Electrohydrodynamics Cleans Probe Cards
Maria A. Lester, Associate Editor -- Semiconductor International, 8/1/2000
 |
|
|
The electrohydrodynamics method is used in a contact resistance reduction system, called microBurst, cleaning probe cards by creating contact resistance. The system atomizes a conducting liquid using an electric field at the tip of a capillary, creating microclusters. The microclusters are then ejected from the capillary in a "hypervelocity" divergent spray, cleaning a probe card without coming into physical contact. To prevent charge buildup on the probe card, the microcluster beam is neutralized by injecting low-energy electrons. •
Click here to read the Applied Precision research.
Click here to read the Texas Instruments study.
IMEC-Clean, RCA Replacement
The RCA-clean is still a favored front-end wet clean. Alternatives have been developed to meet more stringent process demands and lower environmental impact. The challenge is to develop a hybrid wet/dry clean that can be integrated into a cluster tool. The IMEC-clean has been studied in more detail over the past few years, for it uses less chemicals and DI water with fewer cleaning steps. It was most recently studied collaboratively at IMEC (Leuven, Belgium), Texas Instruments (Dallas) and STEAG MicroTech (Pliezhausen, Germany).
This cleaning concept is a two-step approach. First, organic contamination is removed and a thin chemical oxide is grown. H2SO4/O3 mixture or DI water can remove organic contamination. Then, a diluted HF/HCl mixture removes chemical oxide as well as particle and metal contamination. This mixture suppresses the Cu-outplating common to metal-contaminated HF-baths. A third step can be added to regrow a thin passivating oxide layer if a hydrophilic surface is needed. For example, dHCl/O3 can be used to make a Si-surface hydrophilic at low pH values while preventing reintroduction of metal contamination.
The particle removal efficiency was tested using positive (Si3N4 and Al2O3) or negative (PSL-sphere) charged particles over the pH-range and particles having the same charge as the SiO2 oxide substrate. The pH and megasonic irradiation during the final rinse was found to have a significant effect on the removal efficiencies. PSL-spheres and SiO2 particles were removed by a rinse only when using megasonic irradiation. Electrostatic interaction and van der Waals forces were overcome. High removal efficiencies were found for most combinations of particle type and cleaning sequences (results presented in Table). The Table also indicates the amount of material etched during the cleaning sequence. The 3.2 nm of thermal oxide etched is comparable to the etching in RCA sequences using heated SC1 solutions. However, TEOS etching is faster than thermal oxide, so a few seconds of HF-dip removes particles with less etching.
Overall, no Si-surface roughening was found, as can occur in RCA-clean. All results indicate very low levels of metals contamination can be achieved. •
| Table. Particle Removal Efficiency | ||||
| Substrate | Etching (nm) | Particle removal efficiency (%) | ||
| Si3N4 | Al2O3 | silica | ||
| Bare silicon | 100 | 100 | 99 | |
| TEOS film | 12.8 | 99 | 100 | 100 |
| Thermal oxide | 3.1 | 100 | 100 | 100 |
| LPCVD nitride | 0.8 | 98 | 100 | 94 |
(Source: IMEC)
