Optical Materials 76 (2018) 287e294

Contents lists available at ScienceDirect

Optical Materials
journal homepage: www.elsevier.com/locate/optmat

Hollow-core silver coated photonic crystal fiber plasmonic sensor

Moriom Rojy Momota, Md. Rabiul Hasan*
Rajshahi University of Engineering & Technology, Department of Electronics & Telecommunication Engineering, Rajshahi, 6204, Bangladesh

a r t i c l e i n f o a b s t r a c t

Article history: We propose a simple hollow-core circular lattice photonic crystal fiber (PCF) based surface plasmon
Received 23 October 2017 resonance (SPR) refractive index sensor. The sensing performance is investigated by using the finite
Received in revised form element method (FEM). Silver is used as the plasmonic material for this design, which is placed on the
23 December 2017
outer surface of the PCF to facilitate the fabrication. The proposed sensor shows a maximum wavelength
Accepted 25 December 2017
sensitivity of 4200 nm/RIU with a sensor resolution of 2.38
105 RIU. Besides, a maximum amplitude
sensitivity of 300 RIU1 and a resolution of 3.33
105 RIU is reported for an analyte refractive index of
1.37. Moreover, the effect of varying structural parameters on the sensing performance such as pitch, air
Keywords:
Surface plasmon resonance
hole diameter and silver layer thickness are also discussed thoroughly. Sensitivity analysis of the pro-
Photonic crystal fiber posed sensor is performed in order to investigate the impact on loss depth and amplitude sensitivity.
Refractive index sensor Thanks to high sensitivity and linearity characteristics, the proposed sensor can be potentially employed
Sensitivity in practical bio-sensing and chemical sensing applications.
Resolution © 2017 Published by Elsevier B.V.

1. Introduction
Over the last three decades, surface plasmon resonance (SPR)
phenomenon has been highlighted as a powerful optical detection
technique due to its highly sensitive behavior, and capability for
label-free sensing [1,2]. SPR sensors have revealed impressive
progress in bio-sensing applications including biological analytes
detection, medical diagnosis, bio-imagining, antibody-antigen
interaction, organic chemical sensing, water testing, and environ-
mental safety [3e7]. Due to ongoing development of SPR technol-
ogy, it has been also employed in optoelectronic devices such as
film thickness monitoring [8,9], optical tunable filters [10,11], SPR
imaging [12,13], and optical modulators [14,15]. Additionally, SPR
technology permits integration of nanoelectronic and nano-
photonic components with the aim of obtaining ultra-compact
optoelectronic devices [16,17].
Recently, photonic crystal fiber (PCF) based SPR sensors have
been extensively studied due to their unusual and appealing optical
characteristics over the conventional optical fibers [18]. Due to
controllable birefringence and confinement loss, PCF SPR sensors
have shown an improved performance in bio-sensing applications.
For instance, by changing the structural parameters such as pitch,
* Corresponding author.
E-mail addresses: rojy.ruet12@gmail.com (M.R. Momota), ruetrabu@gmail.com
(Md.R. Hasan).
air hole dimension, and number of air hole rings, it is possible to
operate the PCF sensor at the optimum condition. In contrast to
conventional optical sensors, PCF-based SPR sensors permit better
control of the evanescent field by varying the structural parame-
ters. Moreover, the size of the PCF SPR sensors is small, making
them suitable for nanosensors and remote sensing applications
[19].
Over the last few years, several PCF SPR sensors have been
explored in order to improve the sensitivity, detection accuracy and
dynamic range of detection. PCF SPR sensors are classified into two
categories. First, internally metal-coated PCF sensors where plas-
monic material is selectively deposited inside the PCF and liquid
analyte is selectively infiltrated into the air hole(s). Second, exter-
nally metal-coated PCF sensors where plasmonic materials are
deposited on the outer surface of the PCF. Very recently, a diamond
ring fiber (DRF) based SPR sensor has been proposed [20] showing
maximum wavelength sensitivity of 6000 nm/RIU and sensing
resolution of 1.67
105 RIU in the sensing range between 1.33 and
1.39. The open-ring channel-based sensor proposed by Liu et al.
[21] can be used to detect low refractive index (RI) analyte between
1.23 and 1.29. The proposed design shows average wavelength
sensitivity of 5500 nm/RIU and maximum amplitude sensitivity of
333.8 RIU1.
Based on internal metal coating, few potential PCF SPR sensors
have been reported in Refs. [22e25]. In practice, deposition of
metal on several air holes while maintaining uniform thickness is
unfeasible from the fabrication point of view. Moreover, selective

https://doi.org/10.1016/j.optmat.2017.12.049
0925-3467/© 2017 Published by Elsevier B.V.
288 M.R. Momota, Md.R. Hasan / Optical Materials 76 (2018) 287e294
infiltration of air holes with analyte is complex and time
consuming. It potentially limits widespread applications in prac-
tical sensing applications. Recently, nanowire-based PCF SPR sen-
sors have been reported in order to eliminate the issues of selective
infiltration [26,27]. Very recently, a hollow-core silver nanowire-
based PCF SPR sensor [28] has been proposed having large detec-
tion range from 1.33 to 1.50. The proposed sensor shows maximum
wavelength sensitivity of 1800 nm/RIU when analyte index is var-
ied from 1.33 to 1.34. Recently, externally-metal-coated PCF SPR
sensors have gained increased attention because of their straight-
forward detection and ease of fabrication. The PCF based SPR sensor
proposed by Hassani et al. was used for the detection of biolayer
thickness [29]. Plasmonic gold and biolayer were deposited outside
the PCF surface. The proposed bimetallic SPR sensor shows
maximum wavelength sensitivity of 4750 nm/RIU and 4300 nm/
RIU for quasi TM and TM fundamental core modes, respectively.
Recently, a multichannel PCF-based sensor has been proposed by
Hameed et al. [30], where plasmonic silver is coated by the gold.
The proposed bimetallic SPR sensor shows maximum wavelength
sensitivity of 4750 nm/RIU and 4300 nm/RIU for quasi TM and TM
fundamental core modes, respectively. The PCF SPR sensor pro-
posed by Otupiri et al. consists of elliptical air holes and four
microfluidic channels [31], which can detect multi-analytes having
different refractive indices. It is reported that maximum wave-
length sensitivity of 4600 nm/RIU and amplitude sensitivity of 425
RIU1 can be achieved for a gold layer thickness of 50 nm. A single-
ring hexagonal PCF SPR sensor with four microfluidic channels has
been demonstrated, where four modes are separately studied for
multi-analyte detections [32]. The proposed sensor shows
maximum wavelength sensitivity of 2400 nm/RIU in y-polarized
HE11 mode. Although multichannel SPR sensors are useful for
multi-analyte detections, obtaining such microfluidic channels is
difficult from a fabrication point of view.
In D-shaped PCF sensors, the top of the cladding is etched out in
order to obtain a flat surface. The plasmonic material and analyte
sample are both placed on that flat surface. Most of the internally
and externally metal-coated PCF sensors are based on the metal
coating on the circular surface. In general, obtaining a uniform
circular surface is difficult since it associates with unwanted surface
roughness. Taking advantage from the flat surface, D-shaped PCF
sensors provide possibilities for homogeneous coating with mini-
mal surface roughness [33]. Very recently, a D-shaped PCF sensor
has been proposed using gold as the plasmonic material and tita-
nium di-oxide (TiO2) for shifting the resonance wavelength from
visible to near infrared [34]. In the sensing range between 1.33 and
1.43, the D-shaped sensor shows maximum amplitude sensitivity
of 1086 RIU-1 with sensing resolution of 9.2
106 RIU. Although
D-shaped PCF sensors provide high sensitivity and eliminate the
issue of uniform thickness, they require accurate polishing of the
predefined section. Very recently, a quasi D-shaped PCF SPR sensor
has been proposed by An et al. [35] using graphene and indium tin
oxide (ITO) layers operating in the near-infrared wavelengths.
Another D-shaped PCF sensor has been reported having rectangular
lattice air-holes [36]. Two larger air holes are placed near the core of
the PCF to create birefringence, which yields a higher figure-of-
merit (FOM) of about 478.3 RIU1.
In this paper, a hollow-core PCF SPR sensor with circular lattice
air holes is proposed. A thin layer of silver is used in the outer
surface of the PCF structure, which simplifies the detection mech-
anism. The performance of the sensor is investigated in terms of
amplitude sensitivity, wavelength sensitivity, sensor resolution,
and linearity. Compared to other designs in the literature, the
proposed sensor exhibits lower confinement loss. Due to such low
confinement loss, the length of the sensor can be extended to
centimeter scale. We also discuss the fabrication tolerance in order
to investigate how sensing performance varies due to fluctuations
of the fiber design parameters.
2. Structural design and plasmonic materials
Cross-section view of the proposed circular lattice PCF sensor is
shown in Fig. 1. There are two air-hole rings with four missing air
holes in the second ring. The size of the air holes in the second ring
is larger than first ring for obtaining better light confinement in the
desired direction [37]. In the first ring, air holes are arranged at 60
anticlockwise progressive rotations and in second ring air holes are
arranged at 30 anticlockwise rotations. In the second ring, four air
holes are missing, which accumulates the evanescent electromag-
netic field at the four sides of the PCF, and it can easily move to-
wards the metal surface. Here, L is the center-to-center distance
between two adjacent air holes, d is the diameter of air holes in the
first ring and d1 is the diameter of air holes in the second ring. The
thickness of silver layer is ts. Outside of silver layer an analyte layer
is placed, which acts as the dielectric medium (sensing medium). In
the proposed sensor, we used fused silica as the background ma-
terial. The RI of fused silica can be obtained by using the Sellmeier
equation [38]. The complex RI of silver can be obtained from the
Palik [39]. Finite element method (FEM) with circularly perfectly
matched layer (PML) boundary condition has been used to simulate
the proposed structure. We used finer element meshing, which
took 64,475 elements to represent the full structure. The PML
thickness and analyte layer thickness were 2.5 mm and 3.5 mm,
respectively. The RI of analyte (na) is varied from 1.33 to 1.37.
The sensing performance of a SPR sensor greatly depends on the
plasmonic material employed in the design. Current plasmonic
sensors mostly use gold because of several advantageous proper-
ties. Gold is chemically inert, long-term stable, and easy to struc-
ture [19]. However, it has higher optical damping and has wider
resonance wavelength peak leading to false positive analyte de-
tections [33]. Copper is another potential plasmonic material hav-
ing almost the same optical damping and interband transition as
gold. Compared to gold and copper, aluminum has not attracted
much attention due to its high optical damping, and high interband
transition loss [19]. From the optical point of view, silver can be
regarded as one of the potential candidates for plasmonic material.
The positive attributes of silver are no interband transition in the
visible wavelength spectrum, low optical damping, narrow
Fig. 1. Cross-section view of the proposed circular lattice PCF sensor with L ¼ 2 mm,
d ¼ 0.57L, d1 ¼ 0.78L and ts ¼ 30 nm.
 M.R. Momota, Md.R. Hasan / Optical Materials 76 (2018) 287e294 289

resonance peak, and a plasma wavelength deep in the ultraviolet

wavelength range [40]. However, in the presence of aqueous so-
lution it creates brittle oxide layers. Such formation of oxide layers
can be strongly prevented by depositing a bimetallic layer on the
top of the silver [41,42].

3. Results and performance analysis

The operating principle of PCF-based SPR sensor is based on the

interaction of evanescent field with the surface electron emitted
from the plasmonic metal. Light wave propagated in the core of the
PCF generates an evanescent field. When the evanescent field
strikes the metallic surface, it produces a surface plasmon wave
(SPW). The SPW propagates along the metal-dielectric surface. If
the frequency of the core-guided fundamental mode and surface
plasmon polariton (SPP) mode is matched, a sharp confinement
loss peak will occur [43]. Any small change of the liquid analyte
causes a significant change of the loss characteristics. An unknown
analyte can be detected by using either the shift of resonance peak
or loss variations. Geometrical parameters of the PCF are highly
responsible for the performance of the SPR sensor. For an easy
interaction between metal surface and evanescent field, design
parameters should be selected carefully. It is possible to increase
the sensitivity by creating strong coupling between SPP mode and
core-guided mode. In order to obtain enhanced sensing perfor-
mance, we used the following design parameters L ¼ 2 mm,
d ¼ 0.57L, d1 ¼ 0.78 L and ts ¼ 30 nm. These design values are kept
constant in the entire simulation unless otherwise stated.
Optical field distribution of the x-polarized core-guided mode,
y-polarized core-guided mode, x-polarized SPP mode and y-
polarized SPP mode are shown in Fig. 2(a), (b), (c) and (d),
respectively. Due to four missing air holes in the second ring,
evanescent field can easily penetrate towards the cladding region. Fig. 2. Optical field distribution of (a) x-polarized fundamental core mode, (b) y-
polarized fundamental core mode, (c) x-polarized SPP mode, (d) y-polarized SPP mode
The effective refractive index (neff) of core-guided x-polarized
at 0.648 mm, and (e) dispersion relation between SPP mode and y-polarized core-
mode, core-guided y-polarized mode, loss spectra of core-guided x- guided mode at na ¼ 1.36, and ts ¼ 30 nm.
polarized mode and core-guided y-polarized mode are shown in
Fig. 2(e) with na ¼ 1.36, and ts ¼ 30 nm. In the same figure, solid
blue line indicates the neff of SPP mode. It is showed that at a shifts from 0.562 to 0.586 mm, 0.586e0.614 mm, 0.614e0.648 mm
wavelength of 0.648 mm, neff of SPP mode intersect with the neff of and 0.648e0.69 mm, respectively. Besides, a monotonic increase of
core-guided mode. This condition is known as phase matching loss depth is noticed by increasing the analyte RI. At analyte RI of
condition, and 0.648 mm is the resonance wavelength. At the 1.33, the lowest confinement loss peak of 1.398 dB/cm is observed
resonance wavelength, a sharp peak is observed, which indicates at 0.562 mm. On the other hand, at analyte RI of 1.37, the highest
maximum energy transfer from the core-guided mode to the SPP confinement loss peak of 3.061 dB/cm is observed at 0.698 mm.
mode. As illustrated in Fig. 2(e), y-polarized core-guided mode The performance of a sensor can be measured with the wave-
shows higher loss depth than that obtained in x-polarized mode. In length interrogation (spectral-based measurement) method and
general, higher loss depth results in stronger interaction of the amplitude interrogation (intensity-based measurement) method.
evanescent field with the SPP mode. As a result, sensing perfor- In wavelength interrogation method, sensitivity can be obtained
mance increases significantly. Since x-polarized mode shows lower from the following expression [31],
confinement loss while having the same resonance wavelength, we
have considered y-polarized mode in the rest of the analysis.
Confinement loss is an important factor for measuring sensor .
performance, which is given by the following expression [31,38], Sl ðnm=RIUÞ ¼ Dlpeak Dna (2)


aðdB=cmÞ ¼ 8:686
k0 Im neff
104 (1) where Dlpeak is the peak wavelength shift due to any change of
analyte RI and Dna is the change of the two successive analyte RIs.
where k0 is the wave number, and Im(neff) is the imaginary part of When analyte RI is varied from 1.33 to 1.34, 1.34 to 1.35, 1.35 to 1.36
neff . The real part of neff of the SPP mode is highly sensitive to the RI and 1.36 to 1.37, the corresponding wavelength peak wavelength
of sensing medium. Due to small variation of analyte RI, the neff of shifts are 24, 28, 34 and 42 nm, respectively. Using Dna of 0.01,
SPP mode will change significantly. As a result, the phase matching theoretical wavelength sensitivity are 2400, 2800, 3400 and 4200
point shifts towards either shorter (blue shift) or longer wavelength nm/RIU. Moreover, average wavelength sensitivity of 3200 nm/RIU
(red shift). The loss spectra due to the change of analyte RI from 1.33 can be obtained in the RI range between 1.33 and 1.37. Sensor
to 1.37 are shown in Fig. 3(a). It is found that increasing analyte RI resolution is another important parameter, which indicates the
results in a red shift. When analyte RI is varied from 1.33 to 1.34, smallest change of RI that can be detected by a sensor. The sensor
1.34 to 1.35, 1.35 to 1.36 and 1.36 to 1.37, resonance wavelength resolution can be found from the following equation [38],
290 M.R. Momota, Md.R. Hasan / Optical Materials 76 (2018) 287e294
Fig. 3. (a) Loss spectra with increasing analyte RI from 1.33 to 1.37, and (b) amplitude sensitivity with increasing analyte RI from 1.33 to 1.37 with L ¼ 2 mm, d ¼ 0.57L, d1 ¼ 0.78L
and ts ¼ 30 nm.
.
RðRIUÞ ¼ Dna
Dlmin Dlpeak
where Dlmin is the minimum spectral resolution and Dlpeak is the
resonant wavelength shift. Assuming that Dlmin ¼ 0.1 nm,
Dna ¼ 0.01, and Dlpeak ¼ 42 nm, the calculated maximum sensor
resolution is 2.38
105 RIU. It should be mentioned that during
the calculation of sensor resolution practical noise influence has not
been considered.
In terms of practical implementation, amplitude interrogation
method is simpler and cost effective because it does not associate
with spectral manipulation. Moreover, it can be measured at a
specific wavelength. Amplitude sensitivity of a sensor can be
calculated by using the following formula [34],

 1 vaðl; na Þ
SA RIU1 ¼ 
aðl; na Þ vna
where a(l, na) is the confinement loss at refractive index of na and
va(l, na) is the confinement loss difference due to two adjacent
analyte RIs. Amplitude sensitivity of the proposed sensor by vary-
ing analyte RI from 1.33 to 1.37 is shown in Fig. 3(b). With analyte RI
of 1.33, 1.34, 1.35, 1.36 and 1.37 we found theoretical amplitude
sensitivity of 152, 180, 215, 260 and 300 RIU1 at 0.594 mm,
0.622 mm, 0.655 mm, 0.698 mm and 0.758 mm respectively. The
proposed sensor shows a maximum amplitude resolution of
3.33
105 RIU assuming that a minimum transmitted intensity of
1% can be detected by the sensor.
Silver layer thickness (ts) has potential impact on the sensitivity
and the loss depth. Fig. 4(a) shows loss spectra for increasing ts from
30 to 50 nm at analyte RI of 1.35 and 1.36. According to the figure,
loss depth monotonically decreases when ts is increased from 30 to
50 nm. The reason is the higher damping loss for thicker silver
layers. Moreover, blue shift is observed when ts is increased from 30
to 50 nm. Amplitude sensitivity for ts of 30, 40 and 50 nm at analyte
RI of 1.35 is shown in Fig. 4(b). It can be seen from the figure that the
highest amplitude sensitivity occurs for ts equal to 30 nm. The
reason can be understood from the confinement loss curve of
Fig. 4(a). Higher loss depth leads to improved confinement of the
guided mode, which enhances the field penetration through the
cladding region. As a result, interaction with analyte increases and
higher exchange of energy occurs from the fundamental mode to
the SPP mode.
The effects of changing pitch (L) values on the performance of
the proposed sensor are shown in Fig. 5. Variation of loss depth for
increasing L from 1.8 to 2.1 mm with analyte RI of 1.35 and 1.36 is
depicted in Fig. 5(a). According to the figure, loss depth decreases
gradually with the increasing of L values. The loss depth decreases
from 5.833 dB/cm to 1.639 dB/cm by increasing L from 1.8 to
2.1 mm at analyte RI of 1.36. The same behavior can be observed for
(3)
analyte RI of 1.35. Moreover, increasing L values results in a
negligible blue shift. Amplitude sensitivity for different L values at
analyte RI of 1.35 is shown in Fig. 5(b). It can be noted that
amplitude sensitivity can be increased by using smaller L values.
Amplitude sensitivity decreases from 248 to 200 RIU1 when L is
increased from 1.8 to 2.1 mm at analyte RI of 1.35. Considering the
length of the sensor and the fabrication feasibility, we have kept L
of 2 mm in the entire simulations.
Fig. 6 represents the performance of the proposed sensor using
gold as the plasmonic material with identical pitch, plasmonic
material thickness and air hole diameters. The aim of this analysis is
showing how loss depth and amplitude sensitivity vary by
employing a different plasmonic material (gold). Loss spectra for
increasing analyte RI from 1.33 to 1.37 with gold layer thickness tg
(4) of 30 nm is shown in Fig. 6(a). The calculated loss depths are 2.055,
2.472, 2.982, 3.71 and 4.672 dB/cm for analyte RI of 1.33, 1.34, 1.35,
1.36 and 1.37, respectively. Fig. 6(b) shows amplitude sensitivity for
different analyte RI with tg ¼ 30 nm. We found amplitude sensi-
tivity of 66, 85, 112, 152 and 200 RIU1 for analyte RI of 1.33, 1.34,
1.35, 1.36 and 1.37 respectively. Using plasmonic gold, a maximum
amplitude sensitivity of 200 RIU1 can be obtained, which can be
improved to 300 RIU1 using silver. Therefore, the proposed sensor
exhibits better performance with silver as the plasmonic material.
High linearity is the indication of a good sensor. Fig. 7 shows
linear fitting of the resonance wavelengths as a function of analyte
RI from 1.33 to 1.37. Linear fitting regression equation is Y ¼ 3.18X e
3.673, where X is the analyte refractive index and Y is the resonance
wavelength. The linear fitting curve shows R2 value of 0.9875,
which provides a good linearity.
In practice, it is not possible to fabricate PCF with exact design
parameters as mentioned for this sensor. Usually ±2% to ±5% var-
iations of the design parameters from their optimum values can
occur during fabrication [44]. To ensure the fabrication accuracy,
we have investigated up to ±10% variations of the design parame-
ters to with respect to their optimum values. Fig. 8(a) shows effect
on loss spectra due to ±5% and ±10% variations of ts from the op-
timum value of 30 nm at analyte RI 1.35 and 1.36. According to the
figure, confinement loss varies about ±0.45 dB/cm and ±1 dB/cm
due to ±5% and ±10% variations from 30 nm at analyte RI of 1.35.
Note that, despite those variations, resonance wavelength remains
at the same position. Fig. 8(b) represents variations of amplitude
sensitivity due to ±5% and ±10% variations of ts. It can be observed
that due to 10% change of ts amplitude sensitivity increases up to
244 RIU1 from the optimum value of 215 RIU1 at analyte RI of
1.35. ±5% and ±10% variations of L from the optimum value of
2 mm at analyte RI of 1.35 and 1.36 is depicted in Fig. 8(c). In this
 M.R. Momota, Md.R. Hasan / Optical Materials 76 (2018) 287e294 291

Fig. 4. (a) Loss spectra with increasing silver layer thickness from 30 to 50 nm at analyte RI of 1.35 and 1.36, and (b) amplitude sensitivity for silver layer thickness of 30, 40 and
50 nm with na ¼ 1.35, L ¼ 2 mm, d ¼ 0.57L, d1 ¼ 0.78L and ts ¼ 30 nm.

Fig. 5. (a) Loss spectra with increasing pitch from 1.8 to 2.1 mm at analyte RI of 1.35 and 1.36, and (b) amplitude sensitivity for different pitch values with na ¼ 1.35, d ¼ 0.57L,
d1 ¼ 0.78L and ts ¼ 30 nm.

Fig. 6. Performance comparison using gold as the plasmonic material (a) Loss spectra with increasing analyte RI from 1.33 to 1.37 and (b) amplitude sensitivity with increasing
analyte RI from 1.33 to 1.37 with L ¼ 2 mm, d ¼ 0.57L, d1 ¼ 0.78L and tg ¼ 30 nm.

case, confinement loss peak varies about ±0.47 dB/cm and ±1 dB/
cm due to ±5% and ±10% variations for analyte RI of 1.35. Unlike
variations of ts, small blue shift occurs due to 5% and 10% increase of
L and small red shift occurs due to 5% and 10% decrease of L.
Fig. 8(d) shows the effect on the amplitude sensitivity due to ±5%
and ±10% variations of L. Amplitude sensitivity does not change
significantly with respect to variations of L. 10% change of L re-
sults in amplitude sensitivity of 228 RIU1 whereas the optimum
value is 215 RIU1.
Fig. 9(a) shows the effect of varying d on the confinement loss by
±5% and ±10% from the optimum value. Due to ±5% and ±10%
variations of d, confinement loss varies about ±0.41 dB/cm and
±0.8 dB/cm for analyte RI of 1.35. Moreover, it is apparent that small
blue shifting occurs due to 5% and 10% decrease of d and small red
shifting occurs due to 5% and 10% increase of d. Dependence of
amplitude sensitivity due to the variation of d by ±5% and ±10%
from the optimum value is shown in Fig. 9(b). According to the
figure, there is a negligible fluctuation of amplitude sensitivity due
to the variation of d. Fig. 9(c) shows the impact on confinement loss
due to the ±5% and ±10% change of d1 from the optimum value. As
illustrated in the figure, there is a noticeable change of loss depth
when d1 is varied. Due to 10% decrease of d1 from the optimum
value, confinement loss is increased from 2 dB/cm to 4.73 dB/cm.
We found a red shift due to this variation. 5% and 10% increases of d1
result in a blue shift. It can be seen from Fig. 9(d) that amplitude
sensitivity is almost independent to ±5% and ±10% variations of d1
from the optimum value.
The distance between second ring and silver coating layer of the
proposed hollow-core PCF sensor plays a significant role in the
manufacturing process. Initially we kept the distance between
second ring and silver coating fixed to 50 nm. In practice, it is
difficult to maintain such small distance between second ring and
silver coating layer. Therefore, we increased this distance to 100,
150, and 200 nm to investigate the effect on sensing performance.
292 M.R. Momota, Md.R. Hasan / Optical Materials 76 (2018) 287e294
Fig. 7. Linear fitting of resonance wavelength with analyte RI from 1.33 to 1.37.
The effects due to increasing this distance on the loss depth and
amplitude sensitivity are shown in Fig. 10. According to Fig. 10(a),
loss depth increases by increasing the distance between second
ring and silver coating layer. The minimum and maximum loss
depth of about 2.435 and 7.74 dB/cm can be obtained when the
distance between second ring and silver coating layer is increased
from 50 to 200 nm for an analyte RI of 1.36. The distance between
second ring and silver coating layer can be further increased, which
results in a significant increase of confinement loss. As a result, it
will be difficult to obtain a measurable signal at the output end to
detect the sample. To maintain the confinement loss lower than
7 dB/cm, the distance between second ring and silver coating layer
should not be longer than 200 nm. It can be observed from
Fig. 10(b) that amplitude sensitivity is reduced from 215 to 203
RIU1 when the distance between second ring and silver coating
layer is increased from 50 to 200 nm for an analyte RI of 1.35. Due to
such low variation of amplitude sensitivity, the distance between
Fig. 8. Investigation of fabrication tolerance on the performance of the sensor (a) Effect on loss spectra due to maximum ±10% variation of ts from the optimum value at analyte RI
1.35 and 1.36, (b) effect on amplitude sensitivity due to maximum ±10% variation of ts from the optimum value at analyte RI 1.35, (c) effect on loss spectra due to maximum ±10%
variation of L from the optimum value at analyte RI 1.35 and 1.36, (d) effect on amplitude sensitivity due to maximum ±10% variation of L from the optimum value at analyte RI
1.35.
second ring and silver coating layer can be enlarged to facilitate the
fabrication.
The proposed PCF consists of circular lattice air holes. The cir-
cular pattern can be realized by employing stack-and-draw, die-
cast and capillary stacking methods [45]. Note that the diameter of
standard PCF is 125 mm; therefore, an extra polishing would be
required to etch out the uncladded silica surface formed outside the
cladding air holes [45]. Moreover, the proposed PCF can be obtained
by using the tapered technique [46]. Due to surface roughness of
the externally-deposited silver nanofilm, it is difficult to obtain a
very smooth surface. The thin silver can be deposited by using
thermal evaporation, sputtering technique, and wet-chemistry
technique [33]. However, minimal surface roughness can be ach-
ieved by employing chemical vapor deposition (CVD) technique
[47]. Due to corrosive nature of the silver and oxidation issue, the
proposed sensor may not be suitable for long-term stability.
However, using graphene on the top of the silver can eliminate
those issues [48,49]. Fused silica has ultra low thermal sensitivity,
which is used as the background material. The RI change of fused
silica with temperature is only 1.28
105/ C. Due to such low RI
change, the effect of temperature variation can be neglected when
the sensor is used in common environments [50]. For instance, the
self-referencing method can be used to eliminate the impact of
instrumental noise and temperature fluctuations [31]. In practical
applications such as in remote sensing, the length of the sensor
should be long. Generally, the sensor length is restricted to milli-
meter scale due to high propagation loss [31,34]. The proposed
sensor shows comparatively lower confinement loss, which
potentially allows extending the sensor length. Additionally, using
three dimensional translation stages and special designed fiber
holder, the sensor length can be further extended to centimeter
scale [51].
Table 1 compares the performance of the existing PCF SPR
sensors with the proposed hollow-core sensor. It can be seen from
the table that the proposed PCF SPR sensor shows improved results
in terms of amplitude sensitivity, wavelength sensitivity and sensor
resolution.
 M.R. Momota, Md.R. Hasan / Optical Materials 76 (2018) 287e294 293

Fig. 9. Investigation of fabrication tolerance on the performance of the sensor (a) Effect on loss spectra due to maximum ±10% variation of d from the optimum value at analyte RI
1.35 and 1.36, (b) effect on amplitude sensitivity due to maximum ±10% variation of d from the optimum value at analyte RI 1.35, (c) effect on loss spectra due to maximum ±10%
variation of d1 from the optimum value at analyte RI 1.35 and 1.36, (d) effect on amplitude sensitivity due to maximum ±10% variation of d1 from the optimum value at analyte RI
1.35.

Fig. 10. Effect on (a) loss depth and (b) amplitude sensitivity due to the variation of distance between second ring and silver layer.

Table 1
Comparison with the existing SPR PCF sensors in the literature.

Ref. Wavelength range (nm) RI Range Wavelength sensitivity (nm/RIU) Amplitude sensitivity (RIU1) Wavelength resolution (RIU) Amplitude resolution (RIU)

[52] 400e1400 1.33 N/A 203 4.5
105 4.9
105

e1.335
[53] 450e750 1.33e1.35 2520 72.47 3.97
105 N/A
[54] 520e820 1.33e1.37 2000 140 5
105 7.1
105
[55] 590e800 1.33e1.37 1000 118 1
104 8.5
105
[25] 1275e1690 1.33e1.35 2000 N/A 5
105 N/A
[56] 480e650 1.33e1.37 3700 216 2.7
105 4.6
105
[57] 550e750 1.33e1.34 2900 120 N/A N/A
[45] 580e720 1.33e1.36 2200 266 4.55
105 3.75
105
This work 500e800 1.33e1.37 4200 300 2.38
105 3.33
105

4. Conclusions analytes. The plasmonic material and sensing layer are placed
outside the PCF surface, which facilitate both fabrication and ana-
A hollow-core circular lattice PCF-based SPR refractive index lyte detection. The proposed sensor shows a maximum wavelength
sensor is proposed for the detection of biological and biochemical sensitivity of 4200 nm/RIU with a wavelength resolution of
294 M.R. Momota, Md.R. Hasan / Optical Materials 76 (2018) 287e294
2.38
105 RIU. The maximum amplitude sensitivity of 300 RIU1
with a high wavelength resolution of 3.33
105 RIU is also ob-
tained for an analyte refractive index of 1.37. The fabrication
tolerance is also analyzed with respect to maximum ±10% varia-
tions of the structural parameters such as silver later thickness,
pitch, and air hole diameter. The proposed sensor can be potentially
employed for the detection of unknown analytes and in lab-on-
fiber technology.
References
[1] E. Helmerhorst, D.J. Chandler, M. Nussio, et al., Real-time and label-free bio-
sensing of molecular interactions by surface plasmon resonance: a laboratory
medicine perspective, Clin. Biochem. Rev. 33 (4) (2012) 161.
[2] H.H. Nguyen, J. Park, S. Kang, et al., Surface plasmon resonance: a versatile
technique for biosensor applications, Sensors 15 (5) (2015) 10481e10510.
[3] C.E. Berger, J. Greve, Differential SPR immunosensing, Sensor. Actuator. B
Chem. 63 (1) (2000) 103e108.
[4] J. Homola, Surface plasmon resonance sensors for detection of chemical and
biological species, Chem. Rev. 108 (2) (2008) 462e493.
[5] V. Koubova , E. Brynda, L. Karasova , et al., Detection of foodborne pathogens
using surface plasmon resonance biosensors, Sensor. Actuator. B Chem. 74
(1e3) (2001) 100e105.
[6] J. Homola, J. Dosta lek, S. Chen, et al., Spectral surface plasmon resonance
biosensor for detection of staphylococcal enterotoxin B in milk, Int. J. Food
Microbiol. 75 (1e2) (2002) 61e69.
[7] Y. Fang, Label-free cell-based assays with optical biosensors in drug discovery,
Assay Drug Dev. Technol. 4 (5) (2006) 583e595.
[8] K.S. Johnston, S.R. Karlsen, C.C. Jung, et al., New analytical technique for
characterization of thin films using surface plasmon resonance, Mater. Chem.
Phys. 42 (4) (1995) 242e246.
[9] T. Akimoto, S. Sasaki, K. Ikebukuro, et al., Refractive-index and thickness
sensitivity in surface plasmon resonance spectroscopy, Appl. Optic. 38 (19)
(1999) 4058e4064.
[10] P.J. Kajenski, Tunable optical filter using long-range surface plasmons, Opt.
Eng. 36 (5) (1997) 1537e1541.
[11] Y. Wang, Voltage-induced color-selective absorption with surface plasmons,
Appl. Phys. Lett. 67 (19) (1995) 2759e2761.
[12] Y.-D. Su, S.-J. Chen, T.-L. Yeh, Common-path phase-shift interferometry sur-
face plasmon resonance imaging system, Optic Lett. 30 (12) (2005)
1488e1490.
[13] L. Wang, R.J.H. Ng, S. Safari Dinachali, et al., Large area plasmonic color pal-
ettes with expanded gamut using colloidal self-assembly, ACS Photon. 3 (4)
(2016) 627e633.
[14] J.S. Schildkraut, Long-range surface plasmon electrooptic modulator, Appl.
Optic. 27 (21) (1988) 4587e4590.
[15] G.T. Sincerbox, J.C. Gordon, Small fast large-aperture light modulator using
attenuated total reflection, Appl. Optic. 20 (8) (1981) 1491e1496.
[16] S.A. Maier, Plasmonics: the promise of highly integrated optical devices, IEEE J.
Sel. Top. Quant. Electron. 12 (6) (2006) 1671e1677.
[17] S.P. Burgos, H.W. Lee, E. Feigenbaum, et al., Synthesis and characterization of
plasmonic resonant guided wave networks, Nano Lett. 14 (6) (2014)
3284e3292.
[18] M.R. Hasan, M.S. Anower, M.I. Hasan, Polarization maintaining highly
nonlinear photonic crystal fiber with closely lying two zero dispersion
wavelengths, Opt. Eng. 55 (5) (2016) 056107.
[19] T. Wieduwilt, A. Tuniz, S. Linzen, et al., Ultrathin niobium nanofilms on fiber
optical tapersea new route towards low-loss hybrid plasmonic modes, Sci.
Rep. 5 (2015).
[20] W.L. Ng, A.A. Rifat, W.R. Wong, et al., A novel diamond ring fiber-based surface
plasmon resonance sensor, Plasmonics (2017) 1e6.
[21] C. Liu, L. Yang, X. Lu, et al., Mid-infrared surface plasmon resonance sensor
based on photonic crystal fibers, Optic Express 25 (13) (2017) 14227e14237.
[22] Z. Fan, S. Li, Q. Liu, et al., High sensitivity of refractive index sensor based on
analyte-filled photonic crystal fiber with surface plasmon resonance, IEEE
Photon J. 7 (3) (2015) 1e9.
[23] D. Gao, C. Guan, Y. Wen, et al., Multi-hole fiber based surface plasmon reso-
nance sensor operated at near-infrared wavelengths, Optic Commun. 313
(2014) 94e98.
[24] W. Qin, S. Li, Y. Yao, et al., Analyte-filled core self-calibration microstructured
optical fiber based plasmonic sensor for detecting high refractive index
aqueous analyte, Optic Laser. Eng. 58 (2014) 1e8.
[25] J.N. Dash, R. Jha, SPR biosensor based on polymer PCF coated with conducting
metal oxide, IEEE Photon. Technol. Lett. 26 (6) (2014) 595e598.
[26] Y. Lu, M. Wang, C. Hao, et al., Temperature sensing using photonic crystal fiber
filled with silver nanowires and liquid, IEEE Photon J. 6 (3) (2014) 1e7.
[27] Y. Lu, X. Yang, M. Wang, et al., Surface plasmon resonance sensor based on
hollow-core PCFs filled with silver nanowires, Electron. Lett. 51 (21) (2015)
1675e1677.
[28] N. Luan, J. Yao, A hollow-core photonic crystal fiber-based SPR sensor with
large detection range, IEEE Photon J. 9 (3) (2017) 1e7.
[29] A. Hassani, M. Skorobogatiy, Photonic crystal fiber-based plasmonic sensors
for the detection of biolayer thickness, J. Opt. Soc. Am. B 26 (8) (2009) 1550.
[30] M.F.O. Hameed, Y.K. Alrayk, A.A. Shaalan, et al., Design of highly sensitive
multichannel bimetallic photonic crystal fiber biosensor, J. Nanophoton. 10 (4)
(2016) 046016.
[31] R. Otupiri, E.K. Akowuah, S. Haxha, Multi-channel SPR biosensor based on PCF
for multi-analyte sensing applications, Optic Express 23 (12) (2015)
15716e15727.
[32] S.I. Azzam, M.F.O. Hameed, R.E.A. Shehata, et al., Multichannel photonic
crystal fiber surface plasmon resonance based sensor, Opt. Quant. Electron. 48
(2) (2016) 142.
[33] A.A. Rifat, R. Ahmed, A.K. Yetisen, et al., Photonic crystal fiber based plasmonic
sensors, Sensor. Actuator. B Chem. 243 (2017) 311e325.
[34] A.A. Rifat, R. Ahmed, G.A. Mahdiraji, et al., Highly sensitive d-shaped photonic
crystal fiber-based plasmonic biosensor in visible to near-IR, IEEE Sensor. J. 17
(9) (2017) 2776e2783.
[35] G. An, S. Li, H. Wang, et al., Metal oxide-graphene-based quasi-d-shaped op-
tical fiber plasmonic biosensor, IEEE Photon J. 9 (4) (2017) 1e9.
[36] L. Peng, F. Shi, G. Zhou, et al., A surface plasmon biosensor based on a D-
shaped microstructured optical fiber with rectangular lattice, IEEE Photon J. 7
(5) (2015) 1e9.
[37] A.A. Rifat, M.R. Hasan, R. Ahmed, et al., Photonic crystal fiber-based plasmonic
biosensor with external sensing approach, J. Nanophoton. 12503 (2018) 1.
[38] A.A. Rifat, G. Mahdiraji, Y.M. Sua, et al., Highly sensitive multi-core flat fiber
surface plasmon resonance refractive index sensor, Optic Express 24 (3)
(2016) 2485e2495.
[39] E.D. Palik, Handbook of Optical Constants of Solids, Academic press, 1998.
[40] P.B. Johnson, R.-W. Christy, Optical constants of the noble metals, Phys. Rev. B
6 (12) (1972) 4370.
[41] S.A. Zynio, A.V. Samoylov, E.R. Surovtseva, et al., Bimetallic layers increase
sensitivity of affinity sensors based on surface plasmon resonance, Sensors 2
(2) (2002) 62e70.
[42] A.A. Rifat, G.A. Mahdiraji, D.M. Chow, et al., Photonic crystal fiber-based sur-
face plasmon resonance sensor with selective analyte channels and graphene-
silver deposited core, Sensors 15 (5) (2015) 11499e11510.
[43] R. Otupiri, E. Akowuah, S. Haxha, et al., A novel birefrigent photonic crystal
fiber surface plasmon resonance biosensor, IEEE Photon. J. 6 (4) (2014) 1e11.
[44] M.R. Hasan, M.I. Hasan, M.S. Anower, Tellurite glass defect-core spiral pho-
tonic crystal fiber with low loss and large negative flattened dispersion over
Sþ Cþ Lþ U wavelength bands, Appl. Optic. 54 (32) (2015) 9456e9461.
[45] M.R. Hasan, S. Akter, A.A. Rifat, et al., A highly sensitive gold-coated photonic
crystal fiber biosensor based on surface plasmon resonance, Photonics 4 (1)
(2017) 18.
[46] J.K. Chandalia, B. Eggleton, R. Windeler, et al., Adiabatic coupling in tapered
air-silica microstructured optical fiber, IEEE Photon. Technol. Lett. 13 (1)
(2001) 52e54.
[47] P.J. Sazio, A. Amezcua-Correa, C.E. Finlayson, et al., Microstructured optical
fibers as high-pressure microfluidic reactors, Science 311 (5767) (2006)
1583e1586.
[48] M. Schriver, W. Regan, W.J. Gannett, et al., Graphene as a long-term metal
oxidation barrier: worse than nothing, ACS Nano 7 (7) (2013) 5763e5768.
[49] V. Kravets, R. Jalil, Y.-J. Kim, et al., Graphene-protected copper and silver
plasmonics, Sci. Rep. 4 (2014) 5517.
[50] X. Dong, H. Tam, P. Shum, Temperature-insensitive strain sensor with
polarization-maintaining photonic crystal fiber based Sagnac interferometer,
Appl. Phys. Lett. 90 (15) (2007) 151113.
[51] B. Shuai, L. Xia, D. Liu, Coexistence of positive and negative refractive index
sensitivity in the liquid-core photonic crystal fiber based plasmonic sensor,
Optic Express 20 (23) (2012) 25858e25866.
[52] A.K. Mishra, S.K. Mishra, B.D. Gupta, SPR based fiber optic sensor for refractive
index sensing with enhanced detection accuracy and figure of merit in visible
region, Optic Commun. 344 (2015) 86e91.
[53] X. Yang, Y. Lu, B. Liu, et al., Analysis of graphene-based photonic crystal fiber
sensor using birefringence and surface plasmon resonance, Plasmonics 12 (2)
(2017) 489e496.
[54] A.A. Rifat, G.A. Mahdiraji, R. Ahmed, et al., Copper-graphene-based photonic
crystal fiber plasmonic biosensor, IEEE Photon J. 8 (1) (2016) 1e8.
[55] A. Rifat, G.A. Mahdiraji, Y. Shee, et al., A novel photonic crystal fiber biosensor
using surface plasmon resonance, Proc. Eng. 140 (2016) 1e7.
[56] J.N. Dash, R. Jha, On the performance of graphene-based D-shaped photonic
crystal fibre biosensor using surface plasmon resonance, Plasmonics 10 (5)
(2015) 1123e1131.
[57] N. Luan, R. Wang, W. Lv, et al., Surface plasmon resonance sensor based on D-
shaped microstructured optical fiber with hollow core, Optic Express 23 (7)
(2015) 8576e8582.