Chain of Custody for Samples of Live Crude Oil Using Visible–
Near-Infrared Spectroscopy
SORAYA S. BETANCOURT, JEP BRACEY, GALE GUSTAVSON, SYRIAC G. MATHEWS, and
OLIVER C. MULLINS*
Schlumberger-Doll Research, Ridgefield, Connecticut 06877 (S.S.B., G.G., O.C.M.); BHP Billiton, Houston, Texas 77056 (J.B.); and
Schlumberger Oilphase-DBR, Houston, Texas 77084 (S.G.M.)
In order to design oil production facilities and strategies, it is necessary to
acquire crude oil samples from subsurface formations in oil wells in so-
called openhole prior to production. In some environments, such as
deepwater production of oil, decisions of huge economic importance are
based on such samples. To date, there has been little quality control to
verify that the crude oils collected in the sample bottles and analyzed up to
a year later in the laboratory have any relation to the actual crude oils in
the subsurface reservoirs. These high-pressure samples can undergo
myriad deleterious alterations. Here, we introduce the chain-of-custody
concept to the oilfield. The visible–near-infrared spectrum of the crude oil
is measured in situ in the wellbore at the point of sample acquisition. This
spectrum is compared with the spectrum measured on putatively the same
fluid in the laboratory at the start of laboratory sample analysis. First,
quantitative assessment is made of whether the fluid in the (high-pressure)
sample bottle remains representative of formation fluids. Second, any
specific changes in the spectrum of the fluid can be related to possible
process control failures. Here, the entire process of chain of custody is
proven. The chain of custody process can rapidly become routine in the
petroleum industry, thereby significantly improving the reliability of any
process that depends on fluid property determination.
Index Headings: Chain of custody; Crude oils; Live crude oils; Ultraviolet–
visible spectroscopy; UV-visible spectroscopy; Near-infrared spectrosco-
py; Gas–oil ratio; Crude oil coloration; High pressure; High temperature.
INTRODUCTION
The properties of the crude oils in an oilfield dictate many of
the critical production parameters of this field.1 For example,
the gas–oil ratio (GOR) of a crude oil determines the
corresponding volumetric ratio in gas and liquid processing.
The GOR of a reservoir crude oil (or live crude oil) is defined
to be the volumetric ratio of gas to liquid under conditions of
one atmosphere and 60 8F. Phase transition pressures, such as
the bubble point, the dew point, or the asphaltene onset
pressure, determine the limits of pressure reduction, thereby
establishing tight constraints on volumetric flow rates and thus
cash flow. Flow assurance issues associated with deposits of
solids in pipelines determine essential mitigation strategies.2 In
a broader context, variations in fluid properties are routinely
used to identify key elements of reservoir architecture such as
the existence of reservoir compartmentalization, i.e., the size of
the producing flow units. This issue remains one of the biggest
problems in oil production in settings such as deepwater, where
history matching of production with reservoir modeling can be
performed only after enormous investment in production
facilities. Recently, Downhole Fluid Analysis (DFA) has been
established as a cost-effective means of identifying fluid
Received 17 August 2006; accepted 22 September 2006.
* Author to whom correspondence should be sent. E-mail: mullins1@slb.
com.
0003-7028/06/6012-1482$2.00/0
1482 Volume 60, Number 12, 2006 Ó 2006 Society for Applied Spectroscopy
property variations in subsurface formations.3 Identification of
compartmentalization is one of the key benefits of DFA.4
In the exploration and appraisal phase of an oilfield, when no
production facilities are yet in place, the requisite crude oil
samples are acquired in so-called ‘‘openhole’’ just after the
formation of interest is drilled and prior to setting any steel
casing in the borehole section. The openhole wellbore pressure
is maintained with drilling fluids weighted with high-density
solid suspensions. To acquire crude oil samples in the
subsurface environment, a sampling tool is lowered into the
well on a cable or ‘‘wireline’’.5 The sampling tool has a probe
that establishes hydraulic communication with the fluid
contents of a permeable zone, thereby enabling fluid removal
from this zone.5 The telescoping probe has a stout steel tube
that is forced against the borehole wall with great force to
establish a hydraulic seal.5 A pressure drop in the flow line
causes formation fluids to flow into the sampling tool.
Reservoir crude oils often contain appreciable dissolved gas;
thus, the pressure drop used to initiate flow could induce a
phase transition. If a phase transition takes place, then the high-
mobility phase flows preferentially. Consequently, if phase
transitions are induced during sampling, it is likely any
acquired sample would be nonrepresentative of the formation
fluid and thus would be invalid. Optical methods can be used to
monitor the flow stream in the downhole environment. Gas-
phase transitions can be detected by index of refraction
measurements.6 Retrograde dew formation is measured using
fluorescence.7 Asphaltene precipitation induced by a pressure
drop can be detected by optical light scattering8,9 in downhole
flow streams, although light scattering is not a unique signature
downhole as it can be in the laboratory.10 Emulsions, gas
evolution, clays, sand, and retrograde dew evolution can also
induce light scattering.
The GOR of a crude oil can be measured using visible–near-
infrared spectroscopy (Vis-NIR) downhole in the sampling tool
at the point of sample acquisition.11–14 The sample is then
stored in a high-pressure sample bottle downhole that is
brought to the well-site surface. The sample is transferred if
necessary to a different high-pressure sample bottle for
transportation to a laboratory for chemical analyses. The harsh
downhole environment and the nature of the operations make
the whole process of sample acquisition with laboratory
analysis susceptible to fouling and inaccuracies. Leaks from
high-pressure sample bottles, errors during transfer of the bottle
contents, and contamination with undesirable fluids could
invalidate the sample. The long delay times between sample
acquisition and analysis, up to one year, do not instill
confidence in this process. Thus, it is very desirable to
introduce the concept of chain of custody for oilfield samples.
Prior to the advent of DFA, there was no convenient means to
perform chain of custody. Now that DFA has become
APPLIED SPECTROSCOPY
established,3 the sample acquisition and analysis process
affords the opportunity to establish the chain of custody for
oilfield samples. A detailed comparison can now be made of
the Vis-NIR spectrum and the GOR determined downhole
versus that measured in the laboratory.
One can determine GOR downhole and in the lab for sample
validation. However, the laboratory measurement of GOR
involves flashing the live crude oil with its dissolved gases at
atmospheric pressure at room temperature in order to determine
the volumes of the liquid and gas phases. Downhole, this
process would be very difficult; thus, Vis-NIR spectroscopy is
used to obtain the GOR.3,9–14 Consequently, simply comparing
GORs determined downhole versus the lab is not preferred.
The Vis-NIR spectrum of a sample provides a fingerprint of the
sample; the Vis-NIR spectrum of a live crude oil acquired
downhole can be compared to that acquired in the lab. This
fingerprint comparison provides the best opportunity for chain
of custody. Furthermore, if differences are found, spectral
analysis may point to the specific problem. For example, lack
of the methane peak in the laboratory measurement of a live
crude oil may point to bottle leakage.
A complication in the downhole environment is that wells
are drilled with so-called drill muds. These muds, which have
either water or oil as the base fluid, provide hydrostatic
weighting to prevent blow outs, provide a means to lift drill
cuttings by mud circulation, lubricate drilling, and perform
other functions. Typically, the borehole is maintained at
pressures exceeding subsurface formation fluid pressures to
prevent the rapid, uncontrolled entry of formation fluids into
the borehole (which can lead to a blow out). Fluid loss of
drilling fluids to the formation is minimized by the addition of
clays and other solids to the mud (thus the term mud), in order
to build an impermeable mud cake at the borehole wall (of
permeable zones). It is essential to differentiate between crude
oil versus oil-based mud (OBM) filtrate in sample acquisition.
To perform this task, Vis-NIR spectra are recorded.3,12,13,15,16
OBM filtrates lack asphaltenes and other heavy ends that
substantially give rise to oil coloration via electronic transitions
of polycyclic aromatic hydrocarbons (PAHs). OBM filtrates
also lack much dissolved methane, and thus their NIR
signatures lack methane bands. Consequently, the Vis-NIR
spectral range is very useful downhole for monitoring sample
acquisition.
The object of this study is to establish and perform a
workflow for initiating the chain of custody for oilfield
samples. We compare the Vis-NIR spectral fingerprint of a
hydrocarbon sample measured downhole during sample
acquisition with the same measurement performed on osten-
sibly the same fluid in the laboratory. The downhole
spectrometer performs sample analysis at the location and
time when the crude oil sample is acquired,15 thus providing
the first direct measurement of the fluid properties when the
sample is at quasi-pristine conditions.12,13 In the laboratory, the
same measurement is done on the collected samples at the same
pressure and temperature as the downhole environment using a
research grade spectrometer. A procedure is described that
relates the spectra obtained with the downhole filter spectrom-
eter with the lab grating spectrometer. In addition to the direct
comparison of the optical spectra, other properties derived from
downhole spectral data such as the GOR and some limited
compositional analysis are compared to lab measurements. We
establish by example that this chain-of-custody process is
readily feasible. Given the recent growing focus on crude oil
production, it behooves the industry to adopt the best practices
that have become routine in other economic sectors.
EXPERIMENTAL
Downhole samples were acquired with a formation sampling
tool.5 The downhole spectrometer utilized is a custom filter
spectrometer commercialized under the name Live Fluid
Analyzer* (LFA)12–14 and is a module of the MDT* (modular
formation dynamics tester) formation sampling tool.5 The LFA
consists of a ten-channel Vis-NIR filter spectrometer and also
has a gas-detection cell based on an index of refraction
measurement. The downhole spectrometer must operate up to
temperatures of 150 8C and sample pressures can be 20 000 psi
or higher. In addition, the spectrometer must fit into a small
package in order to go into oil wells. These constraints are
demanding and preclude many spectral approaches utilized in
the laboratory.
Laboratory spectral measurements were performed using a
Cary 5 ultraviolet (UV)-visible-NIR spectrometer manufac-
tured by Varian Inc. A custom high-pressure optical cell was
built to hold the hydrocarbon samples under borehole
conditions. The custom optical cell, made by DBR, Edmonton,
Canada, has sapphire windows similar to the optics cell of the
downhole tool and has a path length of 2.6 mm. A custom fiber-
optics assembly was built to optically couple the Cary 5
spectrometer with the high-pressure cell.11,14 This optical cell is
placed inside an oven, where the temperature is adjusted to the
temperature of the downhole measurement for that particular
sample. A flow line connects the optical cell with the sample
bottle. The sample bottle has a floating piston separating the
hydrocarbon sample from hydraulic oil. The back side of the
bottle is connected in turn to a high-pressure hydraulic pump.
The cell and flow line dead volume is less than 5 cc.
Prior to sample transfer in the lab, samples are reconditioned
in the high-pressure bottles to reservoir pressure and
temperature with agitation for five days. Sample transfer is
accomplished with the setup schematically depicted in Fig. 1.
The flow line is filled with N2 at a pressure comparable to the
sample bottle pressure. N2 is an inert gas and is not NIR active.
Then, the nitrogen bottle valve is closed and the sample bottle
valve is opened, and the purge valve is opened slightly to
perform sample transfer. Sufficient sample is transferred to
purge the entire dead volume twice. Nitrogen was used as a
transfer fluid rather than evacuating the flow line in order to
avoid sample flashing and possible window motion.
The HPHT (high-pressure, high-temperature) cell must be
clean, properly assembled, and pressurized to 1000 psi to
ensure proper seating of the sapphire windows. Prior to each
run, the Vis-NIR spectrum (300–2500 nm) of a reference
hydrocarbon (spectroscopy grade n-heptane) was measured in a
standard 2 mm glass cuvette and in the HPHT sapphire cell at
room pressure and temperature. In both cases the instrument
baseline is set to be the particular empty cell (cuvette or HPHT
cell). The HPHT cell is placed in the oven and connected to the
fluid transfer lines that lead to the sample bottle and the ultra
purity nitrogen. The system is heated to the desired
temperature, filled with nitrogen, and the pressure is increased
to the sample pressure (which is set to the formation pressure).
The valve to the sample bottle is opened and fluid enters the
flow line and the cell. The optical spectrum is measured at
APPLIED SPECTROSCOPY 1483

FIG. 1. A schematic showing the laboratory apparatus for laboratory
measurement of Vis-NIR spectra of high-temperature, high-pressure crude oil
samples. High-pressure fluid transfer lines connect the sample bottle to the
high-pressure, high-temperature optics cell situated inside an oven. The optics
cell is also connected to a high-pressure bottle filled with nitrogen for sample
transfer purposes. The pressure of the high-pressure sample and nitrogen bottles
is controlled by a hydraulic pump. The optics cell is optically coupled to the
Cary 5 spectrometer via custom fiber-optics transfer optics.
various stages of the sample transfer process for monitoring
purposes.
Data Analysis. The processing of the spectral data is
performed using sequential steps in the following order: (1)
correction for reflective losses, (2) path length correction, (3)
simulation of filtered spectra from high resolution spectra, (4)
elimination of scattering in the downhole signal, and (5) GOR
calculation. Each step is described below.
(1) Reflective Losses. The baseline for optical measurements
is recorded with an empty sample holder (HPHT sapphire cell
or glass cuvette). The empty cell has a larger Fresnel loss than
with fluid-filled cells. These reflection or Fresnel losses at each
interface can be calculated to a first-order approximation using
the Fresnel formula for normal incidence:


n1
n2 2
RF ¼ ð1Þ
n1 þ n2
where n1 and n2 are the refractive indexes of the two materials
(e.g., sapphire and fluid A). The optical density (OD) is equal
to
log(1
2RF). For each sample cell there are two internal
interfaces (e.g., sapphire–fluid–sapphire) and the total reflec-
tion loss is to the zeroth-order 2 3 RF. For a single n-heptane (n
¼ 1.388) and quartz glass (n ¼ 1.544) interface, RF ¼ 0.0028 or
0.28% of the light reflected at each interface. For the two
sample cell interfaces 2 3 RF ¼ 0.0057, giving an OD of
0.0025. In contrast, with air in the cell, the zeroth-order
reflective losses in terms of I0 are 4.6% at the first interface and
4.4% at the second, giving a total loss of ;9.0%, and the
corresponding OD is 0.041. The difference between an air-
1484 Volume 60, Number 12, 2006
filled cell and an n-heptane filled cell is ;0.04. Thus, when
subtracting the air-filled spectrum, the baseline is negative for
the fluid-filled cell because of its index matching of the cell.
For n-heptane and sapphire (n ¼ 1.74) the reflection losses are
1.27% per interface, giving a reflective loss from the two
window–fluid interfaces of 0.011. For sapphire–air, the
reflective losses are ;7.3% per interface, giving an OD of
0.066, again yielding a negative baseline for fluid-filled cells.
The actual reflective losses in the spectra can be measured at
wavelengths where no absorption is expected. The largest
negative value of absorption for the n-heptane spectrum in the
1000–1300 nm range is used to define the reflective losses.
This measured value correlates reasonably to the calculated
value for the glass (quartz) cuvette. A typical measured
reflective loss for the HPHT cell (sapphire) is OD ¼
0.06.
Reflective losses do not scale with path length and must be
subtracted from the spectral data before doing path length
correction.
(2) Path Length Calculation. The downhole spectrometer
and the optical cell used in the laboratory have different path
lengths. For comparison purposes, the lab spectrum must be
converted to the path length of the downhole tool. Downhole
tools are calibrated by measuring the absorption spectrum of
J26 hydraulic oil at room pressure and temperature, and the
path length is derived by comparison with the spectrum
measured in a reference 2 mm path length cuvette in the
laboratory. After correcting both spectra for reflective losses,
we calculate the path length of the downhole tool as:
ðOD1725
OD1600 ÞLFA
hLFA ¼ 3 2:0 ð2Þ
ðOD1725
OD1600 Þ2mm
In the case of the lab measurements, it is necessary to calculate
the path length of the HPHT cell every time it is assembled as a
quality check. This is done by measuring the absorption
spectrum of n-heptane in a 2 mm cuvette and in the HPHT cell
at room temperature. It is necessary to seat the sapphire
windows with application of modest pressure before making
this measurement. Both spectra are corrected for reflective
losses and the cell path length is calculated as above. Finally,
the lab spectrum (without Fresnel losses) is multiplied by the
ratio hdownhole/hlab cell.
(3) Simulation of Filtered Spectra from High Resolution
Spectra. The downhole spectrometer is a filter spectrograph,
while the lab spectrometer is a high-resolution double-grating
spectrometer system. For the downhole spectrograph, each
detector has an associated bandwidth between 10–30 nm
depending on wavelength to account for spectral shifts.14 In
contrast the research-grade spectrometer in the laboratory
measures the absorption spectra at 1 nm intervals with a
spectral bandwidth of 2 nm.
To convert the lab spectrum to an equivalent downhole
spectrum, a Gaussian filter is applied to the lab data:
X
n=2
ODlab ðk þ iÞexpð
i2 =2Þ
ODLFA ðkÞ ¼
i¼
n=2
ð3Þ
X
n=2
expð
i2 =2Þ
i¼
n=2
where ODLFA* is the simulated downhole response at
wavelength k, with bandwidth n, and ODlab is the lab
measurement. Other filter functions could be used.
(4) Scattering Calculation. The final step before plotting the
downhole and lab spectra is to calculate the scattering in the
downhole signal. Often downhole data exhibit scattering as the
fluid being analyzed carries along mud and solid particles that
may form a coating on the spectrometer windows. Large
scattering cross-sections are associated with wavelength
independence; here we presume wavelength independent light
scattering. Some scattering can be wavelength dependent (more
difficult to extract from the signal) and is handled separately.16
After accounting for Fresnel losses and correcting for path
length, the following relation holds for samples that have been
preserved properly:
ODLFA ¼ ODLFA
Scattering ð4Þ
(5) GOR Calculation. The two-stretch overtone of the
various CH oscillators occurs at ;1.7 lm. The CH4 absorption
profile is quite distinct from the –CH2– spectral signature.11,14
More complicated measurement algorithms can be derived for
lighter hydrocarbons with their significant concentration of
non-methane hydrocarbon gases.14 The GOR of a crude oil can
be obtained by measuring the methane and methylene peak.
The GOR calculation is obtained as a variant on a first
principles algorithm.11–13 A central feature of this algorithm for
downhole application is that temperature and pressure effects
predominantly enter through mass density effects alone.17 The
field version of the algorithm is given by:12,13
OD1670
OD1600
ROD ¼ ð5Þ
OD1725
OD1600
mo 0:125
0:741 3 ROD
¼ ð6Þ
mg 0:59 3 ROD
1:177
mo
mg
GOR½scf=stb ¼ 8930 3 m ð7Þ
1
0:193 3 mo
g
The unit of GOR is standard cubic feet (scf) of gas divided by
the number of stock tank barrels (stb) of crude oil all at 60 8F
and at one atmosphere. Thus, GOR is a volumetric ratio. To
obtain GOR in cubic meters of gas divided by cubic meters of
oil, divide GOR in scf/stb by 5.6.
To perform this NIR spectral analysis of a crude oil, it is
often necessary to remove the spectral effect of electronic
transitions of PAHs. The NIR spectral range contains the
vibration peaks and a broad sloping electronic absorption
profile. This profile is largely invariant for all crude oils.18 For
colorless crude oils, the Urbach tail is in the UV;18 however,
for formation hydrocarbons with no aromatics, there would be
no Urbach tail. Because crude oils have such different
intensities of coloration, using a broad spectral range ensures
that the coloration of almost all crude oils will be captured.
This exponential decay of the optical density for crude oils is
reminiscent of the Urbach tail in semiconductor absorption
edges and is given by the equation below; Ep is the photon
energy and E0 is the decay constant. In heavy hydrocarbons
this behavior impacts considerably the GOR calculation and
thus must be accounted for (de-coloration) prior to calculating
the GOR.3,14 Optical channels with k , 1600 nm and OD ,
3.0 are fit to the following equation to obtain the coefficients C
and E0.


Ep
OD½k ¼ C exp ð8Þ
E0
Using Eq. 8, we decolorize the spectra, and then calculate the
GOR using the three channels k ¼ 1600, 1670, and 1725 nm.
RESULTS AND DISCUSSION
Two chain-of-custody examples are presented in this section.
Example #1 corresponds to a downhole sample acquired from
an offshore oil field. The absorption spectra measured in the lab
and downhole are in excellent agreement. Figure 2 shows the
comparison of (1) the downhole discrete spectral data, (2) the
lab continuous spectrum, and (3) the simulated discrete data
from the continuous lab spectrum. Both spectra were measured
at ;15 000 psi and 250 8F. The agreement is excellent. The
diagnostic analysis here indicates that this sample has been
well preserved and could be used for further analyses in the
laboratory. Spectral data with ODs larger than 1.5 are not good
due to light losses in the fiber connections and are not used for
comparison.
The comparison of downhole and lab data can be made
graphically as shown in Fig. 3. All channels show excellent
agreement and fall on the X ¼ Y line. Slight differences are
noted and give the magnitude of our error bar.
In the second example, the fluid sample was also acquired
from an offshore oil field and analyzed in situ with a downhole
spectrometer. In this case both spectra were measured at 20 000
psia and 200 8F. Figure 4 shows the comparison of the
spectrum acquired downhole with that of the simulated discrete
data obtained from the laboratory measurement, again on
ostensibly the same hydrocarbon fluid. Overall the agreement
is quite good. The coloration is nearly identical, showing that
the heavy end content is the same. The NIR bands are in close
agreement but do show some subtle differences.
Figure 5 plots the channel comparison for the downhole and
lab data. First, the methane channel (1671 nm) shows a
lowered absorption in the lab relative to downhole, while the
FIG. 2. Example #1. The spectra acquired downhole (discrete circles) are
compared with simulated discrete data (triangles) from the lab spectrum of
ostensibly the same fluid. Excellent agreement is obtained: the lab sample is good.
APPLIED SPECTROSCOPY 1485
FIG. 3. Example #1. Correlation of laboratory and downhole absorption
spectra for Fig. 2.
methylene peak at 1725 nm shows a slight increase in the lab.
Moreover, the slope obtained fitting all the points deviates from
a unit slope line, indicating loss of light components and a
darker fluid in the laboratory. The GOR computed from the
downhole spectrum (580 scf/stb) is indeed higher than the one
obtained with the lab spectrum (320 scf/stb). A fluid
comparison algorithm19 was used to analyze the two spectra
and the probability of similarity between these two samples.
This algorithm accounts for spectral and other error bars to
determine a statistical likelihood of fluid differences.19 Here,
the fluids are computed to be 93% likely to be different. The
conclusion is that this sample has been compromised. Caution
should be exercised if further lab analysis is performed on this
sample. It is likely there was a bottle leak with differential loss
of light ends.
FIG. 4. Example #2. The spectra acquired downhole (discrete circles) are
compared with simulated discrete data (triangles) from the lab spectrum of
ostensibly the same fluid. Overall, the agreement is good. However, some
subtle differences are noted in the NIR peaks.
1486 Volume 60, Number 12, 2006
FIG. 5. Example #2. Correlation of laboratory and downhole absorption
spectra for Fig. 4. Overall agreement is good, but some differences are noted.
CONCLUSION
Optical spectroscopy is a robust technique and is well suited
for chain of custody of live crude oil fluid samples. Vis-NIR
spectroscopy is used routinely in the acquisition of crude oil
samples in situ in oil wells, especially with the construct of the
recent breakthrough technology Downhole Fluid Analysis.
Comparison of Vis-NIR spectra obtained at the time and
location of sample acquisition versus spectra obtained in the
laboratory at much later times enables proof of sample
validation, i.e., chain of custody. This process is warranted
due to the known difficulties associated with sample handling,
storage at high pressures and temperatures, and long distances
between oilfields and laboratories, along with numerous
potentially deleterious hydrocarbon phase-transition and labile
components. The annual upstream two trillion dollar oil
business should adopt practices that are standard in other
settings in order to avoid expensive miscalculations.
1. L. Dake, Fundamentals of Reservoir Engineering (Elsevier Science, The
Netherlands, 1978).
2. O. C. Mullins, E. Y. Sheu, A. Hammami, and A. G. Marshall, Eds.,
Asphaltenes, Heavy Oils and Petroleomics (Springer, New York, 2006).
3. G. Fujisawa and O. C. Mullins, ‘‘Live fluid Sampling and Downhole Fluid
Analysis’’, in Asphaltenes, Heavy Oils and Petroleomics, O. C. Mullins, E.
Y. Sheu, A. Hammami, and A. G. Marshall, Eds. (Springer, New York,
2006), Chap. 22.
4. O. C. Mullins, G. Fujisawa, M. N. Hashem, and H. Elshahawi,
‘‘Determination of coarse and ultra-fine scale compartmentalization by
downhole fluid analysis,’’ International Petroleum Tech Conference SPE
Paper 10036 (2005).
5. T. Zimmerman, J. MacInnis, J. Hoppe, J. J. Pop, and T. Long,
‘‘Applications of Emerging Wireline Formation Testing Technologies’’,
paper OSEA 90105, Eighth Offshore South East Asia Conference
(Singapore, December 4–7, 1990).
6. O. C. Mullins, R. J. Schroeder, and P. Rabbito, Appl. Opt. 33, 7963 (1994).
7. S. S. Betancourt, G. Fujisawa, O. C. Mullins, K. O. Eriksen, C. Dong, J.
Pop, and A. Carnegie, ‘‘Exploration applications of downhole measure-
ment of crude oil composition and fluorescence,’’ SPE #87011, Asia, Pac.
Tech. Conf. (2003).
8. A. Hammami and H. C. Phelps, T. Monger-McClure, and T. M. Little,
Energy Fuels 14, 14 (2000).
9. N. B. Joshi, O. C. Mullins, A. Jamaluddin, J. L. Creek, and J. McFadden,
Energy Fuels 15, 979 (2001).
10. A. Hammami and J. Ratulowski, ‘‘Precipitation and Deposition of
 Asphaltenes in Production Systems: A Flow Assurance Overview’’, in
Asphaltenes, Heavy Oils and Petroleomics, O. C. Mullins, E. Y. Sheu, A.
Hammami, and A. G. Marshall, Eds. (Springer, New York, 2006), Chap.
23.
11. O. C. Mullins, T. Daigle, C. Crowell, H. Groenzin, and N. B. Joshi, Appl.
Spectrosc. 55, 197 (2001).
12. O. C. Mullins, G. Beck, M. Y. Cribbs, T. Terabayshi, and K. Kegasawa,
‘‘Downhole determination of GOR on single phase fluids by optical
spectroscopy,’’ SPWLA 42nd Annual Symposium, Paper M (Houston,
Texas, 2001).
13. C. Dong, O. C. Mullins, P. S. Hegeman, R. Teague, A. Kurkjian, and H.
Elshahawi, ‘‘In-situ contamination monitoring and GOR measurement of
formation samples,’’ SPE 77899, SPE Asia Pacific Meeting (Melbourne,
Australia, 2002).
14. G. Fujisawa, M. Van Agthoven, P. Rabbito, and O. C. Mullins, Appl.
Spectrosc. 56, 1615 (2002).
15. A. Crombie, F. Halford, G. Melbourne, M. N. Hashem, R. I. McNeil, E. C.
Thomas, and O. C. Mullins, Oilfield Rev. 10(3), 26 (1998).
16. O. C. Mullins and J. Schroer, ‘‘Real-time Determination of Filtrate
Contamination During Openhole Wireline Sampling by Optical Spectros-
copy’’, SPE Annual Technical Conference and Exhibition, Paper# 63071
(Dallas, TX, 2000).
17. O. C. Mullins, N. B. Joshi, H. Groenzin, T. Daigle, C. Crowell, M. T.
Joseph, and A. Jamaluddin, Appl. Spectrosc. 54, 624 (2000).
18. O. C. Mullins, S. Mitra-Kirtley, and Y. Zhu, Appl. Spectrosc. 46, 1405
(1992).
19. L. Venkataramanan, G. Fujisawa, O. C. Mullins, R. R. Vasques, and H.-P.
Valero, Appl. Spectrosc. 60, 653 (2006).
APPLIED SPECTROSCOPY 1487