Reactivity Trends in Furan and Alkyl Furan ... - ACS Publications

Subscriber access provided by Northern Illinois University

Article

Reactivity trends in furan and alkyl furan combustion Mazen A Eldeeb, and Benjamin Akih-Kumgeh Energy Fuels, Just Accepted Manuscript • DOI: 10.1021/ef501181z • Publication Date (Web): 03 Sep 2014 Downloaded from http://pubs.acs.org on September 5, 2014

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Energy & Fuels is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Reactivity trends in furan and alkyl furan combustion Mazen A. Eldeeb and Benjamin Akih-Kumgeh∗ Department of Mechanical and Aerospace Engineering, Syracuse University, 263 Link Hall, Syracuse, NY 13244, USA E-mail: [email protected]

Fax: +1 (315)443-6999

Abstract A systematic study of the ignition behavior of furan and the substituted furans, 2-methyl furan (2-MF) and 2,5-dimethyl furan (DMF), is presented. Ignition delay times are measured over a temperature range of 977 K to 1570 K and pressures up to 12 atm for lean, stoichiometric, and rich mixtures of fuel, oxygen and argon. It is found that when the equivalence ratio, φ , the argon to oxygen ratio, D, and pressure, p, are kept constant over a range of temperatures, T , DMF generally has the longest while 2-MF has the shortest ignition delay times and furan shows intermediate reactivity. Ignition delay times decrease with increasing equivalence ratios, except for DMF, which does not show a conclusive trend over the temperature range investigated. The experimental data are also found to agree with published ignition data, showing differences in some cases partly related to disparities in endwall and sidewall ignition measurements. The ignition delay times of 2-MF and DMF are compared to predictions using furan chemical kinetic models by Sirjean et al. 1 and Somers et al. 2 The models show qualitatively that DMF has longer ignition delay times than 2-MF under similar conditions of φ , D, p, and T , as revealed by the experiments. Quantitatively, the model predictions agree with experimental data at conditions similar to those used in their development and deviations from ∗ To

whom correspondence should be addressed

1 ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

experiment at other conditions are mostly related to unmatched temperature sensitivities over a wider temperature range, revealed by varying pressure and reduced dilution. The reported experimental dataset contributes toward further understanding and improved modeling of the combustion of furans, a promising class of alternative fuels.

Introduction There is increased interest in utilizing furan-based compounds as alternative fuels for spark-ignition (SI) engines. This is particularly attractive since there is a great potential for mass production of furans from sugars, which can be obtained from second-generation biomass. A study by RománLeshkov et al. 3 has demonstrated the synthesis of 2,5-Dimethyl furan (DMF, C6 H8 O) through selective dehydration of fructose (C6 H12 O6 ) and subsequent catalytic processes. Similar methods for DMF production from glucose have been proposed by Tong et al. 4 and Chidambaram et al. 5 Further, Mascal et al. 6 have demonstrated the conversion of cellulose to DMF and 2-Methyl furan (2-MF, C5 H6 O). Compared to more established biofuels, such as ethanol, furans have more favorable properties. For instance, DMF has a higher energy density of 30 MJ/L, compared to an energy density of 20 MJ/L for ethanol. DMF has also been found to have a higher research octane number (RON) value of 119 compared to the 110 RON of ethanol, 7 and furans are generally less soluble in water than ethanol, 1,2 thus reducing the risk of water contamination. In view of these advantages, it is important to further investigate, characterize, and model the combustion properties of furans for engine applications. A number of engine studies have been carried out to evaluate the use of furans as alternative SI engine fuel or fuel additive. An engine study of DMF combustion by Zhong et al. 7 has shown that it has very similar combustion and emissions behavior to those of gasoline-fueled engines, thus indicating the possibility of DMF-fueled SI engines without major engine modifications. Daniel et al. 8 compared the carbonyl emissions of DMF combustion with those of gasoline and bio alcohols; and DMF was observed to produce the lowest emissions of the harmful carbonyl compounds, formaldehyde and acetaldehyde. Also, a study by Rothamer et al. 9 has shown that adding DMF 2 ACS Paragon Plus Environment

Page 2 of 25

Page 3 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

to gasoline exhibits better autoignition resistance than conventional gasoline. The other furan of interest is 2-MF. According to an engine test by Thewes et al., 10 2-MF shows better resistance to autoignition and lower NOx , HC, and PM emissions than RON 95 gasoline, especially for leaner mixtures. Similar results were obtained for furan derivatives and 2-MF in another engine study by Gouli et al., 11 showing that furan-based additives impart better knock resistance to gasoline, compared to other additives, such as MTBE, along with reduction in CO and HC emissions. 12 Apart from oxidation studies, experimental studies of pyrolysis are also being undertaken, such as the flow reactor studies of DMF pyrolysis by Cheng et al. 13 Fundamental studies of the combustion properties of furans are beginning to emerge in the literature, and the rapid growth of literature on pyrolysis, quantum chemical calculations, flames, and ignition data highlight interest in this class of fuels. Ignition delay measurements of DMF, 2-MF, and furan are still limited in the literature and these are crucial to the assessment of the relative ignition propensity of each fuel as well as validation of detailed and reduced chemical kinetic models. An experimental and modeling studies of DMF has been reported by Sirjean et al., 1 which includes shock tube ignition delay measurements for mixtures of DMF and oxygen in argon at pressures of 1.0 and 4.0 atm, over a temperature range of 1300 K - 1831 K, and at equivalence ratios range of 0.5 - 1.5. The experimental datasets are employed in the validation and improvement of the detailed chemical kinetic model presented therein. Another experimental and modeling study of DMF has been provided by Somers et al., 2 including ignition delay times for DMF in argon mixtures of equivalence ratios of 0.5 - 2.0 from 1350 K to 1800 K at 1 atm, and higher pressures of 20 atm and 80 atm. Although their ignition datasets are in good agreement with the new kinetic model, the authors show that the previous model by Sirjean et al. 1 predicts longer ignition delay times than the new model. The model by Somers et al. 2 incorporates the furan sub chemistry presented in the work of Sirjean et al. 1 With regard to other furans, ignition delay times for 2-MF have been reported in studies by Somers et al., 14 and later by Wei et al. 15 The study by Somers et al. 14 includes ignition delay


Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

measurements for 2-MF mixtures in argon of equivalence ratios of 0.5, 1.0 and 2.0 at atmospheric pressure over a temperature range of 1200 K – 1800 K. It also includes a chemical kinetic model which is in reasonable agreement with the ignition data. Later ignition data obtained by Wei et al. 15 at equivalence ratios of 0.25 to 2.0, pressures from 1.25 to 10.65 bar, and temperatures from 1120 K to 1700 K, are in reasonable agreement with the predictions of the model by Somers et al. 14 at lower pressures, but under-predicts ignition delay at higher pressures. Recently, Uygun et al. 16 investigated the ignition of tetrahydrofuran and 2-MF in a shock tube, also focusing on the dynamics of the ignition process through schlieren imaging. Ignition delay times for stoichiometric 2-MF/air mixtures are obtained at 40 atm and compared with model predictions. The authors observed that ignition predictions using models by Somers et al 2 agree with their experimental data to within 20%, in contrast to shorter ignition predictions observed by Wei et al. 15 at higher pressures than atmospheric. The authors observe that the model predicts longer ignition delay times than the experimentally observed trend at lower temperatures. For furan characterization, ignition delay time measurements have been obtained by Wei et al. 17 at temperatures of 1320 K - 1880 K, and pressures of 1.2 - 10.4 atm for dilute mixtures with equivalence ratios of 0.5 - 2.0. The results show good agreement with predictions using a furan chemical kinetic model by Tian et al. 18 However, the model is observed to under predict ignition delay times for lean mixtures and especially at lower temperatures. Although ignition delay studies have been performed for each of these furans, direct comparison of ignition delay times various furan-based molecular systems has not been carried out to establish structure-reactivity trends. This is useful for further model development and for practical engine applications, where fuel technology focuses on relative combustion properties. In the study by Uygun et al., 16 the authors assume that 2-MF has similar ignition behavior as DMF, using this as a basis to scale DMF ignition data at 80 atm from a previous study 2 for demonstration. Further, fuel technology for combustion engines is largely focused on relative ignition behavior. In the production of furan-based fuels, blends of the three fuels can result with varying compositions. It is therefore of great interest to investigate the relative ignition behavior of these systems with the


Page 4 of 25

Page 5 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

aim of better understanding and informing the blending options for improved fuel performance. Such a comparison would have to be done under defined constraints, preferably chosen to reflect those used in engine operation. In modeling the kinetics of furan combustion, challenges are encountered given the limited database of kinetic parameters. To bridge this gap, computational chemistry studies have been carried out. 1,2,19 Pathway explorations are performed to guide the mechanistic description of the oxidation process. The models by Somers et al. 2 and Sirjean et al., 1 have extensively used results of such computational studies to assign reaction pathways and reaction rates. The underlying electronic structure calculations are mostly based on density functional theory with composite methods used to extrapolate to complete basis sets (CBS QB3, Gn , etc). These tools are very useful in the study of these oxygenated ring compounds. Models based solely on these studies still face challenges, such as the appropriate technique to incorporate uncertainties in key rate constants, and the incomplete understanding of the oxidation mechanisms. Systematic experimental studies which examine key combustion properties over a wide range of conditions for several furans are needed to aid model development as well as guide engine development and operation processes. H H O

O H

H

H C

C C

C

C H

C

C

C

H

C

H

H

H

Furan

2-Methylfuran H H

H H O

C

C H

C

C H

C

C

H

H

2,5-Dimethylfuran Figure 1: The three fuels investigated in this work: furan, 2-methyl furan (2-MF), and 2,5-dimethyl furan (DMF). 5 ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

In this work, this need is addressed by investigating the ignition delay times of furan, 2-MF, and DMF (see Figure 1) under conditions of similar equivalence ratios, φ , ratio of argon to oxygen, D, and nominal pressure, p, over a range of temperatures. In addition to the preferred sidewall ignition delay time measurements, endwall measurements are taken for most of the experiments to elucidate differences in ignition delay times and enable direct comparison with data obtained using one or the other approach. We thus seek to establish the reactivity trends in furan combustion based on measurements of ignition delay times.

Experimental approach Experiments are carried out behind reflected shock waves in a newly constructed shock tube facility. The current set-up of the stainless steel shock tube consists of a 4.0 m driven section and a 2.7 m driver section. It will subsequently be extended to a total length of 11.0 m with a 4.0 m driver. The inner diameter of the mechanically polished tube is 10 cm. Shock velocities are measured using four fast-response PCB transducers mounted 40 cm apart. Post-reflected shock temperatures are determined from the one-dimensional shock equations. Postreflected shock pressures are compared to measurements and generally show excellent agreement. The temperature behind the reflected shock is calculated using the CANTERA software package 20 along with the Caltech shock and detonation toolbox. 21 Both sidewall and endwall ignition delay time measurements in the new facility can be simultaneously performed by means of mounted optical fibers, connected to photodiodes with appropriate narrow band filters. Differences between endwall and sidewall ignition delay measurements have been extensively discussed by Petersen, 22 with a recommendation in favor of sidewall over endwall ignition measurements. The photodiodes are equipped with 430±10 nm narrow band filters to obtain CH chemiluminescence signals for ignition delay time determination. The ignition delay time is defined as the time from the arrival of the reflected shock wave at the optical fiber port to the intersection of line of maximum gradient with the initial baseline signal. Ignition delay simulations


Page 6 of 25

Page 7 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

are performed using the CANTERA software package. 20 In this study, mixtures are prepared manometrically in a 150-liter tank using a high precision MKS Baratron pressure transducer. Research grade samples of furan, 2-MF, and DMF (SigmaAldrich, at least 99%) are used for the study. Ultra high purity oxygen, argon, and helium gases (Airgas, >99.999%) are used in the experiment. Figure 2 is an example of the pressure and CH chemiluminescence signals used to determine the ignition delay times. The ignition delay times presented in this work are under 2 ms, so that the ignition delay times are not sensitive to pressure rise during the induction period which would otherwise have to be accounted for in simulations. The shock tube ignition experiment consists of two sub experiments; the first is the rapid compression and heating of the combustible mixture by the shock wave and the second is the evolution of this hot, pressurized non-equilibrium homogeneous reactor towards ignition. Uncertainties in the overall experiment can be related to the initial conditions for the reactor or to the actual determination of the ignition delay time from chemiluminescence signals. The greatest source of uncertainty is the post-reflected shock temperature, to which chemical reaction rates are very sensitive. Uncertainties arise from uncertainties in shock velocity measurements (exact sensor separation and time intervals), shock attenuation rate, and in mixture compositions. We have established that the actual ignition delay times determined from CH chemiluminescence signals are relatively more accurate with errors in slope fit and extrapolation within 3%. The temperature corresponding to this measurement is uncertain to about 20 - 30 K, based on error propagation analysis. Assuming an Arrhenius-type ignition delay dependence on temperature, an uncertainity of 25 K would translate into an ignition delay time uncertainty of up to 30% at 1000 K. However, it is assumed that these are fairly systematic and do not compromise the result of comparative studies. Further, ignition data obtained from the new shock tube are in agreement with those obtained at other shock tube facilities under similar conditions. In this work, we illustrate the shock tube validation by comparing ignition delay times of DMF and 2-MF to published data from other groups. The three furans, DMF, 2-MF, and furan are studied at equivalence ratios, φ , of 0.5, 1.0, and 2.0 over a range of pressures up to 12 atm. The mixtures


Energy & Fuels

1.0

Pressure CH emission Tangent to point of max. gradient Avg. photodiode signal prior to ignition

0.8 Relative signal

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 25

0.6 τ = 1424 µs

0.4

0.2

0.0 0

500

1000 Time [µs]

1500

Figure 2: Representative ignition delay time measurement. Shown are the sidewall pressure and CH radical chemiluminescence measurements with corresponding ignition delay time, τ , for a DMF/O2 /Ar mixture with φ = 1.0, argon/oxygen ratio, D = 3.76, p = 1.9 atm and T = 1258 K. investigated are shown in Table 1. Table 1: Mixtures investigated using constraints of φ and D

φ 1.0 2.0 0.5 * 1.0 * 1.0 2-MF 1.0 2.0 0.5 1.0 * 1.0 Furan 1.0 2.0 0.5 Fuel DMF

*

D 3.76 3.76 3.76 12.2 16.6 3.76 3.76 3.76 12.2 15.5 3.76 3.76 3.76

Fuel % 2.56 5.32 1.38 1.00 0.76 3.38 6.54 1.72 1.22 1.00 4.46 8.54 2.28

O2 % 20.47 19.90 20.72 7.49 5.63 20.28 19.63 20.65 7.48 6.00 20.07 19.21 20.53

Ar % 76.97 74.78 77.90 91.51 93.62 76.34 73.83 77.63 91.30 93.00 75.47 72.25 77.19

Mixtures used for validation experiments.

Results and discussion Experimental results are presented below. Simulations presented in this work are carried out using the CHEMKIN homogeneous constant volume reactor, except for simulations of rich mixtures 8 ACS Paragon Plus Environment

Page 9 of 25

4

10

DMF, this study, sidewall DMF, this study, endwall DMF, Sirjean et al.

3

10 τ [µs]

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

2

10

φ=1.0; D=12.2; p=4.0 atm 1

10

0.60

0.65

0.70 1000/T [1/K]

0.75

0.80

Figure 3: Comparison of ignition delay times for stoichiometric DMF/O2 /Ar at a pressure of 4.0 atm with published data by Sirjean et al. 1 Endwall and sidewall ignition delay times are plotted, indicating the longer endwall ignition delay times as discussed by Petersen. 22 The dashed line represents Model 1, 1 while the solid line represents Model 2. 2 at 12 atm, which have been determined using CANTERA, because of convergence problems in CHEMKIN. Ignition delay times of DMF and 2-MF are first compared to previous data. Results of the comparative study ignition study are then presented for stoichiometric mixtures at various pressures. Variations in the post-reflected shock pressures of individual experimental realizations from the nominal pressure are accounted for in the figures using a power law of the form τ ∝ pn . The exponent n is determined from linear regression of stoichiometric data with the same argon/oxygen ratios. The exponents are found to be −0.59 for furan, −0.83 for 2-MF, and −0.69 for furan. The unscaled ignition data are provided as supplementary information.

Experimental validation results The shock tube has been validated using literature data for n-heptane, iso-octane, and furans; comparisons of DMF and 2-MF are presented in this work including model predictions. New DMF ignition data under conditions similar to those of Sirjean et al. 1 are compared in Figure 3. The sidewall ignition measurements from this study are shorter than the literature data. The endwall data are in closer agreement to the literature data, which were obtained using endwall measurements. A representative temperature uncertainty of 25 K as discussed previously is also indicated


Energy & Fuels

in the figure. Sensor separations can be measured to 1 mm accuracy and time resolution to 1 µ s; these factors yield a shock velocity uncertainty of approximately 5 m/s at conditions studied, translating to approximately 15 K before other factors are considered. It should thus be noted from the trend of the experimental data that uncertainty is mostly systematic and applies to all data points. Ignition data are measured for stoichiometric DMF/O2 /Ar mixtures with D = 16.6, similar to mixtures studied by Somers et al. 2 The delay times are measured at pressures ranging from 1.5 atm to 5.7 atm and scaled to 1.0 atm for comparison in Figure 4. It is observed that the current dataset are in agreement with the literature data, which were also obtained using sidewall ignition detection. Also included in Figure 3 and Figure 4 are model predictions using two detailed chemical kinetic models - the Sirjean et al. model, 1 denoted as Model 1, and the Somers et al. 2 version denoted as Model 2. Model 1 has 1457 elementary reactions among 294 species and Model 2 has 2766 reactions among 545 species. In Figure 3, Model 1 predicts longer ignition delay times than Model 2, and Model 1 is in closer agreement with literature data used in its development while Model 2 is in better agreement with current sidewall measurements over the temperature range considered. In Figure 4, the deviation between the two model predictions is preserved but narrows 4

DMF, this study, scaled from 1.5−5.7 atm using p−0.685 DMF, Somers et al.

10

3

10 τ [µs]

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 25

2

10

φ=1.0; D=16.6; p=1.0 atm 1

10

0.55

0.60

0.65 0.70 1000/T [1/K]

0.75

0.80

Figure 4: Comparison of ignition delay times for stoichiometric DMF/O2 /Ar at a pressure of 1.0 atm and D = 16.6 with published data by Somers et al. 2 The current data are scaled from data obtained at pressures from 1.5 atm to 5.7 atm. Dashed lines: Model 1 by Sirjean et al., 1 solid lines: Model 2 by Somers et al. 2


Page 11 of 25

at higher temperatures, while remaining within the bounds of the experimental data. Further comparison is carried out for 2-MF. Ignition data are obtained for stoichiometric mixtures of 2-MF/O2 /Ar with D = 15.5, similar to conditions studied by Wei et al. 15 In Figure 5 ignition delay times at 10.65 atm are compared while Figure 6 shows data at 1.25 atm. It is observed that the current data are in good agreement with those reported by Wei et al. 15 It should also be noted that further 2-MF data obtained by Wei et al. 15 are in agreement with the results of Somers et al. 14 at a pressure of 1.0 atm. An independent chemical kinetic model for 2-MF was developed by Somers et al., 14 which is incorporated in the later expanded DMF model by Somers et al. 2 The latter has been used in this work to simulate both 2-MF and DMF ignition. Compared to 2-MF experimental data, model predictions using the model by Somers et al. 2 are in close agreement with the experimental data, while the 2-MF sub model in the DMF model by Sirjean et al. 1 predicts shorter ignition delay times at both pressures. The above comparisons show that ignition delay times obtained from the new facility are in agreement with other shock tube facilities.

This study Wei et al. 3

10

τ [µs]

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

2

10

2−MF; φ=1.0; D=15.5; p=10.65 atm

0.70

0.75

0.80 1000/T [1/K]

0.85

0.90

Figure 5: Comparison of ignition delay times for stoichiometric 2-MF/O2 /Ar at a pressure of 10.65 atm and D = 15.5 with published data by Wei et al. 15 Dashed lines: Model 1 by Sirjean et al., 1 solid lines: Model 2 by Somers et al. 2


Energy & Fuels

This study Wei et al. 3

10

τ [µs]

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 25

2

10

2−MF; φ=1.0; D=15.5; p=1.25 atm 0.65

0.70 0.75 1000/T [1/K]

0.80

Figure 6: Comparison of ignition delay times for stoichiometric 2-MF/O2 /Ar at a pressure of 1.25 atm and D = 15.5 with published data by Wei et al. 15 Dashed lines: Model 1 by Sirjean et al., 1 solid lines: Model 2 by Somers et al. 2

Comparative Ignition Study The ignition behavior of the three furans was investigated at stoichiometric conditions and nominal pressures of 2.0, 5.0, and 10 atm with argon/oxygen ratio maintained at 3.76. This is a much lower dilution level than the cases above but it is in line with engine combustion, whereby air is used as an oxidizer with a nitrogen/oxygen ratio, D, of 3.76. The ignition delay times are presented together with an Arrhenius fit for better delineation of trends. To further distinguish among datasets, Arrhenius fits have been added to all figures from Figure 7 to Figure 12. The fit parameters are derived from linear regression of each dataset in the figure. Figure 7 shows that at a nominal pressure of 2.0 atm, DMF has the longest ignition delay times, while 2-MF ignites fastest. Furan ignition delay times lie between DMF and 2-MF, with a possible cross-over between DMF and furan at temperatures above 1410 K. This trend is a key finding in our study which has not yet been captured by individual studies of these fuels. The relative ignition behavior of DMF and 2-MF also differs from the observation by Uygun et al., 16 where the authors assume that DMF and 2-MF have similar ignition delay times. They use this as a basis to scale and plot DMF data together with 2-MF. The pressure scaling used, τ ∝ p−1.4 , seems to have an exponent outside the range usually observed in hydrocarbon ignition, 23 and also observed in oxygenated hydrocarbons. 24,25 12 ACS Paragon Plus Environment

Page 13 of 25

2,5−DMF 2−MF Furan 3

τ [µs]

10

2

φ=1.0; D=3.76; p=2.0 atm

10

0.70

0.75

0.80

0.85

1000/T [1/K]

Figure 7: Relative ignition behavior of furan, 2-methyl furan (2-MF), and 2,5-dimethyl furan (DMF) at nominal pressure of 2.0 atm. The fuel/O2 /Ar mixture is stoichiometric and the Ar/O2 ratio, D, is maintained at 3.76. From a chemical perspective, it means that double alkylation of furan confers greater chemical stability while the mono alkylated 2-MF introduces higher reactivity. Although a number of quantum mechanical calculations have been carried out for each of these furans, 1,2,26 it is difficult to comment on the reactivity trends observed among these furans only based on these calculations. However, abnormally high reactivity has been observed for methoxyfuran, which Simmie et al. 27 attribute to the very weak methoxy (C–O) bond. The trend shown in Figure 7 is also observed at a higher pressure of 5.0 atm as shown in

2,5−DMF 2−MF Furan 3

10 τ [µs]

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

φ=1.0; D=3.76; p=5.0 atm 2

10

0.75

0.80

0.85 1000/T [1/K]

0.90

0.95

Figure 8: Relative ignition behavior of furan, 2-methyl furan (2-MF), and 2,5-dimethyl furan (DMF) at nominal pressure of 5.0 atm. The fuel/O2 /Ar mixture is stoichiometric and the Ar/O2 ratio, D, is maintained at 3.76. 13 ACS Paragon Plus Environment

Energy & Fuels

DMF 2−MF Furan 3

10 τ [µs]

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 25

φ=1.0; D=3.76; p=10 atm

2

10

0.80

0.85 0.90 1000/T [1/K]

0.95

Figure 9: Relative ignition behavior of furan, 2-methyl furan (2-MF), and 2,5-dimethyl furan (DMF) at nominal pressure of 10.0 atm. The fuel/O2 /Ar mixture is stoichiometric and the Ar/O2 ratio, D, is maintained at 3.76. Figure 8. Increasing the pressure to 10 atm as in Figure 9 preserves the established reactivity trend. However, Figure 9 shows that the reactivity difference between DMF and furan is not as significant as at the other two conditions, while 2-MF is consistently the most reactive. The reactivity trends observed at stoichiometric conditions are also observed at rich conditions of φ = 2.0 and pressures of 5.0 atm and 12 atm. At lean conditions of φ = 0.5, 2-MF still has the shortest ignition delay times and DMF has the longest while the ignition delay times of furan are closer to those of DMF. Other studies on reactivity trends and equivalence ratio effects which focus on keeping fuel mole fractions constant are less well suited to reveal these trends which are of interest to engine research and development. For Spark-Ignition engines, for instance, combustion is preferably organized around stoichiometric fuel/air mixture conditions; thus fuel comparisons at similar equivalence ratios and dilution ratios reveal trends which can be expected in practical devices. These conclusions can also be tested and compared with investigations of reactivity trends based on laminar burning velocities of freely propagating fuel/air mixtures which constrain the equivalence and the diluent/oxygen ratios.


Page 15 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Equivalence ratio effect The effect of equivalence ratio on ignition behavior of the three furans is explored. Experiments for the lean and rich mixtures are conducted at nominal pressures of 12 atm, and the stoichiometric data at 10 atm are scaled to 12 atm using a pressure scaling with exponents as discussed previously. Figure 10 shows the equivalence ratio effect on 2-MF ignition; it is observed that ignition delay times decrease with increasing equivalence ratios over the investigated temperature range. This means that constraining the ratio of the number of inert molecules to oxygen molecules and increasing fuel concentration results in higher reactivity. It should be noted that in other studies, for instances, Somers et al., 2 equivalence ratio effects are investigated by keeping the fuel concentration constant. This leads to the opposite observation that lean mixtures are more reactive. At play in such a scenario is actually the higher oxygen concentration (lower dilution) in the lean mixtures. Further, although differences are observed in our approach, they are usually within 80% of the stoichiometric delay time, making the stoichiometric ignition delay times at given pressure, dilution, and temperature, a very useful reference to estimate ignition delay times at other equivalence ratios. When fuel/air comparative studies are conducted, the results are then similar to our observations here. In most hydrocarbon/air ignition at high temperatures, the ignition delay times decrease with increasing equivalence ratio, as in Ciezki and Adomeit, 28 due to increased oxygen-linked reactivity through peroxy chemistry. The same equivalence ratio effect for 2-MF is observed for furan, as shown in Figure 11. Ignition delay times reduce with increasing equivalence ratio. In Figure 12 the equivalence ratio effect on DMF ignition is less distinct for stoichiometric and rich mixtures. Rich mixtures ignite more readily than stoichiometric at high temperatures while at temperatures less than 1130 K the rich mixtures become less reactive.

Comparison of 2-MF and DMF data with model predictions The experimental results on the ignition of the three furans are compared to predictions of recently published chemical kinetic models by Sirjean et al. 1 and Somers et al. 2 These models focus on 15 ACS Paragon Plus Environment

Energy & Fuels

φ = 1.0 φ = 2.0 φ = 0.5 3

τ [µs]

10

2

10

2−MF; D=3.76; p=12 atm 0.75

0.80

0.85

0.90 0.95 1000/T [1/K]

1.00

1.05

Figure 10: Equivalence ratio effect on 2-MF ignition delay times for 2-MF/O2 /Ar mixtures at nominal pressures of 12.0 atm. Ignition delay times decrease with increasing equivalence ratio over this temperature range. DMF but also have 2-MF and furan as sub models. The furan sub models have not been tested against experiments and their predictions are not included in this discussion. In Figure 13 ignition delay times of the three furans at stoichiometric conditions and nominal pressure of 2.0 atm are compared with model predictions. The figure reveals that Model 2 2 captures the ignition behavior of DMF with reasonable accuracy, while Model 1 1 over-predicts the ignition delay times. With regard to 2-MF, both models predict shorter ignition delay times than the actual measurements. φ = 1.0 φ = 2.0 φ = 0.5 3

10 τ [µs]

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 25

2

10

Furan; D=3.76; p=12 atm 0.75

0.80

0.85 0.90 1000/T [1/K]

0.95

1.00

Figure 11: Equivalence ratio effect on furan ignition delay times for furan/O2 /Ar mixtures at nominal pressures of 12.0 atm. Similar to 2-MF, ignition delay time decreases with increasing equivalence ratio.


Page 17 of 25

φ = 1.0 φ = 2.0 φ = 0.5 3

τ [µs]

10

2

10

DMF; D=3.76; p=12 atm 0.75

0.80

0.85 0.90 1000/T [1/K]

0.95

1.00

Figure 12: Equivalence ratio effect on DMF ignition delay times for DMF/O2 /Ar mixtures at nominal pressures of 12.0 atm. The ignition delay times at 10 atm in Figure 14 represents a shift in the direction of lower temperatures with respect to Figure 13. Predictions of both Model 1 1 and Model 2 2 are longer than measured delay times. Because of the lower temperatures and the temperature-sensitivity displayed by the 2-MF predictions in Figure 13, the models predict longer delay times in Figure 14, with the Model 2 in closer agreement at the higher temperature end. Rich mixtures (φ = 2.0) of DMF and 2-MF are compared at a pressure of 12 atm with model predictions in Figure 15. Model 2 2 is in better agreement with the experimental data, while Model

DMF 2−MF

3

10 τ [µs]

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

2

10

φ=1.0; D=3.76; p=2.0 atm 0.70

0.75

0.80 1000/T [1/K]

0.85

Figure 13: Experimental and model predictions of ignition delay times for stoichiometric mixtures of DMF, 2-MF and furan at nominal pressures of 2.0 atm. Dashed lines: Model 1 by Sirjean et al., 1 solid lines: Model 2 by Somers et al. 2


Energy & Fuels

DMF 2−MF

τ [µs]

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 25

3

10

φ=1.0; D=3.76; p=10.0 atm 2

10

0.80

0.85 0.90 1000/T [1/K]

0.95

Figure 14: Experimental and model predictions of ignition delay times for stoichiometric mixtures of DMF, 2-MF and furan at nominal pressures of 10.0 atm. Dashed lines: Model 1 by Sirjean et al., 1 solid lines: Model 2 by Somers et al. 2 1 1 predicts longer ignition delay times for both DMF and 2-MF. Both models accurately predict that DMF has longer ignition delay times than 2-MF. However, when the furan sub models in the comprehensive models are used, the ignition predictions are much longer than those measured. Compared to the other predictions, both models predict a reactivity trend, such that furan is the least reactive and 2-MF the most reactive. This is at variance with the experiment, where DMF is consistently the least reactive under all investigated conditions. For models to capture the trends observed in this study, further improvement of the furan sub models and overall improvement of the temperature sensitivity over the wider temperature range is needed. From the comparison above, it is observed that the models predict ignition delay times to a varying degree of success. The current dataset is obtained at high-temperature conditions. Further experiments at lower temperatures are needed to verify if these conclusions hold consistently over a wider temperature range. A key open question is whether these fuels exhibit strong Negative Temperature Coefficient (NTC) at lower temperatures. The presence of NTC behavior would explain the non-Arrhenius behavior observed at lower temperatures as indicated in lean mixtures in Figure 10. Although good agreement between model predictions and experimental data is not observed at all conditions, the models capture the reactivity trend between 2-MF and DMF. Sensitivity analyses 18 ACS Paragon Plus Environment

Page 19 of 25

4

10

τ [µs]

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

DMF 2−MF

3

10

2

10

φ=2.0; D=3.76; p=12.0 atm 0.70

0.75

0.80

0.85 0.90 1000/T [1/K]

0.95

1.00

Figure 15: Experimental and model predictions of ignition delay times for rich mixtures of DMF, 2-MF and furan at nominal pressures of 12.0 atm. Dashed lines: Model 1 by Sirjean et al., 1 solid lines: Model 2 by Somers et al. 2 are carried out and results of reaction pathway analyses are discussed as an attempt to understand differences in the DMF and 2-MF models. Sensitivity analyses of the two models based on ignition of stoichiometric fuel/O2 /Ar mixtures with D = 3.76 at 1150 K and 10 atm. A-factors of the reaction rate constants are multiplied by 10 to establish their effect on ignition delay times. The logarithmic sensitivities of the elementary reactions are determined and sorted in order of importance. The 16 most important reactions are plotted below. As expected, most of the reactions are C0 - C1 oxidation reactions, with some key fuel specific reactions. The fuel specific-reactions include decomposition, H-abstraction reactions from the fuel, and further reaction of primary fuel radicals. The sensitivity analyses for DMF (appears as C6 H8 O) are shown in Figure 16 for the Somers et al. model or Model 2 2 and in Figure 17 for Sirjean et al. model or Model 1. 1 It is observed that in both models H + O2 ⇀ ↽O + OH is highly sensitive as expected. It is less sensitive in Model 1. 1 The reaction C6 H8 O + O2 ⇀ ↽C6 H7 O + HO2 is more sensitive in Model 2 than in Model 1. These factors contribute to the observed reduced global reactivity of Model 1. The reaction C6 H7 O + O2 ⇀ ↽C6 H6 O + HO2 has a very high sensitivity in Model 1 while it does not appear among the 16 most sensitive reactions in Model 2. The sensitivity analyses for 2-MF are shown in Figure 18 for Model 2 and in Figure 19 19 ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 25

C6H8O + O2 ⇔ C6H7O + HO2

H + O2 ⇔ O + OH

CH3 + C6H7O ⇔ C7H10O

CH3 + HO2 ⇔ CH4 + O2

h45de2o3j ⇔ h35de2o1j CH3 + O2 ⇔ CH2O + OH

C6H7O + CH3O2 ⇔ C6H7O2 + CH3O C6H8O + CH3 ⇔ C6H7O + CH4

C6H8O + H ⇔ C4H6 + CH3CO

che21o4j ⇔ chde241o + H C6H8O (+ M) ⇔ C6H7O + H (+ M) HO2 + HO2 ⇔ H2O2 + O2

H + O2 (+ M) ⇔ HO2 (+ M)

C6H8O + HO2 ⇔ C6H7O + H2O2

CH3 + HO2 ⇔ CH3O + OH che21o4j ⇔ che21o5j −0.4

−0.3

−0.2

−0.1 0 0.1 L.S. = (∆τ/τo)/(∆ kj/kj)

0.2

0.3

Figure 16: The 16 most important reactions from the sensitivity analysis of the DMF mechanism by Somers 2 for a stoichiometric DMF/O2 /Ar mixture at 10 atm with a dilution, D of 3.76 at a temperature of 1150 K. The unperturbed ignition delay time is 1089 µ s.

H + O2 ⇔ O + OH

C6H7O + O2 ⇔ C6H6O + HO2

C6H8O + O2 ⇔ C6H7O + HO2

CH3 + CH3 (+ M) ⇔ C2H6 (+ M) CH3 + HO2 ⇔ CH4 + O2

C6H7O + H (+ M) ⇔ C6H8O (+ M) CH3 + O2 ⇔ HCHO + OH

H + O2 (+ M) ⇔ HO2 (+ M)

CH3OO + CH3 ⇔ CH3O + CH3O

C6H8O + OH ⇔ CH3CO + CH3CHO + C2H2 C6H7O ⇔ C6H6O + H

C4H6 + CH3 ⇔ i−C4H5 + CH4

HCHO + CH3 ⇔ CHO + CH4 R4C6H7O ⇔ R6C6H7O

M5F−2yl + O2 ⇔ M5F−2ylO + O

C6H8O + CH3 ⇔ FurylCH2 + CH4

−0.4

−0.3

−0.2

−0.1 0 0.1 L.S. = (∆τ/τo)/(∆ kj/kj)

0.2

0.3

Figure 17: The 16 most important reactions from the sensitivity analysis of the DMF mechanism by Sirjean 1 for a stoichiometric DMF/O2 /Ar mixture at 10 atm with a dilution, D of 3.76 at a temperature of 1150 K. The unperturbed ignition delay time is 1519 µ s.


Page 21 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

p34de1o2j−c2 ⇔ p14de1o3j−c1 H + O2 ⇔ O + OH C5H5O + CH3 ⇔ C6H8O

C5H6O + CH3 ⇔ C5H5O + CH4

C5H6O + H ⇔ C5H5O + H2

CH3 + HO2 ⇔ CH3O + OH

C5H6O + OH ⇔ C5H5O + H2O

C5H5O + HO2 ⇔ C5H5O2 + OH C5H6O + O2 ⇔ C5H5O + HO2

C2H3 + O2 ⇔ CH2O + HCO

C5H4O2 + CH3 ⇔ C5H3O2 + CH4

C5H6O + HO2 ⇔ C5H5O + H2O2

C2H3 + O2 ⇔ CH2CHO + O C5H6O + H ⇔ C5H7O

n−C4H5 + O2 ⇔ CH2CHCHCHO + O n−C4H5 ⇔ C4H4 + H

−0.4

−0.3

−0.2

−0.1 0 0.1 L.S. = (∆τ/τo)/(∆ kj/kj)

0.2

0.3

Figure 18: The 16 most important reactions from the sensitivity analysis of the 2-MF mechanism by Somers 2 for a stoichiometric 2-MF/O2 /Ar mixture at 10 atm with a dilution, D of 3.76 at a temperature of 1150 K. The unperturbed ignition delay time is 1089 µ s. for Model 1. It is observed that the main branching reaction H + O2 ⇀ ↽ O + OH is very sensitive in Model 2, while its sensitivity is reduced in Model 1. Also, the propagation reaction CH2 CCHCHCHO⇀ ↽CH2 CHCHCHCO is the most sensitive reaction in Model 2, while it is not included in Model 1. On the other hand, the H-abstraction reaction C5 H6 O + HO2 ⇀ ↽C5 H5 O + H2 O2 is the most sensitive in Model 1 while its shows reduced sensitivity in Model 2. The isomerization reaction C5 H6 O⇀ ↽CH3 COCHCCH2 is very sensitive in Model 1 while it is not among the 16 most sensitive reactions in Model 2. As expected, reactions which favor the formation of stable molecules tend to reduce reactivity and increase the ignition delay times in all cases. Reaction pathway analysis are performed for an ignition process of stoichiometric fuel/O2 /Ar mixtures with D = 3.76 at a temperature of 1150 K and pressure of 10 atm. For DMF, the system is analyzed at 10, 500 µ s, and close to ignition. Fuel consumption was found to initially proceed mainly through H-abstraction reactions by O2 , with contributions from CH3 and H radicals. In contrast to the model by Somers et al., 2 analysis of the model by Sirjean et al. 1 shows that 27.5% of (2,5)-DMF consumption at 10 µ s proceeded through isomerization to 2,4-DMF. This partly ex21 ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 25

C5H6O + HO2 ⇔ FurylCH2 + H2O2

FurylCH2 + HO2 ⇔ FurylCH2O + OH C5H6O ⇔ CH3COCHCCH2

C5H6O + OH ⇔ M5F−2yl + H2O

H2O2 (+ M) ⇔ OH + OH (+ M) HO2 + HO2 ⇔ H2O2 + O2

H + O2 ⇔ O + OH

HO2 + OH ⇔ H2O + O2

OCHCHCHCCH2 ⇔ OCCHCHCHCH2 n−C4H5 ⇔ C2H2 + C2H3

HO2 + HO2 ⇔ H2O2 + O2

CH3 + HO2 ⇔ CH4 + O2

M5F−2yl + O2 ⇔ M5F−2ylO + O FurylCH2 + H ⇔ C5H6O

C2H3CHO + HO2 ⇔ CH2CHCO + H2O2

C2H3 + O2 ⇔ HCHO + CHO −0.4

−0.3

−0.2

−0.1 0 0.1 L.S. = (∆τ/τo)/(∆ kj/kj)

0.2

0.3

Figure 19: The 16 most important reactions from the sensitivity analysis of the DMF mechanism by Sirjean 1 for a stoichiometric 2-MF/O2 /Ar mixture at 10 atm with a dilution, D of 3.76 at a temperature of 1150 K. The unperturbed ignition delay time is 481 µ s. plains the reduced reactivity of the model, since this stable fuel molecule still needs to be attacked by radicals before ignition. Further, fuel-derived radicals are produced more slowly as compared to Model 2. At later times, fuel consumption proceeds by H-abstraction by OH, CH3 , and H. Whereas in Model 2, the abstraction is mostly by OH and H radicals, in Model 1 the abstracting partner is mostly CH3 . This suggests that Model 1 focuses more on a pyrolytic fuel consumption pathway, with oxygen-containing radicals becoming important only close to ignition. Fuel radicals are consumed subsequently by beta-scission and radical abstraction reactions. In the case of 2-MF, ignition delay times are 481 µ s for Model 1 and pathways are analyzed at 10, 200, and 450 µ s while the delay time is 347 µ s for Model 2 and reaction pathways examined at 10, 150, and 300 µ s. Similar to DMF, in both models fuel is initially consumed through reactions with O2 to yield HO2 and primary fuel radicals. For Model 1, direct ring-opening of MF to CH3 COCHCCH2 radical. CH3 COCHCCH2 decomposes to release CH3 which is a major abstraction partner unlike in Model 2, where abstraction is mostly by OH radicals. In summary, reaction pathway analyses suggest that Model 1 is more focused on pyrolytic reactions, which yield 22 ACS Paragon Plus Environment

Page 23 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

other stable molecules, thereby retarding reactivity. In addition to these reaction pathway and sensitivity analyses, thermodynamic properties could also partially account for observed differences. Model 2 is an improved version of Model 1 but further work is needed to improve the prediction of experimental data reported here as well as flame and flow reactor datasets in the literature.

Conclusion This study systematically investigates the ignition of a class of oxygenated cyclic hydrocarbons, furans, which are of interest as alternative transportation fuels or fuel additives. Important trends are revealed and the results are compared with model predictions, revealing some weaknesses in model performance. Ignition delay times obtained at conditions of previous studies agree with those datasets and model predictions at conditions where the models have been tested. With regards to structure-reactivity trends, the results show a non-monotonic trend with respect to chemical structure, whereby DMF is the least reactive while 2-MF is the most readily ignitable. Two chemical models also predict 2-MF to be more reactive than DMF, although quantitative agreement varies over the range of conditions investigated. Equivalence ratio effects for each furan are also investigated. Generally, ignition delay times decrease with increasing equivalence ratios, except for DMF whose rich mixtures show reduced reactivity at lower temperatures. Although individual studies of these fuels have been reported in the literature, the main contribution of this work is to reveal trends which cannot be gleaned from individual experimental, modeling, and quantum chemical calculations. The dataset will be useful for further understanding and modeling of furan combustion.

Acknowledgement Acknowledgment is made to the Donors of the American Chemical Society Petroleum Research Fund, for support of this research.


Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

References (1) Sirjean, B.; Fournet, R.; Glaude, P.-A.; Battin-Leclerc, F.; Wang, W.; Oehlschlaeger, M. A. J. Phys. Chem. A 2013, 117, 1371–1392. (2) Somers, K. P. et al. Combust. Flame 2013, 160, 2291–2318. (3) Román-Leshkov, Y.; Barrett, C. J.; Liu, Z. Y.; Dumesic, J. A. Nature 2007, 447, 982–985. (4) Tong, X.; Ma, Y.; Li, Y. Appl. Catal., A 2010, 385, 1–13. (5) Chidambaram, M.; Bell, A. T. Green Chem. 2010, 12, 1253–1262. (6) Mascal, M.; Nikitin, E. Angew. Chem. 2008, 120, 8042–8044. (7) Zhong, S.; Daniel, R.; Xu, H.; Zhang, J.; Turner, D.; Wyszynski, M. L.; Richards, P. Energy Fuels 2010, 24, 2891–2899. (8) Daniel, R.; Wei, L.; Xu, H.; Wang, C.; Wyszynski, M. L.; Shuai, S. Energy Fuels 2012, 26, 6661–6668. (9) Rothamer, D. A.; Jennings, J. H. Fuel 2012, 98, 203–212. (10) Thewes, M.; Muether, M.; Pischinger, S.; Budde, M.; Brunn, A.; Sehr, A.; Adomeit, P.; Klankermayer, J. Energy Fuels 2011, 25, 5549–5561. (11) Gouli, S.; Lois, E.; Stournas, S. Energy Fuels 1998, 12, 918–924. (12) Christensen, E.; Yanowitz, J.; Ratcliff, M.; McCormick, R. L. Energy Fuels 2011, 25, 4723– 4733. (13) Cheng, Z.; Xing, L.; Zeng, M.; Dong, W.; Zhang, F.; Qi, F.; Li, Y. Combust. Flame 2014, –. (14) Somers, K.; Simmie, J.; Gillespie, F.; Burke, U.; Connolly, J.; Metcalfe, W.; BattinLeclerc, F.; Dirrenberger, P.; Herbinet, O.; Glaude, P.-A.; Curran, H. Proc. Combust. Inst. 2013, 34, 225–232. 24 ACS Paragon Plus Environment

Page 24 of 25

Page 25 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

(15) Wei, L.; Tang, C.; Man, X.; Huang, Z. Energy Fuels 2013, 27, 7809–7816. (16) Uygun, Y.; Ishihara, S.; Olivier, H. Combust. Flame 2014, –. (17) Wei, L.; Tang, C.; Man, X.; Jiang, X.; Huang, Z. Energy Fuels 2012, 26, 2075–2081. (18) Tian, Z.; Yuan, T.; Fournet, R.; Glaude, P.-A.; Sirjean, B.; Battin-Leclerc, F.; Zhang, K.; Qi, F. Combust. Flame 2011, 158, 756–773. (19) Davis, A. C.; Sarathy, S. M. J. Phys. Chem. A 2013, 117, 7670–7685. (20) Goodwin, D. G. An open-source, extensible software suite for CVD process simulation. Chemical Vapor Deposition XVI and EUROCVD 14, ECS Proc., 2003; pp 155–162. (21) Browne, S.; Ziegler, J.; Shepherd, E. Numerical solution methods for shock and detonation jump conditions; 2008. (22) Petersen, E. L. Combust. Sci. Technol. 2009, 181, 1123–1144. (23) Davidson, D. F.; Hanson, R. K. Int. J. Chem. Kinet. 2004, 39, 510–523. (24) Noorani, K.; Akih-Kumgeh, B.; Bergthorson, J. Energy Fuels 2010, 24, 5834–5843. (25) Akih-Kumgeh, B.; Bergthorson, J. Combust. Flame 2011, 158, 1037–1048. (26) Somers, K. P.; Simmie, J. M.; Curran, H.; Metcalfe, W. K. Physical Chemistry Chemical Physics 2014. (27) Simmie, J. M.; Somers, K. P.; Yasunaga, K.; Curran, H. J. International Journal of Chemical Kinetics 2013, 45, 531–541. (28) H. K. Ciezki, G. Adomeit, Combust. Flame 1993, 26, 421–433.

Supporting Information Available This material is available free of charge via the Internet at http://pubs.acs.org/. 25 ACS Paragon Plus Environment

Reactivity Trends in Furan and Alkyl Furan ... - ACS Publications

Recommend Documents