Quantitative evolved gas analysis: Winchcombe in comparison with other CM2 meteorites

Two bulk Winchcombe along with six other CM2 meteorite samples were subjected to quantitative evolved gas analysis. The observed release patterns for almost all volatile species demonstrate close similarity for all the samples and especially between those for Winchcombe. This can be considered as a fingerprint for this petrological type of meteorites. We identified several gases including H2, H2O, O2, CO, CO2, and SO2 released in different temperature ranges. The sources and mechanisms of their release were also established. Some of the gases, H2, CO, and CO2, are released as a result of oxidation of macromolecular organic material from oxygen derived from oxygen‐bearing minerals (a part of CO2 is also released as a result of decomposition of carbonates). The others, O2 and H2O, are associated with the phase transformation/decomposition of phyllosilicates and (oxy)hydrates, while a high‐temperature release of SO2 is associated mostly with the decomposition of sulfides and in few cases also with sulfates. A low‐temperature release of SO2 is due to evaporation and oxidation of elemental sulfur from the meteoritic matrix and organic material. The total concentrations of H (mostly represented by H2O), C, and S, calculated according to calibration of the quadrupole mass spectrometer with reference gases and decomposition of solid samples (CaSO4·2H2O and NaHCO3) are in reasonable agreement with those determined by independent methods. Variations in the ratio of the carbon amounts released as CO2 and CO ( CCO2 /CCO) between the samples could be an indicator of their terrestrial weathering.


INTRODUCTION
Evolved gas analysis (EGA) is a widely used technique in a large number of applications in science and industry, from investigations of chemical reactions to food production (see Risoluti & Materazzi, 2018 and references therein).In geo-and cosmochemistry EGA is less common, with a possible exception of the investigation of phase transformations in minerals under heating where thermogravimetric analysis is used in combination with EGA (Maøgorzata & Zdzisøaw, 1989).In its pure form, EGA is used to study volatile species released from rocks and minerals during linear heating.The gases are released according to their volatility, location in the mineral structure, or as a result of chemical reactions between coexisting mineral phases in the case of rocks.Thus, EGA gives an outgassing spectrum for each specific gas as a function of temperature, unique for each individual material.The spectra are usually difficult to interpret because they are a superposition of several processes that occur within samples when heated, such as diffusion, decrepitation, phase transformations, and chemical reactions.However, deciphering the processes provides valuable information about the origin and nature of the analyzed samples in terms of their associated volatiles.
Evolved gas analysis has been applied to a number of Apollo samples to determine their volatile budgets (Gibson & Moore, 1972a, 1972b, 1973) and continuing through to investigations of Martian rocks by the Curiosity rover (Clark et al., 2021;Ming et al., 2014).
Recently, we developed a quantitative EGA (QEGA) technique and applied it to a selection of meteorites and Apollo lunar samples (Verchovsky, Anand, & Barber, 2020;Verchovsky, Anand, Barber, Sheridan, et al., 2020).The method is based on calibration of a quadrupole mass spectrometer (QMS) with gas flow and makes use of a precise capacitance manometer (MKS baratron ® ).In addition, calibrations have been performed using chemical compounds decomposing into simple gases when heated, for example NaHCO 3 or CaSO 4 Á2H 2 O.
In the present study we used QEGA to analyze two Winchcombe samples (see the Samples Section) along with several other CM2 samples for comparison.

ANALYTICAL PROCEDURES Experimental Technique
Full details of the relevant experimental procedures are given in Verchovsky, Anand, and Barber (2020) and Verchovsky, Anand, Barber, Sheridan, et al. (2020).Briefly, the apparatus consists of a reference system, a precise pressure measurement device, that is, capacitance monometer (MKS baratron ® ), an extraction furnace and the QMS (Hiden 3F/PIC) used for registration of signals from gas species at different m/z.The reference system contains a reservoir with reference gases at 4-5 bar pressure.The reservoir can be filled with pure gases or mixtures of gases with known relative abundances (Table 1).The flow rate of gases from the reservoir is controlled through capillaries (stainless steel, 0.89 mm external diameter) with crimps.
The reference and sample gases were analyzed on the QMS where signal intensities at a number of m/z in the range from 2 to 140 in dynamic mode using peak jumping and ion counting are recorded.The meteorite samples and the reference compounds, NaHCO 3 and CaSO 4 Á2H 2 O, are wrapped in clean Pt foil and heated linearly using a 12°C min -1 heating rate in the extraction double-wall furnace (with quartz tube inside and ceramic tube outside the space between which is pumped to 10 -3 mbar) from 100 to 1400°C.The evolved gases are continuously passed through the QMS and analyzed in the same way as the reference gases.No carrier gas was used during calibration and sample analyses.Blanks are measured before each sample and the corresponding correction is applied.The blanks seen with a new empty quartz extraction tube are different from those made after sample analyses, mainly in the amount of O 2 released at very high (>1300°C) temperature: oxygen is observed only in the latter cases.This is due to decomposition of SiO 2 vapor originating from the quartz tube at such a high temperature, catalyzed by Pt foil (Verchovsky, Anand, & Barber, 2020).The more Pt is present the higher the amount of oxygen generated.The sample silicates can also contribute to the molecular oxygen budget the same way as suggested for quartz, but to a much lesser extent due to their far smaller mass/surface area compared to the mass/surface area of the quartz tube in the heating zone of the extraction furnace.All the measurements are performed in an automatic computercontrolled mode.

Flow Rate Calibration
Calibration consists of determining the gas flow rate by measuring the pressure of the reference gas in the capacitance monometer using a range of accumulation times and is expressed in cm 3 STP per second.To simplify automation of the calibration procedure the reference gas passes through three-parallel capillaries each set for different flow rates.By switching between individual capillaries and their combinations, seven different flow rates of the same gas can be obtained.The results of the calibrations are present in Figure 1.

QMS Calibration with Reference Gases
Two gas mixtures prepared by Air Products have been used for the calibration (Table 1).The QMS calibration lines represent dependence of signals (cps) collected at each m/z on flow rate (Figure 2).We fitted the lines using a power law, which gives a better fit than a linear fit, though the difference is generally insignificant.This is important for extrapolation of the fitting lines when calculations are required for the signals outside the range of data points collected during the calibrations.Calibrations with gas mixtures and pure gases give consistent results indicating that the presence of other gases does not significantly affect the calibration results, except when the gases present in the mixtures give signals at the same m/z, which is observed, for instance, at m/z = 28 where there are contributions from N 2 , CO, and CO 2 .In this case the second-order lines, from fragments of the gas ions, usually give consistent results.Therefore, it is important to identify all secondorder signals for the analyzed gases and their relationships with the main signal that is achieved by analyses of pure gases (Figure 3).The second-order signals are formed as a result of dissociation of the gas molecules into fragment ions in the ion source of the QMS.The relationship between the main and the second-order signals depends on QMS settings, mostly on electron acceleration voltage, and can vary within an order of magnitude.
The calibration lines are used to convert the sample signals in cps into flow rates expressed in cm 3 STP per second.In order to find the absolute amount of gas released during linear heating the released curves are integrated over the time of release.Since the number of points collected during linear heating from 100 to 1400°C is large (∼5000) a good determination of the integral is found as a sum of all collected intensities expressed in terms of flow rate (cm 3 STP per second) divided by their number and multiplied by the total duration (s) of heating.
Calibration for H 2 O and SO 2 with NaHCO 3 and CaSO 4 Á2H 2 O Decomposition of the pure compounds occurs according to their stochiometric composition: H 2 O is released at low temperatures (100-200°C) from both compounds while SO 2 is released at T > 1000°C (Figure 4).The integral of the peaks at m/z = 17 (from OH, the secondary signal from water) and m/z = 64 over time gives the total number of counts for the amounts of H 2 O and SO 2 respectively, which can be calculated from the weights of the reference sample and the reaction equations.This determines the relationship (conversion factor) between the amounts of gases and the integral number of counts which is then used to find the amounts of H 2 O and SO 2 in the meteorite samples from the corresponding peak integrals.The strong water signal in the QMS (Figure 4b) indicates that despite the pipes between the extraction furnace and QMS being maintained at room temperature, water is only partially condensed.It is assumed that this partial condensation of water occurs the same way during both calibration and sample analyses.Three calibrations made with two aliquots of CaSO 4 Á2H 2 O and one aliquot of NaHCO 3 gave consistent results for the conversion factor: 2.57, 2.45, and 2.67 (in ×10 -14 g per count).Reference gas was accumulated in the section with the capacitance monometer.The amount of accumulated gas was found from the known volume of the section and the pressure recorded by the capacitance monometer The plots represent the results for two different gas flow rates (1.02 × 10 -6 on the left and 2.5 × 10 -6 on the right in cm 3 STP per second) set by crimps on different capillaries.

SAMPLES
In this study we analyzed two fragments (BM.2022,M1-85 and BM.2022, M1-861a-86; 0.419 and 0.507 mg, respectively) of the Winchcombe CM2 meteorite.The full description of the fragments and aliquots (in this paper W1 and W2, respectively) is given by Greenwood et al. (2023) and King et al. (2022), but to summarize: W1 is an example of the highly altered, ∼CM2.1 'Lithology B' (Suttle et al., 2023) and W2 is a complex breccia composed of cataclastic matrix and a variety of highly altered clasts, dominated by ∼CM2.2, 'Lithology A' material (Greenwood et al., 2023;Suttle et al., 2023).To understand whether Winchcombe is a typical CM2 in terms of evolved species composition and release pattern we also analyzed six bulk samples of other CM2

RESULTS
The results of the EGA for W1, W2, and six other CM2s are presented in a series of similar plots  showing release patterns of the most important species recorded at m/z = 2, 12, 14, 16, 17, 28, 29, 30, 32, 44, and 64 with temperature in the same format for all the samples.The total pressure as measured on the ion pump which pumps the evolved gases through QMS is also recorded.Except for some minor differences the pressure variations as well as the release patterns have many similar features for all the samples analyzed.These are listed below: 1.The total pressure has maximum value of up to 10 -6 mbar in the temperature range of 600-700°C  12g).7.All the samples show a maximum of signal at m/z = 2 between 600 and 800°C (Figures 5-12c).In two samples (Figures 7c and 10c) the signals at m/z = 2 also show an additional low-temperature release peak.8.The release pattern of the species at m/z = 17 is smooth with a maximum between 300 and 600°C (Figures 5-12c).The lower temperature release is observed in the samples where the low-temperature release at m/z = 2 is also observed.9. Some samples have a strong signal at m/z = 32 at T > 1200°C where it is correlated with the signal at m/z = 16 (Figures 8 and 10-12d).In addition, most of the samples have a low-temperature peak at m/z = 32 in the range 500-600°C (Figures 5-12d).
The release pattern of this low-temperature component is very similar to that observed at m/z = 17 (Figures 5-12c).10.The signal at m/z = 64 is generally observed in two temperature ranges 200-800°C and 1200-1380°C (Figures 5-12d and 13).11.Apart from many similarities with the above observations both Winchcombe samples have an almost identical release pattern for all the analyzed species, reproducing each other in detail (Figures 5  and 6).
We also note that the results obtained earlier by QEGA for the Murchison (also CM2) sample (Verchovsky, Anand, Barber, Sheridan, et al., 2020) are very similar to those listed above.

Identification of the Evolved Gas Species
When there is no interference between the species with the same m/z, including the secondary order signals, identification of the species is straightforward.Thus, the signal at m/z = 17 corresponds to H 2 O (assuming constant OH + /H 2 O + ratio), at m/z = 44 to CO 2 and at m/z = 64 to SO 2 .
Thermal decomposition of water occurs at 2200°C and can be reduced to 1300-1400°C by Pt catalyst (Jellinek & Kachi, 1984).Therefore, it is possible that part of the molecular hydrogen is the result of recombination of H 2 O fragments formed during its dissociation in the QMS where there is a hot (∼2000°C) filament and ionizing electrons.However, the releases of water and H 2 are not at all similar to each other: the release of H 2 O is quite smooth while the H 2 release has many features observed at different temperatures (Figures 5-12c).Therefore, it is difficult to imagine how recombination of H 2 O molecule fragments could lead to formation of H 2 by such a complicated and very specific process without correlation with the release of the main signal at m/z = 17.Moreover, some features of the H 2 release coincide with those for CO 2 and CO.Thus, if water produces molecular hydrogen its contribution is small and most of the observed H 2 comes from other sources (see next section).Since most of the water is released by 1000°C its dissociation in the furnace at 1300-1400°C, if it occurs, makes an insignificant contribution to H 2 .Water partly turns into OH (at m/z = 17) as evidenced by the nature of its release, which is identical to H 2 O (at m/z = 18) as can be seen from the experiment with gypsum (Figure 13).
Recombination of H 2 O fragments also results in the formation of O 2 , and since the signal at m/z = 32 is much smaller than for H 2 in the samples analyzed (Figures 5-12c,d) its influence on the O 2 signal can potentially be significant.The results obtained for gypsum (Figure 13), used for calibration for water and SO 2 indicate that in the temperature range where most water is released no signal from oxygen is observed, that is, the process of fragmentation of water molecules does not occur in our experiments, at least to the extent that can be detected.Oxygen is only released at high temperature where it is expected to appear together with SO 2 as a result of decomposition of CaSO 4 .Small amounts of H 2 released at low and middle T appear to be an impurity in the gypsum.Thus, the results for oxygen agree with those for hydrogen discussed in the previous paragraph.
Some contribution at m/z = 32 is also expected from the second-order signal of SO 2 (Figure 3).Since we did not analyze pure SO 2 gas at the same time as the meteorite samples and used different QMS settings during the analysis, the relationship between the main and the second-order signals for SO 2 may differ slightly from what it would be if pure SO 2 were analyzed with the same QMS settings as used for the samples.The hightemperature release of SO 2 from Nogoya (Figure 14) can also be used to determine the relationships between the main and secondary signals for SO 2 , since it does not seem to mix with any other signals and, in particular, oxygen from decomposition of SiO 2 of the sample tube.There are at least three second-order signals associated with SO 2 that closely and in detail correlate with that at m/z = 64: at m/z = 16, 32, and 34 (the signal at m/z = 48 was not measured).As expected, the ratios between the signals are similar but not exactly the same as determined for pure SO 2 (Figure 3).The ratio between signals at m/z = 32 and 34 (∼24) indicates that the signal at m/z = 32 is a 32 S + ion and not 16 O 18 O + ion since the ratio of molecular oxygen with a mass 32 ( 16 O 16 O) to that with a mass 34 ( 16 O 18 O) is ∼250 and the 32 S/ 34 S ratio is 22.This result is also similar to what was determined for pure SO 2 .The ratio of the signal at m/z = 64 to the signal at m/z = 32 ( 32 S + ) is ∼10 (for pure SO 2 it is ∼50).Note that the ratio of SO 2 to O 2 signals for the gases formed during decomposition of calcium sulfate is about 0.6 (Figure 13), indicating that the release of SO 2 from Nogoya shown in Figure 14 is not a result of decomposition of a sulfate (see next section).For the W1 and W2 samples, a significant contribution from SO 2 in the signal at m/z = 32 is observed at low-temperature (250-300°C) where a small amount of O 2 is released while the signal at m/z = 64 is relatively high (Figures 5 and 6d).For the main O 2 release this contribution is almost negligible since the signal at m/z = 64 is significantly lower than that at m/z = 32.For ALH 82100 (Figure 7d) most of the signal at m/z = 32 is from SO 2 since both signals are well correlated and the ratio of the signal at m/z = 64 to that at m/z = 32 is ∼10.For ALH 83100, ALH 88045 and Murray (Figures 8-10d) most of the signal at m/z = 32 corresponds to O 2 since the ratio of the signal at m/z = 64 to the signal at m/z = 32 is much less than 10.For Nogoya and Cold Bokkeveld (Figures 11 and 12d) there is a small contribution of the second-order signal from SO 2 at m/z = 32 since the signal of SO 2 is higher than at m/z = 32, although the ratio between them is less than 10 and the signals are not correlated.Thus, in most samples, the majority of the signal at m/z = 32 corresponds to O 2 .
For m/z = 28 the situation is more complicated.Apart from the interference between CO and N 2 there is also a contribution at this m/z from the secondary signal of CO 2 .Pure CO 2 gives secondary signals at m/z = 16, 28 and 12 which are smaller than the main line at m/z = 44 by factors of 5, 10, and 20, respectively (Figure 3).So, a significant contribution from CO 2 is possible only when the CO 2 signal is significantly higher than that at m/z = 28.At the same level of the main signals the contribution from CO 2 at m/z = 28 is only 10%.For our samples a moderate contribution from CO 2 is possible only in the case of Nogoya in the temperature range 200-600°C (Figure 11b) and at the very low temperature (∼100°C) for ALH 82100 (Figure 7b).For all other samples and temperature ranges the contribution from CO 2 at m/z = 28 is negligible.Thus, the signal at m/z = 28 is potentially created by CO and N 2 .To assess the contribution from each of them the relationship between the signals at m/z = 28, 14, and 12 for the pure gases has to be considered.Pure N 2 gives a second-order signal at m/z = 14 that is a factor of 15 times smaller than that at m/z = 28 (Figure 3).Pure CO gives second-order signals at m/z = 14 and 12 at factors of 100 and 43 smaller than the main signal correspondingly.Thus, at the same level of the main signals at m/z = 28 N 2 gives a higher signal at m/z = 14 than CO gives the signal at m/z = 12 (Figure 3).In fact, the signal at m/z = 14 is significantly smaller than that at m/z = 12 (Figures 5-12b) which means that CO is the only gas creating the signal at m/z = 28 for all our samples.As mentioned in the previous section, the relationship between signals at m/z = 28, 29, and 30 (Figures 5-12g) also supports this conclusion.
We also considered whether the signal at m/z = 14 might be associated with nitrogen.We concluded that it was not, since this signal is well-correlated with that at m/z = 28 (Figures 5-12h).That, together with the association of the signal at m/z = 28 with CO, as discussed in the previous paragraph, means that it comes from CO.It also follows from the relationships between signal at m/z = 14 for pure CO and N 2 (Figure 3).And although the signal at m/z = 14 cannot be formed directly  Although CO dominants the m/z = 28 signal, it is not responsible for the signal at m/z = 12 as might be expected.In fact, the signal at this mass is much better correlated with CO 2 than CO (Figures 5-12e,f), that is, the signal at m/z = 12 comes mainly from CO 2 .This also follows from the relationships between signals at m/z = 12 for pure CO 2 and CO (Figure 3).
In all meteorite samples an m/z = 14 peak is observed at 400 AE 50°C (Figures 5-12h).It is not, however, related to the release of N 2 , as it might seem at first glance.This peak is accompanied by the signal at m/z = 29, released exactly at the same temperature as that at m/z = 14 (Figures 5-12g).An important point is that this peak at m/z = 29 is not correlated with either m/z = 28 or m/z = 30.Therefore, the only reasonable explanation for the observation is to suggest that these peaks represent hydrocarbons (CH 2 and C 2 H 5 ) formed as a result of decomposition of a specific organic compound present in all the samples.Similar explanations can be given to the very pronounced peak observed at 200°C for the signal at m/z = 14 only for ALH 82100 (Figure 7h), since the peak is also accompanied by a pronounced peak at m/z = 30.Thus, this observation can also be explained by the presence of a specific organic compound in this sample which decomposes to form CH 2 and C 2 H 6 .Of course, other hydrocarbons can also be released, for example, methane, which could also contribute to m/z = 16.However, if present, it is obscured by the high signal at m/z = 16 from CO 2 (see Figures 3 and 5-12d).We note here that the amounts of carbon and hydrogen released from the meteorites in the form of hydrocarbon appears to be relatively small; most of which is released in the form of CO 2 and CO and molecular hydrogen correspondingly.
A clear signal at m/z = 34 was detected for all the samples, but it seems not to be associated with H 2 S since it is well correlated with the release of oxygen and not with release of H 2 or sulfur.Figure 15 shows it for Nogoya, but similar correlations are also observed for all other samples where a pronounced broad signal at m/z = 32 is observed (Figures 5d,6d,8d,9d,11d,12d).The relationship between the signals at m/z = 32 and m/z = 34 (∼10, Figure 15) suggests that the latter is not due to molecule of 16 O 18 O since for oxygen the ratio of 16 O 16 O/ 16 O 18 O is ∼250.Thus, a plausible choice for the signal at m/z = 34 is H 2 O 2 .The correlation of the m/z 34 signal, in addition with O 2 , also with the release of water (Figures 5-12c) confirms that formation of the molecule can indeed occur.It is important to note here that no signal at m/z = 34 is observed during release of H 2 O from CaSO 4 Á2H 2 O reference sample, suggesting that H 2 O 2 is not formed in QMS but rather associated with chemical reactions occurring during heating of the meteorite samples.

Carbon Monoxide and Dioxide
Most CO 2 is released at a relatively low temperature in contrast to CO, most of which is released at higher temperature (Figures 5-12b).The fact that the lowtemperature maximum of CO 2 release is observed at 550-650°C suggests a contribution from calcium/magnesium/ iron carbonates often present in CM2 meteorites, which decompose to CO 2 in this temperature range.This is confirmed by stepped combustion data for the CM2 samples (Figure 16) and for Winchcombe in particular (King et al., 2022).The difference in the observed decomposition temperature may be due to the predominant type of carbonate present in each sample.Formation of the rest of CO 2 and CO likely occurs as the result of a chemical reaction between organic material in the meteorites and oxygen-bearing minerals.Some variations in the release pattern of the species can be associated with the difference in mineralogy and organic material distribution in the samples that can affect efficiency of the organic matter oxidation.Nitrogen, which is also a relatively abundant element in carbonaceous chondrites, is closely associated with organic macromolecular material.It can be released as a result of oxidation but has a negligible contribution to the signal at m/z = 28 compared to that derived from CO and at m/z = 14 from CO 2 fragments, as discussed above.

Water
Water (m/z = 17) is released in a wide temperature range with a peak at 300-600°C (Figures 5-12c).According to thermogravimetric analysis (TGA) water from CM meteorites can be released upon heating from different sites and by different mechanisms including release of adsorbed molecular water and as a result of decomposition of (oxy)hydroxides and phyllosilicates, with the latter expected to be its dominating source (Garenne et al., 2014).Our data are generally consistent with these results.However very little, if any, adsorbed water is released below 200°C which is less than expected (Garenne et al., 2014) and this is true for the sample ALH 83100, which was also analyzed by TGA (Garenne et al., 2014).This may be due to loss of the lowtemperature water while the samples were under high vacuum prior to EGA.The maximum water release at ∼300°C from ALH 82100 and Murray apparently suggests that it comes mostly from (oxy)hydroxides.In all other samples most of the water is released in the temperature range 400-700°C corresponding to decomposition of phyllosilicates.The close relationship between release of water and decomposition of phyllosilicates is established by the correlation between the total hydrogen content (mostly in the form of H 2 O, see Table 2) and the phyllosilicate fraction in CM2s (Suttle et al., 2021).

Oxygen
The correlation between O 2 and H 2 O releases (Figures 5,6,8,9,11,and 12d,e) suggests a similar mechanism for their source and evolution.Interestingly this is not observed if H 2 O is released at a much lower temperature (∼300°C), that is, from (oxy)hydroxides, as is seen for ALH 82100 and to some extent for Murray (Figures 7 and 10d,e), but only when H 2 O is released in the 400-700°C temperature range.Thus, it seems that the molecular oxygen that appears along with H 2 O is also associated with phyllosilicate decomposition but not with FIGURE 16.Release pattern of C and δ 13 C variations during stepped combustion of bulk Winchcombe and other CM2 meteorites.In almost all samples, and most clearly for Winchcombe which was analyzed with a high-temperature resolution, a carbon peak is observed at 600°C (less obviously for Nogoya and ALH 82100) with enhanced δ 13 C, that corresponds to decomposition of carbonates.The δ 13 C values at 600°C mainly depend on the mixing ratio of the isotopically light (-25‰ Ä -10‰) meteoritic organic carbon (Sephton, 2014) and isotopically heavy (up to +70‰) carbonate carbon (Grady et al., 1988) that, in turn, depends on the experiment temperature resolution and release patterns of these two components.The presence of carbonates with isotopically heavy carbon is a common feature of carbonaceous chondrites (e.g., Grady et al., 2002).The isotopically heavy carbon observed at T > 1000°C is a signature of presolar SiC seen in almost all carbonaceous meteorites (Grady et al., 2002).(Color figure can be viewed at wileyonlinelibrary.com)Eiler and Kitchen (2004).
h Rubin et al. (2007).decomposition of (oxy)hydroxides and the second-order signal from SO 2 as discussed in the previous section.The release of O 2 at very high temperature (>1300°C) is due to an artifact and not related to the meteorite samples, as was mentioned in the Experimental Technique Section.

Hydrogen
A broad H 2 release within a 200-1200°C interval suggests that it mostly comes from decomposition/ oxidation of the organic macromolecular material, since this is the temperature range where most carbon is released in the form of CO and CO 2 (release of CO 2 from carbonates is seen on the broad total CO 2 release as an additional narrow peak at 600-700°C; Figures 5-12b,e).In more detail this can be seen in Figure 17.For W1, W2, Nogoya, and Cold Bokkeveld most of the local peaks of H 2 are observed at the same temperature as for CO while for ALH83100 and ALH88045 some of them correlate with the release of CO 2 , instead, and for ALH 82100 and Murray, with the release of water.The correlation between H 2 , CO, and CO 2 peaks is pointing to the same release mechanism for these species, which appears to be destruction (oxidation) of the macromolecular material.At the same time the release patterns are generally not correlated with each other, that is, the proportions between the local peaks differ significantly, which can be due to variations in the H/C ratios in different types of organic matter degraded/oxidized at different temperatures.This is confirmed by H/C ratios (Table 2) deduced from the total concentrations of C and H, released as H 2 (see Concentration of the Volatile Species Released During EGA Section) that are in reasonable agreement (except for Murray and Cold Bokkeveld, see below) with the H/C ratios determined in insoluble organic matter (IOM) obtained by dissolution of bulk CM2 meteorites in CsF-HF-HCl (Alexander et al., 2007).The separate lowtemperature peak of H 2 observed for ALH 82100 and Murray present at the same temperature where the water associated with (oxy)hydroxides is released (Figure 17) indicates that this hydrogen may also be associated with decomposition of these minerals.The low yield of oxygen observed in the temperature range (Figures 7d and 10d) indicates that thermal decomposition of water does not play a significant role for the release of hydrogen (see Identification of the Evolved Gas Species Section).

Sulfur Dioxide
Release of SO 2 at high (>1100°C) temperature shows some differences between the meteorite samples (Figure 18).In some samples (W1, W2, Nogoya, and ALH 88045) it has a single peak with maximum at T > 1350°C and is accompanied with a little molecular oxygen including that coming from decomposition of quartz.For the other (ALH 82100, 83100, Cold Bokkeveld, and Murray) samples it is shifted to the lowtemperature side, bimodal (except for Murray) and accompanied with large amounts of O 2 associated with decomposition of quartz reaction tube (see the Experimental Technique Section).However, in general there is no correlation between release of this O 2 and SO 2 , that is, oxygen is not directly related to the release of SO 2 .
The relatively low-temperature SO 2 release (at T < 900°C) has a significantly smaller signal at m/z = 64 than that observed at high temperature and shows highly variable release patterns and temperature ranges for different meteorites.For Nogoya, Cold Bokkeveld, and ALH 88045 SO 2 is evolved between 500 and 900°C, for all other samples it is from 200 to 600°C, and all samples exhibit a series of peaks at different temperatures (Figures 5-12d).
All these observations can be explained in terms of chemical forms of sulfur present in CM2 meteorites.It is known (Airieau et al., 2005;Bates et al., 2020;Burgess et al., 1991;Gao & Thiemens, 1993;Howard et al., 2011Howard et al., , 2015;;Kaplan & Hulston, 1966;Lee, 1993;Monster et al., 1965;Rubin et al., 2007) that in carbonaceous chondrites, and in CM2s in particular, there are three principal chemical forms of sulfur: sulfides (mainly troilite, pyrrhotite, and pentlandite), elemental S and sulfates (gypsum, bloedite, or epsomite).The elemental S is present in both the meteoritic matrix and in the organic material.Sulfide reduced form of sulfur is generally dominated, though in some studies of CM meteorites (Burgess et al., 1991;Gao & Thiemens, 1993;Kaplan & Hulston, 1966) the oxidized (sulfate) sulfur was reported to be the most abundant its form.As it is highly volatile, elemental S is probably released first upon heating compared to its other forms.Since it is detected in the form of SO 2 there must be sufficient oxygen (in different forms) present to oxidize sulfur.Indeed, release of carbon as CO 2 and CO in this temperature range confirms this since these gases are produced from organic or elemental carbon.Elemental sulfur from the organic matter (e.g., sulfonic acid) seems to be released first, followed by release from the matrix when it starts to degrade at higher temperature.This is confirmed by correlations between peaks of SO 2 , CO 2 , and H 2 observed for the releases at 200-300°C.Oxidation of this S and C appears to occur from oxygen present in the organic matter.Release of SO 2 in the 400-800°C temperature range is not clearly correlated with any other species and appears to result from oxidation of the elemental sulfur in the matrix.The complex character of SO 2 release in this temperature interval, with several individual peaks, can be explained by occurrence of the elemental sulfur in different energetic positions in the matrix.However, a contribution of SO 2 from some sulfides cannot be ruled out.
The high-temperature release of SO 2 is evidently related to sulfide and/or sulfate decomposition which Correlation of H 2 releases with CO, CO 2 , and H 2 O for the studied meteorites.For W1, W2, Nogoya, and Cold Bokkeveld most of the local peaks of H 2 are observed at the same temperature as for CO while for ALH83100 and ALH88045 some of them correlate with the release CO 2 , and for ALH 82100 and Murray, with the release of water.(Color figure can be viewed at wileyonlinelibrary.com) occurs at 1100-1300°C depending on its chemical forms and the presence of catalysts.As was mentioned in the Identification of the Evolved Gas Species Section, decomposition of pure gypsum gives O 2 along with SO 2 at ∼1200°C with SO 2 /O 2 signal ratio of ∼0.6 (Figure 13).Therefore, for the samples where the yield of oxygen in the range of SO 2 release (1000-1400°C) is relatively low the release of SO 2 can be attributed to sulfides assuming other types of sulfate also decomposed to SO 2 and O 2 with similar proportions as for CaSO 4 .This is clearly observed for W1, W2, Nogoya, and ALH88045 and to certain extent for Murray where the O 2 signal is significantly lower than that for SO 2 and represent the second-order signal from SO 2 (see previous section).For ALH 82100, 83100, and Cold Bokkeveld this is observed only for the first (relatively low temperature) peak of SO 2 at ∼1250°C (Figure 18).The high-temperature peak of SO 2 release for ALH 83100, 82100, and Cold Bokkeveld is associated with a large release of O 2 most of which is related to decomposition of quartz sample tube.However, the releases of O 2 for ALH 83100 and 82100 clearly show a second release peak of O 2 around 1300°C superimposed on the release from the quartz tube that indicates decomposition of sulfate given the high estimated ratio of O 2 /SO 2 (>1).For Cold Bokkeveld the presence of O 2 associated with the SO 2 release is not as pronounced as for ALH 83100 and 82100 but can be suggested by analogy with the two samples.Thus, SO 2 associated with sulfide (troilite) decomposition is clearly present in all the samples analyzed while SO 2 related to sulfates is observed only in few samples.It is worthwhile to note here that in the case of lunar soils where troilite is the major sulfur compound the release of SO 2 is observed at high (1200-1300°C) temperature (Verchovsky, Anand, & Barber, 2020).The difference in the observed release temperature of the sulfide-associated SO 2 may reflect differences in the morphology and/or sulfide mineralogy of CM meteorites compared to lunar soils.On the other hand, the association of the relatively low-temperature release of SO 2 for ALH 82100, 83100, Cold Bokkeveld, and Murray with the large amounts of molecular oxygen from quartz (Figure 18) may suggest a catalytic effect reducing the sulfide decomposition temperature, since this intensive release of O 2 is promoted by enhanced catalytic action of Pt in the sample bucket that may also play such a role in the decomposition of not only sulfides but also sulfates.An exception is Cold Bokkeveld where the sulfate-associated peak of SO 2 is observed at as high temperature as for the samples with the low oxygen yield.

Concentration of the Volatile Species Released During EGA
The concentrations of the analyzed gases in Winchcombe and other CM2 samples are presented in Table 2 along with those determined by other independent methods for the same or similar meteorites.

Carbon
The total carbon content has been calculated as a sum of CO and CO 2 releases, taking into account the contribution of the second-order signal at m/z = 28 from that at m/z = 44.The correction is based on the results of pure CO 2 measurement (Figure 3) from which the ratio between the main and the second-order signals at m/z = 44 and 28 can be found.The carbon emanates from a mixture of an oxidized meteoritic organic component with δ 13 C of -25 to -10‰ (Sephton, 2014) plus gas from the decomposition of carbonates.CM2 chondrites are known to contain carbonates with δ 13 C up to ∼+70‰ (Grady et al., 1988).The contribution of carbonate to the total carbon content can be assessed from the stepped combustion results (Figure 16) where it is seen as a release of CO 2 at 500-700°C and a shift in δ 13 C to higher values.This was most clearly revealed for Winchcombe when analyzed with a high-temperature resolution.
The relationship between CO 2 and CO is significantly variable (0.67-5.6) in different meteorites (Table 2).In Winchcombe it is the lowest, while in the Antarctic meteorites it is much higher with the highest value in ALH 82100.The non-Antarctic meteorites are all observed falls, hence unweathered as well as Winchcombe.In contrast, the Antarctic meteorites are finds, and known to be affected to various extents by terrestrial weathering, especially the formation of carbonates with δ 13 C ∼ 0 AE 10‰ (Grady et al., 1991).The study found that CM2 chondrites from Antarctica could have as much as twice the carbonate content as non-Antarctic meteorites.This observation suggests that the CO 2 /CO variations could result from the presence of terrestrial weathering products, specifically the decomposition of Antarctic carbonates to CO 2 .
There is good agreement between the concentrations of carbon obtained by QEGA and by stepped combustion (Table 2), except for ALH 83100, where the former gives 2.6 times higher value than the latter.It might be a result of the sample heterogeneity given that a very small aliquot (0.3 mg) was used for QEGA.With the possible exception of Antarctic meteorites, carbonate decomposition does not significantly contribute to the variations in the H/C ratios calculated for total carbon and hydrogen (as H 2 ) contents, discussed in the next paragraph.

Hydrogen
According to our results Winchcombe and other CM2 meteorites contain hydrogen associated with organic material and water from phyllosilicates.Hydrogen from the former source accounts only few percent of its total amount by weight.Most of the hydrogen is related to water.The water, as well as the total hydrogen contents, reasonably agree with the results obtained by independent methods (Table 2).Since we believe that hydrogen released as H 2 and the total carbon (except for that related to carbonate) are associated with meteoritic organic matter, the calculated H/C ratio for our samples should be comparable with that of the organic matter.The IOM in CM meteorites has this ratio within 0.59-0.71(Alexander et al., 2007).For most of our samples the H/C ratios are within or close to this range.Contribution from carbonate or other carbon phases not related to IOM can reduce the ratio to a certain extent but probably not by a factor of 2 as it is observed for Nogoya, Murray, and Cold Bokkeveld.The reason for the discrepancy is not entirely clear but may be related to the formation of hydrocarbons, in which form a part of H 2 is released.This may also be due to the oxidation of part of the molecular hydrogen during the release process.In this case the H/C ratio will be affected much more than the H 2 O content due to the high H 2 O/H 2 ratio, such that, even accounting for the entire discrepancy in H/C ratio reported here compared to previously published ratios by other techniques would result in changes to the reported H 2 O abundance of ∼0.3 wt%, well within the reported uncertainties.

Sulfur
In general, the total sulfur concentrations calculated from the QEGA results are within the range known for CM2 meteorites (Table 2).Most of it is associated with sulfides.Sulfates are identified in three of the seven samples.The concentration of sulfur related to sulfate in the meteorites can be estimated as 1.1, 1.0, and 1.2 wt%, respectively (Table 2).These concentrations are somewhat higher than those associated with gypsum (0.2-0.6 wt% of S) found in few other CM meteorites (Bates et al., 2020;Howard et al., 2015).Sulfates could be products of aqueous alteration on the CM2 parent bodies, although, according to some studies (Losiak & Velbel, 2011;Velbel & Palmer, 2011) a significant part of the sulfur may be due to terrestrial contamination.However, the isotopic composition of oxygen in sulfates from CM and CI carbonaceous chondrites (Airieau et al., 2005) is compatible with extra-terrestrial origin of the minerals.The contribution from the elemental sulfur accounts for from 4% to 32% of its total amount, which is also in the range determined by the independent methods.

FIGURE 1 .
FIGURE 1. Flow rate calibration with gas mixture 1.The slope of the lines gives the gas flow rate in cm 3 STP per second.Reference gas was accumulated in the section with the capacitance monometer.The amount of accumulated gas was found from the known volume of the section and the pressure recorded by the capacitance monometer The plots represent the results for two different gas flow rates (1.02 × 10 -6 on the left and 2.5 × 10 -6 on the right in cm 3 STP per second) set by crimps on different capillaries.

FIGURE 4 .FIGURE 5 .FIGURE 6 .FIGURE 7 .FIGURE 8 .FIGURE 12 .
FIGURE 4. Release of SO 2 (a) and water (b) from known aliquots of CaSO 4 Á2H 2 O versus time at m/z = 64 and 17, respectively.Integration of the curves over time gives the total cps for the known amounts of the gases from which the relationship between the total number of counts and the amount of the gases is established.
FIGURE 17. Correlation of H 2 releases with CO, CO 2 , and H 2 O for the studied meteorites.For W1, W2, Nogoya, and Cold Bokkeveld most of the local peaks of H 2 are observed at the same temperature as for CO while for ALH83100 and ALH88045 some of them correlate with the release CO 2 , and for ALH 82100 and Murray, with the release of water.(Color figure can be viewed at wileyonlinelibrary.com)
FIGURE 14. Release of SO 2 from Nogoya along with releases at m/z = 16, 32, and 34 indicates that the latter represent the second-order signals resulting from the dissociation of SO 2 into ions of 16 O + , 32 S + , and 34 S + .The additional rather abundant second-order signal should be 32 S 16 O + , which was not measured.(Color figure can be viewed at wileyonlinelibrary.com) by dissociation of CO; it could be a result of a combination of 12 C + with H 2 + to form CH 2 + ion or due to formation of a double charged CO ++ ion in the QMS.
FIGURE 15.Release patterns for the species registered at m/z = 17, 32, and 34 in Nogoya.The relationships between the signals suggests that the signal at m/z = 34 appears to represent H 2 O 2 (see text).(Color figure can be viewed at wileyonlinelibrary.com)

TABLE 2 .
Concentrations of volatile species in CM2 meteorites.
The high-temperature release of SO 2 along with the signal at m/z = 16 and 32.The samples corresponding to the upper row plots show weak signals of O 2 from decomposition of quartz, while the samples associated with the bottom row plots display strong such O 2 signals at T > 1300°C.In the samples with the low O 2 signal from decomposition of quartz and particularly for Nogoya, the release curves for m/z = 32 and 16 follow exactly that for m/z = 64 and therefore can be the secondorder signals from SO 2 (see the Identification of the Evolved Gas Species Section for more detail discussion about this).(Color figure can be viewed at wileyonlinelibrary.com)