Download Theoretical Study of Phase Interrogated Surface Plasmon

Theoretical Study of Phase Interrogated Surface
Plasmon Resonance Fiber Optic Sensors with
Metallic and Oxide Layers
H. Moayyed, I. T. Leite, L. Coelho, J. L. Santos, D. Viegas

Abstract— This work reports the theoretical investigation of
optical fiber surface plasmon resonance (SPR) sensors
incorporating an internal metallic layer (Ag) covered with an
oxide layer. This research is supported by the application of an
effective analytical model combining geometrical optics with the
transfer-matrix theory for stratified optical media. Different
oxide materials (TiO2, SiO2, and Al2O3) are considered aiming to
achieve increased/enhanced sensitivity to refractive index
variations of the external medium, particularly when addressing
phase interrogation. It is shown that the combination of a 50 nm
thickness silver inner layer with a dielectric titanium oxide layer
of a specific thickness enables high performance phase sensitivity
reading and is compatible with tailoring the sensor working
region to the third telecommunication wavelength window
around 1550 nm.
Index Terms— Surface plasmon resonance, fiber-optic sensors,
phase interrogation, metallic-oxide coupled layers.
I. INTRODUCTION
T
resonant interaction of light with free electrons in
metals originates loss bands in the light spectrum with a
central wavelength that depends on the properties of the metal
as well as of the surrounding medium, particularly its
refractive index. This dependence is the principle of optical
sensing based in plasmonics where, similarly to sensing
structures based on fiber Bragg gratings (FBG) and distributed
feedback (DFB) fiber lasers, the measurand information is
encoded on the wavelength position of a spectral signature,
with all the benefits which result from this characteristic [1].
HE
This work is part-funded by National Funds through the FCT – Fundação para
a Ciência e a Tecnologia (Portuguese Foundation for Science and
Technology) within project PTDC/FIS/119027/2010 (Plasmonics Based Fibre
Optic Sensing with Enhanced Performance). L. Coelho acknowledges the
support from FCT grant SFRH/BD/78149/2011. This work is also part-funded
by the ERDF – European Regional Development Fund through the
COMPETE Programme (operational programme for competitiveness) and by
National Funds through the FCT within project «FCOMP - 01-0124-FEDER022701».
H. Moayyed, I. T. Leite, L. Coelho and J. L. Santos are with INESC TEC
(coordinated by INESC Porto), Rua Campo Alegre 687, 4169-007, Porto,
Portugal and also with the Department of Physics and Astronomy, Faculdade
de Ciências da Universidade do Porto, Rua Campo Alegre 687, 4169-007
Porto, Portugal; D. Viegas is with INL - International Iberian Nanotechnology
Laboratory, Av. Mestre José Veiga - 4715-310 Braga, Portugal (e-mail:
[email protected];
[email protected];
[email protected];
[email protected], [email protected]).
Plasmonics based optical sensors were first developed
supported in bulk optics elements (prisms) and the measurand
information obtained by monitoring the angle of the incident
light on an interface (where a metal thin film is deposited) that
resulted into a minimum reflected light. With the introduction
of the optical fiber as a platform for plasmonic optical sensors,
the angular interrogation became inappropriate and was
replaced by spectral interrogation, either directly through the
determination of the central wavelength of the loss band, or by
monitoring the optical power at specific wavelengths on the
two sides of the spectral loss band, therefore recovering the
advantages of wavelength encoded sensor operation [2].
When dealing with Plasmonics and sensing two types of
surface plasmon resonances (SPRs) need to be considered: i)
propagating surface plasmon-polaritons (SPPs) and (ii) nonpropagating localized SPRs (LSPRs), in both cases involving
evanescently confined electromagnetic fields [3]. For the
situation of SPPs, the plasmons propagates tens to hundreds
micrometers along the metal surface with an associated
electric field that decays exponentially from the surface
(normal to the dielectric-metal interface). A variation in the
refractive index of the medium above the metal shifts the
plasmon resonance condition. For the case of LSPRs, it is
involved plasmons resonantly excited in metal nanoparticles
and around nanoholes or nanowells in thin metal films. The
spectral position and the magnitude of the LSPR depend on
the size, shape and composition of these structures, as well as
of the surrounding environment.
Independently of the consideration of one or another type of
resonance in a specific SPR sensor design, its performance has
benefited tremendously from the remarkable progresses
Plasmonic has undergone in the past 10 years. Surely, the
subject is not new, with its origin traced to the pioneering
works of Sommerfeld in 1899 [4] and Wood in 1902 [5], the
first sensing application of the SPR technique being reported
in 1983 oriented to gas detection [6]. Along all this period the
interest of the scientific community on this subject has been
diverse, but its intrinsic fundamental relevance and the
recognition of its high technological potential were always
without dispute. Eventually, the reason for such constancy is
routed to the Plasmonic core physical phenomenon, which is
an interaction process between electromagnetic radiation and
conduction electrons at metallic interfaces or in small metallic
2
nanostructures. This coupling between photons and electrons
implies the relevant length scale is no more the light
wavelength but a much smaller effective value, essentially
determined by the electron dynamics in the metal, with the
light energy carried as a package of electron oscillations.
These hybrid electromagnetic modes (surface plasmon
resonances or simply plasmons) have properties that enable
two unique characteristics: i) the concentrated optical field in
the metal surface means that if light is first coupled to a
plasmon any further interaction (for example with adjacent
molecules or with a non-linear material) will be much more
effective; ii) the concentrated optical field permits to confine
light to sub-wavelength spaces, well below the usual
diffraction limit. These remarkable characteristics turn
plasmonics so relevant, particularly in a period where the
advances of the Nanosciences and Nanotechnologies permit to
engineer material structures (metamaterials) with tailored and
enhanced Plasmonic characteristics, enabling their high
performance application in several optical domains and, in
particular, in sensing [7].
All this environment has been inducing a burst of R&D
activity in surface plasmonic resonance oriented for sensing,
looking for highly sensitive and selective detection of several
physical, chemical, and biochemical parameters [8,9]. Indeed,
ever since the introduction of various optical methods in the
excitation of the SPR at a metal-dielectric interface, it has
been widely recognized that such effect can be utilized to
achieve sensing or monitoring of various interfacial
phenomena with ultrahigh sensitivity. These include, for
example, chemical and biological sensing [10,11], filmthickness sensing [12], temperature sensing [13] and angular
measurement [14]. It has been demonstrated that the SPR
technique applied to chemical and biological sensing can
achieve resolutions down to 10-7 refractive index units (RIUs),
values not accessible to other optical sensing technologies
[15].
Up to a few years ago, the large majority of sensing SPR
structures relied in classical sensing heads based on a prism
coupling system in Kretschmann’s configuration. Even so,
soon has been recognized the obvious advantages of using the
optical fiber as the SPR sensing platform, starting by the fact
that propagation in these waveguides in most of the cases is
supported by the effect of total internal reflection, also central
in plasmonic phenomena. Therefore, it has been observed a
trend towards the consideration of the optical fiber as the
privileged platform for SPR sensing [2]. Consequently, most
of the recent developments in this field involve the
combination of the optical fiber with nanostructures whose
manufacturing became possible with the advances of the
Nanosciencies/Nanotechnologies. It is the case of SPR layouts
involving the combination of metal and oxide layers on the
optical fiber surface, enabling the tailoring of the sensor
performance, most notably its sensitivity and wavelength
operation region [16].
In this work we investigate the characteristics of optical fiber
based SPR sensors incorporating the double layer structure
proposed in [16], i.e an internal metallic (silver) layer covered
by an oxide layer, focusing now on phase interrogation. We
start in section 2 by further developing some general points on
SPR sensor interrogation, then in section 3 we move to the
identification of the general sensor layout and presentation of
the study approach based on the application of an effective
analytical model combining geometrical optics with the
transfer-matrix theory for stratified optical media, section 4
presents the main results obtained from this analysis and
section 5 delivers final remarks.
II. SPECTRAL AND PHASE INTERROGATION
NDEPENDENTLY of the specific structure of SPR sensors,
it is advantageous to consider their interrogation in the
spectral domain to benefit from wavelength encoded
operation. This can be done directly by monitoring the deep of
the SPR loss band, which requires a spectral analyzer or a
tunable laser to scan a wavelength range that contains the
extreme of the SPR signature. An indirect approach relies on
reading the optical power that propagates through the sensor in
two wavelengths, one in each side of the loss band. This can
be achieved using two lasers, or by illuminating the sensor
with broadband light and the spectral interrogation lines be
defined by two FBGs [17]. Previous calibration permits to
obtain from these two readings a parameter only dependent on
the spectral position of the SPR loss band and highly sensitive
to its displacement.
I
The resolution of a SPR sensing head depends not only on its
intrinsic sensitivity to the measurand, but also on the
performance of the interrogation technique considered. To
benefit from the wavelength encoded characteristics of these
sensors, the interrogation approach above indicated or variants
shall be adopted. However, when higher/fine resolutions are
on demand this approach can be complemented by the
consideration of an additional interrogation technique which
shows better performance. Ultimately, the most favorable one
would be phase reading, associated with the fact that light
propagates through the sensor at two distinct locations of the
SPR spectral loss band will experience different phase delays.
Therefore, phase interrogation can be achieved by illuminating
the sensor with a laser line with emission in the region of the
SPR loss band where the slope of phase variation with the
measurand induced band spectral shift has its maximum value.
Considering that only the TM polarization component can
couple to surface-plasmons and give rise to SPR, the TE wave
will accumulate a SPR immune phase delay when propagating
through the sensor structure, therefore providing a reference
level to the sensing effective TM wave. A possible
classification for this interrogation approach is based on the
way of generating these waves. When using a polarizer to
obtain, from a single optical source, the TE and TM
components which are made to interfere on the photodetector,
the process is known as polarimetric interrogation. In the case
3
where the TE and TM waves are provided by two distinct light
sources with slightly different optical frequencies and are
made to interfere resulting into a beat signal with frequency
low enough to be resolved by the photodetection and
amplification electronics, we are dealing with heterodyne
interrogation. Finally, if the reference wave propagates along a
different spatial path and brought to interfere with the TM
light on the photodetector, the approach is usually known as
interference interrogation (though all the three processes rely
on interference phenomena).
Phase interrogation has not been widely explored in the
context of SPR sensors. Surely, it is more complex to
implement, requiring more know-how on optical components,
modulation instruments and signal processing techniques, but
eventually the main reason is the circumstance that most of the
R&D community which has been working until recently in
SPR sensing had a biochemical background, therefore with a
tendency to use in their research spectroscopic equipments.
This is becoming to change since plasmonics is now moving
to fields other than sensing, such as imaging and data storage,
bringing to the subject researchers with diverse backgrounds.
A consequence of this dynamics is an improvement of the
resolution values obtained for each of the three described
phase reading approaches. Indeed, for the case of heterodyne
interrogation, a value of 2.8×10-9 RIU was reported in 2008
[18], which seems to be the highest refractive index resolution
achieved so far; when considering polarimetric interrogation,
under optimized conditions, it was obtained a resolution of
3.7×10-8 RIU [19]; for interferometric interrogation, a
resolution down to 2.2×10-7 RIU was reported in 2011 [20].
Most of this research was performed with classical sensing
heads based on a prism coupling system in a Kretschmann
configuration. The SPR induced phase change would be
greatly amplified if the light undergoes multiple attenuated
total reflections, as happens when the sensing platform is an
optical fiber. This advantage is attenuated due to the
propagation characteristics of the fiber. In the case of a multimode optical fiber, each guided mode generates its own
resonance dip at a specific wavelength, different from the
other propagating modes, resulting in a broadening of the full
SPR profile, factor that decreases the phase sensitivity (as well
as the sensitivity expectable from the application of different
interrogation approaches). On the other hand, if in the sensing
head a standard single-mode fiber is considered, its small
numerical aperture limits the incidence angles from
approximately 86.5 to 89.5 degrees (in the simplified picture
of geometrical optics), which are far from the critical angle,
meaning a reduced phase shift occurs in each reflection.
increase of the interaction area, which by itself increases the
intrinsic sensitivity of the sensor, and not on the tailoring of
the guidance mechanisms of the fiber to increase also its
intrinsic sensitivity from this side. Eventually, a compromise
needs to be found between these two mechanisms, a topic that
certainly will be the target of future R&D activity. A possible
path would be to consider for the sensing head a special fiber
with material and geometric characteristics compatible with
few modes light propagation associated to incident angles
closer to the critical angle, providing a large interaction area of
the optical field with the measurand in a region where it can
have ready access, coupled to input/output standard single
mode fibers, if necessary with a coupling transition region to
reduce the connection loss. Desirably, the sensor shall operate
in reflection not only because in such configuration the
illumination fiber is additionally the return fiber, but also in
view of the light double-pass in the sensing head, which
further increases its intrinsic sensitivity.
Another approach compatible with actual fiber-optic
technology is based on the combination of long period
gratings (LPG) and plasmonic resonance. The LPG couples
light that propagates in the fiber core to a cladding mode with
a certain oscillation angle, typical of multimode operation,
which interacts in each reflection at the cladding-metal
interface in a condition of plasmonic resonance. An effective
configuration consists in considering TM light and operation
in reflection, so the radiation propagates twice through the
sensing region, with the LPG coupling light to the cladding
and recoupling back to the core. In this way accumulates a
phase difference between this light and the fraction that
always propagated in the core which provides the reference
beam of the interferometric layout. The recovery of the
interferometric phase shall allow obtaining the component
introduced by the plasmonic resonance, therefore with
information on the refractive index of the surrounding
medium. An approach of this type is quite promising
considering it is fully compatible with single-mode fiber-optic
technology while not compromising the performance of SPR
measurement since, in the sensing region, the light is allowed
to propagate with oscillation angles close to the critical angle
where the phase shift introduced by the SPR phenomenon is
maximum, i.e. in a situation of maximum sensitivity. The
theoretical analysis of such structure has been performed [21]
but further developments need to be undertaken to assess its
main characteristics for sensing, particularly in what concerns
the experimental investigation component.
III. SENSOR LAYOUT AND ANALYSIS
This feature indicates the need to consider a different type of
sensing head in fiber-optic SPR sensing systems when the
target is to achieve very high resolutions. Indeed, up to now no
development effort has been done towards optical fibers well
suited to SPR sensing, as opposed to the large endeavors done
along the years to optimize the prism approach. Even in the
recent appearance of microstructured optical fibers in SPR
sensing, the focus was on the improvement of the access
conditions of the measurand to the optical field and in the
T
HE structure of the sensing head under analysis is
schematically illustrated in Fig. 1. The sensing region is
an uncladded fiber coated with a metallic layer followed by a
layer of oxide material.
The length of the sensing region is assumed to be L, ρ0 the
radius of the core and θ0 the angle of incidence of the optical
4
√
(for TM waves),
(3b)
where εk and μk are the dielectric constant and relative
magnetic permeability, respectively. The reflectance Rs,p of the
multilayer structure for each polarization state is obtained as:
|
Figure 1. Schematic illustration of the SPR sensing head under analysis. The
cladding of the optical fiber has been removed in the sensing region, and
replaced by a metallic film covered with an oxide overlayer.
| ,
where rs and rp are the Fresnel reflection coefficients which
are calculated directly from the transfer-matrix elements in (1)
and the parameters η of the incidence and transmission media:
ray with respect to the normal to the core-cladding interface. A
geometric ray methodology combined with the transfer-matrix
formalism for stratified optical media is applied here,
following the approach of Gupta and co-workers [16, 22, 23].
In order to evaluate the optical response of the sensing
structure, we used the multilayer transfer-matrix theory
applied to a four-layer system comprised of: (i) fiber-core, (ii)
inner-metal, (iii) outer-metal, and (iv) probe-medium [23, 24].
This methodology is applied to compute the propagation of the
radiation through the layer system using the Maxwell’s
equations subject to boundary conditions between two
adjacent layers. From the basis of this formalism the
electromagnetic quantities, the tangential components of the
electric and magnetic field vectors (E and H, respectively) at
the first and final boundaries, are related by [25]:
[
]
[
][
],
(1)
(4)
.
(5)
The normalized transmitted power through the sensing region
of the fiber is given by: [26]
∫
⁄
[
]
∫
.
⁄
(6)
As shown in (4), Rp is the reflectivity of the multilayer
structure and θcr = sin-1(n1/n0) is the critical angle for the light
confinement inside the fiber (n0 and n1 are respectively the
refractive indices of the core and the cladding of the fiber).
Nref ( ) is the number of reflections the optical ray with θ0
undergoes in the sensing region, and is given by
re
.
tan
(7)
Moreover,
where (U,V) = (Ey,Hz) for the TE modes or (U,V) = (Hy,-Ez)
for the TM modes, and [mij] is the transfer matrix of the
multilayer structure comprised of N interfaces:
[
]
∏
[
],
(2a)
with δk being the phase in the kth layer (k = 0 and k = N are the
first and final semi-infinite media, respectively), given by:
√
,
(2b)
where dk, nk and θk are respectively the thickness, refractive
index and propagation angle (with respect to the normal to the
direction of stratification) of the kth layer, λ is the free-space
wavelength of the light propagating in the system, and nonmagnetic media are assumed (i.e. µk = 1). The parameter ηk in
(2a) is defined as a function of the polarization state as:
sin
(
cos
cos
is the intensity distribution between the continuum of guided
modes. In order to access the performance of phase
interrogation of SPR based sensors supported by fiber
structures, it is convenient to express the amplitude reflection
coefficients rs and rp in the polar form as:
| |
| |
and
(for TE waves),
(3a)
.
(9)
Then the phase di erence variation δϕp,s resulting from a
single reflection at a given incident angle between the p and
the s polarization components is:
,
(10)
and the total phase difference variation considering all the
reflections inside the sensing region is given by:
re
√
(8)
)
.
(11)
At this stage, it is useful to define the sensitivity of the sensing
structure to the refractive index ns of the surrounding medium,
which can be regarded as the basic parameter quantifying the
performance of the sensor. A change in the refractive index
ns = (εs)1/2 of the surrounding dielectric produces a significant
5
variation in the propagation constant of the surface plasmon,
resulting in the modification of the SPR coupling condition.
This can be observed as a change in one or more of the
properties of the light transmitted through the sensing
structure. The sensitivity can be defined as:
set for the analysis that provided the results delivered in the
following.
(12)
where δξ is the change in a given property of light (e.g.
intensity, resonance wavelength or phase) due to a variation of
δns in the refractive index of the surrounding medium. We will
be particularly concerned with the case of phase interrogation,
where the property of light of interest is the phase difference
Δ p,s between the p and s polarization components. In this
case, the phase sensitivity to refractive index variations is then
defined as Sn,ϕ = δΔϕp,sp,s/δns.
IV. RESULTS AND DISCUSSION
THE analysis detailed in the previous section was applied to
the study of the characteristics of a fiber optic SPR
metal/oxide double-layer sensing structure for measurement of
surrounding medium refractive index variations. A multimode
silica fiber with 100 µm core diameter and numerical aperture
of 0.24 was considered. In the sensing region the uncladded
length was assumed to be L = 1cm.
The choice of the metal for the internal layer was determined
by the best performance achieved concerning the phase
sensitivity of the double layer structure. Silver was found to be
the most adequate metal, with the shape of the SPR resonance
versus thickness of the silver layer (without the presence of
the external oxide layer) shown in Fig. 2.
Figure 2. Resonance of a fiber optic SPR structure with a single silver layer
(thickness in nanometers; refractive index of the external medium: 1.33).
When the silver layer was combined with an outer layer of
oxide, the thickness of 50 nm for the silver layer was the one
that provided the best sensor performance, so this value was
Figure 3. Spectral behavior of the double layer fiber optic SPR structure for
different thicknesses of the oxide layer (TiO2, SiO2 and Al2O3 oxide
materials are associated with (a), (b) and (c), respectively; refractive index of
the external medium: 1.33).
6
Three different oxide materials were considered for the outer
oxide layer, titanium dioxide (TiO2), silicon dioxide (SiO2)
and aluminum oxide (Al2O3). It is known that the introduction
of this layer modifies substantially the characteristics of the
SPR resonance, most notably the spectral region where it
appears [16]. This can be checked by the observation of Fig. 3
which shows the location and shape evolution of the resonance
for different thicknesses of the oxide layer. The SPR sensing
head is assumed to be surrounded by a non-dispersive
external medium with refractive index of ns = 1.33.
Figure 5 shows the normalized transmittance and the phase
di erence Δϕp,s versus ns for layouts with TiO2 and Al2O3
layers of 15 nm and 30 nm, respectively, which enable to
operate the sensor in the first telecommunications window (the
data is shown for λ = 830 nm; from the data shown in Fig. 3-b,
this wavelength is not accessible by considering a SiO2 layer).
The inclusion of a TiO2 layer permits the largest displacement
of the spectral response of the SPR resonance, turning possible
to locate it within the 830, 1300 and 1550 nm
telecommunication bands by adjusting the layer thickness
within an interval technically feasible. This has also
implications in the width of the resonance, which consistently
becomes larger when it moves to longer wavelengths, as is
shown in Fig. 4 which displays the spectral width of the
resonance versus the thickness of the oxide layer.
When addressing phase interrogation of the fiber optic based
SPR sensing structure, a single oscillation angle of the light in
the fiber, associated to a specific propagation mode, shall be
considered (oscillation angle defined by the fiber axis and the
light ray). The higher the numerical aperture of the fiber the
larger this angle can be, meaning a smaller incident angle in
the interface core-cladding of the fiber. In the fiber length
where SPR occurs this condition favors the introduction of a
larger phase di erence Δϕp,s between the TM and TE
polarizations of the light in the reflection process, increasing
the sensitivity to variations of the refractive index of the
surrounding medium, ns. Therefore, it is advantageous to
consider an incident angle in the core-metallic layer interface
just near the critical angle in the cladded fiber, which is
θc ≈ 80.5° for a fiber with numerical aperture of 0.24. As such
the phase interrogation analysis was performed assuming
θ = 80.5°.
Figure 5. Normalized transmittance and phase difference (between p and s
polarizations) as function of the refractive index of the surrounding medium
for an inner silver layer of 50 nm combined with a TiO2 layer of 15 nm (a) and
a Al2O3 layer of 30 nm (b). The interrogation wavelength is 830 nm.
Figure 4. Evolution of the spectral width of the SPR resonance (full width at
the power level 6 dB above the minimum) as function of the thickness of the
oxide layer. The refractive index of the external medium is considered to be
1.33.
The analysis of this data shows that the maximum of the
refractive index phase sensitivity, S ,ϕ = δ(Δϕp,s)/δ s, occurs for
ns = 1.330 and for the case of the Titanium Dioxide layer
reaching a value of 3.2×106 degrees/RIU, while for the
situation of Aluminum Oxide layer the value is
3.0×106 degrees/RIU for ns = 1.328. In both cases the
condition of maximum phase sensitivity is on the SPR
resonance, where the transmitted optical power is rather low.
This means almost no light to detect, therefore decreasing
substantially the signal-to-noise ratio. As such, a compromise
is needed so that the phase sensitivity is degraded to a certain
extent while the signal-noise-ratio in the detection process
increases, therefore reaching a situation of maximum
resolution in the measurement of ns.
For the sensing head to operate in the second (1300 nm) and
7
third (1550 nm) telecommunications spectral windows, the
observation of Fig. 3 indicates the need to consider an external
layer of TiO2 with thicknesses of 40 and 55 nm, respectively.
The normalized transmittance and Δϕp,s versus ns are shown in
Fig. 6.
noise ratio to a point that such high sensitivity does not have
impact in the final phase reading resolution. Therefore, shall
be considered to operate in a wavelength where there is
enough transmitted optical power, which means in principle it
can be located in either side of the SPR transmittance curve.
To investigate this aspect, it was analyzed the achievable
phase sensitivity for different levels of transmittance in both
sides of the SPR resonance, for the case being selected the
situation associated with Fig. 6-a, i.e. relative to an operating
wavelength of 1300 nm. The results obtained are shown in
Figure 7.
Figure 7. Phase sensitivity versus normalized transmittance for a fiber optic
SPR structure with an inner silver layer of 50 nm combined with a TiO2 layer
of 40 nm (interrogation wavelengths of 1300 nm).
Figure 6. Normalized transmittance and phase difference (between p and s
polarizations) as function of the refractive index of the surrounding medium
for an inner silver layer of 50 nm combined with a TiO2 layer of 40 nm (a) and
a TiO2 layer of 55 nm (b), associated to interrogation wavelengths of 1300 nm
and 1550 nm, respectively.
From these results it turns out the maximum phase sensitivity
has de value of 5.6×105 degrees/RIU (at ns = 1.334) and
3.3×105 degrees/RIU (at ns = 1.332) when the interrogation
wavelengths are 1300 and 1550 nm, respectively.
When considering SPR sensors supported by optical fiber
platforms it is advantageous to operate within these
telecommunications windows, particularly the one in the
1550 nm region, due to a wide availability of high
performance optical fiber technology at a relatively low cost,
allowing the design and implementation of sensing structures
with rather good characteristics.
As mentioned before, maximum phase sensitivity occurs in a
region close to the dip of the SPR resonance, therefore with
low level of optical power, which will degrade the signal-
With the wavelength fixed, to access different transmittance
levels means different refractive index of the surrounding
medium. As expected from the observation of Fig. 3-a, due to
the asymmetry of the SPR resonance, its left side provide
limited freedom to the choice of the operating pair
(transmittance,sensitivity), indicating the preference to work
out in the right side of the resonance.
Optical sensors based in the SPR phenomenon are inherently
highly sensitive to variations of the external medium refractive
index, permitting therefore to detect very small changes of this
measurand, for example variations induced by biochemical
processes. Therefore it is in the context of biochemical sensing
that these devices are rather appellative, pointing out the
relevance of optimizing their performance in the situation
where the external medium is water. Consequently, for each
type of the oxide layer it was determined the appropriate
thickness in order to attain maximum sensitivity at the
refractive index of water evaluated at the sensor operating
wavelength.
Figure 8 shows the maximum phase sensitivity obtained when
the operation wavelengths are 630 nm (a) and 830 nm (b). The
most immediate outcome of the analysis of this data is that for
the case of operation at 630 nm it is possible to find the
required thickness for all oxide materials considered (TiO2,
SiO2 and Al2O3). However, when moving to 830 nm, the layer
8
of SiO2 leaves to be an option, as the results shown in Fig. 3b) already indicated.
Figure 9. Same as Fig. 8 but now with interrogation wavelengths of 1300 nm
(a) and 1550 nm(b).
Figure 8. Thickness of the oxide layer required to operate with maximum
sensitivity to changes of the refractive index of water, determined by the
intersection of the horizontal line (that defines in the y-axis the water
refractive index at the wavelength of operation) with the curves associated
with different materials for the oxide layer. The interrogation wavelength is
630 nm (a) and 830 nm(b).
Figure 9 gives similar results but now considering as operating
wavelengths 1300 nm (a) and 1550 nm (b). Again, the
titanium dioxide option for the external layer permits to
address operation at both wavelengths, eventually being
feasible to consider a layer of aluminum oxide for 1300 nm
interrogation, associated with a tackiness width of ~86 nm,
which is still technically feasible.
V. FINAL REMARKS
The analysis presented here permitted to identify the main
characteristics of SPR fiber optic sensing structures with a
double layer of metal/oxide, particularly when phase
interrogation is under concern. Two of these characteristics are
particularly relevant, one the possibility to tune the sensor
sensitivity to variations of the refractive index of the
surrounding medium, the other the amenability to tune the
measurement operation to a specific wavelength region, most
notably the one around 1550 nm to benefit from the large
portfolio of high performance fiber optic components
available at these wavelengths. It was found the inclusion of
an internal silver layer of 50 nm thickness permits the better
performance for any of the three different oxide materials
considered for the external layer (TiO2, SiO2, and Al2O3).
Also, the results obtained shows the better flexibility is
provided by the inclusion of a layer of titanium dioxide,
approach that is compatible with tuning the sensor operation to
a large wavelength window that includes operationally
important wavelengths (630, 830, 1300 and 1550 nm).
9
For this fiber optic based plasmonic structure to be operative
as a sensing device one of its characteristics, that show
dependence to the refractive index of the surrounding medium,
shall be selected and adequate approaches developed to read
its variations with the better performance possible. The main
features of this process, known as sensor interrogation, were
outlined in the first part of this work, with particular focus
when the characteristic selected is the phase di erence, Δϕp,s,
between the light with TM and TE polarizations that
propagates through the sensing head. This choice oriented
most of the theoretical research performed and the main
results obtained and presented here.
[2]
The experimental validation of this study can be done by
application of techniques developed along the years for the
interrogation of polarimetric fiber sensing interferometers
illuminated by monochromatic light [27]. Following these
approaches, laser light is injected into the multimode fiber at
an angle such that is excited the selected core mode. At the
input of the sensing head, a fiber polarizer element shall be
included in order to have light in the TE and TM polarizations,
recombined at the output of the measurement region by means
of a second polarizer. This will generate an interferometric
signal with a phase proportional to the optical path difference
accumulated by the propagation of the TE and TM
polarizations, therefore including the phase variations of the
light in the TM polarization induced by changes of the
surrounding refractive index.
[8]
To recover the phase of this signal, a robust and sensitive
technique involves to add to the interferometric phase a
sinewave component, which can be generated by sinewave
modulation of the wavelength of the laser light. After
photodetection and incorporation of signal processing, which
includes a feedback loop to the laser source to tune the DC
value of the light wavelength, the information on the quasistatic phase of the polarimetric sensing interferometer, and
therefore on the refractive index of the external medium, can
be obtained from the feedback signal [28]. This approach is
compatible with a DC phase resolution of the order of 1 mrad,
which means from the results presented here refractive index
measurement resolutions of the order 10-7 RIU shall be
possible.
In conclusion, a theoretical investigation of optical fiber
surface plasmon resonance sensors incorporating an internal
metallic layer (Ag) covered with an oxide layer was reported.
Different oxide materials (TiO2, SiO2, and Al2O3) were
considered and the main characteristics of these structures for
refractive index sensing determined. Globally, the results
presented in this work permit to guide system design in order
to achieve optimized performance when phase reading of SPR
sensors is under concern.
[3]
[4]
[5]
[6]
[7]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
REFERENCES
[1]
S. Enoch, N. Bonod, Plasmonics: From Basics to Advanced Topics.
Springer, 2012.
[28]
A. K. Sharma, R. Jha, B. D. Gupta, Fiber-optic sensors based on surface
plasmon resonance: A comprehensive review. Sensors Journal, 7, 11181129, 2007.
S. A. Maier, Plasmonics: Fundamentals and Applications, Springer
2007.
A. Sommerfeld, Uber die Fortpflanzung Electrodynamischer Wellen
Langs Eines Drahtes, Annalen der Physik und Chemie 67, 233-290,
1899.
R. W. Wood, On a Remarkable Case of Uneven Distribution of Light in
a Diffraction Grating Spectrum, Proceedings Physical Society London,
18, 269-275, 1902.
B. Liedberg, C. Nylander, I. Lunström, Surface Plasmon Resonance for
Gas detection and Biosensing, Sensors and Actuators, 4, 299–304,
1983.
Editorial, Surface plasmon resurrection. Nature Photonics, 6, 707-794,
2012.
R. C. Jorgenson, S. S. Yee, A fiber-optic chemical sensor based on
surface plasmon resonance, Sensors and Actuators B, 12, 213-220,
April, 1993.
H. A. Atwater, The promise of plasmonics, Scientific American, 296,
56-63, 2007.
J. Homola, S. S. Yee, G. Gauglitz, Surface plasmon resonance sensors:
A Review, Sensors and Actuators B, 54, 3-15, 1999.
J. Homola, Present and future of surface plasmon resonance biosensors,
Analytical and Bioanalytical Chemistry 377, 528-539, 2003.
H. Suzuki, M. Sugimoto, Y. Matsui, J. Kondoh, Effects of gold film
thickness on spectrum profile and sensitivity of a multimode-opticalfiber SPR sensor, Sensors and Actuators B, 132, 26-33, 2008.
S. K. Ozdemir, G. Turhan-Sayan, Temperature effects on surface
plasmon resonance: Design considerations for an optical temperature
sensor, Journal of Lightwave Technology 21, 805-814, 2003.
H. R. Gwon, S. H. Lee, Spectral and angular responses of surface
plasmon resonance based on the Kretschmann pri m co figu io ,
Materials Transactions, 51, 1150-1155, 2010.
R. Naraoka, K. Kajikawa, Phase detection of surface plasmon resonance
using rotating analyzer method, Sensors and Actuators B, 107, 952-956,
2005.
P. Bhatia, B. D. Gupta, Surface-plasmon-resonance-based fiber-optic
refractive index sensor: sensitivity enhancement, Applied Optics, 50,
2032-2036, 2011.
N. Díaz-Herrera, D. Viegas, P. Jorge, F. M. Araújo J.L. Santos, M.C.
Navarrete. A. González-Cano, Fiber-optic SPR sensor with a FBG
interrogation scheme for readout enhancement, Sensors and. Actuators
B, 144(1), 226-231, 2010.
Y. C. Li, C., Y. F. Chang, L. C. Su, C. Chou. Differential-phase surface
plasmon resonance biosensor, Analytical Chemistry, 80, 5590-5595,
2008.
H. P. Chiang, J. L. Lin, Z. W. Chen. High sensitivity surface plasmon
resonance sensor based on phase interrogation at optimal wavelengths.
Applied Physics Letters 88: 141105, 2006.
S. P Ng, C. M. L. Wu, S. Y. Wu, H. P. Ho. White-light spectral
interferometry for surface plasmon resonance sensing applications.
Optics Express 19, 4521-4527, 2011
Y. J. He, Y. L. Lo, J. F. Huang. Optical-fiber surface-plasmonresonance sensor employing long-period fiber gratings in multiplexing.
Journal of Optical Society of America, 23, 801-811, 2006.
B. D. Gupta, C. D. Singh. Evanescent-absorption coefficient for diffuse
source illumination: uniform- and tapered-fiber sensors. Applied Optics,
33, 2737-2742, 1994.
A. K. Sharma, B. D. Gupta. On the performance of different bimetallic
combinations in surface plasmon resonance based fiber optic sensors,
Journal Applied Physics 101, 093111, 2007.
M. Kanso, S. Cuenot and G. Louarn, “Sensitivity o optical iber sensor
based on sur ace plasmon resonance: Modeling and experiments,”
Plasmonics, vol. 3, pp. 49-57, 2008.
M. Born and E. Wolf, Principles of Optics, seventh (expanded) edition,
Cambridge University Press, 1999.
A. Ankiewicz, C. Pask and A. W. Snyder, “Slowly varying optical
ibers,” J. Opt. Soc. Am., vol. 72, pp. 198-203, 1982.
B. Culshaw, “Inter erometric optical ibre sensors,” Journal o Optical
Sensors, vol. 1, pp. 237-252, 1986.
C. Jauregui, F. M. Araujo, L. A. Ferreira, J. L. Santos and M. LópezHiguera, “Interrogation o low-finesse Fabry-Pérot cavities based on
modulation of the transfer function of a wavelength division
multiplexer,” J. Lightwave Technol., vol. 19, pp. 673-681, 2001.