- Email: [email protected]
Hydrogen depth profiling by p–p scattering in nominally anhydrous minerals
Nuclear Instruments and Methods in Physics Research B 231 (2005) 524–529 www.elsevier.com/locate/nimb
Hydrogen depth profiling by p–p scattering in nominally anhydrous minerals M. Wegde´n a,*, P. Kristiansson a, H. Skogby b, V. Auzelyte a, M. Elfman a, K.G. Malmqvist a, C. Nilsson a, J. Pallon a, A. Shariff a a
Department of Nuclear Physics, Lund Institute of Technology, Lund University, P.O. Box 118, Solvegatan 14, S-221 00 Lund, Sweden b Swedish Museum of Natural History, Department of Mineralogy, P.O. Box 50007, S-104 05 Stockholm, Sweden Available online 16 March 2005
Abstract Hydrogen has been shown to occur as a trace element in many nominally anhydrous minerals. The presence of hydrogen in several of the major minerals in the EarthÕs mantle has received attention, due to the possibility that these phases provide a significant hydrogen reservoir. Recently an experimental and analytical procedure for hydrogen measurements in thin mineral samples by proton–proton scattering has been developed at the Lund Nuclear Microprobe facility. An annular surface barrier detector, divided in two insulated halves, is used to detect the scattered proton and the recoiled proton in coincidence. The summed energy of each detected proton pair can be used to produce depth profiles if the individual scattering angles are known. The easiest case is when only a small difference in energy between each detected proton pair is allowed, i.e. scattering angles very close to 45. This limitation criterion considerably reduces the statistics. For this reason the analytical method has been expanded to use the full detector area (35–55) and to identify the scattering angles individually for each hydrogen event. Nominally anhydrous minerals, both synthesized and of natural occurrence, with hydrogen concentrations from 10 to 100 ppm have been analysed. Hydrous minerals, as well as Mylar foils were used as standards. Depth profiles show that intrinsic hydrogen can be distinguished from surface contaminations, e.g. water adsorbed on the sample surfaces. 2005 Elsevier B.V. All rights reserved. PACS: 25.40.Cm; 29.30.Ep; 29.90.+r Keywords: Anhydrous minerals; Depth profiling; Hydrogen analysis; p–p scattering
1. Introduction *
Corresponding author. Tel.: +46 46 222 76 35; fax: +46 46 222 47 09/03 44. E-mail address: [email protected] (M. Wegde´n).
In recent years it has come to knowledge that several nominally anhydrous minerals (NAM) in the EarthÕs mantle store low but significant
0168-583X/$ - see front matter 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.nimb.2005.01.111
M. Wegde´n et al. / Nucl. Instr. and Meth. in Phys. Res. B 231 (2005) 524–529
concentrations of hydrogen (usually not exceeding a few hundred wt ppm H2O), structurally bound as hydroxyl ions. Hydrogen in upper mantle minerals (olivine, garnet, ortho- and clinopyroxene) is of specific interest, since they provide a mantle-water storage mechanism. Due to the vast dimension of the upper mantle (below the EarthÕs crust down to approximately 410 km), the total amount of water stored in these phases seems comparable to the amount being present in the oceans [1–3]. Much of the understanding of the evolution of the mantle and the internal water cycle of the Earth comes from studies performed in the past ten years, but the knowledge is still limited. Quantification of hydrogen concentrations in mantle minerals, combined with studies on dynamic processes in the mantle (e.g. diffusion, hydrogen exchange reactions and solubility) is a step towards more detailed comprehension of this complex system. The main analytical method for studies of hydrogen in nominally anhydrous minerals has been Fourier transform infrared spectroscopy (FTIR). This method has high sensitivity and yields information about the orientation of the OH ion in the crystal structure, and can thereby distinguish structurally incorporated hydrogen from adsorbed water, fluid inclusions etc. However, for quantitative hydrogen analysis FTIR requires calibration against an independent complementary technique and can for this purpose be combined with e.g. an ion beam technique [4]. When it comes to ion beam analysis, elastic proton–proton scattering [5] is considered to be the most suitable method for trace concentrations of hydrogen with micrometer depth resolution [6]. The method has high cross section and, compared to other available methods e.g. Energy Recoil Detection (ERD) or Nuclear Reaction Analysis (NRA), the lowest possible irradiation damage effects [7–9]. These factors are crucial to get the sensitivity required for hydrogen analysis in nominally anhydrous minerals, without significantly changing the hydrogen content in the investigated area. Recently a procedure for hydrogen analysis and depth profiling has been developed at the Lund Nuclear Microprobe [10]. An annular detector with a large solid angle, divided into two insulated halves, is used to detect the scattered and the recoiled proton in coincidence. The analytical proce-
525
dure for depth profiling was previously limited to accept individual scattering angles close to 45 only. This limitation criterion considerably reduces the statistics. For this reason the procedure has been modified to accept all scattering angles covered by the detector (approximately 35–55). A series of thin geological samples, both hydrous and nominally anhydrous, have been analysed, and the depth profiles were used to distinguish intrinsic hydrogen from surface contaminations.
2. Sample preparation The scattering geometry and the limited range of the 2.8 MeV protons restricted the analysis to thin self-supported samples. Millimetre-sized grains of the mineral were crystallographic oriented and then founded in plastic. One surface was polished at a time until the sample thickness was 10–15 lm. This was difficult to achieve, as the samples were brittle and often started to break around 15–20 lm. The remaining plastic was dissolved with acetone and the resulting plane parallel crystal was transferred and attached to an aperture. A selection of hydrous and nominally anhydrous natural minerals, as well as a synthesized sample with zonations, was included in the investigation. The synthesized sample was produced by a flux-growth method using KVO3 as flux compound, which involved slow cooling of a saturated melt from 1200 to 800 C. Hydrogen was afterwards diffused into the Fe3+-containing sample by heating in hydrogen atmosphere at 800 C, following the redox reaction Fe3+ + O2 + 1/2H2 = Fe2+ + OH [11]. The synthesized sample was cut and polished along the diffusion direction. Mylar foils, C10H8O4, of different thicknesses (single or multiple layers) were also mounted on sample holders for use as standard samples.
3. Experimental method and data analysis The experimental work was conducted at the Lund Nuclear Microprobe, with a single-ended NEC 3 UH accelerator. A beam (5–10 lm) of 2.8 MeV protons in normal incidence was scanned
526
M. Wegde´n et al. / Nucl. Instr. and Meth. in Phys. Res. B 231 (2005) 524–529
over the sample, with a current of 200 pA. The special surface barrier detector, described in [10,12] was positioned in the forward direction to detect the scattered proton and the recoiling target hydrogen nucleus in coincidence for effective background suppression. There is no need for the detector to cover a particular angle, since the cross section for p–p scattering is approximately constant down to an angle of 15 in the forward direction [12,13], but it is desirable that as large angle interval as possible is covered to speed up the analysis and reduce irradiation damage. Our detector covers scattering angles from 35 to 55 relative to the sample normal when placed in an ideal position. Since the hydrogen concentrations are not determined by direct measurement, but relative homogeneous standard targets, the detector alignment is not too critical as long as its position is the same during both standard and sample analysis. Events from the two detector halves were read out simultaneously. The maximum allowed count rate was 20 kHz in each half, to reduce pile-up and a too high amount of random events. The energy signal from each detector half was amplified and the elastic scattering events were identified and picked out with a SCA (Single Channel Analyser). Both branches of the signal were then tested for coincidence within a 200 ns time window, and the relative detection time was measured with a time to pulse-height converter (TPHC). Time and energy spectra were acquired with peak-sensitive ADCÕs and recorded in list mode. In addition, the beam charge was measured by a Faraday cup and off-axis STIM was conducted simultaneously with the hydrogen analysis for thickness determination and control of the sample homogeneity. The mineral samples required data collection times of 1–2 h for good statistics and more than 6 h was spent on the synthesized sample. The experimental data was later analysed with the Kmax software package, Sparrow Corporation [14]. The signal to background ratio could be enhanced, by putting restraints on the coincidence time (typically <20 ns). From the time spectra it was deduced that accidental coincidences contributed to a background of approximately 1 wt ppm. Depth information was obtained from the total
energy loss in the sample and for this purpose the sum of the kinetic energy of every coincident proton pair was determined. Scattering events that originate from the exit surface undergo the smallest energy loss. The energy loss increases with depth (seen from the exit surface) because of the increased path length of the scattered protons and their sharing of energy. Due to the difficulty to thin down the samples to less than 10–15 lm, proton pairs from the entrance surface have too short range and will not reach the detector. Thus depth profiling could only be performed to a limited depth. Since there was no direct information about the individual scattering angles, except that two protons are always emitted in the forward direction with a relative angle of 90, an indirect approach had to be used to determine the depth where the p–p scattering took place. One alternative, which simplifies the depth calculations, is to accept no, or only a small energy difference between two detected protons, i.e. individual scattering angles of 45 [10]. Under this condition, the proton pair equally divides the available energy of the incident proton and both protons travel the same path length through the sample to the detector. This limitation criterion however considerably reduces the statistics. In this work, the scattering angles and the depth location for every hydrogen event were identified from the summed energy combined with the energy difference, and thus all angles covered by the detector could be used. The hydrogen concentrations were determined relative to the homogeneous Mylar samples (4.17 wt% H). Mylar foils of the same mass thickness as the geological samples were chosen, and thus no knowledge about the sample constituents was necessary. The depth profiles were derived from tabulated stopping data of protons in Mylar [15] with an expected depth resolution of 5–10% of the sample thickness [12,16].
4. Results Depth profiles for three minerals with varying hydrogen concentrations are presented with micrometer resolution in Fig. 1(a)–(c). All samples were too thick for depth profiling through the
M. Wegde´n et al. / Nucl. Instr. and Meth. in Phys. Res. B 231 (2005) 524–529 4500
(a)
4000
Amphibole
H conc. (wt-ppm)
3500 3000 2500 2000 1500 1000 500 0 0
0.2 0.4 0.6 0.8
1
1.2 1.4 1.6 1.8
Depth (mg/cm2) 450 400
(b)
Orthopyroxene
H conc. (wt-ppm)
350 300 250 200 150 100 50 0 0
0.5
1
1.5
2
2.5
Depth (mg/cm2) 450
(c)
400
Diopside
H conc. (wt-ppm)
350 300
527
hydrogen, except in the hydrous sample because of its high bulk concentration. The bulk concentrations were determined as the weighted mean for every mineral, starting at some distance from the surface peak. Error bars in the figures only include statistical errors in the data points. The bulk concentrations range from 25 wt ppm hydrogen in orthopyroxene (NAM) (b) to 1855 wt ppm in amphibole (a). Fig. 2(a) shows a 2D image, 512 lm · 512 lm, of the bulk hydrogen events in an inhomogeneous synthesized sample, with horizontal diffusion direction. Two larger regions have been selected in Fig. 2(a) and their hydrogen concentrations are presented as depth profiles in Fig. 2(c), averaged over the separate regions. Fig. 2(a) shows a higher hydrogen concentration in the left part (region 1) of the sample, which is clearly visible in Fig. 2(b) where all hydrogen events have been projected onto the x-axis. This can also be observed when comparing depth profiles of the two selected regions. This exemplifies the possibility to perform 3D hydrogen microscopy with micrometer depth resolution also at the ppm level. However, 3D profiling requires good statistics and thus very long collection times. Hence there is a delicate balance with time and radiation damage effects on one side and spatial resolution and detection limits on the other side.
250 200
5. Summary
150 100 50 0 0
0.5
1
1.5
2
2.5
Depth (mg/cm2)
Fig. 1. Depth profiles for three natural minerals with different hydrogen concentrations. The intrinsic hydrogen can easily be distinguished from the surface contamination, and the weighted mean bulk concentration is given for every mineral: (a) amphibole (hydrous mineral from Austria) 1855 ± 10 wt ppm; (b) orthopyroxene (nominally anhydrous mantle mineral from Kilbourne Hole, New Mexico) 25 ± 1 wt ppm; (c) diopside (NAM), 58 ± 1 wt ppm.
whole sample thickness. The bulk hydrogen can easily be distinguished from surface adsorbed
An analytical procedure has been developed for hydrogen depth profiling by coincident proton– proton scattering. The summed energy of every detected proton pair and their difference in energy is used in an indirect approach to determine the depth location for every hydrogen event. In this fashion, the full detector area (35–55) can be used for the analysis, and restraints on the scattering angles are not necessary. The procedure has been applied to geological samples to determine bulk concentrations of hydrogen in both hydrous and nominally anhydrous minerals. The depth profiles are used to distinguish surface adsorbed hydrogen from bulk hydrogen with micrometer depth resolution.
528
M. Wegde´n et al. / Nucl. Instr. and Meth. in Phys. Res. B 231 (2005) 524–529
Fig. 2. (a) A 2D image, 512 lm · 512 lm, of hydrogen events inside the bulk of a synthesized sample with zonations. The left part of the scanned sample area shows a higher hydrogen concentration, which can also be observed in (b) where the number of hydrogen events has been projected onto the x-axis; (c) shows the depth profiles, averaged over the two selected regions in (a).
Acknowledgment The work was financially supported by the Swedish Research Council for natural and engineering sciences.
References [1] H. Skogby, NATO Adv. Sci. Instr. Ser. C, Math. Phys. Sci. 543 (1999) 509. [2] J. Ingrin, H. Skogby, Eur. J. Mineral. 12 (2000) 543. [3] G.R. Rossman, Phys. Chem. Minerals 23 (1996) 299.
[4] G.R. Rossman, Nucl. Instr. and Meth. B 45 (1990) 41. [5] B.L. Cohen, C.L. Fink, J.H. Degnan, J. Appl. Phys. 43 (1972) 19. [6] G. Dollinger, P. Reichart, G. Datzmann, A. Hauptner, H.-J. Ko¨rner, Appl. Phys. Lett. 82 (2003) 148. [7] P. Reichart, G. Dollinger, A. Bergmaier, G. Datzmann, A. Hauptner, H.-J. Ko¨rner, Nucl. Instr. and Meth. B 197 (2002) 134. [8] P. Reichart, G. Dollinger, G. Datzmann, A. Hauptner, R. Hertenberger, H.-J. Ko¨rner, Nucl. Instr. and Meth. B 210 (2003) 135. [9] W.A. Lanford, Nucl. Instr. and Meth. B 66 (1992) 65. [10] M. Wegde´n, P. Kristiansson, Z. Pastuovic, H. Skogby, V. Auzelyte, M. Elfman, K.G. Malmqvist, C. Nilsson, J. Pallon, A. Shariff, Nucl. Instr. and Meth. B 219–220 (2004) 550.
M. Wegde´n et al. / Nucl. Instr. and Meth. in Phys. Res. B 231 (2005) 524–529 [11] H. Skogby, Amer. Mineral. 79 (1994) 240. [12] K.A. Sjo¨land, P. Kristiansson, M. Elfman, K.G. Malmqvist, J. Pallon, R.J. Utui, C. Yang, Nucl. Instr. and Meth. B 124 (1997) 639. [13] D.J. Knecht, S. Messelt, E.D. Berners, L.C. Northcliffe, Phys. Rev. 114 (1959) 550.
529
[14] Sparrow Corporation. Available from:. [15] J.F. Ziegler, SRIM 2003, Version-2003.20. Available from: . [16] J.F. Ziegler et al., Nucl. Instr. and Meth. 149 (1978) 19.
Hydrogen depth profiling by p–p scattering in nominally anhydrous minerals M. Wegde´n a,*, P. Kristiansson a, H. Skogby b, V. Auzelyte a, M. Elfman a, K.G. Malmqvist a, C. Nilsson a, J. Pallon a, A. Shariff a a
Department of Nuclear Physics, Lund Institute of Technology, Lund University, P.O. Box 118, Solvegatan 14, S-221 00 Lund, Sweden b Swedish Museum of Natural History, Department of Mineralogy, P.O. Box 50007, S-104 05 Stockholm, Sweden Available online 16 March 2005
Abstract Hydrogen has been shown to occur as a trace element in many nominally anhydrous minerals. The presence of hydrogen in several of the major minerals in the EarthÕs mantle has received attention, due to the possibility that these phases provide a significant hydrogen reservoir. Recently an experimental and analytical procedure for hydrogen measurements in thin mineral samples by proton–proton scattering has been developed at the Lund Nuclear Microprobe facility. An annular surface barrier detector, divided in two insulated halves, is used to detect the scattered proton and the recoiled proton in coincidence. The summed energy of each detected proton pair can be used to produce depth profiles if the individual scattering angles are known. The easiest case is when only a small difference in energy between each detected proton pair is allowed, i.e. scattering angles very close to 45. This limitation criterion considerably reduces the statistics. For this reason the analytical method has been expanded to use the full detector area (35–55) and to identify the scattering angles individually for each hydrogen event. Nominally anhydrous minerals, both synthesized and of natural occurrence, with hydrogen concentrations from 10 to 100 ppm have been analysed. Hydrous minerals, as well as Mylar foils were used as standards. Depth profiles show that intrinsic hydrogen can be distinguished from surface contaminations, e.g. water adsorbed on the sample surfaces. 2005 Elsevier B.V. All rights reserved. PACS: 25.40.Cm; 29.30.Ep; 29.90.+r Keywords: Anhydrous minerals; Depth profiling; Hydrogen analysis; p–p scattering
1. Introduction *
Corresponding author. Tel.: +46 46 222 76 35; fax: +46 46 222 47 09/03 44. E-mail address: [email protected] (M. Wegde´n).
In recent years it has come to knowledge that several nominally anhydrous minerals (NAM) in the EarthÕs mantle store low but significant
0168-583X/$ - see front matter 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.nimb.2005.01.111
M. Wegde´n et al. / Nucl. Instr. and Meth. in Phys. Res. B 231 (2005) 524–529
concentrations of hydrogen (usually not exceeding a few hundred wt ppm H2O), structurally bound as hydroxyl ions. Hydrogen in upper mantle minerals (olivine, garnet, ortho- and clinopyroxene) is of specific interest, since they provide a mantle-water storage mechanism. Due to the vast dimension of the upper mantle (below the EarthÕs crust down to approximately 410 km), the total amount of water stored in these phases seems comparable to the amount being present in the oceans [1–3]. Much of the understanding of the evolution of the mantle and the internal water cycle of the Earth comes from studies performed in the past ten years, but the knowledge is still limited. Quantification of hydrogen concentrations in mantle minerals, combined with studies on dynamic processes in the mantle (e.g. diffusion, hydrogen exchange reactions and solubility) is a step towards more detailed comprehension of this complex system. The main analytical method for studies of hydrogen in nominally anhydrous minerals has been Fourier transform infrared spectroscopy (FTIR). This method has high sensitivity and yields information about the orientation of the OH ion in the crystal structure, and can thereby distinguish structurally incorporated hydrogen from adsorbed water, fluid inclusions etc. However, for quantitative hydrogen analysis FTIR requires calibration against an independent complementary technique and can for this purpose be combined with e.g. an ion beam technique [4]. When it comes to ion beam analysis, elastic proton–proton scattering [5] is considered to be the most suitable method for trace concentrations of hydrogen with micrometer depth resolution [6]. The method has high cross section and, compared to other available methods e.g. Energy Recoil Detection (ERD) or Nuclear Reaction Analysis (NRA), the lowest possible irradiation damage effects [7–9]. These factors are crucial to get the sensitivity required for hydrogen analysis in nominally anhydrous minerals, without significantly changing the hydrogen content in the investigated area. Recently a procedure for hydrogen analysis and depth profiling has been developed at the Lund Nuclear Microprobe [10]. An annular detector with a large solid angle, divided into two insulated halves, is used to detect the scattered and the recoiled proton in coincidence. The analytical proce-
525
dure for depth profiling was previously limited to accept individual scattering angles close to 45 only. This limitation criterion considerably reduces the statistics. For this reason the procedure has been modified to accept all scattering angles covered by the detector (approximately 35–55). A series of thin geological samples, both hydrous and nominally anhydrous, have been analysed, and the depth profiles were used to distinguish intrinsic hydrogen from surface contaminations.
2. Sample preparation The scattering geometry and the limited range of the 2.8 MeV protons restricted the analysis to thin self-supported samples. Millimetre-sized grains of the mineral were crystallographic oriented and then founded in plastic. One surface was polished at a time until the sample thickness was 10–15 lm. This was difficult to achieve, as the samples were brittle and often started to break around 15–20 lm. The remaining plastic was dissolved with acetone and the resulting plane parallel crystal was transferred and attached to an aperture. A selection of hydrous and nominally anhydrous natural minerals, as well as a synthesized sample with zonations, was included in the investigation. The synthesized sample was produced by a flux-growth method using KVO3 as flux compound, which involved slow cooling of a saturated melt from 1200 to 800 C. Hydrogen was afterwards diffused into the Fe3+-containing sample by heating in hydrogen atmosphere at 800 C, following the redox reaction Fe3+ + O2 + 1/2H2 = Fe2+ + OH [11]. The synthesized sample was cut and polished along the diffusion direction. Mylar foils, C10H8O4, of different thicknesses (single or multiple layers) were also mounted on sample holders for use as standard samples.
3. Experimental method and data analysis The experimental work was conducted at the Lund Nuclear Microprobe, with a single-ended NEC 3 UH accelerator. A beam (5–10 lm) of 2.8 MeV protons in normal incidence was scanned
526
M. Wegde´n et al. / Nucl. Instr. and Meth. in Phys. Res. B 231 (2005) 524–529
over the sample, with a current of 200 pA. The special surface barrier detector, described in [10,12] was positioned in the forward direction to detect the scattered proton and the recoiling target hydrogen nucleus in coincidence for effective background suppression. There is no need for the detector to cover a particular angle, since the cross section for p–p scattering is approximately constant down to an angle of 15 in the forward direction [12,13], but it is desirable that as large angle interval as possible is covered to speed up the analysis and reduce irradiation damage. Our detector covers scattering angles from 35 to 55 relative to the sample normal when placed in an ideal position. Since the hydrogen concentrations are not determined by direct measurement, but relative homogeneous standard targets, the detector alignment is not too critical as long as its position is the same during both standard and sample analysis. Events from the two detector halves were read out simultaneously. The maximum allowed count rate was 20 kHz in each half, to reduce pile-up and a too high amount of random events. The energy signal from each detector half was amplified and the elastic scattering events were identified and picked out with a SCA (Single Channel Analyser). Both branches of the signal were then tested for coincidence within a 200 ns time window, and the relative detection time was measured with a time to pulse-height converter (TPHC). Time and energy spectra were acquired with peak-sensitive ADCÕs and recorded in list mode. In addition, the beam charge was measured by a Faraday cup and off-axis STIM was conducted simultaneously with the hydrogen analysis for thickness determination and control of the sample homogeneity. The mineral samples required data collection times of 1–2 h for good statistics and more than 6 h was spent on the synthesized sample. The experimental data was later analysed with the Kmax software package, Sparrow Corporation [14]. The signal to background ratio could be enhanced, by putting restraints on the coincidence time (typically <20 ns). From the time spectra it was deduced that accidental coincidences contributed to a background of approximately 1 wt ppm. Depth information was obtained from the total
energy loss in the sample and for this purpose the sum of the kinetic energy of every coincident proton pair was determined. Scattering events that originate from the exit surface undergo the smallest energy loss. The energy loss increases with depth (seen from the exit surface) because of the increased path length of the scattered protons and their sharing of energy. Due to the difficulty to thin down the samples to less than 10–15 lm, proton pairs from the entrance surface have too short range and will not reach the detector. Thus depth profiling could only be performed to a limited depth. Since there was no direct information about the individual scattering angles, except that two protons are always emitted in the forward direction with a relative angle of 90, an indirect approach had to be used to determine the depth where the p–p scattering took place. One alternative, which simplifies the depth calculations, is to accept no, or only a small energy difference between two detected protons, i.e. individual scattering angles of 45 [10]. Under this condition, the proton pair equally divides the available energy of the incident proton and both protons travel the same path length through the sample to the detector. This limitation criterion however considerably reduces the statistics. In this work, the scattering angles and the depth location for every hydrogen event were identified from the summed energy combined with the energy difference, and thus all angles covered by the detector could be used. The hydrogen concentrations were determined relative to the homogeneous Mylar samples (4.17 wt% H). Mylar foils of the same mass thickness as the geological samples were chosen, and thus no knowledge about the sample constituents was necessary. The depth profiles were derived from tabulated stopping data of protons in Mylar [15] with an expected depth resolution of 5–10% of the sample thickness [12,16].
4. Results Depth profiles for three minerals with varying hydrogen concentrations are presented with micrometer resolution in Fig. 1(a)–(c). All samples were too thick for depth profiling through the
M. Wegde´n et al. / Nucl. Instr. and Meth. in Phys. Res. B 231 (2005) 524–529 4500
(a)
4000
Amphibole
H conc. (wt-ppm)
3500 3000 2500 2000 1500 1000 500 0 0
0.2 0.4 0.6 0.8
1
1.2 1.4 1.6 1.8
Depth (mg/cm2) 450 400
(b)
Orthopyroxene
H conc. (wt-ppm)
350 300 250 200 150 100 50 0 0
0.5
1
1.5
2
2.5
Depth (mg/cm2) 450
(c)
400
Diopside
H conc. (wt-ppm)
350 300
527
hydrogen, except in the hydrous sample because of its high bulk concentration. The bulk concentrations were determined as the weighted mean for every mineral, starting at some distance from the surface peak. Error bars in the figures only include statistical errors in the data points. The bulk concentrations range from 25 wt ppm hydrogen in orthopyroxene (NAM) (b) to 1855 wt ppm in amphibole (a). Fig. 2(a) shows a 2D image, 512 lm · 512 lm, of the bulk hydrogen events in an inhomogeneous synthesized sample, with horizontal diffusion direction. Two larger regions have been selected in Fig. 2(a) and their hydrogen concentrations are presented as depth profiles in Fig. 2(c), averaged over the separate regions. Fig. 2(a) shows a higher hydrogen concentration in the left part (region 1) of the sample, which is clearly visible in Fig. 2(b) where all hydrogen events have been projected onto the x-axis. This can also be observed when comparing depth profiles of the two selected regions. This exemplifies the possibility to perform 3D hydrogen microscopy with micrometer depth resolution also at the ppm level. However, 3D profiling requires good statistics and thus very long collection times. Hence there is a delicate balance with time and radiation damage effects on one side and spatial resolution and detection limits on the other side.
250 200
5. Summary
150 100 50 0 0
0.5
1
1.5
2
2.5
Depth (mg/cm2)
Fig. 1. Depth profiles for three natural minerals with different hydrogen concentrations. The intrinsic hydrogen can easily be distinguished from the surface contamination, and the weighted mean bulk concentration is given for every mineral: (a) amphibole (hydrous mineral from Austria) 1855 ± 10 wt ppm; (b) orthopyroxene (nominally anhydrous mantle mineral from Kilbourne Hole, New Mexico) 25 ± 1 wt ppm; (c) diopside (NAM), 58 ± 1 wt ppm.
whole sample thickness. The bulk hydrogen can easily be distinguished from surface adsorbed
An analytical procedure has been developed for hydrogen depth profiling by coincident proton– proton scattering. The summed energy of every detected proton pair and their difference in energy is used in an indirect approach to determine the depth location for every hydrogen event. In this fashion, the full detector area (35–55) can be used for the analysis, and restraints on the scattering angles are not necessary. The procedure has been applied to geological samples to determine bulk concentrations of hydrogen in both hydrous and nominally anhydrous minerals. The depth profiles are used to distinguish surface adsorbed hydrogen from bulk hydrogen with micrometer depth resolution.
528
M. Wegde´n et al. / Nucl. Instr. and Meth. in Phys. Res. B 231 (2005) 524–529
Fig. 2. (a) A 2D image, 512 lm · 512 lm, of hydrogen events inside the bulk of a synthesized sample with zonations. The left part of the scanned sample area shows a higher hydrogen concentration, which can also be observed in (b) where the number of hydrogen events has been projected onto the x-axis; (c) shows the depth profiles, averaged over the two selected regions in (a).
Acknowledgment The work was financially supported by the Swedish Research Council for natural and engineering sciences.
References [1] H. Skogby, NATO Adv. Sci. Instr. Ser. C, Math. Phys. Sci. 543 (1999) 509. [2] J. Ingrin, H. Skogby, Eur. J. Mineral. 12 (2000) 543. [3] G.R. Rossman, Phys. Chem. Minerals 23 (1996) 299.
[4] G.R. Rossman, Nucl. Instr. and Meth. B 45 (1990) 41. [5] B.L. Cohen, C.L. Fink, J.H. Degnan, J. Appl. Phys. 43 (1972) 19. [6] G. Dollinger, P. Reichart, G. Datzmann, A. Hauptner, H.-J. Ko¨rner, Appl. Phys. Lett. 82 (2003) 148. [7] P. Reichart, G. Dollinger, A. Bergmaier, G. Datzmann, A. Hauptner, H.-J. Ko¨rner, Nucl. Instr. and Meth. B 197 (2002) 134. [8] P. Reichart, G. Dollinger, G. Datzmann, A. Hauptner, R. Hertenberger, H.-J. Ko¨rner, Nucl. Instr. and Meth. B 210 (2003) 135. [9] W.A. Lanford, Nucl. Instr. and Meth. B 66 (1992) 65. [10] M. Wegde´n, P. Kristiansson, Z. Pastuovic, H. Skogby, V. Auzelyte, M. Elfman, K.G. Malmqvist, C. Nilsson, J. Pallon, A. Shariff, Nucl. Instr. and Meth. B 219–220 (2004) 550.
M. Wegde´n et al. / Nucl. Instr. and Meth. in Phys. Res. B 231 (2005) 524–529 [11] H. Skogby, Amer. Mineral. 79 (1994) 240. [12] K.A. Sjo¨land, P. Kristiansson, M. Elfman, K.G. Malmqvist, J. Pallon, R.J. Utui, C. Yang, Nucl. Instr. and Meth. B 124 (1997) 639. [13] D.J. Knecht, S. Messelt, E.D. Berners, L.C. Northcliffe, Phys. Rev. 114 (1959) 550.
529
[14] Sparrow Corporation. Available from: