Professional Documents
Culture Documents
Minimizing Errors of Four Point Probe Measurements On Circular Wafers
Minimizing Errors of Four Point Probe Measurements On Circular Wafers
This content has been downloaded from IOPscience. Please scroll down to see the full text.
(http://iopscience.iop.org/0950-7671/44/1/312)
View the table of contents for this issue, or go to the journal homepage for more
Download details:
IP Address: 147.8.31.43
This content was downloaded on 01/10/2015 at 15:08
Abstract. The errors introduced in the use of the four point probe determination of the
resistivity of circular semiconductor wafers are examined. An optimum choice of inter-
probe spacing is found to be dependent on the ratio of the probe spacing to both the radius
and the thickness of the disk. The calculations allow the optimum probe spacing for any
set specimen geometry to be obtained.
1. Introduction If the probes PI, Pz,P3 and P4 are displaced distance 8x1
The use of the four point probe method is well established Sxz 8xgand 8x4 along the line of contact, then the increase
for the measurement of the resistivity of silicon wafers in the voltage appearing between P2 and P3will be
(Hargreaves and Millard 1962). On the assumptions of
uniform resistivity and ohmic current flow calculations have
been made to relate the measurements to the resistivity of
the material. For thin wafers, the resistivity is often derived
from the formula (Uhlir 1955) Substitution of (1) into (2), with s, = sy = s, = s, gives
TW AV
P=mI
where w is the thickness of the wafer, I is the current which
flows through the outer probes and AV is the voltage devel-
If the displacements are assumed to be random with a mean
oped across the inner probes. For this expression to give a
close approximation to the correct value of resistivity, the displacement cc the probable error in the measurement be-
following conditions must be fulfilled: comes
&AV) - 45 a
(i) the interprobe spacing s must be constant; AV 21n2s’ (4)
(ii) the probes must be sdiciently far from the periphery
of the wafer for it to have a negligible effect;
(iii) the wafer must be sui%ciently thin for there to be
negligiblepotential variation across the thickness of the wafer.
It is proposed to consider each of these conditions in turn
r““1
and to estimate the optimum interprobe spacing so as to
minimizethese effects for a wafer of given dimensions.
(5)
53
R. Hull
For s/w >, 1 the calculated values can be closely approxi-
mated by the formula
Apparent resistivity
True resistivity
=1 +
2.63 exp ( - 3 4 s / w ) (8)
so that the apparent fractional increase in resistivity may
be written
9 = 2-63 exp ( - 3 . 4 4 4 ~ ) . (9)
P
5. Optimum probe spacing
If the combined effect of crystal periphery and thickness
and interprobe spacing and wander is neglected then the
apparent resistivity calculated by equation ( 1 ) has a total
error of
1.61a 2 . 1 6 ~ ~
Figure 2. Image system for probes. - S t- U2 +
2.63 exp ( - 3 4 4 s l w ) . (10)
X
(a2 - t y + (U2 - (p + $s) (p - +S)}Z - U4 + t Z (2pZ + 2sp + 2.9)
(U2 - t2)Z + (U2 - (p + $s) (p + +S)}Z - a4 + tz (2p2 + 4sp + Q Z ) (6)
- ( 4 9 + 4pz - 3sZ)d
U8
U8 -(49 +4 2 + 3 s y l-'12
*