The Fibre Optical Sensors

Recently, fibre optical sensors (FOS) have gained increased popularity
and market acceptance. In comparison to conventional sensors they
offer a number of distinct advantages which makes them unique for
certain types of applications, mainly where conventional sensors are
difficult or impossible to deploy or can not provide the same wealth
of information.

According to the spatial distribution of the measurand (the quantity to
be measured), FOS can be classified as...

Point sensors: the measurement is carried out at a single point
in space, but possibly multiple channels for addressing multiple points.
Examples are Fabry-Perot sensors and single Fibre Bragg Grating (FBG)

Integrated sensors: the measurement averages a physical
parameter over a certain spatial section and provides a single value.
An example is a deformation sensor measuring strain over a long base

Quasi-distributed or multiplexed sensors: the measurand is
determined at a number of fixed, discrete points along a single fibre
optical cable. The most common example are multiplexed FBG's.

Distributed sensor: the parameter of interest is measured with
a certain spatial resolution at any point along a single optical cable.
Examples include systems based on Rayleigh, Raman and Brillouin


Completely passive: can be used in explosive environment.

Immune to electromagnetic interference: ideal for microwave

Resistant to high temperatures and chemically reactive environment:
ideal for harsh and hostile environment.

Small size: ideal for embedding and surface mounting.

High degree of biocompatibility, non-intrusive nature and
electromagnetic immune: ideal for medical applications like
intra-aortic balloon pumping.

Can monitor a wide range of physical and chemical parameters.

Potential for very high sensitivity, range and resolution.

Complete electrical insulation from high electrostatic potential.

Remote operation over several km lengths without any lead
sensitivity: ideal for deployment in boreholes or measurements
in hazardous environment.

Multiplexed and distributed sensors are unique in that they
provide measurements at a large number of points along a single
optical cable: ideal for minimising cable deployment and cable
weight, or for monitoring extended structures like pipelines, dams

In what follows we give a brief explanation of the working principles
of optical fibres and each type of sensors.

An optical fibre consists of a thin, low-loss glass wire with a
centre or core region having a slightly higher refractive index
than its surrounding region or cladding.

Figure 1 - Schematic of step-index optical fibre

Figure 1 shows a schematic of a step-index optical fibre. Light is
guided inside the core region by total internal reflection at the
core-cladding interface. Depending on the size of the core region,
one single or multiple light paths (modes) are permitted to
propagate, referred to as single-mode or multimode fibre. Typically,
the bare optical fibre has an outer diameter of 125µm with a
core diameter of 9µm in the case of single-mode fibres and 50µm
or 62.5µm for multimode fibres.

Different protective coatings are applied to protect the fibre
from possible mechanical damage.

Consider the schematic of an optical pressure transducer shown
in Figure 2.

Figure 2 - Schematic of pressure transducer based on Fabry-Perot

Essentially, it consists of a pair of parallel mirrors separated by
an air gap Ls. This arrangement is referred to as a Fabry-Perot
(FP) cavity or sensing interferometer. A semi-reflective mirror 1
is formed by depositing a dielectric layer at the end of the optical
fibre. Mirror 2 is formed by a diaphragm mounted in front of the
optical fibre. Exposing the diaphragm to the pressure p to be
measured changes the gap Ls. Hence, by measuring Ls the applied
pressure p can be determined. Different pressure ranges can be accommodated by appropriately selecting thickness and diameter
of the diaphragm to keep the maximum deflection of similar value
and maintain a linear relation between pressure and deflection.

The preferred light source employed to measure the gap Ls
optically is a so-called white-light or broad-band light source.
It emits light at various colours (equivalent to a broad band of
wavelengths) simultaneously. The resulting light therefore
appears as having no specific colour, i.e. it looks white. In
contrast to a laser, the light is generated as a large number
of short pulses. These pulses are emitted in a random fashion
with no fixed phase relation between them. As a result, they
do not interact or interfere with each other and for the following
it is sufficient to consider a single pulse only.

Next, consider what happens if the FP cavity is illuminated by
a white-light source. The incident light guided in the optical
fibre towards the FP cavity is partially reflected at the first
mirror. The remaining light is transmitted and subsequently
reflected by the second mirror. Hence, the original light pulse
is split into two return pulses with the second pulse being
delayed by t = 2Ls / c with respect to the first one, c
denoting the speed of light.

Interference (and hence, a signal containing information about
Ls) only occurs if the two pulses generated from the same
original pulse can be brought back to overlap again. This is
achieved by employing a second (or readout) interferometer.

Figure 3 - Schematic of white-light sensor system

Here for instance, the interferometer consists of two non-parallel
mirrors, operating in transmission. As can be seen, the air gap
Lr(x) depends on the position x along the mirrors and a maximum
interference signal is generated at the position x0 where Lr(x0)
is exactly matched to the gap Ls of the sensing interferometer.
This position x0 is easily determined by a CCD array mounted
behind the mirrors.

In practice, it is beneficial to replace the two non-parallel mirrors constituting the readout interferometer with a birefringent wedge arrangement, details of which can be found in downloadable
information - click

Employing the same basic principle, a whole range of transducers
for measuring different quantities can be constructed. Examples
include temperature, displacement, strain, force and refractive
index as shown below. To download this information, click

Schematic of position transducer based on polarization

Schematic of Strain/Force transducer based on Fabry-Perot

Schematic of temperature transducer based on polarization

An example of an integrated sensor is the deformation sensor, a
schematic of which is shown in Figure 4.

Figure 4 - Schematic of integrated strain sensor

It is based on the same basic principle of white-light interferometry
as described in the section on point sensor above.
Light from a white-light source is transmitted via optical fibre to
the optical sensor. Here the sensor consists of a fibre optical coupler branching out into two fibres of different length and having a
miniature mirror attached at each end. This configuration is referred
to as a Mach-Zehnder or sensing interferometer. The incident light
pulse is split into two pulses by the coupler and on return the two
pulses are separated in time by t = 2nLs / c, with n denoting the
refractive index of the glass fibre, Ls the length difference of the
fibres and c the speed of light. One fibre is attached to the
structure under test whereas the other fibre is in close proximity
but not attached.

A deformation of the structure leads to a change in path difference
2nLs. In this case it is measured by employing a scanning mobile
mirror in the second (receiving) interferometer housed in the reading
unit. As before, a maximum interference signal only occurs if the
path difference of the sensing interferometer 2nLs exactly matches
that of the receiving interferometer 2Lr. The sensor is temperature independent as any change in temperature has the same effect on
both fibres, leaving the path difference effectively unchanged. The
distance between the anchoring points at which the fibre is attached
to the structure is called base-length. It can be set between
10cm and 10m, resulting in the average strain over the base-length
being measured.

One of the most common quasi-distributed optical sensors is based
on Fibre Bragg Gratings (FBG). A FBG is formed along a short section
of the optical fibre by introducing a periodic modulation of pitch
on the refractive index.
At each period a tiny fraction of light is reflected back, leading to
a strong reflection at a certain wavelength called Bragg wavelength.
The Bragg wavelength
lB is given by lB = 2nL, with n being the
refractive index of the fibre. Only at this wavelength all fractions
add up in phase resulting in a strong reflection signal.

Figure 5 - Schematic of multiplexed sensor employing Fibre Bragg

When the FBG is strained or exposed to heat, both grating pitch
and refractive index n are affected and the Bragg wavelength is
shifted accordingly. This provides a measure for strain and
temperature. As both effects occur simultaneously, additional
measures need to be taken to distinguish between them. For
example, when measuring strain a second FBG not attached to
the structure under test can be deployed next to the first FBG
to provide temperature compensation.

The great benefit of FBG is that a number of FBGs each with a
different Bragg wavelength
l1, l2, …lN can be deployed along
the fibre. This provides N measurement points within a single
cable. A tuneable laser source is used to illuminate the array of
sensors. During the scan, each time the laser wavelength
matches one of the Bragg wavelengths
li a strong backreflected
signal is recorded providing information about temperature and
strain at location i.

In a distributed sensor the parameter of interest is measured
with a certain spatial resolution at any point along a single
optical cable. The basic underlying physical processes for
realising a distributed sensor are provided by various scattering
processes. As the laser light is propagating along the optical
fibre, a small amount of light is continuously scattered back
at each location along the fibre. Three basic scattering
processes are important in silica fibre:

Rayleigh scattering due to reflections at random
inhomogeneities of the refractive index frozen in during
manufacture of the fibre.

Raman scattering due to interaction with molecular
vibrations and rotations in the glass.

Brillouin scattering due to interaction with inhomogeneities
created by sound waves in the fibre (acoustic phonons).

When analysing the backscattered light in the wavelength domain,
one finds that the Rayleigh scattering component is of the same
l0 as the incident light. There are two Raman
components shifted by the same amount above (Stokes component)
and below
l0 (Anti-Stokes component). Similarly, the Brillouin
backscatter consists of two components shifted below and above

Figure 6 - Scattering processes in optical fibre

The fact what makes these scattering processes interesting for
optical sensing is that the property of the backscattered light
depends on strain and/or temperature in the fibre. As indicated
in Figure 6, the intensity of the Raman Anti-Stokes component
increases with increasing temperature T whereas the Stokes
component can be regarded as temperature independent. Hence,
by taking the ratio between them one excludes other (common
to both Stokes and Anti-Stokes components) possible causes
of intensity variations like fibre bending losses and the
temperature can be determined unambiguously. In the case
of Brillouin scattering it is the wavelength shift of the scattered
components with respect to the Rayleigh wavelength that
changes with both temperature T and strain
e . Hence, by
extracting this wavelength shift from the backscattered light
a sensor for strain and temperature can be realised.

Additional measures have to be taken to separate strain and
temperature dependence, like the installation of a reference
cable not rigidly bound to the structure and therefore, measuring temperature only.

The most common way to extract the spatial distribution is the
use of pulsed light and recording the backscattered light
characteristics against time (Optical Time Domain Reflectometry
or OTDR). In this way, it is possible to extract the temperature
and/or strain profile in space.

Figure 7 - Schematic of distributed Brillouin sensor

[1] J.M. Lopez-Higuera, 'Introduction to fibre optic sensing technology',
in Handbook of optical fibre sensing technology, J.M. Lopez-Higuera Ed, Wiley, 2002.

Article presented by...
Dr Ralf Pechstedt
Manager, Fibre Optic Instrumentation
Accurate Controls Ltd.
25 Cowley Road, Nuffield Industrial Estate, Poole, Dorset BH17 0UJ UK
Tel: +44 (0) 1202 678108
Fax: +44 (0) 1202 670161


For details of Sensor Suppliers, click here...

Home - Website - Search - Suppliers - Links - New Products - Catalogues - Magazines
Problem Page - Applications - How they work - Tech Tips - Training - Events - Jobs