You are on page 1of 9

1282

Energy & Fuels 2004, 18, 1282-1290

Comparison of the Impact of Intake Oxygen Enrichment


and Fuel Oxygenation on Diesel Combustion and
Emissions
Juhun Song, Vince Zello, and Andre L. Boehman*
The Energy Institute, The Pennsylvania State University, 405 Academic Activities Building,
University Park, Pennsylvania 16802

Francis J. Waller
Air Products and Chemicals, Inc., Allentown, Pennsylvania 18195
Received December 16, 2003. Revised Manuscript Received May 27, 2004

Among continuing efforts to develop low-emission combustion engines, oxygen-enhanced


combustion has long been considered a promising approach. A number of investigations have
focused on the effects of oxygen addition on soot formation and oxidation by using various oxygen
introduction techniques, such as blending different oxygen-containing fuels or direct oxygen
addition into the intake air stream. The present study of oxygen addition was performed on a
Volkswagen 1.9 L TDI turbodiesel engine to investigate and compare the relative effect of two
oxygen addition methods on diesel emission and combustion: oxygen enrichment of the intake
air and oxygenation of the fuel. The oxygen enrichment was accomplished by connecting an oxygen
generator to the intake air surge tank, while fuel oxygenation was accomplished using two
compounds with different cetane number and molecular structure. The key observations are that
both intake oxygen enrichment and fuel oxygenation via linear structure oxygenated molecules
are effective for reduction of diesel particulate matter, yielding even greater reductions in PM
emissions than for fuel oxygenation via ring-structured oxygenated molecules. However, NOx
emissions are greatly increased with intake oxygen enrichment, owing to either increased
availability of atomic oxygen or attainment of a higher temperature during leaner combustion,
which enhances the kinetics for thermal NOx formation. Comparison between the addition of
two substantially different oxygenated fuels, a mixture of glycol ethers and 1,3-dioxolane, has
also shed light on the mechanisms of soot reduction via oxygen addition. With their linear
structure, the glycol ethers were shown to be far more effective for soot reduction than an
equivalent oxygen addition via dioxolane, which has a ring structure, despite no significant
difference in heat release rate.

Introduction
Addition of oxygen to assist the diesel combustion
process is by no means a new idea. Many researchers
have been actively investigating the effects of oxygenates on diesel combustion, and reviews of many relevant
papers can be found in refs 1 and 2. In ongoing work by
Boehman and co-workers, a number of studies have
considered oxygenation of diesel fuel through additives
and alternative fuels.1,3-5 Oxygen enrichment of intake
air has also been considered as a measure to control the
* Corresponding author. E-mail: boehman@ems.psu.edu.
(1) Litzinger, T.; Stoner, M.; Hess, H.; Boehman, A. Effects of
Oxygenated Blending Compounds on Emissions from a Turbo-charged
Direct Injection Diesel Engine. Int. J. Engine Res. 2000, 1, 57-70.
(2) Natarajan, M.; Frame, E.; Naegli, D. W.; Asmus, T.; Clark, W.;
Garbak, J.; Gonzalez, M. A.; Liney, E.; Piel, W.; Wallace, J. P.
Oxygenates for Advanced Petroleum-Based Diesel Fuels: Part 1.
Screening and Selection Methodology for the Oxygenates. Society of
Automotive Engineers Technical Paper No. 2001-01-3631, 2001.
(3) Hess, H. S.; Roan, M. A.; Bhalla, S.; Butnark, S.; Zarnescu, V.;
Boehman, A. L.; Tijm, P. J. A.; Waller, F. J. Reduction of Particulate
Emissions from a Single-Cylinder IDI Engine with Oxygenated Diesel
Fuels. In Chemistry of Diesel Fuels; Song, C., Hsu, C. S., Mochida, I.,
Eds.; Taylor & Francis: New York, 2000; Chapter 14, pp 255-288.

PM emissions, to improve thermal efficiency, and to


improve the ignition quality of diesel fuels. An example
of such work is the efforts at Argonne National Laboratory where researchers have developed a membrane
that efficiently separates standard air into oxygen and
nitrogen.6 Some discussion of the relevant literature on
these two subjects follows.
The consensus in the literature is that incorporation
of oxygen into diesel fuel reduces particulate emissions.
Where there has been disagreement is on the role of
oxygenate structure in controlling the amount of PM
reduction and the mechanism by which oxygenated fuel
(4) Hess, H. S.; Boehman, A. L.; Tijm, P. J. A.; Waller, F. J.
Experimental Studies of the Impact of CETANERTM on Diesel
Combustion and Emissions. Society of Automotive Engineers Technical
Paper No. 2000-01-2886, 2000.
(5) Chapman, E. M.; Bhide, S. V.; Boehman, A. L.; Tijm, P. J. A.;
Waller, F. J. Emission Characteristics of a Navistar 7.3L Turbodiesel
Fueled with Blends of Oxygenates and Diesel. Society of Automotive
Engineers Technical Paper No. 2000-01-2887, 2000.
(6) Poola, R. B.; Longman, D. E.; Stork, K. C.; Sekar, R.; Nemser,
S.; Callaghan, K. Membrane-Based Nitrogen-Enriched Air for NOx
Reduction in Light-Duty Diesel Engine. Society of Automotive Engineers Technical Paper No. 2000-01-0228, 2000.

10.1021/ef034103p CCC: $27.50 2004 American Chemical Society


Published on Web 07/10/2004

Oxygen Enrichment and Diesel Combustion/Emission

reduces PM emissions. In some recent work by researchers at Sandia National Laboratory and co-workers, both
of these issues have been clarified. Dec and co-workers
have considered the impact of oxygenated fuels on the
diesel spray combustion process, through a combination
of analytical and experiment information.7 Chemical
kinetic analysis and consideration of the fuel-air mixing
process during combustion of n-heptane/methanol and
n-heptane/dimethyl ether mixtures showed that the
percentage mass of carbon existing as soot precursors
falls off with increasing O/C atomic ratio and that CO
concentrations will be elevated. With more of the fuel
carbon bound up as CO, less carbon is available to form
particulate precursors. While this CO formation alone
cannot account for the reductions in soot observed for
many oxygenated fuels, the combination of the CO
formation from oxygen in the fuels, CO liberated
through bond cleavage in the oxygenate (e.g., for the
glycol ethers), the scavenging of H by CO and CO2, and
the attack of soot precursors by OH that is produced by
the H + CO2 reaction, all contribute to the suppression
of soot. Recently, Mueller and co-workers have used an
optically accessible DI diesel engine to examine the
impact of oxygenate structure on the soot formation
process, providing additional support for the model of
soot suppression by oxygenates and demonstrating that
oxygenate structure does affect soot suppression.8,9 In
this work, Mueller and co-workers compared di-butyl
maleate and tri-propylene glycol methyl ether at fixed
oxygen ratio, defined as the amount of oxygen available in the reactants divided by the amount required
for stoichiometric combustion.9,10 The oxygen ratio is
shown to be a more effective basis for comparison than
equivalence ratio, O/C ratio, and oxygen mass fraction.
In the work presented here, however, the more traditional O/C ratio will be used as the basis for comparison.
From laminar flame studies where gaseous oxygen
is added to the fuel stream, oxygen addition effects on
soot formation and oxidation were identified to provide
some implications for diesel engine combustion: (1) a
dilution effect resulting from the change in the amount
of carbon per unit mass of gaseous fuel mixture that
will later involve soot precursor formation, (2) a thermal
effect due to flame temperature change upon diluent
addition, and (3) a direct chemical suppression or soot
formation increase.11 In some experiments, thermal and
dilution effects by certain oxygen-containing additives
are believed to be major factors leading to suppression
(7) Flynn, P. F.; Durrett, R. P.; Hunter, G. L.; zur Loye, A. O.;
Akinyemi, O. C.; Dec, J. E.; Westbrook, C. K. Diesel Combustion: An
Integrated View Combining Laser Diagnostics, Chemical Kinetics, and
Empirical Validation. Society of Automotive Engineers Technical Paper
No. 1999-01-0509, 1999.
(8) Mueller, C. J.; Martin, G. C. Effects of Oxygenated Compounds
on Combustion and Soot Evolution in a DI Diesel Engine: Broadband
Natural Luminosity Imaging. Society of Automotive Engineers Technical Paper No. 2002-01-1631, 2002.
(9) Mueller, C. J.; Pitz, W. J.; Pickett, L. M.; Martin, G. C.; Siebers,
D. L.; Westbrook, C. K. Effects of Oxygenates on Soot Processes in DI
Diesel Engines: Experiments and Numerical Simulation. Society of
Automotive Engineers Technical Paper No. 2003-01-1791, 2003. (Also,
JSAE 20030193).
(10) Upatnieks, A.; Mueller, C. J. Investigation of the Relationships
Between DI Diesel Combustion Processes and Engine-Out Soot Using
An Oxygenated Fuel. Society of Automotive Engineers Technical Paper
No. 2004-01-1400, 2004.
(11) Glassman, I. Sooting Laminar Diffusion Flames: Effect of
Dilution, Additives, Pressure, and Microgravity. Twenty-Seventh Symposium (International) on Combustion; The Combustion Institute:
Pittsburgh, 1998; pp 1589-1596.

Energy & Fuels, Vol. 18, No. 5, 2004 1283

of soot formation. However, there is still controversy


about the presence of the chemical suppression effect,
as well as its variation in effectiveness of oxygenates
with different structure at reducing soot even though
other experimental studies have thoroughly demonstrated the suppression of soot formation by alcohol
addition and provided information on the apparent
influence on the chemical kinetic pathways by which
soot is suppressed.12,13 With increasing oxygen content
in the oxidizer stream, increase of soot formation
together with modest reductions in soot yields have been
reported in laminar premixed diffusion flame studies.14,15 They attributed this soot suppression effect of
increasing the oxygen concentration on the oxidizer side
entirely to a thermal effect. The corresponding dilution
and chemical effects were found to be negligible.
Instead, examination of oxygen addition via fuel
oxygenation in a practical diesel engine presents experimental difficulties in isolating such effects. Since
changes in fuel properties such as volatility and cetane
number can cause the additional thermal and dilution
effects through air entrainment during the ignition
delay period, it is hard to attribute the observed
emission changes purely to the chemical effect of oxygen
addition without a careful consideration of fuel property
effects at the same time. Furthermore, the typically
lower heating value and higher density of the oxygenated fuel might change the timing of the combustion
process as the engines control system makes injection
timing adjustments in response to fuel property changes
which compounds the oxygen addition effect, unless fuel
injection timing can be held fixed. In this regard, the
recent work of Mueller and co-workers provides invaluable insights into the mechanism of soot suppression
by fuel oxygenation through their ability to perform
precisely controlled experiments and to apply sophisticated laser diagnostic techniques to examine the incylinder processes.8,9 A key insight reported by Mueller
and Martin is that differences of a factor of 2 in the
spatially integrated natural luminosity of the diesel
spray flame (a relative measurement of soot volume
fraction) were observed between di-butyl maleate (a diester compound) and tri-propylene glycol methyl ether,
despite using fuel blends containing the same amount
of oxygen at the lift-off position.8 The follow-up studies
by Mueller et al., which included experiments in a
constant-volume combustion vessel and numerical simulation, further support the assertion that the effect of
oxygen addition to the fuel spray plume is to reduce
soot-precursor levels.9 Furthermore, their calculations
and experiments show that for some oxygenates, specifically di-butyl maleate, not all of the oxygen is available
(12) Frenklach, M.; Yuan, T. Proceedings of the 16th International
Symposium on Shock Tubes and Waves; VCH: Weinheim, 1987; pp
487-493.
(13) Ni, T.; Gupta, S.; Santoro, R. Suppression of Soot Formation
in Ethene Laminar Diffusion Flames by Chemical Additives. TwentyFifth Symposium (International) on Combustion; The Combustion
Institute: Pittsburgh, 1994; pp 585-592.
(14) Du, D. X.; Axelbaum, R. L.; Law, C. K. The Influence of Carbon
Dioxide and Oxygen as Additives on Soot Formation in Diffusion
Flames. Twenty-Third Symposium (International) on Combustion; The
Combustion Institute: Pittsburgh, 1990; pp 1501-1507.
(15) Lee, K.-O.; Megaridis, C. M.; Zelepouga, S.; Saveliev, A. V.;
Kennedy, L. A.; Charon, O.; Ammouri, F. Soot Formation Effects of
Oxygen Concentration in the Oxidizer Stream of Laminar Coannular
Nonpremixed Methane/Air Flames. Combust. Flame 2000, 121, 323333.

1284 Energy & Fuels, Vol. 18, No. 5, 2004

for soot precursor reduction, and for others, specifically


tri-propylene glycol methyl ether, all of the oxygen is
available for soot precursor reduction.
Some review of the literature on oxygen enrichment
is provided by Ghojel et al.16 and Desai et al.17 In an
early paper on oxygen enrichment, Karim and Ward
used the variation of intake composition to study the
relationship between the physical and chemical time
scales in the direct injection compression ignition engine.18 They also observed greatly reduced smoke emissions, higher peak cylinder pressure, increased power
out, and significantly shortened ignition delay with
oxygen enrichment up to 55 vol % O2 in the intake air.
They observed increased smoke emissions as the oxygen
content in the intake air was decreased, with degraded
combustion occurring at 7-8 vol % O2. The impact of
both nitrogen (N2) and oxygen (O2) enrichment have
been considered in recent work. At Inha University, the
impact of N2, carbon dioxide (CO2), and argon (Ar)
enrichment have been shown to significantly affect PM
and NOx emissions, with nitrogen enrichment leading
to a reduction in NOx emissions but a moderate increase
in PM.19 In studies by researchers at Argonne National
Laboratory, nitrogen enrichment has been shown to
reduce NOx emissions but increase PM emissions,6
although the PM emissions increase is similar to that
observed with EGR. However, across a range of engine
platforms, oxygen enrichment has been observed to have
beneficial effects on PM but mixed effects on NOx. With
significant optimization, Poola, Sekar and co-workers
have shown that oxygen enrichment can yield reduction
in both NOx and PM in locomotive diesel engines
through use of modest oxygen enrichment and retarded
fuel injection timing.20,21 Others have made similar
observations that retarded injection timing can counteract the adverse impact on NOx emissions from oxygen
enrichment.22-26 In an investigation by Schmidt and
(16) Ghojel, J.; Hilliard, J. C.; Levendis, J. A. Effect of Oxygen
Enrichment on the Performance and Emissions of I.I.I Diesel Engines.
Society of Automotive Engineers Technical Paper No. 830245, 1983.
(17) Desai, R. R.; Gaynor, E.; Watson, H. C.; Rigby, G. R. Giving
Standard Diesel Fuels Premium Performance Using Oxygen-Enriched
Air in Diesel Engines. Society of Automotive Engineers Technical Paper
No. 932806, 1993.
(18) Karim, G. A.; Ward, G. The Examination of the Combustion
Processes in a Compression-Ignition Engine by Changing the Partial
Pressure of Oxygen in the Intake Charge. Society of Automotive
Engineers Technical Paper No. 680767, 1968.
(19) Li, J.; Chae, J. O.; Park, S. B.; Paik, H. J.; Park, J. K.; Jeong,
Y. S.; Lee, S. M.; Choi, Y. J. Effect of Intake Composition on
Combustion and Emission Characteristics of DI Diesel Engine at High
Intake Pressure. Society of Automotive Engineers Technical Paper No.
970322, 1997.
(20) Assanis, D. N.; Poola, R. B.; Sekar, R.; Cataldi, G. R. Study of
Oxygen-Enriched Combustion Air in Locomotive Diesel Engines. J.
Eng. Gas Turbines Power 2001, 123, 157-166.
(21) Poola, R. B.; Sekar, R. Reduction of NOx and Particulate
Emissions by Using Oxygen-Enriched Combustion Air in a Locomotive
Diesel Engine. J. Eng. Gas Turbines Power 2003, 125, 524-533.
(22) Iida, N.; Suzuki, Y.; Sato, G. T.; Sawada, T. Effects of Intake
Oxygen Concentration on the Characteristics of Particulate Emissions
from a D. I. Diesel Engine. Society of Automotive Engineers Technical
Paper No. 861233, 1986.
(23) Watson, H. C.; Milkins, E. E.; Rigby, G. R. A New Look at
Oxygen Enrichment (1) The Diesel Engine. Society of Automotive
Engineers Technical Paper No. 900344, 1990.
(24) Sekar, R. R.; Marr, W. W.; Schaus, J. E.; Cole, R. L.; Marciniak,
T. J.; Eustis, J. N. Cylinder Pressure Analysis of a Diesel Engine Using
Oxygen-Enriched Air and Emulsified Fuels. Society of Automotive
Engineers Technical Paper No. 901565, 1990.
(25) Virk, K. S.; Kokturk, U.; Bartels, C. R. Effects of OxygenEnriched Air on Diesel Engine Exhaust Emissions and Engine
Performance. Society of Automotive Engineers Technical Paper No.
931004, 1993.

Song et al.

Van Gerpen of the mechanism of PM reduction with


biodiesel, an oxygen enrichment test was conducted to
simulate the oxygen content associated with the esters
in biodiesel. Similar soot reduction from oxygen enrichment was observed along with an increase on the NOx
emission.27 They also compared methyl soyate blends
with straight chain alkane blends with the intention of
isolating the effect of the cetane number and long chain
hydrocarbons on particulate emissions.
At the University of Wisconsin-Madison, Donahue
and Foster compared the impact of fuel oxygenation and
intake oxygen enrichment.28 They found that effectiveness for particulate matter reduction depended on the
local concentration of oxygen in the fuel plume, indicated by the local equivalence ratio. In contrast, they
found that the global thermal and dilution effects from
oxygen enrichment led to a far more significant effect
on NOx emissions than did fuel oxygenation. In subsequent work by Foster and co-workers, the use of
auxiliary gas injection for late cycle oxygen enrichment
was simulated in a heavy duty engine and was shown
to provide 80% soot emission reduction with negligible
effect on NOx through controlled enhancement of incylinder mixing.
This recent work and the earlier work on oxygen
enrichment has shed some light on the effects of intake
charge composition, but left many unanswered questions that served as motivation for the work presented
here. The objective of the present study was to pursue
studies on the impact of oxygen enrichment on combustion and emissions in a Volkswagen 1.9L TDI turbodiesel engine to clarify the relative effects of O2 enrichment and fuel oxygenation. To that end, a vacuum swing
absorption (VSA) oxygen generator (Air Products VSA
type A-040/120L Oxygen Generator) was attached to the
intake air management system of a diesel engine to
compare oxygen enrichment with oxygenation of the
fuel. Clarification of the impact of oxygenation and
oxygen enrichment can be accomplished by addition of
a cetane improver (ethyl hexyl nitrate, EHN) to match
the ignition delay of the baseline diesel fuel and the
oxygenated fuel whenever necessary, so as to clarify the
impacts of oxygen addition on pollutant formation.
Experimental Section
The experimental system consists of an engine test cell
(dynamometer, controller), an engine, combustion analysis
instrumentation, and emissions analyzers. Figure 1 presents
a schematic diagram of the VW TDI 1.9 L turbodiesel engine
and test cell instrumentation. Table 1 presents specifications
for the TDI test engine. The facility and experimental procedures for engine operation and emissions tests have been
described earlier.29 Steady-state engine operation at 75% load
(140 ft-lbs torque) and 1900 rpm was chosen for this study
because the effect of oxygenates was found to be more
pronounced at medium and high load than at low load. A
(26) Marr, W. W.; Sekar, R. R.; Cole. R. L.; Marciniak, T. J.;
Longman, D. E. Oxygen-Enriched Diesel Engine Experiments with a
Low-Grade Fuel. Society of Automotive Engineers Technical Paper No.
932805, 1993.
(27) Schmidt, K.; Gerpen, J. V. The effect of Biodiesel Fuel Composition on Diesel Combustion and Emissions. Society of Automotive
Engineers Technical Paper No. 961086, 1996.
(28) Donahue, R. J.; Foster, D. E. Effects of Oxygen Enhancement
on the Emissions from a DI Diesel via Manipulation of Fuels and
Combustion Chamber Gas Composition. Society of Automotive Engineers Technical Paper No. 2000-01-0512, 2000.

Oxygen Enrichment and Diesel Combustion/Emission

Energy & Fuels, Vol. 18, No. 5, 2004 1285

Figure 1. Schematic of test engine setup.


Table 1. Test Engine Specifications
engine

1997 VW Passat 1Z

bore stroke
displacement
compression ratio
rated power
peak torque
injection system
low idle speed
configuration

79.5 95.5 mm
1.9 liter
19.5:1
66 kW at 4000 rpm
202 N m at 1900 rpm
direct injection, EDC
903 rpm
turbocharged,
Intercooler (air-air),
electronically controlled EGR
with actuator

pressure transducer (Kistler 6053C) with 130 kHz frequency


response, mounted in the glow-plug hole in the first cylinder,
was used with a shaft encoder (Kistler 2611) to provide timeresolved pressure traces for heat-release rate calculations.
Pressure data were measured and processed at increments of
0.5 crank angle degrees over 100 consecutive cycles. The
variable error in the pressure measurement from 100 cycles
during the combustion period is less than 8%, which propagates into the subsequent heat release rate calculations
creating a similar level of uncertainty. Therefore, the associated experimental uncertainty has little impact on comparison
of heat release rate between the different fuels and resolution
of the premixed burn spike.
Within the heat release rate program, an ensemble-averaged
pressure was used to calculate the apparent heat release rate
through a high frequency rejection filtering algorithm, referred
to as a Butterworth filter that removes frequencies greater
than the threshold frequency. Since this filtering can slightly
shift the combustion phase, such as the start of combustion
(SOC), an additional correction was made to return to the
phase of the original unfiltered data. Determination of injection timing such as dynamic start of injection (SOI) was made
on the PC monitor by communicating with the ECM via the
onboard diagnostics (OBD) software. Emissions of particulate
matter, NOx, CO, CO2, and HC, were measured, respectively,
via Sierra Instruments BG1 dilution tunnel, a Nicolet Magna
550 FTIR spectrometer, and a California Analytical HFID
hydrocarbon analyzer.
(29) Song, J.; Cheenkakorn, K.; Wang, J.; Perez, J.; Boehman, A.
L.; Young, P. J.; Waller, F. J. Effect of Oxygenated Fuel on Combustion
and Emissions in a Light-Duty Turbo Diesel Engine. Energy Fuels
2002, 16, 294-301.

The intake air flow rate was directly measured via an


electronic mass air flow sensor. Fuel consumption rate was
recorded via fuel tank weight loss using a precision weighing
scale, and real-time fuel consumption rate was also measured
using two separate flowmeters (Micromotion D06) on the fuel
flow to the engine and fuel return from the engine. For oxygen
enrichment of the intake air, modification of the intake air
supply line was required. The oxygen was supplied by an
oxygen generator (Air Products VSA type A-040/120L) into
the intake air surge tank through a pressure regulator, a
metering valve, and flowmeter. Fumigation into the intake
surge tank was chosen to allow thorough mixing of the
oxygen-air mixture through the subsequent components of
the intake system, such as the air filter, turbocharger, and
intercooler, before entering the intake manifold. With known
air flow rates at the desired engine operation condition, the
required oxygen flow rate was adjusted to obtain the desired
volume percent oxygen enrichment. To verify the level of
oxygen enrichment, the oxygen-air mixture was sampled at
the air filter by using a Rosemount O2 analyzer (Model 755R).
Fuel oxygenation was accomplished using two compounds
with widely different cetane number and molecular structure.
A mixture of glycol ethers, 20 vol % monoglyme and 80 vol %
diglyme (referred to as CETANER in several publications3-5,29),
with a cetane number of 100 represented a cetane-improving
oxygenate. The glycol ethers tested in this study are reported
to be potential teratogens, raising health concerns about
genotoxicity that causes birth defects.30 In contrast, 1,3dioxolane with a cetane number of 30 represented a cetanesuppressing oxygenate. Both compounds were soluble in diesel
fuel and provided a means of distinguishing between cetane
number and oxygen addition effects. Figure 2 shows the
molecular structures of the oxygenates used in this study.
Tables 2-4 show the properties of the reference fuels and their
blends.

Results and Discussion


Effect of Different Oxygen Addition Methods.
The experiments presented here were intended to
investigate and compare the relative effect of two oxygen
addition methods on diesel combustion and emissions,
(30) McGregor, D. B. Genotoxicity of Glycol Ethers. Environ. Health
Perspect. 1984, 57, 97-103.

1286 Energy & Fuels, Vol. 18, No. 5, 2004

Song et al.

Table 2. Properties of the Oxygenated and Alkane Compounds


formula
cetane number
boiling point (C)
oxygen (wt %)
specific density (g/cm3)
heating value (MJ/kg)
a

base diesel

monoglyme

diglyme

dioxolane

n-heptane

n-dodecane

C15H27.66a

C4H10O2
100
85
36.6
0.867
31.6

C6H14O3
112
162
35.8
0.943
30.3

C3H6O2
30
78
43
1.060
35

C7H16
56
98.4
0
0.684
44.9

C12H26
87.6
216
0
0.749
44.5

55.4
180/370
0
0.833
45.8

This chemical formula is used to calculate oxygen-to-carbon ratio.

Figure 2. Molecular structures of the oxygenates used in the


present study: (a) monoglyme, (b) diglyme, and (c) 1,3dioxolane.

Figure 3. Particulate matter emissions for different means


of oxygen addition, g/kWh basis.
Table 3. Properties of Base Test Fuels
base diesel
cetane number
heating value (MJ/kg)
specific density (g/cm3)
Saybolt viscosity (SUS) @38 C
sulfur (wt ppm)
aromatics (wt %)
oxygen (wt %)

55.4
45.8
0.833
34.3
39.0
24.3
0.00

via intake oxygen enrichment and fuel oxygenation. To


directly compare both oxygen enrichment of the intake
air and oxygenation of the fuel under the same extent
of oxygen addition, the oxygen-to-carbon ratio (O/C) in
the overall mixture was kept constant. An elementary
analysis was performed to match the oxygen-to-carbon
ratio on the basis of the overall in-cylinder gas compositions as shown in Table 4, in the same manner as used
by Donahue et al.28
As seen in Figure 3, whether it comes from intake
oxygen enrichment or via oxygenated fuel, oxygen
addition reduces soot, while the extent of PM reduction

Figure 4. Cylinder pressure and heat release rates for


different means of oxygen addition at O/C ) 5.1 condition.

varies remarkably with method of oxygen addition. For


the equivalent level of oxygen addition based on an
oxygen-to-carbon ratio of 5.1, oxygen enrichment of the
intake air was as effective for reducing particulate
matter emissions as fuel-bound oxygen addition via the
glycol ether blends. The difference in PM reduction was
not statistically significant between intake air enrichment and the glycol ethers blend, as seen in Figure 3
where the error bars indicate the 90% confidence
interval for the data by accumulating all error sources
evolved in this test. However, it is evident that fuel
oxygenation via the 1,3-dioxolane blends is the least
effective for reductions in PM emissions.
This significant effectiveness of intake oxygen enrichment might be explained by combined effects of the
thermal and chemical interaction arising from intake
oxygen addition. The heat release rate in Figure 4
indicates that oxygen enrichment lowered the peak
premixed burn, which leads to lower soot precursor
formation, without any shift in combustion phase from
the baseline. The resulting increased heat release in the
mixing-controlled combustion phase will also contribute
to increased soot oxidation. A similar trend of heat
release characteristics has been observed previously for
intake oxygen enrichment by Iida et al.,22 Desai et al.,17
Donahue et al.28 and others. It is well-known that cetane
improvers enhance ignition quality by generating a
radical pool at a lower temperature than the components in the base fuel. Recent work by Yonei et al. has
demonstrated through quantum mechanical calculations of radical decomposition and reaction that the best
radical species for improving cetane number are those
that promote the generation of an alkyl peroxy radical
from an alkyl radical through an oxygen addition
reaction.31 This oxygen addition reaction can presumably also be enhanced by having a greater concentration
of O2 available. So oxygen enrichment can achieve the

Oxygen Enrichment and Diesel Combustion/Emission

Energy & Fuels, Vol. 18, No. 5, 2004 1287

Table 4. Calculation of Diluted Properties of Base Fuel and Blended Fuels at O/C ) 4.8 Condition
fuel blends

blend ratio
(wt %)

O/C

mol %
of blend

C%
of blend

aromatics
(wt %)

sulfur
(ppm)

Mair
(g/s)

Mf
(g/s)

base fuel
glycol ethers
n-heptane/n-dodecane
dioxolane
n-heptane

15
15
10
10

4.6
4.8
4.6
4.8
4.6

17.9
15.0
23.8
18.7

9.8
14.7
5.9
9.7

24.3
20.6
20.6
21.9
21.9

39
33
33
35
35

39.9
39.9
39.9
39.9
39.9

1.735
1.823
1.629
1.757
1.629

Table 5. Combustion Event Analysis for Only Different


Oxygenate Blends at O/C ) 5.1 Condition
fuel blends

dynamic SOI
(ATDC)

SOC
(ATDC)

ignition delay
(CA)

base fuel
2 vol % O2 addition
30 wt % glycol ethers
20 wt % dioxolane

-4.7
-4.7
-5.9
-5.7

-0.8
-0.9
-1.9
-1.2

3.9
3.8
4.0
4.5

same end as a cetane improver, by shifting the explosion


limits for the reactant mixture to lower temperature and
lower pressure. Another possible explanation is that the
presence of additional OH and O in the mixingcontrolled phase will more likely attack the soot precursors, which results in lower acetylene concentration as
postulated by Gulder.32 The free oxygen supplied by the
intake air can be the key species affecting the soot
formation chemistry, thereby reducing soot formation
and growth and increasing the oxidation of the soot.
Table 5 presents fuel injection timing changes that
occurred with the different test fuels. The oxygenated
fuels were seen to advance the start of injection timing
in Figure 4 and Table 5 because the engine maintains
power at the requested level through injection timing
adjustment to compensate for the higher density and
lower heating value of the oxygenated fuel as compared
to the baseline fuel, as summarized in Table 3. Specifically, injection timing for the glycol ether blend was
more advanced than that for 1,3-dioxolane, but the high
cetane number of the glycol ether blend compensates
this advanced injection timing so that ignition delay was
similar to the baseline and 23% intake oxygen enrichment, as shown in Table 5. For the same ignition delay,
the glycol ether blend was also observed to lower the
peak premixed burn as with intake oxygen enrichment,
which leads to the lower soot formation. The heat
release characteristics of the glycol ethers shown in
Figure 4 are consistent with the observations reported
by Beatrice and co-workers, that the premixed burn rate
decreases while the diffusion heat release rate increases
at high load. The higher diffusion heat release was
attributed to increased oxidation of the combustion
intermediates by oxygen present in the additive.33
Litzinger et al. also reported the same trend of a lower
premixed burn rate when diglyme is blended with
diesel.1,34
(31) Yonei, T.; Hashimoto, K.; Arai, M.; Tamura, M. Quantum
Chemical Study of Cetane Improvers. Energy Fuels 2003, 17, 725730.
(32) Gulder, O. L. Effects of Oxygen on Soot Formation in Methane,
Propane and n-Butane Diffusion Flames. Combust. Flame 1995, 101,
302-310.
(33) Bertoli, C.; Del Giacomo, N.; Beatrice, C. Diesel Combustion
Improvements By the Use of Oxygenated Synthetic Fuels. Society of
Automotive Engineers Technical Paper No. 972972, 1997.
(34) Stoner, M.; Litzinger, T. Effects of Structure and Boiling Point
of Oxygenated Blending Compounds in Reducing Diesel Emissions.
Society of Automotive Engineers Technical Paper No.1999-01-1475,
1999.

Figure 5. NOx emissions for different means of oxygen


addition, g/kWh basis.

On the other hand, the dioxolane blend produces a


longer ignition delay compared to the glycol ether blend
so that a greater premixed burn occurs, which is one of
the contributing factors for reducing soot. This longer
ignition delay was expected since the dioxolane blend
has a lower boiling point and lower cetane number, in
addition to yielding advanced injection timing. An
increased ignition delay presumably could account for
a portion of the soot reduction observed with dioxolane,
since the longer ignition delay gives the fuel and air time
to mix, thereby resulting in leaner combustion during
the fuel-rich premixed phase.11 Furthermore, the high
volatility of the dioxolane blend may also contribute to
the reduced soot formation since the formation of a fuelrich mixture may not be as likely in the fuel spray
plumes due to enhanced air entrainment.9,34 Therefore,
the soot reduction with the dioxolane blends may result
directly from leaner combustion through fuel-air mixing, rather than through oxygen addition to the fuel.
Consequently, for dioxolane the oxygen content is far
less effective for reducing soot. In the recent oxygenated
fuel study by Litzinger et al.,1 they postulated that a
key issue in understanding possible effects of the
oxygenated compounds is the chemical structure of
parent fuel molecules from which the oxygen (or oxygenbearing fragments) will be released when these compounds decompose. So differences in the molecular
structure of the oxygenated fuels can partly explain the
differences in PM reduction between the glycol ether
and dioxolane blends, as will be discussed in the next
section.
As shown in Figure 5, the unfavorable effect of oxygen
addition on NOx emissions becomes greater in the order
of 1,3-dioxolane, glycol ethers and intake oxygen enrichment. This trend follows the typical PM-NOx tradeoff
behavior, i.e., the lower the PM, the higher the NOx.35
In other words, the increased oxygen concentration
(35) Heywood, J. B. Internal Combustion Engine Fundamentals;
McGraw-Hill: New York, 1988.

1288 Energy & Fuels, Vol. 18, No. 5, 2004

Figure 6. CO emissions for different means of oxygen


addition, g/kWh basis.

causes a NOx increase in all cases, but the extent of the


increase depends strongly on how the oxygen is added.
Intake oxygen enrichment leads to increases in NOx
emissions by 53 and 126% with respect to baseline for
22 vol % and 23 vol % oxygen concentration, respectively. Considering that NOx is more likely to form in a
post flame zone where the mixing controlled regime
dominates, heat release in the mixing controlled phase
of combustion might be one of the key factors. Under
the present engine operation where mixing controlled
combustion occurs, the higher heat release rate of the
mixing controlled combustion phase leads to higher
temperatures for the intake oxygen enrichment case
shown in Figure 4. This thermal effect contributes to
the NOx increase in combination with the local
(equivalence ratio) effect. According to local theory,
the available oxygen in the high-temperature flame
region may shift this locally rich mixture toward a
slightly lean mixture where the maximum equilibrium
O-atom mole fraction lies.28 Consequently, the high
equilibrium concentration of oxygen atom along with
super-equilibrium O-atom results in higher kinetically
formed NOx. Van Gerpen et al. have also noticed a NOx
increase with oxygen addition to the intake air.27 It has
been speculated that the higher oxygen concentration
may increase the flame temperature in otherwise fuelrich regions of the combustion chamber, which increases
the rate of formation of NOx.
The incorporation of the fuel-bonded oxygen mitigates
the adverse impact of leaner combustion conditions
compared to free oxygen atom from intake oxygen
enrichment, causing a smaller NOx increase compared
to oxygen-enriched intake air. The glycol ether blend
yields a moderate NOx increase of 10 and 23% with
respect to the baseline for 15 wt % and 30 wt % glycol
ether addition. The dioxolane blends yield an even
smaller NOx increase of 7 and 15% with respect to the
baseline for 10 wt % and 20 wt % dioxolane addition.
This less pronounced effect of dioxolane blends on NOx
emissions may be due to a reduced temperature in the
mixing controlled regime with oxygen addition and less
oxygen being liberated due to differences in molecular
structure compared to the glycol ether blends, as
explained earlier.
Figure 6 shows that trends in CO emissions correlated
well with PM emissions. The effectiveness of CO reduction is similar to that of PM reduction, in the order of
intake air enrichment, glycol ethers, dioxolane, although

Song et al.

Figure 7. Cylinder pressure and heat release rates for


comparison of oxygen addition effects with glycol ether blend
and the equivalent alkane blend (O/C)4.8).

there is an anamolous result with the 10 wt % dioxolane


blend. The reduced CO emissions also support the idea
that there is a higher temperature in the mixingcontrolled regime, which will enhance the oxidation of
any soot, CO, and unburned hydrocarbon.
Isolating the Effect of Oxygen Addition. Tests
with alkane addition at the same levels as the addition
of glycol ethers and dioxolane were conducted to first
isolate the dilution effect of base fuel components, such
as the aromatic and sulfur content. The alkane blends
were selected to secondarily match the boiling points
of the oxygenated blend components, and thereby to
examine the volatility effect on soot formation by
comparison with the baseline results. Whenever matching of the combustion phase between the alkane blend
and the oxygenate blend is necessary, an addition of
EHN was employed. As summarized in Table 4, 15 wt
% glycol ether blend addition and 10 wt % dioxolane
addition were required for O/C ) 4.8 tests. n-heptane
and n-dodecane were selected to match the volatility of
the oxygenated blends. For example, a mixture of 20
wt % n-heptane and 80 wt % n-dodecane was blended
into the base diesel fuel by 15 wt % to match the 15 wt
% glycol ether blends, which is a mixture of 20 wt %
monoglyme and 80 wt % diglyme. Thus, the observed
change between the oxygen addition (glycol ethers or
dioxolane blends) and no oxygen addition (alkane blends)
might be attributed to the combined effects of thermal
and chemical suppression resulting from the chemical
structure of the oxygenated fuels. Furthermore, if there
is no significant shift in premixed and mixing controlled
phases of heat release between the oxygenated fuel
blends and alkane blends, then any soot reduction is
attributable to only the chemical effect from oxygenate
addition.
Figure 7 shows that there is little difference in
combustion pressure and peak burn rate between the
15 wt % glycol ether blend and the equivalent alkane
blend. Unlike the earlier heat release rate comparison
at higher oxygen-to-carbon ratio (O/C ) 5.1), this lower
oxygen-to-carbon ratio (O/C ) 4.8) caused no change in
heat release rate between the oxygenated and nonoxygenated fuel blends, as summarized in Table 6.
Therefore, changes in heat release are not expected to
be major factors in the comparison of PM reduction
between the glycol ether blends and the alkane blends.

Oxygen Enrichment and Diesel Combustion/Emission

Figure 8. Oxygen addition effect of glycol ethers blend


(O/C)4.8) on PM, NOx and CO emissions, g/kWh basis.

Energy & Fuels, Vol. 18, No. 5, 2004 1289

Figure 10. Cylinder pressure and heat release rates for


comparison of oxygen addition effects with 1,3-dioxolane blend
and the equivalent alkane blend (O/C ) 4.8).

Figure 9. Oxygen addition effect of 1,3-dioxolane blend


(O/C ) 4.8) on PM, NOx and CO emissions, g/kWh basis.
Table 6. Combustion Event Analysis for Different
Oxygenate and Alkane Blends at O/C ) 4.8 Condition

fuel blends
base fuel
1 vol % O2 addition
15 wt % glycol ethers
10 wt % dioxolane
15 wt % n-heptane/n-dodecane
10 wt % n-heptane

ignition
dynamic SOI
SOC
delay
(ATDC)
(ATDC)
(CA)
-4.7
-4.7
-5.3
-5.3
-5.3
-5.1

-0.8
-0.9
-1.4
-1.0
-1.4
-1.2

3.9
3.8
3.9
4.3
3.9
3.9

As a result, the reduction of PM emissions by 12%


shown in Figure 8 between the baseline fuel and the
alkane blend can be attributed to overall effects of heat
release and dilution by fuel property changes from the
alkane addition. The reduction of soot by alkane addition compared to the baseline fuel also may be caused
by lowering of the boiling range which alters the fuelair mixture preparation as discussed earlier. Finally,
the additional 20% PM reduction for the glycol ether
blend can be attributed to the chemical effect of oxygen
addition to the fuel, where again it should be noted that
there were no distinct changes in heat release between
the alkane and glycol ether blends. NOx and CO emission do not vary significantly with oxygen addition via
the glycol ethers when compared to the alkane blend.
The above reasoning can be applied to explain the
chemical effect of oxygen addition for the dioxolane
blends as well. As seen in Figure 9, PM and CO
emissions with alkane addition (n-heptane) differ very
little from the baseline fuel, whereas NOx increases by

Figure 11. Cylinder pressure and heat release rates for


different means of oxygen addition at O/C ) 4.8 condition.

8%. Considering the higher volatility of dioxolane, less


PM emissions are expected compared to the base fuel
as in the case of the glycol ether blends, but the result
is the opposite. With no distinct difference in heat
release between the dioxolane blend and the alkane
blend as shown in Figure 10, the PM emission results
indicate that chemical effect of oxygen addition via
dioxolane blends is modest, yielding only 5% PM reduction and only a slight change in NOx and CO emissions.
For a direct comparison, Figure 11 shows that there
is no distinct difference in heat release rate between the
15 wt % glycol ether and 10 wt % dioxolane blends.
Recalling that both blends dilute the base fuel an
equivalent amount in terms of oxygen-to-carbon ratio
allows us to directly compare PM emission effects on
the basis of the chemical structure of oxygenate. Therefore, the effect on the soot reduction by oxygenates
depends on the chemical structure of the oxygenate. In
this study, the linear glycol ethers were more effective
for PM reduction than dioxolane, which contains a ring
structure. This result is consistent with the recent result
of Mueller and co-workers where DBM (ester group) can
directly produce CO2, thereby wasting some of its
oxygen, while TPGME(ether group) does not directly
produce CO2, and its chemistry does not produce as
many unsaturated species that can lead to soot.8,9

1290 Energy & Fuels, Vol. 18, No. 5, 2004

Conclusions
While oxygen enrichment of intake air reduces diesel
PM significantly, fuel oxygenation can provide PM
reduction with only a modest affect on NOxemissions.
With their linear structure, the glycol ether mixture
considered here was shown to be far more effective for
soot reduction than equivalent oxygen addition via
dioxolane with its ring structure. The observations
related to oxygenated fuels and oxygen enrichment
reported here are seen to be consistent with much of

Song et al.

the literature. On one point of continuing debate, the


present work provides further evidence that fuel oxygenation by certain oxygenates with particular structures can provide larger soot suppression.
Acknowledgment. We thank Air Products and
Chemicals, Inc., and especially Tarik Nahieri and P.
John Young for supporting this research project.
EF034103P