Examensarbete vid Institutionen för geovetenskaper
Degree Project at the Department of Earth Sciences
ISSN 1650-6553 Nr 424
Characterizing Magnetic Susceptibility and Remanent Magnetization of Magnetite and Hematite Rich Drill-
Core Samples at Blötberget
Karaktärisering av magnetisk susceptibilitet och remanens hos magnetit och hematit- rika borrkärnor från Blötberget
Andreas Björk
INSTITUTIONEN FÖR
Examensarbete vid Institutionen för geovetenskaper
Degree Project at the Department of Earth Sciences
ISSN 1650-6553 Nr 424
Characterizing Magnetic Susceptibility and Remanent Magnetization of Magnetite and Hematite Rich Drill- Core Samples at Blötberget
Karaktärisering av magnetisk susceptibilitet och remanens hos magnetit och hematit- rika borrkärnor från Blötberget
Andreas Björk
ISSN 1650-6553
Copyright © Andreas Björk
Abstract
Characterizing magnetic susceptibility and remanent magnetization of mag- netite and hematite rich drill-core samples at Blötberget
Andreas Björk
Laboratory magnetic measurements are used to develop a methodology to characterize the Kiruna-type Rare Earth Elements (REE) bearing apatite iron-oxide deposits at Blötberget in central Sweden. This high-grade ore deposit is known to have sharp boundaries between lens shaped main ore bodies of magnetite-rich ore, and a complex hematite-rich ore associated with pegmatites and skarn formation.
The thesis covers laboratory magnetic measurements of 37 samples originating from eight drill cores and reference samples from previously mined area. It focuses on on-covering how the samples relate in terms of magnetic susceptibility, further its dependency on temperature, frequency, field and the orienta- tion. The results are correlated with petrographic analysis previously performed on accompanying thin sections.
The measurements show that magnetite with strong susceptibility contribution overshadow the hematite contribution in the samples. Transition changes in susceptibility are noticeable when crossing the Ver- wey temperature; -153°C, Curie temperature; 580°C and Néel temperature; 680°C. The Morin tempera- ture appears at -60°C, or is missing. Linear relationships are identified between the magnitude difference in susceptibilities across transitions at high temperature and wt% magnetite and hematite have been identified. The Blötberget skarn and hematite-rich ore samples have a higher degree of susceptibility anisotropy than the other ore-types.
Blötberget samples are dominated by multidomain characteristics in remanence, saturation and coerciv- ity. High temperature measurements have shown that the magnetite is close to pure. The low temperature measurements suggest hematite is impure or bears a petrological footprint. The study also shows that rich iron ore samples sometimes can be at risk of being overlooked with standard methods of measuring susceptibility.
Keywords: Verwey, Morin, Curie, Néel, magnetite, hematite, alteration, iron-ore, quality control, kappabridge, hysteresis, susceptibility, anisotropy
Degree Project E in Geophysics, 1GE029, 30 credits Supervisor: Bjarne Almqvist
Department of Earth Sciences, Uppsala University, Villavägen 16, SE-752 36 Uppsala (www.geo.uu.se) ISSN 1650-6553, Examensarbete vid Institutionen för geovetenskaper, No. 424, 2018
The whole document is available at www.diva-portal.org
Populärvetenskaplig sammanfattning
Karaktärisering av magnetisk susceptibilitet och remanens hos magnetit och hematit -rika borrkärnor från Blötberget
Andreas Björk
Laboratorietekniska metoder kan användas som ett komplement till malmgeologi och geofysisk prospek- tering. I denna metodstudie karaktäriseras apatitjärnmalm från Blötberget, nära Grängesberg. En fyn- dighet bestående av linsformade malmkroppar rika på magnetit och ofta avskilda men komplexa hematitrika stråk. Studien är gjord 37 prover från totalt 8 borrkärnor, och lokaler som tillhörde produktion från gruvverksamhet under 1900-talet. Mätmetoderna fokuserar på att kartlägga malmens magnetiska egen- skaper, och hur temperatur, frekvens, fältstyrka samt riktning påverkar dessa. Resultaten jämfördes med tidigare petrografisk studie av tillhörande tunnslip
Resultaten visar att magnetit står för merparten av susceptibiliteten i proverna, men att även hematit kan urskiljas och kvantifieras. Temperaturberoende har påvisats vid övergångar för Verwey-temperatur; -153
°C, Curie-temperatur; 580 °C, och Néeltemperatur; 680 °C. Den förväntade Morin-temperaturen vid -14
°C, påträffades vid -60 °C eller saknas helt för flera av de hematitrika proverna. Magnetiskt anisotropa prover återfinns bland prover som identifierats som skarn eller hematitrika.
Magnetisk granulometri visar karaktär av multidomäntyp med låg magnetisk coercivitet och hög sat- ureringsförmåga. Högtemperaturmätningar av susceptibilitet visar på ren magnetit för prover från Blöt- berget. Samtidigt visar lågtemperaturemätningar att hematit sannolikt har inblandning av titanium eller bär på ett mer komplext förflutet. Studien visar också att det finns en risk i att enbart förlita sig på bulksusceptibilitet för prover rika på malm.
Nyckelord: Verwey, Morin, Curie, Néel, magnetit, hematit, oxidation, järnmalm, kvalitetskontroll, kappabridge, hysteresis, susceptibilitet, anisotropi
Examensarbete E i geofysik, 1GE029, 30 högskolepoäng Handledare: Bjarne Almqvist
Institutionen för geovetenskap, Uppsala Universitet, Villavägen 16, 752 36 Uppsala (www.geo.uu.se) ISSN 1650-6553, Examensarbete vid Institutionen för geovetenskaper, No. 424, 2018
Hela publikationen finns tillgänglig på www.diva-portal.org
List of Figures
1 a) Simplified geological map of Scandinavia and Finland (Koistinen et al. 2001) showing the locations of the two mining sites that has been studied within the ERA-MIN Start- GeoDelineation project. b) 3D visualization drill holes at Blötberget (Ludvika Mines), which is the focus of the current study. Colour coded localities; Sandel (yellow), Hugget
& Flygruvan (red & magenta) and Kalvgruvan (blue), (Courtesy of Nordic Iron Ore Sweden) c) Distribution of drill holes within the Nordic Iron Ore Sweden exploration license; origin of drill core samples used in this thesis. . . 4 2 Thin sections in Plane Polarized Light (PPL). a) to h) Accompanying thin sections show
high grade ore of disseminated, banded and massive type grains in varying size ranges.
Magnetite (grey) dominates a) to d) with other ferroxides and occasional sulphides. In e) to h) in lesser towards the right and showing varying presence, shape and distribution of hematite (greyish-beige). . . 5 3 Susceptibility and resulting magnetization a) opposing c) parallel, to the applied field.
Susceptibility as a function of temperature b) constant d) proportional to 1/T. Modified after on Lowrie, W. (2007). . . 9 4 Coordination of Fe cations with O−2anions (green) in magnetite. For A-sublattice (tetra-
hedral Fe(III)) in blue and B-sublattice (octahedral Fe(II,III in red). The unit cell of the of the spinel crystal structure is shown by the dashed lines. Modified from Butler, R. F.
(1992b) . . . 10 5 Exchange couplings for ferromagnetism; a) truly ferromagnetic, b) antiferromagnetic,
c) spin-canted antiferromagnetic and d) ferrimagnetic materials. Their net magnetiza- tion ’magnetic moment’ is shown below the coupling of magnetic moment in magnetic minerals. Modified after Butler, R. F. (1992b), and Lowrie, W. (1997). . . 11
6 Sketch of a hysteresis loop; Ms- saturation magnetization; Mrs- Saturation remanence magnetization; Hc- Coercivity; Hcr- Coercivity of remanence. When the back field (ap- plied in the -H direction) is switched off, the original remanent acquired in the positive (H+) direction, after saturation, has been removed. The value of the back field required to nullify the saturation remanent magnetization is the coercivity of remanence (Hcr).
Modified after on Lowrie, W. (2007) . . . 12 7 Sketch of relation between relaxation and grain sizes and shapes for superparamagnetic,
single domain and multidomain for magnetite. Redrawn after McElhinny, 1973. . . 14 8 Ternary compositional diagram of the iron-titanium oxides. . . 16 9 Sketch cation distributions in normal spinel and inverse spinel. A and B are sublattices
and arrows indicating magnetic moments caused by the cations. Redrawn after McEl- hinny (1973) . . . 16 10 Range of susceptibilities for magnetic iron oxides and iron ore. Redrawn and modified
after Clark (1997). . . 17 11 The MFK1-FA laboratory instrument developed by AGICO (Advanced Geoscience In-
struments Company). Picture from high temperature susceptibility (Bjork 2016). . . 21 12 The 7400-S Series vibrating sample magnetometer (VSM) developed by Lake Shore
Cryotronics. During measurement at Ångström Laboratory (Bjork 2016, photo collage). 22 13 a) Project slabs approximately 1 cm high by 2 cm. b) Sawing sample intended for
measurement of magnetic anistropy. c) Sample P20b put aside for measurements of magnetic anisotropy. . . 23 14 Illustration of how the magnetic susceptibility typically changes as a function of tem-
perature in samples from Blötberget. . . 26
15 Illustration of how the magnetic susceptibility changes at low temperatures for a hematite rich sample from Blötberget. The Verwey transition occurs at a temperature (TV) of ca 120 °K (-153 °) at which magnetite experiences a change in crystal structure. Suscep- tibility of hematite is temperature dependent through the Morin transition temperature, which is referred to as -11 °C for pure stochimetric hematite (Morrish 1994). However both temperatures (especially TM) can shift e.g. due to impurities, and in this sample the TM is approximately -60 °C (Morrish 1994; Özdemir 1997). . . 28 16 Sketch over susceptibility dependent on frequency and ratio of grain size. Modified after
Hrouda (2011) . . . 31 17 Measured magnetic susceptibility ’k’ [-]. . . 36 18 Measured bulk susceptibilities (obtained with equation 5). Samples are color themed for
ore body locality and drill hole number or outcrops/claims . . . 37
19 Examples of Curie temperature transitions below 600°C, with decreasing Bulk suscep- tibility∆κ [-]; Samples a) P4 b) P2 c) P38 d) P17. From these graphs Curie temperature has been defined as the average in temperature between that of the peak value before the transition and the first value after it. . . 39 20 Examples of high-temperature transitions between 600°C and 700°C, ordered with de-
creasing∆κ [-]; Samples a) P4 b) P2 c) P38 d) P17. From these graphs Néel temperature has been defined as the average in temperature between that of the peak value before the transition and the first value after it. . . 40 21 Low-temperature transitions; with low and peak value, and bulk susceptibility ∆κ [-];
Verwey transitions: a) P4: high grade magnetite; b) P17: High hematite grade sample;
Verwey and Morin transition: c) Hre fred: small cut cube from a Kidney type hematite, and d) The Island of Elba: A pure hematite crystal for reference . . . 44 22 Low-temperature transitions; Morin, with peak and low value, and bulk susceptibility
∆κ [-]; Samples a) P23 b) P17 c) P20. . . 45
23 Susceptibility at low fields and stepwise change in frequency from 976 to 3904 and 15616 Hz. a) The magnetic susceptibility as measured in lab ’k’. b) bulk susceptibility κ (corrected for volume; eqn. 5) . . . 49
24 Difference in susceptibility measured in frequency 976 and 3904 Hz. a) expressed in percentage. b) expressed as the difference in bulk susceptibility after correcting for volume [-] . . . 50 25 Frequency dependence of a subset of samples all have the same range settings when
measured in frequency 3904 and 15616 Hz. a) expressed in percentage. b) expressed as the difference in bulk susceptibility after correcting for volume [-] . . . 51 26 Flinn diagram showing Magnetic Lineation and Magnetic Foliation in samples . . . 53 27 Jelinek plot; Degree of anisotropy as a function of Bulk susceptibility [-] . . . 54 28 Examples of hysteresis loops at Blötberget, with a) P7: high grade magnetite: b) P7
closup c) P16 d) P16 closup. . . 56 29 Continued examples of hysteresis loops at Blötberget, with a) P25 b) P25 closup c) high
grade hematite P17 d) P17 closup. . . 57 30 Hysteresis curves for reference samples, with a) The Island of Elba: A pure hematite
crystal b) Elba closup c) Langban hematite crystal d) Langban closup . . . 58 32 Bulk susceptibilities κ (obtained with equation 5), for 21 drill core samples, and 3 sam-
ples from outcrops, as a function of vol% magnetite. Legend show sample drill hole No/name or sample outcrop. Note: The bulk susceptibility is plotted as 1st order mag- nitude (i.e. 100) [-] . . . 62 33 Bulk susceptibilites (obtained with equation 5) put into a semilog plot, with 21 drill core
samples, and 3 samples from outcrops, as a function of vol% magnetite. Legend show sample drill hole No/name or sample outcrop. The symbol and colour coding denotes what ore body locality the sample is coming from. . . 63 34 Mass susceptibilities χ, for 21 drill core samples, and 3 samples from outcrops, as a
function of wt% magnetite. Legend show sample drill hole No/name or sample outcrop. 64
35 Mass susceptibility χ put into a semilog plot, with 21 drill core samples, and 3 samples from outcrops, as a function of vol% magnetite. Legend show sample drill hole No/name or sample outcrop. The symbol and colour coding denotes what ore body locality the
sample is coming from. . . 65
37 Change in mass susceptibility χ at high temperature, a) magnetite (∼580 °C), and b) hematite (∼680 °C). Samples are labelled with name and colour coded for ore body designation. Also included is the linear relation between susceptibility and wt%. . . 68
38 Change in temperature dependent susceptibility in sample P17, caused by reduction of hematite to magnetite. Note: data in this graph has not been corrected for the furnace sample holder, approximately -150 E-06 . . . 71
39 Examples of ’humps’ in susceptibility temperature dependency, between 100-575 °C; Samples a) P30 b) P35 c) P36 d) P39. . . 72
40 Examples of ’humps’ in susceptibility dependent in the temperature span between 100- 575 °C, susceptibilities are normalized; Samples a) P19 b) P26. . . 73
41 XR parameter in Blötberget samples. a) XFD(1,16) versus χ1 plot b) XR versus XFD(1,16) plot and c) XFV(4,16)versus XFV(1,4) . . . 76
42 XRparameter in Blötberget samples. XFV(4,16)versus XFV(1,4) . . . 77
45 Micromagnetic property ratio Hcr/Hc in relation to wt% magnetite. . . 80
46 Micromagnetic property ratio Hcr/Hc in relation to wt% hematite. . . 81
47 Linear fit for Magnetic saturation ’Ms’ (normalized by mass) in relation to wt% mag- netite. a) wt% (BGU, 2015) b) estimated wt% derived from temperature susceptibility correlation. . . 82
48 Residual Case Order Plot for Magnetic saturation ’Ms’ (normalized by mass) in relation to wt% magnetite. a) wt% (BGU, 2015) b) estimated wt% derived from temperature susceptibility correlation. . . 83
A1 Show bulk susceptibility normalized by weight χ [cm3g−1]. . . 92
A2 Difference in susceptibility measured in frequency 3904 and 15616 Hz. a) expressed in percentage. b) expressed as the difference in bulk susceptibility after correcting for volume [-] . . . 93 A3 Change in magnetic susceptibility at high temperature, a) magnetite Curie temperature
(580 °C), and b) hematite Néel temperature (680 °C). Note for that for a) sample P17 has been excluded. This increases the fit from R2= 0.54 to R2= 0.62. For b) hematite R2= 0.83. . . 94
Table of Contents
List of Figures iii
1 Introduction 2
1.1 Aim . . . 3
1.2 Sample overview . . . 3
2 Theory 7 2.1 Magnetism . . . 7
2.2 Mineral Magnetism . . . 8
2.3 Micromagnetic properties . . . 13
2.3.1 Domain theory . . . 13
2.3.2 Remanent Magnetism . . . 14
2.4 Iron-titanium oxides and their magnetic properties . . . 15
2.4.1 Titanomagnetite . . . 17
2.4.2 Titanohematites . . . 18
2.4.3 Oxidation of ferromagnetic minerals in continental rock and petrological effects . 18 2.5 Anisotropy of Magnetic Susceptibility . . . 19
3 Measurement equipment 21 3.1 MFK1-FA with semi-automatic Rotator, CS4 furnace and CS-L cryostat . . . 21
3.2 7400-S Series VSM . . . 22
4 Methodology 23
4.1 Sample preparation . . . 23
4.2 Magnetic methods: unless specified measured at 15-25 °C . . . 25
4.2.1 Bulk susceptibility . . . 25
4.2.2 High temperature-dependent susceptibility: 25 to 700 °C . . . 25
4.2.3 Low temperature-dependent susceptibility: -194 to 0 °C . . . 27
4.2.4 Field dependence . . . 28
4.2.5 Frequency dependence . . . 30
4.2.6 Anisotropy of magnetic susceptibility . . . 32
4.2.7 Micromagnetic properties . . . 33
5 Results 34 5.1 Bulk susceptibility . . . 34
5.2 High temperature susceptibility: 25 to 700 °C . . . 37
5.3 Low temperature susceptibility: -194 to 0 °C . . . 43
5.4 Field dependent susceptibility . . . 46
5.5 Frequency dependence measurements . . . 48
5.6 Anisotropy of magnetic susceptibility . . . 52
5.7 Micromagnetic properties . . . 54
6 Discussion 61 6.1 Bulk susceptibility . . . 61
6.2 High temperature susceptibility: 25 to 700 °C . . . 65
6.3 Low temperature susceptibility: -194 to 0 °C . . . 73
6.4 Field dependence . . . 75
6.5 Frequency dependence measurement . . . 75
6.6 Anisotropy of magnetic susceptibility . . . 77
6.7 Micromagnetic properties . . . 80
7 Conclusions 85
8 Outlook 86
9 References 88
10 Acknowledgements 91
Appendices 92
Abbrevations and units
List of abbrevations and units frequently used in this thesis.
Coercivity of remanence Hcr (Hcr) [A/m]
Magnetic lineation L= k1/k2
Magnetic foliation F= k2/k3
Bulk susceptibility κ [-]
Disseminated ore D
Hematite hem
Magnetic field H [A/m]
Magnetite magn
Magnetic moment m [A m2]
Magnetization by mass Ω [A m2/ kg]
Magnetization by volume M [A/m]
Magnetic coercivity Hc (Hc) [A/m]
Massive ore MO
Magnetic saturation Ms (Ms) [Am2]
Mass susceptibility χ [cm3g−1]
Multidomain MD
Pseudo single domain PSD
Saturation remanent Mrs [A m2]
Coercivity of remanence Hcr [A/m]
Saturation remanent magnetization Mrs (Mrs) [A m2]
Single domain SD
Spectrolitic spectr
Total susceptibility kT [-]
Volume vol [cm3]
Frequency dependent susceptibility (low and high frequency) XFD [-]
Frequency dependent susceptibility (between two frequencies) XFV [-]
Frequency dependent susceptibility (comparative ratio factor) XR [-]
1 Introduction
The Blötberget deposit of iron ore bodies contains magnetite and varying degrees of hematite. During the middle to late 1900th century it played an importart role in Swedish iron-production during, and in 2011 Nordic Iron Ore Sweden received rights to prospect the area for re-commencement. The Company has since, together with partners in the ERA-MIN StartGeoDelineation project, made efforts to reveal the vertical extent of the deposit.
The Grängesberg ductile deformation (Eklöf, 2014) suggests that the nearby the host rock at Blötberget also has gone through strong deformation. However petrography and geochemistry from both Gränges- berg (Weis et al 2013) and Blötberget (Jiao, 2011) suggest that the deposit formed mainly through crystallization from a magma. What can still be debated is if the intrusive rock as well as the metavul- canic host rock has gone through green schist or amphibole metamorphism. It could very well have consequences for magnetic remanence (e.g Dunlop, D. & Özdemir, O. (1997)).
In this thesis laboratory magnetic measurements provide the exploration program with new knowledge to drill core samples that previously were subject to thin section analysis. Statistical relationships between magnetic properties and petrography are used as a tool for systematic investigation as to how to make rapid estimates of wt% magnetite and hematite in iron ore. The results further pave the way for targetted research on the Blötberget ore bodies, to explore differences in magnetic parameters such as magnetic remanence, coercivity and degree of anisotropy. An perturbed Morin transition raise questions as to whether the Blötberget hematite hold impurties of titanium or if the cause can be traced to an event in the ore formation.
In a magnetite-rich ore it is expected that the magnetite should be the main contributor to the suscepti- bility by orders of magnitude. This thesis characterizes and quantifies magnetite and hematite in mixed iron ore samples. Techniques are used that overcome the limitations of hand held susceptibility magne- tomers where susceptibility of magnetite overshadow other magnetic minerals by orders of magnitude.
Laboratory magnetic measurement that previously have proved useful in paleomagnetism may effec- tively complement ore geology and geophysical studies. Developed practices in this work can be used to for targetting further research or implemented as part of future ore analysis to maintain ore grade and quality in production.
1.1 Aim
At present there are few studies on magnetic samples with more than one ferromagnetic mineral occuring together. This thesis seeks to address this research gap by using a set of high grade ore samples that are rich in such minerals. It uses several research methods based on theory developed by reputable scientists who have studied magnetic properties. Thus, it includes:
a) a petrophysical study on iron-oxides at Blötberget to characterize magnetic minerals in support of exploration.
b) a correlation exercise using previous results from an independent petrographical study of vol% and wt% which are based on microscopy point counting provided by (BGU, 2015). It estimates the vol%
and wt% of other previously unknown samples by applying a linear trend with additional magnetic susceptibility measurements.
1.2 Sample overview
The Grängesberg area is dominated by fine grained leucocratic metamorphic rock composed mainly of quartz and feldspar and thought to be of volcanic origin. Another name for it is ’leptite’, which in this case also holds darker minerals and is relatively rich in CaO. Agglomerates are common and may have eased impregnation of hematite and formation of skarn. At Blötberget the magnetite has been considered to be of apatite type. The overall strike of the deposit in the area is NE to SW. The two main ore bodies Kalvgruvemalmen and Flyggruvemalmen have traditionally been defined as close to solely containing magnetite and hematite ore respectively (Figure 1a). In these ore-bodies skarn occur as slivers within the magnetite-rich leptite, while the hematite concentrated within cross cutting pegmatites. The ores that are more of impregnation type have been judged as more complex (Geijer and Magnusson, 1944).
(a) (b)
�
�
��
� �
�"BB12008"
"BB12007"
"BB12006"
"BB12005"
"BB12004"
BB 73-246"
"BB 73-233 and BB 73-234
0 0.2 0.4 0.8Kilometers
(c)
Figure 1: a) Simplified geological map of Scandinavia and Finland (Koistinen et al. 2001) showing the locations of the two mining sites that has been studied within the ERA-MIN StartGeoDelineation project. b) 3D visualization drill holes at Blötberget (Ludvika Mines), which is the focus of the current study. Colour coded localities; Sandel (yellow), Hugget & Flygruvan (red & magenta) and Kalvgruvan (blue), (Courtesy of Nordic Iron Ore Sweden) c) Distribution of drill holes within the Nordic Iron Ore Sweden exploration license; origin of drill core samples used in this thesis.
For this study material were collected from five new drill cores, with samples down to 500 m drill
depth (Table B5), three older drill cores, blast samples from the open pit and reference samples from eight hand samples gathered from adjacent mining fields (see arcs in Geijer and Magnusson, 1944).
Traditional names are used for the ore bodies; Hugget, Flygruvan, Kalvgruvan, Sandel, Guldkannan, Fly-Kalv. For hand samples, sometimes the names of the claim is used directly, such as Mossgruvan and Buskgruvan. The provided set contained 37 slabs of approximately 3.0 × 2.0 × 0.3 cm. Accompanying thin sections have been studied together with the magnetic properties (Figure 2). Reflected-light show high grade ore, from massive and banded to disseminated type. Magnetite dominates with varying presence of crystalline and secondary hematite. According to Jiao (2011) Blötberget samples have in polarized light a matrix that is varying intermediate metamorphed mafic, felsic and metasedimentary minerals. Furthermore brittle deformation includes fissures, shear-planes, and brecciation.
50 μm 100 μm
100 μm 500 μm 500 μm 500 μm
500 μm 500 μm
500 μm
a) b) c) d)
e) f) g) h)
Figure 2: Thin sections in Plane Polarized Light (PPL). a) to h) Accompanying thin sections show high grade ore of disseminated, banded and massive type grains in varying size ranges. Magnetite (grey) dominates a) to d) with other ferroxides and occasional sulphides. In e) to h) in lesser towards the right and showing varying presence, shape and distribution of hematite (greyish-beige).
Prior to this work, a mineralogical study using the same sample set was completed by Berg och Gruvun- dersökningar AB in 2015. The results from that study are used here for comparison and evaluation of laboratory magnetic measurements (Sample list and summary of previous studies are shown in Table 1).
Table 1: Sample list and data from report by Berg och Gruvundersökningar AB in 2015. The volumetric percentage data is calculated from point counting performed during optical microscopy. These results serve as a basis for discussion and comparison with measured bulk susceptibility data. The observed banding in the samples were also used in interpretation of magnetic anisotropy. Abbreviations: Massive
’M, Disseminated ’D’, volume ’vol’, weight ’wt’, magnetite ’magn’, hematite ’hem’, martite ’mart’, spectrolitic ’spectr’. Banding (x) denotes strong foliation or banding for non-ore minerals.
Name Grade vol% vol% vol% vol% wt% wt% wt% wt% banded
magn hem mart spectr magn hem mart spectr
P1 M 49.4 1.3 62.2 1.6 x
P2 M 55.6 30.6 3.3 58.2 32 3.5 (x)
P3 D 32.1 12.8 1.6 42 16.8 2.1
P4 M (complete) 70.8 81.4
P5 D (rich) 48.3 63 (x)
P6 D (poor) 11.9 19.9 x
P7 D (rich) 65.5 77.4 (x)
P8 D
P9 M
P10 D
P11 D (rich)
P12 M
P13 D (x)
P14 D (poor) (x)
P15 M 30.3 0.2 2.6 43 0.3 3.8
P16 M 25.2 56.6 5.8 26.6 59.7 6.1
P17 M 3.6 76.6 2.2 3.9 82.9 2.4
P18 Skarn-iron-ore 15.6 3.2 25.5 x
P19 Skarn-iron-ore 49.0 64.5
P20 Skarn-iron-ore 17.6 31.0 1.4 22.5 39.7 1.8 (x)
P21 50.3 2.6 1.3 63.2 3.3 1.6 x
P22 56.6 11.3 2.4 65.2 13 2.8
P23 23.1 46.3 2.8 26.3 52.7 3.2 (x)
P24 19.3 30.5
P25 M 10.7 21 19 13.8 27.2 24.6
P26 32.6 0.8 44.1 1.1 1.1 (x)
P27 D (very sparse) P28
P29
P30 D (very poor)
P32 D (M) 55.8 3.8 0.9 67.5 4.6 1.1
P34 M
P35 M
P36 M
P37 D (rich) P38 D (rich)
P39 M
2 Theory
2.1 Magnetism
This second chapter summarizes the magnetic methodology used in this work. When comparing this thesis to other work it is important to be able to separate units and general concepts of magnetism. If need be, there are numerous reputable reviews on the definitions of Faraday’s law, the concepts used to describe magnetic properties of minerals and rocks. Work by Butler (1992b), Lowrie (1997) and Tauxe et al (2016) are used as foundational knowledge for this study. Static magnetic fields emphasizes two concepts, which uses different units: 1. the unit system (Systeme Internationale), and 2. The CGS:
system of units (centimeters, gram, seconds). In this thesis measurement data in CGS system (for theory see Butler 1992b and Tauxe et al., 2016) has been converted to SI (Table 2).
The magnetic moment (m) is a vector: the direction and magnitude that results from the magnetic field, that is generated by an electrical loop or a dipole. The magnetization M is the sum of all magnetic mo- ments within the volume of a substance. The ability, or the magnetic permeability, to which a substance magnetize is named the magnetic flux. In SI units the magnetic flux of air is unit-less. The magnetic flux in a substance is µ= (1 + k), where k is unit-less proportionality factor named magnetic susceptibility.
When a field H is applied to a substance the result is the magnetic induction B. If it is assumed air does not magnetize, or this effect is calibrated for, the magnetic flux and the magnetic susceptibility can be calculated from
B= µ0(H+ M) = µ0H(1+ k) ; (1)
B= µµ0H, (2)
where µ= (1 + k) is the magnetic permeability or the ability of the substance to become magnetized.
It is because the magnetization is related to magnetic field H and the magnetic properties of the sub- stance that these properties are a unit-less parameter by volume, known as bulk susceptibility. The bulk
Table 2: Conversion between the SI and CGS units
Parameter cgs Conversion SI
Magnetic moment (m) emu 103 A m2, joule per tesla
Magnetization by volume (M) emu/cm3 103 A/m
Magnetization by volume (4πM) G 103/4π A/m
Magnetization by mass (Ω) emu/g 1 A m2/kg
Magnetic field (H) oersted (Oe) 103 A/m
Magnetic Induction (B) Gauss (G) 10−4 tesla (T), Wb/m2 Magnetic susceptibility (by volume) dimensionless 4π dimensionless Magnetic Susceptibility (total: kT) dimensionless 4π dimensionless Magnetic Susceptibility (by mass: χ) cm3/g, emu/g 4π×10−3 m3/kg
susceptibilityκ, and thought of as a scaling factor on magnetic permeability. Assuming there is no mag- netization outside the substance we can simplify magnetization as the product of bulk susceptibility κ and the applied field (equation 3).
M= κH (3)
2.2 Mineral Magnetism
All minerals have magnetic properties and these are fundamentally linked to their crystallography and the electron distribution within. It is sufficient to use a few basic terms of types of magnetism to cat- egorize minerals and rocks. A mineral may exhibit all of these types of magnetism and the degree of susceptibility of one type may overwhelm others by orders of magnitude.
For natural elements their orbiting electrons of atoms respond to an applied magnetic field with a mag- netization in the opposite (negative) direction i.e. diamagnetism (Fig. 3a). This type of magnetization is related to the orbital moment of electrons around the nucleus of the atom. Pure diamagnetic min- erals e.g. quartz and calcite have a negative susceptibility on the order of -10−5 [-]. For solids with incomplete electron shells, the un-paired electron spins can align in the direction of the applied field.
This produces a positive magnetization parallel to the applied field, called paramagnetism (Fig. 3b).
Unlike diamagnetism, it is dependent on temperature. Both diamagnetic and paramagnetic properties are field-dependent and change linearly with the applied field (Lowrie, 1997)).
Diamagnetism
Magnetization J [Am2]
Magnetic f eld H [T]
Magnetic susceptibility χ Temperature T
χ<0 M = χH
χ = const
Paramagnetism
J
H
χ
T χ>0
M = χH
χ 1/ T
a) b) c) d) Figure 3: Susceptibility and resulting magnetization a) opposing c) parallel, to the applied field. Sus- ceptibility as a function of temperature b) constant d) proportional to 1/T. Modified after on Lowrie, W.
(2007).
Electrons may also be coupled between adjacent atoms. For Fe (Ni and Co), situated in two neighbouring crystal sub-lattices, electrons can exchange or move from one Fe-slot to the other (e.g. between Fe cations in a magnetite unit cell Fig. 4). Depending on crystal structure, the exchange can hold an uneven distribution of magnetic moments and have an inherent magnetic moment (i.e., a magnetization in the absence of an applied field). It has a remanent magnetization after the external field is removed and is the definition of ferromagnetism (sensu lato) (Fig. 5).
Figure 4: Coordination of Fe cations with O−2anions (green) in magnetite. For A-sublattice (tetrahedral Fe(III)) in blue and B-sublattice (octahedral Fe(II,III in red). The unit cell of the of the spinel crystal structure is shown by the dashed lines. Modified from Butler, R. F. (1992b)
Unlike iron, where the magnetic moments all align in the same direction, the electron exchange in minerals is often anti-parallel i.e. anti-ferromagnetic. This means that the magnetic moments are of equal size, and if in opposite direction they will cancel each other out. In anti-ferromagnetic minerals, direct electron exchange between Fe cations is not possible but occurs through an electron ’cloud’, over the oxygen anion. Hence, it is an indirect exchange. Adjacent atoms get opposite intrinsic magnetic moments with a resulting weak and positive susceptibility without a magnetic remanence. This electron exchange breaks down at the Néel temperature (TN). If the magnetic moments are not fully antiparallel they create a combined vectorial magnetic moment i.e, spin-canted antiferromagnetism. This should not be confused with magnetic moment caused by parasitic magnetization due to impurities in the crystal lattice (Lowrie, 1997). If the opposing sets of coupled magnetic moments are uneven in size they result in a magnetic moment dipole i.e. ferrimagnetism. The ferrimagnetic property breaks down at the Curie temperature (TC), which is at 580°C for magnetite (Butler 1992b; Dunlop and Özdemir 1997). How strongly ferrimagnetic iron-oxides are magnetized i.e. the susceptibility is directly dependent on crystal grain size.
antiparallel
coupling antiparallel
coupling: layers of unequal M ferromagnetism antiferromagnetism spin-canted
antiferromagnetism ferrimagnetism
parallel coupling
zero Net
magnetic moment:
parallel coupling
Figure 5: Exchange couplings for ferromagnetism; a) truly ferromagnetic, b) antiferromagnetic, c) spin- canted antiferromagnetic and d) ferrimagnetic materials. Their net magnetization ’magnetic moment’
is shown below the coupling of magnetic moment in magnetic minerals. Modified after Butler, R. F.
(1992b), and Lowrie, W. (1997).
Ferromagnetism (s.l.) is with exception of very low fields, non-linearly field-dependent (Dunlop, and Özdemir, 1997). By increasing the field strength more magnetic moments align until the magnetization eventually becomes saturated (Ms). By switching the field direction, the full non-linear magnetization behaviour results are revealed in a hysteresis loop (Fig. 6).
saturation magnetization
isothermal remanent magnetization
coercive force remanent
coercivity Hcr
Hc Mrs Ms
-Hc
H [T]
m [Am2]
Figure 6: Sketch of a hysteresis loop; Ms - saturation magnetization; Mrs - Saturation remanence magnetization; Hc - Coercivity; Hcr- Coercivity of remanence. When the back field (applied in the -H direction) is switched off, the original remanent acquired in the positive (H+) direction, after saturation, has been removed. The value of the back field required to nullify the saturation remanent magnetization is the coercivity of remanence (Hcr). Modified after on Lowrie, W. (2007)
When turning the field off, the induced magnetization disappears momentarily but in ferromagnetic minerals a part of the magnetization remains. The saturation remanent magnetization (Mrs) is defined as what remains after saturation and is one of the parameters that can be obtained from the hysteresis loop. It is only by inducing an oppositely directed field that the original level of magnetization can once again be reached. However because an opposite field has to be applied before relaxation, this procedure becomes a process of trial and error. This is referred to as demagnetization, successively increasing negative fields are applied after achieving saturation. The Hc parameter is defined as the field needed to reduce the saturation magnetization to zero. The width of the hysteresis loop is theoretically 2 × Hc. The demagnetizing field needed to completely remove the saturation remanent magnetization is known as coercivity of remanence (Hcr).
These parameters are commonly used as ratios; Hcr/Hcad Mrs/Ms, which through empirical work (e.g.
Day et al. 1977 and Dunlop 2002) has become common in diagnostics of grain-size-like magnetic properties and theory for behaviour of magnetite. This simplification works well on single grains but
tends to average grains in mixed samples (Tauxe et al. 2016). Furthermore, because the theory only strictly applies to magnetite mixed samples containing e.g. hematite are expected to distort hysteresis curves (Tauxe et al. 1996) and distorted ratios. Albeit the theory is unable to make size predictions for mixed natural samples magnetic properties but may characterize the magnetic properties of the whole sample when making comparisons within a group. The Blötberget are ore grade mixed samples with varying wt% and vol% magnetite and hematite (BGU, 2015).
2.3 Micromagnetic properties
2.3.1 Domain theory
The non-linear dependency of ferromagnetic (s.l.) grains are connected by the parameters of saturation, remanence, and coercivity (see section above). These are also the building blocks for is known as domain theory. Magnetic domains are small (1-100 µm) and single grains can contain one or more domains. Through the continued work of Néel (1955), Day et al. (1977) and Dunlop (2002), it is possible to divide magnetic behaviour into different domain size ranges. Most magnetic grains in an iron ore sample will have grains that are too large enough to contain only a single magnetic domain.
The grain is too large for the magnetic energy to maintain a uniform magnetization and it splits up into parts with individual magnetic domains (Lowrie, 1997). It is said that these behave as ’multidomain’
(MD) grains. This gives the MD grains several degrees of freedom for change in the magnetic moment to align to the field. In grains that contain one single domain (SD) the only available way to align the magnetic moment is to rotate it (although this does not imply a physical rotation of the grain itself!). For MD grains this option will only be used after other easier options. In effect the SD grains are ’harder’ to align with the external field than the ’softer’ MD-grains. The theoretical size range for a magnetite SD grain is between 0.03 to 0.1 µm (Dunlop, D. & Özdemir, O. 1997). The exact size will depend on the saturation magnetization and the shape of the grain. For hematite, this transition starts at a much larger size of 15 µm. There are also size ranges that fall outside the strict definitions of SD and MD. In between there is a transition zone, the so called pseudo-single domain ’PSD’ size. These grains are believed to contain a few domain walls, which are not allowed to move freely (Lowrie, 1997).
The magnetization is unstable for grains that are smaller than SD. At this level thermal energy overcomes the uniform magnetization within the grain, and a stable magnetic remanence does not exist. As an effect,
the uniformed magnetization of the grain fluctuates and becomes frequency dependent. The behaviour is theoretically analogous to paramagnetism (Lowrie, 1997; Dearing et al. 1996, Dunlop and Özdemir 1997). The domain state for grains smaller than SD grains are therefore called superparamagnetic.
2.3.2 Remanent Magnetism
The natural remanence magnetization (NRM) is a material’s remanence when measured without any magnetic or thermal treatment (hence "natural"). The stability of this ’locked’ magnetism is, as men- tioned earlier, dependent on grain size but also subsequently time. The period before the grains’ mag- netic moment re-aligns or remagnetizes is called ’relaxation time’ and is a probabilistic change in mag- netization over time. The relaxation time for multidomain ’MD’ is very short on a geological time scale and the remanence is considered unstable (Fig. 7). However, the remanence of SD grains can be stable over billions of years.
10
1
0.1
0.01
10
1
0.1
0.01
Grain length (μm) Grain length (μm)
acicular grain
spherical grain
multidomain
single domain
superparamagnetic Coercivity (T)
0.0 0.2 0.4 0.6 0.8 1.0
0.25 0.20 0.15 0.10 0.05 т = 10 9 yr
т = 100 s
Axial ratio (b/a)
Figure 7: Sketch of relation between relaxation and grain sizes and shapes for superparamagnetic, single domain and multidomain for magnetite. Redrawn after McElhinny, 1973.
In nature, the sufficient energy provided to push a long term metastable grain to another energy level
depends on the geological history of the rock (e.g., metamorphism). Continental rocks are often older than 1 Ga, and may have experienced metamorphism more than once. Furthermore, a magnetic grain in nature is not an isolated environment. Metamorphic conditions and hydrothermal fluids can also play a role. The NRM can thereby be modified by effects of temperature and viscous overprinting.
Thermoviscous magnetization ’TRM’: alignment of magnetic moments, occur as a function of both time and temperature. Chemical magnetization occur through chemical alteration (precipitation) during diagenesis or metamorphism (Dunlop, D. & Özdemir, O., 1997). A more complex geological history, involving many metamorphic events, may make it harder to distinguish the original remanence that formed during the formation of the rock (Butler, 1992b). The Grängesberg ductile deformation (Eklöf, 2014) suggests that the close-by the host rock at Blötberget also has gone through strong deformation.
However petrography and geochemistry from both Grängesberg (Weis et al 2013) and Blötberget (Jiao, 2011) suggest that the deposit formed mainly through crystallization from an magma. What can still be debated is if the intrusive rock as well as the metavulcanic host rock has gone through green schist or amphibole metamorphism, because such may very well have consequences for magnetic remanence.
According to Dunlop, D. & Özdemir, O. (1997), rocks that metamorphosed above middle amphibolite facies are likely to loose all the original remanence recorded by magnetite but may keep some recorded by hematite. However, if the temperatures reached above 605 °C and/or in other way were exceeded blocking temperature, this initial remanence within hematite would also be lost, dependent on time, temperature or if inclusions, of for example CO2or H2O, are let free and are enough to tilt the chemical balance over. This aside, remanent magnetization can also be secondary (overprinted) and acquired gradually even in weak magnetic fields. This is typical, since the Earth’s magnetic field is relatively weak. To what extent this viscous remanent magnetization (VRM) blur the initial remanence depends on the relaxation time of the grains. Thereby, large MD-grains are prone be more affected by VRM.
Considering that grains in Blötberget samples are of the order of 100 µm, VRM will likely play a very large role for the NRM of Blötberget samples.
2.4 Iron-titanium oxides and their magnetic properties
Iron and titanium oxides are closed packed structure with cations of oxygen and interstitial anions of Fe2+, Fe3+and Ti+4arranged in octahedral or rhombohedral form (Dunlop, D. & Özdemir, O., 1997).
Susceptibility varies depending on the net magnetic moment of these cations. Two solid-solution series, between ulvöspinel and magnetite, and between ilmenite and hematite, are illustrated with a ternary
diagram (Fig. 8). These may also be bridged through oxidation/reduction. Spinel groups (Fig. 9) represents the simplified arrangement of cations and their resulting magnetic moments. It conceptually explains how ferromagnetic (s.l.) properties work (Butler, 1992). The range of magnetic susceptibilities for different iron oxides is shown in Figure (10).
oxidation
FeO Fe2O3
Rutile TiO2
Ulvöspinel Fe2TiO4
Ilmenite FeTiO3
Magnetite
Fe3O4 Hematite
Maghemite Titanomag
netit es Titanohema
tites
Figure 8: Ternary compositional diagram of the iron-titanium oxides.
A B
ZnFe2O4
Zn+2 Fe+3Fe+3 Normal Spinel
A B
Fe3O4
Fe+3 Fe+3Fe+2
A B
Fe2TiO4
Fe+2 Fe+2Ti+4 Inverse Spinel
Ferrimagnetic Antiferrimagnetic
Figure 9: Sketch cation distributions in normal spinel and inverse spinel. A and B are sublattices and arrows indicating magnetic moments caused by the cations. Redrawn after McElhinny (1973)
(titano) magnetite/maghemite
hematite
Banded Iron formation (anisotropic)
Ilmenite/titanohematite
Skarn
Magnetite skarn Magnetite rich Hematite rich
Iron ore
SD MD
Iron oxides
Dispersed grains
Massive grains
10 - 3 10 - 2 10 - 1 1 10
[SI]
Figure 10: Range of susceptibilities for magnetic iron oxides and iron ore. Redrawn and modified after Clark (1997).
2.4.1 Titanomagnetite
Magnetite and ulvöspinel are end members of the solid solution series for Titanomagnetite (Fe3−xT ixO4, 0 ≤ x ≤ 1) and have inverse spinel crystal structure. Titanomagnetite (Fe3−xTixO4) are increasingly paramagnetic with a higher content of Ti and has no ferrimagnetism with x= 0.8 at room temperature (Butler, 1992). Within one unit cell there are two sub lattices, one tetrahedral ’A’ and one octahedral
’B’. Iron can share electrons across these two sub lattices (Fig. 9). For ulvöspinel the charge is Fe+3in both A and B. However for magnetite there is an excess charge from the additional Fe+2in B. Because of this excess charge a net magnetic moment is created and as a consequence magnetite is ferrimagnetic.
Magnetite has the strongest susceptibility (k ≈ 1 to 10 [-]) of any iron-oxide at room temperature, and has a saturation magnetization of 480 G (4.8×105 A/m). The bulk susceptibility in rocks is to large extent proportional to the amount of magnetite, when magnetite is present even in very low amounts (<1 %) (Lowrie 1997). Furthermore, while a magnetite-rich sample becomes a strongly magnetic and remanence-bearing, a ulvöspinel (Ti-rich) rocks approaches paramagnetic conditions. Magnetite is ferri- magnetic (s.s.) space and the TCis 580 °C. Ulvöspinel is antiferromagnetic, has a TNof -153 °C (Butler
1992b), and is paramagnetic at room temperature. Both Msand TC are lowered with increased amount of Ti and follow the cation substitution nearly linearly (Nagata, 1961).
2.4.2 Titanohematites
Hematite (αFe2O3) and Ilmenite (FeTiO3) are the end members of the solid solution series of titanohe- matites (Fe2−xTixO3). Hematite can crystallize in environments of 600-700°C (Ramdohr, 1980) although it is commonly thought of as a secondary mineral. Hematite has a rhombohedral lattice structure (e.g.
Dunlop & Özdemir, 1997). Magnetic moments between the layers are approximately anti-parallel to each other. However, a small deviation from 180° in the opposing magnetic moments generated by the Fe3+cations create a weak magnetic moment. Hematite thereby obtains a spin-canted antiferromagnetic moment with susceptibility of ∼10−3[-], ∼0.5% of the susceptibility of magnetite (Butler, 1992). With pure samples in room temperature, spontaneous saturation Ms is ∼2G (2.2 × 103 [A/m]; Clark, 1997;
Lowrie, 1997). Titanohematites with Ti4+composition of x ranging from 0 to 0.45 are antiferromagnetic (Butler, 1992). When Ti replaces Fe, it removes the antiparallel coupling between magnetic moment and make the titanohematite ferrimagnetic (Dunlop & Özdemir, 1997). The TN shifts due to the temperature dependency which occurs when the titanomagnetites goes through ionic substitution from Ti4+to Fe3+. The TN for pure hematite is 675 °C and for ilmenite (TN is approximately 50 °K (Lowrie, 1997) (-218
°C, Butler 2002a)). Ilmenite can also be categorized with analytical methods, but it is outside the scope of this thesis. At room temperature ilmenite is a paramagnetic mineral.
2.4.3 Oxidation of ferromagnetic minerals in continental rock and petrological effects
Petrological events, especially oxidation, can affect the magnetic properties of iron-oxides. Here follows a brief summary:
1. Slowly cooled intrusive volcanic rocks, with grain sizes exceeding 100 µm have far shorter relaxation times and lower coercivity. These are ’soft’ and are more prone to rearrange to the present magnetic field with time (Dunlop, D. & Özdemir, O., 1997).
2. Exsolution of titanomagnetites in slowly cooling plutonic rocks at ≈ 600 °C or alteration through unmixing can subdivide regions into Ti-rich and Ti-poor and subsequently affect temperature dependence. It also decreases the grain size (Butler, 2002a).
3. Oxidation, for which there are two types a) deuteric and b) low-temperature. The deu-
teric oxidation drives the composition towards the right in the ternary diagram (Fig. 8), increasing the Fe3+/ Fe+2ratio. The term martite is used for platy hematite grains that form along octahedral planes.
Formation usually follow grain boundaries, fissures and the octahedral plane but can be quite irregular.
Martitization can occur by weathering or if there is oxygen available during heating. It can be formed as a pseudomorph from magnetite (Lowrie, 1997). Martite is generally thought of as an alteration process occurring during the last stages of ore formation while the temperature is decreasing (Ramdohr, 1980).
According to Dunlop (Dunlop & Özdemir, 1997 ; Lowrie, 1997), martite can replace mag- netite at ambient temperatures: low temperature oxidation. Through prolonged oxidation magnetite turns it into a metastable mineral: maghemite (γFe2O3). Because the oxidation vacancies are produced in the crystal lattice, the crystal structure is the same as that of magnetite but the chemical formula is that of hematite. Maghemite can revert to hematite (martite) when heated to ≥ 200 °C (Dunlop & Özdemir, 1997); or 300-350 °C (Lowrie, 1997).
2.5 Anisotropy of Magnetic Susceptibility
In bulk susceptibility measurements, the sample is considered to be isotropic and defined by a single measurement. To derive anisotropy of magnetic susceptibility the induced magnetization is measured in several different directions. With the field already known, the second rank symmetric tensor of suscep- tibility can then be determined. In other words this tensor (ki j) considers the applied field vector Hjand the magnetization vector Mi.
Mi = ki j∗ H j (4)
M and H are the vectors of magnetization and the applied field, respectively. As M and H are vectors, k is naturally a second rank tensor. This method requires the sample itself to be as close to spherical as possible so as not cause an shape anisotropy that influences the inherent magnetic anisotropy. Because preparing a spherical sample is time consuming, a cylinder shape or cubic shape sample are commonly used. Symmetry relations requires measurement of magnetic susceptibility along a minimum of 6 unique axes in order to produce the complete susceptibility tensor. In practice the sample is inserted in an in- duction coil in six or more independent orientations. The second-ranked tensor can then be represented by three eigenvectors and eigenvalues, or the principal axes k1≥ k2≥ k3. These principal axes are used in different ways to present characteristics of anisotropy (see more in methodology section). There are several reasons why a sample is magnetically anisotropic, both on the crystal scale and the rock scale.
1.) Magnetocrystalline anisotropy; For example rhombohedral hematite has a magnetic sym-
metry related to its crystal structure. Crystalline hematite can also be strongly anisotropic (Ku ≈103 J m−3, where Ku i the uniaxial anisotropy constant) (Lowrie, 1997). In contrast crystalline magnetite has a cubic symmetry that is the same in all six directions, and therefore very low magnetocrystalline anisotropy.
2.) Shape anisotropy: Needle shaped grains of ferromagnetic minerals with high spontaneous or saturation magnetization (i.e., magnetite) can be strongly anisotropic. The elongated grains have its maximum susceptibility perpendicular to the long axis on the grain. Conceptually this is explained by magnetic charge distribution where the longer opposite sides of a grain can have a build up of more
’positive’ and ’negative’ charges than two shorter opposite sides (see e.g Butler 2002a).
3.) Distribution anisotropy, related to the distribution of grains in a sample. Although suscep- tibility can be related to grain size and shape, in natural samples it is based on an averaged magnetization in a range of grains. Strongly magnetic samples, such as iron ore, are more likely to reflect linear or planar alignment (Dunlop, D. & Özdemir, O., 1997). Sedimentary rocks are prone to have alignment of grains and layering. A clear example of this is banded iron ore which has magnetite-rich bands with low magnetic bands in between. The susceptibility normal to the banding is self-demagnetized by the next layer yet susceptibility remains parallel to the banding (Clark, 1997).
3 Measurement equipment
3.1 MFK1-FA with semi-automatic Rotator, CS4 furnace and CS-L cryostat
The MFK1-FA (Fig. 11) is a laboratory instrument that enables semi-automatic magnetic suscepti- bility measurements in room temperature. The equipment has been developed by AGICO (Advanced Geoscience Instruments Company) The field strength is measured with the empty holder and with the sample present. The MFK1-FA can perform measurements in the MFK1 standard frequency 976 Hz, but also in 3904 Hz and 15616 Hz. Field strengths from 2 to 700 A/m can be applied. Furthermore, measurement techniques are available using attachments for specific purposes. One such add-on is the semi-automatic rotator. Instead of performing discrete measurements, measurements with this attach- ment reduces the number of manual steps. The MK1-FA software control the 3D-rotator and can register a large number of measurements while the sample is rotating, which in turn can save time and improve accuracy. The MFK1-FA is augmented to measure magnetic susceptibility at high and low temperature.
The CS4 furnace attachment is a high temperature electric furnace. The sample is heated up to 700 C and cooled back to room temperature. The induced magnetic field in the MFK1-FA chamber is measured as with other measurements. In obtaining the actual susceptibility of the sample, the chamber signal has to be subtraced from the raw measurement. Similarly the CS-L cryostat attachment is together with the MFK1-FA. In this case, liquid nitrogen is used to cool the sample to -192 C and while the sample is heating to 0 C, measurements are performed. To avoid oxidation during measurements both furnace and cryostat allow for venting samples with argon gas supplied from a gas tube. The AGICO software Sufyte4 is used for control of temperature dependent measurements.
Figure 11: The MFK1-FA laboratory instrument developed by AGICO (Advanced Geoscience Instru- ments Company). Picture from high temperature susceptibility (Bjork 2016).
3.2 7400-S Series VSM
In this thesis micromagnetic properties of samples are mapped using a 7400-S Series vibrating sample magnetometer (VSM) in the solid state physics section of the Ångströms Laboraty at Uppsala University (Fig. 12). The equipment is developed by Lake Shore Cryotronics (Lake Shore Cryotronics, 2017). A sample is magnetized in a uniform field and an induced oscillating magnetic field is created through mechanical vibration (Lake Shore, 2017). The sample is attached to a non-magnetic rod and vibration is controlled by a linear motor, or connected to a speaker. The oscillating magnetic field generates a DC- current in detection coils which can be measured (Cullity, and Graham, 2009). The VSM measures the magnetic moment of the sample and by mapping the full non-linearity of ferrimagnetism i.e. the magne- tization. Similar to the Kappabridge, the VSM work at room temperature but can also be augmented by adding furnace or cryostat. Fields of up to 1436 kA/m (1.8 T) were used to determine magnetic satura- tion. To determine the coercivity of remanence (Hcr), an initial 1197 kA/m (1.5 T) field was applied in a positive direction, and afterwards the resulting magnetization was progressively removed by applying increasing reversed fields with strengths of -4, -12, -16, -20, -24, -32, -40, -48, -56, 64 kA/m (-5, -15, -20. -25, -30, -40, -50, -60, -70 and -80 mT).
Figure 12: The 7400-S Series vibrating sample magnetometer (VSM) developed by Lake Shore Cry- otronics. During measurement at Ångström Laboratory (Bjork 2016, photo collage).
4 Methodology
4.1 Sample preparation
The sample set in this study consists of 31 samples, originating from drill cores and an additional 8 from hand samples (Table 3). For this project slabs of approximately 1 cm high by 2 cm (domino shaped pieces, Fig. 13a ) were used. These had previously been left overs from thin section preparation and each slab was thereby accompanied by a thin section. From these slabs, one third were marked for analysis as powder, one third for further sawing and polishing, and the final third as spare material.
Powdered samples (Table 5) were ground with a ceramic mortar and pestle. A hand held contour saw, equipped with a diamond cutting wheel (Fig. 13b), was used to produce 3x3 millimetres cubes (Fig.
13c). Dimensions of the samples were measured with a digital calliper, and thereafter weighed with a high-resolution balance (Table B3). The 38 thin sections provided were used for an overview of the described petrology and identified mineral, previously performed by Berg och gruvundersökningar (2005). Table 3 provides an overview of the methods used for the respective samples. The best coverage of measurements are on the samples originating from drill cores. The choice was made to focus on trends within the ore body and identify samples deviating from these trends. A small set of samples from old mine localities at the area of Blötberget were also used to put some perspective on the drill core samples. An additional set of 3 hematite samples, from the Swedish Museum of Natural History, were also used for comparative purposes.
Figure 13: a) Project slabs approximately 1 cm high by 2 cm. b) Sawing sample intended for measure- ment of magnetic anistropy. c) Sample P20b put aside for measurements of magnetic anisotropy.
Table 3: Overview of methods for respective sample, origin of depth, and ore body. In parenthesis mea- surements that were only successful in part (remanent coercivity test did not succeed; details in section 5.7). Abbreviations: Drill core sample ’Dc s.’, bulk susceptibility ’Bulk’, frequency dependency ’Freq’, field dependency ’Field’, Anisotropy of magnetic susceptibility ’AMS’, high temperature dependency
’HighT’, low temperature dependency ’LowT’, Micromagnetics ’Mn’, ’Dc s.’ Drill core sample, ’Blast s.’ Blast sample, ’Spec’ Specimen
Name HoleNoName Depth[m] Type Ore body Bulk Freq Field AMS HighT LowT Mn
P1 BB12004 406.9 Dc s. Flygruvan x x x x x
P2 BB12005 409.9 Dc s. Flygruvan x x
P3 BB12004 417.5 Dc s. Flygruvan x x x x x x x
P4 BB12004 447.0 Dc s. Kalvgruvan x x x x x x x
P5 BB12004 456.1 Dc s. Kalvgruvan x x x x x x
P6 BB12005 458.2 Dc s. Kalvgruvan x
P7 BB12004 463.1 Dc s. Kalvgruvan x x x x x x
P8 BB12005 433.8 Dc s. Flygruvan x x x x
P9 BB12006 446.0 Dc s. Flygruvan x
P10 BB12007 457.6 Dc s. Flygruvan x
P11 BB12005 461.9 Dc s. Flygruvan x x x x
P12 BB12005 474.3 Dc s. Kalvgruvan x x x x x x
P13 BB12006 484.5 Dc s. Kalvgruvan x
P14 BB12007 492.4 Dc s. Kalvgruvan x
P15 BB-73-246 18.0 Dc s. Hugget x x x x x x
P16 BB-73-246 26.4 Dc s. Hugget x x x x x x x
P17 BB-73-246 29.9 Dc s. Hugget x x x x x x x
P18 BB-73-234 75.3 Dc s. Sandel x x x x x x
P19 BB-73-234 74.6 Dc s. Sandel x x x x x (x)
P20 BB-73-233 43.8 Dc s. Sandel x x x x x x (x)
P21 BB-73-233 46.8 Dc s. Sandel x x x x x x
P22 BB-73-233 151.9 Dc s. Hugget x x x x x x
P23 BB-73-233 159.9 Dc s. Hugget x x x x x x x
P24 BB-73-233 145.8 Dc s. Hugget x x x x x x
P25 Iviken h.pile Blast s Hugget x x x x x x x
P26 Iviken m.pile Blast s Gulkannan x x x x x
P27 BB12005 323.0 Dc s. Flygruvan HW x x x x
P28 BB12006 391.4 Dc s. Flygruvan HW x x x x
P29 BB12007 470.7 Dc s. Fly-Kalv MA x
P30 BB12008 504.8 Dc s. Kalvgruvan FW x x x x
P31 BB12015 Upper lvl Dc, m. Kalvgruvan
P32 Spec. Flygruvan SW x x x x
P33 Spec. unknown
P34 Spec. Neptungruvan x
P35 Spec. Bolagsgruvan x x x x
P36 Spec. Mossgruvan x
P37 Spec. Buskgruvan x x x x
P38 Spec. Köpmannagruv. x
P39 Spec. N Tremänningen x x x x
4.2 Magnetic methods: unless specified measured at 15-25 °C
4.2.1 Bulk susceptibility
Bulk susceptibility is a fast measurement, and takes only a few seconds per sample at ambient room temperature (about 20 °C). The induced magnetic field H, and the magnetization M are measured si- multaneously. Because both H and M are given by units of A/m, it makes the bulk susceptibility (κ) a dimensionless ratio (eqn. 3). The MFK1-FA has a predefined nominal volume (V0) set at 10 cm3. The result are rescaled when smaller samples are used (eqn. 5, originally by Jelínek, 1977).
κ = (V0/V) ∗ k (5)
The bulk susceptibility, (κ) is the nominal sample size divided by the actual, and multiplied with the measured susceptibility k. The bulk susceptibility, being corrected for volume, makes it easier to com- pare different samples.
The mass susceptibility (χ) is defined as the bulk susceptibility κ normalized by mass:
χ = κ/mass (6)
4.2.2 High temperature-dependent susceptibility: 25 to 700 °C
Magnetic susceptibility measurements were carried out as a function of temperature, using the MFK1- FA susceptibility bridge, equipped with the CS4 owen . The set contained 38 rock powders of 0.4 grams, each put in a glass test-tube and flushed with argon during measurements avoid oxidation. Some oxidation is unavoidable during cooling from high temperatures and a specific sample can therefore only can be measured once. In this study, the samples were exposed to stepwise heating and cooling;
up to 700 and back to 40 °C. The measured magnetic susceptibility is the cumulative susceptibility of minerals with strong susceptibility contribution (i.e. magnetite), which overshadows those with much lower susceptibility (e.g. hematite). However, the susceptibility drops readily when reaching the Curie temperatures (TC) and Néel temperature (TN), respectively for the two minerals (Fig. 14).
Susceptibility [1E-3 SI]
Temperature (°C)
0 580
0 1 2 3 5
4
6 TC
Susceptibility [1E-4 SI]
Temperature (°C)680
-1 0
660 640
TN Magnetite
Hematite
Figure 14: Illustration of how the magnetic susceptibility typically changes as a function of temperature in samples from Blötberget.
The measured Curie temperature for magnetite can be somewhat shifted within a measured group and when compared to a reference. This may reflect small compositional differences between the samples.
As previously mentioned (section 2.4.1) the Curie temperature for titanomagnetites has a close to linear dependency on Ti content. However there may also be several other causes for a higher or lower tran- sition temperatures. In this case these have been defined as simply the average of temperature value at start and end of the step-like transitions. The reason for this chosen approach is ease to replicability of method. However it does demand that transitions are well defined. Because the temperature-dependent susceptibility measurements of the MFK1-FA are done stepwise, the measurement and accuracy are only as good as the length between measured points during heating and cooling. In this study the drop in susceptibility typically involves several measurement points during the measurements of the Blötberget samples, which provides good enough accuracy to define the TC.
4.2.3 Low temperature-dependent susceptibility: -194 to 0 °C
An additional set of measurements of magnetic susceptibility as a function of temperature, were con- ducted using the MFK1-FA together with the CSL cryostat. The subset of 8 powdered rock samples from Blötberget of 0.4 grams each, and 2 additional reference samples, were cooled to -194 °C with liq- uid nitrogen. The nitrogen was then cleared from the chamber with argon gas and within two hours, the samples were warmed back to room temperature. During this time the susceptibility was measured re- peatedly. The method is non-destructive and provides the susceptibility as a function of low temperature, which can be used to identify magnetic minerals. Magnetite displays a temperature dependent change in susceptibility at -153 °C (120 °K; Verwey transition). The exact cause for the transition has been a subject of debate in solid state physics since it was discovered (Verwey, 1939). A somewhat simpli- fied explanation is that the cubic crystal structure becomes distorted to a monoclinic (originally Verwey suggested a orthorhombic system). The crystallographic transition causes magnetite to become less fer- rimagnetic and susceptibility is thereby lowered. For the SD grain the main change is caused by the main contributor in susceptibility shifting crystallographic axis. From dominated by shape anisotropy to be caused only by crystallography. But in MD grains, the magnetic properties is largely dependent on the domain walls and their ability to create parallel or perpendicular twin domains (Kasama, 2010). The 180° twin walls are easier to move than the 90° resulting in a softer magnetization. The above properties together with concentration, composition and oxidation state determines the degree of magnetization.
As discussed in the theoretical section, a weak magnetic moment of hematite originates from a canted sublattice create a small permanent magnetization in the crystal basal plane. However when hematite is cooled to approximately -15 °C the magnetocrystalline anisotropy changes sign (Liebermann & Baner- jee 1971), and the spins align with the crystal c-axis (which is perpendicular to the basal plane), and the effect of the spin-canting is lost, which causes the the sublattices to become anti-parallel, i.e. the Morin transition (Morin, 1950). The definition of the exact temperature for the Morin transition tem- perature varies in literature, and summarized by (Tarling & Hrouda 1993). The main reason for this is that different reference samples have been used. Also referred to in this thesis is the pure stoichiometric hematite from the Island of Elba, with a TM approximately equal to -11 °C (Morrish, 1994). In samples from Blötberget minerals with strong susceptibility contribution (i.e. magnetite) overshadows those of much lower susceptibility (i.e. hematite). However when the relative abundance and purity of the more weakly magnetic mineral is sufficiently large it can be distinguished (Fig. 15).
