In the last years a variety of fiber optic Raman probes emerged, which are only partly suited for in vivo applications. The in vivo capability is often limited by the bulkiness of the probes. The size is associated with the required filtering of the probes, which is necessary due to Raman scattering inside the fibers. We employed in-line fiber Bragg gratings (FBG) as notch filter for the collection path and integrated them in a novel type of Raman probe. Multicore singlemode fibers (MCSMF) were designed and drawn integrating 19 singlemode cores to achieve better collection efficiency. A Raman probe was assembled with one excitation fiber and six MCSMF with inscribed FBGs as collection fibers. The probe was characterized regarding Raman background suppression, collection efficiency, and distance dependence. First Raman measurements on brain tissue are presented.
© 2012 OSA
Molecular diagnostics by means of Raman spectroscopy has been attracting a great deal of attention due to the possibility of non-invasive, in situ analysis of living organisms . The use of fiber based probes for optical biopsies is beneficial for both, the patient and the healthcare system as it prevents (unnecessary) resection of tissue from an (internal) organ for purely diagnostic purposes. The discrimination between normal and pathological tissues including cancer [2–5] has emerged as one of the main objectives for Raman based biomedical applications. The combination with methods like IVUS (intra-vascular ultra sound) and OCT (optical coherence tomography), which give by themselves very limited information about the chemical composition, can greatly improve non-invasive diagnostics. The use of optical fibers makes the positioning and the spectral collection much easier and more convenient compared to free beam optics. Because of the advantages of optical fibers and the still growing variety of fiber types, many different optical fiber probe configurations have been suggested and employed for different applications. The diagnosis and screening of cancer tissues by fiber optic Raman spectroscopy has recently been reviewed . Except for applications in the high-wavenumber region it is commonly accepted, that fiber optic Raman probes need special filtering in the excitation as well as in the collection path. A band pass (BP) or a short-pass (SP) filter has to be placed in the excitation path to ensure that the Raman background of the excitation fiber is suppressed. A second filter serves as a notch or long-pass (LP) filter that blocks Rayleigh scattered light to prevent the generation of Raman background in the collection fibers. The elastically scattered light of the excitation laser is several orders of magnitude more intense than the inelastically scattered Raman signals. Usually, thin film or holographic filters are used. As these filters have to be mounted in front of the excitation or the collecting fibers, respectively, careful positioning and proper fixations are needed. Subsequently the costs for micro mechanics and assembly are high. Also, the costs of high quality filters may avoid widespread application of fiber optic probes. Another issue, aggravated by the filters, is the outer diameter of such probes. In many medical applications the accessibility of the tissue of interest is restricted because of anatomy, e.g. during colonoscopy or cardiovascular endoscopy. To be able to incorporate Raman probes into standard cardiovascular catheters or the working channel of an endoscope, thin [7, 8] Raman probes have to be developed, with outer diameters of preferably 2 mm or less. However, for intravascular applications even smaller diameters are desirable. Miniaturized probes with outer diameters of less than 1 mm which may fit within the instrument channel of a standard medical endoscope have been reported , but require a complex assembly procedure . Moreover, these devices are not sufficient robust to be re-used for a large number of patients in future applications which increases the costs per patient. For safe clinical applications probes need to withstand rigorous sterilization procedures. Another big issue is the interchangeability of the probes. Accordingly running tests revealed residual probe influence on the Raman spectra , even after data filtering. Since a more rigorous filtering can also cause loss of relevant biological information, probes with more reproducible performance are needed. Chemometric algorithms, like principal component analysis (PCA), linear discrimination analysis (LDA) or cluster analysis , which are deployed for spectral decomposition and spectra identification, are quite sensitive to intensity changes.
For these reasons, alternatives of filtering are still of interest. Replacement of a standard notch filter by a fiber Bragg grating (FBG), directly inscribed into the fiber core was already suggested , as FBGs can have high rejection factors over a wavelength range of less than 0.5 nm. These components are highly effective only in singlemode fibers (SMF) with core diameters of typically 4-10 µm, which complicates efficient collection of Raman scattered light. Volume Bragg gratings written in photo-thermo-refractive glass proved to be very valuable for ultralow frequency measurement below 10 cm−1 with single monochromator stage systems . So far a few patents were written which suggest using FBGs in fiber probes . However, the low collection efficiency of strictly singlemode fibers may be one reason why these probes did not succeed on the market or were never realized.
The aim of this work is to deploy FBGs as notch filters in the collection path of fiber optic probes. FBGs represent ideal spectral filters for the above mentioned tasks. They only modify the transmission profile in a very small region, typically 0.5 to 1 nm and do not influence the transmission profile for the rest of the spectrum of interest which makes highly reproducible probes possible. Outside of the narrow band the transmission of the fiber in not effected, which is superior to normally used interference filters [7, 8]. This solution poses several advantages in comparison to the “standard” setup. Using FBG inline filtering would decrease the number of necessary parts and hence the size of the probe. Besides, the inscription process takes only several minutes per grating thus saving a considerable amount of time for probe assembly. Because of this and the relatively low production costs of optical fibers, single use fiber probes are conceivable.
In the next paragraphs FBGs as notch filters will be introduced and collection efficiency based on singlemode fibers estimated. Subsequently, an outline of the design and fabrication of multicore fibers will be given. After an initial characterization of a fiber Raman probe comprising 6 multicore collection fibers, first measurements of model substances will be presented.
1.1 Fiber Bragg gratings as notch filter
An optical fiber Bragg grating can be defined as a periodic modulation of the refractive index along a section of a fiber core [14, 15]. The FBG reflects a wavelength λB, which depends on the period ΛΒ of the structure and the effective refractive index neff of the fiber mode:16–18]. The shift ΔλB of the Bragg wavelength λB depends on the thermo-optic effect19]14].
Here, we use the FBGs not in the commonly applied reflection mode, but in transmission as notch filter. For the intended application in Raman spectroscopy, the filter efficiency or optical density (OD) should be at least OD 2, what means, less than 1% of the Rayleigh scattered light re-entering the fiber actually passes through the fiber and may increase fiber Raman background. Outside this narrow resonance, the spectrum of the transmitted light is unaltered.
1.2 Collection efficiency of singlemode fiber probes
A fiber probe consisting of six singlemode collection fibers surrounding a multimode excitation fiber was assembled. For excitation a multimode fiber was used because of the lower power density and the reduced risk of photo damaging the sample compared to a singlemode fiber. The collection fibers were prepared with a FBG located 1 cm in front of the distal ends. All fibers were embedded into a FC/PC plug and polished. The collection efficiency of the probe is very low and can be roughly estimated in relation to multimode fibers by the product of the solid angle and effective collection area ratios . The distance of the collecting area to the excitation will be neglected. The collection area ratio is proportional to d12/d22. This means, with a 100 µm ( = d1) multimode and 5 µm ( = d2) singlemode core diameter, a collection area ratio of 400:1 is achieved. Using the half-angle of collection θ calculated from the numerical aperture (NA) of the fibers the projected solid angle can be estimated usingFig. 1(a) . Then main bands of cyclohexane at 801, 1028, 1157, 1266 and 1444 cm−1 are visible with 100 s exposure time and agree with published data . Figure 1(b) shows the comparison of Raman spectra of directly illuminated SM fibers with and without an inscribed FBG. The fiber based background of the fiber without the FBG is ten times larger than the background of the fiber with the FBG, demonstrating the effective suppression of the Rayleigh scattered light.
The collection efficiency will not increase considerably when the normally applied six-around-one setup is used. Therefore, a new type of micro-structured fiber was developed which consists of a multitude of singlemode cores hexagonally arranged in one fiber. In the first attempt, 19 singlemode cores were grouped in a 125 µm outer diameter fiber. Employing our setup, FBGs can be inscribed simultaneously in all 19 cores. With this new fiber type the collection efficiency can be increased by a factor of 19 without increasing space requirements. To avoid core coupling, which may counteract filter efficiency, a proper fiber design (i.e. numerical aperture of single core and core distances) was developed.
2. Design and simulation of a MCSMF
Our design considerations will start from a single core singlemode fiber (SCSMF) with core diameter d, exhibiting a V-parameterEq. (5). Thus, a grating in a singlemode core will have a single well-defined reflection wavelength. However, the situation changes when other singlemode cores are located in the near field of that core.
Therefore, we transform the geometry into a MCSMF with N identical singlemode cores with core-to-core distance (pitch) Λ in a hexagonal arrangement like shown in Fig. 2 . Neglecting the interaction of the separated cores we find one degenerated, effective refractive index for the N fundamental modes, resulting in a single peak in the reflection spectrum of the fiber. However, the interaction of the electromagnetic fields between adjacent cores cannot be neglected for a realistic pitch Λ, which is usually in the order of some wavelengths. Applying the coupled mode theory shows that the interaction of the cores affects the propagation characteristics of the MCSMF. Without going into details, it can be found that the mutual perturbation of the guided modes will cause splitting of the initially degenerated effective refractive indices. Furthermore the phases of the fiber core modes are locked, resulting in the formation of so-called supermodes . Each of these supermodes has its own effective refractive index, which enables them to couple energy into any other propagating supermode back and forward. Consequently one Bragg grating results in numerous peaks in the reflection spectrum.
Using the generalized form of Eq. (1) for two arbitrary counter-propagating fiber modes22]. Here, k and m denote the fiber modes indices.
One can easily show that (due to the splitting of neff) the mode coupling process can yield up to M Bragg reflections, where M is the sum of all N cores:
We start our calculations with the design of fiber cores, which guide only the fundamental mode. Using Eq. (5) and inserting the upper limit for singlemode operation V* = 2.405 yieldsEq. (8) at the operating wavelength λ = 785 nm. Due to the complex structure we decided to use commercial finite element method software  to calculate neff. The index of silica n2 = 1.45358 at 785 nm was taken from literature [23, 24].
Figure 3 shows the results of the simulations. With increasing d/Λ the splitting of the effective refractive index (left scale) will increase for all four MCSMFs, resulting in a broadened reflection spectrum of the Bragg grating (right scale). However, the effect decreases with larger core diameters. The grating period Λ was calculated using Eq. (6) and is approximately 270 nm for all MCSMFs investigated here. The results show that a larger core diameter provides a narrower reflection spectrum. Typical spectral widths of FBGs written with our setup are in the range of 0.2 nm. Thus, we decided to set the limit for the spectral splitting Δλ = λ1-λ2 to 0.1 nm. The vertical blue lines in Fig. 3 show the ratios d/Λ corresponding to this splitting.
The increase of the core diameter needs lower refractive index difference Δn, corresponding to a smaller NA. This leads to an increase of bending losses, which make the fiber unfeasible. We decided that an NA≈0.10 is the lower limit for a practicable operation. The MCSMFs in our simulations exhibit an NA of 0.15, 0.12, 0.10, and 0.08 for core diameters of 4, 5, 6 and 7 microns, respectively. In summary, we found that the core diameter has to be less than 6 µm to exhibit an NA>0.10 in the singlemode operation regime. On the other hand, the simulations showed that the spectral broadening for smaller cores due to the split-up of neff is too large for high d/Λ, which are needed for high collection efficiency. Figure 4 shows the calculated phase diagram. Combinations of d and Λ resulting in the red shaded region exhibit a spectral splitting larger than 0.1 nm and are not useful for our application. Since we want to avoid high bending losses, we have to choose a core diameter with an NA of at least 0.10. Thus, we think that an optimized MCSMF geometry for the operation wavelength of 785 nm has a core diameter between 5 and 6 microns.
3. Experimental work
3.1 Preparation of a MCSMF
A MCSMF with 19 cores was fabricated by combined preform stretching and stack-and-draw technique [25, 26]. According to the simulation results, a pitch of 10 µm with a core diameter of 5.5 µm was chosen (blue point in Fig. 4). This geometry corresponds to d/Λ = 0.55 and a filling factor of 27%. The cut-off wavelength of the single core was adapted to the excitation wavelength of 785 nm. Appropriate corresponding parameters are the single core diameter of 5.5 µm and a NA = 0.11. As these fibers are specifically designed for high performance notch filters, an elevated photosensitivity is needed. High photosensitivity is generally achieved by high germanium concentrations codoped with boron and a reducing collapsing atmosphere during the preform preparation . Beyond that the boron codoping serves the adjustment of the refractive index. High germanium doping increases the refractive index, which has to be reduced again by boron codoping to ensure the singlemode behavior of the fiber at given geometrical dimensions. The initial preforms were prepared by the Modified Chemical Vapor Deposition (MCVD) process with a core composition of 6 mol% GeO2 / 7 mol% B2O3 and a resulting NA of 0.11. To meet the geometrical parameters (core diameter, pitch and package geometry) and to control the high thermal tension we approximated internal mechanical tensions, diffusion effects of the dopants during the different thermal processing steps of preforms and fiber. Figure 5(a) is a scanning electron microscope image of the fiber. The added diameter values reflect the still present core diameter variability. During the experimental test, no disturbing influences of this variability were noticed.
Our measurement showed that the refractive index difference Δn between the single cores and the cladding is 4.0∙10−3. For the calculation of the cutoff-wavelength of the first higher order mode Eq. (6) in combination with the singlemode condition V* = 2.405, resulting in a critical wavelength λc = 775 nm. Above this wavelength the cores of the MCSMF can guide only the fundamental mode (HE11). The cutoff wavelength of the drawn fiber is at a wavelength of 771 nm, which was experimentally determined (Fig. 5(b)) by analyzing the wavelength dependent bending induced attenuation.
3.2 Fiber Bragg grating inscription in MCSMF
FBGs are produced by irradiation of photosensitive fibers using an UV laser and an interferometer setup [16, 28], which makes Bragg grating inscription very flexible. Bragg wavelength , grating strength, number, and mutual distance of the FBGs can be adapted to the demands of the application. To increase photosensitivity the samples were hydrogen loaded at 200 bar for several days. FBGs were inscribed in separate fiber pieces (2m length each) of the MCSMF. The inscription process was supervised by transmission measurements via butt coupling a singlemode fiber. For the fabrication of the fiber Bragg gratings we used an interferometric setup with a Kr:F excimer laser  with a pulse power of 200 mJ. The 248 nm UV radiation is divided into two beams by a 1060 nm period phase mask. Two revolvable mirrors steer these beams to interfere in the fiber. The length of the fiber Bragg gratings is around 6 mm. When target wavelength and grating strength were reached, the inscription process was stopped, which was usually after 12 to 14 minutes. The transmission spectrum of a typical FBG in a MCSMF is shown in Fig. 6 . The pronounced transmission structure on the short-wavelength side of the Bragg peak, which is associated with radiation mode coupling, is known to appear in highly photosensitive fibers [31,32]. This feature is only observable in the transmission spectrum; viewed in reflection, only the main peak appears.
Due to the hydrogen loading  of the fibers a wavelength shift of about 0.6 nm to lower wavelength will occur after heating for hydrogen removal. This was accounted for during the FBG inscription. After the hydrogen out-diffusion (90 °C, 12 hours) the optical density of the gratings was determined with the laser of the Raman setup. To this end, light was coupled into the fiber (10 × objective, NA = 0.25, Leica) and the transmitted intensity was compared with that of a fiber of the same length but without inscribed Bragg grating. As already mentioned, the filter optical density (OD) for Raman collection fibers should be higher than 2, which means that 99% of the collected excitation wavelength will be reflected. Six fibers with an OD of 3 (0.1% transmission) for the target wavelength are used for ongoing experiments.
3.3 Probe assembly
Six fibers with inscribed FBGs of OD 3 were chosen as collection fibers to prepare a Raman probe as shown in Fig. 7 . As excitation fiber the step-index multimode fiber AFS105 (Thorlabs / Germany) with a core diameter of 105 µm and a numerical aperture of 0.22 was used. This fiber was chosen to keep the power density in contact mode as low as possible to not damage the sample. The fibers were fed through a steel capillary with an adapted inner diameter. The correct positioning of the fibers, i.e. multimode fiber in the center, was microscopically checked. The fibers were fixed inside the capillary with a cyanoacrylate adhesive (UHU endfest 300). The total outer diameter of this probe is 1.5 mm but can be reduced to 375 µm by direct fiber bonding or gluing. The excitation fiber was equipped with FC/APC, the proximal ends of the six collection fibers with individual FC/PC connectors.
3.4 Raman system
The Raman setup is depicted in Fig. 8 . A Kaiser RXN1 (Kaiser Optical Systems, USA) Raman system is used, which consists of a fiber beam collimation with a notch filter, a refocusing unit towards the 100 µm slit that is imaged to a transmitting holographic grating which is blazed for two wavelength regions for the full spectral range from 100 to 3600 cm−1 at 785 nm excitation. The grating illuminates the low and high wave number region to an Andor I-Dus CCD with 1024 × 256 pixels. Both regions are read out at the same time and clipped together.
This setup was modified with a custom made (FCC FibreCableConnect GmbH, Berlin, Germany) six fiber collection port (all fibers in a row) which was adjusted parallel to the slit and the fibers were imaged onto the CCD. As excitation laser a Toptica Xtra 785 nm frequency and temperature stabilized singlemode laser (Toptica, Germany) was used.
3.5 Probe characterization
3.5.1 Distance dependency
For the working-distance characterization of the introduced fiber probe, Raman spectra of polystyrene were recorded changing the mutual distance between sample and fiber probe stepwise (see Fig. 9(a) , ). The Raman band intensities for polystyrene (1000 cm−1), quartz (490 cm−1) and fluorescence background (950 cm−1) are plotted in Fig. 9(b).
The optimal distance between sample and fiber probe is, as expected [33,34], close to the sample at 0.6 – 0.9 mm. The generated quartz background has the same characteristic. The experiment in the next section will show if the background is generated inside the excitation fiber or in the collection fibers.
3.5.2 Probe Background and filter efficiency
To figure out which pathway generates most of the quartz background, a low pass for 800 nm (LP800) or a high pass for 800 nm (HP800) were used as reflectors for the Raman probe. With the LP800 all light with a wavelength shorter than 800 nm (corresponding to 250 cm−1) is transmitted through the filter while the remaining part will be reflected and collected by the collection fibers. Using the HP800 the background signal generated in the collection path will be detected. The results (depicted in Fig. 10(a) ) show that the quartz background is generated in the excitation pathway. Figure 10(b) compares polystyrene spectra measured with the fiber probe and the Raman microscope. It confirms that the transmitted light in the MCSMF is not altered. The more intense background in the spectrum obtained by the MCSMF probe with FBGs (called iProbe) is due to the quartz as evident from the band near 480 cm−1.
3.5.3 Tissue samples
A Raman spectrum from porcine brain tissue was collected to demonstrate the applicability of the iProbe to medical relevant samples (Fig. 11 ). Spectra were collected with 10 and 30 seconds acquisition time in contact mode with 200 mW excitation power. To keep the sample – probe distance sensitivity small, the probe was used without any focusing optics. Because of similar setup, the spatial resolution is comparable with a “standard” six-around-one probe . The spatial resolution was estimated in the following way: the core diameter of the excitation fiber is 100 µm with an NA = 0.23 which corresponds to a half angle of collection of 13.3 °. With an assumed penetration depth of 300 µm a mean value of 200 µm is obtained for the spatial resolution. This value depends on the tissue type as the penetration depth depends on the tissue type.Bands near 800 and 1060 cm−1 are assigned to quartz. The signal near 2000 cm−1 is an uncorrected cosmic spike. Bands near 1440, 1660, and 2900 cm−1 are assigned to lipids and protein in brain tissue. These spectra share many similarities with previously collected Raman spectra of brain tissue (Fig. 1 in ). The relative intensities of lipid and protein bands were found to distinguish normal brain tissue from tumors. The similarity also includes spectral contributions from quartz because the tissue was covered by a quartz slide to prevent drying. Here, quartz signals originated from the excitation fiber. It is evident from Fig. 11 that the signal intensities and signal-to-noise ratio increase if the accumulation time is extended from 10 to 30 s. It can be expected that at even longer accumulation time signal intensities and signal to noise ratio further improve and more bands become visible. Suppression of quartz bands by data processing or filters in the excitation fiber would also contribute to better spectra. This work is under progress.
4. Discussion and conclusion
We described the simulation, construction, and application of a new type of fiber, which for the first time can be named as background free Raman collection fibers. The introduced fiber is all solid which has remarkable advantages compared to fiber probes with dielectric filter assemblies for in vivo measurements. Fiber optic probes based on MCSMF with inscribed Bragg gratings as collection fibers can be used for efficient Raman signal collection. The core to core coupling can be suppressed and high reflective gratings with an OD higher than 3 are possible. The multicore structure does not show any adverse effects on the quality of the gratings.
The presented concept has high potential for miniaturized probes down to several 100 µm which are interesting for intravascular measurements and plaque detection as well as plaque identification. It is of cause preferable to increase the number of cores within the MCSMF to increase the collection efficiency. The extension to more than 50 singlemode cores in one fiber seems feasible. With such fibers acquisition times, acceptable for clinical applications, of 10 s or less would be possible.
The developed fibers can, in conjunction with inscribed Bragg gratings, be used for several applications like background free Raman collection fibers, spectral filters for fluorescence probes and other applications where a small fraction of the spectrum needs to be suppressed due to background effects generated by fibers.
One obvious challenge is the background generation in the excitation pathway. This can be solved by e.g. micro-structured fibers which show a reduced background generation [36, 37], chirped fiber Bragg gratings [38, 39] which transmit only the laser wavelength and reflect lower and higher wavelength and singlemode fibers which will generate quartz background but in much lower intensities.
Financial support of the Institute of Photonic Technology e.V. (IPHT) within the project iProbe is kindly acknowledged.
References and links
1. C. Krafft, B. Dietzek, and J. Popp, “Raman and CARS microspectroscopy of cells and tissues,” Analyst (Lond.) 134(6), 1046–1057 (2009), http://www.ncbi.nlm.nih.gov/pubmed/19475129. [CrossRef] [PubMed]
2. C. Krafft, G. Steiner, C. Beleites, and R. Salzer, “Disease recognition by infrared and Raman spectroscopy,” J Biophotonics 2(1-2), 13–28 (2009), http://www.ncbi.nlm.nih.gov/pubmed/19343682. [CrossRef] [PubMed]
3. Q. Tu and C. Chang, “Diagnostic applications of Raman spectroscopy,” Nanomed.- Nanotechnology 8, 545–558 (2011), doi:. [CrossRef]
4. T. Meyer, N. Bergner, C. Bielecki, C. Krafft, D. Akimov, B. F. M. Romeike, R. Reichart, R. Kalff, B. Dietzek, and J. Popp, “Nonlinear microscopy, infrared, and Raman microspectroscopy for brain tumor analysis,” J. Biomed. Opt. 16(2), 021113 (2011), doi:. [CrossRef] [PubMed]
5. Z. Huang, H. Zeng, I. Hamzavi, D. I. McLean, and H. Lui, “Rapid near-infrared Raman spectroscopy system for real-time in vivo skin measurements,” Opt. Lett. 26(22), 1782–1784 (2001). [CrossRef] [PubMed]
6. C. Krafft, S. Dochow, I. Latka, B. Dietzek, and J. Popp, “Diagnosis and Screening of Cancer Tissues by fiber optic probe Raman spectroscopy,” Biomed. Spectrosc. Imaging 1, 39–55 (2012), doi:. [CrossRef]
8. J. T. Motz, M. Hunter, L. H. Galindo, J. A. Gardecki, J. R. Kramer, R. R. Dasari, and M. S. Feld, “Optical fiber probe for biomedical Raman spectroscopy,” Appl. Opt. 43(3), 542–554 (2004), doi:. [CrossRef] [PubMed]
10. I. A. Boere, T. C. Bakker Schut, J. van Den Boogert, R. W. F. de Bruin, and G. J. Puppels, “Use of fibre optic probes for detection of Barrett’s epithelium in the rat oesophagus by Raman spectroscopy,” Vib. Spectrosc. 32(1), 47–55 (2003), doi:. [CrossRef]
11. P. R. Stoddart and D. J. White, “Optical fibre SERS sensors,” Anal. Bioanal. Chem. 394(7), 1761–1774 (2009), http://www.springerlink.com/content/51k1n72204m22153/. [CrossRef] [PubMed]
12. A. Glebov, O. Mokhun, V. Smirnov, L. Glebov, B. Roussel, H.-J. Reich, and F. Adar, “Novel volume Bragg grating notch filters for ultralow-frequency Raman measurements,” (2010), http://www.optigrate.com/.
13. M. J. Pelletier, “Fiber optic probe with integral optical filtering”, Kaiser Optical Systems, Inc., US 5,862,273 (1999).
14. K. O. Hill and G. Meltz, “Fiber Bragg grating technology fundamentals and overview,” J. Lightwave Technol. 15, 1263–1276 (1997), http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=618320.
15. A. D. Kersey, M. A. Davis, H. J. Patrick, M. LeBlanc, K. P. Koo, C. G. Askins, M. A. Putnam, and E. J. Friebele, “Fiber grating sensors,” J. Lightwave Technol. 15(8), 1442–1463 (1997), http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=618377. [CrossRef]
16. W. Ecke, I. Latka, R. Willsch, A. Reutlinger, and R. Graue, “Fibre optic sensor network for spacecraft health monitoring,” Meas. Sci. Technol. 12, 974–980 (2001), doi:. [CrossRef]
17. K. Schroeder, W. Ecke, J. Apitz, E. Lembke, and G. Lenschow, “A fibre Bragg grating sensor system monitors operational load in a wind turbine rotor blade,” Meas. Sci. Technol. 17(5), 1167–1172 (2006), doi:. [CrossRef]
18. I. Latka, T. Habisreuther, and M. Zeisberger, “Fiber Bragg grating based spatially resolved characterization of flux-pinning-induced strain of disk-shaped bulk YBCO samples,” Cryogenics 49(7), 340–345 (2009), doi:. [CrossRef]
19. I. Latka, W. Ecke, B. Höfer, T. Habisreuther, and R. Willsch, “Fiber optic Bragg gratings as magnetic field-insensitive strain sensors for the surveillance of cryogenic devices,” Cryogenic 49(9), 490–496 (2009), doi:. [CrossRef]
21. E. Kapon, J. Katz, and A. Yariv, “Supermode analysis of phase-locked arrays of semiconductor lasers,” Opt. Lett. 9(4), 125–127 (1984). [CrossRef]
22. T. Erdogan, “Fiber grating spectra,” J. Lightwave Technol. 15(8), 1277–1294 (1997), http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=618322. [CrossRef]
25. K. Busch, S. Lölkes, R. B. Wehrspohn, and H. Föll, Photonic Crystals: Advances in Design, Fabrication, and Characterization (Wiley-VCH Verlag GmbH & Co. KGaA, 2004), Chap. 14.
26. K. Schuster, K. Gerth, J. Kirchhof, J. Kobelke, “Verfahren zur Herstellung von strukturhomogenen mikrooptischen Fasern,“ Institut für physikalische Hochtechnologie e.V., DE102004059868B3 (2006).
27. H. Bartelt, K. Schuster, S. Unger, C. Chojetzki, M. W. Rothhardt, and I. Latka, “Single-pulse fiber Bragg gratings and specific coatings for use at elevated temperatures,” Appl. Opt. 46(17), 3417–3424 (2007). [CrossRef] [PubMed]
28. M. Becker, J. Bergmann, S. Brückner, M. Franke, E. Lindner, M. W. Rothhardt, and H. Bartelt, “Fiber Bragg grating inscription combining DUV sub-picosecond laser pulses and two-beam interferometry,” Opt. Express 16(23), 19169–19178 (2008). [CrossRef] [PubMed]
29. E. Lindner, M. Becker, M. Rothhardt, and H. Bartelt, “Generation and characterization of first order fiber Bragg gratings with Bragg wavelength in the visible spectral range,” Opt. Commun. 281(18), 4612–4615 (2008), doi:. [CrossRef]
30. M. W. Rothhardt, C. Chojetzki, and H. R. Mueller, “High-mechanical strength single-pulse draw tower gratings,” Proc. SPIE 5579, 127–135 (2004). [CrossRef]
31. V. Mizrahi and J. E. Sipe, “Optical Properties of Photosensitive Fiber Phase Gratings,” J. Lightwave Technol. 11(10), 1513–1517 (1993). [CrossRef]
32. A. Othonos, “Fiber Bragg gratings,” Rev. Sci. Instrum. 68(12), 4309–4341 (1997), http://link.aip.org/link/RSINAK/v68/i12/p4309/s1&Agg=doi. [CrossRef]
33. I. Latka, S. Dochow, C. Krafft, B. Dietzek, H. Bartelt, and J. Popp, “Development of a fiber-based Raman probe for clinical diagnostics,” Proc. SPIE 8087, 80872D, 80872D-8 (2011). [CrossRef]
34. T. F. Cooney, H. T. Skinner, and S. M. Angel, “Comparative study of some fiber optic remote Raman probe designs. Part I: Model for liquids and transparent solids,” Appl. Spectrosc. 50(7), 836–848 (1996). [CrossRef]
35. C. Krafft, S. B. Sobottka, G. Schackert, and R. Salzer, “Near infrared Raman spectroscopic mapping of native brain tissue and intracranial tumors,” Analyst (Lond.) 130(7), 1070–1077 (2005). [CrossRef] [PubMed]
36. S. O. Konorov, C. J. Addison, H. G. Schulze, R. F. B. Turner, and M. W. Blades, “Hollow-core photonic crystal fiber-optic probes for Raman spectroscopy,” Opt. Lett. 31(12), 1911–1913 (2006). [CrossRef] [PubMed]
37. K. M. Tan, G. P. Singh, C. S. Herrington, and C. T. A. Brown, “Near-infrared Raman spectroscopy using hollow-core photonic bandgap fibers,” Opt. Commun. 283(16), 3204–3206 (2010). [CrossRef]
38. M. Abtahi, A. D. Simard, S. Doucet, S. LaRochelle, and L. A. Rusch, “Characterization of a linearly chirped FBG under local temperature variations for spectral shaping applications,” J. Lightwave Technol. 29(5), 750–755 (2011). [CrossRef]
39. O. Frazão, M. Melo, P. V. S. Marques, and J. L. Santos, “Chirped Bragg grating fabricated in fused fibre taper for strain-temperature discrimination,” Meas. Sci. Technol. 16(4), 984–988 (2005), doi:. [CrossRef]