Abstract
A large amount of CO2 is stored in the deep waters of Lake Nyos, a volcanic crater lake in Cameroon. The lake is meromictic and thus anoxic in the deeper areas, where dissolved iron exists as Fe2+. Since 2001, a controlled degassing of the lake has been underway. The degassing brings deep water containing Fe2+ to the lake surface as a fountain. This resulted in the formation of Fe(OH)3 precipitates and turned the lake surface red-brown. This coloration was accelerated after the addition of two further degassing pipes in 2011. The Fe(OH)3 precipitates sink to deeper parts of the lake, re-dissolve and are reduced back to Fe2+, which is thought to be precipitating as siderite.
The rates of dissolution and precipitation of siderite in Lake Nyos were examined. Fragments of siderite crystals were covered with gel (Epikote) and placed in lake water for 54 h at several depths characterized by chemical features. The change in the thickness of the crystal surface before and after the reaction was analysed by an interferometer in the laboratory. The siderite dissolves at a rate of −0.09 µm a−1 for samples placed at a depth of 50 m, whereas precipitation took place in deeper waters. The precipitation rate of siderite increased by 0.29 µm a−1 with an increase in depth from 100 to 200 m, and decreased to 0.25 and 0.09 µm a−1 at depths of 208 and 210 m, respectively. However, the calculated saturation indices of siderite in the lake waters increased with depth. The objective of this study is to examine the observed rates of dissolution and precipitation of siderite and to compare them with the saturation index of siderite calculated from the chemistry of the lake water. We also discuss the change in Fe species from the oxidation of Fe2+ to Fe3+ at the surface and reduction to Fe2+ and consequent precipitation as siderite in deeper parts of the lake by degassing of CO2 from bottom water in Lake Nyos.
In August 1986 a large amount of CO2 was released from Lake Nyos, NW Cameroon, killing close to 1800 people (Sigurdsson et al. 1987; Sigvaldason 1989). This limnic eruption was caused by the excessive accumulation of magmatic CO2 in the deep waters of the lake followed by a sudden outburst of dissolved gases (e.g. Kling et al. 1987; Kusakabe et al. 1989). After the disaster, much research was conducted in the domains of geology, geochemistry and geophysics (Sigvaldason 1989). The temporal and spatial variation of CO2 concentrations in the lake is an important parameter for forecasting a potential limnic eruption at the lake, thus the CO2 content has been continuously monitored since 1986 (e.g. Kusakabe et al. 2008; Kusakabe 2015). In 2001 an artificial degassing system was installed to remove CO2 from the lake. Since the rate at which gas was removed by a single pipe was too slow to reduce the gas content to a safe level within a short period of time (Kusakabe et al. 2008), two more degassing pipes were added to accelerate the rate of gas removal from the lake. The degassing brings deep water that contains Fe2+ to the lake surface as a fountain. The oxidation of Fe2+ on contact with atmospheric oxygen resulted in the formation of Fe(OH)3 precipitates and turned the lake surface red-brown (Fig. 1).
Lake Nyos (photo 17 December 2011). Oxidation of Fe2+ upon contact with atmospheric oxygen resulted in the formation of Fe(OH)3 precipitates and turned the lake surface red-brown.
Although the geochemistry of Fe species in Lake Nyos has been reported (e.g. Bernard & Symonds 1989; Teutsch et al. 2009), the detailed behaviour of iron species, and thus the hydrogeochemical controls on the provenance, mobility and distribution of siderite in the lake, is little known. The purpose of this study is to investigate the rate of siderite precipitation in CO2-rich Lake Nyos and to discuss the change in Fe species caused by degassing of CO2 from the bottom waters of Lake Nyos.
Methods of study
Field methods
The rates of siderite precipitation
The dissolution and precipitation rates of siderite were determined by using the techniques described by Satoh et al. (2007) and Ozawa et al. (2010). A siderite crystal was placed on a glass plate (Fig. 2). Half of the siderite surface was coated with a gel (Epikote resin) to expose 50% of the crystal surface to lake water, with the remaining 50% kept intact as a control. That was done to observe the crystal growth or dissolution in the lake's environment. The glass plate with the crystal was held in a perforated plastic syringe to facilitate interaction with lake water. The syringe was put into the lake water at the depths of 50, 100, 150, 200, 204, 208 and 210 m for 54 h (Fig. 3). When the plastic syringe was retrieved, the crystal was detached from the glass plate and washed with ethanol to ensure complete dehydration of the crystal surface. The samples were kept dry for later observation with interferometry.
Photograph of siderite crystals before and after the reaction with lake water. The height difference between the reacted and unreacted surface was measured using the interferometer in the areas indicated by the squares.
Experimental system to determine the deposition or precipitation of siderite in Lake Nyos. Siderite and calcite crystals were fixed on a glass slide. A plastic syringe with holes is used as the crystal holder. Lake water enters inside the syringe through the holes.
Collection of water samples for major ion and iron measurements
A newly designed fluid sampler was constructed to measure CO2 concentrations in Lake Nyos (Fig. 4). This sampler was modified from a previous design that was used to sample geothermal fluids at high temperature and high pressure without vapour–liquid separation (Ozawa et al. 2010). The sampler consisted of two stainless steel bottles, one small (10 ml in volume for CO2 analysis) and one large (200 ml for the analysis of major chemical components) (Fig. 4). Pressurized nitrogen gas was used to operate pneumatic valves through a capillary tube. The details of the analytical procedure will be published elsewhere.
Water sampling system without gas–liquid separation. The internal pressure can be controlled with N2 gas through a capillary tube on the surface.
Lake water collected in the 200 ml stainless steel bottles was transferred into two 100 ml plastic bottles. The pH, electric conductivity (EC) and oxidation–reduction potential (ORP) of the collected water samples were measured in the field using a pH meter (Shindengen, KS-701), EC meter (Horiba, B-173) and ORP meter (TOA, RM-12P).
Laboratory analyses and treatment
Interferometric measurements of siderite growth
The surface profiles of the siderite were measured with an interferometer before immersing the sample in the lake water. This interferometer is an accurate device for measuring the surface structure (c. 1 nm) of mineral surfaces using a white light (Ueda et al. 2005; Satoh et al. 2007; Ueta et al. 2013). Areas ranging from 100 to 500 µm2 of the crystal surface were selected for measurement. The areas of the control (standard plane) and those (reacted plane) not covered by gel were observed simultaneously and the height difference between them was measured. Three areas in one specimen were analysed for the height differences, as shown in Table 1.
After the reaction with lake water, the siderite crystal was examined under a microscope after the removal of the gel and set in the interferometer. This device observes the surface of a mineral repeatedly under a microscope with white light and, using a laser, provides an interference picture of the same area. The heights of the measured points on the mineral surface (for a 300×300 µm area) were simultaneously measured as digital data (680×512 pixels). The height differences between the surface that reacted with lake water and the area coated with Epikote were observed with the interferometer. As the vertical resolution of this interferometer is c. 1 nm (Satoh et al. 2007; Ueta et al. 2013), comparable to atomic force microscopy (AFM), the monomolecular steps of the crystals are observable directly. After exposing the crystal to lake water for 54 h, the retrieved sample was observed for any changes of the crystal surface. The difference in the surface profiles of the reacted and unreacted (control) areas was analysed. The results were shown in Table 1. The rates of dissolution and precipitation were calculated by using the mean observed height difference and the reaction time. The analytical uncertainty was c. 5–30%.
Dissolution and precipitation rates of siderite in Lake Nyos
Chemical analyses of lake water
The CO2 concentrations in the lake water were analysed at the Analytical Center, Mitsubishi Materials Techno Corporation (Table 2). Water samples were introduced into a 200 ml flask under a vacuum and acidified to expel the total CO2. The extracted CO2 was purified in a vacuum line at −114.5°C using an alcohol slush. The amount of CO2 was then measured manometrically. The volume of extracted CO2 from Lake Nyos was about 60 ml (at standard temperature and pressure) and separated into a smaller aliquot for volume measurement. Duplicate samples were taken to check the analytical uncertainty. The CO2 concentrations obtained were 6072 and 6292 mg l−1 at 180 m depth and 9988 and 9900 mg l−1 at 204 m depth (Table 2). The analytical accuracy of the CO2 concentration using this method was within ±5%. The 13C/12C ratio of the extracted CO2 gas was then determined with a mass spectrometer (VG-Optima) with an analytical accuracy of ±0.2‰.
Chemical composition and concentration of CO2 (aqueous CO2+HCO3−) in water samples from Lake Nyos
The major cations (Na, K, Ca, Mg) and anions (Cl, SO4) in the filtered samples collected in December 2011 were analysed as shown in Table 2. The analysis was made with an ion chromatograph (Metrohm 761 Compact IC) at the Analytical Center, Kyuden Sangyo Co., Inc. The analytical uncertainties were within ±5%. The chemical compositions of the lake water samples collected in March 2012 and analysed at Tokai University are shown in Table 2. At Tokai University anion concentrations were measured with ion chromatography (DIONEX ICS900). Na and K were analysed with an atomic absorption spectrometer (Analytikjena contrAA700) and other cations including Fe with an inductively coupled plasma mass spectrometer (Thermo Scientific iCAPq).
Results and discussions
The dissolution and precipitation of siderite that was investigated with phase-shift interferometry is shown in Table 1 and Figure 5. The rates of precipitation for siderite were calculated using the data in Table 1. The chemical compositions of the lake water samples are shown in Table 2.
Dissolution and precipitation rates of siderite in Lake Nyos. The precipitation rates are calculated from the measured values at three different areas of the same siderite crystal.
Sources of Fe and CO2 in Lake Nyos
Iron (Fe2+) is the dominant cation (c. 13 mmol kg−1) in Lake Nyos (Kusakabe et al. 2008) as shown in Table 2. Figure 6 shows depth profiles of Fe and CO2 concentrations and Eh (electrical potential) in the lake water on 1 March 2012. Both concentrations drastically increase at the first chemocline between 50 and 100 m depth and again at the second one between 204 and 208 m depth. Below the first chemocline, the Eh value decreases from oxidation to reduction states (up to −50 mV), where Fe3+ in the water is reduced to Fe2+. The simultaneous high Fe and CO2 contents in the lake reflect the incidence of water–CO2–rock interactions in and beneath Lake Nyos.
Depth profiles of (a) Fe and (b) CO2 concentrations and (c) Eh on 1 lake water at Lake Nyos. The Fe, CO2 and Eh were measured on 1 March 2012 in this study.
What is going on beneath the Lake Nyos may be a natural analogue of CO2–rock interactions around room temperature – as supported by the work of Kharaka et al. (2009), who report that Fe-rich groundwater similar to the lake water in Lake Nyos is formed during CO2 sequestration into the sedimentary layers of the Frio Fm, USA. In order to examine the mechanism for the formation of Fe-rich fluids and siderite precipitation in the bottom of the lake, the behaviour of Fe in the lake water is discussed below. The rocks in the diatreme beneath Lake Nyos are quartz monzonite and alkali basalts containing peridotite fragments (e.g. Dalrymple & Lockwood 1990; Aka et al. 2001). The mineral compositions of granite xenoliths in pyroclastic materials surrounding Lake Nyos are mainly quartz (50%), plagioclase (20%), biotite (c. 15%), orthoclase (c. 5%), microcline (c. 5%) and accessory minerals such as pyroxene, apatite and opaque oxides (Bilong et al. 2011). Schenker & Dietrich (1986) reported the total iron in granites from the western shore and southern end of Lake Nyos to be 2.85 (the wt% of Fe2O3) and 3.58, respectively. Moreover, they also report that soil from the southwestern delta of the Lake after the gas outburst contained 9.84 wt% Fe2O3.
The possible mechanisms for the formation of Fe-rich water and siderite in Lake Nyos are proposed as follows. In the first, CO2-rich fluid at depth reacted with magnetite (Fe3O4) in the quartz monzonite and leached iron to form siderite by the following reaction:
(1)
Equation (1) shows that CH4 acts as a reducing agent of magnetite to siderite. Under the observed concentration (1.7–3 mmol kg−1) of methane at the bottom of Lake Nyos (Evans et al. 1993; Issa et al. 2013) equation (1) shifts to the right; the reducing condition dominates, thus favouring siderite formation.
Laterite, commonly observed around Lake Nyos, is a soil type in hot and wet tropical regions. It is rich in ferric iron oxide and aluminum oxide. It is conceivable that laterite is deposited as sediment at the lake bottom. Such laterite can act as the source of siderite formation when reacted with CO2-rich fluids (Sigurdsson et al. 1987). The fluid partially dissolves the hematite component (Fe2O3) of the laterite to form siderite by the following reaction:
(2)
The reaction requires hydrogen gas to precipitate siderite. At the bottom of Lake Nyos, methane-reducing bacteria are active and may produce H2 as well as CH4. Recently, H2-producing bacteria were found at the bottom of Lake Nyos (Tiodjio et al. 2014). These results support the reaction in equation (2) for the formation of siderite at the bottom of the lake as the second possibility.
During CO2 sequestration into sedimentary rocks of the Frio Fm, USA, CO2- and Fe-rich groundwaters were observed, withFe concentrations ranging from 0 to 18 mmol kg−1 with increasing HCO3− ion concentrations from 0 to 50 mmol kg−1 (Kharaka et al. 2009). The following reaction was proposed to explain Fe enrichment in the groundwater;
(3)
where acetic acid is the reducing agent. Equation (3) suggests another mechanism that could form the observed Fe2+-rich water in Lake Nyos. However, this reaction is unlikely to occur due to no evidence of a supply of acetic acid to the fluids underneath the lake.
Calcium concentrations in Lake Nyos increases from 7.4 to 63 mg kg−1 with increasing depth (Table 2). Mito et al. (2008) reported an increase in Ca from c. 400 to 800 mg l−1 in sedimentary brines at Nagaoka, Japan, due to the dissolution of calcite during the injection of CO2; however, the Fe concentration only increased from 0.1 to 1.3 mg l−1. In Lake Nyos the Ca in the lake water may originate from rocks through the interaction with CO2-rich fluids beneath the lake bottom. Although the lake water is enriched in Ca, it is still undersaturated with respect to calcite. In contrast, siderite is saturated in the lake water due to high Fe concentrations.
Precipitation and dissolution rates of siderite
Figure 5 shows the observed precipitation and dissolution rates for siderite against water depth. Although siderite crystals dissolve at 50 m depth, it precipitates below 100 m depth. The precipitation rate increases with depth, reaching a maximum value of 0.29 µm a−1 at 204 m depth (Fig. 5).
The precipitation and dissolution of minerals in a solution can be evaluated by the use of saturation index (SI) defined by
(4)
where Q and K are the activity product (aFe2+×aCO32–) and solubility product ((Fe2+)(CO32–)), respectively, of a mineral – in this case, siderite. If the SI>0, the water is saturated with respect to the mineral, if it is <0, the water is undersaturated with respect to the mineral. The chemical code PhreeqC (Parkhurst & Appelo 1999) was used for the calculations. The SI values for siderite (FeCO3) were related to the dissolution or precipitation rates (Table 2). The siderite precipitation rate is expected to increase with depth as the SI increases. The results are presented in Figure 7. Siderite becomes saturated at depths greater than 50 m. Figure 8 shows the relationship between the siderite SI values and the rate of precipitation in the lake. The plot suggests a tendency for the rate of siderite precipitation to increase with the SI value except at 50 m, where siderite dissolves at a rate of −0.09 µm a−1. This can be translated to 9.7×10−15 mol cm−2 s−1 (−14.0 in log scale) for the whole surface of the crystal.
Saturation indices of siderite v. water depths.
Saturation index (Log (Q/K)) v. precipitation rate for siderite. The precipitation rates were obtained at different depths. The saturation index is calculated by using chemical data in Table 2.
Effect of controlled degassing on siderite reactions
The chemical compositions of major dissolved species are shown in Table 2. After installation of two additional degassing pipes in 2011, the surface water of Lake Nyos turned red-brown due to the formation of Fe(OH)3 precipitates in surface waters, as shown in Figure 1. The precipitates were formed by oxidation through reactions between atmospheric oxygen, siderite and Fe2+ that was initially dissolved in deep water. The oxidation reaction of siderite is represented by equation (5).
(5)
Ford & Pedley (1996) and Jones & Peng (2012) indicated that these precipitates are commonly found in CO2-rich hot springs and in production wells of geothermal plants in the world (see also Corsi 1986; Yoshida 1991; Gallup 2009).
In December 2011 and March 2012 the lake water was clear from 100 to 210 m depth. Fe(OH)3 can be assumed to dissolve into the lake water by the reverse reaction of equation (5). This model is consistent with our dissolution experiments, which found that siderite dissolves at the depth of 50 m depth as seen in Figure 5 and precipitates at 100 m depth and deeper. Above the chemocline at c. 50 m depth the lake water is dominated by the input of river and rain waters. Lake water below this depth is almost in a steady state (Kusakabe et al. 2008). Dissolved oxygen data from the lake water (Evans et al. 1994) show a reducing condition below the chemocline. These results mean that Fe(OH)3 is reduced to ferrous ion and that some of the Fe2+ is precipitated as siderite.
Teutsch et al. (2009) reported the Fe isotopic compositions of Lake Nyos water, and estimated the fluxes of Fe and water into the lake to be 112 mg l−1 and 18 l s−1, respectively. These values correspond to an annual Fe flux of 1.1×106 mol a−1. Controlled degassing started in March 2001 at Lake Nyos, and Kusakabe et al. (2008) estimated the CO2 flux into the lake to be 1.2×108 mol a−1 based on the change in the amount of CO2 accumulated in the lake from 1986 to 2001. The Fe/CO2 ratio in the water recharging the lake is calculated to be 9.1×10−3 ((1.1×106 mol a−1)/(1.2×108 mol a−1)). The observed Fe/CO2 ratio at the bottom (208 m in 2001) of the lake water was 9.6×10−3 (3.21 mmol kg−1/333 mmol kg−1; Kusakabe et al. 2008), which agrees with the estimate in this study. This ratio remained unchanged before the controlled degassing of the lake. The ratio started to increase after the degassing started and the CO2 concentration reduced at a higher rate. The pH was 5.5 in the bottom water and c. 7 at the surface after mixing of the bottom and surface waters (Kusakabe et al. 2008). By accelerating the decrease in CO2 concentration in the bottom water with three pipes, a remarkable change in the siderite chemistry of the lake water could occur. This could be because an increase in pH is expected with the decrease in CO2 concentration in the lake, favouring an increase in the SI of siderite. From this viewpoint it is expected that siderite precipitation will increase in the lake in the near future. This phenomenon also needs to be monitored.
Conclusions
The precipitation rates of siderite in the waters of Lake Nyos were measured. The rate increased with depth and reached a maximum value of 0.29 µm a−1 at 200 m. Lake water is undersaturated with respect to siderite at a depth of 50 m, where siderite dissolution was observed. There are two possible mechanisms for siderite to precipitate: the interaction of CO2 with laterite in an environment with high pressure H2 gas, which is produced by reducing bacteria in sediments at the bottom of the lake and the oversaturation of iron extracted from monzonite during the ascent of CO2-rich fluids from beneath the lake bottom.
Acknowledgments
This research is supported by the SATREPS project entitled ‘Magmatic fluid supply into Lakes Nyos and Monoun, and mitigation of natural disasters through capacity building’ by the Japan Science and Technology Agency, the Japan International Cooperation Agency and the Institute for Geological and Mining Research, Cameroon. Special thanks are due to Y. Yoshida (JICA coordinator to the SATREPS project) for his kind arrangements in our field work. The authors also thank to members of IRGM for their help during sampling.
- © 2017 The Author(s). Published by The Geological Society of London. All rights reserved