2.1 Fundamentals of Noble Gas Cosmochemistry – The Sun

The importance of noble gases in the Earth and Planetary sciences has been excellently explained in two previous volumes of the Geochemical Perspectives series by Manuel Moreira and Bernard Marty (Moreira, 2013; Marty, 2020). As Manuel explains, they are geochemical tracers par excellence because they are chemically (almost) inert and thus unaltered by chemical and biological reactions. They are also exceedingly rare in planetary environments, hence tiny additions from a wide variety of processes can be recognised much more easily than for other elements. As Bernard explains, noble gases are also important for understanding the geochemical cycles of other volatiles, including nitrogen and carbon, making them excellent tools to study geological and geochemical processes throughout the history of the Earth and the Solar system. I will come back to some of these topics. To begin with, I will focus on our home star, sketching the early history of analyses of samples containing noble gases from the Sun.

Knowing the abundances of the elements in the Sun is of paramount importance for geo- and cosmochemistry (Palme, 2018). The Sun contains more than 99.8 % of the mass of the solar system and is therefore representative of the composition of the material from which the planets and their building blocks formed. Today, the abundances of most elements in the Sun can be determined by the strength of characteristic absorption lines in the spectrum of the solar photosphere, i.e. the thin surface layer from which we receive the Sun’s light (Asplund et al., 2021). However, already in the 1920s and 1930s, Victor Moritz Goldschmidt used stony meteorite data to tabulate “cosmic” abundances of the elements, recognising that – unlike the Earth – many meteorites were not affected by melting and crystallisation. Indeed, it is one of the most remarkable facts in cosmochemistry that, for most elements, the abundances in one particular and extremely rare class of meteorites, the CI chondrites, agree spectacularly well with modern spectroscopically-derived photospheric values (Lodders, 2020; Fig. 2.1). CI chondrites are very friable and therefore only very rarely survive passage through the Earth’s atmosphere, but their parent bodies are not rare in the (outer) asteroid belt, as indicated by reflectance spectroscopy (DeMeo and Carry, 2013) and recently also by the Hayabusa 2 and Osiris Rex missions to the asteroids Ryugu and Bennu. Therefore, the asteroid belt hosts material which has conserved the elemental composition of the solar nebula, the disk of gas and dust from which the Sun and planets of our solar system formed.

While ultimately our preferred values for solar element abundances should be based on measurements relating directly to the Sun, the CI abundances are often used as a proxy by cosmochemists because for many elements they are more precise than the spectroscopic abundances. However, for a few elements this does not work, as is shown in Figure 2.1. Notably, noble gas abundances in CI chondrites (as well as in any other meteorite class) are many orders of magnitude below their values in the Sun. Also the abundances of H, C, N, and O, which occur mainly in highly volatile forms in the nebula, are depleted in CI chondrites relative to solar values. Actually, Figure 2.1 is slightly deceptive, since the solar noble gas abundances on the abscissa are not actual measured values in the photosphere. Noble gases do not produce absorption lines in the solar photospheric spectrum because energies of the relevant atomic transitions are not reached at the temperature of the photosphere (He was discovered in the Sun, but in the spectrum of the hotter chromosphere during an eclipse). Therefore, other means are required to estimate the noble gas abundances in the Sun. Knowing the isotopic composition of the noble gases in the Sun is also crucial for geo- and cosmochemistry, because – unlike most other elements – the tiny amounts of noble gases in meteorites are not representative of the solar isotopic abundances. As we will see in Section 2.5.5, this also holds for oxygen and nitrogen, which became the two highest priority elements for the Genesis mission that collected solar wind atoms for analyses in terrestrial laboratories (Section 2.5.5).

What are the possibilities for determining noble gas elemental and isotopic abundances in the Sun? Helium is a special case. It is the second most abundant element in the Sun after hydrogen, accounting for almost a quarter of the Sun’s mass (compared to ~74 % for H and a meager ~1.5 % for all elements heavier than He). Because the fusion of hydrogen to 4He is the Sun’s energy source, solar modelling allows for an accurate estimation of the Sun’s He abundance by fitting the models to the Sun’s luminosity. In addition, the Sun is an oscillating ball of gas, and these seismic oscillations provide an accurate He abundance in the Sun’s outer convective zone (the outermost nearly 30 % of the solar radius, which is also relevant for the He abundance in the solar wind). The current best value for the He abundance in the outer convective zone provided by so called helioseismology is ~24.7 % (Basu and Antia, 1995). This is about 10 % less than the initial solar He abundance obtained from solar models, i.e. the abundance 4.6 Ga ago before the onset of fusion of hydrogen into helium in the Sun’s core. This indicates a certain amount of gravitational settling of He (and heavier elements) from the solar surface region towards its interior.

The isotopic composition of the noble gases in the Sun can be reliably deduced by studying the solar wind, as discussed below. However, the solar wind is not a very good candidate for obtaining elemental abundances because the noble gas abundances in the solar wind do not perfectly represent the unaltered photospheric abundances. Understanding the relevant fractionation processes is a major topic in solar physics but requires independent estimates of the abundances in the solar wind source region. For Kr and Xe this can be done by interpolating between the CI chondritic abundances of elements and nuclides with neighbouring mass. The modern understanding of stellar nucleosynthesis (how elements are formed in stars) helps to select particularly well suited nuclides for such interpolations. Modern estimates for Kr and Xe often rely on models of the slow-neutron capture process (“s-process”) that operates in evolved low to intermediate mass stars (Lodders, 2020). For Ne and Ar, interpolations are less reliable and their abundances in the Sun are therefore, e.g., inferred from values obtained spectroscopically in much hotter stars, in particular B-type dwarfs whose abundances of other elements are similar to those of the Sun (Lodders, 2020).

2.2 The Solar Wind – a Cosmochemist’s Perspective

The solar wind (SW) is a continuous stream of charged particles emitted from the Sun’s outer atmosphere, the chromosphere and the corona. The solar wind as a concept was postulated by Ludwig Biermann in 1951 to explain why comet tails always point away from the Sun (Biermann, 1951; Parker, 1958). It consists of electrons, protons (hydrogen nuclei), and ions of all other elements present in the Sun. Solar wind ions have speeds in the range of ~200–800 km/s, with an average flux near the Earth’s orbit of about 2 × 108 protons/(cm2 × s). The solar wind is intensively studied by solar physicists, often with in situ analyses by spacecraft, for example the Solar and Heliospheric Observatory (SOHO). An improved understanding of the solar wind is also a major goal of the Parker Solar Probe, named in honour of Eugene N. Parker, the visionary heliophysicist. In the years to come, this spacecraft will make multiple flybys much closer to the Sun than any previous mission (e.g., Kasper et al., 2021). Determining elemental and, in a few cases, isotopic compositions of the solar wind are only one of many goals of such missions. Mass spectrometers in space can distinguish compositional differences as a function of solar wind speed and other parameters such as different phases in the 11 year solar cycle. Some space missions also measure so called solar energetic particles, ions ejected from the Sun with much higher energies than the solar wind but with much lower fluxes (Reames, 2018). A few additional details of the physics of the solar wind are mentioned in Section 2.5.5 which is devoted to the Genesis mission. While data obtained in situ allow the solar physics community to obtain an ever better understanding of the Sun, even the newest generations of space borne mass spectrometers often do not provide elemental and isotopic composition data precise enough to be truly useful for cosmochemists. This is no wonder since the fluxes of heavier elements are much lower than the proton flux, scaling approximately with their abundance ratio to hydrogen. The Ar flux in the solar wind, for example, is about 300,000 times lower by number than the proton flux.

For cosmochemical purposes, a much larger number of solar wind atoms need to be analysed than is possible in space. The way to do this is by analysing in terrestrial laboratories samples that were directly exposed to the solar wind. Solar wind ions with energies on the order of one keV per nucleon are implanted several tens of nanometres deep into solid matter, sufficient to trap them securely. Suitable artificial samples are the aluminum foils exposed on the Moon by the Apollo astronauts and the ultraclean targets of the Genesis mission (Burnett and Genesis Science Team, 2011). The other possibility are natural samples exposed to the solar wind at the surface of atmosphere-less bodies like the lunar or asteroidal regoliths. With a few exceptions, from the solar wind only noble gases can be detected in natural samples, because for other elements the relative amounts added by the solar wind are too small. The ultrapure Genesis targets, on the other hand, allow for the analysis of a number of elements that is competitive with what mass spectrometers in space can provide but do so at higher accuracies.

Around 1960, the first analyses of implanted solar wind noble gases in several meteorites were reported by laboratories in the Soviet Union and Germany (Gerling and Levskii, 1956; Zähringer and Gentner, 1960; Fig. 2.2). However, the exceptionally high concentrations, especially of the light noble gases He and Ne, were not immediately recognised as being of solar wind origin. They were initially referred to as “Uredelgase” or “primordial” noble gases. A first step towards the recognition of the solar wind origin of this component was made by Signer and Suess (1963), who distinguished two primordial noble gas components in different meteorites (Fig. 2.2b). One component has elemental abundances roughly similar to those expected for the Sun, while in the other the lighter gases are strongly depleted relative to Kr and Xe and expected solar abundances. Signer and Suess termed the first component “solar”, and the meteorites with sometimes orders of magnitude higher concentrations of He, Ne, and Ar than their non-gas-rich counterparts became known as “gas-rich” meteorites. Notably, Signer and Suess (1963) deliberately omitted the vague term “primordial noble gases” for the gas-rich component to avoid the appearance that it is necessarily associated with the early history of the solar system. Because the second component had abundance patterns somewhat resembling those of the terrestrial atmosphere, Signer and Suess called it “planetary”. While this name was a perfectly reasonable choice at the time, it later caused confusion when it was realised that the noble gas abundance patterns of “planetary” gases in meteorites have no direct relevance to the Earth’s atmosphere. As explained in Section 3.2, at least some “planetary” gas components are carried by presolar grains, hence predate the solar system.

In a next step, Suess et al. (1964) proposed that the “solar” component of Signer and Suess indeed represents noble gases from the Sun, implanted by the solar wind into grains on the surface of larger bodies. In particular, they speculated that cometary nuclei might be covered with chondritic material of a “sandbank-like” structure. The comets’ interior ices would carry trapped noble gases in “planetary” proportions, whereas the grains in the sandy surface regions would have trapped the solar wind. Gas-rich meteorites often show a distinct dark-light pattern, with only the dark portions containing solar noble gases (Suess et al., 1964). Eberhardt et al. (1965) provided further strong evidence for an origin of the solar component by ion implantation. Noble gas concentrations in the dark portions of the aubrite Khor Temiki negatively correlate with grain size, i.e. they are proportional to the surface-to-volume ratio of the grains, and gas concentrations in grain size separates etched to varying degrees progressively decrease. As an important detail, this did not hold for the isotopes 3He and 21Ne, which are mainly produced by cosmic ray interactions throughout the entire grain volumes. In an influential review paper, Pepin and Signer (1965) stated that the “planetary” component represents “a residue of primeval gases [occurring in the solar nebula very early in the history of the solar system] modified to varying degrees by diffusive loss”, while the “solar” component represents a relatively late addition from the solar wind.

Hence, thanks to the noble gas pioneers, by 1965 it became accepted that meteorites would allow us to study matter from the Sun directly in our laboratories. Since the Sun constitutes the vast majority of matter in the solar system, this is obviously particularly important not only for noble gases but also for other elements, like O and N, whose elemental and/or isotopic abundances can neither be directly measured nor inferred from meteorites. Trapped solar wind in extraterrestrial samples also allows us, in principle, to study solar history, e.g., potential variations in intensity and composition of the solar wind over time. However, the question of when gas-rich meteorites trapped their solar wind noble gases has remained partly controversial up to this day. Early on, it was often more or less tacitly assumed that all meteorites contain “ancient” solar wind from the early solar system, although evidence for this has never been clear. Yet, today, solar wind noble gases in extraterrestrial samples largely lost the attribute “primordial”, as Signer and Suess (1963) and Pepin and Signer (1965) foresaw.

In any case, meteorites provided a means of measuring the elemental and isotopic composition of noble gases emitted by the Sun at a time when the first studies by space probes yielded parameters such as particle density and energy distribution of the solar wind but not yet much about its elemental or isotopic composition. Signer et al. (1965) proposed to collect solar wind ions in foils exposed on manned spacecraft missions planned in preparation for the Apollo lunar landings, or, alternatively, on the Moon itself. They pointed out that recent developments in noble gas analysis, driven at the time primarily by studies of meteorites, would certainly permit analyses of He, Ne, and Ar from the solar wind. This estimate was based on the flux of hydrogen in the solar wind measured by space missions and on the assumption of the then best estimates of element abundances in the Sun for the solar wind. As mentioned above, this idea of my doctoral thesis advisor-to-be Peter Signer eventually led to the very successful Apollo Solar Wind Composition (SWC) experiment by the University of Bern, led by Johannes Geiss (Fig. 2.3). Interestingly, Signer et al. (1965) also considered applying a high voltage to the foils, which would have greatly increased the amounts of ions collected and might have allowed the detection of Kr, Xe, and even H, N, and O. This idea was later realised in the Genesis mission (see Section 2.5.5).

2.3 My First Studies on Apollo Samples

The days and months after the first Apollo samples reached laboratories around the world must have been extremely thrilling for everyone involved. There they were, the first extraterrestrial samples ever brought back by humans, waiting to reveal the secrets of the Moon and more. Somewhat unfortunately for me, I missed that time by just a few years. Therefore, in my early years, I was sometimes a little envious when Herbert Funk and Heiri Baur told how, after degassing the first samples, they would stare at the chart recorder while waiting for signals on masses four or twenty, and how excited they were when the pen went off scale. Phone calls to other noble gas laboratories, for example in Bern or Mainz, confirmed their observations. The lunar regolith is indeed full of noble gases and the solar wind is obviously their main source (e.g., Pepin et al., 1970; Hohenberg et al., 1970; Fig. 2.4a).

Already the first data from the aluminum foils exposed by the Apollo 11 astronauts showed that the experiment had been successful. It yielded a value for the 4He flux in the solar wind (Bühler et al., 1969), while foil data from later Apollo missions allowed the determination of precise elemental and isotopic compositions of He, Ne, and Ar in the solar wind (Eberhardt et al., 1972; Geiss et al., 2004; Fig. 2.3). Perhaps the most important result of the Apollo solar wind composition experiment was the reliable determination of the 3He/4He ratio in the solar wind. In the 1960s, the D/H ratio in terrestrial oceans of ~1.6 × 10-4 was thought to represent the “cosmic” or protosolar deuterium abundance. In this case, the 3He/4He ratio in the solar wind should have been higher than 10-3, because deuterium had been converted to 3He in the very young Sun. However, meteorites thought to contain implanted solar wind noble gases yielded lower 3He/4He values on the order of 4 × 10-4. These lower values were confirmed by the Apollo foils, leading Geiss and Reeves (1972, 1981) to conclude that deuterium in Earth’s water is enriched by almost an order of magnitude relative to the protosolar composition, due to low temperature reactions in interstellar clouds. This finding set new constraints on how the Earth acquired its water and other volatiles.

Analyses of samples from further lunar missions confirmed the Apollo 11 results and showed that the regolith is extremely rich in solar wind noble gases to a depth of at least 2.4 m but very likely further down (e.g., Bogard et al., 1973). However, these studies also made it clear that the noble gas record in the lunar regolith is not easy to read. Isotopic compositions of the solar wind component were similar but not identical to those measured in the Bern aluminum foils and elemental ratios in the regolith samples indicated losses of the lighter noble gases. Some minerals, such as ilmenite, are relatively retentive, while others, like plagioclase, are very leaky, especially for He and Ne. For me this meant that by the time I began to work on lunar samples in 1976, the initial excitement about the Apollo samples had already faded a bit. I could take the solar wind origin of the gases we extracted from bulk soil samples and mineral separates as an obvious fact, and it was only later that I realised that this finding had been a major scientific break through only a decade earlier. Thus, I soon found myself on the sometimes winding road to an ever improved reading of the noble gas record conserved in the lunar regolith.

The first scientific paper I was involved in discussed the mineral-specific retentivities of implanted solar noble gases in the major minerals in lunar soils. Unforgettable (on this and many later occasions) are the numerous discussions and editing sessions with the entire group in Peter’s office, which was known as “The Smokehouse”. At the end of a session, the ashtray was filled with the remnants of several loads of pipe tobacco, each actually only smoked to a minor extent. Perhaps the ambience was not really conducive to a pleasant smoking experience. In any case, it led to an important contribution (Signer et al., 1977) that made it possible to better quantify elemental and isotopic fractionation processes of noble gases in the lunar regolith. However, to me this seemed a bit boring at the time. Also, a few years later at my first Lunar and Planetary Science Conference (LPSC, or “The Houston Meeting” to me) I had the impression that work on noble gases in Apollo samples was no longer considered to be at the scientific forefront, which by then was mainly represented by studies of noble gases in tiny acid resistant residues of bulk meteorites. This feeling was reinforced when the Zürich team visited the California Institute of Technology (Caltech) in Pasadena after the meeting, where Gerald (Jerry) Wasserburg was throwing a party for his Swiss friends. At that moment, I certainly did not realise what an honour it was to be invited to the home of the perhaps greatest geo- and cosmochemist of the time. The living room was decorated for the occasion by a three metre wide Bern flag, a reminder of Jerry’s sabbatical in Bern a few years earlier. What I did realise, though, was that Jerry outright criticised Peter’s current research as “pedestrian science”. We did not like that, of course, but I believe this criticism helped spark my ambition to do more than pedestrian science. A dozen years later, at another Houston Meeting, I happened to be sitting next to Jerry, waiting to present our first direct analyses of primordial noble gases (the “Q-component”) by in vacuo online etching of meteorite samples (Section 3.2.1). When I returned to my seat after my presentation, Jerry whispered: “A very good talk!”. At that moment, I remembered the other evening in Pasadena. In retrospect, it gives me a really good feeling to note that noble gas research on lunar samples is thriving again to this day, in stark contrast to my impression in the late 1970s. Obviously, people who were not born in the days of Apollo still find the treasure trove filled with Apollo samples exciting.

My first LPSC talk in 1979 addressed a topic that I continue to work on up to this day: what can the lunar regolith (and perhaps gas-rich meteorites) tell us about possible long term variations in the Sun that would be reflected in the composition of the solar wind? In short, as summarised by Wieler (2016), I now believe that most of the claimed evidence for temporal changes in the solar wind composition has not withstood the test of time. Even the one temporal change I still defended in 2016, namely the about two fold increase in the Kr/Xe ratio in the solar wind over the past few billion years, may need to be reconsidered. I will in the following refer to such possible very long term variations as “secular” changes. But let’s start at the beginning.

Three major problems we face in reading the lunar record are (i) how can we know when a sample acquired its share of solar noble gases, i.e. the “antiquity” of a sample, a term coined by John Kerridge (Kerridge, 1980), (ii) how can we be sure that noble gases (or nitrogen, see next section) residing at grain surfaces really came from the solar wind and not from some other source, and (iii) how can we be sure that differences in the composition of different samples reflect differences in solar wind properties and not modifications upon or after trapping. These problems will guide us in the following considerations.

My first publication as first author was the abstract that accompanied the LPSC presentation just mentioned (Wieler et al., 1979). Its title began with “The solar wind half an aeon ago:…”. We studied mineral separates and bulk samples from about 2 m below the surface of the lunar regolith. Samples from the Apollo 15 deep drill core and other subsurface samples all contain solar wind noble gases, although the penetration depth of solar wind ions into solid matter is only a few tens of nanometres. The noble gases thus testify to the fact that the regolith is a very dynamic environment: almost every grain taken at any depth down to at least 2.4 m had once been at the immediate surface of the Moon. At the Apollo 15 site, this vigorous mixing must have taken place no later than about 500 Ma ago, since G. Price Russ, Donald Burnett, and Jerry Wasserburg (Russ et al., 1972) had shown that the regolith at the Apollo 15 landing site had been undisturbed for at least almost 500 million years, except for the topmost few cm. The Caltech team had measured isotopes of Sm and Gd, whose abundances get modified by interactions with the galactic cosmic radiation (details in Section 3.1). Hence the samples we studied contained solar wind implanted at least half a billion years ago. The trapping could also have occurred earlier and not all grains of a sample were necessarily irradiated at the same time. However, no atom from the solar wind in these samples was implanted later than about 500 Ma ago, which is already a remarkable constraint (consult Box 2.1on the antiquity of lunar regolith samples for more details on the important problem of how old the solar wind in a given lunar sample may be). The conclusion of Wieler et al. (1979) was perhaps rather unspectacular: the relative He, Ne, and Ar abundances and the 20Ne/22Ne ratio in the solar wind half an aeon ago (or earlier) were not different from present day values outside the limits of uncertainty. Nevertheless, I felt that this conclusion of solar wind constancy deserved some attention because at the same time several other notable secular variations in solar wind composition were postulated (Kerridge, 1980). The most important of these will be discussed in the next section.

Text Box 2.1 – The Antiquity of Lunar Regolith Samples

Studies attempting to obtain information about possible long term variations in solar wind composition face the problem of how to determine when in the past a lunar regolith sample trapped its share of solar wind atoms. John Kerridge called this the “antiquity” of a sample (Kerridge,1980). This is a difficult problem. Kerridge et al. (1991) discuss several antiquity measures, most of which are semi-quantitative at best. In a few favourable cases, cosmic ray exposure ages (see Section 3.1) will provide accurate upper bounds on the antiquity of a sample. This is the case when several samples from a rim of a relatively fresh crater have the same exposure age, which then very likely dates the formation age of the crater. Such samples were brought from larger depths to near the surface of the regolith upon crater formation, so all of their grains will have trapped their complement of solar wind between the time indicated by the cosmic ray exposure age and today. All such known samples have relatively young antiquities of ~2–100 Ma. Many solar wind-bearing lunar samples are regolith breccias, that is, compacted soil. Many of these breccias were undoubtedly irradiated by the solar wind billions of years ago, as can be deduced with the semi-quantitative antiquity indicator 40Ar/36Ar discussed in the next paragraph.

“Parentless” 40Ar is found in the surface layers of grains from every solar windbearing lunar soil. It is commonly believed to be radiogenic 40Ar from the decay of 40K (T1/2 = 1.27 Ga) that has been degassed from the lunar interior. Atmospheric 40Ar atoms can become ionised and accelerated by the solar wind-induced electromagnetic field (Manka and Michel, 1970). Many of them will get lost from the atmosphere within a few months but some will be implanted into grains at the lunar surface along with solar wind ions, including 36Ar. The 36Ar serves as a measure of the time a sample was exposed at the lunar surface. So, if the 40Ar concentration in the lunar atmosphere shows a secular variation, 40Ar/36Ar should vary with a sample’s antiquity. Eugster et al. (2001) and Joy et al. (2011) present calibrations of the 40Ar/36Ar ratio as a function of solar wind antiquity based on a series of samples whose antiquity was inferred by various independent methods, including 39Ar-40Ar ages. Some remarkable samples also contain parentless xenon from the decay of the now extinct nuclides 129I and 244Pu and must have been irradiated very early. Quite remarkably, the trend described by the 40Ar/36Ar ratio roughly follows the curve describing the decay of 40K, with values of around 20 for the samples with the highest antiquities, but only about 0.5 for the samples thought to have trapped solar wind within the last few or last few ten Ma. This may be more surprising than is often admitted, as I discuss in Wieler (2016). The more or less parallel decrease of the 40Ar/36Ar ratio and the 40K decay curve seems to suggest that the 40Ar in the lunar atmosphere comes entirely from the instantaneous complete degassing of some part of the Moon, and that the degassing volume has hardly changed over most of the Moon’s history. At the same time, the flux of solar 36Ar would have to have remained rather constant over the past several billion years. The latter requirement may not be too problematic. However, as far as 40Ar is concerned, I believe that one might as well expect that – all other things being equal – the amount degassing from the lunar interior should to first order be proportional to the total 40Ar produced (and thought to be still largely present) in the crust. If so, the 40Ar/36Ar ratio in regolith samples should be higher today than it was in the past. Obviously this expectation is not reflected in the data at all. It is thus questionable to me whether the observed secular decrease of 40Ar/36Ar does indeed reflect the decay of 40K rather than a large temporal decrease of the degassing efficiency of the lunar crust, e.g., due to a lower frequency of moonquakes or impacts today or decreasing diffusion in a secularly cooling crust. In my view, the evolution of the 40Ar/36Ar ratio in lunar samples of different antiquities is still poorly understood. Nonetheless, the 40Ar/36Ar ratio is a useful semi-quantitative antiquity indicator. It is also important to recognise that a sample’s antiquity may be poorly defined due to multiple episodes of exposure at the surface.

2.4 The Lunar Nitrogen Saga

Spectacular variations in the nitrogen isotopic composition in lunar regolith samples have been reported since the mid-1970s. John Kerridge, then at the University of California in Los Angeles (UCLA), and Richard Becker and Robert Clayton at the University of Chicago observed that the 15N/14N ratios in different samples varied by some 15 % and correlated with the sample’s cosmic ray exposure age (Kerridge, 1975; Becker and Clayton 1977). As it seemed reasonable to assume that the cosmic ray exposure age of a sample is a rough indicator of its antiquity – old exposure ages meaning old solar wind – it seemed that the proportion of the heavy isotope 15N in nitrogen trapped in lunar samples increased dramatically over the past one or several billion years. Later analyses, in which the nitrogen was released in several steps by gradually increasing the extraction temperature, revealed two different trapped N components. The first, released at relatively low temperature, was high in 15N, the second, released at higher temperature, was lower in 15N. The difference in the 15N/14N ratios (δ15N) was as much as 30 % (Becker and Clayton 1977; Thiemens and Clayton, 1980). Figure 2.4b shows nitrogen isotope release patterns of regolith breccia 79035 obtained at the University of Minnesota by Urs Frick and co-workers together with data obtained earlier in Chicago by Mark Thiemens and Bob Clayton. These release patterns were not easily explained, but the leading hypothesis at the time – and later – was that the 15N/14N ratio in the solar wind had increased dramatically over time. One presumed mechanism was the break up (spallation) of oxygen atoms in the Sun’s outer convective zone by energetic particles (Kerridge, 1975; Clayton and Thiemens, 1980). One of the main products would be nitrogen with a 15N/14N ratio much higher than the original solar value, leading to a secular increase of δ15N in the solar wind (this should not be confused with cosmogenic nitrogen produced at the lunar surface and manifested in the highest temperature release fractions in Fig. 2.4b).

In Zürich we had always been sceptical of this interpretation. The idea that the variable δ15N in lunar soils reflected a secular change in the solar wind seemed to contradict another observation: the proportion of nitrogen in lunar samples relative to that of Ar from the solar wind is about an order of magnitude higher than the ratio assumed for the Sun, i.e. (N/Ar)Moon >> (N/Ar)Sun. If this problem was addressed at all, it was usually explained in terms of a much better “retentivity” of N compared to that of Ar, i.e. in lunar samples Ar from the solar wind was assumed to be much less tightly bound than solar wind nitrogen. However, while our work (Signer et al., 1977, see above) had shown that the He and Ne retentivities are very mineral specific, Ar was retained equally well in all major minerals, and the Ar/Kr ratio was close to the accepted solar ratio. This would not have been expected if most of the solar wind Ar had been lost in a way that would have changed the N/Ar ratio so dramatically.

So, that same year, 1979, when I first attended the LPSC, Peter Signer encouraged me to visit the conference “The Ancient Sun – Fossil Record in the Earth, Moon, and Meteorites”. This was the first time I attended one of these small topical meetings which often are so rewarding, especially for young scientists. The meeting in Boulder, CO, allowed me for the first time to get in direct touch with many of the leading figures in N and noble gas research, such as John Kerridge, Robert (Bob) Pepin, Robert (Bob) Clayton, Kurt Marti, and many others. Nitrogen in lunar samples was one of the major topics. Kerridge and Clayton pointed out that the 15N/14N variations lacked a satisfactory explanation but again argued that spallation of oxygen in the solar surface region was the best bet (Kerridge, 1980; Clayton and Thiemens, 1980). If I remember correctly, I did not dare to contradict this view in the public discussion. After all, we had not performed any nitrogen analyses ourselves in Zürich (and have not done so to this day). However, in a private discussion with Bob Clayton, I mentioned our reservation, probably without impressing him too much. Nevertheless, he suggested that I ask NASA for a few grams of lunar regolith breccias 79035 and 79135, which he estimated to contain solar wind implanted perhaps 2.5 Ga ago (Clayton and Thiemens, 1980). These two samples have become cornerstones of my own work on lunar samples, although not primarily for N but for noble gases.

Nitrogen in lunar samples remained a controversial topic. Jim Ray was another young colleague I met in Boulder and hiked up to the Continental Divide with after the meeting. Unlike me, he had something to present on lunar nitrogen. He and his PhD thesis supervisor Dieter Heymann (Ray and Heymann, 1980) proposed that the young Sun might have been polluted by nitrogen from a planetary nebula. Geiss and Bochsler (1982) dismissed this idea, as well as a secular change of the N composition in the solar wind due to oxygen spallation or thermonuclear reactions. Instead, they proposed that an isotopically very light additional N component is added to the lunar regolith. At the 1989 “Sun in Time Conference” in Tucson, AZ, Geiss and Bochsler (1991) suggested that nitrogen from Earth’s upper atmosphere might have reached the early Moon as an “Earth wind”, a suggestion later also made by Ozima et al. (2005). The “Sun in Time Conference” was a meeting in the same spirit as “The Ancient Sun” ten years earlier, and in a paper published in the proceedings volume of this meeting, Kerridge et al. (1991) still had to conclude that lunar nitrogen remained a conundrum.

Such was the situation in 1998, when Bernard Marty kindly invited me to work for a while in his laboratory at the Centre de Recherches Pétrographiques et Géochimiques (CRPG) in Nancy. I proposed to attempt to measure argon and nitrogen concentrations and isotopic compositions in single ilmenite grains from an Apollo 17 soil. Franck Humbert and Bernard Marty had developed an analytical protocol that allowed the simultaneous analysis of tiny amounts of noble gases and nitrogen by static mass spectrometry. “Static” means that vacuum pump valves are closed before a sample’s gas is introduced into the mass spectrometer. For noble gases this is routine because chemical “getter pumps” can be used to reduce the background of chemically active gases (H2, CH4, CO2, etc.) to a level that does not compromise the analyses of minute amounts of noble gases. This is much more difficult with nitrogen which reacts with other elements and would therefore be taken up by the getter pumps before it could be measured. But obviously, the very small amounts of N contained in single grains and the requirement to measure not only its isotopic composition but also its amount made static N analyses a necessity. In 1981, Urs Frick and Bob Pepin were the first to develop static N analysis for meteorite samples and later to apply this method also to lunar soils. The system developed a few years later by Humbert and Marty in Nancy remains one of the few facilities allowing static nitrogen analysis, and was ideally suited for the lunar grain studies I had in mind.

The lunar ilmenite measurements were not my first single grain analyses. As described in the next section, I had already analysed He, Ne, and Ar in single grains from two lunar soils for my doctoral thesis, and later in Charles Hohenberg’s laboratory in St. Louis we measured all five noble gases in a further series of grains (Section 2.5.4.). But in 1998 in Nancy I made my first (and only) own nitrogen measurements, with great support from Franck Humbert and Laurent Zimmermann. I very much enjoyed the hospitality of Bernard and the whole team at CRPG, which was celebrated, among other things, with mirabelle plum liqueurs at Nancy’s beautiful Place Stanislas. The data provided clear evidence in support of the suspicion that the overwhelming part of the nitrogen in lunar soils does not come from the solar wind (Wieler et al., 1999). The 14N/36Ar ratios in the single grains varied by more than a factor of 400. Also the 36Ar amounts varied by more than two orders of magnitude from grain to grain, reflecting their individual exposure histories to the solar wind. But in contrast to Ar, the N amounts in different grains varied by no more than a factor of six. Since the earlier analyses in St. Louis on single grains from the same samples had shown almost constant ratios of Ar/Kr/Xe from the solar wind (Wieler et al., 1996), this indicated that most of the N in the grains must have come from some other source. Therefore, we concluded that the hypothesis that “the lunar regolith testifies to a secular variation of the N isotopic composition in the solar wind of ~30 % becomes thus highly questionable” (Wieler et al., 1999). I was pleased to find that this conclusion soon became widely accepted. A few years later, the work in Nancy was continued with Ko Hashizume, who did detailed stepwise heating analyses of 14,15N and Ar in single lunar grains and proposed a micrometeorite origin for the non-solar nitrogen component (Hashizume et al., 2002). Previously, Ko – together with Marc Chaussidon – had performed nitrogen isotopic analyses by secondary ion mass spectrometry (SIMS) using the instrument in Nancy and had argued for a low abundance of 15N in the solar wind, with a δ15N value of less than -240 ‰ (Hashizume et al., 2000). This conclusion was brilliantly confirmed a few years later by Genesis, as discussed in the next few paragraphs.

Given that it was becoming increasingly clear that most nitrogen in lunar soils had a source other than the solar wind, and the undisputed fact that even meteorites rich in solar wind noble gases were not suitable for measuring the solar 15N/14N ratio, the Genesis solar wind collection mission became a unique possibility to determine this ratio, a value of fundamental importance for cosmochemistry. I will discuss Genesis in detail in Section 2.5.5, but it is appropriate to emphasise here that – despite the crash landing of the sample return capsule – among many other successes, the mission achieved its stated two main goals, the determination of the oxygen and nitrogen isotopic compositions in the solar wind. I had been involved in the preparation of Genesis since the late 1990s when Don Burnett, the mission’s Principal Investigator, established the annual Genesis Science meetings on the Sundays prior to the LPSC. So I knew that determining the nitrogen isotopic composition of the solar wind was the second most important goal of the mission, and I suggested to Bernard Marty that he contact Don and join the team. So he did, which perhaps became one of my most important contributions to planetary sciences. I am quite proud of the fact that Bernard acknowledges that I was the one who introduced him to cosmochemistry, after exclusively studying terrestrial volatiles in his early scientific career.

In a first study, Marty and co-workers (Marty et al., 2010) measured N isotopes in the gold plated steel cross used to mount the different collector materials in the Concentrator target of Genesis. As explained in Section 2.5.5, in this target the solar wind ion flux is enhanced several ten fold by electrostatic fields. Nevertheless, at best some 4 % of the measured nitrogen came from the solar wind; the rest was terrestrial contamination. However, Marty and co-workers could combine the N data with analyses of amounts and isotopic composition of solar wind Ne measured along arms of the “gold cross” in Nancy and by Veronika Heber in Zürich. This allowed them to extrapolate the measured N data to the solar wind N/Ne ratio and clearly showed that the 15N/14N ratio in the solar wind is considerably lower than in the terrestrial atmosphere. Instead, the best estimate of (2.26 ± 0.67) × 10-3 was found to be consistent with the Jupiter value. In a second study, Marc Chaussidon, Bernard Marty, and colleagues (Marty et al., 2011) used the Cameca 1280 SIMS instrument at CRPG in Nancy to analyse a SiC target from the Concentrator. This allowed them to obtain a much more precise value of (15N/14N)SW = (2.18 ± 0.02) × 10-3, which is about 40 % lower than the terrestrial atmospheric value. These beautiful data allowed Marty et al. (2011) to note that: “This result demonstrates the extreme nitrogen isotopic heterogeneity of the nascent solar system and accounts for the 15N-depleted components observed in solar system reservoirs”. This heterogeneity is reviewed by Füri and Marty (2015). Hence, the first part of the lunar N conundrum has been definitively solved: neither the 15N/14N ratios observed in some lunar samples which are higher than the terrestrial value, nor the low δ15N values of around -200 ‰ found in other lunar samples represent the pure solar wind. However, what does or do the non-solar nitrogen component or components ubiquitous in lunar samples represent? This question is still being debated. Meteoritic or cometary contributions remain viable candidates (e.g., Hashizume et al., 2002; Mortimer et al., 2016). Another possibility is a contribution from Earth’s upper atmosphere, as proposed by Geiss and Bochsler (1991) and Ozima et al. (2005). It is almost a truism that documented samples from the far side of the Moon would be highly desirable for further progress, as noted in many papers in this and related contexts. Note that reviewer Bernard Marty informed me that Frank Podosek and he did not find any systematic variations between N/36Ar and δ15N values in a series of 12 lunar meteorites. As about half of those should originate from the backside, this does not support a substantial nitrogen contribution on the lunar near-side from an Earth Wind.

Did I just say that the conundrum of the lunar nitrogen is largely solved? While I believe that the overwhelming majority of the scientists in the field would agree, the person who collected some of the crucial samples for this discussion sees it differently. In 2019 – forty years after the Boulder meeting on the Ancient Sun – probably the largest crowd ever in the same lecture hall at an LPSC meeting was listening to NASA astronaut and geologist Harrison H. Schmitt’s lecture celebrating the 50th anniversary of the Apollo 11 landing. Jack Schmitt is one of the two last humans (so far) to walk on the Moon and his talk was very entertaining. However, when he turned to the findings about the Sun obtained from nitrogen isotopes in lunar samples, I became uneasy. He repeated at length the old view that the isotopic composition of solar wind nitrogen has varied by 30 % and even suggested that the cause of this change may have triggered the Cambrian explosion of life on Earth some 550 Ma ago. He dismissed the Genesis data by Bernard Marty and co-workers as being irrelevant, since they disagreed with the lunar samples. Toward the end of the lecture I was getting pretty nervous, because probably not too many people in the lecture hall were as closely involved in the topic as I was. I was relieved when Don Burnett (who was sitting next to me) grabbed a microphone first and got things right, and in a much more elegant way than I could have done.

2.5 The SEP Myth, the FIP Effect, and Genesis

Let me jump back to the days in 1980 when I was about to finish my doctoral thesis. The Sun emits not only solar wind ions with speeds of a few hundred km/s, corresponding to energies of ~1 keV/amu, but also so called solar energetic particles (SEP) with much higher energies of tens of MeV/amu. These are not emitted continuously like the solar wind, but as Coronal Mass Ejection events during solar flares. Solar energetic particles were observed as early as the late 1970s with mass spectrometers on space missions. In lunar samples and meteorites, the effects of these particles are known in the form of so called solar flare tracks, i.e. lattice damages created by SEPs mainly of iron group elements (Fig. 2.5) that can be made visible by chemical treatment in the laboratory. Solar energetic particles of these heavy elements penetrate hundreds of microns deep into solid matter, much deeper than solar wind ions, but the generation of latent solar flare tracks still requires grain exposure directly at the regolith surface. It therefore made sense to compare track densities and solar wind noble gas concentrations in the same samples. Gérard Poupeau, then at the Centre des Faibles Radioactivités in Gif sur Yvette (France), measured track densities, and I analysed noble gases in mineral separates from many samples as well as in single grains from two samples (Wieler et al., 1980). Later, Gérard also taught me how to make tracks visible by etching and how to count their densities by light or electron microscopy (Wieler et al., 1983). The comparison showed – at least on a whole grain scale – that solar wind Ar (and Kr and Xe) concentrations are not in saturation even in heavily irradiated lunar samples, i.e. do not reach an equilibrium value where for each newly implated atom from the solar wind another one is lost, for example by grain surface sputtering. I would not any longer put my money on another conclusion we reached in the 1980s. We claimed that the flux ratio of high energy solar particles to solar wind decreased by perhaps a factor of two over the past 1 – 3 billion years.

The first two publications on the Ne isotopic composition of SEP events by space missions reported nearly identical 20Ne/22Ne ratios of ~7.7 (Mewaldt et al., 1979; Dietrich and Simpson, 1979), much lower than the solar wind value of 13.7 as measured in the Apollo SWC aluminum foils. Although the SEP data had large uncertainties, it seemed clear that the high and low energy solar corpuscular radiation differed in their isotopic composition. Already before the first space mission data became available, there was controversy about whether the noble gases trapped in lunar soils and gas-rich meteorites represent just the solar wind with its known energy of around 1 keV/amu, or – in addition – a second, higher energy component implanted at larger depth than the common solar wind ions. An early advocate of this view was David Black, then at the University of Minnesota and later director of the Lunar and Planetary Institute in Houston. He studied gas-rich meteorites and lunar samples by stepwise heating (Black and Pepin, 1969; Black, 1972). At relatively low temperatures, Ne with a 20Ne/22Ne ratio around 12.5 was released, which he labelled component Ne-B and identified it with solar wind Ne (Fig. 2.7). As we will see below, Ne-B eventually turned out to be solar wind Ne isotopically fractionated upon implantation and grain surface sputtering. While it is thus not a pure component sensu stricto, Ne-B is still important, in particular in terrestrial noble gas studies (e.g., Moreira, 2013). Higher temperature steps, however, fell on a straight line (marked as a light grey band in Fig. 2.7) between a point labelled by Black as Ne-C in the upper left hand and cosmogenic Ne in the lower right hand corner. In such three isotope diagrams, data points falling on a straight line are usually interpreted as a mixing line of two end member components at or somewhere beyond the two ends of the line. Hence, Black concluded that Ne-C with 20Ne/22Ne = 10.6 ± 0.3 must represent a true solar component isotopically different from solar wind Ne and implanted with higher energy than the solar wind. He assigned Ne-C to the low energy fraction (1–10 MeV/amu) of energetic particles emitted during solar flare events, and concluded that SF-Ne, as it was then called, has a lower 20Ne/22Ne ratio than the solar wind. When the first Ne isotopic data of SEPs directly measured in space indeed gave significantly lower 20Ne/22Ne ratios than the solar wind value, David Black’s hypothesis that lunar samples and gas-rich meteorites contain measurable amounts of higher energy solar ions in addition to the solar wind became very popular. However, there was a serious problem with this. Ne-C accounted for up to several tens of percent of the total trapped Ne in the samples, which meant that the flux of SF-Ne would have to be orders of magnitude higher than what could be expected from extrapolations of the high energy SEP flux measured in space. Could this be the case?

2.5.1 Noble Gas Release by in Vacuo Etching

We decided that the most promising way to solve this puzzle would be to improve the quality and depth resolution of concentration profiles of trapped noble gases in lunar grains. So far, depth profiling had been attempted mostly by gas extraction in several temperature steps. In some cases also grain surface layers of different thickness were removed by etching multiple aliquots of the same sample. Information about gas concentrations and compositions in the etched layers had to be inferred by difference. We wanted to improve this by developing a device allowing stepwise etching of grain separates under “vacuum” in a gas extraction line connected directly to the mass spectrometer, as this would allow measuring separately the gases released in each etching step. This decision may have shaped my own research agenda more than any other of our technical developments. In short, for almost twenty years, our team gradually became perhaps the best known advocates of the idea that the lunar and meteoritic regoliths testify to a very substantial contribution from a high energy solar noble gas component. And then, in 2006, all of a sudden we had to abandon our beloved idea, based on the initial analysis of a sample flown as part of the Genesis mission that had been specifically selected to study the high energy component. What was comforting was that we were able to do this in a paper in Science, actually the first article reporting scientific data from Genesis. How this all came about is described next. The in vacuo etching technique later also provided a unique way to study primordial noble gases in meteorites. This will be described in one of the later sections.

It may sound crazy to connect a noble gas extraction system containing strong acids to a mass spectrometer, with only a few valves and some cold traps and chemical getters in between. Perhaps it is indeed crazy, in particular because one of the acids we often use is HF, whose atomic mass of 20 equals that of the main Ne isotope. However, in all these years, we have had to deal with a serious increase of the background on mass 20 only once. This was after a rather silly mishap (if I remember correctly, we forgot refilling liquid nitrogen in a cold trap), but with no major consequences. Our first in vacuo etching line was made of glass and allowed etching of plagioclase and pyroxene with nitric acid. The first data obtained with this line were published by Wieler et al. (1986), and we often refer to this gas release and analysis technique as CSSE (for Closed System Stepped Etching). However, we also wanted to analyse ilmenite, the most noble gas retentive among the major minerals in lunar soil. This required a line allowing HF as the etching agent. Therefore, around 1990 we built the “gold line”, which is still in use today. Both lines are described in a paper presented by Peter Signer in 1991 at a meeting in Durango, CO, honouring Alfred O. C. Nier, the “Father of Mass Spectrometry”. As mentioned above, Peter had worked in Al’s lab in Minneapolis for about seven years, and he greatly admired him. A Festschrift in his honour was therefore the right place to present the technical details of our CSSE lines (Signer et al., 1993). The lines are shown in Figure 2.8. A typical reaction of dear friends of mine around the globe is: “Oh, only the Swiss can afford that, with all their bank vaults filled with gold”. Of course, it’s not quite like that. We had initially planned to use steel tubes, gold plated on the inside. But then Heiri Baur argued that pure gold tubes would not only allow for simpler construction and possibly lower blanks, but would probably also be more economical, since the gold from parts to be replaced could be recycled. Heiri was right. It turned out that the overall costs of the gold over all those years adds up to no more than what scientists spend for a few electronics parts of their instruments. The CSSE technique has also become a great success thanks to the excellent technical staff in our group, mainly Mathias Gerber and – more than anyone else – Urs Menet. The gold line even motivated Urs to become an amateur jeweler. Some of his friends received a gold wedding ring made by Urs.

Closed system stepped etching is not an analytical technique for the impatient scientist. The data shown in Figure 2.9 (left panel) were obtained using nitric acid as the etchant. Etch times per step ranged from less than an hour to several days, resulting in total run times of several weeks. Other analyses even took up to many months. The reward for this time consuming procedure is the fact that gases are released by dissolving the sample, so no elevated temperatures are needed that could lead to isotopic fractionation during gas release. Better depth resolution compared to stepwise pyrolysis or stepwise combustion (the latter in the presence of small amounts of oxygen) can also be expected. Below we will also see that CSSE allows a carrier-selective gas extraction.

Figure 2.9 includes the data from two of several etch runs of plagioclase and pyroxene separates (Wieler et al., 1986) and an ilmenite separate measured by Benkert et al. (1993). The first steps of the plagioclase analyses released Ne with a 20Ne/22Ne ratio somewhat below the value of 13.7 for solar wind Ne, known from the Apollo aluminum foils. Most remarkably, later steps always fell on a well defined straight line, pointing towards the composition of cosmogenic Ne in the lower right and extrapolating to a 20Ne/22Ne ratio of ~11.3 in the upper left. A pyroxene separate also etched by HNO3 showed a different but closely related pattern. All data points aligned between a point only slightly below the Apollo solar wind composition and the 11.3 point seen in the plagioclase runs. Apparently, plagioclase was more or less completely dissolved by HNO3, which in later steps led to large contributions of cosmogenic Ne sited in the grain volumes, while HNO3 attacked only the surface layers of pyroxene grains containing implanted solar Ne.

In 1984 Jean-Paul Benkert and Anselmo Pedroni contacted us for doctoral student positions and in 1990 Christoph Murer also joined the group as graduate student. This trio allowed us to expand the work on solar noble gases and was among the first of many doctoral students I was privileged to supervise, either as principal or co-supervisor, or – in the beginning – less formally as scientific advisor. Jean-Paul Benkert continued my own work on lunar samples, focusing on CSSE studies on ilmenite and pyroxene separates from two lunar samples, one relatively recently exposed to the solar wind and the other perhaps several billion years in the past (Benkert et al., 1993). He used the gold line to etch ilmenite grain separates from various lunar soils with hydrofluoric acid. Figure 2.9 shows that he obtained very similar Ne data patterns to those I had found for the pyroxene separate. All but the last few etch steps fell on a straight line, starting almost exactly at the 20Ne22Ne ratio observed in the Apollo foils (a fact discussed below), and ending close to the 11.3 value known from plagioclase and pyroxene. The last few steps trend slightly toward cosmogenic Ne produced by galactic cosmic rays (GCR Ne).

Anselmo Pedroni and Christoph Murer studied samples from solar gas-rich meteorites with CSSE. Christoph analysed iron-nickel separates in Acfer111, Fayetteville, and Noblesville, Anselmo bulk samples and mineral separates from Kapoeta and Fayetteville. The latter data showed a basically similar pattern as the lunar samples (Pedroni, 1989). Christoph Murer built a separate CSSE line consisting of pyrex glass that allowed him to use copper chloride (CuCl2) as etchant, or more exactly as oxidising agent. This was based on previous work in Mainz by Else Vilczek and Heinrich Wänke who showed that copper-chloride selectively dissolves metallic Fe-Ni without affecting silicates. As expected, metallic Fe-Ni retains the light noble gases from the solar wind even better than lunar ilmenite, with He/Ar and Ne/Ar ratios being essentially identical to the solar wind values in the aluminum foils exposed on the Moon. He/Ar and Ne/Ar ratios were essentially constant throughout the entire CSSE runs, unlike what was observed in any of the lunar minerals studied (Murer et al., 1997).

All these studies convinced us that the “11.3-Ne” indeed represents a second solar Ne component, isotopically distinct from the known solar wind Ne, and essentially identical to David Black’s Ne-C, albeit with a nominally slightly higher 20Ne/22Ne ratio. It seemed clear to us that the straight lines with correlation coefficients of up to 0.999 defined by the lunar plagioclase data could not mimic isotopically fractionated solar wind Ne that had diffused into the grain interiors. In addition, the “11.3-Ne” also showed up in pyroxene, ilmenite, and Fe-Ni metal, which all have much lower diffusivities for Ne than plagioclase. However, the problem which had already puzzled David Black in 1972 remained: this second solar component, thought to represent higher energy particles than the solar wind, appeared to be orders of magnitude too abundant to be consistent with the known fluxes of what was believed to be emitted during solar flares. We therefore concluded that the 11.3-Ne must represent particles with higher energies than the solar wind but nowhere close to those detected during solar flares. We thus decided to call this component “SEP-Ne”, for solar energetic particles, a somewhat unfortunate choice, since we did not pay enough attention to the fact that the acronym SEP had already been adopted by the space physics community for truly high energy particles. Probably the term “suprathermal solar ions” as suggested by Johannes Geiss would have been a better choice for the 11.3-Ne and would have avoided confusion. Yet, the “SEP component” was widely adopted by the cosmochemistry community, and the CSSE data of Benkert et al. (1993) seemed to show that SEP-He and SEP-Ar also existed, enriched in the heavy isotopes relative to the solar wind composition, just like SEP-Ne. Further analyses of the isotopic composition of solar energetic particles by space physicists partly revealed 20Ne/22Ne ratios similar to that of the SEP-Ne component in lunar samples, i.e. higher than the initial space mission value of ~7.7 (Mewaldt et al., 1979), which seemed to provide further support for the existence of that component, albeit it was later found that the isotopic compositions of the SEP of space physicists is highly variable, depending, e.g., on charge state distributions (Leske et al., 2007). Nevertheless, by the late 1990s I was almost completely convinced of the reality of our SEP component in lunar (and meteoritic) samples, and I think many colleagues shared this view. “Almost completely” refers to the lingering problem of the large fraction of SEP gases in the total solar noble gas inventory, up to 50 % in some cases. This was not easy to digest, even considering that the very surficial solar wind component might very well have been strongly depleted relative to the more deeply sited SEPs due to diffusive losses or surface sputtering. We suggested that the data indicate a periodic increase of solar activity in the past relative to present values. The Genesis mission, selected by NASA in 1997 to collect solar wind ions for about two years between 2002-2004, therefore provided a unique opportunity to test this hypothesis, as discussed next.

2.5.2 The End of the SEP Myth

Here I only recount how Genesis helped put the SEP component to rest; our overall involvement in this mission will be addressed in more detail in Section 2.5.5. Don Burnett of Caltech in Pasadena – the mission’s Principal Investigator – proposed to fly a sample on Genesis that would allow very homogeneous etching with nitric acid over the entire analysed area and thus controlled depth profiling of the implanted solar noble gases. The selected sample was a Bulk Metallic Glass (BMG; Figure 2.10) which Veronika Heber in her doctoral thesis at ETH had shown to be etched very homogeneously. The BMG sample was mounted on the hinge holding the large Genesis sample panels, which is probably why it survived the crash landing of the Genesis sample return capsule essentially intact, the only sample that did so besides the concentrator target mentioned in Section 2.4 on nitrogen isotopes. In his doctoral thesis, Ansgar Grimberg measured He, Ne, and Ar in BMG using the CSSE technique. The Ne data published in Science (Grimberg et al., 2006; Fig. 2.10) showed 20Ne/22Ne ratios in the first steps above the solar wind value of 13.8, followed by a gradual decrease with increasing etching duration to a value very close to the “SEP-Ne” value of 11.3. This data pattern was in many ways strikingly similar to that previously observed in CSSE runs of lunar pyroxene and ilmenite separates. One obvious difference was that the data of the last steps did not deviate toward cosmogenic Ne with its high 21Ne abundance, which is essentially not present in Genesis targets after only about two years of exposure to galactic cosmic rays. Another difference was that the 20Ne/22Ne ratio in the first steps was higher than the solar wind value, which, as we will see in Section 2.5.4, is explained by the fact that lunar soil grains, but not the Genesis samples, reach a steady state between solar wind ion implantation and grain surface sputtering. Indeed, the gas release pattern shown in Figure 2.9 (left diagram) was almost perfectly reproduced by that predicted by the SRIM ion implantation programme for a single solar wind Ne component, and the total Ne released had an isotopic composition almost identical to that measured for solar wind Ne in other Genesis targets, with 20Ne/22Ne = 13.8. Ansgar also showed that the lunar data shown in Figure 2.9 (left diagram), including the straight line part between SEP and GCR, could be well reproduced by the single solar wind Ne component measured in the BMG onto which a simulated galactic cosmic ray produced (GCR) component was superposed. Furthermore, measurements carried out by the Advanced Composition Exporer (ACE) mission in parallel with Genesis made it clear that the latter had not collected large fluences of suprathermal ions above solar wind energy.

All this allowed us to conclude – in the first publication reporting scientific results from Genesis – that “no extra high energy component is required and that the solar neon isotopic composition of lunar samples can be explained as implantation fractionated solar wind”. Such was the end of the SEP component, the characterisation of which I had for years considered one of my major contributions to noble gas cosmochemistry. At least I could take comfort in the fact that not everyone succeeds in retracting one of their favourite ideas in a prestigious journal like Science. Grimberg et al. (2008) additionally showed that the He and Ar data also measured in the BMG are compatible with a single implantation fractionated solar wind component and do not require higherenergy particles. In Wieler et al. (2007) we discuss some further consequences of the elimination of “SEP” from the “noble gas alphabet” of cosmochemistry.

While the investigation of the SEP conundrum had been the primary motivation for developing CSSE, I also started using that technique to investigate primordial noble gases in meteorites. In parallel I turned to other meteorite studies, including their cosmic ray exposure histories, and I continued to work on lunar samples, with an emphasis on the isotopic and elemental composition of the heavy noble gases Kr and Xe. To recount all these activities – all of which, of course, were done in collaboration with numerous colleagues – I dial back the clock once again.

2.5.3 Some Personal Reminiscences from the Days After my Doctoral Exam

It was during the early months of the Covid 19 pandemic, more precisely on April Fools’ Day of 2020, that my partner Catherine Jakob and I celebrated the fortieth anniversary of our first beer together and our first kiss. In April 1980 I was about to submit my doctoral thesis and prepare its defense. My slightly stressful behaviour these days was a good test of our personal relationship. This was also the time when I submitted my only written scientific job application ever! It was for a postdoctoral position in the famous noble gas laboratory of John Reynolds in Berkeley. I would have succeeded Uli Ott, who was returning to Mainz. Probably my application was not convincing; in any case I did not get the job. And perhaps one reason for my unconvincing application was that I was not really keen to go to Berkeley, because Catherine was far too independent to easily give up – or even temporarily interrupt – her own professional life to follow me. So John Reynolds’ negative answer was at least as much of a relief to me as it was – nevertheless – a disappointment. So I was all the more pleased when Peter Signer offered me an opportunity to stay in his group as a postdoc. In addition to other projects, this allowed me to continue the work on meteorites and lunar samples that I just alluded to above. A few years later, Peter asked me if I would consider becoming a permanent member of his group. This offer came with a rather large “IF” attached: IF our Department and the ETH administration would agree. Of course I did not think twice about Peter’s offer and happily accepted. And ETH accepted too! Perhaps it was not as easy for Peter to push this promotion through as I imagined at the time. My contract stated that ETH would not guarantee my continued employment after Peter’s retirement, but that seemed so far away that I hardly worried. So, in the mid-1980s I became a permanent staff member in the Department of Earth Sciences at ETH Zürich until my formal retirement in 2014, and I am still a guest scientist there today. When Peter retired in 1994, perhaps the administration had simply forgotten what they had written into the contract some 10 years earlier. It seems worthwhile telling this story, as such a smooth career path does not appear possible anymore. Remarkably, no substantial paper-work on my part nor any formal evaluations were necessary. Further down, I will talk about the years during which I – together with Heiri Baur – took over the management of the laboratory after Peter Signer’s retirement.

2.5.4 Light and Heavy Solar Noble Gases and the FIP Effect

In Section 2.5.2 we noted that although the hypothesis of a high energy solar noble gas component in lunar samples and gas-rich meteorites did not with-stand the test of time, the work undertaken resulted in other important findings. Jean-Paul Benkert found that the first steps of mild etching (sometimes just a few minutes exposure to hydrofluoric acid vapour) always released Ne with an isotopic composition essentially identical to that of the Ne in the aluminum foils exposed on the Apollo missions (Benkert et al., 1993). We concluded that ilmenite grains at their top surface retained the true isotopic composition of noble gases in the solar wind. The value based on the initial etching steps of the ilmenite runs of (20Ne/22Ne)SW = 13.8 ± 0.1 has been adopted for many years by most workers, many primarily interested in noble gases from Earth’s interior. It is perhaps somewhat ironic that for about 15 years, Benkert et al. (1993) was the most cited paper with my name on the author list mostly thanks to just this one number. Fortunately it has withstood the test of time, as we will see in the next section. That this should be so is far from straightforward, as we noted in the previous section that isotopic fractionation upon ion implantation into Genesis targets leads to isotopically lighter compositions at grain surfaces (Grimberg et al., 2006, 2008). That the lunar samples do not show this effect has been explained by an equilibrium between ion implantation and grain surface removal by sputtering (Becker, 1998; Wieler, 1998; Vogel et al., 2011). This idea was later confirmed, when the solar wind isotopic compositions of not only Ne, but also Ar, Kr, and Xe that had been derived largely with CSSE data from lunar soils (Benkert et al., 1993; Wieler and Baur, 1994; Pepin et al., 1995), were found to be essentially in agreement with those derived from Genesis (next section).

In the 1990s, I focused my own solar wind studies mainly on the two heaviest noble gases Kr and Xe. Besides the solar wind isotopic composition of Kr and Xe obtained with CSSE (Wieler and Baur 1994), another long standing problem attracted my attention: to what extent can the elemental abundances of the heavy noble gases in the solar wind be derived from lunar (and meteorite) samples? Since about 1980, elements that are relatively easily ionised were known to be several times more abundant in the solar wind and in solar energetic particles relative to their known or assumed abundances in the solar photosphere (e.g., Reames, 2018). These are the so called low FIP elements, with a First Ionisation Potential less than ~10-11 eV. Figure 2.11a shows a “FIP diagram” for solar energetic particles measured by space missions, but the FIP effect is also observed in the solar wind, as discussed below. The FIP effect is understood to arise upon separation of ions from neutrals when the particles expand from the solar chromosphere up into the corona (Reames, 2018). Krypton and xenon are of particular interest in understanding the FIP effect, as their first ionisation potentials are only slightly above the nominal low FIP threshold (Fig. 2.11b). Their abundances in the solar photosphere cannot be measured spectroscopically, but can still be derived with reasonable confidence (e.g., Lodders, 2020). Therefore, the question was: can reliable abundance ratios of (Ar)/Kr/Xe in the solar wind be derived from lunar samples? And if so, are they similar or different from inferred solar values? First measurements had shown that the Xe/Kr ratio was several times higher than the value then assumed for the Sun (Eberhardt et al., 1972; Bogard et al., 1973). Because the N/Kr and N/Xe ratios were several times higher than the solar value – and N was widely considered at the time to be of solar wind origin – the prevailing view was that both Kr and Xe are not well retained in lunar samples, and that losses of Kr exceed those of Xe. Our online etching experiments on lunar mineral separates seemed to contradict this pessimistic view. The ratio Kr/Xe and essentially also Ar/Kr remained largely constant over several etch runs, unlike He/Ar and Ne/Ar which strongly increased with depth, as expected for diffusive loss of the two lightest noble gases. We therefore concluded that the relative abundances of the heavy noble gases in the solar wind are faithfully mirrored in lunar samples. Plotted on a FIP diagram, and anchored to Ar, this showed that Xe in the solar wind behaves almost like a low FIP element, although its first ionisation potential of 12.1 eV is above the low FIP threshold of ~10 eV observed for other elements (Wieler and Baur, 1995; Fig. 2.11b). Geiss and Bochsler (1985) had previously suggested that the First Ionisation Time (FIT) may be an important parameter governing FIP related fractionation. The FIT is a measure of how fast a sizeable fraction of the atoms ionise in the chromosphere, ranging between less than a second for Mg and up to several minutes for He (Geiss and Bochsler, 1985). FIT correlates well with FIP but is atypically short for Xe. Our data in Figure 2.11b therefore agreed well with Geiss and Bochsler’s FIT hypothesis. Moreover, the FIP fractionation of Xe (especially relative to that of Kr) seemed to have been larger in high antiquity samples, suggesting a secular change in solar wind composition.

These findings required confirmation. The opportunity arose when Charles Hohenberg in St. Louis invited me to visit his laboratory in 1995 (Fig. 2.12). I proposed to analyse the elemental abundances of Ar, Kr, and Xe in individual grains of various mineral types from lunar regolith samples with different antiquities. So, during two months of St. Louis summer heat I was a guest on the famous “fourth floor” at the McDonnell Center for the Space Sciences at Washington University. Charles had been on sabbatical in Peter Signer’s lab a couple of years before I started my doctoral work there. The two mass spectrometers in St. Louis reminded me in many ways of our two “Minneapolis machines”, the instruments Peter Signer had brought to Zürich from Al Nier’s’’ laboratory (see above). The scientific life on the fourth floor was very inspiring, reflecting the spirit of Bob Walker, with Ernst Zinner, Tom Bernatowicz, Ghislaine Crozaz, Sachiko Amari, Frank Podosek, Larry Nittler and many more, of course not forgetting the many dogs that regularly accompanied their human friends to work. We worked hard, as many grains needed to be analysed. Not the least motivating for long working hours was the fact that it was simply too hot outside. Charles showed up in the lab almost every day, and few days went by that he did not modify at least one command in the machine control software that he had written himself. With a lot of help from Karl (Charly) Kehm and Alex Meshik, we produced a nice data set. Alex had just arrived from Russia and we lived next door to each other. I remember our first joint shopping trip. For Alex, a US style grocery store was even more overwhelming than it had been for me when I first visited the US. I couldn’t convince him to buy just a small box with a few chicken wings for both of us, it had to be the largest size. So we ate chicken wings for four or five dinners in a row, and our initial supply of orange juice also lasted quite a while. Alex, his wife Olga Pravdivtseva (who arrived later) and their daughter Xenia (guess why her parents chose that first name for her!) are still in St. Louis and have long since adapted to the American lifestyle.

The data obtained in St. Louis were published in Nature (Wieler et al., 1996; Fig. 2.13). Within each sample, all grains showed essentially constant Ar/Kr/Xe ratios, regardless of mineral type and widely varying gas concentrations. This would not have been expected if the gas abundances had been altered by loss processes. And we also confirmed the difference in Xe/Kr of roughly a factor of two between high and low antiquity samples that had been observed in the in vacuo etch analyses. Hence, we concluded that the lunar regolith does indeed faithfully reflect abundances of the heavy noble gases implanted by the solar wind. This meant that the differences in Xe/Kr between samples that had captured their share of solar wind relatively recently, i.e. perhaps somewhere during the last 100 Ma or so, and samples with solar wind antiquities of billions of years, were not simply a result of element fractionation processes in the regolith after trapping that affected old and young samples differently.

Further confirmation that lunar regolith samples – or at least the clean mineral grains therein – correctly reflect the relative Kr and Xe abundances in the solar wind came with Genesis. As discussed in more detail in the next section, Xe/Kr ratios measured in various Genesis samples (Vogel et al., 2011; Meshik et al., 2014) are only marginally lower than the values in the low antiquity lunar samples, both when measured by in vacuo etching (CSSE) and single grain analyses. Vogel et al. (2011) and Wieler (2016) argued that this strengthened the conclusion for a secular decrease of Xe/Kr in the solar wind over the past several billion years. As I noted in my 2016 paper, this would – in my view – be the only remaining secular change in solar composition that had been proposed by John Kerridge in 1980. However, no clear explanation has been proposed as to how such a change could have happened, although I speculated in 2016 that it might be related to a hypothetical secular decrease in solar wind flux. Over the past few years, I have had many discussions with Peter Bochsler on this topic. If anything, Peter would rather expect Xe/Kr to have been lower in the early solar wind if ionisation in the chromosphere was dominated by EUV (extreme ultra-violet) photons. We discuss this in two LPSC contributions (Wieler and Bochsler, 2020; 2022). Does the lunar regolith perhaps tell us about processes unrelated to a secular change in solar wind composition?

We now explore the hypothesis that the high antiquity samples may contain Xe from the Earth’s atmosphere carried to the Moon. The idea of an “Earth Wind” was mentioned earlier in Section 2.4., as Geiss and Bochsler (1982) and later Ozima et al. (2005) had suggested that the unexpectedly high nitrogen abundance in lunar soils may be due to a contribution from the terrestrial atmosphere. In the case of Xe, the idea has been fostered on the one hand by recent observations by the Nancy group led by Bernard Marty that Xe in the ancient terrestrial atmosphere has become isotopically heavier until about two billion years ago, suggesting a substantial loss of atmospheric Xe (e.g., Avice et al., 2018). On the other hand, Kevin Zahnle and co-workers (Zahnle et al., 2019) proposed that Xe could be efficiently ionised in a hydrogen-rich early atmosphere by resonant charge exchange with protons, a process that would not work for any other noble gas with their higher first ionisation potentials. If some of these Xe ions reached the lunar surface, they could have been implanted into grain surface layers together with solar wind Xe. Wieler and Bochsler (2022) argue that this might explain the higher Xe/Kr ratios in high antiquity samples, but we also point out that the process must be modelled more realistically. For the time being, Earth Wind Xe on the Moon remains a hypothesis. As for many other open questions, samples from the far side of the Moon would be highly desirable for further tests.

2.5.5 The Genesis Mission

The Genesis mission collected solar wind ions in space at Lagrange point L1 (about 1.5 million km sunwards of Earth) for about two years and four months in various materials that were returned to Earth to be analysed by a large number of laboratories. As such, it was a follow up of the Solar Wind Composition Experiment using aluminum foils exposed on the Moon by the Apollo astronauts (Section 2.2), though with a much broader scope. NASA selected Genesis in 1997 as the fifth of its low cost “Discovery” missions, with Donald Burnett from Caltech being the Principal Investigator. Perhaps Don (Fig. 2.10) should more aptly be called The Soul of Genesis, a role for which he received in 2012 the Leonard Medal, the Meteoritical Society’s highest award. Like all space missions, Genesis was a huge team effort, but I allow myself to highlight two persons who certainly were among Don’s most dedicated allies: Amy Jurewicz and Roger Wiens. For our group, Genesis provided much fresh motivation to extend our studies of noble gases (and other elements) in the solar wind, parts of which have been mentioned above. This work resulted in the doctoral thesis of Ansgar Grimberg and a multitude of projects by Veronika Heber and Nadia Vogel. Beyond this, Genesis has led to a large number of collaborations and much scientific exchange with colleagues around the world, far too many to mention them all here. The “Genesis Science Team Meetings” that Don organised on the Sundays before LPSC, and which still take place remotely since 2021, have brought together many old and new friends of mine for more than two decades and continue to be highlights in my personal agenda. The meetings confirm one of the main predictions in Don Burnett’s mission proposal. As a sample return mission, Genesis has the advantage of enabling measurements using the best laboratory-based analysis techniques available for much of the 21st century (e.g., Burnett and Genesis Science Team, 2011).

Veronika Heber was a particularly important person involved in Genesis in our team and later when working with Kevin McKeegan at UCLA. She started her work at ETH in 1998 as the third doctoral student for whom I acted as main supervisor after having succeeded Peter Signer as head of the noble gas cosmoand geochemistry group (we will meet the first two of “my” graduate students, Henner Busemann and Jörg Schäfer, later). Before Veronika became involved in the preparatory studies to select the Bulk Metallic Glass sample to be flown on Genesis and eventually studied by Ansgar Grimberg (Section 2.5.2), she investigated another long standing problem with lunar samples. The question was whether the lunar regolith provided evidence for a (modest) secular increase in the 3He/4He ratio in the solar wind. I mentioned in Section 2.3 the perhaps most important result of the Apollo Solar Wind Composition experiment: definitive evidence that the D/H ratio in Earth’s oceans is about an order of magnitude higher than the protosolar value before deuterium had been burned to 3He in the very young Sun. Geiss and Reeves (1972) explained this in terms of equilibrium reactions between hydrogen and hydrogen compounds in the solar nebula and suggested that D/H values in the solar system could vary widely. There remained one open question, however: does the 3He abundance in the outer convective zone of the Sun, the source region of the solar wind, correctly represent the protosolar (D + 3He)/H ratio? It seemed possible that some 3He, freshly synthesised in the solar interior as transient product in the reaction chain converting H into 4He, was brought into the outer convective zone. Early lunar sample analyses had suggested this (Geiss, 1973). However, Veronika was able to disprove this with a series of very high resolution stepwise etch analyses of both low and high antiquity samples (Heber et al., 2003). Therefore, the 3He/4He ratio of the present day solar wind does not need to be corrected for a putative addition of freshly synthesised 3He. Nevertheless, caution is still required, because the He isotopic fractionation between the Sun and the solar wind is still not perfectly known (see also later in this section), and also because different values are used for the protosolar 3He/4He ratio, the other important parameter for the calculation of protosolar D/H. Most workers adopt the value in Jupiter’s atmosphere (Mahaffy et al., 1998), which is also close to the value in the present day Local Interstellar Medium measured directly with foils exposed on the MIR space station (Busemann et al., 2006; Section 3.3.4). Asplund et al. (2021) discuss these uncertainties in more detail.

Veronika Heber’s paper of 2003 in the Astrophysical Journal was the last major study of our group on solar noble gases based on analyses of lunar (or meteoritic) samples, as we now concentrated on the upcoming Genesis mission. Veronika and I became early members of the Genesis science team.

The history of Genesis has been told many times, including the crash landing in 2004 because a sensor which should have triggered the parachutes when the sample return capsule entered the atmosphere was installed upside down (e.g., Burnett and Genesis Science Team, 2011). In Zürich on September 8, 2004, we had gathered in our coffee room to watch NASA’s Live TV show from the landing site in the Utah desert, which was to culminate in helicopters catching the sample return capsule in mid-air. With us was a team from Swiss TV and I was quite well prepared to tell the public how happy we all were that we could finally soon start doing interesting science on matter returned from the Sun. Since at least the older generations in my country would still remember the days of the famous Apollo Solar Wind Composition experiment some 30 years ago, I also hoped to tell them that Genesis would allow a much wider range of measurements by a much larger community. But instead of helicopters and open parachutes, the NASA camera showed only a blurry image of something resembling a tumbling discus. And then nothing! Also the NASA commentator was unable to explain what he was seeing, but it was crystal clear that something had gone terribly wrong. I was shocked. We all were. My first worry was how to explain to Ansgar Grimberg that we now needed to activate plan B for his doctoral thesis, although I did not really have a plan B worthy of the name. Fortunately, as we saw above, we eventually did not have to rely on a contingency plan. My second problem that evening was what to say to the TV people and the public. I decided to hide my true feelings and play the optimist. I explained that, although we did not know any details at this moment, the precious samples were back on Earth, and even if they – most likely – were shattered to pieces, they might still allow for some analyses to be made. I added that although solar wind ions were implanted to only very shallow depths, they were not residing at the immediate surfaces of sample collectors, so it might just be a matter of selecting intact pieces and doing a thorough cleaning to be able to detect solar wind atoms. Although at the time I did not really believe what I was saying, it turned out to be very close to the truth, and I later learned that Don Burnett had made very similar statements to both the media and NASA officials. I am sure that, unlike me, Don was convinced about what he was saying already immediately after the crash. It took us a while to learn that the Bulk Metallic Glass – the target most important to us at the time – had remained essentially intact except for many scratches. Judy Allton and other members of NASA’s Genesis sample curation team provided us with a part of the BMG sample so that Ansgar could produce a fine doctoral thesis and the first publications reporting Genesis data (Grimberg et al., 2006; 2008), as explained in the Section 2.5.2. Figure 2.14 shows Genesis sample collectors pre-flight and broken upon the crash.

Our further work on Genesis was mainly carried out by Veronika Heber and Nadia Vogel (Fig. 2.15). Veronika’s dissertation has already been mentioned, and Nadia’s doctoral work in my group on primordial noble gases in meteorites is presented in Section 3.2.3. After postdoctoral positions at the Open University in Milton Keynes (Veronika), UC Berkeley (Nadia), and the University of Bern (Veronika and Nadia), they both returned to ETH. First, we focused on determining the elemental abundances and isotopic compositions of He, Ne, and Ar and the elemental abundances of Kr and Xe. The latter two are much rarer in the solar wind than the lighter gases He, Ne, and Ar, and therefore more difficult to measure. Our group was one of four teams analysing noble gases in Genesis targets. The others were led by Bob Pepin in Minneapolis, Alex Meshik in St. Louis, and Sarah Crowther and Jamie Gilmour in Manchester. Because of the controlled exposure conditions, Genesis data are less prone to unrecognised systematic errors than, for example, lunar sample analyses. In Zürich we used two different types of collector materials, DOS (Diamond-like carbon On a Silicon substrate, Fig. 2.15a) for He, Ne, and Ar and mainly silicon wafers for Kr and Xe (Heber et al., 2009; 2012; Vogel et al., 2011; 2019). The gases were extracted by UV laser ablation. DOS has the particular advantage that essentially no correction for losses of implanted solar wind ions by back scattering is required due to the low atomic mass of carbon.

A major goal was to establish if the solar wind noble gas compositions determined over many years on lunar samples could be confirmed. They were confirmed. For example, Veronika Heber’s 20Ne/22Ne ratio of 13.78 ± 0.03 (Heber et al., 2009) agrees perfectly with Jean-Paul Benkert’s lunar value of 13.8 ± 0.1 (Benkert et al., 1993). The backscatter-loss-corrected value of ~13.8 ± 0.2 determined with the Genesis Bulk Metallic Glass sample also agrees very well with the DOS and lunar values, although this was not a primary goal of the BMG analyses (Grimberg et al., 2008). As explained in the previous section, the excellent agreement of the isotopic composition of solar wind noble gases determined on lunar samples with the new Genesis values was far from trivial. The consistency confirmed the hypothesis (Becker, 1998, Wieler, 1998) that the outermost layers of lunar regolith mineral grains reflect the true isotopic composition of implanted solar wind. In the paragraph following the next one, this observation will again be important for the heavy gases Kr and Xe. Heber et al. (2009) also provided accurate isotopic compositions of solar wind He, Ne, and Ar as well as the elemental ratios He/Ne and Ne/Ar, which are now widely adopted by the scientific community.

Just as important as obtaining accurate values for the composition of the noble gases in the solar wind is obtaining reliable estimates of the isotopic and elemental fractionation between the Sun and the solar wind. For cosmochemists, the solar wind is often just a “proxy” for the composition of the Sun (or the solar nebula), which is their primary interest. For solar physicists, fractionation effects between the solar wind and its source region are of interest in understanding solar wind formation. Genesis therefore not only collected bulk solar wind during the entire 2.3 year collection period but also solar wind of three different “regimes”, with the respective sample panels exposed according to an algorithm based on various parameters measured onboard (Neugebauer et al., 2003; see Fig. 2.14). These were “Fast” or “coronal hole” solar wind, “Slow” or “interstream” solar wind and “CME”, i.e. solar wind from so called Coronal Mass Ejections (cf.Reisenfeld et al., 2013). These three regimes only allowed considerably less detailed studies of fractionation effects than data obtained by space-based instruments, but the Genesis regimes provided much higher accuracies, for elemental as well as isotopic abundances. Veronika Heber analysed isotopic fractionation of He, Ne, and Ar among the three Genesis regimes by carefully bracketing analyses of one regime with those of another regime (Heber et al., 2012). The isotopic fractionation strongly decreases from He to Ne to Ar in all regimes, from about 6 %/amu for He to 0.25 %/amu for Ar. Roland Bodmer and Peter Bochsler in Bern had investigated the role of Coulomb collisions in the acceleration of minor species in the solar wind (Bodmer and Bochsler, 2000). Their predicted isotopic fractionation effects agreed well with the observed isotopic fractionation factors of He-Ar between the Fast and Slow Genesis regimes as measured by Heber et al. (2012). However, extrapolating the measured He isotope data to the solar H/He ratio as proposed by Gloeckler and Geiss (2000) yields a substantially different solar He isotopic composition than the value obtained according to the Inefficient Coulomb Drag hypothesis (ICD), a problem discussed by Asplund et al. (2021).

The observation that the ICD hypothesis well predicts noble gas isotope fractionations between Fast and Slow solar wind regimes also became important for correcting the measured oxygen isotopic composition in Genesis targets for fractionation between Sun and solar wind (McKeegan et al., 2011). Because of its paramount importance to cosmochemistry, determining the isotopic composition of oxygen in the solar wind was the most important goal of Genesis. The team led by Kevin McKeegan at UCLA brilliantly achieved this by combining the front part of a secondary ion mass spectrometer (SIMS) with an accelerator mass spectrometer, called the MegaSIMS instrument. When the measured solar wind data are corrected by the ICD formalism, the resulting solar oxygen isotope composition in a classical oxygen three isotope diagram falls almost exactly on the straight line defined by data from calcium-aluminum-rich inclusions (CAI), the first solar system condensates (McKeegan et al., 2011). This is in line with a prediction explaining oxygen isotope compositions of CAIs to result from fractionation processes induced by isotope selective self-shielding during ultra-violet photolysis of CO in the solar nebula (Clayton, 2002). Veronika Heber and Nadia Vogel also provided noble gas data that were essential for correcting the oxygen isotope data for fractionation induced by the trapping process on board Genesis. Oxygen (as well as nitrogen, see Section 2.4) was measured in a circular “concentrator target” mounted at the centre of a focusing ion telescope in order to enhance the ion fluence by several ten times (Wiens et al., 2003; Fig. 2.14). This led to an isotope fractionation dependent on the position of the analysed spots, which was corrected by Ne data measured in Zürich along a traverse of one of the SiC quadrants of the concentrator target (Heber et al., 2011; Fig. 2.16). The corrected data define the isotopic composition of oxygen in the solar wind with high accuracy.

Nadia Vogel investigated the elemental abundances of Ar, Kr, and Xe, first in bulk solar wind samples from Genesis (Vogel et al., 2011). The very low concentrations of Kr and especially Xe required particularly careful analyses. As noted in Section 2.5.4, the Xe/Kr ratio measured in Si wafers sampling Genesis bulk solar wind was only marginally lower than the ratio determined in lunar samples of low antiquity. Also Ar/Kr in Genesis samples agreed with lunar values. This is an important confirmation that the lunar regolith correctly records the relative abundances of Ar, Kr, and Xe in the solar wind, implying that Xe is enhanced in the solar wind relative to Kr and Ar, presumably reflecting the lower first ionisation potential of Xe or its relatively short first ionisation time, as discussed in Section 2.5.4. That section also addressed the implications of this with respect to the observed secular decrease of Xe/Kr in lunar samples of different solar wind antiquities. After the bulk solar wind targets, Nadia analysed Ar, Kr, and Xe elemental abundances in regime targets, work that was eventually published after a “minor” delay of about eight years (Vogel et al., 2019; Fig. 2.17). The enrichment of Xe relative to Ar and Kr and solar abundances is slightly more pronounced in the Slow solar wind (~12 %) than in the Fast solar wind, in accord with the idea that the first ionisation time of the heavy noble gases in the solar chromosphere at least partly governs their fractionation in the solar wind.

Whereas we had focused on determining the elemental abundances of the very rare gases Kr and Xe in Genesis targets, Sarah Crowther and Jamie Gilmour in Manchester and – in a particularly heroic effort – Alex Meshik and his team in St. Louis, were able to measure the isotopic compositions of these two elements as well (Crowther and Gilmour, 2013; Meshik et al., 2014; 2020). I was very pleased to find that the Genesis data essentially confirmed the compositions deduced some 20 years earlier from lunar regolith data, to a large extent those we had obtained with closed system stepped etching analyses (Section 2.5.4.). As for the light noble gases He, Ne, and Ar, the agreement between lunar and Genesis data was not a trivial result.

Around the time of my official retirement in 2014 we performed a last set of analyses related to solar wind noble gases, but this time neither on lunar samples nor on Genesis targets. Together with our colleagues Peter Bochsler and Fritz Bühler from the University of Bern we reanalysed some of the aluminum foils exposed to the solar wind by the Apollo astronauts (Fig. 2.3). Fritz and Peter had been involved in the original measurements in Bern during the Apollo days (Geiss et al., 1972; 2004). Forty years later, Nadia Vogel measured several pieces of the Apollo 15 foil in Zürich (Vogel et al., 2015), after the samples had first been checked for dust contamination by Addi Bischoff and M. Horstmann in Münster. The major conclusion of this study was that mean isotopic and elemental compositions of He, Ne, and Ar in the solar wind have not significantly changed between the Apollo and Genesis mission periods. This conclusion was less trivial as it may sound, since short term variations in solar wind composition are well known from spacecraft data and, taken alone, also the different Apollo foil data could not rule out a minor temporal variability.

Genesis, however, has stuck with me to this day. In 2019, Don Burnett invited me as visiting scholar to Caltech in Pasadena. Don and I had received permission from Veronika Heber to finally publish another important data set of Genesis that she had produced between 2008 and 2013 while working at UCLA with Kevin McKeegan’s group. With crucial contributions from Amy Jurewicz from Arizona State University and Yunbin Guan and Don Burnett at Caltech – among many others – Veronika had measured the abundances of several major elements in Genesis bulk solar wind and regime targets using SIMS at UCLA and Caltech. But after returning to Switzerland to the Paul Scherrer Institute, she had not found time to publish the results comprehensively. The abundances of major elements in Genesis targets are very important to understand element fractionation in the solar wind, both for low and high FIP elements, again because Genesis allows a much higher precision than spacecraft data. When I arrived in Pasadena on February 1st 2020, Covid 19 had already reached Europe (although it did not lock down the continent until around mid-March) but seemed very far from the US West Coast. Based on a first draft by Veronika, Don and I prepared a manuscript and we made good progress, until, almost overnight it seemed, Covid 19 also arrived at the West Coast and I had to return to Switzerland prematurely by mid-March. Barely two weeks earlier, Don and I had attended a hockey game between the LA Kings and the Toronto Maple Leafs, amidst perhaps 10,000 other hockey aficionados. On the day I left, the driver who took me to LAX airport said he had never made that trip so quickly. No traffic jams, obviously out of fear of the virus.

My early return led to a delay, but the paper was finally published in the Astrophysical Journal in early 2021 (Heber et al., 2021). The FIP diagrams in Figure 2.18 display the abundances of the elements measured by SIMS, including the noble gas data from Vogel et al. (2019). The ordinate shows the fluence ratios of the elements with respect to that of Mg, normalised to the respective abundance ratios in the solar photosphere. The well known enhancement of low FIP elements is clearly visible in both the bulk solar wind and regimes, although the differences between regimes are remarkably small. As expected from solar physics and observations by space missions (e.g., von Steiger et al., 2000), the Fast regime is always less fractionated than the Slow and CME regimes. The differences in the Genesis regimes are remarkably small, however, which is new and potentially important and could only be recognised because of the high precision of Genesis data. In some cases, the data presented by Heber et al. (2021) approach the accuracy required to test the extent to which elemental abundances of CI chondrites approach solar abundances, although further analytical progress by the Genesis Science team is required to definitively answer some still lingering questions such as whether the data points in the left part of the diagram form a plateau, i.e. whether the low FIP elements (K – Fe) are unfractionated among themselves. The data in Figure 2.18 are consistent with this hypothesis within their uncertainties, but they are also consistent with an increase of fractionation with decreasing FIP.

Work on Genesis samples is ongoing, not only on this topic, as was again made clear at the 2021 and 2022 virtual Genesis science meetings. There is no question that Genesis was and is a success story, although it seems that some people only remember the images of the crashed landing capsule in Utah, and hence, erroneously, consider the mission a failure. This is one reason I felt it was important to write this section. Also for me personally, my long journey with solar noble gases, which started in 1976, is, I hope, not yet over. For example, I am following with great interest the work on new lunar samples brought back by the Chinese Chang’e 5 mission. Additionally, “Revealing the record of the ancient Sun and our astronomical environment” is one of the seven science objectives of NASA’s Artemis programme, with the first crewed missions after Apollo planned to reach the surface of the Moon, hopefully around 2025.