The negative capacitance operation of a ferroelectric material is not only an intriguing materials science topic, but also a property with important technological applications in nanoscale electronic devices. Despite growing interest for possible applications, the very existence of negative capacitance is still actively debated, even because experimental results for ferroelectric capacitors with or without a metal interlayer led to quite contradicting indications. Here we present a comprehensive analysis of NC operation in ferroelectric capacitors and provide new insights into the discrepancies observed in experiments. Our models duly account for the three-dimensional nature of the problem and show a good agreement with several aspects of recent experiments. Our results also demonstrate that traps at the ferroelectric-dielectric interface play an important role in the feasibility of stable negative capacitance operation in ferroelectric capacitors.

The authors regret that the value of ß in the caption of Fig. 4 was incorrectly given as 4.5 × 109 m5 C-2 F-1. The correct value of ß is 2.25 × 1010 m5 C-2 F-1. Fig. 4, along with the full corrected caption, is displayed below. The Royal Society of Chemistry apologises for these errors and any consequent inconvenience to authors and readers. (Figure Presented).

We present a revised analysis of Negative Capacitance (NC) in ferroelectric-insulator capacitors, and particularly of the difference between systems with and without a metal interlayer. We develop a model accounting for the three-dimensional electrostatics and report analytical and numerical results based on Landau-Ginzburg equations. Our results explain the lack of NC operation in capacitors having an interlayer metal, compare well with recent experiments and enlighten the role of traps at the ferroelectric-oxide interface.