Mud-gas logging for H2 while drilling

Summary

Mud-gas logging is drilling’s built-in gas detector: as the bit cuts rock, gas trapped in the formation is stirred loose into the drilling fluid (“mud”), carried up the wellbore, and released at surface, where a trailer-mounted lab pulls a slipstream of that gas off and analyzes it continuously, producing a real-time readout of what gas is coming out of the ground at every depth as the well is drilled.

Because some version of this equipment rides on essentially every oil-and-gas well, it is easy to assume that “the rig had a mud logger” means hydrogen was being watched, but that is usually not the case. The classic mud-logging detector, a flame ionization detector, cannot detect hydrogen at all (H2 has no carbon to burn), so a large share of the world’s mud-logging fleet reports no signal for a hydrogen show regardless of its size.

Instruments that can detect hydrogen face a second, subtler problem: drilling itself generates hydrogen, both from steel drill pipe corroding in contact with water and from friction and heat at the bit tearing apart rock and water molecules, so a measured “H2%” curve has to be separated into a natural-formation piece and a drilling-artifact piece before it says anything trustworthy about the reservoir below.

With the correct detector chemistry, careful lag-time depth correction, and a disciplined artifact-subtraction protocol, mud-gas logging is the only technique that watches the entire open borehole in real time during drilling, the one operation every well must undergo, and it provided supporting evidence at each of the natural-hydrogen discoveries reviewed for this page, from Mali to Kansas to South Australia. Done carelessly, it is a quick way to produce a confident but wrong resource estimate.

At a glance — published performance envelope

Detection limit

5–1,000 ppm

Range

5–990,000 ppm

Response t₉₀

90–3,600 s

Recovery

n/a

Selectivity

medium

Drift

medium

Operating T

n/a

Humidity

n/a

Power

n/a

Unit cost

$500k–$2000k

TRL

9

Main interferents: CH4 CO2 N2 He H2S co-evolved species require chromatographic or mass separation, CO and ethene propene as drill-bit-metamorphism DBM proxy markers not true interferents but correlated confounders, mud temperature affects Henry's-law extraction efficiency for all dissolved gases, mud type OBM vs WBM changes which artifact pathway dominates and which proxy correction applies · poisons: helium carrier gas in TCD systems causes non-monotonic inverted W-shaped H2 peak above ~8-10pct H2 a detector-configuration failure mode not a chemical poison, MS background at m, drill-string and casing steel corrosion is a continuous parasitic H2 source independent of drilling activity. Ranges span the published devices/variants covered below — per-claim citations are in the body; data: src/data/modalities/mud-gas-logging.yaml.

Mechanism & physics

Not a sensor — a four-stage measurement chain

Every other page in this compendium describes a single transduction physics. Mud-gas logging has none of its own: it is a measurement chain — extraction, transport, detection, and depth-registration — and each stage has independent governing physics that can add or subtract hydrogen signal that has nothing to do with the formation.

The USGS’s 2025 geologic-hydrogen prospectivity assessment states the historical consequence: “drilling mud gas logging tools do not routinely measure hydrogen, and hydrogen commonly is used as a carrier gas in laboratory gas analyses,” a dual blind spot that let decades of accidental hydrogen shows in conventional oil-and-gas wells go unreported or get dismissed as noise [1].

Retrofitting the chain for deliberate, quantitative H2 detection is now an active, well-documented engineering problem spanning at least one dedicated conference paper, one industry patent, and a growing peer-reviewed literature [1][2][3].

Gas–liquid partitioning at the extractor: a Henry’s-law framework

Downhole, H2 exists as some combination of free gas and gas dissolved in the water or oil phase of the mud. At the surface degasser, the measured gas-phase concentration relates to the dissolved concentration through a temperature- and salinity-dependent partition coefficient. A simplified Henry’s-law formulation, validated against experimental H2–water/brine solubility data, gives:

$C_{gas} = \frac{C_{liquid}}{H(T,S)}, \qquad H(T,S) = H_{ref}\,\exp\!\left[-\frac{\Delta H_{sol}}{R}\left(\frac{1}{T}-\frac{1}{T_{ref}}\right)\right]\cdot f(S)$

where $H(T,S)$ is the dimensionless Henry’s constant, $\Delta H_{sol}$ the molar enthalpy of dissolution, $T$ the mud temperature, and $f(S)$ a salinity correction — H2 solubility falls with increasing brine salinity, the same “salting-out” behavior seen for CO2 and CH4 [4].

Because the surface degasser operates far from downhole $T$ and $P$, it never recovers 100% of the dissolved gas: it recovers a fraction set by residence time, agitation intensity, and how close the trap’s local gas–liquid equilibrium approaches the bulk value. This fraction is the extraction efficiency, $\eta = C_{trap}/C_{true}$, and it is gas-species-dependent.

Light, weakly soluble H2 might be expected to partition into the gas phase faster than heavier hydrocarbons, but a 2023 Schlumberger patent on mud-gas hydrogen logging reported the opposite in practice: H2 tends to remain dissolved and desorbs unusually slowly from drilling fluid, even during active recirculation, requiring more aggressive extraction (heated impeller degassers, recirculation loops, headspace probes) than conventional hydrocarbon gas-trap designs provide [5].

The same patent quotes an approximate H2 solubility of only ~1–2 ppm by weight in water at surface conditions, meaning even small absolute extraction losses are large relative to the anomaly sizes relevant to exploration [5].

Two extraction paradigms, two error budgets

Industry uses two extractor architectures (taxonomy in Variants, below); their governing physics differs. A constant-volume trap (CVT) is a stirred, partially evacuated headspace in equilibrium with a metered mud sample — it approximates a well-mixed flash-equilibrium stage, and its calibration is a lumped efficiency factor per gas species, mud type, and temperature [5][6][7].

A semi-permeable membrane extractor instead exploits Fickian diffusion of dissolved gas across a polymer membrane directly in the flowline, governed by a steady-state diffusive flux $J = P_m \cdot \Delta C / \delta$ (permeability $P_m$, membrane thickness $\delta$), calibrated against each hydrocarbon component’s solubility in the specific mud system rather than a single lumped constant [6][7][8].

The two diverge most sharply for methane: a CVT tends to over-read methane relative to its true dissolved concentration because agitation strips more of it than a diffusive membrane would pass, so membrane systems apply a distinct, component-specific solubility correction that CVT systems do not need in the same way [6][8].

Transport lag and depth-registration smearing

Gas liberated at the bit does not appear at surface instantaneously. The bulk transport delay (lag time) is set by the annular fluid volume above the bit divided by pump flow rate, growing from minutes at shallow depth to hours at total depth in deep wells; it is tracked in real time via pump-stroke counts and recalibrated against physical lag markers (lost-circulation material, cavings, lithology breaks) [9] (unverified — see reference note).

Superimposed on this bulk advective delay is dispersive smearing: because cuttings and gas-laden mud parcels travel the annulus at slightly different velocities (affected by hole deviation, eccentricity, and rheology), a single depth of origin arrives at surface spread over a finite time window rather than as a sharp pulse.

This has been formally modeled for directional wells as a cuttings-lag-depth distribution problem, and the depth resolution of any interval-scale mud-gas reading is fundamentally limited by this dispersion, not by analyzer response time [10]. For an anomaly the width of a single reservoir bed, dispersive smearing can blur or erase a thin high-grade H2 interval entirely — one reason single-point “highest reading” reporting is unreliable without back-calculating true interval boundaries.

The artifact-generation term: why “H2” in the log is a sum, not a signal

The key point for H2 mud-gas logging, absent from conventional hydrocarbon mud logging, is that the measured signal is a superposition:

$C_{H_2}^{measured} = C_{H_2}^{natural} + C_{H_2}^{DBM} + C_{H_2}^{corrosion} + C_{H_2}^{additive} + \varepsilon_{extraction,transport}$

Two of these artifact terms have well-characterized proxy relationships that allow partial subtraction. Drill-bit metamorphism (DBM) — thermal cracking and mechanochemical reaction of rock, steel, and drilling fluid at the hot bit face — co-generates CO, ethene (C2H4), propene, and other species not found in natural thermogenic or microbial gas; in oil-based mud systems the DBM-hydrogen fraction correlates well enough with these markers to support a subtraction correction.

A wellsite dataset in the 2023 SLB patent shows an approximately linear H2–CO relationship with slope ≈6 in a clean-sandstone drilling interval, and increasing heating power at the bit produced a linear rise in hydrogen output — consistent with a thermally/mechanically driven generation process rather than formation degassing [5][11][12].

Galvanic/corrosion hydrogen arises from the classic hydrogen-evolution half-reaction of anaerobic steel corrosion — the drill string, casing, and steel fines act as the anode in contact with formation or mud water, with $2H^+ + 2e^- \rightarrow H_2$ at cathodic sites — and this pathway is documented independently of any oil-and-gas context: percussion- and auger-drilled groundwater monitoring wells have been shown to generate up to several volume-percent H2 in the borehole headspace purely from steel–water reactions, with no petroleum or geologic-hydrogen target involved at all [13][14].

For interpretation, the difficulty is that DBM/corrosion hydrogen and “natural” hydrogen from serpentinization or radiolysis are generated by chemically similar redox chemistry (oxidation of Fe(II)-bearing minerals or metallic iron by water, liberating H2): the artifact is the same reaction as the target, occurring on a drill bit or drill string rather than in the reservoir rock. For this reason it cannot be separated by chemistry alone and instead requires proxy-species correlation, trend analysis, and calibrated background subtraction [13][14].

Mechanochemical H2 generation: a statistical, not closed-form, model

Unlike catalytic or electrochemical sensors elsewhere in this compendium, there is no first-principles closed-form rate law analogous to Arrhenius kinetics for DBM hydrogen; the working models are empirical correlations built from wellsite drilling parameters. A 2024 peer-reviewed study of pre-salt carbonate wells in Brazil’s Santos Basin found that DBM severity — quantified via an ethene/(ethene+ethane) ratio and a “dryness index” — peaked when drilling igneous and carbonate intervals with diamond-impregnated bits, and that during severe DBM “the gas composition is completely changed,” rendering those intervals unusable for direct hydrocarbon (or, by extension, hydrogen) show evaluation without correction [12].

A related 2025 workflow applies machine learning (dimensionality reduction plus kernel ridge regression) to fuse drilling parameters, advanced mud-gas composition, mud properties, and geological context into a continuous “DBM severity curve,” specifically so H2 readings from wells targeting natural hydrogen can be de-risked before they drive completion or testing decisions [15]. The practical governing “equation” for H2 exploration mud logging is therefore closer to a statistical discriminant/regression than a physical law:

$\hat{C}_{H_2}^{natural} = C_{H_2}^{measured} - \hat{k}_{DBM}\cdot C_{marker} - \hat{k}_{corr}\cdot f(\text{drill-string age, mud chemistry, ROP})$

with $\hat{k}_{DBM}$ fit per well/bit/lithology combination from co-logged marker species (CO, C2H4) and $\hat{k}_{corr}$ informed by trend analysis (background stability, correlation with mechanical drilling parameters) rather than derived from first principles [5][11][12][15].

Variants & subtypes

By extraction method

• Agitated constant-volume gas trap (CVT). A metal box or cylinder mounted in the shale-shaker possum belly, with mud entering/leaving through lower ports and an agitator continuously churning the mud while ambient air is drawn through the headspace; the resulting air/gas mixture is pulled to the logging unit by a vacuum/suction line [6][17]. This is the default, decades-old workhorse of conventional hydrocarbon mud logging. Extraction efficiency is reported to range roughly 30–70% and is sensitive to pump rate, mud level in the trap, and mud rheology — none independently controlled, so raw trap output is treated as a qualitative “total gas” trend rather than an absolute concentration unless separately calibrated [6]. Bench-scale optimization studies of gas-liquid separation in these traps report the highest, most stable extraction efficiency at specific stirrer-speed/flow-rate combinations (on the order of ~1680 rpm agitator speed, ~1.0 L/min mud flow), underscoring how strongly efficiency depends on operating point rather than being a fixed instrument constant [18].
• Constant-volume, constant-thermodynamic-condition degassers (modern quantitative CVT). Second-generation systems — the lineage behind SLB/Geoservices FLAIR/FlairFlex — sample drilling fluid from both the inbound (drill-pipe/kill line) and outbound (flowline) sides under controlled, repeatable thermodynamic conditions, apply a proprietary per-gas-species extraction-efficiency calibration, and correct for recycled/contaminating gas, to deliver quantitative (not just relative) formation-gas composition [7][19][20].
• Semi-permeable membrane extractors. A polymer membrane probe inserted directly into the mud flow line; dissolved gas diffuses across it according to its own solubility/permeability in that mud system, and downstream analysis (historically paired with high-speed TCD gas chromatography) back-calculates true dissolved concentration using per-component solubility rather than a single lumped trap-efficiency number [6][8]. This approach entered the mid-2000s literature specifically because it avoids the CVT’s tendency to over-represent light, easily stripped gases like methane relative to heavier components [6][8].
• Headspace/vacuum/centrifugal/impeller degassers with heating, tested explicitly for hydrogen recovery in the 2023 SLB patent because standard trap geometries under-recover the slowly desorbing H2 fraction; the patent’s preferred configurations add heat and/or more aggressive mechanical agitation than legacy hydrocarbon-tuned traps use [5].

By analyzer/detector technology (the “back end” of the chain)

• Flame ionization detector (FID) — the historical default, and unable to detect H2. FID ionizes hydrocarbon molecules in a hydrogen/air flame and detects carbon–hydrogen bond ionization current; because H2 contains no carbon, it produces no FID signal at all [21]. Nearly all legacy and many current commercial hydrocarbon mud-logging units are FID-based total-gas analyzers reporting “methane-equivalent” gas, and unable to detect hydrogen at any concentration [17][21].
• Thermal-conductivity detector (TCD), typically paired with a GC. TCD responds to any species whose thermal conductivity differs from the carrier gas, including H2, He, N2, and CO2, and is the traditional workaround for non-hydrocarbon gas detection in mud logging [3][6]. It has a well-known, non-obvious failure mode for hydrogen: with a helium carrier gas (a very common GC configuration), the thermal conductivity of a He/H2 mixture is non-monotonic with H2 concentration — it decreases, passes through an inversion point (~8–10% H2 in He), then increases again, producing a distorted “W”-shaped, non-integrable peak instead of a clean, quantifiable one [22]. The standard fixes are switching carrier gas to nitrogen or argon (in which H2’s TCD response behaves normally) or using a mixed He/H2 carrier so the sample sits entirely on one side of the inversion point [22]. Any GC-TCD system nominally reporting hydrogen must be checked for which carrier gas it uses, because a helium-carrier TCD system not specifically reconfigured can silently misreport H2.
• Quadrupole mass spectrometry (single-stage or tandem). Wellsite mass spectrometers such as SLB’s DQ1000 provide real-time, cycle-time-limited (order ~90 s) semiquantitative analysis of C1–C10 hydrocarbons plus inorganic species (CO2, He, N2, O2, sulfur-bearing volatiles) [23]. For H2 specifically, mass spectrometry faces a background problem: H2 sits at mass-to-charge $m/z=2$, close to instrument background from residual water-vapor cracking and other sources, so the 2023 patent’s protocol requires demonstrating <1000 ppm background noise (via dry-air flushing) before a spectrometer is deemed usable for quantitative H2, followed by a three-point calibration (0.1%, 1.0%, 10% H2 mixtures) fit to a linear-log or exponential curve, targeting $R^2 > 0.999$ [5]. Dedicated GC-MS methods report accurate quantification from roughly 5–5000 ppm using electron-ionization detection with proper chromatographic separation, addressing the m/z=2 background problem that plagues simple residual-gas-analyzer (RGA)-style mass spec, which lacks chromatographic separation and therefore cannot reliably quantify H2 admixed with other low-mass species [3].
• Gas chromatography–thermal conductivity detection (GC-TCD) as a dedicated H2 method — the most widely cited “workhorse” quantitative approach in current natural-H2 mud-logging literature, valued because it directly separates H2 from He, N2, and hydrocarbons on a molecular-sieve or equivalent column before detection, avoiding both the FID-blindness problem and the MS background problem, at the cost of slower cycle time than a continuous-flow analyzer [2][3][24].
• Cavity-enhanced Raman spectroscopy (CERS) and related Raman analyzers — the emerging, still-immature alternative. Raman scattering gives simultaneous multi-species quantification (H2, CH4, CO2, N2, H2S, He) from a single optical measurement with no consumable carrier gas; published cavity-enhancement work claims sub-ppm detection limits with ~1000× Raman signal gain from multi-pass optical cavities, and industry papers explicitly propose CERS as a “dedicated natural H2 logging analyzer” direction. Online multi-gas Raman systems combined with machine-learning spectral deconvolution (e.g., CNN-LSTM-AM architectures) have been demonstrated for mud-logging-relevant gas mixtures, though as of this writing these remain research/early-commercial systems rather than an established field-proven service line — see Raman spectroscopy and Enhanced Raman for the underlying transduction physics [3][25].

By reporting philosophy

Simple “total gas” trend logging (qualitative, uncorrected, historically FID-based, adequate only for lithology/pressure-trend work) versus quantitative fluid/formation-gas logging (calibrated extraction efficiency, per-species solubility correction, recycling/contamination correction, and — for H2-capable services — an explicit DBM/corrosion correction layer) [5][7][19][20].

Performance envelope

The table below is a method-comparison envelope rather than a single-instrument datasheet, because mud-gas logging’s “performance” is really the composite of extractor efficiency, analyzer chemistry, and depth-registration accuracy — no single number characterizes the whole chain.

Parameter	Value / range	Notes	Ref
Gas-trap (CVT) extraction efficiency	~30–70%	Strongly dependent on agitator speed, mud flow rate, mud level, rheology; not a fixed instrument constant	[6][18]
H2 solubility in surface water/mud (approx.)	~1–2 ppm by weight	Quoted in the SLB mud-gas-hydrogen patent as the scale that makes extraction losses significant relative to exploration-relevant anomalies	[5]
DQ1000 wellsite mass spec — general HC sensitivity	~0.1 ppm (C4+) baseline; 90 s cycle time	Not H2-specific; illustrates general wellsite-MS speed/sensitivity class	[23]
MS background requirement for quantitative H2 (m/z=2)	<1000 ppm equivalent (post air-flush)	Gatekeeping criterion before an MS unit is deemed usable for H2 quantification	[5]
MS 3-point H2 calibration range	0.1%, 1.0%, 10% H2 in calibration gas	Linear-log/exponential fit, target $R^2>0.999$	[5]
GC-MS quantitative H2 range (dedicated method)	~5–5000 ppm	Electron-ionization GC-MS developed to directly quantify H2 despite background at m/z=2	[3]
Mud-gas MS quantitative range (broad instrument class)	~50 ppm – 15–20 vol%	Range established after custom calibration work on a wellsite MS for a Colombian H2 exploration well	[24]
FID total-hydrocarbon sensitivity	~5 ppm (as methane-equivalent)	Structurally zero sensitivity to H2 regardless of concentration	[17][21][26]
GC-TCD H2 response with He carrier	Non-linear/inverted (“W”-peak) above ~8–10% H2 in He	Must switch carrier to N2/Ar or use mixed He/H2 carrier for valid H2 quantification	[22]
Lag time (bulk transport delay)	Minutes (shallow) to several hours (deep, e.g. >4000 m)	Scales with annular volume / pump rate; recalibration against lag markers is standard industry practice but not confirmed by [9] (unverified)	[9]
Depth-resolution smearing	Interval-scale (bed-thickness dependent); worse in deviated/horizontal wells	Governed by cuttings/gas-transport velocity dispersion, not analyzer response time	[10]
Bourakebougou (Mali) accumulation-defining threshold	>1 mol% H2 over >20 m interval thickness	Field-specific operational threshold used by SEMM/Geoservices logging	[27]
Bourakebougou discovery well composition	98% H2, 1% N2, 1% CH4	Confirmed by wellsite GC (SEMM Logging)	[27]
Colombia (Llanos Basin) raw mud-gas H2	Locally >40 vol% (raw, uncorrected)	Large fraction identified as drilling-artifact (DBM/corrosion), not formation H2	[14]
Colombia downhole formation-water-normalized H2 vs. mud-gas-normalized H2	~1.1 vol/vol vs. ~1.3 vol/vol (at equivalent depth)	Cross-check between wireline downhole sampling and corrected mud-gas log	[24]
Kansas (HyTerra, Nemaha/McCoy-1)	Up to 96.1% (samples), 83% (McCoy-1 real-time log, TD 1696 m)	Independent lab-verified vs. real-time mud gas	[28][29]
Nebraska (Hoarty NE-3, 2019)	Reported >30% H2 “in mud bubbles” deep in well; “considerable uncertainty” per outside 2022 review	Illustrates raw-log ambiguity absent rigorous artifact correction	[30]
Ramsay-1 (South Australia)	73.3% H2 at 240 m TVD (mud-gas-calibrated), vs. 76% recorded in a 1931 well at the same depth	Real-time calibrated mud gas used to derive a “pseudo hydrogen saturation log”	[31][32]
Ramsay-3 (South Australia, 2025)	Up to 97% H2 purity	Latest well in the same program	[33]
Mud-logging unit capital cost class	~US$500,000–2,000,000 per fully equipped unit (unverified)	Day-rate service model (equipment + personnel), not typically sold outright	[16]

Analyzer capability vs. exploration need

The main point is a three-to-four-order-of-magnitude gap between analyzer capability and exploration need: dedicated GC-MS/GC-TCD methods can resolve single-digit-ppm H2 — just above the ~0.53 ppm atmospheric background and well into the 1 ppm–percent anomaly range relevant to Hestia (see Why H2 is hard and Geologic context for that background figure) — but legacy FID-based rig mud loggers, still the majority installed base on non-dedicated wells, report no signal for H2 at any concentration, and naive GC-TCD setups with a helium carrier can report a qualitatively wrong, non-monotonic signal above roughly 8–10% H2.

As a result, “the rig had a mud logger” is not evidence of H2 capability. The analyzer chemistry must be checked explicitly, contract by contract.

Environmental covariance & cross-sensitivity

Rock type is the first-order control on drill-bit-metamorphism (DBM) severity. DBM hydrogen and its co-generated markers (CO, ethene, propene) are not uniform background noise — they are strongly lithology- and bit-type dependent. The 2024 Santos Basin study found DBM severity (via the ethene/(ethene+ethane) ratio and “dryness index”) peaked specifically when drilling igneous and carbonate rock with diamond-impregnated bits, to the point that gas composition was “completely changed” in affected intervals [12].

Independent literature on wet-grinding experiments shows mechanochemical H2 generation is strongly enhanced in ultramafic rocks and phyllosilicates relative to carbonates or quartz/feldspar-dominated lithologies, because availability of Fe²⁺ and reactive Si–O/Al–O–Si bond rupture at freshly ground mineral surfaces drives the reaction — the same mineralogy that makes serpentinite-hosted settings attractive natural-H2 targets also makes DBM artifact hydrogen worst in those settings, an unfortunate coincidence for interpretation [14].

Mud type governs which artifact pathway dominates. With oil-based mud (OBM), DBM hydrogen is generated alongside thermally cracked hydrocarbon byproducts (alkenes, especially ethene/propene) that correlate well enough with the H2 artifact fraction to support a subtraction correction [1][5][14].

With water-based mud (WBM) or plain water, drilling-challenges literature reports “unexpected hydrogen generation through water decomposition during drilling operations” — mechanical/frictional energy at the bit driving direct water splitting or Fe-mediated water reduction — without the same convenient hydrocarbon-marker correlation available in OBM systems, making WBM-drilled intervals harder to correct [34].

The same source reports that hydrogen production increases stepwise with drilling parameters — rotary speed (RPM) and weight-on-bit (WOB) — and that bit vibration (monitored via triaxial accelerometers), especially with PDC bits in hard formations like granite/basalt, accelerates the effect, directly linking mechanical power input at the bit to artifact generation rate [34].

Drill-string and casing corrosion is a parallel, independent artifact channel that requires no drilling action at all — it is a function of exposed steel surface area, water chemistry, and time. Percussion/auger-drilled groundwater monitoring boreholes with no petroleum objective have shown up to several volume-percent H2 in borehole headspace purely from steel–water reaction [13], and the same mechanism (formation water or mud filtrate reacting with drill-string steel, casing, or steel cuttings/shavings) has been explicitly identified as a contributor — up to ≤4.3 vol% in one cited borehole case — in natural-hydrogen exploration wells [14].

Because this pathway is decoupled from lithology and drilling mechanics, it does not correlate with the same DBM marker species (CO, ethene) and must instead be diagnosed through trend behavior: persistence during non-drilling shut-in periods, correlation with drill-string metallurgy/age, and consistency across pipe connections rather than depth-localized spikes.

Additive chemistry interacts asymmetrically with different gas species — a favorable finding for H2. H2S scavengers (zinc oxide, copper nitrate, potassium permanganate) added to water-based mud specifically to control H2S have been characterized for scavenging capacity — permanganate reaches ~307.5 mg H2S/g saturation capacity, roughly 2.1× the capacity of copper nitrate and 8.8× the capacity of zinc oxide respectively — and for their effect on other gases: the same laboratory study found the base mud (and, by extension, its scavenger additives) showed no measurable affinity for CO2, inlet and outlet CO2 concentrations were unchanged, implying scavenger chemistry targeted at H2S is unlikely to selectively strip H2 either. This is reassuring for H2 quantification but has not been directly verified for H2 in the same rigorous breakthrough-capacity format [35].

This matters operationally because H2S and H2 frequently co-evolve from the same reducing, Fe-rich, or biogenic settings discussed in Geologic context, and H2S scavengers are near-universal on wells where any sour-gas risk exists.

Non-H2 co-evolved species act as both confounders and diagnostic tools. CO2, CH4, He, and H2S routinely appear in the same mud-gas stream. CH4 is not a spectral or chemical interferent for TCD/GC-based H2 methods (they are chromatographically separated), but it is the historical focus of FID-based rig loggers, so a driller watching only the legacy “total gas” trace will see CH4-driven shows and miss H2-only anomalies entirely.

CO2 solubility (and H2S solubility, where present) in the mud is strongly pH- and temperature-dependent, so trap-based readings of these species drift with mud-chemistry changes (e.g., pH-treatment additions) independent of any change in formation gas — a reminder that every gas-trap reading is filtered through mud chemistry, not just extraction hardware.

He is chemically inert and does not participate in DBM or corrosion chemistry, making a He/H2 ratio a potentially useful, though not yet standardized, diagnostic: a natural, deep, radiogenic-He-associated H2 show should show a more stable He/H2 relationship than a shallow, artifact-driven H2 spike with no matching He source.

Temperature and pressure act primarily through solubility, not through analyzer physics. Because H2’s Henry’s-law partition coefficient is temperature- and salinity-dependent [4], the same true formation H2 content will produce a different trap-gas reading depending on mud temperature at the point of extraction — a variable that changes with season, circulation rate, and depth, and that quantitative services correct for by co-logging mud temperature at the extraction point [6]. See Environmental covariance & cross-sensitivity for the cross-modality treatment of temperature and pressure confounds this page’s discussion feeds into.

Failure modes to flag for field crews

• A helium-carrier GC-TCD reporting a suspiciously flat or inverted H2 trend at high concentration is a detector artifact, not a real plateau — check carrier gas before trusting the number [22].
• A rig with only a legacy FID total-gas analyzer will show no H2 response at any concentration — absence of a “gas show” on that trace is not evidence of absence of hydrogen [17][21].
• H2 and CO (or ethene) rising and falling together, tracking ROP/WOB/RPM rather than lithology breaks, is the classic DBM signature and should be treated as artifact-dominated until proven otherwise [5][11][12][34].
• A steady, slowly varying H2 background that persists through connections and shut-in (non-drilling) periods, uncorrelated with drilling mechanics, is more consistent with corrosion than DBM and should be checked against drill-string exposure time/metallurgy [13][14].

Use for geologic H2

Field verdict

Every well drilled with a natural-hydrogen target should carry an H2-capable mud-gas logging package from spud to TD — it is the only real-time, whole-wellbore-coverage measurement available while drilling.

A standard oil-and-gas mud-logging package is unsuitable for H2 unless explicitly re-geared: FID-based units report zero H2 regardless of true concentration (a false negative that could cause a real show to be missed), and improperly configured TCD/He-carrier or uncalibrated MS setups can report large fabricated or distorted H2 numbers — false positives up to tens of vol% in documented cases [5][14][24].

Contracting a “mud logging” line item without independently verifying detector type, carrier gas, and DBM/corrosion correction methodology is a documented way to get a badly wrong answer at the moment the decision to core, sample, or test is being made.

Exploration and appraisal wells (the core use case) — recommended, but only with H2-specific QC. See the verdict above; the H2-specific QC is where the risk lies.

Key risks

Risk 1 — the artifact fraction can exceed the natural signal. In the Colombia case, raw mud-gas H2 exceeded 40 vol% locally, with large fractions of that (tens of vol% in places) later attributed to drilling-related artifact generation [14]. Any raw, uncorrected mud-gas H2 reading should be treated as an upper bound requiring artifact subtraction, not a formation value, especially in oil-based mud or in Fe-rich/ultramafic lithologies where both DBM and corrosion pathways are most active [12][13][14].

Risk 2 — quantitative calibration is well-specific and not yet standardized industry-wide. Extraction efficiency, DBM correction slopes, and MS/GC calibration curves are all fit per well (sometimes per bit run), using proxy species (CO, ethene) whose own quantification is subject to the same mud-type and lithology dependencies described above.

There is no single published, universally applicable correction factor Hestia can apply off the shelf; each campaign needs its own calibration campaign, ideally cross-checked against an independent method (downhole wireline sampling, spot mud-gas can samples sent to an offsite lab) [1][5][24]. This is exactly the kind of contract-level discipline Calibration & QA/QC is meant to enforce.

Interval resolution is coarser than the geology. Lag-time dispersion smears thin, high-grade zones, particularly in deviated or horizontal sections [10]; a mud-gas log alone cannot substitute for wireline logs or core/cuttings analysis when precise reservoir-interval boundaries matter for completion design.

Not a standalone go/no-go tool. Every well case reviewed here — Mali, Kansas, Nebraska, South Australia, Colombia — paired mud-gas logging with independent confirmation (gas chromatography on discrete samples, downhole wireline sampling, or historical well comparison) before committing to a resource characterization [5][24][27][28][31]. Mud-gas logging’s value is as a continuous screening and decision-support layer that tells the wellsite team where and when to spend the much more expensive time on wireline logging, sidewall coring, or formation testing, rather than as the final word on formation H2 content.

Complementary, not competing, with soil-gas surveys and continuous monitoring stations. Soil-gas surveys probe shallow, undrilled or minimally perturbed ground and have their own, mechanistically distinct drilling-artifact problem (shallow probe insertion generating mechanochemical H2 in the probe hole itself, discussed in Geologic context); continuous monitoring stations operate on completed, cased wells with no active mud circulation.

Mud-gas logging fills the gap between those two: it is the only technique that observes the entire open-hole interval in real time during the one operation (drilling) that necessarily happens on every well, regardless of what other instrumentation is deployed.

Vendors & products

The table compiles vendors and services actually named in the H2 mud-gas-logging literature reviewed — a mix of commercial SKUs, a foundational patent, and contractor-attributed field cases where the underlying equipment was not fully disclosed.

Vendor / lineage	Product / service	Extraction method	Analyzer	H2 capability	Notes	Ref
SLB (via Geoservices, acquired 2010)	FlairFlex (“Advanced real-time fluid logging and analysis service”)	Two extraction units (inbound drilling-fluid line + outbound flowline), controlled constant-thermodynamic-condition sampling	GC-MS	Explicitly lists measurement of “nonhydrocarbons such as carbon dioxide, helium, and hydrogen” in current product literature	Quantifies C1–C6, qualifies C7–C8; proprietary extractor-efficiency calibration and recycling/contamination correction; used in formation evaluation, geosteering, and (per current marketing) natural-hydrogen wells	[19][20]
SLB	DQ1000 wellsite mass spectrometer	N/A (analyzer only; paired with rig degasser)	Quadrupole MS	General C1–C10 + inorganics (CO2, He, N2, O2, S-species); H2-specific quantification requires the calibration protocol below	~90 s cycle time, ~0.1 ppm baseline sensitivity for C4+ (not an H2-specific spec)	[23]
SLB (Schlumberger entities)	Patented method: “Mud logging of natural hydrogen” (WO2023192219A1)	Vacuum / centrifugal / impeller (heated) degassers; recirculation loop; headspace probe	MS (protocol-agnostic in principle)	Core subject of the patent: 3-point calibration, background-noise gating, CO/ethene-based DBM correction	Not a commercial SKU per se — describes the methodology underlying SLB’s H2-capable service offerings	[5]
Halliburton (Sperry Drilling / surface data logging)	GasFACTS advanced gas analysis service	Constant volume and temperature gas extraction (combined)	Baseline 8900 GC, Baseline 9000 Total Hydrocarbon Analyzers; IsoLogger GC-C-IRMS for isotope work	Not confirmed H2-specific in materials reviewed; service family targets formation pressure, geology, and hydrocarbon fluid analysis	Documented in a Gulf-of-Mexico real-time well-prognosis application using stable carbon isotope analysis, not an H2 case	[36]
Excellence Logging (Exlog)	Fluid logging service (branded “SAFE” fluid-logging platform per company materials)	Modular, upgradeable gas chain (standard-to-fluid-logging upgrade without full equipment swap)	Not fully specified in public materials reviewed	Company states it has “know-how and proprietary technology in Natural Hydrogen exploration” and was invited to the first Natural Hydrogen Worldwide Summit for its H2 detection/quantification experience	Independent (non-major) surface logging contractor increasingly positioning around natural-H2 wells	[37]
Geolog	Modular Platform Drilling (MPD) formation-gas extraction/quantification literature	CVT-class, calibrated extraction efficiency by mud type	GC-based (per company literature)	Not confirmed H2-specific in the source reviewed	Company states real-time detection of alkenes and H2 at surface informs drilling decisions in its broader “Exploration” service positioning	[7]
SEMM Logging	Wellsite logging at the Bourakebougou (Mali) field	Standard mud-gas extraction (contemporaneous with gamma/neutron/density probes)	Gas chromatograph	Confirmed 98% H2 / 1% N2 / 1% CH4 composition at the discovery well	Earliest large-scale field case in the literature reviewed where mud-gas logging directly defined a commercial H2 accumulation threshold (>1 mol% over >20 m)	[27]
(Contractor not specified in sources reviewed)	Real-time mud gas logging, HyTerra Nemaha/McCoy-1 (Kansas) and Geneva (Nebraska) wells	Not detailed in sources reviewed	Not detailed in sources reviewed	Real-time H2 and He readings reported; independent lab analysis used to verify (up to 96.1% H2, ~5% He)	Wells later converted to long-term monitoring wells for continued surface pressure/gas data	[28][29][38][39]
(Contractor described only as “calibrated real time mud gas analysis” in Gold Hydrogen materials)	Ramsay-1/2/3 (South Australia)	Not detailed in sources reviewed	Not detailed in sources reviewed	Derived a “pseudo hydrogen saturation log”; 73.3% H2 at 240 m TVD, consistent with a 76% reading in a 1931 well at the same depth; up to 97% at Ramsay-3	Explicitly described by the operator as applying “dedicated hydrogen detection tools” distinct from standard oilfield mud logging	[31][32][33]
Generic GC/TCD/MS component suppliers (Agilent, and others referenced for detector physics)	Micro-GC TCD modules, mixed-carrier-gas H2 methods	N/A	TCD	General-purpose H2 quantification method (mixed He/H2 carrier or N2/Ar carrier) applicable to, but not marketed specifically for, mud logging	Cited here for underlying detector physics rather than as a mud-logging product	[22]

Two structural observations for the vendor landscape. First, the market is transitioning from “does the contractor’s standard package happen to include a TCD or MS channel” to explicit H2-branded service tiers — FlairFlex’s current marketing language explicitly names hydrogen, and Excellence Logging is positioning around the H2-Nat Summit community — and Hestia should expect and demand this branding rather than accept legacy hydrocarbon-only packages.

Second, none of the public materials reviewed disclose a standardized, published DBM/corrosion correction factor as a simple lookup table; every documented well case built its own calibration, which argues for Hestia negotiating explicit deliverables (raw vs. corrected H2 curves, calibration methodology, marker-species co-logs) into any surface-logging services contract rather than accepting a single “H2%” curve at face value [5][12][14][24].

Open questions

The natural-hydrogen-specific literature here is very recent and concentrated in 2022–2026, sitting on top of much older foundational work: the corrosion-hydrogen pathway was documented in environmental groundwater science as early as 1994 [13], and the DBM pathway has an even longer pedigree in conventional petroleum mud logging, formalized in a widely cited 2009 SPE paper on drill-bit metamorphism’s impact on hydrocarbon show evaluation [11]. Current frontiers include:

• Moving from single-marker (CO or ethene) linear correction toward multivariate/machine-learning DBM discrimination that incorporates drilling mechanics directly [15][34].
• Maturing cavity-enhanced Raman and other optical multi-gas analyzers toward field-deployable, contractor-branded products [3][25].
• Standardizing extraction-efficiency and solubility calibration across mud systems and temperatures [4].
• Building industry-wide, published (rather than proprietary/per-well) correction methodologies and QC standards, a gap the USGS’s 2025 prospectivity assessment identifies at the policy/methodology level [1].

Specific open items for Hestia:

1. No independently verified, published extraction-efficiency table exists for H2 specifically, analogous to the well-known per-hydrocarbon-component tables used in quantitative fluid logging. Every H2 recovery-fraction number found in this research came from a single-well or single-patent example [5][24]; treat vendor-quoted H2 extraction efficiencies as unverified until the vendor discloses its calibration methodology, and ideally until Hestia runs its own spiked-mud recovery test at anticipated field mud temperatures.
2. No standardized industry threshold exists for “this is DBM/corrosion, not formation gas.” The Mali >1 mol%-over->20 m rule [27] and the various CO/ethene correlation slopes [5][11][12] are field- or well-specific. Track whether a generalizable, cross-basin correction protocol emerges from the current EAGE/SPE conference literature before committing to a single correction methodology across multiple prospects.
3. Full technical specifications (detection limit, response time, cost) for several named services could not be obtained — Halliburton GasFACTS’ H2 capability specifically, Geolog’s H2 capability specifically, and Excellence Logging’s SAFE platform specs — because vendor pages and PDFs returned access errors during this research. Obtain these directly from vendors (RFI/RFQ) rather than assuming absence of capability.
4. Isotopic discrimination (δD, ³He/⁴He) as an artifact-vs-natural diagnostic was not found documented in the mud-gas-logging literature reviewed, despite being conceptually attractive (corrosion/DBM hydrogen should have a distinct isotopic signature from deep serpentinization- or radiolysis-derived H2). This may be an open research gap rather than a settled non-starter — worth a targeted literature check before assuming it is infeasible at wellsite turnaround timescales.
5. He/H2 ratio as a real-time DBM discriminant (proposed qualitatively above) has not been validated against a published dataset in the sources reviewed. If Hestia has access to raw mud-gas logs from any of its own wells, this would be a cheap, high-value internal analysis to run.
6. Quantify the actual gas-trap-to-final-report latency budget (extraction + transport lag + analyzer cycle time + QC processing) for a realistic Hestia target well depth, so real-time drilling decisions (“stop and run wireline now”) can be planned against a known total-latency figure rather than an assumed “real-time” label.
7. Verify current (2026) FLAIR/FlairFlex, GasFACTS, and Geolog literature directly with vendors for explicit H2 detection-limit and QC-workflow claims before writing them into procurement specifications — the public marketing language reviewed here is suggestive but not a substitute for a datasheet-level H2 specification.
8. Cross-check the Nebraska “considerable uncertainty” finding [30] against any more recent (post-2022) independent reassessment, since it illustrates exactly the raw-vs-corrected H2 ambiguity this page warns about and would make a good internal training case study if updated data become available.

Related pages

• Why H2 is hard — the selection-rule and transduction-physics framing behind why no single detector “just works” for H2
• Geologic context — the broader transport-physics and artifact-controversy backdrop for soil-gas work that parallels the drilling-artifact problem here
• Gas chromatography — the separation technique underlying GC-TCD and GC-MS mud-gas methods
• Mass spectrometry — the DQ1000-class wellsite MS detector chemistry and its m/z=2 background problem
• Thermal conductivity & acoustic sensors — the TCD physics and carrier-gas inversion artifact discussed above
• Raman spectroscopy and Enhanced Raman — the physics behind the emerging cavity-enhanced Raman mud-gas analyzer direction
• Soil-gas surveys — the complementary shallow, undrilled-ground technique
• Downhole deployment — continuous monitoring on completed, cased wells, downstream of the drilling phase this page covers
• Calibration & QA/QC — the discipline that must govern any mud-gas H2 service contract
• Environmental covariance & cross-sensitivity — the cross-modality treatment of the temperature/pressure/mud-chemistry confounds summarized here

References

Full reference list (0 entries)

[1] Gelman, S.E., Hearon, J.S., Ellis, G.S. “Prospectivity Mapping for Geologic Hydrogen.” U.S. Geological Survey Professional Paper 1900 (2025). DOI: 10.3133/pp1900. https://pubs.usgs.gov/publication/pp1900/full

[2] “Key Role of Regearing Mud Gas Logging For Natural H2 Exploration.” SPWLA Annual Logging Symposium (SPWLA-2022). https://onepetro.org/SPWLAALS/proceedings-abstract/SPWLA22/3-SPWLA22/D031S001R003/487788

[3] “Advanced Mud Gas Logging Techniques for Hydrogen Exploration.” SPE Middle East Oil and Gas Show and Conference (SPE-25MEOS), 2025. https://onepetro.org/SPEMEOS/proceedings-abstract/25MEOS/25MEOS/D011S018R002/790177

[4] “Hydrogen solubility in water and brine: A simplified Henry’s law method verified against experimental data.” Fuel (2026). DOI: 10.1016/j.fuel.2026.139371. https://www.sciencedirect.com/science/article/pii/S0016236126011257

[5] Strapoc, D., Abolins, N.I., Ammar, M., Gligorijevic, A., Suarez, J., Fowlow, H. (Schlumberger). “Mud logging of natural hydrogen.” Patent WO2023192219A1, filed March 28, 2023, published October 5, 2023. https://patents.google.com/patent/WO2023192219A1/en

[6] AAPG Wiki. “Mudlogging: gas extraction and monitoring.” https://wiki.aapg.org/Mudlogging:_gas_extraction_and_monitoring

[7] Geolog. “Representative Formation Gas extraction and accurate quantification” (product/technical literature). https://www.geolog.com/files/pdf/MPD_accurateformationgas.pdf

[8] “Advances in Chromatographic Analysis of Hydrocarbon Gases in Drilling Fluids — The Application of Semi-Permeable Membrane Technology to High Speed TCD Gas Chromatography.” SPWLA Annual Logging Symposium (SPWLA-2005). https://onepetro.org/SPWLAALS/proceedings-abstract/SPWLA-2005/All-SPWLA-2005/SPWLA-2005-WWW/27665

[9] drillingforgas.com. “Mud Log Gas (background, drilled, swab, connection, trip).” https://drillingforgas.com/technical/well-control/planning/mud-return-gas-background-drilled-connection-trip (does not support the lag-time/pump-stroke/lag-marker claim)

[10] “Cuttings-Lag Distribution in Directional Wells Affects Depth Resolution of Mud Logging.” Journal of Petroleum Technology / SPE. https://jpt.spe.org/cuttings-lag-distribution-directional-wells-affects-depth-resolution-mud-logging

[11] Wenger, L.M., Pottorf, R.J., Macleod, G., Otten, G., Dreyfus, S., Justwan, H., Wood, E.S. “Drill-Bit Metamorphism: Recognition and Impact on Show Evaluation.” SPE-125218-MS, SPE Annual Technical Conference and Exhibition, New Orleans, 2009. https://onepetro.org/SPEATCE/proceedings-abstract/09ATCE/09ATCE/SPE-125218-MS/147297

[12] de Lima Leon, J.A., Penteado, H.L.B., Ellis, G.S., Milkov, A.V., Mendonça Filho, J.G. “Recognition of Artificial Gases Formed during Drill-Bit Metamorphism Using Advanced Mud Gas.” Energies 17(17), 4383 (2024). DOI: 10.3390/en17174383. https://www.mdpi.com/1996-1073/17/17/4383

[13] Bjornstad, B.N., McKinley, J.P., Stevens, T.O., Rawson, S.A., Fredrickson, J.K., Long, P.E. “Generation of Hydrogen Gas as a Result of Drilling Within the Saturated Zone.” Groundwater Monitoring & Remediation 14(3), 140–147 (1994). DOI: 10.1111/j.1745-6592.1994.tb00492.x. https://ngwa.onlinelibrary.wiley.com/doi/10.1111/j.1745-6592.1994.tb00492.x

[14] “Drilling-induced hydrogen: A challenge for natural hydrogen exploration.” International Journal of Hydrogen Energy (2025). https://www.sciencedirect.com/science/article/pii/S0360319925030885

[15] de Lima Leon, J.A. et al. “Machine Learning Approach to Support Mud Gas Evaluation and Detect DBM in Oil Wells.” First EAGE Workshop on Surface Logging, 2025. DOI: 10.3997/2214-4609.202535038. https://www.earthdoc.org/content/papers/10.3997/2214-4609.202535038

[16] Wikipedia contributors. “Mud logging.” https://en.wikipedia.org/wiki/Mud_logging (does not support the capital-cost claim)

[17] Drilling Manual. “Drilled Gas In Mud Logging” (mud logging unit equipment and formation gas). https://www.drillingmanual.com/mud-logging-unit-equipment-formation-gas/

[18] “Gas-Liquid Separation Processes for Mud Logging Systems.” ResearchGate record. https://www.researchgate.net/publication/354466584_Gas-Liquid_Separation_Processes_for_Mud_Logging_Systems

[19] Geoservices (A Schlumberger Company). “FlairFlex: Advanced real-time fluid logging and analysis service” (product brochure), 2021. https://www.slb.com/-/media/files/drilling/brochure/flairflex-br.ashx

[20] SLB. “FlairFlex Advanced Real-Time Fluid Logging and Analysis Service” (product page). https://www.slb.com/products-and-services/innovating-in-oil-and-gas/well-construction/measurements/surface-logging/flairflex-advanced-real-time-fluid-logging-and-analysis-service

[21] Horiba. “Hydrogen Flame Ionization Detection Method (FID).” https://www.horiba.com/usa/process-and-environmental/measuring-principles/fid/

[22] Agilent Technologies. “Hydrogen Detection with a TCD using Mixed Carrier Gas on the Agilent Micro GC.” Application Note 5991-3199EN. https://www.agilent.com/cs/library/applications/5991-3199EN.pdf

[23] SLB. “DQ1000 Wellsite Mass Spectrometer Gas Analyzer” (product page). https://www.slb.com/products-and-services/innovating-in-oil-and-gas/drilling/surface-and-downhole-logging/mud-logging-services/dq1000-spectrometer

[24] “First downhole sampling of a natural hydrogen reservoir in Colombia.” Frontiers in Energy Research (2024). DOI: 10.3389/fenrg.2024.1443269. https://www.frontiersin.org/journals/energy-research/articles/10.3389/fenrg.2024.1443269/full

[25] “On-line multi-gas component measurement in the mud logging process based on Raman spectroscopy combined with a CNN-LSTM-AM hybrid model.” PubMed (2023). https://pubmed.ncbi.nlm.nih.gov/37100477/

[26] SLB. “The Defining Series: Mud Logging” (Oilfield Review). https://www.slb.com/resource-library/oilfield-review/defining-series/defining-mud-logging

[27] “Characterization of the spontaneously recharging natural hydrogen reservoirs of Bourakebougou in Mali.” Scientific Reports 13, 11876 (2023). https://www.nature.com/articles/s41598-023-38977-y

[28] “Natural Hydrogen Breakthrough: HyTerra Reports 83% Concentration in Kansas Drilling.” Hydrogen Industry Leaders. https://hydrogenindustryleaders.com/natural-hydrogen-breakthrough-hyterra-reports-83-concentration-in-kansas-drilling/

[29] “Record 96.1% Hydrogen Found in HyTerra’s Kansas Well.” Discovery Alert. https://discoveryalert.com.au/hyterra-breakthrough-96-1-hydrogen-concentration-kansas/

[30] “Is Nebraska Geologic Hydrogen Site a Black Swan for Clean Energy?” thinc.blog (2023). https://thinc.blog/2023/04/22/is-nebraska-geologic-hydrogen-site-a-black-swan-for-clean-energy/

[31] “Ramsay 1 and Ramsay 2, learnings from the first natural hydrogen exploration wells in Australia.” The APPEA Journal, CSIRO Publishing, EP23119 (2023). https://www.publish.csiro.au/ep/EP23119

[32] Gold Hydrogen. “Significant Concentrations of Hydrogen and Helium Detected in the Ramsay 1 Well” (ASX release, October 31, 2023). https://www.goldhydrogen.com.au/asx-releases/significant-concentrations-of-hydrogen-and-helium-detected-in-the-ramsay-1-well/

[33] Fuel Cells Works. “Gold Hydrogen Confirms Up to 97% Purity at Ramsay 3 Well as South Australia Drilling Completes” (2025). https://fuelcellsworks.com/2025/12/02/green-hydrogen/gold-hydrogen-confirms-up-to-97-percent-purity-at-ramsay-3-well-as-south-australia-drilling-completes

[34] SPE Energy Stream. “Drilling Challenges in Exploration for Subsurface Natural Geologic Hydrogen” (webinar). https://streaming.spe.org/drilling-challenges-in-exploration-for-subsurface-natural-geologic-hydrogen-2

[35] “H2S Scavenging Capacity and Rheological Properties of Water-Based Drilling Muds.” PMC (2020). https://pmc.ncbi.nlm.nih.gov/articles/PMC7711946/

[36] Halliburton. Surface data logging / GasFACTS advanced gas analysis service materials. https://www.halliburton.com/en/subsurface/formation-evaluation/surface-data-logging

[37] Excellence Logging (Exlog). “Fluid Logging” service page. https://www.exlog.com/services/surface-data-logging/formation-evaluation/fluid-logging

[38] “HyTerra provide update on Nemaha Project.” Global Hydrogen Review (2025). https://www.globalhydrogenreview.com/hydrogen/11062025/hyterra-provide-update-on-nemaha-project/

[39] HyTerra. Projects overview (Geneva Project, Nebraska — reported up to 44% H2, 13% He from well testing). https://hyterra.com/projects/