Ultradeep drilling beyond 10 km revealing new insights into Earth systems and resources

Abstract

Ultradeep drilling beyond nine to ten kilometers has progressed from an engineering challenge to a powerful approach for investigating deep Earth processes and subsurface resources. Early continental drilling showed that the deep crust is not cold or impermeable, but contains fluids, active heat flow, and ongoing chemical reactions. This review presents an integrated synthesis of scientific results from drilling that exceeds ten kilometers, with emphasis on recent Chinese programs. These wells document methane-rich fluids, high pressures, and high temperatures in sedimentary and crystalline rocks, indicating that hydrocarbon systems, deep fluids, and permeability persist far deeper than traditionally assumed. Integrating observations from cores, formation fluids, stress measurements, and geochemical analyses, we show that revised geothermal gradients, fault-controlled fluid pathways, and fluid–rock reactions such as serpentinization, carbonation, and hydrocarbon cracking sustain permeability at extreme depths. Ultradeep drilling thus provides insights into Earth system dynamics, deep energy resources, carbon storage, and mineral systems.

Introduction

Drilling has long been one of humanity’s most powerful approaches to exploring Earth’s interior, providing direct access to rocks, fluids, and pressure–temperature regimes otherwise beyond the reach of laboratory simulation or surface observation¹. Boreholes yield cores, logs, and fluids that illuminate lithology, mineralogy, geochemistry, stress states, and thermal gradients, thereby testing fundamental models of plate tectonics, mountain building, volcanism, and fluid–rock interaction². Ocean drilling programs have already validated seafloor spreading, reconstructed paleoclimate histories, and traced geomagnetic field reversals³. Yet for continental settings, the pursuit of ultradeep drilling represents a scientific imperative: only such access can provide coherent lower-crustal sections, mantle samples, and ground-truth data that bridge theoretical models with physical reality^4,5.

The historical trajectory of ultradeep projects underscores both discovery and challenge. The Russia’s Kola Superdeep Borehole (SG-3; 12,262 m) revealed that crystalline basement is fractured, fluid-bearing, and thermally dynamic, in contrast to the “sealed crust” envisioned from seismic models^4,6. Germany’s KTB project (9101 m) advanced seismic imaging and pioneered stress and conductivity measurements⁷. China’s Deep Geological Survey Project (launched 2016) expands the SinoProbe vision with integrated deep seismic reflection, wide-angle/refraction, magnetotellurics, and scientific drilling across key basins and orogens⁸. More recently, China has redefined the frontier with SDTK-1 (10,910 m) in the Tarim Basin and X-1 (11,156 m) in the Junggar Basin, the first vertical >10 km boreholes drilled specifically for petroleum exploration, which confirmed hydrocarbon flows at extreme depths. These milestones highlight drilling as both a technological and scientific gateway, demonstrating that the deep crust is far more dynamic and resource-bearing than once assumed.

Ultradeep drilling is driven by three overarching scientific priorities: (1) resolving the structure and evolution of the continental lithosphere, (2) characterizing deep fluid circulation and element cycling, and (3) redefining the physical–chemical limits of hydrocarbon, hydrogen, and geothermal systems. These themes form the backbone of this study and are emphasized here to clarify the narrative framework.

The scientific rationale extends beyond engineering achievement. To date, no borehole has penetrated the Mohorovičić discontinuity (Moho), leaving Earth’s mantle—the largest portion of the planet’s volume—unsampled in situ⁹. Direct recovery of unaltered peridotite would provide essential ground truth for mantle composition, heterogeneity, and magma-generation processes. Such access also has planetary implications, refining estimates of bulk composition, heat budgets, and volatile cycles while testing the persistence of microbial life in extreme mantle-related environments¹⁰. These efforts align with broader questions of global geodynamic cycles, including the Wilson cycle, crustal growth, subduction, plume activity, and supercontinent assembly, all of which leave signatures only recoverable through deep drilling^1,11.

Ultradeep drilling also intersects directly with resource frontiers. Discoveries in China’s Tarim Basin indicate that petroleum systems can remain viable at depths exceeding 10 km, with the Cambrian Yurtus Formation sustaining hydrocarbon generation under temperature and pressure conditions that extend beyond the conventionally inferred lower depth limit of ~8 km^12,13,14,15. These findings reshape models of oil and gas preservation, fracture behavior, and deep fluid migration.

Parallel advances in geothermal energy also highlight ultradeep drilling’s role in zero-carbon transitions. Enhanced geothermal systems, particularly in hot, dry crystalline rocks, could provide sustainable baseload power; Canada, for example, is positioned to lead in this domain by leveraging petroleum drilling expertise¹⁶. Likewise, ultradeep wells intersect natural hydrogen accumulations and magmatic systems, linking drilling to hydrogen resources and carbon sequestration^17,18.

However, drilling beyond 10 km confronts formidable challenges. Extreme pressures and temperatures (>140 MPa, >200–250 °C) cause tool wear, fluid instability, and rapid equipment failure¹⁹. The concept of “extreme drilling conditions” captures the limits imposed by high hydrostatic pressures, unstable wellbores, and bottomhole temperatures approaching the endurance thresholds of drill strings, assemblies, and drilling fluids². Progress has been enabled by high-temperature polycrystalline diamond compact (PDC) bits, advanced alloys, high-performance drilling fluids, and automated rigs²⁰. Yet persistent gaps remain in high-temperature electronics, rotary-steerable tools, and ultrahigh-temperature/pressure drilling fluids. Future advances in AI-driven monitoring, digital-twin well construction, and closed-loop pressure management will be pivotal to sustaining operations under these conditions.

The scientific significance of ultradeep drilling is thus multi-dimensional. Boreholes provide direct measurements of heat flow, stress, permeability, and seismic velocities at depths where geophysical models diverge, constraining theories of mantle convection, lithospheric strength, and plate behavior²¹. They clarify fluid–rock reactions and element cycling, illuminating the origins of deep fluids, ore mineralization, and long-term carbon dynamics²². They also extend the depth limits of petroleum system models, while enabling new frontiers in hydrogen, geothermal, and CO₂ sequestration²³. Equally, deep boreholes serve as observatories for seismicity, stress evolution, and fluid migration, offering real-time data essential for hazard forecasting^24,25.

Despite major advances achieved by deep-well programs worldwide, critical scientific gaps remain. Landmark projects—including the Kola Borehole, Germany’s KTB, and China’s SK-2, SDTK-1, and X-1—transformed understanding of crustal geology but were constrained by the analytical technologies available at the time. Early campaigns lacked the capacity to resolve nanoscale porosity, organic–inorganic interactions, deep-biosphere signatures, and transient fluid pathways that are now identifiable using modern isotopic, microstructural, and in situ sensing methods. Key datasets—such as deep-crustal fluid inclusions, fracture permeability, and ultrahigh-temperature geochemical anomalies—therefore remain only partially interpreted, and could yield new insights through re-analysis with techniques such as µCT, atom-probe tomography, U–Pb nano-geochronology, and distributed fiber-optic sensing. In addition, emerging concepts such as deep exudation-type uranium systems sourced from uranium-rich source rocks^{26, 27} and hydrogen generation in crystalline rocks²⁸ were unknown during early drilling efforts, implying that certain intervals may require reinterpretation to evaluate overlooked metallogenic or energy implications. Together, these considerations underscore the value of integrating legacy cores and logs with contemporary analytical capabilities to fully realize the scientific potential of past ultradeep wells.

Accordingly, this study synthesizes scientific discoveries, technological advances, and resource implications associated with ultradeep drilling, with particular emphasis on China’s >10 km wells as representative examples at the current exploration frontier. By clarifying key scientific motivations—including crustal architecture, deep fluid circulation, and constraints on resource systems—this introduction establishes a coherent framework for evaluating ultradeep drilling as an important tool for advancing Earth science and sustainable energy research.

Drilling to extremes: progress, records, and global distribution

Ultradeep drilling represents one of the most ambitious frontiers in Earth science and resource exploration. Since the 1970s, boreholes exceeding 7–10 km have been attempted across diverse tectonic settings, yielding breakthroughs in crustal structure, heat flow, and resource potential (Table 1).

Table 1 Drilling findings and significance of global ultradeep wells

Global distribution of ultradeep wells

Figure 1 illustrates the global distribution of ultradeep wells, highlighting key scientific drilling and exploration projects across major continental regions. The deepest penetration is the Russia’s Kola Superdeep Borehole, which revealed fractured, water-bearing basement rocks and unexpectedly high heat flow, while Germany’s KTB pioneered seismic, stress, and conductivity measurements and established a long-term observatory^29,30,31. In North America, ultradeep activity includes offshore Gulf of Mexico wells (>9–10 km) that advanced high-pressure, high-temperature (HPHT) drilling, and onshore wells such as the Bertha Rogers (9583 m) that extended petroleum system models³². Mexico, Central Europe, Turkey–Middle East, and Scandinavian programs have also drilled >8–9 km or deep experimental boreholes, targeting hydrocarbons, geothermal resources, seismic–tectonic research, and basement evolution and fluid systems². The Tarim and Junggar ultradeep wells ‒ SDTK-1 and X-1 ‒ confirmed hydrocarbon accumulations beyond 10 km and placed China at the forefront of ultradeep exploration. These projects confirm seismic models, show deep fluid circulation and hydrocarbons beyond 10 km, refine tectonic and crust–mantle models, and spur advances in drilling tools, fluids, pressure control, and borehole observatories. Looking ahead, ultradeep drilling is expected to expand toward hydrocarbons, hydrogen, geothermal energy, and even mantle penetration, linking fundamental Earth science with the energy transition.

Fig. 1: Global distribution and drilling depths of representative ultra-deep wells.

a Global map showing the geographic distribution of selected ultra-deep wells (>7–10 km km) drilled for scientific research and petroleum exploration (locations modified from ref. ²). Colored circles denote different well types: production wells (red), appraisal wells (orange), prospecting wells (blue), and parametric wells (cyan). b Bar chart summarizing the maximum measured drilling depths of the wells shown in (a), with numerical values (in meters) indicated above each bar and colors corresponding to the same well-type classification as in (a).

Russia’s Kola superdeep borehole: Scientific framework and discoveries

The Russia’s Kola Superdeep Borehole (SG-3) remains the deepest vertical borehole ever attempted, reaching 12,262 m between 1970 and 1989, though the original goal was 15 km (Fig. 2a)²⁹. Drilling was abandoned in 1992 when unexpectedly high bottom-hole temperatures of 212 °C induced plastic rock flow that repeatedly sealed the hole. Despite not penetrating the Moho, SG-3 yielded the most extensive direct dataset on the crystalline continental crust.

Fig. 2: . Location and lithologic architecture of the Kola Superdeep Borehole (SG-3), northwestern Russia.

a Geographic location of the Russia’s Kla Superdeep Borehole (SG-3) on the Kola Peninsula, northwestern Russia. b Generalized 12-km lithologic profile along the borehole trajectory, illustrating Proterozoic supracrustal cover rocks, Archean crystalline gneiss basement, major fault zones, and the slightly deviated drilling path (Modified from refs. ^29,30).

The SG-3 well fundamentally reshaped understanding of continental crustal structure and deep geologic processes. Drilling demonstrated that the classical granite–basalt two-layer crustal model is invalid in the Fennoscandian Shield, revealing instead a continuous transition from Proterozoic supracrustal sequences into Archean granite–gneiss basement and showing that major seismic discontinuities reflect metamorphic and rheological contrasts rather than lithologic boundaries^33,34. SG-3 provided the first complete section of the Pechenga supracrustal complex (0–6842 m) and unexpectedly encountered no basaltic amphibolite layer, disproving the existence of a Conrad surface near 7 km and instead identifying deconsolidated zones and thrust structures that redefine the architecture of the Lapland–Pechenga megablock³⁵. The lithologic profile (Fig. 2b) comprises (i) Proterozoic sandstones, tuffites, phyllites, dolostones, diabases, altered intrusives, and schists, and (ii) deeper Archean mica–plagioclase gneisses, with major faults such as the Poritash Fault acting as highly permeable, fluid-rich conduits to 6–8 km³⁶. These intervals revealed a far more hydraulically active crystalline crust than previously assumed, capable of long-distance fluid migration, episodic overpressure, and localized thermal–chemical alteration—an analogue directly relevant to >10 km modern drilling.

SG-3’s in situ measurements provided unprecedented constraints on deep-crustal conditions: temperatures reached 180 °C at 10 km and 212 °C at 12 km, far exceeding pre-drilling predictions and forcing revision of geothermal gradients and heat-flow models³⁷. Circulating mineralized brines, hydrogen-rich fluids, and reworked fossil plankton at ~6 km indicate sustained hydrothermal transport rather than deep marine deposition, demonstrating that fault systems can store and convey biological material over geologic timescales³⁰. Economic mineralization—including Cu–Ni sulfides (1.6–1.8 km), Au–Ag enrichment (9.5–11 km), and non-biogenic light hydrocarbons (C₁–C₄) within Archean gneisses—supports theories of deep hydrocarbon generation beyond methane alone³⁸. Vertical zoning of U, Th, and K, together with improved 3-D stress models, refined assessments of crustal deformation and tectonic hazards³⁹

Stable isotope data showed that whole-rock δ¹⁸O and δ²H values at ≤12 km mostly retain protolith signatures; only major fault zones record disturbances from interaction with metamorphic or Paleoproterozoic seawater–like fluids, confirming that deep-crustal fluid flow is highly channelized and that most cratonic crust behaves as a closed isotopic system⁴⁰. These insights have direct operational implications: critically stressed faults produce both inflow risks and unique sampling opportunities for deep fluids, ancient seawater, and hydrogen-generating reactions⁴¹; preserved isotope signatures validate surface analogues for deep-crust compositions; and drilling strategies require zonal isolation, careful pressure management, and integration of isotope thermometry with geothermometry⁴⁰.

Collectively, SG-3 overturned long-standing geophysical interpretations, exposed the limitations of seismic extrapolation without ground-truth drilling, and established technological benchmarks that influenced all subsequent ultradeep projects⁴². The early termination of SG-3 represented a lost opportunity to directly probe the Mohorovičić discontinuity and the mantle, and to advance understanding of global geohazards³⁵. Nevertheless, SG-3 remains a cornerstone of ultradeep geoscience, shaping the scientific priorities and safety strategies of modern >10 km drilling initiatives.

Germany’s KTB superdeep borehole: structure, discoveries, and legacy

Germany’s Continental Deep Drilling Program (KTB), carried out between 1987 and 1995 near Windischeschenbach in Bavaria, stands as one of Europe’s most ambitious scientific drilling initiatives. The program consisted of a 4 km pilot hole (KTB-VB) followed by the 9101 m main borehole (KTB-HB), which at the time became the world’s deepest scientifically accessible vertical drill site, with downhole temperatures reaching 265–270 °C⁴³. Notably, borehole deviation remained within 15 m over the upper 7.5 km, ensuring an exceptionally straight trajectory.

The pilot borehole (KTB-VB) yielded 3594 m of core and critical datasets on seismic velocities, stress regimes, and thermal gradients. The main borehole (KTB-HB), drilled between 1990 and 1994, extended these insights to 9 km depth. Large-scale hydraulic and stress experiments in the uncased section revealed high fluid pressures, fracture connectivity, and a transition from brittle to ductile deformation, with induced microseismicity ceasing near the present-day brittle–ductile transition (~9 km depth)^7,44.

Figure 3 presents a cross-section through the KTB site in eastern Bavaria. The borehole lies within the Variscan basement of the Bohemian Massif, an ancient collision zone later modified by post-orogenic extension. The surface is covered by thin wedges of Upper Cretaceous, Triassic, and Permo-Carboniferous sediments, underlain by crystalline units including gneiss, a variegated unit of interlayered metamorphic and igneous rocks, and amphibolite bodies at ~7–9 km depth. A granitic intrusion occurs to the northeast. The Franconian lineament forms a major tectonic boundary across the section, while the borehole path intersects gneiss, the variegated unit, and amphibolite, documenting the steep dips, shear zones, and lithological heterogeneity characteristic of the Variscan crust⁴⁵.

Fig. 3: Geological cross-section across the KTB borehole (KTB-HB) in eastern Bavaria, Germany, showing sedimentary cover (Upper Cretaceous, Triassic, Permo-Carboniferous) overlying Variscan basement units (gneiss, variegated unit, amphibolite) intruded by granite.

The borehole penetrates to 9.1 km, crossing the Franconian lineament (modified from ref. ⁴⁵).

KTB’s discoveries were profound. Downhole measurements confirmed a higher-than-expected geothermal gradient (28 °C/km) and unexpectedly high temperatures at depth, requiring revised models of heat flow that account for paleoclimate effects—particularly the influence of repeated Pleistocene glaciation and deglaciation cycles on near-surface temperature profiles, thermal diffusion, and lithospheric stress redistribution⁴⁶. Pumping tests between the pilot and main holes demonstrated hydraulic connectivity of fractures across kilometers, maintaining near-hydrostatic fluid pressures at mid-crustal levels⁴⁵. High-resolution seismic imaging identified steep reflectors linked to shear zones, while resistivity experiments revealed that graphite and sulfide phases control the electrical conductivity of the crust⁷.

The KTB project revealed that the deep continental crust is a dynamic, permeable, fluid-rich system in which seismic reflectors mark anisotropic, fluid-bearing shear zones and heat flow is influenced by advective fluids, thereby providing rare ground-truth constraints that fundamentally recalibrate models of crustal evolution and tectonics^7,31,45. KTB also laid key institutional foundations: it spurred the creation of the GFZ German Research Centre for Geosciences (1992) and the International Continental Scientific Drilling Program (ICDP) (1996), ensuring its enduring global impact.

Anadarko Basin: ultradeep burial diagenesis and hydrocarbon persistence

The Anadarko Basin of Oklahoma is among the deepest known Paleozoic basins on the North American craton and a globally important natural laboratory for understanding ultradeep burial diagenesis, hydrocarbon systems, and the limits of petroleum preservation³². Structurally, the basin is bounded by the Wichita Uplift to the south, the Ouachita Orogen to the east, and the Nemaha Uplift to the north (Fig. 4a). Its sedimentary pile reaches up to 12 km, making it one of the deepest cratonic depocenters in North America. Cross-sectional profiles (Fig. 4b) reveal a pronounced synclinal downwarp north of the Wichita Uplift that localized extraordinary thicknesses of Pennsylvanian clastics and older carbonate platforms. These successions preserve both the record of late Paleozoic orogenesis and the textural and geochemical responses of sandstones and carbonates to extreme burial and heating.

Fig. 4: Structural framework and ultradeep burial architecture of the Anadarko Basin, southern North American craton.

a Generalized geologic map of the Anadarko Basin, one of the deepest Paleozoic basins of the North American craton. Contours (in km) show the depth to the base of the Paleozoic sequence. Key structural features include the Wichita Uplift, Nemaha Uplift, Ouachita Orogen, and Mississippi Embayment. Locations of studied wells, including Bertha Rogers and Baden, are indicated. b Geologic cross-section (B–B′) of the Anadarko Basin, illustrating (1) the synclinal downwarp on the north flank of the Wichita Uplift and (2) thick accumulations of terrigenous clastics and carbonate strata in the adjacent basin (Modified from refs. ^32,47).

Two landmark boreholes—Lone Star No. 1 Baden (9.16 km, 1970) and Bertha Rogers No. 1 (9.58 km, 1974)—set world depth records before being surpassed by the Russia’s Kola Superdeep. Both encountered bottom-hole temperatures of 200–246 °C and pressures approaching 250 MPa, conditions near the lower greenschist-facies boundary. Yet, contrary to prevailing expectations, neither Simpson Group sandstones nor Hunton Group carbonates showed metamorphic mineral assemblages⁴⁷. The rocks retained diagenetic rather than metamorphic features: sandstones showed burial-related mechanical textures, carbonates displayed plastic deformation without forming new metamorphic minerals, and dolostones preserved up to 20% porosity at 7.9 km, indicating superior compaction resistance^32,47.

These observations fundamentally reshaped models of deep burial diagenesis. They demonstrated that temperature and depth alone—even at conditions near regional metamorphism—do not necessarily drive mineralogical transformation in sedimentary rocks. Instead, mechanical compaction, stress-driven recrystallization, and pressure-solution are the dominant controls on fabric evolution. This insight has major implications for reservoir quality prediction in ultradeep basins: dolostones can retain pore networks well into the lower crustal pressure–temperature regime, while limestones and sandstones typically become tight but not metamorphosed.

Perhaps the most striking discoveries came from the Bertha Rogers well, where geochemical analyses identified significant concentrations of liquid hydrocarbons—including C₁₂⁺ paraffins, condensates, and anomalous bitumen—at depths and temperatures where only methane or graphite were expected⁴⁸. These results show that hydrocarbons can be generated and preserved at far greater depths and temperatures than expected, and that organic-matter type and depositional setting exert stronger control on hydrocarbon composition than maturity alone, prompting a reassessment of deep petroleum system models.

Technologically, the Baden and Bertha Rogers wells were pioneering feats in super-deep drilling. They employed ultra-heavy casing strings exceeding 150 MPa, advanced high-pressure mud circulation systems, and record-setting land rigs⁴⁹. These innovations expanded the practical depth limits of rotary drilling and provided critical engineering data later applied to global ultradeep exploration.

SDTK-1, Tarim Basin—scientific significance & discoveries

The SDTK-1 well in the Tarim Basin NW China marks a global milestone in ultradeep drilling, showcasing both technical achievement and scientific impact. Extending to 10,910 m with bottom-hole conditions of ~220 °C and >145 MPa, it was completed with the world’s first 12-km automated land rig, confirming China’s capability to penetrate lower crustal sedimentary successions under extreme HP–HT regimes^50,51. The well successfully crossed 12 stratigraphic units, providing rare direct data on source rock continuity, reservoir quality, and petroleum system functioning at depths previously regarded as beyond hydrocarbon limits (Fig. 5).

Fig. 5: Geological framework of the Tarim Basin (NW China) highlighting the Cambrian petroleum system and the SDTK1 ultradeep borehole.

a Structural map showing uplifts, sags, hydrocarbon fields (red = oil, yellow = gas), and deep wells, with SDTK1 in the eastern-central basin (Modified from ref. ⁵¹). b Cross-section (A–A′) across SDTK1 illustrating Cenozoic–Mesozoic cover above a thick Paleozoic succession, with Cambrian units (Єᵧ, Єₓ, Єw, Є_s, Єₐ) deformed by NE-verging folds and faults. Reservoirs, sources, and seals are marked in red, brown, and gray (Modified from ref. ¹⁴). c SDTK1 well log showing gamma-ray, resistivity, gas logging, lithologies, and petroleum system elements. Dolomites (oolitic, algal, crystalline, argillaceous) dominate, interbedded with gypsum, marl, and mudstone. Results demonstrate thick Cambrian source rocks overlain by dolomitic reservoirs and evaporitic seals, confirming effective ultradeep petroleum systems (Modified from refs. ^15,51).

The SDTK-1 well has delivered transformative insights into ultradeep petroleum systems, most notably through the identification of multiple reservoir–seal assemblages beneath thick Cambrian evaporites. The Awati gypsum capping back-reef dolomites forms a proven trap configuration, while the Wusonggeer–Xiaoerbulake carbonate succession provides another effective reservoir–seal system. Extending even further, the Sinian Qigebulake dolomite paired with Lower Cambrian shale opens exploration opportunities into pre-Cambrian intervals, underscoring the multi-tiered architecture of Tarim’s ultradeep petroleum systems^14,51. These results demonstrate that evaporite and gypsum seals retain integrity under pressures exceeding 145 MPa, ensuring the secure preservation of hydrocarbons at depths once thought geologically prohibitive.

In August 2025, the SDTK-1 well achieved a landmark breakthrough as the first global demonstration of commercial-scale hydrocarbon flow from depths exceeding 10 km, producing tens of thousands of cubic meters of dry natural gas per day from Cambrian strata. Gas composition data (Table 2) reveal a clear depth-controlled transition: Upper Ordovician (O₁p, 8560–8846 m) gases are wet and mixed, with ~73–75% CH₄, 8–9% C₂, ~6% C₃, up to 3–5% C₄s, measurable pentanes/C₆, and CO₂ as high as 6–9%—a liquid-rich condensate system. By 10,030 m (Cambrian ∈₂s), methane increases to ~79% as C₂–C₃ drop (C₂ ~5.7%, C₃ ~2.8%, C₄ ≤ 1.3%), though CO₂ remains ~9%, marking a shift toward drier but still moderately wet gas. Below ~10,770 m (Lower Cambrian ∈₁x), the system becomes ultra-dry methane-dominant (C₁ 92–100%, C₂ ≤ 0.14%, C₃ ≤ 0.01%, negligible higher hydrocarbons, CO₂ < 1%) (Table 2), defining a dry-gas end-member at depths previously regarded as beyond petroleum viability.

Table 2 Natural gas composition and carbon isotopic composition of various strata in the SDTK-1 well (Modified from ref. ¹⁵)

Carbon isotope and gas composition data from the SDTK-1 well document a clear oil–gas transition with increasing depth (Table 2). In Ordovician reservoirs (O₁–₂y and O₁p; ~8.2–8.8 km), methane dominates but ethane and propane remain abundant (C₁ ~65–75%; C₂–C₃ ~10–15%), consistent with wet gas. At depths approaching 10 km and into the Lower Cambrian, gas becomes progressively drier, with methane exceeding 96–99% and C₂⁺ hydrocarbons reduced to trace levels or absent. Methane carbon isotopes (δ¹³C₁), available only for the deepest Cambrian intervals, show strong enrichment, shifting from values typical of mature thermogenic gas (–47‰ to –49.8‰ in shallower systems) to much heavier compositions (–34 to –32‰) below ~10.7 km¹⁵ (Table 2).

The coupled drying of gas and enrichment of δ¹³C₁ indicate advanced thermal cracking of liquid hydrocarbons to secondary methane under increasingly closed-system conditions. In upper intervals, mild isotopic reversal (δ¹³C₁ > δ¹³C₂,₃) is consistent with wet gas, whereas this pattern disappears in the deepest dry-gas zones, reflecting late-stage methane generation after exhaustion of C₂⁺ species and challenging conventional methane isotope–maturity models. As temperature and burial increase, residual liquids and precipitated pyrobitumen become progressively ¹³C-depleted, causing newly generated methane to inherit heavier isotopic signatures. These observations demonstrate that stable dry gas systems can persist well beyond 10 km, requiring reassessment of petroleum system limits and supporting large ultradeep gas resources in the northern and central Tarim Basin (~6.6 × 10¹² m³ across ~120,000 km²)⁵².

Geochemical analysis established the Lower Cambrian Yuertusi Formation as a key source rock system (Fig. 6). In SDTK-1, 52 m of effective source rock were logged, dominated by dark mudstone and calcareous limestone intervals, including an 18.5 m subunit with TOC values up to 6.0% and strong hydrocarbon potential. Despite >10 km burial, organic matter is overmature yet retains geochemical signatures of hydrocarbon generation, demonstrating that kerogen-rich intervals remain functional as source rocks under extreme burial^12,53. Comparison with well LT1 (~6.5 km) shows consistent TOC enrichment, high Rock-Eval S₂ yields, and trace-metal signatures (Mo, U, V, Ni/Co ratios), reflecting euxinic depositional settings favorable for long-term preservation of organic matter. Collectively, these findings confirm the regional persistence of Cambrian source rocks across Tarim, validating their role in sustaining ultradeep petroleum systems.

Fig. 6: Geochemical and lithological profiles of Cambrian source rocks in the Tarim Basin from wells SDTK-1 (left) and LT1 (right). SDTK-1 (>10 km) shows dolomitic micrite and mudstones with residual TOC, S₂, and HI values, indicating overmature but detectable hydrocarbon potential.

LT1 (~6.4–6.6 km) penetrates the Yuertusi Formation with higher TOC, strong S₂ peaks, and trace-element enrichment, reflecting euxinic deposition and excellent source potential. Together, the wells confirm the Yuertusi Formation as a regionally extensive Cambrian source rock system, from preserved kerogen at shallower depths to geochemical signatures retained at >10 km burial.

Reservoir development is equally striking. Core samples from the Cambrian Shayilike Formation (Fig. 7) document 96 m of effective dolomite reservoirs with preserved porosity at >10 km. Intercrystalline pores, dissolution vugs, and fractures form significant storage space, supported by evidence of paleo-hydrocarbon charging such as pyrobitumen residues. Textures include algal-bound and fine-crystalline dolomites, characteristic of platform-margin shoal environments, where early diagenetic dissolution and dolomite brittleness enhanced porosity retention under HP–HT stress^15,54. While pressure-solution seams and argillaceous laminae locally reduced porosity, the overall carbonate system preserved effective reservoir quality at depths previously considered beyond productive limits. This redefines the lower boundary for effective carbonate reservoirs and illustrates the role of early diagenetic processes in sustaining ultradeep porosity.

Fig. 7: Core photographs of the Shayilike Formation from the SDTK-1 well (after ref. ¹⁵), illustrating reservoir characteristics at different depths.

a 10,045.09–10,045.27 m, algal-laminated dolomite with abundant bedding-parallel dissolution pores; b 10,043.02–10,043.19 m, centimeter-scale caverns up to 1.5 cm and dissolution pores along bedding planes; c 10,045.56–10,045.75 m, fractured dolomite with reticular fractures intersecting to form brecciated cavities; d 10,045.81 m, pore spaces filled with black dry bitumen indicating ancient hydrocarbon charging; e 10,043.19–10,043.30 m, high-angle vertical fractures enlarged by dissolution forming favorable storage space; f 10,042.65–10,042.98 m, granular dolomite with gray-black siliceous bands cut by dissolution-enlarged fractures; g 10,183.61–10,183.84 m, micritic dolomite with pressure-solution seams and argillaceous laminae showing poor porosity; and h 10,186.57–10,186.62 m, algal-laminated dolomite with wavy lamination typical of shallow-marine algal mats.

The broader scientific implications are profound. SDTK-1 and X-1 confirm that petroleum systems can persist well below 10 km, contrary to earlier assumptions that high maturity, compaction, and thermal stress would obliterate source, reservoir, and seal integrity. Instead, evidence from geochemistry, petrophysics, and core analyses proves that viable petroleum systems—complete with effective source, reservoir, and seal—exist under HP–HT conditions once considered prohibitive¹³. These results demand revision of global petroleum system models, particularly in cratonic basins with thick Paleozoic successions, where analogous plays may host substantial ultradeep resources.

Ultradeep drilling as ground truth for lithosphere dynamics: structure, heat flow, chemo-mechanical feedbacks, stress, and seismicity

Ultradeep boreholes transform geophysical anomalies into physical explanations by retrieving rocks, fluids, and in-situ logs from the lower crust and upper mantle. These records ground-truth seismic and magnetotelluric images, sharpen models of heat flow and lithospheric rheology, and expose fluid-driven chemo-mechanical feedbacks that control stress, permeability, and seismicity.

Ground-truthing crust–mantle structure, fluids, and tectonic processes

Understanding Earth’s crustal structure requires linking surface observations with the processes that operate deep underground. The mantle is compositionally heterogeneous and dynamically active, with seismic discontinuities at ~410 and ~660 km depth reflecting mineral phase transitions such as olivine to wadsleyite and ringwoodite to bridgmanite plus magnesiowüstite⁵. These transitions regulate mantle convection, slab dynamics, and plume ascent, coupling deep Earth processes to plate tectonics at the surface. Geochemical signatures in basalts further reveal distinct mantle reservoirs that preserve billions of years of crust–mantle recycling. While ultradeep drilling cannot directly access such depths, it provides critical ground truth for calibrating seismic and geochemical models by sampling the lower crust and uppermost mantle, offering unique constraints on heat flow, fluid migration, and mass transfer⁵.

The rationale for drilling beyond 10 km lies in the limitations of geophysics. Seismic and magnetotelluric imaging provide indirect contrasts in velocity and conductivity, but these cannot distinguish between lithologic changes, shear zones, or fluid-filled fractures. Boreholes overcome this ambiguity by recovering rocks, fluids, and downhole logs, allowing one-to-one comparison between signals and physical causes^7,39. This link has repeatedly overturned assumptions. For example, strong reflectors once thought to mark granite–gabbro contacts were shown instead to represent metamorphic layering or deconsolidated fault gouge⁵⁵. Such corrections improve regional interpretations of crustal architecture and tectonic history.

Major ultradeep projects have transformed understanding of crustal stratigraphy. Russia’s SG-3 and Germany’s KTB traversed depths traditionally associated with the Conrad discontinuity. Rather than simple granite–basalt layering, both revealed thick successions of metamorphic rocks cut by shear zones and fluidized fracture corridors^7,30. High reflectivity often arose from graphite- and sulfide-rich horizons, which also explained high conductivities without invoking partial melt. These findings emphasized that many geophysical anomalies reflect mechanical fabrics or mineralogical concentrations rather than compositional boundaries.

Core-scale studies from these boreholes anchor the brittle–ductile transition. Microstructural changes—cataclasis giving way to crystal-plastic flow and localized mylonitization—show how strength decays with depth. Thermobarometry constrains pressure–temperature–time histories of orogenic crust, shedding light on thickening, re-equilibration, and exhumation⁵⁶. Isotopic and geochemical tracers document fluid–rock interactions that drive metasomatism, control metamorphic reactions, and concentrate ore-forming metals⁵⁷. These results establish how mechanical and chemical processes couple to govern crustal rheology.

SG-3’s purpose was to interrogate the continental crust directly, and it overturned entrenched models. Drilling through Archean basement on the Kola Peninsula disproved the “granite–basalt” model by showing that velocity steps were caused by metamorphic rocks and structural fabrics at shallow depth⁵⁸. Even more striking was the discovery of highly mineralized fluids in fractured crystalline rocks at depths where a sealed, impermeable crust had been expected. These fluids, interpreted as mineral-bound water released under extreme pressure and temperature, forced a rethinking of seismic interpretations and revealed new roles for deep brines in mass transfer, deformation, and ore concentration³⁹.

Germany’s KTB yielded complementary lessons. Its pilot and main holes revealed elevated geothermal gradients, widespread hydraulic connectivity, and a brittle–ductile transition marked by waning microseismicity and localized viscous shear⁷. Conductivity proved to be controlled by graphite and sulfide phases, while steep seismic reflectors were tied to shear zones. Together, SG-3 and KTB established that the deep crust is not a static, impermeable block but a stressed, permeable medium where fluids circulate episodically, altering both its mechanical properties and geophysical signatures.

Perhaps the most consequential realization from SG-3 and KTB is that porosity, permeability, and fluids persist far deeper than expected. This explains seismic anisotropy, high conductivity, and anomalous heat flow in the crust: aligned microcracks, graphite films, saline brines, and advective circulation provide mechanisms inconsistent with “dry crust” assumptions. Fluids in turn weaken rocks, localize strain, and redistribute metals, feeding back into cycles of rifting, collision, magmatism, and erosion that build and modify continents⁵⁹.

Ultradeep drilling is indispensable for understanding Earth’s interior. By sampling across seismic discontinuities, logging stress and pore pressure, and retrieving rocks and fluids at in situ conditions, boreholes turn geophysical contrasts into physical causes. They show that crustal “boundaries” are often shear zones or fluid pathways rather than sharp compositional steps, that tectonic models are best tested where drilling intersects major structures³⁰. These insights refine crustal rheology, fluid–rock interaction, and heat transport models, while improving hazard assessment by revealing how stress and fluids interact at depth^56,57. Ultimately, ultradeep drilling integrates direct sampling and measurement with geophysics to build a coherent picture of continental evolution—from mantle phase transitions driving convection to Archean basement that stays permeable, and from shear zones controlling conductivity to ultra-high-pressure belts recording deep subduction. It provides the essential ground truth, making inaccessible processes testable and anchoring tectonic theory in the reality of rocks, fluids, and structures at depth.

Geothermal gradients, heat flow, and fluid circulation in the deep continental crust

Ultradeep drilling provides the only direct means of measuring the thermal structure of the continental lithosphere, transforming how scientists understand heat transfer, crustal strength, and fluid circulation at depth. Near-surface geothermal gradients typically average ~25–30 °C/km, but extrapolating these values to the lower crust is unreliable because shallow surveys miss vertical changes in thermal conductivity, transient heat regimes, and the effects of circulating fluids. Deep boreholes yield continuous, high-resolution temperature profiles across the brittle–ductile transition, revealing that geothermal gradients are not linear but curved and variable^7,39. These observations recalibrate global models of lithospheric geotherms and overturn the long-held assumption of simple, conductive heat transfer.

A global synthesis drawing on more than 30,000 core measurements and over 100 deep temperature logs from 15 deep boreholes showed that, although conduction accounts for 70–100% of total heat transfer, its relative contribution changes markedly with depth⁶⁰. In many cases, shallow sections (0–4 km) showed heat flow up to twice that of deeper levels due to conductivity contrasts, and ultradeep wells recorded 30–130% higher heat flow than shallow estimates—evidence that traditional averaging systematically underestimates terrestrial heat flow. Such calibrated geotherms are vital: they redefine the brittle–ductile transition, revise predictions of crustal strength and earthquake depth limits⁵³, and underpin accurate petroleum system modeling, where misplacing oil and gas windows can lead to over- or underestimation of hydrocarbon potential. For unconventional resources, realistic thermal histories determine the viability of overmature shale gas, residual oil, and hydrogen generation. They also guide geothermal energy strategies, pinpointing fracture-controlled heat anomalies and deep hot-dry-rock reservoirs invisible to shallow exploration⁶¹.

Equally transformative is the realization that the deep crystalline crust is neither dry nor impermeable. Geochemical evidence reveals meteoric recharge, metamorphic devolatilization, and mantle inputs, proving that fluids persist as active agents redistributing heat^62,63. These fluids catalyze metamorphic reactions, mobilize ore metals, alter redox states controlling hydrocarbon and hydrogen stability⁵⁷, and strongly influence tectonic behavior. At KTB, stress measurements linked pore-pressure variations to seismicity and stress orientation, demonstrating how permeability and pressurized fluids modulate earthquake nucleation⁶⁴.

Collectively, these findings reveal a thermally complex, fluid-active, and mechanically responsive deep crust. Ultradeep wells integrate temperature logs, geochemical tracers, and in-situ sampling to replace abstract models with direct evidence of heat flow, permeability, and tectonic coupling. They show that crustal “boundaries” are often shear zones or fluid pathways rather than simple compositional steps, and that the lithosphere’s strength and seismic hazard potential must be understood in this context.

The Songliao Basin continental drilling program illustrates these advances. SK-2 recovered a continuous, centimeter-scale Cretaceous core preserving climate variability and mass extinction signals tied to Deccan volcanism and the Chicxulub impact. Geophysical imaging revealed previously unrecognized Triassic strata and refined the basin–mountain tectonic framework. Downhole temperatures reached ≥150 °C at ~4.4 km, corresponding to ~84 mW/m² heat flow—a clear indicator of hot-dry-rock geothermal potential⁸. At ~7 km, drilling intersected a pre-Mesozoic metamorphic complex of mylonites, cataclasites, and fractured lithologies, documenting both regional and dynamic metamorphism⁶⁵. Together, these results underscore how ultradeep drilling anchors thermal models, refines tectonic reconstructions, and expands frontiers for energy development, critical resource assessment, and Earth system science.

Stress–fluid–chemical–seismic coupling in the deep crust

Extreme pressures and temperatures below ~9 km drive hydration, carbonation, and thermal cracking reactions in mafic and ultramafic rocks. Serpentinization and carbonation can increase solid volume by 40–50%, generating crystallization pressures that exceed 300 MPa—far greater than typical differential stresses in the mid-crust⁵⁷. These pressures can fracture intact rocks, reopen sealed microcracks, and sustain fluid pathways over geological timescales. Cores from SG-3 and SDTK-1 contain dense vein networks of serpentine, carbonate, and sulfides, as well as shattered feldspar and partially altered olivine, directly confirming that chemical reactions drive fracturing in situ⁶⁶. These interactions are strongly temperature dependent: serpentinization-related failure peaks near ~260 °C, carbonation near ~185 °C, and both occur within the temperature ranges recorded in >9–10 km boreholes⁶⁷. Because viscous creep cannot relax stresses fast enough under these conditions, brittle failure dominates. Even relatively small drilling disturbances—such as mud losses, water inflow, or CO₂-bearing fluids—can initiate cracking and microseismicity. Fluid-injection tests at KTB demonstrated that pore-pressure perturbations of less than 1 MPa were sufficient to trigger hundreds of microearthquakes⁶³.

Hydrocarbon systems reveal similar stress–fluid–reaction coupling. In SDTK-1, a sharp transition from viscous oil to methane at ~9.5 km reflects thermal cracking, pyrobitumen formation, and chemically induced overpressure that modify fracture stability. Dolomite reservoirs maintain fracture porosity under >145 MPa and >220 °C due to early diagenetic fabrics and brittle reactivation, whereas shales deform by pressure solution and tend to seal¹⁵. These processes explain how thermogenic methane remains stable and producible far below traditional limits for petroleum preservation^12,68.

Direct downhole stress measurements have reshaped understanding of lithospheric strength. Compiled data indicate that the brittle crust maintains Coulomb failure equilibrium down to ~10–12 km depth, with strength governed by Byerlee’s friction (μ ≈ 0.6–1.0)⁵⁶. Pore pressures remain closer to hydrostatic than lithostatic, implying slow strain rates consistent with geodetic constraints and contradicting earlier models that invoked near-lithostatic overpressures. Drilling and laboratory studies from Kola and KTB reveal that permeability evolves nonlinearly under stress, temperature, and chemical influence. Experimental results show that permeability decreases with increasing confining pressure but becomes highly anisotropic, with flow parallel to foliation greatly exceeding perpendicular flow⁶. Heating initially seals microcracks but reopens them between ~200–500 °C, producing permeability inversions that allow episodic migration of brines and gases at depth. Geological modeling indicates that prograde metamorphism maintains sufficiently high permeability for widespread fluid flow up to ~600 °C, whereas retrograde metamorphism at ~180–320 °C collapses matrix permeability and diverts fluids into faults and shear zones⁶⁹. These conditions explain why ultradeep fluids are commonly detected in discrete, fault-bounded reservoirs rather than uniformly distributed across the crust.

Deep stress profiles from KTB⁶⁴ show a stable maximum horizontal stress orientation (~N160°E) and magnitudes consistent with frictional failure. The crust remains critically stressed, with narrow mud-weight windows during drilling. Small deviations in pore pressure can promote slip on optimally oriented faults, requiring real-time geomechanical monitoring and anisotropy-aware well design to maintain wellbore stability.

Ultradeep drilling also transforms understanding of earthquake processes. Boreholes that penetrate the brittle–ductile transition provide direct observations of the depth range where most large earthquakes initiate. Downhole seismic monitoring yields high-fidelity waveforms, constrains seismic velocities and attenuation⁷⁰, and improves earthquake hazard models. The KTB project demonstrated exceptional sensitivity of deep crustal rocks to fluid perturbations: fluid injection at ~9 km and ~260 °C triggered more than 400 microearthquakes with pore-pressure changes <1 MPa⁶³. Most events displayed strike-slip mechanisms consistent with the regional stress field, confirming that brittle failure persists even near the brittle–ductile transition.

Fault-zone drilling further clarifies rupture mechanics. The San Andreas Fault Observatory at Depth (SAFOD), although only ~3 km deep, showed that major faults can accommodate slip within ultrathin (<3 m), clay-rich gouge zones that possess extremely low friction²⁵. These results imply that deeper major faults may host similarly weak, fluid-rich zones, presenting both drilling hazards and valuable scientific targets for future >10 km projects.

Deep shield drilling, including SG-3 and Scandinavian boreholes, confirms that seismic reflectivity commonly originates from graphite-, sulfide-, and fluid-rich shear zones rather than simple lithologic layering^71,72. These structures also explain deep conductive anomalies imaged beneath mountain belts^7,58.

Fluid-mediated weakening is widespread in the deep crust. Highly mineralized brines and fluid inclusions recovered from >9 km depth in Kola and KTB^7,39 demonstrate that deep fluids reduce effective normal stress, catalyze mineral reactions, and localize deformation. These processes are crucial for understanding natural seismicity and predicting induced seismicity associated with CO₂ sequestration, hydrogen storage, geothermal systems, and ultradeep petroleum operations.

Insights from KTB and Kola provide essential design principles for future ultradeep drilling. Both projects showed that wellbore stability at >8–10 km is controlled by critically stressed rock, strong permeability anisotropy, and thermally activated crack networks. These findings indicate that future >10 km wells must incorporate continuous downhole stress imaging, real-time pore-pressure tracking, and thermal–hydraulic modeling to anticipate permeability inversions and fault slip. KTB in particular demonstrated that small pressure oscillations during circulation or injection can trigger slip on optimally oriented faults, underscoring the need for pressure-managed drilling, narrow mud-weight windows, and high-frequency geophysical monitoring. Meanwhile, Kola revealed that chemically reactive zones—especially where hydration or carbonation is active—can rapidly destabilize wellbores unless casing programs and drilling mud chemistries are adapted to evolving rock–fluid reactions. Together, these lessons form a benchmark for designing safer, more resilient ultradeep wells and the next generation of downhole monitoring technologies.

The ultimate depth of drilling and resource potential

Ultradeep drilling has opened an entirely new domain for resource science, revealing that the Earth’s crust remains chemically, thermally, and hydrologically active well beyond 9–10 km. Once considered inaccessible, these depths now represent frontiers for hydrocarbons, natural hydrogen, mineral resources, and geothermal energy.

Frontier of hydrocarbon systems

Ultradeep drilling has become a critical frontier in global hydrocarbon exploration as conventional reserves decline and mature fields reach depletion. Traditional petroleum system models define the oil window at ~60–160 °C and the gas window to ~200 °C⁷³, but drilling beyond 10 km shows that hydrocarbons can persist far deeper under P–T conditions exceeding 200–250 °C and 200 MPa. Results from China’s SDTK-1 and X-1 wells confirmed that carbonaceous matter and light hydrocarbons survive where classical models predict destruction, demonstrating that the deep crystalline crust remains chemically and hydrologically active and that petroleum systems extend well beyond previous depth limits.

The Tarim Basin, long known for its supergiant Ordovician carbonate fields sourced from the Cambrian Yurtus Formation¹², has provided key evidence. The Luntan-1 well at 8,260 m flowed oil commercially from source rocks with TOC > 10% and trace-metal enrichment (Mo, U, V)⁵³. More recently, SDTK-1 in Tarim and X-1 in Junggar extended exploration past 10 km and produced commercial hydrocarbons. Gas from SDTK-1 (Table 2) is consistent with oil-cracking methane, proving that petroleum generation and preservation continue into the >10 km domain and redefining the lower boundary of viable petroleum systems.

Laboratory experiments complement these findings. Pyrolysis and basin modeling show that overmature kerogen and bitumen can keep generating methane and hydrogen under extreme conditions, while heavier hydrocarbons crack to pyrobitumen or crystalline carbon^74,75. These reactions are strongly modified by mineral catalysis, confining pressure, and fluid–rock interactions, meaning hydrocarbon stability is governed not by fixed thermal limits but by dynamic, water- and pressure-dependent processes. Hydrous pyrolysis studies show that high pore-water pressures accelerate some reactions while delaying others⁷⁶; at depths >9 km, pore pressures routinely exceed 100 MPa, and water remains reactive at high temperature. Consequently, common maturity indicators such as vitrinite reflectance can misjudge hydrocarbon windows if water–rock reactions and pressure effects are ignored⁴⁸. These insights explain why basins considered overmature may still host methane- or hydrogen-rich gases and support a petroleum–hydrogen continuum that merges deep methane preservation with natural hydrogen generation⁷⁷.

Ultradeep drilling has redefined models of thermal structure and fluid circulation, key to hydrocarbon preservation. Borehole data reveal nonlinear geothermal gradients and higher heat flow than shallow surveys predict, forcing recalibration of basin models and hydrocarbon windows⁶⁰. Recent >10 km drilling allows clearer separation between direct observations and process-based interpretations regarding why hydrocarbons and natural hydrogen persist under extreme P–T conditions. Direct measurements from SDTK-1 and X-1 show in situ temperatures of >220–240 °C, pressures exceeding 145–160 MPa, methane-rich compositions, and the presence of water-rich, mineralized fluids—all unequivocal physical observations. Interpretations based on laboratory and numerical models suggest that high confining pressure suppresses the breakdown of methane and stabilizes C–H bonds; that water-rock reactions (serpentinization, radiolysis, and Fe-rich mineral alteration) continuously generate H₂; and that mineral surfaces catalyze hydrocarbon cracking while simultaneously buffering redox conditions⁷⁷. These mechanisms collectively explain persistence of thermogenic methane and active generation of hydrogen at depths where classical models predict hydrocarbon destruction. The distinction between measured in situ data and inferred mechanistic controls is critical, as it clarifies that the existence of deep hydrocarbons and H₂ is an observational fact, whereas the processes governing their stability are model-supported but not yet fully resolved.

China’s ultradeep programs have pushed this frontier further, revealing multiple vertically stacked petroleum systems below 10 km, including Jurassic–Triassic coal-bearing systems, Cambrian subsalt–Sinian dolomite systems, and Carboniferous–Permian foreland plays¹⁴. Across Tarim, Sichuan, Junggar, Qaidam, and Ordos, systematic studies identify 11 major ultradeep prospective domains with trillion-cubic-meter potential (Table 3). The Tarim Basin dominates with Lower Cambrian subsalt dolomite–evaporite systems (18,000 km², trillion-m³ gas potential), Ordovician fracture–cavity reservoirs with km-scale oil columns, and Carboniferous–Permian carbonates in long-mature platforms now viewed as strategic replacements. The Sichuan Basin offers Sinian–Cambrian paleo-uplifts, Permian–Triassic reef-shoal complexes, and Dengying–Longwangmiao systems with up to 2 trillion m³ gas potential. The Ordos Basin hosts pre-salt Ordovician plays with low exploration maturity. In the Junggar Basin, Jurassic coal measures and Permian Lucaogou sources could form trillion-m³ gas provinces, while Permian Fengcheng systems offer large-scale potential. The Qaidam Basin shows Jurassic–Paleogene reservoirs with deep, porous Jurassic coals favorable for ultradeep gas field development.

Table 3 Favorable exploration regions, resource potential, and exploration targets for ultradeep (>10 km) major gas field groups in China

Together, these discoveries push the depth limit of viable petroleum systems well beyond 10 km, show that hydrocarbons and hydrogen can coexist in the deep crust, and redefine global exploration strategy. Ultradeep drilling turns once-inaccessible “blind zones” into a new energy frontier, linking petroleum geoscience with emerging plays in natural hydrogen, deep methane, and high-pressure reservoirs, while providing essential ground truth for thermal, geochemical, and tectonic models of Earth’s interior.

Emergence of natural hydrogen resources

Early ultradeep drilling projects, including Russia’s Kola SG-3 and Germany’s KTB, established both the technical feasibility of drilling beyond 9–12 km and the geological reality of deep hydrogen, recovering H₂-rich fluids from crystalline basement rocks^7,30. These pioneering wells provided the first borehole-scale confirmation that hydrogen can be generated, migrate, and persist in the deep crust. Modern ultradeep programs build on this foundation by deploying high-temperature drilling fluids, reinforced casing and mud systems, and advanced downhole testing tools that enable in situ measurement of hydrogen generation, transport, trapping, and sealing—constraints that cannot be resolved from surface observations alone. Recent ultradeep drilling campaigns, particularly those exceeding 9–10 km, now provide the pressure–temperature and lithologic conditions required to directly test competing hydrogen-generation models at depths previously accessible only through inference. Without such borehole validation, hypotheses invoking thermogenic hydrogen from overmature organics^74,77, mantle-derived fluxes^78,79, or a large global hydrogen endowment⁸⁰ remain weakly constrained.

A critical advance enabled by ultradeep wells is the ability to discriminate among multiple hydrogen-generation pathways operating under distinct lithologic and thermal regimes. Serpentinization of olivine-rich ultramafic rocks generates H₂ through Fe²⁺ oxidation to magnetite, sustaining fluxes over kyr–Myr timescales⁸¹. Radiolysis of water driven by U–Th–K decay dominates in granitic and felsic crust, producing hydrogen continuously over hundreds of millions of years⁷⁸. A third pathway—thermogenic hydrogen—occurs when kerogen reaches extreme maturity (Ro >3.5–5%), releasing H₂ during advanced aromatization and concurrent methane cracking⁷⁷. Laboratory pyrolysis and basin modeling predict that overmature shales may remain significant hydrogen sources⁷⁶, but only ultradeep boreholes can verify whether these yields persist under natural pressure–temperature conditions. Helium serves as a robust co-tracer: its mobility, isotopic composition (³He/⁴He), and association with H₂ help fingerprint deep sources and migration pathways⁷⁸.

Global borehole observations substantiate these mechanisms. SG-3 recovered fluids containing up to 20.8% H₂ with trace He at depths >11 km in Archean gneiss⁸², while the KTB well detected up to 8.6% H₂ and 0.08% He at ~9.1 km⁸³. In China, the Songke-2 scientific borehole in the Songliao Basin logged hydrogen anomalies up to 2.39% at depths of 2.8–6.8 km, interpreted as products of crust–mantle interaction and deep fault-controlled migration^17,84. Mining legacy wells in France further identified 26 H₂-bearing intervals, including concentrations up to 52 vol% in the Dogger aquifer of the Paris Basin along deep-seated feeder faults⁷⁹. Although most reported H₂ occurrences remain shallower than 10 km, new ultradeep drilling programs—particularly in stable cratonic settings—are now positioned to test whether hydrogen systems persist, accumulate, or are consumed under extreme burial conditions. Collectively, these observations indicate that hydrogen systems can adopt petroleum-style architectures, with identifiable sources, migration pathways, reservoirs, and seals preserved beneath evaporite or clay-rich barriers. While global in-place hydrogen resources may reach ~5.6 × 10⁶ Mt⁸⁰, uncertainties surrounding generation rates, consumption by redox reactions, and long-term retention underscore the need for ultradeep borehole constraints.

Ultradeep drilling further enables high-resolution sampling of ultramafic, granitic, and overmature sedimentary intervals, allowing real-time evaluation of serpentinization kinetics, radiolytic yields, and thermogenic gas compositions. Noble gas and stable isotope measurements (³He, ²⁰Ne, δ²H) refine source attribution and leakage–retention balances^78,79, while microstructural imaging (SEM, FIB-SEM) documents porosity and fracture evolution in minerals and organic matter, improving predictions of hydrogen storage capacity. These datasets are increasingly achievable only in the context of modern ultradeep wells, which extend hydrogen research from indirect geochemical inference to direct process-based validation. From an engineering perspective, H₂-rich drilling environments pose challenges including metal embrittlement, corrosion, blowout risk, and mud–gas management, necessitating specialized alloys, continuous monitoring, and real-time geomechanical control².

Future exploration focuses on two end-member play concepts. The first involves ultradeep organic “hydrogen kitchens,” where overmature marine shales generate thermogenic H₂ contemporaneously with late-stage methane cracking, potentially forming mixed gas accumulations^74,77. The second targets basement- and fault-controlled systems, in which serpentinization, radiolysis, and mantle-derived fluxes supply hydrogen to traps above ultramafic or radiogenic crystalline rocks^78,79. Natural hydrogen commonly co-occurs with deep hydrocarbons not because of a shared mantle source, but because independently generated fluids exploit the same fault networks, lithologic contrasts, and trapping geometries. In this framework, ultradeep drilling provides the critical means to link hydrogen occurrence with deep fluid circulation, structural connectivity, and thermal evolution, positioning H₂ not only as a potential energy resource but also as a sensitive tracer of deep Earth processes.

Frontier of mineral resource exploration

Ultradeep drilling has opened a transformative frontier in mineral resource exploration, providing direct access to P–T regimes once inferred only from geophysics and surface analogues. Traditional methods capture limited signals of mineralizing systems, but boreholes beyond 9–10 km deliver cores, fluids, and in situ geophysical logs, revealing how extreme P–T conditions control mineral stability, metal transport, and ore concentration. One of the most important scientific advances is the deep carbon cycle and its mineralogical expression. The Russia’s Kola SG-3 intersected graphite-rich zones, demonstrating that carbon is actively transformed under extreme conditions and may connect to diamond precursors⁸⁵. Coupled processes such as serpentinization and carbonation in ultramafic rocks illustrate the link between energy and mineral systems: they generate hydrogen and magnetite, precipitate carbonates, and create stresses exceeding 300 MPa that fracture rocks, maintaining long-term permeability^57,86. These chemo-mechanical feedbacks position serpentinite systems as natural ore-forming reactors and hydrogen factories, while also revealing critical geohazards such as mechanical instability and fluid overpressure—factors essential to anticipate in safe ultradeep drilling.

Mineralogical studies also clarify how phase transitions control deep resource systems. Common serpentine polymorphs—lizardite and chrysotile (Mg₃Si₂O₅(OH)₄)—form under contrasting conditions: lizardite during olivine hydration accompanied by magnetite formation and H₂ release, and chrysotile in fluid-rich, stress-relaxed zones⁸⁶. At >300 °C, both transform into antigorite and brucite, marking a transition where hydrogen release, fault weakening, and stress redistribution converge. Ultradeep boreholes allow these processes to be observed under true in situ conditions, quantifying hydrogen flux, porosity evolution, and their role in ore metal transport, seismic weakening, and crustal fluid pathways. Such insight is essential both for understanding metallogenic systems and for predicting geomechanical risk during deep exploration.

From an applied perspective, ultradeep drilling proves that the mineral frontier extends far below conventional exploration depths. SG-3 intercepted a Cu–Ni sulfide body enriched in platinum-group elements within Proterozoic metavolcanics at 1.6–1.8 km, with nickel up to 1.2%⁸⁷. The KTB borehole recorded metasomatic mineral assemblages, clarifying the structural and thermal controls on deep mineralization⁸⁸. Unexpected temperatures—180 °C at 10 km and 212 °C at 12 km—challenged models of the Baltic Shield’s thermal state, proving that ore mineralization and hydrothermal activity persist at great depths. SG-3 gamma-ray logs further mapped uranium, thorium, and potassium zonation, refining models of Archean–Proterozoic crustal evolution and metallogeny³⁶. Beyond Eurasia, deep geophysical arrays in South China’s Qinzhou–Hangzhou belt linked Moho depth, Curie isotherms, and orogen structure to the distribution of Fe–Cu and W–Sn–Au–Pb–Zn metallogenic belts⁸, demonstrating how crustal architecture controls deep mineral systems.

Looking forward, ultradeep drilling offers a blueprint for exploring concealed, strategic mineral systems in diverse tectonic settings. Targets include magmatic intrusions hosting Cu–Ni–PGE (Platinum group elements) deposits, alkaline complexes enriched in rare earth elements, granitoid provinces with porphyry Cu–Mo–Au systems, and banded iron formations recording early Earth redox and biospheric evolution²¹. As surface-accessible deposits decline, critical mineral security will depend on characterizing these buried systems. Ultradeep boreholes are indispensable for calibrating geochemical and structural models, testing fluid–rock reaction kinetics, and mapping mineralizing gradients at depth. By integrating drilling results into global exploration strategies, science and industry can jointly expand the resource base while refining theories of crustal evolution. In this way, ultradeep drilling functions both as a scientific laboratory and a strategic exploration tool, bridging fundamental geoscience with the practical need for sustainable mineral supply in the twenty-first century.

Future of crustal geothermal energy

Ultradeep drilling is reshaping how scientists and engineers evaluate geothermal energy by revealing the true thermal and hydrological behavior of the crust under extreme P–T conditions. Traditional geothermal development has relied on shallow boreholes that tap naturally convecting hydrothermal systems, but these typically supply only modest heat and electricity. Even advanced initiatives—such as the Icelandic Deep Drilling Project, which targets magma-related reservoirs—suggest geothermal could meet ~10% of electricity demand, an important but still limited share of the global energy transition⁵⁸. Boreholes such as SG-3 and KTB showed that deep continental crust can store and transmit both heat and fluids, providing a foundation for reimagining geothermal exploitation when combined with modern multi-well designs, advanced reservoir stimulation, and high-temperature drilling technologies.

A defining challenge for ultradeep geothermal development is the extreme downhole environment. Bottomhole conditions in deep wells typically reach 100–300 °C with pressures of hundreds of megapascals^2,89. Temperature gradients are shaped not only by conduction but also by tectonic faults, magmatic intrusions, and reactive mineralogy, creating localized thermal anomalies that may act as natural geothermal “sweet spots.” These zones combine high heat flux with open fracture networks capable of sustaining hydrothermal circulation—ideal for engineered geothermal systems.

Enhanced geothermal systems offer the most promising route to tap this potential, relying on ultradeep drilling to create and sustain permeable reservoirs within hot, dry crystalline rock, typically at depths greater than 5 km¹⁶. These projects act not only as energy sources but also as natural laboratories for testing rock mechanics, fracture propagation, and fluid circulation under extreme P–T conditions. For countries with established oil and gas expertise—such as Canada—this represents a strategic opportunity: existing directional drilling, high-pressure well control, and geotechnical engineering skills can accelerate geothermal deployment. From an energy perspective, ultradeep enhanced geothermal systems could supply baseload, zero-carbon power to complement intermittent renewables. From a scientific standpoint, they provide unique datasets on heat flow, stress evolution, and fluid–rock interactions, improving both resource management and fundamental geoscience. From a production perspective, ultradeep wells also help constrain the deliverability of geothermal systems by quantifying reservoir temperature, permeability, and fluid recharge potential—parameters essential for evaluating whether heat can be effectively extracted and converted into usable energy.

The convergence of energy innovation and Earth science defines the next frontier of ultradeep drilling. Penetrating igneous and metamorphic basement at >9–10 km enables direct testing of models for heat flow, crustal rheology, seismicity, and permeability, while cores and logs document how thermal conductivity, pore pressure, and fracture networks interact to shape geothermal systems. These insights refine global geothermal models but also reveal significant operational risks: hydrogen-rich fluids can induce steel embrittlement, chemically aggressive brines accelerate casing corrosion, and fractured or overpressured intervals heighten risks of wellbore instability or blowouts.

Commercialization hinges on overcoming several technological barriers. Extreme downhole temperatures exceed the limits of most electronics, logging tools, and rotary-steerable systems; high-temperature brines and acidic fluids rapidly degrade steel; low permeability in hot, dry basement requires reliable stimulation; and drilling costs rise steeply beyond 7–10 km. Mitigation strategies include ultrahigh-temperature electronics, corrosion-resistant alloys and ceramics, pressure-managed drilling, and closed-loop or multilateral well designs, supported by AI-based monitoring of fractures and wellbore stability. Emerging techniques—such as laser or plasma drilling and high-temperature downhole power generation—may further reduce cost and enable routine access to ≥10 km systems.

In the context of the energy transition, ultradeep drilling positions geothermal as a dispatchable, low-carbon energy source capable of tapping heat and fluid flux from the mid- to lower crust. By demonstrating that geothermal potential extends far beyond shallow hydrothermal systems, it suggests that most continental regions could host engineered geothermal projects if drilling costs decline and high-temperature well integrity improves. This is critical for decarbonization, as geothermal offers continuous baseload power with minimal land footprint. Ultimately, the scalability of ultradeep geothermal energy will depend on cost reduction, material advances, and integrated power-generation strategies, reinforcing geothermal’s potential to become a science-driven cornerstone of a resilient, low-carbon global energy mix^16,58.

Summary and future perspectives

Ultradeep drilling has evolved from an engineering milestone into a strategic frontier for both Earth system science and resource security. Since the mid-20th century, scientific drilling has advanced from purely exploratory efforts to a multidisciplinary platform addressing four grand challenges: lithospheric geodynamics, geo-resources, geohazards, and long-term climate–environmental change⁹⁰. Pioneering programs such as the Russia’s Kola SG-3 and Germany’s KTB proved that crystalline basement is permeable and chemically reactive—with active fluid–rock interactions involving brines, gases, and mineral alteration—and that these deep crustal zones are not isolated or “sealed” as once assumed, but instead can host circulating fluids and significant resource potential. China’s recent SDTK-1 and X-1 well pushed this frontier further, producing the first confirmed hydrocarbons below 10 km and extending the lower limits of petroleum systems. Collectively, these breakthroughs show that ultradeep drilling is the only direct way to test lithospheric models, measure stress and strain at depth, and observe the fate of fluids, carbon, and heat under extreme P–T conditions.

One of the greatest scientific impacts of ultradeep drilling is its ability to turn abstract geodynamic models into testable process histories. Cores and geophysical logs across seismic boundaries such as the Conrad and Moho show that many deep reflectors represent metamorphic layering, fluid-filled fractures, and mylonitic shear zones rather than sharp compositional changes. Stress data from KTB and other deep wells demonstrate that the brittle regime extends to 8–10 km and that the crust is critically stressed with near-hydrostatic pore pressures, reconciling borehole geomechanics with plate-scale strain rates. These findings challenge long-held assumptions about the depth of the brittle–ductile transition, showing it to be deeper, more variable, and strongly influenced by fluids and reactions.

Thermal structures have also been redefined. High-resolution temperature logs from SG-3, KTB, and newer ultradeep wells reveal curved geothermal gradients and conductive heat flow that often exceed shallow predictions by 30–130%. Such data recalibrate models of crustal strength, heat transport, and geothermal potential. Borehole observations show that the deep crust is neither dry nor impermeable. Fluids move through critically stressed faults, where they trigger chemo-mechanical reactions such as serpentinization and carbonation, which produce crystallization pressures >300 MPa, fracture surrounding rocks, sustain permeability, and link deep hydrogen generation, CO₂ mineralization, and ore-forming processes. Together, these processes demonstrate that the deep crust operates as an active, evolving system, with fluid circulation and chemical reactions continuously reshaping its physical and mechanical properties.

Future research must deliberately target these “blind zones” to refine models of crustal strength, permeability evolution, and heat flow. Deep drilling observatories should integrate long-term stress and fluid monitoring with high-resolution isotopic tracers, helping resolve how earthquakes nucleate, how permeability changes over time, and how deep reservoirs of hydrocarbons, hydrogen, and ore fluids persist or leak. A major scientific opportunity lies in quantifying the rates and thresholds of reaction-driven fracturing and fluid pressure build-up—processes that govern both geohazards and deep resource systems.

Perhaps the most transformative resource insight is the redefinition of hydrocarbon limits. Classical petroleum system models predicted oil destruction beyond 8 km, but SDTK-1 and X-1 confirm methane persistence and commercial gas flow beyond 10 km, driven by mineral catalysis, confining pressure, and water activity rather than temperature alone. These findings demand that basin models incorporate pressure-dependent kinetics, catalytic mineral surfaces, and water–rock reactions to correctly predict hydrocarbon and hydrogen potential at great depth.

The natural hydrogen frontier is emerging as an equally disruptive field. Multiple deep H₂ generation pathways—serpentinization of ultramafics, radiolysis of water in radiogenic crust, and thermogenic hydrogen from overmature organics—operate under distinct lithologic and thermal regimes. Noble gases such as helium provide critical tracers for migration and trapping. Ultradeep drilling remains the only way to validate generation rates, assess seal integrity, and determine the long-term stability of hydrogen accumulations, which are increasingly recognized as a potential clean energy resource.

Deep mineral systems are likewise being reframed. Sulfides, graphite-rich horizons, and metasomatic assemblages encountered in SG-3 and KTB link devolatilization and fluid migration to ore formation. Serpentinization and carbonation couple hydrogen production with metal precipitation, showing that processes generating energy resources also drive mineralization and mechanical weakening. Integrating these findings into exploration models could unlock concealed ore systems critical for future mineral supply.

Meeting these scientific and industrial goals will require major technical advances. Ultradeep wells operate at 200–300 °C, >100 MPa, and in chemically aggressive fluids that challenge tools, cements, elastomers, and electronics. Innovations are needed in high-temperature polycrystalline diamond compact bits, thermally stable muds and cements, corrosion-resistant alloys, and memory or fiber-optic logging tools. Pressure-managed drilling and AI-enabled real-time monitoring will be essential to maintain wellbore stability in critically stressed, fluid-active crust.

Equally important is designing wells as long-term scientific observatories. KTB and SAFOD proved the value of installing geophones, strain meters, fluid samplers, and repeat-access completions, enabling decades of stress, seismicity, and fluid monitoring. Future >10 km projects should integrate established traffic-light seismic protocols—operational frameworks that adjust drilling or injection activity based on real-time seismic thresholds—together with robust zonal isolation across major faults and coupled geochemical–geomechanical monitoring. Such measures go beyond current practice, where seismic monitoring is often limited to passive observation, and provide a proactive system for managing induced seismicity and safely investigating active deep-crustal processes.

Strategically, a balanced global portfolio of ultradeep targets is emerging. Overmature organic “hydrogen–hydrocarbon kitchens” can calibrate deep fluid generation; basement fault corridors with helium anomalies can test migration and sealing; and ultramafic–granitic hybrids can quantify serpentinization, radiolysis, and ore-forming metasomatism. Standardized downhole measurements—gases, noble-gas isotopes, water chemistry, microbiology, fracture permeability—will enable cross-basin comparisons and predictive global models.

Finally, ultradeep drilling should be coordinated internationally, similar to the International Ocean Discovery Program (IODP). Synergies between continental ICDP initiatives and offshore IODP expeditions could maximize scientific return, address hazards and resources jointly, and unify studies of deep-time tectonics, magmatism, and carbon cycling. Priority targets include large igneous provinces, fossil and active subduction zones, rift systems, and Archean–Proterozoic cratons, which offer natural laboratories for plate tectonics and early Earth evolution.

By reframing ultradeep drilling as both a technological and scientific enterprise, future programs can deliver breakthroughs in deep energy systems, hydrogen and mineral resources, geothermal potential, and earthquake physics while advancing fundamental geoscience. If pursued collaboratively and strategically, the next generation of >10 km observatories will turn Earth’s inaccessible crust into a permanent laboratory, transforming inference into direct observation and reshaping our understanding of planetary dynamics and sustainable resource use in the 21st century.