distillation produces water

Membrane distillation uses a temperature difference to create a vapour-pressure difference across a porous, hydrophobic membrane. Water evaporates on the warm feed side, and water vapour passes through the membrane pores. On the colder side, the vapour condenses and can be collected as purified water. The membrane is intended to maintain a liquid-vapour boundary: it should allow vapour transport without becoming wetted by the feed liquid.

In AGMD, an air gap lies between the membrane and the condensing surface. This configuration separates the condensate from direct contact with the membrane. However, the air gap adds resistance to vapour transport, and the module’s design involves balancing the available diffusion area against heat transfer and operating conditions.

The module examined was a prototype, so its performance should be understood in that development context. The study focused on verifying the produced water and comparing selected design geometries; it did not assess every aspect of system performance, such as energy use or production flow across the simulated geometries.

Testing water purity with particle and wafer analysis

Dynamic light scattering

The study used dynamic light scattering (DLS) to assess particle size in the AGMD-produced water. Spherical particles with a diameter of 20 nanometres were added to the purified water for examination. The reported result was that the measured particle size did not exceed 20 nanometres. This provides evidence relevant to the study’s stated particle-size target, although it does not by itself establish every aspect of water quality.

Drying and rinse tests on silicon wafers

To examine what the water left on a wafer, samples of UPW, tap water and distilled water were placed on silicon wafers and allowed to dry. The surfaces were then inspected using a scanning electron microscope (SEM). No drying spots or visible particles were reported for the UPW sample. Tap water left visible drying spots and many particles on the surface. Distilled water left no visible drying stains, but particles could still be seen under SEM examination.

A further test examined how well the waters rinsed away 5% hydrofluoric acid from a wafer. After rinsing with UPW, no particles were detected on the examined surface. Particles were found after the acid was rinsed with tap water. These comparisons illustrate why a clear-looking dry surface is not enough to confirm that a rinse leaves no detectable residue.

Dilution and storage considerations

The study also compared the amount of water required to dilute a quantity of 5% hydrofluoric acid to a neutral pH of 6–7. Approximately 200 millilitres of UPW was used. The tap-water comparison required about four times as much water to dilute the same amount of acid. This was an initial comparison within the study, rather than a complete assessment of acid-use or process efficiency.

Because UPW can be affected by contact with surrounding materials, the work examined the ability of four container materials to store it while maintaining purity: white borosilicate glass, brown borosilicate glass, ethylene chlorotrifluoroethylene (ECTFE) and polyvinylidene fluoride (PVDF). The available findings describe the materials investigated but do not establish a preferred storage material. For plant and system designers, this reinforces the need to consider storage and wetted components as part of water-quality control, not just the purification unit.

Module geometry involves a production trade-off

COMSOL simulations compared temperature profiles for four membrane geometries. The largest area studied, 255 × 255 millimetres, showed a less favourable profile: temperature fell substantially on the hot side and rose on the cold side. Smaller areas produced more even temperature profiles, with 180 × 100 millimetres giving the best profile among the geometries examined.

That result does not mean the smallest geometry is automatically the best overall design. A smaller membrane area also reduces the diffusion area available for vapour transport, which can reduce UPW production. The design decision therefore involves balancing temperature behaviour against the area available for water production. The simulations examined temperature profiles, not output flow for each geometry.

Practical implications for industrial water systems

The study demonstrates a useful verification approach for evaluating AGMD-produced water for semiconductor cleaning: combine particle-size analysis with direct wafer tests, including SEM inspection after drying and rinsing. It also highlights that the module geometry and water-contact materials deserve attention alongside the membrane separation process.

For engineers and technical decision-makers, the central takeaway is to treat purity verification as a system-level task. Water quality must be checked at the point and in the context of its intended use, while module design must balance temperature profile and available diffusion area. The reported results support further development of AGMD for UPW production, but prototype evaluation and additional testing remain important before drawing broader conclusions about industrial performance.

Access Official Technical PDF

About The Author