8.5 Sspd_chapter 6_part 4_the instability of mos in 60s and its

SSPD_Chapter 6_Part 4_The instability problem of MOS devices and its solution

In 1965 a very important paper was published "Investigation of Thermally Oxidized Silicon Surfaces Using Metal-Oxide-Semiconductor Structures" in Solid State Electronics in 1965. Due to electrical instability of Gate Oxide in MOS devices, commercial MOS ICs were not available in 60s. To overcome this instability, extensive studies were made of gate oxides using C-V measurements method.

Figure 6.32 MOS C-V measurement method under steady state DC gate voltage bias.

MOS capacitance is Aluminum Gate contact deposited on dielectric SiO ₂ thermally grown on P type Si-substrate. Here we assume that there are very little charges or defects in the oxide or at the interface of Si/SiO ₂ or in the P type Si-Substrate.

There are four kinds of charges/defects present in Si/SiO ₂ system which have been identified through C-V measurement curves namely:

1. Fixed Oxide Charge (Q _f ) – this is about 10 ⁹ to 10 ¹¹ cm ^-2 . This lies in the gate oxide 2nm distant from the interface. It likely to be associated with the transition from Si to SiO ₂ . It is positive fixed charge and does not vary during the operation of the device.
2. Interface trapped Charge ( Q _it ) also known as surface states – this is possibly the incompletely oxidized Si atoms creating unsatisfied or dangling bond in the oxide layer but very near to the interface. They act as traps which can trap the mobile carrier of the channel. It will trap electrons in NMOS hence act as (-)ve charge and it will trap holes in case of PMOS acting as (+)ve charge. These act very much like bulk deep-level traps. The energy associated with the interface traps lie all along the forbidden gap but more so near the conduction band and the valence band. Oxidizing Si causes Q _it = 10 ⁹ to 10 ¹¹ cm ^-2 eV ^-1 . This is of the same order as fixed oxide charge and it is experimentally established that both have the same origin hence higher fixed oxide charge implies higher interface trapped charges.
3. Mobile Oxide Charge ( Q _m ) – these are generally sodium (Na ⁺ ) and potassium (K ⁺ ) ions. These were the main culprits which were making MOS ICs technically unviable in 60s. These mobile charges were causing the instability in the threshold voltages. These were controlled by Phosphate doped SiO ₂ (P-glass) and by replacing the tungsten filament for electron beam deposition of Aluminum Contact. Tungsten filament was the main source of Alkali ion contamination. It was also experimentally established that the shift in V _th is inversely proportional to C _ox hence in subsequent years with scaling as C _ox increased the instability of the threshold voltage became less of a problem.
4. Oxide trapped Charges ( Q _ot ) – these are defects in the oxide layer which can trap electron or holes or both. These defects are caused due to broken Si and O bond due to ionizing radiation or due to process steps such as plasma etching or ion-implantation. These can be annealed and thereby minimized. But if they are not repaired they will cause trapped charges which will effect the threshold voltage of the MOS device. With scaling there is hot carrier effect and injection of hot carriers into oxide which may be trapped in these defects or the injected hot carriers may themselves cause defects leading to trapping of charges. These trapped charges will cause threshold instability.