Spatially and temporally resolved, concentration measurements of important intermediate species in chemically reacting flows, such as flames and plasmas, provide a key insight into the physical and chemical nature of such systems. Noninvasive, laser-based spectroscopic techniques such as laser-induced fluorescence (LIF) are widely used for such measurements [1,2]. In many situations, LIF offers sensitive detection as well as the ability to extend to two-dimensional (2D) imaging by using planar LIF. LIF-based measurements of key atomic species such as H, O, and N as well as molecular species such as CO require multiphoton excitation, as the frequencies of single-photon electronic transitions fall in the vacuum ultraviolet region, for which the medium becomes optically thick in most practical devices [3]. The relatively weaker multiphoton excitation cross-sections necessitate the use of high-energy UV pulses. Such high-energy UV photons can photodissociate certain other molecules in the medium, generating the species being probed. For example, in two-photon excitation LIF (TPLIF) detection of atomic hydrogen, a substantial amount of additional H can be produced via photodissociation of vibrationally excited water vapor and methyl (${\mathrm{CH}}_3$) radicals [4]. In earlier work, improved TPLIF detection has been reported in H and O using picosecond (ps) duration pulses as opposed to the traditionally used nanosecond (ns) duration pulses [3,5].

In the present study, for the first time to our knowledge, we demonstrate nearly photolytic-interference-free femtosecond (fs) TPLIF line imaging of atomic hydrogen in flames. By using high-peak-power but low-energy fs pulses, the photolytic interferences are virtually eliminated while substantially increasing the two-photon excitation efficiency. Additionally, by using amplified Ti:sapphire-based laser systems, the measurement bandwidths can be increased to the 1–10 kHz regime, enabling the study of the spatiotemporal dynamics of turbulent reacting flows. Hydrogen is a key intermediate element in hydrocarbon combustion because of its high reactivity and diffusivity, playing a major role in ignition/extinction and heat release [3,6,7]. In another example, understanding the spatial distribution of H and O atoms is critical for effective use of biomedical plasmas in the rapidly growing area of plasma medicine [8].

The excitation dynamics of the TPLIF process of H are shown in Fig. 1. As shown in Fig. 1(a), two-photon excitation of the $n=1\to \to n=3$ transition with 205 nm photons is used to populate the $n=3$ level, and the subsequent ${\mathrm{H}}_{\alpha }$ fluorescence at 656 nm from the $n=3\to n=2$ decay is detected. Also shown are several other simultaneous processes, including photolytic production of H, photoionization, stimulated emission (SE), and collisional quenching, that can complicate the quantitative interpretation of the measured LIF signal. Photodissociation of numerous H-containing flame molecules can produce substantial quantities of additional H in the medium, as shown in previous studies [3,4,9]. In these conditions, fs-duration pulses become favorable for TPLIF because the signal scales as the laser irradiance squared, whereas single-photon-induced photodissociation processes scale linearly. Additionally, the broad bandwidth of nearly Fourier transform–limited (TL) fs pulses contributes to enhanced excitation through the combination of a large number of photon pairs as shown in Fig. 1(b). In TL pulses, all photons of different colors—corresponding to different frequencies—have the same spectral phase, thus collectively contributing to two-photon excitation [10].

 

Fig. 1. (a) Two-photon-excited LIF detection scheme for atomic hydrogen and (b) efficient two-photon excitation of H using multiple pairs of femtosecond photons.

Download Full Size | PDF

The experimental apparatus consists of a 1 kHz repetition rate, amplified fs Ti:sapphire laser system (Spectra-Physics Solstice) pumping an optical parametric amplifier (OPA) (Coherent OPerA Solo). The peak pulse energy of the pump laser was 2.5 mJ at 800 nm; parametric conversion of the pump beam followed by frequency mixing and then upconversion of the idler beam resulted in up to 8 μJ of UV radiation at 205 nm. The 205 nm beam was collimated to a diameter of approximately 1 mm and transmitted through a $\mathrm{\Phi }=1.2$, ${\mathrm{CH}}_4/{\mathrm{O}}_2/{\mathrm{N}}_2$ Bunsen flame stabilized over a 5 mm diameter underexpanded jet. Two $+50\mathrm{mm}$ focal length $f/1.2$ camera lenses (Nikon Nikkor AIS) were used in the conjugate configuration to collect the fluorescence signal and focus it back onto the detection system. A 50 mm diameter ${\mathrm{H}}_{\alpha }$ bandpass filter (Semrock FF01-655/40-50) was mounted in between the two camera lenses.

Two types of camera systems were used on either side of the flame to record the LIF line image orthogonal to the direction of propagation of the laser beam. An intensified charge-coupled device (ICCD) camera (Princeton Instruments PI-MAX2) recorded single-laser-shot images at a rate of 30 Hz, while a dual-stage high-speed intensifier (LaVision High Speed IRO with S25 visible-enhanced photocathode) coupled to a complementary metal-oxide semiconductor (CMOS) camera system (Photron FastCAM SA5) recorded single-shot images at 1 kHz. The CCD-based camera system was used for initial characterization of the system and for power dependence studies, as well as to ensure that the CMOS-based system was free of nonlinear response issues at the configurations used. The corresponding spatial resolutions were $30\mathrm{\mu m}/\mathrm{pixel}$ (ICCD) and $35\mathrm{\mu m}/\mathrm{pixel}$ (CMOS) respectively, and were uniform in both horizontal and vertical directions. For both camera systems, the spatial response functions across the image plane were determined by recording a fixed LIF signal while translating the burner along the direction of propagation of the excitation laser beam. It was experimentally verified that—for a given flame condition—after proper calibrations, the LIF line profiles recorded using both camera systems were nearly identical.

Shown in the bottom of Fig. 2 are the spatially corrected atomic H LIF line profiles generated for a range of laser pulse energies by vertically integrating the line images at $h=7.5\mathrm{mm}$ above the nozzle exit (a representative example of which is shown in the top of Fig. 2). Note that all LIF line profiles plotted are normalized to the peak signal value near the flame front, although the absolute signal counts increased by a factor of approximately 900 as the laser pulse energy was increased by a factor of nearly 30. As evident from Fig. 2, no apparent change is observed in the relative shape of the line profiles over the range of laser pulse energies used. This observation verifies, even at the highest pulse energy used, that there is no evidence of photolytically produced H, as had been observed in previous ps and ns TPLIF experiments [3]. In those experiments, photolytic production of H was clearly visible in the flame front as well as in the flame products region. In the flame front, the ${\mathrm{CH}}_3$ radical is the primary photolytic precursor, whereas in the flame products region, photodissociation of vibrationally excited water (${\mathrm{H}}_2\mathrm{O}$) vapor becomes the primary interference.

 

Fig. 2. Peak-normalized TPLIF line profiles of H from the right half of the flame for various values of laser pulse energy. A flame-luminosity image is shown in the inset.

Download Full Size | PDF

Furthermore, we observed that the total integrated LIF signal in Fig. 2 scales linearly with the square of the laser energy. Goldsmith [11] has observed a higher-than-quadratic energy dependence when significant photolytic production of H occurs in lean H–O flames. In the absence of photolytic production of H, excited state population loss through laser-energy-dependent nonradiative processes of photoionization and SE, as depicted in Fig. 1(a), can result in subquadratic energy dependence. Since we have concluded that our measurements are photolytic free, the observed quadratic energy dependence suggests that photoionization and SE are negligible in the flames investigated in this work. Furthermore, we observed no SE signal through a fast photodiode placed behind a 656 nm bandpass filter along the pump-beam path after the burner. A similar configuration was used to detect SE in previous ps and ns TPLIF experiments [3].

Interference-free, H-atom TPLIF profiles were recorded in the same flame at four different heights above the nozzle exit using 4 μJ of pump energy. Atomic hydrogen line profiles derived from single-shot images at 1 kHz were compared with 256-shot averaged line profiles obtained using the ICCD camera system as well as with numerical calculations. Spatially resolved temperature and species mole-fraction calculations were performed using the UNICORN code, a time-dependent, 2D mathematical model developed for the simulation of unsteady reacting flows [12]. The experimental profiles were corrected for quenching by ${\mathrm{N}}_2$, ${\mathrm{H}}_2$, ${\mathrm{O}}_2$, ${\mathrm{H}}_2\mathrm{O}$, CO, ${\mathrm{CO}}_2$, ${\mathrm{CH}}_4$, and ${\mathrm{C}}_2{\mathrm{H}}_2$ using the calculated species number densities and using the quenching cross-sections given in [13]. The quenching correction procedure is described in detail in [7]. The total quenching rate increases by up to a factor of 7 from the hot flame front to cold reactants or products. As can be seen in Fig. 3, an excellent match is obtained between the shapes of the experimental and calculated H atom profiles. In a separate experiment, we also obtained H-atom TPLIF line images in a ${\mathrm{CH}}_4/{\mathrm{H}}_2/\text{air}$ turbulent diffusion flame. Such measurements, especially when extended to full 2D imaging, can provide critical experimental data for validating complex turbulent combustion models.

 

Fig. 3. Experimentally measured and calculated H-atom profiles at four different heights above the nozzle.

Download Full Size | PDF

In this Letter, the OPA-based 205 nm generation scheme had an overall conversion efficiency of approximately 0.3%. We have now developed a direct frequency-quadrupling scheme to obtain 205 nm radiation from amplified Ti:sapphire pulses at 820 nm with $>1\%$ overall conversion efficiency for 2D, single-shot imaging of H under realistic flame conditions. The current fs-TPLIF scheme is also being applied to photolytic-interference-free, kHz-rate imaging of O, N, and CO. Furthermore, TPLIF measurements of inert gas tracers such as Kr and Xe are planned for mixture-fraction imaging in combustion-related turbulent flows [14]. In summary, the fs-TPLIF scheme described in this Letter has a wide potential for high-repetition-rate, photolytic-interference-free, 1- and 2D imaging of several key species in gas-phase reacting flows and plasmas.