Reaction Pathway Analysis of Ethyl Levulinate and 5

Sep 22, 2015 - Department of Chemical and Environmental Sciences, University of Limerick, Limerick, Ireland. Energy Fuels , 2015 .... Dong-Shik Kim. R...
16 downloads 11 Views 632KB Size
Subscriber access provided by UNIV OF LETHBRIDGE

Article

Reaction Pathway Analysis of Ethyl levulinate and 5Ethoxymethylfurfural from D-Fructose Acid Hydrolysis in Ethanol Thomas Flannelly, Stephen Dooley, and James J. Leahy Energy Fuels, Just Accepted Manuscript • Publication Date (Web): 22 Sep 2015 Downloaded from http://pubs.acs.org on October 9, 2015

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 36

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

1 2

Reaction Pathway Analysis of Ethyl levulinate and 5-Ethoxymethylfurfural from DFructose Acid Hydrolysis in Ethanol

3

Thomas Flannelly, Stephen Dooley, J.J Leahy

4 5 6

Department of Chemical and Environmental Sciences, University of Limerick, Ireland.

7 8

Abstract

9

This study utilises numerical modelling to provide a mechanistic discussion of the synthesis

10

of the advanced biofuel candidates, ethyl levulinate and 5-ethoxymethylfurfural, from α/β-D-

11

fructopyranose (D-fructose) in a condensed phase homogeneous ethanol system at 351 K

12

catalysed by hydrogen cations. A mechanistic comprehension is pursued by detailed

13

measurements of reactant, intermediate and product species temporal evolutions, as a

14

function of H2SO4 (0.09, 0.22, 0.32 mol/L) and

15

concentration, also considering the addition of water to the ethanol media (0, 12, 24 mass %

16

water in ethanol).

17

levulinate, and several other intermediate species are quantified as major species fractions at

18

45-85 % of the initial D-fructose mass. To inform the mechanistic discussion mass-conserved

19

chemically authentic kinetic models and empirical rate constants are derived each assuming a

20

first order relationship to the hydrogen cation concentration. The optimal synthesised

21

fractions of ethyl levulinate and 5-ethoxymethylfurfural considered as fuel components,

22

achieve a mass yield of 63 % with respect to the fructose mass and a volumetric energy

23

valorisation (∆HCombustion, kcal/mL) of 215 % with respect to the ethanol consumed, indicating

24

the viability of the synthesis.

D-fructose,

D-fructose

(0.14, 0.29, 0.43 mol/L)

5-hydroxymethylfurfural, 5-ethoxymethylfurfural, ethyl

25 26

Keywords. D-fructose, 5-ethoxymethylfurfural, ethyl levulinate, reaction mechanism, kinetic

27

model.

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

Page 2 of 36

28

1. Introduction

29

Presently there is a growing effort to find renewable and sustainable alternatives to petroleum

30

derived bulk chemicals and fuels. The catalytic conversion of biomass derived cellulose and

31

hemicellulose to platform chemicals has been widely recognised as an opportunity to develop

32

a carbohydrate based chemical industry.1-2 Lignocellulosic derived hexose and pentose sugars

33

have potential as a sustainable alternative to the carbohydrates derived from edible crop

34

matter. The United States Department of Energy have identified promising renewable

35

chemical building blocks that may be produced from such biomass derived sugars.3 Ethyl

36

levulinate and 5-ethoxymethylfurfural, Figure 1, are two such promising furan derived

37

chemicals for potential use as transportation fuels.

38 39

Ethyl levulinate has received a significant amount of attention purporting its potential use as a

40

fuel.4-5 Chemical systems reporting its synthesis include; ethylation of levulinic acid, furfuryl

41

alcohol and 5-chloromethylfurfural.6–12 Ethyl levulinate production from biomasses have

42

been reported13–15 however; yields have been modest at 40-50 %. There have also been

43

considerable difficulties in its synthesis from glucose and cellulose with 44.8 mol % the

44

highest yield reported to date using glucose as a starting material in an ethanol/H2SO4

45

system.16

46 47

5-ethoxymethylfurfural has received more limited suggestions as a fuel component17 despite

48

its high volumetric density of 1.099 g/mL (298 K). Gruter and Dautzenberg18 suggest an

49

enthalpy of formation of 120.1 kcal/mol, corresponding to an enthalpy of combustion of 7.87

50

kcal/g (see Table 1). These terms correspond to a volumetric energy density of 8.66 kcal/mL

51

(36.24 MJ/L at 298 K), thus being advantageous over those of other oxygenated fuel

52

components, such as ethanol (7.11 kcal/mL), and ethyl levulinate (7.53 kcal/mL). Indeed the

2 ACS Paragon Plus Environment

Page 3 of 36

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

53

value is comparable to petroleum derived diesel and gasoline (8.53 kcal/mL for the gasoline

54

primary reference fuel, iso-octane). Mascal and Nikitin19 use 5-chloromethylfurfural to

55

convert 5-hydroxymethylfurfural to 5-ethoxymethylfurfural. Other reports use aluminium

56

chloride as a catalyst to transform glucose to 5-ethoxymethylfurfural in an ethanol/water

57

medium20 and also from using a mixture of Sn-Beta and amberlyst catalysts.21

58 59

Like ethyl levulinate, the production of 5-ethoxymethylfurfural from glucose or other

60

cellulosic sugars presents challenges. The desired etherification of glucose is suppressed by

61

side reactions of various polymerisations and acetalizations producing the recalcitrant humic

62

substances.22,23 40-50 % of lignocellulosic biomass is made-up of cellulosic glucose

63

polymers, accounting for the largest proportion of hexoses that may be obtained from

64

biomass.24 As such glucose-like hexose sugars are cheaper and more readily available than

65

pentose sugars.

66 67

It is well understood that the steric and electronic configurations of the hydroxyl groups of

68

sugars, significantly affect yields of esterification products from monosaccharides.25 As a

69

consequence, fructose is much easier to convert into furan related products than glucose. The

70

desired isomerisation from glucose is known to be facilitated by an aqueous/organic media

71

where the equilibrium population have high proportions of labile α and β‐fructofuranose

72

structures.26 There is a general consensus that in order for glucose to be efficiently converted

73

into furanic derivatives, it must initially tautomerize into fructose species.24,27 There are

74

several recent reports describing reaction conditions that promote glucose isomerisation to

75

fructose in chemical media.11,28–30 For example, Despax et al.30 reported on the use of

76

heterogeneous catalysts in organic solvent mixtures showing ~ 68 % conversion of glucose to

77

fructose. The successful isomerisation of glucose to fructose in an alcohol medium by the

3 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 4 of 36

78

deliberate formation of methyl fructoside species as intermediates between glucose

79

conversions to fructose is of particular potential significance for the production of ethyl

80

levulinate and 5-ethoxymethylfurfural. Saravanamurugan et al.30,32

81

conversion from glucose to fructose by a one hour reaction in methanol at 120 C using an H-

82

USY zeolite. This suggests that an alcohol may be employed as both a solvent and alkylating

83

agent simultaneously, whilst also facilitating the required isomerization of glucose to

84

fructose. In this way, ethyl levulinate and 5-ethoxymethylfurfural may be synthesised

85

directly, rather than relying on the intermediary ethylation of the levulinic acid produced in

86

an aqueous system.

achieved 55 %

87 88

In addition to this sugar inter-conversion, the subsequent mechanism of fructose consumption

89

is the obvious further limiting step in achieving viable yields of furanic derivatives.33 Given

90

this prevailing position in the literature an improved mechanistic understanding of ethyl

91

levulinate and 5-ethoxymethylfurfural synthesis from fructose, as well as from glucose is

92

sought. There are recent reports of kinetic studies conducted on the purported H+ (hydrogen

93

cation) homogeneously catalysed dehydration of D-fructose to 5-hydroxymethylfurfural in

94

water.34–36 However, little is known of the analogous D-fructose dehydration in the presence

95

of ethanol and H+. Plausible reaction pathways have been suggested,27,37 but no quantitative

96

kinetic data for 5-ethoxymethylfurfural and ethyl levulinate synthesis have been reported.

97 98

We pursue an improved mechanistic comprehension employing a hierarchical modelling

99

approach that studies one sugar sub-mechanism at a time. In this context, we study the bottle-

100

neck α/β-D-fructopyranose (D-fructose) sub-mechanism initially, which once understood

101

would allow the more complex D-glucose and cellulose sub models to be developed in a

102

hierarchical manner. In order to develop realistic reaction kinetics, it is necessarily to limit

4 ACS Paragon Plus Environment

Page 5 of 36

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

103

the modelling complexity. To do so, we choose a homogeneous catalytic system of α/β-D-

104

fructose/H2SO4/ethanol, thereby minimising the mass transfer complexities of multiphase

105

heterogeneous reactions. The aim is to establish the main mechanistic relationships between

106

reactant, intermediate and intended product species such as to inform mechanistic discussion

107

and to test the validity of a viable reaction mechanism by the derivation of empirical rate

108

constants. By so doing the viability of preferentially producing one proposed fuel component

109

over the other, or to what extent this is achievable may be determined in a rigorous scientific

110

manner.

111 112 113

2. Experimental Materials

114

Ethanol, normal-octanol, acetone, (99 % purity), α/β-D-fructopyranose (CAS 57-48-7, 99%

115

purity), α/β-D-glucopyranose (CAS 50-99-7, 99 % purity) α/β-D-mannopyranose, (3458-28-4,

116

99 % purity), hence forth “D-fructose” “D-glucose” and “D-mannose” respectively, sulphuric

117

acid (H2SO4, 95-97% purity), 5-hydroxymethylfurfural (CAS 67-47-0, 99 % purity) furfural

118

(CAS 98-08-1, 98 % purity),and ethyl levulinate (CAS 539-88-8, 97 % purity) are each

119

obtained from Sigma Aldrich Ireland. Ethyl-α-D-glucopyranoside (CAS 34625-23-5, 98 %

120

purity) is obtained from Carbosyth Ltd. UK, and 5-ethoxymethylfurfural (CAS 1917-65-3,

121

96-97 % purity) is purchased from Akos Organics Gmbh, Germany.

122 123

Experimental Design and Procedure

124

Reactions are performed in a 20 cm3 spherical reactor at isothermal conditions of 351 ±1 K at

125

atmospheric pressure. The reactor is heated by an external oil bath. The reaction temperature

126

is independently controlled and monitored by a thermocouple array (Stuart™ SCT1

127

temperature controller) and an in-situ magnetic propeller ensures that the reaction mixture (D-

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

Page 6 of 36

128

fructose/H2SO4/ethanol) is well mixed and homogeneous. Atmospheric pressure is regulated

129

by fitting the main reactor exit with an open-ended condensing unit (~ 277 K), thus allowing

130

the reaction to be at reflux. For the test conditions reported here, Table 2, a heating time of 16

131

minutes is required for the reacting mixture to be heated from ambient to the prescribed

132

reaction temperature of 351 ±1 K. Reaction conditions are selected (Table 2) to parameterise

133

the influence of [H2SO4] and [D-fructose] on the reaction mechanism, whilst also considering

134

three scenarios of ethanol/water as reaction media. Reaction progress is monitored by

135

removing and analysing a 50 mg sample of the bulk reaction (0.104 g of D-fructose in 15.78 g

136

of ethanol) every hour for 480 minutes, resulting in a small cumulative perturbation to the

137

overall system mass. Control tests at the most severe conditions of Table 2, replacing D-

138

fructose with ethyl levulinate, show ethyl levulinate degradation to be within the estimated

139

experimental uncertainty, indicating it as a stable end-product. Control reactions are also

140

performed substituting 5-hydroxymethylfurfural and 5-ethoxymethylfurfural as starting

141

materials for the purposes of identifying the origins of various intermediate species, as

142

elaborated later.

143 144

Analytical Methods

145

The concentrations of ethyl levulinate and 5-ethoxymethylfurfural, are analysed by gas

146

chromatography (GC, Agilient Technologies 7820 A GC system) fitted with a Restek

147

Stabilwax capillary column (30 m, 0.25 mm ID, 0.25 µm), employing hydrogen carrier gas

148

and a flame ionisation detector. Species are identified by matching retention-times to known

149

standards, and quantified by calibration of detector response to known concentrations (using

150

n-octanol as internal standard). The injection port is maintained at 523 K, a temperature

151

sufficiently high to ensure the full vaporisation of the expected reaction components. A

152

temperature program of 40 K increasing to 493 K at a rate of 20 K per minute, remaining

6 ACS Paragon Plus Environment

Page 7 of 36

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

153

isothermal at 493 K for 5 minutes is found to achieve adequate separation of these species

154

from the ethanol/water media. GC-MS analysis is also employed for the identification of

155

sample species using an Agilient 5975C MSD, which uses a HP-5MS column (30 m, 0.25

156

mm ID, 0.25 µm) otherwise employing the same variables as for GC–FID analysis. For GC

157

analysis, a known mass (50 ± 5 mg ) of analyte is extracted from the reaction media into 0.4

158

g of room temperature acetone and 0.8 g of 0.16 mg/g n-octanol in acetone, this is followed

159

by the neutralisation of any remaining acid by the addition of 50 mg of NaHCO3. This

160

dilution and cooling procedure ensures that the chemical reaction is effectively quenched.

161

This sample is then filtered through 13 mm thick, 2 µm pore size syringe filters (Acrodisc) to

162

remove any insoluble humic substances that may have been formed, and 1µl of the resulting

163

solution is injected into the sample inlet port of the GC.

164 165

Identification and quantification of D-fructose, D-glucose, 5-hydroxymethylfurfural and the

166

various sugar-type derivatives is performed on an ion exchange liquid chromatography

167

system (IC) system (Dionex Corp., Sunnydale, CA) equipped with a pulsed amperometric

168

detector (AS, 10 µL sample loop, Dionex Corp., Sunnydale, CA). Analysis is performed at

169

291 K by isocratic elution with deionised water (18.2 MΩ.cm at a flow rate of 1.1 mL/min)

170

using a Dionex CarboPac PA1 carbohydrate column. The column is reconditioned using a

171

mixture of 0.4 mol/L sodium hydroxide and 0.24 mol/L sodium acetate after each analysis. A

172

25 mg portion of the sampled reaction media is diluted with 1.0 g of deionised water. As

173

before, 50 mg of NaHCO3 is added to neutralise any acid present. This sample is filtered as

174

described above before being analysed. D-fructose, D-glucose, and 5-hydroxymethylfurfural

175

concentrations are determined by detector calibration to mass prepared standard solutions.

176

7 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 8 of 36

177

In all experiments the quantity of “humins” formed is very small, and so are only determined

178

at the completion of each reaction, when the reaction mixture is filtered through glass fibre

179

paper (Whatman, grade GF/B 2.7 µm). The filter paper is subsequently washed with ethanol

180

and placed in an oven for 24 hours at 378 K. The mass of the material remaining on the paper

181

is determined by difference and referred to as “humins”.

182 183

The pH of the reaction samples is also determined in order to measure hydrogen cation

184

concentrations [H+]. An Orion pH Ag/AgCl glass electrode fitted to a VWR Symphony

185

SB70P pH meter is employed. For pH measurements, 0.4 g samples taken from the reaction

186

media are diluted with 10 g of deionised water. The pH meter is calibrated against buffer

187

solutions (VWR 32032.291) of known [H+]. The pH of samples is variable with reaction time

188

and condition but is always in the range of 1.7-2.4.

189 190

Measurement Uncertainties

191

A reproducibility and repeatability study of Test #1 shows the overall experiment-to-

192

experiment variability to be ± 12 %, which is comparable to the majority of the uncertainty

193

estimates below. For GC analysis, experimental measurement uncertainties are; ethyl

194

levulinate (± 9.6 %), 5-ethoxymethylfurfural (± 8.2 %),

195

hydroxymethylfurfural (± 10.1 %). Uncertainties in reported [H+] are generally ± 4.5 %. In

196

addition to the several identified chemical species discussed in Section 3.1, five discrete

197

components separated and detected by IC analysis cannot be identified by retention time

198

matching to expected sugar derivatives for which analytical standards are available. Figure

199

SI1 marks these detections at 1.66, 1.81, 2.11, and 2.41 and 15.27 minutes for a

200

representative chromatogram. These species are termed “unknown # 1-5” for the purposes of

201

discussion. By testing and elimination, it is determined that these detections are not due to the

D-fructose

(± 5.6 %) and 5-

8 ACS Paragon Plus Environment

Page 9 of 36

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

202

following compounds; ethyl levulinate, 5-ethoxymethylfurfural, levulinic acid, formic acid,

203

dihydroxyacetone, ethyl formate, ethyl α/β-D-glucopyranoside, H2SO4, furfural, or any

204

species that result from a series of dummy reactions comprising; H2SO4/ethanol, and each of

205

5-hydroxymethylfurfural, 5-ethoxymethylfurfural, and ethyl levulinate at 351 K for 480

206

minutes at the most extreme reaction conditions listed in Table 2. In this way, it is determined

207

that the unknown species originate from the reaction of D-fructose. As they account for a

208

considerable amount of the total ion chromatograph signal (see Table 3), their identity is

209

worthy of some speculation.

210 211

To separate carbohydrates, the PA1 Carbopac column exploits their weakly acidic nature. At

212

high pH values (supplied by the sodium hydroxide mobile phase) the carbohydrates are

213

partially ionised and can be separated by the anion exchange mechanisms embedded on the

214

column. More acidic carbohydrates bind more strongly to the column and are retained for

215

longer time. Table SI1 in the Supporting Information demonstrates the correlation of sugar

216

pKa to retention time for a series of standard carbohydrates tested.

217 218

Of the five unknowns marked in Figure SI1 (1.61, 1.81, 2.11, 2.41 and 15.27 mins), it may

219

thus be concluded that #1-4 are of higher pKa than D-fructose or D-glucose. By analogy to the

220

glucose/methanol/acid studies of Saravanamurugan et al.38, who provide evidence of the

221

formation of various methylated pyranosides and furanosides as intermediate species; it is

222

speculated that the species that are eluted before 2 minutes are ethyl fructopyranoside or

223

fructofuranoside species. It is clear from the temporal evolution of these identities that they

224

are intermediates in the formation of the desired fuel components. Only Ethyl α-D-

225

glucopyranoside (CAS 34625-23-5) and ethyl β-D-glucopyranoside analytical standards are

226

presently available (Carbosynth Ltd.) and show very similar retention times of 1.68 min and

9 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 10 of 36

227

1.72 minutes respectively under the conditions of separation. This behaviour is consistent

228

with the elutions at 1.61 and 1.81 minutes being similar such C8H16O6 ethyl pyranoside or

229

analogous ethyl furanoside isomers (for which standards are not available). As only very

230

small quantities of D-glucose are detected in any of the experimental tests, it suggests that

231

these entities are more likely fructose derived ethyl fructofuranoside and/or fructopyranoside,

232

C8H16O6 isomers. Henceforth these substances will be termed “ethyl fructosides”. As ethyl

233

fructoside analytical standards are unavailable, these moieties are quantified by standard

234

preparations of ethyl α-D-glucopyranoside and ethyl-β-D-glucopyranoside, which show

235

exactly equivalent response factors on the pulsed amperometry detector (see SI).

236 237

Correlation of pKa to retention time infers that unknown #5 (15.27 mins) is more acidic in

238

nature than

239

speculate that unknown #5 is a C6 sugar species intermediate between

240

hydroxymethylfurfural. A detailed discussion supporting this speculation is provided as SI.

D-fructose

(12.5 mins) and is thus likely be a stable sugar intermediate. We D-fructose

and 5-

241 242

The corresponding quantitative uncertainties are estimated as; unknowns #1 & #2 (ethyl

243

fructosides, ±5 %), unknowns #3 & #4 (±24 %) and unknown #5, ±12 %.

244 245 246

3. Results and Discussion Experimental Observations and Reaction Mechanism

247

Figure 2 shows a general reaction mechanism for the formation of ethyl levulinate and 5-

248

ethoxymethylfurfural from

249

measurements reported in this study and those conducted by others. Figures 3-5 show the

250

temporal evolution of the major species for representative ethanol and ethanol/water

D-fructose

that is derived from both the experimental

10 ACS Paragon Plus Environment

Page 11 of 36

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

251

conditions. The data for the other tests closely follow this detail and are available as

252

Supporting Information (SI3-SI6).

253 254

Trace Species, Not Considered in Mechanistic Analysis

255

Trace amounts of D-mannose, D-glucose, dihydroxyacetone, and 5.5’(oxybis(methylene)bis-

256

2-furfural are observed in each test condition #1-7. Levulinic acid and furfural are also

257

present at