Copyright 2006, Society of Petroleum Engineers

This paper was prepared for presentation at the 2006 SPE Asia Pacific Oil & Gas Conference
and Exhibition held in Adelaide, Australia, 1113 September 2006.

This paper was selected for presentation by an SPE Program Committee following review of
information contained in an abstract submitted by the author(s). Contents of the paper, as
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at
SPE meetings are subject to publication review by Editorial Committees of the Society of
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
for commercial purposes without the written consent of the Society of Petroleum Engineers is
prohibited. Permission to reproduce in print is restricted to an abstract of not more than
300 words; illustrations may not be copied. The abstract must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.

Abstract
PVT fluid properties are an integral part of determining the
ultimate oil recovery and characterization of a reservoir, and
are a vital tool in our attempts to enhance the reservoir's
productive capability. However, as the experimental
procedures to obtain these are time consuming and expensive,
they are often based on analyses of a few reservoir fluid
samples, which are then applied to the entire reservoir.
Therefore, it is of utmost importance to ensure that
representative samples are taken, as they are fundamental to
the reliability and accuracy of a study.
Critical to the successful sampling of a reservoir fluid is
the correct employment of sampling procedures and well
conditioning prior to and during sampling. There are two
general methods of sampling; surface and subsurface
sampling. However, within these there exist different methods
that can be more applicable to a particular type of reservoir
fluid than to another. Further to this, well conditioning can
differ depending on the type of reservoir fluid. Sampling
methods for each reservoir type will be discussed with an
emphasis on scenarios where difficulties arise, such as near
critical reservoir fluids and saturated reservoirs. Methods
including single-phase sampling and iso-kinetic sampling,
which have been used increasingly in the last decade, will also
be discussed with some detail as will preserving the
representativeness of other components in the sample
including asphaltenes, mercury and sulphur compounds.
The paper presents a discussion aimed at better
understanding the methods available, concepts behind the
methods, well conditioning and problems involved in
obtaining representative fluid samples.




Introduction
Reservoir fluid samples are obtained for a number of reasons
including:
PVT analysis for subsequent engineering calculations.
Determination of the components that exist in a
particular reservoir to have an understanding of the
economic value of the fluid.
To know the content of certain components that exist in
the reservoir fluid for further planning and future
drilling programs, such as the content of sulphur
compounds and carbon dioxide and the corrosiveness
of the fluid. This will impact the material used for
casing, tubing and surface equipment that may be
necessary.
To obtain a knowledge of the fluids ability to flow
through production tubing, pipelines etc. and possible
problems that may arise because of viscosity changes
due to precipitation of solids such as wax and/or of
asphaltene.
To determine the contaminating components that affect
plant design such as the mercury content, sulphur
components and radioactive components.

Mostly the samples are required to obtain a better
knowledge of a combination of these effects, however, it must
be kept in mind that often the sample is not required to solve
all of these.
The properties of fluid samples from oil and gas reservoirs,
at original conditions and as found at different stages during
production, have been of great interest for as long as the oil
industry has had reservoir engineers. The quality of these
samples is of utmost importance to make the right reservoir
management decisions and the constantly improving
compositional data is helping the downstream industry with
planning facilities. Further improvement of more detailed fluid
properties is being expected and the fact that many of the
specialists in this field are being renamed flow assurance
experts highlights the downstreams need for good quality
compositional, fluid property and phase behaviour data.
Bottom-hole sampling goes back more than 60 years
1, 2

when most operators came up with their own design of
bottom-hole samplers. These were either flow through
samplers, purging fluid constantly or previously evacuated
chambers, that when activated by clocks or other mechanical
devices trapped downhole fluid samples.

SPE 101037
Reservoir Fluid Sampling RevisitedA Practical Perspective
J. Bon, SPE, and H.K. Sarma, SPE, U. of Adelaide; J.T. Rodrigues, SPE, Santos Ltd.; and J.G. Bon, Petrolab Pty. Ltd.
2 SPE 101037
There was more consensus in surface sampling and most
everyone agreed on the way to sample from separators and in
some cases from the wellhead. From very early on the
engineers were aware of occasional carry over of entrained
liquid droplets in the gas phase and early attempts of iso-
kinetic sampling were initially superseded by improving the
separation properties of the separators, increasing their size
and/or lowering the flow rates to allow more time for
equilibrium
2
.
The biggest challenge lies in improving the quality of the
ever more popular open hole wireline samples. These samples
were completely discredited by the fluid experts in the early
days because of conditioning constraints and poor quality of
samples taken with the tools used then. However, since then
the technology has been greatly improved and these samples
are of irreplaceable importance today.
Since then all these issues have been addressed again and
changes and improvements are being made wherever possible
and will continue to be ongoing.
Representative samples taken with typical single-phase
bottom-hole samplers are often taken for reservoir fluid
composition, PVT analysis and, depending on the reservoir,
asphaltene studies. However, these tools are often not suitable
for measuring the content of sulphur compounds or mercury
content. A good knowledge of the wax content and problems
that may arise due to wax can be obtained from studies
performed on stock tank oil. Sulphur compounds and mercury
are mostly present in the gas stream so analysis can be
performed on separator or wellhead gas streams. While ideally
all the parameters could be obtained from the same sample, in
practice different samples can suit different needs and first and
foremost, the need for the sample should be established so the
correct samples can be taken. This can avoid extra costs of
sampling for parameters that will not be measured, or the cost
of having to go back and sample again. But key to a successful
sampling program is to first answer the question: What are
you sampling for?.
This paper aims at discussing different sampling methods
and what they are most suited for. The focus is on obtaining
representative samples of live reservoir fluid for the aim of
characterizing the reservoir production, however also touches
on some of the downstream issues.

Understanding our Reservoir Processes
When a well is drilled it is typical to for some drilling fluid to
enter the pore spaces in the near well bore vicinity. Further to
this the well bore itself is full of drilling fluid. Therefore, wells
first go online with a cleanup flow period of reasonably hard
flow to push these fluids out of the near well bore region and
the well bore
1, 3-7
.
The region in the reservoir from the well bore at bottom-
hole flowing pressure to the nearest point away from the well
bore at initial reservoir pressure is termed the drainage radius.
Depending on how under saturated the reservoir is (sometimes
not at all) the pressure drop induced to create the flow may
result in a draw down of pressure below the saturation
pressure. This causes either dissolved gas or condensate to
come out of solution the subsequent effects are discussed
further for each reservoir type.

Undersaturated Oil Reservoirs
In principle, sampling undersaturated oil reservoirs should be
the simplest. Ideally, if the reservoir is greatly undersaturated,
a drawdown never pulls the bottom-hole flowing pressure
below the saturation pressure, and thus the well bore will have
virgin reservoir fluid flowing through it. In attempting to clean
the well of other fluids, the large flow rate can result in a draw
down below the saturation pressure. If for this or any other
reason the reservoir is drawn down below the bubble point
pressure then two-phase fluid needs to be driven out with
virgin reservoir fluid. This can be done by flowing at a lower
rate hence minimizing the drawdown so flowing bottom hole
pressure is still above the saturation pressure. Initially, due to
capillary pressure effects, released solution gas will remain
immobile in the pore spaces until a critical gas saturation is
achieved, and after that an excess of gas may flow. At the
surface, monitoring the GOR can indicate when the reservoir
is producing virgin reservoir fluid. Initially the GOR will be
too low. A rate should be sought such that a stable GOR is
attained after prolonged flow. At the correct flow rate, the
fluid entering the well bore should be single-phase and the
best samples will be collected with bottom-hole samplers.

Saturated Oil Reservoirs
Sampling saturated reservoirs is not simple since flowing the
reservoir results in a draw down which will pull the bottom
hole flowing pressure below the bubble point pressure
1, 6-8
. As
mentioned before, this results in the release of solution gas
from the reservoir fluid.
Shutting in the reservoir to build-up pressure again will re-
dissolve gas into the oil
7
. Obtaining a stable GOR is more
difficult for saturated oil reservoirs. However, by flowing at a
lower rate a stabilized flow can eventually be reached and a
constant GOR can be observed. The fluid, although flowing in
two-phases, carries both the liquid and gas at the same GOR of
the virgin reservoir fluid.
Bottom-hole samples are not recommended as they may
collect a disproportionate amount of liquid or gas and samples
should be taken at the separator
7
. However, if it is the only
option, they should be taken at a trickle flow so the drawdown
is small, the solution gas in the pores is undisturbed and the
liquid carries as much as possible gas still in solution. If a gas-
cap exists the pressure at the gas-oil contact (GOC) should be
used as the saturation pressure for the reservoir fluid if
correction is required.
If considerable water is flowing together with the oil, the
flow may not be great enough to push the water from the well
bore. To avoid sampling water a static gradient should be run
prior to sampling.

Alternatives:
If the samples attained have a bubble point pressure above the
reservoir pressure, a representative sample can still be
attained. Once in the lab the sample can be charged into a PVT
cell, equilibrium established at the reservoir pressure, and the
excess gas blown off leaving a sample saturated at the
reservoir pressure.
If the samples obtained have a low GOR and separator gas
or wellhead gas is available, additional gas can be added to
raise the GOR to reservoir pressure. This results in a slightly
SPE 101037 3
different composition as the gas will have a different
composition (lighter) than the solution gas. However, the
difference in composition and subsequent PVT parameters
measured should be small and hence the PVT properties
obtained from a study should be within an acceptable error
margin. This of course depends on how much gas is required
to increase the pressure and therefore how far off the virgin
reservoir fluid composition the final recombined fluid will be.
The most extreme case would be using stock tank oil and gas
from the reservoir. An equation of state (EOS) or correlations
can be used to help calculate the volume of gas required to
raise the bubble point pressure.

Gas-Condensate Reservoirs
Similarly to an undersaturated oil reservoir, if a gas
condensate reservoir is undersaturated, provided the well is
not drawn down below the dew point pressure condensate will
stay in solution.
If the reservoir fluid is saturated or reservoir pressure is
near the dew point pressure, a flow period resulting in a
drawdown will inevitably result in condensate coming out of
solution. Similarly to saturated oil, the condensate that initially
comes out of solution will fill pore space until a critical
saturation is reached. This fluid needs to be flushed out with
virgin reservoir fluid. After this, condensate will flow with the
gas through the well bore. For surface sampling, another
problem arises; the flow rate must be large enough to carry the
condensate to the surface, but if the flow is too high stable
flow will not be achieved and perhaps even poor separation at
the separator because the fluid has little time to reach
equilibrium.
The condensate dropout in the near well bore region has
other effects, particularly in low permeability reservoirs; as
condensate drops out and fills pore spaces, the permeability is
reduced, sometimes severely affecting flow.
Reducing the flow rate results in and increase in the
bottom-hole flowing pressure and the pressure in the drainage
radius. This causes some of the condensate in the pore spaces
to re-vaporise decreasing the GOR below that of the virgin
reservoir fluid. After a rate reduction a period of days or even
months for low permeability reservoirs must be waited prior to
sampling
3
.
If the collected sample has too high a dew point and CGR,
a representative sample can still be obtained. Similarly to
saturated oil reservoirs, this is done by establishing
equilibrium in PVT cell at reservoir conditions. Once
equilibrium is achieved and the excess condensate is in the
liquid phase the gas phase is sub-sampled from the PVT cell.
This process yields a gas sample which is saturated at the
reservoir conditions.

Well Conditioning
The general recommendation for oil reservoirs conditioning is
to alternate between flowing periods and shut-in periods with
successively lower rates. This is continued until a stable GOR
is established.

For conditioning of reservoirs for conventional sampling (no
asphaltenes) first draw a phase envelope, then:

1. Obtain only fluid that has not been inside the two phase
region.
2. If that is not possible obtain fluid after re-dissolving
freed solution gas or condensate (second best).

For reservoirs that may have asphaltene problems, draw a
phase envelope with the Onset of Asphaltene Precipitation
(OAP) envelope
9
. Bear in mind that some studies have shown
this curve to be decreasing with increasing temp while others
show it to be increasing with increasing temp:

1. Obtain only fluid that has not been below the OAP line
and maintain the sample in single-phase.
2. If that is not possible, fluid sampled needs to
reconditioned to become same as fluid above OAP line.

There is no consensus on the method to recondition
samples with asphaltenes that have been below the OAP or
even if it can be done. However, it is often achieved or
attempted to be achieved by a prolonged period (a week) of
maintaining the sample at reservoir pressure and temperature
or above and inverting the sample periodically.
So, if possible, samples are taken at ideal conditions
(single-phase), but if this is not possible, we have to accept
this fact and endeavor to take whatever we can and recombine
or recondition the sample to get the best possible and most
representative sample.
For gas condensate reservoirs, the well is produced at
successively higher rates. Retrograde liquid dropout will occur
once the reservoir is drawn down in two phases. This liquid
will accumulate in the near well bore region but eventually
with continued flow the fluid entering those pore spaces and
leaving will be have a constant CGR. Samples should be taken
at the lowest rate that produces a stable CGR.
It is globally accepted that sampling is best achieved as
early as possible in the life of a reservoir.
Generally speaking conditioning is achieved at equilibrium
situations usually, but not always at low flow rates. The
conditioning process should preferably be monitored to ensure
its progress. Due to time and cost restraints, assumptions are
usually made as to its completion and if sampled a different
way as well, comparisons can support assumptions.
For slickline bottom-hole sampling, pressure surveys for
both absolute pressure and gradients resulting in depth of fluid
levels are important monitoring tools.
Additionally the monitoring of the mole weight or the C
7
+

fraction of produced gases, can improve the quality of these
samples
3
.
The conditioning principle extends into conditioning or
cleaning / purging of the well bore, separators, sampling lines,
equipment and cylinders of all foreign material.
The more on-site checks to evaluate the overall quality of
sampled material are made the easier it is to decide to change
the conditioning process for improved or repeated sampling.

Reservoir Fluid Sampling Methods
Reservoir fluid sampling techniques can be generalized into
two groups sub-surface and surface.


4 SPE 101037
Sub-surface Sampling Techniques
1. Slickline run tools usually in cased holes using
conventional flow through or pistonned samplers. This
method is recommended for normal undersaturated oil
and gas reservoir fluids
2. Slickline run tools, again, usually in cased holes with
pressure compensated single phase fluid samplers. This
is recommended for undersaturated oil and gas reservoir
fluids with possible solids (wax / asphaltenes)
precipitation problems.
3. Exothermic samplers: These samplers not only keep the
sample pressure compensated but also temperature
compensated to further reduce any chance of solids
precipitation prior to sample transfer and analysis.
4. Wireline Formation Testers run with electrical line, tied
into the logging string, usually taken in open holes prior
to casing being set. These include the proprietary
samplers such as the MDT, RDT, RCI and RFT tools.
There are only subtle differences in tools from different
companies and these types of tools will be referred to
throughout the paper as Wireline Formation Testers
(WFT).
5. Pressure compensated WFT.
6. Samples taken during open hole drill stem testing (DST).
7. Samplers taken with a sub carried in the drillstring
accommodating the samplers described in the above
described points 1, 2 and 3.

Surface Sampling Techniques
1. Separator samples, from test or production separators.
Often recommended for gas condensate or saturated
black and volatile oil reservoir fluids. The equilibrium
separator liquid and gas are recombined in their
produced ratio to obtain representative reservoir fluid
samples.
2. Wellhead samples, for undersaturated fluids still in
single phase at well head conditions.
3. Split-phase sampling or iso-kinetic sampling scaled
down reconditioned homogeneous two phase flow
through a mini-separator.
4. From pipelines or plant flow lines.

Whatever way we choose to take reservoir fluid samples, it
is essential to ensure that the actual fluid obtained is as close
as possible to the virgin fluid discovered in the reservoir to
ensure it is as representative as possible. To do that, all wells
usually have to be conditioned to bring such fluid into the well
bore or separator. Since this process is more complicated than
it sounds it is recommended to take samples in more than one
way, to therefore obtain differently conditioned fluids from the
same reservoir to better define its properties.
The advantage of awareness of all new developments,
should also allow us to evaluate simpler situations with
simpler sampling solutions while at the same time provide
solutions for more complex problems when required.

Surface Sampling
Separator Sampling
A good guideline to collecting good separator sampling is to
flow the well through the separator at a stable pressure,
temperature and GOR for two to three separator volumes to
ensure that fluid collected in the separator at any other
equilibrium state is flushed out. At equilibrium, at any
pressure and temperature (within the two-phase conditions)
the gas and liquid coexist in equilibrium with most
components coexisting in both phases. If either pressure or
temperature is changed drastically, some of the component
that exists in the gas phase may condense or conversely, some
of the component that exists in the liquid phase may vaporise.
When the samples are taken the pressure and temperature of
the separator are recorded as well as, as accurately as possible,
the gas and liquid flow rates. Furthermore, the GOR should be
based on oil flow rate at separator conditions and the
correction factors used should be noted
10
. No matter how good
the sampling and well conditioning are, if the rates measured
and subsequent GOR are incorrect, the recombined sample
will be unrepresentative.
Separator samples can be taken by the following methods,
for gases:
Sampled directly into an evacuated cylinder (preferred
method)
If evacuated cylinders are not available, the air in the
cylinder can be removed by dilution. Filling the bottle
and emptying prior to sampling.
For liquids:
Brine displacement (preferred method)
Evacuated cylinder and sample is flashed into cylinder
Separator gas displacement. Sample cylinder is first
filled with separator gas, this is then displaced as the
fluid enters the bottle (from the bottom)

Wellhead Sampling
Sampling at the wellhead is only done if it is known that the
wellhead pressure and temperature is above the reservoir
fluids saturation pressure. Dry gases can be sampled at the
wellhead as well as some lean gases and low GOR oils.
However if the sample is flowing in two-phases at the
wellhead then an unrepresentative sample will be taken as the
two individual phases will not flow into the cylinder in the
same ratio as they exist in the reservoir.

Split-Phase Sampling
Isokinetic or split-phase sampling can be used to sample two-
phase well-head fluid or to measure the volume of entrained
liquid in a gas stream
11-13
. Once a stable GOR is established,
the two-phase fluid mixed using a mixing head. The mixing
head is designed to homogeneously spread the liquid droplets
carried through with the vapour. A probe then collects the gas
and liquid sample at the corresponding GOR. An online
laboratory ensures that GOR is stable prior to sampling and a
mini-separator separates the fluid into the correct ratio.
Split-phase sampling is applicable particularly for
condensate reservoirs that produce at low rates with low CGR.
It is also applicable for separator gas streams with a large
volume of entrained liquid to correct the CGR and improve the
separator test data; in this case it is not replacing the separator
test but improving the data. However, this method of sampling
is not always economic and also not always better than
simultaneously obtained separator samples.
SPE 101037 5
Split-phase sampling technology is over 60 years old
2
,
however in recent years has been used increasingly. The
technology has improved, particularly in the measurement of
GOR as it is distributed through the cross section of the pipe
(not actually homogenous). Previously it was assumed that the
mixing head perfectly distributed the liquid droplets across the
pipe cross section. This assumption was found to be not
exactly correct, so to improve the GOR measurement and
minimize the error due to this effect, the probe now samples at
three different points across the pipe and averages the result
13
.
A simple alternative to needing to measure the volume of
liquid carry over is to flow at a lower rate, if possible, so that
the fluid has more time to reach equilibrium in the separator
2
.
Another simple alternative is to take the separator gas
samples as normally done. These, will bring the entrained
liquid along with them. Once in the lab, the gas samples are
returned to separator conditions by heating to separator
temperature. The cylinder is placed vertically upside down
allowing the entrained liquid to dropout and flow to the
bottom of the cylinder. This liquid can then be removed,
analyzed and the volume measured and then both
mathematically and physically added to the total well stream
composition.

Bottom-Hole Sampling Tools
Wireline Formation Tools
Representative samples are often taken with Wireline
Formation Tools (WFT)
14
. If this is the case the added expense
of bottom-hole samples later can be avoided. So, rather than
assuming the samples to be bad, they should be assumed good
and quality checked prior to a decision being made on their
validity. However, if there is any doubt with respect to how
representative the sample is, samples should be taken during a
Drill Stem Test (DST) as not having a solid grasp on the PVT
properties of the reservoir fluid may have much larger
repercussions over the long term. WFT are run open hole and
insert a probe into the formation. The fluid flows into the
chamber at a low drawdown. The sample taken at this stage is
completely virgin prior to any flow and drawdown in the
reservoir. More modern WFT can also supply back pressure
and maintain the sample in single phase.
For gas-condensate reservoirs, the sample volume
collected may be too small for good analysis on the reservoir
fluid.
The WFT has a limited capacity to clear out the mud
filtrate from the formation. More recent WFT tools have a
pump out option, so fluid initially pumped from the reservoir
can be dumped into the well rather than collected. If Oil Based
Mud (OBM) has been used, the samples will be contaminated
limiting the accuracy on subsequent analysis performed on the
sample.
WFT have the key advantage that they are taken before
any other form of sampling can be taken. This is particularly
vital for saturated reservoirs.
However, due to the draw down to which the samples are
exposed to, the samples may not be of good quality. This is
particularly an issue for gas condensates as the dew point
pressure is non-monotonic it does not constantly increase
with increasing pressure and temperature, it increases and then
decreases after the cricondenbar. Thus, an unrepresentative
sample may have a dew point pressure that is higher, lower or
equal to the actual dew point pressure
2
.
Also, the samplers often return to surface with drilling
fluid and little sample.
Improved downhole sensors to monitor other fluid
properties could improve the sample quality, unfortunately, a
pressure draw down cannot be eliminated but only minimized.
Remaining convinced of future improvements, we believe
the present position of these tools is quite acceptable and more
often than not, samples obtained this way give reliable fluid
properties data, both for up- and down-stream petroleum
industry. These improvements will be especially needed for
problem fluids from which solid material precipitates with
changes in temperature and pressure as the reservoir is
exposed to less change at the WFT sampling stage.
In stead of bringing up and subjecting them to changes, it
may be worth to combine the race for better MDT / RCI /
RDT sensors with looking at the down-hole samples in-situ
and simulate a down-hole depletion at different depths
(temperatures) to better evaluate the solid phase envelope part.
To better evaluate the contaminants of interest to the
production engineers, down-hole sensors have been developed
to give indications but are improving in accuracy.

Flow-Through Samplers / Conventional Bottom-Hole
Samplers
The flow-through type bottom hole samplers are the simplest
and oldest bottom-hole sampling tool. The tool is run into the
well on a wireline with the valves open. As it travels through
the well bore, the fluid present in the well bore flows through
the sampler. When it reaches the sampling depth the valves
close by some mechanism, be it mechanical, electrical or
triggered by a clock. The sampler is then retrieved and the
sample is transferred.
When the sample is retrieved, the temperature decreases
and therefore the pressure of the sample decreases, for oil
samples this almost always draws the sample into two phases.
At the surface the solution gas must be dissolved back into the
liquid prior to transferring the sample to a cylinder. The old
fashioned method pressurised the oil with mercury, however,
due to occupational health and safety measures, this is
seldom/not used anymore. The great advantage of mercury is
that accurate validity checks can be done, and as a mixing
mechanism when returning the sample to single phase mercury
is unsurpassable. The alternative is to transfer the sample back
in the lab this is not good practice as you are then sampling
blind and should not be done unless absolutely necessary. The
other option is to use acidified brine, but due to the low
difference in density with oil, it is not great as a mixing
mechanism and the sample cannot be analysed for its water
content. In order to avoid the use of mercury in the field and
the issues that using brine may incur, samplers with pistons
were designed.
Quality checks should always be performed on site for
bottom-hole samples, be it only an opening pressure. These
important quality checks cannot be performed on single-phase
bottom-hole samples hence the need for more sophisticated
on-site analysis such as compositions and saturation pressures
in an on-site lab. Otherwise, the quality of the single-phase
sample relies solely on the opening pressure of the tool.
6 SPE 101037
Flow through samplers will not work well for heavy and
extra heavy oil (API < 10) samples and since piston slickline
bottom-hole samples need to be brought down on sucker rods,
the extra long delay in clock setting and tool filling up justifies
planning bottom hole sampling of these fluids only by WFT.

Pistonned Samplers
Pistonned samplers are in principle the same as flow through
samplers. However, the sample chamber has a piston so that
sample transfer into a cylinder for transportation can be done
without the use of mercury or brine. Instead the piston is
pushed with glycol to displace the reservoir fluid from the
sample chamber. However, to re-establish the sample into the
original single phase, some sort of mixing mechanism should
exist in the sample chamber to speed up equilibrium and
ensure that all solution gas, condensate and/or wax is re-
dissolved.
The tool is run in the hole closed and triggered open once
at depth. The sample is collected, the valves closed again and
the sampler is then retrieved.
Since mercury is rarely being used these days, most liquid
sample cylinders and pistonned sampling chambers have
become overly complicated. An array of O-rings
accommodating pistons have made these expensive appliances
unsuitable for long-term storage through interchanging of
fluids across the seals through osmosis / adsorption. If
conventional cylinders are not acceptable, alternative sample
cylinders, using inert diaphragms have to be considered for
long-term storage.

Single-Phase Samplers
Single-phase samplers were designed for sampling of
reservoirs which may have asphaltene problems, to keep the
asphaltenes in solution so they can be studied in a laboratory
15
.
Since asphaltene precipitation may not be reversible, the idea
is to not allow the asphaltenes to come out of solution. They
supply a back-pressure against a piston in the sample chamber
with nitrogen, maintaining the sample well above reservoir
pressure. By over pressuring, the idea is that even when the
sample cools off, it will remain above the conditions for onset
of asphaltene precipitation.
Too many expensive single phase samples are taken in
reservoirs with no solid precipitation problems, and this is
something that perhaps should be addressed rather than
accepted.
An occupational health and safety issue arises for high
pressure reservoirs. Reservoirs with pressures around the
15,000psi mark, would in principle require the nitrogen
chamber to be charged to 17,000psi or more. These pressures
begin to be very dangerous to operate and are at the upper
limit of the equipment rating. Charging the nitrogen chamber
with a lower pressure is safer however will not overpressure
the sample and the sample may still be drawn below the OAP.
This is where exothermic samplers will help as the
temperature compensation will also help keep the sample
away from the OAP.

Exothermic Samplers
As for single-phase samplers, exothermic samplers were
designed for sampling of reservoirs which may have
asphaltene problems, to keep the asphaltenes in solution so
they can be studied in a laboratory. However, further to only
pressure compensation, the exothermic sampling tool also
maintains the temperature high. The single-phase sampler
avoids the issue of the temperature dependence of asphaltene
precipitation by keeping the sample over pressured. However,
temperature also affects asphaltene precipitation and therefore
the exothermic sampler was designed. Problems have been
found with the temperature maintenance and the samplers do
cool off however, the temperature is kept higher than that of
any other tool. Once at the surface the sample is kept hot using
a battery operated heating jacket and the samples taken back to
the laboratory immediately where they are transferred hot and
kept at reservoir pressure and temperature or above. Further
development is still required and ongoing to perfect the tool.

Other Components of Interest
Asphaltenes
If sampling a reservoir and the engineer is very confident that
asphaltenes will not be a problem and no study is intended to
be done for solids, then single phase or exothermic sampling
tools are surplus.
However, in fields where asphaltene problems may arise
and the samples are intended for asphaltene analysis, the
samples should be taken single-phase so to preserve the
asphaltenes in solution for subsequent analysis. Although the
sampling and analysis of asphaltenes are costly, a prior
knowledge of the problems the field may have can be of
economic benefit in the long run. However, if analysis has
already been done and/or no asphaltene analysis is intended
for the samples, single-phase sampling may not be required.
Asphaltene problems should also be kept in mind in
particularly if gas injection at some stage in the reservoirs life
is intended. It may be too late to take representative samples
later.
Analysis for asphaltenes performed on live single-phase
fluid is performed in laboratories with specialized equipment.
However, some analysis can be performed onsite, if required,
on stock tank fluids on site such as weight percent of
asphaltenes in stock tank oil by n-heptane precipitation or
Saturates, Aromatics, Resins and Asphaltene (SARA)
analysis.
The reversibility of asphaltene precipitation is
questionable. For this reason single-phase and exothermic
samplers were designed, so that the sample does not cross the
OAP prior to analysis.

Wax
Wax can be treated and controlled by maintaining the
temperature and pressure high. If any indication of wax issues
are encountered during WFT sampling the transfer of samples
to cylinders for further analysis becomes more difficult. The
solidified wax is a third phase and needs to be re-dissolved
into the fluid prior to transferring the sample into sample
cylinders. This is performed with a prolonged period of
heating at reservoir temperature and stirring (even as long as
12 hours for severe cases). If a mixing mechanism exists in the
sample chamber, this time can be shortened otherwise stirring
has to be done by gravity, lifting one end of the chamber for a
half hour, and then the other end. Patience is required.
SPE 101037 7
Sulphur Compounds
Sampling for analysis of the sulphur content present in the
reservoir fluid is not easy and still developing. The key issue is
that sulphur compounds such as H
2
S, mercaptans (RSH) and
carbonyl sulphide (COS) are adsorbed by the walls of
cylinders or sample chambers, particularly those made of
stainless-steel
16
. Alternate materials are recommended to
preserve the content of the sulphur compounds, this is
particularly important if analysis is to be done in a laboratory
as the time required for the sample to reach the laboratory (be
it only a week) is enough for the analysis to measure greatly
reduced H
2
S contents.
This problem is solved as best possible with current
technology by using cylinders made from other materials such
as NP-35N, Monel or Inconnel. Further to this coatings have
been made to reduce the reaction of conventional stainless
steel cylinders with H
2
S. These tend to be either PTFE based
or the latest generation are silicon based such as Silicosteel or
Sulfinert
16
.
When taking WFT samples, adsorption often occurs in the
flow lines into the sample chambers and, most of all at the
pumping unit.
The error margin due to adsorbed sulphur is less if there is
a higher sulphur content in the sample. This is because, even
though a larger volume is adsorbed, it is a smaller percentage.
Some analysis can also be performed on-site. Again, an
issue still remains as to whether the sulphur compounds are
adhering to the walls of the tubing, casing and surface
equipment prior to being sampled for analysis. Typical
analysis for sulphur compounds are as follows
16
:
Potentiometric titration
Iodometric titration (highly accurate but time
consuming)
Tutwiler method (higher concentrations)
Gas Chromatography
Gas detection tubes such as Drger, Kitagawa or
GasTec tubes.

Further to this, downhole sensors now exist to at least give
an indication of whether surface measured sulphur compound
concentrations are wrong or in the ball park. The future of
measurement of concentrations of sulphur compounds is likely
to be in improvement of this technology.

Mercury Content
The mercury content in the well-stream is also important to
measure to determine whether extra requirements are
necessary in the refining process. Mercury can cause metallic
parts to fail by amalgamating and changing the mechanical
properties of the metal, particularly aluminium or copper.
Similarly to sulphur compounds, the measurable mercury
content decreases with time as it is adsorbed by the walls of
stainless steel cylinders and sample chambers. to the walls of
the cylinder. On-site analysis can be performed on the
separator gas stream. Some of the mercury compounds that
may naturally occur in a reservoir fluid are Hg, RHgCl,
RRHg.
The mercury compounds are removed by passing the gas
through a permanganate solution. The solution is analised in a
lab by atomic absorption spectroscopy (AAS). This can also
be performed on stock tank liquid/condensates. The samples
should be kept in glass vials with minimal head space for
transportation to the lab.
The mercury content in gas can also be measured using a
Gold Leaf type device. These machines suck a volume of
gas through the sample probe which consists of a film of gold
which adsorbs the mercury.

Oil Based Mud
For WFT sampling, the use of Oil Based Muds (OBM) can
contaminate samples. If this is the case for oil wells,
depending on the severity of the contamination, the
composition can be back calculated if the OBM composition is
known. This is done by determining the composition of the
contaminated oil sample, then a sample with a known added
volume of OBM, and the full OBM composition. Based on
this the OBM components can be mathematically removed
from the contaminated oil composition
13
.
With gas-condensates a much smaller contamination will
have more sever effects. Not only in terms of the composition
but measured dew points can be far too high. However, the
same principle as for oils can be applied to gases.

Case Studies
The following section describes case studies where sampling
was performed in a non-conventional manner and the reasons
behind it.

Case 1
A low permeability saturated gas-condensate reservoir
(<1mD), was initially attempted to be separator sampled.
However, as the flow rate was through separator and CGR of
the well stream was low, there was insufficient condensate to
collect a sample. The separator pressure was lowered, however
there was still not enough condensate to collect a sample.
Split-phase sampling was economically unviable, therefore, it
was decided to collect bottom-hole samples. Saturated gas
samples are rarely taken bottom-hole, however in this case it
was the most viable option. Iso-kinetic sampling would have
proved better in this case, however the economics of the field
could not justify iso-kinetic sampling.

Case 2
A bottom-hole gas sample was collected which had too high a
saturation pressure (above reservoir pressure). The samples
were taken after conditioning with the well shut in as any flow
was creating too high a drawdown. A static pressure gradient
was run to ensure that there were no liquids in the well bore
prior to sampling. It was concluded that during conditioning
the reservoir fluid had been drawn into two-phases and the
well had not been shut in long enough to re-establish
equilibrium. As a result is was likely that condensate mist was
taken as well as the reservoir gas sample. Compositional
analysis was performed by equilibrating the sample in a PVT
cell at the reservoir pressure and temperature, allowing the
surplus condensate to drop out, and flashing only the gas.

8 SPE 101037
Case 3
An unexpected zone was encountered during drilling. Drilling
was ceased until it was determined whether it was a separate
reservoir that had been discovered or if the targeted zone was
at a shallower depth than expected. Bottom hole samples were
taken ASAP only to measure API gravity. In this case, well
conditioning, contamination with drilling mud and asphaltenes
were not an issue for the sampling program. Conventional
bottom-hole samples were taken as they are quick and easy to
run and re-run. The samples, although not representative PVT
samples, did suit the purpose of the sampling job.

Case 4
A volatile oil well from a tight formation was sampled at the
separator. Drawdown from the DST caused the bottom-hole
flowing pressure to be well below the saturation pressure. The
GOR at the test separator was observed to be moving
continuously, first a dip and then increasing gradually, but not
stabilizing. As the increase was slow, separator gas and oil
samples were taken, however recombining at the separator
GOR would have given too high a saturation pressure so using
an Equation of State package the GOR was back calculated
based on a saturation pressure. This method although not
perfect will still yield representative samples.


Conclusions
Sample as early as possible in the life of the reservoir.
Generally speaking, the earlier you sample the higher
the likelihood of obtaining representative samples.
Well conditioning is of utmost importance. Ensure
the well is conditioned as best possible prior to
sampling. The preparation of a good conditioning
program will improve sample quality.
Keep your WFT samples under pressure and transfer
them into a cylinder. It is common for these samples
to be released to stock tank conditions and only some
basic analyses performed. However, advanced
technology in this area has greatly improved the
quality of these samples which should be considered
representative until proven not so, rather than the
opposite. The chance to collect bottom hole or
separator samples later may not come. However, bear
in mind that WFT samples still require a drawdown
and suffer problems from drilling fluid issues and
may not be representative. But dont dump them until
some analysis has been done.
When in doubt take both surface and bottom-hole
samples.
If no asphaltene or solids present you may be able to
use a cheaper alternative to single-phase sampling
If asphaltenes may be a problem, take single phase
samples.
More research needs to be done on temperature
dependence of asphaltenes and this be translated into
better design of exothermic samplers. Since single-
phase sampling has advanced, comparative studies
should be performed on conventional samples versus
single-phase samples for different reservoir and
asphaltene types.
For difficult situations, a good hindsight quality
check is to compare the laboratory or on-site live
reservoir fluid density with the one calculated from
the pressure gradients during MDT logging. The new
developments in the field are excellent advances and
should be used to improve sampling. However, the
problem should not be over complicated and complex
solutions should be used only when necessary.

Acknowledgments
The authors would like to thank their affiliated companies,
Australian School of Petroleum in the University of Adelaide,
Santos Ltd. and Petrolab Pty. Ltd. for providing their support
knowledge and time.

References
1. Reudelhuber, F.O.: Sampling Procedures for Oil reservoir
Fluids paper SPE 816-G, Journal of Petroleum
Technology, (December 1957) 15
2. Fevang, and Whitson, C.H.: Accurate In-Situ
Compositions in Petroleum Reservoirs, paper SPE 28829
prepared for presentation at the European Petroleum
Conference, London, U.K., 25-27 October 1994
3. McCain Jr., W.D. and Alexander R.A.: Sampling Gas-
Condensate Wells paper SPE 19729, SPE Reservoir
Engineering, (August 1992) 358
4. El-Banbi, A.H. and McCain Jr., W.D.: Sampling Volatile
Oil Wells paper SPE 67232 was prepared for presentation
at the SPE Production and Operations Symposium held in
Oklahoma City, Oklahoma, 2427 March 2001
5. Cobenas, R.H. and Marcelo A. Crotti, M.A.: Volatile Oil.
Determination of Reservoir Fluid Composition From a
Non-Representative Fluid Sample paper SPE 54005
prepared for presentation at the 1999 SPE Latin American
and Caribbean Petroleum Engineering Conference held in
Caracas, Venezuela, 2123 April 1999
6. Strong, J., Thomas, F.B. and Bennion, D.B.: Reservoir
Fluid Sampling and Recombination Techniques for
Laboratory Experiments paper No. CIM 93-54 prepared
for presentation at the CIM 1993 Annual Technical
Conference in Calgary, May 9-12, 1993
7. Towler, B.F., Reservoir Engineering Aspects of Bottom
Hole Sampling of Saturated Oils for PVT Analysis paper
SPE 19438, SPE 1990
8. Dake, L.P.: The Practice of Reservoir Engineering,
Elsevier Science B.V., Amsterdam, (1994) 33
9. Jamaluddin, A.K.M., Creek, J., Kabir, C.S., McFadden,
J.D., DCruz, D., Joseph, M.T., Joshi, N. and Ross, B.: A
Comparison of Various Laboratory Techniques to Measure
Thermodynamic Asphaltene Instability, paper SPE 72154
prepared for presentation at the SPE Asia Pacific Improved
Oil Recovery Conference, Kuala Lumpur, Malaysia, 8-9
October 2001
10. Williams, J.M., Getting the Best Out of Fluid Samples
paper SPE 29227, Journal of Petroleum Technology
(September 1994) 752
11. Amyx, J.W., Bass, D.M. and Whiting, R.L.: Petroleum
Reservoir Engineering, McGraw-Hill Book Company ,
New York, (1960) 360
12. Riley, M. D., Walters, R. P., Cramer, S. D. and McCawley,
F. X., Isokinetic Technique for Sampling Geothermal
Fluid In Two-Phase Flow SPE 7885
SPE 101037 9
13. Dybdahl, B.: A Systematic Approach to Sampling During
Well Testing, paper prepared for presentation at the ENI
E&P Division and ENI Technology Conference, 23
February 2006
14. Michaels, J., Moody, M., and Shwe, T.: Wireline Fluid
Sampling paper SPE 30610 prepared for the SPE Annual
Technical Conference & Exhibition, Dallas, Texas, USA,
22-25 October 1995
15. Jamaluddin, A.K.M., Ross, B. and Daigle, T.: Single-
phase Reservoir Sampling: Is it a Necessity or Luxury?
paper CIM 99-36 to be presented at the 1999 CSPG and
Petroleum Society Joint Convention, Digging Deeper,
Finding a Better Bottom Line, in Calgary, Alberta, Canada,
June 14 18, 1999
16. Elshahawi, H. and Hashem M.: Accurate Measurement of
the Hydrogen Sulfide Content in Formation Fluid Samples
Case Studies paper SPE 94707 prepared for presentation
at the 2005 SPE Annual Technical Conference and
Exhibition held in Dallas, Texas, U.S.A., 9-12 October
2005

SI Metric Conversion Factors
bbl x 1.589 874 E-01 = m
3

ft
3
x 2.831 685 E-02 = m
3

scf/stb x 5.6146 E+00 = m
3
/m
3

psi x 6.894 757 E+00 = kPa

*Conversion factors are exact