Open Access
Add to BibSonomyAbstract
This paper presents an optical diagnostic technique based on seeded thermographic phosphor particles, which allows the simultaneous two-dimensional measurement of gas temperature, velocity and mixture fraction in turbulent flows. The particle Mie scattering signal is recorded to determine the velocity using a conventional PIV approach and the phosphorescence emission is detected to determine the tracer temperature using a two-color method. Theoretical models presented in this work show that the temperature of small tracer particles matches the gas temperature. In addition, by seeding phosphorescent particles to one stream and non-luminescent particles to the other stream, the mixture fraction can also be determined using the phosphorescence emission intensity after conditioning for temperature. The experimental technique is described in detail and a suitable phosphor is identified based on spectroscopic investigations. The joint diagnostics are demonstrated by simultaneously measuring temperature, velocity and mixture fraction in a turbulent jet heated up to 700 K. Correlated single shots are presented with a precision of 2 to 5% and an accuracy of 2%.
©2012 Optical Society of America
1. Introduction
Laser-based measurement techniques are crucial to improve the physical understanding of complex turbulent flows, as they allow the non-intrusive measurement of important scalar and vector variables with high spatial and temporal resolution. By combining different optical techniques, multiple fluctuating quantities can be measured simultaneously. The correlations provided by these combined diagnostics are particularly beneficial in the case of turbulent combustion. Modern combustors are typically designed to burn fuel-lean mixtures in a highly turbulent flow in order to achieve low raw pollutant emissions and high power densities. In such flames, there is a strong coupling between the flow field and the heat release from chemical reactions. Multi-dimensional joint measurements of velocity and several scalar quantities such as temperature or major species concentrations are therefore required to gain insight into the underlying physics of these interactions. In addition, such measurements are also particularly useful when compared to the results of numerical models such as large eddy simulation (LES), which are often used in the simulation of turbulent reacting flows [1].
For the measurement of velocity fields, particle image velocimetry (PIV) is widely applied. In reacting flows, micrometer-size tracer particles made of a non-combustible material must be used, but the large Mie scattering cross-section of such particles strongly interferes with most scalar measurement techniques based on Raman and Rayleigh scattering [2]. This interference can be reduced by imaging through a spectrally sharp molecular or atomic filter which blocks the narrowband Mie scattering, but transmits the wings of the temperature-broadened Rayleigh scattering signal [3]. This principle has been applied to perform Rayleigh thermometry simultaneously with PIV [4]. However, this filtered Rayleigh scattering technique is associated with considerable experimental complexity and is restricted to premixed flames where the gas composition, used to evaluate the scattering cross section, is a known and monotonic function of temperature.
Laser induced fluorescence (LIF) thermometry allows planar measurements of temperature even in the presence of solid particles and is based on either naturally occurring or seeded species. A method based on the two-line atomic fluorescence (TLAF) of indium was used over a wide temperature range (800 – 3000 K) to obtain time-resolved measurements in an internal combustion engine [5] and in a diffusion flame [6]. However, the measurements were limited to post ignition events or regions, as neutral indium is only activated within the flame. Likewise, the existence of OH, which is produced naturally in the vicinity of the flame front, limits measurements to the post-flame region [7]. Two line excitation of acetone has been used for thermometry in turbulent flows [8] and for instantaneous measurements of temperature and residual gas composition in engines [9], and is less sensitive to the bath gas composition than aromatic hydrocarbon tracers. Even though acetone can be used for measurements up to 1000 K, such tracers eventually burn and consequently do not allow measurements in the post-flame regions. Finally, NO, when seeded to the flow, is present in both burnt and unburnt regions of the flame. However, the experimental difficulties associated with two-line NO-LIF, which is sensitive to the temporal fluctuations of the narrow modes produced by the two lasers required and to photochemical interferences caused by excitation in 220 - 250 nm range, often does not allow accurate time-resolved measurements in reacting flows [10,11]. For increased accuracy and versatility, a multi-line thermometry technique can be applied to a wide range of steady flames or repetitive systems, but requires time or phase averaging as many rotational lines must be probed [12].
In some cases, such as in the so-called thin flame regime of premixed flames, binarized temperature information is sufficient for the conditioning of velocity data. This is often done using methods based on PIV and either the planar LIF of flame radicals such as OH, which appear only in the burnt gas, or the particle number density which decreases sharply at the flame front (conditioned PIV) [13]. However, such methods based on the binarized identification of the local flame structure are only applicable in certain flames and do not give access to the entire temperature field.
Joint planar measurements of temperature or mixture fraction and velocity in turbulent reacting flows have therefore proven difficult to achieve, and the development of a versatile technique applicable to a broad range of reacting flows remains highly desirable. One possibility is that since velocimetry requires the seeding of solid refractory particles, the use of a material with temperature-dependent properties would allow the simultaneous measurement of temperature, eliminating the need for an additional tracer.
Thermographic phosphors are ideal tracers for this concept. They consist of a ceramic host doped at low concentrations with rare-earth ions, which act as luminescent activator centers. These materials exhibit a temperature-dependent radiative emission following UV excitation. Unlike some of the previously mentioned LIF tracers, many thermographic phosphors are able to survive high temperature environments [14], are chemically inert, and are insensitive to pressure and the local gas composition [15,16]. When used as a gas-phase tracer, this latter feature is a great advantage over several LIF techniques, where the signal is usually strongly influenced by quenching from collision partners, which depends on the unknown local bath gas composition. Thermographic phosphors exhibit a wide range of sensitivity and emit light at distinct wavelengths for an extended duration, allowing spectral and temporal discrimination against background signals. Many phosphors have a broad absorption spectrum in the region near 266 nm or 355 nm, allowing excitation using a single solid-state laser and avoiding the significant photochemical interferences generated by excitation with shorter wavelengths. Many also emit light in the visible spectral region, which can be recorded by non-intensified cameras. In conclusion, a combined thermometry and velocimetry technique based on thermographic phosphor tracer particles requires a comparatively simple experimental setup.
Thermographic phosphors have been widely used to measure surface temperatures and comprehensive review articles can be found [17,18]. Following excitation, on subsequent relaxation to the ground state these materials emit red-shifted light for prolonged times, which, depending on the phosphor, can be up to several milliseconds. Temperature influences both the phosphorescence decay time and emission spectrum, offering two distinct ways to perform thermometry.
First, the rate of non-radiative decay (quenching) is strongly dependent on temperature due to the increased vibrational energy of the phosphor, promoting multi-phonon emission and crossover relaxation [19]. As a result, for many phosphors the phosphorescence decay time varies with temperature by two or three orders of magnitude allowing very precise temperature determination. Using the lifetime method, point measurements [20,21] and also planar measurements [22,23] of surface temperature in combustion environments have been performed with a precision as high as 2 K. This technique has also been used to measure temperature in single droplets and sprays [24,25] where the phosphor was added in powder form to the liquid. 2D lifetime imaging can be realized using a CMOS camera with a maximum frame rate approaching MHz, allowing readout of up to 100,000 frames per excitation pulse and therefore exploiting the very wide range over which the lifetime varies for high-temperature phosphors. However, in that case, the measurement times are too long for turbulent flow investigations, and both a phosphor with a lifetime varying below the microsecond scale over the temperature range of interest and an even faster intensified framing camera would be required.
Second, the distribution of electrons among excited states is governed by the temperature-dependent Boltzmann distribution, and so the emission spectrum also changes with temperature [26]. Therefore, some thermographic phosphors can be used for thermometry by simultaneously measuring the change of the emission spectrum using two different interference filters, chosen to include emission lines or spectral regions that exhibit relative temperature dependency. The superposition and division of the two images results in an intensity ratio map, which can then be converted to a two-dimensional temperature image using a calibration curve previously obtained at known temperatures. This two-color technique has also been used to measure temperature on surfaces [27,28] and in sprays [29]. Some phosphors exhibit a strong change in the emission spectrum with temperature and can therefore be used for very precise thermometry over a small temperature region, limited by the dynamic range of the detector [30,31]. In practice, ratio-based methods are more straightforward to apply to planar measurements over a wide temperature range, using only two detectors and a single exposure.
For gas phase measurements in turbulent flows the time resolution must be short enough to avoid averaging over the rapidly fluctuating quantities, therefore limiting the integration time of the measurement. Because of this, the use of phosphors with short lifetimes in the microsecond range is clearly desirable in order to maximize collection of the phosphorescence emission during the short exposure time. Also, the tracer particles must be small enough to accurately trace the flow, but, together with a low seeding density, these restrictions on particle diameter and exposure time severely limit the phosphorescence signal. In the gas phase, thermographic phosphors have been used to measure instantaneous temperature fields during the ignition phase of an internal combustion engine [32]. Combined temperature and velocity measurements were performed in a turbulent heated jet using Mg4FGeO6:Mn4+, but the signal level was reported to be too low to allow correlated single shot measurements [33]. In another study, an up-conversion phosphor Y2O2S:Er2+,Yb3+ was employed to measure temperature in a similar heated jet, but 8 μm diameter particles were seeded to the flow [34].
In the following, we describe the use of the phosphorescence emission of seeded BaMgAl10O17:Eu2+ (BAM:Eu) phosphor particles for ratio-based planar measurements of gas temperature. Mie scattering from the same 2 μm particles is simultaneously recorded to measure the velocity field in a conventional PIV approach. Also the phosphorescence intensity, proportional to the tracer particle number density, is locally conditioned for temperature to simultaneously determine the mixture fraction. Theoretical models are used to address fundamental restrictions on the particle diameter to ensure accurate tracing of temperature and velocity. Spectroscopic analysis of the phosphor is carried out to characterize
the temperature dependence of the emission spectrum. Preliminary experiments are performed in the gas phase to investigate non-linearity of the phosphorescence emission with increasing laser intensity. Finally, the joint technique is applied to a turbulent jet electrically heated to temperatures between 300 and 700 K. Correlated single shots of temperature, velocity and mixture fraction are presented.
2. Phosphorescent tracer particles
2.1 Thermographic phosphor
2 μm BAM:Eu thermographic phosphor particles (KEMK63/UF-P1, Phosphor Technology) were used as a tracer. The ceramic material is chemically inert and has a high melting point of 2200 K [35]. The phosphorescence emission consists of a single broad and featureless band around 445 nm, characteristic of the strong coupling between the host and activator in phosphors doped with divalent rare-earths [36]. This emission has a favorable lifetime of 1 μs at room temperature owing to the parity-allowed transition between the 4f65d → 4f7 energy levels [37]. When the camera exposure time is limited by the flow timescales under investigation, the choice of a phosphor with such a short lifetime ensures that good signal levels are achieved from the small disperse particles in the gas phase.
With increasing temperature, the emission band broadens and shifts toward the UV as shown in Fig. 1 . To obtain the spectra, the phosphor powder was placed in an optically accessible temperature-controlled oven (Eurotherm 2416CG, Lenton Furnaces) and excited at 355 nm with a fluence of 1 mJ/cm2 using the third harmonic of a 5 Hz pulsed Nd:YAG laser (Quanta-Ray GCR-150, Spectra-Physics). The emission spectra were measured using a 300 mm focal length spectrometer (Acton SP-2300i, Princeton Instruments) with a grating groove density of 300 g/mm, f/4 optics and an interline transfer CCD camera (Imager Intense, LaVision). The entrance slit width was 100 μm, providing a spectral resolution of 1 nm. The transmittance/quantum efficiency of the complete detection system was previously calibrated using the reference spectrum of a tungsten halogen lamp (LS-1, Ocean Optics).
Fig. 1 BAM:Eu emission spectrum following 355 nm excitation, normalised to the emission band peak and recorded at 50 K intervals (1 nm spectral resolution). The measured transmission curves of the filters used in this study are superimposed on the spectrum.
Previous studies have reported evidence of thermal degradation occurring in BAM:Eu powder samples after extended (1 hour) exposure to temperatures exceeding 800 K [38], leading to an irreversible decrease in the room temperature emission intensity with increasing annealing temperature. When exposed to high temperatures the oxidation state of the europium atoms increases to Eu3+, and weak emissions characteristic of trivalent europium can be observed after annealing. Also, Eu2+ ions migrate from their original sites within the lattice to a location where excitation is only possible using wavelengths below 200 nm [38]. This lack of thermal stability requires further investigation to determine whether the extremely short residence times found in flames are sufficient to adversely affect the luminescent properties of BAM:Eu, and consequently its temperature tracing ability. For the present calibration, no differences were observed between spectra obtained during the heating and cooling phases for a maximum oven temperature of 700 K.
2.2 Particle tracing properties
This combined experimental technique indirectly determines the gas properties by measuring the movement and luminescent properties of seeded phosphor particles. Consequently the measurement accuracy relies on how rapidly the tracer particles adjust to changes in gas temperature and velocity.
2.2.1 Velocity tracing
For micrometer-sized particles, the Reynolds number is below unity for a wide range of gas viscosities and slip velocities, and a Stokes flow can be assumed. Considering a step change in gas velocity, the slip between the particle and the gas decays exponentially and can be characterized using the particle relaxation time constant [39]:
where ρ is the density, μ the dynamic gas viscosity and d the particle diameter, with subscripts p and g referring to the properties of the particle and gas respectively.The relaxation time scales with the square of the particle diameter (see Fig. 2 ). It also increases with a reduction in gas temperature owing to the decrease in viscosity, and scales with the tracer material density. Most phosphor materials such as BAM:Eu (3.70 g/cm3) and YAG-based phosphors (4.55 g/cm3) have a density similar to tracers widely used in PIV such as TiO2 (4.23 g/cm3) and Al2O3 (3.94 g/cm3), resulting in very similar response times. The 3τ (95%) relaxation times for 2 μm particles range between 53 and 65 μs at a typical mean flame temperature of 1150 K. Other analyses use the frequency of the turbulent velocity fluctuations to determine the appropriate particle size [40,41]. In this case, 2 μm particles of a similar density were reported to follow turbulence frequencies of 1 kHz in a lean methane-air flame [40]. If oscillations of higher frequency or second-order statistics in turbulent premixed flames need to be measured, smaller particles may be required [42].
Fig. 2 (a) Particle (YAG) velocity after a step change in gas velocity at 2000 K. (b) Particle (YAG) temperature following a step change in gas temperature from 300 to 2000 K.
2.2.2 Temperature tracing
This thermometry technique assumes thermal equilibrium between gas and particles, so that the gas temperature can be inferred from the temperature-dependent emission properties of the phosphor. Across a flame front, the gas temperature rapidly increases due to the reaction heat release. However, owing to the volumetric heat capacity of the particle, a large transfer of thermal energy through the poorly conductive gas is required in order to achieve thermal equilibrium. Consequently, an analysis of the particle temperature response is necessary before the technique is applied for turbulent flow diagnostics.
Assuming the tracer particles follow the flow, heat transfer to the particle is solely by conduction and radiation. A transient, one-dimensional, heat conduction model was used to simulate the particle response. The system consists of a spherical particle in an infinite gas, at different initial temperatures Tp0 and Tg0. The particle thermal conductivity kp, of the order 10 - 20 W/mK for some well characterized host materials such as YAG or Y2O3 [43], is several orders of magnitude larger than the gas conductivity and so the temperature inside the sphere can be considered uniform at all times.
To account for local temperature-dependent thermal properties of the gas and radiative heat transfer from the particle surface to the surroundings, a numerical model was developed using an implicit finite-difference scheme in spherical coordinates in order to investigate the temperature history of the sphere. To validate this procedure, the results from this numerical model were compared to analytical solutions of Konopliv and Sparrow [44]. In their work, the appropriate heat conduction equation assuming uniform gas thermal properties was solved using Laplace transforms and contour integration, resulting in solutions taking the form of integral expressions. In this work, these integral-form solutions were evaluated and compared with the results of the numerical model. Under the same conditions of uniform gas thermal properties and no radiative heat transfer, the agreement between the theoretical response times is excellent, as shown in the first two rows in Table 1 .
Table 1. Temperature Time Response and Steady State Radiation Error of YAG Particles
To investigate the temperature tracing ability of different particle sizes in practical applications, the initial particle temperature was set at 300 K and the initial gas temperature was varied. The particle thermal properties were assumed constant and were based on an undoped YAG crystal at 300 K, for which thermal properties are available [43].
The results are presented in Table 1. The response time t95 is the time when the normalized sphere temperature given by:
reaches 95% of its steady state value. If the initial gas temperature increases, the particle response time decreases because the gas thermal diffusivity increases with temperature. Also, when the temperature dependencies of the gas thermal properties are modeled, the response time increases for the case of a cold particle in a hot flow. This is because initially the gas in the particle vicinity is cooled down, which reduces the gas thermal diffusivity and limits the conductive heat transfer. Radiative heat transfer to the cold surroundings has also been investigated by considering the particle as a blackbody. Radiative heat loss has very little influence on the temperature response time but causes a small steady state error εT that increases with the initial gas temperature and the particle diameter. This error will be even lower for a more realistic emissivity of less than one. Comparing these results with those from the previous section, for the same YAG based phosphor material the temperature response is slightly faster than the response to a change in velocity, and it can be concluded that, in general, these particles are suitable for gas-phase temperature tracing. However, the particle heat capacity grows dramatically with particle diameter. For example, for larger particles of 8 μm diameter the response time is already 589 μs at 2000 K (Fig. 2), and therefore the accuracy with which temperature fluctuations are followed is strongly affected.3. Experimental setup
3.1 Heated jet configuration
A diagram of the experimental setup used for the measurements is shown in Fig. 3 . The test case was a turbulent free jet of air with a Reynolds number of approximately 10,000, electrically heated using temperature-controlled heaters (750 W in-line heater and 1256 W heating tape, Omega Engineering). A thermocouple positioned 15 mm above the 8 mm diameter nozzle was used to monitor the flow temperature and could be translated horizontally and vertically across the measurement plane using micrometer stages. The central jet was seeded with the phosphor and shielded by an 80 mm diameter coflow with a constant velocity of 0.3 m/s, seeded with 2 μm non-luminescent Al2O3 particles. Both powders were dried for 2 hours at 400 K in an oven before seeding into the flow using a reverse cyclone seeder [45] developed in-house to separate large agglomerates from the bulk flow. Furthermore, the powder was treated with a SiO2 coating to reduce cohesion forces between particles. These steps were essential to meet the particle size restrictions on flow tracing ability and also to provide the consistent, spatially homogeneous particle seeding required for the mixture fraction measurements.
Fig. 3 Experimental setup. λ/2: half wave plate; BSpol: polarizing beamsplitter; BS: beamsplitter; DM: dichroic mirror; SO: sheet optics; IF: interference filters and PBS: plate beamsplitter.
3.2 Phosphorescence imaging
The phosphor particles were excited at 355 nm using the 3rd harmonic of an Nd:YAG laser (Quanta-Ray GCR-150, Spectra-Physics) triggered at 5 Hz. The beam was expanded and focused using f = −50 mm and f = 400 mm cylindrical lenses to form a tall light sheet with a thickness of approximately 400 μm above the nozzle, as measured using burn marks on photosensitive paper. To achieve homogeneous illumination of the area of interest, the f = −50 mm lens was positioned some distance from the nozzle. The phosphorescent emission was detected by two hardware-binned (4 x 4) non-intensified interline transfer CCD cameras (Imager Intense, LaVision) with 50 mm f/1.4 Nikon lenses fitted with interference filters and a 50:50 plate beamsplitter, AR-coated on the reverse side. Two interference filters at 420-30 (Chroma Technology) and 460-10 nm (Edmund Optics) were used (notation is CWL-FWHM) to exploit the temperature dependent shift and broadening of the emission line for the ratio-based temperature measurement (see Fig. 1). The timing of the laser-camera system was controlled using a trigger clock (PTU9, LaVision). The variable laser-camera delay caused by synchronization between the camera clock and trigger clock was experimentally determined to be around 50 ns. To avoid the effects of this jitter on the intensity ratio, and to collect the entire phosphorescence decay, the camera exposure time was set to 5 μs, beginning 1 μs before the laser pulse. By seeding the flow with non-luminescent Al2O3 particles, it was also verified that the phosphorescence signal was unaffected by scattered 355 nm/532 nm light due to the high attenuation (optical density of 5) of the applied filters outside the passbands.
For this phosphor, the signal is in the spectral region near 500 nm where non-intensified CCD cameras have peak quantum efficiency. Interline transfer CCDs can achieve exposure times down to 1 μs with acceptable level of jitter, obviating the need for the fast gating offered by image intensifiers, which are associated with non-linearity and higher noise levels. However, in highly luminous flames, the rejection ratio after exposure and the minimum exposure time of interline CCD cameras may not be sufficient to avoid contributions from chemiluminescence or blackbody radiation from hot particles, requiring the use of intensified cameras. It may also be advisable to use a phosphor where the emission spectrum is both well separated from spectral lines of species chemiluminescence and at sufficiently short wavelengths in order to reduce the influence of thermal soot emissions. Laser-induced incandescence of soot particles can also be avoided by delaying the camera exposure after the laser pulse.
3.2.1 Phosphorescence linearity investigations
The linearity of the phosphorescence intensity with increasing excitation fluence was initially investigated in the gas phase at room temperature. The phosphorescence emission was recorded with one camera equipped with the 420-30 nm interference filter (see Fig. 3). The UV laser energy, progressively increased during the experiment, was recorded on a shot to shot basis using a calibrated energy monitor (EMV9, LaVision). In order to correct the phosphorescence signal for temporal variations in seeding density, the particle were also illuminated by a frequency-doubled Nd:YAG laser (Quanta-Ray LAB-150, Spectra-Physics) and the Mie scattering signal was recorded by the reflection camera fitted with a 532-10 nm interference filter (LaVision).
The results are presented in Fig. 4 . The phosphorescence intensity visibly departs from the linear regime at fluences above 2 mJ/cm2. Similar behavior was observed when performing an analogous test using bulk powder samples, but here the emission intensity keeps increasing with laser fluence, since, with this increase in fluence, the laser beam penetrates further into the bulk powder. This is different from the gas phase experiment where the number of phosphorescing particles is determined by the seeding density and the laser sheet volume. For these flow conditions, the large fluctuations in seeding density lead to the high noise level in the gas phase data.
Fig. 4 Phosphorescence emission intensity with increasing laser fluence in the gas phase and in the bulk powder.
The dependence of the intensity ratio on laser intensity was also tested in the gas phase using fluences up to 300 mJ/cm2 and no changes were observed, a finding that supports the results of previous investigations using the same phosphor [46]. For phosphorescence imaging in the gas phase, it is preferable to operate in the non-linear regime to maximize the signal and reduce the measurement uncertainty due to shot noise. However, in this regime the phosphorescence images cannot directly be corrected for spatial variations in laser illumination in order to determine the tracer number density, used to evaluate the mixture fraction. For this reason, the beam was expanded beyond the necessary height to obtain homogeneous illumination across the field of view. The FWHM of the beam profile in the vertical direction was 130 mm, measured using fluorescence from Coumarin 47 dye solved in methanol contained in a glass cell. The beam was positioned so that the maximum was vertically located in the center of the field of view, and in this configuration the laser illumination across the field of view varied by less than 2.5%. For all subsequent measurements the fluence was set to 100 mJ/cm2.
3.2.2 Thermometry
The intensity ratio technique requires division of two simultaneously acquired images. Their superposition is generally achieved using mapping software to correct for translation, magnification and rotation of the detectors and any other distortion caused by components on the optical path. Here, the use of two cameras and a 50:50 plate beamsplitter enables the entire field of view to be imaged in focus and with uniform magnification. Also, the measurement plane is observed from the same direction, avoiding any effects arising from inhomogeneous refractive index fields on the optical path. To eliminate any optical mismatch between the camera collection paths, one camera was mounted on micrometer stages, allowing precise matching of position in the x, y and z dimensions (see Fig. 3) using a target placed in the measurement plane and aligned with the laser sheets. CCD chip rotation could also be accounted for by altering the vertical angle of the beamsplitter. Alignment with this level of accuracy allowed direct division of the raw images prior to any software image mapping.
After acquisition, background images were recorded and subtracted from the image pairs which were then digitally filtered and smoothed using a 5 x 5 unweighted moving average filter (no binning). The final resolution was 400 μm, measured using an identically smoothed image of a resolution target (1951 USAF) positioned in the measurement plane. This corresponds to approximately 70 x 60 independent measurements across the 25 x 20 mm field of view. Then, a flat field correction was performed to correct for spatial non-uniformity in light collection efficiency, which was found to originate from the angular dependence of the transmission of the filters and beamsplitter. The spectral transmission profiles, sensitive to incident angle, are slightly broadened toward the UV even for a point directly on the optical axis because uncollimated light is imaged through the filter. Also, the incident angle of off-axis points in the measurement plane varies spatially, causing a slight variation in the effective filter transmission profiles over the field of view. This angular effect also depends on the spectrum of the incident light. Because of this, performing a white field correction using, for example, a narrow line LED source, leads to an over-estimation of the angular dependence in comparison with the actual phosphorescence, as the shift in the filter transmission profile results in an exaggerated intensity variation across the field of view. A pellet made from compressed BAM:Eu was positioned behind two diffusive ground glass screens and excited using the laser to generate a Lambertian light source with the correct spectrum. Because of the wavelength dependence of the white field correction, a temperature calibration curve for each individual pixel would be necessary. However, for the phosphor used in this study, the gradients in the emission spectrum (shown in Fig. 1) are small and do not vary significantly with temperature, so a flat field correction performed at room temperature was used to correct the actual non-isothermal measurements.
Finally, the processed ratio images were converted to temperature using a quadratic fit to in-flow calibration data. The data points were obtained by averaging over a volume measuring 1.2 x 0.8 x 0.4 mm in the x, y and z directions and centered at the thermocouple position during steady operation of the jet.
3.2.3 Mixture fraction
Within the limitations described in Section 2.2.1, small seeded particles can be used to track the turbulent mixing of two fluid streams because they follow the turbulent motion with reasonable accuracy. Each phosphor has a specific spectral signature which can be used to distinguish between different streams. In the present study, the central jet and co-flowing stream were seeded with phosphor and non-fluorescent alumina particles respectively, and the mixture fraction is determined from the phosphorescence intensity. However, a single phosphorescence image is not sufficient to determine the mixture fraction as the particle number density is also dependent on the local fluid density. In flows with large density variations, the simultaneous measurement of temperature is therefore required. Here, the simultaneously acquired temperature is used to condition each image recorded by one of the phosphorescence imaging cameras, before normalizing to a region near the jet exit.
To infer the particle number density from the phosphorescence signal in a non-isothermal flow, the temperature dependence of the filtered phosphorescence emission intensity σph must also be known. This calibration was performed using the bulk powder in the oven with the previously described spectroscopic setup. The transmission spectrum (see Fig. 1) of the interference filter was also measured using the tungsten halogen lamp in order to evaluate σph. The dependence of this filtered phosphorescence intensity on temperature is shown in Fig. 5 , which indicates thermal quenching of the radiative transition with increasing temperature.
The mixture fraction, defined at a given spatial position as the mass fraction of gas originating from the central jet, was calculated using:
where the terms account for the recorded phosphorescence intensity, the temperature dependent filtered emission intensity and the variable gas density respectively. A reference location containing a known mixture fraction is required, and the subscript ref designates the instantaneous quantities averaged over a region used to account for shot to shot fluctuations in the overall seeding density and laser intensity (first term) and exit jet temperature (last term). This volume had dimensions of 1.2 x 0.8 x 0.4 mm in the x, y and z directions and was located on the jet axis 6 mm from the nozzle exit.Equation (3) assumes uniform laser intensity, light collection efficiency, pressure and seeding density. Due to the non-linearity of the phosphorescence signal with laser intensity, it is not possible to correct the images using a dye cell and so uniform illumination of the particles was required, and achieved as described above. To reduce both temporal fluctuations and spatial inhomogeneities in laser illumination an efficient alternative would be the use of a beam homogenizer [47]. However, it should be noted that neither the thermometry nor the velocimetry needs a uniform beam profile, which is a considerable advantage in comparison with other scalar measurement techniques. Using the same Lambertian target (section 3.2.2) with a previously measured intensity distribution, the collection efficiency of the transmission camera was corrected for. The pressure is assumed uniform since the jet is subsonic.
While temporal fluctuations in the seeding density are accounted for in the first term of Eq. (3), spatially uniform seeding depends on the overall seeding quality and the size of the field of view. For this small imaged area, homogeneous seeding is difficult to achieve, and the single pixel standard deviation of the instantaneous seeding density was found to be 12% at 530 K, which limits the precision of the mixture fraction technique.
3.3 Particle image velocimetry
The same Nd:YAG laser used for the phosphorescence excitation was used to provide the first pulse for the PIV using the remaining 532 nm light after the third harmonic generation. The energy of this beam was further adjusted using a λ/2 plate and a polarizing beamsplitter. The second green pulse was provided by another frequency-doubled Nd:YAG laser (Quanta-Ray LAB-150, Spectra-Physics). The two green beams were combined using a 50:50 beamsplitter before overlapping with the UV beam using a dichroic mirror. The same cylindrical lenses were used for the UV and green light sheets and the alignment of all three beams was monitored both where the beams were superimposed and in the far field.
Mie scattering from the Al2O3 and phosphor particles was detected for PIV using an interline transfer CCD camera (Image Intense, LaVision) operating in double shutter mode, positioned on the opposite side of the measurement plane (see Fig. 3). The time between the laser pulses was 1 μs. Light was collected using a 105 mm f/2.8 Nikon lens with the f-stop at f/11 and a 532-10 nm interference filter (LaVision) to reject the phosphorescence emission. When the camera was positioned on the same axis as the phosphorescence detection system, light emitted from the particles and reflected by the PIV filter was found to interfere with the temperature measurement. For this reason the PIV camera was positioned at an angle and a Scheimpflug adaptor between the camera and lens was used to eliminate off-axis defocusing across the field of view. Perspective distortion of the particle images was corrected and then the images were scaled to the unprocessed phosphorescence images using a mapping program.
Processing of the particle images was carried out using a multi-pass cross-correlation algorithm (DaVis 7.2, LaVision) with an interrogation window size of 32 x 32 and 50% overlap, resulting in a final vector spacing that matched the true resolution of the temperature/mixture fraction data.
4. Results and discussion
In Fig. 6 average temperature images are shown for different jet conditions. Measurements in the core region were used to compute the intensity ratio-temperature response shown in Fig. 7 and the accuracy was evaluated using the mean deviation from the fit applied to these data points. Data was gathered from different measurement sequences recorded on different days and the repeatability was found to be excellent. Using this in-flow calibration method, an accuracy of 2% was achieved.
Fig. 6 Averages of 100 single shot temperature images for different jet conditions. (a)-(d) jet temperature T = 293, 483, 583 and 683 K.
Fig. 7 (a) Intensity ratio response measured in the gas phase. (b) Comparison of horizontal temperature profiles measured during steady operation of the jet at 530 K.
For comparison, jet temperature profiles were measured by translating a thermocouple through the flow 15 mm above the nozzle during steady operation. These data were compared to time-averaged temperature profiles extracted at the same vertical position from a single line of independent measurements obtained using the two-color phosphorescence technique, as shown in Fig. 7. The mean deviation from the thermocouple measurements is 8.5 K (1.6%). As the signal decreases in the shear layer away from the jet axis, the measured temperature deviates slightly from the thermocouple trace until the signal drops below the applied cutoff filter threshold. In this experiment, the two flows were deliberately seeded with different material to distinguish the two streams. Consequently, at a given position in the shear layer the temperature was sampled only from events when particles originating from the central jet are present. At the same position, colder, unseeded coflowing air that is unmixed or mixed such that the signal is below the filter threshold was not sampled, so the average of the sampled values is biased toward higher temperatures. This problem would not occur if both streams were seeded with thermographic phosphor.
Statistics to determine the single-pixel temperature precision were gathered from 600 measurements at each steady jet temperature, as shown in Fig. 8 . At room temperature, the single pixel standard deviation of the 600 sampled measurements is 4.1 K. As the temperature increases, the signal to noise ratio decreases due to thermal quenching of the radiative transition and the reduction in particle number density. As a result, lower signal levels are recorded in the central part of the jet, and the measurement precision drops to 32.7 K at 683 K.
Fig. 8 Histograms of 600 independent temperature measurements for each steady jet temperature recorded near the nozzle exit. The number of samples at 293 K has been reduced by a factor of 4 for improved visualisation.
Examples of simultaneously acquired single shots of temperature, velocity and mixture fraction are shown in Figs. 9 and 10 . In Fig. 9 the average velocity is subtracted from the instantaneous velocity to display the fluctuating component, and the single shot shows regions where the hot central jet is visibly contorted by the turbulence. Scalar quantities are transported by the turbulent velocity field and a good correlation between cooler regions and areas where the two streams mix is evident. Unlike the ratio technique used for the temperature measurement, the mixture fraction method is also dependent on local spatial inhomogeneities in the particle seeding density as mentioned in section 3.2.3, leading to a higher noise level. This results in a single pixel precision of 20% at 530 K.
Fig. 9 (a) Instantaneous temperature and velocity fluctuation image for a jet temperature of 530 K. (b) Average temperature and velocity fields compiled from 100 single shots.
Fig. 10 (a) Instantaneous mixture fraction image corresponding to Fig. 9. (b) Average mixture fraction image, determined from the same time series shown above.
5. Conclusions
This work demonstrates that precise, combined, single shot measurements of temperature, velocity and mixture fraction can be achieved using seeded thermographic phosphor particles. Theoretical models were used to characterize the ability of these solid particles to trace the gas temperature and velocity. It was found that the temperature response is slightly faster than the velocity response. Based on these results, a suitable particle diameter of 2 μm, also a common choice in PIV experiments, was identified as a compromise between tracing properties and signal intensity. Because of the high radiative transition rate of the chosen BAM:Eu phosphor, the decay lifetime of 1 μs at room temperature was well matched to the short exposure required and produced reasonable signal intensities, proving this phosphor to be a suitable candidate for turbulent flow measurements. Measurements of signal linearity also indicated saturation effects occurring at fluences as low as 2 mJ/cm2, but saturation has no effect on the thermometry technique. Evidence suggests that europium ions reside in specific sites within the BAM lattice [38]. Further work will focus on using a different excitation wavelength to access more dopant ions that occupy different sites, which would increase the saturation limit.
The experimental method for combined measurements was presented in detail. A two-color imaging system consisting of two precisely positioned non-intensified CCD cameras and a plate beamsplitter was employed, allowing good particle image registration with minimum relative distortion, together with a high detector sensitivity and linearity. One of the advantages of thermographic phosphors is that different fluid streams can be distinguished using their distinct emission lines, and in this study this property was exploited to simultaneously measure the mixture fraction. The effects of variable gas density and thermal quenching of the phosphorescence emission were accounted for by conditioning with the measured local temperature.
Single shot and average planar measurements of a turbulent heated jet test case were presented, with a temperature accuracy of 2% and a precision between 2 and 5%. This study demonstrates the utility of this new technique for precise vector-scalar measurements on a single shot basis, using a single tracer and simple instrumentation. In addition, a tracer material can be chosen from the wide variety of thermographic phosphors with different emission characteristics and integrated with the demonstrated technique to suit specific temperature ranges or applications. These simultaneous diagnostics are therefore promising for the future investigation of turbulent flows involving heat transfer or chemical reactions.
Acknowledgments
The authors gratefully acknowledge the financial support of the UK Engineering and Physical Sciences Research Council (EPSRC).
References
1. D. Geyer, A. Kempf, A. Dreizler, and J. Janicka, “Turbulent opposed-jet flames: A critical benchmark experiment for combustion LES,” Combust. Flame 143(4), 524–548 (2005). [CrossRef]
2. A. C. Eckbreth, Laser Diagnostics for Combustion Temperature and Species, 2nd ed. (Gordon and Breach Publishers, 1990).
3. J. N. Forkey, N. D. Finkelstein, W. R. Lempert, and R. B. Miles, “Demonstration and characterization of filtered Rayleigh scattering for planar velocity measurements,” AIAA J. 34(3), 442–448 (1996). [CrossRef]
4. D. Most and A. Leipertz, “Simultaneous two-dimensional flow velocity and gas temperature measurements by use of a combined particle image velocimetry and filtered Rayleigh scattering technique,” Appl. Opt. 40(30), 5379–5387 (2001). [CrossRef] [PubMed]
5. C. F. Kaminski, J. Engtröm, and M. Alden, “Quasi-instantaneous two-dimensional temperature measurements in a spark ignition engine using 2-line atomic fluorescence,” Proc. Combust. Inst. 27, 85–93 (1998).
6. P. R. Medwell, Q. N. Chan, P. A. Kalt, Z. T. Alwahabi, B. B. Dally, and G. J. Nathan, “Instantaneous temperature imaging of diffusion flames using two-line atomic fluorescence,” Appl. Spectrosc. 64(2), 173–176 (2010). [CrossRef] [PubMed]
7. R. Giezendanner-Thoben, U. Meier, W. Meier, J. Heinze, and M. Aigner, “Phase-locked two-line OH planar laser-induced fluorescence thermometry in a pulsating gas turbine model combustor at atmospheric pressure,” Appl. Opt. 44(31), 6565–6577 (2005). [CrossRef] [PubMed]
8. M. C. Thurber, F. Grisch, and R. K. Hanson, “Temperature imaging with single- and dual-wavelength acetone planar laser-induced fluorescence,” Opt. Lett. 22(4), 251–253 (1997). [CrossRef] [PubMed]
9. M. Löffler, F. Beyrau, and A. Leipertz, “Acetone laser-induced fluorescence behavior for the simultaneous quantification of temperature and residual gas distribution in fired spark-ignition engines,” Appl. Opt. 49(1), 37–49 (2010). [CrossRef] [PubMed]
10. B. K. McMillin, J. L. Palmer, and R. K. Hanson, “Temporally resolved, two-line fluorescence imaging of NO temperature in a transverse jet in a supersonic cross flow,” Appl. Opt. 32(36), 7532–7545 (1993). [CrossRef] [PubMed]
11. W. G. Bessler, F. Hildenbrand, and C. Schulz, “Two-line laser-induced fluorescence imaging of vibrational temperatures in a NO-seeded flame,” Appl. Opt. 40(6), 748–756 (2001). [CrossRef] [PubMed]
12. W. G. Bessler and C. Schulz, “Quantitative multi-line NO-LIF temperature imaging,” Appl. Phys. B-Lasers 78(5), 519–533 (2004). [CrossRef]
13. S. Pfadler, F. Beyrau, and A. Leipertz, “Flame front detection and characterization using conditioned particle image velocimetry (CPIV),” Opt. Express 15(23), 15444–15456 (2007). [CrossRef] [PubMed]
14. M. Yu, G. Särner, C. C. M. Luijten, M. Richter, M. Aldén, R. S. G. Baert, and L. P. H. de Goey, “Survivability of thermographic phosphors (YAG:Dy) in a combustion environment,” Meas. Sci. Technol. 21(3), 4 (2010). [CrossRef]
15. J. P. Feist, A. L. Heyes, and S. Seefeldt, “Oxygen quenching of phosphorescence from thermographic phosphors,” Meas. Sci. Technol. 14(5), N17–N20 (2003). [CrossRef]
16. J. Brübach, A. Dreizler, and J. Janicka, “Gas compositional and pressure effects on thermographic phosphor thermometry,” Meas. Sci. Technol. 18(3), 764–770 (2007). [CrossRef]
17. S. Allison and G. Gillies, “Remote thermometry with thermographic phosphors: Instrumentation and applications,” Rev. Sci. Instrum. 68(7), 2615–2649 (1997). [CrossRef]
18. M. Aldén, A. Omrane, M. Richter, and G. Särner, “Thermographic phosphors for thermometry: A survey of combustion applications,” Prog. Energ. Combust. 37(4), 422–461 (2011). [CrossRef]
19. G. Blasse and B. C. Grabmaier, Luminescent Materials (Springer-Verlag, 1994).
20. J. P. Feist, A. L. Heyes, and S. Seefeldt, “Thermographic phosphor thermometry for film cooling studies in gas turbine combustors,” Proc. Instn. Mech. Engrs Part A: J. Power and Energy 217(2), 193–200 (2003). [CrossRef]
21. J. Brübach, M. Hage, J. Janicka, and A. Dreizler, “Simultaneous phosphor and CARS thermometry at the wall-gas interface within a combustor,” Proc. Combust. Inst. 32(1), 855–861 (2009). [CrossRef]
22. A. Omrane, F. Ossler, and M. Aldén, “Two-dimensional surface temperature measurements of burning materials,” Proc. Combust. Inst. 29(2), 2653–2659 (2002). [CrossRef]
23. J. Brübach, T. Kissel, and A. Dreizler, “Phosphor thermometry at an optically accessible internal combustion engine,” in Laser Applications to Chemical, Security and Environmental Analysis, (Optical Society of America, 2010), paper LWA5.
24. A. Omrane, G. Särner, and M. Aldén, “Two-dimensional temperature imaging of single droplets and sprays using thermographic phosphors,” Appl. Phys. B-Lasers 79, 431–434 (2004). [CrossRef]
25. A. Omrane, G. Juhlin, F. Ossler, and M. Aldén, “Temperature measurements of single droplets by use of laser-induced phosphorescence,” Appl. Opt. 43(17), 3523–3529 (2004). [CrossRef] [PubMed]
26. J. P. Feist and A. L. Heyes, “The characterization of Y2O2S:Sm powder as a thermographic phosphor for high temperature applications,” Meas. Sci. Technol. 11(7), 942–947 (2000). [CrossRef]
27. L. P. Goss, A. A. Smith, and M. Post, “Surface thermometry by laser-induced fluorescence,” Rev. Sci. Instrum. 60(12), 3702–3706 (1989). [CrossRef]
28. A. Heyes, S. Seefeldt, and J. Feist, “Two-color phosphor thermometry for surface temperature measurement,” Opt. Laser Technol. 38(4-6), 257–265 (2006). [CrossRef]
29. J. Brübach, A. Patt, and A. Dreizler, “Spray thermometry using thermographic phosphors,” Appl. Phys. B-Lasers 83(4), 499–502 (2006). [CrossRef]
30. G. Särner, M. Richter, and M. Aldén, “Two-dimensional thermometry using temperature-induced line shifts of ZnO:Zn and ZnO:Ga fluorescence,” Opt. Lett. 33(12), 1327–1329 (2008). [CrossRef] [PubMed]
31. B. Li, J. Linden, Z. S. Li, A. A. Konnov, M. Aldén, and L. P. H. de Goey, “Accurate measurements of laminar burning velocity using the heat flux method and thermographic phosphor technique,” Proc. Combust. Inst. 33(1), 939–946 (2011). [CrossRef]
32. R. Hasegawa, I. Sakata, H. Yanagihara, G. Särner, M. Richter, M. Aldén, and B. Johansson, “Two-dimensional temperature measurements in engine combustion using phosphor thermometry,” SAE Paper, 2007–01–1883, 1797–1803 (2007).
33. A. Omrane, P. Petersson, M. Aldén, and M. Linne, “Simultaneous two-dimensional flow velocity and gas temperature measurements using thermographic phosphors,” Appl. Phys. B-Lasers 92, 99–102 (2008). [CrossRef]
34. A. Rothamer and J. Jordan, “Planar imaging thermometry in gaseous flows using upconversion excitation of thermographic phosphors,” Appl. Phys. B-Lasers O. 106(2), 435–444 (2012). [CrossRef]
35. D. Ravichandran, R. Roy, W. B. White, and S. Erdei, “Synthesis and characterization of sol-gel derived hexa-aluminate phosphors,” J. Mater. Res. 12(03), 819–824 (1997). [CrossRef]
36. B. Henderson and G. F. Imbusch, Optical Spectroscopy of Inorganic Solids (Oxford Science Publications, 1989), 2nd ed.
37. W. Yen, S. Shionoya, and H. M. Yamamoto, Phosphor Handbook, 2nd ed. (CRC Press, 2006).
38. Y. H. Wang and Z. H. Zhang, “Luminescence thermal degradation mechanism in BaMgAl10 O17:Eu2+ phosphor,” Electrochem. Solid St. 8(11), H97–H99 (2005). [CrossRef]
39. M. Raffel, C. Willert, S. Wereley, and J. Kompenhans, Particle Image Velocimetry: A Practical Guide, 2nd ed. (Springer, 2007).
40. F. Durst, A. Melling, and J. H. Whitelaw, Principles and Practice of Laser-Doppler Anemometry, 2nd ed. (Academic Press, 1981).
41. A. Melling, “Tracer particles and seeding for particle image velocimetry,” Meas. Sci. Technol. 8(12), 1406–1416 (1997). [CrossRef]
42. F. Picano, F. Battista, G. Troiani, and C. M. Casciola, “Dynamics of PIV seeding particles in turbulent premixed flames,” Exp. Fluids 50(1), 75–88 (2011). [CrossRef]
43. P. H. Klein and W. J. Croft, “Thermal conductivity, diffusivity, and expansion of Y2O3, Y3 Al5O12, and LaF3 in the range 77°K–300°K,” J. Appl. Phys. 38(4), 1603 (1967). [CrossRef]
44. N. Konopliv and E. M. Sparrow, “Transient heat conduction in non-homogeneous spherical systems,” Heat Mass Transfer 3, 197–210 (1970).
45. M. Glass and I. Kennedy, “An improved seeding method for high temperature laser doppler velocimetry,” Combust. Flame 29, 333–335 (1977). [CrossRef]
46. J. Lindén, N. Takada, B. Johansson, M. Richter, and M. Aldén, “Investigation of potential laser-induced heating effects when using thermographic phosphors for gas-phase thermometry,” Appl. Phys. B-Lasers 96(2-3), 237–240 (2009). [CrossRef]
47. S. Pfadler, F. Beyrau, M. Löffler, and A. Leipertz, “Application of a beam homogenizer to planar laser diagnostics,” Opt. Express 14(22), 10171–10180 (2006). [CrossRef] [PubMed]
