Quantitative and Spatially Resolved Measurement of Atomic Potassium in Combustion Using Diode Laser

Qiang Gao(高强)$^{1}$, Wu-Bin Weng(翁武斌)$^{2}$, Bo Li(李博)$^{1\hyperlink{s}{**}}$, Zhong-Shan Li(李中山)$^{1,2}$

doi:10.1088/0256-307X/35/2/024202

⇦Chinese Physics Letters, 2018, Vol. 35, No. 2, Article code 024202⇨ Quantitative and Spatially Resolved Measurement of Atomic Potassium in Combustion Using Diode Laser * Qiang Gao(高强)1, Wu-Bin Weng(翁武斌)2, Bo Li(李博)1**, Zhong-Shan Li(李中山)1,2 Affiliations 1State Key Laboratory of Engines, Tianjin University, Tianjin 300072 2Division of Combustion Physics, Lund University, Lund SE-22100, Sweden Received 10 October 2017 ^*Supported by the National Natural Science Foundation of China under Grant Nos 91541119 and 91541203.
^**Corresponding author. Email: boli@tju.edu.cn Citation Text: Gao Q, Weng W B, Li B and Li Z S 2018 Chin. Phys. Lett. 35 024202 Abstract A compact optical setup for quantitative and spatially resolved measurement of atomic alkali concentration in combustion is demonstrated. Tunable diode laser absorption spectroscopy and laser-induced fluorescence are combined using a single continuous wave diode laser to measure the line-integration concentration and the relative distribution simultaneously, thereby obtaining the absolute concentration distribution along the laser beam. The results indicate the good performance of this method for one-dimensional quantitative measurement. DOI:10.1088/0256-307X/35/2/024202 PACS:42.62.Fi, 42.55.Px, 88.20.jj, 87.64.kv © 2018 Chinese Physics Society Article Text Alkali compounds are of concern in coal and biomass combustion. As a consequence of their low melting points and high reactivity, alkali compounds are involved in the corrosion of plant materials.[1,2] For example, in a gas turbine, alkali species with a concentration above 24 ppb will prominently decrease the turbine lifetime,[3] and thereby a fast on-line quantitative measurement system is required to monitor the alkali species concentration. Tunable diode laser absorption spectroscopy (TDLAS) is a high-sensitivity, high-speed, on-line monitoring technique[4-7] that is widely used in combustion diagnosis.[8-10] TDLAS is usually used to monitor the absolute concentration of the species of interest along the laser beam and to estimate the amount of the species released.[11,12] However, in some cases, like the combustion in gas turbine, furnace or engine, knowing the absolute concentration with good spatial resolution of the species is beneficial for optimizing the combustors, and therefore, increases their lifetime.[13-15] The spatially resolved information is also significant for turbulent combustion research and valuable for validation of various computational fluid dynamic (CFD) models.[16,17] It is a challenge to acquire the data with high spatial resolution using TDLAS. Such as TDLAS-based tomography system[18,19] and CT-TDLAS[20] have been developed to obtain the concentration distributions. The spatial resolution of these methods depends on how many laser beams are used. Laser-induced fluorescence (LIF), in contrast, is powerful in terms of achieving high spatially resolved measurements such as species concentration,[21-24] but it is complicated to obtain the absolute concentration. Because the parameters such as the fluorescence quantum yield, laser line shape, molecular absorption line profile, temperature etc., have to be measured or calibrated first. It is, however, relatively easy to obtain the relative concentration distribution from the LIF measurement. We thereby try to combine the advantages of TDLAS and LIF to realize a quantitative and spatially resolved measurement. However, a practical problem is that, to achieve high signal-to-noise ratio, LIF requires high-power laser source such as pulsed lasers. Continuous wave (cw) lasers are not normally qualified due to their relatively low power. Recently, cw lasers with higher power of up to several watts and with wider tunable range in wavelength were developed.[25,26] More and more species can be detected, such as alkali metal atoms that have strong absorption lines in visible wavelength range, and it is easy to induce the resonance fluorescence of alkali metal atoms by using relatively low power cw lasers. This provides a possibility to use TDLAS and LIF together. In this Letter, a compact method for quantitative and spatially resolved measurement of atomic potassium (K) in the post flame region is demonstrated. We combine TDLAS and LIF using an external cavity diode laser to acquire the absolute line-integration concentration and relative concentration distribution simultaneously, and the absolute concentration distribution of atomic K along the laser beam is obtained. Figure 1 shows the schematic diagram of the setup. A single-mode external cavity diode laser (DL 100, portable Toptica Photonics) was used as the light source. The laser was controlled by a laser controller (Toptica Photonics) including a temperature control module (DTC 110), a current control module (DCC 110) and a scanning control module (SC 110). The central wavelength was set at 769.9 nm with the output power of 40 mW. The mode-hop free scanning range was 25 GHz at a repetition rate of 100 Hz, which covered the D1 absorption line of atomic K. The laser beam passed through the post flame over a burner, and then the laser was received by a photon detector (DTE 210, Thorlab). A B-type thermocouple was inserted in the post flame region, 2 mm above the laser beam, for temperature monitoring. The signals from the detector and the thermocouple were received by a data acquisition card and were transferred to a computer. At a right angle to the laser beam, an infrared-enhanced ICCD camera (IRY-1024, Princeton) was used for recording the fluorescence from atomic K. A home-made analysis program based on LabVIEW™ was used for on-line data processing.

Fig. 1. Schematic diagram of the setup for simultaneous absorption and laser-induced fluorescence measurement.

The burner used was a specially designed multi-jet burner, 65 mm in diameter. It can provide relatively homogenous laminar gas flows within the temperature range of 1000–2000 K. Detailed information of the burner has been described elsewhere.[16] The gases of CH$_{4}$/O$_{2}$/N$_{2}$ were well mixed in the plenum chamber at the bottom of the burner before being introduced into each jet, and a premixed flame is stabilized on each jet nozzle. Each jet is surrounded by six small holes that supply a homogeneous co-flow of N$_{2}$. The co-flow can dilute the burned gas of the jet flames and thus can control the temperature when varying the flow rate. A shielding ring on top of the burner was used to shield the post-flame region from the ambient air. The flame condition and related parameters are listed in Table 1. The potassium hydroxide (KOH) water solution was used as the source of atomic K, and it was nebulized into a fog of micrometer droplets using an ultrasonic nebulizer, and was fed into the jets of the burner with N$_{2}$ as carrier gas. The atomic K was released from KOH and was measured in the post flame.

**Table 1.** Flame conditions and its post-flame compositions calculated using CHEMKIN (SL=liter at standard temperature and pressure).
Jets flow rate (SL/min),		Co-flow
Flame ($\phi=0.9$)		(SL/min)
CH$_{4}$	O$_{2}$	N$_{2}$	N$_{2}$
1.2	2.67	4.7	4
Post-flame composition (mole fraction)$^{\ast}$
CO$_{2}$	H$_{2}$O	O$_{2}$	N$_{2}$
0.095	0.191	0.022	0.692

To measure the absolute atomic K concentration, TDLAS was employed. The Beer–Lambert law is $$ \ln\frac{I(v)}{I_0}=-{\alpha}(\nu)L,~~ \tag {1} $$ where $I_{0}$ and $I(\nu)$ are the input and output laser intensities, respectively, $L$ is the optical path length that is assumed to be equal to the burner diameter, and $\alpha(\nu )$ is the absorption coefficient that is proportional to $N$, the number density of atomic K per unit volume. Hence, we obtain $$ \alpha(\nu)=N\sigma (\nu),~~ \tag {2} $$ where $\alpha(\nu)$ is the absorption cross section of the atomic K. It could be calculated as follows: $$ \mathop \int \limits_{-\infty}^\infty \sigma (\nu)dv=\frac{B_{12} hv}{c},~~ \tag {3} $$ where $h$ is Planck's constant, $c$ is the speed of light, and $B_{12}$ is the Einstein absorption coefficient. Here we have $$\begin{align} B_{12} =\,&\frac{g_2}{g_1}B_{21},~~ \tag {4} \end{align} $$ $$\begin{align} B_{21} =\,&\Big(\frac{c^3}{8\pi hv^3}\Big)A_{21},~~ \tag {5} \end{align} $$ where $A_{21}$ is the Einstein spontaneous emission coefficient, $B_{21}$ is the stimulated emission coefficient, and $g_{i}$ is the degeneracy of level $i$. For the D$_{1}$ line of atomic K, $A_{21}=3.75\times10^{7}$ S$^{-1}$, and $g_{1}=g_{2}=2$. According to Eq. (3), the absorption cross section can be deduced, and the absolute concentration is determined by $$ N=-\frac{8\pi v^2}{LA_{21} c^2}\mathop \int \limits_{-\infty}^\infty {\rm \ln}\frac{I(\nu)}{I_0}dv.~~ \tag {6} $$ Figure 2 shows the spectra of the D1 line of the atomic K obtained from the experiment. The laser mode-hop free scanning range is 25 GHz. According to Eq. (6), the concentration of the atomic K can be deduced.

Fig. 2. Experimental data of D1 line of atomic potassium: (a) scanning spectrum, (b) transmission spectrum, and (c) absorption spectrum.

The atomic K concentration was measured as shown in Fig. 3. When the ultrasonic nebulizer was turned on, KOH fog was fed into the flame, and the atomic K formed immediately. The concentration of the KOH solution is 42 ppm and the atomic K in the post flame was measured to be $\sim$13 ppb, indicating that the atomic K conversion rate is 0.03%. In Fig. 3, the standard deviation of 0.1 ppb was obtained, which is the sensitivity of the TDLAS system.

Fig. 3. The release history of atomic K in the post flame when the ultrasonic nebulizer was switched on first (before 350 s) and off later (after 350 s).

Fig. 4. (a) LIF images and (b) normalized LIF intensity distributions of atomic K at the heights of 15, 20, 25, 30 mm, respectively, above the burner.

During the TDLAS measurements, the LIF images were recorded simultaneously. The scanning repetition rate of the laser was 100 Hz, and when its wavelength arrived at the peak absorption of the D1 line, the resonant fluorescence was recorded by the ICCD with an exposure time of 1 ms. A band pass filter with a central wavelength at 766 nm and a full width at half maximum of 10 nm was equipped to the camera to block the stray light. Figure 4(a) shows the LIF images of the atomic K at the heights above the burner (HAB) of 15, 20, 25 and 30 mm, respectively, in the post flame of the burner. From these results, we could know where the atomic K exists and how long the optical path length is. A feature of the multi-jet burner is that the composition in the post flame is nearly the same. The temperature decreases gradually of about 200 K from 15 mm HAB to 30 mm HAB. Under these conditions, the quenching coefficient of atomic K fluorescence was almost the same at different heights in the post flame. Hence, the LIF intensity distributed along the laser beam can represent the relative number density of atomic K. Figure 4(b) shows the normalized LIF intensity distributions along the laser beam. It can be seen that the LIF intensity distributions at different heights have a similar shape, but the optical path length is different and shorter than 65 mm owing to the ambient air intrusion. In Fig. 4, at the heights of 15 and 20 mm, part of the fluorescence on the left side of the images were missed by the ICCD, but based on the trend of the lines in Fig. 4(b), it was compensated. Using Eq. (6) and the optical path length, the accurate concentration can be obtained. Figure 5(a) shows the normalized concentrations (red circle) and their corresponding normalized integrated LIF intensities (black triangle) at the four heights. An excellent agreement between them was obtained, and the relative deviation was less than 1%. This result indicates that the integrated LIF intensity of the atomic K is proportional to its concentration.

Fig. 5. The relationship between the total number of atomic K and its integrated fluorescence intensity (a), and the relative deviation (b).

Since the absolute concentration of atomic K was obtained from TDLAS, and the relative concentration distribution along the laser beam was obtained from LIF, we could deduce the absolute concentration distribution along the laser beam by combining them. Figure 6(a) shows the calculated absolute concentration distribution at four heights. It can be seen that the concentration distribution of the atomic K is not homogeneous, and the region with the highest concentration is near the rim of the burner. The temperatures at the heights of 15 and 30 mm, respectively, along the laser beam direction were measured using a thermocouple, and the results are shown in Fig. 6(b). Except for the two end points of each curve, the temperature difference is within 100–200 K at different heights, and the temperature is relatively homogeneous in the middle area, and the highest temperature region is also near the rim of the burner, like the concentration. It can be seen from Fig. 6 that the concentration of atomic K increases with temperature, which indicates that the quantity of the atomic K released from KOH has a strong correlation with the temperature, especially when the temperature is higher than 1400 K. The reason may be that certain potassium-release chemical reactions tend to happen above 1400 K, and they produce more atomic K at higher temperatures. The detailed quantitative relationship between the atomic K released and the temperature will be studied in the future work. Since the atomic K release in the flame is very sensitive to the temperature, this technique may be used for post-flame temperature visualization.

Fig. 6. Concentration distributions (a) and temperature distributions (b) along the laser beam in the post flame at different heights.

In summary, we have demonstrated a method for quantitative and spatially resolved atomic K concentration measurement in flame. The absolute concentration of the atomic K in the measurement volume is obtained by TDLAS using a signal mode external cavity laser with the sensitivity of 0.1 ppb, and the same laser is also used for LIF of the atomic K simultaneously to obtain the relative concentration distribution. We find that in the post flame, where the temperature variation is within 200 K, and the composition is nearly the same, the quenching coefficient of the atomic K fluorescence is approximately the same. Hence, the atomic K fluorescence intensity distribution along the laser beam can represent its relative concentration distribution. By combining the results of TDLAS and LIF, the absolute concentration distribution is obtained along a line, and two-dimensional concentration distribution can be expected by the laser beam being changed into a laser sheet, where a more powerful laser may be needed. References Biomass combustion in fluidized bed boilers: Potential problems and remediesBiomass Co-Firing for Boilers Associated with Environmental ImpactsSimultaneous CO concentration and temperature measurements using tunable diode laser absorption spectroscopy near 2.3 μmTunable multi-mode diode laser absorption spectroscopy for methane detectionDiffuse reflectivity measurement using cubic cavityTunable diode laser measurements of CO, H2O, and temperature near 1.56 µm for steelmaking furnace pollution control and energy efficiencyTunable diode laser spectroscopy as a technique for combustion diagnosticsSimultaneous measurement on gas concentration and particle mass concentration by tunable diode laserIn situ detection of potassium atoms in high-temperature coal-combustion systems using near-infrared-diode lasersSimultaneous diode-laser-based in situ detection of multiple species and temperature in a gas-fired power plantQuantitative measurement of atomic sodium in the plume of a single burning coal particleIn-situ Measurement of Sodium and Potassium Release during Oxy-Fuel Combustion of Lignite using Laser-Induced Breakdown Spectroscopy: Effects of O ₂ and CO ₂ ConcentrationOn-line and in situ monitoring of oxygen concentration and gas temperature in a reheating furnace utilizing tunable diode-laser spectroscopyHigh-speed tunable diode laser absorption spectroscopy for sampling-free in-cylinder water vapor concentration measurements in an optical IC enginePost-flame gas-phase sulfation of potassium chlorideTemporal release of potassium from pinewood particles during combustionTunable diode laser absorption spectroscopy-based tomography system for on-line monitoring of two-dimensional distributions of temperature and H ₂ O mole fractionAlgebraic tomographic reconstruction of two-dimensional gas temperature based on tunable diode laser absorption spectroscopyTime resolved 2D concentration and temperature measurement using CT tunable laser absorption spectroscopyLIBS measurements and numerical studies of potassium release during biomass gasificationTwo-dimensional visualization of the flame front in an internal combustion engine by laser-induced fluorescence of OH radicalsExperimental investigation of the structure and the dynamics of nanosecond laser-induced plasma in 1-atm argon ambient gasLight-induced atom desorption for cesium loading of a magneto-optical trap: Analysis and experimental investigationsWatt-class high-power, high-beam-quality photonic-crystal lasersHigh power and widely tunable Si hybrid external-cavity laser for power efficient Si photonics WDM links

[1]	Khan A A, Jong W D, Jansens P J and Spliethoff H 2009 Fuel. Process. Technol. 90 21
[2]	Demirbas A 2005 Energy Sources 27 1385
[3]	Scandrett L and Clift R 1984 J. Energy Ins. 57 391
[4]	Sane A, Satija A, Lucht R P and Gore J P 2014 Appl. Phys. B 117 7
[5]	Gao Q, Zhang Y G, Yu J, Wu S H, Zhang Z G, Zheng F, Lou X T and Guo W 2013 Sens. Actuators A 199 106
[6]	Yu J, Zhang Y G, Gao Q, Hu G, Zhang Z G and Wu S H 2014 Opt. Lett. 39 1941
[7]	Wu Q, Thomson M J and Chanda A 2005 Metall. Mater. Trans. B 36 53
[8]	Bolshov M A, Kuritsyn Y A and Romanovskii Y V 2015 Spectrochim. Acta Part B 106 45
[9]	Wang F, Cen K F, Li N, Huang Q X, Chao X, Yan J H and Chi Y 2010 Flow Meas. Instrum. 21 382
[10]	Schlosser E, Fernholz T, Teichert H and Ebert V 2002 Spectrochim. Acta. Part. A 58 2347
[11]	Ebert V, Fernholz T, Giesemann C, Pitz H, Teichert H, Wolfrum J and Jaritz H 2000 Proc. Combust. Inst. 28 423
[12]	Eyk P J, Ashman P J, Alwahabi Z T and Nathan G J 2008 Combust. Flame 155 529
[13]	He Y, Zhu J J, Li B, Wang Z H, Li Z S, Aldén M and Cen K F 2013 Energy Fuels 27 1123
[14]	Sandström L and Malmberg D 2002 Spectrochim. Acta Part. A 58 2449
[15]	Witzel Q, Klein A, Wagner S, Meffert C, Schulz C and Ebert V 2012 Appl. Phys. B 109 521
[16]	Li B, Sun Z W, Li Z S, Aldén M, Jakobsen J G, Hansen S and Glarborg P 2013 Combust. Flame 160 959
[17]	Zhang Z H, Song Q, Alwahabi Z T, Yao Q and Nathan G J 2015 Combust. Flame 162 496
[18]	Xu L J, Liu C, Jing W Y, Cao Z, Xue X and Lin Y Z 2016 Rev. Sci. Instrum. 87 013101
[19]	Song J, Hong Y, Wang G and Pan H 2013 Appl. Phys. B 112 529
[20]	Deguchi Y, Kamimoto T and Kyota Y 2015 Flow Meas. Instrum. 46 312
[21]	Fatehi H, He Y, Wang Z H, Li Z S, Bai X S, Aldén M and Cen K F 2015 Proc. Combust. Inst. 35 2389
[22]	Suntz R, Becher H, Monkhouse P and Wofrum J 1988 Appl. Phys. B 47 287
[23]	Ma Q L, Motto-Ros V, Bai X S and Yu J 2013 Appl. Phys. Lett. 103 204101
[24]	Zhang P F, Li G, Zhang Y C, Guo Y, Wang J and Zhang T 2009 Phys. Rev. A 80 053420
[25]	Hirose K, Liang Y, Kurosaka Y, Watanabe A, Sugiyama T and Noda S 2014 Nat. Photon. 8 406
[26]	Lee J H, Shubin I, Yao J, Bickford J, Lou Y, Lin S, Djordjevic S S, Thacker H D, Cunningham, J E, Ray K, Zheng X Z and Krishnamoorthy A V 2014 Opt. Express 22 7678