The structure and composition of granites provide clues to the nature of silicic volcanism, the formation of continents, and the rheological and thermal properties of the Earth's upper crust as far back as the Hadean eon during the nascent stages of the planet’s formation1,2,3,4. The temperature of granite crystallization underpins our thinking about many of these phenomena, but evidence is emerging that this temperature may not be well constrained. The prevailing paradigm holds that granitic mineral assemblages crystallize entirely at or above about 650–700 degrees Celsius5,6,7. The granitoids of the Tuolumne Intrusive Suite in California tell a different story. Here we show that quartz crystals in Tuolumne samples record crystallization temperatures of 474–561 degrees Celsius. Titanium-in-quartz thermobarometry and diffusion modelling of titanium concentrations in quartz indicate that a sizeable proportion of the mineral assemblage of granitic rocks (for example, more than 80 per cent of the quartz) crystallizes about 100–200 degrees Celsius below the accepted solidus. This has widespread implications. Traditional models of magma formation require high-temperature magma bodies, but new data8,9 suggest that volcanic rocks spend most of their existence at low temperatures; because granites are the intrusive complements of volcanic rocks, our downward revision of granite crystallization temperatures supports the observations of cold magma storage. It also affects the link between volcanoes, ore deposits and granites: ore bodies are fed by the release of fluids from granites below them in the crustal column; thus, if granitic fluids are hundreds of degrees cooler than previously thought, this has implications for research on porphyry ore deposits. Geophysical interpretations of the thermal structure of the crust and the temperature of active magmatic systems will also be affected.
Access optionsAccess options
Rent or Buy article
Get time limited or full article access on ReadCube.
All prices are NET prices.
Subscribe to Journal
Get full journal access for 1 year
only $3.90 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
All data for this manuscript are reported in the Extended Data. This research was supported by the Carnegie Institution of Washington’s Postdoctoral Fellowship Program and the NASA Astrobiology Institute (grant NNA09DA80A) to the New York Center for Astrobiology. We thank N. Ackerson, M. J. Ackerson and V. Ackerson for assistance in the field, E. Bullock and J. Armstrong for their technical assistance in analysing quartz, and D. Coleman, J. Reimink, L. Wagner, H. Le Mével, S. Shirey and P. Ulmer for conversations during manuscript preparation.
Extended data figures and tables
Geological map modified from ref. 10.
Extended Data Fig. 2 Secondary fluorescence analyses of Ti concentrations in quartz as a function of distance from an adjacent titanite crystal.
Analyses of the Ti of a quartz crystal from the Glen Aulin tonalite (Kga) show a rimward increase in Ti towards an adjacent titanite crystal. This increase in Ti content is due to a secondary fluorescence of Ti from the nearby titanite crystal. Error bars are 1σ.
Extended Data Fig. 3 Blue-filtered cathodoluminescence images of quartz from the TIS demonstrate both primary crystallization features and features that suggest alteration of the initial Ti content.
a, Quartz from the Glen Aulin tonalite. b, Quartz from the equigranular Half Dome granodiorite has a centre-to-rim decrease in Ti content consistent with lower-temperature crystallization. c, d, In contrast, other quartz crystals from the equigranular Half Dome granodiorite exhibit complex cathodoluminescence features suggesting that the primary Ti content of the quartz has been altered by syn-magmatic deformation. e, f, Quartz from the Cathedral Peak granodiorite displays a series of primary magmatic growth features suggesting thermal cycling and dissolution-recrystallization of quartz. Lines are microprobe traverses. Source Data
a, Calibration by Thomas et al.18; b, calibration by Huang and Audétat20. Although the Huang and Audétat20 calibration could yield near-solidus temperatures for 20–40 p.p.m. Ti in quartz, these high temperatures cannot explain the sharp Ti concentration zones in quartz crystals, which could only be retained if the quartz crystallized at low temperatures (Fig. 3).
Extended Data Fig. 5 Core–rim Ti concentration profiles in quartz compared with a constant-cooling rate diffusion model.
Diffusion model from Fig. 3b. a, Cathodoluminescence (452 nm) map image of a quartz crystal from the Cathedral Peak granodiorite from approximately the core (upper left) to the rim (right and bottom). Coloured lines are profiles seen in b. b, Normalized concentration gradients from profiles across the grain combined with the diffusion model from Fig. 3b indicate that initial crystallization temperatures must have been below the traditional solidus at about 700 °C for almost the entirety of the observed quartz crystal. For simplicity, high concentrations are all plotted on the left, regardless of position within the crystal. At the scale of these profiles, this simplification does not noticeably change the modelled profiles (that is, it does not matter if the high concentration is coreward or rimward). A second example from the Cathedral Peak granodiorite also contains core (top) to rim (bottom) cathodoluminescence (c) and corresponding concentration gradient profiles that indicate low temperature crystallization for the entire crystal (d).
Extended Data Fig. 6 The influence of cooling rate on the results of constant-cooling rate diffusion models.
a–c, Diffusion modelling at 10 °C per million years (10 °C Myr−1; a), 36.36 °C Myr−1 (b) and 100 °C Myr−1 (c). Faster cooling rates will lead to less diffusion. However, even at rapid cooling rates (for example, 100 °C Myr−1), the modelled profiles do not match the observed Ti concentrations.
Extended Data Fig. 7 Estimating the volume of low-temperature quartz crystallization when diffusion modelling is intractable.
a, Cathodoluminescence (452 nm) image of a centre-cut quartz crystal from the Half Dome granodiorite. This crystal has internal Ti concentrations as high as 101 p.p.m. Concentric circles are contours showing the percentage of total volume crystallized at that radius, assuming spherical quartz growth. Using 60 p.p.m. Ti as a rough approximation for high-T quartz crystallization (denoted with a bold dot-dashed polygon), this volume approximation demonstrates that only around 10% of the quartz could have crystallized at solidus or supersolidus temperatures. b, Radius of crystal (approximately 590 µm) versus the total volume of quartz crystallized at a given radius for a spherical quartz crystal with r = 590 µm.
In addition to quartz, feldspar compositions and oxygen isotope fractionation between quartz and zircon results in subsolidus temperatures. Two-feldspar temperatures are calculated using three different calibrations (equations (27a), (27b) and a global regression from ref. 49). Zircon–quartz oxygen isotope fractionation temperatures are calculated using averaged oxygen isotope compositions from quartz and zircon from individual units in the TIS19 and an experimentally calibrated zircon–quartz fractionation relationship50.