Professional Documents
Culture Documents
Atom to Transistor
01.15.2003
I Vacuum Level
0eV E
VG VD
Gate
Empty EC
S INSULATOR
D
O L R
U
R CHANNEL
A
I
EF
C
E
z
N
EV
INSULATOR
fo
x 0 1
I Vacuum Level
VG VD 0eV E
EC
Gate
S INSULATOR
Empty
D
O L R
U
R
CHANNEL A
I
EF
C
E z
N EV
INSULATOR
f0
x 0 1
• Consider a simple system with wide spacing of electron energy levels (a small
molecule)
• 4 electrons in the system considering spin up and down fill the first 2 energy
levels.
• The Fermi energy is a level for which at T=0K the levels below it are full and the
levels above it are empty. For our system at 0 K, the Fermi level will be half way
between levels 2 and 3.
Fixed Small
ID VD
• The smaller the system, the more the
spacing between the energy levels. VG
I Vacuum Level
VG VD
Gate
µ1=Ef + qVD/2
S INSULATOR D
O
U
L R
qVD
A
µ2=Ef - qVD/2
R CHANNEL
I
C N
E
z
INSULATOR
x
• When VD is applied between the source and • The reason why the shift in
drain, system is driven out of equilibrium; the Fermi levels is symmetric will
Fermi energies in the source and drain will be discussed later. For now
not be at the same level any more. consider the fact that division
• Total energy difference between the two of voltage is very important for
Fermi levels (chemical potentials) µ1 and µ2 is large bias but not so important
qVD. for small bias which is in fact
• i.e.01.15.2003
1V x q = 1eV = 1.6 x 10-19J the case here.
Current Flow due to 16:59
Difference in Agenda
Vacuum Level
E 0eV Top Levels
μ1 = EF + qVD / 2 E
μ2 = EF − qVD / 2
Bottom Levels
• If you do not have any • Why don’t bottom and top levels contribute
levels between µ1 and µ2 to current? Since we have two different Fermi
then current will not flow levels, we have two different Fermi functions.
• If you have a level between Consider their value for high energies and for
µ1 and µ2 current will flow low energies; you’ll find that they have the
because left side (at μ1 ) same value of 0 (For high E) and 1 (For low
keeps filling up the level and E) - In other words, both contacts have the
right side (atμ 2 ) keeps same agenda for these levels; hence no
emptying it.
01.15.2003 current flow due to these levels.
Average Number of 24:19
Electrons
01.15.2003
Current Flow Left and 28:28
Right
Vacuum Level
• Use Taylor expansion to get an expression for current at
small voltages: f1(ε) = f0(ε - µ1) 0eV
f2(ε) = f0(ε - µ2) therefore:
f1- f2 = (df0 / dE) (µ2 - µ1) I1 I2
= -(df0 / dE) qVD
Therefore for small voltages:
2q γ1γ 2 2 q 2 γ 1γ 2 ⎡ df 0 ⎤ µ1
I= [ f1 (ε) − f 2 (ε)] =V ⎢ − dE ⎥
h γ1 + γ 2 h γ1 + γ 2 ⎣ ⎦ E =ε − Ef ε
Use E = ε-Ef since µ1= Ef + qVD/2 and µ2= Ef – qVD /2
µ2
• Note: 2q 2 , dimensions of conductance
h
γ 1γ 2
, dimensions of energy
γ1 + γ 2 ⎡ df 0 ⎤
⎢ dE ⎥ , dimensions of inverse energy
01.15.2003 ⎣ ⎦ E =ε − Ef
46:01
No Upper Limit?
df 0
What does − look like? • From the equation
dE
2 q 2 γ 1γ 2 ⎡ df 0 ⎤
I =V −
E E 1/(4k T) h γ 1 + γ 2 ⎣ dE ⎥⎦ E = ε − E f
⎢
B
Ef it would appear that there is no upper
limit on the conductance
2 q 2 γ 1 γ 2 ⎡ df 0 ⎤
−
f0 h γ 1 + γ 2 ⎣ dE ⎥⎦ E = ε − E f
⎢
1 −
df 0
• However, this is not true. An upper
dE limit does occur due to energy
broadening. The maximum value of
Peak ≅ 1/(4kBT) ; Width ≅ kBT Area conductance for one level device is:
= 1. 2q 2
≅ 77 . 4 μ S ≅ 1
As tÆ0, it can be approximated as a h 12 . 9 k Ω
delta function.
01.15.2003 …to be discussed in more detail next