The stable isotopic signature of biologically produced molecular hydrogen (H 2 )

Abstract. Biologically produced molecular hydrogen (H2) is characterised by a very strong depletion in deuterium. Although the biological source to the atmosphere is small compared to photochemical or combustion sources, it makes an important contribution to the global isotope budget of H2. Large uncertainties exist in the quantification of the individual production and degradation processes that contribute to the atmospheric budget, and isotope measurements are a tool to distinguish the contributions from the different sources. Measurements of δ D from the various H2 sources are scarce and for biologically produced H2 only very few measurements exist. Here the first systematic study of the isotopic composition of biologically produced H2 is presented. In a first set of experiments, we investigated δ D of H2 produced in a biogas plant, covering different treatments of biogas production. In a second set of experiments, we investigated pure cultures of several H2 producing microorganisms such as bacteria or green algae. A Keeling plot analysis provides a robust overall source signature of δ D = −712‰ (±13‰) for the samples from the biogas reactor (at 38 °C, δ DH2O= +73.4‰), with a fractionation constant vH2-H2O of −689‰ (±20‰) between H2 and the water. The five experiments using pure culture samples from different microorganisms give a mean source signature of δ D = −728‰ (±28‰), and a fractionation constant vH2-H2O of −711‰ (±34‰) between H2 and the water. The results confirm the massive deuterium depletion of biologically produced H2 as was predicted by the calculation of the thermodynamic fractionation factors for hydrogen exchange between H2 and water vapour. Systematic errors in the isotope scale are difficult to assess in the absence of international standards for δ D of H2. As expected for a thermodynamic equilibrium, the fractionation factor is temperature dependent, but largely independent of the substrates used and the H2 production conditions. The equilibrium fractionation coefficient is positively correlated with temperature and we measured a rate of change of 2.3‰ / °C between 45 °C and 60 °C, which is in general agreement with the theoretical prediction of 1.4‰ / °C. Our best experimental estimate for vH2-H2O at a temperature of 20 °C is −731‰ (±20‰) for biologically produced H2. This value is close to the predicted value of −722‰, and we suggest using these values in future global H2 isotope budget calculations and models with adjusting to regional temperatures for calculating δ D values.


Abstract. Biologically produced molecular hydrogen (H 2 )
is characterised by a very strong depletion in deuterium. Although the biological source to the atmosphere is small compared to photochemical or combustion sources, it makes an important contribution to the global isotope budget of H 2 . Large uncertainties exist in the quantification of the individual production and degradation processes that contribute to the atmospheric budget, and isotope measurements are a tool to distinguish the contributions from the different sources. Measurements of δD from the various H 2 sources are scarce and for biologically produced H 2 only very few measurements exist.
Here the first systematic study of the isotopic composition of biologically produced H 2 is presented. In a first set of experiments, we investigated δD of H 2 produced in a biogas plant, covering different treatments of biogas production. In a second set of experiments, we investigated pure cultures of several H 2 producing microorganisms such as bacteria or green algae. A Keeling plot analysis provides a robust overall source signature of δD = −712 ‰ (±13 ‰) for the samples from the biogas reactor (at 38 • C, δD H 2 O = +73.4 ‰), with a fractionation constant ε H 2 -H 2 O of −689 ‰ (±20 ‰) between H 2 and the water. The five experiments using pure culture samples from different microorganisms give a mean source signature of δD = −728 ‰ (±28 ‰), and a fractionation constant ε H 2 -H 2 O of −711 ‰ (±34 ‰) between H 2 and the water. The results confirm the massive deuterium depletion of biologically produced H 2 as was predicted by the cal-culation of the thermodynamic fractionation factors for hydrogen exchange between H 2 and water vapour. Systematic errors in the isotope scale are difficult to assess in the absence of international standards for δD of H 2 .
As expected for a thermodynamic equilibrium, the fractionation factor is temperature dependent, but largely independent of the substrates used and the H 2 production conditions. The equilibrium fractionation coefficient is positively correlated with temperature and we measured a rate of change of 2.3 ‰ / • C between 45 • C and 60 • C, which is in general agreement with the theoretical prediction of 1.4 ‰ / • C.
Our best experimental estimate for ε H 2 -H 2 O at a temperature of 20 • C is −731 ‰ (±20 ‰) for biologically produced H 2 . This value is close to the predicted value of −722 ‰, and we suggest using these values in future global H 2 isotope budget calculations and models with adjusting to regional temperatures for calculating δD values. contributes significantly to atmospheric chemistry (Novelli et al., 1999;Hauglustaine and Ehhalt, 2002;Rahn et al., 2003 Ehhalt andRohrer, 2009). By reaction with the hydroxyl radical ( q OH), hydrogen indirectly increases the atmospheric lifetimes of other trace gases that also react with q OH, for example, CH 4 and carbon monoxide (CO) (Prather, 2003;Schultz et al., 2003;Jacobson et al., 2005;Jacobson, 2008) and, therefore, acts as an 'indirect' greenhouse gas. In the stratosphere, oxidation of H 2 is a source of water vapour, which is important for the radiative properties of the stratosphere and also forms the substrate for polar stratospheric clouds, which are key ingredients in the formation of the polar ozone holes (Tromp et al., 2003;Warwick et al., 2004;Feck et al., 2008;Jacobson, 2008).
H 2 is considered as a promising future energy carrier. It can be produced chemically, physically and biologically. The shortage, increase in cost and climate impact of fossil fuels leads to increased interest in sustainable and clean production of H 2 . One possible source to accommodate the expected energy demand might be biologically produced H 2 , e.g., via fermentation or photosynthesis.
Numerous studies in the past have addressed the global atmospheric budget of H 2 , but still none of the individual source or sink strengths is constrained to better than ±25 % (Ehhalt and Rohrer, 2009). Additional information is expected to come from the analysis of the H 2 isotopic composition (δD), because the different sources of H 2 have a very different deuterium content. δD is defined as the relative deviation of the D / H ratio in a sample from the same ratio in the international reference material Vienna Standard Mean Ocean Water (VSMOW). Also the kinetic fractionation in the two main removal processes, soil deposition and reaction with OH, is different.
Tropospheric H 2 is enriched in deuterium with δD ∼ +130 ‰, (Gerst and Quay, 2001;Rhee et al., 2006;Rice et al., 2010;Batenburg et al., 2011) compared to surface emissions from fossil fuel combustion and biomass burning (δD approximately −200 ‰ and −300 ‰, respectively) (Gerst and Quay, 2001;Rahn et al., 2002;Röckmann et al., 2010a;Vollmer et al., 2010). As originally proposed by Gerst and Quay (2001) from budget closure, the photochemical sources of H 2 are also enriched in deuterium with δD between ∼ +100 ‰ and +200 ‰, (Rahn et al., 2003;Röckmann et al., 2003Röckmann et al., , 2010bFeilberg et al., 2007;Nilsson et al., 2007Nilsson et al., , 2010Pieterse et al., 2009). Biologically produced H 2 has the most exceptional isotopic composition. Biochemical reactions take place in the aqueous phase and, therefore, the isotopic composition of biologically produced H 2 should reflect the thermodynamic isotope equilibrium between H 2 and H 2 O. Bottinga (1969) calculated fractionation factors for isotope equilibrium in the system H 2 -water vapour. He predicts ε H 2 -H 2 O values for biologically produced H 2 of −737 ‰ to −693 ‰, relative to the water, in the main biological relevant temperature range between 10 • C and 40 • C. Up to now only few individual measurements have been carried out to experimentally determine the isotopic composition of biologically produced H 2 and confirm the extremely depleted values calculated by Bottinga (1969). Rahn et al. (2002) measured headspace samples from a jar of termites with a value of δD = −778 ‰ at a mixing ratio of 1.8 ppm (parts-permillion, µmole/mole), and δD = −690 ‰ at a mixing ratio of 4 ppm from a water headspace sample taken from an eutrophic pond. For none of these values the isotopic composition of the water was reported and it appears that the equilibrium isotope effect between H 2 and H 2 O has never been experimentally verified. Today recent global modelling studies have incorporated biological sources with an isotopic composition of δD = −628 ‰ (Price et al. 2007;. Although biologically produced H 2 is only responsible for approximately 10 % of the annual global H 2 source (Novelli et al., 1999;Hauglustaine and Ehhalt, 2002;Ehhalt and Rohrer, 2009;Pieterse et al., 2011) the extreme deuterium depletion relative to ambient atmospheric H 2 makes it a quite important contributor to the isotope budget (Price et al. 2007;Pieterse et al. 2011). An increasing demand and anthropogenic biological production of H 2 by e.g., industrial fermentation of biogenic waste material is associated with an expected release to the atmosphere because of leakage during production, storage, transport and use. This may increase the contribution of highly deuterium-depleted H 2 to the atmosphere.
Here we present the first systematic experimental evaluation of the isotope source signature of biologically produced H 2 , which is then compared to the theoretical calculations of Bottinga (1969). We measured the isotopic composition of fermentative produced molecular H 2 in biogas, using different production conditions and substrates. Additionally we investigated H 2 produced from pure cultures of fermentative bacteria (Caldicellulosiruptur saccharolyticus, Escherichia coli, and Clostridium acetobutylicum) and of one N 2 -fixer (Azospirillum brasiliensis). We also measured photosynthetically produced H 2 from the common green algae Chlamydomonas reinhardtii.

Samples from a biogas plant
Samples were provided from a biogas plant in Freising, Germany, where also the experiments were conducted. Experiments were carried out with batch cultures (2 l Merck glass bottles) and continuous cultures (30 l glass container). Both were fed with different substrates from surrounding agricultures such as corn, sunflower, cellulose, grass, wheat or mixtures of these substrates. For both treatments the same inoculum was used. It was provided from a pilot-plant scaled plant (3500 l volume). An overview about used substrates and different treatments is given in Table 1.
The batch cultures consist of 1600 ml inoculum and were fed once with 50 g substrate (organic dry substance, oDS) and incubated at a stable temperature of 38 • C. After 35 days, headspace gas samples were taken with gas tight syringes into evacuated 12 ml glass tubes with an overpressure of approximately 1 bar.
The continuous cultures consist of 30 l inoculum and were fed daily in the morning and incubated at temperatures of 38 • C to 60 • C depending on the treatment. The treatments also differ in the amount of substrate between 2 and 3.5 kg organic dry substance/day (oDS/d). Approximately 4 h after feeding, samples were taken at a syringe port at the fermenter with gas tight syringes into evacuated 12 ml glass tubes with an overpressure of approximately 1 bar.
In total, three samples from batch cultures and 13 samples from continuous cultures were measured (see Table 1). Some samples were measured in duplicate or more. The headspace of pure inoculum was also sampled and measured.
E. coli and C. saccharolyticus were grown with 10 or 20 mM glucose and 0.2 g l −1 yeast extract in the medium as described in Stams et al. (1993). These bacteria were grown in 120-ml vials with 50 ml medium or 250-ml bottles with 100 ml medium, and a gasphase of N 2 / CO 2 (80/20). E. coli was grown at 37 • C and C. saccharolyticus at 70 • C. C. acetobutylicum was grown at 37 • C as described by Nimcevic et al. (1998). The gas phase was N 2 . Gas samples were taken from the cultures at the end of growth by gastight syringes and injected in sterile vacuum vials, previously flushed with pure nitrogen.
12 ml of preincubated A. brasiliensis (strain SP7) was used to inoculate 600 ml of ampicillin medium in a closed 2 l borosilicate bottle. Three replicates and one control were incubated for 5 days at 30 • C. The headspace gas volume was sampled into a pre-evacuated 1 l glass container (NOR-MAG, Illmenau, Germany) sealed with two polychlorotrifluoroethylen (PCTFE) valves.
The green algae C. reinhardtii was cultivated in a sulfatelimited Tris-Acetate-Phosphate (TAP) medium as part of an experiment conducted in Switzerland and described in more detail by Haus et al. (2009). For the batch used in our isotope study, a N 2 headspace technique in a glass bottle was applied. After approximately 8 days of incubation, several ml of headspace gas were extracted using a gastight syringe and injected into a pre-evacuated 1 l glass container of the same type as mentioned above. Synthetic air, further purified from traces of H 2 using a catalyst (Sofnocat 514, Molecular Products, Thaxted, UK) was added (to 1.9 bar total pressure) to dilute the sample and, thereby, making it suitable for H 2 measurements. Initial H 2 mixing ratio measurements were conducted at Empa before transferring the sample to IMAU for detailed H 2 and δD analysis. Results for H 2 mixing ratios of Empa and IMAU are in agreement within the error bars and the direct comparison is not shown here.

Determination of H 2 mixing ratio and isotopic composition
The mixing ratio and isotopic composition of molecular H 2 was determined by using the experimental setup developed by Rhee et al. (2004) and modified as described in Röckmann et al. (2010b). Due to a lack of international isotope standards for H 2 , calibration is a critical issue. Our calibration scale has been described in Batenburg et al. (2011). Samples were measured randomly and within 35 days after collection. The measurements consist of the following steps: (1) The sample is cryogenically separated at −240 • C, which means that all gaseous compounds, with the exception of H 2 and some noble gases, are condensed; (2) The non-condensed fraction of the sample (including H 2 ) is preconcentrated using a 5Å molecular sieve at −210 • C; (3) H 2 is focused on a capillary gas chromatographic column (5Å molecular sieve) and chromatographically purified from remaining contaminants at 50 • C; (4) the D / H ratio of molecular H 2 is determined by continuous flow isotope ratio mass spectrometry using a ThermoFinnigan Delta Plus XL instrument.
The analytical system is designed for measurement of air samples with H 2 mixing ratios in the range of typical atmospheric air samples (e.g., Röckmann et al. 2003Röckmann et al. , 2010bRhee et al. 2006;Batenburg et al. 2011). The samples obtained from the biogas reactor and the individual cultures have extremely high H 2 molar mixing ratios between 10 ppm and 1.4 % (see Table 1), which are outside the measurement range of our instrument. To a certain degree, the analytical system has some flexibility as regards high H 2 mixing ratios because simply smaller samples can be inserted into the sample volume; however, in this study the values were that high that the samples had to be diluted. Two dilution methods were adapted for samples of pure cultures and biogas samples.
Several samples from the pure cultures were expanded into 2 l electropolished stainless steel canisters that are routinely used in our laboratory for airborne air sampling (Kaiser et al., 2006;Laube et al., 2008Laube et al., , 2010 and referenced herein). They were diluted by a factor of approximately 2000 with H 2 -free synthetic N 2 -O 2 mixtures. The mixtures were then measured as normal air samples. This procedure induces errors from the dilution itself (for the mixing ratios) and from unquantifiable blank levels of H 2 in the dilution gas. Another disadvantage is that no reference gases are available in the Table 1. Molecular hydrogen (H 2 ) mixing ratio and δD (vs. VSMOW) from different biogas production treatments and pure cultures. Columns 4 and 5 give the raw (i.e. measured) values for mixing ratio and δD, which are used in the Keeling plot in Fig. 1 region of the extremely deuterium-depleted samples, and the isotope scale has to be extrapolated very far outside the range that was used for calibrating the reference gas (−9.5 ‰ to +205 ‰) . Therefore, for the samples from the biogas plant and the N 2 fixer (A. brasiliensis), a standard addition method was developed. Small amounts of a sample (usually approximately 1 ml) were added manually with a gas tight syringe to air from the laboratory reference air cylinder (H 2 mixing ratio = 546.2 ppb, δD = +71.4 ‰) . For the biogas samples following this procedure the measured mixing ratios after dilution were between 575 ppb and 2510 ppb, and δD values were between +35 ‰ and −535 ‰ (Table 1). This means that in the measurement procedure itself a "Keeling plot analysis" is involved, because the H 2 and HD measured in the isotope ratio mass spectrometer is then a mixture of the well-known reference air and the unknown sample ( Fig. 1 for the biogas samples). The isotopic composition of the original sample is then inferred by extrapolation of the linear fit to the correlation between δD and inverse mixing ratio to 0 (y-axis intercept). On the one hand, this introduces an error from the extrapolation, but on the other hand the measured δD values are much closer to the range that was used for calibration of the reference gas. The manual injection of the reference gas with a syringe leads to a relatively high error for the reproducibility of mixing ratios for the original biogas samples (±4.1 %), whereas the reproducibility for δD is not much worse than for normal atmospheric air samples (±2.4 %), since an error in mixing ratio only changes the location of the mixture on the mixing line, but not the y-axis intercept (see Fig. 1). The error for measurement reproducibility is given as the average of absolute deviation of data from their mean. The isotopic composition of the water used in the incubation experiments was determined by Hydroisotop GmbH, Schweitenkirchen, Germany. investigate whether these differences are significant, but this would be an interesting task for the future. In the absence of further information, it may not be appropriate to simply average the results from this to some degree arbitrary selection of samples to obtain a representative mean. The result from the biogas samples is best constrained, however, this value is determined for a temperature range of 38 • C to 60 • C. Including only biogas samples at 38 • C (inoculum and treatments at higher temperatures are excluded from the Keeling plot, Fig. 1) we end up with a δD of −712 ‰ ± 13 ‰ and a fractionation constant ε H 2 -H 2 O of −689 ‰ ± 20 ‰. This value is our best estimate for a fractionation constant ε H 2 -H 2 O at 38 • C. Although systematic errors in the isotopic scale cannot be excluded due to a lack of international isotope standards, the good agreement with the theoretically calculated value of δD = −695 ‰ (Bottinga, 1969) provides strong support for the validity of our results. Bottinga (1969) also reports the temperature dependence of ε H 2 -H 2 O . We determined this temperature dependence experimentally over the incubation range 45 • C-60 • C with otherwise identical conditions (same inoculum and substrate, 30 % grass, 30 % maize, 40 % cereals). As expected for an enzymatic-catalysed reaction in this temperature range, the mixing ratio of H 2 is increasing with increasing temperatures (Fig. 2a). Figure 2b shows that ε H 2 -H 2 O increases with increasing incubation temperature from −713 ‰ at 45 • C to −680 ‰ at 60 • C, thus, by 2.3 ‰ / • C. Gray diamonds in Fig. 2b indicate the theoretically predicted temperature dependency from Bottinga (1969), which is slightly smaller with 1.4 ‰ / • C over the same temperature range. The measurements show a distinct offset of 28 ‰ at 45 • C reducing to 16 ‰ at 60 • C, relative to the theoretical results over this temperature range (Fig. 2). This offset is slightly larger than our estimated experimental uncertainty and remains at present unexplained. Possible contributing factors in the measurements are the potential errors in the absolute isotope calibration  or a nonlinearity in the isotope scale at very low δD values, which is not obvious from the Keeling plot. Nevertheless, the overall temperature dependence is in good qualitative agreement with the calculations of Bottinga (1969), and we conclude that the experimental techniques are sufficiently advanced now to detect such small changes in the region of very depleted isotope values.
In order to derive a revised δD value for H 2 from biological sources that can be used in global models or isotope budget calculations, we calculated ε H2-H2O at a mean temperature of 20 • C using the measured value at 38 • C and our experimentally determined temperature dependence, yielding a value of ε H 2 -H 2 O = −731 ‰ (±20 ‰). For calculating a global average δD value of H 2 from biological sources we used a global average value of δD of precipitation of δD = −37.8 ‰ (Hoffmann et al. 1998;Bowen and Revenaugh, 2003), and then calculate δD = −741 ‰ (±20 ‰). Using the theoretical temperature dependence of Bottinga, we calculate ε H 2 -H 2 O = −715 ‰ (±20 ‰) and δD = −726 ‰ (±20 ‰). Our experimental values are in a very good agreement to the predicted value by Bottinga (1969) of ε H 2 -H 2 O = −722 ‰, which gives a δD = −733 ‰ (20 • C, δD of precipitation of −37.8 ‰).

Summary, conclusions and outlook
The isotopic composition of biologically produced H 2 was investigated systematically and our measurements confirm the massive deuterium depletion as predicted by Bottinga (1969). Using a Keeling plot analysis, we establish an overall source signature of δD = −712 ‰ (±13 ‰) for biologically produced H 2 , with a fractionation constant of  Bottinga (1969). Results of the measured fit lines are: (a) y = 1990ln(x) − 7453; R 2 = 0.96; (b) y = 2.3x − 817.4; R 2 = 0.98. Note: With respect to enzymatic catalised production of hydrogen, a logarithmic fit is chosen for the relation between temperature and mixing ratio. Note: the error bars given are not representing the reproducibility of the duplicate measurements, but also take into account the general uncertainties by using mean relative errors of 2.9 % for δD and 5.4 % for the mixing ratio.
ε H 2 -H 2 O − 689 ‰ ±20 ‰ between the H 2 and the source water at 38 • C and a δD H 2 O of −73.4 ‰. The temperature dependence of ε H 2 -H 2 O has also been determined, and accounting for the temperature effect the fractionation constant is extrapolated to ε H 2 -H 2 O = −731 ‰ (±20 ‰) at 20 • C. This gives an experimentally received source signature of approximately δD = −741 ‰ (±20 ‰) for biologically produced H 2 at mean temperatures (20 • C) and mean δD of precip-itation (−37.8 ‰). Thus, we suggest using these values in global models, rather than the value of −628 ‰ that has been assumed in recent global model studies (Price et al., 2007;Pieterse et al., 2011).
As expected for a thermodynamic equilibrium, the isotopic fractionation is independent of used substrates in the samples from the biogas plant. Samples from individual microorganism cultures confirm the depletion in general, but show even slightly lower δD values; whereas H 2 produced from a nitrogen fixing species had slightly higher δD values. These differences could be caused by extremely high mixing ratios and dilution effects, but this needs further detailed investigation.
Due to its extreme deuterium depletion, biological H 2 , thus, has a high leverage in the global atmospheric H 2 isotope budget. Biological H 2 accounts for only ∼ 10 % of the total H 2 source, but this fraction is depleted by ∼ 772 ‰ relative to the ambient reservoir of ∼ +130 ‰ (note that δ values do not add linearly), so including this source or not makes a huge difference of > 70 ‰ in the atmospheric isotope budget.
The new results imply that the δD values of biological H 2 are distinctly lower than what was included in the two recent global model studies of δD H 2 (Price et al., 2007;. First studies by Pieterse et al. (2011) included a sensitivity test for a change in the isotope source signature from −628 ‰ to −700 ‰ and found that this would change the atmospheric isotope budget by −4 ‰. The demand and production of biologically produced H 2 is expected to increase in the future, and a small increase in the production and release to the atmosphere of e.g., 1 Tg yr −1 would lead to an observable decrease in δD in atmospheric H 2 and can influence the global isotope budgeting.
Despite the large advance in H 2 measurement techniques, the isotopic signature of this gas is still challenging to measure. Intercomparison experiments within the European "EUROHYDROS" project reveal mean deviations of < 1 % in H 2 and provide confidence in the reproducibility of the mixing ratios , but a lack of international isotope standards could still cause systematic errors in the isotopic scale. This should be taken into account when interpreting isotopic data.