16th Int Symp on Applications of Laser Techniques to Fluid Mechanics
Lisbon, Portugal, 09-12 July, 2012
Influence of Key Biofuels on the DISI Spray Evaporation Rate
by Phase Doppler Anemometry
Tobias Knorsch1,2*, Markus Heldmann1,2, Tim Hagedorn1, Michael Wensing1,2 ,
Alfred Leipertz1,2
1: Department of Engineering Thermodynamics, University of Erlangen-Nuremberg (FAU), Germany
2: Erlangen Graduate School in Advanced Optical Technologies, Germany
* correspondent author: [email protected]
Abstract Biofuels and alternative fuels are increasingly being blended to conventional gasoline fuel to
decrease overall CO2 emissions. The atomization and evaporation of gasoline with bio-components differ
depending on the respective alternative fuel specific physiochemical properties. This work focuses on
estimating the biofuel evaporation rate of gasoline sprays at stratified charge conditions used in modern DISI
engines. One specific spray plume is analyzed in terms of local droplet size distributions. Depending on the
operating conditions different physiochemical properties were found to dominate the atomization and
evaporation processes: For moderate ambient temperature and pressure high-boiling components show a
strong influence on the droplet size distribution found in the sprays studied in an injection chamber.
However, at elevated temperature the behavior changes completely. Due to a high degree of evaporation
taking place, the evaporation cooling effect dominates the droplet size distributions found in different
positions in the sprays. In the center of the spray plume larger droplets are detected resulting from the higher
droplet density in those areas compared to the more dilute outer positions. Smaller droplets as a measure of
progressed evaporation are found there. In low pressure measurements the boiling point is found to dominate
the droplet size. However, at very high gas temperature and pressure, fuel mixtures with higher evaporation
enthalpy show larger droplet size – even if these fuels have a higher boiling point. Overall, it can be stated
that for the droplet evaporation at high level supercharged conditions, the evaporation enthalpy is the
dominating physiochemical property.
1. Introduction
In modern IC engines combustion is increasingly controlled by the spray of the Direct-Injection
Spark-Ignition (DISI) process. The complex mechanisms of unsteady high pressure spray
atomization and evaporation are still not well understood for the complete reaction chain – even for
conventional gasoline. In a context where the bio component of gasoline blends is set to increase by
law over the coming years, there is an urgent need to understand the spray mechanisms of both
gasoline and blends of fossil fuels with alternative fuels. That requires a detailed analysis of the
processes which atomize fuels within some milliseconds into hundred thousands of droplets and
simultaneously change their temperature by hundreds of degrees.
The complexity of the spray processes of ordinary gasoline fuels is most often simplified by using
single-component surrogate fuels which do not represent realistic fuel behavior, as laser diagnostic
detection of different fuel fractions in a realistic multi component mixture is complex [1]. Of
current special interest are blends of biofuels that contain ethanol, n-butanol and isobutanol. Each of
these blends shows significant differences in physiochemical properties which change the
atomization and evaporation rate. For fuel stratification conditions or injection during catalyst
heating with late injection timing, an inhomogeneous temperature distribution throughout the
cylinder and the spray occurs. This determines local inhomogeneities in the evaporation ratios and,
therefore, mixture formation and ignitability. Spatial-resolved information of the droplet size during
-1-
16th Int Symp on Applications of Laser Techniques to Fluid Mechanics
Lisbon, Portugal, 09-12 July, 2012
and after injection is required to locally determine the evaporation rate and evaporation cooling in
the spray as well as the affinity to pre-ignition or soot formation. The droplet sizes are a measure for
fast conversion of the fuel into the vapor phase. Small droplet sizes are easier to evaporate as the
respective surface is a quadratic function of the droplet radius. For a fast atomization and
momentum exchange between the liquid spray and the surrounding gas, the hot gas entrainment can
be enhanced. The degree of gas entrainment is the key to rapid mixing of this ambient gas and the
gasified fuel for achieving a proper ignitable mixture with low raw emissions.
In general the spray plumes of DISI injectors cool off the surrounding ambient gas during
evaporation whereby the spray cone centers are typically expected to show lower temperatures due
to the higher droplet density. In outer regions the spray density is lower due to stronger air
entrainment. The spatial distributions of droplet diameters depend on the injector type and distance
to nozzle tip as well as specific initial fuel temperature and pressure, the fuel quantity per injection,
gas pressure, evaporation enthalpy, and local concentrations and temperatures. This work focuses
on 0-D/2-D optical measurements of biofuels that are likely to be relevant in the future, and their
influences on the spray atomization of the mainly applied multi-hole solenoid injectors.
2. Measurement Techniques and Comparison Strategy
Spray measurements are carried out using an optically accessible high temperature / high pressure
spray chamber and applying a setup with 0-D/2-D optical measurement techniques (figure 1, right).
To study the general spray propagation we applied different Mie light scattering methods. First,
planar Mie scattering measurements using a frequency-doubled Nd:YAG laser (532 nm, Quantel
Brilliant B) of the liquid phase using gasoline E10 as fuel are carried out to analyze the liquid spray
distribution and determine relevant 0-D positions for the subsequent phase Doppler anemometry
(PDA) measurements. The determination of 0-D PDA positions based in 2-D Mie scattering images
proves to be a strong tool to explain the complex phenomena in dense sprays.
The PDA system (Dantec HiDense P80) is supplied by an Ar* laser (9W, 488 nm/ 514.5 nm,
Coherent Innova 300). The complete setup is presented in figure 1. Due to dense spray close to the
nozzle tip subsequently both the couple in laser power for the Ar* laser partial beams of the PDA
emitting probe and the Doppler signal to be gained by the receiving probe become poorer. With
respect to this finding, the minimum PDA measurement volume distance to the nozzle tip will be
set to 15 mm, the second PDA plane is set to 30 mm for estimating the droplet atomization and thus
estimate evaporation rate at positions where the momentum exchange of the droplets with the
ambient gas is largely progressed.
-2-
16th Int Symp on Applications of Laser Techniques to Fluid Mechanics
Lisbon, Portugal, 09-12 July, 2012
Figure 1: Experimental setup with PDA measurement axis along one spray plume
For the droplet size and velocity measurements, gasoline free of aromatic compounds (Aspen) is
examined as a reference fuel and compared to two three-component biofuel mixtures, ‘3K’ with
RON46 [1] (n-hexane 35% by volume, isooctane 45%, n-decane 20%) and ‘3ZK’ with RON95, (nhexane 10% by volume, ethanol 75%, n-butanol 15%). The model fuels ‘3K’ and ‘3ZK’ and their
boiling points are presented in figure 2 and table 1. As single component fuels isooctane as well as
the pure biofuels n-butanol and ethanol are tested. The spray measurements are conducted in an
optical-accessible high pressure / high temperature spray chamber under variation of injection
conditions, thus, all modern DISI conditions can be investigated (figure 1).
Figure 2: Boiling point diagram of gasoline [2] with indicated boiling points of the fuel components
for ‘3K’ (blocks) and ‘3ZK’ (dashed/arrows)
First, the suitability of this nozzle type for early and late injection times at low and increased
ambient pressure and temperature are tested. In this work, the investigation of cold start split
injection (20 MPa) at low, increased and very high cylinder pressure and temperature (0.04 MPa0.8 MPa and 293K-673K, see table 2) is of particularly interest. Furthermore, measurements of the
impact of different physiochemical properties on the evaporation rate are carried out. Figure 2
shows all tested fuels and their physiochemical properties whereby the focus of this work is the
comparison of the role of boiling point vs. evaporation enthalpy for different boundary conditions.
The injection test chamber in which the spray measurements are carried out (figure 1) is heated by a
-3-
16th Int Symp on Applications of Laser Techniques to Fluid Mechanics
Lisbon, Portugal, 09-12 July, 2012
constant air flow which scavenges the spray from one injection to another (flow velocity <0.2 m/s),
the injection repetition is 1 Hz. The fuel is conditioned to 353K to simulate cold and hot engine
conditions.
Table 1: Investigated fuel mixtures and fuel components with key physiochemical parameters
at ambient conditions (20°C / 1 bar) [5-7]
fuel
isooctane
n-butanol
ethanol
3K
3ZK
gasoline E10
n-hexane
n-decane
boiling
point / °C
99
118
78
69-174
69-118
~46-207
69
174
RON /
100
96
109
~46
~95
95
19
-41
kin. viscosity /
10-6m²/s
0.79
3.74
1.42
~0.57
0.46
1.9
evap. enthalpy/
kJ/kg
305.4
637.7
938.2
334.7
844.6
~350
363.6
352.5
density/
g/cm³
0.70
0.80
0.80
0.67
0.79
~0.76
0.66
0.72
Table 2: Applied conditions
OP
#1
#2
#3
#4
#5
pGas /
MPa
0.04
0.04
0.56
0.80
0.80
TGas/
K
293
373
473
473
673
pFuel/
MPa
20
20
20
20
20
Τ Fuel/
K
353
353
353
353
353
tinj./
ms
1
1
1
1
1
3. Results and Discussion
For the droplet sizes, five different PDA measurement positions through one specific spray plume
give information about the droplet size from the spray outer region, spray plume center and spray
intermediate region. The droplet size distributions are presented in spatially and temporally resolved
diagrams for droplet size d10 and d32 (SMD) and velocity. Furthermore, bar diagrams are carried out
to demonstrate the difference of two different boundary conditions for each fuel.
First, a variation of the chamber temperature from 293K to 373K (table 2: OP#1 vs. OP#2) is
carried out. The chamber pressure remains constant at 0.04 MPa. Figure 3 shows the droplet size for
OP#1 (blue bars) and OP#2 (green bars) on each of the 15 mm and 30 mm plane. The values for all
tested fuels are values of Sauter mean diameter (SMD) averaged over all five measurement
positions through one spray plume. The drop size differences resulting from the gas temperature
difference become visible as differences between the bars. As the gas pressure is 0.06 MPa below
ambient pressure, the boiling points of the fuels are much lower. Applying these two different
temperatures, non-flashboiling conditions for OP#1 are present while for OP#2 flash-boiling
conditions result. The flashboiling phenomenon is known to enhance evaporation, however, the
targeting may change in comparison with non-flashboiling conditions [4]. All fuels with exception
of n-butanol show a decrease in droplet size for the higher gas temperature case at constant
measurement positions in the spray. On the 15 mm plane, this effect for n-butanol is even reversed.
-4-
16th Int Symp on Applications of Laser Techniques to Fluid Mechanics
Lisbon, Portugal, 09-12 July, 2012
The reason for this behavior is the high boiling point of n-butanol, which is 373K even at a low
pressure of 0.04 MPa. As the droplets of n-butanol are quite large, there is no significant
evaporation. The droplets are even smaller for OP#1 due to higher atomization obtained by higher
aerodynamic interaction as the ambient gas density is larger for the cold condition case. For the
30 mm measurement plane, the momentum exchange is already quite progressed. Thus, the effect of
higher atomization and evaporation counteract for n-butanol resulting in almost equal d32 SMD
(figure 3). As the boiling points are reduced in comparison to ambient pressure, ethanol behaves
like a low-boiling point fuel because it’s boiling point at 0.04 MPa is 335K and thus very low –
despite a very high evaporation enthalpy.
pgas=0.4 barabs.
°C
°C
%=
Figure 3: Droplet sizes (d32) for two different gas temperatures OP#1 (20°C, blue bars) vs. OP#2
(100°C, green bars) for constant low gas pressure = 0.4 bar; Δm = relative mass difference based on
arithmetically averaged droplet size
Due to the boiling point, ethanol droplets show similar large droplets like gasoline or 3K which
both contain quite high amount of low-boiling point fuels. For all multi-component fuels, ‘3K’,
‘3ZK’ and at gasoline E10, the low-boiling fractions decrease the droplet diameters for 373K gas
temperature. The differences of the SMD presented in bar graphs seem quite small. However, the
associated average droplet mass shows that for ‘3ZK’ on the 15 mm plane is 14% lower in the hot
gas environment compared to the cold gas temperature. The extent of the total drop size change
resulting from the gas temperature difference of 80K is exactly 82% for ‘3ZK’ of the total drop size
change of ethanol. This corresponds to the 75% by volume fraction of ethanol in ‘3ZK’ which
influences the evaporation behavior.
Thus, for low pressure environments the boiling points dominate the evaporation behavior. Ethanol
shows droplet sizes much larger than gasoline due to a high evaporation enthalpy, which limits the
evaporation rate after a certain time. For OP#3 and OP#4 this trend is less pronounced.
-5-
16th Int Symp on Applications of Laser Techniques to Fluid Mechanics
Lisbon, Portugal, 09-12 July, 2012
ϑgas=200°C
Figure 4: Droplet sizes (d32) for two different gas pressures OP#3 (5.6 bar, green bars) vs. OP#4
(8 bar, blue bars) for constant gas temperature = 200°C; Δm = relative mass difference based on
arithmetically averaged droplet size
Investigations at moderate pressure conditions OP#3 at 0.56 MPa (figure 4, green bars) show an
increasing influence of the evaporation enthalpy compared to previous tests [1] indicated that at low
pressure ethanol behaves like a low-boiling point component; the evaporation enthalpy influence
seemed to be reduced. The droplet sizes of ethanol at low pressure were found to be approximately
20% smaller than those of isooctane. However, in the case of OP#4 they are bigger due to the high
evaporation enthalpy which is found to become more important. While for the measurement plane
15 mm the droplet size difference of 1 µm between isooctane and ethanol is not that significant, at
30 mm downstream to the nozzle tip the evaporation enthalpy of ethanol limits the evaporation rate
(figure 4, right) while isooctane shows 3 µm smaller droplets. Figure 4 also displays that in addition
to n-butanol, ethanol behaves more like a high-boiling component and shows larger droplets despite
the relatively low boiling point. N-butanol shows the highest difference in droplet mass as its
boiling point is 447K at 0.56 MPa and 464K at 0.08 MPa, which is close to the gas temperature
(473K) in the measurement chamber. Thus, for OP#4 n-butanol shows a lower evaporation rate
compared to OP#3. This effect can also be seen for ‘3ZK’ due to its small amount of n-butanol. The
evaporation behavior for these boundary conditions is found to be more influenced by the boiling
temperature than by the evaporation enthalpy.
The droplet velocities of all investigated fuels do not show very large differences. However, the
high boiling point and large viscosity fuel n-butanol shows the lowest droplet velocities at most of
the points in time (figure 5, left). N-butanol also exhibits the largest droplets for all times (figure 5,
right). The larger the droplets, the slower they propagate in the measurements. The histogram
distribution with high statistical accuracy shows that there are only larger droplets for n-butanol for
increased gas pressure and temperature [3].
-6-
16th Int Symp on Applications of Laser Techniques to Fluid Mechanics
Lisbon, Portugal, 09-12 July, 2012
Figure 5: Time-resolved axial droplet velocities (left) and droplet sizes d10 (right)
for OP#4 (0.8MPa, 473K)
At increased ambient pressure conditions for OP#5 with 0.08 MPa / 673K (figure 6, blue bars) the
main influences change significantly. The evaporation behavior of ethanol is now closer to that of
the high-boiling n-butanol, which implies that the influence of the evaporation enthalpy increases
(figure 6-8). For higher gas temperature an increase for the droplet size difference of ethanol to
gasoline and isooctane occurs (figure 6, 7). The main result of the space resolved distribution
(figure 7, right) is that the largest droplet diameters are present for n-butanol over the complete
plume cross-section for both measuring planes. In general, the droplet sizes are larger for the
30 mm positions as collision effects may take place and very small droplets may evaporate before
reaching this distance.
pgas=8 bar
Figure 6: Absolute SMD for two different gas temp. OP#5 (400°C, green bars) vs. OP#4 (200°C,
blue bars) at gas pressure of 8 bar; Δm = arith. change in droplet mass (temp. difference of 200K)
For the very high pressure and temperature of OP#6 figure 7 presents all investigated fuels resolved
in time and space. The 3-component ‘3K’ fuel with RON46 matches the droplet size distribution of
gasoline for all distributions (figure 6-8) as it contains low (n-hexane), middle (isooctane) and high
-7-
16th Int Symp on Applications of Laser Techniques to Fluid Mechanics
Lisbon, Portugal, 09-12 July, 2012
boiling point (n-decane) components [1]. This verifies the similar atomization and evaporation
behavior for ‘3K’ and gasoline for this high gas pressure and temperature environment, which was
also demonstrated for another injector at different conditions (1.5 MPa, 473K) [1]. N-butanol as a
higher viscosity fuel with the highest boiling point of the investigated fuels and a relatively high
evaporation enthalpy shows the largest droplet mean diameters for all points in time (figure 7, left)
and for all measurement positions. During the measurements the lowest droplet velocities could
also be observed for this fuel, similar to OP#4. Isooctane worked well as a surrogate fuel for
gasoline for this specific operating point. ‘3ZK’, however, behaves more like a high-boiling point
fuel, as it consists only of 10% by volume n-hexane thus it shows droplet sizes that are closer to
those of n-butanol.
Figure 7: Temporal resolved droplet size d10 distribution at elevated temperature/pressure for
various fuels at a measurement distance of 30 mm from the nozzle tip (left) and space resolved
SMD through one spray plume for all tested fuels (30 mm) (right) for OP#5 (0.8MPa, 673K)
Despite its large amount of ethanol, the droplet sizes of ‘3ZK’ are larger than for the pure ethanol
fuel. This confirms the large effect of the high boiling point fuel components despite their low
fraction in the mixtures [1]. For the low volatility fuel isooctane the increase of 200K (figure 6,
D200°C vs. D400°C) causes a reduction in the arithmetic drop mass by 63.9% (figure 6, right).
However, the mass of ethanol drops in this hot environment decreases only 47.1% due to its high
evaporation enthalpy. Basically, the strong influence of the high-boiling point n-butanol in the 3component fuel ‘3ZK’ is observed – even though only 15% by volume of it is contained in ‘3ZK’.
Its main component ethanol has less effect on the evaporation on the 30 mm measuring plane where
the momentum exchange with the ambient gas is almost completed. Even those low amounts of
high-boiling point components show a great influence on the droplet evaporation for all operating
points. In low atmospheric pressure measurements the boiling point was found to be the main
influencing physiochemical property. For moderate pressure conditions at elevated temperature the
influences of the properties boiling point and evaporation enthalpy counteract to this behavior
(OP#4). At the very high gas pressure and temperature conditions (OP#5), the evaporation is found
to be strongly dominated by the evaporation enthalpy as proved simply by comparing the fuels
‘3ZK’, ethanol and n-butanol.
4. Conclusions
In this study a variety of single-component fuels and mixtures are tested at different ambient
conditions to identify the most significant fuel properties and their effects on the atomization and
-8-
16th Int Symp on Applications of Laser Techniques to Fluid Mechanics
Lisbon, Portugal, 09-12 July, 2012
evaporation rate. The local droplet size distributions found as a function of spray momentum,
mixture formation by gas entrainment as well as the chemical data of evaporation enthalpy and
boiling points can explain the global spray propagation. For various investigated engine conditions
different physiochemical properties were found to become dominant. In low pressure measurements
(0.04 MPa) it was observable that the boiling point is the main influencing physiochemical
property. A high degree of evaporation rate could be observed especially for ethanol and the multicomponent fuels which contain low boiling point fractions.
For moderate ambient conditions such as 0.56 MPa / 473 K isooctane, gasoline and ethanol and
‘3K’ produce similar evaporation by similar droplet sizes, despite large differences in the chemical
data of boiling point and evaporation enthalpy of the single component fuel. The chemical data of
boiling point and evaporation enthalpy neutralize each other. N-butanol as a very high boiling fuel
with relatively high evaporation enthalpy exhibits the largest droplets for this operating point,
followed by ‘3ZK’ (RON95). Even though the three-component mixture ‘3ZK’ contains only 15%
by volume of n-butanol, the behavior appeared more similar to n-butanol than to ethanol, which is
contained in ‘3ZK’ by 75% by volume. Thus, the high boiling components show a very strong
effect on the spray behavior and droplet distribution for moderate operating conditions.
At the elevated operating conditions of 0.8 MPa / 673 K, the spray behavior and main influences
change completely. Due to a high degree of evaporation which took place, the studied fuels were
found to exhibit a stronger difference in droplet size. Isooctane with a higher boiling point than
ethanol shows much smaller droplets than ethanol. Due to the very high evaporation enthalpy of
ethanol, the droplets evaporation rate is limited. Thus, ethanol behaved much more like a high
boiling point fuel component such as n-butanol. The phenomenon of ‘3ZK’ mentioned above
remains, the component n-butanol with 15% by volume in the mixture delivers average droplet
sizes similar to n-butanol rather than its main component ethanol. However, the evaporation cooling
of ethanol must be very high due to the specific evaporation enthalpy. This can be seen for nbutanol with an evaporation enthalpy of 637.7 kJ/kg compared to ethanol with 938.2 kJ/kg. The
trend for the droplet sizes correspond to the evaporation enthalpy values.
Thus, it can be stated that for stratified charge conditions with high pressure and temperature, the
evaporation enthalpy is seen to be the dominating physiochemical property for more-distant spray
locations where the secondary breakup is already finished. The findings of this work will be
verified by planar laser-induced fluorescence vapor temperature measurements in further
experiments. The knowledge of the atomization and evaporation rate in dense sprays can be used
for both model development and validation as well as to accelerate the optimization of injection and
combustion strategies for modern fuels.
-9-
16th Int Symp on Applications of Laser Techniques to Fluid Mechanics
Lisbon, Portugal, 09-12 July, 2012
Acknowledgements
The authors acknowledge the Bavarian Research Foundation (BFS) for supporting the project
“WiDiKO” in which this work was carried out. The authors also gratefully acknowledge the
financial support for parts of this work from the Erlangen Graduate School in Advanced Optical
Technologies (SAOT) within the framework of the German Excellence Initiative by the German
Research Foundation (DFG).
Abbreviations
Ar*: Argon Ion
d10: Arithmetic mean droplet diameter
d32: Sauter mean diameter
DISI: Direct Injection Spark Ignition
Nd:YAG: Neodym Yttrium Aluminium Granat
PDA: Phase Doppler Anemometry
RON: Research Octane Number
SMD: Sauter mean diameter
SOI: Start of injection
vSOI: Visible start of injection
‘3K’: Three-component model fuel consisting of n-hexane, isooctane and n-decane (RON 46)
‘3ZK’: Three-component model fuel consisting of n-hexane, ethanol and n-butanol (RON 95)
5. References
[1] Zigan, L., Schmitz, I., Flügel, A., Knorsch, T., Wensing, M., Leipertz, A., Energy&Fuels
2010, 24, 4341–4350.
[2] Pischinger, S. Verbrennungsmotoren, Band I, 21st edition, RWTH Aachen, 2000.
[3] Knorsch, T., Zigan, L., Trost, J., Wensing, M. and Leipertz, A., Biofuel Droplet Evaporation
Rate of a DISI Spray by Laser-induced Fluorescence and Phase Doppler Anemometry, 12th
Triennial Int. Conference on Liquid Atomization and Spray Systems (ICLASS), Heidelberg,
Germany, September 2 - 6, 2012.
[4] Heldmann, M., Knorsch, T., Schmitz, I., Wensing, M., Leipertz, A., Investigation of
significant Spray Rotation Phenomena under Flash-Boiling Conditions studied on a Multihole DISI Injector for Bio-Ethanol E85 and Gasoline E5, 24th Annual Conference on Liquid
Atomization and Spray Systems (ILASS), Estoril, Portugal, September 5 - 7, 2011.
[5] Institut für Arbeitsschutz der Deutschen Gesetzlichen Unfallversicherung: GESTISStoffdatenbank. website: http://gestis.itrust.de, pageview 12/11/2011.
[6] Bronkhorst High-Tech B.V.: FLUIDAT on the web. website: http://www.fluidat.com,
pageview 12/19/2011.
[7] Eyidogan, M., Ozsezen, A.N., Canakci, M., and Turkcan, A., Fuel 89: 2713-2720 (2010).
- 10 -