Access Official Technical PDF

Share Summary:LinkedInEmail

As MOS devices are scaled, shrinking a silicon dioxide gate dielectric can increase leakage as carriers tunnel through the increasingly thin insulating layer. A high-k material can provide useful capacitance in a physically thicker film. Research into a zirconium oxide gate dielectric examined plasma-deposited ZrO2 nanolayers on silicon and their electrical behavior in metal-oxide-semiconductor (MOS) capacitors.

The work is relevant to engineers and device developers assessing alternative gate-insulating materials. It connects deposition conditions and film morphology with measured leakage current and capacitance, while also identifying the conduction mechanisms observed in the capacitors.

Why investigate zirconium oxide as a high-k dielectric?

Zirconium oxide (ZrO2) is one of several materials explored as a replacement for silicon dioxide in MOS technology. The paper describes ZrO2 as having a high dielectric constant, a wide band gap and stability with silicon substrates and gate-electrode metals. These properties make it a candidate for further investigation, although the electrical behavior of a finished device also depends on how its dielectric film is fabricated.

That process dependence matters when comparing materials. A dielectric’s nominal properties alone do not establish device performance: film continuity, surface condition and interfaces can all affect the electrical measurements. The study therefore considered both the nanolayers’ morphology and the behavior of MOS capacitors built using them.

Dense plasma focus deposition of ZrO2 nanolayers

To form the films, the researchers used a modified dense plasma focus (DPF) device. Focused, hot argon plasma interacted with a solid ZrO2 target, ablating material that was transported to silicon substrates. The substrates were positioned 4.0 cm and 5.0 cm from the top of the anode, and each received two plasma bursts.

The resulting film thickness depended on substrate position. Ellipsometry measurements found approximately 30 nm at 4.0 cm and 20 nm at 5.0 cm. Scanning electron microscopy (SEM) showed uniform layers with nanoscale structures averaging about 26 nm and 19 nm, respectively. Atomic force microscopy (AFM) supported the observed morphology, showing continuous films with low root-mean-square roughness: approximately 0.13 nm for the 4.0 cm sample and 0.26 nm for the 5.0 cm sample.

The authors attributed the smaller structures at the farther position to fewer material ions reaching the substrate, consistent with the reported fountain-like trajectories of the ablated ions. The two placements therefore produced films that differed not only in thickness but also in measured morphology.

MOS capacitor electrical results

For electrical evaluation, aluminum top electrodes were deposited as dots on the ZrO2 films, with an aluminum back contact on the silicon side. The resulting Al-ZrO2-Si structures were tested using current-voltage (J-V) and high-frequency capacitance-voltage (C-V) measurements.

At 1 V, the capacitor using the film deposited at 4.0 cm had a leakage current density of about 1.72 × 10-8 A/cm2. The 5.0 cm sample measured about 7.34 × 10-8 A/cm2 at the same voltage. In these tests, the closer-position sample therefore showed lower leakage.

At a measurement frequency of 1 MHz, capacitance density was approximately 2.2 µF/cm2 for the 4.0 cm sample and 1.7 µF/cm2 for the 5.0 cm sample. Notably, the 4.0 cm film was thicker yet had higher capacitance density. The paper suggests that lower defect levels or greater crystallinity may help explain this result; it does not establish either explanation as a definitive cause.

What the conduction analysis indicates

The J-V data for the 4.0 cm sample were analyzed to assess how current moved through the capacitor. The results indicated different mechanisms in different electric-field regimes:

  • Low-field regime: Field-assisted or tunnel emission was identified as the likely conduction mechanism.
  • High-field regime: The near-linear behavior in the analyzed plot was consistent with Schottky emission.

This distinction is useful when interpreting dielectric measurements: leakage behavior does not necessarily follow a single mechanism across the full applied-field range. The study also observed higher leakage in the accumulation region than in inversion, which the authors linked to possible differences in material properties and conduction at the Al/ZrO2 and ZrO2/Si interfaces.

Practical implications for dielectric evaluation

The results show why process conditions and electrical characterization should be assessed together when evaluating a high-k dielectric. In this study, changing the substrate distance in the DPF process coincided with differences in film thickness, nanoscale features, roughness, leakage and capacitance. The 4.0 cm sample had the lower measured leakage and higher capacitance density of the two tested positions.

These findings are specific to the reported deposition setup and MOS capacitor structures. They do not by themselves establish how the films would perform in a complete transistor or under other fabrication and operating conditions. For device development, the practical next step is to treat deposition geometry and film quality as variables to validate alongside electrical properties and interface behavior.

Conclusion

Plasma-based deposition produced uniform ZrO2 nanolayers on silicon that could be integrated into MOS capacitors. Under the reported test conditions, the film deposited 4.0 cm from the anode had a leakage current density of about 1.72 × 10-8 A/cm2 at 1 V and a capacitance density of about 2.2 µF/cm2. The study also identified field/tunnel emission at low fields and Schottky emission at high fields. For engineers assessing zirconium oxide gate dielectrics, the central takeaway is that deposition conditions, morphology and interface-related electrical behavior all warrant attention—not just the material choice.

Access Official Technical PDF

About The Author