Effects of dimensions on the sensitivity of a conducting polymer
microwire sensor
Cheng Luo
Ã
, Anirban Chakraborty
a
Institute for Micromanufacturing, Louisiana Tech University, 911 Hergot Avenue, Ruston, LA 71272, USA
article info
Article history:
Received 26 August 2008
Received in revised form
20 November 2008
Accepted 24 November 2008
Keywords:
Conducting polymers
Microwire sensors
Surface-to-volume ratio
Sensitivity
Intermediate-layer lithography
abstract
It is commonly considered that the sensitivity of a microsensor increases with its increasing surface -to-
volume ratio. However, it is not exactly clear how the surface-to-volume ratio affects the sensitivity of a
conducting polymer microsensor. The change in any of the three geometrical dimensions (i.e., length,
width and thickness) of a microsensor changes the surface-to-volume ratio. In designing a microsensor
of desired sensitivity, it is important to know the effect of each dimension on the sensitivity for properly
defining the sizes and shapes of the microsensor. As such, in this work, we have investigated the effects
of each individual dimension on the sensitivity of a conducting polymer microwire sensor. Polypyrrole
(PPy) and Poly (3,4-dimethlydioxythiophene) poly(styrenesulfonate) (PEDOT–PSS) microwire sensors of
different dimensions were fabricated using an intermediate-layer lithography (ILL) method. They were
further employed to detect methanol and acetone vapors at concentrations in the range of 0.6–7.1 parts
per thousand (ppt). The corresponding three relationships between the three geometrical dimensions

for various chemical analytes including water vapor [5,9,10,18],
volatile organic gases [5,7,8,11–13,16,18,19] (such as methanol,
acetone, alcohol, and ethanol), industrial gases [14] (such as
ammonia, NO
x
,CO
x
,SO
2
,H
2
S, O
2
and H
2
), glucose [15], and
antigens [17].
Compared to film sensors, microsensors generally have
exhibited higher sensitivity in detecting analytes of low concen-
trations. It is normally considered that the higher sensitivity is
induced by the higher surface-to-volume ratio of the micropat-
terns. However, it is not exactly clear how the surface-to-volume
ratio affects the sensitivity of a conducting polymer microsensor.
A recently developed intermediate-layer lithography (ILL) enables
us to properly fabricate conducting polymer microsensors. There-
fore, in this work, PPy and PEDOT–PSS microwires of different
dimensions have been fabricated using the ILL method and
subsequently applied to detect methanol and acetone vapors of
concentrations in the range of 0.6–7.1 parts per thousand (ppt).
ARTICLE IN PRESS

conducting polymer microsensor actually depends on the
sensitivity of a unit block, and then discuss ways to increase the
sensitivity of the unit block. As what has been done by many
researchers [8,10–13,17,18], the sensitivity index (SI) is defined as
(R
exposure
ÀR
Base
)/R
Base
, where R
Base
and R
exposure
represent the
resistances of a sensor before and after exposure to a target,
respectively. The SI indicates how large the sensor response is
to a particular concentration, and is used as a measure of the
sensitivity of the corresponding sensor.
Geometrically, the sensing area of a film sensor can be
modeled to be made up of multiple microwires of unit width,
connected in parallel between the opposite edges at the electro-
des. These microwires may be further divided into blocks of unit
top area along the entire length (Fig. 1). These blocks of unit area
may be regarded as individual ‘‘chemiresistor’’ elements with the
base resistance r
Base
i;j
, which respond to the various concentrations
of analytes with unique changes in conductivity. These ‘‘chemir-

be (n/m) Â r
0
i,j
. The SI for the film sensor would be
ð
D
R=RÞ
Total
¼ f½ðn=mÞÂr
0
i;j
Àðn=mÞÂr
Base
i;j
=½ðn=mÞÂr
Base
i;j
g
¼ð
D
R=RÞ
Block
. (1)
It is observed from Eq. (1) that the SI of a film sensor equals that
of a single unit block, which does not depend on how many unit
blocks this film sensor has.
In deriving Eq. (1) for a film sensor, it is assumed that the unit
blocks have the same sensitivity. This assumption holds when the
top surface of the film is much larger than the side surfaces.
During the detection, a film sensor has five surfaces exposed to a

affect the ratio much for a fixed t. For example, when t ¼ 1
m
m,
a ¼ 10
m
m and b ¼ 10
m
m, the ratio is 1.4
m
m
À1
. The reduction of
both a and b by half yields a new ratio of 1.8
m
m
À1
, while the
reduction of t by half leads to a new ratio of 2.4
m
m
À1
. In short, the
thickness is the most important dimension among the three in
affecting surface-to-volume ratio. In this work, we explored the
effects of the surface-to-volume ratio on the sensitivity of a
microsensor. We further examined the effects of each individual
dimension on the sensitivity of a microsensor. We particularly
ARTICLE IN PRESS
Base
r

Conducting polymer micropatterns were generated using the
ILL method [20–22] as follows (Fig. 3): (i) a layer of multiple
conducting polymer coatings and a layer of a non-conducting
polymer polymethyl methacrylate (PMMA) are heated up to the
printing temperature, which is above the glass transition
temperatures (T
g
) of all polymers (Fig. 3a), (ii) a Si mold of
desired patterns and the substrate are brought into physical
contact by applied pressure, followed by subsequent cooling
(Fig. 3b), and (iii) they are separated when their temperatures
are below the lowest T
g
of all polymer materials, completing the
pattern transfer from the mold to the conducting polymer
layer (Fig. 3c). The three-step patterning process of the ILL is
identical to that of the hot-embossing process [23]. The critical
difference is that the substrate in the hot-embossing process has
only the layer of the material to be printed, while the substrate in
the ILL approach involves an additional intermediate layer
of a non-conducting polymer. As a result of this difference,
the conducting polymer patterns would be electrically isolated
over the insulating intermediate layer, and patterns would be
imprinted on the conducting polymer layer even if there were
height differences existing between the features of the mold
[20–22].
PMMA was chosen as the intermediate-layer material, because
it is a good hot-embossing material. The PMMA has small thermal
expansion coefficient of $5.0 Â 10
À5

on these conducting polymers. The mold was slowly inserted into
the substrate to avoid the dynamic effects in the embossed
polymer. The silicon molds were fabricated using conventional
ultraviolet lithography and deep reactive ion etch. The embossing
temperature and pressure were 150 1C and 50 MPa, respectively.
Fig. 4 shows a representative set of generated PPy microwires
which have been used for sensing. Every sensor comprised six PPy
or PEDOT–PSS microwires which were connected in parallel.
Ag epoxy was placed at the two ends of these microwires as
contact pads for electrical connection. The dimensions of micro-
wires were changed to vary the surface-to-volume ratios of the
microwires. One type of film and five types of microwire sensors
were fabricated using the ILL method for either conducting
polymer. Tables 1 and 2 give the corresponding dimensions of
these sensors.
ARTICLE IN PRESS
Si mold
Conducting
polymer layer
Intermediate polymer
layer
PMMA substrate
Convex mold
structure
Concave mold
structure
Fig. 3. The three-step procedures to fabricate polymeric patterns using the proposed ILL method: (a) heating of the substrate, (b) insertion of the mold into the two polymer
layers, and (c) separation of the mold and the substrate.
Overall embossed area
100µm

for methanol to 120 s for acetone, since according to preliminary
tests the PEDOT–PSS sensors responded to acetone exposure
within 120 s. After a test, the chamber was purged by nitrogen and
vented. For the next round of testing, the chamber was closed and
the above procedure was repeated for detecting vapors of different
concentrations.
The masses of methanol and acetone were calculated from
their known volumes (i.e., the evaporated volumes) and their
densities at room temperature. The mass of air was calculated
from the known volume (i.e., the volume of the chamber) and the
density of air at room temperature. The concentration of the
methanol was calculated from the ratio between the mass of
methanol and that of air inside the test chamber. The same
applied to acetone. The concentrations of methanol vapor ranged
from 1.3 to 6.4 ppt. This range of methanol concentrations was
about the same as the one reported in [8], which varied from
about 1.5–5.0 ppt. The detection of methanol of lower concentra-
tions (0.049–1.059 ppt) was reported in [13]. The acetone
concentration of this work was varied from 0.6 to 5.8 ppt, which
was lower than the concentration of 12.7 ppt considered in [7]
(that is, 5% of acetone vapor pressure at 21 1C) and below the
range of 104–416 ppt in [19].
5. Sensing results
5.1. Exposure of PPy sensors to methanol vapor
When the PPy film and microwire sensors were exposed to
methanol vapor, response currents at 10 V varied with time in a
wave-like form (Fig. 6). For the microwire sensors of different
surface-to-volume ratios, the peak currents were reached be-
tween 60 and 120 s after the methanol droplet had evaporated,
and the response current varied between 1.5 Â 10

m
m)
Thickness of
PPy layer
(
m
m)
Surface-to-
volume ratio
(
m
m
À1
)
Film 10,000 10,000 0.19 5.155
Microwire Type I 300 5000 0.19 5.162
Microwire Type II 100 5000 0.19 5.175
Microwire Type III 100 2000 0.19 5.176
Microwire Type IV 100 2000 0.13 7.655
Microwire Type V 100 2000 0.25 3.974
Table 2
Dimensions of the PEDOT–PSS sensors using in the tests.
PEDOT–PSS Width
(
m
m)
Length
(
m
m)

2
outlet
N
2
inlet
Microwire
sensor
Fig. 5. Experimental setup to determine the sensitivity of PPy and PEDOT–PSS sensors in detecting methanol and acetone, respectively.
C. Luo, A. Chakraborty / Microelectronics Journal ] (]]]]) ]]]–]]]4
Please cite this article as: C. Luo, A. Chakraborty, Effects of dimensions on the sensitivity of a conducting polymer microwire ,
Microelectron. J (2009), doi:10.1016/j.mejo.2008.11.064
was 5.00 Â10
3
O
. Hence, the effect of contact resistance was also
neglected in considering the PEDOT/PSS sensors. For the PPy
film sensor, the response current reached a peak between 120
and 160 s, and the response current varied between 5 Â10
À7
and
5.29 Â10
À7
A(Fig. 6b). The corresponding resistances varied
between 2.00 Â10
7
and 1.89 Â 10
7
O
. The transient nature of the
response current was due to the fact that the methanol molecules

in determining the corresponding SI, since this gave a
much larger SI compared with the case of adopting the steady
current to calculate R
exposure
.
Except for Type V microwires, the sensitivity increased in the
order: FilmoType IoType IIoType IIIoType IV (Fig. 7). For the
lowest methanol concentration of 1.3 ppt, the sensitivity of PPy
film sensor was 1.6% for a surface-to-volume ratio of 5.155
m
m
À1
,
as compared to PPy microwire Type IV with sensitivity of 36.44%
for a surface-to-volume ratio of 7.655
m
m
À1
. Similarly, at the
highest acetone concentration of 6.4 ppt, the PPy film sensitivity
was 10.4% and Type IV microwire was 55.6%. These results
indicate that in general the sensitivities of these sensors increase
with the increasing surface-to-volume ratios.
PPy film and microwires of Types I, II and III had the same
thickness of 0.194
m
m(Table 1). The sensitivities of the PPy film
and microwire sensors had an approximately linear relationship
with increasing methanol concentrations (Fig. 7). At the lowest
methanol concentration of 1.3 ppt, the sensitivities of PPy film

respectively.
The PPy thicknesses were varied for Types III, IV and
V microwire sensors with their lengths and widths kept constant.
This was done to study the effects of the PPy thicknesses on the
sensitivity responses of the microwires. The thicknesses of the
PPy layers were 0.131 and 0.253
m
m for Types IV and
V microwires, respectively. When the PPy thicknesses were varied,
there were large variations in the surface-to-volume ratios.
Type IV microwires had the highest surface-to-volume ratio and
the highest sensitivities at all the methanol concentration levels
(Fig. 7). The sensitivity of Type IV microwires at the lowest
methanol concentration was 36.4% and at the highest concentra-
tion was 55.6%. These results imply that for the PPy microwires
their thicknesses may have larger effects on sensitivity than the
length and width.
It is also worth pointing out that, although the width of Type I
wires was three times as large as that of Type II wires (they have
the same length and width), their surface-to-volume ratios only
differed by 0.013. Similarly, the 2.5-times difference in the widths
between Types II and III led to only 0.001 difference in their
surface-to-volume ratios. These two comparisons support the
point raised in Section 2. That is, the changes of the length and
width do not cause much change in the surface-to-volume ratio of
a microsensor. However, it is interesting to see from Fig. 7 that
these three types of microwires still had several percents of
difference in their sensitivities of detecting methanol.
It is noted that PPy sensors that other researchers used have
demonstrated different sensitivities. For example, as indicated in

first decreased and then increased back to a steady value a little
lower than its original value. At the initial stage, exposure of
PEDOT–PSS to acetone reduced the conductivity of the PEDOT–PSS
microwires. According to Ruangchuay et al. [26], acetone being a
polar molecule, it dispersed inside the PPy matrix by hydrogen
bonding. This mechanism disrupted the ordered structure and
hence reduced the conductivity of PPy. A similar mechanism may
be playing a role in reducing the conductivity of the PEDOT–PSS
microwires in our case. Alternatively, acetone molecules diffused
inside PEDOT–PSS, expanding the matrix, hindering the flow of
charge carriers and thereby reducing conductivity of the micro-
wires. As the response current reached a minimum, the acetone
vapor diffused out of the PEDOT–PSS due to the fact that the
acetone concentration in the PEDOT–PSS was higher than that in
the environment. This caused the increase in the current. The
sensing response of the PEDOT–PSS to acetone was different from
that in the case when PPy sensors were used to detect methanol.
The sensor current decreased to a minimum after about 90 s of
exposure. The response times of film and microwire sensors were
about the same. The same transient phenomenon was also found,
for example, in [5] when the PEDOT–PSS sensors were employed
to detect methanol and ethanol. However, due to the same reason
addressed in Section 5.1, their steady currents were much
different from the original currents, while in our case the steady
value was just a little lower than the original value. Thus, the
minimum current was used in this work to calculate R
exposure
in
determining the corresponding SI, since this gave a much larger SI
compared with the case of adopting the steady current to

m(Table 2). The responses of the
PEDOT–PSS microwires were more closely placed in the sensitiv-
ity scale than the PPy microwires, while the overall trend was
similar. For the PEDOT–PSS microwires (Fig. 9), the sensitivity
increased from 0.05% for film sensor to 3% for Type III microwires,
ARTICLE IN PRESS
60
55
50
45
40
35
30
25
20
15
10
5
0
1234567
Methanol concentration (ppth)
Sensitivity (%)
Type III
Type II
Type IV
Type I;
Type V
Film
Fig. 7. Sensitivity responses of the PPy sensors at various concentrations of
methanol exposure.

to detect acetone, whose concentrations ranged from 104 to
416 ppt. The corresponding sensitivities varied from 3% to 9%. The
nanowires were synthesized using anodic aluminum oxide
membranes. These results imply that our sensors also generally
have higher sensitivities than the nanowire sensors. It has been
indicated in [19] that, compared with film sensors, these nanowire
sensors do not show higher sensitivities. They considered this was
due to the impact of substrate roughness during film formation of
the film sensors. In other words, different manufacturing
approaches generate different features, making sensors have
different sensitivities, as discussed in Section 5.1. In this work,
ARTICLE IN PRESS
Response current (10
-3
A)
0.60
0.58
0.56
0.52
0.48
0.50
0.66
0.64
0.62
Time (s)
0.54
PEDOT-PSS microwire sensors
Response current (10
-3
A)

C. Luo, A. Chakraborty / Microelectronics Journal ] (]]]]) ]]]–]]] 7
Please cite this article as: C. Luo, A. Chakraborty, Effects of dimensions on the sensitivity of a conducting polymer microwire ,
Microelectron. J (2009), doi:10.1016/j.mejo.2008.11.064
the same manufacturing approach (as well as the same sensing
material) has been used to generate the five types of sensors.
Therefore, the manufacturing effect (as well as the sensing
material) is not a concern here in comparing the sensitivities of
these five types of sensors.
6. Statistical analysis of sensing data
Based on the method of least square [27], a statistical program
SAS has been run to fit the data points for further analyzing the
sensing results and examining the effects of each individual
dimension. We intended to find the relationship of the SI with the
three geometrical dimensions and the vapor concentration. It was
noticed that surface-to-volume ratio is a linear combination
of the inverse of the three geometrical dimensions, and that the
sensitivity should increase as this ratio increases. Therefore, we
assumed that the SI was related with the inverse of these three
dimensions. In addition, it was found from Figs. 7 and 9 that the
SI had an approximately linear relationship with the vapor con-
centration. Therefore, the SI was assumed to be directly related to
the vapor concentration, instead of its higher orders. Let x
1
, x
2
, and
x
3
represent the inverse of the length, the inverse of the width,
and the inverse of the thickness of a microwire, respectively. Set x

þ 1:7x
4
À 10:8. (3)
The related r
2
was 0.76. To see clearly the effects of each
individual dimension on the sensitivity from the above two
equations, let’s consider an example. In designing a conducting
polymer microwire sensor, the initially chosen dimensions could
be 1000
m
m  100
m
m  0.1
m
m. Based on these dimensions, next
we considered how the changes in these three dimensions affect
the sensitivity. Let alternative length, width and thickness be
1000m,100n, and 0.1 l, respectively, where m, n and l were three
positive constants and their values determine the final values of
the three dimensions. Substituting the inverse of these three
dimensions into Eqs. (2) and (3) for x
1
, x
2
, and x
3
, we had
y ¼ 26:5=m þ 2:7=n þ 99=l þ 1:9x
4

that the length had more effects than the width. To see this
clearly, we compared the effects of the length with those of the
width via the detection of acetone using PPy sensors. PPy film and
Types I, II and IIII microwire sensors were chosen to detect
acetone vapors, whose concentrations were 1.3, 3.2, 4.5, 5.8, and
7.1 ppt (Fig. 10). The same setup and testing procedure as
described in Section 4 were used. The overall trend of the sensor
responses was similar to what has been found in the previous two
sets of experiments (Figs. 7 and 9). The sensitivity increased from
3.6% for film sensor to 13.0% for Type III microwires at a
concentration of 1.3 ppt and from 10.1% for film sensor to 29.0%
for Type III microwires at a concentration of 7.1 ppt. For a
particular concentration, the sensitivity increased in the order:
FilmoType IoType IIoType III. Since these sensors had the same
thickness of 0.194
m
m, we could directly compare the effects of the
length and width. The corresponding fitting result was
y ¼ 19301:5x
1
þ 920:0x
2
þ 2:1x
4
À 5:0. (5)
The related r
2
is 0.94. Following the same line of reasoning that
was used to obtain Eq. (4) from Eqs. (2) and (3), by Eq. (5) we had
y ¼ 19:3=m þ 9:2=n þ 2:1x

25
30
35
1.3
Acetone cencentration (
pp
th)
Sensitivity (%)
Type III
Type II
Type I
Film
3.2 4.5 5.8 7.1
Fig. 10. Sensitivity responses of the PPy sensors at various concentrations of
acetone.
C. Luo, A. Chakraborty / Microelectronics Journal ] (]]]]) ]]]–]]]8
Please cite this article as: C. Luo, A. Chakraborty, Effects of dimensions on the sensitivity of a conducting polymer microwire ,
Microelectron. J (2009), doi:10.1016/j.mejo.2008.11.064
For PEDOT–PSS, the surface-to-volume ratio varied from 0.890 to
4.873
m
m
À1
. The PPy film and microwire sensors were exposed to
methanol vapor whose concentrations ranged from 1.3 to 6.4 ppt.
Methanol exposure increased the response current of the PPy
sensors. The PEDOT–PSS film and microwire sensors were
exposed to acetone vapor whose concentrations ranged from 0.6
to 5.8 ppt. The response current of the sensors was reduced upon
exposure to acetone vapor. In general, the sensitivities of the

polymer composite-based chemiresistor, Analytical Chemistry 72 (2000)
1532–1542.
[7] G.A. Sotzing, S.M. Briglin, R.H. Grubbs, N.S. Lewis, Preparation and properties
of vapor detector arrays formed from Poly(3,4-ethylenedioxy)thiophene-
Poly(styrenesulphonate)/Insulating polymer composites, Analytical Chemis-
try 72 (2000) 3181–3190.
[8] M.F. Marbrook, C. Pearson, M.C. Petty, Inkjet printed PPy thin films for vapor
sensing, Sensors and Actuators B 115 (2006) 547–551.
[9] W.A. Daoud, J.H. Xin, Y.S. Szeto, Polyethylenedioxythiophene coatings for
humidity, temperature and strain sensing polyamide fibers, Sensors and
Actuators B 109 (2005) 329–333.
[10] J.H. Cho, J.B. Yu, J.S. Kim, S.O. Sohn, D.D. Lee, J.S. Huh, Sensing behaviors of
polypyrrole sensor under humidity condition, Sensors and Actuators B 108
(2005) 389–392.
[11] C.P. de Melo, B.B. Neto, E.G. de Lima, L.F. B de Lira, J.E.C. de Souza, Use of
conducting Polypyrrole blends as gas sensors, Sensors and Actuators B 109
(2005) 34 8–354.
[12] H.K. Jun, Y.S. Huh, B.S. Lee, S.T. Lee, J.O. Lim, D.D. Lee, J.S. Huh, Electrical
properties of polypyrrole gas sensors fabricated under various pre-treatment
conditions, Sensors and Actuators B 96 (2003) 576–581.
[13] L. Jiang, H.K. Jun, Y.S. Huh, J.O. Lim, D.D. Lee, J.S. Hoh, Sensing characteristics
of polypyrrole-poly(vinyl alcohol) methanol sensors prepared by in situ vapor
state polymerization, Sensors and Actuators B 105 (2005) 132–137.
[14] Q. Ameer, S. Adeloju, Polypyrrole-based electronic noses for environmental
and industrial analysis, Sensors and Actuators B 106 (2005) 541–552.
[15] T. Yamauchi, K. Kojima, K. Oshima, M. Shimomura, S. Miyauchi, Glucose-
sensing characteristics of conducting polymer bound with glucose oxidase,
Synthetic Metals 102 (1999) 1320.
[16] L. Ruangchuay, A. Sirivat, J. Schwank, Electrical conductivity response of
polypyrrole to acetone vapor: effect of dopant anions and interaction

polypyrrole to acetone vapor: effect of dopant anions and interaction
mechanisms, Synthetic Metals 140 (2004) 15–21.
[27] R.L. Plackett, The discovery of the method of least squares, Biometrika 59
(1972) 239–251.
[28] E.F. Fama, J.D. MacBeth, Risk, return and equilibrium: empirical tests, Journal
of Political Economy 71 (1973) 607–636.
ARTICLE IN PRESS
C. Luo, A. Chakraborty / Microelectronics Journal ] (]]]]) ]]]–]]] 9
Please cite this article as: C. Luo, A. Chakraborty, Effects of dimensions on the sensitivity of a conducting polymer microwire ,
Microelectron. J (2009), doi:10.1016/j.mejo.2008.11.064