By Aymen Alramadhan and Dr. Yildiray Cinar
Reservoir Description & Simulation Department, Saudi Aramco


The ability of an oil reservoir to hold oil and water within its pore structure is called capillary pressure (Pc) and it is a function of interfacial tension, capillary sizes (conduits created between the grains of the rock), wettability, rock texture, density difference, and formation saturation. The capillarity does not only control the initial distribution of fluids (oil/water/gas) within a given reservoir but it also drives hydrocarbon recovery processes through the interaction of the three forces: capillary imbibition, viscous and gravity forces. Therefore, accurate assessment of hydrocarbon in-place and recovery is directly linked to an accurate knowledge of capillary pressures for different rock types within a reservoir.

Due to the complex heterogeneity in carbonate reservoirs and its strong impact on Pc , there is always a need for reliable and accurate laboratory measurements to assess Pc at core scale. There are three standard laboratory tests for Pc assessment: multi-speed centrifuge tests, semi-permeable membrane (porous plate) tests, and mercury injection capillary pressure (MICP) tests.

To the best of our knowledge, there is no such study reported in the open literature that examines variation in derived imbibition capillary pressure (Pci) from a given multi-speed centrifuge test using different interpretation methods similar to the Society of Core Analysits (SCA) study. Part of this study looks at different models in transforming imbibition centrifuge data to obtain a representative Pci to integrate with the MICP and Quantitative Evaluation of Minerals by Scanning Electron Microscopy (QEMSCAN) data.


Experimental Data. A large number of data sets from a carbonate reservoir are examined in this study. Each data set contains end-trim petrographical data (thin section, scanning electron microscopy, QEMSCAN), conventional core data (porosity, permeability, and grain density), multispeed centrifuge data, and MICP data. The rock samples used in this study cover a wide range of porosity, permeability, and rock types.

Mineral and Textural Characterization Using QEMSCAN. QEMSCAN analysis was performed on end-trims for a detailed mineralogical imaging and textural and mineralogical characterization. For each sample, QEMSCAN outputs include:

  • Quantitative bulk mineralogical abundance data (expressed as both mass and area %) 
  • Mean mineral size data for each reported mineral 
  • Calculated grain and bulk density data 
  • Estimated macroporosity data 
  • Mineralogical images (i.e., mineral maps)

Conventional Core Analysis. Reservoir plugs of 1 in. in diameter and approximately 1 in. in length were used for conventional core analysis, centrifuge, and MICP tests. Plugs were cleaned as per the approach proposed by Masalmeh and Jing (2007) and dried in an oven at 105C. Porosity, permeability, and apparent grain density were obtained according to API-RP40 (1998).

Multispeed Centrifuge Data. After conventional core analysis tests, core plugs were saturated under vacuum with a synthetic brine of 200 k ppm. Then, the plugs were put in centrifuge for draining brine against air at a Pc of 130 psi to establish an immobile (connate) water saturation. Air was replaced by dead crude oil taken from the field (0.89 g/cm3 and 7 cp at 60C), and several pore volumes of crude oil were flushed through the plugs. Then, to restore wettability, the plugs were immersed in crude oil for static aging at 60C for at least 4 weeks. Although no wettability tests were conducted on the aged plugs, a weak oil-wettability is expected in the plugs based on the previous wettability studies on the samples from this reservoir. Then, several pore volumes of crude oil were flushed through the plugs to establish the initial conditions before the centrifuge tests.

MICP Data. After centrifuge tests, the plugs were cleaned, dried, and then subjected to MICP tests up to 60,000 psi to determine the pore-throat size distribution (PTSD) and to classify samples based on modality and pore-throat radii. The number of peaks on PTSD determines the modality of the rock sample (one peak represents a monomodal pore system, two peaks a bimodal system, and three peaks a trimodal). These data sets make it possible to directly link Pci end-point to the pore systems.

Linking Centrifuge with MICP and QEMSCAN Data. To assess the effect of the microporosity, porosity modality, and mineral content on Pci, we attempt to link Pci end-point to MICP and QEMSCAN data. We do this in three stages: First, a representative Pci end-point from the solution space (six solutions for each centrifuge Pci test) is determined. Second, the impact of pore systems and micro-macro interaction on Pci end-point is explored. Third, the impact of variation in mineralogy on Pci end-point is investigated. The first stage involves the assessment of numerical solutions transformed through history matching the smoothed and equilibration-corrected experimental data for representativeness. For analytical solutions, Pci end-point obtained from two formulations are compared using crossplots and deviations from the 45o line. Representative Pci end-point solutions are averaged at a reference negative Pci (–30 psi) to study its variation across different samples. To address how pore systems and micro-macro interaction control Pci end-point, diagnostic plots linking variation in properties and internal pore structure of the cores to variation in Pci end-point are constructed for investigation. To address how the variation in mineralogy influences Pci end-point, samples from the previous analysis are clustered based on the mineralogy data from QEMSCAN to determine outliers. Results are detailed in the next section.

Results and Discussion

The QEMSCAN data shows that calcite is the most abundant mineral within the examined reservoir with several samples of a large dolomite content. All other minerals occur in minor to trace quantities.

A qualitative assessment of the end-point saturation of Pci curve (Pci end-point) vs. MICP of bi-modal samples demonstrates that the degree of communication between micro- and macropores has some control on Pci end-point. Figure 1 shows the pore-throat size distribution and Pci curves. The medium-permeability samples have a better communication between micro- and macropores (evident with samples’ higher fraction of intermediate pores between micro- and macropore peaks) compared to the high-permeability samples that have a relatively lower amount of intermediate pores. The better micro-macro communication for the medium permeability samples is thought to be responsible for higher Pci end-point (which indicates a better recovery efficiency) in comparison to lower Pci end-point for the high-permeability samples.

Another observation from Figure 1 is that the peaks of micro- and macropores are closer for the medium permeability samples than those for the high-permeability samples. This might be another factor for a better imbibition recovery efficiency of the medium permeability samples.

Significance of the communication between the micropore and macropore systems.

Figure 1

A search for a relation between Pci end-point and many other parameters including porosity, permeability, initial water saturation, dolomite content, and the ratio of peak micro- to macropore radii (rm/rM) shows a weak correlation.

An in-depth investigation reveals that samples with high dolomite, high heavy mineral, and high microporous calcite form outliers on Pci end-point vs rm/rM plots. With outliers removed, a better correlation is achieved for bi-modal rock (Figure 2).

Correlation between Pci end-point and micro to macro pore throat ratio with the outliers removed.

Figure 2


Acquiring Pci, QEMSCAN, and MICP data on the same rock sample adds value in assessing carbonate waterflood behavior. The study shows that, for multiporosity carbonate rock, the connectivity between micro- and macropores appears to affect the efficiency of oil recovery significantly. The connectivity depends on the amount of intermediate pores between micro- and macropores, and how close the peak radii of micro- and macropores.

The QEMSCAN data shows that bimodal carbonate samples without dolomite, heavy minerals (rutile/anatase) or microporous calcite exhibit a better correlation between Pci end-point and rm/rM. We recommend that this information should be considered in carbonate rock typing.

Leave a comment

Your email address will not be published. Required fields are marked *