Catalytic syntheses of amino acids and their significance for nebular and planetary chemistry




Abstract– It has been intermittently debated whether some of the organic compounds we find in meteorites, which show a general relationship to interstellar precursors in their isotopic enrichments, could also be formed ab initio from simple gases in nebular and/or parent body processes. Spurred by divergent findings for the organic composition of different stones of the Tagish Lake meteorite, we studied the likelihood of Fisher Tropsch type syntheses of amino acids from CO, H2, and NH3 in the presence of different meteoritic minerals as catalysts and report that amino acids and amines can be produced efficiently under these conditions. Products differed in their molecular distribution depending on the catalyst used, with α-aminoisobutyric acid synthesized preferentially by Murchison and magnetite powders.


Carbonaceous chondrite (CC) meteorites contain complex and diverse suites of soluble organics that differ in their overall composition and relative distributions of molecular components. For example, the organic compounds of Murray-type chondrites (CM) come in several molecular groups characterized by complete isomerism, of up to at least C-7 in some cases, and have their larger abundance in solvent soluble-hydrocarbons (e.g., Pizzarello et al. 2006); in the Renazzo-type (CR) meteorites, on the other hand, it is ammonia, amino acids, and water-soluble species that are predominant, but with a sharp abundance drop for longer carbon chains (Pizzarello et al. 2008; Pizzarello and Holmes 2009).

For the most part, the origin of these meteoritic compounds has been explained by a general theory of formation that involves interstellar as well as parent body processes (e.g., Cronin and Chang 1993). By this hypothesis, icy asteroidal bodies accreted with abundant volatiles including water and deuterium-rich interstellar organics that, upon warming and a subsequent period of aqueous phase chemistry, yielded the various soluble organic compounds of meteorites. Support to this theory comes from the finding of D, 13C, and 15N enrichments in organic compounds of several CMs (Pizzarello et al. 2006) and more recently in CRs. For example, a δD value of +3700‰ was determined for the amino acid isovaline of the GRA 95229 meteorite (Pizzarello et al. 2008), which is comparable to those found for interstellar molecules (e.g., Roueff and Gerin 2003).

The possibility of parent body aqueous syntheses seems also confirmed by several mineralogical (e.g., Bunch and Chang 1980; Zolensky and McSween 1988; Grossman and Brearley 2005) and chemical data; e.g., at least some of Murchison amino acids were probably formed via a Strecker-like reaction of precursor aldehydes and ketones, ammonia, and HCN (Peltzer and Bada 1978), which is supported by the finding in the meteorite of comparable suites of α-amino and α-hydroxy linear acids and the likely byproducts of these syntheses, imino acids (Pizzarello and Cooper 2001).

However, there are isotopic as well as molecular trends within the Murchison organic suite that reveal significant formative distinctions between individual compounds (e.g., Pizzarello and Shock 2010), and no simplified model can be applied even to this extensively studied meteorite. We should also expect that a variety of secondary physico-chemical processes further affected the composition of small C-containing parent bodies and/or their meteoritic fragments throughout their residence in the solar system.

A noteworthy record of such occurrences has been offered by the analyses of the Tagish Lake (TL) meteorite. TL fell in 2000 and some of its stones were quickly recovered from the fall site and maintained frozen right up to the time of analysis (Gilmour et al. 2001; Pizzarello et al. 2001). The one stone analyzed then for organic materials, classified as ungrouped type-2 CC (Zolensky et al. 2002), showed a simpler organic composition than CMs, with fewer soluble compounds and abundant oxidized molecular species such as dicarboxylic acids and O-containing hydrocarbons (Pizzarello et al. 2001; Pizzarello and Huang 2002). Amino acids were found in low amounts and with a distinct distribution, which it was suggested could be secondary, e.g., due to in situ catalytic syntheses from gases released by oxidative decomposition of TL organic materials. At the time, it appeared that TL samples had experienced a rather wide range of parent body conditions, e.g., practically all of the carbon was bound in carbonates in some samples, while this is not the case in other CM lithologies (Zolensky et al. 2002).

Recently, however, another pristine TL stone was analyzed after storage and revealed a composition that, although somewhat depleted in overall species, resembles those of other CMs, e.g., in containing amino acids (Herd et al. 2011). From these new data, it appears that the large asteroidal fragment that created the meteorite shower of Tagish Lake was an unusual stone that had experienced different levels of organic alteration and decomposition throughout its mass, but in sufficient proximity to allow a record of their possible correlation. It follows therefore that this meteorite may offer not only unique insights into the secondary alteration pathways undertaken by soluble organic compounds in meteorites but also the possible recognition of independent formation pathways for some of the various components of the more abundant organic suites, which are not yet fully characterized (e.g., Pizzarello and Huang 2005).

On the prospect of diverse pathways, the idea of ab initio catalytic syntheses of meteorite organics in the cooling solar nebula had been proposed since the early analyses of CC (Anders 1964), gained impetus after the fall of Murchison (e.g., Anders et al. 1973), and was followed by several studies and sometimes heated debate. At the molecular level, this was argued mainly on the comparison between the types of hydrocarbons produced by Fisher Tropsch type (FTT) syntheses, which are mainly linear in their chain length, and those found in meteorites (e.g., see Cronin and Pizzarello 1990). However, it was the isotopic distribution in some meteorite organics that at the time seemed to exclude a major catalytic contribution to the origin of meteoritic compounds in solar nebular processes (Wright and Gilmour 1990; Yuen et al. 1990).

On the basis of our initial observation during TL analyses that some other aliphatic compounds of the meteorite besides amino acids, e.g., carboxylic acids and n-alkanes, displayed a distinct linear chain preference, suggesting a formation by catalytic surface processes rather than random combination of radicals as predicted for interstellar precursors, we address in this work the possibility that some amino acids could be formed in meteorites by catalytic processes upon release of simple gases, as probably occurred in some TL stones, under thermolytic conditions.

We report here the results of FTT syntheses of amino acids from CO, H2, and NH3 in the presence of different meteoritic minerals as catalysts: iron/nickel, Murray-type matrix, magnetite, and Bentonite clay. We found that in some cases, amino acids and amines can be produced efficiently under catalytic conditions and that the products differed in their distribution depending on the catalyst used, with α-amino isobutyric acid (aib) synthesized preferentially by Murchison and magnetite powders.


The catalysts employed in the experiments were: the iron meteorite Santiago Papasquiaro (ASU Center for Meteorite Studies [CMS], cuttings, 7.48% Ni), the CM Murchison meteorite (CMS, powdered), magnetite (Spinel Black, XLine Michigan, 30–100 mesh), and Bentonite clay (CMS clay standard, Montmorillonite #23 Chambers AZ, coarse powders). The Murchison powders were first extracted in water at 100 °C for 24 h and then heated at 550 °C overnight in air prior to use; all other catalytic materials were similarly heated. Reactions were conducted in a closed system: 500 mg of powders was added to cylindrical quartz tubes, 17 cm long and 7.5 cm diameter of approximately 700 mL capacity, closed at one end; the other end had a ball/socket joint opening for the filling of powders followed by a glass stopcock; both openings were fitted with Viton® O-rings. The tubes were evacuated, filled with CO, H2, and NH3 gases in 1:1:1 ratio and with the powder spread along one side. In most experiments (Table 1), they were heated for three-fourth of the length inside a furnace at 370 °C for 24 h. After cooling, the tubes were vented and powders transferred with approximately 5 ml of triply distilled water to 9 × 2 cm vials, which were evacuated and sealed, powders were then extracted at 100 °C for 24 h. The tubes were annealed at 550 °C overnight prior to each run. Blank analyses were conducted both with gases and no meteorite powder, and vice versa. Preliminary experiments also used Pirex tubes of the same dimensions and showed similar results and blanks, except for an isopropylamine contaminant in the latter case (vide infra). One hour 25 °C extracts with stirring in air, which had been performed to analyze the products of a Santiago Papasquiaro sample, showed far lower yields (approximately 1:40) than the subsequent higher T extract, and were not repeated for the other catalytic experiments. All extracts were concentrated by rotary evaporation and analyzed by both cation-exchange liquid chromatography (LC) with postcolumn OPA derivatization and fluorescence detection and gas chromatography mass spectroscopy (GC-MS). For GC-MS analysis, the extracts were dried and reacted in sequence with acidified isopropanol and trifluoroacetic anhydride to give the amino acids-O-isopropyl N-TFA derivatives (e.g., Pizzarello et al. 2004). Amounts reported in Table 1 were calculated from LC analyses because of their higher sensitivity (approximately 1:100 compared to GC-MS) and quantitative reliability. GC-MS analyses confirmed the presence of the LC-detected molecular species; small amounts (0.8–1.1 nmol g−1) of γ-aminobutyric acid were detected in the liquid chromatograms of Santiago Papasquiaro, Murchison and magnetite experiments’ extracts, but not confirmed by GC-MS, and are not entered in the table, albeit that the detections appeared reliable in comparison with standards. Isovaline detection by LC in magnetite experiments, although not supported by GC-MS, was confirmed by its marked increase in fluorescence intensity with higher postcolumn derivatization temperature (Cronin et al. 1979; Cronin and Pizzarello 1986).

Table 1.   . Amino acids synthesized from CO, H2, and NH3 with meteorite powder catalysts; in nanomole per grama of catalyst.
Amino acidSant. Papas.b,c
25 °C extr.
Sant. Papas.c
100 °C extr.
100 °C extr.
100 °C extr.
100 °C extr.
  1. aQuantified by LC with fluorescence detection.

  2. bSantiago Papasquiaro.

  3. cPowders annealed at 550 °C overnight in air before the experiments.

2-amino butyric1.167<0.120.3
2-amino isobutyric<0.12848.9540.00.2


Results are listed in Table 1, which gives amino acid amounts obtained with various catalysts and extraction conditions (in nmol g−1of catalyst powder). It shows that, under the conditions used, several catalytic syntheses can take place and produce amino acids efficiently. It is interesting that, at least for the Ni/Fe catalyst, these amino acids remained mostly bound to it after water rinses at room temperature and were released only after extraction at 100 °C. Data from early experiments (J. R. Cronin and S. Pizzarello, unpublished data) showing a significant increase in amino acids’ yield with the temperature of water extractions come to mind in this context.

Different substrates were found to affect the syntheses of amino acid products with different distributions; only the clay did not catalyze significant amounts of amino acids. The iron/nickel powders catalyze mainly linear compounds up to four-carbon (4C) chain length (Fig. 1), while α-amino isobutyric acid, a branched amino acid, is the most abundant product with Murchison (Fig. 2) and magnetite powders; in the case of a magnetite catalyst, even some isovaline was also observed. Not listed in Table 1 are repeat experiments with Santiago Papasquiaro catalyst that were performed with two separate quartz tubes with a Viton O-ring joint connection where powders were kept either outside or inside the furnace; these produced comparable amounts of amino acids in both cases. Also, quick (25–370 °C) and brief (5 min) heating of reagents and Ni/Fe catalyst, followed by quenched cooling, still produced amino acids, but in lesser amount (approximately by 1/10). Amines up to at least 4C chain length were formed in large amounts by Ni/Fe and magnetite catalysts, but not by Murchison and bentonite powders. These compounds were observed only by LC with fluorescent detection and could not be measured accurately because of a preceding large ammonia peak; however, their combined amounts appear at least equivalent to the amino acids’ in those experiments. All blank runs were void of products except for a few nanomoles of isopropyl amine produced in the case of pyrex tubes.

Figure 1.

 Amino acids produced from CO, H2, and NH3 with iron meteorite Santiago Papasquiaro catalyst.

Figure 2.

 (a) Amino acids produced from CO, H2, and NH3 in the presence of previously annealed Murchison powders, by ion exchange chromatography with fluorescence detection. (b) Amino acid standard.


Results from the analyses of a pristine TL stone analyzed in 2001 suggested that the meteorite’s amino acids, which were few and included only linear-chain species, could have formed in situ from gases released from organic materials undergoing oxidation and degradation. How relevant are the above laboratory experiments toward interpreting the origin of meteoritic amino acids? Isotopic data accumulated so far relate the formation of at least a good portion of these compounds to water processing of highly deuterated interstellar precursors (Pizzarello and Huang 2005; Pizzarello et al. 2008); in addition, isotopic and chiral data point to the possibility that α-substituted amino acids were produced by separate processes (Cronin and Pizzarello 1997; Pizzarello et al. 2004). Yet, it is intriguing that magnetite appears to catalyze the syntheses of amino acids with the closest similarity to those observed in Murchison and other chondrites where α-amino isobutyric acid is predominant. If the combined experimental evidence is hard to reconcile with the likelihood of amino acid in meteorites deriving predominantly from surface catalyzed syntheses, it does not exclude that some fraction of these meteoritic compounds could have been synthesized solely through catalytic processes, nebular or presolar.

In fact, the influence of minerals on the syntheses of meteoritic amino acids as well as other meteoritic compounds cannot be considered a priori an exclusively solar occurrence. The synthetic relationship of the organic and inorganic materials we find in meteorites is almost completely unknown, but we have to assume that, at least for their organic precursor molecules, proximity and synthetic opportunities with minerals must have started early in their cosmic history. In dense interstellar clouds, a variety of predominantly organic gas-phase molecules are known to become adsorbed onto the surface of dust grains (e.g., Herbst and van Dishoeck 2009) and, during the subsequent stages of star formation, we may envision different regimes of temperature and pressure to lead to more complex chemistry involving adsorption/desorption of molecular species, including their cooperation with catalytic inorganic materials. For example, recent isotopic analyses of ammonia released from the insoluble organic materials of several meteorites with different classification (CI, CM, CR, CV, and TL) revealed that this compound is released by all meteorites analyzed (with the exception of the highly metamorphosed CV) and with δ15N values that match closely within meteorite types (Pizzarello and Williams 2012). As meteorites’ classification is based on elemental and mineralogical traits and the 15N enrichments observed for the ammonia are well above solar values, i.e., appear presolar, the study’s results led us to question whether presolar precursors could have come into the solar nebular with their own complement of minerals.

To conclude, although most cosmochemical events may not have met model FTT conditions, this study would suggest that upon accretion, simple adsorption, or ab initio syntheses of simple precursor species, different mineralogical substrates may have substantially affected the direction and speed of organic reactions during solar as well as presolar water processing. The amino acid alpha-carbon could be the main site of this effect, as indicated by the alanine/aib ratios we obtained with different catalysts and the amino acids’ hydrogen/deuterium exchange dependence upon rock/water ratios observed for amino acids by Lerner (1995). These FTT data also add to the study by Lerner et al. (1993), who previously assessed the influence of Allende and Murchison powder over the Strecker syntheses of amino acids. The mineral surfaces’ ability to bind to amino acids shown in our FTT experiments further allows us to speculate on their possible capability to preserve these compounds on early Earth environments upon delivery.


  1. [Correction added on 03 July 2012 after online publication. The spelling of Herd et al (2011) reference was corrected.]

Editorial Handling— Dr. Scott Sandford