• hydrothermal organic geochemistry
• organic catalysis

Organic compounds are practically ubiquitous in natural hydrothermal environments, in deep sedimentary systems, in subduction zones, at spreading centers, and at continental hot spots. They are critical constituents in the deep branch of Earth’s global carbon cycle (1). Hydrothermal organic reactions affect petroleum formation, degradation, and composition (2, 3), provide energy and carbon sources for deep microbial communities (4, 5), and may be important in the origin of life (6, 7). The essential ingredients that control the chemical reactions of organic material in hydrothermal systems are the organic molecules, hot pressurized water, and associated mineral assemblages. To date, there have been many studies of organic reactions in water at high temperatures and pressures (8⇓–10). Relatively few of these, however, have incorporated the inorganic mineral components present in natural systems (11⇓⇓–14). Furthermore, studies of the ways in which individual minerals control organic reactions at the mechanistic level are virtually nonexistent. Geochemical organic reactions tend to generate complex product mixtures (15, 16), which can obscure the fundamental mechanistic understanding required to establish guiding principles on which to build predictive models of organic reactivity under relevant conditions. Here, we describe an efficient and highly specific catalytic effect of the sphalerite (ZnS) mineral surface on a fundamental process in organic chemistry: carbon–hydrogen bond breaking and making. Sphalerite is a common precipitate in sedimentary exhalative base metal deposits (i.e., black smokers), along with other common sulfides (CuFeS₂, PbS, FeS₂, FeS) (17, 18), and has been the focus of recent origins-of-life investigations (19, 20).

The idea of mineral-surface promoted organic reactions is not new (refs. 6 and 21⇓–23, but see also reviews in refs. 7 and 12) and has been the subject of several recent studies related to the origin of life and to hydrothermal systems. For example, studies have demonstrated that (Ni, Fe) sulfides are necessary to promote reactions of peptides (23) and thioesters (22) at elevated temperatures (∼100 °C); however, these studies did not directly assess the catalytic role of the metal sulfides. Still other studies have shown that experiments with chromium-bearing minerals enhance the production of small hydrocarbons (C₂, C₃) relative to experiments using FeO (21), although they note some enhancement also occurred with FeO alone. Cody et al. present more compelling evidence for metal sulfide catalysis of carboxylation reactions wherein they present estimates of surface site turnovers for a range of metal sulfides that exhibit different product yields (11); these authors note an important question that remains to be answered—specifically, “which catalysts do what”? Mineral catalysis may turn out to be the dominant factor that controls organic transformations under geochemically relevant conditions, and within the organic chemistry community this idea has, to date, been essentially unrecognized. Clarifying the scope of such mineral catalysis and how it occurs at the molecular level is essential to understand organic transformations in a variety of geological environments. The work presented here demonstrates mineral catalysis of the most fundamental component of an organic reaction, the breaking and making of a covalent bond, and finds dramatically enhanced reactivity compared with water alone.

Recent work on hydrothermal reactions of the model alkanes cis- and trans-1,2-dimethylcyclohexane at 300 °C and 100 MPa in water alone, revealed very slow reactions (<5% conversion over 2 wk), and the formation of a complex mixture of isomeric products including alkanes, alkenes, ketones, and aromatic functional groups (24). Key findings of this previous work were that the functional group interconversions were reversible, that aromatic xylenes began to accumulate at the expense of other products at longer reaction times, and that reaction equilibrium had not been attained even on week-to-month timescales.

Hydrothermal reaction of either the cis- or the trans-1,2-dimethylcyclohexane in the presence of sphalerite, however, yields very different results. First, the rate of the reaction is dramatically increased in the presence of the mineral (Fig. 1). Second, essentially only one product is formed: the corresponding stereoisomer (cis- is formed from trans- and vice versa). Small amounts of xylenes are also formed in the reaction with sphalerite, but 2 to nearly 4 times less than what is observed in the water-only experiments, this despite much higher conversions over a similar time period (see Tables S1 and S2 for detailed product distributions). These simple cycloalkanes have relevance beyond our experiments because they not only provide useful stereochemical markers for monitoring reaction pathways but also represent some of the most abundant structures found in hydrothermal systems (13).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Percentage conversion of cis-1,2-dimethylcyclohexane in water alone (blue points), and in the presence of sphalerite (ZnS) (red points), under aqueous hydrothermal conditions (300 °C and 100 MPa). The reaction in water alone generates multiple organic products and never approaches equilibrium over the experimental reaction times. The reaction in the presence of sphalerite forms the trans-stereoisomer almost exclusively, and the conversion of the cis-isomer stabilizes at ca. 74% after 14 d, at which time equilibrium is established. The curve through the sphalerite data represents an exponential fit of the time dependence (see text), the curve through the water-alone data is only for guidance. The specific data values, together with estimated errors, are given in Tables S1 and S2.

We adopt the premise that in order for a mineral-surface promoted reaction to be catalytic, the mineral must not participate as a reactant. Sphalerite increases the rate of reaction, but formation of one stereoisomer from the other does not add or take away any atoms from the sphalerite (i.e., the products of the reaction contain neither Zn nor S); therefore, sphalerite is not a reagent and acts as a catalyst. The stereoisomerization reaction requires either carbon–carbon bond cleavage, followed by bond reformation, or carbon–hydrogen bond cleavage and reformation. To determine which bond-breaking process is responsible for the reaction, experiments were performed with the cis- and trans-1,2-dimethylcyclohexane reactants in D₂O with sphalerite. Deuterium incorporation was found in all of the isomerized products as well as to some extent in the starting material after 24 h (see Fig. S1). The products primarily contained a single deuterium, as did the starting material that was found to incorporate the isotope. These results are consistent with a mechanism in which sphalerite catalyzes breaking of a carbon–hydrogen bond to form an intermediate that can either revert to the same starting structure or form the corresponding isomer by reformation of the carbon–hydrogen bond. The hydrogen that adds to the intermediate is not the original hydrogen atom but must be derived from the solvent, since the products always incorporated at least one deuterium. A schematic illustration of this process is given in Fig. 2 (Lower), which is based on previously studied metal-catalyzed C−H bond-breaking mechanisms (25). Small amounts of products with two deuterium atoms (see Fig. S1) were also observed for longer reaction times, indicating replication of the exchange process with sequential incorporation of a single deuterium. Reeves et al. (26) recently reported deuterium incorporation in alkanes under hydrothermal conditions, which they attributed to the addition of solvent-derived deuterium to the corresponding alkenes observed under equilibrium conditions. The selective incorporation of a single deuterium in the dimethylcyclohexanes shows that formation of an alkene is not necessary for deuterium incorporation in the presence of sphalerite (moreover, alkenes are never detected in our sphalerite experiments), confirming catalytic breaking and making of single C−H bonds in the presence of the mineral.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

(Upper) Plot of percent conversion for hydrothermal reaction of trans-1,2-dimethylcyclohexane over a 24-h reaction period at 300 °C and 100 MPa, as a function of the surface area of added sphalerite (ZnS). Zero surface area represents a water-only experiment. The line represents a linear fit to the data, but no meaning is assigned to the slope. The essentially linear dependence supports the proposed surface-catalyzed reaction mechanism, and is consistent with an excess of reactant molecules compared with surface active sites. The data values for the plot together with estimated errors are given in Table S3. (Lower) Schematic mechanism for ZnS catalyzed C–H bond breaking and making for trans-1,2-dimethylcyclohexane, via surface-bound intermediates, showing exchange of hydrogen with the solvent, and generation of both stereoisomers.

The reaction rate is increased sufficiently in the presence of sphalerite that thermodynamic equilibrium is readily achieved; i.e., the mineral catalyzes the approach to equilibrium. Starting with the cis-1,2-dimethylcyclohexane, the ratio of cis- to trans-isomers attains a constant value of 0.354 by day 14 (Fig. 3). Starting with the trans-isomer, essentially the same cis- to trans- ratio of 0.341 is observed for the same reaction time. The apparent equilibrium constant, Keqapp, must equal the activity product for the reaction: trans ⇄ cis; therefore, if activity is equated with concentration (i.e., assuming activity coefficients for the isomers are equal) Keqapp = [cis]/[trans] = 0.348 (the average experimental ratio). The rate law for approach to equilibrium (Fig. 3) contains both the forward (k) and the reverse (k′) rate constants. Starting with the trans-isomer: d[trans]/dt = −k[trans] + k′[cis]. At infinite time, this expression reduces to [trans]_∞ = [trans]₀ k′/(k + k′) and [cis]_∞ = [trans]₀ k/(k + k′); thus Keqapp= [cis]_∞/[trans]_∞ = k/k′, where ∞ and 0 refer to infinite time and zero time, respectively. Analogous expressions can be written starting with the cis-isomer. The best-fit values to the time-resolved data of Fig. 3 are, k = 0.0078 h⁻¹ and k′ = 0.0215 h⁻¹, yielding: k/k′ = 0.363. The slight difference between Keqapp derived from the kinetic model (0.363) and Keqapp obtained from the measured concentration ratio (0.348) is probably due to the small amounts of xylenes formed in the reaction. The free energy difference between the cis- and trans-isomers is thus determined to be 4.8 kJ mol⁻¹ at the experimental conditions, with the trans-isomer being more stable than the cis-isomer. Simple conformational analysis gives an estimate of the free energy difference between the cis- and trans-isomers of 7 kJ mol⁻¹ at room temperature (27), which suggests the 4.8 kJ mol⁻¹ value is reasonable.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Percentage of the reaction mixture for isomerization of 1,2-dimethylcyclohexane in water with sphalerite (ZnS) at hydrothermal conditions (300 °C and 100 MPa), for reaction starting with the trans-isomer (red symbols/curves), and reaction starting with the cis-isomer (gray symbols/curves), showing approach to the same equilibrium ratio of the isomers. The squares are for the trans-isomer products, and the circles are for the cis-isomer products. The curves represent the best exponential kinetic fit to the entire data set; see text.

Experiments were performed to distinguish between heterogeneous (surface) and homogeneous catalysis mechanisms. Starting with trans-1,2-dimethylcyclohexane, the percent conversion of the starting material over a 24-h time period was found to increase essentially linearly with the mass of sphalerite loaded into the reaction container (Fig. 2). Higher mineral loadings result in more rapid approach to equilibrium, so that at the highest loading, the reaction has almost attained equilibrium after just 24 h. The equilibrium concentrations of the stereoisomers are the same at different mineral loadings; equilibrium is just attained more rapidly with more mineral. The available surface area increases linearly with added mineral, whereas the activity of any mineral-derived dissolved species at equilibrium is independent of the amount of mineral; this observation argues strongly for a surface catalyzed reaction. Nevertheless, we investigated a possible role for dissolved ions. Using data and parameters from refs. 28 and 29, we calculate an equilibrium concentration of aqueous Zn²⁺ that would be present as a result of sphalerite dissolution at 300 °C and 100 MPa to be 4.4 × 10⁻⁶ mol L⁻¹. Experiments performed with no mineral, but in the presence of 0.6, 6.0, and 60 mg L⁻¹ of ZnCl₂, which correspond to 1, 10, and 100 times the calculated equilibrium concentration of Zn²⁺ ions, respectively, gave results that were indistinguishable from water alone, indicating that the sphalerite catalysis was not due to the aqueous Zn²⁺ ions. We acknowledge there could be an effect due to aqueous sulfide species; however, our experimental setup does not allow us to control sulfide speciation during the reaction. However, in experiments with iron sulfide minerals, we did not observe the same results as we did with sphalerite (ZnS); these studies generally resulted in only minimal rate enhancement and very complex product suites. We assume the solubility of Zn sulfides and Fe sulfides under our experimental conditions (300 °C, 100 MPa) are fairly similar, and infer that dissolved sulfide alone is insufficient to generate the catalytic effect we observe with ZnS.

The surface area of the sphalerite used in these experiments, measured via a N₂ Brunauer–Emmett–Teller (BET) isotherm (30), is 12.68 m²·g⁻¹, and so the total mineral surface area available in the experiments shown in Fig. 3 is 0.11 m². A gas-phase energy-minimized structure was calculated for trans-1,2-dimethylcyclohexane using the B3LYP/6–31+G* method [Spartan'08; Wavefunction, Inc. (31)]. Using the molecular dimensions from this structure, the area occupied by a single molecule is estimated to be ca. 28 Ų. The total area occupied by 50 μmol of 1,2-dimethylcyclohexane is ca. 8.3 m². Thus, there are many more reactant molecules than can be accommodated by the mineral surface, which means there must be many more molecules than surface active sites, consistent with the essentially linear dependence on surface area (32). A conservative estimate of the mineral surface area available to the dimethylcyclohexane, as it is probably smaller, is that it is the same as that available to nitrogen. A conservative estimate of the effective area of an active site, as it is probably larger, is the area of a dimethylcyclohexane molecule. From these, we obtain a number of active sites on the mineral of less than 3 × 10¹⁷. The number of reactant molecules is ca. 3 × 10¹⁹; given that approach to equilibrium is reversible, and that the rates of the forward and reverse reactions differ by a factor of only ca. 3 (see above), essentially all of these must react at least once by the time equilibrium is reached, especially when starting from the cis-stereoisomer. This means that each active site must catalyze at least 100 reactions, and probably many more, as equilibrium is approached.

In water at the elevated temperatures and pressures found deep in the crust, where more than 90% of Earth’s carbon-cycling occurs, sphalerite specifically activates a single carbon–hydrogen bond toward reaction in these model alkanes. The presence of the mineral increases the reactivity of the alkanes compared with water alone, to such an extent that studies of the hydrothermal chemistry of such organic structures in the absence of minerals may be completely irrelevant to actual geochemical systems. This work provides mechanistic insight into the ways that minerals can direct the reaction of organic compounds on Earth and, potentially, in other planetary systems. The mechanistic study of mineral catalysis of hydrothermal organic reactions is a new field in geochemistry, which also has implications for green chemistry. The catalysis of carbon–hydrogen bond activation, for example, has been the subject of extensive research, and a wide range of potential catalysts have been synthesized, mostly based on organometallic chemistry (33, 34). Minerals, however, are inexpensive and robust and require no synthesis compared with organometallic catalysts. The results described here suggest that appropriately optimized naturally occurring Earth-abundant minerals may represent a new approach to the development of heterogeneous catalysts for a wide range of organic transformations in the context of green chemistry.