planetary penetrators: their origins, history and future

Available online at www.sciencedirect.com

www.elsevier.com/locate/asr

Advances in Space Research 48 (2011) 403–431

Planetary penetrators: Their origins, history and future

Ralph D. Lorenz ⇑

Johns Hopkins University, Applied Physics Laboratory, Laurel, MD 20723, USA

Received 6 January 2011; received in revised form 19 March 2011; accepted 24 March 2011Available online 30 March 2011

Abstract

Penetrators, which emplace scientific instrumentation by high-speed impact into a planetary surface, have been advocated as an alter-native to soft-landers for some four decades. However, such vehicles have yet to fly successfully. This paper reviews in detail, the originsof penetrators in the military arena, and the various planetary penetrator mission concepts that have been proposed, built and flown.From the very limited data available, penetrator developments alone (without delivery to the planet) have required �$30M: extensiveanalytical instrumentation may easily double this. Because the success of emplacement and operation depends inevitably on uncontrol-lable aspects of the target environment, unattractive failure probabilities for individual vehicles must be tolerated that are higher than thetypical ‘3-sigma’ (99.5%) values typical for spacecraft. The two pathways to programmatic success, neither of which are likely in an aus-tere financial environment, are a lucky flight as a ‘piggyback’ mission or technology demonstration, or with a substantial and unprec-edented investment to launch a scientific (e.g. seismic) network mission with a large number of vehicles such that a number of terrain-induced failures can be tolerated.� 2011 COSPAR. Published by Elsevier Ltd. All rights reserved.

Keywords: Penetrators; Space missions; Landers; Moon; Mars

1. Introduction

Penetrators have been proposed as planetary explora-tion vehicles for four decades, but have yet to fly success-fully. Penetrators are self-contained vehicles designed tofunction after traversing some distance in a solid targetusing the kinetic energy of their arrival. This definition(where ‘some distance’ is understood to imply a lengthscale comparable with or greater than the vehicle length)excludes self-hammering drills such as the ‘mole’ carriedon Beagle 2, and instruments (‘penetrometers’) to measuresurface mechanical properties which are part of a largervehicle, such as those on Venera landers or the Huygensprobe (e.g. Lorenz et al., 1994). All of these systems are dis-cussed in two proceedings volumes (Komle et al., 2001;Kargl et al., 2009) following workshops in Graz, Austriain 1999 and 2006, and although the terms ‘penetrometer’(an instrument) and ‘penetrator’ (a vehicle) are often

0273-1177/$36.00 � 2011 COSPAR. Published by Elsevier Ltd. All rights rese

doi:10.1016/j.asr.2011.03.033

⇑ Tel.: +1 443 778 2903.E-mail address: [email protected].

sloppily interchanged, we would urge rigor in maintainingthe distinction. The definition, by requiring a functionpost-impact, also excludes inert impactors such as cannonshot, upper stages used to impact the moon, or the cometimpactor on the Deep Impact mission, whose only functionis to deposit kinetic energy in a target with no post-impactoperation.

Although penetrator advocates often cite extensive mil-itary experience with such systems, the history of militarydevelopments and its relationship to the planetary programhas not been laid out in the literature previously. Thereforethis paper discusses the military developments that led tothe advocacy of penetrators as planetary explorationvehicles.

The paper then discusses in detail the various planetarypenetrator mission concepts (e.g. Fig. 1) that have beenproposed, and the smaller number that have been devel-oped and/or flown: the timeline of developments is shownin Fig. 2. Finally, the relationship of the probability of fail-ure of an individual vehicle’s emplacement and the real orperceived likelihood of programmatic success of the

rved.

http://dx.doi.org/10.1016/j.asr.2011.03.033

mailto:[email protected]

http://dx.doi.org/10.1016/j.asr.2011.03.033

Fig. 2. A schematic of the development history of various penetrator projects. ‘S’ denotes study, while ‘D’ represents more substantial development,including penetration tests, hardware construction and flight.

Fig. 1. A selection of penetrator vehicles shown to the same scale. (A) An ACUSID air-dropped acoustic monitor used in the Vietnam conflict, (B) theoriginal Comet Rendezvous Asteroid Flyby (CRAF) penetrator – note the wide flare, (C) the Mars-96 penetrator (D), NASA proposed 1970’s Marspenetrator, (E) Lunar-A penetrator, and (F) DS-2. Note the remarkably small size of DS-2 compared with the other vehicles.

404 R.D. Lorenz / Advances in Space Research 48 (2011) 403–431

mission overall is discussed, and the (narrowing) scientificniche for penetrators is noted. A detailed discussion of

the scientific capabilities of instrumentation on penetratorswill be the subject of a separate paper.

1 These larger dams were not vulnerable to the bouncing bombs usedagainst the smaller concrete dams for which Barnes Wallis is more widelyknown).

R.D. Lorenz / Advances in Space Research 48 (2011) 403–431 405

2. Terrestrial penetrators

The origination of penetrators as planetary explorationvehicles is inevitably connected with certain military devel-opments, which we will review here to set the planetaryapplications in context. There are also a few examples ofcivilian or dual-use instrumentation deployed in thismanner.

2.1. Early penetration mechanics

Probably the first quantitative assessment of projectilepenetration into the ground was by Benjamin Robins,Engineer-General to the East India Company, notably inhis treatise ‘New Principles of Gunnery’ (Robins, 1742),which applies Newton’s (1687) analytic methods to artil-lery. Robins developed a number of remarkable innova-tions in instrumentation, notably he devised the ballisticpendulum to measure the launch velocity of projectiles,as well as a whirling-arm apparatus to measure aerody-namic drag. Even viewed purely as an aerodynamicist, hiscareful investigations were far ahead of his time. He firstrecognized the utility of streamlined projectiles and bycareful experimental gunnery (using specially-prepared pre-cision cannonballs, and accurately-measured powdercharges) he was able to detect the effect of spin on projec-tiles, namely the Robins–Magnus effect, often ratherunfairly referred to only by the latter, whose experimentstook place some 150 years later (e.g. Lorenz, 2006). Heconfirmed the dependence of drag on the square of velocityand even detected the increase the transonic rise in dragcoefficient (the ‘sound barrier’).

From his experiments, in particular with the ballisticpendulum, Robins’ determined that the penetration of shotinto typical solid materials varies as the square of velocity.This essentially means the kinetic energy of delivery is dis-sipated by work against a constant strength target (kineticenergy per unit area has the same dimensions as strengthtimes penetration distance – equivalently the kinetic energyper unit volume of ‘tunnel’ swept by the projectile is a mea-sure of target strength.)

2.2. World War 2 ‘earthquake bombs’

Despite these rigorous beginnings (Robins work becamea standard reference, translated into German by Euler)gunnery and penetration mechanics remained a largelyempirical business for nearly two centuries (and indeed,to some extent remains so, as evidenced by widely-usedequations by Wayne Young of Sandia (Young, 1997)), inpart due to the transition from the use of solid shot (wherepenetration more or less is the desired effect) to explosiveprojectiles, where penetration is generally less important.

That said, in World War II aerial bombing reached alevel of fidelity where empirically-designed bombs weretailored to specific targets: notably, for heavy concrete orfor the steel decks of battleships, armor-piercing bombs

(‘Panzersprengbombe’, Fleischer, 2003) with more care-fully engineered casings (more slender-nosed, harder steel,and a larger casing-to-explosive ratio) were used by theLuftwaffe. A radio-controlled armor-piercing bomb (thePC-1400, or Fritz-X, e.g. Wolf, 1988 or the recent detailedaccount by Bollinger (2010)) was also fielded. Since pene-trating weapons typically must be close to their targets,the introduction of guidance is a key development and thisand similar radio-controlled weapons subsequently devel-oped in the USA Azon and Razon (e.g. Schmitt, 2002)set the stage for future precision-guided weapons, to whichwe will momentarily return.

Of particular interest in connection with planetary pen-etrators (and better documented in the English-languageliterature), however, are the large unguided weapons thatwere developed in Britain to address a class of militaryand industrial targets such as large earthen dams1 such asthe Sorpe dam which could not be practicably breachedby conventional bombing. The British engineer BarnesWallis proposed (e.g. Murray, 2009) that a large penetrat-ing bomb could efficiently couple its explosive energy to thedam by seismic means (hence ‘earthquake bomb’). In factthe mechanism is largely one of explosively creating anunderground cavity (‘camouflet’) into which the targetstructure subsequently collapses.

This mission required not only a weapon with a largeexplosive warhead, but also a hardened nose such thatthe weapon could burrow to depth. In addition to thestructural strength required, the explosive filling had to tol-erate the impact loads without detonating prematurely.Furthermore, the fuzing systems had to function afterimpact – often three separate fuzes were installed in orderto ensure that at least one would work (although even then,at least one bomb failed to explode – one Tallboy bombwas found in the mud at the Sorpe dam when the reservoirwas drained in 1958).

Compared to other bombs, these weapons (first the Tall-boy 5443 kg, then the Grand Slam, 10,160 kg – even largerUS weapons such as the T-12 were developed later, seeFig. 3) were fairly slender and streamlined – the Tallboywas 6.4 m long. This assured both effective penetration ofthe target (for a constant target strength and impactenergy, following the considerations in Section 2.1, thepenetration distance will be maximized by reducing theprojectile cross-sectional area) but also maximized theimpact velocity.

In order to reach a high enough impact speed, theseweapons had to be dropped from high altitude, 6000 mor more. As well as challenging the lift capability of thelargest aircraft available (the Lancaster) this demandedsophisticated bomb-aiming, using electromechanical com-puters to correct for speed variation with altitude, winddrift, etc. Trials noted that good penetration performance

Fig. 3. A T-12 ‘Cloudmaker’ bomb – a larger version of the WWIITallboy and Grand Slam weapons, designed to penetrate into the earthand damage a target seismically. Note the canted tail fins to effect spin forstabilization.

2 Durandal, poetically, was the sword of the Frankish warrior Roland,Charlemagne’s chief paladin in the fight against the Saracens in the 8thcentury . Roland was also the name for a comet lander to be deployed bythe Rosetta mission: the lander was subsequently renamed Philae.


required that the bomb be closely aligned with its velocityvector at impact – a slight angle of attack would cause thebomb to veer in its subsurface trajectory. Such deviationsnot only reduced depth performance, but were also morelikely to lead to structural failure of the vehicle due tothe side-loads (see also Section 3.8).

Achieving an adequate attitude for penetration was dif-ficult, as drop tests showed that the bomb tended to expe-rience pitch oscillations in flight, which also degraded theimpact speed. Performance was enhanced considerably bycanting the tail fins by 5� to cause a stabilizing spin duringflight – at impact the spin reached 300 rpm, and impactspeed reached the speed of sound (�335 m/s).

We see that these weapons anticipate several aspectsconfronted by planetary penetrators, namely the challengein setting up the vehicle’s velocity and attitude state atimpact and in surviving impact, as well as the improvedseismic coupling afforded by this type of vehicle comparedwith simple surface delivery.

More significantly than the dams for which they wereoriginally conceived, these penetrating weapons were usedsuccessfully against a number of different hardened targets.Some notable examples include the armored deck of theTirpitz battleship, the U-boat facility at Farge (coveredwith 8m of concrete), the massive concrete dome of theV-2 rocket complex (‘La Coupole’) at Wizernes, the rail-way tunnels at Saumur, railway bridges and the never-usedmultibarrel V-3 ‘Supergun’ installation at Mimoyecques.The success of the weapons led to their emulation in theUSA, with the 10,000 kg Amazon penetrating bomb, andthe guided Tarzon bomb, and anticipates the attack of

hardened targets by penetrating weapons in the Gulf Warnearly 50 years later.

2.3. Modern earth-penetrating weapons (‘Bunker Busters’)

The post-WWII era has seen a vigorous application ofscientific principles to the performance of military systems,and thus penetrating weapons for specific missions havereached a high degree of development, supported bysophisticated computational methods (finite-element mod-els and hydrocodes), together with airgun and rocket-sledtesting. The focused nature of such weapons requires adelivery that is high speed and accurate. Two classes ofweapon are of note.

First is the somewhat exotic runway-cratering munitionexemplified by the French Durandal,2 an early version ofwhich was used by Israel in the Six Day War (althoughin fact a number of rocket-assisted bombs were developedin Germany in WWII). An effective way to limit the effec-tiveness of an enemy air force is to disable the runwaysfrom which it might take off. Runways are somewhat slen-der targets and thus unguided weapons must be droppedfrom low altitude in order to have a good probability ofhitting the tarmac (as also the case for railways – see Sec-tion 2.4). However, general-purpose bombs, especially ifdropped from low altitude, tend to produce only shallowcraters that are easily repaired. The Durandal weapon isa small, slender bomb that is first decelerated by a para-chute – this allows a delivery from aircraft at high speedand low altitude, reducing the aircraft’s vulnerability andminimizing the dispersion of impact points due to winddrift etc. The parachute deceleration effects a gravity turn,allowing the weapon to hang more vertically over the tar-get. At this point, the parachute is released and a rocketmotor is ignited, which accelerates the hard-nosed projec-tile into the tarmac. After penetrating some 40 cm into con-crete, the warhead is detonated, forming a 5 m deep and16 m wide crater

The other way of achieving an accurate high speed deliv-ery is by high altitude drop, but with active guidance toensure a small miss distance. A tactical need for deep pene-trating weapons emerged during the First Gulf War in1990, when it was recognized that key Iraqi facilities (chem-ical weapons sites, command and control bunkers, etc.)were protected from air attack by thick concrete sheltersand/or burial under sand.

By this time, guidance systems for air-dropped muni-tions had become readily available, and indeed laser-guid-ance kits (whose development in fact began in the 1960s –e.g. Gillespie, 2006) could be retrofitted in a modular wayonto existing ‘dumb’ bombs, it was relatively straightfor-ward to implement an earth-penetrating weapon. Indeed,


the GBU-28 air-dropped precision penetrating munition(Fig. 4) was developed in haste (�2 weeks) during the GulfWar to attack buried command and control facilities. Theoriginal bodies of this slender (35 cm diameter, 7.6 m long)weapon were in fact machined from old 8-inch artillery gunbarrels. The weapon’s mass was 2270 kg – about half thatof the Tallboy above, but the penetration performance(30 m of earth, or 6 m of concrete) is reported as being sim-ilar. A variety of precision penetrating weapons hasemerged, from the very large (the 13,600 kg Massive Ord-nance Penetrator, GBU-57A/B) to the quite small (the250 kg Isreali MPR-500) – the latter being attractive fromthe point of view of causing minimal collateral damageby virtue of its modest warhead, yet still able to penetrate1 m of reinforced concrete, or four concrete floors.

In an interesting parallel to the application of acceler-ometers on planetary vehicles to measuring layering inthe subsurface, the most modern weapons of this type areequipped with a ‘Hard Target Smart Fuze’ which use accel-erometers to ‘count floors’ to detonate at a specified flooror depth, whereas the fuzing on the original Tallboys sim-ply actuated some seconds or minutes after impact.

Finally, although functionally rather similar to the con-ventional penetrating weapons above, a separate categoryof nuclear penetration weapons should be noted, if for noother reason than they have been documented in somedetail. There was some recent controversy over proposalsduring Bush administration to develop a penetratingnuclear weapon, the RNEP (Robust Nuclear Earth Pene-trator). As described in a National Academies report(NRC, 2005), the US Department of Defense (DoD) esti-mates that there are some �10,000 Hard and Deeply-Bur-ied Targets (HDBTs) worldwide that are essentiallyinvulnerable to conventional attack and can only be ‘heldat risk’ by a penetrating nuclear weapon. Penetration to�3 m enhances shock coupling strongly: penetration>3 m decreases probability of survival of weapon to deto-nation point. A ‘selling point’ offered in support of thesesystems is that they might produce less radioactive falloutthan a more conventional nuclear weapon.

Fig. 4. An F-15 fighter drops a GBU-28 penetrating bomb, developedduring the 1st Gulf war to attack deeply-buried targets.

2.4. Instrumented penetrators

WWII saw development of a variety of air-droppablesystems with instrumentation more elaborate than a simplemechanical time or impact fuze, notably mines for use atsea. There also exists at least one land-based example.The ‘Stabo’ (‘Stachebomben’) variant of the LuftwaffeSC-250 bomb (Fleischer, 2003) had a spike nose and avibration-sensitive fuse. This weapon was intended specifi-cally for attacking railways – the spike nose aided implan-tation in the ground and avoided bounce from the lowaltitudes from which precision drops had to occur. Useof a delayed-action vibration-sensitive fuse increased theprobability that the explosion might damage a train as wellas the railway line itself. While not quite a penetrator forseismic studies, the parallels are evident.

A related development in WWII that bears on instru-mented penetrators was the origination of expendableair-dropped packages for the remote detection of sub-marine acoustic emissions, i.e. sonobuoys. The conceptwas originated (Holler et al., 2006) by Blackett in theUK, but along with the transfer of other technology suchas radar, was pursued industrially in the USA. The firstsonobuoy was manufactured by RCA in 1941 and thefirst successful experiment was conducted in March1942 off New London, Connecticut: the propeller soundsof the submarine S-20 were detected at a range of threemiles by the K-5 blimp which could receive the radio sig-nals from the (ship-deployed) buoy up to five miles away.By July of that year the first air-dropped models weredeveloped and tested from B-18 bombers. Procurementof 1000 buoys and 100 radio receivers followed by theBureau of Ships that October and over 150,000 had beenordered by the end of the war. Sonobuoys continued toplay a significant role in the Cold War, and remain acentral part of the Anti-Submarine Warfare arsenal.While the sensors and penetration mechanics differ some-what, the functional requirements of instrumented pene-trators and sonobuoys are clearly rather similar andsonobuoy development likely aided the availability ofcomponents and methods for penetrators (notably radiotelemetry).

2.5. Igloo White

The largest and best-known (e.g. Correll, 2004: see alsoRosenau, 2001) instrumented penetrator project is ‘IglooWhite’, which saw the creation of a virtual barrier acrossthe Ho Chi Minh trail in 1968. This �$1.5B project duringthe Vietnam war air-dropped seismic, acoustic, magneticand chemical instrumentation (Fig. 5) around the trail inan attempt to cue strikes to interdict supply operations.

The Infiltration Surveillance Center at Nakhon Phanomin Thailand was the largest building in S. Asia and housedsome 400 Air Force personnel in high-security conditions.In addition it housed two IBM 360-65 computers, at thetime the most powerful computers available.

Fig. 5. A set of instrumented penetrators from the Igloo White project ofthe Vietnam era, to be dropped from aircraft to monitor the Ho Chi Minhtrail. Some are acoustic monitors, some seismic, some a combination.Note the large antennas, shaped to resemble tree branches.


Tens of thousands of sensor units were deployed, somewith acoustic sensors and others with chemical sensors(famously subject to spoofing) to detect sweat and urine.Some were designed to hang in tree foliage, whereas otherswere intended to implant themselves into the ground. Chiefamong these was the most widely used sensor platform, theADSID – Air-Delivered Seismic Intrusion Detector. Thispackage had a mass of 20 kg and was 1.3 m long. A similarsystem listened for acoustic emissions (e.g. Fig. 6). For allthese systems, battery life was several weeks, with databeing relayed to Nakhon by dedicated relay aircraft thatmaintained station above the area.

The ADSIDs were typically dropped in strings of �6 inthe expectation that �3 might work and thereby assurecoverage of a target site. While the actual statistics werebetter (about 80% of the systems returned data), the differ-ence in practical expectations of success rate of these sys-tems should be noted.

Although the target environment (jungle) is complicatedcompared with many planetary surfaces, the delivery prob-lem was somewhat straightforward in that the antenna

Fig. 6. Cross-section of an ACUSID (Acoustic Seismic Intrusion Detec-tor) from the Igloo White project.

acted as a drag brake to stabilize and slow the vehicle. Geo-phones could tolerate a 2000 g impact pulse (few ms long)from air-drop deployment, yet were sensitive enough todetect human footsteps �30 m away. It may be noted thatthis project saw the relatively early introduction of on-board data compression in a distributed sensing systemto reduce communications and processing bandwidthrequirements (e.g. the ADSID could be set not to transmitindication of a pedestrian ‘encounter’ until 32 sequentialfootsteps were detected.) This data volume considerationis common to many planetary penetrator systems.

2.6. Tabasco

In one of the most intriguing ‘untold stories’ of the ColdWar (only the most fragmentary accounts exist in the openliterature – the project has not been unclassified, even 40years later) is that of the Tabasco U-2 overflights of Chinain the late 1960s (e.g. Pocock, 2004). The intent was to dropa surveillance pod (with a spike and parachute) to monitorChinese nuclear tests in the western deserts near Lop Nor.These pods contained sensors to measure the air pressurepulse, seismic waves and light flash from a nuclear explo-sion – taken together they should indicate the distance tothe detonation, the yield, and whether the explosion wasan air burst or ground burst. The pods, 3.1 m long and130 kg in mass were designed and constructed by SandiaNational Labs, using information from the CIA on themost likely soil types in the TaklaMakan desert.

After release, the pods were to deploy a parachute (pre-sumably to assure vertical orientation) which was then tobe jetissonned at an altitude of �1000 ft – this would allowfree-fall to provide a desired impact velocity (presumablyneeded in order to achieve adequate seismic coupling) as wellas distancing the easy-to-spot parachute from the sensorpod. In case of discovery, they were marked with “China Sci-ence Institute: Do not touch” in Chinese characters.

Although one pair of pods was dropped successfully,notelemetry was ever recovered. One was battery-powered,the other solar powered. The pods were to be interrogatedby radio, for which they had a 3.1 m telescoping antenna.Nominally communications were to be by HF radio fromlistening stations in friendly territory, but when noresponse was received a risky 9-hour overflight with a U-2 equipped with a trailing wire antenna was attempted, inthe hope that a signal too weak to detect from a distancecould be recovered.

Faced with lack of success and an increasing surface-to-air missile threat to the U-2s, the monitoring mission wasinstead executed by other means. Nonetheless, the storyserves to set the paradigm of instrumented air-droppedpenetrating vehicles in this period.

2.7. Sea ice penetrometer

In the early 1970s, perhaps motivated to broaden theapplication of their penetrator technology, Sandia labs


explored the possibility of measuring the thickness of seaice using an air-dropped penetrator (unhappily for the dis-tinctions advocated in the present paper, they (Young andKeck, 1971) refer to a ‘sea ice penetrometer’ whereas wewould rather refer to it as a penetrator). Although thisapplication might have some geophysical relevance, mili-tary applications are obvious (e.g. establishment of icethickness for safe landing of aircraft, or for surfacing ofballistic missile submarines).

The approach was to modulate the output of an acceler-ometer onto a radio signal in real time – knowing theimpact velocity, one can integrate the equation of motionto determine the depth as a function of time, and the thick-ness of the ice is determined from the depth at which thedeceleration drops to zero (this approach relies, however,on complete perforation of the ice sheet).

An initial variant (e.g. Young, 1974) used a 42in. long50lb vehicle with a trailing wire antenna with a small dro-gue parachute at its end. However, this layout presentedsome difficulties. A second variant used a detachable aftbody which would be left behind at the surface on impact.The acceleration record in tests indeed appeared to be asuccessful indicator of ice strength and thickness, but thetechnique appears not to have been widely adopted.

2.8. Recent military systems

Technological developments which permit small systemsto perform advanced signal processing and classificationhave stimulated interest (e.g. Vick et al., 2001) in Unat-tended Ground Sensors (UGS), some of which can beair-deployed like the Igloo White systems, although moregenerally they may be manually installed for perimeterdefense or monitoring of roads. Unlike Igloo White andTabasco, modern systems can exploit wireless networking(allowing one sensor to cue others, for example) and satel-lite communications, without requiring dedicated and vul-nerable radio relay aircraft.

A Sandia National Laboratories concept circa 2003 wasnamed Steel Eagle, intended for air-drop from F-15, F-16and A-10 aircraft. This concept was extended by the AirForce into ARGUS (Advanced Remote Ground Unat-tended Sensor), whose minimum performance specifica-tions require finding, fixing, and tracking targets within500 m of the sensor and recognizing them within 200 m.The tested prototype design weighs 38 kg. For production,this weight was to be reduced by integrating advanced bat-teries and miniaturized electronics into a smaller body.Although systems similar to this are likely under develop-ment by a number of nations at present, for obvious rea-sons there are few details in the public domain.

2.9. Antarctic seismic penetrators

Geophysical studies of the Earth sometimes mention‘penetrators’, referring to devices to penetrate sea floor sed-iments to measure geothermal heat flow. However, the only

application of instrumented high-speed penetrators ofwhich this author is aware is the work in Japan on helicop-ter-dropped penetrators for making explosion seismicexperiments to study the crustal and mantle structureunder East Queen Maud Land, Antarctica (e.g. Shibuyaet al., 1991) with the intent of deploying an array of some40–50 expendable seismic stations at intervals of 5–10 km.Inert penetrators were dropped into glazed surface snow onthe Antarctic ice sheet, determining its penetrability. Animportant consideration in the interpretation of seismicmeasurements is the accuracy of location knowledge ofthe penetrators – differential GPS was used to record theposition of the deployment helicopter after it had des-cended to hover over the visually-identified impact point(Shibuya et al., 1993). GPS is useful also for assuring syn-chronization of clocks on the deployed penetrators so thatthe arrival times of seismic signals could be measured tobetter than 0.1 ms (previous Antarctic seismic studiesrequired the laborious carriage of 4 master clocks in vehi-cles across the snow to over 30 stations, Shibuya et al.,1991). After these initial developments, the full-scale imple-mentation of a penetrator-emplaced seismic survey appearsnot to have been realized, however.

3. Planetary penetrator missions

In this section, I discuss the programmatic history ofvarious penetrator missions. Likely the earliest consider-ations were given to penetration mechanics on the Moonas a means of determining whether the surface would sup-port a manned lander. However, in the end this questionwas resolved by landing dynamics and soil mechanicsinvestigations by the Surveyor landers rather than pene-trating vehicles. There have doubtless been many proposalsof varying levels of maturity – the discussion that follows isrestricted to missions or proposals that received significanteffort and are reasonably well-documented or are otherwisenoteworthy. A broadly chronological arrangement is fol-lowed in the discussion, although it must be recognizedthat parallel developments in various countries and thevarying fortunes of different missions mean the chronologyis not strict.

3.1. Early Mars proposals

The first detailed planetary penetrator vehicle discussionappears to be a Sandia Labs Mars Penetrator design in1974, perhaps stimulated in part by work in the 1960sand 1970s by that same institution on terrestrial penetra-tors including nuclear weapons (Sandia Laboratories oper-ate under contract with the US Department of Energy) andinstrumented penetrators for monitoring Chinese nucleartests. An early popular account (Anonymous, 1976) notesthat G. Simmons of Sandia had proposed in 1969 a free-falling penetrator to be released from �2000 m altitudeas an augmentation to the Viking mission to search forsubsurface water. This experiment was not considered


compatible with the mission and so was not selected forflight, but perhaps opened the door for subsequent studies:Ames Research Center began working in 1973 with Sandiaon Mars penetrator concepts (e.g. Lumpkin, 1974) A ratherdetailed analysis (Hughes, 1976) was also made of how aMars penetrator might be delivered by a spacecraft derivedfrom the Pioneer Venus Orbiter spacecraft (it will berecalled that Pioneer Venus was under development byHughes at this time, with the project being managed byNASA Ames.)

The Mars penetrator concept received further attentionafter the Viking mission, which while a scientific successoverall was disappointing for seismologists (the instru-ments were added as an afterthought and were very suscep-tible to wind noise: while genuine seismic events would beidentifiable by near-simultaneous detection on the twolanders, the failure of the seismometer on Viking 1 to unc-age prevented this approach). Furthermore, the inability ofthe Viking soil sampling arm to obtain material from below�10 cm was a recognized limitation.

Thus, penetrators seemed a logical next step, and wereadvocated by Sandia (e.g. Simmons, 1997). The scientificscope of a penetrator mission was considered by a NationalResearch Council body, the Ad Hoc Surface PenetrationScience Committee chaired by James Westphal (NRC,1976) which recommended

“We therefore recommend that a minimum viable mission

must consist of at least four penetrators, and that each of

these penetrators must carry a seismometer, and afterbody

imager, and at least one of the following additional experi-ments: (a) chemical composition (b) total water measure-

ment(c) heat flow (d) afterbody meteorology. In our

opinion, with reasonable effort it will be possible to fly all

these experiments plus a few others.”It was anticipated in this report that a penetrator mis-

sion might be developed in time for the 1981 Mars oppor-tunity, and considerable (now somewhat forgotten) effortwas devoted to developing engineering designs for such amission and to investigating the scientific capabilities ofinstrumentation emplaced this way.

A 1977 NASA Ames report (Manning, 1977) describes apenetrator system (Fig. 7), then conceived as a point designfor the 1984 Mars opportunity, complementary to an orbi-ter and rover. Six penetrators, each of 38 kg plus a 16 kgentry system, 8 kg launch tube and 7 kg rocket motor,would be independently launched from the carrier space-craft during hyperbolic approach to Mars. The motor(from the TOW missile) would accelerate the penetratorto 60 m/s within the launch tube, thereby minimizingplume impingement on the orbiter. This launch DV alloweda variety of selected landing sites to be targeted with a100 � 80 km footprint.

Each penetrator would deploy a hypersonic decelerator,imagined to resemble an umbrella, to slow from arrivalspeed – this would then be separated and a small ribbeddrogue would assure the terminal velocity at impact wouldbe in the range 135–165 m/s, and 10 cm diameter, 1.4 m

long penetrator would see an impact load of 1500–2000 g. It should be noted that the details of stability andaerothermodynamics during entry do not seem to havereceived much attention.

A �2.5 kg payload was anticipated (Fig. 8), including asoil moisture detector, an Alpha-Proton-X-ray (APX)experiment, a seismometer, heat flow experiment, magne-tometer and meteorology experiment. A drill system wouldbe used to acquire samples. The heat flow experiment wasanticipated to use a series of thermocouples on the umbil-ical between forebody and afterbody, the latter carrying themagnetometer, antenna and meteorology experiments. A400 MHz UHF radio link at 1 W RF power would trans-mit data to the orbiter at 2500 bps for a �10 min pass(1.5 Mbit) per day.

Power was to be provided by a 20 W-hr primary battery,plus NiCd secondary batteries trickle-charged by a480 mW RTG. The primary battery was principally to per-form the soil sampling (drill, APX, moisture) experimentsand seismometer leveling.

The mission failed to materialize – not only did the late1970s and early 1980s see almost no new starts in planetaryscience overall (e.g. Ulivi and Harland, 2009), but smallRTGs were not developed. A 1981 report (Murphy et al.,1981) summarizes much of the design considerations fromthe 1977 studies, together with references to various scien-tific (heat flow, sample modification, seismic coupling) andengineering field tests. In fact a rather large number ofextensive field tests were performed on penetrator emplace-ment (Figs. 9 and 10), as well as tests on scientific aspects ofpenetration, such as the effects of frictional heating on tar-get material.

3.2. Galilean satellites and Mercury

Essentially the same basic Mars penetrator vehicle wasconsidered (Friedlander et al., 1976) for application atthe Galilean satellites, although delivery to the surfacerequired rather more elaborate propulsion and attitudecontrol, since the Galilean satellites lack atmospheres.Consideration was given to two basic delivery strategies –a more or less rectilinear trajectory from a hyperbolic tra-jectory with respect to the relevant satellite, and deliveryfrom an elliptical orbit around it. The study correctly notedthat the guidance and navigation problems weresignificant.

A more or less parallel study considered Mercury pene-trators, where the delivery problem was essentially thesame. (Both the Mercury and Galilean studies were con-ducted by the same team at Science Applications Interna-tional Corporation).

3.3. Titan

Penetrators were similarly considered for Titan, whichwas known at the time (1976) to have a methane-bearingatmosphere, but it was not known how thick this atmo-

Fig. 7. NASA 1970’s Mars penetrator mission concept. The carrier spacecraft would launch the penetrator by rocket from a tube. An umbrella-likedeployable fabric decelerator would be used to slow and stabilize the penetrator, which would leave an aftbody antenna at the surface.

Fig. 8. Cross- Section of the 1970s NASA Mars Penetrator concept. Note that the RTG was anticipated in the forebody.


sphere was (i.e. what was the surface pressure). Thus mis-sion concepts had to be adaptable to thick (a few bar)and thin (perhaps even Mars-like of �10 mbar) atmo-spheres. Against this backdrop of ignorance, a conceptnamed the ‘penetrobe’ was suggested – a combination ofatmospheric probe and penetrator. This used anumbrella-like decelerator, much like some of the Mars pen-etrator variants, but the decelerator would be collapsed or

jetissoned if the entry and descent profile warranted, e.g. ifthe atmosphere were to prove thick.

Once the Voyager 1 flyby of Titan in 1980 determinedthe atmospheric conditions at the surface, studies tendedto focus on probes and balloons (e.g. Lorenz, 2009),although penetrators were mentioned as possible elementsto be deployed from an airship. However, no detaileddesign of these penetrators was performed.

Fig. 9. The 1970s Mars penetrator effort saw extensive field testing atseveral locations in the USA. In these tests at Tonopah, New Mexico, thetunnel carved by the penetrator during impact was exposed by boring awell. Emplacement geometry, and material modification during entry wasstudied.

Fig. 10. Inspections such as that in Fig. 9 allowed the effect of rockobstacles on penetration performance to be determined.


Interestingly, the 1976 NRC report seems to have con-sidered penetrators unviable at Venus owing to the thick,hot atmosphere which would transfer heat rapidly to aslender vehicle such as a penetrator. In fact the same prob-lem would apply for Titan (although there the heat transfer

would be from the vehicle rather than to it: heating onTitan to compensate would be somewhat easier to imple-ment than the cooling that would be required on Venus,however). On both bodies, the atmospheric density is suchas to make terminal velocity rather lower than is typical forpenetrators, and the velocity might need to be augmentedby a rocket motor, much as runway-cratering weaponssuch as the Durandal. These thermal and delivery compli-cations make Titan and Venus unappealing applications ofpenetrator vehicles, despite occasional claims of their suit-ability (e.g. Collinson, 2008)

3.4. Comet nucleus penetrator/CRAF

In the 1980s, which saw overall relatively little in-situplanetary mission development overall, penetrators simi-larly seem to have lapsed. They returned to considerationwith the development in the mid-late 1980s of a comet pen-etrator, which was seen as an element of a Comet Rendez-vous/Asteroid Flyby (CRAF) mission. CRAF was to flypast two asteroids before inserting into orbit around orat least alongside a comet for close remote sensing andin-situ measurements of the cometary environment (dust,gas and plasma). In addition, it was desired to havein situ measurements from the surface of the cometnucleus.

Any in situ interaction with a comet suffers serious andpoorly-understood challenges of low-gravity, unknownsurface roughness, texture, outgassing from jets and soon which makes the design and operation of soft landersvery challenging. One approach that mitigates these effectssomewhat (although far from completely) is to use a pene-trator that would be anchored in the surface by virtue of itshigh-speed impact (Fig. 11).

The concept overall is described by Swenson et al.(1987), drawing in part on an industrial study (Adamset al., 1986). This latter 240-page document is remarkablydetailed – including extensive modeling of the delivery atti-tude kinematics and even a component-level parts list: theconfiguration designed at that time is shown in Fig. 12. Theinitial concept of the penetrator was principally to carry aGamma Ray Spectrometer to determine the elementalcomposition of cometary material (Evans et al., 1986). Thiswould be complemented by an Alpha-Proton-X-ray spec-trometer (which, among other elements, would determinethe abundance of Titanium and Aluminium, whoseGamma Ray signature would be masked by the penetratorstructure itself). Accelerometers and temperature sensorswould contribute to understanding of the physical proper-ties of the nucleus.

Because it would not be desired to place the ‘mothership’ on a high-speed impact trajectory (which would allowa simple penetrator design, with simple release by a spring,but would require a rushed avoidance manoeuvre and com-plicate the relay geometry) the penetrator was to be accel-erated towards the comet by its own rocket motor. Sincethe low-gravity cometary surface could in principle have

Fig. 11. Concept of the Comet Rendezvous Asteroid Flyby (CRAF)mission, with remote investigations on a large orbiter complemented by asmall set of in situ measurements made by a penetrator.


a very low strength, the penetrator had a prominent aftflare to prevent over-penetration. It is reported (Boynton,personal communication) that some penetration and sam-pling tests were made by dropping penetrator models fromthe Arizona Wildcats Football stadium on the Universityof Arizona campus (close to the Lunar and Planetary Lab-oratory building).

The concept in 1986 (Swenson et al. (1987)) was for a1 m-long 20.35 kg vehicle, with a 7 kg support module onthe CRAF orbiter including the relay receiver and a spin-eject mechanism. The mechanism would provide a spin rateof 116 rpm (rather high compared with most spin-stabilizedprobes) and a linear velocity of 0.7 m/s (which is quite typ-ical). The high spin rate was to minimize the nutation anglefrom any tip-off errors during the release event, since these

Fig. 12. Original, detailed design of the CRAF penetrator, from

would translate into delivery errors and ultimately mightcause an unacceptable angle-of-attack at impact. At a dis-tance of �1 km from the comet nucleus, a rocket motorwould accelerate the vehicle to a speed of 40 m/s.

The gamma ray instrument required a minimum of 3,and nominally 5, days of operation. Significantly, its detec-tor had to be maintained below 120 K: this was planned tobe achieved by pre-cooling the detector before release fromthe orbiter spacecraft and thermally isolating the detectorfrom the penetrator body (although it might be noted thatthe ambient temperature was expected to be lower than160 K).

The mission was expected to generate 170 kbits of rawdata over the 5 days, to be sent at intervals on a 1.5 GHzlink with radiated power of 200 mW at 1 kbit/s, even allow-ing for 1m of cometary material burying the antenna. Itmay be noted that the telemetry opportunities at a cometcannot be reliably anticipated before arrival – since thecomet nucleus rotation will likely not be known before arri-val at the comet, nor can the motion of the penetratorantenna boresight across the celestial sphere. Similarly,the comet nucleus mass will not be known a priori, sothe orbital period for a given distance of the relay space-craft cannot be anticipated. These and other aspects ofthe delivery and relay geometry are discussed by Stetson(1988).

The CRAF mission evolved in the late 1980s into thefirst of the Mariner Mark II program, which was to use amodular multimission bus for a variety of outer solar sys-tem missions. Two initial missions were selected, Cassini–Huygens and CRAF. By using a common bus (structure,radioisotope power, propulsion systems etc.) it was claimedthat the recurring costs of missions could be reduced. Each

a 1986 Martin Marietta NASA study (Adams et al., 1986).


mission had instruments and elements adapted to its partic-ulars (e.g. Cassini would be much further from Earth, andso had a larger high-gain antenna than CRAF). CRAF wasseen to be an important precursor to a future Comet Sam-ple Return (at the time, this mission was named Rosettaand was to be a joint NASA-ESA endeavour).

The penetrator for CRAF was competed via anannouncement of opportunity (AO) released by NASA injust the same way as instruments for the CRAF spacecraft.Two proposals were submitted, and that led by W. Boyn-ton of the University of Arizona was selected (Boynton,personal communication, 2011). The scientific scope wasexpanded beyond the 1986 CNP concept to include mea-suring the volatile composition of the nucleus, and thedesign evolved (Fig. 13), to include more extensive instru-

Fig. 13. A later (circa 1990) concept of the CRAF penetrator, with moreelaborate instrumentation and a more modest flare.

mentation. An important new feature of the CRAF pene-trator was a hook-shaped sample acquisition scoop onthe side of the vehicle (see also Bickler et al., 1997 for dis-cussion of penetrator sample acquisition). It was hopedthat this scoop would permit the ingestion of a sample ofsubsurface material from the comet (and indeed, testsinvolving the drop of mockup penetrators from the largefootball stadium at the University of Arizona suggested ithad a good probability of doing so, Boynton, personalcommunication, 2011), which would then be examined bya DSC/EGA instrument (Differential Scanning Calorime-ter/Evolved Gas Analyzer). Programmed heating of thesample (e.g. Boynton et al., 1997) allows the detection ofphase changes such as those associated with ices, and detec-tion of gases (e.g. evolved from clathrate ices) would beimportant in understanding the volatile content and originof cometary material. A similar, but larger, instrument wasflown by this same Arizona group on the ill-fated MarsPolar Lander in 1999, and then the Phoenix lander in 2007.

The Committee on Planetary and Lunar Exploration(COMPLEX) of the Space Studies Board (SSB) of theNational Academy of Sciences was requested by NASAin 1990 to evaluate various CRAF descope options. Theirreport (COMPLEX, 1990) considered the deletion of thepenetrator to have a serious science impact, noting theirearlier evaluation of “the unique ability of the penetratorinstruments to address in situ the highest priority scienceobjective ... namely, determination of the dust and volatilecomposition, state, and physical properties of the [comet]nucleus”, and proposed instead a broader range of payloadcuts, including capping the cost of the penetrator.

An interesting remark in the report is that “The Com-mittee received no information concerning either the man-agement or technical risk of the penetrator, but theprincipal investigator’s statement that it is not inherentlymore risky than several of the other experiments was notcontested by the JPL or NASA representatives at theCOMPLEX review.”

The total FY91-FY95 cost of the penetrator investiga-tion, in real year numbers in appendices to the report,was $66.49M, of which science coordination, engineeringoversight and mission design contributed about $1.5Mand the thermal shield and cover (which COMPLEX pro-posed deleting) was $2M. Thus the penetrator representedabout half of the total ($137M) science cost on CRAF.Examining the estimated cost savings anticipated bydescoping various instruments on the penetrator suggeststhat the payload costs on the penetrator were of the orderof $14M.

Unfortunately, faced with budgetary pressure on theMMII program, the comet penetrator was the first elementto be descoped. Subsequent more severe pressures in theFY92 budget led to cancellation of the CRAF missionentirely, and significant descopes to the Cassini missionwhich became the only flight of the MMII ‘series’.

Comet penetrator missions (e.g. Boynton and Reinert,1995), with simple carrier/relay spacecraft rather than the


elaborately-instrumented CRAF orbiter, have been subse-quently proposed to the NASA Discovery program(wherein Principal-Investigator-led concepts could beentertained) but so far without success.

The US withdrew from the Rosetta Comet NucleusSample Return mission, which was then pursued by ESAalone as a Comet Rendezvous mission, retaining the samename. Initially in this rescoped ESA mission, 2 small land-ers (Champollion and Roland) were anticipated, someinstrumentation on which drew on CRAF work; these wereeventually consolidated into one lander now named Philae.Although not penetrators by the definition in this paper,Philae does carry a self-hammering thermal probe, as wellas anchoring harpoons instrumented with thermometers(Paton et al., 2010) and accelerometers (e.g. Kargl et al.,2001; Komle et al., 1997), which have some methodologyoverlap with true penetrators. The Champollion namewas retained by a mission (Space Technology-4 or ST-4,in the NASA New Millennium technology validation pro-gram – see Section 3.9) that did consider comet anchoringand penetration – e.g. Steltzner and Nasif, 2000 – before ittoo was cancelled. Rosetta and Philae were launched suc-cessfully in 2003 and a rendezvous with a comet is antici-pated in 2015.

3.5. VESTA

Around the same time as the buildup to CRAF, theEuropean Space Agency was considering mission candi-dates for M2, the second ‘medium class’ mission in its Cos-mic Vision program. One candidate was named Vesta, andwas a joint project of the Russian Space Research InstituteIKI, the European Space Agency, and the French SpaceResearch Agency CNES, following earlier bilateralFrench/Soviet study of a multiple flyby mission. The

Fig. 14. Concept of the VESTA mission. A subspacecraft (‘decelerator modurelease them. The mission would be limited by the duration of the communica

mission (ESA, 1988), to be launched in 1994 on a RussianProton launcher, would fly past Mars, and an asteroid enroute to the large asteroid 4 Vesta.

During the asteroid flyby, the mission would release twopenetrators. In contrast to the CRAF mission at a smallbody where a motor would accelerate the penetrator intothe comet, the mission profile for VESTA saw an asteroidflyby velocity of some 2.2 km/s, and thus the penetratorswere carried (Fig. 14) in a separate 500 kg ‘decelerationmodule’ which would use bipropellant propulsion toinstead reduce the impact velocity to an acceptable level(� 100 m/s). The 25 kg penetrators would communicatewith the mother spacecraft (which would fly past the aster-oid at a distance of �500 km) with an L-band radio link at64 bps over a mission duration nominally of 1 h, althoughthe visibility window would depend on the rotation periodof the asteroid chosen (nominally 7 Iris or 46 Hestia).

Unusually, the two penetrators were allocated differentinstrument payloads, each totaling just under 4 kg. Bothpenetrators would carry an accelerometer/gravimeter, ther-mal array probe, alpha- and X-ray spectrometers, and agamma ray spectrometer. One would also carry a permittiv-ity meter, neutron detector and television camera, while theother would carry a seismometer and gas chromatograph.

In 1989 ESA selected the Huygens mission as the M2mission, and work on Vesta ceased, although Russia didconsider a variant ‘Mars-Aster’ in the mid-1990s (Uliviand Harland, 2009).

3.6. Mars-96

The Mars-96 mission (e.g. Surkov and Kremnev, 1998,see also Surkov, 1997) was a large orbiter vehicle, some-what derived from the Phobos mission of circa 1988. (Infact what flew as Mars-96 actually began development as

le’) would slow the penetrators to an acceptable impact speed, target andtions geometry with the main spacecraft. Credit: ESA.


Mars-94, with the original Mars-96 to have carried a bal-loon. Economic difficulties led to the cancellation of thatmission, and the delay of Mars-94 by two years). Thespacecraft, small stations and penetrators were developedand manufactured at the Lavochkin Research and Opera-tional Association (NPO). The small stations were designedin cooperation with the Space Research Institute (IKI), thepenetrators in cooperation with the Vernadsky Institute(GEOHI). The mission included scientific contributionsfrom no less than 20 countries (Harvey, 2007).

The mission featured two small stations (semi-hardlanders, with self-righting by petal opening, somewhatresembling Luna-13), two penetrators and an orbiter.While the surface stations were delivered from the hyper-bolic arrival trajectory, the penetrators were to be releasedfrom orbit (Fig. 15) around Mars. The orbiter wouldrelease each penetrator in turn which would fire its own ret-romotor to provide a 30 m/s deorbit burn. The 1.7 m-longpenetrator vehicle would then inflate a 3.6 m diameterinflatable decelerator (ballute) to enter the Mars atmo-sphere �20 h later and assure correct orientation at impact.This decelerator (Fig. 16) provided the required drag areato slow from the hypersonic entry to a tolerable impactspeed.

The forebody was 14 cm in diameter, while the aftbodywas a 17 cm cylinder mated to a cone with a base diameterof 80 cm (Fig. 17). The principal structural material wastitanium (Ulivi and Harland, 2009). The estimated impact

Fig. 15. Delivery geometry of the Mars-96 penetrators, on the first post-inindividually by the orbiter, and then fire small motors to achieve a shallow en

speed would be 60–80 m/s, leading to a penetration depthof up to �5 m. Deceleration loads on the aftbody wouldbe limited to 500 g by a shock attenuation system. Penetra-tion tests including drops from helicopters, as well as drop-ping 60 m down the lift shaft of the Moscow AviationInstitute (Harvey, 2007).

The penetrators (120 kg at release, of which 8.8 kg wasscience payload: though note that there appears to be someconfusion in the literature over the breakdown of the massbetween the two penetrators and their delivery systems)had relatively large instrument complements. Theyincluded a camera system, magnetometer, gamma ray spec-trometer and meteorology package on the aftbody, as wellas a seismometer, and alpha, neutron and X-ray spectrom-eters for mineralogical studies and to detect water or ice.

Power was to be supplied for the year-long planned life-time by a small (0. 5W electric) radioisotope thermoelectricgenerator in the aftbody, supplemented for the initial mea-surements by a 150 W-hr Lithium primary battery. Com-munications would be at 8 kb/s by a UHF relay to Mars-96 or the Mars Global Surveyor (on which a relay receiverhad been installed to support Russian in-situ vehicles atMars: this ‘Mars Balloon Relay’ would later be used forthe DS-2 penetrator mission, see next section), with over-flight communications sessions lasting �6 min (Surkovand Kremnev, 1998).

The Mars-96 mission was launched on a Proton boosteron November 16, 1996, but a failure of the Block D-2

sertion orbit around Mars. The penetrators are spun up and releasedtry into the Martian atmosphere while the orbiter flies overhead.

Fig. 16. Concept of the entry and descent of the Mars-96 penetrators. Asolid thermal protection material insulates the vehicle from the heat ofentry, while an inflatable decelerator reduces the velocity at impact to�80 m/s.

Fig. 17. Cross-section of the deployed Mars-96 penetrator, showing theextensive instrumentation and the split into fore- and aft-bodies connectedby umbilical. Note the RTG is in the aftbody where it can effectively rejectheat.


fourth stage prevented departure from Earth and so themission was lost, a brutal blow which effectively terminatedthe Russian planetary program for the next decade and ahalf (though many European payloads developed for thePhobos and Mars-96 missions were subsequently reflownon the ESA Mars Express mission). The Mars-96 space-craft assembly re-entered over South America, althoughno parts (including the radioisotope generators) are knownto have been recovered.

3.7. MARSNET

MARSNET was an ESA study of a potential networkmission to Mars (Chicarro et al., 1991). This envisaged 3or 4 penetrators, each of �65 kg, to perform geochemical,seismic and meteorological studies. The penetrator archi-tecture was considered in parallel with a ‘semi-hard lander’type, which the trade study favored, but the study is notedhere as having a few distinctive features. First, the penetra-tor would make its terminal descent under a parachute,with a terminal velocity of about 80 m/s. Second, the aftpart of the penetrator - which would remain fixed to thecylinder/cone part – had a relatively large diameter in orderto accommodate a conical solar array (Fig. 18) for powergeneration (radioisotope power sources being unavailablefor ESA, although recall Mars 94 penetrators were underdevelopment at this time). Third, a cylindrical impactattenuator made from a crushable material (Fig. 19) suchas aluminum honeycomb was mounted around the

penetrator body ‘to limit loads in soft material’. The logichere appears to be that in a hard target, the penetratorwould enter some distance and the lander would be rigidlyattached, with some small deformation of the crushablematerial. In a softer target, the penetrator alone wouldenter too deeply, but if the aft structure impacted the sur-face without the forebody slowing the assembly down,the loads on the aft structure would be too large – hencethe crushable material to limit these loads. These differentlander concepts were studied by industry under the ESATechnology development program, including finite-elementsimulations of impact kinematics (e.g. Steckemetz et al.,1993) and testing of impact absorption properties of hon-eycomb materials (e.g. Doengi et al., 1988).

Despite these studies, residual concerns over the robust-ness of the landing approach remained, and a more con-ventional semi-hard lander concept was favored in afollow-on study (‘INTERMARSNET’) and ultimately inthe Beagle-2 mission which saw flight with an airbagimpact-attenuation system.

Fig. 18. Concept of the ESA Marsnet hard lander on the Martian surface.Note the large conical solar array. Credit: ESA


3.8. Lunar-A

The Japanese Lunar-A mission was a focused geophys-ical mission to the moon (Fig. 20), to use penetrators todeploy seismometers and make a heat flow measurementon the moon: at least one penetrator would be deployedto the far side of the moon, which has a significantly differ-ent crustal structure from the nearside where the Apollomissions made geophysics measurements. The missionwas initially developed by the Japanese scientific spaceagency ISAS (Institute of Space and Astronautical Sci-ence), which was subsequently assimilated into the largeragency JAXA (Japanese Aerospace EXploration Agency)when that was formed in 2003.

After initial studies (e.g. Mizutani et al., 1990, 1995) thedesign of a prototype model began in 1993, with manufac-turing of the flight units to follow in time for environmentaltests in April 1996 for an anticipated launch in 1997 (Nak-ajima et al., 1996) with three penetrators. Early workemphasized shock tolerance of components (it was noted– Hayashi et al., 1993 – that plastic-encapsulated commer-cial components were often more impact-tolerant thanceramic mil-spec parts), use of low-voltage electronics(while 3.3 V logic is now common, in the early 1990s 5 Vwas almost always used) and encapsulation in epoxy resinfilled with glass microballoons to reduce density.

Originally, communication via a telescopic VHFantenna deployed after penetration was planned, but afterimpact tests showed reliability problems with deployment

in sand (Nakajima et al., 1996) a fixed UHF system wasto be used. Another design feature of note is that the thrustprofile of the solid de-orbit motor was tuned to be triangu-lar instead of flat since the spin dynamics of the latter(more common) profile would lead to amplified mis-align-ment of the thrust vector during the burn.

Lacking an atmosphere to provide deceleration, theimpact speed would be basically as high as tolerable (tominimize the demands on propulsion) and the impact speedwas specified to be at a speed V < 300 m/s, which with thisvehicle configuration and mass could lead to 10,000 gimpact loads. The allowable angle of attack (AoA, or a)is a crucial parameter which appears to have undergonesome evolution – Mizutani et al., 1990 suggests penetrationis possible up to a �20�, yet Hinada et al., 1993 give a limitof only 60�. An extensive series of scale-model tests (Shirai-shi et al., 2008) using a 155 mm gun provide some explana-tion – the resting attitude of the penetrator is a strongfunction of impact AoA. Specifically, if AoA exceeds 10�,the final rest attitude of the penetrator will exceed 45degrees from vertical, which will degrade the heat flowmeasurement (which relies on measuring the vertical tem-perature gradient).

The energy source for the mission was to be LithiumThionyl Chloride batteries with an energy density of430 Whr/kg, and the nominal duration was to be 1 year.Communications would take place at approximately twoweek intervals, during overflights of the carrier spacecraft:essentially, the orbit plane of the carrier remains fixed,while the moon rotates underneath, bringing the penetratorsite under the orbit track twice per lunar day.

The attitude dynamics of the penetrator is of course keyto its successful penetration (Fig. 21). This was recognizedvery early (e.g. Morita et al., 1992.) Challenges includeknowing and controlling the attitude during the de-orbitburn, as well as precessing the long (spin) axis around tothe vertical during the relatively short free-fall time (theso-called ‘rhumb line’ manoeuvre). And these manoeuvresmust be conducted with a slender vehicle, which thereforehas its smallest principal moment of inertia around its longaxis (and hence is not stable in the long term).

The initial configuration (e.g. Morita et al., 1992) mini-mized these attitude challenges at the cost of requiring hea-vier propulsion, and was to use a de-orbit motor for eachpenetrator (released from at the 10 Rm apoapsis of a highlyelliptical orbit) to put it on an impact trajectory, and a sep-arate retromotor would brake the penetrator and be jetis-onned just before impact. The back-to-back motorassembly was strapped like a backpack onto the penetra-tor, and had masses on two arms to create an assemblywith a stable moment of inertia.

Later the configuration evolved (e.g. Morita et al., 1997)to a rather simpler system, wherein the penetrators wouldbe released at a much lower altitude (25 km) from which,after a motor burn to null the orbital velocity, it wouldfree-fall to the lunar surface. From 25 km the impact veloc-ity was low enough to permit survival of impact. While this

Fig. 19. Anatomy of the Marsnet hard lander – the penetrator part is surrounded by a large honeycomb impact attenuator to limit loads on the uppercone. Credit: ESA

Fig. 20. Artists impression of the Japanese Lunar-A mission: the carrierspacecraft in lunar orbit has just released a penetrator. Photo: JAXA.


required only one motor (which was now mounted at therear of the penetrator itself, Fig. 22), it required more atti-tude agility. The spin rate would need to be 120 rpm (2 Hz)to achieve adequate stability during the motor burn. The

system would then use a sun sensor and a cold gas jet toprecess the long axis to the vertical, and to actively sup-press nutation. The system furthermore needs to autono-mously adapt to its performance (specifically the delaybetween commanding the cold gas valve to open and theimpulse being achieved). The system was therefore testedin a balloon drop test, and then in a sounding rocket flightin 1997, an earlier sounding rocket test in 1995 havingfailed (Harvey, 2000.)

A number of electronics failures occurred during thepenetrator development, simple potting with epoxy loadedwith glass microspheres being inadequate to impact-hardensome components. In part these issues are attributed to theside-loads being larger than anticipated – indeed, when theangle of attack at a 305 m/s impact was 11�, the side-loadspeaked at 5000–6000 g, while the axial loads only reached4000 g (Shiraishi et al., 2008). Unfortunately logistical con-siderations (tests used Sandia labs airgun facilities in NewMexico, USA) that forced tests to be scheduled 10 monthsapart (Normile, 2007), and as a result the inevitable

Fig. 21. The Lunar-A delivery concept. The free-fall phase after the motorburn lasts only a few tens of seconds, during which time the vehicle mustbe precessed around to the vertical. Note that the whole vehicle penetratesthe surface (no aftbody), and communicates through the lunar regolith.

Fig. 23. Constructed Lunar-A penetrator, ready for flight. Photo: JAXA.


trial-and-error in design and test was very slow. Otherdelays and development difficulties (which led to thedescoping from three to two penetrators) included

Fig. 22. Cross-section of the Lunar-A penetr

concerns in a 1997 design review, poor illumination andpower geometry for a 1999 launch (Harvey, 2000).

ISAS/JAXA doggedly pursued development of instru-mentation, delivery system and penetrator vehicles andthe mission was close to flight (Fig. 23) in 2004 (e.g. Mizu-tani et al., 2003 – indeed NASA issued a proposal call for asmall number of US scientists to participate in the mission,to which the present author responded), but the valves inthe propulsion system had to be replaced (Mizutaniet al., 2005; Shiraishi et al., 2008) following a recall bythe manufacturer after similar valves had malfunctionedon another project. Furthermore, the robustness of thecommunication system was called out as an issue of con-cern in a launch readiness review (citing Beagle 2 andDS-2 experience – see next section) which necessitated fur-ther impact tests. At that point, a delay of two years ormore was anticipated, with possible launch in 2007.

In this period, a mechanical switch system was devel-oped to disconnect the vehicle Central Processing Unit(CPU) from its power, in order to reset it after impact (Shi-raishi et al., 2008). This followed a failure to communicate

ator vehicle and attitude control system.

Fig. 25. Artist’s impression of the delivery of DS-2: after release from theMars Polar Lander, the vehicle simply enters the atmosphere anddecelerates, progressively turning towards the vertical at impact.


with the CPU after an impact test in 2003. Although thecause of this was not identified, ESD (Electrostatic Dis-charge) was noted as a plausible scenario, and ESD eventsoriginating from triboelectric charging during the penetra-tion event had been documented in impact tests in supportof the DS-2 mission (Lorenz and Shandera, 2002). In addi-tion to the reset switch, assembly procedures were reviewedand circuit board layouts were redesigned to reduce electri-cal noise from the data processing unit from coupling intothe radio frequency (RF) systems, and thereby improve thecommunication system signal-to-noise.

Unfortunately, despite these efforts, and an investmentof some $132 million, JAXA formally cancelled theLunar-A project on January 15, 2007 (Normile, 2007). Itseems that in principle all the development difficulties andperformance uncertainties had finally been addressed toa large extent, so it is somewhat heartbreaking that themission was not launched after all. A contributing factorto this decision was the deterioration of the carrier space-craft (Fig. 24 – even though it had been in storage in drynitrogen after the propulsion valve replacement). Elementsof the launch and delivery system at this point exceededtheir qualification lifetime (recall launch on an all-solidvehicle with a custom upper stage was originally foreseenin 1995), and indeed the originator of the mission (H.Mizutani) had retired in 2005. Concerns remained overthe robustness of the communications system, although afinal end-to-end penetration test was planned in late 2007.

3.9. DS-2

The New Millennium Deep Space 2 mission (aka ‘MarsMicroprobes’) is the only penetrator mission to date tohave at least reached its destination. The concept, discussed

Fig. 24. Lunar-A Penetrators mounted on the carrier spacecraft. Photo:JAXA.

by Keese and Lundgren (1994), entailed integrating a pen-etrator vehicle within a frangible heat shield, such that thepenetrator would punch through it on impact. This archi-tecture (Fig. 25) neatly avoids the complications of para-chute deployment and heat shield separation andprovides passive aerodynamic deceleration and orientationfor acceptable impact velocity and attitude. The mission’searly development – as part of NASA’s New Millenniumtechnology development and validation program – isdescribed in Gavit and Powell (1996). In addition to dem-onstrating the novel entry, descent and impact architecture,the mission was to validate other technologies includinglow-temperature, impact-tolerant lithium batteries andhighly-integrated microelectronics. These included a micro-controller, mixed analog-digital ASICs (Application-Spe-cific Integrated Circuits), a programmable radiotransceiver (‘telecom on a chip’), power-switching electron-ics and science instruments: these technologies were all seenas enabling future network science missions.

In fact, some early considerations were given towards amore geophysically-oriented mission, with inclusion of aseismometer payload, towards low latitudes. However,when the opportunity emerged for delivery to Mars as asecondary payload along with the Mars Polar Lander, acombination of factors led towards a focus on water detec-tion. First, the polar layered terrain, where the penetratorswould impact, was believed to have a substantial water(ice) content. Second, the low temperatures of the polarregions and the low elevation of the sun, compromise thelongevity of a penetrator without radioisotope heating/power. Thus the principal science objective (and scienceobjectives were explicitly secondary to the technology dem-onstration goals of successful delivery and operation)became the search for subsurface ice on Mars, which hadat the time been hinted at from geomorphological andother studies.

Fig. 27. The truck-mounted airgun in New Mexico used for DS-2penetration tests. Over 60 shots were performed: the one shown here wasinto a chilled ice target.


Initially, a 1.2 kg landed mass (300 g aftbody and 900 gforebody, connected by an umbilical) was anticipated(Gavit and Powell, 1996) but the configuration evolved,in particular to accommodate the batteries in the aftbody,such that the flight configuration had a 670 g forebody anda 1740 g aftbody. This fore/aftbody mass ratio rather com-plicated the kinematics of separation at impact (see e.g.Smrekar et al., 2001; Lorenz et al., 2000). Considerabledesign iteration took place (Stone and Rivellini, 1998) todetermine the aftbody configuration that would preventrolling or skipping in a range of targets for a range of angleof attack. This configuration evolution (Fig. 26), as well asimpact qualification of components, necessitated a pro-gram of some 63 impact tests using a Sandia-operated air-gun (Fig. 27) at the Energetic Materials Research andTesting Center (EMRTC) operated by New Mexico Techat Socorro, New Mexico. However, unlike for Lunar-A,the implementing organization (the Jet Propulsion Labora-tory in Pasadena, CA) and the test facility were somewhatclose (a couple of hours flying time) which permitted a rel-atively rapid turnaround to permit iterative design and test– at the peak of the test program some 20 tests were per-formed over 100 days.

The small size of the DS-2 spacecraft necessitated closecoupling of the aerodynamic and terradynamic (impact)aspects of the mission. Given a fixed entry body size (deter-mined early in the project as part of the interface with theMars Polar Lander mission) the mass of the penetratorwould determine the overall ballistic coefficient (mass perunit drag area) and thus the impact speed and, since thevehicle would still be decelerating from its hypersonic arri-val and would not reach terminal velocity, the flight pathangle (the angle between the velocity vector and the verti-

Fig. 26. Artist’s impression of the DS-2 penetrator deployed on thesurface. The forebody and aftbody are connected with a flexible tapeumbilical.

cal, also termed ‘angle of incidence’). Monte Carlo simula-tions were performed (e.g. Braun et al., 1999) to evaluatethe probability, given specified variations in Mars atmo-spheric properties, for example, of staying within specified

Fig. 28. The DS-2 penetrator had a fairly squat aftbody, in which theforebody was held by shear pins. The penetrator vehicle occupied the frontinterior of a larger frangible entry shell, which shattered on impact.Maintaining a forward center of gravity for entry and descent stability wasa key design driver.

Fig. 29. Main part of the DS-2 forebody, with case removed – 40 mm indiameter. Circuit boards containing an advanced microcontroller andpower electronics are mounted in a prism shape around the motor for thesoil sampler, for which a crown gear is visible at top.

Fig. 30. The extreme miniaturization of DS-2 is evident in this assemblyphoto, which resembles watchmaking. Note the umbilical. Press fittolerances and use of adhesives meant nondestructive disassembly wasgenerally impossible.


ranges. constraints. The design requirements on penetra-tion conditions were surface impact velocity(140 < Vp < 210 m/s), total angle of attack (0� < a < 10�),and incidence angle (�30� < i < 30�).

Additionally, the fore- and aftbody configuration deter-mined the center-of-mass position of the entry system(Fig. 28), which influenced its aerodynamic stability. Theaerodynamic properties in turn map into the angularmotions which control the probability of angle-of-attackexcursions. The need for a strongly forward center-of-massfor entry stability (the DS2 vehicles relied on passive, ‘shut-tlecock’, stability and were not spun upon release, as isimplemented on many entry vehicles) drove the overall aft-body configuration, as well as the incorporation of a heavytungsten nose on the forebody.

The entry system was a frangible shell 35 cm in diame-ter. The front part was a 45-degree half-angle cone (witha spherical nose) made of SIRCA-SPLIT (silicon-impreg-nated, reusable ceramic ablator – secondary polymerlayer-impregnated technique), a ceramic material able totolerate 2000 �C, as verified in arcjet tests at NASA Ames.The structural rigidity was provided by silicon nitride, ontowhich the SIRCA was attached. The rear part was a hemi-sphere (to minimize any aerodynamic torques). The flightcharacteristics of the entry system were evaluated largelyby computational fluid dynamic (CFD) studies, togetherwith a very small number of ballistic and wind tunnel testsat Eglin Air Force Base in Florida and in Kaliningrad,Russia. Ballistic tests also ensured that the entry shellwould break at impact.

The power supply (activated by a microswitch on sepa-ration of the probes from the carrier ring on which theywere mounted) comprised two sets of four half-‘D’ cells,each less than 40 g in mass. Design and construction ofLithium Thionyl-Chloride cells able to deliver adequateenergy at �80 �C and able to tolerate 80,000 g shockswas a significant technology development (e.g. Deligianniset al., 1997) The impact tolerance derived in part from care-ful seal redesign and welding procedures, while the majorfactor in achieving adequate low-temperature performancewas the use of a tetrachlorogallate-doped electrolyte (i.e.0.5 M LiGaCl4 in SOCl2), which yielded a capacity of�0.7 A-hr/cell (about 40% of the room-temperature capac-ity) at �80 �C with a cell voltage of �2.5 V.

The systems on DS-2 were exceptionally compact(Fig. 29) – the forebody (containing microcontroller, sam-pling drill and soil analysis experiment) was only 40 mm indiameter and 105 mm long. It was noted that volume was amore challenging constraint than mass, and that volumewas generally not available for fasteners such that partshad to be press-fit or secured with adhesives, making non-destructive disassembly and rework essentially impossible(Fig. 30).

The payload during development was described byBlue (1999), including an atmospheric pressure sensor.Particular effort was devoted (e.g. Catling, 1998; Reynoldset al., 2000) to assuring high performance of this tiny

capacitive diaphragm sensor – not only to survive theimpact, but to retain a precision calibration. Unfortu-nately, a late design change to the microprobes requiredsubstitution of discrete components for the ‘radio-on-a-chip’ system, and the required circuit board space requireddeletion of the pressure sensor.

The mission overall as finally implemented, with partic-ular attention to the operations sequence and sciencemeasurements, is described by Smrekar et al. (1999). Dur-ing entry, a small two-axis accelerometer (Analog Devices


ADXL250, digitized �25 to +25g with 38 mg resolution at20 Hz) would record the drag deceleration, allowing anatmospheric density profile reconstruction down to the sur-face. The impact, at 200 m/s, would result in an estimatedpenetration of 0.5–1 m of the forebody, the deceleration ofwhich would be recorded by a separate single-axis acceler-ometer (Endevco 7570A, digitized to 10g resolution over –10,000 to +50,000 g at 25 kHz). This deceleration profilewould allow estimation of penetration depth (e.g. Lorenzet al., 2000), as well as the mechanical hardness and possi-ble layering in the target material which would be an indi-cation of ice content. It may be noted that theinterpretation of the deceleration record was substantiallycomplicated by the forebody/aftbody separation kinemat-ics at impact. Forebody cooling would be monitored bytwo platinum resistance thermometers to deduce the ther-mal conductivity of the target material (e.g. Smrekaret al., 2001).

The aftbody contained (Fig. 31) the batteries and radiosystem, as well as a single flexible solar cell used as a sunpresence detector. The S-band radio system as flown,weighing 50 g and using 64 cm2 of board space, used 2 Wof electrical power to transmit data at 7 kbps, and 0.5 Win receive-only mode (NASA, 1999). Communications wereeffected via an antenna that nominally pointed verticallyfrom the aftbody – this was a rigid structure with flexible‘whiskers’. The aftbody would in principle be able to con-duct some rudimentary communications even if the fore-body electronics, or the umbilical, failed. The umbilical –

Fig. 31. Cross-section of the deployed DS-2 penetrator, identifying majorcomponents.

for which volume was a particularly challenging constraint– was a concertina’d kapton film 0.6 m long stored in theback of the forebody.

The forebody, made of a high-strength alloy, containedcustom power management electronics (Fig. 29) and anadvanced microcontroller (80C51 architecture running at10 MHz with a 32 channel analog to digital converterand 128 K each of RAM and ROM: the system used only6 mW operating, 0.5 mW sleep.) The forebody also con-tained the �1.5 W Soil Water Detection Experiment, inwhich a motor would turn a sideways augur to bring in�100 mg of soil into a cup comprising a heater and therm-istors to detect latent heat of ice melting, while any evolvedwater vapor would be detected by a tiny tunable diode laser(operating wavelength 1.37 lm) absorption cell, defined bya �2.6 cm optical path (e.g. Smrekar et al., 1999). Thecool-down of the forebody to ambient conditions wouldbe monitored by two platinum resistance thermometersto deduce the thermal properties of the adjacent regolith(e.g. Smrekar et al., 2001).

The cost of the DS-2 project was �$28M (NASA, 1999),plus $1.6M support for the science team. The two DS-2microprobes and MPL to which they were mounted:Figs. 32 and 33) were launched from Kennedy Space Cen-ter at 3.21pm on 3rd January 1999 on a Delta II rocket. Nocontact with the microprobes was possible before impact.Nominally, an overflight by the Mars Global Surveyor(MGS) spacecraft was to interrogate the microprobes somehours after arrival, but neither on this nor subsequentpasses was any response received. After a certain periodthe microprobes would have autonomously transmitted

Fig. 32. The DS-2 entry shell was held by a ‘spider’ mounting, whichreleased it on command, without inducing spin.

Fig. 33. DS-2 capsules mounted on the separation ring around the MarsPolar Lander Entry shell.


at intervals (e.g. to guard against receiver failure) but nei-ther MGS nor a search for signals by a radio telescope atStanford University detected any transmissions. There existseveral failure modes for Mars Polar Lander (which wasalso lost without trace) and DS-2 (e.g. JPL, 2000) – amongmost salient for the latter are the possibility of inadvertentactivation and battery depletion on the launch pad, a sep-aration anomaly, electrical breakdown around theantenna, or failure at impact, most likely of the batteriesor the radio system. The review board noted that no end-to-end test had been conducted in the very rapid develop-ment program wherein operations took place in situ afteran airgun shot.

3.10. Sampling penetrators

Development work has been performed for several yearson various sampling penetrator concepts, which in princi-ple allow acquisition of surface material from a planetarybody without requiring the main vehicle (an asteroid orcomet rendezvous spacecraft, or a balloon on Titan) tocontact the surface.

In the airless body case, a relatively large coring penetra-tor was developed under the NASA Planetary InstrumentDesign and Development Program (PIDDP) by W. Boyn-ton of the University of Arizona. This device demonstrated(Lorenz et al., 2003, 2006) the ability to acquire samplesfrom frozen ice targets. Significantly, the work demon-strated an ability to preserve the stratigraphy in a �3 cmwide, 30 cm long cryogenic (190 K) layered ice target withairgun launch to �20 m/s.

A spring-actuated mechanism sealed forward knifedoors of a sample canister. The canister could then be with-drawn or ejected through back of penetrator. The missionscenario envisaged that a telescoping catcher system (Konget al., 2006) could snatch the canister in free space andintroduce it into a sample return capsule.

An important aspect to the success of kinetic coring isthat the hole is flared i.e. the sample encounters an ever-widening duct. This prevents the material from blocking

the duct. The vehicle shape is unusual for a penetrator, inthat there is an exceptionally wide flare at the aft end. Thisflare has two functions. First, it prevents the vehicle frompenetrating too deeply and thus assuring a clear path forthe sample canister ejection. Second, the wide ring providesfor a large transverse moment of inertia, allowing the vehi-cle to be passively spin-stabilized (unlike most penetratorswhich are slender).

Other work has explored the use of a smaller coring pen-etrator to sample surfaces (Backes et al., 2008), with gravity-dropped or pyrotechnically-actuated tethered harpoons ableto acquire�1 cm3 cryogenic ice samples from�15 m ranges.

3.11. Luna-Glob

A mission advocated from time to time in Russia, capital-izing somewhat on the Mars-96 development (although per-haps now, if it is to happen at all, to involve cooperation withthe Japanese), is a multi-penetrator mission to the Moon,named Luna-Glob. The use of penetrators to deploy seis-mometers to terrestrial planetary bodies has been long advo-cated (e.g. Bogdanov et al., 1988). The original Luna-Globconcept (Surkov et al., 1999; Galimov, 2005) anticipated acarrier spacecraft which would first release a ring-shapedvehicle which in turn would dispense 10 small high-velocitypenetrators to form a small aperture seismic network. Laterduring the cruise to the moon, the carrier would sequentiallydispense two large penetrators, each with an attitude controland retropropulsion system amounting to 250 kg of whichthe penetrator itself comprises 35 kg. Power of �0.5 Wwould be supplied by a small RTG, and the payload wouldinclude neutron and gamma ray sensors, a camera system,a mass spectrometer and seismometer.

The small high-speed penetrators deserve mention, eventhough no significant details have been published. It hasbeen suggested that these could survive impact onto thelunar surface at 2.5 km/s, by using a ‘supercavitating’ con-figuration, with an erodible mast at the nose. This extendednose causes a separation bubble around the vehicle, whichtherefore experiences much less drag (nonetheless, deceler-ations of some 10,000g were claimed). This supercavitationapproach is known to succeed in underwater vehicles (e.g.Ashley, 2001), specifically the rocket-propelled VA-111‘Shkval’ missile which can travel at some 200 knots. Notests or detailed models have been reported on perfor-mance in solid targets, however. At the time of writing thispaper (March 2011), industry materials appear to suggestperhaps carrying four Lunar-A derived penetrators, withno mention of the supercavitating devices, but a morerecent report on the near- and mid-term Russian Lunarprogram (Zelenyi et al., 2011) describes a Luna-Glob con-cept with no penetrators at all.

3.12. Polar night

At least one NASA Discovery proposal (e.g. Mosherand Lucey, 2006) has advocated penetrators to the moon,

Fig. 34. Entry, Descent and Landing concept of the Metnet lander: theMars-96 heritage is evident, although Metnet features an inflatable heatshield for the hypersonic entry phase.

Fig. 35. Mission Lead Ari-Matti Harri (FMI) presents the prototype ofthe Mars MetNet Lander in FMI space laboratory. Photo: Antonin Halas,FMI.


notably to search for lunar volatiles such as water ice,which had been suggested by bistatic radar studies byClementine and later by Gamma Ray and Neutron surveysby the Lunar Prospector (both orbiters relying on indirectmeans to infer water). The Polar Night Discovery missionsuggested three penetrators released from a low 30 kmorbit, and planned to use an agile hydrazine propulsionsystem to achieve a relatively modest impact speed of75 m/s, leading to 300 g impact loads (separation fromthe propulsion system would occur only 20 s beforeimpact). This proposal was not selected for flight. Othershave advocated similar concepts (e.g. Lo et al., 2006), butit should be noted that the scientific imperative for this mis-sion may have been lessened by the LCROSS mission inOctober 2009 which detected water and other compoundsin the lunar regolith via hypervelocity impact (e.g. http://lcross.arc.nasa.gov/).

3.13. MoonLITE

Interest in the UK arose during the late 2000s in imple-menting a Lunar penetrator mission. This followed in partfrom the apparent abdication of JAXA from seeing Lunar-A through completion (leaving the opportunity for lunarseismic discovery open) and the strong perceived publicsupport of planetary exploration following the (unsuccess-ful) Beagle-2 Mars mission.

MoonLITE (e.g. Gao et al., 2008) was anticipated tohave a small lunar orbiter with four penetrators to belaunched in 2010–2011 and to see one year of lunar surfaceoperations. In addition to a Lunar-A like payload of seis-mometer, heat flow (and tilt sensor and impact accelerom-eter), a water/volatiles detector, descent imager and an X-ray spectrometer was planned to fit in a rather modest 2 kgscience payload allocation. (An important distinction isthat a micromachined silicon seismometer would be carriedon MoonLITE, with poorer long-period seismic perfor-mance than Lunar-A, but with a much lower mass(300 g) and volume.) The penetrator vehicle was to havea total mass of 13 kg, plus a rather optimistic propulsionstage of 23 kg to perform a de-orbit from 100 km to a40 km periapsis, then braking there to free-fall and turnto vertical for impact.

The penetrator work advanced to the point of conduct-ing a small number of full-scale penetration tests with somecomponents (including microseismometer parts) in arocket-sled test at the Pendyne range in Wales. However,despite this progress and enthusiasm within the UK spacecommunity for the mission (e.g. Collinson, 2008), thefinancial crises of the late 2000s appear to have stalled pro-gress on the mission.

3.14. Europa/Ganymede/Enceladus

Studies in NASA and ESA of a joint Europa-JupiterSystem Mission (EJSM) featuring a NASA Jupiter/EuropaOrbiter and an ESA Ganymede orbiter led to consideration

of possible surface elements, such as a penetrator (see theSurface Element appendix in Greeley et al. (2010)). The sci-entific goals would be geophysical (notably, seismic studies)and astrobiological – the search for biomarkers in the ice.Industrial studies were pursued in the UK under ESAsponsorship, and JPL studies were also performed.Whether the resource envelopes of these missions allow asurface element, and whether that surface element shouldbe a penetrator or some other architecture is not yet deter-mined. In so far as the landing problem for the Galileansatellites is essentially the same as for the moon in dynam-

http://lcross.arc.nasa.gov

http://lcross.arc.nasa.gov


ics terms, the case for a penetrator over semi-hard landersor other vehicles is not strong. Additionally, emplacementinto the very cold ice surfaces of these worlds poses a sig-nificant thermal management challenge. Also, while anyEJSM element will likely have to use radiation-tolerantelectronic parts, and a penetrator must use shock-tolerantparts, a penetrator must use parts that are both radia-tion-tolerant and shock-tolerant, which likely limits therange of parts that can be used, or at least imposes a testingoverhead.

It may be noted also that penetrators were advocated forEnceladus as part of the TandEM (Titan and EnceladusMission) concept proposed to ESA in 2007 (Coustenis,2008). However, little detailed work on the concept wasperformed.

3.15. Metnet

At the time of writing, only one penetrator projectappears to be under active development, although its flightprospects are not clear. This project is, perhaps surpris-ingly, led by Finland – details are at metnet.fmi.fi. TheFinnish Meteorological Institute (FMI) has been a partici-pant in many planetary missions, including Huygens andthe Phoenix Lander. The combination of this activity(which included participation in the unsuccessful Mars-96mission) and a national debt situation wherein the RussianRepublic was required to provide goods or services to theRepublic of Finland led to pursuit of a Mars-96-like entryand descent system with an inflatable ballute and penetra-tor impact (Fig. 34).

The science payload is considerably more austere thanMars-96, with (as the mission’s name suggests) a focuson meteorological measurements. Additionally a smallcamera, UV sensor and magnetometer is planned. Signifi-cant hardware development has been performed(Fig. 35), including penetration tests, balloon drops andarcjet testing of thermal protection material.

An agreement was signed in 2001 between the BabakinSpace Center in Russia and FMI with a space test antici-pated in 2005. Thinking at the end of the 2000s centeredon flying one or more MetNET vehicles on the RussianPhobos-Grunt mission in 2011, although no recentannouncements have been made on this topic.

4. Considerations for mission success

It is conventional spacecraft engineering practice todesign to ‘3-sigma’ values, such that unacceptable parame-ters are encountered less than 0.5% of the time, althoughthe empirical success rate for planetary missions is ratherlower than this. Yet because a penetrator is so stronglycoupled to its environment, and dependent on an absenceof rocks and steep slopes to achieve tolerable angle ofattack, it is difficult to see how such a reliability level canbe attained, even if shock tolerance of individual compo-nents can be adequately verified. Other environmental

issues such as the fidelity of scientific measurements,acceptability of the thermal environment and performanceof communications systems will be discussed elsewhere, buthave broadly similar considerations.

A notable exercise in the 1977 Mars study was the con-sideration that sufficiently small rocks would not disturbthe penetration process, and sufficiently large rocks behaveas a semi-infinite target. But rocks of a size comparablewith the diameter of the penetrator (R times smaller to R

times larger) could cause the penetrator to ricochet side-ways and break up due to sideloads. The study examinedViking lander measurements of the rock size distributionon the Martian surface to deduce that a probability of fail-ure of �9% derived for R = 5, or �14% for a more conser-vative R = 10.

It was also noted that a 5� angle-of-attack constraint inthat study could be met (assuming level terrain) as long aswinds remained below 14 m/s. Although no probabilisticcalculations were performed on this question, inspectionof Viking data (e.g. Lorenz, 1996) suggests the likelihoodof such winds, even during ‘windy’ seasons, is �5%.

Similarly, examination of topography data generatedfrom photoclinometry and stereo imagery of the surfaceof Europa (e.g. Schenk, 2009) suggest that slopes on thatworld exceed 10� (roughly the AoA allowed for Lunar-A)on as much as 60% of some types of surface. Even the mostbenign regions (crater floors) exceed 10 degrees over 10% ofterrain. Presumably the combination of a global high-reso-lution topography dataset from a future orbiter mission,and highly accurate guidance, might be able to target spe-cific sites where the terrain is flatter than this, but for pres-ent planning purposes, it seems that a probability for asingle vehicle failing to penetrate successfully due to terrainslopes of the order of 10% or more must be tolerated.

Even where the terrain slope and/or rock size distribu-tions of the target area are perfectly known on an airlessbody, it should be noted that the incidence and angle ofattack statistics for a typical vehicle will depend on the(often poorly-known) impulse variations for retro motors.In other words, it is the variation in performance of theengineered systems, not their performance itself, that deter-mines the success of delivery. This makes margin controland risk engineering a particular challenge.

Faced with 10–15% probabilities of encountering intol-erable conditions, simple statistics requires that in orderto achieve a 99.5% probability of at least partial success(i.e. at least one penetrator being emplaced successfully),assuming the failure probabilities are uncorrelated, needsthree penetrators to be delivered to the surface. This 3:1reliability overhead exceeds the number of vehicles thathave been developed in missions to date (DS2, Mars-96and Lunar-A all had two vehicles) and challenges claimsof the cost-effectiveness of penetrators for single-sitemeasurement.

This probabilistic terrain overhead becomes much lesssevere for network science missions (seismology, meteorol-ogy and magnetotelluric investigations) where several vehi-


cles are required to operate simultaneously e.g. to deter-mine seismic source locations. A common ‘minimum net-work’ for seismic studies (e.g. as described in the Marsnetstudy) is of four stations. In order to have a 99.5% chanceof success of emplacing four stations if the probability ofeach is 90%, then only eight stations need to be sent – i.e.the reliability overhead falls from 200% to 100%. For ameteorological network, a 99.5% chance of 12 operatingstations is achieved for 19 attempts – i.e. an overhead ofjust over 50%. Of course, these calculations are stronglydependent on the single-attempt success probability – butthis just makes the whole enterprise contingent upon (usu-ally poorly-known) angle-of-attack tolerances and rockand slope distributions.

For network missions, then, the statistics are only mod-estly forbidding. This situation stands in some contrast tothe military arena in which penetrator technology firstevolved, wherein if a ‘mission’ such as bombing a damfailed, it could be quickly re-attempted. Overhead factorsof at least several are common. A notable counter-example,which proves the point, is the small number of attempts todeposit surveillance penetrators at Lop Nor: like the plan-etary missions to date, this effort did not succeed.

5. Conclusions

Due in part to real or perceived technical challenges,most proposed planetary penetrator missions have not pro-gressed beyond the study phase. Among those that have,programmatic factors have led to project termination(e.g. Lunar-A and CRAF) and only two (DS-2 andMars-96) have reached the launch pad. Neither of thesesucceeded, although in one or possibly both cases, the pen-etration emplacement was not the cause of the failure.

Some of the original scientific justification for Moon andMars penetrators, notably the search for water ice in thesubsurface of the polar regions, has been progressivelyeroded over recent years by scientific progress on thesequestions by other means, e.g. the Phoenix lander on Mars,and the LCROSS impactor at the moon. However, Marsperhaps remains the most attractive target for penetratorinvestigation, since subsurface materials (organics, ratherthan water) are of astrobiological interest, and seismicstudies of its interior are still wanting. Further, on Marsthe atmosphere can fulfil the important and expensivedeceleration and attitude control functions that must beimplemented by propulsion hardware on a lunar or Gali-lean mission.

Yet, despite persistent calls for Mars network science,the programmatic commitment to network missions hasbeen lacking (also for the moon). In part this may be dueto the perpetual erosion of network size in response todeclining margins during development (as designs arerefined – q.v. Lunar-A – the number of stations that canbe accommodated declines and thus the small network sizecompromises scientific capability and the project’s robust-ness to failure becomes reduced), and in part a reluctance

to invest substantially in a large mission for which the sci-entific return may appear risky (the history above, fewinstruments, unfamiliar instruments, modest total datareturn compared with an orbiter, etc.). It would likely beeasier to ‘sell’ a network mission after a successful flightof a single ‘demonstration’ penetrator and its instruments.However, without promising a large scientific capabilitywith adequate reliability, a single penetrator mission (suchas might be easily affordable in a PI-led framework such asNASA’s Discovery program) is not likely to fly. Thus thereis a ‘chicken and egg’ syndrome, due largely to program-matic risk-aversion, such that stand-alone penetrator mis-sions seem unlikely to proceed without some riskreduction. Since the Moon and Mars have received detailedexamination by orbiters with radars, thermal mappers,high-resolution imagers and laser altimeters, it may be thattarget areas can be identified that are large enough to bewithin the capability of modest delivery systems, and thathave gentle slopes and are sufficiently free of rocks thatthe single-station impact survival probability can beincreased above values discussed in the previous section.This question requires detailed examination. For Galileantargets, such data do not exist.

Penetrators may also be attractive for small body targets(comets, asteroids) since penetrators address the anchoringproblem that must otherwise be addressed another way.For the Galilean satellites, the requirements of deliveryare rather similar to the moon, but emplacement in a coldtarget presents considerable thermal challenges. It may bethat other architectures such as semi-hard landers (whichafter all led to the first successful lunar landing – Luna 9)are more efficient.

Empirically, the adoption of novel exploration plat-forms has been largely via ‘add-on’ elements to larger mis-sions, such that if the new element (balloon, rover,penetrator, etc.) fails, the mission overall can still beclaimed to have been a success. Thus, the first planetaryballoons (at Venus, dropped off by the VeGA missions enroute to comet Halley) were secondary payloads with verysmall scientific payloads. Similarly, the first Mars rover(Sojourner, on Mars Pathfinder) was considered strictlyexperimental and carried only a single science instrument.Indeed, the paradigm under which NASA implemented apenetrator mission (DS-2) was as an add-on technologydemonstration. However, add-on elements are always firstto be descoped when resource limits are met during devel-opment (q.v. the CRAF penetrator) and thus without arobust programmatic commitment to network science theprospects even for an add-on implementation of a penetra-tor mission are not especially hopeful. That said, theauthor would be glad to be proven wrong on this point.

Acknowledgements

Some of this work was stimulated by Europa surface ele-ment discussions related to the prospective Europa JupiterSystem Mission (EJSM) and by small mission studies in


support of the Planetary Decadal Survey. The authorthanks Dwayne Day for drawing his attention to the Ta-basco project and to Chris Popock for discussions thereon.Two anonymous referees contributed comments on this pa-per: referee 1’s contributions were particularly cogent.

References

Adams, G.L., Amundsen, R.J., Beardsley, R.W., Cash, R.H., Clark, B.C.,Knight, T.C.D., Martin, J.P., Monti, P., Outteride, D.A., Plaster,W.D., Simon, W.E.. Study of Comet Nucleus Gamma-Ray Spectrom-eter Penetration System. NASA CR-177393. Martin Marietta DenverAerospace, Colorado, 1986.

Anonymous A New Concept in Planetary Probes – The Mars Penetrator.Spaceflight 18, 432–434, 1976.

Ashley, S. Warp Drive Underwater. Scientific American, May 2001.Backes, P., Zimmerman, W., Jones, J., Gritters, C. Harpoon-based

sampling for planetary applications, in: IEEE Aerospace Conference,IAC Paper 1120, Big Sky, MT, March 2008.

Bickler, D., Slimko, E., Stone, J.A. Sub-surface soil sampling withpenetrators, in: International Symposium on Artificial Intelligence,Robotics and Automation in Space, 14–16 Tokyo, July 1997. <http://trs-new.jpl.nasa.gov/dspace/handle/2014/22023> (downloaded24.11.2010).

Blue, R.C. Mars microprobe project instrumentation package. ActaAstronautica 45, 585–595, 1999.

Bogdanov, A.V., Nikolaev, A.V., Serbin, V.I., Skuridin, G.A., Khavr-oshkin, O.B., Tsyplakov, V.V. Method for analyzing terrestrialplanets. Cosmic Research 26, 505–515, 1988.

Bollinger, M.J. Warriors and Wizards: The Development and Defeat ofRadio Controlled Glide Bombs of the Third Reich. Naval InstitutePress, 2010.

Boynton, W.V., Reinert, R.P. The cryo-penetrator: an approach toexploration of icy bodies in the solar system. Acta Astronautica 35(Suppl.), 59–68, 1995.

Boynton, W.V., D’Uston, L.C., Young, D.T., Lunine, J.I., Waite, J.H.,Bailey, S.H., Berthelier, J.J., Bertaux, J.L., Borrel, V., Bruke, M.F.,Cohen, B.A., McComas, D.H., Nordhold, J.E., Evans, L.G., Trom-bka, J.I. The determination of ice composition with instruments oncometary landers. Acta Astronautica 40, 663–674, 1997.

Braun, R.D., Mitcheltree, R.A., Cheatwood, F.M. Mars microprobeentry-to-impact analysis. Journal of Spacecraft and Rockets 36, 412–420, 1999.

Catling, D.C. High-sensitivity silicon capacitive sensors for moderate gaspressures. Sensors and Actuators A 64, 157–164, 1998.

Chicarro, A., Coradini, M., Fulchignoni, M., Liede, I., Lognonne, F.,Knudsen, J.M., Scoon, G.E.N., Wanke, H. MARSNET AssessmentStudy Report, ESA Publication SCI(91)6, European Space Agency,Noordwjik, The Netherlands, January 1991.

Collinson, G.The UK Penetrator Consortium Planetary penetrators: thevanguard for the future exploration of the solar system. Journal of theBritish Interplanetary Society 61, 198–202, 2008.

COMPLEX. On the Scientific Viability of a Restructured CRAF SciencePayload: Letter Report. Committee on Planetary and Lunar Explo-ration, National Research Council, Washington, DC, USA. ISBN: 0-309-12295-3, 1990. Electronic copy downloaded 11/20/2010 from<http://www.nap.edu/catalog.php?record_id=12330>.

Correll, J.T. Igloo White. Air Force Magazine, 56–61, 2004.Coustenis, A., 154 co-authors. TandEM; Titan and Enceladus mission.

Experimental Astronomy: Astrophysical Instruments and Methods,23, 893–946, 2008.

Deligiannis, F., Frank, H., Davies, E., Russell, P.G., Reddy, T.B. Ultralow temperature batteries for martian applications, in: Paper 97161IEEE Aerospace Conference, 1997 (See also Frank, H., Kindler, A.,Deligiannis, F., Davies, E., Blakevoort, J., Ratnakumar, B.V.,Surampudi, S. NASA Battery Workshop, Huntsville, AL, October27, 2008).

Doengi, F., Burnage, S.T., Cottard, H., Roumeas, R. Lander shock-alleviation techniques. ESA Bulletin 93, 51–60, 1988.

ESA. Vesta: A Mission to the Small Bodies of the Solar System. Report onthe Phase A Study, ESA SCI(88)6. European Space Agency, Noo-rdwijk, The Netherlands, October 1988.

Evans, L.G., Trombka, J.I., Boynton, W.V. Elemental analysis of a cometnucleus by passive gamma ray spectrometry from a penetrator. Journalof Geophysical Research 91, D525–D532, 1986.

Fleischer, G. German Air-Dropped Weapons to 1945. Motorbuch Verlag,Stuttgart, 2003.

Friedlander, A.L., Niehoff, J.C., Davis, D.R. Penetrator Mission Con-cepts for Exploration of the Galilean Satellites, in: AIAA 76-800,AIAA/AAS Astrodynamics Conference, San Diego, CA, August 18–20, 1976.

Gao, Y., Phipps, A., Taylor, M., Crawford, I.A., Ball, A.J., Wilson, L.,Parker, D., Sweeting, M., da Silva Curiel, A., Davies, P., Baker, A.,Pike, W.T., Smith, A., Gowen, R. Lunar Science with affordable smallspacecraft technology: MoonLITE and Moonraker. Planetary andSpace Science 56, 368–377, 2008.

Galimov, E. Luna-Glob project in the context of the past and presentlunar exploration in Russia. Journal of Earth System Science 114, 801–806, 2005.

Gavit, S.A., Powell, G. The new millennium program’s Mars microprobemission. Acta Astronautica 39, 273–280, 1996.

Gillespie, P.G. Weapons of Choice: The Development of Precision GuidedMunitions. University of Alabama Press, Tuscaloosa, 2006.

Greeley, R., Dougherty, M., Pappalardo, R., Lebreton, J.-P. EuropaJupiter System Mission (EJSM): Exploring the Emergence of Habit-able Worlds Around Gas Giants. JPL D-67959, Jet PropulsionLaboratory, Pasadena, CA, 15 November 2010.

Harvey, B. The Japanese and Indian Space Programmes: Two Roads intoSpace. Springer-Praxis, Berlin, 2000.

Harvey, B.. Russian Planetary Exploration: History, Development,Legacy and Prospects. Springer-Praxis, Berlin, 2007.

Hayashi, T., Saito, H., Orii , T., Masumoto, Y. MiniaturizationTechnology for Lunar Penetrator Mission, IAF/IAA-93-U.5.575, in:44th Congress of the International Astronautical Federation, Graz,Austria, October 16–22, 1993.

Hinada, M., Ninomiya, K., Mizutani, H., Kohno, M., Nakajima, T.,Saito, H., Kawaguchi, J. Lunar exploration by penetrators, in: 43rdCongress of the International Astronautical Federation, Washington,DC, August 28–September 5, 1993.

Holler, R., Horbach, A., McEachern, J. Not ready for retirement: theSonobuoy approaches age 65. Sea Technology, 2006.

Hughes. Pioneer Mars Surface Penetrator Mission: Mission Analysis andOrbiter Design, NASA CR-137568. Hughes Aircraft Company, 1976.

JPL Special Review Board. Report on the Loss of the Mars Polar Landerand Deep Space 2 Missions, D-18709. Jet Propulsion Laboratory,California Institute of Technology, 22nd March 2000.

Kargl, G., Macher, W., Kmole, N.I., Thiel, M., Rohe, C., Ball, A.J.Accelerometry measurements using the Rosetta Lander’s anchoringharpoon: experimental set-up, data reduction and signal analysis.Planetary and Space Science 49, 425–435, 2001.

Kargl, G., Komle, N.I., Ball, A.J., Lorenz, R.D. (Eds.), Penetrometry inthe Solar System II. Austrian Academy of Science Press, 2009.

Keese, D.L., Lundgren, R.G. Integrated Reentry and PenetratorVehicle (IRPV) for subsurface soil collection and analysis onMars, in: Utah State University Conference on Small Satellites,Logan, Utah, 1994.

Komle, N.I., Ball, A.J., Kargl, G., Stocker, J., Thiel, M., Jolly, H.S.,Dziruni, M., Zarnecki, J.C. Using the anchoring device of a cometlander to determine surface mechanical properties. Planetary andSpace Science 45 (12), 1515–1538, 1997.

Komle, N.I., Kargl, G., Ball, A.J., Lorenz, R.D. (Eds.), Penetrometry inthe Solar System. Austrian Academy of Science Press, 2001.

Kong, K.Y., Gorevan, S., Mukherjee, S., Wilson, J. Telescoping samplecanister capture mechanism (TSCCM). Acta Astronautica, 1008–1013,2006.

http://trs-new.jpl.nasa.gov/dspace/handle/2014/22023

http://trs-new.jpl.nasa.gov/dspace/handle/2014/22023

http://www.nap.edu/catalog.php?record_id=12330


Lo, A.S., Eller, H., Sollitt, L. Penetrator Science – Making an Impact onPlanetary Composition Science, AIAA 2006-7423, in: Space 2006Conference, San Jose, CA, September 2006.

Lorenz, R.D. Martian Surface Windspeeds, described by the Weibulldistribution. Journal of Spacecraft and Rockets 33, 754–756, 1996.

Lorenz, R.D. Spinning Flight: Dynamics of Frisbees, Boomerangs,Samaras and Skipping Stones. Springer/Copernicus, 2006.

Lorenz, R.D., Shandera, S.E. Target effects during penetrator emplace-ment: crushing. Heating and Triboelectric Charging Planetary andSpace Science 50, 141–157, 2002.

Lorenz, R.D., Boynton, W.V., Turner, C.F. Demonstration of cometsample acquisition by penetrator, in: ESA Proceedings, IAA Confer-ence on Low-Cost Planetary Exploration, SP-542, 2003.

Lorenz, R.D., Boynton, W.V., Turner, C.F. Demonstration of cometsample acquisition by penetrator. Acta Astronautica 59, 1000–1007,2006.

Lorenz, R.D. A review of titan mission studies. Journal of the BritishInterplanetary Society 62, 162–174, 2009.

Lorenz, R.D., Bannister, M., Daniell, P.M., Krysinski, Z., Leese, M.R.,Miller, R.J., Newton, G., Rabbetts, P., Willett, D.M., Zarnecki, J.C.An impact penetrometer for a landing spacecraft. MeasurementScience and Technology 5, 1033–1041, 1994.

Lorenz, R.D., Moersch, J.E., Stone, J.A., Morgan Jr., A.R., Smrekar, S.E.Penetration tests on the DS-2 Mars microprobes: penetration depthand impact accelerometry. Planetary and Space Science 48, 419–436,2000.

Lumpkin, C.K.. Mars Penetrator: Subsurface Science Mission. SAND-74-0130. Sandia National Laboratories, 1974.

Manning, L.A. (Ed.). Mars Surface Penetrator – System Description,NASA TM 73243, June 1977.

Mizutani, H., Kohno, M., Tsukamoto, S., Kawaguchi, J., Hinada, H.,Ninomiya, K. Lunar interior exploration by lunar penetrator mission,IAF 90-039, in: 41st Congress of the International Federation,Dresden, GDR, October 6–12, 1990.

Mizutani, H., Kohno, M., Nakajima, S., Fujimura, A., Kawaguchi, J.,Saito, H., Hinada, M. Japanese Lunar Mission, Lunar-A. ActaAstronautica 35 (Suppl.), 323–327, 1995.

Mizutani, H., Fujimura, A., Tanaka, S., Shiraishi, H., Nakajima, T.Lunar-A mission: goals and status. Advances in Space Research 31,2315–2321, 2003.

Mizutani, H., Fujimura, A., Tanaka, S., Shiraishi, H., Nakajima, T.Lunar-A mission: outline and status. Journal of Earth System Science114, 763–768, 2005.

Morita, Y., Kawaguchi, J., Hashimoto, T., Nakajima, T., Baba, K.,Terada, H. Attitude control for the Lunar-A penetrator system and itsflight test, AAS 97-475, in: 7th International Space Conference of thePacific Basin Societies, Advances in the Astronautical Sciences, vol. 96,pp. 853–863, 1997.

Morita, Y., Kawaguchi, J., Nakajima, T., Hinada, H., Naide, T. Attitudedynamics of the Lunar-A penetrator, AIAA 92-4370, in: AIAA/AASAstrodynamics Conference, Hilton Head Island, SC, USA, August 10–12, 1992.

Mosher, T.J., Lucey, P. Polar night: A lunar volatiles expedition. ActaAstronautica 59, 585–592, 2006.

Murphy, J.P., Reynolds, R.T., Blanchard, M.B., Clanton, U. S., 1981.Surface Penetrators for Planetary Exploration: Science Rationale andDevelopment Program. NASA Technical Memorandum 81251.

Murray, I. Bouncing-Bomb Man: The Science of Sir Barnes Wallis.Haynes Publishing, Yeovil, UK, 2009.

Nakajima, T., Hinada, M., Mizutani, H., Saitoh, H., Kawaguchi, J.,Fujimura, A. Lunar Penetrator Program: Lunar-A. Acta Astronautica39, 111–119, 1996.

NASA. Mars Polar Lander/Deep Space 2 Press Kit, November 1999.Newton, I. Philosophi� Naturalis Principia Mathematica, 1687.NRC. NASA Ad Hoc Committee on Planetary Surface Penetrator

Missions, 1975–1976. National Academy of Science, 1976.NRC. Effects of Nuclear Earth-Penetrator and Other Weapons. National

Academy of Science, 2005.

Normile, D. Lunar-A Moon Mission gets an ‘F. Science 315, 445–446,2007.

Paton, M., Kargl, G., Ball, A.J., Green, S.F., Hagermann, A., Komle,N.I., Towner, M., Zarnecki, J.C. Computer modeling of a penetratorthermal sensor. Advances in Space Research 46, 337–345, 2010.

Pocock, C. 50 Years of the U-2: The Complete Illustrated History of theDragon Lady. Schiffer Publishing, Atglen, PA, 2004.

Reynolds, J.K., Catling, D., Blue, R.C., Maluf, N.I., Kenny, T. Packaginga Piezoresistive pressure sensor to measure low absolute pressures overa wide sub-zero temperature range. Sensors and Actuators 83, 142–149, 2000.

Robins, B. New Principles of Gunnery: Containing the Determinations ofthe Force of Gun-powder and Investigations of the Difference in theResisting Power of the Air to Swift and Slow Motions. J. Nourse,London, 1742.

Rosenau, W. Special Operations Forces and Elusive Enemy GroundTargets: Lessons from Vietnam and the Persian Gulf War. MonographMR1408. RAND Corporation, Santa Barbara, 72p., 2001.

Schenk, P. Slope characteristics of Europa: constraints for landers andradar sounding. Geophysical Research Letters 36, L15204, 2009.

Schmitt, V. Controlled Bombs and Guided Missiles of the World War IIand Cold War Eras: An Inside Story of Research and DevelopmentPrograms. SAE Press, 2002.

Shibuya, K., Kaminuma, K., Ito, K., Yamada, I., Ikami, A.,Mizutani, H., Fujimura, A., Tsukamoto, S., Kanazawa, T. Plannedexplosion seismic experiments in East Queen Maud Land using apenetrator. Proceedings NIPR Symposium on Antarctic Geoscience5, 29–38, 1991.

Shibuya, K., Kasahara, M., Hayakawa, M., Fujimura, A., Koyama, J.,Kaminuma, K., Mizutani, H. Application of GPS differential posi-tioning for the development of the antarctic penetrator. Journal ofPhysics of the Earth 41, 291–304, 1993.

Shiraishi, H., Tanaka, S., Fujimura, A., Hayakawa, H. Present status ofthe Japanese Penetrator Mission: Lunar-A. Advances in SpaceResearch 42, 386–393, 2008.

Simmons, G.J. Surface penetrators – a promising new type of planetarylander. J. Brit. Interplanetary Soc. 30 (7), 243–256, 1997.

Smrekar, S.E, Lorenz, R.D., Urquhart, M. The Deep-Space-2 penetratordesign and its use for accelerometry and estimation of thermalconductivity, in: Komle, N.I., Kargl, G., Ball, A.J., Lorenz, R.D.(Eds.), Penetrometry in the Solar System. Austrian Academy ofScience Press, pp. 109–124.

Smrekar, S.E., Catling, D., Lorenz, R., Magalhaes, J., Meyer, M.,Moersch, J., Morgan, P., Murphy, J., Murray, B., Presley- Holloway,M., Yen, A., Zent, A. The DS-2 Mars Microprobe Mission. Journal ofGeophysical Research 104 (E11), 27013–27030, 1999.

Steckemetz, B., Weinss, W., Bradbury, B., Pickett, A., Natenbruck, P.,Roumeas, R. Flight and impact dynamics of landers for marsexploration, IAF-93-Q.1.379, in: 44th Congress of the InternationalAstronautical Federation, Graz, Austria, October 16–22, 1993.

Steltzner, A.D., Nasif, A.K. Anchoring technology for in situ exploration ofsmall bodies, in: IEEE Aerospace Conference, Big Sky, MT, March2000.

Stetson, D. Near-nucleus trajectory design for comet penetrator-landerdata relay. AIAA, 88–4267, 1988.

Stone, J.A., Rivellini, T.P. Configuration evolution of the DS-2 penetra-tor, IAA-L98-0303, in: 3rd IAA International Conference on Low-Cost Planetary Missions, Pasadena, CA, April 1998.

Surkov, Y.A.. Exploration of the Terrestrial Planets from Spacecraft:Instrumentation, Investigation, Interpretation. Wiley, Chichester, UK,1997.

Surkov, Y.A., Moskaleva, L.P., Shchegov, O.P., Sheretov, E.P., Kremnev,R.S., Pichkhadze, K.M., Akulov, Y.P., Dolgopolov, V.P. Lander andscientific equipment for exploring of volatiles on the Moon. Planetaryand Space Science 47, 1051–1060, 1999.

Surkov, Yu.A., Kremnev, R.S. Mars-96 mission: Mars exploration withthe use of penetrators. Planetary and Space Science 46, 1689–1696,1998.


Swenson, B., Squyres, S.W., Knight, T.C.D. A proposed comet nucleuspenetrator for the comet rendezvous asteroid flyby mission. ActaAstronautica 15, 471–479, 1987.

Ulivi, P., Harland, D.M.. Robotic Exploration of the Solar System, Part 2:Hiatus and Renewal 1983–1986. Praxis-Springer, 2009.

Vick, A.J., Moore, R.M., Pirnie, B.R., Stillion, J., 2001. AerospaceOperations Against Elusive Ground Targets (Appendix A:) MR-1398.Rand Corporation, Santa Monica, 192pp., 2001.

Wolf, W. German Guided Missiles Henschel HS 293 and RuhrstahlSD1400 X “Fritz X”. Merriam Press, 1988.

Young, C.W. Penetration of sea-ice by air-dropped projectiles, in: IEEEOceans Conference, New York, 1974.

Young, C.W., Keck, L.J.. An Air-Dropped Sea Ice Penetrometer, SC-DR-71-0729. Sandia National Laboratories, Albuquerque, New Mexico,1971.

Young, C.W.. Penetration Equations, SAND97-2426. Sandia NationalLaboratories, Albuquerque, 1997.

Zelenyi, L., Khartov, V.V., Mitrofanov, I.G., Martynov, M.B. Short- andMid-term Russian Lunar Program, Abstract #1804, in: 42nd Lunarand Planetary Science Conference, Houston, TX, March 2011.

planetary penetrators: their origins, history and future

Documents