Carbon isotopic signatures of methanol and acetaldehyde emitted from biomass burning source

Authors


Abstract

[1] We present carbon isotopic signatures of methanol and acetaldehyde emitted from biomass burning sources using laboratory experiments. The respective δ13C of methanol and acetaldehyde emitted from burning experiments of five plant materials (three C3 and two C4 plants) were −20–−46‰ and −11–−25‰. The variation in δ13C of methanol depends on the δ13C of the fuel biomass and burning conditions, but the variation in δ13C of acetaldehyde depends on the δ13C of fuel biomass and is independent of burning conditions. Combining these observations with previously reported global distributions of fire types, burning conditions, amounts of biomass burned, vegetation types, and emission factors, we estimated the global isotopic signature for biomass burning source as −33 ± 16‰ for methanol.

1. Introduction

[2] The global troposphere contains considerable amounts of oxygenated volatile organic compounds (OVOCs) [e.g., Lewis et al., 2005; Singh et al., 1995, 2001, 2004]. Their occurrence influences oxidant distributions in the atmosphere and the atmospheric chemistry of important trace gases and particles. Consequently, it affects Earth's climate. Nevertheless, knowledge of their sources and sinks, and their atmospheric chemistry remains limited.

[3] Stable isotopic signatures of OVOCs can provide information related to respective source and sink processes and the balance of source and sink strengths, i.e., their budgets, as demonstrated for other atmospheric trace gases such as methane (CH4) [e.g., Quay et al., 1999] and chloromethane (CH3Cl) [Keppler et al., 2005]. However, consideration should be given to the faster isotope modifications of atmospheric OVOCs because of their shorter lifetimes than those of CH4 and CH3Cl. Another possibility is their use for determination of the photochemical age of air masses from a certain source, as proposed by Rudolph and Czuba [2000], if kinetic isotope effects (KIEs) associated with photochemical losses of OVOCs are better understood. However, isotopic signatures on atmospheric OVOCs and each source and sink process have been limited to date [Goldstein and Shaw, 2003, and references therein]. To develop isotopic studies on understanding of each source process of OVOCs, budget estimation of atmospheric OVOCs and estimation of the photochemical age of air mass, characterization of isotopic signature of each source is needed first.

[4] Biomass burning is an important source of atmospheric OVOCs [e.g., Holzinger et al., 1999; Yokelson et al., 1999]. In this study, we characterized the first carbon isotope signatures of methanol and acetaldehyde emitted from biomass burning sources (BB) and elucidated factors controlling their variation using small chamber burning experiments of five plant materials (three C3 and two C4 plants). We also ascertained whether methanol and acetaldehyde emitted from BB are isotopically distinguishable from those emitted from fresh plant leaves, another important source known as plant growth source (PG). Methanol and acetaldehyde were chosen because of their important roles in tropospheric chemistry, as described below.

[5] Methanol is the second most abundant organic molecule in the atmosphere after methane (CH4). It is the predominant OVOC in the middle to upper troposphere [Singh et al., 2001, 2004]. It can serve as an abundant source of HOX, particularly in the free and upper troposphere [Tie et al., 2003]. It is also an important source of important atmospheric constituents such as formaldehyde [e.g., Riemer et al., 1998] and carbon monoxide [e.g., Duncan et al., 2007]. These engender a critical effect on the oxidative capacity of the atmosphere and the radiative budget. Several observational and model studies have estimated the global atmospheric methanol budget [e.g., Jacob et al., 2005]. Nevertheless, the uncertainties of those estimates are large. Methanol from BB has been estimated as constituting 10% of all sources.

[6] Acetaldehyde is also ubiquitous in the troposphere and plays an important role in tropospheric HOX chemistry. The global atmospheric acetaldehyde budget has been reported by Singh et al. [2004], but the uncertainties of the estimates are large. Biomass burning directly emits acetaldehyde [Holzinger et al., 1999] and precursors of acetaldehyde as a secondary product of oxidation.

2. Materials and Methods

[7] In all, 25 biomass burning experiments were conducted using a small closed chamber (262 l in volume). Grasses (three kinds) and wood chips (two kinds) (ca. 1 to 30 g of dry weight) were burned as fuel biomass. The burnings generally lasted a few minutes. At 20 min after burning, well mixed air in the chamber was collected into four polytetrafluoroethylene (PTFE) septum-sealed 20 ml glass vials at each burning experiment. Some air in the chamber was also collected into a 100 ml glass canister through a 1/4″ OD PTFE tube packed with water removal reagent (magnesium perchlorate, 8–24 mesh; Wako Pure Chemical Industries, Ltd., Osaka, Japan) using a diaphragm pump for measurement of CO2 and CO concentrations. Plant materials used for this study were categorized by their plant photosynthetic pathways: some were C3 plants such as Japanese Red Pine (wood chips; Pinus densiflora), Japanese Cypress (wood chips; Chamaecyparis obtusa), and oat hay grass (Avena sativa); others were C4 plants such as maize leaves (Zea mays) and Bermuda grass (Cynodon dactylon). These fuel biomass plants were dried in an oven at 60°C for more than one week until burning experiments.

[8] Concentrations of CO2 and CO in dried air samples were analyzed using a gas chromatograph equipped with a packed column (3 m × 3 mm i.d. stainless tube packed with Unibeads C; 60/80 mesh, GL Sciences Inc., Tokyo, Japan) and a TCD detector. Detection limits were 1 ppm for CO2 and 0.1 ppm for CO. The precision of replicate measurements was ±2% for both CO2 and CO.

[9] Measurement of the carbon isotope ratios of methanol and acetaldehyde in air samples was conducted using a gas chromatograph – combustion – isotope ratio mass spectrometer (GC-C-IRMS) combined with solid-phase microextraction (SPME). The instrumentation consisted of a GC-combustion interface (ThermoQuest GC Combustion III; Thermo Finnigan, Bremen, Germany), and an isotope ratio mass spectrometer (ThermoQuest DeltaplusXL; Thermo Finnigan). The GC-combustion interface consisted of a gas chromatograph (model HP6890) equipped with a split/splitless injector and a capillary column (CP-PoraBOND Q, 25 m × 0.32 mm i.d., 5 μm film; Varian, CA, SA), a combustion furnace (ceramic tube (Al2O3) packed with CuO, NiO, and Pt wires, 320 mm × 0.5 mm i.d.), a Nafion dryer, and an open split. The SPME was conducted using an 85-μm fiber coated with carboxen/polydimethylsiloxane (Carboxen/PDMS StableFlex™; Supelco, Bellefonte, PA, USA). The SPME fiber was inserted into the 20 ml glass vial; compounds were extracted for 50 min at 30°C. After extraction, the SPME fiber was introduced into the GC injection port of the GC-C-IRMS system at 220°C for 5 min under a splitless condition.

[10] Carbon isotope ratios are expressed in the conventional δ-notation relative to the international standard of Vienna Peedee Belemnite (VPDB) calculated using the following equation.

equation image

Before measurement of air samples from biomass burning experiments, we estimated the accuracy and precision of the δ13C determination using artificial gas mixtures of methanol and acetaldehyde reagents. The δ13C for the reagents were measured beforehand by a conventional off-line method and calibrated against VPDB using the international standard IAEA-RM8563. The method was certified by measurements of IAEA-CH-6 [Coplen et al., 2006]. Using the method, δ13C can be determined with an uncertainty of ±0.6‰ for more than 3 ppmv of atmospheric methanol and ±0.7‰ for more than 2 ppmv of atmospheric acetaldehyde.

[11] Measurement of the δ13C of fuel biomass (δ13CFuel Biomass) was conducted using an elemental analyzer-IRMS (EA-IRMS). The instrumentation consisted of an EA interface and an isotope ratio mass spectrometer (Finnigan MAT 252; Finnigan MAT, Bremen, Germany). The EA interface was an EA (EA 1110; CE instruments Ltd.) and a flow control unit (ConFlo II; Finnigan MAT). Isotopic standardization was accomplished using an international standard USGS40; it was certified by measurements of IAEA-CH-6 [Coplen et al., 2006]. The uncertainty of measurement results was ±0.4 ‰.

[12] The respective δ13C values of methanol and acetaldehyde emitted from plant leaves was determined for PG. Two sheets of fresh leaves detached from a tree (Ligustrum japonicum) were placed in a 20 ml glass vial with a PTFE septum. After around 12 h incubation at room temperature, δ13C of methanol and acetaldehyde in the air sample in the glass vial were measured using SPME-GC-C-IRMS. In all, six samplings were conducted during May–August: Japan's growing season.

3. Results

[13] A summary of measured δ13C of methanol and acetaldehyde in air samples collected from biomass burning experiments (δ13CMethanol-BB and δ13CAcetaldehyde-BB) is presented in Table 1 along with δ13CFuel Biomass, concentrations of CO2 and CO emitted (Δ[CO2] and Δ[CO]), and modified combustion efficiency (MCE, Δ[CO2]/(Δ[CO] + Δ[CO2])) as an index of burning conditions [Ward and Radke, 1993]. Molar emission ratios of methanol and acetaldehyde relative to CO emitted (ERMethanol and ERAcetaldehyde) are presented also.

Table 1. Measurement Results for Biomass Burning Experimentsa
Fuel BiomassAmount of Fuel Biomass Burned (g)Δ[CO2]b (ppm)Δ[CO]b (ppm)MCEcδ13CFuel Biomassd (‰ VPDB)δ13Cmethanol-BBe (‰ VPDB)δ13CAcetaldehyde-BBe (‰ VPDB)ERMethanol (mol/mol-CO)ERAcetaldehyde (mol/mol-CO)
  • a

    ND, not determined.

  • b

    Values were calculated by subtraction of background concentrations from measured concentrations.

  • c

    Values were calculated using the following equation: Δ[CO2]/{Δ[CO2] + Δ[CO]}.

  • d

    Uncertainty propagated from the standard deviations for multiple measurements of sample and standard.

  • e

    Uncertainty propagated form the standard deviations for measurements of 3–4 air samples and the uncertainties for correction of isotopic fractionations associated with SPME-GC-IRMS, except for the first line (maize leaves) and the last line (oat hay) which were one sample measurements; their uncertainties were assigned as the uncertainties of correction of isotopic fractionations associated with SPME-GC-IRMS.

Maize leavesNDNDNDND−13.0 ± 0.4−22.9 ± 0.6−12.7 ± 0.7NDND
Maize leaves4.834982 ± 2481130 ± 250.97 ± 0.01 −22.2 ± 0.6−10.8 ± 0.70.005 ± 0.0010.0028 ± 0.0000
Pine (wood chips)NDNDNDND−26.2 ± 0.4−37.5 ± 0.6−21.8 ± 0.7NDND
Pine (wood chips)1.51305 ± 1295 ± 10.93 ± 0.01 −38.4 ± 0.6−20.3 ± 0.70.027 ± 0.0030.0108 ± 0.0009
Pine (wood chips)0.82489 ± 1973 ± 10.97 ± 0.01 −32.0 ± 0.6−20.0 ± 0.70.023 ± 0.0030.0032 ± 0.0009
Pine (wood chips)1.43786 ± 30258 ± 20.94 ± 0.01 −38.4 ± 0.6−21.5 ± 0.70.015 ± 0.0010.0032 ± 0.0003
Pine (wood chips)0.82367 ± 19158 ± 60.94 ± 0.01 −37.4 ± 0.6−20.7 ± 0.70.008 ± 0.0010.0027 ± 0.0003
Cypress (wood chips)NDNDNDND−22.8 ± 0.4−33.5 ± 0.6−22.6 ± 0.7NDND
Cypress (wood chips)10.430675 ± 219863 ± 150.97 ± 0.01 −30.8 ± 0.6−20.3 ± 0.70.015 ± 0.0010.0026 ± 0.0001
Cypress (wood chips)7.620020 ± 140515 ± 130.97 ± 0.01 −29.2 ± 0.6−20.2 ± 0.70.013 ± 0.0030.0033 ± 0.0004
Cypress (wood chips)10.930274 ± 2141272 ± 270.96 ± 0.01 −32.0 ± 0.6−22.4 ± 0.70.022 ± 0.0020.0026 ± 0.0001
Cypress (wood chips)17.348047 ± 3341903 ± 410.96 ± 0.01 −33.5 ± 0.6−22.5 ± 0.70.025 ± 0.0020.0013 ± 0.0001
Cypress (wood chips)8.9NDNDND −33.0 ± 0.6−21.8 ± 0.70.016 ± 0.0010.0004 ± 0.0001
Bermuda grass27.122705 ± 1672536 ± 530.90 ± 0.01−13.3 ± 0.4−27.0 ± 0.6−12.5 ± 0.70.007 ± 0.0010.0013 ± 0.0001
Bermuda grass10.025670 ± 3251364 ± 330.95 ± 0.02 −20.0 ± 0.6−12.5 ± 0.70.008 ± 0.0010.0032 ± 0.0002
Bermuda grass13.1NDNDND −24.6 ± 0.6−12.7 ± 0.7NDND
Bermuda grass20.538565 ± 2673582 ± 830.92 ± 0.01 −24.5 ± 0.6−12.2 ± 0.70.007 ± 0.0000.0013 ± 0.0002
Bermuda grass5.011039 ± 81617 ± 150.95 ± 0.01 −25.4 ± 0.6−12.4 ± 0.70.007 ± 0.0010.0034 ± 0.0004
Bermuda grass2.96179 ± 43405 ± 70.94 ± 0.01 −23.5 ± 0.6−11.2 ± 0.70.007 ± 0.0010.0050 ± 0.0004
Oat hay13.817029 ± 1311961 ± 400.90 ± 0.01−25.3 ± 0.4−42.4 ± 0.6−24.8 ± 0.70.011 ± 0.0010.0021 ± 0.0003
Oat hay6.69042 ± 651307 ± 330.87 ± 0.01 −44.1 ± 0.6−24.4 ± 0.70.016 ± 0.0010.0046 ± 0.0004
Oat hay19.832419 ± 2303641 ± 820.90 ± 0.01 −39.6 ± 0.6−24.5 ± 0.70.008 ± 0.0000.0025 ± 0.0002
Oat hay7.217972 ± 2931155 ± 270.94 ± 0.02 −39.3 ± 0.6−24.9 ± 0.70.016 ± 0.0000.0050 ± 0.0006
Oat hay5.515928 ± 1441027 ± 260.94 ± 0.01 −37.1 ± 0.6−23.8 ± 0.70.007 ± 0.0010.0041 ± 0.0003
Oat hay2.73.033 ± 22525 ± 150.85 ± 0.01 −48.5 ± 0.6−24.3 ± 0.70.0130.0060
Geometric Mean       0.012 ± 0.0070.003 ± 0.002

[14] The ERMethanol was 0.005–0.027. The geometric mean was 0.012 ± 0.007 (Table 1). The range and the mean value are consistent with previous reports of laboratory experiments and field observations [e.g., Holzinger et al., 1999; Christian et al., 2003; Singh et al., 2004; Yokelson et al., 1999, 2003]. The ERAcetaldehyde were 0.0004–0.0108; the geometric mean was 0.003 ± 0.002 (Table 1). Both are consistent with field observations compiled by Andreae and Merlet [2001]. Our results for burning experiments are representative of BB in general.

[15] The δ13CMethanol-BB for C3 plant burning (−29.2–−45.8‰) was lower than those for C4 plant burning (−20.0–−27.0‰), depending on the δ13CFuel Biomass. The δ13CMethanol-BB varied up to 8.6‰ despite the use of an identical fuel biomass (in case of oat hay burning). Similar to δ13CMethanol-BB, δ13CAcetaldehyde-BB depended on the δ13CFuel Biomass, showing −20.0–−24.9‰ and −10.8–−12.7‰ for C3 and C4 plant burnings, respectively. Variation in δ13CAcetaldehyde-BB for the use of an identical fuel biomass (ca. 2‰) was smaller than that for δ13CMethanol-BB.

[16] The δ13C of methanol emitted from fresh plant leaves (δ13CMethanol-PG) were −73.3–−76.1‰; the average was −74.6 ± 1.2‰, which is remarkably lower than that of the bulk δ13C of plant leaves (−27.8‰). The δ13C values of acetaldehyde emitted from fresh plant leaves (δ13CAcetaldehyde-PG) were −25.0–−30.7‰. The average was −28.9 ± 2.1‰, which was the same as the bulk δ13C of plant leaves. These observations were consistent with data reported by Keppler et al. [2004]: the δ13CMethanol-PG for C3 and C4 fresh plant leaves were, respectively, −68.2 ± 11.2‰ and −52.9 ± 0.8‰, which were lower values than those of bulk δ13C of respective plant leaves. Their study demonstrated that methanol was produced from the isotopically highly depleted pectin methoxyl pool (−63.6 ± 7.9‰ for C3 plants and −40.5 ± 0.6‰ for C4 plants) in leaf tissue with almost no fractionation. The δ13CAcetaldehyde-PG of −24.9 ± 2.2‰ for C3 plants and −9.3 ± 1.3‰ for C4 plants resemble values for bulk δ13C of respective plant leaves, indicating that acetaldehyde comes from precursor compounds that reflect the δ13C of the bulk plant material. These demonstrate that δ13CMethanol-PG is significantly smaller than δ13CMethanol-BB, although δ13CAcetaldehyde-PG is similar to δ13CAcetaldehyde-BB for the same plant type (i.e., C3 or C4).

4. Discussion

[17] As described above, one factor controlling the variation in δ13CMethanol-BB and δ13CAcetaldehyde-BB is variation in δ13CFuel Biomass. However, the δ13CMethanol-BB and δ13CAcetaldehyde-BB varied up to ca. 9‰ and ca. 2‰, respectively, in spite of the identical δ13CFuel Biomass. Additional controlling factors of variations in δ13CMethanol-BB and δ13CAcetaldehyde-BB should be considered. We evaluated burning conditions as another controlling factor of variations in δ13CMethanol-BB and δ13CAcetaldehyde-BB below.

[18] To offset the influence of variations in δ13CFuel Biomass, we calculated the isotopic differences between δ13CMethanol-BB and δ13CAcetaldehyde-BB and δ13CFuel Biomass, which represents the apparent isotopic fractionation occurring during biomass burning and which is expressed as ɛ (‰), as defined by the following equation.

equation image

[19] The ɛMethanol-BB and ɛAcetaldehyde-BB were, respectively, −6.0‰–−21.0‰ and 0.2‰–6.3‰. Factors controlling the variation in ɛ would be mainly an isotopic fractionation that is inversely correlated with temperature in the first pyrolysis process and the isotopic fractionation associated with the subsequent methanol and acetaldehyde loss process involving enrichment in 13C in the residual methanol and acetaldehyde, same as CH4 [Chanton et al., 2000; Yamada et al., 2006] and some non-methane hydrocarbons [Czapiewski et al., 2002] emitted from biomass burning, indicating that ɛ is expected to correlate with burning conditions.

[20] Plots of ɛ versus MCE, a numerical indicator of burning conditions, for all data in this study are portrayed in Figure 1. The ɛMethanol-BB was high when MCE was high (i.e., closer to complete combustion). When MCE was low (i.e., higher contribution of incomplete combustion), the ɛMethanol-BB was low. Linear regression of ɛMethanol-BB data revealed strong correlations (r = 0.9) irrespective of fuel biomass, indicating that the variation in ɛMethanol-BB is explained by MCE variation. The relation is expressed as the following equation.

equation image
Figure 1.

Here ɛMethanol-BB and ɛAcetaldehyde-BB are portrayed as a function of modified combustion efficiency (MCE) for biomass burning experiments. The regression line is expressed as ɛMethanol-BB = (−112.2 ± 10.6) + (107.3 ± 11.4) × MCE (R = 0.92). Boxes at the left side show isotopic differences between emitted methanol and acetaldehyde and heated biomass calculated for heating experiments of plant biomass conducted by Keppler et al. [2004]: −31 ± 9‰ for methanol and +0.5 ± 2.5‰ for acetaldehyde.

[21] Keppler et al. [2004] reported the δ13C of methanol emitted from heating (not burning) of leaf biomass at 225°C. According to their data, the isotopic difference between emitted methanol and heated biomass, which corresponds to ɛMethanol-BB in our study, was −13.8‰–−45.0‰; the mean was −30.7 ± 8.9‰. They also revealed that methanol was released from the pectin methoxyl pool in leaf tissue and that the release was completed by 300°C. Furthermore, they suggested that the lignin methoxyl pool, which is isotopically heavier by ca. 3‰ than the pectin methoxyl pool, contributes to methanol production for biomass burning. Combining these findings with our results, variations in δ13CMethanol-BB can be explained as follows; methanol having small δ13C by ca. −31 ± 9‰ compared to its fuel biomass is released from the pectin methoxyl pool in fuel biomass at the initial stage of biomass burning. When increasing the temperature, the proportion of methanol from the lignin methoxyl also increases, resulting in a slight enrichment in 13C in methanol. Subsequently, a part of methanol itself burns as a fuel during biomass burning. The methanol loss process involves enrichment in 13C in the residual methanol and the extent of enrichment in 13C depends on the extent of loss, namely burning conditions, according to the relation between ɛ and MCE observed in this study.

[22] The plot for ɛAcetaldehyde-BB shows that ɛAcetaldehyde-BB, except for the data of pine burning, was constant (0.1–2.7‰) irrespective of MCE (Figure 1). According to results reported by Keppler et al. [2004], the isotopic difference between emitted acetaldehyde and heated biomass was −4.2‰–+3.2‰; the mean was +0.5 ± 2.5‰, being similar to the ɛAcetaldehyde-BB observed in this study. The reasons for the constant ɛAcetaldehyde-BB irrespective of MCE and the different ɛAcetaldehyde-BB of pine remain unclear.

[23] Results of this study indicate that the ɛMethanol-BB can be expressed as a function of MCE. The global distribution of MCE might reflect the global distribution of the ɛMethanol-BB when the generality of the relation is assumed for various fires on a global scale. Furthermore, when the global distribution of δ13CFuel Biomass is assigned, the global average of isotopic signatures for BB can be inferred from equation (2).

[24] As a case study, the global average of isotopic signature of methanol emitted from BB was inferred based on regional and the global distributions of MCE [Hao and Ward, 1993], C3 and C4 vegetation as fuel biomass [Still et al., 2003], methanol emissions from BB calculated from the amount of biomass burned [Hao and Ward, 1993], and the emission factor of methanol [Andreae and Merlet, 2001] using the same manner as that used for estimation of regional and the global averages of isotopic signature of CH4 emitted from BB conducted by Yamada et al. [2006]. Estimation of the global methanol emission from BB was calculated as 9 Tg/y, which is in line with estimations summarized by Jacob et al. [2005]. The estimated global average of methanol for BB is −33 ± 16‰. The uncertainty was estimated from the uncertainty of ɛMethanol-BB, as calculated using equation (3). This value indicates that methanol from BB is distinguishable from that from PG, although the calculation is based on isotopic measurements for limited number of plant species. Furthermore, large geographical and temporal heterogeneity of biomass burnings must be considered and the calculation parameters such as global distributions of C3/C4 vegetation and MCE must be refined for more accurate estimation.

Acknowledgments

[25] The authors thank two reviewers for their valuable comments on the manuscript. This work was supported by a Grant-in-Aid for Scientific Research (C) (19510005) and the Global Environment Research Fund (B-094) of the Ministry of the Environment, Japan.

Ancillary