An abrupt switch in magmatic plumbing taps porphyry copper deposit-forming magmas


 Porphyry-type deposits are a vital source of green technology metals such as copper and molybdenum. They typically form in subduction-related settings from large, long-lived magmatic systems. The most widely accepted model for their formation requires that mantle-derived magmas undergo a multi-million year timescale ramp-up in volatiles and ore-forming constituents in mid- to lower-crustal reservoirs, however this does not explain why porphyry deposits are absent from the vast majority of arc magmatic systems. To address this, we have carried out geochemical and geochronological studies on the tilted, ~8 km depth equivalent, cross-section through the classic Yerington magmatic system, Nevada. Here we show that the magmas underwent a major and abrupt change in chemistry over a period of 100 kyrs which is coincident with the initiation of ore formation. This is attributed to a wholesale switch in the magmatic plumbing system whereby volatile-rich granitic melts were extracted from an estimated ~30 km depth and transported to shallow levels (~3-8 km) where exsolving fluids were focussed through highly permeable pathways to form porphyry deposits. The change in magma chemistry is documented across the entire plutonic to volcanic record. Its rapidity suggests that the increase in a magma’s ore-forming potential is not solely driven by tectonic factors, that occur over multi-million year scales, but through internal processes within the melt evolution zone, operating at more than an order of magnitude faster than previously envisaged. This short timescale narrows the temporal-geochemical footprint of magmas associated with porphyry mineralisation which will aid in targeting the next generation of ore deposits.

timescale ramp-up in volatiles and ore-forming constituents in mid-to lower-crustal reservoirs, however this 23 does not explain why porphyry deposits are absent from the vast majority of arc magmatic systems. To 24 address this, we have carried out geochemical and geochronological studies on the tilted, ~8 km depth 25 equivalent, cross-section through the classic Yerington magmatic system, Nevada. Here we show that the 26 magmas underwent a major and abrupt change in chemistry over a period of 100 kyrs which is coincident 27 with the initiation of ore formation. This is attributed to a wholesale switch in the magmatic plumbing system 28 whereby volatile-rich granitic melts were extracted from an estimated ~30 km depth and transported to 29 shallow levels (~3-8 km) where exsolving fluids were focussed through highly permeable pathways to form 30 porphyry deposits. The change in magma chemistry is documented across the entire plutonic to volcanic 31 record. Its rapidity suggests that the increase in a magma's ore-forming potential is not solely driven by 32 tectonic factors, that occur over multi-million year scales, but through internal processes within the melt 33 evolution zone, operating at more than an order of magnitude faster than previously envisaged. This short 34 Introduction 38 The shift to new and green technologies is driving the increasing requirement for metals 1,2 . Copper demand 39 is forecast to increase by 140-350% from 2010 to 2050 3,4 . Porphyry-type deposits provide more than 70% of 40 global copper, around 95% of molybdenum (~95%) and important amounts of gold (20%) and other metals 5 . 41 They form from hydrothermal fluids produced by large and long-lived calc-alkaline to slightly alkaline, water-42 rich and relatively oxidising trans-crustal magmatic systems, mostly in subduction-related settings e.g. [5][6][7][8][9] . Whilst 43 such magmas are arguably relatively common, porphyry-, and particularly large porphyry-type deposits are 44 extremely rare and increasingly difficult to find 7 . Their formation probably requires a series of specific 45 conditions and events during the evolution of the magmatic-hydrothermal system. 46

47
In the drive to discover new ore-deposits, there have been many recent attempts to develop whole-rock and 48 mineral geochemical indicators to assess the ore-forming potential, or the fertility, of magmatic systems 10 . 49 Their main advantage compared with conventional exploration techniques is that they are cheap and have 50 low environmental impact. Most indicators reflect the geochemical signatures of amphibole fractionation and 51 plagioclase suppression due to the hydrous nature of the magmas from which porphyry-type deposits 52 form e.g.7,11-20 . 53 54 The current paradigm is that the hydrous magmas that form porphyry-deposits result from an extremely long 55 (multi-million year), arc-scale, tectonically driven "ramp-up" in volatiles and ore-forming constituents in mid-56 to lower-crustal magmatic reservoirs 11, 16,[21][22][23][24][25] . Here, mantle-derived melts develop their hydrous ore-forming 57 geochemical signatures over protracted time scales due to cyclical fractionation and re-fertilisation by mafic 58 magmas, before emplacement into the upper crust. The macro-scale tectonic regime has been suggested to 59 progressively deepen the melt evolution zone and slow the upwards migration of magmas through the 60 crust 22,24,26 . In an alternative model, it is suggested that magmas acquire their ore-forming geochemical 61 signatures during evolution within an upper crustal staging ground e.g. 5,27,28 . Such conceptual models for the 62 evolution of magmatic systems linked to porphyry-style mineralisation are however greatly limited by a 63 paucity of vertically extensive exposure over the crustal windows of porphyry ore-forming systems 29 . We 64 therefore have a fragmented understanding of the timescales of porphyry-deposit formation and the 4-D solidification textures (USTs), but are also cross-cut by mineralised veins (Fig. 3c-f & S3). The quartz USTs 119 within the aplites is likely to indicate rapid temperature or pressure fluctuations 45 and fluid exsolution via first-120 type boiling 46 , suggesting that the mineralising aplite dykes were strongly undercooled having been emplaced 121 rapidly to shallow depths. Given the aplite dykes host mineralised miarolitic cavities which are closely 122 associated with early mineralised veins (A-type 44 ) (Fig. 3e & 3f), they capture the nature and timing of 123 magmatic-hydrothermal fluid exsolution and mineralisation, as well as having acted as "crystal mush" 124 conduits for mineralising fluids from deep portions of the LHG 32 . 125 126 Field relations indicate that some parts of the Fulstone volcanics were cogenetic with the emplacement of 127 porphyry dykes associated with the LHG 40,41 . Propylitic alteration is ubiquitous across the Fulstone Spring 128 Volcanics (e.g. epidote replacing primary plagioclase and chlorite replacing mafic minerals) (Fig. S4), 129 indicating that the hydrothermal system was active for some time after volcanism. The lack of more acidic 130 alteration (e.g. advanced argillic) indicates that these volcanics, if related, were distal to ore-forming 131 hydrothermal activity. 167.365 ± 0.041 Ma and 167.275 ± 0.027 Ma, respectively (Fig. 4). These ages define the maximum repose 146 time between the McLeod QMD/BH and LHG of 215 ± 59 kyrs. When the 167.440 ± 0.039 Ma age of a 147 mineralised porphyry dyke (AC25) is considered, which cross-cuts the LHG cupola ~1 km higher in the 148 system than AM18LHG, then this repose time must be even shorter (140 ± 57 kyrs, or ~ 100 kyrs). The 149 crosscutting relationships imply a protracted emplacement of the exposed LHG over >~150 kyrs. These new 150 timescales also indicate that construction of the Yerington batholith was significantly longer than previously 151 estimatedmeaning that long-term emplacement rates by protracted episodes of magma recharge were 152 ~two times slower than the minimum rates previously estimated 31  Myrs, rather than a single post-ore volcanic event 34 . 159

160
The timing of porphyry-style Cu-Mo mineralisation has been constrained from cross-cutting relationships and 161 Re-Os molybdenite ages (Fig. 4). Ages for the mineralised porphyry dykes that cross-cut the cupolas of the  162   LHG and the youngest ages for the McLeod QMD and Bear QM both constrain the onset of ore formation at  163 ~167.4 Ma. The multiple generations of aplite dykes (AM63A and AM13BAP) capture the nature and timing 164 of magmatic-hydrothermal fluid exsolution and mineralisation, and acted as conduits for the transport of 165 mineralising fluids into the ore-forming environment 32 (Fig 3 & S3). As the youngest zircon growth contained 166 within the aplite dykes likely crystallised as part of the magmatic assemblage, the U-Pb ages of 167.282 ± 167 0.040 Ma and 167.045 ± 0.057 Ma (Fig. 4)   There is no major difference in zircon composition between the LHG and aplite dykes. Zircon data of the 215 porphyry dykes show no notable trace element differences between dyke generations of the "early" and "late" 216 mineralised dykes from both the Ann Mason and Yerington porphyry copper deposits, in agreement with 217 previous zircon data from the Yerington system 36  indicating no discernible assimilation of previous crust after the melts were extracted. This indicates that the 271 mineralising melts evolved within an environment where they were only exposed to a single, homogenous, 272 mantle-derived source melt prior to emplacement. This supports different evolution zones for the pre-and 273 inter-mineralisation melts. These could be either disparately within the crust, or within the same "hot-zone" 54  Hf concentration (Fig. 6 & 7). This abrupt shift in the dominant fractionating assemblage is constrained to 311 within ~100 kyrs and is coincident with the onset of porphyry mineralisation. Following this, the mineralising 312 magmatic system was stable for potentially in excess of 1 Myrs, with the youngest magmatic activity being 313 documented within the propylitically altered volcanics, that bear the same zircon geochemical signatures as 314 the inter-mineralisation LHG samples (Fig. 4, 7 & S4). 315

316
The abrupt change in the magmatic chemistry to more hydrous and amphibole-dominated signatures can be 317 explained by a phase boundary change in the magma extracted from a single magma evolutionary zone, 318 during its longer-term progressive evolution. Within the lower crust, fractionation of anhydrous phases in the 319 relatively dry magmas, as magmatism continued over 1.6 Myrs, progressively ramped up the concentrations 320 of volatiles. The sudden change reflects the point at which the "amphibole-in" line was suddenly crossed due 321 to the volatile build up, or due to the injection of new melts into a lower crustal clinopyroxene cumulate pile 322 or "sponge", that reacts with new melt to become progressively replaced by amphibole 62 . The dated porphyry 323 dyke that sits at the temporal onset of mineralisation (AC25; Fig. 4) and has zircon geochemistry appearing 324 to straddle the pre-and inter-mineralised signatures (Fig. 7 & S8-10) could mark this threshold being crossed. 325

326
Although we cannot rule out this transitional model to describe the evidence, there are a number of features 327 that do not support that they resulted from progression within a single magma evolution zone. For example, 328 the porphyry dykes hosted by and therefore relatively later than the LHG also contain zircon with pre-329 mineralisation chemistry. The sharp contacts between the mineralogical distinct McLeod QMD and LHG 330 plutonics 27 and the temporally abrupt stepwise change in magma geochemistry following 1.6 Myrs of pre-331 cursor magmatism is suggestive of a fundamental shift in where the melt is extracted from, rather than a 332 transition of a single melt extraction zone within the system. A scenario where a single melt extraction zone 333 suddenly changes fractionating assemblage is also challenging to reconcile given the shift in the isotopic 334 data from a heterogeneous signature, which indicates interaction with crustal components, to a 335 homogeneous more juvenile, mantle-derived signature particularly over the short timescale the data identify 336 (Fig. 8). 337

338
The abrupt, ~100 kyrs change after the >1.6 Myrs period of emplacement of the pre-mineralisation plutons 339 is better explained by a rapid switch in the magmatic plumbing system, probably within the mid-to lower-340 crustal melt evolution zone from where the magmas feeding the upper crustal reservoir were derived (Fig.  341 10). We envisage that the earlier, pre-mineralisation stage magmas were derived from the mid-to lower-342 crust (~15-20 km; Fig. S11). During protracted storage and evolution, these assimilated crustal material. It is 343 feasible that the contrasting nature of the McLeod QMD and Bear QM 27 , yet in-part contemporaneous 344 emplacement periods (Fig. 4), reflect slightly different zones of storage and evolution during this period of 345 batholith growth. After the switch, magmas in the lower crustal "hot-zone" 54 (~20-40 km; Fig. 9 & 10) evolved 346 to more volatile-rich compositions which initiated the crystallisation of amphibole and/or the reaction of the 347 more hydrous melts with clinopyroxene in a mush or pre-existing cumulate pile. The more evolved melts 348 tapped from this zone had only interacted with precursor mantle-derived magmas, and possibly cumulates, 349 explaining their lack of crustal isotopic signatures (Fig. 8). In this scenario the pre-mineralisation geochemical 350 signature of the zircon cargo of the porphyry dykes would be acquired as they punched up through the pre-351 cursor magmatic system on route to their emplacement levels. 352

353
Although there is little distinction between the LHG and porphyry dykes in terms of whole-rock geochemistry 354 ( Fig. 5 & S6), from differences in their zircon trace element compositions and isotopes (Fig. 7 & 8), 355 comparable melt minima barometry (Fig. S11) along with the paucity of field evidence, it is very unlikely that 356 the porphyry dykes were derived from the upper parts of the LHG (as per previous models e.g.27 ). From textural 357 evidence (Fig. S2), once emplaced at shallow crustal levels, the magmas underwent further magmatic 358 differentiation (at ~ 3-8 km depth, based on melt minima plots; Fig. 9) to form the more evolved and volatile-359 rich melts that were episodically injected as undercooled aplite dykes over a period of at least ~400 kyrs. The proposed time-period for porphyry ore formation, which is likely to have exceeded 1 Myrs post 367 emplacement of the LHG, is not uncommon for medium to large scale, composite porphyry systems e.g.24,63-368 65 . Given the similar zircon trace element geochemistry between mineralised porphyry dykes in the Ann 369 Mason and Yerington porphyry deposits 36 (Fig. S8-10), along with their petrographic similarities 27 , they are 370 most likely co-genetic. It is probable that these are also co-cogenetic with porphyry dykes in the Yerington 371 districts' two other known porphyry deposits: Bear and MacArthur. It is therefore salient for future numerical 372 models and computational simulations of batholith construction and porphyry mineralisation to include fluids 373 derived from across all porphyry centres, which would produce a considerably larger copper endowment. 374 375 Genetic implications for porphyry deposit-forming magmatic 376 systems 377 The apparent change in geochemistry (whole-rock and zircon) as the Yerington system began to produce 378 porphyry deposits is consistent with observations across a wide range of global localities where precursor 379 magmatism and syn-mineralisation intrusions have been examined e.g. 19,20,22,25,66 . Typically these changes 380 have been interpreted as being due to long-term, arc-scale, transitional "ramp-ups" towards ore-formation 381 over millions of years e.g.11,16,21,23 , however this model was constrained by limited exposure in most systems 29 . 382 Our detailed studies of the considerably better-exposed Yerington batholith suggest that the switch to 383 porphyry deposit-producing magmatism was at least an order of magnitude faster than previously thought. 384 The short timescale of the geochemical changes suggested here does not necessarily contradict longer-term 385 progressions towards hydrous, ore-forming arc magmas seen in other magmatic systems. The longer 386 durations documented in other systems between precursor and ore-related magmatism, and their 387 corresponding changes in geochemistry, may simply relate to the juxtaposition of upper crustal magmatic 388 expressions over the protracted duration of the magmatic system. For example, when temporally comparing 389 dyke to host rock in Yerington, there is a difference of up to ~1.7 Myrs for dykes that intruded up through the 390 LHG and into shallower levels of the McLeod QMD. This means the footprint of ore-bearing systems may 391 differ with exposure levelat shallower levels the record would look discrete and appear to develop over 392 longer timescales, whereas at depth the system appears more concurrent. 393 394 We, however, recognise that a wholesale change in the deeper magmatic system occurs at a resolution 395 which can be used to isolate geochemical signatures in the rock record. The timescale of this process 396 appears to be beyond a multi-million year magmatic "ramp-up" driven by the macro-scale tectonic regime 397 11,16,21-26 . Instead, our results indicate how the magmatic systems can much more rapidly develop their ability 398 to form porphyry copper deposits and that this must relate to processes that are driven internally by the 399 magmatic processes of melt generation and the extraction of evolved melts within the system itself (Fig. 10). 400

401
The recognition of a rapid ~100 kyrs switch within the magmatic plumbing system requires a new approach 402 in the interpretation of plutonic processes in ore-forming systems. The magmas responsible for ore-formation 403 underwent different routes of evolution and were tapped from spatially independent melt zones within the 404 lower crust compared to the magmas which formed the pre-cursor plutons. We suggest from this that the 405 processes and evolution histories of early intruded plutons cannot necessarily be used to infer whether other 406 parts of the batholith may have produced porphyry-type deposits. 407 408 Our research offers new constraints on the depth of different melt evolution zones below the Yerington 409 porphyry district, and the abrupt nature of the change between zones of melt extraction at the onset of 410 mineralisation. This has important implications for models of batholith construction, the formation of porphyry-411 type deposits and the development of porphyry exploration indicators. This is mainly because the short-412 timescale over which the geochemical signatures associated with mineralisation appeared throughout the 413 magmatic system, in plutons, dykes and volcanics, significantly narrows the temporal window for ore 414 formation. This increases the potential efficacy of using these geochemical signatures to isolate areas 415 prospective for porphyry-style mineralisation. This greater confidence in the resolution of these signatures is 416 important in aiding discovery of the next generation of porphyry deposits, which are likely to be deeper and 417 often under cover and so will be more difficult to find 10  boundaries. Samples are submerged in a dielectric process medium such as water, which is more resistive 482 than solids at these pulse rise times, resulting in the discharge being forced through the relatively conductive 483 solid and along internal phase boundaries such as mineral-mineral contacts. Each discharge event is a 484 movement of electrons from the working electrode to the ground electrode as a plasma channel 71,73 . The 485 rapid formation of this plasma channel causes explosive expansion within the material along the discharge 486 pathway 71,72 . In addition to direct breakage from the plasma channel, this explosion creates a shockwave that 487 propagates through the material. Varying elasticity moduli between minerals results in shear stresses being 488 focussed on mineral contact surfaces, causing intra-mineral breakage and disaggregating the rock. This 489 tensile intra-mineral breakage is less damaging to individual minerals which are liberated from the rock larger 490 and more intact than mechanical crushing. 491

492
The treatment was conducted using the 'Lab', a laboratory scale EPF device for the batch processing of to the sample, with treatment conditions for this work listed in Table 1. Further information on the Lab system 500 can be found in 74  the SELFRAG open process vessel. Appropriate sieve aperture diameter is generally equal to 10x the target 507 particle diameter. A series of 100 pulses were applied to the sample followed by visual inspection of the 508 remaining sample; if >10 % if the sample remained above the sieve, another cycle of 100 pulses were 509 administered. When >90 % of sample material had passed through the sieve, treatment was stopped, and 510 the sample recovered from the process vessel collection cup before drying at 70°C. Whole-rock XRF geochemistry was used to calculate CIPW normative mineralogy (method of 70 , after 75 ). 529 Normative mineralogy data was then plotted on the H2O-saturated melt minima ternary plot 61 to estimate the 530 pressures of melt differentiation 54 of H2O-saturated melts. Assuming lithostatic conditions, pressures from 531 this plot were used to equate approximate depths of melt differentiation using P = ρgh and assuming an 532 average overburden density of 2.5 g/cm 3 . 533 534 Zircon Separation 535 Zircons were separated from disaggregated samples at the British Geological Survey, Keyworth, using the 536 sequentially described circuit: Sieve to <500 μm using a Fritsch automatic sieve; Pass the <500 μm fraction 537 over a Gemini water table, twice; Separate non-magnetic minerals using a Frantz isodynamic separator -538 subsequent paramagnetic charges of 0.1 A, 0.3 A, 0.7 A, 1.1 A and 1.7 A were used to reduce the bulk 539 material in stages; Perform gravity separation utilising methylene iodide (ca. 3.32 SG) as a density medium. 540 The final zircon (amongst other phases) separate was thermally annealed at 900°C for 12 hours. Annealed 541 zircon grains were then picked by hand and prepared as polished blocks. Cathodoluminescence (CL) images 542 of these were generated by SEM-CL, using an FEI Quanta 650F FEG-SEM equipped with a Gatan 543 monochrome CL detector at the University of Exeter's Environment and Sustainability institute operating at 544 an accelerating voltage of 20 kV, as well as using a CITL Mk5 electron source, operating at approximately 545 250 uA and 10 kV. For the latter, images were captured using a Nikon DS-Ri2 camera, attached to a 546 petrographic microscope, and operated using NiS-elements software. Images were captured in a darkened 547 room, with an exposure time of 2 seconds. Keyworth. After thermal annealing at 900°C zircon were chemically abraded at 190 °C for 12 hours following 562 76 . The methodology for all other analytical procedures, instrumental conditions, corrections and data 563 reduction follows that outlined in detail in 77 using the ET(2)535 tracers 78,79 . Isotope ratio measurements were 564 made using a Thermo Triton thermal ionization mass-spectrometer (TIMS), with the U decay constants of 80  Standard sample cones and X-skimmer cones were used, giving a typical signal of c. 800-1000 V/ppm Hf. 604 Correction for 176 Yb on the 176 Hf peak was made using reverse-mass-bias correction of the 176 Yb/ 173 Yb ratio 605 empirically derived using Hf mass-bias corrected Yb-doped JMC475 solutions 88 . 176 Lu interference on the 606 176 Hf peak was corrected by using the measured 175 Lu and assuming 176  Os bearing a normal isotope composition) were placed into a carius tube and digested with 3mL HCl and 628 6mL HNO3 at 220°C for 23 hrs. Osmium was isolated and purified using solvent extraction (CHCl3) and micro-629 distillation methods, with the resulting Re-bearing fraction purified using NaOH-Acetone solvent extraction 630 and anion chromatography 90,91 . Although negligible in comparison to the Re and Os abundance in the 631 molybdenite, the final Re-Os data are blank corrected. A full analytical protocol blank run parallel with the 632 molybdenite analysis yields 3.9 pg Re and 0.5 pg Os, the latter possessing a 187 Os/ 188 Os composition of 0.21 633 ± 0.2. Data treatment follows that outlined in 91  Temporal relations in the Yerington magmatic system Field photographs of a cross cutting relations of multiple porphyry dyke generations which cut the LHG cupola b lobate contacts and evidence for mingling of co eval magmas between an aplite dyke and porphyry dyke Secondary copper staining prevalent in the aplite dyke c d, multiple generations of aplite dykes hosting pegmatitic segregations and mineralised miarolitic cavities (MC The aplite dykes sharply cross cut the cupola zone of the LHG and a porphyry dyke Both the aplite and porphyry dykes lie palaeo vertically beneath the Ann Mason porphyry deposit e, cupola zone of LHG cut by an aplite dyke hosting a chalcopyrite ( mineralised miarolitic cavity Qtz quartz f, drill core from the Ann Mason porphyry deposit showing LHG cut by an aplite dyke hosting miarolitic cavities and early chalcopyrite bornite quartz (Ccp Bn Qtz) A type, nomenclature after 44 veins, which locally truncate at the dyke's margin B, e f after 32 Figure 4 Geochronological framework for the Yerington porphyry system Zircon single grain U Pb CA ID TIMS and molybdenite Re Os geochronological framework for samples spanning the Yerington magmatic system Pre and inter mineralisation intrusive samples grouped and plotted in order of approximate palaeo depth 27 Sample details in S upplementary Data 1 We take the weighted mean of the youngest population of zircon dates that formed a statistically acceptable Mean Square Weighted Deviation ( or chi squared) as the best approximation for the crystallisation of the host magma Porph porphyry M S mass spectrometry Error bars at 2 σ  Zircon trace element signatures through the Yerington porphyry system Zircon LA ICP MS trace element data from samples spanning, temporally and spatially, the Yerington porphyry system ..'Pre mineralisation' and 'inter mineralisation' elds shaded Only zircon rim data have been plotted See Fig S 7 10 for full sample breakdown along with zircon core data Depth of different magma sources CIPW normative mineralogy (method of 70 from whole rock XRF data for LHG and aplite dykes plotted on the H O saturated haplogranitic melt minima plot 61 Cotectic lines and eutectics are a function of pressure and therefore the whole rock data can be used to provide constraints for the pressure of magma differentiation 54 from which depth can be approximated Other units plotted in Fig S 11

Figure 10
A rapid switch in magmatic plumbing to tap porphyry mineralising magmas Simpli ed system paragenesis and conceptual cross section through the porphyry system A long lived 1 6 Myrs) evolution and contemporaneous emplacement of precursor plutonics, with volcanic activity, was followed by a rapid 100 kyrs) switch in magmatic plumbing to tap fertile porphyry deposit forming magmas from a 20 40 km deep lower crustal staging ground where they predominately underwent amphibole dominated fractional crystallisation From this zone of melt evolution, fertile magmas were emplaced into the shallow crust to form plutons and porphyry stocks, and underwent further differentiation at 3 8 km depth, with episodic upward injection of multiple generations of aplite dykes for 400 kyrs, which acted as crystal mush conduits for mineralising uids 32 As mineralising uids exploited these conduits, porphyry deposit formation continued episodically for potentially in excess of 1 Myrs, and may have been co eval with volcanism M cavs miarolitic cavities USTs unidirectional solidi cation textures Vein nomenclature after 44 Modi ed after 5 32 54 Supplementary Files This is a list of supplementary les associated with this preprint. Click to download.