New temperature and oxygen fugacity data of Martian nakhlite from Northwest Africa (NWA) 5790 and implications for sulphur degassing depth

Newly analysed titanomagnetite–ilmenite (Tim–Ilm) intergrowths from Martian nakhlite meteorite Northwest Africa (NWA) 5790 yielded crystallisation temperature up to 1032°C and oxygen fugacity (fO 2 ) up to ΔQFM + 1.6, notably higher than previous estimates for nakhlite magmas (temperature < 950°C, fO 2 = ΔQFM-1 to ΔQFM + 1). To interpret how the magma was reduced from ΔQFM-1 to ΔQFM + 1.6, we used D-Compress to model the sulphur degassing process. For fO 2 to signi�cantly decrease in this extended range, the sulphur-rich Martian magma had to degas all the sulphur species at a certain �nal degassing pressure, which was 2–4 bar for NWA 988 and Lafayette and < 0.7 bar for Y-000593 and Nakhla. These �nal degassing pressure data are in good agreement with the Martian nakhlite burial depth estimated by other petrological and geochemical methods. These estimates are also comparable with the excavation depth of ~ 40 m based on the small (6.5 km in diameter) impact crater over the Elysium lava plain. The fO 2 -controlled sulphur degassing pressure may constitute a method for estimating the burial depth of sulphur-rich lava �ows on Mars.


Main Text
Introduction Determining the oxygen fugacity (fO 2 ) of Martian shergottite-nakhlite-chassignite meteorites is important to understand the volcanic eruption process in ancient Mars.Both shergottite and chassignite have relatively reduced fO 2 , ranging from ΔQFM-4 (4 logfO 2 units below Quartz-Fayalite-Magnetite buffer) to ΔQFM-0.5 (Hewins et al., 2020;Udry et al., 2020).However, fO 2 in this range does not appear to represent magma erupted to the surface of Mars due to the high degree of crystallisation of shergottite and chassignite.Nakhlite has a petrographic texture similar to terrestrial basalts, believed to have been excavated from lava piles buried at the shallow surface of Mars (Mikouchi et al., 2006;Shuster and Weiss, 2005;Mikouchi et al., 2012;Righter et al., 2008;Cohen et al., 2017).Therefore, the fO 2 of nakhlite samples can directly provide constraints on lava ows erupted from ancient Martian volcanoes.
There is a considerable variation of fO 2 measured from Fe-Ti oxide phenocrysts in nakhlite samples, ranging from ΔQFM-1 to ΔQFM+1 (Szymanski et al., 2010).However, the actual variation range of fO 2 in nakhlites is still unclear.This is because the equilibrium temperature corresponding to the highest observed fO 2 is less than 950°C, much lower than the crystallisation temperature of clinopyroxene phenocrysts (1100°C, Righter et al., 2008).These low-temperature records may be because most studies applied geothermometers and oxybarometers to ilmenite-titanomagnetite pairs that occurred as veinlets and marginal growth along the cracks, which are considered to be late features (e.g., Szymanski et al., 2010;Treiman and Irving, 2008).There have been reported ilmenite exsolution lamellae intergrowth occurred in Fe-Ti oxide phenocrysts (e.g., Balta et al., 2017;Righter et al., 2014;Imae et al., 2005).
However, none of these intergrowths has been successfully analysed by electron probe microanalysis (EPMA) due to the inadequate thickness of the ilmenite lamellae.
We found titanomagnetite-ilmenite intergrowths in nakhlite NWA 5790, which were wide enough (> 2 μm) to locate a single EPMA spot.Then, we determined the equilibrium temperature and fO 2 .We conducted sulphur degassing modelling based on the newly obtained fO 2 data and the extended variation range of fO 2 .The modelling results helped us constrain the degassing pressure during the crystallisation of Fe-Ti oxide phenocrysts and the emplacement depth (i.e., burial depth) of different nakhlites.

Petrography and P-T estimations
The NWA 5790 sample in this study contains clinopyroxene (55.7%), olivine (6.1%), Fe-Ti oxides (0.4%) and mesostasis (37.8%) (Fig. 1a).As shown in Fig. 1b, the Fe-Ti oxide grains are subhedral to anhedral, with a major axis of about 1 mm and a minor axis of about 0.5-1mm.There are two occurrence domains in the titanomagnetite-ilmenite intergrowths: (1) pervasively distributed exsolution textures characterised by thin ilmenite lamellae occurring in titanomagnetite; and (2) ilmenite veinlets and rim regions along the cracks of titanomagnetite host crystals.EPMA was used to analyse ilmenites from these two domains and their host titanomagnetites.Ilmenite veinlets and rim regions are typically different in composition from lamella, with high TiO 2 , MgO, and MnO contents and low Al 2 O 3 , Cr 2 O 3 , and FeO contents (Table 1).
Ttanomagnetite-ilmenite geothermometer and oxybaromenter (Sauerzapf et al., 2008) were used to calculate the equilibrium temperature and fO 2 of Tim-Ilm lamellae pairs, which are 1005°C to 1032°C and ΔQFM + 1.60 to ΔQFM + 1.44, respectively (Table 1).As shown in Fig. 2a, the equilibrium temperature and fO 2 obtained in this study are remarkably higher than that calculated for Tim-Ilm pairs from other nakhlites, including Nakhla, Lafayette, Y-000593, and NWA 998 (Boctor et al., 1976;Bunch and Reid, 1975;Sautter et al., 2002;Szymanski et al., 2010;Treiman and Irving, 2008).The temperature of 1032°C narrowed the gap between the crystallisation temperature of Fe-Ti oxide phenocrysts and clinopyroxene phenocrysts, which is estimated to be 1100°C (Righter et al., 2008).It also suggests that the clinopyroxene phenocrysts crystallised at relatively high fO 2 close to ΔQFM + 1.60.As shown in Table 1, the veinlet and rim ilmenites in the same sample yield an equilibrium temperature of 848°C to 901°C and fO 2 of ΔQFM + 1.26 to ΔQFM + 1.49, which are consistent with previous studies (Fig. 2a).This comparison con rms that the high-Ti veinlets and rim regions grew at a later stage, in contrast to the ilmenite lamellae.

Sulphur degassing modelling
During the evolution of silicate magma, several processes change the Fe 3+ / ΣFe ratios that control the observed uctuations in fO 2 of nakhlite magma (Kress and Carmichael, 1991).The crystallisation of pyroxene and olivine increases the Fe 3+ / ΣFe ratio in the residual melt (Brounce et al., 2017).Degassing of H 2 and CO 2 can also lead to the oxidation of H 2 O-CO 2 bearing residual liquids (Holloway, 2004;Mathez and Delaney, 1981).However, neither of these processes is eligible to reduce the nakhlite magma.
Sulphur dissolves in silicate melts and takes the forms of sulphates (S 6+ ) or sulphides (S 2− ), while sulphur is present as SO 2 (S 4+ ) or H 2 S (S 2− ) in volcanic gas (Metrich et al., 2009).Therefore, sulphur degassing can lead to either oxidation or reduction of the residual melt phase, depending on the following reactions (Metrich et al., 2009): and Martian basalts have a sulphur content in the range of 4000-7000 ppm, much more abundant than Earth (Gaillard and Scaillet, 2009).Thus, we modelled the effect of sulphur degassing on fO 2 changes in nakhlite magma (Figs.2b and 2c).The modelling was based on a progressive degassing model of the C−O−H−S system using the software package D-Compress (Burgisser et al., 2015), which executes the gas−melt equilibrium model of Iacono-Marziano et al. (2012).Suppose that the system has some constraints (i.e., initial H 2 O and CO 2 contents, magmatic fO 2 ).In that case, based on the experimental calibration of the solubility of H 2 , H 2 O, CO, CO 2 , SO 2 , H 2 S, and S 2 in the silicate melt and the calculation of homogeneous equilibria in the gas phase for the same species, D-Compress calculates the concentration and speciation of C, H, O, and S in coexisting gas and silicate melt as a function of pressure and temperature (Brounce et al., 2017).
In this study, the nakhlite parent magma composition "NPM05" recently proposed by Sautter et al. (2012) was used.The initial pCO 2 was 400 mbar (Franz et al., 2019), the H 2 O content was 1.44 wt.% (Weis et al., 2017), and the initial fO 2 was ΔFMQ + 1.6 at 1100°C.As D-Compress calculates the initial sulphur content of the melt based on implemented solubility laws (Longpre et al., 2017), and sulphur degassing is most sensitive to pressure variation (Gaillard and Scaillet, 2009), the initial sulphur content of the melt (4000-7000 ppm) was controlled by adjusting the pressure.Moreover, calculating the "sulphur content at sul de saturation" (SCSS, O'Neill (2020)) of "NPM05" yielded 4787 ppm, which is in the range of 4000-7000 ppm.Thus, the initial melt sulphur contents used in this study were 4000 ppm, 4787 ppm, 6000 ppm, and 7000 ppm, corresponding to the pressures of 235 bar, 245 bar, 280 bar, and 335 bar, respectively.
The main results of degassing modelling are: (1) In all realistic open and closed sulphur degassing systems, a net decrease is observed in the melt fO 2 upon decompression (Fig. 2c).( 2) Sulphur degassing is predicted to begin after sulphur reaches saturation, but more effective sulphur degassing is achieved only at low pressures (< 50 bar) with a signi cant drop in fO 2 (Fig. 2b).
(3) With increasing initial melt sulphur content, the melt becomes more reduced, and open-system degassing tends to promote this process (Fig. 2b and 2c).( 4) The melt fO 2 can decrease to ΔQFM-1 at < 1bar (Fig. 3), and the fO 2 variation from ΔQFM + 1.6 to ΔQFM-1 can cover all the fO 2 data recorded in nakhlite titanomagnetite−ilmenite intergrowths (Fig. 1).These results support that decompression-driven sulphur degassing from nakhlite magma caused a signi cant reduction.

Implications on burial depth of nakhlite magma
As shown in Figure 2c, fO 2 has a nonlinear positive correlation with pressure.Thus, the minimum fO 2 recorded in the Fe-Ti oxides of nakhlite samples re ects its nal emplacement depth allowing the degassing to proceed to the minimum fO 2 .Natural deduction means that the nal emplacement depth is equal to the burial depth of the meteorite samples in lava piles.Therefore, a further inference from the degassing curve is that the burial depth of nakhlites can be estimated from their minimum fO 2 .
These emplacement depth constraints are broadly consistent with burial depth in previous studies (Mikouchi et al., 2003;Mikouchi et al., 2006;Mikouchi et al., 2012).However, nakhlites are unlikely to reside at a depth of less than 0.4 m (0.04 bar), as 0.04 bar is the upper limit of the Martian atmospheric pressure (Fanale, 1982).Due to insu cient data, the burial depth of NWA 5790 cannot be constrained in this study.However, the pressure range of 235-335 bar obtained by D-Compress re ects the crystallisation depth of phenocrysts (2.4-3.4 km).Thus, a parental melt of nakhlite was formed and began to crystallise ilmenite lamellae in Fe-Ti oxide phenocrysts at a depth of > 2.4-3.4 km.The melt then ascended to a depth of < 40 m or erupted onto the surface, forming late-stage veinlets and rim ilmenites.
The shallow burial depth constrained by fO 2 , coupled with sulphur degassing pressure, indicates that nakhlites originated from the shallow levels of a lava lake or a lava ow at different depths.This occurrence is also consistent with the excavation depth of ~40 m using the small impact crater (6.5 km in diameter) over the Elysium lava plain (Cohen et al., 2017).

Concluding remarks
Based on the newly analysed data of NWA 5790 and the modelling of sulphur degassing, the following conclusions can be reached: (1) The titanomagnetite-ilmenite intergrowth in nakhlite NWA 5790 in this study extended the magma temperature record to 1032°C and fO 2 record to ΔQFM + 1.60, respectively.
(2) The burial depths obtained from the fO 2 -controlled sulphur degassing pressure are consistent with the previously estimated burial depths of various nakhlites, about 20-40 m for NWA 998 and Lafayette and < 7 m for Y-000593 and Nakhla.
(3) Our results suggest that analysis of fO 2 coupled with modelling of sulphur degassing pressure may constitute a method for estimating the burial depth of sulphur-rich lava ows on Mars.