Radon emanation from brittle fracturing in ... - Wiley Online Library

3 downloads 23 Views 4MB Size Report
Aug 4, 2014 - The radioactive gas radon-222 (half-life 3.82 days) is naturally produced in ..... dioxide and radon-222 discharge in Central Nepal, Earth Planet.
PUBLICATIONS Geophysical Research Letters RESEARCH LETTER 10.1002/2014GL061095 Key Points: • Mechanical and thermal damage effects on radon release are studied in granites • Measurements of radon concentration in pores are coupled with a triaxial cell • Mechanical damage connects isolated pores leading to transient radon signals

Supporting Information: • Readme • Figure S1 • Figure S2 Correspondence to: A. Nicolas, [email protected]

Citation: Nicolas, A., F. Girault, A. Schubnel, É. Pili, F. Passelègue, J. Fortin, and D. Deldicque (2014), Radon emanation from brittle fracturing in granites under upper crustal conditions, Geophys. Res. Lett., 41, 5436–5443, doi:10.1002/2014GL061095. Received 2 JUL 2014 Accepted 18 JUL 2014 Accepted article online 23 JUL 2014 Published online 4 AUG 2014

Radon emanation from brittle fracturing in granites under upper crustal conditions Aurélien Nicolas1, Frédéric Girault1, Alexandre Schubnel1, Éric Pili2, François Passelègue1, Jérôme Fortin1, and Damien Deldicque1 1

Laboratoire de Géologie, École Normale Supérieure, CNRS UMR 8538, Paris, France, 2CEA, DAM, DIF, Arpajon, France

Abstract

Radon-222, a radioactive gas naturally produced in the Earth’s crust, informs us about the migration of fluids and is sometimes considered as a potential earthquake precursor. Here we investigate the effects of mechanical and thermal damage on the radon emanation from various granites representative of the upper crust. Radon concentration measurements performed under triaxial stress and pore fluid pressure show that mechanical damage resulting from cycles of differential stress intensifies radon release up to 170 ± 22% when the sample ruptures. This radon peak is transient and results from the connection of isolated micropores to the permeable network rather than new crack surface creation per se. Heating to 850°C shows that thermal fracturing irreversibly decreases emanation by 59–97% due to the amorphization of biotites hosting radon sources. This study, and the developed protocols, shed light on the relation between radon emanation of crustal rocks, deformation, and pressure-temperature conditions.

1. Introduction The radioactive gas radon-222 (half-life 3.82 days) is naturally produced in rocks by alpha decay of radium-226. Easily measurable in the field, it is a precious asset that has given information about fluid migration in several volcanically [Richon et al., 2003; Cigolini et al., 2007] and tectonically active areas [Trique et al., 1999; Perrier et al., 2009]. Furthermore, radon has been considered as a potential earthquake precursor [Virk and Singh, 1994; Igarashi et al., 1995]. However, most of the reported observations have highlighted contradictory signals [e.g., Ghosh et al., 2009] and remain questionable in many aspects [Geller, 2011]. It is generally accepted that earthquake results from a sudden rupture of the crust due to the gradual increase of tectonic stress [e.g., Scholz, 2002]. Coseismic mechanical and possible thermal processes induce sudden changes of the hydraulic properties of the fault zone [Miller et al., 2004; Manga et al., 2012]. All these changes may allow the degassing of volatile elements or affect the natural release of fluid if already present [e.g., Toutain and Baubron, 1999]. However, these effects on the radon emanation from rocks are still poorly understood under upper crustal conditions. Laboratory experiments under controlled conditions can provide interesting insights. First, the effect of stress on radon emanation has been studied using uniaxial experiments on radium-rich rocks using punctual [Tuccimei et al., 2010; Mollo et al., 2011] and continuous radon measurements [Holub and Brady, 1981; Scarlato et al., 2013]. These studies have shown that radon emanation decreases during the elastic loading phase and is maximal after failure. Between these two stages, i.e., when the increase of stress induces irreversible damage, the radon emanation either decreases or does not change significantly [Mollo et al., 2011]. However, the association of confined pressure conditions and controlled pore fluid pressure may lead to new information. Second, high temperature can influence radon emanation, as evidenced in hydrothermal and metamorphic regions [Girault et al., 2012]. In the laboratory, the heating of U-bearing minerals, sediments, and clay samples has shown an irreversible decrease of radon emanation above a given temperature threshold [Garver and Baskaran, 2004; Jobbágy et al., 2009; Sas et al., 2012]. However, no experiment has been reported on granite, the major constituent of the upper continental crust. In this paper, we report new laboratory experiments on four different granites. First, radon emanation from thermally and mechanically fractured samples was compared with that from intact samples. Second, a new setup placing the sample under natural conditions was used to investigate radon release during mechanical damage.

2. Samples Description Four granites (porosity 1%) were studied (Table 1): a leucogranite (LG), Ailsa Craig (AC), La Peyratte (LP), and Westerly (W). Average grain size was determined by optical microscopy. LG shows the highest average grain

NICOLAS ET AL.

©2014. American Geophysical Union. All Rights Reserved.

5436

NICOLAS ET AL.

©2014. American Geophysical Union. All Rights Reserved.

La Peyratte granite, France (200 ± 40 μm)

Westerly granite, USA (500 ± 100 μm)

LP

W

n.m.: Not measured.

Ailsa Craig granite, Scotland (20 ± 4 μm)

AC

a

Leucogranite, Portugal (2000 ± 400 μm)

Rock Type, Location (Average Grain Size)

LG

Sample Name

Quartz-feldspar Feldspar Biotite Iron oxides and other minerals

Bulk rock

Quartz-feldspar Feldspar Biotite-muscovite 60% feldspar, 25% biotite

Bulk rock

Bulk rock

Quartz-feldspar K-Feldspar Muscovite Biotite with traces of chlorite

Bulk rock

Location of Analyses

29.4

n.m.

a

19.4

20.3

Th (ppm)

232

0.5 0.1 3.0 14.2

1.3

3.2 2.2 12.6 12.4

3.1

4.9

8.2 7.4 13.6 52.0

23.9

U (ppm)

238

≤10 ≤10 50 150

10

50 30 240 200

60

60

120 100 190 970

310

Ra 1 (Bq kg )

226

Table 1. Petrophysical Properties, Composition, and Radon Parameters of the Granites

( Ra)/ ( U)

0.30

0.74

0.47

0.50

238

226

2.19 ± 0.16

2.78 ± 0.22

5.42 ± 0.40

33.3 ± 1.7

ECRa 1 (Bq kg )

21.9 ± 1.6

4.63 ± 0.37

9.03 ± 0.66

10.75 ± 0.36

Intact Sample

22.1 ± 2.9 (+0.9 ± 0.1%)

4.92 ± 0.31 (+6.3 ± 0.6%)

10.63 ± 0.72 (+18 ± 2%)

10.61 ± 0.71 (1.3 ± 0.1%)

Mechanically Fractured

6.43 ± 0.35 (72 ± 2%)

2.78 ± 0.35 (34 ± 5%)

5.36 ± 0.50 (35 ± 4%)

7.31 ± 0.47 (32 ± 2%)

550°C

2.45 ± 0.60 (88 ± 22%)

1.91 ± 0.17 (59 ± 3%)

0.245 ± 0.085 (97 ± 34%)

1.140 ± 0.033 (89 ± 4%)

850°C

Thermally Fractured

Radon Emanation Coefficient E (%) (Difference Against Intact)

Geophysical Research Letters 10.1002/2014GL061095

5437

Geophysical Research Letters

10.1002/2014GL061095

100 µm Zrn

20 µm

Zrn

Mnz

Mnz

Bt Qz

Bt a

Radium

20 µm

20 µm

c

Uranium

b

Thorium

d

Figure 1. (a) Large-scale scanning electron microscope (SEM) image and detailed element mapping analyses of biotites showing high (b) radium, (c) uranium, and (d) thorium concentrations in zircons and monazites for the LG leucogranite. Qz: quartz; Bt: biotite; Mnz: monazite; Zrn: zircon.

size (2000 μm), AC the smallest (20 μm), and LP and W intermediate values of 200 and 500 μm, respectively. Mineralogy was studied with a scanning electron microscope (SEM) and with X-ray diffraction (XRD). 232Th and 238U concentrations and 226Ra concentration were measured on bulk and mineral fractions using highresolution inductively coupled plasma–mass spectrometry and alpha spectrometry, respectively (Table 1). Energy-dispersive X-ray spectrometry was used to detect radium-rich inclusions in minerals. While bulk 232Th concentration remained relatively similar in all samples, bulk 238U and 226Ra concentrations varied from 1.3 to 24 ppm and from 10 to 310 Bq kg1, respectively. In general, 238U, 232Th, and 226Ra are mainly concentrated in zircons and monazites entrapped in biotites (Figure 1), which present concentrations in these elements 3–5 times larger than in the bulk sample. The (226Ra)/(238U) activity ratios of the studied samples, in the range 0.3–0.7, indicate that the uranium decay chain is not at secular equilibrium. Some radium must have been lost relatively to uranium during some open-system processes that could occurred at several time since rock formation. All experiments were performed on cylindrical cores of diameter 4 cm and length 8 cm. Samples were kept at 40°C for at least 1 week before the start of an experiment.

3. Experimental Methods 3.1. Radon-222 Emanation From Intact, Thermally, and Mechanically Fractured Samples Only a fraction of 226Ra atoms are able to produce 222Rn atoms that reach the pore space and then the surface of the sample [e.g., Nazaroff, 1992]. This fraction is called the emanation coefficient E and depends on the spatial distribution of radium, the properties of the porous network, moisture, and temperature [e.g., Sakoda et al., 2011]. The radon source term (effective radium concentration, ECRa), expressed in Bq kg1, is the product of the bulk radium concentration (CRa) and E. The ECRa value of intact, thermally, and mechanically fractured samples were determined in the laboratory using the accumulation method. The sample was placed in an accumulation chamber connected to an intercalibrated ionization chamber (Alphaguard™, Saphymo, Germany), which measured continuously the radon concentration in-growth C(t) at 1 h interval following Ferry et al. [2002]. The ECRa value was calculated as a function of accumulation time t and possible leakage in the experiment using [Perrier and Girault, 2012] EC Ra ðtÞ ¼

Va C ðtÞ ð1 þ aV Þ ’ m 1  eλð1þaV Þt

(1)

where Va is the total air volume available, m is the mass of the sample, λ is the decay constant of radon-222 (2.1 × 106 s1), and aV = λV/λ is the normalized leakage rate determined by a least squares fit of the data, NICOLAS ET AL.

©2014. American Geophysical Union. All Rights Reserved.

5438

Geophysical Research Letters

a

10.1002/2014GL061095

b

200 µm

c

d

e

f

g

h

Figure 2. SEM images of (a, c, e, and g) intact and (b, d, f, h) 850°C heat-treated granites sorted by decreasing average grain size: from top to bottom, (Figures 2a and 2b) leucogranite (LG), (Figures 2c and 2d) Westerly (W), (Figures 2e and 2f) La Peyratte (LP), and (Figures 2g and 2h) Ailsa Craig (AC).

where λV is the volumetric leakage rate. The total uncertainty on ECRa (8% on average) corresponds to the punctual uncertainty related to accumulation curve (