Integrating Fiber Fabry-Perot Cavity Sensor into 3-D ... · Heriot -Watt University . Research Gateway. Heriot-Watt University Integrating Fiber Fabry-Perot Cavity Sensor into 3-D

Heriot-Watt University Research Gateway

Heriot-Watt University

Integrating Fiber Fabry-Perot Cavity Sensor into 3-D Printed Metal Components for ExtremeHigh-Temperature Monitoring ApplicationsMathew, Jinesh; Hauser, Carl; Stoll, Philipp; Kenel, Christoph; Polyzos, Dimitrios;Havermann, Dirk; MacPherson, William Neil; Hand, Duncan Paul; Leinenbach, Christian;Spierings, Adriaan; Koenig-Urban, Kamilla; Maier, Robert R JPublished in:IEEE Sensors Journal

DOI:10.1109/JSEN.2017.2703085

Publication date:2017

Document VersionPeer reviewed version

Link to publication in Heriot-Watt University Research Gateway

Citation for published version (APA):Mathew, J., Hauser, C., Stoll, P., Kenel, C., Polyzos, D., Havermann, D., ... Maier, R. R. J. (2017). IntegratingFiber Fabry-Perot Cavity Sensor into 3-D Printed Metal Components for Extreme High-Temperature MonitoringApplications. IEEE Sensors Journal, 17(13), 4107-4114. DOI: 10.1109/JSEN.2017.2703085

http://dx.doi.org/10.1109/JSEN.2017.2703085

https://pureapps2.hw.ac.uk/portal/en/publications/integrating-fiber-fabryperot-cavity-sensor-into-3d-printed-metal-components-for-extreme-hightemperature-monitoring-applications(963c11e2-c792-4538-9d6d-3c9710b50d62).html

1530-437X (c) 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/JSEN.2017.2703085, IEEE SensorsJournal

Sensors-17040-2017 1

Abstract—This paper reports methods of embedding into 3D

printed metal components a fused silica capillary designed to

accept an in-fibre Fabry-Perot cavity based extreme high

temperature sensor. The components are manufactured in

stainless steel (SS316) by additive manufacturing using selective

laser melting (SLM). The temperature sensor consists of a

standard single mode optical fibre with the F-P sensor located at

the distal end of the fibre with the fibre being inserted into the

capillary. The capillary is either directly embedded into the

structure during the SLM build process or brazed into the

structure in between the SLM build process and the advantages

and disadvantages of these two manufacturing approaches are

discussed. Temperature sensing of up to 1000°C inside the metal

with accuracy better than ±10°C is reported. The capillary can

be directly embedded in the component which needs to be

monitored, or it can be embedded in a metal coupon which can

be attached to a component by conventional welding technology,

including the use of Laser Metal Deposition (LMD). In the case

of LMD, the sensor coupon can also be fully encapsulated by over

cladding the coupon.

Index Terms— Additive layer manufacturing, Fabry-Perot

cavity, fibre optic sensor, selective laser melting, sensor

packaging, smart metal, temperature sensor

I. INTRODUCTION

ONDITION monitoring of structural components and

machinery at elevated temperatures is a critical issue for

both economic and safety reasons. Many industrial

The work was supported by the European Union’s Seventh Framework

Program for research, technological development and demonstration project

under Grant 310279.

J. Mathew, D. Polyzos, D. Havermann, W. N. MacPherson, D. P. Hand, and R. R. J. Maier are with the Heriot Watt University, SUPA, School of

Engineering & Physical Sciences, Edinburgh, United Kingdom (e-mail:

[email protected], [email protected]). C. Hauser, is with TWI Technology Centre Yorkshire, Rotherham, S60

5TZ, UK (e-mail: [email protected]).

P. Stoll, A. Spierings, are with Inspire AG, icams, Lerchenfeldstrasse 5, St. Gallen, Switzerland (e-mail: [email protected]).

C. Kenel and C. Leinenbach, are with Empa-Swiss Federal Laboratories

for Materials Science and Technology, Überlandstrasse 129, 8600 Dübendorf, Switzerland (e-mail: [email protected]).

K. Koenig-Urban is with the General Electric Gas Power Systems,

Zentralstrasse 40, CH-5242 Birr, Switzerland (e-mail: [email protected]).

applications require sensors for real-time process condition

monitoring and asset management to ensure reliable

operations. For example, monitoring stress corrosion cracking

and spalling in high temperature steam systems, condition

monitoring of risers and wellheads of oil and gas industry,

monitoring thermal or mechanical load on blisks, blades,

vanes and heat shields of turbines, etc. The ability to

continually monitor component temperature has the prospect

of allowing an operator to run a power plant at optimum

elevated temperatures with extended maintenance intervals

whilst avoiding unexpected shutdowns through continuous

knowledge of component health. Sustained high temperature

turbine operation leads towards power plant efficiency gains

[1]. Operation at increased temperatures also enhances its

efficiency, however higher temperatures increase issues with

component life time, accelerating its wear and wear rate.

Therefore, metal embedded high temperature sensors are

gaining increasing importance in the energy industry.

Fibre optic sensors can operate safely in harsh and

hazardous environments under the presence of

electromagnetic fields, ionizing radiation and in areas with

risks of explosion due to their all optical nature.

Many of the above applications require sensors that must be

capable of sustained high temperature operation in excess of

many hundreds of degrees, and hence fused silica optical fibre

sensors are an obvious choice; the high strain point of 990-

1100°C (depending on its water content) of fused silica [2]

means that fibre optic sensors have the potential to operate at

temperatures up to, or slightly in excess of, 1000°C. Fabry-

Perot (F-P) type in-fibre optic sensors are a design concept

which is suitable for temperature measurement in this range

[3, 4]. We have shown in [3] that such F-P sensors have good

long-term stability at high temperature (above 700°C) over

extended periods of time.

Selective laser melting (SLM) is an additive manufacturing

process that can be used for many different applications [5-9].

Additive manufacturing involves building up structures layer

by layer and this opens up the prospect of incorporating

valuable internal features or components into parts during

their manufacture. Fused silica fibre can be directly

embedded in to metallic components using this process,

however the application temperature range is limited to 450°C

Integrating fibre Fabry-Perot cavity sensor in to

3D printed metal components for extreme high

temperature monitoring applications

Jinesh Mathew, Carl Hauser, Philipp Stoll, Christoph Kenel, Dimitrios Polyzos, Dirk Havermann,

William N. MacPherson, Duncan P. Hand, Christian Leinenbach, Adriaan Spierings, Kamilla

Koenig-Urban and Robert R. J. Maier

C



Sensors-17040-2017 2

due to the mismatch of the coefficient of thermal expansion

and hence fibre detachment or breakage [10-14]. In order to

circumvent this problem a fused silica capillary is embedded

into the metal component to enable the installation of high

temperature compatible F-P fibre sensors. The techniques

used for embedding the capillary are direct embedding using

SLM and a hybrid approach using brazing and SLM.

Preliminary results were reported at the conference OFS24

[15]; here we present a significantly extended study including

enhanced sensor packaging; sensor response to thermal

cycling; analysis of the measurement repeatability; and

stability at high temperature over extended periods of up to

two months. It is also shown that the sensor coupon survives

the high laser heat of Laser Metal Deposition (LMD) process

such that it is possible to weld or attach such sensor

configuration to other metallic parts by using LMD process.

In this work the temperature measurements are independent

of the induced strain from the host metallic structure, therefore

it is suitable only for temperature measurements. If strain

measurement at high temperature is a key objective, then the

solution explained in [16] can be used. Also, a combination of

both these methods can be used to monitor temperature and

strain simultaneously.

II. EXPERIMENTAL RESULTS AND DISCUSSIONS

A. F-P sensor

The sensor concept is shown schematically in Fig. 1. The

in-fibre F-P sensor is formed between two reflections along a

length of optical fibre. The first reflection point (M1) is an in-

fibre reflective splice, using a thin layer of Chromium [3].

The fibre is cleaved at a distance of about 50-100 µm from the

splice to form a low finesse optical cavity between the

reflective splice and the Fresnel reflection (M2) at the fibre

end. Details of the sensor fabrication and its temperature

response are given in [3]. The typical cavity reflection

spectrum of the sensor is shown in Fig. 2. In order to

incorporate our sensor into engineering structures, and to

avoid any strain transfer during temperature measurement, the

fibre optic sensor is freely-floating inside a fused silica

capillary.

B. Direct embedding of capillary using SLM

To manufacture the smart metallic test component, a fused

silica capillary (Z-FSS-150240, Postnova, inner and outer

diameters 150 µm and 240 µm respectively) of length ~150

mm is sealed at one end using an electric arc. Approximately

60 mm of the polyimide jacket (10 µm thick) is subsequently

removed from the sealed end by dipping it in hot (200°C)

sulphuric acid. A ~80 mm long metallic jacket is deposited on

the capillary in order that it can be ‘welded’ into place using

the SLM process without damaging the fibre. A silver (Ag)

layer was first deposited using an electro-less deposition

process [17-18]; a solution containing silver nitrate (Tollen’s

reagent) and a reducing sugar (glucose) react to form Ag,

which is deposited on the capillary. The Ag layer acts as the

anode for a subsequent nickel (Ni) electroplating process in a

Ni-sulphamate bath [10-11]. The thickness of the Ag and Ni

layers of the jacket are 4±1 µm and 140±5 µm respectively. A

Ni layer of this thickness is required to act as a thermal barrier

during the SLM embedding process.

Fig. 1. Schematic structure of the F-P cavity. M1 is partial reflective Cr

splice and M2 is cleaved end of fibre. Light is guided in the core region, and

the outer fibre diameter is typically 125 µm. The sensor is interrogated from

the lead in/out fibre on the right.

Fig. 2. Spectra recorded from the F-P cavity before (black) and after (red)

inserting it inside the SLM sample

The basic approach to embedding is to use the SLM process

to manufacture a ‘U’-shaped groove with dimensions tailored

to the diameter (~500 µm) of the metalized capillary (see Fig.

3). The capillary is placed in the groove, and the build process

of the part is continued to completion.

Fig. 3. Embedding process in three steps: (1) Building the ‘U’-shaped groove

with SLM on SS 316 substrates, (2) Nickel coated capillary placed in the

groove and (3) final encapsulation by continuing SLM process.



Sensors-17040-2017 3

Fig. 4. (a) & (c) Surface profile image of a typical U- groove measured using

3D surface profiling microscope (Alicona IFM G4). (Note: The white patches

are unresolved data from the microscope processing, and are not related to any

specific sample artefacts). (b) & (d) Typical cross section of the U-groove

measured perpendicular to and parallel to the axis of the groove respectively

using Alicona IFM G4.

The surface of the U-groove was examined using a 3D

surface profiling microscope Alicona IFM G4. The measured

surface profile of a typical U-groove, together with an

example cross section is shown in Fig. s 4(a) and 4(b)

respectively. The bottom surface of the U-Groove was

measured to have a peak to peak roughness (Rz) of 9 µm. The

edges of the SLM part are always slightly higher than the

interiors with a Δz between 40-100 µm due to surface tension

(Fig. 4. (c) and (d)). This can result in significant point like

stress on the capillary at the ingress/egress point to the

embedded structure [19] and it also prevents proper fitting of

the capillary in to the groove. To avoid such problems, the U-

groove is designed to be 200 µm wider and deeper towards the

end of the structure. A 3D printed structure containing such

an embedded sensor is shown in Fig. 5.

Fig. 5. Image of a 3D printed smart metallic structure (test component) after

removing its baseplate by milling

After the embedding process the other end of the capillary

is cleaved using a fibre cleaving tool. The cleaved end is then

heated using a small electric arc to slightly melt the sharp

edges in order to prevent them from damaging the sensor

fibre. A 125 µm diameter F-P sensor with a cavity length of

~70 µm and the fibre jacket removed over a length of 150 mm

is inserted into the capillary. A sufficient gap (10 mm) is

maintained between the sensor and the end of the capillary to

avoid back reflection coupling into the fibre. The capillary-

fibre interface is sealed using a 40 mm long fibre splice

protector to encapsulate the capillary-fibre interface. This also

provides vibration damping and an environmental seal,

protecting the fibre/capillary from damage and contaminants.

A schematic diagram of the F-P sensor inside metalized

capillary is shown in Fig. 6. The F-P spectra were identical

before and after embedding (Fig. 2) implying that no

transmission loss has been introduced during the embedding

process. It is worthwhile to mention here that the capillary to

fibre sealing used in this work should remain outside the high

temperature process environment for its stable operation. This

is possible by increasing the length of capillary as required.

Another possible solution is to use a high temperature

compatible adhesive, e.g. a mix of water glass and glass frits,

to seal the capillary to fibre interface, perhaps over coated

with metal.

The cross-sectional image of the direct embedded sensor

test component is shown in Fig. 7. Although well-attached at

the top, this approach tends to result in voids below the

capillary as shown, which could weaken the part. An

alternative approach using a brazing process was therefore

developed, as described below.



Sensors-17040-2017 4

Fig. 6. Schematic of an SLM built SS316 test component with embedded F-P sensor suitable for high temperature applications. (NB not to scale). OD-outer

diameter, ID-inner diameter, SLM-selective laser melting, SMF-standard single mode fibre.

C. Embedding capillary using brazing technique and SLM

In order to minimize any gaps around the capillary high

temperature brazing in a vacuum furnace was used. Initially a

SS316 U-groove was built using a commercial SLM machine

as explained in [20]. The U-groove made (SLM1 part) is

identical to the one described above, built using the custom

SLM system (Fig. 3). Ridges are created on top of SLM1-

part (see Fig. 8) at either side of the U-groove to prevent

brazing material from flowing out of the cavity during brazing

process. These ridges are cut down to restart the SLM2 (see

Fig. 8) process. To remove potential surface contamination

and entrapped gases, the SLM part is degassed in vacuum at

1025°C for 10 min prior to brazing.

This method was tested with a metallized capillary of

diameter 315±5 µm and a Ni-Cr-Si-B-Fe filler metal powder

(Amdry 770, Oerlikon Metco, solidus temperature 971°C,

liquidus temperature 999°C) in the form of a paste produced

by mixing the alloy powder with a cellulose nitride-based

binder [21]. The smaller metallic jacket used in this method is

because additional brazing material layer will be deposited on

top of it before the SLM process. To remove volatile solvents

and to temporarily fix the capillary in the groove the entire

setup was dried at 100°C for 1 hour in air. Finally, the

capillary was vacuum brazed to the coupon at 1000°C for 10

min. The temperature was controlled using thermocouples

placed in the heating zone next to the metal coupon. Upon

melting, the braze alloy wets the fibre Ni metallization layer as

well as the walls and bottom of the groove thereby embedding

the capillary into the material and providing a sound bond.

The top SLM layers (SLM2 in Fig. 8) of the part are also built

on top of the brazed coupon after placing the coupon back into

the SLM machine and restarting the build process. A cross

sectional analysis of the brazed sample is shown in Fig. 8.

Even though the capillary was broken for this sample, the

results shows great reduction in the gap underneath the

embedded capillary. The remaining small void may be

formed either by shrinkage upon melting of the loosely packed

powder, outward diffusion of Si and/or B into the steel coupon

or due to the incorporation of a cavity upon placement of the

capillary and paste in the first place. It is also visible that the

top part of the capillary is damaged during the second SLM

process. This might be avoided by increasing the depth of the

U-groove and bringing the capillary further down in the U-

groove such that there is sufficient thickness of brazed alloy to

overcome the thermal load of the SLM process. It is also

possible to overcome this damage by increasing the thickness

of the metallic jacket around the capillary.

Fig. 7. Cross section of coated capillary embedded in stainless steel test

component.

Fig. 8. Cross section of coated capillary brazed in to the stainless steel test

component. The capillary is damaged during the second SLM process

D. Attaching the sensor test component to functional

components using LMD process

In order to demonstrate the feasibility of attaching the

sensor configuration to a real functional component (e.g:

blisks, blades or heat shields of turbines) a laser metal

deposition (LMD) process was applied. LMD uses a laser

beam to form a melt pool on a metallic substrate, into which

powder is fed. The powder melts to form a deposit that is



Sensors-17040-2017 5

fusion bonded to the underlying substrate. The required

geometry is built up layer by layer. Both the laser and nozzle,

from which the powder is delivered, are manipulated using a

gantry system. A fibre embedded within an SLM build

coupon was bonded to the surface, or incorporated into the

bulk geometry of a component using LMD. The SLM

encapsulation of the sensor configuration gives additional

protection from the high thermal load of the LMD process.

LMD work was carried out using a Trumpf DMD505 laser

deposition system. The process parameters used are laser

power 750 W, spot size 1.2 mm, scanning velocity of 400 mm

per minute and with a standoff focal distance of 7 mm. The

feed rate of metal powder, carrier gas and argon shielding

were 1.75 g per minute, 3 liters per minute and 3 liters per

minute respectively. An image of the LMD embedded SLM

test component and a schematic of its cross section are shown

in Fig. 9. A high temperature F-P fiber sensor (see ΙΙ.A) was

inserted into the LMD embedded SLM test component. In

order to avoid the droplet like material deposited at the corners

of SLM sensor coupon to SS316 base component interface

(see Fig 9(a)) and hence to improve the quality of the welding

process an alternate welding procedure is used as shown in

Fig. 10 (a) and (b). In Fig.10 only the lower part of the two

sides of the sensor test component is welded to a turbine heat

shield element of General electric (GE) using the LMD

process. Spectra of the F-P sensor before and after integrating

inside the LMD embedded SLM test component were

identical, demonstrates that the silica capillary and the SLM

test component survived the laser heat during the LMD

process. Sensors integrated by this method were tested to

monitor in-process high temperature.

Fig. 9 (a). Image of the LMD embedded SLM test component

Fig. 9 (b). Schematic of the cross section of the LMD embedded SLM test

component (NB: not to scale)

Fig. 10 (a). Image of the LMD welded (2) SLM test component (1) on a GE

turbine heat shield (3). Only a partial image is shown not to break the GE

intellectual property rights.

Fig. 10 (b). Schematic of the cross section of red dotted line region of the

LMD welded SLM test component (NB: not to scale)

E. Temperature testing results

Thermal cycling of the sensor test component was carried

out in a tube furnace (Carbolite). The reflected spectrum was

recorded using a swept wavelength interrogator (sm125,

Micron Optics, 1510-1590nm, 5pm step, 1Hz) and analyzed in

LabVIEW software [3]. Extraction of the sensor temperature

involves multiple stages of data manipulation, including non-

linear least square fitting of a sinusoidal function to the

spectral data [3, 22]. The fundamental frequency of the

spectra and the phase shift of the sinusoidal fitting were

extracted from this data while the furnace temperature was

varied. An N-type thermocouple attached to the metallic

coupon, 2 mm away from the F-P sensor provided an

independent measurement of temperature using a PREMA

3040 precision thermometer.

Thermal cycling of the embedded sensor was carried out

with an initial cycle from ~30°C to 200°C and back to ~30°C.

After each cycle the upper temperature of the next cycle was

increased by 100°C until it reached 900°C following which

the step increment was reduced to 50°C. The heating and

cooling rates were set to 3°C/minute and 2°C/minute

respectively. The sensor was kept at each upper temperature

for approximately 20 minutes before the cooling process

started to ensure thermal equilibrium and monitor drift. The



Sensors-17040-2017 6

temperature calibration curve obtained during the cycling

process is given in Fig. 11. A second order polynomial fits

well to the calibration curve as shown in Fig. 11. The non-

linear response of the sensor is due to the nonlinearity of the

thermo-optic coefficient of fibre material SiO2 [23].

Fig. 11. Temperature response of the embedded sensor and a second order

polynomial fit to the temperature calibration curve of the sensor.

The calculated repeatability of the temperature cycling up to

950°C is ±9°C (see Fig. 12). Further studies are required to

understand the variations in run-to-run, which shows a kind of

periodic nature in Fig 12. The observed good repeatability

during several initial thermal cycles is better than our previous

observation where the sensors showed some drift. The

reduced drift here is due to the sensor sitting inside a sealed

silica capillary which isolates it from external environmental

and mechanical influences. Sensor drift was ~25°C after

cycling to 1000°C. The measured temperature sensitivity of

the sensor at 400°C is ~7.7 mrad/°C which corresponds to a

temperature resolution of 3°C for the sensor interrogation

system used with a measurement speed of 0.5 Hz, including

the signal processing time. Sensors with different cavity

lengths were manufactured and tested, and results are as

shown in Fig. 13, the similar sensitivity of the packaged

sensor in the metal compared to the sensitivity curve of free

sensors given in Fig. 13 demonstrates that the embedded

sensor is well isolated from any external strain influences.

The very smooth (typically 0.1-0.2 nm rms roughness) inside

walls of the capillary avoids any frictional force on the sensor

making it strain free. The embedded capillary, meanwhile is

under compression that prevents its damage due to CTE

mismatch during the thermal cycling process.

Detailed studies of the long-term stability of the F-P

sensors are reported in [3] demonstrating long term stability of

~10°C over 300 hours. The long-term stability of the

embedded sensor at 1000°C given in Fig. 14 shows that drift

of the embedded sensor is in accordance with the stability

results given in [3]. After 10 hours at 1000°C, the temperature

is increased to 1050°C but then the sensor starts showing some

random response because of the breakage of Ni jacket and the

capillary ~25 mm away from its SLM ingress/egress point.

Considering the stability of the Ni jacket above 1000°C, the

strain point of fused silica (around 1000°C) and melting

temperature of the Ag coating (962°C), it is worth mentioning

here that for any critical applications this sensor configuration

is not recommended for use above 950°C.

Fig. 12. Repeatability of sensor response inside the metal and during the

temperature cycling experiments. (a) For thermocouple temperature of 50°C

and (b) for temperature of 400°C. (NB: star marks of the temperature values

are overlapped by the dots of phase values)

Fig. 13. Theoretical and experimental values of sensitivity of F-P sensors as a

function of cavity length.



Sensors-17040-2017 7

The cross-sectional image of the embedded sensor test

component were taken after the temperature cycling

experiments shown in Fig. 7. The test component was subject

to over 1000°C for ~50 hours. It is clear from the picture that

the silica capillary is not damaged due to the CTE mismatch of

the host metal. Cracks in the Ni jacket are visible in Fig. 7,

which we believe is due to the Phosphorous content [14]

entered in it from the Ni-Sulphamate bath during its

deposition.

Fig. 14. Stability of the embedded sensor at 1000°C. Apparent temperature

drift given in right side of the plot is also from the sensor.

Fig. 15. Temperature response of the F-P sensor inside LMD embedded SLM

test component. (Inset) Spectra of the sensor after 4 hours at 800°C and 61

days at 800°C

The LMD embedded SLM test component with integrated

F-P sensor was also temperature tested by heating it from

room temperature (20°C) to 800°C with a heating rate of

2°C/minute. It was then maintained at this temperature for

~1500 hours (Fig. 15). The sensor showed small levels of

drift over time which was assumed to be due to the annealing

effects and dopant diffusion. The sensor spectra also showed

a small phase change (inset of Fig. 15). Additionally, a

reduction in the peak to peak amplitude of the sensor signal

was observed (inset of Fig. 15) over time. This did not affect

performance for temperature sensing because of the particular

signal analysis routine used to extract phase information of the

signal. The total temperature drift of the sensor over the 1500

hours at 800°C was approximately 50°C. This drift is mainly

because of the initial annealing of the fibre during heating,

once at sufficient time at high temperature the stability is

much better as explained in [3]. These results show the

possibility of attaching an SLM test component to any

functional surfaces of a component by LMD for real time

structural health monitoring.

III. CONCLUSIONS

In conclusion, a robust method to strain isolate and integrate

an in-fibre F-P cavity based high temperature sensor in to

SLM built metallic components is described. The sensor

consists of an optical fibre end and a reflective splice to form

the optical cavity. Silica capillary encapsulation isolates the

sensor from any external mechanical influences and prevents

fibre breakage due to CTE mismatch while sensing inside

structures. Additive manufacturing via SLM is used to embed

a metal-coated capillary into metallic structures and the sensor

is inserted into the capillary and sealed to permit in-situ

temperature monitoring at highly elevated temperatures. The

metallic jacket provides thermal shield and protects the

capillary or sensor during the SLM process. The repeatability

and then the accuracy of the embedded sensor are found to be

better than ±10°C for thermal cycling. Compared to direct

embedding the brazing method reduces the gap underneath the

embedded capillary. The results reported in this paper

demonstrate the feasibility to monitor temperature up to

1000°C inside 3D printed metallic components. The sensor

packaging technique demonstrated is useful for protecting

fibre optic sensors during the embedded sensing application in

high temperature environments. The feasibility of attaching

the packaged sensor configuration to functional metallic

components using the laser metal deposition process is also

demonstrated in this paper.

Compared to previously reported metal embedded fibre

optic temperature sensors by other authors [10-14] the smart

metals demonstrated in this paper have twice upper

temperature range of operation. The feasibility of

incorporating embedded intelligence into the components of

power generation industry exposed to very high operational

temperatures for their process optimization has also been

demonstrated by a SLM-LMD hybrid process. 3D printing is

believed to be the manufacturing technology of the future.

This paper has demonstrated methods to integrate optical fibre

sensors into 3D printed components giving them functionality.

ACKNOWLEDGMENT

The work was supported by the European Union’s Seventh

Framework Program for research, technological development

and demonstration project under Grant 310279. The authors

thank Alicona for the provision of the IFM G4 instrument at a

discounted rate. The authors also thank Niel Preece for the

LMD process. CL and CK thank H.-R. Elsener for vacuum

brazing and provision of the binder.



Sensors-17040-2017 8

REFERENCES

[1] http://www.oxigen-project.eu/

[2] N. P. Bansal and R. H. Doremus. (2013). Handbook of glass properties.

Elsevier. [Online]. Available:

http://www.sciencedirect.com/science/book/9780080523767

[3] J. Mathew, O. Schneller, D. Polyzos, D. Havermann, R. Carter, W. N. MacPherson, D. Hand, and R. R. J. Maier. (2015). In-Fibre Fabry-Perot

Cavity Sensor for High Temperature Applications. J. Lightwave

Technol. [Online]. 33(12), pp. 2419-2425. Available: https://www.osapublishing.org/jlt/abstract.cfm?uri=jlt-33-12-2419

[4] C. E. Lee, R. A. Atkins, and H. F. Taylor. (1988). Performance of a

fibre-optic temperature sensor from −200 to 1050 C. Opt. Lett., 13(11), pp. 1038-1040.

[5] S. Bremen, W. Meiners and A. Diatlov. (2012). Selective laser melting.

Laser Technik Journal. 9(2), pp. 33-38. [6] M. Agarwala, D. Bourell, J. Beaman, H. Marcus and J. Barlow. (1995).

Direct selective laser sintering of metals. Rapid Prototyping Journal.

1(1), pp. 26-36. [7] J. P. Kruth, P. Mercelis, J. Van Vaerenbergh, L. Froyen and M.

Rombouts. (2005). Binding mechanisms in selective laser sintering and

selective laser melting. Rapid prototyping journal. 11(1), pp. 26-36. [8] S. Dadbakhsh, L. Hao and N. Sewell. (2012). Effect of selective laser

melting layout on the quality of stainless steel parts. Rapid Prototyping

Journal. 18(3), pp.241-249. [9] D. Wang, Y. Liu, Y. Yang and D. Xiao. (2016). Theoretical and

experimental study on surface roughness of 316L stainless steel metal

parts obtained through selective laser melting. Rapid Prototyping Journal, 22(4), pp. 706-716.

[10] D. Havermann, J. Mathew, W. N. MacPherson, R. R. J. Maier and D.

Hand. (2015). Temperature and Strain Measurements with Fibre Bragg Gratings Embedded in Stainless Steel 316. J. Lightwave Technol.,

33(12), pp. 2474-2479.

[11] X. Li, “Embedded sensors in layered manufacturing.” PhD dissertation, Stanford University, 2001.

[12] H. Alemohammad and E. Toyserkani. (2011). Metal embedded optical

fibre sensors: Laser-based layered manufacturing procedures. J. Manuf. Sci. E., 133(3), pp. 031015.

[13] C. E. Lee, W. N. Gibler, R. A. Atkins, J. J. Alcoz and H. F. Taylor.

(1991). Metal-embedded fibre-optic Fabry–Perot sensors. Opt. Lett., 16(24), pp. 1990-1992.

[14] K. Wysokiński, T. Stańczyk, K. Gibała, T. Tenderenda, A. Ziołowicz,

M. Słowikowski, M. Broczkowska and T. Nasiłowski. (2014). New Methods of Enhancing the Thermal Durability of Silica Optical Fibres.

Materials, 7(10), pp. 6947-6964.

[15] J. Mathew, D. Havermann, D. Polyzos, W. N. MacPherson, D. Hand, and R. R. J. Maier, “SS316 structure fabricated by selective laser

melting and integrated with strain isolated optical fibre high temperature

sensor,” in Proc. SPIE 9634, 24th International Conference on Optical Fibre Sensors, (2015), pp. 96340Q.

[16] J. Mathew, R. R. J. Maier, et.al., United Kingdom patent application no.

1622357.0. [17] C. Xu, R. Zhou, H. Chen, X. Hou, G. Liu and Y. Liu. (2014). Silver-

coated glass fibres prepared by a simple electroless plating technique.

Journal of Materials Science: Materials in Electronics, 25(10), pp. 4638-4642.

[18] N. Chitvoranund, S. Jiemsirilers and D. P. Kashima. (2013). Effects of

surface treatments on adhesion of silver film on glass substrate fabricated by electroless plating. J. Aust. Ceram. Soc., 49(1), pp. 62-

69.

[19] H. K. Kang, J. W. Park, C. Y. Ryu, C. S. Hong and C. G. Kim. (2000). Development of fibre optic ingress/egress methods for smart

composite structures. Smart. Mater. Struct., 9(2), pp. 149.

[20] P. Stoll, J. Mathew, A. Spierings, T. Bauer, R. R. J. Maier and K. Wegener, “Embedding fibre optical sensors into SLM parts,” presented

at the annual international Solid Freeform Fabrication symposium, The

University of Texas in Austin, Texas, USA, Aug. 8-10, 2016. [21] C. Leinenbach, R. Transchel, K. Gorgievski, F. Kuster, H. R. Elsener

and K. Wegener. (2015). Microstructure and Mechanical Performance of Cu-Sn-Ti-Based Active Braze Alloy Containing In Situ Formed Nano-

Sized TiC Particles. Journal of Materials Engineering and Performance,

24(5), pp. 2042-2050. [22] M. Han, “Theoretical and experimental study of low-finesse extrinsic

Fabry-Perot interferometric fiber optic sensors,” Ph.D. dissertation,

´Bradley Dept. Elect. Eng., Virginia Tech, Blacksburg, VA, USA, 2006.

[23] G. M. Flockhart, R. R. Maier, J. S. Barton, W. N. MacPherson, J. D.

Jones, K. E. Chisholm, L. Zhang, I. Bennion, I. Read and P. D. Foote. (2004). Quadratic behavior of fibre Bragg grating temperature

coefficients. Appl. Opt., 43(13), pp. 2744-2751.

Documents

Integrating Fiber Fabry-Perot Cavity Sensor into 3-D ... · Heriot -Watt University . Research Gateway. Heriot-Watt University Integrating Fiber Fabry-Perot Cavity Sensor into 3-D