1 Introduction

The primary application of the Lu-Hf isotopic system is to trace crustal evolution and growth through time. Its direct applications to regional mineral systems include, but are not restricted to, mineralisation styles directly related to crustal growth via addition of juvenile mantle material rich in compatible elements (e.g. most Ni and platinum-group element deposits, and kimberlites), and those developed at large-scale tectonic interfaces between isotopically disparate crustal blocks (e.g. porphyry Cu-Au and orogenic Au deposits). The Lu-Hf system can be used to isotopically map and identify prospective crustal blocks and boundaries, to provide a combined geochemical-geophysical view of the lithosphere, and to act as a ‘paleo-geophysical’ tool (Hartnady et al. 2018) to reconstruct crustal and lithospheric changes through time. When integrated with geological, geochemical, structural and geophysical datasets, the Lu-Hf isotope system can be a powerful tool for refining mineral systems modelling (McCuaig et al. 2010) and narrowing the search space for a range of mineralisation styles.

The Lu-Hf isotopic system can be used to track crustal growth and evolution through time in a similar fashion to the Sm-Nd isotopic system (Champion and Huston 2023). One of the most important advantages of the Lu-Hf system is the ability to measure Lu-Hf in individual zircons. This is possible because the combination of high Hf (1–2 wt%) and low Lu (ppm-ppb) in zircon ensures that the preserved (present-day) Hf isotopic composition is close to initial (crystallisation) values; it requires little correction for ingrowth of radiogenic (post-crystallisation) Hf. It also means that Lu-Hf measurements can be securely linked to U-Pb measurements directly constraining the age of the zircon. Measuring Lu-Hf on zircon also allows direct integration with other zircon measurements beyond U-Pb—such as oxygen isotopes and trace elements—allowing zircon to be ‘fingerprinted’ in multiple systems.

In this chapter, the fundamentals of the Lu-Hf system are outlined to provide the basis for application to exploration and metallogenic investigations. This includes better integration into exploration programs, whether by designing new data collection campaigns and their interpretation, or evaluating the quality and impact of existing datasets. Lutetium-Hf isotope data are useful in early planning stages to identify the search space for province selection and, depending on the commodity of interest and geological setting, may also be valuable at the district scale.

2 The Lutetium-Hafnium Isotopic and Geochemical System

The Lu-Hf isotopic system is based on the beta decay of 176Lu to 176Hf, with a half-life of about 37 billion years (Söderlund et al. 2004). Lutetium and Hf are fractionated differently during crust formation processes (Vervoort 2014). During mantle melting, Hf is more incompatible than Lu, so fractionation processes enrich the crust in Hf relative to Lu, and different crustal reservoirs (e.g., continental vs oceanic) take on different Lu-Hf isotopic characteristics. As 176Lu decays, reservoirs with higher Lu/Hf will have a correspondingly high radiogenic 176Hf abundance relative to the non-radiogenic and stable 177Hf isotope (Eq. 1; Fig. 1).

There is widespread agreement amongst the Hf-isotope community on a value of 1.867 × 10−11 yr−1 for the 176Lu decay constant (Scherer et al. 2001; Söderlund et al. 2004; Thrane et al. 2010). However, older published literature may use an alternate decay constant (e.g., Sguigna et al. 1982; Bizzarro et al. 2003). When compiling data from different sources, each author’s choice of decay constant should be confirmed. This is particularly important in older terranes, or for high-Lu samples, where more 176Hf will have accumulated from 176Lu decay.

The specific equation for the evolution of the Hf isotope composition is:

$$ \left( {\frac{{{}^{176}Hf}}{{{}^{177}Hf}}} \right)_{now} = { }\left( {\frac{{{}^{176}Hf}}{{{}^{177}Hf}}} \right)_0 + \frac{{{}^{176}Lu}}{{{}^{177}Hf}}\left( {e^{\lambda t} + 1} \right) $$
(1)

where 176Hf/177Hf and 176Lu/177Hf are the present-day (measured) values, (176Hf/177Hf)0 is the initial Hf isotope composition at time t in the past, and λ is the decay constant of 176Lu.

Crustal differentiation processes yield reservoirs with differing initial Lu/Hf values, and the decay of 176Lu will result in diverging 176Hf/177Hf between reservoirs over time. However, the long half-life of 176Lu means that isotopic variations between reservoirs are generally very small, and distinguishing between them requires high analytical precision and accuracy.

2.1 Epsilon Hf (ɛHf)

The isotope 176Hf continues to accumulate in all reservoirs over time, so the same 176Hf/177Hf value can have very different interpretations at different geologic times and in different geologic settings. The εHf parameter normalizes 176Hf/177Hf values against that of the chondritic uniform reservoir (CHUR), which serves as an estimate of the bulk Earth 176Hf/177Hf. This system is easier to work with than the isotopic ratios, because differences in 176Hf/177Hf are geologically significant at the fourth decimal place. The equation for calculating εHf at a time (t) is:

$$ \rm{\varepsilon }{\text{Hf}}_{\text{t}} = { 1}0^{4} \times \, \left[ {\left( {^{{176}} {\text{Hf}}/^{{177}} {\text{Hf}}} \right)_{{\text{sample}},{\text{ t}}} /\left( {^{{176}} {\text{Hf}}/^{{177}} {\text{Hf}}} \right)_{{\text{CHUR}},{\text{ t}}} -{ 1}} \right] $$
(2)

2.2 Model Ages

As with the Sm-Nd system, model ages of crustal reservoirs can be calculated from Lu-Hf data in order to estimate the time since extraction from the mantle (i.e. crust formation events). There are several assumptions that must be made to deduce these model ages, and so it is important not to confuse them with accurate, precise determinations of geologically-meaningful rock-forming event ages such as igneous crystallisation U-Pb or Ar/Ar ages. Model ages are estimates of average mantle extraction age, and do not directly date geologic events, so they should primarily be used as petrogenetic tools for comparing rock suites (Spencer et al. 2020).

The simplest type of model age is a single-stage model age, which assumes the sample was extracted directly from the mantle and has evolved as a closed system ever since, and that the Lu/Hf value of the analysed material reflects bulk-rock values. This is calculated by using the measured 176Lu/177Hf and 176Hf/177Hf to track backwards in time until the depleted mantle model curve is intercepted (Fig. 1). In this case, it is not necessary to know the crystallisation age of the rock. There are significant problems with this method, particularly when measuring Lu-Hf in zircon because zircon fractionates Lu and Hf extremely strongly, leading to a Lu/Hf in zircon that can be orders of magnitude lower than the protolith from which the zircon formed. The result is that Hf single-stage model ages calculated from zircons are always minimum estimates of the model age of extraction of associated crust from the mantle.

Fig. 1
figure 1

(a) 176Hf/177Hf and (b) epsilon-Hf (εHf) plots showing main principles of the Lu-Hf isotope system as applied to measurements in zircon. To calculate the 176Hf/177Hf (a) or εHf (b) at the time of crystallisation (εHft), the measured 176Hf/177Hf of a zircon (Hf/Hf)zrn,0 is corrected for the ingrowth of radiogenic 176Hf since crystallisation based on the 176Lu/177Hf in the zircon. To calculate a single-stage model age (TDM), this trend is extrapolated until it intercepts the depleted mantle (DM) curve. A two-stage model age (T2DM) uses an estimate for the 176Lu/177Hf composition of the source to extrapolate from the crystallisation event until the depleted mantle intercept. CHUR, an estimate of bulk earth composition continues to accumulate 176Hf as a function of 176Lu abundance. On an εHf plot (b), CHUR is defined as zero throughout Earth history. [Hf/Hf]CHUR,0 is the 176Hf/177Hf of CHUR at present day. [Hf/Hf]DM,0 is the 176Hf/177Hf of the depleted mantle at present day (t=0).

Full size image

A better approximation of the time of extraction from the mantle (crustal residence age) is derived from a two-stage model age, in which the measured isotopic ratios are traced back in time the same way as the single-stage model age to the time of zircon crystallization, at which point an estimated Lu/Hf for the protolith or source region is used to calculate the intersection point with the depleted mantle curve (Fig. 1). This gives a more reasonable model age estimate, particularly when using measured values from zircons, but it does also entail more assumptions.

Firstly, the age of the rock (or analysed mineral grain) must be well-constrained. This is generally straightforward for zircon-bearing rocks, as the U-Pb age is usually determined on the same zircons, either beforehand or simultaneously with the Lu-Hf isotopic measurements.

Secondly, it is necessary to assign an initial Lu/Hf value to the precursor source material(s). Given that these sources are commonly destroyed or substantially modified during the crustal differentiation process responsible for forming the existing rock or minerals, it can be difficult to appropriately characterise their Lu/Hf, particularly when these sources are mixed, or at depth. Bulk 176Lu/177Hf is most often assigned the value of 0.015 (Griffin et al. 2002), estimated from average continental crust, though other researchers use other values based on other evidence or models for a different source, such as mafic or felsic crust (Amelin et al. 1999; Kemp et al. 2006; Vervoort and Kemp 2016).

Thirdly, there are multiple models for depleted mantle composition through time, and other suggested source regions, such as that similar to island arc crust (Dhuime et al. 2011; Vervoort and Kemp 2016).

Because of the large differences between different methods used to estimate model mantle evolution curves, the reader should be careful to understand which method the author has reported and why, particularly when compiling data from multiple sources. Figure 1 demonstrates the potential differences between the two types of model ages. The two-stage model age is so strongly preferred in recent literature that the ascribed notation for this model age is often TDM which can easily be confused with the single-stage model age notation.

2.3 Hf Isotope Evolution Models

The isotopic composition of Earth reservoirs evolves as more 176Hf is accumulated through decay of 176Lu (Fig. 1). This means that material extracted from the mantle reservoir at different times has different isotopic compositions. It is these differences which form the basis of the Lu-Hf system for characterising the nature of the crust and mantle lithosphere.

When a melt from a depleted-mantle source is extracted and emplaced at shallower levels, this material has an isotopic composition that lies on the depleted mantle curve, and is considered ‘radiogenic’ due to its high proportion of radiogenic 176Hf relative to non-radiogenic 177Hf. Some authors also describe these compositions as ‘juvenile’. In contrast, where melt is extracted from crustally-derived material, the corresponding terminology is ‘unradiogenic’ and ‘evolved’.

Figure 2 shows a synthesis of various ways that geological processes influence 176Hf/177Hf and eHf. Crust-mantle differentiation begins in the early Earth, marked on the figure by the initial divergence of the CHUR and depleted mantle curves. At position t1, partial melting in the depleted mantle generates new crust. If there is no more mantle-derived material added to the crust after this point (i.e., crust is stable or internally reworked), the extracted crustal material simply accumulates 176Hf proportional to its 176Lu abundance. If there is a second partial melting event in the depleted mantle at t2, any mixing between crust from the t1 and t2 events will have εHf intermediate between the two isotopic compositions.

Fig. 2
figure 2

Epsilon-Hf (εHf) plot showing generalised geological processes that influence Hf-isotope signatures. Material derived directly from a depleted mantle source will fall along the depleted mantle curve at the time of extraction. As this material resides in the crust, its Hf composition will become unradiogenic relative to depleted mantle, and remelting of these crustal sources will have an εHf signature that lies along the same trajectory indicated by shaded grey bars. Mixing between multiple sources extracted from the mantle at different times leads to intermediate εHf between the various sources, with exact signature a complex function of the proportions and Hf content of the disparate sources.

Full size image

Confidence in interpreting these types of events stems from the fact that the Lu-Hf system is robust, particularly in zircon (Cherniak and Watson 2003). This is because zircons themselves are generally robust: hard, refractory and resistant to alteration, thus able to remain intact during sedimentary, metamorphic and even some igneous processes. By comparison, the bulk-rock Lu-Hf isotope system is more susceptible to weathering- or alteration-related isotopic disturbance.

Not all influences on εHf come from geological sources. Analytical and data reduction artifacts can introduce apparent trends which do not reflect a geological process. Vervoort and Kemp (2016) and Spencer et al. (2020) provide a detailed synopses of many of the potential pitfalls and issues in collecting Lu-Hf isotopic data, as well as advice on analytical and interpretive approaches to avoid these problems.

3 Analytical Methods

The Lu-Hf isotopic system reflects similar geological processes to the Sm-Nd system, but has historically been more difficult to measure (Vervoort 2014). Bulk-rock Lu-Hf systematics can be interpreted in largely the same way as Sm-Nd systematics (Kinny and Maas 2003; Vervoort 2014).

The low abundance of Lu and Hf in Earth materials, the very small differences in isotopic ratios that mark geologically significant variation, and the large isobaric interferences on the isotopes of interest, pose difficulties for precise isotopic measurements (Halliday et al. 1995). Thermal ionisation mass spectrometry (TIMS) was the first technique used to collect high-precision bulk-rock Lu-Hf data (Patchett and Tatsumoto 1980; Patchett et al. 1982; Corfu and Noble 1992), but the high ionisation potential of Hf makes these measurements difficult (Vervoort 2014). The high Hf content of zircon makes analysis of mineral separates easier than whole-rock analysis (Kinny and Maas 2003). As inductively-coupled plasma mass spectrometry (ICP-MS), and in particular multi-collector (MC) capability (Walder and Freedman 1992; Halliday et al. 1995; Blichert-Toft and Albarède 1997), and laser ablation (LA) functionality (Thirlwall and Walder 1995; Woodhead et al. 2004) has become more accessible, use of the Lu-Hf system has increased dramatically.

Lu-Hf data are now most routinely collected by laser ablation microanalysis (Kinny and Maas 2003), which allows close links to the age information necessary for appropriately determining time-corrected parameters (Amelin et al. 1999, 2001; Gerdes and Zeh 2006). Microanalysis also permits greater understanding of how multiple components in a single sample contribute to the isotopic signature, and therefore the history of the sample (Woodhead et al. 2004; Mole et al. 2014; Kirkland et al. 2015). Data collection is very rapid, however sample preparation involves mineral separation and preparing grain mounts in addition to the crushing steps of whole-rock analysis, and requires further mineral characterisation techniques such as cathodoluminescence imaging (Kinny and Maas 2003).

LA-MC-ICP-MS analysis has enabled rapid (on the order of 1–2 min per analysis; Thirlwall and Walder 1995; Woodhead et al. 2004) collection of data from individual minerals while sampling very small volumes of material (usually below 100 µm diameter ablation pit and excavating 0.5–1.0 µm per second; Kinny and Maas 2003). By targeting individual mineral grains, it is possible to gain extra insights from ‘unmixing’ different components of a rock. Several of the case studies presented in this chapter take advantage of this ability.

Laser ablation analyses are increasingly benefiting from split-stream technology, whereby material ablated from the target mineral can be ‘split’, and directed to multiple mass-spectrometer systems for simultaneous measurement. This permits truly synchronous acquisition of Lu-Hf using a multicollector mass spectrometer while using a second mass spectrometer to measure U-Pb, and trace elements (Woodhead et al. 2004; Yuan et al. 2008). This is beneficial not only because of the increased richness of data but also allows real-time assessment of Lu-Hf analysis quality. Oxygen isotopic analysis is also commonly paired with Hf and U-Pb analysis, and provides information on alteration and crustal recycling processes (Valley 2003; Kemp et al. 2007; Harrison et al. 2008). These oxygen analyses are usually done by ion-probe before laser analysis.

The technical aspects of Lu-Hf data collection (such as sensitivity improvements, down-hole fractionation and resolution of triple-interferences at mass 176) have been thoroughly reviewed by other authors (e.g., Kinny and Maas 2003; Fisher et al. 2014; Vervoort and Kemp 2016; Spencer et al. 2020). These reviews provide useful discussion of the power and pitfalls of Lu-Hf analyses, and also provide further criteria for determining data quality and for aiding interpretation of a diverse range of data trends—not all of which are geologically meaningful. Other papers cover the specifics of Hf analysis of zircons by LA-ICP-MS (Amelin et al. 2000; Woodhead et al. 2004; Xie et al. 2008; Yuan et al. 2008; Gerdes and Zeh 2009).

4 Geological Applications of the Lu-Hf Isotope System

Analysis of magmatic rocks or minerals can be used to characterise crustal blocks through time (Mole et al. 2014; Cross et al. 2018; Waltenberg et al. 2018) and to identify subsurface source regions (Flowerdew et al. 2009; Dolgopolova et al. 2013). The meaning and origin of geophysical and paleo-thermal anomalies can also be investigated (Hartnady et al. 2018; Siegel et al. 2018; Waltenberg et al. 2018). The Lu-Hf isotope system can be used to understand the origins of ore-bearing fluids in hydrothermal systems (Westhues et al. 2017). Further, Lu-Hf isotopic analysis can determine the metal transport pathways, and mechanisms of mineralisation caused by magmatic processes (Murgulov et al. 2008; Mole et al. 2014; Kirkland et al. 2015) and refine mineral systems models (Hou et al. 2015; Kirkland et al. 2015).

Analysis of metamorphic zircons can reveal information about the source of fluids involved in the metamorphic process. If the metamorphic domains have a different Hf signature, this may indicate an exotic source for metamorphic fluids, including the possibility of mantle-derived fluids. If the signature is the same as earlier magmatic domains, the source of metamorphic material is more likely to be proximal (e.g. Wu et al. 2009; Kirkland et al. 2015).

One of the strengths of the Lu-Hf system as applied to mineral separates is the ability to retrieve fossil isotopic information even when characteristics of the original rock have been lost to weathering and erosion. Lutetium-Hf studies of detrital minerals (e.g., zircon) allow investigation of the sources of detritus and characterized minerals can be used as indicators to detect possible nearby intrusions, for example in zircons derived from kimberlites in nearby creek catchments (Griffin et al. 2000; Belousova et al. 2001; Batumike et al. 2009). Analysis of detrital material in a catchment can be used to survey the entire provenance spectra, and narrow down targeting of regions. Isotopic ‘fingerprints’ can be used to match sediments with a library of magmatic data to determine provenance, and correlate sedimentary basins (Doe et al. 2013). This can be expanded by also collecting O-isotopes and split-stream U-Pb and trace elements on the same zircons to characterise sources in more detail (e.g., Belousova et al. 2002; Griffin et al. 2006; Lu et al. 2016; Purdy et al. 2016).

5 Target Materials and Minerals

The dominant method of Hf-isotope determination is analysis of zircons by MC-ICPMS methods. Zircon substitutes Hf in the Zr site in its structure due to identical charge and similar ionic radius between the two elements—most terrestrial zircons have 0.5–2.0 wt% HfO2 (Speer and Cooper 1982), making zircon the primary host of Hf in most rocks (Kinny and Maas 2003). Zircon also excludes Lu, making it ideal for Lu-Hf isotopic studies (Amelin et al. 1999). Because it incorporates Hf so effectively, the original 176Hf/177Hf at the time of zircon formation is well-preserved within zircon crystals, and only a small correction is required for ingrown 176Hf (Kinny and Maas 2003). This means that zircons preserve ‘fossil’ isotopic markers and are used to determine Hf isotope composition at the time of zircon growth. Zircon is informative for characterising a range of rock-forming processes including magmatic, metamorphic and detrital processes (Kinny and Maas 2003; Vervoort et al. 1996). Individual zircon grains are commonly complexly zoned and can robustly retain inherited cores and magmatic and metamorphic overgrowths, which can provide constraints on the growth of the zircon, and host rocks—one of the primary reasons why zircon is such a popular mineral for both geochronology and isotopic studies.

Rutile LA-ICPMS analytical methods have been developed (Luvizotto et al. 2009; Ewing et al. 2011) for igneous, metamorphic and detrital applications (e.g., Choukroun et al. 2005; Zack et al. 2011; Ewing et al. 2014). Ewing et al. (2011) documented an example of zircon and rutile Hf data from the same rock recording information on different parts of metamorphic P-T-t history and suggested that these two minerals give complementary information about metamorphism. Ewing et al. (2014) demonstrated that rutile preserves 176Hf/177Hf even through dissolution-reprecipitation events during metamorphism, provided rutile is not replaced by another mineral such as ilmenite. These authors proposed that rutile may be especially useful in UHT metamorphic rocks where metamorphic zircon is limited, and in other zircon-poor lithologies.

Likewise, baddeleyite analysis by LA-MC-ICPMS methods have been developed (Xie et al. 2008; D'Abzac et al. 2016) and these techniques have been applied to sediment-hosted baddeleyite grains (e.g., Schärer et al. 1997; Bodet and Schärer 2000), and igneous applications (Söderlund et al. 2005, 2006). Like zircon, baddeleyite contains very little Lu in its structure (Söderlund et al. 2005), so the correction for 176Hf ingrowth since crystallisation is similarly small. Baddeleyite rarely preserves inheritance and populations in magmatic rocks are predominantly cogenetic, so there may be benefit in analysing baddeleyite rather than zircon in silica-undersaturated samples with complex histories to retrieve unambiguously magmatic Lu-Hf signatures from the rock.

This chapter is focused on the use of Lu-Hf as an isotopic tracer, however minerals that incorporate the parent (176Lu) isotope can be used as geochronometers (e.g., apatite, garnet, lawsonite; Vervoort 2014). Vervoort (2014) present an overview of the geochronological applications of the Lu-Hf system.

6 Application of the Lu-Hf Isotope System to Mineral Deposit Research and Exploration

The nature and configuration of crustal blocks is critically important to understand where and when mineral deposits form. The primary value of Lu-Hf isotopic data in mineral exploration is currently in focussing the search space, and guide the extension of exploration out from regions of known mineralisation. The structures of the deep crust and lithosphere are major controls on Lu-Hf isotopic signatures, and these are also strong controls on the occurrence of many mineral deposits. Geophysical techniques can be used to image these sub-surface regions in their present-day configuration, but an isotopic approach means that it is possible to map crustal blocks and lithospheric architecture through time (Champion and Huston 2016, 2023).

Several studies have demonstrated a strong link between isotopic signatures and lithospheric thickness in young terranes (e.g., Nash et al. 2006; Zhu et al. 2011; Yang et al. 2014; Hou et al. 2015) but isotopic information is especially valuable in older regions (e.g., Australia, Canada), where the environments and crustal configurations prevailing during mineral deposit formation differ significantly from those of the modern Earth. Isotope systems such as Sm-Nd and Lu-Hf can be used as paleo-geophysical tools (Hartnady et al. 2018), to reconstruct the Earth’s dynamic history and place mineralisation in a contemporaneous context. The isotopic signature of pre-existing materials is preserved even though the pre-existing material may have been lost or is inaccessible.

Direct application of the Lu-Hf system to deposit-scale problems is an emerging field and there are not many publicly available case studies. However, larger-scale studies have value to explorers, in much the same way as regional geophysical datasets provide context. Regional studies are useful because they provide ‘baseline’ data against which camp-scale results can be compared. For example, identification of unusually radiogenic εHf in a region of non-radiogenic, old crust may indicate mantle-derived input in that area, and thus increased potential for particular types of mineralisation. Investigating temporal variations in isotopic character may indicate periods of mantle input, and enable the explorer to narrow the time window of interest and constrain timing of mineralisation. The key is to keep in mind the mineral systems approach (McCuaig et al. 2010), what components are required for the mineral deposit style of interest, and how these might be expressed isotopically.

The first three of the following case studies demonstrate some applications of the Lu-Hf isotope system in understanding metallogenesis, and emphasise the close links between Lu-Hf isotopes and lithospheric thicknesses and boundaries, both of which strongly influence mineral occurrence and endowment. The fourth case study demonstrates the power of the Lu-Hf isotope system applied to detrital minerals to act as pathfinders for kimberlite-hosted mineral deposits.

6.1 Lithospheric Controls on Mineralisation Style in the Lhasa Terrane

The Lu-Hf isotope system assists in understanding regional trends in lithospheric character and the resulting mineral potential for magmatic-associated systems. Hou et al. (2015) compiled new and published LA-ICP-MS zircon Hf and other geochemical data to investigate why magmatic-hydrothermal ore deposits occur in specific tectonic environments. These authors produced Hf model-age maps for the Lhasa Terrane of the Himalayan–Tibetan Orogen which demonstrate a correlation between the occurrence of several styles of mineralisation formed during the Jurassic-Miocene, and the Hf-signature of mineralisation-related igneous rocks.

The Hf model-age maps vary between the three distinct E-W trending crustal blocks—a central Proterozoic microcontinent with bounding Phanerozoic blocks to the north and south (Fig. 3). The Hf data also suggested the existence of two concealed N-S trending lithospheric-scale faults that cross-cut all three blocks. The authors argued that the configuration of these blocks and faults exert a first-order control on the formation of various magmatic-hydrothermal mineral deposits across the Lhasa Terrane.

Fig. 3
figure 3

Two-stage model-age map of the Lhasa terrane. Reproduced with permission from Hou et al. (2015); Copyright 2015 Society of Economic Geologists. The maps show the close spatial relationship between ore deposits and sub-terranes and boundaries mapped out by Lu-Hf isotopes. A central Proterozoic microcontinent (blue) is bounded to the north and south by two Phanerozoic blocks (red-yellow), and the Hf data (n=4762 analyses) are also suggestive of the existence of two concealed N-S trending lithospheric-scale faults (white dashed lines) crosscutting the blocks. The location of mineral deposits is closely linked to blocks of specific isotopic character, or to isotopic boundaries that may indicate regions of lithospheric weakness. A, B and C denote the author’s delineations of three domains, based on spatial differences in Hf model age. NLS = northern Lhasa subterrane, CLS = central Lhasa subterrane, SLS = southern Lhasa subterrane, BNSZ = Bangong-Nujiang suture zone, IYZSZ = Indus-Yarlung-Tsangpo suture zone, SNMZ = Shiquan River-Nam Tso Mélange Zone, LMF = Luobadui-Milashan Fault. Red lines denote Miocene normal fault systems; white dashed lines denote inferred basement faults.

Full size image

Hou et al. (2015) interpreted that many of the deposit types in the region form in proximity to terrane margins and lithosphere-scale faults (as expressed by regions with the highest Hf-isotope gradients) because these regions of structural weakness promote transport of mantle-derived material to provide the metalliferous and heat/energy inputs needed to form ore deposits. Mantle-derived magma ascending through lithospheric faults sourced metal for skarn Fe-Cu ore deposits, whereas away from terrane boundaries the lack of structural conduits limits transport of mantle-derived material—so the potential for skarn Fe-Cu deposits is low.

Where mineral deposits occur away from boundaries and major faults, the authors also demonstrated that the major mineralisation style varies based on the age and character of the crust. Porphyry Cu systems are associated with relatively young crust (T2DM < 1200 Ma), but Mo-dominated porphyry systems occur in older crust (T2DM > 1200 Ma). Fe-Cu skarns are associated with young crust, however Fe-only skarns are more strongly correlated with older crust. Granite-related Pb-Zn deposits are associated with older crustal blocks. In addition to the spatial correlation, the authors also demonstrated that many rocks in the region with high zircon εHf are rich in Cu (Fig. 4).

Fig. 4
figure 4

Whole-rock copper content vs zircon εHf for igneous rocks of the Lhasa Terrane. Reproduced with permission from Hou et al. (2015); Copyright 2015 Society of Economic Geologists. Juvenile rocks tend to contain the highest Cu-content, demonstrating a link between positive εHf values and increased copper mineralisation potential.

Full size image

Paired Lu-Hf and U-Pb age data from zircons enables further temporal constraints to be applied to tectonic reconstructions and crustal evolution models, including periods of crustal reworking and mantle input into crust. Integration of this isotopic information into reconstructions has enabled better understanding of the processes that formed the Lhasa Terrane (Zhu et al. 2011; Hou et al. 2015) and the development of a geodynamic model for its mineralisation (Hou et al. 2015).

6.2 Cautions and Considerations

Hou et al. (2015) showed that the Lu-Hf isotope system can be used as a mapping tool to identify crustal blocks of different character and origins and that boundaries between these blocks tend to be significant hosts of mineralisation. This study showed the power of high data density (4762 analyses total) to understand subtle differences across terranes. This example also demonstrates that there is no single ‘golden’ Hf-signature to look for—mineralisation can occur in old or young crust, but different styles tend to occur in regions of specific crustal character. The relative differences in Hf isotope signatures are as important as the absolute isotopic values—isotopic gradients can indicate the most prospective regions due to their correlation with lithosphere-scale faults acting as a conduit for mantle-derived material. The background isotopic signature of a terrane will be important information to consider when conducting a more localised or camp-scale study.

Hou et al. (2015) illustrated the Hf-signatures using both εHf and model-age (TDM) maps; however, mapping using the εHf parameter is not considered best-practise for comparing isotopic signatures between samples of different age, because εHf values are not time-independent (Fig. 1). The depleted mantle source continues to increase in 176Hf/177Hf, which means that the same absolute εHf value can have different implications at different times. In cases where there is significant age variation in the analysed rocks, a more robust approach is to use T2DM (Champion and Huston 2016).

7 Mapping Lithospheric Evolution Through Time and Implications for Ni Mineralisation

The previous Mesozoic-Cenozoic example illustrated the strong relationship between Lu-Hf isotopes, lithospheric structure and transport of metals into the crust. Isotopic characterisation also provides unique insight into lithospheric processes. There is a strong link between isotopic mapping and geophysical imaging techniques if the lithospheric configuration at the time of igneous emplacement is preserved. In older or dynamic terranes where current-day configurations might be very different to those of the past, isotopes can preserve signatures of configurations through time, including pre-, syn- or post-mineralization. This ability to reconstruct past configurations can help guide mineral exploration and targeting (Champion and Huston 2016; Collins et al. 2011; Mole et al. 2014).

Mole et al. (2014) conducted a zircon LA-ICP-MS Lu-Hf isotopic study in the Eastern Goldfields Province in Western Australia to understand the isotopic architecture and evolution of the lithosphere for the purposes of understanding controls on komatiite volcanism and associated Ni occurrences. Hafnium isotope data were used to generate a series of isotopic time-slices across the region. Isotopic variations in both space and time were interpreted to reflect gross lithospheric architecture. The researchers found that two older crustal blocks and one younger block existed in the time interval 3050–2820 Ma, and the configuration of these blocks controlled the location of komatiite emplacement in the Forrestania and Lake Johnston greenstone belts at 2.9 Ga (Fig. 5). The 2820–2720 Ma time slice provides evidence for the formation of the Eastern Goldfields crustal block, along with two other new blocks and the three previously existing blocks. This time period is interpreted to be associated with significant crustal reworking and minimal komatiite emplacement. The third time slice (2720–2600 Ma) identified by Mole et al. (2014) coincides with the timing of the most voluminous komatiite emplacement (c. 2.7 Ga) and is characterised by the presence of only two major crustal blocks—the Eastern Goldfields and the West Yilgarn. In both periods of komatiite generation (2.9 Ga and 2.7 Ga), the volcanism is constrained to radiogenic (juvenile) crustal blocks and margins.

Fig. 5
figure 5

εHf mapping of the Yilgarn Craton. Reproduced with permission from Mole et al. (2014); Copyright 2014 National Academy of Science. The subset of zircons 3050–2820 Ma in age show variation across the craton, reflecting the configuration of different crustal blocks at this time. (A) Relationship between komatiite locations and radiogenic (juvenile) εHf. (B) Interpreted configuration of the crustal blocks at this time; a juvenile east-west trending block separates two blocks of reworked crust. The probability density plots for each region are shown as insets.

Full size image

Mole et al. (2014) mapped the changing lithospheric configuration through time, and produced geodynamic models for the generation of the Ni camps of Forrestania, Lake Johnston and Kambalda (Fig. 6). These models, along with complementary evidence from geochronology and geochemistry of magmatic rocks, infer the presence of thin lithosphere based on the radiogenic isotopic signatures, and propose that these regions of thinner lithosphere facilitated ascent of mantle-derived komatiitic magma metals into the upper crust.

Fig. 6
figure 6

Interpreted cross-section at 2.9 Ga in the southern Youanmi Terrane. Reproduced with permission from Mole et al. (2014); Copyright 2014 National Academy of Science. (A) Measured εHf values plotted for analyses in the 3050–2820 Ma age range, as per Figure 5 and location of cross section A-A’. (B) Isotopic cross-section between the Hyden Block and the Lake Johnston Block. The average Hf-isotope signature is significantly different across the two blocks and the interface is interpreted as a craton margin between the two blocks. (C) Interpreted lithospheric architecture at c. 2.9 Ga as interpreted from the isotopic results. Komatiite eruption is facilitated by plume-related extension at the interface of the two crustal blocks, and metals are preferentially deposited in the crust in the Lake Johnson Block where the lithosphere is thinner.

Full size image

7.1 Cautions and Considerations

Mole et al. (2014) combined U-Pb and Lu-Hf isotopes to map ancient craton boundaries through time, and track the progressive development of Archean crust. The earliest cratonic configurations no longer exist, but the isotopic signatures and mineralisation associated with these structures remain intact. The power of isotopes to look back through time is amplified by the zircon LA-ICP-MS method, as used in this study. A range of different zircon growth periods can be targeted and used to get a time-series of events—from a single sample. Mole et al. (2014) remarked upon the abundance of inherited zircons available in the analysed samples, and this allowed isotopic extraction of not only the magmatic population (the magmatic data considered in isolation is similar to whole-rock Sm-Nd data in the region), but also the isotopic signatures of earlier magmatic events as preserved in the inherited zircon population. This approach was especially valuable here, because direct exposure of granites older than c. 2800 Ma are rare in this region (Mole et al. 2014). There is an implicit assumption that the inherited material is from a relatively local source to the emplaced igneous body, and not transported laterally before incorporation into the magma.

Compromises must be made to reduce the complexity of multi-spot analytical data to enable meaningful spatial portrayal. The choice of this simplification (e.g., means vs medians, εHf vs T2DM) can have strong impacts on subsequent interpretations of the data. Mole et al. (2014) used median values from each population, but have provided a range of other portrayals in their supplementary material. These plots demonstrate that the overall trends remain the same for a range of data reduction techniques, but it is an important reminder that decisions on data portrayal have the potential to impact isotopic interpretations.

8 Metamorphism and Gold Mineralisation in the Tropicana Zone

Kirkland et al. (2015) examined potential models for gold deposit formation in the regolith-covered Archean Tropicana Zone, in the Albany-Fraser Orogen on the eastern margin of the Yilgarn Craton, Australia. They applied zircon LA-ICP-MS Lu-Hf analyses, zircon SIMS U-Pb and pyrite TIMS Re-Os geochronology, and whole-rock geochemistry to contextualise the gold mineralisation that occurs within the Tropicana Zone, and to draw comparisons with the adjacent, well-endowed Eastern Goldfields province.

As a component of the Albany-Fraser Orogen, the high-grade Tropicana Zone preserves a complex Proterozoic thermal history. The rocks investigated by Kirkland et al. (2015) are predominantly granulite-facies gneissic rocks, with sanukitoids as the protolith. Kirkland et al. (2015) used micro-analytical U-Pb and Lu-Hf analytical techniques to extract information from metamorphic, magmatic and inherited regions in zircons from these rocks to characterise fractionation events, and also to compare with other lithospheric blocks.

Kirkland et al. (2015) found that there are strong similarities in Hf-isotope character between the Tropicana Zone and the Eastern Goldfields Terrane. The authors interpret the Tropicana Zone was originally deep Archean crust that was structurally emplaced in its current configuration at or before a 1780–1760 Ma thermal overprinting event.

Timing of gold mineralisation in the Tropicana Zone (c. 2520 Ma and c 2100 Ma; Doyle et al. 2015) is much younger than that in the Eastern Goldfields (2660–2630 Ma; Vielreicher et al. 2015). Kirkland et al. (2015) pointed out that there is a known association between Archean sanukitoids and gold mineralisation, and argued that the previous configuration of the Tropicana Zone enabled the remobilisation of metalliferous (Au-bearing) fluids derived from sanukitoids at depth, and subsequent concentration of those fluids into fracture systems.

8.1 Cautions and Considerations

This study demonstrated the power of microbeam isotopic analysis on complex rocks; an advantage that zircon Lu-Hf has over more traditional whole-rock Sm-Nd analysis. By targeting zones within zircon grains, it is possible to disentangle the influence of multiple components of a rock and better understand the processes that generated the rocks and associated mineralisation. As well as magmatic zircons, metamorphic rims and inherited cores were also analysed, which enabled tracking of source fluids and magmas through different geological events, even high-grade metamorphism. This wealth of information is particularly valuable in regions where samples are sparse or difficult to collect, such as regolith-covered regions away from outcropping basement rocks.

This study highlighted the importance of CL imaging to delineate different growth zones in complex zircon grains. Without appropriate characterisation of the internal structure of the zircon grains, it would be impossible to distinguish the different growth phases, and so the isotopic data would be a geologically meaningless mixture of multiple growth phases.

The metamorphic rims in this study have a Lu-Hf signature that is more akin to recycled crust than juvenile crust—which is interpreted to mean that no new mantle-derived material was introduced to the system during the metamorphism that generated the metamorphic zircon rims. However, the authors noted that the Lu-Hf isotopic values in zircon rims appears to be derived more strongly from resorbed inherited zircon cores than primary magmatic zircon. This leaves open the possibility of a mantle contribution that was not preserved in the metamorphic zircon rims.

9 Kimberlite Pathfinding for Diamond Exploration

Schärer et al. (1997) and Batumike et al. (2009) applied ID-TIMS U-Pb, Hf and trace elemental analysis on mineral separates from modern drainage systems to characterise crustal and magmatic evolution in the central Congo-Kasai Craton of central Africa, and to investigate sources of alluvial diamonds in the region. It had previously been assumed that alluvial diamonds were derived solely from the Angolan kimberlite field which intruded Cretaceous sandstone, imposing a maximum age of c. 120 Ma on these occurrences (Batumike et al. 2009). However, some alluvial diamonds in the region contain natural irradiation features characteristic of ancient diamonds (Shmakov 2008), and at least some kimberlites have interacted with older crust, as evidenced by a diamond-hosted zircon inclusion that yielded a U-Pb age of 628 ± 12 Ma (Kinny and Meyer 1994).

To better understand the age, origin and mantle source characteristics of kimberlitic diamonds, Schärer et al. (1997) performed isotope-dilution U-Pb dating and Lu-Hf analysis on zircon and baddeleyite megacrysts with high-pressure formation characteristics (indicative of crystallisation at great depth) from the Mbuji-Mayi kimberlite in the Democratic Republic of Congo. The researchers derived ages of both zircon and baddeleyite of c. 70 Ma. The εHf values of both minerals were strongly positive: c. + 8 for zircon and + 5 to + 10 for baddeleyite, in line with a moderately to strongly depleted mantle source.

Batumike et al. (2009) characterised zircons from the sediments from the Luebo region based on the trace-element classification system of Belousova et al. (2002). This sediment-sampling approach allowed them to gather isotopic and trace-element information from any rock units that eroded heavy minerals into the catchment. Zircons were filtered by εHf and trace element composition to constrain the protolith composition, potentially identifying kimberlite occurrences that had not yet been discovered (Fig. 7). The Hf-isotopes from samples which were classified as kimberlite-derived had relatively positive εHf, consistent with mantle-derived material interacting with late Archean lithosphere during magma ascent. The zircon U-Pb ages comprise three groups: late Archean, Neoproterozoic and Cretaceous. The authors proposed three distinct episodes of diamondiferous kimberlite magmatism in the region despite previous assumptions that diamonds in the region were all sourced in a single time period. This increases the range of geology which may be host to kimberlitic intrusions from post-120 Ma into the Archean.

Fig. 7
figure 7

176Hf/177Hf data and interpreted rock-types from trace element data. Reproduced with permission from Batumike et al. (2009); Copyright 2009 Elsevier. Three kimberlite-related age groups (filled diamonds) were interpreted from sediment-hosted zircons in central Africa, as identified by zircon trace-element data. Associated εHf data shows the radiogenic (juvenile) nature of these zircons relative to the bulk sedimentary zircon load (all other symbols), providing evidence for three distinct periods of kimberlite emplacement rather than a single event as previously thought.

Full size image

9.1 Cautions and Considerations

Hf-isotope data in combination with geochronology and trace-element geochemistry (all of which can be performed on the same zircons) can be used to understand kimberlite genesis and improve pathfinding in diamond exploration. Drainages sample heavy minerals from across a catchment and so provide an efficient method for capturing information on a range of nearby lithologies from a single sediment. This approach can provide first-order information in under-explored regions, including from lithologies that may not be accessible to surface sampling due to vegetative or alluvial cover. In this application, where contextual information on the source protolith is limited, Hf isotope data is most useful when accompanied by complementary geochronological and geochemical datasets.

10 Discussion, Conclusions, Future Developments

The Lu-Hf isotope system is now used in a well-developed and robust metholology for characterising magma and fluid sources, particularly crust-mantle interactions. Many studies using the Lu-Hf isotopic system investigate large-scale lithospheric processes, but there is plenty of opportunity within this framework for more focused investigations to aid mineral exploration.

The four case studies highlight the utility of Lu-Hf isotope studies to increase the understanding of lithospheric processes that control how and where mineral deposits form. This enables greater understanding of which crustal blocks and boundaries may have this highest potential for particular ore deposits (Hou et al. 2015). The Lu-Hf isotopic system preserves syn-mineralization information even if crustal configurations have been subsequently re-arranged, because zircons, even inherited ones, preserve Hf signatures from the time of their formation (Batumike et al. 2009; Kirkland et al. 2015; Mole et al. 2014; Schärer et al. 1997). The case studies demonstrate the wide applicability of this isotopic system across a range of commodities, including Cu-Mo-Fe-Pb-Zn (Hou et al. 2015), Ni (Mole et al. 2014), Au (Kirkland et al. 2015) and diamonds (Batumike et al. 2009; Schärer et al. 1997).

The Lu-Hf isotope system is best used as part of a toolkit that includes other geochemical and isotopic systems such as U-Pb, Sm-Nd, O-isotopes, and trace elements. The information from the Lu-Hf isotope system complements that from geophysical, structural and petrological investigations and can improve understanding of how geological systems and settings have changed through time.

Instrumentation, data quality criteria, interpretations and ideas continue to evolve rapidly in this dynamic field of research. Instrumental improvements are enabling a growing richness of data and improving the precision and spatial resolution of data, as well as enabling linked acquisition of additional elemental and isotopic data. With the increased ease and speed of data acquisition comes a need for improved data reduction and statistical treatment, and data storage systems to deal with large datasets easily.

Similarities between the Lu-Hf and Sm-Nd systems may see integrated datasets in the future, which leverage large existing Sm-Nd datasets and combine them with rapidly growing Lu-Hf data holdings. This will allow more comprehensive coverage and make it easier to identify regions of interest—be they isotopically distinct regions or isotopic boundaries.

Although there are large overlaps between the Sm-Nd and Lu-Hf systems, there are applications in which the Lu-Hf system provides new information. For example, it is possible to use the Lu-Hf isotopic system to date certain minerals that yield only poor ages with Sm-Nd or none at all (e.g. garnet, phosphate and carbonate minerals—see above). The big advantage of the Lu-Hf isotopic system is that Hf is hosted by zircon, which is both dateable via U-Pb, and sufficiently durable that it can be preserved during petrogenetic processes.