Applications of a Nuclear Technique for Depth-Sensitive Hydrogen Analysis: Trapped H in Lunar Samples and the Hydration of Terrestrial Obsidian Citation Leich, Douglas Albert (1974) Applications of a Nuclear Technique for Depth-Sensitive Hydrogen Analysis: Trapped H in Lunar Samples and the Hydration of Terrestrial Obsidian. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/BVB9-FJ36. https://resolver.caltech.edu/CaltechTHESIS:11172015-101124496 Abstract The resonant nuclear reaction 19 F(p,αy) 16 O has been used to perform depth-sensitive analyses for both fluorine and hydrogen in solid samples. The resonance at 0.83 MeV (center-of-mass) in this reaction has been applied to the measurement of the distribution of trapped solar protons in lunar samples to depths of ~1/2µm. These results are interpreted in terms of a redistribution of the implanted H which has been influenced by heavy radiation damage in the surface region. Fluorine determinations have been performed in a 1-µm surface layer on lunar and meteoritic samples using the same 19 F(p,αy) 16 O resonance. The measurement of H depth distributions has also been used to study the hydration of terrestrial obsidian, a phenomenon of considerable archaeological interest as a means of dating obsidian artifacts. Additional applications of this type of technique are also discussed. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Physics) Degree Grantor: California Institute of Technology Division: Physics, Mathematics and Astronomy Major Option: Physics Thesis Availability: Public (worldwide access) Research Advisor(s): Tombrello, Thomas A. Thesis Committee: Unknown, Unknown Defense Date: 20 November 1973 Funders: Funding Agency Grant Number NDEA UNSPECIFIED Caltech UNSPECIFIED NSF GP-28027 Record Number: CaltechTHESIS:11172015-101124496 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:11172015-101124496 DOI: 10.7907/BVB9-FJ36 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 9281 Collection: CaltechTHESIS Deposited By: Dan Anguka Deposited On: 17 Nov 2015 21:08 Last Modified: 29 Jul 2024 22:54 Thesis Files Preview PDF - Final Version See Usage Policy. 29MB Repository Staff Only: item control page

APPLICATIONS OF A NUCLEAR TECHNIQUE FOR DEPI'H-SEHSITIVE HYDROGEH ANALYSIS: TRAPPED H IN LUNAR SAVPLES AND THE HYDRATION OF TERRESTRIAL OBSIDiru~ Thesis by Douglas Albert Leich In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy California Institute of Technology Pasadena, California 1971~ (Submitted November 20, 1973) -ii- ACKNOWLEDGEr-.mNTS I vrish to express my sincere appreciation to Professor Thomas ,~ . Tombr ello for his much needed guidances advice, and constant interest throughout the course of this research. I also g ratefully a.cknovledc;e Professor Donald s. Burnett for his intense interest and indi spensi ble help in the lunar sample investigations, as 1·rell as for his nany helnful suc;gestions in all phases of this 1-rork, 1 / ·" an" r!,embers of the Kellogg Radiation Laboratory i·rere directly or {ndirectl:r helpful to this uork, and particular thanks go to >1ess ers. Richard !1, Goldberg and Roland R. Lee, It is a pleasure to acknOi·!ledce the entire Kellogg staff for the friendly and stimulating atmosphere the~r have Drovided and for valuable support and assistance. The author •rould like to thank Professors G, ,J. 1-lasserburg, :3 , l:;nstein and H, P, Taylor, Jr., Dr. I. Friedman and Hr. J. E. Ericson for their support and for many useful discussions, I am grateful for the financial assistance l')rovided by an IIDEA felloivship and by various assistantships awarded to me by the California Institute of Technology. This research was supported in part by the National Science Foundation grant r,p- 28027. Fine.ll~t, I vish to thank my wife, Sandras and her mother, 'Irs. Virginia Franklin, •rho collaborated in the t~rping of this thesis and '..rho have been patient, understandings and helpful in many ways. -iii- ABSTRttCT The resonant nuclear reaction 19F(p,ay)l6o has been u s ed to perform depth-sensitive analyses for both fluorine and hydrogen in solid sam:9les. The resonance at 0.83 ~1eV (center-of-mass) in this reaction has been applied to the measurement of the distribution of trapped solar protons in lunar samples to depths of ~~ ~m. c.. These results are interpreted in terms of a redistribution of the implanted H which has been influenced by heavy radiation damage in t he surface reg ion. Fluorine determinations have been performed in a 1-~m surface layer on lunar and meteoritic samples using the same l9F(p,ay) 1 6o resonance. The measurement of H depth distributions has also been used to study the hydration of terrestrial obsidian, a nhenomenon of considerable archaeological interest as a means of dating obsidian artifacts. Additional applications of this type of technique are also discussed. -iv- 'rADLE OF CONTENTS SECTIOH I. II. TI'l'LE PAGE IHTRODUC'l'ION 1 EXPERH1ENTAL :.~E.'THODS 5 A. Depth Analysis for Hydrogen Using lu( 1 9F,ay) 16 o D. Experimental Apparatus 10 c. Dat a Collection and Reduction 15 D. Implantation Tests 19 E. Depth Analysis for F, Na, and Af Using 20 Proton-Induced Nuclear Reactions III. 5 LUHI'u'-\ SAl'lPLE A.l'iALYSIS 23 n. Experimental Results 24 1. Apollo 11 and Apollo 15 coarse fines 24 2. Apollo 15 glass-coated rod:s 15015 34 and 15059 3. r.. Apollo 16 glass-coated rocks 64455 and 65315 40 4. Apollo 16 ALSRC samples 48 5. Simulation exneriments 56 6. Apollo 16 LSCRE foil 513 Discussion 60 l. Depth distributions of H in lunar sar!lples 60 2. Depth distributions of P in lunar samples 82 -v- SECTION IV. v. TITLE PAGE OBSIDIAN HYDRATION PROFILES 85 A. Experimental Results 86 B. Discussion 89 COITCLUSION 92 A. Lunar Sample and Obsidian Analyses 92 B. Additional Applications of Nuclear Techniques for Depth-Sensitive Analysis APPENDIX A. Depth-Sensitive Analysis with Charsed Particle Induced Nuclear Reactions B. Special Procedures for Clean UHV Systems 101 c. Analyzed Standards and Detection Efficiency 104 D. Calculation of Stopping Powers of Complex Substances E. r. 105 Imf)lantation of Interplanetary Ions in Lunar Samples 107 Lunar Sample Inventory 112 niPl'!lU~NCES 116 TABLES 122 FIGUR.8S 131~ -1I. INTRODUC'l'IOi~ T':1e use of resonant nuclear reactions to determine the concentration of certain light elements as a function of depth in solid samples has developed over the last decade as a result of increasing interest in properties of surfaces and in surfacerelated phenomena. Work begun by Amsel and Samuel resonances in the 18 o(p,a) 15N and (1962) using 27AR.(p,y) 28si reactions to study anodic oxidation has led to a continuing program of investigation of the processes of oxygen diffusion in solids using the reaction (Choudhury et al., 1965; Palmer, reaction (Amsel et al., 1968). 18o(p,a) 1 5N 1965) and the 16O(d,p) l1 0 Another group (Ollerhead et al., 1966) has used the 11 o( 3He,a) 16 o reaction in a study of oxygen diffusion and oxidation. 'rhe idea of using resonant nuclear reactions to study the depth distributions of implanted ions in solids (Porat and ramavataram, 1960) has been applied to the 15u range distributions (Phillips and Tiead, 1963) 18 and, more recently, 0 range distributions (Whitton et al., 1971) measurement of . . . th e reac t 1ons . 15-r·r ( p,ay )l2c at various implantat1on energ1es, us1ng and 180( p,a ) 15.. ~, respec t.1vel y . Resonances in the 19 F(p,ay) 16 0 reaction have also been used ( Holler and Starfelt-, 1967; Pada-vrer, 1970) to measure the concentration profiles of fluorine contarr~ ination in metals. The purpose of the present experinental study is to apnly a new variation of this t ype of technique to the measurement of hydrogen depth distributions in solids by reversing -2- the roles of projectile and target in the 19F + p systen. The motivation for this study has mainly been supplied by the lunar science program. Large noble gas concentrations were observed in the preliminary examination of the fine lunar soil material (LSPET, 1969). In order to prove that this large rare gas component was the result of solar wind ion implantation, several groups performed experiments to show that the rare gases were located near the surfaces rather than uniformly distributed throughout the volume of the grains, Measurement of the rare gas content of soil samples after removal of a surface layer by chemical etching showed that the rare gases were indeed located near the surfaces of the grains, confirming evidence supplied by the observation of a distinct anti-correlation between grain size and rare gas content per gram in the Apollo 11 soils (Eberhardt et al., 1970; Hintenberger et al., 1970; Kirsten et al., 1970). However , the implied thicknesses (ranging from 0.2 - 8 ~m) of the gas-rich surface layers were significantly larger than typical solar wind ion ranges of 0.01 - 0.05 ~. suggesting extensive modification of the distribution of the implanted ions. The resonant nuclear reaction depth analysis technique, using the reaction 1 H( 1 9F,ay) 16 o, made possjbl P. a direct measurement of the depth dis+ribution of im ~ J n nted hydrogen in lu~ar samples, providing an important check on the ~omewhat ments. surprisin~ results of the chemical Due to jts excellent denth resolution et~hin~ (~0.02 ~m) . experithis -3- technique is particularly well suited to the investigation of the historical record of solar wind and other particle radiations contained in extraterrestrial materials and of the processes which modify the distribution of the implanted ions. A s~cond problem to which this technique has been successfUlly applied is the study of obsidian hydration. An archaeological technique developed for dating of obsidian artifacts depends on the correlation of the thickness of a hydration "rind" on the surface of a given obsidian artifact, which can be observed in thin-section, with the age of the obsidian surface (Friedman and Smith, 1960). High resolution hydrogen depth profile measurements on hydrated obsidian samples, made possible by the 1 H( 1 9F,ay)1 6o resonant nuclear reaction technique, represent a valuable aid toward understanding the hydration process. Such an understanding is essential to the establishment of the obsidian hydration dating technique as a reliable chronometer. In Section II of this thesis I shall discuss the experimental methods and apparatus along with the results of implantation experiments performed as a test of the analytical technique. Results of measurements on lunar samples, including measurements of fluorine depth distributions using the 19 F(p,ay)16 o reaction, are presented in Section III. Section IV is concerned with measurements of obsidian hydration profiles. In conclusion, Section V includes an evaluation of the technique and its applications, and a discussion -4of further promising applications of nuclear reactions as depthsensitive analvsis probes. 'l'he development of the 1 H( 1 9F,ay) 1 6o technique discussed in Section II has been previously described by Leich and Tombrello (1973). l'1any of the lunar sample results and ideas presented in Section III have also been published (Leich, Tomhrello and Burnett, 1973a) or are in press (Leich, Tombrello and Burnett, 1973b). A manuscript dealing with the obsidian hydration measurements discussed in Section IV is being prepared for publication (Lee et al., 1973). -5- II. EXPERIMENTAL METHODS In this section the experimental methods, techniques, and apparatus are treated in detail. A. Depth Analysis for Hydrogen Using 1 H( 1 9F,ax) 16 o The applicability of the resonant nuclear reaction depth analysis technique depends on the existence of an isolated resonance in a nuclear reaction involving an isotope of the element to be analyzed. In Appendix A the general case of a nuclear reaction A(a,b)B is considered in some detail, with regard to its use as a depth analysis probe for nuclei A, when such a resonance exists. The existence of a strong, narrow, isolated resonance in the reaction 1 H( 1 9F,ay) 16 o has been exploited in this study to determine hydrogen depth profiles in solid samples, an l9F energy The resonance employed occurs at ER = 16.45 MeV with a total width r (FWHM) of ~9 keV, corresponding to a proton energy of 0.872 MeV and width of 4.7 keV for the same resonance in 19 F(p,ay) 16 0 (Ajzenberg-Selove, 1972). The peak cross section a R is "'0.6 barn for emission of an a particle leading to one of three excited states of the residual 16 o nucleus at excitation energies of 6.1, 6.9, and 7.1 MeV (Ajzenberg-Selove and Lauritsen, 1959). All three states de-excite by emission of a prompt y ray directly to the ground state. At 19F energ ies different from the resonance energy b,y more than a few the cross section for the production of the high energy y rays is r -6- ne~lip,ible. Thus, if the surface of a material containin~ hydro~en is irradiated with 19 F ions at energies sufficiently greater than the resonance energy ER• the 19 F ions will gradually slow down, due to electronic collisions, until at a depth xR the resonance energy is reached and the reaction will occur at a rate proportional to the hydrogen concentration in a thin layer at xR. At greater depths the 19F energy will fall below the resonance energy where the cross section is again negligible compared to ~· Hence hydrogen located outside the layer at :x;R :t r/[-dE/dx] contributes only a negligible amount to the total reaction yield. Since the stopping power, dE/dx (a negative quantity), is nearly independent of energy (within ~1%) (Northcliffe, 1963) in the relevant range of l9F energies, the depth xR can be related to the l9F beam energy E0 by the linea.r relation = Eo- ER [-dE/dx] (1) Selection of the 19F beam energy E0 is equivalent to specifying the depth xR at which the hydrogen concentration is to be determined~ and measuring the y-ray production rate as E0 is varied gives a direct indication of hydrogen concentration as a function of depth in the target. The relation between H(xR)' the hydrogen concen- tration at deoth xR~ and Y(E0 ) , the reaction yield per incident l9p ion of energy E , is shown in Appendix A (eauation A.7) to be 0 -7given by n OR r Y(E ) = 2(-dE/dx)R 0 (2) H(xR), using the Breit-Wigner dispersion equation, a(B) (3) to describe the dependence of the cross section a on 19 F energy E in the vicinity of the resonance at ER. A second strong resonance at an 19F energy of 17.64 MeV producing the same characteristic y rays as the first resonance (although in slightly different proportions) limits the depth range over which the first resonance can be used in the straightforward manner outlined above to about 0.4 ~m. For H distributions extend- i ng to greater depths, information can still be extracted from t h e excitation function Y(E0 ) for E0 > ER 2 = 17.64 MeV. (Since we must now deal with two resonances simultaneously, the subscripts l and 2 are used in the following to indicate the resonances at 16.45 HeV and 17.64 MeV respectively.) The yield per incident 19F ion is now the sum of the yield from the first resonance due to hydrogen at depth ~l = f ~l dE [dE/dx] and a contribution from the second Eo dE . for resonance due to hydrogen at depth xR 2 = f ER2 [dE/dx] , ~.e., Eo (4) -8where 0 Rl fl 2[-d.E/dx]Rl 1r H(xRl) (5) and 0 R2 f2 2[-dE/dx]R 2 1r But since xR 2 = xRl - y (E ) 2 0 ~x, (6) where ~x n °R2 r2 = 2[-d.E/dxJR2 /1u dE , we have 2 [dE/dx] - ~ H(xRl - ~x), (7) and hence, Y2(Eo) = ( r2 [dE/dx]n1 [ y E - ~E)' d.E/dx]R2 1 o rl Rl 0 R2 0 (8) ~X with 6E =! [-dE/dx]dx. The ~5% variation of the stopping power 0 over the extended range of useful 1 9F beam energies may be approximated quite accurately by a linear decrease vrith energy. With this [dE/dx] 0 approxir.1ation, we obtain ~E = ~~ [dE/dx]R ''here [dE/dx] 0 is 2 the value of the stopping power at E0 and ER 2 - ERl• Substituting equation 8 into equation 4 leads to the expression -9- 0 R2 r2 (dE/dx]Rl ___;~..,r,;;.~(..-d.E~"~/~dx-].,..;;.;;;;. Y1 (Eo- 0 Rl 1 R2 from a similar determination of Y1 at a lower beam energy, E0 (If (E0 - liE)< (9) l:!,.E'). t,E. - (En 2 - r 2 ), Y1 (E0 - liE) is measured directly.) 'rhe hyurogen concentration H(xR1 ) is then inferred from the determination of Y1 (E 0 ) in the usual way, using equation 5. Hydrogen depth profiles can be determined, using the procedures outlined above, with a depth resolution ox depending on the width r 1 and on the spread in energy of the l9F ions as they penetrate the srunple. Using equation A.9 of Appendix A, we can estimate the resolution by (10) where ll~0 is the FWHM spread in beam energy (a few keV) and n(~ 1 ) is the F'\·lH.M eqergy straggling at a depth xRl. Figure 1 shows the calculated resolution ox as a function of depth in a quartz sample using an estimate of the energy straggling based on theoretical treatments of atomic collision processes (Maccabee et al., 1968). The 0.02-].Jm resolution available near the surface is gradually degraded by the energy straggling to 0.03 ].Jm at a depth nf nearly 2 J.Jm. '1.'his represents an order of magnitude iraprovement in reso- -10- lution compared to chemical etching and other destructive sectioning techniques. The sensitivity of nuclear reaction analysis techniques depends to a large extent on the importance of competing reactions; i.e.; reactions produced by the ion beam on other constituents of the sample. Some of these reactions may result in reaction products which are difficult to distinguish from the products of the reaction of interest. For 1 H( 19F,ay) 160 there is very little interference of this sort (until the l9F energy is raised above ~20 MeV) since the production of the 6-7 MeV y rays is quite copious. Above 20 MeV, the sensitivity falls rapidly to zero because of the rapid increase in background due to what appears to be neutron production by ( 1 9F,n) and/or ( 1 9F,2n) reactions on oxygen as the Coulomb barrier height is approached. This interference has been the determining factor in the ~2-~m depth limit of the present measurements. B. Experimental Apparatus The extreme sensitivity of measurements of this kind to surface contamination necessitated the design and construction of a special ultrahigh vacuum (UHV) scattering chamber. 'I'he primary concerns as sources of hydrogen contamination were diffusion pump oil vapors (since accelerator beam tubes are normally pumped by oil diffusion pumps) and water vapor. The solution of this problem has been to equip the scattering chamber with a clean, self-contained getter-ion pumping system, and to baffle the accelerator-scattering chamber interface allowing passage of the ion beam with a minimal -11flm·r of contaminant vapors from the accelerator beam tube. A schematic drawing of the scattering chamber vacuum system is shown in Figure 2. The scattering chamber itself consists of two stainless steel UHV flanges bolted together with a copper gasket metal-tometal seal to form a cylindrical exterior 20 em in diameter and 4.5 em ueep. The interior faces of the flanges were machined to form a cavity 15 em in diameter and 3 em deep. A direct drive rotary feedthrough was mounted on a port at the rear of the scattering chamber so that its shaft extended into the cavity along the chamber axis. A 10-cm diameter aluminum target wheel was mounted on this shaft using glass spacers to electrically insulate the wheel from the rotary drive shaft. Twelve 16.5-mm diameter by 6-mm deep holes equally spaced along a 7-cm diameter circle on the target wheel provide mounts for samples to be analyzed. One of two ports on the front face of the chamber serves as the connection to the accelerator beam tube and is placed 3.5 em off-axis to line up with the centers of the target mounts. The second port is fitted with a window and allows precise positioning of the target wheel using degree markings on the edge of the wheel and a vernier mounted on the chamber wall, and also allows the targets to be viewed while in their mounts in the sealed scattering chamber. A sliding electrical contact fabricated from beryllium-copper alloy makes contact with the aluminum target wheel and connects with one lead of a ceramicsealed electrical feedthrough, thus allowing the beam current to b e -12- collected from the target wheel with negligible leakage to ground. The scattering chamber vacuum is maintained by a getterion pumping system with a pumping speed of 50 liters per second for air. A Varian sublimation pump provides most of the pumping speed by means of the gettering action for chemically reactive gases of a thin layer of titanium coating the interior walls of the water-cooled stainless steel pump chamber. The titanium layer is periodically renewed by sublimation from one of three titanium alloy filaments in the center of the pump chamber. A small pumping speed for inert gases is n~intained by a standard 15~/s Varian Vac-Ion pump. All vacuum connections in the scattering chamber and vacuum system are made using copper-gasket metal-to-metal seals, allowing the entire system to be baked at temperatures up to 300°C to drive trapped gases from the interior surfaces. A straight- through 1-cm bore gold seal valve is used as a be&n-line isolation valve, while a second gold seal valve is used on the roughing line. A molecular sieve sorption pump is used to obtain the necessary lo- 2 Torr vacuum to start the ion pump. The getter-ion pumping system, together with the use of low vapor pressure bakeable materials in the construction of the scattering chamber, enables a vacuum of <10- 9 Torr to be routinely maintained, and eliminates internal sources of hydrocarbon contamination. To minimize contamination from pump oil and other vapors in the accelerator beau1 tubes, the connection of the scattering chamber to the beam line is interfaced with a 40-cm long, 1-cm bore -13li~uid nitrogen cooled baffle. The trapping of condensable vapors on the refrigerated walls of this low-conductance tube enables the getter-ion pumping system to maintain a vacuum of the scattering chamber with beam on the target. ~o-9 Torr in Thus, sources of hydrogen contamination are minimized while another potential ~roblem, the build-up of carbon deposits on the targets from the cracking of hydrocarbon vapors during irradiation with an ion beam, is virtually eliminated. The 1 9F beam necessary for hydrogen depth analysis is provided by the Caltech tandem accelerator. Fluorine ions are extracted from a duoplasmatron ion source using freon-14 (CF ), 4 diluted with helium, as the source gas. l9F- ions are accelerated to the terminal where they are stripped to a variety of positive charge states. Following the second stage of acceleration, the l9p 4+ ~he resulting 19F4+ beam is monoenergetic to within a few parts in 104 • The beam c har r,e state is selected by a 90° magnetic analyzer. energy can be determined precisely using a digital nuclear magnetic resonance gaussmeter (Alpha Scientific, Inc. Model 3193) to measure the field between the two poles of the analyzing magnet, and a calibration performed by Mann, et al. (1973). The beam is then steered into the target room beam tube (S 10° leg) and focused through two sets of collimating slits located 50 em apart at the end of the flight tube as indicated in Figure 2. With the beam-line gate valve and the scattering chamber isolation valve open, the l9p4+ bean! passes through the 1-cm diameter tube of the liquid -14nitrogen cooled baffle and into the scattering chamber where the 19p ions are stopped in .the target. The target wheel is held at a potential of +300 V to retain secondary electrons, so each ion deposits a net charge q of +4e. This charge is continuously collected by the sliding electrical contact on the target wheel and the resulting current is fed out the electrical feedthrough, through the 300 V bia s battery and a 1 MQ limiting resistor, and delivered to the input of a Brookhav en Nuclear Instruments Model 1000 current integrator. This instrument amplifies the input current and puts out a digital pulse each time the integrated current reaches a preset charge, typically 6.0 x lo-1 ° Coulomb. A 7.6-cm x 7.6-cm lead shielded Nai(Ti) .scintillation detector placed outside the scattering chamber at about 30° scattering angle and about 7 em from the target is used to detect y radiation from the excited 1 6o nuclei. A photomultiplier converts the scintillations into voltage pulses which are then amplified and analyzed using a 400-channel pulse-height analyzer (RIDL 34-27 series). The pulse- height s p ectrum s hmm in Figure 3 is the result of a 10-~C irradiation of an analy zed sample of Belvidere Hountain chlorite (1. 64% H by weight) with a 30-nA beam of 17.2 HeV l9F 4+ ions. A 3.7-to 7.4-MeV counting window includes the full -energy peaks and single and double escape peaks, but excludes the tail of the Compton distribution in the response of the Nai(Ti) crystal to the three y rays. With this energy window, the counted fraction n of all y rays emitted is estimated, using the chl orite standard, to be about 0.022 (see Appendix C). -15- c. Data Collection and Reduction The procedure for measurement of the excitation function 1 (~ 0 ) involves counting the number of y-rays emitted for a given number of incident 1 9F 4+ ions at a series of beam energies E • Y 0 First the beam energy E0 must be selected and the beam steered and focused onto the target wheel. Then a beam-chopping magnet is turned on, steering the beam out of the beam tube, and the sample to be analyzed is set in position by manipulating the rotary feedthrough and lining up the correct degree mark on the target wheel with the vernier. The total amount of beam charge Q to be collected per data point is preset on a mechanical register, and the run is started by activating a gate control which simultaneously starts the counting electronics and switches off the beam-chopping magnet, allowing the beam to irradiate the target. Counts are stored in the pulse-height analyzer during the entire "beam on" period, while a digital register tracks the integrated beam charge by tall~ring the pulses from the current integrator. \·lhen this register reaches the value preset on the accompanying mechanical register, a gate is automatically activated which turns on the beam-chopping magnet, thereby stopping the beam, and also blocks further input into the pulse-height analyzer. The count totals in the channels correspond- ing to the selected 4-7 HeV energy window are then summed and read out. This number, N(E0 ), is recorded along with the energy E0 , the total charge Q, and the total "beam on" time t, normally 3 - 10 minutes for beam currents in the 10- 30 nA range with Q = 6 ~ C. 'I'he analyzer memory and current integration register are then cleared, -16a new beam energy is selected, and the sequence is repeated for the next dattm. The optimum beam energy sequence spaces data points b;r an amount consistent with the spatial resolution of the measurements, or about 50 keV. The range of energ ies used begins slightly below the resonance energy ~l and continues up to a maximum energy determined by the depth of the distribution being measured. Once data have been collected in this manner, they can be transformed to a depth profile by plotting the number of counts N(E ) for a given total beam charge 0 Q vs. the beam energy :ti:0 , and performing simple linear scale transformations. becomes a depth scale using the conversion xRl The energy scale - En1 = Eo [ -dE/dx] • The data 1J (E 0 ) include a background contribution consisting of a beam independent portion (room background) primarily due to cosmic rays, and a beam dependent background due partially to competing reactions 1 and partially to 1·reaker resonances in the reaction 1 H( 9F ,a.y ) at lower energies. 16 o The background can be taken into account by writing N(E 0 ) = ~[nY(E ) +A]+ Bt q 0 (ll) vrhere n is the detection efficiency, q is the charge per incident ion(= 4e), A represents the beam-dependent background (counts per incident ion) and B the beam-independent background (counts per second). :Jince D is independent of Q and E , the contribution Bt 0 can be held constant by holding the time t constant, or equivalently holding the beam current constant, for each data point. The beam -17dependent background AQ/q will also be constant if A is independent of E0 over the energy range used. This assumption has been verified (to vi thin :t 10/; ), by measurement of hydrogen-free samples, and samples with constant volume concentrations of H such as the chlorite standard, up to an energy of "'20 ~1eV. Thus, if the beam current is held constant, equation 11 becomes S.q nY(E0 ) + K where the constant K (equal to (AQ/q)+ Bt) is evaluated from data taken at beam energies below ERl where Y(E0 ) goes to zero. E0 < ~2 , (1 2) For the scale transformation used to give H(xR1 ) is obtained by substituting equation 5 into equation 12: 2[-dE/dx]Rlq H(xru) = - - - - - - - [N(E ) - K] • n n aRl r 1 Q o (13) and subtracted from the data to give (14) following the procedures outlined in Part A of this section. for ~l < E0 < Hence, ER 2 , the N(E0 ) data scale is converted to a hydrogen concentration scale with the zero point determined by the constant background contribution Ks and the scaling factor given as in -18equation 13 above. For E 0 > En 2 , H2 (E0 ) must be subtracted from the data as in equation 14 in order to extract the hydrogen concentration for the depth ~ 1• Hare sophisticated numerical deconvolutions of the resonance shape from the excitation function to yield a more accurate depth profile are possible, in principle, but have not proved necessary in practice. Rather than rely on uncertain values for oRl and r 1 and an estimate of n based on these same values, a more direct approach is used in 1vhich the weight fraction of hydrogen WH is obtained by direct comparison with counting rates obtained from irradiation of the chlorite standard: (15) where [dE/d(px)]Rl = ~ [dE/dx]Rl' [clli/d(px)]Fn = t,[dE/dx]'Rl' p and p' are the densities of the target medium and the chlorite standard respectively, [~/dx]~ 1 is the stopping power of the chlorite for 19F ions of energy EHl' and H' and K 1 are the total counts and baclcground c,o unts, respectively, for data taken for a total beam charge Q' per data point at energies between Enl and ER 2 • The densities p and p' need not be known to use equation 15 since the quantities [dE/d(px)]Rl and [dE/d(px)]Rl can be calculated -19directly from tabulated stopping powers (Northcliffe, 1963; Horthcliffe and Schilling, 1970) and chemical analyses of the samples as described in Appendix D. D. Implantation Tests In order to establish the sensitivity and reliability of this technique, implantation experiments have been carried out on fused silica, crystalline quartz, and Ca-rich feldspar samples using 12-keV protons with doses of about 1016 cm- 2 • The target materials were chosen for their chemical and physical similarity to the returned lunar samples (silicate glasses and rocks). Targets were chemically cleaned and then baked in the scattering chamber to remove surface contamination prior to implantation. A magnetically analyzed proton beam with a flux of about 1ol 2 s-1 -cm- 2 (~0.1 ~ into a l cm2 area) was obtained for implantation from a duoplasmatron 2 ion source. Following implantation of a set of targets the scattering chamber was transferred to the tandem accelerator beam line where a 0.1-~ beam of 19p 4+ ions •ras directed into a 4-mm spot for the H analysis. Figure 4 shows data for a typical implanted target and a target which was not implanted but was otherwise identical. Figure 5 shows the implantation profile obtained by subtracting the data for the blank from the data for the implanted sample. The depth and H concentration scales are calculated using the stopping power of Si0 2 for 1 9F ions calculated from data in Northcliffe (1963) and Horthcliffe and Schilling (1970). The experimentally determined distributions show -20- some deviation from the theoretical range distribution (Lindhard et al., 1963). A comparison between the measured and theoretical mean range and full-width at half-maximum (FWHM) is shown in Table 1. The differences may, in part, represent the effects of diffusion; accelerated, perhaps, by localized heating from the implantation beam. Repeated analyses of a single implanted sample show gradual but consistent inward shifting and broadening of the distribution, suggesting a continuation of the same process by which the range distribution is distorted during t he implantation, although at a slower rate. With the exception of this effect, the profiles are reproducible; thus, there is no reason to doubt that the actual proton distribution is being measured. These results als o establish the quantitative retention, within a factor of two, of amorphous and crystalline materials for implanted protons. Similar experiments with aluminum targets gave negative results in this regard, with less than 5% retention under similar experimental conditions. E. Depth Anaylsis for F 2 Na 2 and AQ. with Proton-Induced Nuclear Reactions Depth distributions of fluorine in solid samples can be measured using a proton beam to produce y radiation from the l9F (p,ay) 16o reaction in a manner entirely analogous to the measurement of hydrogen distributions with a 1 9F beam. The same resonance is used in both analyses, occurring at a proton ener gy Ern = 0. 87 2 MeV ,.i th FllHll r 1 = 4. 7 keV in the frame of reference in which the 1 9F target is initially at rest. The proton beam energy E0 is varied to -21measure the excitation function Y1 (E0 ), and the fluorine concenERl dE tration F(xRl) at depth xRl = f [dE/dx] Eo is found from (16) in analogy with equation 5. Here [dE/dx]Rl is the stopping power of the medium for protons of energy E = 0.872 MeV. Rl yield due to the resonance at ~ The y-ray 2 = 0.935 MeV (r 2 = 8.1 keY) can be unfolded in the same manner, and in this way the depth distributions can be measured beyond the 1-~m depth corresponding to an energy loss of 6ER = ER 2 - ~l = 63 keY. The proton beam energies necessary for this analysis can also be obtained from the Caltech tandem accelerator, and the measurements can be performed with the same apparatus and electronic set-up as is used for the hydrogen analysis. Since the same characteristic y rays are detected with the same efficiency n in each case, even the discriminator levels remain the srune. All that need be done is to switch from a 1 6 .45 + HeY l9p beam to a 0.872 + MeV proton beam, and the distribution being measured becomes that of fluorine rather than hydrogen. Aluminum and sodium distributions can also be measured by detecting high energy y ra~rs from (p,y) reactions. the 2 7A~(p,y) 28 For aluminum, si resonance used occurs at a proton energy of 0.992 l·~ eV •rith a •ridth of 0.1 keV resulting in a maximum energy y ray of 10.8 J.1eV vrith a branching ratio of 0 .76. Sodium measure- -22ments can he performed using a 2 3Ha.(p,y) 24Mg resonance at 1.318 MeV proton energy and a width of 1.4 keY resultine; in 11.6 and 13.0 HeY y rays (Endt and Vander Leun, 1 967). Counting windows must be adjusted for the particular y rays to be detected for AR- and Na analyses, thereby altering the detection efficiency n, but otherwise the procedure for obtaining depth distributions follows the same procedures described for hydrogen and fluorine measurements using a single resonance. Table 2 lists the reactions used in this study to analyze for H, F , :aa, and AR-, along with the ion energy at resonance, the energy of the principal y rays, and the reaction sensitivity for each resonance. The latter quantity is defined as the thick target yield per incident ion for a quartz sample containing one ppm by weight of the particular target atoms. The resolution ox for measurement of depth profiles in quartz is shown in Figure 6 for each reaction. The rapid degradation of the resolution with depth for F, Na , and AR- is a result of the relatively large energy straggling for protons as compared to heavier ions of comparable velocity. The resolution for H depth profiles suffers little from energy straggling of the l9p ions and remains excellent to a depth of ~2 ~m. -23- III. LUNAR SAHPLE ANALYSIS Results of hydrogen and fluorine depth distribution measurements on lunar samples are presented and discussed in this section~ as well as a simulation experiment performed to determine the possible extent of penetration of terrestrial contaminants into the radiation damaged surface layers of lunar samples. The results of a search for II implanted into a platinum foil exposed to a solar / flare during the Apollo 16 mission are also presented. Knowledge of the depth distribution of hydrogen in lunar samples is important for understanding the origin of the observed H. Implantation of solar wind protons should lead to large surface (within a few hundred angstroms) concentrations of H. The more energetic solar flare protons would result in a distribution extending over a much greater depth range (up to a few millimeters). Indigenous lunar H~ incorporated into rocks at the time of their formation~ would be distributed throughout the volume of the lunar samples; however~ solar wind H inherited from a pre-irradiated parent material could result in a similar volume distribution. Terrestrial contamination~ particularly surface absorption of H20~ is a likely source for observable amounts of H. Consequently~ the basic problem in identifying the origin of the lunar H lies in distinguishing the lunar component from terrestrial contamination. -24- A. Experimental Results Samples 1vi th relatively large ( "' 2 mm x 2 mm) smooth surfaces of uniform composition are desirable for analysis by the resonant nuclear reaction techniques employed in this study in order to simplify interpretation of the measurements. Large samples also allow low beam-current densities (typically lOnA of l9p 4+ into a O.l-cm2 area) to be used, thereby minimizing the effects of beam heating. These considerations are best satisfied by the lunar glasses, which are a major constituent of the lunar regolith. The samples selected for this study, therefore, are predominantly lunar glass fragments and glass coated rocks. Samples were obtained in a series of separate allocations. For convenience, however, the samples can be collected into four ~oups determined by similarities in the particular sample handling and analysis procedures used and by similarities in the samples themselves. These groups are: (1) Apollo 11 and Apollo 15 coarse fines, (2) Apollo 15 glass-coated rocks, (3) Apollo 16 glass-coated rocks, and ( l~) Apollo 16 samples returned in a vacuum-sealed sample container. 1. Apollo 11 and Apollo 15 coarse fines The first set of samples obtained for this study consisted of three 2-3 mm sized lunar glass fragments selected on the basis of their size and smoothness from two separate aliquots of Apollo 11 "course fine s " (l-4 mm fragments) numbered 10085 ,1 and 10085 ,31. The parent sampl e 10085 consists of the 1-10 mm fragments separated fr om -25- the Apollo 11 bulk sample 10002 at the Lunar Receiving Laboratory (LRL). Separate allocations (or other subdivisions) of material from a partieular parent sample are indicated by the number following the parent sample number and separated from it by a comma, e.g. 10085,1 and 10085,31. \\'here such a sample has been further subdivided in our laboratory, an additional identifying number or letter is assigned to each fragment and is preceded by a hyphen, e.g. 10085,31-9. The 10085 brown glass samples are typical Apollo ll brown glass as described by many authors (see, for example, Keil et al., 1 97 0). Samples were subjected to a microscope examination in a clean laboratory atmosphere prior to analysis. Two of the samples, 10085,31-9 and 10085,31-12, had been given an ultrasonic rinse in high-purity acetone -- a treatment which is essential in many cases to remove layers of lunar dust from the samples. In order to determine the extent of surface H contamination associated with this procedure, several fused silica discs were given an acetone rinse using exactly the same procedure as was used with the lunar samples. These discs had previously been chemically cleaned and vacumn baked at 300°C and showed no detectable (<10 1 4 atoms/cm2 ) surface concentration of H. Following the acetone rinse, the silica discs were remounted in the scattering chamber and remeasured with the l9F beam. contamination ranging from only ~1o 14 Surface to 101 5 H atoms/cm 2 were observed, and even these small surface concentrations disappeared a lmost entirely after a few times 1014 ions/cm2 of illumination 'vith the l9F beam. Although these results show that acetone contamination -26is not a problem for normal glass surfaces, the exotic surfaces of lunar sruaples could conceivably react quite differently to the acetone. However, measurement of the hydrogen distribution on the 10085,1 r,lass fragment and on a sample from a subsequent allocation (15533,1~-1), both before and after exposure to an acetone rinse, resulted in no observable increase in the H content of the sample. He conclude, therefore, that the acetone rinse is not likely to be an important source of H contamination of the lunar samples. The three 10085 brown glass fragments were mounted in the scattering chamber using aluminum foil to support the sample in the desired position, chosen to expose a smooth, dust-free face to the 19p beam, and then placing a clean, H-free fused silica collimator with a 2-3 mm circular aperture directly over the sample. The resulting target assembly is kept in place in the target ,.,heel with a copper alloy retaining ring. Hydrogen distribution measurements were performed according to the procedures outlined in the preceding section using a 10-15 nA beam of l9p 4+. The beam was collimated to about the same size as the smallest sample (~2 mrn) so that most of the 1 9F ions would strike the lunar samples, while the fused silica collimators prevent the halo of the 19p beam from illuminating the target holder. Tests have shown that this fused silica n~terial retains very little H after baking under vacuum at 300°C, so l9F ions striking the fused silica instead of the lunar sample contribute a negligible amount to the y-ray y ield. However, since these colli- mators are fastened to the target wheel the integrated beam current includes a contribution from the ions incident on the collimator as -27well u~> those hi ttint::. the lunar sample. Using a sample of the chlorite standard (1.64 percent H) mounted behind one of these collimators, we have determined the fraction of the 1 9F beam illuminating the collimator to be normally about 0.2 for a collimator uith a 2.4 mm aperture. The uncertainty associated with estimating this fraction for a particular distribution measurement does not affect the measured shape of the distribution (provided the fraction remains constant for each data point), but only the estimate of the absolute H content. Although the density of the target medium need not be known in order to obtain a depth scale for the measured H profile in units of ~g/cm2 , it is generally more convenient to deal with depths measured in the usual distance units. Calculation of the stopping power dE/d(px) proceeds as in Appendix D resulting in a value of -9.0 keV-cm 2-ug-l for an 11 average 11 lunar composition (46% Si0 2 , 4 7~ Ti0 2 , 15% At 2 o3 , 14% FeO, 11~; CaO, 9% MgO, 0.6J~ Na2 o, and 0.3/~ K<-0 0) . 'rhe densit y of a glass of such a composition is estimated to be ~2.6g/cm 3 by extrapolating data on commerical glasses and using empirical rules for calculating the density of a glass from its chemical composition (Norey , 1954; Stevels, 1948; Huggins and Sun , 1943). Although the den sity p increases as the FeO and Ti0 2 ~ontents increase at the expense of Si0 2 and At 2 o3 , the higher stopping cross sections of the heavier atoms tend to counterbalance the increasing densit~r so that neither dE/d(px) nor d.E/dx are as sensitive to th e composition as P. Hence by assuming a nominal density of 2.6g/cm3 -28and using the value 9.0 keV-cm2 -~g- 1 for -dE/d(px) for all lunar glasses, without considering the specific composition of a particular s~nple, the uncertainty of ~10% in the calculation of the depth scale is dor:J.inated by the uncertainties in the estimates of p and dE/d(px) for the "average" lunar glass and not by the unknown composition. ~or simplicity, we have used the same scale conversion for crystalline samples also. The heavy radiation damage associated with long-term solar wind bombardment should transform any exposed crystalline surfaces to a ~uasi-amorphous state (metamictized) to depths of at 0 least 500-1000 A as observed by Bibring, et al. (1972) for lunar dust grains. Consequently, the density of the surface layer (to depths of at least ~o.l ~m, and possibly somewhat deeper if radiation damage is still quite heavy below this layer) might be better approximated by the density of a glass rather than that of the relatively undamaged crystalline material so~ewhat deeper within the lunar rocks. One of the brown glass fragments, 10085,31-12, revealed a large surface-correlated H content, with a peak density in excess of 2 x 1021 II atoms per cm3 (more than 1500 ppm H by weight) at a depth of 0.11 ~m (Fig. 7). The H concentration profile drops sharply near a depth of 0.20 ~m followed by a more gradual decrease with increasing depth to the limit of the measurement at ~0.45 ~m. The residual H concentration of more than 200 ppm at this depth is well in excess of bulk hydrogen analyses of Apollo 11 lunar materials, typically 50-100 p:om H (D 1 Amico et al., 1970; Epstein and 'l'aylor, -291970; Friedman et al., 1970); however, the distribution has not yet begun to level off at this depth. The measurement of the distribution was repeated twice on this sample. Each repetition shmred a reduction in the H content of roughly twenty percent in the outer 0.2 ~ and five percent between 0.2 ~m and 0.4 ~m deep as compared to the previous measurement of the distribution. Data from a remeasurement of the profile (third run) are plotted in Figure 7 to illustrate this gradual reduction in hydrogen content observed QUring the measurement of this sample as well as subsequent samples. This means, of course, that the measured profile is a slightly distorted representation of the initial distribution because the mobilization of the H due to the irradiation by the 19p beam is continuously causing a slight modification of the distribution. However, similar profiles were obtained from several other lunar samples even though the sequence of beam energies chosen for the analysis was different for each sample. This indicates that the distortion is not great enough to change any of the essential features of the distribution, with the exception of the first few hundred angstroms. An increase in the H content within a few hundred angstroms of the surface has been observed on most samples following a few days of atmospheric exposure, but usually this surface H concentration is rapiQly depleted during the first few minutes (~lol4 ions/cm2 ) of remeasurement with the l9p beam. It is likely that this extremely mobile surface layer is primarily a very superficial H2 0 contamination. PROPERTY OF KELLOGG RADIATION LABORATORY -30Sample 10085,31-12 was large enough to permit a relatively large (3.2-mm aperture) collimator to be used, allowing essentially all of the 1 9F beam to illuminate the sample. The smaller size of the 10085,31-9 and 10085,1 fragments necessitated the use of smaller collimators with apertures of 2.4 mm and 2.0 mm respectively. A beam alignment problem encountered during the analysis of 10085,31-9 limits the usefulness of the data from this sample, but a definite enrichment in the H content of a layer roughly 0.2-0.3 l-Im deep was observed suggesting a profile with approximately the same shape, although with a much lower absolute H content, as the distribution measured on 10085,31-12. 'r he H distribution obtained from the 10085,1 glass fragment (Fig. 8) also exhibits a similar depth dependence, with a total H content (in the 0.4 l-Im deep measured region) '\.l/3 as great as the value for 10085,31-12. The most obvious difference in the shapes of the distributions in Figures 7 and 8 (aside from the difference in absolute H content) is the much greater clarity with which the 10085,31-12 distribution shows a twocomponent shape due to the relative enhancement of the shallower, 0.02-l-lm deep component in the 10085,31-12 sample as compared to the somewhat flatter distribution in the 10085,1 sample. Ultimately, it is the continuous mobilization of the H distribution which prevents the extension of these profile measurements to depths ereater than '\.0.45 l-Im. Beyond this depth, the contribution of the 17.G4 !'.leV (19F energy) resonance in the 1H( 1 9F,ay) 1 6o r eaction becomes significant , and the relatively simple unfolding procedure -31described in Section II of this thesis can only be used in cases where the profile is reasonably stable and reproducible. However~ since the most interesting portion of the H distribution appears to consistently lie within the 0.45-~m range of the present measurements, it has not seemed necessary or fruitful to attempt to extend the measur~1ents to greater depths. A second sample allocation consisted of three Apollo 15 coarse fines. Handling and analysis procedures were the same as for the 10085 samples as described above. Sample 15413,5-2 is a 2 x 4-mm partially glazed pyroxene-rich crystalline rock fragment with a l x 3-mm area rich in pyroxene on one face. The sample was mounted, using a 2.4-mm aperture fused silica collimator~ to expose this pyroxene-rich area to the 19F beam. The H distribution measure- ment resulted in the profile shown in Figure 9. Once again there is a significant enhancement of H within about 0.2 ~m of the surface. The shape of the profile is quite similar to the 10085 distributions , with an absolute H content comparable to that of the 10085~1 glass fragment. Sample 15413,5-5 is a highly fractured fragment which appeared to be essentially all plagioclase. Due to its relatively small size, it was mounted using a 2.0-mm collimator. The usual analysis procedure resulted in a distribtuion which showed only a small enhancement of H, just barely above statistical uncertainties, in a layer a gain about 0.2 ~m deep. In order to determine whether this apparent enhancement was due to a real, although small, H content -32qualitatively similar to the distributions measured on the previous samples or merely the effect of statistical fluctuations, the counting statistics were improved by increasing the beam current to 50 nA and collecting 20 ~C of beam charge per data point. The higher beam current is expected to cause a more rapid dissipation of the H content; nevertheless, the measurement resulted in a reproducible profile quite similar in shape to the distributions observed in 15413,5-2 and the 10085 samples, but at a concentration smaller by more than an order of magnitude than the H concentration over the same depth region in 10085,31-12. Sample 15533,4-l is a glass coated breccia. fragment. glass is probably a glaze rather than a splash glass. but non-uniform in color and transparent in spots. The It is light The measured H distribution was quite flat with only a slight suggestion of an enhanced H content at ~o.l ~m deep and a peak at zero depth corresponding to a small (~5 x 101 4 cm- 2 ) surface H concentration. The profile obtained from these samples shows a wide variation in absolute hydrogen concentration, although the profile shapes are quite similar. The distribution measurements are summarized in the first part of Table 3. The limits shown on surface (zero-depth) concentrations pertain to the measured distributions. It is probable that somewhat larger surface concentrations may have existed prior to illumination with the l9F beam. The total hydrogen content of a 0-0.4 11m deep region (excluding this surface concentration) is also given in units of lol5 H atoms/cm 2 • The H contents -33at depths of 0~1 ~m and 0.4 ~m are given in ppm H by weight. The measured profiles are also characterized by the depth xp at vrhich the peak H content is observed, and the full-width at half-maximum of the H distribution obtained by subtracting (in quadrature) the vidth of the resonance from the width of the peak in the measured excitation function. Since the uncertainties in the values given for the H content at various depths in a particular sample are affected by the aperture size of the fused silica collimator used with that sample, the aperture size is also indicated in Table 3. The mobility of the H distributions observed in these lunar samples, especially in the outer 0.2 ~' is much greater than had been observed for the proton implantation profiles in feldspar, quartz, and fused silica test samples. One possible interpretation of this behavior is that we may be observing a distribution of terrestrial u2 o contamination absorbed into a porous surface region. ~pstein and Taylor (1970) have shown that terrestrial H2 0 exists in returned lunar soil samples in amounts comparable to their solar wind hydrogen content. Unless carefully cleaned, laboratory glass and other test samples exhibit a contamination hydrogen peak at z ero depth but show no penetration of the surface, within the resolution of the measurement. However, this may not be true for the much more shocked and highly radiation-damaged lunar sample surfaces. A number of tests have been carried out to investigate t his poss ib i lity. Samples 10085,31-12 and 1 5413, 5- 2 wer e subjected to a series of heating experiments with the i d ea that ba king the -34samples in a vacuum under certain conditions might drive off any superficial H2 o contamination without disturbing the implanted hydrogen. Both samples were baked for 100 hours at 150°C (simulating the conditions of lunar noon) at a vacuum of ~lo-9 Torr. Sample 10085,31-12 was subsequently baked for 24 hours at 290°C. Neither sample exhibited any significant change in the II distribution after baking. Hovrever, H2 0, apparently of terrestrial origin, is still retained in bulk lunar soil samples even at temperatures of 400-500°C (Epstein and Taylor, 1970). The H distributions in three of the above samples (the 10085,1 fra~nent and samples 1 5413,5-5 and 15533,4-1) were remeasured after a prolonged storage period (~6 months) in a vacuum desiccator. All three s~1ples had developed a relatively large (~1016 H atoms/cm2) surface conta!llination, but this contamination appeared to be confined to the surface with little, if any, penetration beyond depths of a fe'" hundred angstroms. Tvro of these samples (10085,1 and 15533,4-1) were also given an ultrasonic rinse in high-purity acetone with no observable change in the H distributions as a result of this treatment. 2. Apollo 15 glass-coated rocks 15015 and 150~2 Samples were obtained from two Apollo 15 glas s-coated breccias (rocks 15015 and 15059) sampled near the Lunar Module (LM) landing site. Sample handling procedures were modified to minimize contamination, especially from atmospheric H2 o . These rocks had been 1rrapped only in unsealed teflon bags before being placed in one of -35the cloth sample collection bags for the return trip from the lunar surface. Hence they were exposed to atmospheric humidity in both the LM and Command ModW.e (CSM) cabins and on earth prior to arrival at the LRL. LRL processing, however, is performed in a dry nitrogen atmosphere. The sealed sample containers received from the LRL w·ere opened in our laboratory in a P 2o -dried, nitrogen-filled dry box 5 and mounted in aluminum sample holders. These sample holders were then sealed in a clean vial and the vial was triply sealed using polyeth;rlene bags and a second larger snap-cap vial. This package ''as then removed from the dry box and transferred to a polyethylene glove bag. The scattering chamber was then backfilled with dry nitrogen, the viewing window removed from the access port, and the · mouth of the glove bag was clamped to the access port allowing the glove bag to inflate with dry nitrogen. bag with dry ~ After flushing the glove 2 gas for several minutes, the sample package was opened inside the nitrogen-filled glove bag and the sample holders installed in their mounts on the target wheel. The glove bag was then removed, the window immediately replaced and sealed, and the scattering chamber evacuated. This procedure eliminates any contact with water vapor, other than small residual amounts in the "dry" nitrogen gas, from the time the rocks were introduced into the LRL processing line. The orientations of both rocks are well documented in lunar surface photography (Sutton et al., 1972). The 15015 samples (15015,39-1 and 15015 , 39-2) are adjacent surface chips from the lunar -36bottoJn side of the rock. Samples 15059,32 and 15059,28 are surface glass chips from the top and bottom, respectively, of rock 15059. The lunar top surface of 15015 is heavily pitted, while there are no detectable microimpact pits on the bottom surface, suggesting that surface exposure has been primarily, possibly entirely, in one orientation (LRL, 1971). The microimpact pit density on the top surface of 15059 is very low indicating a short surface residence time. There are no detectable microimpact pits on the lunar bottom of 15059 (Horz, 1971). Our own examination of samples 15015,39-2 and 15059,28 under a high power binocular microscope revealed no impact p its. Both samples are bubbly glass with some dust in cavities separated by smooth areas of a few millimeters square. Impact pits as small as ~50 ~m would have been easily observed on these smooth glass surfaces. These samples were analyzed in the same manner as the Apollo 11 and 15 coarse fines; however, a glass collimator (3.2-mm anerture) was needed only for 15059,32 since the area of surface glass was large enough (0.2-0.5 cm2) on the other samples to insure that only a negligible portion of the l9p beam would be "spilled" on the aluminum holders. The distributions obtained from the two 15015 chips are qualitatively similar (Fig s. 10 and 11), but are not at all like those obtained from the coarse fine fragments discussed above. lihile the latter samples showed evidence of a surface H concentration which disappeared rapidly with illureination by the flourine beam, these -37two rock chips displayed a larger and more persistent surface 0 concentration (within about 200 A below the surface) decreasing 0 rapidly with depth to a flat distribution within 1500 A of the surface 0 for 15095,39-l and 800 A for 15015,39-2. It should be noted that for 15015,39-2 the dimensions of the beam spot were increased to more than double the usual size (2-3 mm). Consequently, the beam current density was decreased to less than half the norn1al value. current was not changed.) (Beam The accompanying reduction in local heating of the target by the beam may partially account for the relative stability of the H distribution on this sample. The two chips from rock 15059 exhibited the same type of H distribution as was observed on the 15015 samples, concentrated mainly within a few hundred angstroms of the surface (Fig. 12). The chip from the top of the rock (15059,32) appeared to show somewhat greater penetration of the H distribution than the sample from the bottom surface (15059,28) reaching as far as 0.25-0.30 vm before decreasing to the same level as the flatter 15059,28 distribution. However, the amount of excess hydrogen in t he region 0.05-0.30 ~m of 15059,32 when compared to 15059,28 is only ~10 1 5 atoms per cm2 , indicating no strong solar wind excess in the top sample. The 15015 samples were turned over (under dry nitrogen) to measure the H distribution on the interior rock surfaces. These surfaces were created in the process of breaking the samples from the narent rock and thus have never been exposed on the lunar surface. Since the chipping procedure was performed after introduction into -3Uthe LRL processing line, these surfaces have received no exposure to atmospheric humidity, but only to dry (20-50 ppm H20) nitrogen. The slightly smaller (~2xlol5 atoms/cm2 ), more superficial surface H concentrations detected on each of these samples must be interpreted as contrunination, probably H20, even though these surfaces were exposed only to a dry nitrogen atmosphere. While the measured distri- butions for the lunar exterior surface glass coatings are not inconsistent with a relatively unmodified solar wind proton implan0 tation profile (with a mean penetration depth of ~200 A), it must be concluded that at least a significant portion (if not all) of the H concentrations observed on these surfaces (within the resolution of ~ 200 0 A) is due to H2o contamination. Data from these samples are also sununarized in Table 3. small (~40 'I'he ppm) H contents observed in the interior rock faces can be taken to represent the volume H content of the rock, since the flux of solar flare protons with sufficient energy to penetrate to these depths (several n1illimeters below the surface) is a negligible effect and since we have no evidence that H20 contamination can penetrate a fresh, undamaged silicate surface to depths greater than a few hundred angstroms. Each of the 15015 samples was removed from the scattering chamber and exposed to the laboratory atmosphere for at least 24 hours. Remeasurement showed no significant change in the H distributions in the lunar exterior surfaces. However, the initial exposure of these rocks to the ~1 cabin atmosphere may have saturated -39an active surface layer so that subsequent exposure would produce no additional contamination. Wi th a solar wind proton flux of 6xlol5 per cm2 per year, the time necessary to accumulate the total H contents observed in the outer 0.4 ~m of these glass-coated samples is on the order of only one year, implying that either these samples have all had extremely short exposure times or the implanted H has been essentially totally lost. In order to determine whether keV protons implanted in the lunar glass should be expected to be retained, the glass coating of 15015,39-2 was implanted with a dose of 2x101 6 protons per cm2 at an energy of ~12 keV. The resulting H distribution was quite similar to the distributions obtained by implanting fused silica and crystalline silicate targets under similar conditions (see Section II, Part D), with quantitative retention verified within the uncertainties of the measurement. The artificially implanted distribution did not exhibit the high mobility under irradiation with the l9F beam characteristic of the II distributions at similar depths in the 10085 and 15413 samples, but behaved in a similar manner to the distributions implanted into terrestrial silicate samples. This implies that the differences in the mobility of the H distributions in the 10085 and 15413 samples as compared with the artificially implanted distributions are probably not the result of bulk physical or chemical differences between the materials in question. -403. Apollo 16 gl ass- coated rocks 64455 and 65315 Rock 64455 is a heavily pitted ellipsoidal glass-coated feldspar-rich crystalline rock (Grieve and Plant, 1973). Sample 64455,24 is a 1-cm chip of the glass coating from the rock's western edGe adjacent to an exposed patch of light-colored rock. Sample 64455,33 consists of two similar ~ em surface chips taken from this 2 light-colored area. Rock 6531 5 is described as friable white anorthosite with patches of relatively thick black glass coatings (LRL, 1973). These patches appear to be remnants of a more extensive glass coating which has been chipped away by repeated impacts to expose the underlying rock. The lunar top and northern faces are the most obviously pitted, but pits may also be present on the bottom face near the northern edge and on the southern face (LRL, 1973). The 65315,6 and 65315,20 chips are 1 em and !ern samples of the patches of black glass coating from 2 the southern and top faces of the rock, respectively. 65315,8 is a 1 ~em chip of the white anorthosite sampled adjacent to the 65315,20 ,_ ') glass chip on the lunar top of the rock. In obtaining samples of glass and crystalline material from the same rock, our primary objective was to investigate whether the crystalline state of the medium would have any significant effects on the retention of implanted solar wind ions. However, prompted by a suggestion from A. Turkevich (1973), it was decided to measure fluorine depth distributions on these samples as well. The in situ chemical analysis of lunar highlands material performed by the -41Surveyor 7 alpha-particle scattering experiment revealed a surprisingly high fluorine content (0.26 ± 0.11 percent F b~r weight) in two soil samples (Patterson et al.~ 1970). (0.07 ± Somewhat less fluorine .2 percent) was detected in a rock analysis although statistical uncertainties were quite large. Analyses of returned lunar soils have not shown such high fluorine contents~ suggesting that fluorine may have been concentrated in micron-thick layers at the surface of the Surveyor 7 samples. Turkevich suggested that this fluorine could be the result of volcanic exhalations or gases from impact volatilization depositing on the surfaces of soil grains~ and that the anal~rsis technique using the 19F(p~ay )lGo reaction could be used to investigate this possibility. Solar wind implantation could be considered as an alternative~ although less likely~ source. Since the Apollo 16 site 11as in a lunar highlands area geolog ically similar to the Surveyor 7 site, the Apollo 16 samples presented a good opportunity to look for surface fluorine enhancements. The atmospheric exposure and sample handling history of the 64455 and 65315 samples was virtually unchanged from the description g iven above for the 15015 and 15059 samples. The H distributions were measured first on the 64455 samples with the results shown in Figures l3 and lll. The "exterior surface" data are measurements on the lunar exterior faces of the samples. As observed previously, the H in the first few hundred angstroms is rather labile, while the relatively small H contents at greater depths are much more stable. 'rhe G4455 distributions are all characterized. by a maximum within tvro -42or three hundred angstroms of the surface, similar to the profiles measured on the 1501 5 and 15059 samples. There is apparently no significant difference between the glass srunple (64455,24) and the rock samples ( 64455,33). the interior samples all shovr a concentration of h;rdrogen within a few hundred angstroms of the surface which is systematically lower, although by less than a factor of two, than on the corresponding lunar exterior surface. For 64455,33-1 the H concentration rer.~ins higher in the exterior srunple to a depth of at least 0.4 ~m, while for 64455,33-2 and the 64455,24 glass, the exterior sample shows no excess below about 0.2 ~m compared to the interior samples. Once again, the relatively large H concentrations observed within a few hundred angstroms of the surfaces of the interior samples must be interpreted as terrestrial contamination since these surfaces were freshly created in chipping the samples from the rock in the LRL processing line. These surfaces have been exposed only to dry nitrogen, providing additional support for the conclusion, made as a result of similar observations on the 15015 and 15059 samples, that even the small residual H')O content of the "dry" nitrogen is c;_ sufficient to significantl:v contruninate these surfaces. All of the 64455 samples as well as the 15015 and 15059 samples, both interior and exterior, show an H content of ~20-50 ppm bv weight at depths of 0.2-0.4 ~m. Although relatively small, these concentrations are distinctly above background. (Compare with the countinG rates for depths below zero in Figures 12-14). For the -43exterior samples, inward diffusion of solar wind hydrogen or direct implantation of 10-100 keV protons are possible sources; however, these cannot explain the results for the interior samples unless diffusion has extended essentially throughout the volume of the rock. It app ears more reasonable to conclude that we are observing either indigenous lunar H o r "inherited" solar wind hydrogen which was pr esent in the materials f rom which 64455 was formed. Following the measurement of the H profiles, a 30-nA proton beam was used to measure l9F depth distributions . It should be noted that the penetration depths of the 16-18 MeV 1 9F ions used in the H analysis (~8 urn) are well beyond the 1-um range of these F distri- bution measurenents and hence should not be expected to contribute to the measured F contents of this 1-um layer. The results of 1 9F depth profile measurements on the 64455 samples are sho;.m in Figures 15 and 16. The 64455,24 glass sample showed a fluor i ne-rich zone at an apparent depth of 0.35 urn and a width of ~0.2 urn. Scanning the proton beam across the surface of the sample showed that this fluorine-r i ch zone is apparently not parallel to the surface, although the profil e was not measured in detail for each beam position. The 64455,33-1 sample showed a peak within 0.1 urn of the surface, but this peak was not reproduced in the "second run" data. For thi s remeasurement the peak had apparently shifted inward to 0.1-0.15 urn deep. This apparent shift could be interpreted as a volat i lization of fluorine from the surface due to irradiation wit h the proton beam, but it could also be due to a build-up of a n -44electrostatic charr;e on the silicate sample as the irradiation pro.::;r('sseu. A. surface potential of a few kilovolts vrould slow down the incident protons enough to cause a significant difference between the measured beam energy of the protons as they strike the sample. (In the calculation of 'the depth scale, it is assumed that this difference is negligible.) Since the average surface potential on the sample necessary to cause an apparent inward shift of 0.1 ~m is 5.5 kV, this explanation does not seem unreasonable. The charging of a 1.5-cm diameter fused silica disc was measured by observing the shift in beam energy (relative to the resonance energy) necessary to obtain the maximum counting rate from a thin (5 ~g/cm2 ) evaporated CaF 2 surface coating. The average potential was determined to be 10 kV with fluctuations of ±5 kV, implying a Inagnitude for this charging effect on the 64455,33-1 sample (probably somewhat smaller than for the fused silica target) in reasonable agreement with the observed shift in the peak of the F distribution. (This effect is not important for the hydrog en measurements since the apparent inward shift of the H distribution resulting from a 10 kV surface potential is less than the 0 resolution of ~200 A.) It would appear more difficult to ascribe the peak in the measured 64455 ,24 F distribution to a surface (zero-depth ) concentration of F , since this would require an average surface potential of ~20 kV for this sample. Although this seems unreasonably high, the possibility cannot be ruled out at this time. Data f or the lunar exterior surface of 64455,33- 2, although i ncomplete , shm• no fluorine enhancement above a constant level of -45~50 ppm applicable to all three 64455 exterior samples. The interior surfaces of these samples also showed relatively uniform F distri0 butions, except for small concentrations within ~500 A of the surface, <rith the 64455,33-2 interior having a relatively high F content of ~100 ppm. The fluorine data for these samples are summarized in Table 4, giving the apparent depth and width (FHI-ll1) of the peak in the distribution and the total F content in the 0-0.5 ~m and 0.5-1.0 ~m deep regions. The H distribution data are again summarized in Table 3. For the 65315 samples, F distributions were measured first to eliminate the remote possibility that a portion of the observed F content could be due to 1 9F ions implanted during the H measurements migrating out to within 1 ~m of the surface. An additional procedure i nvolving measurement of the yield of 11-MeV y rays from 27Ai(p,y)28si near the 992-keV (proton energy) resonance was also introduced. This reaction, although much weaker than 1 9r(p,ay) 1 6o, produces most of the counting background for the F analysis measurements due to the relatively high Ai content of the lunar samples. If the Ai content is reasonably uniform in a 'Vl-JJm surface layer, a step in the 11-Mev y-ray yield is observed as the proton energy is increased past the 992-keV resonance. Any observable shift in the location of this step can be attributed to charging of the silicate sample, assuming that thin (~0.1 ~m), A2-free coatings can be ruled out. If a significant surface potential exists during the F distribution measurements, it should be detected by this procedure since the experimental conditions are identical, except for a ten percent difference in the proton beam energy . -46r1easurement of the F distribution in the 65315,6 glass chip revealed a much larger F content (Fig. 17) than was observed in the 64455 samples. An upper limit of ~5 kV can be placed on the surface potential during irradiation as a result of the 27A~(p,y) 2 8Si measurement described above; however, this limit is not strong enough to rule out the possibility that the observed peak is actually a surface (zero-depth) concentration apparently shifted to a depth of 0.1 ~m by a surface potential of ~5 kV. The peak concentration corresponds to ~2000 ppm F, with an average over the outer 0.5-~m layer of ~1000 ppm. A remeasurement of the distribution resulted in a n apparent decrease in the F content of this depth region by about a factor of 2, implying that F may have been volatilized by the proton beam. This discrepancy could also be due to a shift in the position of the beam spot on the sample, if large local variations in F content exist over the surface of the sample. However, a scan across the sample revealed no areas with a large enough F content to reproduce the ori8inal neasurement. The lunar interior side of this sample contained little F -- about an order of magnitude less than the exterior surface. The interior surface analyzed was mostly glass with small patches of white anorthosite. Interior and exterior surfaces of the 65315 ,20 glass chip were also analyzed for F, as well as an interior face of the white anorthosite chip 65315,8. These surfaces contained only small amounts ofF , compared with the G5315,6 exterior glass surface. Howev er, the -47near-surface enrichments of ~lo1 5 F atoms/cm 2 observed on the interior faces of both of these samples indicated that they have most likely been subjected to some form of terrestrial F contamination. Contam- ination with F has ~been observed on any of the fused silica or other test samples or on fresh interior surfaces of meteorites prepared in this laboratory. chondritic It can be concluded from this observation that either the lunar samples (both exterior and interior surfaces) have an exceptional affinity for some unidentified form of F contaminant associated with the sample handling procedures used in this study for the lunar, meteoritic, and terrestrial samples alike; or the contamination is introduced in the LRL processing line. A likely source is the freon used to clean the tools, work areas, and containers used on the Apollo missions and at the LRL. The teflon bags in which samples are packaged and sealed are another possible source of F contamination. The F data for these 65315 samples are again summarized in Table 4. H distributions were also measured and the data summarized in Table 3. Although the data for both exterior glass samples (65315,6 and 65315,20) show significant penetration of the H to depths UIJ to ~o. 3 l.l m, the prior irradiations to measure F distributions may have significantly depleted the original H content of these samples, rendering the measured profiles less accurate representations of the initial distributions than if the H measurements had been performed first. The large size of the 65315, 8 sample prohibits mounting of the sample with the present sample holders to irradiate the lunar exterior surface. Rather than break the sample into smaller pieces, it uas decided to store this sample until it could be mounted in a new, larger scattering chamber. 4. Apollo 16 ALSRC samples For a relatively small fraction of the returned lunar samples, the atmospheric exp osure in the 1M and CSll! cabins and on Earth prior to introduction into the LRL processing line has been avoided by sealing the samples in an indium-sealed aluminum vacuum container (ALSRC) on the lunar surface. Thus, although these samples are still subject to possible H 0 contamination from the astronauts' cooling 2 systems on the lunar surface and from the residual H20 content of the "dry" nitrogen gas to which they are subsequently exposed, exposure to normal atmospheric humidity is eliminated as a source of H20 contamination. Four Apollo 16 samples returned in one of these vacuum-sealed sample containers were obtained for this study. Three of these (66044,8; 68124,3; and 68124,10) are 4-10 mm coarse fines and the fourth (68815,27) is a chip from a large breccia boulder. The sample handling procedures used are the same as those initiated with the 1 5015 samples, avoiding any exposure to the laboratory atmosphere. Rock 68815 is a glassy, dark-matrix, dark-clast breccia in the classification of Wilshire et al. (1973). Track gradients, rare gas spallation ages, and microimpact pit densities give a concordant exposure age of ~2 x 106 years for this rock, which is itself a chip -49from the top of a large boulder (Behrmann et al., 1973). Sample 68815,27 is a surface chip with an exposed 2-mm x 4-mm light-colored region surrounded by darker gray material. A thin section prepared from small chips of the sample showed that the dark gray m.a terial was a very fine-grained matrix, perhaps devitrified glass, and the lightcolored material was mostly crystalline plagioclase. The H distri- bution measured on this sample (Fig. 18), characterized by a broad peak extending to ~0.2 urn followed by a more gradual decrease in the H concentration with increasing depth, is qualitatively quite similar to the distributions measured in the Apollo ll and 15 coarse fines, most notably 10085,31-12; 10085,1 and 15413,5-2. The total H content of 2.3 x 1016 atoms per cm2 is second only to the 5.5 x 1016 per cm2 observed in the 10085,31-12 brown glass fragment. The lunar interior surface shows only a superficial surface H concentration, similar to the distributions measured on the interior surfaces of the other lunar rock chips, in addition to a uniform H content of only ~20 ppm in the 0.4-um deep measured region. Figure 19 illustrates the evolution of the distribution in 68815,27 during the irradiation w~th the l9p beam. Three successive data runs are presented with individual data points for a given run connected by straight line segments, except at the peak where a some,.,hat arbitrary curve has been drawn to estimate its shape. Despite the low beam current density (< 1014 ions/cm2 in about 10 minutes for each data point), the H content is significantly reduced for each successive r:1easurement. A slight suggestion of a shoulder in the -50profile at ~0.3 ~m in peak by the third run. the first run has evolved into a definite second The slope changes more abruptly at a progres- sively shallower point from run to run, suggesting that the first peak represents a relatively loosely bound H component superimposed on a deeper, broader, more tightly bound distribution. This is the same sort of evolution of the profile under bombardment observed on Apollo ll and 15 coarse fine aamples. The light-colored region on the lunar exterior surface of 68815, 27 is predominantly feldspar, so a significant fraction of the exposed material (possibiy beneath a thin metamictized layer) is crystalline. A scan across the surface of 68815,27 showed no significant variation in the H content below ~a.l ~m, indicating no observable difference in the H distribution between the feldspar and the glass matrix. Using a similar technique employing a resonance in the reaction lH(7Li,y)8Be (Padawer and Schneid, 1969), a research group at Grumman Aerospace Corporation has also detected a concentration of H within ~o.4 ~m of the surface of a sample from rock 68815 ( Pada,·Ter, 1973; Stauber et al. , 1973) • The sample subjected to this analysis (688l5,25) was quite similar to 688l5,27 in that a lightcolored clast was embedded in the analyzed exterior surface. Since this technique is only ~6 percent as sensitive (in terms of counts per incident ion per ppm H) as the lH(l9p,ay)l6o technique, in addition to having a poorer depth resolution (by a factor of ~6), statistical uncertainties and resolution limitations permit only qualitative -51comparisons with the results of this study. Over the same depth ranBe, the 1H(7Li,y) 8 Be data (5 data points) are not inconsistent with the 68815,27 profile measured in this study, however the measured H content is lower by a factor of ~3 relative to the 68815,27 results. Although variations of this magnitude across the surface of rock 68815 are not unreasonable, it is probable that the H content of 68815,25 prior to the 7Li irradiation was somewhat higher than the measurements indicate. The mobility of the distributions observed in the present study may be greatly magnified b~r the much higher ion fluxes and total doses per datn point used by Stauber et al. The analyzed area of 68815,25 sustained a power density input of ~10 watts/cm2 (Stauber et al., 1973), ,.,hile the analysis of 68815,27 in the present study involved only ~0.5 watt/cm2 • Stauber et al. also subjected a splash-glass chip from rock 61016 (61016,135) to the 1 H(7Li,y)8Be analysis. A rather large (~2 x 1016 atoms/cm2 ) concentration of H was observed within ~0.1 ~m of the surface, with an H content of 100-200 ppm from ~0.1 ~m to ~2 ~m deep. The three coarse fine samples from the Apollo 16 ALSRC obtained for this study consisted of two crystalline anorthosite fragments (66044,8 and 68124,10) and a dark brown glass spherule (68124,3). •rwo opposite surfaces were analyzed on each sample. Surface A on each sample was analyzed for fluorine before performing the H distribution measurements, while the H analysis was done first on the B surface. -52Sample 66144,8 is a 5-mm blocky anorthosite fragment, Surface A includes a l-mm2 gray area which may be glass, while surface B appears to be essentially all anorthosite, usine a 3.2-mm glass collimator. The sample was mounted Neither surface revealed an H distribution which was remarkable in any way. A superficial surface H concentration of ~2 x 1015 per cm2 on B was not observed on A; however, the previous irradiation of surface A for F analysis probably dissipated any superficial surface H from this surface. Only slight enhancements of the H content at depths greater than ~o.l ~m were observed on these surfaces, Surface A of 68124,3, a 5-mm glass spherule, appeared to be quite smooth, Hhile several small pits could be seen in a cursory inspection of surface B. The H profiles measured on these two opposite surfaces are virtuall~r identical, and quite similar to the data shmm in F' i gure 18 for the 68815,27 interior surface. Similar results "1-rere obtained for the two analyzed surfaces of 68124,10, a 5- mm x 7-mm anorthosite fragment. The only notable difference in the H analvsis for the 68124 samples is the slightly higher uniform H content in the 0,4-~m deep measured region for 68124,10 (30 ppn) as compared Pith 68124,3 (10 ppm). Figure 20 shows the F distributions measured on both exterior and interior surfaces of 68815,27. The dominant feature is a narrow (.:; 0.1 ~m) peak ~0.15 ~n deep with a tail extending to o,8 the exterior sarrrole. Hoving the proton beam across the surface indicated that this sharp peak apparentl~r has a sizeabl e (~ ~m deep 1 mm) in -53lateral extent at a relatively constant magnitude. This thin F-rich la7 er may be a surface (zero-depth) F concentration which appears to be shifted inward due to a surface potential of several kilovolts, as discussed earlier. The excellent reproducibility of this distribution {data from t'.-10 consecutive energy scans are plotted in Fi~;ure 20), requiring an extremel;.r stable surface potential, mal,es this interT'retation aprear somewhat unlikely. Hmrever, the 27AQ.(p,y) 28si measurer.1ent uas not performed on this sample, so no definite conclusions can be nade concerning the extent of the c har g ing effect. Contamination appears to be a problem Hith fluorine even •·ri th samples from the sealed sample container. The F profile for the 68815,27 interior has a peak ,.Tithin 0.1 -urn of the surface. Further, a remeasurement of this distribution after an interval of about one '·reek resulted. in an observed F content twice as large as in the ori::;inal meas ureraent sho•m here. During this interval the sample . re1nained undisturbed in the scattering chamber and ,.ras exposed onl ~r to dr~r ni troc;en gas 61~ 45 5 (for "-2 hours) and a vacuum of "-lo-9 Torr. The samples, however, showed little change over this same per iod. ao H contamination accompanied this increase in F content. The source o.f the appar e nt contamination has not been identified ; hovever, it may be that this apparent increase is due to large local variatlons in F' c oncentration from prior contamination, with an area of higher F contamination, not 1·rithin the beam s pot for the first analysis, illuminated during the reroeasurement. An exceptionally high F content was observed on 66044,8-n, avera~ing nearly 2000 p~m Fin a layer 0.5 ~m deep (Figure 21). High r contents (1400 ppm) continue into the sample to at least 1.0 ~m deep. An intentional shift in the position of the beam spot on the sample shows a sip;nificant variation in the F distribution, but the distribut io n men.sured on a particular spot appears to be reasonably stable under continual bombardment. 66044,8-A contains somevrha.t less F than the 66044,8-B surface, but the 820 ppm F within 0.5 ~m of the surfac e i s st i ll 0uite high compared with most of the other samples listed in Ta~le 4. Although the peak F concentration in 66044,8-A appears to be located a.t a somewhat greater depth than in 66044,8-B (0.10 ~m com:.nared with 0.03 ~m), this difference ca.n be explained entirely by the surface potentials measured using the 992 keV resonance in 27A~(p,y) 2 8Si a.nd interpreting the peaks a.s true surface (zero-depth) concentrations. Average potentials of 7 ± 3 kV a.nd 2 ± 2 kV vrere measured for surfaces A and B respectively, as shown in Figure 22. The sharper sten observed for surface B indicates that fluctuations about the average potential were not as great a.s for surface B. The resonance Hidth is only '\.0.1 keV; however, the estimated energy spread of the incident proton beam ('\.l keV) makes a significant contribution to the '.vidth of the step observed for 66044,8-B. Although additional hroadenine could be caused by a. large depletion of A~ within a few hundred an~stroms of the surface in a particular sample, this alternative appears quite unlikely. Thus, additional broadening beyond that due to the energy spread of the proton beam must be interpreted as due to kilovolt fluctuations in the surface potential -55on the sample during irradiation with the proton beam. mo~t Although the likely interpretation for the peaks in the F distribution appears to be surface contamination, the large (1400 ppm) amounts of F observed below this surface concentration on 66044,8-B must be interpreted as an exceptionally large volume concentration of F. It is unlikely that such large a~mounts of terrestrial F contamination could penetrate so deeply (up to at least 1. 0 lJm), especially considering the lack of a penetrating tail on the surface peak. Conse~uently, the best inter- pret ation for the observed large concentration of F in the outer micron of the 66o4l+, 8-B surface is that it is lunar F. However, a small particle of teflon lodged on this surface could conceivably account for t hese results. The F distributions measured on the two opposite surfaces of 68124,3, the glass ' spherule, are qualitatively quite similar (Fig. 23); however, the amounts differ by about a factor of two. The apparent depth (""0.1 l-Im) of the peak F concentration can be explained entirel~r b~r interpreting the peak as a surface concentration with an apparent imrard shift of more than 0.1 l-Im due to the surface potentials measured usinp: 2 7AR.(p,y)28si (see Figure 22). For sample G8l24,lO, comparatively little F was observed in either face, but small peaks were observed in both. 0.17-l-1~ The apparent depth of the peak in 68124,10-B can again be explained by the surface potential due to charging inferred from the 27AR.(p,y)28si measurement (althou~h insufficient data make the 14 kV estimate highly uncertaiL), interpreting the peak as a surface concentration. The PROPERTY OF K:"l LOGG RADIATION LABORATORY -56surface potential determination was not performed for 68124,10-A; however, the apparent 0.09 ~m depth of the peak could easily be explained by shifting of a surface concentration due to a potential of "-5 kV. 5. Simulation experiments Borg et al. (1971) have observed metamictized coatings 0 (rendered amorphous by radiation damage) 200-1000 A thick on lunar soil grains by high voltage electron microscopy. Although the boundary between the amorphous layer and the underlying crystalline grain generally appears quite sharp, high nuclear particle track densities (> 1011 tracks/cm2 ) indicate that radiation damage may still be quite severe (but below saturation levels) to depths of a few tenths of a micron. If this type of radiation damage can render a silicate material porous to water, then the deep (-v0.2 ~m) H concentrations observed in several of the lunar samples, most notably 10085,31-12 and 68815,27, might be explained as terrestrial H20 contamination ~enetrating the surface of the lunar sample to the limit of the porous region. In order to check the possibly greater extent of H2 o !)enetration into heavily radiation damaged silicate naterials relative to undamaged samples, a set of fused silica samples was first baked at 300° C in vacuum to drive off surface contaminants and then irradiated "rith 86 keV l6o ions. The irradiation doses of 2 x 101 6 2 to l. 4 x 101 7 ions I cm (into a ~ cm2 area) should be more than enour:ll -57to drunage a layer 0.20-0.30 )Jm deep to saturation (Hines and Arndt, 1960 ; vlinterbon et al., 1970). Some of the artificially damaged fused silica surfaces were exposed to H2o and then subjected to H distribution measurements. One set of irradiated samples was exposed to H20 in va:ror form only (laboratory atmospheric humidity for "-l 'reek) and a second set was given a 24-hour exposure to distilled ti,.,O in addition to atmospheric exposure. A third set was kept under '- vacuum, except for a two-hour exposure to dry n gas . 2 Unirradiated samples vrere included with each set as controls. Results of the subsequent H distribution measurements indicate that a small amount of E 0 (101 5 to 2 x 101 5 H atoms/cm2) 2 has penetrated to depths of up to 0.10 )Jill (Fig . 24) in all of the irradiated samples exposed to water in either liquid or vapor form. The distributions are similar to the H profiles measured on the 1 5015 , 15059 , and 64455 lunar samples, except the concentrations are smaller 1)y less than a factor of 2 than the surface concentrations on thes e lunar samnles and drop to zero by NQ.lO )Jill deep in even the most heavily irradiated samples. '.L'he depth of the H2 0 penetration is not comparable to the expected extent of the damage saturated region; howev e r, scanning laterally across the irradiated samples indicated that penetration of the H o had occurred mainly in the area irra2 diated by the lGo beam. The samples exposed only to dry N2 for a short time showed no surface n2 o (~ 101 4 H atoms/cm2 ). These res ults provide additional evidence for the identifi cation of the near-surface H concentrations in 15015, 15059, and -5864455 lunar samples as terrestrial H2o. It seems improbable, in this light, that the large 0.2-wm deep concentrations on 68815,27 and some of the Apollo ll and 15 coarse fines could be terrestrial 6. Apollo 16 LSCRE foil The hydrogen distribution was measured on both sides of an aluminum-coated platinum foil from the Apollo 16 Lunar Surface Cosmic Ray Experiment (LSCRE). The side with the aluminum coating (~0.1 ~m thick) had been exposed to the April 18, 1972 solar flare event, which lasted for the first half of the mission. Both sides showed a large 0 concentration of hydrogen within ~200 A of the surface -- about 6 x 101 G per cm2 for the exposed side and 3 x 101 6 per cm2 for the other side. This difference could be due partially to the solar flare exposure, but unequal surface contamination should be expected, especially considering that the two surfaces are of different metals. There is no evidence of any implanted hydrogen above that from H contamination, but, using the H observed at depths greater than 0.1 ~~ an upper limit can be set on the exposure to 10-40 keV protons if one assumes that all such particles are retained i n the platinum at the end of their range. ener~J The y-ray yield is quite flat over the l9F range corresponding to the implantation depths for l0-40 keV protons. Assuming that the difference in yield between the two sides of the ~oil over this range is due entirely to implanted solar flare protons, the implied integrated flux of protons in the l0-40 keY range -59is 101 5 per cm2 • At least part of the difference, how·ever, can be attributed to the off-resonance ~rield from the higher surface concentration of hydrogen on the exposed side. So at best we obtain an upper limit of about 101 5 protons per cm2 • This limit is questionable, however, due to the doubtful assumption that all implanted protons are retained at the end of their range. An artificial implantation of 20-keV protons was performed to test this assumption and resulted in about 40% retention at the implantation depth vrhen the distribution was measured several hours after the implantation. Under illumination by the l9F beam, however, the diffusion process was much more rapid than had been observed in implanted glass samples, so the effects of diffusion during the measurement of the distribution may have led to considerable losses of implanted H. Data from the Apollo 16 Lunar Surface Cosmic Ray Experiment (Burnett et al., 1973) imply that if an inverse cubic proton energy spectrum is extrapolated from the MeV range down to 10 keV or lower, the integrated flux of 10-40 keV protons on the piece of aluminumcoated platinum foil should have exceeded lo16;cm2 • The upper limit of 101 5/cm2 suggests that the spectrum diverges from an inverse cubic behavior at an energy some;.rhat higher than 10 keV. Our implantation results, hmrever, suggest that the rE1tention of the solar flare protons at depths corresponding to 10-40 keV proton ranges may have been poor. Consequently, the above conclusion on the energy spectrun cannot be strongly defended. -60B. Discussion l. Depth distributions of H in lunar samples In sunman·, three major features have been observed in the H distributions of one or another of the lunar samples analyzed in 0 this study: (1) a narrow concentration peaking within ~200 A of the surface, (2) a broad distribution extending from the surface to ~0.20 ~m deep, and (3) a still broader distribution apparently peaking at ~0.25 ~m and These features can extending at least to a depth of 0.45 ~m. coexist in a given sample, ~-Ti th ( 3) appearing as a shoulder on the tail of (2) followed by a more gradual decrease in H content with increasing depth. Samples 10085,31-12 (Fig. 7) and 68815,27 (Figs. 18 and 19) show these deeper features most clearly, while the 64455 samples (Figs. 13 and 14) exhibit typical narrow surface (within a few hundred angstroms) concentrations. The obser- vation of this same feature on the "interior surfaces" of the lunar rock samples is the primary basis for the interpretation of this t~pe of distribution as terrestrial contamination, probably H2 o. The consistently higher (except for 65315,20) surface H concentrations on the exterior surfaces compared to the interior surfaces, although possibly due to small amounts of implanted solar wind H, can be satisfactorily explained by a greater affinity for H2 0 resulting from solar wind radiation damage even apart from the more extensive exuosure to terrestrial H2 0 experienced b~ the exterior surfaces. Cadenhead et al. (1972) T;JOint out that a t~rpical gas will achiev e monolaYer coverage in seconds even at pressures as low as -61lo-G ~orr. Hence, the observation of approximate H2 o monolayer eauivalents ("-2 x 101 5 H atoms/cm2 ) on the surfaces of samples from the sealed rock box (ALSRC), 1·T hich have been exposed to "dry" nitrogen (20-50 ~pm H2 o), although not to room atmosphere, is not remarkable. 'i'he distributions observed on the interior surfaces of the luna r rock samples, 1vhich are almost certainly due to terrestrial :r2 o conta'Tlination, are all consistent with coverage in the monolayer rec:ion. (Cor.r;: 'are the calculated appearance of an H2 o monolayer shown in F i gs. 1 3 and 14.) Thus, it should be expected that coverage (ph:rsisorption) in the monolayer region should have taken place on the exterior surfaces as well. Water vapor adsorption studies (Cadenhead et al., 1973; Cadenhead et al., 1972; Fuller et al., 1971) ir.1ply that a brief exposure to air results in an irreversible a dsorpt ion of >rater vapor onto previously pristine lunar sample surfaces. This irreversible adsorption is interpreted as a chemi- sorption since temperatures above 300°C are required to release the adsorb ed 11 2 0 ; however, the extent of the chemisroption also appears to be limited to about a monolayer equivalent, implying that the chemisorption does not significantly penetrate more than a few atomic la;rers. Fuller et al. (1971) find no evidence of inherent porosity 0 in the 500-A thi ck metamict coatings on lunar dust grains observed b~r Borg et al. (1971). Hence, the somewhat larger, more strongly bound ::>urface H concentrations observed on the exterior surfaces of rocks 15015, 1 5059 a nd 64455 as compared to interior roc k surfaces can be interpreted as H2 o chemisorbed by a surface layer rendered -62si.r:nificantlv more reactive than the f r esh interior surfaces as a resul i, of solar 'l·rind. exposure. (Some other effect of the reducing env irorunent of the lunar surface may be necessar~r to explain the apnarent cheraisorntion on samples from the unpitted bottoms of rocks 15015 and 150)9 if, indeed, these samples have had no solar Hind exTlosure.) A :!Jhysisorbed coverage comparable to that observed on the interior surfaces is probabl:v superimposed on the more s tronc;lv bound. chenisorbed component. This interpretation is supported by the results of the tests performed in this study to simulate the surface conditions due to radiation damage from ions of keV energi es . Although the irradiations of fused silica samples 1-r:i.th 5 .1~ keV I ar.m 16o i ons are expected t o cause severe radiation damage over a depth range of ~0.25 ~m, exposure to atmosphere resulted in an adsorption o f H2 0 in monolayer ouantities with penetration confined essenti all:v to a re~ion evidence of penetration to a depth of 0.1 0 0-500 A deep. ~ Only slight was observed. The observed II concentrations were somewhat larger and more strongly hound on the irradiated samples exposed to H2 o than on samples which had not been irradiated. (A similar observation has already been made concerning the H distributions on exterior surfaces of lunar rocks as C01:t-parcd to fresh interior surfaces.) A chemisorption has apnarcntly occurred on these simulated surfaces as well as the actual lunar sample surfaces ; however, the penetration of the che~isorbed H00 does not agree vith the radiation damage rang e of the 16o ions. '-· '.l.'h is sugr;ests that whatever chemical alteration of the samples is -63involved, it is essentially a surface effect of the irradiation, rather than a volume effect. A likely candidate is sputtering. The surfaces of lunar samples have been described as non-stochiometric and o~rgen-deficie nt (Gaw~ge and Becker,l971; Epstein and Taylor, 1<)71) and this reduced state has been attributed, in part, to fractional vaporization (by solar wind sputtering and/or impact volatilizat ion) of oxygen relative to heavier atoms. has probabl~ This effect a lso depl eted the surfaces of the irradiated fused silica samples in 0 relative to Si (vrhile an oxygen excess results at r,r eater denth due to the use of 16 o in the irradiation). 'l'he chemisorntion of H2 o, then, may be part of an oxidation process in which stochiometr:v is a t least partially restored to the surface rer;ion from which the sputtered atoms vrere removed. 'i'he lack of a strong lv bound surfFJ.ce H concentration on the 10085, 15413 anu 1 )533 sarnnles, 1rhich had been exposed to room atmosphere for a feH months nrior to anal~sis, ma~ he the result of this oxidation r>ro cess f.!;Oinp; to completion with the release of the H in the form of H?. c;as . "'or the 6531 5 samples, the prior irradiation (for F analysis) vas probably sufficient to dissi pate an:r surface H concentration. r: o'.;ever , the 681 24 sa.1'1lples also appeared to lack a strongly bound surfac e H con:'Jonent, s ince the sr.1all (lol 5 E/cm2 ) surface concentrat ions diss i r>ate<l rapi<il:'r on illumination wi tll the 19F beam. This nrovides an encouraging suggestion that perhap s the "dry" N gas to 2 uhi ch these sampl es 1.rere exnosed has a low enough vrater content to prevent a si gnificant chemisor:r,Jtion of a2o . -64The deeper features (~500 A deep) observed in the H distributions of the 68815,27 rock sample, the three 10085 glass fragments, and the 15413,5-2 rock fragment (and to a lesser extent in a few other samples) are almost certainly the result, either directly or indirectly, of exposure to solar radiation; although it is not clear exactly how these features are related to the irradiation history of these surfaces. Four possibilities will be considered here: (1) adsorption of a terrestrial contaminant (H 2 o) by a region which has been heavily damaged by solar radiation, (2) an inward bulk volwne diffusion of solar wind H, perhaps reaching an equilibrium with the surface erosion rate, (3) population of a distribution of radiation d~~age traps by solar wind H, with diffusion acting only as a means of populating traps and not directly influencing the shape of the H distribution, and (4) a relatively unmodified (except by erosion) implantation distribution for a "suprathermal" solar proton component. ('~~5 - 100 keV) Clearly, combinations of these proposed mechanisms should also be considered. The possibility that a terrestrial contaminant, most likely H2 o, has penetrated beyond the region a fe1-r hundred angstroms deep, which might be affected by sputtering or other fractional vaporization effects, and into a region made porous or highly reactive to depths of '110. 4 \.liD by extreme radiation damage in these fe1-r samples cannot be ruled out; but it appears unlikely for a number of reasons. First, the identification as n o contamination of H concentrations 2 essentially confined to the 0-500 K deep region even on exterior -65surfaces of heavily pitted rocks such as 64455 indicates that prolonged exposure to solar radiation has ~made these surfaces porous to H2 0 contamination. Second, exposure of artificially damaged fused silica surfaces to n2 o, even in liquid form, provides no evidence for the penetration of H 0 to depths greater than ~0.1 ~m. 2 Third, protection against exposure to atmosphere appears to have prevented the chemisorption of H2 0 on the 68124 samples. The H distribution on the pitted 68124,3-B surface was essentially identical to that on the interior surface of 68815,27 indicating that only monolayer physisorption appears to have occurred, with this small surface conentration dissipating rapidly under illumination with the 1 9F beam. Since H2 o contamination has had such a small effect on these samples, it appears improbable that it could have had such a profound effect on the 68815,27 exterior surface as would be implied by interpreting the observed distribution as extensive penetration of terrestrial H0<- 0. Tal:ing this evidence at face value, the bulk of the H content 0 at depths rr,reater than ~500 A is best interpreted as true lunar II rather than terrestrial contamination. The concentration of this H near (,.,ithin ~0.5 ~m) the surfaces of the samples is consistent with a solar wind origin; hm.rever, the details of the depth distributions imply extensive modification of the implantation distribution. Processes which ma:r have significant effects include saturation, solid state diffusion, surface erosion, and radiation damage. Since the flux of solar wind protons, ~6 ~ 101 5 cm- 2 - yr-1 -66(Geiss et al., 1970; Bame et al., 1970), is sufficient to saturate an exposed surface '..rith H in a very short time (10 - 100 yr), the turnover of the top soil layers should insure that most soil fragments have been exposed long enough to reach saturation at some time in lunar history. Hm·rever, very few grains (other than ilmenite) are actually found vlith saturated surface layers (Eberhardt et al., 1970; Kirsten et al., 1971), implying that diffusive losses may be extensive even for concentration levels well below saturation. \ihile many lunar rocks have apparently been tumbled so that all faces have been exposed for times > 105 yr, a significant number exhibit no microimpact p its on their lunar bottom surfaces and hence have apparently been exposed in only one orientation (Morrison et al., 1972). Rocks 15015 and 1)059 appear to fit this latter category; however, the surface of rock 68815 from which sample 68815,27 was taken has an exposure a ge of ~2 x 106 y ears, and has therefore certainly had adequate exposure to reach a balance between the rate of implantation and the rate at -vrhich H is lost from the surface by diffusion and erosion. The apparent Penetration of rare gases to depths of up to a few microns in lunar soil grains has been attributed to solid state diffu sion of solar ui nd ions (Kirsten et al., 1970). Unfortunately, little is l:no;.m about diffusion of gases in lunar materials, except by inference from the kno~ properties of terrestrial materials. However, if a bulk diffusion process is assumed for the redistribution of implanted solar 'vind protons, an equilibrium distribution ma:r eventually be reached due to a balance between the rate at vrhich the -67diffusion front progresses into the material and the rate at vrhich the surface is beinp: ·e roded by solar wind sputtering. 'i'he characteristic depth of such a distribution would be given by D/v where D is the diffusion coefficient defined by Fick's law and v is the atomic erosion 0 rate. Using an erosion rate of 0.5 A/yr and the characteristic depth of 0.2 ~m for the measured distributions, an estimate of i s obtained for D. lo-1 3 cm2 /yr This is to be compared vrith the following Jr...nown diffusion coefficients for H in fused silica at temperatures of l00-l50°C (lunar daytime temperatures): ~lo-2 cm2 /yr for interstitial diffusion of H2 , "-l0-4 cm2/yr for diffusion of H+ to form OH-, and -vlo-9 cm2/yr for diffusion of OH- (Bruckner, 1971). Hence it appears -- that bulk diffusion should be much too fast to account for the observed distributions. However, the mobility of any diffusing snecies is influenced by local radiation damage, with the result that the diffusing particles can be "trapped" in radiation damage defects at low temperatures and released in the healing process at higher temperatures (~ 500° C) (Hatzke, 1966). A weak form of trapping , with a finite mean trapping time, might result in a diffusion coefficient of the right order of magnitude (lo-l3 cm2/yr at "-100° C); however, the shape of the H depth distribution would be determined b~r diffusionerosion equilibrium onl~r if the distribution of traps were reasonably uniform as a function of depth. 0 The observation of 200-1000 A thick metamict coatings on lunar dust grains (Borg et al., 1971) indicates that radiation damage is quite severe in this depth range. However, hi [!h densities of nuclear narticle tracks (> 1011 tracks/cm2 ) we re -68also observed in the underlying grain. In a bright-field micrograph of an Apollo 11 grain (Fig. 3 of Borg et al., 1971) the authors point out that "very fevr tracks with length greater than o. 5 ~m can be seen." This suegests that the radiation damage distribution may be quite steep over a 0.5-~m depth range, resulting in a distribution of traps which decreases rauidly with increasing depth over this range. Hence, the distribution of radiation damage traps may be a more important factor than diffusion in determining the depth distribution of H in lunar samples. A number of studies concerned with the trapping of rare gases in radiation damaged solids have led to a partial understanding of the interactions of the implanted rare gas atoms with radiation damage (Kelly and Brmm, 1965; !.{atzke, 1966; Hatzke and Whitton, 1966; Kelly et al., 1968; Ducati et at., 1972). At low implantation doses, gas release is governed by normal diffusion kinetics. However, if the doses are high enough (~1016 ions/cm2 ) three new processes appear. A small but significant fraction of the gas is released at temperatures between the bombar~ent temperature and the onset of release by normal volume diffusion. This is interpreted as diffusive motion which is short-circuited towards the surface by interconnected regions of radiation damage, and is given the name "damage di:ff'usion" (stage I). Normal volume diffusion (stage II A) is replaced by diffusion with •Teak trapping (stage II B) at high doses, characterized by a higher temperature release. Gas atoms are believed to be trapped in defects produced by the heavy radiation damage, possibly small vacancy clusters. -69Large activation energies are required for escape from a trap; however, trapped gas atoms are released when the temperature is sufficiently high to allovr healing (annealing) of the disordered regions and, hence, elimination of the traps. Thermal release profiles may be different for the various rare gases if escape from traps is important at lower temperatures, but release at annealing temperatures is common to all of the rare gases. A third process, bubble diffusion (stage III), is important at much higher temperatures and involves the motion of gas-filled bubbles. The rare gas contents of lunar samples are only a small residue of the amounts implanted during surface exposure. Ducatiet al. (1972) suggest that most of the solar wind gas atoms are lost from the heavily damaged metamict coatings at lunar daytime temperatures by the "damage diffusion" mechanism discussed above (stage I). The remaining gases ~ay uiffuse inward to populate traps associated with the high track densities observed beneath the metamict coatings by Borg et al. (1971), or may be retained in the metamict layer in those traps which are not connected ~vith the surface. Duplication of the lunar sample He release patterns for synthetic glass implanted with relatively low He+ doses (< 101 5 ions/cm2 ) but pre- or post-damaged with high proton doses (6 x 101 6 ions/cm2 ) represents convincing evidence for the proposed dependence on gas-damage interactions, since samples implanted only with a low dose of He+ show a much different release pattern . At low doses normal volume diffusion results in a low-temperature peak in the differential release profile, while at -70hit~h doses a second peak appears at a higher temperature (consistent with temperatures necessary for partial annealing) at the expense of the first peak. Observation of this same effect with relatively lmv He+ doses in pre- and post-damaged samples argues strongly against interpreting the higher temperature release as simply the effect of having high rare gas concentrations, independent of the radiation drunage. For hydrogen, the situation is complicated by the possibility of chemical effects. Epstein and Taylor (1970, 1971, 1972) have shm·m that l1 exists in lunar samples in forms which are thermally released as both H2o and H • 2 The high D/H ratio in the released H o (up to 133 ppm) as compared to the H2 ( as lmv aa 18 ppm) has 2 led them to identify the H o as primarily terrestrial (typically 2 ~150 ppm D) contamination while most of 'the H2 is undoubtedly related to the solar '"ind ( < 10 ppm II). The H2 /H 2o ratio ranges between 2 and 5; hmvever, the variation of D/H ratios · with temperature indicates that there is significant exchange of terrestrial and lunar H between the tvo chemical forms. Lord (1968) has shown that artificial implantation of 2-keV protons into terrestrial silicate minerals (olivine and enstatite) is follovred b:v thermal release of H2 gas. A retention coefficient of 0.9 for doses as high as 5 x 1016 ions/cm2 (~0.1 saturation dose) was obtained by comparing the ~uantity of H2 released to the total proton dose. 'l'he implications of this result for the lunar samples, that imT'lanted solar ,.rind H would probably be released as H2 rather than -71H2 0 (or some other product of chemical interaction with the medium), are consistent with the interpretation based on the different D/H ratios of the two thermally released chemical forms. However, infrared absorption lines characteristic of OH and OD groups have been observed in silicate targets following bombardment with MeV protons and deuterons, respectively (Zeller et al., 1968), suggesting that formation of hydroxyl groups by interaction of implanted solar wind protons and the lunar silicates may be an important process. The release of implanted H primarily in the form of H2 from lunar and terrestrial samples indicates that either only a small fraction of the retained H was in the form of hydroxyl, or that the 0-H bonds are preferentially broken in the thermal release experiments. Since most of the H2 is released from lunar samples at about the same temperatures as He (Gibson and J.ohnson, 1971) and in agreement with temperatures ( ""500° C) at 1o1hich partial healing of radiation damage is expected {i1atzke, 1966), one might expect that He and H gas 2 are trapped together in the same sites. Indeed, small gas bubbles 0 (-v80 A diameter) have been observed in breccia and soil grains and attributed to solar wind (Phakey et al., 1972). Gentle room- temperature crushing of a breccia sample (Funkhouser et al., 1971), however, resulted in appreciable (-vlo- 3 of total content) rare gas release but very little II 2 (2H 2 /He "' 0.01 compared to -v8 for thermally released gases) and no detectable H2o. A similar result was obtained by crushinp, a sample of fines (Epstein and Taylor, 1972) vrhich released normal amounts of H2 on heatinB. This suegests that H does -72not exist as molecular n2 gas populating the same voids, bubbles and grai n boundaries as He, but rather in an atomic or ionic state vrith chemical binding preventing release at room temperatures. For- nation of ~1yclroxyl group s is one possibility, but the diffusion of Oil in silicates appears to be too fast to retain OH near the surface v:i thout significant radiation daJnage trapping effects. It appears likely that both H and He are trapped in vacancies and vacancy clust ers, vri th the H forming a chemical bond with one of the surrounding atoms, while the trapping for He is purely a physical effect. Hence, room- temperature crushing would release He but not H, while the healing of defects at ~500° C would result in the release of both H and He from their substitutional sites, with the H atoms combining to form H2 • Since radiation damage appears to influence the mobility of implanted ions so strongly it is reasonable to suppose that the observed II depth distributions could reflect the concentration of traps as a function of depth. (~500 The thickness of the metamict coatings 0 A) has been correlated with the radiation damage range of ions of mean solar wind velocities. The formation of these metamict lay ers is attributed to interconnection of a high density of radiation damage islands, s.o most traps may be interconnected with each other and w·ith the surface, providing an explanation for the almost total loss of solar wind H from this region. One would expect, then, to find a peak concentration of isolated traps below this damagesaturated layer where the radiation damage due to solar wind particle s with higher-than-average velocities is still quite heavy, but not -73severe enough to interconnect the isolated traps. Calculations of radiation damage ranBes from theoretical considerations (Winterbon et al., 1970, Schi~tt, 1966) agree well with experimentally determined damage ranges (Hines and Arndt, 1960) for ions of solar wind velocities in quartz, and indicate that damage from solar wind He should extend deeper than damage due to either H or heavier solar wind ions of the same velocity . Using optical techniques to measure the thickness of a laye r in vrhic h saturated radiation damage has alt ered the refrac t ive index, Hines and Arndt (1960) found a radiation damage range of 0.19 ~m for 3.8 keV/amu He+ ions. Bulk solar wind energies as hi gh as 3.5 keV/amu have been detected in 3 hour averages of satellite data (vlolfe, 1972). !l content Hence the observation of a rapid decrease in Hith increasing depth at "'0.2 ~m deep in several of the s anroles a nal;rzed in this study is consistent with an interpretation in which the observed depth of the H-rich layer is related to the maxim~ ranBe over which solar wind is effective in producing radiation damage. A snall fraction of the implanted sola r wind H may di f fuse inward to populate these traps, rather than e s cap e t o th e s urface 1·Ti th the bulk of the solar wind H. The more gradual d ecrea ses with increasing depth between 0. 2 ~m and o.4 ~m deep may represent a slm·r escape of H from these traps with a subsequent imvard diffusion, or it may reflect a more g entle gradient in the concentration of traps. Since this region is beyond the damage range of solar wind ions and since solar flare ions are too penetr ating to res ult in a noticeable damage gradient in this depth region, such -74a gradient could only be due to ions of intermediate energies (~10-100 keV) such as the suprathermal protons observed by Frank ( 1970). The possibility of an appreciable flux of suprathermal protons suggests an alternative explanation for the observed H depth distributions, that of direct implantation of suprathermal protons resulting eventually in an equilibrium with surface erosion and involving only minor modifications due to diffusion. Such an equilibrium ;vould be reached in a time given by Rmaxlv where Rmax is the maximum range of the implanted ions and v is the erosion rate. For Rmax ~ 0.5 ~m and v ~ 0.5 A/yr the required time is ~104 yr. In Apnendix E, the procedure for calculating the depth distribution resultinG from a spectrum of incident energies is outlined. A convolution of the flux spectrum with a function representing the eouilibrium depth profile due to a monoenergetic portion of the snectrum must be performed. The form of this function for solar protons as 1.rell as for normal incidence and an isotropic angular distribution (neglecting the effects of range straggling) are shmm in Figure 25. Since the exposure age of 2 x 106 yr for 68815,27 is vell in excess of the time necessary to reach equilibrium, this sample was chosen as representative of an equilibrium distribution. Fieure 26 shows an example of the kind of long-term flux spectrum needed to give such a distribution, with suprathermal proton data from Fr ank (1970 ) p lotted for comparison. It should be noted, however, that these data represent the peak flux during one of several events -75h·imilar to ~olar flares ) ohserved by Frank and, consequently, :-:;hould not be considered as representative of a long-term average. Figure 27 shmvs the II distribution resulting from the hypothetical lon~-term spectrum compared to the initial set of data from 68815,27, as ·vell as the tail of a distribution chosen to overestimate the nenetration of present-day solar '.rind protons ( "-108 cm- 2-sec-1 behreen 1 and 2 keV and "-107 cm- 2 -sec-1 between 2 and 3 keV) calculated for normal incidence and including theoretical estimates of the stra~~line in ~rejected rang e (Schi~tt, 1966). The implied long - term suprathermal nroton fluxes are lm.rer than the fluxes in the events o1)served by Frank b~r almost an order of magnitude, indicating that this proposed ex:olanation of the observed H profiles deserves serious consideration. IIo•rever, the long-term average flux in this ranGe ma~r well be many orders of maGnitude lower than the peak fluxes indicated by the data in Figure 26 . Consideration of the various processes lvhich might affect the shA.pe of the meA.sured profiles has led to the follovring conclusions: (1) Essentially all .of the implanted solar Hind H has been lost 0 :from the ~500 -A thick amorphous la~rer in which it is stopped •. 0 (2) \lhile a portion of the observed H contents of the 0-500 A deep surface la;rer may be a remnant of implanted solar -vrind, most of the Ir observed in this region has been attributed to H2 o contamination, both nh:rsisorbed and cheaisorbed. It is doubtful that the extensive renetration necessary to explain the deeper features of the II profiles -76of some samples is possible, especially for the vacuum-sealed sample 68815,27. (3) The deeper features may reflect a steep gradient in the concentration of radiation damage traps,with the traps populated by diffusion of solar-wind (and/or suprathermal) protons. The high mobility of the H under bombardment with the 1 9F beam, especially in the 0-0.2 ~m region, suggests that this interpretation, involving high levels of radiation damage to depths up to about 0.2 ~m, is more probable than the direct implantation of suprathermal protons unaffected by radiation damage or diffusion. The extent of radiation 0 damage apparently decreases from saturation near "-500 A to a fairly lovr level at "-0. 2 ~m deep. The damage to these depths is probably due primarily to He ions from high velocity solar wind streams which can cause radiation damage to significantly greater depths than protons of the same velocity and are much more abundant than the heavier ions ( ivhich have comparable ranges). The thermal release of H2 , apparently of solar ivind origin, at "-500° C is consistent with the annealing of these traps. The mobility of the I-I in the 0-0.2 ~m deep region under illumination vith the 1 9F beam may not be a temperature effect but may actually be due to collisions between fast 1 9F ions (101 4 ions/cm2 for each data point) and trapped H atoms, knocking the H out of traps; or, perhaps, may be due to the interaction of secondary electrons produced by the 1 9F ions with the chemical bonds, resulting in the release of some of the previously bound H. Once released from isolated traps, the II may be able to escape to the surface through inter- -77connected regions of radiation damage. This hypothesis would explain why the distribution is stable under a bakeout to 300°C, but highly mobile under irradiation even though the surface temperature of the irradiated spot is probably somewhat lower than 300°C. (4) The portion of the H distribution extending from "'0.2 \.liD deep to depths greater than 0. 4 \.l m, characterized by a more gradual decrease with increasing depths, is best interpreted as either an inward diffusion of H that has escaped from traps in the highly damaged 0 - 0.2 \.liD deep region into a region with relatively uniform, low-level damage from solar flare ions; a population of traps b:r diffused solar wind and/or suprathermal II reflecting a gradient in the trap concentration caused by damage from suprathermal ions; or a direct implantation of suprathermal protons. The first inter- pretation does not depend on any long-term suprathermal ion flux, but does require a rather low diffusion coefficient. This possibility is not inconceivable, however, since the diffusion of OH may be strongly dependent on the chemistry of the medium, or weak trapping in solar-flare damage tracks may slow the diffusion process significantly. However, suprathermal ions have been detected and track gradients over the 0 - o. 5 \.lm ~epth range have been observed in lunar soil grains, so such a diffusion process would probabl~r be strongly affected by the resulting gradient in the concentration of traps. Hieher long-term fluxes ("'1013 protons/cm2 - vr) are needed in order to explain the observed distribution in terms of direct implantation of suprathermal protons (-vrith only minor diffusion -78effects) as compared with the fluxes necessary to create a sufficient density of traps; hm-Tever, even these fluxes are almost an order of magnitude lower than the events detected by Frank (1970). Thus, further investigations are necessary in order to distinguish between the possibilities implied by these results. Although the shape of the H profiles between a few hundred angstroms and "-0.4 ~m deep, with the alternative interpretations discussed above, is quite consistent; the absoulte H content in this region shows a rather large sample-to-sample variation (Table 3). The profile shapes are all consistent with the existence of a small uniform volume concentration (averaging about 40 ppm H by weight), underlying varying amounts of surface contamination and a "deep" (0.05 - 0.4 ~m) component, with a relatively well defined shape, 1-rhich has been attributed to implanted solar protons. Variations in the absolute amount of this component must either be due to differe nces in exposure time or differences in the implanted material. Although we cannot rule out a short (possibly zero) exposure age as an explanation for low II content in unpitted surfaces, the cratering rate for rocks (10 - 100 pits with spall diameters > 1 mm per cm2 per million years) implies that any surface with an appreciable number of visible pits ought to have reached an equilibrium between the incident proton flux and the rate of loss of H by diffusion and erosion. Thus, differences in the retention properties of the lunar materials must account for the bulk of the observed variation. Ilmenites have been found to retain solar wind rare -79gases much better than the silicate minerals. Glasses are more open structures than crystallized minerals, and should, therefore, be subject to larger diffusive losses. However, Ti-rich glass fragments were found to retain rare gases better than silicate minerals, although not nearly as well as ilmenite (Kirsten et al., 1972). Most of the glass samples analyzed in the present study are dark in color and could probably be classified as Ti-rich. However, the degree of devitrification may w·ell be important in determining the trapping efficiency of a particular sample. Hence even the extremely fine crystalline grains in a devitrified glass such as the 68815,27 matrix material may provide large numbers of radiation damage traps that the onen structure of the actual glass would not. If this interpretation is correct, the low "deep" (0.1 - 0.4 ~m) H contents of exposed splash glasses, such as the coating on 15059,32, may be due to rapid quenching resulting in an extremely low degree of crystallization. Such a uniformly open structure may be incompatible with the formation of traps by radiation damage. The observed widths of the H distributions (about 0.2 ~m) in samples such as 68815,27, the 10085 brown glass fragments, and 15413,5-2 are in good agreement with the chemical etching results of Eberhardt et al. (1970) on ilmenite grains. These authors showed that a 50 percent reduction in the content of each of the rare gas isotopes 4He, 20Ne, 36Ar, 86Kr and 132xe could be obtained by removing a surface layer 0.15 - 0.20 ~m thick by etching •rith HF. The observation that the depth of the distribution is essentially -80independent of the mass of the implanted atom is a stronc; arc;ument against any pro7osed redistribution mecha.nism vrllich depends on normal volume diffusion, since diffusion coefficients for rare Gases are quite sensitive to the atomic weight. A ~opulation of radiation damage traps by all of the rare gases seems favorable. An alternative origin for some of the "deep" (> 0. 1 ~m) hydrogen is that it is indigenous to the lunar sample rather than being incorporated as a result of the lunar surface exposure of the sample. This interpretation is particularly suggestive for the 64455 samples in which the "dee~" H concentrations, although small, are similar for the interior and exterior samples. On the other hand, for 68815,27 the large difference in H content at "'0.4 ~m deep between the interior and exterior samples strongly argues for a solar particle origin for most of the H at this depth in the exterior s~nple. The deep H in the Apollo 11 and 15 coarse fine samples appears to be more like that of 68815, although here we do not have interior and exterior samples for comparison. The possibility of H in interior samples unrelated to the recent surface exposure of the rock can be tested by bulk H analysis of interior samples. Rock 64455 is regarded as a crystalline rock (Grieve and Plant, 1973; Wilshire et al., 1973) although it is quite possible it has been formed by metamorphism of a breccia . Consequently if, prior to metamorphism, the rock contained a of surface irradiated com~onent material, then it is possible that a fraction of the H from this material has survived the heating and thus has been "inherited" b~r -81641155 in its present :form. (••acdour~all High voltage electron microscope studies et al ., 1973; Hutcheon et al., 1972) have shown that essentially all Apollo 14 breccias show evidence for such a preirradiated component. An intriguing alternative hypothesis is suggested by the observations of rusty regions ("goethite") around metal grains on a thin section of an interior sample of 64455 by Grieve and Pilant (1973). f) imilar observations have been made on other Apollo 16 rocb=;, notably ()6095 ('l'a~rlor, 1973; El Gorsey et al. , 1973). Grieve and Plant are;ue that, because the rust was seen in the interior portion of the rock, it is lunar in oriGin; however, it would appear difficult to rule out the possibility that the rust formed during the thin-section rmkinc :9rocess . Hevertheless , our observation of the roughlJ• uniform deep II concentrations in both interior and exterior samples of 64455 are consistent Hith the conclusions of Grieve and Plant and the hypothesis of indigenous lunar "goethite," but do not establish a lunar origin. The techniques employed in the present study are potentially useful to the investigation of this problem, particularly if an attempt is made to extend the measurements to greater depths on interior samples , because H can be detected in the interior portions of srunples which are shielded from atmospheric exposure with a sensitivity of "-20 ppm. -822. Depth distributions of F in lunar samples The results of ~ depth distribution measurements on Apollo 1 6 samples provide evidence for an enhanced F content within one micron of the surface, compared to bulk values as measured by Jovanovic and Reed (1973) on other Apollo 16 rocks (<50 ppm?) and soils (50- 100 ppm F). The 1~ content of one surface of the anorthosite fraoment 66044,8 was nearly 2000 ppm in the outer 0.5 ~m and ~1400 ppm in the underlying 0.5 um region. The F content of the opposite face '"as less than half these values indicating that F distributions are characterized by large local variations, F contents of 400 - 1000 ppm in the 0 - 0.5 um deep layer were found in three samples: glass chip 65315, 6 glass spherule 68124,3, and rock chip 68815,27. Some of these samples had received doses of 1 9F ions previously for the H analysis,hut the amounts ofF observed are much larger than the previous ion dose, In addition, the range of the 17-MeV l9F ions is about 8 urn (Northcliffe and Schilling, 1970) and it is extremely doubtful that any appreciable amount of the implanted 19F could migrate to within one micron of the surface. The details of the distributions were highly variable from samule to sample, with 68815,27 (Fig , 20) showing one of the most striking profiles, In light of the surface potentials measured on other samples using the 27Ai(p,y) 2 8Si reaction, it appears quite possible that charging of the sample by the proton beam could cause a high enough surface potential to account for the observed peak as a surface (zero-depth) F concentration, The 8-lcV surface potential -83necessary to account for the apparent depth of the peaks assuming it is actually at zero depths is not unreasonable. However, the excel- lent reproducibility of this distribution (data from two consecutive energy scans are plotted together in Figure 20). requiring an extremely stable surface potential,makes this interpretation appear somewhat unlikely. In spite of this reservations charging seems to be a slightly more probable explanation than any of the alternatives. For the 68124 samples, the apparent depth of the peak F content agrees reasonably well with the expected position of a surface concentration due to the measured surface potentials of 8 and 10 kV on the bvo 68124,3 surfaces and 14 kV on 68124,10-B. onl~r For 66044,8-B, a small ( 'V2 kV) surface potential was detected. but a narrow }leak ,,ras observed ap:narently much closer to the surface than the similar feature in 68815,27. Although the measured potentials do not correspond as 1vell to the apparent depths of the peak concentrations for the 65315 samples, the discrepancies are not so great that surface (zero-depth) concentrations can be ruled out. The unresolved question of contamination makes it premature to identif~r the observed F contents in these samples as true lunar F ivith absolute certainty although the large F contents observed in several samnles beneath a fairly narrow peak near the surface are nrobabl;r real. It is interesting to note that three of the four sample s richest in F were from the sealed rock box (ALSRC) and that the interior s urface of one of these, 68815.27 s shovred relatively large amounts of F , apparently due to contamination. Nevertheless, -84the F contents of interior rock surfaces are all less than 200 ppm, so a lunar ori gin for much of the observed F in the exterior surfaces does not appear unreasonable. Although the concentrations in the outer 0.5 ~m of all the samples analyzed in this study are somewhat lower than the Surveyor 7 fluorine levels, the value of 1900 ppm (0.19%) for 660411,8-B is within the uncertainties of the Surveyor measurements. Hence t he results of this study are not incompatible Hith the interpretation of the high F contents measured in the Surveyor 7 soils as a real surface-correlated lunar fluorine concentration. The contamination question will have to be answered more definitel~r before firmer conclusions can be drawn. -85- IV. OBSIDiru~ HYDRATION PROFILES Exposure of fresh surfaces of natural obsidian to ambient ,.,ater, either as ground moisture or atmospheric humidity, results in a slow diffusion process with the eventual formation of a hydration rind (Friedman and Smith, 1960). Microscopic examination of a thin- section cut perpendicular to such a surface reveals a well-defined band up to ~20 ~ in thickness. The thickness of this hydration rind can be measured under the microscope with a resolution of ~0.2 ~ and correlated with the age of a given surface if the hydration rate is known. The hydration rate is believed to be essentially independent of the relative humidity (above some minimma value of a fraction of one percent). The temperature and the chemical composition are observed to be much more important rate controlling variables. Using radiocarbon dating techniques as a calibration, a hydration rate is estimated for a given area, 1·T here a single effective temperature and obsidian chemistry can be assumed. The result is a relatively simple technique for archae- ological dating of obsidian artifacts, since fresh surfaces are exposed in the chipping process used by the artisan to fashion the desired implement. The 1n(l9F,ay) 16o technique for depth analysis o:e' hydrogen represents a unique probe for the direct measurement of obsidian hydration profiles! Results of such measurements on a variety of obsidian samples are presented and discussed in this section. -86A. Experimental Results A number of obsidian samples with estimated (from known exposure age) or optically measured hydration bands of < 2 11m ,.,ere collected, along 1-1i th several much older samples with thicker hydration layers. Samples were prepared by selecting a clean surface and cutting or chipping a ~l-cm2 sa.nple of exterior surface. The samples vere then degreascd vri th trichloroethylene and rinsed in methanol nrior to mounting in the scattering chamber. Ii:rdration rrofile measurements were performed by first collecting ra1-1 data in the form of Y-ray counts per 3 JJC of l9p4+ vs. E0 , the l9F bombarding energy, as described in Section II. The counting rate Y(E0 ) was related to water concentration with the assumption that all of the detected H could be identified as 11 vrater." The results of obsidian hydration profiles measured in this manner are given in Table 5. Samples l - 4 had sufficiently narrow hvdration layers that the entire hydration profile could be measured without encountering the resonance at ~ 2 = 17.64 NeV, as shown in Pigure 28 for samples 2 and 4. The hydration layer thicknesses listed in Table 5 for these samples were obtained by subtracting (in quadrature) the resonance width r 1 from the measured energy-width of the excitation function and converting the result to a depth using the calculated stopping :pmver, dE/d(px) (Appendix D), and the density = 2.4 g/cm2 • 'l'he H2 o content (in weight uercent) of the thin h~rdrated la~rer i s calculated f rom the peak counting rate •rith a P correction for the finite vTidth r 1 of the resonance, i.e. -87y max vrhere 1-1 is the E 2 o content, Ymax is the peak :rield, and h is tl1e hydration la:rer thickness. (This expression follm·rs directlv frorr. the equations of A::mendix A for the case o~ a distribution confined to a thin la:rer.) These distributions '\.rere not nearl~r as labile under irradiation vrith the l9p beam as the H distributions in lunar sarnnles. Excellent reproducibility ~as found for repeated measure- ments. G8J11.Dles 1-3 were hydrat ed artificially by heating :freshly exnosed s urface s o:f a single obsidian sampl e for neriods of 1, 2 and 4 davs , resnecti.velv, at 75°C in a humid atmosphere ( Friedman, 1973). A progressive increase in the thickness o:f the hydration laver with time is observed , in qualitative agreement 1.ritll the expected trend o:f the hydration la:-rer thickness p roportional to the square root the h;,rdration time (Friedman and Smith, 1960). o:f Sample 4 is a surface 1-rith an exposure age of '\.10 years at normal temperatures. Figure 29 illustrates the un:foldinb procedure described in Section II for a sample 'Hi th a h;,rdration la,ver '\.l. 2 (sample 5 ). ~m thick ;)at a :from an unhydrated obsidian saJnple (!I 2 v content uniformly 0 . 3;,) were subtracted uoint b~y p o int from the raw data for the hydrated sampl e to minimize the e ffects of beam- dependent bac kground above '\,20 '1eV 1 9F ener gy . The profile Has then unf olued as described in Section II to l eav e only the contribution o:f the resonance at ~l = 16 .4 5 '1eV . These reduced data were then fi tt ed vrith -88t\l~ calculated denth and water content scales, adding back the 0. 3?~ H2 0 1-li1ich 1·ras subtracted in a previous step. The result is the hy- dration profile shovm in F i 5 ure 29b. The same procedure was used to obtain hydration profiles for samples 6 - 10, with two of these ~rofiles shown in F i g ure 30. Samples 5 - 10 have all hydrated naturall y and their hy dration bands have all been measured by the optical thin-section techni q ue described previously. In gener a l, the hydration-band thicY..nesses measur ed in this wa~r show good ar-;reement vrith the depth at which the n o concentration gradient is 2 a ~~ximum. Samples 6 - 9 are artifacts collected from the same archaeolog ical stratum and. thus are probably of the same approximate age and have had the same temperature history . Consequently, differ- en c es in the hydration profiles may be controlled primarily by sample to sample di ff e rences in chemical composition. In addition to the samples listed in Table 5, several samples •.;ith much greater exposure ages were also obtained. These samples have h~rdration rinds much thicker than the 2-JJm limit of the present measurement technique, but. the H o content of the out e r 2-JJm layer 2 should be representative of the final saturation H o content for a 2 ~ iven sample . Table 6 lists the measured H2 0 content in thes e samples and in freshly exposed surfaces from the same obsidian sample. Using the resonance at a proton energy of 1.318 MeV in the reaction 23u a(p,y) 24 Hg , we have also measured the sodium depth distribution s to a dep th of ~1 JJm in two samples with hydration layer s less than 0.2 J.lm thick (samples 1 and 4 of Table 5 ) and in ancient and -89fresh surfaces from the same obsidian source. This was done to check a hypothesized ionic-exchange diffusion mechanism predicting a sodium depletion in the hydrated layer (Bikerman, 1970). No significant variation of Ha content (typically -v3% Na2 o) vras detected; however, the resolution was limited by counting statistics to a detection threshold of about a ten percent variation in the Na content. B. Discussion The detailed hydration profiles measured by this technique can be used to obtain information about the mechanism of water diffusion into obsidian and the factors which influence hydration. The general shape of the measured profiles agrees qualitatively with the idealized profile suggested by Friedman et al. (1966), characterized by a saturated hydration plateau followed by a steep diffusion front, rather than the more conventional exponential profile suggested by ~1arshall ( 1961). However, the presence of a second step in some of the hydration profiles (Fig. 30) suggests that more than one mechanism of vrater diffusion and binding may be important in the hydration process. The good general agreement between the optically measured hydration band thickness and the depth at 1-1hich the H2o concentration gradient is a maximum verifies the interpretation of T"riedman et al. ( 1966); 1-Ti th the border betVTeen the hydrated and unhydrated reg ions made visible in ordinary light due to a differenc e i_n index of refraction bet>·reen the hydrated and unhydrated region. The brir;ht ap1)earance of this border under crossed polarized li8ht -90ls due to stress birefringence at the point of maximum stress (due to the change in volume associated with hydration), and it is not surprisinB that this point should be correlated with the H2o concentration gradient. It should be noted, in this regard, that H depth distributions have also been measured for a number of tektite samples - , ( Hhich do not have visible hydration bands); and these distributions ' vrere characterized by a gently-sloping exponential diffusion profile, rather than the steep diffusion fronts observed in hydrated obsidians. Evidence for the effects of chemical composition can be found by inspection of the data in Table 6 and for samples 6 - 9 in Table 5. A comparison of the n2o content of hydrated and unhydrated samples from the same source shows a consistent correlation of the final saturation level with the intrinsic H o content. 2 It is also apparent from the data for a set of obsidian samples from the same source, and presumably with the same exposure history, that the saturation H2o content and the thickness of the hydration layer are ,.,eakly correlated (Table 5, samples 6 - 9). This suggests that the chemical factors •·rhich control the saturation level also control the rate of growth of the hydration rind. Finally, a progressive increase in the thickness of the artificially hydrated layer with the exposure time at 75°C is observed (Table 5, samples 1 - 3). The ability of this technique to resolve and accurately neasure such thin hydration layers as can be prepared under controlled laboratory conditions on a reasonable time scale makes this type of experiment particularly promising. 'rhe investigation -91of the effects of chemical composition and temperature on the hydration rate can be carried out under controlled conditions. The contribution of such an investigation toward understanding the hydration process could have great potential value in terms of the establishment of a precise hydration rate for a particular set of conditions. This rate could then be used to obtain a more accurate date for obsidian artifacts than is currently possible. -92- V. A. CONCLUUION Lunar Sample and Obsidian Analyses Despite the implicit ambiguity in the interpretation of the orirrin of the H observed in lunar samples, the 1 H( 1 9F,a.y) 1 6o depth analysis technique has been shown to provide a reasonably accurate and reproducible measurement of its distribution, limited only b~,r the mobilit~r of the hydrogen under irradiation with the 19p beam. Simply decreasing the beam current density will help to control this problem, as has been demonstrated for one of the larger glass-coated rock chips (15015,39-2). For smaller samples, significant decrease in beam current density could only be achieved at a sacrifice in countinG rate, leading to large statistical uncertainties, unless the ITai (TQ.) detector can be placed closer to the irradiated sample. This is not possible vrith the present scattering chamber, but has l)een successful l~r done vi th a new UHV scatter inc; cha.I'lber of an im!.lroved desit:;n . Nevertheless, the results of this study have pro- vided an insight into the interaction of solar corpuscular radiation ••ith lunar materials \vhich could not be obtained by more conventional technioues . The potential usefulness of the ln( 1 9F,ay) 16 o depth analysis techninu e for the study of solid state diffusion processes involving H has been demonstrated by the obsidian hydration profile measurements ~erformed in this study. The reproducible profiles obtained on a variety of obsidian samples have verified the correspondenc e -93hetween the location of the stress birefrineence line visible under the nicroscore, and the depth at which the H concentration gradient is a maximum, for hydration rinds up to 2 l-JID thick. This technique appears to be narticularl:r well suited to the measurement of very thin ( < 0.5 pm) hydration layers where the order-of-magnitude inprovement in resolution over optical techniques can be used to best advantace . An investigation of the dependence of the hydration rate on chemical com~osition using obsidian samples hydrated under controlled conditions appears to be a particularly promising anplication of this technique. The results of such a study would be valuable in establishing the reliability of the age assigned to an obsidian artifact on the basis of the thickness of its hydration rind. The measurement of fluorine depth distributions using the 1 9P(n,ay)l6o reaction has already been demonstrated 0tarfelt , 1 9G7). ( ~oller and The application of this technique to lunar samples has l1een somewhat inconclusive due to the suggestion of significant contamination and also due to the problems caused by charging of the :::;j_licate samples with the ion beam. The zero-point of the depth scale could be redefined based on the 27At(p,y) 28si measurements, if these measurenents were ~erformed with sufficient care to improve the uncertainties in the surface-potential determinations. Even thrm, hmvever, the large potential fluctuations apparent on several samples ivould cause a significant degradation in the depth resolution. 'l'he charp;in(3; nroblem could be solved quite easily by depositing a thin conducting film onto the sample surfaces to be analyzed; however , it is not clear that elimination of the uncertainties caused by sample charging would justify this destructive procedure. i..Teve rtheless, if further investigations are able to demonstrate that contamination can account for only a small portion of the large F contents observed in a 1-~m thick surface layer on several of the lunar samples, the results of this study will have an unambiguous inter~r etat ion as true lunar F , providing ~ossible evidence for recent exhalations of volcanic gases from the moon. B. 1\.d.ditional Applications of Nuclear Techniques for Depth-Sensitive AnalYsis The 1 9F(p,ay) 16 o technique is also being used to determine the 1" content of chondritic meteorite samples. The carbonaceous chondrites are of special interest because their compositions are presumed to reflect cosmic elemental abundances. Samples with fresh surfaces were obtained by chipping from three carbonaceous chondrites (Allende, Eralcot and ~-!urchison). Preliminary results on three samples of Allende matrix material and two samples from tlurchison indicate that a range of "'40 ppm to "'70 ppm F may be applicable to both of these meteorites. The F determination for one Erakot sample also fell in this range. si r~n ificantly These measurements are lo,·rcr than most previous F determinations in carbo- naceous chondrites, except for the 66 ppm determined by Reed (1964) in the carbonaceous chondrite Lance. A more extensive data set -95must be obtained from a particular meteorite in order to establish a value representative of the meteorite as a whole; however 9 the sensitivity of this technique appears to be at least competitive ,.rith the n-activation 9 y-activation and emission spectrography techniques used previously (Reed 9 1971). Another application of resonant nuclear reaction techniques tal<:es advantage of their particular sensitivity to surface contamination, using a variety of reactions to identify contaminants on surfaces for ,.,hich cleanliness is critical. A particular example is a series of measurements performed on electroplated lead samples in an effort to identif~r the source of residual RF losses in superconducting lead cavities being developed for a heavy ion accelerator. In a previous study (Tombrello and Leich, 1971) 9 the alpha particle yield from the 0.2-t1eV wide resonance in the reaction 16o( 3He,a) 1 5o at 2.3G MeV( 3 He energy) was measured to determine the surface concentration of l6o to be ~3 x 1016 cm- 2 Subsequent measurements have used the resonance in the reaction 12c(d 9 py) 13c at a deuteron encrey of 1.1146 !!leV (Aj zenberg-Selove 9 1970) to determine carbon concentrations and have also included F and H determinations using the 10 16 . - F(p,ay) 0 react~on. The results for untreated lead samples 2 show 5 x 1016 C/cm 2 , 2 x 1016 H/cm and 5 x 101 5 F/cm2 • In each 0 case the thickness of the contaminated layer was shown to be < 300 A. Samples treated in a chelating agent (trade name Versene) show decreases in the C and H concentrations by factors of 2 and 5 respectivelv, but F is i ncreased by a factor of 2. A vacuum bakeout -96n t 2 ') fPC:, however, decrenRed the C, II, and F concentrations hy factors of 5, 10, and 100, respectively, suggestine that if dielectric losses are a limiting factor in the performance of the superconducting r esonators, a vacuum-baked resonator should show a marked inprovement in Q value compared to cavities that have not been baked. However, the thickness of dielectric implied by these measurements is much too s~all, even for the untreated samples, to account for the observed losses (Dick, 1973). Additional possibilities for the application of these techniques are numerous. As an extention of our investigation of deuth distributions of implanted solar ions in lunar materials, the reaction 4 IIe( 10 B,n) 13u looks r:Jarticularly promising as a depth 0 analysis probe for I·i e 1-rith a depth resolution of 'V300 A. The H/He ratio of ~ for lunar soils implies that neutron counting rates from the resonance at 3.78 ~·IeV( 10B energy) (Ajzenberg-Selove, 1970) 1-rill probably be an order of magnitude lower than the y-ray counting rates from 1 u( 19F,ay) 16 ous limitation the He distributions are mobilized b~r the ion if o on lunar samples. beam to the same extent as H. This may represent a seri- The best possibility for studying implanted solar 1.rind 12c involves the 12c( 3He,n) 14 o reaction and the detection of the delayed y rays following the ~+ decay of the residual l4o nucleus. (The 12c(d,py) 1 3c reaction discussed earlier is not useful due to copious background from silicate targets.) The ln( 1 5N,ay) 12c reaction represents an alternative way to meaRure U distributions. The estimated deptl1 resolution using -97the resonance at 6.39 HeV ( 1 51'1 energy) (Ajzenberg-Selove, 1971) 0 is better than 100 A near the surface, and the maximum depth which can be studied •ri thout interference from other resonances in the ln(l~N,ay) 12c reaction is about 3 ~m. improvements in Hence, significant resolution and depth range could be obtained at an order-of-magnitude sacrifice in counting rate compared 16 1 to the present technique using 1 H( 9F,ay) o. The importance of resonant nuclear reaction depth analysis techniques for the investigation of solid-state diffusion problems has been demonstrated by our study of hydration profiles in obsidian. Clearly, a more general class of solid-state diffusion, weathering and corrosion problems involving chemical surface reactions can be studied using similar techniques. -98- APPENDIX A. Depth-Sensitive Analysis with Charged Particle Induced Nuclear Reactions Consider the general case of a nuclear reaction A(a,b)B involving compound nucleus formation with an entrance channel consisting of a charged particle projectile ~and a target nucleus ~, and an exit channel characterized by an emitted particle or .quantum £ and a residual nucleus !• Let the reaction cross section be given by o(E) where E is the kinetic energy of the projectile ~· Suppose that nuclei ~ are distributed near the surface of a solid medium with number density A(x) for a given depth x in the sample, If the sample surface is irradiated with a beam of particles ~ of energy E0 , the gradual slowing down of the a ions due to electronic collisions, characterized by an energy dependent stopping power [dE/dx] (a negative quantity) results in an ion energy at depth x' given hy (A,l) (Normal incidence is assumed and energy straggling is neglected for the moment.) '~ithout loss of generality we may write the reaction yield per incident particle ~ of energy E0 as -99- (A. 2 ) Hence, the excitation function Y(E 0 ) is, in a sense, a convolution of the reaction cross section with the distribution function A(x). Although it is possible, in principle, to unfold (or deconvolute) the runction A(x) from a measurement of the excitation function Y(E 0 ) for a more general form of o(E), the process is greatly simplified if the cross section is dominated by a single resonance at ener~r ~· Then, following the treatment of Fowler, Lauritsen, anu Lauritsen (1948), the cross section is given by the Breit- Hi gner dispersion equation, (A.3) o(E) vhere oR = o(ER) is the cross section at the resonance energy and r(<<~) is the full-vidth of the resonance at half-maximum intensity (Ji'HID!). (The quantity oR includes wavelength and barrier penetration factors 1-1hich need not be considered explicitly.) Since this form of the cross section consists of a sharp spike at ER• the effect of equation (A.2) for E 0 > ER is to pick out the value of A(~) -100- 00 A(xR)J dx' (E(x') 0 ~rov ided t hat A(x) varies slowly over distances of order f j[-dE/dx] near ~, i.e. , 11. dA A dx r [ dE/dx] I « 1. (A. 5) f.Juhstitution of equation (A.l) in equation (A.ld and a chang e of variables leads to [1 iT2 -1 + t an (A. 6) The subscript R on t he stopping pow·er factor means that it is to be evaluated at ER . E0 - F or beam energ ies sufficiently greater than ~ , i.e. , BR >> f/2, equation (A.6) reduces to (A. 7) Hence the :function A (xR ) is obtained directly from the excitation function Y(E0 ) using equation (A.7) and the relation dE [dE/dx] -101- In nractice, the stopping povrer may vary only slightly over the ra.n~e from the resonance energy up to the maximum beam energy E0 , so that a useful approximation is (A.8) The inequality A. 5 is an expression of the finite depth resolution (due to the resonance 1.Jidth r) associated with the determination of the distribution A(x) from a measurement of the excitation function Y(E0 ). Setting A(x) =constant obtain an estimate of the resolution ox "' r 2[-d.E/dx] x IS(x-x') we from the half'-width at half-raaximum of the distribution inferred from measurement of the excitation function as described above. The effects of ener~r stra~gling and the finite energy width of the ion beam (both of which have been neglected to this point) on the resolution can be expressed by: ox = •:-here F\-JH~.f B. oE0 2[-dE/dx] , (A. 9) is the r.vnN spread in the beam energy and n(xR) is the energy straggling at depth xH. :3pecial Procedures for Cl ean UHV Systems Special techniques necessary to raaintain the cleanliness required in UllV systems include procedures applying to the assembly -102- ami oneration of such a system. Selection of materials is one of the first considerations, since low vapor pressure materials must be used exclusivel y. For this reason, the UIN scattering chamber described in Section I I of this thesis is constructed primaril y of type 304 stainless steel vith metal-to-metal seals using OFHC copper gaskets . The target wheel, however, was machined from a corrunon aluminum alloy, and the electrical connections make us.e of glass insulators, a sliding beryllium-copper alloy contact, a copper wire vrith ceramic insulation, and a ceramic-sealed kovar electrical feedthrough. All permanent joints are heliarc welded. All lJB.rts must be chemically cleaned before assenbly using nrocedures described in various sources on vacuum technology (see, for example, Espe , 1966). These procedures include a degreasing step using an organic solvent such as trichloroethylene to dissolve oils and grease, one or more acid baths to remove surface layers ,.,hich may be rich in adsorbed gases, a thorough deionized water rinse to remove all traces of acid, and an ultrasonic rinse in methanol or acetone. This sequence can be followed by drying with a hot air blower, after which parts are placed i n clean polyethylene bags until all parts are prepared for assembly. Clean polyeth~rlene gloves should be 'vern for the assembly as any fingerprints can act as essentially infinite gas sources at ultrahigh vacuum. After assembly, the entire system is baked at ~300°C to drive adsorbed gases from the vacuum system walls. to min imize this major source of gas. This procedure is important -103The most important operational procedures are concerned with op e nin~ the system to atmospheric pressure and the subsequent pumpdown. The specific procedures are tailored to the particular system, but a common objective is to minimize both the pumpdown time and the base pressure. This is most effectively done by backfilling the system with dry nitrogen gas and minimizing exposure to room atmosphere. the present system the procedure is as follows: For A liquid nitrogen trapped gas line is connected to the roughing line, evacuated, and then filled i-Tith dry nitrogen to a pressure slightly above one atmosphere. The metal sealed roughing valve is then opened allowing dry n2 gas to fill the system. The viewing window is then removed from the scattering chamber and the mouth of a polyethylene glove hug is clamped to the resulting access port. The glove bag, containing all of the necessary targets and tools, is allowed to inflate and flush with dry N before new targets are exchanged with the old ones. 2 lVhen t his operation is completed, the glove bag is removed, the viewing port is immediately replaced and sealed, and the N gas line is disconnected. 2 The system is then pumped down to a pressure of ~lo- 2 Torr using the molecular sieve sorption !='ump refrigerated vrith liquid nitrogen. ion pump is started at this pressure. 'l'he The titanium sublimation pump is typically started at ~lo-5 Torr i.rith a duty cycle decreasing from 1 at pressures s.bove 10-6 Torr to less than 0.01 below 10-8 Torr. normally three or four days are needed to attain a vacuum of "-lo-9 ",'orr. Periodic bakeouts are helpful not only to maintain a clean -104vacuum system but also to remove surface contamination from target samples. Baking to 300°C has been found to eliminate surface con- taminants far more effectively than any other procedure. C. Analyzed Standards and Detection Efficiency The following samples '\-Tith their analyses by weight percent were used as standards to establish the conversion from counting rate to absolute concentration: 1) Belvidere Mountain chlorite (Albee, 1971): HgO (33.88), Si0 2 (32.18), A1 2 o (16.07), H o (14.64), FeO (1.38), plus minor 3 2 constituents (< 1% each) including Fe 203, CaO and co 2 • 2) Durango apatite (Young et al., 1969): CaO (54.02), P 2o5 (40.78), F (3.53), RE 2o (1.43), plus minor constituents(< 1%) 3 including Ct, so 3 , Si0 2 , and Na2o. 3) CaF 2 : 300 ~g/cm2 CaF 2 deposited by vacuum evaporation on a tantalum substrate. The peak counting rate from a 300 ~g/cm2 CaF 2 target due to the 872 keV resonance in 1 9F(p,ay) 1 6o is calculated to be ~5% less than the counting rate from an infinitely thick target. Cor- recting for this thickness effect, the measured counting rate can be compared directly with the total thick target yield from CaF 2 determined by Chao et al. (1950) to be 3.7 x 10-7 per incident proton. The result implies a detection efficiency n = 0.0230. Using the relative stopping powers of CaF 2 and apatite, calculated as in Appendix D, the countine rate measured from the apatite -105- n of 0.0215. Using the chlorite n is obtained. He take the mean of standard results in a value for standard, a similar value for these results to be the best estimate of the detection efficiency, i.e. D. 11 = 0.022. Calculation of Stop-ping Pow·ers of Complex Substances The stopping power of a compound YnZm for a given energetic ion can be calculated from the atomic stopping powers of the elements Y and Z using Bragg's rule of additivity of atomic stopping cross sections: (D .1) II vhere £ = N" 1 di' dE voll.U!le. Since the stopping power data of :Northcliffe (1963) and and N is the number of atoms or molecules per unit Northcliffe and Schilling (1970) are in the form dE/d(px), it is convenient for purposes of calculation to rewrite equation D.l as nAy[~] :t mAz [~]. (D. 2 ) n Ay + m Az uhere Ay and Az are the atomic weights of Y and Z respectively. Generalizing this expression to a sample with a complex composition, we obtain -106- dE [ d(px) ~.rhere f. 1 the J samp1 e == fl. [ d(dE px) ~ i Ji (D. 3) is the Height fraction of the ith element, [dE/d(px)]. is 1 stornin~ newer for a pure substance of element i, and the stoppinr,: no,·rer for the particular sample composition summin~ is obtained by the product of these two auantities over all chemical constituents of the sample. A GOOd estimate to dE/d(px) for a sample of complex campocition can be made simply by calculating the average atomic weight and atomic number for the sample composition and finding dE/d(px) by interpolation using these average values (Schi¢tt, 1970). For tvri.cal silicates, stopring !lOwers obtained in this >.ray differ by lees than 5~ from the calculations using eauation D.3. Using equation D.3 and stopping powers taken from Northcliffe (1963) and rforthcliffe and Schilling (1970) the stopping powers of a variety of lunar sample compositions for 1 6.5 MeV l9F ions vrer e found to vary from -8.7 keV-cm -~g-l to -9.3 keV-cm2 ~g- 1 • 2 The mean of -9.0 keV-cm 2 -~~ -l >ras taken as a reasonable esti.m.ate for all of the lunar samples analyzed in this study. ,,m.s found for unh~rdrated A value of -9.4 keV-cm 2 -~g-l terrestrial obsidian, wit h the additional H2 0 content in the hydrated re g ion resulting in a stopping p m.rer of -9 . o~ keV-cm2 -~r; -l • • _.A,_ stoppJ.ng pov1er o f -0. 21 keV -em 2- ~p; -l 1.ras calculated for 0 . 372 ~1eV protons in lunar materials. A relative uncertainty of about 5 percent should be -107- associated vith the calculated stopping powers due primarily to uncertainties in the stopping po-vrer data and interpolations used to obtain the [dE/d(px)]. in equation D,3, Although 10 percent 1 devj ations from Bra~s 's rule have been reported for Si0 2 (Thompson and Hackintosh, 1 971) , other studies indicate that deviations are not larger than -v2 percent for Si0 1 973) . b~en 2 or At o (Nicolet and Feng, 2 3 'l'hus the !'Ossibility of deviations from Bragg's rule have disregarded for the purposes of this stud~r. ". ImDlantation of Interplanetary Ions in Lunar Samples ]; · Since most implantation experiments in the laboratory are performed 1-!ith a monoenergetic berun at normal incidence and a limited exposure , the resulting depth d istributions of the implanted ions are usually considered to represent a distribution in projected range . This is not the case for lunar samples, where even neglecting modifications due to diffusion, the distributions of implanted ions are governed h;r the combined effects of a n extended spectrum of ir.1nlantation energies, the angular distribution of the incident ions, and the gradual erosion of sample surfaces, In the following, we ¥i l l consider t hese effects, assuming that diffusion plays a negligible role so that all implanted ions are frozen into the solid at the e nd of their range. First, consider the case of a monoenergetic beam of protons normally incident on a lunar sample with energy E 0 and a corresponding nrojected range Ra · Neglecting range straggling and the effects of -108- eroRion, the resultant H distribution would be given by ¢0 To(x - R0 ) 1.,rhere ¢ 0 is the incident flux, the depth in the sample. T (E.l) is the irradiation time, and x is The effects of erosion can be accounted for by replacing x by x - v(t - t 0 ) where v is the erosion rate (assumed constant) and t is the time extending from the beginning of the irradiation at t = 0 to its termination at t = t 0 • Inte- grating over t gives hence, = ~U(xR0 + vt 0 ) - U(x- n0 )], v vrhere U(x) is the · Heaviside step function. (E.2) Assuming that the irradiation time is long enough to erode a layer of thickness greater than R0 (i.e., t 0 > R0 /v), an equilibrium (independent of t 0 ) dis- tribution 1.,rill result, given by (E.3) 'rhus, the equilibrium distribution for a monoenergetic, normally incident proton beam is a constant H content (= ¢0 /v) extending from the surface to the projected range R • 0 straggl in~ The effect of range on this distribution is to smooth out the drop from the constant (¢ 0 /v) content for x < R0 to zero for x > R0 by -109- I'Cl>laclng ll(x - H0 ) in equation E.3 by an error function: where o0 is the range straggling parameter. Normal incidence for a sample from one of the Apollo sites is a good approximation for ions reflected toward the moon from the Earth's bm1shock. For ions incident from a solar direction, however, the rotation of the moon results in the equivalent of isotropic incidence in a half-plane. Once again neglecting range straggling and erosion , the resultant H distribution 1-10Uld be 11 n8 (x) = :~1 2 n0 cos8)d6 , 6(x _.:!L, 2 vhere 8 is the angle of incidence measured from the normal. Inte- r:;ratinc; gives <Po ' '"> [U(x) - U(x- R0 ], (E.4) - x"" and including erosion in the same manner as above results in lL, (x) ::.> = ~0 2v (E. 5) -110- For completeness, we also consider the opposite extreme from normal incidence, that of an isotropic angular distribution in three dimensions. 'de take cp steradians at energy E0 • to be the proton flux per 41f 0 n eglecting straggling and erosion, R 0 w~i ch cos0) sinS d0 results in [U(x) - U(x- n )]. (E.6) 0 1·lith erosion, this leads finally to the equilibrium distribution cp H (x) = I 0 2v (1 - 2S...) [U(x) - U(x - R0 )] R0 • (E.7) The equilibrium distributions resulting from these conditions (neglecting range straggling) are plotted in Figure 25 for comparison. It should be noted that the shape of the equilibrium distribution for monoenergetic protons incident from a solar direction (II (x)) is intermediate between the corresponding distri8 butions for normal (~ ~(x)) and isotropic (HI(x)) incidence. To treat an extended spectrum of energies, one has only to select the appropriate source function and replace R0 by a rangeenergy rela tion TI (E) and $ 0 by (d~.dE) and integrate over the -111spectrum of incident energies. Hence the H distribution II(x) would be siven by J 00 H(x) = P(x,E) d4l dE (E.8) dE 0 vrhere 1/v P(x,E) = [U(x) - U(x- R(E))] [ 1 - ; sin-1 (nCE))]/2v (E.9) [1 - RCE) ]/2v for normal, solar anele, and isotropic incidence respectively. The effects of straggling complicate the mathematics considerably, but if a gaussian range distribution defined by the straggling parameter cr(E) is assumed, analytical expressions can be derived for P(x,E) for both normal and isotropic incidence. For normal incidence, P( x,E) = Uw [ l + erf (R(E) - X )\l {2 cr(E) ~ (E.lO) For isotropic incidence the expression is considerably more complicated and is not g iven here. -112F. Lunar Sample Inventory Following is a list of all lunar samples allocated for this study: 10085,1: One brown glass fragment with 2-mm x 1.5-mm analyzed surface characterized by rough texture and small pits. 10085,31: Two brown glass fragments. 10085,31-9 appears to be a ~- em glass spherule. chip from a 2 The analyzed surface is a 2.5-mm diameter convex surface of shiny glass. 10085,31-12 was an angular fragment ~3-mm x 5-mm which broke into two pieces subsequent to the analysis of a 3.2-mm, shiny, slightly concave surface. 15083,2: One 1.5-mm x 1.5-mm white crystalline grain (probably plagioclase) which broke into two pieces on mounting and was not analyzed. 15413,5: Two samples. 15413,5-2 is a 2-mm x 4-mm partially glazed pyroxene-rich crystalline rock fragment. The analyzed surface contained an area ~1-mm x 3-mm rich in nyroxene. 15413,5-5 is a highly fractured 1.5~ x 2.5-mm fragment which appeared to be essentially all plagioclase. 15533 ,4: One 3-rnm x 2.5-mm glass coated breccia fragment. -11315015,39: T,.,o surface glass chips from the unpitted lunar bottom of rock 15015, a glass coated breccia. The glass surfaces are 5 ram x 4 rnm for 15015,39-1 and 7 mm x 7 mm for 15015,39-2. The analyzed interior surfaces are dark matrix breccia. 15059,28: One surface glass chip from the unpitted lunar bottom of rock 15059, a glass coated breccia similar to 15015. A 5-mnl x 6-mm glass coated area was obtained for analysis by breaking the original 15059,28 sample into two roughly equal pieces. 15059,32: One surface glass chip from the lightly pitted lunar top of rock 15059. The 4-mm x 4-mm glass coated area was analyzed. 64455,24: One surface glass chip from rock 64455. is 1-cm x 1-cm smooth black glass. Exterior surface Interior surface is light colored rock. 64455,33: Two light colored anorthositic rock surface chips, 0.38 grams total weight. 65315,6: Three black surface glass chips from rock 65315, only one analyzed. Exterior 12-mm x 7-mm surface contained two possible impact pits visible to the naked e~e. Interior surface is mostly white anorthosite with some -114patche:1 of black ~lass. The 1-nun thick glass cracked in its mount , but the pieces stayed together for exterior surface analysis. Interior surfaces of two of the three pieces were analyzed separately. 65315,7: One large (13-mm x 16-mm) rounded surface anorthosite chip not mounted for analysis because of large size. 65315,13: One large (10-nun x 10-mm x 7-mm), blocky surface anorthosite chip. The surface anal~rzed is a clear vthite interior surface. The sample was too large to mount for anal~sis G5315,20: of the exterior surface. Two black surface glass chips, only one (~ x 4 mm) analyzed, showing some white anorthosite on both interior anu exterior surfaces • . One 3-~m x 5-mm blocky light-colored anorthosite fragment returned in sealed rock box. arc~n. Surface A contains one gray '·rhici1 rna~r be glass, while :3urface B is essentially all anorthosite. One 5-rnrn dark brown glass spherule in sealed rock box. (~ 0 .2 grrun ) returned Surface B has a few small pits, but Surface A is smooth and shiny. -ll5@ l24 ,lO: One 7-mm x 5-nua white anorthosite fragment with translucent areas >vhich may be single feldspar crystals. This samnle ~as also returned in the sealed rock box. 688l5, 27: One l-cm. x 7-cm surface breccia chip from rock 688l5 returned in sealed rock box. The exterior surface con- tains a light-colored but inhomogeneous region about 1. 5-mm x 4-mm surrounded by dark matrix material. interior surface is similar dark matrix material. The -116- REFEHENCE3 A.jzenberc;- Selove, F, and Lauritsen, T, 1959, Nucl , Phys, 11, 1. 1\jzenber t;- Selove, F. 1970, IJucl, Phys, &h.2, 1, A.lzenberg-Selove, Ti', 1971, Nucl, Phvs. Al66, 1. Ajz enberg-Selove, F. 1972, Nucl. Phys. Al90, 1. A.lhee , A, L. 1 971 , private communication. Amsel, G, and Samuel, D. 1962, J , Phys. Chem. Solids Amsel, ~~ 1707. r. ., Bcranger, G., de Gelas, B, and Lacombe, P. 1968, J. Appl, Phys, 12• 2246, Bame, S. J. , Asbridge, J. R. , Ilundhausen, A. J, and Montgomery, M. C. 1970, J. 0eODbyS• Res. 75, 6360. Behrman, c., Crozaz, G., Drozd, D., Hohenberg, c ., Ralston, c., \-lalker, TI, and Yuhas, D, 1973, in Lunar Science IV, Lunar Science Institute, Houston, p, 54. Bibring, ,T, P., Duraud, J. P,, Durrieu, L,, Jouret, and Heunier, H, 1972, Science 175, 753. c., Maurette, H . Bikerman, J. ,J, 1970, Physical Surfaces, Academic Press , He-vr York, p . 1 90 . Borg, J., ~1nurette, H., Durrieu, L, and Jouret, C. 1971, Proc. Second Lunar Sci. Conf. 2027. l• BrUckner, n , 1971, J. Non-Crystalline Solids it 177. Burnett, D., Hohenberg, c., Maurette, M., Honnin, M., Walker, R, and \foolum, D, 1973, in Apollo 16: Preliminary Science Report, NASA, \fnshington, D. C. , p. 15-19. Cadenhead, D. A., Wagner, N.J., Jones, B. R, and Stetter, J. R. 1972, Proc, Third Lunar Sci, Conf. 2243. l• Cadenhead, D, A., Jones, B. R., Buerget, W, G. and Stetter, J, R, 1973, in Lunar Science IV, Lunar Science Institute, Houston, p. 109. -117Chao, c. Y., Tollestrup, A. v., Fowler, w. A. and Lauritsen, c. c. 1950, Phvs. Rev. ]2, 108. Choudhury, A., Palmer, D. w., Amsel, G., Curien, H. and Baruch, P. 1965, Solid State Communications l• 119. D'Amico, J., DeFelice, J. and Fireman, E. L. 1970, Proc. Apollo 11 Lunar Sci. Conf. ~· 1029. Dick, r.. J. 1973, private communication. Ducati, II., Kalbitzer, s., Kiko, J., Kirsten, T. andMtUler, H. W. 1972, preprint, Hax-Planck-Institut fur Kernphysik, Heidelberg. Eberhardt, P., r.eiss, J., r.raf, H., Grogler, N., KrahenbUhl, U., Schwaller, H., Sch,.rarzmuller, J. and Stettler, A. 1970, Proc. Apollo 11 Lunar Sci. Conf. £, 1037. ---El ~oresy, A., Randohr, P. and Hedenback, o. 1973, in Lunar Science IV, Lunar Science Institute, Houston, p. 222. Endt, P. !·1. and Vander Leun, C. 1967, Nucl. Phys. Al05, 1. Epstein, s. and Taylor, H. P., Jr. 1970, Proc. Apollo 11 Lunar Sci. ~· ~. 1085. Epstein, s. and Taylor, H. P., Jr. 1971, Proc. Second Lunar Sci. Conf. g_, 1421. Epstein, s . and Taylor, H. P., Jr. 1972, Proc. Third Lunar Sci. Conf. g_, 1429. Ericson, J. E. 1973, private communication. Espe, ':T. 19GG, ~-1aterials of High Vacuum Technology, Vol. 1, "Metals and :letalloids, 11 Pergamon Press, NeH York. F01·rler, H. A., Lauritsen, c. c. and Lauritsen, T. 1948, Rev. Hod. T'hys. ~. 236. Frank, L. A. 1970, J. <ieophys. Res. 1.2., 707. Friedman, I. and Smith, R. L. 1960, American Antiguity ~, 476. Friedman, I., Smith, R. L. and Long, H. D. 1966, Geol. Soc. America ~· 77, 323. -118Friedman, I. • Gleason, J. D. and Hardcastle, K. G. 1970, ~· Apollo 11 Lunar Sci . Conf. ~. 1103. Friedman , I. 1973, private communication. Fuller, E. L., Jr., Holmes, H. P. , Gammage, R. B. and Becker, K. 1971 Proc. Second Lunar Sci. Conf. _l, 2009. Funkhouser, J., Jessberger, E., !1uller, o. and Zahringer, J. 1971, Proc. Second Lunar Sci. Conf. ~. 1381. Gammage, n. B. and Becker, K. 1971, Proc. Second Lunar Sci. Conf. l• 2057. r.eiss, J., Eberhardt , P., BUhler, F., Meister, J. and Signer, P. 1970, J. Geophys . Res . Il, 5972. r.ibson, E. K., Jr., and Johnson, s. M. 1971, Proc. Second Lunar Sci. ~· ~. 1351. Grieve, R. A. F. and Plant, A. G. 1973, in Lunar Science IV, Lunar Science Institute, Houston, p. 317. Hines, R. L. and Arndt , R. 1960, Phys. Rev. 119, 623. Hintenberger, H., \-Ieber , H. W.,Voshage, H., Wanke, H., Begemann, F. and Wlotza, F. 1970, Proc. Apollo 11 Lunar Sci. Conf. ~. 1269. Harz, F. 1971, private communication. Iius~ins, !Jf. L. and Sun, K.-H., 1943 , J. Am. Cerrun. Soc. 26, 4. Hutcheon, I. D., Phakey , P. P. and Price, P. B. 1972, Proc. Third Lunar Sci. Conf. _l, 2845. Jovanovic, S. and Reed, G. W., Jr. 1973, in Lunar Science IV, Lunar Science Institute, Houston, p. 418. Keil, K., Bunch, T. E. and Prinz, M. 1970, Proc. Apollo 11 Lunar Sci. Conf . 1_, 561. Kelly, n . and Brmm, F. 1965, Acta Hetallurgica 13, 169. Kelly, R., Jech, c. and Matzke, Hj. 1968, Phys. Stat. Sol.~. 641. Kirsten, T. • 11Uller, o., Steinbrunn, F. and Zahringer, J. 1970, Proc. Apollo 11 Lunar Sci. Conf. ~. 1331. ( -119- ICirsten, Tq Steinbrunn, R. and Z'~hringer, J. 1971, Proc. Second Lunar Sci. Conf. ~. 1651. Kirsten, T ., Deubner, J. , Horn, P., Kaneoka, I., Kiko, J., Schaeffer, 0. A. and Thio, s. K. 1972, Proc. Third Lunar Sci. ~· ~. 1865. Lee , R. , Leich, D ., Tombrello, T., Eric son, J •. and Friedman, I. 1973, in preparation. Leich, D. A. and Tombrello, T. A. 1973, Nucl. Inst. and Meth. 108, 67. Lcjch, D. A., To~brello, T. A. and Burnett, D. Planet Sci. Lett. !£, 305. s. 1973a, Earth Leich, "J. A., Tombrello, ·r. A. and Burnett, D. s. 1973b, to be published in Proc. Fourth Lunar Sci. Conf. ~· Linclhard, J., Scharff, 1!. and Schi¢tt, H. E. 1963, Hat. Fys. ~1edd. Dan . Videnskab. Selskab .1h_ no. 14, l. Lord, H. C . 1968 , J. Geoph;vs. Res. 73, 5271. LHL (Lunar rteceiving Laboratory) 1971, Lunar Sample Information Catalog , Anollo 15 1 nASA report HSC 03209. LP.L (Lunar Rece iving Laboratory) 1973, private coimnunication. LSPET (Lunar Sample Preliminary Examination Team) 1969, Science 165, 1211. !vlaccabee , H. D., Raju, !1 . R . and Tobias, C. A. 1968, Phys. Rev. 165, 4G9 . ' ~::tcdour,all , D., Rajan, R. S .,Hutcheon, I. D. and Price, P. B. 1973, in Lunar Sc ience IV, Lunar Science Institute, Houston, p. 490. ' -'1ann , F . :.1., Pak, H.-B. and Denkin, N.H. 1973, private communication. l,~arshall , R . R . 1 961, rreol. Soc. American Bull. 72, 1493. 'latzke, H;J . 196G , Phvs. Stat. Gol. 18, 285. 'latzl>:.e , li,j. and 1.Jhitton, J. L. 1966 , Can. J. Phvs. 44, 995. ' ~eU ler, E . and Starfelt, N. 1967, Nucl. Inst. and T!eth. ~. 225. '·.:orev , r; . \·T. 1954, The Pronerties of Glass, Reinhold Publishing Cornoration , Ne•,r York. -120o rrison, D. A., r1cKay , D. s . and Heiken, G. H. 1912, Proc, Third Lunar Sci. Conf . 2, 2767. ' 1 lfi colet, 11 . A. and Feng, J. 1913, private communication. Horthcliffe, L. C. 1963 , Ann. Rev. Nucl. Sci. 13, 67. Horthcliffe, L. c . and Schilling. Section A 233 . l• ~. F. 1910, Nuclear Data Tables Ollerhead, ~. '.l, , /"lnqvist, ....... and Kuehner, J. A. 1966, J. Appl. Ph:rs. 31, 21~ l~o. P3.dm.'cr, r: . :•, and Gchneid , E . ,J . 1969, 1\.mer. Nucl. Soc. 'l'rans. 12, 1,93. Padn.vPr , r. . 1-1 . 1 910, l;ucl . Anul. and Tech. r>adauer , 2, 1356. r. . !1 , 1913, :nrivate conununication. Pn.1ner, I"J , '.: . 19115 , ;;ucl. Instr. and ~1eth, 38, 181. T"ntterson , ,J. li ,, 't:'url:evich , A. L., Franzgrote, E. J., Economou~ T. E. anc'!. So-vrinski , K . P. 1910, Science 168, 825. Phake;r , P . P., !Iutcheon, I. D ., Rajan, R. s. and Price, P. B. 1912, Proc. Third Lunar Sci . Conf. 2, 2905. Phi11iT's , 1:. 'i . and ~ead, F. II . 1963 , Proc. Phvs. Soc. 81, 1. J'orat , D . I. and Haw.avataram, K. 1960, Proc. Phys. Soc. 16, 438. Heed , r.. \!,, Jr. 1964, Geochim. Cosmochim. Acta 213, 1129. l~ced , ~.J ., .Jr. 1911, in Handbook of Elemental Abundances in :a. ··~ason , ed., Gordon and Breach Science Publishers, llew York, ~· 1 03 . r;. r~eteor i tcs, :JchigStt , u. "!"':, 1966, 1·1at. Fys . '1edd. Dan. Videnskab, Selskab l2_, no. 9, 1. 8chj¢tt , rt , :1':. 1 910, ;{adiation I:ffects §., 101. M. , Stauber, ' ·1 • C., P ada•N"er, r;, Brandt, W., D'Agostino , ~1. D., Ka.mykm.rski , J: . and Young , D. A, 1913, submitted for publication in Proc. Fourth Lunar Sci. Conf. -121~;tevels , .J. 't. lllli£3, Progress in the Theory of the Physical Properties of Glass , Elsevier Publishinc; Company, New York. Sutton , R . L., iiait, ~1 . Ii., Larson, K. B., Swann, G. A., Reed, V. s. and i3chaber, G. G. 1972, Documentation of Apollo 15 Samples, U.S.G.3. Interagency Report: Astrogeology 47. 'l'a~rlor, L. A. 1973, in Lunar Science IV, Lunar Science Institute, Houston, p . 715. 'l'hompson , D . fl. . and '·1ackintosh, W. D. 1 971, J. Anpl. Phys. 42, 3969. ~ombrello, T. A. and Leich, D. A. 1971, IEEE Trans . on Nucl. Sci. ns-18, lG4. 'l'urkevich, A . L. 1973, private communication. \.Jhitton, J. L. , T1 itchell, I. v. andHinterbon, K. D. 1971, r.;o.n . ,T . Phvs. 49, 1 225 . :Tilshire, ;r . r: ., Stuart-Alexander, D. E. and Jackson, E . D. 1973, in Lunar Science IV, Lunar Science Institute, Houston, p. 784. 1 Hinterbon, K. B., Sigmund, P. and Sanders, ,T. D. 1970, ~at. Fvs. ~Ae dd. Dan. Videnskab . Selskab ]l, no. 14, 1. 'tlolfe , ,J. H. 1972, in Solar Hind, NASA SP-308, Hashington, D. C. Youn g , E . J ., ~ Ayers, A. T. , Munson, E. L. and Conklin , N. 1\1 . 1 969, U. S .G. S . Professional Paner 650-D, p. 84. Z.e ller, E. J ., Dreschhoff, G. and Kevan, L. 1968, Hodern Geology !_, 1111. 150 200 200 200 6 8 60 100 200 5 9 10 9 9 5 10 10 9 16 Imnlantation Dose (xio 15 ions/cm2 ) 75 150 150 200 200 Imnlant at ion Current (nA) hp 33 34 40 34 50 35 35 35 39 34 24 39 ( f]g/crr,2) n 43 38 44 33 45 33 41 35 39 37 35 44 RH (f1g/~m2) F\'7llN 45 46 50 35 47 35 59 23 47 41 35 45 (f]g/cm2) (1966) predict a of 25 f1g/cm2 and 26 f1g/cm2 , repectively (see text, page 20). mean range of 34 ).1g/cm2 for Si0 2 and 35 f]g/cm 2 for the feldspar composition, with FWHH's for samples implanted with ~12 keV protons. Calculations from Schi~tt ',lost probable range (Rp)·, median range (RM) and full-width at half-maximum (FHIU!) are given (Ca-~ich) Feldsnar Fused Silica Quartz Cr~rstalline Sar.rnlc T~T?e TABL.C: l I I f\) f\) f-' -123- TABLE 2 Reaction Used ER (MeV) y-ray Energy (MeV) II lE(l9F,ay)l60 16.45 6.1, 6.9, 7.1 51 F' l 9F(p,ay)l60 0.872 6.1, 6.9, 7.1 6.5 Na 23Na (p,y)24Mg 1.318 11.6, 13.0 0.03 A'l. 27A'l.(p,y )28Si 0.992 10.8 0.01 Element Analyzed Sensitivity (xlo-13) Resonant nuclear reactions used for analysis of H, F, Na and A'l. are given alone 1vith the incident ion energy and the energy of the principal characteristic y-ray reaction products. sensitivit~r , In addition, the reaction defined as the thick target yield per incident ion for a ouartz target containing l ppm by weight of the particular target atoms, is also given for each reaction (see text, page 22). -124TABLE 3 Hydrogen depth profile data summary. The "surface" H concen- trations pertain to observed peaks within ~300 Aof the surface . The total H content of the 0-0.4 ~ deep measured region, expressed as a surface density, does ~ include the amounts identified as "surface" H, so the total H observed is the sum of the values in the first two columns. The H content, expressed in npm H by weight (100 ppm is equivalent to 1.55 x 1020 H atoms/cm3 assuming a nominal density of 2.6 g/cm3), is also given for depths of 0.1 ~m and 0.4 ~m. The profiles are charac- terized by the depth Xp at which the peak H content is observed and the full-width at half-maximum (FWH11) of the distribution. (The apparent broadening due to the resonance width r 1 has been removed in the calcul ation of the latter quantity.) The size of the fused silica coll imator used with each target is also listed. The fraction of the beam hitting the sample is estimated as 0.6 ± .2 for the 2.0-mm size, 1,0 + _ 0,2 o.8 ± .2 for the 2.4-mm size, for the 3.2-mm size, and 1,0 for the 4 .5-mm size and the samples without collimators. tration are ± ± Uncertainties in the H concen- lO% with lm·rer limits of ± 20 ~pm (see text, Section III, Part A). 2 0.5 x lo1 5/cm and - 125- TABLE 3 H content Sample Surface (xlol5cm-2) 0-0.4)Jm (x1ol5cm-2) at x=O.l)Jm at x=0.4\lm (npm) (ppm) 120 80 300 80 30 60 10085,1 10085,31-Sla l0085,31-12a 151113, 5-2a 15413,5-5 1 5533,4-1 <2 <0.5 <3 <1 <0.5 0.5 19 9 55 17 5 370 160 1500 4oo 70 80 15015,39-1 1501 5 , 39-1(int . ) 1')01:i,3'i-2 l5015 , 30-2(int.) 15059 , 28 "-10 2 2 2 ll 4 2 3 2 3 3.5 2 100 50 30 40 70 120 20 40 30 40 50 40 3 2.5 3 2.5 2.5 2.2 90 l10 100 50 (,4455 , 33- ~ Gll45~ 1 33 -2~int . ~ 3 1 G ll 3 2 40 30 30 50 30 30 ;20 G5 315 , 6* G5315 , h(int . )*h h5315 , 8(int. )* (5315 , 20* 6 :231~ 1 20~int . ~* 0.5 1.5 0.5 1.5 3 5 100 110 10 150 80 40 30 0 10 30 (,(iOllll ' 8-.1\ ·Y.· (,(,o44,8- B (,131?11 ' 3-/\·~ :;s1?.l1, 3-:R (,8124 , 10- 1\* GG124 1 10-B 0. 5 2 l 1 <0 . 5 1 5 90 6 50 <100 10 10 30 ~ 100 30 30 60 60 GD815 , 27 G8812 ~ 2I( int . ~ "-3 1.5 23 1 500 30 150~2 1 3 ? ()4455 . ~l4 6l145) , 2ll(int . ) (.1~4 55 ' 33-1 ()1~1155 , 33- 1( int . ) ~~ 2 1 7 ~ 2 2 3 r;o ~0 150 20 -126TABLE 3 (continued) Xp ( ).Jm) ri@1 ( )Jm) Collimator size (mrn) Sample 0.09 ± .03 0.13 ± .05 0.11 ± .02 0.09 ± .03 0.07 ± .03 'VO.l 0.27 ± . 07 0 . 22 ± .10 0.22 ± . 03 0.20 ± .04 0 . 20 ± .05 2.0 2.4 3.2 2.4 2.0 2.4 10085,1 10085,31-9 10085,31-12 15413,5-2 15413,5-5 15533 4-l <0 . 02 <0.02 <0.02 <0.02 <0.02 <0 .03 <0,08 <0,05 o.o4 ± .02 <0.05 0,02 t ,02 0.03 t .03 3.2c 15015,39-l 15015,39-l(int.) 15015,39-2 l5015,3')-2(int.) 15059,28 15059.32 <0 .02 <0.02 <0.02 <0.02 <0.03 <0.03 <0,05 <0.05 0.05 t .03 <0.05 <0.05 <0,05 64455,24 64455,24(int.) 64455,33-1 64455,33-l(int.) 64455,33-2 64455.33-2(int.) 0 .11 ± .05 <0.03 <0 .03 <0.05 <0 .03 0.22 t .05 <0.05 <0.05 0.15 t .05 o .os ± .o4 65315,6 65315,6(int.) 65315,8(int.) 65315,20 65315,20(int.) <0.05 <0.03 <0 .03 <0.03 <0 .05 <0 .05 0.15 ± .03 >0.4 <0.05 <0,05 >0.4 0.12 ± .05 0.03 ± .02 <0.02 0.18 ± .02 <0.05 >0.1~ 3.2c 3.2c 4.5 4.5 66044,8-A 66044,8- B 68124,3-A 68124 ,3-B 68124,10-A 68124,10-B 68815,27 6881 5 2'(( int.) -127- TABLE 3 (continued) F distribution measured prior to H distribution. a Ultrasonic rinse in high-purity acetone prior to analysis. b Values c;i ven are averages of measurements on tvro separate pieces of this sample. c Collimator material is pyrex rather than fused silica. The small H content of the pyrex may contribute to the measured H for these samples. int.: interior The F content is 10% vith a lower limit of ± 20 ppm, The apparent depth xp at •rhich Uncertainties are The surface potential of the sample under F distribution measured prior to H analysis. not determined n,d,: * interior int,: for several of these samples (see text, Section III, Part A). irradiation, as measured using the reaction 27A~(p,y) 28 si, is also given the measured distribution. the peak F concentration is observed is f;iven along 1-rith the ?HIT': of ± a layer extending from ~0.5 ~m deep to ~l ~m deep. given for a layer extending from the surface to ~0.5 ~m deep and for F distribution data summary for Apollo 16 samples. TABLE 4 I c:> f\) I I-' 480 50 50 50 120 540 1400 50 120 130 100 210 120 1000 100 140 90 160 820 1900 410 850 260 220 460 180 65315,6* 65315,6 (int.)* 65315,8 (int,)* 65315,20* 65315,20 (int.)* 66044,8-A* 66o4lr,8-B 68124,3-A* 68124,3-B 68124,10-A* 68124,10-B 68815,27 68815,27 (int.) 70 50 50 20 40 llO 0. 5-l. 0 \Jm (ppm) 270 70 210 60 4o 100 0-0,5 \JID (:ppm) 64455,24 64455,24 (int,) 64455,33-l 64455,33-l (int.) 64455,33-2 641155,33-2 (int,) Sarnnle F Content TABLE 4 n.d, n,d, 0,08 0,18 0.15 0.05 7 ± 3 2 :t 2 10 ± 3 8 ± 3 n,d, (14) -vo.4 0,08 0.13 0,13 0,20 0,20 0,10 0.03 0,13 0,12 0,09 0,17 n.d, n.d. ~5 3 :t 2 ~5 0,17 0,30 0,20 -vo.8 0,18 0,10 0.12 0.12 0,12 0.05 Surface Potential (kV) n,d, n,d, n.d. n.d. n.d. < 10 ( um) F'Hlf l 0.35 0.22 0.18 0.05 0,05 0,10 0,05 0.18 no :peak observed no peak observed Xp ( lJm) \0 I I f-' r0 The quoted statistical uncertainties do not include a The measured hydration layer The uncertainties listed indicate the for comparison. nesses of thin-sections taken from the same obsidian samples are included Optically measured hydration band thick- 4 where the uncertainties are determined by the resolution of the 1H( 1 9F,ay) 16o technique. for samples 1 - extent of the depth region in vrhich the diffusion front is located, except gradient is observed, is also given. thickness h, determined by the depth at which the maximum H2o concentration detection efficiency and stopping power. possible systematic error (of up to 5%) introduced by the estimates of Figs. 29 and 30). The values given for H2o content are hydration plateau values (compare w·ith ' 1easurenents of hydration profiles for a selection of obsidian samples. 'l'ABLE 5 I 1--' I 0 w Initial exposure August 1963 4 i3ig r'.lass \1ountain, 7t tunapa, :Nayarit, Mexico ''!exico Artifact {2.15 ± .1 1.56 ± .05 1.9 ± .5 6 .ll.ma"9a, Nayarit, Artifact 2. 5 ± .1 5 Big Obsidian Flow, Oref,on California 0.9 ' .1 } 1. 77 ' ± .27 1. 55 ± • 50 1.2 ± .1 0.19 ± .02 1.7 ± .2 1.4 ± .2 1.2 ± .2 0.39 n.d. 0.14 ± .02 2.45 ± .14 Artificially hydrated at 75°C for 4 days 3 :Big Obsidian Flovr, Oregon 2.16 ± .12 n.d. 02 • n.d. O.ll ± .02 2.08 ± .13 ± Optical Techniq.uelf Artificially hydrated at 75°C for 2 days Oregon 0.09 Technioue 1H( 1 9r, a Y) 160 2 Big Obsidian Flow, Oregon 1 2.11 ± .14 II20 Content (nercent by wei ght) Artificially h~rdrated at 75°C for 1 day Description h(\.lm) 3it, Obsidian flov, Samnle Source TA:::JL.r: 5 I 1--' I 1--' w Artifact Chipping waste 9t Amapa, l{ayarit, "exico 10 Borax Lake, California 1.13 :t .5 0.7 :1: .5 2 .1 :1: .2 1.82 ± .25 0.53 • ,l} Optical Technique* 1. 8 ± . 2 o 1.8 :1: .2 Techniaue 1 n(l9F ,ay) 16 h(um) The higher H2 o content applies from t his der,th and the larger den.th. the surface to the first denth and the smaller content renresents the :rr 2 o level betveen anpearance of t•ro plateaus in the h~rdration profile. tTwo values are given for H2 o content and the hydration layer thickness h due to the estimated hydration rate (Friedman, 1973). not an actual measurement , but an estimate based on the kno\m exposure time and an The value of h given for sample 4 is 2.8 :1: .1 {2. 65 :t .10 2.1 :1: .1 2.3 :1: .1 H20 Content (nercent by weight) *Data are from Ericson (1973) and Friedman (1973). 8 Artifact Description Amapa, ITayarit, '·' exico Sample Source TABLE 5 (continued) I 1\) I t--' w -133- TABLE 6 H2 0 Content (percent by vreight) Intrinsic Hydrated Bodie Hills, California 0.21 ± • ot~ 2.31 1 .20 Coso, California 0.23 1 .07 2.68 1 . 25 Borax Lake, California 0.35 .09 3.32 ± . 32 East Dago Valley, California o. 66* ± .o6 3.47 ± . 23 Source ± *Mean of three samples from same source. o.6o , 0.69, and 0.10% H2 o. Individual analyses Here Heasurement of H2 o content are r;iven for hydrated a nd unh;.rdrated samples from each of four California opsidian sources. ~~uoted uncertainties reflect random errors in the measurements but do not include a possible systemat i c error (of up to 5%) introduced by the estimates of detection efficiency and stopping pmrer (see text, page 88). text, page 9 ). I !::' vridth of the resonance, and the .e nergy spread of the 1 9F ion beam (see I ...... w The solid curve is calculated from equation A.9 in ADpendix A for a quartz sample including the effects of energy straggling, the energy of depth. Depth resolution of the 1H( 1 9F,ay) 16o H analysis technique as a function FIGURE 1 -135- NOilnlOS3~ (w-rl) 0 0 z C\J 0 ._ 0 0 l{) 0 ::::::> CD _J .,_ 0 >-..(/) ts w lJ...- l{) en a:: I I - ._ - 0 0 v E Cl.. ~ w E 0 ::1.. Io .__; - t\1 0\ rl o rt') ;~ 0 ::1.. ._I CL w CL w 0 oo 0 C\J l{) d 0 0 co U) v C\J <aw~;o-rl) NOilnlOS3~ r:~ ,..) ('" 1-1 r, , The insert shows the arrangement of target mounts. text, page 11.) electronics. (See Schematic d.ra,dng of the target chamber vacuum system and detection FIGURE 2 I I-' I Ci\ w GATE VALVE ROTARY TARGET WHEEL ~SPOT VERNI~F BEAM DIFFUSION PUMP SLITS ACCELERATOR BEAM LINE LN2 LN2 :r:;:;:c IN - LINE TRAP 2 ION PUMP PHOTOTUBE Nai(TI) CRYSTAL JARGET WHEEL I I-' I -:J w The 2 3 are also shown (the double-escape peaks are lost in the y 1 full- The peak at channel 58 is room backeround due to the decay of 4°K by electron capture to the 1.46 ~1eV state of 40Ar. ground is negligible above channel 100.) (Back- The relatively flat spectrum between channels 100 and 200 is due to and y Compton scattering of the three high-energy y rays in the Nai(Ti) crystal, energy peak). peaks from y The positions of the single-escape ray are easily resolved at the positions indicated by y1 -mc 2 and y 1 -2mc 2 , respec- tively, ••here m is the rest mass of an electron. y single-escape and double-escape peaks following pair production from the 6.1-HeV 6.1-, 6.9- and 7.1-MeV y rays are indicated by y1 , y2 and y 3 , respectively. The positions of the full-energy peaks due to the A 30-nA beam of 1 9F 4+ was used •Ti th a total integrated charge of 10 ~C (see text, page 14). :-fountain chlorite. Pulse-height spectrum for 17.2-MeV l9F ions incident on a sample of Belvidere FIGURE 3 I I-' I co w -139- w ra::: 0 _J ::r: (.) ~- LL... a::: 0 r0 ) ~ow-~..<~ow- ~.A- z m ,, ~.,(- ,.J" \, !.<- 0 0 0 ~ow-!,{-'··· .. .. .. .. • • • > .. . , .. . .,, 0 1.() C\J ; ,·.. '· .. ~owz- !,{- LL... a::: w 0(]) .., I 0~ C\J=> ., '·f} ~ ::> a::: r(.) z p~ (f) a.. (!)T r- -o t--0 ~, ·~ ::r: zz ::>_ { w ::r: I w u 03: \ z3: (9 • ::> a_ . . . . ..'· >to~- 0 0 l',, 0 1.() 0 0 1'- 0 0 (.£) 0 0 1.() l3NNVHC> 0 0 0 0 ¢ r0 ~3d SlNnOC> ~::> w H (_) () (f) _J _J oz l.()z -<l ::r: '( w (Y) .~ 0 c r, . -140- FIGURE 4 The y-ray yield versus bombarding energy for an implanted fused silica sample and for an identical non-implanted (blank) sample. The energy scale shows the resonance energy F~ subtracted from the l9F beam energy E. A calculated depth scale is also shown. Error bars show statistical uncertainties. (See text, pa ge 19 .) PROPERTY OF r~ ' LOGG RADIATION LABORATORY -141- DEPTH (ANGSTROMS) (J)I5 z 0 1000 2000 3000 I I I I 4000 1 • IMPLANTED • BLANK ! f f! ! ~ - 0 lL "'0 10 - - 0::: LIJ Cl. (/) .,.._ z (._) I t t .J- I I ! 0 - ! 65 I ! I I I ! ' I ! ! I I I 0 500 1000 E- ER (keV) ? I ~URE 4 -142- FIGURE 5 Implantation profile for fused silica sample implanted "rith -vl2-keV protons. Experimental points •rere obtained b:v sub- tracting the data for the blank from the data for the implanted sample shown in Figure 4. (See text, page 19) -14 3- DEPTH (fLg/cm2) - 0 f(') E4~ I 20 I 40 60 100 80 I I I 1- ~ CJ) E 0 0 0 '03 -z - !! ! t t ~ 0 t-<l ~2 f- - z w (.) z 0 (.) - zl r- w <.!) 0 a:: 0 >- Io t ! ! ! f I 0 I I I I 1000 2000 3000 4000 DEPTH (ANGSTROMS) ! -144- FIGURE 6 Depth resolution as a function of depth for F, Na, and Ai in quartz. The resolution curve for each of these elements is indicated by the reaction used in the depth analysis measurement for that element (see text, page 22). The depth resolution for H (Figure 1) is also shown for comparison. -145- DEPTH 0 .5 0 (fLm) 1.0 1.5 2.0 35 0 .12 30 I 0 .10 I I I I E I ::t I 0.08 z I Z20 0 0 t- 1-- :J 0 0 0 .06 ~ :J _.J _.J (f) (f) w a:: .I .:..: I i .. 0.04 i/ ..if I! if 5· ---- -- \60 'r\t\9f,ay) --------- -- --- - ------- --- 0.02 0 o':----l_.o_o_ _2. .o-o---=-3o~o--4-oL-o------=5--~:o-o ... _ __, 0 DEPTH (fLg/cm 2 ) FIGUHE G -146- FIGURE 7 Hydrogen concentration versus depth for lunar glass sample 10085,31-12. The top scale is the difference between the l9F beam energy, E, and the resonance energy ~· scale indicates the y-ray counting rate. The right-hand Data points for this figure and those following are number of counts with their associated statistical (lo) uncertainties plotted against l9F beam energy. Hydrogen concentration and depth scales are cal- culated using a nominal density of 2.6 g/cm3 • The zero point of the hydrogen concentration scale corresponds to the background counting rate, indicated by the displacement of the zero point from the bottom border of the graph. The conversion factors used to obtain the calculated scales are uncertain by about 10% due mainly to the uncertainty in the electronic stopping power of the lunar material for 1 9F ions. Open squares are data from the third measurement of the distribution. Uncertainties are com- parable to those on the data points for the original measurement, but error bars are omitted for clarity. "First run" data were taken in 50-keV steps from below the resonance energy to about 1-HeV above the resonance. reverse order. "Third run" data were taken in the (See text, page 28.) E-ER (keV) -E -200 0 200 It) 400 600 1000 800 (.) ........ tr\i- U) E2 0 -z 0 10085,31-12 •- FIRST RUN o- THIRD RUN } c.\1 0 0 (f) z 1- 0 <( 0 1- z 0 (.) z z L1J ~ 0 ~0 ~ L&.J CL '~ 0 0 --0 \ 0 oo!---~~ (.) (.!) lJ.. 10 (7) t 0 L&.J 0 \ 0 ~ I 15 t,, 0 I 0 5 -~ -~ (f) 1- z :::) 0 r' o (.) D 0 1000 2000 3000 DEPTH (ANGSTROMS) FIGUHE 7 4000 0 FIGURE 8 Hydrogen concentration versus depth for lunar glass fra{gllent 10085.1. "First run" data were taken in 200-keV steps decreasing in energy (depth) • lines) "Second run" data (connected by straight were taken in the reverse order. The H concentration is expressed in terms of ppm by weight as well as in atoms/cm3. 'rhese scales have been calculated assuming the fraction of the beam hitting the sample was 0.6 with the remainder striking the 2.0-mm aperture fused silica collimator and 9 consequently, are uncertain by about 20%. (See text, page 30.) - 10085,1- GLASS FRAGMENT IC') E •-FIRST RUN o-SECOND RUN ~ (/) -~ 0 f 6 0 t\1 0 -z4 I 0 ~ <! a:: ~ z 2. w u 100 z 0 u IO 0 0 1000 2000 3000 DEPTH (ANGSTROMS) FIGURE 8 I 4000 -l ) O- FIGUR~ 9 Hydrogen concentration versus depth for pyroxene-ric h lunar roc}, fragment 15413,5-2. "First run" data were taken in steps decreasin13 in energ:v ( de:tJth). "Second run" data (connected b;r straight lines) were taken in the opposite direction. The fraction of the be~ striking the 2.4~nm aperture fused silica collimator was estir.1ated at 0.2 in the calculation of the E concentration scales (with a 20% uncertainty). pa ge 31.) (Gee text, - 151- 15413,5-2 •-FIRST RUN o-SECOND RUN -6 400 I(') E (.) ........ (/) E 0 0 0 t\1 0 -z 4 ~~~~ ~ 300 I 0 I 200 E a. a. k"~~! ~ <(2 a:: ~ z w (.) z ~~---~ 100 0 (.) I 0 0 0 1000 2000 3000 QEPTH (ANGSTROMS) FIGUTIE 9 4000 -152- FHmRE 10 Hydrogen concentration ver::ms depth for glass-coateu lunar rock chip l5015s39-l. rock 15015. This sample comes from the lunar bottom of The counting rate for the initial data noint (taken at the resonance energys corresponding to zero depth) was a factor of 2 too high to be included in the fip;ure, " ><irst run" data are in increasing energy (depth) stepss with the direction reversed for the "second run" data , (See text, page 36 .) -153- 1 -E 5 15015,39-1 •-FIRST RUN o-SECOND RUN ~ rt) (.) .......... U) 300 E -4 0 0 0 "'0 - 200 z 3 0 I ....... <l a:: E ....... a. a. z 2 w 0 z 100 0 0 z w (!) 0 a:: /~ } ~ ----~2-~ 0 >-o I 0 1000 DEPTH 2000 3000 (ANGSTROMS) Flr~UIU:: 10 0 4000 -154- FIGURE 11 Hydrogen concentration versus depth for glass-coated lunar rock chip 15015,39-2, from the lunar bottom of rock 15015. " First run" data were taken in increasing energy (depth) steps ;lith the direction reversed for the "second run" data. page 36.) (See text, -155- - 4 15015,39-2 IC') ~~ E 0 ......... (/) E 0 0 0 (.\,1 0 z ~ ~ <t ~I z 0 I \ I 0 (.) 200 \ 2 a:::: z w •- FIRST RUN o- SECOND RUN 100 E a.. a.. A( l"'rr~--+----~ ;2 2 (.)0 0 -::r: 0 1000 2000 3000 DEPTH (ANGSTROMS) F I GURE 11 4000 -1 56- FIGURE 12 Hy drogen concentration versus depth for glass-coated lunar rock chips 15059,32 and 15059,28, from the lunar top and bottom of rock 15059, respectively. The initial data point for each sample (taken at the resonance ener€Qr, corresponding to zero depth) was too large, by almost a factor of 2, to be included in the figure. The same energy was repeated for the second data point, with increasing energy (depth) steps for subsequent "first run" data on each sample. "Second run" data are in step s of decreasing energy (depth). (See text, page 37.) - 157- E- ER (MeV) (). ~ ().~ 1.0 ()_E) ~----~----~--- -E 5 i 15059,32 (TOP) I •-FIRST RUN o- SECOND RUN 15059,28 (BOTTC>M) ~00 •-FIRST RUN RUN !'f) A- SECOND u ......... (/) (/) z E4 0 - 0 0 300 0 t\1 -z ro -0 0 3 0 06. ~ <t a:: ~2 z ! a:: 20() ; (..) z 0 0 (..) z 6. 0 L&J (/) ~ ~ z 0 .l 0 6. t (!) 0 0 >-() L&J Q.. 0 L&J a:: LL (7) 0 <}. :::J 0 (..) ~k 6. ) 100 0 6. 0 :I: ~----~----~~----~----~------~~() 0 10()0 2()00 3000 DEPTH (ANGSTROMS) FIGURE 12 ~()()() -158- FIGURE 13 Hydrogen concentration versus depth for feldspar-rich lunar rock chip 64455,33-1, lunar exterior and interior surfaces. Error bars on the "interior surface" data are omitted for clarity, but are of the same size as for the "exterior surface" data. Straight lines have been drawn to connect the "second run" data and the last three points from the "first nn" for the exterior surface. 'rhe dotted curve shows the calculated appearance (includine background) of a stable H 0 monola~er (~9 A2 per molecule corre2 1 sponding to ~2 x 10 5 H atoms/cm2 ) on the surface of the sample. The width of the peru• at zero depth reflects the resolution of this technique. du e to counts from the resonance at 17.64-HeV 1 9F enere;y . text, page 41.) 0 The curve risE·s at depths ~reater than ~!~000 A ( See - 159- 644 55,33-1 FELDSPAR-RICH ROCK e- FIRST RUN / •- SECOND RUN I EXTERIOR SURFACE 0 0-INTERIOR SURFACE 400 • •• •• • H2 0 , MONOLAYER (CALCULATED) E u ....... 0 VI 300 E 0 0 c N 0 I z E 0 200 ~ 0::: ~ z w (.) z 0 (.) 2 100 I 0 I I ··· ··... ·········.... ·· 0 0 0 0 ················································ 1000 2000 3000 DEPTH (ANGSTROMS) FIGUHE 1 3 ... 4000 0 ~ -1 )0- FIGLRE 14 Hydrogen concentration versus depth for glass-coated lunar rod: chip 64455,24 and feldspar-riel. lunar rocl~ chip Gld1)5,33-::. "lat a from "interior surfaces" are also plotted for both sa.r.tples , vith error bars omitted for clarity. The calculated appearance of a stable H20 monolayer is again jncluded for comparison. text, page 41.) (See - 1 61 - 64455,24 5 GLASS •-EXTERIOR SURFACE 6 -INTERIOR SURFACE - 300 r tl E 64455,33-2 ';;;4 e -EXTERIOR u ROCK SURFACE 0-INTERIOR SURFACE E 0 0 0 :::3 z 0 1-- (22 1-- z w (.) z 200 I ... .. . . ..... ..... . : :o E ~: . : 6 0 · •·••· H20 MONOLAYER (CALCULATED) 0 N Q. Q. 100 • (.) 0 6 I ' .. · 0 ····g)··· 0 0 ..... i I ................. ..... .......... .............. .. . 0 1000 2000 3000 DEPTH (ANGSTROMS) FIGURE 1 4 4000 omitted for clarity . Error bars for "interior surface" data are (See text, page 43.) for the exterior surface. 64455,24. Straight lines have been drawn connecting "first run" data Fluorine concentration versus depth for glass-coated lunar rock chip FIGURE 15 I I 1\) C\ 1-' (1) lL. 8 0~ !:> 0 » 0.2 0 (fLm) 0.6 0 0.8 ~f 0-INTERIOR SURFACE 1.0 e-FIRST RUN ~ EXTERIOR A-SECOND RUN SURFACE 64455,24 1 ~ ?Ir;LBJ::; 15 0.4 DEPTH ~ 0 ! a::: ~ z w 0 z j!~\ ;! ! i\ I 0.4 ~ 0.2 ('""'\ z - 0 0 0 -2 E (f) .......... E (.) ,;) 0.6 E a. a. t.o ~200 ~ r"'\ 400 w za::: 1600 I w I I-' C\ for clarity. ' chips 64455,33. (See text, page 43.) Error bars for "interior surface" data are omitted .;::- I C\ Fluorine concentration versus depth for two feldspar-rich lunar rock FIGURE 16 ~ lL (.) 0 z w (.) z I- 0::: <{ 0 I- 0.2 0 z ...- 0 t\J 0 0 0 - E fJ') ~ E 0.4 -0.6 !'(') I 0 f k ! }! Q. }{ 0.2 f 0 0 t } :::'IGU!1E 1 6 b. 0.8 0 ! b. • - EXTERIOR SURFACE b.- INTERIOR SURFACE 0.4 0.6 DEPTH (JLm) ~ ..f! ± <A I tf}J t 64455,33-2 b. e-FIRST RUN EXTERIOR • - SECOND RUN SURFACE 0- INTERIOR SURFACE 64455,33-1 1.0 r A a. w -iO L1. too~ I E 1400 Cl. -1600 I I Vl 0\ ...... r;>age 46.) (See text, Error bars for "second run" data are omitted for clarit~r. Straight lines have been drawn connecting the data points. energy (depth) • Data for "first run" and "second run" were taken in steps of increasing Fluorine concentration versus depth for lunar glass chip 65315,6. FIGURE 17 I C\ I t-' C\ - 167- .:J JJdd 0 0 0 0 0 0 0 1.0 1.0 0 z (/) (/) <l: _j (.!) <.0 ~ 1.0 r<> 1.0 Z=> =>cr cr co d 0 r-z U)0 cru -W LL. (f) I I • - <.OE 0 0 - :::1.. I <.0 ....... ~Q_ .w Oo C\.1 0 - --0 0,0 0-- -o- - 0 --._ 90 0 ~0 0 1.0 0 (£w:>;swolo ozOI) NOI.l tt~l.N3~NO~ ..:f 0 ['-~~ -1 )8- FIGLRE 18 Hydrogen concentration versus depth for lunar breccia chip 68815 ,27. Data from an intericr surface of the same chip are also plotted with error bars on.i tted for clari t:r. The first two data points (taken at the resor.ance energy, corresnonding to zero depth) for each surf'ace show a comparable decrease in countinc rate. Subsequent "first run" J>Oints are in increasinG energy (depth) steps , vrith the directjon revers ed for " second run" dat a . (See text, page 49.) -1"0- Curves drawn through each of three consecutive data sets (runs) for the 68815,27 hydrogen depth distribution (Figure 18) show the evolution of the profile due to irradiation vith the 1 9F beam. Straight lines have been drawn connecting the data points except near the peak, where a somewhat arbitrary curve has been drmm as an estimate of its shape for each run. The dash-dot curve connects data for the interior surface, with the vertical dashdot line at zero depth connecting the three data points taken at the resonance energy, each having a successivel;.r lmrcr yield . (See text, page 49.) -171- 68815,27 12 --FIRST RUN - - - SECOND RUN ········THIRD RUN ·-·-·INTERIOR SURFACE f "\ 1 \ 10 I I - I I I I / I • ,E (.) ~8 E I : I ; 0 I; I. - : 0 0 I; I: I; I: I; C\1 0 ::::.s z \ \ \ \ .···.I \\\ •• ·. \ ·.. \ ... \ ••• \ . ' ·.. \ •• ' ·.. ' ·. ·~ r- 1 : 1: I: w ,: 0:::4 rz Q. Q. \ \ \ \ • \ ·... .... ;"'~'"----- \······ ' ·.. . I: \ ·········-- li (..) ' .. 1: 1 1: I z 0 02 1: . I : I:" :r: ...., .. "" ... I i 1/ 1.· .........;;·· 0 \ ·.~ : <t E ·~ ... I: 0 I /·,.-.-.-. -·.-. -·-·-·-·-·-·-·- -·-·-. 0 1000 DEPTH 2000 3000 (ANGSTROMS) T'IGlJIH~ 19 4000 -172- FIGUHE 20 Fluorine concentration versus depth for breccia chip 68815,27, both lunar exterior and interior surfaces. :L;ach curve is dravm through two consecutive runs, showing the excellent reproducibility of these measurements. The dotted curve represents the data for the hydrogen concentration on the same sanrnle multiplied by 1/10 and plotted on the same depth s cale for comparison. 'l'he difference beb.reen the proton bea.m energy E and the resonance energy ER (= 872 keV) is shown on the to:o scale. The fluorine content is also expressed in 1;1})::1 on the rieht-hund s cale. sam:9le error bars are shmm. (See text, pae;c 52 .) Onl;r ~.'he data re1_)resented b:r t he solid circles v ere taken follovring are shmm . (See text, rage 54 ,) the <lata sho"m b~.r the open circles had been taken. Onl y sample error bars a ~ - P"'.m shift in the center of the :nroton be~ spot on t he S8Jrt"9le aft er G6o44 , 8. Fluorine concentration versus depth for lunar anorthosite fragment FIGURE 21 I f-' --:) -175- ~ wdd 0 0 0 0 0 l{) 0 0 l{) 0 ... I Q 00 d <l: CD - w E ·:::t. 0....._ I .._...._ • • I .,__ 0 ~a... .w oo C\J d 0 l{) 0 C\J C\J l{) (£W:>/SWOJO 0 ~01) Q 0 C) NOll.V~J.N3QNOQ .::J l{) rl ('J -176- FIGURE 22 27Ai(p,y) 2 8si measurements on lunar samples. This composite figure shm•s data for 11-i'·!eV y-ray counts per 6 )JC of incident protons versus proton energy near the 992-keV resonance (indicated by the solid line at ER) for six lunar sam~l e s urfaces . The estimated shape of the step (due to a presumably uniform aluminum content in a particular sample) is shmm for each sample as a solid curve, with an arrovr drawn to the center of the step indicating the magnitude of the shift in energy from ~~ The implied average surface potential is g iven for each sample. The more gentle slopes (e.g. 68124,3) probably indic ate large (a few keV) fluctuations about the average potential, while the steeper slopes (e.g. 65315,6) may be largel~r due to the spread in energy of the proton beam. (See text, page 54.) -177- t 68124,10-B 68124,3-A ~ U) a:: 100 w 0.. 200 en ...... z 66044,8-A :::> 0 u 66044,8-B 65315,6 (interior) 980 + 990 1000 PROTON ENERGY FIGURE 22 1030 (See text, ~age 55.) Only sampl e error bars are shown. utive runs are plotted together. surface A, data from two consec- ~or data respectively for surface B. Solid circles and open circles represent "first run" and "second run" Fluorine concentration versus depth for lunar glass spherule 68124~3. FIGURE 23 I c:> I I-' --l -17 9- wdd .::J 0 0 0 0 0 l{) 0 0 0 l{) • 0 \• I <( a:l LIJ LIJ q- ~ ~ f'() ... (.) \I • !/I • (.) C\J 0:: 0:: CJ) (/) - ::> ::> CX) . (.0 • • 0 •I • I I /' 0 C\J ~~-/ 0 \ ~tef·~· ~. ----. 0 -0 •' \0 I 0 C\J ~ 0 l{) 0 - :1.. Ow /. / / tef • 0 ·E I ~t. a_ • • / 0 <D- • 0 CD 0 {£W~/oz01) NOLL'J~.LN3JN0J .:J 01 (\J ·~ r; :-:J c· H f; I -1130- F'FiURE 24 Fused silica simulation experiment results. Data sho•m are representative of a set of swnples subjected first to radiation damage and H2 o exposure tests. Two of the samples vere damar;ed bv irradiatinG them ,.,i th 86-keV l<lo- ions for 4 hours to a tota.l dose of l. 4 x 10 17 2 ions/ cm • One of these (solid circles) vras subsequently exposed to H2 o in both liquid (submer,?;ed in distilled water for 24 hr) and vapor (laboratory atmosphere for one Heelc) form, ,.,bile the other (solid triangles) ,.,as ex:noned only to dry N gas for 2 hr. 2 A third sample (open circles) 1vas not radi- ation damaged but was given the same H2 o ex:nosure an the first sample. Onl;r sample error bars are shmm on the data 110ints obtained during the subse0uent H analysis, performed to determine the extent of H,..,O penetration. L 'I'he solid curve represents t;r:nical results for a clean fused silica sample 1-rith a normal (for this hatch) H content of -v?.O p:nm. (Sec text, 11n.gP. )'r.) -1131- 3 I -E Jt) ~ V) E - 0 •- damaged, exposed to liquid H20 o-undamaged, exposed to liquid H20 •-damaged, no H20 exposure • • 0 c FUSED SILICA 2 N 200 150 0 • • z 0 ..... 0 <t a:: ! I E 100 a. a. • ..... z •o w (.) z 0 0 (.) • 50 Bo I - 0 0 ... t 0 0 1000 2000 3000 DEPTH (ANGSTROMS) F IGURE 2 4 Straggling is neglected, so each The distributions shown are for normal incidence (HH(x)), an page 74 and Annendix E, paGe 110.) solar direction (H8 (x)); 1-rhere x is the depth in t!'le sample. (See text, isotropic angular distribution (H1 (x)), and protons incident fro~ the direction. the mean nrojected range of protons of energy E0 ) along its incident incident proton is assumed to penetrate a distance R0 (corresponding to to a constant atomic erosion rate v. monoenergetic proton source of flux ~ 0 for lunar surface samples subject Equilibriun H distribution functions resulting from implantation of a FIGURE 25 I i0 I c::- 1---' - 183- 0 I I I I I . . ... 01 ~ -:r: > . .. X z co 0 0 ~a:: ......·~,~ Q ..·· \~ . ........ .··X'~ ~----~~--------~--------~~------~~~------~~---------~() 0 0 0 () X FIGURI.!; 26 " Suprathermal" proton flux spectrw::1 d<j>/dE ver sus proton cnere;;r E. Data points with associated error bars are taken from satellite observations reported by Frank (1970). A spectrUJ:l adjusted to g ive a rough fit to the observed H distribution in 6881 ) , 27 , 0 assuming an atomic erosion rate of 0. 5 A/yr , is indicat ed 'o ;' the solid lines. ( See text, page 71+.) -185- 11 -1 - 11 SUPRATHERMAL PROTONS co 0 I -> 0 I z Q) ~ I ~ 0 a::: I ~ Q) CJ) I (\1 ....J 0 (/) I E 0 I CJ) c: 0 0 '- a. w ~ -e.. "0 103~----~~--~~~~~--~~~~~~~WU~--~ 1.0 10 PROTON 100 ENERGY (keV) FIGURE 26 -186- FIGURE 27 Ir.1plantation of solar protons in lunar sanples. are from sample 68815,27 (Figure 18). 'l'he data points The solid curve is the distribution resulting from the flux spectrum indicated by the solid lines in Figure 26, assuming an atomic erosion rate of 0 0.5 A/yr, and calculated using equations E.8 and E.9 of Appendix E. The spectrum "'as chosen to give a rough fit to the data, usinc a proton range-energy relation derived from Schi~tt (196G) and neglecting range straggling and diffusion. 'i'he dashed curve indicates the limit of penetration of the present-day solar '•ind, including the effects of range strage;ling. Hith no diffusive losses, the peak H content at the surface would be c:reater than 10 23 H atoms/cm3 -- more than bm orders of magnitude higher than the observed H content near the surface of sample 68815,27. text, page 75. ) (See -1137- IMPLANTATION SOLAR - 10 OF PROTONS i- 68815,27 ro E (.) 'E 8 (/) I ~ - I ~ 0 0 0 C\J 0 z 6 I I I I I I I ~ )7,.-\ ~ !\ 0 ._ I <l lo 'r ._a:: 4 I)> z 1::0 (_) l:E w I ,,z z 0 0 r-~ IU> 2 -I ,o ' ' \ \ \ 00 1000 2000 DEPTH 3000 4000 (ANGSTROMS) Fir.URE 21 I -188- FIGURE 28 Data :from two obsidian samples with thin hydration laver3. Sample 2 (a) vas artificially hydrated at 75°C :for 2 days. Sample 4 (b) uas hydrated in a normal l aboratory environment (room temperature) :for 10 years. Data are y-ray counts per 3 ]JC o:f l9F4+ plotted against the incident 19F ion enerp;~r E 0 • The depth scale is calculated :from the stopping rm·rer of the obsidian :for 1 9F ions, w·i th the zero point corresponding to the resonance energ:-r (16.1~5 MeV). 'l'he data impl~r ]leak H 0 contents 2 o:f "-2% by veie;ht, ,,,i th hydration layer thicknesses of 0.11 and 0.19 ]Jm :for samples 2 and 4, respectively. page 86.) (See text, lJID -189- Eo (MeV)' 17 16 (a) 500 + <;j- SAMPLE 2 DATA 16 (b) 17 18 SAMPLE 4 DATA 400 LL CJ) (_) ::t 300 ('() '-.... U) 1--- z :J 0 200 (_) 0 0.5 DEPTH (fLm) FIGURE 28 0 0 .5 -190- FIGURE 29 Unfolding procedure for obsidian sample number 5. In (a) raw data are ~lotted for sample 5 and for a freshly exposed, unhydrated interior sample (0.3% H 0). 2 The hydration profile in (b) is obtained by subtracting the counting rate for the unhydrated sample (dashed curve) from the data for sample 5 at each enerh~, and then performing the unfolding procedure described in the text (Section II, Part A) in vrhich the y-ray counts due to the resonance at 17.64 ''l(eV are subtracted, leaving only the yield due to the resonance at 16.45 MeV. The resulting reduced data are then fitted vrith an H2o content scale (the 0.3% H o is added back in 2 the placement of the zero point of this scale) and a depth scale using the calculated stopping pouer of obsidian for 1 9F ions. (See text, page 87.) -191- Eo (MeV) 800,_ + - o;;;t ! I..L (j) SAMPLE 5 !--+ UNHYDRATED _ (.3% H 2 0) SAMPLE (_) 600,_ ::i. r0 ........... (j) I- 400f- z => 0 (_) - 200f- l 0 3 N :r: (b) _l J SAMPLE 5 HYDRATION PROFILE 1- z w 2 (_) 0:: w Q_ II <..9 w 3: F'IC";URE 29 1 -192- FIGURE 30 Hy dration profiles for obsidian samples 9 (a) and 1 (b). 'J'he se nrofiles were obtained in the same manner as shown i n Fi Gure 29 f or sample ·5 . Error bars ·are not shown, but a re of c o1:1parabl e magnitude to those in Figure 29 (b) (la statistical uncertaint ies) . Samples 7 and 9 both app ear to have ~ :plateau l ev els in the i r hydration ~rofiles. (See text, uage 08.) -193- DEPTH (f-Lm) 0 .5 0 (a) 0 1.0 1.5 2 .0 SAMPLE 9 HYDRATION PROFILE 3 N I 1-- z w u 0::: 2 w o._ I- I 0 w 3 (b) 0 SAMPLE 7 HYDRATION PROFILE 3 N I 1-- z w u 0::: w 2 o._ I- I 0 w 3 ot~~~~·~'_J·-~~~~~~~~~~~~~ 0 0 .5 - 1.0 DEPTH (f-Lm) fo' Ir";URE 30 1.5 2 .0