Renewable Energy 104 (2017) 30e39

Contents lists available at ScienceDirect

Renewable Energy
journal homepage: www.elsevier.com/locate/renene

Regeneration of diesel particulate lters: Effect of renewable fuels

Jose
ndez*, Magn Lapuerta, Jess Sa

Rodrguez-Ferna
nchez-Valdepen
~ as
Escuela T

ecnica Superior de Ingenieros Industriales, Universidad de Castilla-La Mancha, Avda. Camilo Jos

e Cela s/n, 13071, Ciudad Real, Spain

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

Article history: Current trends in diesel transport anticipate that in the upcoming future a range of renewable fuels will
Received 23 May 2016 be necessary to comply with emission and sustainability legislations. Exhaust after-treatment devices
Received in revised form such as diesel particulate lters eDPFse will have to operate satisfactorily with this pool of biofuels. In
5 October 2016
particular, DPF regeneration is crucial to cut the fuel penalty and guarantee an acceptable lifetime for this
Accepted 30 November 2016
device. In the present work, an automotive diesel engine was run with fossil fuel and three renewable
Available online 30 November 2016
fuels: a conventional biodiesel, a fuel manufactured through Fischer-Tropsch eFTe process and a HVO
biofuel. The DPF was loaded and regenerated through an active process with fuel post-injections.
Keywords:
Renewable diesel fuels
Additionally, soot samples were investigated with thermo-gravimetry (TGA) and calorimetry (DSC) to
Soot properties conrm whether these techniques obtain relevant information for explaining DPF behavior. Both
Particle lter regeneration methods proved that biodiesel leads to a more economical regeneration being the biodiesel soot, more
Diesel engine reactive than the other samples, the main reason. DPF regenerations with parafnic fuels (FT-derived and
Biofuels HVO) did not reveal strong differences compared to diesel, though TGA and DSC results suggested that
soot from parafnic biofuels is more reactive than that from diesel. The exhaust gas temperature and
composition are behind this apparent discrepancy.
2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction
The increase on the sale and use of diesel vehicles in Europe [1]
and the effect of their emissions on human health [2,3] and the
environment [4e6] have forced EU institutions to act. On one hand,
transport emission legislation [7,8], recently set tighter limits on
particulate matter/NOx and will adopt a more demanding driving
cycle (WLTP - Worldwide Harmonized Light Vehicles Test Proced-
ure) and tougher ambient conditions in the future. On the other
hand, a broad diesel fuel mix combining fossil fuels, synthetic
parafnic diesel and oxygenated biofuels [9,10], is being promoted
in the transport sector to both tackle the GHG emissions (mainly
CO2) and increase the renewable energy share up to 10% by 2020
[11].
Under this regulatory framework, manufacturers have intro-
duced advanced injection strategies and efcient aftertreatment
techniques, and they will have to adapt their diesel vehicles to run
on multiple fuels. A typical approach consists in running with high
EGR ratios [12] to reduce NOx formation at the expense of
increasing emission of particulate matter, which is afterwards
* Corresponding author.

ndez).
E-mail address: Jose.RFernandez@uclm.es (J. Rodrguez-Ferna
trapped and removed in a diesel particle lter e DPF [13]. Euro 6
Standard, with a further 55% reduction of NOx with respect to Euro
5 limits, obliges to depurate this scheme by including a NOx after-
treatment system [14e17].
Different DPF congurations are available [18,19], but wall-ow
type lters are the most widely used. Wall-ow lters are honey-
comb monoliths with parallel channels plugged alternately at each
end to force the exhaust gas through the porous lter wall where
soot is retained. Therefore, backpressure increases in the exhaust
line penalizing the fuel consumption [20,21], but the high ltering
efciency (about 90%) makes these lters essential for complying
with emission standards. When the backpressure reaches a
threshold, fuel post-injections are launched to increase the exhaust
temperature and oxidize the soot collected (lter regeneration)
[22]. Apart from temperature, other factors such as the exhaust gas
(composition, ow rate) and the physicochemical properties of soot
affect the regeneration process [23e26]. Indeed the properties of
soot modify its reactivity in a remarkable way [27,28]. The term
reactivity is used hereinafter to refer to the soot ability to be
oxidized at higher rates and/or under a lower temperature envi-
ronment, which leads to a more efcient regeneration.
The production routes, properties and performance/emissions
of engines and vehicles fueled with alternative and renewable
diesel fuels have been broadly studied [9,29e31]. Briey, biodiesel

http://dx.doi.org/10.1016/j.renene.2016.11.059
0960-1481/ 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

ndez et al. / Renewable Energy 104 (2017) 30e39
J. Rodrguez-Ferna 31

Nomenclature LTFT low temperature Fischer e Tropsch

m total particle mass
A total particle surface MLRTmax temperature for the maximum mass loss rate
BET break-even temperature NEDC new European driving cycle
BTL biomass to liquid R particle radius
CFPP cold lter plugging point SMPS scanning mobility particle sizer
DPF diesel particulate lter TDC top dead center
DSC differential scanning calorimetry TEM transmission electron microscopy
ECU electronic control unit TGA thermogravimetric analyzer
EGR exhaust gas recirculation V particle volume
FT Fischer-Tropsch WLTP worldwide harmonized light vehicles test procedure
GHG greenhouse gas XRD X e ray diffraction
HRRTmax temperature for the maximum heat release rate r particle density
HVO hydrotreated vegetable oil 4 particle concentration

(obtained from oils through transesterication reaction) and syn-
thetic parafnic fuels (hydrogenation of bio oils and Fischer-
Tropsch process are typical routes) lead to similar engine ef-
ciency, fuel economy that scales inversely with the heating value
and a benet in pollutant emissions, though the effect of biodiesel
on NOx is controversial and depends markedly on its composition.
However, the effect of these future fuels on the performance of
advanced after-treatment technologies such as DPFs is a recent
issue that will raise the concern of manufacturers.
One research line involves the study of the soot oxidation pro-
cess and its effect on lter regenerations. Some works characterized
commercial soot surrogates [32,33], but this approach is invalid
when the effect of fuel on soot reactivity is to be assessed. Others
used analytical techniques (TEM [26,34e37], XRD [26,36,38e40] or
Raman spectroscopy [26,39,41e44]) to study real soot generated
running engines with different fuels. Generally, the fuel used
modies the characteristics (structure, primary particle size,
graphitization) of soot generated [26,28,45]. This may affect the
soot reactivity and therefore oxidation rate in the DPF. As a rule, a
more disordered soot nanostructure presents more distance be-
tween graphene layers and a higher number of active sites where
oxidants, such as O2 or NO2, can be adsorbed initiating the soot
oxidation [27,39]. The presence of oxygenated moieties on the soot
surface [26] affect the soot reactivity as well, as the oxygen con-
tained in these groups may participate in the carbon oxidation
reactions. A higher concentration of oxygenated groups has been
reported for biodiesel soot, which may be behind its higher reac-
tivity (compared to fossil diesel soot) [24,26,27].
Temperature-programmed oxidation and thermal analysis
techniques [24,26,28,39,46], mainly TGA and DSC, under oxidant
atmospheres have been used to directly evaluate soot reactivity in
laboratory, but without a validation in a diesel DPF. Finally, few
researchers have carried out engine tests to evaluate soot reactivity
through the DPF pressure drop trace during regeneration, once the
DPF was previously loaded with soot [24,26,47]. Some authors
[24,48], showed that biodiesel or alternative fuels enhanced the
regeneration process, while others [25,26,47] concluded that bio-
diesel did not improve the regeneration process because engine
calibration (ECU) was not adapted for fuels other than conventional
diesel.
In the present work, a Euro 5 diesel engine tted with a DPF has
been used to carry out load and regeneration processes in the lter
with four different fuels, including three renewable fuels. Tem-
perature and pressure drop in the lter were monitored to evaluate
soot reactivity. Furthermore, soot samples collected during the
loading process were analyzed in laboratory with TGA and DSC to
conrm whether the trends obtained in laboratory match those in
the DPF. The information obtained is relevant to improve the per-
formance of modern after-treatment systems under new biofuels.
2. Experimental setup
Engine tests were carried out in a 4-cylinder, 4-stroke, turbo-
charged, intercooled, common-rail, 2.0 L Nissan diesel engine
(model M1D), fullling Euro 5 Standard. Ambient temperature was
xed at 21e23
C. The engine is equipped with cooled exhaust gas
recirculation (EGR), and its temperature was externally controlled.
As after-treatment devices, a diesel oxidation catalyst eDOCe and a
regenerative wall-ow type diesel particulate lter eDPFe (see
lter characteristics in Table 1) are tted in the engine exhaust
system. The main specications of the engine are given in Table 2.
The engine was coupled to an asynchronous electric brake Schenck
Dynas III LI250, equipped with speed and torque sensors that allow
measuring and controlling the engine speed and effective torque,
respectively. The INCA PC software and the ETAS ES 591.1 were used
for the communication and management of the electronic control
unit (ECU). This allows dening the injection process (timing,
pressure, strategy) as well as measuring and recording temperature
and pressure in different locations of the engine (temperature
upstream of the turbine, temperature and pressure upstream and
downstream of the DPF, oil temperature and ambient pressure,
among others).
NOx emissions were measured using a chemiluminescence
Topaze 32 M analyzer able to distinguish between NO and NO2.
Furthermore, oxygen concentration was determined with a non-
dispersive infrared analyzer MIR2M. Particulate matter (PM) was
collected using a partial-ow dilution micro tunnel (Horiba DLS
2300). This equipment sucks a portion of the exhaust gas through a
heated probe into the tunnel where it is diluted with ltered
ambient air, as required by emission directives, at a dilution ratio of
10:1. Particulate matter is then collected on Whatman GF/F glass
microber lters (47 mm diameter, 0.7 mm pore side). To obtain the
Table 1
DPF characteristics.
Material Silicon Carbide (SiC)
Dimensions Width (190 mm)/Length (240 mm)
Conguration 16/200
Catalyst Coated with Pt (0.0032 g/cm3)
Volume 4.1 L
Channel width 1.39 mm
Specic ltering surface 758.91 m2/m3
32
ndez et al. / Renewable Energy 104 (2017) 30e39
J. Rodrguez-Ferna

Table 2 using a thermal ramp of 1 C/min up to 650
C (it was previously

Diesel engine characteristics. veried that at this temperature all soot mass has been consumed).
Fuel injection system DI, common-rail In the case of the TGA, the weight loss is recorded as a function of
Cylinders 4 the temperature, while in the case of DSC heat release is recorded
Valves 16 instead.
Bore 84 mm
Stroke 90 mm
Compression ratio 16:1
Displacement 1994 cm3
Maximum power 111 kW @ 4000 rpm
Maximum torque 323 Nm @ 2000 rpm 3. Test procedure

The velocity-time sequences of the New European Driving Cycle

mass particle the lters were weighed before and after collection in
a Sartorius SE2 precision balance. Prior weighing, lters were
conditioned in a climatic chamber (Dycometal CM O/216) at 22
C
and 45% relative humidity. The particle size distributions were
measured using a Scanning Mobility Particle Sizer (SMPS) TSI
model 39136L10. The conguration used in the SMPS counts par-
ticles from 12.6 nm to 552.3 nm, distributed in 105 diameter ranges.
To avoid saturation of the SMPS counting system and to decrease
the temperature below the maximum limit permitted in the
specications, the exhaust gas sample was diluted with ltered
compressed air in two Dekati diluters DI-1000 based on ejection
dilution (total dilution ratio of 64:1). The accuracy of the measuring
instruments, alongside other details of the equipment, is listed in
Table 3.
The engine exhaust system was extended (as illustrated in Fig. 1)
to include a by-pass line with a stainless steel cylindrical case
previously lled with stainless steel mesh (which acts as a ltering
medium). When the exhaust gas ows through the by-pass line, the
soot is trapped in the mesh. The aim of this modication is to
enable the collection of soot in a more accessible way than in the
original conguration. Once the mesh has retained a sufcient
amount of soot (indicated by the pressure drop measured with two
pressure sensors mounted up- and downstream of the mesh), the
structure is disassembled and the mesh is removed and softly
shaken to separate the soot, letting it fall onto an aluminum sheet,
and the soot is kept for analysis.
Soot reactivity was investigated through two thermal analysis
techniques: a TGA (Thermogravimetric Analyzer) TA Q500 and a
DSC (Differential Scanning Calorimetry) TA Q20. In both in-
struments the soot sample is located in a crucible inside of a small
furnace (to reduce the resistance associated with the heat-transfer
process). A soot mass of 3 mg was used and the oxidation process
followed was previously developed [46]. Firstly, a desorption pro-
cess under nitrogen atmosphere takes place (at 400
C for 1 h).
Afterwards the sample is cooled down to 100
C. Then the oxidation
process of the soot sample is carried out under air atmosphere
(NEDC) were converted to steady operation modes (dened by
constant engine speed and torque) following a procedure previ-
ously published [49]. Among all of them, two modes (here denoted
as U9 and REU8) were selected to illustrate the loading and the
regeneration regimes in the DPF, respectively. NEDC is the cycle
that light-duty vehicles must follow for certication in Europe. It is
comprised by four replicates of an urban sub-cycle (which simu-
lates a city driving pattern with low speed and frequent idle se-
quences) followed by one extra-urban sub-cycle (highway driving
pattern with more aggressive modes and a maximum velocity of
120 km/h).
The loading mode (U9) is a low-load mode with low exhaust
temperature, representative of urban driving conditions. Its high
EGR ratio (see Table 4) leads to high soot emissions that contribute
to load the DPF. The regeneration mode (REU8) is a variation from a
medium-load mode (EU8) that belongs to the extra-urban sub-
cycle. Fixing the speed and torque of EU8 mode, in REU8 an active
regeneration is user-commanded in the electronic control unit. In
response, the control algorithm changes the injection timing, the
injection strategy and the EGR ratio to the settings listed in Table 4.
In the EU8 mode the exhaust temperature is not enough to sustain
the oxidation of soot in the DPF at a sufcient rate, therefore
switching to REU8 mode is necessary for an active regeneration.
To assess the reactivity of soot two different engine tests have
been performed. In the rst one, the engine was run in the U9 mode
and the soot was collected in the mesh of the stainless steel case
previously described. The soot was later analyzed in the thermog-
ravimetric analyzer (TGA) and the calorimeter (DSC) to evaluate
reactivity. The second test evaluated soot reactivity directly in the
DPF by measuring pressure drop and temperature trends during i) a
loading process and ii) a regeneration event. This experimental
procedure is illustrated in Fig. 2. For the regeneration test, the DPF
was previously loaded with soot generated in the U9 mode until
pressure drop increased up to 80 hPa, approximately. Then the
engine was switched to REU8 mode to trigger the active regener-
ation of the DPF.

Table 3
Measuring equipment characteristics.

Equipment Principle of operation Manufacturer and model Precision

NOx analyzer Chemiluminescence Environnement. Topaze 32 M Accuracy: 1%

Oxygen analyzer Infrared absorbance Environnement. Accuracy: 1%
MIR 2 M
Partial e ow dilution micro tunnel Particle collection in lters; subsequent weighing Horiba. Uncertainty: 5%
DLS 2300
Double dilution system Ejection Dekati. Accuracy: 1%
DI e 1000
Scanning mobility particle sizer Electric mobility TSI. Resolution: 1 particle/cm3
19136L10
Thermogravimetric analyzer Mass loss measurement TA Instruments. Weighing precision: 0.01%
Q500 Resolution: 1 mg
Differential scanning calorimeter Heat ux TA Instruments. Calorimetric precision: 0.05%
Q20 Resolution: 0.04 mW

ndez et al. / Renewable Energy 104 (2017) 30e39
J. Rodrguez-Ferna 33

STAINLESS
STEEL STRUCTURE EXHAUST
GASES

DOC
INTAKE
AIR FILTER DIESEL
INTAKE PARTICULATE FILTER
HOT WIRE
AIR SENSOR
TOPAZE32M
Intake Air Temperature Co ntrol
MIR2M

Fuel
tank
INTAKE AIR
INTERCOOLER EGR
Temp erature Control MICRO TUNNEL SMPS
HORIBA DLS 2300 TSI 3936 L10

EG R CO O L ER

FUEL GRAVIMETRIC Fuel Temperature TDC Sensor

BALANCE Control
(AVL 733S)

ASYNCHRONOUS BRAKE
(SCHENCK DYNAS LI 250)

COMUNICATION ELECTRONIC
CONTROL AND HARDWARE CONTROL UNIT (ECU)
ACQUISITIOMSYSTEM
(INCAPC) ETASES591.1

Fig. 1. Engine test bench.

Table 4 4. Test fuels

Operating parameters of U9 and REU8 modes.
Engine speed (rpm)
Effective torque (Nm)
EGR ratio (%)
Air owrate (kg/h)
Start of pilot injection (CA bTDC)
Start of main injection (CA aTDC)
Start of post injection 1 (CA aTDC)
Start of post injection 2 (CA aTDC)
Fuel injected during pilot injection (mL/inj)
Fuel injected during post injection 1 (mL/inj)
Fuel injected during post injection 2 (mL/inj)
Injection pressure (bar)
Four different fuels were used to carry out the engine tests. The
U9 mode REU8 mode
rst one is a rst-ll fossil diesel fuel that meets the EN-590
1667 1972
Standard with no biodiesel content. It is used in cars exported
78 94
22e23 0 abroad to avoid lter plugging in harsh weather and was donated
z78 z151 by Repsol Corporation (Spain). This diesel fuel was used as a
12.7 13.6 reference to compare the results with the obtained with two
5 11 parafnic biofuels (HVO and Fischer-Tropsch-derived) and a bio-
e 62.6
diesel fuel. HVO was produced by hydro-treating of a blend of oils
e 135.1
1.86 3.08 including palm, camelina and other residual oils. This fuel was
e 2.59 donated by Neste Corporation (Finland) and contains lubricity ad-
e 11.28 ditives. Moreover, the parafn fraction of HVO was further iso-
660 745
merized to improve the cold ow properties. Hydro-treating of
vegetable oils is a novel way to produce high-quality renewable
diesel fuels at a reduced cost (it can be combined with other fuel
processes in current reneries). The other parafnic fuel studied
was produced by Sasol Corporation through a low temperature
Fischer Tropsch process (LTFT). It was produced from natural gas,
therefore it is here denominated GTL (gas-to-liquid), but its prop-
erties are common to fuels produced through Fischer-Tropsch
process from renewable feedstock (BTL, biomass-to-liquid).
Finally, the biodiesel used in this study was an animal fat methyl
ester, which is currently among the most widely used domestic
feedstock for biodiesel production in southern European countries.
This biodiesel fuel meets the EN-14214 Standard and was supplied

s (Spain). The composition and main properties of
by Stock del Valle
the tested fuels are presented in Table 5. Furthermore, Table 5 also
shows the standards for the four fuels tested: UNE EN 590 for fossil
diesel, UNE EN 15940 for parafnic fuels (HVO/GTL) and UNE EN
14214 for biodiesel. All fuels tested meet its respective Standard.

Fig. 2. Experimental procedure.

34
ndez et al. / Renewable Energy 104 (2017) 30e39
J. Rodrguez-Ferna

Table 5
Fuel properties.

Properties Diesel HVO GTL Biodiesel UNE EN 590 UNE EN 15940 UNE EN 14214

Min Max Min Max Min Max

3
Density at 15 C (kg/m ) 811 779.6 774 874.3 820 845 765 800 860 900
Viscosity at 40
C (mm2/s) 2.02 2.99 2.34 4.5 2 4.5 2 4.5 3.5 5
Lubricity (WS1.4) (mm) 226 334 211 187 e 460 e 460 e e
Lower heating value (MJ/kg) 43.04 43.96 44.03 37.26
Lower heating value (MJ/L) 34.91 34.27 34.08 32.58
CFPP (C) 30 21 7 9
Aromatic content (% w/w) 17.2 0 0 0 e 1.1
Cetane number 67 94.8 89.2 68.8 51 e 70 e 51 e
Acid number 0.16 0.06 0.20 0.34 e 0.5
Sulfur (mg/kg) <10 <10 <10 <10 e 10 e 5 e 10
Water (mg/kg) 60 19.2 20 102 e 200 e 200 e 500
C (% w/w) 85.74a 84.68a 84.82a 76.45d
H (% w/w) 14.26a 14.53 15.18a 12.36d
O (% w/w) 0a 0a 0a 11.19d
Mean molecular weight (g/mol) 187.4b 196.1b 238.9b 285.4d
Molecular formulac C13.39H26.72c C13.95H28.7c C16.89H36.26c C18.19H35.05O2d
Stoichiometric fuel/air ratio 1/14.70 1/14.79 1/14.92 1/12.48
a
Elemental composition measured with a CHNS analyzer.
b
Calculated with software HYSYS.
c
Calculated through elemental composition and mean molecular weight.
d
Calculated from the ester prole (EN-14103).

5. Results and discussion
5.1. Soot analysis results
An example of thermograms directly obtained in the TGA and
DSC is shown in Fig. 3 a) and b), respectively. The shape of the
curves for all fuels is rather similar, though the fuel affects the curve
location on the temperature axis and therefore the soot oxidative
reactivity (the higher the temperature, the lower the oxidative
reactivity). As detailed in Ref. [46], several characteristic tempera-
tures can be dened from the thermograms to determine the effect
of the fuel. One of these is the temperature for the maximum of the
trace, denoted as MLRTmax (temperature for the maximum mass
loss rate) in the case of TGA or HRRTmax (temperature for the
maximum heat release rate) in the case of DSC, as presented in
Ref. [28].
Fig. 4 represents the MLRTmax and HRRTmax for the soot
generated with all fuels tested. As can be seen, the same trend is
observed from mass loss rate and heat release: parafnic fuels and,
especially, oxygenated biodiesel, shift the soot oxidation towards
lower temperatures. The observed consistency between TGA and
DSC is expected since both share the same working principle and
the samples were tested under equal temperature-atmosphere
programs. The minor differences in the temperature values from
TGA and DSC are understandable since the furnace conguration
(which affect the thermal inertia) and the thermocouple location
are different in both instruments.
Soot generated from diesel fuel is the least reactive sample
(higher temperature indicates lower reactivity). The presence of
aromatic compounds in diesel fuel, well known as soot precursors
[50], may lead to a higher number of soot nuclei formed in the
combustion chamber that, by coalescent coagulation [51], may
form larger primary particles. Previous publications show larger
primary particle and agglomerate diameter in the case of soot from
diesel fuels [26,27]. Larger particles and higher fractal dimensions
possess lower specic surface area, which may be responsible for a
lower reactivity. In addition, the lower porosity of diesel soot [28]
hinders the accessibility of oxygen to active sites, therefore
reducing the oxidative reactivity. Biodiesel fuel generated the most
reactive soot. Again, this may be explained by the lowest primary

Fig. 3. Typical soot oxidation thermograms in TGA (a) and DSC (b).

ndez et al. / Renewable Energy 104 (2017) 30e39
J. Rodrguez-Ferna
Fig. 4. MLRTmax (TGA) and HRRTmax (DSC) during soot oxidation.
particle diameter [27] since the oxygen in the ester molecule favors
the oxidation of soot nuclei. The presence of oxygenated func-
tionalities on the biodiesel soot surface has been related to an
enhanced reactivity [28] as well.
To support this theorized relationship between soot reactivity
and surface area, particle size distributions were measured with the
SMPS and are presented in Fig. 5 a). Using the apparent density
correlation proposed in Ref. [52], calculated assuming spherical
particles, the total particle mass may be estimated (Eq. (1)). Addi-
tionally the surface area may be calculated assuming spherical
particles as well (Eq. (2)). The resulting specic surface area (sur-
face area divided by mass) is shown in Fig. 5 b). A more precise
computation of surface area should involve the use of the exact
geometry of the particle agglomerates, which are fractal-shaped
rather than spherical. Nevertheless, this is not expected to change
the fuel effect manifested in Fig. 5 b), since the main parameter
involved in the surface area calculation is the number and size of
the agglomerates. As observed, predicted surface area correlates
well with soot reactivity in TGA/DSC: biodiesel soot, the most
reactive sample, exhibits the largest specic surface area, while
Fig. 5. Particle size distributions (a) and specic surface area (b).
35
diesel soot is just the opposite.
X
n
m fi $Vi $ri (1)
i1
X
n
A fi $Ai (2)
i1
where 4i is the particle concentration for each diameter range i; Vi
and Ai are the volume and the surface area of a particle (assuming
sphericity, Vi 4/3$p$R3i and Ai 4$p$R2i ); ri is the particle density
calculated using the correlation proposed in Ref. [52]; m and A are
the total particle mass and total particle surface (per unit of exhaust
volume); n is the number of diameter ranges (in the SMPS model
used, n 105).
5.2. Engine test results
As described in the test procedure, the DPF was rstly loaded by
running the engine in the U9 mode and then it was regenerated
(REU8 mode). Fig. 6 shows the exhaust backpressure during the
loading process. For all fuels, the pressure drop rapidly increased at
the beginning of the loading because of the heating up process of
the ltering substrate (which increases the volumetric owrate and
thus the velocity of the exhaust ow) and the ltering regime,
which is initially dominated by the rapid clogging of the pore dis-
tribution (the so called intrinsic ltering regime [53]).
Once this initial ltering regime ceases, the drop pressure across
the DPF keeps increasing at a constant rate in the case of diesel fuel
but reaches a steady value in the case of parafnic and biodiesel
fuels. This means that the latter fuels reached the temperature at
which soot accumulation balances soot oxidation in the DPF
channels, known as break-even temperature (BET). The lowest
equilibrium temperature was found for the biodiesel fuel, followed
by both parafnic fuels (see Fig. 6). Contrarily, diesel fuel did not
reach the BET. Signicant benets are expected for the alternative
fuels here tested, since a lower pressure drop means an increased
fuel economy as the engine pumping work is reduced [21] and a
lower frequency of active regeneration is required (the need of
post-injections introduce an additional fuel penalty).
The main reason for explaining BET/pressure drop trends is the
different oxidative reactivity of the trapped soot. However, soot
reactivity cannot be evaluated directly from the pressure drop
36
J. Rodrguez-Ferna
Fig. 6. a) DPF pressure drop during the loading process. b) nal DPF pressure drop and DPF temperature (a high emission variation of the operation mode was used for biodiesel and
parafnic fuels, as detailed in the text).
progress, since the exhaust gas composition and the particulate
emissions (which depend greatly on the fuel tested) also play a role.
For example, both NO2 and O2 participate as oxidants during the
oxidation process. NO generated can react with O2 generating NO2
by a reversible reaction that takes place on catalytic sites in the DPF
wall [54]. The reactivity of NO2 towards soot is much higher at low
temperatures (150e350
C) [55] than that of O2, and some authors
have studied the activation energy when soot reacts with different
oxidants obtaining that this energy is lower with NO2 than with O2
[56]. The oxidation pathway is different as well (the reaction be-
tween NO2 and soot generates NO, CO2, CO, N2 and N2O [57]). Also
NO could participate signicantly in soot oxidation reactions, but
only at temperature above 600
C [54]. Hence, the highest NOx
emissions for the biodiesel fuel, represented in Fig. 7, may be
responsible for enhancing the biodiesel soot oxidation. In addition,
the lowest particulate emissions for the biodiesel fuel (Fig. 8)
Fig. 7. NOx emissions during the loading test.

ndez et al. / Renewable Energy 104 (2017) 30e39
facilitated to reach the equilibrium between particle generation
and oxidation in the DPF (i.e. to reach its BET temperature).
Oppositely, the highest particulate emission with diesel fuel (much
higher than that of the rest of fuels) explains that diesel fuel was not
able to reach its BET during the whole loading process.
As commented above, BET is reached for parafnic and biodiesel
fuels and this is the reason why the pressure drop across the DPF
reaches a steady value, as shown in Fig. 6 a). To achieve a pressure
drop across the DPF high enough to start an active regeneration
process (about 80 hPa), U9 mode had to be modied to a higher
emission condition (increasing EGR ratio) during the loading test
when running with biodiesel and parafnic fuels. Temperature and
backpressure across the DPF at the end of the modied loading
process are shown in Fig. 6 b). Starting the regeneration tests with a
similar backpressure allows reducing the uncertainties when
determining the effect of the fuel on the subsequent regeneration
Fig. 8. Particulate emissions during the loading test.

ndez et al. / Renewable Energy 104 (2017) 30e39
J. Rodrguez-Ferna 37

process.
The regeneration process in the REU8 mode is presented in
Fig. 9, with an enlarged window focused on the detail of diesel and
parafnic fuel traces. It can be observed that soot generated from
diesel fuel takes more time to regenerate than those of the other
fuels (longer regeneration processes is an indication of lower
reactivity). This result agrees with those obtained with TGA and
DSC. Nevertheless, there is no much difference between fossil
diesel and parafnic fuels, even though the temperature upstream
of the DPF when the engine is running with parafnic fuels is a bit
higher than that with diesel (a higher temperature favors a higher
oxidation rate). The regeneration of soot from parafnic fuels,
slower than expected according to TGA/DSC, can be partially
explained by the lower NOx emissions of these fuels, especially in
the case of HVO (see Fig. 10). As discussed above, as NOx emissions
increase, the regeneration process is enhanced [54]. Additionally,
the higher pressure drop after the loading process for HVO (Fig. 6)
may explain its relatively slow regeneration.
The temperature upstream of the DPF is a decisive variable in a
DPF regeneration. The lowest temperature was reached with the
biodiesel fuel, in accordance with its lowest volumetric heating
Fig. 10. NOx emissions during the regeneration test.
value (see Table 5), since the volume of fuel in the post-injections
was the same for all fuels. For the other three fuels, heating value
and temperature differences were smaller. However, the higher
volumetric heating value of diesel fuel does not explain its lower
upstream DPF temperature. This may be a consequence of the
combustion of the rst post-injection, around 60 CA aTDC
(Table 4), which is more fully burnt inside the combustion chamber
in the case of the parafnic fuels because of their markedly higher
cetane number. Therefore, cylinder exhaust temperature increases
with these fuels and contributes to a higher efciency in the
oxidation catalyst, leading to a higher upstream DPF temperature.
The fastest regeneration process was found for the biodiesel fuel
(Fig. 9), despite its lowest exhaust temperature. Biodiesel regen-
eration was completed in less than 400 s, around half of the time
required for the regeneration with the other fuels. In a previous
work [25,26] in the same installation, the even lower exhaust
temperature reached with biodiesel (around 60
C lower than that
of diesel) slowed down the regeneration process with respect to
that of diesel fuel. Differently, in the present work the temperature
difference between diesel and biodiesel fueling, only 30
C
approximately, was not sufcient for bringing diesel regeneration
closer to that of biodiesel. In addition to the enhanced biodiesel
soot reactivity, the increased biodiesel NOx emissions and, espe-
cially, exhaust O2 concentration (because of its stoichiometry) also
Fig. 11. Oxygen concentration during the regeneration test.
favors a faster regeneration process when biodiesel is used (Fig. 11).

Fig. 9. Effect of the fuel on the pressure drop across the DPF and upstream temperature during the regeneration test.
38
ndez et al. / Renewable Energy 104 (2017) 30e39
J. Rodrguez-Ferna
6. Conclusions
A modern diesel automotive engine was run with four con-
ventional and renewable fuels to study the DPF regeneration pro-
cess and the properties of soot. Two methods have been applied for
this: i) soot accumulation in a real particle lter followed by an
engine-commanded active regeneration and ii) soot analysis in
laboratory through thermal analysis (TGA and DSC) under a
controlled environment and temperature program. In general, both
methods agreed and concluded that DPF regeneration ability is
dominated by the oxidative reactivity of soot and sorted this way:
biodiesel > parafnic > fossil diesel. Accordingly, two practical
conclusions may be stated. First, the higher soot oxidative reactivity
with alternative renewable fuels will entail a reduction in the fuel
penalty imposed by both the soot accumulation process on the
lter surface (increasing pumping losses with the pressure drop)
and the lter active regeneration through modied injection stra-
tegies. Second, thermal analysis for predicting soot reactivity in real
diesel lters emerges as a powerful tool, which increases the in-
terest in the use and development of these techniques.
Diesel lter regeneration at realistic engine conditions are
unique for manufacturers to assure the compatibility of new ve-
hicles and after-treatment systems with present and future bio-
fuels, but introduces uncertainties in the experiments (temperature
values, gas composition) that must be considered in the discussion.
Designed thermal analysis experiments do not represent such a
genuine approach, but experimental uncertainties can be mini-
mized and results compare well with those of real regenerations.
Acknowledgements
This study has been carried out under the framework of project
ENE2013-48602-C3-1-R, nanced by the Spanish Ministry of
Economy and Competitiveness. The authors also wish to acknowl-
edge Repsol (Spain), Stock del Valle
s (Spain) Sasol (South Africa)
and Neste (Finland) for supplying the fuels tested, and Nissan

rica (Spain) for the donation of the engine. Junta de
Motor Ibe
Comunidades de Castilla-La Mancha is also acknowledged for
nancing the fellowship of Jess S

anchez-Valdepen ~ as.
References
[1] European Automobile Manufacturers Association, ACEA, Share of Diesel in
New Passenger Cars, 2015. Available at: http://www.acea.be/statistics/tag/
category/share-of-diesel-in-new-passenger-cars.
[2] D.W. Dockery, J. Schwartz, J.D. Spengler, Air pollution and daily mortality:
associations with particulates and acid aerosols, Environ. Res. 59 (1992)
362e370.
[3] J. Pourazar, A.J. Frew, A. Blomberg, R. Helleday, F.J. Kelly, S. Wilson, Diesel
exhaust exposure enhances the expression of IL-13 in the bronchial epithe-
lium of healthy subjects, Respir. Med. 98 (2004) 821e825.
[4] P. Goyal, N. Jaiswal, A. Kumar, J.K. Dadoo, M. Dwarakanath, Air quality impact
assessment of NOx and PM due to diesel vehicles in Delhi, Transp. Res. Part D
Transp. Environ. 15 (2010) 298e303.
[5] M.Z. Jacobson, Global direct radioactive forcing due to multicomponent
anthropogenic and naturals aerosols, J Geophys Res. 106 (2001) 1551e1568.
[6] D. Koch, Transport and direct radiative forcing of carbonaceous and sulphate
aerosols in the GSISS GCM, J Geophys Res. 106 (2011) 20311e20332.
[7] Directive 136/2014/CE of the European Parliament and of the Council, 2014.
[8] Directive 715/2007/CE of the European Parliament and of the Council, 2007.
[9] M. Lapuerta, O. Armas, J. Rodrguez-Ferna
ndez, Effect of biodiesel fuels on
diesel engine emissions, Prog. Energy Combust. Sci. 34 (2008) 198e223.
[10] O. Armas, M.R. Garca, A. Ramos, Impact of alternative fuels on performance
and pollutant emissions of a light duty engine tested under the new European
driving cycle, Appl. Energy 107 (2013) 183e190.
[11] Directive 28/2009/CE of the European Parliament and of the Council, 2009.
[12] A. Abu-Jrai, J. Rodrguez-Fern
andez, A. Tsolakis, A. Megaritis, K. Theinnoi,
R.F. Cracknell, Performance, combustion and emissions of a diesel engine
operated with reformed EGR. Comparison of diesel and GTL fuelling, Fuel 88
(2009) 1031e1041.
[13] B. Guan, R. Zhan, H. Lin, Z. Huang, Review of the state-of-the-art of exhaust
particulate lter technology in internal combustion engines, J Environ. Manag
154 (2015) 225e258.
[14] S.Y. Park, C. Rutland, Analysis of SCR performance differences caused from
ow characteristics of wall ow and ow through type substrate: a simulation
study, Chem. Eng. Sci. 88 (2013) 69e78.
[15] M. Koebel, M. Elsener, M. Kleeman, Urea-SCR: a promising technique to
reduce NOx emissions from automotive diesel engines, Catal. Today 59 (2000)
335e345.
[16] P. Forzatti, L. Lietti, I. Nova, E. Tronconi, Diesel NOx aftertreatment catalytic
technologies: analogies in LNT and SCR catalytic chemistry, Catal. Today 151
(2010) 202e211.
[17] S. Benramdhane, C.N. Millet, E. Jeudy, J. Lavy, V.B. Aube
, M. Daturi, Impact of
thermal and vehicle aging on the structure and functionalities of a lean NOx-
trap, Catal. Today 176 (2011) 56e62.
[18] J.P.A. Neeft, M. Makkee, J.A. Moulijn, Diesel particulate emission control, Fuel
Process Technol. 47 (1996) 11e69.
[19] D. Fino, V. Specchia, Open issues in oxidative catalysis for diesel particulate
abatement, Powder Technol. 180 (2008) 64e73.
[20] A.J. Torregosa, J.R. Serrano, F.J. Arnau, P. Piqueras, A uid dynamic model for
unsteady compressible ow in wall-ow diesel particulate lters, Energy 36
(2011) 671e684.
[21] M. Lapuerta, J. Rodrguez-Ferna
ndez, F. Oliva, Effect of soot accumulation in a
diesel particle lter on the combustion process and gaseous emissions, Energy
47 (2012) 543e552.
[22] C. Beatrice, S. Di Iorio, C. Guido, P. Napolitano, Detailed characterization of
particulate emissions of an automotive catalyzed DPF using actual regenera-
tion strategies, Exp. Therm. Fluid Sci. 39 (2012) 45e53.
[23] A.L. Boehman, J. Song, M. Alam, Impact of biodiesel blending on diesel soot
and the regeneration of particulate lters, Energy Fuels 19 (2005) 1857e1864.
[24] J. Song, M. Alam, A.L. Boehman, Impact of alternative fuels on soot properties
and DPF regeneration, Combust. Sci. Technol. 179 (2007) 1991e2037.
[25] M. Lapuerta, J.J. Hern
andez, F. Oliva, Strategies for active diesel particulate
lter regeneration based on late injection and exhaust recirculation with
different fuels, Int. J Engine Res. 15 (2014) 209e221.
[26] M. Lapuerta, F. Oliva, J.R. Agudelo, A.L. Boehman, Effect of fuel on the soot
nanostructure and consequences on loading and regeneration of diesel par-
ticulate lters, Combust. Flame 159 (2012) 844e853.
[27] O.M. Bhardwaj, B. Lers, B. Holderbaum, T. Koerfer. Utilization of HVO fuel
properties in a high efciency combustion system. Part 2: relationship of soot
characteristics with its oxidation behavior in DPF. SAE Paper 2014-01-2846.
[28] M. Lapuerta, J. Rodrguez-Ferna
ndez, J. Sa
nchez-Valdepen ~ as, M.S. Salgado,
Multi-Technique analysis of soot reactivity form conventional and parafnic
diesel fuels, Flow Turbul. Combust. 96 (2016) 327e341.
[29] V. Rathore, B.L. Newalkar, R.P. Badoni, Processing of vegetable oil for biofuel
production through conventional and non-conventional routes, Energy Sus-
tain Dev. 31 (2016) 24e49.
[30] M. Lapuerta, J. Rodrguez-Ferna
ndez, J.R. Agudelo, A.L. Boehman, Blending
scenarios for soybean oil derived biofuels with conventional diesel, Biomass
Bioenergy 49 (2013) 74e85.
[31] S.S. Gill, A. Tsolakis, K.D. Dearn, J. Rodrguez-Ferna
ndez, Combustion charac-
teristics and emissions of Fischer-Tropsch diesel fuels in IC engines, Prog.
Energy Combust. Sci. 37 (4) (2011) 503e523.
[32] J.P.A. Neeft, T.X. Nijhuis, E. Smakman, M. Makkee, J.A. Moulijn, Kinetics of the
oxidation of diesel soot, Fuel 76 (12) (1997) 1129e1136.
[33] B.R. Stanmore, P. Gilot, The inuence of sample containment on the ther-
mogravimetric measurement of carbon black reactivity, Thermochim. Acta
261 (1995) 151e164.
[34] R.L. Vander Wal, V.M. Bryg, M.D. Hays, Fingerprinting soot (towards source
identication): physical structure and chemical composition, J Aerosol Sci. 41
(2010) 108e117.
[35] T. Ishiguro, N. Suzuki, Y. Fujitani, H. Morimoto, Microstructural changes of
diesel soot during oxidation, Combust. Flame 85 (1991) 1e6.
[36] J. Song, J. Wang, A.L. Boehman, The role of fuel-borne catalyst in diesel par-
ticulate oxidation behavior, Combust. Flame 146 (2006) 73e84.
[37] R.L. Vander Wal, A. Yezerets, N.W. Currier, D.H. Kim, C.M. Wang, HRTEM study
of diesel soot collected from diesel particulate lters, Carbon 45 (2007)
70e77.
[38] A. Liati, P.D. Eggenschwiler, Characterization of particulate matter deposited
in diesel particulate lters: visual and analytical approach in macro-, micro-
and nano- scales, Combust. Flame 157 (2010) 1658e1670.
[39] K. Al-Qurashi, A.L. Boehman, Impact of exhaust gas recirculation (EGR) on the
oxidative reactivity of diesel engine soot, Combust. Flame 155 (2008)
675e695.
[40] A. Braun, N. Shah, F.E. Huggins, K.E. Kelly, A. Sarom, C. Jacobsen, H. Francis,
J. Ilavski, G.E. Thomas, G.P. Huffman, X-ray scattering and spectroscopy
studies on diesel soot from oxygenated fuel under various engine load con-
ditions, Carbon 43 (2005) 2588e2599.
[41] J. Song, M. Alam, A.L. Boehman, U. Kim, Examination of the oxidation behavior
of biodiesel soot, Combust. Flame 146 (2006) 589e604.
[42] A. Sadezky, H. Muckenhuber, H. Grothe, R. Niessner, U. Po schl, Raman
microspectroscopy of soot and related carbonaceous materials: spectral
analysis and structural information, Carbon 43 (2005) 1731e1742.
[43] S.K. Sze, N. Siddique, J.J. Sloan, R. Escribano, Raman spectroscopic character-
ization of carbonaceous aerosols, Atmos. Environ. 35 (2001) 561e568.
[44] M. Lapuerta, F. Oliva, J.R. Agudelo, J.P. Stitt, Optimization of Raman spectros-
copy parameters for characterizing soot from different diesel fuels, Combust.

ndez et al. / Renewable Energy 104 (2017) 30e39
J. Rodrguez-Ferna
Sci. Technol. 183 (2011) 1203e1220.
[45] M. Salamanca, F. Mondrago
n, J.R. Agudelo, A. Santamara, Inuence of palm oil
biodiesel on the chemical and morphological characteristics of particulate
matter emitted by a diesel engine, Atmos. Environ. 62 (2012) 220e227.
[46] J. Rodrguez-Fern
andez, F. Oliva, R.A. V
azquez, Characterization of the diesel
soot oxidation process through an optimized thermogravimetric method,
Energy Fuels 25 (2011) 2039e2048.
[47] D. Buono, A. Senatore, M.V. Prati, Particulate lter behavior of a diesel engine
fueled with biodiesel, Appl. Therm. Eng. 49 (2012) 147e153.
[48] A.L. Boehman, J. Song, M. Alam, Impact of biodiesel blending on diesel soot
and the regeneration of particulate lters, Energy Fuels 19 (2005) 1857e1864.
[49] M. Lapuerta, J.J. Hern
andez, F. Gime
nez, Evaluation of exhaust gas recircula-
tion as a technique for reducing diesel engine NOx emissions, J Automob. Eng.
214 (2000) 85e93.
[50] D.R. Tree, K.I. Svensson, Soot processes in compression ignition engines, Prog.
Energy Combust. Sci. 33 (2007) 272e309.
39
[51] H. Omidvarborna, A. Kumar, D.S. Kim, Recent studies on soot modeling for
diesel combustio
n, Renew. Sustain Energy Rev. 48 (2015) 635e647.
[52] M. Lapuerta, O. Armas, A. Go
mez, Diesel particle size distribution estimation
from digital image analysis, Aerosol Sci. Technol. 37 (2003) 369e381.
[53] P. Eastwood, Critical Topics in Exhaust Gas Aftertreatment, 1 st ed, Research
Studies Press LTD, Baldock, Hertfordshire, England, 2000.
[54] M. Schejbal, M. Marek, M. Kub cek, P. Ko
ci, Modelling of diesel lters for
particulate removal, Chem. Eng. J. 154 (2009) 219e230.
[55] B.R. Stanmore, J.F. Brilhac, P. Gilot, The oxidation of soot: a review of exper-
iments, mechanisms and models, Carbon 39 (2001) 2247e2268.
[56] A. Messerer, F. Niessner, U. Poschl, Comprehensive kinetic characterization of
the oxidation and gasication of model and real diesel soot by nitrogen oxides
and oxygen under engine exhaust conditions measurements, Langmuir e
Hinshelwood, and Arrhenius parameters, Carbon 44 (2006) 307e324.
[57] B.R. Stanmore, V. Tschamber, J.F. Brilhac, Oxidation of carbon by NOx, with
particular reference to NO2 and N2O, Fuel 87 (131) (2008) 46.