Fuel composition effect on the electrostatically-driven atomization of bio-butanol containing engine fuel blends

Energy Conversion and Management 60 (2012) 28–35

Contents lists available at SciVerse ScienceDirect

Energy Conversion and Management journal homepage: www.elsevier.com/locate/enconman

Fuel composition effect on the electrostatically-driven atomization of bio-butanol containing engine fuel blends Maria S. Agathou, Dimitrios C. Kyritsis ⇑ Department of Mechanical Science and Engineering, University of Illinois at Urbana-Champaign, Urbana IL 61801, United States

a r t i c l e

i n f o

Article history: Available online 20 March 2012 Keywords: Bio-butanol Bio-fuel blends Engine applications Electrospray

a b s t r a c t Electrostatically assisted sprays of three fuel blends of bio-butanol, ethanol and heptane were studied experimentally. Mixture composition was selected such that electrical conductivity and surface tension were kept constant for all three mixtures. In this manner, the effect of fuel composition was investigated in a context that broadens the classical focus on the effective decrease of surface tension through the action of electrostatic fields. High-speed visualization was used in order to capture e-spray morphology. In addition, probability density functions of the e-spray droplet size and velocity were measured using Phase-Doppler Anemometry for a variety of flow rates and applied voltages. The dependence of droplet average diameter on both flow rate and applied electric field was highlighted. Polydisperse sprays were observed which was rationalized through the calculation of droplet Weber numbers that pointed to the possibility of a secondary droplet break-up. Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction Recent research revealed the possibility of producing butanol efficiently from agricultural sources with the use of clostridia through fermentation processes [1–3]. The feasibility of butanol as a potential alternative to currently used bio-ethanol is particularly attractive because of the higher energy content of butanol. Specifically, the energy densities of butanol and ethanol are 36.4 and 24.8 MJ/kg respectively, whereas the energy content of gasoline is 44.9 MJ/kg. An additional attractive feature of bio-alcohols is that they contain the –OH bond in their molecules, which favors the electrostatic atomization process because it is strongly polarized. Hence, the electrospray technology is particularly appropriate for alcohols and alcohol-containing blends. For relatively low conductivity liquids, there is a regime of stable electrospray operation where the e-spray consists of three major parts [4,5]. First, the conical meniscus, which is formed at the outer surface of the nozzle and is also known as ‘‘Taylor cone’’ [6]. The Taylor cone is followed by the liquid ligament, which further downstream breaks up due to Rayleigh instability into a spray of nearly uniform-size droplets [7]. After the droplets are generated, they repel each other due to Coulombic forces, forming the droplet ‘fan’. This mode of operation is also known as the ‘‘cone-jet’’ regime and can be potentially advantageous for applications that relate to power generation. This is because in this regime the electrospray can then produce practi⇑ Corresponding author. Tel.: +1 217 333 7794. E-mail address: [email protected] (D.C. Kyritsis). 0196-8904/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.enconman.2011.09.021

cally monodisperse sprays the droplet size of which is controlled by the mass flow rate per spray [4,5]. For, butanol [8–10], gasoline [11,12] and gasoline-ethanol [13] sprays, promising results regarding electrostatically assisted atomization have already been reported [8–14], whereas butanol and bio-fuel blends were investigated in engine environments [15–20]. Experimental and computational studies addressing the possibility of automotive application of the electrospray technology were presented in [21–24]. The potential of electrostatic effects for combustion applications was pointed out in [25] and in [26,27] it was shown that electric charge could be injected in liquids of relatively low conductivity such as hydrocarbons with the use of sharp electrodes. The fluid mechanics of droplet emission from liquid cones was analyzed in [28]. The morphology of the electro-sprays was extensively studied by Cloupeau and PrunetFoch, both in the ‘‘cone-jet’’ mode regime and in other modes of operation [29,30]. Electrical conductivity, is one of the controlling properties in the formation of the e-spray and determines to a significant extent the Coulombic forces that will atomize the liquid. The electrospray forms as a result of the competition between these Coulombic forces with the cohesive forces acting on the surface of the liquid, because of surface tension. Thus, surface tension constitutes the second fundamental property for the creation of an electrospray. Our purpose in this paper was to investigate experimentally the electrosprays of three alcohol-containing fuel blends with the same electric conductivity and surface tension in order to check the extent to which it is really just these two properties that determine e-spray behavior. High speed visualization was used in

29

M.S. Agathou, D.C. Kyritsis / Energy Conversion and Management 60 (2012) 28–35

Nomenclature English d G MW N [P] Q T u UAVE V W We x y z

e e0 e1

molecular diameter (m) binary interaction parameter molecular weight number of carbon atoms Parachor of component i volumetric flow rate (m3/s) temperature (K) mobility of the ion (m2/V s) average velocity (m/s) voltage (V) intermediate parameter in G calculation Weber number mol fraction in liquid phase mol fraction in vapor phase order of magnitude estimate

k

l v

q r RD

Constants k Boltzmann’s constant (1.3806503  1023 m2 kg/s2 K) NA Avogadro’s number (6.0221415  1023) e electron charge (1.60217646  1019 C) Greek a

Subscripts 1 component 1 in a binary mixture 2 component 2 in a binary mixture A Avogadro i mixture component i j mixture component j L liquid m mixture V vapor Superscripts ⁄ non-dimensional parameter

molecular radius (m)

tandem with droplet size and velocity measurements, using PhaseDoppler Anemometry, for a broad range of flow rates and applied voltages. Particular focus is given to the determination of the fundamental causes for possible differences among the three blends under investigation. Other properties, namely viscosity and density play a secondary role in the e-spray formation and were also considered. Finally, several modes of operation for the established sprays were recorded and are presented in the sections that follow. 2. Experimental methodology

A schematic of the experimental configuration is presented in Fig. 1. The apparatus consisted of a syringe pump that fed the injector with the liquid fuel under consideration and controlled the flow rate. The injector was charged at a potential on the order of several kV provided by a high voltage power supply. The charged spray was collected at an electrical ground in the form of a steel mesh placed several mm from the capillary exit. The liquid was drained in a liquid beaker placed beneath the electrical ground. Three mixture combinations of butanol, ethanol and heptane were considered (conductivity enhancer Stadis was not added to heptane), which had the same surface tension and electrical conductivity. Each mixture contained two substances. The composition, as well as the calculated physical properties of the created mixtures, are presented in Table 1. The mass fraction of each pure substance in the mixture was selected such that the resulting mixtures had the same electrical conductivity k and surface tension r. 2.2. Calculation of electrical conductivity and surface tension of the mixtures In order to calculate the electrical conductivity of a mixture, the concept of the mobility u of an ion was used, as introduced in [31]:

ze ; 6pla

where ze is the charge of an ion, l is the dynamic viscosity and a is the molecular radius, which is calculated through [31]:

ffi vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u 2 k  T  MW t ; 2a ¼ d ¼ 3l N A  p3

ð1Þ

ð2Þ

where d is the molecular diameter, k is the Boltzmann’s constant, NA the Avogadro’s number and T the temperature. The dynamic viscosity l was approximated using the method of Grunberg and Nissan [32] for a binary mixture:

ln lm ¼ x1 ln l1 þ x2 ln l2 þ x1 x2 G12 ;

2.1. Electrospray experimental set-up

u¼

liquid dielectric constant electric permittivity of the surrounding medium (s/Xm) electric permittivity of the liquid (s/Xm) electric conductivity (1/Xm) dynamic viscosity (kg/ms) kinematic viscosity (m2/s) density (kg/m3) surface tension (kg/s2) total group contributions for Gij

ð3Þ

where G12 is the binary interaction parameter, given in Eq. (4), where the intermediate parameters RD, W and N for the two mixture components i and j, are obtained from tables in [33]:

G12 ¼ RDi  RDj þ W; W¼

0:3161  ðNi  Nj Þ2  0:1188  ðN i  Nj Þ; ðNi þ Nj Þ

ð4Þ ð5Þ

For each mixture, i was selected using priority rules in [33] and j became the second mixture component. The values of Gij for the three mixtures, along with the intermediate parameters from (4) and (5), are given in Table 2. Ni is the number of carbon atoms in i and RD was calculated based on the contributions of a group of bonds in the molecule. For example: RD–CH3 = 0.100 and RD–CH2– = 0.096. After calculating the mobility through (1), the electrical conductivity of the mixture under consideration is given by:

k ¼ z  e  u  NA ;

ð6Þ

In order to calculate the surface tension of the mixtures the Macleod–Sugden correlation was used [34]:

rm1=4 ¼

n X i¼1

½Pi   ðqLm xi  qv m yi Þ;

ð7Þ

30

M.S. Agathou, D.C. Kyritsis / Energy Conversion and Management 60 (2012) 28–35

Fig. 1. Electrospray configuration.

Table 1 Compositions and calculated physical properties for mixtures 1, 2 and 3.

Mixture 1 Mixture 2 Mixture 3

Butanol (% per volume)

Ethanol k (1/Om)

Heptane

Electrical conductivity

Surface tension

r (kg/s2)

m (m2/s)

q (kg/m3)

0.48 0.45 0.00

0.52 0.00 0.70

0.00 0.55 0.30

4.311E04 4.311E04 4.338E04

2.330E02 2.333E02 2.331E02

Kinematic viscosity

Density

1.908E06 1.248E06 1.335E06

7.99E+02 7.38E+02 7.56E+02

Table 2 Gij values and intermediate calculated parameters. Mixture 1

RD N W Gij

Mixture 2 j: Ethanol

i: Butanol

j: Heptane

i: Ethanol

j: Heptane

0.255 4 0.027 0.255

0.027 2 0.615 0.08

0.255 4 1.472 1.165

0.28 7

0.027 2

0.28 7

In the last equation Pi is an intermediate term called the Parachor and its value was obtained from tables in [33]. In these tables, the structural contributions for each bond group in the molecule are found. For the three liquids under consideration: [PBUTANOL] = 205.3, [PETHANOL] = 125.3 and [PHEPTANE] = 311.0. In addition, xi and yi is the mole fraction of i in liquid and vapor phases, whereas qLm and qv m are the liquid-mixture and the vapormixture density respectively. The density of the vapor term was neglected and the equation reduced to:

r

1=4 m

¼ qLm

n X xi  r1=4 i i¼1

Mixture 3

i: Butanol

qLi

! ;

ð8Þ

2.3. Experimental techniques High speed visualization was achieved with a Vision Research Phantom v7.0 unintensified CCD high-speed camera. Three Nikon extension tubes totalling 82.5 mm in length were used to increase magnification along with a Tiffen 52 mm zoom lens. Camera exposure time was set to 96 ls and the zoom and f-stop settings were adjusted to provide a clear view of the entire spray. Laser light from an Oxford Laser Systems LS 20-50 Copper-Vapor laser operating at 510.6 nm and having a frequency of 10 kHz was directed toward the doublet of the concave spherical – cylindrical lens that was used to generate the laser sheet. Both the high-speed camera and the laser were synchronized by the Berkeley Nucleonics 500C042 pulse

generator. To acquire images of a 2-D cross section of the spray, the created laser sheet was aligned so that the top would be directly below the capillary tip and the bottom would be approximately 20 cm below the orifice, so that the whole spray was captured. The thickness of the laser sheet was approximately 1 mm. For spray droplet size and velocity measurements, a Dantec Fiber Phase Doppler Anemometer (PDA) was used. It was powered by a 5 W Ar-ion laser and was run in 1-D PDA mode at 514.5 nm and a 71.4° forward scattering angle. In the PDA software, the refractive index for the mixtures was set to 1.38 [35]. A number of 1000 counts was sufficient for adequate statistics and it was applied to all spray cases investigated here, avoiding high collection times. Droplet size and velocity measurements were performed at a location on the central spray axis 10 mm from the capillary exit. The collection point location is also shown in Fig. 1. Droplet size measurements were obtained at only one location, since the goal remains to investigate the possibility of monodisperse sprays for the liquids under consideration. 3. Results and discussion 3.1. Electrospray phenomenology Imaging results for the three fuels under consideration are presented in Fig. 2 for a flow rate of 10 ml/h. In the figure, each row corresponds to the different mixture, named after Table 1 and each column corresponds to a different applied voltage: 4 kV, 5 kV and

31

M.S. Agathou, D.C. Kyritsis / Energy Conversion and Management 60 (2012) 28–35

Fig. 2. Electrospray imaging results for mixtures 1–3, for a flow rate 10 ml/h.

2000

MIXTURE 1: Single-jet, intermittent MIXTURE 1: Single-jet, continuous MIXTURE 2: Single-jet, continuous

1500

Q*

6 kV. Additional results can be found at [36]. Each of the images was selected to be characteristic of the particular spray-movie it described and it corresponded to the screenshot when most details of the spray structure were apparent. The overall electrospray behavior of mixtures with the same conductivity, the same surface tension, but different composition did reveal some similarities among their actual modes of operation, however, substantial differences were demonstrated as well. In particular, atomization happens either through a ‘‘whip’’-like instability that breaks the liquid jet in the form of long ligaments and is termed kink instability, or through multiple sprays. The multiple jets seem to develop Rayleigh-type instabilities. In the conejet regime, the spray break-up occurs due to Rayleigh instabilities on the liquid ligament that is extracted from a conical meniscus, resulting in uniform size droplets. It is noted, that as long as the instability that will ultimately lead to the breakup of the liquid column and the generation of the spray has not developed fully, i.e. when the voltage is low (4 kV) the sprays exhibit similar behavior. Once the instability occurs (>5 kV), there are occasional differences. The sprays were mostly generated through kink instabilities [30]; hence the electrosprays were unstable and ramified the liquid column into streams of liquid droplets. In comparing these unstable behaviors, it is not reasonable to expect a precise similarity for initial conditions that are very closely similar – this would violate the very definition of instability. The exact regime for each set of conditions was characterized by the same randomness that corresponded to the ways instability could propagate, i.e. unstable behaviors between sprays of different mixtures were as similar as sprays coming from the same mixture.

MIXTURE 3: Single-jet continuous MIXTURE 1: Multi-jet MIXTURE 2: Multi-jet MIXTURE 3: Multi-jet

1000

500

0

0

0.2

0.4

0.6

0.8

1

1.2

V* Fig. 3. Electrospray regimes for mixtures 1–3, based on non-dimensional flow rate Q⁄ and voltage V⁄.

In other words, the fundamental properties in the electrospray creation, namely electrical conductivity and surface tension did influence the onset of the instability but did not determine the exact temporal evolution of the resulting spray pattern. Similarities have to be sought only in the qualitative sense. This explains why the mixtures have increased similarities for the same applied voltage. For example at lower voltages, the results suggest that whipping jets prevail, whereas when the voltage increases (6 kV), multi-jet spraying occurs. The spray modes of operation observed through the high-speed videos were categorized with the help of nondimensional analysis and are presented in the section that follows.

M.S. Agathou, D.C. Kyritsis / Energy Conversion and Management 60 (2012) 28–35

30 25

Percentage [%]

30

Mixture 1 Q=10ml/hr

4.5kV 5kV 5.5kV 6kV 6.5kV

20 15 10

15 10 5

0

5

10

15 20 25 Diameter [µm]

30

35

0

40

Mixture 2 Q=10ml/hr

4.5kV 5kV 5.5kV 6kV 6.5kV

5

10

20 15 10

35

40

4.5kV 5kV 5.5kV 6kV 6.5kV

15 10 5

0

5

10

15

20

25

30

35

0

40

0

5

10

15

20

25

30

35

40

Diameter [µm]

Diameter [µm] 30

30 5kV 5.5kV 6kV 6.5kV 7kV

20 15 10

5kV 5.5kV 6kV 6.5kV 7kV

Mixture 3 Q=30ml/hr

25 20

Percentage [%]

Mixture 3 Q=10ml/hr

25

Percentage [%]

30

20

5

15 10 5

5 0

15 20 25 Diameter [µm]

Mixture 2 Q=30ml/hr

25

Percentage [%]

25

Percentage [%]

0

30

30

0

4.5kV 5kV 5.5kV 6kV 6.5kV

20

5 0

Mixture 1 Q=30ml/hr

25

Percentage [%]

32

0

5

10

15

20

25

30

35

0

40

0

5

10

15

20

25

30

35

40

Diameter [µm]

Diameter [µm]

Fig. 4. Distribution of the droplet diameter of mixtures 1–3 for flow rates 10 ml/h and 30 ml/h.

3.2. Non-dimensional controlling parameters



d ¼ According to [37] the non-dimensional droplet diameter is a function of the non-dimensional voltage and flow-rate as well as of the dielectric constant e of the liquid: 

d ¼ f ðQ  ; V  ; eÞ;

ð9Þ ⁄

⁄

⁄

Thus, for a specific liquid, d is a function of V and Q . The formulation of the non-dimensional parameters is [37]:

Q ¼

qkQ ; e  e0  r

ð10Þ

V ¼

V  V0 ; V0

ð11Þ

q  k2  d3 ðe  e0 Þ2  r

!1=3 ;

ð12Þ

In the above relations, q is the density of the liquid, e0 is the electrical permittivity of air and V0 is the onset voltage, which was taken to be 3 kV for this analysis. Conducting a surface fit over the non-dimensional data, an empirical relation was generated that correlated d⁄, V⁄ and Q⁄, for each fuel blend. These expressions, Eqs. (13a–c), were second order polynomials with regression coefficients R2 = 84.72%, R2 = 82.71% and R2 = 79.55%, for mixtures 1–3 respectively. All coefficients of the polynomials are given with 95% confidence bounds. 

d ¼ 4:039  4:039  V  þ 0:003496  Q  þ 4:748  V 2  0:006029  Q   V  þ 2:171e  006  Q 2 ;

ð13aÞ

33

M.S. Agathou, D.C. Kyritsis / Energy Conversion and Management 60 (2012) 28–35

60

35 4.5kV 5kV 5.5kV 6kV 6.5kV

MIXTURE 1_10ml/hr

25

Mixture 1 50 40

20

We

Percentage [%]

30

15

20

10 5 0

30 10ml/hr 15ml/hr 20ml/hr 25ml/hr 30ml/hr 35ml/hr

10

0

1

2

3

4

5

6

0 3.5

Velocity [m/s]

4

4.5

5

5.5

6

6.5

7

7.5

6.5

7

7.5

6.5

7

7.5

Voltage [kV] 35

MIXTURE 2_10ml/hr

4.5kV 5kV 5.5kV 6kV 6.5kV

25

60

20

40

15

30

10

20

5 0

0

1

2

3

4

5

35

0 3.5

MIXTURE 3_10ml/hr

25

4

4.5

5

5.5

6

Voltage [kV]

5kV 5.5kV 6kV 6.5kV 7kV

30

10ml/hr 15ml/hr 20ml/hr 25ml/hr 30ml/hr 35ml/hr

10

6

Velocity [m/s]

60

Mixture 3

50

20

40

15

We

Percentage [%]

Mixture 2

50

We

Percentage [%]

30

10

30

5

20 0

0

1

2

3

4

5

10ml/hr 15ml/hr 20ml/hr 25ml/hr 30ml/hr 35ml/hr

6

10

Velocity [m/s] Fig. 5. Velocity distribution of the three Mixtures for a flow rate of 10 ml/h.

0 3.5

4

4.5

ð13bÞ



d ¼ 2:918 þ 0:8557  V  þ 0:002241  Q  þ 0:2201  V 2  0:001499  Q   V   2:443e  007  Q 2 ;

5.5

6

Voltage [kV]



d ¼ 26:01  25:22  V   0:009935  Q  þ 8:715  V 2 þ 0:004556  Q   V  þ 2:617e  006  Q 2 ;

5

ð13cÞ

In addition, non-dimensional parameters can be used in order to classify flow regimes. Fig. 3 presents an overview of the electrospray phenomenology. Three different electrospray regimes are presented in the figure, namely intermittent single-jet, continuous single-jet and multi-jet, corresponding to the modes observed with the high-speed videos. Notably, all of them are unsteady. The intermittent regime is favored in the region of intermediate values of Q⁄ and V⁄. Notably, this regime appears only for mixture 1. In this plot,

Fig. 6. Effect of voltage on We numbers of the three mixtures.

it can be also seen that multi-jet regimes appear exclusively at high V⁄. At intermediate V⁄ and low Q⁄ the multi-jet regime is favored as well. Finally, at low V⁄ continuous single-jets appear. 3.3. Droplet size measurements The electrospray droplet size distributions of the mixtures under consideration are presented in Fig. 4 as a function of mass flow rate and applied voltage. Results are presented for flow rates 10 and 30 ml/h and five applied voltages between 4 and 7 kV. As presented in Fig. 4, the mixture electrosprays show a polydisperse behavior. The 10% ratio of the standard deviation of the distribu-

34

M.S. Agathou, D.C. Kyritsis / Energy Conversion and Management 60 (2012) 28–35

tions over the mean droplet diameter reported in previous studies [5] is not observed here. The polydispersion revealed by the high-speed movies was verified through the droplet size measurements. For mixtures 1 and 3, the droplet size distribution is almost equally narrow for all applied voltages and it develops around the same average value for most values of voltage. The droplet size shows a higher sensitivity to the electric field in the case of mixture 2, where the distributions for each applied voltage move towards lower average droplet sizes for a constant flow rate. This is similar to the findings of the previous studies [8–10] where the average droplet diameter decreased with increasing voltage. The increase in size dispersion is observed only for mixture 2 and this occurs for decreasing voltage for a constant flow rate. It is noted that mixtures with the same surface tension and electrical conductivity produce different droplet size distributions. This means that, despite their dominant role, these two properties do not determine fully the spray behavior. In addition to electrical conductivity, liquid dielectric constant e is a property that incorporates effects rising from Coulombic forces, thus constituting a controlling factor in the establishment of an electrospray. As presented in Table 1, its value is substantially lower for heptane compared to alcohols and as a result, the formed mixtures have different dielectric constants. This could explain the possible differences at the formation of the e-sprays as well as the increased similarities between mixtures 2 and 3, which both contain heptane. In addition, the mixtures are characterized with different viscosities and densities, properties that influence the cohesive forces developed on the droplet forming at the capillary tip. Hence, the electrical conductivity and surface tension do not assure similar e-spray formation since this also depends on dielectric constant, density and viscosity. 3.4. Velocity measurements Probability density functions of the velocities corresponding to the spray of each blend, were provided using the Phase Doppler Anemometer. These distributions of the established electrosprays for the three mixtures as a function of mass flow rate per spray and applied voltage, are shown in Fig. 5. Results are presented for a flow rate of 15 ml/h and applied voltages between 5 and 7 kV with an increment of 500 V. The non-uniformity observed both through high speed imaging and droplet size measurements, is shown also through the velocity measurements. In the plots, the distributions develop similarly as a function of velocity for the three mixtures and for the various flow rates and they are not always characterized by a single peak. This study is extended to calculate Weber numbers for the above presented cases. We numbers are used to study the aerodynamic effects of the sprays, since they offer useful information on the spray break-up mechanisms and are calculated:

We ¼

q  U 2AVE  d ; r

ð14Þ

where UAVE and d is the droplet average velocity and diameter respectively, m is the kinematic viscosity, q is the density and r the surface tension. As shown in Eq. (14), the We number is a ratio of inertial to surface tension forces and therefore controls the relative importance of inertial and surface tension effects on atomization. In particular, at large We numbers, droplets that have been generated by the Rayleigh instability (which is controlled by surface tension) can break up further because they are intensely shear in the flow field. This phenomenon is termed secondary droplet break-up and will of course destroy the monodispersity of an electrospray. The calculated We numbers as a function of flow rate are presented in Fig. 6 for the three mixtures.

According to [38], droplet break-up occurs in break-up regimes and could be responsible for the polydispersion observed in the previous section since they do not favor the creation of uniform size droplets. The critical Weber number under which monodisperse aerosols are most probably produced is We 6 12. In the current study though, We P 30, for all cases under investigation. Thus, aerodynamic effects could also be responsible for the creation of polydisperse sprays. 4. Summary and conclusions In summary, it can be stated that sprays of alcohol-containing blends are amenable to electrostatic manipulation. However, the monodispersion that was perceived as an important advantage of previous e-spray applications (e.g. analytic chemistry) is practically non-achievable for the range of droplet sizes, velocities and nozzle configurations that should pertain to automotive applications. This was supported by high-speed visualization of electrospray phenomenology and droplet size measurements in three alcohol-containing fuel blends that had the same electrical conductivity and surface tension, and demonstrated a polydisperse behavior. It was also shown that electrical conductivity and surface tension do not determine fully the spray behavior, fact that was rationalized through the increased differences among the three investigated fuel blends. It is conceivable that the different liquid dielectric constants, viscosities and densities of the mixtures, could explain the possible differences at the formation of the e-sprays. Furthermore, non-dimensional analysis was performed to classify flow regimes, as well as to produce empirical correlations, for each of the fuel blends. Finally, velocity distributions were presented and the calculated We numbers revealed the possibility of droplet secondary break-up, thus pointing to an additional source of spray polydispersion. Acknowledgment The authors would like to acknowledge the support of the US Department of Energy through the Graduate Automotive Technology Education (GATE) Center of Excellence at the University of Illinois. References [1] Eseji T, Qureshi N, Blaschek HP. Production of acetone–butanol–ethanol in a continuous flow bioreactor using degermed corn and Clostridium beijernickii. Process Biochem 2007;42:34–9. [2] Tashiro Y, Shinto H, Hayashi M, Baba S, Kobayashi G, Sonomoto KJ. Novel highefficient butanol production from butyrate by non-growing Clostridium saccharoperbutylacetonicum N1–4 (ATCC 13564) with methyl viologen. Biosci Bioeng 2007;104:238–40. [3] Tashiro Y, Takeda K, Kobayashi G, Sonomoto KJ. High production of acetone– butanol–ethanol with high cell density culture by cell-recycling and bleeding. Biotechnol 2005;120:197–206. [4] Tang KQ, Gomez A. Monodisperse electrosprays of low electric conductivity liquids in the cone-jet mode. J Colloid Interface Sci 1996;184:500–11. [5] Gomez A, Tang KQ. Charge and fission of droplets in electrostatic sprays. Phys Fluids 1994;6:405–14. [6] Taylor GI. Disintegration of water drops in an electric field. Proc Roy Soc London 1964;280:383–97. [7] Fernández de la Mora J, Navascues J. Generation of submicron monodisperse aerosols in electrosprays. J Aerosol Sci 1990;21:673–6. [8] Agathou MS, Powell JW, Lee CF, Kyritsis DC. Preliminary experimental study of butanol electrosprays for power generation. Society of Automotive Engineers 2007; SAE Paper 2007-24-0020. [9] Agathou MS, Kyritsis DC. A Comparative experimental study of butanol electrosprays through phase-doppler anemometry, Paper Nr. C10-12, The Meeting of the Central States Section of the Combustion Institute, March 2010, UIUC University, Urbana, IL. [10] Agathou MS, Kyritsis DC. An experimental study of electrostatic sprays of biobutanol at low flow rates, Paper Nr. 2011–205-693, AIAA-ASME2011/49th AIAA Aerospace Sciences Meeting, January 2011, Orlando, FL.

M.S. Agathou, D.C. Kyritsis / Energy Conversion and Management 60 (2012) 28–35 [11] Anderson EK, Carlucci AP, de Risi A, Kyritsis DC. Electrostatic effects on gasoline direct injection in atmospheric ambiance. Atom Sprays 2007;17: 289–313. [12] Anderson EK, Carlucci AP, de Risi A, Kyritsis DC. Experimental investigation of the possibility of automotive gasoline spray manipulation through electrostatic field. Int J Vehicle Design 2007;45:61–79. [13] Anderson EK, Kyritsis DC. Experimental investigation of combustion of electrostatically charged ethanol blended gasoline droplets. The 5th Joint meeting of the US Sections of the Combustion Institute. San Diego, CA, 2007; Paper G11. [14] Agathou MS, Kyritsis DC. An experimental comparison of non-premixed biobutanol flames with the corresponding flames of ethanol and methane. Fuel. 2010. doi: 10.1016. [15] Rakopoulos CD, Dimaratos AM, Giakoumis EG, Rakopoulos DC. Study of turbocharged diesel engine operation, pollutants emissions and combustion noise radiation during starting with bio-diesel or n-butanol diesel fuel blends. Appl Energy 2011;88:3905–16. [16] Rakopoulos DC, Rakopoulos CD, Hountalas DT, Kakaras EC, Giakoumis EG, Papagiannakis RG. Investigation of the performance and emissions of a bus engine operating on butanol/diesel fuel blends. Fuel 2010;89:2781–90. [17] Rakopoulos CD, Rakopoulos DC, Giakoumis EG, Dimaratos AM. Investigation of the combustion of neat cottonseed oil or its neat bio-diesel in a HSDI diesel engine by experimental heat release and statistical analyses. Fuel 2010;89: 3814–26. [18] Rakopoulos CD, Dimaratos AM, Giakoumis EG, Rakopoulos DC. Investigating the emissions during acceleration of a turbocharged diesel engine operating with bio-diesel or n-butanol diesel fuel blends. Energy 2010;35(12):5173–84. [19] Rakopoulos DC, Rakopoulos CD, Kakaras EC, Giakoumis EG. Effects of ethanol– diesel fuel blends on the performance and exhaust emissions of heavy duty DI diesel engine. Energy Convers Manage 2008;49(11):3155–62. [20] Rakopoulos CD, Antonopoulos KA, Rakopoulos DC. Multi-zone modeling of diesel engine fuel spray development with vegetable oil, bio-diesel or diesel fuels. Energy Convers Manage 2006;47(11–12):1550–73. [21] Shrimpton JS. Pulsed charged sprays: application to DISI engines during early injection. Int J Numer Meth Eng 2003;58:513–36. [22] Shrimpton JS, Yule AJ. Characterization of charged hydrocarbon sprays for application in combustion systems. Exp Fluids 1999;26:460–9.

35

[23] Hetrick RE, Parsons MH. Electrospray for fuel injection. Society of Automotive Engineers. 1997; SAE Paper 972987. [24] Thomas ME, DiSalvo R, Makar P. Electrostatic atomization insertion into compression ignition engines. Society of Automotive Engineers. 2002; SAE Paper 2002-01-3053. [25] Thong KC, Weinberg FJ. Electrical control of the combustion of solid and liquid particulate suspensions. Proc Roy Soc Lond 1971;324:201–15. [26] Kim K, Turnbull RJ. Generation of charged drops of insulating liquids by electrostatic spraying. J Appl Phys 1976;47:1964–9. [27] Kelly AJ. The electrostatic atomization of hydrocarbon oils. J Inst Energy 1984;1:312–20. [28] Fernández De La Mora J. The effect of charge emission from electrified liquid cones. J. Fluid Mech. 1992;243:561–74. [29] Cloupeau M, Prunet-Foch B. Electrostatic spraying of liquids in cone-jet mode. J Electrostat 1988;22:135–59. [30] Cloupeau M, Prunet-Foch B. Electrohydrodynamic spraying functioning modes: a critical review. J Aerosol Sci 1994;25:1021–36. [31] Atkins P, de Paula J. Physical chemistry. 8th ed. New York: W.H. Freeman and Co.; 2006. [32] Grunberg L, Nissan AH. The additivity of T 3c for some non-linear molecules. Trans Faraday Soc 1948;44:1013–21. [33] Reid RC, Prausnitz JM, Poling BE. The properties of gases and liquids, 4th ed. New York: McGraw-Hill; 1987. [34] Macleod DB. On a relation between surface tension and density. Trans Faraday Soc 1923;19:38. [35] Lide DR. CRC handbook of chemistry and physics 75th ed. London: CRC Press; 1994. [36] Agathou MS. Bio-butanol fuel atomization and combustion processes [dissertation]. Urbana, IL, USA: University of Illinois at Urbana-Champaign; 2010. [37] Fernández De La Mora J. The effect of charge emission from electrified liquid cones. J Fluid Mech 1992;243:561–74. [38] Pilch M, Erman CA. Use of breakup time data and velocity history data to predict the maximum size of stable fragments for acceleration-induced breakup of a liquid drop. Int J Multiphase Flow 1987;13(6):741–57.