Download as pdf or txt
Download as pdf or txt
You are on page 1of 12

Fuel 343 (2023) 127885

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Full Length Article

Effect of dimethyl ether on ignition characteristics of ammonia and


chemical kinetics
Yifan Jin , Xin Li , Xin Wang , Zhihao Ma *, Xianglin Chu
School of Vehicle and Traffic Engineering, Henan University of Science and Technology, Luoyang, Henan 471003, People’s Republic of China

A R T I C L E I N F O A B S T R A C T

Keywords: The contribution of dimethyl ether(DME) to the ignition delay times(IDTs) of ammonia(NH3) was investigated
Ammonia behind reflected shock waves. The experiments were performed at a pressure of 0.14/1.0 MPa, temperature
Dimethyl ether range of 1150–1950 K, equivalence ratio of 0.5/1.0/2.0, and NH3/DME mixing ratios of 100/0, 95/5, 90/10, and
Shock tube
70/30. It was observed that the addition of DME decreased the IDTs and promoted the reactivity of NH3. With the
Ignition delay time
increase of DME, the effect of the equivalence ratio on the IDTs of NH3 decreased. Under higher temperature and
Chemical kinetics
pressure conditions, the promoting effect of DME on the ignition of NH3 was weakened. An updated mechanism
is proposed to reveal the promoting effect of DME on the ignition of NH3. Mechanisms from the literature were
compared against the measurements, and the updated kinetic mechanism was validated with experimental data.
Good agreement between measurements and simulations were shown. Chemical kinetic analyses were performed
to interpret the interactions between DME and NH3 during fuel ignition. The numerical analysis indicated that
the promotion effect of DME is primarily due to an increase of the rate of production and concentration of the
radical pool, especially the OH radical pool. The large number of OH radicals generated by the reaction HO2 +
CH3 = OH + CH3O during the early oxidation of DME is key to the NH3 consumption and early initiation of the
chain reaction.

1. Introduction DME binary fuel. The results show that NOX emissions increase with
increasing ammonia molar fraction, but soot emissions are extremely
In recent years, global warming has become a major global chal­ low. Reiter et al. [7] found that replacing a portion of diesel with
lenge. There is a tendency to search for fuels that can reduce greenhouse ammonia can significantly reduce CO2 emissions. Recent studies have
gas emissions. Ammonia (NH3) can be produced using renewable energy reported the blending of hydrocarbon fuels with pure ammonia [8–10],
sources. Ammonia is considered as an effective carrier of H2 owing to its NH3/H2 [11–13], NH3/CH4 [14–15], and NH3/n-heptane [16] to
higher energy density, convenient transportation, and storage [1–3]. improve ammonia combustion. Among them, Mathieu et al. [8]
Ammonia has received increasing attention as a carbon-free alternative measured ignition delay time(IDTs) behind reflected shock waves at
fuel for engines that can effectively reduce CO2 emissions. However, temperatures of 1560–2455 K, pressures of 1.4, 11, and 30 atm, and
ammonia has certain demerits as a fuel, such as a slow combustion rate, equivalence ratios of 0.5, 1.0, and 2.0 for mixtures of ammonia highly
high ignition energy, and poor combustion stability [4]. Numerous diluted in Ar (98–99 %). Otomo et al. [13] improved the reaction
studies have demonstrated that blending some reactive fuels with mechanism and predicted the IDTs and laminar flame speed for NH3/H2.
ammonia can significantly improve their combustion characteristics. To The ammonia oxidation reaction mechanism was developed based on
avoid the demerits of ammonia as a fuel and to expand opportunities for Song et al. [17], and the performance of the improved mechanism was in
the practical application of ammonia in available combustion systems, good agreement with that proposed by Mathieu et al. [8]. Yu et al. [16]
binary combustion using hydrocarbon fuels blended with ammonia has combined the n-heptane mechanism proposed by Zhang et al. [18] and
been investigated⋅NH3 and H2 mixtures as a fuel can eliminate carbon the NH3 mechanism proposed by Glarborg et al. [19] to obtain an n-
oxide emissions from engines [5]. Gross et al. [6] investigated the per­ heptane/NH3 co-blending mechanism. This mechanism can qualita­
formance characteristics of a compression-ignition engine using NH3/ tively analyze the variation of the IDTs with the NH3 blending ratio,

* Corresponding author.
E-mail address: mazhihao@haust.edu.cn (Z. Ma).

https://doi.org/10.1016/j.fuel.2023.127885
Received 7 December 2022; Received in revised form 12 February 2023; Accepted 16 February 2023
Available online 25 February 2023
0016-2361/© 2023 Elsevier Ltd. All rights reserved.
Y. Jin et al. Fuel 343 (2023) 127885

Fig. 1. Schematic diagram of shock tube apparatus.

Table 1
Composition of the experimental gas mixture.
Mixture ϕ NH3/% DME/% O2/% Ar/% P (MPa)

1 0.5 2.457 0 3.686 93.857 0.14


2 0.5 2.064 0.108 3.745 94.083 0.14
3 0.5 1.751 0.194 3.793 94.262 0.14
4 0.5 0.961 0.412 3.912 94.715 0.14
5/13 1.0 4.376 0 3.282 92.342 0.14/1.0
0
6/14 1.0 3.721 0.195 3.377 92.707 0.14/1.0
7/15 1.0 3.193 0.354 3.458 92.995 0.14/1.0
8/16 1.0 1.799 0.771 3.662 93.768 0.14/1.0
9 2.0 7.181 0 2.693 90.126 0.14
10 2.0 6.226 0.327 2.825 90.622 0.14
11 2.0 5.422 0.602 2.936 91.040 0.14
12 2.0 3.190 1.367 3.247 92.196 0.14

P
Fig. 3. Comparison of the IDTs of NH3/O2/Ar mixtures between the measured
data and the predictions with different models.

DME has a high cetane number, low ignition temperature, short IDTs,
and minimal soot formation during combustion, making it an attractive
alternative to diesel [21]. DME has a promoting effect on low-activity
fuels under various conditions. The addition of DME to methane accel­
erates ignition and increases the burning velocity over a wide temper­
ature range [22]. Propane combustion can be promoted by DME through
the increased concentration of free radicals [23–24].
DME has performed well as an additive to promote ammonia com­
bustion. Due to its polarity, DME can be mixed with NH3 and the mixture
remains stable. Therefore, NH3/DME binary mixtures are promising
t alternative fuels for the internal combustion engine. Issayev et al. [25]
determined the IDTs of NH3/DME mixtures at pressures of 2.0–4.0 MPa,
Fig. 2. Schematic diagram of the definition of the ignition delay time(IDT). temperatures of 649–900 K, and the DME blending ratios of 5 to 50 %.
The laminar flame speeds of NH3/DME mixtures were measured at DME
temperature, and equivalence ratio, but overpredicts the IDTs at tem­ blending ratios of 18 to 47 %, pressures of 0.1, 0.3, and 0.5 MPa,
peratures below 720 K. equivalence ratios of 0.8–1.3, and temperature of 300 K. The Issayev
Dimethyl ether (DME) has considered as a promising fuel. As a mechanism [25] is a modified version of the Shrestha mechanism [26]
structurally simple ether that remains in the gaseous state at atmo­ with the addition of the O2CH2OCH2O2H submechanism. The results
spheric temperature and pressure, DME has physical properties similar show that the mechanism can efficiently predict the IDTs of the NH3/
to those of LPG and can be produced from renewable energy sources, DME mixture, but the predictions are biased under low-pressure con­
making it one of the most promising renewable energy sources [20]. ditions. Murakami et al. [27] investigated the effect of DME

2
Y. Jin et al. Fuel 343 (2023) 127885

Fig. 4. Comparison of the IDTs of NH3/DME/O2/Ar mixture between the Fig. 5. Comparison of the calculated results of IDTs using the Shrestha-Burke
measured data and the predictions with different assembled models. mechanism and the final mechanism in this study.

Based on experimental data, the available mechanism was validated and


Table 2
modified to develop the NH3/DME mechanism to better understand the
Rate coefficients for selected reactions in the updated model.
effect of oxygenated fuels on ammonia ignition characteristics. The
Reaction A/ n Ea/ Source
developed mechanism can efficiently predict the IDTs of NH3/DME
Mechanism (cm3⋅(mol⋅s)− 1) (cal⋅mol− 1)
mixtures over a wider range of conditions. Sensitivity, reaction path, and
1 CH3OCH3 + NH2 = 1.8E00 3.610 4353 Dai [14] rate of production analyses were performed using the NH3/DME
CH3OCH2 + NH3
2 CH3 + NH2(+M) = 7.2E12 0.420 0 Jodkowski
mechanism. Experiments were conducted to further understand the
CH3NH2(+M) [37] combustion characteristics of NH3/DME fuel, to provide references for
3 CH3OCH2 + 7.2E12 0.420 0 est developing reasonable NH3/DME alternative fuels, to construct their
NH2(+M) = kCH3+NH2 chemical reaction kinetic models, and to provide reliable experimental
CH3OCH2NH2(+M)
data for validating the kinetic mechanism.
4 CH2O + NH2 = 6.3E04 3.000 3770 Li [38]
CHO + NH3
5 CH3O + NH2 = 3.3E06 1.94 1150 Gao [39] 2. Experimental setup
NH3 + CH2O
6 CH3O + NH2 = 6.8E14 0.11 2332 Dean [40]
H2NO + CH3 2.1. Experimental method
7 CH3OCH3 + NO2 = 5.8E01 3.500 23,755 Guan [41]
CH3OCH2 + HONO The experiments were performed in a stainless-steel chemical shock
8 CH3OCH3 + NO2 = 6.5E02 3.000 23,176 Guan [41] tube with an inner diameter of 100 mm and a wall thickness of 17.5 mm.
CH3OCH2 + HNO2
The tube consisted of a driver section of 4.0 m long and a driven section
of 5.5 m long. Polycarbonate diaphragms were used to separated the
concentration and equivalence ratio on the oxidation and reactivity of driving gas from the test mixture and punctured using a needle-punched
NH3/DME mixtures in flow reactors, and the trends of hot flame loca­ diaphragm-breaking mechanism to satisfy the test requirements under
tions were confirmed using mechanism simulations. Dai et al. [28] different conditions. The test setup is shown in Fig. 1, where fast-
measured the IDTs of NH3/DME mixtures using a rapid compression response piezoelectric pressure transducers (PCB113B24) were
machine at equivalence ratios of 0.5 to 2.0, pressures of 1.0 to 7.0 MPa, installed at the end of the driven section, PCB3 was located 20 mm from
and temperatures of 610 to 1180 K. Based on the DTU mechanism [26] the end wall, and the distance between each pressure transducer was
and the Glarborg mechanism [19], a modified mechanism for the NH3/ 200 mm to measure the velocity of the incident shock wave. A photo­
DME mixture was proposed. The results showed that the addition of electric multiplier (PMT, Hamamatsu CR131) were mounted in the same
trace amounts of DME can shorten the IDTs of NH3. Yin et al. [30] axial position as the PCB3 pressure transducers and a 307 ± 7.5 nm
determined the laminar flame velocity of the NH3/DME mixture at high bandpass filter located at the front of the PMT was used to record the OH
pressure and temperature at different equivalence ratios. The Dai [28] radical emissions.
and Zhang mechanisms [31] were coupled, and the simulated values The mixtures were allowed to mix for 12 h to ensure spatial homo­
were in good agreement with the experimental data. geneity. The inner walls of the mixing tank and the shock tube were
Reported studies on the IDTs of NH3/DME mixtures have mainly passivated with NH3 before the preparation of the gas mixture and each
focused on low-temperature chemistry below 1000 K. Further studies experiment to prevent NH3 adsorption. The passivation method was
are needed to validate the applicability of the mechanism under a wider performed by introducing about 20 kPa of NH3 into the mixing tank and
range of test conditions or to understand certain reaction paths. There­ shock tube for at least 15 min, then pumping to vacuum for a continuous
fore, this study used a shock tube to measure the IDTs of the NH3/DME/ time of 5 min (absolute pressure less than 10 Pa and gas leakage rate less
O2/Ar mixture at 1150–1950 K, 0.14/1.0 MPa, 0.5/1/2 equivalence than 1 Pa/hour for the whole unit) and then filling the fuel mixture for
ratio, 100/0, 95/5, 90/10, and 70/30 blending ratios of NH3 to DME. the test. The test mixture was prepared according to Dalton’s law of
partial pressure. The purity of the various substances used in this test

3
Y. Jin et al. Fuel 343 (2023) 127885

Fig. 6. Comparison of the IDTs of NH3/DME/O2/Ar mixture between the measured data and the predictions with different models.

was 99.99 % NH3, 99.999 % O2, 99.999 % Ar, 99.99 % DME and 99.999
VS
% He. Helium and nitrogen were used as the driving gases in this study, M = √̅̅̅̅̅̅̅̅̅̅ (2)
γRT1
and all fuel mixtures were diluted with 80 % argon to use an argon/
oxygen (79:21) mixture as the synthesis air. The molar composition of √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
( )2 ( ̅
the mixture is shown in Table 1. 1 − Δz )2
δVS = δΔz + δΔt (3)
The IDT used in this study was defined as the time interval between Δt Δt2
pressure jump from PCB3 caused by the reflected shock wave and the
rise of OH * chemiluminescence, as shown in Fig. 2. The reflected shock ∂2 T5 ( ) δVS
δT5 = δM = 2AM − 2CM − 3 √̅̅̅̅̅̅̅̅̅̅̅̅ (4)
temperature and pressure were calculated using Gaseq [32], which is a ∂M γ1 RT1
software for chemical equilibrium according to the reflected shock wave
speed. The standard square root (RSS) method [33] was used to assess where T5 is the reflected shock temperature (K); T1 is the initial tem­
the uncertainty, perature (K); γ is the adiabatic exponent; vS is the velocity of the incident
shock wave (m/s); and R is the universal gas constant(kJ/kmol.K). The
T1 [2(γ− 1)M2 + (3 − γ)][(3γ − 1)M2 − 2(γ − 1)] main source of measurement uncertainty is T5, which is estimated to be
T5 = = AM 2 + B + CM − 2
(γ + 1)2 M 2 no more than 20 K using Eqs. (1)–(4), leading to a maximum uncertainty
(1) in IDTs within ±10 %, depending on the experimental conditions.

4
Y. Jin et al. Fuel 343 (2023) 127885

2.2. Kinetic mechanism

CHEMKIN Pro is used for the kinetic analysis of combustible gas


mixtures in a constant-volume homogeneous reaction system. In this
study, the combined mechanism of the NH3/DME mixture is established
and used for calculations. A comparison of the experimental data of
NH3/O2/Ar mixture IDTs with the literature mechanisms at an equiva­
lence ratio of 1.0, and a pressure of 0.14/1.0 MPa is performed, as shown
in Fig. 3. The experimental data are closer to those of the Shrestha [26]
and Glarborg mechanisms [19]. The Zhang mechanism [34] shows
better agreement at 1.0 MPa, and the Okafor mechanism [35] over-
predicted the IDTs for pure ammonia combustion under all conditions.
Generally, although there are some differences in the literature mech­
anisms, all of the reported mechanisms can reasonably predict the IDTs
obtained from the experimental data. Considering the uncertainties in
different mechanisms, to better compare the carbon and nitrogen sub­
mechanisms, the extensively validated Burke [36] and Hashemi [29]
mechanisms are chosen as the candidate mechanisms for the DME re­
action subset. The Shrestha [26], Zhang [34], and Glarborg [19]
mechanisms were chosen as the candidate mechanisms for the NH3 re­
action subset. These mechanisms were combined separately, and
representative results are shown in Fig. 4. The Glarborg-Hashemi
mechanism under-predicts IDTs at both pressures of 0.14 and 1.0
MPa. The Shrestha-Hashemi and Glarborg-Burke mechanisms predict
well at high pressure, but these two mechanisms over-predicted IDTs by
30 % at a pressure of 0.14 MPa and high temperatures. The Zhang-Burke
and Zhang-Hashemi mechanisms over-predict IDTs by 55 % at a pres­
sure of 1.0 MPa and low temperature conditions. In conclusion, although
there are some differences between the above combined mechanisms, all
of them can reasonably predict the IDTs of NH3/DME. Considering it
comprehensively, the combined mechanism of Shrestha’s [26] and
Burke’s mechanism [36] is used as the basic mechanism. The Shrestha-
Burke mechanism provides optimum reproducibility of the IDTs
measured for NH3/DME mixture combustion.
In this study, new reaction paths have been added to improve the
prediction of the kinetic mechanism. The cross-reaction between DME
and NH3 in the Dai mechanism [28] is used to describe the low-
temperature oxidation process of DME and its interaction with NH3.
The amine (NH2) subset is crucial for this kinetic mechanism. The H-
abstraction reaction of CH3OCH3 + NH2 = CH3OCH2 + NH3 by DME
through NH2 radicals to produce CH3OCH2 is one of the most important
reactions. The pre-exponential factor Ea from ab initio theory in the Dai
mechanism [28] is cited. The reactions of NH2 with hydrocarbon radi­
cals have been added. The rate coefficient for the reaction CH3 +
NH2(+M) = CH3NH2(+M) is taken from the study of Jodkowski et al.
[37], and the rate coefficient for the reaction CH3OCH2 + NH2(+M) =
CH3OCH2NH2(+M) is estimated by this value. Among the reactions to
generate stable species, the rate coefficients for CH2O + NH2 = HCO +
NH3 is taken from the theoretical study by Li et al. [38] and the rate
coefficients for CH3O + NH2 = NH3 + CH2O is taken from the study by
Gao et al. [39]. Other simple C–N interaction reactions are selected
from the Glarborg et al. [19]. Some of the key reactions are shown in
Table 2. The comparison between the simulations of IDTs of NH3/DME
mixtures using the Shrestha-Burke mechanism and the developed
mechanism of this study is shown in Fig. 5. It can be found that the final
mechanism predicts better IDTs than the Shrestha-Burke mechanism at
low temperature conditions.
To further validate the predictive ability of the proposed mechanism,
a comparison with Issayev [25] and Dai mechanisms [28] for the pre­
diction of the IDTs at an equivalence ratio of 1.0, DME molar fraction of
5/25/40 %, and pressures of 0.14 to 4.0 MPa respectively is conducted,
Fig. 7. Comparison laminar burning velocity of NH3/DME/air mixture between
as shown in Fig. 6. The Dai mechanism [28] over-predicts the IDTs under
the measured data and the predictions with different models.
high-temperature conditions. The Issayev mechanism [25] remains in
good prediction with the experimental results at 1.0 MPa, but the pre­
diction is higher at the pressure of 0.14 MPa. At low temperatures, Dai
mechanism and Issayev mechanism under-predict the IDTs with 5 %

5
Y. Jin et al. Fuel 343 (2023) 127885

Fig. 8. The IDTs of NH3/DME/O2/Ar mixtures with different DME content and pressure.

DME addition and pressure of 4.0 MPa. Issayev over-predicts the IDTs at and 100 % of the total fuel, respectively. The NH3-DME mechanism
25 % DME addition. It should be noted that in Fig. 6(b-d), the heat loss of predicts the experimental data reasonably well. It is observed that the
rapid compression machine in simulations were not considered in this IDTs decrease with increasing temperature. The IDTs of the NH3/DME
work. Small differences are expected in the predictions of the IDTs. mixture show a strong dependence on DME molar fraction. The addition
Overall, the proposed mechanism can efficiently predict the IDTs of of a small amount of DME can promote the ignition and significantly
NH3/DME blended fuels under different conditions. shorten the IDTs, and the promotion effect becomes more evident with
Fig. 7 shows the comparison between the mechanism predictions an increase in the molar fraction of DME. For example, at an equivalence
(this study, Issayev et al. [25] and Xiao et al. [42]) and experimental ratio of 1.0 and a pressure of 0.14 MPa, the D5, D10, and D30 mixtures
data (Issayev et al. [25] and Xiao et al. [42]) for the NH3/DME/air reduce the IDTs by 63, 72, and 92 %, respectively, relative to the D0
laminar burning velocity at temperature of 298 K, pressure of 0.1/0.3/ mixture at 1500 K. The IDTs at an equivalent ratio of 1.0 and a pressure
0.5 MPa, NH3/DME blending ratio of 80/20 and 50/50 and different of 1.0 MPa are shown in Fig. 8(d) and the pressure dependence of the
equivalence ratios. Among them, The experimental data of Issayev et al. IDTs is found to be significant. At a pressure of 1.0 MPa, the reactivity of
[25] has a DME blending ratio of 47. The differences between the pre­ the mixture increased with the addition of DME. At an equivalence ratio
dictions of laminar burning velocity by the three mechanisms are of 1.0 and pressure of 1.0 MPa, the D5, D10, and D30 mixtures shortened
smaller at low DME molar fraction. However, Issayev mechanism over- the IDTs by 59, 70, and 89 %, respectively, relative to the D0 mixture at
predicted the laminar burning velocity with increasing DME molar 1500 K. The DME promoting effect decreases with increasing pressure.
fraction. In conclusion, the present model can predict NH3/DME/air
laminar burning velocity well. 3.2. Effect of DME blending

3. Results and discussion Fig. 9 shows the variation in IDTs with temperature for the NH3/
DME/O2/Ar mixture at.014 MPa and different equivalence ratios. The
3.1. Effect of equivalence ratios and pressure IDTs of pure NH3 increase with an increase in the equivalence ratio. The
change in the equivalence ratio leads to a weaker change of the IDTs
Fig. 8 shows variation of IDTs with temperature for the NH3/DME/ when DME is blended. Specifically, the IDTs do not vary significantly
O2/Ar mixture under different DME blending ratios; the solid lines with the equivalence ratio at lower temperatures, whereas it increases
denote model simulations with the NH3-DME mechanism. D0, D5, D10, slightly with an increase in the equivalence ratio at higher temperatures.
D30, and D100 denote the molar fractions of DME blending of 0, 5, 10, 30, At 1500 K, for the D5 mixture, a reduction in the equivalence ratio from

6
Y. Jin et al. Fuel 343 (2023) 127885

Fig. 9. The IDTs of NH3/DME/O2/Ar mixtures with different equivalent ratios.

2.0 to 1.0 resulted in a 21 % reduction in the IDTs, and a reduction from where, τ is the ignition delay time, Si is the sensitivity coefficient of the
1.0 to 0.5 resulted in a 15 % reduction in the IDTs. For the D30 mixture, elementary reaction to the ignition delay time, and ki is the reaction rate
the reduction in the equivalence ratio from 2.0 to 1.0 resulted in a 32 % coefficient of the ith elementary reaction.
reduction in the IDTs and from 1.0 to 0.5, resulted in a 43 % reduction in Fig. 10 shows the sensitivity coefficient analyses of the IDTs of D5,
the IDTs. In contrast to pure NH3, the IDTs of the NH3/DME mixture D10, and D30 mixtures at 1300 and 1600 K. First, under the current
exhibit weakly equivalence ratio dependence, similar to the IDTs pattern conditions, the chain-branching reaction R39 (H + O2 = O + OH) is
of pure DME [36]. The reactions related to DME oxidation, especially the always an ignition-promoting reaction with the largest negative sensi­
low-temperature chain branching reactions, are crucial for the com­ tivity coefficient among all NH3/DME mixtures, and its sensitivity co­
bustion and IDTs of the mixture. As the temperature increases, a gradual efficient decreases with increasing DME molar fraction. Similarly, R257
transition to nitrogen chemistry dominates the reaction. Therefore, in (NH2 + NO = NNH + OH) promotes the reactivity of the whole system
the high-temperature region, the variation in the IDTs with the equiv­ by consuming the intermediate NH2 to produce a large number of OH
alence ratio is more pronounced than that in the low-temperature radicals and sustains the thermal De-NOX mechanism in ammonia
region. chemistry. The addition of DME promotes the development of the free
radical pool at the beginning of the reaction, especially OH radicals,
3.3. Mechanism analysis leading to reduced sensitivity coefficients of R39 and R257. N2H2 pro­
duces reactive radicals, such as H2O2 and H, through the H-abstraction
3.3.1. Sensitivity analysis reaction R1397 (N2H2 + HO2 = NNH + H2O2) and the decomposition
A previous analysis showed that DME has a significant promoting reaction R1393 (N2H2 + M = NNH + H + M), along with the production
effect on the ignition of NH3, which is especially evident in the low- of the major intermediate, NNH. Similar to the sensitivity analysis for
temperature region. Therefore, to identify the key reactions that domi­ pure ammonia, reactions R275 (HNO + O2 = HO2 + NO) and R1497
nate the ignition of NH3 blending with DME, a sensitivity analysis of the (H2NO + O2 = HNO + HO2) promote ignition and have relatively high
IDTs is conducted for NH3 at an initial pressure (0.14 MPa), initial sensitivity coefficients. However, the sensitivities of R275 and R1497
temperatures (1300 and 1600 K), and equivalence ratios (0.5, 1.0, and decrease with increasing molar fraction of DME. The sensitivity coeffi­
2.0). The sensitivity coefficient was defined as follows: cient of the DME unimolecular decomposition reaction R632 (CH3OCH3
+ M = CH3 + CH3O + M) increases with increasing DME. Reactive
τ(2ki) − τ(0.5ki)
Si = (5) carbon-containing species can react with intermediate species to pro­
1.5τ(ki)
duce more reactive radicals such as OH, which can further promote the

7
Y. Jin et al. Fuel 343 (2023) 127885

subsequently decreases as the molar fraction of DME increases. The


chain termination reaction R256 (NH2 + NO = N2 + H2O) competes
with ignition-promoting reaction R257 for NO, thereby inhibiting igni­
tion. As the molar fraction of DME increases, the sensitivity to R256
increases substantially. In reactions with large sensitivity coefficients,
where C–N interactions are less frequent, DME promotes NH3 ignition,
primarily by promoting the development of free radical pools. Sensi­
tivity analysis shows that for D5 mixture, the absolute sensitivity co­
efficients for reactions such as R39 and R275 that consume O2 and
generate OH and HO2 reactive radicals decrease with increasing tem­
perature, while the absolute sensitivity coefficients for DME decompo­
sition reactions also decrease substantially. The inhibition of O2 and
DME promotion at high temperatures is the key reason for the reduced
catalytic effect of DME on ammonia fuel ignition at high temperatures.
The sensitivity coefficients of the reactions decreased at 1600 K, whereas
the sensitivity coefficients of the inhibited reactions R278 and R170 in
the D5 mixture are similar to those of the D30 mixture at 1300 K. Thus,
the sensitivity coefficient of the inhibition reaction R231 (NO2 + H =
NO + OH) relating to NO2 increased with the increase in temperature.
The sensitivity coefficient of the carbon chemistry reaction dramatically
is reduced; therefore, the promotion effect of DME on the system is
primarily in the low-temperature region.
As shown in Fig. 10(b), to reveal the effect of the equivalence ratio on
the IDTs and kinetics, a sensitivity analysis was performed for the D30
mixture at a pressure of 0.14 MPa, a temperature of 1300 K, and
equivalence ratios of 0.5, 1.0, and 2.0. The sensitivity coefficients of the
reactions to compete for O2 show a large variation as the equivalence
ratio increases. Among them, the sensitivities of R39 and R171 (HCO +
O2 = HO2 + CO) increase with an increasing equivalence ratio. As the
equivalence ratio increase, excess NH3 and H2 consume more OH radi­
cals, leading to DME-relevant reactions R104 (OH + CH2O = HCO +
H2O) and reaction R102 (OH + CO = H + CO2) showing a significant
decrease in the sensitivity coefficients. However, the sensitivity of other
key reactions remain almost constant as the equivalence ratio changed.
The experimental IDTs data are weakly dependent on the equivalence
ratio, and the proposed mechanism in this study provides a good
prediction.

3.3.2. Reaction path


To further investigate the reason for the promotion of DME on NH3
ignition, a reaction path analysis of the NH3/DME mixture at a fuel
consumption of 20 %, a temperature of 1300 K, a pressure of 0.14 MPa,
and different DME molar fractions were conducted, as shown in Fig. 11.
First, almost all NH3 is consumed by H-abstraction reaction to produce
NH2⋅NH3 reacts primarily with OH radicals via R278 to produce NH2
and H2O, which are the most sensitive inhibitory reactions under all
studied conditions. The addition of DME increases the reaction flux of
NH2 through the oxidation reaction to produce HNO, but the reaction
flux of NH2 through the dehydrogenation reaction to produce NH re­
Fig. 10. Sensitive analysis of the IDTs time at different DME blending ratios(P mains almost unchanged. NH via NO reduction reactions R240 (NH +
= 0.14 MPa, ϕ= 0.5/1.0/2.0, T = 1300/1600 K). NO = N2O + H) and R242 (NH + NO = N2 + OH) to generate N2O or N2.
The generated N2O reacts with CO to generate CO2 and N2. In addition,
system activity and shorten the IDTs. another important path of NH consumption is consumed by reaction
The main H-abstraction channels during ignition such as reactions R1396 (NH + NH = N2H2), which subsequently produces NNH.
R278 (NH3 + OH = NH2 + H2O) and R634 (CH3OCH3 + H = CH3OCH2 DME produces CH3OCH2 primarily through H-abstraction reactions,
+ H2) inhibit ignition. Almost all the OH radicals are consumed by R278 while a small proportion produces CH3 and CH3O through unimolecular
and form stable species H2O, resulting in the whole reaction system decomposition reactions. CH3O radicals are easily generate H radicals
maintaining a relatively low OH radical molar fraction until ignition. As through the decomposition reaction R58 (CH3O + M = H + CH2O + M).
the molar fraction of DME increases, more OH radicals that are provided Therefore the reaction fluxes associated with H radicals increased
to the system in the early stage of the reaction further reduce the significantly with the increasing DME molar fraction. For D5, the the
sensitivity of R278. The H radicals are consumed by R634 in the early proportion of DME via H-abstraction reactions with NH2 to produce
stage of the reaction, but when the fuel blends consume 20 %, DME CH3OCH2 was 8.3 %. For D30, DME is consumed primarily through H-
consumption is over 90 %, and the system has abundant H radicals. The abstraction reaction R634, due to the higher DME molar fraction, only
reaction R273 (HNO + H = H2 + NO) and the unimolecular decompo­ 1.8 % of DME is consumed by the reaction with NH2. The oxidation of
sition reaction R170 (HCO + M = H + CO + M) inhibit the system ac­ CH3 to CH2O produce a large number of OH radicals. R122 replaces the
tivity, but the sensitivity of HNO to the overall reaction system pathway of CH3 reacting with NO2 to produce CH3O.

8
Y. Jin et al. Fuel 343 (2023) 127885

Fig. 11. Reaction path of NH3/DME/O2/Ar. The colors differentiates the pathways for the three blends: black (D5 mixture), red (D10 mixture), and blue
(D30 mixture).

It can be seen from the reaction path that NO is mainly produced by rates exceeded 90 % at 20 % of total fuel consumption. At higher DME
the oxidation of NH, HNO, and N. Different DME molar fractions lead to molar fraction, H, OH, and HO2 in the radical pool developed earlier. To
different paths for NO production. HNO is primarily produced by the further understand the effect of active radicals in the system, the trend of
oxidation reaction R275 (HNO + O2 = HO2 + NO) for D5. The propor­ the molar fraction of active radicals with the consumption rate of fuel
tion of the reaction path that produces NO through oxidation decreases with the addition of DME was plotted at an equivalence ratio of 1,
for D30, and the amounts of NO produced by the reactions R273 (HNO + pressure of 0.14 MPa, and temperature of 1300 K, as shown in Fig. 12. As
H = H2 + NO) and R1561 (CH3 + HNO = CH4 + NO) increase. As seen from Fig. 12(a–c), the molar fraction of O/H/OH radicals increase
increasing DME molar fraction, more O2 produces CO through the significantly with increasing DME molar fraction for a given proportion
oxidation of HCO, which competes with R275. Reaction R273 replaces of fuel consumed. This means that there is a chemical environment to
R275 as the most NO-producing reaction path, and owing to the rapidly consume NH3 before the main ignition. Notably, the OH, O, and
increasing CH3 molar fraction in the system, R1561 also becomes an H radical molar fractions of the D5 mixture increase by an order of
important part of NO production by consuming HNO. The NO con­ magnitude at the beginning of the reaction compared to those of the D0
sumption path almost remains unchanged with the addition of DME. NO mixture. This further confirms that the IDTs of NH3 can be significantly
is primarily consumed through reactions with radicals such as NH2 and reduced by adding a trace amount of DME. HO2 radical molar fraction
NH to produce N2. NO reacts with NH2 to produce NNH, which is further increases rapidly with increasing DME at the beginning of the reaction.
reduced to N2. NO is readily converted to NO2 by oxidation reactions. HO2 radical molar fraction produces more reactive CH3O and OH radi­
From the reaction path analysis, it is observed that NO and NO2 exhibit cals in the early stages of the reaction, primarily through R122 (HO2 +
good performance in promoting NH3/DME combustion, which promotes CH3 = OH + CH3O) in the consumption of CH3. Therefore, DME is the
the reactivity of the system. The results suggest that the addition of DME key reason for promoting the IDTs of NH3 mainly by accelerating the
changes the NO production reaction path and partially affects NO early accumulation of radical pools.
emissions. Compared with NO, the production pathway of CO2 is rela­ The rate of production of NH3 and OH radicals is also analyzed under
tively simple. As shown in Fig. 11, CO2 is mainly produced through the the same conditions, as shown in Fig. 13. As the reaction proceeds, the
sequence of DME → CH3OCH2 → CH2O → HCO → CO → CO2. As the chain-branching reaction R39 (H + O2 = O + OH) becomes the main
DME molar fraction increases, the system obtaineds more OH radicals, source of the OH radicals. Under pure ammonia conditions, reactions
leading to an increase in CO consumption via the reaction with OH and a R278 (NH3 + OH = NH2 + H2O) and R85 (OH + H2 = H + H2O) become
gradual decrease in CO consumption via reactions with N2O and HNO. the primary consumption channels for OH radicals. R278 changes
Small molecules and radicals in nitrogen chemistry play an important slightly by the variation of DME molar fraction. For all the key reactions
role in influencing the kinetics of DME oxidation but have a negligible shown in Fig. 13(a), the ROP peaks occur near the moment of ignition.
effect on the CO2 production pathway. With the increase of DME molar fraction, OH radical consumption and
Prior to ignition, the early oxidation of DME is the main source of OH production rates through reactions R39 and R85 increase, and the
radicals, initiating NH3 decomposition and promoting the development ignition temperature shifts to the lower temperature region. It is note­
of the radical pool. Therefore, the oxidation mechanism of DME is key to worthy that R122 (HO2 + CH3 = OH + CH3O) and R86 (H2O2 + M =
promoting the reaction activity and shortening the IDTs. 2OH + M) produce large amounts of OH radicals at the beginning of
ignition, rather than at the moment of ignition. This indicates that the
3.3.3. Radical mole fraction free radical pool is pre-established at the initial stage of ignition due to
Early oxidation of DME was observed, i.e. the DME consumption DME addition, promoting the consumption of NH3, especially the

9
Y. Jin et al. Fuel 343 (2023) 127885

Fig. 12. Changes of small molecules’ species with fuel consumption ratio during ignition.

reaction initiation. As the molar fraction of DME increases, a new OH 4. Conclusions


radical-consuming reaction path is added, R104 (OH + CH2O = HCO +
H2O), which is also an ignition-promoting reaction with a large sensi­ In this study, to extend the database of NH3/DME combustion and to
tivity factor in the D30 mixture. Fig. 13(b) shows that NH3 is primarily further develop the detailed kinetic mechanism, the IDTs of NH3/DME
consumed by the H-abstraction reactions. Compared to pure ammonia, were measured at pressures of 0.14 and 1.0 MPa, temperatures ranging
the H-abstraction reactions remain the most competitive ammonia from 1150 to 1950 K, equivalence ratios of 0.5/1.0/2.0, and NH3/DME
consumption reaction with the addition of DME, and the rate of NH3 mixture ratios of 100/0, 95/5, 90/10, and 70/30 were measured by a
consumption increases significantly with increasing DME molar frac­ shock tube. The Shrestha-Burke mechanism is used as the basic mech­
tion. In the presence of DME, the chain propagation reaction R122 anism, and the interaction mechanism between C1 ~ C2 species and N-
provides a different reaction path from pure NH3 combustion. R122 related species is updated to add cross-reactions between DME and NH3
produces a large number of OH and CH3O radicals during the initial subsets and to extend the key reactions of low-temperature DME
ignition stage. The produced CH3O radicals are decomposed by reaction oxidation and its interactions with NH3. The final mechanism can better
R58 (CH3O + M = H + CH2O + M) to produce H radicals, which then predict the ignition delay times and laminar burning velocities for a
promotes OH radicals via R39. The generated OH radicals further pro­ wide range of conditions.
mote the combustion of NH3 via R278, which explains the large increase Sensitivity analysis, reaction path analysis, and rate of production
in the sensitivity coefficient of reaction R122 with the increasing molar analysis were performed on the blended fuel using the established
fraction of DME in Fig. 10. This path increases the system reactivity and mechanism. The results showed that the effect of DME on NH3 presents
promotes the production of active radicals in the low-temperature range considerable temperature and pressure dependence and is particularly
during the initial ignition stage. As the molar fraction of DME increases, pronounced at lower temperatures and higher pressures. DME promoted
the fuel consumption rate increases, and the development of the free the overall reaction activity by accelerating the development of the
radical pool becomes faster and shifts to a lower temperature range. radical pool in the early stages of ignition, particularly the rates of
Therefore, the early oxidation process of DME is vital for promoting production and molar fractions of O, H, and OH radicals.
reactivity. In the presence of DME, combustion produces more reactive inter­
mediate species, which accelerate the radical pool development. As DME
molar fraction increased, R122 (HO2 + CH3 = OH + CH3O) provided a

10
Y. Jin et al. Fuel 343 (2023) 127885

1.0E-4

5.0E-5

0.0E+0

-5.0E-5

-1.0E-4
1300 1350 1400 1450

Fig. 13. Rate of productions of OH radicals and NH3.

different reaction path from pure NH3 combustion, generating a large (H + O2 = O + OH). Abundant OH radicals were consumed via R278
number of OH and CH3O radicals. The generated CH3O radicals easily (NH3 + OH = NH2 + H2O), which further accelerated the combustion of
decomposed via the reaction R58 (CH3O + M = H + CH2O + M) to NH3.
produce reactive H radicals and promote OH radical formation via R39

11
Y. Jin et al. Fuel 343 (2023) 127885

CRediT authorship contribution statement [16] Yu L, Zhou W, Feng Y, et al. The effect of ammonia addition on the low-
temperature autoignition of n-heptane: An experimental and modeling study.
Combust Flame 2020;217:4–11.
Yifan Jin: Data curation, Writing – original draft, Methodology. Xin [17] Song Y, Hashemi H, Christensen JM, et al. Ammonia oxidation at high pressure and
Li: Conceptualization, Investigation. Xin Wang: Project administration, intermediate temperatures. Fuel 2016;181:358–65.
Funding acquisition. Zhihao Ma: Supervision, Project administration, [18] Zhang K, Banyon C, Bugler J, et al. An updated experimental and kinetic modeling
study of n-heptane oxidation. Combust Flame 2016;172:116–35.
Writing – review & editing. Xianglin Chu: Software, Validation, [19] Glarborg P, Millr JA, Ruscic B, et al. Modeling nitrogen chemistry in combustion.
Methodology. Prog Energy Combust Sci 2018;67:31–68.
[20] De Vries J, Lowry WB, Serinyel Z, et al. Laminar flame speed measurements of
dimethyl ether in air at pressures up to 10 atm. Fuel 2011;90(1):331–8.
Declaration of Competing Interest [21] Arcoumanis C, Bae C, Crookes R, et al. The potential of di-methyl ether (DME) as
an alternative fuel for compression-ignition engines: A review. Fuel 2008;87(7):
The authors declare that they have no known competing financial 1014–30.
[22] Chen Z, Qin X, Ju Y, et al. High temperature ignition and combustion enhancement
interests or personal relationships that could have appeared to influence by dimethyl ether addition to methane–air mixtures. Proc Combust Inst 2007;31
the work reported in this paper. (1):1215–22.
[23] Dames EE, Rosen AS, Weber BW, et al. A detailed combined experimental and
theoretical study on dimethyl ether/propane blended oxidation. Combust Flame
Data availability 2016;168:310–30.
[24] Hu E, Zhang Z, Pan L, et al. Experimental and modeling study on ignition delay
Data will be made available on request. times of dimethyl ether/propane/oxygen/argon mixtures at 20 bar. Energy Fuel
2013;27(7):4007–13.
[25] Issayev G, Giri BR, Elbaz AM, et al. Ignition delay time and laminar flame speed
Acknowledgement measurements of ammonia blended with dimethyl ether: A promising low carbon
fuel blend. Renew Energy 2022;181:1353–70.
This work was supported by the National Natural Science Foundation [26] Shrestha KP, Seidel L, Zeuch T, et al. A Detailed kinetic mechanism for the
oxidation of ammonia including the formation and reduction of nitrogen oxides.
of China (Grant No. 51906061) Energy Fuel 2018;32(10):10202–17.
[27] Murakami Y, Nakamura H, Tezuka T, et al. Effects of mixture composition on
References oxidation and reactivity of DME/NH3/air mixtures examined by a micro flow
reactor with a controlled temperature profile. Combust Flame 2022;238:111911.
[28] Dai L, Hashemi H, Glarborg P, et al. Ignition delay times of NH3/DME blends at
[1] Giddey S, Badwal SPS, Munnings C, et al. Ammonia as a renewable energy high pressure and low DME fraction: RCM experiments and simulations. Combust
transportation media. ACS Sustain Chem Eng 2017;5(11):10231–9. Flame 2021;227:120–34.
[2] Zamfirescu C, Dincer I. Using ammonia as a sustainable fuel. J Power Sources 2008; [29] Hashemi H, Christensen JM, Glarborg P. High-pressure pyrolysis and oxidation of
185(1):459–65. DME and DME/CH4. Combust Flame 2019;205:80–92.
[3] Chiuta S, Everson RC, Neomagus HWJP, et al. Reactor technology options for [30] Yin G, Li J, Zhou M, et al. Experimental and kinetic study on laminar flame speeds
distributed hydrogen generation via ammonia decomposition: A review. Int J of ammonia/dimethyl ether/air under high temperature and elevated pressure.
Hydrogen Energy 2013;38(35):14968–91. Combust Flame 2022;238:111915.
[4] Valera-Medina A, Amer-Hatem F, Azad AK, et al. Review on ammonia as a [31] Zhang X, Moosakutty SP, Rajan RP, et al. Combustion chemistry of ammonia/
potential fuel: from synthesis to economics. Energy Fuel 2021;35(9):6964–7029. hydrogen mixtures: Jet-stirred reactor measurements and comprehensive kinetic
[5] Mørch CS, Bjerre A, Gøttrup MP, et al. Ammonia/hydrogen mixtures in an SI- modeling. Combust Flame 2021;234:111653.
engine: Engine performance and analysis of a proposed fuel system. Fuel 2011;90 [32] Morley C. Gaseq: a chemical equilibrium program for Windows. Ver. 0.79, 2005.
(2):854–64. [33] Petersen EL, Rickard MJA, Crofton MW, et al. A facility for gas-and condensed-
[6] Gross CW, Kong SC. Performance characteristics of a compression-ignition engine phase measurements behind shock waves. Meas Sci Technol 2005;16(9):1716.
using direct-injection ammonia–DME mixtures. Fuel 2013;103:1069–79. [34] Zhang Y, Mathieu O, Petersen EL, et al. Assessing the predictions of a NOX kinetic
[7] Reiter AJ, Kong SC. Demonstration of compression-ignition engine combustion mechanism on recent hydrogen and syngas experimental data. Combust Flame
using ammonia in reducing greenhouse gas emissions. Energy Fuel 2008;22(5): 2017;182:122–41.
2963–71. [35] Okafor EC, Naito Y, Colson S, et al. Experimental and numerical study of the
[8] Mathieu O, Petersen EL. Experimental and modeling study on the high-temperature laminar burning velocity of CH4–NH3–air premixed flames. Combust Flame 2018;
oxidation of Ammonia and related NOx chemistry. Combust Flame 2015;162(3): 187:185–98.
554–70. [36] Burke U, Somers KP, O’Toole P, et al. An ignition delay and kinetic modeling study
[9] Mei B, Zhang X, Ma S, et al. Experimental and kinetic modeling investigation on the of methane, dimethyl ether, and their mixtures at high pressures. Combust Flame
laminar flame propagation of ammonia under oxygen enrichment and elevated 2015;162(2):315–30.
pressure conditions. Combust Flame 2019;210:236–46. [37] Jodkowski JT, Ratajczak E, Fagerström K, et al. Kinetics of the cross reaction
[10] Shu B, Vallabhuni SK, He X, et al. A shock tube and modeling study on the between amidogen and methyl radicals. Chem Phys Lett 1995;240(1–3):63–71.
autoignition properties of ammonia at intermediate temperatures. Proc Combust [38] Li QS, Lü RH. Direction dynamics study of the hydrogen abstraction reaction CH2O
Inst 2019;37(1):205–11. + NH2→ CHO+ NH3. Chem A Eur J 2002;106(41):9446–50.
[11] Pochet M, Dias V, Moreau B, et al. Experimental and numerical study, under LTC [39] Gao CW, Allen JW, Green WH, et al. Reaction Mechanism Generator: Automatic
conditions, of ammonia ignition delay with and without hydrogen addition. Proc construction of chemical kinetic mechanisms. Comput Phys Commun 2016;203:
Combust Inst 2019;37(1):621–9. 212–25.
[12] Chen J, Jiang X, Qin X, et al. Effect of hydrogen blending on the high temperature [40] Dean A M, Bozzelli J W. Combustion Chemistry of Nitrogen. 1rd Ed. New York:
auto-ignition of ammonia at elevated pressure. Fuel 2021;287:119563. Springer New York,2000:125-341.
[13] Otomo J, Koshi M, Mitsumori T, et al. Chemical kinetic modeling of ammonia [41] Guan Y, Liu R, Lou J, et al. Computational investigation on the reaction of dimethyl
oxidation with improved reaction mechanism for ammonia/air and ammonia/ ether with nitric dioxide. II. Detailed chemical kinetic modeling. Theor Chem Acc
hydrogen/air combustion. Int J Hydrogen Energy 2018;43(5):3004–14. 2020;139(1):1–12.
[14] Dai L, Gersen S, Glarborg P, et al. Autoignition studies of NH3/CH4 mixtures at high [42] Xiao H, Li H. Experimental and kinetic modeling study of the laminar burning
pressure. Combust Flame 2020;218:19–26. velocity of NH3/DME/air premixed flames. Combust Flame 2022;245:112372.
[15] Tian Z, Li Y, Zhang L, et al. An experimental and kinetic modeling study of
premixed NH3/CH4/O2/Ar flames at low pressure. Combust Flame 2009;156(7):
1413–26.

12

You might also like