1. * Rosalind.wynne@villanova.edu; phone 610-519-4294
Invited Paper: Fluorescence monitoring with steering wheel photonic
crystal fiber
Alpha Mansaray, Rosalind Wynne+*
Department of Mechanical Engineering, Villanova University, Villanova, PA/USA,
+
Department of Electrical and Computer Engineering, Villanova University, Villanova, PA/USA,
ABSTRACT
The development of a chemical sensor based on steering-wheel photonic crystal fiber (SW-PCF) and a
NanoSpectrometerTM
from Nano-Optics Devices, LLC can benefit industrial process-monitoring and environmental
sensing applications. This chemical sensor can potentially result in a compact, image-based sensor with enhanced spectral
resolution (~0.15nm) for applications such as environmental monitoring of water quality or quality control of
pharmaceutical production. A nanospectrometer is a planar spectrometer-on-chip that can be combined with a number of
light sources. The chip diffracts incident light to a series of wavelength dependent spatially addressed units that can be
imaged and collected with a CCD camera. It is compact in size (10 mm x 15 mm x 0.5 mm) and has a high spectral
resolution of 2x10-5
um. This study is an extension of a previous investigation of water-filled SW-PCF spectroscopy.
Instead of analyzing water samples fluorescent dyes were tested. Different types of dyes that absorbed and emitted light
in the same spectral window as the chip were identified. Spectroscopy measurements for nile blue perchlorate dye are
presented in this conference paper.
A 70 mW laser at 637nm was employed to demonstrate the fluorescence spectroscopy capability of SW-PCF enhanced
spectroscopy with a nanospectrometer. It was demonstrated that the SW-PCF is suitable for spectroscopy of dyes with a
conventional optical spectrum analyzer and a nanospectrometer. A 5 microliter sample of dye was loaded into a 14cm long
SW-PCF. The fluorescence-spectroscopic data was compared to an un-filled SW-PCF. Absorption and emission spectra
for the dye were measured near 637nm.
Keywords: Photonic crystal fiber, nanospectrometers, fluorescence spectroscopy
1. INTRODUCTION
Compact sensing systems with extended interaction lengths offer enhanced detection sensitivities. Photonic crystal fibers
(PCF) promote extended interaction lengths without sacrificing valuable sample volumes.1
A steering-wheel photonic
crystal fiber (SW-PCF) based chemical detector is presented for spectroscopic use. A 14cm length of SW-PCF was
investigated for spectroscopy applications. The fiber featured in this investigation was filled with less than 20uL of the
florescent dye, nile-blue perchlorate. The dye and light source were selected for operation in the spectral window (near
637nm) of a nanospectrometer. The absorption and emission spectra of the dye-filled PCF were not in agreement with
reference spectra. The motivation for this study is provided in section 1. Methods including SW-PCF based fluorescence
spectroscopy and digital planar holography nanospectrometer detection will be discussed in section 2. The data analysis
and conclusions of the results are provided in sections 3 and 4, respectively.
1.1 Motivation
Like pharmaceutical production, the food industry faces strict regulatory demands in terms of quality control, safety and
traceability.2
Considering the limitations of existing analytical methods (i.e. off-line laboratory analysis), such regulation
is a major challenge. The high degree of variation and parameters involved in biological processes is difficult to analyze
in real-time. Introducing real-time process monitoring capabilities with spectroscopic sensors will shift the industry -
moving from inferential monitoring and control- towards continuous measurement of core quality parameters. Multivariate
data collection through in-line and on-line analytical techniques (such as fluorescence spectroscopy with optical fiber) for
continuous learning and information collection offers the advantage of enabled control as well as optimized utilization of
raw materials. Additionally, this approach limits waste; leads to less variation in the final product quality and process cycle
times are reduced while replacing costly/slow laboratory testing. The most significant attribute of such analytical
techniques is the support of system intelligence, which promotes process and product innovation.
2. Industries dependent on regulation compliance create a market for compact, inexpensive process monitoring equipment.3
The proposed photonic crystal fiber nanospectrometer device is a scalable solution for small processing facilities to large
corporation facilities, alike. According to recent U.S. food regulation agency reports, of approximately 92,646 domestic
food-processing facilities nearly 40,235 of them do not meet proposed regulation guidelines. This creates an annual market
of $395 million dedicated to regulation and control costs. Monitoring and verification systems are a significant portion of
those costs at $82 million/year as in the food industry.4
Similarly, the inefficient use of chemical monitoring data for
decision-making in a typical pharmaceutical plant has severe impacts on production costs.5
A system featuring a nanospectrometer and a SW-PCF designed to further reduce the device size and response times of
chemical detection systems is proposed. The unique arrangement of SW-PCF makes them a good candidate for housing,
relatively, very small volumes of sample materials for spectroscopic analysis in the optical regime. This offers a unique
platform for sample housing and optical integration that is physically compatible with the digital planar holography (DPH)
integrated circuit topology.6
An overview of the fluorescence spectroscopy system technology and data analysis is
provided in following sections.
2. METHODS
Spatially resolved molecular spectroscopy is a mature technique. One of the earliest recorded descriptions involved the
use of a microscope coupled with an infrared (IR) spectrometer, collecting the first spatially resolved IR spectra.7
Since
that time, improvements in optical design and advancements in computers, software, and automation technology have
enhanced the speed, performance, and utility of spectroscopic mapping instruments. However, these instruments still
utilize motorized components and mechanical acquisition modes that limit data collection to relatively slow rates.
Additionally, the fact that moving parts are not optimal for at-line or on-line instrumentation, limits the usefulness of this
type of system for process monitoring and quality control for industrial applications.8
The featured sensing configuration
is free of mechanical motion processes providing the benefits of imaged-based spectroscopy with real-time accuracy.
A system consisting of a commercially available nanospectrometer with a specialty photonic crystal fiber is featured as a
chemical sensor that would improve the immediate real-time testing of pharmaceuticals, waterways and food processing
plants. This would, for example, greatly improve drug development, environmental monitoring and quality control of food
distribution. Such a system offers an automated passive real-time chemical analysis that requires low sample-volumes
and yields high sensitivity. Current methods for monitoring in such industries are based on optical detection methods but
these processes require large sample volumes and processing periods of hours/days/weeks for analysis.9
The featured
system configuration is illustrated in fig.2.
Fig.2. Proposed system configuration.
2.1 SW-PCF Fluorescence Spectroscopy
Fluorescent materials contain molecules that absorb energy of a specific wavelength and then re-remit energy at a different
specific wavelength. The intensity and wavelength of the emitted energy depend on both the fluorescent material and the
chemical environment of the material.10
The design of the fiber lends itself to promote increased sensitivity in an
fluorescence spectroscopy configuration due to the increased interaction length of the material with the evanescent waves.
The magnitude of fluorescent intensity is dependent on both intrinsic properties of the compound and on controlled
experimental parameters, including the intensity and energy of the absorbed light in addition to the concentration of the
fluorescent material in solution. The intensity of emitted light, F, is described by the relationship
F = ϕI0[1-exp(-ecx)], (1)
Spectra
Signal
Processing
PCF/Nano
spectro-
meter with
Sample
Quality
Control
3. where ϕ is the quantum efficiency of the fluorescent material, I0 is the incident radiant power, e is the molar absorptivity,
x is the path length of the cell, and c is the molar concentration of the fluorescent dye.11
Steady-state measurements were
made for a SW-PCF filled with the fluorescent sample, nile blue perchlorate. This fluorescent material absorbs light at
631nm and emits light 660nm. The nile blue perchlorate was diluted with sodium hydroxide (NaOH). A set of five aliquots
of 0.1 M NaOH were prepared, each 0.5mL in volume. About 5 microliters of the fluorescein stock was added to one of
these NaOH aliquots, to make 10 micromoles concentration of the dye. Forward fluorescent and transmission
characteristics were collected.
The long optical path (>cm), low transmission losses and unprecedented access to the optical field enhance the
performance of the SW-PCF sensors according to the relationship in (1). The effective radius of the microstructure
facilitates filling with low-viscosity fluids.12-17
(See fig. 3.a.) The SW-PCF provides a cross-section advantageous for
measuring optical properties of gases and fluids. The solid core has an effective diameter of 3.3 μm, while the radius of
the microstructure is 22 μm. The result is a large surface area for material to interact with the evanescent waves propagating
outside of the core.
2.2 NanoSpectrometer
The integrated circuit nature of the detection system eliminates the need for manual/mechanical spectrometer calibration.
The core technology of the nanospectrometer is digital planar holography (DPH).18-20
DPH-devices are integrated optics
components, capable of processing light, propagating inside planar optical waveguides. (See fig. 3.b.) A transfer function
is defined by the locations of millions of nano-features, embedded into the waveguide and resonantly interacting with
transmitted light. The technology utilizes spatially controlled separation and combination of light waves in planar
waveguides, enabling arbitrary discrete spectral and spatial signal arrangement and distribution. It combines the flexibility
of digital holography with well-developed microlithography mass-production techniques. The selectivity is performed in
the passive optical-domain producing a spatially-addressed spectroscopic signature that can be image-processed on a
computing device.
a) b)
Fig. 3. a) Cross-section of SW-PCF and b) Image of the DHP chip.
The Nano SpectrometerTM
, based on DPH technology, allows for managing, processing, and guiding light inside miniature
integrated optics devices. Specifically, in the NSM the DPH chip performs spectral dispersion of input light and focusing
spectral components at a linear detector array. Digitized signals are transferred via a USB link to a computer for processing
and displaying. (See figs. 4 and 5.)
4. Fig. 4. Image of the packaged Nano SpectrometerTM
is displayed.
Fig. 5. View of the internal components of the Nano SpectrometerTM
is shown.
3. DATA ANALYSIS
Early investigations of an unpackaged DPH chip for spectroscopic applications were performed.21
A custom image-
processing algorithm was developed and demonstrated, resulting in the successful detection of the DPH-chip-produced
diffraction pattern. The spatially-addressed pattern produced by the DPH chip was imaged and collected with a CCD
camera. The size of the DPH chip was 10 mm x 15 mm x 0.5mm. The collected image pattern was acquired with data
acquisition software via a desktop computer. MATLAB tools were employed to allow for the implementation of intelligent,
automatic detection of the relevant sub-patterns in the diffraction patterns and subsequent extraction of the parameters
using region-detection algorithms such as the generalized Hough algorithms which detects specific shapes within the
image.22
Fig. 6. Block diagram of an unpackaged DHP chip configuration for preliminary measurements.
The MATLAB script imported the image from the CCD camera and found the points via various image-analysis
algorithms. These raw images contained noise and other obstructions. The script automatically cropped and rotated the
image so as to focus solely on the diffraction pattern. The Hough transform was applied to the primary object in figure 7a,
which identified and outlined (circles) the diffraction pattern to demonstrate automated recognition by the algorithm as
shown in figures 7b and c. The DPH chip demonstrated enhanced resolution (~0.15nm) compared to the convention optical
spectrum analyzer (OSA) with a 0.2nm resolution. Although the chip was functional, the transmission losses related to
directing the light into the chip through the bulk-optics configuration were significant such that SW-PCF spectroscopy
measurements could not be supported. A packaged version of the DPH chip (Nano Spectrometer) was employed to address
the attenuation limitations.
5. Fig. 7 a) Captured diffraction pattern (digitally-enhanced contrast) of a fixed mounted chip with a 635nm laser
source, b) the Hough transform was applied to identify and outline the diffraction pattern(red circles) and c) a
comparison of the spectrum generated with the MATLAB processing tools for the spatially-addressed diffraction
pattern for a range of 632.2nm to 633.6nm (blue curves) and the lower resolution reference spectrum from the
conventional spectrum spectrum analyzer (green curve).
Fluorescent spectroscopic measurements from the featured chemical sensor based on steering-wheel photonic crystal fiber
(SW-PCF) and a packaged DPH chip (nanospectrometer) were performed. The feasibility of SW-PCF spectroscopy for
dyes and other low viscosity materials was explored. The experiments were performed for two distinct spectrometers. The
nanospectrometer results were compared to the conventional spectrometer results for identical experimental conditions.
For example, one study involved a 14cm long SW-PCF coupled to a 635nm laser source operating at 2.5mW and the
transmitted light was directed to a conventional optical spectrum analyzer (OSA) via a single mode optical fiber (SMF).
A 5 microliter sample of 10 micromoles concentration of nile blue perchlorate dye was loaded into the SW-PCF. (See fig.
8.) The spectroscopic data was compared to spectra of an un-filled SW-PCF. As shown in figures 9 and 10, where
absorption was measured near 637nm and 638.5nm. The transmission peaks were shifted to the right (~2nm) that can be
attributed to the refractive index of the detector material in the conventional spectrometer since it affects both the empty
and fluid filled fibers. This behavior is not characteristic of the dye absorption and emission profiles at 631nm and 660nm.
The linewidth of the transmission profile was 0.98nm. The experiment was repeated with an increased laser power of
17mW to promote absorption and emission behavior.
Fig. 8. Diagram of the sensing configuration.
A single-mode fiber-pigtailed laser diode operating at 637nm was used to excite the nile blue perchlorate dye in the
configuration described in fig. 8. The laser operated at an output power of 17mW. The attempt produced unique spectra
that were not in agreement with predicted absorption and emission values for the dye. (See fig. 11.) The spectra were
shifted to the right (~2nm), similar to the previous trial, as an artifact of the detector material. The linewidth of the
transmission profile was 2.75nm. In this case, the dye-filled SW-PCF emitted light near the 640nm and 642.5nm spectral
Fluorescent Dye
Laser
SW-PCF SMF
Conventional
Spectrometer
6. regions. (See fig. 12.) This behavior may be attributed to scattering processes and other spontaneous emission phenomena
that occur when the dye is excited with similar optical power.
Fig. 9. Transmission spectra for empty (blue dashed line) and fluorescent filled (red solid line) PCF with a
conventional OSA excited with a 635nm laser at 2.5 mW.
Fig.10. The measured absorption of the dye with peaks near 637nm and 638.5nm is shown.
0
10000
20000
30000
40000
50000
60000
70000
80000
635 637 639 641 643 645
Intensity(Counts)
Wavelength (nm)
0
10000
20000
30000
40000
50000
60000
70000
635 637 639 641 643 645
Intensity(Counts)
Wavelength (nm)
7. Fig. 11. Transmission spectra for empty (blue dashed line) and fluorescent filled (red solid line) PCF with a
conventional OSA excited with a 637nm laser at 17mW.
Fig. 12. Emission spectra with peaks at 640.4nm and 642.2nm measured from figure 11 with 17mW laser using
a conventional OSA.
A single-mode fiber-pigtailed laser diode operating at 637nm was used to excite the nile blue perchlorate dye in the
configuration employing a DPH chip Nano Spectrometer described in fig. 13. The laser operated at an output power of
17mW. The attempt produced unique spectra that were not in agreement with predicted absorption and emission values
for the dye. (See fig. 14.) The spectra were not shifted unlike the conventional spectrometer results. In this case, the dye-
filled SW-PCF emitted light near 637nm. (See fig. 12.) The linewidth of the transmission profile was 1.03nm. This
behavior may be attributed to scattering processes and other spontaneous emission phenomena that occur when the dye is
excited with similar optical power.
0
10000
20000
30000
40000
50000
60000
70000
80000
635 637 639 641 643 645
Intensity(Counts)
Wavelength (nm)
0
10000
20000
30000
40000
50000
60000
635 637 639 641 643 645
Intensity(Counts)
Wavelength (nm)
8. Fig. 13. Fluorescent spectroscopy configuration with a packaged DPH chip (Nano Spectrometer).
Fig. 14. Transmission spectra for empty(blue dashed line) and fluorescent filled (red solid line) PCF with a
nanospectrometer excited with a 637nm laser at 17mW.
4. CONCLUSION
A system featuring a SW-PCF and nanospectrometer was investigated using a fluorescent dye sample. A 5uL sample of
10 micromoles of nile blue perchlorate was housed in the SW-PCF structure. The experimental results obtained from
employing a conventional OSA with the SW-PCF configuration were compared to the reference spectra characteristic of
the nile blue perchlorate dye. The transmission spectra for this configuration produced were inconsistent with the standard
absorption and emission peaks for the dye that occur at 631nm and 660nm, respectively. It was observed that relatively
low intensity light (~2.5mW) was absorbed by the dye in the 637nm and 638nm regions. However, for increased optical
power the dye emitted light in the 640nm and 642nm regions. The experimental results collected from the configuration
featuring a nanospectrometer produced emission spectra with a modest intensity that exceeded baseline transmission
intensities near 637nm. The associated linewith was consistent with the 637nm 17mW laser source. In summary, it appears
that the nanospectrometer provided unaltered transmission characteristics in comparison to a conventional OSA with a
0.2nm resolution. It is recommended that a more suitable light source operating near the absorption window of the dye be
employed with sufficient intensity to promote fluorescence. Ultimately, with the recommended improvements these
sensors may be networked to provide feedback to monitoring agencies for environmental monitoring or food quality
control.
5. REFERENCES
[1] T. M. Monro ; Y. Ruan ; H. Ebendorff-Heidepriem ; H. Foo ; P. Hoffmann ; R. C. Moore; Antibody
immobilization within glass microstructured fibers: a route to sensitive and selective biosensors. Proc. SPIE 7004,
19th International Conference on Optical Fibre Sensors, 70046Q (May 16, 2008); doi:10.1117/12.801887.
0
5000
10000
15000
20000
25000
30000
35000
40000
45000
633 635 637 639 641 643
Intensity(Counts)
Wavelength (nm)
Fluorescent Dye
637nm
Laser
(17mW)
SW-PCF SMF
NanoSpectrometer
9. [2] Frans van den Berg, Christian B. Lyndgaard, Klavs M. Sørensen, Søren B. Engelsen, “Process Analytical
Technology in the food industry,” Trends in Food Science & Technology, Volume 31, Issue 1, May 2013, Pages
27-35, ISSN 0924-2244, http://dx.doi.org/10.1016/j.tifs.2012.04.007.
[3] Hofmann, Dietrich, et al. "Smart Instrumentation for Mobile Diagnosis and Quality Assurance in Industry,
Biology and Medicine."Biomed Tech 57 (2012): 1.
[4] Food Safety and Modernization Act (FSMA) Proposed Rule for Preventive Controls for Human Food: “Current
Good Manufacturing Practice and Hazard Analysis and Risk-Based Preventive Controls for Human Food” Docket
Number: FDA-2011-N-0920
[5] Bakeev, K. A. (2007) “Future Trends in Process Analytical Chemistry,” in Process Analytical Technology (ed
K. A. Bakeev), Blackwell Publishing Ltd, Oxford, UK. doi: 10.1002/9780470988459.ch12
[6] C. Peroz, C. Calo, A. Goltsov, S. Dhuey, A. Koshelev, P. Sasorov, I. Ivonin, S. Babin, S. Cabrini, and V. Yankov,
"Multiband wavelength demultiplexer based on digital planar holography for on-chip spectroscopy applications",
Opt. Lett. 37, 695-697 (2012).
[7] Barer, R.; Cole, A.R.H. & Thompson, H.W., “Infra-Red Spectroscopy with the Reflecting Microscope in Physics,
Chemistry and Biology;” Nature 1949, 163, 198–200.
[8] Lewis, E. N., Schoppelrei, J. W., Lee, E. and Kidder, L. H. (2007) “Near-Infrared Chemical Imaging as a Process
Analytical Tool,” in Process Analytical Technology (ed K. A. Bakeev), Blackwell Publishing Ltd, Oxford, UK.
doi: 10.1002/9780470988459.ch7
[9] Zengping Chen, David Lovett and Julian Morris, “Process analytical technologies and real time process control
a review of some spectroscopic issues and challenges,” Journal of Process Control, Volume 21, Issue 10,
December 2011, Pages 1467–1482 http://dx.doi.org/10.1016/j.jprocont.2011.06.024
[10]Sauer, M., Hofkens, J. and Enderlein, J. (2011) Front Matter, in Handbook of Fluorescence Spectroscopy and
Imaging: From Single Molecules to Ensembles, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany.
doi: 10.1002/9783527633500.fmatter
[11]George G. Guilbault. (1990) Practical Fluorescence, Second Edition, CRC Press, New York, New York.
[12]R. M. Wynne, K. Creedon, B. Barabadi, S. Vedururu, J. Merritt and A. Ortega, "Simultaneously sensing
multiple gases using a single length of hollow-core photonic bandgap fiber with sub-minute response times",
Proc. SPIE 7056, Photonic Fiber and Crystal Devices: Advances in Materials and Innovations in Device
Applications II, 70560W (August 28, 2008); doi:10.1117/12.794226
[13]J.A. Curcio and C. C. Petty, "The near infrared absorption spectrum of liquid water," J. Opt. Soc. Am. 41, 302-
302, 1951.
[14]W. Demtroder, Laser Spectroscopy: Basic Concepts and Instrumentation, 3rd ed. Berlin: Springer, 2003, pp. 370-
391.
[15]R. M. Wynne, B. Barabadi, K. Creedon, and A. Ortega, “Sub-minute Response Time of a Hollow-core Photonic
Bandgap Fiber Gas Sensor,” IEEE Journal of Lightwave Technology, Vol. 27, No. 11, June 2009.
[16]Kristian Nielsen et al “Selective filling of photonic crystal fibres,” 2005 J. Opt. A: Pure Appl. Opt. 7 L13
doi:10.1088/1464-4258/7/8/L02
[17]Y. Hoo, W. Jin, C. Shi, H. Ho, D. Wang, and S. Ruan, "Design and Modeling of a Photonic Crystal Fiber Gas
Sensor," Appl. Opt. 42, 3509-3515 (2003).
[18]C. Peroz, C. Calo, A. Goltsov, S. Dhuey, A. Koshelev, P. Sasorov, I. Ivonin, S. Babin, S. Cabrini, and V. Yankov,
"Multiband wavelength demultiplexer based on digital planar holography for on-chip spectroscopy applications",
Opt. Lett. 37, 695-697 (2012).
[19]Babin, S. et al. “Digital optical spectrometer-on-chip”, Applied Physics Lett. 95, 041105-041105-041103 (2009).
[20]Yankov, V. et al. “MC. Peroz, S. Dhuey, A. Goltsov, M. Volger, B. Harteneck, I. Ivonin, A. Bugrov, S. Cabrini,
S. Babin, V. Yankov, Digital spectrometer-on-chip fabricated by step and repeat nanoimprint lithography on pre-
spin coated films, Microelectronic Engineering, Volume 88, Issue 8, 2092-2095, August 2011
[21]M. Reimlinger; E. Battinelli; R. Wynne; “Photonic crystal fiber nanospectrometer.” Proc. SPIE 8346, Smart
Sensor Phenomena, Technology, Networks, and Systems Integration 2012, 83460T (April 26, 2012);
doi:10.1117/12.915207.
[22]D.H. Ballard, “Generalizing the Hough transform to detect arbitrary shapes,” in Pattern Recognition, vol. 13 (2),
pp. 111-122, 1981.