Zircon - Structure, crystal chemistry, and provenance

Zircon - Crystallography, Crystal Chemistry, Morphology and Provenance Potential

Zircon is one of the most common Zr minerals in the crust

It forms in some mafic to silicic igneous and metamorphic rocks.
It is one of the three most refractory heavy mineral - together with rutile and tourmaline. Basis for the ZTR index.
It is the basis for robust geochronology of the source rock crystallization age (high isotopic resetting T of 800 C)
It provides an excellent basis for examining the age range of the zircon-bearing sources and, thus, the tectonics

Zircon formula and structure

Zircon is zirconium orthosilicate, ZrSiO₄, and has a stoichiometric composition of 67.2 wt % ZrO₂ and 32.8 wt % SiO₂. The structure of zircon contains two cation sites; the 4-coordinated Si-site and the distorted 8-coordinated Zr-site. Both Si and Zr are tetravalent and have ionic radii of 0.84 Å and 0.26 Å, respectively.

Substituent cations for Zr⁴⁺ include Hf⁴⁺ (most abundant), Y³⁺, REE³⁺, P⁵⁺, U⁴⁺, Th⁴⁺and several others.

Only hafnon (HfSiO₄) and zircon have complete solid-solution. The extent of this solid solution in natural zircon is typically restricted although variation of the Zr/Hf ratio in zircon away from the chondritic value of ~37 is common. A compilation of published Hf values in zircon up to 1969 by Ahrens and Erlank (1969) showed a range in HfO₂ values from ~0.7-8.3 wt % with a mean of 2.0 wt %. Large variation above and below this 1969 mean value has now been reported. Given that the most Hf-enriched zircon occur in evolved rock-types, it appears that the Hf abundance of zircon increases with magmatic differentiation.

Hafnium (wt %) variation in zircon populations from the Sweetwater Wash pluton, southeastern California, U.S.A. Each data point is an analysis of a single crystal, or the mean of multiple analyses on a single crystal. Columns of data points represent the spread of Hf abundance in a single zircon population. Data from Wark and Miller (1993).

Compilation of REE patterns for meteorite, lunar and terrestrial zircon.

(a) meteorite zircon from the Vaca Muerta mesosiderite and the H5 chondrite Simmern (Ireland and Wlotzka 1992)

(b) lunar zircon from the Apollo 14 landing site (Hinton and Meyer 1991, Snyder et al. 1993)

(d) zircon from Blind Gabbro, Australia, and Mawson igneous charnockite, Antarctica (Hoskin and Ireland 2000);

(e) ophiolite and plagiogranite zircon (Hoskin and Ireland 2000);

(f) zircon from two zones, diorite and aplite, of the Boggy Plain Zoned Pluton, BPZP (Hoskin et al. 2000)

Minor and trace elements with CART statistics (Belousova et al. 2002)

Trace element composition of zircons from different rock types

The fields of zircon compositions used as discriminants for different rock types.

‘Granitoids’ include:

1 aplites and leucogranites;

2 granites;

3 granodiorites and tonalites

CART tree for the recognition of zircons from different rock types.

Terminal nodes indicate predicted rock type, estimated probability and number of observations. Error rate matrix for this classification is in Table 7

Average Hf and Y concentrations in zircon relative to the fields of zircon composition defined by Shnukov et al. (1997), where

a data from this study,

b data from previous work.

Morphology of zircon

Zircon has a highly variable external morphology. Often crystals are faceted with combinations of prism ({100} and {110}) and pyramid forms ({211}, {101} and {301}).

(a) top row: cathodoluminescence image (left) of zircon from an adamellite revealing changes in the external morphology (represented schematically, right) of the crystal during growth. After Hoskin (2000).

(b) bottom row: cathodoluminescence image of zircon from an intermediate alkali volcanic rock, northeast Turkey. Both crystals exhibit well developed oscillatory and sector zoning as a result of heterogeneous traceelement distribution.

External morphology of igneous zircon.

A typologic scheme relating the relative development of crystal forms with temperature and host-rock type was published by Pupin (1980) Although this scheme was initially widely accepted and is still used three subsequent and ongoing observations have seen it rarely used

zircon from a single rock and single age population can have widely varying morphologies
zircon from widely different rock-types can have similar to identical morphologies with no systematic measurable differences despite claims to the contrary
the external morphology of a single crystal can change a number of times during a single growth event as a result of kinetic factors, such as diffusion rates and adsorption, which effect the growth rates of crystal faces and therefore control the morphology of a growing crystal

External morphology of metamorphic zircon.

Zircons in low-grade metamorphic rocks are usually inherited from the protolith and may show signs of resorption or metamorphic overgrowth.
High-grade rocks such as granulites may contain zircon that grew during metamorphism in the presence or absence of an anatectic melt. Typical morphologies are rounded or ovoid shapes which are interpreted to be formed by resorption by a zircon-undersaturated intergranular fluid

Zircon in granulite-grade rocks, Vosges Mountains, eastern France:

(a) rounded-ovoid grain; (b) “soccer-ball” zircon; (c) prismatic zircon from the same granulite as the grain in (a), but occurring on thegrain-boundaries of rock-forming minerals.

Internal textures of metamorphic zircon.

Internal textures are observed by CL and BSE imaging allow for visual distinction between metamorphic growth zones, preserved igneous cores, recrystallized domains and domains showing other types of structural reorganization.

A metamorphic zircon illustrating the commonly observed sequence of growth structures: (1) low-luminescence center; (2) sector zoned domain; (3) oscillatory zoned outer part. The bright-CL area to the right of domain 1 is an area of localized recrystallization around a 40-um mineral inclusion. Figure after Schaltegger et al. (1999).

Basis for U-Pb geochronology using zircon

Long-lived radioactive decay systems provide the only valid means of quantifying geologic time.

The uranium-lead decay system has always played a central role for several reasons.

Minerals that contain very high U concentrations, although rare, are well known and easily obtained.
The half lives of the natural U isotopes ²³⁸U and ²³⁵U are long enough to span all of Earth’s history but short enough that both parent and radiogenic daughter elements could be measured in such minerals, even with the methods of a century ago.

Basic decay system equations for U-Pb. The decay equations of these two decay schemes yields the concordia line - a curve for a theoretically closed system

If the system has been closed to mobility of parent or daughter these two ages should agree, thus furnishing an internal test on the accuracy of the age.

Further, the chemical coupling of the decays allows the age of the radiogenic daughter to be determined solely from its isotopic composition without knowing the parent-daughter ratio, a more difficult and less reliable parameter to measure.

Assuming a simple two-stage history, the upper concordia intersection of a best-fit line through a data array should give the crystallization age of the zircon while the lower intersection should give the age of isotopic disturbance

Multi-grain zircon geochronology

U-Pb concordia diagram showing episodic and continuous diffusion (Tilton 1960) Pb loss models fitted to moderately discordant zircon data from igneous rocks in the Rainy Lake area (Hart and Davis 1969).

The concordia age is calculated using presently accepted U decay constants.

Improvements of the technique

(A) U-Pb data on abraded zircon from igneous rocks in the Rainy Lake area, NW Ontario, Canada by Davis et al (1989). Compare to Hart and Davis (1969) data on the same units in Fig 2.

(B) U-Pb data on single detrital zircon grains from Archean metasandstone in the Rainy Lake area.

CL image of polyphase zircon from the Cowra granite, Australia, showing pits produced by the SHRIMP primary ion beam. Concordia data show that the overgrowth is about 600 Myr younger than the core.

Images of zircon from Archean gneisses in the Winnipeg River belt, NW Ontario, Canada.

(A) BSE image of a grain with a core from a ca. 2700 Ma gneiss. BSE is sensitive to the average atomic number so trace element enriched areas are brighter.

(B) CL image of the grain in A. CL is sensitive to both elemental composition and structural state. Emission is suppressed by radiation damage so areas that are bright in BSE are often dark in CL.

(C) BSE image of a zircon from a 2880 Ma old gneiss with 3060 Ma inheritance. The overgrowth has been cracked by differential expansion of the higher U core. The core is altered (dark regions) and contains bright inclusions that are probably REE phosphate.

(D) BSE image of zircon from the older gneiss with four phases of growth. Lower-U phases are preferentially cracked.

Basis for zircon as a provenance mineral

Detrital zircon studies can be applied to

determine maximum age of stratigraphic successions and to help recognize time gaps in the geologic record
determine provenance characteristics such as age and composition,
test regional paleogeographic reconstructions via provenance analysis
unravel facets of Earth history locked in the mineral chemistry of detrital zircon.

Binned frequency diagrams are common method for displaying age data where modes, ranges and proportions may relate to the timing, duration and relative significance of geologic events.

Frequency histograms for detrital zircons from three quartzites deposited at ~3.0 Ga. Bin width for histograms is 25 Myr.

(A) Grains analyzed using SHRIMP from the Orange Grove Quartzite Formation, Witwatersrand Supergroup, South Africa (data from Barton et al. 1989).

(B) Grains analyzed using SHRIMP from Buhwa quartzite, Zimbabwe (data from Dodson et al. 1988).

(C) Grains analyzed by TIMS from meta-conglomerate, Jack Hills, Western Australia (data from Amelin 1998). Nelson (2001) reported a single grain at ~3.1 Ga and Wilde et al. (2001) reported grains up to 4.404 Ga from different samples from the Jack Hills. These are not shown on this frequency histogram.

Two critical limitations.

the histograms are based only on the age measurement and the inherent errors in the ages are discarded. Thus a measurement with a ±100 Myr standard error could appear in the same bin as one with a ±1 Myr standard error even though the two measurements are strictly not comparable.
the size of the bins themselves is arbitrary with values in the literature including 5 Myr

To circumvent these two limitations to histograms use of probability density distributions has become widespread. These diagrams incorporate the errors in the age data and produce a probability distribution of the entire sample based on Gaussian kernels that vary with each individual age.

Probability distribution diagrams of zircon ages for modern sediments on the eastern coast of Australia. n is the number of grains analyzed.

Modified from Sircombe (1999).

The kilometers-thick Paleoproterozoic Hurwitz Group in the western Churchill Province of northern Canada provides an excellent example where detrital zircon ages, together with bulk rock isotopic analyses have revealed a time gap of approximately 200 Myr across a cryptic internal unconformity within the succession.

Figure 6. Stratigraphic context of the Hurwitz Group, Canada, showing a 200 Myr disconformity within the stratigraphic section, revealed in part by detrital zircon geochronology.

Modified from Aspler et al. (2001).

REE plots for zircons derived from compositionally disparate sources.

The plots show that zircon composition is not dramatically different regardless of composition.

Consequently, REE typically cannot fingerprint provenance. Note particularly the similarity of patterns for aplite, diorite, and high-grade gneiss from Sri Lanka. Such lithologies dominate the major composition of the continental crust and so are likely to represent a large fraction of detrital grains in continental sedimentary deposits.

One of the earliest papers to evaluate provenance by examining U-Pb ages of detrital zircons is that of Gaudette et al. (1981).

They took a sample of Cambrian Potsdam Sandstone from the eastern flank of the Adirondack Dome, New York State, and separated four populations of zircon crystals based on color, crystal habit, and morphology and analyzed, for the first time, single zircon grains and small groups from the same population to extract defined provenance ages.

Flow diagram illustrating the different zircon provenance ages and the possible delivery pathways into the Potsdam Sandstone (Gaudette et al. 1981).

The geochronology results portrayed in this diagram demonstrated for the first time the power of isolating zircon populations in provenance reconstructions.

Detrital zircon geochronology is a powerful tool for provenance analysis, particularly for helping to constrain paleogeography, tectonic reconstructions, and crustal evolution

U-Pb TIMS dating of detrital zircons was applied to assess the provenance of fluvial sandstones of the Shaler Supergroup, an early Neoproterozoic (~1.0-0.75 Ga) intracratonic basin succession preserved on the northwestern (present coordinates) margin of the North American craton (Laurentia) in the Canadian Arctic.

These and correlative sandstones exhibit consistent northwesterly paleocurrents.
Following the ideas of Potter (1978) concerning “big river” systems, Young (1978) suggested that the source for these sediments may have lain in the Grenville province on the opposite side of the continent

Northwestern Canada showing location of inliers containing fluvial quartz-arenites of early Neoproterozoic (1.0-0.75 Ga) age, considered to be remnants of an extensive river system originating from the Grenville orogen, some 3000 km to the southeast. Arrows represent generalized paleocurrents based on measurements of cross-bedding (minimum of 20 readings for each arrow). Modified from Rainbird et al. (1992).

Subsequent detrital zircon geochronology supported this assertion with a significant proportion of the ages matching that of the adjacent Archean Slave province and its marginal Paleoproterozoic orogenic belts.