Investigation Method

## **Chapter 2**

## Imaging Diagnostics for Jet Breakup into Droplets: A Review

*Anu Osta*

## **Abstract**

A concise review of the recent developments in some of the standard optical diagnostics applied for primary jet breakup studies has been presented here. Primary breakup is the core breakup of liquid jets and sheets into droplets upon its interaction with the ambient gaseous atmosphere. This phenomenon is encountered in various aerodynamic, fluid dynamic, and combustion situations. The imaging diagnostics reviewed here include photography, high-speed imaging, shadowgraphy, digital holography, ballistic imaging, jet core illumination, thermal imaging, Mie imaging, x-ray phase contrast imaging, and laser-induced fluorescence. The advantages and limitations of each technique, their success, and future developmental trend are discussed.

**Keywords:** atomization, diagnostics, liquid-jet, visualization

## **1. Introduction**

Liquid atomization is a phenomenon associated with liquid fuel combustion processes, industrial and agricultural sprays, and our daily life activities such as when using body or hair sprays. When a liquid column or a sheet issuing from a source (nozzle or channel), interacts with the ambient atmosphere instabilities develop inside the liquid sheet or the column core. Instabilities also develop at the atmosphere-liquid interface in the form of flow structures like surface waves and ligaments. Flow conditions prevailing inside the source boundary such as void cavities or turbulence also affect instability development and the resulting bulk liquid disintegration. **Figure 1** depicts typical atomization in a liquid jet that is subjected to a crossflowing fluid like air. The liquid disintegration is also referred as the 'primary breakup' process. Primary jet breakup is often regarded as the first step in a jet atomization process. The instabilities are mostly of the Kelvin–Helmholtz (KH) and Rayleigh Taylor (RT) type. The primary droplets formed in the initial stage of atomization may then undergo subsequent secondary breakup stages.

Complete characterization of spray atomization would require one to analyze it at both the macroscopic and microscopic scales. Macroscopic characteristics would reveal the spray volume and its penetration into the surrounding atmosphere (both axially and radially), spray cone angle, mass flow rate, spray momentum flux, and the mass distribution of the spray fluid further downstream after atomization. The microscopic properties would reveal droplet sizes, droplet velocities, droplet number density, droplet distribution, and the temperature field of the spray.

**Figure 1.** *Schematic representation of primary breakup of a liquid jet in crossflow.*

Over the past few decades, various techniques [1–5] have developed both intrusive and non-intrusive to visualize the primary breakup process. These are based on physical jet interaction, thermal response, electrical response, and optical imaging techniques. No single diagnostic can completely characterize the entire spray structure. A combination of diagnostics is often applied to obtain detailed information about the spray characteristics. Newer developments in the field of optics, and electronics have vastly improved the traditional imaging and probing techniques yielding significantly better results.

The effectiveness of any visualization technique depends on several factors. The optical setup, illumination quality, light source, dynamic range and spatial resolution, light sensitivity, frame speed, and signal-to-noise ratio of the camera sensors – all play an important role. Care should be exercised to minimize the errors associated with non-uniform or unstable illumination, curvature effects, reflections, and shadows. When dealing simultaneously with structurally different entities like liquid core, liquid surface structures, and spatial droplet distribution, various challenges present themselves. They include the construction of three-dimensional atomization map from two-dimensional images, optical inaccessibility of dense jet breakup region, high-speed imaging without sacrificing high resolution, signal loss due to high noise, diffraction blurring of small droplets, overcoming multiple scattering, optical aberrations, and attenuation to name a few. The rest of this chapter will discuss some standard techniques for visualizing the primary breakup region.

## **2. Light sources and imaging devices**

All optical techniques are based on certain lighting schemes, such as direct lighting, diffused lighting, flood lighting, trans-illumination, reflective, etc. In the *transillumination scheme,* a light beam is passed through the spray or sheet and is imaged on

#### **Figure 2.**

*Some of the commonly used light sources for imaging purposes. A-gas arc lamp used as stroboscopic source, B - incandescent light source, C- fluorescent light source, D - Nd:YAG laser light source.*

the other side. In the *internally illuminated scheme,* a light source is located inside the spray and light is transmitted through the spray core to the surrounding. In a *reflective scheme,* the incident light reflects off the liquid surface, while in *fluorescence*, a thin light sheet illuminates a planar section of the spray. The light sources used for illumination could be either coherent or non-coherent type (**Figure 2**). Incoherent sources include strobe, incandescent, halogen, arc, and fluorescent lamps among others. Coherent sources include lasers and LEDs.

### **2.1 Strobe light**

Strobe lighting involves producing flashes of light for a short period of time at regular intervals. The typical flash lasts for 200 μs and may be synchronized with the framing rate of a suitable camera. With stroboscopic illumination, one can prolong the source lifetime, operate at increased light intensities, 'freeze' the motion of a fastmoving object, and time the pulses.

## **2.2 Incandescent light**

Incandescent lighting works by producing light by heating a wire filament. They have a wavelength of 300 nm – 1500 nm, flicker between 60 and 120 Hz, have an orange-yellow color cast, and are considered harsh for imaging purposes. They used to be widely used in the past for scientific imaging owing to their color being similar to natural sunlight. However low lamp efficacy (lumens per watt), luminaire efficiency, poor controllability of the light source, and incoherency presented challenges. Tungsten filament lamps, Halogen lamps, and Xenon arc lamps are some examples of incandescent light sources.

### **2.3 Fluorescent light source**

When electricity is passed through mercury vapor in a glass tube the radiation emitted interacts with the glass coating to produce white light. The illumination is high and this is known as fluorescent lighting. However, most of them cannot be dimmed and they may Fluorescent light commonly has wavelengths in the visible spectrum of 400–700 nm and frequencies of 10 kHz to 100 MHz. They have a negative reputation in photography due to the blue or green color cast that they produce. Flicker correction using electronic ballasts and radio interference removal using suitable filter circuits might need to be applied.

### **2.4 Lasers**

Lasers have come to be regarded widely as the most versatile source of light for imaging applications due to their many desirable properties like high brightness, stability, longevity, narrow spectral bandwidth, narrow beam divergence, high degree of spatial and temporal coherence, and well-defined polarization properties. A disadvantage of using laser beams is that they exhibit a Gaussian intensity profile, revealing particles in the middle region while particles at the edges get concealed. Some of the common types of lasers used in laboratories for research purposes have been discussed below.

#### *2.4.1 Solid state lasers*

These are one of the most commonly used laser systems, e.g. Nd:YAG Laser. They operate in the infrared 1064 nm wavelength regime, but pulses can be frequency doubled to generate wavelengths at 532 nm, or higher harmonics, 355 nm/266 nm. They can be operated in both the pulsed and the continuous mode and have an average power density of 5 <sup>10</sup><sup>3</sup> W/cm<sup>2</sup> /10s.

### *2.4.2 Gas lasers*

In gas lasers, an electric current is discharged through a gas or plasma to produce coherent light in the ultraviolet (excimer or nitrogen lasers) and visible range (He-Ne or ion-gas lasers). The excimer lasers typically operate at 193/248/308/351 nm with a pulse repetition rate of 100–200 Hz and a pulse duration of 10 ns. The He-Ne lasers typically operate at 632.8 nm. These lasers can produce beams, with a near-Gaussian/super-Gaussian profile. Molecular gas lasers emit in the infrared region popularly known as infrared (IR) lasers. These lasers can emit between 2 and 1000 μm range, at a rate of 300 GHz and10 THz. Since the optical depth of the jet breakup region in the infrared regime is smaller than in the visible spectrum, the infrared lasers can probe dense regions of the spray more effectively than their visible and ultraviolet counterparts [6].

#### *2.4.3 Ionized-metal vapor lasers*

These are an important tool for high-speed flow visualization, e.g. copper vapor lasers. Copper vapors are used as the lasing medium. The emitted pulses are in the green/yellow spectral range (510 nm/578 nm) with a pulse width in the range of 5–60 ns.

#### *2.4.4 Mode-locked lasers*

Mode locking or phase locking is a technique used for achieving ultrashort pulses, on the order of picoseconds or femtoseconds. Operating modes periodically interfere

constructively to produce an intense burst of light with peak power several orders of magnitude higher than the average power.

Lasers are also classified as continuous wave and pulsed lasers. In continuous wave lasers, the emitted light intensity is constant as a function of time. In pulsed lasers, the energy is released in the form of a short pulse of light.

## **2.5 LED**

LED illumination has recently gained popularity in a number of imaging applications. LED arrays produce intense, even illumination over a given object or area, have relatively low power requirements, generate very little heat, have a very long life and their pulse widths can be finely controlled. LEDs can produce coherent light under specific conditions such as in laser diodes and can be operated in continuous or pulsed mode. Their emissions spectra are broader than that of the lasers.

An integral component of any imaging system is the camera. Almost all of the imaging nowadays is carried out by digital cameras of either the CCD type (charged coupled device) or CMOS type (complementary metal–oxide–semiconductor). A comparison between the two sensor types is presented in **Table 1**. **Figure 3**(a) shows a simplified CCD sensor architecture and **Figure 3**(b) shows a simplified CMOS sensor architecture.

#### **2.6 CCD**

A CCD is an image sensor whose working principle is based on the photoelectric effect. It produces electrical charges proportional to the light intensities incident at different locations on the sensor which are then converted to digital values by "shifting" the signals one at a time between stages within the device. A photoactive capacitor array captures a two-dimensional picture of the scene. Each capacitor transfers its charges to its neighbor after which the charge is converted into voltage. The total number of image frames acquired is limited by the on-chip storage capacity. The transfer voltage signals get dampened at very high frequencies. Operating the CCD at very high frame rates leads to heating of the sensor, with accompanying thermal noise. The current practical


#### **Table 1.**

*Comparison of operating characteristics, of CCD and CMOS [7–9].*

#### **Figure 3.**

*Schematic of a typical (a) CCD sensor (left) and (b) CMOS sensor architecture (right) [7].*

frame rate limit is of the order of 1000 fps which could be extended to up to 100 Mfps (mega frames per second) having a spatial resolution of 312 260 pixels per frame [10, 11]. *Intensified CCD* (ICCD) is CCD coupled with an image intensifier for achieving high sensitivity in ultra-low-light-level conditions. They provide better temporal resolution and are suitable for capturing transient events.

## **2.7 CMOS**

A CMOS image sensor consists of an integrated circuit containing an array of pixel sensors, each pixel containing a photodetector and an active amplifier. Each pixel can be read individually. This enables fast clocking speeds (time taken to read the charge of the sensor) and high frame rates. The CMOS devices are highly immune to noise and have low static power consumption. Since the CMOS sensors have readout transistors at every pixel, most of the photons falling on the chip hit the transistors instead of the photodiode, lowering the light sensitivity of the CMOS chip. CMOS is more suited for faint/low light conditions and requires taking longer exposures. Some of the latest CMOS cameras can reach frame rates of 285,000 fps at reduced resolution<sup>1</sup> or the high definition<sup>2</sup> , 1080 HD resolution up to 2000 fps.

In order to effectively visualize the liquid breakup region under various illumination and density constraints, it is important to have a good understanding of the essential basics of *photography*. Some of these parameters e.g. field of view, frame rate, exposure, aperture, magnification, depth of field, depth of focus, and the dynamic range shown in **Figure 4**, are discussed below.

### **2.8 Field of view**

The field of view in the case of optical instruments is the solid angle through which the camera is sensitive to light. It defines the area that the camera is able to record, and is a function of the working distance, the focal length of the lens, and the sensor/film area.

<sup>1</sup> MotionBLITZ EoSens® mini2

<sup>2</sup> Photron Fastcam BC2

**Figure 4.** *Schematic of a typical camera and the different imaging parameters.*

### **2.9 Frame rate**

Frame rate is the number of consecutive images recorded per unit time by an imaging device. It is mostly expressed in frames per second (FPS) or hertz (Hz). For a fast event like atomization, a low frame rate may result in jerky and less fluidic motion while a very high frame rate would require high processing power and storage to yield sharp high-quality images of the phenomena in motion.

## **2.10 Exposure**

It is the amount of light collected by the sensor of the camera during single image capture. Too long exposures lead to excessive light collecting on the film, which results in a washout appearance. Too short exposures lead to insufficient light collection resulting in a dark image. Exposure depends on the frame rate and camera aperture. The exposure time in CCD and CMOS cameras is set by an electronic shutter that is controlled either manually, electronically, or by software.

### **2.11 Aperture**

Aperture is the size of the lens opening which limits the amount of light entering the camera and falling on the image plane. It controls the depth-of-field in the sense that the background can be blurred with a wide aperture keeping just the object in focus or alternatively keeping everything in focus by using a narrow aperture. A wide aperture results in a higher degree of optical aberrations (distortions) and vignetting (falling intensity toward the edges of the picture).

### **2.12 Magnification**

Magnification is the degree of scaling (enlarging or diminishing) of a subject on the image plane. There are two ways to represent magnification, (a) Linear or transverse magnification - it is represented by Y/X where Y is the image length and X the subject length and (b) Angular magnification - angle subtended by the object at the focal point.

### **2.13 Depth of field**

It is the range of distance along the optical axis between the nearest and farthest objects in a scene that appear to be in focus (sharp) in the photograph.

#### **2.14 Depth of focus**

This is the limit of the image plane displacement at which the image will appear sharp. Depth of focus refers to image space, while the depth of field refers to object space.

#### **2.15 Dynamic range**

It is the ratio of the saturation level of a pixel to its signal threshold or simply put the ratio between the maximum and minimum measurable light intensities. Since the dynamic range of sensors is far less than the human eye, local tone mapping and dynamic range adjustment are used.

The *spatial resolution* of an imaging system is its ability to distinguish separate objects within its field of view. For the same sensor size, increasing inthe field of view decreases the image resolution.

Illumination setup plays a vital role in determining the clarity of an image. Various lighting techniques exist [5]. The most commonly used method is frontal lighting in which the camera and illuminating source are placed in front of the jet and the light reflected from the breakup entities is recorded. *Frontal lightening* gives a threedimensional appearance that helps in visualizing the liquid surface features. This arrangement fails to yield any inner details of the jet breakup. Illumination of a section of the liquid by a light sheet is followed mostly in cases of axisymmetric flows. Weak elastic scattering by the droplets at the breakup location helps to reveal the inner details of the breakup location in the plane of illumination. However in this case, multiple scattering is often a major drawback. *Backlighting*, is a technique in which the light source is placed behind the object, and translucent glass is located between the object and the light source. The translucent screen diffuses the flash uniformly over a wider region illuminating the entire breakup location section. This diffuse light can produce diffuse reflections from the object being backlit resulting in soft and blurry edges. *Bidirectional lighting* is when two light sources are placed at an angle of 120° from the camera's line of sight, illuminating the breakup location.

## **3. Diagnostics**

Different optical techniques are used to characterize the primary breakup process; e.g. photography, shadowgraphy, holography, ballistic imaging, jet core illumination, laser-induced fluorescence, thermal imaging, Mie imaging, and X-ray phase contrast imaging. **Table 2** summarizes the application situations of the different diagnostics and the type of information they provide.

#### **3.1 Photography**

Photography in its simple form consists of an illumination source illuminating the object and a camera recording the images of the object. Photography of a jet breakup is *Imaging Diagnostics for Jet Breakup into Droplets: A Review DOI: http://dx.doi.org/10.5772/intechopen.107370*


**Table 2.**

*A comparison of the different diagnostics available for visualizing primary breakup.*

used for getting measurements on breakup lengths, drop sizes, droplet distribution density, jet and droplet velocities, fluid flow behavior, etc. In some cases, a diffuse screen may help scatter the light incident on the object, and digital display help in visualizing the image in real-time as shown in **Figure 5**. Imaging could be in the single-shot, low speed, and high-speed modes. Illumination could be provided by any of the schemes discussed previously, e.g. strobe synchronized with the camera frame rate in the forward light scattering configuration [12–14]. Incandescent or pulsed laser sources could also be used.

**Figure 6** shows a high-speed photograph of a jet surface undergoing primary breakup in still air [15] where the jet was back-lighted at 45° from the horizontal and shielded from ambient light (**Figure 7**).

**Figure 5.** *Optical setup of the photography technique.*

**Figure 6.**

*Photographic image of jet surface primary breakup in still air [15].*

**Figure 7.** *Laser backlit illumination [16].*

#### *Imaging Diagnostics for Jet Breakup into Droplets: A Review DOI: http://dx.doi.org/10.5772/intechopen.107370*

With increasingly dense sprays, the intensity of the back-lit illumination should increase to distinctly reveal the droplets in the breakup location. Magnification is pivotal in distinguishing and measurement of atomization entities. Droplets, ligaments, and surface features in the size range of micrometers would require significant magnification. Higher magnification comes at the expense of the depth of field and this presents a challenge toward achieving a focused image of the small non-stationary liquid surface. Optical systems have been designed to make it flexible such as by using a phase mask [16] consisting of a combination of Fresnel lenses (FL) (**Figure 8**) where each FL works in tandem with the primary lens to produce a sharp image for a unique object plane. A micro-actuator can be used to translate the detector along the optical axis during image integration [18]. Image processing algorithms [19] applying the focus stacking method can modify the phase of incoherent light wavefront to produce a point-spread-function (PSF) over a large region of focus yielding an extended depth-of-field. For visualization of the breakup phenomena over a wide range of distances, a large field of view with a high depth of focus might be desirable when the magnification is low.

Spatial resolution is often denoted by lines per inch (lpi) or μm. It represents how closely two lines can be resolved in an image. Film-based photography has a spatial resolution of 100 μm, while digital photographs could achieve resolutions up to 10 μm. Digital image resolution is limited by pixel noise and pixel cross-talk. The spatial resolution also depends on the conical angle subtended by the object at the lens aperture. Lens resolution for narrow lens aperture is affected by diffraction, and for large apertures affected by optical aberration. For imaging small structures at widely separated points with a high resolution, a multiple-segment long-distance microscope in combination with micro-lens and aperture array has been demonstrated [17] (**Figure 8**).

The finite size of a camera lens with a circular aperture leads to the diffraction of parallel light rays passing through them to form a diffraction pattern in the image. It usually has an appearance of a central bright spot and surrounding bright rings, separated by dark nulls. This two-dimensional far-field diffraction pattern is called '*Airy disc*', (**Figure 9**.). Its angular radius measured from the center to the first null is equal to sin<sup>1</sup> (1.22λ/*D*), where λ is the wavelength of light passing through and *D* is the aperture diameter. The diameter of the first dark circle (width of the disk) defines the theoretical maximum resolution for an optical system such that if two objects imaged by the lens start to have overlapping airy discs by the virtue of their closeness, a blurring effect occurs.

**Figure 8.** *Optical setup of Bauer [17].*

#### **Figure 10.** *Tsai calibration grid.*

*Camera calibration* is done both geometrically and photometrically. Geometric calibration involves using a Tsai grid (two planes at the right angle with checkerboard patterns as in **Figure 10**) by which the scaling factors between the image and the actual target dimensions are achieved. In photometric calibration relation between digital counts and luminance is sought by capturing test patches with known relative luminance. Photography also has some drawbacks. Mainly its inability to resolve a threedimensional perspective of the object accurately, inability to see through optically dense droplets or liquid structures, and limitations with respect to depth of focus/magnification. For cameras, high spatial resolution comes at the expense of reduced pixel size and therefore a reduced light-sensitive area and reduced signal-to-noise ratios. The absence of cameras and light sources with very high repetition rates, high enough to capture fast motion continuity is another limitation in photography. **Figure 11** shows a setup [20]

**Figure 11.** *High magnification video setup [20].*

for obtaining high-magnification video images of a jet breakup. In this setup, a highspeed CCD camera fitted with a long-distance microscope lens faces the spray which is itself illuminated with a flat-faced halogen lamp from the rear. A computer records and stores the high-speed movie at the rate of 2000 frames/sec. The field of view, in this case, was 13.9 mm x 8.9 mm, with an image resolution of 29.6 pixels/mm.

#### **3.2 Shadow imaging**

Shadowgraphy is a technique of imaging the shadow of the refractive index field. When light passes through a region of varying refractive index (**Figure 12**), it experiences retardation proportional to the material's density and bends toward the region of higher refractive index. The angular deflection and displacement of the rays are small. The local beam intensity is not significantly affected, but the angular deviation is enough to produce a focusing effect above the higher refractive index regions as the beam propagates beyond the fluid layer. This coupled with the retardation causes the wavefronts to turn, due to which rays converge and diverge into bright and dark regions.

The refractive deflection of rays causes a shadow effect (spatial modulation of the light-intensity distribution) in the recording plane which is then imaged. The image intensity thus depends on the variations in the optical density of the object media. A portion of incident light refracts at the fluid-interface boundary and may produce a darker boundary region. The portion that does not interact with the object produces a homogenous background. The dark regions (shadow) in a shadowgram mark the boundaries of the object. Light transmittance is affected by scattering and absorption too.

**Figure 13** shows a shadowgraph image [21] of quenching oil jet spray from a 1.2 mm atomizer nozzle at 15 m/s and 38°C. Shadowgraphy has been used to visualize liquid jet breakup [23, 24] at different length scales in-order to determine ligamentdroplet sizes, and speeds. Shadowgraphs can be subjected to further image analysis for detecting the liquid contour and its individual features based on the comparison of the RGB (Red Green Blue) intensity levels to a pre-established threshold level or identification of local changes of RGB intensity, which are their greatest at the fluid boundaries. Shadowgraphy is different from backlit photography despite a light source being located behind the object in both cases. For backlit photography, the background lighting is diffusivein nature. In addition, the liquid frontal surface is illuminated and the reflected light is directly photographed with a camera. The shadow image formed

**Figure 12.** *Formation of shadow image by relative deflection of rays.*

**Figure 13.** *Shadowgrams of (a) quenching oil jet spray [21] and (b) water jet spray [22].*

on a screen is imaged by the camera. For optically dense objects the intensity variation inside the image is insignificant compared to the surrounding bright background intensity. In such cases, a dark shadow image with a bright background may result and therefore any kind of density variations analysis would be irrelevant. Shadowgraphs provide phase information (refractive index gradients) [21, 25] or simply a monochromatic projected shadow image of the object [22, 26–28]. It can help in visualizing convective flows [29].

**Figure 14** shows a typical laboratory setup using incident parallel laser light forming a focused shadowgraph image onto a screen. It is known as "focused shadowgraphy". Light from the source is passed through a spatial filter, then collimated by using a collimating lens, and allowed to pass through the object. After passage, the beam is focused using a relay lens which forms a real inverted image onto a translucent or a ground glass screen or a photographic film. The camera is focused on this image from the other side of the translucent screen. Alternatively, the primary shadowgraph formed by the relay lens can be recorded directly with a camera simply by focusing the camera lens on the plane of the primary shadowgram. This allows variable magnification to be achieved as well. The relay lens somewhat limits the field of view of focused shadowgraph but sizes to the order of micro scales can be obtained using the appropriate magnification. The use of photographic films yields better resolution [26, 30, 31].

**Figure 14.** *'Focused shadowgraphy' setup (shadowgraph is formed on glass plate or film).*

#### **Figure 15.**

*Shadowgraph's inability to image entities hidden or obscured by the projection of the larger structures.*

Shadowgraphy is easy and inexpensive to carry out, has a high spatial resolution for most practical purposes, is independent of the shape and material of the phase media and its relatively large field of view is well suited for size, shape, position, and velocity determination.

A disadvantage of shadowgraphy is that since it gives the projected image of the object, any overlapping ligament or droplet information is lost if its projection does not fall along the jet periphery or is obscured by a larger structure in its line of sight (see **Figure 15**). It fails to give details of the inner fluid structure of the primary breakup region high-density region. It is also unable to distinguish between fuel vapor and small droplets. Image sensitivity, blur, and feature size are always competing issues in shadowgraphy which makes precise measurements difficult. The formation of caustics in shadowgrams is a particular disadvantage when the caustics become confused with some other phenomenon. For instance, a boundary layer can act as a cylindrical lens, focusing light into a bright line.

Basic shadowgraphy has been adapted to other imaging techniques such as Microscopic shadowgraphy, Stereoscopic shadowgraphy, and Holographic shadowgraphy [32].

Spark-shadowgraph technique [33, 34], High-speed shadowgraphy (HSS) [35, 36], and Specialized Imaging Shadowgraph (SIS) are some of the other related developments.

The tomographic Shadowgraphy technique [37] is based on a multiple-view camera setup and is capable of resolving the liquid jet core both spatially and temporally (**Figure 16**). The spray shadowgraphs are obtained using a pulsed LED and four synchronized double frame cameras angled 30° to each other. The spray is then reconstructed by using a line-of-sight reconstruction technique. Stereoscopic shadowgraphy [38] uses four parabolic mirrors to form two inclined intersecting beams and the object is placed at the intersection. Two cameras synchronized in the masterslave configuration record the shadowgraph pairs automatically. Upon image reconstruction, three-dimensionality is achieved.

#### **3.3 Holography**

Holograms are made by the interference between a wave field scattered from the object and a coherent background, called the reference wave [39–41]. They can focus

**Figure 16.** *Tomography shadowgraphy setup [37].*

**Figure 17.** *The optical system of parallel phase-shifting digital holography [43].*

simultaneously in three-dimensional spaces [42]. A sketch of a digital in-line holography setup is shown in **Figure 17**. The optical setup consists of an incident light source (laser). The incident beam is collimated using a beam expander (objective lens + spatial filter) and convex lens combination followed by splitting it into the object and the reference beams using a beam splitter. The object beam is passed through the object where it suffers diffraction and a phase change. Both the reference and the object beams are adjusted for intensity before being combined via a beam splitter. The resulting interference pattern is then recorded on a CCD sensor at the hologram plane which is at a finite distance from the true image plane. Magnification can be introduced by using a convex lens as a relay lens after the object beam has passed through, the object, and this magnified hologram would then be recorded on the CCD sensor. The recorded hologram is then reconstructed using a computer program based on the convolution type approach which solves the Rayleigh Sommerfeld formula for the reconstruction of a wave field [40, 44] using the Fast Fourier Transform algorithm (or the Fresnel-Kirchhoff for high numerical aperture or any of the other Fraunhofer transform/wavelet transform schemes).

Holography eases the limitation of the depth of field and provides a threedimensional image made up of two-dimensional image planes focused at different distances along the axis perpendicular to the image plane [45]. Unaffected by the nonspherical droplets and ligaments that are usually encountered very close to the injector exit, it also has a virtually unlimited depth of focus, and high magnification even for a *Imaging Diagnostics for Jet Breakup into Droplets: A Review DOI: http://dx.doi.org/10.5772/intechopen.107370*

**Figure 18.** *Hologram reconstruction coordinate system.*

small field of view. Depending on the angle between the reference wave and the object wave, a zero reference angle setup is called inline holography and a non-zero reference angle setup is called an off-axis holography. They can be recorded using a photographic film [46], or digitally using a CCD [47–49]. Though film-based holography provides a high spatial resolution the time involved in carrying out the wet processing of film makes it disadvantageous to be employed for measuring dynamic phenomena.

Digital holography enables quick recording and retrieval of hologram information in real time. In **Figure 18**, ρ is the distance between a point in the hologram plane and a point in the reconstruction plane, *ξ*, *η* are coordinates in the reconstruction image plane, and *x*, *y* are coordinates in the recording hologram plane. The hologram is reconstructed to different spanwise distances so as to focus on the particles centered at those distances from the camera in the jet breakup field. Different modifications to the reconstruction process have been adopted by various researchers to serve specific purposes. For example viewing the object at different viewing angles and with different focal lengths within the Fresnel domain [50], reconstructing the image of a diffusively reflecting object using only the phase data of the complex amplitude [51], for digitally compensating the aberrations that arise during normal holographic reconstruction process [40] (by expanding the recorded wavefront to increase the spatial resolution), and to overcome the disruption in the visibility of the nearby focused objects due to the diffraction effects caused by the out of focus objects [52]. The relay lens used for increasing magnification works well at low magnification but causes aberration and noise at higher magnification.

In such a situation digital holographic microscopy [53] (DHM) is commonly adopted (**Figure 19**). This setup can successfully image liquid breakup elements as small as 5 μm. Only one beam is expanded with an objective lens and this serves both as a reference and an object beam. It is then passed directly through the liquid breakup region onto a CCD. The incident beam, in this case, is a coherent spherical wave emanating from a point source. This is achieved by passing the incident beam through a microscope objective followed by a pinhole. The pinhole size is of the order of the laser wavelength. A sample hologram of the liquid jet surface and its reconstructed image is shown in **Figure 20**. The resolution in DHM depends on (1) pin-hole size this controls the spatial coherence and illumination cone; (2) numerical aperture - this is a function of the size and position of the CCD sensor; (3) pixel density and dynamic range - this controls fringe resolution and hologram noise level; and (4) wavelength of the laser light. Even though the depth of focus is small in this case owing to the large numerical aperture the digital reconstruction would allow one to focus at different depths of the object. Various modifications of DHM exist. In Digital Image Plane Holography (DIPH) [54], a plane of the fluid is illuminated with a laser sheet.

**Figure 19.** *Digital holographic microscopy setup.*

**Figure 20.**

```
Reconstructed holograms of a jet surface at different reconstruction distances.
```
The image of the illuminated fluid plane is then made to interfere with a reference beam on the CCD sensor which is imaged (**Figure 21**). The depth of focus is limited to the laser sheet thickness. Numerically reconstructing the holograms requires enormous processing power and memory. With Fast Fourier transform algorithms, the reconstructing time for a single hologram has been brought down to within a few tenths of a second on personal workstations. Another issue is adopting holography as a 3D particle characterization tool is its long depth of field [55]. Angular aperture in holography is limited to only a few degrees and this results in poor depth resolution. The finite pixel size of the image sensors limits the angle between the object and reference wave leading to the virtual and real images not being fully separated during hologram reconstruction. The field of view is also limited to the sensor size (or less if using a divergent beam).

The limited resolution capability of digital image sensors leads to the low spatial resolution of the hologram image. Some deconvolution methods based on the Weiner filter and iterative routines with point spread functions are used for overcoming the depth of focus limitation [56]. Other drawbacks include the reconstruction of spurious twin images of particles, leading to ghost images and multiple focusing around the actual depth location of the particles [57].

*Imaging Diagnostics for Jet Breakup into Droplets: A Review DOI: http://dx.doi.org/10.5772/intechopen.107370*

**Figure 21.** *Digital image plane holography (DIPH) setup [54].*

#### **3.4 X-ray phase contrast imaging**

X-ray-based diagnostics have recently found increasing applicability in the study of liquid breakup due to some of its inherent advantages. The high penetrability of dense breakup regions and the ability to probe the internal mass distribution of breakup regions with a good time and spatial resolution makes it appealing. The x-ray beam experiences a significantly lower wave-shift as compared to visible light because of the large value of the index of refraction at optical wavelengths. This means the xrays can be absorbed in specific wavelengths but they do not undergo appreciable scattering as in the case of visible light. Phase contrast imaging is based on utilizing the coherence properties of x-ray to enhance the edges of near-field Fresnel diffraction. Upon interaction with liquid jets, the x-rays suffer absorption along with a corresponding phase shift due to changes in phase media. Recording this phase shift on a CCD sensor enables us to distinguish between phase boundaries owing to improved image contrast and is termed as phase contrast imaging [58].

In a typical set-up, a synchrotron x-ray source is used whose undulator provides an extremely brilliant white monochromatic x-ray beam required for ultra-fast imaging purpose (**Figure 22**). The x-ray beam directly illuminates the breakup region but is tuned to the photon energy that does not get absorbed. When the x-rays pass through the sample they suffer attenuation and phase shift [59–62]. This creates small variations in the speed and direction of propagation of the x-rays. Interference effects are produced at the feature boundaries which have significant contrast. After interaction with the sample, the different wavefronts of the x-ray beam get diffracted by the sample and start to overlap and interfere with each other, giving rise to a contrasting pattern. This contrast depends on the Laplacian of the phase shift undergone by the beam upon its passage through the sample [63]. A fast scintillator crystal (LYSO:Ce or YAG:Ce) converts the x-ray phase contrast pattern into visible light which is then imaged into a CCD. Around the edges of features (ligaments and drops), an evolving pattern of light and dark fringes develop. This makes this technique particularly

**Figure 22.**

*X-ray phase contrast imaging experimental setup.*

#### **Figure 23.**

*(a) Phase contrast image of jet surface structures, and simulated images of (a) air bubble in water (c) water bubble in air respectively [59].*

sensitive to boundaries and interfaces of different phase media [64]. Therefore a bubble (gas surrounded by liquid) will have a dark/bright outer edge while a droplet (liquid surrounded by gas) will have a bright/dark outer edge as illustrated in the simulations shown in **Figure 23**(b and c). The CCD camera is coupled to the scintillator using a microscope objective and a 45° mirror and records the images at a high speed. The exposure time (150 ps) for each image is achieved by shuttering and timing the pulsed x-ray beam. In order to reduce the heat power in the beam, shuttering is usually employed whose synchronized operation could cut off more than 99% of the beam heat power.

The phase contrast images need to be normalized to exhibit the contrast/intensity gradients more efficiently. This technique is particularly suitable for studying dense primary breakup near-injector regions. The image resolution is generally a function of the detector resolution and the phase sensitivity, is a function of the sourcespecimen/specimen-detector distance. Phase contrast imaging has been applied to the study of spray breakup [58, 65] and as a simulation tool [66].

Some disadvantages of this technique are that the x-rays are difficult to manipulate and oftentimes the experiment needs to be designed in a way so as to be remotely controllable owing to safety issues. The breakup surface features e.g. ligaments and

drops located at different axial distances are indistinguishable in the phase contrast xray images.

## **3.5 Ballistic imaging**

Ballistic imaging [67–75] is a line-of-sight two-dimensional imaging technique that can overcome the effect of multiple scattering in dense liquid breakups. It can provide good quality images of the dense liquid breakup region which would otherwise be opaque to simpler imaging techniques. It derives its name from the term 'ballistic photons', the group of photons that do not suffer scattering when transmitted through highly turbid media because they travel the shortest path. This technique is based on a photon time of flight selection in which the transmitted photons are temporally selected, filtering out the multiply scattered photons which are time delayed depending on the optical length of the probed medium.

When light passes through a highly turbid medium, scattering occurs and the emerging beam comprises ballistic, snake, and diffuse photons (**Figure 24**).

### *3.5.1 Ballistic photons*

These photons pass straight through the medium without scattering. They exit within the same solid angle that they entered and so travel the shortest path, and exit first.

## *3.5.2 Snake photons*

These photons undergo negligible scattering, i.e. from a single scattering event to up to four scattering events and exit the medium along the same incident axis but with a larger solid angle than the ballistic photons.

## *3.5.3 Diffuse photons*

These photons exit the medium after scattering multiple times (5 times or more). It has a large photon number density, is scattered into 4π steradian and exits last. They do not retain the memory of the structure within the material.

A typical Ballistic imaging setup [68], is shown in **Figure 25**. Ultra short light pulses (100 fs), isolated from all of the other background and noise photons (by using polarizer) are used as the imaging light to illuminate the break-up region. A timeresolved image is recorded using an ultra-fast time gate (full optical system such as the

**Figure 24.** *Scattering through a dense media showing ballistic, snake, and diffuse photons [70].*

**Figure 25.** *Schematic of ballistic imaging setup.*

optical Kerr gate), short enough to select photons with respect to their time of flight by separating them into ballistic, refractive, and scattered light which is then followed by constructing the image using just the selected class of photons.

The optical Kerr gate could be obtained by inducing a temporary birefringence in a Kerr active liquid between two crossed polarizers. The ballistic photons are attenuated by the large liquid structures inside a dense breakup region via absorption or refraction thus creating the intensity modulation required to produce an image.

These are used to form an undistorted, diffraction-limited image of structures inside the optically dense medium. The snake photons are also used along with the ballistic photons to improve the signal-to-noise ratio. Since highly turbid media leads to significantly more diffuse photons, they must be separated from the ballistic and snake photons. In **Figure 25**, a light source (1-kHz Ti-Sapphire regenerative amplifier, seeded with a Ti-Sapphire mode-locked laser oscillator) generates 150 femtosecond pulses centered around 800 nm). The light exiting the amplifier is linearly polarized and is split into an optical Kerr effect (OKE) switching beam and an imaging beam using a beam splitter. The OKE is a time gate consisting of a very fast shutter that selects only the leading edge of the image pulse containing ballistic and snake photons. To introduce the beams into the breakup region the linearly polarized imaging beam is made to pass through a polarizer followed by a rotation of the beam polarization by 45°. The imaging beam is then passed through a telescope which controls the beam size as it crosses the breakup region. The relay optics then focuses the beam through the OKE switch which serves as a shutter (2 ps) and is triggered by the switching pulses. When the switching pulse is absent image transfer to the display screen does not occur. The first polarizer in the OKE gate which is also the second polarizer used in the imaging beam, allows the polarization orientation of the imaging beam to pass through. The imaging beam is then focused into the Kerr active medium and upcollimated again. The second OKE polarizer is oriented normally with respected to the first one to block the unperturbed imaging beam. When the intense electric field of the switching pulse arrives at the Kerr active medium it rotates the polarization of the imaging beam, allowing most of it to pass through the second polarizer. After the imaging beam passes through the OKE gate, the image is relayed through a short-pass filter and displayed on a screen which is then imaged using a camera. By adjusting the length of the time delay segment of the imaging beam, it would be possible to

*Imaging Diagnostics for Jet Breakup into Droplets: A Review DOI: http://dx.doi.org/10.5772/intechopen.107370*

**Figure 26.** *Ballistic image of a jet breakup [2].*

introduce a delay that would control the temporal overlap between the switching and imaging pulses when they arrive at the OKE gate for optimum time-gating.

The ballistic image thus obtained (**Figure 26**) reveals the main jet liquid column, droplet distribution, voids, and jet profile structures. The dark areas shown represent the liquid phase, the light areas are the gas phase, and the speckles and other spurious features are caused by diffraction.

If the droplets in the breakup region are smaller than the limit of spatial resolution for the setup (40–50 μm) they go undetected which is a disadvantage. The spatial resolution is limited by the CS2 liquid contained in the CS2 cell located in the Fourier plane which acts as a spatial filter to reduce the scattering noise.

#### **3.6 Liquid jet Core illumination**

Optical connectivity or liquid jet core illumination is another relatively new technique for studying atomization. Here, a light guide illuminates the liquid jet from within the injecting nozzle in a direction parallel to the flow [76, 77].

Here the liquid jet acts as an optical fiber through which the light propagates and is interrupted only at the breakup region (**Figure 27**). The continuous portion of the liquid jet is made to fluoresce by adding a fluorescing dye (Rhodamine WT dye). The laser beam is steered by means of a light-guide tube or an optical fiber through the liquid jet nozzle. The fluorescing dye in the liquid emits fluorescent light along the laser beam path and is interrupted at the breakup region. The intensity of the beam should be sufficient to identify this position. The advantage of introducing the laser beam from within the liquid nozzle is that it is minimally attenuated and multiple scattered and maintains a low light intensity beyond the surface of the liquid jet. The addition of the dye aids in the emission of longer wavelengths of fluorescing light which can be detected separately from the incoming laser beam thus avoiding background noise on images due to the scattered light. The fluorescing light is emitted from everywhere inside the liquid core, associated with the volume of the liquid. The liquid jet surface features act as minute mirrors to focus the laser beam which produces bright spots along the liquid surface and could be identified as locations for significant changes in the surface structure. Short-duration laser pulses allow the

**Figure 27.** *(a) Liquid jet core illumination [76] (b) principle of optical connectivity [2].*

image to be frozen in time. The imaging system is a CCD camera fitted with a low-pass filter to suppress the scattered light from the laser beam.

This diagnostic promises improved measurements for atomization occurring at the nozzle exit since scattering (caused by the diffusion of the laser light by the droplets) persists further downstream of the breakup region. Light propagation in narrow ligaments connecting the detaching masses of liquid and the surrounding droplets causes the light to become diffuse and its intensity to be significantly reduced after the core breakup region. The detached droplets do not fluoresce and the area surrounding the breakup point does not illuminate.

As the jet becomes increasingly turbulent, the transition of the smooth jet surface to a rough and wavy one results in the angle of incidence between the laser light rays and the jet surface becoming less than the critical angle for total internal reflection. This leads to increased scattering, loss of fluorescence intensity, and bright spots created due to the focusing of laser light by jet surface features which act as minute mirrors. In certain situations like at high Reynolds numbers, the liquid core can suddenly change thickness drastically, become very thin, and then thick again. The fluid constriction thus created can cause significant loss of laser light. Moreover, it does not work for a situation requiring droplet size measurements or location determination, or their velocity measurements as is often possible with other techniques.

#### **3.7 Mie scattering**

Mie scattering is an elastic scattering that occurs when the size of the particles in the liquid breakup is of the order of the scattered wavelengths. Mie-scattering intensity is proportional to the total liquid surface area. A number of researchers have used Mie imaging for investigating fuel sprays [78–84]. An instantaneous Mie scattering image of a jet in crossflow [78] is shown in **Figure 28**, where the inset represents the original Mie image which is then inverted to yield a dark image of the jet against a white background.

A Mie scattering setup for two-dimensional spray visualization [82–86] is shown in **Figure 29**. A wide planar laser sheet is formed using a cylindrical lens, made to pass

*Imaging Diagnostics for Jet Breakup into Droplets: A Review DOI: http://dx.doi.org/10.5772/intechopen.107370*

**Figure 28.** *Instantaneous Mie scattering image [78].*

**Figure 29.** *Mie scattering setup for two-dimensional spray visualization.*

through a particular plane of the liquid jet. The scattered signal normal to the illuminated plane is imaged onto an ICCD camera fitted with a narrowband interference filter. The images are averaged for noise removal followed by background subtraction, uniform field correction, and intensity normalization. The clarity of the Mie image depends on the extent of multiple scattering. When the droplet concentration is low and the breakup dimension is small, the detected signal does not suffer attenuation. However, when droplet concentration is high the incident beam becomes attenuated as it travels the breakup region, 'secondary scattering' from droplets lying between the incident beam and the detector and 'multiply scattering' by the droplets in the surrounding atomized region result in "extraneous light" to be detected. Since the Mie intensity is proportional to the square of the particle diameter ( *d*<sup>p</sup> 2 ) a CCD sensor having a large dynamic range is required to simultaneously image the large liquid core and the smallest droplets. Because the scattered cross section is always larger than the corresponding ligament/droplet diameter, it is difficult to carry out accurate size measurements and distributions from the Mie signals. Spatial resolution is also a significant limitation of Mie imaging.

Comparisons are often made between shadowgraphy and Mie imaging. However, while Mie imaging is the visualization of elastically scattered light normal to the plane of the incident light sheet passing through the breakup region, shadowgraphy is the image of the shadow of the breakup region formed by a collimated beam of light.

## **3.8 Planar laser-induced fluorescence**

Planar laser-induced fluorescence (PLIF) is a technique [87–93] where a plane laser light sheet is incident on a cross-section of the fluid mixed with a fluorescent dye. The dye causes the liquid plane to fluoresce and this fluoresced pattern is then imaged by a camera. A part of the laser excitation lies within the absorption spectrum of the dye. **Figure 30** shows a typical PLIF setup.

In the fluorescence linear regime, the fluorescence intensity is proportional to the laser excitation intensity. The dye has a large separation between the absorption and emission spectra (e.g. Rhodamine WT/6G/B, and Fluorescein). A narrow-band optical filter allows only the fluorescence wavelengths to be imaged by the camera. **Figure 31** shows a PLIF image of a jet primary breakup. The characteristic decay time (time necessary for 86.5% of fluorescence intensity to be emitted) and fluorescence intensity require calibration [20, 89]. The PLIF technique suffers from a non-uniform

**Figure 30.** *PLIF setup.*

**Figure 31.** *PLIF image showing the liquid jet sheet breakup [88].*

*Imaging Diagnostics for Jet Breakup into Droplets: A Review DOI: http://dx.doi.org/10.5772/intechopen.107370*

**Figure 32.** *(a) SLIPI beam setup [2] (b) SLIPI Mie scattering planar image of hollow cone spray [2].*

intensity profile of the planar laser sheet, and light intensity reduction due to absorption by the background dye concentration. High dye concentrations produce nonlinearities in fluorescence because of absorption-related changes in the excitation intensity. The camera exposure time depends on the spectral sensitivity of the camera's sensor at the fluorescing wavelength.

Light emitted from the object plane in PLIF could be either Mie scattered radiation or fluorescence light. To reduce the scattering, the Structured Laser Illumination Planar Imaging (SLIPI) technique is carried out (**Figure 32**a). SLIPI [94–96] generates a recognizable signature in the incident planar beam (by time-based modulation of the excitation source) that yields signal photons. The multiply scattered photons are discarded due to their lack of this signature. **Figure 32**(b) shows a SLIPI Mie scattering planar image in a hollow cone spray.

#### **3.9 Thermal imaging**

Thermal imaging is used for temperature visualization in the breakup regime. It uses planar laser-induced emission (fluorescence/phosphorescence) of a suitable excitation wavelength in the UV region and thermographic phosphors or organic tracer molecules. The two primary methods under thermal imaging are thermographic fluorescence and thermographic phosphorescence.

#### *3.9.1 Thermographic fluorescence*

In this technique [7, 97–100], a fluorescent additive is added to the hydrocarbon liquid in the form of a monomer (M), and an exciplex forming an organic molecule called quencher (N0). When this sample is irradiated, M is electronically excited to form M\* which then reacts with N0, which is in the ground state to form the organic exciplex (E\*) which is an excited state complex. The reaction is represented by

$$\mathbf{M}^\* + \mathbf{N}\_0 \boldsymbol{\Theta} \boldsymbol{\Theta} \mathbf{E}^\* \tag{1}$$

Fluorescence is observed from both M\* and E\*. Typical fluorescence lifetimes are less than 100 nanoseconds and the populations of M\* and E\* are strongly temperature dependent. By comparing the decay lifetimes or the intensity ratio between two or several emission lines to a calibrated standard, the liquid breakup droplet temperature distribution can be obtained. A typical experimental setup resembles the PLIF setup [97]. An Nd:YAG laser source produces a light sheet which is incident on a liquid plane to excite the monomer/quencher pairs leading to exciplex formation and fluorescence from the breakup region. This is photographed at right angles to the plane of the laser light pulse. The liquid and vapor phase images are imaged with a single ICCD camera using UV objective and a stereoscope. Different emission lines correspond to different temperature ranges. The spatial profile of the breakup, droplets, and vapor are distinguishable and measurable. For a multi-component fuel mixture, a two-color laserinduced fluorescence technique can be adopted [101, 102] in which the additive (pyrromethene) exhibits a highly different temperature-sensitive fluorescence spectrum for two spectral bands of detection. Rhodamine B is often used as a fluorescent temperature sensor as it exhibits significant temperature dependence. This technique can measure the thermal transport properties (mean and fluctuating characteristics of the dynamic and thermal fields) of a turbulent heated jet undergoing atomization.

#### *3.9.2 Thermographic Phosphorescence*

Thermographic phosphors (e.g. La2O2S:Eu or Mg4FGeO6:Mn) are temperaturesensitive materials that when excited by UV light exhibit temperature-dependent emission characteristics. Thermographic phosphors are mixed with the liquid forming the jet. A thin laser sheet is then directed at the liquid plane to be imaged. This causes phosphor excitation and subsequent emission from the breakup region [103, 104]. After excitation, the lifetime of a particular emission line is monitored to obtain temperature measurements.

The emitted light is passed through an interference filter, and recorded by a fastframing ICCD camera. After image acquisition, the pixel positions of the ICCD detectors are correlated with one another. Using an exponential fitting procedure, the lifetime and temperature information is then extracted. A two-dimensional temperature image is thus obtained, as shown in **Figure 33** It shows a water spray temperature distribution image. Since the temperature of the droplets of different sizes after jet breakup changes due to evaporation and convection in comparison to the main jet core, their detection becomes possible using this method.

This technique has high accuracy; the advantage of remote detection and a high signal yield; and is independent of pressure variations. The excitation energies for both fluorescence and phosphorescence should be below a threshold value as high excitation energies lead to luminescence saturation. Fluorescence lifetime is on the order of nanoseconds whereas phosphorescence lifetime ranges between the orders of microseconds to milliseconds. The decay time is often a very sensitive function of

**Figure 33.** *Single shot water spray image using thermographic phosphors [103].*

**Figure 34.** *Experimental setup of thermal imaging by the spectral method.*

temperature [105]. So any phenomenon is limited to being observed if only it occurs within the excited state period of the phosphor and the speed is essentially dependent on the camera frame rate (MHz). Combining both the fluorescence and phosphorescence techniques has shown to be successful [106, 107] in understanding the breakup and mixing of liquid fuels at high pressure and temperature conditions similar to those in combustion chambers. **Figure 34** shows the experimental setup of thermal imaging by the spectral method.

## **4. Conclusion**

The need to develop newer spray characterization techniques has always existed due to an increasing emphasis on understanding the different unsteady processes involved in liquid atomization, primarily for the aerodynamics and combustion industry. This involves complete characterization of flow fields, in terms of density, temperature, pressure, and flow velocity. Among the techniques discussed here, four are transillumination techniques (shadowgraphy, holography, x-ray phase contrast, ballistic) where an incident beam is passed through the liquid breakup and imaged on the other side, and one is an internal illumination technique (jet core illumination), one is a scattering (Mie) technique, and the rest are planar laser fluorescence/phosphorescence technique. Besides these high-speed ordinary imaging is based on reflection. From a study of the experimental diagnostics that are presently pursued by the scientific community, we can observe that even though the experimental techniques have remarkably improved over time in providing us with more information about the breakup process they are still limited in their applicability as far as meeting the complete demands is concerned. Some of the alternative non-intrusive laser-based diagnostic techniques that exist at presents such as Phase Doppler Particle Analyzer (PDPA), Phase Doppler Anemometry (PDA), Laser Doppler Anemometry (LDA), Laser Diffraction particle sizing, Laser-Induced Incandescence (LII), and Chemiluminescence Imaging, work mostly in the gaseous/vapor phase [108] or are mainly

concerned with determining the particle sizes, velocities, and volume fraction. Almost all of the present diagnostics are restricted to low Reynolds number flows to yield accurate Spatio-temporal measurements. For some very small length and time scale fluid flow situations, the spatial and temporal resolution, as well as frequency, is beyond the resolution of present-day high-speed video cameras. There is a technological need for developing ultra-high-speed cameras of sufficient dynamic range/resolution and even techniques to capture the high-speed atomization process with extremely high spatial and temporal resolution.

## **Author details**

Anu Osta Rowan University, Glassboro, USA

\*Address all correspondence to: anu.osta@hotmail.com

© 2022 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

## **References**

[1] Coghe A, Cossali GE. Quantitative optical techniques for dense sprays investigation: A survey. Optics and Lasers in Engineering. 2012;**50**:46-56

[2] Linne M. Imaging in the optically dense regions of a spray: A review of developing techniques. Progress in Energy and Combustion Science. 2013; **39**:403-440

[3] Fansler TD, Parrish SE. Spray measurement technology: A review. Measurement Science and Technology. 2015;**26**(012002):34

[4] Soid SN, Zainal ZA. Spray and combustion characterization for internal combustion engines using optical measuring techniques - a review. Energy. 2011;**36**:724-741

[5] Sridhara SN, Raghunandan BN. Photographic investigations of jet disintegration in Airblast sprays. Journal of Applied Fluid Mechanism. 2010;**3**(2): 111-123

[6] Parker T, Rainaldi LR, Rawlins WT. A comparative study of room-temperature and combusting fuel sprays near the injector tip using infrared laser diagnostics. Atomization and Sprays. 1998;**8**(5):565-600

[7] Khalid AH, Kontis K. Thermographic phosphors for high temperature measurements: Principles. Current State of the Art and Recent Applications, Sensors. 2008;**8**:5673-5744

[8] Thoroddsen ST, Etoh TG, Takehara K. High-speed imaging of drops and bubbles. Annual Review of Fluid Mechanics. 2008;**40**:257-285

[9] Bachalo WD. Spray diagnostics for the 21st century. Atomization and Sprays. 2000;**10**:439-474

[10] Vu Truong Son D, Goji Etoh T, Tanaka M, Hoang Dung N, Le Cuong V, Takehara K, et al. Toward 100 megaframes per second: Design of an Ultimate Ultra-High-Speed Image Sensor. Sensors. 2010;**10**:16-35

[11] Crua C, Shoba T, Heikal M, Gold M, Higham C. High-speed microscopic imaging of the initial stage of diesel spray formation and primary breakup. SAE International. 2010;**01**:2247

[12] Lasheras JC, Villermaux E, Hopfinger EJ. Break-up and atomization of a round water jet by a high-speed annular air jet. Journal of Fluid Mechanics. 1998;**357**(1):351-379

[13] Zhu Y, Oguz HN, Prosperetti A. On the mechanism of air entrainment by liquid jets at a free surface. Journal of Fluid Mechanics. 2000;**404**: 151-177

[14] Varga CM, Lasheras JC, Hopfinger EJ. Initial breakup of a smalldiameter liquid jet by a high-speed gas stream. Journal of Fluid Mechanics. 2003;**497**:405-434

[15] Hoyt JW, Taylor JJ. Turbulence structure in a water jet discharging in air. Physics of Fluids. 1977;**20**: S253-S257

[16] Ben-Eliezer E, Marom E, Konforti N, Zalevsky Z. Experimental realization of an imaging system with an extended depth of field. Applied Optics. 2005;**44**: 2792-2798

[17] Bauer D, Chaves H, Brucker C. A multiple-segment long-distance microscope for flow visualization and measurements. Measurement Science and Technology. 2009;**20**:1-4

[18] Nagahara H, Kuthirummal S, Zhou C, Nayar S. Flexible depth of field photography. Proceedings of European Conference on Computer Vision. 2008; **4**:60-73

[19] Dowski ER, Cathey WT. Extended depth of field through wave-front coding. Journal of Applied Optics. 1995; **34**(11):1859-1866

[20] Leong MY, McDonell VG, Samuelsen GS. Effect of ambient pressure on an airblast spray injected into a crossflow. Journal of Propulsion and Power. 2001;**17**(5)

[21] Castrejon-Garcia R, Castrejon-Pita JR, Martin GD, Hutchings IM. The shadowgraph imaging technique and its modern application to fluid jets and drops. Revista Mexicana de Fisica. 2011; **57**:266-275

[22] Osta AR, Sallam K. Nozzle-geometry effects on upwind-surface properties of turbulent liquid jets in gaseous crossflow. Journal of Propulsion and Power. 2010;**26**(5):936-946

[23] Timmerman BH, Bryanston-Cross PJ, Skeen AJ, Tucker PG, Jefferson-Loveday RJ, Dunkley P, et al. High-speed digital Shadowgraphy for high frequency shock tracking in supersonic flows. In: 7th Symposium on Measuring Techniques in Transonic and Supersonic Flows in Cascades and Turbomachines. Sweden: Stockholm; 2004

[24] Sinha A, Prakash RS, Madan Mohan A, Ravikrishna RV. Airblast spray in crossflow – Structure, trajectory and droplet sizing. International Journal of Multiphase Flow. 2015;**72**:97-111

[25] Bougie B, Tulej M, Dreier T, Dam NJ, Meulen JJ, Gerber T. Optical diagnostics of diesel spray injections and combustion in a high-pressure high-temperature cell. Applied Physics B. 2005;**80**(8): 1039-1045

[26] Lee K, Aalburg C, Diez FJ, Faeth GM, Sallam KA. Primary breakup of turbulent round liquid jets in uniform crossflows. AIAA Journal. 2007;**45**(8): 1907-1916

[27] Klein-Douwel RJH, Frijters PJM, Somers LMT, de Boer WA, Baert RSG. Macroscopic diesel fuel spray shadowgraphy using high speed digital imaging in a high pressure cell. Fuel. 2007;**86**:1994-2007

[28] Watanabe H, Okazaki K. Visualization of secondary atomization in emulsified-fuel spray flow by shadow imaging. Proceedings of the Combustion Institute. 2013;**34**:1651-1658

[29] Trainoff SP, Cannell DS. Physical optics treatment of the shadowgraph. Physics of Fluids. 2002;**14**(4):1340-1363

[30] Wu PK, Kirkendall KA, Fuller RP, Nejad AS. Breakup processes of liquid jets in subsonic crossflows. Journal of Propulsion and Power. 1997;**13**(1):64-73

[31] Sallam K, Aalburg C, Faeth GM. Breakup of round nonturbulent liquid jets in gaseous crossflow. AIAA Journal. 2004;**42**(12):2529-2540

[32] Settles GS. Schlieren and Shadowgraph Techniques: Visualizing Phenomena in Transparent Media. Berlin, Heidelberg, Germany: Springer; 2001. p. 376. Available from: https://doi. org/10.1007/978-3-642-56640-0

[33] Muirhead JC, McCallum FL. Improvement in spark shadowgraph technique. The Review of Scientific Instruments. 1959;**30**:830-831

[34] Fuller RP, Wu PK, Kirkendall KA, Nejad AS. Effects of injection angle on *Imaging Diagnostics for Jet Breakup into Droplets: A Review DOI: http://dx.doi.org/10.5772/intechopen.107370*

atomization of liquid jets in transverse airflow. AIAA Journal. 2000;**38**(1):64-72

[35] Biss MM, Settles GS, Hargather MJ, Dodson LJ, Miller JD. High-speed digital Shadowgraphy of shock waves from explosions and gunshots. Shock Waves. 2009;**II**:91-96

[36] Volpe JA, Settles GS. Laser-induced gas breakdown as a light source for Schlieren and shadowgraph particle image velocimetry. Optical Engineering. 2006;**45**(8):080509

[37] Klinner J, Willert C. Tomographic Shadowgraphy for three-dimensional reconstruction of instantaneous spray distributions. Experiments in Fluids. 2012;**53**:531-543

[38] Wang Q, Zhang Y. High speed stereoscopic shadowgraph imaging and its digital 3d reconstruction. Measurement Science and Technology. 2011;**22**(6):1-9

[39] Gabor D. Holography, 1948-1971. Science, New Series. 1972;**177**(4046): 299-313

[40] Schnars U, Juptner W. Digital recording and numerical reconstruction of holograms. Measurement Science and Technology. 2002;**13**:R85-R101

[41] Frauel Y, Naughton TJ, Matoba O, Tajahuerce E, Javidi B. Threedimensional imaging and processing using computational holographic imaging. Proceedings of the IEEE. 2006; **94**(3):636-653

[42] Leith EN, Upatnieks J. Holograms their properties and uses (hologram photography, principles, techniques and application). SPIE Journal. 1965;**4**:3-6

[43] Kubota T. 48 years with holography. Optical Review. 2014;**21**(6):883-892

[44] Milgram JH, Li W. Computational reconstruction of images from holograms. Applied Optics. 2002;**41**: 853-864

[45] Muller J, Kebbel V, Juptner W. Characterization of spatial particle distributions in a spray-forming process using digital holography. Measurement Science and Technology. 2004;**15**: 706-710

[46] Santangelo PJ, Sojka PE. A holographic investigation of the nearnozzle structure of an effervescent atomizer-produced spray. Atomization and Sprays. 1995;**5**(2):137-155

[47] Burke J, Hess CF, Kebbel V. Digital holography for instantaneous spray diagnostics on a plane. Particle and Particle Systems Characterization. 2003; **20**:183-192

[48] Miller B, Sallam KA, Bingabr M, Lin K-C, Carter CD. Breakup of aerated liquid jets in subsonic crossflow. Journal of Propulsion and Power. 2008;**24**(2): 253-258

[49] Lee J, Miller B, Sallam KA. Demonstration of digital holographic diagnostics for the breakup of liquid jets using a commercial-grade CCD sensor. Atomization and Sprays. 2009;**19**(5): 445-456

[50] Yu L, An Y, Cai L. Numerical reconstruction of digital holograms with variable viewing angles. Optics Express. 2002;**22**(10):1250-1257

[51] Yamaguchi I, Yamamoto K, Mills GA, Yokota M. Image reconstruction only by phase data in phase-shifting digital holography. Applied Optics. 2006;**45**:975-983

[52] Monnom O, Dubois F, Yourassowsky C, Legros JC. Improvement in visibility of an In-focus reconstructed image in digital holography by reduction of the influence of out-of-focus objects. Applied Optics. 2005;**44**(18):3827-3832

[53] Garcia-Sucerquia J, Xu W, Jericho SK, Klages P, Jericho MH, Kreuzer HJ. Digital In-line holographic microscopy. Applied Optics. 2006;**45**: 836-850

[54] Palero V, Arroyo MP, Soria J. Digital holography for micro-droplet diagnostics. Experiments in Fluids. 2007;**43**:185-195

[55] Choo YJ, Kang BS. The characteristics of the particle position along an optical Axis in particle holography. Measurement Science and Technology. 2006;**17**:761-770

[56] Latychevskaia T, Gehri F, Fink HW. Depth-resolved holographic reconstructions by three-dimensional deconvolution. Optics Express. 2010; **18**(21):22527-22544

[57] Gire J, Denis L, Fournier C, Thiebaut E, Soulez F, Ducottet C. Digital holography of particles: Benefits of the 'Inverse Problem' approach. Measurement Science and Technology. 2008;**19**(7):074005

[58] Wang YJ, Im K, Fezzaa K, Lee WK, Wang J. Quantitative x-ray phasecontrast imaging of air-assisted water sprays with high Weber numbers. Applied Physics Letters. 2006;**89**:151913

[59] Osta AR, Lee J, Sallam KA, Fezzaa K. Study of the effects of the injector length/ diameter ratio on the surface properties of turbulent liquid jets in still air using X-ray imaging. International Journal of Multiphase Flow. 2012;**38**(1):87-98

[60] James RW. The Optical Principles of the Diffraction of X-Rays. Ithaca, NY, USA: Cornell University Press; 1965

[61] El-Ghazaly M, Backe H, Lauth W, Kube G, Kunz P, Sharafutdinov A, et al. X-Ray Phase Contrast Imaging at MAMI. European Physical Journal A: Hadrons and Nuclei. 2006;**28**(S01):197-208

[62] Powell CF, Yue Y, Poola R, Wang J. Time-resolved measurements of supersonic fuel sprays using synchrotron X-rays. Journal of Synchrotron Radiation. 2000;**7**:356-360

[63] Guigay JP, Langer M, Boistel R, Cloetens P. Mixed transfer function and transport of intensity approach for phase retrieval in the Fresnel region. Optics Letters. 2007;**32**:1617-1619

[64] Wilkins SW, Gureyev TE, Gao D, Pogany A, Stevenson AW. Phasecontrast imaging using polychromatic hard X-rays. Nature. 1996;**384**(6607): 335-338

[65] Moon S, Liu Z, Gao J, Dufresne E, Fezzaa K, Wang J. Ultrafast X-ray phasecontrast imaging of high-speed fuel sprays from a two-hole diesel nozzle. In: 22nd Annual Conference on Liquid Atomization and Spray Systems. Cincinnati, OH: ILASS; 2010

[66] Peterzol A, Berthier J, Duvauchelle P, Ferrero C, Babot D. Xray phase contrast image simulation. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms. 2007;**254**(2):307-318

[67] Sedarsky D, Paciaroni M, Berrocal E, Petterson P, Zelina J, Gord J, et al. Model validation image data for breakup of a liquid jet in crossflow: Part I. Experiments in Fluids. 2010;**49**(2): 391-408

[68] Sedarsky D, Gord J, Carter C, Meyer T, Linne M. Fast-framing ballistic *Imaging Diagnostics for Jet Breakup into Droplets: A Review DOI: http://dx.doi.org/10.5772/intechopen.107370*

imaging of velocity in an aerated spray. Optics Letters. 2009;**34**:2748-2750

[69] Linne MA, Paciaroni M, Hall T, Parker T. Ballistic imaging of the near field in a diesel spray. Experiments in Fluids. 2006;**40**(6):836-846

[70] Linne MA, Paciaroni M, Berrocal E, Sedarsky D. Ballistic imaging of liquid breakup processes in dense sprays. Proceedings of the Combustion Institute. 2009;**32**(II):2147-2161

[71] Paciaroni M, Linne M. Single-shot, two-dimensional ballistic imaging through scattering media. Applied Optics. 2004;**43**:5100-5109

[72] Paciaroni M, Hall T, Delplanque JP, Parker T, Linne M. Single-shot twodimensional ballistic imaging of the liquid Core in an atomizing spray. Atomization and Sprays. 2006;**16**(1): 51-70

[73] Idlahcen S, Rozé C, Méès L, Girasole T, Blaisot JB. Sub-picosecond ballistic imaging of a liquid jet. Experiments in Fluids. 2012;**52**:289-298

[74] Sedarsky D, Paciaroni M, Linne M, Gord J, Meyer T. Velocity imaging for the liquid–gas interface in the near field of an atomizing spray: Proof of concept. Optics Letters. 2006;**31**:906-908

[75] Idlahcen S, Mees L, Roze C, Girasole T, Blaisot J. Time gate, optical layout, and wavelength effects on ballistic imaging. Journal of the Optical Society of America. A. 2009;**26**(9): 1995-2004

[76] Charalampous G, Hardalupas Y, Taylor A. Novel technique for measurements of continuous liquid jet Core in an atomizer. AIAA Journal. 2009;**47**(11):2605-2615

[77] Charalampous G, Hadjiyiannis C, Hardalupas Y. Proper orthogonal decomposition analysis of photographic and optical connectivity time resolved images of an atomizing liquid jet. In: 24th Ann. Conf. Liquid Atom. Spray Sys. Portugal: ILASS; 2011. pp. 1-8

[78] Elshamy OM, Tambe SB, Cai J, Jeng SM. Structure of liquid jets in subsonic crossflow at elevated ambient pressures. In: 44th AIAA Aerospace Sciences Meeting and Exhibit. Reno: AIAA; 2006. pp. 1-11

[79] Smallwood GJ, Gulder OL, Snelling DR. The structure of the dense Core region in transient diesel sprays. 25th Symposium (International) on Combustion. 1994;**25**(1):371-379

[80] Fansler TD, Drake MC. "Designer diagnostics" for developing directinjection gasoline engines. Journal of Physics Conference Series. 2006;**45**:1-17

[81] Sick V, Drake MC, Fansler TD. Highspeed imaging for direct-injection gasoline engine Research and Development. Experiments in Fluids. 2010;**49**(4):937-947

[82] Stenzler JN, Lee JG, Santavicca DA. Penetration of liquid jets in a crossflow. Atomization and Sprays. 2006;**16**(8): 887-906

[83] Tambe S, Elshamy O, Jeng S. Spray properties of liquid jets injected transversely into a shear layer. In: 43rd AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit. Cincinnati, OH: AIAA; 2007. pp. 2007-5695

[84] Rachner M, Becker J, Hassa C, Doerr T. Modelling of the atomization of a plain liquid fuel jet in crossflow at gas turbine conditions. Aerospace Science and Technology. 2002;**6**(7):495-506

[85] Han D, Mungal MG, Orozco V. Gross-entrainment behavior of turbulent jets injected obliquely into a uniform crossflow. AIAA Journal. 2000;**38**(9): 1643-1649

[86] Siebers, D.L., (1998) Liquid-Phase Fuel Penetration in Diesel Sprays, International Congress & Exposition, Detroit, MI, Paper No. 980809.

[87] Niederhaus CE, Champagne FH, Jacobs JW. Scalar transport in a swirling transverse jet. AIAA Journal. 1997; **35**(11):1697-1704

[88] Cloeter MD, Qin K, Patil P, Smith B. Planar laser induced fluorescence (PLIF) flow visualization applied to agricultural spray nozzles with sheet disintegration; influence of an oil-in-water emulsion. In: 22nd Annual Conference on Liquid Atomization and Spray Systems. Cincinnati, OH: ILASS; 2010

[89] Pastor JV, Lopez JJ, Julia JE, Benajes JV. Planar laser-induced fluorescence fuel concentration measurements in isothermal diesel sprays. Optics Express. 2002;**10**:309-323

[90] Davidson MJ, Pun KL. Weakly Advected jets in Cross-flow. Journal of Hydraulic Engineering. 1999;**125**(1): 47-58

[91] Wieske P, Wissel S, Grünefeld G, Pischinger S. Improvement of LIEF by wavelength-resolved acquisition of multiple images using a single CCD detector– Simultaneous 2D measurement of air/fuel ratio, temperature distribution of the liquid phase and qualitative distribution of the liquid phase with the multi-2D technique. Applied Physics B. 2006; **83**(2):323-329

[92] Webster DR, Longmire EK. Jet pinch-off and drop formation in

immiscible liquid–liquid systems. Experiments in Fluids. 2001;**30**(1):47-56

[93] Milosevic IN, Longmire EK. Pinchoff modes and satellite formation in liquid/liquid jet systems, Int. Journal of Multiphase Flow. 2002;**28**:1853-1869

[94] Kristensson E, Berrocal E, Ritcher M, Pettersson S-G, Aldén M. High-speed structured planar laser illumination for contrast improvement of two-phase flow images. Optics Letters. 2008;**33**(23):2752-2754

[95] Kristensson E, Berrocal E, Richter M, Alden M. Nanosecond structured laser illumination planar imaging for singleshot imaging of dense sprays. Atomization and Sprays. 2010;**20**(4): 337-343

[96] Berrocal E, Kristensson E, Hottenbach P, Alden M, Grunefeld G. Quantitative imaging of a noncombusting diesel spray using structured laser illumination planar imaging. Applied Physics B: Lasers and Optics. 2012;**109**:683-694

[97] Desantes JM, Pastor JV, Pastor JM, Julia JE. Limitations on the use of the planar laser induced Exciplex fluorescence technique in diesel sprays. Fuel. 2005;**84**(18):2301-2315

[98] Melton LA, Verdieck JF. Vapor/ Liquid Visualization in Fuel Sprays. International Symposium on Combustion. 1985;**20**(1):1283-1290

[99] Murray AM, Melton LA. Fluorescence methods for determination of temperature in fuel sprays. Applied Optics. 1985;**24**(17):2783-2787

[100] Lemoine F, Antoine Y, Wolff M, Lebouche M. Simultaneous temperature and 2d velocity measurements in a turbulent heated jet using combined

*Imaging Diagnostics for Jet Breakup into Droplets: A Review DOI: http://dx.doi.org/10.5772/intechopen.107370*

laser-induced fluorescence and LDA. Experiments in Fluids. 1999;**26**:315-323

[101] Depredurand V, Miron P, Labergue A, Wolff M, Castanet G, Lemoine F. A temperature-sensitive tracer suitable for two-color laserinduced fluorescence thermometry applied to evaporating fuel droplets. Measurement Science and Technology. 2008;**19**(10):105103

[102] Castanet G, Lavieille P, Lebouché M, Lemoine F. Measurement of the temperature distribution within monodisperse combusting droplets in linear streams using two-color laserinduced fluorescence. Experiments in Fluids. 2003;**35**(6):563-571

[103] Omrane A, Sarner G, Alden M. 2d-temperature imaging of single droplets and sprays using thermographic phosphors. Applied Physics B: Lasers and Optics. 2004;**79**:431-434

[104] Yi SJ, Kim KC. Phosphorescencebased multiphysics visualization: A review. Journal of Visualization. 2014;**17**: 253-273

[105] Allison SW, Gillies GT. Remote thermometry with thermographic phosphors: Instrumentation and applications. The Review of Scientific Instruments. 1997;**68**(7):2615-2650

[106] Tran T, Kochar Y, Seitzman JM. Measurements of acetone fluorescence and phosphorescence at high pressures and temperatures. In: Paper AIAA-2006-831-356, 44th AIAA Aerospace Science Meeting. Reno, NV: AIAA; 2006

[107] Ritchie BD, Seitzman JM. Simultaneous imaging of vapor and liquid spray concentration using combined acetone fluorescence and phosphorescence. In: Paper

AIAA-2004-0384, 42nd AIAA Aerospace Science Meeting. Reno, NV: AIAA; 2004

[108] Wigley G, Goodwin M, Pitcher G, Blondel D. Imaging and PDA analysis of a GDI spray in the near-nozzle region. Experiments in Fluids. 2004;**36**(4): 565-574

Section 3

## Fundamental Research on Liquid Droplet Behaviors

## **Chapter 3**

## Behaviors of Multi-Droplets Impacting on a Flat Wall

*Hongliang Luo and Feixiang Chang*

## **Abstract**

Microscopic characteristics of fuel spray are very important for atomization and mixture formation. The droplet size, number density, velocity distribution as well as minimum distance reveal the quality of spray and atomization, which affects the subsequent combustion and emissions for different engines such as vehicle, marine and aircraft. Moreover, in the internal combustion engine, the spray-wall impingement is difficult to avoid, which is the main source for soot emissions. Nowadays, regulations for emissions become straight by governments. Therefore, it is urgent for us to alleviate the energy and emissions crisis. In this study, the droplets behaviors will be characterized under the related engine working state. Firstly, the experimental setup and measurement were explained in detail. Then, images process method was induced to calculate the droplet size, velocity and distance among them. Finally, results of the impinging spray were presented. One thing should be noted, as the dense region is not available to detect the droplets by the observation. Therefore, a spray "slicer" was designed and applied to cut the spray slim. Finally, multi-droplets were generated, and the results can be concluded as well. All the results could provide insights into the impacting behaviors for better understanding the droplet dynamics.

**Keywords:** multi-droplets, impacting behaviors, energy conversion, microscale analysis

## **1. Introduction**

Droplet-wall interactions are widely used in the gas-turbine and internal combustion engines under a range of different conditions [1]. After injected into the cylinder, droplets with various velocities travel along the spray axis to impact on the piston head or cylinder wall to form fuel film as well as breakup into small ones splashing off the wall [2]. Moreover, the rebounding and splashing droplets coalesce and collide with each other in the air by the help of vortex, which may impinge on the cylinder wall again or eventually adhere on it [3]. Especially, for the cold-start operation, the liquid film on the wall surface significantly contributes to the soot emissions [4]. Therefore, understanding the mechanism of impinging spray is essential for better design and optimization on the engine performance. Specially, the investigations on droplet characteristics could contribute to model the impinging behaviors for separating the deposit/splash droplets under a series of pressure and temperature conditions.

Recently, droplets impacting on the wall becomes a hotpot for investigations, especially after Moreira et al. [5] put forward the question that "How much of single droplet impact research is useful?". As a fundamental study, the single drop-wall impacting can clearly explain the relationship between the splashing behavior and non-dimensional parameters, such as Reynolds, Weber, Ohnesorge and K numbers [6]. Previously, Walzel [7] firstly described the droplets impinging behavior. Then, the splashing characteristics was formulated by Yarin and Weiss [8]. Followed by them, Mundo et al. [9] firstly applied K number to separate droplets behaviors of "splashing" and "no-splashing". Then, Bai et al. [10] characterized the different impingement regimes of the incident drops by Weber number. Then it was refined and developed well by other scholars [11–13] Recently, Cen et al. [14] studied the dynamic break-up of the alternative biofuel impacting on the hot wall under the film boiling regime with different Weber number. They found that droplet breakup was dominated by Rayleigh instability, and the timing of jet break-up agreed well with the theory of Rayleigh instability. Qin et al. [15] investigated the splashing behaviors of an impacting drop on the hot surface to compare the effects of fuel viscosity and surface roughness. Liu et al. [16] studied the spreading dynamics of a single droplet impacting on a heated surface. They found that impacting behavior such as "depositing", "rebounding" and "breakup with atomization" were sensitive to both surface temperature and Weber number.

According to the literature surveys above, although many investigations were performed on the drop-wall interaction. It is still necessary to conduct an in-depth study on the multi-droplets wall-impingement. This study aims to fill the research gap. The target is to investigate the characteristics of multi-drops impingement on the wall experimentally under various conditions. Particle image analysis (PIA) technique was implemented to obtain the microscopic observations. Firstly, the spray droplets were checked by a single hole injector. Then, a spray "slicer" was introduced to make the multi-droplets impinging on the wall. Moreover, the thickness of "slicer" was defined. Finally, 0.4 and 0.04 mm in thickness of the "slicer" were determined to test at the spray tip and quasi-steady state.

## **2. Experimental system and image processing method**

The experimental setup is shown in **Figure 1**. It consists of three parts, including an injection system, an optical system and one constant volume chamber. Fuel was injected into the chamber, then developed and impinged on a flat wall by a high-pressure common rail. The high-pressure constant volume chamber was equipped with four glass-windows to transmit the observation light. Similar to the shadow method, particle image analysis (PIA) was applied in this research, which was also used in our published literatures [17, 18]. Therefore, here only a brief introduction is shown. A double pulsed Nd: YAG laser of 532 nm wavelength was used to illuminate the spray. And microscopic behaviors of spray can be observed through a charge-coupled device (CCD) camera (Flowtech Research Inc., FtrNPC). Moreover, a microscope was used to connect the camera to obtain the tiny droplets clearly with three teleconverters (Kenko Tokina, N-AF 1.4X TELEPLUS MC4\*3) connected to amplify the microimage. Besides, owing to the dense drops along the spray axis, a house-designed spray "slicer" was designed and applied to get the slim spray, as depicted in **Figure 2**. The spray "slicer" made of steel with length of each slicer at 30 mm and height at 3 mm. And the distance between two slicer was defined as the thickness, which could be adjusted from 0.04 to 1 mm.

*Behaviors of Multi-Droplets Impacting on a Flat Wall DOI: http://dx.doi.org/10.5772/intechopen.105007*

**Figure 1.** *Experimental apparatus.*

**Table 1** lists the experimental conditions. On mini-sac injector with a single hole of 0.135 mm in diameter and 0.65 mm in hole length was used, resulting in the lengthto-diameter (L/D) ratio at 4.8. Toluene was used as the tested fuel with density being 866 kg/m3 under the room temperature. The properties such as viscosity and surface tension under the room temperature are 5.89 × 10–4 N·s/m2 and 0.0285 N/m, respectively. The injection mass of toluene was kept constant at 4.0 mg. The chamber was filled with nitrogen to conduct the experiment under the non-evaporation condition. And the ambient pressure was changed between 0.1 and 0.5 MPa, leading to the gas


#### **Table 1.**

*Experimental conditions.*

density at 1.19 and 5.95, respectively. Besides, the injection pressure was set among 10, 20 and 30 MPa with injection duration at 2.9,2.1 and 1.7 ms. A flat wall made of quartz glass was used as the impingement wall, as shown in **Figure 3**. The impingement distance from the nozzle exit to the wall was decided at 22 mm. And the impingement angle was determined at 45 deg. between the spray axis and the wall surface. One thing should be noted that the roughness of the flat wall was measured at Ra7.0 μm by a portable high-performance surface roughness and waviness measuring instrument (Kosaka Laboratory Ltd., SE300).

The image processing is shown in **Figure 3**. Firstly, the resolution can be calculated by the help of a micro-scale ruler. For filtering the spherical droplets, some refined criteria should be defined, such as roundness, pixel number and diameter. After detecting the spherical droplets, the distance between droplets can be obtained. Furthermore, with the interval timing between these two frames at 0.4 μs, the velocity of each droplet can be calculated. Moreover, in order to make the statistical results reasonable, at least 100 repeated tests were conducted under the same condition to make sure more than 5000 particles are available for statistics. Finally, the droplet

*Behaviors of Multi-Droplets Impacting on a Flat Wall DOI: http://dx.doi.org/10.5772/intechopen.105007*

**Figure 3.** *Image processing.*

size, velocity, distance and non-dimensional results can be gotten. The details about the algorithm can be seen in the article of Wang et al. [19].

## **3. Results and discussion**

#### **3.1 Impinging spray and droplet behaviors**

Microscopic spray development before impingement under 0.1 and 0.5 MPa are presented in **Figure 4**. It can be seen that all the images are arranged along the spray axis and then rotated 45 deg. at counterclockwise connected. Under Pamb = 0.1 MPa, when fuel spray at the near field region, less droplets can be seen. With development at 2–4 mm from the origin, it is clear to see the liquid column. With longer penetration development at 8–10 mm, better atomization can be seen that liquid sheet located in the core with droplets at the periphery. At further locations, more ligaments and clear droplets can be captured. Besides, increased the ambient pressure to 0.5 MPa, more atomized droplets can be obtained because of the strong sheer force to promote droplets collision, leading to dense spray when compared to that of 0.1 MPa. In this case, although PIA could obtain the microscopic spray clearly at periphery, it is still difficult to clear observe the drops at the dense spray regions.

Then, typical morphologies at (10, 10) and (14, 14) are shown in **Figure 5**. Before impingement at (10, 10), both ligaments and droplets can be seen at 0.1 MPa due to the primary breakup. Under 0.5 MPa, intense interactions result in more small size droplets. Transferred to being impingement location at (14, 14). The white solid line represents the surface of the flat wall. At the near wall region, it is too dense to identify the droplets along the spray axis when compared to (10, 10), resulting in only droplets at periphery can be observed.

**Figure 4.** *Microscopic behaviors of spray and droplets.*

Generally, the atomized droplets are the outcomes of competitions between aerodynamic force and surface tension. The friction in the gas-fuel interface results in the disintegration and instable deformation, finally contributing to the formation of smaller droplets. Therefore, *We*g can be used to describe instability of the droplet formations and predicts the behaviors of droplet breakup furtherly, defined as:

$$\text{We}\_{\mathfrak{g}} = \frac{\rho\_{\mathfrak{g}} \boldsymbol{v}^2 d}{\sigma} \tag{1}$$

*Behaviors of Multi-Droplets Impacting on a Flat Wall DOI: http://dx.doi.org/10.5772/intechopen.105007*

**Figure 5.** *Spray morphology under different ambient pressures.*

where ρg is the ambient gas density with σ being the liquid surface tension. *v* is the relative velocity between droplet and ambient gas. *d* represents the droplet diameter.

Previously, Wierzba [7] found the vibrational breakup occurs at a relatively low *We*g, and droplets are easy to break up into smaller ones. Therefore, *We*g can be applied to indicate the degree of atomization during development, which is divided into four parts: when 0<*We*g<1 and 1<*We*g<10, it suggests good atomization; when 10<*We*g<14, the transient section indicates the further atomization can be expected; when *We*g>14, the large droplets still exist.

The diameter-velocity distribution is presented in **Figure 6**. The scatter number in each sub-figure suggests the number of available droplets. The horizontal axis is droplet size (6–80 μm), and vertical axis is velocity (0–200 m/s). At 0.1 MPa, compared between (10, 10) and (14, 14) in Figures. (a) and (b), results are similar. However, more droplets with larger *We*g can be seen when ambient pressure increased to 0.5 MPa in Figures. (c) and (d), suggesting these droplets have the potential for further atomization. Besides, in contrast to Figures. (c) and (d), the number of large *We*g decreases significantly, indicating that higher ambient pressure favors fuel breakup and atomization. In other words, it can be concluded that fuel spray development is the process of *We*g decreasing with better atomization.

The number probability of droplet size is calculated in **Figure 7**. Before impingement (10, 10), two curves are similar but with different peak values, and the peak value of 0.1 MPa is larger than that of 0.5 MPa. However, difference becomes not obvious. Furthermore, the droplet size corresponding to the peak value of curves is the same at 12 μm. One possible reason may be that the scale of atomization is mainly decided by the injector itself, resulting in less relationship with the ambient pressure. In order to discuss the velocity distribution, the probability of velocity is presented in

**Figure 6.** *Diameter–velocity distribution with gas phase weber number.*

**Figure 7.** *Droplet diameter probability curves.*

**Figure 8**. With the increase in ambient pressure, the maximum velocity reduces from 200 to 150 m/s. Besides, the velocity corresponding to the peak value decreases at elevated ambient pressure decreases from 65 to 25 m/s at (10, 10), from 45 to 10 m/s at (14, 14). Besides, different to droplet size distribution, the peak value in velocity decreases sharply under high ambient pressure owing to more energy dispersion.

*Behaviors of Multi-Droplets Impacting on a Flat Wall DOI: http://dx.doi.org/10.5772/intechopen.105007*

**Figure 8.** *Droplet velocity probability.*

**Figure 9.** *(a) SMD and (b) averaged velocity.*

All in all, it can be seen that velocity is sensitive to the location and ambient pressure significantly compared to diameter. And both ambient pressure and spray development decelerate droplets.

Sauter mean diameter (SMD = ∑D3/D2) is defined as the diameter of a droplet with the same ratio of volume to surface area, which is widely applied to evaluate the atomization, as shown in **Figure 9**. Here, SMD and averaged velocity under each condition are compared. SMD decreases with elevated ambient pressure and spray development, resulting in better atomization. The ambient shear force and high gas density should be the main reasons for it. Moreover, the averaged velocity decreases as well because of the resistance force. And the minimum distance of the selected droplets is defined as the shortest distance among them, as shown in **Figure 10**, which can be applied to represent the spray dispersion. The results of mean minimum distance (MMd) are calculated and depicted here. Similar to SMD and averaged velocity, MMd decreases with ambient pressure increasing but with different reason. That high ambient pressure compacts droplets, making them together due to the high ambient density should be one possible reason for it.

**Figure 10.**

*Definition and results of mean minimum distances.*

## **3.2 Sliced spray behaviors**

In this part, the spray behaviors will be discussed at the spray tip for clear observation. Moreover, the spray "slicer" was applied here with the thickness at 0.4 mm.

Firstly, **Figure 11** shows the morphology of spray tip at two locations under the injection pressure of 10 and 30 MPa, respectively. At (10, 10), clear droplets, ligaments and even liquid sheets can be seen, indicating the spray slicer improves the atomization. Moved to the impinging location at (14, 14), the splashing crown structures can be seen clearly, which previously reported in the single drop impacting experiment [20, 21]. Besides, the photographing timing is recorded in each image. With an increase in injection pressure, shorter timing is needed due to high momentum energy from the drops.

**Figure 11.** *Spray morphology under different injection pressures.*

*Behaviors of Multi-Droplets Impacting on a Flat Wall DOI: http://dx.doi.org/10.5772/intechopen.105007*

**Figure 12.** *"Fish-shape" distribution.*

Droplet diameter-velocity distribution is presented in **Figure 12**. The horizontal axis is diameter, and vertical axis is velocity. Besides, SMD is depicted in each image. It is interesting to find the "fish-shape" distribution under all conditions with big "head" and small "tail". Furthermore, increased the injection pressure from 10 to 30 MPa, distribution becomes wider with higher velocity because of high kinetic energy. In contrast to location effect, the influence of injection pressure plays a much more important role on not only velocity but droplet size. It is evident that increasing injection pressure reduces SMD significantly, when compared it between (10, 10) and (14, 14).

The size probability is presented in **Figure 13**. It is interesting to find the almost half droplet diameter is in the range of 12–20 μm. Moreover, the peak value corresponding to diameter locates in only 12–16, indicating the droplet size is determined largely by the injector itself, instead of the injection pressure. Besides, the peak value increases with spray tip movement and injection pressure owing to better atomization. Furthermore, injection pressure shows lager effect on droplets atomization than that of measurement location.

**Figure 14** shows the velocity distribution along x and y directions to describe both velocity and direction. Four quadrants are separated to define the velocity direction. Quadrant I indicates the incident droplet moving along the spray axis; Quadrant II indicates droplet splashing off the wall in the direction perpendicular to spray axis; Quadrant III indicates droplet rebounding into the air along the spray axis; Quadrant IV indicates droplet rushing to the wall in the completely opposite direction of Quadrant II. Moreover, the absolute velocity is the distance from one certain dot to the origin. Because of the spray tip, all the droplets show similar rushing direction to the wall in Quadrant I. And almost the same shape can be seen under 10 MPa. But increasing the injection pressure to 30 MPa, droplet velocity can reach to 200 m/s.

**Figure 13.** *Droplet size distribution.*

**Figure 14.** *Droplet velocity direction.*

**Figure 15** shows the diameter-minimum distance. As explained before, minimum distance (Md) could be used to judge the droplet dispersion with larger Md to indicate the better dispersion. All results show similar distribution with little difference. One thing can be concluded from distributions that increased the injection enlarges Md at a certain level. Therefore, in order to discuss the Md clearly, mean minimum distance (MMd) is used in **Figure 16** as well as averaged velocity. For averaged velocity, increasing injection pressure accelerates droplet greatly from 83.3 to 137.0 m/s at (10, 10). But this increasing tendency becomes moderate at (14, 14). In addition, averaged velocity decreases slightly with tip movement due to the air drag force. For MMd, both movement and injection pressure enlarge MMd, suggesting better atomization. However, one interesting thing is that different to velocity, MMd is more sensitive to the location instead of injection pressure. Droplets disperse well with the spray development should be one possible reason for it.

*Behaviors of Multi-Droplets Impacting on a Flat Wall DOI: http://dx.doi.org/10.5772/intechopen.105007*

**Figure 15.** *Droplet diameter-Md distribution.*

*(a) Averaged velocity and (b) MMd.*

## **3.3 Multi-droplets impacting characteristics**

In order to explore the multi-droplets impacting behaviors deeply, the thickness of "slicer" was decreased to 0.04 mm, as shown in **Figure 17**. At spray tip, droplets can be recognized at thickness of 0.4 and 0.1 mm. However, changed to quasi-steady state, the thickness should be decreased much more. Finally, the thickness was determined at 0.04 mm. And in this part, multi-droplets can be achieved through the "slicer". Moreover, the impacting and splashing behaviors of droplets will be discussed in detail.

**Figure 18** presents the diameter-velocity distribution at (16, 14) and (18, 14). The horizontal axis is the droplet diameter, with vertical axis being as the velocity. Results of 10, 20 and 30 MPa are depicted by the black, red and blue data, respectively. For the velocity, at (16, 14), the velocity ranges from 0 to 110 m/s. But it decreases to only 50 m/s when the location is transferred to (18, 14). And it is supposed that the energy

#### **Figure 17.**

*Comparison in the clear observations with different thickness of "slicer".*

**Figure 18.** *Droplet diameter-velocity distribution.*

disperses with the interaction of droplets and fuel film on the wall during the injection, which will be discussed more in the following part. Moreover, for the droplet size, it is clear to see that diameter increases significantly when transferring from (16, 14) to (18, 14). Droplets collision with each other after impacting on the wall should be one possible reason for it.

In order to analyze the size distribution statistically, the probability of the diameter is applied and shown in **Figure 19**. At (16, 14), the peak value corresponding to the diameter increases with higher injection pressure from 20–25%. In contrast to (18, 14), peak value declines obviously, suggesting that droplet size increases from (16, 14) to (18, 14) with spray movement but decreases with increasing injection pressure. Finally, SMD is calculated and depicted in **Figure 20**. It shows that SMD decreases slightly with an increase in injection pressure at (16, 14). However, larger influence of injection pressure can be seen at (18, 14). The larger-size drops are easy to break up into smaller ones owing to high kinetic energy. While, by comparing the *Behaviors of Multi-Droplets Impacting on a Flat Wall DOI: http://dx.doi.org/10.5772/intechopen.105007*

**Figure 19.** *Probability of diameter at different locations. (a) @(16, 14) (b) @(18, 14).*

**Figure 20.** *SMD at different locations.*

location effect, SMD is enlarged with movement. After impingement on the wall, the splashing droplets may collide and coalesce by others can explain this phenomenon. More details will be discussed through the velocity analysis.

**Figure 21** shows the velocity-direction distribution. X and Y coincide with the direction presented in **Figure 4**. And the distance from the origin to one selected droplet presents the magnitude velocity. At (16, 14), almost all the droplets rush into the wall along the spray jet direction. Besides, high injection pressure accelerates droplets. While, changed to (18, 14), droplets change direction in Quadrant II, indicating the splashing direction. And this is the main reason for the low velocity and large size. After impacting on the wall, droplets splashing off the wall results in energy loss. Additionally, some droplets may collide and change the direction as well as coalescence into larger ones. One thing should be pointed that at (16, 14) and (18, 14), droplets locate in all these four quadrants because of the vortex.

Next, the averaged velocity was calculated and presented in **Figure 22**. It shows that at (16, 14), velocity increases significantly with higher injection pressure. Compared with after impingement at (18, 14), only half velocity can be seen due to

**Figure 21.**

*Droplet velocity distribution at different locations.*

**Figure 22.** *Averaged velocity at different locations.*

the interaction consuming energy. Finally, the mechanisms of multi-droplets impacting on the wall is conceived. After impingement, the secondary droplets are generated by the irregular crown structure with the help of the interaction between fuel film and droplets. On one hand, the splashing droplets after impingement collide with each other and then coalesce into larger one. On the other hand, the splashing crown structures also collide owing to the multi-droplets crash into the fuel film, leading to droplets generated from the rim of the crown body. Moreover, the crown bodies collide, then generate droplets splashing off the structure to increase the totally calculated SMD.

## **4. Conclusions**

In this study, droplet behaviors were investigated before, during and after impingement on the wall. For clear observation, free and impinging spray (with/ without "slicer") at different thickness of 0.4 and 0.04 mm were characterized to

## *Behaviors of Multi-Droplets Impacting on a Flat Wall DOI: http://dx.doi.org/10.5772/intechopen.105007*

evaluate the droplet size, velocity and distance. Besides, different influences including measurement locations, injection pressures and ambient pressures were considered to better understand the behaviors of multi-droplets impacting on the wall. The following conclusions can be drawn as:


## **Acknowledgements**

The author would like to acknowledge National Natural Science Foundation of China [Grant 51909037] and Foundation of State Key Laboratory of Engines [No. K2022-12].

## **Conflict of interest**

The authors declare no conflict of interest.

## **Author details**

Hongliang Luo1,2\* and Feixiang Chang2

1 Foshan University, Foshan, P.R. China

2 Hiroshima University, Higashi-Hiroshima, Japan

\*Address all correspondence to: luo@hiroshima-u.ac.jp

© 2022 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

*Behaviors of Multi-Droplets Impacting on a Flat Wall DOI: http://dx.doi.org/10.5772/intechopen.105007*

## **References**

[1] Luo H, Uchitomi S, Nishida K, Ogata Y, Zhang W, Fujikawa T. Experimental investigation on fuel film formation by spray impingement on flat walls with different surface roughness. Atomization and Sprays. 2017;**27**:611-628. DOI: 10.1615/ AtomizSpr.2017019706

[2] Hiroyasu H, Arai M. Structures of fuel sprays in diesel engines. SAE Technical Paper. 1990:900475. DOI: 10.4271/900475

[3] Luo H, Nishida K, Uchitomi S, Ogata Y, Zhang W, Fujikawa T. Microscopic behavior of spray droplets under flat-wall impinging condition. Fuel. 2018;**219**:467-476. DOI: 10.1016/j. fuel.2018.01.059

[4] Luo H, Nishida K, Uchitomi S, Ogata Y, Zhang W, Fujikawa T. Effect of temperature on fuel adhesion under spray-wall impingement condition. Fuel. 2018;**234**:56-65. DOI: 10.1016/j. fuel.2018.07.021

[5] Moreira ALN, Moita AS, Panao MR. Advances and challenges in explaining fuel spray impingement: How much of single droplet impact research is useful? Progress in Energy and Combustion Science. 2010;**36**:554-580. DOI: 10.1016/j. pecs.2010.01.002

[6] Cossali GE, Coghe ALDO, Marengo M. The impact of a single drop on a wetted solid surface. Experiments in Fluids. 1997;**22**:463-472. DOI: 10.1007/ s003480050073

[7] Wierzba A. Deformation and breakup of liquid drops in a gas stream at nearly critical weber numbers. Experiments in Fluids. 1990;**9**:59-64. DOI: 10.1007/ BF00575336

[8] Yarin AL, Weiss DA. Impact of drops on solid surfaces: Self-similar capillary waves and splashing as a new type of kinematic discontinuity. Journal of Fluid Mechanics. 1995;**283**:141-173. DOI: 10.1017/S0022112095002266

[9] Mundo CHR, Sommerfeld M, Tropea C. Droplet-wall collisions: Experimental studies of the deformation and breakup process. International Journal of Multiphase Flow. 1995;**21**:151-173. DOI: 10.1016/0301-9322(94)00069-V

[10] Bai CX, Rusche H, Gosman AD. Modeling of gasoline spray impingement. Atomization and Sprays. 2002;**12**:1-27. DOI: 10.1615/AtomizSpr. v12.i123.10

[11] Sivakumar D, Tropea C. Splashing impact of a spray onto a liquid film. Physics of Fluids. 2002;**14**:85-88. DOI: 10.1063/1.1521418

[12] Kalantari D, Tropea C. Spray impact onto flat and rigid walls: Empirical characterization and modelling. International Journal of Multiphase Flow. 2007;**33**:525-544. DOI: 10.1016/j. ijmultiphaseflow.2006.09.008

[13] Roisman IV, Breitenbach J, Tropea C. Thermal atomisation of a liquid drop after impact onto a hot substrate. Journal of Fluid Mechanics. 2018;**842**:87-101. DOI: 10.1017/jfm.2018.123

[14] Cen C, Wu H, Lee CF, Liu F, Li Y. Experimental investigation on the characteristic of jet break-up for butanol droplet impacting onto a heated surface in the film boiling regime. International Journal of Heat and Mass Transfer. 2018;**123**:129-136. DOI: 10.1016/j. ijheatmasstransfer.2018.02.059

[15] Qin M, Guo Y, Tang C, Zhang P, Huang Z. Spreading and bouncing of liquid alkane droplets upon impacting on a heated surface. International Journal of Heat and Mass Transfer. 2020;**159**:120076. DOI: 10.1016/j. ijheatmasstransfer.2020.120076

[16] Liu H, Si C, Cai C, Zhao C, Yin H. Experimental investigation on impact and spreading dynamics of a single ethanol–water droplet on a heated surface. Chemical Engineering Science. 2021;**229**:116106. DOI: 10.1016/j. ces.2020.116106

[17] Luo H, Jin Y, Nishida K, Ogata Y, Yao J, Chen R. Microscopic characteristics of impinging spray sliced by a cone structure under increased injection pressures. Fuel. 2021;**284**:119033. DOI: 10.1016/j. fuel.2020.119033

[18] Luo H, Nishida K, Ogata Y, Zhang W. Microscopic characteristics of nearnozzle spray at the initial and end stages. Fuel. 2021;**283**:118953. DOI: 10.1016/j. fuel.2020.118953

[19] Wang C, Wang Y, Zhu J, Long W, Nishida K. Design and application of an algorithm for measuring particle velocity based on multiple characteristics of a particle. Fuel. 2020;**271**:117605. DOI: 10.1016/j.fuel.2020.117605

[20] Krechetnikov R, Homsy GM. Crownforming instability phenomena in the drop splash problem. Journal of Colloid and Interface Science. 2009;**331**(2): 555-559. DOI: 10.1016/j.jcis.2008.11.079

[21] Rioboo R, Bauthier C, Conti J, Voue M, DeConinck J. Experimental investigation of splash and crown formation during single drop impact on wetted surfaces. Experiments in Fluids. 2003;**35**(6):648-652. DOI: 10.1007/ s00348-003-0719-5

## **Chapter 4**

## Interaction and Transport of Liquid Droplets in Atmospheric Pressure Plasmas (APPs)

*Muhammad M. Iqbal and Mark M. Turner*

## **Abstract**

The transport of liquid droplets in atmospheric pressure plasma (APP) has been recognized as a two-phase flow. The liquid droplet introduces a perturbation in APP and under several constraints, the behavior of this perturbation is not clear during transport. As soon as the droplets interact with the APP, they become charged, which causes the droplets to evaporate. Using 1D normalized fluid model, we first describe how the liquid droplets charge and interact with helium APP. The impact of this droplet-plasma interaction is then discussed and described using 2D coupled fluid-droplet model. The numerical modeling outcomes suggested that the evaporation of droplets has emerged as a primary mechanism in plasma; however, the mutual interactions, such as grazing and coalescence between the droplets, are dominant at higher precursor flow rates (>100 μl min<sup>1</sup> ). To demonstrate the importance of different liquid precursors during dropletplasma interaction, we analyzed the spatiotemporal patterns of discharge plasma while considering the effects of HMDSO, n-hexane, TEOS, and water. Variable gas and liquid precursor flow rates are used to further examine the features of discharge plasma. Comparing penning ionization to other ionization processes, it is found to be the prime activity along the pulse of droplets, demonstrating the significance of small nitrogen impurities. Using a laser diffraction particle size analysis approach as part of an APP jet deposition system, the validation of the numerical simulations is proven by comparison with experimental observations of droplet size distributions.

**Keywords:** atmospheric pressure plasma, droplet charging, evaporation, interaction, droplet model, two-phase flow, precursor flow, coupled model, grazing, coalescence, helium gas, N2 impurities, penning ionization, deposition system, plasma simulations

## **1. Introduction**

It has been widely accepted that atmospheric pressure discharge plasmas are recognized as an ideal source for achieving advanced desired requirements of coatings in industry, including surface modification, plasma-polymerized thin films, and engineering of mechanical components. Using nonequilibrium atmospheric pressure

plasmas (APPs) with spraying precursors, recent studies in biology and medical science provided a clear path toward further innovation and improvement in these fields [1–5]. High-pressure discharges have been shown to be extremely useful for industrial applications such as sterilization of medical instruments, biomaterials for surface modifications, tissue engineering, blood coagulation, and therapy of skin disorders. While these applications operate at atmospheric pressure, the impact of small traces of impurities, such as gases, liquid vapors, and dust particles, cannot be ignored during processing because the interplay between these elements and their ramifications is controlled by the characteristics of the carrier gas mixtures [6–12]. Through the combination of liquid precursors in gas mixtures and discharge plasma, it is possible to produce material surfaces with desired properties. A direct injection of liquid precursors into discharge plasma can further complicate the occurring of chemical processes due to the shift from uniphase to two-phase flow. During this activity, two types of monomers were introduced into discharge plasma, such as vapors and liquid aerosols. Using both of these as a catalyst, the surface furnishings of the substrate were altered at a nanometer scale [13, 14].

Experiments perform an important role to test the theoretical concepts and, in some cases, also help to improve a perception with the use of outcomes from both fields. Using numerical models, this chapter addresses distinct characteristics of twophase flow, including mixing, ignition, chemical reactions, and heat transfer in the chamber by exploring various aspects of two-phase flow [15, 16]. By investigating the spray dryers containing flue gases from coal fire power plants and diesel engine fuel, O<sup>0</sup> Rourke and Dukowicz [17, 18] developed a numerical model of a multiphase flow which demonstrates the interaction of gas with liquid droplets. When a droplet comes in contact with the surface, its functionality can be altered in several ways, such as by altering the size, radius, morphology, and temperature of the droplet. Many studies on two-phase flow have been performed that examined both gas mixtures and droplets in a closely coupled plasma-droplet environment involving such factors as geometry, gas and precursor flow rates, temperature, and injection of mono- and polydisperse droplet sizes [19–22]. Droplets play a powerful role in a long discharge plasma channel due to their strong combination of external and internal forces, although sheath formation, around the droplet, has profound effects on how carriers are distributed in the plasma [23, 24].

The direct introduction of liquid precursors into atmospheric pressure plasma jets has proved an effective method for many applications in the past [2, 9, 25, 26], from industrial plasma processes to medical treatments, for instance, plasma-enhanced chemical vapor deposition, sterilization of medical equipment, and direct treatment of skin diseases. In [27], Navier–Stokes and energy equations have been solved to explain droplet deformation, solidification, and energy transfer on the substrate. It might be desirable for smooth plasma surface deposition to evaporate droplets completely in the plasma during transport, and particularly undesirable for partially evaporated droplets to reach substrate surfaces [28]. Surface deposition has numerous physical and chemical properties influenced by a variety of characteristics, especially the precursor and gas flow rates and the composition of the gas mixture, that is primarily responsible for the performance of the coating because a variety of chemical interactions take place during evaporation of droplets as well as particulate nucleation [19]. It has been determined that penning ionization occurs in the presence of trace amounts of nitrogen impurities in helium gas mixtures [29]. Previous research showed that Radio-frequency (RF) plasma sources ranging from 0.1 to 100 kHz produce

*Interaction and Transport of Liquid Droplets in Atmospheric Pressure Plasmas (APPs) DOI: http://dx.doi.org/10.5772/intechopen.105010*

satisfactory and promising results in controlling the relevant features of APP during processing within the last two decades [2, 30–32].

The bulk of the literature deals with gas dynamics in the presence of liquid droplets and the authors in [17, 18] elucidated the behavior of evaporation of droplets as well as mutual interactions between them in the case of small Weber numbers during liquid– gas interaction. In experiments, PlasmaStream™ atmospheric pressure plasma jet deposition systems [14] can be adapted to achieve uniform deposition of coatings on silicon substrate surfaces by adjusting the flow rate of liquid precursors. Inductively coupled plasmas are revitalized by the influence exerted by droplet breakup, desolvation, and coalescence, as demonstrated in [21]. In order to explore the vital characteristics of interaction among liquid droplets and flowing gas in the peculiar applications of combustion, the Eulerian–Lagrangian method is applied, and these features have been verified by the spectroscopic observations involving signal fluctuations in [15, 33–35]. Researchers found that a full packet of liquid droplets dissolved into discharge plasma resulted in more accurate spectroscopic results than fragmentary vaporization of droplets [36]. It has been shown that droplet dynamics are governed by several forces, such as electrical, aerodynamic, surface tension, internal viscous, and gravity during discharge plasma downward drag [23]. According to recent research investigations [19, 24, 29, 43, 56], homogeneous deposition coatings can be achieved by dissolving liquid droplets in atmospheric pressure plasmas at various precursor and gas flow rates.

## **2. Charging of liquid droplets in atmospheric pressure plasma**

To understand the charging of liquid droplets in APP, we consider a fluid model in spherical coordinates in one dimension for a recombination-dominated equilibrium helium discharge plasma at atmospheric pressure. The radius of the sphere is (*r*), and it measures the distance from the symmetry axis. The initial densities of electrons and ions are assumed as *n*<sup>e</sup> (0) and *n*<sup>+</sup> (0), respectively, while the normalized fluid model is able to predict and identify the dynamic interaction among the droplets and atmospheric pressure plasma. Suppose, a spherical liquid droplet of radius (*rd*) (*rd* ≪ Debye length (*λD*)) is immersed in the discharge plasma and the recombination starts occurring on the surface of droplet after charging. As a result, the negative potential is acquired on the surface of droplet due to the higher thermal flux of electrons than ionic species to balance the flow of electrons and ions around the droplet [1, 2]. The growth of this process can be demonstrated in three possible steps (I, II, and III) as exhibited in **Figure 1**. The surface of liquid droplet is neutral before entering into discharge plasma as shown by step I. When the droplet interacts with atmospheric pressure plasma, the electrons rushed toward the droplet due to high thermal speed and formed a layer around the surface of the droplet at step II, which transform a neutral droplet into completely negatively charged surface of the droplet. This shows that the electrons develop a strong veneer on the surface of droplet and ultimately form a sheath around the droplet as expressed in step III for the single isolated droplet in the stationary discharge plasma. For simplicity, we consider direct ionization and dissociative recombination of helium ions and electrons for the numerical modeling.

In equilibrium, *k*i*n*He*n*<sup>e</sup> *k*R*n*e*n+ =* 0, where *k*<sup>i</sup> is the ionization rate of reaction, *k*<sup>R</sup> is the recombination rate of reaction in the helium plasma, and *n*<sup>e</sup> and *n+* are the densities of electrons and ions, while *n*He is the neutral helium gas density. We assume a constant temperature of electrons to sustain the equilibrium in this 1D model,

#### **Figure 1.**

*Development of floating potential on the surface of droplets in APP.*

although this assumption will require a further investigation in the future. Since the discharge plasma is assumed as quasi-neutral (*n*ð Þ <sup>0</sup> <sup>þ</sup> <sup>¼</sup> *<sup>n</sup>*ð Þ <sup>0</sup> *<sup>e</sup>* ), the Boltzmann distribution is considered as suitable to describe the electrons in APP, and mathematically it can be written as *ne* <sup>¼</sup> *<sup>n</sup>*ð Þ <sup>0</sup> <sup>þ</sup> exp *<sup>e</sup><sup>Φ</sup>=kTe* , where *kT*<sup>e</sup> is the electron temperature expressed in eV and Φ is the electrostatic potential. We assumed an equilibrium state where the ionization and recombination rates in the discharge plasma were in balance. Further, the description of one-dimensional normalized fluid model equations for the discharge plasma is available in [37], which are used for the plasma modeling in the current scenario.

#### **2.1 Effect of perturbation**

The occurrence of perturbation is essential after the introduction of nebulizer droplets in the atmospheric pressure plasma, while the role of perturbation can be expressed by using a power series method [38]. We applied a classic technique to explore the behavior of liquid droplet using power series as *<sup>n</sup>* <sup>¼</sup> *<sup>n</sup>* <sup>þ</sup> *<sup>α</sup>*<sup>1</sup>*n*ð Þ<sup>1</sup> <sup>þ</sup> *<sup>α</sup>*<sup>2</sup>*n*ð Þ<sup>2</sup> <sup>þ</sup> … . This approach is applicable for simple cases and becomes more complex at higher orders of magnitude. For lowest order, the density and velocity of discharge species after simplifications can be written as:

$$
\pi^{(0)} \approx 1 + \eta\_0 \frac{e^{-\rho}}{\rho} \tag{1}
$$

$$
v^{(0)} \approx \eta\_0 e^{-\rho} \left(\frac{1}{\rho} + \frac{1}{\rho^2}\right) \tag{2}$$

where *ρ* is the normalized radius of droplet and *n*<sup>0</sup> =1+ *η*0, and *η*<sup>0</sup> is the arbitary constant. The above system of equations is reformulated by letting *τ* � *>* 0 and after simplication, the solution of these equations is discussed around and at the surface of droplet in APP.

When the fluid model equations in normalized form are solved numerically under the specific boundary conditions in the discharge plasma, the parameters, such as floating potential, ionic species flux, the electric field, and ionic species density, are evaluated to describe the characteristics of APP in the presence of water droplets.

The electrons in discharge plasma are responded rapidly than heavy ionic species when the liquid droplet enters from the nebulizer. These electrons immediately

### *Interaction and Transport of Liquid Droplets in Atmospheric Pressure Plasmas (APPs) DOI: http://dx.doi.org/10.5772/intechopen.105010*

interact with the liquid droplet and accumulate as a layer on the surface of droplet due to their higher thermal flux than ions [39, 45]. The APP is positively charged as compared to the deposition of negative charge on the surface of droplet due to the accumulation of electrons. The negative potential on the surface of the droplet is screened because Debye shielding confines the potential variations in this particular small region, called as a thin layer. The formation of a thin layer is termed as a sheath, and it forms a potential barrier to extinguish further spread of negative charge. The distribution of potential is progressed around the droplet when the dimensionless radius of discharge plasma channel alters from 0.01 to 0.1 as shown in **Figure 2(a)**. The sharply rising edge in electric potential around the droplet exhibits that the floating potential is immediately strengthened after entrance into atmospheric pressure plasma because of charging. Eventually, this develops a balance of electronic and ionic fluxes around the droplet, and this is verified by the presence of constant flux in the inner region, when the normalized radius (*ρ*) varies from 0.01 to 0.1 approximately as appeared in **Figure 2(b)**. It exhibits that the dynamic behavior of discharge participating species is evolved in the mentioned domain of normalized radii of droplet from 0.01 to 10. Therefore, it is evident from **Figure 2** that the characteristics of charge carriers are changed in the presence of nebulizer droplets and effectively perform an important role for the charging of the droplet. Based on this analysis, it is concluded that system of fluid model equations becomes simplified in the inner region due to the development of constant flux.

To distinguish the attributes of charge carriers around and on the surface of the droplet, the distributions of drift velocity and electric field are compared as shown in **Figure 3**. Considering the flux balance conditions around the droplet, the strength of magnitudes of drift velocity and electric field is altered from higher to lower values. As the intense charged layer is formed at the surface of the droplet, which is responsible for the organization of distorted electric field at the surface of the droplet. It is apparent from **Figure 3(a)** that the lowering in the drift velocity and electric field is linearly varying from *ρ* = 0.01 to 1, while a sharp fall appears at higher normalized radii *ρ*. This shows that the magnitudes of drift velocity and electric field are prominently reduced to very small values away from the surface of the droplet. This situation is entirely transformed at the surface of droplet, when the velocity of ionic species and strength of electric field are certainly boosted to the higher values as a function of droplet radius as exhibited in **Figure 3(b)**. The comparison of both situations reveals that the tremendous modification occurs at the surface of droplet

**Figure 3.** *(a, b) Spatial distributions of normalized drift velocity and electric field around and at the surface of droplet in APP.*

due to the formation of sheath including the accumulation of electrons. This can be verified from the sharp and steady change in the drift velocity and electric field as viewed from **Figure 3**. The sharp distortion in electric field from 0.0001 to 0.001 exhibits that the sheath developed around the droplet due to the deposition of negative charge on the surface of the droplet.

**Figure 4(a)** shows that the normalized magnitude of the droplet floating potential at the surface continuously changes over time due to the strong perturbations caused near the surface by ions and electrons. It is also evident from the mathematical expressions in Eqs (1) and (2) that the spread of perturbation is uniformly homogeneous around the droplet. To explore the interaction of droplets during the transport in the plasma channel, the surface charge on the droplet is evaluated from the floating potential (*ψ*4*πε*0*rd*) [40] as well as from the Rayleigh method (8*π* ffiffiffiffiffiffiffiffiffiffiffi *ε*0*γr*<sup>3</sup> *d* q ) [41]. Based on **Figure 4(b)**, it can be demonstrated that the surface charge of the droplet is appropriately less than the charge determined by Rayleigh's criterion. In comparison with droplet charging, droplets remain intact during residence time and are not decomposed into smaller particles. Therefore, the scenario of droplet charging can be supported. The characteristics of floating potential are distinguished

*Interaction and Transport of Liquid Droplets in Atmospheric Pressure Plasmas (APPs) DOI: http://dx.doi.org/10.5772/intechopen.105010*

from **Figures 2(a)** and **4(a)** showing the strong behavior around and on the surface of the droplet. This shows that the evaporation is not a very important mechanism in certain domain of droplet radii (>5 μm), and the droplets survive successfully during the passage in the plasma channel, which act as the seed for nonuniform surface depositions. Therefore, the behavior of droplets in helium APP can be described by the spatio-temporal characteristics using the nonequilibrium Langumir-Knudsen law [42] and normalized one-dimensional normalized fluid model [37] in spherical coordinates.

## **3. Multidimensional coupled fluid-droplet model for PlasmaStream system**

Now the multidimensional modeling approach is preferred to explain and interpret the liquid-plasma interaction, while the multiphase flow can be described by the Eulerian–Lagrangian numerical scheme. In this approach, the Lagrangian methods are used to investigate droplet dynamics, and Eulerian methods are used to investigate the kinetics of the He-N2 gas mixture, in addition to major collision interactions initiated by chemical reactions. As a result of adding the coupling terms to the conservation equations of mass, momentum, and energy of the gas and liquid phases [43, 44], the fluid model of the plasma gas and stochastic liquid droplet model of the liquid precursor droplets were integrated including mutual interactions. The following set of equations is used for the modeling of flowing gas mixture as:

$$\frac{\partial \rho\_{\text{g}}}{\partial t\_{1}} + \rho\_{\text{g}} \overrightarrow{\nabla}.\overrightarrow{u}(r, z, t) = \text{S} \tag{3}$$

$$\frac{\partial \left(\rho\_{\mathcal{g}} \overrightarrow{u}\left(r, z, t\right)\right)}{\partial t\_1} = \overrightarrow{\nabla}P + \overrightarrow{\nabla}.\overrightarrow{\Gamma}\_{\mu}\nabla\overrightarrow{u}\left(r, z, t\right) + \text{S}\_{\mathsf{M}}\tag{4}$$

$$\frac{\partial \left(\rho\_{\text{g}} h\right)}{\partial t\_{1}} = \overrightarrow{\nabla} \cdot \frac{k}{C\_{\text{p}}} \overrightarrow{\nabla} h + \overrightarrow{j} \, \overrightarrow{E} + \text{S}\_{\text{E}} \tag{5}$$

$$
\rho\_{\text{g}} \frac{\partial Y}{\partial t\_1} = \overrightarrow{\nabla} . D\_{\text{AB}} \nabla Y + \text{S} \tag{6}
$$

**Table 1** lists the terminology and description of variables for the above set of equations. Eqs (3–6) represent the fluid model equations of mass density, momentum, and energy of the gas mixture, while the models are coupled through source terms (S, SM, and SE). In this case, Implicit Continuous Eulerian (ICE) method is used because this has already been tested for the numerical solution of gas flows of any Mach number [33]. This has also considered as the effective for the calculation of terms containing momentum exchange between two phases. For the presence of species in APP, the kinetics behavior of discharge species (e�, He+ , He2 + , N2 + , He\* , and He2 \* ) are determined by solving the generalized continuity equation as written as follows:

$$\frac{\partial n\_{\rm sp}(r, z, t)}{\partial t\_2} + \overrightarrow{\nabla} \cdot \left( \overrightarrow{\Gamma}\_{\rm sp}(r, z, t) + n\_{\rm sp} \overrightarrow{u}(r, z, t) \right) = \mathbf{S}\_{\rm sp} \tag{7}$$

where the index **sp** corresponds to the distinctive species (electrons, ions, neutrals, and radicals), *Γ* ! spð Þ¼ *<sup>r</sup>*, *<sup>z</sup>*, *<sup>t</sup>* sgn qsp *<sup>μ</sup>*sp*<sup>E</sup>* ! *n*sp � *D*sp∇ ! *n*sp indicates the flux of discharge


*Interaction and Transport of Liquid Droplets in Atmospheric Pressure Plasmas (APPs) DOI: http://dx.doi.org/10.5772/intechopen.105010*


#### **Table 1.**

*Variables in fluid-droplet model (2D and 3D).*

species including drift and diffusion components, μsp and Dsp correspond to mobility and diffusion coefficient of discharge species, respectively, while the Einstein relation (Dsp = μspkBTsp/qsp) is used for the calculation of numeric values. The discharge species density is denoted by *n*sp, and *S*sp exhibits to the source term for the construction and destruction of discharge species based on the chemical reactions as investigated in [45]. The electron mean energy plays a vital role in the transport coefficient and reaction rate coefficient of electron impact reactions that are updated at every time step in the plasma model indicating a strong coupling between the two-phase flow. The electron mean energy (*ε*) is determined by using the electron energy density balance equation as:

$$\frac{\partial n\_{\varepsilon}(r,z,t)}{\partial t\_{2}} + \overrightarrow{\nabla} \cdot \left( \overrightarrow{\Gamma}\_{\varepsilon}(r,z,t) + \varepsilon n\_{\varepsilon} \overrightarrow{u}(r,z,t) \right) = \mathbb{S}\_{\varepsilon} \tag{8}$$

The electron energy density is calculated using *nε*ð Þ¼ *r*, *z*, *t ne*ð Þ *r*, *z*, *t* ̄*ε*ð Þ *r*, *z*, *t* , where *n*<sup>e</sup> exhibits the electron density and *Γ* ! *<sup>ε</sup>* corresponds to the flux. Considering all of the terms in [46] together, the impact of joule heating, electron energy losses, and implicit correction terms is denoted by the energy source term (*S*ε). To solve the particle balance equation and electron energy density equation numerically, the alternating

direction implicit (ADI) solver is used [47]. For the estimation of the electric field (E! ), Poisson's equation is coupled with the plasma fluid model, and it can be expressed as:

$$\overrightarrow{\nabla} . \left( \varepsilon\_0 \overrightarrow{E} (r, z, t) \right) = -\sum\_{\text{sp}} q\_{\text{sp}} n\_{\text{sp}} (r, z, t) \tag{9}$$

where *ε*<sup>0</sup> is the permittivity of the free space, and the space charge density is the product of species charge (*q*sp) and density (*n*sp). From the gradient of the electric potential (�*∂V=∂r*, � *<sup>∂</sup>V=∂z*) in two dimensions, we calculate the electric field strengths along r and z. Here, the numerical solution to Poisson's equation is efficiently achieved by using the successive over relaxation (SOR) method as applied in [48]. The local field approximation is used to calculate the effective electric field for the transport of ions in discharge plasma, in which the transport and reaction rate

coefficients are tied to the electric field [49]. In this numerical model, secondary emission of electrons is not involved because ions have small energies compared to electrons, according to [50].

### **3.1 Geometry of a plasma chamber and experimental chamber**

A range of siloxane monomers including hexamethyldisiloxane (HMDSO), *n*-hexane, tetraethyl orthosilicate (TEOS), and water were used to deposition coating on a PlasmaStream system using the laser diffraction particle size analysis technique as shown in **Figure 5(a)**. For plasma generation, a 7.5-cm-long polytetrafluoroethylene (PTFE) tube with a 1.5-cm inner diameter is used. **Figure 5(b)** shows the schematic geometry for the simulation region used in modeling the experiment, in which an oscillating sinusoidal potential (*V*) is imposed to a thin metal electrode at a specific frequency (*f*), and the substrate is grounded. Experimental investigations are performed on silicon substrates with resistivities ranging from 0 to 100 V cm<sup>1</sup> (450 and 300 mm thick), for the deposition of coatings [19]. Assuming axisymmetry at the inlet boundary, we introduce the laminar profile of gas flow velocity. **Figure 5(b)** shows a pin metal electrode separated by two gas flows (Q1 and Q2 l min<sup>1</sup> ) on the top side of the chamber (AB). The nebulizer droplets are introduced from the central location (B) with a flow rate Q3 μl min<sup>1</sup> and are mixed with APP. At the central channel of APP, the droplets of different types of precursors which are served as discrete parcels are propelled along a stream of He-N2 (99% He and 1% N2) carrier gas until they reach the gaseous state (**Table 2**).

#### **Figure 5.**

*(a) Symbolic diagram of experimental chamber of PlasmaStream™ atmospheric pressure plasma jet deposition system with laser source [19] and (b) schematic geometry for the fluid-droplet model.*

*Interaction and Transport of Liquid Droplets in Atmospheric Pressure Plasmas (APPs) DOI: http://dx.doi.org/10.5772/intechopen.105010*


**Table 2.**

*Fluid-droplet model (2D) using initial and boundary conditions.*

#### **3.2 Stochastic liquid droplet model**

For a single liquid droplet falling vertically under the combined influence of a number of forces, the following equations described in [17] estimate the transient position of the droplet, its velocity, radius, and temperature.

$$\frac{d\overrightarrow{\dot{x}\_p}}{dt\_1} = \overrightarrow{\dot{u}\_p} \tag{10}$$

$$\frac{\mathbf{d}\overrightarrow{u}\_p}{\mathbf{d}t\_1} = \overrightarrow{F} \tag{11}$$

$$\frac{\mathrm{d}r\_p}{\mathrm{d}t\_1} = \left(\frac{\lambda\_\mathrm{g}}{\rho\_\mathrm{l}c\_{\mathrm{p}\_\mathrm{g}}}\right) \left(\frac{\mathrm{Nu}\_\mathrm{g}}{2r\_\mathrm{p}}\right) \left(\frac{Y\_{V\_\mathrm{s}} - Y\_V}{1 - Y\_{V\_\mathrm{s}}}\right) \tag{12}$$

$$\frac{\mathbf{d}T\_p}{\mathbf{d}t\_1} = \left( \left( \frac{3\lambda \mathbf{l} N \mathbf{u}\_1}{2\rho\_l c\_l r\_\mathbf{p}^2} \right) + 3 \frac{\mathbf{d}r\_p / \mathbf{d}t\_1}{r\_\mathbf{p}} \right) \left( T\_\mathbf{s} - T\_\mathbf{p} \right) \tag{13}$$

where the index *p* corresponds to a particular parcel and the details of variables in Eqs (10–13) are available in **Table 1**. In this model, an x-squared distribution (*R r*ð Þ¼ <sup>d</sup> ð Þ 1*=r*<sup>d</sup> *e* �*r*d *<sup>r</sup>*<sup>d</sup> ) of droplets in various parcels is introduced by taking the rectangular computational mesh, where *rd* and *r*<sup>d</sup> are the radius and mean radius of droplets. The index *i* and *j* of each cell increases along the radial and axial directions. The accuracy of the numerical simulations is obtained by mesh independence, and the mesh size is implemented as (*r*, *z*) = (20, 30) in these simulations, which are similar as used in [17, 21]. In the above equations, the dynamic nature of droplets is similar in each parcel, but collisions occur between different parcels. The formula for the evaluation of the collision frequency (νc) is mentioned as follows:

$$
\mu\_{\rm c} = n\_{\rm p\_2} \pi (r\_{\rm col} + r\_{\rm con})^2 \Big| \overrightarrow{u}\_{\rm r} \Big| / V\_{\rm ijk} \tag{14}
$$

In the above Eq. (14), the collector and contributor droplets are shown by *r*col and *r*con, while the relative velocity between parcels is denoted by *u* ! <sup>r</sup> ¼ *u* ! p1 � *u* ! p2 . The number of droplets in the associated parcel is exhibited by *n*p2 , and the volume of the cell containing both parcels is represented by *<sup>V</sup>*ijk. *<sup>P</sup>*<sup>0</sup> <sup>¼</sup> *<sup>e</sup>*�*<sup>n</sup>* is employed to calculate the probability of absent collision. Here, *n* ¼ *νc*Δ*t*<sup>1</sup> is the mean number of parcels and Δ*t*<sup>1</sup> is used the computational time step in the case of Lagrangian approach. The collision and critical impact parameters (*b*, *b*cr) are described below to express the collisions between the droplets by the following relations as:

$$\begin{aligned} b &= (r\_{\text{col}} + r\_{\text{con}}) \sqrt{Y} \\ b\_{\text{cr}} &= (r\_{\text{col}} + r\_{\text{con}}) \sqrt{E\_{\text{coal}}} \end{aligned} \tag{15}$$

where *E*coal is the coalescence efficiency and defined as *Ecoal* ¼ min 1ð Þ *:*0, 2*:*4*ζ γ*ð Þ*=We* . The complex function *ξ* (*γ*), is approximated by the polynomial for simplicity [17] as *ξ* (*γ*) = *γ* 3 *–*2.4 *γ* <sup>2</sup> *- γ* and *γ* = *r*con/*r*col. The parameter (*Y*) is determined by a random number that exists within the interval (0, 1). In case the condition (*b* ≥ *b*cr) is true, then the result of a collision is grazing, which arises within a brief period following an injection of droplet pulse. Collisions caused by grazing occur when the droplets in the parcels and collectors collide so that they preserve their size radii and temperatures, but this collision also changes their velocities. The droplets in the plasma chamber coalesce when (*b* < *b*cr) occurs during downward displacement, leading to the formation of a new droplet with a higher radius. Nevertheless, the main step is to combine stochastic parcel method and ICE technique, in which gas flow rates and droplet velocities are integrated. For the discretization of mathematical equations, it is recommended to use first-order accuracy of finite difference method (FDM). A detailed description of the discretization procedure for these models is given in [17, 46]. Accordingly, **Tables 3–5** contain the description of the fluid-droplet model, including parameters and constants from an online database [51].



#### **Table 3.**

*Fluid-droplet model initialization.*


#### **Table 4.**

*Parameters used in fluid-droplet model (2D).*


#### **Table 5.**

*Density of electrons and vapor species using three-dimensional fluid-droplet model (3D).*

#### **3.3 Initial and boundary conditions**

Assumptions are made regarding laminar profiles of the axial, radial, and swirl velocities (*u*, *v*, *w*) in the PlasmaStream model. For grounded electrodes and solid metal substrates, the flux of electrons, heavy charged particles, neutral gas species, and electron energy density are calculated by considering the modified boundary conditions during the alternate cycles of sinusoidal voltage as mentioned in [52, 53]. The expression *<sup>v</sup>*th,sp <sup>¼</sup> ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 8*kBT*sp*=πm*sp p is used for the evaluation of thermal velocity. The freeware Boltzmann solver (BOLSIG+) [54] was employed for the estimation of electron mobilities as well as reaction rate coefficients of excitation and ionization processes. Moreover, two different time steps are implemented to control the liquid phase of droplets (Δ*t*1) and plasma in the fluid-droplet model (Δ*t*2). Plasma equations can be numerically solved with a small time step constraint imposed on the plasma model. Courant–Friedrichs-Levy (CFL) condition is met when time step is selected for the case plasma model using an inequality, *Δt*<sup>2</sup> < *ε*0*=* P spj*q*spj*μ*sp*n*sp � � [37] as compared to the liquid droplet phase. On the other hand, the coupling source terms are used to

solve the numerical solution of the liquid and gas phases by using the average values of electric field. For the gas and precursor droplets, we take into account small values of swirl velocities, since they have no significant effect on the numerical solution of twodimensional equations. **Table 3** summarizes the operating conditions for the fluiddroplet model as described in the experiment.

## **4. Dynamics of interaction in two-phase flow**

Utilizing a coupled fluid-droplet model, this section investigates the dynamics of interactions between different precursor droplets and APP. By displaying a quantitative understanding of droplet transit in the plasma chamber, we are able to explore the temporal mean profiles of droplet radii and temperatures, droplet count and size distributions, and vapor species density. To establish the authenticity of modeling outcomes, we performed a comparison of simulation results with experimentally measured size of HMDSO droplets using the laser diffraction particle size analysis technique in which the plasma discharge tube was positioned between the laser source and the capture lens as rationalized in detail in [19]. In order to evaluate the effectiveness of the results derived using the fluid-droplet model, the effect of evaporation is compared by taking four liquid precursors into account: HMDSO, n-hexane, TEOS, and water. The time for the evaporation of the entire pulse of droplets is dependent on many factors, while we address an impact of precursor and gas flow rates to describe the characteristics of APP. A similar profile of gas velocity is assumed as provided by the experimental measurements for the numerical simulations [19]. The initial injection velocity of droplets is distinctive in various sections of this chapter.

### **4.1 Behavior of two-phase flow**

Immediately after the liquid droplets interact with discharge plasma, the evaporation of the droplets begins. This occurs due to multiple forces that are simultaneously applied to the liquid droplets, such as electrical (*F* ! ¼ *qE* ! ), aerodynamic drag (*F* ! ¼ *Kp u* ! � *u* !0 � *u* ! *d* � <sup>1</sup>*=ρ<sup>l</sup>* ð Þ<sup>∇</sup> ! *p*), internal viscous (*μv*), surface tension, and gravitational (*F* ! ¼ *m g*! *<sup>z</sup>*) during this interactive process in the PlasmaStream jet deposition system. The effect of interaction between two-phase flow is investigated at different time instants, such as 3, 6, and 9 ms as shown in **Figure 6**. Because the volatile HMDSO liquid droplets are exposed to atmospheric pressure plasmas within discharge plasma pulses, the vaporization rate of liquid droplets increases rapidly from 0 to 3 ms, which can be proved by a sharp peak in vapor species density along the pulse of droplets. In discharge plasma, the impact of droplet vaporization increases the lifetime of remaining droplets. The lifetime of the survived droplets in discharge plasma is increased by the impact of vaporization of droplets that further decelerates the evaporation of the remaining droplets in the plasma chamber. This effect expands spatially with time because the vapor species density (*ρ*g*Y*) increases along the evaporating pulse of droplets as shown in **Figure 6(a–c)**. Due to the higher temperatures of the plasma gas mixture in comparison with the liquid droplets, vaporization of droplets dominates in nonequilibrium APP. Based on the increase in vapor species density, the droplets vaporize fairly rapidly between 0 and 6 ms; however, this process impels to slow down with a reduction in the number of droplets in the APP at 9 ms, as discussed in [43].

*Interaction and Transport of Liquid Droplets in Atmospheric Pressure Plasmas (APPs) DOI: http://dx.doi.org/10.5772/intechopen.105010*

**Figure 6.**

*Top row (a, b, c) shows vapor species density and bottom row (a*<sup>0</sup> *, b*<sup>0</sup> *, c*<sup>0</sup> *) exhibits temperature of the gas at 3, 6, and 9 ms using HMDSO precursor flow rate = 100 μl min*�*<sup>1</sup> , gas flow rate = 5 l min*�*<sup>1</sup> , f = 20 kHz, and V0 =* �*13.5 kV.*

A strong gradient along the pulse of evaporating droplets is evident in the spatial distributions of gas temperature (*T*g) as illustrated in **Figure 6(a**<sup>0</sup> **–c**<sup>0</sup> **)**. Vapor species density is produced as an additional species that is incorporated after droplets transfer into vapor after phase shifting. Due to evaporation, the temperature of the gas surrounding the HMDSO droplets has dropped sharply. As can be seen from the spatial distributions of gas temperatures, the area around the droplets pulse has been significantly affected by the evaporation as they fall downward. In the plasma chamber, as indicated clearly at three distinct points in time (3, 6, and 9 ms), the temperature of gas falls as the droplets evaporate, but the impact of the vaporization increases. Consequently, this can result in a lowered gas temperature which is useful and effective for the sensitive biomedical and industrial applications in different operating conditions of gas flow and precursor flow rates. Observations of gas temperature profiles indicate that distortions, caused by droplet dispersal, contribute significantly to the formation of a cold plasma at atmospheric pressure. Therefore, the diminish in gas temperature along with the increase in species density endorses the turbulent nature of the liquid precursor of the HMDSO in He-N2 atmospheric pressure plasma as shown in **Figure 6**.

## **4.2 Temporal evolution of a pulse of HMDSO droplets**

We analyzed the temporal mean profiles of radius, axis velocity, and temperature of HMDSO droplets in APP in order to understand the role of evaporation in APP. **Figure 7** shows the mean density of the vapor species during the entire lifetime of a pulse of droplets in the plasma chamber. A higher temperature of the APP enables the droplets to gain energy, resulting in evaporation around the surface of the droplets. While both the number of droplets and the temperature of the parcels are unchanged at the start, their radii and temperatures are reduced due to strong evaporation in domain I. Droplets in parcels contract in size due to evaporation, and some of them acquire nanometer-sized domain radii. For the various kinds of precursors in gas mixtures, the precedent for vaporization of droplets has been examined [17], and a similar benchmark is employed when trying to explain the evaporation behavior of droplets in APP. During interaction, a parcel is removed from the parcel distribution if its entire bundle of droplets evaporates. Droplet counts in parcels evaporate continuously, and its overall effect can be observed in the distribution of their mean radii. Due to drop-down of average temperature, the temperature of each droplet amplifies after gaining energy from the discharge plasma; however, due to the strong vaporization of droplets at the start, the rise in temperature of droplets diminishes in each parcel. Thus, the mean temperature of the remaining parcels actively curtails during the initial period of 1 ms. This is illustrated by a red dotted line in **Figure 7** which indicates that the mean concentration of vapor species increases moderately. This occurs because of evaporation of droplets in the domain I rather than being converted to vapor.

In domain II, the droplets start enhancing in temperature as they attain partial steady states in the plasma chamber, as indicated by a black dashed line. Droplet-plasma interactions ultimately result in vapor species being transferred into discharge plasma, and this in turn drives a greater density of He-N2 gas mixture. As a result, when intensive evaporation of droplets occurs, their momentary temperature quenches while in the case of reduced evaporation, it starts to increase, as demonstrated by the green dashed line in **Figure 7** for domains I and II. It is seen at 8 ms as highlighted in the

#### **Figure 7.**

*Mean distributions of droplet radius (solid line), droplet axial velocity (dashed line), droplet temperature (dashdotted line), and mean vapor species density (dotted) of parcels using HMDSO precursors at flow rate = 100 μl min<sup>1</sup> , gas flow rate = 5 l min<sup>1</sup> , f = 20 kHz, and* V*<sup>0</sup> = 13.5 kV.*

*Interaction and Transport of Liquid Droplets in Atmospheric Pressure Plasmas (APPs) DOI: http://dx.doi.org/10.5772/intechopen.105010*

domain II with an arrow that the evaporation accelerates sharply because the number of liquid droplets becomes very small at this stage. Droplets are distributed in a new manner after parcels of droplets are eliminated by evaporation, as indicated by the small spikes in the profile of mean radius of droplets. According to **Figure 7**, it is obvious that the density of the mean vapor species grows until it reaches a stationary state just before the phase shift of the entire pulse of droplets as indicated by the dotted red line.

#### *4.2.1 Effect of HMDSO,* n*-hexane,TEOS, and water*

Through the incorporation of a relevant precursor, the characteristics of surface deposition coating can be adjusted, resulting in a discharge plasma that was altered fundamentally by evaporation and finite-shared interactions between the droplets. We consider three kinds of liquid precursor droplets, for example, HMDSO, *n*-hexane, TEOS, and water to analyze the properties of APP and the size radii of droplets exists in the range from 4 to 24 μm. As a result of strong intermolecular interactions within their molecules, the HMDSO liquid precursor evaporates completely within 4 ms, while the *n*-hexane, TEOS, and water droplets survive for longer. It suggests that HMDSO is more reactive and volatile than *n*-hexane, TEOS, and water under similar operating conditions when interacting with He-N2 discharge plasma. By avoiding the areas near the cathodes and grounded electrodes at the maximum value of discharge current density after 2 ms, the spatial profiles of vapor species and electron density are exhibited in the center of the plasma chamber, as shown in **Figure 8**. In the analysis of the electronic and vapor species distributions, the central section of the plasma chamber is used due to dominant effect of evaporation in APP rather than the area near the electrodes since it has higher density peaks. As demonstrated in **Table 4**, all of these precursors have varying physical and chemical characteristics.

#### **Figure 8.**

*Top row showed the spatial profiles of vapor species density and bottom row illustrated electron density, using (a, a*<sup>0</sup> *) HMDSO, (b, b*<sup>0</sup> *)* n*-hexane (c, c*<sup>0</sup> *) TEOS, and (d, d*<sup>0</sup> *) water at precursor flow rate = 100 μl min*�*<sup>1</sup> , gas flow rate = 5 l min*�*<sup>1</sup> , f = 20 kHz, and* V*<sup>0</sup> =* �*13.5 kV.*

**Figure 8(a–d)** reveals that HMDSO precursor has higher vapor species density than the three other precursors, including *n*-hexane, TEOS, and water because of the high level of evaporation. Several factors influence the rate of evaporation when droplets interact with discharge plasma, such as surface tension, dynamic viscosity, mass density, and boiling point of the precursor. Droplets'surface tension provides a strong resistive force against evaporation, and since HMDSO has a very small surface tension in contrast to other precursors, this implies rapid evaporation of the droplets (as shown in **Table 4**). The temperature of gas mixtures amplifies due to higher value of electron mean energy (*ε* � 0.4 eV) around the pulse of droplets than other discharge species in APP. A convection effect occurs during the active phase of the discharge current pulse in which the precursor droplets can vaporize as a result of the energy gain in the He-N2 gas mixture. This happens due to the impact of electric field and neutral gas density which is evolved by the exchange of energies in two-phase flow through coupling source terms as represented in the group of Eqs (3–6). Since water droplets have a high mass density value, they are pulled downward by gravity to a much greater extent than liquid precursor droplets. According to **Figure 8(a–d)**, the evaporating pulse of water droplets is more likely to pull downward than other precursors. HMDSO and *n*-hexane liquid precursors do not exhibit significant spatial differences along the pulse due to their small differences in mass density. Hence, the vaporization of droplets causes changes in the density of the vapor species and neutral gas, influencing the discharge properties directly.

**Table 6** shows that the numerical values of vapor density for different species diminish from HMDSO to water. Due to a high rate of evaporation, the gradient of electron density along the pulse of droplets in HMDSO and *n*-hexane precursors is relatively larger than that of TEOS and water precursors as shown in **Figure 8(a**<sup>0</sup> **–d**<sup>0</sup> **)**. This develops conductive pathways and can be grasped by HMDSO or *n*-hexane as active channels, rather than TEOS or water. Moreover, water (H2O) is regarded as the best example of hydrogen bonding because it is polar molecule [55]. As well as having intensive intermolecular forces between their molecules, water droplets have a higher surface tension and density. These factors all contribute to the increased lifetime of water droplets in APP and a weak rate of evaporation of water relative to HMDSO, *n*hexane, and TEOS precursors. So, there is a higher probability of smashing of water droplets on the substrate surface due to survival in the plasma and the survived droplets responsible to build a nonuniform surface deposition. Based on the numerical values of vapor species and electron densities for these precursors, it can be seen that the trend is reducing as exhibited in **Figure 8**. Compared to other precursors, the electron density gradient in water droplets is sluggish, which provides an explanation for small vaporization of water droplets in APP. In turn, the simulations contrast emphasizes the importance of tailoring plasma deposition parameters for different liquid monomers in order to obtain a uniform coating.


#### **Table 6.**

*Maximum values of vapor species density and electron density.*

*Interaction and Transport of Liquid Droplets in Atmospheric Pressure Plasmas (APPs) DOI: http://dx.doi.org/10.5772/intechopen.105010*

#### *4.2.2 Impact of precursor flow rates*

In assessing the impact of different flow rates to achieve the desired features of the surface coating deposition, the initial size distribution of the HMDSO droplets is pivotal. In discharge plasma, droplets are introduced with similar size distributions in the domain (4 μm ≤ *r*<sup>d</sup> ≤ 18 μm) at the start. Using the laser diffraction particle size analysis technique as employed by atmospheric pressure plasma jet deposition, similar initial size distributions of droplets in the domain (4 μm ≤ *r*<sup>d</sup> ≤ 18 μm) are introduced in the discharge plasma. According to numerical simulations [56], there is only a small amount of mutual interaction between the droplets at low precursor flow rates (<100 μl min<sup>1</sup> ). However, these interactions are important for altering the structure of discharge plasma at higher precursor flow rates. To demonstrate the significance of collective collision interactions between droplets and explore the effects on the atmospheric pressure plasma distribution, the fluid-droplet model is investigated at 200 and 500 μl min<sup>1</sup> under identical numerical framework.

The droplet size distributions at 0 ms are represented with black bars, whereas the density bars at 200 and 500 μl min<sup>1</sup> are depicted with blue and red bars, respectively, as shown in **Figure 9(a)**. Vaporization of droplets is responsible for the shrinkage of their size domains, along with the events of mutual interactions involving grazing and coalescence, respectively, at 200 μl min<sup>1</sup> at 1 ms. Based on the comparison of size distributions of droplets with blue and red bars, it is apparent that the collision events are substantially lower at 200 than 500 μl min<sup>1</sup> . The droplet radii increment more rapidly as the coalescence of droplets continues to amplify at 500 μl min<sup>1</sup> , and it clear from the size domain (2.5 μm ≤ *r*<sup>d</sup> ≤ 40 μm) of droplet radii relative to the initial size of droplets as shown in **Figure 9(a)** with black bars. Droplets scatter and settle during descent in the plasma chamber as they pass through these collision events gradually diminishing with time.

In the previous research work [56], they observed that the droplets with larger radii are pulled toward the substrate surface by gravity and electric potential that attract the droplets toward substrate surface. The evaporation of droplets continues to squeeze the radii of droplets in the plasma chamber as the droplets attain the steady state during downward drag at 500 μl min<sup>1</sup> . In spite of mutual interactions making a significant contribution to droplet coalescence, the size domain of droplets has

#### **Figure 9.**

*(a) Black bars corresponds to 0 ms and blue as well as red at 1 ms. (b) Distribution of droplets radius at 12 ms using hexamethyldisiloxane using 200 and 500 μl min<sup>1</sup> precursor flow rates, 5 l min<sup>1</sup> gas flow rate, and 20 kHz as frequency (f) and* V*<sup>0</sup> = 13.5 kV.*

become smaller, and it occurs within the range (100 nm ≤ *r*<sup>d</sup> ≤ 18 μm) at 12 ms in both cases as marked in **Figure 9(b)**. The smaller domain of droplet radii is caused by the ejection of higher radii of droplets from the plasma chamber indicating with blue bars at 200 μl min�<sup>1</sup> with a smaller population as well as size domain in contrast to red bars at 500 μl min�<sup>1</sup> as shown in **Figure 9(b)**. A large part of the droplets cross the plasma channel without evaporating completely and the remaining droplets found a decent opportunity to evaporate strongly in APP during downward drag. In the plasma chamber, droplets are found in the nanometer range prior to phase transition to vapors as evidenced by the multiple bar distributions.

## *4.2.3 Spatial distributions of electrons and N2 ions*

It has been observed that many nonthermal APP characteristics are associated with the presence of trace quantity of N2 impurities in the helium gas because these impurities trigger a reduction of the ignition potential by the action of Penning ionization [29, 45]. To determine whether Penning ionization dominates in forming charge carriers along the pulse of droplets, the spatial profiles of charge carriers can provide a better understanding at the peak value of discharge current density at 12 ms to highlight its impact. As compared to the plasma chamber, He+ and He2 <sup>+</sup> ions densities are greater near the momentary cathode electrode during the breakdown phase in contrast to N2 + ions around the pulse of droplets indicating coercive effect of Penning ionization.

**Figure 10(a, a**0 **)**showed the 3D spatial structure of electrons and N2 <sup>+</sup> ion species at the maximum value of discharge current density along the radial and axial axes at 200 and 500 μl min�<sup>1</sup> , respectively. This happens by the evaporation of droplets around the pulse. As depicted in **Figure 10**, the mass flow of liquid precursor in discharge plasma at 500 μl min�<sup>1</sup> is greater than 200 μl min�<sup>1</sup> because of broader effective area of evaporation at 500 μl min�<sup>1</sup> . The density of electrons surrounding the droplets is slightly greater than the density of N2 <sup>+</sup> ions, implying that electrons generate through multiple channels, such as direct ionization, Penning ionization, and stepwise ionization. Unlike N2 <sup>+</sup> ions, He and He2 <sup>+</sup> ions densities are smaller around the pulse of droplets in the plasma chamber which highlights an impact of Penning ionization. A uniform layer is formed on the substrate surface if the evaporation effect is spread evenly along the radial axis. There is a higher probability at 500 μl min�<sup>1</sup> because of the formation of uniform sheet of electrons along the radial axis as opposed to 200 μl min�<sup>1</sup> . According to the spatial distribution of charge carriers, a faster precursor flow rate results in greater density values at 500 μl min�<sup>1</sup> , underlining why higher precursor flow rates result in higher density values. This discussion clarifies that different precursor flow rates in APP influence the kinetics of charge carriers.

#### *4.2.4 Impact of gas flow rates*

A classification of the behavior of complex interactions between liquid droplets and APP based on similar initial size distributions of droplets in the domain is made by considering different gas flow rates with the domain's initial size distributions (5 μm ≤ *r*<sup>d</sup> ≤ 24 μm). At the inlet boundary of the PlasmaStream system, the gas flow is laminar, but the time it takes before droplets acquire settling velocity differs with gas flow rates during the downward fall. In APP, the properties of droplets and charge carriers are dynamically modified at different gas flow rates, such as 5, 10, and 20 l min�<sup>1</sup> that are exhibited with the spatial distribution of vapor species and electron densities as shown in **Figure 11**. Evaporation length is longer at 20 l min�<sup>1</sup> than at 5

*Interaction and Transport of Liquid Droplets in Atmospheric Pressure Plasmas (APPs) DOI: http://dx.doi.org/10.5772/intechopen.105010*

#### **Figure 10.**

*3D profiles of electron density (a, b) and N2 <sup>+</sup> ion density (a*<sup>0</sup> *, b*<sup>0</sup> *) for HMDSO precursor droplets at flow rates of 200 and 500 μl min*�*<sup>1</sup> , gas flow rate = 5 l min*�*<sup>1</sup> , f = 20 kHz, and* V*<sup>0</sup> =* �*13.5 kV.*

and 10 l min�<sup>1</sup> gas flow rates. With greater gas flow rates, this amplification develops because evaporation and convection of gas mixtures occur in the plasma chamber. It takes around 2 ms for an entire pulse of droplets to evaporate at a gas flow rate of 20 l min�<sup>1</sup> , which is smaller than the gas flow rates at 5 and 10 l min�<sup>1</sup> . The spatial profiles of the densities of the vapor species in **Figure 11(a–c)** demonstrate how the rate of cooling shifts quickly toward the exit axis when the gas flow rate increases.

In order to determine the effect of different gas flow rates on electron distribution during droplet-plasma interactions, **Figure 11(a**<sup>0</sup> **–c**<sup>0</sup> **)** exhibits the distribution of electrons during the droplet-plasma interaction. In plasma channel, the evaporation rate is greater locally when droplets travel at a low gas flow rate, such as 5 l min�<sup>1</sup> . When the surrounding region of droplets is saturated due to evaporation, the remaining droplets in the pulse evaporate at a lower rate due to the impact of this saturated environment. As time passes in discharge plasma, the effect of evaporation spreads across it, increasing the rate of evaporation again. This process is relatively slow at 5 l min�<sup>1</sup> as compared to 10 l min�<sup>1</sup> and 20 l min�<sup>1</sup> . Since gas has a short residence time, its convection is agile at higher gas flow rates, causing a sharp evaporation of droplets. This means that there are smaller chances of saturation due to higher gas flow rates, such as 20 l min�<sup>1</sup> which can be confirmed from the increment in the density of electrons from �3.50 � 1010 cm�<sup>3</sup> to 4.50 � 1010 cm�<sup>3</sup> , when the gas flow rate varies from 5 to 20 l min�<sup>1</sup> . Due to the intense gravitational pull, evaporation of large radius drops in the

#### **Figure 11.**

*Distributions of (a, b, c) vapor species density and (a*<sup>0</sup> *, b*<sup>0</sup> *, c*<sup>0</sup> *) electron density at precursor (HMDSO) flow rate at 100 μl min*�*<sup>1</sup> using gas flow rates = 5 l min*�*<sup>1</sup> , 10 l min*�*<sup>1</sup> , and 20 l min*�*<sup>1</sup> , f = 20 kHz, and* V*<sup>0</sup> =* �*13.5 kV.*

plasma chamber occurs within a short period of time, which causes their journey to be accelerated. However, the small duration of evaporation at 20 l min�<sup>1</sup> indicates the fast conversion of liquid material into vapor phase. According to these simulation outcomes of vapor species and electron densities, it is clearly observed that the characteristics of surface deposition can be readjusted by the control of the gas flow rates.

#### *4.2.5 Effect of gas flow rates on gas temperature and electron mean energy*

By applying similar conditions, the spatial patterns of gas temperature and electron mean energy are investigated at distinctive gas flow rates. While falling, the droplets reach a stationary state, indicating that the initial velocity is most important for mutual interactions, rather than after settling in the plasma chamber. In the process of transport of two-phase flow, the local mass flux ratio (*ρ*d*u*d/*ρ*gas*u*gas) alters until the droplets are all converted into vapor phase. This emerges as the reduction of gas temperature in the vicinity of droplets pulse by the evaporation process. **Figure 12(a–c)** reveals that as the rate of gas flow amplifies, the perturbation caused by the droplets in the mixture enhances. A change in the rate of gas flow can modify the interaction between the droplets and discharge plasma; however, the presence of impurities in the gas in two-phase flow can also play a significant role at atmospheric pressure. Impurities such as nitrogen (N2) and oxygen (O2) can

*Interaction and Transport of Liquid Droplets in Atmospheric Pressure Plasmas (APPs) DOI: http://dx.doi.org/10.5772/intechopen.105010*

#### **Figure 12.**

*Contour distributions of (a, b, c) temperature of gas and (a*<sup>0</sup> *, b*<sup>0</sup> *, c*<sup>0</sup> *) electron mean energy using HMDSO droplets at different gas flow rates = 5*, *10, and 20 l min*�*<sup>1</sup> , precursor flow rate = 100 μl min*�*<sup>1</sup> , f = 20 kHz, and* V*<sup>0</sup> =* �*13*.*5 kV in He-N2.*

enhance the rate of ionization because they can lead to an increase in the vaporization rate of droplets. The gas mixture ingests heat energy from the discharge plasma, and thereafter, it results as a substantial improvement in the evaporation of droplets in APP. As a result, the temperature of the gas falls along with the pulses of droplets from 5 to 20 l min�<sup>1</sup> and the duration of evaporation is increased with an increment in the gas flow rate.

The temperature of electrons is calculated by the numerical solution of electron energy density equation considering the elastic and inelastic collision energy losses as well as joule heating as employed in [43, 45]. The maximum values of discharge current density during alternate cycles of discharge current pulses are observed near the thin cathode electrode and grounded substrate when the electron mean energy reaches the peak, although the prime focus lies in this case is to gain an understanding for the dynamic updates in the structure of electron mean energy along the pulse of droplets. The regions adjacent to the cathode and grounded substrate are excluded in this scenario in order to investigate the effects of evaporation on the electrons in the bulk of APP. The mean energy of electrons is reduced near the evaporating pulse of HMDSO droplets, as demonstrated by the spatial distributions in **Figure 12(a**<sup>0</sup> **–c**<sup>0</sup> **)**. Because, the spatial spread of electron mean energy is wider at 20 l min�<sup>1</sup> than at 5 and 10 l min�<sup>1</sup> , resulting a quicker cooling transfer downward in the plasma chamber at a higher gas flow rate. This demonstrates that the distribution of discharge plasma changes throughout droplet transit and evaporation. The spatial structures of gas

temperature and electron mean energy show that when the gas flow rate increases from 5 to 20 l min<sup>1</sup> , the length of cooling increases in the plasma chamber.

## *4.2.6 Significance of He and He-N2 gases*

**Figure 13(a)** exhibited the average distribution of generation rates of stepwise ionization in the pure He and He-N2 gases. The main sources of ionization are direct and stepwise mechanisms in He gas, while Penning ionization is more important in the He-N2 gas mixture than other generation rates of ionization along the pulse of droplets. The stepwise ionization process is illustrated as a dotted line in this case and is responsible for ionization in He gas due to the abundance of a high density of metastables. The numeric value of the rate of stepwise ionization is 2.0 1014 cm<sup>3</sup> <sup>s</sup> <sup>1</sup> in pure helium gas, and it curtails to 1.0 <sup>10</sup><sup>12</sup> cm<sup>3</sup> <sup>s</sup> <sup>1</sup> in He-N2 gas mixture. **Figure 13(b)** shows the rate of net ionization in atmospheric pressure plasmas, which is the outcome of direct and Penning ionization processes. Through Penning ionization, He-N2 gas mixture experiences a faster rate of metastable destruction. It eliminates the vast majority of metastables, increasing the rate of ionization in large parts of APP. As a result, the density of metastables is reduced to a small amount in He-N2 gas. The net ionization rate in He-N2 discharge plasma is imperatively greater than pure He gas, while its numeric values 2.0 1015 cm<sup>3</sup> <sup>s</sup> 1 in He-N2 diminishe to 1.0 1012 cm<sup>3</sup> <sup>s</sup> <sup>1</sup> in pure He gas as shown in **Figure 13(a, b)**. The imbalance in the rate of net ionization is due to Penning ionization caused by the presence of trace quantity of nitrogen impurities, and this is highlighted in **Figure 13(b)** by a solid line. Droplets in the plasma chamber vaporize rapidly due to the impact of net ionization in He-N2 gas mixture, and therefore their lifetime is reduced. Based on the above outcomes, it is clearly observed that the rate of evaporation is amplified in He-N2 gas mixture as compared to pure helium gas, which is consistent with the previous [19] simulation modeling results.

To explore the behavior of plasma species, **Figure 14** showed the line-averaged distributions electrons, He2 <sup>+</sup> ions, N2 <sup>+</sup> ions, and metastables (He<sup>∗</sup> ) species density over various cycles in the pure helium using solid black lines plus symbols, while showing the red and green lines plus symbols in He-N2 gas mixture. These distributions are obtained by neglecting the areas near the cathode and grounded substrate to explore the performance of droplet-plasma interaction during transport in the plasma

#### **Figure 13.**

*Distribution of averaged generation rates of (a) stepwise ionization and (b) net ionization in the pure He and He-N2 gases using gas flow rate = 10 l min<sup>1</sup> and HMDSO precursor flow rate = 100 μl min<sup>1</sup> , f = 20 kHz, and* V*<sup>0</sup> = 13*.*5 kV.*

#### *Interaction and Transport of Liquid Droplets in Atmospheric Pressure Plasmas (APPs) DOI: http://dx.doi.org/10.5772/intechopen.105010*

chamber. The direct excitation is diagnosed as the central mechanism in the pure helium and He-N2 gas mixtures, but the relevant contribution of metastables becomes divergent in both gas mixtures. As can be seen from the line distribution of species density in **Figure 14(a, b)**, electrons and molecular helium ions are responsible for keeping quasi-neutrality in pure helium gas. On the other hand, the presence of nitrogen impurity molecules in He-N2 gas changes the situation, resulting in the existence of N2 <sup>+</sup> ions and electrons in APP, as shown in **Figure 14(a**<sup>0</sup> **, d**<sup>0</sup> **)**. This clearly shows that the Penning process is the primary ionization mechanism for supplying electrons along the pulse of droplets in He-N2 discharge plasma. The line-averaged behavior of electrons suggests that the ionization rate is amplified by the evaporation of droplets because of the high impact of Penning ionization than direct and stepwise ionization rates. The temporal distribution of metastables shows that they have a higher density in pure helium than in He-N2 gas mixture, as seen by the black line with hollow triangle and red line plus solid triangles in **Figure 14(c, c**<sup>0</sup> **)**. This occurs due to higher destruction rate of metastables through Penning ionization as compared to stepwise ionization. Therefore, the above discussion showed that the major role of interaction between the droplets and plasma depends on the chemical reactions that are happening during two-phase flow.

Under comparable operating conditions, **Figure 15** shows a clear distinction in the spatial distributions of electron density in pure helium and He-N2 gas mixtures. At the highest value of discharge current density, the contour spatial distributions of electron density are exhibited in both gas mixtures. In the plasma chamber, the distributions of electron density are shown in **Figure 15(a, b)**, whereas the bottom row depicts the distribution of HMDSO droplets throughout the pulse as illustrated in **Figure 15(a**<sup>0</sup> **, b**0 **)**, neglecting the thin cathode electrode and grounded substrate. The electron dynamics along the pulse of evaporating droplets are made clearer using this method. The electron density is significantly greater near the thin cathode electrode than along the pulse of droplets in the case of pure He, but its magnitude falls from 1.91 � 1011 to 1.22 � 1010 cm�<sup>3</sup> as shown in **Figure 15(a, a**<sup>0</sup> **)**. The similar trend in the contraction of electron density is observed in He-N2 gas mixture in which it alters from 3.15 � 1011 to 5.55 � 1010 cm�<sup>3</sup> as displayed in **Figure 15(b, b**<sup>0</sup> **)**. It is clear from the shift in the magnitudes of the electron density that its value (2.595 � 1011 cm�<sup>3</sup> ) in He-N2 mixture

#### **Figure 14.**

*Comparison of line-averaged species density (electrons, He2 <sup>+</sup> ions, N2 <sup>+</sup> ions, and He ) in the pure helium and He-N2 gas mixtures using HMDSO droplets at precursor flow rate = 100 μl min*�*<sup>1</sup> , gas flow rate = 10 l min*�*<sup>1</sup> , f = 20 kHz, and* V*<sup>0</sup> =* �*13*.*5 kV.*

is greater than the value (1.788 � 1011 cm�<sup>3</sup> ) in the pure helium gas. This suggests that the rapid rate of evaporation is accelerated by the significant ionization effect in the He-N2 gas mixture, which accelerates the desolvation of droplets in the APP. As a result, this creates a feasible environment in discharge plasma to achieve the desired uniform deposition coating.

According to simulation outcomes, it has been rarely identified the decomposition of HMDSO precursor droplets at small precursor flow rates (10–100 μl min�<sup>1</sup> ), which can be confirmed by considering the criterion of Rayleigh limit. In APP, the magnitude of surface charge on the droplets in the mentioned size domain of radii is less than Rayleigh limit as discussed in [24]. The area of ionization around the pulse of droplets is large in He-N2 gas, while it squeezes in pure helium gas combination, according to the spatial profiles of electron density. This means that the composition

*Profiles of electron density in the plasma chamber (a, b), electron density around the pulse of HMDSO droplets (a*<sup>0</sup> *, b*<sup>0</sup> *) in the pure He and He-N2 gases at gas flow rate = 10 l min*�*<sup>1</sup> , HMDSO precursor flow rate = 100 μl min*�*<sup>1</sup> , f = 20 kHz, and* V*<sup>0</sup> =* �*13*.*5 kV.*

### *Interaction and Transport of Liquid Droplets in Atmospheric Pressure Plasmas (APPs) DOI: http://dx.doi.org/10.5772/intechopen.105010*

of the operating gas mixture affects droplet evaporation. If the gas mixture contains small amount of impurities that are capable of modifying the properties of discharge plasma, and ultimately, the evaporation and mutual interactions of droplets are effected. In APP, the droplets acted as the perturbing agent at the beginning; however, they became part of the operating gas mixture after phase transformation later as vapors. The entire set of interactions is glued through the coupling source terms providing a clear description of heat transfer in two-phase flow. The characteristics of cold plasma can be manipulated by the control of precursor and gas flow rates, because the rate of evaporation inflates with an increment of precursor and gas flow rates. This kind of cold APP is identified as suitable for the medical and industrial applications [6, 9, 25, 26]. In case of incomplete evaporation of droplets, they are not considered as appropriate for the coating deposition applications due to the presence of droplets in discharge plasma. As can be seen from the above contrast, N2 impurities are extremely helpful in enhancing ionization activities during transport of carrier of gas in the plasma chamber, which therefore enhances vaporization of droplets, resulting in a homogenous plasma environment appropriate for surface coatings.

#### *4.2.7 Comparison of model and experimental observations*

To verify the numerical simulation results, it is critical to do a comparison with experimental measurements using a separate setup of laser diffraction particle size analysis technique with the PlasmaStream atmospheric pressure jet deposition system. At location B in the PlasmaStream system in **Figure 5(b)**, a pulse of HMDSO precursor droplets is injected within the size range of 1 μm ≤ *r*<sup>d</sup> ≤ 6 μm in APP. At small precursor flow rates (≤100 μl min<sup>1</sup> ) [56], the mutual interactions between HMDSO droplets, such as grazing and coalescence, are minimal. The fluid-droplet model developed in this chapter's simulation study had similar experimental settings as those listed in **Table 3**, and the initial injection velocity of droplets was assumed to be the same as 1.5 103 cm s<sup>1</sup> in this case. The radii of droplets are reduced steadily by the main contribution of evaporation on the surface during downward fall in the plasma chamber.

The initial distribution of droplets exists in the range (≥1 μm) before introducing into plasma chamber, and the size radii of droplets continuously contract due to evaporation as clearly observed by the presence of droplets within nanometer range (100 nm ≤ *r*<sup>d</sup> ≤ 900 nm) as shown in **Figure 16(a)** at 7 ms. This demonstrates that the HMDSO liquid precursor in He-N2 discharge plasma is highly volatile. The experimental size distributions are recorded in two different experiments in the plasma chamber as shown in different colors in **Figure 16(b)**, but the size domain of droplet radii is similar in both cases. It is evident from the size domain of droplets as mentioned in dotted curly brackets in **Figure 16(a, b)** that the entire bunch of droplets lies in the domain (500 nm ≤ *r*<sup>d</sup> ≤ 5 μm) at 7 ms, which exhibits a good agreement between the coupled fluid-droplet model and experimental size distributions. In the experimental measurements, the split at 0.5 μm is a result of the limiting resolution of the laser diffraction imaging lens arrangement. However, as seen in the size distribution of droplets, the fluid-droplet model also offers information regarding the occurrence of minimum feasible radii of droplets in the nanoscale range as highlighted in **Figure 16(a)**. The foregoing comparison clearly demonstrates that under similar operational constraints, the numerical model and experimental measurements are synchronized nicely. As a result, the similarity of results increases confidence that the numerical simulations utilizing the 2D-coupled fluid-droplet model can accurately describe the complicated interaction between two-phase flow.

**Figure 16.**

*Comparison of (a) fluid-droplet model and (b) experimental observations of size distribution of droplets using HMDSO precursor droplets at flow rate = 100 μl min*�*<sup>1</sup> and gas flow rate = 5 l min*�*<sup>1</sup> , f = 20 kHz, and* V*<sup>0</sup> =* �*13.5 kV.*

#### *4.2.8 3D profiles of droplet-plasma interaction*

The dynamic characteristics of evaporation utilizing TEOS and HMDSO precursors under the same limitations are described in this section using three-dimensional profiles species density in APP. The validity and legitimacy of multidimensional (2D and 3D) numerical modeling outcomes have already been described in [43] by comparing with experimental measurements. **Figure 17** shows the iso-contours of electrons (red) and vapor species density (green) at three consecutive time instants (*t*<sup>1</sup> = 0.6, *t*<sup>2</sup> = 1.6, and *t*<sup>3</sup> = 2.6 ms) during the evaporating pulse of two distinct precursor droplets (TEOS and HMDSO). In the case of both precursors, the iso-contours of vapor species density revealed that their volumes grow with time; however, in the discharge plasma, TEOS expands in volume faster due to higher evaporation of droplets than HMDSO. Because the mass density of TEOS and HMDSO droplets differs, the bunch of TEOS droplets experiences a stronger gravitational attraction than the HMDSO droplets, and eventually, it manifests itself as a disparity in their volumetric spread as clearly contrasted in the top and bottom rows of **Figure 17**. As shown by the blue dashed line, the affected volume along the pulse of droplets extends and pulls downward. On the other hand, the rate of vaporization of HMDSO droplets is greater than TEOS in He-N2 discharge plasma because of its less surface tension and boiling point as compared to TEOS. This is supported by **Table 5** that showed greater values of vapor species density for HMDSO than TEOS, while these features are in accordance with the previous results discussed in the previous subsections using two-dimensional-coupled fluid-droplet model.

At t1, the conducting channel is forming an intensive ionization along the pulse of both precursor droplets, as shown in **Figure 17(a, a**<sup>0</sup> **)**. At *t*2, **Figure 17(b, b**<sup>0</sup> **)** exhibits how the overlapped density layers of electrons (red) and vapor species density (blue) are expanding, with a growth of vapor species continuing until the full pulse of droplets dissolves in the discharge plasma. In terms of electron kinetics, they are present throughout the plasma chamber at the highest value of negative discharge current density, as illustrated in **Figure 17** by the red isocontours. This demonstrates a progressive increase in electron density in the plasma chamber. The numerical values listed in **Table 5** indicate that the electron

*Interaction and Transport of Liquid Droplets in Atmospheric Pressure Plasmas (APPs) DOI: http://dx.doi.org/10.5772/intechopen.105010*

#### **Figure 17.**

*Spatio-temporal iso-contours of electrons (red) and vapor species density (green) using TEOS (top row, a, b, c) and HMDSO (bottom row, a*<sup>0</sup> *, b*<sup>0</sup> *, c*<sup>0</sup> *) at t1, t2, and t3 ms at precursor flow rate = 100 μl min*�*<sup>1</sup> , gas flow rate = 5 l min*�*<sup>1</sup> , f = 20 kHz, and V0 =* �*13.5 kV [44].*

density around the pulse of HMDSO droplets is higher than TEOS. HMDSO has a smaller effect on the volume of vapor species compared with TESO due to the robust evaporation rate in APP. Additionally, the numerical simulation outcomes highlighted that the duration of evaporation for the entire pulse of TEOS droplets is almost twice as long as its counterpart for HMDSO droplets. Thus, TEOS droplets scatter more than HMDSO droplets causing the volume of vapor species to increase. It is evident that these observations differ sharply from those of the top and bottom rows as shown in **Figure 17**. Following the discussion above, it is evident that liquid droplets respond very differently to discharge plasma, depending on the type of precursor. As a result, the numerical modeling outcomes are deemed to be adequate for describing a good understanding of complex interaction of droplets with plasma in APP. These quasi-volumetric characteristics suggest that the internal structure of discharge plasma in two-phase flow is highly complicated, which may be further rationalized by considering a complex chemistry set comprising chemical reactions between vapors and discharge species.

## **5. Conclusion**

In this chapter, we discussed how droplet-plasma interaction is initiated in the atmospheric pressure plasmas. For describing the properties of plasma and explain droplet transport and interaction with the plasma, the one-dimensional normalized fluid model is developed considering Boltzmann distribution of electrons around the droplets. Based on the simulation outcomes, it is concluded that the evaporation and charging of droplets are considered as the most dominant mechanisms within the plasma channel. After developing and testing multidimensional fluid models (2D and 3D) in two-phase flow under distinct precursor and gas flow rates, we examined the implications of complicated interaction between the droplets and plasmas. The characteristics of droplet-plasma interaction manifest that the evaporation of droplets is noticed as the prime mechanism, which has been verified by using different types of liquid precursors (HMDSO, n-hexane, TEOS, and water). A deeper observation at the mean profiles of droplet radii showed that the mutual interactions between the droplets are a major factor to modify the structure of discharge plasma in the limit of higher precursor flow rates (≥200 μl min<sup>1</sup> ). Additionally, the spatial and temporal distributions of droplets and APP are explored to understand the interactive behavior of various precursors, showing that evaporation is remarkably intensive in the case of HMDSO and n-hexane compared to TEOS and water. In the plasma chamber, the convection of cooling is much shifted downward at 20 l min<sup>1</sup> compared with smaller gas flow rates, justifying the significance of the evaporation process at different gas flow rates (5, 10, and 20 l min<sup>1</sup> ). To establish the validity of the numerical simulation results, we compared the outcomes from the 2D fluid-droplet model with experimental observations. Finally, a 3D view of droplet-plasma interaction is presented using spatio-temporal iso-contours of electrons and vapor species density using HMDSO and TEOS to show the volumetric spread in APP.

## **Acknowledgements**

This work is supported by Science Foundation Ireland under grant 08/SRC/I1411.

## **Author details**

Muhammad M. Iqbal1 \* and Mark M. Turner<sup>2</sup> \*

1 CCT College Dublin, Dublin, Ireland

2 School of Physical Sciences, Dublin City University, Dublin, Ireland

\*Address all correspondence to: miqbal@cct.ie and miles.Turner@dcu.ie

© 2022 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

*Interaction and Transport of Liquid Droplets in Atmospheric Pressure Plasmas (APPs) DOI: http://dx.doi.org/10.5772/intechopen.105010*

## **References**

[1] Fridman A. Plasma Chemistry. New York, ISBN: 13 978–0–521-84735-3 Hardback: Cambridge University Press; 2008

[2] Becker KH, Kogelschatz U, Schoenbach KH, Barker RJ. Non-Equilibrium Air Plasmas, Series in Plasma Physics. Bristol and Philadelphia, ISBN: 0 7503 0962 8: IOP Publishing; 2005

[3] Reece Roth J. Industrial Plasma Engineering. Vol. 2. Bristol and Philadelphia, ISBN: 0 7503 0544 4: Institute of Physics Publishing; 2001

[4] Sanchez-Estrada FS, Qiu H, Timmons RB. Molecular tailoring of surfaces via RF pulsed plasma polymerizations: biochemical and other applications. IEEE Conference Record - Abstracts. 2002 IEEE International Conference on Plasma Science (Cat. No.02CH37340). 2002. p. 254. DOI: 10.1109/PLASMA. 2002.1030530

[5] Kolb JF, Mohamed A-AH, Price RO, Swanson RJ, Bowman A, Chiavarini RL, et al., Cold atmospheric pressure air plasma jet for medical applications. Applied Physics Letters. 2008;**92**: 241501

[6] Laroussi M. Nonthermal decontamination of biological media by atmospheric-pressure plasmas: review, analysis, and prospects, in IEEE Transactions on Plasma Science. Aug 2002;**30**(4):1409-1415. DOI: 10.1109/TPS.2002.804220

[7] Stoffels E, Kieft IE, Sladek RE J. Superficial treatment of mammalian cells using plasma needle. Journal of Physics Department. 2003;**36**:2908– 2913. DOI: 10.1088/0022-3727/36/23/007 [8] Pierre-Luc G-L, Mwale F, Iordanova M, Demers C, Desjardins P, Wertheimer MR. Atmospheric Pressure Deposition of Micropatterned Nitrogen-Rich Plasma-Polymer Films for Tissue Engineering. Volume 2, Issue 3, Special Issue: Highlights from the Ninth International Symposium on High-Pressure, Low-Temperature Plasma Chemistry 2004 in Padua, Italy. March 31, 2005, pp. 263-270

[9] Fridman G, Peddinghaus M, Ayan H, Fridman A, Balasubramanian M, Gutsol A, et al., Blood coagulation and living tissue sterilization by floating electrode dielectric barrier discharge in air. Plasma Chemistry and Plasma Processing. 2006; **26**:425-442. DOI: 10.1007/s11090-006- 9024-4

[10] Yu Babaeva N, Kushner MJ, Intracellular electric fields produced by dielectric barrier discharge treatment of skin. Journal of Physics Department. Applid Physics. 2010;**43**:185206

[11] Massines F, Rabehi A, Decomps P, Gadri RB, Segur P, Mayoux C. Experimental and theoretical study of a glow discharge at atmospheric pressure controlled by dielectric barrier. Journal of Applied Physics. 1998;**83**:2950. DOI: 10.1063/1.367051

[12] Wang Q, Doll F, Donnelly VM, Economou DJ, Sadeghi N, Franz GF. Experimental and theoretical study of the effect of gas flow on gas temperature in an atmospheric pressure microplasma. Journal of Physics Department. Applied Physics. 2007;**40**:4202

[13] Babayan SE, Beach H, Selwyn GS, Fe S, Hicks RF. Patent No: US 6,194,036 B1, Feb. 27, 2001, Published in Official Gazette of the United States Patent and

Trademark Office. Patents February 27, 2001;**1243**(4). Published weekly by authority of congress. Available from: https://books.google.ie/books?id= xg4ETKL3IzYC&printsec= frontcover#v=onepage&q&f=false; https://patentimages.storage.googleapis. com/e4/6a/d4/d8279b5f5dcb8b/ US6194036.pdf

[14] Anthony P, Herbert F, O'Neill L, Jaroszynska-Wolinska J, Stallard C, Ramamoorthy A, et al. A Comparison between Gas and Atomized Liquid Precursor States in the Deposition of Functional Coatings by Pin Corona Plasma. 22 Mar 2011;**8**(3):230-238. DOI: 10.1002/ppap.201000119

[15] Amsden AA, O'Rourke PJ, Butler TD. KIVA-II: A computer program for chemically reactive flows with sprays. United States: Np. 1989. DOI: 10.2172/ 6228444. Available from: https://www. osti.gov/biblio/6228444-kiva-iicomputer-program-chemically-reactiveflows-sprays

[16] Fauchais P, Vardelle A, Vardelle M. Modelling of plasma spraying of ceramic coatings at atmospheric pressure. Ceramics International. 1991:**17**(6):367-379

[17] O'Rourke PJ. Collective Drop Effects in Vaporizing Liquid Sprays. Ph.D Thesis 1532-T, Princeton University, Department of Mechanical and Aerospace Engineering. 1981

[18] Dukowicz JK. A particle-fluid numerical model for liquid sprays. Journal of Computational Physics. Apr 1980;**35**(2):229-253. DOI: 10.1016/0021- 9991(80)90087-X

[19] Stallard CP, Iqbal MM, Turner MM, Dowling DP. Investigation of the Formation Mechanism of Aligned Nano-Structured Siloxane Coatings Deposited Using an Atmospheric Plasma Jet,

Plasma Processes and Polys. Oct 2013;**10** (10):888-903. DOI: 10.1002/ ppap.201300056

[20] Yoshida T, Akashi K. Particle heating in a radio-frequency plasma torch. Journal of Applied Physics. 1977; **48**(6):2252. DOI: 10.1063/1.324036

[21] Shan Y, Mostaghimi J. Numerical simulation of aerosol droplets desolvation in a radio frequency inductively coupled plasma, Spectrochimica Acta Part B: Atomic Spectroscopy. 21 November 2003;**58**(11): 1959-1977. DOI: 10.1016/j. sab.2003.09.003

[22] Clayton TC, Schwarzkopf JD, Sommerfeld M, Tsuji Y. Multiphase Flows with Droplets and Particles. Second Edition, ISBN: 978-1-4398-4050- 4. Boca Raton: CRC Press, Taylor & Francis Group, NW; 2011

[23] Leiby CC, Oskam HJ. Volume Forces in Plasmas. The Physics of Fluids. 1967; **10**:1992. DOI: 10.1063/1.1762397

[24] Iqbal MM, Turner MM, Liquid Droplets in Atmospheric Pressure Plasmas. Proceeding in 63rd Annual GEC and 7th ICRP. Paris: France. Oct 2010;**55**: 7

[25] Thomas M, Mittal KL. Atmospheric Pressure Plasma Treatment of Polymers: Relevance to Adhesion. Hoboken, New Jersey: John Wiley & Sons, Inc. 2013. ISBN 978-1-118-59621-0

[26] Nelea V, Li Luo CN, Demers J, Antoniou A, Petit S, Lerouge MRW, et al. Selective inhibition of type X collagen expression in human mesenchymal stem cell differentiation on polymer substrates surface-modified by glow discharge plasma, Volume 75A, Issue 1, Pages 216-223, 1 October 2005. DOI: 10.1002/jbm.a.30402

### *Interaction and Transport of Liquid Droplets in Atmospheric Pressure Plasmas (APPs) DOI: http://dx.doi.org/10.5772/intechopen.105010*

[27] Pasandideh-Fard M, Bhola R, Chandra S, Mostaghimi J. Deposition of tin droplets on a steel plate: simulations and experiments. International Journal of Heat and Mass Transfer. 1998;**41**(19): 2929-2945

[28] Al-Bataineh SA, Szili EJ, Gruner PJ, Priest C, Griesser HJ, Voelcker NH, et al. Fabrication and operation of a microcavity plasma array device for microscale surface modification. Plasma Processes and Polymers. 2012;**9**(7):638-646

[29] Zhang P, Kortshagen U. Twodimensional numerical study of atmospheric pressure glows in helium with impurities. Journal of Physics Department Applied Physics. 2006;**39**:153

[30] Este G, Westwood WD. A quasidirect-current sputtering technique for the deposition of dielectrics at enhanced rates. Journal of Vacuum Science & Technology. 1988;**A6**:1845. DOI: 10.1116/1.575266

[31] Birmingham JG, Moore RR. Reactive Bed Plasma Air Purification. US Patent no. 4954 320, 1990

[32] Gerenser LJ, Grace JM, Apai G, Thompon PM. Surface chemistry of nitrogen plasma-treated poly(ethylene-2,6-naphthalate): XPS, HREELS and static SIMS analysis, Surface and Interface Analysis. Volume 29, Issue 1, Pages 12-22, January 2000. DOI: 10.1002/(SICI)1096-9918(200001)29:1 <12::AID-SIA687>3.0.CO;2-7

[33] Harlow FH, Amsden AA. A numerical fluid dynamics calculation method for all flow speeds. Journal of Computational Physics. Volume 8, Issue 2, Pages 197-213, October 1971. DOI: 10.1016/0021-9991(71)90002-7

[34] Han Z, Fan L, Reitz RD. Multidimensional Modeling of Spray Atomization and Air-Fuel Mixing in a Direct-Injection Spark-Ignition Engine. SAE Paper 970884, SAE Transactions, Journal of Engines. 01 Jan 1997;**106**(3): 1423-1441

[35] Olesik JW, Smith LJ, Williamsen EJ. Signal fluctuations due to individual droplets in inductively coupled plasma atomic emission spectrometry, Analytical Chemistry. 15 Sep 1989;**61**(18):2002-2008. DOI: 10.1021/ac00193a003

[36] Hobbs SE, Olesik JW. The effect of desolvating droplets and vaporizing particles on ionization and excitation in Ar inductively coupled plasmas. Spectrochimica Acta Part B: Atomic Spectroscopy. May-June 1993;**48**(6-7): 817-833. DOI: 10.1016/0584-8547(93) 80087-B

[37] Meyyappan M. Computational Modelling in Semiconductor Processing, Pages: 363. Published by Artech House. 1994. ISBN 10: 0890067074; ISBN 13: 9780890067079. Available from: https:// uk.artechhouse.com/Computational-Modeling-in-Semiconductor-Processing-P95.aspx

[38] Arfken G. Mathematical Methods for Physicists. 3rd Edition. January 1, 1985, Copyright: © Academic Press 1985, eBook ISBN: 9781483277820. Available from: https://www.elsevier.com/books/ mathematical-methods-for-physicists/ arfken/978-0-12-059820-5

[39] Iqbal MM, Turner MM, Oral presentation. Dynamic Interaction Between Liquid Droplets and Atmospheric Pressure Discharge, 37th EPS Conference, 21–25 June, 2010, Vol. 34A, ISBN 2-914771-62-2

[40] Coppins M. Electrostatic Breakup in a Misty Plasma. Physical Review Letters. 104, 065003 - Published. 11 February 2010. DOI: 10.1103/PhysRevLett.104. 065003

[41] Rayleigh L. On The Instability Of Jets. Proceedings of the London Mathematical Society, Volume s1-10, Issue 1, Pages 4-13, November 1878. DOI: 10.1112/plms/s1-10.1.4

[42] Yue Y, Sun J, Gunter KL, Michalek DJ, Sutherland JW. Character and Behavior of Mist Generated by Application of Cutting Fluid to a Rotating Cylindrical Workpiece, Part 1: Model Development. Journal of Manufacturing Science Engineering. August 2004;**126**(3):417-425. DOI: 10.1115/1.1765150

[43] Iqbal MM, Stallard CP, Dowling DP, Turner MM. Two-Dimensional Integrated Model for Interaction of Liquid Droplets with Atmospheric Pressure Plasma. Plasma Processes and Polymers. Volume12, Issue 11, Pages 1256-1270, November 2015. DOI: 10.1002/ppap.201500003

[44] Iqbal MM, Turner MM. Investigations of Droplet-Plasma Interaction using Multi-Dimensional Coupled Model, Contributions to Plasma Physics. Volume 55, Issue 9, Pages 627-642, October 2015. DOI: 10.1002/ctpp.201500048

[45] Iqbal MM, Turner MM. Spatial Uniformity of Atmospheric Pressure Discharges: A Simulation Study. Contributions to Plasma Physics. Volume 54, Issue 9, Pages 756-771, October 2014. DOI: 10.1002/ ctpp.201400002

[46] Hagelaar GJM. Modeling of microdischarges for display technology. [Phd Thesis 1 (Research TU/e / Graduation TU/e), Applied Physics]. Eindhoven: Technische Universiteit, 2000. DOI: 10.6100/IR538700

[47] Peaceman DW, Rachford HH Jr. The Numerical Solution of Parabolic and

Elliptic Differential Equations. Journal of the Society for Industrial and Applied Mathematics. March 1955;**3**(1). DOI: 10.1137/0103003

[48] Adler SL. Over-relaxation method for the Monte Carlo evaluation of the partition function for multiquadratic actions. Physical Review D. Volume 23, Issue 12, Pages 2901, 15 June 1981. DOI: 10.1103/PhysRevD.23.2901

[49] Lymberopoulos DP, Economou DJ. Modeling and simulation of glow discharge plasma reactors. Journal of Vacuum Science & Technology. 1994**A12**(4):1229-1236

[50] Yuan X, Raja LL. Computational Study of Capacitively Coupled High-Pressure Glow Discharges in Helium. IEEE Transactions On Plasma Science. Aug 2003;**31**(4):495-503

[51] WolframAlpha computational intelligence, ©2022 Wolfram Alpha LLC, Wolfram for Education. Available from: https://www.wolframalpha.com

[52] Hagelaar GJM, Kroesen GMW, Van Slooten U, Schreuders H. Modeling of the microdischarges in plasma addressed liquid crystal displays. Journal of Applied Physics, Volume 88, Issue 5, pp 2252, 20 March 2000. DOI: 10.1063/1.1287529

[53] Alves LL, Gousset G, Ferreira C. Self-contained solution to the spatially inhomogeneous electron Boltzmann equation in a cylindrical plasma positive column. Physical Review E. Volume 55, Issue 1, pp 890, 1 January 1997. DOI: 10.1103/PhysRevE.55.890

[54] Hagelaar G, Pitchford L. Solving the Boltzmann equation to obtain electron transport coefficients and rate coefficients for fluid models, Plasma Sources Science and Technology. Volume 14, Number 4, pp. 722-733,

*Interaction and Transport of Liquid Droplets in Atmospheric Pressure Plasmas (APPs) DOI: http://dx.doi.org/10.5772/intechopen.105010*

2005. Available from: http://stacks.iop. org/PSST/14/722

[55] Petrucci RH, Geoffrey Herring F, Madura JD, Bissonnette C. General chemistry principles & modern applications. Upper Saddle River Pearson, 2007. ISBN: 0135334985

[56] Iqbal MM, Turner MM, Behavior of HMDSO Liquid Precursor Droplets in APD Plasma and Gas Mixture, Proceedings in 31st ICPIG, July 14-19, 2013, Granada, Spain

Section 4
