Available online at www.sciencedirect.com

www.elsevier.com/locate/asr

Advances in Space Research 44 (2009) 1019–1029

c-ray spectroscopy in Mars orbit during solar proton events

M.S. Skidmore *, R.M. Ambrosi

Space Research Centre, Department of Physics and Astronomy, University of Leicester, University Road, Leicester LE1 7RH, UK

Received 10 March 2009; received in revised form 4 June 2009; accepted 4 June 2009

Abstract

Understanding the evolution of Mars requires determining the composition of the surface and atmosphere of the planet. The Euro-pean Space Agency’s ExoMars rover mission, which is expected to launch in 2016, is part of the Aurora programme. The instruments onthe rover will search for evidence of life on Mars and will map a sub-section of the Martian surface, extracting compositional informa-tion. Currently our understanding of the bulk composition (and mineralogy) of Mars relies on orbital data from instruments on-boardsatellites such as 2001 Mars Odyssey, Mars Reconnaissance Orbiter and Mars Express, in addition to in-situ instrumentation on roverssuch as Spirit and Opportunity. c-ray spectroscopy can be used to determine the composition of Mars, but it has yet to be successfullycarried out in-situ on Mars. This study describes some of the results obtained from the c-ray spectrometer on 2001 Mars Odyssey duringsolar proton events and discusses whether the increased emissions are useful in c-ray spectroscopy. The study shows that althoughincreased c-ray emissions were expected from the Martian surface during a solar proton event, they were not detected from orbit prob-ably due to insufficient signal-to-background. However, this does not preclude the possibility of measuring changes in c-ray flux corre-sponding to changes in solar activity on the surface of the planet.� 2009 COSPAR. Published by Elsevier Ltd. All rights reserved.

Keywords: Solar proton events; c-ray spectroscopy; Mars; Composition; 2001 Mars Odyssey

1. Introduction

1.1. Planetary c-ray emission

Particles (mainly protons and electrons) are ejected fromthe Sun and contribute to the solar wind. Their energies aretypically in the range of eV to keV (Hundhausen, 1995), andthey have speeds of approximately 400 km s�1 when theyreach the Earth (Barnes, 1992). When the solar particlesdirectly interact with a planetary atmosphere or surface, acomplex series of interactions occur (including spallation,ionisation and absorption; these and other reactions aredetailed in Krane (1988)), producing secondary particlessuch as neutrons and high-energy c-rays (see Fig. 1). Simi-larly, galactic cosmic radiation (GCR) that is composed pri-marily of protons (87%) with energies ranging between 0.1and 10 GeV/nucleon may also interact with planetary atmo-

0273-1177/$36.00 � 2009 COSPAR. Published by Elsevier Ltd. All rights rese

doi:10.1016/j.asr.2009.06.004

* Corresponding author. Tel.: +44 116 252 3519.E-mail address: mss16@star.le.ac.uk (M.S. Skidmore).

spheres and surfaces, thus contributing to the secondary par-ticle fluence (Reedy and Arnold, 1972). The secondaryparticles created consist primarily of neutrons and c-rayswhich are attenuated by the planetary surface and atmo-sphere. Neutron interactions with matter can include inelas-tic scattering or neutron capture. Boynton et al. (2004) haveestimated that �10 neutrons are produced per primary par-ticle. The type of interaction is dependent on the neutronenergy, as discussed in Masarik and Reedy (1996) and Boyn-ton et al. (2004). The interaction of the neutron with thenucleus (via capture, inelastic scatter or non-elastic scatter)can leave the nucleus in an excited state; the de-excitationprocess often results in the emission of a particle or c-rayand in some cases the re-release of a neutron at a lowerenergy (Kaplan, 1963). c-ray energies can vary between200 keV and 10 MeV (Boynton et al., 2004), and the energyof the emitted c-ray is characteristic of the element it has beenemitted from. The high-energy protons can undergo similarscattering and capture reactions with planetary surfaces andatmospheres to produce c-rays. c-ray emission may also

rved.

Fig. 1. Diagram showing the possible interactions of protons with aplanetary surface.

1020 M.S. Skidmore, R.M. Ambrosi / Advances in Space Research 44 (2009) 1019–1029

occur as a result of the decay of radioactive nuclei, such as U,Th and K. A suitable c-ray spectrometer can be used todetect the c-rays on the surface of a planet or in orbit.

The Martian atmosphere, although thin, significantlyattenuates the proton flux below 100 MeV, so few protonsare able to interact with the surface (A Monte Carlo modelbeing developed at the University of Leicester (UoL) that sim-ulates the Martian atmosphere and surface has indicated thatnearly all of the sub- 100 MeV protons are attenuated in theatmosphere, several kilometers above the surface). However,the neutrons and secondary particles created in the atmo-sphere via these proton reactions are able to penetrate intothe surface. As a result normal solar proton emission fromthe Sun does not create a significant amount of c-ray emissionfrom the surface of Mars compared with the GCR interactionwith Mars. Unlike the solar protons, the GCR protons areable to directly interact with the sub-surface because their typ-ical energies are in excess of 100 MeV instead of keV. Typicalproton fluences of GCR and solar origin detected by the spaceenvironment monitors (SEM) on the geostationary opera-tional environmental satellites (GOES) at 1 AU are of theorder of 10 protons cm�2 sr�1 s�1 for 1–10 MeV protons,�5 protons cm�2 sr�1 s�1 for 10–100 MeV protons and�1 proton cm�2 sr�1 s�1 for >100 MeV protons (NGDC,2008). Since the total GCR flux can be assumed to be 1–2 cm�2 sr�1 s�1, with energies ranging from 0.1 to 10 GeV/nucleon (Reedy and Arnold, 1972), the protons >100 MeVare assumed to be of GCR origin. During a solar proton event(SPE), protons can be accelerated out from the Sun at energies>100 MeV; and the overall proton flux is enhanced. Theseprotons will be able to directly interact with the Martianand lunar surfaces and sub-surfaces, and could lead toincreased c-ray emission from the planet. This study discusseswhether this increased c-ray emission could be exploited inorder to obtain a deeper understanding of the interactionmechanisms and composition of a planetary surface.

1.2. c-ray spectroscopy

The composition of the surface and sub-surface of a pla-net can be extracted from the c-ray spectra collected, given

that the c-rays emitted by these mechanisms are elementspecific. In addition, the intensity of a c-ray line is propor-tional to the concentration of the associated element in thesurface, sub-surface or atmosphere of a planet.

c-ray spectroscopy (GRS) can be carried out from orbitif the planetary body of interest has a thin enough atmo-sphere. It has been carried out in the orbit of Mars (Mars5, Phobos 2 and 2001 Mars Odyssey, hereafter called MO)but has yet to be successfully applied in-situ (Surkov,1997). Detection of c-rays from orbit is dependent on thepresence and density of an atmosphere. The atmosphericattenuation varies with atmospheric thickness, and theatmosphere itself will contribute to the c-ray spectra(deconvolving an orbital c-ray spectrum into its atmo-spheric and surface contributions has yet to be carriedout). On the Martian surface or sub-surface the c-ray fluxcan be significantly higher than that in orbit about Mars asa result of the attenuation by the atmosphere (the averagecount rate from the surface of Mars detected by the c-rayspectrometer on MO is �190 s�1 (Evans et al., 2007)).

A research team led by R. Ambrosi at the UoL is currentlydeveloping instrumentation for geophysical applications.The instrumentation includes a geophysical package for afuture in-situ planetary science missions that can simulta-neously carry out GRS, c-ray densitometry and radiometricdating of a planetary surface. The geophysical package couldbe installed on a lander or (ideally) a sub-surface probe thatwould carry out measurements at depths of up to 5 m. Theadvantages, goals and challenges of a sub-surface GRS mis-sion are detailed in Skidmore et al. (2009). Current designsfor the c-ray portion of the detector feature a lanthanum bro-mide (LaBr3(Ce)) scintillator detector, which is compact,radiation tolerant, operates at a range of temperatures, islow in mass and has a low power requirement compared tothe cooling power requirements of a high-purity germanium(HPGe) detector (Skidmore et al., 2009).

Currently the geophysical package will rely on naturalc-ray emissions caused by the proton interaction with the sur-face. The addition of a neutron generator could increase thesurface c-ray emission further. An increase in the c-ray fluxwould reduce the accumulation time required to get a spec-trum with sufficient counts in it. This was discussed in detailby Ambrosi et al. (2005). Although neutron generators arebecoming more compact (e.g. a 252Cf source could weigh� grams (Ambrosi et al., 2005)), they require shielding to pro-tect the rest of the lander/probe, which increases the mass aswell as posing a radiation hazard. The additional proton fluxinteracting with a planetary atmosphere, surface and sub-sur-face during a SPE results in an increase in the neutron fluxwithin the surface, and an increase in the c-ray emission fromthe surface. This may positively affect the counting statistics.The mechanism that creates a SPE is discussed in Section 1.3.

1.3. Solar proton events

A SPE refers to an enhancement in solar proton emis-sions from the Sun, but generally speaking these events

Fig. 2. Relative amounts (%) of secondary particles in the MCNPXsimulated Martian surface. The plot indicates where the number ofsecondary particles reaches a maximum for protons of different energies.The energy stated in the brackets in the legend next to the particle typeindicates the energy of protons that were interacting with the simulatedsurface. For protons of >500 MeV, the number of secondary particlesappears to reach a maximum within the Martian regolith.

M.S. Skidmore, R.M. Ambrosi / Advances in Space Research 44 (2009) 1019–1029 1021

include heavy ions, which form a similar proportion of thetotal number of emitted particles as the heavy ions in theGCR particle distribution (Morthekai et al., 2007). Solarprotons dominate the spectrum and are accelerated to highenergies (>1 MeV) by shocks associated with coronal massejections (CMEs), magnetospheric and bow shocks, andco-rotating interaction regions (Reames, 1999a). Themechanism that generates the most intense events is theCME, which is the result of magnetic energy being releasedfrom reconnecting coronal fields. This process can causethe acceleration of protons to energies in excess of>100 MeV, even reaching energies of several GeV. Theseenergies are reached via collision-less shock waves thatare driven out from the solar corona. CMEs can occur(on average) 2.5 times a day at solar maximum; however,only about 1–2% of these actually form shocks (Reames,1999a). Proton acceleration occurs when protons are scat-tered back and forth across the shock many times by mag-netic turbulence upstream and downstream from the shock.Reames (1999a) provides a more detailed explanation ofthe acceleration mechanisms.

The protons accelerated by the shock can stream out-wards from it, and tend to follow the magnetic field linesthat make up the interplanetary magnetic field (IMF)(Cleghorn et al., 2004). Protons of 10 MeV can have speedsup to 1 AU/h (Reames, 1999b); higher energy protons tra-vel faster, therefore the onset of the CME can be detectedvery quickly via radiation monitors in orbit about theEarth. The shock wave itself also travels outwards fromthe Sun at speeds of 750–2500 km s�1 (Reames, 1999b).Protons within the shocks generate resonant waves that‘trap’ other protons in the shocks paths; these areaccelerated in the same way as the protons within theshocks. These shocks can spread over a wide angular range,reaching 180� width at 1 AU (Cane, 1988). The peak flux ofenergetic protons is usually observed when the shock wavepasses (Reames, 1999a), even for protons at energies of>500 MeV. If a planet is well connected to the CME viaone of the IMF lines, then an observer at 1 AU could seea sharp rise in particle in particle flux (associated withthe SPE) approximately 12 h after the onset of the event(Futaana et al., 2008). If the planet is not well connectedto the CME then the particle flux will increase more grad-ually, or not at all. The location of the CME emission pointon the Sun, the width of the shock wave and the speed ofthe shock all determine whether the SPE could affect theplanet or not.

The high-energy solar protons associated with a SPE areable to penetrate further into a planetary atmosphere orsurface before an interaction occurs. In the case of Mars,if the particles were able to reach the denser surface mate-rial of the planet, the probability of a nuclear reactionoccurring would increase, therefore more secondary parti-cles would be produced than in any interaction with theMartian atmosphere. This should result in an increase inc-ray emission from the planetary surface. The increasein c-ray emission would result in increased signal-to-back-

ground and will positively impact the counting statistics.Improved counting statistics would be of great benefit toa lander-based c-ray spectrometer. Often lander-basedinstruments have to share limited resources with otherinstruments thus reducing data acquisition times.

Studying the effects of a SPE on a planetary surface is ofgreat interest, not only for the purposes of planetary com-position studies, but also in understanding the radiationhazard astronauts could face on the surfaces of other plan-ets. The Earth’s atmosphere is sufficient to protect us fromthe energetic particles released in a SPE, but the Martianatmosphere is not, and there would be no protection atall on the lunar surface. Models of the Martian surfaceradiation environment have been developed in order tostudy the dose rate that astronauts would be exposed toduring quiet solar conditions and during SPEs. Morthekaiet al. (2007) modelled the dose rate with depth in the Mar-tian regolith using GEANT-4 and compared it to the esti-mated dose rate at the surface determined by MARIE, aMartian radiation environment monitor on MO. Theydetermined that GCR and solar event protons are onlyslightly attenuated in the Martian atmosphere; most aredeposited in the regolith. From balloon experiments car-ried out in the upper atmosphere on Earth, we know thatthe secondary particle production (including neutronsand c-rays) as a result of the nuclear interaction of protonswith planetary material reaches a maximum at approxi-mately 50 g cm�2 (Morthekai et al., 2007). Given that theMartian atmospheric thickness is approximately 16 g cm�2

(Morthekai et al., 2007), this maximum would occur withinthe Martian regolith. This is in agreement with the MonteCarlo model being developed at the UoL. Fig. 2 shows pre-liminary results from the UoL Monte Carlo model, whichshows the flux of neutrons and c-rays as a function of

1022 M.S. Skidmore, R.M. Ambrosi / Advances in Space Research 44 (2009) 1019–1029

depth in the Martian surface. The Figure demonstrates thatthe secondary particle flux of particles at energies>500 MeV reaches a maximum at tens of centimetersbeneath the surface.

2. The MO c-ray spectrometer

MO entered Mars orbit in October 2001 and is still inoperation. The c-ray spectrometer suite was designed todetermine the composition of the surface of Mars fromorbit, by measuring the c-ray and neutron emissions fromthe Martian surface. It became operational in February2002. The instrument suite included the gamma subsystem(hereafter referred to as the c-ray spectrometer), which wasdesigned to measure the c-rays emitted from Mars, the neu-tron spectrometer (NS), which was designed to monitorepithermal and thermal neutron emissions, the high-energyneutron detector (HEND), which was designed to monitorfast and epithermal neutrons detectable from orbit and thecentral electronics box (Boynton et al., 2004).

The c-ray spectrometer consists of a large, n-type, HPGecrystal enclosed in a Ti structure (Boynton et al., 2004). It isseparated from the spacecraft by a 6 m boom to minimise thecontribution of the background from the spacecraft. The NSis a B-loaded plastic scintillator, that uses the differentialvelocity of the neutrons relative to the spacecraft in orderto discriminate between thermal (E < 0.4 eV) and epithermalneutrons (0.4 eV < E < 500 keV), and uses a double pulsecoincidence technique to measure fast neutrons up to�10 MeV (Boynton et al., 2004). The HEND is composedof five different detectors that measure neutrons in the energyrange 0.4 eV to 15 MeV (epithermal and fast neutrons).Three of the detectors incorporate 3He proportional coun-ters and two additional detectors are composed of a singlescintillator that can detect neutrons of energies higher than1 MeV. The HEND instrument is described in detail by Lit-vik and Mitrofanov (2006).

3. Detection of SPEs at the Earth and Mars

There has been a documented correspondence betweenSPEs detected at the Earth by GOES and at Mars byMARIE on 2001 Mars Odyssey. The MARIE instrumentwas a charged particle directional spectrometer that mea-sured particles above 30 MeV, and ceased functioning in2003. It was designed to measure the Martian radiationenvironment. Cleghorn et al. (2004) reported six SPEsbetween March and mid-September 2002 that weredetected by both spacecraft; however the location of thetwo planets in relation to the CME meant that the space-craft observed different effects. For example in the 16thto the 18th March 2002 SPE, GOES-8 experienced a smallenhancement of the high-energy proton flux, whereasMARIE showed that the Martian environment experienceda sharp rise in energetic proton levels (Cleghorn et al.,2004). The small enhancement is consistent with>10 MeV particles diffusing across magnetic field lines,

whereas Mars was well connected to a flux tube originatingat the point of the CME. In the events that occurred on the23rd April, it was MARIE that observed a gradualenhancement and GOES that observed the sharp rise. Cleg-horn et al. concluded that increases in particle flux could bedetected at both planets during a SPE, depending on howwell the planets were connected to the CME and the asso-ciated shock wave. Futaana et al. (2008) studied the SPEthat occurred in December 2006 using instruments on MarsExpress in Mars orbit, Venus Express in Venus orbit andthe GOES spacecraft in Earth orbit. The Venus Expressand Mars Express ASPERA (see Futaana et al., 2008 fora more detailed description of the instrument) instrumentsobserved a large enhancement in the background count lev-els in December 2006 at their respective locations, the tim-ing of the enhancement was consistent with the time itwould take for the energetic particles to reach Venus andMars if they travelled along the field lines of the IMF. Evi-dently some SPEs can be detected at multiple locations inthe solar system, depending on how well they are connectedto the CME event.

3.1. SPE effects on the MO c-ray spectrometer

Evans et al. (2003) reported that a SPE causes increasesin the c-ray and background count rates and can lead tothe distortion of a c-ray spectrum collected by MO. Theincrease in background dominated the signal reducing theability to identify any spectral lines in the data. In somecases the flux levels saturated the system and the detectordead time rose significantly. During the largest SPEs theMO c-ray spectrometer is switched off due to pile-up andleakage current increases induced by high ion and c-rayfluxes (Evans et al., 2003). This implies that the data duringlarge SPEs, such as the July 2002 and November 2003events were not collected. The data for these time periodsare not available on the NASA Planetary Data System(PDS) Geosciences node (http://pds-geosciences.wustl.edu). However data was collected during SPEs that didnot disable the c-ray spectrometer (therefore we can assumethat pile-up effects and leakage current increases were notsignificant) and did not have a negative impact on thec-ray data. Lawrence et al. (2004) demonstrated that thelunar spectra from the Lunar Prospector c-ray spectrometerincreased during a moderate SPE, then decreased to normallevels over a period of several days; however, it was not con-clusive about whether the increased gamma emission origi-nated from the planet or the spacecraft. Lawrence et al. alsosuggested that using these enhanced peaks in the spectra todetermine the elemental composition of the surface couldprovide additional information about the Moon, but nofurther discussion on this topic followed. This study buildson from the work carried out by Lawrence et al., by explor-ing data MO collected during SPEs to ascertain if any usefulinformation may be inferred from it.

Fig. 3 gives an example of the MO c-ray spectrometerspectra taken over a SPE that occurred in January 2005,

Fig. 3. c-ray spectra from the MO c-ray spectrometer collected over a SPE that occurred in January 2005 (Boynton, 2002).

M.S. Skidmore, R.M. Ambrosi / Advances in Space Research 44 (2009) 1019–1029 1023

which was also detected by the GOES spacecraft in orbitaround the Earth. The spectra increased sharply in inten-sity then took several days to return to the original back-ground levels. According to Evans et al. (2007) changesin the c-ray spectrum caused by the solar event of July2002 resulted in a distortion and broadening of the elemen-tal peaks. Fig. 3 highlights the enhancement and broaden-ing of a carbon c-ray peak at 4438 keV as a result of theSPE that took place in January 2005. The aim of Evanset al. was to generate summed and averaged peak intensi-ties for each spectral element over a period of 650 days;the inclusion of the data associated with the SPE wouldhave introduced errors in the estimated concentration ofelements on the Martian surface. For this reason all datacollected over SPEs were excluded from the publishedspectra.

4. SPE data

The globally summed c-ray spectra collected by the c-ray spectrometer on MO were analysed to investigatehow a SPE affects the c-ray peaks of specific elements inthe spectra. Data on the solar proton fluence was obtainedfrom the GOES satellite data node (NGDC, 2008). Thedata collected indicated that SPE events occurred duringthe primary MO mission (2002–2006).

4.1. SPE data for MO c-ray spectrometer

As mentioned above, a SPE detected at the Earth doesnot always affect the Martian radiation environment. Sincethe malfunction of MARIE in 2003, there have been nooperational particle monitors in orbit around Mars to bet-ter study the effects of SPEs on the planet.

A solution was to examine HEND data, and the c-raybackground levels of the c-ray spectrometer spectra. The

HEND on-board MO is sensitive to high-energy protons,as well as fast and epithermal neutrons (however, there isno way to distinguish between the signals generated bythe two particles). The HEND can be used as an indicatorof when there is increased particle flux at Mars, i.e. whenthe SPE protons have reached Mars orbit. The c-ray back-ground in the c-ray spectrometer spectra also increased,indicating that high-energy protons were interacting withthe spacecraft and planet, generating c-rays. The contin-uum was measured at two energies that did not appear tohave any lines of planetary or spacecraft origin; 1070 and7196.25 keV. These two indicators, combined with the datafrom GOES were used to determine whether a SPE hadoccurred and whether it had reached Mars.

Fig. 4 shows the proton fluence measured by the SEMon GOES over 2005. Two high-energy events were identi-fied in January and September. Fig. 5 shows the proton flu-ence as detected by the SEM on GOES, the particle fluxdetected by the fast neutron detector on the HEND andthe c-ray continuum at 1070 and 7196.25 keV detected bythe c-ray spectrometer when the Earth and Mars were innear opposition (January 2005). Fig. 6 shows the samevariables for when the Earth and Mars were closely aligned(August–September 2005). It was concluded that all threeinstruments detected solar event protons or other particlesinduced by the solar event protons, therefore the two plan-ets could have been well connected to the CME. Unfortu-nately this does limit the usefulness of the data, since thistechnique is reliant on a SPE affecting both the Earthand Mars systems in order to verify that the increases inthe HEND particle flux and c-ray background in the c-ray spectrometer are due to energetic solar protons reach-ing Mars.

Ambrosi et al. (2005) determined that protons of ener-gies less than 20 MeV would not penetrate the Martianatmosphere, therefore only the >15 MeV proton fluence

Fig. 4. EPS (15–500 MeV) and HEPAD (510 MeV to >700 MeV) data from GOES-11 for 2005 (corrected proton channel data from GOES-11) (NDGC,2008).

Fig. 5. The EPS and HEPAD measured proton fluence from GOES-11, the HEND particle flux and the c-ray spectrometer continuum count rate at 1070and 7196.13 keV from MO in January 2005 (NDGC, 2008; Boynton, 2002, 2004).

1024 M.S. Skidmore, R.M. Ambrosi / Advances in Space Research 44 (2009) 1019–1029

is included in this study for the case of Mars. This studyand the current Monte Carlo model being developed atthe UoL aim to refine the initial calculation by Ambrosiet al. (2005).

5. c-ray data analysis

The corrected c-ray spectra data from the MO c-rayspectrometer and the HEND data were obtained fromthe NASA PDS Geosciences node (Boynton, 2002). The

data consists of cumulative counts of c-rays over 19.7 speriods binned into 16,384 channels. The corrected c-rayspectra have been processed so that they have a commonenergy scale in order to co-add or compare spectra (Crom-bie et al., 2003). Co-adding spectra often occurs in GRS inorder to get a single spectrum with good counting statistics.A Java code was developed to extract the data from thecorrected c-ray data binary files and convert them intoASCII format. The c-ray spectra for the SPEs and theirdecay phases in January and September 2005 were co-

Fig. 6. The EPS and HEPAD measured proton fluence from GOES-11, the HEND particle flux and the c-ray spectrometer continuum count rate at 1070and 7196.13 keV from MO in September 2005 (NDGC, 2008; Boynton, 2002, 2004).

M.S. Skidmore, R.M. Ambrosi / Advances in Space Research 44 (2009) 1019–1029 1025

added to produce a single spectrum with enough counts init to ensure that uncertainties in the peaks were lower than10%.

5.1. Spectral analysis

The spectral analysis was carried out using a peak-fittinganalysis package called OriginTM (Version 7.5). The SPEspectrum was analysed using the package, the peaks wereidentified and the associated count rates were calculated.In the case of the MO spectra, the peaks were identifiedwith the help of data reduced by Evans et al. (2007). Thepeaks were fitted with Gaussian, multiple Gaussian orasymmetric Gaussian (with a low energy tail) functions.Low energy tailing occurs when an HPGe detector hasradiation damage. This is discussed further in Evanset al. (2007). OriginTM is also capable of producing a residu-als plot, and these were examined to determine whether anypeaks had been overlooked.

When a SPE occurs, it often has a large impact on theentire spectrum. This was usually in the form of an inten-sity shift in the whole continuum but changes in the shapesand sizes of the elemental peaks were also seen. Fig. 3 dis-plays these changes; the background in the MO c-ray spec-trometer spectra increases over a SPE, and several of thepeaks become both larger and broader.

6. Results

Table 1 shows the peak count rates extracted from theMO c-ray spectrometer data. In order to compare theMO c-ray spectrometer count rates in Table 1 to normal

solar condition data, readers are directed to Table 1 inEvans et al. (2007).

The c-ray lines included in Table 1 are those that havean uncertainty of less than 10% and have statistically sig-nificant increases compared to the normal solar conditiondata (the data was considered significant if the count rateincrease was greater than a 3-sigma limit to the normalsolar condition data). The uncertainty was set at this levelto ensure that the peaks were easily distinguishable againstthe increased background and that they could be fittedaccurately using the peak-fitting programme.

7. Discussion

When comparing the results of this study to Table 1 inEvans et al. (2007), two aspects are clear:

(a) Many of the lines believed to be of planetary origin(such as many of the Cl lines) could not be identifiedin the SPE data. The majority of lines that could bepositively identified and were statistically significantwere ones that are suspected to have a significant con-tribution from the spacecraft itself (such as Fe, Mg,Ti, Ge, Si, Al, C, O, Zn, F, K, V, Cr, Ca and otherelements produced as a result of activation of thespacecraft by the high-energy particle interactions,e.g. the 48Ti(p,n)48V or the 48Ti(n,p)48Sc reaction),or the atmosphere (C and O).

(b) The identified lines often became broader or sufferedfrom low energy tailing as a result of the SPE, this isevident from the larger full width half maximum ofthe peaks.

Table 1MO c-ray spectrometer peak data analysis results for the summed SPE data over 2005. The reference energies [1] were obtained from Evans et al. (2007).Key: The reactions that take place for the c-rays to be created in the list above are inelastic scatter reactions (n,ng) unless stated otherwise; NC, neutroncapture; NE, non-elastic reaction; �, multiple reactions to create the c-ray; EC, c-ray created as a result of electron capture; SE, single escape peak; DE,double escape peak.

Energy (keV) Possible source (s) Reference energy (keV) [1] SPE spectra

Counts (min) Uncertainty (%) FWHM (keV)

808 70-Ge NC/73-Ge NC 808.1 0.78 0.9 5.719835.4 54-Mn 835.8 1.454 0.7 6.825841 54-Mn 840.8 1.467 0.7 4.77846.9 56-Fe* 846.7 1.121 0.8 5.539872.6 69-Ge EC 873.7 0.366 1.3 5.702882.3 69-Ge EC 882.5 2.72 0.5 8.219936.8 52-Cr*/52-Mn 935.5 0.94 0.8983.6 48-Ti*/48-V 983.5 7.737 0.3 4.547991.2 64-Zn* 991.6 2.25 0.5 4.981014.5 27-Al* 1014.4 0.333 1.4 5.2951021.9 10-C/511 keV SUM 1021.7 0.424 1.2 6.1111039.2 48-Sc/70-Ge 1039 1.769 0.6 6.051049 66-Ga EC 1048.9 1.208 0.7 6.4341077.4 68-Zn* 1077.3 0.305 1.4 5.7121091.9 70-Ge NC 1095.8 0.251 1.6 8.1221108.2 69-Ge EC 1108 0.43 1.2 4.5211129.6 26-Mg* 1129.6 1.355 0.7 11.7751147 47-Sc* 1147 0.456 1.2 5.5771157 44-Sc/44-Ca* 1157 1.754 0.6 6.3351204.3 73-Ge NC 1204.2 0.268 1.5 5.1361230 1.96 0.6 14.1291237.8 56-Fe*/214-Bi 1238.3 0.877 0.8 9.4191285.1 47-Ti* 1284.9 0.512 1.1 4.711312.7 48-Ti*/48-V 1312.1 2.354 0.5 7.0241333.8 60-Co 1332.5 1.07 0.8 7.7021346.6 69-Ge EC 1347 0.82 0.9 7.1541368.5 24-Mg* 1368.6 20.513 0.2 9.5471410 54-Fe*/55-Fe* 1408.1/1408.4 0.946 0.8 8.2421454.7 58-Ni* 1454.3 0.659 1 6.1491525.3 40-Ca* 1524.7 0.754 0.9 8.8581553.9 50-Ti* 1553.8 0.456 1.2 8.5051611.9 56-Fe NC 1612.8 4.185 0.4 29.0821634.5 20-Ne*/20-F 1633.9 1.861 0.6 13.6571723.9 48-Ti NC SUM 1723.5 0.271 1.5 8.4181727.5 56-Fe NC 1725.3 0.494 1.1 17.6831760.5 0.594 1 15.4241794 48-Ti NC 1793.5 0.295 1.5 9.6111808.7 26-Mg* 1808.7 2.599 0.5 11.9511810.8 56-Fe* 1810.8 0.78 0.9 10.1651824.5 48-V + 511 keV 1823.1 0.205 1.8 5.0631982 18-O* 1982 1.207 0.7 30.1622009.4 46-Ti* SUM 2009.8 0.985 0.8 22.5222028.2 29-Si* 2028.1 1.324 0.7 33.2032108.3 56-Fe 2112.9 1.141 0.7 65.3392211 27-Al 2211 0.482 1.1 17.6612295.1 48-Ti* SUM 2295.6 0.265 1.5 11.9392300.4 48-V SUM EC 2300.6 0.184 1.9 5.5032312.6 14-O 2312.6 0.339 1.4 7.7622347.2 0.369 1.3 16.8372376.4 48-Ti* 2375.2 0.988 0.8 17.6862517 28-Si NC DE 2517 2.265 0.5 38.1512571.3 2.229 0.5 38.8552599 56-Fe* 2600 0.632 1 18.3732638.7 48-Ti NC 2635.7 2.984 0.5 41.7342719.4 56-Fe NC 2721.2 0.282 1.5 14.8192750.2 16-O 2741.5 3.719 0.4 37.5252867.9 24-Mg* 2869.5 0.535 1.1 21.1933222.6 48-Ti NC SE 3222.6 0.244 1.6 25.918

1026 M.S. Skidmore, R.M. Ambrosi / Advances in Space Research 44 (2009) 1019–1029

Table 1 (continued)

Energy (keV) Possible source (s) Reference energy (keV) [1] SPE spectra

Counts (min) Uncertainty (%) FWHM (keV)

3370 3369.7 0.929 0.8 53.33410.4 56-Fe NC 3413.1 1.211 0.7 33437.5 56-Fe NC 3436.7 2.024 0.6 57.7673475.1 48-Ti NC 3475.6 3.993 0.4 89.4953539 28-Si NC 3539 0.762 0.9 38.8553735 48-Ti NC/40-Ca* 3733.6/3736.5 0.39 1.3 23.4173863.1 48-Ti NC DE 3859.4 0.346 1.4 21.7173934.8 12-C* SE 3927 3.125 0.4 72.2994151.5 0.54 1.1 49.734218 56-Fe NC 4218.3 0.408 1.2 23.2154247.8 16-O NE DE 4247.2 0.791 0.9 28.694433.1 12-C* 4438 8.388 0.3 77.5955089.8 5088.9 0.72 0.9 47.0195107.6 16-O DE 5106.6 0.607 1 13.6395239.3 15-O* 5239.9 0.438 1.2 34.7345268.4 16-O NE 5269.2 0.681 1 30.1495619.4 16-O* SE 5617.6 1.985 0.6 17.5135629.9 16-O 5617.6 1.966 0.6 90.3825868.5 28-Si NC SE 5868.8 0.945 0.8 96.9825906.3 48-Ti NC SE 5907.4 0.382 1.3 23.776127.4 16-O 6128.6 3.638 0.4 34.6446139 16-O 6128.6 1.653 0.6 116.7656365.2 28-Si SE 6366.9 0.435 1.2 61.1026421.9 48-Ti NC 6418.4 0.974 0.8 37.7436611.2 16-O SE 6604.2 0.77 0.9 81.3126874.7 28-Si 6877.9 0.19 1.8 41.8926901.3 6903 0.489 1.1 62.6156927.4 16-O 6915.5 0.506 1.1 100.5027117.6 56-Fe NC SE 7120.1 1.188 0.7 73.4317131.1 16-O 7115.2 0.386 1.3 37.913

M.S. Skidmore, R.M. Ambrosi / Advances in Space Research 44 (2009) 1019–1029 1027

The lines of both planetary and spacecraft origin thatwere ‘lost’ in the SPE spectra were often close to a line thatunderwent a strong enhancement/broadening as a result ofthe SPE, making that line unidentifiable. It is suspectedthat the weaker lines could have been masked by theincreased c-ray background from the spacecraft.

The conclusion of this study is that it cannot be con-firmed that any of the c-ray peaks identified in the SPEspectra originated from the surface of Mars. A few S andCl lines were identified in the spectra at 3370.0 keV (S),5089.8 keV (Cl) and 6901.3 keV (Cl) and are highlightedin Table 1. However, it is suspected that the peak increaseswere due to other sources other than these planetary ele-ments, because other S and Cl peaks could not be identifiedin the rest of the spectra. The two Cl peaks are escapepeaks, but the full-photo peak could not be identified inthe spectra, agreeing with the theory that these peaks mustbe due to other sources. The alternative sources could be48Ti(n,nc) or 56Fe(n,nc) for the 3370.0 keV line, 52Cr(p,c)for the 5089.8 keV line, 52Cr(p,c) for the 6901.3 keV line.

The fact that ‘pure’ planetary lines could not be identi-fied in the MO c-ray spectra does not necessarily imply thatthe increased c-ray flux associated with a SPE would not beuseful for the purposes of in-situ Martian GRS. The space-craft contained a number of different elements that wereactivated by the SPE protons, and since the spacecraft iscloser to the MO c-ray spectrometer than the planet, the

c-ray lines from it are stronger, masking any planetarysignal. An in-situ c-ray spectrometer would have the bene-fit of a larger signal from the surface, therefore the signal-to-background ratio should be higher. It would also beessential to know the exact composition of the rest of thepenetrator/lander/rover so the c-ray contribution from itcould be subtracted from the c-ray spectra, leaving onlyc-rays of planetary origin. If a c-ray spectrometer wereintegrated into a sub-surface probe, e.g. as part of a heatflow and physical properties package (Ambrosi et al.,2006), the surface material around it could shield it fromsome of the surface radiation environment.

An essential factor in accurately modelling the c-ray emis-sion from the Martian surface and sub-surface is knowledgeof the surface radiation environment. The Mars Science Lab-oratory mission that will be launched in 2009 and arrive atMars in 2011, contains the radiation assessment detector(RAD). The RAD is designed to fully characterise the ener-getic particle spectrum (charged particles, neutrons andc-ray fluxes up to 100 MeV energies) that is incident on theMartian surface, including those created in the Martianatmosphere and regolith. This will give us a very detailedview of the radiation environment on the Martian surfaceand will lead to the development of a more accurate plane-tary radiation environment model (Wimmer-Schweingruberet al., 2006). If a c-ray spectrometer were selected for an in-situ planetary science mission, having a proton monitor on

1028 M.S. Skidmore, R.M. Ambrosi / Advances in Space Research 44 (2009) 1019–1029

the surface of Mars at the same time would be extremelyvaluable; the information could be applied to the radiationenvironment model to accurately determine the elementalcomposition of the surface/sub-surface.

The increased c-ray signal collected by an in-situ c-rayspectrometer implies that the concentration of elementsin that region will be determined with greater accuracythan from orbit. It may also be possible to detect more ele-ments that have previously been undetectable from space,such as Br, Sm and Gd. Br has been detected in large quan-tities on the Martian surface by X-ray spectrometers on thesurface of Mars, but not by the MO c-ray spectrometer(Gellert et al., 2004). If a c-ray spectrometer is deployedin a sub-surface probe, the higher energy solar proton dis-tribution of a SPE will mean that protons will be able topenetrate and induce c-ray emission at greater depths, per-haps below the oxidant extinction depth, leading to a moreaccurate estimate of the bulk composition of Mars (Hager-mann et al., 2005). The addition of a small c-ray source willallow the bulk density to be determined by the c-ray back-scatter method. The ability to measure the composition ofa planetary surface and sub-surface by GRS as well as thebulk density (and through these measurements the plane-tary heat flow, see Ambrosi et al., 2006) implies that anin-situ c-ray spectrometer would be a powerful instrumentin planetary surface and evolution studies.

8. Summary

This study has shown that there is a correspondencebetween the proton fluence emitted from the Sun and thec-ray emission detected in Mars orbit, but it is inconclusivewhether this causes increased c-ray emission from the planetand whether it may be of use in an in-situ GRS mission. Theelemental peaks detected by the 2001 Mars Odyssey c-rayspectrometer can all be attributed to elements that are pres-ent on the spacecraft. From this study it is clear that for afuture in-situ GRS mission it would be essential to minimisethe elements present on the in-situ lander/probe to avoidinterferences with planetary lines and to fully characterisethe contribution of the lander/probe to the c-ray spectrumcollected. It has also highlighted the need to fully character-ise the radiation environment on the surface of a planetarybody to correctly interpret the c-ray emission data and con-vert it into meaningful estimates of the concentration of eachelement on the surface. A neutron generator would signifi-cantly reduce acquisition times of c-ray spectra; howeverthere is a trade-off between the desired speed of spectralacquisition, against the mass, volume and power constraintsof the lander/probe.

Acknowledgements

The authors thank EPSRC for funding this work, Dr.William Boynton and the GRS team at the University ofArizona and Dr. David Lawrence at the Johns HopkinsUniversity Applied Physics Laboratory.

References

Ambrosi, R.M., Ball, A.J., Nelms, N., Talboys, D., Bentley, M.S., Spohn,T., Nadalini, R., Hagermann, A., Richter, L., Calcutt, S., Fraser,G.W., Bridges, J., Zarnecki, J.C., Cockell, C.S., Grady, M.M., Simon,H., Siffert, P., Skidmore, M.S. X-ray and c-ray Spectrometers for theHeat Flow and Physical Properties Instrument (HP3) on the ExoMarsGeophysical and Environmental Package. Presented at the 4thInternational Planetary Probe Workshop, Pasadena, 27–30 June2006, Conference Proceedings, 4th International Planetary ProbeWorkshop, 2006.

Ambrosi, R.M., Talboys, D., Sims, M.R., Bannister, N.P., Makarewicz,M., Stevenson, T., Hutchinson, I.B., Watterson, J.I.W., Lanza, R.C.,Richter, L., Mills, A., Fraser, G.W. Neutron activation analysis, c-rayspectrometry and radiation environment monitoring instrument con-cept: GEORAD. Nucl. Instr. Meth. A 539 (1–2), 198–216, 2005.

Barnes, A. Acceleration of the solar wind. Rev. Geophys. 30, 43–55, 1992.Boynton, W.V. Mars 2001 Odyssey, c-ray Spectrometer Corrected

Gamma Spectra V1.0, ODY-M-GRS-4-CGS-V1.0, NASA PlanetaryData System Geosciences Node, 2002.

Boynton, W.V. Mars 2001 Odyssey Gamma Ray Spectrometer DerivedHigh Energy Neutron Detector Data V1.0, ODY-M-GRS-4-DHD-V1.0, NASA Planetary Data System Geosciences Node, 2004.

Boynton, W.V., Feldman, W.C., Mitrofanov, I.G., et al. The Mars Odysseyc-ray spectrometer instrument suite. Space Sci. Rev. 110, 37–83, 2004.

Cane, H.V. The large-scale structure of flare associated interplanetaryshocks. J. Geophys. Res. 93, 1–6, 1988.

Cleghorn, T.F., Saganti, P.B., Zeitlin, C.J., Cucinotta, F.A. Solar particleevents observed at Mars: dosimetry measurements and model calcu-lations. Adv. Space Res. 33 (12), 2215–2218, 2004.

Crombie, M.K., Harshman, K.P., McArthur, G.K., Shinohara, C.,Boynton, W.V. 2001 Mars Odyssey c-ray Spectrometer Science DataProducts. Lunar Planet. Sci. XXXIV, Abstract No. 1671, 2003.

Evans, L.G., Reedy, R.C., Boynton, W.V. Effects of solar particles on theMars Odyssey c-ray spectrometer. Lunar Planet. Sci. XXXIV,Abstract No. 1591, 2003.

Evans, L.G., Reedy, R.C., Starr, R.D., Kerry, K.E., Boynton, W.V.Analysis of c-ray spectra measured by Mars Odyssey. J. Geophys. Res.111, E03S04, doi:10.1029/2005JE002657 (printed 112(E3), 2007), 2007.

Futaana, Y., Barabash, M., Yamauchi, S., et al. Mars Express and VenusExpress multi-point observations of geoeffective solar flare events inDecember 2006, Planet. Space Sci. doi:10.1016/j.pss.2007.10.014, 2008.

Gellert, R., Rieder, R., Anderson, R.C., Bruckner, J., Clark, B.C.,Dreibus, G., Economou, T., Klingelhofer, G., Lugmair, G.W., Ming,D.W., Squyres, S.W., d’Uston, C., Wanke, H., Yen, A., Zipfel, J.Chemistry of rocks and soils in Gusev Crater from the alpha particleX-ray spectrometer. Science 305 (5685), 829–832, 2004.

Hagermann, A. Planetary heat flow measurements. Philos. Trans. R. Soc.Lond. A 363 (2), 2777–2791, 2005.

Hundhausen, A.J. The solar wind, in: Kivelson, M.G., Russell, C.T.(Eds.), Introduction to Space Physics. Cambridge University Press, pp.91–128, 1995.

Kaplan, I., second ed Nuclear Physics, vol. 16 Addison Wesley PublishingCompany, Inc., USA, 1963.

Krane, K.S.. Introductory Nuclear Physics, vol. 11. John Wiley and Sons,1988.

Lawrence, D.J., Maurice, S., Feldman, W.C. Gamma-ray measurementsfrom lunar prospector: time-series data reduction for the gamma-rayspectrometer. J. Geophys. Res. Planets 109, E07S05, doi:10.1029/2003JE002206, 2004.

Litvik, M.L., Mitrofanov, I. Mars 2001 Odyssey c-ray spectrometer High-energy Neutron Detector data processing Version 1.2. Available from:http://pds-geosciences.wustl.edu/geodata/ody-m-grs-4-dhd-v1/odgr1_xxxx/document/grs_hend_processing.pdf, 2006.

Masarik, J., Reedy, R.C. c-ray production and transport in Mars. J.Geophys. Res. 101 (E8), 18891–18912, 1996.

Morthekai, P., Jain, M., Dartnell, L., Murray, A.S., Bøtter-Jensen, L.,Desorgher, L. Modelling of the dose-rate variations with depth in the

M.S. Skidmore, R.M. Ambrosi / Advances in Space Research 44 (2009) 1019–1029 1029

Martian regolith using GEANT4. Nucl. Instr. Meth. A 580 (1), 667–670, 2007.

NGDC, NOAA, Space Physics Interactive Data Resource (SPIDR).Available from: <http://spidr.ngdc.noaa.gov/spidr/>, 2008.

Reames, D.V. Particle acceleration at the Sun and in the heliosphere.Space Sci. Rev. 90, 413–491, 1999a.

Reames, D. Solar energetic particles: is there time to hide. Radiat. Meas.30, 297–308, 1999b.

Reedy, R.C., Arnold, J.R. Interaction of the solar and galactic cosmic rayparticles with the Moon. J. Geophys. Res. 77, 537–555, 1972.

Skidmore, M.S., Ambrosi, R.M., Simon, H. Hybrid X-ray and c-rayspectrometer for in-situ planetary science missions. Nucl. Instr. Meth.A 604 (3), 592–603, 2009.

Surkov, Y. Section 2.1 – c-ray Spectrometric Studies, Explorationof Terrestrial Planets from Spacecraft, Instrumentation, Investi-gation, Interpretation, second ed Wiley-Praxis, pp. 140–244,1997.

Wimmer-Schweingruber, R.F., Hassler, D.M., Kortmann, O., et al. TheMars Science Laboratory Radiation Assessment Detector, ThirdEuropean Space Weather Week, Brussels, 13–17 November, 2006.