May 1, 2015 / Vol. 40, No. 9 / OPTICS LETTERS
2041
Temperature dependence of sapphire fiber Raman scattering Bo Liu,1,* Zhihao Yu,1 Zhipeng Tian,1 Daniel Homa,2 Cary Hill,2 Anbo Wang,1 and Gary Pickrell2 1
2
Center for Photonics Technology, Bradley Department of Electrical and Computer Engineering, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061-0287, USA
Department of Materials Science and Engineering, Virginia Polytechnic Institute and State University, 213 Holden Hall, Blacksburg, Virginia 24060-0111, USA *Corresponding author: [email protected] Received February 20, 2015; revised April 6, 2015; accepted April 7, 2015; posted April 8, 2015 (Doc. ID 235074); published April 27, 2015
Anti-Stokes Raman scattering in sapphire fiber has been observed for the first time. Temperature dependence of Raman peaks’ intensity, frequency shift, and linewidth were also measured. Three anti-Stokes Raman peaks were observed at temperatures higher than 300°C in a 0.72-m-long sapphire fiber excited by a second-harmonic Nd YAG laser. The intensity of anti-Stokes peaks are comparable to that of Stokes peaks when the temperature increases to 1033°C. We foresee the combination of sapphire fiber Stokes and anti-Stokes measurement in use as a mechanism for ultrahigh temperature sensing. © 2015 Optical Society of America OCIS codes: (190.4370) Nonlinear optics, fibers; (290.5860) Scattering, Raman; (060.2370) Fiber optics sensors; (120.6780) Temperature. http://dx.doi.org/10.1364/OL.40.002041
Raman scattering in grade-index multimode fiber has long been recognized as a mature method of distributed temperature sensing. [1–3]. This technique is demonstrated in single-mode fiber both theoretically and experimentally [4,5]. However, fused silica is not able to withstand temperatures in excess of 800°C, prohibiting use in critical temperature monitoring for applications such as power production and coal gasification. Sapphire fiber has attracted more interest in the Raman sensing field in recent years due to its broad transmission range, high temperature melting point (∼2053°C), and corrosion resistance. The temperature dependence of Raman active mode frequency, width [6], and pressure [7] in bulk sapphire have been reported. When the sapphire fiber is contaminated, it is difficult to extract the intensity change required for temperature sensing using Raman Stokes peak analysis due to fiber loss. Perhaps more significantly, the Stokes peaks are less sensitive than the anti-Stokes peaks. However, few have reported observations about the Raman anti-Stokes peaks because they require much higher laser power to observe, and such power is very close to the damage threshold of the sapphire fiber. In this Letter, for the first time to our knowledge, we report the observation of the anti-Stokes component in a sapphire fiber and its temperature dependence on Raman scattering intensity, frequency, and width. We foresee that the combination of anti-Stokes and Stokes technique can be used for distributed ultrahigh temperature sensing. When photons are scattered from an atom or molecule, a small fraction of the scattered photons have frequencies different from the excitation frequency. The scattered photons having lower frequency than incident photons are labeled Stokes photons, and scattered photons having higher frequencies are called anti-Stokes photons. From a quantum-mechanical viewpoint, the incident photon first excites the molecule from ground state to a virtual state. When the molecule goes back to a vibrational state, certain Raman modes will generate 0146-9592/15/092041-04$15.00/0
low-frequency Stokes photons, and when the molecule is excited from a vibrational state to the virtual state and then back to the ground state, higher (anti-Stokes) photon energy is released. Since the anti-Stokes requires the presence of vibrational states, the anti-Stokes are usually much weaker than Stokes. The Raman intensity of the anti-Stokes and Stokes components is proportional to its differential cross-section given by [8] dσ AS 1 1 i h ; (1) ≅ dΩ x λ4AS exp hcΔυ − 1 K B Tx dσ S 1 1 i; h ≅ dΩ x λ4S 1 − exp − hcΔυ K B Tx
(2)
where h is Plank’s constant, c is the velocity of light in vacuum, K B is Boltzmann’s constant, λ is the wavelength, and T is the absolute temperature at locale position x. With increasing temperature, the intensity of anti-Stokes components increases almost linearly, and the intensity of Stokes components does not increase linearly, as predicted by Eqs. (1) and (2). The equations also predict that the anti-Stokes components are more sensitive to temperature variation than Stokes components. For this reason, Stokes components are usually used for calibration purpose. The setup for the experiment is depicted in Fig. 1. The light source is a Nd:YAG laser (Continuum, SLIII-10) at 532 nm. The laser pulse repetition rate was 10 Hz with a pulse energy of about 500 μJ and a pulse width (FWHM) of 24 ns. It is well known that 1 in 108 photons is scattered by the Raman scattering; such weak scattering will be easily masked by blackbody radiation background. A shorter wavelength laser could efficiently avoid the blackbody radiation background because the blackbody radiation intensity increases rapidly with respect to the wavelength less than 2 μm while Raman intensity is inversely proportional to the fourth power of the © 2015 Optical Society of America
2042
OPTICS LETTERS / Vol. 40, No. 9 / May 1, 2015
Fig. 1. Schematic of the experimental setup for Raman scattering detection.
wavelength. Strong fluorescence absorption at 694.3 nm in sapphire fiber was also found due to Cr3 ion impurities with a broad background when the laser light is intense [9–11]. A laser wavelength of 532 nm avoided the strong background from thermal radiation [12] as well as most of the fluorescence. The beam first passed through a laser clean-up filter centered at 532 nm. The beam was then coupled into the sapphire fiber using a concave lens. The sapphire fiber was protected in a ceramic tube (with inner diameter 1.6 mm and outer diameter 6.75 mm) as it was inserted into the furnace. A thermocouple was located beside the ceramic tube to monitor the temperature distribution along the fiber. The backward scattering light was reflected by the beam splitter, passed through a laser reject filter at center wavelength 532 nm, and was last coupled into a 105-μm core multi-mode fiber and detected with a spectrometer (Ocean Optics, USB4000). The sapphire fiber was 0.72 m long, 75 μm in diameter, and was grown by the laser heated pedestal growth method [13]. A dark spectrum was immediately measured before the Raman spectrum by blocking the light into the spectrometer, which was later subtracted from the Raman spectrum measurement. The sapphire fiber’s temperature distribution is shown in Fig. 2. The sapphire fiber was heated at a rate of 10°C/min, and the temperature was kept constant during
Fig. 2. Temperature distribution along the length of the fiber was measured inch-by-inch using a standard thermocouple, indicating that the furnace set point was reached in the center of the fiber, cooling gradually toward the ends of the tube furnace.
the measurements within 1°C. Each measurement was taken after a given temperature was maintained for 30 minutes. The thermocouple beside the ceramic tube was used to measure the point temperature inch by inch to discern the temperature gradient along the fiber. The temperature profile showed a 10-cm-long plateau corresponding to the center of the furnace and dropped quickly from its center. When the fiber’s segment was at the outside the furnace, the temperature dropped quickly to room temperature. Sapphire crystal belongs to space group D63d and has irreducible representation for the optical modes [7,12] of Γ 2A1g 2A1u 3A2g 2A2u 5E g 4E u . Only two A1g and five E g modes are Raman active, and therefore seven phonon modes are expected in Raman spectrum of sapphire with peaks located at 379, 418, 431, 450, 578, 645, and 750 cm−1 . When the laser propagates dominantly perpendicular to C-plane, only 418, 431, 450, 578, and 750 cm−1 Raman modes are excited [7]. Because 418, 431, and 450 cm−1 are too close to each other and 431 and 450 cm−1 are less than 6% of the amplitude of 418 cm−1 [7], it is usually only the 418, 578, and 750 cm−1 Stokes components that are distinguished in sapphire fiber [14]. The experimental result is shown in Fig. 3. Thirty accumulations, each with an integration time 10 s, were averaged in each plot, and high-frequency noise was filtered. At room temperature, only one antiStokes peak (418 cm−1 ) was visible. When the temperature increased to 300°C, two anti-Stokes peaks (418 and 750 cm−1 ) were identified. At temperatures above 500°C, the intensity of the third peak at 578 cm−1 became measurable along with the previous two peaks. The anti-Stokes peak position and Stokes peak position were symmetric as expected. To better predict the peak amplitude, position, and width (Full Width at Half Maximum) on each spectrum, a Lorentz fitting technique was applied to each peak to represent the original peak. The thermal radiation background was recorded by simply turning off the laser. The thermal radiation background was later subtracted in the following figures. Figure 4 shows the temperature dependence of Raman components’ intensity after removing the thermal background. The intensity of the
Fig. 3. Raman peak intensities increase greatly with temperature.
May 1, 2015 / Vol. 40, No. 9 / OPTICS LETTERS
2043
Temperature dependence of sapphire Raman position
25
−1
Frequency change (cm )
20 15
A 751 A 418 S 418 S 751
10 5 0 −5 −10 −15 0
anti-Stokes peak at 751 cm−1 at room temperature was set as zero because this peak was not yet visible, so the Raman intensity were normalized to their value at 300°C. The results demonstrate that the anti-Stokes components are more sensitive to temperature variation than Stokes components, and the temperature response of Stokes components are not as linearly dependent as antiStokes, as predicted by the theoretical equations. If we take into account the fiber loss and laser power depletion, the Raman power received by the spectrometer is proportional to integration differential cross section along the fiber as Z P AS;S
L 0
Z x dσ AS;S exp −2 αxdx dx dΩ x 0
(3)
where L is the sapphire fiber’s length, and α is the attenudσ AS;S denotes differation of the fiber at locale position x, dΩ ential cross-section of anti-Stokes or Stokes, and P AS;S denotes the receiving power of anti-Stokes or Stokes in arbitrary unite. It was noted that the Stokes peaks were the combination of Raman scattering and fluorescence background. A strong fluorescence background centered at 694.3 nm is observed with the broad side lobe extending into the Stokes peak region, reaching its maximum at 300°C. At higher temperatures, the intensity of the fluorescence peak and its broad side lobe gradually reduced. At temperature above 1000°C, the fluorescence background almost diminished, resulting in a peak intensity drop. Since the anti-Stokes peaks were located at shorter wavelengths compared with the Stokes peaks’ wavelengths, the anti-Stokes were less affected by the fluorescence background. Temperature dependence of Raman frequencies of the Stokes components in sapphire was studied by Schauer [15]; thermal expansion of the lattice was believed to be responsible for this dependence. In Fig. 5, the temperature dependence of Raman frequency shift of both antiStokes and Stokes peaks were observed for the first time. The frequency shifts were normalized to their value at room temperature. The shift of the Stokes and
200
400 600 800 Max temperature (oC)
1000
Fig. 5. Temperature dependence of sapphire Raman frequency is strong enough to be utilized in sensing techniques.
anti-Stokes peaks tended to be symmetric, which is predicted by the phase-matching condition of Raman scattering. The Stokes peaks have less standard deviation because the peaks are stronger than anti-Stokes and have a higher signal-to-noise ratio. The result in Fig. 5 predicts that the temperature dependence of Raman frequency shift on any Raman peak is capable of sensing temperature. Temperature dependence of Raman line-width in sapphire fiber for anti-Stokes and Stokes components are also observed for the first time, as shown in Fig. 6. Large Raman shifts undergo a faster broadening speed. It is noted that since the spectrometer has lower bandwidth, the Raman peak width is broadened compared with others’ measurement results [6]. Klemens [16] and Ashkin et al. [6] attempted to explain Raman linewidth expansion by the perturbation theory. The peak of 751 cm−1 is not visible yet at room temperature. Although it becomes measurable at high temperature, the peak intensity is still very small, and the variation of the width is much larger than others, so the Raman width of 751 cm−1 is not presented. Temperature dependence of sapphire Raman width
100 A 418 S 418 S 751
90 Frequency (cm−1)
Fig. 4. Temperature dependence of sapphire Raman intensity. Solid curve shows the theoretical predictions on anti-Stokes 418 cm−1 (blue) and Stokes 418 cm−1 (red) indicated by Eq. (3).
80 70 60 50 40 30 0
Fig. 6.
200
400 600 800 Max temperature (oC)
1000
Sapphire Raman width is temperature dependent.
2044
OPTICS LETTERS / Vol. 40, No. 9 / May 1, 2015
Although our results were similar to theirs, an additional explanation takes the entire fiber’s temperature profile into consideration. Since part of the sapphire fiber was at low temperature and part was at high temperature, one side of peak remained unshifted, while the other side shifted; it is likely that part of Raman linewidth expansion is due to the combination of all Raman line shifts along the sapphire fiber at different temperatures. Raman linewidth-broadening phenomenon can also be used for temperature sensing and is independent of the fiber loss and contamination. For point temperature sensing, tracking the Raman frequency shift is an accurate method but not appropriate for distributed temperature sensing. For distributed sensing, the Raman intensity-based method is more suitable. Providing dual information from anti-Stokes and Stokes peaks, the Raman intensity method is more sensitive, accurate, and physically meaningful. In addition by using a more stable laser, the accuracy of the intensity based method will be greatly improved. In summary, we reported the observation of anti-Stokes peaks in sapphire fiber for the first time. Temperature dependence of sapphire Raman line intensity, frequency shift, and linewidth were experimentally measured up to 1033°C. These experiments demonstrated that the antiStokes components of sapphire fiber are more sensitive to temperature compared with Stokes components, and the intensity becomes comparable to the Stokes components at higher temperature. We foresee the combination of Stokes and anti-Stokes measurement in use as a mechanism for temperature sensing. Particularly, Stokes components can be used for self-calibration or reference deducing the impact of fiber-bending loss or contamination, which is more reliable than relying only on Stokes components. The consistent and predictable shifting of the Raman line frequency as temperature changes could potentially be used for point temperature sensing. An additional explanation on Raman linewidth expansion was given by taking the integration of Raman line along the fiber at different temperatures into account. The authors would like to acknowledge financial support from the National Energy Technology Lab (NETL) at the U.S. Department of Energy (DOE) under contract DE-FE0012274.
This Letter was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference hereinto any specific commercial product, process, or service by trade name, trade mark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof. References 1. K. S. Chiang, Opt. Lett. 17, 352 (1992). 2. N. B. Terry, T. G. Alley, and T. H. Russell, Opt. Express 15, 17509 (2007). 3. A. Polley and S. E. Ralph, IEEE Photon. Technol. Lett. 19, 218 (2007). 4. J. P. Dakin, J. D. Pratt, G. W. Bibby, and J. N. Ross, Fiber Opt. Laser Sensors III 566, 249 (1985). 5. S. D. Dyer, M. G. Tanner, B. Baek, R. H. Hadfield, and S. Woo Nam, Appl. Phys. Lett. 99, 034102 (2011). 6. M. Ashkin, J. H. Parker, Jr., and D. W. Feldman, Solid State Commun. 6, 343 (1968). 7. G. H. Watson, W. B. Daniels, and C. S. Wang, J. Appl. Phys. 52, 956 (1981). 8. M. Hobel, J. Ricka, M. Wuthrich, and T. Binkert, Appl. Opt. 34, 2955 (1995). 9. R. S. Krishnan, Proc. Indian Acad. Sci. 26, 399 (1947). 10. Q. Ma and D. R. Clarke, Acta Metall. Mater. 41, 1811 (1993). 11. C. Zuo, Part II. Luminescence in the Raman Spectra of Aluminum Oxide (West Virginia University, 2002). 12. E. Zouboulis, D. Renusch, and M. Grimsditch, Appl. Phys. Lett. 72, 1 (1998). 13. R. K. Nubling and J. A. Harrington, Appl. Opt. 36, 5934 (1997). 14. C. Raml, X. He, M. Han, D. R. Alexander, and Y. Lu, Opt. Lett. 36, 1287 (2011). 15. A. Schauer, Can. J. Phys. 63, 523 (1964). 16. P. G. Klemens, Phys. Rev. 148, 845 (1966).
2041
Temperature dependence of sapphire fiber Raman scattering Bo Liu,1,* Zhihao Yu,1 Zhipeng Tian,1 Daniel Homa,2 Cary Hill,2 Anbo Wang,1 and Gary Pickrell2 1
2
Center for Photonics Technology, Bradley Department of Electrical and Computer Engineering, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061-0287, USA
Department of Materials Science and Engineering, Virginia Polytechnic Institute and State University, 213 Holden Hall, Blacksburg, Virginia 24060-0111, USA *Corresponding author: [email protected] Received February 20, 2015; revised April 6, 2015; accepted April 7, 2015; posted April 8, 2015 (Doc. ID 235074); published April 27, 2015
Anti-Stokes Raman scattering in sapphire fiber has been observed for the first time. Temperature dependence of Raman peaks’ intensity, frequency shift, and linewidth were also measured. Three anti-Stokes Raman peaks were observed at temperatures higher than 300°C in a 0.72-m-long sapphire fiber excited by a second-harmonic Nd YAG laser. The intensity of anti-Stokes peaks are comparable to that of Stokes peaks when the temperature increases to 1033°C. We foresee the combination of sapphire fiber Stokes and anti-Stokes measurement in use as a mechanism for ultrahigh temperature sensing. © 2015 Optical Society of America OCIS codes: (190.4370) Nonlinear optics, fibers; (290.5860) Scattering, Raman; (060.2370) Fiber optics sensors; (120.6780) Temperature. http://dx.doi.org/10.1364/OL.40.002041
Raman scattering in grade-index multimode fiber has long been recognized as a mature method of distributed temperature sensing. [1–3]. This technique is demonstrated in single-mode fiber both theoretically and experimentally [4,5]. However, fused silica is not able to withstand temperatures in excess of 800°C, prohibiting use in critical temperature monitoring for applications such as power production and coal gasification. Sapphire fiber has attracted more interest in the Raman sensing field in recent years due to its broad transmission range, high temperature melting point (∼2053°C), and corrosion resistance. The temperature dependence of Raman active mode frequency, width [6], and pressure [7] in bulk sapphire have been reported. When the sapphire fiber is contaminated, it is difficult to extract the intensity change required for temperature sensing using Raman Stokes peak analysis due to fiber loss. Perhaps more significantly, the Stokes peaks are less sensitive than the anti-Stokes peaks. However, few have reported observations about the Raman anti-Stokes peaks because they require much higher laser power to observe, and such power is very close to the damage threshold of the sapphire fiber. In this Letter, for the first time to our knowledge, we report the observation of the anti-Stokes component in a sapphire fiber and its temperature dependence on Raman scattering intensity, frequency, and width. We foresee that the combination of anti-Stokes and Stokes technique can be used for distributed ultrahigh temperature sensing. When photons are scattered from an atom or molecule, a small fraction of the scattered photons have frequencies different from the excitation frequency. The scattered photons having lower frequency than incident photons are labeled Stokes photons, and scattered photons having higher frequencies are called anti-Stokes photons. From a quantum-mechanical viewpoint, the incident photon first excites the molecule from ground state to a virtual state. When the molecule goes back to a vibrational state, certain Raman modes will generate 0146-9592/15/092041-04$15.00/0
low-frequency Stokes photons, and when the molecule is excited from a vibrational state to the virtual state and then back to the ground state, higher (anti-Stokes) photon energy is released. Since the anti-Stokes requires the presence of vibrational states, the anti-Stokes are usually much weaker than Stokes. The Raman intensity of the anti-Stokes and Stokes components is proportional to its differential cross-section given by [8] dσ AS 1 1 i h ; (1) ≅ dΩ x λ4AS exp hcΔυ − 1 K B Tx dσ S 1 1 i; h ≅ dΩ x λ4S 1 − exp − hcΔυ K B Tx
(2)
where h is Plank’s constant, c is the velocity of light in vacuum, K B is Boltzmann’s constant, λ is the wavelength, and T is the absolute temperature at locale position x. With increasing temperature, the intensity of anti-Stokes components increases almost linearly, and the intensity of Stokes components does not increase linearly, as predicted by Eqs. (1) and (2). The equations also predict that the anti-Stokes components are more sensitive to temperature variation than Stokes components. For this reason, Stokes components are usually used for calibration purpose. The setup for the experiment is depicted in Fig. 1. The light source is a Nd:YAG laser (Continuum, SLIII-10) at 532 nm. The laser pulse repetition rate was 10 Hz with a pulse energy of about 500 μJ and a pulse width (FWHM) of 24 ns. It is well known that 1 in 108 photons is scattered by the Raman scattering; such weak scattering will be easily masked by blackbody radiation background. A shorter wavelength laser could efficiently avoid the blackbody radiation background because the blackbody radiation intensity increases rapidly with respect to the wavelength less than 2 μm while Raman intensity is inversely proportional to the fourth power of the © 2015 Optical Society of America
2042
OPTICS LETTERS / Vol. 40, No. 9 / May 1, 2015
Fig. 1. Schematic of the experimental setup for Raman scattering detection.
wavelength. Strong fluorescence absorption at 694.3 nm in sapphire fiber was also found due to Cr3 ion impurities with a broad background when the laser light is intense [9–11]. A laser wavelength of 532 nm avoided the strong background from thermal radiation [12] as well as most of the fluorescence. The beam first passed through a laser clean-up filter centered at 532 nm. The beam was then coupled into the sapphire fiber using a concave lens. The sapphire fiber was protected in a ceramic tube (with inner diameter 1.6 mm and outer diameter 6.75 mm) as it was inserted into the furnace. A thermocouple was located beside the ceramic tube to monitor the temperature distribution along the fiber. The backward scattering light was reflected by the beam splitter, passed through a laser reject filter at center wavelength 532 nm, and was last coupled into a 105-μm core multi-mode fiber and detected with a spectrometer (Ocean Optics, USB4000). The sapphire fiber was 0.72 m long, 75 μm in diameter, and was grown by the laser heated pedestal growth method [13]. A dark spectrum was immediately measured before the Raman spectrum by blocking the light into the spectrometer, which was later subtracted from the Raman spectrum measurement. The sapphire fiber’s temperature distribution is shown in Fig. 2. The sapphire fiber was heated at a rate of 10°C/min, and the temperature was kept constant during
Fig. 2. Temperature distribution along the length of the fiber was measured inch-by-inch using a standard thermocouple, indicating that the furnace set point was reached in the center of the fiber, cooling gradually toward the ends of the tube furnace.
the measurements within 1°C. Each measurement was taken after a given temperature was maintained for 30 minutes. The thermocouple beside the ceramic tube was used to measure the point temperature inch by inch to discern the temperature gradient along the fiber. The temperature profile showed a 10-cm-long plateau corresponding to the center of the furnace and dropped quickly from its center. When the fiber’s segment was at the outside the furnace, the temperature dropped quickly to room temperature. Sapphire crystal belongs to space group D63d and has irreducible representation for the optical modes [7,12] of Γ 2A1g 2A1u 3A2g 2A2u 5E g 4E u . Only two A1g and five E g modes are Raman active, and therefore seven phonon modes are expected in Raman spectrum of sapphire with peaks located at 379, 418, 431, 450, 578, 645, and 750 cm−1 . When the laser propagates dominantly perpendicular to C-plane, only 418, 431, 450, 578, and 750 cm−1 Raman modes are excited [7]. Because 418, 431, and 450 cm−1 are too close to each other and 431 and 450 cm−1 are less than 6% of the amplitude of 418 cm−1 [7], it is usually only the 418, 578, and 750 cm−1 Stokes components that are distinguished in sapphire fiber [14]. The experimental result is shown in Fig. 3. Thirty accumulations, each with an integration time 10 s, were averaged in each plot, and high-frequency noise was filtered. At room temperature, only one antiStokes peak (418 cm−1 ) was visible. When the temperature increased to 300°C, two anti-Stokes peaks (418 and 750 cm−1 ) were identified. At temperatures above 500°C, the intensity of the third peak at 578 cm−1 became measurable along with the previous two peaks. The anti-Stokes peak position and Stokes peak position were symmetric as expected. To better predict the peak amplitude, position, and width (Full Width at Half Maximum) on each spectrum, a Lorentz fitting technique was applied to each peak to represent the original peak. The thermal radiation background was recorded by simply turning off the laser. The thermal radiation background was later subtracted in the following figures. Figure 4 shows the temperature dependence of Raman components’ intensity after removing the thermal background. The intensity of the
Fig. 3. Raman peak intensities increase greatly with temperature.
May 1, 2015 / Vol. 40, No. 9 / OPTICS LETTERS
2043
Temperature dependence of sapphire Raman position
25
−1
Frequency change (cm )
20 15
A 751 A 418 S 418 S 751
10 5 0 −5 −10 −15 0
anti-Stokes peak at 751 cm−1 at room temperature was set as zero because this peak was not yet visible, so the Raman intensity were normalized to their value at 300°C. The results demonstrate that the anti-Stokes components are more sensitive to temperature variation than Stokes components, and the temperature response of Stokes components are not as linearly dependent as antiStokes, as predicted by the theoretical equations. If we take into account the fiber loss and laser power depletion, the Raman power received by the spectrometer is proportional to integration differential cross section along the fiber as Z P AS;S
L 0
Z x dσ AS;S exp −2 αxdx dx dΩ x 0
(3)
where L is the sapphire fiber’s length, and α is the attenudσ AS;S denotes differation of the fiber at locale position x, dΩ ential cross-section of anti-Stokes or Stokes, and P AS;S denotes the receiving power of anti-Stokes or Stokes in arbitrary unite. It was noted that the Stokes peaks were the combination of Raman scattering and fluorescence background. A strong fluorescence background centered at 694.3 nm is observed with the broad side lobe extending into the Stokes peak region, reaching its maximum at 300°C. At higher temperatures, the intensity of the fluorescence peak and its broad side lobe gradually reduced. At temperature above 1000°C, the fluorescence background almost diminished, resulting in a peak intensity drop. Since the anti-Stokes peaks were located at shorter wavelengths compared with the Stokes peaks’ wavelengths, the anti-Stokes were less affected by the fluorescence background. Temperature dependence of Raman frequencies of the Stokes components in sapphire was studied by Schauer [15]; thermal expansion of the lattice was believed to be responsible for this dependence. In Fig. 5, the temperature dependence of Raman frequency shift of both antiStokes and Stokes peaks were observed for the first time. The frequency shifts were normalized to their value at room temperature. The shift of the Stokes and
200
400 600 800 Max temperature (oC)
1000
Fig. 5. Temperature dependence of sapphire Raman frequency is strong enough to be utilized in sensing techniques.
anti-Stokes peaks tended to be symmetric, which is predicted by the phase-matching condition of Raman scattering. The Stokes peaks have less standard deviation because the peaks are stronger than anti-Stokes and have a higher signal-to-noise ratio. The result in Fig. 5 predicts that the temperature dependence of Raman frequency shift on any Raman peak is capable of sensing temperature. Temperature dependence of Raman line-width in sapphire fiber for anti-Stokes and Stokes components are also observed for the first time, as shown in Fig. 6. Large Raman shifts undergo a faster broadening speed. It is noted that since the spectrometer has lower bandwidth, the Raman peak width is broadened compared with others’ measurement results [6]. Klemens [16] and Ashkin et al. [6] attempted to explain Raman linewidth expansion by the perturbation theory. The peak of 751 cm−1 is not visible yet at room temperature. Although it becomes measurable at high temperature, the peak intensity is still very small, and the variation of the width is much larger than others, so the Raman width of 751 cm−1 is not presented. Temperature dependence of sapphire Raman width
100 A 418 S 418 S 751
90 Frequency (cm−1)
Fig. 4. Temperature dependence of sapphire Raman intensity. Solid curve shows the theoretical predictions on anti-Stokes 418 cm−1 (blue) and Stokes 418 cm−1 (red) indicated by Eq. (3).
80 70 60 50 40 30 0
Fig. 6.
200
400 600 800 Max temperature (oC)
1000
Sapphire Raman width is temperature dependent.
2044
OPTICS LETTERS / Vol. 40, No. 9 / May 1, 2015
Although our results were similar to theirs, an additional explanation takes the entire fiber’s temperature profile into consideration. Since part of the sapphire fiber was at low temperature and part was at high temperature, one side of peak remained unshifted, while the other side shifted; it is likely that part of Raman linewidth expansion is due to the combination of all Raman line shifts along the sapphire fiber at different temperatures. Raman linewidth-broadening phenomenon can also be used for temperature sensing and is independent of the fiber loss and contamination. For point temperature sensing, tracking the Raman frequency shift is an accurate method but not appropriate for distributed temperature sensing. For distributed sensing, the Raman intensity-based method is more suitable. Providing dual information from anti-Stokes and Stokes peaks, the Raman intensity method is more sensitive, accurate, and physically meaningful. In addition by using a more stable laser, the accuracy of the intensity based method will be greatly improved. In summary, we reported the observation of anti-Stokes peaks in sapphire fiber for the first time. Temperature dependence of sapphire Raman line intensity, frequency shift, and linewidth were experimentally measured up to 1033°C. These experiments demonstrated that the antiStokes components of sapphire fiber are more sensitive to temperature compared with Stokes components, and the intensity becomes comparable to the Stokes components at higher temperature. We foresee the combination of Stokes and anti-Stokes measurement in use as a mechanism for temperature sensing. Particularly, Stokes components can be used for self-calibration or reference deducing the impact of fiber-bending loss or contamination, which is more reliable than relying only on Stokes components. The consistent and predictable shifting of the Raman line frequency as temperature changes could potentially be used for point temperature sensing. An additional explanation on Raman linewidth expansion was given by taking the integration of Raman line along the fiber at different temperatures into account. The authors would like to acknowledge financial support from the National Energy Technology Lab (NETL) at the U.S. Department of Energy (DOE) under contract DE-FE0012274.
This Letter was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference hereinto any specific commercial product, process, or service by trade name, trade mark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof. References 1. K. S. Chiang, Opt. Lett. 17, 352 (1992). 2. N. B. Terry, T. G. Alley, and T. H. Russell, Opt. Express 15, 17509 (2007). 3. A. Polley and S. E. Ralph, IEEE Photon. Technol. Lett. 19, 218 (2007). 4. J. P. Dakin, J. D. Pratt, G. W. Bibby, and J. N. Ross, Fiber Opt. Laser Sensors III 566, 249 (1985). 5. S. D. Dyer, M. G. Tanner, B. Baek, R. H. Hadfield, and S. Woo Nam, Appl. Phys. Lett. 99, 034102 (2011). 6. M. Ashkin, J. H. Parker, Jr., and D. W. Feldman, Solid State Commun. 6, 343 (1968). 7. G. H. Watson, W. B. Daniels, and C. S. Wang, J. Appl. Phys. 52, 956 (1981). 8. M. Hobel, J. Ricka, M. Wuthrich, and T. Binkert, Appl. Opt. 34, 2955 (1995). 9. R. S. Krishnan, Proc. Indian Acad. Sci. 26, 399 (1947). 10. Q. Ma and D. R. Clarke, Acta Metall. Mater. 41, 1811 (1993). 11. C. Zuo, Part II. Luminescence in the Raman Spectra of Aluminum Oxide (West Virginia University, 2002). 12. E. Zouboulis, D. Renusch, and M. Grimsditch, Appl. Phys. Lett. 72, 1 (1998). 13. R. K. Nubling and J. A. Harrington, Appl. Opt. 36, 5934 (1997). 14. C. Raml, X. He, M. Han, D. R. Alexander, and Y. Lu, Opt. Lett. 36, 1287 (2011). 15. A. Schauer, Can. J. Phys. 63, 523 (1964). 16. P. G. Klemens, Phys. Rev. 148, 845 (1966).
