Using X-Ray Fluorescence (XRF) withGeostatistics in Exploration Drilling forOil and Gas

A powerful approach for identifying lithologies in geological studies.

Abstract

Exploration drilling in the oil and gas industry traditionally relies on a combination of geophysical, geological, and geochemical data to identify and evaluate potential hydrocarbon reservoirs. One critical aspect of this process is the analysis of elemental composition in drill cuttings. X-ray fluorescence (XRF) spectroscopy serves as a powerful tool to support geochemical characterization. It provides elemental compositions of rock or soil samples, which can be correlated with lithological characteristics. Geostatistics leverages spatial relationships and variability in this data to map and classify lithologies across a study area, enhancing interpretation of geological units. This whitepaper explores the operational, economic, and technical benefits of deploying portable and benchtop XRF instruments in exploration drilling for oil and gas. It also gives real world applications, methodology, examples, and important considerations for XRF measurements in combination with Geostatistics.

1. Introduction

The search for economically viable oil and gas deposits depends heavily on accurate subsurface characterization. While traditional laboratory-based geochemical techniques such as ICP-MS and XRD provide detailed data, they are often time-consuming and logistically demanding. XRF, a chemically non-destructive analytical technique in conjunction with geostatistical methods, offers a rapid and more economical alternative to traditional laboratory techniques.

In this whitepaper, we discuss:

• The principles and instrumentation of XRF
• Applications in oil and gas drilling as well as other fields
• Methodology for using Geostatistics with XRF data for lithology identification
• Integration with subsurface data
• Economic and operational benefits
• Specific considerations and future developments for petroleum drill cuttings
• Example workflow

2. Principles of XRF Technology

X-ray fluorescence works by bombarding a sample with primary X-rays, causing inner-shell electrons to be ejected. The resulting electron transitions emit secondary (fluorescent) X-rays that are characteristic of the elements present.

Key advantages:

• Chemically non-destructive testing
• Relatively minimal sample preparation
• Rapid results, within 60 seconds in some cases after sample preparation
• Capable of detecting a wide range of elements (Na to U)
• Suitable for solids, powders, and pressed pellets

Types of XRF instruments:

• Portable XRF / Handheld XRF (pXRF/HHXRF): Ideal for field use, real-time decision making
• Benchtop XRF: Greater precision and sensitivity, often used in field labs or core analysis centers

3. Applications in Exploration Drilling

3.1. Geochemical Fingerprinting

XRF provides elemental fingerprints that help:

• Differentiate lithologies
• Identify stratigraphic markers
• Detect hydrocarbon-bearing formations (via trace elements like V, Ni, and Mo)

3.2. Mud Logging and Cuttings Analysis

Near real-time XRF analysis of drill cuttings:

• Accelerates decision-making on well trajectory and target formations
• Reduces reliance on delayed lab results
• Enhances safety by identifying H2S indicators or toxic metal concentrations

3.3. Core and Sidewall Sample Evaluation

XRF allows detailed vertical geochemical profiling:

• Enables rapid reservoir quality assessments
• Helps map mineralogy changes that affect porosity and permeability
• Supports source rock evaluation by assessing elements like S, Fe, and Mn

3.4. Elemental Mapping and Chemostratigraphy

Using XRF data in multivariate statistical models improves:

• Chemostratigraphic zonation
• Basin modeling
• Paleoenvironment reconstruction

3.5 Mineral Exploration

XRF provides rapid means for mapping “pathfinder” elements which assist in identifying lithologies hosting mineralization

3.6 Reservoir Characterization

Using XRF measurements allow a geologist or geochemist to differentiate sedimentary facies in:

• Hydrocarbon reservoirs
• Groundwater reservoirs

3.7 Environmental Studies

XRF allows a user to identify trace lithological controls on soil geochemistry for contamination studies.
3.8 Archaeology

3.8 Archaeology

XRF provides a rapid means for mapping source lithologies of artifacts based on elemental signatures

4. Methodology for Using Geostatistics with XRF Data for Lithology Identification in Petroleum Drill Cutting Analysis

4.1 Sample Collection and Preparation

The objective of this method is correlating lithology units between different methods. The logging geologist is key in selecting these intervals in the field.

Cutting Collection Preparation / Logistics

• The logging geologist must rapidly assess intervals for slower drilling and higher sample interval rates.
• Correlation intervals should be selected where distinctly different lithologies are observed. Marker beds should be multiple lithologies that can be tracked with multiple methods.
• Once the logging geologist identifies the correlation interval, he must notify the drilling supervisor to slow down drilling rate for higher resolution sample collection.
• Correlation intervals are recommended to be 100-150 feet in length with a sample frequency of 5-10 feet between samples.
• It is recommended that all samples be collected before logging and visual cutoff be assessed during sample collection.
• Collect sufficient volume of sample to satisfy the needs of logging, x-ray analysis methods,
• and library samples.

Drill Cutting Collection

• Collect drill cuttings at regular depth intervals (e.g. every 5-10 feet) during higher resolution sample collection areas where drilling is intentionally slowed. Otherwise, collect drill cuttings at regular depth intervals (e.g. every 10-30 feet) during normal drilling operations. Ensure samples are representative of the drilled interval, accounting for lag time due to mud circulation.
• Record precise depth and geospatial coordinates (if applicable, e.g. for deviated or horizontal wells).

Sample Preparation

• Wash cuttings to remove drilling mud and contaminants, following standard protocols (e.g. water or solvent rinsing).
• Dry samples at low temperatures (<60°C, preferably using an air flow dryer) to preserve mineral integrity.
• Pulverize a subset of cuttings to a fine powder (<100 μm) for XRF analysis to ensure homogeneity and minimize matrix effects.
• For portable or handheld XRF (pXRF/HHXRF) measurements on unpulverized cuttings, outcrops, or larger samples, prepare the sample material by selecting representative fragments. Ensure that flat surfaces will be oriented toward the beam for analysis.

4.2 XRF Data Acquisition

Instrumentation

• Use a laboratory XRF analyzer (e.g. wavelength-dispersive or energy-dispersive systems) for high-precision analysis or a pXRF/HHXRF for rapid, on-site measurements.
• Calibrate the XRF instrument with matrix-matched standards specific to sedimentary rocks (e.g. shales, sandstones, carbonates) to account for matrix effects in drill cuttings.

Elemental Analysis

• Measure major elements (e.g. Si, Al, Ca, Mg, Fe, K, Na) and trace elements (e.g. Zr, Sr, Ba, V, Ni) relevant to lithological discrimination in petroleum reservoirs.
• Collect multiple measurements per sample (e.g. 3-5 replicates) to assess variability and improve data reliability.
• Record elemental concentrations in weight percent (wt%) for major elements and parts per million (ppm) for trace elements.

4.3 Data Preprocessing

Quality Control

• Identify and remove outliers caused by contamination (e.g. drilling additives) or instrument errors using statistical tests (e.g. Grubbs’ test) or visual inspection of depth profiles.
• Check for consistency by comparing XRF data with mud logs or wireline logs (e.g. gamma- ray, resistivity) if available

Normalization

• Normalize elemental concentrations to account for dilution effects (e.g. by volatile content or organic matter) or use ratios (e.g. Si/Al, Ca/Mg) to highlight lithological signatures.
• Apply log-transformation to trace element data if distributions are skewed.

Depth Alignment

• Correct for depth inaccuracies due to lag time or mixing in the drilling process, using mud logging data or time-depth correlations.
• Assign a single representative depth to each sample for geostatistical analysis

4.4 Geostatistical Analysis

Exploratory Data Analysis

• Plot elemental concentrations versus depth to identify trends (e.g. increasing Si indicating sandstone, high Ca for carbonates).
• Use multivariate techniques like Principal Component Analysis (PCA) to reduce dimensionality and identify elemental combinations that discriminate lithologies (e.g. shale vs. sandstone vs. limestone).

Variography

• Compute one-dimensional variograms along the well trajectory (depth) for key elements or elemental ratios to quantify spatial continuity.
• Model variograms (e.g. spherical, exponential) to capture vertical lithological transitions, accounting for cyclicity in sedimentary sequences.
• If multiple wells are available, compute 3D variograms to assess lateral continuity, though this is less common with drill cuttings due to sparse spatial coverage.

Interpolation

• Apply ordinary kriging or co-kriging to interpolate elemental concentrations at unsampled depths, producing continuous depth profiles.
• Use co-kriging to leverage correlations between elements (e.g. Si and Zr for quartz-rich sandstones) for improved accuracy.
• Alternatively, use conditional simulation to generate multiple realizations of elemental distributions, capturing uncertainty in lithological boundaries.

4.5 Lithology Classification

Supervised Classification

• Train a machine learning model (e.g. random forest, support vector machine) using XRF data from cuttings with known lithologies (e.g. from core samples or wireline log interpretations).
• Input features include elemental concentrations, ratios, or PCA scores; output classes are lithologies (e.g. shale, sandstone, limestone, dolomite).
• Validate the model using cross-validation or a holdout dataset from adjacent wells.

Unsupervised Clustering

• Apply clustering algorithms (e.g. k-means, hierarchical clustering) to group cuttings based on elemental signatures when lithological labels are unavailable.
• Determine the optimal number of clusters using metrics like silhouette scores or geological knowledge.

Integration with Other Data

• Combine XRF-derived classifications with mud logs, gamma-ray logs, or cuttings descriptions (e.g. color, texture) to refine lithology assignments.
• Correlate elemental patterns with depositional environments (e.g. high Al for clay-rich shales in marine settings, high Sr for evaporitic carbonates).

4.6 Mapping and Interpretation

Depth Profiles

• Create lithology logs by plotting classified lithologies or interpreted elemental concentrations versus depth.
• Highlight formation boundaries, reservoir intervals, or seals based on lithological transitions (e.g. sandstone reservoirs capped by shale).

Reservoir Characterization

• Identify pay zones by mapping lithologies with favorable reservoir properties (e.g. porous sandstones) using elemental proxies (e.g. high Si, low Al).
• Assess seal integrity by detecting impermeable lithologies (e.g. shales with high Al and Fe).

3D Modeling (if applicable)

• For multi-well datasets, integrate lithology logs into a 3D geological model using software like Petrel or GSLIB, interpolating lithologies between wells with geostatistical techniques.
• Use sequential indicator simulation to model categorical lithology distributions, preserving spatial heterogeneity.

4.7 Validation

Ground-Truthing

• Compare XRF-based lithology predictions with core samples, petrographic thin sections, or wireline log interpretations (e.g. density, neutron porosity).
• Use mineralogical data (e.g. XRD) to confirm elemental-lithology correlations (e.g. high Ca linked to calcite in limestones).

Error Quantification

• Assess classification accuracy using confusion matrices or misclassification rates.
• Evaluate interpolation uncertainty by comparing kriging results with held-out samples or alternative datasets.

4.8 Tools and Software

• XRF Analysis: Bruker Tracer, Thermo Fisher Niton, or Olympus Vanta for pXRF / HHXRF; PANalytical or Rigaku for lab XRF.
• Geostatistics: R (gstat, geoR), Python (PyKrige, scikit-learn), or commercial software (Isatis, Surfer).
• Visualization and Modeling: Petrel, ArcGIS, or open-source tools like QGIS for mapping; matplotlib or ggplot2 for depth profiles.

5. Integration with Subsurface Data

XRF enhances multi-disciplinary subsurface evaluation by:

• Complementing petrophysical logs (gamma ray, resistivity)
• Calibrating and constraining geophysical models
• Supporting machine learning applications for reservoir prediction

Data from XRF can be seamlessly integrated into GIS platforms, geological models, and real-time dashboards for decision support.

6. Economic and Operational Benefits

7. Specific Considerations for Petroleum Drill Cuttings

7.1 Challenges

Despite its advantages, XRF does have limitations:

• Cannot detect light elements (e.g. H, He, C, N) critical inorganic geochemistry
• Drill cuttings are often mixed or contaminated, requiring careful cleaning and quality control.
• Accuracy may be affected by sample heterogeneity and surface roughness
• Fine-grained cuttings (e.g. shales) may dominate samples, masking coarser lithologies (e.g. sandstones)
• Requires calibration and matrix corrections for quantitative results
• Depth resolution is coarse compared to wireline logs, limiting precision in thin-bedded sequences.
• Regulatory acceptance varies by region and project

7.2 Solutions

However, several known best practices may help mitigate the XRF limitations, including:

• Using certified standards for calibration
• Using high-frequency sampling (e.g. every 5-10 feet) to capture rapid lithological changes.
• Combining XRF with other analytical methods (e.g. TOC, pyrolysis)
• Integrating XRF with real-time mud logging or drilling parameters (e.g. rate of penetration) to improve depth accuracy.
• Routine instrument validation and maintenance
• Focus on robust elemental ratios (e.g. Si/Al for sandstone vs. shale) less sensitive to contamination.

7.3 Petroleum Relevance

• Identify reservoir lithologies (e.g. porous sandstones) and seals (e.g. shales) to guide completion strategies.
• Detect diagenetic features (e.g. cementation indicated by high Ca or Fe) affecting reservoir quality.
• Map source rock potential by targeting organic-rich shales with elevated trace elements (e.g. V, Ni).

7.4 Ongoing and Future Trends

• Machine Learning Integration: Using AI to correlate XRF data with petrophysical and production outcomes
• Miniaturization and Automation: Drone-mounted XRF or inline sensors for core scanning
• Cloud-Connected Devices: Enabling real-time data sharing and centralized monitoring
• Expanded Element Libraries: Improved software algorithms for complex matrix correction

8. Example Workflow

In a clastic petroleum reservoir:

1. Collect cuttings every 5 feet from a vertical well, wash, and analyze with pXRF/HHXRF for Si, Al, Ca, Fe, Zr, and V.
2. Normalize data, plot Si/Al vs. depth, and observe high Si/Al at 4875’-5200’, suggesting sandstone.
3. Compute a variogram for Si/Al, fit a spherical model, and use kriging to interpolate across the interval.
4. Apply k-means clustering to group cuttings into sandstone (high Si, Zr) and shale (high Al, V) clusters.
5. Validate with gamma-ray logs showing low readings in the sandstone interval.
6. Generate a lithology log highlighting a 325’ thick sandstone reservoir, informing perforation zones.

9. Conclusion

XRF is a valuable, underutilized technology in the exploration drilling process for oil and gas. By offering rapid, reliable elemental data at the wellsite or nearby. It empowers geologists and drilling engineers to make more informed, timely decisions. When integrated with traditional datasets, XRF can significantly reduce costs, improve well placement, and enhance reservoir understanding—contributing to safer and more efficient exploration.