A Sagnac loop interferometer based on polarization-maintaining photonic crystal fiber was built and analyzed. Mainly the temperature dependence of the Sagnac loop filter function was measured and analyzed. By measuring the filtering function of the Sagnac loop as a function of the temperature over 200 °C, we deduced an unambiguous temperature dependent birefringence coefficient, dΔn/dT = -2.0 × 10-9 /K. Over the full temperature swing, the maximum peak shifts was less than 10% of the filter period. For comparison, a standard Sagnac loop was built with the exact same length and experimental condition, where we deduced dΔn/dT = -7.0 × 10-8 /K.
© 2004 Optical Society of America
A polarization maintaining fiber (PMF) Sagnac loop interferometer exhibits low insertion loss, polarization independence to input light, broad useful spectral bandwidth and high resistance to environmental changes. These unique characteristics make it a promising wavelength-selective comb filter for multi-wavelength generation in fiber lasers . Nevertheless, however small, PMF Sagnac filters are affected by environmental changes such as temperature, which results in a slow drift of the peak transmission wavelengths [2,3]. Recently, photonic crystal fibers become available with wide ranges of different fabrication designs including polarization-maintaining photonic crystal fiber (PM-PCF). A few papers report that PM-PCF may exhibits extremely low differential temperature sensitivity between two orthogonal guided modes [4–6]. This property makes Sagnac filters based on PM-PCF extremely insensitive to temperature variations compared to that of standard PMF based Sagnac filters. In this work, using a commercially available PM-PCF, we built a PM-PCF Sagnac loop and analyzed its temperature dependence.
2. Principle and theory
The Sagnac loop interferometer is made up of a conventional single mode 50:50 fiber coupler and a PM-PCF. Figure 1 shows the schematic of the fiber Sagnac loop interferometer. Transfer matrix of the device can be calculated using the standard Jones matrix representation and from it, the optical intensity transmission, t, can be simply written as:
where L is the length of PM-PCF, λ is the center wavelength of the light source, and Δn(T) is the birefringence, n x(T) - n y(T) of the PM-PCF. The equation can be Taylor expended around a temperature, T0 and can be written as:
The wavelength spacing of the peaks, Δλ, is inversely proportional to the length and birefringence of PMF as follows:
And the position of the transmission peaks will depend on the second term of the equation where a full period shift occurs when
Thus, in order to decrease the temperature dependency of a fiber Sagnac loop one can either use a shorter loop or use a fiber with less dΔn/dT value. For a normal elliptically stress-induced 1-meter PM fiber with dΔn/dT ~ -1×10-7 /K, one can expect a full period shift with temperature drift of only 13.8 °C.
3. Experiment and results
A Sagnac loop configuration was built using a standard 50/50 fused fiber coupler with the SMF-28 fiber pigtails and a 1-meter long commercial elliptical-core PM-PCF fiber from the Thorlabs (PM-1550-01, loss < 1.5 dB/km) (see Fig. 1). The two ends of fiber coupler were fusion spliced to the PM-PCF forming the Sagnac loop.
A portion (L’ = 0.6m) of both PM-PCF and PMF from the two different Sagnac loop was placed in a temperature-controllable furnace (Yamato, DX 300). An optical incoherent broadband CW light source (HP 83437A) was used as the light input. The light is launched into the single mode 50:50 fiber coupler via a commercial single mode fiber (SMF-28) patch cord, then Temperature dependence was measured from 40 °C to 240 °C for both Sagnac loops in steps of 10 °C. The each output of the interferometer was connected to an optical spectrum analyzer (Ando AQ6135A, resolution=0.05 nm). Another Sagnac loop using an elliptically stress-induced PMF from 3M (FS-PM-7811) with the same length was built for comparison.
Figure 2 shows filtering spectra and its shift due to temperature change for the PMF based Sagnac loop. Only few examples are shown for easiness of reading. The dotted line was put as the guide to the peak shift. Temperature increment in steps of 2 °C was also measured for the PM version due to the large temperature dependence of the PM version. (see Fig. 3)
The wavelength spacing of the peaks (Δλ) was 6.40 nm and the corresponding birefringence (Δn) obtained from Eq. (3) was 0.00037. Fig. 4 shows filtering spectra and its shift due to temperature change for the PM-PCF based Sagnac loop. The wavelength spacing of the peaks (Δλ) was 3.12 nm and the corresponding birefringence (Δn) obtained from Eq. (3) was 0.00076.
As can be seen from Fig. 2 and Fig. 4, peak shift suddenly changes from red-shift to blue-shift above 200 °C. At that temperature the jackets of fibers are observed to change their color from clear to brown. Thus this sudden change is believed due to burning of the fibers. Temperature dependent peak shift was analyzed in mostly linear region from 40 °C to 150 °C. (See Fig. 5). dλ/dT in this region are -0.41 nm/K and -0.0025 nm/K for PMF and PM-PCF, respectively. PM-PCF showed 164 times less temperature dependency in the peak wavelength shift than PMF did. An internal stress birefringent fiber-based Sagnac interferometer was reported to have dλ/dT of -0.99 nm/K  and this is consistent with our measurements was well. Based on the Eq. (4), the relationship between dΔn/dT and dλ/dT can be derived and can be written as
Figure 2 and Fig. 4 were fitted to Eq. (2), letting the T0 as 40 °C (shown in Fig. 5). Calculated values for dΔn/dT of PMF and PM-PCF are -7.0 × 10-8 /K and -2.0 × 10-9 /K, respectively. The experimental error, which comes from the uncertainly in the measurement of wavelength spacing, wavelength shift and the temperature, was calculated to be ±2×10-9/K and ±1×10-10/K for the PMF and PM-PCF, respectively. This conclude that the birefringence of the PM-PCF measured exhibited 35 times less temperature dependence than that of the standard PM fiber. Szpulak, et. al, calculated sensitivity of modal birefringence to temperature for the wavelength range from 500 nm to 1300 nm . They showed dΔn/dT of standard elliptical core fiber (a/b=5) and elliptical core holey fiber (2a/Λ=0.7) to be -9.0 × 10-8 /K and 1.2 × 10-8 /K at center wavelength 1300 nm, respectively. Our experimental results show that PM-PCF has much lower sensitivity to temperature than PMF does.
In PM-PCF, the core and the cladding of fiber are made of the same pure-silica, thus when temperature is applied, it does not exhibit differential thermal expansion of the core and the cladding, which result in the internal birefringence. Furthermore, pure silica has a very low thermal expansion coefficient and the change in the birefringence due to the thermal expansion of photonic crystal structure is insignificant.
We have experimentally investigated temperature dependence of the birefringence of polarization maintaining photonic crystal fiber using Sagnac interferometer configuration. Our results show that polarization maintaining photonic crystal fiber has 45 times smaller temperature dependence of birefringence that that of standard polarization maintaining fiber. The temperature dependence of birefringence caused 290 times less peak displacement. Photonic crystal fiber’s exceptionally small dependence on environmental change (temperature, in this paper) can open new possibility of building very robust temperature insensitive fiber Sagnac loop interferometers.
References and links
1. C.S. Kim, R.M. Sova, and J.U. Kang: “Tunable multi-wavelength all-fiber Raman source using fiber Sagnac loop filter,” Opt. Commun. 218, 291–295 (2003). [CrossRef]
2. Y. Han, Q. Li, X. Liu, and B. Zhou: “Architecture of high-order all-fiber birefringent filters by the use of the Sagnac interferometer,” IEEE Photon. Technol. Lett. 11, 90–92 (1999). [CrossRef]
4. M. Szpulak, W. Urbanczyk, T. Martynkien, J. Wojcik, and W. J. Bock: “Temperature sensitivity of photonic crystal holey fibers,” in Optical Fibers and Their Applications VIII, J. Dorosz and R.S. Romaniuk, eds., Proc. SPIE 5028, 108–114 (2003). [CrossRef]
5. T. Ritari, T. Niemi, H. Ludvigsen, M. Wegmuller, N. Gisin, J. R. Folkenberg, and A. Petterson: “Polarization-mode dispersion of large mode-area photonic crystal fibers,” Opt. Comm. 226, 233–239 (2003). [CrossRef]
6. M. Szpulak, T. Martynkien, W. Urbanczyk, J. Wójcik, and W. J. Bock: “Influence of temperature on birefringence and polarization mode dispertion in photonic crystal holey fiber,” in Proceedings of 4th International Conference on Transparent Optical Networks and 1st European Symposium on Photonic Crystals, (Warsaw, Poland, 2002) Vol. 2, pp. 89–92.
7. A.N. Starodumov, L.A. Zenteno, D. Monzon, and E. De La Rosa: “Fiber Sagnac interferometer temperature sensor,” Appl. Phys. Lett. 70, 19–21 (1997). [CrossRef]