Quantification of ion-induced molecular fragmentation of isolated 2-deoxy-D-ribose molecules

Abstract

Recent experiments on low energy ion-induced damage to DNA building blocks indicate that ion induced DNA damage is dominated by deoxyribose disintegration (Phys. Rev. Lett., 2005, 95, 153201). We have studied interactions of keV H⁺ and He^q+ with isolated deoxyribose molecules by means of high resolution time-of-flight spectrometry. Extensive statistical fragmentation of the molecules is observed. The fragment distribution is found to follow a power law dependence. The exponent can be used to characterize and quantify the molecular damage.

---

1. Introduction

Ionization and fragmentation of DNA and its constituents is a primary step in biological radiation damage. When a charged particle crosses the cell, secondary particles such as low energy electrons, radicals and (singly/multiply charged) ions are formed within the track. The interaction of these secondary particles with biologically relevant molecules is responsible for a large fraction of the induced biological damage in the cell.

In this context, ions can be of importance as both, primary and secondary particles. Molecules within a cell can be subject to core ionization by a primary particle. Subsequent Auger-cascades lead to formation of multiply charged ions. These ions in turn can interact e.g. with DNA constituents.

As primary particles, for instance fast protons are successfully used in the treatment of small tumors localized near critical structures (e.g. uveal melanoma)² whereas fast C^q+ ions are used in therapy of deeply seated tumors.³ Furthermore, it is known from environmental studies that inhaled radon and its decay products emit energetic alpha particles (He²⁺) which can, in turn, have large damage potential.

The damage inflicted upon DNA can be classified as either single strand breaks (SSBs) or double strand breaks (DSBs). In living organisms, SSBs are usually repaired without further consequences but DSBs can cause cell death or mutations. Furthermore, multiple close ionization events along the particles’ track can form complex clusters of lesions which are even more difficult to repair.⁴ Lacombe et al. have shown that keV ions can induce SSBs and DSBs in plasmid DNA.⁵

Our studies on the interaction of keV ions with isolated molecules, mainly nucleobases in the gas phase, have revealed that fragmentation patterns and fragment formation are strongly influenced by the projectile.^6–10 We studied the dependence of ionization and fragmentation of various molecules on atomic number Z, charge state q and velocity v of the projectiles. These studies left open questions regarding the strong dependence of the fragmentation process on the properties of the target molecule.

For the isolated nucleobases uracil and thymine we have shown that their response to the interaction can be divided into three regimes: (i) non-dissociative ionization (mainly due to electron capture), (ii) multi-fragmentation and (iii) two-body break up.¹⁰ The fragments originating from the collisions can have energies higher than 10 eV. Very recently, Huels and coworkers found that even ions with kinetic energies of this range can produce damage to fundamental building blocks of DNA in films.¹¹ Their results indicate that usual models of biological radiation damage might underestimate the damage at the end of the Bragg peak.¹

Besides the studies in condensed phase, only the response of gas phase deoxyribose upon dissociative electron attachment has yet been studied.¹² It should be noted that gas phase studies allow a detailed exploration of the molecular mechanisms involved, but by nature neglect any effects of the chemical environment such as modified ionization energies or dissipation of the excitation energy.

In this paper, we study the fragmentation of gas phase deoxyribose after irradiation with different projectile ions. We will show that the molecular response is dominated by statistical fragmentation and that fragment distribution can be described by a power law with exponent τ. This exponent can be used to quantify the damage.

2. Experimental

The ions used in this study were extracted from our 14 GHz electron cyclotron resonance (ECR) ion source. The source was floated on potentials ranging from 2 to 20 kV for He⁺ and H⁺ and from 2 to 10 kV for He²⁺. The ion kinetic energies ranged from 2 to 20 keV for the singly and doubly charged ions. Mass-over-charge ratios were selected by deflection of the ion beam in a 110° magnet. The ions were deflected from the central beam line into the setup by a 45° magnet. The base pressure in the beam transport sections was about 1 × 10⁻⁸ mbar.

Fig. 1 displays a schematic of the experimental setup. The ion beam was pulsed by keeping one chopper plate at 90 V and switching the other one between 0 and 200 V, thus periodically deflecting the beam over a diaphragm (just before the focusing lenses). This way, ion pulses of less than 10 ns duration were generated. Ion beam pulses with a length of a few ns only are a requirement for good mass resolution. The chopped beam was collimated by two diaphragms 205 mm apart, and focused into the extraction region by means of an electrostatic lens system. In the collision chamber the ion beam pulses intersected the gaseous deoxyribose target.

Fig. 1 Sketch of our experimental setup.

2-Deoxy-D-ribose (C₅H₁₀O₄ purity ≥99%) by FLUKA was purchased from Sigma-Aldrich and employed without further purification. The powder was evaporated from a stainless steel oven with a nozzle of 1 mm diameter kept at 95 °C. The oven was placed at a distance of ca. 20 mm away from the collision center. The pressure inside the setup during the experiments stayed below 5 × 10⁻⁷ mbar and the base pressure was always below 2 × 10⁻⁸ mbar. In order to keep the collision chamber free from residual gas, a stainless steel plate mounted close to the collision region and kept at liquid nitrogen temperature served as a cryo-trap.

After the interaction of the ion beam with the gaseous target, the charged fragments were extracted from the collision region by means of a static electric field (600 V cm⁻¹). The field was provided by opposite voltages on two stainless steel discs of 50 mm diameter, mounted 10 mm apart. Positively charged collision products were extracted through a 5 mm diaphragm into a reflectron time-of-flight (TOF) spectrometer with resolution m/Δm ∼ 1500 at 720 amu.¹³ The ions were detected by a Microchannel plate (MCP) detector. Our data acquisition system was used in single ion detection mode as well as in coincidence mode. In the latter, two charged fragments stemming from one collision event are detected.

The temperature for sublimation of the deoxyribose (dR) from the oven was chosen such that the signal-to-noise ratio was optimal. This is particularly important for a precise quantification of the weaker peaks in the spectrum, such as the parent molecular ion (C₅H₁₀O⁺₄). The temperature, therefore, was chosen just above the melting point of deoxyribose (∼85 °C), but below the threshold for polymerization into long molecular chains. No clusters or fragments bigger than the parent molecule were observed.

Three TOF spectra of dR fragments after bombardment with 7 keV He²⁺ ions at different oven temperatures T = 70, 80 and 95 °C are shown in Fig. 2. The spectra are normalized to the total integral. Obviously, the overall fragmentation pattern does not change in this range of temperatures. The major differences are (i) the intensity increase of the peak around 13 μs assigned to a mass of 18 Da (H₂O⁺) and (ii) the intensity decrease of the peak at 22 μs assigned to a mass of 44 Da (possibly partly CO⁺₂).

Fig. 2 TOF spectra of dR fragments after collisions with 7 keV He²⁺ for different T.

The peak labelled O²⁺ can help to understand the difference. A doubly charged atomic fragment ion, such as O²⁺, is usually not produced in He²⁺ collisions with large molecules. It is thus not very likely that O²⁺ stems from dR. Probably, this fragment stems from collisions with residual gas, namely H₂O or CO₂. It is known that after interactions of He²⁺ with neutral water the fragment products are O⁺, OH⁺ and O²⁺. At 7 keV projectile energy the ratio O²⁺/H₂O⁺ resulting from these collisions is around 10%.¹⁴ In Fig. 2, the ratio O²⁺/H₂O⁺ is only 5% for all oven temperatures. This implies that more than half of the H₂O⁺ observed comes from dR itself and the rest corresponds to residual gas, probably originating from contaminations in the dR powder which is highly hygroscopic.

Residual CO⁺₂ follows the opposite trend than the H₂O⁺: it decreases with T. It is relatively strong at low T (panel (a) in Fig. 2) where the vapor pressure of the deoxyribose is low. At higher T, the dR vapor pressure is higher so the CO⁺₂ relative intensity decreases and the fragment coming from dR at this mass (C₂H₄O⁺) starts to be important. This can be understood by looking again to the change of the O²⁺ from (a) to (b), there the H₂O⁺ is more or less constant. The O²⁺ formed from the CO⁺₂ decreases as the residual gas gets less important.

Note that a similar increase of the H₂O⁺ peak with temperature was observed by Ptasińska et al.¹² after 70 eV electron impact and higher dR temperatures up to 110 °C. There, the formation of H₂O⁺ was attributed to deoxyribose fragmentation only.

In Fig. 2, there is also a slight increase in the peak at 17 μs, assigned to mass 29 Da (CHO⁺). As the residual gas is less important, the fragments originating from the collisions with the deoxyribose itself start to be more important. Then, at 95 °C the CHO⁺ peak becomes the strongest fragment after the H₂O⁺. The intensity change in the other peaks is negligible.

To conclude, an oven temperature of 95 °C ensures sufficient signal-to-noise ratio without noticeable thermal modifications. Fragment ions with very low relative intensity can thus be clearly distinguished from the background.

We can not, however, rule out a negligible contribution of thermal fragmentation of deoxyribose. Since ion-induced fragmentation turns out to be a rather violent process, this contribution can be neglected.

3. Results and discussion

In Fig. 3 we compare mass spectra for 5 keV amu⁻¹ He⁺ induced ionization/fragmentation of the nucleobase adenine C₅H₅N₅¹⁵ (a) and deoxyribose C₅H₁₀O₄ (b). Striking differences are seen. The most remarkable difference is the intensity of the surviving parent molecules. For adenine we observe a very strong parent molecule peak. For deoxyribose, on the other hand, the parent peak is negligible (for some projectiles’ energies even zero). The spectrum shown in panel (a) is typical for nucleobases; similar characteristics were observed for uracil⁶ and for thymine⁷ after irradiation with carbon ions in different charge states (for other projectile ions see ref. 10).

Fig. 3 Mass spectra of adenine (a) and deoxyribose (b) fragments after collisions with 5 keV amu⁻¹ He⁺ ions. As an inset in each graph the structure of the correspondent molecule is shown.

The higher stability of adenine as compared to dR manifests itself in the appearance energies (AE) of different fragments. For example, for group 8 from Fig. 3 in the case of adenine only 3.34 eV are needed in addition to the ionization energy (8.20 eV) to form the fragment at m/z = 108.¹⁶ In the case of deoxyribose only 0.3 eV of excess energy are needed in order to form the fragment at m/z = 116.¹² Some fragment cations are even formed at energies below the dR ionization potential!

In general, the nucleobase spectra are very structured with clearly defined groups of peaks. These groups—numbered from 1 to 8 in Fig. 3a—are due to fragments with different numbers of “heavy” atoms (carbon and nitrogen for adenine) and a variable number of hydrogen atoms. It is important to note that not all groups are present. For example, in adenine fragmentation, a group containing 9 heavy atoms (the parent molecule missing only one nitrogen or one carbon) was not observed. Furthermore, the fragmentation patterns for adenine and other nucleobases follow a bimodal distribution with decreasing intensities down to a minimum (around group 7 for adenine) and increasing again up to the parent molecule.

For targets like nucleobases and C₆₀, the formation of stable parent molecular cations is usually assigned to resonant electron capture into the unoccupied states of the projectile and therefore depends on the projectile electronic structure. The resonant electron capture processes are considered comparably gentle, i.e. not accompanied by transfer of excitation energy. In collisions with keV ions, resonant electron capture mainly involves electrons from the highest occupied molecular orbitals of the target molecule. In contrast, the formation of small fragments (with low masses) is associated with more violent close collisions involving mainly direct ionization accompanied by electronic and vibrational excitation. These collisions can involve virtually any target electron. The transferred energy mainly stems from kinetic energy of the projectile ion.^9,17

For the nucleobases the surviving molecular ions represent an appreciable fraction of the total yield so that relative fragmentation cross sections can be obtained and relative yields for the different groups of fragments can be evaluated.⁶ The relative fragmentation yield for a molecule after impact with different projectiles can be assumed to be inversely proportional to the electron capture cross section. Recently, Bacchus-Montabonel et al. showed that semiclassical calculations based on ab initio potential energy curves can explain the experimentally observed trends in resonant charge transfer between uracil and C^q+ ions.¹⁸

In contrast to the nucleobase case, dR shows a very small parent ion peak. No groups are missing and the overall peak intensities monotonically decrease with m/z. These characteristics hold for all the projectiles as shown in Fig. 4.

Fig. 4 Mass spectra of deoxyribose fragments after collisions with 5 keV amu⁻¹ projectiles.

Mass spectra of the dR fragmentation, obtained from collisions with H⁺, He⁺ and He²⁺ of equal velocity are shown in Fig. 4. The spectra resemble the mass spectra from the NIST database¹⁹ and from ref. 12 for 70 eV electron impact. The ratio between the large fragments (m ≥ 65) and small fragments (m ≤ 62) changes with projectile, being three times smaller for He⁺ than for H⁺ and He²⁺.

The deoxyribose ionization potential is 10.51 eV.¹² In Fig. 5, we compare the vertical IP of dR with the electronic structure of the projectiles. Obviously, for He⁺ resonant capture from the dR HOMO is energetically ruled out, inhibiting the most gentle electron removal mechanism. This explains the more severe fragmentation for He⁺ than for H⁺ and He²⁺ where resonant capture is possible: Only in the case of He⁺, in electron capture processes the potential energy of the projectile has to be (partly) transferred to the dR molecule. For He⁺, we thus only look at events of high energy deposition: direct ionization or electron capture accompanied by potential energy transfer of the order of ∼15 eV. For He²⁺ and H⁺, on the other hand, resonant capture into excited states leave most of the projectile potential energy on the projectile ion.

Fig. 5 Relevant energy levels of the different projectiles in comparison with the vertical ionization energy of deoxyribose. The dashed rectangle contains the levels where electron resonant capture is possible.

In general, we can point out some noticeable differences between the spectra in Fig. 4 as follows:

• The peaks of C⁺_n with n = 1, 2 and 3 and of C_nH⁺_m with m = 1, 2 and 3 for H⁺ (a) are very low compared to the same peaks in He⁺ (b) and He²⁺ (c). This was observed for all investigated ion velocities. This implies that on average the energy deposited to the dR molecules is less for protons than for He⁺ and He²⁺ ions. This might be attributed to the lower electronic stopping of H⁺.²⁰

• In the shaded area of Fig. 4, ranging from m/z = 46 Da to m/z = 55 Da, the intensity of the peaks changes with projectile, being highest for He⁺.

• For He²⁺ the fragments m/z = 65 Da (not observed for films¹ and very weak in electron studies^12,19) and m/z = 104 Da (C₄H₈O⁺₃) are present for most energies but not for the one shown in Fig. 4. They are absent in (a) and (b) and they were not observed for these projectiles at any energy.

• For H⁺ (a) and He²⁺ (c) the fragments at m/z = 81 Da (C₅H₅O⁺) and 117 Da (C₅H₉O⁺₃) are observed. For He⁺ (b) these fragments are negligible.

• The group of peaks at m/z = 97–99 (dashed circled) is present for H⁺ projectile (a) and He²⁺ projectile (c) but absent or very weak for He⁺ (b). This is valid for the whole energy range under study.

The change in peak intensity with projectile (stronger for He⁺ than for H⁺ and He²⁺) implies that certain fragments, i.e. the group 46–55 Da, originate from scission of rather strong bonds. This process is only possible after high energy deposition and thus more prominent for He⁺ projectiles. Meanwhile, fragments formed in more gentle collisions (for example the ones negligible for He⁺ (b)) are not stable after high energy depositions in the molecule.

As in the condensed phase results of Deng et al.,¹ H₃O⁺ (m/z = 19 Da) is observed in our spectra as well. This fragment was not observed in our studies on nucleobases. According to Deng et al., in the condensed phase this fragment is a clear evidence of abstraction of two adjacent hydrogen atoms from the dR molecule by a highly reactive OH⁺ radical. It is thus the signature of an indirect effect. The presence of the H₃O⁺ peak in our spectra shows that in the gas phase it is possible to generate this fragment directly from the dR without the assistance of reactive radicals. The ratio between the relative intensities of the peaks H₃O⁺ and H₂O⁺ gives a clear indication that it is not an isotopic effect (see Table 1).

Table 1 Ratios of the relative intensities of the fragments at m/z = 18, 19 and 20 for different projectiles

%	H^+	He^+	He^2+
Ratio 19/18	2.8	8.4	2.5
Ratio 20/18	0.2	0.2	0.2

---

Note that in Fig. 4 there is also one peak at m/z = 20. This peak can be unambiguously assigned to H₂¹⁸O⁺ because its ratio corresponds to the natural abundance of isotopic ¹⁸O (0.20%). In Table 1 the ratio 20/18 is given for the different projectiles.

Even though Miaskiewicz and Osman could show that for certain C-sites, H abstraction from dR requires about 4 eV only,⁴ H⁺ loss (at m/z 133 Da) was not observed in our study for any projectile. It has been observed for nucleobases, for example see Fig. 3a and for thymine after electron impact.²²

For deoxyribose the yield of surviving parent molecules is usually less than 0.01% and the fragmentation yields are thus always very close to 100%. We hence have to find another parameter (equivalent to the relative fragmentation cross sections) to quantify the damage inflicted upon the deoxyribose by the different projectiles.

A more detailed analysis of the fragmentation patterns discussed in Fig. 4 can be done by evaluation of the relative yield for each m/z (see Fig. 6). Every point represents the relative yield of the respective fragment with the corresponding statistical error. The line shown is an apparent fit (not weighted) to the data. Deviations from the fit are not statistical but an inherent property of the molecular fragmentation. They reflect fragment stabilities, ionization energies, bond strengths within the molecule, etc. The deviations will be discussed in detail later in this section.

Fig. 6 Integral points of a deoxyribose spectrum with their corresponding statistical error. The apparent linear fit of the points (dashed line) gives the trend of the fragmentation and a value for the characteristic exponent τ = 2.0.

Numerous studies in a variety of fields have already shown that a fragment mass distribution following a power law n(M) ∼ M^τ (n: number of fragments, M: fragment mass, τ: characteristic exponent) is a signature of statistical fragmentation processes. Examples are the fragmentation of nuclei,²³ size distributions of asteroids,²⁴ fragmentation of solid and liquid matter²⁵ or sputtering of clusters.²⁶ The power law-like behavior is often linked to critical phenomena, for instance at the liquid–gas phase transition. However, other fragmentation processes, such as weathering, can exhibit a power-law behavior as well.²⁷ Almost independent of the nature of the system, exponents of the order of τ = 2 are found.

During the last decade, statistical fragmentation has also been observed for atomic clusters, for instance in H⁺_n collisions with atoms²⁸ or in ion–fullerene collisions.²⁹ For these systems, a general requirement for statistical fragmentation is a rather hot intermediate system, i.e. the excitation energy has to clearly exceed the binding energy of the system. Within the clusters, the energy is quickly distributed over all available electronic and vibrational degrees of freedom. This can lead to evaporation of fragments (a sequential process governed by rate constants) or to multifragmentation, if the excitation energy is sufficient. Multifragmentation is an entropy-driven process during which the system tries to reach a distribution of products that occupies the maximum volume in accessible phase-space. The fragment distribution is thus dominated by smaller fragments. This is exactly what we observe in the fragmentation of deoxyribose (Fig. 6).

The power law fit in Fig. 6 for dR fragmentation yields an exponent τ = 2.0 for this projectile at this energy.

Fragments with yields exceeding the power law fit could be due to high stability or low IP. In their electron impact ionization experiment, Ptasińska et al. have observed that indeed all the fragments marked by a framed label in Fig. 6 have low appearance energies between 10 and 14 eV.¹² These energies are given in Table 2. The fragments below the fit, therefore probably correspond to high appearance energies or are very unstable.

Table 2 Appearance energy of cations produced by electron impact of neutral deoxyribose¹²

Cation	Mass/Da	AE value/eV
CH^+3	15	13.31 ± 0.50
CHO^+	29	12.18 ± 0.49
CH3O^+	31	12.36 ± 0.14
C3H^+3	39	13.01 ± 0.42
C2H4O^+	44	12.50 ± 0.02
C3H5O^+	57	11.63 ± 0.06
C2H4O^+2	60	11.28 ± 0.16
C4H6O^+	70	10.67 ± 0.03
C4H9O^+	73	10.74 ± 0.03
C5H5O^+	81	11.72 ± 0.50
C4H6O^+2	86	11.41 ± 0.02
C5H6O^+2	98	10.12 ± 0.15
		12.36 ± 0.15
C4H7O^+3	103	11.25 ± 0.08
C5H8O^+3	116	10.82 ± 0.05
C5H10O^+4	134	10.51 ± 0.11

---

Note that in order to obtain the parameter τ, we neglected the contribution of H₂O⁺ and its fragments because of the unknown contribution due to contaminating H₂O.

By fitting a power law to all the measured fragmentation spectra, the characteristic exponent τ can be obtained for H⁺, He⁺ and He²⁺ colliding with deoxyribose over the whole velocity range under study. The results are displayed in Fig. 7. Obviously, lowest exponents that increase linearly with v are observed for H⁺. For He²⁺ projectiles, higher values are found and the exponents increase linearly with v, as well. For He⁺ projectiles, on the other hand, largest characteristic exponents are observed which exhibit a linear decrease with v. The explanation for this behavior has already been sketched in the discussion of Fig. 4 (and in ref. 6 for collisions of C^q+ with uracil): whereas H⁺ and He²⁺ can resonantly capture an electron from the deoxyribose HOMOs, this is energetically ruled out for He⁺ projectiles. In the latter case, the measured fragment ions almost exclusively originate from collisional ionization processes at small impact parameters. The violent nature of the ionization process is reflected in the mass spectrum by a fragmentation spectrum characterized by a high τ.

Fig. 7 Value of the parameter τ for deoxyribose spectra after collisions with different projectiles at different kinetic energies.

In most ion–molecule collision systems, if energetically possible, resonant electron transfer from the highest molecular orbitals is usually the channel with the largest cross section (H⁺, He²⁺). For He⁺, where resonant capture from the HOMOs is ruled out, different scenarios are possible. Non resonant capture from HOMOs would lead to excess energy sufficient for release of an additional electron from the dR. Electron capture from lower lying states is also possible, leaving the molecule in a strongly excited state. Both scenarios could explain the observation of much stronger fragmentation for He⁺ and its high characteristic exponent. The increase in fragmentation when going from H⁺ to He²⁺ can be explained by the fact that a doubly charged projectile usually induces more double ionization than a singly charged one. Double ionization in turn will lead to much more severe fragmentation than single ionization.

The velocity dependence of the characteristic exponents can be easily explained within the same framework. In case resonant electron capture is possible (H⁺, He²⁺) the projectiles can still induce collisional excitation of the electronic system of the target molecule. This process is usually called electronic stopping of the projectile. For ion–fullerene collisions and v = 0.1–1 au, it has been shown that electronic stopping depends strongly on the impact parameter scales and increases with v.^17,30 If ion–deoxyribose collisions follow a similar trend, the energy deposited into the target molecule by a given projectile increases with v and therefore the characteristic exponent increases as well. In case resonant capture is ruled out (He⁺) only small impact parameters are involved. At small impact parameter electronic stopping is highest because the projectile traverses a high electron density.

However, with increasing v non-resonant processes become possible. This facilitates electron capture into excited projectile states.^31–33 Gentle electron capture at larger impact parameters thus becomes increasingly important with increasing v. This is reflected in a characteristic exponent decreasing with v (see Fig. 7).

Itoh et al.³⁴ showed that for collisions of MeV Si^q+ ions with C₆₀ the characteristic exponent can be used to quantify the exact amount of electronic stopping deposited into the target at a given projectile velocity. For ion–fullerene collisions τ appears to follow a simple exponential dependence on the deposited energy. Even though the exact relation between characteristic exponent and deposited energy is unknown for ion–deoxyribose collisions, τ can still be used for a qualitative analysis of the fragmentation process. For a quantitative interpretation dependence of the characteristic exponent energy deposition theoretical studies on this collision system would be needed.

4. Conclusions

We have studied the ionization and fragmentation of gas-phase deoxyribose molecules by keV H⁺, He⁺ and He²⁺ impact over a wide range of velocities. Virtually complete disintegration of the molecule is invariably observed. The fragment distribution is found to follow a power law—a fingerprint of statistical fragmentation. Deviations from the power law trend can be explained by the different appearance energies of the fragments, in line with the results of Ptasińska et al.¹² The characteristic exponent of the power law fit is obtained for each collision system as a function of projectile velocity. For H⁺ and He²⁺, low exponents are observed which linearly increase with projectile v. For He⁺, the characteristic exponent is larger and decreases with v. These findings can be explained in the framework of charge exchange and electronic stopping. The characteristic exponent is found to be a suitable parameter for quantitative studies of dR fragmentation. Most probably, it is also applicable to other fragile molecules.

Acknowledgements

This project is part of the research program of the Stichting voor Fundamenteel Onderzoek der Materie (FOM) which is financially supported by the Nederlandse Organisatie voor Wetenschapelijk Onderzoek (NWO). We acknowledge support from the EC within the COST P9 action “Radiation Damage in Biomolecular Systems”. TS acknowledges support by the Royal Netherlands Academy of Arts and Sciences (KNAW).

References

1. Z. W. Deng, I. Bald, E. Illenberger and M. A. Huels, Phys. Rev. Lett., 2005, 95, 153201.
2. A. Wambersie, in Atomic and molecular data for radiotherapy and radiation research, IAEA, 1995, ch. 1, pp. 7–45.
3. G. Kraft, M. Scholz and U. Bechthold, Radiat. Environ. Biophys., 1999, 38, 229.
4. B. M. Sutherland, P. V. Bennett, O. Sidorkina and J. Laval, Biochemistry, 2000, 39, 8026–8031.
5. S. Lacombe, C. L. Sech and V. A. Esaulov, Phys. Med. Biol., 2004, 49, N65–N73.
6. J. de Vries, R. Hoekstra, R. Morgenstern and T. Schlathölter, J. Phys. B: At. Mol. Opt. Phys., 2002, 35, 4373–4381.
7. J. de Vries, R. Hoekstra, R. Morgenstern and T. Schlathölter, Eur. Phys. J. D, 2003, 24, 161–164.
8. J. de Vries, R. Hoekstra, R. Morgenstern and T. Schlathölter, Phys. Rev. Lett., 2003, 91, 053401.
9. T. Schlathölter, R. Hoekstra and R. Morgenstern, Int. J. Mass Spectrom. Ion Processes, 2004, 233, 173–179.
10. T. Schlathölter, F. Alvarado and R. Hoekstra, Nucl. Instrum. Methods Phys. Res., Sect. B, 2005, 233, 62–69.
11. Z. W. Deng, M. Imhoff and M. A. Huels, J. Chem. Phys., 2005, 123, 144509.
12. S. Ptasińska, S. Denifl, P. Scheier and T. D. Märk, J. Chem. Phys., 2004, 120, 8505–8511.
13. O. Hadjar, R. Hoekstra, R. Morgenstern and T. Schlathölter, Phys. Rev. A, 2001, 63, 033201.
14. F. Alvarado, R. Hoekstra and T. Schlathölter, J. Phys. B: At., Mol. Opt. Phys., 2005, 38, 4085–4094.
15. S. Bari, A. Lecointre, F. Alvarado, R. Hoekstra and T. Schlathölter, 2006, to be published.
16. H. W. Jochims, M. Schwell, H. Baumgärtel and S. Leach, Chem. Phys., 2005, 314, 263–282.
17. T. Schlathölter, O. Hadjar, R. Hoekstra and R. Morgenstern, Phys. Rev. Lett., 1999, 82, 73.
18. M. C. Bacchus-Montabonel, M. Łabuda, Y. S. Tergiman and J. E. Sienkiewicz, Phys. Rev. A, 2005, 72, 052706.
19. NIST chemistry web book http://webbook.nist.gov/chemistry/.
20. J. Opitz, H. Lebius, S. Tomita, B. A. Huber, P. Moretto-Capelle, D. Bordenave-Montesquieu, A. Bordenave-Montesquieu, A. Reinköster, U. Werner, H. O. Lutz, A. Niehaus, M. Benndorf, K. Haghighat, H. T. Schmidt and H. Cederquist, Phys. Rev. A, 2000, 62, 022705.
21. K. Miaskiewicz and R. Osman, J. Am. Chem. Soc., 1994, 116, 232–238.
22. S. Ptasińska, S. Denifl, V. Grill, T. D. Märk, E. Illenberger and P. Scheier, Phys. Rev. Lett., 2005, 95, 093201.
23. R. W. Minich, S. Agarwal, A. Bujak, J. Chuang, J. E. Finn, L. J. Gutay, A. S. Hirsch, N. T. Porile, R. P. Scharenberg, B. C. Stringfellow and F. Turkot, Phys. Lett. B, 1982, 118, 458–460.
24. J. S. Dohnanyi, J. Geophys. Res., 1969, 74, 2531.
25. J. A. Åström, B. L. Holian and J. Timonen, Phys. Rev. Lett., 2000, 84, 3061–3064.
26. T. Schlathölter, M. W. Newman, T. R. Niedermayr, G. A. Machicoane, J. W. McDonald, T. Schenkel, R. Hoekstra and A. V. Hamza, Eur. Phys. J. D, 2000, 12, 323–327.
27. J. Hüfner and D. Mukhopadhyay, Phys. Lett. B, 1986, 173, 373–376.
28. B. Farizon, M. Farizon and M. J. Gaillard, Int. J. Mass Spectrom., 1999, 192, 259–266.
29. H. Tsuchida, A. Itoh, K. Miyabe, Y. Bitoh and N. Imanishi, J. Phys. B: At. Mol. Opt. Phys., 1999, 32, 5289–5297.
30. T. Kunert and R. Schmidt, Phys. Rev. Lett., 2001, 86, 5258.
31. R. Hoekstra, D. Ciric, A. N. Zinoviev, Yu. S. Gordeev, F. J. de Heer and R. Morgenstern, Z. Phys. D: At. Mol. Clusters, 1988, 8, 57.
32. R. K. Janev and H. Winter, Phys. Rep., 1985, 117, 265–387.
33. W. Fritsch and C. D. Lin, Phys. Rep., 1991, 202, 1–97.
34. A. Itoh, H. Tsuchida, K. Miyabe, T. Majima and Y. Nakai, Phys. Rev. A, 2001, 64, 032702.

---

Footnote

† Presented at the Bunsen Discussion on Structure and Dynamics of Free Clusters and Nanoparticles using Short Wavelength Radiation, Bad Honnef, 7–9 Sept 2005.

---