Ancient Alpine rocks secretly melted twice, zircon crystals reveal

Deep in the Swiss Alps, some of the oldest rocks in the Aar Massif have been hiding a fiery secret for more than 300 million years. New research published in the Swiss Journal of Geosciences shows that Ordovician migmatites and granites—rocks that first formed when molten crust cooled around 455 million years ago—were melted a second time during the Variscan orogeny, producing a previously unrecognized generation of anatectic granites. The discovery hinges on the remarkable memory of zircon, a mineral so durable that it can carry the chemical fingerprints of multiple melting events inside a single crystal.

The study focused on two gneiss units within the Aar Massif, one of the External Crystalline Massifs that expose fragments of the European continental basement in the Central Alps. In the northern part of the massif lies the Erstfeld Gneiss Complex, a suite of stromatic metatexites and diatexites with meta-igneous textures. In the south-eastern part, near the Oberalp Pass and the Val Strem, sits the Strem Granite unit, which contains metatexites and diatexites that grade into true granites. To the naked eye, and even under the microscope, the rocks in both units look strikingly similar: quartzo-feldspathic gneisses with layered light and dark bands, formed when the crust partially melted. Yet the two units tell fundamentally different stories about when they last saw molten conditions.

The key to unraveling this history was in-situ uranium-lead geochronology of zircon, performed with laser ablation inductively coupled plasma mass spectrometry at the Universities of Lausanne and Bern. Before any dating, researchers imaged the internal structure of individual zircon crystals using cathodoluminescence on a scanning electron microscope. This technique reveals growth zones—concentric bands that record successive episodes of crystallization. By carefully placing laser spots on specific zones rather than averaging across whole grains, the team could distinguish Ordovician growth domains from younger Variscan rims that overgrew them. In the Erstfeld Gneiss Complex, 151 spot analyses on sector- and oscillatory-zoned domains yielded a dominant age group of 455.1 ± 0.8/7.2 Ma, with no evidence of any Variscan zircon growth in the two studied samples.

The Strem Granite told a very different story. Zircon from three samples spread across dates ranging from roughly 2000 million years ago down to about 300 million years ago, reflecting a complex population of inherited cores and multiple growth episodes. Two main age populations emerged: oscillatory and sector-zoned domains of late Ordovician age at around 450 million years, and low-luminescence overgrowths dated to 311.2 ± 1.7/5.3 Ma. These younger rims are interpreted as zircon that crystallized from a new melt during the late Carboniferous. Crucially, the team also dated euhedral allanite crystals associated with large feldspar grains, which returned a U-Pb age of 315.1 ± 3.7 Ma—indistinguishable from the Variscan zircon domains. The high proportion of common lead in the allanite, between 20 and 80 percent, further supports an anatectic origin, meaning the mineral grew directly from a partial melt rather than from solid-state metamorphism.

Field relationships reinforce this interpretation. The Strem Granite shows mostly transitional contacts with the surrounding migmatites and gneisses rather than sharp intrusive boundaries, suggesting the melt never traveled far from its source. Fragments of the roughly 335-million-year-old Giuv Syenite are found as xenoliths within the Strem Granite, which brackets the remelting event between the intrusion of that syenite and the emplacement of the much younger Central Aar Granite around 298 million years ago. Previous geochronology on the Strem Granite had been inconclusive: multigrain zircon fractions analyzed in the 1990s gave an inaccurate age of about 350 million years because of unresolved inheritance, and later laser ablation spot dates scattered over 450 million years without yielding a coherent age. Only by combining cathodoluminescence imaging with targeted in-situ dating could the two overlapping melting events be separated.

Hafnium isotope analysis of zircon added another layer of evidence. The initial εHf values of Ordovician zircon domains range from -5 to 0, indicating a crustal, non-radiogenic source consistent with melting of clastic sediments. Variscan zircon rims show a similarly restricted range from -4 to +1, with one outlier at -7. This near-identical isotopic signature means the Variscan melt recycled the same chemical inheritance that had already been processed during the Ordovician event, with essentially no addition of new juvenile material. The detrital zircon cores, which span ages from roughly 500 to 720 million years with two outliers near 1.9 and 2.1 billion years, point to an accretionary wedge of Cambrian to lower Ordovician age that hosted Neoproterozoic and Cambrian detritus shed from the Cadomian orogeny along the northern margin of Gondwana.

Perhaps the most provocative finding concerns the physical conditions under which this remelting occurred. The researchers found no evidence of dehydration melting—no anhydrous peritectic minerals such as sillimanite or garnet that would normally form when hydrous minerals break down at high temperatures. Instead, they propose that the Ordovician migmatites were remelted at temperatures just above the solidus of the quartz-feldspar system, through fluid-assisted melting. Experimental data suggest that adding roughly 3.5 weight percent external water at 700 degrees Celsius and 5 kilobars would be sufficient to generate 50 percent melt in a haplogranitic composition. The team hypothesizes that this water was advected through numerous trans-crustal shear zones that were active during the late stage of the Variscan orogeny, when the thickened crust was extending and transtensional faults were slicing the orogenic root into discrete blocks.

The broader tectonic picture places this event within the final assembly of the supercontinent Pangea. During the Variscan orogeny, which ran from roughly 380 to 330 million years ago, the External Crystalline Massifs were caught up in continental collision between Laurussia and peri-Gondwanan terranes as the Rhenohercynian ocean closed. Late-collisional granitoids such as the Giuv Syenite and the Punteglias granite suite intruded around 335 million years ago. Afterward, as the subducted slab was lost through breakoff or thermal erosion, the thickened crust underwent extension and the underlying asthenosphere upwelled, providing the heat source for the remelting recorded in the Strem Granite. This late-orogenic phase, between roughly 315 and 300 million years ago, also saw the emplacement of the Fruttstock and Haslital intrusive groups, whose zircon shows increasingly juvenile εHf values up to +5, reflecting a growing mantle contribution to the magmatic budget.

The Erstfeld Gneiss Complex, by contrast, emerges as a lonely survivor from the Ordovician. Its zircon records only the Cenerian melting event at 455 million years, with no Variscan overgrowth whatsoever. A microdioritic dyke that crosscuts the Silberberg inlier at Susten Pass was dated by chemical abrasion isotope dilution thermal ionization mass spectrometry to between 313 and 323 million years, providing a minimum age for the metamorphic overprint of the northern unit. The researchers propose that the stark contrast between the two parts of the Aar Massif reflects late Variscan shearing and faulting, which juxtaposed blocks that once sat at different depths within the Variscan crust. The southern blocks experienced fluid-flushed remelting; the northern blocks underwent only solid-state metamorphism.

What makes this study resonate beyond the Alps is its methodological lesson. Rocks that look identical in the field can carry entirely different ages, and only the internal architecture of accessory minerals can reveal the truth. Zircon, with its resistance to chemical dissolution and mechanical abrasion, preserves successive melting events in zircon-saturated magmas that would otherwise erase each other. The recognition that Ordovician migmatites were quietly remelted into Variscan granites—without any obvious field expression—suggests that similar hidden episodes of crustal recycling may await discovery in other ancient orogenic belts, wherever water found its way into hot, deep crust along the faults that stitched continents together.

Subject of Research: Remelting of Ordovician migmatites and granites during Variscan high-grade metamorphism (Aar Massif, Central Alps, Switzerland)

Article Title: Remelting of Ordovician migmatites and granites during Variscan high-grade metamorphism (Aar Massif, Central Alps, Switzerland)

Article References: Remelting of Ordovician migmatites and granites during Variscan high-grade metamorphism (Aar Massif, Central Alps, Switzerland). (n.d.). https://doi.org/10.3897/sjg.119.183954

Image Credits: AI Generated

DOI: 10.3897/sjg.119.183954

Keywords: Remelting, Ordovician, migmatites, granites, during, Variscan, high-grade, metamorphism, Massif, Central, Alps, Switzerland