Formation of nitrogenated organic aerosols in the Titan upper atmosphere

Abstract

Many aspects of the nitrogen fixation process by photochemistry in the Titan atmosphere are not fully understood. The recent Cassini mission revealed organic aerosol formation in the upper atmosphere of Titan. It is not clear, however, how much and by what mechanism nitrogen is incorporated in Titan’s organic aerosols. Using tunable synchrotron radiation at the Advanced Light Source, we demonstrate the first evidence of nitrogenated organic aerosol production by extreme ultraviolet–vacuum ultraviolet irradiation of a N₂/CH₄ gas mixture. The ultrahigh-mass-resolution study with laser desorption ionization-Fourier transform-ion cyclotron resonance mass spectrometry of N₂/CH₄ photolytic solid products at 60 and 82.5 nm indicates the predominance of highly nitrogenated compounds. The distinct nitrogen incorporations at the elemental abundances of H₂C₂N and HCN, respectively, are suggestive of important roles of H₂C₂N/HCCN and HCN/CN in their formation. The efficient formation of unsaturated hydrocarbons is observed in the gas phase without abundant nitrogenated neutrals at 60 nm, and this is confirmed by separately using ¹³C and ¹⁵N isotopically labeled initial gas mixtures. These observations strongly suggest a heterogeneous incorporation mechanism via short lived nitrogenated reactive species, such as HCCN radical, for nitrogenated organic aerosol formation, and imply that substantial amounts of nitrogen is fixed as organic macromolecular aerosols in Titan’s atmosphere.

Nitrogen is an essential biotic element, yet difficult to fix into biotic/prebiotic molecules directly from N₂. Despite N₂ being the most abundant constituent in the Earth’s atmosphere through most of its history, only limited fixed nitrogen is available for the biosphere. In the thick N₂ atmosphere of Titan with a trace amount of methane, atmospheric chemistry leads eventually to heavy organic gaseous species and aerosol particles. Understanding the formation chemistry and the resulting chemical structure of possible nitrogenated organic aerosols in the Titan atmosphere might help in constraining the atmospheric contribution in abiotic nitrogen fixation processes relevant to the origin and evolution of early life.

One of the most surprising results from the Cassini Ion Neutral Mass Spectrometer (INMS) and the Cassini Plasma Spectrometer is the formation of complex organic molecules over 3,000 atomic mass units in the ionosphere of Titan (1, 2). This observation clearly demonstrated the importance of complex organic chemistry in the upper atmosphere induced by extreme ultraviolet (EUV)–vacuum ultraviolet (VUV) photons and Saturn’s magnetospheric electrons. The INMS data show the prominent existence of unsaturated hydrocarbons in the gas phase (1, 3, 4). Although the nitrogenous compounds are definitely suggested as significant components (5), bulk nitrogen fixation in Titan’s atmosphere is not well constrained at present (6).

Because of the earlier observations of gaseous nitriles (HCN, HC₃N, and C₂N₂) in the stratosphere (7), it has been generally believed that the CN unit plays a central role in Titan’s nitrogen chemistry (8) (Fig. 1). The ground state N(⁴S) atom with CH₃ and C₂H₅ radicals is the probable origin of such CN functional groups. However, a substantial amount of the excited N(²D) atom is also generated from N₂ photodissociation and photoionization via subsequent dissociative electron- recombination (9). Recently, crossed molecular beams studies of the N(²D) reactions with CH₄, C₂H₄, and C₂H₂ reveal the products of CH₂NH (10), C₂H₃N isomers (11), and HCCN (12), respectively. These highly reactive species may be important precursors for the nitrogenated complex molecules and organic aerosols. However, speculative reaction pathways and rates, such as HCCN loss schemes and chemical pathways to aerosols, have been assumed in numerical models (13–17) (Fig. 1).

Fig. 1.

Schematic of nitrogen chemistry in Titan’s atmosphere based on previous models (13, 14). Stable species are shown in double rectangles. Dotted arrows represent photolysis, and red arrows indicate the enhanced channels by C₂H₂ and C₂H₄. The thick dashed box contains the stable CN-unit species of heavy nitriles (C + N)₄ (HC₃N, C₂H₃CN, and C₂N₂) observed in Titan’s atmosphere. Hypothetical chemical pathways toward nitrogenated organic aerosols, in double arrows, are also shown (14, 16); (I) HCN nitrile polymer (19), (II) R-CN nitrile heteropolymers (19), (III) Copolymers (18), and (IV) CH₂NH imine polymer (14). Among each mechanism (II) and (III), the dominant channels (17) are shown as examples. Our study suggests another mechanism as direct HCCN radical incorporation to organic aerosol via heterogeneous chemistry.

To our knowledge, no experimental investigation of nitrogenated organic aerosol formation from EUV-VUV irradiation of N₂/CH₄ has been conducted. Most UV photochemical investigations for nitrogenated organics have been restricted to studies using HCN and HC₃N in the initial gas mixtures (18, 20). Therefore, the role of N(²D) atoms from N₂ molecules in the subsequent complex chemistry is largely unknown. We have succeeded in generating photolytic aerosol samples from a N₂/CH₄ gas mixture at wavelengths sufficient for dissociation and ionization of N₂, demonstrating the important chemical pathway via N(²D)/HCCN for the nitrogenated organic aerosol formation.

Results

Our experiments are conducted using a synchrotron radiation source (Chemical Dynamics Endstation, 9.0.2) at the Advanced Light Source (ALS). The experimental setup is described in our previous papers (21, 22). A gas mixture of N₂/CH₄ (equal to 95/5) flows through the windowless photocell chamber during continuous irradiation by EUV-VUV photons separately at 60 and 82.5 nm, photon energies above and below the N₂ ionization potential, respectively. The neutral gas species are analyzed in situ using a quadrupole mass spectrometer (QMS) with electron impact ionization (EI). The solid samples are kept under vacuum and subsequently analyzed with laser desorption/ionization–Fourier transform ion cyclotron resonance–mass spectrometry (LDI-FTICR-MS). The experimental conditions are described in Materials and Methods.

QMS Analysis of Gaseous Species Generated at 60 and 82.5 nm.

The obtained neutral mass spectra naturally show complex overlaps from isobaric compounds and fragmentation ions from the EI source (70 eV) (Fig. 2A). At 60 nm irradiation, a prominent peak of benzene is observed, which further confirms our previous work on the formation of unsaturated hydrocarbons via N₂ catalytic photoionization (21, 22). The benzene mixing ratio is 3 ± 1 ppm based on our calibration with standard mixtures of N₂/benzene. At 82.5 nm irradiation, less abundance of unsaturated hydrocarbons are observed. The presence of NH₃ at both irradiation wavelengths is clearly observed from the ion (m/z = 8.5) which allows us to derive the abundance of NH₃ with mixing ratios of 800 ± 300 ppm and 200 ± 70 ppm at 60 and 82.5 nm, respectively.

Fig. 2.

QMS analyses of gaseous species from 60 and 82.5 nm irradiation. (A) Averaged mass spectra of neutral gaseous species (EI = 70 eV) obtained by 60 nm (thick line, average of 25 scans) and 82.5 nm (thin line, average of 74 scans) EUV-VUV irradiation of a N₂/CH₄ (equal to 95/5) gas mixture. The MS at 82.5 nm irradiation is vertically offset by a factor of 1,000 for visualization. Formation of unsaturated hydrocarbons (red circles) are much reduced at 82.5 nm irradiation. (B) An MS-deconvolution analysis for 60 nm irradiation. A synthetic mass spectrum from the estimated neutral species (blue bars) with the experimental data (red dots) with standard deviations. (C) For the 60 nm irradiation, the neutral species considered in our MS-deconvolution analysis and the derived mixing ratios in N₂ for major components (red), hydrocarbons (orange), and amines/nitriles (purple), and impurities (gray).

An MS deconvolution analysis with a singular-value decomposition algorithm (3) was utilized to identify and quantify the neutral species from the complex mass spectra (SI Text). A result of MS deconvolution obtained at 60 nm irradiation (Fig. 2 and Table S1) shows that the obtained MS is consistent with the dominance of hydrocarbons without invoking significant nitrogen-bearing species. Two nitriles, HCN and CH₃CN, are not demanded to improve the fitting quality enormously, and the derived mixing ratios should be treated as upper limits. Even though this type of analysis is generally model dependent (SI Text), the unsaturated hydrocarbon ratios of 60 nm over 82.5 nm for C₂H₂, C₂H₄, and C₆H₆ are, at least, larger than 4 ± 2, 3.8 ± 2.4, and 30, respectively.

Search for Nitrogenated Gaseous Species from ¹³C and ¹⁵N Isotope Labeled Initial Nitrogen/Methane Gas Mixtures at 60 nm EUV Irradiation.

One direct way to constrain the upper limit of nitrogenous products is to compare the obtained mass spectra of nitrogen/methane irradiation from separately ¹³C and ¹⁵N isotope labeled initial gas mixtures. Fig. 3A shows the two mass spectra obtained from 60 nm irradiation of nitrogen/methane (95/5) gas mixtures, and . The shift toward heavier mass with ¹³C labeling clearly indicates the incorporation of carbon from the initial gas mixture. In particular, the prominent peak at m/z = 78 from irradiation shifts to m/z = 84 for N₂/ irradiation, confirming the assignment of C₆H₆.

Fig. 3.

QMS analysis of gaseous products from 60 nm irradiation of the isotope substituted nitrogen/methane (equal to 95/5) gas mixtures. The QMS intensities are normalized for N₂ peak intensities. (A) Comparison of and (EI = 35 eV). (B) Comparison of and (EI = 70 eV). Inserted figures are shown in linear scale for the C₆-C₇ mass spectral windows.

In contrast, the comparison of the mass spectra from gas mixtures of and irradiated at 60 nm shows no obvious shift in the observed MS spectra (Fig. 3B) with only a few exceptions: (i) A clear shift from m/z = 91 to m/z = 93 suggests the presence of ion that overlaps with the peak for . If CN^- is the daughter fragment from the parent neutral upon EI, its origin could be C₆H₃N₃ = (HC₂N)₃. (ii) Shifts from m/z = 17/18 to m/z = 18/19 are consistent with the presence of NH₃. The upper limit of the contribution of nitrogenated species to the spectral intensity is 10% for the m/z range larger than 50.

Evidence of N₂ Dissociation from EUV-VUV Irradiation of a Gas Mixture.

Even though the neutral gas products are dominated by hydrocarbons, direct evidence of N₂ dissociation is demonstrated from the irradiation of a gas mixture. An equimolar gas mixture is suddenly irradiated at a given photon wavelength, and the time variations of the ¹⁴N-¹⁵N molecules ion signal is observed with the QMS (Fig. S1). The scrambled formation of ¹⁴N-¹⁵N molecules is the definitive evidence of the initial N₂ molecule dissociation, and the equilibrium value of ¹⁴N-¹⁵N over the initial value provides a first-order approximation of dissociation fraction under the assumption of random recombination from N atoms to N₂ molecules. Fig. 4 indicates ∼1% dissociation of the molecules between 80–100 nm, which corresponds to the total N₂ ( and ) photodissociation rate of 3–5 × 10¹⁵ molecules/s in our narrowband UV irradiation at the ALS. At wavelengths shorter than 80 nm irradiation, approximately 3–5% of total N₂ molecules experience dissociation. At wavelengths shorter than 66 nm, where the N₂ molecule experiences continuum photoabsorption (23), the photon flux is equal to the N₂ dissociation rates under our optically thick condition and is estimated to be 1–1.5 × 10¹⁶ photons/s, which is consistent with previous measurement (24).

Fig. 4.

EUV-VUV wavelength dependence of N₂ dissociation fraction estimated from the production of ¹⁴N-¹⁵N molecule (m/z = 29) from photolysis of a (equal to 50/50) gas mixture at total pressure of 0.11 mbar. A gas filter of either argon (circles) or helium (squares) is used for reducing the high-energetic overtone light of the undulator beam at the Chemical Dynamics Beamline (Endstation 9.0.2) at the Advance Light Source.

LDI-FTICR-MS Analysis of Organic Aerosols Generated at 60 and 82.5 nm Irradiation.

The LDI-FTICR-MS analyses of the accumulated aerosol samples at 60 and 82.5 nm irradiation (hereafter, UV_600 and UV_825 samples) reveal more than 5,000 peaks between 200–700 m/z in the mass spectra. This ultrahigh mass resolution analysis (M/ΔM = 3 × 10⁷) enables the unambiguous assignments of CHNO formulae of several thousand molecules in the mass range (SI Text). Unlike the gaseous sample, most of the observed ions from the solid contain nitrogen, and only a few pure hydrocarbon ions are observed in these samples (Table S2). The MS-intensity weighted ratios of H/C and N/C in the observed mass region clearly indicate the production of heavily unsaturated and heavily nitrogenated solids. Based on our calibration comparing the MS weighted N/C ratio and the combustion elemental analysis for various types of organic macromolecules (SI Text), the estimated bulk C/N ratios in the UV_600 and UV_825 are 1.6 ± 0.3 and 1.1 ± 0.2, respectively.

In order to investigate the functional diversity and the possible precursors, all identified ion species are plotted with the aid of van Krevelen diagrams (25, 26) (Fig. 5A). This perspective clearly reveals the distinct nitrogen incorporation in organic macromolecules at 60 and 82.5 nm irradiation, implying the different chemical reaction pathways responsible for the solid product formation. Most species at 82.5 nm irradiation are distributed around H/C = 1 and N/C = 1, implying HCN as the important precursor. The wide distribution around that of HCN suggests that the species observed is the result of the covalent or polymeric mixing of HCN and other moieties, such as -CN, -H, -NH_x, and -CH_x. The N/C ratio appears to converge with higher mass (Fig. 5B), which probably reflects the reduction of statistical dispersion in the inclusion of various precursors.

Fig. 5.

Ultrahigh-mass-resolution LDI-FTICR-MS analyses of photolytically generated organic aerosols at 60 nm (UV 600) and 82.5 nm (UV 825) irradiation of a N₂/CH₄ (equal to 95/5) gas mixture. (A) Van Krevelen diagrams (N/C vs H/C) of all identified CxHyNz⁺ compounds. (B) Variation of N/C ratios of all identified compounds as a function of mass. (C) Cumulative MS intensities shown as van Krevelen diagrams.

On the other hand, the distribution in the UV_600 is more carbon rich with only a few peaks around H/C = 1 and N/C = 1, indicative of HCN not being the major seed at 60 nm. Although the UV_600 sample shows a wide diversity in constituents, the integration of MS intensity in the van Krevelen diagram (Fig. 5C) suggests the dominant chemical constituents are the congeners with the bulk ratio of C₂H₂N, which provide the constraint in the combinations of dominant precursors, such as HCCN + H, C₂H + NH, and CH₂ + CN. However, almost no detection of pure hydrocarbons suggests that dominant precursors at 60 nm contain both carbon and nitrogen.

Atomic hydrogen liberated from CH₄ dissociation certainly collides with the surface of the evolving solid. Considering the previous experimental demonstration that hydrogenation is much faster than H abstraction for H-atom heterogeneous reactions of this type of material (27), the dominant precursors probably contain less hydrogen than those forming the major peak location of (C₂H₂N)_n in the UV_600 sample. Therefore, the HCCN radical as the dominant precursor at 60 nm irradiation is consistent with these observations.

Discussion

HCCN Radical as an Important Precursor for Nitrogenated Organic Solid Production in the EUV Irradiation of N₂/CH₄.

The efficient formation of unsaturated hydrocarbons is observed in the gas phase without abundant nitrogenated neutrals at 60 nm, whereas the solid components are dominated with heavily nitrogenated species and few pure hydrocarbons. This strongly suggests a direct nitrogen incorporation mechanism of short-lived nitrogenated reactive species into solid products via heterogeneous chemistry. The chemical pathways of (II) and (III) in Fig. 1 require those heavy (C + N)₄ nitriles as intermediates, and therefore, are unlikely the dominant mechanism for the UV_600 generation. The ultrahigh-mass-resolution study of N₂/CH₄ photolytic solid products at 60 and 82.5 nm define the distinct nitrogen incorporations, and the integrated MS intensities centered at the elemental abundances of H₂C₂N and HCN, respectively, are suggestive of important roles of H₂C₂N/HCCN and HCN/CN in their formations. The much faster production rate of solid sample at 60 nm than that at 82.5 nm seems correlated to the efficient formation of gaseous unsaturated hydrocarbons. This implies a key role of small unsaturated hydrocarbons in the formation of nitrogenated solid products (Fig. 1).

The formation of N(²D) atoms is expected at both 60 and 82.5 nm irradiation (SI Text). The production of HCCN radical from N(²D) is active at 60 nm irradiation, whereas it becomes minor at 82.5 nm irradiation (SI Text and Table S3). The generated HCCN radical species at 60 nm irradiation is most likely lost through subsequent chemical reactions. Nondetection of (C + N)₄ nitriles species in the gas phase at 60 nm does not support the proposed bimolecular reactions (15, 18, 29) (Table 1) as the major HCCN loss channels in our experimental condition. The competition of gas-phase and heterogeneous reactions can be discussed by introducing γ, a reaction probability, and ν_coll, a collision frequency at the product aerosol substrate/chamber wall (SI Text). Those possible competitive loss channels via bimolecular reactions remain minor if the heterogeneous loss of HCCN is represented by γ much larger than 0.1 (SI Text and Table S3). Although the reaction probability of HCCN incorporation is not known, such a polyatomic biradical would be expected to have γ closer to one. Our observations of the UV_600 seem to support this efficient HCCN radical incorporation into organic macromolecules via heterogeneous chemistry.

Table 1.

Loss reactions of HCCN radicals

Possible reactions	Reaction rate*
Upper limits measured by Adamson et al. (28)
HCCN + CH4(/C2H4/C2H2/H2)	< 1 × 10-13
Proposed schemes by Yung (15)
HCCN + N → C2N2 + H	1 × 10-12 (est.)
HCCN + HCCN → C4N2 + H2	5 × 10-11 (est.)
Proposed schemes by Osamura and Petrie (29)
HCCN + H → CCN + H2	3 × 10-11 (est.)
HCCN + CH3 → C2H3CN + H	3 × 10-11 (est.)
HCCN heterogeneous incorporation to solid
HCCN + surface → growth	γ νcoll†

^*Units are in cm³ s^-1 for the bimolecular reaction rates and in s^-1 for the heterogeneous reaction rate.

^†γ, reaction probability; ν_coll, collision rate.

Even though most N(²D) are converted to CH₂NH (Table S3) at 82.5 nm irradiation, HCN is likely to be the major precursor reactant (Fig. 5 and Fig. S2), which is consistent with the chemical pathway (I) in Fig. 1. Because of the underestimation of CH₂NH loss process to match the INMS observations in Titan’s ionosphere, gas-phase conversions of CH₂NH to HCN/H₂CN have been recently proposed (Fig. 1) (17). Our observations seem to support these mechanisms. However, the proposed CH₂NH polymerization [Fig. 1 (IV)] (17) is only consistent if hydrogen could efficiently be removed during polymerization reaction which is unlikely (27).

Finally, significant amount of HCN formation is also expected even at 60 nm irradiation via the N(⁴S)/CH₃ radical pathway (30) or via CH₂NH channels (17) (Table S3). However, negligible contribution of a (HCN)_n center in the UV_600 sample suggests that HCN [Fig. 1 (I)] is not as efficient as HCCN radical as a precursor for solid production. In fact, the production rate of the UV_825 sample is much slower over that of the UV_600. The diradical or carbene nature of HCCN (31) allows for efficient insertion into single bonds or addition to unsaturated and radical centers. These strong differences and the differences between solid growth rates in our experiments cast doubt on the HCN polymerization pathway [Fig. 1 (I)] (16, 19) as the dominant nitrogenated organic aerosol production mechanism in Titan’s atmosphere.

Implication for Titan’s Upper Atmospheric Chemistry.

In the upper atmosphere of Titan, the presence of both N(²D) atom and unsaturated hydrocarbons most likely leads to the formation of HCCN (12) and C₂H₃N (11). The Cassini Ultraviolet Imaging Spectrometer observation reveals aerosol particles widely distributed in Titan’s thermosphere/mesosphere (32). Here, we discuss the nitrogen incorporation mechanism for aerosol growth, with particular emphasis on whether heterogeneous reaction of HCCN radicals could be plausible in the dusty environment in Titan’s thermosphere/mesosphere.

It has long been assumed that HCCN radical reacts with N(⁴S) and HCCN itself, to generate C₂N₂ and C₄N₂, respectively (15) (Table 1). The latter pathway has been used estimation of the primary responsible pathway to organic aerosols via C₄N₂ and C₃N [Fig. 1 (III)] in the upper atmosphere (17). All numerical models adapting such HCCN loss chemical schemes predict 10–100 times larger C₄N₂ abundance than C₂N₂ around 1,000 km altitudes (13–15, 17). Such preponderance of C₄N₂ is not observed from the Cassini INMS observations (3, 4), implying other chemical pathways for the HCCN radical. One possible channel might be the reactions of HCCN radical with H and CH₃ (29) (Table 1). Unfortunately, no experimental validation nor reaction rate measurements have been conducted for such systems.

A third possible loss channel for HCCN is heterogeneous reactions occurring on the surface of aerosol. This can be described as a first-order process with the rate constant, (1/4)γvA_V, where v is the mean thermal velocity and A_V is the aerosol surface area per unit volume (cm² cm^-3) (8). Previous models (32, 33) give a surface area density (A_V) of the order of 10^-9–10^-8 cm²/cm³ in the thermosphere/mesosphere. The loss rates of N(²D) and HCCN radicals via bimolecular reactions and heterogeneous reactions are compared in Fig. 6. The neutral density profiles and the aerosol surface area density profile are based on the recent numerical models (33, 34) even though the effect of heterogeneous chemistry on neutral/aerosol densities are not self-consistently coupled in these models.

Fig. 6.

Comparison of loss rates of (A) N(²D) atoms and (B) HCCN radicals via gaseous bimolecular reactions and heterogeneous reactions in Titan’s upper atmosphere. The reaction probabilities, γ, of N(²D) and HCCN heterogeneous incorporation are shown for values of 1, 0.1, 0.01, and 0.001.

The photodissociation of N₂ molecules occurs at altitudes above 700 km (35). N(²D) loss via reactions with C₂H₂/C₂H₄/CH₄ dominate below 1,100 km altitudes where those chemical reactions rates exceed the radiative loss of N(²D) (1.6 × 10^-5 s^-1) (36) (Fig. 6A). Thus, formation of HCCN most likely occurs via bimolecular reactions between 700–1,100 km altitudes. If the reaction probability for HCCN radical incorporation is larger than 0.1 as suggested from our experiment, the HCCN loss thorough heterogeneous reaction becomes larger than the competing gas-phase reactions with N(⁴S) atom and HCCN radical itself at altitudes below 900 km (Fig. 6B). This heterogeneous loss of HCCN radical can even be 10% that of the hypothetical loss reaction with H atom. Even if HCCN reaction with H atom is dominant, the expected CCN radical unit would eventually be incorporated into organic aerosols (37). The column mass production rate via heterogeneous gas-to-particle conversion of HCCN radical between 700–900 km altitude ranges can be up to 1 × 10^-15 g cm^-2 s^-1 (based on a sticking coefficient of one), which is much greater than other estimated chemical pathways [Fig. 1 (I–IV)] at these altitudes (17).

The above discussion is definitely dependent on the aerosol particle density/shape. The heterogeneous chemical growth of particles and particle coagulation may affect the altitude variation of A_V, which is not considered in Fig. 6. However, the constraint for formation and heterogeneous loss of HCCN radicals is robust between A_V = 10^-10–10^-8 cm²/cm³. This condition is probably satisfied at the altitude range above 700 km, however further study of the coupled chemical heterogeneous growth and resulting particle size/shape evolution is needed in the future.

In the upper atmosphere of Titan where substantial amounts of excited N(²D) are generated accompanied with unsaturated hydrocarbons, nitrogen incorporation into organic aerosols efficiently proceeds via N(²D) and HCCN. The HCCN heterogeneous incorporation may have great impact on the entire nitrogen chemistry, beause a significant fraction of excited N(²D) is decoupled from the CN-unit-based chemistry (Fig. 1, dashed box). Organic aerosols in the upper atmosphere of Titan might be a hidden N sink, which eventually accumulate on the surface of Titan with chemical potential for prebiotic evolution.

Materials and Methods

A gas mixture of N₂/CH₄ = 95/5 flows through the windowless photocell chamber of 70-cm length during continuous irradiation of the centerline with wavelength tunable EUV-VUV photons using a synchrotron radiation source (Chemical Dynamics Endstation, 9.0.2) at the ALS (The Lawrence Berkeley National Laboratory) (21, 22). The photon beam was nominally 0.5 mm², with a flux near 10¹⁶ photons s^-1 at ∼1 nm bandwidth (24). Two series of experiments flowing a premixed N₂/CH₄ (5.01% CH₄) gas mixture (Matheson Tri-Gas) at 5 × 10¹⁷ molecules/s were conducted at a pressure of 0.13 mbar for 3 h at 60 nm and for 8 h at 82.5 nm photon irradiation. A gas filter of either argon at 82.5 nm or helium at 60 nm is used for reducing the high-energy overtone light of the undulator beam (24). The ¹³C and ¹⁵N isotope labeling studies are conducted at 0.066 mbar using premixed mixtures of (equal to 95/5) (¹³C 99.9%, Cambridge Isotope Laboratories) and (equal to 95/5) (¹⁵N 98%, Isotech Inc.). A continuous flow of an equimolar gas mixture at total pressure of 0.11 mbar was made by adjusting flows of (Research Purity, Matheson Tri-Gas) and (¹⁵N 98%+, Cambridge Isotope Laboratories). The gas species produced by photochemistry are analyzed using the QMS (CIS 200, Stanford Research System). The solid sample is accumulated on a copper plate as a thin film, whose optical thickness growth is monitored with UV-visible reflectance interferometric spectroscopy. A tungsten-halogen lamp source (LS-1, Ocean Optics) illuminates the growing film with a 3 mm focused diameter spot, and the specularly reflected light is continuously measured with a CCD-array spectrometer (300–850 nm, USB2000, Ocean Optics). The solids production rate at 60 nm irradiation is, at least, 7–10 times larger than at 82.5 nm.

The solid samples are kept under vacuum without exposure to the ambient air, and are subsequently analyzed with a LDI-FTICR-MS [Bruker Daltonics, apex ultra 12 Tesla with MALDI source (355 nm)]. The accuracy of the mass-to-charge ratio (m/z) is within 0.3 ppm of ΔM/M after external and self-improved calibration. The identification of CxHyNz⁺ ions are based on the comparison of the measured mass and the calculated mass with 0.5 ppm tolerance. The MS-intensity weighted N/C ratios are calibrated by conducting the same LDI-FTICR-MS analyses and combustion elemental analyses for N₂/CH₄ plasma polymer samples and black HCN polymer. The detailed procedure is given in SI Text.