Specialty optical fibers, in particular microstructured and multi-material optical fibers, have complex geometry in terms of structure and/or material composition. Their fabrication, although rapidly developing, is still at a very early stage of development compared with conventional optical fibers. Structural characterization of these fibers during every step of their multi-stage fabrication process is paramount to optimize the fiber-drawing process. The complexity of these fibers restricts the use of conventional refractometry and microscopy techniques to determine their structural and material composition. Here we present, to the best of our knowledge, the first nondestructive structural and material investigation of specialty optical fibers using X-ray computed tomography (CT) methods, not achievable using other techniques. Recent advances in X-ray CT techniques allow the examination of optical fibers and their preforms with sub-micron resolution while preserving the specimen for onward processing and use. In this work, we study some of the most challenging specialty optical fibers and their preforms. We analyze a hollow core photonic band gap fiber and its preforms, and bond quality at the joint between two fusion-spliced hollow core fibers. Additionally, we studied a multi-element optical fiber and a metal incorporated dual suspended-core optical fiber. The application of X-ray CT can be extended to almost all optical fiber types, preforms and devices.
© 2014 Optical Society of America
The fabrication of conventional optical fibers has undergone more than 40 years of continuous development and the various preform fabrication techniques and fiber-drawing processes used to produce the hundreds of millions of kilometers of low-loss transmission fiber installed each year are now highly engineered and optimized. However, there is ever increasing interest in specialty fibers, in particular microstructured optical fibers (MOFs) and multimaterial optical fibers (e.g. incorporating. different glass, metals and polymer elements) (MMOFs) which are used for an ever increasing array of fiber optic applications, and which frequently require very complex internal structure to achieve their desired optical properties. The fabrication of MOFs and MMOFs is substantially more complex than that of conventional fibers [1, 2], and typically involves a multi-stage fabrication process that requires the manual stacking and drawing of preform assemblies (that often comprise materials with very different thermal and mechanical properties) first to canes and then, in a second step, to fibers. There is, therefore, significant potential to introduce defects, stacking errors and contaminants into the fabrication process that can ultimately result in compromised fiber performance and yield . In order to understand the origin and impact of the effects of such problems during the fabrication process it is highly desirable to develop non-destructive techniques capable of accurately providing high resolution 3-D structural images within the complex preforms and canes used within the fabrication process, and indeed within the final fibers themselves. To date this has not been possible, however, thanks to recent advances in X-ray computed tomography (CT), the non-destructive study of preforms, canes and fibers of the most complex specialty fibers, covering a very wide range of dimensions is now viable. This technique can be applied to virtually all types of fibers i.e. both specialty and conventional.
In this paper we present what is, to the best of our knowledge, the first use of X-ray CT to image the internal structure of optical fibers and preforms. We begin with a short outline of the conventional methods used to inspect fibers and preforms, and then briefly explain the concepts and operating principle of X-ray CT. We then report the demonstration of a number of exemplar applications of the technique on a range of fiber types that highlight the power and utility of the approach. In particular, we describe the non-destructive structural characterization of Hollow Core Photonic Bandgap Fibers (HC-PBGFs), arguably one of the most structurally complex optical fiber types produced to date, and their preforms. We then use the technique to understand the internal 3D-deformations caused during the splicing process of HCPBGFs, before applying it to determine the internal arrangement of fiber elements in a multi-element fiber, and the consistency of metal rods in a metal incorporated fiber.
2. Conventional methods of identifying structural non-uniformities
Conventional solid optical fibers are generally analyzed using either refractometry or microscopy techniques . Refractometry suits simple structures; however, the equivalent is not yet available for complex structures. Optical and electron microscopy are the most common techniques for the structural inspection of complex fibers. These techniques require a clear line of sight to the target under inspection. Therefore, to image the internal structure of the fiber or preform, one should create a flat cross section normal to the fiber axis, e.g. by cleaving, sawing, or polishing them, which in most cases is destructive and prevents further use, particularly in the case of preforms. Additionally, for some fiber types, mainly multi-material fibers, achieving a clean cleave is difficult due to the different material characteristics, simply because the crack does not propagate across the different materials uniformly. Furthermore, in the case of a rapidly changing structure the cleaving approach ultimately limits the spatial resolution at which the structural change can be analyzed, the spacing between cuts is limited in practice to the mm range . Moreover, the presence of orientation dependent features, such as twists, is difficult to capture using sequential cleaves.
3. Tomographic methods
CT is a method to obtain a three dimensional (3D) density map of a specimen. There are various ways of doing CT, such as acoustic, optical and X-ray .We used X-ray CT because it is the most versatile in terms of structure, materials, scale factor and resolution. Using X-ray CT a 3D map of the density of the sample is reconstructed from a collection of conventional 2D radiographs acquired from hundreds or thousands of different angles. Voxel is a term used to define a pixel in 3D space representing a small cube with known dimensions. The 2D images are then reconstructed into a 3D data set with each voxel containing a value of the density of that location in space. This method can capture complex internal geometries where there is sufficient contrast between material densities .
X-ray CT is used prominently in medical applications, however it is finding increasing scientific applications in the field of material science, archaeology, and marine biology [7, 8]. Recent advances in CT techniques employing X-ray optics allow voxel sizes smaller than 50nm . X-ray CT allows nondestructive observation of the internal structure of preforms, canes and fibers, allowing us to investigate their structural uniformity in three dimensions.
X-ray CT requires a number of radiograph images of an object to create its tomographic images, i.e. its stack of virtual slices. Radiographs are monochromatic images. The pixel values in radiographs represent the total attenuation along that X-ray path length - the line integral from the X-ray source, through everything in the way, to the detector. Each intensity level (grey shade value) represents a specific density and directly depends on the atomic number of the elemental composition of the specimen. As the width of a voxel is much larger than a single atom, the captured intensity represents the average density of all compositional elements within the window of the voxel; therefore, compound materials are identified with a specific attenuation level.
An X-ray CT scanner is used to acquire the radiograph images. This machine has three main parts: an X-ray Source, a stage (sample holder) and an X-ray detector. A common configuration of the parts is shown in Fig. 1. The source emits X-rays that propagate toward the detector while en-route they interact with the specimen . The detector collects a radiograph image (either 1D or 2D image depending on the detector array) of the specimen at each time the source illuminates the specimen from a new angle. Tomographic images of the specimen are reconstructed from the collected radiographs, using reconstruction methods such as the back projection method .
The detector can be a 1D array where a virtual cross-sectional 2D image is reconstructed; by repeating 1D-sampling along the third dimension of the specimen, a sequence of 2D images is produced allowing reconstruction of the 3D map. In the case that the detector is a 2D array, then the reconstruction process produces a 3D image of the specimen directly.
In this work, we performed a set of X-ray CT scans at the µVIS Centre at the University of Southampton using three different X-ray CT machines: (a) 225kV/450kV Hutch: Nikon/Metris, (b) 225kV Nikon/Metris HMX ST, and (c) Zeiss Xradia Versa 510 . In this paper, we provide the first demonstration of the potential of X-ray CT for both preform and fiber analyses, showing that it enables non-destructive measurements that simply cannot be made by other means.
The setups we used for our measurements use a fixed source, a fixed 2D detector array and a rotational stage similar to Fig. 1. The Hutch has been designed to accommodate specimens as large as 1 × 1 × 1.5m weighing up to 100kg. It has been equipped with two X-ray sources (20-225kV and 100-450kV). We use the low energy source that can provide a minimum resolution of 3µm in order to study preforms. The HMX scanner is designed to scan specimens ranging in dimension from ~5mm to 300mm with a lowest achievable resolution of 3µm. The HMX is suitable for the study of small preforms and canes. The Versa scanner is used to scan relatively smaller specimens covering dimensions from 0.7mm to 50mm with a minimum resolution of 0.7µm. We used the Versa to scan canes and fibers. As seen in Fig. 1, the specimen is fixed firmly on the rotational stage to avoid any vibration while the stage rotates. A typical full scan of our objects takes ~30minutes to 12 hours depending on the size and required resolution.
Analysis of the reconstructed data is done in post-processing software. The generated radiographs usually have a bit-depth of 16-bit, i.e. in the tomographic images there are 216 shades of grey (intensity levels) representing the accumulated attenuation along the lines between the X-ray source and detector pixels. A useful tool in post-processing of CT data is the image histogram. All the examples discussed next are analyzed using this feature at some point. A histogram is a graph showing how many pixels/voxels are binned at each intensity level. One can use image histogram as a basic tool to separate different materials within the full data set.
4. Exemplar Cases
In this section, we demonstrate four different exemplar applications of the X-Ray CT technique on a range of fiber types. The applications have been selected to highlight the power and utility of the approach and we have opted to use a diverse range of fiber types to emphasize its general applicability.
4.1 Exploring structural integrity and looking for contamination in HC-PBGFs and Preforms
HCPBGFs belong to the family of microstructured fibers. They have a complex cross-sectional structure that consists a central large hole as their core and a microstructured cladding with a few hundreds holes surrounded by sub-µm nodes connected via very thin (~100 nm) glass membranes [Fig. 2(a)]. These features extend uniformly along the entire length of fiber. HC-PBGFs represent a strong candidate to replace conventional optical fiber in next generation optical transmission networks, offering 1000 times lower nonlinearity, ~50% less latency and the prospect of an order of magnitude lower loss compared to conventional fiber . In addition, they offer novel solutions for sensing, power delivery, medicine and fundamental science. The electromagnetic resonances of the node and strut features of their structure determine the guiding mechanism and therefore their consistency and uniformity both axially and transversally are paramount .
HCPBGFs are fabricated through a complex labor-intensive fabrication process. A plurality of mm size high purity glass capillaries are stacked in a hexagonal lattice and inserted into a glass tube to form the first stage preform [Fig. 2(c)], which typically has an outer diameter (OD) of a few cm. The first stage preform is drawn down to a microstructured cane with an OD of a few mm [Fig. 2(b)], which is then inserted into another jacketing tube to form the second stage preform. This final preform is then drawn on a fiber-drawing tower to produce the final HC-PBGF with an OD ranging from 100 to 300 microns depending on the operating wavelength of choice .
Longitudinal consistency of both preform and cane over their entire length and absence of defects is required to produce a long length of uniform fiber (i.e. free from defects or inconsistencies both of which can significantly degrade the transmission quality of the fiber) . Such fiber defects can alter the microstructure significantly over distance scales ranging from a few centimeters along the fiber and they can be initiated from sub millimeter defects (e.g. contaminants/cracks etc.) in preforms and canes. CT provides a tool to image the full structure of preforms, cane and fiber allowing the longitudinal consistency to be examined.
To highlight the above capabilities we illustrate CT scans from each key fabrication stage of a typical HCPBGF (preform, cane, fiber) in Figs. 3(a)-3(c). Tomographic imaging of the preform, Fig. 3(a), can allow the identification of longitudinal inconsistencies in the arrays of capillaries (i.e. twist, bend or misplaced elements), if present. In addition, one should be able to identify contamination (based on shape or material density), and/or failure of any of the glass capillaries (cracks). This may enable substantially improve time and cost management of the fiber fabrication process. Figure 3(a) shows an example of a preform with some arrangement drift across the preform structure in the very outer ring of holes as highlighted by the yellow oval. Note that this dislocation is not visible from the free ends of the preform. The voxel size and scan length are 18.4 µm and 36.8 mm respectively. In Fig. 3(b) the X-ray CT shows the image of a cane approximately 10 times smaller in diameter, as compared to the preform. The voxel size and scan length are 3.35µm and 3.39mm respectively. The cane we imaged presents an excellent consistency and longitudinal uniformity. In Fig. 3(c) an image of a 19cell HCPBGF is shown with a quarter cut-away during the post processing of the data in order to show the internal structure of the fiber. The voxel size here is set to 0.418µm and the scan length is 407µm. In general, any structural defects or contaminants from small foreign objects are easy to identify as long as they are larger than a voxel.
Next, we show how the density- dependent response of this technique can be explored for the study of contamination within the preforms, canes and fibers. Figure 4(a) shows a high-density contaminant discovered in between the outer ring of the microstructure cladding and the jacketing glass tube in one of the canes we examined. The density of this defect would seem to indicate it is a high-density material, (we speculate a metal particle), which may have inadvertently been incorporated during the preform assembly. The deformation induced by the particle on the wall of its adjacent hole is shown in Fig. 4(b). If such a structural deformation were to happen in the vicinity of the core, it would induce a substantial undesirable change in the transmission characteristic of the resultant fiber.
4.2 Exploring structural distortions during the splicing of HCPBGFs
Fusion splicing is the standard method of joining two optical fibers together with low loss. In the case of specialty fibers, this method is still applicable however often presenting more complications in terms of the splice loss, fragility and sensitivity. Splicing microstructured fibers is particularly challenging because of structural deformation caused by the fusion splicing process; for instance, we have developed a special procedure to join HCPBGFs together , and to fuse an HC-PBGF to a solid fiber [14–16]. Low loss and robust “HCPBGF to HCPBGF” or “HC-PBGF to solid fiber” splices have been demonstrated  but careful choice of the splice recipe in order to avoid damage to the microstructure is essential. The recipe has been designed to fuse glass regions together while avoiding the collapse of the delicate microstructure. Some qualitative information on the effect of the splice can be gained e.g. from side imaging of the splice or by carefully breaking and inspecting the splice point , however a more detailed nondestructive method is preferable for the optimization of the splice process. In this case, 3D-CT imaging offers the opportunity for non-destructive analysis of a splice allowing, for example, to investigate the quality of the bond and to explore the induced structural deformation with unprecedented detail.
To illustrate this capability Fig. 5 shows images of a 19-cell HC-PBGF which has been cleaved and fusion spliced to itself. Figures 5(a) and 5(b) are dark-field and bright-field optical images respectively, highlighting a discontinuity. This can be due to small lateral and rotational misalignments as well as to deformation of the microstructure due to the fusion splicing. The heat during the splice causes the microstructure to retract slightly into the respective fibers leaving a funnel-like cavity in the cladding region, at the end face of each fiber. Figure 5(c) is an electron microscope image of the HC-PBGF cross section showing deformation at the splice point. The SEM is obtained by breaking the splice, but clearly the ability to image the splice non-destructively is important.
The splicing parameters were improved based on the optical images accessible through the splicing machine. In order to assess quantitatively the impact of the splicing process on the structure of the fibers on either side of a splice we used X-ray CT. In this study, we used the Xradia Versa with a voxel size of 0.714µm to study the structural changes induced in fusion splicing of two HCPBGFs.
Figure 6 shows the result of the splice scan where a quarter of the structure is clipped out to expose the internal structure of the fiber. The outer diameter of the fibers is 211µm ± 0.7µm. The position of the joint where the jacketing glass was fused together is visible in the center of the image. The distance between the two microstructured claddings varies between 0 and 21µm across the gap.
The 3D data set of the splice contains vast amounts of information that can be post-processed in many different ways, some of which are shown in the examples below. Micron scale misalignment of the cladding boundaries can be clearly seen in the inset of Fig. 7(a). The angular misalignment (~51°) between the microstructures of the fibers is seen in Fig. 7(b) observed through virtual cross-sections obtained below and after splice point without physically cutting the fiber. The concentricity of the core-to-core alignment is also measurable. Structural deformations due to the heating process can be detected up to 165µm before and after the splice point. For example, in this splice the core has expanded from a nominal 38µm to 41µm near the joint over this distance scale. Yellow lines in Fig. 7(a) determine the core boundary. At the point where the two fibers meet, on the left of the image the core boundary looks straight without any noticeable deviation from the ideal shape. However, a strong structural deformation is seen along the right core boundary. Moreover, we can observe the impact of the cleave quality on the strength of the bond; the uneven surface generated by the cleaving crack has caused a gap (dark patch) in the bond between the fibers [Fig. 7(c)]. Any deviations from ideal structure are bad for a number of reasons and must be minimized using often-complex splice parameters in the recipe. This analysis thus provides a very powerful way to assess the impact of individual steps in the splice recipe.
4.3 Investigating structure and polymer cladding quality in Multi Element Fibers
Multi-Element Fibers (MEFs) are made of several individual small diameter optical fibers (elements) embedded in a common polymer coating to form a single strand of fiber . They represent a novel technological approach proposed to enable spatial channel multiplexing, known as SDM, for next generation multiplexing technologies in optical transmission networks. Figure 8 shows a few examples of MEFs. MEF ensures signal multiplexing or demultiplexing with ultra-low crosstalk between spatial channels while providing easy end access similar to conventional fashion without requirement for special fan-in/out devices .
Arrangement of the elements is a determining factor on the structural strength of MEFs. Additionally, active MEFs (doped with Erbium/ Ytterbium) with cladding-pump amplification arrangement require optical contact between the pump and signal fiber-elements to be maintained for efficient pump transfer. Also macro bending and trapped bubbles in the polymer-coating can affect the transmission characteristic and performance of these fibers.
Cleaving MEFs to produce a flat cross-section for structural inspection is non-trivial as the polymer coating does not allow the crack to propagate across all elements. The micrographs in Fig. 8 have been achieved through many attempts with a high precision saw and focus stacking (that combines micrographs taken at different focus depth to form a sharp image with longer depth of field) which do not assure the fidelity of small details. Consequently, locating element positions within the polymer coating and obtaining a clean cross-section image of the fiber are tedious and difficult. Analysis of homogeneity of the coating and longitudinal element arrangement is also not straight forward using conventional methods. CT can provide a solution to the above problems, as shown as follows.
We used X-ray CT to look at a passive 7-element MEF with a hexagonal arrangement of elements. The achieved voxel resolution in this measurement was 0.71 µm. Figure 9(a) shows a virtual cross-sectional image of the MEF. Data related to the fiber elements has been removed from the MEF [Fig. 9(b)] in order to separate the polymer coating region from the glass and to help with identifying features more precisely. This allows the uniformity of the polymer to be checked, i.e. to determine whether any air bubble or contaminant has been trapped within it. Such analysis (not shown here) confirmed a uniform polymer coating without any trapped bubble (within the resolution of this measurement). Figure 9(c) is a 3D view of the sample, where half the data is cut away to illustrate the post-processing functionalities and to visualize the physical features of this structure.
The fiber’s diameter was measured as 400 µm and the cladding diameter of each element is ~80 µm. The average core size of the elements is about 8.5 µm. A dual coating process was used in the fabrication of this MEF. The fiber cores were doped with ~3.7 mole% of Ge representing ∆n = 0.0055. Despite the low concentration of Ge, the sensitivity of X-ray CT to material density allows the core and cladding of each element to be distinguished and also the two polymer coating materials. Figure 10(a) compares the desired arrangement against the actual fiber. The data across the dashed white line is drawn with a red curve in Fig. 10(b) showing the relative intensity for different regions in the radiographs. Measurement artifacts (phase contrast edge enhancement), highlighted by the yellow ellipse, have produced a sharp contrast at the transition interfaces between air-polymer and polymer-glass. The histogram curve of this measurement, Fig. 10(c), allows separation of these regions. Each Gaussian distribution and its corresponding peak in the histogram curve represent a region with a particular X-ray beam attenuation, which corresponds to a specific region of the specimen. The phase enhancement has produced two unwanted peaks in the histogram curve; highlighted with arrows. The information we obtained through CT scans provides valuable feedback for optimization of the MEF fabrication process and its mechanical and optical characteristics.
4.4 Examining element and structural integrity in Metal-Incorporated Fibers
Metal-Incorporated Fibers (MIFs) open the possibility of in-line control of fiber properties via electrical signals [19–22]. The fabrication of a novel type of dual-core fiber, where the relative position of the cores can be controlled by an actuation mechanism based on electrostatic forces, is currently under development [19, 23]. Figure 11 shows, a fiber that exhibits a dual suspended-core structure at the center surrounded by four metal electrodes embedded in the fiber cladding. The fiber structure is based on lead-silicate glass (Schott F2) and low melting temperature tin. A dual-core preform was fabricated by the extrusion technique, and then drawn down to a smaller dimension to form the fiber cane  [the red part in Fig. 11(a)]. The dual-core cane is placed in the middle of a Schott F2 glass jacket tube [Fig. 11(a): the yellow tube] and is surrounded by four glass tubes filled with tin (the blue and grey parts respectively). The structure is packed with some extra F2 rods (the green parts). Figure 11(b) shows the actual preform. Figure 11(c) shows an SEM picture of the fiber with the two suspended cores clearly seen in the center. Each core is ~2.0 × 3.1 µm and they are separated by an air gap of 200 nm and suspended on two 500 nm thick glass membranes. The average diameter of the four tin wires is ~50 µm.
Uniformity, continuity and adherence of the metal rods as well as the detailed structure are important in determining the optical switching performance. Assessing these qualities requires a clear cross sectional image of the fiber to investigate how the metal has filled the glass tube. Because of the different stiffness of the glass and metal, cleaving this structure is very difficult and does not produce a clear cross section that can be used to assess the consistency of the metal rods. Here, CT is capable of providing images of the internal structure without requirement for any cleaves.
Figures 12 and 13 show the result of the CT scan of the MIF described above. A virtual cross-sectional image of the fiber is seen in Fig. 12(a); it allows geometrical investigation of the different parts of the structure. Full reconstruction of the 3D structure of fiber from the CT data is seen in Fig. 12(c), where the air, metal rods and glass materials were separated based on the histogram in Fig. 12(b). The CT data set allows structural investigation of metal rods. Figure 13(a) shows an isolated rod with its surface topology where one can observe surface features like bumps and dents and the overall uneven shape of the rod. Moreover, this data set provides the necessary ground for investigation of the homogeneity of the tin rods (i.e. presence of contamination and trapped air bubbles). A possible air void is seen in Fig. 13(b). The achieved voxel size in this measurement is 0.706 µm. CT scans with higher resolution can provide more detail about the cores and membranes in this structure. The information we obtained from CT examination of this sample has provided us with unprecedented detail about the metal rods and their formation. In addition, it will help us to optimize the switching capability of this device by improving the overall structure and its fabrication process.
We have presented the first application of non-destructive X-ray CT to inspect the internal features of various specialty fibers and preforms. We have shown how this non-destructive technique can be utilized to visualize and measure macroscopic and microscopic deformations within a multi-material fiber (e.g. polymer-glass, glass-tin), or within a splice. This technique unveils structural changes induced by the splice process hidden to other inspection methods. Our results illustrate the capability and value of using X-ray CT techniques in the ongoing refinement of various aspects of Microstructured and Multi-Material fiber fabrication and technology.
We wish to thank µVIS Centre for Computed Tomography at the University of Southampton. This work was supported by the EU 7th Framework Programme (grant agreement 228033; MODE-GAP) and by the UK EPSRC (grants EP/I01196X/1 (Hyperhighway) and EP/H02607X/1).
References and links
1. F. Poletti, M. N. Petrovich, and D. J. Richardson, “Hollow-core photonic bandgap fibers: technology and applications,” P. Soc. Photo-Opt. Ins. 2, 315 (2013).
2. F. Benabid and P. J. Roberts, “Linear and nonlinear optical properties of hollow core photonic crystal fiber,” J. Mod. Opt. 58(2), 87–124 (2011). [CrossRef]
3. S. R. Sandoghchi, T. Zhang, J. P. Wooler, N. Baddela, N. V. Wheeler, Y. Chen, G. T. Jasion, D. R. Gray, E. Numkam Fokoua, J. Hayes, M. Petrovich, F. Poletti, and D. J. Richardson, “First Investigation of Longitudinal Defects in Hollow Core Photonic Bandgap Fibers,” in Optical Fiber Communication Conference, OSA Technical Digest (online) (Optical Society of America, 2014), M2F.6. [CrossRef]
4. R. Hui and M. S. O'Sullivan, Fiber Optic Measurement Techniques (Elsevier/Academic Press, 2009).
6. R. Hanke, T. Fuchs, and N. Uhlmann, “X-ray based methods for non-destructive testing and material characterization,” Nucl. Instrum. Meth. A 591(1), 14–18 (2008). [CrossRef]
7. A. E. Scott, M. Mavrogordato, P. Wright, I. Sinclair, and S. M. Spearing, “In situ fibre fracture measurement in carbon–epoxy laminates using high resolution computed tomography,” Compos. Sci. Technol. 71(12), 1471–1477 (2011). [CrossRef]
8. J. Dinley, L. Hawkins, G. Paterson, A. D. Ball, I. Sinclair, P. Sinnett-Jones, and S. Lanham, “Micro-computed X-ray tomography: a new non-destructive method of assessing sectional, fly-through and 3D imaging of a soft-bodied marine worm,” J. Microsc. 238(2), 123–133 (2010). [CrossRef] [PubMed]
9. P. R. Shearing, D. J. L. Brett, and N. P. Brandon, “Towards intelligent engineering of SOFC electrodes: a review of advanced microstructural characterisation techniques,” Int. Mater. Rev. 55(6), 347–363 (2010). [CrossRef]
10. J. Hsieh, Computed Tomography: Principles, Design, Artifacts, and Recent Advances, 2nd ed. ed. (Wiley Interscience; Bellingham, Wash.: SPIE Press, Hoboken, N.J., 2009).
11. Southampton Imaging, “microCT” (University of Sothampton, 16/07/2014, 2014), retrieved 15/08/2014, 2014, http://www.southampton.ac.uk/southamptonimaging/instruments/microct.page.
12. F. Poletti, N. V. Wheeler, M. N. Petrovich, N. K. Baddela, E. Numkam Fokoua, J. R. Hayes, D. R. Gray, Z. Li, R. Slavik, and D. J. Richardson, “Towards high-capacity fibre-optic communications at the speed of light in vacuum,” Nat. Photonics 7(4), 279–284 (2013). [CrossRef]
13. J. Wooler, D. Gray, F. Poletti, M. Petrovich, N. Wheeler, F. Parmigiani, and D. J. Richardson, “Robust Low Loss Splicing of Hollow Core Photonic Bandgap Fiber to Itself,” in Optical Fiber Communication Conference/National Fiber Optic Engineers Conference 2013, OSA Technical Digest (online) (Optical Society of America, 2013), OM3I.5. [CrossRef]
14. F. Couny, F. Benabid, and P. S. Light, “Reduction of Fresnel Back-Reflection at Splice Interface Between Hollow Core PCF and Single-Mode Fiber,” IEEE Photonic. Tech. L. 19(13), 1020–1022 (2007). [CrossRef]
15. F. Benabid, P. J. Roberts, F. Couny, and P. S. Light, Light and Gas Confinement in Hollow-Core Photonic Crystal Fibre Based Photonic Microcells (Journal of the European Optical Society, 2009), Vol. 4.
16. J. P. Wooler, S. R. Sandoghchi, D. R. Gray, F. Poletti, M. N. Petrovich, N. V. Wheeler, N. K. Baddela, and D. J. Richardson, “Overcoming the Challenges of Splicing Dissimilar Diameter Solid-Core and Hollow-Core Photonic Band Gap Fibers,” in Workshop on Specialty Optical Fibers and their Applications, (Optical Society of America, 2013), W3.26. [CrossRef]
17. J. P. Wooler, F. R. Parmigiani, S. R. Sandoghchi, N. V. Wheeler, D. R. Gray, F. Poletti, M. N. Petrovich, and D. J. Richardson, “Data transmission over 1km HC-PBGF arranged with microstructured fiber spliced to both itself and SMF,” in Optical Communication (ECOC 2013), 39th European Conference and Exhibition on, 2013), 1–3. [CrossRef]
18. S. Jain, V. J. F. Rancaño, T. C. May-Smith, P. Petropoulos, J. K. Sahu, and D. J. Richardson, “Multi-Element Fiber Technology for Space-Division Multiplexing Applications,” Opt. Express 22(4), 3787–3796 (2014). [CrossRef] [PubMed]
19. Z. Lian, M. Segura, N. Podoliak, X. Feng, N. White, P. Horak, and W. Loh, “Electrical current-driven dual-core optical fiber with embedded metal electrodes,” in Optical Fiber Communication Conference, OSA Technical Digest (online) (Optical Society of America, 2014), Tu3K.3. [CrossRef]
20. M. Fokine, L. E. Nilsson, Ã. Claesson, D. Berlemont, L. Kjellberg, L. Krummenacher, and W. Margulis, “Integrated fiber Mach-Zehnder interferometer for electro-optic switching,” Opt. Lett. 27(18), 1643–1645 (2002). [CrossRef] [PubMed]
21. H. W. Lee, M. A. Schmidt, H. K. Tyagi, L. P. Sempere, and P. S. J. Russell, “Polarization-dependent coupling to plasmon modes on submicron gold wire in photonic crystal fiber,” Appl. Phys. Lett. 93(11), 111102 (2008). [CrossRef]