Influence of pressure, temperature and gas phase composition on biacetyl laser-induced fluorescence
- First Online:
- Cite this article as:
- Guibert, P., Modica, V. & Morin, C. Exp Fluids (2006) 40: 245. doi:10.1007/s00348-005-0064-y
- 185 Downloads
The purpose of this work is to get fundamental knowledge and to understand experimentally the fluorescence phenomenon usually used in laser-based imaging diagnostics. A complete review of thermodynamic effects (pressure, temperature, oxygen and tracer concentration) on biacetyl fluorescence is presented for a wide range of temperature and pressure. Biacetyl fluorescence increases with pressure and then levels off at high pressure when the relaxation is complete. The influence of the temperature is explained by a competition between the intersystem crossing and fluorescence. The effect of oxygen fluorescence quenching is important at high pressure and is discussed.
List of symbols
light velocity (m s−1)
tracer concentration (molecule cm−3)
laser fluence (J cm−2)
Planck constant (J s)
light intensity (J)
Boltzmann constant (J K−1)
optical path length (cm)
tracer volume (μl)
collection volume (cm3)
Lennard–Jones collision frequency (cm3 s−1)
fluorescence quantum yield
excitation wavelength (nm)
overall efficiency of the collection optics
absorption cross section (cm² molecule−1)
Laser induced fluorescence (LIF) technique allows quantitative determination of scalar fields (concentration, temperature), however, the aerothermodynamic surrounding conditions involve complex fluorescence phenomena. Before setting quantitative fuel concentration measurement by LIF on engines, the influence of preponderant parameters (concentration of oxygen, tracer, temperature, pressure, laser power, combination of tracer and excitation wavelength...) on the fluorescence signal must be characterized. The objective of this study is thus to investigate the LIF experiments in a constant volume vessel at high temperature and high pressure to underline the influencing parameters over a wide range of pressure, temperature, tracer concentration, in conditions close to those encountered in internal combustion (IC) engines.
Laser induced fluorescence presents many applications in spectroscopy. This technique serves as a sensitive monitor for the absorption of laser photons in fluorescence excitation spectroscopy; it is suited to gain information on molecular states and allows the spectroscopic study of collision processes (Demtröder 1982). Moreover, planar laser induced fluorescence (PLIF) has been widely developed for imaging fuel distributions in IC engines (Zhao and Ladommatos 1998), for concentration (Lozano et al. 1992; Thurber and Hanson 2001) and temperature measurements (Thurber et al. 1997; Koch and Hanson 2003). For fuel concentration measurement by LIF, the technique usually adopted is fluorescence from a tracer molecule, which presents satisfactory spectroscopic and thermodynamic properties similar to the fuel. For instance, a tracer must present a limited absorption of the available laser wavelength, satisfactory fluorescence yield, low quenching rate by oxygen, low toxicity, good stability and solubility in fuel, with a boiling point and a latent heat of vaporisation similar to the fuel. A number of studies have been conducted to find suitable tracers for IC engines (Zhao and Ladommatos 1998). The required absorption wavelength in the visible and UV region restricts the tracers to two main categories: aromatic and carbonyl compounds. The fluorescence intensity of the aromatic tracer is much stronger than that of the carbonyl compounds (Krämer et al. 1998). The lowest-lying transitions of aromatic hydrocarbons are of the π–π* type and are characterized by high molar absorption coefficients and relatively high fluorescence quantum yields (Valeur 2002). A π orbital is formed from two p atomic orbitals overlapping laterally, that involves a π bond. The transition π–π* corresponds to an absorption of a photon with appropriate energy, which promotes one of the π electrons to an antibonding orbital π*. Aromatic compounds fluoresce in the broad range of 270–640 nm, with an intense fluorescence for unsubstituted aromatics in the ultraviolet (benzene, toluene, naphthalene) or visible region (anthracene). As the degree of conjugation increases (an increase in the extent of the π-electron system), the intensity of fluorescence often increases and a shift to longer wavelengths is observed (Valeur 2002). Among the carbonyl compounds, aldehydes, ketones and diones are used. These compounds fluoresce in the spectral region of 300–550 nm as a result of the n–π* transition in the C=O group (Grossmann et al. 1996; Ossler and Alden 1997; Yuen et al. 1997), which corresponds to a promotion of a non-bonding electron to an antibonding orbital. The transition can be realised with sufficient coupling between the magnetic moments of orbital and spin. Transitions of the n–π* type are characterized by molar absorption coefficients which are at least a factor of 102 smaller than those of π–π* transitions (Valeur 2002). For example, ketones are excited from a singlet ground electric state (S0, n) to a first excited singlet state (S1, π*), and then transit to an excited triplet state (T1) whose energy level is lower than S1 by intersystem crossing without radiation. The fluorescence results from the relaxation from S1 to S0. For carbonyl compounds, the transition from S1 to T1 by intersystem crossing has a high probability, whereas the probability of S1 to S0 is very low. The intersystem crossing to the triplet state is the most important process for the excited singlet ketone (Shortridge et al. 1971). On the contrary, for aromatic compounds the intersystem crossing probability is low and hence they emit strong fluorescence. Quenching by molecular oxygen makes the fluorescence intensities inversely proportional to oxygen concentration (Koban et al. 2002). Since the lifetime of excited states is controlled by a rapid intersystem crossing, the influence of collisional quenching of the S1 state by O2 is higher for aromatic compounds than for carbonyl compounds (Einecke et al. 2000). Thus, aromatics such as toluene or xylene are strongly quenched by oxygen (Reboux et al. 1994; Fujikawa et al. 1997) and the oxygen quenching rate is particularly elevated at high pressure.
Ketones commonly used are mainly acetone and 3-pentanone (Yuen et al. 1997; Thurber et al. 1998; Thurber and Hanson 1999). As shown by Fujikawa et al. (1997), both tracers have higher fluorescence yields than aldehydes (e.g. propionaldehyde, n-valeraldehyde). Moreover, ketones are preferred over aldehydes as tracers due to their low reactivity under high pressure and high temperature environment such as in an IC engine (Zhao and Ladommatos 1998). Acetone and 3-pentanone are relatively chemically stable and dissociate only at high temperatures in the flame front (Grossmann et al. 1996). These compounds are commonly studied for pressures up to 1.5 MPa and in the temperature range of 293–723 K (Fujikawa et al. 1997; Thurber et al. 1998). In a pressure range of 0.1–1.5 MPa, the fluorescence intensity per ketone molecule increases with increasing pressure and excitation wavelength. The increase in the fluorescence quantum yield with pressure may be the result of a combination of incomplete vibrational relaxation in the excited singlet level and the dependence of the singlet–triplet intersystem crossing rate on the vibrational level in the singlet manifold (Reboux et al. 1994; Yuen et al. 1997). High-excitation wavelengths involve weaker pressure and composition dependences of the fluorescence yield. The observed decrease in the fluorescence quantum yield with increasing temperature (Yuen et al. 1997; Thurber and Hanson 1999; Koch and Hanson 2003) is the result of the increase in the intersystem crossing rate for vibrational levels with higher energy in the excited singlet manifold.
Diones, such as biacetyl (2,3-butanedione), present longer excitation and fluorescence wavelengths (Horowitz et al. 2001). The fluorescence emission is broadband in the blue with a lifetime of 15 ns, the phosphorescence is long-lived (lifetime of 1.5 ms). When excited by a laser beam at 355 nm, biacetyl is generally selected as the tracer of iso-octane (2,2,4-trimethylpentane), a reference compound of gasoline (Baritaud and Heinze 1992; Deschamps et al. 1994; Itoh et al. 1995). This tracer decomposes slowly and polymerizes in presence of impurities, that involves a decrease of the fluorescence efficiency with time. Biacetyl has been widely used as a PLIF tracer in low-pressure (0.1–0.5 MPa) and low-temperature (293–373 K) environments (Heicklen 1959; Van Cruyningen et al. 1990; Lawrenz et al. 1992; Deschamps et al. 1994). The fluorescence efficiency is very low when the excitation wavelength is shorter than 320 nm.
For applications in IC engines (for fuel distribution imaging and study of the mixture homogeneity), tracers usually used are: ketones, aldehydes, aromatics and diones. Acetone and 3-pentanone are generally selected because of their property of showing the weakest temperature and pressure dependence of LIF intensity compared to all other tracers. However, their pumping requires the use of lasers at low wavelength (KrF laser at 248 nm, Nd:YAG laser at 266 nm, XeCl laser at 308 nm), in comparison with biacetyl excited at 355 nm by third harmonic Nd:YAG laser and ionized argon Ar++ laser. This continuous laser source was used for concentration measurements in a pressurized and heated gas mixture flow (Guibert et al. 2002) and for LIF application in engine (Rochelle and Perrard 1999). Biacetyl presents longer excitation and fluorescence band than ketones and has thermodynamic properties similar to iso-octane: at atmospheric pressure, the boiling point is 361 K for biacetyl, 329 K for acetone and 373 K for iso-octane.
Pioneering studies of biacetyl photophysics (Okabe and Noyes 1957; Heicklen 1959) have detected little dependence of biacetyl fluorescence on pressure and incident wavelength, over a weak domain. Heicklen (1959) reported the increase of the fluorescence intensity with increasing pressure when the excitation wavelength was 365.5 nm. On the other hand, at 435.8 and 404.7 nm, fluorescence efficiencies were independent of pressure. Amongst the numerous experimental and theoretical studies on the biacetyl emission characteristics (Brand and Mau 1974; Kamei et al. 1986; Senent et al. 1994), Ten et al. (2002) have compared the experimental spectra of the biacetyl absorption and fluorescence with theoretical spectra calculated from the method of hybrid atomic orbitals and semi-empirical correlations. With a satisfactory agreement between experimental and theoretical spectra, the model shows the structural and dynamical changes occurring in molecules under electronic excitation. Recent results (Guibert et al. 2002) have shown the effects of the temperature on the fluorescence intensity. These effects are similar to those observed with ketones and the LIF signal decreases with increasing temperature. Concerning the phosphorescence characteristics, biacetyl triplets phosphoresce with high efficiency, except in an oxygenated environment where the oxygen molecule quenches the biacetyl phosphorescence rapidly and strongly (Jiang-Bang et al. 1988).
In the following sections, the experimental setup (a high temperature and high pressure facility) developed to follow instantaneously the fluorescence signal is described with the optical system. The fluorescence signal is studied for biacetyl. Absorption and emission spectra are characterized. Moreover, a parametric study with the influence of different parameters (laser power, pressure, temperature, mole fraction of tracer, percentage of oxygen) on the fluorescence signal is presented, in order to underline the significant parameters for an optimal application of biacetyl LIF diagnostics. The influence of each parameter is discussed separately by underlining the important effects of the biacetyl photophysics.
2 Experimental setup
2.1 High temperature and high pressure facility
The maximum operating pressure of the experimental facility is 5 MPa and the maximum temperature is 773 K. The vaporization of the liquid tracer is achieved by reaching the tracer vapor pressure through a void pump. The liquid tracer is injected by means of different types of syringes at airtight piston. An electrical heating system regulates the temperature of the chamber. Temperature is controlled inside the chamber by three thermocouples that ensure uniformity of the temperature field. During the LIF experiments, the following uncertainties are encountered on measuring the tracer volume, the ambient temperature and the ambient pressure: ΔV/V=1%, 0.05%<ΔT/T<0.08%, 0.5%<ΔP/P<20%.
2.2 Optical system
The chamber is equipped with three optical quartz windows for absorption and fluorescence signal measurements. Two types of excitation sources are used. The first one used for fluorescence measurements is a continuous argon laser in an UV domain (ionized argon Ar++, discrete wavelengths of 352.9 and 365.2 nm, maximum power output of 100 mW) coupled to an optical fiber of 200 μm diameter. The second one used for the absorption measurements is a high-pressure xenon lamp with a wide wavelength range (from 250 to 1,100 nm) coupled to a liquid optical fiber (Series 380, Lumatec) with a core diameter of 2 mm, a light conducting liquid core, transparent from 270 to 720 nm and more efficient than quartz fiber for transmission in UV range. A focusing lens system allows obtaining an incident coherent light beam of 1 mm diameter approximately in the chamber center. A second optical fiber of 400 μm diameter collects the emitted fluorescent signal along the axis perpendicular to the incident light beam and is coupled to a high sensitivity miniature spectrometer with a permanently installed diffraction grating (2,048 pixels, a spectral range of 200–850 nm), UV detector upgrade, longpass sorting filter and entrance slit of 10 μm. The spectrometer is connected to a PC, that allows to analyze the absorption and fluorescence spectra accurately.
2.3 Experimental conditions
In the current work, biacetyl (CH3–(CO)2–CH3) is tested as a fluorescence tracer for its satisfactory spectroscopic qualities and thermodynamic properties (boiling point at atmospheric pressure of 361 K, density of 980 kg m−3, molecular weight of 86×10−3 kg mol−1, saturated vapour pressure at 293 K of 6,960 Pa). The experiments are conducted in nitrogen or reconstituted air (80% vol. N2 + 20% vol. O2) for a constant biacetyl volume and for different biacetyl mole fractions. The constant biacetyl volume is fixed at 400 μl, that corresponds to 3.57×1018 molecules cm−3, a value close to those used in the literature (for example: fluorescence of various tracers with a concentration between 1015 and 1018 molecules cm−3 (Grossmann et al. 1996; Bryant et al. 1997), fluorescence of acetone and 3-pentanone with a concentration of 3.1×1017 molecules cm−3 (Krämer et al. 1998).
4.6×10−3 to 257.6×10−3
Air and nitrogen
Air and nitrogen
Air and nitrogen
Air and nitrogen
2.4 Spectra of absorption and emission
Normalized standard deviation in percent at given pressure and temperature obtained from 15 integrated fluorescence signals in nitrogen for a constant biacetyl volume of 400 μl (3.57×1018 molecules cm−3)
3 Influence of different parameters on the biacetyl fluorescence signal
3.1 Linear fluorescence regime
As expressed in Eq.1, the fluorescence signal depends on excitation wavelength. To explain the observation of decreasing fluorescence quantum yield with decreasing excitation wavelength, a multi-step model was developed by Thurber and Hanson (1999) and discussed by Schulz and Sick (2005). The fluorescence quantum yield is decreased with increasing vibrational excitation in the excited state, which gets preferentially populated with short excitation wavelengths. The relaxation of the laser-coupled upper state is treated as a multi-step process, which depends on fluorescence, vibrational relaxation, intersystem crossing and collisional quenching. The wavelength dependence of biacetyl LIF is not studied in this work where the excitation wavelength is fixed by the ionized argon laser.
3.2 Influence of the tracer mole fraction on fluorescence
For a temperature of 423 K and a pressure of 0.1 MPa, the frequency of biacetyl–biacetyl collisions is 1.65×109 s−1 for a biacetyl mole fraction of 0.2 against 6.35×109 s−1 for a biacetyl mole fraction of 0.8.
For the following studies, in order to avoid the saturation phenomenon, the fluorescence signal will be studied for different concentrations of biacetyl with a maximum value of 3.57×1018 molecules cm−3.
3.3 Influence of the pressure on fluorescence
With increasing pressure, the collisional vibrational relaxation will allow to return more rapidly from molecules with high-lying, initially-excited singlet states to molecules with less vibrational states. Since the intersystem crossing is slower from these lower-lying levels, the fluorescence yield and the fluorescence signal will be higher.
3.4 Influence of the temperature on fluorescence
For the biacetyl molecule which presents a transition of type n–π* (molecular orbital n and orbital π*) between the singlet ground electric state S0 and the first excited singlet state S1, the intersystem crossing is efficient. For an excited triplet state T1 whose energy level is lower than S1, the intersystem crossing from the state S1 to the state T1 can be followed by a thermal excitation from the triplet state T1 to the state S1. This transfer from T1 to S1 occurs when the energy difference between both levels is low and the lifetime of T1 is long. In this case, the fluorescence presents two parts: the first one with a classic deexcitation time, the second one with a slightly lower lifetime than the state T1. The fluorescence intensity then increases with temperature.
On the other hand, at a high temperature, the biacetyl molecule at the singlet ground state will acquire more vibrational energy, to reach a state S1 than the triplet state T2, higher than the state T1. In this case, the intersystem crossing between S1 and T2 is activated thermally and is in competition with the fluorescence when the temperature increases, which will be translated into a decrease in fluorescence signal.
3.5 Influence of the oxygen concentration on fluorescence
The fluorescence signal is normalized by the atmospheric pressure value for every biacetyl concentration. As plotted in Fig. 8, it increases linearly with increasing pressure when the ambient gas is nitrogen. On the other hand, with air bath gas, from a pressure of 1 MPa, the fluorescence signal does not increase linearly with pressure, and it is strongly quenched by oxygen molecules, with fluorescence data for air rising very slightly above 1 MPa. For a pressure of 4 MPa, a temperature of 473 K and a biacetyl mole fraction of 2.9×10−3 (equivalence ratio of 0.6), the maximum LIF intensity is twice higher in nitrogen than in air. Experimental results for the three equivalence ratios differ significantly in nitrogen, which would characterize a self-quenching and define the influence of biacetyl–biacetyl collisions.
Oxygen often presents a quenching behavior different from other molecules: oxygen occurs naturally as a ground-state triplet, with two low-lying singlet states. Nau and Scaiano (1996) have studied oxygen quenching of excited aliphatic ketones (acetone, 2-pentanone) and diketones (biacetyl). They propose an oxygen-assisted intersystem crossing, which proceeds through an intermediate encounter complex, as the mechanism for oxygen quenching of emission from the excited singlet state of ketones. The formation of a charge-transfer complex with the ketone involves the electronegative oxygen atom of the excited carbonyl bond and electrons of the double bond.
According to the excitation wavelength, the temperature dependence of the fluorescence signal differs strongly (cf. Fig. 11). For example, Thurber et al. (1998) have shown a continuous transition from a steady increase with temperature at 320 nm to a steady decrease at 248 nm for the acetone fluorescence signal at atmospheric pressure in nitrogen. Grossmann et al. (1996) have studied the temperature dependence of 3-pentanone in synthetic air using 277 and 312 nm excitation wavelength at atmospheric pressure: in the temperature range of 383–573 K, the fluorescence intensity decreased by 15% per 100 K for 277 nm excitation wavelength, but increased by 80% for 312 nm excitation. On the other hand, for each excitation wavelength, Fujikawa et al. (1997) have underlined a decrease of the acetone and 3-pentanone fluorescence signal with increasing temperature in air at atmospheric pressure.
The experimental study on the LIF allowed to characterize the influence of parameters such as pressure, temperature, concentration of tracer and oxygen, on the fluorescence signal intensity of biacetyl. This step is necessary before applying the LIF technique on IC engines in order to obtain quantitative measurements of fuel concentration over the chamber. The fluorescence of biacetyl, usually selected as the dopant of iso-octane and excited with Ar++ laser, is studied for a pressure range of 0.1–4 MPa, a temperature range of 373–573 K, and for different biacetyl mole fractions in air and nitrogen in a high pressure and high temperature facility. The studied regime is a linear fluorescence regime with a spectral intensity of the incident laser light strongly lower than the saturation intensity. Due to the reabsorption process, the fluorescence signal reaches a saturation level from a biacetyl mole fraction of 3.13×10−1 at 423 K and 3.87×10−1 at 523 K. The fluorescence is then studied for different concentrations of biacetyl with a maximum value of 3.57×1018 molecules cm−3.
The results obtained with biacetyl are in agreement with previous studies in the literature about the pressure dependence of acetone: the fluorescence signal increases at low pressure and decreases slightly or stagnates at high pressure. The increase of biacetyl fluorescence with pressure is attributed to the vibrational relaxation of levels with higher fluorescence quantum yields. At high pressure, the vibrational relaxation is faster than the intersystem crossing. As far as the temperature effect is concerned, the fluorescence signal increases and then decreases with increasing temperature resulting from the competition between the intersystem crossing and the fluorescence. From a pressure of 1 MPa, the excited singlet state of biacetyl is strongly quenched by collisions with the oxygen molecule, which involves a decrease of the fluorescence signal. In this case, the vibrational relaxation, which explains the increase of fluorescence with pressure is not more significant, compared to the fluorescence quenching effect of oxygen.
Concerning the application of PLIF diagnostics, the results observed in this work for the variation of fluorescence with pressure, temperature, concentration of tracer and oxygen should allow to optimize experimental measurements. The next step of this study consists of the proposal of an analytic model in order to process direct fluorescence images obtained in specific thermodynamic conditions in order to extract quantitative information.
Vincent Modica is supported by a joint PhD grant from the French Ministry of National Education, Research and Technology.