Using a titanium-in-quartz geothermometer for crystallization temperature estimation of the Palaeoproterozoic Suursaari quartz porphyry

The Suursaari volcanic sequence represents volcanic activity related to Wiborg Batholith rapakivi intrusions in the southern part of the Fennoscandian Shield. The estimated pressure conditions for batholith granitic rocks are 15 kbar and crystallization temperatures range from 670 to 890 °C. To describe the temperature regime of the Suursaari volcanic system, a rock sample was taken from the Mäkiinpäällys Mountain outcrop and analysed with laser ablation inductively coupled plasma mass spectrometry. Sample spots were selected from quartz phenocrysts and groundmass. Quartz crystallization temperatures were calculated by the Ti-in-quartz method that takes into account rutile equilibrium and Ti activity in each phase. The calculated crystallization temperatures of the Suursaari quartz porphyry are in the range of 647738 °C. The results show that the Suursaari quartz porphyry contains two generations of quartz which can be distinguished on the basis of crystallization temperatures: phenocrysts crystallized at higher and groundmass quartz at lower temperature.


INTRODUCTION
Suursaari Island lies in the middle of the Gulf of Finland, approximately 40 km from the coast of Finland and 55 km from the Estonian mainland (Fig. 1).It provides a unique example for studying crystalline basement rocks of the southern margin of the Fennoscandian Shield.Moreover, volcanic products of the Proterozoic rapakivi-type association are well exposed on the island.In fact, Suursaari has yielded the best evidence for volcanic activity associated with the Wiborg batholith (Rämo et al. 2010).
The 1650-1620 Ma rapakivi granites of the Wiborg batholith and its satellites are relatively high-level, epizonal plutons that were emplaced into the ~ 1.9 Ga Svecofennian crust in an extensional tectonic setting.The current erosional level of the Wiborg batholith corresponds to a palaeodepth of 1-5 kbar (Rämö et al. 2010).Batholith intrusion was associated with a relatively thinned crust, swarms of basaltic and silicic dikes, as well as rare volcanic rocks as those on Suursaari Island.The preserved volcanic sequence on the island is about 200 m thick.
High-precision isotope dilution-thermal ionization mass spectrometry U-Pb zircon data on the pyroclastic rhyolitic units from Suursaari imply upper intercept crystallization ages of 1633 ± 2 Ma (Rämö et al. 2010).Epsilon-Nd values of lavas and pyroclastic rocks from Suursaari are slightly negative: around -0.5 for basalts and about -2 for silicic rocks (Rämö et al. 2010).The zircon ages of volcanic products of Suursaari, complying with those of the early and main intrusive phases of the Wiborg rapakivi complex, show that concomitant, quite extensive bimodal volcanism was associated with the emplacement of the Wiborg batholith.It is interesting to note that Estonian rapakivi-type plutons (Soesoo & Niin 1992;Soesoo 1993) fall within the age range similar to that of the Wiborg (U-Pb zircon age 1.650-1.625Ga; Vaasjoki et al. 1991) varieties: granodiorite of the Märjamaa pluton has yielded a U-Pb zircon age of 1.65-1.63Ga, Neeme granitoids have given ages of 1.634 and 1.648 Ga and Ereda granitoids 1.642-1.627Ga (Soesoo et al. 2004;Kirs et al. 2009;Soesoo & Hade 2010).
The age relationships between different units of the Wiborg batholith and associated sequences are well established.However, knowledge on the primary magma emplacement conditions, more specifically temperature and pressure conditions, is still insufficient.Different methods are available to estimate P-T conditions, several of which use mineral assemblages or mineral pairs for P-T calculations.In order to estimate P-T conditions across the entire evolution of a complex magmatic system, single-grain mineral analysis may be helpful, but the number of investigations using single-grain mineral methods is limited.For example, the method estimating the crystallization temperature and pressure based on clinopyroxene composition (e.g.Putirka et al. 1996;Soesoo 1997) cannot be applied to granitic compositions because of the absence of clinopyroxene in the rock.
Recently, a titanium-in-quartz geothermometer (TitaniQ) was developed (Wark & Watson 2006).Due to its novelty and a wide range of applications it has been used in generation temperature research for many rock types.Important applications of the thermometer are the studies on plutonic rocks by Johnson et al. (2009Johnson et al. ( , 2011) ) and Wiebe et al. (2007), and on volcanic rocks by Wark et al. (2007), Bachmann (2010), Reid et al. (2011), Smith et al. (2010) and Wilcock et al. (2009).Müller et al. (2008) provided an overview of the thermometry of rapakivi granites.Girard & Stix (2010) and Shane et al. (2008) used the thermometer to understand better the magma chamber processes.
The aim of this paper is to use the titanium-in-quartz thermometer on Suursaari volcanic rocks for refinement of temperature conditions during their formation.Possible applications of the method in the studies of rapakivitype rocks are assessed.

RAPAKIVI AND RELATED ROCKS -CRYSTALLIZATION TEMPERATURES IN THE MAGMATIC SYSTEM
Temperature and pressure conditions of rapakivi formation have been evaluated on many complexes of the Fennoscandian Shield.The estimated temperatures usually mark the range of 600-900 °C, while pressure conditions for the rock formation range between 1 and 6 kbar (Table 1; Eklund & Shebanov 1999).
An overview of the formation parameters of rapakivi intrusions in the Fennoscandian Shield has been given by Eklund & Shebanov (1999).Temperature conditions between different complexes vary, but usually do not exceed 800 °C, whereas temperatures below 780 °C seem to dominate.There are some exceptions towards higher, up to 850 °C temperatures, especially in granite-related monzonites and granite varieties that are affected by simultaneous mafic magmatism.
Rapakivi granites contain at least two generations of K-feldspar, plagioclase and quartz (Nekvasil 1991;Eklund & Shebanov 1999), although the majority of the reported P-T determinations reflect averaged intensive parameters estimated from all generations of major phases in the magmatic system.To give more complete information about this system, it is important to apply thermobarometry directly to different rock/mineral generations (incl.megacrysts, inclusions in megacrysts and groundmass).Eklund & Shebanov (1999) have shown that P-T conditions may differ between megacrysts and groundmass.The core zones in feldspar ovoids showed a pressure of about 5-6 kbar and temperature of 680-750 °C, while the cores of quartz megacrysts showed 4.5-6.5 kbar and 720-780 °C, respectively.The matrix of the same rock type gave the values 1-2.5 kbar and 650-750 °C (Eklund & Shebanov 1999).
For example, the two-stage growth of zircon during crystallization of the rapakivi parental magma is consistent with the evidence for two distinct mineral assemblages in amphibole-biotite rapakivi granites of the Salmi complex.These were formed in temperature intervals of 740-780 °C and 680-720 °C (Shebanov et al. 1996).
A detailed overview of Wiborg batholith rapakivi crystallization conditions has been given by Elliott (2001).Hornblende-plagioclase thermometry and aluminiumin-hornblende barometry within wiborgite record crystallization temperatures between 670 and 800 °C, at pressures of 2.5-5.4 kbar (Elliott 2001).Amphibole data from mafic magmatic enclaves and hybrid rocks record a wide range of temperatures and pressures between 710 and 890 °C, at pressures of 0.7-3.1 kbar (Elliott 2001).Väisänen et al. (2000) described P-T conditions of post-collisional magmatism in SW Finland.In this region the Palaeoproterozoic mineral assemblages attained equilibrium at average P-T values of 4.1 kbar and 680 °C (Väisänen et al. 2000).The post-collisional intrusions intruded at a pressure of at least 4.1 kbar, corresponding to a minimum depth of 14-15 km.
Summarizing the results of mineral crystallization regime investigations on the Fennoscandian rapakivitype rocks, it is well evident that the rapakivis have formed in polybaric and -thermal conditions (Table 1).In order to gain a better understanding of various stages of the evolutionary sequence, a detailed method is necessary enabling mineral grain-based estimates.

TitaniQ: TITANIUM-IN-QUARTZ GEOTHERMOMETER
Titanium (Ti) is one of many trace elements that substitutes silica (Si) in quartz (Larsen et al. 2000;Flem et al. 2002;Müller et al. 2003aMüller et al. , 2003b;;Götze et al. 2004Götze et al. , 2005)).A titanium-in-quartz (TitaniQ) geothermometer is based on the idea that Ti concentration in quartz is related to the mineral formation temperature.Higher crystallization temperature means also that Ti concentration in rock is higher (Wark & Watson 2006).In igneous rocks Ti can substitute for Si without having to be charge balanced by coupled substitution of another element, because of the tetravalent nature of both the Tiand Si-cations (Götze et al. 2001).The activity of Ti in many systems is fixed by the presence of a nearly pure TiO 2 phase (typically rutile, Wark & Watson 2006).Consequently, the chemical potential of Ti, and hence the extent of Ti substitution for Si in quartz, should vary systematically with temperature.The titanium concentration in quartz typically ranges between 1 and 100 ppm, in the quartz of high-temperature rocks the content can be even higher.
The TitaniQ geothermometer is based on the same thermodynamic principles as trace element thermo-meters Ti-in-zircon and Zr-in-rutile (Wark & Watson 2006).At equilibrium conditions the quartz-rutile exchange reaction can be written as rutile qtz 2 2 TiO TiO .

=
(1) Wark & Watson (2006) showed experimentally that the Ti concentration in quartz increases exponentially with the reciprocal temperature, and quartz crystallization temperatures, if we do not account for pressure conditions, are calculated by the equation where qtz Ti X is the Ti concentration in quartz and 2 TiO α is the activity of TiO 2 of the system.In the presence of rutile the 2 TiO α value is ~ 1 (Wark & Watson 2006).In rocks where Ti concentration is low and mineral rutile has not formed, the 2 TiO α value is < 1.In this case the Ti activity must be independently estimated and activity fluctuation by ± 0.2 gives an error of ± 10 °C in the final calculation (Kohn & Northrup 2009;Spear & Wark 2009).
The advantage of this method is the variable use range, because almost every rock contains quartz that is stable in different pressure and temperature conditions.The TitaniQ geothermometer is especially useful for analysing rocks that crystallized at temperatures above 500 °C.

MATERIAL AND METHODS
Quartz porphyry samples were collected in the NW part of Suursaari Island (during the field work in 1987) at the Mäkiinpäällys Mountain outcrop.The samples belong to the Museum of Geology of the University of Tartu.The main minerals in the sample are quartz, perthitic orthoclase, plagioclase with sericite, epidote and chlorite.Accessory minerals are represented by apatite, zoisite, sphene, calcite, fluorite and magnetite.
Before analysing the samples with laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS), they were studied with a petrographic microscope and scanning electron microscope (SEM) to determine the possible presence of titanium-containing inclusions and avoid them while analysing quartz.The in-house SEM instrument Zeiss EVO MA15, equipped with an Oxford INCA energy-dispersive spectrometer, was used.
Quartz phenocrysts in the rock are relatively large and surrounded by grey-brown sub-microscopic groundmass (Fig. 2).The petrographic microscope revealed in places embayments of the groundmass inside quartz crystals (Fig. 3).To clarify for further examination whether the quartz crystals contain impurities, the samples were studied by SEM.It is extremely important to make sure that the quartz crystals selected for LA-ICP-MS analysis do not contain impurities or inclusions, especially of rutile.The SEM analyses confirm that quartz is pure and has no rutile inclusions (Fig. 4).In some cases, plagioclase, feldspar and accessory magnetite were determined in quartz grains.
Titanium in quartz was measured by using inductively coupled plasma mass spectrometer (a quadrupole  X-Series 2 ICP-MS, Thermo Scientific) equipped with the laser ablation system (UP213nm, NewWave).
A total of 49 spot analyses from porphyritic quartz grains and 40 spot analyses from groundmass in an uncoated thin-section were subjected to LA-ICP-MS analysis.To ablate quartz in both phenocrysts and groundmass, a 40 µm laser beam during 60 s was used.Phenocrysts were ablated on a straight line and a spot was used for groundmass.Before analysing the sample surface was cleaned by ablating it during 1 s.The isotope Si 29 was used as an internal standard.External calibration was done by using three multi-element silicate glass reference materials produced by the National Institute of Standards and Technology (NIST): SRM 610 (consists of 437 ppm Ti), SRM 612 (50.1 ppm Ti) and SRM 614 (3.1 ppm Ti).

RESULTS AND DISCUSSION
The Ti content of quartz from phenocrysts and groundmass of the Suursaari quartz porphyries together with calculated rock crystallization temperatures are presented in Table 2 and Figs 5 and 6.The average Ti concentration in phenocrysts is approximately 204 ppm and slightly lower in the groundmass, averaging at 187 ppm.
The Ti concentration in phenocrysts ranges between 160.9 and 250.6 ppm (δ ± 14.5-26.9).The crystallization temperature of the phenocrysts of the Suursaari quartz porphyry calculated according to Eq. ( 2) and assuming Ti activity of 0.8 varies between 669 and 726 °C (± 24 °C) (Table 2).Even though the samples do not contain the rutile phase, the Ti activity of 0.8 is suggested on the basis of the minor presence of apatite and titanomagnetite (as Ti-bearing phases) in the studied samples.The distribution of the measured temperature values of phenocrysts is presented in Fig. 5.The main clustering of the measured values in phenocryst grains falls between 680 and 720 °C, with two temperature peaks around 680 and 700 °C.There may exist a third cluster, pointing towards higher (up to 740 °C) crystallization temperature values.
The obtained data show that quartz in groundmass has generally crystallized at an approximately 20 °C lower temperature than the phenocryst variety.This, indeed, is a logical assumption as groundmass crystallizes later.As seen from Fig. 6, the main range of groundmass quartz crystallization temperature stays within 670-700 °C.The quartz in groundmass shows gradual temperature distribution with the main peak at 680-690 °C.However, the higher-temperature values, up to 750 °C, also exist, which may be derived from early magmatic crystallization processes.This tendency can also be seen in phenocrysts.Some groundmass quartz grains show even higher crystallization temperatures compared to phenocryst quartz.More analyses are needed to understand these relationships.
As shown in Fig. 7, quartz in rapakivi granitic rocks crystallizes mostly at temperatures 700-850 °C, with the tendency towards lower temperature values in lower pressure conditions.The analysed quartz crystallization temperatures from the Suursaari quartz porphyry are close Fig. 7. Suursaari rapakivi intrusion P-T conditions in comparison with previously published data on other rapakivi intrusions in the southern slope of the Fennoscandian Shield (for legend references see Table 1).

T, °C
to the lower values of wiborgite crystallization temperatures from the Wiborg and Salmi plutons (Shebanov et al. 1996;Elliott 2001).According to Elliott (2001), the Wiborg batholith granites emplaced at relatively shallow levels in the crust, with P-T values of 0.7-5.4kbar and 670-890 °C (Table 1).A higher P-T crystallization regime (4-8 kbar and 740-840 °C) is characteristic of monzonitic rocks of the Åland batholith (Eklund & Shebanov 2005).However, it should be mentioned that volcanic Suursaari quartz porphyries have undergone intensive hydrothermal alteration (albitization, epidotization, etc.) processes, which may have had some effect on the Ti content in quartz.
In general, quartz varieties in granites, monzonites and diorites that have crystallized at higher temperatures contain more trace elements in its crystal structure than the quartz formed at lower temperature (Larsen et al. 2000).The most widespread trace elements in quartz aside from Ti are Mg, Ca and Cr in natural less-fractioned pegmatites, and Fe, Li and B in more-fractioned pegmatites.Larsen et al. (2004) found Al, P, Li, Ti, Ge and Na in the order of > 1 ppm in pegmatitic quartz, whereas K, Fe, Be, B, Ba, Sr and their trace elements were below the detection limit of LA-ICP-MS.Therefore, it is important to study the Suursaari quartz porphyry for all possible trace elements in order to assess the calculated geothermometry data.
This pilot study shows that the TitaniQ geothermometer may give a valuable input into understanding the magmatic history of complex magmatic systems, such as rapakivi formations.A single-crystalbased method is able to distinguish different crystal generations even in a small sample.However, the knowledge about pressure conditions of the magmatic system will greatly expand the understanding of the magmatic history.
Recently, Thomas et al. (2010) studied how pressure influences Ti solubility in quartz.They used the same material (distilled water, TiO 2 and quartz powder (< 22 µm) or SiO 2 glass) to synthesize quartz crystals as Wark & Watson (2006).The results of the experiment show that if in the equilibrium phase defined by Eq. ( 1) the substitution of Ti 4+ for Si 4+ is implicit, the site on which Ti resides in the quartz structure is not specified.It is conceivable that Ti may reside on tetrahedral sites or may dissolve into interstitial sites and thermodynamic variables are unique for each solubility mechanism (Thomas et al. 2010).In case the pressure can be constrained to within ± 1 kbar, the temperature can be constrained to approximately ± 20 °C (Thomas et al. 2010), because the Ti activity increases when pressure decreases.Thus, the method may have important applications in understanding the complex magmatic systems, which have evolved through different pressure and temperature ranges.

CONCLUSIONS
Rapakivi granites are crystallized in polybaric magmatic systems, where pressure and temperature conditions vary between different rapakivi complexes.In a large scale the rock formation temperatures vary in a range of 600-900 °C and pressure conditions are around 1-6 kbar, in some cases even higher.The Suursaari quartz porphyry is a good example of a volcanic rapakivi rock, containing idiomorphic quartz phenocrysts.Analysis of quartz porphyries with LA-ICP-MS using the TitaniQ geothermometer allows measuring each quartz phenocryst and quartz in groundmass separately, thus having an advantage of providing information on various quartz generations.The crystallization temperatures of the Suursaari quartz porphyry phenocrysts range from 669 to 726 °C, and those of the groundmass vary between 647 and 738 °C.The analysed quartz crystallization temperatures are close to lower values of wiborgite crystallization temperatures from the Wiborg (670-890 °C) and Salmi (680-780 °C) rapakivi batholiths.To analyse pressure variations in rapakivi polybaric magmatic systems, more quartz grain measurements from different rapakivi granitic types have to be made.

Table 2 .
Titanium content (measured with laser-ablation ICP-MS) and calculated crystallization temperatures of quartz phenocrysts (49 analyses) and groundmass (40 analyses) in the Suursaari quartz porphyry