I have found several interpretations of the sub threshold swing.One is " The subthreshold swing of a device is defined as the change in gate voltage which must be applied in order to create a one decade increase in the output current ". The other one is that the amount of gate voltage that is needed to bring down the sub threshold current by one decade . which of them is correct and in which context. Are these two interpretations same but for the former case device is in strong inversion and the latter case the device is in weak inversion ? What i understood is to have large change in the output current low gate voltage should be enough which implies sub threshold swing should be smaller.If this is true then how are tunneling FETs better over the conventional MOSFETs.(where in tunneling FETs we say the sub threhold swing is not limited?)
I'll give it a shot. The subthreshold swing definition describes an exponential behavior of the current as a function of voltage. The sign of the parameter should depend on the type of the device (n- or p-), so it's easier to just say that some voltage difference is required for a decade of change in current. Except for the sign it's the same number up or down (if +50mV gives 10 times more current, -50mV should give ten times less). As for the question which device is better... that depends. If you need very low threshold voltages, you don't get a choice, there will be leakage. If you need high operating voltages, then you need thick oxides, which means little leakage and the effect doesn't play much of a role. That used to be the case for most mosfets in the past, but at the current integration density we don't get the next to ideal devices of the past any longer, whether we like that, or not.
Subthreshold swing is also the inverse of the subthreshold slope. On a graph of Ids(Vgs) with logarithmic (base 10) axis for Ids the subthreshold slope is found as the straight-line approximation of the subthreshold current, normally expressed in units of decades/mV. Subthreshold swing is expressed in units of mV/decade. Both express the same, however subthreshold swing is more often used.
Are these two interpretations same but for the former case device is in strong inversion and the latter case the device is in weak inversion ?
When subthreshold swing is calculate and quoted it is usual that only the magnitude of the subthreshold swing is of interest, so the sign from the slope is removed. Both cases are for weak inversion (=subthreshold)
What i understood is to have large change in the output current low gate voltage should be enough which implies sub threshold swing should be smaller. If this is true then how are tunneling FETs better over the conventional MOSFETs.(where in tunneling FETs we say the sub threhold swing is not limited?)
Subthreshold swing should normally be minimized when it is allowed by the application. Small subthreshold swing means better channel control, e.g. improved Ion/Ioff, which usually means less leakage, and less energy. For subthreshold circuits it also means better performance. Silicon MOSFETs have a theoretical minimum subthreshold swing of about 60mV/decade for room temperature. FETs using tunneling in some manner can achieve much better (I have seen 5mV/decade quoted), how is another story altogether and I suppose they are not without problems, otherwise what are they waiting for?