Investigation of Fiber Optic Sensor for Monitoring of Ammonia

Proceedings of IMECE2008
2008 ASME International Mechanical Engineering Congress and Exposition
October 31-November 6, 2008, Boston, Massachusetts, USA
IMECE2008-67940
INVESTIGATION OF FIBER OPTIC SENSOR FOR MONITORING OF AMMONIA
Sistla S Shastry, Abdeq M. Abdi and A. G. Agwu Nnanna
Purdue Calumet Water Institute
Purdue University Calumet, Hammond, IN 46323
Phone: 219 989-2071; Fax: 219 989 2898
Email: shridharsastry@gmail.com
IMECE2008-67940
1. Abstract
Detection and characterization of chemical contaminants in
water network is paramount for water quality and water
security. The current trend of monitoring the presence of
contaminants is the batch sampling technique, where sample
of water is collected and analyzed in the laboratory. While this
technique is accurate, it fails to provide immediate
information. In this work, the authors investigate the
effectiveness of utilizing a fiber optics based sensor for
detecting ammonia in water. In order for the system to sense
ammonia, a small portion of the cladding of the fiber optic
cable is stripped and replaced by a porous polymer material. A
novel procedure of etching the glass cladding is reported. The
modified cladding when interacts with ammonia causes a
change in intensity of the electromagnetic wave flowing
through the cable. The change in intensity caused by the
modified cladding is studied parametrically which will help in
forming a correlation between concentration of ammonia and
absorbance.
2. Introduction
Providing contaminant free water is of paramount importance
for a water utility service. The current trend of detection is
based on online monitoring of various water related
parameters such as conductivity, turbidity, pH, dissolved
oxygen (DO), free chlorine (FC), etc. Even measuring these
parameters help in detection of anything alien that maybe
present in water, it really gives no clue as to the specific
contaminant present in water. The authors in this paper are
trying to bridge this gap by developing a fiber optics based
ammonia sensor which is highly selective to ammonia. Effort
has been put to develop a correlation between the output of the
signal and the concentration of the contaminant that maybe
present in water. Of the chemical sensors (viz. Enzyme
sensors, Catalytic sensors, CHEMFETS, Chemiresistors,
Optical Sensors) Optical Chemical Sensors and CHEMFETs
are the most commonly used sensor technologies. In the recent
years, Fiber Optics based sensors are being preferred over
conventional l chemical sensors, such as Chemiresistors and
Transistor based sensors, because of their large dynamic
range, selectivity, sensitivity, simplicity and cost effectiveness
[1]-[4].[4]. Fiber optics based sensors can be used to respond to
external physical, chemical, biological changes [5]-[9].
Cladding Modification Technique (CMT) was adopted to
make the fiber cable. CMT involves removing the passive
cladding of the fiber cable and replacing it with a material
which is active towards ammonia. Polyaniline (PAni) is a
conducting polymer which has found many applications in
semiconductor industry due to its conducting nature.
Polyaniline shows a strong absorption in the blue and green
region of visible spectrum when exposed to Ammonia and
Hydrochloric acid (HCl), respectively which can be utilized
for detection of ammonia and HCl. The amount of change in
absorption can be utilized to find out the concentration of
ammonia or HCl. Other than the developed fiber cable,
components like a light source and spectrometer are required.
Detection of contaminant is achieved by observing the shift in
absorbance curve. . Whereas the concentration of contaminant
can be observed by the correlation curve developed between
absorbance and concentration of contaminant.
3. Theory
Based on the Refractive Index of the active cladding material
the fiber optic based sensor can either be an evanescent wave
based sensor or a leaky evanescent wave based sensor. If the
refractive index (RI) of the sensing material is less than that of
the core then the resulting sensor is an evanescent wave based
sensor and if the RI of modified cladding is more than the RI
of core then the sensor developed is a leaky evanescent wave
based sensor. Figure 1, shows the evanescent wave based
sensor with the evanescent wave. w
Fig.1 Evanescent Wave based Fiber Optic Sensor
( ncore > n mod clad
)
1
Copyright © 2008 by ASME

Evanescent wave is the property of a fiber cable wherein some
part of the electromagnetic wave flowing inside the fiber cable
leaks out of the core and is present in the cladding of the fiber
cable. Evanescent wave is the result of the conservation of
energy at the core-clad clad interface and decays exponentially
along the radial direction, the value becoming almost zero as
the wave reaches the cladding-surrounding interface. When
the normal passive glass cladding of the fiber cable is replaced
by an active sensitive sitive material, the energy associated with the
evanescent wave is absorbed by the sensing material and
hence the energy associated with the electromagnetic wave
flowing inside the core of the fiber is reduced. When the
sensor is exposed to a contaminant it interacts with the
contaminant chemically, and causes further change in output.
But this change is in a specific wavelength region which can
be identified as the signature of the contaminant. The main
drawback with this kind of design is that the sensitivity of
such a sensor is very small. The evanescent wave carries at
most 1% of the total energy and the change in that 1% is even
small, typically the change in output due to evanescent wave
based sensors is about 0.02%. On the other hand in case of
Leaky Evanescent wave based sensors, the drop in the output
is considerably large. Figure 2 shows the structure of such a
sensor.
4. Optical Property of Polyaniline
The polyanilines lines refer to a class of different polymers which
can be considered as being derived from a polymer, the base
form of which known as emeraldine base consists of alternate
oxidised and reduced states. The average oxidation state can
be varied continuously from 0 to 1. An oxidation state of 0.5 is
known as emeraldine state, the deprotanated form of which
known as emeraldine base is blue in color whereas the
protanated form known as emeraldine salt is green in color.
Emeraldine state polyaniline used in this case as the sensing
material with a RI of 2.43 in HCl and 1.94in Ammonia, as
found in [11], which are both greater than water. Using the
previous data given in the literature [10], it was estimated that
the minimum and maximum absorption coefficient of
Polyaniline to be around 0.3 um -1 when exposed to HCl and
0.6 um -1 when exposed to ammonia, respectively. This shows
that 30% and 60% of energy is absorbed by a 1
µ m
Polyaniline layer exposed to HCl and Ammonia respectively.
Emeraldine state Polyaniline polymers are usually prepared by
chemical [10]-[14][14] oxidation of aniline hydrochloride by
ammonium persulphate. Of interest for development of optical
sensor is the reversible protanation and deprotanation reaction
between the base and salt as seen in Fig.3. This change from
base to salt and back is accompanied by a shift in
λ
max
from
900nm for HCl to 780nm for ammonia as seen from Fig. 6,
where Friedman’s statistical smoothing is done for clarity of
the plot. The change in UV-VIS spectrum characteristics of
Polyaniline when exposed to ammonia and HCl is being used
for detection.
Fig. 2 Structure of Leaky Evanescent Wave based Fiber Sensor
In this case, the evanescent wave phenomenon takes place
along with multiple reflections. When the RI of modified
cladding is greater than that of the core ( n
mod clad
> ncore
) the
electromagnetic wave gets refracted into the modified
cladding but since RI of the modified cladding is more than
that of surrounding (RI of water=1.33< n
core
< n mod clad
) light
is reflected back into the modified cladding, again undergoing
refraction and reflection at the modified cladding -core
interface and this process keeps on repeating throughout the
length of the sensing region. Also the absorption coefficient of
the sensing material causes the intensity of light to decrease
after every reflection which adds to the loss due to refraction.
Figure 2 shows this process for just one light ray, when the
complete electromagnetic spectrum is taken into account the
phenomenon of multiple reflections is considerable.
Considering the above discussion it was decided that the leaky
Evanescent wave based sensor be developed for increased
sensitivity.
Absorbance, au
Fig.3 Polyaniline in Emeraldine Base and Emeraldine salt forms
2% HCl
2% Ammonia
polymer
1.0
2% Ammonia
0.8
2% HCl
0.6
0.4
polymer
400
600 800 1000
Wavelength, nm
Fig.4 Characteristic plot of Absorbance as a function of Wavelength
for Polyaniline
2
Copyright © 2008 by ASME

5. Preparation of Fiber Optic Sensor
5.1 Preparation of Fiber cable surface
A multimode silica fiber cable, with numerical aperture (NA)
0.22 and diameter 200 µ m /220 µ m /240 µ m for core,
cladding and buffer jacketing was used. Required length of the
fiber cable 2 to 5 cm is chosen to be stripped by using the wet
etching technique, where a 50% Sodium Hydroxide (NaOH)
solution is used as the etching reagent. The fiber cable is
dipped into the solution with the temperature maintained at
240 o F . The 200 µ m fiber cable with the cladding thickness
of 20 µ m is dipped into the solution for 30 minutes after
which only the core is left, hence giving an etch rate of
660nm/min. NaOH solution was used because of the relative
ease of operation compared to standard buffered HF solution.
Intensity of electromagnetic wave received at spectrometer
when the sensor is exposed to DI water, Intensity of
electromagnetic wave received at the spectrometer due to stray
light, respectively. DI water is placed in a container and
placed over a magnetic stirrer as shown in fig.5. To increase
the concentration of ammonia in the container ammonia is
added from top and thoroughly mixed using the stirrer; this
gives a uniform sample of required concentration.
Spectrometer
5.2 Polyaniline Coating
The stripped fiber cable is now introduced into the Polyaniline
solution using a dipping machine for 15 minutes. This process
gives a uniform coating of thickness less than 1 µ m . The
Polyaniline solution is made by mixing two solutions prepared
by mixing 0.777g of Aniline Hydrochloride and 0.1925g of
Ammonium Persulphate in 15ml of water as in [12]. The
Polyaniline layer thus formed is hydrophobic in nature and
hence can be used in water but gradually deteriorates and the
sensitivity of the sensor decreases. Also the polymer layer
reacts with any base or acid that maybe present in water which
makes it less selective. To increase the selectivity and
sensitivity of the polymer sensor a thin layer of Polydimethyl
Siloxane is applied on top of the polymer layer. The fiber
cable is then air dried at 50 o C
6. Experimental Setup
for 3 days.
The experimental setup shown in figure 5 consists of the fiber
optic sensor along with a light source and spectrometer both
supplied by Ocean Optics, Inc operating in the wavelength
range of 200-1100nm.The light source is a broad band type
with deuterium and halogen light bulbs which operate in the
210-400nm and 300-1500nm range respectively. Data from
the spectrometer is processed using the software Spectrasuite
provided by Ocean Optics Inc. A plot of Absorbance vs.
wavelength is plotted where Absorbance is calculated as
⎛ S
A = −
⎜
λ
log
⎝ R
λ
λ
− D
λ
− D
λ
⎞
⎟
⎠
Water Sample
Magnetic
with Ammonia
Stirrer
Fig. 5 Experimental Setup
Also the container is always kept closed to avoid any escape
of ammonia vapors. The fiber sensor is passing through the
container immersed in the sample hence its detecting ammonia
in water at various concentrations.
7. Results
Light Source
Fig. 6 shows the plot of Absorbance as a function of
wavelength for various concentrations of Ammonia. Plot A
was taken with the sensor in DI water. While preparation of
the ammonia sensor a small amount of 0.1M HCl is added to
the solutions used to prepare Polyaniline, this makes sure that
the polymer sensing region is in Emeraldine Salt form which
is green in color. This can be observed from the dip in the
absorbance curve of plot A at 530nm which belongs to the
green region of visible spectrum.
Where, A , S , R ,
λ
λ
λ
Dλ
represent Absorbance at a specific
wavelength, Intensity of electromagnetic wave received at
spectrometer when the sensor is exposed to contaminant,
3 Copyright © 2008 by ASME

Absorbance, au
1.1
0.9
0.7
A
B
C
D
E
F
G
H
I
Absorbance min. at 480nm
when exposed to 18ppm ammonia
The plots 6 and 7 shown above give a relative measurement of
absorbance compared to the initial plot in DI water.
Plot 8 shows plot of percentage change in absorbance as a
function of wavelength for ammonia. This helps in
understanding the effect of the interaction between the sensing
region and the contaminant. Percentage change in absorbance
is defined as
0.5
0.3
Absorbance min. at 540nm
Absorbance min. shifted to 516nm when
exposed to 2.25ppm ammonia
A
conta min ant
A
− A
DIwater
DIwater
400 600 800 1000
Wavelength, nm
Fig. 6 Plot of Absorbance vs. Wavelength when sensor S1 was
exposed to A: DI water; B: 2ppm; C: 4ppm; D: 6ppm; E: 8ppm; F:
10ppm; G: 12ppm; H: 14ppm; I: 16ppm solutions of ammonia.
Hence the sensor developed can be used to detect ammonia to
a concentration down to 18ppm. Another sensor with the same
configuration was made but without the 0.1M HCl solution.
This gives a uniform blue colored layer of Polyaniline in its
Emeraldine base form. Such a sensor can be used to detect any
acid that maybe present in water. In this case HCl was used as
the contaminant. Fig. 7 shows the plot between absorbance
and wavelength for various concentrations of HCl as
contaminant. Plot A is obtained when the sensor is introduced
into DI water and as concentration of HCl contaminant is
added the absorbance minimum shifts from 430nm to 530 nm
which belongs to green region of visible spectrum. The
amount of HCl required to saturate the sensor can be seen
from fig. 7 to be around 116ppm. This suggests that less
amount of ammonia is required to saturate the sensor than
HCl. It is speculated that the presence of gas permeable layer
over the Polyaniline material is the cause of this variation. The
+
gas permeable layer is partially blocking some of the H 3
O
ions from interacting with the Polyaniline material, hence
increasing the selectivity of the sensor towards Ammonia.
1.1
A
B
C
D
E
F
G
Where,
A
DIWater
and
conta ant
A
min
are absorbance values of
the 2 sensors when exposed to DI water and contaminant
solution, respectively.
Normalized Absorbance
0 .4
0 .2
0 .0
- 0 .2
- 0 .4
A
B
C
D
E
F
G
H
I
4 0 0 6 0 0 8 0 0 1 0 0 0
W a v e le n g th , n m
P e a k a t 7 6 0 n m
Fig. 8 Plot of %change in Absorbance vs. Wavelength when sensor
S1was exposed to A: DI water; B: 2ppm; C: 4ppm; D: 6ppm; E:
8ppm; F: 10ppm; G: 12ppm; H: 14ppm; I: 16ppm solutions of
ammonia.
As was observed from fig. 8, the sensor starts saturating when
exposed to 18ppm of ammonia and above and is completely
saturated at a concentration of 30ppm. This plot shows that the
maximum change of 0.15 in absorbance is obtained at the
wavelength of 760nm. A similar plot for HCl was obtained as
shown in fig. 9.
0.4 A
B
C
D
E
F
G
Absorbance
0.9
% change in absorbance
0.2
0.0
0.7
-0.2
0.5
400 600 800 1000
Wavelength
400 600 800 1000
W avelength
Fig. 7 Plot of Absorbance vs. Wavelength when sensor S2 was
exposed to A: DI water; B: 58ppm; C: 116ppm; D: 174ppm; E:
232ppm; F: 290ppm; G: 348ppm solutions of HCl.
Fig. 9 Plot of %change in Absorbance vs. Wavelength when sensor
S2 was exposed to A: 58ppm; B: 116ppm; C: 174ppm; D: 232ppm;
E: 290ppm; F: 348ppm; G: 408ppm solutions of HCl.
4 Copyright © 2008 by ASME

Here, the maximum change is observed at 610nm but
saturation is achieved much later unlike in case of ammonia
which emphasizes the assumption that the gas permeable layer
+
is blocking O ions. Figure 10 and 11 show percentage
H 3
change in absorbance as a function of concentration. It can be
seen that sensor which is exposed to ammonia saturates at a
concentration of 30ppm for ammonia and in case of HCl the
saturation is not reached even at 400ppm. Two different test
runs were performed for the ammonia sensor as shown in
Fig.11; the second test was done 24 hours later from the first
one.
Normalized Absorbance for hcl
Normalized Absorbance for Ammonia
0
-0.05
-0.1
-0.15
-0.2
-0.25
-0.3
Fig. 10 Plot of normalized change in absorbance as a function of
concentration for HCl.
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0 200 400 600
Fig. 11 Plot of normalized change in absorbance as a function of
concentration for Ammonia.
8. Conclusion
y = -0.0005x - 0.0803
R² = 0.9266
Concentration
R² = 0.9834
R² = 0.9794
610nm
Linear
(610nm)
First run
Second run
Poly. (First run)
Poly. (Second run)
0 10 20 30 40 50
Concentration (ppm)
Two different sensors are developed one each for ammonia
and HCl, respectively. The sensor was saturated by ammonia
with a concentration as small as 30ppm whereas in case of
HCl the saturation was not reached even at 400ppm suggesting
that the gas permeable layer on Polyaniline is blocking all the
ions. Ammonia and HCl could be detected to a concentration
as low as 2ppm and 58ppm, respectively. Further work needs
to be done to determine the minimum concentration that can
be detected by bending the fiber cable.
9. References
1. M. A. El-Sherif, “An Apparatus and a Method
Comprising an Optical Fiber Modulator, Coupler,
Switch, Sensor, and Distribution System,” U.S.
Patent 5 060 307, Oct. 22, 1991.
2. J. Yuan, “Development of Smart Structures Utilizing
Chromogenic Materials for Optical Fiber Sensor,”
M.S. thesis, Drexel Univ., Philadelphia,PA, 1997.
3. M. A. El-Sherif, “On-Fiber sensor and modulator,”
IEEE Trans. Instrum. Meas., vol. 38, pp. 595–598,
Apr. 1989.
4. M. A. El-Sherif and J. Yuan, “Development of a
novel class of fiber optic sensors for environmental
field measurements,” in International Conference on
Agropoles and Agro-industrial Technological Parks,
Barretos, Sao Paulo, Brazil, Nov. 15–21, 1999.
5. A. D. Kersey, “A review of recent developments in
fiber optic sensor technology,” Opt. Fiber Technol.,
vol. 2, pp. 291–317, 1996.
6. T. G. Giallorenzi, J. A. Bucaro, A. Dandrige, G. H.
Sigel, J. H. Cole Jr, and S. C. Rashleigh, “Optical
fiber sensor technology,” J. Quantum Electron., vol.
QE-18-4, pp. 626–665, 1982.
7. R. Narayanaswamy, “Current developments in
optical biochemical sensors,” Biosens. Bioelectron.,
vol. 6, pp. 467–475, 1991.
8. Xiaojing Liu and Weihong Tan, “A Fiber-Optic
Evanescent Wave DNA Biosensor Based on Novel
Molecular Beacons”, Anal. Chem. 1999, 71, 5054-
5059.
9. M. A. El-Sherif, J. Yuan, and A. G. MacDiarmid,
“Fiber optic sensors and smart fabrics,” J. Intell.
Mater. Syst. Structures, vol. II, no. 5, pp. 407–414,
May 2000.
10. Jaroslav Stejskal, Irina Sapurina, “Polyaniline: Thin
Films and Colloidal Dispersions”, Pure and Applied
Chemistry, Vol.77, No.5, pp. 815-826, 2005.
11. Jianming Yuan and Mahmoud A. El-Sherif, “Fiber-
Optic Chemical Sensor Using Polyaniline as
Modified Cladding Material”, IEEE Sensors Journal,
Vol. 3, No. 1, February 2003.
12. Zhe Jin, Yongxuan Su, Yixiang Duan, “An improved
optical pH sensor based on Polyaniline”, Sensors and
Actuators B 71 (2000) 118-122.
13. Stejskal J.; Sapurina I.; Prokes J.; Zemek J., “In-Situ
Polymerized Polyaniline films”, Synthetic Metal,
Vol.105, No.3, Sep. 1999, pp.195-202
5 Copyright © 2008 by ASME