Open Access
Add to BibSonomyAbstract
A photonic crystal fibre has been fabricated with a photonic crystal that is surrounded by a number of silica cores. Bending of the fibre induces an interaction between the core and photonic crystal areas, resulting in a highly wavelength-dependent loss of the core modes. White-light transmission experiments are presented which show that the colour of the transmitted light changes as a function of the fibre-bending radius. We compare the results to a simple model and find agreement.
©2000 Optical Society of America
1. Introduction
The first air-silica microstructure fibre was produced at Bell Labs in 1972 [1], a project that was ahead of its time. In the light of today’s demand for specialty fibre with very specific optical properties, air-silica structure fibres have been rediscovered as an extremely versatile and promising new class of fibres, with highly tailorable optical properties. The fibres are fabricated from pure silica, the guiding properties arise from a pattern of microscopic air holes that run down the entire length of the fibre. The unusual optical properties of these photonic crystal fibres include single-mode guiding at all optical wavelengths [2], zero group-velocity dispersion at wavelengths around 800 nm [3,4], and the possibility of designing fibres with an almost wavelength-independent dispersion [5]. In strong contrast to conventional optical fibres, where total-internal reflection is essential for guiding, a photonic crystal fibre (PCF) can guide light in a low refractive-index region such as air, relying solely on photonic bandgap confinement [6]. More commonly, however, PCFs are designed with a silica core in the centre of a hexagonal pattern of air holes, allowing guiding to be aided by an average refractive-index effect [2].
We present the results of a study of a fibre that has been fabricated with a number of silica cores surrounding a photonic crystal region. When the fibre is bent, coupling is induced between the core and photonic crystal regions, modifying the guiding properties. We find that this interaction causes highly wavelength-dependent losses, resulting in the observation of bending-induced colouring of the transmission.
2. The fibre
The fibre was fabricated by stacking silica tubes of 1-mm diameter in a larger silica tube. This stack was then fused and drawn into long lengths of fibre on a drawing tower. The resulting air-hole size and regularity of the hexagonal pattern depends on the preform temperature, the drawing speed and the air pressure in the preform. At first no attempt was made to control the pressure during the drawing process, resulting in a fibre with a non-uniform distribution of air holes and sizes as reported in [9–11]. A second draw was performed while monitoring and releasing the pressure in the preform, producing the fibre described in this paper. We generally draw hundreds of meters of fibre in a single run with a conventional single polymer coating, the air-hole structure remaining virtually unchanged over lengths of about 10 meters.
Fig.1. Optical micrograph of the cross section of the fibre with five cores near the edge, surrounding a photonic crystal area in the centre. The fibre diameter is 147 µm.
Figure 1 shows an optical micrograph of a cross section of the fibre. Experiments on about one metre of this fibre showed that five cores near the edge of the fibre (numbered areas in Fig. 1) guided laser light at 1.5 µm and 0.633 µm with relatively low losses. The cores guide light through an average index effect, and could be excited individually by butt-coupling to an SMF28 fibre. The transmission through the cores showed a high sensitivity to bending, as a result of the silica bridges linking the silica islands (cores) with their surroundings (outer cladding and photonic crystal area), providing a coupling between the different areas in a way similar to that reported in [9–11]. The photonic crystal in the centre of the fibre consisted of an hexagonal pattern of air holes with an average hole spacing of Λ=6.7 µm and hole diameters ranging from 0.8 to 1.5 µm. The fibre was found to have no individual guiding cores in the central region, and the hexagonal air hole just to the right of core 1 (see Fig. 1) did not guide light at any wavelength.
3. Bending-induced colouring of transmission
Fig. 2. a) and b) show coloured transmission through the five cores when a white light source is launched into the fibre. The two pictures show different colours in the different cores as a result of the different fibre bending radius. c) and d) show the coloured transmission through the central region (overexposing cores 2 and 3 to allow the low intensity of the light in the centre to become visible).
White light was coupled directly into the fibre with a microscope objective. All five cores transmitted white light when the fibre was lying loosely on the table (bending radius R>10 cm). Bending of the fibre resulted in a colouring of the transmitted light in the cores, as can be seen in Fig. 2 a) and b). A close study of the low intensity light visible in the central region sometimes showed a distribution of colours as in Fig. 2 c) and d). Highly dependent on the bending conditions, the entire colour distribution would change through small changes in the bending, resulting in different colours at different positions [12].
The observation of coloured transmission in photonic crystal fibres has in the past been directly associated with the presence of photonic band gaps [6,8,13,14]. When a photonic crystal fibre is designed to have a considerable fraction (or all) of the guiding strength arising from the photonic crystal structure, a colouring of the transmission is observed, reflecting the presence of a photonic bandgap. We observe colouring of the transmission only when the fibre is bent, inducing an interaction between the light in the cores and the photonic crystal.
3.1 Wavelength selection through Bragg reflection
We describe the bending-induced colouring with a model that is based on the angular Bragg grating confinement picture presented in [7], utilising the observation that, qualitatively, many aspects of photonic crystal fibres can be understood in terms of grating diffraction properties. In Fig. 3 a) we depict a core that is separated a distance L from a photonic crystal area, which consists of a hexagonal pattern of air holes spaced by a distance Λ. Fig. 3 b) depicts a side-view of the fibre when bent at a radius R.
Fig. 3. a) Model of the fibre with a core at a distance L from a photonic crystal area with an air hole spacing Λ. b) side-view of the fibre when bent at a radius R. The red dashed lines indicate the wavelength for which the guiding is enhanced by satisfying the Bragg condition.
The angle θ at which the light is incident on the photonic crystal depends on the bending radius R, as shown in Fig. 3 b). A certain wavelength, λB, will satisfy the Bragg condition at that angle. This particular wavelength, for which the guiding is aided by Bragg reflection, experiences much less bending losses compared to all other wavelengths, giving rise to the observed coloured transmission. The wavelength λB is given by mλ B=2n eff Λ sin θ where n eff is the effective refractive index of the photonic crystal, taking into account the refractive index of the air holes, and m is the diffraction order. Using the relation (R+L) cos θ=R we find that
This equation is illustrated by Fig. 4, where we have plotted the dependence of λB on R. It can be seen that the first-order Bragg reflection (m=1) reaches visible wavelengths (λ>400 nm) at bending radii below ~6 cm. It must be noted, however, that the bending losses -and therefore the colouring effects - are still very small for such weak bending. In the experiments we have observed a largely white transmission spectrum for all bending radii above ~3.5 cm, although there was occasionally a faint hint of blue. Therefore we have indicated this region in Fig. 4 as ‘white’. Going towards stronger bending (smaller R) from this region, we observe a sequence of colours. Yellow around R=3.5 cm, orange around R=3.1 cm, and red from 3 to 2 cm. Then we observe a gap, corresponding to going from first to second-order Bragg reflection, between 2.1 and 1.6 cm. For the second-order Bragg reflection (m=2) we find blue and green transmission at 1.6 and 1 cm bending radius respectively. Below 1 cm the bending losses were found too severe to observe any transmission in the visible (indicated by the area marked ‘dark’ in Fig. 4). Also, bending the fibre below R=1cm often resulted in fibre breakage.
We conclude from our model that we expect to see a change of transmitted colour from white to yellow, orange, red, blue and green as the bending radius is decreased. We have neglected that the colours will depend on the bending direction and fibre-twist angle. Further, the model neglects the additional complexity one would expect from a 2-dimensional array grating. Therefore only a qualitative comparison with the experimental observations will be possible.
Fig. 4. Wavelength λB of the light that experiences the lowest loss plotted as a function of the bending radius. The curves have been calculated from Eq. (1), with the parameters corresponding to core 4: L=13.4 µm, neff=1.43, and Λ=6.7 µm.
3.2 Observed bending-induced colours in core 4
To illustrate the bending-induced colouring effect and to verify the model presented above, we have focussed on the transmission of core number 4. White transmission was observed for weak or no bending as shown in Fig. 5 a). Colours were observed to be independent of the position of the bend in the fibre; bending the fibre directly after the launching platform, in the middle or near the microscope has the same effect. Further, we have found that the observed colours were independent of the launching conditions.
Because of the dependence of the colours on bending direction, and the difficulty to determine which direction the fibre is bent with respect to the axis between core and photonic crystal, no attempt has been made to accurately record what colours appear at the various bending radii. A particular bending radius can correspond to nearly any of the observed colours, dependent on the bending direction. Twisting the fibre can also induce a change of colour. These effects make it very hard to characterise the fibre quantitatively. We will therefore limit ourselves to a qualitative description of the relation between the observed colours and bending radius.
Some bending directions only introduce losses, and the transmission fades without any colouring effects. When a colour does appear, it is nearly always yellow that is first to appear (at weak bending around 3.5 cm radius). When a series of colours is obtained (by fortunate choice of bending direction) it initially follows the sequence white, yellow, orange, red. This sequence is illustrated by Fig. 5 a) to d). In most of the attempts, we observed no transmission when the fibre was bent beyond the red transmission point d), either because the first-order Bragg condition is satisfied for infrared wavelengths (invisible to the eye), or because the induced losses were too great. Very occasionally, however, a blue or green transmission was found for strong bending conditions, as illustrated by Fig. 5 e) and f). No reproducible method was found to obtain blue or green; and it would generally take some time to achieve the right bending conditions for these colours.
Fig. 5. The transmission through core 4 changes colour as the fibre is bent further, starting with white in a) for an ‘unbent’ fibre and ending with green in f) for the tightest bending radius (R~1 cm).
We find the appearance of colours in agreement with the simple model put forward in Sec. 3.1. The observed colour sequence versus bending radius agrees with the predictions. This is further corroborated by the observation that the colours are independent of the launching conditions and bending position [15]. The difficulty that was experienced going from red to blue/green transmission, corresponds to the calculated change of diffraction order (from m=1 to m=2), as can be seen in Fig. 4.
Figure 6 shows a measurement of the transmission spectrum, corresponding to the case of red transmission as observed in Fig. 5 d). The spectrum was recorded with an optical spectrum analyser with a 10 nm resolution bandwidth, covering the range from 400 to 1750 nm. Due to the low power of the white-light source the signal-to-noise ratio is rather low, though the general features of the transmission spectrum are clear. The spectrum shows a transmission band at 715 nm with a width of about 70 nm, and a number of weak side-bands due to modal interference. We attribute the width of the peaks to local irregularities in the photonic crystal air-hole spacing and/or to non-uniform bending of the fibre. Transmission spectra were obtained for different colours, which showed generally a similar structure.
Fig. 6. Spectrum of the transmission through the bent fibre corresponding to the case of red transmission as in Fig 5 d).
4. Conclusions
Photonic crystal fibres clearly have a bright future ahead of them, with surprising new properties and rapidly growing scope for applications. We have presented an air-silica microstructure fibre of an unusual geometry, with separate core and photonic crystal areas. White-light transmission experiments showed that the fibre can be tuned to transmit any specific colour of the visible spectrum by bending the fibre, which induces highly wavelength-dependent losses.
References and links
1. P. Kaiser, E.A.J. Marcatili, and S.E. Miller, “A new optical fiber,” Bell Sys. Tech. J. 52, No. 2, 265–269 (1973).
2. T.A. Birks, J.C. Knight, and P. St. J. Russell, “Endlessly single-mode photonics crystal fibre,” Opt. Lett. 22, 961–963 (1997). [CrossRef] [PubMed]
3. J. Ranka, R.S. Windeler, and A.J. Stentz, “Visible continuum generation in air-silica microstructure optical fibers with anomalous dispersion at 800 nm,” Opt. Lett. 25, 25–27 (2000). [CrossRef]
4. W.J. Wadsworth, J.C. Knight, A. Ortigosa-Blanch, J. Arriaga, E. Silvestre, and P.St.J. Russel, “Soliton effects in photonic crystal fibres at 850 nm,” Electron. Lett. 36, 53–55 (2000). [CrossRef]
5. J. Broeng, D. Mogilevstev, S.E. Barkou, and A. Bjarklev, “Photonics crystal fibers: A new class of optical waveguides,” Opt. Fiber Tech. 5, 305–330 (1999). [CrossRef]
6. R.F. Cregan, B.J. Mangan, J.C. Knight, T.A. Birks, P.St.J. Russel, P.J. Roberts, and D.C. Allen. “Single-mode photonic bandgap guidance of light in air,” Science 285, 1537–1539 (1999). [CrossRef] [PubMed]
7. J. Canning, “Grating confinement in a photonic crystal fibre,” Opt. Commun. 176, 121–124 (2000). [CrossRef]
8. See e.g. “Focus issue: photonic crystals,” Opt. Express13, 3–34 (1998) http://www.opticsexpress.org/v3n1cvr.htm.
9. T. Ryan, J. Canning, M. Kristensen, and K. Lytkkainen, “Multiple-core air-silica structure optical fibre,” Opto-Electronics and Communications Conference OECC’2000, Chiba, Japan, (2000).
10. J. Canning, M.A. van Eijkelenborg, T. Ryan, M. Kristensen, and K. Lyytikainen, “Complex mode coupling within air-silica structures optical fibres and applications,” submitted for publication (2000).
11. M. Kristensen, J. Canning, and T. Ryan, “Mode coupling in photonic crystal fibres with multiple cores,” Conference on Lasers and Electro Optics CLEO, Nice, France (2000).
12. All photographs presented in this paper were taken with a Kodak DC210 zoom camera placed on top of an Olympus BH2 microscope. The various colours observed in all the photographs accurately represent what was observed simultaneously looking down the microscope.
13. J.C. Knight, J. Broeng, T.A. Birks, and P.St.J. Russell, “Photonic band gap guidance in optical fibers,” Science 282, 1476–1478 (1998). [CrossRef] [PubMed]
14. A.R. Parker, R.C. McPhedran, D.R. McKenzie, L.C. Botten, and N.A. Nicorovici, “Nature’s finest photonic crystal,” submitted for publication (2000).
15. The fibre reported in [9–11] showed no indications of colouring at all: white light experiments showed transmission at all wavelengths, independent of bending of the fibre.
