-
Oxidation Characteristics of Soot
from a Gasoline Direct-Injection (GDI) Engine
Seungmok Choi, HeeJe Seong, Kyeong O. Lee
Transportation Technology R&D Center
Argonne National Laboratory
2014 DOE CLEERS Workshop
Dearborn, Michigan, USA
April 30, 2014
-
Background and Objective
Background
– Higher PN of GDI engines: gasoline particulate filter
(GPF)
– Most of GDI particulate emissions are formed during cold start
(warm-up) and
transient periods.
• For GPF regeneration, the oxidation characteristics of cold
soot are important.
– Soot oxidation reactivity: Printex U (flame soot) < Diesel
soot < GDI soot
– Diesel soot oxidation reactivity depends on engine conditions
(speed, load,
EGR, inj. timing): changes in carbon nanostructure and chemical
properties
(organic fractions, SFG)
– Few studies about GDI soot oxidation characteristics.
Objectives
– Investigating GDI soot oxidation characteristics in relations
to engine operating
conditions and TWC effects.
– Proposing a kinetic correlation relevant to GDI soot which can
be used for
simulation of GPF regeneration
2
-
Experimental setup with 2.4L 4-cylinder NA GDI
engine – homogenous/stoichiometric charge strategy
3
-540 -480 -420 -360 -300 -240 -180 -120 -60 0 60 120 180
Crank angle [°ATDC]
IntakeExhaust
Singleinjection
Sparktiming
-
0
5
10
15
20
25
30
35
0 1 2 3 4
Soo
t m
ass
[μg/
cyc]
Ash fraction [mass %]
1250rpm-25%
1500rpm-50%
3000rpm-50%
Cold idle
Ash fraction is much higher in GDI soot than in diesel
soot, primarily due to lower soot mass emissions of
the GDI engine
Ash fraction (in mass %) in engine-out soot will vary as
function of soot mass emission and lube oil consumption.
– Ash fraction in soot tends to increase with lower soot mass
calibration.
4
Adv (330)
Adv (330) – postTWC
Rtd (190) Adv (330)
Rtd (190)
Adv (330) – postTWC
Default Adv (330) Adv (330) – postTWC
GDI soot at advanced or retarded IT shows similar levels of soot
mass and ash fraction (order of 0.1%) to diesel soot.
GDI soot with default calibration contains over 3% of ash due to
very low soot mass.
TWC decreases total soot mass, but increases ash fraction in
soot.
At cold start, GDI soot lies in diesel-like regime with high
soot mass and moves to the GDI engine regime as the engine warms
up.
GDI engine regime
Conventional diesel engine regime
High lube oil consumption
Low lube oil consumption
-
GDI soot oxidation reactivity is significantly enhanced
with increased ash fraction in soot
Catalytic effect of ash is one of the driving factors that
enhances oxidation reactivity of GDI soot.
5
0
20
40
60
80
100
0 50 100 150 200
Printex U
LD Diesel_Ash 0.36%GDI_Ash 0.10%GDI_Ash 0.58%GDI_Ash
1.35%GDI_Ash 3.36%GDI_Ash 17.27%
m/m
0 [
%]
Time [min.]
Increase of
ash fraction
0
20
40
60
80
100
120
0 5 10 15 20
50% Conv.
90% Conv.
Tim
e [
min
]
Ash fraction [%]
PU (90% Conv.)
PU (50% Conv.)
Default inj. Timing (Low soot, high ash)
Advanced inj. Timing (High soot, low ash)
TGA isothermal oxidation (600 °C, 8% O2, pre-treated by N2)
-
6
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0 0.2 0.4 0.6 0.8 1
Printex U
LD Diesel_A0.33%
GDI_Hot Steady_A0.10%
GDI_Hot Steady_A0.58%
GDI_Hot Steady_A3.36%
m
/min
s [
(mg
/min
)/m
g]
Conversion ()
Specific soot oxidation rates
GDI soot has unique oxidation characteristics different
from flame or diesel soot
A
C
B
The oxidation rates of GDI soot are significantly promoted with
increase of ash fractions.
Without ash, oxidation reactivity of GDI soot is lower than that
of Printex U.
GDI soot shows three-staged oxidation which differs from Printex
U/diesel soot.
– Initial stage (A): higher oxidation rate due to additional
oxidation of soluble organics and weakly bonded carbons (WBC)
– Intermediate stage (B): lower oxidation rate after completion
of SOF and WBC oxidation (sole carbon oxidation)
– Final stage (C): soot oxidation rate becomes higher, because
of additional catalytic effects of ash by higher ash-to-soot ratio
in the remaining sample.
-
TWC decreases SOF/WBC and increases the ash fraction
in GDI soot, and improves oxidation reactivity in overall
7
0
0.05
0.1
0.15
0.2
0 0.2 0.4 0.6 0.8 1
Hot Steady_Eout_ash0.58%
Hot Steady_TWCout_ash1.35%
m
/min
s [
(mg
/min
)/m
g]
Conversion ()
Engine-out
TWC-out
Ash
SOF/WBC
Three reasons of ash fraction increase after TWC
– Catalytic oxidation of organics, WBC, and carbon soot in the
TWC
– Physical loss (attachment on the TWC wall) of soot
particles
• Particles loss was measured by SMPS.
– Separation of catalyst supporting materials: SEM-EDS data
• Higher fractions of Mg, Al, and Si were measured from TWC-out
ashes.
Oxidation reactivity of TWC-out soot is enhanced by the increase
in ash fraction.
Specific soot oxidation rates
-
Cold condition GDI soot: ash plays less catalytic roles,
resulting in lower oxidation reactivity
Low ash (0.1%) soot represents intrinsic GDI soot oxidation
reactivity. (negligible ash effects)
The oxidation rates of cold GDI soot in the intermediate stage
is close to intrinsic GDI soot oxidation rate.
8
Hypothesis 1: GDI soot seems to have an “Intrinsic carbon
oxidation reactivity” which is unchanged at different engine
conditions.
Hypothesis 2: Ash in GDI soot may have different forms, offering
different levels of catalytic effects, depending on hot and cold
engine conditions.
Specific soot oxidation rates
With similar ash fractions, the oxidation rate is much slower
for cold idle soot than hot steady state soot.
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0 0.2 0.4 0.6 0.8 1
GDI_Hot Steady_A0.10%
GDI_Cold Idle_A0.50%
GDI_Hot Steady_A0.58%
m
/min
s [
(mg
/min
)/m
g]
Conversion ()
-
Hypothesis 1: HR-TEM and Raman spectroscopy revealed that
GDI soot nanostructures are well defined and unchanged by
engine conditions: “Constant intrinsic carbon soot
reactivity”
9
0
0.2
0.4
0.6
0.8
1
800 1000 1200 1400 1600 1800 2000
1800rpm-15%1800rpm-50%1800rpm-75%
No
rmali
ze
d In
ten
sit
y
Raman shift (cm-1
)0
0.2
0.4
0.6
0.8
1
1000 1500 2000
GDI: cold idle
GDI: 1250rpm-25% load
GDI: 1500rpm-50% loadN
orm
ali
zed
in
ten
sit
y
Raman shift (cm-1
)
Diesel soot nanostructures change significantly with engine
conditions. Different carbon soot reactivity
GDI soot nanostructures do not depend on engine conditions.
Constant carbon soot reactivity
2°CA advanced Inj.
2°CA Retarded Inj.
Cold idle
1500 rpm-50%
[TEM images: Yehliu, Combustion and Flame, 2013]
ST
SLδ ~ 0.1 mmT ~ Tad ~ 2700K
Unburned mixture
(λ=1)
Burned gas
[Source: Dec, SAE 970873, 1997]
Conventional diesel engine: spray combustion GDI engine (NA,
λ=1): premixed flame propagation
Flame structures and temperatures of diesel spray combustion
strongly depend on ambient P & T at injection, dilution (EGR),
and swirl.
Turbulence increases reaction surface area with highly wrinkled
and convoluted flame sheet. Flame sheet structure (thickness) and
temperature do not change much at different engine conditions.
-
Hypothesis 2: Three different states of ashes are
proposed for GDI soot oxidation
Combustion-derived ash precursor (Ash_C)
– Metallic oxide nano-particles generated during in-cylinder
combustion and soot formation processes.
– Tight-contact with soot particles in nano-scale, offering
strong catalytic effects on soot cake oxidation.
– Converted to oxidation-derived ash by sintering during soot
cake oxidation.
Unburned oil-derived ash precursor (Ash_U)
– Unburned oil additives (e.g., ZDDP or calcium sulfonate)
discharged to exhaust in cold engine condition.
Nearly no catalytic effects.
– Converted to oxidation-derived ash by oxidation and sintering
during soot cake oxidation.
Oxidation-derived ash (Ash_O)
– Metallic oxide micron-particles generated from the ash
precursors with soot cake oxidation.
– Loose-contact with soot particles, offering weak catalytic
effects on soot cake oxidation.
– Contribution increases at the final soot oxidation stage with
increase in ash-to-soot ratio.
10
Soot primary particle
Ash_C
Ash_U
Ash_O
Exhaust emissions Soot cake
-
A global GDI soot oxidation mechanism is proposed
which includes the effects of organics/WBC and ash
Assumptions in mechanism
– Constant intrinsic carbon soot oxidation reactivity
– Three states of ashes
– Conversion of ash precursors to Ash_O at the same rate as soot
conversion
11
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
m
/min
s [
(mg
/min
)/m
g]
Conversion ()
Measured oxidation rate
Carbon oxidation shifted up by
Ash_C
Carbon oxidation assisted by Ash_O
Intrinsic carbon soot oxidation
Oxidation of SOF/WBC
Carbon oxidation assisted by Ash_C
0
0.01
0.02
0.03
0.04
0.05
0.2
0.4
0.6
0.8
1
0 0.2 0.4 0.6 0.8 1
Ox
ida
tio
n d
eri
ve
d a
sh
co
nte
nt
[mas
s %
]
Conversion ( )
0.10%
0.58%1.35%
17.3%
-
A modified kinetic correlation for GDI soot is derived
in consideration of SOF, WBC and ash effects
12
𝛼 = 𝑚𝐶𝑜𝑛𝑣/𝑚0
𝑟 =𝑑𝛼
𝑑𝑡= 𝐴 × 𝑒−𝐸𝑎/𝑅𝑇 × (1 − 𝛼)𝑛
𝑟 =𝑑𝛼
𝑑𝑡= (𝑚𝑆−𝑊,0/𝑚𝑠𝑜𝑜𝑡,0) ∙ 𝑟𝑆−𝑊 + (𝑚𝐶,0/𝑚𝑠𝑜𝑜𝑡,0) ∙ 𝑟𝐶
𝑟𝑆−𝑊 =𝑑𝛼𝑆−𝑊
𝑑𝑡= 𝐴𝑆−𝑊 × 𝑒
−𝐸𝑎,𝑆−𝑊/𝑅𝑇 × (1 − 𝛼𝑆𝑂𝐹)𝑛𝑆−𝑊
𝑟𝐶 =𝑑𝛼𝐶𝑑𝑡
= 𝐴𝐶 × 𝑒−𝐸𝑎,𝐶/𝑅𝑇 × (1 − 𝛼𝐶)
𝑛𝐶+[𝐴𝑠ℎ_𝑂 𝑎𝑠𝑠𝑖𝑠𝑡𝑒𝑑]
𝐴𝑠ℎ_𝑂 𝑎𝑠𝑠𝑖𝑠𝑡𝑒𝑑 =𝑑𝛼𝐴𝑠ℎ_𝑂 𝑎𝑠𝑠𝑖𝑠𝑡𝑒𝑑
𝑑𝑡= (1 − 𝛼𝐶) ∙ 𝑎 ∙ exp 𝑏 ∙ 𝑇 ∙ 𝑓𝐴𝑠ℎ_𝑂,𝑖
Reaction order (n) of soot samples
The typical kinetic correlation doesn’t hold for GDI soot
oxidation.
– Activation energy (Ea) and reaction order (n) change with
conversion, due to the effects of organics/WBC and ash.
A modified kinetic correlation has been developed for accurate
prediction of soot oxidation.
– The effects of SOF/WBC and ash are taken into account.
– Soot oxidation rates can be predicted at different engine
conditions without changing kinetic parameters.
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
m
/m [
(mg
/min
.)/m
gin
sta
nt]
Conversion ( )
GDI soot oxidation mechanism
-11
-10
-9
-8
-7
-2.5 -2 -1.5 -1 -0.5 0
ln(dα
/dt)
ln(1-α)
Printex U
GDI-1
GDI-3
GDI-4
GDI-5
-
0
0.005
0.01
0.015
400 450 500 550 600 650 700
ExperimentTypical kinetic correlationModified kinetic
correlationCarbon oxidation by Ash_CCarbon oxidation by
Ash_OSOF/WBC oxidation
m
/m0 [
(mg
/min
.)/m
g]
Temperature [degC]
0
0.01
0.02
0.03
0.04
0.05
0.2 0.4 0.6 0.8 1
ExperimentTypical kinetic correlationModified kinetic
correlationCarbon oxidation by Ash_CCarbon oxidation by
Ash_OSOF/WBC oxidation
m
/min
s [
(mg
/min
.)/m
g]
Conversion ()
The modified kinetic correlation predicts accurate
oxidation rate of GDI soot at a wide ranges of
conversion and temperature.
13
Typical kinetic correlation
Modified kinetic correlation
Typical kinetic correlation
Modified kinetic correlation
Carbon oxidation by Ash_C
SOF/WBC oxidation Carbon oxidation
By Ash_O
Carbon oxidation by Ash_C
SOF/WBC oxidation
Carbon oxidation By Ash_O
Isothermal oxidation (600 °C, 8% O2) Non-isothermal
oxidation
(increased by 1°C/min, 8% O2)
-
Summary
Major characteristics of GDI soot oxidation
– High ash fraction: an order of magnitude higher ash fraction
than in diesel soot.
– Catalytic effects of ash: dominantly enhances soot oxidation
reactivity.
– Three-staged oxidation: additional SOF/WBC oxidation at
initial, and Ash_O assisted carbon oxidation at final.
– Constant intrinsic carbon soot oxidation reactivity,
independent to engine conditions.
– TWC effect: enhances soot oxidation reactivity due to
increased ash fraction.
– Lower oxidation reactivity of cold condition soot: unburned
oil-derived ash precursor.
A general GDI soot oxidation mechanism and a modified kinetic
correlation have been proposed.
– The oxidation rates of GDI soot have been accurately predicted
at wide ranges of conversion/temperature and engine conditions,
without changing kinetic parameters.
14
-
Acknowledgement
Funding
– U.S. DOE Office of Vehicle Technologies
Industrial partners
– Hyundai Motor Company
– Corning Inc.
15
-
Thank you for your attention!
Contact Seungmok Choi
[email protected]
16
-
Technical support pages
17
-
Significance of cold and transient condition soot in
the GDI engine
18
0.004
0.057
0.022
0.206
0
0.1
0.2
0.3
Hot steady(1500rpm-25%)
Cold steady(1500rpm-25%)
Hot transient(Single ramp)
Cold transient(Single ramp)
Soo
t m
ass
[mg/
g Fu
el]
-
Contributions of SOF/WBC and ashes on GDI soot
oxidation of different engine conditions
19
0
0.5
1
1.5
2
GDI1 GDI2 GDI3 GDI4 GDI5
No
rmal
ized
So
ot
Oxi
dat
ion
Rat
e
SOF
Carbon_Ash_O
Carbon_Ash_C
Carbon_intrinsic
-
TWC effects on soot mass/particle size distribution
and ash composition
20
SEM-EDS (Compositions of ash)
0.E+00
2.E+06
4.E+06
6.E+06
8.E+06
1.E+07
10 100 1000
dN
/dlo
gDp
[#/
cm3]
Mobility diameter [nm]
EGout_1500-25
TWCout_1500-25
0.E+00
2.E+06
4.E+06
6.E+06
8.E+06
1.E+07
10 100 1000
dN
/dlo
gDp
[#/
cm3]
Mobility diameter [nm]
EGout_1500-75
TWCout_1500-75
MSS and SMPS (Soot mass and particle size distribution)
Soot mass: 1.0 0.2 mg/m3
Soot mass: 2.5 2.2 mg/m3
1.5%14.3%
0.5%0.3%0.9%
22.0%
3.2%40.7%
16.8%
F0.6%
Na18.5%
Mg4.3% Al
1.2%
Si8.6%
P25.0%
S2.0%
Ca30.8%
Zn9.0%
Engine-Out
TWC-Out
-
Validation of the modified kinetic correlation –
Isothermal 600˚C, 8% O2
21
0
0.01
0.02
0.03
0.04
0.05
0.2 0.4 0.6 0.8 1
GDI-1
ExperimentTypical kinetic correlation (10-90% Param.)Modified
kinetic correlationCarbon oxidation by Ash_CCarbon oxidation by
Ash_OSOF/WBC oxidation
m
/min
sta
nt
[(m
g/m
in)/
mg
]
Conversion ( )
0
0.01
0.02
0.03
0.04
0.05
0.2 0.4 0.6 0.8 1
GDI-2
ExperimentTypical kinetic correlation (10-90% Param.)Modified
kinetic correlationCarbon oxidation by Ash_CCarbon oxidation by
Ash_OSOF/WBC oxidation
m
/min
sta
nt
[(m
g/m
in)/
mg
]
Conversion ( )
0
0.01
0.02
0.03
0.04
0.05
0.2 0.4 0.6 0.8 1
GDI-3
ExperimentTypical kinetic correlation (10-90% Param.)Modified
kinetic correlationCarbon oxidation by Ash_CCarbon oxidation by
Ash_OSOF/WBC oxidation
m
/min
sta
nt
[(m
g/m
in)/
mg
]
Conversion ( )
0
0.02
0.04
0.06
0.08
0.1
0.2 0.4 0.6 0.8 1
GDI-4
ExperimentTypical kinetic correlation (10-90% Param.)Modified
kinetic correlationCarbon oxidation by Ash_CCarbon oxidation by
Ash_OSOF/WBC oxidation
m
/min
sta
nt
[(m
g/m
in)/
mg
]
Conversion ( )
0
0.02
0.04
0.06
0.08
0.1
0.2 0.4 0.6 0.8 1
GDI-5
ExperimentTypical kinetic correlation (10-90% Param.)Modified
kinetic correlationCarbon oxidation by Ash_CCarbon oxidation by
Ash_OSOF/WBC oxidation
m
/min
sta
nt
[(m
g/m
in)/
mg
]
Conversion ( )
-
Validation of the modified kinetic correlation –
Non-isothermal 1°C/min, 8% O2
22
0
0.005
0.01
0.015
400 450 500 550 600 650 700
GDI-1
ExperimentTypical kinetic correlation (10-90% Param.)Modified
kinetic correlationCarbon oxidation by Ash_CCarbon oxidation by
Ash_O
SOF/WBC oxidation
m
/m0 [
(mg
/min
)/m
g]
Temperature [degC]
0
0.005
0.01
0.015
400 450 500 550 600 650 700
GDI-2
ExperimentTypical kinetic correlation (10-90% Param.)Modified
kinetic correlationCarbon oxidation by Ash_CCarbon oxidation by
Ash_OSOF/WBC oxidation
m
/m0 [
(mg
/min
)/m
g]
Temperature [degC]
0
0.005
0.01
0.015
400 450 500 550 600 650 700
GDI-3
ExperimentTypical kinetic correlation (10-90% Param.)Modified
kinetic correlationCarbon oxidation by Ash_CCarbon oxidation by
Ash_OSOF/WBC oxidation
m
/m0 [
(mg
/min
)/m
g]
Temperature [degC]
0
0.005
0.01
0.015
400 450 500 550 600 650 700
GDI-4
ExperimentTypical kinetic correlation (10-90% Param.)Modified
kinetic correlationCarbon oxidation by Ash_CCarbon oxidation by
Ash_OSOF/WBC oxidation
m
/m0 [
(mg
/min
)/m
g]
Temperature [degC]
0
0.005
0.01
0.015
400 450 500 550 600 650 700
GDI-5
ExpSim_10-90
Param.Sim_ModelSim_Carbon_Ash_tSim_Carbon_Ash_lSim_SOF
m
/m0 [
(mg
/min
.)/m
gto
tal]
Temperature [degC]