UNIVERSITY OF CALIFORNIA College of Engineering Department of Electrical Engineering and Computer Sciences EE 130 Spring 2007 Homework Assignment

#9 Solutions Problem 1: Qualitative BJT Questions Change a Increase base width, WB b Increase minority-carrier lifetime in the base, c Increase base doping, NB d Increase emitter doping, NE Effect on Decrease No Effect (if W<<LB) Decrease Increase Effect on Decrease Increase Decrease Slight Increase
T

Prof. Liu

Effect on dc Decrease Increase Decrease Increase

Explanations: dc is defined to be the ratio of IC to IB in active mode. If IB increases for a given IC, then dc decreases. IB is primarily comprised of two main components: (1) carriers injected from the base into the emitter, and (2) carriers recombining with minority carriers injected from the emitter. If either of these components increases, then dc decreases. An increase in (1) results in a decrease in emitter injection efficiency . An increase in (2) corresponds to a decrease in base transport factor T. If IC increases for a given IB, then dc increases. In a short-base BJT (W<<LB), IC is proportional to the magnitude of the derivative of the excess minority-carrier concentration in the quasi-neutral base region, which in turn is inversely proportional to the quasi-neutral base width W.

a) An increase in WB results in an increase in the quasi-neutral base width W. Since the derivative of
the excess minority-carrier concentration in the quasi-neutral base region decreases in magnitude with increasing W (assuming that W<<LB), the minority-carrier diffusion current in the base decreases. This is the current due to carriers injected from the emitter. Thus, the emitter injection efficiency decreases. The longer W is, the more likely that the injected minority carriers will recombine before they reach the collector-junction depletion region to be swept into the collector, so the base transport factor T decreases. Since the rate of carrier injection from the emitter into the base decreases, and the percentage of these carriers which survive long enough to reach the collector decreases, so IC decreases and hence dc decreases.

b) An increase in

B results in an increase in the average distance that a minority carrier diffuses before it is annihilated, i.e. the minority-carrier diffusion length LB. By increasing LB relative to

W, the probability that a minority carrier will survive long enough to reach the collectorjunction depletion region increases, so T increases. Typically, the base of a BJT is short (i.e. W<<LB), so that the minority-carrier diffusion current in the base is strongly dependent on W rather than LB. Thus, an increase in B has little effect on . (If the base is not short, however, then an increase in B (hence LB) will decrease the magnitude of the derivative of the excess minority-carrier concentration in the quasi-neutral base region, so that the minority-carrier diffusion current in the base decreases. This is the current due to carriers injected from the emitter. Thus, decreases.) Since the percentage of carriers which survive long enough to reach the collector increases, IC increases and hence dc increases.

c) An increase in base doping NB results in more carrier injection from the base into the emitter, so
that the total current flowing across the emitter junction increases; therefore, (ratio of current due to carrier injection from the emitter into the base, to the total current flowing across the emitter junction) decreases. Because the minority-carrier mobility decreases with increasing NB, the minority-carrier diffusivity DB decreases (due to the Einstein relationship), so that the average distance that a minority carrier diffuses before it is annihilated due to recombination decreases. Thus, the probability that a minority carrier will recombine before reaching the collector-junction depletion region increases, so T decreases. Since there is more carrier injection from the base into the emitter, IB increases. Also, since it is more likely that a minority carrier will recombine in the base before reaching the collectorjunction depletion region, IC decreases while IB increases. Thus, dc decreases.

d) An increase in emitter doping NE results in more carrier injection from the emitter into the base,
so that increases. An increase in NE will result in a slightly higher built-in potential of the emitter junction and hence a very slightly larger width of the emitter-junction depletion region. As a result, W will decrease very slightly, making it somewhat more likely that a minority carrier will survive long enough to reach the collector-junction depletion region, so T increases very slightly. (This effect is small.) Since carrier injection into the base increases, IC increases and hence
dc

increases.

Problem 2: BJT Current Components and Performance Parameters Given an npn BJT biased as follows: kT V BE = ln[ 10 10 ] = (60 mV )(10 ) = 0.60V q a) First we calculate the built-in potential for each junction:

V BC = 0V

Vbi , BE

NB kT 10 16 = 0.56 + ln[ ] = 0.56 + .026 ln( 10 ) = 0.92V q ni 10 N N kT 10 1610 15 ln[ B C ] = .026 ln( ) = 0.66V q ni 10 10

Vbi , BC =

Then we can obtain the widths of the depletion regions on the n-side and p-side of each junction:

x n, BE = [2 x p , BE = [2 x p , BC = [2 x n , BC = [2
Therefore,

Si N B 1 ( )(Vbi , BE V BE )]1 / 2 = 2.03 10 7 cm = 2.03nm q NE NB + NE

Si N E 1 ( )(Vbi , BE V BE )]1 / 2 = 2.03 10 5 cm = 0.2m q NB NB + NE

Si N C 1 ( )(Vbi , BC V BC )]1 / 2 = 88.5 10 7 cm = 0.089 m q N B N B + NC

Si N B 1 ( )(Vbi , BC V BC )]1 / 2 = 0.885 10 4 cm = 0.89 m q NC N B + NC

W ' E = WE xn ,BE = 1 m 2.03nm WE = 1 m

W ' B = WB x p , BE x p , BC = 0.5 m 0.2 m 0.089 m 0.21 m

W ' C = WC xn, BC = 2 m 0.89 m 1.11 m

b) We need to determine the minority-carrier diffusivity and diffusion length for each of the quasineutral regions of the BJT, in order to calculate the various current components:
E = 150 cm 2 / V s
D E = E kT = 150 cm 2 V 1 s 1 0.026 V = 3.9 cm 2 / s q

L E = D E E = 3.9 10 6 cm 2 =1.97 10 3 cm = 20 m

B = 1248 cm 2 / V s
D B = B kT = 1248 cm 2 V 1 s 1 0.026 V = 32 .3 cm 2 / s q

L B = D B B = 32 .3 10 6 cm 2 = 5.68 10 3 cm = 60 m

C = 458 cm 2 / V s
DC = C kT = 458 cm 2 V 1 s 1 0.026 V =11 .8 cm 2 / s q

LC = DCC = 11 .8 10 5 cm 2 = 0.011 cm =110 m

The excess minority-carrier concentrations at the edges of the depletion regions are

nB 0 (e qV BE / kT 1)

10 20 10 10 = 1014 cm 3 17 10 10 20 10 10 = 1012 cm 3 1018 10 20 0 (e 1) = 0cm 3 10 15

pE 0 (e qV BE / kT 1)

pC 0 (e qV BC / kT 1) =

I Ep = qA

DE cosh( WE ' / LE ) qV BE / kT pE 0 e 1 LE sinh( WE ' / LE )

)
( )

I Ep = 1.6 10 19 C (10 7 cm 2 )

cosh( 1 10 4 cm / 20 10 4 cm ) 3.9cm 2 / s 10 2 cm 3 10 10 1 20 10 4 cm sinh( 1 10 4 cm / 20 10 4 cm )

I Ep = 6.2 10 10 A = 0.62 nA

I En = qA

DB nB 0 LB

cosh( W / LB ) sinh( W / LB )

1 (e qV BE / kT 1) sinh( W / LB ) e qV BC / kT 1

)]
( )

I En = 1.6 10 19 C (10 7 cm 2 )

cosh( 0.21 10 4 cm / 60 10 4 cm ) 32 .3cm 2 / s 4 10 cm 3 [ 10 10 1 0] 4 4 4 60 10 cm sinh( 0.21 10 cm / 60 10 cm )

I En = 2.5 10 6 A = 2.5A

I Cp = qA

DC cosh( WC ' / LC ) qV BC / kT pC 0 (e 1) LC sinh( WC ' / LC )

I Cp = qA

DC cosh(WC ' / LC ) 0 ( 0) (e 1) LC sinh(WC ' / LC )

I Cp = 0 A

I Cn = qA

DB nB0 LB

1 sinh( W / LB )

W L (e qV BE / kT 1) cosh (W // LBB )) e qV BC / kT 1 sinh(

)]
( )

I Cn = 1.6 10 19 C (10 7 cm 2 )

32 .3cm 2 / s 4 1 10 cm 3 [ 10 10 1 0] 6 10 3 cm sinh( 0.21 10 4 cm / 6 10 3 cm )

I Cn = 2.5 10 6 A = 2.5A

c) I E = I En + I Ep = 2.5A + 0.62 nA 2.5A I C = I Cn + I Cp = 2.5A + 0 A = 2.5A

I B = I E I C = 0.64 nA
d)

(Including all significant digits!!)

I En 1 I En + I Ep

T
dc

I Cn 1 I En IC 3882 IB

Problem 3: Ebers-Moll Model

Note that the diode current I is the BJT emitter current IE. From the Ebers-Moll equation for IE, I = I E = I F 0 (e qVEB / kT 1) R I R 0 (e qVCB / kT 1) (Equation 1) However, we need to express I as a function of VA, rather than VEB and VCB, i.e. I should be expressed as a function of (e qV A / kT 1) . First, lets find a relationship between (e qV EB / kT 1) and (e qVCB / kT 1) so that we can simplify Equation 1. We can do this by noting that IB = 0 so that IE = IC. Equating the Ebers-Moll expressions for IE and IC:

I F 0 (e qVEB / kT 1) R I R 0 (e qVCB / kT 1) = F I F 0 (e qVEB / kT 1) I R 0 (e qVCB / kT 1) (1 F ) I F 0 (e qVEB / kT 1) = (1 R ) I R 0 (e qVCB / kT 1)

Therefore (e qVEB / kT 1) = (e qVCB / kT 1) where

(1 R ) I R 0 (1 F ) I F 0

(Equation 2)

Next, lets find a relationship between VCB and VA: V A = VEC = VEB +VBC = VEB VCB VEB = V A +VCB so

e qVEB / kT 1 = e qVA / kT eVCB / kT 1 = e qVA / kT (e qVCB / kT 1) + (e qVA / kT 1)

Eliminating VEB in the above equation by applying Equation 2, and then combining terms, we obtain

(e qV A / kT + )( e qVCB / kT 1) = (e qV A / kT 1) (e qVCB / kT 1) = (e qVEB / kT 1) = (e qV A / kT 1) (e qV A / kT + )

qV A / kT

(Equation 3)

(e 1) qV A / kT (e +)
(Equation 4)

Applying Equations 3 and 4 to Equation 1, and rearranging terms, we obtain

(1 R F ) I R 0 I F 0 (e qVA / kT 1) I= (1 F ) I F 0 e qVA / kT + (1 R ) I R 0
Note that in this configuration, the BJT does not exhibit ordinary diode-like characteristics:

Under forward bias with VA greater than several kT/q, the second term in the denominator is small relative to the first term in the denominator, so that

(1 R F ) I R 0 I R0 (1 F )

Thus, the forward diode current is essentially the reverse-bias current of the collector pnjunction. For large reverse bias with |VA| larger than several kT/q,

(1 R F ) I F 0 I F 0 (1 R )

Thus, the reverse diode current is essentially the reverse-bias current of the emitter pn-junction. Therefore, very little current flows in either direction for a BJT connected in this configuration.

Problem 4: BJT Deviations from the Ideal a) The Early voltage (VA) is given by the following equation:

VA =
where CJC has units of Farads per cm2:

qN BW C JC

C JC =
Thus,

Si
Wdep , BC

x p , BC + xn , BC

Si

(10 12 F / cm ) = 1.110 8 F / cm 2 (0.089 + 0.89 ) 10 4 cm

VA =

(1.6 10 19 C )(10 16 cm 3 )( 0.2 10 4 cm ) = 3.1V 1.1 10 8 F / cm 2

b) Mirroring Exercise 11.8 in Pierret for an npn BJT, we arrive at the following equation for VBC at punchthrough:

VBC = Vbi , BC

(WB x p , BE ) 2 2 si NC ( ) q N B ( NC + N B )
(0.5 10 4 cm 0.2 10 4 cm ) 2 2 1.0 10 12 F / cm 1015 cm 3 ( 16 ) 1.6 10 19 C 10 cm 3 (1.1 1016 cm 3 )

VBC = 0.66V

VBC = 6.8V
Therefore,
V PT V CE , PT =V CB +V BE = 6.8V +0.6V = 7.4V

c) The breakdown voltage (VCE0) is determined as follows: First, calculate the reverse breakdown voltage of the base-collector junction (VCB0), considering that it breaks down when the peak electric field

max

=5 105 V/cm:

max =

qN B q NC N B x p , BC = [ 2 ( )(Vbi , BC + VCB 0 )]1 / 2 Si Si N C + N B

VCB 0 =

Si ( N C + N B ) ( max ) 2 Vbi , BC 2qN C N B

VCB 0

1 10 12 F / cm(1.1 1016 cm 3 ) = (5 10 5 V / cm ) 2 0.66V 19 15 3 16 3 2 1.6 10 C 10 cm 10 cm

VCB 0 897V
Therefore,

VCE 0 =

VCB 0 897V = 14.4V dc + 1 3882 + 1

d) Comparing the answers obtained in parts (b) and (c), we see that the punchthrough voltage (VPT) is smaller (in magnitude) than the (common-emitter) breakdown voltage (VCE0). Therefore, punchthrough will be observed before breakdown in the common-emitter output characteristics of this BJT.