Saturation Monitoring With the

RST Reservoir Saturation Tool

Bob Adolph
Christian Stoller
Houston, Texas, USA
Jerry Brady
ARCO Alaska, Inc.
Anchorage, Alaska, USA
Charles Flaum
Montrouge, France
Chuck Melcher
Brad Roscoe
Ridgefield, Connecticut, USA
Amal Vittachi
DeWayne Schnorr
Anchorage, Alaska
For help in preparation of this article, thanks to Darwin
Ellis and Jeff Schweitzer, Schlumberger-Doll Research,
Ridgefield, Connecticut, USA; and Mohamed Watfa, Abu
Dhabi, U.A.E.
In this article, ELAN (Elemental Log Analysis), CNL
(Compensated Neutron Log), Gradiomanometer, RST
(Reservoir Saturation Tool), GST (Induced Gamma Ray
Spectrometry Tool), Dual-Burst and TDT (Thermal Decay
Time) are marks of Schlumberger. Macintosh is a mark of
Apple Computer, Inc. VAX is a mark of Digital Equip-
ment Corporation.
1. Steinman DK, Adolph RA, Mahdavi M, Marienbach E,
Preeg WE and Wraight PD: “Dual-Burst Thermal
Decay Time Logging Principles,” paper SPE 15437,
presented at the 61st SPE Annual Technical Confer-
ence and Exhibition, New Orleans, Louisiana, USA,
October 5-8, 1986.
2. Woodhouse R and Kerr SA: “The Evaluation of Oil
Saturation Through Casing Using Carbon-Oxygen
Logs,” paper SPE 17610, presented at the SPE Interna-
tional Meeting on Petroleum Engineering, Tianjin,
China, November 1-4, 1988.
The RST Reservoir Saturation Tool combines the logging capabilities of tra-
ditional methods for evaluating saturation in a tool slim enough to pass
through tubing. Now saturation measurements can be made without killing
the well to pull tubing and regardless of the well’s salinity.
Determining hydrocarbon and water satura-
tions behind casing plays a major role in
reservoir management. Saturation measure-
ments over time are useful for tracking reser-
voir depletion, planning workover and
enhanced recovery strategies, and diagnos-
ing production problems such as water
influx and injection water breakthrough.
Traditional methods of evaluating satura-
tion—thermal decay time logging1 and car-
bon/oxygen (C/O) logging2—are limited to
high-salinity and nontubing wells, respec-
tively. The RST Reservoir Saturation Tool
overcomes these limitations by combining
both methods in a tool slim enough to fit
through tubing.3 The RST tool eliminates
the need for killing the well and pulling tub-
ing. This saves money, avoids reinvasion of
perforated intervals, and allows the well to
be observed under operating conditions
(next page ). Moreover, it provides a log of
the borehole oil fraction, or oil holdup,
even in horizontal wells. To understand the
operation and versatility of the RST tool
requires an overview of existing saturation
measurements and their physics.4
3. Audah T and Chardac J-L: “Reservoir Fluid Monitoring
Using Through-Tubing Carbon/Oxygen Tools,” Trans-
actions of the SPWLA 34th Annual Logging Sympo-
sium, Calgary, Alberta, Canada, June 13-16, 1993,
paper LL.
Stoller C, Scott HD, Plasek RE, Lucas AJ and Adolph
RA: “Field Tests of a Slim Carbon/Oxygen Tool for
Reservoir Saturation Monitoring,” paper SPE 25375,
presented at the SPE Asia Pacific Oil and Gas Confer-
ence and Exhibition, Singapore, February 8-10, 1993.
Scott HD, Stoller C, Roscoe BA, Plasek RE and Adolph
RA: “A New Compensated Through-Tubing
Carbon/Oxygen Tool for Use in Flowing Wells,”
Transactions of the SPWLA 32nd Annual Logging
Symposium, Midland, Texas, USA, June 16-19, 1991,
paper MM.
The Saturation Blues
In a newly drilled well, openhole resistivity
logs are used to determine water and hydro-
carbon saturations. But once the hole is
cased, saturation monitoring has to rely on
tools such as the TDT Dual-Burst Thermal
Decay Time tool or, for C/O logging, the
GST Induced Gamma Ray Spectrometry
Tool, which can “see” through casing.
The Dual-Burst TDT tool looks at the rate
of thermal neutron absorption, described by
the capture cross section Σ of the formation,
to infer water saturation (terms in bold are
explained in “Gamma Ray Spectrometry at
a Glance,” page 38 ). A high absorption rate
indicates saline water, which contains chlo-
rine, a very efficient, abundant thermal-neu-
tron absorber. A low absorption rate indi-
cates fresh water or hydrocarbon.
The TDT technique provides good satura-
tion measurements when formation water
salinity is high, constant and known. But oil
production from an increasing number of
reservoirs is now maintained by water injec-
tion. This reduces or alters formation water
salinity, posing a problem for the TDT tool.
Roscoe BA, Stoller C, Adolph RA, Boutemy Y, Cheese-
borough JC III, Hall JS, McKeon DC, Pittman D, See-
man B and Thomas SR: “A New Through-Tubing Oil-
Saturation Measurement System,” paper SPE 21413,
presented at the SPE Middle East Oil Show, Bahrain,
November 16-19, 1991.
4. For more on nuclear logging:
Ellis DV: Well Logging for Earth Scientists. New York,
New York, USA: Elsevier, 1987.
Tittman J: Geophysical Well Logging. Orlando,
Florida, USA: Academic Press, Inc., 1986.
Ellis D, Grau J, Schweitzer J and Hertzog R: “Basics of
Nuclear Logging,” Oilfield Review 35, no. 3 (July
1987): 4-10.

January 1994 29
 Near Carbon/ Far Carbon/ Volumetric In low-salinity water (less than 35,000 parts
Oxygen Ratio Oxygen Ratio Analysis per million), the tool cannot accurately dif-
ferentiate between oil and water, which
Depth, ft

Borehole Oil Sw Water

have similar neutron capture cross sections.
Shut-In Shut-In Holdup Shut-In Shut-In Oil When the salinity of the formation water
Borehole Oil Sw Limestone is too low or unknown, C/O logging can be
Flowing Flowing Holdup Flowing Flowing
0 1.0 0.1 0.6 -20 120 100 0 Dolomite used. C/O logging measures gamma rays
emitted from inelastic neutron scattering to
determine relative concentrations of carbon
X700
and oxygen in the formation. A high C/O
ratio indicates oil-bearing formations; a low
C/O ratio indicates water- or gas-bearing
formations (next page, top).
The major drawback to C/O logging tools
has been their large diameters. Producing
wells must be killed and production tubing
removed to accommodate tools with diame-
ters of nearly 4 in. [10 cm]. In addition, the
tools have slow logging speeds and are
more sensitive to borehole fluid than forma-
tion fluid, which affects the precision of the
X800
saturation measurement.

As Easy as RST
The RST tool directly addresses these short-
comings and can perform either C/O or
TDT logging (see “Logging the RST Tool in
Prudhoe Bay,” page 32 ). It comes in two
diameters—111/16 in. (RST-A) and 21/2 in.
(RST-B)—and can be combined with other
production logging tools ( next page,
bottom ). The RST-A tool logs up to four
X900
times faster than the GST tool. The RST-B
nComparing RST logs run in the same Middle East well during shut-in (red) and flowing tool—the only C/O tool that can log flowing
(blue) conditions. Production is from a vertical, 6-in. diameter, openhole completion. wells—makes passes at speeds comparable
Tracks 1 and 2 show the carbon/oxygen logging ratio curves for both the near and to the GST tool.
far detectors. Tracks 3 and 4 show the RST interpretation (borehole oil holdup and Both versions have two gamma ray detec-
water saturation logs). Track 5 is the volumetric analysis. tors. In the RST-A tool, both detectors are on
Both the near and far C/O ratios show a sharp increase at X848 ft, indicating an oil-
water interface in the borehole. Above 850 ft, the C/O ratios from both detectors the tool axis, separated by neutron and
increase steadily, showing the depths at which the oil is produced. gamma ray shielding. In the RST-B tool, the
The borehole oil holdup during flowing indicates that most of the oil is produced from detectors are offset from the tool axis and
the interval X728 to X750. The water saturation curves separate from X770 to X850 ft, shielded to enhance the near detector’s
indicating that oil from the borehole reinvaded the formation while the well was shut-
in. After the shut-in period, when the well flowed, the formation water flushed out the borehole sensitivity and the far detector’s
oil, as shown by the increased water saturation and verified by the flowing borehole oil formation sensitivity. This allows the forma-
holdup. The volumetric analysis from openhole ELAN Elemental Log Analysis interpre- tion oil saturation and borehole oil holdup
tation indicates substantial oil saturation in the upper half of the reservoir. to be derived from the same RST-B C/O
measurement. Because of size constraints,
such detector shielding is not possible with
the RST-A tool. An independent determi-
nation of borehole fluid holdup is then
needed, for example from the Gradioma-
nometer tool run on the same logging suite
or by logging shut-in.
For both tools, the detector crystal is
cerium-doped gadolinium oxyorthosilicate
(GSO), one of a new generation of scintilla-
tion crystals that outperforms the sodium-
(continued on page 34)

30 Oilfield Review
a Depth, ft

X150

X200

X250
aa 50
Open Hole

Connate Water

Movable Oil

Residual Oil

Pore Volume
p.u.

RST-A Sonde
111/16 in.
0

nThe dual-detector RST-A and RST-B tools.

January 1994
C/O Near
Ratio

C/O Far
Ratio
Carbon/Oxygen

Borehole Oil Holdup

0 -20
%

Side View

Electronics
0.5

120 50

Photomultiplier
tube

GSO detector
(far)
Electronics

Photomultiplier
tube
GSO detector
(near)
Shielding

Neutron
generator

Top View

Far

Near
Mixed Water

Remaining Oil

Pore Volume
p.u. 0 30

RST-B Sonde
2 1/2 in.
Carbon/Oxygen and ∑

∑ From RST
c.u.
Connate Water

Injection Water

10 50
Remaining Oil

Pore Volume
p.u. 0
nCombining the diagnostic capabilities
of C/O logging and TDT logging with the
RST-B tool in a Middle East observation
well. From left to right: Track 1 shows
openhole fluid analysis. Track 2 shows
C/O logs from the near and far detectors
and oil holdup. Track 3 shows the fluid
analysis based on C/O logging. Track 4
shows the log of Σ. Track 5 shows a com-
bined C/O and Σ fluid analysis.
The observation well contained water in
the borehole over the entire interval, so
the oil holdup value was set to 0. The C/O
fluid analysis, used to distinguish between
oil and water, shows only a small oil
depletion from X190 ft to X250 ft and a
large oil depletion from X140 ft to X170 ft.
By itself, however, C/O logging cannot
differentiate between injection and con-
nate water. To accomplish this, a log of Σ
is used in conjunction with the C/O logs.
The Σ log can differentiate between oil
and salt water but not oil and fresh water.
The revised interpretation indicates for-
mation water from X170 to X210 ft and a
water injection breakthrough from X140 to
X170 ft. Interpretation based on Σ alone
would have identified the water injection
breakthrough as oil.

31

Prudhoe Bay, the largest oil field in North Amer-
ica, contained 20 billion stock tank barrels when
it was discovered in 1968.1 Most of the hydrocar-
bons are in the Ivishak Sandstone. The main
recovery methods used in Prudhoe Bay are grav-
ity drainage, gas injection, aquifer influx, water-
flooding and miscible gas flooding. Over time,
the expansion of gas above and water below the
oil column has produced mobile oil lenses that
are elusive to tap. The salinity of injected water
is low, creating conditions suited to C/O logging
with the RST tool.2
As the Prudhoe Bay field matures, increasing
gas and water production is exceeding the capac-
ity of surface handling facilities, thus limiting oil
production. No pipeline exists for the gas, so it
must be reinjected. Wells with declining oil pro-
duction and increasing gas and water production
are typically worked over or produced intermit-
tently. Extending the life and economic viability
of these marginal wells relies on reducing gas
and water production and maximizing oil produc-
tion by producing bypassed oil zones.
RST data were used to reduce water cut in a
BP well with a 250-ft [76-m] oil-bearing sand-
stone interval (right). Openhole logs from 1983
(not shown) marked the original oil-water contact
1. Shirzadi SG and Lawal AS: “Multidisciplinary Approach for Tar-
geting New Wells in the Prudhoe Bay Field,” paper SPE 26093,
presented at the SPE Western Regional Meeting, Anchorage,
Alaska, USA, May 26-28, 1993.
2. Vittachi A and Schnorr D: “Reservoir Oil Saturation Monitoring
Through Casing With New Carbon/Oxygen Measurements,”
Transactions of the SPWLA 34th Annual Logging Symposium,
Calgary, Alberta, Canada, June 13-16, 1993, paper HHH.
Dupree JH and Cunningham AB: “The Application of
Carbon/Oxygen Logging Technology to the Ivishak Sandstone,
Prudhoe Bay, Alaska,” paper SPE 19615, presented at the 64th
SPE Annual Technical Conference and Exhibition, San Antonio,
Texas, USA, October 8-11, 1989.
32
Logging the RST Tool in Prudhoe Bay
Connate Water
(OWC) at X1250 ft and the original gas-oil contact
Borehole Oil (GOC) at X1050 ft.
Holdup Flowing Injection Water
and Gas By 1992, production rates fell, indicating
50 p.u. 0
Gas severe reservoir depletion. The well was produc-
Depth, ft Oil ing 3750 B/D of fluids with a 94% water cut. Pro-
Gamma Ray
Porosity duction was 220 BOPD and the gas/oil ratio
0 GAPI 100 40 p.u. 0
(GOR) was 1115 ft3 of gas per barrel of oil. The
X1000
RST tool was run to evaluate hydrocarbon distri-
bution and locate fluid contacts.
C/O data measured with the well flowing were
combined with CNL Compensated Neutron Log
data as input to the ELAN Elemental Log Analysis
program (right track). The resulting fluid analysis
confirmed oil depletion over both perforated
intervals and identified the current GOC at X1110
X1100
ft. The borehole oil holdup log showed oil pro-
GOC
duction from perforations above X1170 ft and was
used to identify the present OWC at that depth.
The lower perforations were not producing any
oil. After these were plugged, the well produced
Perfs
OWC 300 BOPD of oil with negligible water cut.
Locating Bypassed Oil
X1200
In early 1992, ARCO drilled and perforated a
Perfs
sidetrack well in an area of Prudhoe Bay under-
going waterflooding. Less than six months later,
production was 90% water with less than 200
BOPD, as expected. The original perforations
extended from X415 to X440 ft (next page). C/O
nReducing water cut. The left track shows the gamma
ray and borehole oil holdup logs measured with the well logging measurements were made In the shut-in
producing. The borehole oil holdup curve shows no oil well with three different tools—the RST tool and
production below 1170 ft.
The right track is an ELAN Elemental Log Analysis
two sondes from other service companies.
output based on openhole and RST fluid data. It indi- The RST results confirmed depletion over the
cates oil depletion below 1170 ft. After the lower perfo- perforated interval (Tracks 2 and 3). Effects of the
rated interval was plugged, production increased from
220 BOPD oil with 94% water cut to 300 BOPD oil with miscible gas flood sweep are apparent through-
negligible water cut. out the reservoir. The total inelastic count rate
ratio of the near and far detectors indicates quali-
tatively the presence of gas in the reservoir. In
Oilfield Review
 Openhole
Data
CRRA
Connate Water Sand
4.0 2.6 Gas
Depth, ft

Gas Bound Water

Saturation
Oil Oil Shale
Gamma Ray Saturation
Porosity Lithology
0 GAPI 150
0 1 40 p.u. 0 0 10

X100
GOC nFinding bypassed oil
zones. ARCO ran three
C/O logging tools with the
well shut-in over the same
interval in search of
bypassed oil. Track 1
includes the gamma ray,
count-rate ratio between
near and far detectors
(CRRA) and the borehole
oil holdup curves. Track 2
X200 shows the RST oil satura-
tion and gas saturation
B curves. Track 3 shows the
RST fluid analysis. Track 4
is openhole log data.
The RST data identified
Zone A as the best zone
for perforating. Interpreta-
tion based on logs from a
A non-RST C/O logging tool
identified Zone B for per-
X300 foration, which the RST
data show to be mostly
water. Zone B was perfo-
rated first, producing 200
BOPD with more than
95% water cut in a matter
of weeks as a result of
water coning into the per-
foration. When Zone A
was perforated, oil pro-
duction increased to an
X400
average of 600 BOPD with
a reduced water cut of
Perfs

90%. All water production

was identified as from
Zone B.

addition, differences between the openhole fluid
analysis and the RST fluid analysis were
assumed to be gas.
One potential bypassed zone, A, was identified
from X280 to X290 ft. A second zone, B, based on
the openhole logs and a C/O log from another ser-
vice company, was proposed from X220 to X230
ft. The RST log shows Zone B to contain more gas
and water than Zone A.
After assessing the openhole logs and the
three C/O logs, ARCO decided to perforate Zone
B. The initial production was 1000 BOPD with a
75% water cut. Production declined to 200 BOPD
with more than 95% water cut in a matter of
weeks. The decline prompted ARCO to perforate
Zone A, commingling production from earlier per-
forations. Production increased to an average of
600 BOPD and the water cut decreased to 90%.
Subsequent production logs confirm that Zone A
is producing oil and gas and Zone B is producing
all of the water with some oil.

January 1994 33
iodide [NaI] crystal conventionally used for
gamma ray detection (see “New Scintilla-
tion Detectors,” page 35 ). Several properties
of GSO allow for a smaller diameter detec-
tor crystal than if NaI were used, and hence
a smaller diameter tool.
Modes of Operation
Flexibility is a key advantage of the RST
tool. It operates in three modes that can be
changed in real time while logging:
•inelastic-capture mode
•capture-sigma mode
•sigma mode.
Inelastic-capture mode: The inelastic-cap-
ture mode offers C/O measurements for
determining saturations when the formation
water salinity is unknown, varying or too
low for TDT logging. In addition to C/O log-
ging, thermal neutron capture gamma-ray
spectra are recorded after the neutron burst.
Elemental yields from these spectra provide
lithology, porosity and apparent water salin-
ity information.
In the inelastic-capture mode, each mea-
surement cycle contains one neutron burst
and three timing gates for collecting spectra
( right ). The first gate measures the total
gamma ray spectrum during the neutron
burst, which contains both inelastic and
capture spectra. The second gate measures
an early capture spectrum following the
neutron burst, which is used to subtract the
capture background from the previous
inelastic spectrum, yielding the net inelastic
spectrum. The C/O ratio from the net inelas-
tic spectrum is used to determine saturation.
The third gate measures a capture spectrum
after the neutron burst, which is used to
determine formation lithology. For example,
the ratio of silicon to calcium is used to
distinguish silicates from carbonates. Log-
ging passes are made at 60 to 100 ft/hr [18
to 30 m/hr].
Capture-sigma mode: The capture-sigma
mode is used to determine lithology and the
capture cross section Σ in the same logging
pass. It simultaneously records capture
gamma ray spectra and total capture gamma
ray count rates. Elemental yields from the
capture spectra can provide lithology,
porosity and apparent water salinity infor-
34
100
80
Counts
C
60
Time, µsec
Counts
40
Net inelastic
B
20
Neutron burst
Counts
Energy
A
0
Energy
nTiming for the inelastic-capture mode of the RST tool. Timing gate A
records inelastic spectra during the neutron burst. Timing gates B and
C record capture spectra after the burst. The net inelastic spectrum is
formed by subtracting a fraction of the early capture spectrum mea-
sured in gate B from the spectrum measured in gate A.
mation as in the inelastic-capture mode. Interpretation
Total count rate measurements are used to The conversion of C/O ratios to saturations
determine Σ for the formation and the bore- relies on an extensive data base designed to
hole. Unlike the inelastic-capture mode, measure RST tool response to a variety of
each measurement cycle in the capture- borehole environments. The RST-A and
sigma mode contains two neutron pulses RST-B tools have been logged in a wide
—a short one and a long one. Total count variety of conditions at Schlumberger’s
rates collected during and after the short Environmental Effects Calibration Facility in
burst are used to determine the borehole Houston, Texas, USA. For each combina-
fluid Σ; total count rates collected after the tion of lithology, casing and cement at the
longer burst are used to determine the for- facility, four C/O ratio measurements were
mation Σ. Logging is usually performed at made to cover the four combinations of oil
600 ft/hr [180 m/hr]. and water in the borehole and formation:
Sigma mode: The sigma mode is used water-water, water-oil, oil-water and oil-oil.
when the salinity of the formation water is The resulting measurements represent the
high enough for TDT logging. It provides largest C/O characterization data base for
capture cross-section data in a fast pass—up any nuclear tool in the oil field today. Simi-
to 1800 ft/hr [550 m/hr]. Although the tim- larly, the largest data base of capture-sigma
ing sequence mimics the capture-sigma measurements was also acquired. Different
mode, only decay-time data and a pulse
height spectrum, for calibrating the tool
gain, are recorded.
Oilfield Review
 New Scintillation Detectors

Running the RST tool through tubing would not be possible without a scintillation crystal called cerium-doped
gadolinium oxyorthosilicate (GSO), one of several new crystals finding its way to the oil patch. Developing
these crystals for borehole detectors requires coupling a knowledge of nuclear physics with sophisticated
crystal-growing techniques.

Comparison of Scintillation Crystals

Crystal Density Effective Refractive Relative Energy Light Flash Rugged Unaffected Dewar
g/cm3 Atomic Number Z Index Light Flash Resolution Decay Time2 By Moisture System
Intensity @662keV1, % ns Required
NaI(Tl) 3.67 51 1.85 100 6.5 230 No No No
BGO 7.13 75 2.15 15 9.3 300 Yes Yes Yes
GSO 6.71 59 1.85 20 8.0 56 and 600 No Yes No
LSO 7.40 66 1.82 75 10 40 Yes Yes Depends on
application

1. Energy resolution refers to the commonly used relative “full width at half maximum” for the 137Cesium gamma ray peak at 662 keV.
Sharply defined spectral peaks indicate good energy resolution.
2. Decay time is the time constant for the exponential decay of the light flash. Faster decay times are preferred.

What Makes a Great Crystal?
Scintillation detectors are so named because
they generate flashes of light when struck by
gamma rays (for details on how scintillator crys-
tals work, see “Gamma Ray Spectrometry at a
Glance,” page 38).1 Most of the properties that
make a crystal desirable for logging are those
that maximize the intensity of these flashes and
the number of counts.2 In addition, crystals
should be rugged, not cracking on impact, and
unaffected by moisture.
A crystal of high density and high atomic num-
ber provides more opportunity for interaction with
incident gamma rays, maximizing count rates.
Crystal volume also affects performance. A big-
ger crystal intercepts more gamma rays, thereby
increasing count rates. It also decreases the
probability that a gamma ray will escape the
crystal after one or two scatterings.
A high light-flash intensity produces large volt-
ages that are easily measured, improving energy
resolution, and makes the detector more sensi-
tive to low-energy gamma rays. If at all possible,
the light-flash intensity and duration should be
relatively unaffected by temperature change. This
eliminates the need for temperature control hard-
ware downhole.
C/O logging and thermal decay time logging
expose the crystal to high instantaneous gamma
ray fluxes during or immediately after the neutron
burst. The light flashes must have a short enough
duration to avoid “piling up” on one another,
which produces false pulse amplitudes and
counting rates.
Once a light flash has been generated, the
goal is to get as much light as possible to the
photomultiplier tube (PMT). To prevent internal
absorption of light, the crystal should be trans-
parent to the light it generates. Because light
generated by the crystal must eventually pass
through the window of the photomultiplier tube,
the index of refraction of the crystal and window
should be similar to maximize light transmission.
The photocathode of the PMT must be highly
responsive to the wavelength of the crystal
light to maximize the number of photoelectrons
it ejects.
The New Detectors
For nearly 45 years, thallium-doped sodium
iodide (NaI) has been the gamma ray detector of
choice for nuclear logging. It is widely used to
determine formation density and chemical com-
position, salinity through casing, and mineralogy.
NaI excels in intensity of light flash and tempera-
ture stability. Although new crystals literally do
not outshine NaI, they do bring increased detec-
tion efficiency, greater ruggedness, reduced sen-
sitivity to humidity, and the ability to handle
much higher counting rates without pileup (see
“Comparison of Scintillation Crystals,” above).
Bismuth germanate (BGO) was first manufac-
tured commercially in the early 1970s, providing
higher counting rates than NaI that result in bet-
1. For a discussion of scintillation spectrometry:
Knoll GF: Radiation Detection and Measurement. New York,
New York, USA: John Wiley and Sons, 1989.
Birks JB and Fry D: The Theory and Practice of Scintillation
Counting. London, England: Pergamon Press, 1964.
2. Melcher CL, Schweitzer JS, Manente RA and Peterson CA:
“Applications of Single Crystals in Oil Well Logging,” Journal of
Crystal Growth 109 (1991): 37-42.

January 1994 35

ter precision (right). Its use in the Hostile Natural
Gamma Ray Sonde (HNGS) offers the choice of
faster logging or improved statistical precision in
estimates of thorium, uranium and potassium.
BGO’s high detection efficiency is also used in a
33/8-in. neutron-induced spectrometry tool.3 In
general, higher count rates and improved spec-
tral peak-to-Compton ratios more than compen-
sate for the main limitation of BGO, low intrinsic
resolution. GSO and BGO crystals generally give
better precision than similarly sized NaI crystals
in estimating element concentrations from neu-
tron-induced capture gamma ray spectra. The
poor temperature response of BGO, however,
requires the use of a Dewar flask. BGO can also
experience pileup problems.
The RST tool is the first commercial logging
tool to use the GSO crystal, which was developed
by the Hitachi Chemical Co., Ltd. in the early
1980s.4 The high density and high atomic number
of GSO make for a smaller diameter detector than
NaI, and hence a smaller diameter tool (top).5
The high density also improves the crystal’s
ability to detect gamma rays, particularly high-
energy ones.
GSO has a detection efficiency nearly as high
as BGO’s but can operate at higher temperatures
without a bulky Dewar flask. It produces signifi-
cantly less pulse pileup during the neutron burst.
Because the light output from GSO is lower than
for NaI, the RST tool employs a newly designed,
sensitive PMT.
A new member of the scintillation detector
family was developed two year ago by scientists
at Schlumberger-Doll Research, Ridgefield, Con-
necticut, USA. They found that cerium-doped
36
nComparison of a
GSO crystal used in
the RST tool and
the larger NaI crys-
tal used in the GST
tool. GSO’s higher
density and higher
atomic number
allow it to function
in the small diame-
ter RST tool.
BGO vs. NaI
Uncertainties BGO NaI
Th 2.82 ppm 4.63 ppm
U 1.51 ppm 2.61 ppm
K 0.51% 0.74%
Th = 12 ppm, U = 6 ppm, K = 2%
nQuantifying BGO’s statistical advantage.
Uncertainties for thorium (Th), uranium
(U) and potassium (K) outputs are pre-
sented for 1-sec accumulations in a stan-
dard shale. Standard deviations of Th, U
and K concentration estimates drop by 30
to 40 % when a BGO detector is used
instead of a NaI detector.
lutetium oxyorthosilicate (LSO), a cousin of GSO,
promises to be the best scintillator for many
applications.6 Although still in development, this
scintillator combines high counting efficiency
with a light-flash intensity nearly that of NaI and
five times that of BGO. It is unaffected by mois-
ture, is rugged and can handle higher counting
rates free of pileup than other scintillators under
consideration for borehole use. The light gener-
ated by an LSO crystal, similar in wavelength
to that of NaI, works well with existing photo-
cathode materials.
Competing crystals are breaking the near
monopoly held by NaI for over four decades,
though the price is high. The pair of GSO scintil-
lators in the RST tool costs about $10,000 at pre-
sent. But an expanding market is likely to result
in improved production methods and reduced
prices. These new scintillators are sure to make
further inroads on NaI’s territory. Crystal research
and development are continuing and new crystals
with special properties may be added to this
expanding list of exotic detectors.—JT, TAL
combinations of formation and borehole
salinities were used to generate over 1000 Σ
measurements for each tool.
The C/O data base is used to calibrate the
C/O ratio (COR) model to a particular log-
ging environment. The COR model relates
measured C/O ratios to the porosity, oil sat-
uration, lithology, oil density, oil holdup,
borehole diameter and casing weight and
diameter. Once calibrated, the model is
inverted to solve for fluid volumes,5 and
these values are further refined using alpha
processing, described below. The oil vol-
umes are then converted to saturations.
Alpha Processing: Windows
Versus Yields
Schlumberger’s traditional approach to pro-
cessing C/O logging data, the C/O yields
approach, is to determine the carbon and
oxygen yields from the measured inelastic
spectrum, and then use their ratio to deter-
mine oil volume, then saturation. The
C/Oyields approach provides good accuracy
but reduced precision.6 Advantages of this
method are ease of use and interpretation.
Because the contribution of carbon to the
total spectrum is relatively small, however,
good statistics require long measurement
times or slow logging speeds. To shorten
measurement times without sacrificing pre-
cision, a “windows” approach has been
adopted. Windows are sections of the spec-
trum most influenced by changes in carbon
and oxygen and least influenced by
changes in other elements (next page, left ).
The ratio of the gamma ray counts from the
“carbon window” to counts in the ”oxygen
window” is then used to determine a more
precise saturation.
The C/Owindows approach provides good
precision but poor accuracy. It improves the
statistics of the measurements because the
total counts in the windows are high. But
interpretation is more difficult. For example,
in a clean, water-bearing sand and a water-
filled borehole, the C/Oyields ratio would be
zero, whereas the C/Owindows ratio is non-
zero. The ratio produced by the windows
method must be calibrated with ratios mea-
3. Jacobson LA, Beals R, Wyatt, DF Jr and Hrametz A: “Response
Characterization of an Induced Gamma Spectrometry Tool
Using a Bismuth Germanate Scintillator,” Transactions of the
SPWLA 32nd Annual Logging Symposium, Midland, Texas,
USA, June 16-19, 1991, paper LL.
4. Takagi K and Fukazawa T: “Cerium-activated GD2SiO5 Single
Crystal Scintillator,” Applied Physics Letters 42 (1983): 43-45.
5. Melcher CL, Schweitzer JS, Utsu R and Akiyama S: “Scintillation
Properties of GSO,” IEEE Transactions of Nuclear Science 37,
no. 2 (April 1990): 161-164.
6. Melcher CL and Schweitzer JS: “A Promising New Scintillator:
Cerium-Doped Lutetium Oxyorthosilicate,” Nuclear Instruments
and Methods in Physics Research A 314 (1992): 212-214.
Oilfield Review
sured under known conditions. Ideally, the
windows-based C/O tool would be logged
first in a known water-bearing zone to deter-
mine a zero carbon value, then logged in a
zone of known oil saturation to obtain a
second calibration point.
Alpha processing for the RST tool com-
bines the accuracy of the yields ratio with the
precision of the windows ratio to obtain satu-
ration results in the minimum time. It calcu-
lates the volume of oil (the product of poros-
ity and oil saturation) from the C/Owindows
method—the windows oil volume, and the
volume of oil from the C/Oyields method—the
yields oil volume. The difference between
these volumes is tracked and used to adjust
the windows oil volume. Alpha processing is
applied over a specified number of levels,
known as the alpha filter length.

nThe RST Job Planner software for planning RST jobs on the Macintosh. The field engi-
neer or client specifies the many input parameters, including RST tool type, lithology,
saturation precision and alpha processing level, to calculate corresponding logging
speeds with the RST tool. CDV stands for oil gravity. Logging speeds are reported for
two levels of confidence, 68% and 95%. The use of alpha processing increases logging
Carbon Oxygen speeds by improving measurement statistics.
Carbon
Counts

Planning the RST Job Future of the RST Tool

Oxygen
0 2 4 6
Energy, MeV
nInelastic spectra measured with the
far detector of the RST tool in an oil
and water tank. Shaded areas indicate
the carbon and oxygen windows used
to improve measurement statistics.
With nuclear tools like the RST tool, careful
job planning can significantly affect the pre-
cision—and the cost—of the final answer.
To facilitate planning an RST job, a software
program called RST Job Planner is used to
determine the logging speed and number of
passes needed to produce results according
8 to clients’ specifications. Versions of RST Job
Planner can run on VAX computers at the
field logging interpretation center (FLIC) or
on a Macintosh computer. The Macintosh
version is available to clients.
RST Job Planner requires a variety of input
parameters, including tool type, lithology,
porosity, oil gravity (referred to as carbon
density value—CDV), casing size and
weight, borehole diameter, number of levels
for alpha processing and number of levels of
vertical averaging. In addition, clients spec-
ify the precision, or standard deviation of
the answer, in terms of a percent saturation.
The output consists of station times and
logging speeds in a table showing two levels
of confidence,7 68% and 95% (above ). This
allows the client to weigh precision needs
against logging costs. Higher precision
requires slower logging speeds. The fastest
logging option, however, may not always be
the most cost-effective for the client. While
the RST-B tool typically logs at less than half
the rate of the RST-A, it can minimize the
time a well is off-line and avoid problems
associated with reinvasion from shut-in.
With the RST tool’s accuracy, precision and
versatility for saturation monitoring
unmatched in the oil field, applications for
the tool continue to grow. The RST tool has
recently provided holdup measurements in
horizontal and through-tubing wells. A
porosity measurement determined from a
ratio of the count rates in the near and far
detectors will soon be added to the sigma
mode. Software is being developed that will
convert between RST and TDT logs,
enhancing the value of RST logs in Σ-moni-
toring programs. In addition, the RST tool’s
response to gas is being studied. These
ongoing projects, plus experimental use of
the RST tool in uncharacterized environ-
ments by innovative field engineers and
clients, promise an exciting future for this
newest saturation monitoring device. —TAL
5. Oil volume is the volume of formation taken up by oil.
It can be expressed as porosity times oil saturation.
6. Accuracy refers to how close the mean value of a
measurement is to the true value. Consider a
speedometer that always registers 10 miles per hour
(mph). It is inaccurate when the car is traveling 60
mph but accurate when the car travels 10 mph.
Precision refers to the spread of individual measure-
ments from the mean. A precise system produces
results with very little spread. The broken speedometer
is very precise, since it always gives the same mea-
surement of 10 mph.
7. A precision of ± one standard deviation corresponds
to 68% confidence; ± two standard deviations corre-
spond to 95% confidence and ± three standard devia-
tions correspond to 99.9% confidence.

January 1994 37
 Gamma Ray Spectrometry at a Glance

The interactions and measurements associated with gamma ray spectrometry may be confusing to nonexperts in nuclear logging.
Simply put, gamma ray spectrometry measures gamma rays (counts) in time and gamma ray energies. Some elements emit
gamma rays naturally; others can be bombarded with neutrons to induce gamma ray emissions. Each element produces
characteristic gamma rays of specific energies. Moreover, the number of characteristic gamma rays produced is proportional to
the abundance of the element. Naturally occurring and induced gamma rays may be counted and sorted according to energy.
This produces a gamma ray spectrum that can be processed, or decoded, to identify the elements and their concentrations.

Downhole Accelerator Inelastic Neutron Scattering In inelastic neutron

n 3
Ion source scattering, the neutron
bounces off the nucleus,
but excites it into quickly
giving off what are
n n
called inelastic gamma
Filament Magnet Target rays. The measurement
of gamma ray energies
The starting point for induced gamma ray from inelastic neutron
1 spectrometry is the generation of neutrons. Like scattering yields the
its predecessors, the RST tool uses a downhole relative concentrations
accelerator to emit pulses of high-energy of carbon and oxygen,
neutrons into the formation. This device creates which are then used to
14-million electron volt (MeV) neutrons by determine water
Inelastic saturation.
accelerating deuterium ions into a tritium target. gamma rays

Elastic Neutron Scattering

Neutron Absorption

Slow Excited
neutron nucleus

The neutrons primarily interact with formation
2 nuclei in three ways:
In elastic neutron-scattering, the neutron
bounces off the bombarded nucleus without
exciting or destablizing it. With each elastic
interaction, the neutron loses energy. Hydrogen,
with the mass of its nucleus equal to that of a
4 excited, typically emitting capture gamma rays. Neutron absorption, or
neutron, is very good at slowing down neutrons.
Hence, how efficiently a formation slows down
neutrons generally indicates the abundance of
hydrogen. Because hydrogen is most abundant in
pore fluids, neutron slowdown indicates porosity.
n
Capture
gamma ray
In neutron absorption, the nucleus absorbs the neutron and becomes
neutron capture, is most common after a neutron has been slowed by
elastic and inelastic interactions to thermal energies of about 0.025 eV.
The measurement of capture gamma ray energies is used to estimate the
abundances of elements most likely to capture a neutron—silicon, calcium,
chlorine, hydrogen, sulfur, iron, titanium and gadolinium.

38 Oilfield Review
 Photomultiplier Tube

Gamma
ray

Scintillation Window
material
Photoelectron
Photocathode
surface
Short burst Long burst
Counts

Collector
0 500 1000 1500
Time, µsec

A decrease in the production rate of capture
5 gamma rays over time, as measured with the
TDT tool, is proportional to the absorption rate of
thermal neutrons. This decay is basically
exponential and accounts for both the absorption
and diffusion of neutrons in the formation. The
slope of a semilog plot of gamma ray counts
versus time yields the capture cross section Σ of
the formation. Expressed in capture units, the
capture cross section expresses the probability
that a neutron passing through a cross-sectional
area of a material will be captured.
A gamma ray is detected when its interactions with the detector crystal create
7 electrons and holes that excite the crystal into generating flashes of light
(scintillations). The intensity of these scintillations is related to the energy of
the bombarding gamma ray. Most gamma rays lose some of their energy on
the way to the detector because they scatter randomly, create new particles
or disappear.
The light flashes pass through a window at one end of the crystal and
fall on the photocathode surface of a photomultiplier tube, liberating
electrons via the photoelectric effect. The tube amplifies the electronic
charges about 200,000 times and provides a current signal large enough to
be analyzed by downhole electronics.

Oxygen
Silicon
Tool background
Calcium
Iron
Counts

Carbon

0 20 100
Time, µsec
Neutron Neutron
burst burst

Inelastic Early Late Inelastic 1 2 3 4 5 6 7 8

capture capture Energy, MeV

A pulsed neutron source and “time-gated” detectors
6 solve the difficulty of distinguishing gamma rays of
nearly the same energy from two different interactions.
Inelastic gamma rays occur during the neutron pulse;
capture gamma rays occur later, after the neutrons have
been slowed down. A typical timing cycle takes
advantage of this by scheduling the first measurement
gate during the neutron pulse and subsequent
measurement gates at later times. Which interactions,
and hence, which elements contribute to the spectra,
depend largely on the sequence and duration of pulsed
neutron bursts and timing gates.
As with any other nuclear logging tool, the gamma ray spectra collected by
8 the RST tool must be processed to identify the elements that contributed to
the spectra and their abundances. This is possible because each element
produces a characteristic set of gamma ray peaks.
A library of standard elemental spectra is used to determine the
individual elemental contributions. Each standard elemental spectrum
represents the response of a tool to a particular element. A computer
determines the linear combination of these elemental standards that best
fits the measured spectrum and calculates the elemental yields. A yield is
the fractional contribution of an element to the total observed spectrum,
after subtracting off the capture background. It is proportional to the
abundance of the element.
39