If the scaling down of basic electronic components such as Field Effect Transistor [FET] is to be continued, reduction of the operating voltage is essential in order to avoid overheating due to too high power-dissipation in the ultra-high-density microelectronic circuits. On the other hand, a minimum voltage is required in FET that will allow a clear on/off current ratio, thus the operating voltage cannot be reduced below the theoretical thermal noise limit (60 mV/decade). A way to solve these conflicting needs is to reduce the gate voltage and use a Negative Capacitor [NC] that delivers amplified gate voltage to the channel. Capacitors have always positive values but using ferroelectric materials, it is possible, under special conditions, to give them negative values, negative capacitance. During switching of a ferroelectric, the screening charge lags behind the switched polarization charge for some nano-seconds and during this time the ferroelectric capacitor manifests transiently negative capacitance. After this short time, domain splitting occurs and the capacitor becomes positive as usual. The world-wide intensive search for negative capacitors, particularly for NC-FET, centers on these transient negative capacitors.
In contrast, we proposed a way to get static negative capacitance, leading to Temporally Invariant Negative-Capacitance FET, [TINCFET]. The concept leads also to a desired low capacitance and insensitivity to temperature variations. Temporally Invariant Negative-Capacitance happens when domain splitting (namely, creation of domain walls) is energetically unfavorable because the wall energy is larger than the depolarization energy. This happens when the surface of the ferroelectric capacitor is very small and the domain wall width and the driving electrode dimensions are of the same scale, i.e. 1-10 nm. In this case, the domain wall occupies a considerable fraction of the nano-capacitor volume and the domain wall energy becomes dominant. In this few nanometers nanocapacitor, domain splitting is therefore unfavorable and the ferroelectric can be stabilized in the negative capacitance regime by in-series connected positive capacitor. In this project we set to demonstrate TINCFET experimentally and identify suitable framework to translate our concept into commercial products.
For the demonstrator we have chosen the classical ferroelectric BaTiO3. We selected LaAlO3 for the dielectric layer to stabilize negative capacitance in BaTiO3. The stalk was grown on SrTiO3 substrate, in order to use the conducting interface between LaAlO3 and SrTiO3 as conductor for the demonstration of the negative capacitance. The elaboration of the epitaxial stalk involved a series of experiments monitoring several parameters, particularly the oxygen pressure during deposition. Toward the end of the project we were able to demonstrate ferroelectric switching in the BaTiO3 film. The concept of the functional measurement that would lead to the demonstration of the stabilized negative capacitance was elaborated. Methods for fine patterning of the conductive interface (“bottom electrode”) were investigated. We have not arrived to conduct this experiment in the 18 months period of the project. In parallel, we also collaborated with other academic groups to reveal the various phases of the ferroelectric HfZrO2 and related materials in view of further stabilizing the ferroelectric phase. These materials are Silicon compatible and suitable for the production of transient negative capacitors, but won’t fit into our stabilized NC concept as long as they are multi-phasic. By mid-time of the project, we came to know of a totally new ferroelectric material, which is potentially compatible with Silicon and semiconducting nitrides and can be made into a mono-crystalline, mono-phasic and mono-domain capacitor. We have pursued this material too, towards understanding of its structure and properties, using thin and thick films that were given to us by collaborating academic groups. We have not arrived to show negative capacitance in this material but believe it is highly promising.
In parallel we studied the most significant and largest market size application, that of Temporally Invariant Negative Capacitance FET (TINCFET). Since it is in a very early stage of maturation, the market analysis was based on the realm of opportunities of FinFET, which is the state-of- the-art technology in today’s miniaturization of FET. Most widely used downstream fields of FinFET, and hence also TINCFET Technology market are: smartphones, computers and tablets, wearables, automotive, and high-end networks. The predication, based on FinFet is that the value of FinFET/TINCFET Technology markets can be > 40,000,000,000 USD by 2024. Assuming that TINCFET is to be validated as a more rigorously stable, easier and cheaper to implement - and much better performing technology – it could break through the glass ceiling that today’s FinFET technology encounters. Based on this, commercialization strategy was elaborated. In the current stage of the Technology Readiness Level (TRL), which is still in the process of basic demonstration, further development from public money was recommended. Nevertheless, contacts with companies/partnering were initiated.
