Metal gate work function engineering on gate leakage of MOSFETs

We present a systematic study of tunneling leakage current in metal gate MOSFETs and how it is affected by the work function of the metal gate electrodes. Physical models used for simulations were corroborated by experimental results from SiO/sub 2/ and HfO/sub 2/ gate dielectrics with TaN electrodes. In bulk CMOS results show that, at the same capacitance equivalent oxide thickness (CET) at inversion, replacing a poly-Si gate by metal reduces the gate leakage appreciably by one to two orders of magnitude due to the elimination of polysilicon gate depletion. It is also found that the work function /spl Phi//sub B/ of a metal gate affects tunneling characteristics in MOSFETs. It is particularly significant when the transistor is biased at accumulation. Specifically, the increase of /spl Phi//sub B/ reduces the gate-to-channel tunneling in off-biased n-FET and the use of a metal gate with midgap /spl Phi//sub B/ results in a significant reduction of gate to source/drain extension (SDE) tunneling in both n- and p-FETs. Compared to bulk FET, double gate (DG) FET has much lower off-state leakage due to the smaller gate to SDE tunneling. This reduction in off-state leakage can be as much as three orders of magnitude when high-/spl kappa/ gate dielectric is used. Finally, the benefits of employing metal gate DG structure in future CMOS scaling are discussed.     

A dual-metal gate integration process for CMOS with sub-1-nm EOT HfO/sub 2/ by using HfN replacement gate

A replacement gate process employing a HfN dummy gate and sub-1-nm equivalent oxide thickness (EOT) HfO/sub 2/ gate dielectric is demonstrated. The excellent thermal stability of the HfN-HfO/sub 2/ gate stack enables its use in high temperature CMOS processes. The replacement of HfN with other metal gate materials with work functions adequate for n- and pMOS is facilitated by a high etch selectivity of HfN with respect to HfO/sub 2/, without any degradation to the EOT, gate leakage, or time-dependent dielectric breakdown characteristics of HfO/sub 2/. By replacing the HfN dummy gate with Ta and Ni in nMOS and pMOS devices, respectively, a work function difference of /spl sim/0.8 eV between nMOS and pMOS gate electrodes is achieved. This process could be applicable to sub-50-nm CMOS technology employing ultrathin HfO/sub 2/ gate dielectric.     

Impact of Gate Electrodes on $hbox{1}/f$ Noise of Gate-All-Around Silicon Nanowire Transistors

The low-frequency (1/f) noise of gate-all-around silicon nanowire transistors (SNWTs) with different gate electrodes (poly-Si gate, doped fully silicided (FUSI) gate, and undoped FUSI gate) is studied in the strong-inversion linear region. It shows that the gate electrodes have a strong impact on the 1/f noise of the SNWTs. The highest noise is observed in the SNWTs with a poly-Si gate, compared to their FUSI-gate counterparts. The observations are explained according to the number fluctuation with correlated mobility fluctuation theory by assuming that the correlated mobility scattering is better screened in the case of an undoped FUSI gate. However, the doped FUSI gate with silicidation-induced impurity segregation at the gate/SiO₂ interface gives rise to extra mobility scattering.     

Abrupt breakdown in dielectric/metal gate stacks: a potential reliability limitation?

In downscaled poly-Si gate MOSFET devices reliability margin is gained by progressive wearout. When the poly-Si gate is replaced with a metal gate, the slow wearout phase observed in ultrathin SiON and HfSiON dielectrics with poly-Si gate disappears, and with it, the reliability margin. We demonstrate for several combinations of dielectric and gate materials that the large abrupt current increase (/spl Delta/I) as compared to poly-Si is not likely due to process issues, but is an intrinsic property of the dielectric/metal gate stack. The occurrence of large /spl Delta/I is a potential limitation for the reliability of metal gate devices.     

Gate Influence on the Layout Sensitivity of $ hbox{Si}_{1 - x}hbox{Ge}_{x} hbox{S/D}$ and $hbox{Si}_{1 - y}hbox{C}_{y} hbox{S/D}$ Transistors Including an Analytical Model

We present a simulation study on the effect of the gate module on the channel stress in Si_1-xGe_x and Si_1-yC_y S/D MOS transistors. Stiff gate materials, such as titanium nitride, lead to a decreased channel stress, while a replacement-gate scheme allows the increase of the effectiveness of the Si_1-xGe_x and Si_1-yC_y S/D techniques significantly, independent of the gate material used. The drawback of using a replacement gate is that the channel stress becomes more sensitive to layout variations. In terms of effect on Si_1-xGe_x/Si_1-yC_y S/D stress generation, using a thin metal gate capped by polysilicon is similar to a full metal gate if the thin metal gate thickness exceeds 10 nm. Even metal gates as thin as 1 nm have a clear influence on the stress generation by Si_1-xGe_x/Si_1-yC_y S/D. Removing and redepositing the polysilicon layer while leaving the underlying metal gate unchanged increases the stress, although not to the same extent as for complete gate removal. A simple analytical model that estimates the stress in nested short-channel Si_1-xGe_x and Si_1-yC_y S/D transistors is presented. This model includes the effect of germanium/carbon concentration, active-area length, as well as the effect of gate length and the Young's modulus of the gate. Good qualitative agreement with 2-D finite element modeling is demonstrated.     

A novel GaAs FET with double camel-like gate structure

Extremely high potential barrier height and gate turn-on voltage of a novel GaAs field-effect transistor with n/sup +//p/sup +//n/sup +//p/sup +//n double camel-like gate structure are demonstrated. The maximum electric field and potential barrier height of the double camel-like gate are substantially enhanced by the addition of another n/sup +//p/sup +/ layers in gate region, as compared with the conventional n/sup +//p/sup +//n single camel-like gate. For a 1/spl times/100 /spl mu/m/sup 2/ device, a potential barrier height up to 2.741 V is obtained. Experimentally, a high gate turn-on voltage up to +4.9 V is achieved because two reverse-biased junctions of the double camel-like gate absorb part of positive gate voltage. In addition, the transistor action shows a maximum saturation current of 730 mA/mm and an extrinsic transconductance of 166 mS/mm.     

Gate engineering for deep-submicron CMOS transistors

Gate depletion and boron penetration through thin gate oxide place directly opposing requirements on the gate engineering for advanced MOSFET's. In this paper, several important issues of deep-submicron CMOS transistor gate engineering are discussed. First, the impact of gate nitrogen implantation on the performance and reliability of deep-submicron CMOSFET's is investigated. The suppression of boron penetration is confirmed by the SIMS profiles, and is attributed mainly to the diffusion retardation effect in bulk polysilicon by the presence of nitrogen. The MOSFET' I-V characteristics, MOS capacitor quasi-static C-V curves, SIMS profiles, gate sheet resistance, and oxide Q_bd are compared for different nitrogen implant conditions. A nitrogen dose of 5×10¹⁵ cm^-2 is found to be the optimum choice at an implant energy of 40 keV in terms of the overall electrical behavior of CMOSFET's. Under optimum design, gate nitrogen implantation is found to be effective in eliminating boron penetration without degrading performance of either p⁺ gate p-MOSFET and n⁺ gate n-MOSFET. Secondly, the impact of gate microstructure on the performance of deep-submicron CMOSFET's is discussed by comparing poly and amorphous silicon gate deposition technologies. Thirdly, poly-Si_1-xGe_x is presented as a superior alternative gate material. Higher dopant activation efficiently results in higher active-dopant concentration near the gate/SiO₂ interface without increasing the gross dopant concentration. This plus the lower annealing temperature suppress the dopant penetration. Phosphorus-implanted poly-Si_1-xGe_x is gate is compared with polysilicon gate in this study     

Electron gate currents and threshold stability in the n-channel stacked gate MOS tetrode

A new structure is given for the n-channel stacked gate MOS tetrode which consists of a polycrystalline silicon buried control gate and thermally grown oxide for the offset gate insulator. As a result of the large band-bending in the offset gate depletion region of an operating tetrode, some drain current electrons surmount the Si-SiO2energy barrier and are injected into the oxide. Since the electron trapping is relatively small in the thermal-oxide offset gate insulator, it was possible to measure gate currents of up to2 times 10^{-4}A/cm². The gate current was measured as a function of the drain current, the drain voltage and the offset gate voltage. The resulting behavior confirms previous models of the tetrode device. Since electron trapping is much less in thermally grown oxide than in deposited pyrolytic oxide which was used formerly, the offset gate threshold voltage shifts less. As a result of this effect the new structure is used to advantage in fabricating the n-channel stacked gate tetrode in that the drain current is comparatively insensitive to changes in the offset gate voltage.     

The influence of gate edge shape on the degradation in hot-carrierstressing of n-channel transistors

The hot-carrier properties of planar and graded gate structures (upturning of the gate edge in the gate overlap region) of n-MOS transistors were examined. It was found that the type of degradation suffered by each type of device depends on the shape of the gate edge. This is interpreted in terms of the degree of gate control of the gate over the region in which the damage takes place in the different devices. The nongraded gate (NGG) devices degrade chiefly by a V_t shift, whereas the graded gate (GG) devices show a pronounced transconductance decay, with practically no V_t shift. It is suggested that the damage is situated in the gate overlap region, and that the different degradations result from a weaker field control of the gate over the degraded region leading to a series resistance type of effect in the case of the GG structure. This is supported by two-dimensional simulations     

Suppression of the boron penetration induced dielectric degradationby using a stacked-amorphous-silicon film as the gate structure forpMOSFET

This work proposes a stacked-amorphous-silicon (SAS) film as the gate structure of the p⁺ poly-Si gate pMOSFET to suppress boron penetration into the thin gate oxide. Due to the stacked structure, a large amount of boron and fluorine piled up at the stacked-Si layer boundaries and at the poly-Si/SiO₂ interface during the annealing process, thus the penetration of boron and fluorine into the thin gate oxide is greatly reduced. Although the grain size of the SAS film is smaller than that of the as deposited polysilicon (ADP) film, the boron penetration can be suppressed even when the annealing temperature is higher than 950°C. In addition, the mobile ion contamination can be significantly reduced by using this SAS gate structure. This results in the SAS gate capacitor having a smaller flat-band voltage shift, a less charge trapping and interface state generation rate, and a larger charge-to-breakdown than the ADP gate capacitor. Also the Si/SiO₂ interface of the p⁺ SAS gate capacitor is much smoother than that of the p⁺ SAS gate capacitor     

Influences of structure around gate-edge on high electric field strength in MISFETs with high-k gate dielectrics

In this paper, dependences of electric field strength around gate-edge in gate dielectrics of MISFETs with high-k gate dielectrics on design parameters are studied. It is newly found that locations of sidewall/gate dielectric interfaces relative to gate electrode edges are critical to electric field strength of high-k MISFETs. Electric field can be as high as 4 MV/cm, which could have large influences on the yield of large scale integrated circuits (LSIs) with high-k gate dielectrics. An explanation of this phenomenon is given by considering discontinuity in electric field at interfaces between two materials with different dielectric constants. It is clarified that an electrical potential of side and top surfaces of gate dielectrics is strongly affected by the discontinuity of electric field strength at interfaces. As a result, electric field strength around gate electrode edges critically depends on locations of sidewall/gate dielectrics interfaces relative to gate electrode edges. Based on the physical considerations, a structure, in which gate sidewalls are also made of high-k materials, is studied from the viewpoint of electric field strength around gate electrode edges. It is shown that this structure effectively suppresses electric field strength around gate edges.     

Self-aligned nickel, cobalt/tantalum nitride stacked-gate pMOSFETs fabricated with a low temperature process after metal electrode deposition

This letter reports the first replacement (Damascene) metal gate pMOSFETs fabricated with Ni/TaN, Co/TaN stacked electrode, where Ni or Co is in direct contact with the gate SiO/sub 2/, to adjust the electrode metal work function and TaN is used as the filling material for the gate electrode to avoid wet etching and CMP problems. The process is similar to the fabrication of traditional self-aligned polysilicon gate MOSFETs, except that in the back end (after the source/drain implants are activated) a few processing steps are added to replace the polysilicon with metal. Our data show that the Ni or Co/TaN gate electrode has the right work function for the pMOSFETs. The metal gate process can reduce the gate resistivity. Thermal stability of the stacked electrodes is studied and the result is reported in this paper. The damascene process flow bypasses high temperature steps (> 400/spl deg/C)critical for metal gate and hi k materials. This paper demonstrates that a low temperature anneal (300/spl deg/C) can improve the device performance. In this paper, the gate dielectrics is SiO/sub 2/.     

A high-speed high-density silicon 8/spl times/8-bit parallel multiplier

An 8/spl times/8-bit parallel multiplier with submicrometer gate lengths has been fabricated using silicon NMOS technology. The multiplication time is 9.5 ns. This corresponds to an average loaded gate delay in the multiplier circuit of 244 ps/gate, which the authors believe is the shortest gate delay for MOS multiplier circuits demonstrated to date. The power dissipation is 600 mW at a supply voltage of 5 V. The multiplier circuit has a total of 1427 transistors in an active area of 0.61/spl times/0.58 mm/SUP 2/, corresponding to a gate density of 1125 gates/mm/SUP 2/.     

Is gate line edge roughness a first-order issue in affecting the performance of deep sub-micro bulk MOSFET devices?

An experimental study of the effects of gate line edge roughness (LER) on the electrical characteristics of bulk MOSFET devices was performed. Device simulation had previously predicted that gate LER causes off-state current to increase. In our experiments, gate LER was deliberately introduced in devices with 40 nm or longer physical gate length, and we found about 3X I/sub OFF/ increase (for 40 nm gate length) in the I/sub OFF/-I/sub ON/ plot. In our devices, Source/Drain (S/D) extensions are produced by implants self-aligned to gate edges. Simulation results indicate that what really matters is the roughness induced of the S/D to channel junctions by gate LER. Implantation scattering and dopant diffusion cause the S/D to channel junctions to be smoother than the gate edges. This will partially reduce the differences in the I/sub OFF/-I/sub ON/ curves caused by differing amounts of gate LER. By optimizing our process flows, we obtained a minimized gate LER (EdgeRMS<2 nm). We believe that the consequence of this minimized LER is secondary to the impact of other process variations across wafer for devices with 40 nm or longer gate length.     

Effects of gate geometry on propagation delay of integrated injection logic (I/SUP 2/L)

The effects of gate geometry on the propagation delay have been investigated for I/SUP 2/L gates with a self-aligned double-diffusion injector (S/SUP 2/L). To improve the switching speed of the I/SUP 2/L gate, the stored charge in the upside-down operated n-p-n transistor in the gate should be minimized. Following this principle, one can straightforwardly find that the reduction of the stored charges in the internal n-p-n base region and in the lateral p-n-p base region is the step to be taken for the further improvement of the speed. This can be realized by simply contracting the geometry of the gate. The minimum delay time realized in the gate was 3.2 ns/gate. Assuming that capabilities of processing the devices with 1-/spl mu/m accuracy become available, it is predicted that 1 ns/gate delay time can be realized with an improved S/SUP 2/L gate.     

Inverter light triggered thyristor with unique arm-structure amplifying gate

A 1200-V 200-A directly light triggered thyristor suitable for inverter application has been developed. A new amplifying gate design with a second amplifying stage was used in achieving a factor of 15 to 50 increase in gate sensitivity without any loss indV/dtcapability and only a small (less than a factor of two) reduction in devicedi/dtrating, despite a ten times smaller initial turn-on line length. In all, three versions were made with gate threshold currents down to 1 mA anddV/dtcapabilities to 1000 V/µs. All three types had 60-Hz di/dt capabilitLes of about 250 A/µs at 125 deg TJand turn-off times of approximately 25 µs. The new light sensitive amplifying gate stage design features a gate thyristor region with extending arms for high gate sensitivity, the inner portion of which is just large enough to accommodate initial on-region spreading duriag the short on-time of the gate stage. The arms increase gate sensitivity while contributing very little to the overalldV/dtcurrent. The turn-on speed can be accounted for by most of the inner region being turned on by the photogate pulse. Like regular electrically fired thyristors, a gate overdrive factor is important. With these devices an overdrive factor of about 3 to 5 is needed for highdi/dtturn-on whereas in an electrically triggered device this factor is closer to 10.     

Enhancement-Mode$hboxSi_3hboxN_4hbox/AlGaN/GaN$MISHFETs

Enhancement-mode$hboxSi_3hboxN_4/hboxAlGaN/GaN$metal–insulator–semiconductor HFETs (MISHFETs) with a 1-$muhboxm$gate footprint are demonstrated by combining$hboxCF_4$plasma treatment technique and a two-step$hboxSi_3hboxN_4$deposition process. The threshold voltage has been shifted from$-$4 [for depletion-mode HFET] to 2 V using the techniques. A 15-nm$hboxSi_3hboxN_4$layer is inserted under the metal gate to provide additional isolation between the gate Schottky contact and AlGaN surface, which can lead to reduced gate leakage current and higher gate turn-on voltage. The two-step$hboxSi_3hboxN_4$deposition process is developed to reduce the gate coupling capacitances in the source and drain access region, while assuring the plasma-treated gate region being fully covered by the gate electrode. The forward turn-on gate bias of the MISHFETs is as large as 7 V, at which a maximum current density of 420 mA/mm is obtained. The small-signal RF measurements show that the current gain cutoff frequency$(f_T)$and power gain cutoff frequency$(f_max)$are 13.3 and 23.3 GHz, respectively.     

MOS characteristics of substituted Al gate on high-/spl kappa/ dielectric

Substituted aluminum (SA) metal gate on high-/spl kappa/ gate dielectric is successfully demonstrated. Full substitution of polysilicon with Al is achieved for a Ti-Al-polysilicon-HfAlON gate structure by a low-temperature annealing at 450/spl deg/C. The SA gate on HfAlON dielectric shows a very low work function of 4.25eV, which is well suitable for bulk nMOSFETs. The SA process is fully free from the Fermi-level pinning problem. In addition, the SA process also shows improved uniformity in leakage current distribution compared to fully silicided metal gate.     

Characteristics of sub-1/4-μm gate surface channel PMOSFET'susing a multilayer gate structure of boron-doped poly-Si on thinnitrogen-doped poly-Si

This paper reports the effects of a new p⁺ gate structure (MBN gate) on the properties of surface channel PMOSFET's with an extremely thin gate oxide. The MBN gate is a multilayer gate structure of boron-doped poly Si on thin nitrogen-doped poly-Si. The thin nitrogen-doped Si layer effectively suppresses boron diffusion, so that the gate poly Si can be doped with boron in high concentration without the fear of boron penetration. Gate depletion effects are well suppressed. Effective hole mobility is improved due to the reduction of the initial interface state density. The hot-hole induced interface state generation is shown to be the dominant clause of degradation in the 1/4-μm level PMOSFET's, and less Gm degradation is found in the MBN-gate PMOSFET's than in conventional p⁺-gate PMOSFET's. Finally, with respect to the reliability of the gate oxide, a conventional p⁺ gate with boron penetration exhibits an increase in short-time defect related breakdown during constant-current FN stressing. Short-time defect-related breakdown is not observed in the MBN gate but a slight decrease in charge to breakdown     

Laser thermal processing of amorphous silicon gates to reduce poly-depletion in CMOS devices

One major challenge in advanced CMOS technology is to have adequate dopant activation at the polycrystalline silicon (poly-Si) gate/gate oxide interface to minimize the poly-Si depletion effect. In this paper, laser thermal processing (LTP) was employed to fabricate single or dual-layer poly-Si-gated MOS capacitors with ultrathin gate oxides. Capacitance-voltage data show that the carrier concentration at the poly-Si gate/gate oxide interface increases substantially when the devices are subjected to LTP prior to a rapid thermal anneal (RTA). Thus, LTP readily reduces the poly-depletion thickness in MOS devices. For p/sup +/-gated capacitors, this is achieved with boron penetration that is equivalent to the control sample with 1000/spl deg/C, 5 s RTA (without LTP). In addition, results from secondary ion mass spectrometry indicate that the concentration of dopants near the critical gate/gate oxide interface increases significantly after a post-LTP anneal, in good agreement with the electrical data. Time-dependent dielectric breakdown studies show that the gate oxide reliability is not degraded even after LTP at high fluences.