Compound-Specific Carbon Isotopic Composition of Ethanol in Brazil

Subscriber access provided by YORK UNIV

Energy and the Environment

Compound-specific carbon isotopic composition of ethanol in Brazil and US vehicle emissions and wet deposition Joseph David Felix, Rachel Thomas, Matt Casas, Megumi Shimizu, Gene Brooks Avery, Robert J. Kieber, Ralph N. Mead, Chad Lane, Joan D. Willey, Amanda Guy, and Maria Lucia Campos Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.8b05325 • Publication Date (Web): 19 Dec 2018 Downloaded from http://pubs.acs.org on December 21, 2018

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 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 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.

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 24

Environmental Science & Technology

1 2 3

Compound-specific carbon isotopic composition of ethanol in Brazil and US vehicle emissions and wet deposition

4 5 6

*J. David Felix1, Rachel Thomas2, Matt Casas2, Megumi S. Shimizu2, G. Brooks Avery2, Robert J. Kieber2, Ralph N. Mead2, Chad S. Lane3, Joan D. Willey2, Amanda Guy2, M. Lucia A. M. Campos4

7 8 9 10 11 12

1. Department of Physical and Environmental Science, Texas A&M University – Corpus Christi, Corpus Christi, Texas, USA, 78412

13 14 15 16 17 18

3. Department of Earth and Ocean Sciences, University of North Carolina Wilmington, Wilmington, NC, USA, 28403

19

2. Department of Chemistry and Biochemistry, University of North Carolina Wilmington, Wilmington, NC, USA, 28403

4. Departamento de Química, Faculdade de Filosofia, Ciencias e Letras de Ribeirão Preto, Universidade de São Paulo, Avenida dos Bandeirantes 3900, 14040-901 Ribeirão Preto, São Paulo, Brazil *Corresponding author: [email protected]

20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 1 ACS Paragon Plus Environment


37 38

ABSTRACT Global atmospheric ethanol budget models include large uncertainties in the magnitude

39

of ethanol emission sources and sinks. In order to apply stable isotope techniques to constrain

40

ethanol emission sources, a headspace solid phase micro-extraction gas chromatograph-

41

combustion-isotope ratio mass spectrometry method (HS-SPME-GC-C-IRMS) was developed to

42

measure the carbon isotopic composition of aqueous phase ethanol at natural abundance levels (1 to 30

43

µM) with a precision of 0.4‰. The method was applied to determine the carbon isotope signatures (δ13C)

44

of vehicle ethanol emission sources in Brazil (-12.8 ± 2.4‰) and the US (-9.8 ± 2.5‰), and to measure

45

the carbon isotope composition of ethanol in wet deposition (-22.6 to -12.7‰). A two end-member

46

isotope mixing model was developed using anthropogenic and biogenic endmembers and fractionation

47

scenarios to estimate ethanol source contributions to wet deposition collected in Brazil and US. Mixing

48

model results indicate anthropogenic sources contribute two and a half to four times more ethanol to the

49

atmosphere than previously predicted in modeled global ethanol inventories. As established and

50

developing countries continue to rapidly increase ethanol fuel consumption and subsequent

51

emissions, understanding the magnitude of all ethanol sources and sinks will be essential for

52

modeling future atmospheric chemistry and air quality impacts.

53 54 55

INTRODUCTION Global ethanol fuel production and consumption have doubled over the previous decade

56

in efforts to reduce greenhouse gas emissions and increase domestic production of fuel.1,2 Sixty-

57

four countries currently have active biofuel mandates or targets that are increasing ethanol

58

production.3,4

59

placed on the effects ethanol production has on a country’s food, water, and energy supply5-7

60

while less effort has been applied to considering the atmospheric impacts of increased ethanol

When determining the pros and cons of production mandates, focus is often

2 ACS Paragon Plus Environment

Page 2 of 24

Page 3 of 24


61

consumption. Studies focusing on atmospheric effects have concentrated on potential decreases

62

in greenhouse gases and VOCs, but there are number of adverse atmospheric impacts associated

63

with ethanol emissions and combustion byproducts.2,8,9

64

Vehicle ethanol emissions can alter the oxidizing capacity of the atmosphere and alter

65

radical and NOx cycling as well as NOy speciation.10,11 Ethanol combustion also leads to an

66

increase in atmospheric concentrations of the secondary pollutant peroxyacetyl nitrate and a

67

variety of VOCs with relatively high ozone reactivities.12 For instance, large increases in

68

formaldehyde and acetaldehyde are associated with blended fuel (i.e. gasoline/ethanol)

69

emissions, both of which have relatively high ozone reactivities13 and many studies modeling

70

hypothetical increases in blended fuel consumption in the US have predicted an increase in

71

national average ozone concentrations.2, 14-16 However, depending on the NOx and VOC

72

sensitivity of ozone formation in individual regions, some regions in the US may experience

73

decreasing ozone concentrations with increasing ethanol fuel consumption.2,14 The increasing

74

effect of ethanol fuel emissions on tropospheric ozone production was recently observed in a

75

“real world” scenario when Brazil experienced large fluctuations in the price of ethanol. This

76

forced the proportion of bi-fuel vehicles (engines capable of running on any proportion of

77

gasoline and ethanol) burning gasohol (i.e E27, where EXX is defined as gasoline-ethanol blend

78

and XX is % ethanol) instead of ethanol, to rise from 14 to 76%. In São Paulo, Brazil this 62%

79

increase in bi-fuel vehicles burning gasohol, resulted in an average of 22% decrease in ground-

80

level ozone.17,18 Despite this observed correlation between ethanol fuel consumption and ozone

81

levels, most research investigating the impacts of ethanol fuel consumption on air quality are

82

modeling efforts based on limited empirical atmospheric ethanol concentration data.



83

The lack of atmospheric ethanol concentration data has led to a deficiency in

84

comprehension of the sources, sinks, and cycling of ethanol in the atmosphere and has been

85

noted in various studies.2,8, 11, 19-23 Subsequently, large discrepancies are reported among

86

estimated atmospheric ethanol inventories. For instance, recent ethanol budget estimates range

87

from 7 to 56 Tg/yr for the total ethanol source to the atmosphere. The reported major source of

88

ethanol (biogenic emissions from plants) ranges from 21% to 92% of total emissions while

89

anthropogenic sources (vehicle and ethanol production facility emissions) range from 8% to 34%

90

of total emissions.8,10,22,24 The large range in estimated ethanol contribution from anthropogenic

91

sources highlights the general lack of knowledge of vehicle ethanol emission contributions to

92

atmospheric ethanol concentrations. Delineating the significance of this anthropogenic source to

93

atmospheric concentrations is vital when debating the viability of the current and future global

94

biofuel mandates.

95

One quantitative approach to determining the significance of anthropogenic versus

96

biogenic emission contributions to atmospheric ethanol concentrations is to measure the carbon

97

isotopic composition of ambient ethanol (δ13CEtOH) in the air or in atmospheric waters (e.g. wet

98

deposition, condensates). The δ13C values of ethanol from the major biogenic source (C3 plants)

99

(-30.9 ± 4.2‰ (n =15)) are highly distinguishable from the vehicle exhaust emission source

100

(direct vehicle sampling: -5.0‰ (n =1); roadside sampling: 10.4 +3.6‰) (n = 25)).21,25 The large

101

difference in δ13C signatures between these two sources is ideal for constructing a two

102

endmember isotope mixing model for quantifying the relative contribution of biogenic versus

103

anthropogenic emission sources to the atmosphere. However, the vehicle emission endmember

104

is based on one measurement from a vehicle produced in 1972 that was not equipped with a

105

catalytic converter and current combustion efficiency technology and single site roadside


Page 4 of 24

Page 5 of 24


106

collections of ambient air used to represent a vehicle fleet.25 For this mixing model approach to

107

be effective in estimating source contribution, more adequate endmember source characterization

108

needs to be undertaken.

109

In order to further constrain this vehicle endmember and apply an isotope mixing model

110

to determine ethanol source contributions to the atmosphere this work aims to 1) combine a

111

headspace solid phase micro-extraction gas chromatograph (HS-SPME-GC) method optimized

112

for the analysis of low aqueous phase ethanol concentration samples with gas chromatograph-

113

combustion-isotope ratio mass spectrometry (GC-C-IRMS), 2) characterize the carbon isotopic

114

composition of ethanol in vehicle emissions, and 3) determine biogenic and anthropogenic

115

emission source contributions to ethanol concentrations in Brazil and US wet deposition,

116

countries that account for ~84% of the world’s ethanol fuel consumption and production.26

117 118

MATERIALS AND METHODS

119

Rain Sample Collection and Preservation

120

Rainwater samples were collected on an event basis in 2016 at the University of North

121

Carolina Wilmington (UNCW) campus (Figure 1). The UNCW rainwater collection site (34°

122

22.0' N, 77° 86.3' W) is an open area of land of approximately 0.01 km2 containing vegetation

123

that is representative of inland coastal regions of southeastern North Carolina including Aristida

124

stricta, Quercus laevis, and Pinus palustris. The site is located approximately 8.5 km from the

125

Atlantic Ocean. Samples were collected utilizing Aerochem-Metrics (ACM) Model 301

126

Automatic Sensing Wet/Dry Precipitation Collectors containing glass beakers. All beakers and

127

glassware used for the collection and filtration were rinsed with Millipore MQ deionized water

128

(resistivity of 18 MΩ.cm @ 25 °C) and combusted in a muffle furnace at 450°C for a minimum



129

of 4.5 hours to eliminate organics prior to use. Immediately after the rain event, samples were

130

returned to the laboratory and filtered using 0.2 μm acid-washed Gelman Supor® polysulfonone

131

filters. Brazil wet deposition were collected using an automated sampler located at a site on the

132

University of São Paulo campus in Ribeirão Preto (21° 11’ S, 47° 48’ W) (Figure 1) and

133

underwent the same treatment as the Wilmington rain, except the glassware used for ethanol

134

analysis was cleaned with Fenton solution, a mixture of Fe2+ salts with hydrogen peroxide.27 All

135

samples were preserved by adding 70 µL of 100 mg/L HgCl2 to 40 mL of sample to eliminate

136

variability in ethanol concentrations that could occur during sample collection and analysis. Previous

137

stability experiments report ethanol concentration stability for up to 160 days when preserved with 100

138

mg/L HgCl2.23

139 140 141

Vehicle Exhaust Sample Collection Emissions from four vehicles in the US consuming E10 fuel and five Brazilian vehicles

142

consuming either E27 or E100 fuel were sampled. A 500 mL fluorinated ethylene propylene

143

resin (FEP) bottle was placed on the tailpipe of the vehicle. The bottle’s bottom was removed

144

and feathered so that it could be fitted to different tail pipe diameters. The feathered end was then

145

sealed on the tailpipe of the car using a hose clamp. On the opposite end, the bottle was attached

146

to a 250 mL low-density polyethylene (LDPE) bottle containing 100 mL of synthetic wet

147

deposition (SWD, pH 4.5 H2SO4). Exhaust was collected from cars while idling in neutral. The

148

exhaust was allowed to bubble through the 100 mL of SWD for ten minutes to ensure ethanol

149

concentrations were well above the HS-GC-C-IRMS analysis limit of detection of 1 µM. All

150

exhaust was 0.2 µm filtered and immediately preserved with 100 mg/L HgCl2 upon collection.

151

Ethanol Concentration Analysis


Page 6 of 24

Page 7 of 24

152


A previously published solid-phase microextraction (SPME) ethanol concentration

153

analysis optimized for low concentration aqueous phase samples was used to analyze vehicle

154

exhaust and wet deposition samples.28 In summary, 3.5 g of NaCl was added to 11.6 mL of

155

vehicle or wet deposition sample, which was then adjusted to a final pH of 4.4 by adding 0.4 mL

156

of a saturated succinic acid solution. The sample was heated at 50 ºC for 10 minutes prior to a

157

20-minute extraction by a 75 µm Carboxen/PDMS SPME fiber in the headspace. Magnetic

158

stirring at a rate of 750 rpm occurred during heating and extraction. The fiber was thermally

159

desorbed after extraction in the injection port of the GC-FID for concentration analysis. The

160

same method was used for ethanol extraction before injection to the GC-C-IRMS.

161 162 163

Carbon Isotope Analysis of Ethanol A Thermo Trace 1310 GC was equipped with a Restek®-5 (30 m x 0.32 mm x 1 μm) or

164

Equity-5 (20 m x 0.32 mm x 0.32 μm) fused-silica capillary column and a Merlin microseal

165

adapter was mounted on the GC injection port. A Taylor-Wharton 25 L liquid nitrogen Dewar

166

was equipped with a liquid withdrawal device connected to the back of the GC oven by a copper

167

pipe to allow for sub-ambient temperatures. Sub-ambient temperatures focused the ethanol upon

168

injection leading to improved chromatography and separation.

169

After ethanol was extracted via the SPME method described previously, the SPME fiber

170

was introduced to the GC-C-IRMS via the Merlin microseal septum encased in the microseal

171

adapter mounted on the GC injection port. The oven temperature was held for 3 minutes at 10oC

172

before being ramped to 60°C at 10°C/min and heated to the final temperature of 200°C at

173

60°C/min and held for 2 minutes. Eluents were carried to instrumentation (e.g.Thermo Isolink II)

174

designed for conversion of analytes to CO2 via the helium carrier gas where they were



Page 8 of 24

175

combusted at 1000°C in the presence of nickel oxide and platinum wires in an alumina ceramic

176

reactor tube. Resulting CO2 was analyzed in continuous-flow mode via an interfaced Thermo

177

Delta V plus stable isotope mass spectrometer at the University of North Carolina Wilmington

178

Isotope Ratio Mass Spectrometry (UNC-WIRMS) Facility.

179

The vacuum system of the IRMS was maintained at a constant operating source pressure

180

of 4.9 x 10-8 mbar. Standards and samples were calibrated daily against a working CO2

181

reference gas versus the international Vienna PeeDee Belemnite (V-PDB) standard (equation 1). δ13Csample = [(13C/12C)sample / (13C/12C)V-PDB -1] x 1,000

182

(1)

183

Two ethanol isotope standards were used to apply a two-point normalization to the V-PDB

184

scale.29 The standards were purchased from the stable isotope reference program at Indiana

185

University's Biogeochemical Laboratories (δ13C values of -10.98 ± 0.02‰ and -27.53 ± 0.02‰).

186

Reference standards were then further verified using a Costech 4010 elemental analyzer

187

interfaced with a Thermo Delta V Plus stable isotope ratio mass spectrometer (EA-IRMS) at the

188

University of North Carolina Wilmington Isotope Ratio Mass Spectrometry (UNC-WIRMS)

189

Facility.

190

Source Apportionment

191

Ethanol source contribution to wet deposition was calculated using a two-endmember

192

mixing model (eq. 2). EPA ISOERROR30 was used to report uncertainty due to the variability in

193

source signatures and deviation of the δ13CEtOH values in wet deposition samples. The uncertainty

194

is reported as the standard error of the mean at 95% confidence level. In order to calculate

195

uncertainty, the ISOERROR user must provide mixture δ13CEtOH values with standard deviations

196

and source endmember δ13CEtOH values with standard deviations. Our anthropogenic endmember 8 ACS Paragon Plus Environment

Page 9 of 24


197

represents both vehicle and ethanol production facility emissions due to the observed overlap of

198

vehicle emission and commercially produced δ13CEtOH values (results section) . The

199

anthropogenic endmembers for the Ribeirão Preto, SP, BR and Wilmington, NC, USA were

200

calculated separately due to differences between ethanol fuel feedstock, fuel content and vehicle

201

engines (further detail in discussion section). The biogenic endmember δ13CEtOH value was

202

calculated from available literature of δ13CEtOH values emitted from C3 plants (-30.9 ± 4.2‰ (n

203

=15)).21,25 C3 plant emissions were chosen to represent the biogenic endmember because C3

204

plant biomass contributes ~95% of the world’s total C3 + C4 biomass.31 It should be noted that

205

this assumes that C3 plants and C4 plants emit ethanol at similar amounts per biomass and future

206

work should focus on quantifying C3 vs. C4 emission rates in an effort to understand the

207

significance of their role in ethanol emission. Other endmembers (e.g. biomass burning, ocean,

208

C4 plants) are acknowledged, but are assumed to contribute minimally. For instance, Kirstine

209

and Galbally 2012 estimate global biomass burning and ocean to contribute 1.9 and 9.5%,

210

respectively. Additionally many of the ethanol source δ13CEtOH values have not been thoroughly

211

characterized.21

212

δ13CEtOH wet deposition =𝑓anthro (δ13CEtOH anthro) + (1-𝑓anthro)( δ13CEtOH biogenic)

(2)

213 214

HYSPLIT Back Trajectory Models

215

Air-mass back trajectories were generated for each wet deposition sample using the

216

Hybrid Single Particle Lagrangian Integrated Trajectory model (HYSPLIT) (Figure 1). The back

217

trajectories were calculated starting at the end of the precipitation event for a 72-hour hind-cast

218

at an altitude of 500 m. Seventy two hours was chosen because the estimated lifetime of ethanol



219

is 2.8 days10,32 and the 500 m level represented the air mass near the well-mixed boundary layer

220

likely to contribute more heavily to in-cloud processes contributing to wet deposition.33

221 222

Figure 1: Left: Sampling sites Ribeirão Preto, SP, BR (purple circle) and Wilmington, NC,

223

USA (green circle). Middle: Wilmington NC, USA (X) wet deposition event air mass back

224

trajectories. Right: Ribeirão Preto, SP, BR (X) wet deposition events air mass back trajectories.

225

Maps created via ESRI34.

226

RESULTS AND DISCUSSION

227

HS-SPME-GC-C-IRMS method

228

Ethanol isotope standards (-10.98 ± 0.02‰ and -27.53 ± 0.02‰) were analyzed at

229

various concentrations to determine isotope fractionation associated with the SPME process. The

230

measured isotopic composition of the isotope standard with the higher δ13CEtOH value ranged

231

between -10.2 and -11.2 ‰ with a mean of -10.9 ± 0.4‰ indicating there was no offset from the

232

value reported by the manufacturer. Additionally, there was no concentration dependent offset

233

variation between 1 and 30 µM (Figure 2). The measured δ13CEtOH of the lower isotope standard

234

(-29.9 ± 0.5) was offset from the manufacturer’s given value by an average of -2.4‰, but the

235

δ13C offset was consistent over a range of concentrations between 1 and 30 µM and any observed


Page 10 of 24

Page 11 of 24


236

variation was not significant (ANOVA, p = 0.4) thus the “offset” does not affect the results when

237

using two-point normalization (Figure 2).

238

This measured offset of the lower standard is similar to the offset (1.4 to 2.2‰) observed

239

by previously reported SPME-GC-C-IRMS ethanol analysis methods35 and has been observed in

240

the analysis of several other polar VOCs (

Compound-Specific Carbon Isotopic Composition of Ethanol in Brazil

Recommend Documents