Completed Research 1995 to 2006

Theme 2 Projects (1999-2002):
Fibre Optic Sensor Technology

THEME 2 OVERVIEW

(All project descriptions are provided as proposed in 1998 as part of the NCE mid-term review)

Long Gauge Fibre Optic Sensor System (T2.1)
Project Leader: Dr. Roderick Tennyson, University of Toronto Institute for Aerospace Studies

This research project interacts strongly with Projects T2.2, T2.3, T2.4 and T2.5. In addition, a collaborative effort to investigate pultrusion effects on the optical performance and life of FBG and FP sensors will be undertaken with Dalhousie University and RocTest Ltd. The research program can be subdivided into three main categories:

1. Long gauge fibre optic sensors.
2. Optical performance and life of FBG and FP sensors.
3. Use of FBG and FP sensors in service.

A major research effort to develop a reusable long gauge fibre optic sensor will be undertaken. At the present time, we have successfully manufactured long gauge sensors using FBG technology [Huang, S., et al., 1996]. However, they are for one time use only because these sensors are currently bonded to the structure. We plan to develop a new configuration that permits their attachment and removal from structures. At the same time, the static, manual compensation instrument that we developed for measuring strains in long gauge sensors will be significantly redesigned using current optoelectronic technology and new concepts to upgrade its performance for automated measurements of static and dynamic strains. The new design will be based on a concept developed by one of our recent Ph.D. graduates from this project.

Our research to date has demonstrated that FBG sensors suffer significant tensile strength reductions due to the treatment of the bare core by ultra violet radiation in the manufacture of FBGs. Continued research will be directed towards modifying the method of manufacture, and the addition of special coatings to minimize strength reductions.

The untreated FBGs have been successfully imbedded in glass/ vinyl ester composite tendons manufactured by the pultrusion process at Dalhousie University. Research is now required to assess the long term optical performance of those FBG sensors, as well as FP gauges, and to develop connector devices for reliable read outs. In addition, the thermal correction for FBG sensors still needs development for field use, when service loads are applied to a structure and their associated strain fields recorded. When a large number of FBGs are incorporated into a structure, either in the form of an array along a single fibre or a series of fibres, multiplexing is required in order to interrogate the structure. Thus, considerable effort will be devoted to testing linear arrays and parallel systems using the instruments under development in Projects T2.3 and T2.4. We are also continuing the development of methods of measuring the strain along an extended FBG sensor. As noted, this technology will provide a means for assessing the integrity of the FBG and FP sensor bonds to the host structure.

See Leslie Street Bridge Field Demo

Back

As described in the Progress Report, a great many installations of short and long gauge fibre optic Bragg gratings (as well as Fabry Perot sensors) have been made across Canada [Maaskant, R., et al., 1996]. It is now intended over the next three years to return to these sites with existing instruments (and those from Projects T2.1, T2.3, T2.4 when they become available) to perform an assessment on the multitude of sensors.

The purpose of this aspect of our research project is to assess the optical performance and operational life of these surface mounted and imbedded sensors. No such data base exists for FBG sensors based on field site installations. This will provide, for the first time, a history on FBG reliability and performance as a function of installation time and environmental cycles, and comparisons with similar data on FP sensors.

We will supply FP gauges and specially prepared optical fibres for the pultrusion work at Dalhousie University. If these sensors survive the pultrusion, the pultruded element will be loaded to ascertain the maximum strain the sensors can survive. Close collaboration will be required between Projects T2.1 and T2.2 in terms of developing a method of interrogating the strings of Bragg grating sensors. We expect to work with Electrophotonics Corporation and RocTest Ltd. on the development of special optical connections for the smart tendons and will also study the types of coating required to give both good strain coupling and long life to the pultruded optical fibre sensors.

We will also explore the possibility that this type of measurement can be used in the development of anchorage systems and may be used for studying strain along smart tendons. This project will also assume the responsibility of working with other ISIS researchers on the installation of FBG, long gauge sensors and FP gauges for both field and laboratory applications.

We shall investigate the concept of distributed sensing using multiple FBGs on optical fibres imbedded in FRP structures /concrete to demonstrate strain distributions along a particular structure or component member. This will involve Projects T2.1, T2.2 and T2.4.

Finally, software will be developed to provide an interface between the response of FBGs and the computer data acquisition systems used by ISIS researchers and industry. It is recognized that special software is required for FBG sensors to provide a means of visually inspecting (ie: in a graphical format) the dynamic response of the sensors after installation for particular field tests.

Back

This research program will interact closely with Projects T2.5 and T2.6 to define the specifications of tunable lasers and wavelength measurement devices needed for the demodulation systems required by T2.1. The overall goal is to design and build:

1. a new 'true' parallel multichannel demodulation system for FBG sensors.
2. a demodulation system for a multiplexed string of FBG sensors.
3. a distributed sensing demodulation system for an extended single FBG.

To interrogate a string of grating strains, expected with the 'smart'tendon to be used in bridge girders, it will probably be necessary to use either wavelength division multiplexing (WDM) or a combination of VV DM with time gating [Coroy, T., et al., 1996]. This will impose a much wider tuning range on the laser diode and this will have to be studied further. We are currently designing a 32 channel fibre optic strain sensing system that will be truly parallel using a DFB laser diode provided by Nortel Technology. This laser source allows a broad wavelength tuning range (l 0nm) and very fast scanning speed (less than 1 ms for a full scan), thus allowing parallel dynamic measurements. Such a laser necessitates the need for accurate measurement of its wavelength. This will be undertaken by means of a simple passive quantum well spectral radiometric technique for a wavelength tuning range of less than 10 nm.

Since the demands on the laser are quite severe it is probably prudent to follow standard practice of positioning a mesa detector adjacent to the laser rather than trying to develop the technology of selective area epitaxy so they can share a common substrate, as originally proposed. The important difference is that in this work, the detection system will monitor both the wavelength and the power of the laser rather than just its power as is the practice. This detection system will include a mesa quantum well detector used either in the passive or active mode. We have recently demonstrated a crude form of such a system. It might also be noted that the primary rational for trying to mount both a quantum well detector and laser on the same substrate was the extreme simplicity and potential small size in the resulting demodulation system.

To date, construction of an automatic characterization system for quantum well electroabsorption devices (AWED) and other optoelectronic components has been completed. This system is capable of working with unpackaged (chip) devices, and can characterize a component's performance in a variety of passive and active wavelength measurement systems. Temperature compensation techniques will be evaluated as a possible means of increasing the resolution/ bandwidth of QWED based wavelength measurement systems. Wavelength measurement systems based on the use of quantum well electroabsorption devices have been demonstrated for single Bragg grating sensors, with a resolution of 3.0µe/üHz or 3.62 pm / üHz and an accessible wavelength range of 90 nm.

Back

This research project will work closely with Projects T2.1, 2.2 and 2.3. The essential difference between this project and Project T2.3 is that this system is focussed on multiple FBG on a single fibre and not on parallel multiplexing. Project T2.3 will provide us with MQW detection systems, components for initial studies of fibre laser design, and much helpful advice and guidance.

There has been considerable work recently on the use of tunable detectors for de multiplexing signals from a FBG sensor network. Most of these previous techniques use some type of wavelength division multiplexing which puts substantial demands on components for systems with a large number of sensors. For our short pulse time division multiplexed system, relatively simple components can be used. One possible configuration would use a MQW electro absorption filter to gate the optical pulses before the detection system.

For initial experimental demonstrations, we will use the broad range, moderate sensitivity, wavelength selective filter for strain demodulation that has already been developed in the ISIS program. We will also explore the use of the same MQW electro absorption filter for both the timegating and the strain reading operations. Finally, we will investigate the capabilities of a fibre grating filter design to enhance sensitivity, and to permit tailoring filter properties to optimize our strain sensing system.

A fibre laser is a source of high intensity, very broadband light when mode locked to generate short pulses. The advent of diode laser pumping of fibre lasers has opened the door for the development of these broadband sources with: compact structure, high efficiency, stable output, and relatively low cost. We propose to build a mode locked fibre laser with a pulse repetition rate equalto or greater than the round trip time of light propagation in the fibre sensor network. The pulse width will be designed in the range of a few hundred femtoseconds to provide > 10 nm bandwidth necessary to fully exploit the range of the strain sensors. (One of the advantages of this short pulse technique is that it is relatively easy to generate pulses with bandwidths of up to 70 nm, allowing great flexibility in system design). Preliminary work on fibre sensor read out will be done using our existing short pulse laser sources.

Once the mode locked fibre laser has been built and tested, and the time gating and wavelength measuring elements have been fully characterized, we will build a breadboard demonstration system to serve as a testbed for the evaluation of our short pulse technique for multiple fibre sensor read out. We expect to demonstrate the capability of reading out greater than 20 serially multiplexed sensors with spacings as close as one metre.

Back

The purpose of this research project is to produce a continuous tunable laser for incorporation in the sensor and optoelectronic demodulation system being developed in Project T2.3. This work will require the development of split electrode MQW InGaAsP distributed feedback (DFB) lasers having band gap shifted (BGS) regions (that is, the regions in which the bandgap energy is increased), and to develop an understanding of defect induced quantum well intermixing for achieving the bandgap modification. For the purposes of tuning, the use of split electrode lasers rather than conventional single electrode lasers can be electrically contacted separately and biased independently. Hence, by independently adjusting the electrode currents it is possible to tune the lasing wavelength while retaining a relatively constant output power. Also, in order to achieve continuous wavelength tuneability combined with a very narrow line width, it is necessary to use DFB lasers, rather than conventional FP lasers [Chen, L., et al., 1997].

The approach for performing the bandgap shifting will be by using SiO2 capping over the region to be modified, with subsequent rapid thermal annealing (RTA) to produce the necessary changes in regions underlying the oxide layer. In order that the BGS technique has a minimal effect on the properties of the DFB grating, it is anticipated that the BGS will be carried out after the first growth stage, i.e. immediately before forming the DFB grating. (We note here that a fabrication facility for producing DFB gratings is being commissioned for this purpose at McMaster University, and will be the only such facility in a university in Canada.) The top cladding and electrical contact layers for the laser will then be grown. The integrity of the DFB grating will be then investigated using photo luminescence and x-ray and TEM structural diagnostics.

When the BGS process has been established, various laser structures will be studied to determine the laser design for producing the optimal tuning. Dr. Mascher of Project T2.6 will assist in the laser fabrication by supplying the reactive etching facilities for defining the laser waveguide and depositing the anti reflecting coatings on the laser facets.

As the ISIS program requires that the strain will be for measurements under dynamic conditions up to frequencies of 100 Hz, it will be necessary to characterize the frequency response of the lasers. Also, the laser wavelength will be calibrated as a function of input current to the split electrodes to establish if the wavelength can be accurately calibrated in terms of the current. If this proves to be the case, then it will eliminate the need for a wavelength detection system which, if successful, will simplify the demodulation system. Finally, we will develop multi electrode BGS DFB lasers with extended cavities. These devices are expected to give wider and finer tuneabilities than the two electrode lasers. The finer tuneability should enhance the possibility of using the laser as both a source and wavelength measuring system.

Back

The goal of this project is to provide the enabling technology to develop, in close collaboration with Project T2.5, an optical wavelength generation and detection module suitable for incorporation in the sensor and optoelectronic demodulation system of Project T2.3. The development of such an integrated module requires a widely and continuously tunable laser source, in the form of split electrode, DFB lasers. An even more novel and interesting approach to designing and fabricating integrated optoelectronic sensor modules is that of the integration of a DFB laser and a non absorbing DBR, where the lasing wavelength can be controlled as a function of injection current into the laser section, and by either injection current or reverse bias in the DBR section [Prosyk, K., et al., 1997].

In order to be able to grow both the DFB and the DBR from the same fundamental structure, it is necessary to selectively modify the band structure of the DFB. This can be accomplished by inducing a process called "quantum well intermixing", whereby the bandgap of a quantum well structure is modified by intermixing the wells with the barrier. Until recently, quantum well intermixing required ion implantation of the device structure, with all the problems associated with the implantation damage and the donor/acceptor characteristics of the implanted species.

A considerably more gentle and thus reliable method is impurity free disordering where a dielectric (either Si02 or SiNxOy) thin film is deposited in certain areas which is followed by rapid thermal annealing. This causes diffusion of vacancies left behind by the group III atoms that migrated into the dielectric cap, thus leading to intermixing. Preliminary experiments using a small radiofrequency based deposition system at McMaster have shown the general validity of the approach. In order to make the process as predictable and reproducible as necessary for industrial applications, however, one has to avoid as much as possible the introduction of defects at the film semiconductor interface. It is thus proposed to use the McMaster electron cyclotron resonance (ECR) chemical vapour deposition (CVD) system to deposit the required films. The main advantages of ECR CVD over conventional plasma enhanced CVD are optimized compositional control, low deposition temperatures, and low ion energies (and thus minimal plasma damage). This latter aspect is particularly important when dealing with the very sensitive (InGaAsP) materials system. For this project, process optimization will be critical and will be achieved through detailed materials and process characterization.

The two other principal objectives of this project are more closely related to conventional processing of DFB laser structures. First, we will continue to provide the facilities to fabricate ridges, facets, trenches and other features to define the laser waveguide, thus transforming the MBE grown structures, after metallization, into optically active devices. To this end we will employ ECR based reactive ion etching (RIE). The main tasks at hand will be the fabrication of high aspect ratio features which will form an essential part of the laser structures employed in the sensor arrangements and the further optimization of the RIE gas chemistries. A second processing step and also the "final touch" provided within this project is the deposition of anti reflection coatings on the laser facets. Over the past years, we have demonstrated our ability to design and fabricate multilayer stacks of SiNxOy with reflectances, R< 10-4, over a wavelength region of more than 120 nm. Real time in-situ ellipsometry will be used for process control to achieve the required precision in the deposition sequence.

Back

Implementation and monitoring of new FOS systems and measuring instruments in field applications, together with development of wireless FOS monitoring technology and software for on-site and remote monitoring of field application.

Back

Distributed fibre optic sensors are unlike Fabry-Perot and Bragg grating sensors in that they allow measurements to be made in many locations simultaneously. Brillouin scattering-based distributed sensors allow both strain and temperature measurements to be taken. The number of measurement points, their locations and their gauge length may be selected at measurement time rather than installation time, offering great flexibility. Relatively long lengths of fibre (up to tens of kilometers) can be used, allowing easy instrumentation of large civil structures in a much more complete manner than possible using other sensor technologies.

The Fibre Optics Group of the University of New Brunswick Physics Department, in association with the Civil Engineering Department, has been developing a Brillouin scattering-based distributed sensor with the intent of using it for structural monitoring applications. The Brillouin system developed to date is based on the work of Bao, et. al. Modifications to the configuration have allowed the improvement of the performance of the system. The system has been used in a structural monitoring application in which a 3 m cantilever beam was instrumented. The strain was successfully determined with an accuracy of better than 40 micro-strain using a 40 cm gauge length. By introducing compound spectrum techniques, a 25 cm spatial resolution with 40 micro-strain accuracy was achieved.

The next phase of this work is to further improve the sensor system performance with the ultimate goal of making it a practical tool for commercial use. The primary goals are:

1. To package the system in a small portable unit suitable for field use.
The existing sensor system was originally built with the intent of performing testing solely in a laboratory environment. It is constructed largely of general purpose, reusable fibre components and several stand-alone measurement instruments including two oscilloscopes and a computer. The use of these stand-alone instruments makes the entire system quite bulky and not well suited to field use.

In order to become a practical tool for field measurements, the sensor system will ultimately have to be packaged into a single compact, rugged unit. This should be possible to accomplish largely by replacing various components of the current system with smaller components that are better suited to their specific task.

2. To further improve the spatial resolution (gauge length) of the system
The spatial resolution of a distributed sensor system is the accuracy to which one can determine the position of a strain measurement. This is essentially the same as the minimum gauge length over which one can measure the average strain in the fibre. The spatial resolution of this sensor system is determined by the length of the laser pulse used to interrogate the fibre.

It has been predicted that the ultimate spatial resolution for a Brillouin scattering- based sensor was limited to approximately one metre. At optical pulse widths less than 10 ns (corresponding to a one metre spatial resolution) the Brillouin spectrum of the fibre, from which the strain is obtained, begins to change shape considerably. As the pulse length is decreased the spectrum becomes broader and the signal intensity decreases significantly, resulting in increased difficulty in determining the strain.

In the course of the experiments achieving 40cm spatial resolution, a broadened Brillouin spectrum was observed as expected. However, it was discovered that as the spatial resolution is further decreased, the theoretical predictions no longer hold. The spectrum narrows rapidly as shorter pulses are used. This discovery is very encouraging as it opens the door for constructing a sensor with spatial resolution on the order of 10 cm or possibly even shorter.

3. To develop a method for simultaneously measuring both strain and temperature.

Sensors based on Brillouin scattering are sensitive to both temperature and strain as both change the Brillouin frequency of the fibre. Conventionally, this is the only parameter measured by the sensor, therefore the effects of temperature and strain are normally indistinguishable. This means that one of the two parameters must be independently determined in order to successfully monitor the other using this type of sensor. The intensity of the Brillouin amplification signal vs. temperature and strain will be closely monitored, as will the Brillouin frequency shift vs. temperature and strain information. With both intensity and Brillouin frequency shift relationships, temperature and strain measurements can be separated.

Back

Completed Research 1995 to 2006