The feasibility of using hollow core photonic crystal fiber (HC-PCF) in conjunction with Raman spectroscopy has been explored for real time monitoring of heparin concentration in serum. Heparin is an important blood anti-coagulant whose precise monitoring and controlling in patients undergoing cardiac surgery and dialysis is of utmost importance. Our method of heparin monitoring offers a novel alternative to existing clinical procedures in terms of accuracy, response time and sample volume. The optical design configuration simply involves a 785-nm laser diode whose light is coupled into HC-PCF filled with heparin-serum mixtures. By non-selectively filling HC-PCF, a strong modal field overlap is obtained. Consequently, an enhanced Raman signal (>90 times) is obtained from various heparin-serum mixtures filled HC-PCFs compared to its bulk counterpart (cuvette). The present scheme has the potential to serve as a ‘generic biosensing tool’ for diagnosing a wide range of biological samples.
© 2011 OSA
Heparin is a blood anti-coagulant whose uncontrolled administration poses severe life threatening risks to patients undergoing surgery. It is thus imperative for physicians to rapidly monitor the amount of heparin in the blood with utmost precision. The conventional methods to quantify heparin level in surgical environment include activated clotting time (ACT), Anti-Xa, and activated partial thromboplastin time (APTT) . Among these, Activated Clotting Time (ACT) has been one of the commonly used techniques for monitoring therapeutic doses of heparin in blood. It is a measure of the anticoagulation effects of heparin obtained by determining the time for blood with heparin to clot when induced by activators. However, its slow response time and qualitative nature makes it necessary to find faster and more accurate methods [2, 3]. Such methods should not only be capable of real-time quantitative monitoring of heparin, but also consume minimal amounts of patient’s blood.
The in-lab techniques for monitoring heparin include fluorescence, surface plasmon resonance (SPR), field effect transistor, and membrane-based ion-selective electrode [4–6]. These methods involve indirect detection by using heparin probes, such as protamine or synthetic cationic polymers. Moreover, they are complicated and based on either surface affinity capture or an automated heparin protamine titration which limits the system sensitivity to detect lower concentrations of heparin in blood. Also, accuracy of such methods depends on the cross reaction of heparin with the labeling agent which may give false results. Based upon these considerations, we have combined Raman spectroscopy with a hollow core photonic crystal fiber (HC-PCF) to monitor heparin in blood-derived serum.
In the recent past, efforts have been made to employ solid core photonic crystal fibers for bio-sensing by filling sample in the cladding holes and monitoring evanescent absorption of the light guided in the core region . The disadvantage of this type of sensing is that the mode-field overlap with sample is just 1%, which results in poor sensitivity.
A number of researchers have used HC-PCF as a biosensor. This was done in two ways: the first is to fill all the holes non-selectively . However, in that case, the photonic band gap (or, the transmission wavelength) of sample filled HC-PCF shifted significantly compared to the empty HC-PCF. Under this situation, the shifted transmission wavelength of HC-PCF did not match the excitation wavelength. Therefore, the light was weakly guided into HC-PCF and was not tightly confined into the sample-filled core region, resulting in weak light-matter interaction. The second method of using HC-PCFs for characterizing sample was by filling it only in the core while sealing off the cladding holes [9, 10]. Again, the photonic band gap is not preserved in this case. Moreover, such selective filling of HC-PCF by the sample is quite cumbersome to perform. Hence, such time-consuming sample filling procedures are not useful in clinical environment which requires simple and rapid methods.
In contrast, our method preserves the photonic band gap property of HC-PCF by matching the excitation wavelength (785 nm) to that of the shifted transmission band maxima of the sample filled HC-PCF. As a result, the interaction of light with the sample is extremely strong which causes a large enhancement of the Raman signal from the analyte.
The present research builds on our previous work that dealt with the optimization of the HC-PCF-based sensing platform [11–14]. The novelty of the current work lies in extending the strong sensing potential of HC-PCF for detecting a weak Raman scatterer (like heparin) which would have been otherwise difficult to detect in a complex biological material (such as serum). It is achieved by exploiting the photonic band gap property of HC-PCF along with efficient collection of the Raman signal. This is the first demonstration of monitoring a clinically important molecule within serum by a HC-PCF-based Raman sensor.
The high sensitivity, accuracy along with a minimal test sample (micro-litre) requirement makes the present optical sensing scheme an attractive candidate for heparin monitoring in clinical setting. Furthermore, Raman spectroscopy offers the inherent advantage of molecular specificity which makes it more relevant in the context of clinical diagnosis where sample screening is performed on bodily fluids (blood, serum) characterized with multi-species biological components.
The organization of the paper is as follows: we begin with the theory and criteria involved in selection of HC-PCF and optical characterization of sample mixtures (heparin-serum). It is followed by a layout of the experimental configuration and description of the adopted methodology. The spectral results and analysis are presented in the last section along with challenges encountered during the performances of the various tasks involved.
The light guiding property of a non-selectively filled HC-PCF changes depending on the refractive index of the filled sample. In this case, the guiding principle is still due to the bandgap effect, but the transmission band supported by the fiber is shifted. The shift in the transmission wavelength of HC-PCF can be determined from the equation given by Russell et al. , as follows:
2.2 Samples and their refractive index measurement
MiniHep was purchased from Leo Pharmaceuticals in a vial of 25000 USP/ml concentration. USP stands for United States Pharmacopeia (USP) and describes the potency of the drug in clinical applications. Blood samples were obtained from a local bovine slaughterhouse. Serum was extracted by centrifuging blood at a speed of 2000 rpm for 30 minutes. Different concentrations of heparin in serum with a total volume of 2 ml were accurately prepared.
As mentioned in Sec 2.1, the light guiding property of the HC-PCF depends on the refractive index, which in turn is related to the concentration of the sample mixture. The in-lab refractive index measurements of heparin-serum mixtures needed to be found experimentally, because the refractive index values were only available for individual constituents (heparin and serum) and not their mixture with varying volume ratios. The refractive index values in the second column of Table 1 were determined from a Reichert AR200 Digital Handheld Refractometer that measured the refractive index of heparin-serum mixtures at the wavelength of 588 nm. Ideally, the refractive index values of the sample (s) filled into HC-PCF holes must be measured at the wavelength of excitation light (785-nm in our case) for calculating the transmission band of HC-PCF. However, the choice of refractometer (as described above) was based on its availability. Moreover, the sample refractive index values measured at the wavelength of 588-nm was believed to be comparable to the one’s measured at 785-nm. As an example, the refractive index of one of the sample mixture (with heparin concentration in serum ~0.5 [vol./vol.]) was found to be 1.3509.
2.3 Simulation of HC-1550-04 with COMSOL
In order to confirm that HC-1550-04 would efficiently guide 785-nm light when filled with heparin-serum mixture, we simulated the fiber geometry using finite element method software (COMSOL Multiphysics V3.3). Firstly, the fundamental mode for the empty HC-1550 at 1535-nm excitation wavelength was determined. In the next step, the refractive index parameter was altered to the value that corresponded to the refractive index of heparin-serum mixture. The simulated mode profile suggested that the fiber would guide 785-nm excitation on being filled with heparin-serum mixture as shown in (Fig. 1 ). This was later confirmed by launching 785-nm laser light and observing the far-field pattern. The single mode pattern of the filled fiber HC-1550 is necessary to obtain maximum overlap of sample and the excitation light resulting in strong Raman scattering from the sample.
2.4 Fiber characteristics
Based on Eq. (1), we found that the HC-1550-04 from NKT Photonics would support light transmission through our sample at 785-nm excitation wavelength. The center wavelength of this fiber was 1535 nm with a numerical aperture (NA) of 0.2. The fiber had a core size of 10.6 μm with a pitch of 4.1 μm and a mode field diameter of 7.5 μm. As a preliminary step, we experimentally verified the bandgap shift property of the fiber by coupling broadband light into the HC-PCF. The broadband spectrum ranging from 400to 1700 nm was recorded from a LS-1 tungsten halogen lamp through an optical spectrum analyzer (OSA). This was in turn launched into the HC-PCF in two ways: empty and filled with heparin serum mixture. The output spectrum from the fiber in both cases was recorded as shown in Fig. 2 . The empty fiber guided wavelength in the IR range. While the fiber is filled with the sample mixture, the fiber guided the red/near-IR part of the input spectrum. It clearly proved that the photonic bandgap wavelength shifted after filling the HC-PCF with the sample solution as described in section 2.1. There can be a situation where the refractive index of the sample is such that the transmission maximum of the sample-filled HC-PCF may not match the excitation wavelength. However, in that case, the role of the HC-PCF would be restricted to merely a micro-litre sample container, and light would barely be confined to the core region due to the loss of the photonic band gap. Consequently, light would be partially guided within the core region and weak light-sample interaction would generate a low Raman signal.
2.5 Experimental configuration
A schematic of the experimental setup is shown in Fig. 3 . A single mode 100-mW, 14-pin butterfly pin lasers was purchased from Innovative Photonic Solutions (I0785SB0100B) and was coupled into the HC-PCF. The laser beam was passed through a bandpass filter (BP) centered at 785-nm (+/− 2nm) to filter out other wavelength components around 785-nm from the laser diode. Then, it was directed through a dichroic filter (R785RDC, Chroma Technologies Corp.) which reflected 785-nm (+/−5nm) at an angle of 45° and transmitted 790-1000 nm band. The dichroic filter acted as a reflector for the laser beam which was further focused onto the tip of the HC-PCF by a 40x microscopic objective lens (L1) with numerical aperture (N.A) as ~0.65. The light coupling efficiency for heparin-serum filled HC-PCF was ~30%. The channels of the fiber were filled through a reservoir as explained in section 2.6. Furthermore, the dichroic filter acted as a high pass filter for the light scattered backward from the sample-filled HC-PCF, thus allowing only the Raman wavelength to pass through it. The filtered Raman light was then imaged onto a fiber bundle (Fiberoptic System Inc., 26 multimode fiber, NA= 0.22) by another 6.3x microscopic objective lens (L2) with N.A as ~0.20. The output of the fiber bundle was interfaced into a Kaiser f/18i Spectrograph with a TE-cooled Andor CCD camera. Andor SOLIS software was used for spectral data acquisition and spectra were monitored on the data acquisition computer. The self-configured optical configuration, as described above, was well optimized for achieving high signal-to-noise Raman signal from heparin.
2.6 Reservoir and HC-PCF filling
The high light power density within the HC-PCF causes evaporation of the filled sample. It results in the formation of air gaps or the occurrence of non-uniform sample distribution within the HC-PCF channels. Under such circumstances, light is partially guided within HC-PCF leading to an overall decline in the Raman signal of the chemical of interest. Therefore, an aluminum-based, in-house sample reservoir was designed that could continuously replenish heparin-serum mixture sample into the HC-PCF channels as shown in Fig. 4 . It can be easily mounted on a XYZ translation stage (Thorlabs Inc.). One end of the reservoir was sealed with a glass slide using water resistant epoxy while the other end was used for HC-PCF fiber insertion. The fiber was fixed in place with a Teflon tube and Swagelok connectors.
The process of manually filling the HC-PCF with heparin-serum sample solution was as follows: The first step involved the collection of sample solution in a clinical syringe that was subsequently connected with the reservoir inlet via a Teflon tube (length~6 cm). The sample solution was transferred from syringe to the reservoir till it was filled to capacity. In the second step, the syringe piston was gradually pushed further, which created a difference in the pressure between the two ends of the HC-PCF for complete filling of HC-PCF. The light guiding end of HC-PCF was also examined under the microscope to cross-validate the complete loading of the HC-PCF with the sample. It’s worth mentioning here that in few iterations of the filling procedure, an optimal pressure on syringe piston was qualitatively ascertained to prevent the sample solution from leaking out of the light guiding end of the HC-PCF. Thus, it was ensured that light must couple to the sample filled HC-PCF efficiently without being affected by the formation of sample droplet at the light guiding end of the HC-PCF.
3. Results and discussion
3.1 System optimization
The objective of our study was to monitor the concentration of heparin in serum using HC-PCF and Raman spectroscopy. The initial phase of the study focused on recording the Raman spectrum of pure heparin in a cuvette with near IR-785-nm light as the excitation source. The Raman spectrum of heparin was quite similar to the one reported previously by Atha et al. . The choice of 785-nm for excitation was obvious owing to the minimal absorption of blood/serum in the NIR wavelength region. The Raman peak of heparin around 1005 cm−1 was considered suitable for quantifying heparin in serum. The sample mixtures were then prepared with different concentration ratios of heparin and serum. It was soon realized that Raman spectra of heparin-serum mixtures exhibited Raman peaks of heparin with poor signal-to-noise ratios. This is because the spectral background of serum completely dominated the Raman signal of heparin. The next challenge was to enhance the Raman signal of heparin and filter out the spectral background of serum to improve the detection sensitivity.
To evaluate the potential of HC-PCF as a ‘Raman signal enhancer’, it was first filled with pure heparin by capillary action and carefully monitored under the microscope. Figure 5(a) and 5(b) present the microscopic view of the empty and heparin filled HC-PCF. The brighter patch in the cross sectional view of the HC-PCF indicates that almost all the micro-channels are filled with heparin whereas the few dark spots indicate unfilled or partially filled channels as shown in Fig. 5(b). The fiber was then carefully inserted inside a reservoir as shown in Fig. 4 (b). A reservoir at one of the fiber ends was essential to overcome the problem related to evaporation of samples inside the fiber channels, which consequently affects the light guidance within the fiber. The fiber was then mounted on a 5-axis flexure stage (Thorlabs Inc.). The experimental set-up was optimized with respect to maximum light coupling efficiency (~30%) within the HC-PCF
3.2 Polynomial background subtraction
The Raman spectra of serum and heparin mixtures exhibited strong spectral background mainly due to serum. The spectral background was fitted to a third-order polynomial and subtracted from the spectra as shown in Fig. 6 (a) and Fig. 6 (b). For post processing data, we selected the range 900-1150 cm−1 as it contained prominent Raman peaks of heparin that can be used for quantifying heparin in serum. Similar data processing was applied to filter the Raman peak of heparin from the spectral data of heparin-serum mixtures in subsequent steps.
3.3 Raman signal enhancement of heparin-serum mixture within HC-PCF
We recorded Raman spectra of sample mixture of heparin-serum (1:1) filled into different lengths of HC-PCF. The major challenge associated with this task was to maintain the same power of light and to ensure homogeneous distribution of sample mixture within different lengths of HC-PCF. Figure 7(a) shows a gradual increase in the intensity of the Raman peaks of heparin (within the wave number range of 910 to 1120 cm−1) as the length of HC-PCF was increased from 3 cm to 10 cm at a regular interval of 1cm. For the sake of clarity, only three Raman spectrums at length 3, 6 and 9 cm are shown in Fig. 7 (a). The Raman peak intensity of heparin around 1005 cm−1 was calculated for different lengths of HC-PCF. Figure 7(a) also shows the Raman signal of heparin-serum (1:1) in bulk (cuvette), which is weak in comparison to that obtained from sample-filled HC-PCF. In conventional cuvette-based Raman spectroscopy, the maximum Raman light is generated from the focal point of the beam within the sample. The focal point can also be considered as a spot where laser light strongly overlaps/confines with the sample. A similar condition of strong light-sample interaction is achievable all along the entire length of the sample-filled HC-PCFs due to their ability to strongly confine light via the photonic bandgap. Hence, samples when filled into HC-PCFs give several orders of magnitude higher Raman signal compared to when in a cuvette. In other words, HC-PCFs can act as a ‘Raman signal enhancer’. An enhancement factor representing the ratio of Raman signal obtained from HC-PCF to that from a cuvette was calculated for different lengths of HC-PCF. Finally, a calibration curve between the enhancement factors versus HC-PCF length was drawn as shown in Fig. 7(b). It’s worth mentioning here that this comparison takes into account similar throughput sample power (~5 mW) while keeping other experimental parameters same.
Based on these results, one can note the following: The enhancement factor is directly proportional to the sample volume in the core region, which is ultimately related to the HC-PCF length. The enhancement factor linearly increases with the HC-PCF length in the range of 3 to 8 cm, and tends to saturate for higher values of HC-PCF length. There are a number of reasons for saturation of the Raman signal of heparin (or enhancement factor) beyond a certain length of HC-PCF, i.e., 8 cm. This includes light absorption as well as viscosity, surface tension from the sample, and fiber bend loss. Thus, the optimal range of HC-PCF length for heparin detection is between 5 to 8 cm. Figure 7(b) clearly indicates an enhancement of ~90 times the Raman signal of heparin at ~1005 cm−1 from the mixture filled within the HC-PCF with respect to that obtained from the cuvette. This demonstrates that HC-PCFs have an inherent potential to enhance the Raman signal and their length can be suitably optimized for sensitive detection of bio-analytes.
3.4 Raman spectra of heparin-serum mixtures
The present study was primarily focused on quantitative measurement of heparin in sample mixtures of heparin and serum. Various sets of heparin-serum mixtures with different percentage compositions of heparin (within the range of 12-20000 USP/ml) were prepared. These were subsequently filled into a 5 cm long HC-PCF. This length was picked so we can obtain significant enhancement and avoid saturation effects. The Raman spectra of heparin-serum mixtures indicate prominent Raman peaks within the spectral range of 900-1070 cm−1 as shown in Fig. 7 (a). These were well correlated to the Raman bands of commercial-grade heparin as assigned by Atha et al. . The clinical-grade heparin that is employed in this study exhibited a well resolved Raman peak at ~1005 cm−1 (C-N stretching). The overlapped Raman bands due to the symmetric SO3 vibration were located at ~1035 cm−1 (N-SO3 vibration), 1045 cm−1 (6-O-SO3 vibration), and 1060 cm−1 (2-O-SO3 vibration). Therefore, the intensity of the Raman peak at ~1005 cm−1 was considered as a parameter to evaluate the quantity of heparin in serum. It varies with increasing concentration of heparin as shown in the inset of Fig. 8 (a) . The inset shows the Raman peak at 1005 cm−1 for lower heparin USP/ml concentrations. The spectral band intensity of the Raman peak of heparin at ~1005 cm−1 was calculated. This calculation was performed for all the sets of spectra corresponding to a particular concentration ratio of heparin and serum and was averaged to obtain a single data point. As a result, error due to fluctuations of the Raman peak intensity was minimized. Finally, a calibration curve was drawn between the averaged spectral band intensity of the Raman peak as a function of the corresponding percentage composition of heparin, as shown in Fig. 8 (b).
The following conclusions have been drawn from the above experimental results. First, the HC-PCF based sensor system is capable of successfully monitoring the heparin concentration in a heparin-serum mixture. Second, the Raman peak intensity of heparin varies linearly with increasing concentration of heparin. This implies that the sensor has high detection sensitivity in a broad concentration range of heparin and can measure heparin concentration as low as 12USP/ml in a serum sample mixture. The HC-PCF sensor detection efficiency was also calculated in terms of limit of detection (L.O.D) from the standard equation of L.O.D expressed as [3*N/S], where N is the standard deviation of the spectral background noise, and S is the slope of the calibration curve. The limit of detection (L.O.D) of the sensor for heparin based on the data collected was calculated as ~10 USP/ml. Also, intensity data points fluctuated little as reflected by the error bars on the calibration curve in Fig. 8 (b), further implying a high degree of accuracy in the measurement. The linear calibration data for the heparin-serum mixture indicates that the HC-PCF-based Raman sensor has a great potential for qualitative as well as quantitative monitoring of heparin in serum.
It should be noted that the lowest concentration of detected heparin is 12USP/ml which falls within the clinical level detection of heparin in patients during surgeries. The entire credit for detecting a weak Raman scatterer (such as heparin) at such low concentration in a complex matrix goes to the HC-PCF because it supports a strong light-matter interaction. However, we are exploring methods to further improve the detection limit of the HC-PCF-based sensor system. The inclusion of nanoparticles to further enhance the Raman signal of heparin can be a way forward. However, synthesizing nanoparticles that exhibits affinity for complex/bulky molecules (like heparin) along with the robustness to retain their optical property in a serum environment will be challenging. Another concern is to find ways to estimate the heparin concentration directly from blood samples instead of from serum. To this end, a porous membrane could be embedded into the sample reservoir that would filter the hemoglobin from blood and allow only the heparin-serum-serum solution to penetrate into micro-channels of HC-PCF .
The present work demonstrates a novel application of hollow core photonic crystal fibers in monitoring the quantity of heparin in serum by Raman spectroscopy. It illustrates a simple procedure of filling the holes of HC-PCF by a sample mixture whose molecular constituents (heparin in our case) can be determined quantitatively by measuring the corresponding Raman peak intensity. The HC-PCF supported strong modal field overlap with the heparin-serum mixture which resulted in ~90 times enhancement of the Raman signal of heparin. Consequently, it enabled us to detect compositions of heparin as low as 12 USP/ml in serum. The present detection scheme has the merit of high detection sensitivity with minimal sample requirement along with molecular specificity as offered by Raman spectroscopy. Hence, it finds relevance in clinical diagnosis which requires the extraction of micro-litre amounts of bodily fluids with precise estimation of its component(s).
This work was supported by Ontario Centres of Excellence (OCE), and The Technology Transfer and Business Enterprise (TTBE). We would also like to thank Gurinder Gill from the Heart Institute for valuable discussion on clinical monitoring of heparin and Louis Tremblay from the Department of Chemical Engineering of the University of Ottawa.
References and links
1. S. A. Spinler, A. K. Wittkowsky, E. A. Nutescu, and M. A. Smythe, “Anticoagulation monitoring part 2: Unfractionated heparin and low-molecular-weight heparin,” Ann. Pharmacother. 39(7), 1275–1285 (2005). [CrossRef] [PubMed]
2. N. M. Milović, J. R. Behr, M. Godin, C. S. Hou, K. R. Payer, A. Chandrasekaran, P. R. Russo, R. Sasisekharan, and S. R. Manalis, “Monitoring of heparin and its low-molecular-weight analogs by silicon field effect,” Proc. Natl. Acad. Sci. U.S.A. 103(36), 13374–13379 (2006). [CrossRef] [PubMed]
4. H. Szelke, J. Harenberg, and R. Krämer, “Detection and neutralisation of heparin by a fluorescent ruthenium compound,” Thromb. Haemost. 102(5), 859–864 (2009). [PubMed]
6. N. M. Milovic, J. R. Behr, M. Godin, C. S. Hou, K. R. Payer, A. Chandrasekaran, P. R. Russo, R. Sasisekharan, and S. R. Manalis, “Monitoring of heparin and its low-molecular-weight analogs by silicon field effect,” Proc. Natl. Acad. Sci. U.S.A. 103(36), 13374–13379 (2006). [CrossRef] [PubMed]
7. H. Yan, C. Gu, C. Yang, J. Liu, G. Jin, J. Zhang, L. Hou, and Y. Yao, “Hollow core photonic crystal fiber surface-enhanced Raman probe,” Appl. Phys. Lett. 89(20), 204101 (2006). [CrossRef]
8. J. B. Jensen, L. H. Pedersen, P. E. Hoiby, L. B. Nielsen, T. P. Hansen, J. R. Folkenberg, J. Riishede, D. Noordegraaf, K. Nielsen, A. Carlsen, and A. Bjarklev, “Photonic crystal fiber based evanescent-wave sensor for detection of biomolecules in aqueous solutions,” Opt. Lett. 29(17), 1974–1976 (2004). [CrossRef] [PubMed]
9. K. Nielsen, D. Noordegraaf, T. Sørensen, A. Bjarklev, and T. P. Hansen, “Selective filling of Photonic Crystal Fibres,” J. Opt. A, Pure Appl. Opt. 7(8), L13–L20 (2005). [CrossRef]
10. Y. Huang, Y. Xu, and A. Yariv, “Fabrication of functional microstructured optical fibers through a selective-filling technique,” Appl. Phys. Lett. 85(22), 5182–5184 (2004). [CrossRef]
11. M. Naji, A. Khetani, N. Lagali, R. Munger, and H. Anis, “A novel method of using hollow-core photonic crystal fiber as a Raman biosensor,” Proc. SPIE 6865, 68650E (2008). [CrossRef]
12. A. Khetani, M. Laferrière, H. Anis, and J. C. Scaiano, “Laser flash photolysis with nanoliter samples: photonic crystal fibers as ultrasmall smart test tubes,” J. Mater. Chem. 18(40), 4769–4774 (2008). [CrossRef]
13. V. S. Tiwari, A. Khetani, M. Naji, and H. Anis, “Study of Surface Enhanced Raman Scattering (SERS) within Hollow Core Photonic Crystal Fiber,” IEEE Sensors 5404, 367–370 (2009). [CrossRef]
14. A. Khetani, M. Naji, N. Lagali, R. Munger, and H. Anis, “A method for using Photonic Crystal Fiber as a Raman biosensor”, US Patent 2010/0014077 (2010).
15. G. Antonopoulos, F. Benabid, T. A. Birks, D. M. Bird, J. C. Knight, and P. St. J. Russell, “Experimental demonstration of the frequency shift of bandgaps in photonic crystal fibers due to refractive index scaling,” Opt. Express 14(7), 3000–3006 (2006). [CrossRef] [PubMed]
17. C. Herrmann, C. Vrancic, A. Fomichova, N. Gretz, S. Hoecker, A. Pucci, and W. Petrich, “In vitro characteristics of a mid-infrared continuous glucose sensor,” Proc. SPIE 7560, 75600E, 75600E-6 (2010). [CrossRef]