Research Article

Cite This: ACS Appl. Mater. Interfaces 2017, 9, 41284-41290 www.acsami.org

Development of Embedded Fiber-Optic Evanescent Wave Sensors

for Optical Characterization of Graphite Anodes in Lithium-Ion
Batteries
AbdulRahman Ghannoum,† Patricia Nieva,*,‡ Aiping Yu,† and Amir Khajepour‡
†
Department of Chemical Engineering, University of Waterloo, 200 University Avenue West, Waterloo, Ontario N2L 3G1, Canada
‡
Department of Mechanical and Mechatronics Engineering, University of Waterloo, Waterloo, Ontario N2L 3G1, Canada

ABSTRACT: The development, fabrication, and embedment

See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

of fiber-optic evanescent wave sensors (FOEWSs) to monitor

Downloaded via NATL UNIV OF SINGAPORE on February 22, 2023 at 08:50:01 (UTC).

the state of charge (SOC) and the state of health (SOH) of

lithium-ion batteries (LIBs) are presented. Etching of
FOEWSs is performed using a solution of 40 wt % ammonium
fluoride (NH4F) and 49 wt % hydrofluoric acid (HF) (6:1),
which is found to be superior to an etching solution containing
just 49 wt % HF. FOEWSs were characterized using glycerol
and found to have the highest sensitivity in a lithium-ion
battery when they lose 92% of their transmittance in the presence of glycerol on their sensing region. The physical effect that the
FOEWS has on the graphite anode is also investigated and is found to be much more significant in Swagelok cells compared to
that in in-house-fabricated pouch cells, mainly due to pressure variation. The FOEWS was found to be most sensitive to the
changes in the LIB when it was completely embedded using a slurry of graphite anode material within a pouch cell. The
optimized fabrication process of the embedded FOEWS demonstrates the potential of using such sensors commercially for real-
time monitoring of the SOC and SOH of LIBs while in operation.
KEYWORDS: fiber-optic sensor, evanescent waves, fiber-optic etching, fiber-optic sensor characterization, lithium-ion batteries,
graphite anode, state of charge

■ INTRODUCTION
Lithium-ion batteries (LIBs) with their relatively high energy
density have supported the development of plug-in hybrid/
electric vehicles.1 A concern of every driver is the distance they
can travel before the depletion of their energy source,
promoting research on more effective and accurate battery
monitoring systems (BMSs). The main goal of a vehicle’s BMS
is to estimate the state of charge (SOC) and state of health
(SOH) of individual batteries in a battery pack to effectively
control the flow of current to achieve a specified optimum, such
as maximum battery life. Sensors are utilized to measure
variables (e.g., voltage, temperature, and current flow) that can
be correlated to the SOC and SOH. In addition to measuring
voltage and current flow, research has extended toward
developing new sensing systems to measure internal material
property changes to improve the accuracy in estimating SOC
and SOH and allow for early detection of cell failure.2−5
A recent study observed that acoustic ultrasonic waves
traveling through an operating battery experiences a shift in
both amplitude and time of flight due to changes in the
mechanical properties of the materials within a battery.4,5
Coupling the acoustic measurements with voltage measure-
ments in a machine learning model, the error of estimating
SOC was reduced to about 1% in a controlled testing
environment with equivalent cycle conditions.5 However, this
preliminary study did not examine the effects of varying charge
rates on the acoustic measurements and further investigations
are needed to determine its viability in a commercial electric
vehicle.
The use of optical fibers to track volume/strain dynamics
within a LIB has also attracted researchers. Optical fibers with
Bragg gratings6 have been used as sensors for temperature and
strain7 because Bragg gratings reflect a narrow band of
wavelengths based on the grating period. This mechanism has
been used to monitor strain in LIBs during cycling in hopes of
also using it as a SOC sensor. However, it requires a costly
sensor interrogator to detect shifts in the wavelength.8
Optical fibers utilizing evanescent waves as a sensing
mechanism have been used in various applications9−13 due to
their fast response, selectivity, and durability.14 A recent study
on the reflectance of commercial graphite anodes and fiber-
optic evanescent wave spectroscopy of electrochemically
lithiated graphite demonstrated a correlation between lithiation
and the transmittance through a fiber-optic evanescent wave
sensor (FOEWS).15 The spectroscopy results demonstrated
significant increase in transmittance in the near-infrared band
(i.e., 750−900 nm) when graphite was lithiated electrochemi-
cally. This illustrated that changes in graphite’s optical
Received: September 5, 2017
Accepted: November 6, 2017
Published: November 7, 2017

© 2017 American Chemical Society 41284 DOI: 10.1021/acsami.7b13464

ACS Appl. Mater. Interfaces 2017, 9, 41284−41290
ACS Applied Materials & Interfaces
Figure 1. Overview of the pouch cell with an optical fiber sensor preparation process starting with (a) optical fiber etching, then (b) spot welding of
the current collectors to the electrodes, and (c) stacking the battery components (laminated aluminum, cathode, separator, anode, and second
laminated aluminum layer). (d) Final assembled pouch cell with a connectorized FOEWS.
properties during lithiation can be observed using a FOEWS.15
This discovery resulted in the simplification of the FOEWS’s
interrogation system, whereby a photodetector sensitive to
near-infrared wavelengths can be utilized instead of a costly
spectrometer to monitor a LIB.16 The lithium content of a
graphite anode is directly related to the SOC of LIB. The
transmittance intensity of a FOEWS, which is affected by the
lithium content in graphite, can hence be used to estimate a
battery’s SOC. Capacity fade in a LIB results in a decrease in
the change of graphite’s lithium content; this results in a
decrease in the amplitude of the FOEWS’s transmittance. This
implies that the amplitude of the FOEWS’s transmittance can
be correlated to a battery’s state of health (SOH).
The LIB environment can be considered a delicate but also a
chemically harsh environment containing an anode, a cathode,
a separator, and an electrolyte. Overcharging, excessive external
stress, and contamination due to poor assembly can all lead to
capacity fade by allowing lithium deposition, electrode
detachment, electrolyte decomposition (side reactions), or
separator pore size reduction.17−20 Consequently, it is essential
that the assembly of a battery is studied and optimized when
introducing a new component to the battery environment such
as a fiber-optic sensor. In this article, we report the studies for
the development of a FOEWS, its embedment into a LIB, and
its performance characterization. Different etching solutions are
studied in their ability to produce durable fiber-optic sensors.
The method for embedding the FOEWS into the LIB and its
characterization has been developed to ensure maximum
battery cell life and improved sensor sensitivity. The minimal
capacity fade observed in all studied LIBs using the embedded
FOEWS also demonstrates its commercial feasibility.
■
EXPERIMENTAL SECTION
Materials. The LIB sensor was fabricated using a step index
multimode optical fiber (AFS105/125Y, Thorlabs) with core and
cladding diameters of 105 and 125 μm, respectively. The fiber was
etched using both 49% hydrofluoric acid (HF, JT9564-6, VWR) and a
buffered oxide etchant (JT5569-3, VWR). The FOEWSs were
connectorized using universal bare fiber terminators (BFTU,
41285
Research Article
THORLABS) with connectors (B10125A, THORLABS) attached.
Glycerol (GX0185-2, EMD Millipore) was used to characterize the
FOEWS’s etched cladding prior to battery testing. Swagelok cells and
pouch/prismatic cells were fabricated using graphite and LiFePO4
electrodes (MTI Corp.) and a trilayer polypropylene−polyethylene−
polypropylene membrane (Celgard) as a separator. As an electrolyte
for the batteries, a mixture of 1:1 (vol %) ethylene carbonate (BASF)
and dimethyl carbonate (BASF) was used, with 1 M lithium
hexafluorophosphate (Purolyte, Novolyte Technologies) as a salt. A
graphite electrode slurry was prepared for embedding the sensors,
which consisted of 26% solids and 74% liquid. The solid mixture
consisted of 85% graphite (43209, Alfa Aesar), 10% poly(vinylidene
fluoride) (182702, Sigma-Aldrich), and 5% carbon black (Lib-SuperP,
MTI Corp.), whereas the liquid component contained 100% N-
methyl-2-pyrolidone. An ethylene-vinyl acetate polymer-based hot
melt adhesive (PLIB-HMA8, MTI Corp.) strip was attached to the
sensors to ensure a proper seal during pouch cell fabrication.
Sensor/Battery System Preparation. The optical fibers are
initially stripped and etched in a polypropylene beaker (1201-0250,
Nalgene), as seen in Figure 1a and described elsewhere.15,21 The
sensors are then tested using glycerol to determine the amount of
transmittance loss, as described in detail elsewhere.21 The Swagelok
assembly configuration has been presented elsewhere.15 Sensors used
in pouch cells had two hot melt adhesive strips (MTI Corp.) attached
about 3 cm from the center of the sensor to ensure a hermetic seal.
The graphite/LiFePO4 electrodes are cut into 3 cm × 2 cm rectangles,
and 1 cm × 2 cm of active material is removed to obtain square
electrodes (2 cm × 2 cm). The region with the removed electrode
material is reserved for current-collecting tabs, which are spot-welded,
as seen in Figure 1b. Aluminum tabs (PLiB-ATC4, MTI Corp.) are
used for the positive electrode (LiFePO4), and nickel tabs (PLiB-
NTA3, MTI Corp.) are used for the negative electrode (graphite). The
electrodes are then used to sandwich a separator, and the sensor is
positioned between the graphite electrode and the separator. The stack
is then sandwiched between two laminated aluminum sheets (alf-100-
210, MTI Corp.), as seen in Figure 1c. Three sides are sealed using a
heat sealer (MSK-140, MTI Corp.). The open end is used for
electrolyte insertion inside a glove box and then sealed using a
compact vacuum sealer (MSK-155A, MTI Corp.). The pouch is then
placed on a three-dimensionally printed case to be held for testing, as
seen in Figure 1d. Pouch cells with embedded sensors were prepared
by placing the sensor on the graphite electrodes; a slurry of graphite is
then poured on top to envelop the sensor and then left to air-dry in a
DOI: 10.1021/acsami.7b13464
ACS Appl. Mater. Interfaces 2017, 9, 41284−41290
ACS Applied Materials & Interfaces Research Article

Figure 2. SEM micrograph of fiber-optic evanescent wave sensors (FOEWSs) etched using (a) concentrated hydrofluoric acid (49%) and (b)
buffered hydrofluoric acid (BHF) (6:1 ratio of ammonium fluoride to hydrofluoric acid).

Figure 3. FOEWS in two LIB configurations: (a) Swagelok and (b) pouch cell. Both the illustrated cells contained three layers: 1, graphite electrode;
2, separator; and 3, lithium iron phosphate (LiFePO4), with the fiber-optic sensor positioned on the graphite electrode.

Figure 4. SEM of a graphite anode extracted from a Swagelok cell that has completed two complete cycles, demonstrating the trench formed by the
presence of a FOEWS on the graphite electrode.

fume hood. The batteries are then cycled using an eight-channel
battery analyzer (BST8-WA, MTI Corp.), and the optical signal is
recorded using a custom optical sensor interrogator with a narrow-
band light-emitting diode concentrated at 850 nm connected to a
computer, as described elsewhere.21
Figure 2a, and circular pits with diameters ranging from 3 to 25
μm were also observed.
Etching solutions consisting of 6:1 to 10:1 mixtures of 40 wt
% ammonium fluoride (NH4F) and 49 wt % HF are utilized

■
RESULTS AND DISCUSSION
Fabrication of Fiber-Optic Evanescent Wave Sensors.
Etching of optical fibers using HF concentrations ranging from
48 to 52 wt % had been performed in the past to reduce the
fiber’s diameter.22 We first examined the use of a 49 wt % HF
solution to etch the fluorine-doped silica cladding off the
sensing region of the FOEWS. However, the resultant sensor
was fragile as it broke easily during handling and battery
fabrication. In addition, the FOEWSs prepared using 49 wt %
HF were found to have significant nonuniformity on their
surface, as seen by scanning electron microscopy (SEM) in
41286
extensively to etch silica in the semiconductor industry and are
referred to as buffered HF (BHF).23 A 6:1 mixture (7.0 wt %
HF and 34.3 wt % NH4F) was tested and was found to produce
FOEWSs that can withstand the LIB fabrication process. As
seen in Figure 2b, FOEWSs prepared using BHF were observed
to have smoother surfaces. The observed difference between
the two fibers is believed to be due to the variance in the
etching mechanism. The solution with a higher concentration
of HF etches quicker23 and favors HF−2 reactions with silica.24
The second solution containing NH4F, however, had a higher
pH and favored HF reactions with silica over HF−2 during the
etching process.24 The mechanism resulting in the nonun-
iformity in the fabricated FOEWSs is beyond the scope of this
DOI: 10.1021/acsami.7b13464
ACS Appl. Mater. Interfaces 2017, 9, 41284−41290
ACS Applied Materials & Interfaces
Figure 5. SEM micrographs of the region where the FOEWS was positioned (between the dotted lines) on the graphite anode extracted from (a) an
in-house pouch cell and (b) a Swagelok cell.
article, but the results indicate that using BHF solutions
produces FOEWSs with higher durability, increasing the
throughput of the sensor/battery fabrication process.
Fiber-Optic Evanescent Wave Sensor in LIBs. Two LIB
configurations were used with the developed FOEWS and
tested: (a) a Swagelok and (b) an in-house pouch cell, as shown
in Figure 3.
The Swagelok configuration allows for repeated testing of the
same FOEWS because the cell can be disassembled and the
electrodes can be replaced. We found that with careful
handling, a sensor can be tested up to three times within one
Swagelok cell. Theoretically, the FOEWS should not break but
because the assembly takes place in a glove box, handling can
be difficult. The Swagelok cell has been reported to have a low
first-cycle efficiency (FCE) of 72%, attributed to the higher-
pressure conditions within the cell when compared to those in
other configurations such as coin cells.15 It is difficult to control
the applied pressure because it depends on the tightness of the
Swagelok nuts. However, decreasing the degree of turning
would increase the risk of contamination by air. Hence, a
second generation of Swagelok cells was fabricated by reducing
the applied pressure and keeping them in an inert environment
to avoid air contamination. With these precautions, a FCE of
83.2 ± 3.00% was achieved, which is comparable to the
reported FCE of coin cells tested with the same electrodes.15
Graphite electrodes were extracted from the Swagelok cells
after cycling, and an SEM was used to observe the structural
effect that the FOEWS has on the graphite electrode. As seen in
Figure 4, a “trench” was formed on the graphite electrode in the
region where the FOEWS was positioned, covering about 1.7%
of the entire area. The increased strain in this trench is expected
to cause a variation in the performance of the electrode in the
region below the sensor compared to that in the bulk region.20
This variation may cause the sensor signal to represent only the
local region but can be correlated to the general SOC as the
local strain is constant. For preliminary testing, the Swagelok
configuration is considered ideal but not sufficient for long-
term testing due to the observed capacity fade after multiple
cycles.
For improved pressure control and long-term testing and
packing efficiency, the pouch cell configuration shown in Figure
3b was chosen over prismatic and cylindrical cells. The
incorporation of a FOEWS requires the use of a polymer-
based hot melt adhesive around the optical fiber’s exit points
from the pouch cell to avoid air from seeping into the pouch.21
Humidity in the air forms PF3O and HF from the electrolyte’s
salt LiPF6, and HF would cause battery degradation.25 As a
41287
Research Article
battery cycles, the applied pressure on the electrodes increases
during charge and decreases during discharge due to volume
change.20 The assembled pouch cells were hence placed in
custom housings to apply an initial pressure of 4 psi, promoting
uniform current distribution and minimizing rippling of the
electrodes, resulting in lower capacity fade.20 Using this in-
house pouch cell configuration, a FCE of 87.2 ± 2.5% was
achieved, which exceeded the performance of both the coin
cells and second-generation Swagelok cells with the same
electrode material. Electrodes from the in-house pouch cells
were also examined using an SEM to determine the degree of
deformation caused by the presence of the FOEWS. As seen in
Figure 5, the graphite electrode in the pouch cells did not
undergo significant deformation as compared to that in the
Swagelok cells, which demonstrates that the Swagelok cells
were cycling at a larger applied pressure, promoting an increase
in the rate of capacity fade. However, the area of the FOEWS
interfacing with the graphite anode in the pouch cell is smaller
because the deformation and indentation occurring in the
electrode are smaller, as seen in Figure 5.
To maximize the interaction between the FOEWS and the
graphite electrode, a slurry of graphite was prepared to
completely embed the FOEWS (see Figure 6).
Figure 6. SEM micrograph of the prepared graphite slurry for
embedding the FOEWS.
Embedding the sensor allowed for easier pouch cell assembly
because the sensor was fastened to the graphite electrode. In an
industrial process, this would ensure a higher throughput in a
production line. The three tested configurations are illustrated
in Figure 7, where the sensor in the Swagelok cell sits deeper
into the graphite electrode when compared with the pouch cell,
as observed by the larger trench in Figure 5. However, the
embedded sensor is positioned on top of the original electrode
DOI: 10.1021/acsami.7b13464
ACS Appl. Mater. Interfaces 2017, 9, 41284−41290
ACS Applied Materials & Interfaces
Figure 7. Schematic illustration of the FOEWS positioning within the
graphite anode in (a) a Swagelok cell (based on Figure 5b), (b) a
pouch cell (based on Figure 5a), and (c) a pouch cell with added
slurry (Figure 6) to embed the sensor.
with a layer of added slurry to encompass the sensor’s diameter.
The results from testing the three configurations are discussed
in the next section.
Fiber-Optic Evanescent Wave Sensor Transmittance
and Glycerol Testing. The fabrication of FOEWS for LIBs
requires a validation step to ensure that the etched optical fibers
are functional and sensitive to changes within the battery. Many
sensors were fabricated and tested with 100% glycerol to ensure
that the optical fibers are not over-/underetched and to
correlate the loss in glycerol to the sensor sensitivity in a LIB.
Transmittance through an optical fiber relies on the refractive
index difference between the core of the fiber and the cladding
(i.e., the layer around the core); the core’s refractive index
needs to be higher than that of the cladding to allow total
internal reflection to occur.6 Adding a temporary layer of
glycerol on the etched sensors would indicate if the cladding
still exists on the fiber and the condition for total internal
reflection still exists because glycerol’s refractive index (i.e.,
1.4656 at 850 nm26) is higher than the refractive index of the
sensor’s core (silica) (i.e., 1.4525 at 850 nm27). If a cladding
does not exist, 100% loss in transmittance is observed when
glycerol is placed on the etched fiber, which translates to a
sensor with no signal in a battery. Etched FOEWSs with about
85 to 96% loss in glycerol were tested in LIBs. To compare the
various cells, the LIBs were cycled from 0 to 100% SOC based
on the battery’s voltage limits (i.e., 2.5−3.7 V), while recording
the optical transmittance through a FOEWS. The sensor’s
baseline (i.e., transmittance at 0% SOC) differed between the
various batteries due to variation in the contact area, optical
fiber coupling, and cladding thickness. To compare the results,
the sensors’ sensitivities were determined using expression
[(transmittance at 100% SOC − transmittance at 0% SOC)/
transmittance at 0% SOC] and were correlated to the FOEWSs’
loss in glycerol (see Figure 8).
On the basis of the results in Figure 8, the Swagelok and
pouch cells with and without embedded sensors demonstrated
the highest sensitivity when the sensor’s transmittance loss in
glycerol was about 92.0, 93.5, and 92.0%, respectively.
However, because only two pouch cells with embedded sensors
were tested, on the basis of the other two configurations, we
conclude that FOEWSs with about 92.0% loss in glycerol
achieve the highest sensitivity in a LIB. In the same figure, a
scatter was observed in the Swagelok cells’ results, which can be
41288
Research Article
Figure 8. Fiber-optic evanescent wave sensor sensitivity in relation to
the loss in 100% glycerol. The presented results include all three
configurations, Swagelok cells, pouch cells, and sensors embedded
within the graphite electrode in a pouch cell.
attributed to the variation in pressure between the cells because
the depth of the trench is dependent on the pressure applied
during assembly. The deeper the trench, the larger the
interaction between the FOEWS and the graphite electrode,
increasing the sensitivity of the sensor. This also explains the
reason why the sensors in the Swagelok cells had larger
sensitivity than that of those in the pouch cells, as seen by the
trench size difference in Figure 5. Embedding the sensor
completely within the graphite electrode as illustrated in Figure
7c within a pouch cell resulted in the highest sensitivity, which
can also be seen when comparing the full cycle transmittance of
all three FOEWS positions (Swagelok cell, pouch cell, and
pouch cell with an embedded FOEWS) (see Figure 9). The
Figure 9. Full charge and discharge transmittance for all three
configurations: Swagelok, pouch, and pouch with an embedded sensor.
The relative sensor signal change is equal to [(transmittance −
transmittance at 0% SOC)/transmittance at 0% SOC].
relative sensor signal change was determined by shifting all
transmittance measurements to start at 0 at 0% SOC and
dividing the transmittance measurements by the initial
transmittance at 0% SOC as a point of reference to permit
comparison between all three configurations.
The transmittance through the FOEWS was observed to
increase when the battery is charged in all three configurations
DOI: 10.1021/acsami.7b13464
ACS Appl. Mater. Interfaces 2017, 9, 41284−41290
ACS Applied Materials & Interfaces
following a similar trend. The slope in the charging signal
significantly decreased as the battery approached 62.5% SOC;
this may indicate the sensor’s limit of interaction with the
graphite particles, due to the sensing depth,28 or the decrease in
lithiation activity in the local region29 of the FOEWS. During
discharge, the transmittance was observed to decrease steadily
as lithium deintercalated, indicating a steady change in the
graphite particles surrounding the FOEWS during discharge
unlike charging. To investigate this further, a method to embed
multiple fibers within one graphite electrode in different
locations/depths is required for complete characterization of
the graphite anode. Nevertheless, the current observed trend in
the signal from an embedded optical fiber within a pouch cell
can be used to estimate SOC based on the optical changes
within a LIB.21 In addition to investigating multiple embedded
FOEWSs within one battery, an optical model of the FOEWS
and a model for lithium transport into graphite are currently
being coupled and fitted to the experimental results and will be
presented in future work. Further development in the
embedding process of FOEWSs will allow the use of such
sensors in cylindrical cells and will further promote their
commercial use in many more applications.
■
CONCLUSIONS
The process of fabricating fiber-optic evanescent wave sensors
and integration of the sensors into Swagelok and pouch battery
cells has been presented. A method for characterization to
achieve higher sensitivity and higher efficiency in the battery
performance has been developed. The experimental results
demonstrate that a buffered hydrofluoric solution of 40 wt %
ammonium fluoride (NH4F) and 49 wt % hydrofluoric solution
(6:1) is superior to a solution of 49 wt % hydrofluoric acid for
the cladding etching process during sensor fabrication. A
qualitative analysis of the graphite electrodes used in both
Swagelok and pouch cells demonstrates a significant reduction
in the applied pressure when using pouch cells, allowing for
battery’s with a longer life. The quantitative analysis of the first-
cycle efficiency of the two battery configurations demonstrates
the superiority of using a pouch cell with an initial applied
pressure of about 4 psi. The fiber-optic sensor was tested within
batteries and was found to have highest sensitivity when the
transmittance loss in 100% glycerol approached 92%. It was
also demonstrated that increasing the area that the FOEWS
contacts the graphite electrodes by embedding the sensor
enhanced the sensor sensitivity. Future studies will incorporate
models of the battery/sensor system to analyze the signal and
further optimize its integration into commercial lithium-ion
batteries.
■ AUTHOR INFORMATION
Corresponding Author
*E-mail: pnieva@uwaterloo.ca.
ORCID
Patricia Nieva: 0000-0002-8819-4334
Aiping Yu: 0000-0002-7422-7537
Author Contributions
All authors have contributed to the completion of the
manuscript and have given approval to the final version of
the manuscript.
Notes
The authors declare no competing financial interest.
41289
Research Article
■ ACKNOWLEDGMENTS
This work was financially supported by the “Green Intelligent
Transportation Systems (GITS)”Ontario Research Fund,
Canada, the “Next Generation Electric Vehicles: Development
of Key Technologies and Full Vehicle Testing”Natural
Sciences and Engineering Research Council of Canada
(NSERC)−Automotive Partnership Canada, and General
Motors. Special thanks to Liliana Zdravkova and Krishna Iyer
for their contribution on the etching and the characterization of
the optical fibers.
■
REFERENCES
(1) Gu, Y.-J.; Zeng, C.-S.; Wu, H.-K.; Cui, H.-Z.; Huang, X.-W.; Liu,
X.-B.; Wang, C.-L.; Yang, Z.-N.; Liu, H. Enhanced Cycling
Performance and High Energy Density of LiFePO4 Based Lithium
Ion Batteries. Mater. Lett. 2007, 61, 4700−4702.
(2) Raghavan, A.; Kiesel, P.; Sommer, L. W.; Schwartz, J.; Lochbaum,
A.; Hegyi, A.; Schuh, A.; Arakaki, K.; Saha, B.; Ganguli, A.; Kim, K. H.;
Kim, C.; Hah, H. J.; Kim, S.; Hwang, G.-O.; Chung, G.-C.; Choi, B.;
Alamgir, M. Embedded Fiber-Optic Sensing for Accurate Internal
Monitoring of Cell State in Advanced Battery Management Systems
Part 1: Cell Embedding Method and Performance. J. Power Sources
2017, 341, 466−473.
(3) Ganguli, A.; Saha, B.; Raghavan, A.; Kiesel, P.; Arakaki, K.; Schuh,
A.; Schwartz, J.; Hegyi, A.; Sommer, L. W.; Lochbaum, A.; Sahu, S.;
Alamgir, M. Embedded Fiber-Optic Sensing for Accurate Internal
Monitoring of Cell State in Advanced Battery Management Systems
Part 2: Internal Cell Signals and Utility for State Estimation. J. Power
Sources 2017, 341, 474−482.
(4) Hsieh, A. G.; Bhadra, S.; Hertzberg, B. J.; Gjeltema, P. J.; Goy, A.;
Fleischer, J. W.; Steingart, D. A. Electrochemical-Acoustic Time of
Flight: In Operando Correlation of Physical Dynamics with Battery
Charge and Health. Energy Environ. Sci. 2015, 8, 1569−1577.
(5) Davies, G.; Knehr, K. W.; Van Tassell, B.; Hodson, T.; Biswas, S.;
Hsieh, A. G.; Steingart, D. A. State of Charge and State of Health
Estimation Using Electrochemical Acoustic Time of Flight Analysis. J.
Electrochem. Soc. 2017, 164, A2746−A2755.
(6) Kawasaki, B. S.; Hill, K. O.; Johnson, D. C.; Fujii, Y. Narrow-
Band Bragg Reflectors in Optical Fibers. Opt. Lett. 1978, 3, 66−68.
(7) Kersey, A. D.; Davis, M. A.; Patrick, H. J.; LeBlanc, M.; Koo, K.
P.; Askins, C. G.; Putnam, M. A.; Friebele, E. J. Fiber Grating Sensors.
J. Lightwave Technol. 1997, 15, 1442−1463.
(8) Sommer, L. W.; Kiesel, P.; Ganguli, A.; Lochbaum, A.; Saha, B.;
Schwartz, J.; Bae, C.-J.; Alamgir, M.; Raghavan, A. Fast and Slow Ion
Diffusion Processes in Lithium Ion Pouch Cells During Cycling
Observed with Fiber Optic Strain Sensors. J. Power Sources 2015, 296,
46−52.
(9) Polynkin, P.; Polynkin, A.; Peyghambarian, N.; Mansuripur, M.
Evanescent Field-Based Optical Fiber Sensing Device for Measuring
The Refractive Index of Liquids in Microfluidic Channels. Opt. Lett.
2005, 30, 1273−1275.
(10) Nath, P. Enhanced Sensitive Fiber-Optic Sensor with Double
Pass Evanescent Field Absorption. Microwave Opt. Technol. Lett. 2009,
51, 3004−3006.
(11) Choudhury, P. K.; Yoshino, T. On The pH Response of Fiber
Optic Evanescent Field Absorption Sensor Having a U-Shaped Probe:
An Experimental Analysis. Optik 2003, 114, 13−18.
(12) Choudhury, P. K.; Yoshino, T. On The Fiber-Optic Chlorine
Sensor with Enhanced Sensitivity Based on The Study of Evanescent
Field Absorption Spectroscopy. Optik 2004, 115, 329−333.
(13) Margalit, E.; Dodiuk, H.; Kosower, E. M.; Katzir, A. Infrared
Fiber Evanescent Wave Spectroscopy for In-Situ Monitoring of Chemical
Processes, Proceedings of SPIE, Los Angeles, CA, 1989; Vol. 1048, pp
145−153.
(14) Messica, A.; Greenstein, A.; Katzir, A. Theory of Fiber-Optic,
Evanescent-Wave Spectroscopy and Sensors. Appl. Opt. 1996, 35,
2274−2284.
DOI: 10.1021/acsami.7b13464
ACS Appl. Mater. Interfaces 2017, 9, 41284−41290
ACS Applied Materials & Interfaces Research Article

(15) Ghannoum, A.; Norris, R. C.; Iyer, K.; Zdravkova, L.; Yu, A.;
Nieva, P. Optical Characterization of Commercial Lithiated Graphite
Battery Electrodes and in Situ Fiber Optic Evanescent Wave
Spectroscopy. ACS Appl. Mater. Interfaces 2016, 8, 18763−18769.
(16) Bae, C.-J.; Manandhar, A.; Kiesel, P.; Raghavan, A. Monitoring
the Strain Evolution of Lithium-Ion Battery Electrodes Using an
Optical Fiber Bragg Grating Sensor. Energy Technol. 2016, 4, 851−855.
(17) Arora, P.; White, R. E.; Doyle, M. Capacity Fade Mechanisms
and Side Reactions in Lithium-Ion Batteries. J. Electrochem. Soc. 1998,
145, 3647−3667.
(18) Peabody, C.; Arnold, C. B. The Role of Mechanically Induced
Separator Creep in Lithium-Ion Battery Capacity Fade. J. Power
Sources 2011, 196, 8147−8153.
(19) Lu, M.; Cheng, H.; Yang, Y. A Comparison of Solid Electrolyte
Interphase (SEI) on the Artificial Graphite Anode of The Aged and
Cycled Commercial Lithium Ion Cells. Electrochim. Acta 2008, 53,
3539−3546.
(20) Cannarella, J.; Arnold, C. B. Stress Evolution and Capacity Fade
in Constrained Lithium-Ion Pouch Cells. J. Power Sources 2014, 245,
745−751.
(21) Ghannoum, A.; Iyer, K.; Nieva, P.; Khajepour, A. In Fiber Optic
Monitoring of Lithium-Ion Batteries, IEEE SENSORS 2016 Proceedings,
Orlando, FL, 2016; pp 868−870.
(22) Lyons, E. R.; Lee, H. P. Demonstration of an Etched Cladding
Fiber Bragg Grating Filter with Reduced Tuning Force Requirement.
IEEE Photonics Technol. Lett. 1999, 11, 1626−1628.
(23) Spierings, G. A. C. M. Wet Chemical Etching of Silicate Glasses
in Hydrofluoric Acid Based Solutions. J. Mater. Sci. 1993, 28, 6261−
6273.
(24) Proksche, H.; Nagorsen, G.; Ross, D. The Influence of NH4F on
the Etch Rates of Undoped SiO2 in Buffered Oxide Etch. J.
Electrochem. Soc. 1992, 139, 521−524.
(25) Goodenough, J. B.; Kim, Y. Challenges for Rechargeable Li
Batteries. Chem. Mater. 2010, 22, 587−603.
(26) Rheims, J.; Köser, J.; Wriedt, T. Refractive-Index Measurements
in The Near-IR Using an Abbe Refractometer. Meas. Sci. Technol.
1997, 8, 601−605.
(27) Malitson, I. H. Interspecimen Comparison of the Refractive
Index of Fused Silica. J. Opt. Soc. Am. 1965, 55, 1205−1209.
(28) Golden, J. P.; Anderson, G. P.; Rabbany, S. Y.; Ligler, F. S. An
Evanescent Wave Biosensor. II. Fluorescent Signal Acquisition From
Tapered Fiber Optic Probes. IEEE Trans. Biomed. Eng. 1994, 41, 585−
591.
(29) Maire, P.; Evans, A.; Kaiser, H.; Scheifele, W.; Novák, P.
Colorimetric Determination of Lithium Content in Electrodes of
Lithium-Ion Batteries. J. Electrochem. Soc. 2008, 155, A862−A865.

41290 DOI: 10.1021/acsami.7b13464

ACS Appl. Mater. Interfaces 2017, 9, 41284−41290