• No results found

Rare earth elements in mud-rich sediments and their use as provenance indicators

N/A
N/A
Protected

Academic year: 2020

Share "Rare earth elements in mud-rich sediments and their use as provenance indicators"

Copied!
222
0
0

Loading.... (view fulltext now)

Full text

(1)

Rare earth elem ents in mud-rich sedim ents and their use as

provenance indicators.

Jane Louise Alexander

Ph.D.

(2)

ProQuest Number: 10045858

All rights reserved

INFORMATION TO ALL USERS

The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed,

a note will indicate the deletion.

uest.

ProQuest 10045858

Published by ProQuest LLC(2016). Copyright of the Dissertation is held by the Author.

All rights reserved.

This work is protected against unauthorized copying under Title 17, United States Code. Microform Edition © ProQuest LLC.

ProQuest LLC

789 East Eisenhower Parkway P.O. Box 1346

(3)

Abstract

Rare earth elem ent (REE) patterns play a key role in understanding the provenance of m udrocks. How ever, there is uncertainty about the extent to which REE are mobile in sedim ents, and the effect this has on the provenance signal. This study focuses on M iocene to Pliocene hemipelagic m udrocks from O cean D rilling Program (G D P ) Site 808, drilled in the toe of the N ankai accretionary prism , SE Japan. A second m udrock suite from the A m atsu Form ation, Boso Peninsula, Japan, was studied as a comparison.

M udrocks from the N ankai accretionary prism all have interm ediate provenance, while those from the A m atsu Form ation on the Boso Peninsula tren d from interm ediate provenance in the lower part of the sequence to mafic provenance in the upper part. There is a correlation betw een REE fractionation and provenance, which holds for b o th m udrock suites. M udrocks w ith interm ediate provenance have m ore fractionated REE patterns and higher concentrations of the light REE than those w ith mafic provenance. Some samples do n o t conform to this trend. Those from the Boso Peninsula appear to have had mixed sedim ent sources.

The anomalous samples from the N ankai accretionary prism are related to fluid- rock interactions at the décollem ent zone, which have altered the provenance signature. M ost have REE, trace and m ajor elem ent concentrations similar to the oth er hemipelagic m udrocks, despite being brecciated. H ow ever, there is a heavy R EE enrichm ent in some samples from the décollem ent that are close to breaks in core recovery. T hey are associated w ith an enrichm ent in calcium, in the form of calcite. The enrichm ent of heavy REE in otherw ise typical m udrocks suggests that they have been transported to the décollem ent in fluids, where they were co­

(4)

Table of Contents

A b s tra c t... 2

Table of C o n te n ts ... 3

List of Figures... 7

List of T a b le s...17

List of P la te s ...18

A cknow ledgem ents...20

C h apter 1. In tro d u c tio n ...21

1.1 Aims and objectives... 21

1.2 Previous w o r k ... 23

1.2.1 N ankai T rough m u d ro ck s... 23

1.2.2 Review of literature on rare earth elem ents in sed im en ts...25

1.2.2.1 General observations...25

1.2.2.2 M ineralogy...26

1.2.2.3 R ock type and size fractio n ... 27

1.2.2.4 Provenance...28

1.2.2.5 D iagenesis...29

1.2.2.6 P ro b lem s... 30

1.2.2.7 C o nclu sio n s...31

C h ap ter 2. Sample preparation and analytical techniques... 32

2.1 Sum m ary of techniques...32

2.2 Sample preparation...32

2.2.1 D rying and crushing... 32

2.2.2 Acid digestion...33

2.2.3 Clay mineral separation...33

2.3 Leaching ex p erim en ts... 34

2.3.1 M e th o d ...34

2.3.1.1 Stage 1...35

2.3.1.2 Stage 2 ...35

2.3.1.3 Stage 3 ...35

(5)

2.3.1.5 Stage 5 ...36

2.4 Sample analysis... 36

2.4.1 IC P-A ES A nalysis...36

2.4.2 IC P-M S A nalysis... 37

2.4.3 A AS A nalysis...37

2.4.4 X R D and Electron M icroprobe A nalysis... 37

2.4.5 E stim ation of E r r o r s ... 38

C hapter 3. Results from the N ankai T ro u g h ...39

3.1 In tro d u ctio n to the N ankai T rough and O D P Site 808...39

3.1.1 Location of O D P Site 808... 40

3.1.2 Stratigraphy at O D P Site 808...41

3.1.3 H y d ro g e o lo g y ...43

3.2 Sampling philo so ph y ...44

3.3 G eochem istry and m ineralogy...50

3.3.1 B ackground hemipelagic m u d ro c k s ... 53

3.3.1.1 M udrocks from the décollem ent z o n e ... 55

3.3.1.2 D ark coloured m udrocks... 57

3.3.1.3 Yellow coloured m udrocks...58

3.3.2 A nom alies...59

3.3.2.1 Anom alous décollem ent m udrocks... 59

3.3.2.2 U m b e rs...66

3.3.2.3 M etalliferous m udrocks... 68

3.3.2.4 A nom alous background m u d ro ck ... 68

3.3.2.5 Anom alous yellow m u d ro c k ... 69

3.4 Leaching ex p erim en ts... 70

3.4.1 Problem s and u n certain ties... 70

3.4.1.1 N on-selectivity of leaching solutions...70

3.4.1.2 R edistribution of trace elem en ts...71

3.4.1.3 C o n ta m in a tio n ... 72

3.4.1.4 W ash w ater d isp o sa l... 72

(6)

3.4.2.1 Stage 1... 73

3.4.2.2 Stage 2 ... 73

3.4.2.3 Stage 3 ... 73

3.4.2.4 Stage 4 ... 76

3.4.2.5 Stage 5 ... 76

3.4.2.6 Stage 6 ... 76

3.4.3 Summary of REE leached at each stage...77

3.4.3.1 Stage 3 ... 77

3.4.3.2 Stage 4 ... 80

3.4.3.3 Stage 5 ... 82

3.4.3.4 Stage 6 ... 82

3.5 M odelling of pore fluid c h em istry ...85

3.5.1 In tro d u c tio n ...85

3.5.2 Overview of m o d e ls... 85

3.5.3 M odelling approach... 86

3.5.4 R e su lts...88

3.5.5 D iscu ssio n ... 90

C hapter 4. Results from the Boso P en in su la... 91

4.1 In tro d u ctio n to the Boso Peninsula field study area...91

4.1.1 Basic stratig rap h y ...93

4.1.2 Field study a re a ... 98

4.2 F ield w o rk ...103

4.3 G eochem istry and m ineralogy... 109

4.3.1 Samples from Location 5 ... 113

4.3.2 Samples from Locations 7 and 8 ...114

4.3.3 Samples from Location 9 ... 115

4.3.4 Samples from Location 12... 119

4.3.5 Samples from Location 14... 120

4.3.6 U m b e r ... 124

C hapter 5. D iscu ssio n ... 125

5.1 M udrock p ro v en an ce... 125

(7)

5.1.1.1 M e th o d ...125

5.1.1.2 Provenance of N ankai T ro u g h sam ples... 128

5.1.1.3 Provenance o f Boso Peninsula sam ples... 133

5.1.2 Relationships betw een REE and p rovenance... 140

5.2 A nom alous REE concentrations in the décollem ent z o n e ... 149

5.2.1 N atu re of the heavy REE anom aly... 149

5.2.2 Possible causes of the heavy REE anom aly...152

5.2.3 Possible sources of the heavy R E E ... 156

5.2.4 The influence of fluids on the décollem ent anom aly... 161

5.3 U m bers and associated se d im en ts... 164

5.3.1 U m bers from the N ankai T ro u g h ...164

5.3.2 M etalliferous m udrocks from the N ankai T r o u g h ... 168

5.3.3 U m ber from the M ineoka O p h io lite ...169

C h apter 6. C onclusions... 171

R eferences... 174

A ppendix I N ankai T rough re s u lts ... 187

A ppendix II Leaching experim ent re su lts ... 208

(8)

List of Figures

Figure 1. M ap of Japan showing location of sampling sites, O D P Site 808 and the Boso Peninsula... 22

Figure 2. C h o n d rite norm alized REE concentrations from the N ankai accretionary prism (data from Pickering etal.y 1993a)...24

Figure 3. REE concentrations in décollem ent samples and um ber, norm alized to background m udrocks from the N ankai accretionary prism (data from

Pickering et al., 1993a)... 25

Figure 4. C h o n d rite norm alized REE concentrations of average shales from different continents: Post-A rchaean A ustralian Average Shale (McLennan, 1989), N o rth Am erican Shale C om posite (G rom et etal., 1984) and European Shales (H askin and FI as kin, 1966)... 26

Figure 5. M ap of southern Japan, showing the locations of the Shikoku Basin,

N ankai T ro u g h and plate boundaries... 39

Figure 6. Schematic cross section showing the location of Site 808 in the N ankai accretionary prism . Section drawn along line X-Y in Figure 5... 40

Figure 7. Stratigraphy of the N ankai accretionary prism at Site 808C. Com piled using data from Shipboard Scientific Party (1991a), Pickering eta l. (1993b) and U nderw ood er iï/. (1993a)...41

Figure 8. M ajor and trace elem ent concentrations versus depth (m bsf). D o tte d red lines represent the top and b o tto m o f the décollem ent zone...51

Figure 9. REE concentrations versus depth (m bsf). D o tte d red lines represent the to p and b o tto m of the décollem ent zo n e... 52

(9)

Figure 11. C h o n d rite norm alized (Taylor and M cLennan, 1985) concentrations of REE in the background hemipelagic m udrocks com pared to the N o rth Am erican Shale C om posite (G rom et etal.^ 1984) and Post-A rchaean

A ustralian Average Shale (M cLennan, 1989)... 54

Figure 12. Variations in concentration of selected REE (ppm) w ith depth (mbsf) th ro u g h the décollem ent zone. The grey lines show the breaks in core recovery. ...56

Figure 13. C h o n d rite norm alized (Taylor and M cLennan, 1985) concentrations of REE in the non-anom alous décollem ent m udrocks com pared to the N o rth Am erican Shale C om posite (G rom et etal.^ 1984) and Post-A rchaean

A ustralian Average Shale (M cLennan, 1989)...56

Figure 14. % Fe2 0 3 and S (ppm) concentrations versus depth (m bsf). Brown triangles represent samples n o ted to be darker in colour, yellow circles

represent samples n o ted to be yellowish in colour, and blue crosses represent the m etalliferous m udrocks. D o tte d red lines represent the top and b o tto m of the décollem ent zone... 57

Figure 15. C h o n d rite norm alized (Taylor and M cLennan, 1985) concentrations of REE in the dark coloured m udrocks com pared to the N o rth A m erican Shale C om posite (G rom et etal., 1984) and Post-A rchaean A ustralian Average Shale (M cLennan, 1989)...58

Figure 16. C h o n d rite norm alized (Taylor and M cLennan, 1985) concentrations of R EE in the yellow coloured m udrocks com pared to the N o rth American Shale C om posite (G rom et etal., 1984) and Post-A rchaean A ustralian Average Shale (M cLennan, 1989)...59

(10)

Figure 18. X RD analysis o f clay-sized fractions of a ‘norm al’ décollem ent m udrock (69R-3, 10-12) and an anomalous décollem ent m udrock (70R1, 10-12). There is very little difference betw een the superim posed patters, suggesting that clay mineralogy is the same in b o th samples. T he only difference is that 70R1, 10-12 has a m uch larger calcite peak...65

Figure 19. C h o n d rite norm alized (Taylor and M cLennan, 1985) concentrations of REE in the anomalous décollem ent m udrocks com pared to the N o rth

A merican Shale C om posite (G rom et etal., 1984) and Post-A rchaean

A ustralian Average Shale (M cLennan, 1989). Average, m aximum and m inim um values for hemipelagic m udrocks and background décollem ent m udrocks are shown for reference... 65

Figure 20. ^oAljOj, % M gO and “/oKjO versus depth (mbsf). Pink stars represent the um bers, blue crosses represent the m etalliferous m udrocks, the grey diam ond represents the anomalous hemipelagic m udrock and the yellow circle represents the anomalous yellow m udrock... 66

Figure 21. % M nO , % C aO and Sr (ppm) versus depth (mbsf). Pink stars represent the um bers, blue crosses represent the metalliferous m udrocks, the grey diamond represents the anomalous hemipelagic m udrock and the yellow circle represents the anomalous yellow m udrock... 67

Figure 22. C h o n d rite norm alized (Taylor and M cLennan, 1985) concentrations of REE in the um bers, metalliferous m udrocks, anom alous background m udrock and anomalous yellow m udrock, com pared to the N o rth A m erican Shale C om posite (G rom et etal., 1984) and Post-A rchaean A ustralian Average Shale (McLennan, 1989). Average, maximum and m inim um values for background hemipelagic m udrocks are show n for reference... 67

Figure 23. G raph of % T i02 versus Z r (ppm). H ig h concentrations o f these

elements indicate the presence of heavy accessory minerals. T he regression line is fitted throu g h all samples w ith detectable levels o f zirconium ... 69

Figure 24. C o n cen tratio n of calcium (ppm) rem oved during each stage of the

(11)

Figure 25. C o n cen tratio n of stro ntiu m (ppm) rem oved during each stage of the sequential leaching experim ents. See text for description of each stage 74

Figure 26. C o n cen tratio n of iron (ppm) rem oved during each stage of the sequential leaching experim ents. See text for description of each stage...75

Figure 27. C o ncen tratio n of manganese (ppm) rem oved during each stage of the sequential leaching experim ents. See text for description of each stage 75

Figure 28. C h o n d rite norm alized (Taylor and M cLennan, 1995) concentrations of REE leached during stage 3, and assum ed to be present in the carbonate phase. ...78

Figure 29. REE concentrations from stage 3, expressed as a percentage of the sum of concentrations from all leaching stages...78

Figure 30. C o n cen tratio n of lanthanum (ppm) rem oved during each stage of the sequential leaching experim ents. See text for description of each stage 79

Figure 31. C on cen tratio n of samarium (ppm) rem oved during each stage of the sequential leaching experim ents. See text for description of each stage 79

Figure 32. C o n cen tratio n of lutetium (ppm) rem oved during each stage of the sequential leaching experim ents. See text for description of each stage 79

Figure 33. C h o n d rite norm alized (Taylor and M cLennan, 1995) concentrations of REE leached during stage 4, and assumed to be present in the oxide phase 81

Figure 34. REE concentrations from stage 4, expressed as a percentage of the sum of concentrations from all leaching stages...81

Figure 35. C h o n d rite norm alized (Taylor and M cLennan, 1995) concentrations of REE leached during stage 5, and assumed to be present in the sulphide phase. 83

Figure 36. REE concentrations from stage 5, expressed as a percentage of the sum of concentrations from all leaching stages...83

Figure 37. C ho n d rite norm alized (Taylor and M cLennan, 1995) concentrations of REE leached during stage 6, and assumed to be present in the aluminosilicate phases, com pared to the N o rth American Shale C om posite (G rom et et al.^

(12)

Figure 38. R EE concentrations from stage 6, expressed as a percentage of the sum of concentrations from all leaching stages...84

Figure 39. V ariation in p H w ith depth m easured in borehole 808C (Shipboard Scientific Party, 1991a). D o tte d red lines represent the to p and b o tto m of the décollem ent zone... 87

Figure 40. V ariation in w ater density w ith tem perature and pressure, calculated using H 2 0 . The red line shows conditions in borehole 808C, w ith the vertical lines being erro r bars... 87

Figure 41. Variations in N a^ and C l' concentrations w ith depth (left) and Ca^"^ and S O / concentrations w ith depth (right). D o tte d red lines represent the top and b o tto m o f the décollem ent zone...88

Figure 42. M odel variations in lanthanum and lutetium spéciation w ith depth.

D o tte d red lines represent the to p and b o tto m o f the décollem ent zone... 89

Figure 43. Location of the Boso Peninsula... 91

Figure 44. Stratigraphy of the sedim entary basins to the n o rth and sou th of the M ineoka C om plex... 93

Figure 45. Geological map of the Boso Peninsula... 94

Figure 46. Geological map of the area around Awa-A m atsu, from w here m udrocks of the A m atsu F orm ation were collected (m odified after N akajim a eta/., 1980). ... 99

Figure 47. Geological map of the M ineoka C om plex and surrounding area (modified after N akajim a et al., 1980)... 100

Figure 48. Field map of the M ineoka M ountains showing sampling locations 1 to 3. Scale is now 1 : 27 200 due to red u ctio n ... 106

Figure 49. M ajor element concentrations in samples from the Boso Peninsula. See Table 7 for descriptions and locations of sample n u m b e rs ... 110

(13)

Figure 51. Selected REE concentrations in samples form the A m atsu Form ation. See Table 7 for descriptions and locations of sample num bers... 112

Figure 52. C h o n d rite norm alized (Taylor and M cLennan, 1995) concentrations of REE in samples from Location 5, com pared to the N o rth American Shale C om posite (G rom et etal.^ 1984) and Post-A rchaean A ustralian Average Shale (M cLennan, 1989)... 113

Figure 53. C h o n d rite norm alized (Taylor and M cLennan, 1995) concentrations of REE in samples from Locations 7 and 8, com pared to the N o rth American Shale C om posite (G rom et etal.^ 1984) and Post-A rchaean A ustralian Average Shale (M cLennan, 1989)... 115

Figure 54. C h o n d rite norm alized (Taylor and M cLennan, 1995) concentrations of REE in samples from Location 9, com pared to the N o rth Am erican Shale C om posite (G rom et etal.^ 1984) and Post-A rchaean A ustralian Average Shale

(M cLennan, 1989)... 116

Figure 55. C h o n d rite norm alized (Taylor and M cLennan, 1995) concentrations of REE in samples from Location 12, com pared to the N o rth Am erican Shale C om posite (G rom et etal.^ 1984) and Post-A rchaean A ustralian Average Shale

(M cLennan, 1989)...120

Figure 56. C h o n d rite norm alized (Taylor and M cLennan, 1995) concentrations of REE in samples from Location 14, com pared to the N o rth Am erican Shale C om posite (G rom et et al., 1984) and Post-A rchaean A ustralian Average Shale

(M cLennan, 1989)...121

Figure 57. C h o n d rite norm alized (Taylor and M cLennan, 1995) concentrations of REE in the um ber, com pared to the N o rth Am erican Shale C om posite

(G rom et etal., 1984) and Post-A rchaean A ustralian Average Shale (M cLennan, 1989)... 124

(14)

Figure 59. D iscrim inant function diagram 2, defining the fields for sedim ents of mafic, interm ediate, felsic and quartzose provenance for sedim ents containing significant biogenic com ponents (after R oser and Korsch, 1988)... 128

Figure 60. D iscrim inant function diagram showing the provenance of hemipelagic m udrocks from the N ankai accretionary prism ... 129

Figure 61. D iscrim inant function diagram showing the provenance of mudrocks from décollem ent zone of the N ankai accretionary prism ... 130

Figure 62. D iscrim inant function diagram showing the provenance of dark and yellow coloured m udrocks from the N ankai accretionary prism ... 131

Figure 63. Plot of iron concentration (% Fe20)) against sulphur concentration

(ppm) for m udrocks from the N ankai accretionary prism . The regression line is fitted thro u g h the data for the dark coloured m udrocks... 132

Figure 64. D iscrim inant function diagram showing the provenance of samples from Location 5 on the Boso Peninsula... 133

Figure 65. D iscrim inant function diagram showing the provenance of samples from Location 8 on the Boso Peninsula... 134

Figure 66. P lot of discrim inant function 1 against distance th ro u g h the sequence, showing a direct correlation. The regression line is drawn thro u g h all samples, except Sample 35 which is obviously anom alous... 135

Figure 67. P lot of discrim inant function 2 against distance th ro u g h the sequence, showing a direct correlation. The regression line is drawn th ro u g h all samples, except Sample 35 which is obviously anom alous... 135

Figure 68. D iscrim inant function diagram showing the provenance of samples from Location 9 on the Boso Peninsula... 136

Figure 69. D iscrim inant function diagram showing the provenance of samples from Location 12 on the Boso Peninsula... 137

(15)

Figure 71. D iscrim inant function diagram showing the provenance of all samples from the Boso Peninsula and their location in the A m atsu F o rm atio n 139

Figure 72. C oncentrations of REE in background hemipelagic m udrocks norm alized to the N o rth Am erican Shale C om posite (G rom et eta l., 1984)... 141

Figure 73. C oncentrations of REE in décollem ent m udrocks norm alized to the N o rth Am erican Shale C om posite (G rom et eta l., 1984)... 142

Figure 74. C oncentrations of REE in dark and yellow coloured m udrocks

norm alized to the N o rth Am erican Shale C om posite (G rom et etal., 1984). 142

Figure 75. P lot of chondrite norm alized (La/Y b)^ against C e^ for samples from the N ankai T rough (chondrite data from Taylor and M cLennan, 1985). The

regression line is fitted th ro u g h all background hemipelagic, décollement, dark and yellow coloured m udrocks... 143

Figure 76. C oncentrations of R EE in samples from the Boso Peninsula, norm alized to the N o rth American Shale C om posite (G rom et etal., 1984)...145

Figure 77. P lot of chondrite norm alized (La/Y b)^ against C e^ for samples from the Boso Peninsula (chondrite data from Taylor and M cLennan, 1985). The

regression line is fitted throu g h all samples except the quartzose sandstone, which is anom alous... 146

Figure 78. P lot of chondrite norm alized (La/Y b)^ (chondrite data from Taylor and M cLennan, 1985) against distance thro u g h the sequence, showing a direct correlation. The regression line is drawn th ro u g h all samples, except Sample 35 w hich is obviously anom alous...147

Figure 79. P lot of chondrite norm alized (La/Y b)^ against C e^ for samples from b o th the N ankai T rough and the Boso Peninsula (chondrite data from Taylor and M cLennan, 1985). The regression line is fitted th ro u g h all samples except the anomalous ones discussed earlier... 149

(16)

Figure 81. Plot of chondrite norm alized (L a/Y b)^ against Yb^ for hemipelagic and décollem ent m udrocks from the N ankai accretionary prism (chondrite data from T aylor and M cLennan, 1985). O n e regression line is fitted through all background hemipelagic and décollem ent m udrocks, and the o th er thro u g h the anomalous décollem ent m udrocks... 151

Figure 82. Plot of calcium concentration (ppm) against y tterb iu m concentration (ppm) for bulk m udrock samples from the N ankai accretionary prism. The regression line is draw n thro u g h all data points except the anomalous

hemipelagic m udrock and the drilling m ud... 153

Figure 83. P lot of calcium concentration (ppm) against lanthanum concentration

(ppm) for bulk m udrock samples from the N ankai accretionary prism 153

Figure 84. Plot of chondrite norm alized (La/Y b)^ against calcium concentration (ppm) for bulk m udrock samples from the N ankai accretionary prism. The regression line is drawn through the anomalous décollem ent m udrocks 154

Figure 85. Plot of calcium concentration (ppm) against ytterb iu m concentration (ppm) for each stage of the leaching experim ents on m udrock samples from the N ankai accretionary prism. The regression line is drawn th ro u g h stage 3 and 4 results for the anomalous samples...155

Figure 86. P lot of calcium concentration (ppm) against lanthanum concentration (ppm) for each stage of the leaching experim ents on m udrock samples from the N ankai accretionary prism. T he regression line is drawn th ro u g h stage 3 and 4 results... 155

Figure 87. Plot of stro ntiu m concentration (ppm) against calcium concentration (ppm) for bulk m udrock samples from the N ankai accretionary prism. The regression line is drawn th ro u g h all background hemipelagic and décollement m udrocks... 158

(17)

Figure 89. Plot of Y b /C a atomrado against Y b/L a atomrado^ to determ ine the genesis of calcite (Parekh and M oiler, 1977). D o tte d lines approxim ate the fields where samples p lo tted in the original study. D ata is p lo tted for stage 3 of the leaching experim ents, where only calcite was rem oved for analysis... 159

Figure 90. REE concentrations in anomalous décollem ent m udrocks, um bers and m etalliferous m udrocks, norm alized to background hemipelagic and

décollem ent m udrocks from the N ankai accretionary prism ... 161

Figure 91. Plot of calcium and ytterbium concentrations (ppm) against depth

(m bsf), showing the shape of the anom aly above the lower core break... 162

Figure 92. Plot of manganese concentrations (ppm) against aluminium

concentrations (ppm) for all samples betw een 1000 and 1100 mbsf. A rrow indicates decreasing co n ten t of clay m inerals... 166

Figure 93. Plot of calcium concentrations (ppm) against aluminium concentrations (ppm) for all samples betw een 1000 and 1100 mbsf. Regression line is drawn th ro u gh the um bers and metalliferous m udrocks... 166

(18)

List of Tables

Table 1. List of samples collected from O D P C ore 808C during the first visit to the repository, w ith notes on samples that are different to the background

hemipelagic m udrocks... 45

Table 2. List of samples selected for sequential leaching experim ents, representing background and anomalous m udrocks from the décollem ent zone...70

Table 3. Summ ary of form ations from the M ineoka G roup. D ata from N akajim a et a l, 1980... 94

Table 4. Summ ary of form ations from the H o ta G roup. D ata from Suzuki et al.^

1990...95

Table 5. Summ ary of form ations from the Sakuma G roup. D ata from Suzuki etal.,

1990...96

Table 6. Summ ary of form ations from the Awa G roup. D ata from N akajim a et al.^

1980 and Suzuki et al., 1990... 97

Table 7. List of samples collected during field study on the Boso Peninsula, Japan. ... 105

Table 8. D iscrim inant function coefficients for the first diagram (Roser and Korsch,

1988) 126

Table 9. D iscrim inant function coefficients for the second diagram (Roser and Korsch, 1988)... 127

(19)

List of Plates

Plate 1. Û Photographs of the archive (upper) and w orking (lower) halves of core 68R-2, which is located directly above the upper core break in the décollement zone. T he to p of the core is to the left. Samples were taken at 3 cm intervals, and are m arked ‘X ’. Scale on ruler is cm ...47

Plate 2. Example of a typical background hemipelagic m udrock (sample 72R-4, 71-73)... 49

Plate 3. L? Typical examples of m udrocks from the décollem ent zone, showing different degrees of brecciation - brecciated (left), very brecciated (centre) and extrem ely brecciated (rig h t)... 49

Plate 4. <P Pyrite and o th er am orphous sulphur minerals on the surface of sample 79R-2, 110-112. Spacing betw een scale lines is 0.6 m m ...49

Plate 5. Û Jarosite and oth er am orphous sulphur minerals on the surface of sample 80R-1, 19-21. Spacing betw een scale lines is 0.6 m m ... 49

Plate 6. Û P h o to showing coastal exposure of the A m atsu m udrocks, taken from A w a-A m atsu looking east tow ards Awa-Kamogawa... 101

Plate 7. Typical thick sem i-tropical forest covering inland hillsides...101

Plate 8. 13 Rice fields, typical of the intensive cultivation of valleys and shallow slopes. Thick forest and bam boo can be seen in the background...101

Plate 9. ■=> Typical Japanese road cutting that has been covered in concrete (“p h o to 17"on Figure 48b)... 102

Plate 10. ^ Small road cutting that has been left w ith o u t concrete, due to by-passing by new road (“loc 5” on Figure 48 b )... 102

Plate 11. 13 Location of um ber in the M ineoka C om plex (“loc 1” on Figure 48a). Sampling was n o t possible due to thick und erg ro w th and extensive weathering.

(20)

Plate 12. û Photograph of Location 5 (see Figure 48b for locality “loc 5”) taken facing tow ards 150°. Sample locations are m arked ‘X ’. There is no obvious correlation betw een beds on either side o f the fault...117

Plate 13. T hin section of Sample 13, shown in plane-polarized light (left) and u n d er crossed polars (right). Clasts are labelled, and the iron-oxide staining of the clay m atrix is clearly visible... 117

Plate 14. Û Photograph of Location 9 (see Figure 48b for locality “loc 9”) taken facing tow ards due no rth . Sample locations are m arked ‘X ’... 118

Plate 15. Û T hin section of Sample 39, shown in plane-polarized light (left) and under crossed polars (right). Lithic clasts, w ith plagioclase phenocrysts are labelled ‘A ’. There is a large plagioclase phenocryst at the centre on the frame. The clast marked ‘B’ appears to be com posed of fine, glassy, tuffaceous

material... 118

Plate 16. Û T hin section of Sample 57, shown in plane-polarized light (left) and under crossed polars (right). Ffigh order interference colours in the matrix indicate a calcite cem ent...122

Plate 17. Û T hin section of Sample 57, shown in plane-polarized light (left) and un d er crossed polars (right). A shell fragm ent is visible near the centre of the fram e...122

Plate 18. Û T hin section of Sample 63, shown in plane-polarized light (left) and un d er crossed polars (right). The grass shards appear black u n d er crossed polars. H ig h order interference colours in the m atrix indicate a calcite cem ent.

122

(21)

Acknowledgements

I have received assistance and support from m any people during my tim e as a research student, and would like to take this o p p o rtu n ity to say thank you.

Firstly, I w ould like to thank my supervisors. D r. Kevin Pickering (U niversity College L ondon), D r. Elizabeth Bailey (U niversity o f N ottingham ) and Prof. Paul H en d erso n (The N atu ral H isto ry M useum ), for coming up w ith the original idea for the project and for their extensive help and support over the past three years.

M y fiance, G raham Hill, has provided endless support, b o th financially and

emotionally, and I w ould like to thank him for making m y time as a student m uch easier than it w ould otherw ise have been. Thanks also go to G raham for setting up and m aintaining my com puter facilities.

M any mem bers of staff in the M ineralogy D epartm ent of the N atu ral H istory M useum have answered questions, provided technical assistance and offered helpful suggestions, th ro u gh o u t the course of this project. In particular, I w ould like to thank M r. Vic D in and M r. Gary Jones for assistance in the chemical laboratory and w ith IC P-A ES analysis; Mr. T erry G reenw ood and M r. T o n y W ight on for making polished th in sections and helping w ith sedim ent crushing; and D r. G ordon C ressey and D r. M eryl Batchelder for X R D analysis.

T he D epartm ent of G eology at the U niversity of Bristol allowed me access to IC P - MS facilities, and I w ould like to th an k D r. T o n y Kemp for all his assistance during m y visits there. Laboratory facilities for the leaching experim ents and atomic absorption analysis were provided by the D epartm ent o f Physiology and Environm ental Science at the U niversity of N o ttin g ham .

(22)

Chapter 1,

Introduction

1.1

Aims and objectives

Rare earth element (REE) patterns play a key role in understanding the provenance of m udrocks (M cLennan, 1989). Shales from around the w orld have very similar REE patterns, w ith little fluctuation th ro u g h tim e since the end of the Archaean. This is a result o f recycling and mixing o f source rocks on the continents. H ow ever, in active tectonic settings, where source rocks are m uch younger, a variety of REE patterns are present in sedim ents (M cLennan, 1989; M cLennan etaLy 1990). These patterns represent the variable chemical com position and m aturity of the source rocks, and allow the provenance of m udrock suites to be determ ined (Bock et al.y

1994).

There is, however, uncertainty about the extent to which REE are mobile in such m udrocks. They can be redistributed during weathering, early diagenesis,

hydrotherm al activity, burial diagenesis or m etam orphosis (H em m ing e t a l . y 1995).

The m obility of each elem ent is determ ined by a com bination of factors, including pore fluid complexation, mineral precipitation and adsorption of REE to mineral surfaces. The relative im portance o f each process varies depending on tem perature, pressure and chemical conditions, and may lead to a fractionation betw een heavy and light REE (R onov e t a l . y 1967). The original provenance signal may therefore be

masked or distorted.

This study focuses on REE associated w ith hemipelagic m udrocks from the O cean Drilling Program , Leg 131, Site 808, drilled in the toe o f the N ankai accretionary prism (Figure 1; Shipboard Scientific Party, 1991a). This area was chosen as previous w ork indicated anomalous R EE patterns associated w ith the décollem ent zone (Pickering e t a l . y 1993a; see Section 1.2.1). M udrocks of similar age and

(23)

Key

Plate boundary

Relative plate motion

Relict spreading centre

JAPAN SEA

PACIFIC OCEAN

Site

Shikoku

Basin S ca le (km)

300

Figure 1. Map o f Japan show ing location o f sampling sites, O D P Site 808 and the B oso Peninsula.

Initially, the project had three main aims:

• evaluate the ways in which REE are/become associated with clay minerals.

• understand the influence of p H /E h and the presence of various chemical species in controlling the mobility of REE in pore waters.

• characterise the differences in mobility between the light and heavy REE in sediments.

(24)

also applicable to oth er situations. R EE associations w ith clay minerals were n o t as im po rtan t as originally expected in controlling REE co n ten t and fractionation w ithin the m udrocks. M ineral precipitation from pore fluids was found to have a m uch greater influence. The effects o f p H /E h and chemical species on REE m obility were modelled (see C hap ter 3). A lthough there were no pore fluids available for analysis to confirm these predictions, the REE concentrations measured in the m udrocks support the conclusions derived from the modelling. Analysis and leaching of sedim ent samples enabled the differences in m obility of light and heavy REE to be determ ined.

F u rth er to these initial aims, the influence of provenance on the REE patterns of m udrocks was investigated. A correlation betw een REE fractionation and

provenance is clear for m udrocks from b o th the N ankai T ro u g h and the Boso Peninsula (see C hapter 5). The chemical com position of um bers and metalliferous sedim ents from b o th localities was studied, as those from the N ankai accretionary prism are a potential source for the heavy REE enrichm ent at the décollement (see C h ap ter 5).

1.2

Previous work

1.2.1 Nankai Trough mudrocks

O cean D rilling Program (O D P ) Site 808 was drilled in 1990 in the toe of the N ankai accretionary prism (Figure 1). The site was chosen as a good m odern example of a clastic accretionary prism , being dom inated by turbiditic and hemipelagic sedim ents (Shipboard Scientific Party, 1991b). The décollement is shallow w hen com pared w ith oth er prism s (—1000 m bsf), and drilling was able to penetrate this zone. The main them es to be investigated during drilling were:

(25)

(Shipboard Scientific Party, 1991b). A variety of physical and chemical techniques were used in combination to study these aspects of accretionary prism development. Geochemical studies included interstitial pore fluid sampling, along with the

determ ination of major and trace element concentrations in sediment samples. They were correlated with other parameters, to provide inform ation about the nature of fluid flow and diagenetic processes (Shipboard Scientific Party, 1991b).

Pickering et al. (1993a) have already shown that there are a variety of discrete REE signatures in the sediments from the Nankai accretionary prism (Figure 2). The décollement is associated with enrichment of heavy REE when compared with typical mudrocks from the core, despite having similar concentrations of other elements. There are also several suspected umbers between 1060 and 1110 mbsf, which are enriched in all REE (Figure 2), as well as calcium, iron, magnesium, manganese, phosphorus, yttrium , strontium , thorium , scandium, copper and barium

(Pickering er i^/., 1993a).

1000

-,

Background mudrocks Décollement mudrocks Umber

5

8 LU

1

CO

Ê

2 10

-I

La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

Figure 2. C hondrite normalized REE concentrations from the Nankai accretionary prism (data from Pickering 1993a).

(26)

shows REE concentrations for the décollement and um ber (Pickering et al.^ 1993a), normalized to the average of the background mudrocks. Clearly, the um ber is more enriched in heavy REE than in light REE, enhancing its case as a potential source.

14

Décollement mudrocks Umber

2 12

10 i

1

2 8

I

I

r

111 o 111

oc

0

La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

Figure 3. REE concentrations in décollem ent samples and umber, normalized to background m udrocks from the N ankai accretionary prism (data from Pickering et aL, 1993a)

This Study aims to answer the questions posed by the initial work, by determining the nature and possible causes of the anomaly in the décollement zone and

investigating the umbers as a possible source of the heavy REE. There has been no published w ork on the chemistry of the mudrocks sampled on the Boso Peninsula, but they are sedimentologically similar to those of the Nankai accretionary prism.

1.2.2 Review of literature on rare earth elem ents in sedim ents

1.2.2.1 G eneral observations

(27)

They are therefore assumed to be good indicators of sediment provenance (McLennan, 1989). However, other studies indicate that the environments of weathering, transport, deposition and diagenesis may also influence the REE composition of the sediment (e.g. Ronov etal., 1967; Hem m ing et al., 1995; Sholkovitz, 1995).

1000 -I

Post-Archaean Australian Average Shale North American Shale Composite European Shales

c

.9

g cœ

c 1 0 0

-8 H I LU

tr

1

i 2

I

1

o

La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

Figure 4. C hondrite normalized REE concentrations o f average shales from different continents: Post-A rchaean Australian Average Shale (McLennan, 1989), N orth American Shale C om p osite

(G rom et et al., 1984) and European Shales (Haskin and Haskin, 1966).

1.2.2.2 M ineralogy

The REE com position of any rock is related to its mineralogy. Some minerals have higher absolute REE concentrations than others and they may fractionate light or heavy REE. The REE tend to sit in 8-fold co-ordination within the mineral structure, and they may substitute for Ca (McLennan, 1989).

Q uartz generally has low total REE concentrations, with most of the REE being present in inclusions (Taylor and McLennan, 1985). Feldspar also has a relatively low REE content, however enrichment of plagioclase in volcanogenic sediments may result in a positive Eu anomaly relative to chondrite (McLennan, 1989). Carbonate minerals, such as calcite and dolomite, also have low REE totals,

(28)

shales (e.g. R onov e t 1967; G irin etal., 1970), but Piper (1974) found that they are slightly depleted. Clay minerals have high to tal REE concentrations, w ith similar patterns to average shale (Taylor and M cLennan, 1985). T hey contribute m ore to the REE com position o f sedim entary rocks than any o th er rock-form ing minerals. Cullers eta l. (1975) suggest that REE distributions vary betw een clay mineral types, bu t this has n o t been confirm ed by o th er w orkers.

The oth er main locations for REE in sedim ents are heavy minerals and iron and manganese oxyhydroxides. H eavy minerals, such as zircon, m onazite, sphene and allanite, have high total REE concentrations (Taylor and M cLennan, 1985;

M cLennan, 1989; G otze and Lewis, 1994). These minerals may be concentrated during sedim entary sorting processes, resulting in higher than expected total REE concentrations in the sedim ent. How ever, each m ineral has a distinctive REE pattern, making it possible to identify the presence of the heavy mineral from the REE p attern of the whole rock (M cLennan, 1989). F o r example, allanite has high light REE concentrations, zircon has high heavy REE concentrations, and m onazite has high m iddle REE concentrations w ith a negative europium anomaly, when com pared to average shales. Iro n and manganese oxyhydroxides may coat o th er minerals in a sedim ent (Cullers etal., 1979). T hey have high total REE

concentrations relative to average shales, due to adsorption to the oxyhydroxide surfaces (Cullers etal., 1979; G otze and Lewis, 1994). Piper (1974) found them to be enriched in Ce.

1.2.2.3 Rock type and size fraction

Rock type influences REE content as a direct result of the relationship w ith mineralogy. Arenaceous rocks are depleted in R EE as they contain a high

p ro p o rtio n of quartz (e.g. 'KonoY et al., 1967; G o tze and Lewis, 1994). H ow ever, if a sandstone contains a high p ro p o rtio n of heavy minerals, the total REE

concentration will be higher, and either heavy or light REE will be enriched,

(29)

clays, b u t slightly lower due to dilution w ith o th er minerals (Taylor and M cLennan, 1985).

The location of R EE w ithin a m ineral and the solubility of the mineral determ ine w hether the R EE are likely to be available for dissolution in a particular fluid. Those enclosed in the structure of heavy minerals and quartz are unlikely to be released in to pore fluids. REE adsorbed to clay surfaces o r mineral coatings and those held in soluble minerals may be available to pore fluids during diagenesis. In the N o rth American Shale C om posite 33% of the heavy REE are contained in heavy minerals. A fu rth er 10% of heavy and 20% of light REE are either adsorbed to surfaces or present in a soluble phase (G rom et etal., 1984).

Some studies on the different size fractions of sedim entary rocks indicate th at small particles have higher total REE concentrations than large particles, independent of m ineralogy (Cullers etal., 1979; Taylor and M cLennan, 1985). This may be because smaller particles have a higher surface area to volum e ratio, giving them a larger area per unit weight to adsorb REE.

1.2.2.4 Provenance

M ost sedim ents preserved in the rock record have similar REE patterns to average shales (see Section 1.2.2.1). This results from an extended recycling histo ry (A ndré

etal., 1986; M cLennan, 1989). H ow ever, there are locations where sedim ent

recycling is inefficient and one lithology may dom inate the provenance, resulting in different R EE patterns. This usually occurs at volcanically active tectonic settings (M cLennan, 1989).

A t active margins, the expected REE p attern in sedim ents is th at of undifferentiated volcanic rocks, w ith low La/Sm and La/Yb ratios and no negative europium anom aly

(30)

input of sedim ents, either from old basem ent rocks (e.g. Japan) o r from the exposed plutonic roots of a m ature arc (e.g. the A leutians). T he REE patterns at such

settings may also be distorted by mixing, rew orking and sorting of sedim ents, masking the provenance signature (Bock et aL^ 1994).

Passive margins usually receive their sedim ent supply from the continental interiors, and are therefore expected to have similar REE patterns to average shales. H ow ever, they are o ften m ore variable than expected, as in the Precam brian to Cam brian shales of the C entral Iberian Zone, Spain (U gidos etal.^ 1997). Back arc basins generally have REE patterns betw een those of active and passive margins, w ith sedim ents derived from a m ixture of arc material and continental crust (M cLennan

etal.^ 1990).

There have been several studies that have tested these theories against sedim ents of know n provenance, such as Floyd and Leveridge (1987) in C ornw all and Williams et al. (1996) in the Southern Uplands. In b o th cases, the results were as predicted.

How ever, there are still to o few studies in to o few areas for the results to be generally accepted for all settings (M cLennan, 1989).

1.2.2.5 Diagenesis

T he effects of diagenesis are difficult to distinguish from provenance signals in REE patterns. Some authors (e.g. M cLennan, 1989) consider the effects to be

insignificant at low tem peratures. O th ers (e.g. Schieber, 1988; M ilodow ski and Zalasiewicz, 1991) have found diagenetic effects to be im portant w hen large prop o rtio n s o f clay and soluble minerals are present.

Pore fluid chem istry influences the m obility and fractionation of R EE during diagenesis. H eavy REE form stable aqueous complexes w ith carbonates and natural organics (M ilodowski and Zalasiewicz, 1991), b u t light REE complexes are m uch less stable. H eavy REE also form hydroxides m ore easily, leading to further fractionation. In general, REE m obility in pore fluids increases from La to Lu

(31)

Small-scale m ovem ent of REE during the form ation of concretions has been described by Schieber (1988) and M ilodowski and Zalasiewicz (1991). A patite concretions act as a sink for any mobile REE, and they are particularly enriched in heavy REE (M ilodowski and Zalasiewicz, 1991). D iagenetic lim estone tends to be enriched in light REE due to the removal of heavy REE in fluids (Schieber, 1988).

1.2.2.6 Problems

M any of the conflicts of opinion betw een authors arise from examining processes at different scales. M cLennan (1989) examined large scale processes, such as sedim ent transport, so REE do n o t appear mobile. M ilodowski and Zalasiewicz (1991) looked at the small scale processes of diagenesis and nodule form ation in which REE are noticeably mobile. It is therefore very difficult to draw any general conclusions th at w ould hold in m ost, or all, situations.

O th e r problem s w ith studies on REE in sedim ents arise from differences in the analytical techniques used. N e u tro n activation techniques analyse all REE in the sedim ent, including those in heavy minerals such as zircon. H ow ever, acid digestion techniques used to prepare samples for IC P-M S analysis may n o t dissolve all heavy minerals. T he R EE they contain will n o t be counted in the results (Condie, 1991). O p en beaker digestion will be incom plete w hen heavy minerals are present. Lithium m etaborate fusion is better, but is still incom plete. It is therefore difficult to

determ ine if some o r all of the heavy minerals have been digested, leading to uncertainty in the results.

(32)

1.2.2.7 Conclusions

REE in sedim entary rocks from around the w orld all have similar chondrite

norm alized signatures, which closely m atch those of the average upper crust. Small differences betw een these signatures can be used to derive inform ation about provenance and environm ental conditions. A uthors disagree about the extent to w hich each process contributes to the overall distribution of REE, but this is often the result of conducting studies on vastly different scales.

M ineralogy has the greatest control on REE signatures, w ith quartz and carbonate being depleted in REE, clays being similar to average shale, and heavy minerals and oxyhydroxides being enriched. M inerals may also fractionate light and heavy REE, and control which REE are available for tran sp o rt in solution. Pore fluid chem istry dictates which REE are taken into solution in the presence of a particular mineral. Analytical techniques for determ ining REE content and location may be

problem atic, and results presented in the literature m ust be viewed w ith these problem s in mind.

A lthough the quantity of literature regarding R EE in sedim ents is large, there are problem s w ith the techniques and approaches used in many of the investigations. Results cannot often be com pared w ith similar studies, as different analytical techniques have been used. Studies on different scales are also difficult to compare, and there has been very little w ork at scales betw een a few centim etres and a few kilom etres. T he small scale processes m ust be linked to large scale features, and the possibility of mobile REE should be taken in to account w hen using REE as

(33)

Chapter 2.

Sample preparation and analytical techniques

2.1

Summary of techniques

Field and O D P samples were processed and analysed using a variety of techniques to determ ine their mineralogy and geochem istry. The size o f the individual O D P samples was lim ited to 10 ml, and this influenced the am ount of w ork that could be done.

Polished th in sections of field samples were used to determ ine mineralogy optically and using an electron m icroprobe. H ow ever, O D P samples were n o t large enough for this, and small am ounts were m ounted and polished for electron m icroprobe analysis. Pow ders were analysed using X -ray diffraction (XRD) to determ ine clay mineralogy.

The samples were analysed using inductively-coupled plasma atom ic emission spectroscopy (ICP-A ES) for m ajor and trace elem ents, inductively-coupled plasma mass spectrom etry (ICP-M S) for REE, uranium and thorium , and atomic

absorption spectroscopy (AAS) for potassium , after they had been crushed and digested in acid. Selected O D P samples were sequentially leached to determ ine the location of various elem ents, particularly the REE.

2.2

Sample preparation

2.2.1 Drying and crushing

All samples from the O D P cores and those collected in the field were processed in the same way. Initially, they were dried for one week at 50 °C. A t higher

(34)

w ith this technique, due to the straight crushing action involved (Fairchild et al.^

1988). Each sample was then crushed for 30 s in a Tem a® disk mill w ith an agate barrel, resulting in a fine pow der (grain size —40 fxm). Silicon and lead are possible contam inants from agate, b u t neither elem ent was used in this study. The agate surfaces were thoroughly cleaned betw een samples, by washing in w ater followed by acetone. Finally, the pow ders were dried for a fu rth er 24 hours at 50 °C and stored in clear plastic tubs.

2.2 .2 Add digestion

Acid digestions were perform ed to allow the samples to be analysed by IC P-M S and IC P-A ES. F o r each sample, 0.2 g of pow der was digested in an open PT FE beaker. Before use, the beakers were washed, rinsed three times in deionized H jO (18 M il), soaked in 5% Aristar™ H N O3, and finally rinsed a further three times in deionized H2O (18 M H ). Aristar™ acids were used th ro u g h o u t the digestion procedure.

C o n cen trated

HNO3

(1.5 ml) was added to each sample and evaporated to dryness on a heated sand-bath, to oxidise any organic m atter present. This was followed by fuming the sample w ith 12 ml H F (40%) and 2 ml

HCIO4

for 2 to 3 hours, until it had reduced to a gel-like mass. This stage decom poses silicates and carbonates, w ith silicon being rem oved as the volatile silicon tetrafluoride complex. T he

HCIO4

ensures oxidising conditions during this process. The residue was then digested in 3 ml concentrated

HNO3

and 18 ml deionized w ater (18 M fl) for 2 hours. The samples were filtered directly in to cleaned volum etric flasks using W hatm an N o . 42 filter papers, made up to exactly 100 ml and stored in H D P E bottles.

2.2.3 Clay mineral separation

(35)

(18 M fl). These were shaken and placed in an ultrasonic hath for 10 m inutes, before being left to settle overnight.

T he top 25 ml of fluid was rem oved from each measuring cylinder using disposable pipettes, and placed in centrifuge tubes. These were centrifuged at 4000 rpm for 10

minutes, and the supernatants rem oved and discarded. The residue fine fractions were dried at 50 °C for a week, and then disaggregated using an agate pestle and m ortar. This was washed and cleaned w ith acetone betw een samples.

2.3

Leaching experiments

Selected m udrock samples from the O D P cores were leached sequentially using a m ethod based on the procedure described by Lewis and M cC onchie (1994). The five leaching stages were designed to extract soluble salts, exchangeable and adsorbed ions, carbonates, reduced species and oxides. The residues, which were mainly silicates, were then digested in acid for analysis as outlined in Section 2.2.2.

Similar chemical leaching procedures have attracted m uch criticism (e.g. Sholkovitz, 1989), as the leaching solutions may n o t extract exactly the phases expected, and there is a risk of re-adsorption to mineral surfaces which may result in REE fractionation. A higher solution to rock ratio than th at suggested by Lewis and M cC onchie (1994) was used here to help minimise these factors. T he problem s associated w ith the extractions and their influence on the results are discussed in Section 3.4.

2.3.1 Method

T he extractions were conducted in polypropylene centrifuge tubes, to minimise sample loss betw een each stage. Each tube contained 2 g o f sample and was

(36)

leaching solution by centrifuging at 1 0 0 0 0 rpm , 2 0 °C for 30 m inutes. The

supernatant was removed using a pipette, and processed for analysis as described at each stage below. The residue was then rinsed w ith 2 0 ml deionized w ater (18 M il), centrifuged at 10 0 0 0 rpm, 2 0 °C for 30 m inutes and this second supernatant was discarded.

2.3.1.1 Stage 1

The sample in each centrifuge tube was agitated in a shaking w ater b ath w ith 2 0 ml deionized w ater (18 M-Q) for 15 m inutes at 20 °C. The supernatant was then boiled dow n to near dryness, and made up to exactly 25 ml using 5%

HNO3.

This stage should leach all soluble salts.

2.3.1.2 Stage 2

The solid residues from stage 1 were agitated in a shaking w ater bath w ith 2 0 ml 0.1 M BaClg/N H ^Cl solution for 1 ho u r at 20 °C. The supernatant from each sample was then boiled down to near dryness, and made up to exactly 25 ml using 5%

HNO3.

This should remove all exchangeable and adsorbed ions (Gillman and Sum pter, 1986).

2.3.1.3 Stage 3

(37)

2.3.1.4 Stage 4

The solid residues from stage 3 were agitated in a shaking w ater hath w ith 2 0 ml o f a solution of 30 gl'^ sodium dithionite and 90gl'^ citric acid for 12 hours at 2 0 ®C. C o n cen trated H N O3 (5 ml) and hydrogen peroxide (5 ml) were added to the supernatant, which was then boiled dow n to near dryness on a h o t plate. Each sample was made up to exactly 25 ml using 5%

HNO3.

This stage should remove all remaining oxides (H olm gren, 1967).

2.3.1.5 Stage 5

The residues from stage 4 were rem oved from the centrifuge tubes and placed in PTFE beakers. A fter drying, 15 ml aqua regia (5 parts concentrated

HNO3

to tw o parts concentrated H C l) was added to each sample. These were placed on a hot plate to boil fo r 15 m inutes, before being diluted w ith 10 ml deionized water

(18 M O) and returned to the centrifuge tubes. The supernatant was boiled down to near dryness and made up to exactly 25 ml w ith 5%

HNO3.

This stage should remove all sulphides and noble metals (Fairchild et al.y 1988).

The residues from stage 5 were dried and then weighed. A 0 .2 g sample of each was digested in an open PT FE beaker using m ethod described in Section 2.2.2.

2.4

Sample analysis

2

.

4.1

ICP-AES Analysis

All acid digestions and leaching solutions were analysed for m ajor and trace

(38)

D rift samples containing all the elem ents of interest were run after every fifth sample to correct for instrum ent drift during each run. Several acid blanks and reference standards (SGR-1 (shale) and NBS1633a (fly-ash)) were also run. The blanks were below the detection limits and reference standards were in good agreement w ith published data (Govindaraju, 1989).

2.4.2 ICP-MS Analysis

All acid digestions and leaching solutions were analysed for the rare earth elem ents, uranium and thorium using a V G Elem ental Plasm aQ uad inductively-coupled plasma mass spectrom eter at the U niversity of Bristol. Each sample was spiked w ith a 100 ppb Re and Ru internal standard to correct for instrum ent drift, by adding 5 ml of 200 ppb Re and Ru in 5% H N O3 to 5 ml digested sample. N o further dilutions o f acid digestions o r leaching solutions were necessary, as concentrations fell w ithin the calibration limits. Results are an average of three replicate runs.

Several acid blanks and reference standards (SGR-1 (shale) and NBS 1633a (fly-ash)) were also run. T he blanks showed no contam ination w ith REE, U o r Th, and

reference standards were in good agreem ent w ith published data (Govindaraju, 1989).

2.4.3 AAS Analysis

Acid digestions were analysed for potassium using a Varian SpectrA A -20 atom ic absorption spectrom eter, at the U niversity of N o ttin g h am . These results were used in preference to the IC P-A ES analysis of potassium , which had unacceptably large errors.

2 .4 .4 XRD and Electron Microprobe Analysis

(39)

S-2500 scanning electron microscope, to examine accessory mineral phases and their influence on the anomalous REE levels in some m udrocks.

2.4 .5 Estimation o f Errors

(40)

Chapter 3.

Results from the Nankai Trough

3.1

Introduction to the Nankai Trough and ODP Site 808

The Nankai Trough and accretionary prism lie to the southeast of Japan, and mark the boundary between the Shikoku Basin and the Southwest Japan Arc (Figure 5). The Shikoku Basin forms part of the Philippine Sea Plate, and is being subducted beneath the Southwest Japan Arc (Eurasia Plate) at a rate of between 2 and 4 cm a'^ (Shipboard Scientific Party, 1991b).

Scale (km)

300

JAPAN SEA

Izu-Hon^u Collision Zone

PACIFIC OCEAN

Site 808 ^

Shikoku

Basin Key

^ P l a t e boundary

M Relative plate motion = = Relict spreading centre

Figure 5. Map o f southern Japan, show ing the locations o f the Shikoku Basin, N ankai Trough and plate boundaries.

The Nankai Trough is relatively shallow, with a w ater depth of less than 4900 m. This is due to the young age of the Shikoku basin and the thick pile of sediments in the trench (Shipboard Scientific Party, 1991b). Sediments are being actively accreted in the accretionary prism, as the Pacific Sea Plate is subducted towards the

(41)

The Shikoku Basin formed as a backarc basin behind the Izu-Bonin Arc, from the Oligocene to the middle Miocene (Shipboard Scientific Party, 1991b). A spreading centre was active in an east-west direction at the centre of the basin until 15 Ma. There is some uncertainty as to w hether the spreading ceased after 15 Ma (O kino,

1994), or continued in a north-south direction until 12 Ma (Cham ot-Rooke et al.,

1987). The relict spreading centre collided with the N ankai accretionary prism at about 15 Ma (Pickering et a i, 1993b), and is currently being subducted close to O D P Site 808 (Figure 5; Shipboard Scientific Party, 1991a).

3.1.1 Location of ODP Site 808

O D P Leg 131, Site 808, was drilled in the toe of the Nankai accretionary prism (Figure 6). It was located on the landward trench slope, less than 150 m above the trench floor. Drilling reached a depth of 1327 metres below sea floor (mbsf), penetrating the frontal thrust, décollement zone and the basaltic oceanic basement

(Shipboard Scientific Party, 1991a). The proximity of the relict spreading centre affects the thermal regime of the subducting sediments (Wanger^z/., 1995). Fieat flow is higher than that measured in other accretionary prisms, affecting the rate of sediment diagenesis.

Site 808

SL

4 9 00 m

/ / Décollement

Compiled using data from M o oref a/ (1991) and Shipboard Scientific Party (1991)

Fault

Sediments Scale (km)

Basement

(42)

3.1.2 Stratigraphy at ODP Site 808

The N ankai accretionary prism is a elastic-dominated accretionary prism (Taira and Pickering, 1991). The thick trench fill shows an overall coarsening upward sequence, typical of subduction systems. The oldest recovered sediments are hemipelagic muds, which gradually become interbedded with turbidites and then coarsen up into entirely terrigenous sands (Pickering etal., 1993b). The stratigraphy at Site 808 can be divided into six discrete lithological units, two of which have been further divided in to subunits (Figure 7). There is some repetition of subunits Ilb and lie across the frontal thrust.

AGE DEPTH LITHOLOGY

Pliocene

200 mbsf

400 mbsf

600 mbsf

800 mbsf

1000 mbsf

1200 mbsf

I Slope apron deposits

lia Upper axial trench wedge

lib Lower axial trench wedge Frontal thrust

lie Marginal trench wedge Trench to basin transition

IVa Upper Shikoku Basin deposits

Décollement zone

IVb Lower Shikoku Basin deposits

V Acid volcaniclastic deposits V I Igneous basement

Figure 7. Stratigraphy o f the Nankai accretionary prism at Site 808C. C om piled using data from Shipboard Scientific Party (1991a), Pickering et (1993b) and U nd erw ood (1993a).

U nit I comprises the slope apron, and is made up of thin sand and mud debris flows, sediment slides and turbidites. It is about 20 m thick (Pickering et al., 1993b;

U nderw ood 1993a).

(43)

sedim ents originated from the Fuji River and were redeposited from sedim ent gravity flows that tran sp o rted sedim ents th ro u gh the Suruga T rough to the N ankai T rough (U nderw ood etal.^ 1993a; U nderw ood and Pickering, 1996). The

sedim entation rate ranged from 785 to 1380 m Ma"^ (Shipboard Scientific Party, 1991a), and may have been as high as 2000 m Ma"^ at the time of arc-arc collision (Taira and Pickering, 1991). U nit II is divided in to three subunits. The youngest subunit, Ila, form s the upper axial trench wedge. It is com posed of thick coarse­ grained sandy turbidites. Subunit lib , forms the lower axial trench wedge, and comprises th in finer-grained sand and silt turbidites, interbedded w ith hemipelagic muds. Subunit lie , the marginal trench wedge, comprises silt turbidites interbedded w ith hemipelagic muds (U nderw ood et al., 1993a).

U n it III marks the transition betw een trench and basin deposits. Hemipelagic muds are the dom inant sedim ents, but there are also silt turbidites as in subunit lie and volcanic ash similar to th at found in subunit IVa (U nderw ood etal., 1993b).

U nit IV comprises the hemipelagic deposits of the Shikoku Basin. It is divided in to tw o subunits. Sedim entation rates were m uch slower than in the trench wedge, and varied betw een 45 and 110 m Ma'^ (Shipboard Scientific Party, 1991a). Subunit IVa comprises volcanic ash and tuff, interbedded w ith hemipelagic muds. Subunit IV b is com posed alm ost entirely of bioturbated hemipelagic muds (U nderw ood et al.,

1993a). H ow ever, there are some m etalliferous sedim ents and um bers lower in the sequence, interpreted as being due to the proxim ity o f the relict spreading centre

(Pickering et al., 1993a). The décollem ent zone is contained w ithin subunit IVb, and is n o t associated w ith any change in lithology (Figure 7).

The acid volcaniclastic deposits of U n it V are com posed of siliceous tuff,

interbedded w ith variegated m udstone. Below this lies U n it VI, the basaltic igneous

(44)

3.1 .3 Hydrogeology

M ost of the evidence for pore fluid circulation in the N ankai accretionary prism comes from variations in pore fluid chem istry (e.g. Gieskes etal.^ 1993; K astner et al., 1993; Y ou etal., 1993). M ost of these variations are related to the mineralogy of

the sedim ents and diagenetic changes (U nderw ood etal., 1993b). There are

noticeable changes associated w ith lithological boundaries, for example, betw een the trench and basin deposits (U nderw ood and Pickering, 1996). This indicates that there is chemical exchange betw een the pore fluids and the sediments.

The m ost striking feature of the pore fluid chem istry is an area of low chloride concentration below the décollem ent, betw een 1040 and 1080 m bsf (Shipboard Scientific Party, 1991a; K astner etal., 1993). A lthough the transform ation of sm ectite to illite results in the form ation of less saline pore fluids, there has n o t been enough clay mineral diagenesis at Site 808 to account for the measured reduction (U nderw ood etal., 1993b; U nderw ood and Pickering, 1996). The low chlorine pore fluid m ust therefore have m igrated from deeper w ithin the prism. It may have been injected along the décollem ent zone around 300 000 years ago, and then m igrated slowly dow n to its current location (Shipboard Scientific Party,

1991a). Alternatively, it may have m igrated upw ard th ro u g h the sedim ents below the décollem ent (K astner et ^ï/., 1993).

Figure

Figure 1. Map of Japan showing location of sampling sites, O D P Site 808 and the Boso Peninsula.
Figure 2. Chondrite normalized REE concentrations from the Nankai accretionary prism (data fromPickering 1993a).
Figure 25. Concentration of strontium (ppm) removed during each stage of the sequential leaching experiments
Figure 27. Concentration of manganese (ppm) removed during each stage of the sequential leaching experiments
+7

References

Related documents

Interaction means ± SE (error bars) for the effects of species (cow, goat) and month (in order of appearance from left to right in Axis X: M, March; A, April; M, May; J, June; J,

Bucknell, sank a career-high four three-pointers and also had five rebounds...has at least five boards in eight games...has made 25 three pointers this season and is second on the

den bzw. weiterhin ausgefüllt werden. Die Auswertung in Kapitel 6 beinhaltet zudem die zusammengefassten Evaluationsergebnisse der Soft Skills Blockseminare, die im Rahmen

The risk categories, credit risk, market risk and liquidity risk are tested by some single factor sensitivity stress tests to assess the resilience of the banks.. In each

In this paper, Beijing Telecom 3G for me study,by integrating theory with practice,Focusing on how to adjust the development strategy of Beijing Telecom 3G realize market

The increase in liver glycogen level was higher in rats administered with hexane and n-butanol fractions of the plant compared with rats given the ethyl acetate fraction..

Bio-Rad Protein Standard I (bovine gamma globulin) was used for the Lowry and Bio-Rad Protein Assays, with bovine serum albumin was used for the Biuret method.. All three assays

As health care entities manage their businesses, they enter into various types of agreements with technology vendors to support the provision of care, including: software