effect of impact angle on vaporization - semantic scholar · 2017. 5. 3. · effect of impact angle...

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 101, NO. E9, PAGES 21,117-21,136, SEPTEMBER 25, 1996

Effect of impact angle on vaporization

Peter H. Schultz

Department of Geological Sciences, Brown University, Providence, Rhode Island

Abstract. Impacts into easily vaporized targets such as dry ice and carbonates generate a rapidly expanding vapor cloud. Laboratory experiments performed ,•n a tenuous atmosphere allow deriving the internal energy of this cloud through well-established and tested theoretical descriptions. A second set of experiments under near-vacuum conditions provides a second measure of energy as the internal energy converts to kinetic energy of expansiol• The resulting data allow deriving the vaporized mass as a function of impact angle and velocity. Although peak shock pressures decrease with decreasing impact angle (referenced to horizontal), the amount of impact-generated vapor is found to increase and is derived from the upper surface. Moreover, the temperature of the vapor cloud appears to decrease with decreasing angle. These unexpected results are proposed to reflect the increasing roles of shear heating and downrange hypervelocity ricochet impacts created during oblique impacts. The shallow provenance, low temperature, and trajectory of such vapor have implications for larger-scale events, including enhancement of atmospheric and biospheric stress by oblique terrestrial impacts and impact recycling of the early atmosphere of Mars.

Introduction

Experimental studies of shock-induced vaporization typically use flat-plate accelerators in order to establish controlled conditions, to provide direct measurements, and to allow comparison with one-dimensional theory [e.g., see Lange and Ahrens, 1982, 1986; Tyburczy and Ahrens, 1986]. The present experiments examine the transfer of energy to the atmosphere produced by spherical projectiles impacting at different angles, thereby revealing possible differences between one- and three-dimensional models as well as revealing phenomena related to the process. Such studies provide a basis for better understanding the role of collisions on the formation or loss of atmospheres on the early Earth, Mars, and Venus.

The surface of Venus preserves a record of the atmospheric response to the impact process in three dimensions, including signatures of the vapor cloud [Schultz, 1992]. Such signatures become increasingly offset downrange as impact angles (from the horizontal) decrease and indicate that the vapor cloud and atmospheric response decouple from later stages of crater excavation. The energy coupled to the atmosphere can be determined from the limits of vapor-cloud expansion. Taylor [1950a] used fixis approach to estimate the energy yield of the first atomic explosion in New Mexico from high-speed pho- tography, and its validity has since been repeatedly continned in numerical codes of strong explosions in the presence of an atmosphere [Brode, 1955; Goldstine and Von Neuman, 1955; Jones and Kodis, 1982; Vickery, 1986]. Free (undecelerated) expansion of a vapor cloud in a vacuum reflects its internal energy density [Zel'dovich and Razier, 1967]. These two observations (one in an atmosphere and the other under vacuum) provide two equations and two unknowns, thereby giving an estimate of the amount of impact-generated vapor. Previous studies using dry ice as a target [Schultz and Gault, 1990] indicated that vaporization increases with decreasing impact angle (from the horizontal). Because shock eft•cts decrease with decreasing impact angle, they concluded that shear heating must play a significant role for easily vaporized targets. Here we examine the process in more detail for a wider variety of targets.

Copyfight 1996 by the American Geophysical Union.

Paper number 96JE02266. 0148-0227/96/96JE-02266 $09.00

Laboratory experiments cannot directly simulate impact- induced vaporization at large scales. Nevertheless, they allow witnessing processes and phenomena that can be identified and applied at much larger scales through numerical or ana- lytical models. Experiments also permit isolating and model- ing relevant processes that otherwise might be either unex- plained from the geologic record or difficult to reproduce in a full-scale computational code. Comparisons between predictions extrapolated from laboratory experiments and the geologic record on the planets therefore can provide a reality check. For example, scaling the laboratory perspective to Ve- nus was consistent with atmospheric containment and deceleration of a decoupled impact-generated vapor/melt cloud that retains a significant fraction of the impactor momentum [Schultz, 1992]. Numerical models allow constraining parame- ters characterizing the downrange vapor cloud based on observing the process in the laboratory [Sugita and Schultz, 1995].

The discussion below first summarizes the physical context and basis for the experiments. Because impact vaporization depends on impact angle, it is necessary next to review phenomenology associated with vapor-cloud evolution. With fixis background, the observed expansion of the vapor cloud is used to determine its internal energy and vaporized mass. The experimental results indicate that the translational motion of both the impactor and downrange impactor ricochet debris plays an important role in target vaporization through frictional shear heating. Although the amount of impact- generated vapor increases with decreasing angle (from the horizontal), its temperature appears to decrease.

Experimental Approach

The laboratory experiments were performed at the NASA Ames Vertical Gun Range (AVGR), a national facility funded by NASA's Planetary Geology and Geophysics Program and operated through NASA Ames Research Center. The AVGR allows impacting targets over a range of angles in 15 ø incre- ments from 0 to 90 ø (from the horizontal). The large target chamber minimizes possible interference from chamber walls, while large side ports permit recording the evolution of the impact event with high-speed imaging systems. The intensity of self-luminous vapor clouds generated during certain impacts was sufficient to expose high-speed film at 35,000 frames per second and an Imacon imaging system up to one

21,117

21,118 SCHULTZ: EFFECT OF IMPACT ANGLE ON VAPORIZATION

million frames per second. This approach was possible only when highly luminous ionized gases were produced (e.g., • AIO +, CaO, CO, Ca)as determined by a diffraction spectrograph. Less luminous clouds were recorded with backlighting, thereby producing a shadowgraph of opaque phases.

In 1950, Sir Geoffrey Taylor described a relatively straight- forward approach for deriving the energy of an explosion by observing the expansion of the shocked air mass expressed as the fireball [Taylor, 1950a,; also see Sedov, 1959]. This approach assumes that the linear dimension of the disturbance expands without radiative energy loss. Under these conditions, dimensional analysis indicates that the radius R of the spherically expanding shocked air mass generated by an explosion of energy Ea expands over time t according to the following equation:

R = S(7)t2/SEla/Sp• 115 (1) where Po is the atmospheric density and S(7) is a function of the ratio 7 of the specific heats of the surrounding atmosphere. Consequently, when S(7) is calculated, the observed expansion can be used to estimate Ea. Taylor [1950b] derived S(7) expressed as K = [S(7)] -5 for different values of 7 for a nuclear explosion in air and found that 7 was reasonably matched by a value characteristic for room temperatures (i.e., 7 = 1.4).

The atmospheric response from an instantaneous surface explosion is fundamentally different from an impact by a hypervelocity, solid projectile. A surface explosion directly couples its energy with the surrounding atmosphere, whereas an impact first transfers most of its kinetic energy (Yd•) to the target, then to the atmosphere through retarding expansion of the induced vapor cloud or drag deceleration of ejecta [see Cooper, 1977; Knowles and Brode, 1977; Schultz and Gault, 1979, 1982, 1990; Schultz, 1992]. Transfer of energy from impact-generated vapor to the atmosphere, however, can be significantly affected by partial containment within both the impact cavity and the early-time ejecta plume. This containment redirects the vapor initially into a jet, rather than a stationary point-source explosion. With sufficient time and atmospheric interaction, however, such a distinction in source geometry may be lost. The effect of containment and energy losses at early times is defined here by a coupling efficiency factor, k = E/KE.

In the absence of an atmosphere, the internal energy of an isentropically expanding vapor cloud will be converted to kinetic energy with a constant mean expansion velocity simply given by

where Ev is the energy of the vapor and my is its mass. As shown by Zel'dovich and Razier [1967], the observable expansion of the front of a vapor cloud Uma x can be related to u, by the following relation:

uoo y'-I (3) = 2 •r' Umax where 7' is the ratio of specific heats in the vapor cloud.

Consequently, the mass of an impact-generated vapor cloud can be estimated by recording the evolution of two identical experiments (same impactor/target conditions) with (equation (1)) and without (equations (2) and (3)) an atmosphere. Such an approach implicitly assumes that the vapor cloud becomes completely coupled with the atmosphere (Ev = Ea) and that vapor phases dominate this coupling process. Figure la illustrates the expected evolution of the vapor-cloud radius with time, and Figure lb illustrates cloud expansion velocity Uma x as a function of cloud radius. While Figure lb depicts the expected effect for two different cloud energies at a given atmos-

atmosphere-reduced

expansion ......./.' R 5 t-2

undecelerated expansion

LOG TIME

Figure la. Schematic evolution of expanding vapor cloud in an atmosphere. At late times, atmospheric resistance to expansion results in the diameter of vapor cloud expanding as the 2/5 power of time as described by Taylor [1950a, b].

phere, Figure lc contrasts the consequences for a given cloud energy under different atmospheric conditions. Figure lb indicates that Umax also could be estimated with the presence of a tenuous atmosphere if the expansion velocity approaches its maximum free-expansion value well before significant decel-

o o

o

E.• E.c non-isentropic expansion

ß .. ß . e e • ITISX e e eeeee

ß . Urea/ ........ '".. "'"' ..... ":' /E isentropic expansion

'.... acoustic velocity C 0 eeeee

LOG (Cloud Radius) Figure lb. Schematic vapor-cloud expansion velocity as a function of cloud size for clouds of two different energies in a given atmosphere. At early times, the vapor cloud eventually expands at a constant rate, reflecting conversion of internal energy to kinetic energy of expansion. With time, the decreased density of the vapor cloud is decelerated by the atmosphere. Higher-energy vapor clouds (more vapor or greater internal energy) expand at a higher rate and do not feel the effect of the surrounding atmosphere until the cloud has expanded to a greater distance. The outward expansion velocity of the vapor cloud continues to decelerate. An air shock eventually decouples from the vapor cloud and travels outward at the speed of sound Co.

SCHULTZ: EFFECT OF IMPACT ANGLE ON VAPORIZATION 21,119

ß

ß

ß

".. 25po 5.Pc P.o ambient atmospheric p

.....

v e,oo,

LOG (Cloud Radius)

Figure l c. Schematic evolution of vapor-cloud expansion velocity as a function of size under different atmospheric densi- ties. Vapor cloud expands at a constant rate until decreased density is affected by atmosphere.

eration due to atmospheric resistance. Regardless, this approach requires that cloud expansion with or without an atmosphere can be documented at sufficiently late stage (far field) when the specific nature of the energy source can no longer be recognized. Such an approach is often used to ap- proximate the cratering process, e.g., the late-stage equiva- lence of Dienes and Walsh [1970] or the point-source ap- proximation of Holsapple [ 1987].

Experiments allow establishing the degree of energy coupling between the atmosphere and impact as a function of impact angle, the amount of vaporization as a function of velocity and impact angle, and early-time phenomena that might be revealed in the planetary crateting record. The following three different camera systems were used: a NOVA camera recording at 8000 frames per second (fps) on 16-mm film; a Cordin camera exposing up to 35,000 fps on a strip of 35-mm film; and an Imacon camera using an image-orthicon tube capable of 2 million fps yielding up to eight images on 4 x 5 sheet film. Tar- gets of aluminum, compacted pumice, and loose sand provided a reference for atmospheric effects without appreciable vaporization. The available impact velocities (up to 7 km/s) are theoretically sufficient to induce only partial melting of targets such as aluminum [Gault and Heitowit, 1963]. Conse- quently, the effect of vaporization was assessed by using targets of dry ice (blocks and powder), powdered carbonate, powdered dolomite, water, and large Iceland spar calcite crystals as summarized in Table 1. Projectiles were typically 0.635-cm aluminum spheres. Aluminum projectiles allowed using a thin mylar diaphragm at the entrance port, thereby permitting different atmospheric pressures in the impact chamber and mini- mizing effects from launch contaminants. High-speed imaging of the projectile in flight with and without this mylar diaphragm documented both the integrity of the projectile and minimal velocity change.

Phenomenology

Application of (1) through (3) depends on the evolution of the vapor cloud and the atmospheric response. It is not only useful but necessary to examine phenomena observed with and without vaporization and to assess possible effects of impact angle on the energy coupled to the atmosphere.

Vertical impacts. Without vaporization, the ejecta plume quickly (within 1 ms) establishes the conical profile characteristic of impacts into particulate targets under vacuum conditions [see Gault et al., 1968; Oberbeck, 1975]. Impact vaporization, however, produces a dramatically different early-time evolution as illustrated by impacts into dry-ice powder (Figure 2). Immediately (within 100 ms), a self-luminous (ionized) vertical jet of vaporized target material is created. Such jets are pronounced with pyrex projectiles that completely fragment and melt at impact, regardless of target. Si- multaneously, a self-luminous cloud expands spherically from a neck above the point of entry at a velocity of about 700 m/s. Within 0.5 ms, a high-temperature (blue) jet emerges from the cloud with an upward velocity of 1.3 km/s and a lateral expansion velocity of about 200 m/s. As the vapor cloud continues to expand and dissipate, a separate cone-shaped plume of ballistic ejecta develops by 50 ms and extends to late time (not shown in Figure 2). These two examples provide reference for following discussions concerning energy coupling between solid-body impact and the atmosphere at different impact angles at early times.

The observed contrast in early-time plume evolution with and without vaporization reflects the initial energy partitioning between impactor and target. Shock-induced vaporization is created in front of the projectile as it penetrates the surface and laterally as the shock expands into the target. The early- time penetration cavity conf'mes and redirects the vapor into a plume which subsequently evolves into a spherically expanding cloud whose center of mass rises above the cavity at about 1 km/s. The observed blue jet is most likely derived from high-temperature vapor temporarily trapped in front of the impactor as it penetrates the target. Similar "cavitation" (collapse and redirection of the vapor) also can be documented in computational codes depicting impacts of natural materials at much higher velocities [e.g., 07•eefe and Ahrens, 1982] as well as more comparable simulations of aluminum spheres impacting plasticene [Thomsen etal., 1979]. Escape of the vapor is delayed until it fills the cavity. This evolution is dis- tinct from the classical fireball rising above large-scale explosion craters that create the distinctive mushroom shape. Fire- balls from such explosions represent the buoyant rise of hot gases in an atmosphere [see Jones and Kodis, 1982], whereas

Table 1. Summary of Experimental Variables and Abbreviations

V ariab le Abbreviation

9 Targets Dry ice, H, = 5.6 x 10 9 DI Dolomite, H, = 9.5 x 10 dolo

9

Carbonate, H, = 9.5 x• 10 carbo Calcite, H• = 9.5 x 10•o calcite Water, H, = 2.3 x 10 H 20- A

•o

Water ice, H, = 2.8 x 10 H•)-B 10 • Silica sand, H, - 1.8 x S

Projectiles aluminum, 2024-T4 AI aluminum, 606 l-T4 sAl pyrex px

nylon n broken pyrex bpx

Atmospheres Argon A Air a

Helium He

//• is the energy required for vaporization (in ergs per gram).

21,120 SCHULZ: EFFECT OF IMPACT ANGLE ON VAPORIZATION

the expanding and rising cloud in Figure 2 represents expansion of gases in a vacuum initially fed by the upward jet.

Early time coupling between the atmosphere and impact products should be significantly different with and without vaporization. Without vaporization, ejecta leave the cavity as part of the flow comprising the conical ejecta sheet. Unless the surrounding atmosphere can decelerate the head of this stream, energy coupled to the atmosphere will be deferred to later times when ejecta trajectories no longer parallel the plume [Schultz and Gault, 1979, 1982]. This contrasts with the vapor cloud, which expands radially against the atmosphere. Equal partitioning between projectile and target limits the internal energy of the vapor cloud to be less than about 0.5 KE i for vertical impacts by aluminum impactors at 6 km/s [Gault and Heitowit, 1963]. Energy expended in crater excavation (target kinetic energy), however, exceeds the energy fraction in the vapor cloud. This statement will be justified in a subsequent section, where the energy contained in the vapor cloud is estimated from its rate of expansion.

The effect of introducing an atmosphere is illustrated in Figures 3 and 4. Without vaporization (ptunice target, Figure

-...

..

Figure 2. (bottom to top) Evolution of plume for a vertical hypervelocity impact into a block of dry ice under vacuum conditions. In the first frame, a vertical high-velocity jet is observed (arrow). CO and CO2 vapor expands spherically from a rising "ource" in the next frame. A higher-temperature (blue and more luminous) cloud indicated by arrow eventually extends above the spherically expanding component. This component is believed to develop from vapor trapped in front of the projectile during penetration and redirected upward by the growing crater cavity into a jet as shown in the first frame. The intefframe time interval is 125 ms.

Figure 3. (bottom to top) Evolution of ejecta plume following vertical impact of a 0.635-cm aluminum sphere into pumice at 5.5 knds in a 0.9-bar atmosphere of argon. Heated atmosphere in front of and trailing the impactor form a small gas parcel (arrow) that rises upward within the wake Time interval between frames is 57 ms.


Figure 4. (bottom to top) Evolution of self-luminous ejecta plume (vapor)produced by the vertical impact of a 0.635-cm aluminum sphere into dry ice under a 0.9-bar atmosphere of argon. The atmosphere prevents expansion of the vapor cloud, which rapidly rises into the downward moving gases in the projectile wake (contrast with Figure 2). The collision between the upward moving vapor cloud and downward wake creates a well-def'med shock front (small arrows). The heated atmosphere and vapor are surrounded by opaque, cooler ejecta. Time interval between frames is 28 gs.

3), the ionized wake trailing the impactor continues to im- pinge and fills the growing transient cavity delineated by the ejecta curtain. High-speed ejecta and vapor ionize the atmosphere with sufficient illumination to be recorded at 35,000 frames per second. Even low atmospheric pressures (20 mbar) significantly enhance visibility of the vapor through ionization. As this self-luminosity rapidly decreases, a small, more brilliant ionized ball moves upward through the wake column. On the basis of spectra, this minifireball represents heated atmospheric gas trapped in front of the projectile during pas- sage through the atmosphere and released after the projectile has transferred most of its kinetic energy and momentum to the target.

Use of easily vaporized dry ice under an atmosphere reveals significant enhancement in luminosity (Figure 4). Without an atmosphere, vertical impacts into dry ice produced only a faint image at first contact using the same exposure (28 gs). With an atmosphere, the visible effects of vaporization increase through rapid heating and ionization of the atmosphere and vapor cloud components, in contrast with ballistic ejecta flow. The vertical ionized jet observed for impacts into dry ice under a vacuum changes under atmospheric conditions. In an atmosphere, the jet is replaced by a fight ballistic fireball which emerges above the cavity with time (Figure 4). As it passes through the column of ionized atmosphere comprising the projectile wake, multiple standoff shocks develop. This interaction demonstrates the rapid inward collapse and downward flow of wake gases behind the projectile as previously described [see Schultz, 1992]. At this scale and atmospheric conditions, the rising ionized ball is completely contained within the developing ejecta curtain.

The evolution of the vapor plume and ejecta under low atmospheric pressure was also examined using back illumination recorded at 35,000 frames per second. Under very low atmospheric pressures, the opaque phases initially form a conical plume which "blooms" as it rapidly changes from a cylin- drically expanding column to a spherically expanding cloud. Eventually, the cloud engulfs the surface as illustrated in the second frame in Figure 2. Under an atmospheric pressure of 2 mbar (air), the opaque conical plume is contained within a less opaque but more vertical plume. The less opaque plume represents the detached mach cone from the projectile, and the base of the less opaque plume forms a small kink that rises with time. As the projectile air shock expands and interacts with the vapor plume, it creates a turbulent eddy. At much later times, the eddy forms a "crown" above an opaque spherically expanding cloud whose center gradually rises.

In summary, the degree of direct coupling between the kinetic energy and the surrounding atmosphere from a vertical impact immediately after contact in laboratory experiments strongly depends on the amount of impact vaporization. Without vaporization, energy coupling is inefficient since it involves deceleration of ballistic ejecta within a flow stream. With vaporization (but without an atmosphere), the rapidly expanding cloud outpaces the advancing ejecta curtain, which reflects energy partitioned to kinetic energy in the target. In a dense atmosphere at laboratory scales, expansion of the "cool" vapor from dry ice is contained within the conical ejecta curtain. Two vapor components can be recognized: a lower temperature cloud that resembles expansion of gases from a noz- zle and a higher temperature cloud corresponding to highly compressed gases initially trapped in from of the projectile. The latter component is redirected by the transient cavity into a vertical plume. At laboratory scales, such interactions occur within the first 0.1 ms, which corresponds to less than 0.1% of the time for crater formation.

Oblique impacts. As impact angle decreases (referenced to horizontal), the early time plume changes significantly. When little vaporization occurs (e.g., impacts into pumice or sand at 15'), the high-frame rate camera (35,000 fps) records only a


small flash at impact (Figure 5a). With vaporization (e.g., use of dry ice or carbonate targets), however, a highly luminous and rapidly expanding cloud moves downrange (Figure 5b). Three separate self-luminous components can be documented under vacuum conditions for a 15' impact. First, a jet with a velocity 2 to 3 times the initial impactor velocity rapidly travels downrange just above the surface as documented in previous experiments [Gault and Heitowit, 1963; Gault et al., 1968] and numerical analyses [e.g., Kief•r, 1977; Vickery, 1993]. ̂ second self-luminous component accompanies the hypervelocity fragments of the original projectile that ricochet downrange. Third, a hemispherical vapor cloud expands and moves downrange at a velocity dependent on impact angle. All three components decouple from the point of impact and evolve well before development of the ejecta curtain for impact angles less than about 45'. The total luminosity for impacts at 15' was obsexxed to be enhanced a factor of about 4 greater than impacts at 30' for water targets.

Evolution of the three-component vapor cloud changes with increasing atmospheric pressure. Under low pressures (<7 mbar) in the experiments, the atmosphere simply retards outward expansion and downrange motion of the hemispherical

cloud. At higher atmospheric pressures, the three components combine downrange in a single, jet-like component (Figure 6). At laboratory scales, the atmospheric pressure retards vapor cloud evolution (expansion and downrange motion) more rapidly than the high-speed jetting and ricochet. Containment and redirection of a portion of the vapor cloud by the evolving impact cavity also occur at low angles. Figure 6 reveals a small jet-like plume uprange that is visible only under high atmospheric pressures. This blue plume emerges at an angle higher than the impactor entrance angle. Consequently, the uprange plume is not controlled by the high-temperature cor- ridor created by the ionized wake of the impactor but more likely corresponds to ionized gases redirected by the evolving penetration cavity. The plume becomes more vertical with time as the cavity shape becomes more symmetrical. In higher- angle impacts, the deepening crater cavity prevents free expansion of the vapor cloud.

In summary, low-angle impacts (<30') into easily vaporized targets produce an early-time vapor cloud that decouples from the later stage crater growth and travels downrange ballisti- cally. Use of a tenuous atmosphere restricts expansion of this cloud but is insufficient to decelerate individual, high-speed, submicron solid ejecta over the same distance.

Figure 5. (a) Oblique impact (15') by 0.635-cm aluminum'sphere at 5.2 km/s into a pumice target under near- vacuum conditions (<0.7 mbar). Only a small transient flash is produced. (b) Oblique impact (15' from fight) by 0.635-cm aluminum sphere at 5.3 km/s into a dry ice target under low atmosphere (1 mbar) revealing brilliant, self-luminous vapor cloud expanding at a velocity of about 2 km/s while moving rapidly downrange (3 km/s). Vapor cloud completely decouples from the crater cavity and later excavation of crater Time interval is 57 gs; sequence shown is from bottom to top.

SCHULZ: 'EFI•ECT OF IMPACT ANGLE ON VAPO••ON 21,123

Vapor-cloud energy. An expanding cloud in the presence of an atmosphere should exhibit growth (cloud radius) depending on the 2/5th power of time (equation (1)). Figure 7 illustrates the typical cloud evolution for 15 ø impacts into a dry-ice target at different velocities. Expansion does not fit the power law growth at early times but reflects unimpeded expansion at early times (Figure lb). The surrounding atmosphere eventually slows expansion to match the predicted rela- tions after 0.1 ms, and its dimensions depend on impact velocity as expected from Figure 1 a.

Although the observed evolution of the disturbed atmosphere is consistent with theory, the energy source needs not reflect a vapor cloud but only efficient coupling between impact products or heating of the surrounding atmosphere. Comparisons between easily vaporized and refractory target materials, spectra of the self-luminous cloud, and use of low ambient pressures (less than 33 mbar) are all consistent, however, with expansion of a vapor. As a further test, hypervelocity debris-cloud collisions were performed in order to maxi- mize coupling with the atmosphere rather than the target, analogous to near-surface explosions [Schultz and Gault, 1982, 1985; 07(eeJb and Ahrens, 1982]. For these experiments, pyrex spheres were catastrophically disrupted by pas- sage through thin (2.54 x 10 -3 cm) paper above the target. The resulting debris cloud expands at a rate controlled by stored strain and rotation rate imparted by the rifling in the launch tube used to separate the projectile from the sabot. Vertical impact by this cloud into an aluminum target establishes the energy directly transferred to the atmosphere principally by heating and radiation since small ejecta will escape deceleration at early times. From the observed growth of the intensely ionized cloud (Figure 8a), only 0.8% of the initial impactor

Figure 6. Oblique impact (15' from right) by 0.635 cm aluminum sphere at 5.3 km/s into a dry-ice target under a 0.9-bar atmosphere of argon (from bottom to top). Presence of atmosphere prevents vapor-cloud expansion but does not signifi- 4 canfly arrest downrange motion. Time interval is 57 gs. Three components of the vapor plume can be recognized. A jet-like plume uprange becomes more vertical with time and most g: likely reflects evolution of the transient cavity. During initial penetration, high-temperature vapor is directed back up the rr trajectory, but as the cavity grows, it opens and redirects vapor upward. Most of the vapor, however, decouples from the growing crater, either moving rapidly downrange with the impactor ricochet or expanding above the impact. Lower temperature particulates form an opaque cloud that partly ob- scures the self-luminous vapor.

Energy Partitioning and Vaporization

Free expansion of the impact-generated vapor cloud in the presence of an atmosphere allows estimating its energy from (1), while expansion in a vacuum for identical impactor conditions permits calculating an upper limit for the vaporized mass from (2)and (3), once the energy of the vapor cloud is known. The following discussion first compares the observed vapor cloud expansion with theoretical predictions, thereby allowing estimates of the contained energy. Second, these results are tested for conservation of energy and varying impact conditions. Estimates of impact-vaporized mass are then com- pared as a function of impact angle.

Time (ms) 0.001 0.01 0.1 1.0

Dry ice • v = 6.2 km/s

P = 10mb (Ar) v = 4.5 km/s

(3 p = 30mb, air ß v = 4.0 km/s

P = 28mb, air

5/2 Log R = 1/2 Log (Ev/Kpo) + Log to

///e •e/

ß

i

-6 -5 -4 -3

Log (time, sec)

Figure 7. Expansion of vapor-cloud evolution produced by oblique (15 ø from horizontal) impacts into dry ice in a tenuous atmosphere. The vapor cloud rapidly expands initially until decelerated by the surrounding atmosphere. The three parallel lines indicate expected expansion rate from theory (see Figure l a) for the three different impact conditions.


oo ß

P=8.7 mb air

-'5 -2

90 ø (vertical)

Ec=0.8% KE i

AI target

i

-3

Log time (sec)

Figure 8a. Expansion of fireball produced by broken pyrex (bpx) debris cloud impacting into a solid aluminum plate. The hypervelocity debris cloud transfers its energy near the surface and creates only small pits in the target. Expansion of the fireball in an atmosphere indicates that it represents only about 0.8% of the initial kinetic energy, thereby indicating that f'me ejecta and atmospheric heating by radiation contribute little energy to the fireball.

kinetic energy is coupled to the atmosphere. Debris cloud impacts into dry ice, however, transfer energy to the atmosphere principally through deceleration of expanding vapor (Figure 8b). These experiments indicate that more than 50% of the impactor energy is transferred to the atmosphere. Note that a vertical impact by a solid aluminum impactor couples only 0.2% of its initial energy. Consequently, the energy represented by the expanding ionized cloud is believed to reflect coupling at a molecular level, consistent with vapor production.

Spectra of the vapor were obtained using a photographic diffraction spectrograph oriented such that the downrange luminous vapor cloud would pass across the slit (Figures 9a and 9b). Because of the translational velocity and expansion of the cloud, the overall exposure corresponds to about 150 gs. As shown in Figure 9a, the vapor cloud produced by impacts into carbonate contains prominent A10 + bands from rapid reactions with the impactor as well as emission lines of neutral species (Ca I) and dissociation products (CaO and CO) from the target. A strong background gray-body/blackbody exists at longer wavelengths (above 550 nm). Impacts into dry ice (Figure 9b) produced simpler spectra with strong A10 + and CO emissions. Consequently, rapid dissociation and mixing between impactor and target can be documented. Both impacts resulted in Na I emissions from contaminants. Such experiments indicate that time-resolved, high-resolution spectrophotometry could be a valuable tool for probing the compositional and thermal evolution of impact-generated vapor clouds.

Crater growth disrupts free expansion of the vapor cloud at high impact angles (>45'). Figure 10a compares the downrange and above-impact components for hypervelocity impacts into dry ice blocks at 45' under low atmospheric pressures, whereas Figure 10b compares them for high atmospheric pressures. At lower pressure, free expansion of the downrange cloud significantly outpaces the above-impact component and detaches from the crater lip about 35 Ixs after impact. Neverthe-

less, the above-impact spherical cloud follows the expected growth rate. At higher pressure, the downrange and above- impact components merge to form a more complicated evolution in advance of the crater lip (Figure 10b). Free expansion is deferred to later times. Vertical impacts even at lower pressures result in partial cavity containment of the vapor that subsequently breaks away and expands spherically in advance of the ejecta plume (Figure 10c).

Figure 11a shows the observed expansion velocity as a function of stage of growth and allows further comparison with theory shown in Figure lb. The expansion velocity is initially very high but decreases to a nearly constant value beyond about 10 impactor radii. Under low atmospheric pressures (---10 mbar), the expanding cloud overwhelms the surrounding gas and continues to expand at a rate dependent on its internal energy. The constant expansion velocities for vapor from impacts into both powdered carbonate and dry ice should be expected for gases with similar bulk composition, i.e., CO or CO 2. Figure 1 lb shows the evolution of velocity as a function of cloud radius for two different atmospheric pressures and again shows that a constant velocity is established after the cloud has expanded about 10 times the projectile diameter. Both Figures 11a and l lb reveal that the surrounding atmosphere eventually decelerates the expanding vapor, consistent with expectations from Figure lb and l c.

As a further test, experiments were performed in the absence of an atmosphere (i.e., pressures below 7 mbar) in order to determine vapor-cloud expansion for different impact velocities (Table 2). Figure 12a shows the expansion velocity of the vapor cloud at different stages of growth for a given impact an-

, / ,

bpx @ 5.2 km/s P=30 mb air

/

AI @ 5 km/s

P=24 mb air

DRY ICE 90 ø

I

-5 -4 -3 -2

Log time (sec) Figure 8b. Comparison of vertical impacts by broken pyrex (bpx) and solid aluminum sphere (0.635 cm) into dry ice. The aluminum impactor generates little coupling with the atmosphere (0.2%), whereas the broken pyrex results in 50% of the initial energy contained in the vapor cloud. This comparison suggests that atmospheric coupling occurs at a molecular level, consistent with a cloud composed of vapor which is enhanced in the broken pyrex experiment. The two different symbols for the solid aluminum represent two different imaging records of the same event at high (dots correspond to 35000 fps) and low (crosses correspond to 9200 fps) framing rates.


wA•E'LENG.TH, nm

-400 420 440 460 .. t.. .I .. ! .

..... ß ::'.. ' :.:;:i' ':: ß ..•':•... •:½•? ....... ' .... :.F '":-:.},;• '• ................ ';':' ........... :'

ti•...,i':i: . .•.. .

,:,.•::.

CaI

'Alurni'num "into CaCO3

480 5:00 .520 540 560 .:580 600 620 640 .i '-:*:.... .......... :.: .... :t::-: '-:.::::.': .........

................... 7'"" :•'•'7•!" :•::

• •5.•: .• .... •.:::..: .:-.•.; :: •..

- :

"?

co NaI

Figure 9a. Spectra produced by a 5.98 km/s impact (0.635-cm sphere) into calcium carbonate at 15' from the horizontal. Slit was oriented perpendicular to downrange motion of the vapor cloud and was exposed for the duration of the event through a side-viewing port. Consequently, variation in intensity along the slit vertically represents the time-integrated position of the vapor cloud above the target. The overall intensity therefore reflects the luminosity for the 150-Bs exposure created by the lateral motion of the vapor. The reference lines on top are from Hg and correspond to the prominent 405/408 nm doublet (left) to the 546 nm and 577/579-nm line pairs (fight). The vapor-cloud spectra reveal that the luminosity is created by strong A10 + bands between 460 and 530 nm and blackbody/gray body radiation at longer wavelengths (above 550 nm). The strong doublet of neutral Na (Na I) can be recognized at the fight (590 nm), while atomic emission Ca I is identified to the left (423 nm).

WAVELENGTH, nm

400 420 440 460 480 500 .520 540 560 580 600 620 640

Figure 9b. Spectra produced by a 5.5 km/s impact (0.635-cm aluminum sphere) into dry ice at 15'. The strong A10 + emission seen for the carbonate target (Figure 9a) is again recognized, thereby indicating that it results from rapid impactor/target chemical reactions in the vapor cloud. The Na doublet at 590 nm is the most prominent emission. Additional emission (and absorption) bands can be recognized, thereby indicating a wealth of additional information that could be obtained from high-speed, high-resolution spectrophotometry.

21,126 SCHULZ: EFFECT OF IMPACT ANGLE ON VAPO••ON

3.0

2.0

1.0

downrange

(free expansion)

-5 -2

? ,,'•above impact

(delayed expansion)

,++-% crater growth

DRY ICE 45'

P= 4 mb i i

Log time (sec)

Figure 10a. Comparison of vapor-cloud expansion (radius R) downrange and above the point of impact with crater growth for a 45 ø impact angle under low atmospheric pressure (3 mm Hg of air). The downrange vapor cloud rapidly decouples from target excavation and undergoes free expansion. The vapor cloud above the impact point contains less energy and interacts with the crater cavity and solid ejecta.

gle (15') and velocity (5 km/s). Different target materials yield slightly different expansion velocities over a range of dis- tances from the point of impact. The nearly constant expansion velocity through time indicates that almost all of the internal energy has been converted to kinetic energy of expansion. Figure 12b reveals that the observed expansion velocity

3.0

2.0

1.0

-/•atm.-retarded free . © expansion expansion

©/ /, combined components

++ '"X_. crater growth

DRY ICE 45'

P= 24 mb i i i i i

-5 -4 -3 -2

Log time (sec) Figure 10b. Vapor cloud for 45 ø impact (5.55 km/s) into dry ice in an atmosphere (18mm Hg air) At early times (< 0.1 ms), crater growth interferes with the vapor cloud but eventually gives way to free expansion until contained by ambient atmosphere.

3.0

1.0

.j br ak•.w y • sp•erlc.a• expansion

partial cavity containment

+

+4- ++4-

DRY ICE 90'

P=4 mb

i

-5 -4 -3

Log time (sec)

Figure 10c. Vapor-cloud expansion for 90 ø impact (5.32 km/s) into dry ice at low atmospheric pressures. Breakaway free expansion occurs at later times and results in "blooming" of the plume.

exhibits little dependence of in-vacuo expansion velocity on impact velocity at a given impact angle above a limit reflecting onset of complete vaporization. This absence of a dependence should be expected if the energy partitioned to vaporization is a constant for a given impact angle while the vaporized mass (in projectile masses) increases as vi 2, which is expected from theory [e.g., O'Keej• and Ahrens, 1977]:

Ev = k KE i (4a)

m u 2 =km v. 2 (4b) v• pt

1.0

R/r 10 20 40 60 100 200

0.5

o.o

-0.5

P-' 26 mb

dolomite target

'"'--•;., \, Q •,(• ,'--- v = 5.5 km/s

\,

f\.GG v = 4.8 km/s • •

6

0.8 '•

0.6" v

0.4

0.2

0.5 1.0 1.5 2.0

Log (R, cm)

Figure 11a. Vapor-cloud expansion rate for two impact velocities at 15 ø under a given atmospheric pressure. The vapor cloud expands initially at very high velocities but eventually approaches a constant value before being decelerated by the surrounding atmosphere (see Figure lb).

SCH[R,TZ: EFFECT OF IMPACT ANGLE ON VAPO••ON 21,127

1.0

R/r

10 20 40 60 100 200 10 20 , 1.0 ,

0.5

-0.5

Dry ice targets

vexl• ~ R 'lJ '"'. P --- 8 mb air • v = 5.1 km/s

Pv = 37rob (.Argon) = 4.9 km/s

!

0.5 1.0 1.5

Log (R, cm)

6

0.6 •

0.4

0.2

2.0

Figure lib. Vapor-cloud expansion rate from 15 ø impacts into dolomite and carbonate under increasing atmospheric pressures as a function of stage of growth (cloud radius). Compare with Figure lc.

0.0

R/rp

40 60 100 200

Dry ice Dolomite

Water ice

15 ø from horizontal

400

i]max(km/s)

ß Q' © © 0OO (:•). 0 0 • 2.2 Dolomite ß .. • 2.0 Dry ice ß •'e e' • 1.8 Waiter ice

I ,I

i 10 100

CLOUD RADIUS (cm)

Figure 12a. Comparison of maximum vapor-cloud expansion velocity Urea x as a function of target type for a given impact angle (15ø). Data represent constant expansion velocities before any effects of atmospheric deceleration and are plotted against (top) cloud radius R scaled to projectile radius r and (bottom) corresponding cloud radius. Different average cloud radii result from relative timing between first paired images and moment of impact. Different target types exhibit different maximum expansion velocities, consistent with energy required for vaporization. Maximum expansion velocity is achieved within about 15 projectile radii.

Table 2. Asymptotic Expansion Velocity of Vapor Cloud After Expansion Beyond 25 Projectile Radii Prior to Atmospheric Containment Under Low Ambient Pressure (<8 mbar)

v, m, Target 0, u u. km/s g, type Type deg KE 10 km/s km/s

860827 4.55 0.376, AI DI 15 3.89 2.00 0.66 860826 4.72 0.376, AI DI 15 4.19 2.00 0.66 860825 5.11 0.376, AI DI 15 4.91 1.80 0.59 870202 5.6 0.376, AI DI 15 5.89 2.00 0.66 870203 6.17 0.375, AI DI 15 7.14 2.00 0.66 890206 5.55 0.376, AI DI 45 5.79 2.30 0.8 870832 5.59 0.376, AI dolo 15 5.86 2.20 0.73 870835 5.54 0.376, AI dolo 15 5.75 2.20 0.73 870836 5.44 0.376, AI dolo 15 5.55 2.00 0.67 870839 5.29 0.376, AI dolo 15 5.29 2.10 0.69 870838 5.24 0.376, AI dolo 15 5.15 2.00 0.66 901201 4.89 0.375, AI dolo 15 4.48 1.70 0.56 901202 4.26 0.375, AI dolo 15 3.4 1.10 0.33 901203 5.33 0.375, AI dolo 15 5.33 2.30 0.76 901204 5.23 0.375, AI dolo 15 5.13 2.30 0.76 901205 5.22 0.375, AI dolo 15 5.1 2.30 0.76 901209 5.34 0.375, AI dolo 15 5.35 2.30 0.76 870833 5.64 0.376, AI dolo 15 6 2.20 0.73 930705 5.19 0.375, AI dolo 30 5.05 2.50 0.82 931202 4.92 0.376, AI dolo 45 4.6 2.80 0.96 931206 5.16 0.376, AI dolo 45 5 2.80 0.96 870215 5.78 0.376, AI carb 45 6.3 2.50 0.82 870640 5.48 0.376, AI carb 30 5.6 2.50 0.82

Abbreviations are defined as follows: v, impact velocity; m, projectile mass including type (see Table 1); target types (see Table 1); 0, impact angle measured in degrees from the horizontal; KE 10, impactor kinetic energy x 10 •ø ergs; Um•, maximum observed expansion velocity of vapor cloud; u., derived constant mean expansion velocity (from (3)) with a value of 7 = 1.28 (cooled (202 vapor).

m--• I-•-I 2 Ev ] (4c) ß •=k'vi • (4d)

where uoo 2 can be simply given in •ms of •e observed ma•m• ex•sion velocity umax of •e vapor cloud for is- endopic conditions wi• •e •e energy •ifioned • va- •dzation, E•t, (equil©ns (2) •d (3)) •at de•nds on •- get mat©hal for a given •pact •gle (Fig• 12a).

ß dry ice 0 dolomite

J . ©

I , ..0 s'.0

Impact velocity (km/s)

Figure 12b. Comparison of maximum vapor-cloud expansion velocity for 15 ø impacts into dry ice and dolomite under low atmospheric pressures. Impacts into dry ice result in a nearly constant expansion rate over a wide range of impact velocities, whereas impacts into dolomite result in a gradual increase in expansion velocity for impact velocities below 5.3 km/s but a constant velocity for higher impact velocities. Re- sults indicate that the energy partitioned to the vapor cloud reaches a maximum above a minimum impact velocity.


Derivation of Ev from (1) requires specifying the constants in (1) and solving from the observations shown in Figures 1 l a and 11 b:

5/2 logR o = 1/21og(EA/KPo) + log t o (5) where Ro, po, and to correspond to observed values and where it is assumed that Ev eventually couples completely to internal energy of the atmosphere E,4. As previously noted, K depends on the adopted ratio of specific heats for the expanding gas above the impact.

Different ambient atmospheric pressures and compositions were used in order to test the repeatability of the derived energies and revealed that ambient pressures must exceed about 5 mbar to provide consistent results owing to measurement uncertainties. The value of K in (5)could depend on the surrounding ambient atmosphere, the expanding vapor cloud, or a mixture of the two. As a test, the energy fraction coupled to the atmosphere was derived for dry ice targets under atmospheres of different compositions. If the value of K reflects only target vapor, then the derived energy should not depend on atmospheric composition. Instead, atmospheric composition does make a difference. Just as Taylor [1950a] concluded, the value of K apparently reflects conditions of the ambient atmosphere, even though the expanding gas may be highly ionized.

Direct coupling between the impact and atmosphere at an impact angle of 15' allows deriving the maximum energy partitioned to vaporization as shown in Figures 13a and 13b and summarized in Table 3 for targets including dry ice blocks, powdered carbonate, Iceland spar, calcite, and dolomite (Ca_Mg(CO3)2). The initial impactor kinetic energy obviously represents the maximum possible energy that can be coupled with the atmosphere and hence partitioned to vaporization. In reality, however, ricochet of the projectile downrange cames away a significant fraction of the initial kinetic energy [Gault and Wedekind, 1978; Schultz and Gault, 1990]. For a 15' impact angle, this fraction is about 40%. Consequently, a maximum 60% of the initial impactor kinetic energy is actually partitioned to the target for vaporization, melting, comminu- tion, and displacement. The vapor cloud from dry-ice and carbonate targets (Figure 13a) represents about 10-15% of the initial impact energy or about 15-25% of the available energy transferred to the target. Figure 13a also reveals that the energy partitioned to vaporization increases with increasing velocity up to a maximum velocity, above which the fraction of

ß 10

o

•' 0

15 ß impact angle

/, Dry Ice '•. jl T J• Dolomite ß

I

Impact velocity (km/s)

Figure 13a. Energy fraction contained in expanding vapor cloud for 15 ø impacts into dry ice and dolomite as a funtion of impact velocity. Vapor cloud (from dry ice targets) represents nearly constant 15% of the initial kinetic energy for impact velocities greater than 4.5 km/s. Error bars represent maximum measurement uncertainty.

Impact Angle 45 ø 30 ø

ß dry ice

© dolomite

@ carbonate

• calcite

15 ø 7.5 ø , 100%

10%

1%

-3 0.1% -0.4 -0.3 -0.2 -0.1 0.0

Log (cos 2 e)

Figure 13b. Energy fraction contained in expanding vapor cloud for different impact angles and different target types. Energy fraction increases with the translational velocity component (cos20), in contrast with expectations if only peak shock pressure (i.e., sin20 for the vertical component) controls the process.

the initial impactor kinetic energy plateaus. The maximum velocity for dry ice is about 4.5 km/s but is about 6 km/s for carbonates. The ratio of peak pressures for these two velocities (1.8) approximates the ratio of energy required for vaporization (1.7) for the two target types.

As a reference for the particulate targets, two experiments used solid carbonates, one, a single large crystal of Iceland Spar calcite, and the other, a block of limestone (Kaibab). The calcite crystal appeared to result in greater vaporization, whereas the limestone block was less than the particulate targets. The enhancement for the calcite may indicate higher pressures created at first contact. The reduction for the limestone block, however, is attributed to incomplete energy transfer since some of the ricochet missed the block downrange. The possible contribution of the impactor ricochet to vaporization will be discussed in a subsequent section.

The derived vapor-cloud energy depends on impact angle for impact velocities a!½ove 4 km/s. Figure 13b reveals that Ev/KE i increases as cos'O for impact angles decreasing from 45' to 30'. This result contrasts with th• expected peak shock pressure, which should decrease as sin 0 as indicated by the sin'O dependence on cratering efficiency [Gault and Wedek- ind, 1978]. For an impact angle of 7.5', the vapor-cloud energy plummets to only 1% for dry-ice targets. In this case, kinetic energy retained by the ricocheting projectile approaches 90% [Schultz and Gault, 1990]. For vertical (90') impacts, the vapor-cloud energy decreases to about 0.05% for carbonates (powdered and calcite crystals) and 0.15% for dry ice (Table 3). The derived energy fraction coupled to the atmosphere for impact angles above 30' is less certain since direct coupling between the vapor cloud and ambient atmosphere is reduced owing to both temporary containment within the transient cavity and the interacting effects of the ejecta curtain. The effect of containment was minimized by impacting small, solid targets (e.g., dry-ice cubes or small particulate samples), thereby allowing measurements of the early-time free expansion of the vapor phase.

In summary, vapor-cloud energy increases with decreasing impact angle, in contrast with the expected effect of decreasing in peak shock pressures. This increase is not believed to reflect enhanced atmospheric coupling due to submicron solid ejecta or thermal effects because complementary experiments maximizing such effects did not exhibit enhanced coupling


Table 3. Impact Energy Coupled to Atmosphere

P, v, rn, Target 0, mm Hg km/s g, type Type deg KE 10 b K E 9

860817 21.5, a 3.01 0.376, AI DI 15 1.7 6.72 0.856 0.84 860818 21.5, a 4.04 0.376, AI DI 15 3.07 7 0.856 3.3 930914 17.5, A 4.31 0.376, A1 DI 15 3.48 7.17 0.54 4.8 930911 22.5, a 4.5 0.376, A1 DI 15 3.8 7.12 0.856 5.7 860819 28, a/A 5.06 0.376, A1 DI 15 4.81 7.13 0.62 9.3

890207 17, a 5.2 0.290, px DI 15 3.92 7.06 0.856 3.3 870202 6.6, A 5.6 0.376, AI DI 15 5.89 7.51 0.54 14 870203 7.5, A 6.17 0.375, A1 DI 15 7.14 7.5 0.54 16 870207 6.6, A 6.45 0.0459, AI DI 15 0.955 7.11 0.54 2.2 860843 7.0, A 5.44 0.376, A1 DI 7.5 5.56 6.95 0.49 0.66 860844 6.6, A 5.64 0.376, A1 DI 7.5 5.98 6.98 0.54 0.76 870637 15.5, a 5.27 0.376, AI DI 30 5.22 7.25 0.856 7.1 890205 18, a 5.13 0.376, A1 DI 45 4.95 6.73 0.856 0.753 931107 18, a 4.45 0.376, A1 DI 45 3.72 6.69 0.856 0.63 931108 18, a 4.95 0.376, AI DI 90 4.6 6.49 0.856 0.26 890203 18, a 5.11 0.375, A1 DI 90 4.9 6.3 0.856 0.11 870636 15.5, a 5.57 0.376, A1 DI 90 5.82 6.3 0.856 0.09 850917 6.6, a 1.83 0.463, n DI-P 90 0.776 5.85 0.856 0.0048 870636 15.5, a 5.57 0.155, n DI 90 2.4 6.3 0.856 0.09 930703 20, a 4.84 0.376, A1 dolo 15 4.4 7.1 0.856 4.6 901210 18, a 5.54 0.375, AI dolo 15 5.76 7.2 0.856 6.6 901213 18, He 4.82 0.375, A1 dolo 15 4.36 6.85 0.54 0.11 871216 22, a 5.2 0.372, sAl dolo 15 5.03 7.12 0.856 5 930704 20, a 5.17 0.376, A1 dolo 30 5.02 7.16 0.856 6.1 931105 18, a 5.05 0.376, AI dolo 45 4.78 6.46 0.856 0.22 931101 18, a 4.3 0.376, A1 dolo 90 3.48 6.39 0.856 0.16 931102 18, a 4.84 0.376, A1 dolo 90 4.4 6.55 0.856 0.33 941011 18, a 4.54 0.376, A1 dolo-A1 15 3.87 7.07 0.856 10 941201 14, a 4.66 0.375, A1 dolo-A2 15 4.08 7.12 0.856 8.7 941203 15, a 5.1 0.376, AI dolo-A2 15 4.88 7.20 0.856 11 941208 16, a 5.98 0.375, AI dolo-B 15 6.71 7.30 0.856 13 941014 18, a 4.98 0.375, AI dolo-A2 30 4.65 7.34 0.856 27 880620 17, a 5.45 0.376, AI H20 15 5.58 7.10 0.856 3.9 870212 7, A 5.82 0.375, AI H20-A 15 6.36 7.28 0.54 3.2 900826 14.5-a 4.71 0.376, A1 H20-B 15 4.17 6.91 0.856 1.4 860820 21.5, a 4.83 0.376, A1 S 15 4.38 6.10 0.856 0.049 860824 6.6, A 5.11 0.376, AI S 15 4.9 6.50 0.54 0.063 870211 7, A 5.94 0.376, AI carb 15 6.62 7.51 0.49 8.5 870640 17, a 5.48 0.375, A1 carb 30 5.63 7.28 0.856 9.00 870215 6.6, a 5.78 0.375, A1 carb 45 6.26 7.03 0.856 1.1 870638 15.5, a 4.96 0.376, AI carb 90 4.62 6.00 0.856 0.22 870639 17, a 5.51 0.375, AI carb 90 5.69 6.34 0.856 0.12 870642 16, a 5.04 0.376, AI calcite 15 4.78 7.28 0.856 8.4 870644 14.5, a 4.93 0.376, AI calcite 90 4.56 6.40 0.856 0.11 870214 7, a 5.86 0.375, A1 AI 15 6.44 6.31 0.856 0.42 870216 7, a 6.44 0.154, n AI 15 3.2 6.85 0.856 0.51 870641 15, a 5.20 0.376, AI A1 15 5.09 6.52 0.856 0.26 900827 13, a 5.37 0.376, AI Kb 15 5.42 6.88 0.856 1.1 930904 6.6, a 5.32 0.290, bpx DI 90 4.22 7.57 0.856 14 930903 23, a 5.18 0.298, bpx DI 90 4.00 7.40 0.856 21 930905 6.6, a 4.56 0.290, bpx AI 90 3.62 6.75 0.856 0.3

Abbreviations are same as Table 2, with P, atmospheric pressure for gases (see Table 1); target materials also include thin dolomite layer 0.3 cm (dolo-A1) and 0.15 cm (dolo-A2) overlying aluminum block and thin 0.6-cm layer of dolomite (dolo-B) suspended in tray made of thin mylar; Kb, limestone block from the Kaibab Formation; derived value of b = 5/21ogR -1ogt used to derive vapor cloud energy (equation (5)•; K, assumed value for atmospheric constant in (5); and E 9, derived vapor cloud energy x 10'ergs.

unless easily volatized targets were used. Consequently, the enhanced vapor-cloud energy most likely reflects enhanced vapor production, consistent with emission spectra revealing dissociated and ionized target species.

Vaporized mass fraction. Comparison of vapor-cloud energy derived from expansion in a tenuous atmosphere (equation (5)) and isentropic vapor-cloud expansion in a vacuum (equation (4))allows calculating the amount of impact-

21,130 SCHULTZ: •CT OF IMPACT ANGLE ON VAPORIZATION

0.5

0.4

0.2

0.1

15 ø

ß dry ice © dolomite

• carbonate

.(9.

30 ø 45 ø

0.635 cm AI

4.9 to 5.8 km/s

,i, i i i i

-0.6 -0.4 -0.2

Log (sin e) Figure 14a. Comparison of maximum velocity of expansion Urea x for different impact angles from the horizontal and targets over a limited range of impact velocities under low atmospheric pressures. Maximum expansion velocity increases with increasing impact angle, thereby indicating greater internal energy in the vapor cloud due to greater peak pressures.

generated vapor. The maximum expansion velocity Uma x in a vacuum should increase linearly with increasing impact angle (sin0) as the impactor becomes better coupled to the target for a given impact velocity (provided that target energy losses due to impactor ricochet are ignored). Increased energy partitioning to the target by the ricochet fraction with increasing impact angle, however, should result in a power law dependence on sin0. Nevertheless, Figure 14a reveals that Uma x does appear to increase with impact angle for a given impact velocity, and Figure 14b confu'ms a dependence on v sin0.

0.5

0.4

0.1

ß dry ice O dolomite

[• carbonate

i ! ! i i 0 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Log (v sin o) Figure 14b. Comparison of maximum velocity of expansion •tma x under low atmospheric pressures as a function of the vertical velocity component (vsin0)of the impactor. The in- creas• in Urnax (corresponding to internal energy) is consistent with increased shock heating of the target. Solid line links data for dolomite and carbonate targets; dashed line links data for dry ice. Figures 14a and 14b suggest that the vertical component of velocity at first contact may affect the temperature of the vapor cloud, whereas impact angle (Figures 13a and 13b) affects the total mass of the vapor.

The derived values of Uma x from Figure 14b are combined with values of Ev to estimate the vaporized mass given in Ta- ble 4. Figures 15a and 15b compare the effect of impact velocity and impact angle, respectively, on the vaporized mass expressed in terms of the impactor mass. As expected (equation (4a) - (4d)), the vaporized mass fraction increases as vi 2, once a critical velocity is reached. This critical velocity approaches 3.5 km/s for impacts into dry-ice blocks at 15' but approaches 4.5 km/s for impacts into dolomite at an angle of 15'. Previous studies concluded that the onset velocity for vaporization in vertical flyer-plate experiments was about 1.2 km/s [Lange and Ahrens, 1986]. The onset velocity of 4.5 km/s for the 15'

Table 4. Vaporized Mass Estimates

0, deg E,/KE i mv/m p IEt/KE i 860817 15 0.049 1.4 0.16 860818 15 0.11 5.0 0.33 930914 15 0.14 7.1 0.42 930911 15 0.15 8.2 0.45 860819 15 0.19 13 0.56 870202 15 0.24 18 0.65 870203 15 0.22 20 0.60 870207 15 0.23 22 0.61 860843 7.5 0.012 1.1 0.040

860844 7.5 0.013 1.3 0.042

870637 30 0.14 6.8 0.32 890205 45 0.015 0.64 0.033 931107 45 0.017 0.56 0.038

931108 90 0.0054 0.18 0.011 890203 90 0.0022 0.076 0.0044

870636 90 0.0016 0.061 0.0030 850917 90 0.00062 0.0045 0.0016 870636 90 0.0038 0.15 0.0072

930703 15 0. i 1 5.1 0.52

901210 15 0.12 6.7 0.53 871216 15 0.099 5.3 0.47

930704 30 0.12 4.4 0.44 931105 45 0.0046 0.13 0.13 931101 90 0.0045 0.011 0.0056

931102 90 0.0074 0.36 0.036

941011 15 0.10 3.6 0.43 941201 15 0.087 3.3 0.37 941203 15 0.11 4.9 0.47

941208 15 0.13 8.3 0.57 941014 30 0.27 12 1.16

880620 15 0.070 5.9 0.96 870212 15 0.051 4.9 0.86

900827 15 0.033 2.1 0.46

870211 15 0.13 8.2 0.57 870640 30 0.16 6.3 0.56 870215 45 0.018 0.62 0.053

870638 90 0.00049 0.012 0.0014

870639 90 0.0021 0.061 0.0059 870642 15 0.18 9.2 0.87

870644 90 0.0024 0.058 0.0069

900827 15 0.020 1.1 0.07 930904 90 0.33 13 0.51

930903 90 0.53 19 0.79

Abbreviations are as follows: 0, impact angle from the horizontal; Ev/KE e derived vapor-cloud energy scaled to impactor energy; m,/mv derived vaporized mass scaled to projectile mass; IE/KE i, total internal energy represented by the vapor cloud including energy for vaporization scaled to initial impactor kinetic energy. Vaporized mass is calculated using (4c) and (3) with the value of Urea x (equation (3)) including the effect of impact velocity (Figures 14a and 14b).

SCHULZ: EFFECT OF IMPACT ANGLE ON VA!•RIZATION 21,131

2.0

ß dry ice © dolomite

[] carbonate

+ water

0.635 cm AI

Velocity (km/s) 2 4 6 8 10

my v2 •,,•

dry ice-'• • '[] • / dolomite ., . [] ----carbonate

,' o.•+o+ / ,•... ----water !

!

/ ß ----- water ice !

ß

lOO

lO

........ o Log (Velocity, km/s)

Figure 15a. Derived vaporized target mass rr½ in terms of impactor mass m e for various target types as a function of impact velocity for a given impact angle (15ø). Vaporization increases approximately as the square of impact velocity (solid line) above a minimum impact velocity for dry ice targets. Va- porization in carbonate/dolomite targets appears to increase with a greater exponent on v, most likely because Ev/KEi has not yet reached a maximum (Figure 13a).

impact angles is consistent with the 1.2 km/s for the 90' impacts if the vertical velocity component affects vaporization. Although the observed amount of vapor released for vertical (90') impacts is also consistent with results of flyer-plate experiments, here the total vaporized mass increases with decreasing impact angle and reaches a maximum between 15' and 30 ø '

There are two additional uncertainties associated with the

derived mass of the vapor cloud: the role of exothermic reactions and the effect of entrained particulates. Their effects on the results, however, are believed to be minimal based on several observations. First, the use of aluminum projectiles helped to illuminate the vapor cloud through the formation of A10 +. Since the oxidation reaction of aluminum is exothermic, this process may contribute chemical energy, i.e., inducing explosive coupling with the atmosphere. Such a reaction could result from any aluminum vapor or dust-size solid/melt debris with a large area:mass ratio. Released chemical energy is not believed, however, to contribute significantly to atmospheric heating observed in the oblique impact experiments based on five observations. First, the amount of aluminum vapor produced is exceedingly small at the velocities available [Gault and Heitowit, 1963]. Second, only a very small fraction (<0.01 me)of the disrupted aluminum projectile could contribute to this process at low impact angles (<30') since significant mass and energy are retained by the larger downrange impactor debris [see Schultz and Gault, 1990]. Exothermic energy release will contribute less than 10% of the observed energy in the vapor cloud if it is assumed that 1% of the projectile mass is involved. Third, such a reaction should become more important with increasing impact angle as the disrupted projectile mass is driven against the target cavity for a longer time. This is the opposite to the observed trends. Fourth, very similar results were obtained using a pyrex impactor impacting into dry ice as well as small amounts of aluminum powder mixed in with carbonates. In the former case, backlighting was necessary in order to capture the opaque vapor cloud. In the latter case, illumination by the oxidation of aluminum gave the same results. Fifth, vertical impacts by pyrex debris clouds into dry ice vertically produced even more vapor, thereby indicating that multiple impacts play a greater role than chemical reactions.

Hence the chemical reaction leading to the brilliant self- luminous cloud provides a convenient tracer of a vapor cloud but is not believed to interfere significantly with its evolution. The role of entrained, nonvolatile fractions during impacts into more chemically complex targets such as dolomite may reduce the expansion rate through viscous energy losses within a two-phase medium. Such a process might reduce the estimate for ttmax, thereby overestimating the vaporized mass fraction. Use of dry ice, however, should minimize such effects. Future experiments will attempt to better constrain the contribution of exothermic reactions and the possible effect of cool particulates on the observations.

Frictional Heating. The increase in vaporization with decrease in impact angle (Figure 15b) suggests that frictional shear heating generated in oblique impacts may play a greater role than waste heat created during shock decompression [Schultz and Gault, 1990]. Spray [1992, 1993] has produced friction melts in silicates at strain rates (10S/s) and loading conditions (10 GPa) much lower than those produced in laboratory impact experiments (106/s and 50 GPa)but over much longer times (10 s v•rsus 1 gs). Friction melting at much higher strain rates (10'/s), however, also has been produced in oblique collisions of flat plates at relatively low velocities [Klopp et al., 1985]. Consequently, the role of shear-generated vapor in easily vaporized targets is a reasonable expectation and is consistent with the dependence of the vapor cloud energy on cos20 for a given impact velocity (Figure 13b).

Figure 16 provides a schematic approach to alternative processes for heating and vaporization. In Figure 16a, impact vaporization is assumed to occur only from shock effects in front of or below the projectile. If vapor is generated in front of the impactor without any reduction in peak pressure, then the effect of impact angle is minimal (case 1, Figure 16a). If vapor depends on the vertical component of impact velocity created only at first contact, then vapor production should decrease as sin20 (case 2). Crateting efficiency also depends on an exponent of the vertical component of velocity [Gault and Wedekind, 1978]. A 15' impact angle in this case should va- porize only 13% of the mass relative to 45' impact angle.

Increasing vaporization with decreasing angle, however, suggests two additional processes (Figure 16b). First, coupling between impactor and target may occur in a volume pro-

+1

•0

7.5 ø 15 ø 30 ø , , ,

ß

ß

ß

o dolomite

[] carbonate

+ water 0.635 cm AI

4.5 to 5.5 km/s

45 ø

10

1

0.1

0.01 -1.0 -0.5 0

Log (sin e)

Figure 15b. Derived vaporized target mass rn v in terms of impactor mass %, as a function of impact angle (0, from horizontal) and target type for a limited range of impact velocities (4.5 to 5.5 km/s). Impact vaporization increases as impact angle decreases from 45 ø to 15 ø, in contrast with expectations for only shock-induced waste heat.

21,132 SCHtLTZ: EFFECT OF IMPACT ANGLE ON VAI•RIZATION

SHOCK

1. v 2 2. v2sin 20

SHOCK/SHEAR

1. v2cot0sin2/30 2. v2cot20cos20

SHOCK/SHEAR/RICOCHeT • "'" '3 :.::':.....

v2cot20cos20

Figure 16. Comparison of impact vaporization processes and implications for impact-angle effects on vaporization effi- ciencies. (a) Shock-induced vaporization affecting a hemispherical zone mound the impactor without (ease 1) or with (ease 2)reduced peak pressures due to impact angle. (b) Shock/shear-induced vaporization affecting the projected impactor area to a depth proportional to penetration depth (case 1) or to impactor diameter (case 2). (c) Vaporization including enhancement due to hypervelocity ricochet impacts.

portional to the projected area of the impactor to a depth controlled by impactor penetration depth. In this case (1 in Figure 16b), the affected area can be viewed as an ellipse with a semi- major axis of rcot0 and semiminor axis of 2r. For a penetration depth proportional to sin2/30 [Gault and Wedekind, 1978], the vaporized zone should depend on cot0sin2/30. Such a dependence predicts, however, decreasing vaporization with decreasing impact angle: a 15' impact generating less (0.6 times) vapor than a 45' impact.

Alternatively, shear heating contributes to vaporization (see case 2, Figure 16b). In this view, the translational kinetic energy of the impactor is transformed into internal energy (IE) through interparticle friction as target material is set in motion downrange (Figure 16b• case 2). If the affected target mass is again an ellipse of area r'cot0 but a thickness proportional to impactor size, then the expended energy can be expressed simply by

2 2

I.•_E., õtr3(v cos 0)(cot0) (6) KE $ r 3v 2

i p

- (cosOXcot0)

In terms of the vaporized target mass, vaporization now increases with impact angle, a 15' impact generating 7 times more vapor than a 45' impact.

If the sheared target mass reflects the deformed impactor as it penetrates (rather than just a projection due to its oblique trajectory), then the affected area can be viewed as a parabolic section with both the downrange and transverse dimension increasing with rcot0. In this case, decreasing angle as internal energy increases (cos20cot20) as depicted in Figure 16c. Now vaporization increases more dramatically, with a 15' impact generating 26 times more vapor than a 45' impact.

As impact angles decrease, however, an increasing fraction of KE i is carded away by impactor fragments following disruption at contact [Schultz and Gault, 1990]. Consequently, the energy actually partitioned to the target (KEo)will be the initial impactor energy (KEi)minus the kinetic energy retained by the downrange impactor debris (KER):

KER/K.E i =jcot20 (7a)

KE o = Kei- KE R (7b)

= rei(1 -jcot20) (7c) where j is an empirically derived constant (taken as 0.05 for impacts into sand). Equation (7)applies only for impact angles from 45' to about 12' from the horizontal. Below about 12', KER/KEi becomes approximately constant with a value of about 0.90. The total energy expended in target heating IE t also must include the heat of vaporization. Consequently, energy partitioned into different target materials for various impactors assuming deformation can be expressed in the following form:

m/KEo = - KE) (8)

- (cot0cos0)2 Figure 17a assesses (8)and reveals two inconsistencies.

First, the observed total internal energy for impacts into water

o/

let >1 KEo •

dry. ice • carbonate calcite dolomite water water ice

KEo = KEi-KE R

1.0

0.1

0.01

-o.s o:o og

Log (cos o cot o)

Figure 17a. Total internal energy let in vaporized target mass (including energy necessary for phase change) relative to the available impactor kinetic energy KEo as a function of frictional heating that includes the energy expended by translational motion v2cos20 multiplied by the contact area affected by the impactor cot20. The available impactor energy incorporates losses due to kinetic energy retained by the impactor, following impact. Impact angles of 30' and 7.5' result in data offset from trends for 15' impact angles. Certain data indicate that the total energy represented by the vapor cloud exceeds the available input energy (KE o = • - KE•), perhaps indicating the additional contribution by the ricochet fraction.

SCHULTZ: F3TECT OF IMPACT ANGLE ON VAPO•TION 21,133

exceeds the available input energy based on (7). This result can be readily understood if the downrange hypervelocity ricochet debris actually contributes to vaporization as well (Figure 16c), i.e., not all of KER should be excluded as assumed in (7a)-(7c)and (8). Second, vertical offsets remain in data for different impact angles (Figure 17a). This inconsis- tency may represent an incomplete physical picture of the process. The peak pressure due to compression transmitted downward into the target depends on the vertical component of velocity. Consequently, the energy coupled to the target KEo actually should be replaced by KEo_L:

(9)

where 0 o refers to a reference value, which is taken as 7.5' based on experiments. Figure 17b incorporates this correction and reveals that it accommodates the data well. Data for very low angle 7.5' impacts now appear to be extensions of the higher angle data, and the observed rollover is expected since IEt/KEoñ cannot exceed unity. A consequence of combining (8) and (9) is that IEt/KE o should depend simply on cos40.

A critical test for frictional heating considers the "skin" depth of vaporization. In vertical impacts, the region of intense shock heating is viewed as a hemispherical zone around and below the impactor [e.g., Croft, 1982; Cintala and Grieve, 1994] as depicted in Figure 16a. In oblique impacts, however, the region of frictional heating occurs along the projectile/target interface, thereby localizing the process near the surface. If the process is controlled only by the peak pressure created below or in front of the impactor, then vaporization should depend on the thickness of the volatile-rich layer; that is, vaporization will decrease as the layer becomes thinner down to some limiting value. As a test, dolomite layers of different thicknesses were placed over solid aluminum blocks and impacted by 0.635-cm aluminum spheres at 15'. Surface layers as thin as 25% the diameter of the projectile had no significant effect on the observed internal energy of the vapor cloud or the derived vaporized mass (Figure 18). Because mul-

45 ø 30 ø 15 ø 7.5 ø 0 1.0

-3

O• •-N• IE• t ~ c0s20 cot20 ' KEo•.

19 KEo• = KEosin20

0.1

0.01

o. oo i

-0.5 0.0 0.5 1.0 1.5

Log (cosecote)

Figure 17b. Total energy partitioned to vaporizing the target IE t normalized by the vertical energy component KEo, = Keosin20, which incorporates the applied vertical stress, as a function of frictional heating as in Figure 16a. Data for different impactor variables and targets fall reasonably on the same relation.

1.0

0.4

Impact Velocity (km/s) 4 5 6

, , ,

15' impact angle

O carbonate •] dolomite

ß dolomite layer ß suspended dolomite jn rnv/rnp .,, v 2

8

10

8

6 •

4

!

0.5 0.6 0.7 0.8 0.9

Log Impact Velocity (km/s)

Figure 18. Derived vaporized mass rnv scaled to projectile mass m•, for 15' impacts into particulate dolomite and carbonate targets as a function of velocity. Open symbols indicate semiinfinite targets, whereas solid symbols indicate thin layers of dolomite overlying a solid aluminum substrate and dolomite suspended in a tray of thin mylar. Dolomite layers as thin as 25% of the projectile diameter did not significantly affect the degree of vaporization, consistent with the effects of shear heating.

tiple shock reflections from the aluminum/dolomite interface might enhance target heating, an additional experiment was performed where the dolomite layer equal to a projectile di- meter was suspended in a tray constructed of thin mylar (2.5 x 10 -3 cm). This configuration also had little effect on the degree of vaporization. The derived vaporized mass (4-8 m•,) exceeds the available target mass directly beneath the impactor (0.5 me) but approaches the mass contained in the projected impactor (an ellipse). More likely, the thin skin depth of interaction reflects vaporization not only from direct contact by the impactor but also from downrange hypervelocity impacts by impactor debris, i.e., "sibling" crateting as previously in- ferred from Figures 17a and 17b. In such experiments, downrange impactor spray from a 15' impact angle does not pene- trate below the original target surface (even for water).

Equations (4a)-(4d) and (9) now result in a relatively simple relation between the vaporized mass and impactor velocity and angle:

mv/m•,- V2COS4O (10) Figure 19a reveals that shear heating expressed by (10) provides a reasonable accounting of the empirical data for dry ice above a critical value. An additional exponent is required for dolomite (Figure 19b) because energy partitioned to vaporization has not reached a constant fraction of the impactor kinetic energy (Figure 13a). Vaporization for 30' impact angles into dolomite appears to be greater than the expected trend for 15' impacts (more than a factor of 2). It is proposed that the aug- mented vaporization for the 30' impacts actually represents reduced vaporization for 15' impacts owing to energy lost to hypervelocity decapitated impactor fragments. For 30' impact angles, most decapitation fragments impact the surface nearby. At 15' impact angles, some fragments fail to contact the surface while others strike at such low angles (<9') that they contribute little to additional vaporization. This interpretation is consistent with an oblique impact into a block of Kaibab limestone (see Table 3), where much of the ricochet debris missed impacting the block downrange (as revealed by nu- merous impact pits below the target plane on the downrange witness plate). The derived vaporized mass was reduced by a

21,134 SCHULTZ: EFFECT OF IMPACT ANGLE ON VAPO••ON

lOO

Dry ice

!

! ß

.J

.. mv/mp ~ v2cos40 /

ß

/ mv/mp (v2cos40) n / ~

ß 45"

i

0.0 0.5 1.0

10

1 'lJ

0.1

Log (v cos2e)

Figure 19a. Vaporized mass m,, for dry-ice targets in terms of projectile mass mp as a function of vcos20. The effects of frictional shear heating should result in mJm•,-.-v2cos40 (scc text). Such a relation holds above a critical value of vcos20, consistent with results in Figure 13a, where minimum velocity is required before EJKE• reaches a maximum.

factor of 4. Consequently, additional experiments isolating the contribution of the ricochet fraction are necessary.

Concluding Remarks

The laboratory experiments suggest that impact generation of vapor (and presumably melt) may be more complex than previously assumed owing to the roles of friction heating and impactor fate. Several •ific conclusions can be drawn for oblique impacts.

1. Four different components of impact-generated vapor can be recognized in high frame rate imaging. In addition to the jetting phase, a downrange-moving vapor cloud accompanies ricochet debris and a slower moving vapor cloud expands above the impact. A fourth vapor plume evolves from containment and redirection by the early-time penetration cavity.

2. Impact angles less than about 30' result in vapor that largely decouples from the crater excavation stage, thereby undergoing relatively unrestricted expansion.

3. Impact generation of vapor is proposed to reflect not only the role of waste heat from peak shock pressures controlled by the vertical component of velocity, but also the role of friction from high shear stresses created during oblique impacts. Although the temperature of the vapor cloud is observed to decrease with decreasing impact angle (lower expansion velocity), the combination of reduced peak pressures at contact and enhanced shear along the projectile/t_arget inter-

. 2 4

face results in impact vaporization depending on v cos 0. 4. Impact angles of 15' result in 50-100 times more vapor

for dry ice targets and 15-20 times more vapor for carbonate targets than in vertical impacts. This enhanced vaporization is consistent with independent measurements of the observed impact-generated magnetic fields [Crawford and Schultz, 1993].

5. Oblique impacts generate vapor with relatively low in- temal energy (i.e., low expansion velocities) that does not increase with impact velocity. The low internal energy corresponds to cool vapor just above the vaporization temperature. Heating may be self-limiting by vaporizing greater amounts of material through turbulent mixing (Kelvin-Helmholtz insta- bilities) rather than increasing temperature.

6. Downrange hypervelocity impacts by the disrupted impactor following first contact with the surface appear to contribute significantly to the vaporization process. This process may double the amount of vaporization for impact angles of 30' (relative to 15').

7. Low-angle impacts generate vapor principally from the upper surface. For 15-30' impact angles, vapor is largely derived from a region in the target that is only 25% the diameter of the impactor.

8. Time-exposed spectra reveal a wealth of compositional and temperature information that should allow remotely probing evolving conditions within the vapor cloud.

Such results could have several implications for larger planetary-scale events. First, the shallow provenance and enhanced vaporization at lower impact angles (15'-30') should increase the release of volatiles stored in upper cmsts. For example, the amount of CO 2, CO, SO n, and I-I•O released from the 3-kin thick sequences of carbonates and anhydrites at Chicxu- lub could be much greater than previous estimates [e.g., Sigurdsson et al., 1992; Pope et al., 1994]. Second, impact release of volatiles accumulated in the upper crest of Mars should be enhanced relative to estimates based on vertical im-

pacts [e.g., Carr, 1989]. The relatively low temperature of the released volatiles would prevent gravitational escape, thereby possibly leading to temporary increases in atmospheric pres- sme by occasional large, low-angle (30') impacts. Episodic climate change seems necessary to account for enhanced gradational epochs resulting in increased erosion rates and nar- row-valley formation [Grant and Schultz, 1990]. Perhaps certain major impacts not only removed the early Martian atmosphere [Schultz, 1986; Melosh and Vickery, 1989], but also re- plenished it.

2 . 100

O carbonate

• dolomite

ß dolomite layer

ß suspended dolomite /

30ø/"a / ?

mv/mp ~ (V2COS40)n

!

0.0 0.5 1.0

Log (v cos2e)

10

0.1

Figure 19b. Vaporized mass fraction as in Figure 18a, but for dolomite/carbonate targets. Impact angles of 30' appear to re- suit in vaporization greater than 15'. This enhancement is believed to indicate the contributing role of additional downrange hypervelocity impacts by debris, following disruption of the impactor at first contact. Enhancement occurs even for a thin layer of dolomite (25% impactor diameter) overlying a solid aluminum block. Impact vaporization for impact angles of 15-30' appears to be enhanced by a factor of 15-20 relative to vertical impacts. The vaporized mass fraction for the targets appears to depend on (v2cos40) n, where n --- 2. This relation most likely reflects energy partitioning to target heating that has not yet reached a maximum for the available velocity range (see Figure 13a).


Acknowledgments. This study would not have been possible without the technical crew at the NASA Ames Vertical Gun Range including Wayne Logsdon, John Vongrey, and Ben Langedyk. In addition, J.T. Heineck provided critical and creative input for the imaging necessary for this analysis. Both D. Crawford and S. Sugita contributed helpful discussions at various stages in this study. Lastly, the formal constructive reviews by R. Schultz and an anonymous reviewer are greatly appreci- ated. This research was supported by NASA Grant NAGW-705.

References

Brode, H.L., Numerical solutions of spherical blast waves, d. •lppl. Phys., 2•, 766-775, 1955.

Cart, M.H., Recharge of the early atmosphere by impact-induced release of CO• Icarus, 79, 311-327, 1999.

Cintala, M., and R. Grieve, Impact-induced melting of planetary sur- faces, in Large Meteorite Impacts and Planetary Evolution, edited by B.O. Dressier, R.A.F. Grieve, and V.L. Sharpton, Spec. Pap. Geol. Soc. •4m. 293, 103-110, 1994.

Cooper, H.F., Jr, A summary of explosion cratering phenomena relevant to meteor impact events, in Impact and Explosion Cratering, edited by D.J. Roddy, R.O. Pepin, and R.B. Merrill, pp. 11-44, Pergamon, Tarrytown, N.Y., 1977.

Crawford, D.A., and P.H. Schultz, The production and evolution of impact-generated magnetic fields, Int. d. Impact Eng., 14, 205-216, 1993.

Croft, S.K., A first-order estimate of shock heating and vaporization in oceanic impacts, in Geological Implications of Impacts of Large Asteroids and Comets on the Earth, edited by L. Silver, and P.H. Schultz, Spec. Pap. Geol. Soc. ,din. 190, 143-152, 1982.

Dienes, J.K., and J.M. Walsh, Theory of impact: Some general principals and the method of Eulerian codes, in High-Velocity Impact Phe- nomena, edited by R. Kinslow, pp. 46-104, Academic, San Diego, Calif., 1970.

Gault, D.E., and E.D. Heitowit, The partition of energy for hypervelocity impact craters formed in rock, in Proceedings, Sixth Hypervelocity Impact Symposium, Cleveland, Ohio, vol. 2, pp. 419-456, 1963.

Gault, D.E., and J.A. Wedekind, Experimental studies of oblique impact, in Proceedings, 9th Lunar and Planetary Science Conference, pp. 3843-3875, Pergamon, Tarrytown, N.Y., 1978.

Gault, D.E., W.L Quaide, and V.R. Oberbeck, Impact cratering mechanics and structures, in Shock Metamorphism of Natural Materi- als, edited by B.M. French and N.M. Short, pp. 87-100, Mono Books, San Francisco, Calif., 1968.

Goldstine, H.H., and J. Von Neumann, Blast wave calculation, Commun. Pure/lppl. Math, 8, 327-354, 1955.

Grant, J.A. and P.H. Schultz, Gradational epochs on Mars: Evidence from west-northwest of lsidis Basin and Electds, Icarus, 84, 166- 195, 1990.

Holsapple, K.A., The scaling of impact phenomena, Int. J. Impact Eng., 5, 343-355, 1987.

Jones, E., and J.W. Kodis, Atmospheric effects of large body impacts; The first few minutes, in Geological Implications of Impacts of Large Asteroids and Comets on the Earth, edited by L. Silver and P.H. Schultz, Spec. Pap. Geol. Soc. ,4m. 190, 175-186, 1982.

Kieffer, S.W., Impact conditions required for formation of melt by jetting in silicates, in Impact and Explosion Cratering, edited by D.J. Roddy, R.O. Pepin, and R.B. Merrill, pp. 751-770, Pergamon, Tar- rytown, N.Y., 1977.

Klopp, R.W., R.J. Clifton, and T.G. Schamki, Pressure-shear impact and the dynamic viscoplastic response of metals, Mech. Mater, 4, 375- 385, 1985.

Knowles, C.P., and H.L. Brode, The theory of cratering phenomena, An overview, in Impact and Explosion Cratering, edited by D.J. Roddy, R.O. Pepin, and R.B. Merrill, pp. 869-896, Pergamon, Tarrytown, N.Y., 1977.

Lange, M.A., and T.J. Ahrens, Impact-induced dehydration of serpen- tine and the evolution of planetary atmospheres, Proc. Lunar

Planet. Sci. Conf. 13th, Part 1, J. Geophys. Res., 87, suppl., A451- A456, 1982.

Lange, M.A., and T.J. Ahrens, Shock-induced CO 2 loss from CaCO3; Implications for early planetary atmospheres, Earth Planet. Sci. Lett., 77, 409-418, 1986.

Melosh, H.J., and A.M. Vickery, Impact erosion of the primordial Mar- tian atmosphere, Nature, 338, 227-236, 1989.

Oberbeck, V.R., The role of ballistic erosion and sedimentation in lunar stratigraphy, Rev. Geophys., 13, 337-362, 1975.

O'Keefe, J.D., and T.J. Ahrens, Impact-induced energy partitioning, melting and vaporization on terrestrial planets, in Proceedings, 8th Lunar and Planetary Science Conference, pp. 3357-3374, Per- gamon, Tarrytown, N.Y., 1977.

O'Keefe, J.D., and T.J. Ahrens, The interaction of the Creta- ceous/Tertiary extinction bolide with the atmosphere, ocean, and solid Earth, in Geological Implications of Impacts of Large •4ster- oids and Comets on the Earth, edited by L. Silver, and P.H. Schultz, Spec. Pap. Geol. Soc. Am. 190, 103-120, 1982.

Pope, K.O., K.H. Baines, A.C. Ocampo, and B.A. Ivanov, Impact winter and the Cretaceous/Tertiary Mass Extinctions: Results of a Chicx- ulub asteroid impact model, Earth Planet. Sci. Lett., 128, 719-725, 1994.

Schultz, P.H, The martian atmosphere before and after the Argyre impact, in Workshop on the Evolution of the Martian Atmosphere, Publ. 86-07, pp. 38-39, Lunar and Planet. Inst., Houston, Tex., 1986.

Schultz, P.H., Atmospheric effects on ejecta emplacement and crater formation on Venus from Magellan, J. Geophys. Res., 97, 16,183- 16,248, 1992.

Schultz, P.H., and D.E. Gault, Atmospheric effects on martian ejecta emplacement, J. Geophys. Res., 84, 7669-7687, 1979.

Schultz, P.H., and D.E. Gault, Impact ejecta dynamics in an atmosphere: Experimental results and extrapolations, edited by L.T. Silver and P.H. Schultz., in Geological Implications oœImpacts of Large ,dster- olds and Comets on the Earth, edited by L. Silver, and P.H. Schultz, Spec. Pap. Geol. Soc. ,dm. 190, 153-174, 1982.

Schultz, P.H., and D.E. Gault, Clustered impacts: Experiments and implications, J. Geophys. Res., 90, 3701-3732, 1985.

Schultz, P.H., and D.E. Gault, Prolonged global catastrophes from oblique impacts, in Global Catastrophes in Earth History:/In Inter- disciplinary Conference on Impacts, Volcanism, and Mass Mortal- ity, edited by V.L. Sharpton and P.D. Ward, Spec. Pap. Geol. Soc. ,dm. 247, 239-261, 1990.

Sedov, L.I., Similarity and Dimensional Methods in Mechanics, 363 pp., Academic, San Diego, Calif., 1959.

Sigurdsson, H., S. D'Hondt' and S. Carey, The impact of the Creta- ceous/Tertiary bolide on evaporite terrain and the generation of major sulfuric acid aerosol, Earth Planet. Sci. Lett., 109, 543-559, 1992.

Spray, J.G., A physical basis for the frictional melting of some rock- forming minerals, Tectonophysics, 204, 205-221, 1992.

Spray, J.G., Viscosity determinations of some frictionally generated sili- cate melts: Implications for fault zone rheology at high strain rates, J. Geophys. Res., 98, 8053-8068, 1993.

Sugita, S., and P.H. Schultz, Dynamical evolution of vapor clouds by oblique impacts on Venus (abstract), Lunar Planet. Sci., XXVI, 1369-1370, 1995.

Taylor, G.T., The formation of a blast wave by a very intense explosion, I, The atomic explosion of 1946, I, Theoretical discussion, Proc. R. Soc. London ,4, 201,4, 175-186., 1950•

Taylor, G.T., The formation of a blast wave by a very intense explosion, I, The atomic explosion of 1946, I, Theoretical discussion, Proc. R. Soc. London,4, 201,4, 159-174., 1950b.

Thomsen, J.M., M.G. Austin, S.F. Ruhl, P.H. Schultz, and D.L. Orphal, Calculational investigation of impact cratering dynamics: Early- time material motions, in Proceedings of the Lunar and Planetary Science Conference loth, pp. 2741-2756, Pergamon, Tarrytown, N.Y., 1979.

21,1 •6 SCHULTZ: EFI•ECT OF IMPACT ANGLE ON VAt•RIZATION

Tyburczy, J..4,, and T.J. Ahrens, Dynamic compression and volatile release of carbonates, J. Geophys. Res., 91, 4730-4744, 1986.

Vickery, A., Effect of an impact-generated gas cloud on the accelera- tion of solid ejecta, J. Geophys. Res., 91, 14,139-14,160, 1986.

Vickery, A.M., The theory of jetting: Application to the origin oftektites, Icarus, 10.5, 441-453, 1993.

Zel'dovich, Y.B., and Y.P. Raizer, Physics of Shock Waves and High-

Temperature Hydrodynamic Phenomena, 464 pp., Academic, San Diego, Calif., 1967.

P.H. Schultz, Department of Geological Sciences, Brown University, Box 1846, Providence, RI 02912. (e-mail: Peter_Schultz•brown.edu)

(Received July 17, 1995; revised July 8, 1996; accepted July 19, 1996.)

effect of impact angle on vaporization - semantic scholar · 2017. 5. 3. · effect of impact angle...

Documents