High-speed, two-dimensional synchrotron x-ray radiography and phase-contrast imaging are demonstrated in propulsion sprays. Measurements are performed at the 7-BM beamline at the Advanced Photon Source user facility at Argonne National Laboratory using a recently developed broadband x-ray white beam. This novel enhancement allows for high speed, high fidelity x-ray imaging for the community at large. Quantitative path-integrated liquid distributions and spatio-temporal dynamics of the sprays were imaged with a LuAG:Ce scintillator optically coupled to a high-speed CMOS camera. Images are collected with a microscope objective at frame rates of 20 kHz and with a macro lens at 120 kHz, achieving spatial resolutions of 12 μm and 65 μm, respectively. Imaging with and without potassium iodide (KI) as a contrast-enhancing agent is compared, and the effects of broadband attenuation and spatial beam characteristics are determined through modeling and experimental calibration. In addition, phase contrast is used to differentiate liquid streams with varying concentrations of KI. The experimental approach is applied to different spray conditions, including quantitative measurements of mass distribution during primary atomization and qualitative visualization of turbulent binary fluid mixing.
© 2017 Optical Society of America
Understanding the liquid distribution in atomizing sprays is crucial to the understanding of spray breakup phenomena and has been the subject of many prior investigations [e.g., 1–3]. The highly dynamic nature of these sprays has prompted the use of myriad diagnostic techniques to interrogate the behavior of primary and secondary breakup [4,5].
Measurements in the spray farfield, where the liquid has disintegrated into a droplet field, have been performed using patternation, interferometry, Mie scattering, shadowgraphy, holography, and laser-induced fluorescence. Liquid distribution and mixing measurements have been performed with mechanical patternation [3,6]. Phase Doppler interferometry  and visible light scattering [8,9] have proven indispensable for acquiring droplet size and velocity distributions. Spray imaging using schlieren and shadowgraphy has enabled mapping of optical density and liquid distribution fields , and multiple sources and detectors have been employed to acquire three-dimensional shadowgraphy . Holography has been applied to spray diagnostics to image three-dimensional droplet and ligament fields [12,13]. Interferometry has been used to measure the liquid sheet thickness in sprays . Laser sheet imaging of laser-induced fluorescence with Mie scattering has been used to measure droplet sizes , as well to differentiate two liquids to measure mixing in liquid-liquid injectors [16–18].
A major challenge with the application of optical diagnostics is the occurrence of multiple scattering events within the spray, making the measurement of quantitative liquid distributions difficult [9,19]. Due to these challenges, a new suite of dense spray diagnostics has been developed to analyze the near injector region of the spray . Ballistic imaging relies on an optical time-gate to discriminate multiply scattered photons [20–22]. Structured light illumination has been developed to spatially encode the light and reduce interferences from diffuse scatter [23,24]. Optical connectivity uses total internal reflection to trap light in the contiguous portion of a dense spray . Strong refraction at the liquid–gas interfaces of complex spray structures, however, can make it very challenging to acquire quantitative information about internal mass distributions.
While optical frequencies are susceptible to multiple scattering events and refraction, the dominant interaction between x-rays and sprays is attenuation. X-ray photon energies that range from 1 to 1000 keV are readily available from commercial laboratory-scale sources and at synchrotron facilities [26–50]. Time-averaged, two- and three-dimensional measurements have been made using broadband tube sources with limited spatio-temporal resolution [4,26–29]. Radiographic measurements have been compared with laser Doppler velocimetry, phase Doppler interferometry, and high-speed photography . An industrial source (with beam filtering to mitigate beam hardening) has been used to collect images for tomographic reconstruction on a time-averaged basis , and a high-flux medical source used for tomography has also been shown . Issues with using a polychromatic source have been addressed and measurements have been compared with a narrowband synchrotron source . A polycapillary tube has been employed to increase spatial resolution for tomographic reconstruction . Time-resolved radiographic imaging has been shown [30,31] and 1-kHz-rate tomography has been shown . While continuous-wave (CW) sources enable high-speed imaging with limited temporal resolution, flash sources offer excellent temporal resolution with relatively few (1–4) sequential images. Flash x-rays have been shown to freeze the flow with nanosecond exposure times of evaporating sprays to visualize the jet core  and of impinging jets to compare spatial and temporal frequencies from a synchrotron source . Recent upgrades at the Sector 7 bending-magnet beamline at the Advanced Photon source have allowed for a high intensity CW source of broadband x-rays capturing the benefits of both CW and flash x-ray tube sources.
Synchrotron x-ray sources, which produce high flux at high repetition rates and are nearly collimated, are ideal sources for x-ray radiography. Prior work has utilized focused x-ray beams for high spatial resolution along a focused path through the spray . They have produced a large quantity of high fidelity, time-resolved radiographic diesel-spray measurements [36–40]. Cavitation inside of an x-ray transparent injector has also been shown . Propulsion sprays have been studied for various injectors including gas-centered liquid swirl injectors , swirl coaxial injectors , and impinging jet injectors [28,34,44]. Synchrotron sources also permit small angle x-ray scattering measurements , phase contrast measurements , and measurements inside the injector . A narrow bandwidth x-ray beam can excite fluorescent tracers in the spray, and their emission can be differentiated to investigate mixing between gases , liquids , and liquids and gases . These studies utilized a filtered, narrowband x-ray spectrum to allow for easier discrimination between scattered and fluorescence photons. However, using a narrowband source limits the total flux available, a prohibitive drawback for high-speed imaging techniques. To increase the total x-ray flux the full spectrum may be utilized introducing issues with beam hardening due to the variation in attenuation coefficient at different x-ray energies.
Recent upgrades to the Sector 7 bending-magnet beamline at the Advanced Photon Source have allowed for a high intensity CW source of broadband x-rays, combining the benefits of synchrotron radiation as well as CW and flash x-ray tube source imaging. These upgrades enable high-fidelity imaging with a larger beam diameter than undulator sources previously used in spray imaging . This new capability is available to the x-ray and spray communities as a user facility. The current work examines this novel imaging ability to investigate the potential spatial resolution, signal-to-noise ratios, and measurement uncertainty when imaging the highly dynamic behavior in sprays. The additional photon flux allows two-dimensional radiography with microsecond or sub-microsecond time resolution. The beam characteristics and their effects on image quality are addressed, and the polychromatic nature of the beam and its effects on quantitative accuracy are discussed. This approach is demonstrated for use in an impinging jet spray and swirl spray. Pathways to future improvements and studies are also expressed.
2. Experimental setup
The current experiments were conducted at the 7-BM beamline of the Advanced Photon Source at Argonne National Laboratory. In previous studies at the 7-BM beamline , the raw x-ray emission was spectrally filtered and spatially focused. A double multilayer monochromator was used to create a monochromatic (ΔE/E = 1–4%) beam, and Kirkpatrick-Baez mirrors focused the beam to ~10 × 10 μm2 . The current work differs in two ways. First, no focusing optics were used; the beam is simply allowed to diverge from the x-ray source point (source size 80 × 200 μm2 FWHM, located 36 m from the experiment). Second, the raw emission from the x-ray source with no narrowband spectral filtering (“white” beam) was used. With the removal of the monochromator, the x-ray flux increases greatly, optical aberrations due to slope errors on the multilayer substrates were removed , and the field of view for imaging expands significantly.
Several challenges were presented when using the white beam that contains both much higher flux and much harder x-rays and a broader spectral range than the monochromatic beam used in previous experiments at the 7-BM beamline [34–50]. First, the raw power of the beam—approximately 0.6 W/mm2 at the scintillator—can easily damage equipment, either through excessive heating or radiation damage to the molecular structure of materials. The white beam also contains a large amount of high-energy x-rays, which tend to undergo scattering (elastic and Compton) as well as absorption. Stray high-energy x-rays scattered along the beam path are occasionally absorbed in a camera pixel. The relatively large amount of energy in each of these photons saturates the pixel, causing abnormally bright pixels (zingers), which degrade image quality.
A schematic of the experiment is shown in Fig. 1, with several features specifically designed to overcome the challenges of white-beam x-ray imaging. Upstream of the equipment in Fig. 1 are movable slits to control the x-ray beam size and reduce parasitic visible light emission form the scintillator, and a set of water-cooled filters to allow the spectrum to be prehardened while keeping high-energy photons scattered from the filters from interfering with the experiment. A chopper wheel was used to prevent x-ray degradation of various materials along the beam path prior to data collection. A lead shield and lead oxide window further reduced the scattered x-ray flux from reaching the camera, and few zingers were seen in the images as a result. After the spray, a 100-μm-thick LuAG:Ce scintillator screen absorbed x-rays and emitted visible light that reflected off a mirror and into the camera objective. Increasing the scintillator thickness would reduce the spatial resolution of the images and absorb a greater number of higher energy x-rays, thereby increasing the visible light output but reducing the image contrast. Water-cooled beam stops (composed of leaded bronze, copper, and lead) were used downstream of the mirror to absorb the remaining beam that transmits through the scintillator.
Two types of measurements were performed to determine the quantitative mass distribution (Type 1) and the qualitative mixing (Type 2). The same LuAG:Ce scintillator and high-speed camera (Photron SA-Z) were used in both measurements. The goal of the first measurement type (1) was to collect quantitative mass distribution data with a larger field of view and an unfiltered beam. The attenuation coefficient was modeled to convert the transmission data to equivalent path length. A 50-mm Nikon macro lens with f/1.2 aperture resulted in images with a magnification of 1.2:1, an image size of 21 μm/pixel, and a spatial resolution of 65 μm, defined as the 10–90% rise distance across a sharp opaque edge in the x-ray beam. The phase effects due to propagation were subpixel and were not considered in this imaging setup. The temporal resolution with the macro lens was defined by the 120 kHz frame rate of the camera with a 350 ns exposure. The nearly linear range of the CMOS camera was verified using white-light source and neutral density filters to vary the signal intensity. The images were dark-current subtracted and normalized by a flat-field image with no spray present. The goal of the second measurement type (2) was to obtain qualitative mixing data in a smaller field of view and a filtered beam. A Mitutoyo 5x objective (NA = 0.14) with the standard Mitutoyo tube lens and 200 mm extension tube was used resulting in a 3.8-μm-image pixel size and a spatial resolution of 12 μm. The phase effects due to propagation were noticeable and were considered in this imaging setup. A 50-μm-thick copper filter was used to preferentially attenuate the softer x-rays; this decreased the attenuation of x-rays through water and increased the contrast at the liquid-gas interfaces for slight phase contrast imaging. The temporal resolution with the microscope was defined by the 20 kHz frame rate of the camera with a 2.5 μs exposure. The low light levels collected with the microscope and filter resulted in pixel intensities in the non-linear region of the CMOS detector. The decreased signal-to-noise ratio (SNR), non-linear detector response, and use of two fluids with varying composition yielded qualitative attenuation measurements that were capable of tracking the internal mixing behavior of the two jets. The attenuation coefficient was not modeled for the filtered beam because the results were qualitative. The spray was 370 mm from the scintillator, resulting in ~1% geometric magnification and a penumbra thickness of 2 μm, which is less that one pixel width for both imaging objectives.
A further challenge is associated with processing data from a polychromatic source with a spatially dependent x-ray energy spectrum. The polychromatic nature of the white beam requires a modeled attenuation coefficient due to beam hardening effects. As the x-rays transmit through material, lower energy x-rays tend to be absorbed more strongly than higher energy photons. As such, the effective attenuation coefficient decreases as the beam transmits through more material. In addition, the signal from the scintillator is caused by the photons absorbed by the scintillator crystal; as such, one must model the spectral absorption of the scintillator to properly account for the x-ray spectrum. Moreover, the spectrum of the white beam varies in the vertical direction, with more high-energy photons found in the center of the beam than along the top and bottom edges.
The x-ray spectra, attenuation of materials, and conversion to visible light were modeled with combined data from NIST  and X-ray Oriented Program (XOP) . XOP provides a database for x-ray spectra given the beamline specifications as a function of angle and for spectrally resolved attenuation coefficients for common materials from the NIST database. The model starts with the x-ray source spectrum from XOP and accounts for the attenuation from windows, filters, and air, as well as for the conversion to visible light in the scintillator. It is assumed the spectrum of visible output from the scintillator does not vary with x-ray energy or intensity. This model is then used to compute the attenuation coefficient of pure water and different water–KI solutions (10–50% KI by mass was added to increase image contrast) as a function of equivalent liquid path length or transmission. The liquids used in the model and experiments will be referred to as water for pure water and 10% KI or 50% KI, for KI solutions. The intensity of the visible emission scales proportionally with the x-ray power absorbed by the scintillator, also affecting the attenuation coefficient as measured by a visible detector. If the detector pixel intensity response is assumed to be linear and consistent between pixels, or if any non-linearity or non-uniformity is accounted for, the variable attenuation coefficient can be determined through a calibrated model. The model accounts for the angular dependence of the x-ray spectrum in the vertical direction, and therefore a variable attenuation coefficient was determined for each row of pixels in the images. The model does not account for the slow decay component of the scintillator emission lifetime–on the order of microseconds. No significant ghosting affects were observed in the images, and a scintillator with a quicker response is recommended for future measurements .
A sample of this model that accounts for x-ray energies from 1 to 200 keV is shown in Fig. 2 that displays the details from 1 to 100 keV (there are no absorption features between 64 and 200 keV). Figure 2(a) shows the attenuation coefficients of iodine, potassium, and water as a function of x-ray energy. An important feature of iodine as a contrast-enhancing agent is the absorption edge at 33.1 keV, allowing much higher absorption at greater x-ray energies compared with water. The x-ray source spectra at the center of the beam and 2.1 mm from the beam (representing the angle of the beam at the upper and lower edges of the images) are shown in Fig. 2(b), denoted as “Center Beam” and “Edges Beam”. Also shown are the effective spectra after the softer x-rays were preferentially absorbed by the scintillator at the center (“Center Scint.”) and edges (“Edges Scint.”) of the beam. These spectra do not account for any spectral changes caused by windows, air, or the spray. The effect of an absorption edge from the scintillator appears at 63.3 keV in Fig. 2(b).
Spectral beam hardening is apparent in Fig. 3(a) by the shift in the beam-averaged spectrum absorbed by the scintillator (denoted as “Source”) to increased x-ray energies as the beam passes though various materials. The “prehardened” spectrum in the absence of the spray was hardened by three 0.25 mm beryllium windows, a 0.1 mm Kapton window, and 700 mm air along the path to the detector. Figure 3 also shows the effect of beam hardening by 1 mm of pure water, 10% KI, and 50% KI, after being prehardened by the windows and air. Because of the shift to higher x-ray energies, the effective attenuation coefficient decreases for low transmissions occurring for longer path lengths. Conversely, the attenuation coefficient increases for high transmissions occurring for shorter path lengths because the relative contribution of softer x-rays, as shown in Fig. 3(b).
Once converted to visible light using the scintillator and captured by the camera, pixel counts represent the intensity of transmitted x-rays through the spray and were converted to an equivalent path length (EPL) of liquid using the modeled attenuation coefficients. The conversion of transmitted light to EPL coupled the beam hardening in the spray, scintillator absorption, and the visible light emission of the scintillator into a single transfer function. The transfer function is embedded in a variable attenuation coefficient that accounts for the spectral changes that are a function of the path length of a liquid of uniform KI concentration and therefore of the detected visible light. In a binary system where it is assumed only the liquid is attenuating the x-rays and the effect of the deficit due to air attenuation is neglected, the dependence can be described as a function of the detected visible light and KI concentration. This is shown in the Beer-Lambert law in Eq. (1).
Therefore the attenuation coefficient (α) is reduced to a function of the detected visible light (I/I0), KI concentration (CKI), and vertical location within the beam (h) such that the EPL can be determined without iteration. The initial model underestimated the experimental results of the path length through a single 0.5 mm jet by approximately 13% for water and 10% KI, as determined through comparison with 0.5 mm liquid jets. The error was accounted for by adjusting the path lengths with a multiplicative correction factor.
Variation in the spectrum of the beam not predicted by the model reached ± 6% uncertainty in the upper and lower 0.5 mm of the images shown with the macro lens. The integrated beam intensity fluctuated ± 0.6% RMS; a small portion of the beam (100 pixels) was sampled to reduce these fluctuations to ± 0.1% RMS. The attenuation as measured across a static cylindrical cuvette of KI solution, fluctuated ± 0.1% RMS at 63% attenuation.
Due to the variable attenuation coefficient and high levels of attenuation the SNR does not increase linearly or monotonically with EPL. The measurement precision was modeled, using the standard deviation of the attenuation values in the normalized images, shown in Fig. 4. The RMS noise in the unattenuated signal (I0) was 4% and 1% for water and 10% KI data, respectively. This attenuation noise was converted to an uncertainty in the calculated EPL, shown as the error bars in Fig. 4(a). As the EPL increases the attenuation increases logarithmically; this causes a decrease in sensitivity at greater levels of attenuation. The SNR was determined for given EPL values and is shown in Fig. 4(b). The SNR rolls off and flattens around 2.5 mm or 50% attenuation.
3. Results and discussion
The synchrotron-based, high-speed x-ray imaging technique was applied to study the quantitative mass distributions (Type 1) and qualitative mixing (Type 2) in an impinging jet spray, as well as quantitative mass distributions in a swirl spray (Type 1).
3.1 Impinging jet spray
The like-doublet impinging jet injector was comprised of two, D = 0.5 mm diameter, liquid jets colliding with an enclosed angle of 60 degrees. The liquid jet velocities of V = 10 m/s corresponded to a jet Reynolds number, ReD = VD/ν, of 5,000, where ν is the kinematic viscosity of the liquid. To measure the overall liquid distribution during impingement, 10% KI was used. As shown in Fig. 5, the turbulent liquid jets impinge in free space, and they interact while spreading into an unstable liquid sheet that disintegrates further into ligands and droplets during primary and secondary breakup, respectively. The high-speed images allowed for temporal tracking of the EPL for features such as impact wave dynamics and subsequent breakup. Note that (dark) regions of high attenuation (i.e., high EPL) in Fig. 5(a) represent long path lengths generated by the liquid spreading into the plane perpendicular to the liquid jets, as shown in Fig. 5(b).
Line plots along the vertical direction in the impinging jets are shown in Fig. 6 at different time instances separated by 50 μs. A series of impact waves, such as a feature at 1.2 mm, can be tracked at different times. The impingement location was defined at 0 mm where the centerlines of the jets would meet in space. At a location of 3.7 mm, the thickness of the liquid sheet has dropped to the level that was very close to zero, with undulations representing the passing of the impact waves.
Note that in prior work, it was possible to track the signal in time with high dynamic range for a single point in space or a two-dimensional image with limited dynamic range at a single time instant . The data in this work capture the two-dimensional evolution in time with high dynamic range and spatio-temporal resolution.
To investigate the mixing between the two jets, one jet contained 50% KI while the opposite jet was water to differentiate the two jets based on contrast from differential attenuation. Changes to the experimental setup include filtering the beam and using the long distance microscope objective. The 50-μm-thick copper filter (located within the evacuated section of the beam) was used to pre-harden the beam to allow greater transmission through the unseeded jet, thereby increasing the contrast between the two liquids. The pre-hardened x-ray spectrum passed through the water jet with minimal attenuation and the KI jet attenuated ~50% of the x-rays. In addition, the five-times magnification of the microscope objective was used to increase the spatial resolution and show slight phase contrast (due to the propagation) apparent at the edges of the unseeded water jet. This was sufficient to identify the liquid jet boundary without compromising the spatial resolution capable of identifying droplets as small as ~12 μm. Of primary interest is how the jets meet and transmit through the opposite jet, revealing via instantaneous two-dimensional images for the first time the mechanism of transmitive mixing that was the subject of speculation from prior time-averaged mass distribution measurements by [3,6,44] and others. The data show that bulk transmission of one jet can occur intermittently just below the point of impingement, simultaneously thinning the other jet and causing subsequent sheet instabilities that propagate downstream (Fig. 7).
3.2 Swirl spray
To illustrate the imaging capability for different spray geometries, a second injector study was conducted using a pressure-swirl atomizer with an internal exit diameter of 3.2 mm and mass flow rates of 25.8 g/s and 43 g/s. Water and 10% KI delivered via a pressurized flow tank were used in the study to investigate effects the attenuation level on image quality. Volumetric flow meters, pressure transducers, and thermocouples were used to compute state properties.
As shown in the images in Fig. 8 and the line plots of Fig. 9, the measurement technique was capable of capturing time-resolved images of water. Contrast in the images was increased significantly by adding KI, substantially increasing x-ray attenuation and SNR. This can be observed by comparing horizontal scans across the images in Fig. 9, which have been processed using with the x-ray attenuation converted to EPL in accordance with the Beer-Lambert Law. Note that the addition of the contrasting agent was accounted for and EPL values of the different scans were found to be nearly identical. This provides further confidence in the methods used to compute the attenuation coefficient despite significant spatial and spectral variations in the source beam.
High-speed, synchrotron-based x-ray radiography and phase contrast imaging were employed to measure time-resolved quantitative mass distributions and qualitative mixing in an impinging jet injector, as well as quantitative mass distributions in a liquid-swirl injector. The measurements were performed using a new polychromatic “white beam” x-ray beamline at the Advanced Photon Source. The polychromatic x-ray spectrum was modeled to determine the attenuation coefficient that was dependent upon the path length of x-rays through the liquid and variations in intensity and spectral characteristics within the beam itself. The uncertainty of 50 μm in EPL through water mixed with 10% KI allows quantitative imaging of the EPL down to very thin layers, while the spatial resolution allowed for visualization of very small droplets.
Two imaging systems were utilized to demonstrate the spatio-temporal resolution of the measurements. A macro lens collected images at a rate of 120 kHz with an exposure of 350 ns and a spatial resolution of 65 μm. A microscope objective collected images at rates of 20 kHz with an exposure of 2.5 μs and a spatial resolution of 12 μm. In addition, the microscope objective had sufficient spatial resolution to resolve the phase effects at the edges of the jets for studies of jet mixing dynamics. These date illustrate the ability of the measurement system to capture both quantitative and qualitative features of complex spray geometries. Future work includes detailed study of primary and secondary breakup processes using this new measurement capability.
National Research Council Post-doctoral Research Associateship award at the Air Force Research Laboratory, Aerospace Systems Directorate, Wright-Patterson AFB; U.S. Department of Energy under Contract No. DE-AC02-06CH11357.
The measurements were performed at the 7-BM beamline of the Advanced Photon Source, Argonne National Laboratory. This manuscript has been cleared for public release by the Air Force Research Laboratory (No. 88ABW-2016-4318).
References and links
1. A. Lefebvre, Atomization and Sprays (Taylor and Francis, 1989).
2. V. Yang and W. Anderson, Liquid Rocket Engine Combustion Instability (AIAA Inc., 1995).
3. N. Ashgriz, W. Brocklehurst, and D. Talley, “Mixing mechanisms in a pair of impinging jets,” J. Propuls. Power 17(3), 736–749 (2001). [CrossRef]
4. T. R. Meyer, M. Brear, S. H. Jin, and J. R. Gord, Formation and Diagnostics of Sprays in Combustion, in Handbook of Combustion (Wiley, 2010).
5. M. A. Linne, “Imaging in the optically dense regions of a spray: a review of developing techniques,” Prog. Energ. Combust. 39(5), 403–440 (2013). [CrossRef]
6. J. H. Rupe, “The liquid-phase mixing of a pair of impinging streams,” Jet Propulsion Laboratory Progress Report 20–195, (1953).
8. R. A. Dobbins, L. Crocco, and I. Glassman, “Measurement of mean particle sizes of sprays from diffractively scattered light,” AIAA J. 1(8), 1882–1886 (1963). [CrossRef]
9. D. G. Talley, A. T. S. Thamban, V. G. McDonell, and G. S. Samuelsen, “Laser sheet visualization of spray structure,” Recent Advances in Spray Combustion: Spray Atomization and Drop Burning Phenomena (AIAA Inc., 1995).
10. G. Settles, Schlieren and Shadowgraph Techniques (Springer, 2001).
11. J. Klinner and C. Willert, “Tomographic shadowgraphy for three-dimensional reconstruction of instantaneous spray distributions,” Exp. Fluids 53(2), 531–543 (2012). [CrossRef]
13. D. R. Guildenbecher, P. L. Reu, H. L. Stuaffacher, and T. Grasser, “Accurate measurement of out-of-plane particle displacement from the cross correlation of sequential digital in-line holograms,” Opt. Lett. 38(20), 4015–4018 (2013). [CrossRef] [PubMed]
14. Y. J. Choo and B. S. Kang, “Parametric study on impinging-jet liquid sheet thickness distribution using an interferometric method,” Exp. Fluids 31(1), 56–62 (2001). [CrossRef]
15. P. Le Gal, N. Farrugia, and D. A. Greenhalgh, “Laser sheet droplet sizing of dense sprays,” Opt. Laser Technol. 31(1), 75–83 (1999). [CrossRef]
16. V. McDonell, V. Phi, S. Samuelsen, A. Nejad, M. Shahnam, C. Guernsey, and R. Carlson, “Structure of sprays generated by unlike doublet injectors,” AIAA Paper 99–2464 in 35th AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit, Los Angeles, CA, 20–23 June, 1999. [CrossRef]
17. K. Jung, H. Koh, and Y. Yoon, “Assessment of planar liquid-laser-induced fluorescence measurements for spray mass distributions of like-doublet injectors,” Meas. Sci. Technol. 14(8), 1387–1395 (2003). [CrossRef]
18. T. Yuan and B. Huang, “Optical analysis of the mixing effect in fully developed like-doublet impinging jet sprays,” At. Sprays 22(5), 391–408 (2012). [CrossRef]
21. J. B. Schmidt, Z. D. Schaefer, T. R. Meyer, S. Roy, S. A. Danczyk, and J. R. Gord, “Ultrafast time-gated ballistic-photon imaging and shadowgraphy in optically dense rocket sprays,” Appl. Opt. 48(4), B137–B144 (2009). [CrossRef] [PubMed]
22. H. Purwar, S. Idlahcen, C. Rozé, D. Sedarsky, and J.-B. Blaisot, “Collinear, two-color optical Kerr effect shutter for ultrafast time-resolved imaging,” Opt. Express 22(13), 15778–15790 (2014). [CrossRef] [PubMed]
23. E. Berrocal, E. Kristensson, M. Richter, M. Linne, and M. Aldén, “Application of structured illumination for multiple scattering suppression in planar laser imaging of dense sprays,” Opt. Express 16(22), 17870–17881 (2008). [CrossRef] [PubMed]
24. E. Kristensson, E. Berrocal, and M. Aldén, “Quantitative 3D imaging of scattering media using structured illumination and computed tomography,” Opt. Express 20(13), 14437–14450 (2012). [CrossRef] [PubMed]
25. G. Charalampous, Y. Hardalupas, and A. M. K. P. Taylor, “Novel technique for measurements of continuous liquid jet core in an atomizer,” AIAA J. 47(11), 2605–2615 (2009). [CrossRef]
26. B. Heine, B. Balewski, and C. Tropea, “Experimental Investigation of the Correlation Between Nozzle Flow and Spray using Laser Doppler Velocimetry, Phase Doppler System, High-speed Photography, and X-ray Radiography,” At. Sprays 20(1), 57–70 (2010). [CrossRef]
27. F. Coletti, M. J. Benson, A. L. Sagues, B. H. Miller, R. Fahrig, and J. K. Eaton, “Three-Dimensional Mass Fraction Distribution of a Spray Measured by X-Ray Computed Tomography,” J. Eng. Gas Turbines Power 136(5), 051508 (2014). [CrossRef]
28. B. R. Halls, T. J. Heindel, A. L. Kastengren, and T. R. Meyer, “Evaluation of X-ray sources for quantitative two- and three-dimensional imaging of liquid mass distribution in atomizing sprays,” Int. J. Multiph. Flow 59, 113–120 (2014). [CrossRef]
29. L. Marchitto, D. Hampai, S. B. Dabagov, L. Allocca, S. Alfuso, C. Polese, and A. Liedl, “GDI spray structure analysis by polycapillary X-ray μ-tomography,” Int. J. Multiph. Flow 70, 15–21 (2015). [CrossRef]
30. K. Kuo, K. C. Hsieh, and J. M. Char, “Observations of Breakup Processes of Liquid Jets Using Real-Time X-ray Radiography,” J. Propuls. Power 6(5), 544–551 (1990). [CrossRef]
31. R. D. Woodward, S. Pal, R. J. Santoro, and K. K. Kuo, “Measurement of core structure of coaxial jets under cold-flow and hot-fire conditions,” in Recent Advances in Spray Combustion: Spray Atomization and Drop Burning Phenomena (AIAA Inc., 1995), pp. 185–209.
32. J. Lim, Y. Sivathanu, and M. Wolverton, “Evaluation of Soft X-Ray Absorption Tomography for the Near Injector Charicterization of Dense Sprays,” in 25th ILASS-Americas Conference (2013).
33. A. Birk, M. McQuaid, and M. Gross, “Liquid core structure of evaporating sprays at high pressures –Flash X-ray Studies,” Report from the Army Research Lab, Aberdeen Proving Ground-TR-901 (1995).
34. B. R. Halls, S. Roy, J. R. Gord, A. L. Kastengren, and T. R. Meyer, “Quantitative imaging of time-resolved liquid distributions in sprays using broadband flash x-ray radiography,” Int. J. Multiph. Flow (Accepted).
35. A. Kastengren, C. F. Powell, D. Arms, E. M. Dufresne, H. Gibson, and J. Wang, “The 7BM beamline at the APS: a facility for time-resolved fluid dynamics measurements,” J. Synchrotron Radiat. 19(Pt 4), 654–657 (2012). [CrossRef] [PubMed]
37. A. G. MacPhee, M. W. Tate, C. F. Powell, Y. Yue, M. J. Renzi, A. Ercan, S. Narayanan, E. Fontes, J. Walther, J. Schaller, S. M. Gruner, and J. Wang, “X-ray imaging of shock waves generated by high-pressure fuel sprays,” Science 295(5558), 1261–1263 (2002). [CrossRef] [PubMed]
38. W. Cai, C. F. Powell, Y. Yue, S. Narayanan, J. Wang, M. W. Tate, M. J. Renzi, A. Ercan, E. Fontes, and S. M. Gruner, “Quantitative analysis of highly transient fuel sprays by time resolved X-radiography,” Appl. Phys. Lett. 83(8), 1671–1673 (2003). [CrossRef]
39. F. X. Tanner, K. Feigl, S. A. Ciatti, C. F. Powell, S.-K. Cheong, J. Liu, and J. Wang, “Structure of high-velocity dense sprays in the near-nozzle region,” At. Sprays 16(5), 579–598 (2006). [CrossRef]
40. A. L. Kastengren, C. F. Powell, Y. Wang, K.-S. Im, and J. Wang, “X-ray radiography measurements of diesel spray structure at engine-like ambient density,” At. Sprays 19(11), 1031–1044 (2009). [CrossRef]
41. D. J. Duke, A. L. Kastengren, F. Z. Tilocco, A. B. Swantek, and C. F. Powell, “X-ray radiography measurements of cavitating nozzle flow,” At. Sprays 23(9), 841–860 (2013). [CrossRef]
42. S. A. Schumaker, A. L. Kastengren, M. D. A. Lightfoot, S. A. Danczyk, and C. F. Powell, “A study of gas-centered swirl coaxial injectors using X-ray radiography,” in Technical Paper AFRL-RZ-ED-TP-2012-070, Air Force Research Lab, Edwards AFB CA Propulsion DIR/Space and Missile Propulsion DIV/Aerophysics Branch 12th ICLASS (2012).
43. C. J. Eberhart, D. M. Lineberry, R. A. Frederick Jr, and A. L. Kastengren, “Mechanistic assessment of swirl coaxial injection by quantitative x-ray radiography,” J. Propuls. Power 30(4), 1070–1079 (2014). [CrossRef]
44. B. R. Halls, T. R. Meyer, and A. L. Kastengren, “Quantitative measurement of binary liquid distributions using multiple-tracer x-ray fluorescence and radiography,” Opt. Express 23(2), 1730–1739 (2015). [CrossRef] [PubMed]
45. K.-C. Lin, M. Ryan, C. Carter, A. Sandy, S. Narayanan, J. Ilavsky, and J. Wang, “Investigation of condensed supercritical ethylene jets using Small Angle X-ray Scattering (SAXS) technique,” Nucl. Instrum. Methods Phys. Res. A 649(1), 219–221 (2011). [CrossRef]
46. K.-C. Lin, C. Rajnicek, J. McCall, C. Carter, and K. Fezzaa, “Investigation of pure- and aerated-liquid jets using ultra-fast X-ray phase contrast imaging,” Nucl. Instrum. Methods Phys. Res. A 649(1), 194–196 (2011). [CrossRef]
47. A. L. Kastengren, F. Z. Tilocco, C. F. Powell, J. Manin, L. M. Pickett, R. Payri, and T. Bazyn, “Engine combustion network (ECN): Measurements of nozzle geometry and hydraulic behavior,” At. Sprays 22(12), 1011–1052 (2012). [CrossRef]
49. C. D. Radke, J. P. McManamen, A. L. Kastengren, B. R. Halls, and T. R. Meyer, “Quantitative time-averaged gas and liquid distributions using x-ray fluorescence and radiography in atomizing sprays,” Opt. Lett. 40(9), 2029–2032 (2015). [CrossRef] [PubMed]
50. A. L. Kastengren and C. F. Powell, “Synchrotron X-ray techniques for fluid dynamics,” Exp. Fluids 55(3), 1686 (2014). [CrossRef]
51. P. J. Eng, M. Newville, M. L. Rivers, and S. R. Sutton, “Dynamically figured Kirkpatrick Baez x-ray micro-focusing optics,” P. Soc. Photo-Opt. Ins. 3449, 145–156 (1998).
52. A. Kastengren, C. F. Powell, D. Arms, E. M. Dufresne, H. Gibson, and J. Wang, “The 7BM beamline at the APS: a facility for time-resolved fluid dynamics measurements,” J. Synchrotron Radiat. 19(Pt 4), 654–657 (2012). [CrossRef] [PubMed]
54. M. Sanchez del Rio and R. J. Dejus, “Status of XOP: v2.4: recent developments of the x-ray optics software toolkit,” SPIE Proc., 814115 (2011).
55. M. E. Rutherford, D. J. Chapman, T. G. White, M. Drakopoulos, A. Rack, and D. E. Eakins, “Evaluating scintillator performance in time-resolved hard X-ray studies at synchrotron light sources,” J. Synchrotron Radiat. 23(Pt 3), 685–693 (2016). [CrossRef] [PubMed]