JID: PROCI

ARTICLE IN PRESS [m;July 21, 2016;17:11]

Available online at www.sciencedirect.com

Proceedings of the Combustion Institute 000 (2016) 1–9

www.elsevier.com/locate/proci

Ignition delay times of Jet A-1 fuel: Measurements in a

high-pressure shock tube and a rapid compression
machine
A.R. De Toni a,∗, M. Werler b, R.M. Hartmann a,b, L.R. Cancino a,
R. Schießl b, M. Fikri c, C. Schulz c, A.A.M. Oliveira a, E.J. Oliveira d,
M.I. Rocha d
a Combustion and Thermal Systems Engineering Laboratory – LABCET/UFSC, Mechanical Engineering Department,
Federal University of Santa Catarina, Campus Florianópolis – Trindade, Florianópolis, SC 88.040-900, Brazil
b Institute of Technical Thermodynamics – Karlsruhe Institute of Technology, KIT, Karlsruhe, Germany
c IVG, Institute of Combustion and Gas Dynamics – Reactive Fluids, University of Duisburg-Essen, Duisburg, Germany
d Petróleo Brasileiro S.A. – PETROBRAS, Brazil

Received 3 December 2015; accepted 6 July 2016

Available online xxx

Abstract

Ignition delay time (IDT) measurements for Jet A-1 fuel samples have been performed with a rapid com-
pression machine (RCM) and a high-pressure shock tube (ST). The IDT measurements span a pressure range
from 7 to 30 bar, a temperature range from 670 K to 1200 K, and fuel/air equivalence ratios φ from 0.3 to 1.3.
Expressions fitting the experimental data sets were obtained, with fitting parameters being provided. The
combined RCM/ST data aimed at providing information on the two-stage ignition behavior and on the tran-
sition from NTC chemistry to high-temperature radical chain-branching, which are important and hard to
meet targets in the development of chemical surrogates.
© 2016 by The Combustion Institute. Published by Elsevier Inc.

Keywords: Jet fuel; Rapid compression machine; Shock tube; Ignition delay time; Fuel surrogate

1. Introduction of the resistance of gasoline to autoignition, while

Autoignition properties of liquid fuels are
accounted for in terms of octane (ON) and cetane
(CN) ratings, with the former being an assessment
∗
Corresponding author.
E-mail address: detoni@labcet.ufsc.br (A.R. De
Toni).
the latter gauges ignition propensity of Diesel fuel.
However, none of these ratings is included in jet
fuel (kerosene) specifications, even though there is
evidence for a correlation between fuel’s derived
cetane number (DCN) and the occurrence of lean
blow-off [1–3].
Measurements of autoignition in shock tubes
(ST) and rapid compression machines (RCM)

http://dx.doi.org/10.1016/j.proci.2016.07.024
1540-7489 © 2016 by The Combustion Institute. Published by Elsevier Inc.

Please cite this article as: A.R. De Toni et al., Ignition delay times of Jet A-1 fuel: Measurements in a high-
pressure shock tube and a rapid compression machine, Proceedings of the Combustion Institute (2016),
http://dx.doi.org/10.1016/j.proci.2016.07.024
JID: PROCI
ARTICLE IN PRESS
2 A.R. De Toni et al. / Proceedings of the Combustion Institute 000 (2016) 1–9
provide insights into the effects of fuel composi-
tion, equivalence ratio, pressure, and temperature
on autoignition-related phenomena. The assess-
ment of these effects is particularly important
when new jet fuel mixtures are proposed, es-
pecially, those containing biofuel components.
Dooley et al. [4,5] and Hui et al. [6] reported mea-
surements of ignition delay time for alternative
jet fuels made from hydro-processed esters and
fatty acids (HEFA) based on vegetable oils and
animal fats, Fischer–Tropsch fuels made from coal
and natural gas (GtL), and surrogate mixtures
for these alternative fuels. Their results indicated
that all alternative fuels presented higher reactivity
than Jet A, with GtL and camelina HEFA being
consistently the more reactive, reinforcing the need
for the assessment of combustion properties of
new fuels, in relation to established fuels, before
the new mixtures are widely used.
The comparison of the behavior of the igni-
tion delay time of a complex fuel with that of
pure species provides general guidelines for the
development of chemical surrogates amenable to
theoretical studies using detailed chemical kinetics
mechanisms. The comparisons are needed in a large
range of temperatures and pressures [2,6], usually
achieved by a combination of measurements in
shock tubes (ST) and rapid compression machines
(RCM). However, several autoignition studies for
jet fuels [6–13] showed that even when using the
same fuel sample the agreement between different
sets of data is limited. While measurements of
Vasu et al. [7,8] using ST and Dooley et al. [9] using
ST and RCM for Jet A (POSF 4658) at 650–1250 K
and 20 bar presented good agreement, the RCM
results of Kumar and Sung [3,10] at 650–770 K
and 15 bar using the same facility of Dooley et al.
[9], compare well with previous ST measurements
at 15 bar but do not present the behavior expected
from the previous RCM measurements obtained
at 7 bar. The experiments at lower temperature
require care in properly vaporizing the fuel since
kerosene contains species with relatively low vapor
pressure. Also, cracking of the fuel molecules may
occur at higher preheating temperatures, an effect
whose extent is hard to evaluate. While it is possible
to work with the liquid fuel injected into the test
chamber of the ST and the RCM [11,14,15], in
most facilities the fuel and oxidant are premixed in
the gas phase before feeding the reactor.
Here, we present measurements of ignition
delay time of a Jet A-1 fuel produced in Brazil in
the 660–1250 K temperature range at pressures of
7, 15, and 30 bar and at equivalence ratios from 0.3
to 1.3, using the high-pressure shock tube at IVG
and the rapid compression machine at KIT. The
measurements aimed at detecting the extent of the
NTC behavior and the occurrence of two-stage
ignition to aid in the selection of an appropriate
surrogate mixture to emulate ignition delay time.
Please cite this article as: A.R. De Toni et al., Ignition delay times of Jet A-1 fuel: Measurements in a high-
pressure shock tube and a rapid compression machine, Proceedings of the Combustion Institute (2016),
http://dx.doi.org/10.1016/j.proci.2016.07.024
[m;July 21, 2016;17:11]
These measurements will be used as one of the
validation targets for the development of chemical
surrogates for the Jet A-1 fuel.
2. Jet fuel and surrogates
The kerosene obtained from fractional distil-
lation of crude oil has a typical carbon number
distribution between 8 and 16 mostly formed by
paraffinic, naphthenic (cycloalkane), and aromatic
hydrocarbons with boiling point in the 200–300 °C
range. Wood et al. [15] developed one of the earlier
surrogate mixtures, a 14-components blend, in an
attempt to match JP-4 and JP-5 boiling range,
atomization quality and combustion behavior.
Since then, surrogate fuels evolved to include from
simpler mixtures of hydrocarbons, adequate to
computational fluid dynamic (CFD) studies, to
more elaborate blends designed to emulate chem-
ical, physical, or combustion properties in detailed
chemical kinetics and experimental investigations.
A surrogate fuel is usually defined in terms of
the number of species in its composition and its be-
havior relative to ‘targets’. Farrell et al. [16] defines
three main types of targets, in order of increasing
complexity: property targets, development targets,
and application targets. A property target refers
to chemical and physical fuel properties, like the
hydrogen-to-carbon ratio and the distillation curve.
Development targets are the chemical kinetics and
thermophysical behavior typically evaluated in
experiments under well-controlled conditions,
such as ignition delay time (IDT) measurements
in shock tubes and rapid compression machines,
species evolution profiles in flow reactors, or
multicomponent spray vaporization. Application
targets are based on measurements obtained from
engine tests in transient and steady-state operation,
such as combustion efficiency, and gaseous and
particulate emissions.
Regarding the availability of validated oxi-
dation mechanisms for relevant hydrocarbons,
Colket et al. [17], Farrell et al. [16], and Pitz et
al. [18] pointed out that, as late as 2006, only
linear alkanes (n-heptane, n-decane, n-dodecane,
n-hexadecane) and one branched alkane (iso-
octane) received sufficient attention to enable
reliable modeling of their oxidation, while mech-
anisms for naphthenic and aromatic compounds
lacked validation. Further studies [5,19] have
indicated that jet fuel is better approximated
by alkanes with one or two branches, instead
of highly-branched molecules like iso-cetane
(2,2,4,4,6,8,8-heptamethylnonane). However,
highly-branched alkanes and linear ones are much
more affordable to formulate fuel mixtures for
large-scale experiments. Therefore, since the earlier
studies with surrogate fuels, it has become common
practice to select one or more ‘reference fuels’ from
JID: PROCI
ARTICLE IN PRESS [m;July 21, 2016;17:11]

A.R. De Toni et al. / Proceedings of the Combustion Institute 000 (2016) 1–9 3

each hydrocarbon class and adjust its proportions real fuel since both MURI surrogates were unable
to match the desired real-fuel properties. For to reproduce the real fuel’s smooth transition from
instance, the six-component Utah/Yale surrogate NTC chemistry to high-temperature radical chain-
[20] (30% n-dodecane, 20% n-tetradecane, 10% branching mechanism. The authors argue that
iso-cetane, 20% methylcyclohexane, 15% o-xylene, none of the a priori evaluated properties (DCN,
5% tetralin by vol.) has two aromatic species to H/C ratio, molecular weight, and TSI) seem to
match jet fuel’s volatility and smoke point; the correlate with the occurrence of this transition.
Drexel/S5 surrogate [20] (26% n-dodecane, 36% In order to select a surrogate mixture for the
iso-cetane, 14% methylcyclohexane, 6% decalin, Jet A-1 fuel of interest here, this work focus on the
18% 1-methylnaphthalene by vol.) was formu- NTC behavior and the occurrence of two-stage ig-
lated to mimic the low-temperature reactivity of nition as detected by ignition delay measurements.
JP-8; and the binary Aachen surrogate [21] (80% Therefore, shock tube and RCM measurements are
n-decane, 20% 1,2,4-trimethylbenzene by weight) combined to span the low to high temperature igni-
focuses on the overall H/C ratio and autoignition tion delay along a few pressures and stoichiometry.
behavior, but has also similar heat-of-combustion
and molecular weight.
The Multi University Research Initiative
3. Materials and methods
(MURI) [4,5,9,19,22–24] extensively tested real
fuels in shock tubes, rapid compression machines,
3.1. High pressure shock tube
flow reactors, counterflow and diffusion flames,
outlined new concepts on the nature of real fuel
The measurements were carried out in a heated
oxidation pathways and proposed approaches to
(125 °C) high-pressure shock tube in the Institute
the formulation of surrogates. One of the impor-
for Combustion and Gas Dynamics (IVG) at the
tant takeaways from these works is the ‘distinct
University of Duisburg-Essen that can be oper-
chemical functionalities’ concept that defines that
ated up to post-combustion pressures of 500 bar.
a surrogate fuel must be able to produce a radical
This facility and the experimental procedures are
pool similar to that of the real fuel. Therefore,
described elsewhere [27–30]. Gas mixtures were
the chemical class composition of surrogate mix-
prepared by injecting liquid Jet A-1 into an evacu-
tures can significantly differ from that of the real
ated, heated (125 °C) stainless-steel mixing vessel.
fuel while the combustion properties of the fully
The total amount of fuel and air were controlled
prevaporized fuel remain remarkably similar. This
manometrically to ensure the desired equivalence
concept was associated with an a priori property
ratio. The mixture was stirred for 120 min to ensure
matching method involving four property tar-
homogeneity before it was filled in the test section
gets: derived cetane number (DCN), H/C ratio,
of the shock tube. The shock speed was mea-
average molecular weight, and threshold sooting
sured over two intervals using three piezo-electric
index (TSI). Based on these principles and exper-
pressure gauges to determine the post-reflected
imental data of the average Jet A sample (POSF
shock temperatures. The estimated uncertainty in
4658), two surrogates mixtures were proposed:
reflected shock temperature is less than ± 25 K.
three-component MURI 1 (42.67% n-decane,
Pressure data were recorded with a time resolution
33.02% iso-octane, 24.31% toluene by mol) and
of 0.1 μs. The ignition delay times were determined
four-component MURI 2 (40.4% n-dodecane,
by extrapolating the steepest increase of the CH∗
29.5% iso-octane, 22.8% n-propylbenzene, 7.3%
chemiluminescence emission (at 431.5 nm) signal
1,3,5-trimethylbenzene by mol) with an associ-
to its zero level on the time axis as shown in Fig. 1.
ated comprehensive reaction mechanism for the
To also determine the delay times of the
former. The overall agreement between the real
first-stage ignition (cool flame), CH2 O∗ chemi-
fuel and both surrogates was significant but some
luminescence was simultaneously detected in the
issues remained. The three-component MURI 1
400–450 nm wavelength range. The experiments
presented a mismatch in average molecular weight
were carried out with synthetic air containing
thus affecting the evaluation of targets that depend
79.5% N2 and 20.5% O2 . As driver gas, tailored
on transport properties, e.g., extinction strain
mixtures of 5–20% Ar in He were used to adjust its
rates. This issue was amended in the formulation
acoustic properties according to [31,32] to ensure
of the MURI 2, which comprises heavier linear
long test times.
alkane and aromatics. As of 2015 the only reaction
mechanisms that include the whole palette of rele-
vant hydrocarbons, i.e., the four species considered 3.2. Rapid compression machine
in MURI 2 surrogate and also naphthenics are
the computationally-generated schemes from the Details of the Rapid Compression Machine fa-
Milano group [25,26]. One of these schemes was cility at the Institute of Technical Thermodynam-
employed by Dooley et al. [24] to study the role ics (ITT) at Karlsruhe Institute of Technology have
of cycloalkanes in the combustion behavior of the been described elsewhere [33]. The fuel-mixture is

Please cite this article as: A.R. De Toni et al., Ignition delay times of Jet A-1 fuel: Measurements in a high-
pressure shock tube and a rapid compression machine, Proceedings of the Combustion Institute (2016),
http://dx.doi.org/10.1016/j.proci.2016.07.024
JID: PROCI
ARTICLE IN PRESS [m;July 21, 2016;17:11]

4 A.R. De Toni et al. / Proceedings of the Combustion Institute 000 (2016) 1–9

Fig. 1. Determination of the ignition delay time from the CH∗ chemiluminescence signal trace in the ST experiments. Jet
A-1 in air, p5 = 16.3 bar, T5 = 704.3 K, φ = 1.0, τ 1 = 6829 μs, τ = 8176 μs.

created in a mixing vessel heated at 140 °C. Homo-
geneity of the fuel mixture is ensured by constantly
mixing with a magnetically coupled stirrer for at
least 120 minutes. The mixture is then filled into the
combustion chamber up to some pre-defined pres-
sure and temperature. An oil bath surrounding the
machine ensures well-defined thermal conditions at
the combustion chamber before compression.
The compression is achieved by a pneumatic
driver that pushes the piston into the combustion
chamber. A knee-lever mechanism is used to pre-
vent further piston motion at top dead center po-
sition. Thus, after compression, constant-volume
conditions prevail. The RCM allows an adjustment
of the geometric compression ratio to provide a
wide temperature range. Each condition was mea-
sured three times and the repeatability was found to
be very good throughout the whole range of mea-
surement conditions, as detailed below.
4. Results and discussion
Following Vasu et al. [7], the chemical composi-
tion of Jet A-1 fuel was assumed as C11 H21 with a
density of 0.81 g/cm3 , resulting in a stoichiometric
air fuel ratio of 16.25. Table 1 presents the condi-
tions of the measurements in the ST and the RCM.
4.1. ST ignition delay time measurements and
simulations
The ignition delay times evaluated from the
CH∗ emission are listed in the Supplementary
material, along with the respective pressures p
and temperatures T, for stoichiometric and lean
Jet A-1/air mixtures. During the ignition delay, a
slight gradual increase in pressure (dp/dt = 3%/ms)
is recorded. Using the maximum pressure increase,
we calculate the resulting variation in temperature
relative to the post-shock temperature T5 based on
an adiabatic compression model. The maximum
temperature increase was found to be 30 K for our
measurement conditions (see the column ±T5
in Table S1). The determined T5 uncertainties
[34] imply an error between ± 11 and ± 16% in the
calculated ignition delay times, to be taken into
account in chemical kinetics modeling.
The measurements are curve-fitted to an equa-
tion of the form τ = A exp(B/T) p–x φ –y , where
x and y are the pressure and stoichiometry ex-
ponents, respectively. Multiple linear regression
analysis using ln(τ ) as the dependent variable and
(1/T), ln(p) and ln(φ) as independent variables
identified the values of x = 0.89 and y = 0.79 using
the data for φ =1.0 and 0.3 listed in Table S1.
For the temperature-and pressure dependence, the
expression
 
14510 ± 490
τ /μs = 102.06±0.22 exp
T/K
× ( p/bar )−0.89±0.12 (φ)−0.79±0.06 (1)
was determined from a fit for the high-temperature
range (925–1195 K).
Numerical simulations of ignition delay times in
CHEMKIN-Pro [35], using the three-component
mechanism of Dooley et al. [9] which comprises
6633 reactions among 1599 species were per-
formed. Dooley et al. acknowledge that a detailed
assessment and validation of the mechanism was
not carried out and that the model should closely
reproduce their flow reactor data while predicting
ignition delays up to a-factor-of-two slower.

Please cite this article as: A.R. De Toni et al., Ignition delay times of Jet A-1 fuel: Measurements in a high-
pressure shock tube and a rapid compression machine, Proceedings of the Combustion Institute (2016),
http://dx.doi.org/10.1016/j.proci.2016.07.024
JID: PROCI
ARTICLE IN PRESS [m;July 21, 2016;17:11]

A.R. De Toni et al. / Proceedings of the Combustion Institute 000 (2016) 1–9 5

Table 1
Investigated conditions (fuel/air equivalence ratio, effective temperature, pressure) for the
ST and RCM measurements.
Equivalence ratio, φ ST RCM
0.3 15 bar, 1000 K/T = 0.83–1.2 15 bar, 1000 K/T = 1.1–1.5
30 bar, 1000 K/T = 0.83–1.25 –
0.7 – 7 bar, 1000 K/T = 1.1–1.5
– 15 bar, 1000 K/T = 1.1–1.5
1.0 –
15 bar, 1000 K/T = 0.8-1.4 7 bar, 1000 K/T = 1.1–1.5
30 bar, 1000 K/T = 0.83-1.11 15 bar, 1000 K/T= 1.45–1.55-
–
1.3 – 7 bar, 1000 K/T = 1.1–1.5

Table 2
Composition of proposed surrogates for Jet A-1 ignition delay simulations and related a
priori targets.
Surrogate A Surrogate B
n-decane (% mol) 33.5 65.4
iso-octane (% mol) 36.5 14.0
toluene (% mol) 30.0 20.6
Derived cetane number (by regression) 42.0 56.5
CH2 /CH3 x [CH2 + CH3 ] mass fraction 1.304 2.462

Two surrogate mixtures were proposed, based
on composition data of the Jet A-1 sample (ASTM
D1319: 79.4% paraffins, 0.8% olefins, 19.8% aro-
matics by vol.) and regarding one of the MURI
surrogate targets (ASTM D6890 – DCN 41.9). The
surrogate ‘A’ (33.5% n-decane, 36.5% iso-octane,
30.0% toluene by mol) was based on regression
analysis of derived cetane number, as employed
by Dooley et al. [9]. Since several mixtures of
these three hydrocarbons can produce the same
DCN, the mixture chosen as surrogate ‘A’ was
the one with similar volume fraction of aromatics
(30.0% toluene by mol i.e. 20.5% by vol.). The
surrogate mixture ‘B’ was formulated consider-
ing the experimental evidence [36–38] that linear
and weakly-branched alkanes are the statistical
‘mode’ of hydrocarbon isomers present in liquid
fuels and this preponderance can be related to a
methylene-to-methyl (CH2 /CH3 ) ratio that glob-
ally correlates with the low-temperature reactivity.
Therefore, surrogate ‘B’ (65.4% n-decane, 14.0%
iso-octane, 20.6% toluene by mol) was formu-
lated regarding the ASTM D1319 composition
and adding this molecular group composition
CH2 /CH3 x [CH2 +CH3 ] metric [19] while relaxing
the DCN matching criteria. Table 2 presents the
composition and target values for the proposed
surrogates while Fig. 2 shows the comparison
between experimental and numerical results of
surrogate ‘B’.
The numerical IDT prediction based on the
composition of surrogate ‘A’ is about 60% higher
than the IDT values predicted by surrogate ‘B’ lead-
ing to a larger deviation compared to the experi-
ments therefore only modeling with surrogate B is
presented. As mentioned before, an overestimation
of IDT for both surrogates is expected since this
chemical kinetics model consistently overestimated
IDT of the original experiments of Dooley et al.
[9], in which the a priori target-matching criteria
was carried out for all four targets. The computed
IDTs of surrogate ‘A’ presented a significant over-
estimation, indicating that a DCN matching us-
ing the strategy of adjusted volume of aromatics
is not enough to characterize the ignition behavior.
Among the likely causes for this lack of agreement
is the limited information provided by the ASTM
D1319 composition evaluation which does not
distinguish aliphatics (linear and branched) from
naphthenics nor does it provide data regarding the
substituted or unsubstituted nature of the aromat-
ics. Such information would be helpful to use an-
other MURI target, i.e., the overall H/C ratio.
On the other hand, surrogate ‘B’ computations
predicted IDTs very close to the experimental
values for low-to-intermediate temperatures (700–
900 K) at 15 bar and stoichiometric mixtures. The
agreement between experiment and computations
for other conditions was qualitative, with lean mix-
tures having longer IDTs than stoichiometric ones
and an increase in pressure resulting in shorter
ignition delays. The faster ignition of surrogate ‘B’
compared to ‘A’ is due to the higher fraction of
n-decane that offers sixteen secondary hydrogen
sites (methylene) where molecular oxygen can
promote an H-abstraction to form radicals. The
other alkane in the formulation, iso-octane, has
only two such sites and they are shielded against O2
attack by the methyl substituents in both adjacent
carbons.

Please cite this article as: A.R. De Toni et al., Ignition delay times of Jet A-1 fuel: Measurements in a high-
pressure shock tube and a rapid compression machine, Proceedings of the Combustion Institute (2016),
http://dx.doi.org/10.1016/j.proci.2016.07.024
JID: PROCI
ARTICLE IN PRESS [m;July 21, 2016;17:11]

6 A.R. De Toni et al. / Proceedings of the Combustion Institute 000 (2016) 1–9

Fig. 2. Comparison of measured (shock tube) and simulated ignition delay time for Jet A-1/air mixtures (detailed kinetics
model for Jet A from Dooley et al. [9]); surrogate ‘B’ was used for the numerical calculations.

In the low-to-intermediate temperatures, sur- time below a few milliseconds) to reasonably

rogate ‘B’ has enough secondary hydrogen sites
to present reactivity similar to the real fuel.
However, in the transition from the NTC region
to high temperature, n-decane is mostly gone
while toluene, after an initial H-abstraction at the
methyl substituent, forms a resonance-stabilized
benzylic radical. Therefore, the surrogate ability
to reproduce the real fuel’s intermediate-to-high
temperature reactivity is limited by the absence of
sufficient methylene units in the compounds other
than n-decane. Dooley et al. also argue that the
model shows an elongated NTC behavior, which
is likely due to an under-emphasis of beta-scission
reactions of the alkyl radicals thus leading to
increased low temperature chemistry processes
instead of high temperature chain branching. This
behavior is also noticeable in our simulations.
4.2. RCM ignition delay time measurements
Figure 3 presents the ignition delay times mea-
sured in the RCM for the Jet A-1 sample at 7 bar
for equivalence ratios ranging from slightly lean
(φ = 0.7) to rich (φ = 1.3). The temperatures shown
in the diagram are effective temperatures represent-
ing an average of the pre-ignition temperature in
the RCM, as derived from the measured pressure
curve employing an adiabatic core hypothesis.
The experimental data and non-reactive pres-
sure histories are available in the Supplementary
material.
A clear two-stage ignition behavior was ob-
served in the range below 760 K. Above 790 K,
the first stage ignition became too fast (delay
assign a pressure and temperature to the ignition
event. The data points for first-stage ignition
above 790 K should therefore be considered as
semi-quantitative. This somewhat fuzzy nature of
the points with ignition delay times below ∼5 ms
becomes evident also by the larger scatter in first-
stage IDT at the nominally-identical measurement
points.
First-stage and main-ignition delay times be-
come increasingly close as temperature decreases.
Below about 660 K at 7 bar, they practically co-
incide. The temperature where the two begin to
“overlap” is lower for the lean mixture (φ = 0.7)
and higher for the rich mixture (φ = 1.3). At 7 bar,
φ = 0.3, no auto-ignition was detected in the RCM
within the studied temperature range. At a pres-
sure of 15 bar, however, auto-ignition could be de-
tected also for φ = 0.3, and again, two-stage igni-
tion was observed, as shown in Fig. 4. At 15 bar,
ignition was one order of magnitude faster than at
7 bar, for both φ = 0.7 and φ = 1.0. Comparison of
Figs. 2 and 4 reveals that RCM and shock tube IDT
match well in the temperature and pressure region
where they overlap i.e. 15 bar, equivalence ratio
1.0, 1000 K/T=1.4; thus the often described phe-
nomenon of non-matching ST and RCM ignition
delay times does not appear in our case, see Fig. S1.
Within the NTC region, a pronounced pres-
sure dependence of the IDT was found, which is
demonstrated in Fig. 5 by IDTs for φ = 0.7 as a
function of temperature. These measurements near
7.2 and 8.1 bar highlight this strong dependence.
In the RCM the post-compression pressure at top
dead center displays slight shot-to-shot variations

Please cite this article as: A.R. De Toni et al., Ignition delay times of Jet A-1 fuel: Measurements in a high-
pressure shock tube and a rapid compression machine, Proceedings of the Combustion Institute (2016),
http://dx.doi.org/10.1016/j.proci.2016.07.024
JID: PROCI
ARTICLE IN PRESS [m;July 21, 2016;17:11]

A.R. De Toni et al. / Proceedings of the Combustion Institute 000 (2016) 1–9 7

Fig. 3. RCM ignition delay times of Jet A-1/air mixtures at 7 bar for various equivalence ratios. Open symbols denote the
first stage ignition, solid symbols denote the main ignition.

Fig. 4. RCM ignition delay times of Jet A-1/air mixtures at 15 bar for various equivalence ratios. Open symbols denote
the first stage ignition (φ = 0.3 only), solid symbols denote the main ignition.

with an amplitude of around 0.1 bar. Because data
points with post-compression pressures from a
certain interval are assigned to the same nominal
pressure in the IDT diagrams, the strong pressure
dependence of IDT creates an apparent scatter in
the diagrams.
The regions of the Arrhenius diagram where
the IDT display a linear behavior are usually
curve fitted using expressions in the form
log(τ ig ) = A + B/T, such as Eq. (1). These expres-
sions, however, cannot describe the NTC region
with a single set of parameters A and B. In this
context Zhou et al. [39] recently proposed fitting
expressions using combinations of different mathe-
matical functions that can reproduce complex igni-
tion delay curve shapes. Here, we present a compact

Please cite this article as: A.R. De Toni et al., Ignition delay times of Jet A-1 fuel: Measurements in a high-
pressure shock tube and a rapid compression machine, Proceedings of the Combustion Institute (2016),
http://dx.doi.org/10.1016/j.proci.2016.07.024
JID: PROCI
ARTICLE IN PRESS [m;July 21, 2016;17:11]

8 A.R. De Toni et al. / Proceedings of the Combustion Institute 000 (2016) 1–9

Fig. 5. Strong pressure dependence of IDT (main ignition stage) for φ = 0.7, pressure range between 7 and 8.3 bar. The
color of the scatter symbols represents the pressure, as indicated by the color scale. The dashed line is a nonlinear fit
through data points near 7 bar. (For interpretation of the references to color in this figure legend, the reader is referred to
the web version of this article)

formula that can also fit the NTC region, given by
 
  θ − a5
log τig = a1 + a2 θ + (a3 + a4 θ ) erf (2)
a6
where erf is the error function, θ is the inverse
temperature and ai are curve-fitted parameters.
This function has some properties that make it
attractive for fitting ignition delay time curves.
For instance, in the limit of small (large) values
of θ, erf attains the value –1 (+1), and the function
approaches the forms:
 
log τig ≈ a1 − a3 + (a2 − a4 ) θ , when θ is small,
 
log τig ≈ a1 + a3 + (a2 + a4 ) θ , when θ is large.
Both are of the “standard” Arrhenius form
A + Bθ with high- and low-temperature values for
A and B. Therefore, the function is “asymptotically
Arrhenius-like” for small and large θ and can
represent the normal branches of ignition delay to
the left and right sides of the NTC region. Between
these regions, it can fit both negative and positive
slopes. A sample fit with this function is included
in Fig. 5, for a pressure of 7 bar. The parameters
for this fit are
a1 = -17.484, a2 = 16.169, a3 = 6.5411,
a4 = -7.8843, a5 = 1.1706, a6 = 0.22398,
using θ = 1000 K/T, and expressing τ ig in millisec-
onds. Fitting parameters for the remaining con-
ditions, including the matching shock tube data
(φ=1, 15 bar), are included in the supplementary
material.
5. Summary and conclusions
Ignition delay times of Jet A-1/air mixtures were
measured in a high-pressure shock tube and a rapid
compression machine spanning a pressure range
from 7 to 30 bar, temperatures of 670–1200 K and
fuel/air equivalence ratios from 0.3 to 1.3. In the
700–870 K range, IDTs show distinct NTC be-
havior at both facilities (7 bar at RCM and 15 bar
for ST) while at temperatures outside this range
the IDTs showed an Arrhenius-like behavior. The
experiments in rapid compression machine show
a two-stage ignition behavior for compressed tem-
peratures below 760 K. Expressions fitting the ex-
perimental data sets were obtained, with fitting pa-
rameters provided in the supplementary material.
Based on limited information about the Jet
A-1 sample hydrocarbon makeup (ASTM D1319:
79.4% paraffins, 0.8% olefins, 19.8% aromatics by
vol.), computations of IDT, using the mechanism
of Dooley et al. [9] were conducted for two different
surrogate compositions. The surrogate ‘B’ (65.4%
n-decane, 14.0% iso-octane, 20.6% toluene by mol),
which was proposed regarding ASTM D1319 com-
position and a low temperature molecular group
metric [19] presented good agreement for stoichio-
metric data at 15 bar, i.e., reproducing the reactivity
plateau in the low to intermediate temperatures.

Please cite this article as: A.R. De Toni et al., Ignition delay times of Jet A-1 fuel: Measurements in a high-
pressure shock tube and a rapid compression machine, Proceedings of the Combustion Institute (2016),
http://dx.doi.org/10.1016/j.proci.2016.07.024
JID: PROCI
ARTICLE IN PRESS
A.R. De Toni et al. / Proceedings of the Combustion Institute 000 (2016) 1–9
Acknowledgments
The authors gratefully acknowledge the sup-
port for this work from Petróleo Brasileiro – S.A.
PETROBRAS (Agreement # 0050.0070218.11.9).
The help of Natascha Schlösser in supporting the
experiments at IVG is greatly appreciated.
Supplementary materials
Supplementary material associated with this ar-
ticle can be found, in the online version, at doi:
10.1016/j.proci.2016.07.024.
References
[1] L.Q. Maurice, H. Lander, T. Edwards, W.E. Harri-
son, Fuel 80 (2001) 747–756.
[2] P. Dagaut, M. Cathonnet, Prog. Energy Combust.
Sci. 32 (2006) 48–92.
[3] K. Kumar, C.J. Sung, Combust. Flame 157 (2010)
676–685.
[4] S. Dooley, S.H. Won, J. Heyne, et al., Combust.
Flame 159 (2012) 1444–1466.
[5] S. Dooley, S.H. Won, S. Jahangirian, et al., Combust.
Flame 159 (2012) 3014–3020.
[6] X. Hui, K. Kumar, C.J. Sung, T. Edwards, D. Gard-
ner, Fuel 98 (2012) 176–182.
[7] S.S. Vasu, D.F. Davidson, R.K. Hanson, Combust.
Flame 152 (2008) 125–143.
[8] S.S. Vasu, D.F. Davidson, R.K. Hanson, Shock
Waves, Springer, Berlin Heidelberg, Göttingen, 2009,
pp. 293–298.
[9] S. Dooley, S.H. Won, M. Chaos, et al., Combust.
Flame 157 (2010) 2333–2339.
[10] K. Kumar, C.J. Sung, Fuel 89 (2010) 2853–2863.
[11] C. Allen, E. Toulson, T. Edwards, T. Lee, Combust.
Flame 159 (2012) 2780–2788.
[12] T. Malewicki, S. Gudiyella, K. Brezinsky, Combust.
Flame 160 (2013) 17–30.
[13] P. Dagaut, F. Karsenty, G. Dayma, et al., Combust.
Flame 161 (2014) 835–847.
[14] D.F. Davidson, D.R. Haylett, R.K. Hanson, Com-
bust. Flame 155 (2008) 108–117.
[15] C. Wood, V. McDonnell, R. Smith, G. Samuelson, J.
Propuls. Power 5 (1989) 399–405.
[16] J.T. Farrell, N.P. Cernansky, F.L. Dryer, et al., in:
Proceedings of SAE World Congress, SAE Interna-
tional, Detroit, 2007, 2007.
[17] M. Colket, T. Edwards, S. Williams, et al., in: Pro-
ceedings of the 45th AIAA Aerospace Sciences
Meeting and Exhibit, AIAA, Reno, 2007, pp. 1–21.
[18] W.J. Pitz, N.P. Cernansky, F.L. Dryer, et al., in:
Proceedings of SAE World Congress, SAE Interna-
tional, Detroit, 2007, 2007.
Please cite this article as: A.R. De Toni et al., Ignition delay times of Jet A-1 fuel: Measurements in a high-
pressure shock tube and a rapid compression machine, Proceedings of the Combustion Institute (2016),
http://dx.doi.org/10.1016/j.proci.2016.07.024
[m;July 21, 2016;17:11]
9
[19] S.H. Won, S. Dooley, P.S. Veloo, et al., Combust.
Flame 161 (2014) 826–834.
[20] S. Humer, a. Frassoldati, S. Granata, et al., Proc.
Combust. Inst. 31 (I) (2007) 393–400.
[21] S. Honnet, K. Seshadri, U. Niemann, N. Peters, Proc.
Combust. Inst. 32 (I) (2009) 485–492.
[22] S.H. Won, S. Dooley, F.L. Dryer, Y. Ju, Combust.
Flame 159 (2012) 541–551.
[23] F.L. Dryer, S. Jahangirian, S. Dooley, et al., Energy
Fuels 28 (2014) 3474–3485.
[24] S. Dooley, J. Heyne, S.H. Won, P. Dievart, Y. Ju,
F.L. Dryer, Energy Fuels 28 (2014) 7649–7661.
[25] E. Ranzi, A. Frassoldati, A. Stagni, M. Pelucchi,
A. Cuoci, T. Faravelli, Int. J. Chem. Kinet. 46 (2014)
512–542.
[26] A. Cuoci, A. Frassoldati, T. Faravelli, E. Ranzi, Proc.
Combust. Inst. 35 (2014) 1621–1627.
[27] M. Fikri, J. Herzler, R. Starke, C. Schulz, P. Roth,
G.T. Kalghatgi, Combust. Flame 152 (2008) 276–281.
[28] L.R. Cancino, M. Fikri, A.A.M. Oliveira, C. Schulz,
Proc. Combust. Inst. 32 (2009) 501–508.
[29] L.R. Cancino, Development and Application of De-
tailed Chemical Kinetics Mechanisms for Ethanol and
Ethanol-Containing Hydrocarbon Fuels, Federal Uni-
versity of Santa, Catarina, 2009.
[30] L.R. Cancino, M. Fikri, A.A.M. Oliveira, C. Schulz,
in: Proceedings of the 20th International Congress
on Mechanical Engineering – COBEM 2009,
ABCM, Gramado, 2009, pp. 15–20.
[31] H. Oertel, Stossrohre, Springer Verlag,
Wien-NewYork, 1966.
[32] H.B. Palmer, B.E. Knox, ARS J. 31 (1961) 826–828.
[33] M. Werler, L.R. Cancino, R. Schiessl, U. Maas,
C. Schulz, M. Fikri, Proc. Combust. Inst. 35 (2015)
259–266.
[34] L.R. Cancino, M. Fikri, A.A.M. Oliveira, C. Schulz,
Energy Fuels 24 (2010) 2830–2840.
[35] CHEMKIN-Pro 15083, Reaction Design, (2009) Re-
action Design, San Diego.
[36] L.M. Shafer, R.C. Striebich, J. Gomach, T. Edwards,
in: Proceedings of the 14th AIAA/AHI Space Planes
Hypersonic Systems and Technologies Conference,
AIAA, Canberra, 2006.
[37] C.A. Moses, Comparative Evaluation of Semi-
Synthetic Jet Fuels. Contract 33415.02-D, Dayton,
2008.
[38] F.L. Dryer, Y. Ju, K. Brezinsky, R.J. Santoro, T.A.
Litzinger, C.J. Sung, Generation of Comprehensive
Surrogate Kinetic Models and Validation Databases
for Simulating Large Molecular Weight Hydrocar-
bon Fuels. Contract FA9550-07-1-0515, Arlington,
2012.
[39] A. Zhou, T. Dong, B. Akih-Kumgeh, in: Proceedings
of the 25th International Colloquim on the Dynam-
ics of Explosions and Reactive Systems, Leeds, 2015.