tesi R. Miscioscia.pdf - EleA@UniSA

26 Working Principles
1.5 Models for the extraction of the channel
current
In the study of OTFTs, it’s necessary to consider the
contribution to Source and Drain currents due to a non ideal
behavior of the gate dielectric which acts as a leaky insulator. This
implies a non-perfect insulation of the gate versus the channel and
then versus S/D contacts which are usually named gate-leakages.
The presence of such parasitic currents which reflects in a non-null
gate current IG ≠0 and then in a static power dissipation in the gate,
makes not straightforward the determination of the channel current
(Ich) which otherwise can be confused with the drain-to-source
current (IDS). This non-ideality makes complicated the derivation of
the channel current (the one which is usually predicted by the
physical models i.e. in eq. 18 or in the ubiquitous utilizations of the
MOSFET equations). In a deeper detail, from static measurements,
we can measure ID, IS and IG but for physical modeling purposes,
Ich is the parameter which cannot be directly measured but which is
the only one, in a static context, which can be thought directly
connected to charge transportation in the semiconductor channel.
Nevertheless, because of the failure of the approximation of
Ich to IDS, performance parameters for the OTFT, transconductance,
field-effect mobility, on/off ratio, threshold voltage, on-set voltage
etc. cannot be directly extracted then also modeling can be
sometimes not physically consistent because of the presence of
these parasitic currents. Then, also the extraction of working
parameters under thermally activated conditions suffers from an
undesired shift in the electrical behavior from the real phenomenon
and also from theoretical models.
For devices having negligible gate-leakages, it’s easy to
suppose IS ≈ ID ≈ IDS ≈ Ich otherwise channel current has to be
estimated in an indirect way. Such kind of non-ideality is much
more relevant for long-channel devices because of a larger leakage
section [10][11]. In fact, Ich grows linearly with the W/L ratio as in
the first-order MOSFET model but the gate current IG is related to
channel’s surface (W·L) then, in a perspective of first
approximation,