Journal of Geochemical Exploration 146 (2014) 17–26

Contents lists available at ScienceDirect

Journal of Geochemical Exploration

j ourna l homepage: www.e lsev ie r .com/ locate / jgeoexp

The relationships between magnetic susceptibility and elemental variationsfor mineralized rocks

Yu Li a,b, Renguang Zuo a,⁎, Yin Bai a,b, Mingguo Yang b

a State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences, Wuhan 430074, Chinab Faculty of Earth Resources, China University of Geosciences, Wuhan 430074, China

⁎ Corresponding author. Tel./fax: +86 27 67885096.E-mail address: zrguang@cug.edu.cn (R. Zuo).

http://dx.doi.org/10.1016/j.gexplo.2014.07.0100375-6742/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 5 December 2013Accepted 12 July 2014Available online 19 July 2014

Keywords:Itrax core scannerMagnetic susceptibilityMineral explorationMineralization

Investigation of the geophysical and geochemical properties of rocks is essential for successful mineralexploration. In the present study, the Itrax core scanner, a fast and nondestructive instrument at ChinaUniversityof Geosciences (Wuhan), was applied to measure the magnetic susceptibility (MS) and elemental variations atspecific intervals of two limestone samples collected from the Jinding sediment-hosted Pb–Zn deposit, YunnanProvince (China). Statisticalmethods, including correlation, cluster, and regression analysis, were used to explorethe relationships between rock MS and elemental variations of Zn, Fe, Pb, Ni, Cr, Si and Ca. The resultingcorrelation analysis demonstrated that concentrations of Fe and Zn were positively correlated with MS, whereasPb, Si and Ca concentrations were negatively correlated with MS. The resulting cluster analysis and stepwisemultiple linear regression analysis betweenMS and elements concentrations show that Zn, and Cr are the crucialfactors responsible for the variability ofMS. These results are probably due to the Fe, Ni, and Cr elements existingin the lattice of sphalerite, which contributes positively to the MS of rocks.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

The relationships between magnetic susceptibility (MS) andelemental variations have been the focus of considerable study in thefields of environmental assessment and resource exploration. Forinstance, Zhang et al. (1998) investigated the subject and concludedthat magnetic parameters are related to heavy metal content based onanalysis of the magnetism of lake and tidal flat sediments. Similarly,Spiteri et al. (2005) studied the relationships between topsoil and thedistribution of heavy metals in the Lausitz region of eastern Germanyand demonstrated that MS can be used as a proxy for soil heavy metalscontaminations. Wang et al. (2009) described an indicative function ofrock MS in the exploration of iron oxide copper–gold (IOCG) depositsin Chile in another research, Deng et al. (2010) delineated rock MSanomalies in East Sichuan (China) in the exploration of IOCG-typedeposits. Cao et al. (2007) explored the relationship between rock MSand gold mineralization in Henan (China). However, few studies todate have investigated the relationships between MS and elementalvariations at the hand specimen scale.

To address this issue,we used the Itrax core scanner – a fast and non-destructive instrument – to measure MS and elemental concentrations

of limestone samples collected from the Jinding Pb–Zn deposit, aworld-class sediment-hosted deposit. In this paper, we discuss therelationships between MS and Pb–Zn mineralization using statisticalapproaches.

2. Study area and data

The Jinding deposit is one of the largest Pb–Zn deposits in China(a metal reserve of greater than 15 Mt deposit, average Zn = 6.08%,Pb/Zn = 1:4.7), located in the Lanping basin in northwestern YunnanProvince (Xue et al., 2003, 2004, 2007). The Lanping basin representsthe northern part of the Lanping–Simao Mesozoic–Cenozoic basin, de-veloped on the Changdu–Simao microplate between the Lancangjiangand Jinshajiang–Ailaoshan tectonic belts (Xue et al., 2003). The depositis composed of six ore blocks (Beichang, Paomaping, Jiayashan, Xipo,Baicaoping, and Fengzishan blocks) and has a total area of less than10 km2 (Hu et al., 1998). The mining area has previously been affectedby processes of sedimentation, thrusting, heat doming, and dome col-lapse, with local extension and heat flow upwelling accompanied bymineralization (Xue et al., 2003). The mineralization age coincides ap-proximately with the initiation of Himalayan alkali magmatic intrusion(i.e., 68Ma; Xue et al., 2003). The strata in themining area include bothautochthons and allochthons: the stratum overlying the flat-lying faultF2 is allochthonous and inverted, whereas that underlying the fault isautochthonous and exhibits a normal sequence (Fig. 1). The orebodiestypically exhibit tabular or vein forms and are hosted by fine sandstones(in the western part of the Beichang block and in the Fengzishan and

Fig. 1. Simplified geological map of the orefield (after Gao, 1989). 1: Limestone—Upper Triassic Sanhedong Formation; 2: Shale and siltstone—Upper Triassic Maichuqing Formation; 3:Siltstone and mudstone—Middle Jurassic Huakaizuo Formation; 4: Sandstone—Lower Cretaceous Jinxing Formation; 5: Sandstone—Middle Cretaceous Hutousi Formation; 6: Siltstoneand mudstone—Paleocene Lower Yunlong Formation; 7: Sandstone—Paleocene Upper Yunlong Formation (sandstone); 8: Zn–Pb ore body; 9: fault; 10: unconformity; 11: geologicalboundary; 12: sampling location.

18 Y. Li et al. / Journal of Geochemical Exploration 146 (2014) 17–26

Xipo blocks) or calcibreccia (in the eastern part of the Beichang blockand in the Jiayashan, Nanchang, and Paomaping blocks; after Zhao,2007).

Samples Nos. 1 (sphalerite 35%, galena 5%, pyrite 3%, marcasite 1%,calcite 35%, quartz 10%) and 2 (sphalerite 20%, galena 10%, pyrite10%,marcasite 10%, calcite 40%, quartz 5%) were collected from the Pb–Znmineralized breccia of Beichang and Paoma Ping in the eastern part ofthe mining area (Fig. 1). The length, width, and thickness of sampleNo. 1 (sample No. 2) were 10.5, 6.5, and 1.3 cm (7, 2.3, and 1.2 cm),respectively. The main metallic minerals in both samples were sphaler-ite and galena, and calcite was the primary transparent mineral; bothsamples contained a small amount of quartz. Galena and sphaleritewere developed primarily along breccia fractures in sample No. 1,whereas two ore-forming stages were observed within sample No. 2:

the first was characterized by a pyrite+ marcasite+ sphalerite assem-blage and was found within the orebodies; the second was associatedwith sphalerite + galena and occurred within calcite veins, and someof the earlier pyrite was replaced by the later sphalerite (Fig. 2).Sphalerite was found to bemore abundant than galena in both samples.

Magnetic susceptibility and elemental concentrations were mea-sured with Itrax core scanner instrument at the State Key Laboratoryof Geological Processes and Mineral Resources at China University ofGeosciences (Wuhan) (Zuo, 2013). The Itrax core scanner is an auto-mated instrument that can synchronously provide four datasets:micro-X-ray fluorescence analysis (μ-XRF) data, high-resolution X-radiographic images, high-resolution optical images, and MS profiles.It can be operated on rock or sediment cores up to 1800 mm long andwith diameters ranging from a few cmup to 12 cm, and the abundances

Fig. 2. Optical microscope photos of minerals from sample No. 1 (A) and No. 2 (B). Gn: galena, Sp: sphalerite, Py: pyrite, Mrc: marcasite, Qtz: quartz, Cal: calcite.

19Y. Li et al. / Journal of Geochemical Exploration 146 (2014) 17–26

of multiple elements (ranging from Al to U) can be measured synchro-nously. The optical line camera system has a maximum resolution of50 μm pixel−1 and the digital X-ray line camera has a pixel resolutionof approximately 20 μm with exposure times ranging from 20 ms upto several seconds. The measurement point size of XRF is 0.2 × 4 mm.The measured MS is a volume magnetic susceptibility k with units of10−5 SI and precision up to 2 × 10−6 SI. More detailed informationabout the underlying principles and operation of the Itrax core scanneris available in the literature (Croudace et al., 2006). One advantage ofItrax core scanner is its fast and nondestructive determination of thephysical properties and elemental variations of rocks with little analyt-ical effort. In the present study, using aMo tube, two rock samples weresubjected to XRF examination andMS analysis at resolutions of 0.2 mmand 2 mm, respectively, with exposure times of 30 s, 30 kV and 50 mA.We obtained 50 and 25 MS measurement points, and 500 and 250 ele-ment concentration points for sample Nos. 1 and 2, respectively (Fig. 3).The XRF andMS resolutionswere different; therefore, we calculated thegeometric mean of the measured XRF data between two MS measure-ment points to ensure that the final XRF analysis results coincided

with the MS measurement points (Xu et al., 2006). The accuracy ofour procedure was determined through analyzing the certified refer-ence material of NIST 1834 (National Institute of Standards and Tech-nology, USA). Table 1 showed the detection limits for the selectedelements.

3. Results and discussion

3.1. MS and elemental concentrations

The minimum, maximum, and median concentrations of Fe, Ni, Zn,Pb, Cr, Si and Cawith theirMS values are presented in Table 2. Themag-netic effects are caused by the orbital and spin angular moment of elec-trons. Various elements have different numbers of unpaired electronorbitals at different oxidation states, resulting in elements and com-pounds that can have positive (paramagnetic) or negative (diamagnet-ic) MS values (Spaldin, 2010). The atoms of most elements have onlyone unpaired electron (odd atomic numbers) or no unpaired electrons(even atomic numbers). Their permanent moments are either small or

Fig. 3.Magnetic susceptibility Measurement locations for samples No. 1 (A) and No. 2 (B).

Table 2Summary statistics of the data measured by Itrax.

Sample Statistical magnitude Maximum Minimum Mean

No. 1 MS (k) × 10−5 SI 0.23 −2.11 −1.02Fe2O3 (%) 2.64 0.13 0.74Zn (%) 12.85 7.88 9.91Pb (%) 3.89 0.09 1.54Ni (%) 0.02 0.01 0.01Cr (%) 0.02 0.007 0.01Si (%) 53.18 0.43 8.60Ca (%) 51.51 0.40 12.40

No. 2 MS (k) × 10−5 SI −1.98 −2.51 −2.21Fe2O3 (%) 26.61 0.06 9.35Zn (%) 11.71 0.04 6.14Pb (%) 12.14 0.01 3.09Ni (%) 0.03 0.001 0.008Cr (%) 1.72 0.06 0.50Si (%) 8.00 0.40 3.05Ca (%) 75.78 0.57 35.74

20 Y. Li et al. / Journal of Geochemical Exploration 146 (2014) 17–26

non-existent, such as Zn, Pb, Si and Ca, therefore, in general, sphalerite,galena, quartz and caltite are diamagnetic minerals. Atoms of transitionmetals like Fe, Ni, and Cr have larger moments that arise from unpaired3d (n=3, l=2) electrons. Compounds of these elementsmay also havelargemoments because the 4s electrons are removedfirst in ionic bond-ing and they exhibit paramagnetism or ferromagnetism (Clark, 1997;Dunlop and Özdemir, 1997). The magnitude of rock MS is assumed tobe affected by rock composition and the properties of minerals (Wu,1988). Aydin et al. (2007) demonstrated a positive linear relationshipbetween rock MS and mineral MS. Minerals can be classified as ferro-magnetic (k NN 0), paramagnetic (k N 0), or diamagnetic (k b 0) accord-ing to their MS values, with ferromagnetic (diamagnetic) mineralscontributing the most (least) to the rock MS (Lang et al., 2011;Sandgren and Snowball, 2002).

Fig. 4 presents plots of MS values against natural logarithm of ele-ments concentrations for samples Nos. 1 (Fig. 4A) and 2 (Fig. 4B). Theoverall variations of MS and the elements of Fe, Ni, Cr, Zn, Si and Cawere similar in sample No. 1, while Pb showed an opposite trend. The

Table 1Detection limits for Itrax core scanner (c.f., Itrax Core Scanner, Cox Analytical Systems,2010).

Element Detection limit (%)

Si 0.3793Ca 0.0046Cr 0.0013Fe 0.0009Ni 0.0007Zn 0.0005Pb 0.0036

The data are based on measurements using standard reference material 1834.Measurement time is 100 s.

variances of Fe and Zn are similar to that of Si and Ca, respectively, prob-ably due tomineral paragenesis of pyrite and quartz, sphalerite and cal-cite (Fig. 2). TheMS of sample No. 1 significantly increased at 16, 34, 60,and 84 mm and a similar trend was observed in the variations of Fe(peaks at 16, 34, and 84 mm), Ni (peaks at 34 and 84 mm), Cr (peaksat 34, 60, and 84 mm), and Zn (peaks at 34, 60, and 84mm). The varia-tion of Zn concentration revealed a strong correlation with MS values.Pb concentration exhibited a negative relationship with MS. The MSvalues of sample No. 2 showed a steep drop in the range of 0–2 mm,changing from−2.26 (10−5 SI) to−2.51 (10−5 SI); the concentrationsof Fe, Zn, Ni, and Cr also displayed a steep drop at this position. The MSvalues gradually increased in the range of 2–48mmaswell as Fe, Cr andZn, while the Pb, Si and Ca exhibited an opposite trend. High concentra-tions of Si and Ca were observed when the MS values were at low level(in the range of 0–25mm). The variations of MS and the elements indi-cated that Fe, Cr and Zn had a positive relationshipwithMS, while Si, Caand Pb exhibited a negative relation.

3.2. Relationship between MS and elemental concentration

Most rock-formingminerals that lack iron are diamagnetic in natureand their MS values are unrelated to temperature or the strength of theearth's magnetic field (Hrouda et al., 2003). Both samples investigatedin the present study exhibited negative MS values, although some para-magneticmaterials may have been present in these samples at very lowconcentrations. The correlation analysis between MS and elementalconcentrations was conducted to further investigate the relationshipbetween the rock MS values and elemental variations. The resultswere classified according to the correlation coefficient r (Yu and Hu,2005), as follows:

0.8 ≤ |r| ≤ 1 suggests a strong correlation;

0.5 ≤ |r| b 0.8 suggests a significant correlation;0.3 ≤ |r| b 0.5 suggests a weak correlation; and|r| b 0.3 suggests an insignificant correlation.

The Pearson correlation coefficients between elemental concentra-tions and MS values and between different elements are presented inTables 3 and 4. The scatter plots of the values are illustrated in Fig. 5.The resulting correlation analysis demonstrated that concentrations ofFe andZnwere positively correlatedwithMS,whereas Pb concentrationwas negatively correlatedwithMS. For sample No. 1, the correlation co-efficients between Zn andMS, Ca andMSwere 0.60 and 0.59 respective-ly, indicating a significant correlation. In contrast, the relationships

Fig. 4. Plots of magnetic susceptibility versus natural logarithm of elements concentrations along sample Nos. 1 (A) and 2 (B).

21Y. Li et al. / Journal of Geochemical Exploration 146 (2014) 17–26

between Si, Fe, Fe–Ni–Cr, and MS were weak (r = 0.45, 0.45, 0.45);Pb concentration was observed to have negative correlation with MS(r = −0.42); and Ni and Cr concentrations were almost uncorrelatedwithMS (r= 0.18, 0.17). For sample No. 2, a strong correlation betweenZn concentration and MS (r = 0.82) was observed; concentrations of

Fe, Cr, and Fe–Ni–Cr were significantly correlated with MS (r = 0.71,0.71, and 0.76, respectively); concentration of Pb, Si and Ca were nega-tively correlated with MS (r = −0.53, −0.76, −0.58); and Ni concen-trations were observed to be weakly correlated with MS (r = 0.32).The correlation coefficients between MS and Si, Ca (r = 0.45, 0.59) in

Fig. 4 (continued).

Table 3Correlation coefficients between magnetic susceptibility and natural logarithm of elements concentrations.

MS LnFe LnNi LnCr Ln(Fe + Ni + Cr) LnZn LnPb Ln(Zn + Pb) LnSi LnCa

No. 1 0.45⁎⁎ 0.18 0.17 0.45⁎⁎ 0.60⁎⁎ −0.42⁎⁎ 0.18 0.45⁎⁎ 0.59⁎⁎

No. 2 0.71⁎⁎ 0.32 0.71⁎⁎ 0.76⁎⁎ 0.82⁎⁎ −0.53⁎⁎ 0.20 −0.76⁎⁎ −0.58⁎⁎

⁎⁎ Correlation is significant at the 0.01 level.

22 Y. Li et al. / Journal of Geochemical Exploration 146 (2014) 17–26

Table 4Correlation coefficients between different elements.

Sample LnFe LnNi LnCr Ln(Fe + Ni + Cr) LnZn LnPb Ln(Zn + Pb) LnSi LnCa

No. 1 LnFe 1LnNi 0.60⁎⁎ 1LnCr 0.17 0.35⁎ 1Ln(Fe + Ni + Cr) 1 0.60⁎⁎ 0.17 1LnZn 0.36⁎ 0.48⁎⁎ 0.78⁎⁎ 0.36⁎ 1LnPb −0.22 0.15 0.09 −0.21 −0.03 1Ln(Zn + Pb) 0.08 0.33⁎ 0.46⁎⁎ 0.08 1 0.46⁎⁎ 1LnSi 0.63⁎⁎ 0.18 −0.05 0.63⁎⁎ 0.17 −0.68⁎⁎ −0.26 1LnCa 0.07 0.36 0.76⁎⁎ 0.07 0.90⁎⁎ 0.19 0.86⁎⁎ −0.14 1

No. 2 LnFe 1LnNi 0.67⁎⁎ 1LnCr 0.41⁎ 0.16 1Ln(Fe + Ni + Cr) 0.99⁎⁎ 0.66⁎⁎ 0.50⁎ 1LnZn 0.84⁎⁎ 0.48⁎⁎ 0.67⁎⁎ 0.84⁎⁎ 1LnPb −0.28 −0.09 −0.6 −0.32 −0.40⁎ 1Ln(Zn + Pb) 0.36⁎ 0.23 −0.02 0.3 0.51⁎⁎ 0.54⁎⁎ 1LnSi −0.92⁎⁎ −0.57⁎⁎ −0.31⁎ −0.91⁎⁎ −0.57⁎⁎ 0.32 −0.31 1LnCa −0.89⁎⁎ −0.65⁎⁎ −0.12 −0.86⁎⁎ −0.44⁎ 0.04 −0.51⁎ 0.84⁎⁎ 1

⁎ Correlation is significant at the 0.05 level.⁎⁎ Correlation is significant at the 0.01 level.

23Y. Li et al. / Journal of Geochemical Exploration 146 (2014) 17–26

sample No. 1 suggested a positive relationship between MS and Si, Caprobably because of mineral paragenesis of pyrite and quartz, andsphalerite and calcite (Fig. 2). We also can observe a positive relation-ship between Fe and Si (r = 0.63), and Zn and Ca (r= 0.90). The min-eral paragenesis of pyrite and quartz, and sphalerite and calcite (Fig. 2)may be led to the pseudo-linear variation of the MS for the measure-ments along the sample No. 1, which should be further studied.

These different correlation coefficients are probably produced fromthe abundance of elements concentrations and theirmagnetism charac-teristics within the different samples. For example, the average Cr con-tent in sample No. 2 was observed as 50 times greater than that insample No. 1, resulting in different correlation coefficients between Crcontentwith themeasuredMS (r= 0.71, 0.17). In both samples,we ob-served Zn to exhibit the strongest positive relationship with the mea-sured MS, whereas Pb exhibited negative. These observations indicatethat the abundance of sphalerite and galena contributed positivelyand negatively to the value of MS, respectively. The elements Fe, Ni,and Cr did not form minerals solely in sample No. 1, mainly existed inthe lattice of sphalerite as isomorphism. In sample No. 2, Fe mainly ex-ists in the form of pyrite and marcasite, less Fe exists in the lattice ofsphalerite; Ni and Cr still mainly exist in the lattice of sphalerite as iso-morphism. The electron microprobe (EMPA) spectra of sphalerite andgalena in the samples revealed the presence of Fe, Ni, and Cr withinthe sphalerite (Fig. 6). Niwas not detected in the sphalerite probably be-cause its concentration was below the minimum detection limit. Thereason that Zn has a positive correlation with MS value is probablydue to Fe2+ (0.74 Å), Ni2+ (0.69 Å), Cr3+ (0.63 Å) existing in the latticeof sphalerite as isomorphism, because they have similar ionic radius tothe Zn2+ (0.74 Å). It is known that Ni2+ and Cr2+ are paramagnetic;Fe2+ in pyrite can be paramagnetic or diamagnetic; whereas, it is para-magnetic in marcasite and sphalerite, and the MS of marcasite is largerthan pyrite (Mohindar and Jagadeesh, 1977). The presence of paramag-netic ions of Fe2+, Ni2+, and Cr2+ in the sphaleritemade a huge positivecontribution to theMS of sphalerite, leading to a positive correlation be-tween Zn and the measured MS. However, the radius of Fe2+ (0.74 Å),Ni2+ (0.69 Å), Cr3+ (0.63 Å) are significantly different from the ionic ra-dius of Pb2+ (1.20 Å), therefore, they hardly enter the lattice of galena(Liu et al., 1984), resulting in a negative correlation between Pb contentwith MS.

Rock MS (κr) is assumed to be proportional to the volume contentand MS (κi) of the minerals that make up the rock (Lang et al., 2011).

Here, we observed that Zn concentration is correlated more stronglywithMS thanwith Pb concentration, and also Zn concentration is great-er than that of Pb.Moreover, Pb concentrationwas greater in sampleNo.2 than in sample No. 1, producing a stronger correlation between Pb andMS in the sample No. 2. We observed similar phenomena for the rela-tionships between Fe and Cr with MS.

The correlation between Zn and MS was observed stronger insample No. 2 (r = 0.82) than in sample No. 1 (r = 0.60), despitethe fact that the abundance of Zn was similar in both samples.The abundances of Fe–Ni–Cr revealed larger increases in sampleNo. 2 than in sample No. 1, leading to a stronger correlation be-tween Zn and Fe–Ni–Cr in sample No. 2 (r = 0.84) than in sampleNo. 1 (r = 0.36). These results suggest that the presence of Fe2+,Ni2+, and Cr2+ in sphalerite has a significant influence on the MSvalue of sphalerite; they could greatly enhance the MS value ofsphalerite. In sample No. 2, Zn concentration was observed nega-tively correlated with Pb concentration, perhaps because sphaler-ite and galena are distributed unequally along the measurementpoints: Zn concentration gradually increased from the startingpoint to the end, whereas Pb concentration decreased. Moreover,a strong positive correlation between Zn and Fe concentrationswas observed due to sphalerite being accompanied by pyrite andmarcasite. The sample likely had a limestone protolith, which isdiamagnetic and exhibited MS values ranging from −18.8 × 10−5

SI to −10.0 × 10−5 SI. Typically, the MS values of rock increasewith enrichment of sphalerite that includes paramagnetic ionsand are typically larger when it is accompanied by paramagneticor ferromagnetic minerals.

3.3. Cluster and stepwise regression analysis

To further investigate the contribution of various elements torock MS, we used the SPSS 19.0 statistical analysis software to con-duct the cluster analysis among MS and various elements and thestepwise regression to regress MS on the elemental concentrations.In the result of cluster analysis (Fig. 7), the abscissa represents thedistance between variables (using Pearson correlation coefficientas the standard of distance), the nearer the distance, the closerthe relationship between variables. In sample No. 1 (Fig. 7A), Zn,Ca and Cr showed the most important relationship with the MS,followed by Fe and Si. The farthest distance between MS and Pb

Fig. 5. Scatter plots of magnetic susceptibility versus natural logarithm of concentrations of Fe + Ni + Cr, Zn + Pb, Zn and Pb for sample No. 1 (A) and No. 2 (B).

24 Y. Li et al. / Journal of Geochemical Exploration 146 (2014) 17–26

indicated the weakest relationship. In sample No. 2 (Fig. 7B), Znshowed a strong relationship with MS, followed by Fe, Cr and Ni,but the group of Ca, Si, and Pb exhibited the further distances withMS. The dendrogram indicated that the MS could be considered asan indicator for the enrichment of sphalerite in the study area.

Selection of factors in the stepwise regression model was based onprobability ≤0.05 (Freund and Littell, 2000). The following model wasused:

y ¼ b0 þ b1 F1 þ b2 F2 þ…þ bn Fn þ ε ð1Þ

Element Weight% Atomic%

S K 11.90 46.60

Pb M 88.10 53.40

Totals 100.00

Element Weight% Atomic%

S K 31.83 49.02

Cr K 0.23 0.22

Fe K 0.27 0.24

Zn K 65.77 49.68

Cd L 1.90 0.83

Totals 100.00

A

B

Fig. 6. The resulting electron microprobe spectra for galena (A) and sphalerite (B) within sample No. 2.

25Y. Li et al. / Journal of Geochemical Exploration 146 (2014) 17–26

where y represents the estimatedmagnetic susceptibility, b0 to bn are co-efficients, F1 to Fn are elemental concentrations (Fe, Ni, Cr, Zn, Pb, Si, Ca),and ε represents the residual error.We selected the best predictivemodelbased on the determination coefficient of R2 (Karimi et al., 2011).

The resulting multiple linear regression analyses between MS andelemental concentrations (Table 5) demonstrated that Zn and Si, andZn, Si, and Cr can explain the majority of the observed changes in MSfor samples No. 1 and No. 2, respectively. In sample No. 1, Zn and Sihad high correlation coefficients (Table 3) while Ca was not includedin the stepwise regression model probably due to the co-linearity andsimilar distribution of Zn and Ca in the sample (Table 4, Fig. 2A). In sam-ple No. 2, Fe also had a high correlation coefficient (Table 3) butwas notincluded in the stepwise regression model, probably due to the co-linearity and similar distribution of Zn and Fe (Table 4, Fig. 2B). The re-sult of multiple linear regression analysis between MS and elementalconcentrations indicated that in the study area rock MS is under thepowerful influence of sphalerite with paramagnetic ions, such as Fe2+,Ni2+, and Cr2+. Moreover, Fe2+ and Cr2+ played essential roles onthe change of magnetic susceptibility of sphalerite. Sphalerite withinparamagnetic ions made a positive contribution to rock MS while thematrix minerals such as quartz made a negative contribution.

4. Conclusions

We utilized the Itrax core scanner, a fast and nondestructive instru-ment, to obtain high-resolution measurements of the magnetic suscep-tibility and elemental variations of the Pb-Zn–mineralization of theJinding Pb–Zn deposit. Based on statistical analyses (correlation analy-sis, cluster analysis and stepwise multiple linear regression analysis),the following conclusions can be drawn.

(1) The MS of mineralized breccia in the study area was posi-tively correlated with concentrations of Zn, Fe, and Fe–Ni–Cr, and negatively with concentration of Pb. The resultingcluster multiple linear regression analysis suggested thatthe sphalerite with paramagnetic ions provide a consider-able influence on the rock MS, and Fe2+ and Cr2+ played es-sential roles on the change of magnetic susceptibility ofsphalerite, and matrix minerals such as quartz makes a neg-ative contribution to rock MS. It should be aware that in na-ture not all sphalerites contain Fe, Ni, Cr contents, meaningthat this conclusion is valid when sphalerite contains suchelements.

Fig. 7. Dendrogram for natural logarithm of elements concentrations with MS for sampleNo. 1 (A) and No. 2 (B).

26 Y. Li et al. / Journal of Geochemical Exploration 146 (2014) 17–26

(2) The Itrax core scanner instrument offers potential for theidentification of different mineralized rocks. In particular, itallows fast and nondestructive collection of multiple datasetsconsisting of high-resolution optical images, high-resolutionX-radiographic images, micro-X-ray fluorescence data, andMS measurements.

Table 5Stepwise regression results.

Model Unstandardizedcoefficients

Standardizedcoefficients

B Std. Error t Sig.

No. 1 (Constant) −7.080 0.964 −7.344 0.000LnZn 2.382 0.427 5.583 0.000LnSi 0.353 0.064 5.554 0.000R2 0.614

No. 2 (Constant) −2.196 0.050 −43.676 0.000LnZn 0.039 0.017 2.297 0.033LnSi −0.044 0.010 −4.442 0.000LnCr 0.071 0.031 2.314 0.031R2 0.849

Acknowledgments

We thank Prof. Changjiang Li and M. Gettings for constructive com-ments and suggestions, which improve the quality of this study. This re-search benefited from the jointfinancial support froma research projecton “Quantitative models for prediction of strategic mineral resources inChina” (201211022) by China Geological Survey, the Fundamental Re-search Funds for the Central Universities, China University ofGeosciences (Wuhan) (Nos. CUG120501 and CUG120116), the NationalNatural Science Foundation of China (No. 41372007), and the Programfor New Century Excellent Talents in University (NCET-13-1016).

References

Aydin, A.,Ferré, E.C.,Aslan, Z., 2007. Themagnetic susceptibility of granitic rocks as a proxyfor geochemical composition: example from the Saruhan granitoids, NE Turkey.Tectonophysics 441, 85–95.

Cao, Y., Li, S., Shen, J., Yao, M., Li, Q.,Mao, F., 2007. Relation between the characteristics ofmagnetic susceptibilities of altered rocks and gold mineralization in the Qianhegold deposit, Henan. Geol. China 34, 1082–1090 (In Chinese with English Abstract).

Clark, D.A., 1997. Magnetic petrophysics and magnetic petrology: aids to geological inter-pretation of magnetic surveys. J. Aust. Geol. Geophys. 17, 83–103.

Croudace, I.W.,Rindby, A., Rothwell, R.G., 2006. ITRAX: description and evaluation of anew multi-function X-ray core scanner. In: Rothwell, R.G. (Ed.), New Techniques inSediment Core Analysis. Geological Society of London, 267, pp. 51–63.

Deng, Y., Fang, W.,Guo, M.,Liu, Y., 2010. The research and trial of magnetic susceptibilityparameters on the iron–copper exploration in Dongchuan. Contrib. Geol. Miner.Resour. Res. 25, 157–162 (In Chinese with English Abstract).

Dunlop, D.J.,Özdemir, Ö., 1997. Rock Magnetism: Fundamentals and Frontiers. CambrigeUniversity Press, UK, p. 573.

Freund, R.J.,Littell, R.C., 2000. SAS System for Regression. SAS Inst., Cray, NC, pp. 199–224.Gao, H., 1989. Some views on strata and sedimentary facies in Jingding lead–zinc deposit.

Earth Sci. J. China Univ. Geosci. 14, 457–465 (In Chinese with English Abstract).Hrouda, F.,Müller, P.,Hanák, J., 2003. Repeated progressive heating in susceptibility vs.

temperature investigation: a new palaeotemperature indicator? Phys. Chem. Earth28, 653–657.

Hu, R., Turner, G., Burnard, P.G., Zhong, H.,Ye, Z., Bi, X., 1998. Helium and argon isotopicgeochemistry of Jingding superlarge Pb–Zn deposit. Sci. China Ser. D 41, 443–448.

Karimi, R.,Ayoubi, S., Jalalian, A.,Hosseini, A.R., Afyuni, M., 2011. Relationships betweenmagnetic susceptibility and heavy metals in urban topsoils in the arid region ofIsfahan, center Iran. J. Appl. Geophys. 74, 1–7.

Lang, Y.,Hu, D.,Liu, C.,Zhang, B., Lu, B.,Wang, P., 2011. Mineralogy study of magnetic sus-ceptibility of rocks along the coast of the northern South China Sea. Chin. J. Geophys.54, 573–587 (In Chinese with English Abstract).

Liu, Y., Cao, L., Li, Z.,Wang, H., Chu, T., Zhang, J., 1984. Element of Geochemistry. SciencePress, Beijing (In Chinese).

Mohindar, S., Jagadeesh, M.S., 1977. Temperature-dependent magnetic susceptibility ofmarcasite (FeS2). Phys. Rev. B 20, 3897–3899.

Sandgren, P., Snowball, I., 2002. Application of mineral magnetic techniques topaleolimnology. Dev. Paleoenviron. Res. 2, 217–237.

Spaldin, N.A., 2010. Magnetic materials: fundamentals and applications, second ed. Cam-bridge University Press, New York.

Spiteri, C.,Kalinski, V.,Rösler, W.,Hoffmann, V.,Appel, E., 2005. Magnetic screening of apollution hospot in the Lausitz area, Eastern Germany: correlation analysis betweenmagnetic proxies and heavy metal contamination in soils. Environ. Geol. 49, 1–9.

Wang, L.,Fang, W.,Zhang, D.,Gao, H., 2009. Research and application of magnetic suscep-tibility measurement of rocks and minerals in Riconada, Chile. Miner. Resour. Geol.23, 473–479 (In Chinese with English Abstract).

Wu, H., 1988. Magnetic fabric and its application to strain analysis. Acta Petrol. Sin. 1,94–98 (In Chinese with English Abstract).

Xu, H.,Jin, Z.,Ou, X., 2006. Lithology determination of rocks from CCSD 100–2000 mmainhole by magnetic susceptibility and density using discriminant function analysis.Earth Sci. J. China Univ. Geosci. 31, 513–519 (In Chinese with English Abstract).

Xue, C.,Chen, Y.,Wang, D.,Yang, J.,Yang, W.,Zeng, R., 2003. Geology and isotopic compo-sition of helium, neon, xenon and metallogenic age of the Jingding and Baiyangpingore deposits, northwest Yunnan, China. Sci. China Ser. D 46, 790–800.

Xue, C.,Liu, S.,Chen, Y.,Zeng, R.,Zhao, S., 2004. Giant mineral deposit and their geodynamicsetting in the Lanping Basin, Yunnan, China. Acta Geol. Sin. 78, 368–374.

Xue, C.,Zeng, R.,Liu, S.,Chi, G.,Qing, H.,Chen, Y.,Yang, J.,Wang, D., 2007. Geologic, fluid in-clusion and isotopic characteristics of the Jinding Zn–Pb deposit, western Yunnan,South China: a review. Ore Geol. Rev. 31, 337–359.

Yu, J.,Hu, X., 2005. Application of data statistical analysis with SPSS. Post & Telecom Press,Beijing, pp. 163–173 (In Chinese).

Zhang, W.,Yu, L.,Xu, Y.,Sun, Z., 1998. The instructions of the sediment's magnetic propertyon ferric reduction and its application to heavy metal contamination. Chin. Sci. Bull.43, 34–43.

Zhao, Z., 2007. The Jinding Pb–Zn deposit of Lanping—a land facies sedex ore deposit.Geol. Yunnan 26, 1–14 (In Chinese with English Abstract).

Zuo, R., 2013. ITRAX: a potential tool to explore the physical and chemical properties ofmineralized rocks in mineral resource exploration. J. Geochem. Explor. 132, 149–155.