SCA2013-020 1/12

AN INTEGRATED APPROACH TO ESTIMATE

THRESHOLD CAPILLARY PRESSURE FROM CORE AND
LOG DATA

K. Djaouti, P. Egermann, P. Hennebelle
Storengy (a GDF SUEZ company), Bois-Colombes, France

This paper was prepared for presentation at the International Symposium of the Society
of Core Analysts held in Napa Valley, California, USA, 16-19 September, 2013.

ABSTRACT
Threshold capillary pressure (P
cd
) is an important parameter to characterize in the context
of reservoir production, basin modeling and geological storage (natural gas or CO
2
).
Various approaches exist to measure this parameter in the laboratory. For caprocks, their
extremely low permeability values makes it difficult to multiply the number of
measurements in order to directly obtain a large scale picture of the structure confinement
property (path of less capillary resistance). It is therefore very attractive to use P
cd

correlations anchored on properties that can be estimated from the logs to conduct this
type of analysis.
This paper presents an example of such integration methodology in a heterogeneous
caprock. The first part describes how a revisited correlation was derived using HPMI
(High Pressure Mercury Injection) data that were obtained on old, poorly preserved
samples of cores. Corrected P
cd
values (from interfacial tension, IFT) were obtained
directly from the HPMI curve and the associated permeabilities were deduced by
standard estimation methods (Swanson et al., 1981, Thomeer et al., 1983, Kamath et al.,
1992). In the second part, the results were then compared with published correlations
(Thomas et al., 1968, Monicard et al., 1975), with well documented experimental
published data and also with results of new experiments obtained on recently acquired
preserved fresh cores. It has enabled us to set up and strengthen a revisited correlation
which appears to be representative and applicable over a wide range of permeability. The
third part describes how the new correlation was used in combination with relevant logs
to obtain a large scale description of P
cd
.
The main result of this study is a new correlation that integrates rock properties variations
in the estimation of P
cd
. It underlines the importance to integrate cores and logs data,
using correlations anchored on representative core measurements.
INTRODUCTION
Evaluating sealing efficiency of caprocks and maximal gas overpressure are a key criteria
in the selection of a geological structure for underground storage (natural gas or CO
2
).
This leads to assess the maximum injection operating pressure gradient to avoid

SCA2013-020 2/12


migration through the caprock into overlying permeable formations. The threshold
capillary pressure characterizes the ability of a porous medium saturated with a wetting
phase to block the flow of a non-wetting phase. Its value corresponds to the minimum
pressure difference between the two phase to promote an invasion of the non wetting
fluid in the porous medium. It is directly related to the size of the largest pore throat in
the porous medium and can be calculated using the following equation:

Laplace-Young

Where P
cd
is the capillary pressure of a rock with the largest throat radius of the rock
( ), the gas/water interfacial tension () and the contact angle between the
gas/water interface and the rock (). Caprocks have high capillary threshold pressure and
low permeabilities, due to very fine pore and pore-throat sizes.

The purpose of this paper as any reservoir characterization is to define a parameter (P
cd
),
that can be integrated with a geological model to display the caprock properties in three-
dimensional space. To allow this, cores capillary pressure data must be related to other
reservoir rock properties, using wireline logs that are calibrated on core measurements.
The first part of this paper presents the basis of the method and how well logs data and
key core measurements (porosity, permeability) can be combined to obtain a P
cd
log. In
the second part, the methodology is applied in a real case using mercury porosimetry
obtained on old poorly preserved cores to calibrate the model. The P
cd
log calculated
from this approach is then compared with relevant P
cd
data recently obtained on fresh
cores in the third part. This enables us to draw some conclusions about the potential of
the proposed approach to obtain early relevant estimation of confinement properties even
when only old core material is available.

METHODOLOGY
Basis of the method
We provide a methodology, using empirical relationships, to produce a capillary pressure
log in three steps. The technique can be applied to wells that have no measured capillary
pressure but for which core porosities, horizontal (K
h
) and vertical (K
v
) permeabilities
have been measured on field cores. This method provides a direct solution for predicting
vertical and lateral variations in reservoir rock properties that are related to variations in
capillary pressure properties. The following approach was followed (Figure 1).

First, a total porosity log is obtained using available log data for porosity computation
(neutron, density, sonic). Then, core relationships ( vs K
h
and K
h
vs K
v
) are used to
determine horizontal and vertical permeability logs. Finally, a P
cd
log is applied by using
an empirical P
cd
-K
v
relationship.

SCA2013-020 3/12



Figure 1 : workflow to construct a threshold capillary pressure log from well log data and
core measurements.
Various P
cd
- K
v
relationships have been suggested in the literature :

Thomas et al. (1968) (2)
Monicard et al. (1975) (3)
Where P
cd
is the threshold capillary pressure (psi) and K
v
the vertical permeability (mD).

Application made in this study
In this study, we had initially only old poorly preserved cores. Therefore the data needed
to calibrate the relationships were derived from mercury porosimetry measurements.
Porosity and P
cd
(after correction) were obtained directly whereas the permeability was
estimated using available methods in the literature (Thomeer et al., 1983; Swanson et al.,
1981, Kamath et al., 1992). This has made possible to calculate a preliminary P
cd
log on a
recently drilled well where fresh preserved cores were retrieved. This P
cd
log was first
used as a guide to sample the best locations. The new P
cd
measurements obtained from
these samples have then been used to quality control a posteriori the representativity of
the P
cd
log.

P
cd
log from old core data
Core description
At the beginning of the study, existing cores were recovered several decades ago during a
former drilling campaign to assess a potential area. Two types of formation were

SCA2013-020 4/12


represented on the caprock: clay rich silty and bioclastic marlstone (rock facies 1), shown
as the dominant lithology on well log/core data and tight carbonates (rock facies 2).
By examining the thin sections, we can see a lack of apparent macro porosity (larger than
5 microns) on both rock facies, implying dominance of mesoporosity (0.5 to 5 microns)
and microporosity (less than 0.5 microns). Low pore throat sizes with high entry capillary
pressure values are expected with associated low permeabilities. Some areas contain a
significant amount of bioclasts (about 50%) beside clays and the microstructure can be
also quite complex on carbonate samples.

Figure 2 : Thin-sections photomicrograph of two caprock facies, clay-rich silty and
bioclastic marlstone (left) and tight carbonate (right).
Carbonates containing multiple pore systems present challenges to obtain representative
capillary pressure data (Almarzouqi M. et al, 2010). The pore geometry can be highly
variable, with zones containing significant amounts of macroporosity interspersed with
zones dominated by meso or microporosity.

Mercury porosimetry measurements to calibrate relationships
Mercury injection is a relatively quick and low cost technique to obtain capillary pressure
data. Sixty-two old core samples recovered from three old wells were prepared for testing
with high pressure mercury injection (HPMI) (also referenced as mercury injection
capillary pressure (MICP) in the literature), which offers a rapid method of developing
capillary pressure curves to very high pressure. The mercury method also has the
advantage to consume a small volume of rock (often only fragments of rocks were
available). It involves injecting mercury in a step-wise manner into a dry core sample
initially placed under vacuum. The threshold P
c
is obtained from the mercury intrusion
pressure after interfacial tension corrections (IFT). The displacement of the mercury can
therefore be used to calculate the bulk volume of the rock, the porosity and the associated
permeability, deduced from estimation methods (Thomeer et al., 1983; Swanson et al.,
1981, Kamath et al., 1992; Comisky et al, 2007). Some limitations of the mercury
porosimetry method (lack of confining pressure, samples preparation procedure without
alteration of the pore structure...), can make this approach very inaccurate but it is very

SCA2013-020 5/12


adequate to collect early "trends" when no other data under in-situ conditions are
available (Egermann et al., 2006).

Result of HPMI method: P
hi
-K relationship
The evolution of the volume of mercury injected (depending on the pressure) gives direct
access to the entry pressure and the pore volume. Figure 3 (left) shows an example of the
distribution of pore size that can be deduced from one measurement on a tight carbonate
caprock sample. In this case it gives an average size of 0.03 microns which confirms the
very tight nature of the environment. The Figure 3 (right) shows a plot of the cumulative
intrusion curve as a function of pressure for the same carbonate sample. Significant
mercury intrusion occurs only above pressure of 3000 psi.


Figure 3 : Pore throat size (left) from the cumulative mercury intrusion (right) of a tight
carbonate sample.
We chose the Swansons model to derive permeability estimates (Swanson et al., 1981).
The porosity, calculated from the total amount of Hg injected during the measurement
and the permeability estimated on sixty-two samples, provides a fair correlation (Figure
4). Although there is a wide scatter in porosity and permeability data for high values that
can be related to an increasing clay content of badly preserved samples.


Figure 4 : Core porosity and permeability relationship (HPMI measurements).

SCA2013-020 6/12


Result of HPMI method: P
cd
-K
v
relationship
Threshold capillary pressure results from HPMI method were compared with the two
correlations previously presented and commonly used by petroleum industry (because of
solid experimental validation) relating threshold Pc and vertical permeability (Kv)
(Thomas et al., 1968; Monicard, 1975) in Figure 5.


Figure 5 : The revisited P
cd
vs K
v
correlation
At rather higher permeability values, the Monicard law appears to better represent the
data which seems better representative. However, the measurements corresponding to
progressively lower permeability seem better represented by the model of Thomas. This
feature is certainly due to the fact that Thomass law was mainly anchored on very shaly
and very low permeability samples.
A composite model initially anchored on Monicards law, then on Thomass law was
introduced to better capture this trend. The permeability value (K>0.00033 mD) from
which the new correlation deflects from Monicards law to the Thomass law is defined
visually on the plot. The revisited correlation is presented as:


(4)
With the following input parameters:
a=9 and b=0.5 for K>0.00033 mD,
a=100 and b=0.2 otherwise.
The new correlation was also compared with other published data obtained on tight
carbonate study (Carles et al., 2010). It shows a similar trend to what was observed with
our samples (Figure 5).

SCA2013-020 7/12


Core to log integration
Determination of a porosity log is the first important part before following the workflow
proposed in Figure 1. In our study, the porosity is estimated from a combination of
porosity logs, in order to correct for variable lithology effects (clay, silt and carbonate).
The neutron density log combination, commonly used in the industry, is chosen to
determine the porosity by taking an average of the two log readings:

Neutron-density porosity (5)

Where
n
and
d
are neutron and density total porosity.

Downhole logging tools and laboratory techniques are sensitive to different portions of
the pore system. Therefore in practice, available wireline logs for this study are
insufficient to distinguish clay bound water and capillary bound water associated with
shale micropores. So, total porosity is considered and represents an approximation. A
simple crossplot of the two measurements from one well shows in Figure 8 a reasonable
agreement between core (CPOR) and log (PHIT) porosity suggesting HPMI total porosity
is approximately equivalent to the total porosity log derived from neutron and density
tools.


Figure 6: HPMI core porosity versus the total porosity log computed on the same well.
Figure 7 shows the P
cd
log estimated for a new well with the calibrated relationships. A
pseudo image log (FMI type image) has also been generated, using the computed log, to
highlight caprock layers with very good confinement properties. As new fresh preserved
cores were acquired on this well, this a priori P
cd
log was used to steer the sampling
strategy in order to collect laboratory data from the different facies in term of
confinement properties. This has therefore also provided to us a kind of blind test very
useful to assess the added value of the proposed approach to provide default P
cd
data or to
populate P
cd
data in non cored wells.

SCA2013-020 8/12




Figure 7: Example of a threshold P
c
log calculated for a well.

Quality control of P
cd
log using standard method on fresh cores
Experimental Results
There are different methods for determining threshold capillary pressures in laboratory
(Boulin et al, 2012). The most commonly used being:
- Standard method: based on the step by step approach (Al-Bazali et al., 2005). Gas
is in contact with the sample surface at the inlet. Initially gas pressure is equal to
pore pressure. Then gas pressure increases by steps. Each pressure amplitude and
step duration depends on the accuracy required on P
c
and the sample permeability.
When the capillary pressure (gas pressure minus pore pressure) is higher than Pcd,
water is displaced out of the sample. The pump placed downstream provides this
information.
- Dynamic method: gas injection pressure imposed just above the value of the entry
pressure. The displacement pressure is deduced from variations in water

SCA2013-020 9/12


production output (Egermann et al., 2006). This method is fast but requires an a
priori assessment of the value of P
cd
.
- Residual method: although simple to implement, it has been shown that this
method consist in balancing two pressure vessels at the inlet and the outlet
through the sample, leading to an underestimation of the displacement pressure
(Hildenbrand and al., 2002).

As part of this study, the standard method was chosen to determine the values of P
cd
. It
was conducted on seven 1 diameter cylindrical core plugs taken in the vertical
direction of preserved cores from a recent well. A schematic description of the device is
provided in Figure 8. The measurements were made using the following protocol. The
experimental setup was first validated on an old core, well consolidated, tight carbonate
sample (sample 1).

The experimental set up used to perform this experiment is composed of a horizontal cell
to confine the sample, fitted at the outlet with a capillary water filled tube, which
measured the water going out of the sample. The monitoring is based on the progression
of the meniscus located at the air/water interface. This approach provides a very good
sensitivity even with a low water production. The pressure of effective stress (also called
confining pressure), calculated on the side of the sample, is applied to the sample in an
isotropic way. All measurements were made using a net confining pressure. The nitrogen
pressure is applied to the front face of the sample with typical pressure increment of 10
bar. The experiments were usually pushed to a pressure of 60-70 bar with the exception
of the last sample where the maximum applied pressure was 90 bar.


Figure 8 : Schematic of the experimental setup.
Pressure breakthrough is confirmed if and only if: gas is observed at the output of the
sample, the applied input pressure drop substantially, and a slope change is observed on
the cumulative water production. Figure 9 (left) shows, for one sample representing rock
facies one, the different levels of gas pressure imposed in entry (dashed curve), while
water production (pushed by gas) is measured at the output (curve points). Once the gas

SCA2013-020 10/12


breakthrough is established, the gas outlet production line was connected to a returned
separator initially filled with water. This device is very efficient to record accurately
small volumes of gas production, in the order of the cubic centimeter, to measure
effective gas permeability after breakthrough.

During experiment, samples of rock facies one (clay rich silty and bioclastic marlstone)
were proved deformable and subject to slow creeping. This feature is observed on Figure
9 left (1) where water production was observed for a shaly sample without gas detection
at the outlet. This water corresponds to the deformation of the sample under the influence
of the confining pressure (confirmed on some samples by water production recorded in
output without gas pressure at the outlet). Water production after a capillary breakthrough
by gas is characterized by a rapid change in the slope (2) related to the increase of input
pressure and bubbles gas detection at the outlet (P
cd
50 bar).

Figure 9 right shows an example of creeping when the confining pressure was initially
applied on the sample. This plot clearly illustrates a displacement of the meniscus without
any gas pressure applied at the inlet. In term of experimental protocol, it is therefore
recommended to wait for stabilization after change in the confining pressure before
applying gas inlet pressure increments, in order to dissociate the two mechanisms that can
lead to a meniscus displacement.

Figure 9 : Monitoring of input pressure levels and water production of one sample (rock
facies 1), and example of creeping effect (right)).
The main results of the standard method experiments are summarized in Table 1 with the
effective gas permeability (Kgas) in the experiments where gas breakthrough has been
observed. The measurement was carried out with the permanent method, by imposing a
differential gas pressure on the sample and measuring the corresponding flow.
Klinkenberg corrected permeability were also computed, to correct for gas slippage effect
that may dominate, and yield an estimate of the equivalent liquid permeability.

Rock creeping

SCA2013-020 11/12


Results shows good confinement properties of caprock samples with particularly high P
cd

values in the basal part (samples 7 and 8). To obtain a breakthrough, the last experiment
(sample 8) was pushed up to 90 bar, but despite this value, no gas was observed at the
outlet. The associated effective gas permeabilities were found very low (from 0.003 to
0.03 D) which is in good agreement with the expected range of permeability
considering the measured P
cd
values.

Table 1 : Summary of results using standard method (*Old core sample for experimental
setup validation).

Comparison with a priori P
cd
log
A blind test of the representativity of the P
cd
log was obtained using the conventional
(standard method) P
cd
measurements (Table 1). It shows quite a good correlation
especially for intervals not affected by hole washouts (without badhole indication). It is
interesting to note that all the measured points are in good agreement with what was
anticipated from the P
cd
log and especially the large variability of the values. It enables to
confirm from this integrated study that the confinement properties of this caprock results
from both a baseline P
cd
value between 30 and 40 bar and several metrics layers with
extremely good confinement properties at the bottom of the caprock (>100 bar).

CONCLUSIONS
Several conclusions can be drawn from this study. A methodology was proposed to
integrated P
cd
core data and log data to obtain a log of P
cd
value. It was successfully
applied in a real case using mercury porosimetry data obtained on old cores to calibrate
the various relationships. Having such P
cd
log on a new well was very useful to steer the
sampling strategy in new fresh preserved cores in order to investigate in the laboratory
the variability of the caprock in term of confinement property. Finally the comparison a
posteriori of the representative P
cd
measurements from fresh cores with the a priori P
cd

log was very good. This suggests that this approach could consist in a fair estimation
approach to assess the confinement properties of a structure when scarce data are
available.

SCA2013-020 12/12


ACKNOWLEDGEMENTS
The authors acknowledge Storengy for giving its permission to publish this paper, the
members of the Geosciences Department for their support and Poweltec for their
contribution to the experimental work.
REFERENCES
1. Al-Bazali T.M., J., Zhang, M.E. Chenevert, and M.M. Sharma, "Measurement of
sealing capacity of Shale Caprocks," SPE, 2005.
2. Almarzouqi M. I. , Bush I. , Griffiths R. , Husser A. , Jeha Z. , Montaron B. ,
Narhari S. R. , Poiriercoutansais X., resolving carbonate complexity, Oilfield
Review, Vol 22 (2010), Issue 2, p. 40-55.
3. Boulin, P.F., P. Bretonnier, N. Gland, and J.M. Lombard, "Low permeability
measurements with water on clay sample. Contribution of steady state method
compared to transient methods," SCA2010-31, 2010.
4. Carles P., Bachaud P., Lasseur E., Berne P. and Bretonnier P. : Confining
properties of carbonated Dogger caprocks (Parisian Basin) for CO2 storage
purpose, Oil & Gas Science and Technology, vol.65 (2010), n3, pp. 461-472.
5. Comisky J.T., Newsham K.E., Rushing J.A., Blasingame T.A., A comparative
study of capillary-pressure-based empirical models for estimating absolute
permeability in tight gas sands, SPE 110050, 2007.
6. Egermann, P., Lombard, J.-M., and Bretonnier, P., " A fast and accurate method
to measure threshold capillary pressure of caprocks under representative
conditions, SCA2006-07, 2006.
7. Hildenbrand A., Schlmer S., Kroos M.:Gas breakthrough experiments on fine
grained sedimentary rocks, Geofluids, 2002, 2, 3-23.
8. Kamath J. : "Evaluation of accuracy of estimating air permeability from mercury
injection data, SPE Formation evaluation, vol 4 (1992), pp 304-310.
9. Monicard R., Caractristiques des roches rservoirs, Editions Technip , 1975.
10. Purcell, W.R.: "Capillary Pressures-Their Measurement Using Mercury and the
Calculation of Permeability" Trans. AIME, 186 (1949), 39-48.
11. Swanson B. F. : "A simple correlation between permeability and mercury
capillary pressures", JPT, December, pp 2498-2504, 1981.
12. Thomeer J. H. M. : "Introduction of a pore geometrical factor defined by the
capillary pressure curve", Trans AIME, March, pp 73-77, 1960.
13. Thomeer J. H. M. : "Air permeability as a function of three pore network
parameters", Trans AIME, April, pp 809-814, 1983.
14. Thomas, L.K., Katz, D.L. and Tek, M.R. :Threshold pressure phenomena in
porous media, Trans AIME, Vol.243, 1968.