Sensing Dissolved Methane in Aquatic Environments - American

Article pubs.acs.org/est

Sensing Dissolved Methane in Aquatic Environments: An Experiment in the Central Baltic Sea Using Surface Plasmon Resonance Cédric Boulart,*,†,‡ Ralf Prien,† Valérie Chavagnac,‡ and Jean-Pierre Dutasta§ †

Leibniz Institute for Baltic Sea Research, Warnemünde, Seestrasse 15, 18119 Rostock, Germany Geosciences Environnement Toulouse, UMR5563 CNRS/UPS/IRD/CNES, 14 Avenue Edouard Belin, 31400 Toulouse, France § Laboratoire de Chimie, École Normale Supérieure de Lyon, 46 Allée d’Italie, 69364 Lyon 07, France ‡

S Supporting Information *

ABSTRACT: A new sensor for in situ, real time methane (CH4) measurements in aqueous environments is based on the refractive index (RI) modulation of a sensitive film composed of a polydimethylsiloxane (PDMS) layer incorporating molecules of cryptophane-A. The RI varies according to the amount of CH4 bound to the cryptophane-A in the film and is determined using surface plasmon resonance (SPR). Tests of the sensor in the summer of 2012 reveal the expansive range of conditions of the Central Baltic Sea with CH4 concentrations varying from 5 nM up to a few hundred nanomolar. The sensor showed detection limits down to 3 nM, sensitivity of 6 to 7 × 10−6 RIU/nM, and response times of 1 to 2 min. Best responses were obtained for concentrations up to 200 nM. Side effects (temperature, cross-sensitivity) are reviewed for future improvements to the sensor design. CH4 values are highest in the Landsort Deep up to 1.2 μM at 400 m depth and lowest in the Gotland Deep with 900 nM at 220 m depth. However, variable values in the upper layers indicate higher mixing rates due to currents and wind driven forces in the Gotland Basin compared with almost constant CH4 values in the Landsort Deep.

■

INTRODUCTION

The current global estimates, varying from 5 to 50 GT CH4 per year1 released into the atmosphere, rely on scarce and discrete measurements, using sampling methods that cannot cope with the strong spatial and temporal variability, as found, for example, in the Baltic Sea6,7 or in the Black Sea.8−10 Furthermore, the description of vertical features throughout the water column, that is, gradients of oxygen, salinity or temperature, the identification and the quantification of the underlying biogeochemical processes are challenging from poor-resolution vertical profiles. In situ dissolved methane sensing is possible today through the use of gas-permeable membrane systems associated with specific detection technologies from semiconductors to laser detectors and mass spectrometry. These sensors and analyzers were reviewed earlier in Boulart et al.11 While these techniques offer satisfying performance in terms of detection limits and specificity, their deployments in strong concentration gradients, for example, variations of several tens of nanomolar over a few meters depth, suffer from relatively slow response and time lags due to diffusion through the membrane, governed by Fick’s Law.12

Identification and quantification of the biogeochemical processes involving methane (CH4) in aquatic systems have been of great interest in the oceanographic community but also for climate studies within the global warming context. While marine contribution to the global atmospheric budget is only 2% of the total emission,1 large uncertainties remain on both the magnitude of marine emissions and the influence of the changing climate on the release of large quantities of methane from marine environments, including gas hydrates,2 and marine permafrost.3 Assessing marine emissions of CH4 requires the comprehension and quantification of production processes, from microbially mediated diagenesis of sediment organic matter to abiogenic production of methane by water-rock interactions and destabilization of gas hydrates, oxidation and transfer mechanisms,1 which are difficult to quantify separately. Yet, many questions regarding production rates versus oxidation rates are still unanswered.4,5 Oxidation is indeed the main process controlling the release of oceanic methane into the atmosphere and leading to marine systems being globally under-saturated.1 However, shallow water bodies such as the Baltic Sea may represent a significant source of CH4 to the atmosphere, especially in winter, with over saturations >800% and fluxes >1.1 nM/m2/s.6 © 2013 American Chemical Society

Received: Revised: Accepted: Published: 8582

March 20, 2013 June 25, 2013 July 1, 2013 July 1, 2013 dx.doi.org/10.1021/es4011916 | Environ. Sci. Technol. 2013, 47, 8582−8590

Environmental Science & Technology

Article

Figure 1. Study area and sampling stations in the Central Baltic Sea. The Baltic Sea is connected to the North Sea via the Kattegat Strait. Bathymetric data were downloaded from http://www.ngdc.noaa.gov/mgg/global/.

selective and reversible affinity for CH4.19 The RI is accurately determined using surface plasmon resonance (SPR) spectroscopy.20 This technology has been thoroughly tested under controlled conditions21 but never in conditions of deployment at sea, where some molecules may compete with CH4 for the encapsulation into the cryptophane-A (e.g., H2S, NH4+). Only deployments in a relevant environment, where CH4 is present at various concentrations together with other dissolved gases, can therefore validate this technology. Here, we present a new CH4 concentration data set from the Baltic Sea with the aim of both validating the technology in realistic conditions and determining the suitability of the method to describe CH4 dynamics across chemical and physical gradients. The sensor data were compared to conventional measurements of dissolved CH4 concentrations from water samples and data quality and sensor performance were evaluated. These experiments revealed the very high variability of CH4 in relation with the physical and chemical contexts and the value of a high-resolution (∼1 min response time, ∼6 nM detection limits) in situ dissolved methane sensor for marine studies.

For real-time, in situ measurements, optical sensors are a favorable solution for many chemicals in seawater.13 They are potentially fast and robust and low power requirements make them suitable for long-term deployments on autonomous platforms. Many techniques based on absorption, fluorescence, scattering, and spectroscopic methods are currently under development for nutrients, metals, and physical parameters measurements.14 Gases are, however, more complicated to detect using simple optical setups due to their chemical properties. As for instance, CH4 is a “dark” solute in water since its infrared absorption is masked by the strong O−H absorption of water (e.g., refs 15 and 16). To overcome this issue, one suggestion is based on Raman spectroscopy and surface-enhanced Raman spectroscopy (SERS), using the property of the CH4 molecule to scatter light in a vibrational mode. Recent developments on these techniques are encouraging, although they address only high concentrations of dissolved CH4 so far (30 mM).17,18 An alternative strategy is allowing dissolved methane to diffuse into a specific hydrophobic medium that changes optical properties in the presence of dissolved methane in the water. This is the approach chosen here, based on refractive index (RI) modulation of a modified polydimethylsiloxane (PDMS) layer incorporating molecules of cryptophane-A, which have a 8583

dx.doi.org/10.1021/es4011916 | Environ. Sci. Technol. 2013, 47, 8582−8590


Article

Table 1. Sampling/Measurement Methods Used for Each Station, Showing Detection Limits and Sensitivity of the Sensor during Deploymentsa

a Note that no data are available for Redox06 because of adverse weather conditions. Values given for Redox07 are calculated for CH4 concentrations below 200 nM. Sensor data are not available (n.a.) at Redox10 because of a problem with the sensor during the deployment. Lab data are given for comparison with field measurements.

■

sensor and the data processing can be found in Boulart et al.21 (see also Figure S1, Supporting Information). Before deployments, the SPR methane sensor was calibrated in the laboratory using a modified version of the “gas-tight calibration rig” used in Boulart et al.21 (see Figure S2, Supporting Information). The resulting calibration curve is presented in Figure S3 (Supporting Information). Previous laboratory tests21 showed no salinity influence on sensor response; temperature is an important factor that sensor response must be corrected for. The refractive index of all materials is indeed changing with temperature variations, increasing RI with decreasing temperature. Here, the RI measured during deployments at sea was compensated for temperature induced changes by adding 2.3 × 10−5 RIU/°C. Sampling Methodology and Sample Analysis. Sampling and sensor measurements were performed on board FS Elisabeth Mann Borgese during the REDOX cruise to the Central Baltic Sea (20/08/12 to 06/09/12). In total, 10 stations were occupied (Figure 1), 4 in and around the Gotland Deep (Redox01 to Redox04), 3 north of the Gotland Basin (Redox05 to Redox07), 1 in the Landsort Deep (Redox08) and two stations south of the Landsort Deep (Redox09 and Redox10). Methane measurements and water sampling were carried out using a pump-CTD system developed by IOW together with the Max Planck Institute for Microbiology Bremen in 2001.26 It consists of a submersible CTD (Seabird SBE911+ including SBE43 O2 sensor) with rosette sampler fitted with a pump unit, an underwater camera and a 350 or 600 m long nylon hose, allowing measurements in the water column with continuous water sampling over a depth of 450 m (maximum depth in the Baltic Sea). During sampling, the pumped water flows (2 l/min rate) through the nylon hose directly to the ship’s lab. While a chemical profile is usually limited to 12 or 24 Niskin bottles on the rosette, such a system enhances the sampling resolution through the water column down to 450 m depth (1 m resolution against 5 m for a “normal” CTD-rosette), and therefore enables the study of chemical parameters in strong chemical and/or physical gradients.26 In order to use the SPR methane sensor online with the pump-CTD, a flow-cell was fabricated and attached to the water outlet of the pump-CTD in the ship’s lab via a bypass to reduce the flow to ∼70 mL/min (see Figure S1, Supporting

STUDY AREA The Baltic Sea is a shallow (∼52 m, average depth), marginal, semienclosed sea, located in the northern part of Europe (Figure 1). It is characterized by a succession of basins delimited by submarine sills,7 each of which exhibits different hydrological conditions. The Baltic Sea is connected to the North Sea through the Skagerrak and the Kattegat Straits, which represent the only pathways for seawater input. Salinities, varying from 1 to 2 (Gulf of Bothnia) to 17−20 (Belt Sea),22 are strongly influenced by land and river runoff of freshwater in the northern and eastern parts, while seawater inflows through the western straits control the salinity of the deeper layers.23 The difference of salinities between surface and deep waters causes the presence of a halocline limiting vertical mixing and exchanges. In the central basin, oxygen can only be replenished below the halocline at 60−80 m depth by North Sea water intrusions.24 The high biological productivity consuming oxygen in surface waters together with the rapid consumption of low oxygen inputs in the deep layers lead to a strong decrease of oxygen from below the halocline to the bottom and the formation of a permanent redoxcline in the deepest basins of the Baltic Sea.23,24 The central basins of the Baltic Sea (Gotland Basin and Landsort Deep) show the strongest enrichments in dissolved methane close to the seafloor, up to 1000 nM in the Landsort Deep and 500 nM in the Gotland Deep. 7 However, concentrations decrease rapidly toward the redoxcline, as low as 4−5 nM in the oxic surface waters, corresponding to 120− 200% saturation with respect to atmospheric equilibrium. In those basins, methane is still mainly produced by microbial degradation of organic matter within organic-rich postglacial sediments25 and is consumed by microbial oxidation in the water column.7

■

EXPERIMENTAL SECTION Methane Sensor. The SPR methane sensor was constructed from the “Big Spreeta” chip (Nomadics) as the RI sensor, able to determine the RI of any material with 10−7 refractive index unit (RIU) of precision. To make the Spreeta chip sensitive for methane, a thin (∼100 μm) layer of a PDMScryptophane-A mixture was coated directly on the Au surface by dip-coating. Details on the preparation of the methane 8584



Article

Figure 2. CTD, oxygen, and dissolved methane profiles. Methane data used here are from Niskin sampling. Note that H2S is only available for Redox02 and Redox07, while there was no Niskin sampling for methane across the water column at Redox04 and Redox05. Note different methane scales. Colors of the labels refer to the described parameter, that is, red for temperature, green for salinity, blue for oxygen, black for methane, and orange for hydrogen sulfide. Blue transparent zones indicate the redoxcline.

Sensor Deployment Strategy. Table 1 lists all the stations and the corresponding sampling/measurement methods. The sampling and measurement strategy relied on the identification of depths of interest above, within and below the main chemical gradients (salinity and oxygen), which provided an expansive range of conditions (temperature, salinity, CH4 concentrations) to characterize the behavior of the in situ detection method in a natural environment. Once the relevant depths were chosen from the initial vertical CTD-O2 profile, the sensor was

Information). This setup allowed a continuous acquisition of data over time and depth. Water samples for verification of sensor results were recovered from the pump-CTD and from the Niskin bottles into 250 mL glass flasks and poisoned (using 100 μL of HgCl2). Dissolved methane was later extracted from the water using the purge-and-trap technique back in the lab.27 The methane content was then determined with a gas chromatograph (GC) fitted with a Flame Ionization Detector (Shimadzu GC-2014). 8585



Article

Figure 3. Methane sensor data from the Central Baltic Sea, corrected for the temperature effects. Note that the sensor signal is expressed as CH4 concentrations (nM CH4 as RIU) calculated from the lab calibration curve or as dRIU. For each profile, CH4 concentration from Niskin sampling is given (black dots, nM CH4 as GC). A, B: Step-by-step profiles at Redox02 and Redox09. The signal at Redox09 was not filtered to highlight the low noise level and the stability of the signal over time. At Redox02, the strategy was to record the sensor signal continuously, including when the CTDrosette frame was lowered; while at Redox09, a file was created for each depth to avoid any perturbation because of the variations of flow by lowering the pump-CTD. C, D: Stationary profiles at Redox04 (100 m depth) and Redox05 (70 m depth) respectively. E: Slow-motion vertical profile at Redox07. The sensor signal (black line) is compared to Niskin samples (black dots). For information, oxygen and temperature profiles are also given. F: Comparison of sensor sensitivities during deployments to the lab calibration curve. Note that for Redox07, only concentrations below 200 nM are depicted to show the similar sensitivities in this range of concentrations. The regression line represents the ensemble of all Redox data. 8586



Article

temperature depends on the sun exposure of the part of the nylon hose that stays on the deck. The difference of concentrations observed between the two methods of measurement, particularly at Redox04 and Redox05 (Figures 3C and 3D), can be hence partly attributed to the calculation of the temperature correction. In order to estimate the possible loss of methane by degassing, water samples taken directly from the pump-CTD outlet and from Niskin bottles were compared at Redox03 (Figure 4). Interestingly and contrary to expectations, the

deployed following a step-by-step pattern at Redox02 and Redox09, allowing a time of 20 min at each depth. At stations Redox04 and Redox05, the sensor was left at 70 and 100 m depth respectively for 1 h, while samples were taken whenever an event was indicated by the sensor signal. Finally, at Redox07 and Redox10, the sensor was left acquiring data while the pump-CTD was lowered to the bottom and back to the surface at 0.1 m/s. Such deployments called “slow-motion profile”, aimed at testing the dynamic sensor response. No data are available at Redox10 due to the degradation of the sensing layer during the deployment. Redox03 and Redox08 were used as reference profiles for CH4 concentrations in the central basin of the Baltic Sea, while Redox01 was only validating the experimental setup on the ship.

■

RESULTS AND DISCUSSION Environmental Conditions. Oxygen, temperature, salinity, and methane profiles (Figure 2) illustrate the variability of the physical and chemical parameters in the Central Baltic Sea. As already described in previous studies,7,26,28 we also mapped the characteristic halocline at 50 to 100 m depth (salinity varying from 6 to 12) and the redoxcline below. Concentrations of O2 varied from anoxia in the deep layers up to 330 μmol/L in surface waters. A suboxic zone could be observed below the redoxcline with O2 concentrations of up to 50 μmol/L, especially at Redox01, Redox02, and Redox04. Temperatures, at this period of the year (late summer), ranged from 15 to 16 °C (surface) down to 3.5 °C above the halocline. In the deeper layers, the temperature was stable at 5.5−6 °C. As expected, CH4 concentrations increased from the surface to the bottom up to 900 nM at 200 m depth at Redox07. In Situ Results and Sensor Performance. Methane timeseries and profiles recorded by the sensor are compared with concentrations measured by purge-and-trap/GC (Figure 3). Note that for Redox02 and Redox09 (step-by-step profiles), we only focused on the transition zone between surface and deep waters, as indicated by the oxygen profiles (Figure 2), where the concentrations of methane are highly variable,7 from 20 to 90 nM at Redox02 and 10 to 27 nM at Redox09. The in situ data at all stations clearly show a strong variability of the sensor signal over depth and time, which is correlated to the concentration of dissolved methane (R2 varying from 0.53 to 0.94, Figure 3F). They also validate the stability of the sensing layer over time since the lab calibration was performed a few months before deployments at sea. These results confirm therefore the ability of the sensor to measure dissolved CH4 concentrations in a natural environment with good accuracy and reversibility up to 200 nM. However, some differences between the sensor data and the GC measurements can be observed, especially at Redox05 (Figure 3D), where the sensor indicated concentrations up to 60−65 nM at the end of the deployment, while water samples remained at ∼25 nM. These differences are likely due to a combination of the temperature effect, the cross-sensitivity to other chemicals, degassing during the ascent of water through the nylon hose, and saturation of the sensor at high concentrations (above 200 nM). In the case of Redox05, a technical issue might be responsible for the deviation of the sensor signal and is still under investigation. After each deployment, the sensor signal was corrected for the temperature effect. Because it was not possible to measure the temperature of the water coming from the nylon hose, we used in situ data directly from the CTD system on the rosette. This might induce variations of 1−2 °C as the water

Figure 4. Comparison of CH4 samples from the pump-CTD and from the Niskin bottles.

pump-CTD sampling seemed to exhibit less losses than sampling from Niskin bottles. For the higher methane concentration sample at station Redox03, the sample from the Niskin bottle shows a concentration of 600 nM whereas the sample from the pump-CTD shows a concentration of methane >700 nM. Yet, the reference methodology is not infallible as illustrated by the data from Redox02 (Figure 3A). Concentrations measured by GC in samples taken at 90 and 110 m depth are much lower than indicated by the sensor and disagree with the previous observations in the same area, that showed relatively uniform CH4 profiles over the Eastern Gotland Basin at those depths.7 This is likely due to a problem of degassing during the GC analysis and confirmed by comparing the profile with the one from Redox03 (Figure 2), very close to Redox02 (∼7.5 km). As described by Schmale et al,5 the concentration of hydrogen sulfide (H2S) constantly increases toward the sediment. This is confirmed by the measurements of H2S performed at Redox02 and Redox07 with concentrations reaching 150 μM at 230 m depth (Figure 2). Although these are high compared to concentrations of CH4 (650 nM), it is unlikely that H2S interferes with the detection of CH4 since the sensing layer is intrinsically specific to methane due to the affinity of the cryptophane-A for CH4 and the van der Waals forces established between the host and the guest.29−32 However, high concentrations of ammonium (NH4+), as high as 2 μM, have been reported close to the seafloor in the Central Baltic Sea.33 At this level of concentration, ammonium cations may be better candidates for the encapsulation into the cryptophane-A,19 although no NH4+-cryptophane-A complexation has been observed neither in an organic matrix nor in 8587



Article

aqueous solutions (Dutasta, Pers. Com.). In fact, what has been observed is the encapsulation of ammonium ions (e.g., RNH3), i.e. with lipophilic groups that make these species more adapted for an encapsulation into the cryptophane-A cavity.34,35 In the present case, the substrate (CH4 or NH4+) has to penetrate the polymer (PDMS) to reach the cryptophane-A cavity. Because the CH4 molecule is more lipophilic than NH4+, NH4+ will remain preferably in the aqueous phase. It is therefore unlikely that ammonium will compete for the encapsulation into the cavity. But other unidentified lipophilic chemicals may be present and interfere with the detection of methane. The Redox07 profile (Figure 3E) shows that the sensor underestimates the actual concentrations measured by GC from 100 m depth downward with possible saturation of 200 nM and higher. This saturation limit is about 100 nM lower than the saturation concentration determined in the lab (i.e., 300 nM).21 In fact, one has to consider two parameters, i.e. i) the thermodynamics of the complexation between methane and the host, and ii) the range of RI that can be measured by the Spreeta chip. In the first case, saturation of the cryptophanes will be related to the thermodynamics, e.g. the quantity of methane present in the aqueous solution and the association constant (Ka) since the encapsulation is a reversible process depending only on the equilibrium between the aqueous phase and the polymer. The second parameter is more a technical issue related to the increase of the RI with the increase of CH4 above the measurement range of the Spreeta chip. Further investigations are required to determine precisely the role of these two parameters. Detection limits and sensitivity were calculated for each station where the sensor was deployed and are indicated in Table 1. Detection limits correspond to three times the standard deviation of the signal (3σ), which was determined from the baseline acquired for 15 min before each deployment. Compared with previous tests in the lab, they are about 10−20 times higher (3.45 to 6.64 nM), which can be explained by the increase of the environmental noise in the ship’s lab (light, flow rate, temperature). It is interesting to note that the sensor signal shows a strong variability over time when staying at the same depth at Redox02, as for instance at 130 m, where strong peaks were observed (Figure 3A). One of the reasons may be the simultaneous collection of sample water from the pump-CTD stream for other analyses, causing variations of pressure in the sensor bypass, and hence on the sensing layer. These peaks, when clearly identified, were removed during the signal processing. However, this was not always the case and these peaks may also reflect the strong variability of methane over time, or are related to other issues as for Redox05 (Figure 3D). Sensitivity of the sensor, determined from the correlation between the sensor signal and the actual concentrations (Figure 3F), is in the range of 6 to 7 × 10−6 RIU/nM, depending on the signal-to-noise ratio. In the current configuration (sensing layer thickness of ∼100 μm), the response time (t90) is ∼1 min, which remains in the same magnitude as during the first tests in the lab.21 Response times for decreasing concentrations are slightly higher (∼2 min) but remain of the same magnitude as for increasing concentrations, which is an improvement compared to tests in the lab. Such response times make highresolution profiles of dissolved methane possible, as depicted at station Redox07 despite the problem of saturation (Figure 3E), although it was necessary to reduce the lowering speed of the CTD from 1 to 0.1 m/s in order to meet the response times. To our knowledge, it is the first time that a dissolved CH4

sensor is able to profile the water column at these levels of response time. Methane dynamics in the Central Baltic Sea. Apart from the increase of concentrations from the redoxcline to the seafloor, one of the most intriguing results is the rapid change in CH4 concentrations within the redoxcline observed at stations Redox02, Redox04, and Redox05, while they seem to be more stable at Redox09. To validate, whether the strong CH4 peaks at Redox02 (Figure 3A) are related to pressure change on the sensor layer, we left the sensor acquiring data at the same depth for 1 h at Redox04 and Redox05. Those long-term deployments showed variations from 75 nM up to 120−130 nM at 100 m (Redox04). Concentrations are lower at Redox05 (70 m depth), located north of the Gotland Deep but the range of concentrations expands from a few nM up to 65 nM. In contrast, in the Landsort Deep (Redox09) stationary profiles of methane remained stable over time, without any strong peaks as observed in the Gotland Basin. The variability of concentrations is not surprising since chemical gradients are known for their various biogeochemical processes involving a strong bacterial activity36 but the rapidity of the change in the concentrations has not been observed before. Schmale et al36 identified a group of methanotrophs and calculated a consumption rate due to aerobic oxidation of 0.28 nM d−1 in the redoxcline of the Gotland Deep, extending from 115 to 135 m depth. At Redox04, if we consider the time lapse between the first peak at ∼125 nM and the stabilization of the concentration at ∼85 nM, the consumption rate would be 10 000 times higher. This sharp decrease is likely to be real, as the concentration measurements from water samples do confirm the drop in concentrations. Such high oxidation rates are obviously not realistic and other parameters and/or processes have to be taken into account to explain the rapid decrease in concentrations in the Gotland Deep. In the Baltic Sea, methane is essentially brought to the water column by vertical diffusion from the deep sediments.7,37 The decrease of CH4 concentrations from the bottom toward the suboxic and oxic layers is mainly due to microbial oxidation36 but one can expect that vertical mixing as well as lateral intrusions play a significant role on methane inputs into the redox-zone, hence explaining the variability of CH4 concentrations. As for example, the vertical stratification of the Gotland Basin is frequently disturbed by intrusion of oxygenated waters, as a result of the energy from winds and internal currents.38,39 However, in the Landsort Deep, the transfer of energy to the deep waters is limited because of the presence of the coastal boundary layer.38 This might explain the differences in the variability observed between Redox02 (Gotland Basin) and Redox09 (Landsort Deep). From the sensor data, it appears that concentrations of CH4 in the redoxcline are patchy, and not as uniform as one could expect from discrete sampling. This clearly demonstrates that vertical mixing plays a major role on the transfer of methane to the oxygenated layers, especially in the Gotland Basin where the energy because of the winds and currents is stronger. Finally these data justify a dedicated experiment and a longterm monitoring in the redox-zone to identify the biological and physical processes responsible for such variations of methane concentrations. Summary and Recommendations. This paper demonstrated the validity of a new technology, based on SPR, to measure in situ dissolved methane concentrations in a dynamic 8588



Article

during the EMB06Z1215 cruise, Dr. Olaf Dellwig (IOW, PI of the cruise), Johann Ruickoldt (IOW) for the CTD-rosette and Pump-CTD operations, and Prof. Peter J. Statham and Ambra Milani for their contribution in the interpretation of the vertical CTD profiles at sea. Finally, the authors acknowledge the efforts of Prof. Gregor Rehder and Stine Kedzior (IOW) for setting up and carrying out the PT/GC measurements, and Olaf Dellwig for carrying out the H2S measurements. The authors would also like to thank the anonymous reviewers for helpful comments and suggestions that no doubt improved this manuscript.

environment. The Baltic Sea, characterized by strong physical and chemical gradients, with CH4 concentrations varying from 5 nM up to a few hundred nanomolar below the redoxcline, proved to be a good testing environment for our new sensor. It responded to the variations of concentrations over time and space, with a response time of ∼1 min and a sensitivity of 6− 7.10−6 RIU/nM, confirming previous data from the lab. Detection limits were determined from the signal-to-noise ratio and varied from 3 to 7 nM. Response times can be improved by reducing and optimizing the thickness of the sensing layer, as well as developing specifically designed sensing materials, with greater resistance to degradation. The future version of the SPR-based methane sensor will include in situ temperature correction to avoid any cross-effects of the temperature. One of the main issues we have to address is to determine the cross-sensitivity to other chemicals, which might reduce significantly the sensor’s range of detection. Two ways are currently investigated: (i) a modification in the design of the cryptophane-A to improve the affinity for CH4 and (ii) differential measurements, that is, comparing the sensor’s response to a reference. Finally, it is also possible to increase the concentration range by increasing the SPR sensors measurement range of RI. Beyond the validation of the principle of the sensor, the high variability of methane concentrations in the redoxcline of the Central Baltic Sea with large peaks in the Gotland Basin in contrast to almost constant values ∼15 nM in the Landsort Deep, was demonstrated. These results imply more mixing (stronger influence of currents and wind) in the Gotland Deep compared with the Landsort Deep. Targeted future experiments and measurements will allow determination and investigation of this variability. This is indeed the first time that dissolved methane concentrations were recorded at such high resolution in time and space, which make this type of sensor a valuable tool to understand the biogeochemical cycle of methane in the oceans.

■

■

(1) Reeburgh, W. S. Oceanic methane biogeochemistry. Chem. Rev. 2007, 107, 486−513. (2) Reagan, M. T.; Moridis, G. J.; Elliott, S. M.; Maltrud, M. Contribution of oceanic gas hydrate dissociation to the formation of Arctic Ocean methane plumes. J. Geophys. Res. 2011, 116, No. C09014. (3) O’Connor, F. M.; Boucher, O.; Gedney, N.; Jones, C. D.; Folberth, G. A.; Coppell, R.; Friedlingstein, P.; Collins, W. J.; Chappellaz, J.; Ridley, J.; Johnson, C. E. Possible role of wetlands, permafrost, and methane hydrates in the methane cycle under future climate change: A review. Rev. Geophys. 2010, 48, No. RG4005. (4) Hakemian, A. S.; Rosenzweig, A. C. The biochemistry of methane oxidation. Annu. Rev. Biochem. 2007, 76, 223−241. (5) Schmale, O.; Blumenberg, M.; Kießlich, K.; Jakobs, G.; Berndmeyer, C.; Labrenz, M.; Thiel, V.; Rehder, G. Microbial methane oxidation at the redoxcline of the Gotland Deep (Central Baltic Sea). Biogeosci. Discuss. 2012, 9, 8783−8805. (6) Gülzow, W.; Rehder, G.; Schneider v. Deimling, J.; Seifert, T.; Tóth, Z. One year of continuous measurements constraining methane emissions from the Baltic Sea to the atmosphere using a ship of opportunity. Biogeosci. Discuss 2012, 9, 9897−9944. (7) Schmale, O.; Schneider von Deimling, J.; Gülzow, W.; Nausch, G.; Waniek, J. J.; Rehder, G. Distribution of methane in the water column of the Baltic Sea. Geophys. Res. Lett. 2010, 37, No. L12604. (8) Amouroux, D.; Roberts, G.; Rapsomanikis, S.; Andreae, M. O. Biogenic gas (CH4, N2O, DMS) emission to the atmosphere from near-shore of the northwestern Black Sea. Estuar. Coast. Shelf Sci. 2002, 54, 575−587. (9) Blumenberg, M.; Seifert, R.; Michaelis, W. Aerobic methanotrophy in the oxic−anoxic transition zone of the Black Sea water column. Org. Geochem. 2007, 38, 84−91. (10) Yakushev, E.; Podymov, O.; Chasovnikov, V. Seasonal changes in the hydrochemical structure of the Black Sea redox zone. Oceanography 2005, 18, 48−55. (11) Boulart, C.; Connelly, D. P.; Mowlem, M. C. Sensors and technologies for in situ dissolved methane measurements and their evaluation using Technology Readiness Levels. Trends Anal. Chem. 2010, 29, 186−195. (12) Newman, K. R.; Cormier, M.-H.; Weissel, J. K.; Driscoll, N. W.; Kastner, M.; Solomon, E. A.; Robertson, G.; Hill, J. C.; Singh, H.; Camilli, R.; Eustice, R. Active methane venting observed at giant pockmarks along the U.S. mid-Atlantic shelf break. Earth Planet. Sci. Lett. 2008, 267, 341−352. (13) Moore, C.; Barnard, A.; Fietzek, P.; Lewis, M. R.; Sosik, H. M.; White, S.; Zielinski, O. Optical tools for ocean monitoring and research. Ocean. Sci. 2009, 5, 661−684. (14) Moore, T. S.; Mullaugh, K. M.; Holyoke, R. R.; Madison, A. S.; Yucel, M.; Luther, G. W. Marine chemical technology and sensors for marine waters: Potentials and limits. Annu. Rev. Mar. Sci. 2009, 1, 91. (15) Buerck, J.; Roth, S.; Kraemer, K.; Scholz, M.; Klaas, N. Application of a fiber-optic NIR-EFA sensor system for in situ monitoring of aromatic hydrocarbons in contaminated groundwater. J. Hazard. Mater. 2001, 83, 11−28. (16) Mizaikoff, B. Mid-infrared evanescent wave sensorsA novel approach for subsea monitoring. Meas. Sci. Technol. 1999, 10, 1185− 1194.

ASSOCIATED CONTENT

S Supporting Information *

Figures showing in situ dissolved methane sensor, principle of the dissolved gas calibration facility, and calibration curve obtained with the dissolved gas calibration facility. This material is available free of charge via the Internet at http://pubs.acs.org.

■

REFERENCES

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Author Contributions

The manuscript was written through contributions of C.B., R.D.P., V.C., and J.-P.D. All authors contributed equally and have given approval to the final version of the manuscript. Funding

Spreeta kit was acquired through the RTRA-MAISOE project (STAE Foundation, Toulouse, France) and deployments at sea were funded through IOW and BMBF-DNS TIEFSEE project (contract no. 03SX276B). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors wish to thank the Captain and the crew of the FS Elisabeth Mann Borgese for their help and assistance at sea 8589



Article

(17) Zhang, X.; Shah, N. C.; Van Duyne, R. P. Sensitive and selective chem/bio sensing based on surface-enhanced Raman spectroscopy (SERS). A Collection of Papers Presented at the 3rd International Conference on Advanced Vibrational Spectroscopy (ICAVS-3), Delavan, WI, U.S.A., 14−19 August 2005, Part 1; Elsevier: Bridgewater, NJ, 2006; Vol. 42, p 2. (18) Zhang, X.; Hester, K. C.; Ussler, W.; Walz, P. M.; Peltzer, E. T.; Brewer, P. G. In situ Raman-based measurements of high dissolved methane concentrations in hydrate-rich ocean sediments. Geophys. Res. Lett. 2011, 38, No. L08605. (19) Brotin, T.; Dutasta, J.-P. Cryptophanes and their complexes. Present and future. Chem. Rev. 2009, 109, 88−130. (20) Homola, J.; Piliarik, M. Surface plasmon resonance (SPR) sensors. In Surface Plasmon Resonance Based Sensors; Homola, J., Ed.; Springer: Berlin, 2006; Vol. 4, pp 45−67. (21) Boulart, C.; Mowlem, M. C.; Connelly, D. P.; Dutasta, J.-P.; German, C. R. A novel, low-cost, high performance dissolved methane sensor for aqueous environments. Opt. Express 2008, 16, 12607− 12617. (22) Feistel, R.; Feistel, S.; Nausch, G.; Szaron, J.; ysiak-Pastuszak, E.; rtebjerg, G. BALTIC: Monthly time series 1900−2005. In State and Evolution of the Baltic Sea, 1952−2005; Feistel, R., Nausch, G., Wasmund, N., Eds.; John Wiley & Sons, Inc.: Hoboken, NJ; pp 311− 336. (23) Ulfsbo, A.; Hulth, S.; Anderson, L. G. pH and biogeochemical processes in the Gotland Basin of the Baltic Sea. Mar. Chem. 2011, 127, 20−30. (24) Nausch, G.; Matthäus, W.; Feistel, R. Hydrographic and hydrochemical conditions in the Gotland Deep area between 1992 and 2003. Oceanologia 2003, 45, 557−569. (25) Leipe, T.; Dippner, J. W.; Hille, S.; Voss, M.; Christiansen, C.; Bartholdy, J. Environmental changes in the central Baltic Sea during the past 1000 years: inferences from sedimentary records, hydrography and climate. Oceanologia 2008, 50, 23−41. (26) Strady, E.; Pohl, C.; Yakushev, E. V.; Krüger, S.; Hennings, U. PUMP-CTD-System for trace metal sampling with a high vertical resolution. A test in the Gotland Basin, Baltic Sea. Chemosphere 2008, 70, 1309−1319. (27) Swinnerton, J. W.; Linnenbom, V. J. Determination of the C1 to C4 hydrocarbons in sea water by gas chromatography. J. Gas Chromatogr. 1967, 5, 570−573. (28) Feistel, R.; Weinreben, S.; Wolf, H.; Seitz, S.; Spitzer, P.; Adel, B.; Nausch, G.; Schneider, B.; Wright, D. G. Density and absolute salinity of the Baltic Sea 2006−2009. Ocean Sci. 2010, 6, 3−24. (29) Garel, L.; Dutasta, J. P.; Collet, A. Complexation of methane and chlorofluorocarbons by cryptophane-A in organic solution. Angew. Chem., Int. Ed. Engl. 1993, 32, 1169−1171. (30) Wu, S.; Zhang, Y.; Li, Z.; Shuang, S.; Dong, C.; Choi, M. M. F. Mode-filtered light methane gas sensor based on cryptophane A. Anal. Chim. Acta 2009, 633, 238−243. (31) Sun, P.; Jiang, Y.; Xie, G.; Du, X.; Hu, J. A room temperature supramolecular-based quartz crystal microbalance (QCM) methane gas sensor. Sens. Actuators, B 2009, 141, 104−108. (32) Tao, C.; Li, X.; Yang, J.; Shi, Y. Optical fiber sensing element based on luminescence quenching of silica nanowires modified with cryptophane-A for the detection of methane. Sens. Actuators, B 2011, 156, 553−558. (33) Nausch, G.; Nehring, D.; Nagel, K. Nutrient concentrations, trends and their relation to eutrophication. In State and Evolution of the Baltic Sea, 1952−2005; Feistel, R., Nausch, G., Wasmund, N.; Wiley: Hoboken, NJ, 2008; pp. 337−366. (34) Garel, L.; Lozach, B.; Dutasta, J. P.; Collet, A. Remarkable effect of the receptor size in the binding of acetylcholine and related ammonium ions to water-soluble cryptophanes. J. Am. Chem. Soc. 1993, 115, 11652−11653. (35) Garcia, C.; Humilière, D.; Riva, N.; Collet, A.; Dutasta, J.-P. Kinetic and thermodynamic consequences of the substitution of SMe for OMe substituents of cryptophane hosts on the binding of neutral and cationic guests. Org. Biomol. Chem. 2003, 1, 2207.

(36) Schmale, O.; Blumenberg, M.; Kießlich, K.; Jakobs, G.; Berndmeyer, C.; Labrenz, M.; Thiel, V.; Rehder, G. Aerobic methanotrophy within the pelagic redox-zone of the Gotland Deep (central Baltic Sea). Biogeosciences 2012, 9, 4969−4977. (37) Bange, H. W.; Bartell, U. H.; Rapsomanikis, S.; Andreae, M. O. Methane in the Baltic and North Seas and a reassessment of the marine emission of methane. Global Biogeochem. Cycles 1994, 8, 465− 480. (38) Axell, L. B. On the variability of Baltic Sea deepwater mixing. J. Geophys. Res. 1998, 103, 21667. (39) Holtermann, P. L.; Umlauf, L. The Baltic Sea tracer release experiment: 2. Mixing processes. J. Geophys. Res. 2012, 117, No. C01022.

8590


Sensing Dissolved Methane in Aquatic Environments - American

Recommend Documents