Fundamental mechanisms of photon and electron-induced desorption

Get Complete Project Material File(s) Now! »

Dust and ice mantles

Dust grains in the ISM lock up a large part of the heavy elements available. They are mostly either silicate grains or carbonated grains, with varied forms of silicates (porous or crystalline, pyroxene and olivine…) and carbon (amorphous, diamondoid, aromatic or aliphatic…). Grains form mostly at the end of the life of a star, in the stellar winds created by the expansion and explosion, which create conditions where stellar material can be accreted in the form of a solid [18]. The carbon abundance in the star will regulate the composition of dust: if the C/O ratio is > 1 mostly carbonaceous grains will be created, while in C/O < 1 conditions carbon is locked up in the formation of CO and silicate grains form. In addition to Si or C, looking at table I.1 these grains should also contain significant amounts of metallic atoms Mg and Fe, along with other refractory material in trace amounts (Al, Na, etc). Only part of these atoms are observed in the gas phase, and the rest is therefore presumed to be locked up in the grains. Models of dust include this material e.g. as nano-inclusions in silicate grains [19]. Grains then evolve with time depending on the conditions they are exposed to: irradiation by UV photons and cosmic rays, or shocks, influence their nature and composition. Models of layered grains are often considered, with for example a core made of amorphous carbon or silicate and a « mantle » of aromatic-rich carbon [19].
The composition of dust can be observed by infrared absorption bands characteristic of either silicates or diﬀerent types of carbon. Also characteristic of dust is a continuous extinction curve going from the deep UV to the near IR, with a few famous features like the 217 nm bump and the diﬀuse interstellar bands (DIBs) in the visible/near-IR. The counterpart is an emission curve in the mid/far-IR. Models of dust attempt to reproduce correctly the features of these extinction and emission curves, constraining parameters such as the size distribution, shape and composition of the grains. In the ISM grains have sizes varying from about 1 nm to 1 µm, with a peak around 100 nm. The smaller grains are therefore basically macromolecules, with probably big PAHs (polycyclic aromatic hy-drocarbons) and fullerenes (C60, C70). Grains are also not particularly spherical: instead evidence from polarization of light by dust grains suggest more asymmetric shapes. In protoplanetary disks, dust grains can start to accrete and grow up to mm sizes. After reaching this size, further growth is diﬃcult and requires other and not yet completely un-derstood mechanisms to explain the formation of planetesimals [20]. See the introduction chapters of [21] for more details on grains.
In the cold regions that are of interest for us, dust grains are covered by an icy mantle of molecules that can go up to a few hundred monolayers in thickness. These molecules either accrete from the gas phase or are formed directly on the grain through surface processes, and stay there when the grain temperature is lower than their sublimation temperature. Only infrared spectroscopy in absorption allows to probe ices, as condensed molecules do not have rotational transitions. As mentioned before, observations are made using either background stars behind the cloud or embedded stellar objects as infrared sources against which absorption features of condensed molecules can be seen. The observations that have been made of icy mantles are reviewed in Boogert [8].
The main species that have been detected in ice mantles are summarized in the graph of fig. I.1a, while an example of dust and ice features in the infrared spectrum of a massive young stellar object is shown in fig. I.1b. Ice mantles are composed mainly of water, with also large amounts of CO2 and CO. Depending on the line of sights CH3OH can be a very abundant molecule or be completely absent. NH3 and CH4 are observed at the few percent range. A number of other molecules have been detected in some line of sights, but whether this is representative of a global composition of ice mantles is not clear yet. According to the classification of Boogert [8], H2CO, OCN− and OCS are « likely detected » while some other molecules likes HCOOH, SO2, NH+4 or CH3CHO are « possibly detected ». In addition, upper limits for a large number of molecules (including very large upper limits on the abundance of hardly detectable species like N2 and O2) have been obtained. Solid phase infrared bands are broad and the sources weak, so that it is diﬃcult to obtain an inventory of molecules as varied and complex as is obtained in the gas phase, although there is undoubtedly numerous other molecules residing in ice mantles, as a result of solid phase (photo)chemistry of the main components mentioned.
The details of the absorption bands also give more information on the exact molecular environment of the molecules. The bands of CO and CO2 in particular have been used to deduce the existence of diﬀerent components in the ice. The most common distinction that is made is between polar and apolar ices, with the polar environment being mostly a water matrix and the apolar one a CO environment. This led to a picture of ice mantles as being roughly separated into two phases, linked with the way the mantles build up. At an early stage for the cloud, when grains are still bare and atoms are still available in the gas phase, the accretion of atoms on grains will lead to the formation of molecules like H2O, CH4 and NH3 by hydrogenation, and CO2. At a later stage when temperature drops so that the abundant gas phase CO freezes out onto the grains, the CO-rich second layer forms, with CO2 also being present. Hydrogenation of CO leads to CH3OH, with H2CO as an intermediate. This picture is coherent with the molecular environment suggested by the shapes of the absorption bands, but also with observations made as a function of extinction (for example CH3OH appears when extinctions are very high, corresponding to the massive CO freeze-out).
How much is the composition of the ice mantles conserved through the processes of star and planet formation is an important question that is diﬃcult to answer. It seems that an important part of the water is conserved as is from the pre-stellar core to the protoplanetary disk [23], while some portion has been sublimated in the protostellar phase and recondensed afterwards. Objects of the outer stellar system may then inherit this ice.

Solar system ices

Many objects in the solar system host molecular ices while being « airless » bodies (with no atmosphere or a very tenuous one), i.e. conditions that are close to the ones we study in the laboratory and in the context of interstellar grains. In the inner solar system (before the asteroid belt), temperatures are usually high, but in some permanently shadowed regions, water ice can exist. Its presence has notably been shown in craters of the Moon [12]. In the outer solar system, there are many more objects of interests: the icy moons of the giant planets, the trans-neptunian objects, and comets.
The well-studied icy moons of Jupiter are Callisto, Europa and Ganymede. Their surfaces are covered with water ice. The surface temperatures are too high to condense more volatile species like CO, N2 or CH4 but other molecules such as SO2 and CO2 have been detected [12]. Around Saturn, Enceladus is the icy moon that has generated the most interest, and aside from the vast parts of the surface that are covered with nearly pure water ice, organics and CO2 have been detected in some places. The other moons, except Titan, are also airless icy bodies for the most part. More details can be found in Bennett et al. [12].
With the Neptunian satellite system, which is mostly its biggest moon Triton, we reach a point where more volatile molecules can be found in icy forms. Thus the surface of Triton is composed of N2, CH4, H2O, CO2, CO and C2H6, with N2 being a very dominant species. The very volatile N2, CO and CH4 are also detected on Pluto. On Charon, only water has been found, although evidence suggests the presence of processed organics at the poles originating from CH4 photolysis [24].
Comets are the most interesting objects to compare to interstellar grains in terms of ice composition, because they are often believed to have conserved a « primitive » composition, representative of what was found in the circumstellar disk of the our Sun prior to planet formation. Indeed, the composition of cometary ices are well studied and the detailed review of Mumma et al. [13] shows that it is quite similar to the kind of composition found for interstellar ices, with some specific diﬀerences [8].

Sources of irradiations

Diﬀerent sources of irradiation prevail in diﬀerent regions of the ISM, which we will review here. Diﬀuse clouds, for example, are permeated by the so-called InterStellar Radiation Field (ISRF). The ISRF is a « background » electromagnetic field resultant from the radiation of an average distribution of stars (at least in the IR to UV spectral range: dust emission dominates far IR contributions and the cosmological background radiation even further in the mm range) at a given point that is not too close from any particular star. The star part has two components, one due to low mass stars and peaking around 1 µm, and another more interesting for us due to massive stars peaking around 150 nm (8 eV), both corresponding roughly to black body radiation at diﬀerent temperatures. The spectrum cuts oﬀ at 13.6 eV, the ionization threshold of H, above which radiation is absorbed. A typical (calculated) ISRF spectrum, from Mathis et al. [27], is shown in fig. I.2 (in blue, the scale is not shown). The integrated UV photon flux from the ISRF (from say 5 to 13.6 eV) is of the order of 108 photons.s− 1.cm−2. The ISRF intensity G0 is often used as a unit to compare the UV radiation flux of diﬀerent environments.
The interior of dense clouds is shielded from the ISRF, but the edges are not. The radiation will be progressively attenuated, giving a layered structure to the edge of the cloud. Although the ISRF is not an extremely strong radiation source, and its eﬀects will quickly die out, there can also be stars near the cloud illuminating it with much higher flux. This is common since molecular clouds are precisely the birthplaces of stars. If the nearby star is a massive star, its UV radiation spectrum will be mostly black-body like and resemble that of the ISRF in terms of spectrum. If it is a lower mass star, the UV spectrum will rather be dominated by atomic lines, especially Ly-α (atomic hydrogen line at 121.6 nm/10.2 eV). The star can also be embedded in the cloud: the protostellar envelope of a young star is significantly irradiated. There is a class of regions called Photon-Dominated Regions or PhotoDissociation Regions (PDRs) to qualify, among others, the previously mentioned regions where the UV photon flux is high. They are called thus because their chemical and physical properties are highly controlled by the eﬀects of the UV irradiation, and because they can be (more or less generically) modelled by so-called PDR codes. Cosmic rays have high enough energy that they interact little with matter (see chapter II) and can penetrate dense clouds with little attenuation. However, they do interact enough with matter to be an important energy source inside clouds. Cosmic rays excite and ionize gas phase molecules (mostly H2), and this in turn creates secondary electrons and secondary UV photons that will subsequently have eﬀects. There is therefore still a UV photon flux inside dense clouds. Calculations of this flux give a number of ∼ 103-104 photons.s−1.cm−2 [28, 29], four orders of magnitude lower than the ISRF. This means that these secondary UV photons dominate the UV flux in a dense cloud when the ISRF is attenuated by a factor of 104, which occurs around Av = 5 (depending on the considered region). The secondary electrons created by cosmic rays are quickly thermalized by interaction with H2 (more will be said on the topic at the end of chapter V). The spectrum of the generated secondary UV photons is very structured since they mostly come from de-excitation of H2 or H: it consists of many spectral lines between 7 and 13.6 eV, with a strong Ly-α component. The spectrum calculated by Gredel et al. [25] is reproduced in fig. I.2 (in black).
In protoplanetary disks the sources of irradiation are complex. The disk is close to a young star, which abundantly produces UV and X-rays. The UV spectrum of the young star TW Hya, which hosts a protoplanetary disk that has been very well studied, is shown in fig. I.2 and was taken from [26]. It is dominated by atomic lines. The X-ray spectrum of TW Hya is displayed in fig. I.4, taken from ref. [30]. In the X-ray spectrum, there is a continuum that originates from a hot plasma (here with two components around 2 MK and 8 MK) emission, and atomic lines embedded in the continuum. Although this is not obvious on the log plot, the contribution of atomic lines is important.
The density of the disk is very high, and the irradiation from the star will not be able to penetrate all the way to the disk midplane. Instead in this midplane again cosmic rays should be the main source of energy, with similarly secondary UV photons (and electrons) playing an important role. There is therefore a layered structure which is schematized in fig. I.5. UV photons dominate in the outermost layer, but they are attenuated more quickly than X-rays that will play an important role in an intermediate layer, before being themselves attenuated. The exact structure will depend on various parameters such as the type of disk. Note that propagation through the gas density will change the UV and X-ray spectrum (for attenuation of the X-ray spectrum, see chapter VI). To further complicate the picture, it has been suggested that disks may be shielded from cosmic rays [31] by magnetic eﬀects. If this is the case, then the low flux of X-ray photons that reach the midplane, but also the external UV (ISRF) and X-ray fields (from nearby young stars [32]), which are less attenuated than those from the stars since they come from a diﬀerent angle and traverse less column density, could become the dominant sources of irradiation instead. Disks are therefore particularly complicated regions.
In the solar system, the solar wind and solar photons play an important role as sources of irradiation. A main-sequence star like our own has a much lower X-ray and deep UV flux than a young star like the ones discussed previously, and therefore the solar X-ray component can be mostly neglected. There is still a significant amount of UV photons, but the spectrum is largely dominated by infrared and visible light. A much more detailed discussion of the solar photon fluxes can be found in ref. [12]. Surfaces that are exposed directly to sunlight are usually heated to temperatures where thermal desorption is much more important than non-thermal processes (e.g. UV photodesorption). However, these surfaces are not always exposed to sunlight, and other surfaces are never exposed at all. Energetic ions and electrons from the solar wind or from external cosmic rays, as well as UV photons from the interplanetary medium (∼ 2 ×108 Ly-α photons.s−1.cm−2 [12]), will become important here.
For the icy moons of giant planets with magnetospheres like Jupiter and Saturn, en-ergetic particles are particularly important: the magnetosphere traps solar wind particles and also accelerates other particles originating from the atmospheres of the planets and moons themselves, and sends a shower of energetic particles on the moons. UV photons are not likely to play a role here. It should be noted that the energy of the particles of the solar wind are typically in the keV/amu range, much lower than cosmic rays.
In the outer solar system, for trans-neptunian objects, solar wind and photons are more and more attenuated and external galactic cosmic rays and UV photons can start playing a role. UV photons may irradiate surfaces hidden from the sun by a resonant scattering mechanism: Ly-α photons (from the sun or the interplanetary medium) are resonantly scattered by interplanetary H and illuminate objects like Pluto or Charon with a non-negligible flux (of the order of 107 photons.cm−2.s−1) [33].

READ Hard Multi-Task Metric Learning for Kernel Regression

Astrophysical context

Role of non-thermal desorption

PID and EID inscribe themselves in the general topic of grain surface processes, one aspect of which is the gas-grain exchanges, as the surface is the interface between these two phases. The question is, what processes regulate gas-grain exchanges, in particular the desorption side? The thermal desorption rate is exponential with temperature: we can basically say that there is a temperature above which thermal desorption will dominate, and below which it is entirely negligible. The intermediate range of temperature is small, although the fact that we are dealing with astronomical time scales (104 to 106 years) and densities means that it is diﬀerent from what a surface scientist may be used to in a vacuum chamber.
Wherever the surface temperature is small enough to have more than one monolayer of molecules condensed, thermal desorption is no longer at play and gas-solid exchanges are dominated by non-thermal desorption processes. In the ISM, these non-thermal processes either prevent some molecules from entirely freezing out of the gas phase, or liberate in the gas phase molecules that are formed on grains but should have stayed there. On icy bodies with no atmosphere in the solar system, they may create an « exosphere », a tenuous gas phase density of molecule in equilibrium above the surface, dominated for the larger bodies by gravity rather than gas phase collisions. They can also erode the icy layers and deplete the molecular content if the desorbed molecules are faster than the escape velocity.

Diﬀerent non-thermal desorption processes

There is a variety of non-thermal desorption processes that can been considered in as-trophysical environments. A non-thermal desorption process requires bringing energy in some form to the ice to break the physisorption bonds. One source of energy are energetic particles: photons, electrons and ions (cosmic rays, CRs), that all have in common that they excite or ionize the ice, bringing electronic energy (although ions can also transfer momentum by knock-on collisions). This is the class of process that we are interested in in this manuscript, and that will be presented in more detail in the next chapter.
Among these, in the ISM, UV photodesorption is certainly the one that is the most taken into account. Early on, the process was thought to be ineﬃcient [34], but it was realized later with new experimental results that it was in fact orders of magnitude more eﬃcient than reported at first [17]. For the last two decades, UV photodesorption has become an important factor in models to explain a number of observations, as will be detailed in the next section. The ubiquity of UV photons in interstellar media is certainly an important factor of this success, along with the existence of laboratory data to constrain the process. In comparison, processes like electronic desorption by electrons or cosmic rays, despite their presence in all regions as well, are often less considered. Sputtering of molecular ices by energetic ions has been experimentally studied extensively early on (see e.g. [35] and references therein), but these studies were more geared towards a planetary science context, and CR desorption is not always taken into account in models. This could be because calculations suggested that secondary UV photons generated by cosmic rays deposit ten times more energy in ices than cosmic rays directly do [29]. But experimental studies continue still to be made, with an interstellar medium context in mind, for example with heavy ions [36, 37], and they tend to show that in fact, secondary UV photon desorption and direct CR desorption have comparable eﬀects. To my knowledge, electron-induced desorption is not really taken into account. I will expand on this at the end of chapter V.
X-ray photodesorption is a process that is not taken into account in models yet either. The relevance of X-ray photodesorption is limited to fewer regions: protoplanetary disks, as mentioned earlier, and molecular clouds exposed to the X-ray irradiation of a nearby young star or an object emitting strong X-rays. Relevant experimental data on the process did not exist yet, and providing these is the object of chapter VI.

Table of contents :

I Context(s)
I.1 Astrophysical context
I.1.1 Molecules in the universe
I.1.1.1 The interstellar medium
I.1.1.2 Dust and ice mantles
I.1.1.3 Solar system ices
I.1.1.4 Sources of irradiations
I.1.2 Role of non-thermal desorption
I.1.2.1 Different non-thermal desorption processes
I.1.2.2 Astrochemical models
I.1.2.3 Observations
I.2 Vacuum technology context
I.2.1 Non-thermal desorption in vacuum dynamics
I.2.2 Non-thermal desorption in the context of accelerators and the LHC
I.2.2.1 Brief description of the LHC
I.2.2.2 The different sources of non-thermal desorption in accelerators
II Fundamental mechanisms of photon and electron-induced desorption
II.1 Position of the problem
II.2 Interaction of photons and electrons with molecular ices
II.2.1 Ices/molecular solids
II.2.2 Electronic transitions
II.2.3 From free molecules to molecular solids
II.2.4 Electron-matter interaction
II.2.4.1 Stopping power
II.2.4.2 Compounds
II.2.4.3 Penetration and energy deposition profiles
II.2.4.4 Low-energy (< 20 eV) electrons
II.2.5 X-ray photons: core excitations and Auger decay
II.2.5.1 EXAFS, Shape resonances
II.3 Historic models of photon- and electron-induced desorption
II.3.1 The MGR model
II.3.2 Non-MGR models
II.4 A case study: desorption mechanisms from rare-gas solids
II.4.1 Electronic excitations in RGS: excitons
II.4.2 Desorption mechanisms
II.5 Desorption from molecular ices
III Experimental methods
III.1 Techniques
III.1.1 Ultra-high vacuum
III.1.2 Mass spectrometry
III.1.2.1 Mass filtering
III.1.2.2 Ionization source
III.1.2.3 Detection
III.1.2.4 Residual gas analysis
III.1.2.5 Kinetic energy analysis with an electrostatic deflector
III.1.3 Ice growth
III.1.4 Temperature-programmed desorption
III.1.5 Infrared spectroscopy
III.1.6 Electron yield
III.1.7 Calibration of EID and PID
III.1.7.1 Absolute calibration methods
III.1.7.2 Relative calibration methods
III.1.7.3 Fragments and reflected light
III.2 Electron-induced desorption studies at CERN
III.2.1 The Multisystem set-up
III.2.2 Measurement procedure
III.3 SPICES II at LERMA
III.4 Synchrotron-based experiments
III.4.1 Synchrotron light and its advantages
III.4.2 Experiments on the DESIRS beamline
III.4.3 Experiments on the SEXTANTS beamline
III.5 Development of a UV laser desorption and spectroscopy set-up in the lab
III.5.1 UV and VUV laser desorption
III.5.1.1 VUV generation
III.5.1.2 Practical implementation
III.5.2 REMPI spectroscopy
III.5.3 Desorption + REMPI set-up
IV VUV photon-induced desorption
IV.1 Pure ice systems
IV.1.1 CO
IV.1.1.1 Recent studies on CO photodesorption
IV.1.1.2 Thickness and deposition temperature dependence of CO photodesorption
IV.1.1.3 Photodesorption mechanisms
IV.1.2 NO
IV.1.2.1 Synchrotron wavelength-resolved study
IV.1.2.2 NO gas phase REMPI
IV.1.2.3 NO desorption + REMPI
IV.1.3 CH4
IV.1.4 H2O
IV.1.4.1 Water ice structure and electronic spectrum
IV.1.4.2 Photodesorption yields and mechanisms in the literature
IV.1.4.3 Experimental results from synchrotron study
IV.1.5 NH3
IV.1.5.1 Photodesorption spectra and yields
IV.1.6 Other organic molecules
IV.1.6.1 Photodesorption from HCOOH ice
IV.1.6.2 Photodesorption from organic molecules
IV.1.7 Perspectives and limits for pure ices
IV.2 Indirect desorption: model layered ices
IV.2.1 CO-induced indirect desorption
IV.2.1.1 Single layers
IV.2.1.2 Multiple layers and other systems
IV.2.1.3 Discussion
IV.2.2 H2O-induced desorption
IV.2.2.1 Results on single and multilayers on H2O and D2O
IV.2.2.2 Discussion
IV.2.3 Other systems
IV.3 Implementation in astrochemical models spectral dependence
IV.4 Conclusions
V Electron-induced desorption
V.1 Chemically inactive pure ices: N2 and Ar
V.1.1 Interpretation of EID yield curves
V.1.2 N2
V.1.3 Ar
V.1.4 Ar mixed with impurities: effects of ice composition
V.2 CO, CO2, H2O and the role of chemistry
V.2.1 CO
V.2.2 CO2
V.2.3 H2O
V.3 Relevance of the data to astrophysical and accelerator contexts
V.4 Conclusions and perspectives on electron-induced desorption
VI X-ray photon-induced desorption
VI.1 H2O X-ray photodesorption
VI.1.1 Ice absorption spectroscopy and structure
VI.1.2 Desorption of neutral species, and astrophysical relevance of X-ray photodesorption
VI.1.3 Desorption of ions
VI.1.3.1 H− desorption
VI.1.3.2 Oxygen fragments desorption
VI.1.3.3 H+ desorption
VI.2 CO X-ray photodesorption
VI.2.1 Effects of the irradiation: TEY evolution
VI.2.2 Desorption of neutral species
VI.2.3 Desorption of ions
VI.2.3.1 Ice charging and ageing
VI.2.3.2 Mass spectrum of cations
VI.2.3.3 Spectral signatures
VI.2.4 Discussion on X-ray induced photochemistry
VI.2.4.1 CO irradiation chemistry
VI.2.4.2 Comparison of different probes of chemistry
VI.2.5 Astrophysical yields
VI.3 Conclusions on X-ray photodesorption
Conclusion