REDOX STATUS OF FE IN SERPENTINITES OF THE COAST RANGE AND ZAMBALES OPHIOLITES

Although, the reduced status of the Earth’s upper mantle is a possible controller of the deep, rock-hosted biosphere, knowledge of the redox state of the mantle is incomplete. Peridotites (mantle rocks) are composed of ultramafic (Fe, Mgrich) minerals such as olivine and pyroxene. During serpentinization, water and ultramafic minerals react, generating a package of secondary minerals dominated by serpentine. This releases hydrogen gas in amounts dependent on system geochemistry and largely controlled by the Fe(II) budget in the protolith, as well as other products. Microbial life can be fueled by the hydrogen produced by serpentinization in environments that are generally not regarded as hospitable to life—cool, dark, low energy, subseafloor settings. Peridotite-hosted vents in the seabed and springs in continental ophiolites reveal active microbial communities at work in these distinctive serpentinization-associated waters. In this study, 16 variably serpentinized peridotite samples from the Coast Range Ophiolite (CRO) (11 core samples and one hand sample) and Zambales Ophiolite (ZO) (four hand samples) were selected for study based on mineralogy. The objective of this study was to understand better the redox status of Fe in these rocks and produce possible H2 generation values for the CRO and ZO. Each sample was analyzed using X-ray diffraction and thin sections (when available) to identify possible Fe bearing minerals (olivine, spinel, serpentine, pyroxene, magnetite, other Fe-oxides). X-ray fluorescence was used to obtain the bulk concentration of Fe in each sample (~28,000 to 51,000 ppm (~3.7 to 6.5 wt% FeO)). Mössbauer spectroscopy was used to determine the percentage of total Fe that is Fe 2+ (~23 to 70%), Fe 3+ (~14 to 65%), and magnetite (~0 to 63%), which is a combination of Fe 2+ and Fe 3+ . The data sets were integrated into a hydrogen generation model. I assumed that each sample was representative of the peridotite units of the corresponding ophiolite. This permitted computation of a range of total hydrogen production possible by the peridotite considered, until serpentinization is complete (~900 to 4800 Tmol H2 or ~2000 to 12,735 Tmol H2 if density is factored into the calculation). The CRO can produce less H2 per rock volume than the ZO because the CRO samples generally have a lower Fe concentration, but the CRO has a greater volume and can produce a larger total amount of H2. Variability in bulk rock Fe concentration and Fe valence states in samples taken in close proximity indicate diverse serpentinization reaction paths even in a single ultramafic unit. Tectonics, emplacement history, age, climate, composition, and hydrology of the ophiolite all influence the redox status in the modern, ophiolitehosted ultramafics.


Introduction
As the search for the deep extent of life on Earth and other planets progresses, it fuels the need to better understand life on Earth and the extreme environments in which life is found. Water-rock chemical reactions often provide the energy source(s) these extremophile microbes need. As peridotite is serpentinized, Fe can be oxidized by water, releasing hydrogen gas, which can be used by microorganisms as an energy source.
The objective of this study was to determine the redox status of the Coast Range Ophiolite (CRO) and Zambales Ophiolite (ZO), based on iron speciation in selected mineral phases. Serpentinized peridotite samples were collected from the CRO and ZO, their general mineralogy was identified, Fe concentration data were obtained, and Mössbauer spectroscopy was utilized. Fe was used as the redox status indicator because of its general abundance in the rocks, its multiple valence states, and its participation in energy production during serpentinization. Possible energy yield in the form of hydrogen released during serpentinization was also calculated.

Deep life can be fueled by serpentinization
Hydrogen is an energy source for various forms of life from man to microbes.
It can be produced naturally through water-rock reactions by rusting or oxidizing the Fe within ultramafic (Mg, Fe-rich) rocks. Serpentinization at its most fundamental level is the hydration of Fe 2+ minerals in ultramafic rocks in an overall reducing, anoxic environment. This process can be represented by a redox transformation of mineral-hosted Fe from Fe 2+ to Fe 3+ : ( The hydrogen so generated can be utilized by microbes as hydrogen gas (H 2 ), often with metabolic reactions coupled to cycling of methane (CH 4 ) other hydrocarbons, and/or complex organic compounds that are also produced in this environment [Ophan and Hoehler, 2011;.
Chemosynthetic ecosystems that are fueled at least in part from hydrogen produced by serpentinization have been discovered in/around sites where there is active serpentinization occurring, including submarine seeps from fault-bounded peridotite blocks like the Lost City hydrothermal field on the Atlantis Massif near the Mid-Atlantic Ridge ; within ophiolite groundwater and/or groundwater springs including the Cedars  and CROMO wells  in California, USA, various locations within the Oman ophiolite [Neal and Stranger, 1983;, the Tablelands ophiolite in Newfoundland [Szponar et al., 2013], the Leka Ophiolite complex in Norway , and Gruppo di Voltri in Italy [Cipolli et al., 2003] to name a few. Serpentinization tied to plate convergence and aqueous alteration of mantle wedge material also fosters submarine mud volcanoes such as the Kumano Mud Volcano, Nankai Trough, Japan  and the South Chamorro Seamount near the Marianas Trench, and mud diapirs. See appendix for a brief overview of microbiological diversity of serpentinization-related fluids.
Because the oxidation of available Fe and attendant H 2 production is expected in serpentinization, the redox status of the upper mantle can be considered an indicator of a deep rock-biosphere, at least until temperatures of 121°C, and perhaps 150°C, are reached . Serpentinites are thus one promising environment for studying endolithic extremophiles ] (microbes that live in rocks and off the energy released as the rocks weather). Subsurface lithoautotrophic microbial ecosystems (SLiMES)  can be fueled by the hydrogen produced during serpentization in an environment that is generally not regarded as hospitable to life-cold, dark, possibly high pressure (compared to surface pressure), anaerobic, and extremely alkaline with high pH (8-12) formation fluids and potentially high levels of metals like Fe, Ni, As and Cr.

Serpentinization as a geologic process
Mantle rocks of the subseafloor can be sampled where exposed by fault action or tectonics; diverse dredged peridotite samples from oceanographic expeditions and many seafloor drilling projects have also sampled peridotite. However, drilling through stratigraphically complete lithosphere to mantle peridotite is exceedingly difficult due to the depths (generally 5+ km) of rock that would need to be drilled, generally submarine, where there are additional complications and pressure and temperature constraints. Geoscientists often rely also on specimens delivered from great depth by magma streams as xenoliths [e.g., Pearson et al., 2014]. In this study, tectonics. In a naturally reducing environment, ultramafic minerals can react with water, a process that often starts on the ocean floor, and alter to serpentine assemblages that are stable at the lower temperatures and pressures of Earth's surface.
Serpentinization is the process in which hydrous fluids react with ultramafic rocks to produce serpentine and other alteration minerals, and is a volume increasing (which can further fracture rocks and expose fresh surfaces), exothermic (adds heat energy back into the system to further catalyze the reaction), and, therefore, can be a self-sustaining (positive feedback) reaction. For equation 1 to be viable, the environment needs to be oxygen-depleted and reducing so that Fe is oxidized by oxygen from water while the hydrogen is released. Ultramafic minerals formed at high temperatures and pressures undergo chemical weathering (serpentinization) to reequilibrate with lower temperatures and pressures. Having a heat source (such as magma) nearby, can expedite the serpentinization process, but serpentinization itself is exothermic, thus heating and priming regionally associated rocks for further serpentinization.
Serpentinization is coupled to plate tectonics. Ophiolites that are or were on plate boundaries ( Figure 2) can record subduction-related plate flexure or regions of tension that yield fractured areas with deep faults that allow sea water access to ultramafic rocks. For example, ophiolites may result from obduction near the trench of oceanic lithosphere formed at a mid-ocean ridge (MOR) .
Ophiolites may also form in supra-subduction zone(s) (SSZ) in environments similar to nascent spreading centers in backarc or fore-arc settings, which have distinct oceanic lithosphere geochemical signatures . All formation mechanisms allow for complete or partial ophiolite sequences to be exposed at Earth's surface.
According to the 1972 GSA Penrose Conference, an ophiolite consists of a partial or complete sequence consisting of, from bottom to top, an ultramafic complex (variably serpentinized mantle peridotite), a gabbroic complex, mafic sheeted dike complex, and mafic volcanic complex dominated by pillow basalt overlain by various sedimentary rocks (see Figure 3) . This definition holds true, whatever the tectono-magmatic origin may be. Seawater is also mineral-bound within the ophiolite when it is uplifted, and continues to serpentinize the ultramafic units, slowly exiting through vents/springs, and variably mixing with meteoric surface and groundwater.

Means of inferring serpentinite redox status and relevance to H 2 yield
The redox status of Fe in variably serpentinized peridotite can give clues to the possible hydrogen yield that has occurred and still has the potential to occur because Fe is oxidized by the oxygen in water (equation 1) and the hydrogen is released. because it will substitute for Mg in mineral structures (see Table 1). Molecular water in the mantle is possibly located mostly in the transition zone between the upper and lower mantle Pearson, 2014] and the mantle wedge in subduction zones where the subducting oceanic slab is releasing water into the mantle.

Importance of Al-rich phases, spinels, in H 2 production
Serpentinization can be catalyzed by spinels and is a complicated and highly variable process that depends on the starting composition of the rock and primary minerals, fluid composition, substitutions between Fe, Mg, and other elements , if and how much Al is present in the system , system pressure and/or temperature , and even the presence of microbes could encourage one mineral to form over another . For example, serpentinization needs olivine and water, and responds to additional mineral phases such as spinel and pyroxene. Different  (Table 1), which help illustrate the complex chemistry observed in some serpentinization systems.
Spinel group minerals, especially Al-bearing spinels, can act as catalysts for hydrogen production.   production could also increase with the increased rate of serpentinization, which may allow for economic gain from serpentinization .

Geologic Setting
Ophiolites are generally structurally intact blocks of the oceanic lithosphere (ocean crust and upper mantle, see Figure 3) that are uplifted and deposited on continental crust, or form (essentially in situ) in extensional environments in the back arc. Partial, segmented, and/or mélange ophiolite sequences are common because the emplacement can be complex. The Coast Range Ophiolite is located on modern tectonic plate boundries as illustrated by Figure 2.

Coast Range Ophiolite (CRO)
The Coast Range Ophiolite (CRO) is exposed at various locations in the midwestern portion of California, USA ( Figure 1). The CRO exposures and sampling locations in this study are near Lower Lake (junction of Lake, Napa, and Yolo  Figure A1 for a geologic map of the CRO. The CRO formed in a supra-subduction zone (SSZ) setting . SSZ tectonic settings may include backarc and forearc basins, and arc volcanism in an intra-oceanic plate convergent margin . The CRO formation (rock age and emplacement) is ~163-170 Ma  and was associated with an oceanic plate back-arc basin and continent collision in a convergent margin. The CRO units are often found between the Franciscan Complex (Middle Jurassic or older peridotite wedges ) to the west and the Great Valley Sequence (Middle Jurassic calc-alkalic and mafic pillow lava, sheeted dikes, and gabbro  to the east. The CRO sequence includes mafic rocks, pyroclastic rocks, gabbro, and peridotite . It is often found as mélange. The McLaughlin Natural Reserve is now a research and education-focused space, administered by the UC-Davis and supported by a joint initiative between the Homestake Mining Inc. Co. (currently engaged in closure of a gold mining operation at the site) and local conservation groups.

Zambales Ophiolite (ZO)
The Philippine Islands are a mixture of island arcs and continental fragments . They are bordered by oppositely-dipping subduction trenches (Manila Trench to the west and Philippines Trench to the east) and have a complicated history of uplift and faulting. The Zambales Ophiolite (ZO), Zambales Range, Luzon, Philippines ( Figure 1) is a SSZ ophiolite and was derived from interactions of an island arc system and back-arc basin  during the Cenozoic era , Eocene epoch (33-56Ma) ), which implies that the rock and emplacement age are roughly the same.
The ZO is broadly composed of two blocks of uplifted, tilted, and strike-slip fault shifted lithosphere that have complete Penrose ophiolite sequences  and vary in geochemistry and thickness of the crustal section . The (1) Acoje block to the north is accreted tholeiitic intraoceanic island arc material, while the (2) Coto block to the south is a typical back-arc basin rock series . See the appendix Figure A2 for a geologic map of the ZO.  . Additional data and sample information are reported in the appendix.

Coast Range Ophiolite Samples
Coast Range Ophiolite Microbial Observatory (CROMO) wells were drilled and cores dominated by serpentinite were collected in 2011 .
The CROMO 2 well located in the Quarry Valley area of the McLaughlin Reserve bottoms out at ~45.7 m (~152 ft) depth from surface in a serpentinized peridotite layer . CROMO 2 core samples were taken from ~44.5 m (~148 ft) to ~45.7 m (bottom of hole). CROMO samples were vibrated in distilled/DI water. The water and suspended clay minerals were poured into an Al-foil boat. The bulk of the samples (not suspended in the water), were also placed on an Al-foil boat. Both parts of the samples were dried at 60 o C over night and were analyzed with XRD, XRF, and MOSS. See appendix for data on the suspended clays.
Additionally, the McLaughlin Natural Reserve has archived cores from when Homestake Mining Company was surveying the area. Some of the boreholes passed through hundreds of feet of variably-serpentinized peridotite, some of which were found to contain relict olivine grains. Samples were collected from 3 cores: M81-167, M81-309, and M81-313 at various depths (see Table 2).
Hand samples from the Stonyford Volcanic Complex  were collected as float (cobbles in the creek and not collected in situ) located in the Hyphus-Little Stony Creek confluence near Stonyford, California.

Zambales Ophiolite Samples
Serpentinized peridotite hand samples were collected in September 2012 from the Poon Bato region of the ZO and were analyzed. See Table 2 for latitude and longitude data, and Figure 4 for pictures of the samples before being powdered for analysis.

XRD
X-ray Diffraction (XRD) analysis determined the bulk mineralogy of the samples. A portable Olympus (formerly InXitu) Terra field XRD instrument, with the specifications equivalent to the CheMin tool developed for Mars exploration as described in , was used for all XRD analyses. The Terra engages a Co X-ray source and a cooled charge-coupled device (CCD) detector arranged in transmission geometry with the sample, with angular range of 5 o to 50 o 2θ with < 0.35 o 2θ resolution . X-ray tube voltage is typically 30 kV, with a power of 10 W, a step size of 0.05°, and an exposure time of 10 s per step. Default settings were used except the number of exposures was 1000 (total run time is about 70 min) and the piezo volume was 70.
Samples were powdered using a percussion mortar and/or agate mortar and pestle; when necessary a Dremel manual drill was used to subsample grains of interest. A total of ~9 g of the focus samples were powdered. Powders were passed through a standard 150 μm sieve (100-mesh) prior to analysis. The portions of each sample selected for analysis were not always completely powdered to ≥150 μm, so some of the harder minerals may be under represented in all experiments that required a powdered sample. About 15 mg of powdered material was transferred with a spatula to the inlet hopper of the standard sample vibration chamber, which continuously mixes the powdered sample for the duration of the analysis. Rotation disks (sample is rotated instead of vibrated) were also used for some samples and all the standards.
The resulting diffractograms were interpreted using XPowder software, which is a commercially available peak search-and-match program that queries the PDF2 database for reference mineral peak information. XPowder allows for identification only (not quantification) of major minerals and trace minerals can be easily missed and/or masked by peaks of other minerals. Diffractograms have °2θ on the x-axis and intensity on the y-axis. An intensity peak is the result of constructive interference when Bragg's law (nλ=2d sin θ, where n is the "order" of reflection, λ is the incident X-rays wavelength, d is spacing between atomic planes in a crystal structure, and θ is the incidence angle) is fulfilled by the incoming X-rays.

Thin section microscopy
Thin section (TS) petrography was used to identify relic, accessory, and trace minerals, confirm XRD analysis, and observe relationships between minerals at the micro-scale.  EPA Method 6200, 2007]. An adapted EPA Method 6200 was used. Samples were analyzed using soils mode, which is tuned for quantification of elements common in soils and is biased toward those of specific interest to the lab analyzing soils. There are other modes (metal, mining, consumer goods) generally used with the hand held XRF, but they were not available at time of analysis.
Samples were sieved to 150 μm (#100 sieve) and ~2 mL of bulk sample was analyzed. Samples were run for 200 s and analyzed three times with the sample being shaken/agitated between each run. The three values were averaged and the mean was normalized using USGS standards described below.
To obtain the normalization factor ( The error for XRF is ±24% based on the discrepancy of DTS-2 after normalizing to DTS-1 ( Most of the samples were powdered using a steel percussion mortar, so it is possible that the samples were contaminated resulting in higher Fe concentrations.
This would not account for the higher values of the USGS standards as they were already powdered.

SEM-EDS
SEM-EDS was used to measure the concentration of most major elements in weight percent (wt%) in specific mineral grains from the thin section of sample 313_329. A JEOL 5900 scanning electron microscope (SEM) with Energy Dispersive X-ray Spectroscopy (SEM-EDS) was used on thin sections to obtain high resolution imagery (Figures 13,18,20,22,(25)(26)(27)(28)(29) and spot analysis compositions of major elements in wt% oxides (Tables 7-11) in individual mineral grains and within the groundmass.
The detection limit for SEM-EDS is 0.5wt%, so the elements with lower concentrations have lower accuracy. The SEM-EDS also has limited sensitivity for elements with atomic numbers below 11 (Na)

Mössbauer spectroscopy (MOSS)
The redox status of a system can be broadly determined from the valence state of Fe-Fe 2+ is reduced and Fe 3+ is oxidized-and MOSS data provide Fe 2+ and Fe 3+ as percentages of total Fe. There are limitations to MOSS: it can't distinguish between similar minerals, and results vary as a function of cation substitution and temperature.
There is better peak resolution at colder temperatures with a recommended temperature of 40 K and the magnetic properties can change at temperatures below ~24 K  Spectra were processed using two software packages from the University of Ghent. Simple paramagnetic doublet (Fe 2+ and Fe 3+ ) spectra were modeled using the Dist3e program. More complex spectra with sextets (magnetite) and doublets were modeled using the MEX FielDD program. While modeling, isomer shifts (IS) and quadrupole splitting (QS) were allowed to vary in unison. Peak widths for doublet sets were allowed to vary in unison with IS and QS

Results and Discussion
In general, data from XRD, TS, XRF, SEM-EDS, and MOSS analyses confirm there are important differences in the proportions of mineral phases and Fe valence state in those phases, within a geographic region and also across all samples.

XRD & Thin section petrography
Minerals identified using XRD and/or TS include serpentine, magnetite, other spinel group minerals, pyroxenes, olivine, chlorite, other unidentified clay minerals, brucite, amphibole, and garnet. On XRD diffractograms, brucite and spinels, including magnetite, have overlapping and/or closely spaced intensity peaks that vary in position depending on chemical composition, which can make differentiating these minerals challenging. Spinels (spinel, magnetite, chromite) and brucite look different under a polarizing light microscope. The spinel series in these samples are often reddishbrown or dark brown to black. Spinel is most often in individual, isolated grains that are reddish brown and semi-transparent ( Figure 12, 15, 20, among others). Magnetite and chromite are opaque and appear black (Figure 9,11,14,15,etc…). They were not distinguished from each other and assumed to be magnetite in thin sections. Individual mineral elemental analysis would be needed to distinguish them. Brucite is often colorless or yellowish and found intermixed with serpentine and/or in veins ( Figure   11, 19). Therefore, thin sections were used to confirm XRD analysis, and identify trace or relict minerals that (1) may not have been included in the ~15 mg of powdered rock used in XRD and/or (2) were present in amounts too small to be detected by the XRD. Thin sections also allow petrographic study of the spatial distribution of mineral grains and textures.

Diffractograms
The main two serpentine phases available for peak comparisons in reference databases were antigorite and lizardite. While each sample had peaks corresponding to one or both of these phases, it is also highly likely that other phases of serpentine like chrysotile and greenalite were also present. The serpentine peaks  generally have the highest intensity in the samples followed by peaks for magnetite, spinel, and olivine near 39 to 44°2θ. The intensity indicates diffraction caused by mineral crystallinity and helps in mineral identification because each mineral has a unique diffraction pattern. Figure 6 is the stacked diffractogram for ZO samples.
PHL_3 has the strongest peaks for olivine found in any of the samples. Figure 7 is the stacked diffractogram for CRO CROMO2 samples. CROMO2_3 has an unknown peak with a 2θ of ~31° (i.e., d-spacing of 3.3471Å). The variation of peaks with a 2θ of ~40-45° is from magnetite and spinel variations.
Overall, XRD data indicate that serpentine, magnetite and/or other spinel are present in all the samples. Pyroxene, olivine, chlorite, brucite, amphibole, garnet, and/or other clay minerals such as smectite are identified in various samples.

Thin section and SEM images
Serpentine is the most abundant mineral based on qualitative assessment of mineral abundance in thin sections. Serpentine often forms a square-like mesh or net-like texture on the micro-and macro-scale; interlocking veins of serpentine surround fractured relict (or variably serpentinized) olivine and/or pyroxene, as seen especially in Figures 12,16,19,20,and 23. This mesh texture is often more easily visible in xpl with the grey serpentine surrounding the more brightly colored (higher birefringence) olivine/pyroxenes. Serpentine is also found in larger veins that show variations in color, texture, and/or composition (TS Figures 11a, 16a, and 23a, and SEM-EDS

Mineralogical results summary
Documenting site-specific mineral assemblages is a first step in characterizing the redox status of Fe because Fe occurs in different valence states in different minerals ( Table 1). The mineral assemblages in the samples are similar to those predicted  and similar to those found in serpentinite cores from Ocean Drilling Project (ODP) sites Beard et al., 2009  is expected, to some extent, in these ophiolite samples .
Garnet was found in samples HLSC_1, 313_356, and maybe PHL_3, indicating possible high pressure/temperature conditions, likely in a subduction zone subsurface environment in the mantle wedge  and/or during emplacement at the continental margin. Andradite garnet (Ca 3 Fe 3+ 2 (SiO 4 ) 3 ) was identified by XRD in HLSC_1 and 313_356. Andradite garnet can be formed at low pressure and high temperatures .
Chemical analysis is needed to confirm the XRD analysis. The garnet grains in 313_356 do not appear to be the well-formed, geometric crystals expected if the garnets were neoformed minerals, instead suggesting that they may have started to undergo dissolution and/or other alteration since their original formation. It is possible that these samples were in a relatively shallow area in a subduction zone but still exposed to high temperatures (prograde metamorphism), and then uplifted (retrograde metamorphism) and emplaced. Sample 313_356 seems to have a more complicated history than some of the other samples as indicated by the garnet and lack of serpentine veins or classic serpenitinite mesh textures. It is also possible that Fe and Mg-rich brucite is present in the samples.
The minerals most likely to contain Fe are magnetite, spinel, olivine, pyroxenes, and serpentine.

2 XRF & SEM-EDS
In general, the Fe concentrations found in the samples are similar to those found in other variably serpentinized peridotite (see Table 6, Figure 24, Table A2 for data from literature, and Table A8 for raw XRF data collected). CRO core 313 and the ZO samples have the highest concentrations of Fe ( Figure 24). They are also the samples with more abundant relict olivine and, therefore, are less serpentinized (see 14,[16][17][18][19][20][21]23). This leads to the conclusion that Fe is reapportioned into different host minerals through water-rock reaction, may have left the system in fluid phases (Fe 2+ ) during serpentinization, and/or CRO core 313 and the ZO samples parent rocks had more Fe than the other samples. It is possible that Fe 2+ could have been oxidized, precipitated, and accumulated in areas other than where the samples were collected, perhaps in the more hydrothermal areas of the ophiolites where Au, Ag, S, and other metals also accumulate. CRO core 313 was located in what was a pit mine, so maybe, it was more Fe-rich because more Fe was being deposited, which may also account for the higher percent of Fe 3+ reported in MOSS. that the Fe concentration in the minerals analyzed is ≤0.02wt% (200 ppm) (see section 2.5 for detection limit discussion-0.5 wt%) for olivine (Tables 7, 9, and 10), ≤11.1wt% (111,000 ppm) for an alteration mineral (probably serpentine) (Table 8) next to an olivine grain (  . Besides spinels, Al can also be hosted in olivine, serpentine, pyroxene, and other clay minerals. Al was identified in olivine by SEM-EDS as having ≤5.89 wt% (589,000 ppm), which was surprising because olivine is not usually thought of as incorporating elements with a 3+ valence state into a 2+ slot.

3 MOSS
MOSS parameters (IS, QS, W, A, and B hf ), chi-squared (Χ 2 ), and areas under the curve (Table 12) are similar to those of serpentine and other minerals reported by others (

MOSS analysis of bulk sample
The amount of magnetite in each sample is just a random chance from sampling and the magnetite peaks are fit so they can be removed and look more directly at the silicates, especially serpentine. Magnetite, however, will be discussed in this section in the attempt to gauge the status of the whole bulk rock and not just the silicates, which will be discussed in the next section ( Figure A39).

MOSS analysis of silicates
To better understand the redox state of the system, the magnetite was factored out ( Variations in Fe valence states were expected because of the varying degree of serpentinization, the mineral contents, and environments even within a few feet. It was assumed that the Fe would be oxidized during serpentinization, so it was expected that samples with trace amounts of olivine would have a higher percentage of Fe 3+ than Fe 2+ ; however, the general trend where the more olivine-rich or the less serpentinized rocks had higher %Fe 3+ . The Fe 3+ could be within the pyroxenes, spinel, and serpentine and/or it is possible that when the Fe 2+ was mobilized, it was carried away in the fluid while the Fe 3+ was deposited (recall that Fe 3+ is not soluble in water at high pH, which would apply in this case). As discussed in section 2.7, Adreani et al.
[2013B] did not see a change in Fe concentration over time, so the Fe is probably staying within close proximity to its parent mineral.

4 H 2 modeling
Hydrogen generation was calculated using the normalized average Fe concentrations, the estimated volumes of the ZO and CRO, the valence states of Fe obtained from MOSS with %magnetite divided into %Fe 3+ and %Fe 2+ values, and the simplified serpentinization reaction equation (1). For each sample's hydrogen yield, it is assumed that the Fe concentration for that sample was the same for all of the peridotite units of the CRO or ZO. Table 13 and 14 (utilizes density into conversion from ppm Fe to possible Tmol H 2 (T, Tera=10 12 )) are summaries of normalized Fe concentration in ppm (see appendix Table A7  Explanations and example calculations starting with CRO sample 167_238 Fe concentration (ppm) to possible H 2 (Tmol) already released (total %Fe 3+ ) are below.
Calculation using mg/L to convert ppm to g: The same conversion, only using mg/kg and density (g/cm 3 ): The total possible hydrogen production (Tmol) normalized to a volume of one 1km 3 (Figure 49 and 51) illustrates that even though the overall volume of the ZO is smaller, it can produce more hydrogen per unit volume than the CRO due to the generally higher concentration of Fe ( Figure 24). The average total hydrogen that could be produced-per 1km 3

Implications of rock data and modeling outputs
Possible hydrogen yield over the lifetime of the ultramafic peridotite blocks considered range from ~2848 Tmol to ~4752 Tmol for the CRO and ~908 Tmol to 1029 Tmol for the ZO (when density is use it is ~2028 to 2758 for ZO and ~7634 to 12735 for CRO), and is modeled here as largely controlled by Fe concentration and the volume of the ophiolite. The likelihood of water causing oxidization of Fe 2+ to Fe 3+ and releasing hydrogen is also a factor, but not directly tied into the calculations.
The range of samples with their different Fe concentrations may give an upper and a lower limit of the total hydrogen generation possible from the peridotite units. It is also possible to have hydrogen produced from other mafic and ultramafic layers such as basalt and gabbros, so the actual hydrogen generation from the entire ophiolite could be higher than the simplified model predicts. The more Al-rich a sample is the more hydrogen can be yielded via serpentinization . Additional analysis of the bulk samples considered in this work and separated mineral grains is required to constrain where and how much Al is in the system at the time, to constrain the relative volumes of minerals (% serpentine, olivine, chlorite, magnetite, spinel, etc…) to help gage the extent of serpentinization (parent vs. alteration minerals), to better identify minerals including trace minerals to understand better the chemistry of the system, and to analyze fluid data, if possible, of current fluids in the system including dissolved gases like hydrogen.
Based on TS, XRD, and MOSS, the CRO and ZO are still reactive as both ophiolites still have budgets of Fe 2+ and serpentinization is not complete. However, in areas that are mostly serpentinized, the remaining Fe 2+ budget has been partitioned into minerals like spinel, serpentine, and magnetite that are generally chemically stable at surface conditions and based on the sample set, %Fe 2+ could stay near 50% of the total Fe.

Conclusion
Serpentinites evolve complexly in the natural environment. Serpentinizing systems on Earth can be used as an analog for other terrestrial planets as we continue to search for the limits of life both on Earth and in the universe.

Implications for future research
A more thorough and accurate chemical (XRF, Mg concentration data for both bulk sample and individual minerals) and mineralogical analysis of the samples presented here, along with additional samples, are needed to better (1) constrain the possible hydrogen generation of the CRO and ZO, (2) understand the Fe and Mg concentration, their relationship within the contexts of peridotites, serpentinized or not, and which minerals they are partitioned into, (3) differentiate between spinel group minerals such as spinel, magnetite, and chromite, (4) obtain mineral specific redox status and possible zoning as a marker of the changes the mineral(s) has undergone as the rock and ophiolite unit were formed, and (5) provide markers to look for as we continue the search for life on Earth and other terrestrial planets.    . Quadrupole splitting (QS) is the separation between two component peaks or the difference between two transition energies . The magnetic hyperfine field (B hf ) accounts for the magnetic field created by the Fe. All peaks had a width ≥0.23 mm/s. Tetrahedral site (T) and octahedral site (M) for molecular structure site occupancies.
x x X Shading used to group drill holes and localities; (?) indicates minerals may be present; cpx is clinopyroxene; opx is orthopyroxene; the most common form of brucite identified by XRD was portlandite, which substitutes Ca for Mg. *No thin section for comparison. **No thin section for direct comparison. Thin section from shallower depths (~2-3m) confirms serpentine, magnetite, and other spinel group minerals.
‡MOSS indicates that the Fe is contained in magnetite <1%, which means that in the portion of sample analyzed by MOSS, there was little to no magnetite and the magnetite indicated by XRD could actually be another spinel group mineral.     *Element present is based on a second look at the intensity peaks. Elements labeled as no, did not actually have a peak and the values reported are often zero. Those labeled as "not likely" or "maybe" may have a presence, but the peaks could be counted as noise between other peaks. Those labeled as yes have distinguished intensity peaks. See appendix Figure A19 for peak images.  Table 8 continued.
*Element present is based on a second look at the intensity peaks. Elements labeled as no, did not actually have a peak and the values reported are often zero. Those labeled as "not likely" or "maybe" may have a presence, but the peaks could be counted as noise between other peaks. Those labeled as yes have distinguished intensity peaks. See appendix Figure A20 for peak images.  Table 9: SEM-EDS elemental concentrations for olivine in 313_329. Image is at 250x shows serpentine and magnetite in veins and olivine grains. Pit is from ICPMS laser oblation (data in appendix). See Figure 27 for larger image.  *Element present is based on a second look at the intensity peaks. Elements labeled as no, did not actually have a peak and the values reported are often zero. Those labeled as "not likely" or "maybe" may have a presence, but the peaks could be counted as noise between other peaks. Those labeled as yes have distinguished intensity peaks. See appendix Figure A21 for peak images.  *Element present is based on a second look at the intensity peaks. Elements labeled as no, did not actually have a peak and the values reported are often zero. Those labeled as "not likely" or "maybe" may have a presence, but the peaks could be counted as noise between other peaks. Those labeled as yes have distinguished intensity peaks. See appendix Figure A22 for peak images.   *Element present is based on a second look at the intensity peaks. Elements labeled as no, did not actually have a peak and the values reported are often zero. Those labeled as "not likely" or "maybe" may have a presence, but the peaks could be counted as noise between other peaks. Those labeled as yes have distinguished intensity peaks. See appendix Figure A23 for peak images.    184 Hydrogen values are calculated assuming that all of the peridotite units of the ophiolite have the same composition as the sample. Tmoles (10 12 = tera T) † To be released values use the total % Fe 2+ in sample (calculated using MOSS Fe 2+ + 1/3 of the % magnetite), and it is assumed that the Fe(II) in the system can still react with water to produce hydrogen. ‡ Released values use the total % Fe 3+ in sample (calculated using MOSS Fe 3+ + 2/3 of the % magnetite), and that all of the Fe(III) has reacted with water to produce hydrogen. * Normalized volume to 1km 3 to eliminate volume as a controlling factor of hydrogen production from the ZO and CRO. ** Estimated volume of the peridotite units in the CRO [Area (~3865km 2 )-Carnevale, 2013; depth (~2km) ] and the ZO [Area (~485km 2 )- depth (~3km)-Hawkins and Evans, 1983]. Values in FeO wt% are in the appendix, Table A7. 494 Hydrogen values are calculated assuming that all of the peridotite units of the ophiolite have the same composition as the sample. Tmoles (10 12 = tera T) † To be released values use the total % Fe 2+ in sample (calculated using MOSS Fe 2+ + 1/3 of the % magnetite), and it is assumed that the Fe(II) in the system can still react with water to produce hydrogen. ‡ Released values use the total % Fe 3+ in sample (calculated using MOSS Fe 3+ + 2/3 of the % magnetite), and that all of the Fe(III) has reacted with water to produce hydrogen. * Normalized volume to 1km 3 to eliminate volume as a controlling factor of hydrogen production from the ZO and CRO. ** Estimated volume of the peridotite units in the CRO [Area (~3865km 2 )-Carnevale, 2013; depth (~2km) ] and the ZO [Area (~485km 2 )- depth (~3km)-Hawkins and Evans, 1983] ***Average density (g/cm^3) of variably serpentinized hazburgites and dunites taken from     and the Coast Range Ophiolite (CRO) near Lower Lake, CA, USA . See appendix Figures A1 and A2       conducted theoretical experiments at 500 bars and 350 and 400°C, which predict that serpentinization should yield 10 to 25 vol% of chlorite, considering peridotites of diverse starting compositions.  (Table A12, Figures A40-42). SEM-EDS chemical data for a-d are reported in Tables 7-10. (c) (d) Figure 14. Thin section images of 313_356 in plane polarized light and cross polarized light. Serpentine (serp), magnetite (mag) and/or spinel (sp), pyroxene (px), garnet(?)(gt) and other Fe-oxide veins and around grains are visible. The yellowing of serpentine in ppl may be due to brucite and yellow grains in xpl may be brucite. Scale bar represents 500μm. The garnet (andradite) indicates that this sample may have undergone metamorphism prior to serpentinization because garnet is generally thought of as a higher pressure and temperature mineral. The garnet grains are not well formed crystals, which mean that they could be remnants from the original peridotite body. This sample seems to have a more complicated history than some of the other samples as indicated by the garnet and lack of serpentine veins or classic serpenitinite mesh textures seen in other samples. grains and highly fractured. Some of the serpentine veins cut through the cpx grains and show zoning in plane-polarized light between light-colored and rust-colored (Feoxide) serpentine. Spinel is often located in veins or on the edges of cpx grains and is likely to be alteration minerals that form as the olivine and cpx serpentinize. Detail of a spinel with an interesting shape that appears to be bisecting a cpx grain. The spinel also has inclusions of the cpx grain in legs of the spinel, which indicates that the cpx grain existed before the spinel, which, therefore, is an alteration mineral. The SEM BSE image of the spinel shows micro-cracks in the spinel and variations in its composition (shading differences).     Table 7. Image is at 250x shows serpentine veins and olivine grains.   Table 9. Image is at 250x shows serpentine and magnetite in veins and olivine grains. Pit is from ICPMS laser oblation.  Table 10. Image is at 150x shows serpentine and magnetite in veins and possibly pyroxene grains.  Table 11. Image is at 150x shows a spinel grain surrounded by serpentine and olivine grains. Pit is from ICPMS laser oblation.
McCollom and Seewald (2013), Serpentinites,hydrogen,and life,Elements,9(2), 129-134.        The first 4 runs were shaken between each run and used to gauge accuracy. The last 3 runs (4-6) were not shaken between each run and used to gauge precision.

STANDARDS
Empty sample holder ESH                          Isomer shift (IS) is in mm/s; Quadrupole splitting (QS) is in mm/s; Peak width (W) is in mm/s; magnetic hyperfine field (B hf ) is in tesla; % Area (A) under the curve; CHIsquared (Χ 2 ), and normalized CHI-squared (|Χ 2 |). Silicates are serpentine, pyroxene, and/or chlorite. *Indicates restricted (fixed) parameter. ‡MOSS curves fit by M.Nelms in Dyar Lab and rest were fit by A.Stander. Figure A24. Mössbauer Spectroscopy 309_84 plot. The data (black dots) were fit using the Ghent program to obtain a best fit curve (red), Fe 3+ (blue) and Fe 2+ (green) curves.  Figure A25. Mössbauer Spectroscopy 309_105_B plot. The data (black dots) were fit using the Ghent program to obtain a best fit curve (red), Fe 3+ (blue) and Fe 2+ (green) curves.  Figure A26. Mössbauer Spectroscopy CROMO1_1_A plot. The data (black dots) were fit using the Ghent program to obtain a best fit curve (red), Fe 3+ (blue and green) and Fe 2+ (brown and purple) curves.  Figure A27. Mössbauer Spectroscopy CROMO1_1_B plot. The data (black dots) were fit using the Ghent program to obtain a best fit curve (red), Fe 3+ (blue and green) and Fe 2+ (brown) curves.  Figure A33. Mössbauer Spectroscopy CROMO2_1B plot. The data (black dots) were fit using the Ghent program to obtain a best fit curve (red), magnetite (blue and green), Fe 3+ (purple) and Fe 2+ (brown) curves.  Figure A34. Mössbauer Spectroscopy CROMO2_3B plot. The data (black dots) were fit using the Ghent program to obtain a best fit curve (red), magnetite (blue and green), Fe 3+ (purple) and Fe 2+ (brown) curves.  Figure A38. Mössbauer Spectroscopy PHL_4 plot. The data (black dots) were fit using the Ghent program to obtain a best fit curve (red), Fe 3+ (blue and green) and Fe 2+ (purple, brown, black) curves. PHL_4 still requires additional fitting because the parameters are not within the general range of values for Fe 2+ and Fe 3+ .  % Fe3+ (Table A10) was used in H 2gas released calculations; % Fe2+ was used in H 2gas to be released calculations. Table A12: IC-PMS concentration data for standards and CRO sample 313-329. Standards analyzed include BCR, BHVO, BIR, GOR, StHls, T1, ML3B, KL2, and San Carlos olivine. Data was collected in 4 different areas of the thin section and the mineral grains are labeled accordingly (sp1 for spinel grain in area 1, opx1 orthopyroxene grain in area 1, ol 1 for olivine grain in area 1, cpx2 for clinopyroxene in area 2, etc…). Ol3-2 is probably actually cpx or opx based on the chemical data below and opx3 maybe cpx (exsolution lamelle) due to the high CaO concentrations measured. The raw data was input into the "LazyBoy" version 3.73 macro spreadsheet developed by Joel Sparks (jwsparks@bu.edu; ©2011; version date 2/1/2013). Figures A40-42 illustrate the mineral grains sampled. Concentrations are in ppm unless indicated as wt%. Electron microprobe data of major cations in spinel(s) (Mg, Al, Cr, Fe) are needed to better assess the IC-PMS data, especially in regards to further analysis of the partitioning of the trace elements, such as V which has multiple valence states and may be another way to understand the redox status of rocks ].  Figure A40: Thin section images of sample 313-329 ICPMS area 1 before laser ablation. Areas of analysis are indicated by red circle; spinel 1 (sp1), orthopyroxene 1(opx1), and olivine 1 (ol1). From right to left, images are in plane polarized light (scale bar is ~500μm), cross-polarized light, and reflected light. See Figure 12 for thin section images after laser ablation and general field of view. ol1 opx1 sp1 Figure A41: Thin section images of sample 313-329 ICPMS area 2 before laser ablation. Area of analysis is indicated by red circle; clinopyroxene (cpx2). From right to left, images are in plane polarized light (scale bar is ~500μm), cross-polarized light, and reflected light.

Body of thesis/ main samples
cpx2 Figure A42: Thin section images of sample 313-329 ICPMS area 3 before laser ablation. Areas of analysis are indicated by red circle; orthopyroxene 3 (opx3), and olivine 3 (ol3). Ol3-2 is probably cpx not ol based on chemical analysis and the opx may be cpx based on the high Ca concentration, or an exsolution lamelle. From right to left, images are in plane polarized light (scale bar is ~500μm), cross-polarized light, and reflected light.