METEOROID IMPACTS ON THE GEMINI Vr'INDOWS
I{ERBERT A. ZOOK, RODERT D. FI-AHERTY lnd DONALD J. KESSLER
Spee Pblsie Division, NASA Mamed Spaceratr Cent.., Houslon, Tqas 7?058,U.S.A.
(Recebed l0 Nothbet
1969,
Abstnct
Fouriar windows of the Gdini spa@ft
*erc cloely @min.d afl.r retum from
spae flight for evidenc of Det@ro:d imprc'. Alrhough a r6ber of himqpil
pnt $ere
foJnd on e{h window. on\ onc oi rlele ?irs appeab ro have bed clied by a dereorord
impet. A m€toroid nu-lfus
ftlation calculal€n froh this siltl. hit is found io be in close
( 1966) anal'sis oi the Explord sd PegaG ftt@roid penetralion
agr@nenl with Nawnn\
exneribdts.
Panicl€. tBs. and da disdburion ons
de d€rivei ftom fte n'rr-mris
reition dd .r pr*nred in graphiel form. PrcbleEs i. inrerprerirg the dara b..aus ol
contaninanls on thc aindows are also di$used.
r. INTRODUCTION
The windowsof the l0 mannedGemini spacecraftthat were orbited during 1965and
1966presented
anexcellentopponunityto searchfor neteoroidimpactevidence
andthereby
to determinethe associated
meteoroidimpactrate. The opportunitywasgood because
the
windows* prcsented a smooth exterior surface that f.cilitated detection of very small
impact cratersand becausethe area-timeproductof 42.8ft -daysin spacefor the 14 windows examined was quite adequate to determine answersto certain questions raised by
previous experimental results about the influx rate of microm€teoroids.
Two problemsdid arise, however,in interpretingthe resultsfrom this experiment.
First, postflightexaminationof the windowsrevealeda numberofmicroscopicpits on both
the interiorand exteriorsurfaces
of the llindows. Therefore,it wasnecessary
to distinguish
betweencratersthat werecausedby impacting meteoroidsand pits that werecausedby other
sources.Second,afterthe spacecraft
had returnedfrom space,a somewhatscalysubstance
was found adheringto portions of the windowsand occasioDally
coveing as much as 80
per centof the windowsurface.The thicknessofth€ scalevariedfrom a fractionofa micron
to as muchas 50 lt- Obviously,if this rnaterialwason the windowswhilein orbit, it would
significandy
modify the resultinganalysisof the meteoroidinfluxrate; lor example,micronsizedmet€oroidswould noi penetratethe thickerscaleand hencewould not be detectedon
portionsof t}le windows.
The followingsectionscontain| (2)A discussion
of the windowcontaminationproblen;
(3) A descriptionof the experinentalprocedureused to examinethe windows; (4) An
analysisof how the wiodow data are relatedto laboratorytests; (5) A discussionof the
probablevalidity of the Geminiwindowdata; (6) A pres€ntation
of t.hecumulativeflux vs.
inass; (7) Additional .esultsofinterest derivedfrom the meteoroidimpact experiments;
(8) Conclusions.
2. II'INDQW CONTAMINATTON
All of the Gerniniwindowswerefound to havebeer coDtaminated
to varyingdegrees
with foreignmaterialduring spaceflight. The sourceofthis contaminationis still a matter
of sone conjecture.However,a crediblereconstructionof the eventsthat causedthe scale
on the windowshasbeenderivedand, becaus€
of its pertinenceto the subjectof this paper,
is presentedbrieflyin the followingparagraphs.f
* only the outer pan€ of the th@-panc Gemini window wa consid..ld in rhis €xaDination.
i This reonsrrucrion wa tl2lMiated to us by G. P. Bonocr ol thc AsrrcnoEy Branch add othss at rhe
NASA Mdned Spaffiaft Cents.
951
954
H. A. ZOOX,R. E. FLAHERTYatutD. J. KESSLER
Aiter the spacecraft
returnedfronr spac€,samplesof thescalefrom severalwindowswere
chemicallyanalyz€d(Bonnerer a/., 1968).The rnajorcons.ituentofthe scatewasfound to
be a siliconoil. The presence
ofa rransparent,oily materialon somespacecrafr
winatows
duringthe orbital pe od becameapparentduring the CT-w flight. Duringexrra-rehicutar
activityin this fli8ht, the pilot inadverrentlybrushedagainstand smearedthe window. The
smearingwasobservedand reportedby tlle commardpilot and is alsoapparentin a photo_
graphtakenof tie CT-IV windowwhilein orbit. Theseobservations
led to the spec;htion
tllat the oil was.eleased
from the ablativenose-cone
fairingwhich protectedtle noseofthc
Gemini sparecraft fiom aerodynamic heating during launch. tr s€;rnedquite possiblethat
the DC-325ablationmat€rialwould heatenoughduring launchto rctea;esiliconoil in a
Iiquid form that could rhenwashback onto the windows.
For this reasoo,the outer windowsof cemini spac€craftVIII throughXII wereprotecteddurinSlaunch with a' additionatcoveringwindowwhich wasdiscarded,as wasrhe
nose-cone
fairing, after o.bir insertion. This precaution,however,only partially reduced
thewindowcontamiuation(largevariationsin contaminationfrom indowto windowmade
it somewhatdifrcult to correlatecllangesin contaminationwith actionstaken to realuce
contamination).Other speculations
concerningadditionalsourcesofcontaminationwere
otrered and found lacking in one respector another.
It waslaternotedthat oil occasionally
appearedon theinsideasw€ltas the outsideofthe
outer Gemini windowpanes.This observationled to the propositionthar silicon oil had
es.apedfrom the porringnateriat { Rl V-90)s hrchua, usj asi seatfor rbeq indowframel,
and had spread onto the surface of rhe glass. To verify this proposirion, thc GT-XII
windowswere bakedin an oven under vacuumconditionsbefor; Bisht and werecleaned
b€forelaunch. Thesewrndows'Id exhjbit signrJicaorty
lessconraminrtionafler recoveq
than did the windowsof the previousCemini spacecraft.
Hen€e,the folowing cemini windowcontaminationhistoryis widelyacc€pted.Silicon
._
oil f.om the RTV-90 potting compound creepsout onro the surfaceofthe wind;w under th€
vacuumconditionsof space. Vacuumgreasefrom the window
tasket (a paper material
satumtedwith vacuum grease)may also do the same. Silicon oil from the ablative nose_con€
fairingquiteprobablyalsogetson rhewindowsduringtaunchunlessthe windowshaveprotectiv€ covers. During reentry, malerial from the rear ablative heat shietd (also DC-32j) as
wellasmaterialfron someof lhe guilloLined
electrical
connections
{rbeuniquecompositional el€mentsofwhich weredetectedin the chemicalanalysisof the window scale)swirls
bacl,uponthe $indowaDdadheres
ro Lhe\iticonoil atread'y
on rhe uindow. The hearol
reentryth€n bakesthis materialtumly on the window, producingthe scaleobservedupon
recovery.A closerlook at how this contamirationaFectsthe mereoroidimDactanalvsi3is
deferreJto Section5.
]. EXPERIMENTALPROCI]DURE
After a smallsampleofthe window scalewasremovedfor chemicalanalysis.the windo$s were subjectedro an oplrcal anatysisro determinethe cbangein oprjial propenies
causedby the contamination. The scalewas then scrapedotr witl a razoi bhd;. a;d rhe
windowsurlac€swerecleaDed
and "ubiecredro a secondoptrcat analysi". Tbewindowsr ere
then examinedin detail for evidenceofmeteoroidimpactcrarering. It shouldbe notedat
tlis time tlnt, althou8hgreatcarewasexercised,
it is dimcult to bo certainwhetler or not
the nzor-bladecleatringwasresponsibte
for any ofrhe pits larerobservedon the windows.
It is believed,[ow€ver, rhat no parricularproblemis pres€ntedwith respectto confusing
this typepit with im!'act craters,as will be explainedin Section4.
Selected
regionsof severalwindows(boththeexteriorand interiorsurfac$)wetescanned
MITLOROID IVPA''TS ON THE CfMTNI WINDOWS
955
at a maSnincationof l00x and all pirs w€rerecord€drhar had diamererslarg€r than 20
p. Bel$€€nfour atrd ten 20 t or largercrar€rsper in! $ere tound on tbe ouisideof tbe
windows,and betweenoneand six of tlese crarerswerefound on ihe insideof tle windows.
Roughly, there were twic€ as many pits of this size on the outside surfacesas on the inside
The entiresurface(85 in ) of eachwindowwasrhenscannedat a maenificationof20 x.
andallc.aters$ererecorded
tbarhaddiameters
ldrgerrhant0O/. A ;fthe:ecrarersand
some of the smallerones were rhen examinedin detail at magnilicationsup to 750x.
Although thereis considemble
variarionamongindividuatf,its, the rario ofpit diameterto
pit depth was typicallyabout 10, regardless
of cratersize_Similarresuttsfor Vycor glass
weie found for cratersproducedin the laboratoryat both high and low impact velociti€s.
.I. ANALI'SIS OF DATA
The flight time in daysand the Dumberofpits 100t or largerin dia. thar werefound on
eachwindowexarnitred
are shownin Tabl€ l. It is seenthat the correlationbetweenlencth
ofspaceflightdndnumberof pilsoDrbeq indoqsir not close-For exampte.
CT-Vjtwa;itr
orbit for 13.8days and had 16 and 15 pirs on rhe outsidesufaces of the lefr and right
windows,rcsp€ctively,
wbileCT-VIA had corresponding
numbersof2l and 9 pits but;as
ontyin orbit for l.l daysofflight. Curiously,therewerefew t 00-t and larSercrarerson rhe
intedor surfacesof t-hewindows; rwo or threeper window wastypical.
T^M
l. NUER
or prs w
lOO/. oN cE cEurNr
krr rindow
GT.IV
GT,V
GT,VTA
GT.VII
GT-IX
GT-X
GT,XI
GT.XII
4.1
79
I l
t3.8
30
2.9
l0
39
2
4
2l
t6
9
R r B h ts i n d o w
9
l5
3
9
l0
3
The pit datacould be analyzedby subtractingthe numberof pits observedon ihe inside
of the windowsfrom thoseobservedon ihe outsideand ascribinrthe ditrerence
to meteoroid impacrcrdtering.Tbis melhodisnor satr.facrory
torar teasifour reuson":
(l) The CT-VIA mission, for example, would give an averageformarion rate for the
l0t!p and larger craters about l0 times gr€at€r tllan would be obtained from rhe GT-VII
rnission. This would iodicateconsjderable
show€ractivity for lhe verv small meteoroids
{-10 10g1.Radromeleordaraindrcate
lhar majorshowiracrivirydeireases
signrficanrly
relative to sporadic activity as tle meteoroid massdecreases.The trend indicated by th€
radio meteordata would predictlittle detectableshoweracrivityat 10-10g. This tr€nd is
contumed by tlrc !'enetntion histories of the Explorer and Pegasuspenetration satellites.
Also, the GT-VII and GT-VIA spacecraftwere in orbit at the sametime.
(2) The left and right windowsof severalofthe spacecraft
do not havesimilarnumbers
of l0Gp and largercmiers. The chanceoccurrence
ofsuch largediscrepancies
is verysmall.
(3) For the 20.p-dia. and larger craten, one window wasexaminedwhich had about the
samedensityofpits (approxirnat€ly
6 per inN)on rheinsideof the wiDdowason the ourside.
956
H, A, ZOOK. R. E. FLAIIERTY and D. J. KESSLER
And, in general,for the 20-p crat€rs,the differenceber\reenthe outsideand insidecrater
densitiesvaried greatly fron 0 per jn, to about 5 per inr. This diff€rencealso had no
particularcorrelationwith the lengthof spaceflight.
(4) With oneexception
, the pits observ€don the windowsdo not havethecharacterisrics
of laboratory-produced
hypervelocityimpacts by small projectiles. The overwh€lming
majority of the pils closelyresembles
the type of pit that is comrnonlyleft after a ghsspolishingoperation.
The lasl reasoncited aboveprovidesthe only reasonable
crirerionfound for derermining wh€theror not a Geminiwindowpit is, in fact, a meteoroidinpact crater. That is, .he
overwhelmingproportion of meteoroidsstriking the spacecraftshould have velociriesin
excessof 7 km/sec(the nininum possiblev€locityof impact equalsthe escapevelocity
minusthe orbitalv€locity^,a kn^ec). Cratersproducedin Vycorglassin thelaboratoryby
projectilestraveling7 km/secor faster have a charactefisticmolten-appearing
or partly
fusedregionin the ccntralpart ofthe crater. This fus€dappcaraoce
in the centralportion of
tlrc crateror pit hasn€verbeenobserved
forlow-v€locityimpactcratersandfor chip-outsor
pits left from a glass-polishing
process.Th€ kinds of pits found on the windowsand the
impactcratersproducedin the labomtorywill be examinedand comparedin the following
paragraphs.
The singlecraterfoundon the Geminiwindowsthat is believedto becaus€dby meteoroid
impact wasfound on the left-handwindowof the GT-V spacecraftand is shown in Fig. I .
The craier is -ll0 p in dia. and 12& deepand hasmany of the characteristics
of a hypervelocityimpactin thissizerange. That js, thereis a molten-appearing
raisedIip surrounding
the indentedcentralregionof the crater. Dxteriorto this regionand roughlyconc€ntricto
it is a conchoidalfractureregion. (The craterdiameterreferredto henceforthwill be rhe
diameterof the coochoidalfractur€ region unlessotherwisespecjfred.)The glassin ihe
conchoidallracture region has chippedaway to a depth of -5 p. Figure2 illustratesa
view of this crater(not to scale).
crosesectional
1-110',
I
['."-]
-
i
12r
Fra, 2. AlRo)<MrE
cRoss
oF fr!
cT-V
Because
GT-VII spentthe longesttime in space,the pits or craterson itswindowsare of
unusualinterest. Each of tle 31 100'p and larger pits was examiredin great detail at
magnificationsup to 750x to searchfor evidenceof hypervelocityimpact. Typical exampl€s
of pits od the GT-VII windowsare shownin Figs. 3-6. Figure 3 showsa chipout or pit
-90 t in dia. and -10/ deep. A smallsatellitecrateris locatednearbyandjustoutofthe
field of view in the photograph. One or more smallsat€llitecraterswerc associated
with a
largefraction ofthe Gemini windowcraters.Cratergroupingis a commonoccurr€[cefor
pits madeduring polishing.
Figure 4 showsan elongatedcrater of lateral dimensionsl00t by l30p and -l0p
deep. A smallsat€llitecmter is locatednearby. Figure5 showsa thjrd chip'out -120 in
dia. and aSain-10 A deep.
F_(i. l. Ponr ^nD
rN rHr LrF,r^rD CT-V wr^urw.
Crar.r dianeler is -110 / and crater d.fth is ll /,.
$i\@$.
FrG. l. CHrP{o
oN ^ GT-vll
The diadeler is -901? and thc deprh is -10
/r.
ljrti .1 CHrr{nr oN A CTvll sr\ars
T h c d l . m c l e ri s - l I 0 / , a n d r h . d . t r h i \ - . 1 0 / , .
F(r. 5. CsP{6
or A (;T-VII \'\D.$
t h c d i a m c t c ir - 1 2 0 / , r n d t h e d c p r hn - . 1 0 / r
I k,
l(,
6 S.R^r(H oN a (;l
vll
$r\Dos
7 T$!) r!p\r !
M^F Ir CoMr\\l
I l r c d r n r c r c f , i r r h e . n r r e r . f u r o i \ . ' 1 0 0 / , n r d r h ed . p r h r i - l t !
trr!.r.rnrcr rs '-2U),/r and rlrcdclrh r\ -1.1i
thcdtameLeroithe
Fro. 8, TBE ryo la]m
oF Frc. 7 ssom
The lip diamteB of lhe crat6
FIc. 9. tuP^cr tuR
Tl'e dimler
^r a BraIfl
are -20 r (smlld)
md -lo
p (large.).
c^u€D By ^ 38zl-p-Dra. PlrB sFtrm m (rNc Nro VycoR Ar A
vIlcn
oF 7.37lo/s.
of the @td (including the spall.d{ul Etion) is -6600 /,.
I r,
l0 INr! r
( : 1 6 0 0r ) L , r\ \ \
(tr 0l7 tm )..
I r c r r x r . r d r a n r c t c \r - 1 1 0 0 ,
't,
I
l.
-F,
'a
l ,
JT
.: "i
it*
t " j t \
(r\LR\L r,{ik,\ (n (R^l,R rN I.b. l0 rr\\\
rr irtr,rnR
Th. I rnr.r.r nf r rc rc.rivclv unbrokef .enrra r.gi\)ni.-ll0/
METEOROIDIMPACIS ON THE CEMINI WINDOWS
951
Scratchesare anoiher type of imperf€€tion,of which severalare includedamong the
100-pand largerpits listedin Table L A very largescratchthat is severalhundredmicrons
long but only a few micronsd€epand 70 or 80 t wide is shownin Fig. 6. This scratchwas
oneofthetwo imperfections
in the GT-VIIwindows that hada lineardimensionover200A.
Occasionally,a chip.out was seenthat was not immediatelyidentifiableas being of nonhypervelocity
impacl origiD. However,d€tailedstudiesof thes€cratersrevealedno moltenappearing central regiotr where the glass appearedto have ffowed under impact heating.
The GT-V craterpreviouslymention€dwas, otcource,the singleexc€ption.
Figure 7 showstwo hyp€rvelocityimpactsinto Vycor glass(Conring 7913)that wer€
prcducedwith the explodinglithium-wiredraS-accelerating
gun of the Martir Company.
Glass spheres-60 t in dia. were used as projectiles but theseFnerally broke into smaller
piecesupon launch. The two craters wer€ made by fragments that werc, perhaps, between
4 and 5 p in dia. Someofthe cratersobtairedat the Martin facility did not showthe striking molten-appearingcentml region and wereassum€dto be causedby trash that camedown
lat€r at a slower velocity- lndicated velocities on sorneof th€ faster projectiles ranged up to
20 km/sec. Craters that have the nolten-appearing central regions were assumedto be
causedby tlrc very fast pmj€cliles becauseit has not been possibleto produce this rype of
crater by any other means,
Figure8 showsthe centralregionsof the sametwo cmtersat a highermagnificationand
illustrateshow strikinglyevidentit is that tlle c€ntralregiononceconsistedof glassflowin8
in tle molten state. (Note: Tbe two craters shown are typical of dozens of the Martin
craters. Thesewereforlunately closeenoughtogetlrcr to exhibit both in a singlephotograph.
Thesetwo cratersare also interesting becausethey are in the samesize range and are similar
in appearanceto the GT-Y clater.)
FiSure9 showsthe resultsof a 384-p-dia.Pyrexprojectileirnpactedinto Vycor at 7.37
km/sec. The cnter is -8600 / (0.86 cm) in dia. and 800 t de€p. The glassaround and
closeto the impactpoint appearsto havebeenshatteredand pa ly fused,indicatingconsiderableheat productionduring the impact. In general,evidenceof meltingis conspicuously presentat impact velocities above 7 krn/secfor glassprojectiles. Below this vetocity,
althoughthe glassappearshighlycrushedat the impactpoint, meltingis rot so evident. At
much lower impacl velocities,therc is no evidenceof meltiogat the impactpoint.
Figure l0 showsthe results of a +-in.-dia. glassprojectile impacted into Vycor at -370
mhec-a low-velocityinpact. The c€ntralregionofthis crateris shownat highermagnificationin Fig. I l. The shatteredglassis blockyandangularand showsnoneof the evidence
offusion tlat is apparentin Figs.7-9.
5. DISCUSsIONOF E)(PERIMEIiIAL REST'LIS
When tlrc pits oo the G€mini windows ate compar€d with craters produced under
experimental conditioos in Vycor glassat both high and low velocities, it is seenthat the
Gemini window pits do not even laidtly resemblecraters made by hypervelocity impacts
(with the one exceptionnoted on the cT-V window). The Cemini window pirs do rot even
resemblecratersmadeby impacting proj ectilesat interm€diateandlow velociries,although the
resemblanc€would improve if all of the incipient spall and crushedglasswerc removedfrom
the low-velocity impact claters. The pits on the cemini windows do resembleone classof
pits very closely; namely those left by the glass-polishingprocess. One may then ask why
tlere are signitrcantlynore pits oDthe outsideof the windows tban on the inside. There is no
certain answerto this question. It has be€n found, however,that by gougingat very
958
H. A. ZoOK, R. E. FLAHERTYdd D. J. KESSLER
minute pGtpolish pits with a pin, one can often substantially enlargethem to the point that
tlrcy would app€a. very similar to tle 100-t and larger pits on the windows. The re€ntry
heating and later scrapingwith a razor blade could haveproduced someof the pits obsefl€d
on the exterior surfac€s. Although not all of the 2Gp and lrrger diameter craters we.e
exhaustivelyexamined becauseof the prohibitively large nunber of them, ihe obsewations
previously mentioned also apply to the smaller craters that were so examined.
Thequestionnow aris$ whetheror not lhe contaminationon the windowshashiddenor
dilguisedimpactsof very smallmeteoroids.It is not possibleto be certlin but an educated
guesscan be made basedon the window coniaminationhistory previoudy postulated.
Becausea ilm of silicon oil from tle window sealsappearsto be tle najor (and on GT-8
tlrough CT-12 p€rhapsthe only) sourceof window conlamination while in orbit, the thicknessof tlrc oil contamination should be greaternear tlrc edgeof the wirdows than the center.
After reenty, the baked-on scalewas observedto be thicker near the edge of some of lhe
windows. In fact, c€ntral regions of mary of the returned wiodows \yere entirely free of
scale,and impact characteristicsof micron-sizedmeteoroidswould not be modified if thes€
window areaswere similarly clear in space. Impacts createdby larger meteoroids would, of
course, be lessaflect€d by a thin film. Hence, it s€emsreasooableto assumethat a signilicant fraction (perhaps oDe-half) of the windows was not contaminated severelyenough to
disguiseimpaclsmade by meteoroidswith diametersas small as 4 p. A +p meteoroid,
traveling at 20 km/sec, should produce a cmter in Vycor -100 p in dia- The detectiotr
thr€shold s€t on tle Gemini windows was 100p. It will thercfore b€ assumedin the following analysis that exacdy one-half of the window surface was available for undistorted
meteoroid impacts producing craten 100p and larger in dia.
5. TLUX OBIAINED FROM METEOROIDIMPACT DATA
To obtain th€ flux of parriclesdown to tlle threshold size, the singleimpact is divided by
th€ exposedarca{irne of 3.98 rnadays (or 42.8 ft -days), divided by 0 63 to allow for Earlh
shielding appropriate to the Gemini flights, and multiplied by 2 to account for the assuned
efect of cortaminatiotr. The rcsult is approximately 0.8 meteoroid impacts per m8-dayof
sufrcientsizeto producea cnter in Vycor at least1001in dia.
The massof a meteoroidcausinga craterin Vycor l00l, in dia. must be estimatedby
extrapolating lower-velocity Iarger-masslaboratory data. The results of impacting Pyrex
spheres
384p ir dia.into sectiorsof the GT-V window(madeofVycor) aregivenin Table2.
The projectileswerequite uniform, varyingfrom the meanby lessthan 5 p in dia. and I ]lrg
in rnass. The averageprojecrile masswas63.I pg and the correspondingavemgedensity was
2 13g/cm3.
TscZIF^oDr
olra
v
80?
746
790
800
788
794
799
741
7.37
734
7-37
7.34
7.14
D
210
205
2-06
2.08
205
2-0a
208
6747
7UA
79{r')
8582
7U5
7t76
8911
8.36
922
1008
10.73
922
8.99
1t.16
'Pmelation depth. P, impacL veleiry. I. Elio of p.ne'Etion deprh ro
proicl,le drao.rd, P/4 @ler didetr,
D, tud nrio ot @t.r dimeH ro p€n+
trario' dcprh, D/P, r6petively,
for 384-pdi{ Pyrx proietile
imp&1ing
METEOROTD
IMPACTSON THE GEMINI WINDOWS
959
The Manned SpacecraitCenter empirical penetration formula is usedto extraPolatefrom
laboratory velociti€s aod densitiesto an averagemeteoroid impact velocity ard an averaS€
meteoroidmass. The equationis
pld : constdllrsplz(v cos0)d3
whereP is the d€pthof p€netrationin cm, dis tle equivalentprojec.il€diarneterin cm, p, is
the projectile densily in &/cm3, / is the projectile velocity in km/sec, and 0 is the angle of
incidence. The constant is evaluatedfrom the data in Table 2 for vycor and is 0 45.
Dohnanyi (1966) derived the dishibution in velocity of meieoroids at tlrc toP of t]le
from an analysisofphotographicmeteorobservationsThisdistribution
Ea h's atmosphere
is given by
Il2
' '.", ', -. i r .l "oYr ' "
.ro',u.u
,/:
166km/rc
t 6 b . t // i 2 z k m h e c
where c is the normalization constant. The coresponding averagevelocity is 19 17 km/sec.
Naumann (1966)has adoptedthe velocity distribution ofDohnatryi as apProPriatefor small
meteoroids detectedby the Pegasusand Explorer Pnetration satellites. He also used the
m€teoroidmassdensitydistribution(somewhatmodified)of verniani and Hawkins(1965)
which they obtain€d from radar meteor observations. The modifreddistribution usedwasas
folows : 5I per cent of the meteoroids have a density of 0 37 gi cm3, 45 per cent haYea
density of2 8 gictrr3,and 4 p€r cent have a density of8 g/cm3- Thesedistributrons and averagesarc assumedappropriate for tle Ger ni window Penetrationdata. Normal imPact will
b€ aslumedso tbat cos 0 : 1.
The averagevalue of ,/P from the data in Table 2 is 9 7, hencea loGp-dia. crater
correspondsto a 10'3-tde€p crater. Using the averagevalues above for meteoroids, this
craterwouldbeproduc€dby a meteoroidwith a diameteroi 3 ?2 pandamassof4TT. 10-u g.
The D/P valuesdo not dep€ndvery muchon cratersize,as sho*n by tle datair Table 3.
D r
T^Eu 3. Iw^cr
om nED ^r Tt! MrrN
co@^N'
D
28
m
l5
u
l8
8
IO
t50
65
62
66
3l
400
170
150
2m
r70
55
130
1050
700
3J0
550
,lOO
200
14.38
8.J0
100
8.80
944
6.47
13.0
7.m
lo.Tl
8t3
9.52
6.45
'Peftt aio! dePth, qater dia$etd, and th. ntio of rh*
for impacre of high-v.lcily
two, Bpetiv€ly,
tla$ fragndts
inro vyloi
These data were taken from all of the hypervelocity-appearingcraters on four separate
targets impact€d at the Martin Company hypervelocity facilily. There is a fair amount of
H. A. ZOOK, R. E. FLAHERTY dd D. J. KESSLER
scatt€r in the ,/P data which is probably causedby the iffegular-shap€d projectiles. The
averageof the D/P data in Table 3 is 9. I , which doesnot difrer gre.tly from the 9.7 value io
Table 2. The latter valueis assumedto apply to the cemini windows.
The point defned by the flux* and meteoroid massderiv€d in the previous paragmph is
shown with 95 per cent statisticalconfidencebars in F_ig.12 wherethe logarithm of th€
cumulative flux of panicles having mass,r or larger is plotted vs. the logarithm of lhe mass.
Since no meteoroids impact€d on the uncontaminat€dwindow surface tlat causedcnters
deeperthan tlle single l2-p-deep crater oll the GT-V window (conesponding to a m€teoroid
diameter of 4.3 I and a massof 7 4 . I0 n g), a 95 per cent slatistical conidence l€vel is used
io establish a rcasonableupper limit to the flux of meteoroids larger than this mass. The
other points shown on the smooth curve arc the Explorcr and Pegasuspenetration data as
reducedby Naumann(1966)at tle MarshallSpaceFlight Centerand labeled'fit II' in his
report.
A
E
o
I
L
\
0.0068
I
-r2
-14
-t2
-10
-2
L o gm / 9
Frc. 12. Ie
F (cuMlulvE
Fun) vs. r.oo n (ErEoRorD Ms).
Thc doe i!di@t6 no imtacc larger rhan ih. i.diet.d 'Ms (w tqr).
The smooth curve is dcfioed by the analytical exprcssionshown in Fig. 12. The parameters in tlis exprossionhave beenadjusted to frt the Pegasusand Explorer flux-masspoints
as reducedby Naumann and to extrapolate smootbly, at larger meteoroid mass€s,into the
r The flui colMtiotr factd
Fveo L{ Naumrm (1966).which allos tor nFr@roidsimpaclingat obliqu.
dgles and q(h v.l@iri6 and-deGnris dififtnt frod avedge var6. is -lo F snr'aDd is-nFtlecr;d.
METEOROIDIMPACTSON THE GEMINI \T'INDOWS
961
meteoroid flux-mass curve adopted by Whippl€ (1963). The averagemass density of the
larger photogmphic meieoroids was arbita ly chos€n to be 0.4 g/clnt so that Whipple's
Equation (5) b€comes
totF: -l341o9m - 9.4
where .lt is the flur of particles incidert from 2z stemd. of meteoroids of massm gm and
larger. It is seentbat the flux-masspoint determinedfrom the cemini windows is in excellent ageement witl the Pegasusand Explorer penetration data.
The Explorer \aIII acousticdata (Mccracken, Aloxander and Dubin, 196l) is in serious
disagreementwith the penetration and window data. Although not showDin Fig. 12, tle
Ariel II data (Jennison, McDonnell and Rodgers, 1967)support the Fnetration data_ We
believethe pen€tratioDdata to b€ much more reliable thad th€ acousticdata and, tlrcrefore,
do not us9the acousticdata to fuiuence tlrc shapeof tbe cumulative flux-masscurve.
7. AI'DITIONAL RESULIS OT INTEREST
Once a cumulative flux-mass curve is obtained, various distribution curves of interest
may be derivcd. The derivativeof Fwith resp€ctto log ,r is plotr€dvs. Iogn in Fig. 13.
3oxlo_1
€
I
2'0
I
3 1 . 5
\
\
I
i.
t-
\
.5
o
-t8
-14
Frc. 13. TiE r'EriBmoN
-12
-ro
oF
The logarithmic derivative gives the distdbution of paiicles as a function of log rr and is
convenient fof gmphical reprEsentation. The area uoder this curve is equal to the total
number ofmeteoroids arriving at the surlaceofthe EarthFrmt{ay.
Thiscurvepeaks, on
962
H . A . Z O O K , R . F . I L A H E R T Y a n d D . J . K t s S SEl R
a logarithmicmassinterval, at about l0 0r g. The Pegasus
and Explorerdata havedetermined only the right side of the curve up to about l0 eg- The left side of the curve is
obtainedby extrapolatirgthe analyticalfunctionintroducedin Fig_12.
The distributionin massofthe incomingmeteoroidsas a functjon oflog m is shownin
Fig. 14. The arm under the curveis equalto the total massimpactingrhe surfaceof the
-10
L o gd , 9
Frc. 14- TiB DtsniBUrIoN o
Earth pcr m'-day. Thevalueis 8 S x l0{ 8/m'-dayswhichis almostidenticalto WhipPlc\
to about 48 tons ofmeteoriticmaterialbeingcollectedbl
1967estimate.This corresponds
the Earth eachday. The curvepeaksat about 10r'gwith significantcontributionsfrom
betweenl0 e and 10g The Photographicdata (not includingthe
meteoroidswith mass€s
Prairie Network data) have determinedthe right side of this curve,and the Pegasusand
Explorerpoints havedeterninedthe left side. The regionunder this curvercpres€ntsthe
for masstransportin the lunar-erosionProcess
massregionthat is primarily responsible
work
of
Gault, Shoenakerand Moore (1963),in which
This observationfollowsfrom the
secondary
ejectaproducedby a hype elocityinPact ts
they establishedthat the massof
of
the
incoming
very nearlypropo.tiooal to lhe mass
Primarymeteoroid. It is clea. that,
involv€d,
most of tlrc imPactsr€sponsibl€
impactitg
masses
becauseofthe genemllysmall
in
the
Lunar
Orbiter Photographs
visible
for lunar erosiondo not producecraters
as a functionoflog /, is
meteoroids
area
ofincoming
The distributionin cross-sectional
in
Vycor is relat€dto the
imPact
crater
dimension
ofan
givenin Fig. 15. Because
the linear
Vycor
crater
is relat€dto the
the
area
ofa
ratio
100/1
7
27,
impactingmet€oroidby the
summing'rnder
730
Hence,
by
by
the
ratio
impacting
meteoroid
cross-sectional
areaofthe
per
mr-day
is obtained
arriving
ar€a
ofmeteoroids
cross-sectional
the curvein Fig. 15, the
region
of
chiPped
the
total
area
by
?30,
By
nultiplying
and is 9'07 x loj crnt/nt-days.
area
is
6
62 x
is
derived.
This
glass
maday
unprotected
Vycor
being produced on
Per
METEOROID IMPACTS ON THE GEMTNI WINDOWS
961
5
\
I
-12
-10
\
-2
-8
L o sn , s
Frc, 15. TE DIrouroN
o
l0-i cmg/rn'days, which correspondsto an erosion rate of 2.4 per cent per hundred yearson
Vycor. Fusedsilicachipsouttoa somewhatIargerdiameterthaodoesVycor; n€vertheless,
meteoritic erosion in spaceon an unprot€cted glasssuface, such as a telescopemi(ror, does
not appear to be a serious problem. A catastrophic impacr, causing the mirror to break,
rnay be more lik€ly to linit the lifetime ofa spacetelescope.
The flux of m€teoroidsirnpacting tle surfaceof th€ Earth (at the top of the atmospherc)
at velocity t/6 is decreas€dat I a.u., awayfrom the Earih, by the factor | - V"elV-' wherc
4 is the escapevelocity from the surface of the Earth and ,/- is the velocity of the net€oroid as measued at the top of tbe atmosphere. To affive at the distribution of particles,
area,or massper unit volumeat I a.u.,it is nec€ssary
to multiply the gravitationaldecrease
factot aboyeby 4l(V-2
r"t)l/t, integrateover the velocitydistributiongivenby Dohnanyi
( 1966),and rnultiplyby the ordinateof th€ appropriatedistributionfunctionin Figs. I 3- I 5.
The factor of 4 results from averaging the flux over 4u steradiansto a[ive at th€ above
spatialdensities.The numberof particlesper cm3,the cross-sectional
areaper cm3,and
the massper cmsprovidedby meteoroidsat I a.u.and in ihe eclipiicplaneare obtainedby
summingover the distributionfunctionsand are shorvnin Table4.
T^br! 4. MmoRorD !RoPR
1 a.u, noM rHE sirN
2.2x10\
l.6xt0s
t5xl0,
964
H. A. ZOOK. R. E. FLAHFRTY and D. J. KESSLER
lntegrationoverthe areadistributioncurvecanbe cornbinedwith an albedoto conpare
it with experimentalobservalionsof the zodiacallight. In particular,if it is asslrmed
that
the spatialdensityof meteoroidsdecreases
as / 16 where' is the distancefrom the Sun, tr
Seometricalbedoof 0.057must be assumedto make the zodiacallight and penehation
measurements
agree. This is not an unreasonable
value of geonet.ic albedofor certain
typesof dust grains. The teometricalb€dofor ast€roidsranSesfrom 0.1 to 0 4.
a. coNclusloNs
The Gernini spacecraft windows have furnished an independent measurementof the
near-Earthpenetrationflux for lery small meteoroidmasses.The contamiDationon the
windowshasmadethemlessthan idealmeteoroiddetectors,but thecontaminationevidence
doessuggestthat a reasonable
fraction of the total window areawas not coatedwhile in
orbit. This givessome,thou8h not cornplete,confdencethat the true cemini flux lies
within the 95 per cent statistical confdence bars shown in FiB. 12. The meteoroid flux
obtained from the witrdows is in agreenent witl reasonableextrapolations of all other
p€netrationexperim€nts.The cumulativeflux-masscurve deierminedby the penetration
exp€riments
is in good agreementwith ihe zodia€alligh! measurements.
REFER.ENCES
BoNtER, G. P., HEDI, M. F., KorLA, C. L., LNrorr, J. A., Novmy, J. E., SIlaR, J. W. and Sroc,
R. C. (1968). Postflitht optiql @luaiior olthe right-hand aDd left-hand $irdoes of Gmid Disios
4,5,6 a\d7. NASA T6h. Nor. No. D4916.
DoNAM,
J. S. (1966). Mod€l disaibution or pbotog.rphic nreroF. B. .onn, Inc: Tech. R.p. No.
66-340.I
G^urr, D. E., SEoll,rxn, E- M.ud MooR!, H.J. (1963). Sprayejeted firo the lma su|fae by meteoroid idpact . NASA Tech. Nor. No. D-r767.
JENMSoN,R, C., McDoM,
J. A. M. and RomEs, I- (1967). The Arie! II micrometeorite poetiatio!
Pro. 4. .Sa. Al0O.
MceacKEN, c- w, AffiER,
w. M. aDd DuBrN, M. (1951). Di@t tt'6illttlMts
of intdpldetdy
dusLparuclA itr thc vimry olrbe Eanh. ^/,lJ,l TAh. Not?No.l>1174,
N^unNN, R. J. (1960. The nd'Edlh
ftreoioid enrilor'jMt.
NAS,I Tech. Nor. No. D-3717.
VERNhN, F. and H^wns,
G. S. (t965). Mss,
rognirudes and deositiB of 320 rdio dgteo6. Ednqd
college obs. snithen. Astnphft. ob6. Rddio Meteo, Ptoj.ct L.s. R.p.No 12.
wElpD,
F. L. (1%3). on retoroiG
dd Fnetration. .t. Eeph!'. R.t. 6,4929.
WsrppD, F. L.0%7).
on Minlaining the;deoriii..{,mp1d.
3inithtua. ,lttrcphlt. Obs. Spec. R.p. No.
239.