- Email: [email protected]
Description of an air temperature sensor based on O2 absorption spectroscopy
Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 210 (2019) 245–250
Contents lists available at ScienceDirect
Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy journal homepage: www.elsevier.com/locate/saa
Description of an air temperature sensor based on O2 absorption spectroscopy Marlon Diaz, Parashu R. Nyaupane, Daniel Monteagudo, Carlos E. Manzanares ⁎ Department of Chemistry & Biochemistry, Baylor University, Waco, TX 76798, United States of America
a r t i c l e
i n f o
Article history: Received 28 June 2018 Received in revised form 4 October 2018 Accepted 13 November 2018 Available online 13 November 2018 Keywords: Temperature sensor Cavity ring down Flowing air Oxygen A-band Optical cavity
a b s t r a c t In this paper, we describe in detail an assembled open-path optical cavity to act as a temperature sensor of air. A metal absorption cell in a temperature-regulated tube furnace is placed at the center of an optical cavity. The optical cavity consists of two mirrors, two nitrogen buffer sleeves, and the open cell. The air is injected through a fitting in one extreme of the metal tube and travels half the tube length through a channel in the wall of the tube. The channel directs the air towards the center of the cell. The air flowing is heated at the temperature of the metal tube in contact with the furnace. The heated air injected at the center of the tube, flows towards the open extremes of the tube. The nitrogen buffer sleeves protect the mirrors from the heated air. The temperature of the air flowing through the tube is determined by measuring the absorption of the A band of oxygen as a function of the wavenumber in the 769–755 nm wavelength range. The absorption technique is phase-shift cavity ring down spectroscopy. To obtain the temperature, the energy of the lower rotational state for eleven selected rotational transitions is linearly fitted to a logarithmic function that contains the relative intensity of the rotational transition, the initial and final rotational quantum numbers and the energy of the transition. Accuracy of the measurement is determined by comparing the calculated temperature from the spectra with the analog reading of the temperature-regulated tube furnace. This technique is proposed for exhaust air temperature measurements of combustion chambers and cooling air after passing through the blades of a turbine. © 2018 Published by Elsevier B.V.
1. Introduction Determination of temperature based on measurements of oxygen rotational transitions around 760 nm have been reported using absorption [1–4], emission [5,6], Raman [7,8], and microwave Rayleigh [9–11] techniques. Oxygen absorption in air for a 67 m path length has been obtained for ambient temperature determinations using a diode laser. The ratio of two selected oxygen lines (two-line thermometry) was used to determine the temperature [1]. In another experiment the rotational line absorption of molecular oxygen is obtained in a shock-heated oxygen flow. The absorption intensities of five rotational lines around 760 nm are fitted to an empirical function that includes a line profile function, the absorption frequency, the temperature, and the pressure. The fitting function gives the temperature and pressure of the sample [2]. Laser remote atmospheric temperature has been obtained by observing resonant absorption of oxygen. The calculation of temperature is done with an iterative equation that uses the measured absorption of one representative oxygen line [3]. The oxygen absorption around 760 nm has been obtained using cavity ring down spectroscopy [CRD] to determine oxygen concentrations in a low-pressure flame [4]. In ⁎ Corresponding author. E-mail address: [email protected] (C.E. Manzanares).
https://doi.org/10.1016/j.saa.2018.11.034 1386-1425/© 2018 Published by Elsevier B.V.
emission experiments, rotational temperatures have been obtained from the intensity distribution of the excited state emission of the Aband around 760 nm in an oxygen glow discharge. The temperatures are obtained and compared with temperatures from coherent antiStokes Raman spectroscopy of oxygen molecules in the ground state [5]. The time evolution of the O2 rotational temperature has been estimated for light emitted around 760 nm after a pulsed discharge by comparison with calculated low-resolution spectra for different gas pressures [6]. In Raman scattering experiments, rotational temperatures have been determined in pure O2 at 1 atm for temperatures from 243 to 343 K and in air at 298 K using Raman spectroscopy in an intracavity cell. The temperature was calculated from the Boltzmann equation as well as intensities and frequencies of the transitions [7]. Coherent anti-Stokes Raman scattering experiments have also been reported. Dual-pump coherent anti-Stokes Raman scattering system has been done for temperature and species (CO2, O2, N2) measurements [8]. Another technique for temperature determinations is the coherent microwave Rayleigh scattering from a plasma created by resonance enhanced multi-photon ionization (REMPI) [9–11]. This technique uses a frequency doubled dye laser whose output is focused on an air or oxygen sample to produce a plasma that is irradiated with a microwave source. The microwave horn that acts as source and receiver, detects the scattered microwave radiation from the plasma. The microwave signal
246
M. Diaz et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 210 (2019) 245–250
is input into an automatic data acquisition program that records the rotational spectrum of oxygen as the laser is tuned. The cell is at the temperature of a tube furnace. The rotational temperatures are determined from Boltzmann plots for rotational transitions and compared with the tube furnace temperature [9]. In subsequent applications of the technique, the temperatures are obtained with empirical equations without the use of high resolution rotational transitions using linewidth, linear, and area fitting of selected bands [10,11]. Our group has used phase shift cavity ring down spectroscopy and oxygen absorption (γ and A bands) to obtain temperatures between 90 K and 296 K in a cryogenic cell [12]. We have used the same technique to obtain temperatures between 298 K and 373 K with a heated open glass tube [13]. In this experiment, we describe in detail an open metal absorption cell inside a temperature-regulated tube furnace, placed at the center of an optical cavity. The temperature of the air sample is obtained from the near-IR absorption of oxygen (A-band absorption) using the Phase Shift Cavity Ring Down technique (PS-CRD). The accuracy of the absorption-based temperature calculation is determined by comparing with the temperature of the furnace. This oxygen absorption-based gas temperature sensor as well as, the other spectroscopic techniques mentioned before, can be applied to determine the temperature in atmospheric air, combustion chambers, flames, discharge tubes, and cold plasmas. We are using this absorption method to obtain the temperatures of cooling air at the input and hot air at the output after passing through the blades of a turbine. These measurements could contribute to the development of more efficient cooling technologies, as well as leading to increased efficiency of the gas turbine engine. 2. Experimental 2.1. Open Cell and Optical Cavity We assembled an open-path optical cavity to act as a temperature sensor of air. An experimental setup of the high temperature system is shown in Fig. 1 (a). A metal absorption cell placed in a temperature regulated furnace tube is at the center of an optical cavity. The optical cavity consists of two mirrors, two nitrogen buffer sleeves, and the open cell. Fig. 1 (b) shows details of the tube inside the furnace. The tube is 41 cm long and the outside diameter is 4.5 cm. The wall thickness is 1.3 cm and the inside diameter is 1.9 cm. The sample from a pressurized
air tank (Praxair) is injected through a Swagelok fitting in a 0.32 cm diameter, 20.5 cm length opening through the wall that directs the air towards the center of the tube. The air flowing is heated at the temperature of the metal tube in contact with the furnace. At the center, the hot air comes out of a small opening and flows towards both extremes the tube. The flowing hot air is not allowed to get close to the mirrors by having a room temperature flow of nitrogen gas injected on both sides of the heated cell using PVC pipes of the same diameter of the cell tube. The optical cavity is completed with two highly reflective mirrors (Los Gatos Research, 99.997%) which are mounted parallel to each other and separated by 93 cm. The air flow at the center of the tube was controlled by a flow meter (Cole Parmer, 014-96-ST). The temperature-regulated furnace provides the temperature to compare with the calculated temperature using the oxygen absorption method. 2.2. Cavity Ring Down Spectrometer The experimental set up for the PS-CRD technique has been given in a previous publication [14]. The complete experimental set up for air temperature measurements is illustrated in Fig. 2. Briefly, a laser system consisting of a continuous wave Ti-Sapphire laser (Coherent 899 ring laser) pumped by a Nd-YVO4 laser (Coherent Verdi) in the range 13,000–13,250 cm−1 was used. Calibration of the laser was done obtaining the opto-galvanic spectrum of a hollow cathode lamp filled with neon. The resolution of the pump laser was 0.17 cm−1. A typical spectrum recorded 3174 points (0.064 cm−1/step) during a scan of around 3600 s. An electro optic modulator (Conoptics model 350-50) was used to modulate the laser beam as a 12 kHz square wave. The modulated beam passes through the optical cavity and is detected with a photomultiplier (Oriel model 70680). The zero phase-shift is obtained by bypassing the cavity. The reference signal for the lock-in amplifier (Stanford Research Systems (SRS) model 830) is obtained from a function generator (SRS model DS335) which also supplies the modulation signal to the driver input of the electro-optic modulator. The phase difference between the signal exiting the cavity and the signal entering the cavity is measured with a dual phase lock-in amplifier. For the measurement of the phase shift angle, the laser is tuned to a wavelength where there is no absorption. Fig. 3 also depicts the phase angle measurement set up. The modulated laser beam bypasses the optical cavity (dashed line) using two movable mirrors and reaches the detector. The strong laser signal is sent to a pre-amplifier (SRS-SR560). This signal is
Fig. 1. (a) Expanded view of an open path optical cavity with highly reflective mirrors. The optical cavity consists of two mirrors, two nitrogen buffer sleeves, and an open tube inside a furnace. (b) Shows dimensions and details of the tube inside the furnace.
M. Diaz et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 210 (2019) 245–250
247
Fig. 2. Phase shift cavity ring down spectrometer used to record the oxygen A-band absorption in an open optical cavity. The reference beam bypassing the cavity is shown.
not amplified (×1) but it is sent to the lock in amplifier. The angle between the reference from the function generator and the bypassed signal is automatically made equal to zero with the lock-in amplifier. When the diverting mirrors are removed, the laser passes through the cavity. The weak signal exiting the cavity is pre-amplified (×1000) and sent to the lock in amplifier that shows the angle. The measured value is the phase shift angle due to delay inside the cavity. The phase shift (ϕ) is related to the time that the light spends in the cavity or ringdown time (τ) by the equation: tanϕ = 2πfτ, where f is the modulation frequency [15,16]. Optimal alignment of the cavity mirrors occurs when the phase shift is around −45° that is a region where the tangent function is linear. In a typical experiment the initial phase shift angle is close to −47°, the presence of the flowing air inside the cavity usually changes the angle to −45° due to scattering. The phase angle is recorded as a function of wave number (σ). For a cavity filled with an absorbing gas the ring-down time is: ℓ τσ ¼ c ð1−RÞ þ α O2 ℓ þ α scatt: ℓ
ð1Þ
The time (τσ) is related to the reflectivity of the mirrors (R), the speed of light (c), the absorption coefficient (αO 2) of the sample and the length (ℓ) of the optical cavity. The scattering losses are represented by αscatt. The observed signal is: α σ ¼ ½c τ σ −1 ¼ α O2 þ ð1−RÞ=ℓ þ α scatt
where ασ is the absorption in units of cm−1. The measured absorption is plotted as a function of wave number (σ, cm−1) to get the absorption spectrum. The reflectivity of the mirrors and the scattering contributions only affect the baseline with the oxygen absorption band on top of it. LabVIEW software was used (National Instruments, version 13.0.1) to control the data acquisition, laser wavelength scanning, processing and displaying data. A National Instruments GPIB-USB-HS controller was used to interface the Lock-in amplifier to a PC via a USB connection. 3. Results and Discussion The rotational quantum numbers for the ground state (3Σg−) are designated as the rotational angular momentum (K) and the total angular momentum (J) that is the sum of the rotational (K) and spin (S) angular momentum. The ground state is split into three levels corresponding to J″ = K″, K″ + 1, and K″ − 1 with only odd values of K″ allowed. The upper electronic state (1Σg+) which has zero spin, is composed of single states with J′ = K′, where only even values of K′ are allowed. The only transitions considered here are the PP transitions that correspond to absorptions from the ground state rotational quantum number J″ = K″ = 1, 3, 5, 7…, with final states J′ = 0, 2, 4, 6…,
ð2Þ
Fig. 3. Phase shift CRD spectrum showing the PP and the PQ transitions at 298 K and 557 K. The PP peak transitions assigned are indicated.
Fig. 4. Comparison between absorption spectra at 298 K (red) and 373 K (blue) for PP transitions with J″ = 13 to 21.
248
M. Diaz et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 210 (2019) 245–250
respectively [17]. Fig. 3 depicts a low-resolution A band of oxygen showing the PP and PQ branches. The signals were obtained with the open path CRD technique at 298 K and at 550 K. A comparison between the spectra at 298 K (red) and 373 K (blue) is shown in Fig. 4 for PP transitions with J″ from 13 to 21. The temperature was measured with the sensor of the furnace. The selected PP transitions are indicated with a vertical line on top and the J″ assignment. The original spectra have been modified by normalizing the signal dividing the intensity at each wavelength by the maximum intensity in the original spectrum and by correcting the baseline to show zero baseline intensity. The baseline correction eliminates the effect of reflectivity of the mirrors and the scattering losses. Seven spectra were obtained at each temperature. The intensity of a rotational transition at a given temperature is given by an equation [18]: Iabs ¼
C abs σ J 0 þ J ″ þ 1 e− F J″ Qr
ðhc=kT Þ
ð3Þ
where Iabs is the absorption intensity, σ is the wavenumber at the peak absorption, Qr is the rotational partition function and FJ″ is the energy of the initial rotational state. Cabs is a constant depending on the change of the dipole moment and total population of the molecules in the initial vibrational level. For a rotational-vibrational band at a given temperature, the factor
C abs Qr
is very nearly constant [18]. The energy of the
lower state is given by the equation [17]: F J″ ¼ B″ J″
J″ þ 1 þ 2λ−γ
ð4Þ
The rotational constant of ground state is B″ = 1.4377 cm−1, λ = 1.984 cm−1 is the spin-spin coupling and γ = −0.084 cm−1 is the spin-rotation interaction constant. The linear fitting method was used to plot: ln ðσ ð J 0IþabsJ″ þ1ÞÞ versus FJ″. Fig. 5 shows the Boltzmann plots of A-band for 550 K and 373 K. Each experimental point in the plot is the average of the peak value obtained for seven normalized spectra. The best straight line was obtained using the transitions and energies beginning with the maximum of the distribution of the PP branch because they are more sensitive to the temperature change. For the 298 K temperature, the plot produces a straight line when using PP transitions originating from J″ = K″ = 3, 5, 7, 9, 11, 13, 15, and 17. The points corresponding to J″ = 1 and 3 depart from the straight line. For the 550 K transition the points selected were J″ = K″ = 7, 9, 11, 13, 15, 17 and 21. The slope (−1/kT) has units of cm. The temperature is calculated by taking the inverse of the slope and multiplying by 1.43879 K/cm−1. The calculated temperatures were (298 ± 14) K, (383 ± 25) and (557 ± 44) K. A summary of the results is presented on Table 1 where the temperatures indicated by the furnace and calculated from the spectra are shown. The assignments are indicated with lower level J″ and wavenumber (σ) of the peak. Peak intensities for the (PP) branch are an average given on the table for of several spectral peaks. The calculated temperatures are shown with their respective standard deviations. We observed that with the oven at a set temperature, an increase in the air flow rate produces a small increase in the baseline and a small decrease of the overall intensities because of scattering effects. After baseline correction and normalization of the signal, the distribution of intensities is similar at the same temperature. Because of the large optical pathlength of the cavity, the spectral bands have a signal to noise ratio N100. The contribution of instrumental noise can be illustrated with a discussion of propagation of errors. A treatment based on propagation of errors of measured intensities is usually very small compared with the calculated error in temperature obtained from the least square treatment of the data. The optical pathlength is 12 km with a time constant around 13 μs. The measured angle changes from 47° (baseline) to a minimum of 42° (highest peak). The lock-in amplifier can measure the angle with a precision of 0.01°. Assuming a very large fluctuation in the angle of 0.5° (not usual), the error in the tangent at θ = 45° is (Δ(tan θ)/
Fig. 5. Boltzmann plot of selected transitions from J″ = 7 to 21 for 550 K and 373 K. The linear fitting method was used to plot: lnðσ ð J0IþabsJ″ þ1ÞÞ versus FJ″.
Table 1 Observed peak assignments are indicated with the lower level J″, upper level J′ and the wavenumber (σ) of the peak. Lower state energy is FJ″. Peak intensities for the (PP) branch are an average of several spectral peaks. Temperature measured (furnace sensor) and calculated from the spectra. J″
J′
σ (cm−1)
FJ″(cm−1)
1 0 13,116.5 6.85 3 2 13,111.8 21.23 5 4 13,105.0 47.11 7 6 13,098.8 84.49 9 8 13,090.5 133.37 11 10 13,083.2 193.75 13 12 13,076.2 265.64 15 14 13,067.6 349.02 17 16 13,059.0 443.91 19 18 13,049.2 550.3 21 20 13,039.9 668.19 Calculated temperature (Boltzmann plot) Measured temperature (furnace sensor)
Iabs/10−6 cm−1 2.03 ± 0.02 2.49 ± 0.02 2.79 ± 0.01 2.38 ± 0.02 2.35 ± 0.02 1.64 ± 0.02 1.60 ± 0.03 1.51 ± 0.03 1.25 ± 0.03 0.79 ± 0.05 (298 ± 14) K (300 ± 5) K
2.10 ± 0.04 2.41 ± 0.03 2.74 ± 0.03 1.90 ± 0.04 1.79 ± 0.04 1.59 ± 0.05 1.36 ± 0.05 (557 ± 44) K (550 ± 5) K
M. Diaz et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 210 (2019) 245–250
tan θ) = 0.037. The value of (Δτ/τ) is approximately 0.037 with an insignificant error in the modulation frequency. For a time constant of 13 μs, the absorption Iabs is (2.56 ± 0.10) × 10−6 cm−1. Experimentally, we get that the maximum baseline noise is 0.05. The error associated with the natural logarithm is mainly (ΔIabs)/Iabs. This error is around 0.04. A typical point for the natural logarithm is (−25.45 ± 0.04). Considering that seven spectra were obtained and averaged at 550 K, calculated temperatures and errors were obtained from the least square treatment of the data. Fig. 5 shows the straight lines with error bars and the calculated error from dispersion of the data. Another spectroscopic method of temperature determination is the ratio of measured absorbance of two rotationally resolved transitions [1,19,20]. Usually the two transitions are selected based on absorption strength, spectral isolation, and temperature sensitivity to maximize the accuracy of temperature determinations. In comparison, the Boltzmann plot allows for a more accurate and precise temperature determination using more experimental points. In a previous study, we have shown that air flow temperatures can be determined below 373 K. In this study the temperatures at 298 K and 550 K have been measured indicating that the method could be used to determine temperatures of exhaust gases originating from combustion or other industrial processes. Similar studies using several techniques have been done in other laboratories using of water, carbon dioxide, and carbon monoxide absorption to determine temperatures of flowing gases [21–25]. The CRD method is very sensitive because of the kilometer pathlength that an optical cavity can afford. Improvements can be implemented using a high-resolution diode laser to resolve the peaks. The technique can be used for measurements of flame temperatures, but it cannot be used to measure temperatures inside a combustion chamber [26] because the complete absence of windows is required. We began this study to measure the temperature of the output air flow cooling the blades of a laboratory turbine. The evaluation of the temperature involving flowing air near the blades of gas turbines is an important parameter to evaluate the design of different blade cooling technologies. Gas turbine engines are used extensively around the world. Due to their favorable power to weight ratios, they have been the engine of choice for aircraft propulsion for decades. With the efficiency improvements that have been made to the engines, they have also taken over a substantial share of the land-based, power generation market. Increasing the operating temperature of the engine leads to increased power production while increasing the pressure ratio within the engine yields an increase in its thermal efficiency [27]. As a result, modern gas turbine engines are operating with mainstream temperatures exceeding 2000 K in the turbine section of the engine. To prevent the turbine blades and vanes from melting at these extreme temperatures, they must be actively cooled. At the increased pressure ratios of modern engines, sophisticated cooling schemes are required to effectively utilize the cooling air supplied at elevated temperatures (N950 K) and reduced flow rates. When implementing advanced cooling technology into the turbine blades and vanes, it is necessary to have an accurate prediction of the coolant to mainstream flow interaction around the airfoils; if the predicted metal temperature is off by only 30 K, the life of the component can be reduced by half. Therefore, engine designers require accurate and highly resolved temperature measurements of both the airfoil wall and the air temperature surrounding the blades and vanes. This coolant to mainstream interaction is modeled in the laboratory to match non-dimensional blowing ratios and density ratios with the actual engines. While many groups have experimentally obtained detailed distributions of the surface temperature for film cooled components, limited data exists to quantify the nearwall fluid temperature above the cooled airfoil. This work will allow researchers to investigate the thermal boundary layer development due to the interaction of the hot mainstream gas and cooling air. This will lead to a more fundamental understanding of how these flows mix in highly turbulent environments. These thermal flow field measurements will be coupled with surface temperature measurements and velocity
249
distributions within the flow to fully characterize mass, momentum, and energy transport in highly turbulent flows [28,29].
4. Conclusion The temperature of air flowing through a heated furnace tube has been determined using a near infrared laser and oxygen absorption in air. Experimental details related to the tube dimensions and the main elements of the optical cavity have been given. A metal absorption cell in a temperature-regulated tube furnace is placed at the center of an optical cavity. The main advantage of the method is that the optical pathlength is of the order of kilometers. The combination of an open optical cavity with highly reflecting mirrors and the phase shift cavity ring down technique ensures that strong signals will be obtained even from weak absorbers at low concentrations. Oxygen temperatures were determined after recording low-resolution spectra of the A-band of oxygen. To determine the temperature from the rotational transitions corresponding to the A-band of oxygen a Boltzmann plot was used. For this plot, we consider the PP (lines with odd values of rotational quantum numbers J″ = K) lines. Values of the of the J″ = 1–23 were selected based on their intensity and their sensitivity to the temperature change. One limitation of the technique is that the mirrors must be continuously flushed with pure nitrogen at room temperature. This is done to protect the mirrors from the high temperature of the air coming out of the furnace and any particles in the laboratory. The temperature measurement is not affected but the baseline increases due to beam scattering. Improvements can be implemented using a high-resolution diode laser to resolve the peaks. The technique can be used for measurements of temperature in a flame, but it cannot be used to measure temperatures inside a combustion chamber because the complete absence of windows is required. Like studies with other molecules, oxygen absorption is an excellent choice to determine the temperatures in a flow system. In our laboratory, the main application of the open-path PS-CRD absorption of oxygen is the possibility to measure the temperature of the output air flow cooling the blades of a laboratory turbine. This would be an added contribution for the development of efficient blade cooling technology.
Acknowledgement This work was supported by the Baylor University Research Committee and the Vice-Provost for research. References [1] T. Hieta, M. Merimaa, Spectroscopic measurement of air temperature, Int. J. Thermophys. 31 (2010) 1710–1718. [2] L.C. Philippe, R.K. Hanson, Laser diode wavelength-modulation spectroscopy for simultaneous measurement of temperature, pressure, and velocity in shock-heated oxygen flows, Appl. Opt. 32 (1993) 6090–6103. [3] J.E. Kalshoven Jr., C.L. Korb, G.K. Schwemmer, M. Dombrowski, Laser remote sensing of atmospheric temperature by observing resonant absorption of oxygen, Appl. Opt. 20 (1967) 1967–1971. [4] A. Goldman, I. Rahinov, S. Cheskis, Molecular oxygen detection in low pressure flames using cavity ring-down spectroscopy, Appl. Phys. B Lasers Opt. 82 (2006) 659–663. [5] M. Touzeau, M. Vialle, A. Zellagui, G. Goussett, M. Lefebvre, M. Pealat, Spectroscopic temperature measurements in oxygen discharges, J. Phys. D. Appl. Phys. 24 (1991) 41–47. [6] P. Macko, P. Veis, Time resolved O2 (b1Σ+ g ) rotational temperature measurements in a low-pressure oxygen pulsed discharge. Simple and quick method for temperature determination, J. Phys. D Appl. Phys. 32 (1999) 246–250. [7] R.S. Hickman, L. Liang, A comparison of temperature measurement of oxygen in the 0-0 and 0-1 rotational bands using laser Raman spectroscopy, Rev. Sci. Instrum. 45 (1974) 1580–1582. [8] T.R. Meyer, S. Roy, R.P. Lucht, J.R. Gord, Dual-pump dual-broadband CARS for exhaust gas temperature and CO2-O2-N2 mole-fraction measurements in model gasturbine combustors, Combust. Flame 142 (2005) 52–61. [9] Y. Wu, Z. Zhang, S.F. Adams, O2 rotational temperature measurements by coherent microwave scattering, Chem. Phys. Lett. 513 (2011) 191–194.
250
M. Diaz et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 210 (2019) 245–250
[10] S.F. Adams, Y. Wu, Z. Zhang, Oxygen rotational temperature determination using empirical analyses of C3Π (v′ = 2) ← X3Σ (v″ = 0) transitions, Appl. Spectrosc. 69 (2015) 1036–1041. [11] Y. Wu, Z. Zhang, S.F. Adams, Temperature sensitivity of molecular oxygen resonantenhanced multiphoton ionization spectra involving the C3Πg intermediate state, Appl. Phys. B 122 (2016) 149 (1–10). [12] Y. Perez-Delgado, C.E. Manzanares, Low temperature cavity ring down spectroscopy with off-axis alignment: application to the A- and γ-bands of O2 in the visible at 90 K, Appl. Phys. B Lasers Opt. 106 (2012) 971–978. [13] P. Nyaupane, Y. Perez-Delgado, D. Camejo, L.M. Wright, C.E. Manzanares, Cavity ring down absorption of O2 in air as a temperature sensor for an open and a cryogenic optical cavity, Appl. Spectrosc. 71 (2017) 847–855. [14] E.K. Lewis, C.J. Moehnke, J. Navea, C.E. Manzanares, Low temperature cell for cavity ring down absorption studies, Rev. Sci. Instrum. 77 (2006) (073107-1-073107-9). [15] J.M. Herbelin, J.A. McKay, M.A. Kwok, R.H. Ueunten, D.S. Urevig, D.J. Spencer, D.J. Benard, Sensitive measurement of photon lifetime and true reflectances in an optical cavity by a phase-shift method, Appl. Opt. 19 (1980) 144–147. [16] R. Engeln, G. von Helden, G. Berden, G. Meijer, Phase shift cavity ring down absorption spectroscopy, Chem. Phys. Lett. 262 (1996) 105–109. [17] J.H. Miller, R.W. Boese, L.P. Giver, Intensity measurements and rotational intensity distribution for the oxygen A-band, J. Quant. Spectrosc. Radiat. Transf. 9 (1969) 1507–1517. [18] G. Herzberg, Spectra of Diatomic Molecules, 2nd ed. Van Nostrand, New York, 1950 126. [19] F.B. Wampler, R.A. Gentry, Determination of gas temperature and density by measuring ultraviolet gas absorption at two wavelengths, J. Appl. Phys. 52 (1981) 1583. [20] W.W. Rice, R.C. Oldenborg, F.B. Wampler, J.J. Tiee, Gas temperature and density of UF6 determined by two-wavelength UV absorption, Appl. Opt. 20 (1981) 2625–2629.
[21] X. Zhou, X. Liu, J.B. Jeffries, R.K. Hanson, Development of a sensor for temperature and water concentration in combustion gases using a single tunable diode laser, Meas. Sci. Technol. 14 (2003) 1459–1468. [22] J.T.C. Liu, G.B. Rieker, J.B. Jeffries, M.R. Gruber, C.D. Carter, T. Mathur, R.K. Hanson, Near-infrared diode laser absorption diagnostic for temperature and water vapor in a scramjet combustor, Appl. Opt. 44 (2005) 6701–6711. [23] T. Werblinski, F. Mittmann, M. Altenhoff, T. Seeger, L. Zigan, S. Will, Temperature and water mole fraction measurements by time-domain-based supercontinuum absorption spectroscopy in a flame, Appl. Phys. B Lasers Opt. 118 (2015) 153–158. [24] A. Farooq, J.B. Jeffries, R. Hanson, CO2 concentration and temperature sensor for combustion gases using diode-laser absorption near 2.7 μm, Appl. Phys. B Lasers Opt. 90 (2008) 619–628. [25] W. Ren, A. Farooq, D.F. Davidson, R.K. Hanson, CO concentration and temperature sensor for combustion gases using quantum-cascade laser absorption near 4.7 μm, Appl. Phys. B Lasers Opt. 107 (2012) 849–860. [26] M.P. Thariyan, A.H. Bhuiyan, S.E. Meyer, S.V. Naik, J.P. Gore, R.P. Lucht, Dual-pump coherent anti-Stokes Raman scattering system for temperature and species measurements in an optically accessible high-pressure gas turbine combustor facility, Meas. Sci. Technol. 22 (2011), 015301. (1–12). [27] J.C. Han, S. Dutta, S. Ekkad, Gas Turbine Heat Transfer and Cooling Technology, Taylor and Francis, New York, 2000. [28] L.M. Wright, S.T. McClain, M.D. Clemenson, Effect of freestream turbulence on film cooling jet structure and surface effectiveness using PIV and PSP, ASME J. Turbomach. 133 (2011), 041023. (1–12). [29] L.M. Wright, S.T. McClain, M.D. Clemenson, Effect of density ratio on flat plate film cooling with shaped holes using PSP, ASME J. Turbomach. 133 (2011), 041011. (1–11).
Contents lists available at ScienceDirect
Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy journal homepage: www.elsevier.com/locate/saa
Description of an air temperature sensor based on O2 absorption spectroscopy Marlon Diaz, Parashu R. Nyaupane, Daniel Monteagudo, Carlos E. Manzanares ⁎ Department of Chemistry & Biochemistry, Baylor University, Waco, TX 76798, United States of America
a r t i c l e
i n f o
Article history: Received 28 June 2018 Received in revised form 4 October 2018 Accepted 13 November 2018 Available online 13 November 2018 Keywords: Temperature sensor Cavity ring down Flowing air Oxygen A-band Optical cavity
a b s t r a c t In this paper, we describe in detail an assembled open-path optical cavity to act as a temperature sensor of air. A metal absorption cell in a temperature-regulated tube furnace is placed at the center of an optical cavity. The optical cavity consists of two mirrors, two nitrogen buffer sleeves, and the open cell. The air is injected through a fitting in one extreme of the metal tube and travels half the tube length through a channel in the wall of the tube. The channel directs the air towards the center of the cell. The air flowing is heated at the temperature of the metal tube in contact with the furnace. The heated air injected at the center of the tube, flows towards the open extremes of the tube. The nitrogen buffer sleeves protect the mirrors from the heated air. The temperature of the air flowing through the tube is determined by measuring the absorption of the A band of oxygen as a function of the wavenumber in the 769–755 nm wavelength range. The absorption technique is phase-shift cavity ring down spectroscopy. To obtain the temperature, the energy of the lower rotational state for eleven selected rotational transitions is linearly fitted to a logarithmic function that contains the relative intensity of the rotational transition, the initial and final rotational quantum numbers and the energy of the transition. Accuracy of the measurement is determined by comparing the calculated temperature from the spectra with the analog reading of the temperature-regulated tube furnace. This technique is proposed for exhaust air temperature measurements of combustion chambers and cooling air after passing through the blades of a turbine. © 2018 Published by Elsevier B.V.
1. Introduction Determination of temperature based on measurements of oxygen rotational transitions around 760 nm have been reported using absorption [1–4], emission [5,6], Raman [7,8], and microwave Rayleigh [9–11] techniques. Oxygen absorption in air for a 67 m path length has been obtained for ambient temperature determinations using a diode laser. The ratio of two selected oxygen lines (two-line thermometry) was used to determine the temperature [1]. In another experiment the rotational line absorption of molecular oxygen is obtained in a shock-heated oxygen flow. The absorption intensities of five rotational lines around 760 nm are fitted to an empirical function that includes a line profile function, the absorption frequency, the temperature, and the pressure. The fitting function gives the temperature and pressure of the sample [2]. Laser remote atmospheric temperature has been obtained by observing resonant absorption of oxygen. The calculation of temperature is done with an iterative equation that uses the measured absorption of one representative oxygen line [3]. The oxygen absorption around 760 nm has been obtained using cavity ring down spectroscopy [CRD] to determine oxygen concentrations in a low-pressure flame [4]. In ⁎ Corresponding author. E-mail address: [email protected] (C.E. Manzanares).
https://doi.org/10.1016/j.saa.2018.11.034 1386-1425/© 2018 Published by Elsevier B.V.
emission experiments, rotational temperatures have been obtained from the intensity distribution of the excited state emission of the Aband around 760 nm in an oxygen glow discharge. The temperatures are obtained and compared with temperatures from coherent antiStokes Raman spectroscopy of oxygen molecules in the ground state [5]. The time evolution of the O2 rotational temperature has been estimated for light emitted around 760 nm after a pulsed discharge by comparison with calculated low-resolution spectra for different gas pressures [6]. In Raman scattering experiments, rotational temperatures have been determined in pure O2 at 1 atm for temperatures from 243 to 343 K and in air at 298 K using Raman spectroscopy in an intracavity cell. The temperature was calculated from the Boltzmann equation as well as intensities and frequencies of the transitions [7]. Coherent anti-Stokes Raman scattering experiments have also been reported. Dual-pump coherent anti-Stokes Raman scattering system has been done for temperature and species (CO2, O2, N2) measurements [8]. Another technique for temperature determinations is the coherent microwave Rayleigh scattering from a plasma created by resonance enhanced multi-photon ionization (REMPI) [9–11]. This technique uses a frequency doubled dye laser whose output is focused on an air or oxygen sample to produce a plasma that is irradiated with a microwave source. The microwave horn that acts as source and receiver, detects the scattered microwave radiation from the plasma. The microwave signal
246
M. Diaz et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 210 (2019) 245–250
is input into an automatic data acquisition program that records the rotational spectrum of oxygen as the laser is tuned. The cell is at the temperature of a tube furnace. The rotational temperatures are determined from Boltzmann plots for rotational transitions and compared with the tube furnace temperature [9]. In subsequent applications of the technique, the temperatures are obtained with empirical equations without the use of high resolution rotational transitions using linewidth, linear, and area fitting of selected bands [10,11]. Our group has used phase shift cavity ring down spectroscopy and oxygen absorption (γ and A bands) to obtain temperatures between 90 K and 296 K in a cryogenic cell [12]. We have used the same technique to obtain temperatures between 298 K and 373 K with a heated open glass tube [13]. In this experiment, we describe in detail an open metal absorption cell inside a temperature-regulated tube furnace, placed at the center of an optical cavity. The temperature of the air sample is obtained from the near-IR absorption of oxygen (A-band absorption) using the Phase Shift Cavity Ring Down technique (PS-CRD). The accuracy of the absorption-based temperature calculation is determined by comparing with the temperature of the furnace. This oxygen absorption-based gas temperature sensor as well as, the other spectroscopic techniques mentioned before, can be applied to determine the temperature in atmospheric air, combustion chambers, flames, discharge tubes, and cold plasmas. We are using this absorption method to obtain the temperatures of cooling air at the input and hot air at the output after passing through the blades of a turbine. These measurements could contribute to the development of more efficient cooling technologies, as well as leading to increased efficiency of the gas turbine engine. 2. Experimental 2.1. Open Cell and Optical Cavity We assembled an open-path optical cavity to act as a temperature sensor of air. An experimental setup of the high temperature system is shown in Fig. 1 (a). A metal absorption cell placed in a temperature regulated furnace tube is at the center of an optical cavity. The optical cavity consists of two mirrors, two nitrogen buffer sleeves, and the open cell. Fig. 1 (b) shows details of the tube inside the furnace. The tube is 41 cm long and the outside diameter is 4.5 cm. The wall thickness is 1.3 cm and the inside diameter is 1.9 cm. The sample from a pressurized
air tank (Praxair) is injected through a Swagelok fitting in a 0.32 cm diameter, 20.5 cm length opening through the wall that directs the air towards the center of the tube. The air flowing is heated at the temperature of the metal tube in contact with the furnace. At the center, the hot air comes out of a small opening and flows towards both extremes the tube. The flowing hot air is not allowed to get close to the mirrors by having a room temperature flow of nitrogen gas injected on both sides of the heated cell using PVC pipes of the same diameter of the cell tube. The optical cavity is completed with two highly reflective mirrors (Los Gatos Research, 99.997%) which are mounted parallel to each other and separated by 93 cm. The air flow at the center of the tube was controlled by a flow meter (Cole Parmer, 014-96-ST). The temperature-regulated furnace provides the temperature to compare with the calculated temperature using the oxygen absorption method. 2.2. Cavity Ring Down Spectrometer The experimental set up for the PS-CRD technique has been given in a previous publication [14]. The complete experimental set up for air temperature measurements is illustrated in Fig. 2. Briefly, a laser system consisting of a continuous wave Ti-Sapphire laser (Coherent 899 ring laser) pumped by a Nd-YVO4 laser (Coherent Verdi) in the range 13,000–13,250 cm−1 was used. Calibration of the laser was done obtaining the opto-galvanic spectrum of a hollow cathode lamp filled with neon. The resolution of the pump laser was 0.17 cm−1. A typical spectrum recorded 3174 points (0.064 cm−1/step) during a scan of around 3600 s. An electro optic modulator (Conoptics model 350-50) was used to modulate the laser beam as a 12 kHz square wave. The modulated beam passes through the optical cavity and is detected with a photomultiplier (Oriel model 70680). The zero phase-shift is obtained by bypassing the cavity. The reference signal for the lock-in amplifier (Stanford Research Systems (SRS) model 830) is obtained from a function generator (SRS model DS335) which also supplies the modulation signal to the driver input of the electro-optic modulator. The phase difference between the signal exiting the cavity and the signal entering the cavity is measured with a dual phase lock-in amplifier. For the measurement of the phase shift angle, the laser is tuned to a wavelength where there is no absorption. Fig. 3 also depicts the phase angle measurement set up. The modulated laser beam bypasses the optical cavity (dashed line) using two movable mirrors and reaches the detector. The strong laser signal is sent to a pre-amplifier (SRS-SR560). This signal is
Fig. 1. (a) Expanded view of an open path optical cavity with highly reflective mirrors. The optical cavity consists of two mirrors, two nitrogen buffer sleeves, and an open tube inside a furnace. (b) Shows dimensions and details of the tube inside the furnace.
M. Diaz et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 210 (2019) 245–250
247
Fig. 2. Phase shift cavity ring down spectrometer used to record the oxygen A-band absorption in an open optical cavity. The reference beam bypassing the cavity is shown.
not amplified (×1) but it is sent to the lock in amplifier. The angle between the reference from the function generator and the bypassed signal is automatically made equal to zero with the lock-in amplifier. When the diverting mirrors are removed, the laser passes through the cavity. The weak signal exiting the cavity is pre-amplified (×1000) and sent to the lock in amplifier that shows the angle. The measured value is the phase shift angle due to delay inside the cavity. The phase shift (ϕ) is related to the time that the light spends in the cavity or ringdown time (τ) by the equation: tanϕ = 2πfτ, where f is the modulation frequency [15,16]. Optimal alignment of the cavity mirrors occurs when the phase shift is around −45° that is a region where the tangent function is linear. In a typical experiment the initial phase shift angle is close to −47°, the presence of the flowing air inside the cavity usually changes the angle to −45° due to scattering. The phase angle is recorded as a function of wave number (σ). For a cavity filled with an absorbing gas the ring-down time is: ℓ τσ ¼ c ð1−RÞ þ α O2 ℓ þ α scatt: ℓ
ð1Þ
The time (τσ) is related to the reflectivity of the mirrors (R), the speed of light (c), the absorption coefficient (αO 2) of the sample and the length (ℓ) of the optical cavity. The scattering losses are represented by αscatt. The observed signal is: α σ ¼ ½c τ σ −1 ¼ α O2 þ ð1−RÞ=ℓ þ α scatt
where ασ is the absorption in units of cm−1. The measured absorption is plotted as a function of wave number (σ, cm−1) to get the absorption spectrum. The reflectivity of the mirrors and the scattering contributions only affect the baseline with the oxygen absorption band on top of it. LabVIEW software was used (National Instruments, version 13.0.1) to control the data acquisition, laser wavelength scanning, processing and displaying data. A National Instruments GPIB-USB-HS controller was used to interface the Lock-in amplifier to a PC via a USB connection. 3. Results and Discussion The rotational quantum numbers for the ground state (3Σg−) are designated as the rotational angular momentum (K) and the total angular momentum (J) that is the sum of the rotational (K) and spin (S) angular momentum. The ground state is split into three levels corresponding to J″ = K″, K″ + 1, and K″ − 1 with only odd values of K″ allowed. The upper electronic state (1Σg+) which has zero spin, is composed of single states with J′ = K′, where only even values of K′ are allowed. The only transitions considered here are the PP transitions that correspond to absorptions from the ground state rotational quantum number J″ = K″ = 1, 3, 5, 7…, with final states J′ = 0, 2, 4, 6…,
ð2Þ
Fig. 3. Phase shift CRD spectrum showing the PP and the PQ transitions at 298 K and 557 K. The PP peak transitions assigned are indicated.
Fig. 4. Comparison between absorption spectra at 298 K (red) and 373 K (blue) for PP transitions with J″ = 13 to 21.
248
M. Diaz et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 210 (2019) 245–250
respectively [17]. Fig. 3 depicts a low-resolution A band of oxygen showing the PP and PQ branches. The signals were obtained with the open path CRD technique at 298 K and at 550 K. A comparison between the spectra at 298 K (red) and 373 K (blue) is shown in Fig. 4 for PP transitions with J″ from 13 to 21. The temperature was measured with the sensor of the furnace. The selected PP transitions are indicated with a vertical line on top and the J″ assignment. The original spectra have been modified by normalizing the signal dividing the intensity at each wavelength by the maximum intensity in the original spectrum and by correcting the baseline to show zero baseline intensity. The baseline correction eliminates the effect of reflectivity of the mirrors and the scattering losses. Seven spectra were obtained at each temperature. The intensity of a rotational transition at a given temperature is given by an equation [18]: Iabs ¼
C abs σ J 0 þ J ″ þ 1 e− F J″ Qr
ðhc=kT Þ
ð3Þ
where Iabs is the absorption intensity, σ is the wavenumber at the peak absorption, Qr is the rotational partition function and FJ″ is the energy of the initial rotational state. Cabs is a constant depending on the change of the dipole moment and total population of the molecules in the initial vibrational level. For a rotational-vibrational band at a given temperature, the factor
C abs Qr
is very nearly constant [18]. The energy of the
lower state is given by the equation [17]: F J″ ¼ B″ J″
J″ þ 1 þ 2λ−γ
ð4Þ
The rotational constant of ground state is B″ = 1.4377 cm−1, λ = 1.984 cm−1 is the spin-spin coupling and γ = −0.084 cm−1 is the spin-rotation interaction constant. The linear fitting method was used to plot: ln ðσ ð J 0IþabsJ″ þ1ÞÞ versus FJ″. Fig. 5 shows the Boltzmann plots of A-band for 550 K and 373 K. Each experimental point in the plot is the average of the peak value obtained for seven normalized spectra. The best straight line was obtained using the transitions and energies beginning with the maximum of the distribution of the PP branch because they are more sensitive to the temperature change. For the 298 K temperature, the plot produces a straight line when using PP transitions originating from J″ = K″ = 3, 5, 7, 9, 11, 13, 15, and 17. The points corresponding to J″ = 1 and 3 depart from the straight line. For the 550 K transition the points selected were J″ = K″ = 7, 9, 11, 13, 15, 17 and 21. The slope (−1/kT) has units of cm. The temperature is calculated by taking the inverse of the slope and multiplying by 1.43879 K/cm−1. The calculated temperatures were (298 ± 14) K, (383 ± 25) and (557 ± 44) K. A summary of the results is presented on Table 1 where the temperatures indicated by the furnace and calculated from the spectra are shown. The assignments are indicated with lower level J″ and wavenumber (σ) of the peak. Peak intensities for the (PP) branch are an average given on the table for of several spectral peaks. The calculated temperatures are shown with their respective standard deviations. We observed that with the oven at a set temperature, an increase in the air flow rate produces a small increase in the baseline and a small decrease of the overall intensities because of scattering effects. After baseline correction and normalization of the signal, the distribution of intensities is similar at the same temperature. Because of the large optical pathlength of the cavity, the spectral bands have a signal to noise ratio N100. The contribution of instrumental noise can be illustrated with a discussion of propagation of errors. A treatment based on propagation of errors of measured intensities is usually very small compared with the calculated error in temperature obtained from the least square treatment of the data. The optical pathlength is 12 km with a time constant around 13 μs. The measured angle changes from 47° (baseline) to a minimum of 42° (highest peak). The lock-in amplifier can measure the angle with a precision of 0.01°. Assuming a very large fluctuation in the angle of 0.5° (not usual), the error in the tangent at θ = 45° is (Δ(tan θ)/
Fig. 5. Boltzmann plot of selected transitions from J″ = 7 to 21 for 550 K and 373 K. The linear fitting method was used to plot: lnðσ ð J0IþabsJ″ þ1ÞÞ versus FJ″.
Table 1 Observed peak assignments are indicated with the lower level J″, upper level J′ and the wavenumber (σ) of the peak. Lower state energy is FJ″. Peak intensities for the (PP) branch are an average of several spectral peaks. Temperature measured (furnace sensor) and calculated from the spectra. J″
J′
σ (cm−1)
FJ″(cm−1)
1 0 13,116.5 6.85 3 2 13,111.8 21.23 5 4 13,105.0 47.11 7 6 13,098.8 84.49 9 8 13,090.5 133.37 11 10 13,083.2 193.75 13 12 13,076.2 265.64 15 14 13,067.6 349.02 17 16 13,059.0 443.91 19 18 13,049.2 550.3 21 20 13,039.9 668.19 Calculated temperature (Boltzmann plot) Measured temperature (furnace sensor)
Iabs/10−6 cm−1 2.03 ± 0.02 2.49 ± 0.02 2.79 ± 0.01 2.38 ± 0.02 2.35 ± 0.02 1.64 ± 0.02 1.60 ± 0.03 1.51 ± 0.03 1.25 ± 0.03 0.79 ± 0.05 (298 ± 14) K (300 ± 5) K
2.10 ± 0.04 2.41 ± 0.03 2.74 ± 0.03 1.90 ± 0.04 1.79 ± 0.04 1.59 ± 0.05 1.36 ± 0.05 (557 ± 44) K (550 ± 5) K
M. Diaz et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 210 (2019) 245–250
tan θ) = 0.037. The value of (Δτ/τ) is approximately 0.037 with an insignificant error in the modulation frequency. For a time constant of 13 μs, the absorption Iabs is (2.56 ± 0.10) × 10−6 cm−1. Experimentally, we get that the maximum baseline noise is 0.05. The error associated with the natural logarithm is mainly (ΔIabs)/Iabs. This error is around 0.04. A typical point for the natural logarithm is (−25.45 ± 0.04). Considering that seven spectra were obtained and averaged at 550 K, calculated temperatures and errors were obtained from the least square treatment of the data. Fig. 5 shows the straight lines with error bars and the calculated error from dispersion of the data. Another spectroscopic method of temperature determination is the ratio of measured absorbance of two rotationally resolved transitions [1,19,20]. Usually the two transitions are selected based on absorption strength, spectral isolation, and temperature sensitivity to maximize the accuracy of temperature determinations. In comparison, the Boltzmann plot allows for a more accurate and precise temperature determination using more experimental points. In a previous study, we have shown that air flow temperatures can be determined below 373 K. In this study the temperatures at 298 K and 550 K have been measured indicating that the method could be used to determine temperatures of exhaust gases originating from combustion or other industrial processes. Similar studies using several techniques have been done in other laboratories using of water, carbon dioxide, and carbon monoxide absorption to determine temperatures of flowing gases [21–25]. The CRD method is very sensitive because of the kilometer pathlength that an optical cavity can afford. Improvements can be implemented using a high-resolution diode laser to resolve the peaks. The technique can be used for measurements of flame temperatures, but it cannot be used to measure temperatures inside a combustion chamber [26] because the complete absence of windows is required. We began this study to measure the temperature of the output air flow cooling the blades of a laboratory turbine. The evaluation of the temperature involving flowing air near the blades of gas turbines is an important parameter to evaluate the design of different blade cooling technologies. Gas turbine engines are used extensively around the world. Due to their favorable power to weight ratios, they have been the engine of choice for aircraft propulsion for decades. With the efficiency improvements that have been made to the engines, they have also taken over a substantial share of the land-based, power generation market. Increasing the operating temperature of the engine leads to increased power production while increasing the pressure ratio within the engine yields an increase in its thermal efficiency [27]. As a result, modern gas turbine engines are operating with mainstream temperatures exceeding 2000 K in the turbine section of the engine. To prevent the turbine blades and vanes from melting at these extreme temperatures, they must be actively cooled. At the increased pressure ratios of modern engines, sophisticated cooling schemes are required to effectively utilize the cooling air supplied at elevated temperatures (N950 K) and reduced flow rates. When implementing advanced cooling technology into the turbine blades and vanes, it is necessary to have an accurate prediction of the coolant to mainstream flow interaction around the airfoils; if the predicted metal temperature is off by only 30 K, the life of the component can be reduced by half. Therefore, engine designers require accurate and highly resolved temperature measurements of both the airfoil wall and the air temperature surrounding the blades and vanes. This coolant to mainstream interaction is modeled in the laboratory to match non-dimensional blowing ratios and density ratios with the actual engines. While many groups have experimentally obtained detailed distributions of the surface temperature for film cooled components, limited data exists to quantify the nearwall fluid temperature above the cooled airfoil. This work will allow researchers to investigate the thermal boundary layer development due to the interaction of the hot mainstream gas and cooling air. This will lead to a more fundamental understanding of how these flows mix in highly turbulent environments. These thermal flow field measurements will be coupled with surface temperature measurements and velocity
249
distributions within the flow to fully characterize mass, momentum, and energy transport in highly turbulent flows [28,29].
4. Conclusion The temperature of air flowing through a heated furnace tube has been determined using a near infrared laser and oxygen absorption in air. Experimental details related to the tube dimensions and the main elements of the optical cavity have been given. A metal absorption cell in a temperature-regulated tube furnace is placed at the center of an optical cavity. The main advantage of the method is that the optical pathlength is of the order of kilometers. The combination of an open optical cavity with highly reflecting mirrors and the phase shift cavity ring down technique ensures that strong signals will be obtained even from weak absorbers at low concentrations. Oxygen temperatures were determined after recording low-resolution spectra of the A-band of oxygen. To determine the temperature from the rotational transitions corresponding to the A-band of oxygen a Boltzmann plot was used. For this plot, we consider the PP (lines with odd values of rotational quantum numbers J″ = K) lines. Values of the of the J″ = 1–23 were selected based on their intensity and their sensitivity to the temperature change. One limitation of the technique is that the mirrors must be continuously flushed with pure nitrogen at room temperature. This is done to protect the mirrors from the high temperature of the air coming out of the furnace and any particles in the laboratory. The temperature measurement is not affected but the baseline increases due to beam scattering. Improvements can be implemented using a high-resolution diode laser to resolve the peaks. The technique can be used for measurements of temperature in a flame, but it cannot be used to measure temperatures inside a combustion chamber because the complete absence of windows is required. Like studies with other molecules, oxygen absorption is an excellent choice to determine the temperatures in a flow system. In our laboratory, the main application of the open-path PS-CRD absorption of oxygen is the possibility to measure the temperature of the output air flow cooling the blades of a laboratory turbine. This would be an added contribution for the development of efficient blade cooling technology.
Acknowledgement This work was supported by the Baylor University Research Committee and the Vice-Provost for research. References [1] T. Hieta, M. Merimaa, Spectroscopic measurement of air temperature, Int. J. Thermophys. 31 (2010) 1710–1718. [2] L.C. Philippe, R.K. Hanson, Laser diode wavelength-modulation spectroscopy for simultaneous measurement of temperature, pressure, and velocity in shock-heated oxygen flows, Appl. Opt. 32 (1993) 6090–6103. [3] J.E. Kalshoven Jr., C.L. Korb, G.K. Schwemmer, M. Dombrowski, Laser remote sensing of atmospheric temperature by observing resonant absorption of oxygen, Appl. Opt. 20 (1967) 1967–1971. [4] A. Goldman, I. Rahinov, S. Cheskis, Molecular oxygen detection in low pressure flames using cavity ring-down spectroscopy, Appl. Phys. B Lasers Opt. 82 (2006) 659–663. [5] M. Touzeau, M. Vialle, A. Zellagui, G. Goussett, M. Lefebvre, M. Pealat, Spectroscopic temperature measurements in oxygen discharges, J. Phys. D. Appl. Phys. 24 (1991) 41–47. [6] P. Macko, P. Veis, Time resolved O2 (b1Σ+ g ) rotational temperature measurements in a low-pressure oxygen pulsed discharge. Simple and quick method for temperature determination, J. Phys. D Appl. Phys. 32 (1999) 246–250. [7] R.S. Hickman, L. Liang, A comparison of temperature measurement of oxygen in the 0-0 and 0-1 rotational bands using laser Raman spectroscopy, Rev. Sci. Instrum. 45 (1974) 1580–1582. [8] T.R. Meyer, S. Roy, R.P. Lucht, J.R. Gord, Dual-pump dual-broadband CARS for exhaust gas temperature and CO2-O2-N2 mole-fraction measurements in model gasturbine combustors, Combust. Flame 142 (2005) 52–61. [9] Y. Wu, Z. Zhang, S.F. Adams, O2 rotational temperature measurements by coherent microwave scattering, Chem. Phys. Lett. 513 (2011) 191–194.
250
M. Diaz et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 210 (2019) 245–250
[10] S.F. Adams, Y. Wu, Z. Zhang, Oxygen rotational temperature determination using empirical analyses of C3Π (v′ = 2) ← X3Σ (v″ = 0) transitions, Appl. Spectrosc. 69 (2015) 1036–1041. [11] Y. Wu, Z. Zhang, S.F. Adams, Temperature sensitivity of molecular oxygen resonantenhanced multiphoton ionization spectra involving the C3Πg intermediate state, Appl. Phys. B 122 (2016) 149 (1–10). [12] Y. Perez-Delgado, C.E. Manzanares, Low temperature cavity ring down spectroscopy with off-axis alignment: application to the A- and γ-bands of O2 in the visible at 90 K, Appl. Phys. B Lasers Opt. 106 (2012) 971–978. [13] P. Nyaupane, Y. Perez-Delgado, D. Camejo, L.M. Wright, C.E. Manzanares, Cavity ring down absorption of O2 in air as a temperature sensor for an open and a cryogenic optical cavity, Appl. Spectrosc. 71 (2017) 847–855. [14] E.K. Lewis, C.J. Moehnke, J. Navea, C.E. Manzanares, Low temperature cell for cavity ring down absorption studies, Rev. Sci. Instrum. 77 (2006) (073107-1-073107-9). [15] J.M. Herbelin, J.A. McKay, M.A. Kwok, R.H. Ueunten, D.S. Urevig, D.J. Spencer, D.J. Benard, Sensitive measurement of photon lifetime and true reflectances in an optical cavity by a phase-shift method, Appl. Opt. 19 (1980) 144–147. [16] R. Engeln, G. von Helden, G. Berden, G. Meijer, Phase shift cavity ring down absorption spectroscopy, Chem. Phys. Lett. 262 (1996) 105–109. [17] J.H. Miller, R.W. Boese, L.P. Giver, Intensity measurements and rotational intensity distribution for the oxygen A-band, J. Quant. Spectrosc. Radiat. Transf. 9 (1969) 1507–1517. [18] G. Herzberg, Spectra of Diatomic Molecules, 2nd ed. Van Nostrand, New York, 1950 126. [19] F.B. Wampler, R.A. Gentry, Determination of gas temperature and density by measuring ultraviolet gas absorption at two wavelengths, J. Appl. Phys. 52 (1981) 1583. [20] W.W. Rice, R.C. Oldenborg, F.B. Wampler, J.J. Tiee, Gas temperature and density of UF6 determined by two-wavelength UV absorption, Appl. Opt. 20 (1981) 2625–2629.
[21] X. Zhou, X. Liu, J.B. Jeffries, R.K. Hanson, Development of a sensor for temperature and water concentration in combustion gases using a single tunable diode laser, Meas. Sci. Technol. 14 (2003) 1459–1468. [22] J.T.C. Liu, G.B. Rieker, J.B. Jeffries, M.R. Gruber, C.D. Carter, T. Mathur, R.K. Hanson, Near-infrared diode laser absorption diagnostic for temperature and water vapor in a scramjet combustor, Appl. Opt. 44 (2005) 6701–6711. [23] T. Werblinski, F. Mittmann, M. Altenhoff, T. Seeger, L. Zigan, S. Will, Temperature and water mole fraction measurements by time-domain-based supercontinuum absorption spectroscopy in a flame, Appl. Phys. B Lasers Opt. 118 (2015) 153–158. [24] A. Farooq, J.B. Jeffries, R. Hanson, CO2 concentration and temperature sensor for combustion gases using diode-laser absorption near 2.7 μm, Appl. Phys. B Lasers Opt. 90 (2008) 619–628. [25] W. Ren, A. Farooq, D.F. Davidson, R.K. Hanson, CO concentration and temperature sensor for combustion gases using quantum-cascade laser absorption near 4.7 μm, Appl. Phys. B Lasers Opt. 107 (2012) 849–860. [26] M.P. Thariyan, A.H. Bhuiyan, S.E. Meyer, S.V. Naik, J.P. Gore, R.P. Lucht, Dual-pump coherent anti-Stokes Raman scattering system for temperature and species measurements in an optically accessible high-pressure gas turbine combustor facility, Meas. Sci. Technol. 22 (2011), 015301. (1–12). [27] J.C. Han, S. Dutta, S. Ekkad, Gas Turbine Heat Transfer and Cooling Technology, Taylor and Francis, New York, 2000. [28] L.M. Wright, S.T. McClain, M.D. Clemenson, Effect of freestream turbulence on film cooling jet structure and surface effectiveness using PIV and PSP, ASME J. Turbomach. 133 (2011), 041023. (1–12). [29] L.M. Wright, S.T. McClain, M.D. Clemenson, Effect of density ratio on flat plate film cooling with shaped holes using PSP, ASME J. Turbomach. 133 (2011), 041011. (1–11).