International Journal of Multiphase Flow 35 (2009) 8–19

Contents lists available at ScienceDirect

International Journal of Multiphase Flow

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / i j m u l fl o w

Flow and breakup characteristics of elliptical liquid jets

T.V. Kasyap, D. Sivakumar *, B.N. Raghunandan
Department of Aerospace Engineering, Indian Institute of Science, Bangalore 560 012, India

a r t i c l e i n f o a b s t r a c t

Article history: This paper presents the results of an experimental study on liquid jets discharging from elliptical ori-
Received 7 December 2007 fices into still ambient air. The experiments were conducted with a set of elliptical orifices of approx-
Received in revised form 7 August 2008 imately same area of cross section but varying orifice aspect ratio using water and water–glycerol
Accepted 21 September 2008
mixture as experimental fluids. The flow behavior of liquid jets was analyzed using their photographs
Available online 27 September 2008
captured by an imaging system. The measurements obtained for the elliptical liquid jets were com-
pared with the circular liquid jets discharging from a circular orifice of the same area of cross section.
Keywords:
Elliptical geometry of the orifice results in a flow process by which the emanating liquid jet period-
Elliptical liquid jets
Breakup
ically switches its major and minor axes as it flows downstream of the orifice. In this paper, we
Instability attempt to characterize the axis-switching process through its wavelength and amplitude. For a given
Axis-switching elliptical orifice, the axis-switching process is dominantly seen in a particular range of flow conditions.
The effects of the orifice aspect ratio and liquid viscosity on the axis-switching process are revealed
through this study. The experimental results on jet breakup show that axis-switching process has a
destabilizing effect on elliptical liquid jets within a particular range of flow conditions and it results
in shorter breakup lengths compared to the circular jet. The extent to which axis-switching destabi-
lizes the jet is dictated by the viscosity of liquid. An increase in orifice aspect ratio destabilizes ellip-
tical liquid jets with low viscosity like water; however, this behavior seems to get obscured in water–
glycerol mixture elliptical jets due to high viscosity.
Ó 2008 Elsevier Ltd. All rights reserved.

1. Introduction disturbance. Haenlien (1932) conducted experiments on liquid

jet breakup with different liquids and circular orifices and con-
Liquid jet breakup is a ubiquitous phenomenon in nature and is structed jet breakup length versus jet velocity curve (breakup
a classic problem in hydrodynamics. It is commonly observed in curve) from the experimental measurements. Jet breakup length
spray and droplet formation processes encountered with several is the length of the coherent portion of the jet measured from
practical applications. Academic interest in the problem of liquid the orifice exit plane (Grant and Middleman, 1966; Lin and Reitz,
jet breakup dated back to the 19th century. Seminal contributions 1998). Breakup length of the circular jet increases linearly with
on the flow behavior of liquid jets discharging from circular and jet velocity in a range of low jet velocities, reaches a peak or critical
noncircular orifices came from the works of Bidone (1829), Savart point, and thereafter decreases with increasing jet velocity. As jet
(1833), Magnus (1855), Plateau (1873), Boussinessq (1877), Ray- velocity increases, the ambient atmosphere starts influencing the
leigh (1879, 1945) and Chandrasekhar (1961). Delightful accounts liquid jet and asymmetric waves or transverse waves begin to grow
of the earlier works on liquid jet breakup can be obtained from the on the jet surface. Weber’s theoretical analysis by considering the
comprehensive review articles by Bogy (1979) and Lin and Reitz effect of ambient medium revealed a peak in the breakup curve
(1998). Lord Rayleigh (1945) carried out a linear stability analysis (Weber, 1931), however, model predictions of the jet velocity at
of an infinite cylindrical column of inviscid liquid and concluded the peak or critical point did not match with experimental results
that such a liquid jet is unstable to axisymmetric disturbances of (Grant and Middleman, 1966). It was concluded by Phinney (1972,
wavelength exceeding the circumference of the unperturbed liquid 1973) that, immediately after the peak, the ambient medium
jet. Weber (1931) included the effects of viscosity and ambient causes a sharp increase in the amplification rate of disturbances
medium in the analysis of liquid jet breakup and found that the ef- on the liquid jet which reverses the increasing trend of breakup
fect of viscosity is to change the wavelength of the most unstable length variation as observed with experiments (Grant and Middle-
man, 1966). Further increase in the jet velocity earmarks the pres-
ence of short wavelength disturbances on the jet surface over a
* Corresponding author. Tel.: +91 80 22933022; fax: +91 80 23600134.
range of jet velocities and the breakup curve shows a positive slope
E-mail addresses: [email protected], [email protected] (D. Siva-
kumar). in this regime (Lefebvre, 1989; Blaisot and Adeline, 2000). Breakup

0301-9322/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijmultiphaseflow.2008.09.002
 T.V. Kasyap et al. / International Journal of Multiphase Flow 35 (2009) 8–19 9

length soon reaches a maximum with increasing jet velocity and Except the above conclusions on the free surface evolution of
the liquid jet operates in the fully developed spray regime or the elliptical liquid jets, the fluid dynamic behavior of such liquid jets
atomization regime at very high jet velocities. is largely overlooked in the literature. For instance published
Despite all these advances, there exist some subtle areas in this works on the breakup phenomena of liquid jets issuing from ellip-
jet breakup problem which have not received adequate attention tical orifices is very scarce in the literature despite the fact that
in past. For instance, effects of orifice geometry on the behavior these orifices have already been studied for potential practical
of liquid jets have never been consistently taken into consideration applications (McHale et al., 1971; Snyder et al., 1989). Hoyt and
(Lin and Reitz, 1998). A liquid jet emanating from an elliptical ori- Taylor (1978) reported from their experiments that they were
fice switches its major and minor axes by right angles in a periodic not able to identify any regular breakup on elliptical water jets.
manner as it flows downstream (Bidone, 1829; Rayleigh, 1879; In this paper, we present experimental results on the breakup of li-
Taylor, 1960). Surface tension causes the elliptical jet cross section quid jets issuing from elliptical orifices of different aspect ratios
to oscillate about a circular figure of equilibrium like a stretched using water and water–glycerol mixture as working liquids. Addi-
membrane and consequentially the elliptical liquid jet executes tional experiments were conducted with a circular orifice of
multiple axis-switching (Rayleigh, 1879). Schematic sketches of approximately same cross sectional area as that of the elliptical
free surface boundaries of a liquid jet discharging from an elliptical orifices for the sake of comparison purposes. This paper is orga-
orifice are given in Fig. 1. Fig. 1(a) and (b) shows the appearance of nized as follows. The experimental apparatus and procedure used
elliptical liquid jet in the major axis plane and the minor axis plane in the study are given in the next section which is followed by
of the elliptical orifice, respectively. The sketches also illustrate the presentation and discussion of the experimental results. In
axis-switching phenomenon observed with the elliptical liquid Section 3, we first describe the visual observations on elliptical jets
jet. Once the jet ejects out of an elliptical orifice, surface tension at various flow conditions, followed by the characterization of the
force tries to minimize the curved surface area of the jet by pulling axis-switching process in detail. We then proceed to describe the
the ends of major axis inwards and pushing the ends of minor axis breakup curves of elliptical and circular jets of both liquids and
outwards. Due to the lateral inertia of the jet, the movements of present evidences to establish the role of axis-switching in the
major and minor axes ends do not stop abruptly at the ideal circu- breakup of elliptical liquid jets. Finally, the major results arrived
lar cross section, i.e., the cross section with minimum surface area, from the experimental study are listed in Section 5.
but overshoot. This causes the outward moving minor axis ends to
be pushed further outwards and the inward moving major axis 2. Experimental details
ends to be pulled further inwards. This kind of geometrical trans-
formation superimposed with the axial motion of jet gives rise to A set of stainless steel orifices containing one circular orifice
the axis-switching profile on the jet as illustrated in Fig. 1. The and five elliptical orifices of various aspect ratios were used for
axis-switching wavelength, kas, is defined as the distance measured the study. The geometric details of the orifices are given in Table
between two consecutive crests or troughs as shown in Fig. 1. 1. Cross sectional areas of all of these orifices were approximately
Rayleigh (1890) developed a mathematical model to calculate the same. Elliptical orifices were manufactured by electro-discharge
dynamic surface tension of liquids from the axis-switching wave- machining process. Note that, owing to the fabrication processes,
length of elliptical liquid jets. This model was further improved there were differences in the cross sectional area between elliptical
by Bohr (1909). Analytical and numerical studies have been re- orifices and the circular orifice and the measured error in this re-
ported on elliptical liquid jets with and without surface tension gard was estimated to be less than 5%. Sufficient care was taken
to understand the spatial evolution of the free surface (Green, to minimize the errors in obtaining the contour for elliptical ori-
1977; Geer and Strikwerda, 1980, 1983). Bechtel et al. (1988a,b, fices. The maximum deviation between the actual orifice contour
1989) studied viscoelastic liquid jets issuing from elliptical orifices and the ideal elliptical contour (the contour obtained based on
and predicted the axis-switching behavior of the liquid jets theo- the major and minor axes dimensions of the elliptical orifice)
retically for the special case of Newtonian jet with constant surface was estimated to be less than 0.078 mm (3.1% of Do of the circular
tension. Later their model was extended to devise a technique for orifice) for all elliptical orifices used in the present study. The sur-
the measurement of dynamic surface tension and elongational vis- face roughness of internal walls of the orifices was characterized
cosity of liquids (Bechtel et al., 1995). by cutting a sample orifice longitudinally and examining it with

Fig. 1. Schematic sketches of an elliptical liquid jet discharging from an elliptical orifice. (a) Jet appearance in the major axis plane of the elliptical orifice, and (b) jet
appearance in the minor axis plane of the elliptical orifice.
10 T.V. Kasyap et al. / International Journal of Multiphase Flow 35 (2009) 8–19

Table 1
Geometrical details for the orifices used in the present study

Elliptical orifices Circular orifice

b a
Orifice index Major axis (mm) Minor axis (mm) Deq (mm) Length (mm) Aspect ratio Orifice index Do (mm) Length (mm) Aspect ratioa
E1 3.02 2.04 2.48 25 1.48 C 2.51 25 1
E2 3.47 1.79 2.49 25 1.94
E3 4.43 1.49 2.57 25 2.97
E4 5.05 1.31 2.57 25 3.85
E5 6.07 1.04 2.51 25 5.84
a
The ratio of major axis to minor axis of the orifice.
b
The geometric mean of the major and minor axes of the elliptical orifice.

a stylus profilometer at various locations. The mean surface rough- Orifice

ness of the internal walls of the orifices was found to be ranging 50 mm
from 0.8 to 1.48 lm. Fig. 2 illustrates the circular orifice and a sam-
ple of elliptical orifices used in the present study. During experi-
mental runs, orifices were fixed in an assembly consisted of a
cylindrical chamber of diameter 16 mm and length 50 mm and a 16 mm 60°
converging section of circular cross section. Care was taken to
avoid any sharp corner between the converging portion and the
orifice passage entrance by rounding off the edge between them.
The details of the entire orifice assembly are given in Fig. 3. A
20 mm
high-pressure liquid flow line comprised of a filter, a pressure Quieting chamber
gauge, and a control valve was used to supply experimental fluids
from a large stainless steel tank to the orifice assembly. The vol- Fig. 3. A schematic of the orifice assembly.
ume of the tank (approximately 80 L) was large enough to neglect
the orifice inlet pressure variations during experimental runs.
Moreover, the tank was pressurized by means of a constant pres- The two normal views were not taken simultaneously because of
sure, regulated, compressed air supply which ensured a constant the limitation of the experimental setup. Since axis-switching is
orifice inlet pressure. a stationary phenomenon, the experimental procedure adopted
Experiments were carried out with two different liquids, water in the present work did not restrict the analysis of axis-switching
and a 1:1 mixture by volume of water and glycerol. The relevant in any form. Due to the limitation of the field of view of the camera,
physical properties of water–glycerol mixture have been taken for each flow condition several images of jets were captured at var-
from the literature (Yang and Leong, 2002) and are given in Table ious axial locations from the orifice exit at different instants which
2. Liquid flow rate from the orifice was estimated by measuring were later patched together to obtain the complete picture of the
the volume of liquid collected in a measuring jar for a fixed time intact jet without any spatial discontinuity. Captured images were
interval. The maximum error in the measurement of flow rate analyzed by means of commercially available image processing
was estimated to be less than 3%. A Nikon D1X camera with a dif- software. In the present study, the jet breakup length is defined
fused backlighting system was used to take photographs of the jet. as the distance measured from the orifice exit to the point of the
The imaging time was less than 15 ls. The pixel resolution of the first break in the jet along the jet axis. Mean breakup length, Lb
camera was 2000  1312 with the viewing field ranging from 30 was estimated from the individual measurements of jet breakup
to 60 mm. A detailed description of the experimental setup used length obtained from photographs available for each flow condi-
in this study can be obtained from Kasyap et al. (2008). Identical tion. The amplitude of axis-switching, Dmax, defined as the width
experimental procedure was followed for both circular and ellipti- of jets measured in the major axis plane, was also obtained from
cal orifices except that for each elliptical orifice flow condition two the photographs. In a similar manner the measurement of axis-
sets of images were taken albeit at different instants, one with the switching wavelength, kas was also obtained from the photographs.
camera viewing plane aligned with the major axis plane of the Sufficient care was taken to ensure the repeatability of experimen-
elliptical orifice and other with the minor axis plane of the ellipti- tal measurements. For each flow condition, experiment was re-
cal orifice to correctly observe the axis-switching phenomenon. peated from 8 to 15 times and mean and standard deviation of

Fig. 2. Typical high resolution images of orifices used in the present study. (a) Circular orifice C, and (b) elliptical orifice E2.
 T.V. Kasyap et al. / International Journal of Multiphase Flow 35 (2009) 8–19 11

Table 2 3. Results
Properties of working fluids

Working fluid ql (kg/m3) ll (kg/m/s) r (N/m) 3.1. Visual observations of elliptical liquid jets with different flow
Water 995 0.802  10 3
0.0728 conditions
Water–glycerol mixture (1:1 by volume) 1107 6.01  103 0.0669
Experimental data obtained for liquid jets discharging from
elliptical orifices reveals that elliptical liquid jets show
distinct flow features in different ranges of flow conditions.
Table 3
Maximum standard deviation of measured parameters
Fig. 4(a)–(c) shows images of waterpjets ffiffiffiffiffiffiffi discharging from the
elliptical orifice E3 for different low We in the range of 0.86–
Liquid Measured We Range of Maximum 1.49. For each We, a pair of images are given in which the top
quantities measured standard deviation
values (cm) (cm)
one corresponds to the major axis plane and the bottom one
to the minor axis plane of the elliptical orifice. The jet width ap-
Water kas <30 <1.5 0.07
pears to be same in both images for all We except in the vicinity
30–225 1.5–3.0 0.10
Dmax 4–100 0.2–0.4 0.02 of the orifice exit. A comparison of the major axis and minor axis
images for each We given in Fig. 4 reveals that the elliptical jet
Water–glycerol kas <15 <1.2 0.1
mixture 15–400 1.2–5.0 0.4 does not exhibit the axis-switching phenomenon. Instead it
Dmax 10–625 0.2–0.4 0.01 degenerates to a circular jet immediately after exiting from the
orifice. The dominance of surface tension forces at these low
We conditions compels the jet to take a circular cross section
to minimize its surface energy and leads to the suppression of
various measured quantities were calculated subsequently. The
axis-switching process. At locations far away from the orifice
standard deviation for the measurements of axis-switching wave-
exit, an elliptical jet develops symmetric radial deformations
length and the amplitude of axis-switching at different We ranges
on its surface and breaks into drops in the exact manner of
for both liquids are given in Table 3.
the Rayleigh breakup of a circular liquid jet (Rayleigh, 1945).
Flow conditions of liquid jets were expressed in terms of nondi-
Fig. 4(d) shows the image pffiffiffiffiffiffiffi of a water jet discharging from the
mensional numbers such as Weber number, We, Reynolds number,
circular orifice C for We ¼ 1:42 which falls within the range
Re and Ohnesorge number, Oh which were estimated from
of We conditions given in Fig. 4(a)–(c). It is interesting to ob-
ql U 2o D serve from the images given in Fig. 4 that the flow behavior of
We ¼ ; ð1Þ elliptical liquid jets at these We conditions resembles that of a
r
q Uo D circular liquid jet and lacks the influence of orifice shape.
Re ¼ l ; ð2Þ As We increases, the lateral inertia of the elliptical liquid jet also
ll
ll increases and becomes comparable to its surface tension force. At
Oh ¼ pffiffiffiffiffiffiffiffiffiffiffi
ffi; ð3Þ this point, according to Rayleigh (1879), the free surface of the
ql Dr
jet behaves like a stretched membrane which executes oscillations
where ql is the liquid density, Uo, the mean axial velocity of the in the lateral direction about a circular figure of equilibrium and
liquid jet at the orifice exit; D, the flow geometrical scale; ll, the li- develops axis-switching. Fig. 5 illustrates the flow behavior of
pffiffiffiffiffiffiffi
quid viscosity, and r, the surface tension. The value of D for the cir- elliptical water jets at moderate We values falling in the range
cular and the elliptical orifices was taken as Do, the orifice exit of 3.93–10.58. A comparison of the free surface of elliptical jets
diameter, and Deq, the geometric mean diameter of the elliptical ori- seen in the near region of the orifice exit between the major axis
fice, respectively. The values of Uo were estimated from the volume and minor axis images for each We reveals that the elliptical jet
flow rate and the orifice exit area. In the present study, the ranges of switches its axes as it flows downstream. The switching process
values of We and Re for water jets were 0.8–13,300 and 450–61,000, is found to be periodic particularly in the lower band of We values
respectively. The corresponding values for water–glycerol mixture given in Fig. 5. At still higher We, the free surface of elliptical jet is
jets were 0.28–1300 and 36–2300. Oh numbers for water jet and not as smooth as seen at lower We probably due to other influenc-
water–glycerol mixture jets were 0.0019 and 0.0139, respectively. ing parameters, like turbulence or the ambient medium. Fig. 6

pffiffiffiffiffiffiffi
Fig. 4. Images (a)–(c) of water jets discharging from the elliptical orifice E3 at three different We conditions: (a) 0.86, (b) 1.12, and (c) 1.49. For each flow condition a pair of
images is given: the top one corresponds to the case in which the camera viewing plane was aligned with the major axis plane of the elliptical orifice and the bottom one to
pffiffiffiffiffiffiffi
the minor axis plane of the elliptical orifice. (d) Image of water jet discharging from the circular orifice C at We ¼ 1:42. The line segment shown on the left top corner of the
figure indicates a scale distance of 5 mm.
12 T.V. Kasyap et al. / International Journal of Multiphase Flow 35 (2009) 8–19

pffiffiffiffiffiffiffi
Fig. 5. Images of water jets discharging from the elliptical orifice E3 at different moderate values of We: (a) 3.93, (b) 5.45, (c) 6.35, (d) 8.06, and (e) 10.58. All other details
remain same as in the caption of Fig. 4.

illustrates water jets discharging from the elliptical orifice E3 for

pffiffiffiffiffiffiffi due to the formation of ruffles on water jets and difficulty in deter-
different high We values falling in the range of 18–31. The ruf- mining the axis-switching
pffiffiffiffiffiffiffi segments precisely. The variation of
fling of the jet surface at these We conditions is evident from the kas/Deq with We for both water jets and water–glycerol mixture
images given in Fig. 6 which hinders the precise identification of jets discharging from elliptical orifices of various aspect ratios is gi-
axis-switching segments. ven in Fig. 9. The value of kas for a given We presented in this plot
The experimental observations on the flow behavior of elliptical corresponds to the mean value of all measurable wavelengths seen
water jets hold good for the elliptical water–glycerol mixture jets in the entire jet. For both water and water–glycerol mixture jets,
pffiffiffiffiffiffiffi
studied here except the significant viscous damping of the axis- there exists a lower limit of We below which axis-switching
switching process and the suppression of ruffling on the jet surface was impossible to be perceived. pIn
ffiffiffiffiffiffiffithe case of water jets, measure-
at higher We present in the latter. Fig. 7 shows the viscous damp- ments of kas were limited by a We above which precise measure-
ing of axis-switching of the water–glycerol mixture jet issuing ment was impossible due to the presence of ruffles on the jet
from the elliptical orifice E4 at We = 39.26. The amplitude of surface. The plots of kas/Deq for both water and water–glycerol mix-
axis-switching decreases drastically along the jet and at larger dis- ture jets show a linear variation which is consistent with the pre-
tances from the orifice exit the elliptical jet degenerates to a circu- vious observations (Rayleigh, 1879;
pffiffiffiffiffiffiffi Geer and Strikwerda, 1983).
lar liquid jet. Owing to the higher viscosity, no ruffling is seen on The linear trend of kas/Deq with We holds true for all orifice aspect
the free surface of water–glycerol jets at higher We. This is clear ratios and the orifice aspect ratio does not seem to influence kas/Deq
from Fig. 8 which illustrates images of water and water–glycerol for moderate values of the orifice aspect ratio because all curves
mixture jets discharging from the elliptical orifice E3 for We  330. collapse to a reasonable extent. However, an examination of the
slope of the curves given for water jets in Fig. 9 reveals a deviation
3.2. Characterization of the axis-switching phenomenon in slope for the data obtained for the elliptical orifices with high
orifice aspect ratio, especially at higher We. Elliptical jets discharg-
The axis-switching phenomenon of elliptical liquid jets is char- ing from orifices with high aspect ratios exhibit distortion in their
acterized in the present study by the wavelength of axis-switching elliptical cross section (Bechtel et al., 1995) which may alter the
(kas) and its amplitude or length of major axis (Dmax). The measure- axis-switching wave length. The effect of viscosity on kas can be
ments of kas and Dmax for water jets were limited to a relatively understood from Fig. 9. Since all data points of water jets and
narrow range of We compared to water–glycerol mixture jets water–glycerol mixture jets collapse into a single in Fig. 9 line

pffiffiffiffiffiffiffi
Fig. 6. Images of water jets discharging form the elliptical orifice E3 at different high We conditions: (a) 18.35 and (b) 30.34. All other details remain same as in the caption
of Fig. 4.
 T.V. Kasyap et al. / International Journal of Multiphase Flow 35 (2009) 8–19 13

Fig. 7. Image of water–glycerol mixture jet discharging from the elliptical orifice E4 at We = 39.26. All other details remain same as in the caption of Fig. 4.

Fig. 8. Images of water jet and water–glycerol mixture jet discharging from the elliptical orifice E3 for We  330. (a) Water jet and (b) water–glycerol mixture jet. The
Ohnesorge numbers are 0.0019 and 0.0139, respectively, for (a) and (b). All other details remain same as in the caption of Fig. 4.

30
Water jets
E1 (1.48)
E2 (1.94)
E3 (2.97) 7
E4 (3.85)
20
E5 (5.84)
eq
/D

6
as

Water-glycerol
λ

mixture jets

10 E1 (1.48)
E2 (1.94)
λas/Deq

5
E3 (2.97)
E4 (3.85) E1(1.48)
E5 (5.84)
E2(1.94)
0 4
E3(2.97)
0 8 16 24 32
E4(3.85)
We 1/2 E5(5.84)
Fig.
3
p ffiffiffiffiffiffiffi9. Variation of nondimensionalized axis-switching wavelength, kas/Deq with
We for water jets and water–glycerol mixture jets discharging from the elliptical 0 2 4 6 8
orifices with different orifice aspect ratio. The solid line is the linear fit of the entire
Axis switching number
data. The error bar for each data point indicates one standard deviation towards
both directions from the mean.
20

we can conclude that the wavelength of axis-switching is indepen-

dent of the viscosity also.
The variation of kas along the axis of the jet is important since 15
some of the previous studies have reported its consequences while
measuring the dynamic surface tension (Bohr, 1909). The measure-
ment of kas along the axis of the jet has been attempted only for
λas/Deq

water–glycerol mixture jets since the ruffles developing on water 10

jets at larger axial distances from the orifice exit makes the accu- We1/2
rate measurement of kas impossible. Measured values of kas along 4.15
the axis of water–glycerol mixture jets have been plotted in
6.63
Fig. 10 in the nondimensional form. The abscissa of the plot is 5
the axis-switching number as shown in Fig. 1. Fig. 10(a) illustrates 12.99
the axial variation of kas for the elliptical orifices of various aspect
pffiffiffiffiffiffiffi 17.65
ratios around a We of 6.0 and Fig. 10(b) depicts pffiffiffiffiffiffiffithe axial variation
of kas for the elliptical orifice E3 at various We conditions. Axis- 0
switching wavelength is initially low and it increases to almost a 0 2 4 6 8
constant value. This is in agreement with the previously observed Axis switching number
trends by Bohr (1909). Two effects may contribute to the axial-var-
Fig. 10. Variation of nondimensionalized axis-switching wavelength, kas/Deq with
iation of kas: one is the gravitational acceleration, g of the jet and axis-switching number for water–glycerol mixture jets. (a) For jets issuing from
pffiffiffiffiffiffiffi
other is the presence of the orifice. To reveal the effect of gravity, various pelliptical orifices at a We ¼ 6:0, and (b) for jets issuing from E3 with
ffiffiffiffiffiffiffi
Froude number, Fr defined by various We.
14 T.V. Kasyap et al. / International Journal of Multiphase Flow 35 (2009) 8–19

U 20 1.6
Fr ¼ ð4Þ E1(1.48)
gDeq
E2(1.94)
has been calculated for flow conditions shown in the plots. Esti-
E3(2.97)
mated value of Fr ranges from 10 to 850, indicating that the flow 1.4
is dominated by inertia. Moreover, if gravitational acceleration af- E4(3.85)
fects kas, a continuous increase in kas is to be expected since the E5(5.84)
jet velocity continues to increase along the jet axis. However, the

Dmax/Deq
plots given in Fig. 10 do not show such a continuous variation. 1.2
Therefore, the plausible cause for the variation of kas near the orifice
may be the presence of orifice itself. Small irregularities involved in
the formation of the jet, like an imperfection in the contour of the
orifice, may cause a change in kas near the orifice, which later dies 1.0
out at larger distances from the orifice, yielding a nearly constant
kas. This point was considered by Bohr (1909). However, a rigorous
study is required to ascertain these facts.
0.8
The variation of Dmax/Deq given in Fig. 11(a) and (b) for water 0 8 12
4
and water–glycerol
pffiffiffiffiffiffiffi mixture jets, respectively, shows a monotonic 1/2
We
increase with We. The value of Dmax given in these figures corre-
sponds to the jet width measured at the first crest of the major axis
image of the jetpas illustrated in Fig. 1. The monotonic variation of 1.6
ffiffiffiffiffiffiffi E1(1.48) E2(1.94)
Dmax/Deq with We is understandable since an increase in We
E3(2.97) E4(3.85)
apparently increases the lateral inertia of the jet which results in
the increase in the magnitude of overshooting of the jet major axis. E5(5.84)
1.4
In contrast to the variation of kas/Deq given in Fig. 9, the orifice as-
pect ratio shows a remarkable effect on the nondimensionalized
amplitude. For a given flow condition, an increase in orifice aspect
Dmax/Deq

ratio causes an increase

pffiffiffiffiffiffiffi in amplitude also. Superimposed curves of 1.2
Dmax/Deq versus We for water and water–glycerol mixture jets
issuing from a typical elliptical orifice are shown in Fig. 11(c). It
is obvious from the plot that for the same flow condition, the
amplitude of axis-switching of water–glycerol liquid jet is less than 1.0
that of the water jet. This is due to intense viscous damping pres-
ent in water–glycerol mixture jets. The same trend has been found
for other aspect ratios also.
Further insights about the effect of viscous damping on Dmax 0.8
can be obtained by nondimensionalizing the measurements of 0 5 10 15 20 25
1/2
Dmax given in Fig. 11(a) and (b) with the major axis dimension We
of the elliptical orifice, Do,major to yield the quantity Dmax/Do,major.
Physically this quantity represents the extent to which Dmax 1.4
recovers to the major axis dimension of the particular elliptical
orifice. If gravity is neglected and jet is of an inviscid liquid, then
Dmax should always recover to the major axis dimension of the
orifice Do,major, thus making the value of Dmax/Do,major equal to
unity. However, in real cases, due to the viscous damping of 1.2
axis-switching and the presence of gravitational acceleration,
Dmax/Deq

Dmax/Do,major will always be less than unity. To study the effect

of the orifice p ffiffiffiffiffiffiffi ratio on Dmax/Do,major, the variation of Dmax/D
aspect
o,major with We is plotted for each orifice aspect ratio in
Fig. 12(a) and (b). The former is for water jets and the latter is 1.0
for water–glycerol mixture jets. It is obvious from Fig. 12(a) and
(b) that liquid jets issuing from orifices of larger aspect ratios Water
show smaller values of Dmax/Do,major and therefore recover less
Water-glycerol mixture
than those issuing from orifices of lower aspect ratios. The nature
of the amplitude decay along the jet axis is more or less exponen- 0.8
tial. Fig. 13 illustrates the exponential nature of amplitude decay 0 5 10 15 20 25
of water–glycerol mixture jet emanating from the elliptical orifice We
1/2

E4. The solid curve in Fig. 13 fits the variation of amplitude with
axis-switching number without considering the amplitude of the Fig.
p ffiffiffiffiffiffiffi11. Variation of nondimensionalized axis-switching amplitude, Dmax/Deq with
zeroth axis-switching which is the major axis dimension of the We. (a) Elliptical water jets, (b) elliptical water–glycerol mixture jets, and (c)
water and water–glycerol mixture jets issuing from the elliptical orifice E3 together.
elliptical orifice, Do,major. If Do,major is introduced in the data set,
the same exponential curve does not fit the experimental mea-
surements. In other words, between the orifice exit and the first
crest of the axis-switching profile in the major axis plane, the aspect ratios also. This shows the possible role of the orifice exit
amplitude decay is no longer same as the exponential decay ob- conditions on the amplitude of axis-switching, similar to the ef-
served afterwards. A similar trend has been observed with other fect of orifice exit on the axis-switching wavelength.
 T.V. Kasyap et al. / International Journal of Multiphase Flow 35 (2009) 8–19 15

1.0 3.3. Breakup of elliptical liquid jets

Measurements of the mean breakup length, Lb of the jet were

obtained for the water jets discharging from the circular and
0.8 elliptical orifices with different flow conditions. Fig. 14 shows
the variation of Lb/Deq for water jets discharging from the circu-
lar and elliptical orifices for the whole range of We and Fig. 15
Dmax/Do,major

depicts the same for water–glycerol mixture jets. For a clear

0.6 view of the nature of breakup curves in the lower range of We
of water jets, a portion of the breakup curves falling in the
pffiffiffiffiffiffiffi
E1(1.48) We range of 0–20 has been magnified and inserted in
Fig. 14. The data identified by filled symbols in Figs. 14 and 15
E2(1.94)
0.4 corresponds to the variation of Lb/Deq for the circular liquid jets,
E3(2.97) which shows a linear variation at first, reaches a peak and there-
E4(3.85) after decreases with increasing We. This trend in the breakup
E5(5.84) curve for liquid jets discharging from the circular orifice C is
0.2 in agreement with the previous works (Grant and Middleman,
2 4 6 8 10 12 1966; Lefebvre, 1989; Blaisot and Adeline, 2000; Sterling and
1/2
We Sleicher, 1975) discussed in Section 1. The data identified by
open symbols in Figs. 14 and 15 corresponds to the variation
1.0 of Lb/Deq for jets discharging from the elliptical orifices with dif-
ferent aspect ratio. At very low We, the breakup curves of ellip-
tical jets for all orifice aspect ratios tend to merge along line A–A
as shown in the inset of Figs. 14 and 15. It indicates that the ori-
0.8 fice geometry (elliptical orifice shape) does not influence the
breakup of elliptical liquid jets at very low We. Images given
in Fig. 4(a)–(c) correspond to flow conditions falling on the line
Dmax/Do,major

A–A. Absence of axis-switching causes elliptical jets to behave in

0.6 a similar manner to that of the equivalent circular jet, which ex-
plains the collapse of breakup curves at low We. The line B–B
E1(1.48) shown in the inset of Figs. 14 and 15 joins points where the
axis-switching starts to become clearly visible for elliptical jets
E2(1.94)
0.4 and it continues to be visible in the region lying to the right
E3(2.97)
of the line B–B. For the region lying to the left of the line B–B,
E4(3.85) the axis-switching is not clearly visible. Thus the transition of
E5(5.84) elliptical jets from no axis-switching mode to axis-switching
0.2 mode occurs between the line A–A and the line B–B. Fig. 16(a)
0 7 14 21 28 35 and (b) portray the flow behavior of elliptical water jets and
1/2
We water–glycerol mixture jets, respectively, in the transition re-
pffiffiffiffiffiffiffi gime with different We. In the case of water jets given in
Fig. 12. Variation of Dmax/Do,major with We. (a) Elliptical water jets, and (b)
Fig. 16(a) axis-switching starts to become visible in the forth im-
elliptical water–glycerol mixture jets issuing from various orifices.
age. The fifth image lies in a flow condition which is immedi-
ately beyond the transition regime. The respective images for
water–glycerol mixture jets are the third and forth of
Fig. 16(b). A close observation of the breakup curves given in
2.0 the inset of Fig. 14 reveals that, as We increases, breakup curves
E4 (3.85)
for elliptical jets start deviating from the line A–A and they soon
Exponentially reach a peak and encounter a change in slope. The peaks, here-
decaying fit after, will be called as critical points. As seen in the inset of
Fig. 14, the critical point and the associated slope change in
1.6 the breakup curves of elliptical jets occur between the lines A–
A and B–B. Thus the transition regime of elliptical jets from no
Dmax/Deq

axis-switching mode to axis-switching mode is characterized

by a change of slope in their breakup curves. The change of
slope is more conspicuous in the case of water jets than in the
1.2 case of water–glycerol mixture jets whose breakup curves are gi-
ven in Fig. 15. However, Fig. 17 reveals that the change of slope
is indeed present in the case of water–glycerol mixture jets.
Within the transition regime, after the critical points, the break-
up lengths of elliptical jets are significantly lesser than that of
0.8 the equivalent circular jet. This trend continues after the line
0 2 4 6 8 B–B till the line C–C in Figs. 14 and 15. Lower breakup lengths
Axis-switching number of elliptical jets indicate that elliptical jets are more unstable
Fig. 13. Variation of Dmax/Deq with axis-switching number for water–glycerol
than the equivalent circular jet. Within this zone of enhanced
mixture jets discharging from the elliptical orifice E4. The solid curve shows the instability, increasing aspect ratio for a given We, destabilizes
exponential decay of Dmax/Deq with axis-switching number. water jets further as seen in Fig. 14. However, such a destabiliz-
16 T.V. Kasyap et al. / International Journal of Multiphase Flow 35 (2009) 8–19

160

120

160
A
o
L /D & L /D
b

80 120
B C

Lb/Deq & Lb/Do

eq

80
b

40 40

A B C
0
0 5 101/2 15 20
We
0
0 20 40 60 80 100
1/2
We
pffiffiffiffiffiffiffi
Fig. 14. Breakup curves (variation
pffiffiffiffiffiffiffi of the nondimensionalized breakup length, Lb/Do and Lb/Deq against We) for water jets. The inset shows the magnified view of the breakup
curves in the lower range of We, 0–20. The symbols used in the plot denote the following: j, C (1); h, E1 (1.48); , E2 (1.94); M, E3 (2.97); O, E4 (3.85); and }, E5 (5.84).

330 up curves of elliptical jets may be due to the onset of axis-

C switching. To confirm this hypothesis further we made attempts
A to understand the effect of ambient medium on the breakup of
elliptical liquid jets at flow conditions near the critical point. Note
that the critical point occurs even in the case of circular jets and
B most of the time it is caused by the ambient medium (Phinney,
220
1973). Earlier studies of circular jets show that the effect of ambi-
ent medium on the jet breakup is negligible if the gas Weber num-
Lb/Deq & Lb/Do

ber, Weg < 0.4 (Pan and Suga, 2006; Reitz, 1978). The gas Weber
number, Weg is defined by the relation

C C(1) qg U 2o D
110 Weg ¼ ; ð5Þ
E1(1.48) r
E2(1.94)
where qg is the density of the ambient gas. The values of Weg
A E3(2.97)
B evaluated at the critical conditions for the elliptical jets studied
E4(3.85) here lie in the range of 0.0024–0.045 which is significantly lower
E5(5.84) than 0.4 and hence rules out the role of ambient medium at the
0 critical point seen in the breakup curves of elliptical water jets
0 10 20 30 40 as well as water–glycerol mixture jets. The critical point on the
1/2
We breakup curve earmarks the onset of the enhanced instability
Fig. 15. Breakup curves found in elliptical jets. Note that within this regime of enhanced
pffiffiffiffiffiffiffi (variation of the nondimensionalized breakup length, Lb/Do
and Lb/Deq against We) for water–glycerol mixture jets. instability, axis-switching is dominantly perceivable on both
water and water–glycerol mixture jets. These findings reveal that
the effect of axis-switching on elliptical jets is to destabilize
them further since it reduces their breakup length significantly
ing effect of orifice aspect ratio has not been observed with rel-
compared to that of the equivalent circular jet under the same
atively more viscous water–glycerol mixture jets (refer Fig. 15).
flow conditions.
For further analysis we consider the We at which the slope
The influence of elliptical orifice geometry on the breakup of
change takes place in the breakup curve, referred as the critical
liquid jet ceases to exist beyond the line C–C in Figs. 14 and 15
point, and the We where axis-switching starts to become visible.
pffiffiffiffiffiffiffi pffiffiffiffiffiffiffi though axis-switching is seen at these flow conditions on both
These are denoted by We and We , respectively. Both elliptical water jets and water–glycerol mixture jets. The increas-
crit min
pffiffiffiffiffiffiffi pffiffiffiffiffiffiffi
ing trend of the breakup length of the circular water jet with We
We and We are found to be nearly matching in the
crit min in Fig. 14, shown by filled symbols, agrees with the previous
case of water jets. Details of this method of analysis of the breakup observations (Grant and Middleman, 1966). Photographs of ellip-
curves can be found in Kasyap et al. (2008). A similar analysis on tical water jets at these flow conditions (Fig. 6) show consider-
water–glycerol mixture jets revealed remarkable agreement be- able ruffling on the jet surface which suggests that the flow is
pffiffiffiffiffiffiffi pffiffiffiffiffiffiffi
tween We and We . In both cases, most of the time, no longer laminar. Estimated values of Re of water jets at these
crit min
flow conditions range from 7500 to 61,000 which shows that
an exact numerical equality was not found. This is indeed so since
pffiffiffiffiffiffiffi turbulence is weakly developed. The values of Weg estimated
the determination of the We , where axis-switching starts to at these flow conditions lie in the range 0.26–12 which indicates
min
become visible, is more or less subjective. The above analysis sug- a strong influence of ambient medium on the jet breakup (Pan
gests that the critical point and the associated change in the break- and Suga, 2006; Reitz, 1978). The effect of axis-switching on
 T.V. Kasyap et al. / International Journal of Multiphase Flow 35 (2009) 8–19 17

pffiffiffiffiffiffiffi
Fig. 16. Images of elliptical water and water–glycerol mixture jets in the transition regime. (a) Water jet issuing from E3 with We: (i) 1.49, (ii) 2.40, (iii) 2.77, (iv) 3.28, and
pffiffiffiffiffiffiffi
(v) 3.94. Image (v) was taken at a flow condition lying immediately after the transition regime. (b) Water–glycerol mixture jets issuing from the elliptical orifice E2 with We:
(i) 1.91, (ii) 2.67, (iii) 3.71, and (iv) 5.53. The flow condition for image (iv) lies immediately after the transition regime. All other details remain same as in the caption of Fig. 4.

250
to merge together after the line C–C. All elliptical water–glycerol
mixture jets exhibit a peak after the line C–C as seen in Fig. 15
and asymmetric breaking was observed in these flow conditions.
200
Transition regime

Note that at larger distances from the orifice exit, elliptical

water–glycerol mixture jets behave effectively like circular jets
due to the damping of the axis-switching process. Ambient med-
Lb/Deq & Lb/Do

150 ium causes these degenerated elliptical jets to breakup in an

asymmetric manner, like in the case of circular jets at high We
conditions. Gas Weber number Weg evaluated in this regime
100 ranges from 0.65 to 1.66 which clearly indicates the influence
of ambient atmosphere near the peak.
Starting of visible axis switching An interesting observation on water jets issuing from elliptical
50 Critical point orifices of larger aspect ratios is that at larger We, transverse waves
C(1) develop in the vicinity of the first switching observed from the ma-
E2(1.97) jor axis viewing plane of the elliptical orifice. This is illustrated in
0 Fig. 18, which shows the images of water jets discharging from
0 5 1/2 10 15 the elliptical orifice E5 with orifice aspect ratio 5.84, at various
We We conditions. The regions surrounded by rectangles marked in
Fig. 17. Breakup curve for the water–glycerol mixture jet issuing from the elliptical the images given in Fig. 18 indicate the presence of transverse
orifice E2 and circular orifice C to illustrate the change of slope at the critical point. waves on the jet surfaces. Note that, at the same axial location in
the minor axis images, no transverse waves are seen. A comparison
the jet breakup may get undermined by the influence of ambient of the jet width at these locations between the major and minor
medium and jet turbulence which can cause the breakup curves axis images indicates that, the jet width is significantly larger in
18 T.V. Kasyap et al. / International Journal of Multiphase Flow 35 (2009) 8–19

Fig. 18. Images of water jets discharging from the elliptical orifice E5 at various We conditions. The values of We are: (a) 3310, (b) 4300, and (c) 6450. Transverse waves are
observed only in the major axis plane images and their locations are marked by rectangles. All other details remain same as in the caption of Fig. 4.

the minor axis plane than in the major axis plane. Thus the ellipti- 8
1/2
cal jet effectively behaves like a thin sheet whose plane is perpen- Wecrit of water jets
dicular to the major axis plane of the elliptical orifice. This liquid 1/2
sheet then tends to develop transverse waves similar to waves Wecrit of water-glycerol mixture jets
seen on plane liquid sheets.

4. Discussion: role of axis-switching 6

The experimental results of water and water–glycerol mix-

1/2

ture jets given in the previous sections clearly show that the on-
We

set of axis-switching process results in a significant reduction of

the breakup length of elliptical liquid jets compared to that of
the equivalent circular liquid jets. This destabilizing effect of 4
axis-switching on liquid jets can be understood further via the
following arguments. It is known that the free surface of a liquid
jet always seek configurations with minimum surface energy for
a given volume. This makes a cylindrical liquid jet or a liquid
ligament spontaneously disintegrate into spherical drops as the
spherical shape has the minimum surface area. While comparing 2
to the equivalent circular liquid jet, the surface area of elliptical 0 2 4 6 8
Orifice aspect ratio
liquid jet with axis-switching segments is larger and such jets
pffiffiffiffiffiffiffi
tends to disintegrate faster. This is manifested in the breakup Fig. 19. Variation of We with orifice aspect ratio for water and water–
crit
curves as the zone of enhanced instability bounded by the lines glycerol mixture jets.
A–A and C–C in Figs. 14 and 15. Within this zone of enhanced
instability, an increase in orifice aspect ratio causes further
destabilization of the elliptical liquid jets with relatively lower
viscosity. Since an increase in orifice aspect ratio results in an the lack of destabilizing effect of aspect ratio in the case of ellip-
increase in amplitude of axis-switching as illustrated in tical water–glycerol mixture jets. The stabilizing effect of viscous
Fig. 11(a) and (b), surface energy of the jet increases and desta- damping is illustrated through
pffiffiffiffiffiffiffiFig. 19 which is a superimposed
bilizes the jet further. However, the presence of viscosity damps plot of the variation of We of water and water–glycerol
crit
axis-switching process quickly for viscous liquids like water– mixture jets with orifice aspect ratio. Fig. 19 clearly reveals that
glycerol mixture and all elliptical jets reduce to a more or less the effect of viscosity is to delay the critical points in the break-
circular one after a short distance from the orifice exit. As a re- up curves.
sult, the effective increase in the jet surface area due to axis-
switching process for viscous liquids is significantly less than 5. Conclusions
that of liquids with relatively low viscosity. This renders axis-
switching process in elliptical water–glycerol mixture jets too An experimental study on liquid jets discharging from elliptical
weak to cause a sharp change of the slope of the breakup curve orifices of various aspect ratios has been reported. Experiments
at critical points. In fact breakup curves for elliptical water–glyc- were carried out with water and water–glycerol mixture of 1:1
erol mixture jets show a positive slope even after the critical by volume as working fluids. Identical experiments were con-
point (refer Fig. 17) and elliptical jets break axisymmetrically ducted with the circular orifice of same area of cross section as that
at these flow conditions like circular jets operating in the of the elliptical orifices. Jet velocity at the orifice exit was esti-
Rayleigh regime as seen in Fig. 7. Nevertheless, axis-switching mated from the measurement of liquid volume flow rate and the
indeed enhances their instability and causes their breakup orifice exit area. The simulated flow conditions were extended to
lengths to be reduced less than that of thep equivalent circular Weber number (We) as high as 104. Photographic techniques were
ffiffiffiffiffiffiffi
jet under the same flow conditions. Though We is affected employed to extract details on the flow behavior and breakup pro-
crit
by the aspect ratio, the presence of viscous damping results in cess of liquid jets. A systematic analysis was carried out by com-
 T.V. Kasyap et al. / International Journal of Multiphase Flow 35 (2009) 8–19 19

paring the experimental results obtained for the elliptical orifices Bechtel, S.E., Lin, K.J., Forest, M.G., 1988b. On the behavior of viscoelastic free jets
with elliptical cross section. J. Non-Newton. Fluid Mech. 27, 87–126.
with that of the circular orifice.
Bechtel, S.E., Cooper, J.A., Forest, M.G., Petersson, N.A., Reichard, D.L., Saleh, A.,
At very low We, liquid jets discharging from the elliptical ori- Venkataramanan, V., 1995. A new model to determine dynamic surface tension
fices behave like a circular jet by limiting the influence of elliptical and elongational viscosity using oscillating jet measurements. J. Fluid Mech.
orifice geometry to the immediate vicinity of the orifice exit plane. 293, 379–403.
Bidone, G., 1829. Experiences sur la forme et sur la direction des veines et des
As We increases, the elliptical cross section of liquid jet switches its courants d’eau lances par diverses ouvertures. Impremerie Royale, Turin, pp. 1–
major and minor axes and the elliptical liquid jet develops axis- 136.
switching segments as it flows downstream. The onset of axis- Blaisot, J.B., Adeline, S., 2000. Determination of the growth rate of instability of low
velocity free falling jets. Exp. Fluids 29, 247–256.
switching process in an elliptical liquid jet is found to be a gradual Bogy, D.B., 1979. Drop formation in a circular liquid jet. Annu. Rev. Fluid Mech. 11,
process occurring within a transition regime covering a range of 207–228.
We values. The axis-switching process is visibly seen on elliptical Bohr, N., 1909. Determination of dynamic surface tension by the method of jet
vibration. Philos. Trans. Roy. Soc. Lond. 209, 281–317.
liquid jets at flow conditions beyond the transition regime. How- Boussinesq, J., 1877. Mem. Acad. Sci. Paris 23.
ever, at very high We, jet turbulence and ambient gas medium de- Chandrasekhar, S., 1961. Hydrodynamic and Hydromagnetic Stability. Clarendon
velop large scale distortions on the jet surface which undermine press, Oxford.
Geer, J.F., Strikwerda, J.C., 1980. Vertical slender jets. J. Fluid Mech. 101, 53–63.
the effect of axis-switching process. The axis-switching process Geer, J.F., Strikwerda, J.C., 1983. Vertical slender jets with surface tension. J. Fluid
of elliptical liquid jets was characterized in the present study by Mech. 135, 155–169.
obtaining the measurements of axis-switching wavelength of the Grant, R.P., Middleman, S., 1966. Newtonian jet stability. AIChE J. 12, 669–678.
Green, A.E., 1977. On the steady motion of jets with elliptical sections. Acta Mech.
jet and maximum amplitude of axis-switching. The wavelength
26, 171–177.
of axis-switching is found to be independent of the viscosity of Haenlien, A., 1932. Disintegration of a liquid jet. NACA TM 659.
the liquid as well as the aspect ratio of the orifice and it varies sig- Hoyt, J.W., Taylor, J.J., 1978. Elliptical water jets. AIAA J. 16, 85–87.
nificantly along the axis of the jet. The amplitude of axis-switching Kasyap, T.V., Sivakumar, D., Raghunandan, B.N., 2008. Breakup of liquid jets
emanating from elliptical orifices at low flow conditions. Atomization Spray 18,
is strongly dependent on the aspect ratio of the orifice as well the 645–668.
viscosity of the liquid. The effect of viscosity is to cause the decay Lefebvre, A.H., 1989. Atomization and Sprays. Taylor & Francis, London.
of axis-switching along the jet and this decay is found to increasing Lin, S.P., Reitz, R.D., 1998. Drop and spray formation from a liquid jet. Annu. Rev.
Fluid Mech. 30, 85–105.
with the aspect ratio of the orifice. The axial variation of the axis- Magnus, G., 1855. Hydraulische untersuchungen. Ann. Phys. Chem. 95, 1–59.
switching wavelength and the non-exponential decay of amplitude McHale, R.M., Nurick, W.H., Clapp, S.D., 1971. Injector design criteria using
near the orifice exit advocates for rigorous studies to understand noncircular orifice geometry. J. Spacecraft Rockets 8, 408–410.
Pan, Y., Suga, K., 2006. A numerical study of the breakup process of laminar liquid
the role played by orifice exit conditions on the behavior of ellipti- jets into a gas. Phys. Fluids 18, 1–11.
cal liquid jets. Phinney, R.E., 1972. Stability of a laminar viscous jet – the influence of initial
Comparison of breakup curves for elliptical liquid jets with that of disturbance level. AIChE J. 18, 432–434.
Phinney, R.E., 1973. Stability of a laminar viscous jet – the influence of an ambient
the equivalent circular liquid jet reveals that the breakup length of gas. Phys. Fluids 16, 193–196.
elliptical liquid jets is significantly less than that of the circular jet Plateau, J., 1873. Statisque experimentale et theorique des liquids soumis aux seules
within a range of We at which the former exhibits visible axis- forces moleculaires. Gauthier-Villars, Paris.
Rayleigh, L., 1879. On the capillary phenomena of jets. Proc. R. Soc. Lond. 29, 71–97.
switching process. This unveils the destabilizing role of axis-switch-
Rayleigh, L., 1890. On the tension of recently formed surfaces. Proc. R. Soc. Lond. 41,
ing process on elliptical liquid jets. At high We, the elliptical orifice 281–287.
geometry lacks its influence on the macroscopic breakup process of Rayleigh, L., 1945. The Theory of Sound, vol. 2. Dover, New York.
liquid jets. The destabilizing effect of axis-switching is found to be Reitz, R.D., 1978. Atomization and other breakup regimes of a liquid jet, Ph.D.
Thesis, Princeton University, NJ.
weaker in the case of relatively more viscous liquid jets like water– Savart, F., 1833. Memoire sur la constitution des veines liquides lancees par des
glycerol mixture jet. Additional studies are needed to quantify the orifices circulaires en mince paroi. Ann. Chim. 53, 337–386.
details of surface atomization and droplet formation processes. Snyder, H.E., Senser, D.W., Lefebvre, A.H., Countinho, R.S., 1989. Drop size
measurements in electrostatic paint sprays. IEEE Trans. Ind. Appl. 25, 720–727.
Sterling, M., Sleicher, C.A., 1975. The instability of capillary jets. J. Fluid Mech. 68,
References 477–495.
Taylor, G.I., 1960. Formation of thin flat sheets of water. Proc. Roy. Soc. Lond. 259A,
Bechtel, S.E., 1989. The oscillation of slender elliptical inviscid and Newtonian jets: 1–17.
effects of surface tension, inertia, viscosity and gravity. Trans. ASME: J. Appl. Weber, C., 1931. Zum zerfall eines flüssigkeitsstrahles. Z. Angrew. Math. Mech. 11,
Mech. 56, 968–974. 136–154.
Bechtel, S.E., Forest, M.G., Holm, D.D., Lin, K.J., 1988a. One-dimensional closure Yang, C., Leong, K.C., 2002. Influences of substrate wettability and liquid viscosity on
models for three-dimensional incompressible viscoelastic free jets: von Karman isothermal spreading of liquid droplets on solid surfaces. Exp. Fluids 33, 728–
flow geometry and elliptical cross section. J. Fluid Mech. 196, 241–262. 732.