Search

With the gate-length scaling, the number of domains in FeFET is reduced to a few or a single domain. In this paper, we investigate the effect of multi-domains versus few/single-domain behavior in FeFET. The abrupt polarization switching behavior of a single-domain is obtained by modifying the Preisach model in which the difference between saturation and remnant polarization (P s− P r) is reduced. We show that for the same program/erase voltage, a two-times higher memory window can be achieved with single/few-domains FeFET than the multi-domain FeFET. Further, at fixed program/erase voltage, the scaling behavior shows improved variability due to increased polarization-induced vertical field with single-domain FeFET. We present an optimized device with a single-domain FeFET having a low operating voltage of ±2.4 V but with the same device performance that can be achieved for multi-domain FeFET having a higher operating voltage of ±5 V, which is highly promising for low power applications. [ABSTRACT FROM AUTHOR]

In this article, the device scaling of FeFET down to 22 nm and its impacts on memory performance and reliability has been investigated in detail using experimentally calibrated data. Lateral nonuniformity along the channel in the polarization of the ferroelectric layer is observed, which is shown to be due to the scaling effects in FeFET. It has been observed that with gate-length scaling, the remnant polarization shows opposite trends, i.e., Pr reduces for program and increases for erase. Also, with scaling, there is a significant rise in the electric field across the interfacial layer for erase state, which is an important concern for reliability. We find a relatively stable memory window with gate-length scaling due to the counter-balancing of ± Pr. Gate-length scaling has resulted in a larger decrease in Vth and higher subthreshold slope (SS) for program state due to a reduction in gate control as compared to erase state. Our results provide important insights into the design optimization of nanoscale FeFET for high-density and low-power applications. [ABSTRACT FROM AUTHOR]

The threshold voltage variability of the select transistors is an important issue in the development of 3-D NAND flash memory. Particularly, pillar-to-pillar variations in threshold voltage (VT) of the ground select transistor (GST) are critical across the wafer. The VT variation is attributed to the nonuniformity in the plug height of the epitaxially grown silicon layers in different pillars. In this article, we propose different techniques to achieve performance uniformity in different strings across the wafer in terms of VT distribution of GST. We show that by optimizing the channel doping and gate metal work function (WF) of the GST, the NAND string VT nonuniformity can be eliminated. It is also shown that VT variability can be further minimized by optimizing GST gate length. Further, we present a two-MOSFET model for the pillar-to-pillar VT variation across the wafer. This study provides important results for designing 3-D NAND memories with higher performance uniformity for the pillar-to-pillar variations in strings across the wafer. [ABSTRACT FROM AUTHOR]

In this article, a novel technique using the work-function (WF) engineering of passing word line (PWL) has been proposed to mitigate zero-failure in Saddle-Fin-Recessed-Channel-Access-Transistor (S-RCAT). The zero-failure caused by the hammered access [i.e., row hammer (RH)] of PWL is shown to be due to the interference between the PWL and the adjacent storage node (SN). We propose a localized introduction of a high WF PWL to minimize the interference between the PWL and the SN significantly. Using TCAD 3-D process simulations, we analyze the mechanism of PWL hammer-induced zero-failure. We show that a high WF PWL reduces the electron current density by 92% near the shallow-trench-isolation (STI) during the accessing of PWL. The lesser electron density near STI isolates the SN potential to rise due to the hammered access of PWL, which results in the improvement of 77% in zero-failure mitigation. The introduction of high WF PWL does not have any substantial effect on active access transistor characteristics (ION, SS, and VT). The results obtained in this article demonstrate the potential of using localized high WF PWL in S-RCAT to mitigate the zero-failure in the future dynamic-random-access-memory (DRAM) technology. [ABSTRACT FROM AUTHOR]

The channel tapering from top to bottom in vertical channel 3-D NAND is a major concern. This leads to nonuniformity in the NAND string performance, including cell current (Icell) and threshold voltage (VT) variation along the vertical NAND string. In this article, we show that the variation in the electric field due to the difference in the channel radius from bottom to top is the root cause behind the VT variation along the string. For the first time, we propose novel techniques to minimize the adverse effects of the channel tapering on VT variations. It is shown that graded channel doping (~1018 cm−3 at the bottom to ~1015 cm−3 at the top) results in the narrowing of the VT variation by ~90% along the vertical string. Further, we propose that a nonuniform block oxide thickness along the string can be used to enhanced uniformity of VT distribution from bottom to top. Additionally, we investigate and show that the uniformity in VT distribution among word line (WL) transistors can also be achieved by optimizing the amplitude and duration of the position-dependent program/erase voltages. The proposed techniques in this article have a high potential for designing variation tolerant and reliable 3-D NAND memories having enhanced uniformity in the program/erase VTand Icell distribution. [ABSTRACT FROM AUTHOR]

In this letter, we investigate row hammering (RH) in recessed channel access transistor (RCAT)-DRAM cells. A novel technique is proposed to minimize the row hammer fail in DRAM cells by selective/local introduction of metal nanoparticles (MNPs) at the gate metal–oxide interface. Using TCAD simulations, we show that the introduction of MNPs with an optimized work function (WF) significantly reduces the storage node leakage of unaccessed DRAM cells (victim cells) due to the hammered access of neighboring DRAM cells (aggressor cells) sharing a common bit line. The proposed technique potentially eliminates row hammer fail in RCAT-DRAM cells at 2X and 1X nodes. The improvement in row hammer fail is due to the dual-energy valley induced between the neighboring DRAM cells by the introduction of MNPs. These valleys prevent diffusion of electrons from the aggressor cell to the victim cell, which improves row hammer fail. This letter shows that the introduction of MNPs is a promising technique to eliminate row hammer fail in DRAM cells. [ABSTRACT FROM AUTHOR]

String read current ( ${I}_{\textsf {read}}$ ) reduction with rising mold height and grain boundary traps is one of the major hurdle in the development of 3-D NAND flash memory. In this paper, we have investigated ${I}_{\textsf {read}}$ with variation in polysilicon channel grain size (GS), grain boundary trap density, and channel thickness ( ${T}_{\textsf {Si}}$ ), using TCAD. We find that under a critical value of GS, ${I}_{\textsf {read}}$ decreases with increase in ${T}_{\textsf {Si}}$. This is attributed to the fact that with smaller GS, the total number of grain boundaries and associated traps are significantly higher. Moreover, there exists a typical value of GS for which ${I}_{\textsf {read}}$ is independent of ${T}_{\textsf {Si}}$ , which is desirable to minimize the deviations in ${I}_{\textsf {read}}$ arising from ${T}_{\textsf {Si}}$ variations. The resulting tradeoff in the design of more efficient 3-D NAND flash is demonstrated and discussed. Further, it is found that ${I}_{\textsf {read}}$ increases significantly by limiting the polysilicon channel grain boundary trap concentration under 1012 cm−2. The results presented in this paper are crucial for optimizing ${I}_{\textsf {read}}$ and program/erase threshold voltage ( ${V}_{T}$ ) window, and serve as key guidelines in the design of 3-D NAND flash memory with better performance. [ABSTRACT FROM AUTHOR]

Development of a reliable 3D integrated system is largely dependent on the choice of filler materials used in through-silicon vias (TSVs). This research paper presents carbon nanotube (CNT) bundles as prospective filler materials for TSVs and provides an analysis of signal integrity for different single-(SWCNT), double- (DWCNT), and multi-walled CNT (MWCNT) bundle based TSVs. Depending on the physical configuration of a pair of TSVs, an equivalent electrical model is employed to analyze the inphase and out-phase delays. It is observed that, using an MWCNT bundle (with number of shells = 10), the overall in-phase delays are reduced by 96.86%, 92.33%, 78.35%, and 32.72% compared to the bundled SWCNT, DWCNT, 4-shell MWCNT, and 8-shell MWCNT, respectively; similarly, the overall reduction in out-phase delay is 85.89%, 73.38%, 45.92%, and 12.56%, respectively. [ABSTRACT FROM AUTHOR]