**1. Introduction**

Numerous morphological, molecular, and geochemical biosignatures have been proposed over recent decades in order to identify records of past life in the 'sedimentary archives' of the ancient Earth or Mars [1–5]. Among these signatures, a variety of mineralogical biosignatures formed directly or indirectly by bacterial activity provide records of biogenesis in certain environments [6–10]. Microbially-mediated mineral precipitation happens in hot springs because microbial activities may change the concentration of ions in the micro-environments and provide nucleation sites for mineralization. However, the characterization of a biosignature based on the morphology of a single mineral often needs to be used cautiously because of the possibly confusing abiogenic imitators of biosignature [11]. As it is insufficient to take only one single mineral as a biosignature, a suite of parameters that may consistently indicate a biological origin must be considered [1]. It was recently proposed that the synthetic features of a mineral assemblage, including size, crystallinity, and morphology, could be a reference for a specific environment with a certain microbial community. In a study on the diversity of the crystal habits of gypsum, Tang et al. [12] described various morphologies and sizes of gypsum that uniquely coexisted in a square-meter sized volcanic hot spring on the Kamchatka Peninsula of Russia, and suggested that it was mainly due to the secular interactions

between microbial metabolism and geochemical environments. Pyrite is also one of the most common biogenic minerals that are observed as euhedral or framboidal crystals in sediments or sedimentary rocks [8,13]. Laboratory studies of the crystallization of pyrite take physicochemical parameters, such as temperature, pressure, and ion concentrations into consideration and have established the relationship between morphologies of pyrite crystals and their depositional environments [14–21]. Though the chemical pathway of pyrite formation is still in a debate [14,22,23], the direct and indirect effects of biological processes on pyrite crystallization is commonly accepted [20,24]. Microorganisms obtain energy from the geochemical environment and release metabolic products that may affect the chemical composition of their aqueous environments and initiate subsequent mineralization [25–28]. For instance, coupling to the oxidation of organic matter, sulfate-reducing bacteria enzymatically reduce sulfate to hydrogen sulfide, which further reacts with iron in euxinic environments and leads to the precipitation of iron sulfides [29]. Thiel et al. [24] identified a novel type of microbial metabolism that favors energy conservation by oxidizing S2<sup>−</sup> in FeS to S<sup>−</sup> in FeS2 as a syntrophy coupling to the hydrogenotrophic methanogenesis (4FeS + 4H2S + CO2→4FeS2 + CH4 + 2H2O). Microorganisms not only reduce the activation energy barrier for mineral nucleation, but also offer their cell walls as substrates to facilitate the nucleation of crystals [25,30]. Thus, microorganisms are important agents that may induce pyrite mineralization. However, in most cases, it is known that microorganisms have little control over the specific crystal habit of pyrite [26,31]; nevertheless, the biological processes that induce the precipitation of pyrite crystals and/or their assemblages should still carry information about past ecophysiological environments.

Framboid and euhedron are the two dominant morphologies of pyrite crystals in low-temperature sedimentary environments [32,33]. Pyrite is also one of the dominant iron sulfides in natural high-temperature sedimentary environments, such as deep-sea hydrothermal vents and terrestrial hot springs, which are analogs to the early environments for life on Earth [34–36], or possibly early Mars [6]. Pyrite nanoparticles of irregular sizes and shapes in deep-sea hydrothermal vents are considered important sources of iron for the deep ocean biosphere [37,38]. Framboidal and euhedral pyrite crystals were assumed to be generated as a consequence of microbial sulfate-reduction in active shallow submarine vents [39]. Microorganisms also thrive in hot springs that are usually characterized by extreme conditions such as high temperature and low pH through deep geological time [40–43]. Combined with the active iron, the production rate of biogenic hydrogen sulfide influences the amount of iron monosulfide in the system, which may eventually transfer to pyrite [19,24]. Many studies on pyrite in hot spring sediments have focused on the sulfur isotopic signatures (biogenic or abiogenic) [44]. Few data are available on the textures and crystal morphologies of pyrite and the relation to their microbially-mediated environments [45].

Looking for mineral biosignatures in the sedimentary rocks of Earth and Mars has long been an effort because biogenic minerals have a much higher chance of surviving the changing planetary environments during their multi-billion-year of evolution [6,7,46]. For instance, the single domain magnetite of 35–120 nm produced by magnetotactic bacteria has clear protein-modulated mineralization mechanisms and has very well-defined ecophysiological significance [47]. However, a completely inorganic process can also produce single domain magnetite with the same morphology [11]. Textural structures [2,9,48] or the complexity represented by a set of crystal habits, including morphology and size (e.g., gypsum, [12]), were suggested as a new type of biosignature of possible past Martian life. In this paper, we provide a detailed description of the high diversity of mineralogical features of pyrite identified from the Kamchatka volcanic hot springs, which we suggest to be a signature of microbiologically-mediated mineral deposition.

#### **2. Materials and Methods**

The Kamchatka Peninsula is located in the transition zone between the Eurasian, North American, and Pacific plates, and is one of the most tectonically active regions on Earth, featuring volcanoes and earthquakes [49]. There are 31 active volcanoes and hundreds of craters in Kamchatka, but hydrothermal

activity is mostly located in the central and eastern volcanic zones [50,51]. As the largest living sulfide ore-forming hydrothermal system in Kamchatka, the Uzon caldera (54◦26 –54◦31 N, 159◦55 –160◦07 E) is located in the center of the eastern volcanic zone with thick Paleogene-Neogene sedimentary rocks. It formed after the collapse of the volcanic crater about 40,000 years ago and is underlain by Pliocene volcanogenic sediments [52–54].

Of the hundreds of hot springs in Kamchatka, five were chosen for this study (Figure 1, [55–57]): Burlyashii, Zavarzin, Thermophile, Jen's Vent, and Oil Pool (Oil Pool lacks data on location and chemistry but is in the same area). Based on geochemical data listed in Table 1, Jen's Vent and Burlyashii hot springs have the highest temperatures, while Thermophile has the lowest among the hot springs studied. All hot springs are predominantly in reduced geochemical conditions and with pH values varying at large scales (pH = 4.4–7.5). Concentrations of soluble Fe and S2<sup>−</sup> species of the Burlyashii hot spring are much higher than the other hot springs. The general geochemistry of these volcanic hot springs can be found in Table 1 and Taran [55].

**Figure 1.** Locations of hot springs in Uzon Caldera, Kamchatka Peninsula (after 56,57).


**Table 1.** The geochemistry of Kamchatka hot springs. All concentration values are in mmol/l unless otherwise noted.

Samples were collected with sterilized bottles by researchers from the University of Georgia [56], transported to the University of Hong Kong with dry ice, and stored at −21 ◦C. For scanning electron microscope (SEM) measurements, samples were dehydrated with anhydrous ethanol several times and spread onto silicon chips. The silicon chips were sputtered with gold/palladium for 20 seconds for electron microscopic observation. A Hitachi S4800 SEM in the Electron Microscope Center of the University of Hong Kong was used for morphological and structural characterizations using the secondary electron mode at low voltage (5 kV). Equipped energy-dispersive X-ray spectroscopy (EDS) was used to measure the in-situ chemical composition of each sample to identify minerals based on the primary results of SEM observations.

For Mössbauer spectroscopic measurements, samples were ground to a 200-mesh powder using agate mortar after being freeze-dried. Each sample was mounted onto an acrylic holder (10 mm2) with a 5 mg Fe/cm2 thickness. The 57Fe Mössbauer spectra were collected at room temperature (293 K) in transmission mode with a 25mCi 57Co/Pb source at the University of Hong Kong. The Mössbauer spectroscopic hyperfine parameters were calibrated by the fitted hyperfine parameters of the spectrum of a 25-μm α-Fe film measured after every a few samples.
