Recitation 7

From n+ p diode to MOS structure

6.012 Spring 2009

Recitation 7: From n+p diode to MOS structure

Consider a p-n diode with very asymmetric doping:

xno xpo xno

= =

2 s B Na q (Na + Nd )Nd 2 s B Nd q (Na + Nd )Na xpo

The SCR is entirely on the lightly doped side. xdo xpo = 2 s B qNa

Q is the charge per unit area (charge/cm2 ) and (x) is the space charge density (charge/cm3 ).

Recitation 7

From n+ p diode to MOS structure

6.012 Spring 2009

Now if we insert an insulator layer (oxide) in between n+ and p, we obtain our MOS structure.

Recitation 7

From n+ p diode to MOS structure

6.012 Spring 2009

Same doping level: Nd = 1020 cm3 , Na = 1016 cm3 How does the depletion region xdo compare with xpo above? We can take some intuitive guess rst. Then let us look at the problem quantitatively. QG = qNa xdo Coulomb/cm2

Electric eld is the integration of charge density/

ox E (0 )

or

ox :

At the boundary,

= =

+ s E (0 ) s ox

E (0 ) Eox , E (0+ ) Es Eox Es 3Es

Recitation 7

From n+ p diode to MOS structure

6.012 Spring 2009

VB,O is potential drop across semiconductor body, while Vox,O is potential drop across oxide. (Note that unit wise: potential, voltage and bias are the same). But B is xed (same as n+ p junction). B = n+ p = Vox,o + VB,O 1 qNa xdo qNa 2 = Eox tox + Es xdo = tox + x 2 s do 2 ox Dene oxide capacitance Cox =
ox

tox

. xdo can be solved from above equation:

s

xdo = Si surface potential:

Cox

1+

2Cox B 1 s qNa 1 qNa 2 x 2 s do

(0) = s = p +

This is very important to calculate n(0), p(0) carrier concentration at the surface!! Now we see that xdo should be less than xpo in the n+ p junction case, because much less potential drop (VB,O compared to B ) a lot of potential dropped across the oxide.

Recitation 7

From n+ p diode to MOS structure

6.012 Spring 2009

Now let us do some calculation: n+ -polysilicon gate, 1020 cm3 and p-type substrate NA = 1017 cm3 . tox = 100 A = 10 nm = 1 106 cm xdo = ? Cox B surface carrier concentration n(0), p(0) 3.9 8.85 1014 F/cm ox = = = 3.45 107 F/cm2 capacitance per unit area tox 1 106 cm = n+ p = 550 mV (420 mV) = 970 mV = 0.97 V
s

xdo = =

Cox

1+

2 2Cox B 1 qN s a

11.9 8.85 1014 Fcm1 3.45 107 Fcm2 1+

1+

2 (3.45 107 )2 (0.97 V) F2 /cm4 1 11.7 8.85 1014 1.6 1019 1017 cm3

= 2.98 106 cm

2 0.97 1 8.85 1.6 102

= 2.98 106 cm (3.83 1) = 8.45 106 cm = 84.5 nm = 845 A On the contrary, if there is no oxide: xpo = 2 s B = qNa 2 11.7 8.85 1014 F cm1 0.97 = 11.2 106 cm = 112 nm 1.6 1019 1 1017

(0) = s = p +

1 1.6 1019 C 1 1017 cm3 1 qNa 2 xdo = 420 mV + times(8.45 106 )2 2 s 2 11.7 8.85 1014 , F cm1 = 0.42 + 0.55 = 0.13 V

n(0) = ni e(q(0)/kT) = 1010 10130 mV/60 mV = 1.46 1012 cm3 n2 i p(0) = = 6.85 107 cm3 n(0)

Recitation 7

From n+ p diode to MOS structure

6.012 Spring 2009

Yesterday, we also talked about what happens when we apply a bias (voltage) VGB means body ground, gate in reference to body (Compare to VD in diode case). It does not matter whether n or p for this denition. Since there is oxide, no matter VGB > 0 or

VGB < 0, there is no current. = still consider like thermal equilibrium n(0) = ni e(q(0)/kT) Tomorrow, we will see more detailed discussion on what will happen to the surface carrier density n(0) or p(0) when we apply VGB . The surface carrier density will determine whether the MOSFET channel can conduct or not.

MIT OpenCourseWare http://ocw.mit.edu

6.012 Microelectronic Devices and Circuits

Spring 2009

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.