We demonstrate a femtosecond two-photon laser-induced fluorescence (fs-TPLIF) technique for sensitive CO detection, using a 230 nm pulse of 9 µJ and 45 fs. The advantages of fs-TPLIF in excitation of molecular species were analyzed. Spectra of CO fs-TPLIF were recorded in stable laminar flames spatially resolved across the flame front. A hot band (1, n) together with the conventional band (0, n) of the B→A transitions were observed in the burned zone and attributed to the broadband nature of the fs excitation. The CO fs-TPLIF signal recorded across the focal point of the excitation beam shows a relatively flat intensity distribution despite of the steep laser intensity variation, which is beneficial for CO imaging in contrast to nanosecond and picosecond TPLIF. This phenomenon can be explained by photoionization, which over the short pulse duration dominates the population depletion of the excited B state due to the high peak power, but only contributes in total a negligible X state depletion due to the low pulse energy. Single-shot CO fs-TPLIF images in methane/air flames were recorded by imaging the broadband fluorescence. The results indicate that fs-TPLIF is a promising tool for CO imaging in flames.
© 2017 Optical Society of America
Carbon monoxide (CO) is one of the most important combustion intermediates and fuels. Nonintrusive spatiotemporally resolved CO detection in combustion environments with effective optical methods is of general interest (see e.g .), where two-photon laser-induced fluorescence (TPLIF) has been the most widely used technique. TPLIF technique using nanosecond (ns) laser has been broadly adopted to visualize CO in flames  and in engines . There has been a large number of works devoting to the selection of proper excitation-detection schemes and the measurement of crucial parameters (cross sections for absorption, quenching, photoionization etc.) for sensitive and quantitative CO detection. A comprehensive summary can be found in a recent publication . TPLIF has also been developed by employing a picosecond (ps) laser  to improve the multi-photon excitation processes and to suppress the photolytic interferences.
Femtosecond (fs) laser has been applied extensively to femto-chemistry , while its application to combustion diagnostic has previously been concentrated mainly on nonlinear coherent techniques, e.g. fs coherent anti-Stokes Raman spectroscopy [7–9] and degenerate four-wave mixing . The extraordinary property of the ultra-high peak power of fs laser pulse is also in favor for a high efficient two-photon excitation, while the relatively low pulse energy facilitates an effective suppression of photolytic interferences generally involved single-photon processes. So far fs-TPLIF has been applied to the detection of atomic hydrogen [11–14], OH , oxygen . Very recently, an fs-TPLIF CO detection works has also been reported .
More systematic studies are needed to explore the potential impact of fs-TPLIF technique in combustion research especially for the detection of molecular species. In this work, we analyzed the advantages of fs-TPLIF in comparison with ns/ps TPLIF regarding the high peak power and broadband excitation. In the fs-TPLIF CO experiment conducted in this work, we report the observation of hot CO band in the fluorescence spectra and a sensitive detection.
2. Advantages of fs-TPLIF for molecular species excitation
Besides the well-described advantage about the suppression of photolytic interferences when employing short-pulse excitation in TPLIF especially at the ultraviolet spectral range, the effective utilization of the whole broadband profile of an fs excitation pulse due to the exist of multi photon-pairs in matching the two-photon resonant excitation has been mentioned . To illustrate more quantitatively the superiority of fs-TPLIF relative to the conventional ns and ps techniques, an analysis of the dependence of the two-photon excitation efficiency on laser linewidth and pulse duration is needed.
The fluorescence signal for a TPLIF can be expressed asEq. (5) in reference , σ can be expressed as
To simplify the expression of the two-photon rate coefficient σ in Eq. (2), the Boltzmann fraction and the rotational line strength are set to 1 by considering the condition of a single-line transition, and the second-order intensity correlation factor of the photon statistics of the excitation pulse is also set to 1 for the case of a Fourier transform limited (FTL) excitation laser . To facilitate a numerical estimation, we assume that the excitation laser has a Gaussian spectral distribution with a half-width-half-maximum (HWHM) of ,Eq. (2) can be simplified toEq. (6). It is worth to note that the linewidth dependence specified here is slightly over estimated due to the fact that the molecular linewidth is treated as infinitely narrow .
A similar argument can be applied to estimate the dependence of the two-photon excitation efficiency on the temporal distribution of the excitation fs laser pulses expressed asEqs. (6) and (9), the two-photon excitation efficiency can be expressed as
So the relative TPLIF efficiency can be estimated by the product of of the excitation laser pulse. For a FTL laser pulse, is a constant and has the smallest value, which favoring the most for the two-photon excitation. Most fs laser pulses are close to FTL, while for most ns or ps laser pulses, the values are generally larger by a factor of around 100. For detection of molecular species, the possibility of a simultaneous excitation of the whole vibration band (or even include some hot bands) forms another extra advantage of fs-TPLIF in contrast to the ns and ps cases, where only a couple of rotational lines can be excited. This effect will be more obvious in flames, where much more rotational lines and even hot vibrational bands being populated.
3. Experimental setup and measurements
The experiments were performed in CH4/air flames burned on a modified McKenna burner, which consists of a central tube (3 mm in inner diameter) and a water-cooled annular porous plug (60 mm in diameter). The tube and the plug can be supplied independently with premixed CH4/air gases through separated mass flow controllers to provide stable Bunsen-type flames for averaging spectroscopic measurements. Three types of laminar CH4/air flames, i.e. premixed flames with hot co-flow, partially premixed flames without hot co-flow, and a diffusion flame, were tested. Conditions of the flames are listed in Table 1.
Shown in Fig. 1 is a schematic of the experimental setup together with a photo of the burner with a typical premixed flame. The laser source for two-photon CO excitation was a Ti:sapphire laser system (Spectra-Physics, Spitfire Ace) pumping an optical parametric amplifier (OPA, Light Conversion, TOPAS-Prime). The OPA output has a maximum pulse energy of 9 μJ (a typical pule to pulse energy variation observed during the experiment was aournd 3%) centered at 230.1 nm with a pulse duration of ~45 fs and a repetition rate of 1 kHz. The laser was focused at the middle of the flame cone 8 mm above the burner surface by a spherical lens (f = 300 mm). CO was electronically excited from the ground state (X1Σ+) to the excited state (B1Σ+) through a two-photon process, as shown in a simplified energy level diagram presented in Fig. 2. For spectral analysis, the LIF signal was collected at the right angle using a spherical lens (f = 100 mm), and the laser-illuminated flame volume (as indicated by the broken line box in the flame photo in Fig. 1) was imaged onto the input slit (10 mm × 250 μm) of a spectrometer (Princeton Instruments, Acton SP-2300i). The slit was horizontally orientated to record spatially resolved spectra across the flame front. The fluorescence dispersed by a grating (300 grooves/mm blazed at 300 nm) was captured at the exit port by an ICCD camera (Princeton Instruments, PI-MAX3:1024i). For CO TPLIF imaging, a Nikon objective (50 mm, f/1.2) was installed into the ICCD camera.
4. Results and discussions
Shown in Fig. 3(a) is a spatially resolved fs-TPLIF imaging spectrum collected from the Φ = 1.5 premixed flame with the y-coordinate representing the radial position across the flame front. A spectral curve integrating along the y-coordinate is shown in Fig. 3(b). For comparison, an fs-TPLIF spectral curve collected at room temperature (5000 ppm CO diluted in N2) is shown in Fig. 3(c). All the peaks in Fig. 3(c) are recognized as the B1Σ+→A1Π bands covering from (0-0) to (0-6), which are similar to those collected from flames in previous studies [5, 20] using either ns or ps pulses regarding the peak positions, except that the line profiles of fs-TPLIF appear slightly broader. The fs-TPLIF spectrum of the flame, however, shows a much broader structure and some extra molecular bands. Similar fs-TPLIF spectra have also been observed by Richardson et al.  in a methane diffusion flame, but the limited S/N ratio seems to mislead some of their line assignments. We assign all the extra bands observed in the burnt zone of the flame to the hot band of the B1Σ+→A1Π transition covering from (1-0) to (1-8) as indicated in Fig. 3(b). The related vibrational states of the X1Σ+, B1Σ+ and A1Π electronic levels are depicted in Fig. 2. The separation between the diagonal bands (0-0) and (1-1) in the B1Σ+←X1Σ+ transition is ~80 cm−1 (2130-2050 cm−1). Since the fs laser pulse has a bandwidth of ~250 cm−1 (measured using a grating spectrometer), a simultaneous excitation of both the diagonal (0,0) and (1,1) bands can be achieved , especially in flames where the X1Σ+ (1) level (~2100 cm−1) is substantially populated. The thermally populated high J rotational levels explain the much broader band structure collected in the burned zone of the flame. This is, however, not possible when using either a ns or a ps laser with a linewidth typically less than 10 cm−1. This broadband excitation enhances the excitation efficiency by including all the thermal populations. Given the knowledge of an accurate spectral profile of the excitation fs laser, this feature can also be used for temperature measurement. In the unburnt zone, fluorescence from CH radical is also observed, which is attributed to the photodissociation of CH4 . The 656.3 nm line of atomic hydrogen is also observed in the burnt zone.
In order to evaluate the potential photolytic interferences to the CO fs-TPLIF measurements, we tuned the OPA output to 228 nm, off-resonance from the CO excitation, and compared the off-resonance (off-line) spectra with the on-resonance (on-line) spectra in different flames, and the results are presented in Fig. 4. The flame photos are also provided as insets in Fig. 4 accordingly, and the arrows indicate the positions where the fs laser beam crossed the flames. It can be seen in Fig. 4 that the CO fluorescence disappears in the offline spectra collected at premixed and partially premixed flames. This result indicates that the fs laser does not generate photolytic CO* fragments in the excited B state directly. As to the photolytic CO fragments in the ground-state, if there were any, their presence through dissociation from the super-excited states of precursor species  will be too ‘late’ to ‘catch up’ the fs laser pulse, and thus will not experience the resonant excitation to contribute to the fs-TPLIF signal, owing to the fact that the fs pulse is even shorter than the photolytic-induced dissociation process. This is a distinct different feature from the photolytic process introduced by ns or ps laser pulses. Hence, the fs-TPLIF technique is essentially interference-free in terms of CO-related photolysis . Photolytic C2 interference was hardly observed in CH4/air premixed flames within the flammability limit (Figs. 4(a) and (b)), which is used to be a serious problem suffered by CO TPLIF technique using ns and ps lasers. Limited C2 interference was observed at sooty environment (Fig. 4(e)). Photolytic atomic H line was observed at 656.3 nm (see Figs. 3(a) and 3(b)), which can be neglected due to its trivial contribution compared with the overall CO fluorescence intensity. Besides, no obvious amplified spontaneous emission was observed in the backward direction along the laser beam.
The CO imaging measurements using fs-TPLIF were performed in a gas mixture of CO/N2 (CO concentration: 0.2 vol. %) at room temperature. A homogeneous laminar gas flow of the CO/N2 mixture was generated above the burner through the 60 mm diameter plug of the McKenna burner. The collected CO fs-TPLIF images are shown in Fig. 5(a), which were obtained using an fs laser pulse of 3, 5 and 9 µJ, respectively, averaging over 1000 laser pulses. Intensity curves vertically integrated of the data in Fig. 5(a) are plotted in Fig. 5(b). The CO fs-TPLIF intensity displays a rather flat plateau, which stretches as long as 10 mm near the laser focal point, despite the rather steep laser intensity distribution as shown in Fig. 5(c), which was measured by a moving blade as shown by the dots together with an estimated intensity distribution based on the adopted optical geometry as shown by the broken line. This kind of ‘saturation’ effect can be explained if the photoionization process during the period of the short laser pulse dominates the population loss of the excited states (over quenching etc.). In such an extreme condition, the excitation efficiency expressed in Eq. (1) can be simplified toEquation (11) shows a linear dependence of the CO fs-TPLIF signal on the laser intensity similar to a single-photon excitation process instead of the traditional quadratic dependence in a ns or ps two-photon process. The independence of the laser focus geometry is of substantial practical advantage in the CO fs-TPLIF visualization over conventional ns-TPLIF, where nonlinear effects related to the laser power density are significant. Also thanks to the short pulse duration and consequently the small pulse energy, the total ionization depletion or bleaching of CO ground state was negligible, as indicated by the result that the fs-TPLIF signal increased with the pulse energy. However, a slight absorption of the excitation laser beam can be seen in Fig. 5(b) when using the 9 µJ pulse for excitation, due to the fact that the two-photon absorption cross section is increased with the laser power density.
Besides, we measured the detection limit of the single-shot CO fs-TPLIF imaging technique using the 9 µJ fs laser pulse. The experiments were performed at room temperature in CO/N2 gas mixtures with various CO concentrations. The lowest possible concentration of CO in the gas mixture with a signal-to-noise ratio (SNR) of the CO image being at least 2 is defined as the detection limit, which was estimated to be ~50 ppm.
Furthermore, the fs-TPLIF technique was applied to visualize CO in laminar premixed CH4/air flames. Typical single-shot images using the 9 µJ laser pulse are illustrated in Fig. 6, where SNR obtained for the premixed flame of Φ = 0.6 (lower) and of Φ = 1.5 (upper) are ~6 and ~25), respectively. Within the flammability limit of the CH4/air premixed flames, the technique is proved to be able to provide a SNR good enough to realize single-shot CO imaging.
In summary, the advantage of fs-TPLIF for detection of molecular species regarding the high peak power, ultrashort pulse duration, moderated pulse energy and broad bandwidth multi-lines excitation have been analyzed in comparison with traditional ns and ps experiments. CO detection using fs-TPLIF has been investigated systematically. Fluorescence spectra were collected spatially resolved across laminar CH4/air flames. Hot vibrational bands and rotational lines were recognized and attributed to the feature of broadband excitation. Potential photolytic interferences from the UV excitation laser were examined and interference-free CO detection in premixed CH4/air flames are confirmed. Only a limited photolytic-induced C2 spectral lines were observed in the sooty part of a CH4 diffusion flame, but the interference to the fs-TPLIF is much less than that in conventional TPLIF using ns and ps lasers. Single-shot CO imaging in CH4/air premixed flames was achieved. The CO fs-TPLIF signal shows a rather flat distribution indicating a linear dependent similar to single-photon LIF experiments. This feature can be explained by the dominating photoionization process and being in favor for practical CO imaging measurements. The hot vibrational bands and rotational lines observed in the fs-TPLIF emission spectra are sensitive to temperature and can potentially be used for simultaneous flame temperature and the CO concentration measurements. The fs-TPLIF detection can also be applied to other molecules, e.g. NH3 and NO, important in combustion diagnostics.
The National Natural Science Foundation of China (91541119, 51320105008, 91541203); the Swedish Energy Agency for the Swedish-Chinese collaboration project.
The authors declare that there are no conflicts of interest related to this article.
References and links
1. B. B. Dally, A. R. Masri, R. S. Barlow, and G. J. Fiechtner, “2-photon laser-induced fluorescence measurement of CO in turbulent non-premixed bluff body flames,” Combust. Flame 132, 272 (2003).
2. K. Smyth and D. R. Crosley, “Detection fo minor species with laser techniques,” in Applied Combustion Diagnostics, K. Kohse-Höinghaus and J. Jeffries, eds. (Taylor & Francis, 2002), p. 9.
3. M. Richter, Z. S. Li, and M. Aldén, “Application of two-photon laser-induced fluorescence for single-shot visualization of carbon monoxide in a spark ignited engine,” Appl. Spectrosc. 61(1), 1–5 (2007). [PubMed]
4. O. Carrivain, M. Orain, N. Dorval, C. Morin, and G. Legros, “Experimental spectroscopic studies of carbon monoxide (CO) fluorescence at high temperatures and pressures,” Appl. Spectrosc. 71(10), 2353–2366 (2017).
5. C. Brackmann, J. Sjoholm, J. Rosell, M. Richter, J. Bood, and M. Alden, “Picosecond excitation for reduction of photolytic effects in two-photon laser-induced fluorescence of CO,” Proc. Combust. Inst. 34, 3541–3548 (2013).
6. A. H. Zewail, “Femtochemistry: Atomic-scale dynamics of the chemical bond,” J. Phys. Chem A. 104, 5660–5694 (2000). [PubMed]
7. A. Bohlin, M. Mann, B. D. Patterson, A. Dreizler, and C. J. Kliewer, “Development of two-beam femtosecond/picosecond one-dimensional rotational coherent anti-Stokes Raman spectroscopy: Time-resolved probing of flame wall interactions,” Proc. Combust. Inst. 35, 3723–3730 (2015).
8. C. N. Dennis, C. D. Slabaugh, I. G. Boxx, W. Meier, and R. P. Lucht, “Chirped probe pulse femtosecond coherent anti-Stokes Raman scattering thermometry at 5 kHz in a Gas Turbine Model Combustor,” Proc. Combust. Inst. 35, 3731–3738 (2015).
9. S. P. Kearney, “Hybrid fs/ps rotational CARS temperature and oxygen measurements in the product gases of canonical flat flames,” Combust. Flame 162, 1748–1758 (2015).
10. T. Hornung, H. Skenderovic, K. L. Kompa, and M. Motzkus, “Prospect of temperature determination using degenerate four-wave mixing with sub-20 fs pulses,” J. Raman Spectrosc. 35, 934–938 (2004).
11. W. D. Kulatilaka, J. R. Gord, V. R. Katta, and S. Roy, “Photolytic-interference-free, femtosecond two-photon fluorescence imaging of atomic hydrogen,” Opt. Lett. 37(15), 3051–3053 (2012). [PubMed]
12. C. A. Hall, W. D. Kulatilaka, J. R. Gord, and R. W. Pitz, “Quantitative atomic hydrogen measurements in premixed hydrogen tubular flames,” Combust. Flame 161, 2924–2932 (2014).
13. W. D. Kulatilaka, J. R. Gord, and S. Roy, “Femtosecond two-photon LIF imaging of atomic species using a frequency-quadrupled Ti:sapphire laser,” Appl. Phys. B-Lasers Opt. 116, 7–13 (2014).
14. B. Li, D. Y. Zhang, X. F. Li, Q. Gao, M. F. Yao, and Z. S. Li, “Strategy of interference-free atomic hydrogen detection in flames using femtosecond multi-photon laser-induced fluorescence,” Int. J. Hydrogen Energy 42, 3876–3880 (2017).
15. H. U. Stauffer, W. D. Kulatilaka, J. R. Gord, and S. Roy, “Laser-induced fluorescence detection of hydroxyl (OH) radical by femtosecond excitation,” Opt. Lett. 36(10), 1776–1778 (2011). [PubMed]
16. B. S. Jacob, L. S. Brian, D. K. Waruna, R. Sukesh, S. James, and R. G. James, “Femtosecond, two-photon laser-induced-fluorescence imaging of atomic oxygen in an atmospheric-pressure plasma jet,” Plasma Sources Sci. Technol. 24, 032004 (2015).
17. D. R. Richardson, S. Roy, and J. R. Gord, “Femtosecond, two-photon, planar laser-induced fluorescence of carbon monoxide in flames,” Opt. Lett. 42(4), 875–878 (2017). [PubMed]
18. M. D. Di Rosa and R. L. Farrow, “Two-photon excitation cross section of the B <- X(0,0) band of CO measured by direct absorption,” J. Opt. Soc. Am. B 16, 1988–1994 (1999).
19. B. R. Marx, J. Simons, and L. Allen, “Effect of laser linewidth on 2-photon absorption rates,” J. Phys. B-Atom. Molec. Opt. Phys. 11, L273–L277 (1978).
20. D. A. Everest, C. R. Shaddix, and K. C. Smyth, “Quantitative two-photon laser-induced fluorescence imaging of CO in flickering CH4/air diffusion flames,” in Twenty-Sixth Symposium (International) on Combustion, Vols 1 and 2 (1996), pp. 1161–1169.
21. P. J. H. Tjossem and K. C. Smyth, “Multiphoton Excitation Spectroscopy of the B1sigma+ and C1sigma+ Rydberg States of Co,” J. Chem. Phys. 91, 2041–2048 (1989).
22. F. Kong, Q. Luo, H. Xu, M. Sharifi, D. Song, and S. L. Chin, “Explosive photodissociation of methane induced by ultrafast intense laser,” J. Chem. Phys. 125(13), 133320 (2006). [PubMed]
23. H. L. Xu, A. Azarm, and S. L. Chin, “Controlling fluorescence from N-2 inside femtosecond laser filaments in air by two-color laser pulses,” Appl. Phys. Lett. 98, 141111 (2011).
24. A. P. Nefedov, V. A. Sinel’shchikov, A. D. Usachev, and A. V. Zobnin, “Photochemical effect in two-photon laser-induced fluorescence detection of carbon monoxide in hydrocarbon flames,” Appl. Opt. 37(33), 7729–7736 (1998). [PubMed]