the fuel or the desired component of a multicomponent fuel)

Engine Combustion

In-cylinder laser-based combustion diagnostics in internal combustion (IC) engines focus on three main fields:

1. 1.
   the observation of the mixture preparation before ignition, which includes fuel injection and evaporation, fuel–air mixing, pre-combustion temperature distributions, and the distribution of residual inert gases from the previous engine cycle and intentionally recirculated exhaust gas
    
2. 2.
   the observation of the ignition and combustion process including flame propagation and potential flame quenching, also considering post-flame temperatures
    
3. 3.
   the observation of pollutant formation, namely nitrogen oxides (NO x ) and soot
    

In the following paragraphs these different topics are treated separately.

Different strategies for optically accessible IC engines have been used over the last years. The design of optical engines requires compromises between the needs of the diagnostics and the level of comparability to production-line engines. Best optical accessibility is obtained with transparent cylinders and extended windows that are inserted into the piston (Fig. 20.28). The thickness of pressure-resistant quartz windows, however, usually prevents the operation of the neighboring cylinder and changes in-cylinder heat transfer considerably due to the extended quartz surfaces. More-advanced concepts sacrifice optical access by using smaller optical sapphire windows that can be thin enough to fit into a multi-cylinder engine block with minor modifications (Fig. 20.29). Recently, micro-optical laser-based probes using endoscope optics and fibers have been developed for in-cylinder investigation in IC engines [20.175,176,177].

Open image in new window

Fig. 20.28

Basic setup of an optical engine with access to the combustion chamber through a quartz cylinder and a piston window [20.173]

Open image in new window

Fig. 20.29

Advanced concept for an optical engine with narrow sapphire ring that allows the operation of full quasi-production-line four-cylinder engines [20.174]

20.4.4.1 Analysis of Mixture Formation

There are a number of quantities that would be desirable to measure with non-intrusive in-situ techniques to characterize the mixing process in the fresh gas completely. These quantities include:

• Fuel concentration
• Fuel–air ratio
• Temperature
• Fuel composition (i.e., the concentration of individual components)
• Residual gas concentration

To visualize the mixing process and to facilitate the interpretation of the results these quantities should be imaged in at least two dimensions with temporal resolution faster than the time scale of mixing and chemical reaction. A recent review summarizes the state of the art [20.11].

Mixing processes of interest can be categorized according to the level of difficulty in terms of quantitative imaging measurements. They are listed here in order of increasing complexity.

• Constant pressure versus temporally fluctuating pressure
• Constant temperature versus temporally fluctuating (spatially homogeneous) temperature
• Homogeneous temperature versus spatially varying temperature
• Temperature approaches the stability limit of the tracer
• Homogeneous versus inhomogeneous bath gas composition (mainly oxygen concentration is relevant)
• Single-component fuel versus multicomponent fuel
• Non-fluorescing versus fluorescing fuel
• Homogeneous (gas phase) versus heterogeneous (two-phase flows)

The ideal tracer should behave exactly like the fluid it is added to (i.e., the fuel or the desired component of a multicomponent fuel) in terms of droplet formation, evaporation, convection, diffusion, reactivity, and reaction rate. It is obvious that these requirements can not be met in full. However, practical tracers are often very similar to the fuel or are components that are present in commercial fuels. Therefore, in some situations, tracers must not be understood as being added to the fuel. Rather, the other fluorescing substances are replaced by non-fluorescing compounds. The modification of the system must be kept to a minimum and the influence of the tracers on a given experimental situation must be critically reviewed.

Ideally, the tracer should yield LIF signal intensities that are directly proportional to the desired quantity and should not be influenced by the ambient conditions. Unfortunately, signals from all fluorescent tracers show at least some dependence on local temperature, pressure and bath gas variation. Therefore, in experiments where ambient conditions change in time or space, the underlying interdependencies with the tracer signal must be understood in order to obtain quantitative results.

20.4.4.1.1 Tracers for Homogeneous Gas-Phase Systems

Tracer-based LIF techniques have been used for experimental studies in fluid mechanics and combustion for several years. Tracers are single components (molecules or atoms) with highly defined spectral behavior that represents the local concentration of the fluid of interest or that allow to remotely measure a quantity of interest (i.e., temperature or pressure). Typically, compounds are chosen that yield strong enough LIF signal intensities to allow two- (or even three-)dimensional visualization of the desired quantity with sufficient temporal resolution to freeze the motion.

There are two opposite cases that require the application of tracers for the measurement in fluids. First, the components of the fluid do not (or only weakly) fluoresce. This is the case for air and typical exhaust gases (H2O [20.178], CO2 [20.179]). At room temperature these species are only excited in the vacuum-UV region and only at high temperatures do their spectra extend into the spectral range that is of practical use for combustion diagnostics. The resulting signal in O2 [20.180] and CO2 [20.179] is then strongly temperature dependent and the practical use for concentration measurements is limited. In the second case the fuel contains too many fluorescing compounds. This is true for commercial fuels. Their fluorescence has been used to obtain qualitative and semiquantitative measurements of fuel vapor concentrations [20.181]. However, because all these compounds have different physical properties in terms of volatility and transport as well as in terms of their spectral response on variations in ambient conditions, the overall signal cannot be quantified. In both situations it is desirable to add a single tracer that can be selectively observed within the fluid. In the case of fuels (gasoline or Diesel-type fuels) this often means replacing the fluorescing compounds of the fuel by non-fluorescing compounds leaving only one, or adding an additional compound that was not part of the original mixture.

Different classes of molecules have been used as tracers. The choice of potential tracers is driven by the desire to add a minimum concentration of foreign material that yields a maximum LIF signal intensity while not perturbing the system under study. To provide high enough seeding concentrations especially at low temperatures (room temperature) the tracer must have a sufficient vapor pressure. While the main part of the following section focuses on the fuel-like hydrocarbon-based tracers, we include other concepts in the following overview.

Atoms have large absorption cross sections and are candidates that emit strong fluorescence upon excitation in the UV and the visible. The atomization of the material, however, requires the high temperatures that are present in flames. Some metal salts (such as thallium or indium chloride) can be dissolved in the fuel. In the flame front metal atoms are then generated that can be used to measure temperature in the burned gas [20.182,183]. The strong transition moments in atoms allow the use of extremely low (and therefore non-perturbing) seeding levels. The strong transitions, in turn, are easily saturated. Excitation laser intensities are therefore limited and signals are weak, despite their favorable spectroscopic properties. For the observation of fuel distributions prior to combustion this class of fluorescing species is not suitable.

20.4.4.1.2 Small Inorganic Molecules

Di- and triatomic inorganic molecules are frequently used in combustion diagnostics. While unstable radicals such as OH, CH, C2 that appear during the combustion process can be used for flame-front localization and combustion diagnostics [20.10], they are not suitable for observations in the mixing process prior to ignition. Strongly fluorescing stable species, however, are potentially interesting as tracers for the airflow. NO has been used despite its toxicity for studies in gaseous mixing processes [20.184,185] and its spectroscopy is well understood for a wide range of possible applications [20.186]. OH and NO can also be photolytically generated in flow systems. While not suited for studying mixing on a large scale, these flow-tagging techniques give detailed insight into the fluid motion within the lifetime of the generated species. OH has been generated following photodissociation of vibrationally hot water [20.167], NO has been produced from NO2 photolysis [20.187] and from O2 photolysis in air [20.188].

Molecular oxygen has been used to trace the air flow [20.189] and to measure temperatures during mixture formation in a Diesel engine [20.180]. It was also used for flow tagging following the excitation of higher vibrational states by stimulated Raman scattering [known as Raman excitation plus laser-induced electronic fluorescence (RELIEF) [20.169]].

Iodine was applied as a fluorescing tracer that can be excited and detected in the visible spectral range [20.190]. Its toxicity, corrosiveness and the difficulty of seeding iodine at constant rates limits its practical applicability. SO2 can be excited at various wavelengths below 390 nm and subsequently emits fluorescence from the UV to the violet [20.191,192,193]. The (corrosive and toxic) gas (boiling point: −10 °C) can either be doped into the flow or generated in a flame from sulfur-containing precursors [20.194]. The latter application was suggested to mark residual burned gases in internal engine combustion and to visualize their mixing with fresh air and fuel. SO2 fluorescence is strongly quenched by many molecules, including N2 [20.192,195,196,197,198]. Its applicability in high-pressure environments is therefore restricted. The LIF properties of high-temperature CO2 upon excitation in the UV [20.179] offer the potential for new diagnostics for the observation of mixing of hot burned gases with air and fuel.

20.4.4.1.3 Organic Molecules

In contrast to the excitation of atoms and di- and several triatomic molecules, polyatomic organic molecules have a high density of states and therefore show broadband absorption spectra with excitation possible at various wavelengths that are often accessible with standard laser sources. The organic tracers are chemically close relatives of hydrocarbon fuels. Some of the molecules that are attractive tracers are present in commercial fuels at the few percent level. Therefore, relatively high tracer concentrations can be applied without significantly disturbing the combustion process.

The chemical similarity between tracer and fuel has the additional advantage that the tracer disappears (burns) together with the fuel close to the flame front. In measurements with limited spatial resolution (≈1  mm) this is a good match to identify and visualize the position of reactive zones. However, since the reaction kinetics of the tracer is not perfectly identical to that of the fuel, tracers are not in general suitable for high-resolution measurements close to the flame front, because their concentration may not represent the fuel concentration accurately in this zone.

Fluorescing organic tracers come in different sizes and structures with different volatilities. According to their boiling points as a first criterion they can be used to represent different volatility classes of multicomponent fuels [20.199]. At the same time, care must be taken to avoid distillation processes that separate fuel and tracer due to non-ideal boiling behavior during the fuel evaporation [20.200,201,202].

Aromatic hydrocarbons are typical components of commercial fuels. These species are responsible for the strong absorption in the UV and the subsequently emitted fluorescence [20.203]. Single-ring aromatics such as benzene, toluene and xylene are part of gasoline fuels on the percent level while two-ring aromatics such as naphthalene and its derivatives are present in Diesel fuels. They typically have high fluorescence quantum yields (toluene: ϕ = 0.17, benzene: ϕ = 0.22, fluorene: ϕ = 0.66, dimethyl anthracene: ϕ = 0.82) and their absorption and emission spectra shift towards the red with increasing size of the aromatic system. The wide variety of molecular sizes (and therefore boiling points) makes this class of molecules attractive as tracers that can be adjusted to the boiling behavior of the fuel or that are representative for boiling classes in multicomponent fuels. For seeding room-temperature gas flows compounds larger than benzene and toluene have too low vapor pressures.

A major drawback of aromatic tracers is the strong quenching by oxygen. The signal intensities therefore do not only depend on the tracer but also (inversely) on the oxygen concentration. This effect was taken advantage of by interpreting the signal as proportional to fuel–air ratios, which are of major practical interest.

The aromatic compounds in commercial fuels have been used for qualitative and semiquantitative measurements, both in the vapor [20.204] and liquid phase [20.205,206]. Benzene as a fuel tracer is typically avoided because of its carcinogenic effects. Toluene is less toxic and not considered carcinogenic. Therefore, it has been most frequently chosen as a fuel tracer [20.207] and recent publications have shed more light on the dependence of its fluorescence on p, T and nO2 [20.208,209]; α-methyl naphthalene was investigated by LIF, see e.g. [20.210], because it is part of a model two-component fuel that is used in experimental and numerical studies as a substitute for diesel or JP8 fuel. Like naphthalene, it is used in combination with N,N,Nʼ,Nʼ-tetra-methyl-p-phenylenediamine (TMPD) in exciplex studies to visualize liquid and vapor phases simultaneously [20.211].

20.4.4.1.4 Aliphatic Compounds

Typical saturated aliphatic hydrocarbons such as alkanes and saturated alcohols do not fluoresce. They have their first absorption bands in the vacuum-UV region and excitation typically leads to photodissociation. Nonsaturated hydrocarbons with extended conjugated systems that would have useful spectroscopic properties are unstable and tend to polymerize.

Fluorescing aliphatic candidates contain chromophores that allow excitation into stable states that subsequently fluoresce. This class of molecules includes ketones (R2CO), aldehydes (R–CHO) and amines (R3N, where R is a saturated aliphatic hydrocarbon chain). The (conjugated) combination of chromophores (as in diketones R–CO–CO–R) typically shifts the absorption and fluorescence spectra to the red.

Ketones are the most frequently used class of fluorescent tracers. Their properties have been extensively studied [20.212,213,214,215,216,217,218] and they have been applied to various practical situations [20.219,220,221,222]. The high vapor pressure makes acetone (bp: 56 °C) an ideal tracer for gaseous flows [20.223,224]. 3-pentanone (bp: 100 °C) [20.213,225,226] or mixtures of 3-pentanone and 3-hexanone [20.200] have been suggested as tracers that mimic the boiling and transport properties of gasoline. In most of those cases, iso-octane was substituting gasoline. This has the advantage that iso-octane is non-fluorescent and as a single component is more amenable to detailed modeling studies. For Diesel fuels, the use of even larger ketones was suggested, although they turned out to have limited stability at high temperatures. The larger aliphatic chains enhance the reactivity of the carbonyl group.

The smallest aldehyde, formaldehyde (HCHO), tends to polymerize and is therefore difficult to handle as a dopant. The next largest homologous molecule, acetaldehyde (CH3CHO), has been used as a tracer in internal combustion engines [20.227]. Its low boiling point (21 °C) allows high seeding concentrations. Its spectral properties are comparable to those of acetone. Because acetone is considered less harmful, acetaldehyde is not frequently used as a tracer substance. Aldehydes with different molecular weight have been used to trace different boiling fractions in multicomponent fuels [20.228].The engine measurements [20.134] (Fig. 20.32) were conducted in a modified production-line single-cylinder two-stroke engine (bore: 80 mm, stroke: 74 mm, compression ratio: 8.6). The original cylinder head was replaced by a quartz ring with a height of 4 mm to allow for the entrance and exit of the laser sheets, and a full-size cylindrical quartz window on top through which fluorescence was monitored. The engine was carburetor-fueled with iso-octane doped with 10% (v/v) 3-pentanone. For the measurements described here the equivalence ratio was kept at ϕ = 0.62 with an ignition timing at −20° crank angle (CA) with respect to top dead center (TDC) at a speed of 1000 rpm. Two excimer lasers, operated with KrF (248 nm) and XeCl (308 nm), were fired with a fixed delay of 150 ns to prevent crosstalk between the LIF signals. The laser pulse energies were adjusted to no more than 50 mJ within a 20 ×0.5  mm2light sheet. The signals were directed via a metal-coated beam splitter to two ICCD cameras that were equipped with f # = 2, f =100  mm achromatic UV lenses.